Portal de Periódicos da CAPES

Analysis of the Quaternary Structure, Substrate Specificity, and Catalytic Mechanism of Valine Dehydrogenase

1997; Elsevier BV; Volume: 272; Issue: 40 Linguagem: Inglês

10.1074/jbc.272.40.25105

1083-351X

A.P. Turnbull, Patrick J. Baker, David W. Rice,

Amino Acid Enzymes and Metabolism

The solution of the three-dimensional structure of Bacillus sphaericus leucine dehydrogenase has enabled us to undertake a homology-based modeling exercise on the sequence differences between the families of leucine (LeuDH) and valine (ValDH) dehydrogenases. This analysis indicates that the secondary structure elements in the core of the two domains of a single subunit of these enzymes are conserved, as are residues directly implicated in the recognition of the nucleotide cofactor and in catalysis. Comparison of the sequences indicates that the residues in the pocket accommodating the side chain of the amino acid substrate are conserved between these two enzymes, suggesting that the small differences in specificity arise from minor changes in molecular structure, possibly associated with shifts of the main chain rather than mutation of residues in the pocket itself. While B. sphaericus LeuDH is an octamer, bothStreptomyces cinnamonensis and Streptomyces coelicolor ValDHs are dimers. The differences in quaternary structure can be understood in terms of the deletion in the latter of a C-terminal loop, which forms important interactions around the four-fold axis in LeuDH. The solution of the three-dimensional structure of Bacillus sphaericus leucine dehydrogenase has enabled us to undertake a homology-based modeling exercise on the sequence differences between the families of leucine (LeuDH) and valine (ValDH) dehydrogenases. This analysis indicates that the secondary structure elements in the core of the two domains of a single subunit of these enzymes are conserved, as are residues directly implicated in the recognition of the nucleotide cofactor and in catalysis. Comparison of the sequences indicates that the residues in the pocket accommodating the side chain of the amino acid substrate are conserved between these two enzymes, suggesting that the small differences in specificity arise from minor changes in molecular structure, possibly associated with shifts of the main chain rather than mutation of residues in the pocket itself. While B. sphaericus LeuDH is an octamer, bothStreptomyces cinnamonensis and Streptomyces coelicolor ValDHs are dimers. The differences in quaternary structure can be understood in terms of the deletion in the latter of a C-terminal loop, which forms important interactions around the four-fold axis in LeuDH. The oxidative deamination of amino acids to their corresponding keto acids is catalyzed by the family of amino acid dehydrogenases and provides a route for the incorporation of ammonia into organic compounds and links the metabolism of carbohydrates and amino acids. Sequence homology between glutamate (GluDH), 1The abbreviations used are: GluDH, glutamate dehydrogenase; LeuDH, leucine dehydrogenase; PheDH, phenylalanine dehydrogenase; ValDH, valine dehydrogenase. leucine (LeuDH), valine (ValDH), and phenylalanine (PheDH) dehydrogenases (1Britton K.L. Baker P.J. Engel P.C. Rice D.W. Stillman T.J. J. Mol. Biol. 1993; 234: 938-945Crossref PubMed Scopus (67) Google Scholar, 2Teller J.K. Smith R.J. McPherson M.J. Engel P.C. Guest J.R. Eur. J. Biochem. 1992; 286: 151-159Crossref Scopus (109) Google Scholar, 3Nagata S. Tanizawa K. Esaki N. Sakamoto Y. Ohshima T. Tanaka H. Soda K. Biochemistry. 1988; 27: 9056-9062Crossref PubMed Scopus (69) Google Scholar, 4Leiser A. Birch A. Robinson J.A. Gene ( Amst. ). 1996; 177: 217-222Crossref PubMed Scopus (19) Google Scholar, 5Takada H. Yoshimura T. Ohshima T. Esaki N. Soda K. J. Biochem. ( Tokyo ). 1991; 109: 371-376Crossref PubMed Scopus (61) Google Scholar) clearly indicates the existence of an enzyme superfamily related by divergent evolution (1Britton K.L. Baker P.J. Engel P.C. Rice D.W. Stillman T.J. J. Mol. Biol. 1993; 234: 938-945Crossref PubMed Scopus (67) Google Scholar). This enzyme family has considerable commercial potential for the production of novel non-proteogenic amino acids for the pharmaceutical industry (6Hanson R.L. Singh J. Kissick T.P. Patel R.N. Szarka L.J. Mueller R.H. Bioorg. Chem. 1990; 18: 116-130Crossref Scopus (47) Google Scholar, 7Patchett A.A. Harris E. Tristram E.W. Wyvratt M.J. Wu M.T. Taub D. Peterson E.R. Ikeler T.J. Broeke J. Payne L.G. Ondeyka D.L. Thorsett E.D. Greenlee W.J. Lohr N.S. Hoffsommer R.D. Johshua H. Ruyle W.V. Rothrock J.W. Aster S.D. Maylock A.L. Robinson F.M. Hirschmann R. Sweet C.S. Ulm E.H. Gross D.M. Vassil T.C. Stone C.A. Nature. 1980; 288: 280-283Crossref PubMed Scopus (742) Google Scholar) and for the diagnosis of genetic diseases of amino acid metabolism including phenylketonuria (8Dooley K.C. Clin. Biochem. 1992; 25: 271-275Crossref PubMed Scopus (24) Google Scholar), maple syrup urine disease (9Livesey G. Lund P. Methods Enzymol. 1988; 166: 3-10Crossref PubMed Scopus (18) Google Scholar, 10Yamaguchi A. Mizushima Y. Fukushi M. Shimizu Y. Kikuchi Y. Takasugi N. Screening. 1992; 1: 49-62Abstract Full Text PDF Scopus (25) Google Scholar), and homocystinuria (11Mudd S.H. Levy H.L. Skovby F. Sciver C.R. Beudet A.L. Sly W.S. Valle D. The Metabolic Basis of Inherited Disease. McGraw-Hill, New York1989: 693-734Google Scholar). A thorough understanding of the way in which members of this family achieve differential substrate specificity might not only enhance our understanding of the relationship between molecular structure and biological function, but may also have important industrial applications. ValDH (EC 1.4.1.8) catalyzes the reversible oxidative deamination ofl-valine to 2-ketoisovalerate, with the corresponding reduction of the cofactor NAD+ (Scheme 1).(CH3*2CHCHNH3+(valine)COO−+NAD++H2O↔(CH3*2CHCOCOO−2­ketoisovalerate)+NADH+NH4++H+SCHEME1This enzyme has been characterized in species ofStreptomyces, where it functions in the catabolism of branched chain amino acids and also plays an important role in providing precursors for the biosynthesis of polyketide antibiotics (12Omura S. Tanaka Y. Mamada H. Masuma R. J. Antibiot. 1983; 36: 1792-1794Crossref PubMed Scopus (35) Google Scholar). The ValDH genes from Streptomyces cinnamonensis (4Leiser A. Birch A. Robinson J.A. Gene ( Amst. ). 1996; 177: 217-222Crossref PubMed Scopus (19) Google Scholar),Streptomyces coelicolor (13Tang L. Hutchinson C.R. J. Bacteriol. 1993; 175: 4176-4185Crossref PubMed Google Scholar) and Streptomyces fradiae (14Tang L. Zhang Y.-X. Hutchinson C.R. J. Bacteriol. 1994; 176: 6107-6119Crossref PubMed Google Scholar) have been cloned and sequenced and contain 358, 364, and 371 amino acids, respectively, with subunitM r of approximately 38,000. Biochemical studies have established that the ValDHs from S. cinnamonensis andS. coelicolor are homodimers (15Priestley N.D. Robinson J.A. Biochem. J. 1989; 261: 853-861Crossref PubMed Scopus (24) Google Scholar, 16Navarrete R.M. Vara J.A. Hutchinson C.R. J. Gen. Microbiol. 1990; 136: 273-281Crossref PubMed Scopus (24) Google Scholar). In contrast, the common quaternary structure for GluDH is based on a hexamer and for LeuDH and PheDH an octamer. There are small differences in substrate specificity between ValDH and LeuDH. Members of the ValDH family (15Priestley N.D. Robinson J.A. Biochem. J. 1989; 261: 853-861Crossref PubMed Scopus (24) Google Scholar, 16Navarrete R.M. Vara J.A. Hutchinson C.R. J. Gen. Microbiol. 1990; 136: 273-281Crossref PubMed Scopus (24) Google Scholar, 17Vančurová I. Vančura A. Volc J. Neužil J. Flieger M. Basařová G. Bêhal V. J. Bacteriol. 1988; 170: 5192-5196Crossref PubMed Google Scholar, 18Vančura A. Vančurová I. Volc J. Fussey S.P.M. Flieger M. Neužil J. Maršálec J. Bêhal V. J. Gen. Microbiol. 1988; 134: 3213-3219PubMed Google Scholar, 19Ohshima T. Soda K. Biochim. Biophys. Acta. 1993; 1162: 221-226Crossref PubMed Scopus (8) Google Scholar) exhibit preferential substrate specificity toward valine; however, other hydrophobic branched chain amino acids are also accepted by these enzymes but with lower activity (Table I). For example, the k cat/K m values forS. cinnamonensis ValDH (15Priestley N.D. Robinson J.A. Biochem. J. 1989; 261: 853-861Crossref PubMed Scopus (24) Google Scholar) when valine and leucine are used as substrate are 21.8 and 3.0 mm−1s−1, respectively. In contrast, LeuDH (20Misono H. Sugihara K. Kuwamoto Y. Nagata S. Nagasaki S. Agric. Biol. Chem. 1990; 54: 1491-1498PubMed Google Scholar, 21Ohshima T. Misono H. Soda K. J. Biol. Chem. 1978; 253: 5719-5725Abstract Full Text PDF PubMed Google Scholar, 22Ohshima T. Nishida N. Bakthavatsalam S. Kataoka K. Takada H. Yoshimura T. Esaki N. Soda K. Eur. J. Biochem. 1994; 222: 305-312Crossref PubMed Scopus (57) Google Scholar, 23Nagata S. Bakthavatsalam S. Galkin A.G. Asada H. Sakai S. Esaki N. Soda K. Ohshima T. Nagasaki S. Misono H. Appl. Microbiol. Biotechnol. 1995; 44: 432-438Crossref PubMed Scopus (23) Google Scholar, 24Schütte H. Hummel W. Tsai H. Kula M.-R. Appl. Microbiol. Biotechnol. 1985; 22: 306-317Crossref Scopus (72) Google Scholar) favors the 1-carbon longer branched chain amino acid leucine, with, for example, the k cat/K m values forCorynebacterium pseudodiptheriticum LeuDH (20Misono H. Sugihara K. Kuwamoto Y. Nagata S. Nagasaki S. Agric. Biol. Chem. 1990; 54: 1491-1498PubMed Google Scholar) with valine and leucine as substrate being 73.3 and 101.7 mm−1 s−1, respectively (TableII). Using a structure-based sequence alignment, we have analyzed the similarities and differences between ValDH, LeuDH, and GluDH and present our findings below, arranged into sections that describe the similarity in amino acid sequence and its consequences for the tertiary structure, quaternary structure, catalytic mechanism, and substrate specificity of ValDH.Table ISubstrate specificities in the oxidative deamination direction for the ValDHs from S. cinnamonensis (15Priestley N.D. Robinson J.A. Biochem. J. 1989; 261: 853-861Crossref PubMed Scopus (24) Google Scholar), S. coelicolor (16Navarrete R.M. Vara J.A. Hutchinson C.R. J. Gen. Microbiol. 1990; 136: 273-281Crossref PubMed Scopus (24) Google Scholar), S. aureofaciens (17Vančurová I. Vančura A. Volc J. Neužil J. Flieger M. Basařová G. Bêhal V. J. Bacteriol. 1988; 170: 5192-5196Crossref PubMed Google Scholar), S. fradiae (18Vančura A. Vančurová I. Volc J. Fussey S.P.M. Flieger M. Neužil J. Maršálec J. Bêhal V. J. Gen. Microbiol. 1988; 134: 3213-3219PubMed Google Scholar), and A. faecalis (19Ohshima T. Soda K. Biochim. Biophys. Acta. 1993; 1162: 221-226Crossref PubMed Scopus (8) Google Scholar)SubstrateS. cinnamonensisS. coelicolorS. aureofaciensS. fradiaeA. faecalisK mk catk cat/K mRelative activity1-aRelative activity is the specific activity of the particular amino acid substrate relative to l-valine, expressed as a percentage.K mRelative activity1-aRelative activity is the specific activity of the particular amino acid substrate relative to l-valine, expressed as a percentage.K mRelative activity1-aRelative activity is the specific activity of the particular amino acid substrate relative to l-valine, expressed as a percentage.K mRelative activity1-aRelative activity is the specific activity of the particular amino acid substrate relative to l-valine, expressed as a percentage.K mmms −1mm −1 s −1%mm%mm%mm%mml-Valine1.32821.810010.01002.51001.01002.3l-Norvaline3.2185.6435.7981.344l-α-Aminobutyrate19.61045.31301714.8693.133l-Leucine4.011.93.08366.3250.773l-Isoleucine10.85.80.528475.0291.272l-Norleucine30.017.90.61115.6526.7161-a Relative activity is the specific activity of the particular amino acid substrate relative to l-valine, expressed as a percentage. Open table in a new tab Table IISubstrate specificities in the oxidative deamination direction for the LeuDHs from C. pseudodiptheriticum (20Misono H. Sugihara K. Kuwamoto Y. Nagata S. Nagasaki S. Agric. Biol. Chem. 1990; 54: 1491-1498PubMed Google Scholar), B. sphaericus (21Ohshima T. Misono H. Soda K. J. Biol. Chem. 1978; 253: 5719-5725Abstract Full Text PDF PubMed Google Scholar), T. intermedius (22Ohshima T. Nishida N. Bakthavatsalam S. Kataoka K. Takada H. Yoshimura T. Esaki N. Soda K. Eur. J. Biochem. 1994; 222: 305-312Crossref PubMed Scopus (57) Google Scholar), B. licheniformis (23Nagata S. Bakthavatsalam S. Galkin A.G. Asada H. Sakai S. Esaki N. Soda K. Ohshima T. Nagasaki S. Misono H. Appl. Microbiol. Biotechnol. 1995; 44: 432-438Crossref PubMed Scopus (23) Google Scholar), and B. cereus (24Schütte H. Hummel W. Tsai H. Kula M.-R. Appl. Microbiol. Biotechnol. 1985; 22: 306-317Crossref Scopus (72) Google Scholar)SubstrateC. pseudodiptheriticumB. sphaericusT. intermediusB. licheniformisB. cereusK mk catk cat/K mRelative activity2-aRelative activity is the specific activity of the particular amino acid substrate relative to l-leucine, expressed as a percentage.K mRelative activity2-aRelative activity is the specific activity of the particular amino acid substrate relative to l-leucine, expressed as a percentage.K mRelative activity2-aRelative activity is the specific activity of the particular amino acid substrate relative to l-leucine, expressed as a percentage.K mRelative activity2-aRelative activity is the specific activity of the particular amino acid substrate relative to l-leucine, expressed as a percentage.K mmms −1mm −1 s −1%mm%mm%mm%mml-Valine0.641.873.3741.7872.45912.5612.5l-Norvaline0.515.528.7413.527282.9l-α-Aminobutyrate2.77.52.81410.08322422.0l-Leucine0.330.5101.71001.01002.01002.11001.5l-Isoleucine0.812.515.4581.8890.4723.3611.0l-Norleucine2.84.21.5106.3761.52-a Relative activity is the specific activity of the particular amino acid substrate relative to l-leucine, expressed as a percentage. Open table in a new tab The close relationship between the sequences of ValDH and LeuDH and the more remote relationship with GluDH are illustrated in the DIAGON plots presented in Fig. 1.S. cinnamonensis ValDH exhibits 49% and 20% identities with Thermoactionomyces intermedius LeuDH andClostridium symbiosum GluDH over the 346 and 349 residues that can be aligned, respectively. Nevertheless, of the 68 residues that are strongly conserved in the GluDH family (25Britton K.L. Baker P.J. Rice D.W. Stillman T.J. Eur. J. Biochem. 1992; 209: 851-859Crossref PubMed Scopus (75) Google Scholar), 33 are also conserved in the ValDH family, indicating that all these enzymes are closely related. The crystal structures of C. symbiosum GluDH (26Baker P.J. Britton K.L. Engel P.C. Farrants G.W. Lilley K.S. Rice D.W. Stillman T.J. Proteins Struct. Funct. Genet. 1992; 12: 75-86Crossref PubMed Scopus (231) Google Scholar, 27Stillman T.J. Baker P.J. Britton K.L. Rice D.W. J. Mol. Biol. 1993; 234: 1131-1139Crossref PubMed Scopus (202) Google Scholar) and B. sphaericus LeuDH (28Baker P.J. Turnbull A.P. Sedelnikova S.E. Stillman T.J. Rice D.W. Structure. 1995; 3: 693-705Abstract Full Text Full Text PDF PubMed Scopus (86) Google Scholar) have been solved and can be seen to be closely related despite having a low sequence identity of 19% (28Baker P.J. Turnbull A.P. Sedelnikova S.E. Stillman T.J. Rice D.W. Structure. 1995; 3: 693-705Abstract Full Text Full Text PDF PubMed Scopus (86) Google Scholar). In both cases, the subunit structure is based on two domains separated by a deep cleft. Domain I is constructed primarily from residues in the N-terminal half of the polypeptide chain and is exclusively involved in subunit assembly. In contrast, Domain II comprises residues from the C-terminal half of the polypeptide chain and is responsible for nucleotide binding. The structure-based sequence alignment of representative members of the ValDH, LeuDH, and GluDH families is presented in Fig.2, and, unless otherwise stated, the sequence numbering and identification of the secondary structural elements in B. sphaericus LeuDH are used to identify equivalent residues in the ValDH sequences throughout this paper. With the exception of residues close to the N and C termini and a region of the polypeptide chain that includes α8, βi, and α9 (all of which lie on the periphery of the molecule), ValDH and LeuDH show considerable sequence similarity over the entire length of the polypeptide chain, implying that their tertiary structures will be almost identical (Fig. 3 a). Three sites of insertion and deletion are found between LeuDH and all ValDHs, including approximately 11 additional residues at the N terminus, 3 additional residues in the loop connecting α8 to βi, and the deletion of approximately 15 residues at the C terminus. Additionally, there are 5 extra residues in the βd–α3 loop and a 1-residue deletion in the α9–βj loop in S. fradiaeValDH. Previous studies have noted that insertions and deletions at these positions occur commonly within the wider enzyme superfamily (1Britton K.L. Baker P.J. Engel P.C. Rice D.W. Stillman T.J. J. Mol. Biol. 1993; 234: 938-945Crossref PubMed Scopus (67) Google Scholar).Figure 3a, a schematic diagram of a B. sphaericus LeuDH subunit, produced using MOLSCRIPT (37Kraulis P.J. J. Appl. Crystallogr. 1991; 24: 946-950Crossref Google Scholar), showing the helices and strands conserved in ValDH colored yellowand portrayed as labeled helical ribbons andarrows, respectively. The helices α8 and α9 and the strand βi in LeuDH, which show low sequence similarity with ValDH, are highlighted in green. Regions of insertions and deletions between LeuDH and ValDH are highlighted in red.b, a view from the 42 symmetry point down the two-fold axis, which relates the dimer, AB, with the four-fold axis vertical.c, close-up of view b. The secondary structural elements implicated in interactions across the two-fold axis relating dimers are labeled. d, a view from the 42 symmetry point down the two-fold axis, which relates the pairs of dimers, AB and A′B′, with the four-fold axis vertical. e, close-up of view d. The interaction of α14 and the C-terminal loop of subunit A (red) with the complementary U-shaped pocket formed by the symmetry-related protruding arm and the loops connecting βb to βc and α3 to βe of subunit B′ (yellow) and α3, α4, and the loop connecting βa to βb of subunit A′ (blue) about the two-fold axis can clearly be seen. The C-terminal residues that form the last turn of α14 and the loop to the C terminus in LeuDH and which are deleted in ValDH are highlighted by a dashed outline for subunits A and B′, respectively.View Large Image Figure ViewerDownload Hi-res image Download (PPT) Crystallographic studies have shown thatC. symbiosum GluDH assembles into a hexamer with 32 symmetry (26Baker P.J. Britton K.L. Engel P.C. Farrants G.W. Lilley K.S. Rice D.W. Stillman T.J. Proteins Struct. Funct. Genet. 1992; 12: 75-86Crossref PubMed Scopus (231) Google Scholar), this being the most common quaternary structure within this enzyme family. In contrast, the structure determination of B. sphaericus LeuDH has shown it to assemble into an octamer with 42 symmetry (Refs. 28Baker P.J. Turnbull A.P. Sedelnikova S.E. Stillman T.J. Rice D.W. Structure. 1995; 3: 693-705Abstract Full Text Full Text PDF PubMed Scopus (86) Google Scholar and 29Turnbull A.P. Ashford S.R. Baker P.J. Rice D.W. Rodgers F.H. Stillman T.J. Hanson R.L. J. Mol. Biol. 1994; 236: 663-665Crossref PubMed Scopus (11) Google Scholar; Fig. 3, b–e). Unfortunately, the complete sequence of the latter is not yet available and the structure is currently based on a partial amino acid sequence and a combination of a sequence determined by inspection of the electron density map and the aligned sequences of other species variants of the enzyme. In LeuDH, the dimer interface consists of three interacting complementary areas (Fig. 3 b). Centrally, strand βa from Domain I interacts with its symmetry-related mate across the two-fold axis to form an antiparallel sheet of 12 β strands spanning both subunits (Fig. 3 c). Another secondary structural element involved in interactions across the two-fold interface is the C-terminal end of α3, which packs against its two-fold related counterpart. The final interaction involves residues from one face of α1, which pack against βa, βb, βd, and α2 in the symmetry-related subunit. The sequence alignment strongly indicates that, with the exception of α1 for which the sequence similarity is low, each of the elements of secondary structure that are involved in these interactions are conserved in ValDH. Furthermore, of the 33 residues that take part in interactions across the two-fold axis in LeuDH, 15 residues are identical in at least six of the eight aligned LeuDH and ValDH sequences, implying that the character of this interface is maintained. Around the two-fold axis relating pairs of dimers (Fig. 3 d), an important feature of the four-fold interface in B. sphaericus LeuDH involves the interaction of a protruding arm, formed by residues at the end of α14 (residues 347–350) and the loop to the C terminus (residues 351–364) from subunit A of an AB dimer, with residues in a complementary U-shaped pocket constructed by the neighboring dimer, A′B′ (Fig. 3 e). One face of this pocket is built from the symmetry-related protruding arm of subunit B′, the base of the pocket is constructed from the loops connecting βb to βc and α3 to βe of the same subunit, and the other face is formed by the loops connecting βa to βb, the N-terminal end of α3, and the C-terminal end of α4 from subunit A′. Calculations of the accessible surface area using the algorithm of Lee and Richards (30Lee B. Richards F.M. J. Mol. Biol. 1971; 55: 379-400Crossref PubMed Scopus (5416) Google Scholar) have shown that the LeuDH monomer possesses an accessible surface area of 16,100 Å2 and on octamer formation 3,400 Å2 (21%) of this surface is buried, of which 1,250 Å2 is buried on dimer formation and the remainder on assembly of dimers to form the octamer (Fig.4). Analysis of the structure-based sequence alignment in Fig. 2 illustrates that the final 15 C-terminal residues in LeuDH are deleted in the ValDH sequences. In LeuDH, these residues contribute 45% (950 Å2 out of 2,150 Å2) of the surface that becomes buried on the assembly of dimers to form the octamer. Therefore in ValDH, the deletion of these residues implies that an octameric arrangement for this enzyme is unlikely, which is consistent with the proposed dimeric quaternary structures of the ValDHs from S. cinnamonensis and S. coelicolor (15Priestley N.D. Robinson J.A. Biochem. J. 1989; 261: 853-861Crossref PubMed Scopus (24) Google Scholar, 16Navarrete R.M. Vara J.A. Hutchinson C.R. J. Gen. Microbiol. 1990; 136: 273-281Crossref PubMed Scopus (24) Google Scholar). Proposals for the molecular basis of the catalytic mechanism of members of the amino acid dehydrogenase superfamily have been suggested following the structure determination of the binary complex of the glutamate dehydrogenase from C. symbiosum with glutamate (27Stillman T.J. Baker P.J. Britton K.L. Rice D.W. J. Mol. Biol. 1993; 234: 1131-1139Crossref PubMed Scopus (202) Google Scholar), coupled to earlier kinetic studies on bovine GluDH (31Rife J.E. Cleland W.W. Biochemistry. 1980; 19: 2328-2333Crossref PubMed Scopus (75) Google Scholar, 32Fisher H.F. Maniscalco S. Singh N. Mehrotra R.N. Srinivasan R. Biochim. Biophys. Acta. 1992; 1119: 52-56Crossref PubMed Scopus (17) Google Scholar). The structural studies have highlighted a number of potentially important residues, which are fully conserved in GluDH, LeuDH, and PheDH. In LeuDH, these key residues include 5 glycines (residues 41, 42, 77, 78, and 290) important in the design of the active site; Lys68, which recognizes the 1-carboxyl group of the amino acid substrate; Asp115, which is believed to be involved in proton transfer to and from the amino acid substrate during catalysis; and Lys80, which has an unusually low pK a and is thought to enhance the nucleophilicity of an essential water molecule involved in the proposed reaction mechanism (Fig. 6 a). Comparison of the sequences for LeuDH and ValDH in Fig. 2 reveals that all of these residues are conserved, indicating that they share an identical catalytic mechanism. LeuDH shows a considerable structural resemblance to other NAD(P)+-linked dehydrogenases in its nucleotide binding domain (Fig.5) and recognizes the nucleotide in a similar manner (28Baker P.J. Turnbull A.P. Sedelnikova S.E. Stillman T.J. Rice D.W. Structure. 1995; 3: 693-705Abstract Full Text Full Text PDF PubMed Scopus (86) Google Scholar, 33Rossmann M.G. Moras D. Olsen K.W. Nature. 1974; 250: 194-199Crossref PubMed Scopus (1194) Google Scholar). Specific interactions between LeuDH and the NAD+ include hydrogen bonds between the adenine ribose hydroxyl groups and the side chain carboxyl group of Asp203. Additionally, the side chain of Asp203forms a hydrogen bond to the main chain peptide NH group of Leu181, which lies in a glycine-rich region (residues 180–185) between βg and the dinucleotide-binding helix, α7 (34Wierenga R.K. De Maeyer M.C.H. Hol W.G.L. Biochemistry. 1985; 24: 1346-1357Crossref Scopus (469) Google Scholar). Furthermore, in LeuDH the pyrophosphate moiety of the NAD+forms hydrogen bonds to the main chain peptide NH groups of Asn183 and Val184 within this region. In ValDH, the presence of an aspartate equivalent to Asp203 and the glycine-rich region can be inferred from the sequence alignment, implying that the interactions with this part of the NAD+cofactor in ValDH and LeuDH are similar. In B. sphaericusLeuDH, the 2′ and 3′ hydroxyl groups of the nicotinamide ribose make hydrogen bonds to the side chain carboxyl group and main chain peptide NH group of Asp261, respectively, a residue that is also conserved in B. stearothermophilus LeuDH. Although this residue is an asparagine in the six remaining LeuDH and ValDH sequences, it is anticipated that it will also be involved in cofactor recognition. In LeuDH, the nicotinamide ring lies in the cleft between the two domains, above α12 which forms one face of the active site pocket, and close to α6. Considerable homology between LeuDH and ValDH can be detected in this area. This includes the conservation of Thr150, which forms a hydrogen bond to the carboxyamide moiety of the nicotinamide ring and which is responsible for determining the conformation of the glycosidic bond that leads to the presentation of the 4-pro-S hydrogen of the NADH toward the active site in LeuDH. The conservation of residues in this region, and in particular Thr150, provides a molecular explanation for the identical stereospecificity of the hydride transfer step observed in S. cinnamonensis ValDH (15Priestley N.D. Robinson J.A. Biochem. J. 1989; 261: 853-861Crossref PubMed Scopus (24) Google Scholar). A comparison of the x-ray structures of C. symbiosum GluDH and B. sphaericus LeuDH (26Baker P.J. Britton K.L. Engel P.C. Farrants G.W. Lilley K.S. Rice D.W. Stillman T.J. Proteins Struct. Funct. Genet. 1992; 12: 75-86Crossref PubMed Scopus (231) Google Scholar, 28Baker P.J. Turnbull A.P. Sedelnikova S.E. Stillman T.J. Rice D.W. Structure. 1995; 3: 693-705Abstract Full Text Full Text PDF PubMed Scopus (86) Google Scholar) has revealed that the discrimination of the amino acid substrate between these two enzymes is achieved by a combination of point mutations at the base of the substrate side chain binding pocket (Lys89 and Ser380 in GluDH, equivalent to Leu40 and Val294 in LeuDH, respectively) and changes in the shape of this pocket resulting from movements in the main chain. In LeuDH, there are 13 residues with a side chain atom lying within 6 Å of the expected position of the leucine substrate determined by molecular modeling using the binding site of glutamate in GluDH as a guide (Table III). These include Leu40 and Val294, which are involved in crucial interactions with the side chain of the amino acid substrate and are thought to be principally responsible for controlling the amino acid specificity. All of these residues are identical in the seven aligned ValDH and LeuDH sequences and in the x-ray LeuDH sequence (Fig.6 a). Given the difference in substrate specificity between ValDH and LeuDH, this is somewhat surprising. However, a number of differences occur in the residues that lie more remote from the substrate binding site. Thus, of the 40 residues with either a side chain or main chain atom between 6 and 10 Å from the leucine substrate, 28 are identical in all the aligned ValDH and LeuDH sequences, including the B. sphaericus LeuDH x-ray sequence (Table IV). Although the sequences at three positions (residues 43, 62, and 261 in LeuDH) differ between some ValDHs and LeuDHs, similarities can also be seen between some members of the two families. Thus, we assume that these positions are less important in the observed differential substrate specificity. The remaining nine positions provide substitutions that are identical within, but not between, members of the LeuDH and ValDH families and include T66S, L76H, T81A, I111V, E114C, V133T, F140N, P146S, and N293Q. The sequence differences are highlighted in Fig. 6 b, where it can be seen that they form the second shell of residues surrounding those on the surface of the substrate binding site. Furthermore, examination of Fig. 6 b also indicates that the C-terminal tail in LeuDH lies directly behind the elements of the secondary structure that form the substrate binding pocket and therefore its absence in ValDH may lead to further subtle changes in the detailed conformation within this region. Taken together, the similarities in the residues which lie directly in the amino acid specificity pocket coupled with the changes in the second shell suggest that the modulation of substrate specificity between these enzymes arises in part through changes in the shape of the amino acid specificity pocket caused by differences in the relative position of residues which line the pocket, rather than their mutation. The values for the relative rates of LeuDH and ValDH, highlighted in Tables I and II, indicate that the energetics of discrimination between the substrates leucine and valine are small and thus minor movements of the side chains within the active site can be sufficient to produce this discrimination. Currently it is impossible to predict the nature of any such subtle changes by homology-based modeling, and structural data are required to understand how these differences provide fine control over the substrate specificity. In the longer term, information gleaned from the comparison of closely related but distinct structures may improve our ability to model these types of changes and advance our understanding of the relationship between structure and function.Table IIIResidues in B. sphaericus LeuDH with at least one side chain atom (or Cα in the case of glycine) residing within 6 Å of the modeled leucine substrateResidue (LeuDH x-ray sequence)No. of side chain atoms <6 ÅASide chain atom nearest substrateClosest substrate atomSeparationÅLeu404Cδ2Cδ24.0Gly41CαCδ14.3Gly42CαN4.2Arg441NH2O15.6Met654CεCδ12.8Lys684CεO24.5Asn692Oδ1Cδ15.3Lys803NζO13.6Ala1131CβCδ23.5Asp1154Oδ2N2.7Thr1341Cγ2Cδ24.9Val2913Cγ2O22.8Val2943Cγ2Cδ23.7These include Leu40 and Val294, critical for the amino acid specificity of LeuDH. All 13 residues are identical in the ValDH sequences, implying that the character of the amino acid binding pocket is preserved. Open table in a new tab Table IVPatterns of sequence substitution between LeuDH and ValDH close to the active siteResidue no. (B. sphaericus x-ray sequence numbering)Residue type in LeuDH (x-ray/Bc/Bl/Bst/Ti)Residue type in ValDH (Sci/Sco/Sfr)43Ala/Thr/Thr/Thr/MetThr62Ala/Ala/Ala/Ala/GlySer/Ala/Ala66ThrSer76LeuHis81ThrAla111IleVal114GluCys133ValThr140PheAsn146ProSer261Asp/Asn/Asn/Asp/AsnAsn293AsnGlnOf the 40 residues in LeuDH with either a side chain or main chain atom lying between 6 and 10 Å from the leucine substrate, 28 residues are identical in all the aligned ValDH and LeuDH sequences: Ala39, Leu61, Gly64, Tyr67, Ala72, Leu74, Gly77, Gly78, Gly79, Tyr110, Thr112, Val116, Gly117, Thr118, Met123, Gly135, Thr150, Asn262, Tyr284, Asn287, Ala288, Gly289, Gly290, Ile292, Ala295, Asp296, Gly297, and Leu298. The sequences at three positions (43, 62, and 261) differ between some ValDHs and LeuDHs, but similarities can be seen between some members of the two families. The remaining nine positions provide substitutions that are identical within, but not between, members of the LeuDH and ValDH families (T66S, L76H, T81A, I111V, E114C, V133T, F140N, P146S, and N293Q). Open table in a new tab These include Leu40 and Val294, critical for the amino acid specificity of LeuDH. All 13 residues are identical in the ValDH sequences, implying that the character of the amino acid binding pocket is preserved. Of the 40 residues in LeuDH with either a side chain or main chain atom lying between 6 and 10 Å from the leucine substrate, 28 residues are identical in all the aligned ValDH and LeuDH sequences: Ala39, Leu61, Gly64, Tyr67, Ala72, Leu74, Gly77, Gly78, Gly79, Tyr110, Thr112, Val116, Gly117, Thr118, Met123, Gly135, Thr150, Asn262, Tyr284, Asn287, Ala288, Gly289, Gly290, Ile292, Ala295, Asp296, Gly297, and Leu298. The sequences at three positions (43, 62, and 261) differ between some ValDHs and LeuDHs, but similarities can be seen between some members of the two families. The remaining nine positions provide substitutions that are identical within, but not between, members of the LeuDH and ValDH families (T66S, L76H, T81A, I111V, E114C, V133T, F140N, P146S, and N293Q). The Krebs Institute is a designated Biomolecular Sciences Center of the Biotechnology and Biological Sciences Research Council.