The structure of Pyrococcus furiosus glutamate dehydrogenase reveals a key role for ion-pair networks in maintaining enzyme stability at extreme temperatures.

BACKGROUND: The hyperthermophile Pyrococcus furiosus is one of the most thermostable organisms known, with an optimum growth temperature of 100 degrees C. The proteins from this organism display extreme thermostability. We have undertaken the structure determination of glutamate dehydrogenase from P. furiosus in order to gain further insights into the relationship between molecular structure and thermal stability. RESULTS: The structure of P. furiosus glutamate dehydrogenase, a homohexameric enzyme, has been determined at 2.2 A resolution and compared with the structure of glutamate dehydrogenase from the mesophile Clostridium symbiosum. CONCLUSIONS: Comparison of the structures of these two enzymes has revealed one major difference: the structure of the hyperthermophilic enzyme contains a striking series of ion-pair networks on the surface of the protein subunits and buried at both interdomain and intersubunit interfaces. We propose that the formation of such extended networks may represent a major stabilizing feature associated with the adaptation of enzymes to extreme temperatures.

Insights into the molecular basis of thermal stability from the structure determination of Pyrococcus furiosus glutamate dehydrogenase.

The structure determination of the glutamate dehydrogenase from the hyperthermophile Pyrococcus furiosus has been completed at 2.2 A resolution. The structure has been compared with the glutamate dehydrogenases from the mesophiles Clostridium symbiosum, Escherichia coli and Neurospora crassa. This comparison has revealed that the hyperthermophilic enzyme contains a striking series of networks of ion-pairs which are formed by regions of the protein which contain a high density of charged residues. Such regions are not found in the mesophilic enzymes and the number and extent of ion-pair formation is much more limited. The ion-pair networks are clustered at both inter domain and inter subunit interfaces and may well represent a major stabilising feature associated with the adaptation of enzymes to extreme temperatures.

Initial formation of a non-covalent enzyme-reagent complex during the inactivation of clostridial glutamate dehydrogenase by Ellman's reagent: determination of the enzyme's dissociation constant for the binary complex with NAD+ from protection studies.

The time-course of reaction between Ellman's reagent (DTNB) and clostridial glutamate dehydrogenase has been investigated over a wide range of reagent concentrations (50-5000 microM) and showed pseudo-first-order kinetics throughout. The reaction was followed both by monitoring loss of enzyme activity and by detection of released thionitrobenzoate through its absorbance at 412 nm, and, when both methods were used for the same DTNB concentration, the pseudo-first-order rate constants were identical within experimental error, suggesting that the two methods detect the same process. The dependence of the rate constants on DTNB concentration clearly shows saturation, with a limiting value of 1.62 x 10(-3) s-1 and a dissociation constant of 1.0 mM governing the formation of the implied non-covalent enzyme-DTNB complex. This information has allowed a detailed analysis of the protection of the enzyme by NAD+, yielding a value of 334 microM for the dissociation constant for the enzyme-coenzyme binary complex. In view of the convenience of protection studies as a means of determining dissociation constants, this study emphasizes the importance of establishing whether a chemical modification reaction follows simple first-order kinetics with respect to the chemical reagent.

CoA-persulphide: a possible in vivo inhibitor of mammalian short-chain acyl-CoA dehydrogenase.

The characteristic green colour of native short-chain acyl-CoA dehydrogenases (EC 1.3.99.2) results from a charge transfer complex between the FAD prosthetic group and a tightly bound molecule of CoA-persulphide. The native enzyme from ox liver mitochondria was found to have about 60% of its FAD cofactor liganded with CoA-persulphide. When artificially fully liganded with CoA-persulphide, this enzyme was inhibited by 90% in comparison to unliganded enzyme. Enzymic activity could be slowly restored by displacing the CoA-persulphide with high concentrations of butyryl-CoA, the enzyme's physiological substrate. The results show that CoA-persulphide is a potent inhibitor of short-chain acyl-CoA dehydrogenase and may have a physiological role in the regulation of beta-oxidation.

A convenient and rapid method for the complete removal of CoA from butyryl-CoA dehydrogenase.

The commercially available gel, 2-pyridyl disulphide hydroxypropyl ether-Sepharose (thiopropyl-Sepharose 6B), can be used to remove bound ligand completely from butyryl-CoA dehydrogenase (EC 1.399.2) in two simple operations. The resultant enzyme forms normal complexes with acetoacetyl-CoA and CoA persulphide, contains no bound CoA as determined by the enzymatic assay for CoA, and retains full catalytic activity.

Functional studies of a glutamate dehydrogenase with known three-dimensional structure: steady-state kinetics of the forward and reverse reactions catalysed by the NAD(+)-dependent glutamate dehydrogenase of Clostridium symbiosum.

Steady-state kinetic properties of glutamate dehydrogenase from Clostridium symbiosum are reported. Rates with NADP(H) are over three hundred times lower than with NAD(H) under identical conditions. The 3-acetyl pyridine and 6-deamino adenine analogues of NAD+, on the other hand, are used almost as well as NAD+ itself. Amino acid specificity is very tight at both pH 7 and pH 9. The best alternative substrate of those tested, L-alpha-amino-gamma-nitraminobutyrate, gave only 0.5% of the rate seen with glutamate. With 400 microM NAD+ a 160-fold variation of the glutamate concentration gave a linear Eadie plot apart from slight inhibition at the highest concentrations. With 40 mM L-glutamate and varied [NAD+], the Eadie plot appeared linear between 1.6 microM and 60 microM and again between 60 microM and 2000 microM, but the slopes of the two lines differed by a factor of 8.4. This striking pattern is not attributable to impurities in the coenzyme or to changes in the state of aggregation of the enzyme. For the high concentration range (greater than 60 microM NAD+), the presence of all four linear terms in the reciprocal form of the initial rate equation indicates a sequential mechanism. Similar measurements made for APAD+ and dnNAD+ show no sign of non-linearity in the Eadie plot over the wide concentration ranges explored. In the reductive amination direction, with NADH as coenzyme, linear reciprocal plots were obtained for all three substrates. Systematic variation of concentrations led via primary, secondary and tertiary plots to all eight possible initial-rate parameters in a linear reciprocal initial-rate equation. Compulsory-order and enzyme-substitution mechanisms appear to be excluded, and a random route to the central complex seems the only possibility compatible with the results.

Kinetic studies on the binding of 1,N6-etheno-NAD+ to glutamate dehydrogenase from Clostridium symbiosum.

The mechanism of the binding of reduced coenzyme (NAD+) to clostridial glutamate dehydrogenase (GDH) was determined by transient kinetics. The fluorescent 1,N6-ethenoadenine analogue of NAD+ (epsilonNAD+) was used as a probe of nucleotide binary and ternary complex formation because the binding of NAD+ is optically silent. The kinetics of epsilonNAD+ binding were consistent with a 3-step binding process. The enzyme was found to oscillate between two conformational forms, termed E1 and E2, in the presence and absence of L-glutamate. However, L-glutamate shifted the equilibrium from 96.8% to 99% of the enzyme in the E1 form. The rapid-equilibrium binding of epsilonNAD+ to the E2 form was rate limited by a slow isomerisation of the ternary complex as the binary complex became saturated with epsilonNAD+. The L-glutamate binary complex had a greater affinity for the coenzyme (Kd = 11 microM) than the free enzyme (Km = 39 microM), indicative of a positive interaction of the substrate and coenzyme binding sites. Steady-state studies were also indicative of a positive interaction in the formation of the catalytic complex, with this complex having a Kd for epsilonNAD+ of 6.8 microM. Consequently, there is stabilization of successive complexes on the reaction pathway.

Correlation of intron-exon organisation with the three-dimensional structure in glutamate dehydrogenase.

The positions of the intron-exon boundaries in the genes for glutamate dehydrogenase from Chlorella sorokiniana rat, and human have been located on the three-dimensional structure of the highly homologous enzyme from Clostridium symbiosum and analysed for their position in the protein structure. This analysis shows no correlation between the positions of these boundaries in the mammalian and Chlorella glutamate dehydrogenase genes and no correlation with units of function in the enzyme and suggests that the present day exons do not represent the protein modules of an ancestral glutamate dehydrogenase. There appears to be no clear preference for the residues at the splice junctions to be either buried or exposed to solvent. However, the frequency with which the introns appear in the loops linking elements of secondary structure, rather than in either the alpha-helical or beta-sheet segments, is higher than predicted on the basis of the proportion of residues in the loops. This is consistent with but not proof of a role for exon modification/exchange in protein evolution since changes at these positions are less likely to disturb the structure and hence maintain function.

Construction of a dimeric form of glutamate dehydrogenase from Clostridium symbiosum by site-directed mutagenesis.

By using site-directed mutagenesis, Phe-187, one of the amino-acid residues involved in hydrophobic interaction between the three identical dimers comprising the hexamer of Clostridium symbiosum glutamate dehydrogenase (GDH), has been replaced by an aspartic acid residue. Over-expression in Escherichia coli led to production of large amounts of a soluble protein which, though devoid of GDH activity, showed the expected subunit M(r) on SDS-PAGE, and cross-reacted with an anti-GDH antibody preparation in Western blots. The antibody was used to monitor purification of the inactive protein. F187D GDH showed altered mobility on non-denaturing electrophoresis, consistent with changed size and/or surface charge. Gel filtration on a calibrated column indicated an M(r) of 87000 +/- 3000. The mutant enzyme did not bind to the dye column routinely used in preparing wild-type GDH. Nevertheless suspicions of major misfolding were allayed by the results of chemical modification studies: as with wild-type GDH, NAD+ completely protected one-SH group against modification by DTNB, implying normal coenzyme binding. A significant difference, however, is that in the mutant enzyme both cysteine groups were modified by DTNB, rather than C320 only. The CD spectrum in the far-UV region indicated no major change in secondary structure in the mutant protein. The near-UV CD spectrum, however, was less intense and showed a pronounced Phe contribution, possibly reflecting the changed environment of Phe-199, which would be buried in the hexamer. Sedimentation velocity experiments gave corrected coefficients S20,W of 11.08 S and 5.29 S for the wild-type and mutant proteins. Sedimentation equilibrium gave weight average molar masses M(r,app) of 280000 +/- 5000 g/mol. consistent with the hexameric structure for the wild-type protein and 135000 +/- 3000 g/mol for F187D. The value for the mutant is intermediate between the values expected for a dimer (98000) and a trimer (147000). To investigate the basis of this, sedimentation equilibrium experiments were performed over a range of protein concentrations. M(r,app) showed a linear dependence on concentration and a value of 108 118 g/mol at infinite dilution. This indicates a rapid equilibrium between dimeric and hexameric forms of the mutant protein with an equilibrium constant of 0.13 l/g. An independent analysis of the radial absorption scans with Microcal Origin software indicated a threefold association constant of 0.11 l/g. Introduction of the F187D mutation thus appears to have been successful in producing a dimeric GDH species. Since this protein is inactive it is possible that activity requires subunit interaction around the 3-fold symmetry axis. On the other hand this mutation may disrupt the structure in a way that cannot be extrapolated to other dimers. This issue can only be resolved by making alternative dimeric mutants.

Contribution of an aspartate residue, D114, in the active site of clostridial glutamate dehydrogenase to the enzyme's unusual pH dependence.

Glutamate dehydrogenase from Clostridium symbiosum displays unusual kinetic behaviour at high pH when compared with other members of this enzyme family. Structural and sequence comparisons with GDHs from other organisms have indicated that the Asp residue at position 114 in the clostridial enzyme may account for these differences. By replacing this residue by Asn, a mutant protein has been created with altered functional properties at high pH. This mutant protein can be efficiently overexpressed in Escherichia coli, and several criteria, including mobility in non-denaturing electrophoresis, circular dichroism (CD) spectra and initial crystallisation studies, suggest a folding and an assembly comparable to those of the wild-type protein. The D114N mutant enzyme shows a higher optimum pH for activity than the wild-type enzyme, and both CD data and activity measurements show that the distinctive time-dependent reversible conformational inactivation seen at high pH in the wild-type enzyme is abolished in the mutant.

The partial amino acid sequence of the NAD(+)-dependent glutamate dehydrogenase of Clostridium symbiosum: implications for the evolution and structural basis of coenzyme specificity.

The amino acid sequence is reported for CNBr and tryptic peptide fragments of the NAD(+)-dependent glutamate dehydrogenase of Clostridium symbiosum. Together with the N-terminal sequence, these make up about 75% of the total sequence. The sequence shows extensive similarity with that of the NADP(+)-dependent glutamate dehydrogenase of Escherichia coli (52% identical residues out of the 332 compared) allowing confident placing of the peptide fragments within the overall sequence. This demonstrated sequence similarity with the E. coli enzyme, despite different coenzyme specificity, is much greater than the similarity (31% identities) between the GDH's of C. symbiosum and Peptostreptococcus asaccharolyticus, both NAD(+)-linked. The evolutionary implications are discussed. In the 'fingerprint' region of the nucleotide binding fold the sequence Gly X Gly X X Ala is found, rather than Gly X Gly X X Gly. The sequence found here has previously been associated with NADP+ specificity and its finding in a strictly NAD(+)-dependent enzyme requires closer examination of the function of this structural motif.

Crystallization of an NAD+-dependent glutamate dehydrogenase from Clostridium symbiosum.

Crystals of a bacterial NAD+-dependent glutamate dehydrogenase (GDHase) have been grown over a wide range of pH values by using the hanging drop method of vapour diffusion with ammonium sulphate as the precipitant. Sodium dodecyl sulphate/polyacrylamide gel electrophoresis of this enzyme together with high pressure liquid chromatography/gel filtration, shows that this GDHase is hexameric like the GDHases of vertebrates. X-ray photographs of the crystals show that they diffract to at least 2.0 A, and an analysis of the diffraction pattern demonstrates that the hexamer is arranged in at least pseudo 32 symmetry.

Evolution of substrate diversity in the superfamily of amino acid dehydrogenases. Prospects for rational chiral synthesis.

We have analysed the sequence homology between glutamate, leucine and phenylalanine dehydrogenases in the light of the solution of the structure of the glutamate dehydrogenase from Clostridium symbiosum. This analysis indicates that the elements of secondary structure comprising the core of the two domains in glutamate dehydrogenase are conserved in the other two enzymes. There is a striking conservation of the residues responsible for the recognition of the nicotinamide ring of the nucleotide cofactor and the backbone of the amino acid substrates. Furthermore, residues involved in a major conformational rearrangement on amino acid binding are preserved, as are those implicated in the catalytic chemistry. In contrast, the pattern of insertions/deletions between these enzymes is consistent with possible differences in quaternary structure. Differential substrate specificity between these enzymes is achieved by critical substitutions at the base of the binding pocket, which accommodates the side-chain of the amino acid substrate. This provides insights into the mutations necessary to produce new catalysts for the chiral synthesis of novel amino acids.

Insights into the mechanism of domain closure and substrate specificity of glutamate dehydrogenase from Clostridium symbiosum.

Comparisons of the structures of glutamate dehydrogenase (GluDH) and leucine dehydrogenase (LeuDH) have suggested that two substitutions, deep within the amino acid binding pockets of these homologous enzymes, from hydrophilic residues to hydrophobic ones are critical components of their differential substrate specificity. When one of these residues, K89, which hydrogen-bonds to the gamma-carboxyl group of the substrate l-glutamate in GluDH, was altered by site-directed mutagenesis to a leucine residue, the mutant enzyme showed increased substrate activity for methionine and norleucine but negligible activity with either glutamate or leucine. In order to understand the molecular basis of this shift in specificity we have determined the crystal structure of the K89L mutant of GluDH from Clostridium symbiosum. Analysis of the structure suggests that further subtle differences in the binding pocket prevent the mutant from using a branched hydrophobic substrate but permit the straight-chain amino acids to be used as substrates. The three-dimensional crystal structure of the GluDH from C. symbiosum has been previously determined in two distinct forms in the presence and absence of its substrate glutamate. A comparison of these two structures has revealed that the enzyme can adopt different conformations by flexing about the cleft between its two domains, providing a motion which is critical for orienting the partners involved in the hydride transfer reaction. It has previously been proposed that this conformational change is triggered by substrate binding. However, analysis of the K89L mutant shows that it adopts an almost identical conformation with that of the wild-type enzyme in the presence of substrate. Comparison of the mutant structure with both the wild-type open and closed forms has enabled us to separate conformational changes associated with substrate binding and domain motion and suggests that the domain closure may well be a property of the wild-type enzyme even in the absence of substrate.

Specificity of coenzyme analogues and fragments in promoting or impeding the refolding of clostridial glutamate dehydrogenase.

NAD+ facilitates high-yield reactivation of clostridial glutamate dehydrogenase (GDH) after unfolding in urea. The specificity of this effect has been explored by using analogues and fragments of NAD+. The adenine portion, unlike the nicotinamide portion, is important for reactivation. Alteration in the nicotinamide portion, in acetylpyridine adenine dinucleotide, has little effect, whereas loss of the 6-NH2 substitution on the adenine ring, in 6-deamino NAD, diminishes the effectiveness of the nucleotide in promoting refolding. Also ADP-ribose, lacking nicotinamide, promotes reactivation whereas NMN-phosphoribose, lacking the adenine, does not. Of the smaller fragments, those containing an adenosine moiety, and especially those with one or more phosphate groups, impede the refolding ability of NAD+, and are able to bind to the folding intermediate though unable to facilitate refolding. These results are interpreted in terms of the known 3D structure for clostridial glutamate dehydrogenase. It is assumed that the refolding intermediate has a more or less fully formed NAD+-binding domain but a partially disordered substrate-binding domain and linking region. Binding of NAD+ or ADP-ribose appears to impose new structural constraints that result in completion of the correct folding of the second domain, allowing association of enzyme molecules to form the native hexamer.

A monomeric mutant of Clostridium symbiosum glutamate dehydrogenase: comparison with a structured monomeric intermediate obtained during refolding.

The refolding of Clostridium symbiosum glutamate dehydrogenase (GDH) involves the formation of an inactive structured monomeric intermediate prior to its concentration-dependent association. The structured monomer obtained after removal of guanidinium chloride was stable and competent for reconstitution into active hexamers. Site-directed mutagenesis of C. symbiosum gdh gene was performed to replace the residues Arg-61 and Phe-187 which are involved in subunit-subunit interactions, as determined by three-dimensional structure analysis. Heterologous over-expression in Escherichia coli of the double mutant (R61E/F187D) led to the production of a soluble protein with a molecular mass consistent with the monomeric form of clostridial GDH. This protein is catalytically inactive but cross-reacts with an anti-wild-type GDH antibody preparation. The double mutant R61E/F187D does not assemble into hexamers. The physical properties and the stability toward guanidinium chloride and urea of R61E/F187D were studied and compared to those of the structured monomeric intermediate.

Novel methylenecyclopropyl-based acyl-CoA dehydrogenase inhibitor.

A novel hexyl-substituted methylenecyclopropyl acetyl-CoA was tested as an enzyme-specific acyl-CoA dehydrogenase inhibitor. Its CoA ester generated in situ from the carboxylic acid and CoASH, displayed marked differences in inhibition specificity as compared to methylenecyclopropyl acetyl-CoA, consistent with the substrate specificities of the target enzymes. Thus methylenecyclopropyl acetyl-CoA inactivated short-chain-specific acyl-CoA dehydrogenase rapidly, medium-chain-specific acyl-CoA dehydrogenase much more slowly and had no effect on long-chain- or very long-chain-specific acyl-CoA dehydrogenases. The hexyl-substituent on the methylenecyclopropyl ring gave an inhibitor which rapidly inactivated MCAD and LCAD whilst VLCAD was inhibited more slowly and SCAD was essentially unaffected. In some cases (e.g. SCAD and MCPA-CoA) inhibition was accompanied by flavin bleaching. In other cases (e.g. LCAD and C6MCPA) less pronounced bleaching suggests a different chemistry of inhibition.

Alteration in relative activities of phenylalanine dehydrogenase towards different substrates by site-directed mutagenesis.

Glycine-124 and leucine-307 of phenylalanine dehydrogenase from Bacillus sphaericus were altered by site-specific mutagenesis to the corresponding residues in leucine dehydrogenase: alanine and valine, respectively. These two residues have previously been implicated from molecular modelling as important in determining the substrate discrimination of the two enzymes. Single and double mutants displayed lower activities towards L-phenylalanine and enhanced activity towards almost all aliphatic amino acid substrates tested compared to the wild-type, thus confirming the predictions made from molecular modelling.

Protein thermostability in extremophiles.

Thermostability of a protein is a property which cannot be attributed to the presence of a particular amino acid or to a post synthetic modification. Thermostability seems to be a property acquired by a protein through many small structural modifications obtained with the exchange of some amino acids and the modulation of the canonical forces found in all proteins such as electrostatic (hydrogen bonds and ion-pairs) and hydrophobic interactions. Proteins produced by thermo and hyperthermophilic microorganisms, growing between 45 and 110 degrees C are in general more resistant to thermal and chemical denaturation than their mesophilic counterparts. The observed structural resistance may reflect a restriction on the flexibility of these proteins, which, while allowing them to be functionally competent at elevated temperatures, renders them unusually rigid at mesophilic temperatures (10-45 degrees C). The increased rigidity at mesophilic temperatures may find a structural determinant in increased compactness. In thermophilic proteins a number of amino acids are often exchanged. These exchanges with some strategic placement of proline in beta-turns give rise to a stabilization of the protein. Mutagenesis experiments have confirmed this statement. From the comparative analysis of the X-ray structures available for several families of proteins, including at least one thermophilic structure in each case, it appears that thermal stabilization is accompanied by an increase in hydrogen bonds and salt bridges. Thermostability appears also related to a better packing within buried regions. Despite these generalisations, no universal rules can be found in these proteins to achieve thermostability.

Protein stability in extremophilic archaea.

Extremophilic microorganisms have adapted their molecular machinery to grow and thrive under the most adverse environmental conditions. These microorganisms have found their natural habitat at the boiling and freezing point of water, in high salt concentration and at extreme pH values. The extremophilic proteins, selected by Nature to withstand this evolutionary pressure, represent a wide research field for scientists from different disciplines and the study of the determinants of their stability has been an important task for basic and applied research. A surprising conclusion emerges from these studies: there are no general rules to achieve protein stabilization. Each extremophilic protein adopts various strategies and the outstanding adaptation to extreme temperature and solvent conditions is realized through the same weak electrostatic and hydrophobic interactions among the ordinary amino acid residues which are also responsible for the proper balance between protein stability and flexibility in mesophilic proteins.