The crystal structure of human phosphoglucose isomerase at 1.6 A resolution: implications for catalytic mechanism, cytokine activity and haemolytic anaemia.

Phosphoglucose isomerase (PGI) is a multifunctional protein, which, inside the cell, functions as a housekeeping enzyme of glycolysis and gluconeogenesis and, outside the cell, exerts wholly unrelated cytokine properties. We have determined the structure of human PGI to a resolution of 1.6 A using X-ray crystallography. The structure is highly similar to other PGIs, especially the architecture of the active site. Fortuitous binding of a sulphate molecule from the crystallisation solution has facilitated an accurate description of the substrate phosphate-binding site. Comparison with both native and inhibitor-bound rabbit PGI structures shows that two loops move closer to the active site upon binding inhibitor. Interestingly, the human structure most closely resembles the inhibitor-bound structure, suggesting that binding of the phosphate moiety of the substrate may trigger this conformational change. We suggest a new mechanism for catalysis that uses Glu357 as the base catalyst for the isomerase reaction rather than His388 as proposed previously. The human PGI structure has also provided a detailed framework with which to map mutations associated with non-spherocytic haemolytic anaemia.

Eight novel mutations and consequences on mRNA and protein level in pyruvate kinase-deficient patients with nonspherocytic hemolytic anemia.

Pyruvate kinase (PK) deficiency (PKD) is an autosomal recessive disorder with the typical manifestation of nonspherocytic hemolytic anemia. We analyzed the mutant enzymes of 10 unrelated patients with PKD, whose symptoms ranged from a mild, chronic hemolytic anemia to a severe anemia, by sequence analysis for the presence of alterations in the PKLR gene. In all cases the patients were shown to be compound heterozygous. Eight novel mutations were identified: 458T-->C (Ile153Thr), 656T-->C (Ile219Thr), 877G-->A (Asp293Asn), 991G-->A (Asp331Asn), 1055C-->A (Ala352Asp), 1483G-->A (Ala495Thr), 1649A-->T (Asp550Val), and 183-184ins16bp. This 16 bp duplication produces a frameshift and subsequent stop codon resulting in a drastically reduced mRNA level, and probably in an unstable gene product. Surprisingly, the existence of M2-type PK could be demonstrated in the patient's red blood cells. The study of different polymorphic sites revealed, with one exception, a strict linkage of the 1705C, 1738T, IVS5(+51)T, T(10) polymorphisms and the presence of 14 ATT repeats in intron 11. Our analyses show the consequences of a distorted structure on enzyme function and we discuss the correlations between the mutations identified and the parameters indicative for enzyme function.

Molecular basis of neurological dysfunction coupled with haemolytic anaemia in human glucose-6-phosphate isomerase (GPI) deficiency.

Glucose-6-phosphate isomerase (GPI) deficiency, an autosomal recessive genetic disorder with the typical manifestation of nonspherocytic haemolytic anaemia, can be associated in some cases with neurological impairment. GPI has been found to be identical to neuroleukin (NLK), which has neurotrophic and lymphokine properties. To focus on the possible effects of GPI mutations on the central nervous system through an effect on neuroleukin activity, we analysed DNA isolated from two patients with severe GPI deficiency, one of them with additional neurological deficits, and their families. The neurologically affected patient (GPI Homburg) is compound heterozygous for a 59 A-->C (H20P) and a 1016 T-->C (L339P) exchange. Owing to the insertion of proline, the H20P and L339P mutations are likely to affect the folding and activity of the enzyme. In the second family studied, the two affected siblings showed no neurological symptoms. The identified mutations are 1166 A-->G (H389R) and 1549 C-->G (L517V), which are located at the subunit interface. We propose that mutations that lead to incorrect folding destroy both catalytic (GPI) and neurotrophic (NLK) activities, thereby leading to the observed clinical symptoms (GPI Homburg). Those alterations at the active site, however, that allow correct folding retain the neurotrophic properties of the molecule (GPI Calden).

Pyruvate kinase of Trypanosoma brucei: overexpression, purification, and functional characterization of wild-type and mutated enzyme.

A procedure was developed for overexpression of Trypanosoma brucei pyruvate kinase in Escherichia coli. The enzyme was purified to near-homogeneity from the bacterial lysate by first removing nucleic acids and contaminating proteins by protamine sulfate precipitation and subsequent passage over a phosphocellulose column. The purified protein is essentially indistinguishable in its physicochemical and kinetic properties from the enzyme purified from trypanosomes. Furthermore, experiments were undertaken to locate the binding site of the allosteric effector fructose 2,6-bisphosphate. Regulation of pyruvate kinase by this effector is unique to trypanosomes and related protozoan organisms. Therefore, a three-dimensional structure model of the enzyme was made, and a putative effector-binding site could be identified in an interdomain cleft. Four residues in this cleft were mutated, and the mutant proteins were produced and purified, using the same methodology as for the wild-type pyruvate kinase. Some mutants showed only minor changes in the activation by the effector. However, substitution of Arg22 by Gly resulted in a 9.2-fold higher S(0.5) for phosphoenolpyruvate and a significantly smaller kcat than the wild-type enzyme. Furthermore, the apparent affinity of this mutant for the allosteric effectors fructose 1,6-bisphosphate and fructose 2,6-bisphosphate was 8.2- and 5.2-fold lower than that of its wild-type counterpart. Effector binding was also affected, although to a lesser extent, in a mutant Phe463Val. These data indicate that particularly residue Arg22, but also Phe463, are somehow involved in the binding of the allosteric effectors.

Cooperativity in Bacillus stearothermophilus pyruvate kinase.

The enzyme pyruvate kinase (PK) from the moderate thermophile Bacillus stearothermophilus has been used as a model system with which to investigate the homotropic and heterotropic cooperative interactions of the enzyme. Cooperative ligand binding by the wild-type enzyme was measured using pre-steady-state and steady-state fluorescence spectroscopy, and steady-state kinetics. The results suggest that the cooperative structural changes induced by the substrate phosphoenolpyruvate (PEP) are distinct from those induced by the allosteric activator ribose- 5-phosphate (R5P). Furthermore the structural transition induced by the binding of saturating amounts of both PEP and R5P is itself distinct. This conclusion was further substantiated by the production of five mutant proteins in which the R5P- and PEP-induced homotropic cooperative transitions were separated. These results suggest that the cooperativity exhibited by pyruvate kinase from B. stearothermophilus does not conform to a simple two-state model. A putative four-state model is proposed.

A general method of domain closure is applied to phosphoglycerate kinase and the result compared with the crystal structure of a closed conformation of the enzyme.

The occurrence of large domain motions associated with the mechanism of action of many proteins is well established. We present a general method of predicting domain closure applicable to proteins containing domains separated by an apparent hinge. The method attempts to allow for natural directional bias within the closing protein by repeatedly applying a weak pulling force over a short distance between pairs of atoms chosen at random in the two domains in question. Appropriate parameters governing the pulling function were determined empirically. The method was applied to the bi-lobal protein PGK and a closed-form activated ternary complex generated for Bacillus stearothermophilus PGK. This model was compared with the recently determined crystal structure of closed-form Trypanosoma brucei PGK. The model predicts the correct hinge regions, although the magnitude of movement at one hinge point was overestimated, and provides a reasonable representation of the closed-form ternary complex.

Structure of apo duck ovotransferrin: the structures of the N and C lobes are in the open form.

The structure of apo duck ovotransferrin (APODOT) has been determined at a resolution of 4.0 A by the molecular replacement method using the structure of duck ovotransferrin (DOT) as the search model. The DOT structure contains two iron binding sites; one in the N-terminal lobe lying between domains N1 and N2 and one in the C-terminal lobe between domains C1 and C2. Both lobes have a closed structure. Models of various forms of both the N and C lobes were used in the search. The final model was refined to give an R factor of 0.22. The comparison of the structure of APODOT with that of DOT shows that both the N and the C lobes are in an open form, where the N2 and C2 domains undergo large rigid-body rotations of 51.6 and 49.9 degrees relative to the N1 and C1 domains, respectively. The interface between the N and C lobes, which is formed by the N1-C1 contact in the core of the molecule does not change significantly. The DOT molecule may be described in terms of three rigid bodies; the N1 and C1 domains as one rigid body forming the static core of the molecule and the N2 and C2 domains as two other rigid bodies which, on the release of iron, move away from the static core of the molecule to form the open structure of APODOT.

Structure of the Fab fragment from a neutralizing monoclonal antibody directed against an epitope of gp41 from HIV-1.

The structure of a Fab fragment of a monoclonal antibody (1583) that neutralizes a broad range of HIV-1 isolates has been solved by X-ray crystallography. This antibody is directed against a poliovirus/HIV-I chimaera which presents a conserved epitope of the envelope protein gp41. Crystals of 1583 were obtained in the space group P2(1)2(1)2(1) and the structure solved by molecular replacement. The model has been refined against all data in the range 10-2.9 A to a final crystallographic R factor of 0.198. The antigen-binding site features a well defined groove, typical of antibodies that bind to small antigens, created in part by a relatively short CDR H3. The variable regions of 1583 were sequenced and, given the hydrophilic nature of the epitope, revealed a surprising lack of charged residues in the CDR's. However, the antigen-binding cleft is indeed very polar, due in part to the presence of two charged residues that emanate from outside the recognized CDR's.

Ligand-induced conformational changes in wild-type and mutant yeast pyruvate kinase.

A mutant form of pyruvate kinase in which serine 384 has been mutated to proline has been engineered in the yeast Saccharomyces cerevisiae. Residue 384 is located in a helix in a subunit interface of the tetrameric enzyme, and the mutation was anticipated to alter the conformation of the helix and hence destabilize the interface. Previous results indicate that the mutant favours the T quaternary conformation over the R conformation, and this is confirmed by the results presented here. Addition of phosphoenol-pyruvate (PEP), ADP and fructose-1, 6-bisphosphate (Fru-1.6-P2) singly to the wild-type and mutant enzymes results in a significant quenching of tryptophan fluorescence (12-44%), and for Fru-1,6-P2, a red shift of 15 nm in the emission maximum. Fluorescence titration experiments showed that PEP, ADP and Fru-1,6-P2 induce conformations which have similar ligand-binding properties in the wild-type and mutant enzymes. However, the Fru-1,6-P2 induced conformation is demonstrably different from those induced by either ADP or PEP. The enzymes differ in their susceptibility to trypsin digestion and N-ethylmaleimide inhibition. The thermal stability of the enzyme is unaltered by the mutation. Far-UV CD spectra show that both enzymes adopt a similar overall secondary structure in solution. Taken together, the results suggest that the Ser384-Pro mutation causes the enzyme to adopt a different tertiary and/or quaternary structure from the wild-type enzyme and affects the type and extent of the conformational changes induced in the enzyme upon ligand binding. A simplified minimal reaction mechanism is proposed in which the R and T states differ in both affinity and kcat. Thus, in terms of the models of cooperativity and allosteric interaction, pyruvate kinase is both a K and a V system.

Structure of diferric duck ovotransferrin at 2.35 A resolution.

The structure of diferric duck ovotransferrin (DOT) has been determined and refined at a resolution of 2.35 A. The DOT structure, which contains two iron binding sites, is similar to the known transferrin and lactoferrin structures. The two iron-binding sites, one in the N-terminal lobe and one in the C-terminal lobe of the molecule, are similar but not identical. The main differences between the three known structures lie in the relative orientations of the N- and C-lobes with respect to each other. In the DOT structure the large aromatic side chain of Phe322 in the N-lobe packs against the conserved residue Gly387 in the C-lobe. This interaction is at the centre of the interface between the two lobes and could play a crucial role in determining their relative orientation. Other differences between the structures occur in the surface loops and in the peptide connecting the two lobes. The final crystallographic model consists of 5309 protein atoms (686 residues), two Fe(3+) ions, two (bi)carbonate ions and three carbohydrate moities. 318 water molecules have been added to the model. The final R factor is 0.22 for 25 400 observed reflections between 10 and 2.35 A resolution.

Refined three-dimensional structure of cat-muscle (M1) pyruvate kinase at a resolution of 2.6 A.

The three-dimensional structure of cat-muscle pyoruvate kinase has been refined at a resolution of 2.6 A. The details of the structure permit interpretation of the original heavy-atom studies and give insight into the importance of conserved residues in pyruvate kinases and the allosteric behaviour of the enzyme. There are a small number of essential residues which determine the relative orientations of domains and the precise nature of intersubunit contacts. Arginine residues are particularly important.

A subunit interface mutant of yeast pyruvate kinase requires the allosteric activator fructose 1,6-bisphosphate for activity.

A variant form of yeast pyruvate kinase (EC 2.7.1.40) with Ser-384 mutated to proline has been engineered in order to study the allosteric properties of this enzyme. Both the mutant and wild-type enzymes were overexpressed in a strain of yeast in which the genomic copy of the pyruvate kinase gene had been disrupted by an insertion of the Ura3 gene. Both enzymes were purified to homogeneity and their kinetic properties characterized. The wild-type enzyme displays sigmoid kinetics with respect to phosphoenolpyruvate (PEP) concentration, and is activated by the allosteric effect fructose 1,6-bisphosphate with concomitant reduction in co-operativity. In contrast, the mutant was found to be dependent on the presence of the effector for catalytic activity and was inactive in its absence. The fully activated mutant enzyme had a kcat. 1.6 times greater than that of the wild-type enzyme. The mutation introduced into the enzyme is in an intersubunit contact which is known to be critical for the allosteric properties of the enzyme, and is far removed from the active site. The major effect of the mutation seems to be to stabilize the low-affinity T state of the apoenzyme, although kcat. is also affected. The S0.5 for PEP and S0.5 for ADP of the wild-type enzyme were 0.22 +/- 0.004 and 0.15 +/- 0.01 mM respectively (means +/- S.E.M.). In the activated mutant enzyme, these kinetic parameters increased to 0.67 +/- 0.03 and 0.43 +/- 0.03 mM respectively. The cooperativity between ADP-binding sites was altered in the mutant enzyme, with the Hill coefficient (h) for ADP increasing to 1.65 +/- 0.07 in the presence of the effector, compared with a value of 0.01 +/- 0.07 for the wild-type enzyme under the same conditions. CD spectroscopy revealed the secondary structure of the mutant enzyme to be little different from that of the wild-type enzyme, indicating that the two enzymes have similar secondary structures in solution. Precise tertiary and quaternary structures such as intersubunit and interdomain interactions may be modified. An improved purification procedure has been devised that allows large quantities of enzyme to be rapidly prepared.

Allosteric activation in Bacillus stearothermophilus lactate dehydrogenase investigated by an X-ray crystallographic analysis of a mutant designed to prevent tetramerization of the enzyme.

The crystal structure of a mutant Bacillus stearothermophilus lactate dehydrogenase, into which an additional loop has been engineered in order to prevent tetramerization of the enzyme, has been solved and refined at 2.4 A. The minimal repeat unit in the crystal is a dimer and the tetramer cannot be generated by any of the crystallographic symmetry operations in P2(1). The loop protrudes out into the solvent, stabilized by a good hydrogen bonding arrangement, and clearly sterically hinders tetramer formation. This is the first structure of B. stearothermophilus lactate dehydrogenase (bsLDH) in which the allosteric activator fructose, 1,6-bisphosphate (FBP) is not present. To investigate the mechanism of allosteric activation in this enzyme we have compared the structure with a ternary complex of B. stearothermophilus lactate dehydrogenase. Many of our observations confirm those reported from a comparison of FBP-bound ternary bsLDH complex with an FBP free LDH from another bacterial source, Bifidobacterium longum. Our results suggest that quaternary structural alterations may have less influence on the mechanism than previously reported. The differences in the quaternary structural behaviour of these two enzymes is discussed.

Key residues in the allosteric transition of Bacillus stearothermophilus pyruvate kinase identified by site-directed mutagenesis.

The structural gene for pyruvate kinase from Bacillus stearothermophilus has been cloned in Escherichia coli and sequenced. The open reading frame from the ATG start codon to the TAG stop codon is 1482 base-pairs and encodes a peptide of relative molecular mass 52,967. In the expression vector pKK223-3, containing the synthetic tac promoter, the gene is overexpressed in E. coli cells to an estimated level of 30% total soluble cell protein. A purification procedure for the overexpressed protein has been established. The construction and characterization of a pair of mutant proteins has given insight into the structural basis of allosteric regulation in the tetrameric enzyme. Substituting tryptophan for tyrosine at position 466 (mutant Trp466-->Tyr) resulted in an activated form of the enzyme, having a reduced K1/2 for the substrate phosphoenolpyruvate. We propose that the characteristics of this mutant might be the result of bulk removal releasing steric inhibition to the formation of an interdomain salt bridge between Asp356 and Arg444. The regulatory behaviour of the double mutant produced by making the additional substitution aspartate for glutamate at position 356 (Trp466-->Tyr/Asp356-->Glu) corroborates this. The position of the salt bridge is such that it might be pivotal to the conformation of a pocket that is proposed to open up when the active R-conformation is adopted. We suggest that the mechanism of activation of B. stearothermophilus pyruvate kinase by ribose-5-phosphate might hinge on an interaction with, or indirectly through, residue Trp466, removing it from the vicinity of the potential salt bridge between Asp356 and Arg444 and thus effecting a closing together of the protein structure concomitant with an opening up of the pocket region.

The cooperative binding of fructose-1,6-bisphosphate to yeast pyruvate kinase.

The cooperative binding of the allosteric activator fructose-1,6-bisphosphate [Fru(1,6)P2] to yeast pyruvate kinase was investigated by equilibrium dialysis and fluorescence quench titration. The results show that yeast pyruvate kinase binds four molecules of Fru(1,6)P2 per tetramer and the observed fluorescence quench follows the binding of the ligand and not the cooperative T to R state transition. Additionally it is shown that the binding of Fru(1,6)P2 to yeast pyruvate kinase is compatible with the model of cooperativity that has been proposed and incorporates an intermediate state, R', with properties between those of the T and R states.

Structure of a ternary complex of an allosteric lactate dehydrogenase from Bacillus stearothermophilus at 2.5 A resolution.

We report the refined structure of a ternary complex of an allosterically activated lactate dehydrogenase, including the important active site loop. Eightfold non-crystallographic symmetry averaging was utilized to improve the density maps. Interactions between the protein and bound coenzyme and oxamate are described in relation to other studies using site-specific mutagenesis. Fructose 1,6-bisphosphate (FruP2) is bound to the enzyme across one of the 2-fold axes of the tetramer, with the two phosphate moieties interacting with two anion binding sites, one on each of two subunits, across this interface. However, because FruP2 binds at this special site, yet does not possess an internal 2-fold symmetry axis, the ligand is statistically disordered and binds to each site in two different orientations. Binding of FruP2 to the tetramer is signalled to the active site principally through two interactions with His188 and Arg173. His188 is connected to His195 (which binds the carbonyl group of the substrate) and Arg173 is connected to Arg171 (the residue that binds the carboxylate group of the substrate).

Purification, characterisation and mutagenesis of highly expressed recombinant yeast pyruvate kinase.

Recombinant yeast pyruvate kinase has been purified from a strain of Saccharomyces cerevisiae expressing the enzyme to very high levels. Expression was from a multicopy plasmid under the control of the yeast phosphoglycerate kinase promoter. The gene was expressed in the absence of the genomically encoded pyruvate kinase, using a strain of yeast in which the pyruvate kinase gene has been disrupted by the insertion of the yeast Ura3 gene. The purification procedure minimised proteolytic artefacts and enabled the convenient purification of 15-20 mg enzyme from 11 culture. The purified enzyme was characterised by a high specific activity and by a lack of proteolytic degradation. Two active-site mutants of yeast pyruvate kinase have been produced, expressed and characterised in this system and preliminary results are described.

Design and synthesis of new enzymes based on the lactate dehydrogenase framework.

Analysis of the mechanism and structure of lactate dehydrogenases is summarized in a map of the catalytic pathway. Chemical probes, single tryptophan residues inserted at specific sites and a crystal structure reveal slow movements of the protein framework that discriminate between closely related small substrates. Only small and correctly charged substrates allow the protein to engulf the substrate in an internal vacuole that is isolated from solvent protons, in which water is frozen and hydride transfer is rapid. The closed vacuole is very sensitive to the size and charge of the substrate and provides discrimination between small substrates that otherwise have too few functional groups to be distinguished at a solvated protein surface. This model was tested against its ability to successfully predict the design and synthesis of new enzymes such as L-hydroxyisocaproate dehydrogenase and fully active malate dehydrogenase. Solvent friction limits the rate of forming the vacuole and thus the maximum rate of catalysis.

Preliminary crystallographic studies on duck ovotransferrin.

Crystals of duck ovotransferrin and duck apo-ovotransferrin have been grown from polyethylene glycol solutions. For both crystals, the space group is P2(1)2(1)2(1), the unit cell dimensions for the ovotransferrin are a = 49.6 A, b = 85.6 A, c = 178.7 A and for the apo-ovotransferrin a = 77.6 A, b = 98.8 A, c = 127.0 A, giving four molecules in the unit cell.