Refined structure of the acetazolamide complex of human carbonic anhydrase II at 1.9 Å

The binding of acetazolamide to human carbonic anhydrase II (HCA II) has been investigated by X-ray crystallography. The atomic positions of the enzyme inhibitor complex have been refined at 1.9 Å resolution using the least squares refinement program package PROLSQ. The crystallographic R-factor is 17.6%. The bound inhibitor is clearly resolved in the active site of the enzyme. The acetazolamide amine group is bound as a fourth ligand to the zinc ion, the other three are all histidine residues. In addition to van der Waals' interactions and the previously described binding of the sulphonamide group, the inhibitor forms a hydrogen bond from the carbonyl oxygen of the acetylamido group to the amino group of Gln 92.     

Structures of partridge egg-white lysozyme with and without tri-N-acetylchitotriose inhibitor at 1.9 A resolution.

The three-dimensional structures of native partridge egg-white lysozyme (PEWL) and PEWL complexed with tri-N-acetylchitotriose inhibitor have been determined crystallographically and refined at 1.9 A resolution. Crystals of native and complexed protein are isomorphous and have space group and cell dimensions that are identical to those of hen egg-white lysozyme (HEWL) under similar crystallization conditions. Full occupancy of the trisaccharide in the inhibitor complex has allowed definitive modeling and refinement of all three sugar residues, located at subsites A, B, and C in the PEWL active site. A comparison has been made with HEWL/inhibitor complexes in which coordinates were either not refined (Blake CCF, et al., 1967, Proc R Soc B 167:378-388) or were refined at partial occupancy (Cheetham JC, Artymiuk PJ, Phillips DC, 1992, J Mol Biol 224:613-628). Although the loop comprising residues 70-75 is located on the surface of the protein and not near the active site, it appears to be affected indirectly by trisaccharide binding such that the loop shifts toward the active site and becomes relatively immobilized. The source of this loop movement appears to be the anchoring of Trp62, located in the active site cleft, as it forms a hydrogen bond with O6 of the N-acetylglucosamine at site C. Good electron density for the trisaccharide in the PEWL complex structure shows that Asp 101 is involved in hydrogen bonding interactions with the terminal sugar residue.     

Complex of rat transthyretin with tetraiodothyroacetic acid refined at 2.1 and 1.8 A resolution

The crystal structure of rat transthyretin (rTTR) complex with 3,5,3',5'-tetraiodothyroacetic acid (T4Ac) was determined at 1.8 A resolution with low temperature synchrotron data collected at CHESS. The structure was refined to R = 0.207 and Rfree = 0.24 with the use of 8-1.8 A data. The additional 8000 reflections from the incomplete 2.1-1.8 data shell, included in the refinement, reduced the Rfree index by 1.3%. Structure comparison with the model refined against the complete 8-2.1 A data revealed no differences in the ligand orientation and the conformation of the polypeptide chain in the core regions. However, the high-resolution data included in the refinement improved the model in the flexible regions poorly defined with the lower resolution data. Also additional sixteen water molecules were found in the difference map calculated with the extended data. The structure revealed both forward and reverse binding of tetraiodothyroacetic acid in one binding site and two modes of forward ligand binding in the second site, with the phenolic iodine atoms occupying different sets of the halogen binding pockets.     

Three-dimensional structure of ribonuclease T1 complexed with guanylyl-2',5'-guanosine at 1.8 A resolution

The enzyme ribonuclease T1 (RNase T1) isolated from Aspergillus oryzae was cocrystallized with the specific inhibitor guanylyl-2',5'-guanosine (2',5'-GpG) and the structure refined by the stereochemically restrained least-squares refinement method to a crystallographic R-factor of 14.9% for X-ray data above 3 sigma in the resolution range 6 to 1.8 A. The refined model consists of 781 protein atoms, 43 inhibitor atoms in a major site and 29 inhibitor atoms in a minor site, 107 water oxygen atoms, and a metal site assigned as Ca. At the end of the refinement, the orientation of His, Asn and Gln side-chains was reinterpreted on the basis of two-dimensional nuclear magnetic resonance data. The crystal packing and enzyme conformation of the RNase T1/2',5'-GpG complex and of the near-isomorphous RNase T1/2'-GMP complex are comparable. The root-mean-square deviation is 0.73 A between equivalent protein atoms. Differences in the unit cell dimensions are mainly due to the bound inhibitor. The 5'-terminal guanine of 2',5'-GpG binds to RNase T1 in much the same way as in the 2'-GMP complex. In contrast, the hydrogen bonds between the catalytic center and the phosphate group are different and the 3'-terminal guanine forms no hydrogen bonds with the enzyme. This poor binding is reflected in a 2-fold disorder of 2',5'-GpG (except the 5'-terminal guanine), which originates from differences in the pucker of the 5'-terminal ribose. The pucker is C2'-exo for the major site (2/3 occupancy) and C1'-endo for the minor site (1/3 occupancy). The orientation of the major site is stabilized through stacking interactions between the 3'-terminal guanine and His92, an amino acid necessary for catalysis. This might explain the high inhibition rate observed for 2',5'-GpG, which exceeds that of all other inhibitors of type 2',5'-GpN. On the basis of distance criteria, one solvent peak in the electron density was identified as metal ion, probably Ca2+. The ion is co-ordinated by the two Asp15 carboxylate oxygen atoms and by six water molecules. The co-ordination polyhedron displays approximate 4m2 symmetry.     

A neutron Laue diffraction study of endothiapepsin: implications for the aspartic proteinase mechanism.

Current proposals for the catalytic mechanism of aspartic proteinases are largely based on X-ray structures of bound oligopeptide inhibitors possessing nonhydrolyzable analogues of the scissile peptide bond. However, the positions of protons on the catalytic aspartates and the ligand in these complexes have not been determined with certainty. Thus, our objective was to locate crucial protons at the active site of an inhibitor complex since this will have major implications for a detailed understanding of the mechanism of action. We have demonstrated that high-resolution neutron diffraction data can be collected from crystals of the fungal aspartic proteinase endothiapepsin bound to a transition state analogue (H261). The neutron structure of the complex has been refined at a resolution of 2.1 A to an R-factor of 23.5% and an R(free) of 27.4%. This work represents the largest protein structure studied to date by neutron crystallography at high resolution. The neutron data demonstrate that 49% of the main chain nitrogens have exchanged their hydrogen atoms with D2O in the mother liquor. The majority of residues resisting exchange are buried within core beta-sheet regions of the molecule. The neutron maps confirm that the protein has a number of buried ionized carboxylate groups which are likely to give the molecule a net negative charge even at very low pH, thereby accounting for its low pI. The functional groups at the catalytic center have clearly undergone H-D exchange despite being buried by the inhibitor occupying the active site cleft. Most importantly, the data provide convincing evidence that Asp 215 is protonated and that Asp 32 is the negatively charged residue in the transition state complex. This has an important bearing on mechanistic proposals for this class of proteinase.     

Crystal structures of phosphate, iodide and iodate-inhibited phospholipase C from Bacillus cereus and structural investigations of the binding of reaction products and a substrate analogue.

The crystal structure of the complex formed between phospholipase C (PLC) from Bacillus cereus and inorganic phosphate (Pi), which is an inhibitor, has been determined and refined to 2.1 A resolution. The final R-factor is 19.7%. We have also studied the binding of two other inhibitors, iodide and iodate, to PLC. X-ray data for these two complexes were collected to 2.8 A resolution during the search for heavy-atom derivatives. A series of screening experiments where PLC crystals have been treated with several reaction products and a substrate analogue were carried out to clarify the question of substrate binding. The results have so far been ambiguous but are discussed briefly. Phosphate and iodate are both found to bind to the three metal ions in the protein molecule, suggesting that these ions are involved directly in the catalytic process and thereby identifying the active site. PLC also binds nine iodide ions, eight of which are on the surface of the molecule and of lower occupancy. The ninth blocks the entrance to the active site cleft and is of higher occupancy. Altogether, these results suggest that the substrate, a phospholipid, is associated directly with the metal ions during catalysis.     

Structure of a pancreatic alpha-amylase bound to a substrate analogue at 2.03 A resolution.

The structure of pig pancreatic alpha-amylase in complex with carbohydrate inhibitor and proteinaceous inhibitors is known but the successive events occurring at the catalytic center still remain to be elucidated. The X-ray structure analysis of a crystal of pig pancreatic alpha-amylase (PPA, EC 3.2.1.1.) soaked with an enzyme-resistant substrate analogue, methyl 4,4'-dithio-alpha-maltotrioside, showed electron density corresponding to the binding of substrate analogue molecules at the active site and at the "second binding site." The electron density observed at the active site was interpreted in terms of overlapping networks of oligosaccharides, which show binding of substrate analogue molecules at subsites prior to and subsequent to the cleavage site. A weaker patch of density observed at subsite -1 (using a nomenclature where the site of hydrolysis is taken to be between subsites -1 and +1) was modeled with water molecules. Conformational changes take place upon substrate analogue binding and the "flexible loop" that constitutes the surface edge of the active site is observed in a specific conformation. This confirms that this loop plays an important role in the recognition and binding of the ligand. The crystal structure was refined at 2.03 A resolution, to an R-factor of 16.0 (Rfree, 18.5).     

Crystallographic refinement and structure of ribulose-1,5-bisphosphate carboxylase from Rhodospirillum rubrum at 1.7 A resolution

The amino acid sequence of ribulose-1,5-bisphosphate carboxylase/oxygenase from Rhodospirillum rubrum has been fitted to the electron density maps. The resulting protein model has been refined to a nominal resolution of 1.7 A using the constrained-restrained least-squares refinement program of Sussman and the restrained least-squares refinement program of Hendrickson & Konnert. The crystallographic refinement, based on 76,452 reflections with F greater than sigma (F) in the resolution range 5.5 to 1.7 A resulted in a crystallographic R-factor of 18.0%. The asymmetric unit contains one dimeric ribulose-1,5-biphosphate carboxylase molecule, consisting of 869 amino acid residues and 736 water molecules. The geometry of the refined model is close to ideal, with root-mean-square deviations of 0.018 A in bond lengths and 2.7 degrees in bond angles. Two loop regions, comprising residues 54 to 63 and 324 to 335, and the last ten amino acid residues at the C terminus are disordered in our crystals. The expected trimodal distribution is obtained for the side-chain chi 1-angles with a marked preference for staggered conformation. The hydrogen-bonding pattern in the N-terminal beta-sheet and the parallel sheet in the beta/alpha-barrel is described. A number of hydrogen bonds and salt bridges are involved in domain-domain and subunit-subunit interactions. The subunit-subunit interface in the dimer covers an area of 2800 A2. Considerable deviations from the local 2-fold symmetry are found at both the N terminus (residues 2 to 5) and the C terminus (residues 422 to 457). Furthermore, loop 8 in the beta/alpha-barrel domain has a different conformation in the two subunits. A number of amino acid side-chains have different conformations in the two subunits. Most of these residues are located at the surface of the protein. An analysis of the individual temperature factors indicates a high mobility of the C-terminal region and for some of the loops at the active site. The positions and B-factors for 736 solvent sites have been refined (average B: 45.9 A2). Most of the solvent molecules are bound as clusters to the protein. The active site of the enzyme, especially the environment of the activator Lys191 in the non-activated enzyme is described. Crystallographic refinement at 1.7 A resolution clearly revealed the presence of a cis-proline at the active site. This residue is part of the highly conserved region Lys166-Pro167-Lys168.     

Conservation of solvent-binding sites in 10 crystal forms of T4 lysozyme.

Solvent-binding sites were compared in 10 different crystal forms of phage T4 lysozyme that were refined using data from 2.6 A to 1.7 A resolution. The sample included 18 crystallographically independent lysozyme molecules. Despite different crystallization conditions, variable crystal contacts, changes due to mutation, and varying attention to solvent during crystallographic refinement, 62% of the 20 most frequently occupied sites were conserved. Allowing for potential steric interference from neighboring molecules in the crystal lattice, this fraction increased to 79% of the sites. There was, however, no solvent-binding site that was occupied in all 18 lysozyme molecules. A buried double site was occupied in 17 instances and 2 other internal sites were occupied 15 times. Apart from these buried sites, the most frequently occupied sites were often at the amino-termini of alpha-helices. Solvent molecules at the most conserved sites tended to have crystallographic thermal factors lower than average, but atoms with low B-factors were not restricted to these sites. Although superficial inspection may suggest that only 50-60% (or less) of solvent-binding sites are conserved in different crystal forms of a protein, it appears that many sites appear to be empty either because of steric interference or because the apparent occupancy of a given site can vary from crystal to crystal. The X-ray method of identifying sites is somewhat subjective and tends to result in specification only of those solvent molecules that are well ordered and bound with high occupancy, even though there is clear evidence for solvent bound at many additional sites.(ABSTRACT TRUNCATED AT 250 WORDS)     

Crystallographic analysis of a complex between human immunodeficiency virus type 1 protease and acetyl-pepstatin at 2.0-A resolution

The mode of binding of acetyl-pepstatin to the protease from the human immunodeficiency virus type 1 (HIV-1) has been determined by x-ray diffraction analysis. Crystals of an acetyl-pepstatin-HIV-1 protease complex were obtained in space group P2(1)2(1)2 (unit cell dimensions a = 58.39 A, b = 86.70 A, c = 46.27 A) by precipitation with sodium chloride. The structure was phased by molecular replacement methods, and a model for the structure was refined using diffraction data to 2.0 A resolution (R = 0.176 for 12901 reflections with I greater than sigma (I); deviation of bond distances from ideal values = 0.018 A; 172 solvent molecules included). The structure of the protein in the complex has been compared with the structure of the enzyme without the ligand. A core of 44 amino acids in each monomer, including residues in the active site and residues at the dimer interface, remains unchanged on binding of the inhibitor (root mean square deviation of alpha carbon positions = 0.39 A). The remaining 55 residues in each monomer undergo substantial rearrangement, with the most dramatic changes occurring at residues 44-57 (these residues comprise the so-called flaps of the enzyme). The flaps interact with one another and with the inhibitor so as to largely preserve the 2-fold symmetry of the protein. The inhibitor is bound in two approximately symmetric orientations. In both orientations the peptidyl backbone of the inhibitor is extended; a network of hydrogen bonds is formed between the inhibitor and the main body of the protein as well as between the inhibitor and the flaps. Hydrophobic side chains of residues in the body of the protein form partial binding sites for the side chains of the inhibitor; hydrophobic side chains of residues in the flaps complete these binding sites.     

Structure of the xylanase from Penicillium simplicissimum.

Despite its relatively low pH and temperature optimum, the xylanase from Penicillium simplicissimum performs exceedingly well under conditions of paper bleaching. We have purified and characterized this enzyme, which belongs to family 10 of glycosyl hydrolases. Its gene was cloned, and the sequence of the protein was deduced from the nucleotide sequence. The xylanase was crystallized from ammonium sulfate at pH 8.4, and X-ray data were collected at cryo-temperature to a crystallographic resolution of 1.75 A. The crystal structure was solved by molecular replacement using the catalytic domain of the Clostridium thermocellum xylanase as a search model, and refined to a residual of R = 20% (R(free) = 23%) for data between 10 and 1.75 A. The xylanase folds in an (alpha/beta)8 barrel (TIM-barrel), with additional helices and loops arranged at the "top" forming the active site cleft. In its overall shape, the P. simplicissimum xylanase structure is similar to other family 10 xylanases, but its active site cleft is much shallower and wider. This probably accounts for the differences in catalysis and in the mode of action of this enzyme. Three glycerol molecules were observed to bind within the active site groove, one of which interacts directly with the catalytic glutamate residues. It appears that they occupy putative xylose binding subsites.     

Cryo-trapping the six-coordinate, distorted-octahedral active site of manganese superoxide dismutase

Superoxide dismutase protects organisms from potentially damaging oxygen radicals by catalyzing the disproportionation of superoxide to oxygen and hydrogen peroxide. We report the use of cryogenic temperatures to kinetically capture the sixth ligand bound to the active site of manganese superoxide dismutase (MnSOD). Synchrotron X-ray diffraction data was collected from Escherichia coli MnSOD crystals grown at pH 8.5 and cryocooled to 100 K. Structural refinement to 1.55 A resolution and close inspection of the active site revealed electron density for a sixth ligand that was interpreted to be a hydroxide ligand. The six-coordinate, distorted-octahedral geometry assumed during inhibition by hydroxide is compared to the room temperature, five-coordinate, trigonal bipyramidal active site determined with crystals grown from practically identical conditions. The gateway residues Tyr34, His30 and a tightly bound water molecule are implicated in closing-off the active site and blocking the escape route of the sixth ligand.     

Structure of phosphate-free ribonuclease A refined at 1.26 A

The structure of phosphate-free bovine ribonuclease A has been refined at 1.26-A resolution by a restrained least-squares procedure to a final R factor of 0.15. X-ray diffraction data were collected with an electronic position-sensitive detector. The final model consists of all atoms in the polypeptide chain including hydrogens, 188 water sites with full or partial occupancy, and a single molecule of 2-methyl-2-propanol. Thirteen side chains were modeled with two alternate conformations. Major changes to the active site include the addition of two waters in the phosphate-binding pocket, disordering of Gln-11, and tilting of the imidazole ring of His-119. The structure of the protein and of the associated solvent was extensively compared with three other high-resolution, refined structures of this enzyme.     

Three-dimensional structure of the complexes between bovine chymotrypsinogen A and two recombinant variants of human pancreatic secretory trypsin inhibitor (Kazal-type)

Variants of the human pancreatic secretory trypsin inhibitor (PSTI) have been created during a protein design project to generate a high-affinity inhibitor with respect to some serine proteases other than trypsin. Two modified versions of human PSTI with high affinity for chymotrypsin were crystallized as a complex with chymotrypsinogen. Both crystallize isomorphously in space group P4(1)2(1)2 with lattice constants a = 84.4 A, c = 86.7 A and diffract to 2.3 A resolution. The structure was solved by molecular replacement. The final R-value after refinement with 8.0 to 2.3 A resolution data was 19.5% for both complexes after inclusion of about 50 bound water molecules. The overall three-dimensional structure of PSTI is similar to the structure of porcine PSTI in the trypsinogen complex (1TGS). Small differences in the relative orientation of the binding loop and the core of the inhibitors indicate flexible adaptation to the proteases. The chymotrypsinogen part of the complex is similar to chymotrypsin. After refolding induced by binding of the inhibitor the root-mean-square difference of the active site residues A186 to A195 and A217 to A222 compared to chymotrypsin was 0.26 A.     

Refined 1.8 A crystal structure of the lambda repressor-operator complex.

The crystal structure of the lambda repressor-operator complex has been refined to an R-factor of 18.9% at 1.8 A resolution. This refinement, using data collected at low temperature, has revealed the structure of the N-terminal arm and shows that the interactions of repressor with the two halves of the pseudo-symmetric operator site are significantly different. The two halves of the complex are most similar near the outer edge of the operator site (in a region where the lambda and 434 repressors make similar contacts), but they become increasingly different toward the center of the operator. There are striking differences near the center of the site where it appears that the arm makes significant contacts to only one half of the DNA site. This suggested a new way of aligning the operator sites in phage lambda. The high resolution structure confirms many of the previously noted features of the complex, but also reveals a number of new protein-DNA contacts. It also gives a better view of the extensive H-bonding networks that couple contacts made by different residues and different regions of the protein, and reveals important new details about the helix-turn-helix (HTH) region, and the positions of many water molecules in the complex.     

Three-dimensional structure of the complex of 4-guanidino-Neu5Ac2en and influenza virus neuraminidase.

The three-dimensional X-ray structure of a complex of the potent neuraminidase inhibitor 4-guanidino-Neu5Ac2en and influenza virus neuraminidase (Subtype N9) has been obtained utilizing diffraction data to 1.8 A resolution. The interactions of the inhibitor, solvent water molecules, and the active site residues have been accurately determined. Six water molecules bound in the native structure have been displaced by the inhibitor, and the active site residues show no significant conformational changes on binding. Sialic acid, the natural substrate, binds in a half-chair conformation that is isosteric to the inhibitor. The conformation of the inhibitor in the active site of the X-ray structure concurs with that obtained by theoretical calculations and validates the structure-based design of the inhibitor. Comparison of known high-resolution structures of neuraminidase subtypes N2, N9, and B shows good structural conservation of the active site protein atoms, but the location of the water molecules in the respective active sites is less conserved. In particular, the environment of the 4-guanidino group of the inhibitor is strongly conserved and is the basis for the antiviral action of the inhibitor across all presently known influenza strains. Differences in the solvent structure in the active site may be related to variation in the affinities of inhibitors to different subtypes of neuraminidase.     

Refined structure of dimeric diphtheria toxin at 2.0 A resolution.

The refined structure of dimeric diphtheria toxin (DT) at 2.0 A resolution, based on 37,727 unique reflections (F > 1 sigma (F)), yields a final R factor of 19.5% with a model obeying standard geometry. The refined model consists of 523 amino acid residues, 1 molecule of the bound dinucleotide inhibitor adenylyl 3'-5' uridine 3' monophosphate (ApUp), and 405 well-ordered water molecules. The 2.0-A refined model reveals that the binding motif for ApUp includes residues in the catalytic and receptor-binding domains and is different from the Rossmann dinucleotide-binding fold. ApUp is bound in part by a long loop (residues 34-52) that crosses the active site. Several residues in the active site were previously identified as NAD-binding residues. Glu 148, previously identified as playing a catalytic role in ADP-ribosylation of elongation factor 2 by DT, is about 5 A from uracil in ApUp. The trigger for insertion of the transmembrane domain of DT into the endosomal membrane at low pH may involve 3 intradomain and 4 interdomain salt bridges that will be weakened at low pH by protonation of their acidic residues. The refined model also reveals that each molecule in dimeric DT has an "open" structure unlike most globular proteins, which we call an open monomer. Two open monomers interact by "domain swapping" to form a compact, globular dimeric DT structure. The possibility that the open monomer resembles a membrane insertion intermediate is discussed.     

Carbohydrate binding sites in a pancreatic alpha-amylase-substrate complex, derived from X-ray structure analysis at 2.1 A resolution.

The X-ray structure analysis of a crystal of pig pancreatic alpha-amylase (PPA, EC 3.2.1.1.) that was soaked with the substrate maltopentaose showed electron density corresponding to two independent carbohydrate recognition sites on the surface of the molecule. Both binding sites are distinct from the active site described in detail in our previous high-resolution study of a complex between PPA and a carbohydrate inhibitor (Qian M, Buisson G, Duée E, Haser H, Payan F, 1994, Biochemistry 33:6284-6294). One of the binding sites previously identified in a 5-A-resolution electron density map, lies at a distance of 20 A from the active site cleft and can accommodate two glucose units. The second affinity site for sugar units is located close to the calcium binding site. The crystal structure of the maltopentaose complex was refined at 2.1 A resolution, to an R-factor of 17.5%, with an RMS deviation in bond distances of 0.007 A. The model includes all 496 residues of the enzyme, 1 calcium ion, 1 chloride ion, 425 water molecules, and 3 bound sugar rings. The binding sites are characterized and described in detail. The present complex structure provides the evidence of an increased stability of the structure upon interaction with the substrate and allows identification of an N-terminal pyrrolidonecarboxylic acid in PPA.     

Structures of D-xylose isomerase from Arthrobacter strain B3728 containing the inhibitors xylitol and D-sorbitol at 2.5 A and 2.3 A resolution, respectively

The structures of D-xylose isomerase from Arthrobacter strain B3728 containing the polyol inhibitors xylitol and D-sorbitol have been solved at 2.5 A and 2.3 A, respectively. The structures have been refined using restrained least-squares refinement methods. The final crystallographic R-factors for the D-sorbitol (xylitol) bound molecules, for 43,615 (32,989) reflections are 15.6 (14.7). The molecule is a tetramer and the asymmetric unit of the crystal contains a dimer, the final model of which, incorporates a total of 6086 unique protein, inhibitor and magnesium atoms together with 535 bound solvent molecules. Each subunit of the enzyme contains two domains: the main domain is a parallel-stranded alpha-beta barrel, which has been reported in 14 other enzymes. The C-terminal domain is a loop structure consisting of five helical segments and is involved in intermolecular contacts between subunits that make up the tetramer. The structures have been analysed with respect to molecular symmetry, intersubunit contacts, inhibitor binding and active site geometry. The refined model shows the two independent subunits to be similar apart from local deviations due to solvent contacts in the solvent-exposed helices. The enzyme is dependent on a divalent cation for catalytic activity. Two metal ions are required per monomer, and the high-affinity magnesium(II) site has been identified from the structural results presented here. The metal ion is complexed, at the high-affinity site, by four carboxylate side-chains of the conserved residues, Glu180, Glu216, Asp244 and Asp292. The inhibitor polyols are bound in the active site in an extended open chain conformation and complete an octahedral co-ordination shell for the magnesium cation via their oxygen atoms O-2 and O-4. The active site lies in a deep pocket near the C-terminal ends of the beta-strands of the barrel domain and includes residues from a second subunit. The tetrameric molecule can be considered to be a dimer of "active" dimers, the active sites being composed of residues from both subunits. The analysis has revealed the presence of several internal salt-bridges stabilizing the tertiary and quaternary structure. One of these, between Asp23 and Arg139, appears to play a key role in stabilizing the active dimer and is conserved in the known sequences of this enzyme.(ABSTRACT TRUNCATED AT 400 WORDS)