In vitro Sequential Selection and Characterization of Human Immunodeficiency Virus Type 1 Variants with Reduced Sensitivity to Hydroxyethylurea Protease Inhibitors

In vitro resistance to the human immunodeficiency virus (HIV) protease inhibitors SC-52151 and SC-55389A was evaluated in an in vitro sequential selection scheme. HIVRF variants were selected for reduced sensitivity to SC-52151 and subsequently passaged in both SC-52151 and a structurally different hydroxyethylurea protease inhibitor, SC-55389A, to select for dual-resistant virus. SC-52151 selection alone resulted in a 23-fold reduction in virus sensitivity whereas selection in both inhibitors resulted in 34- and eightfold reductions in virus sensitivity to SC-52151 and SC-55389A, respectively. Sequence analysis of the protease gene revealed that SC-52151 -resistant virus had a Gly to Val substitution at residue 48 (G48V) and, in 58% of subclones, an accompanying Val to Ala substitution at residue 82 (V82A). Dual-resistant virus had both G48V and V82A substitutions present and, in the majority of subclones, an lle to Thr and/or Leu to Pro substitution at residues 54 and 63, respectively. Drug susceptibility assays with limiting dilution-cloned HIVRFR (G48V/V82A) and HIVRFRR (G48V/154T/L63P/V82A) viruses demonstrated moderate to high-level cross-resistance to additional structurally non-related protease inhibitors. Recombinant HIVHXB2 proviral clones with G48V, L63P and V82A substitutions showed that one active site mutation was permissible, but the presence of both G48V and V82A substitutions together significantly reduced infectious virus production. Insight into the contributions of the observed substitutions to drug resistance is presented in molecular modelling studies.

Investigation of aminopyridiopyrazinones as PDE5 inhibitors: Evaluation of modifications to the central ring system

Efforts to improve the potency and physical properties of the aminopyridiopyrazinone class of PDE5 inhibitors through modification of the core ring system are described. Five new ring systems are evaluated and features that impart improved potency and improved solubility are delineated.

Structure-based design and synthesis of pyrazinones containing novel P1 ‘side pocket’ moieties as inhibitors of TF/VIIa

We describe the structure-based design, synthesis, and enzymatic activity of a series of substituted pyrazinones as inhibitors of the TF/VIIa complex. These inhibitors contain substituents meta to the P(1) amidine designed to explore additional interactions with the VIIa residues in the so-called 'S(1) side pocket'. A crystal structure of the designed inhibitors demonstrates the ability of the P(1) side pocket moiety to engage Lys192 and main chain of Gly216 via hydrogen bond interactions, thus, providing additional possibility for chemical modification to improve selectivity and/or physical properties of inhibitors.

A novel mechanism of cyclooxygenase-2 inhibition involving interactions with Ser-530 and Tyr-385

A variety of drugs inhibit the conversion of arachidonic acid to prostaglandin G2 by the cyclooxygenase (COX) activity of prostaglandin endoperoxide synthases. Several modes of inhibitor binding in the COX active site have been described including ion pairing of carboxylic acid containing inhibitors with Arg-120 of COX-1 and COX-2 and insertion of arylsulfonamides and sulfones into the COX-2 side pocket. Recent crystallographic evidence suggests that Tyr-385 and Ser-530 chelate polar or negatively charged groups in arachidonic acid and aspirin. We tested the generality of this binding mode by analyzing the action of a series of COX inhibitors against site-directed mutants of COX-2 bearing changes in Arg-120, Tyr-355, Tyr-348, and Ser-530. Interestingly, diclofenac inhibition was unaffected by the mutation of Arg-120 to alanine but was dramatically attenuated by the S530A mutation. Determination of the crystal structure of a complex of diclofenac with murine COX-2 demonstrates that diclofenac binds to COX-2 in an inverted conformation with its carboxylate group hydrogen-bonded to Tyr-385 and Ser-530. This finding represents the first experimental demonstration that the carboxylate group of an acidic non-steroidal anti-inflammatory drug can bind to a COX enzyme in an orientation that precludes the formation of a salt bridge with Arg-120. Mutagenesis experiments suggest Ser-530 is also important in time-dependent inhibition by nimesulide and piroxicam.

Synthesis and X-ray crystal structures of substituted fluorobenzene and benzoquinone inhibitors of the tissue factor VIIa complex

Multistep syntheses of substituted benzenes and benzoquinone inhibitors of tissue Factor VIIa are reported. The benzene analogues were designed such that their substitution pattern would occupy and interact with the S(1), S(2), and S(3) pockets of the tissue Factor VIIa (TF/VIIa) enzyme. The compounds exhibited modest potency on TF/VIIa with selectivity over Factor Xa and thrombin. The X-ray crystal structures of the targeted fluorobenzene 12a and benzoquinone 14 inhibitors bound to TF/VIIa were obtained and will be described.

Design, Synthesis, and Crystal Structure of Selective 2-Pyridone Tissue Factor VIIa Inhibitors

Targeted 2-pyridones were selected as tissue Factor VIIa inhibitors and prepared from 2,6-dibromopyridine via a multistep synthesis. A variety of chemical transformations, including regioselective nucleophilic addition, selective nitrogen alkylation, and a Suzuki coupling, afforded the targeted tissue Factor VIIa inhibitors. The pyridone core was selected as a replacement for the pyrazinone core of noncovalent tissue Factor VIIa inhibitors and designed such that their substitution pattern would occupy and interact with the S(1), S(2), and S(3) pockets of the tissue Factor VIIa enzyme. These compounds were tested in several serine protease enzyme assays involved in the coagulation cascade exhibiting modest activity on tissue Factor VIIa with excellent selectivity over thrombin and Factor Xa. Finally, an X-ray crystal structure of inhibitor 14a bound to tissue Factor VIIa was obtained and will be described.

Design, parallel synthesis, and crystal structures of pyrazinone antithrombotics as selective inhibitors of the tissue factor VIIa complex

Structure-based drug design (SBDD) and polymer-assisted solution-phase (PASP) library synthesis were used to develop a series of pyrazinone inhibitors of the Tissue Factor/Factor VIIa (TF/VIIa) complex. The crystal structure of a tripeptide-alpha-ketothiazole complexed with TF/VIIa was utilized in a docking experiment to identify the pyrazinone core as a starting scaffold. The pyrazinone core could orient the substituents in the correct spatial arrangement to probe the S1, S2, and S3 pockets of the enzyme. A multistep PASP library synthesis was designed to prepare the substituted pyrazinones varying the P1, P2, and P3 moieties. Hundreds of pyrazinone TF/VIIa inhibitors were prepared and tested in several serine protease enzyme assays involved in the coagulation cascade. The inhibitors exhibited modest activity on TF/VIIa with excellent selectivity over thrombin (IIa) and Factor Xa. The structure-activity relationship of the pyrazinone inhibitors will be discussed and X-ray crystal structures of selected compounds complexed with the TF/VIIa enzyme will be described. This study ultimately led to the synthesis of compound 34, which exhibited 16 nM (IC50) activity on TF/VIIa with >6250 x selectivity vs Factor Xa and thrombin. This potent and highly selective inhibitor of TF/VIIa was chosen for preclinical, intravenous proof-of-concept studies to demonstrate the separation between antithrombotic efficacy and bleeding side effects in a nonhuman primate model of electrolytic-induced arterial thrombosis.

Polymer-assisted solution-phase library synthesis and crystal structure of alpha-ketothiazoles as tissue factor VIIa inhibitors

A solution-phase synthesis of an alpha-ketothiazole library of the general form D-Phe-L-AA-Arg-alpha-ketothiazole is described. The five-step synthesis is accomplished using a combination of polymeric reagents and polymer-assisted solution-phase purification concepts, including reactant-sequestering resins, reagent-sequestering resins, and tagged reagents. The multistep synthesis affords desired alpha-ketothiazole products in excellent purities and yields. A variety of L-amino acid inputs were used to probe the S2 pocket of tissue Factor VIIa enzyme to influence both potency and selectivity. An X-ray crystal structure of compound 10k bound to the TF/VIIa complex was obtained that explains the observed selectivity. The alpha-ketothiazoles were found to be potent, reversible-covalent inhibitors of tissue Factor VIIa, with some analogues demonstrating selectivity over thrombin.

Synthesis and Crystal Structures of Substituted Benzenes and Benzoquinones as Tissue Factor VIIa Inhibitors

Several multistep syntheses of substituted benzenes are reported. The benzene analogues were designed such that their substitution pattern would occupy and interact with the S(1), S(2), and S(3) pockets of the tissue Factor VIIa enzyme. A variety of chemical transformations including nucleophilic additions, reductive aminations, Stille couplings, and polymer-assisted solution-phase (PASP) techniques were used to prepare key intermediates and final products. The initial analogues identified some weakly active compounds which ultimately led to a 340 nM (IC(50)) tissue Factor VIIa inhibitor with selectivity over other related enzymes. The structure-activity relationship of these inhibitors and the synthetic progression from the discovery of the lead compound to the development of potent analogues will be discussed. The X-ray crystal structures of fluorobenzene 50c and benzoquinone 54 inhibitors complexed with the TF/VIIa enzyme will also be described.

Polymer-assisted solution-phase (PASP) parallel synthesis of an alpha-ketothiazole library as tissue factor VIIa inhibitors

A solution-phase synthesis of an alpha-ketothiazole library of the general form D-Phe-L-AA-L-Arg-alpha-ketothiazole is described. The five-step synthesis is accomplished using a combination of polymeric reagents and polymer-assisted solution-phase purification protocols, including reactant-sequestering resins, reagent-sequestering resins, and tagged reagents. The multi-step synthesis affords the desired alpha-ketothiazole products in excellent purities and yields. A variety of L-amino acid inputs were used to probe the S2 pocket of the tissue factor (TF) VIIa enzyme to influence both potency and selectivity. An X-ray crystal structure of compound 10e bound to the TF/VIIa complex was obtained that explains the observed selectivity. The alpha-ketothiazoles were found to be potent, reversible-covalent inhibitors of tissue factor VIIa, with some analogues demonstrating selectivity versus thrombin.

Structure-based drug design of pyrazinone antithrombotics as selective inhibitors of the tissue factor VIIa complex

Structure-based drug design coupled with polymer-assisted solution-phase library synthesis was utilized to develop a series of pyrazinone inhibitors of the tissue factor/Factor VIIa complex. The crystal structure of a tri-peptide ketothiazole complexed with TF/VIIa was utilized in a docking experiment that identified a benzyl-substituted pyrazinone as a P(2) surrogate for the tri-peptide. A 5-step PASP library synthesis of these aryl-substituted pyrazinones was developed. The sequence allows for attachment of a variety of P(1) and P(3) moieties, which led to synthesis pyrazinone 23. Compound 23 exhibited 16 nM IC(50) against TF/VIIa with >6250x selectivity versus Factor Xa and thrombin. This potent and highly selective inhibitor of TF/VIIa was chosen for pre-clinical intravenous proof-of-concept studies to demonstrate the separation between antithrombotic efficacy and bleeding side effects in a primate model of thrombosis.

The crystal structure, mutagenesis, and activity studies reveal that patatin is a lipid acyl hydrolase with a Ser-Asp catalytic dyad

Patatin is a nonspecific lipid acyl hydrolase that accounts for approximately 40% of the total soluble protein in mature potato tubers, and it has potent insecticidal activity against the corn rootworm. We determined the X-ray crystal structure of a His-tagged variant of an isozyme of patatin, Pat17, to 2.2 A resolution, employing SeMet multiwavelength anomalous dispersion (MAD) phasing methods. The patatin crystal structure has three molecules in the asymmetric unit, an R-factor of 22.0%, and an R(free) of 27.2% (for 10% of the data not included in the refinement) and includes 498 water molecules. The structure notably revealed that patatin has a Ser-Asp catalytic dyad and an active site like that of human cytosolic phospholipase A(2) (cPLA(2)) [Dessen, A., et al. (1999) Cell 97, 349-360]. In addition, patatin has a folding topology related to that of the catalytic domain of cPLA(2) and unlike the canonical alpha/beta-hydrolase fold. The structure confirms our site-directed mutagenesis and bioactivity data that initially suggested patatin possessed a Ser-Asp catalytic dyad. Alanine-scanning mutagenesis revealed that Ser77 and Asp215 were critical for both esterase and bioactivity, consistent with prior work implicating a Ser residue [Strickland, J. H., et al. (1995) Plant Physiol. 109, 667-674] and a Ser-Asp dyad [Hirschberg, H. J. H. B., et al. (2001) Eur. J. Biochem. 268, 5037-5044] in patatin's catalytic activity. The crystal structure aids the understanding of other structure-function relationships in patatin. Patatin does not display interfacial activation, a hallmark feature of lipases, and this is likely due to the fact that it lacks a flexible lid that can shield the active site.

Solution structure and backbone dynamics of the catalytic domain of matrix metalloproteinase-2 complexed with a hydroxamic acid inhibitor

MMP-2 is a member of the matrix metalloproteinase family that has been implicated in tumor cell metastasis and angiogenesis. Here, we describe the solution structure of a catalytic domain of MMP-2 complexed with a hydroxamic acid inhibitor (SC-74020), determined by three-dimensional heteronuclear NMR spectroscopy. The catalytic domain, designated MMP-2C, has a short peptide linker replacing the internal fibronectin-domain insertion and is enzymatically active. Distance geometry-simulated annealing calculations yielded 14 converged structures with atomic root-mean-square deviations (r.m.s.d.) of 1.02 and 1.62 A from the mean coordinate positions for the backbone and for all heavy atoms, respectively, when 11 residues at the N-terminus are excluded. The structure has the same global fold as observed for other MMP catalytic domains and is similar to previously solved crystal structures of MMP-2. Differences observed between the solution and the crystal structures, near the bottom of the S1' specificity loop, appear to be induced by the large inhibitor present in the solution structure. The MMP-2C solution structure is compared with MMP-8 crystal structure bound to the same inhibitor to highlight the differences especially in the S1' specificity loop. The finding provides a structural explanation for the selectivity between MMP-2 and MMP-8 that is achieved by large inhibitors.

Molecular docking and high-throughput screening for novel inhibitors of protein tyrosine phosphatase-1B

High-throughput screening (HTS) of compound libraries is used to discover novel leads for drug development. When a structure is available for the target, computer-based screening using molecular docking may also be considered. The two techniques have rarely been used together on the same target. The opportunity to do so presented itself in a project to discover novel inhibitors for the enzyme protein tyrosine phosphatase-1B (PTP1B), a tyrosine phosphatase that has been implicated as a key target for type II diabetes. A corporate library of approximately 400 000 compounds was screened using high-throughput experimental techniques for compounds that inhibited PTP1B. Concurrently, molecular docking was used to screen approximately 235 000 commercially available compounds against the X-ray crystallographic structure of PTP1B, and 365 high-scoring molecules were tested as inhibitors of the enzyme. Of approximately 400 000 molecules tested in the high-throughput experimental assay, 85 (0.021%) inhibited the enzyme with IC50 values less than 100 microM; the most active had an IC50 value of 4.2 microM. Of the 365 molecules suggested by molecular docking, 127 (34.8%) inhibited PTP1B with IC50 values less than 100 microM; the most active of these had an IC50 of 1.7 microM. Structure-based docking therefore enriched the hit rate by 1700-fold over random screening. The hits from both the high-throughput and docking screens were dissimilar from phosphotyrosine, the canonical substrate group for PTP1B; the two hit lists were also very different from each other. Surprisingly, the docking hits were judged to be more druglike than the HTS hits. The diversity of both hit lists and their dissimilarity from each other suggest that docking and HTS may be complementary techniques for lead discovery.

Structural insights into the stereochemistry of the cyclooxygenase reaction

Cyclooxygenases are bifunctional enzymes that catalyse the first committed step in the synthesis of prostaglandins, thromboxanes and other eicosanoids. The two known cyclooxygenases isoforms share a high degree of amino-acid sequence similarity, structural topology and an identical catalytic mechanism. Cyclooxygenase enzymes catalyse two sequential reactions in spatially distinct, but mechanistically coupled active sites. The initial cyclooxygenase reaction converts arachidonic acid (which is achiral) to prostaglandin G2 (which has five chiral centres). The subsequent peroxidase reaction reduces prostaglandin G2 to prostaglandin H2. Here we report the co-crystal structures of murine apo-cyclooxygenase-2 in complex with arachidonic acid and prostaglandin. These structures suggest the molecular basis for the stereospecificity of prostaglandin G2 synthesis.

A mutation in human immunodeficiency virus type 1 protease at position 88, located outside the active site, confers resistance to the hydroxyethylurea inhibitor SC-55389A

The hydroxyethylurea human immunodeficiency virus type 1 (HIV-1) protease inhibitors SC-55389A and SC-52151 were used to select drug-resistant variants in vitro. One clinical HIV-1 strain (89-959) and one laboratory HIV-1 strain (LAI) were passaged in peripheral blood mononuclear cells or CEMT4 cells in the presence of SC-55389A. Resistant isolates from both strains consistently had a mutation to serine for asparagine at amino acid 88 (N88S) in the protease gene either alone or in combination with a change to phenylalanine at position 10. The N88S mutation, recreated by oligonucleotide-mediated site-directed mutagenesis in HXB2, was sufficient to confer resistance to SC-55389A. In contrast, SC-52151-resistant variants selected from the monocytotropic strain SF162 had multiple substitutions in the protease gene (I11V, M461, F53L, A71V, and N88D), and the N88D mutation, re-created by oligonucleotide-mediated site-directed mutagenesis in HXB2, did not confer resistance to SC-52151. The potencies of L735,524 and Ro31-8959 were not reduced when these compounds were assayed against variants with either the N88S or N88D substitution. Position 88 is in a helix that lies behind the substrate binding pocket and may indirectly influence inhibitor binding through interactions with the amino acid at position 31. The selected mutations were persistent in the viral populations after more than 20 passages in the absence of drugs. Passaging of virus first in SC-55389A alone and then in combination with SC-52151 resulted in the accumulation of more mutations in the protease gene (L10F, D35E, D37M, I47V, 154L, A71V, V82I, and S88D) and in the selection of a variant that was cross-resistant to multiple protease inhibitors. These results indicate that a mutation in the HIV-1 protease at a position that is located outside of the substrate binding pocket confers resistance to a protease inhibitor and that mutations in the protease gene accumulate with increasing selection pressure and can persist in the absence of selection pressure.

Structural basis for selective inhibition of cyclooxygenase-2 by anti-inflammatory agents

Prostaglandins and glucocorticoids are potent mediators of inflammation. Non-steroidal anti-inflammatory drugs (NSAIDs) exert their effects by inhibition of prostaglandin production. The pharmacological target of NSAIDs is cyclooxygenase (COX, also known as PGH synthase), which catalyses the first committed step in arachidonic-acid metabolism. Two isoforms of the membrane protein COX are known: COX-1, which is constitutively expressed in most tissues, is responsible for the physiological production of prostaglandins; and COX-2, which is induced by cytokines, mitogens and endotoxins in inflammatory cells, is responsible for the elevated production of prostaglandins during inflammation. The structure of ovine COX-1 complexed with several NSAIDs has been determined. Here we report the structures of unliganded murine COX-2 and complexes with flurbiprofen, indomethacin and SC-558, a selective COX-2 inhibitor, determined at 3.0 to 2.5 A resolution. These structures explain the structural basis for the selective inhibition of COX-2, and demonstrate some of the conformational changes associated with time-dependent inhibition.

Three-dimensional structure of human cytomegalovirus protease

Herpesviruses encode a serine protease that specifically cleaves assembly protein. This protease is critical for replication, and represents a new target for antiviral drug design. Here we report the three-dimensional structure of the protease from human cytomegalovirus (hCMV) at 2.27 angstroms resolution. The structure reveals a unique fold and new catalytic strategy for cleavage. The monomer fold of the enzyme, a seven-stranded beta-barrel encircled by a chain of helices that form the carboxy terminus of the molecule, is unrelated to those observed in classic serine proteases such as chymotrypsin and subtilisin. The serine nucleophile at position 132 is activated by two juxtaposed histidine residues at positions 63 and 157. Dimerization, which seems to be necessary for activity, is observed in the crystals. Correlations of the structure with the sequences of herpesvirus proteases suggest that dimerization may confer specificity and recognition in substrate binding.

Using sampling techniques in protein crystallization

The crystallization of homogeneous or highly purified macromolecules depends on many variables such as precipitant, pH, choice of buffer, protein concentration, temperature, the participation of different mono- and divalent ions, as well as the presence of minute amounts of detergent and organic molecules. Finding the best combination among these many parameters is a multi-variable optimization problem. This kind of problem can be treated mathematically by sampling techniques. We have used this technique for protein crystallization. The iterative procedure starts with random sampling, followed by quantitative evaluation and cycles with weighted sampling. A simple procedure, derived from this concept and called MON48, has been successfully applied to many protein crystallization problems.

Structure-function in Escherichia coli iron superoxide dismutase: comparisons with the manganese enzyme from Thermus thermophilus

The crystal structure of dimeric Fe(III) superoxide dismutase (SOD) from Escherichia coli (3006 protein atoms, 2 irons, and 281 solvents) has been refined to an R of 0.184 using all observed data between 40.0 and 1.85 A (34,879 reflections). Features of this structure are compared with the refined structure of MnSOD from Thermus thermophilus. The coordination geometry at the Fe site is distorted trigonal bipyramidal, with axial ligands His26 and solvent (proposed to be OH-), and in-plane ligands His73, Asp156, and His160. Reduction of crystals to the Fe(II) state does not result in significant changes in metal-ligand geometry (R = 0.188 for data between 40.0 and 1.80 A). The arrangement of iron ligands in Fe(II) and Fe(III)SOD closely matches the Mn coordination found in MnSOD from T. thermophilus [Ludwig, M.L., Metzger, A.L., Pattridge, K.A., & Stallings, W.C. (1991) J. Mol. Biol. 219, 335-358]. Structures of the Fe(III) azide (40.0-1.8 A, R = 0.186) and Mn(III) azide (20.0-1.8 A, R = 0.179) complexes, reported here, reveal azide bound as a sixth ligand with distorted octahedral geometry at the metal; the in-plane ligand-Fe-ligand and ligand-Mn-ligand angles change by 20-30 degrees to coordinate azide as a sixth ligand. However, the positions of the distal azide nitrogens are different in the FeSOD and MnSOD complexes. The geometries of the Fe(III), Fe(II), and Fe(III)-azide species suggest a reaction mechanism for superoxide dismutation in which the metal alternates between five- and six-coordination. A reaction scheme in which the ligated solvent acts as a proton acceptor in the first half-reaction [formation of Fe(II) and oxygen] is consistent with the pH dependence of the kinetic parameters and spectroscopic properties of Fe superoxide dismutase.

Comparison of the crystal structures of genetically engineered human manganese superoxide dismutase and manganese superoxide dismutase from Thermus thermophilus: differences in dimer-dimer interaction

The three-dimensional X-ray structure of a recombinant human mitochondrial manganese superoxide dismutase (MnSOD) (chain length 198 residues) was determined by the method of molecular replacement using the related structure of MnSOD from Thermus thermophilus as a search model. This tetrameric human MnSOD crystallizes in space group P2(1)2(1)2 with a dimer in the asymmetric unit (Wagner, U.G., Werber, M.M., Beck, Y., Hartman, J.R., Frolow, F., & Sussman, J.L., 1989, J. Mol. Biol. 206, 787-788). Refinement of the protein structure (3,148 atoms with Mn and no solvents), with restraints maintaining noncrystallographic symmetry, converged at an R-factor of 0.207 using all data from 8.0 to 3.2 A resolution and group thermal parameters. The monomer-monomer interactions typical of bacterial Fe- and Mn-containing SODs are retained in the human enzyme, but the dimer-dimer interactions that form the tetramer are very different from those found in the structure of MnSOD from T. thermophilus. In human MnSOD one of the dimers is rotated by 84 degrees relative to its equivalent in the thermophile enzyme. As a result the monomers are arranged in an approximately tetrahedral array, the dimer-dimer packing is more intimate than observed in the bacterial MnSOD from T. thermophilus, and the dimers interdigitate. The metal-ligand interactions, determined by refinement and verified by computation of omit maps, are identical to those observed in T. thermophilus MnSOD.

High-level production of active HIV-1 protease in Escherichia coli

High levels of active HIV-1 protease (PR) were produced in Escherichia coli, amounting to 8-10% of total cell protein. High production levels were achieved by altering the following parameters: (1) codon preference of the coding region, (2) A+T-richness at the 5' end of the coding region, and (3) promoter. To circumvent the toxicity of HIV-1 PR in E. coli, the gene was expressed as a fusion protein with two different proteolytic autocleavage sequences. In both the cases, the fusion protein could be cleaved in vivo to give an active molecule with the native sequence at the N terminus.

Structure and topological symmetry of the glyphosate target 5-enolpyruvylshikimate-3-phosphate synthase: a distinctive protein fold

5-enol-Pyruvylshikimate-3-phosphate synthase (EPSP synthase; phosphoenolpyruvate:3-phosphoshikimate 1-carboxyvinyltransferase, EC 2.5.1.19) is an enzyme on the pathway toward the synthesis of aromatic amino acids in plants, fungi, and bacteria and is the target of the broad-spectrum herbicide glyphosate. The three-dimensional structure of the enzyme from Escherichia coli has been determined by crystallographic techniques. The polypeptide backbone chain was traced by examination of an electron density map calculated at 3-A resolution. The two-domain structure has a distinctive fold and appears to be formed by 6-fold replication of a protein folding unit comprising two parallel helices and a four-stranded sheet. Each domain is formed from three of these units, which are related by an approximate threefold symmetry axis; in each domain three of the helices are completely buried by a surface formed from the three beta-sheets and solvent-accessible faces of the other three helices. The domains are related by an approximate dyad, but in the present crystals the molecule does not display pseudo-symmetry related to the symmetry of point group 32 because its approximate threefold axes are almost normal. A possible relation between the three-dimensional structure of the protein and the linear sequence of its gene will be described. The topological threefold symmetry and orientation of each of the two observed globular domains may direct the binding of substrates and inhibitors by a helix macrodipole effect and implies that the active site is located near the interdomain crossover segments. The structure also suggests a rationale for the glyphosate tolerance conferred by sequence alterations.

Manganese superoxide dismutase from Thermus thermophilus. A structural model refined at 1.8 A resolution

The structure of Mn(III) superoxide dismutase (Mn(III)SOD) from Thermus thermophilus, a tetramer of chains 203 residues in length, has been refined by restrained least-squares methods. The R-factor [formula: see text] for the 54,056 unique reflections measured between 10.0 and 1.8 A (96% of all possible reflections) is 0.176 for a model comprising the protein dimer and 180 bound solvents, the asymmetric unit of the P4(1)2(1)2 cell. The monomer chain forms two domains as determined by distance plots: the N-terminal domain is dominated by two long antiparallel helices (residues 21 to 45 and 69 to 89) and the C-terminal domain (residues 100 to 203) is an alpha + beta structure including a three-stranded sheet. Features that may be important for the folding and function of this MnSOD include: (1) a cis-proline in a turn preceding the first long helix; (2) a residue inserted at position 30 that distorts the helix near the first Mn ligand; and (3) the locations of glycine and proline residues in the domain connector (residues 92 to 99) and in the vicinity of the short cross connection (residues 150 to 159) that links two strands of the beta-sheet. Domain-domain contacts include salt bridges between arginine residues and acidic side chains, an extensive hydrophobic interface, and at least ten hydrogen-bonded interactions. The tetramer possesses 222 symmetry but is held together by only two types of interfaces. The dimer interface at the non-crystallographic dyad is extensive (1000 A2 buried surface/monomer) and incorporates 17 trapped or structural solvents. The dimer interface at the crystallographic dyad buries fewer residues (750 A2/monomer) and resembles a snap fastener in which a type I turn thrusts into a hydrophobic basket formed by a ring of helices in the opposing chain. Each of the metal sites is fully occupied, with the Mn(III) five-co-ordinate in trigonal bipyramidal geometry. One of the axial ligands is solvent; the four protein ligands are His28, His83, Asp166 and His170. Surrounding the metal-ligand cluster is a shell of predominantly hydrophobic residues from both chains of the asymmetric unit (Phe86A, Trp87A, Trp132A, Trp168A, Tyr183A, Tyr172B, Tyr173B), and both chains collaborate in the formation of a solvent-lined channel that terminates at Tyr36 and His32 near the metal ion and is presumed to be the path by which substrate or other inner-sphere ligands reach the metal.(ABSTRACT TRUNCATED AT 400 WORDS)

Structure-function relationships in iron and manganese superoxide dismutases

Using the complete sequences for MnSOD from Thermus thermophilus and for FeSOD from E. coli, structural models for both oxidized enzymes have been refined, the Mn protein to an R of 0.186 for all data between 10.0 and 1.8 A, and the Fe protein to an R of 0.22 for data between 10.0 and 2.5 A. The results of the refinements support the presence of a solvent as a fifth ligand to Mn(III) and Fe(III) and a coordination geometry that is close to trigonal bipyramidal. The putative substrate-entry channel is comprised of residues from both subunits of the dimer; several basic residues that are conserved may facilitate approach of O2-, while other conserved residues maintain interchain packing interactions. Analysis of the azide complex of Fe(III) dismutase suggests that during turnover O2- binds to the metal at a sixth coordination site without displacing the solvent ligand. Because crystals reduced with dithionite show no evidence for displacement of the protein ligands, the redox-linked proton acceptor (C. Bull and J.A. Fee (1985), Journal of the American Chemistry Society 107, 3295-3304) is unlikely to be one of the histidines which bind the metal ion. Structural, kinetic, titration, and spectroscopic data can be accommodated in a mechanistic scheme which accounts for the differential titration behaviour of the Fe(III) and Fe(II) enzymes at neutral and high pH.

Crystallization and preliminary X-ray characterization of a soybean seed lipoxygenase

An isoenzyme of soybean (Glycine max L. Merrill cv. Provar) lipoxygenase (EC 1.13.11.12) has been crystallized using the vapor diffusion method. Crystals were grown from solutions of the protein (7 mg/ml) using 10 to 20% (w/v) polyethylene glycol 8000 in citrate/phosphate buffer (pH 5.7) containing 0.5% (w/v) n-octyl-beta-D-glucopyranoside. The crystals reached maximum dimensions of 0.3 mm x 0.2 mm x greater than 2 mm. The enzyme crystallized in space group C222(1) with unit cell dimensions a = 246 A, b = 193 A and c = 75 A. A calculated Vm value of 2.35 A3/dalton was obtained assuming two molecules per asymmetric unit. The density of the crystals was found to be 1.16 g/ml, which confirmed the presence of two molecules per asymmetric unit and indicated a solvent content of 47.5%.

Iron superoxide dismutase. Nucleotide sequence of the gene from Escherichia coli K12 and correlations with crystal structures

The nucleotide sequence of the iron superoxide dismutase gene from Escherichia coli K12 has been determined. Analysis of the DNA sequence and mapping of the mRNA start reveal a unique promoter and a putative rho-independent terminator, and suggest that the Fe dismutase gene constitutes a monocistronic operon. The gene encodes a polypeptide product consisting of 192 amino acid residues with a calculated Mr of 21,111. The published N-terminal amino acid sequence of E. coli B Fe dismutase (Steinman, H. M., and Hill, R. L. (1973) Proc. Natl. Acad. Sci. U.S.A. 70, 3725-3729), along with the sequences of seven other peptides reported here, was located in the primary structure deduced from the K12 E. coli gene sequence. A new molecular model for iron dismutase from E. coli, based on the DNA sequence and x-ray data for the E. coli B enzyme at 3.1 A resolution, allows detailed comparison of the structure of the iron enzyme with manganese superoxide dismutase from Thermus thermophilus HB8. The structural similarities are more extensive than indicated by earlier studies and are particularly striking in the vicinity of the metal-ligand cluster, which is surrounded by conserved aromatic residues. The combined structural and sequence information now available for a series of Mn and Fe superoxide dismutases identifies variable regions in these otherwise very similar molecules; the principal variable site occurs in a surface region between the two long helices which dominate the N-terminal domain.

Synthesis, structure, and hydrolysis of esters of strained and unstrained N-phosphonylureas

Compounds 1–5 were prepared to compare reactivity patterns of cyclic and acyclic phosphonylurea esters. The rates and products of reactions of phosphonylurea esters (1–3) with hydroxide in aqueous acetonitrile were analyzed. In these compounds the phosphonate moiety is in a strained five-membered ring, which also contains the ureido group. Structural determination of 1 by X-ray crystallography indicates that the five-membered ring is planar and the internal ring angle at phosphorus is 93.1°. The endocyclic N—C—N angle of the ureido group is 111°. The compounds undergo hydrolysis in alkaline aqueous acetonitrile at 35 °C with a rate about 106 times that of analogues (4, 5) in which the phosphonate group is exocyclic to the ureido ring. Compound 1 undergoes alkaline hydrolysis (k = 9.0 × 103 M−1 s−1) to release the phenoxy group to give 6. The hydrolysis of alkyl esters 2 (k = 2.4 × 104 M−1 s−1)and 3(k = 1.3 × 103 M−1 s−1) leads to cleavage of the endocyclic P—N bond, producing 7 and 8 respectively. The exocyclic alkyl esters (4 and 5) also cleave at the P—N bond with respective rate constants of 6.5 × 10−3 M−1 s−1 and 4.4 × 10−2 M−1 s−1. The data are consistent with a mechanism in which hydroxide adds to 1 to form a pentacoordinate phosphorus intermediate with the phenoxy group in an equatorial position and the ureido ring in apical and equatorial positions (with nitrogen apical). The departure of the urea group is slower than pseudorotation of the intermediate and expulsion of phenoxide. In the isomerized intermediate, phenoxy is apical but the methylene group of the ring, which has low apicophilicity, must also be apical. Reactions of 2 and 3, which have more basic oxygen leaving groups, occur with P—N cleavage because expulsion from the isomerized intermediate in those cases is not sufficiently fast. These results fit reaction patterns at phosphorus that are determined by ring strain and electronegativity of ligands. Contributions from effects due to antiperiplanar interactions between bonding and nonbonding electrons are not detected.

The structure of manganese superoxide dismutase from Thermus thermophilus HB8 at 2.4-A resolution

An atomic model of tetrameric manganese superoxide dismutase from Thermus thermophilus HB8 has been built into an electron density map at 2.4 A resolution, using chemical sequences of Mn dismutases from Thermus aquaticus and Bacillus stearothermophilus. The monomer fold is structurally very similar to the fold of iron dismutase and comprises two domains, each contributing two ligands to the metal. The Mn(III) ion is bound by protein ligands assigned as His 28, His 83, Asp 165, and His 169. Near neighbors in the metal-ligand environment include a series of hydrophobic residues, Phe 86, Trp 87, Trp 131, and Trp 167. The hydroxyl groups of two Tyr residues, at 36 and 182, are less than 7 A from the metal, as is His 32. Gln 150 forms a bridge between Tyr 36 and Trp 131. These ligands and nearby residues are strongly conserved in the known sequences of Mn dismutases. Only one of the two oxygens of Asp 165 has been assigned as a metal ligand, so that in the current model four protein atoms bind Mn(III). These ligand atoms form part of an approximate trigonal bipyramid in which water may occupy an axial position on the side opposite His 28. The conformation of the protein is unusual in the vicinity of the first ligand, His 28, as a consequence of the insertion of an extra residue in an alpha-helix. The distortion of the helix allows His 32 to stack against the ligand, His 169, and brings Tyr 36 close to the Mn ion. Across one of the dimer interfaces, the two Mn ions are separated by about 18 A, and active center residues from adjoining subunits interdigitate; Tyr 172 interacts with His 32 of the neighboring chain and Glu 168 with the backbone of 168 and with the ligand His 169 from the opposite subunit. Only one other dimer interface occurs in the tetramer; it involves residues 55-62 and sequences near 140 and 156. The center of the oligomeric molecule is filled with solvent.

Intercalation model for DNA-cross linking in a 1-nitro-9-aminoacridine derivative, an analog of the antitumor agent "ledakrin" (nitracrine)

Ledakrin (nitracrine), C-283, is a 1-nitro-9-aminoacridine derivative that is used in Poland as an antitumor agent. In order to investigate the basis of the activity of this compound the structure of another analog, [9-(3-dimethyl-1-methylpropylimino)-1-nitro-9,10-dihydroacridin e], C-829, that has similar activity, was determined by X-ray crystallographic techniques and was compared with that of ledakrin, already reported in the literature. In both molecules the proximity of the 1-nitro to the substituted 9-aminoacridine group causes extensive distortions. These compounds are believed to act, after metabolic "activation", by cross-linking DNA. Such cross-linking does not occur in the absence of the 1-nitro group or if the nitro group is moved to the 2-, 3- or 4-position. Computer-assisted model-building has been used to test possible intercalative models. It has shown that functional groups on C-829 or C-283 are, when the acridine portion of the molecule is intercalated as in a proflavine dinucleoside phosphate complex, in positions suitable for DNA cross-linking by activated 1-nitro-9-aminoacridine derivatives.

Manganese and iron superoxide dismutases are structural homologs

The crystal structure of a tetrameric manganese superoxide dismutase from a thermophilic bacterium, Thermus thermophilus HB8, has been determined at 4.4-A resolution by local averaging of electron density maps calculated by isomorphous replacement. The spatial arrangement of the principal secondary structural features of iron superoxide dismutase is conserved in manganese dismutase. The structural homology is displayed by orienting the polypeptide chain of Escherichia coli Fe dismutase in the electron density map of Mn dismutase. Densities corresponding to bound Mn3+ occur at locations equivalent to the Fe3+ positions in iron dismutase, indicating one metal binding site per chain, or four sites per tetramer. The Mn tetramer, with 222 symmetry, is approximately rectangular in shape and appears to be constructed with only two unique interfaces. One set of interchain contacts closely resembles the dimer interface of Fe dismutase, but the other interface utilizes an inserted polypeptide segment that has no equivalent in Fe dismutase.

Iron superoxide dismutase from Escherichia coli at 3.1-A resolution: a structure unlike that of copper/zinc protein at both monomer and dimer levels

The structure of iron superoxide dismutase (EC 1.15.1.1) from Escherichia coli has been determined at 3.1-A resolution. The dimeric molecule is constructed from identical subunits, which are two-domain polypeptides. The NH2-terminal domain is composed of two antiparallel crossing helices and the COOH-terminal domain is a three-layered structure characterized by mixed alpha/beta secondary structural features. The active center iron atoms, separated by 18 A and located near the monomer-monomer interface, are coordinated by two amino acid residues from each domain. Azide binding has been investigated by using difference Fourier techniques. Consistent with the notion of the independent evolution of the copper/zinc dismutase gene, the iron dismutase structure resembles the copper/zinc protein at neither the monomer nor the dimer level.

Crystallographic studies of Escherichia coli citrate synthase

The citrate synthase from Escherichia coli B has been crystallized in a cubic space group with a unit cell spacing of 220 A. X-ray diffraction, electron microscopy, symmetry considerations, and low resolution projection Patterson syntheses are consistent with a model proposed in which 24 tetrameric molecules of Mr = 188,000 +/- 12,000 occupy the unit cell. The space group is apparently P23, although at low resolution the observed systematic absences in reflections are consistent with the space group P43n, a space group not allowed for asymmetric molecules. Estimates of VM suggest that in the true space group, P23, two tetrameric molecules occupy the asymmetric unit.

Characterization of crystals of tetrameric manganese superoxide dismutase from Thermus thermophilus HB8

The tetrameric manganese superoxide dismutase from the extreme thermophile Thermus thermophilus HB8 crystallizes in space group P41212 (or its enantiomorph) with a = b = 147.5 A, c = 55.9 A. The diffraction patterns extent to 1.4 A, and the crystals are very resistant to decay induced by x-irradiation. Measurements of the crystal density in Ficoll gradients are consistent with an asymmetric unit containing the entire tetramer (Mr = 80,000).

Molecular structure and intermolecular interactions of N1'-methoxycarbonylbiotin methyl ester: a model for carboxybiotin

The crystal structure of N1'-methoxycarbonylbiotin methyl ester, a model for N1'-carboxybiotin, has been determined. The ureido carbonyl bond has more double bond (keto) character than does the corresponding bound in free biotin, which has single bond (enolate) character. In addition, there is an interesting intermolecular interaction between the ureido carbonyl oxygen and a methyl group. Comparison of the molecular structure and crystal packing with those of free biotin suggests that the coenzyme may have evolved with the incorporation of the ureido moiety because the electronic configuration of this region of the molecule is sensitive to N1' carboxylation. On decarboxylation, the ureido carbonyl bond becomes more polarized (C-O-), thereby facilitating the deprotonation of N1' and increasing its nucleophilicity. As a result, carboxylation can occur readily. On carboxylation, the carbonyl bond is depolarized (C = O), allowing the carboxylated coenzyme to interact with nonpolar groups and carboxylate them. Thus, the carboxylation and decarboxylation of biotin appear to act as a mechanistic switch, turning off and on the polarization of the ureido carbonyl bond as well as modulating the nucleophilicity of N1'.

Structure of a dinucleoside phosphate--drug complex as model for nucleic acid--drug interaction

The crystal structure of a 3:2 complex of the frameshift mutagen proflavine with the dinucleoside phosphate cytidylyl-3'5'-guanosine has been determined. The complex has one drug molecule intercalated between Watson--Crick base pairs of the nucleotide duplex. The other two proflavine molecules are bound to the exterior of the miniature double helix. The orientation of the base pairs in this miniature double helix has aspects similar to that found in RNA 11.