Structure of a small-molecule inhibitor complexed with GlmU from Haemophilus influenzae reveals an allosteric binding site

N-Acetylglucosamine-1-phosphate uridyltransferase (GlmU) is an essential enzyme in aminosugars metabolism and an attractive target for antibiotic drug discovery. GlmU catalyzes the formation of uridine-diphospho-N-acetylglucosamine (UDP-GlcNAc), an important precursor in the peptidoglycan and lipopolisaccharide biosynthesis in both Gram-negative and Gram-positive bacteria. Here we disclose a 1.9 A resolution crystal structure of a synthetic small-molecule inhibitor of GlmU from Haemophilus influenzae (hiGlmU). The compound was identified through a high-throughput screening (HTS) configured to detect inhibitors that target the uridyltransferase active site of hiGlmU. The original HTS hit exhibited a modest micromolar potency (IC(50) approximately 18 microM in a racemic mixture) against hiGlmU and no activity against Staphylococcus aureus GlmU (saGlmU). The determined crystal structure indicated that the inhibitor occupies an allosteric site adjacent to the GlcNAc-1-P substrate-binding region. Analysis of the mechanistic model of the uridyltransferase reaction suggests that the binding of this allosteric inhibitor prevents structural rearrangements that are required for the enzymatic reaction, thus providing a basis for structure-guided design of a new class of mechanism-based inhibitors of GlmU.     

Structure of the YibK methyltransferase from Haemophilus influenzae (HI0766): a cofactor bound at a site formed by a knot

Despite the suitability of various lattice geometries for coarse-grained modeling of proteins, the actual packing geometry of residues in folded structures has remained largely unexplored. A strong tendency to assume a regular packing geometry is shown here by optimally reorienting and superimposing clusters of neighboring residues from databank structures examined on a coarse-grained (single-site-per-residue) scale. The orientation function (or order parameter) of the examined coordination clusters with respect to fcc lattice directions is found to be 0.82. The observed geometry, which may be termed an incomplete distorted face-centered cubic (fcc) packing, is apparently favored by the drive to maximize packing density, in a fashion analogous to the way identical spheres pack densely and follow fcc geometry. About 2/3 of all residues obey this packing geometry, while the remainder occupy other context-dependent positions. The preferred coordination directions show relatively small variations over the various amino acid types, consistent with uniform residue viewpoint. Both the extremes of solvent-exposed and completely buried residue neighborhoods approximate the same generic packing, the only difference being in the numbers (and not the orientations) of coordination sites that are occupied (or left void for solvent occupancy). We observe the prevalence of a rather uniform (tight) residue packing density throughout the structure, including even the residues packed near solvent-exposed regions. The observed orientation distribution reveals an underlying, intrinsic orientation lattice for proteins.     

Crystal structure of YecO from Haemophilus influenzae (HI0319) reveals a methyltransferase fold and a bound S-adenosylhomocysteine.

The crystal structure of YecO from Haemophilus influenzae (HI0319), a protein annotated in the sequence databases as hypothetical, and that has not been assigned a function, has been determined at 2.2-A resolution. The structure reveals a fold typical of S-adenosyl-L-methionine-dependent (AdoMet) methyltransferase enzymes. Moreover, a processed cofactor, S-adenosyl-L-homocysteine (AdoHcy), is bound to the enzyme, further confirming the biochemical function of HI0319 and its sequence family members. An active site arginine, shielded from bulk solvent, interacts with an anion, possibly a chloride ion, which in turn interacts with the sulfur atom of AdoHcy. The AdoHcy and nearby protein residues delineate a small solvent-excluded substrate binding cavity of 162 A(3) in volume. The environment surrounding the cavity indicates that the substrate molecule contains a hydrophobic moiety and an anionic group. Many of the residues that define the cavity are invariant in the HI0319 sequence family but are not conserved in other methyltransferases. Therefore, the substrate specificity of YecO enzymes is unique and differs from the substrate specificity of all other methyltransferases sequenced to date. Examination of the Enzyme Commission list of methyltransferases prompted a manual inspection of 10 possible substrates using computer graphics and suggested that the ortho-substituted benzoic acids fit best in the active site.     

Structure of YciA from Haemophilus influenzae (HI0827), a hexameric broad specificity acyl-coenzyme A thioesterase

The crystal structure of HI0827 from Haemophilus influenzae Rd KW20, initially annotated "hypothetical protein" in sequence databases, exhibits an acyl-coenzyme A (acyl-CoA) thioesterase "hot dog" fold with a trimer of dimers oligomeric association, a novel assembly for this enzyme family. In studies described in the preceding paper [Zhuang, Z., Song, F., Zhao, H., Li, L., Cao, J., Eisenstein, E., Herzberg, O., and Dunaway-Mariano, D. (2008) Biochemistry 47, 2789-2796], HI0827 is shown to be an acyl-CoA thioesterase that acts on a wide range of acyl-CoA compounds. Two substrate binding sites are located across the dimer interface. The binding sites are occupied by two CoA molecules, one with full occupancy and the second only partially occupied. The CoA molecules, acquired from HI0827-expressing Escherichia coli cells, remained tightly bound to the enzyme through the protein purification steps. The difference in CoA occupancies indicates a different substrate affinity for each of the binding sites, which in turn implies that the enzyme might be subject to allosteric regulation. Mutagenesis studies have shown that the replacement of the putative catalytic carboxylate Asp44 with an alanine residue abolishes activity. The impact of this mutation is seen in the crystal structure of D44A HI0827. Whereas the overall fold and assembly of the mutant protein are the same as those of the wild-type enzyme, the CoA ligands are absent. The dimer interface is perturbed, and the channel that accommodates the thioester acyl chain is more open and wider than that observed in the wild-type enzyme. A model of intact substrate bound to wild-type HI0827 provides a structural rationale for the broad substrate range.     

Crystal structure of homoserine transacetylase from Haemophilus influenzae reveals a new family of alpha/beta-hydrolases

Homoserine transacetylase catalyzes one of the required steps in the biosynthesis of methionine in fungi and several bacteria. We have determined the crystal structure of homoserine transacetylase from Haemophilus influenzae to a resolution of 1.65 A. The structure identifies this enzyme to be a member of the alpha/beta-hydrolase structural superfamily. The active site of the enzyme is located near the end of a deep tunnel formed by the juxtaposition of two domains and incorporates a catalytic triad involving Ser143, His337, and Asp304. A structural basis is given for the observed double displacement kinetic mechanism of homoserine transacetylase. Furthermore, the properties of the tunnel provide a rationale for how homoserine transacetylase catalyzes a transferase reaction vs hydrolysis, despite extensive similarity in active site architecture to hydrolytic enzymes.     

From structure to function: YrbI from Haemophilus influenzae (HI1679) is a phosphatase

The crystal structure of the YrbI protein from Haemophilus influenzae (HI1679) was determined at a 1.67-A resolution. The function of the protein had not been assigned previously, and it is annotated as hypothetical in sequence databases. The protein exhibits the alpha/beta-hydrolase fold (also termed the Rossmann fold) and resembles most closely the fold of the L-2-haloacid dehalogenase (HAD) superfamily. Following this observation, a detailed sequence analysis revealed remote homology to two members of the HAD superfamily, the P-domain of Ca(2+) ATPase and phosphoserine phosphatase. The 19-kDa chains of HI1679 form a tetramer both in solution and in the crystalline form. The four monomers are arranged in a ring such that four beta-hairpin loops, each inserted after the first beta-strand of the core alpha/beta-fold, form an eight-stranded barrel at the center of the assembly. Four active sites are located at the subunit interfaces. Each active site is occupied by a cobalt ion, a metal used for crystallization. The cobalt is octahedrally coordinated to two aspartate side-chains, a backbone oxygen, and three solvent molecules, indicating that the physiological metal may be magnesium. HI1679 hydrolyzes a number of phosphates, including 6-phosphogluconate and phosphotyrosine, suggesting that it functions as a phosphatase in vivo. The physiological substrate is yet to be identified; however the location of the gene on the yrb operon suggests involvement in sugar metabolism.     

Solution structure of HI1506, a novel two-domain protein from Haemophilus influenzae

HI1506 is a 128-residue hypothetical protein of unknown function from Haemophilus influenzae. It was originally annotated as a shorter 85-residue protein, but a more detailed sequence analysis conducted in our laboratory revealed that the full-length protein has an additional 43 residues on the C terminus, corresponding with a region initially ascribed to HI1507. As part of a larger effort to understand the functions of hypothetical proteins from Gram-negative bacteria, and H. influenzae in particular, we report here the three-dimensional solution NMR structure for the corrected full-length HI1506 protein. The structure consists of two well-defined domains, an alpha/beta 50-residue N-domain and a 3-alpha 32-residue C-domain, separated by an unstructured 30-residue linker. Both domains have positively charged surface patches and weak structural homology with folds that are associated with RNA binding, suggesting a possible functional role in binding distal nucleic acid sites.     

Crystal structure of YbaK protein from Haemophilus influenzae (HI1434) at 1.8 A resolution: functional implications

Structural genomics of proteins of unknown function most straightforwardly assists with assignment of biochemical activity when the new structure resembles that of proteins whose functions are known. When a new fold is revealed, the universe of known folds is enriched, and once the function is determined by other means, novel structure-function relationships are established. The previously unannotated protein HI1434 from H. influenzae provides a hybrid example of these two paradigms. It is a member of a microbial protein family, labeled in SwissProt as YbaK and ebsC. The crystal structure at 1.8 A resolution reported here reveals a fold that is only remotely related to the C-lectin fold, in particular to endostatin, and thus is not sufficiently similar to imply that YbaK proteins are saccharide binding proteins. However, a crevice that may accommodate a small ligand is evident. The putative binding site contains only one invariant residue, Lys46, which carries a functional group that could play a role in catalysis, indicating that YbaK is probably not an enzyme. Detailed sequence analysis, including a number of newly sequenced microbial organisms, highlights sequence homology to an insertion domain in prolyl-tRNA synthetases (proRS) from prokaryote, a domain whose function is unknown. A HI1434-based model of the insertion domain shows that it should also contain the putative binding site. Being part of a tRNA synthetases, the insertion domain is likely to be involved in oligonucleotide binding, with possible roles in recognition/discrimination or editing of prolyl-tRNA. By analogy, YbaK may also play a role in nucleotide or oligonucleotide binding, the nature of which is yet to be determined.     

Roles of histidines 154 and 189 and aspartate 139 in the active site of serine acetyltransferase from Haemophilus influenzae

A crystal structure of serine acetyltransferase (SAT) with cysteine bound in the serine subsite of the active site shows that both H154 and H189 are within hydrogen-bonding distance to the cysteine thiol [Olsen, L. R., Huang, B., Vetting, M. W., and Roderick, S. L. (2004) Biochemistry 43, 6013 -6019]. In addition, H154 is in an apparent dyad linkage with D139. The structure suggests that H154 is the most likely catalytic general base and that H189 and D139 may also play important roles during the catalytic reaction. Site-directed mutagenesis was performed to mutate each of these three residues to Asn, one at a time. The V1/Et value of all of the single mutant enzymes decreased, with the largest decrease (approximately 1240-fold) exhibited by the H154N mutant enzyme. Mutation of both histidines, H154N/H189N, gave a V1/Et approximately 23700-fold lower than that of the wild-type enzyme. An increase in K Ser was observed for the H189N, D139N, and H154N/H189N mutant enzymes, while the H154N mutant enzyme gave an 8-fold decrease in K Ser. For all three single mutant enzymes, V1/Et and V1/K Ser Et decrease at low pH and give a pKa of about 7, while the V1/Et of the double mutant enzyme was pH independent. The solvent deuterium kinetic isotope effects on V 1 and V1/K Ser decreased compared to wild type for the H154N mutant enzyme and increased for the H189N mutant enzyme but was about the same as that of wild type for D139N and H154N/H189N. Data suggest that H154, H189, and D139 play different catalytic roles for SAT. H154 likely serves as a general base, accepting a proton from the beta-hydroxyl of serine as the tetrahedral intermediate is formed upon nucleophilic attack on the thioester carbonyl of acetyl-CoA. However, activity is not completely lost upon elimination of H154, and thus, H189 may be able to serve as a backup general base at a lower efficiency compared to H154; it also aids in binding and orienting the serine substrate. Aspartate 139, in dyad linkage with H154, likely facilitates catalysis by increasing the basicity of H154.     

Structure of recombinant Haemophilus influenzae e (P4) acid phosphatase reveals a new member of the haloacid dehalogenase superfamily

Lipoprotein e (P4) from Haemophilus influenzae belongs to the "DDDD" superfamily of phosphohydrolases and is the prototype of class C nonspecific acid phosphatases. P4 is also a component of a H. influenzae vaccine. We report the crystal structures of recombinant P4 in the ligand-free and tungstate-inhibited forms, which are the first structures of a class C phosphatase. P4 has a two-domain architecture consisting of a core alpha/beta domain and a smaller alpha domain. The core domain features a five-stranded beta-sheet flanked by helices on both sides that is reminiscent of the haloacid dehalogenase superfamily. The alpha domain appears to be unique and plays roles in substrate binding and dimerization. The active site is solvent accessible and located in a cleft between the two domains. The structure shows that P4 is a metalloenzyme and that magnesium is the most likely metal ion in the crystalline recombinant enzyme. The ligands of the metal ion are the carboxyl groups of the first and third Asp residues of the DDDD motif, the backbone carbonyl of the second Asp of the DDDD motif, and two water molecules. The structure of the tungstate-bound enzyme suggests that Asp64 is the nucleophile that attacks the substrate P atom. Dimerization appears to be important for catalysis because intersubunit contacts stabilize the active site. Analysis of the structural context of mutations engineered for vaccine studies shows that the most promising mutations are located in the dimer interface. This observation suggests a structure-based vaccine design strategy in which the dimer interface is disrupted in order to expose epitopes that are buried in dimeric P4.     

The HI0073/HI0074 protein pair from Haemophilus influenzae is a member of a new nucleotidyltransferase family: structure,sequence analyses,and solution studies

The crystal structure of HI0074 from Haemophilus influenzae, a protein of unknown function, has been determined at a resolution of 2.4 A. The molecules form an up-down, four-helix bundle, and associate into homodimers. The fold is most closely related to the substrate-binding domain of KNTase, yet the amino acid sequences of the two proteins exhibit no significant homology. Sequence analyses of completely and incompletely sequenced genomes reveal that the two adjacent genes, HI0074 and HI0073, and their close relatives comprise a new family of nucleotidyltransferases, with 15 members at the time of writing. The analyses also indicate that this is one of eight families of a large nucleotidyltransferase superfamily, whose members were identified based on the proximity of the nucleotide- and substrate-binding domains on the respective genomes. Both HI0073 and HI0074 were annotated "hypothetical" in the original genome sequencing publication. HI0073 was cloned, expressed, and purified, and was shown to form a complex with HI0074 by polyacrylamide gel electrophoresis under nondenaturing conditions, analytic size exclusion chromatography, and dynamic light scattering. Double- and single-stranded DNA binding assays showed no evidence of DNA binding to HI0074 or to HI0073/HI0074 complex despite the suggestive shape of the putative binding cleft formed by the HI0074 dimer.     

Novel structure and nucleotide binding properties of HI1480 from Haemophilus influenzae: a protein with no known sequence homologues

The crystal structure of the Haemophilus influenzae protein HI1480 was determined at 2.1-A resolution. The amino acid sequence of HI1480 is unique, having no homology with other known protein sequences. The protein adopts a novel alpha+beta fold, and associates into a dimer of tightly associated dimers. The tight dimers are formed by intermolecular interactions that are mediated by an antiparallel beta-barrel involving both monomers. Helical regions of two dimers mediate the tetramer formation. The helical region contains a four-helix bundle that has been seen only in the anticodon binding domains of class I tRNA synthetases. A cluster of four residues, Tyr18, Arg134, Glu26, and Lys12 is located in a depression formed at the four-helix bundle/ beta-barrel interface. The arrangement is suggestive of an active center, possibly a catalytic site. The HI1480 gene is located within the Mu-like prophage region of H. influenzae, has no homology to bacteriophage genes, and is flanked by transposases. Hence, this is an example of horizontal transfer from an unknown organism. Gel mobility shift assays revealed that HI1480 binds DNA and RNA molecules. Double-stranded DNA is favored over single-stranded DNA, and longer DNA molecules are bound better than shorter ones.     

Solution structure of the highly acidic protein HI1450 from Haemophilus influenzae, a putative double-stranded DNA mimic

The solution structure of the acidic protein HI1450 from Haemophilus influenzae has been determined by NMR spectroscopy. HI1450 has homologues in ten other bacterial species including Escherichia coli, Vibrio cholerae, and Yersinia pestis but there are no functional assignments for any members of the family. Thirty-one of the amino acids in this 107-residue protein are aspartates or glutamates, contributing to an unusually low pI of 3.72. The secondary structure elements are arranged in an alpha-alpha-beta-beta-beta-beta order with the two alpha helices packed against the same side of an anti-parallel four-stranded beta meander. Two large loops, one between beta1 and beta2 and the other between beta2 and beta3 bend almost perpendicularly across the beta-strands in opposite directions on the non-helical side of the beta-sheet to form a conserved hydrophobic cavity. The HI1450 structure has some similarities to the structure of the double-stranded DNA (dsDNA) mimic uracil DNA glycosylase inhibitor (Ugi) including the distribution of surface charges and the position of the hydrophobic cavity. Based on these similarities, as well as having a comparable molecular surface to dsDNA, we propose that HI1450 may function as a dsDNA mimic in order to inhibit or regulate an as yet unidentified dsDNA binding protein.