Alexander Wlodawer

Catalytic triads and their relatives

Interactions among the residues in the serine protease Asp-His-Ser catalytic triad, in the special environment of the enzyme-substrate complex, activate the nucleophilic potential of the seryl O gamma. In the subtilisin and trypsin families, the composition and arrangement of the catalytic triad do not vary significantly. However, the mechanisms of action of many other hydrolytic enzymes, which target a wide range of substrates, involve nucleophilic attack by a serine (or threonine) residue. Review of these enzymes shows that the acid-base-ser/thr pattern of catalytic residues is generally conserved, although the individual acids and bases can vary. The variations in sequence and organization illustrate the adaptability shown by proteins in generating catalytic stereochemistry on different main-chain frameworks.

Structure of bovine pancreatic trypsin inhibitor. Results of joint neutron and X-ray refinement of crystal form II

The structure of form II crystals of bovine pancreatic trypsin inhibitor has been investigated by joint refinement of X-ray and neutron data. Crystallographic R factors for the final model were 0.200 for the X-ray data extending to 1 A resolution and 0.197 for the 1.8 A neutron data. This model was strongly restrained, with 0.020 A root-mean-square (r.m.s.) departure of bond lengths from their ideal values and 0.019 A r.m.s. departure of planar groups from planarity. The resulting structure was very similar to that of crystal form I (r.m.s. deviation for main chain atoms was 0.40 A); nevertheless larger deviations were observed in particular regions of the chain. Twenty out of 63 ordered water molecules occupy similar positions (deviation less than 1 A) in both models. Eleven amide hydrogens were found to be protected from exchange after three months of soaking the crystals in deuterated mother liquor at pH 8.2. Their locations were in excellent agreement with the results obtained by two-dimensional nuclear magnetic resonance, but the rates of exchange are much lower in the crystalline state.

Comparison of two highly refined structures of bovine pancreatic trypsin inhibitor

The high resolution structures of bovine pancreatic trypsin inhibitor refined in two distinct crystal forms have been compared. One of the structures was a result of new least-squares X-ray refinement of data from crystal form I, while the other was the joint X-ray/neutron structure of crystal form II. After superposition, the molecules show an overall root-mean-squares deviation of 0.40 A for the atoms in the main chain, while the deviations for the side-chain atoms are 1.53 A. The latter number decreases to 0.61 A when those side-chains that adopted drastically different conformations are excluded from comparison. The discrepancy between atomic temperature factors in the two models was 6.7 A2, while their general trends are highly correlated. About half of the solvent molecules occupy similar positions in the two models, while the others are different. As expected, solvents with the lowest temperature factors are most likely to be common in the two crystal forms. While the two models are clearly similar, the differences are significantly larger than the errors inherent in the structure determination.

1,2,3-triazole as a peptide surrogate in the rapid synthesis of HIV-1 protease inhibitors.

Given the ubiquitous nature of the peptide linkage in biological molecules, replacement of the amide bond with isosteres in potential drug candidates has been a continual goal of many laboratories. Successful replacements will provide improved stability, lipophilicity, and absorption. Many surrogates have been introduced already, yet the synthesis of many of these isosteres in a combinatorial way is difficult and requires several steps. Thus, the discovery of new peptide surrogates with easier syntheses is an important achievement that could open new opportunities for the study of amide-containing molecules and the development of inhibitors with novel physicochemical properties. We have used the copper(i)-catalyzed azide–alkyne [3+2] cycloaddition as a straightforward reaction for the preparation of inhibitor libraries. Over 100 compounds were synthesized in microtiter plates and screened in situ. Two of these compounds—AB2 (pdb-1zp8) and AB3 (pdb-1zpA)—showed the best activity against wild type and mutant HIV-1 proteases (Table 1). AB2 and AB3, were then computationally docked by using AutoDock3. The docking simulation produced two conformations of approximately equal energy. One conformation placed the triazole in the position normally adopted by the peptide unit—between P2’ and P1’—in peptidomimetic compounds. Furthermore, the central nitrogen of the triazole was perfectly positioned to form a hydrogen bond with the water molecule normally found under the protease flaps. This water molecule also formed a hydrogen bond with the sulfonamide as seen in the crystallographic structure of amprenavir when bound to HIV-1 protease. The other conformation positioned the compounds in a similar place, but with the triazole rotated by 180 8. This allowed for a slightly better fit of the triazole substituent but sacrificed the hydrogen bond with the water molecule. In this work we have solved the ambiguity in binding conformation by solving the crystal structure of two inhibitors derived from a library of triazole compounds with HIV-1 protease. Interestingly, the two structures show that the triazole ring is an effective amide surrogate that retains all hydrogen bonds in the active site (Figure 1). HIV-1 protease (3 mgmL 1 in 0.025m sodium acetate pH 5.4, 10 mm dithiothreitol, 1 mm EDTA) was combined with inhibitor (32 mm in 50% (v/v) dimethylsulfoxide and 2-methylpentane2,4-diol) at 4 8C to give a 2:1 molar ratio of inhibitor to protein, and the mixture was centrifuged to remove the precipitate. The complex was crystallized by the hanging-drop vapor-diffusion method by mixing 9.6 mL of protease solution with 4 mL of crystallization buffer (1.34m ammonium sulfate, 0.1m sodium acetate, pH 4.8–5.4). Plates were sealed at 20 8C for one to two weeks. Data were collected from frozen crystals at the Argonne National Laboratory SER-CAT beamline 22-ID and with a Rigaku Table 1. Binding constants of 1,2,3-triazole compounds to HIV-1 protease.

Improved molecular replacement by density- and energy-guided protein structure optimization

Molecular replacement procedures, which search for placements of a starting model within the crystallographic unit cell that best account for the measured diffraction amplitudes, followed by automatic chain tracing methods, have allowed the rapid solution of large numbers of protein crystal structures. Despite extensive work, molecular replacement or the subsequent rebuilding usually fail with more divergent starting models based on remote homologues with less than 30% sequence identity. Here we show that this limitation can be substantially reduced by combining algorithms for protein structure modelling with those developed for crystallographic structure determination. An approach integrating Rosetta structure modelling with Autobuild chain tracing yielded high-resolution structures for 8 of 13 X-ray diffraction data sets that could not be solved in the laboratories of expert crystallographers and that remained unsolved after application of an extensive array of alternative approaches. We estimate that the new method should allow rapid structure determination without experimental phase information for over half the cases where current methods fail, given diffraction data sets of better than 3.2 Å resolution, four or fewer copies in the asymmetric unit, and the availability of structures of homologous proteins with >20% sequence identity.

Structure of form III crystals of bovine pancreatic trypsin inhibitor.

The structure of bovine pancreatic trypsin inhibitor has been solved in a new crystal form III. The crystals belong to space group P2(1)2(1)2 with a = 55.2 A, b = 38.2 A, c = 24.05 A. The structure was solved on the basis of co-ordinates of forms I and II of the inhibitor by molecular replacement, and the X-ray data extending to 1.7 A were used in a restrained least-squares refinement. The final R factor was 0.16, and the deviation of bonded distances from ideality was 0.020 A. Root-mean-square discrepancy between C alpha co-ordinates of forms III and I are 0.47 A, whilst between forms II and III the discrepancy is 0.39 A. These deviations are about a factor of 3 larger than the expected experimental errors, showing that true differences exist between the three crystal forms. Two residues (Arg39 and Asp50) were modeled with two positions for their side-chains. The final model includes 73 water molecules and one phosphate group bound to the protein. Sixteen water molecules occupy approximately the same positions in all three crystal forms studied to date, indicating their close association with the protein molecule. Temperature factors also show a high degree of correlation between the three crystal forms.

The catalytic domain of avian sarcoma virus integrase: conformation of the active-site residues in the presence of divalent cations

BACKGROUND Members of the structurally-related superfamily of enzymes that includes RNase H, RuvC resolvase, MuA transposase, and retroviral integrase require divalent cations for enzymatic activity. So far, cation positions are reported in the X-ray crystal structures of only two of these proteins, E. coli and human immunodeficiency virus 1 (HIV-1) RNase H. Details of the placement of metal ions in the active site of retroviral integrases are necessary for the understanding of the catalytic mechanism of these enzymes. RESULTS The structure of the enzymatically active catalytic domain (residues 52-207) of avian sarcoma virus integrase (ASV IN) has been solved in the presence of divalent cations (Mn2+ or Mg2+), at 1.7-2.2 A resolution. A single ion of either type interacts with the carboxylate groups of the active site aspartates and uses four water molecules to complete its octahedral coordination. The placement of the aspartate side chains and metal ions is very similar to that observed in the RNase H members of this superfamily; however, the conformation of the catalytic aspartates in the active site of ASV IN differs significantly from that reported for the analogous residues in HIV-1 IN. CONCLUSIONS Binding of the required metal ions does not lead to significant structural modifications in the active site of the catalytic domain of ASV IN. This indicates that at least one metal-binding site is preformed in the structure, and suggests that the observed constellation of the acidic residues represents a catalytically competent active site. Only a single divalent cation was observed even at extremely high concentrations of the metals. We conclude that either only one metal ion is needed for catalysis, or that a second metal-binding site can only exist in the presence of substrate and/or other domains of the protein. The unexpected differences between the active sites of ASV IN and HIV-1 IN remain unexplained; they may reflect the effects of crystal contacts on the active site of HIV-1 IN, or a tendency for structural polymorphism.

Structural comparisons among the short-chain helical cytokines

BACKGROUND Cytokines and growth factors are soluble proteins that regulate the development and activities of many cell types. One group of these proteins have structures based on a four-helix bundle, though this similarity is not apparent from amino acid sequence comparisons. An understanding of how diverse sequences can adopt the same fold would be useful for recognizing and aligning distant homologs and for applying structural information gained from one protein to other sequences. RESULTS We have approached this problem by comparing the five known structures which adopt a granulocyte-macrophage colony-stimulating factor (GM-CSF)-like, or short-chain fold: interleukin (IL)-4, GM-CSF, IL-2, IL-5, and macrophage colony-stimulating factor. The comparison reveals a common structural framework of five segments including 31 inner-core and 30 largely exposed residues. Buried polar interactions found in each protein illustrate how complementary substitutions maintain protein stability and may help specify unique core packing. A profile based on the known structures is not sufficient to guarantee accurate amino acid sequence alignments with other family members. Comparisons of the conserved short-chain framework with growth hormone define the optimal structural alignment. CONCLUSIONS Our results are useful for extrapolating functional results among the short-chain cytokines and growth hormone, and provide a foundation for similar characterization of other subfamilies. These results also show that the placement of polar residues at different buried positions in each protein complicates sequence comparisons, and they document a challenging test case for methods aimed at recognizing and aligning distant homologs.

Protein crystallography for non-crystallographers, or how to get the best (but not more) from published macromolecular structures

The number of macromolecular structures deposited in the Protein Data Bank now exceeds 45 000, with the vast majority determined using crystallographic methods. Thousands of studies describing such structures have been published in the scientific literature, and 14 Nobel prizes in chemistry or medicine have been awarded to protein crystallographers. As important as these structures are for understanding the processes that take place in living organisms and also for practical applications such as drug design, many non‐crystallographers still have problems with critical evaluation of the structural literature data. This review attempts to provide a brief outline of technical aspects of crystallography and to explain the meaning of some parameters that should be evaluated by users of macromolecular structures in order to interpret, but not over‐interpret, the information present in the coordinate files and in their description. A discussion of the extent of the information that can be gleaned from the coordinates of structures solved at different resolution, as well as problems and pitfalls encountered in structure determination and interpretation are also covered.

Filamentous phage infection: crystal structure of g3p in complex with its coreceptor, the C-terminal domain of TolA.

BACKGROUND Infection of male Escherichia coli cells by filamentous Ff bacteriophages (M13, fd, and f1) involves interaction of the phage minor coat gene 3 protein (g3p) with the bacterial F pilus (primary receptor), and subsequently with the integral membrane protein TolA (coreceptor). G3p consists of three domains (N1, N2, and CT). The N2 domain interacts with the F pilus, whereas the N1 domain--connected to N2 by a flexible glycine-rich linker and tightly interacting with it on the phage--forms a complex with the C-terminal domain of TolA at later stages of the infection process. RESULTS The crystal structure of the complex between g3p N1 and TolA D3 was obtained by fusing these domains with a long flexible linker, which was not visible in the structure, indicating its very high disorder and presumably a lack of interference with the formation of the complex. The interface between both domains, corresponding to approximately 1768 A2 of buried molecular surface, is clearly defined. Despite the lack of topological similarity between TolA D3 and g3p N2, both domains interact with the same region of the g3p N1 domain. The fold of TolA D3 is not similar to any previously known protein motifs. CONCLUSIONS The structure of the fusion protein presented here clearly shows that, during the infection process, the g3p N2 domain is displaced by the TolA D3 domain. The folds of g3p N2 and TolA D3 are entirely different, leading to distinctive interdomain contacts observed in their complexes with g3p N1. We can now also explain how the interactions between the g3p N2 domain and the F pilus enable the g3p N1 domain to form a complex with TolA.