Statistical and conformational analysis of the electron density of protein side chains

Protein side chains make most of the specific contacts between proteins and other molecules, and their conformational properties have been studied for many years. These properties have been analyzed primarily in the form of rotamer libraries, which cluster the observed conformations into groups and provide frequencies and average dihedral angles for these groups. In recent years, these libraries have improved with higher resolution structures and using various criteria such as high thermal factors to eliminate side chains that may be misplaced within the crystallographic model coordinates. Many of these side chains have highly non-rotameric dihedral angles. The origin of side chains with high B-factors and/or with non-rotameric dihedral angles is of interest in the determination of protein structures and in assessing the prediction of side chain conformations. In this paper, using a statistical analysis of the electron density of a large set of proteins, it is shown that: (1) most non-rotameric side chains have low electron density compared to rotameric side chains; (2) up to 15% of chi1 non-rotameric side chains in PDB models can clearly be fit to density at a single rotameric conformation and in some cases multiple rotameric conformations; (3) a further 47% of non-rotameric side chains have highly dispersed electron density, indicating potentially interconverting rotameric conformations; (4) the entropy of these side chains is close to that of side chains annotated as having more than one chi(1) rotamer in the crystallographic model; (5) many rotameric side chains with high entropy clearly show multiple conformations that are not annotated in the crystallographic model. These results indicate that modeling of side chains alternating between rotamers in the electron density is important and needs further improvement, both in structure determination and in structure prediction.     

A pathway of structural changes produced by monastrol binding to Eg5

Monastrol is a small molecule inhibitor that is specific for Eg5, a member of the kinesin 5 family of mitotic motors. Crystallographic models of Eg5 in the presence and absence of monastrol revealed that drug binding produces a variety of structural changes in the motor, including in loop L5 and the neck linker. What is not clear from static crystallographic models, however, is the sequence of structural changes produced by drug binding. Furthermore, because crystallographic structures can be influenced by the packing forces in the crystal, it also remains unclear whether these drug-induced changes occur in solution, at physiologically active concentrations of monastrol or of other drugs that target this site. We have addressed these issues by using a series of spectroscopic probes to monitor the structural consequences of drug binding. Our results demonstrated that the crystallographic model of an Eg5-ADP-monastrol ternary complex is consistent with several solution-based spectroscopic probes. Furthermore, the kinetics of these spectroscopic signal changes allowed us to determine the temporal sequence of drug-induced structural transitions. These results suggested that L5 may be an element in the pathway that links the state of the nucleotide-binding site to the neck linker in kinesin motors.     

Conformation of the Drosophila motor protein non-claret disjunctional in solution from X-ray and neutron scattering

The quaternary structures of monomeric and dimeric Drosophila non-claret disjunctional (ncd) constructs were investigated using synchrotron x-ray and neutron solution scattering, and their low resolution shapes were restored ab initio from the scattering data. The experimental curves were further compared with those computed from crystallographic models of one monomeric and three available dimeric ncd structures in the microtubule-independent ADP-bound state. These comparisons indicate that accounting for the missing parts in the crystal structures for all these constructs is indispensable to obtain reasonable fits to the scattering patterns. A ncd construct (MC6) lacking the coiled-coil region is monomeric in solution, but the calculated scattering from the crystallographic monomer yields a poor fit to the data. A tentative configuration of the missing C-terminal residues in the form of an antiparallel beta-sheet was found that significantly improves the fit. The atomic model of a short dimeric ncd construct (MC5) without 2-fold symmetry is found to fit the data better than the symmetric models. Addition of the C-terminal residues to both head domains gives an excellent fit to the x-ray and neutron experimental data, although the orientation of the beta-sheet differs from that of the monomer. The solution structure of the long ncd construct (MC1) including complete N-terminal coiled-coil and motor domains is modeled by adding a straight coiled-coil section to the model of MC5.     

Allosteric transition and substrate binding are entropy-driven in glucosamine-6-phosphate deaminase from Escherichia coli

Glucosamine-6P-deaminase (EC 3.5.99.6, formerly glucosamine-6-phosphate isomerase, EC 5.3.1.10) from Escherichia coli is an attractive experimental model for the study of allosteric transitions because it is both kinetically and structurally well-known, and follows rapid equilibrium random kinetics, so that the kinetic K(m) values are true thermodynamic equilibrium constants. The enzyme is a typical allosteric K-system activated by N-acetylglucosamine 6-P and displays an allosteric behavior that can be well described by the Monod-Wyman-Changeux model. This thermodynamic study based on the temperature dependence of allosteric parameters derived from this model shows that substrate binding and allosteric transition are both entropy-driven processes in E. coli GlcN6P deaminase. The analysis of this result in the light of the crystallographic structure of the enzyme implicates the active-site lid as the structural motif that could contribute significantly to this entropic component of the allosteric transition because of the remarkable change in its crystallographic B factors.     

P.versicolor龙虾D-甘油醛-3-磷酸脱氢酶的X射线晶体学研究——分子置换法结构测定

由P.versicolor龙虾尾肌提取的HOIO-D-甘油醛-3-磷酸脱氢酶(GAPDH),已长出可供Χ射线衍射用的晶体。初步Χ射线晶体学研究确定:此酶晶体属於C2空间群,不对称单位内含有半个分子,分子坐落在二重轴上。以Homarus Amercanus龙虾GAPDH结构为模型结构,应用分子置换技术进行了低分辨率Χ射线结构分析,结果表明:分子内亚基排列具有222对称性,分子Q轴平行于晶体学二重轴b,分子P和R轴分别平行于晶体学a和c轴。按分子置换法推出的结构模型算得5A分辨率的晶体学R因子为0.46。并获得了一套5A。分辨率的电子密度图。此酶的几种同晶型晶体,特别是荧光NAD衍生物晶体的较高分辨率的结构分析工作正在进行中。     

Anti-insulin antibody structure and conformation. I. Molecular modeling and mechanics of an insulin antibody.

A knowledge-based three-dimensional model of an anti-insulin antibody, 125, was constructed using the structures of conserved residues found in other known crystallographic immunoglobulins. Molecular modeling and mechanics were done with the 125 amino acid sequences using QUANTA and CHARMm on a Silicon Graphics 4D70GT workstation. A minimal model was made by scaffolding using crystallography coordinates of the antibody HyHEL-5, because it had the highest amino acid sequence homology with 125 (84% light chain, 65% heavy chain). The three hypervariable loop turns that are longer in 125 than in HyHEL-5 (L1, L3, and H3) were modeled separately and incorporated into the HyHEL-5 structure; then other amino acid substitutions were made and torsions optimized. The 125 model maintains all the structural attributes of an antibody and the structures conserved in known antibodies. Although there are many polar amino acids (especially serines) in this site, the overall van der Waals surface shape is determined by positions of aromatic side chains. Based on this model, it is suggested that hydrogen bonding may be key in the interaction between the human insulin A chain loop antigenic epitope and 125.     

Receptor rigidity and ligand mobility in trypsin-ligand complexes

The trypsin-like serine proteases comprise a structurally similar family of proteins with a wide diversity of biological functions. Members of this family play roles in digestion, hemostasis, immune responses, and cancer metastasis. Bovine trypsin is an archetypical member of this family that has been extensively characterized both functionally and structurally, and that preferentially hydrolyzes Arg/Lys-Xaa peptide bonds. We have used molecular dynamics (MD) simulations to study bovine trypsin complexed with the two noncovalent small-molecule ligands, benzamidine and tranylcypromine, that have the same hydrogen-bond donating moieties as Arg and Lys side-chains, respectively. Multiple (10) simulations ranging from 1 ns to 2.2 ns, with explicit water molecules and periodic boundary conditions, were performed. The simulations reveal that the trypsin binding pocket residues are relatively rigid regardless of whether there is no ligand, a high-affinity ligand (benzamidine), or a low-affinity ligand (tranylcypromine). The thermal average of the conformations sampled by benzamidine bound to trypsin is planar and consistent with the planar internal geometry of the benzamidine crystallographic model coordinates. However, the most probable bound benzamidine conformations are +/-25 degrees out of plane, implying that the observed X-ray electron density represents an average of densities from two mirror symmetric, nonplanar conformations. Solvated benzamidine has free energy minima at +/-45 degrees , and the induction of a more planar geometry upon binding is associated with approximately 1 kcal/mol of intramolecular strain. Tranylcypromine's hydrogen-bonding pattern in the MD differs substantially from that inferred from the X-ray electron density. Early in simulations of this system, tranylcypromine adopts an alternative binding conformation, changing from the crystallographic conformation, with a direct hydrogen bond between its amino moiety and the backbone oxygen of Gly219, to one having a bridging water molecule. This result is consistently seen with the CHARMM22, Amber, or OPLS-AA force fields. The trypsin-tranylcypromine hydrogen-bonding pattern observed in the simulations also occurs as the crystallographic binding mode of the Lys15 side-chain of bovine pancreatic trypsin inhibitor bound to trypsin. In this latter cocrystal, a bridging crystallographic water does reside between the side-chain's amino group and the trypsin Gly219 backbone oxygen. Furthermore, the trypsin-tranylcypromine simulations sample two different stable noncrystallographic binding poses. These data suggest that some of the electron density ascribed to tranylcypromine in the X-ray model is rather due to a bound water molecule, and that multiple tranylcypromine binding conformations (crystallographic disorder) may be the cause of ambiguous electron density. The combined trypsin-benzamidine and trypsin- tranylcypromine results highlight the ability of simulations to augment protein-ligand complex structural data by deconvoluting the effects of thermal and structural averaging, and by finding energetically optimal ligand and bound water positions for weakly bound ligands.     

Mutagenic, electrochemical, and crystallographic investigation of the cytochrome b5 oxidation-reduction equilibrium: involvement of asparagine-57, serine-64, and heme propionate-7

A gene coding for lipase-solubilized bovine liver microsomal cytochrome b5 has been synthesized, expressed in Escherichia coli, and mutated at functionally critical residues. Characterization of the recombinant protein revealed that it has a reduction potential that is approximately 17 mV lower than that of authentic wild-type protein at pH 7 (25 degrees C). Structural studies determined that the recombinant protein differed in sequence from authentic wild-type cytochrome b5 owing to three errors in amidation status in the published sequence for the protein on which the gene synthesis was based. The structural origin of the lower reduction potential exhibited by the triple mutant has been investigated through X-ray crystallographic determination of the three-dimensional structure of this protein and is attributed to the presence of Asp-57 within 3.3 A of heme vinyl-4 in the mutant. In addition, the model developed by Argos and Mathews [Argos, P., & Mathews, F.S. (1975) J. Biol. Chem. 250, 747] for the change in cytochrome b5 oxidation state has been studied through mutation of Ser-64 to Ala. In this model, Ser-64 is postulated to stabilize the oxidized protein through H-bonding interactions with heme propionate-7 that orients this propionate group 6.2 A from the heme iron. Spectroelectrochemical studies of a mutant in which Ser-64 has been changed to an alanyl residue demonstrate that this protein has a reduction potential that is 7 mV lower than that of the wild-type protein; moreover, conversion of the heme propionate groups to the corresponding methyl esters increases the potential by 67 mV.(ABSTRACT TRUNCATED AT 250 WORDS)     

Structural characterization of the active and inactive states of Src kinase in solution by small-angle X-ray scattering

Src kinase plays an important role in several signaling and regulation mechanisms in vivo. Enzymatic activity is tightly regulated through the phosphorylation and dephosphorylation of tyrosine 527, which is placed at the C-terminal tail. Here, we have addressed domain rearrangements involved in the regulation mechanism of Src kinase in solution using small-angle X-ray scattering. In the phosphorylated wild-type form of Src kinase corresponding to the inactive state of the protein, a single conformation compatible with a closed crystallographic structure was found in solution. In the Y527F point mutant representing the active state, analysis of scattering data reveals an equilibrium between two differently populated conformations differing in the radius of gyration by 5 Å. The major species (85% of the total population) presents a closed conformation indistinguishable from the crystallographic structure of the inactive state. The minor species (15% of the total population) is an open conformation similar to the crystallographic structure in the active state. The latter structure has the SH3, SH2, and SH2-catalytic domain linker assembled as a pseudo-two-domain protein. The regulation model emerging from this study, including at least three different conformational states, allows the tight regulation of the enzyme without compromising fast response in the presence of natural targets.     

Oligomeric state of lipocalin-1 (LCN1) by multiangle laser light scattering and fluorescence anisotropy decay

Multiangle laser light scattering and fluorescence anisotropy decay measurements clarified the oligomeric states of native and recombinant tear lipocalin (lipocalin-1, TL). Native TL is monomeric. Recombinant TL (5-68 microM) with or without the histidine tag shows less than 7% dimer formation that is not in equilibrium with the monomeric form. Fluorescence anisotropy decay showed a correlation time of 9-10 ns for TL (10 microM-1 mM). Hydrodynamic calculations based on the crystallographic structure of a monomeric TL mutant closely concur with the observed correlation time. The solution properties calculated with HYDROPRO and SOLPRO programs from the available crystallographic structure of a monomeric TL mutant concur closely with the observed fluorescence anisotropy decay. The resulting model shows that protein topology is the major determinant of rotational correlation time and accounts for deviation from the Stokes-Einstein relation. The data challenge previous gel filtration studies to show that native TL exists predominantly as a monomer in solution rather than as a dimer. Delipidation of TL results in a formation of a complex oligomeric state (up to 25%). These findings are important as the dynamic processes in the tear film are limited by diffusional, translational as well as rotational, properties of the protein.     

Expression, crystallization and derivatization of the complete extracellular domain of the beta(c) subunit of the human IL-5, IL-3 and GM-CSF receptors.

The major signalling entity of the receptors for the haemopoietic cytokines granulocyte-macrophage colony stimulating factor (GM-CSF), interleukin-3 (IL-3) and interleukin-5 (IL-5) is the shared beta(c) receptor, which is activated by ligand-specific alpha receptors. The beta(c) subunit is a stable homodimer whose extracellular region consists of four fibronectin domains and appears to be a duplication of the cytokine receptor homology module. No four domain structure has been determined for this receptor family and the structure of the beta(c) subunit remains unknown. We have expressed the extracellular domain in insect cells using the baculovirus system, purified it to homogeneity and determined its N-terminal sequence. N-glycosylation at two sites was demonstrated. Crystals of the complete domain have been obtained that are suitable for X-ray crystallographic studies, following mutagenesis to remove one of the N-glycosylation sites. The rhombohedral crystals of space group R3, with unit cell dimensions 186.1 A and 103.5 A, diffracted to a resolution of 2.9 A using synchrotron radiation. Mutagenesis was also used to engineer cysteine substitution mutants which formed isomorphous Hg derivatives in order to solve the crystallographic phase problem. The crystal structure will help to elucidate how the beta(c) receptor is activated by heterodimerization with the respective alpha/ligand complexes.     

Enzyme kinetics of muscle glycogen phosphorylase b

Interest in the kinetics of glycogen phosphorylase has recently been renewed by the hypothesis of a glycogen shunt and by the potential of altering phosphorylase to treat type II diabetes. The wealth of data from studies of this enzyme in vitro and the need for a mathematical representation for use in the study of metabolic control systems make this enzyme an ideal subject for a mathematical model. We applied a two-part approach to the analysis of the kinetics of glycogen phosphorylase b (GPb). First, a continuous state model of enzyme-ligand interactions supported the view that two phosphates and four ATP or AMP molecules can bind to the enzyme, a result that agrees with spectroscopic and crystallographic studies. Second, using minimum error estimates from continuous state model fits to published data (that agreed well with reported error), we used a discrete state model of internal molecular events to show that GPb exists in three discrete states (two of which are inactive) and that state transitions are concerted. The results also show that under certain concentrations of substrate and effector, ATP can activate the enzyme, while under other conditions, it can competetively inhibit or noncompetitively inhibit the enzyme. This result is unexpected but is consistent with spectroscopic, crystallographic, and kinetic experiments and can explain several previously unexplained phenomena regarding GPb activity in vivo and in vitro.     

The simultaneous binding of lanthanide and N-acetylglucosamine inhibitors to hen egg-white lysozyme in solution by 1H and 13C nuclear magnetic resonance.

Lanthanide ions and the N-acetylglucosamine (GlcNAc) sugars are able to bind simultaneously to hen egg-white lysozyme (EC 3.2.1.17). The present study characterizes the properties of the ternary complexes with lysozyme, which involve up to seven paramagnetic lanthanides and two diamagnetic lanthanides, together with alpha GlcNAc, beta GlcNAc, alpha MeGlcNAc and beta MeGlcNAc. pH titrations and binding titrations of the GlcNAc sugars with lysozyme-La(III) complexes show that the GlcNAc sugars bind to at least two independent sites and that one of them competes with La(III) for binding to lysozyme. Given the known binding site of lanthanides at Asp-52 and Glu-35, the competitive binding site of GlcNAc is identified as subsite E. A simple analysis of the paramagnetic-lanthanide-induced shifts shows that the GlcNAc sugar binds in subsite C, in accordance with crystallographic results [Perkins, Johnson, Machin & Phillips (1979) Biochem. J. 181, 21-36]. This finding was refined by several computer analyses of the lanthanide-induced shifts of 17 proton and carbon resonances of beta MeGlcNAc. Good fits were obtained for all the signals, except for two that were affected by exchange broadening phenomena. No distinction could be made between a fit for a two-position model of Ln(III) binding with axial symmetry to lysozyme, according to the crystallographic result, or a one-position model with axial symmetry where the Ln(III) is positioned mid-way between Asp-52 and Glu-35. Although this work establishes the feasibility of lanthanide shift reagents for study of protein-ligand complexes, further work is required to establish the manner in which lanthanides bind to lysozyme in solution.     

Crystallographic and spectroscopic studies of peroxide-derived myoglobin compound II and occurrence of protonated FeIV O

High resolution crystal structures of myoglobin in the pH range 5.2-8.7 have been used as models for the peroxide-derived compound II intermediates in heme peroxidases and oxygenases. The observed Fe-O bond length (1.86-1.90 A) is consistent with that of a single bond. The compound II state of myoglobin in crystals was controlled by single-crystal microspectrophotometry before and after synchrotron data collection. We observe some radiation-induced changes in both compound II (resulting in intermediate H) and in the resting ferric state of myoglobin. These radiation-induced states are quite unstable, and compound II and ferric myoglobin are immediately regenerated through a short heating above the glass transition temperature (<1 s) of the crystals. It is unclear how this influences our compound II structures compared with the unaffected compound II, but some crystallographic data suggest that the influence on the Fe-O bond distance is minimal. Based on our crystallographic and spectroscopic data we suggest that for myoglobin the compound II intermediate consists of an Fe(IV)-O species with a single bond. The presence of Fe(IV) is indicated by a small isomer shift of delta = 0.07 mm/s from M?ssbauer spectroscopy. Earlier quantum refinements (crystallographic refinement where the molecular-mechanics potential is replaced by a quantum chemical calculation) and density functional theory calculations suggest that this intermediate H species is protonated.     

Conformational changes in nitric oxide synthases induced by chlorzoxazone and nitroindazoles: crystallographic and computational analyses of inhibitor potency

Nitric oxide is a key signaling molecule in many biological processes, making regulation of nitric oxide levels highly desirable for human medicine and for advancing our understanding of basic physiology. Designing inhibitors to specifically target one of the three nitric oxide synthase (NOS) isozymes that form nitric oxide from the L-Arg substrate poses a significant challenge due to the overwhelmingly conserved active sites. We report here 10 new X-ray crystallographic structures of inducible and endothelial NOS oxygenase domains cocrystallized with chlorzoxazone and four nitroindazoles: 5-nitroindazole, 6-nitroindazole, 7-nitroindazole, and 3-bromo-7-nitroindazole. Each of these bicyclic aromatic inhibitors has only one hydrogen bond donor and therefore cannot form the bidentate hydrogen bonds that the L-Arg substrate makes with Glu371. Instead, all of these inhibitors induce a conformational change in Glu371, creating an active site with altered molecular recognition properties. The cost of this conformational change is approximately 1-2 kcal, based on our measured constants for inhibitor binding to the wild-type and E371A mutant proteins. These inhibitors derive affinity by pi-stacking above the heme and replacing both intramolecular (Glu371-Met368) and intermolecular (substrate-Trp366) hydrogen bonds to the beta-sheet architecture underlying the active site. When bound to NOS, high-affinity inhibitors in this class are planar, whereas weaker inhibitors are nonplanar. Isozyme differences were observed in the pterin cofactor site, the heme propionate, and inhibitor positions. Computational docking predictions match the crystallographic results, including the Glu371 conformational change and inhibitor-binding orientations, and support a combined crystallographic and computational approach to isozyme-specific NOS inhibitor analysis and design.     

Crystal structure analysis and refinement at 2.5 A of hexameric C-phycocyanin from the cyanobacterium Agmenellum quadruplicatum. The molecular model and its implications for light-harvesting

The crystal structure of the light-harvesting protein-pigment complex C-phycocyanin from the cyanobacterium Agmenellum quadruplicatum has been determined by Patterson search techniques on the basis of the molecular model of C-phycocyanin from Mastigocladus laminosus. The crystal unit cell (space group P321) contains three (alpha beta)6 hexamers centred on the crystallographic triads. The hexamer at the origin of the unit cell exhibits crystallographic 32 point symmetry. The other two hexamers (independent of the former) show crystallographic 3-fold and local 2-fold symmetry. The 3-fold redundancy of the asymmetric unit of the crystal cell was used in the refinement process, which proceeded by cyclic averaging, model building and energy-restrained crystallographic refinement. Refinement was terminated with a conventional crystallographic R-value of 0.20 with data to 2.5 A resolution. The two independent hexamers of the unit cell are identical within the limits of error at all levels of aggregation. Two trimers, which closely resemble the M. laminosus C-phycocyanin, are aggregated head-to-head to form the hexamer. Both trimers fit complementarily and are held together by polar and ionic interactions. Conservation of the amino acid residues involved in protein-chromophore and intermonomer interactions suggests common structural features for all biliproteins. Most probably, the hexameric aggregation form present in the crystals is closely related to the discs of native phycobilisome rods. All tetrapyrrole chromophores are extended but with different geometries enforced by different protein surroundings. In particular, interactions of the propionic side-chains with arginine residues and of the pyrrole nitrogen atoms with aspartate residues define configuration and conformation of the chromophores. Relative chromophore distances and orientations have been determined and a preferential pathway for the energy transfer suggested. Accordingly, within a hexamer the absorbed energy is funneled to chromophore B84 and then transduced via B84 chromophores along the phycobilisome rods.     

Quantum mechanical modeling of aspartic proteinase interactions: difference in binding of diastereomeric statine models

Quantum mechanical calculations were carried out for the interaction of two diastereomeric model inhibitors of aspartic proteinases with a model for the active site, based on crystallographic coordinates of endothiapepsin. The model inhibitor is formamide N-(2-hydroxy 3-methyl propane) and the active site is represented by the full backbone and most of the side chains of the two partial sequences D32-T33-G34-S35 and D215-T216-G217-T218. Those calculations demonstrate that the best binding mode for this short inhibitor is consistent with the X-ray experiments and somewhat stronger with the inhibitor in a 2(S) configuration, compared to 2(R). Another binding mode is possible for this model inhibitor only in the 2(S)-configuration, and is weaker than the first.     

Crystal structure of bovine pancreatic phospholipase A2 covalently inhibited by p-bromo-phenacyl-bromide

Bovine pancreatic phospholipase A2 covalently inhibited by p-bromo-phenacyl-bromide was crystallized from 50% (v/v) 2-methyl-2,4-pentanediol. The space group was P3(1)21 with cell dimensions a = b = 46.73 A and c = 102.5 A (1 A = 0.1 nm). Diffraction data were collected by oscillation photography from one single crystal of dimensions 0.2 mm x 0.2 mm x 0.2 mm. The crystal structure was determined to a resolution of 2.5 A by crystallographic refinement of a starting model, which consisted of native bovine pancreatic phospholipase A2 positioned and oriented in the P3(1)21 cell as in the bovine pro-phospholipase A2. The crystallographic R-factor decreased from 0.378 to 0.197 after 70 refinement cycles. For the greater part the three-dimensional structure was very similar to that of native phospholipase. The inhibitor group shows up clearly. However, as in solution, there is no calcium ion bound any more in the active site, and this causes a significant conformational change in the loop from residue 59 to 73. This loop is remote from the calcium binding site. Interestingly, this is the same loop that also shows different conformations in other phospholipase A2 molecules. The inhibitor molecule has hydrophobic interactions with Phe5 and Cys45. Rational design of specific and potent inhibitors of phospholipase A2 catalysis is discussed on the basis of the present three-dimensional structure.