Extending the Applicability of the Nonlinear Poisson−Boltzmann Equation: Multiple Dielectric Constants and Multivalent Ions

Poisson-Boltzmann calculations of nonspecific salt effects on protein-protein binding free energies

The salt dependence of the binding free energy of five protein-protein hetero-dimers and two homo-dimers/tetramers was calculated from numerical solutions to the Poisson-Boltzmann equation. Overall, the agreement with experimental values is very good. In all cases except one involving the highly charged lactoglobulin homo-dimer, increasing the salt concentration is found both experimentally and theoretically to decrease the binding affinity. To clarify the source of salt effects, the salt-dependent free energy of binding is partitioned into screening terms and to self-energy terms that involve the interaction of the charge distribution of a monomer with its own ion atmosphere. In six of the seven complexes studied, screening makes the largest contribution but self-energy effects can also be significant. The calculated salt effects are found to be insensitive to force-field parameters and to the internal dielectric constant assigned to the monomers. Nonlinearities due to high charge densities, which are extremely important in the binding of proteins to negatively charged membrane surfaces and to nucleic acids, make much smaller contributions to the protein-protein complexes studied here, with the exception of highly charged lactoglobulin dimers. Our results indicate that the Poisson-Boltzmann equation captures much of the physical basis of the nonspecific salt dependence of protein-protein complexation.

On the lower susceptibility of oseltamivir to influenza neuraminidase subtype N1 than those in N2 and N9

Aiming to understand, at the molecular level, why oseltamivir (OTV) cannot be used for inhibition of human influenza neuraminidase subtype N1 as effectively as for subtypes N2 and N9, molecular dynamics simulations were carried out for the three complexes, OTV-N1, OTV-N2, and OTV-N9. The three-dimensional OTV-N2 and OTV-N9 initial structures were represented by the x-ray structures, whereas that of OTV-N1, whose x-ray structure is not yet solved, was built up using the aligned sequence of H5N1 isolated from humans in Thailand with the x-ray structure of the N2-substrate as the template. In comparison to the OTV-N2 and OTV-N9 complexes, dramatic changes were observed in the OTV conformation in the OTV-N1 complex in which two of its bulky side chains, N-acethyl (-NHAc) and 1-ethylproxy group (-OCHEt2), were rotated to adjust the size to fit into the N1 catalytic site. This change leads directly to the rearrangements of the OTV's environment, which are i), distances to its neighbors, W-178 and E-227, are shorter whereas those to residues R-224, E-276, and E-292 are longer; ii), hydrogen bonds to the two nearest neighbors, R-224 and E-276, are still conserved in distance and number as well as percentage occupation; iii), the calculated ligand/enzyme binding free energies of -7.20, -13.44, and -13.29 kcal/mol agree with their inhibitory activities in terms of the experimental IC50 of 36.1-53.2 nM, 1.9-2.7 nM, and 9.5-17.7 nM for the OTV-N1, OTV-N2, and OTV-N9 complexes, respectively; and iv), hydrogen-bond breaking and creation between the OTV and neighborhood residues are accordingly in agreement with the ligand solvation/desolvation taking place in the catalytic site.

Electrostatically driven protein aggregation: beta-lactoglobulin at low ionic strength

The aggregation of beta-lactoglobulin (BLG) at ambient temperature was studied using turbidimetry and dynamic light scattering in the range 3.8<pH<5.2 in 0.0045 M NaCl, and in the ionic strength range 0.0045-0.5 M at fixed pH=5.0. The initial rate of aggregation, taken as the initial slope of turbidity vs time, (dtau/dt)0, indicated maximum aggregation near pH 4.6 (below the isoelectric point of 5.2), but the dependence of the initial rate of aggregation on pH was highly asymmetric. At pH 5.0, (dtau/dt)0 strongly increased with a decrease in ionic strength I from 0.1 to 0.0045 M and was found to be nearly linear with 1/I. DLS measurements revealed an increase in particle size with time, with the appearance of bimodal distributions in which the fast and slow modes corresponded, respectively, to a BLG dimer and to larger aggregates in the 100-800 nm range. At conditions of slower aggregation, DLS revealed the consumption of dimers to form higher order aggregates with no intermediate species. Computer modeling (Delphi) was used to visualize the electrostatic potential around the dimer to elucidate the pH and ionic strength dependence of the initial aggregation rates. The aggregation process appears to comprise an initial fast consumption of the dimer, whose dependence on pH and I arises from the interaction of the positive and negative domains of interacting dimers, followed by the slow formation of much larger aggregates with relatively little sensitivity to pH and I. The open-ended nature of BLG aggregation is thought to arise from the asymmetry of the dimer charge distribution.

Solution Structure of the Hdm2 C2H2C4 RING, a Domain Critical for Ubiquitination of p53

Regulation of the transcriptional response to the tumor suppressor p53 occurs at many levels, including control of its transcriptional activity, and of its stability and concentration within the cell. p53 stability is regulated by the protein Hdm2, an E3 ubiquitin ligase that binds to p53 and promotes its ubiquitination and degradation. The C-terminal domain of Hdm2, which is critical for this activity, has been classified as a RING domain on the basis of sequence homology, although it lacks the canonical set of zinc ligands (RING domains typically have C3HC4 or C4C4 zinc coordination). Here, we report the solution structure of the C2H2C4 RING domain of Hdm2(429-491), which reveals a symmetrical dimer with a unique cross-brace zinc-binding scheme. Each subunit has one Cys4 Zn site and one His2Cys2 Zn site. The global fold of each subunit is similar to those reported for other RING domains, with a compact betabetaalphabeta fold, a small hydrophobic core, and two Zn ions, which are essential for maintaining the domain structure. The dimer structure is maintained by an extensive interface that buries a large hydrophobic area on each subunit. It has been proposed that Hdm2 and its homologue HdmX form a stable heterodimer through their RING domains, resulting in a synergistic increase in observed E3 activity. To test this proposal, we prepared an HdmX RING construct and showed by NMR titration that it forms a tight 1:1 complex with the Hdm2 RING. The resonances most perturbed by heterodimer formation are located within the subunit interface of the homodimer, far removed from the surface expected to form the docking site of the E2 ubiquitin-conjugating enzyme, providing a structure-based rationale for the function of the RING domains in p53 ubiquitination.

Modeling electrostatic effects in proteins

Electrostatic energies provide what is perhaps the most effective tool for structure-function correlation of biological molecules. This review considers the current state of simulations of electrostatic energies in macromolecules as well as the early developments of this field. We focus on the relationship between microscopic and macroscopic models, considering the convergence problems of the microscopic models and the fact that the dielectric 'constants' in semimacroscopic models depend on the definition and the specific treatment. The advances and the challenges in the field are illustrated considering a wide range of functional properties including pK(a)'s, redox potentials, ion and proton channels, enzyme catalysis, ligand binding and protein stability. We conclude by pointing out that, despite the current problems and the significant misunderstandings in the field, there is an overall progress that should lead eventually to quantitative descriptions of electrostatic effects in proteins and thus to quantitative descriptions of the function of proteins.

The role of electrostatic interactions in the antitumor activity of dimeric RNases

The cytotoxic action of some ribonucleases homologous to bovine pancreatic RNase A, the superfamily prototype, has interested and intrigued investigators. Their ribonucleolytic activity is essential for their cytotoxic action, and their target RNA is in the cytosol. It has been proposed that the cytosolic RNase inhibitor (cRI) plays a major role in determining the ability of an RNase to be cytotoxic. However, to interact with cRI RNases must reach the cytosol, and cross intracellular membranes. To investigate the interactions of cytotoxic RNases with membranes, cytotoxic dimeric RNases resistant, or considered to be resistant to cRI, were assayed for their effects on negatively charged membranes. Furthermore, we analyzed the electrostatic interaction energy of the RNases complexed in silico with a model membrane. The results of this study suggest that close correlations can be recognized between the cytotoxic action of a dimeric RNase and its ability to complex and destabilize negatively charged membranes.

High resolution structure of BipD: an invasion protein associated with the type III secretion system of Burkholderia pseudomallei

Burkoldheria pseudomallei is a Gram-negative bacterium that possesses a protein secretion system similar to those found in Salmonella and Shigella. Recent work has indicated that the protein encoded by the BipD gene of B. pseudomallei is an important secreted virulence factor. BipD is similar in sequence to IpaD from Shigella and SipD from Salmonella and is therefore likely to be a translocator protein in the type-III secretion system of B. pseudomallei. The crystal structure of BipD has been solved at a resolution of 2.1 A revealing the detailed tertiary fold of the molecule. The overall structure is appreciably extended and consists of a bundle of antiparallel alpha-helical segments with two small beta-sheet regions. The longest helices of the molecule form a four-helix bundle and most of the remaining secondary structure elements (three helices and two three-stranded beta-sheets) are formed by the region linking the last two helices of the four-helix bundle. The structure suggests that the biologically active form of the molecule may be a dimer formed by contacts involving the C-terminal alpha-helix, which is the most strongly conserved part of the protein. Comparison of the structure of BipD with immunological and other data for IpaD indicates that the C-terminal alpha-helix is also involved in contacts with other proteins that form the translocon.

Insights into protein-protein interfaces using a Bayesian network prediction method

Identifying the interface between two interacting proteins provides important clues to the function of a protein, and is becoming increasing relevant to drug discovery. Here, surface patch analysis was combined with a Bayesian network to predict protein-protein binding sites with a success rate of 82% on a benchmark dataset of 180 proteins, improving by 6% on previous work and well above the 36% that would be achieved by a random method. A comparable success rate was achieved even when evolutionary information was missing, a further improvement on our previous method which was unable to handle incomplete data automatically. In a case study of the Mog1p family, we showed that our Bayesian network method can aid the prediction of previously uncharacterised binding sites and provide important clues to protein function. On Mog1p itself a putative binding site involved in the SLN1-SKN7 signal transduction pathway was detected, as was a Ran binding site, previously characterized solely by conservation studies, even though our automated method operated without using homologous proteins. On the remaining members of the family (two structural genomics targets, and a protein involved in the photosystem II complex in higher plants) we identified novel binding sites with little correspondence to those on Mog1p. These results suggest that members of the Mog1p family bind to different proteins and probably have different functions despite sharing the same overall fold. We also demonstrated the applicability of our method to drug discovery efforts by successfully locating a number of binding sites involved in the protein-protein interaction network of papilloma virus infection. In a separate study, we attempted to distinguish between the two types of binding site, obligate and non-obligate, within our dataset using a second Bayesian network. This proved difficult although some separation was achieved on the basis of patch size, electrostatic potential and conservation. Such was the similarity between the two interacting patch types, we were able to use obligate binding site properties to predict the location of non-obligate binding sites and vice versa.

Detailed microscopic study of the full zipA:FtsZ interface

Protein-protein interaction networks are very important for a wide range of biological processes. Crystallographic structures and mutational studies have generated a large number of information that allowed the discovery of energetically important determinants of specificity at intermolecular protein interfaces and the understanding of the structural and energetic characteristics of the binding hot spots. In this study we have used the improved MMPB/SA (molecular mechanics/Poisson-Boltzmann surface area) approach that combining molecular mechanics and continuum solvent permits to calculate the free energy differences upon alanine mutation. For a better understanding of the binding determinants of the complex formed between the FtsZ fragment and ZipA we extended the alanine scanning mutagenesis study to all interfacial residues of this complex. As a result, we present new mutations that allowed the discovery of residues for which the binding free energy differences upon alanine mutation are higher than 2.0 kcal/mol. We also observed the formation of a hydrophobic pocket with a high warm spot spatial complementarity between FtsZ and ZipA. Small molecules could be designed to bind to these amino acid residues hindering the binding of FtsZ to ZipA. Hence, these mutational data can be used to design new drugs to control more efficiently bacterial infections.

A hydrophobic gate in an ion channel: the closed state of the nicotinic acetylcholine receptor

The nicotinic acetylcholine receptor (nAChR) is the prototypic member of the 'Cys-loop' superfamily of ligand-gated ion channels which mediate synaptic neurotransmission, and whose other members include receptors for glycine, gamma-aminobutyric acid and serotonin. Cryo-electron microscopy has yielded a three-dimensional structure of the nAChR in its closed state. However, the exact nature and location of the channel gate remains uncertain. Although the transmembrane pore is constricted close to its center, it is not completely occluded. Rather, the pore has a central hydrophobic zone of radius about 3 A. Model calculations suggest that such a constriction may form a hydrophobic gate, preventing movement of ions through a channel. We present a detailed and quantitative simulation study of the hydrophobic gating model of the nicotinic receptor, in order to fully evaluate this hypothesis. We demonstrate that the hydrophobic constriction of the nAChR pore indeed forms a closed gate. Potential of mean force (PMF) calculations reveal that the constriction presents a barrier of height about 10 kT to the permeation of sodium ions, placing an upper bound on the closed channel conductance of 0.3 pS. Thus, a 3 A radius hydrophobic pore can form a functional barrier to the permeation of a 1 A radius Na+ ion. Using a united-atom force field for the protein instead of an all-atom one retains the qualitative features but results in differing conductances, showing that the PMF is sensitive to the detailed molecular interactions.

Simple electrostatic model improves designed protein sequences

Electrostatic interactions are important for both protein stability and function, including binding and catalysis. As protein design moves into these areas, an accurate description of electrostatic energy becomes necessary. Here, we show that a simple distance-dependent Coulombic function parameterized by a comparison to Poisson-Boltzmann calculations is able to capture some of these electrostatic interactions. Specifically, all three helix N-capping interactions in the engrailed homeodomain fold are recovered using the newly parameterized model. The stability of this designed protein is similar to a protein forced by sequence restriction to have beneficial electrostatic interactions.

Factors influencing the energetics of electron and proton transfers in proteins. What can be learned from calculations

A protein structure should provide the information needed to understand its observed properties. Significant progress has been made in developing accurate calculations of acid/base and oxidation/reduction reactions in proteins. Current methods and their strengths and weaknesses are discussed. The distribution and calculated ionization states in a survey of proteins is described, showing that a significant minority of acidic and basic residues are buried in the protein and that most of these remain ionized. The electrochemistry of heme and quinones are considered. Proton transfers in bacteriorhodopsin and coupled electron and proton transfers in photosynthetic reaction centers, 5-coordinate heme binding proteins and cytochrome c oxidase are highlighted as systems where calculations have provided insight into the reaction mechanism.

Benchmarking pK(a) prediction

Background

pK_a values are a measure of the protonation of ionizable groups in proteins. Ionizable groups are involved in intra-protein, protein-solvent and protein-ligand interactions as well as solubility, protein folding and catalytic activity. The pK_a shift of a group from its intrinsic value is determined by the perturbation of the residue by the environment and can be calculated from three-dimensional structural data.

Results

Here we use a large dataset of experimentally-determined pK_as to analyse the performance of different prediction techniques. Our work provides a benchmark of available software implementations: MCCE, MEAD, PROPKA and UHBD. Combinatorial and regression analysis is also used in an attempt to find a consensus approach towards pK_a prediction. The tendency of individual programs to over- or underpredict the pK_a value is related to the underlying methodology of the individual programs.

Conclusion

Overall, PROPKA is more accurate than the other three programs. Key to developing accurate predictive software will be a complete sampling of conformations accessible to protein structures.

Continuum Solvation Models in the Linear Interaction Energy Method

The linear interaction energy (LIE) method in combination with two different continuum solvent models has been applied to calculate protein-ligand binding free energies for a set of inhibitors against the malarial aspartic protease plasmepsin II. Ligand-water interaction energies are calculated from both Poisson-Boltzmann (PB) and Generalized Born (GB) continuum models using snapshots from explicit solvent simulations of the ligand and protein-ligand complex. These are compared to explicit solvent calculations, and we find close agreement between the explicit water and PB solvation models. The GB model overestimates the change in solvation energy, and this is caused by consistent underestimation of the effective Born radii in the protein-ligand complex. The explicit solvent LIE calculations and LIE-PB, with our standard parametrization, reproduce absolute experimental binding free energies with an average unsigned error of 0.5 and 0.7 kcal/mol, respectively. The LIE-GB method, however, requires a constant offset to approach the same level of accuracy.

Predicting DNA-binding sites of proteins from amino acid sequence

Background

Understanding the molecular details of protein-DNA interactions is critical for deciphering the mechanisms of gene regulation. We present a machine learning approach for the identification of amino acid residues involved in protein-DNA interactions.

Results

We start with a Naïve Bayes classifier trained to predict whether a given amino acid residue is a DNA-binding residue based on its identity and the identities of its sequence neighbors. The input to the classifier consists of the identities of the target residue and 4 sequence neighbors on each side of the target residue. The classifier is trained and evaluated (using leave-one-out cross-validation) on a non-redundant set of 171 proteins. Our results indicate the feasibility of identifying interface residues based on local sequence information. The classifier achieves 71% overall accuracy with a correlation coefficient of 0.24, 35% specificity and 53% sensitivity in identifying interface residues as evaluated by leave-one-out cross-validation. We show that the performance of the classifier is improved by using sequence entropy of the target residue (the entropy of the corresponding column in multiple alignment obtained by aligning the target sequence with its sequence homologs) as additional input. The classifier achieves 78% overall accuracy with a correlation coefficient of 0.28, 44% specificity and 41% sensitivity in identifying interface residues. Examination of the predictions in the context of 3-dimensional structures of proteins demonstrates the effectiveness of this method in identifying DNA-binding sites from sequence information. In 33% (56 out of 171) of the proteins, the classifier identifies the interaction sites by correctly recognizing at least half of the interface residues. In 87% (149 out of 171) of the proteins, the classifier correctly identifies at least 20% of the interface residues. This suggests the possibility of using such classifiers to identify potential DNA-binding motifs and to gain potentially useful insights into sequence correlates of protein-DNA interactions.

Conclusion

Naïve Bayes classifiers trained to identify DNA-binding residues using sequence information offer a computationally efficient approach to identifying putative DNA-binding sites in DNA-binding proteins and recognizing potential DNA-binding motifs.

A generalized Born formalism for heterogeneous dielectric environments: application to the implicit modeling of biological membranes

Reliable computer simulations of complex biological environments such as integral membrane proteins with explicit water and lipid molecules remain a challenging task. We propose a modification of the standard generalized Born theory of homogeneous solvent for modeling the heterogeneous dielectric environments such as lipid/water interfaces. Our model allows the representation of biological membranes in the form of multiple layered dielectric regions with dielectric constants that are different from the solute cavity. The proposed new formalism is shown to predict the electrostatic component of solvation free energy with a relative error of 0.17% compared to exact finite-difference solutions of the Poisson equation for a transmembrane helix test system. Molecular dynamics simulations of melittin and bacteriorhodopsin are carried out and performed over 10 ns and 7 ns of simulation time, respectively. The center of melittin along the membrane normal in these stable simulations is in excellent agreement with the relevant experimental data. Simulations of bacteriorhodopsin started from the experimental structure remained stable and in close agreement with experiment. We also examined the free energy profiles of water and amino acid side chain analogs upon membrane insertion. The results with our implicit membrane model agree well with the experimental transfer free energy data from cyclohexane to water as well as explicit solvent simulations of water and selected side chain analogs.

Unraveling the Importance of Protein−Protein Interaction: Application of a Computational Alanine-Scanning Mutagenesis to the Study of the IgG1 Streptococcal Protein G (C2 Fragment) Complex

Alanine-scanning mutagenesis of protein-protein interfacial residues is a very important process for rational drug design. In this study, we have used the improved MM-PBSA approach that combining molecular mechanics and continuum solvent permits one to calculate the free energy differences through alanine mutation. To identify the binding determinants of the complex formed between the IgG1 (immunoglobulin-binding protein G) and protein G, we have extended the experimental alanine scanning mutagenesis study to both proteins of this complex and, therefore, to all interfacial residues of this binding complex. As a result, we present new residues that can be characterized as warm spots and, therefore, are important for complex formation. We have further increased the understanding of the functionality of this improved computational alanine-scanning mutagenesis approach testing its sensitivity to a protein-protein complex with an interface made up of residues mainly polar. In this study, we also have improved the method for the detection of an important amino acid residue that frequently constitutes a hot spot--tryptophan.

High-resolution Structures of Escherichia coli cDsbD in Different Redox States: A Combined Crystallographic, Biochemical and Computational Study

Escherichia coli DsbD transports electrons from cytoplasmic thioredoxin to periplasmic target proteins. DsbD is composed of an N-terminal (nDsbD) and a C-terminal (cDsbD) periplasmic domain, connected by a central transmembrane domain. Each domain possesses two cysteine residues essential for electron transport. The transport proceeds via disulfide exchange reactions from cytoplasmic thioredoxin to the central transmembrane domain and via cDsbD to nDsbD, which then reduces the periplasmic target proteins. We determined four high-resolution structures of cDsbD: oxidized (1.65 A resolution), chemically reduced (1.3 A), photo-reduced (1.1 A) and chemically reduced at pH increased from 4.6 to 7. The latter structure was refined at 0.99 A resolution, the highest achieved so far for a thioredoxin superfamily member. The data reveal unprecedented structural details of cDsbD, demonstrating that the domain is very rigid and undergoes hardly any conformational change upon disulfide reduction or interaction with nDsbD. In full agreement with the crystallographic results, guanidinium chloride-induced unfolding and refolding experiments indicate that oxidized and reduced cDsbD are equally stable. We confirmed the structural rigidity of cDsbD by molecular dynamics simulations. A remarkable feature of cDsbD is the pKa of 9.3 for the active site Cys461: this value, determined using two different experimental methods, surprisingly was around 2.5 units higher than expected on the basis of the redox potential. Additionally, taking advantage of the very high quality of the cDsbD structures, we carried out pKa calculations, which gave results in agreement with the experimental findings. In conclusion, our wide-scope analysis of cDsbD, encompassing atomic-resolution crystallography, computational chemistry and biophysical measurements, highlighted two so far unrecognized key aspects of this domain: its unusual redox properties and extreme rigidity. Both are likely to be correlated to the role of cDsbD as a covalently linked electron shuttle between the membrane domain and the N-terminal periplasmic domain of DsbD.

Insights on HIV-1 Tat:P/CAF bromodomain molecular recognition from in vivo experiments and molecular dynamics simulations

Structural and functional studies indicate that, through its bromodomain, the cellular acetyltransferase P/CAF binds the acetylated Tat protein of human immunodeficiency virus type 1 (HIV-1) and promotes transcriptional activation of the integrated provirus. Based on the NMR structure of P/CAF complexed with an acetylated Tat peptide, here we use molecular dynamics simulations to construct a model describing the interaction between full length Tat and the P/CAF bromodomain. Our calculations show that the protein-protein interface involves hydrophobic interactions between the P/CAF ZA loop and the Tat core domain. In particular, tyrosines 760 and 761 of P/CAF, two residues that are highly conserved in most known bromodomains, play an essential role for the binding. Fluorescence resonance energy transfer (FRET) experiments performed in this work demonstrate that P/CAF proteins in which these tyrosines are mutated into hydrophilic residues neither bind to Tat inside the cells nor mediate Tat transactivation. The combination of theoretical and in vivo studies provides new insights into the specificity of bromodomain recognition.

Surface Enhanced Second Harmonic Generation from Macrocycle, Catenane, and Rotaxane Thin Films: Experiments and Theory

Surface enhanced second harmonic generation (SE SHG) experiments on molecular structures, macrocycles, catenanes, and rotaxanes, deposited as monolayers and multilayers by vacuum sublimation on silver, are reported. The measurements show that the molecules form ordered thin films, where the highest degree of order is observed in the case of macrocycle monolayers and the lowest in the case of rotaxane multilayers. The second harmonic generation activity is interpreted in terms of electric field induced second harmonic (EFISH) generation where the electric field is created by the substrate silver atoms. The measured second order nonlinear optical susceptibility for a rotaxane thin film is compared with that obtained by considering only EFISH contribution to SHG intensity. The electric field on the surface of a silver layer is calculated by using the Delphi4 program for structures obtained with TINKER molecular mechanics/dynamics simulations. An excellent agreement is observed between the calculated and the measured SHG susceptibilities.

Characterization of retinaldehyde dehydrogenase 3

RALDH3 (retinal dehydrogenase 3) was characterized by kinetic and binding studies, protein engineering, homology modelling, ligand docking and electrostatic-potential calculations. The major recognition determinant of an RALDH3 substrate was shown to be an eight-carbon chain bonded to the aldehyde group whose kinetic influence (kcat/Km at pH 8.5) decreases when shortened or lengthened. Surprisingly, the b-ionone ring of all-trans-retinal is not a major recognition site. The dissociation constants (Kd) of the complexes of RALDH3 with octanal, NAD+ and NADH were determined by intrinsic tryptophan fluorescence. The similarity of the Kd values for the complexes with NAD+ and with octanal suggests a random kinetic mechanism for RALDH3, in contrast with the ordered sequential mechanism often associated with aldehyde dehydrogenase enzymes. Inhibition of RALDH3 by tri-iodothyronine binding in competition with NAD+, predicted by the modelling, was established kinetically and by immunoprecipitation. Mechanistic implications of the kinetically influential ionizations with macroscopic pKa values of 5.0 and 7.5 revealed by the pH-dependence of kcat are discussed. Analogies with data for non-phosphorylating glyceraldehyde-3-phosphate dehydrogenase from Streptococcus mutans, together with the present modelled structure of the thioacyl RALDH3, suggest (a) that kcat characterizes deacylation of this intermediate for specific substrates and (b) the assignment of the pKa of the major ionization (approximating to 7.5) to the perturbed carboxy group of Glu280 whose conjugate base is envisaged as supplying general base catalysis to attack of a water molecule. The macroscopic pKa of the minor ionization (5.0) is considered to approximate to that of the carboxy group of Glu488.

Predicting protein interaction sites: binding hot-spots in protein–protein and protein–ligand interfaces

MOTIVATION

Protein assemblies are currently poorly represented in structural databases and their structural elucidation is a key goal in biology. Here we analyse clefts in protein surfaces, likely to correspond to binding 'hot-spots', and rank them according to sequence conservation and simple measures of physical properties including hydrophobicity, desolvation, electrostatic and van der Waals potentials, to predict which are involved in binding in the native complex.

RESULTS

The resulting differences between predicting binding-sites at protein-protein and protein-ligand interfaces are striking. There is a high level of prediction accuracy (< or =93%) for protein-ligand interactions, based on the following attributes: van der Waals potential, electrostatic potential, desolvation and surface conservation. Generally, the prediction accuracy for protein-protein interactions is lower, with the exception of enzymes. Our results show that the ease of cleft desolvation is strongly predictive of interfaces and strongly maintained across all classes of protein-binding interface.

Structure of the Regulatory Subunit of Acetohydroxyacid Synthase Isozyme III from Escherichia coli

The enzyme acetohydroxyacid synthase (AHAS) catalyses the first common step in the biosynthesis of the three branched-chain amino acids. Enzymes in the AHAS family generally consist of regulatory and catalytic subunits. Here, we describe the first crystal structure of an AHAS regulatory subunit, the ilvH polypeptide, determined at a resolution of 1.75 A. IlvH is the regulatory subunit of one of three AHAS isozymes expressed in Escherichia coli, AHAS III. The protein is a dimer, with two beta alpha beta beta alpha beta ferredoxin domains in each monomer. The two N-terminal domains assemble to form an ACT domain structure remarkably close to the one predicted by us on the basis of the regulatory domain of 3-phosphoglycerate dehydrogenase (3PGDH). The two C-terminal domains combine so that their beta-sheets are roughly positioned back-to-back and perpendicular to the extended beta-sheet of the N-terminal ACT domain. On the basis of the properties of mutants and a comparison with 3PGDH, the effector (valine) binding sites can be located tentatively in two symmetrically related positions in the interface between a pair of N-terminal domains. The properties of mutants of the ilvH polypeptide outside the putative effector-binding site provide further insight into the functioning of the holoenzyme. The results of this study open avenues for further studies aimed at understanding the mechanism of regulation of AHAS by small-molecule effectors.

Fluoride inhibition of enolase: crystal structure and thermodynamics

Enolase is a dimeric metal-activated metalloenzyme which uses two magnesium ions per subunit: the strongly bound conformational ion and the catalytic ion that binds to the enzyme-substrate complex inducing catalysis. The crystal structure of the human neuronal enolase-Mg2F2P(i) complex (enolase fluoride/phosphate inhibitory complex, EFPIC) determined at 1.36 A resolution shows that the combination of anions effectively mimics an intermediate state in catalysis. The phosphate ion binds in the same site as the phosphate group of the substrate/product, 2-phospho-D-glycerate/phosphoenolpyruvate, and induces binding of the catalytic Mg2+ ion. One fluoride ion bridges the structural and catalytic magnesium ions while the other interacts with the structural magnesium ion and the ammonio groups of Lys 342 and Lys 393. These fluoride ion positions correspond closely to the positions of the oxygen atoms of the substrate's carboxylate moiety. To relate structural changes resulting from fluoride, phosphate, and magnesium ions binding to those that are induced by phosphate and magnesium ions alone, we also determined the structure of the human neuronal enolase-Mg2P(i) complex (enolase phosphate inhibitory complex, EPIC) at 1.92 A resolution. It shows the closed conformation in one subunit and a mixture of open and semiclosed conformations in the other. The EPFIC dimer is essentially symmetric while the EPIC dimer is asymmetric. Isothermal titration calorimetry data confirmed binding of four fluoride ions per dimer and yielded Kb values of 7.5 x 10(5) +/- 1.3 x 10(5), 1.2 x 10(5) +/- 0.2 x 10(5), 8.6 x 10(4) +/- 1.6 x 10(4), and 1.6 x 10(4) +/- 0.7 x 10(4) M(-1). The different binding constants indicate negative cooperativity between the subunits; the asymmetry of EPIC supports such an interpretation.

The structure of the TRAPP subunit TPC6 suggests a model for a TRAPP subcomplex

The TRAPP (transport protein particle) complexes are tethering complexes that have an important role at the different steps of vesicle transport. Recently, the crystal structures of the TRAPP subunits SEDL and BET3 have been determined, and we present here the 1.7 Angstroms crystal structure of human TPC6, a third TRAPP subunit. The protein adopts an alpha/beta-plait topology and forms a dimer. In spite of low sequence similarity, the structure of TPC6 strikingly resembles that of BET3. The similarity is especially prominent at the dimerization interfaces of the proteins. This suggests heterodimerization of TPC6 and BET3, which is shown by in vitro and in vivo association studies. Together with TPC5, another TRAPP subunit, TPC6 and BET3 are supposed to constitute a family of paralogous proteins with closely similar three-dimensional structures but little sequence similarity among its members.

Experimental and theoretical study of the adsorption of fumaramide [2]rotaxane on Au(111) and Ag(111) surfaces

Thin films of fumaramide [2]rotaxane, a mechanically interlocked molecule composed of a macrocycle and a thread in a "bead and thread" configuration, were prepared by vapor deposition on both Ag(111) and Au(111) substrates. X-ray photoelectron spectroscopy (XPS) and high-resolution electron-energy-loss spectroscopy were used to characterize monolayer and bulklike multilayer films. XPS determination of the relative amounts of carbon, nitrogen, and oxygen indicates that the molecule adsorbs intact. On both metal surfaces, molecules in the first adsorbed layer show an additional component in the C 1s XPS line attributed to chemisorption via amide groups. Molecular-dynamics simulation indicates that the molecule orients two of its eight phenyl rings, one from the macrocycle and one from the thread, in a parallel bonding geometry with respect to the metal surfaces, leaving three amide groups very close to the substrate. In the case of fumaramide [2]rotaxane adsorption on Au(111), the presence of certain out-of-plane phenyl ring and Au-O vibrational modes points to such bonding and a preferential molecular orientation. The theoretical and experimental results imply that the three-dimensional intermolecular configuration permits chemisorption at low coverage to be driven by interactions between the three amide functions of fumaramide [2]rotaxane and the Ag(111) or Au(111) surface.

A molecular model of the Vibrio cholerae cytolysin transmembrane pore

The cytotoxic activity of some of the most pathogenic strains of Vibrio cholerae is associated with a cytolysin protein (VCC), which forms oligomeric transmembrane pores and changes the permeability of intestinal cells. We present here a model structure of the transmembrane pore of VCC based on sequence comparison with other pore-forming toxins. VCC is suggested to form a transmembrane beta-barrel pore with a relatively large trans vestibule region. Calculations of the electrostatic profile within the pore lumen provide a rationale for the low conductance and selectivity of the VCC ion channel.

Molecular dynamics and free energy calculations of the B and Z forms of C8-arylguanine modified oligonucleotides

Arylhydrazines found in the mushroom Agaricus bisporus have been shown to be carcinogenic. Upon metabolic activation, arylhydrazines are transformed into aryl radicals, forming 8-arylpurines, which may play a role in arylhydrazine carcinogenesis. These adducts are poorly read and inhibit chain extension but do alter the conformational preferences of oligonucleotides. We have shown that C8-phenylguanine modification of d(CGCGCG*CGCG) (G*= 8-phenylguanine) stabilizes it in the Z-DNA conformation (B/Z-DNA=1:1, 200 mM NaCl, pH 7.4). Here we have conducted molecular dynamics and free energy calculations to determine the sources(s) of these conformational affects and to predict the affect of the related C8-tolyl and C8-hydroxymethylphenyl guanine adducts on B/Z-DNA equilibrium. Force field parameters for the modified guanines were first developed using Guassian98 employing the B3LYP method and the standard 6-31G* basis set and fit to the Cornell 94 force field with RESP. Molecular dynamics simulations and free energy calculations, using the suite of programs contained in Amber 6 and 7 with the Cornell 94 force field, were used to determine the structural and thermodynamic properties of the DNA. The principal factors that drive conformation are stacking of the aryl group over the 5'-cytosine in the phenyl and tolyl modified oligonucleotides while hydrogen bonding opposes stacking in the hydroxymethylphenyl derivative. The phenyl and tolyl-modified DNA's favored the Z-DNA form as did the hydroxymethylphenyl derivative when hydrogen bonding was not present. The B-DNA conformation was preferred by the unmodified oligonucleotide and by the hydroxymethylphenyl-modified oligonucleotide when hydrogen bonding was considered. Z-DNA stability was not found to directly correlated with carcinogenicity and additional biological factors, such as recognition and repair, may also need to be considered in addition to Z-DNA formation.

Conformational pathway for the kissing complex-->extended dimer transition of the SL1 stem-loop from genomic HIV-1 RNA as monitored by targeted molecular dynamics techniques

HIV-1 retroviral genomic RNA dimerization is initiated by loop-loop interactions between the SL1 stem-loops of two identical RNA molecules. The SL1-SL1 unstable resulting kissing complex (KC) then refolds irreversibly into a more stable complex called extended dimer (ED). Although the structures of both types of complex have been determined, very little is known about the conformational pathway corresponding to the transition, owing to the difficulty of observing experimentally intermediate conformations. In this study, we applied targeted molecular dynamics simulation techniques (TMD) to the phosphorus atoms for monitoring this pathway for the backbone, and a two-step strategy was adopted. In a first step, called TMD(-1), the dimer structure was constrained to progressively move away from KC without indicating the direction, until the RMSD from KC reaches 36A. A total of 20 TMD(-1) simulations were performed under different initial conditions and different simulation parameters. For RMSD ranging between 0 and 22A, the whole set of TMD(-1) simulations follows a similar pathway, then divergences are observed. None of the simulations leads to the ED structure. At RMSD=22A, the dimers look like two parallel Us, still linked by the initial loop-loop interaction, but the strands of the stems (the arms of the Us) are positioned in such a manner that they can form intramolecular as well as intermolecular Watson-Crick base-pairs. This family of structure is called UU. In a second step (TMD simulations), 18 structures were picked up along the pathways generated with TMD(-1) and were constrained to move toward ED by decreasing progressively their RMSD from ED. We found that only structures from the UU family are able to easily reach ED-like conformations of the backbones without exhibiting a large constraint energy.

Structural characterization of Lyn-SH3 domain in complex with a herpesviral protein reveals an extended recognition motif that enhances binding affinity

The Src homology 3 (SH3) domain of the Src family kinase Lyn binds to the herpesviral tyrosine kinase interacting protein (Tip) more than one order of magnitude stronger than other closely related members of the Src family. In order to identify the molecular basis for high-affinity binding, the structure of free and Tip-bound Lyn-SH3 was determined by NMR spectroscopy. Tip forms additional contacts outside its classical proline-rich recognition motif and, in particular, a strictly conserved leucine (L186) of the C-terminally adjacent sequence stretch packs into a hydrophobic pocket on the Lyn surface. Although the existence of this pocket is no unique property of Lyn-SH3, Lyn is the only Src family kinase that contains an additional aromatic residue (H41) in the n-Src loop as part of this pocket. H41 covers L186 of Tip by forming tight hydrophobic contacts, and model calculations suggest that the increase in binding affinity compared with other SH3 domains can mainly be attributed to these additional interactions. These findings indicate that this pocket can mediate specificity even between otherwise closely related SH3 domains.

Unraveling a hotspot for TCR recognition on HLA-A2: evidence against the existence of peptide-independent TCR binding determinants

T cell receptor (TCR) recognition of peptide takes place in the context of the major histocompatibility complex (MHC) molecule, which accounts for approximately two-thirds of the peptide/MHC buried surface. Using the class I MHC HLA-A2 and a large panel of mutants, we have previously shown that surface mutations that disrupt TCR recognition vary with the identity of the peptide. The single exception is Lys66 on the HLA-A2 alpha1 helix, which when mutated to alanine disrupts recognition for 93% of over 250 different T cell clones or lines, independent of which peptide is bound. Thus, Lys66 could serve as a peptide-independent TCR binding determinant. Here, we have examined the role of Lys66 in TCR recognition of HLA-A2 in detail. The structure of a peptide/HLA-A2 molecule with the K66A mutation indicates that although the mutation induces no major structural changes, it results in the exposure of a negatively charged glutamate (Glu63) underneath Lys66. Concurrent replacement of Glu63 with glutamine restores TCR binding and function for T cells specific for five different peptides presented by HLA-A2. Thus, the positive charge on Lys66 does not serve to guide all TCRs onto the HLA-A2 molecule in a manner required for productive signaling. Furthermore, electrostatic calculations indicate that Lys66 does not contribute to the stability of two TCR-peptide/HLA-A2 complexes. Our findings are consistent with the notion that each TCR arrives at a unique solution of how to bind a peptide/MHC, most strongly influenced by the chemical and structural features of the bound peptide. This would not rule out an intrinsic affinity of TCRs for MHC molecules achieved through multiple weak interactions, but for HLA-A2 the collective mutational data place limits on the role of any single MHC amino acid side-chain in driving TCR binding in a peptide-independent fashion.

An assessment of the accuracy of methods for predicting hydrogen positions in protein structures

The addition of hydrogen atoms to models or experimental structures of proteins that contain only non-hydrogen atoms is a common step in crystallographic structure refinement, in theoretical studies of proteins, and in protein structure prediction. Accurate prediction of the hydrogen positions is essential, since they constitute around half of the atoms in proteins and hence contribute significantly to their energetics. Many computational tools exist for predicting hydrogen positions, although to date no quantitative comparison has been made of their accuracy or efficiency. Here we take advantage of the recent increase in ultra-high-resolution X-ray crystal structures (< 0.9 A resolution), as well as of a number of relatively high-resolution neutron diffraction structures (< 1.8 A resolution), to compare the quality of the predictions generated by a large set of commonly used methods. These include CHARMM, CNS, GROMACS, MCCE, MolProbity, WHAT IF, and X-PLOR. The hydrogen atoms that lack a rotational degree of freedom are mostly, but not always, accurately predicted. For hydrogens with a rotational degree of freedom, all the methods give much less accurate predictions. The predictions for the hydroxyl hydrogens are analyzed in detail, particularly those buried within the protein, and some explanation is provided for the errors observed. The results provide a means to make informed decisions regarding the choice and implementation of methodologies for placing hydrogens on structures of proteins. They also point to shortcomings in current force fields and suggest the need for improved descriptions of hydrogen bonding energetics.

A structural model of a seven-transmembrane helix receptor: The Duffy antigen/receptor for chemokine (DARC)

The Duffy antigen/receptor for chemokine (DARC) is an erythrocyte receptor for malaria parasites (Plasmodium vivax and Plasmodium knowlesi) and for chemokines. In contrast to other chemokine receptors, DARC is a promiscuous receptor that binds chemokines of both CC and CXC classes. The four extracellular domains (ECDs) of DARC are essential for its interaction with chemokines, whilst the first (ECD1) is sufficient for the interaction with malaria erythrocyte-binding protein. In this study, we elaborate and analyze structural models of the DARC. The construction of the 3D models is based on a comparative modeling process and on the use of many procedures to predict transmembrane segments and to detect far homologous proteins with known structures. Threading, ab initio, secondary structure and Protein Blocks approaches are used to build a very large number of models. The conformational exploration of the ECDs is performed with simulated annealing. The second and fourth ECDs are strongly constrained. On the contrary, the ECD1 is highly flexible, but seems composed of three consecutive regions: a small beta-sheet, a linker region and a structured loop. The chosen structural models encompass most of the biochemical features and reflect the known experimental data. They may be used to analyze functional interaction properties.

Equilibrium solvation in quadrupolar solvents

We present a microscopic theory of equilibrium solvation in solvents with zero dipole moment and nonzero quadrupole moment (quadrupolar solvents). The theory is formulated in terms of autocorrelation functions of the quadrupolar polarization (structure factors). It can be therefore applied to an arbitrary dense quadrupolar solvent for which the structure factors are defined. We formulate a simple analytical perturbation treatment for the structure factors. The solute is described by coordinates, radii, and partial charges of constituent atoms. The theory is tested on Monte Carlo simulations of solvation in model quadrupolar solvents. It is also applied to the calculation of the activation barrier of electron transfer reactions in a cleft-shaped donor-bridge-acceptor complex dissolved in benzene with the structure factors of quadrupolar polarization obtained from molecular-dynamics simulations.

Universal method for computation of electrostatic potentials

A computational approach to determine electrostatic interaction and gravitational potentials by performing direct numerical integration is presented. The potential is expanded using finite-element functions of arbitrary order. The method does not involve any solutions of systems of linear equations. The potential is instead obtained as a sum of differential contributions. Thus, no boundary conditions for the potential are needed. It is computationally efficient and well suited for parallel computers, since the innermost loops constitute matrix multiplications and the outer ones can be used as parallel indices. Without using prescreening or other computational tricks to speed up the calculation, the algorithm scales as N4/3 where N denotes the grid size.

Comparative study of generalized Born models: protein dynamics

In this work, we compare the results of molecular dynamics simulations involving the application of three generalized Born (GB) models to 10 different proteins. The three GB models, the Still, HCT, and modified analytical generalized Born models, were implemented in the computationally efficient gromacs package. The performance of each model was assessed from the backbone rms deviation from the native structure, the number of native hydrogen bonds retained in the simulation, and the experimental and calculated radius of gyration. Analysis of variance (ANOVA) was used to analyze the results of the simulations. The rms deviation measure was found to be unable to distinguish the quality of the results obtained with the three different GB models, whereas the number of native hydrogen bonds and radius of gyration yielded a statistically meaningful discrimination among models. Our results suggest that, of the three, modified analytical generalized Born yields the best agreement between calculated and experimentally derived structures. More generally, our study highlights the need both to evaluate the effects of different variables on the results of simulations and to verify that the results of molecular dynamics simulations are statistically meaningful.

Structural elements in IGP synthase exclude water to optimize ammonia transfer

In the complex pathway of histidine biosynthesis, a key branch point linking amino acid and purine biosynthesis is catalyzed by the bifunctional enzyme imidazole glycerol phosphate (IGP) synthase. The first domain of IGP synthase, a triad glutamine amidotransferase, hydrolyzes glutamine to form glutamate and ammonia. Its activity is tightly regulated by the binding of the substrate PRFAR to its partner synthase domain. Recent crystal structures and molecular dynamics simulations strongly suggest that the synthase domain, a (beta/alpha)(8) barrel protein, mediates the insertion of ammonia and ring formation in IGP by channeling ammonia from one remote active site to the other. Here, we combine both mutagenesis experiments and computational investigations to gain insight into the transfer of ammonia and the mechanism of conduction. We discover an alternate route for the entrance of ammonia into the (beta/alpha)(8) barrel and argue that water acts as both agonist and antagonist to the enzymatic function. Our results indicate that the architecture of the two subdomains, most notably the strict conservation of key residues at the interface and within the (beta/alpha)(8) barrel, has been optimized to allow the efficient passage of ammonia, and not water, between the two remote active sites.

Computational studies on imidazole heme conformations

Density functional theory computations of heme with ionized propionic acid groups, axially coordinated with two imidazoles, were performed for different mutual orientations of the imidazole planes. Environmental influences from water or protein were considered with a continuum dielectric medium by solving the Poisson equation. In vacuum, optimized geometries yielded imidazole-heme conformations where the NH groups of imidazoles are oriented toward the heme propionic groups in agreement with data from crystal structures of heme proteins. Conformational free-energy dependencies of the mutual orientation of axially ligated imidazoles calculated in protein (epsilon=10) and water (epsilon=80) environments confirmed the vacuum results, albeit the energy difference between the preferred and the 180 degrees opposite orientations of the imidazole ligand decreased from 3.84 kcal/mol in vacuum to 2.35 and 2.40 kcal/mol in protein and water, respectively. Two main factors determine the imidazole orientation: (1) the direct intramolecular electrostatic interactions of propionic groups with the polar NH groups of imidazole and (2) the electrostatic interaction of the total dipole moment of the imidazole-heme complex with the reaction field. In vacuum, only the first type of interaction is present, while in a dielectric medium the latter effect becomes competitive at high dielectric constant, resulting in a decrease of the orientational preference. Interestingly, the orientational preference of the imidazole axially ligated to heme becomes even more pronounced, if the negatively charged propionates are neutralized by counter charges that mimic salt bridges or protonation of the propionates.

Protein misfolding and amyloid formation for the peptide GNNQQNY from yeast prion protein Sup35: simulation by reaction path annealing

We study the early steps of amyloid formation of the seven residue peptide GNNQQNY from yeast prion-like protein Sup35 by simulating the random coil to beta-sheet and alpha-helix to beta-sheet transition both in the absence and presence of a cross-beta amyloid nucleus. The simulation method at atomic resolution employs a new implementation of a Langevin dynamics "reaction path annealing" algorithm. The results indicate that the presence of amyloid-like cross-beta-sheet strands both facilitates the transition into the cross-beta conformation and substantially lowers the free energy barrier for this transition. This model systems allows us to investigate the energetic and kinetic details of this transition, which is consistent with an auto-catalyzed, nucleation-like mechanism for the formation of beta-amyloid. In particular, we find that electrostatic interactions of peptide backbone dipoles contribute significantly to the stability of the beta-amyloid state. Furthermore, we find water exclusion and interactions of polar side-chains to be driving forces of amyloid formation: the cross-beta conformation is stabilized by burial of polar side-chains and inter-residue hydrogen bonds in the presence of an amyloid-like "seed". These findings are in support of a "dry, polar zipper model" of amyloid formation.

Are acidic and basic groups in buried proteins predicted to be ionized?

Ionizable residues play essential roles in proteins, modulating protein stability, fold and function. Asp, Glu, Arg, and Lys make up about a quarter of the residues in an average protein. Multi-conformation continuum electrostatic (MCCE) calculations were used to predict the ionization states of all acidic and basic residues in 490 proteins. Of all 36,192 ionizable residues, 93.5% were predicted to be ionized. Thirty-five percent have lost 4.08 kcal/mol solvation energy (DeltaDeltaG(rxn)) sufficient to shift a pK(a) by three pH units in the absence of other interactions and 17% have DeltaDeltaG(rxn) sufficient to shift pK(a) by five pH units. Overall 85% of these buried residues (DeltaDeltaG(rxn)>5DeltapK units) are ionized, including 92% of the Arg, 86% of the Asp, 77% of the Glu, and 75% of the Lys. Ion-pair interactions stabilize the ionization of both acids and bases. The backbone dipoles stabilize anions more than cations. The interactions with polar side-chains are also different for acids and bases. Asn and Gln stabilize all charges, Ser and Thr stabilize only acids while Tyr rarely stabilize Lys. Thus, hydroxyls are better hydrogen bond donors than acceptors. Buried ionized residues are more likely to be conserved than those on the surface. There are 3.95 residues buried per 100 residues in an average protein.

One- and two-body decomposable Poisson-Boltzmann methods for protein design calculations

Successfully modeling electrostatic interactions is one of the key factors required for the computational design of proteins with desired physical, chemical, and biological properties. In this paper, we present formulations of the finite difference Poisson-Boltzmann (FDPB) model that are pairwise decomposable by side chain. These methods use reduced representations of the protein structure based on the backbone and one or two side chains in order to approximate the dielectric environment in and around the protein. For the desolvation of polar side chains, the two-body model has a 0.64 kcal/mol RMSD compared to FDPB calculations performed using the full representation of the protein structure. Screened Coulombic interaction energies between side chains are approximated with an RMSD of 0.13 kcal/mol. The methods presented here are compatible with the computational demands of protein design calculations and produce energies that are very similar to the results of traditional FDPB calculations.

Crystal Structure of Amine Oxidase from Bovine Serum

Copper-containing amine oxidase extracted from bovine serum (BSAO) was crystallized and its three-dimensional structure at 2.37A resolution is described. The biological unit of BSAO is a homodimer, formed by two monomers related to each other by a non-crystallographic 2-fold axis. Each monomer is composed of three domains, similar to those of other amine oxidases from lower species. The two monomers are structurally equivalent, despite some minor differences at the two active sites. A large funnel allows access of substrates to the active-site; another cavity, accessible to the solvent, is also present between the two monomers; this second cavity could allow the entrance of molecular oxygen necessary for the oxidative reaction. Some sugar residues, bound to Asn, were still present and visible in the electron density map, in spite of the exhaustive deglycosylation necessary to grow the crystals. The comparison of the BSAO structure with those of other resolved AO structures shows strong dissimilarities in the architecture and charge distribution of the cavities leading to the active-site, possibly explaining the differences in substrate specificity.