Nanoinformatics knowledge infrastructures: bringing efficient information management to nanomedical research.

Nanotechnology represents an area of particular promise and significant opportunity across multiple scientific disciplines. Ongoing nanotechnology research ranges from the characterization of nanoparticles and nanomaterials to the analysis and processing of experimental data seeking correlations between nanoparticles and their functionalities and side effects. Due to their special properties, nanoparticles are suitable for cellular-level diagnostics and therapy, offering numerous applications in medicine, e.g. development of biomedical devices, tissue repair, drug delivery systems and biosensors. In nanomedicine, recent studies are producing large amounts of structural and property data, highlighting the role for computational approaches in information management. While in vitro and in vivo assays are expensive, the cost of computing is falling. Furthermore, improvements in the accuracy of computational methods (e.g. data mining, knowledge discovery, modeling and simulation) have enabled effective tools to automate the extraction, management and storage of these vast data volumes. Since this information is widely distributed, one major issue is how to locate and access data where it resides (which also poses data-sharing limitations). The novel discipline of nanoinformatics addresses the information challenges related to nanotechnology research. In this paper, we summarize the needs and challenges in the field and present an overview of extant initiatives and efforts.

Nanoinformatics: developing new computing applications for nanomedicine.

Nanoinformatics has recently emerged to address the need of computing applications at the nano level. In this regard, the authors have participated in various initiatives to identify its concepts, foundations and challenges. While nanomaterials open up the possibility for developing new devices in many industrial and scientific areas, they also offer breakthrough perspectives for the prevention, diagnosis and treatment of diseases. In this paper, we analyze the different aspects of nanoinformatics and suggest five research topics to help catalyze new research and development in the area, particularly focused on nanomedicine. We also encompass the use of informatics to further the biological and clinical applications of basic research in nanoscience and nanotechnology, and the related concept of an extended "nanotype" to coalesce information related to nanoparticles. We suggest how nanoinformatics could accelerate developments in nanomedicine, similarly to what happened with the Human Genome and other -omics projects, on issues like exchanging modeling and simulation methods and tools, linking toxicity information to clinical and personal databases or developing new approaches for scientific ontologies, among many others.

High-resolution crystallography and drug design.

Ultra-high-resolution X-ray crystallography of macromolecules (i.e. resolution better than 0.8 Angstroms) is a rising field that promises to provide new insight into the structure-function relationships of biomacromolecules. The picture emerging from macromolecular structures at this resolution is far more complex than previously understood, requiring for its study improved tools for structure refinement, analysis and annotation. Some of these problems were highlighted during the recent High Resolution Drug Design Meeting (Bischenberg-Strasbourg, France, 13-16 May 2004). We will review here some of the results and discussions that took place during that meeting and elaborate on the trends and challenges ahead in this emerging new field of research.

Ultrahigh resolution drug design I: details of interactions in human aldose reductase-inhibitor complex at 0.66 A.

The first subatomic resolution structure of a 36 kDa protein [aldose reductase (AR)] is presented. AR was cocrystallized at pH 5.0 with its cofactor NADP+ and inhibitor IDD 594, a therapeutic candidate for the treatment of diabetic complications. X-ray diffraction data were collected up to 0.62 A resolution and treated up to 0.66 A resolution. Anisotropic refinement followed by a blocked matrix inversion produced low standard deviations (<0.005 A). The model was very well ordered overall (CA atoms' mean B factor is 5.5 A2). The model and the electron-density maps revealed fine features, such as H-atoms, bond densities, and significant deviations from standard stereochemistry. Other features, such as networks of hydrogen bonds (H bonds), a large number of multiple conformations, and solvent structure were also better defined. Most of the atoms in the active site region were extremely well ordered (mean B approximately 3 A2), leading to the identification of the protonation states of the residues involved in catalysis. The electrostatic interactions of the inhibitor's charged carboxylate head with the catalytic residues and the charged coenzyme NADP+ explained the inhibitor's noncompetitive character. Furthermore, a short contact involving the IDD 594 bromine atom explained the selectivity profile of the inhibitor, important feature to avoid toxic effects. The presented structure and the details revealed are instrumental for better understanding of the inhibition mechanism of AR by IDD 594, and hence, for the rational drug design of future inhibitors. This work demonstrates the capabilities of subatomic resolution experiments and stimulates further developments of methods allowing the use of the full potential of these experiments.

Subatomic and atomic crystallographic studies of aldose reductase: implications for inhibitor binding.

The determination of several of aldose reductase-inhibitor complexes at subatomic resolution has revealed new structural details, including the specific interatomic contacts involved in inhibitor binding. In this article, we review the structures of the complexes of ALR2 with IDD 594 (resolution: 0.66 angstrom, IC50 (concentration of the inhibitor that produced half-maximal effect): 30 nM, space group: P2(1)), IDD 393 (resolution: 0.90 angstrom, IC50: 6 nM, space group: P1), fidarestat (resolution: 0.92 angstrom, IC50: 9 nM, space group: P2(1)) and minalrestat (resolution: 1.10 angstrom, IC50: 73 nM, space group: P1). The structures are compared and found to be highly reproductible within the same space group (root mean square (RMS) deviations: 0.15 approximately 0.3 angstrom). The mode of binding of the carboxylate inhibitors IDD 594 and IDD 393 is analysed. The binding of the carboxylate head can be accurately determined by the subatomic resolution structures, since both the protonation states and the positions of the atoms are very precisely known. The differences appear in the binding in the specificity pocket. The high-resolution structures explain the differences in IC50, which are confirmed both experimentally by mass spectrometry measures of VC50 and theoretically by free energy perturbation calculations. The binding of the cyclic imide inhibitors fidarestat and minalrestat is also described, focusing on the observation of a Cl(-) ion which binds simultaneously with fidarestat. The presence of this anion, binding also to the active site residue His110, leads to a mechanism in which the inhibitor can bind in a neutral state and then become charged inside the active site pocket. This mechanism can explain the excellent in vivo properties of cyclic imide inhibitors. In summary, the complete and detailed information supplied by the subatomic resolution structures can explain the differences in binding energy of the different inhibitors.

A new addition to the structural bioinformatics toolbox: 3D models of short bioactive peptides from multiple sequences using feedback restrained molecular dynamics (FRMD).

Modem molecular biology techniques allow the use of new approaches for the 3D mapping of 1D information. Molecular biology techniques are capable of producing large amounts of 1D information (sequences) from a number of different sources (phage displays, ESTs, etc.). In this work, we present a technique that takes advantage of large sets of 1D information and increasingly available computer power to create 3D models. For the purpose of validation the technique is first applied to the modeling of an erythropoietin analog of known 3D structure from 1D information only. The technique is then used to model the immunoreactive region of echinococcus granulosus AgB based on phage raised mimotopes for which there is no previous structural information. The technique here presented is of general application to similar problems where 1D information is available and structure activity relationships (SAR) is needed.

Ink-jet printer heads for ultra-small-drop protein crystallography.

Mass-produced automated piezoelectric driven picoliter delivery systems (printer heads) are fast, inexpensive, and reliable devices that are capable of delivering a very large range of volumes and are ideally suited for high-throughput protein crystallography studies. We used this technology to set up under-oil crystallization experiments with drop sizes from the 200-nL to 3-microL volume range, commonly used in protein crystallography, and show its application in setting ultra-small (2 nL) drops, the smallest drop volume reported to date for this type of assay.

Structure and inhibition of plasmepsin II, a hemoglobin-degrading enzyme from Plasmodium falciparum.

Plasmodium falciparum is the major causative agent of malaria, a disease of worldwide importance. Resistance to current drugs such as chloroquine and mefloquine is spreading at an alarming rate, and our antimalarial armamentarium is almost depleted. The malarial parasite encodes two homologous aspartic proteases, plasmepsins I and II, which are essential components of its hemoglobin-degradation pathway and are novel targets for antimalarial drug development. We have determined the crystal structure of recombinant plasmepsin II complexed with pepstatin A. This represents the first reported crystal structure of a protein from P. falciparum. The crystals contain molecules in two different conformations, revealing a remarkable degree of interdomain flexibility of the enzyme. The structure was used to design a series of selective low molecular weight compounds that inhibit both plasmepsin II and the growth of P. falciparum in culture.

Inhibition and catalytic mechanism of HIV-1 aspartic protease.

The structure of the HIV-1 protease in complex with a pseudo-C2 symmetric inhibitor, which contains a central difluoroketone motif, has been determined with X-ray diffraction data extending to 1.7 A resolution. The electron density map clearly indicates that the inhibitor is bound in a symmetric fashion as the hydrated, or gemdiol, form of the difluoroketone. Refinement of the complex reveals a unique, and almost symmetric, set of interactions between the geminal hydroxyl groups, the geminal fluorine atoms, and the active-site aspartate residues. Several hydrogen bonding patterns are consistent with that conformation. The lowest energy hydrogen disposition, as determined by semiempirical energy calculations, shows only one active site aspartate protonated. A comparison between the corresponding dihedral angles of the difluorodiol core and those of a hydrated peptide bond analog, calculated ab-initio, shows that the inhibitor core is a mimic of a hydrated peptide bond in a gauche conformation. The feasibility of an anti-gauche transition for a peptide bond after hydration is verified by extensive molecular dynamics simulations. The simulations suggest that rotation about the C-N scissile bond would readily occur after hydration and would be driven by the optimization of the interactions of peptide side-chains with the enzyme. These results, together with the characterization of a transition state leading to bond breakage via a concerted exchange of two protons, suggest a proteolysis mechanism whereby only one active site aspartate is initially protonated. The steps of this mechanism are: asymmetric binding of the substrate; hydration of the peptidic carbonyl by an active site water; proton translocation between the active site aspartate residues simultaneously with carbonyl hydration; optimization of the binding of the entire substrate facilitated by the flexible structure of the hydrated peptide bond, which, in turn, forces the hydrated peptide bond to assume a gauche conformation; simultaneous proton exchange whereby one hydroxyl donates a proton to the charged aspartate, and, at the same time, the nitrogen lone pair accepts a proton from the other aspartate; and, bond breakage and regeneration of the initial protonation state of the aspartate residues.

Molecular dynamics simulations of water using a floating polynomial force field and an interpolating electrostatic field representation.

The ability to accurately describe the force field of a molecule is of great importance in spectroscopic and drug design studies. However, the fitting of accurate potential energy functions has proved to be a highly complex task. The description through a simple generic formula of all conformations of a molecule has proved to be a seldom reliable procedure, while more complex representations are increasingly difficult to fit, slower to compute, and difficult to program. In this work, alternative procedures are explored: (1) the intramolecular force fields are expanded in a floating polynomial representation; (2) a fast treatment for the non-bonded interactions is applied. The advantage of these treatments is in their ability to describe highly accurate representations of molecules in a very efficient manner. The main difficulty is a heavy trade off in computer memory usage. Some of the more frequently used force fields for water, and a first principle force field are used as a test of these techniques.

Novel procedure for structure refinement in homology modeling and its application to the human class Mu glutathione S-transferases.

Glutathione S-transferases (GST) are a major class of phase II detoxifying enzymes that conjugate glutathione to electrophiles. Their involvement in the degradation of chemotherapeutic agents, which contributes to drug resistance, makes this family of enzymes potential targets for therapeutic agents. This study generates, by homology modeling, a 3-D structure of three GST human isozymes of the Mu class, M1b-1b, M2-2 and M3-3, using the Rat3-3 GST structure as a template. The high percentage of identity among these enzymes and the lack of insertions and deletions make the system ideally suited to the technique of homology modeling. A novel technique for the modeling of protein structures was applied. The structure of the template was used to generate a low-resolution crystallographic map in which the initial coordinates of the structure to be modeled were placed. The structure was then annealed within this envelope. In addition, a feedback-restrained molecular dynamics procedure was adopted to scale the template restraints during the simulations. Three independent validation procedures were applied. To assess the reliability of the methods, an identical series of simulation steps to those used in the refinement were applied to the template structure (self modeling). Further, a homology structure for the Rat3-3 template was generated, starting from the modeled M1b-1b structure (reverse modeling). To assess the reasonableness of the modeled structures, two recently developed methodologies to verify protein structures based on statistics of the nonbonded interactions were applied. Overall, the structures appear to be consistent.

Crystal structures of native and inhibited forms of human cathepsin D: implications for lysosomal targeting and drug design.

Cathepsin D (EC 3.4.23.5) is a lysosomal protease suspected to play important roles in protein catabolism, antigen processing, degenerative diseases, and breast cancer progression. Determination of the crystal structures of cathepsin D and a complex with pepstatin at 2.5 A resolution provides insights into inhibitor binding and lysosomal targeting for this two-chain, N-glycosylated aspartic protease. Comparison with the structures of a complex of pepstatin bound to rhizopuspepsin and with a human renin-inhibitor complex revealed differences in subsite structures and inhibitor-enzyme interactions that are consistent with affinity differences and structure-activity relationships and suggest strategies for fine-tuning the specificity of cathepsin D inhibitors. Mutagenesis studies have identified a phosphotransferase recognition region that is required for oligosaccharide phosphorylation but is 32 A distant from the N-domain glycosylation site at Asn-70. Electron density for the crystal structure of cathepsin D indicated the presence of an N-linked oligosaccharide that extends from Asn-70 toward Lys-203, which is a key component of the phosphotransferase recognition region, and thus provides a structural explanation for how the phosphotransferase can recognize apparently distant sites on the protein surface.

Ion channels in icosahedral virus: a comparative analysis of the structures and binding sites at their fivefold axes.

An analysis of the crystallographically determined structures of the icosahedral protein coats of Tomato Bushy Stunt Virus, Southern Bean Mosaic Virus, Satellite Tobacco Necrosis Virus, Human Rhinovirus 14 and Mengovirus around their fivefold axes is presented. Accessibilities surfaces, electrostatic energy profile calculations, ion-protein interaction energy calculations, free energy perturbation methods and comparisons with structures of chelating agents are used in this study. It is concluded that the structures built around the viral fivefold axes would be adequate for ion binding and transport. Relative ion preferences are derived for the binding sites, using free energy perturbation methods, which are consistent with the experimental data when available. In the cases where crystallographic studies determined the existence of ions on the fivefold axes, our results indicate that they would correspond to ions in crystallization or purification buffers. The environment of the fivefold axes are rich in polar residues in all icosahedral viral structures whose atomic coordinates are available, including some that are not being analyzed in detail in this work. The fivefold channel-like structures have most of the basic properties expected for real ion channels including a funnel at the entrance, a polar internal environment with frequent alternation of acidic and basic residues, ion binding sites, the capability to induce ion dehydration and ion transit from the external viral surface to the binding sites.

Recognition in cell adhesion. A comparative study of the conformations of RGD-containing peptides by Monte Carlo and NMR methods.

Many of the proteins that mediate cell adhesion processes processes-fibronectin, fibrinogen, vitronectin, von Willebrand factor, osteopontin, laminin and various collagens--contain the amino acid sequence Arg-Gly-Asp. Short peptides that include this sequence have been shown to inhibit the interactions of cell adhesion proteins with their receptors and to have dramatic effects on developmental processes involving cellular recognition. In order to determine which conformations are accessible to Arg-Gly-Asp containing peptides, we analyzed tri-, tetra- and pentapeptides using molecular mechanics and Monte Carlo methods, and studied the solution conformations of the pentapeptide Gly-Arg-Gly-Asp-Ser using nuclear magnetic resonance techniques. The Monte Carlo method was used to: (a) identify the low energy conformations of the peptides and (b) evaluate their thermodynamic properties. In the case of the pentapeptide Gly-Arg-Gly-Asp-Gly, the four stable conformations include three with reverse turns and one open structure. The conformations found in this analysis are compatible with the nuclear magnetic resonance (nuclear Overhauser effect) data.

Ion channels in southern bean mosaic virus capsid.

The study of southern bean mosaic virus protein coat high resolution model revealed a structure with properties of a natural protein-ion channel. Coat protein pentamers form a 30-A long channel and the amino acid composition of its wall bears some homology with the pentameric structure proposed for the nicotinic acetylcholine receptor channel. Ion transport properties were analyzed by computing ion-protein interaction energies on the basis of quantum chemistry methods. Energy maps show a channel attractive for cations, fully permeable to Li(+) and a narrow barrier for other cations and water. The energy profiles found are similar to the profiles determined for the K(+) channel of the sarcoplasmic reticulum. Comparisons with other icosahedral virus structures, including picornaviruses, suggest that ion channels would be a common feature of viral capsids. Biological roles for these channels are proposed.