Biomedical subjects
Publications and source records attributed to P G Mezey.
A computational technique is proposed for the study of electron density variations within a distribution of molecular conformations. These variations are defined in terms of the deviations of individual electron densities from the average density associated with the average of conformations within a conformational range of a molecule.
In this article we report the application of the Promolecular Atomic Shell Approximation (Promolecular ASA) to the graphical representation of the density function (DF) of large macromolecular systems. Promolecular ASA DF, constructed from previously computed and fitted atomic densities, provides a fast and practical representation of Molecular IsoDensity Contours (MIDCOs). These representations can be extended to macromolecular systems composed by > 1000 atoms easily and with low computational costs, allowing the visualization of protein DF. The method is at first presented with a small molecule (2,4,6-trinitrophenol), comparing the resulting ASA MIDCOs with direct ab initio contours. For macromolecular tests the Promolecular ASA densities are also applied to the generation of macromolecular density surfaces of two proteins: myoglobin (2541 atoms) and gene V protein (1362 atoms).
The Maillard (browning) reaction involving the polycondensation of sugars and amino acids is believed to be an important abiotic pathway for humic substance formation in nature. However, a major drawback is that the Maillard reaction is extremely slow at temperatures encountered under normal environmental conditions. In order to elucidate some details of this process molecular shape analysis was applied to investigate the initial reaction between D-glucose and glycine to form the Amadori compound fructosylglycine which is an intermediate product in the Maillard reaction. The structure of the Amadori compound was optimized at a quantum mechanical level and its ground state electron energy calculated. Molecular Iso-Density Contours (MIDCO's), electron density contour surfaces of constant electron density, were constructed for D-glucose, glycine and fructosylglycine in order to study the steric conditions for the reaction. The calculations indicate that the Amadori compound and water on one hand and the separate entities D-glucose and glycine on the other hand are very similar to each other in terms of their ground state energy. This agrees with the experimental observation that the reaction between D-glucose and glycine to form the Amadori compound is slow.
In mixture experiments, one may be interested in estimating not only main effects but also some interactions. Main effects and significant interactions in a mixture may be estimated through appropriate mixture experiments, such as simplex-centroid designs. However, for mixtures with a large number of factors, the run size for these designs becomes impractically large. A subset of a full simplex-centroid design may be used, but the problem remains regarding which factor-level settings should be selected. In this paper, we propose a solution that considers design points with either one or p individual nonzero factor-level settings. These fractional simplex designs provide a means of screening for interactions and of investigating the behavior of many-component mixtures as a whole while greatly reducing the run size compared with full simplex-centroid designs. The means of construction of the design arrays is described, and designs for < or = 31 factors are presented. Some of the proposed methodology is illustrated using generated data.
In this research we test and compare three possible atom-based screening functions used in the heuristic molecular lipophilicity potential (HMLP). Screening function 1 is a power distance-dependent function, bi/[formula: see text] Ri-r [formula: see text] gamma, screening function 2 is an exponential distance-dependent function, bi exp(-[formula: see text] Ri-r [formula: see text]/d0), and screening function 3 is a weighted distance-dependent function, sign(bi) exp[-xi [formula: see text] Ri-r [formula: see text]/magnitude of bi)]. For every screening function, the parameters (gamma, d0, and xi) are optimized using 41 common organic molecules of 4 types of compounds: aliphatic alcohols, aliphatic carboxylic acids, aliphatic amines, and aliphatic alkanes. The results of calculations show that screening function 3 cannot give chemically reasonable results, however, both the power screening function and the exponential screening function give chemically satisfactory results. There are two notable differences between screening functions 1 and 2. First, the exponential screening function has larger values in the short distance than the power screening function, therefore more influence from the nearest neighbors is involved using screening function 2 than screening function 1. Second, the power screening function has larger values in the long distance than the exponential screening function, therefore screening function 1 is effected by atoms at long distance more than screening function 2. For screening function 1, the suitable range of parameter gamma is 1.0 < gamma < 3.0, gamma = 2.3 is recommended, and gamma = 2.0 is the nearest integral value. For screening function 2, the suitable range of parameter d0 is 1.5 < d0 < 3.0, and d0 = 2.0 is recommended. HMLP developed in this research provides a potential tool for computer-aided three-dimensional drug design.
In this contribution we suggest a heuristic molecular lipophilicity potential (HMLP), which is a structure-based technique requiring no empirical indices of atomic lipophilicity. The input data used in this approach are molecular geometries and molecular surfaces. The HMLP is a modified electrostatic potential, combined with the averaged influences from the molecular environment. Quantum mechanics is used to calculate the electron density function rho(r) and the electrostatic potential V(r), and from this information a lipophilicity potential L(r) is generated. The HMLP is a unified lipophilicity and hydrophilicity potential. The interactions of dipole and multipole moments, hydrogen bonds, and charged atoms in a molecule are included in the hydrophilic interactions in this model. The HMLP is used to study hydrogen bonds and water-octanol partition coefficients in several examples. The calculated results show that the HMLP gives qualitatively and quantitatively correct, as well as chemically reasonable, results in cases where comparisons are available. These comparisons indicate that the HMLP has advantages over the empirical lipophilicity potential in many aspects. The HMLP is a three-dimensional and easily visualizable representation of molecular lipophilicity, suggested as a potential tool in computer-aided three-dimensional drug design.
In this work we apply a recently developed method for characterizing the shape of the tertiary structure of proteins. The approach is based on a combination of graph- and knot-theoretical characterizations of Cartesian projections of the space curve describing the protein backbone. The proposed technique reduces the essential shape features to a topologically based code formed by a sequence of knot symbols and polynomials. These polynomials are topological invariants that describe the overcrossing and knotting patterns of curves derived from the molecular space curve. These descriptors are algorithmically computed. The procedure is applied to describe the structure of the carboxy terminal fragment of the L7/L12 chloroplast ribosomal protein (CTF L7/L12) and the potato carboxypeptidase A inhibitor protein (PCI), which has a set of three disulfide bridges. In the former case, we describe the protein's shape features in terms of its alpha-helices, and a backbone simplified by considering helices without internal structure. An extension of the methodology to describe disulfide bridges is discussed and applied to PCI. Changes in the knot-theoretical characterization due to possible uncertainties in the resolution of the X-ray structure, as well as the inclusion of low-frequency motions of the backbone, are also discussed.
The ribbon model of chain macromolecules is a useful tool for analyzing some of the large-scale shape features of these complex systems. Up to now, the ribbon model has been used mostly to produce graphical displays, which are usually analyzed by visual inspection. In this work we suggest a computational method for characterizing automatically, in a concise and algebraic fashion, some of the important shape features of these ribbon models. The procedure is based on a graph-theoretical and knot-theoretical characterization of three well-defined projections of a space curve associated with the ribbon. The labeled graphs can be characterized by the handedness of the crossovers in the ribbon that are the vertices of the graph. The method can be used to provide a fully algebraic representation of the changes occurring when a molecule, such as a protein, undergoes conformational rearrangements (folding), as well as to provide a shape comparison for a pair of related molecular ribbons. This algebraic representation is well suited for easy storage, retrieval, and computer manipulation of the information on the ribbon's shape. Illustrative examples of the method are provided.
Electron density calculations were undertaken on several atoms in a series of 3-substituted-4-phenyl-3-buten-2-ones in order to gain insight into the molecular features which affect charge densities. The results indicate that substituents at position 3 alter the electron densities of the olefinic group but have little effect on the acetyl function. The compounds were tested against L1210 cells in vitro, and the results suggest that electronic--but not steric--factors are important in affecting cytotoxicity. The most active compound was 3-phenylmethylene-2,4-pentanedione (1c) with an ED50 value of 1.06 x 10(-8) M.
Each complete, boundaryless molecular electron density is fully determined by any nonzero volume piece of the electron density cloud. This inherent feature of molecules, called the "holographic" property of molecular electron densities, provides a strong foundation for the local, quantum chemical shape analysis of various functional groups, pharmacophores, and other local molecular moieties. A proof is presented for the relevant molecular shape theorem, the "holographic electron density shape theorem", and the role of this theorem in quantum chemical, quantitative shape-activity relations (QShAR) is discussed. The quantum chemical methods of molecular shape analysis can be extended to ab initio quality electron densities of macromolecules, such as proteins, as well as to local molecular moieties, such as functional groups or pharmacophores, based on the transferability and additivity of local, fuzzy density fragments and the associated local density matrixes within the framework of the ADMA (Adjustable Density Matrix Assembler) approach. In addition to new results on chemical bonding and the development of macromolecular force methods, the new methodologies are also applicable to QShAR studies in computer-aided drug discovery and in toxicological risk assessment.