Computational observation of an ion permeation through a channel protein.

The ion permeation process, driven by a membrane potential through an outer membrane protein, OmpF porin of Escherichia coli, was simulated by molecular dynamics. A Na+ ion, initially placed in the solvent region at the outer side of the porin channel, moved along the electric field passing through the porin channel in a 1.3 nsec simulation; the permeation rate was consistent with the experimentally estimated channel activity (10(8)-10(9)/sec). It this simulation, it was indicated that the ion permeation through the porin channel proceeds by a "push-out" mechanism, and that Asp113 is an important residue for the channel activity.

Change in conformation by DNA-peptide association: molecular dynamics of the Hin-recombinase-hixL complex.

The Hin-DNA complex is a molecular complex formed by the C-terminal 52mer peptide of the Hin-recombinase and a synthetic 13-bp hixL DNA. The peptide has three alpha-helices, the second and third of which form the helix-turn-helix motif to bind to the major groove. Both termini of the peptide reside within the minor groove. Three molecular dynamics simulations were performed based on the crystal structure of the Hin-DNA complex: one for the free Hin peptide, one for the free hixL DNA, and one for the complex. Analyses of the trajectories revealed that the dynamic fluctuations of both the Hin peptide and the hixL DNA were lowered by the complex formation. The simulation supported the experimental observation that the N-terminus and the helix-turn-helix motif were critical for formation of the complex, but the C-terminus played only a supportive role in DNA recognition. The simulations strongly suggested that the binding reaction should proceed by the induced fit mechanism. The ion and solvent distributions around the molecules were also examined.

Molecular dynamics simulations of trp apo- and holorepressors: domain structure and ligand-protein interaction.

Molecular dynamics simulations of the apo- and holo-forms of the trp-repressor protein were performed under extensively solvated conditions in order to elucidate their dynamic structures and ligand-protein interactions. The root mean square fluctuations calculated from the trajectories agreed with those calculated from X-ray temperature factors. Distance, distance fluctuation, and dynamic cross-correlation maps were drawn to provide information on the dynamic structures and communications among the domains. A three-domain format has been proposed for the crystal structure (Zhang et al., Nature 327:591-597, 1987); namely, helices A-C and F of both subunits make up a central core, and D and E of each subunit forms a DNA binding head. The results of the simulations were mostly consistent with the three-domain format. However, helix F was more flexible and freer than other parts of the central core. The turn DE, the helix-turn-helix DNA binding motif, was free from interactions and correlations with other domains in both forms of the repressor. A comparison of the simulations of the aporepressor and holorepressor showed that tryptophan binding made the DNA-binding helix D more flexible but helix F less flexible. Several amino acid residues in contact with the bound tryptophan were identified as making concerted motions with it. Interaction energies between the corepressor and the amino acid residues of the protein were analyzed; the results were mostly consistent with the mutational experiments.

Enantiotopic differentiation in horse-liver alcohol-dehydrogenase-catalyzed oxidoreduction studied with novel substrates having organometallic moieties.

Horse-liver alcohol-dehydrogenase-catalyzed oxidation of 1,2-bis(hydroxymethyl)ferrocene (1) gave (1R)-(+)-1-formyl-2-hydroxymethylferrocene (3) (86 +/- 2% enantiomeric excess, e.e.), while the reduction of the corresponding dialdehyde 1,2-diformylferrocene (2) gave the antipode (1S)-(-)-3 (94 +/- 2% e.e.). This fact indicates that the pro-R group in both 1 and 2 was preferentially converted by the enzyme. When one of two substituents on the substrate was replaced by a methyl group or moved to the beta-site, the stereoselectivity in the reaction decreased as evidenced by the enantiomeric purity of the products (5-64% e.e.). Treatment of racemic 1-hydroxyethylferrocene (14) with horse-liver alcohol dehydrogenase (HLADH) gave optically pure (R)-(-)-14 together with acetylferrocene. The reduction of 2 with HLADH, NAD and (2H6)ethanol gave (-)-(1S,2R)-1-formyl-2-[(R)-hydroxy(2H1)methyl]ferrocene and that of 1,2-di[(2H)formyl]ferrocene with HLADH, NAD and ethanol gave (-)-(1S,2R)-1-(2H)formyl-2-[(S)-hydroxy(2H1)methyl]ferrocene. These configurations indicate that the enzymic reduction occurred on the re-face of pro-R formyl group. The re-face selectivity was also found in the enzymic reduction of (eta 6-benzaldehyde)tricarbonylchromium and its (2H)formyl analogue. Docking of 2 into the active site of HLADH was examined using computer graphics. It has been suggested that the enantioselectivity to the pro-R side in the oxidoreduction of 1 and 2 by HLADH is a natural consequence of the re-face selectivity, which is caused by a steric interaction between the ligand and the side chain of Phe-93 or the Zn complex and strengthened by an interaction between the unreactive polar alpha-substituent and the protein, probably by hydrogen bond formation.

Purification and properties of 4-hydroxycyclohexanecarboxylate dehydrogenase from Corynebacterium cyclohexanicum.

4-Hydroxycyclohexanecarboxylate dehydrogenase, which requires NAD as a cofactor, was detected in crude soluble extracts of Corynebacterium cyclohexanicum grown on cyclohexanecarboxylic acid as the sole carbon source. The dehydrogenase was purified from extracts to an electrophoretically homogenous state by ammonium sulfate precipitation and chromatography on DEAE-650s, agarose-NAD and hydroxyapatite. The enzyme consisted of two identical subunits and had a native relative molecular mass of 53,600. There were two residues each of cysteine and tryptophan in the enzyme molecule. Oxo acid rather than hydroxy acid was routinely used as substrate for assay of the enzyme. The enzyme is highly specific for 4-oxocyclohexanecarboxylic acid: the carboxyl group is essential and the position of carbonyl group is important; neither the 2-oxo nor the 3-oxo homologue was used as substrate. A methyl substitution on the ring of 4-oxocyclohexanecarboxylate resulted in an almost complete loss of its activity. The reduction product was identified as trans-4-hydroxycyclohexanecarboxylic acid by gas-liquid chromatography and mass spectrometry. It was used as a substrate for the reverse reaction in the presence of NAD but not its cis-isomer. The enzyme was specific for the B-side (pro-S) hydrogen of NADH in the hydrogen transfer from NADH to 4-oxocyclohexanecarboxylate. The Km values for 4-oxocyclohexanecarboxylate and NADH in the reduction reaction at pH 6.8 were 0.50 mM and 0.28 mM, respectively, whereas those for trans-4-hydroxycyclohexanecarboxylate and NAD in the oxidation reaction at pH 8.8 were 0.51 mM and 0.23 mM, respectively. The equilibrium constant of the reaction was 1.79 x 10(-10) M. The enzyme was strongly inhibited by N-bromosuccinimide.

Cyclic oxyphosphoranes as model intermediates during splicing and cleavage of RNA: ab initio molecular orbital calculations on the conformational analysis.

Ab initio molecular orbital calculations have been carried out on hydrated adducts of methyl ethylene phosphate as a model intermediate during cleavage of RNA. Upon rotating the apical methoxyl group two kinds of stable conformers and two kinds of rotational transition states are located, the most stable conformation being gs-G where the dihedral angle between the apical methyl group and the basal ring oxygen is calculated to be 76 degrees. In this gs-G conformation one of the lone pairs on the apical oxygen is oriented antiperiplanar to the basal ring ester bond. The torsional energy required to rotate the methyl group about the phosphorus-apical oxygen bond leading to ts-C conformation, where the methyl group is eclipsed with the ring oxygen, is calculated to be 5.2 kcal/mol. Judging from the published substrate's coordinates in the RNase environment, the expected pentacoordinate-intermediate/transition state during the cleavage of RNA appears to be, in fact, the most stable gs-G conformation.

Energetics of RNA cleavage: implications for the mechanism of action of ribozymes.

A new class of ribozymes produce 2',3'-cyclic phosphate upon self-catalyzed cleavage of RNA molecules, similar to those observed during enzymatic (RNase-catalyzed) as well as non-enzymatic hydrolyses of RNAs. This product suggests that the reaction intermediate/transition state is a pentacoordinated oxyphosphorane. In order to elucidate the energetics of these RNA cleaving reactions, the reaction coordinate has been simulated and a pentacoordinated intermediate has been characterized via ab initio molecular orbital calculations utilizing the dianionic hydrolysis-intermediate of methyl ethylene phosphate as a model compound. The calculated reaction coordinate indicates that the transition state for the P-O(2') bond cleavage is lower in energy than that for the P-O(5') bond cleavage under uncatalyzed conditions. Thus, the dianionic pentacoordinated phosphorus intermediate tends to revert back to the starting RNA by cleaving the P-O(2') bond rather than productively cleaving the P-O(5') bond. In order for ribozymes to effectively cleave RNA molecules, it is therefore mandatory to stabilize the leaving 5'-oxygen, e.g. by means of a divalent magnesium ion.

Free energy perturbation study on a Trp-binding mutant (Ser88-->Cys) of the trp-repressor.

The Ser88-->Cys mutant of the trp-repressor showed a lower affinity for the corepressor than the wild-type repressor [delta delta G = 1.7 +/- 0.3 kcal/mol, Chou and Matthews (1989) J. Biol. Chem., 264, 18314-18319]. A molecular dynamics/free energy cycle perturbation study was performed to understand the origin of the decreased affinity. A value (delta delta G = 1.58 +/- 0.28 kcal/mol) comparable with the experimental value was obtained by the simulation. Free energy component analysis revealed that destabilization of the van der Waals interaction between Ser88 and Trp109 (corepressor) mainly contributed to the decreased affinity of the mutant. The rotational transition of the hydroxyl (sulfhydryl) group of Ser88 (Cys88) during the simulations affected the contributions of Arg84 and water to the free energy change in the aporepressor and those of Arg84 and Trp109 to that in the holorepressor. However, the contributions from different residues compensated each other, and the total free energy changes were almost invariable in the various simulations.

Molecular dynamics simulation of trp-aporepressor in a solvent.

Molecular dynamics simulations of Escherichia coli trp-aporepressor were carried out in the absence and presence of explicit water molecules. The vacuum simulations resulted in significant deformation of the initial X-ray structure. A solvated simulation with a nonbonded cut-off radius of 9 A gave a better result, and the most satisfactory result was obtained when electrostatic interactions within a cut-off radius of 18 A were considered by a twin-range method. The trajectory from the last simulation was used to analyze the dynamical properties of the aporepressor. The root-mean-square fluctuations of the residues showed the rigidity of the central core and the flexibility of the DNA-binding sites, consistent with the X-ray temperature factors. The dynamical cross-correlation map indicated a significant negative correlation between the central core and the two DNA-binding sites, and thus reproduced the three-domain format (a central core and two DNA-binding heads) from a dynamical point of view. The core region showed weak, but many, intra- and inter-molecular correlations, while the helix-turn-helix DNA-binding motifs were free from correlations with other regions.

Computational design of a substrate specificity mutant of a protein.

The wild-type trp repressor of E. coli bound 5-methoxytryptophan, a Trp analogue, less tightly than Trp. A mutant repressor (Val58-->Ala) that should bind 5-methoxytryptophan preferentially to Trp was computationally designed by free-energy calculations accompanied by free-energy decomposition. The designed mutant was demonstrated by experiments to bind 5-methoxytryptophan more tightly than Trp, consistent with the computational prediction. This success indicates the usefulness of free energy decomposition in protein design.

A molecular dynamics study of solvent behavior around a protein.

The solvent structure and behavior around a protein were examined by analyzing a trajectory of molecular dynamics simulation of the trp-holorepressor in a periodic box of water. The calculated self diffusion coefficient indicated that the solvent within 10 A of the protein had lower mobility. Examination of the solvent diffusion around different atoms of different kinds of residues showed no general tendency. This fact suggested that the solvent mobility is not influenced significantly by the kind of the atom or residue they solvated. Distribution analysis around the protein revealed two peaks of water oxygen: a sharp one at 2.8 A around polar and charged atoms and a broad one at approximately 3.4 A around apolar atoms. The former was stabilized by water-protein hydrogen bonds, and the latter was stabilized by water-water hydrogen bonds, suggesting the existence of a hydrophobic shell. An analysis of protein atom-water radial distribution functions confirmed these shell structures around polar or charged atoms and apolar ones.

Theoretical analyses on the role of Mg2+ ions in ribozyme reactions.

To elucidate the role of the Mg2+ ion in ribozyme reactions, we carried out ab initio molecular orbital investigations on dianionic trimethoxyphosphorane A and its Mg2+ complex (overall a neutral molecule) as a model system for the reaction center of Tetrahymena-type ribozyme. Although dianionic oxyphosphorane A concentrates its negative charges on the equatorial phosphoryl oxygens, the coordination of the Mg2+ ion between these two oxygens is unlikely. Geometry optimizations of the complex and the electrostatic potential of A both suggest that Mg2+ coordination preferably occurs in the region between the axial oxygen and the equatorial phosphoryl oxygen. The considerations of electrostatic potential rationalize the geometries of carboxylate-metal and phosphate-metal interactions extracted from the Cambridge Structural Database as well. Consequently, the Mg2+ ion at the active site of Tetrahymena-type ribozyme most likely lies in the regions between the axial and equatorial oxygens. The axial-equatorial coordinations of Mg2+ ions conceivably increase the electronegativities of the axial oxygens and facilitate cleavage of the phosphodiester bond located at the junction of the intron and the exon. It is thus likely that the Mg2+ ions play the key role in the phosphodiester cleavage reactions mediated by ribozymes.

RNA hydrolysis via an oxyphosphorane intermediate.

From calculations of a model reaction scheme for base-catalyzed RNA hydrolysis, a pentacoodinate dianionic intermediate 2a (Storer, et al., J. Am. Chem. Soc., 1991, 113, 5216-5219) as well as two transition states, TS1 and TS2, to the intermediate have been located by ab initio calculations at the 3-21G* level. Although the intermediate, which has the well depth on the order of kBT, is unlikely to be kinetically significant, the overall rate-limiting transition state structure TS2 obtained at 3-21G* level is very close to the corresponding structure at the STO-3G level; it has an extended P-O(5') bond breaking character. These gas-phase calculation results are used to qualitatively interpret mutagenesis results of Barnase and RNase T1 where water molecules are absent from the active site.

Rate limiting P-O(5') bond cleavage of RNA fragment: ab initio molecular orbital calculations on the base-catalyzed hydrolysis of phosphate.

In order to examine the energetics in base-catalyzed hydrolysis of RNA, a tentative pentacoordinated intermediate (3) has been characterized by molecular orbital calculations. Ab initio studies at the level of 3-21G* indicate that, under the Cs symmetry restricted conditions, the P-O(2) bond possessing antiperiplanar (app) lone pair electrons (Ip) on the equatorial oxygen (O(3)) can be cleaved with almost no barrier (TS1 transition state; 0.08 kcal mol-1), from the pentacoordinated intermediate (3) of base-catalyzed hydrolysis of phosphate, compared to the P-O(5) bond (TS2 transition state; 28.9 kcal mol-1) which lacks app lp assistance from O(3). The dianionic intermediate, however, loses the TS1 transition state thus its property as an intermediate when the Cs restriction is removed. The analysis of the entire potential energy surface enables us to conclude that, in a related system examined by Lim and Karplus [1990) J. Am. Chem. Soc., 112, 5872-5873) for attack by OH- on ethylene phosphate monoanion, the TS1 transition state had also been lost and thus no intermediate had been found. These results further support our earlier conclusions (Taira et al. (1990) Protein Engineering, 3, 691-701) of rate limiting transition state possessing extended P-O(5') bond breaking character (the TS2 transition state) in the base-catalyzed hydrolysis of RNA. Finally, although the lack of 2',3' -migration of phosphate moieties in basic condition appears to be in accord with the short-lived intermediate, it really does not prove the absence of the intermediate. The detail will be discussed in the text.

Preferential chelation of cationic ligands to axial-equatorial oxygens over equatorial-equatorial dianionic oxygens: implication to the mechanism of action of ribozymes.

Effects of chelation of H2O, H+, and Mg2+ to two kinds of potential pentacoordinate intermediates of ribozyme reactions were investigated by ab initio molecular orbital calculations. Unexpectedly, in all cases examined, axial-equatorial chelations were found to be more stabilizing than equatorial-equatorial chelations. These results support the hypothesis that Mg2+ ion is bound to the equatorial phosphoryl oxygen and the axial leaving/attacking oxygen in the transition state of ribozyme reactions.

Pentacoordinate oxyphosphorane intermediate always exists in aqueous solution.

Gas-phase ab initio calculations indicate that dianionic pentacoordinate oxyphosphoranes do not have a kinetically meaningful intermediate. The simplest oxyphosphorane PO5H3(2-) has the least tendency to have a pentacoordinate intermediate. However, it does have a pentacoordinate intermediate when it is solvated with six water molecules. These results support the hypothesis that the phosphoryl transfer reactions take place via pentacoordinate intermediate not only in acidic but also in basic media.

Prediction of endocrine disruptors based on a new structure-activity relationship for sex and environmental hormones using chemical hardness concept.

Classification of the relationship between electronic structures and biological activities of endocrine disruptors (so-called environmental hormones) was attempted using the parameters of absolute hardness (eta), absolute electronegativity (chi), and global softness (S), approximately defined as eta=1/2(epsilonLUMO-epsilonHOMO), chi=-1/2(epsilonHOMO+ epsilonLUMO), and S=1/eta, respectively, based on the hardness concept. The strength of binding affinity and toxicity of the chemicals were approximately proportional to the absolute hardness, and laterally toxic chlorinated PCDDs, PCBs, and DDTs are classified as chemically soft. Here we found that the electronic structures of environmental hormones can be classified into four main groups: 17beta-estradiol type (group I), testosterone type (group II), thyroxine type (group III), and HCH (hexachlorocyclohexane) type (group IV). Therefore, if we can predict the coordinate (chi, eta) of the electronic structure of one chemical on the eta-chi activity diagram, we would be able to predict the receptor with which the chemicals (environmental hormones) interact. For instance, 2,3,7,8-tetrachlorodibenzo-p-dioxin (2,3,7,8-TCDD) is classified in group II, therefore, it would bind with the thyroid receptor more than the estrogen receptor (group I). It appears that dibutyl phthalate would not interact with estrogen receptor because it does not belong to group I. In addition, the coordinates of these four groups do not complementarily overlap with the electronic structures of 20 natural amino acid residues. The eta-chi activity diagram is a new tool for the prediction of the toxicity and biological activity of environmental hormones.

Theoretical analyses on the role of metal cations in RNA cleavage processes.

Previously, we found that locations of metal cations in biomolecules, such as proteins and nucleotides, were in accord with predictions based on electrostatic-potential calculations of the nearest polar functional groups. Electrostatic potential of a cyclic oxyphosphorane dianion 2, which is a model compound for an intermediate or a transition state for the hammerhead ribozyme reaction, indicates that Mg2+ coordinations at the regions between axial (bridging) and equatorial (phosphoryl) oxygens are most favorable. If we consider the fact that negative charge is concentrated on two of the equatorial phosphoryl oxygens in the transition state 2, the axial-equatorial coordination is an unexpected result. Nevertheless, experimental results also support such an axial-equatorial coordination. Therefore, hammerhead ribozyme is a metalloenzyme in which at least one Mg2+ ion acts as a base and possibly a second Mg2+ ion acts as a Lewis acid. The double Mg2+ ion catalysis may be a common mechanism among various types of ribozyme and some protein enzymes as well.

A high performance system for molecular dynamics simulation of biomolecules using a special-purpose computer.

GRAPE (GRavity PipE) processors are special purpose computers for simulation of classical particles. The performance of MD-GRAPE, one of the GRAPEs developed for molecular dynamics, was investigated. The effective speed of MD-GRAPE was equivalent to approximately 6 Gflops. The precision of MD-GRAPE was good judging from the acceptable fluctuation of the total energy. Then a software named PEACH (Program for Energetic Analysis of bioCHemical molecules) was developed for molecular dynamics of biomolecules in combination with MD-GRAPE. Molecular dynamics simulation was performed for several protein-solvent systems with different sizes. Simulation of the largest system investigated (27,000 atoms) took only 5 sec/step. Thus, the PEACH-GRAPE system is expected to be useful in accurate and reliable simulation of large biomolecules.