Recent advances in genetic engineering have led to a growing acceptance of the fact that enzymes work like other catalysts by reducing the activation barriers of the corresponding reactions. However, the key question about the action of enzymes is not related to the fact that they stabilize transition states but to the question to how they accomplish this task. This work considers the catalytic reaction of serine proteases and demonstrates how one can use a combination of calculations and experimental information to elucidate the key contributions to the catalytic free energy. Recent reports about genetic modifications of the buried aspartic group in serine proteases, which established the large effect of this group (but could not determine its origin), are analyzed. Two independent methods indicate that the buried aspartic group in serine proteases stabilizes the transition state by electrostatic interactions rather than by alternative mechanisms. Simple free energy considerations are used to eliminate the double proton-transfer mechanism (which is depicted in many textbooks as the key catalytic factor in serine proteases). The electrostatic stabilization of the oxyanion side of the transition state is also considered. It is argued that serine proteases and other enzymes work by providing electrostatic complementarity to the changes in charge distribution occurring during the reactions they catalyze.

No abstract

The modulatory action of Ca2+-calmodulin on multiple targets is inhibited by trifluoperazine, which competes with target proteins for calmodulin binding. The structure of calmodulin crystallized with two trifluoperazine molecules is determined by X-ray crystallography at 2.74 A resolution. The X-ray data together with the characteristic and distinct signals obtained by circular dichroism in solution allowed us to identify the binding domains as well as the order of the binding of two trifluoperazine molecules to calmodulin. Accordingly, the binding of trifluperazine to the C-terminal hydrophobic pocket is followed by the interaction of the second drug molecule with an interdomain site. Recently, we demonstrated that the two bisindole derivatives, vinblastine and KAR-2 [3"-(beta-chloroethyl)-2",4"-dioxo-3, 5"-spirooxazolidino-4-deacetoxyvinblastine], interact with calmodulin with comparable affinity; however, they display different functional effects [Orosz et al. (1997) British J. Pharmacol. 121, 955-962]. The structural basis responsible for these effects were investigated by circular dichroism and fluorescence spectroscopy. The data provide evidence that calmodulin can simultaneously accommodate trifluoperazine and KAR-2 as well as vinblastine and KAR-2, but not trifluoperazine and vinblastine. The combination of the binding and structural data suggests that distinct binding sites exist on calmodulin for vinblastine and KAR-2 which correspond, at least partly, to that of trifluoperazine at the C-terminal hydrophobic pocket and at an interdomain site, respectively. This structural arrangement can explain why these drugs display different anticalmodulin activities. Calmodulin complexed with melittin is also able to bind two trifluoperazine molecules, the binding of which appears to be cooperative. Results obtained with intact and proteolytically cleaved calmodulin reveal that the central linker region of the protein is indispensable for simultanous interactions with two molecules of either identical or different ligands.

The aspartic residue at the base of the substrate-binding pocket of trypsin was replaced by serine (present in a similar position in chymotrypsin) through sitedirected mutagenesis. The wild-type (with in the mature trypsin sequence) and mutant (Ser-189) trypsinogens were expressed in Escherichia coli, purified to homogeneity, activated by enterokinase, and tested with a series of fluorogenic tetrapeptide substrates with the general formula succinylAla-Ala-Pro-Xaa-AMC, where AMC is 7-amino-4-methylcoumarin and Xaa is Lys, Arg, Tyr, Phe, Leu, or Trp. As compared to [Asp'l8trypsin, the activity of [Ser'"1trypsin on lysyl and arginyl substrates decreased by about 5 orders of magnitude while its Km values increased only 2-to 6-fold. In contrast, [Serl89]trypsin was 10-50 times more active on the less preferred, chymotrypsin-type substrates (tyrosyl, phenylalanyl, leucyl, and tryptophanyl). The activity of [Ser"]9]trypsin on lysyl substrate was about 100-fold greater at pH 10.5 than at pH 7.0, indicating that the unprotonated lysine is preferred. Assuming the reaction mechanisms of the wild-type and mutant enzymes to be the same, we calculated the changes in the transition-state energies for various enzyme-substrate pairs to reflect electrostatic and hydrogen-bond interactions. The relative binding energies (E) in the transition state are as follows: El, > EPP > EPA > EIP EIA, where I = ionic, P = nonionic but polar, and A = apolar residues in the binding pocket. These side-chain interactions become prominent during the transition of the Michaelis complex to the tetrahedral transition-state complex.The binding of substrates or inhibitors to the specificity pocket of an enzyme involves a combination of chemical forces including hydrogen bonds and electrostatic, hydrophobic, and steric interactions. The complexity of the interactions involved in the substrate specificity of an enzyme is exemplified by trypsin. The three-dimensional structures of trypsin bound to pancreatic trypsin inhibitor (PTI) (1)(2)(3)(4) or to the pseudosubstrate benzamidine (5, 6) suggest that the carboxylate of Asp-189, at the base of the trypsin binding pocket, is largely responsible for the specificity of binding of the enzyme to positively charged amino acid side chains.The major role of electrostatic interactions in the trypsin binding pocket has been analyzed by measuring (7) and calculating (8) the stabilization energies of binding between a series of benzamidine analogs and trypsin. In addition, the high degree of structural similarity of the trypsin and chymotrypsin binding pockets (9, 10) is consistent with the experimental observations that aromatic side chains may form favorable hydrophobic interactions with the trypsin binding pocket (10-13 by using a series of synthetic fluorogenic substrates with various amino acids in the C-terminal (P1) position in order to compare the electrostatic interactions of the different enzyme-substrate pairs. MATERIALS AND METHODSMaterials. Tetrapeptide substrates with the fluorogenic leaving...

The complexes of pig muscle 3-phosphoglycerate kinase with the substrate MgATP and with the nonsubstrate Mg(2+)-free ATP have been characterized by binding, kinetic, and crystallographic studies. Comparative experiments with ADP and MgADP have also been carried out. In contrast to the less specific and largely ionic binding of Mg(2+)-free ATP and ADP, specific occupation of the adenosine binding pocket by MgATP and MgADP has been revealed by displacement experiments with adenosine and anions, as well as supported by isothermal calorimetric titrations. The Mg(2+)-free nucleotides similarly stabilize the overall protein structure and restrict the conformational flexibility around the reactive thiol groups of helix 13, as observed by differential scanning microcalorimetry and thiol reactivity studies, respectively. The metal complexes, however, behave differently. MgADP, but not MgATP, further increases the conformational stability with respect to its Mg(2+)-free form, which indicates their different modes of binding to the enzyme. Crystal structures of the binary complexes of the enzyme with MgATP and with ATP (2.1 and 1.9 A resolution, respectively) have shown that the orientation and interaction of phosphates of MgATP largely differ not only from those of ATP but also from the previously determined ones of either MgADP [Davies, G. J., Gamblin, S. J., Littlechild, J. A., Dauter, Z., Wilson, K. S., and Watson, H. C. (1994) Acta Crystallogr. D50, 202-209] or the metal complexes of AMP-PNP [May, A., Vas, M., Harlos, K., and Blake, C. C. F. (1996) Proteins 24, 292-303; Auerbach, G., Huber, R., Grattinger, M., Zaiss, K., Schurig, H., Jaenicke, R., and Jacob, U. (1997) Structure 5, 1475-1483] and are more similar to the interactions formed with MgAMP-PCP [Kovári, Z., Flachner, B., Náray-Szabó, G., and Vas, M. (2002) Biochemistry 41, 8796-8806]. Mg(2+) is liganded to both beta- and gamma-phosphates of ATP, while beta-phosphate is linked to the conserved Asp218, i.e., to the N-terminus of helix 8, through a water molecule; the known interactions of either MgADP or the metal complexes of AMP-PNP with the N-terminus of helix 13 and with Asn336 of beta-strand J are absent in the case of MgATP. Fluctuation of MgATP phosphates between two alternative sites has been proposed to facilitate the correct positioning of the mobile side chain of Lys215, and the catalytically competent active site is thereby completed.

A comparative study of the pH-dependent redox mechanisms of several members of the cytochrome c3 family has been carried out. In a previous work, the molecular determinants of this dependency (the so-called redox-Bohr effect) were investigated for one species using continuum electrostatic methods to find groups with a titrating range and strength of interaction compatible with a mediating role in the redox-Bohr effect. Here we clarify these aspects in the light of new and improved pKa calculations, our findings supporting the hypothesis of propionate D from heme I being the main effector in the pH-dependent modulation of the cytochrome c3 redox potentials in all the c3 molecules studied here. However, the weaker (but significant) role of other titrating groups cannot be excluded, their importance and identity changing with the particular molecule under study. We also calculate the relative redox potentials of the four heme centers among the selected members of the c3 family, using a continuum electrostatic method that takes into account both solvation and interaction effects. Comparison of the calculated values with available data for the microscopic redox potentials was undertaken, the quality of the agreement being dependent upon the choice of the dielectric constant for the protein interior. We find that high dielectric constants give best correlations, while low values result in better magnitudes for the calculated potentials. The possibility that the crystallographic calcium ion in c3 from Desulfovibrio gigas may be present in the solution structure was tested, and found to be likely.

Crystal structure of the ternary complex of pig muscle phosphoglycerate kinase (PGK) with the substrate 3-phosphoglycerate (3-PG) and the Mg(2+) complex of beta,gamma-methylene-adenosine-5'-triphosphate (AMP-PCP), a nonreactive analogue of the nucleotide substrate, MgATP, has been determined by X-ray diffraction at 2.5 A resolution. The overall structure of the protein exhibits an open conformation, similar to that of the previously determined ternary complex of the pig muscle enzyme with beta,gamma-imido-adenosine-5'-triphosphate (AMP-PNP) in place of AMP-PCP (May, Vas, Harlos, and Blake (1996) Proteins 24, 292-303). The orientation and details of interactions of the nucleotide phosphates, however, show marked differences. The beta-phosphate is linked to the conserved Asp 218, i.e., to the N-terminus of helix 8, through the Mg(2+) ion; the previously observed interactions of the metal complex of AMP-PNP or ADP with the conserved Asn 336 and the N-terminus of helix 13 are completely absent. These structural differences are maintained themselves in solution studies. Inhibition and binding experiments show a slightly weaker interaction of PGK with MgAMP-PCP than with MgAMP-PNP: at pH 7.5, the K(d) values are 1.07 +/- 0.18 and 0.41 +/- 0.08 mM, respectively. The difference is further enhanced by 3-PG: the K(d) values are 2.80 +/- 0.66 and 0.68 +/- 0.11 mM, respectively. Thus, the previously observed weakening effect of 3-PG on nucleotide binding (Merli, Szilágyi, Flachner, Rossi, and Vas (2002) Biochemistry 41, 111-119) is more pronounced with MgAMP-PCP. The discordance between substrate analogues also shows up in thiol reactivity studies. In their binary complexes, both ATP analogues protect the fast-reacting thiols of PGK in helix 13 against modification to similar extent. In their ternary complexes, however, which also contain bound 3-PG, the protective effect of MgAMP-PCP, but not of MgAMP-PNP, is largely abolished. This indicates a much smaller effect of MgAMP-PCP on the conformation of helix 13, which is in good correlation with its altered mode of phosphate binding and the ensuing increase in the flexibility of helix 13, as shown by elevated crystallographic B-factors. The possible existence of alternative site(s) for binding of the nucleotide phosphates may have functional relevance.

We present a generalization of the reaction coordinate driven method to find reaction paths and transition states for complicated chemical processes, especially enzymatic reactions. The method is based on the definition of a subset of chemical coordinates; it is simple, robust, and suitable to calculate one or more alternative pathways, intermediate minima, and transition-state geometries. Though the results are approximate and the computational cost is relatively high, the method works for large systems, where others often fail. It also works when a certain reaction path competes with others having a lower energy barrier. Accordingly, the procedure is appropriate to test hypothetical reaction mechanisms for complicated systems and provides good initial guesses for more accurate methods. We present tests on a number of simple reactions and on several complicated chemical transformations and compare the results with those obtained by other methods. Calculation of the reaction path for the enzymatic hydrolysis of the substrate by dUTPase for an active-site model with 85 atoms, including several loosely bound water molecules, indicates that the method is feasible for the study of enzyme mechanisms.