Synthesis and characterization of Desulfovibrio gigas rubredoxin and rubredoxin fragments.

The 52-residue Desulfovibrio gigas rubredoxin peptide chain has been synthesized and a procedure for chain folding around iron(II) developed. The folded, stable synthetic rubredoxin can be subjected to purification, and reversibly oxidized and reduced. Ultraviolet/visible absorption and CD spectra of both forms show all the same features as native D. gigas rubredoxin, and the symmetric and asymmetric Fe-S stretching bands in the resonance Raman spectrum can be identified. In addition, the matrix-assisted laser desorption mass spectrum of a peptide sample exposed to trace amounts of iron is dominated by a peak at 5735Da very close to the value for the calculated molecular mass. Details in the ultraviolet/visible bandshape and mass spectrum, however, indicate remaining impurities. In comparison, a previously synthesized 25-residue rubredoxin fragment with the non-conserved positions 13-35 and 51-52 omitted and Val5-Glu50 anchored via glycine folds gives the correct molecular mass and ultraviolet/visible spectrum, but is much more labile than the 52-residue protein. This shows that non-conserved residues are crucial in protein folding and that chemical metalloprotein synthesis offers alternative prospects to microbiological protein engineering.

Synthesis and characterization of a 25-residue rubredoxin(II)-like metalloprotein and its valine-leucine mutant.

An iron-sulfur metalloprotein containing the 5-12 and 35-50 residues of Desulfovibrio gigas rubredoxin has been synthesized by Fmoc solid phase peptide synthesis and subsequent peptide folding. A Gly links the two residue chains between Val-5 and Glu-50. Sybyl Tripos structure optimization indicates only minor structural changes of the folded synthetic protein compared to the similar residue positions in the native protein. The UV-VIS spectrum of the reduced synthetic protein is very similar to that of native D. gigas rubredoxin and the molecular mass determined by laser mass spectrometry has the expected value (+/- 2D). No metal is transferred to the gas phase by the laser beam merely by mixing the peptide and iron(II), substantiating that the folding procedure is a necessary pre-requisite for protein formation. The Val-->Leu41 chemical mutant has also been synthesized and behaves in a closely similar fashion.

Resonance effects in strongly exothermic long-range electron transfer and their possible implications for the behaviour of site-directed mutant proteins.

Long-range electron transfer investigations of hemoproteins, blue copper and iron-sulphur proteins frequently rest on electronically excited metal centres. When the excitation energy approaches the oxidation or reduction potentials of intermediate residues the superexchange view normally used, however, fails and a variety of new dynamic features arise. These all involve population of the intermediate cation or anion residue states which can be partially or wholly vibrationally relaxed. We discuss suitable views and a new theoretical formalism for these phenomena. We also note some important implications for site-directed mutagenesis in long-range, strongly exothermic electron transfer processes.

Conformational dynamics and solvent viscosity effects in carboxypeptidase-A-catalyzed benzoylglycylphenyllactate hydrolysis.

We have used a new approach to the dynamics of hydrolytic metalloenzyme catalysis based on investigations of both external solvent viscosity effects and kinetic 2H isotope effects. The former reflects solvent and protein dynamics, and the nuclear reorganization distribution among damped protein motion and intramolecular friction-free nuclear motion. The isotope effect represents proton tunnelling and reorganization in the hydrogen bond network around the active site. We illustrate the approach by new spectrophotometric and pH-titration data for carboxypeptidase-A-catalyzed benzoylglycyl-L-phenyllactate hydrolysis. This substrate exhibits both a significant inverse fractional power law viscosity dependence over wide ranges controlled by glycerol and sucrose, and a kinetic 2H isotope effect of 1.65. The analogous benzoylglycylphenylalanine hydrolysis has a smaller isotope effect (1.3) and no viscosity dependence. Viscosity variation has no effect on the CD spectra in the 180-240-nm range. In terms of stochastic chemical rate theory, the data correspond to an enzyme-peptide substrate complex with a 'tight' structure protected from the solvent. In comparison, the enzyme-ester substrate complex is 'softer', strongly coupled to the solvent, and the rate-determining step is accompanied by proton transfer or by substantial reorganization in the hydrogen bonds near the active site.

Substrate specificity of solvent viscosity effects in carboxypeptidase A catalyzed peptide hydrolysis.

We have investigated the viscosity of carboxypeptidase A catalyzed Bz-Gly-Phe hydrolysis at pH 7.5 (Tris) and 0.5 mol.l-1 NaCl over the range 10-100 mp, varied by addition of glycerol or sucrose. In contrast to previous reports of strong viscosity effects on the corresponding Cbz-Ala-Ala-Ala hydrolysis, both the catalytic constant and the Michaelis constant are virtually independent of viscosity over the 10-fold range investigated. Furthermore, the CD spectra of carboxypeptidase A in the high-viscosity media point to no change in the alpha-helix and beta-sheet structure in these media. The data are compatible either with a compacter, more rigid enzyme-substrate structure or with a more prominent role of intramolecular nuclear reorganization compared to protein reorganization for Bz-Gly-Phe than for Cbz-Ala-Ala-Ala. These views can be given a preciser frame in terms of stochastic chemical rate theory.