Structure and dynamics of MarA-DNA complexes: an NMR investigation.

An unanswered question regarding gene regulation is how certain proteins are capable of binding to DNA with high affinity at specific but highly degenerate consensus sequences. We have investigated the interactions between the Escherichia coli transcription factor, MarA, and its diverse binding sites using NMR techniques. Complete resonance assignments for the backbone of the MarA protein complexed with DNA oligomers corresponding to its binding sites at the mar, fumC, micF and the fpr promoters were obtained. Secondary structure analysis based on chemical shifts reveals that regions identified as helical in the X-ray structure of the MarA-mar complex are present in the solution structure, although some of the helices are less well defined. The chemical shift differences between the four complexes confirm that helix 3 and helix 6 constitute the major DNA-binding elements. However, in striking contrast with the X-ray data: (i) the protein appears to be present in two or more conformations in each of the complexes; (ii) no slowly exchanging N(zeta)H(2) protons (indicative of hydrogen bonded groups) were observed by NMR for the two arginine residues proposed to form crucial hydrogen bonds in the X-ray structure; and (iii) regions at the N terminus, not observed in the X-ray structure, may be involved in DNA-binding. Taken together, the NMR results indicate that MarA in its complexes with DNA target sites is in a highly dynamic state, allowing for small but significant rearrangements of the side-chains and/or backbone to bind to the different DNA sequences.

Voltammetric probes of cytochrome electroreactivity: the effect of the protein matrix on outer-sphere reorganization energy and electronic coupling probed through comparisons with the behavior of porphyrin complexes.

Using surface-modified electrodes composed of omega-hydroxyalkanethiols, an experimentally based value for the inner-sphere reorganization energy of the bis(imidazole)iron porphyrin system has been obtained by examining the solvent dependence of the reorganization energy of bis(N-methylimidazole)meso-tetraphenyl iron porphyrin. The value obtained (0.41 +/- 0.06 eV) is remarkably similar to values we have recently reported for the reorganization energy of cytochrome b(5) (0.43 +/- 0.02 eV) and cytochrome c (0.58 +/- 0.06 eV). This strongly suggests that the protein matrix mimics the behavior of a low dielectric solvent and effectively shields the heme from the solvent. The effect of the orientation of the heme relative to the electrode was also explored by sytematically varying the steric bulk of the axial ligands. On the basis of a good linear correlation between the electronic coupling and the cosine of the angle between the heme plane and the surface of the electrode, it is suggested that a parallel orientation of the heme yields a maximum in the electronic coupling. Relevance to interheme protein electron transfer is discussed.

The origin of differences in the physical properties of the equilibrium forms of cytochrome b5 revealed through high-resolution NMR structures and backbone dynamic analyses.

On the basis of a comparison of high-resolution solution structures calculated for both equilibrium forms of rat ferrocytochrome b5, differences in reduction potential and thermodyanmic stability have been characterized in terms of significant structural and dynamic differences between the two forms. The dominant difference between A and B conformations has long been known to be due to a 180 degrees rotation of the heme in the binding pocket about an axis defined by the alpha- and gamma-meso carbons, however, the B form has not been structurally characterized until now. The most significant differences observed between the two forms were the presence of a hydrogen bond between the 7-propionate and the S64 amide in the A form but not the B form and surprisingly a displacement of the heme out of the binding pocket by 0.9 A in the B form relative to the A form. The magnitude of other factors which could contribute to the known difference in reduction potentials in the bovine protein [Walker, F. A., Emrick, D., Rivera, J. E., Hanquet, B. J., and Buttlaire, D. H. (1988) J. Am. Chem. Soc. 110, 6234-6240], such as differences in the orientation of the axial imidazoles and differences in hydrogen bond strength to the imidazoles, have been evaluated. The dominant effector of the reduction potential would appear to be the lack of the hydrogen bond to the S64 amide in the B form which frees up the propionate to charge stabilize the iron in the oxidized state and thus lower the reduction potential of the B form. The structure we report for the A form, based on heteronuclear NMR restraints, involving a total of 1288 restraints strongly resembles both the X-ray crystal structure of the bovine protein and a recently reported structure for the A form of the rat protein based on homonuclear data alone [Banci, L., Bertini, I., Ferroni, F., and Rosato, A. (1997) Eur. J. Biochem. 249, 270-279]. The rmsd for the backbone atoms of the A form is 0.54 A (0.92 A for all non-hydrogens). The rmsd for the backbone of the B form is 0.51 A (0. 90 A for all non-hydrogen atoms). An analysis of backbone dynamics based on a model-free analysis of 15N relaxation data, which incorporated axially symmetric diffusion tensor modeling of the cytochrome, indicates that the protein is more rigid in the reduced state relative to the oxidized state, based on a comparison with order parameters reported for the bovine protein in the oxidized state [Kelly, G. P., Muskett, F. W., and Whitford, D. (1997) Eur. J. Biochem. 245, 349-354].

Characterization of a site-directed mutant of cytochrome b5 designed to alter axial imidazole ligand plane orientation.

Mutants of cytochrome b5 were designed to achieve reorientation of individual axial imidazole ligands. The orientation of the axial ligand planes is thought to modulate the reduction potential of bis(imidazole) axially ligated heme proteins. The A67V mutation achieved this goal through the substitution of a bulkier, hydrophobic ligand for a residue, in the sterically hindered hydrophobic heme binding pocket. Solution structures of mutant and wild-type proteins in the region of the mutation were calculated using restraints obtained from 1H and 15N 2D homonuclear and heteronuclear NMR spectra and 1H-15N 3D heteronuclear NMR spectra. More than 10 restraints per residue were used in the refinement of both structures. Average local rmsd for 20 refined structures was 0.30 A for the wild-type structure and 0.38 A for the A67V mutant. The transfer of amide proton resonance assignments from wild-type to the mutant protein was achieved through overlays of 15N-1H heteronuclear correlation spectra of the reduced proteins. Side chain assignments and sequential assignments were established using conventional assignment strategies. Calculation of the orientation of the components of the anisotropic paramagnetic susceptibility tensor, using methods similar to procedures applied to the wild-type protein, shows that the orientation of the in-plane components are identical in the wild-type and mutant proteins. However, the orientation of the z-component of the susceptibility tensor calculated for the mutant protein differs by 17 degrees for the A-form and by 11 degrees for the B-form from the orientation calculated for the wild-type protein. The rotation of the z-component of the susceptibility tensor (toward the delta meso proton) is in the same direction and is of the same magnitude as the rotation of the H63 imidazole ring induced by mutation.