Integrity of thermus thermophilus cytochrome c552 synthesized by Escherichia coli cells expressing the host-specific cytochrome c maturation genes, ccmABCDEFGH: biochemical, spectral, and structural characterization of the recombinant protein

We describe the design of Escherichia coli cells that synthesize a structurally perfect, recombinant cytochrome c from the Thermus thermophilus cytochrome c552 gene. Key features are (1) construction of a plasmid-borne, chimeric cycA gene encoding an Escherichia coli-compatible, N-terminal signal sequence (MetLysIleSerIleTyrAlaThrLeu AlaAlaLeuSerLeuAlaLeuProAlaGlyAla) followed by the amino acid sequence of mature Thermus cytochrome c552; and (2) coexpression of the chimeric cycA gene with plasmid-borne, host-specific cytochrome c maturation genes (ccmABCDEFGH). Approximately 1 mg of purified protein is obtained from 1 L of culture medium. The recombinant protein, cytochrome rsC552, and native cytochrome c552 have identical redox potentials and are equally active as electron transfer substrates toward cytochrome ba3, a Thermus heme-copper oxidase. Native and recombinant cytochromes c were compared and found to be identical using circular dichroism, optical absorption, resonance Raman, and 500 MHz 1H-NMR spectroscopies. The 1.7 A resolution X-ray crystallographic structure of the recombinant protein was determined and is indistinguishable from that reported for the native protein (Than, ME, Hof P, Huber R, Bourenkov GP, Bartunik HD, Buse G, Soulimane T, 1997, J Mol Biol 271:629-644). This approach may be generally useful for expression of alien cytochrome c genes in E. coli.

NMR investigation of ferricytochrome c unfolding: detection of an equilibrium unfolding intermediate and residual structure in the denatured state

Horse ferricytochrome c (cyt c) undergoes exchange of one of its axial heme ligands (Met-80) for one or more non-native ligands under denaturing conditions. We have used (1)H NMR spectroscopy to detect two conformations of paramagnetic cyt c with non-native heme ligation through a range of urea concentrations. One non-native form is an equilibrium unfolding intermediate observed under partially denaturing conditions and is attributed to replacement of Met-80 with one or more Lys side chains. The second non-native form, in which the native Met ligand is replaced by a His, is observed under strongly denaturing conditions. Thermodynamic analysis of these data indicates a relatively small DeltaG (17 kJ/mol) for the transition from native to the Lys-ligated intermediate and a significantly larger DeltaG (47 kJ/mol) for the transition from native to the His-ligated species. Although CD and fluorescence data indicate that the equilibrium unfolding of cyt c is a two-state process, these NMR results implicate an intermediate with His-Lys ligation.

Secondary structure extensions in Pyrococcus furiosus ferredoxin destabilize the disulfide bond relative to that in other hyperthermostable ferredoxins. Global consequences for the disulfide orientational heterogeneity

The single cubane cluster ferredoxin (Fd) from the hyperthermophilic archaeon Pyrococcus furiosus (Pf) possesses several unique properties when compared even to Fds from other hyperthermophilic archaea or bacteria. These include an equilibrium molecular heterogeneity, a six- to seven-residue increase in size, an Asp rather than the Cys as one cluster ligand, and a readily reducible disulfide bond. NMR assignments and determination of both secondary structure and tertiary contacts remote from the paramagnetic oxidized cluster of Pf 3Fe Fd with an intact disulfide bond reported previously (Teng Q., Zhou, Z. H., Smith, E. T., Busse, S. C., Howard, J. B. Adams, M. W. W., and La Mar, G. (1994) Biochemistry 33, 6316-6328) are extended here to the 4Fe oxidized cluster WT (1H and 15N) and D14C (1H only) Fds with an intact disulfide bond and to the 4Fe oxidized WT Fd (1H and 15N) with a cleaved disulfide bond. All forms are shown to possess a long (13-member) alpha-helix, two beta-sheets (one double-, one triple-stranded), and three turns outside the cluster vicinity, each with tertiary contacts among themselves as found in other Fds. While the same secondary structural elements, with similar tertiary contacts, are found in other hyperthermostable Fds, Pf Fd has two elements, the long helix and the triple-stranded beta-sheet, that exhibit extensions and form multiple tertiary contacts. All Pf Fd forms with an intact disulfide bond exhibit a dynamic equilibrium heterogeneity which is shown to modulate a hydrogen-bonding network in the hydrophobic core that radiates from the Cys21-Cys48 disulfide bond and encompasses residues Lys36, Val24, Cys21, and Cys17 and the majority of the long helix. The heterogeneity is attributed to population of the alternate S and R chiralities of the disulfide bond, each destabilized by steric interactions with the extended alpha-helix. Comparison of the chemical shifts and their temperature gradients reveals that the molecular structure of the protein with the less stable R disulfide resembles that of the Fd with a cleaved disulfide bond. Both cluster architecture (3Fe vs 4Fe) and ligand mutation (Cys for Asp14) leave the disulfide orientational heterogeneity largely unperturbed. It is concluded that the six- to seven-residue extension that results in a longer helix and larger beta-sheet in Pf Fd, relative to other hyperthermostable Fds, more likely serves to destabilize the disulfide bond, and hence make it more readily reducible, than to significantly increase protein thermostability.

Solution structure of oxidized Saccharomyces cerevisiae iso-1-cytochrome c

The solution structure of oxidized Saccharomycescerevisiae Cys102Ser iso-1-cytochromechas been determined using 1361 meaningful NOEs (of 1676 total) after extending the published proton assignment [Gao, Y., et al. (1990) Biochemistry 29, 6994-7003] to 77% of all proton resonances. The NOE patterns indicate that secondary structure elements are maintained upon oxidation in solution with respect to the solid state and solution structures of the reduced species. Constraints derived from the pseudocontact shifts [diamagnetic reference shift values are those of the reduced protein [Baistrocchi, P., et al. (1996) Biochemistry 35, 13788-13796]] were used in the final stages of structure calculations. After restrained energy minimization with constraints from NOEs and pseudocontact shifts, a family of 20 structures with rmsd values of 0.58 +/- 0.08 and 1.05 +/- 0.10 A (relative to the average structure) for the backbone and all heavy atoms, respectively, was obtained. The solution structure is compared with the crystal structure and the structures of related systems. Twenty-six amide protons were detected in the NMR spectrum 6 days after the oxidized lyophilized protein was dissolved in D2O (pH 7.0 and 303 K); in an analogous experiment, 47 protons were observed in the spectrum of the reduced protein. The decrease in the number of nonexchanging amide protons, which mainly are found in the loop regions 14-26 and 75-82, confirms the greater flexibility of the structure of oxidized cytochrome c in solution. Our finding of increased solvent accessibility in these loop regions is consistent with proposals that an early step in unfolding the oxidized protein is the opening of the 70-85 loop coupled with dissociation of the Met80-iron bond.

Three-dimensional solution structure of Saccharomyces cerevisiae reduced iso-1-cytochrome c

Two-dimensional 1H NMR spectra of Saccharomyces cerevisiae reduced iso-1-cytochrome c have been used to confirm and slightly extend the assignment available in the literature. 1702 NOESY cross-peaks have been assigned, and their intensities have been measured. Through the program DIANA and related protocols (Güntert, 1992), a solution structure has been obtained by using 1442 meaningful NOEs and 13 hydrogen-bond constraints. The RMSD values with respect to the mean structure for the backbone and all heavy atoms for a family of 20 structures are 0.61 +/- 0.09 and 0.98 +/- 0.09 A, the average target function value being as small as 0.57 A2. The larger number of slowly exchanging amide NHs observed in this system compared to that observed in the cyanide derivative of oxidized Ala 80 cytochrome c suggests that the oxidized form is much more flexible and that the backbone protons are more solvent accessible. Comparison of the present structure with the crystal structures of reduced yeast cytochrome c and of the complex between cytochrome c peroxidase and oxidized yeast cytochrome c reveals substantial similarity among the backbone conformations but differences in the residues located in the region of protein-protein interaction. Interestingly, in solution the peripheral residues involved in the interaction with cytochrome c peroxidase are on average closer to the position found in the crystal structure of the complex than to the solid state structure of the isolated reduced from.

Three-dimensional solution structure of the cyanide adduct of a Met80Ala variant of Saccharomyces cerevisiae iso-1-cytochrome c. Identification of ligand-residue interactions in the distal heme cavity

The 1H NMR spectrum of the the cyanide adduct of a triply mutated Saccharomyces cerevisiae iso-1-cytochrome c (His39Gln/Met80Ala/Cys102Ser) in the oxidized form has been assigned through 1D NOE and 2D COSY, TOCSY, NOESY, and NOE-NOESY experiments; 562 protons out of a total of 683 have been assigned. The solution structure, the first of a paramagnetic heme protein, was determined using 1426 meaningful NOE constraints out of a total of 1842 measured NOEs. The RMSD values at the stage of restrained energy minimization of 17 structures obtained from distance geometry calculations are 0.68 +/- 0.11 and 1.32 +/- 0.14 A for the backbone and all heavy atoms, respectively. The quality, in terms of RMSD, of the present structure is the same as that obtained for the solution structure of the diamagnetic horse heart ferrocytochrome c [Qi, P. X., et al. (1994) Biochemistry 33, 6408-6419]. The secondary structure elements and the overall folding in the variant are observed to be the same as those of the wild-type protein for which the X-ray structure is available. However, the replacement of the methionine axial ligand with an alanine residue creates a ligand-binding "distal cavity". The properties of the distal cavity seen in this solution structure are compared to those of other heme proteins.

pH-dependent equilibria of yeast Met80Ala-iso-1-cytochrome c probed by NMR spectroscopy: a comparison with the wild-type protein

BACKGROUND

Cytochrome c has five distinct pH-dependent conformational states, including two alkaline forms of unknown structure. It is believed that in both of the alkaline forms a Lys residue is ligated to the heme, but the identity of the Lys residue is different. Exchange between these forms would require extensive structural rearrangement. Mutation of the heme axial ligand (Met80) to Ala in Saccharomyces cerevisiae iso-1-cytochrome c yields a protein (Ala80cyt c) capable of binding exogenous ligands such as dioxygen and cyanide. We have analyzed the 1H NMR spectra of this mutant at various pH values in the hope of gaining insight into the structure of the acidic and alkaline forms of native cytochrome c.

RESULTS

The pH dependence of the 1H NMR spectrum of ferriAla80cyt c is consistent with the high-spin/low-spin transition (pKa = 6.5) observed by absorption spectroscopy. The T1 values for the low-spin form are consistent with OH ligation, as inferred previously from absorption and electron paramagnetic resonance spectroscopic results. The pH-dependent equilibria of ferriAla80cyt c differ from those of the wild-type protein. Both Ala80 and wild-type ferricyt c appear to have the same iron coordination at low pH (approximately equal to 2), while only one alkaline form of Ala80cyt c (versus two for WTcyt c) was detected.

CONCLUSIONS

The differences between the pH dependence of the 1H NMR spectra of Ala80cyt c and those of the wild-type protein demonstrate that the heme axial ligands influence the relative energies of the conformational states of cytochrome c. The results are consistent with the notion that a large rearrangement is required to switch between the two alkaline forms.

Structurally engineered cytochromes with unusual ligand-binding properties: expression of Saccharomyces cerevisiae Met-80-->Ala iso-1-cytochrome c

A strategy has been developed to express and purify a recombinant, nonfunctional axial-ligand mutant of iso-1-cytochrome c (Met-80-->Ala) in Saccharomyces cerevisiae in quantities necessary for extensive biophysical characterization. It involves coexpressing in the same plasmid (YEp213) the nonfunctional gene with a functional gene copy for complementation in a selective medium. The functional gene encodes a product with an engineered metal-chelating dihistidine site (His-39 and Leu-58-->His) that enables efficient separation of the two isoforms by immobilized metal-affinity chromatography. The purified Met-80-->Ala protein possesses a binding site for dioxygen and other exogenous ligands. Absorption spectra of several derivatives of this mutant show striking similarities to those of corresponding derivatives of horseradish peroxidase, myoglobin, and cytochrome P450. The use of a dual-gene vector for cytochrome c expression together with metal-affinity separation opens the way for the engineering of variants with dramatically altered structural and catalytic properties.