J. Phys. Chem. B 2009, 113, 2193–2200
2193
Vibrational Dynamics of Iron in Cytochrome c Bogdan M. Leu,†,‡ Tom H. Ching,† Jiyong Zhao,‡ Wolfgang Sturhahn,‡ E. Ercan Alp,‡ and J. Timothy Sage*,† Department of Physics and Center for Interdisciplinary Research on Complex Systems, Northeastern UniVersity, Boston, Massachusetts 02115, and AdVanced Photon Source, Argonne National Laboratory, Argonne, Illinois 60439 ReceiVed: July 24, 2008; ReVised Manuscript ReceiVed: December 11, 2008
Nuclear resonance vibrational spectroscopy (NRVS) and Raman spectroscopy on 54Fe- and 57Fe-enriched cytochrome c (cyt c) identify multiple bands involving vibrations of the heme Fe. Comparison with predictions from Fe isotope shifts reveals that 70% of the NRVS signal in the 300-450 cm-1 frequency range corresponds to vibrations resolved in Soret-enhanced Raman spectra. This frequency range dominates the “stiffness”, an effective force constant determined by the Fe vibrational density of states (VDOS), which measures the strength of nearest-neighbor interactions with Fe. The stiffness of the low-spin Fe environment in both oxidation states of cyt c significantly exceeds that for the high-spin Fe in deoxymyoglobin, where the 200-300 cm-1 frequency range dominates the VDOS. This situation is reflected in the shorter Fe-ligand bond lengths in the former with respect to the latter. The longer Fe-S(Met80) in oxidized cyt c with respect to reduced cyt c leads to a decrease in the stiffness of the iron environment upon oxidation. Comparison with NRVS measurements allows us to assess assignments for vibrational modes resolved in this region of the heme Raman spectrum. We consider the possibility that the 372 cm-1 band in reduced cyt c involves the Fe-S(Met80) bond. 1. Introduction Cytochrome c (cyt c) is a relatively small heme protein (104 amino acids, MW ) 12 400 Da) that mediates electron transfer between two membrane-bound protein complexes in the mitochondrial electron-transport chain1-4 and also plays a crucial signaling role in cellular apoptosis.5,6 During electron transfer, the heme iron cycles between Fe(II) and Fe(III). In both oxidation states, the protein ligates the heme iron by a methionine (Met80) and a histidine (His18). The heme prosthetic group connects to the protein via two additional bonds: the thioether linkages between vinyl groups of the porphyrin ring and two cysteines (Cys14 and Cys17; Figure 1). Both oxidation states adopt the same fold at neutral pH, and the structures are very similar, both in solutions7-11 and in crystals.12-16 Structural changes include a small increase of the Fe-S(Met80) bond length upon oxidation.17 Solution structures derived from NMR measurements reveal a 140 pm rms deviation between backbone atoms in Fe(II) and Fe(III) cyt c from horse.7,8 Small-angle X-ray scattering studies were interpreted as evidence for an increased radius of gyration for the oxidized state in low ionic strength solution.18 However, more recent analysis of X-ray scattering from cyt c solutions attributed oxidation state differences in the small-angle signal from low ionic strength solutions to unscreened interparticle interactions.19 Consistent with this result, a crystal structure of Fe(III) cyt c at low ionic strength16 resembles that determined at high salt concentration. Hydrodynamic, thermodynamic, and spectroscopic properties, as well as ion binding and chemical reactivity, are influenced by the oxidation state (ref 20 and references therein). * Corresponding author: e-mail
[email protected], phone (617)-373-2908; fax (617)-373-2943. † Northeastern University. ‡ Argonne National Laboratory.
Fe-ligand vibrations are useful probes of the metal environment in heme proteins but can be difficult to identify in congested biomolecular vibrational spectra, even with the selectivity provided by Raman scattering in resonance with the heme. For cyt c, resonance Raman spectroscopy is a useful tool for identifying conformational changes,21-24 but structural interpretations of the individual bands remains challenging. Nuclear resonance vibrational spectroscopy (NRVS) is emerging as an even more selective vibrational probe, revealing the complete vibrational density of states (VDOS) for 57Fe25-30 or other Mo¨ssbauer nuclei.31,32 NRVS investigations of myoglobin (Mb) have revealed Fe-ligand vibrations that contribute weakly to the Raman signal,33-35 presumably because of the restrictive selection rules that apply to the highly symmetric heme. In this study, complementary data provided by NRVS and resonance Raman spectroscopy on reduced 54Fe- and 57Feenriched cyt c identify bands in the rich Raman spectra that are due to heme Fe vibrations. We use Raman frequency shifts upon isotope substitution to predict the contributions of these modes to the Fe VDOS. Unexpectedly, comparison with the measured VDOS reveals that ∼70% of the NRVS signal above 300 cm-1 corresponds to vibrations resolved in the Soret-enhanced Raman spectrum. On the other hand, NRVS determines the contributions of these modes to the stiffness, an effective force constant expected to correlate with Fe-nearest-neighbor bond lengths. The dominant contributions to the deoxyMb VDOS appear at lower frequencies (200-300 cm-1), which leads to a lower stiffness of the heme iron environment in the high-spin deoxyMb with respect to low-spin cyt c. As expected, stiffness is inversely related to Fe-ligand bond lengths, the longer Fe-N bonds in deoxyMb with respect to cyt c being reflected in the lower stiffness of the former with respect to the latter. On the other hand, the longer Fe-S bond in Fe(III) cyt c with respect to
10.1021/jp806574t CCC: $40.75 2009 American Chemical Society Published on Web 01/27/2009
2194 J. Phys. Chem. B, Vol. 113, No. 7, 2009
Leu et al.
Figure 1. Left: cytochrome c (Protein Data Bank89 code 1HRC).90 This structure was generated with the program RasMol.91 Right: cytochrome c heme. Hydrogen atoms were omitted for clarity. This structure was generated with the program Molekel.92 Color scheme: orange ) iron, green ) nitrogen, red ) oxygen, blue ) sulfur, gray ) carbon.
Figure 2. Upper and middle panels: VDOS of Fe(II) cyt c and Fe(III) cyt c. Lower panel: VDOS of Fe(II), Fe(III) cyt c, and deoxyMb represented in terms of the integrand D(νj)νj2 of the mean Fe-ligand bond stiffness (eq 5). The area determines the Fe-ligand bond stiffness.
Fe(II) cyt c is the most important contributor to the decreased stiffness upon oxidation. The maximum NRVS signal of reduced cyt c corresponds to the biggest shift in the Raman data (372 cm-1) upon isotope substitution and may involve the Fe-S bond. Stretching of the Fe-N bond to the histidine may contribute to several vibrations, a situation we encountered previously in six-coordinated proteins and model compounds.36 In a separate publication,37 we used the 57Fe-reconstituted protein to evaluate the mechanical effect of the thioether links. In the present paper, we explore the vibrational structure in greater detail, using NRVS measurements on the 57Fe-enriched cyt c together with frequency shifts in the Raman spectrum upon Fe isotope substitution.
2. Experimental Methods 2.1. Cyt c Reconstitution with 57Fe and 54Fe. Because the heme c in cytochrome c is covalently bonded to the protein via thioether linkages to Cys14 and Cys17, it cannot easily be removed as in myoglobin and hemoglobin. We employed the ferrous sulfate-hydrochloric acid method38 to extract the iron from the heme. We adapted metal reinsertion procedures from the literature39-41 to insert isotopes 57Fe and 54Fe in the resulting porphyrin cyt c. The complete preparation procedure is described elsewhere.42 2.2. Spectroscopic Measurements and Analysis. 2.2.1. NRVS Measurements. 57Fe NRVS measurements were performed at sector 3-ID-D of the Advanced Photon Source at Argonne National Laboratory as described in detail elsewhere.26 Frozen solutions were loaded into polyethylene sample cups and mounted on a cryostat cooled by a flow of liquid He, with X-ray access through a beryllium dome. The incident monochromatic 14.4 keV X-ray had a flux of ∼109 Hz, and the experimental resolution was 8 cm-1 (1 meV). Results presented here are averages of 14-26 energy scans, and comparison of individual scans found no spectroscopic changes due to radiation damage. Data analysis described below confirmed temperatures of 87 K for Fe(II) cyt c and 68 K for Fe(III) cyt c. 2.2.2. Raman Measurements. Raman measurements on native and 57Fe- and 54Fe-reconstituted cyt c were performed at room temperature. Previous Raman measurements on cyt c found significant line narrowing with reduced temperature, but reported no frequency changes,43 supporting our comparison of room temperature Raman measurements with cryogenic NRVS measurements. Raman scattering was excited by the 413.1 nm line of a krypton laser and detected using a J-Y LabRam HR Raman microscope. The beam power at the sample was 15 mW. The frequency calibration was verified using the rich Raman spectrum of fenchone as a standard. The instrument resolution was 2.8 cm-1. 2.2.3. Data Analysis. The measured NRVS signal consists of a central resonance due to recoilless excitation of the 57Fe nuclear excited state at E0 ) 14.4 keV, accompanied by a series of sidebands corresponding to creation or annihilation of vibrational quanta of frequency νj, displaced from the recoilless absorption by an energy hcνj.27,44 The program PHOENIX,45 working under the assumptions that samples are isotropic, harmonic, and Debye-like at low energies, removes temperature
Vibrational Dynamics of Iron in Cytochrome c
J. Phys. Chem. B, Vol. 113, No. 7, 2009 2195
factors and multiphonon contributions to yield the normalized one-phonon contribution S1′(νj) to the NRVS signal and the Feweighted vibrational density of states (VDOS)
[
( )]
ν¯ S1′(ν¯ ) hcν¯ D(ν¯ ) ) 3 1 - exp ν¯ R f kBT
(1)
which defines the vibrational properties at all temperatures for a harmonic system.27 The sample temperatures are confirmed by requiring that the ratio S′(νj)/S′(-νj) equal the Boltzmann factor exp(hcνj/kBT). Measurements on a randomly oriented sample such as a solution or polycrystalline powder yield the total VDOS D(νj) summed over the three Cartesian directions. Each mode R contributes to D(νj) an area ejR2, with j ) Fe, equal to the fraction of the kinetic energy associated with motion of the iron atom.28 The total density of states is given by
D(ν¯ ) )
∑ ejR2L (ν¯ - ν¯ R)
(2)
with
∫ D(ν¯ ) dν¯ ) 3
(3)
Mode composition factors ejR2 can be estimated from reported frequency shifts in mode R resulting from small changes in the mass of atom j according to the formula28
1 - ejR2 ) d(ln ν¯ R)/d(ln mj) 2
(4)
Equation 4 is generally applicable in the harmonic approximation, as long as the mode character is insensitive to mass changes, and provides a useful method to estimate mode character from isotopic frequency shifts. 3. Results and Discussion 3.1. Stiffness of the Fe Environment. Figure 2 (upper and middle panels) presents the VDOS derived from NRVS data recorded on reduced and oxidized cyt c. The dominant features appear in the 300-400 cm-1 region, a frequency range comparable to other low-spin heme complexes36,46,47 but much higher than for high-spin complexes.33,47 Inspection reveals that this feature shifts down from 340-400 cm-1 in reduced cyt c to 320-380 cm-1 in oxidized cyt c. Comparison with Raman scattering measurements, in the following section, reveals significant unresolved structure in this region. However, we begin by considering the mean stiffness48,49 of the iron environment
ks ) mFe〈ω2〉 ) mFe
1 3
∫ D(ω)ω2 dω
(5)
as determined directly from the VDOS and reported in Table 1. (Here, ω ) 2πcνj.) The stiffness is the force constant for displacement of the Fe nucleus along the direction of the X-ray beam, with other nuclei fixed at their equilibrium positions. The factor of 3 in eq 5 acknowledges an averaging over all directions for the randomly oriented molecules in the frozen solutions considered here. The stiffness can also be determined directly
TABLE 1: Average Fe-Ligand Bond Stiffness and Selected Bond Lengths for Horse Heart Cyt c and DeoxyMb Fe(II) cyt c
Fe(III) cyt c
deoxyMb
stiffness (pN/pm)
322 ( 17
284 ( 17
190 ( 20
Fe-S(Met80) (pm) Fe-Nε (pm) Fe-Npyr (pm)
229 (2)17 200 (2)17 199 (2)17
233 (2)17 198 (3)17 198 (2)17
220 (4)55 205 (1)55
from the measured excitation probability48 if needed, without invoking the assumption of harmonic behavior. The quantity νj2D(νj) displayed in the lower panel of Figure 2 is proportional to the integrand of eq 5. These stiffness spectra illustrate the dominant contribution of the 300-400 cm-1 frequency region to the stiffness of the Fe environment in cyt c. The corresponding spectrum for deoxyMb is included for comparison and reflects a dominant contribution centered near 250 cm -1, a smaller area, and a correspondingly lower stiffness (Table 1). In contrast, frequencies below 100 cm-1 largely determine the resilience kr ) mFe〈ω-2〉-1, a distinct force constant that characterizes the magnitude of thermal fluctuations of the Fe, as described in a separate publication.37 The nearest neighbors of the Fe atom can be expected to have the primary influence on the stiffness, which thus probes the local structure. This expectation implicitly motivates many applications of vibrational spectroscopy, and a quantitative inverse correlation between frequency and bond length50,51 may apply to vibrations localized on a single bond.52,53 For example, empirical valence bond parameters used to reproduce NRVS data on rubredoxin featured a significantly decreased force constant for stretching of the Fe-S bond in the reduced state, consistent with a reported 5 pm increase in bond length.54 We anticipate that the stiffness may monitor changes in Fe-ligand bond length even in complex situations where it is not possible to assign vibrational frequencies to individual bonds and where empirical approaches may not provide unique force constants. Comparison of the stiffness with structural parameters derived from EXAFS measurements17,55-57 and summarized in Table 1 encourages this expectation. In particular, both the equatorial Fe-Npyr bonds and the axial Fe-Nε(His) bond are significantly longer in deoxyMb, in which the Fe is high spin, than in cyt c, which has a low-spin Fe. The increased bond lengths correspond to a large downshift of the dominant NRVS signal in deoxyMb (Figure 2, lower panel) and a correspondingly reduced stiffness. Reduced vibrational frequencies in the high-spin state are commonly observed for iron complexes and contribute to the entropy changes responsible for the temperature dependence of spin-crossover complexes.58,59 The decreased stiffness observed upon oxidation of cyt c (Table 1) is smaller, but still significant, reflecting more modest structural changes. The Fe-S bond to Met80 exhibits the biggest change upon oxidation17 (Table 1). Other studies reveal some quantitative variation in the absolute Fe-S bond length, with a value as large as 241 pm reported for Fe(III) cyt c,56 but the qualitative increase upon oxidation appears to be reproducible.17,56,57 In contrast, smaller (1-2 pm) decreases reported for the five remaining Fe-N bonds17 are not consistently reproduced in other studies56,57 and may lie within the experimental uncertainty. Thus, the reduced stiffness upon oxidation of cyt c tracks small structural changes that challenge the sensitivity of structural techniques, with the increased Fe-S bond length likely to make the primary contribution. Chemical criteria also suggest that the Fe-S(Met80) bond is stronger in the reduced form, where it is more resistant to displacement by exogenous ligands (ref 20 and references therein).
2196 J. Phys. Chem. B, Vol. 113, No. 7, 2009
Leu et al.
Figure 3. Raman spectra of 57Fe and 54Fe(II) cyt c and their spectral difference demonstrate the presence of numerous Fe modes in the 340-420 cm-1 region.
NRVS measurements on cyt f also found that the stiffness of the low-spin Fe environment was significantly larger than in high-spin forms of myoglobin.49 However, stiffness differences between oxidized and reduced states of cyt f were within the experimental uncertainty. Although replacement of the Met sulfur in cyt c with the nitrogen of the amino terminus as axial ligand in cyt f might be expected to affect the Fe environment, the slightly reduced stiffness of Fe(II) cyt c (Table 1) compared to the 342 ( 18 pN/pm value reported for Fe(II) cyt f49 is not statistically significant. The stiffness spectra (Figure 2, lower panel) clearly identify frequency regions that reflect the structure of the Fe environment. NRVS measurements on deoxyMb identified vibrations of both axial and equatorial Fe-N bonds in the 230-260 cm-1 range.33 For cyt c, vibrational markers for Fe-ligand bond strength are most likely to appear in the 300-420 cm-1 region. Because cyt c NRVS signals appear to be more congested than those previously reported, it is more challenging to resolve features associated with individual vibrations. Nevertheless, the stiffness responds to variations in Feligand bond length (Table 1) without the need to impose an empirical model or to assign individual vibrations. Moreover, comparison with higher resolution measurements on cyt c may identify specific frequencies sensitive to the strength of individual bonds. 3.2. Fe Vibrational Decomposition for Fe(II) Cyt c. Comparison with isotope shift measurements provides more detailed information on the vibrations contributing to the stiffness. Figure 3 presents Raman spectra of reduced 54Fe- and 57 Fe-reconstituted cyt c, excited at a 413.1 nm wavelength. Subtraction of these spectra reveals significant Fe isotope shifts for numerous bands in the 300-420 cm-1 region, while the 447, 479, 521, 538, 551, and 570 cm-1 bands do not shift upon 57 Fe f 54Fe isotope substitution, consistent with the absence of NRVS signal above 440 cm-1 (Figure 2). Figure 4 compares the Fe VDOS for Fe(II) cyt c (upper panel) with the Raman spectra for 57Fe- and 54Fe-reconstituted Fe(II) cyt c (middle and lower panels). The Raman band widths are relatively narrow in comparison with typical heme proteins in solution at room temperature, as noted previously for cyt c,43
Figure 4. VDOS of 57Fe(II) cyt c (upper panel) and Raman spectra of 57 Fe(II)- and 54Fe(II)-reconstituted cyt c (lower panels). The error bars represent the experimental statistical error in the NRVS experiment. Dashed curves indicate components of a fit to the Raman data, with the frequencies and mode composition factors compiled in Table 2. Solid and dot-dash traces in the upper panel indicate contributions of these modes to the Fe VDOS, predicted from the tabulated parameters, assuming that either one or two modes, respectively, contribute to the 320 cm-1 feature.
TABLE 2: Frequencies, Line Widths, and Mode Composition Factors of the Raman Bands for Cyt ca 57
Fe(II)
54
57
Fe(II)
Fe(III)
νj (cm-1)
fwhm (cm-1)
νj (cm-1)
fwhm (cm-1)
eFe2
307 348 358 372 382 392 401 414 422 447
9 10 9 8 10 8 12 7 10 16
307 348 359 377 383 394 403 415 423 447
8 8 7 10 15 6 11 8 9 16
0.11 0.06 0.14 0.47 0.08 0.18 0.13 0.06 0.1 0
νj (cm-1)
fwhm (cm-1)
305 349 363 375 383
11 10 9 8 7
399 414 420 447
10 7 8 15
∑eFe2 ) 1.33 a
The iron mode composition factors were determined from eq 4.
which facilitates deconvolution into individual lines. We used the Levenberg-Marquardt algorithm to fit the Raman spectra in the 300-450 cm-1 region with 10 Lorentzian bands, with all fitting parameters free to vary. Table 2, columns 1-5, lists their positions, line widths Γ, and the mode composition factors eFe2 determined from the isotope shifts according to eq 4. Attempts to fit to Gaussians required additional bands to obtain comparable fitting quality. Our frequencies for these 10 bands for natural abundance Fe(II) cyt c (not listed) agree within 1 cm-1 with other reported measurements.60,61 The mode composition factors estimated from the Raman shifts according to eq 4 determine the expected contribution to
Vibrational Dynamics of Iron in Cytochrome c
Figure 5. Experimentally determined VDOS for Fe(III) cyt c (upper panel) and Raman spectra of native Fe(III) and 57Fe(II) cyt c (lower panels). The error bars represent the experimental statistical error in the NRVS experiment. Dashed curves in the upper panel were generated from the band positions and line widths of native Fe(III) cyt c and the isotope shifts for 57Fe(II) cyt c (Table 2). Dashed curves in the middle and lower panels are the peaks corresponding to data fitting with Lorentzian functions.
the VDOS. We use line widths ΓNRVS ) ΓRaman + 5.2 cm-1, where ΓRaman are the line widths of the peaks in the Raman spectrum (second column in Table 2) to account for the 5.2 cm-1 difference between the resolutions of the two experiments. Figure 4, upper panel, shows the predicted individual (Lorentzian) peaks and their sum. The total eFe2 estimated for the peaks listed in Table 2 is 1.33, accounting for 70% of the integrated NRVS signal from 300 to 450 cm-1. 3.3. Vibrational Effects of Heme Oxidation. Previous interpretations60-62 have usually assumed a one-to-one correspondence between Fe(II) and Fe(III) cyt c vibrational modes in the 300-420 cm-1 region, based on similarities between Raman spectra in this region and their shifts due to isotope substitutions.62 We use the Raman spectra of the oxidized native and 57Fe-reconstituted reduced cyt c to test this assumption. Figure 5 and Table 2, columns 1 and 6, indicate putative correspondences60-62 between the peaks in the reduced and oxidized Raman spectra. We use the frequencies and line widths determined from fitting the Raman data for native Fe(III) cyt c (Figure 5, middle panel) and the eFe2 values for Fe(II) cyt c (Table 2) to predict the NRVS signal due to the Raman active vibrations and compare it with the experimentally determined VDOS (Figure 5, upper panel). The comparison is not inconsistent with similar mode character for most corresponding modes. The 375 cm-1 mode in Fe(III) cyt c is an exception, where the predicted NRVS signal clearly exceeds the experimentally measured VDOS. Comparison of the results in the top panels of Figures 4 and 5 indicates that the Fe amplitude of the 375 cm-1 mode in Fe(III) cyt c is reduced in comparison with the 372 cm-1 mode in Fe(II) cyt c. On the other hand, and in contrast with Fe(II) cyt c, the
J. Phys. Chem. B, Vol. 113, No. 7, 2009 2197 maximum of the Fe(III) cyt c VDOS near 340 cm-1 does not correspond to any observed Raman mode. 3.4. Summary and Comparison with Other Heme Proteins. NRVS and resonance Raman spectroscopy present complementary solutions to the problem of spectral congestion, which is central to vibrational studies of complex macromolecules. The NRVS and Raman results presented here selectively reveal vibrations of the heme active site of cyt c, free from spectral interference from the surrounding polypeptide and solvent. All vibrations involving Fe contribute to the VDOS resulting from NRVS measurements. On the other hand, Raman measurements provide higher spectral resolution at the present time. Quantitative comparison of these two methods reveals a more complete picture of the heme vibrations than is available from either method individually. For Fe(II) cyt c, we identify at least nine vibrational frequencies in the 300-450 cm-1 range that contribute to both NRVS and Raman signals (Table 2). Raman measurements on cyt c isotopomers resolve individual vibrational components of the NRVS signal. These do not completely account for the observed NRVS signal, which therefore contains contributions from additional Fe vibrations, particularly near 340 cm-1. All the Raman bands in the 300-420 cm-1 region shift upon isotope substitution (Figure 4), indicating that these modes involve Fe motion. As a consequence, there is a large (70%) Raman mode contribution to the NRVS signal above 300 cm-1. A previous study on reduced cyt f49 supplemented eight Raman lines observed in this region with seven additional peaks in order to fit the NRVS signal and also concluded that all Raman lines involve iron motion. However, no Fe isotope shift data were presented. Raman measurements on myoglobin have occasionally identified vibrational frequencies sensitive to Fe isotope.34,63,64 Identification of the Fe-His frequency in the Raman spectrum of deoxymyoglobin,63 as well as a mode with Fe-NO stretching character in MbNO,34 was facilitated by observation of a single Fe isotope-sensitive mode in the expected frequency region. In contrast, essentially all modes in the 300-420 cm-1 region of the Raman spectrum of Fe(II) cyt c exhibit varying degrees of sensitivity to the Fe isotope (Figures 3 and 4, Table 2). Fe-S vibrations of Cys-ligated hemes have been reported in this frequency region,65,66 and mixing of axial ligand vibrations with porphyrin vibrations might contribute to the observed vibrational complexity. However, notice that Fe motion along one direction accounts for only one-third of the Fe VDOS,33 whose total area is normalized to 3 (eq 3). Since ∑eFe2 > 1 for the modes listed in Table 2, Fe motion perpendicular to the axial bonds must also contribute to both Raman and NRVS signals in this region and may represent the primary contribution. For a 4-fold D4h-symmetric Fe environment, in-plane Fe modes occur as degenerate pairs with equal eFe2 values and orthogonal Fe motion, and they are Raman inactive in resonance with the heme Soret band. However, difference measurements reveal that in-plane Fe-Npyr vibrations contribute measurably to the MbNO Raman signal.34 Several factors reduce the nominal 4-fold symmetry of the heme. Splitting of the FeO tilting mode in hydroxymetmyoglobin as a result of oxygen protonation35 demonstrates that asymmetry in the axial ligand can lift the degeneracy of in-plane vibrations. However, frequency splittings induced by heme asymmetry in cyt c may be too small to resolve for some modes (see next section). The asymmetric thioether linkages not only perturb the approximate 4-fold symmetry of the heme but also induce substantial distortion of the porphyrin from its nominal planar
2198 J. Phys. Chem. B, Vol. 113, No. 7, 2009 geometry. The reduced symmetry may enable observation of vibrational modes that would otherwise be forbidden by selection rules and contribute to the increased vibrational complexity. A complex vibrational signal is observed in the same frequency region for cyt f,49 although the detailed structure differs, because of axial ligation with the nitrogen atom of the terminal amino group rather than a methionine. Thus, we expect that such vibrational complexity is probably a general feature of thioether-linked hemes. On the other hand, nitrophorin67 and horseradish peroxidase68,69 exhibit less vibrational complexity in this region, in spite of hemes with significantly higher distortion from planarity than myoglobin. As a result, we suggest that the thioether links play a more significant role in reducing the degree of vibrational symmetry than does the distortion of the heme. 3.5. Assessment of Previous Mode Assignments. Previous Raman studies have suggested various assignments for specific features in the 340-420 cm-1 region. Early studies suggested that some bands might be difference combinations70 or overtones or combinations and out-of-plane pyrrole tiltings.71 We showed28,47 that the area contributed to the NRVS signal by an overtone or a combination is smaller than the areas of the corresponding fundamentals by a factor comparable to (νj/νjR)/eFe2, with νjR ) 15.8 cm-1 for 57Fe. The absence of intense features due to vibrational fundamentals below 250 cm-1 clearly rules out the possibility that overtones and/or combinations contribute significantly to the NRVS signal in the 320-420 cm-1 region. We conclude that all frequencies in Table 2 correspond to vibrational fundamentals. The 307 and 358 cm-1 bands were previously assigned as degenerate pairs of modes (ν51 and ν50, respectively),62 involving asymmetric stretching of the Fe-Npyr bonds along two orthogonal in-plane directions. To test this idea, we double the area of the 307 cm-1 peak. The new prediction matches the Fe VDOS better (Figure 4, dash-and-dot line in the upper panel). Therefore, our results are consistent with the assignment of the 307 cm-1 band to a degenerate pair of modes, possibly with a ν51 character. By doubling the predicted area of the 358 cm-1 peak, agreement with the NRVS signal is not significantly improved or worsened. This analysis is not inconsistent with assignment as a degenerate mode pair with ν50 character.62 However, we cannot exclude assignment to a single mode, possibly involving stretching of the axial Fe-N and Fe-S bonds.72 The relatively weak sensitivity of the 348 cm-1 mode to Fe isotope substitution (eFe2 ) 0.06, Table 2) is consistent with its assignment to the ν8 vibration,62 which involves symmetric stretching of the four Fe-Npyr bonds to the heme and thus leaves the Fe atom motionless in 4-fold symmetric molecules.73 This disagrees with the cyt f study mentioned above,49 which concluded that involvement of Fe motion in this mode makes the ν8 label meaningless. Although there may be a small amount of mixing with out-of-plane vibrations, the Fe amplitude remains small, and we conclude that the ν8 label remains meaningful for cyt c. The 392, 401, 414, and 422 cm-1 bands are sensitive to N, Ca, and Cb isotope substitutions (Figure 1) and were assigned as (Cβ-Ca-Cb) and (Cβ-Ca-S) bending modes of the thioether groups.62 Because of the complex character of these modes (involving motions of both the pyrrole nitrogens and the peripheral groups), it is not unexpected that they involve Fe motion as well. Indeed, we find significant eFe2 values for these modes (Table 2). In the same work,62 the 372 and 382 cm-1 bands were assigned as propionate bending modes (δ(Cβ-Cc-Cd)). They
Leu et al. are nearly insensitive to N, Ca, and Cb isotope substitution, and the assignment was made by analogy with myoglobin, where modes at nearby frequencies were assigned on the basis of isotope shift upon propionate deuteration.74 The 382 cm-1 band involves a relatively small iron motion, not inconsistent with substantial contributions from propionate vibrations. However, we find that the 372 cm-1 band displays the biggest shift upon 57 Fe f 54Fe substitution and, consequently, the largest mode composition factor eFe2 (Table 2). This frequency corresponds to the maximum NRVS signal, thus making substantial contribution from propionate vibrations unlikely. 3.6. Identification of Structural Markers. The 320-420 cm-1 region of the Soret-enhanced Raman spectrum of cyt c is significantly richer than the corresponding region of proteins such as myoglobin, which lack the thioether links to the protein. This region is highly sensitive to protein conformation and has been a useful fingerprint for identifying structural changes of the heme environment upon mutation,21 binding to ions71 or lipid membranes,22,75 or unfolding.22,23 In particular, changes in axial ligation lead to spectral perturbations in this region.22,24,71,76,77 However, there are no generally accepted assignments of observed Raman frequencies to vibrations of individual Feligand bonds in cyt c. Although one might consider the possibility that such structurally sensitive modes are not enhanced in resonance with the Soret transition of the heme, the results presented here suggest that this is not the case. Observed Raman modes account for a large fraction of the NRVS signal above 300 cm-1, at least for Fe(II) cyt c, and these vibrations make the primary contribution to the stiffness of the Fe environment in both oxidation states (Figure 2, Table 1). One specific possibility that should be considered in future investigations is that the 372 cm-1 mode in Fe(II) cyt c corresponds to the Fe-S stretching vibration, which may shift down to ∼340 cm-1 in Fe(III) cyt c and no longer contribute to the Raman signal excited in resonance with the Soret transition. The frequency shift would be consistent with the modest increase in Fe-S bond length reported for the oxidized protein17 (Table 1) and would contribute to the reduced stiffness of the oxidized protein. This frequency range is consistent with previous observations of Fe(III)-S stretching vibrations near 350 cm-1 in cyt P45065 and chloroperoxidase.66 On the basis of a 1.5 cm-1 shift upon 56Fe f 54Fe isotope substitution, the Fe-S band was assigned at 328 cm-1 in [Fe(TPP)(S(CH3)2]ClO4.78 The absence of a significant frequency shift upon deuteration of the methionine methyl group,62 together with the large eFe2 (Table 2), might indicate that the 372 cm-1 vibration is relatively localized on the iron and sulfur atoms. The total eFe2 ) 0.47 for this mode exceeds the value eFe2 ) mS/(mFe + mS) ) 0.36 expected for a two-body Fe-S oscillator. Significant vibrational alterations have been observed in this frequency range upon replacement of the methionine sulfur ligand with the nitrogen atom of the protein amino terminus in cyt f.49,79 The maximum of the Fe VDOS decreases considerably upon oxidation (Figure 2), consistent with the longer Fe-S(Met80) bond in oxidized cyt c (Table 1). We suggest that the 372 cm-1 band is the Fe-S(Met80) stretching mode in reduced cyt c. This band does not have an obvious correspondent in the Raman spectrum of oxidized cyt c. The doubling of the area of the 358 cm-1 peak has a minimal effect on the predicted NRVS signal, which may indicate that it has an out-of-plane character. This mode was previously assigned as a stretching of the Fe-N and Fe-S bonds.72 Its mode composition factor eFe2 ) 0.14 lies within a range of
Vibrational Dynamics of Iron in Cytochrome c values 0.11-0.18 calculated for a two-body oscillator, assuming various distributions of Fe, Met, and His masses. Given the relatively large value of the iron mode composition factor, a NRVS measurement on an 15N(His) and/or 34S(Met) isotopically labeled cyt c might clarify the assignment of this mode. Vibrations of the Fe-N bond to the histidine also must contribute to the NRVS signal. Although Fe-histidine vibrations are not observed to contribute significantly to the Raman signal for six-coordinated heme proteins, five-coordinated heme proteins exhibit strongly enhanced Fe-His vibrations in the 200-250 cm-1 region.63,80 As a result, we expect that the Fe-His vibration may contribute to one or more of the features observed in the 100-300 cm-1 range of the VDOS (Figure 2). A band between 176 and 183 cm-1 in the resonance Raman spectra of several mitochondrial cyt c was assigned as a ν(Fe-His) mode coupled to a ν(Fe-S(Met)) mode.81,82 However, we acknowledge the possibility that vibrations may resist simple description in terms of motion localized to individual bonds. In a thorough study of low-spin iron porphyrins with imidazole and CO ligands, we identified several vibrations involving both Fe and imidazole.36 Hence, it may not always be possible to identify a single feature with vibration of each Fe-ligand bond. Naturally occurring porphyrins have a rich vibrational structure, and comparison with quantum chemical calculations may provide valuable guidance for establishing the sensitivity of vibrational features to the strength of particular Fe-ligand bonds.36,46,47,83-86 4. Conclusions As for other 57Fe-containing proteins,29,33,34,49,87,88,54 NRVS provides a site-selective view of active site vibrations in cyt c that complements information available from well-established vibrational techniques such as resonance Raman spectroscopy. NRVS reveals the complete VDOS of the Fe atom, allowing evaluation of effective force constants. The stiffness measures the force required to displace the Fe relative to its nearest neighbors and decreases slightly upon oxidation of cyt c (Table 1).48 This contrasts with the resilience, which is related to the mean-squared displacement of the Fe atom in response to thermal fluctuations.37 The resilience is relatively insensitive to the oxidation state of cyt c, and the decreased resilience observed for myoglobin may reflect the absence of the thioether bonds.37 In contrast, the stiffness of cyt c decreases slightly upon oxidation, reflecting the weakened Fe-S bond. Comparison with Raman measurements on cyt c isotopically labeled at the Fe enables a more detailed description of the vibrational structure. Previous NRVS studies of heme proteins have revealed Fe-ligand modes not reported in resonance Raman spectra, such as the Fe-Npyr bonds to the heme pyrrole ligands.33 For Fe(II) cyt c, however, we find that vibrations observed in the Raman spectrum account for a significant fraction of the NRVS signal above 300 cm-1. This indicates that both axial and equatorial Fe ligand vibrations contribute to the Raman signal. We suggest that observation of many of these modes is forbidden by selection rules in proteins with more symmetric heme groups. Both the asymmetrically placed thioether links and the resulting heme nonplanarity may contribute to allow these modes to couple to the electronic Soret transitions of the heme in cyt c. The rich vibrational dynamics of Fe reported here for cyt c form an intriguing comparison with earlier measurements on Mb. For deoxyMb, 57Fe NRVS data resolved three primary features that we associated with stretching of axial and equatorial Fe-N bonds,33 and comparison with DFT predictions supports
J. Phys. Chem. B, Vol. 113, No. 7, 2009 2199 a similarly simple interpretation of high-frequency Fe vibrations in ferryl Mb.35 In contrast, more than 10 vibrational modes contribute to determining the stiffness of the Fe environment in Fe(II) cyt c, although there are only three chemically distinct bond types to the heme Fe. Further investigation should enable the use of this rich vibrational information to monitor conformational changes of cyt c upon complex formation with phospholipid membranes, as proposed to take place within mitochondria75 or during cellular apoptosis.6 Acknowledgment. We thank Prof. Abel Schejter for useful discussions and acknowledge generous support of this research by the National Science Foundation (PHY-0545787). Use of the Advanced Photon Source was supported by the U.S. Department of Energy, Basic Energy Sciences, Office of Science, under Contract DE-AC02-06CH11357. References and Notes (1) Berg, J.; Tymoczko, J.; Stryer, L. Biochemistry, 5th ed.; Freeman and Co.: New York, 2002. (2) Scott, R. A.; A. G. Mauk, E. Cytochrome c: A Multidisciplinary Approach; University Science Books: Sausalito, 1996. (3) Moore, G. R.; Pettigrew, G. W. Cytochromes c: Biological Aspects; Springer-Verlag: New York, 1987. (4) Moore, G. R.; Pettigrew, G. W. Cytochromes c: EVolutionary, Structural, and Physicochemical Aspects; Springer-Verlag: New York, 1990. (5) Jiang, X.; Wang, X. Annu. ReV. Biochem. 2004, 73, 87–106. (6) Bayir, H.; Fadeel, B.; Palladino, M. J.; Witasp, E.; Kurnikov, I. V.; Tyurina, Y. Y.; Tyurin, V. A.; Amoscato, A. A.; Jiang, J.; Kochanek, P. M.; DeKosky, S. T.; Greenberger, J. S.; Shvedova, A. A.; Kagan, V. E. Biochim. Biophys. Acta 2006, 1757, 648–659. (7) Banci, L.; Bertini, I.; Gray, H. B.; Luchinat, C.; Reddig, T.; Rosato, A.; Turano, P. Biochemistry 1997, 36, 9867–9877. (8) Banci, L.; Bertini, I.; Huber, J. G.; Spyroulias, G. A.; Turano, P. J. Biol. Inorg. Chem. 1999, 4, 21–31. (9) Feng, Y.; Roder, H.; Englander, S. W. Biochemistry 1990, 29, 3494– 3504. (10) Gao, Y.; Boyd, J.; Pielak, G. J.; Williams, R. J. P. Biochemistry 1991, 30, 1928–1934. (11) Qi, P. X.; Beckman, R. A.; Wand, A. J. Biochemistry 1996, 35, 12275–12286. (12) Louie, G. V.; Brayer, G. D. J. Mol. Biol. 1990, 214, 527–555. (13) Berghuis, A. M.; Brayer, G. D. J. Mol. Biol. 1992, 223, 959–976. (14) Takano, T.; Dickerson, R. E. J. Mol. Biol. 1981, 153, 95–115. (15) Takano, T.; Dickerson, R. E. J. Mol. Biol. 1981, 153, 79–94. (16) Sanishvili, R.; Volz, K.; Westbrook, E.; Margoliash, E. Structure 1995, 3, 707–16. (17) Cheng, M.-C.; Rich, A. M.; Armstrong, R. S.; Ellis, P. J.; Lay, P. A. Inorg. Chem. 1999, 38, 5703–5708. (18) Trewhella, J.; Carlson, V. A. P.; Curtis, E. H.; Heidorn, D. B. Biochemistry 1988, 27, 1121–1125. (19) Tiede, D. M.; Zhang, R.; Seifert, S. Biochemistry 2002, 41, 6605– 6614. (20) Schejter, A. Cytochrome c: A Multidisciplinary Approach; University Science Books: Sausalito, CA, 1996; Chapter 8, p 335. (21) Hildebrandt, P.; Pielak, G. J.; Williams, R. J. P. Eur. J. Biochem. 1991, 201, 211–216. (22) Droghetti, E.; Oellerich, S.; Hildebrandt, P.; Smulevich, G. Biophys. J. 2006, 91, 3022–3031. (23) Takahashi, S.; Yeh, S.-R.; Das, T. K.; Chan, C.-K.; Gottfried, D. S.; Rousseau, D. L. Nat. Struct. Biol. 1997, 4, 44–50. (24) Yeh, S.-R.; Rousseau, D. L. J. Biol. Chem. 1999, 274, 17853– 17859. (25) Sturhahn, W.; Toellner, T. S.; Alp, E. E.; Zhang, X.; Ando, M.; Yoda, Y.; Kikuta, S.; Seto, M.; Kimball, C. W.; Dabrowski, B. Phys. ReV. Lett. 1995, 74, 3832–3835. (26) Zeng, W.; Silvernail, N. J.; Scheidt, W. R.; Sage, J. T. In Applications of Physical Methods to Inorganic and Bioinorganic Chemistry; Scott, R. A., Lukehart, C. M., Eds.; John Wiley & Sons, Ltd.: Chichester, UK, 2007; pp 401-421. (27) Sturhahn, W. J. Phys.: Condens. Matter 2004, 16, S497–S530. (28) Sage, J. T.; Paxson, C.; Wyllie, G. R. A.; Sturhahn, W.; Durbin, S. M.; Champion, P. M.; Alp, E. E.; Scheidt, W. R. J. Phys.: Condens. Matter 2001, 13, 7707–7722. (29) Bergmann, U.; Sturhahn, W.; Linn, D. E.; Jenney, F. E.; Adams, M. W.; Rupnik, K.; Hales, B. J.; Alp, E. E.; Mayse, A.; Cramer, S. P. J. Am. Chem. Soc. 2003, 125, 4016–4017.
2200 J. Phys. Chem. B, Vol. 113, No. 7, 2009 (30) Chumakov, A. I.; Sergueev, I.; van Bu¨rck, U.; Schirmacher, W.; Asthaller, T.; Ru¨ffer, R.; Leupold, O.; Petry, W. Phys. ReV. Lett. 2004, 92, 245508. (31) Tse, J.; Klug, D.; Zhao, J.; Sturhahn, W.; Alp, E.; Baumert, J.; Gutt, C.; Johnson, M.; Press, W. Nat. Mater. 2005, 4, 917–921. (32) Giefers, H.; Tanis, E. A.; Rudin, S. P.; Greeff, C.; Ke, X.; Chen, C. F.; Nicol, M. F.; Pravica, M.; Pravica, W.; Zhao, J.; Alatas, A.; Lerche, M.; Sturhahn, W.; Alp, E. Phys. ReV. Lett. 2007, 98, 245502. (33) Sage, J. T.; Durbin, S. M.; Sturhahn, W.; Wharton, D. C.; Champion, P. M.; Hession, P.; Sutter, J.; Alp, E. E. Phys. ReV. Lett. 2001, 86, 4966–4969. (34) Zeng, W.; Silvernail, N. J.; Wharton, D. C.; Georgiev, G. Y.; Leu, B. M.; Scheidt, W. R.; Sturhahn, W.; Alp, E. E.; Sage, J. T. J. Am. Chem. Soc. 2005, 127, 11200–11201. (35) Zeng, W.; Barabanschikov, A.; Zhang, Y.; Zhao, J.; Sturhahn, W.; Alp, E. E.; Sage, J. T. J. Am. Chem. Soc. 2008, 130, 1816–1817. (36) Leu, B. M.; Silvernail, N. J.; Zgierski, M. Z.; Wyllie, G. R. A.; Ellison, M. K.; Scheidt, W. R.; Zhao, J.; Sturhahn, W.; Alp, E. E.; Sage, J. T. Biophys. J. 2007, 92, 3764–3783. (37) Leu, B. M.; Zhang, Y.; Bu, L.; Straub, J. E.; Zhao, J.; Sturhahn, W.; Alp, E. E.; Sage, J. T. Biophys. J. 2008, 95, 5874–5889. (38) Morell, D. B.; Stewart, M. Aust. J. Exp. Biol. Med. Sci. 1956, 34, 211–217. (39) Vanderkooi, J. M.; Adar, F.; Erecinska, M. Eur. J. Biochem. 1976, 64, 381–387. (40) Dickinson, L. C.; Chien, J. C. W. Biochemistry 1975, 14, 3526– 3534. (41) Vanderkooi, J. M.; Landesburg, R.; Hayden, G. W.; Owen, C. S. Eur. J. Biochem. 1977, 81, 339–347. (42) Leu, B. M. Nuclear resonance vibrational spectroscopy: A quantitative picture of iron dynamics in heme proteins and model compounds. PhD Thesis, Northeastern University, 2006. (43) Schomacker, K. T.; Champion, P. M. J. Chem. Phys. 1989, 90, 5982–5993. (44) Seto, M.; Yoda, Y.; Kikuta, S.; Zhang, X. W.; Ando, M. Phys. ReV. Lett. 1995, 74, 3828–3831. (45) Sturhahn, W. Hyperfine Interact. 2000, 125, 149–172. (46) Leu, B. M.; Zgierski, M. Z.; Wyllie, G. R. A.; Scheidt, W. R.; Sturhahn, W.; Alp, E. E.; Durbin, S. M.; Sage, J. T. J. Am. Chem. Soc. 2004, 126, 4211–4227. (47) Leu, B. M.; Zgierski, M. Z.; Wyllie, G. R. A.; Ellison, M. K.; Scheidt, W. R.; Sturhahn, W.; Alp, E. E.; Durbin, S. M.; Sage, J. T. J. Phys. Chem. Solids 2005, 66, 2250–2256. (48) Lipkin, H. J. Phys. ReV. B 1995, 52, 10073–10079. (49) Adams, K. L.; Tsoi, S.; Yan, J.; Durbin, S. M.; Ramdas, A. K.; Cramer, W. A.; Sturhahn, W.; Alp, E. E.; Schulz, C. J. Phys. Chem. B 2006, 110, 530–536. (50) Badger, R. M. J. Chem. Phys. 1935, 3, 710–714. (51) Dyer, R. B.; Woodruff, W. H. In Applications of Physical Methods to Inorganic and Bioinorganic Chemistry; Scott, R. A., Lukehart, C. M., Eds.; John Wiley & Sons, Ltd.: Chichester, UK, 2007; pp 001-024. (52) Green, M. T. J. Am. Chem. Soc. 2006, 128, 1902–1906. (53) Xu, C.; Spiro, T. G. J. Biol. Inorg. Chem. 2008, 13, 613–621. (54) Xiao, Y.; Wang, H.; George, S. J.; Smith, M. C.; Adams, M. W. W.; Jenney, F. E.; Sturhahn, W.; Alp, E. E.; Zhao, J.; Yoda, Y.; Dey, A.; Solomon, E. I.; Cramer, S. P. J. Am. Chem. Soc. 2005, 127, 14596–14606. (55) Chance, M. R.; Miller, L. M.; Fischetti, R. F.; Scheuring, E.; Huang, W.-X.; Sclavi, B.; Hai, Y.; Sullivan, M. Biochemistry 1996, 35, 9014– 9023. (56) Hsu, I.-J.; Shiu, Y.-J.; Jeng, U.-S.; Chen, T.-H.; Huang, Y.-S.; Lai, Y.-H.; Tsai, L.-N.; Jang, L.-Y.; Lee, J.-F.; Lin, L.-J.; Lin, S.-H.; Wang, Y. J. Phys. Chem. A 2007, 111, 9286–9290. (57) Giachini, L.; Francia, F.; Cordone, L.; Boscherini, F.; Venturoli, G. Biophys. J. 2007, 92, 1350–1360. (58) Brehm, G.; Reiher, M.; Schneider, S. J. Phys. Chem. A 2002, 106, 12024–12034. (59) Ronayne, K. L.; Paulsen, H.; Hofer, A.; Dennis, A. C.; Wolny, J. A.; Chumakov, A. I.; Schunemann, V.; Winkler, H.; Spiering, H.;
Leu et al. Bousseksou, A.; Gutlich, P.; Trautwein, A. X.; McGarvey, J. J. Phys. Chem. Chem. Phys. 2008, 8, 4685–4693. (60) Hildebrandt, P. Cytochrome c: A Multidisciplinary Approach; University Science Books: Sausalito, CA, 1996; pp 285-314. (61) Schomacker, K. Absorption, resonant Rayleigh and resonance Raman properties of cytochrome c. PhD Dissertation, Department of Physics, Northeastern University, 1987. (62) Hu, S.; Morris, I. K.; Singh, J. P.; Smith, K. M.; Spiro, T. G. J. Am. Chem. Soc. 1993, 115, 12446–12458. (63) Argade, P. V.; Sassaroli, M.; Rousseau, D. L.; Inubushi, T.; IkedaSaito, M.; Lapidot, A. J. Am. Chem. Soc. 1984, 106, 6593–6596. (64) Hirota, S.; Mizoguchi, Y.; Yamauchi, O.; Kitagawa, T. J. Biol. Inorg. Chem. 2002, 7, 217–221. (65) Champion, P. M.; Stallard, B.; Wagner, G. C.; Gunsalus, I. C. J. Am. Chem. Soc. 1982, 104, 5469–5472. (66) Bangcharoenpaurpong, O.; Champion, P. M.; Hall, K.; Hager, L. P. Biochemistry 1986, 25, 2374–2378. (67) Maes, E. M.; Walker, F. A.; Montfort, W. R.; Czernuszewicz, R. S. J. Am. Chem. Soc. 2001, 123, 11664–11672. (68) Heering, A.; Smith, A. T.; Smulevich, G. Biochem. J. 2002, 363, 571–579. (69) Gruia, F.; Kubo, M.; Ye, X.; Ionascu, D.; Lu, C.; Poole, R. K.; Yeh, S.-R.; Champion, P. M. J. Am. Chem. Soc. 2008, 130, 5231–5244. (70) Friedman, J. M.; Hochstrasser, R. M. Chem. Phys. 1973, 1, 457– 467. (71) Hildebrandt, P. Biochim. Biophys. Acta 1990, 1040, 175–186. (72) Hon-Nami, K.; Kihara, H.; Kitagawa, T.; Miyazawa, T.; Oshima, T. Eur. J. Biochem. 1980, 110, 217–223. (73) Abe, M.; Kitagawa, T.; Kyogoku, Y. J. Chem. Phys. 1978, 69, 4526–4534. (74) Hu, S.; Smith, K. M.; Spiro, T. G. J. Am. Chem. Soc. 1996, 118, 12638–12646. (75) Berezhna, S.; Wohlrab, H.; Champion, P. M. Biochemistry 2003, 42, 6149–6158. (76) Dopner, S.; Hildebrandt, P.; Rosell, F. I.; Mauk, A. G. J. Am. Chem. Soc. 1998, 120, 11246–11255. (77) Hirota, S.; Suzuki, M.; Watanabe, Y. Biochem. Biophys. Res. Commun. 2004, 314, 452–458. (78) Oshio, H.; Ama, T.; Watanabe, T.; Nakamoto, K. Inorg. Chim. Acta 1985, 96, 61–66. (79) Desbois, A. Biochimie 1994, 76, 693–707. (80) Wells, A. V.; Sage, J. T.; Morikis, D.; Champion, P. M.; Chiu, M. L.; Sligar, S. G. J. Am. Chem. Soc. 1991, 113, 9655–9660. (81) Cartling, B. Biophys. J. 1983, 43, 191–205. (82) Desbois, A.; Lutz, M. Proceedings of the XIth International Conference on Raman Spectroscopy, Tokyo; The Chemical Society of Japan: Tokyo, 1984; pp 480-481. (83) Vogel, K. M.; Kozlowski, P. M.; Zgierski, M. Z.; Spiro, T. G. J. Am. Chem. Soc. 1999, 121, 9915–9921. (84) Kozlowski, P. M.; Spiro, T. G.; Zgierski, M. Z. J. Phys. Chem. B 2000, 104, 10659–10666. (85) Rush, T. S., III.; Kozlowski, P. M.; Piffat, C. A.; Kumble, R.; Zgierski, M. Z.; Spiro, T. G. J. Phys. Chem. B 2000, 104, 5020–5034. (86) Paulat, F.; Praneeth, V. K. K.; Nather, C.; Lehnert, N. Inorg. Chem. 2006, 45, 2835–2856. (87) Keppler, C.; Achterhold, K.; Ostermann, A.; van Bu¨rck, U.; Potzel, W.; Chumakov, A. I.; Baron, A. Q.; Ru¨ffer, R.; Parak, F. Eur. Biophys. J. 1997, 25, 221–224. (88) Achterhold, K.; Keppler, C.; Ostermann, A.; van Bu¨rck, U.; Sturhahn, W.; Alp, E. E.; Parak, F. G. Phys. ReV. E 2002, 65, 051916. (89) Berman, H. M.; Westbrook, J.; Feng, Z.; Gilliland, G.; Bhat, T. N.; Weissig, H.; Shindyalov, I. N.; Bourne, P. E. Nucleic Acids Res. 2000, 28, 235–242. (90) Bushnell, G. W.; Louie, G. V.; Brayer, G. D. J. Mol. Biol. 1990, 214, 585–595. (91) Sayle, R.; Milner-White, E. J. Trends Biochem. Sci. 1995, 20, 374. (92) Portman, S.; Lu¨thi, H. P. Chimia 2000, 54, 766–770.
JP806574T