Electron transfer in ruthenium-modified cytochromes c. .sigma

Proton-Coupled Electron Flow in Protein Redox Machines. Jillian L. Dempsey , Jay R. Winkler , and Harry B. .... Theory and Practice of Electron Transf...
0 downloads 9 Views 616KB Size
Download PDF
J. Phys. Chem. 1993,97, 13073-13077

13073

Electron Transfer in Ruthenium-Modified Cytochromes c. a-Tunneling Pathways through Aromatic Residuest Dado R. Casimiro, John H. Richards,' Jay R. Winkler,' and Harry B. Gray' Beckman Institute, California Institute of Technology, Pasadena, California 91 I25 Received: June 3, 1993; In Final Form: July 22, 1993'

The rates of intramolecular electron-transfer (ET) reactions from the ferrohemeto bis(2,2'-bipyridine)(imidazole)ruthenium(II1) complexes bound to genetically engineered histidines ( 5 8 and 66) on the surface of yeast iso1-cytochrome c (cyt c ) have been measured by using a laser flashquench technique. The crystal structure of the wild-type protein indicates that the E T pathways involve aromatic side chains: Ru(His58)cyt c includes a bridging tryptophan at position 59, and Ru(His66)cyt c has a tyrosine at 67. A variant in which the bridging Tyr67 in the His66 mutant had been replaced with a phenylalanine also was examined. The Fez+ Ru3+ET rate constants (25 O C , pH 7.0) are as follows: 5.2(5) X lo4 ( A E O = 0.69(5)), Ru(His58)cyt c; 1.0(1) X 106 (AEO = 0.72(5)), Ru(His66)cyt c; and 3.1(3) X lo6 s-I (&To = 0.77(5) eV), Ru(His66Phe67)cyt c. The experimentally derived electronic coupling constants [H~e(His%)= 0.014; H~a(HiS66)= 0.060 cm-'1 are in closer agreement with the lengths of a-tunneling pathways than with the direct donor-acceptor distances, and there is no indication that the u orbitals of intervening groups enhance any of these couplings. Maximum ET rates in the modified cytochromes drop by 2 orders of magnitude for every 6.3-A increase in the a-tunneling length. Analysis of the results also suggests that an internal water molecule in Ru(His66Phe67)cyt c plays a role in linking the Ru(His66) group to the heme.

-

Introduction Studies of electron-transfer (ET) reactions in metalloproteins have suggestedthat thedistant donor-acceptor electronic coupling is a functionof the structureof the bridging polypeptide medium.lJ One question of particular interest is whether the u orbitals of aromatic groups in the intervening medium influence these electronic couplings to a significant extent.- To examine this matter directly, we have studied site-directed mutants of Saccharomycescereuisiae iso- 1cytochrome c (cyt c ) in which surface histidine residues for ruthenium labeling have been introduced at positions 58 and 66.l A pathway searching algorithms and the three-dimensional structure of the wild-type protein9 show that a tryptophan (at position 59) or a tyrosine (at position 67) is along the ET coupling route to the heme. In addition, a phenylalanine side chain can be substituted for Tyr67, thereby changing the ET path in a well-defined manner.

Experimental Section Materials. (His58)-, (His66)-, and (His66Phe67)cyt c were isolated from recombinant yeast cells and purified to homogeneity.1° Distilled water passed through a Barnstead Nanopure purification system (specific resistance > 18 MS2-cm-I) was used to prepare all aqueous solutions. Sodium phosphate (NaPi) solutions were prepared from analytical grade reagents. 4,4'Dipyridyl disulfide was obtained from Aldrich. Hexaammineruthenium trichloride (Aldrich) was recrystallized from 1 M HC1 before use. Electrochemistry. Direct electrochemistry of 0.4-0.6 mM protein solutions in p 100 mM NaPi, pH 7.0, was performed by using an 8-mm gold disk electrode (Pine Instruments) modified with 4,4'-dipyridyl disulfide." Before use, the electrode was dipped in a saturated solution of the promoter for 2 min and washed extensively with water. Cyclic voltammetry was carried out at 25 OC with a Ag/AgCl/KCl(saturated) reference electrode (Ingold Products) and a platinumcounter electrode. The potential was controlled by a Princeton Applied Research Model 173 Dedicated to the memory of Gerhard Closs. Abtract published in Advance ACS Abstracts, November 1, 1993.

0022-3654/93/2097-13073$04.00/0

potentiostat and the current output was recorded on a Houston Instruments Model 2000 X-Y recorder. A potential range of 50-400 mV vs NHE was swept at 5 mV/s. RutheniumModification. Each protein was modified with bis(2,2'-bipyridine) (imidazole)ruthenium(11)(Ru(bpy)z(im)Z+)by standard proced~res.~*J3 All preparations and manipulations of the modified proteins were conducted with the exclusion of room light. Ru(bpy)z(CO~)was added in 10-fold excess to 0.2 mM protein solutions in p 100 mM NaPi, pH 7.0, and the reactions were carried out under argon. Subsequent reactions with imidazole involved passing the protein through Sephadex G-25 preequilibratedwith 0.5 M imidazole, pH 7.8, eluting the proteins with the same imidazole buffer, and leaving the protein solutions under argon for 36 h at 25 OC. Excess imidazole was removed by gel-filtrationchromatography.The samples were then loaded onto a Mono-S HR 16/10 column and the proteins were eluted with a slow &1 M NaCl gradient in p 50 mM NaPi, pH 7.0. A 1:l ruthenium/protein adduct was indicated by a 2:3 ratio of the 290 (largely bipyridine u u*)to 410-nm (Soret) absorbance.'Z Fractions containing the 1:l Ru/Fe products were collected, pooled, and run repeatedly by using the same ion-exchange separation procedure until the bands were completely resolved. Cyclic voltammetric measurements were made on 0.2-0.4 mM solutions of the modified proteins following the same protocol as that used for unmodified proteins. Spectroscopic Characterization. Solutions in the concentration range 0.10-15 mM of the unmodified and modified proteins in p 100 mM NaPi, pH 7.0, were transferred into 0.010-cm cells for circular dichroism measurements in the far-UV region ( 194250 nm). An averageof 10 scans was taken for each spectrum. Protein solutions, 0.2 mM, in a 0.10-cm cell were employed to obtain the near-UV CD spectra (250-350 nm). Flash-Quench I(ineti~s.~3 The Ru-modified proteins were reduced with excess dithionite and desalted through Sephadex G-25. Solutions, 5-20 pM, of reduced proteins in p 100 mM NaPi, pH 7.0,7 mM hexaammineruthenium(III),were prepared, deoxygenated,and equilibrated under argon in vacuum cells with 1-cm fluorescence cuvette side arms. The samples were excited with 2-mJ, 204s pulses (480 nm) generated by a XeCl excimerpumped dye laser (Lambda Physik LPX 210i, FL-3002), and

-

Q 1993 American Chemical Society

Casimiro et al.

13074 The Journal of Physical Chemistry, Vol. 97, No. 50, 1993

-0.41

I

240

I

I

260

I

280

I

I

300

I

I

320

I

I

340

I

I

380

1

Wavelength, nm Figure 1. Difference circular dichroism specta between unmodified and modified cytochromes in the near-UV region. The spectra reflect the stereochemistries of bound ruthenium labels on (His66)cyt c (- -), (His58)cyt c (- -), and (His55Phe67)cyt c (-). These Ru-modified proteins were prepared by using racemic Ru(bpy)2(CO3).

-

-

transient absorption was measured at 306, 504, and 550 nm. Kinetic traces represent an average of 1000 laser shots. UV/vis spectra of the samples were taken before and after the kinetics measurements to test for protein degradation. Molecular Modeling. Three-dimensional models of the modified cytochromes were constructed from the 1.23-A crystal structure of reduced iso- 1-cytochrome c,9a using BIOGRAF, version 2.2 (Molecular Simulations, Inc.), implemented on a VaxStation3500/Evans and Sutherland PS390 system. The native residue at 58 or 66 was replaced by a R~(bpy)2(im)(His)~+ group built from crystallographic data on R~(bpy)zC12.'~Minimized structures of the modified histidines were obtained after searching the conformational spaces generated by rotating the C,-Cp and C& , bonds. Only those van der Waals interactions within 9 A of the ruthenium-labeled histidine were included in the calculations. Intramolecular distances were calculated from the minimized structures.ls

Results and Discussion Rutbenium Modification and Protein Characterization. We have constructed three mutants of yeast iso- 1-cytochrome c (His58, His66, and His66Phe67) and have shown that these variants are isofunctional and isostructural with the wild-type protein.10 Following established protocol~,~2,*3 the surface histidines in thesecytochromes were modified with R~(bpy)z(im)~+. Modificationdoes not appear to perturb the secondary structure of the protein (far-UV CD spectra; not shown), and the heme reduction potentials of the ruthenium derivatives (0.290(5), Ru(His58)-; 0.255(5), Ru(His66)-; and 0.215(5) V vs NHE, Ru(His55Phe67)cyt c) are similar to those of the corresponding unmodified proteins (0.272(5), (His58)-; 0.258(5), (His66)-; and 0.210(5) V, (His66Phe67)cyt ~16.17). Difference CD spectra between the modified and unmodified proteins in the near-UV region (250-350 nm) suggest a slight preference for the righthanded isomer of the Ru(bpy)z(im)(HisX)2+ complex (Figure 1).1* Kinetics. ET kinetics were measured by using a flash-quench procedure:13

-

0

2

1

Time, p s

Figure 2. Transient absorption decays monitored at 550 nm following laser flash excitation of mixtures of the Ru-modified protein (20 pM) and Ru(NH3)s3+(7 mM): (a) Ru(His58)cyt c; (b) Ru(His66)cyt c and Ru(His66Phe67)cytc. Theresultsof fitting the data tosingle-exponential decay functions (smooth lines) are 5.2(5) X l(r ( A E O = 0.69(5)), Ru(His58)cyt c; 1.0(1) X 106 (AEO = 0.72(5)), Ru(His66)cytc; and 3.1(3) X 106 s-l ( A E O = 0.77(5) eV), Ru(His66Phe67)cytc.

The Ru(bpy)~(im)(HisX)~+ complex is excited with a laser pulse (eq l), generating an excited state that, in the absence of ET quenching processes, decays with a lifetime of 70 ns. In the presence of a quencher (Q = Ru(NH3)s3+,5-10 mM), a fraction of the excited Ru complexes decays by a bimolecular ET route (eq 2), forming transient Ru3+-Fe2+*yt c in less than 100 ns. The intramolecular ET reaction (eq 3) is monitored by transient absorptionspectroscopy at wavelengths characteristicof the heme (550 nm) and Ru (306, 504 nm) oxidation states. The lowdriving-force bimolecular recombination between the reduced quencher and the ferriheme (eq 4) proceeds on a time scale of several seconds to regenerate the original reactants. Flash-quench transient-absorptionkinetics (550 nm, Figure 2) for Ru(His58)cyt c, Ru(His66)cyt c, and Ru(His66Phe67)cyt c are described by first-order decays and are independent of protein concentration (5-20 pM). Data recorded at 306 and 504 nm (not shown) demonstrate that Ru(bpy)2(im)(HisX)'+ is reduced at the same rate that the ferroheme is oxidized. Maximum Rates and Electronic Couplings. According to semiclassical ET theory,19 the rates of long-range nonadiabatic electron transfer depend upon reaction driving force (-AGO), the extent of nuclear reorientation accompanying electron transfer (A), and an electronic matrix element (HAB)coupling reactants and products at the transition state (eq 5).

hv

Ru2+-Fe2+*yt c *Ru2+-Fe2+-cyt c

+Q

Ru3+-Fe2+-cyt c Ru2+-Fe3+*yt c

*Ru2+-Fe2+*yt c

-

(1)

Ru3+-Fe2+*yt c + Q- (2) Ru2+-Fe3+*yt c

+ Q-- Ru2+-Fe2++yt c + Q

(3) (4)

At the point where the reaction driving force equals the reorganization energy (-AGO = A), the rate (k,& is limited by the electronic coupling, HAB.The Marcus cross relation predicts that A = 0.8(2) eV for the ET reactions of Ru(HisS8)cyt c and Ru(His66)cyt c.2 The,k and HABvalues (calculated by using eq 5) arecompared with thoseobtained for fourother Ru-modified cytochromes c in Table 1.20121

The Journal of Physical Chemistry, Vol. 97, No. 50, I993 13075

Electron Transfer in Modified Cytochromes c

TABLE I: ET Parameters for Ru(bpy)z(im)(HisX)cytochromes X

,,k s-1 3.3 x 106 2.7 X 106 66 1.1 x 106 72 9.4 x 105 58 6.0 X 104 62 1.0 x 104 a Reference 15. 39 33

c

HAB,cm-1 1.1 x 10-1 9.7 x 10-2

d (edge-to-edge),@ A

d (Ru-to-Fe),@ A

a1 (edge-to-edge),A

al (Ru-to-Fe), A

12.3 11.1

20.3 17.9

19.6 19.5

6.0 X 1k2 10-2

13.3 8.4

13.2 14.5

19.6 24.6

5.9 x 10-3

18.9 13.8 20.2

28.0 27.9 25.2 30.2 29.8 37.2

5.7 x 1.4 X

20.2

21.4 28.8

R uHis66

I IO

IS

10

IO

11

m

+

I

A d (edge-tc-edw), A Figure 3. ET rates as a function of donor-acceptor separation distance (referenceline, @ = 1.4 A-1): (a) log,k vs edge-to-edge distance and (b) log,k vs Ru-to-Fe distance for Ru(bpy)2(im)(HisX)cytc (X= 33, 39, 58, 62, 66, 72). d (Ru-twFe),

Long-range electron transfer in proteins can be described as

an electron-tunneling process: when nuclei reach the transitionstategeometry, thedonor electronmust tunnel through a potentialenergy barrier to reach the acceptor. A homogeneous medium (or vacuum) between redox sites imposes a tunneling barrier for which the electronic coupling should fall off expontentially with increasing donor-acceptor separation, d (eq 6).19 HA,

= H A B O exp{-p/2(d-

3 A))

In eq 6, HAB’is the electronic matrix element when the donor and acceptor groups are a t van der Waals contact (d = 3 A) and /3 is the distance decay constant. Support for a homogeneousbarrier model has come from Dutton’s analysis of several activationless protein ET rates in terms of an exponential decay function with /3 = 1.4 A-l and an adiabatic limit (kmaxat d = 3 A) of 10’3 S - I . ~ ~ The k,,, values that we have obtained for Ru(bpy)2(im)(His)-modified cytochromes, however, fall below Dutton’s line and are not well correlated with edge-to-edge separations (Figure 3a). It is striking, for example, that Ru(His66)cyt c and Ru(HisS8)cyt c have roughly the same edgeto-edge separation, yet the rates differ by 20-fold. It is possible that the discrepancies between Dutton’s analysis and the cytochrome c ET data arise from the use of the edgeto-edge distance scale. This distance scale implies that, when the edges of the two redox sites are in van der Waals contact, the ET reaction will be adiabatic (kmX= lOI3s-l). Electronic structure calculations, however, suggest that the electron self-exchange reactions of Ru(NH&3+/2+ are nonadiabatic because there is very little delocalization of donor and acceptor metal orbitals onto the ligands.23 If the Fez+ Ru3+ intramolecular ET reactions involve primarily metal-localized donor and acceptor states, then metal-to-metal separationswould be more appropriate in rate us distance correlations. Examination of Figure 3b, however, shows that a shift to a Ru-to-Fe distance scale does not improve matters; ,k values are systematically high and substantial scatter of the data points is still evident. Our conclusion is that the Fez+ Ru3+rates in modified cytochromes c cannot be accounted for by any homogeneous-barrier model; indeed, the

-

-

I

Figure 4. Proposed ET pathways (highlighted by dark lines) in Ru-

(His58)cyt c and Ru(His66)cyt c. The edge-to-edge pathway for Ru(His66)cytc consists of 11 a bonds and a 3.2-A hydrogen bond between Tyr67 and Met80. Theed e-to-edge pathway for Ru(His58)cytcconsists of 13 u bonds and a 3.0-1 hydrogen bond between Trp59 and a heme propionate group. Dashed lines denote hydrogen bonds and dotted areas representsolvent-accessiblesurfacesof the Ru(bpy)2(im)(His)2+ groups. available evidence shows that the distant electronic couplings depend strongly on the details of the structure of the intervening medium. a-TunnelingPathways. Several theories have been developed to address how the heterogeneous character of a bridging medium may influence the electronic coupling between a donor and an acceptor site.3,8,2k27 Beratan and Onuchic have proposed that electron tunneling between two redox sites in a protein is mediated by specific bonded and nonbonded interactions in the intervening polypeptide.8 This model gives HA, as a product of decay factors for each element in the path between redox sites. The problem is made tractable by assuming that, in proteins, there are just three types of pathway elements: u covalent bonds, hydrogen bonds, and through-space jumps. Using thesecriteria in a pathway searching algorithm,8f we can examine a protein structure for optimum electronic-coupling pathways. According to the model, the best ET pathway for Ru(His58)cyt c (Figure 4) consists of 13 u bonds and a hydrogen bond; the pathway is drawn from C, of His58 to the nearest edge of the porphyrin ring (C2A).28On the other hand, the pathway for ET in Ru(His66)cyt c (Figure 4) is two bonds shorter than that of Ru(HisS8)cyt c, which is consistent with the larger coupling constant for His66. u-tunneling lengths (ul) are calculated by multiplying the effective number of bonds in a pathway (defined as the (nonintegral) number of covalent bonds that give the same decay as the actual pathway) by an average bond length of 1.4 A. The current parameterization of the pathway model imposes a coupling-decay factor of 0.6 for each covalent bond.8 This factor defines 8’ = 0.13 A-I, where p’ is the exponential-decay constant for ,k as a function of uI; this value implies that the intramolecular ET rate drops by 2

Casimiro et al.

13076 The Journal of Physical Chemistry, Vol. 97, No. 50, 1993

\

I

\

I ’0

01 (edge-tpedge),

A

IO

Y

m XI 01 (Ru-to-Fe), A

I

m

Figure 5. ET rates in His66- and HisSI-modified mutants and other Ru-modified cytochromes as functions of edge-to-edge (a) and metalto-metal (b) dunneling lengths. Solid lines are constrained linear fits (8’ = 0.73 A-1) to the data.

orders of magnitude for every 6.3-A increase in the u-tunneling length.29 In a linear fit with constrained @’, we find that the activationless rate constants for Ru(His58)cyt c, Ru(His66)cyt c, and other modified cytochromescorrelate reasonably well with theeffectivelengthsof edge-to-edge u-tunneling pathways (Figure sa). Though the edge-to-edge u-tunnelingmodel predicts the weaker coupling in Ru(His58)cyt c than in Ru(His66)cyt c, the maximum ET rate for Ru(His58)cyt c is slower than expected for its calculated tunneling length (Figure sa). Part of the discrepancy again could be due to the edge-to-edgescale, where it is implicitly assumed that the coupling involves donor and acceptor orbitals that are delocalized over the ligands. It has been shown that HAB for two contacting metal complexes depends upon the square of the metal-ligand orbital mixing ~oefficient.~’If the metal-ligand mixings in Ru-modified cytochrome c are relatively small, then the u tunnels should extend to the metal atom redox centers rather than stopping at the ligands. A metal-to-metal pathway is in better accord with the coupling in Ru(His58)cyt c, since the pathway is extended by the u bonds of the porphyrin ring. Figure 5b presents a plot of log,k values versus the Ru-to-Fe tunneling lengths, Le., adding u bonds of the ligands and the metal-ligand bonds to the edge-to-edgepaths. A constrained-@’fit of log hx vs Ru-to-Fe ul shows only a marginally improved correlation over the edge-to-edge ul fit. The primary distinction between the two distance measures is found in the one-bond-limit ET rates: 1.7 X 1OI2 s-l for the edge-to-edge scale and 3.9 X 1014 s-1 for the Ru-to-Fe scale (Figure 5b).30 Regardless of the model used to analyze the data, our results show that the electroniccouplings for intramolecularET reactions in Ru(His58)- and Ru(His66)cyt care not enhanced by aromatic residues in the intervening media. Instead, the rates and the couplings can be described adequately by a model that invokes only u interactions.* The correlation of ET rates with u-tunneling lengths does not preclude a coupling role for the ?r orbitals of the aromatic groups in the pathway, but it does indicate that, in the Ru-modified cytochromes that we have examined, they are no more efficient in mediating the coupling than is the a-bonded framework.3’ Hence, the presence of aromatic groups in the medium between redox sites does not necessarily result in faster ET than in a purely aliphatic medium. The Fez+ Ru3+ ET rate to the surface Ru(His66) group exhibits a slight increase upon replacement of the intervening Tyr67 with a phenylalanine. This observation is suprising since the electronic coupling between Phe67 and Met80 would be expected to be weak, owing to the Phe67-Met80 space jump created in the u-tunnelingpathway. Preliminary crystallographic studies on the Phe67 mutant of iso-1-cytochrome c, however, show that a water molecule occupies the space vacated by the Tyr67 hydroxyl group.” Since the water molecule exhibits

-

relatively high thermal factors in both oxidation states of the protein, ET could be coupled to the motions of this water molecule and other nearby groups in the heme pocket.32 Many complex issues are involved in understanding electronic coupling in long-range electron transfer. The models presently available to address this problem require extreme simplifications. A homogeneous-barriermodel has been shown to describe a large number of ET systems22 but fails to account for many of our results. A model that explicitly treats the inhomogeneities of the intervening medium is clearly called for, and the Beratan-Onuchic a-tunneling pathway model is an important first step.8 We find a reasonable correlation of maximum ET rates with u-tunneling lengths, but there are some discrepancies. For example, the electroniccouplings between Ru(His58)- and Ru(His72)cyt cdo not correlate with the respective tunneling lengths (Figure 5a). Interestingly, extended-Huckel calculations on these modified cytochromesgive donor-acceptorcouplingsin better accord with the experimental values.33 While these systems were evaluated on the basis of a static crystal model, the dynamics of the residues in the ET pathways may contribute to the effective electronic couplings. In the case of Ru(His72)cyt c, this could manifest itself as shortening of the through-space gap in the proposed pathway that leads to a better donor-acceptorcoup1ing.h Recent theoretical efforts to describe electronic-coupling pathways in electron donor-acceptor systems also suggest important roles for interferences among competing pathways and stereoelectronic effects in the bridging medium and at the redox sites, as well as competition and crossovers between u and 7r systems.26.27It is likely that higher order effects of this type must be taken into account in order to explain fully the electronic couplings in Rumodified proteins. Acknowledgment. We thank Prabha Siddarth, Rudy Marcus, David Beratan, and Jos6 Onuchic for helpful discussions; Bill Durham for a preprint of reference IC;Albert Berghuis, Terence Lo, and Gary Brayer for information about the crystal structure of reduced yeast iso- 1-cytochromec and the Tyr67 Phevariant; and Michael Smith for the gift of the cytochrome c expression system. The assistance of I-Jy Chang, Bruce Bowler, Thomas Sutherland, John Racs, and Keith Herman in some of the experimental work is acknowledged. This work was supported by the National Science Foundation, the National Institutes of Health, and the Arnold and Mabel Beckman Foundation (this is contribution no. 8678 from the Beckman Institute).

-

References and Notes (1) (a) Winkler, J. R.; Gray, H. B. Chcm. Reu. 1992, 92, 369. (b) Casimiro, D. R.; Wong, L.-L.; Colh, J. L.; Zcwert, T. E.; Richards, J. H.; Chang, LJ.; Winkler, J. R.;Gray, H. B. J. Am. Chcm.Soc. 1993,115,1485. (c) Scott, J. R.; Willie, A.; McLean, M.; Stayton,P. S.;Sligar, S.G.; Durham, B.; Millett, F. J. Am. Chem. Soc. 1993, 115, 6820. (2) (a) Wuttke,D. S.;Bjerrum,M.J.; Winkler, J.R.;Gray,H.B.Scicnce 1992,256, 1007. (b) Wuttke, D. S.; Bjerrum, M. J.; Chang, I.-J.; Winkler, J. R.; Gray, H. B. Biochim. Biophys. Acra 1992, 1101, 168. (3) (a) Christensen,H. E. M.; Conrad, L. S.;Hammerstad-Pcdemn,J. M.;UIstrup,J.FEBSLerr. 1992,296,141. (b) Christensen,H.E.M.;Conrad, L. S.; Mikkelsen, K. V.; Nielscn, M. K.; Ulstrup. J. Inorg. Chcm. 1990, 29. 2808. (c) Christensen, H. E. M.; Conrad, L. S.;Mikkelacn, K. V.; Ulstrup, J. J . Phys. Chem. 1992,96,4451. (d) Broo, A,; Larsson, S. J. Phys. Chcm. 1991, 95,4925. (4) (a) Farver, 0.;Pecht, I. J. Am. Chcm. Soc. 1992, 114, 5764. (b) Farver, 0.; Skov, L. K.; Pascher, T.; Karlsson, B. G.; Nordling, M.; Lundberg, L. G.; Vinngard, T.; Pecht, I. Biochemistry 1993,32,7317. (c) Canters, G. W.; van de Kamp, M. Curr. Opin. Srrucr. Biol. 1992,2,859. (d) Farid, R. S.;Moser, C. C.; Dutton, P. L. Curr. Opin. Srrucr. Biol. 1993,3, 225. (e) Wuttke, D. S.; Gray, H. B. Curr. Opin. Srrucr. Biol. 1993, 3, 555. (5) (a) Sykes, A. G. In Metal Ions in Biological Sysfcms; Sigel, H., Sigel, A., Eds.; Marcel Dekker: New York, 1991; Vol. 27, pp 291-321. (b) Lloyd, E.;Tomkinson, N.P.; Sykes, A. G. J. Chem.Soc.,Dalron Trans. 1992, 753. (c)Govindaraju,K.; Chrintenacn,H.E. M.;Lloyd, E.;Olscn, M.;Salmon, G. A.; Tomkinson,N.P.; Sykes, A. 0 . Inorg. Chcm. 1993,32,40. ( 6 ) (a) Hazzard, J. T.;Mauk, A. G. Arch. Biochcm. Biophys. 1992,298, 91. (b) Everest, A. M.; Wallin, S. A.; Stemp, E. D. A.; Nocek, J. M.; Mauk, A. G.; Hoffman, B. M. J. Am. Chcm. Soc. 1991,113,4337. (c) Bowler, B.

Electron Transfer in Modified Cytochromes c E.; Meade, T. J.; Mayo, S.L.; Richards, J. H.; Gray, H. B. J. Am. Chem. Soc. 1989,111, 8751. (d) Davies, M.; Sligar, S . G.Biochemistry 1992,31,

11383. (7) The numbering is based on the tuna cyt c sequence (which lacks the first five N-terminal amino acids of yeast iso-1-cytochrome c). (8) (a) Onuchic, J. N.; Beratan, D. N.; Winkler, J. R.; Gray, H. B. Annu. Rev.Biophys. Biomol. Srrucr. 1992. 21, 349. (b) Beratan, D. N.; Betts, J. N.; Onuchic, J. N. Science 1991,252, 1285. (c) Betts, J. N.; Beratan, D. N.; Onuchic, J. N. J. Am. Chem. Soc. 1992,114,4043. (9). (a) Louie, G.V.; Brayer, G.D. J. Mol. Biol. 1990, 214, 527. (b) Berghuis, A. M.; Brayer, G.D. J. Mol. Biol. 1992, 223, 959. (10) Casimiro, D. R.; Richards, J. H.; Winkler, J. R.; Gray, H. B., in preparation. All mutants contain CyslO2 Ser (to prevent oxidative dimerization) and His39 Gln (to eliminate a reactive surface histidine). The amino acids at positions 58 and 66 in the wild-type protein are a leucine and a glutamic acid, respectively. (11) (a) Allen, P. M.; Hill, H. A. 0.;Walton, N. J. J. Electround. Chem. 1984,178,69. (b) Rafferty, S.P.; Pearce, L. L.; Barker, P. D.; Guillemente, J. G.;Kay, C. M.;Smith, M.; Mauk, A. G.Bfochemistry 1990, 29, 9365. (12) Durham, B. D.; Pan, L. P.; Hahm, S.; Long, J.; Millett, F. In ACS Advunces in Chemistry Series; Johnson, M. K., King, R. B., Kurtz, D. M., Kutal, C., Norton, M.L.,Scott, R. A., Eds.;American Chemical Society: Washington, DC, 1990; Vol. 226, pp 180-193. (13) Chang, L J . ; Gray, H. B.; Winkler, J. R. J . Am. Chem. Soc. 1991, 113,7056. (14) Eggleston, D. S.;Goldsby, K. A.; Hodgson, D. J.; Meyer,T. J. Inorg. Chem. 1985,24,4573. (1 5 ) A rigid-geometry search on the Ru-modified histidine was performed by successively rotating the two side-chain dihedral angles by 18O from 0 to 360° (Shih, H. H.-L.; Brady, J.; Karplus, M. Proc. Natl. Acud. Sei. U.S.A. 1985, 82, 1695). The 20 most stable conformers were examined for any unfavorable van der Waals (d < 3 A) contacts. Because interconversions among different conformers are expected to be extremely rapid (Brooks, C. L., 111; Karplus, M.; Pettitt, B. M. Adv. Chem. Phys. 1988, 71, 95). the shortest donor-acceptor distance are given for the modified proteins. Similar calculations were performed for other Ru-modified cytochromes2 using appropriate crystal structures: the 1 . 9 4 4 structureof horse heart cytochrome c (Bushnell, G.W.; Louie, G. V.; Brayer, G.D. J . Mol. Biol. 1990,214,585) forRu(His33)- and Ru(His72)cytcand the yeasti~l-cytochromecstructureA for Ru(His62)- and Ru(His39)cyt c. (16) The decrease in heme potential upon introduction of Phe67 into the His66 variant is comparable to that observed when the same change is made in the CyslO2 Thr pseudo-wild-type protein (A& 2 = -55 mV).” The crystal structure of the Tyr67 Phe, Cys102 khr mutant of iso-lcytochrome c reveals that the polypeptide conformation is virtually identical with that of the CyslO2 Thr pseudo-wild-type protein.l’ (17) McLendon, G.;Hickey, D.; Berghuis, A,; Sherman, F.; Brayer, G. In ACS Advunces in Chemistry Series; Bolton, J. R., Mataga, N., McLendon,

-

-

-

-

-

-

The Journal of Physical Chemistry, Vol. 97, No. 50, 1993 13077 G., Eds.; American Chemical Society: Washington, DC, 1991; Vol. 228, pp 179-190. (18) Mason, S.F. Inorg. Chim. Acta Rev. 1968, 2, 89. (19) Marcus, R. A.; Sutin, N. Biochim. Biophys. Acto 1985,811,265. (20) We have not been able to measure directly the Ru3+/2+potentials for the Ru(bpy)Z(im)(His) cytochromes. Instead, we have used the potential of a model compound Ru(bpy)a(im)2C12 (0.980(5) Vvs N H E 25 OC, p 100 mM NaPi, pH 7.0) to estimate driving forces. With this approximation, there is a fO.05 eV uncertainty in the ET driving force. (21) The uncertainty in A produce some uncertainty in the values of k-, but the relative values of ,k are not affected, since variations in A are likely to be quite small among the different Ru(bpy)z(im)-modified cytochromes. In correlations of k , with d or ul, the uncertainty in A will only affect the extrapolated value of ,k at close contact. (k)Moser, C. C.; Keske, J. M.; Warncke, K.; Farid, R. S.;Dutton, P. L. Nature 1992, 355, 796. (23) Newton, M. D. J . Phys. Chem. 1988,92,3049. (24) (a) Siddarth, P.; Marcus, R. A. J. Phys. Chem. 1990,94,8430. (b) Siddarth, P.; Marcus, R. A. J. Phys. Chem. 1992, 96, 3213. (c) Siddarth, P.; Marcus, R. A. J. Phys. Chem. 1993,97.2400. (d) Siddarth, P.; Marcus, R. A. J. Phys. Chem. 1993, 97,6111. (25) Goldman, C. Phys. Rev. A 1991, 43, 4500. (26) Liang, C.; Newton, M. D. J. Phys. Chem. 1992, 96, 2855. (27) Gruschus, J. M.; Kuki, A. J . Phys. Chem. 1993, 97, 5583. (28) Since -AGO A for these ET reactions, the transition state should r a m b l e the reactant conf~guration.~~ Hence, theelectron-tunnelingpathways were determined by using the crystal structure of the reduced protein.” (29) We note with admiration the many important contributions C l m made to our understanding of long-range electron transfer through u bonds in his studies of synthetic donor-acceptor complexes (Closs, G. L.; Miller, J. R. Science 1988, 240, 440). (30) The edge-to-edge limiting ET rate (01 = 1.4 A) of 1.7 X 101z s-1 is reasonably close to the limiting rates (>5 X IOi2 s-l) extracted from studies of synthetic donor-acceptor complexes.The one-bond limit in the metalto-metal scale corresponds to a hypothetical Ru-Fe dimer; given the strong coupling expected for such a complex, it is difficult to evaluate the significance of the extrapolated ET rate. It is of interest to note that Barbara and coworkers have observed ET rates near loi) s-I in CN-bridged Ru dimers (Tominaga, K.; Kliner, D. A. V.; Johnson, A. E.; Barbara,i’. F. J. Chem. Phys. 1993,98, 1228). (31) Theoretical analyses of long-range ET in proteins suggest that in certain cases aromatic residues between donors and acceptors can enhance rates viu improved coupling and lower activation barriers.)’ (32) Casimiro, D. R.; Berghuis, A. M.; Lo, T. P.; Brayer, G. D.; Richards, J. H.; Winkler, J. R.; Gray, H. B., in preparation. (33) Siddarth, P.; Marcus, R. A. J. Phys. Chem. 1993, 97, this issue.

-