Jun 8, 1982 - Tomoko Ohnishi, Russell LoBrutto, John C. SalernoS, Robert C. Bruckner, ... according to Blokzijl-Homan and Van Gelder (12) and Wilson et al.
THEJOURNAL OF BIOLOGICAL CHEMISTRY Val. 257, No. 24, Issue of December 25, pp. 14821-14825, 1982 Printed m U.S.A.
Spatial Relationship between Cytochrome a and a3* (Received for publication, June 8, 1982)
Tomoko Ohnishi, Russell LoBrutto, John C. SalernoS, RobertC. Bruckner, and TerrenceG. Frey From the University of Pennsylvania, Department of Biochemistry and Biophysics, Schoolof MedicinelG4, Philadelphia, Pennsylvania 19114 and the +Departmentof Biology, Rensselaer Polytechnic Institute, Troy, New York 12181
We havestudiedthespatialrelationship between according to Blokzijl-Homan and Van Gelder (12) and Wilson et al. dithionite in the cytochromes a and a3by the enhancement of the spin (7) by treating oxidasewithhydroxylamineand potentiometric titration vessel under anaerobic conditions (13).After relaxation of cytochrome EPR signals by the 3 min of incubation at room temperature, redox mediator dyes were paramagnetic a heme at 15 K. An Fe-Fe distanceof 12- added, andthe oxidation reduction potential(Eh) of the NO-liganded 19 A is estimated from the absence of dipolar broad- oxidase preparation was raised by the addition of KaFe(CN)G. ening and from the observation of spin relaxation enEPR spectrawere obtained using a Varian E109 spectrometer. Low hancement in the a3-N0 complex. When this result is temperatures were attained with a variable temperature heliumflow combinedwithresonance x-ray diffractiondata re- cryoslat system (Air Products LTD-3-110). Temperatures were measported by Blasie et al. (Blasie, J. K., Pachence, J. M., ured with a calibrated carbon resistor locatedjust below the sample. a Nicoletmodel 1024 signal Tavormina, A., Dutton, L., P. Stamatoff, J., Eisenberger, Spinquantitation wasconductedon P.,andBrown, G . (1982) Biochim. Biophys. Acta 679, averager. Microwave powersaturation analysis and estimationof the saturation parameter ( P I ~were z ) conducted as previously described 188-197) and the contribution from the exchange inter(14),which is not the empirical “half-saturation”power. action is considered, we can limit the iron-iron distance to 12-16 A and estimate the angle between the Fe-Fe RESULTS vector and mitochondrial membrane normal as 30-60”. We also consider the possible effects of CUA on cytoRecently, theusefulness of the NO ligand as a probe of the chrome a3-NO. cytochrome a3 heme has been demonstrated by a number of investigators (7,15-18). Nitroxideforms a stable complex with the reduced hemea3 and remains liganded over a wide range of redox potentials. This property provides an excellent opportunityto analyzespin-spin interactionsbetweencytochromes a3 and a. We obtained the Q-NO complex in the presence of paramagnetic cytochrome a and CuA by poising the sample at an E h value of 357 mV (sample I) and in the presence of diamagnetic cytochrome a and CUA by poising at 100mV (sample 11). Fig. 1 shows EPR spectra A and B of samples I and I1 of themembranouscytochrome oxidase preparation, respectively, recordedin a wide span of the magnetic field (Ho) with 5 milliwatts microwave power setting, at a sample temperature of 10.4 K. Potentiometric analysis of cytochrome a and CUA, monitored with signal amplitudes at g = 3.03 and g = 1.99, respectively, showed that both redox centers are more than 85% oxidized at 357 mV and completely reduced at 100 mV. Spectrum C was obtained by subtracting spectrumB from spectrumA. It clearly showsthe cytochrome a absorption withprincipal g values of 3.03, 2.25, and 1.47, and the CuA resonance in gthe = 2 region. Since the EPR conditions used forFig. 1were chosen for the cytochrome a and CUAsignals, the as-NO signal in SpectrumB is partially saturated the Ua-NO signal in Spectrum A is hidden by the CUA signal. The spectra in Fig. 2A show the detailed lineshape of the MATERIALSANDMETHODS nonsaturated aa-NO spectra of samples I and I1 at 26 K. A modulation amplitude (2 x tesla) and an exTwodifferentkinds of cytochrome c oxidase preparations were smaller these spectra. The as-NO used in this study, membranous oxidase prepared according to Frey panded field scale were used for et al. (IO) anddetergent-solubilizedcytochromeoxidaseprepared spectrum of the oxidase (sample I) with paramagnetic cytoaccording to Hatefi (11). Cytochrome aa-NO complex was prepared chrome a and CuA is represented in the solid line, and the spectrum of sample I1 with diamagnetic cytochrome a and * This work was supported by Research Grant PCM 81-17284 from CUA isrepresented by the dashed line. Atnonsaturating the National Science Foundation and Research Grants GM-25052 power levels( T,,, the allowed area for r and 8 must fall under curve A (or curve A'; if we leave room for a factor of four error in the ratio of the measured P I / Zvalues). The a3-N0 spectrum has multiple turningpointscorresponding to the orthogonal axes of the g tensor coordinate system and the largest numerical principal value of the A
a and a3
tensor coordinate system, oriented at an intermediate direction. All of these turning points cannotbe near a magic angle direction. In addition, even when a species such as cytochrome a with no resolved h y p e r h e splitting is oriented so that all three principal axes are along "magic angle" directions, splitting from dipolar terms would be observed near g, if coupling were tight enough. Preliminary simulationssuggest that lineshape changes could be easily observed in these species regardless of their orientation with respect to the ma3vector if they were closer than 12 A. The limits on r and 8 obtained from EPR data alone are rather broad. However, Blasie et a,?. (25) have applied the novel technique of resonance x-ray diffraction to oriented multilayers of oxidase reincorporated into liposomes. That study provided a crucial piece of information: The z (membrane-normal) coordinates of the two heme irons differ by either 8 8, or 38 A. Since it is unlikely that any relaxation enhancement would beobserved at 38 hi, we willconsider only the case in which Az = 8 (k1.4) A. Straightforward geometry leads to the expression r
=
A21 I cos0 I
(5)
Because we have used g = 1.98 for os,8 represents both the angle between the interiron vectorand themembrane normal and also the angle between the same vector and the applied magnetic field direction. Equation 5 generates the curve C, D, and E in Fig. 5. The allowed area of r and 8 must lie between curves C and E . Because of the opposing dependencies of r on cos8 in the xray and EPR data, we can greatly narrow the range of the possible values of both r and 8, as illustrated with the shaded area of the figure. Thus 12 A < r < 14 hi and 30 < 8 e 50" for dipolar coupling alone. Errors discussed below might allow r and 8 to be somewhat larger. Several assumptions underlie this distance calculation. Perhaps themost important is that theeffect of the CUA"center on the spin-lattice relaxation time of cytochrome as"-NO is not significant. This is not unreasonable, since the P1/2for CUA'I is much smaller (P1/2= 5.4 milliwatts) than that of cytochrome a'" at 15 K. But because there is substantial overlap between the CuAand cytochrome as"-NO spectra, the B term effects are maximized, and the CUA center must be considered as a possible source of error. It is very difficult to estimate the effect of CUA" because the orientation of its g axes with respect to the membrane normal in the multilayer system is unknown. In addition, the oriented multilayer data on cytochrome a;'-NO have not been checked by simulation. A moderate error (5-10") in the orientation of the g = 1.98 axis could conceivably generate arather large error in Aw. The effect of this error is greatly mitigated, though, by the fact that the sixth root of r is taken at theend of the calculation. Simulations are under way to c o n f m the choice of g value in the cytochrome a;'-NO spectrum. 0 20 40 60 80 If exchange coupling is important, the details of the calm8 ( Degree ) lation are somewhat different but the results are much the FIG. 5. Fe-Fe distance of cytochrome c oxidase as a function same. The isotropic exchange matrix elements have no anguof the angle between the Fe-Fe vector and the membrane lar factor. The magnitude of the B type terms (the exchange normal. Curves A , A', and B drawn in solid lines are based on spin terms in S1+S2and S,-S2+) must be comparable to thedipolar ) ~curves . relaxation data wherer6= y?y:k2/8Au2 ( T I , / T I ~ ) (CI -O~S ~ ~ In A and B, TzJT~fratioof 0.4 and 10, respectively, were assumed based terms already treated. We expect interactions at distances on the measured values. Curve A' corresponds to a case in which much over 10-12 8, to be dipolar dominated for the sake of PI/*is in error by a factor of 4; other conditions are the same as for argument we allowthe upper limitfor an exchange interaction curve A. The horizontal dashed line is the minimum r value for of this magnitude to be 15-16 A, since this depends on the which no spin-spin splitting is discernible. Curves C, D, and E shown relative orientation of the orbitals of the paramagnetic molein dotted lines give r versus 0 from resonance x-ray diffraction data cules (26). Then theprincipal difference between this and the of Blasie et al. (25). In curve C, D,and E , r = 9.4/ I cos0 1 , r = 8.0/ I cos0 I , and r = 6.6/ I cos0 I were used, respectively. The shaded dipolar coupled case is the loss of the minimum at the SOarea represents the range of r and 0 values consistent with both x-ray called magic angle, which allows 6' to be as large as 60" and r to be as large as 16 A. The lower limits on 8 and r are and EPR data.
Spatial Relationship
between Cytochrome a and
Normal to the Membrane
I
I ( g = 2.10)
I
I
x ( 1.5)
I
a3
14825
information is the orientation of the heme normals with O as indicated by the upper respect to the u ~ ~ ~ - u " - Nvector, arrows in Fig. 6 showing rotation about membrane normals. Both heme normaIs must, of course, lie in the plane of the membrane. The major fiiding of this study is the presence of magnetic coupling between cytochrome a and the cytochrome aa-NO complex. We hope to refine our interpretation of the data by obtaining TI values frompulsed EPR studies in the near future. Acknowledgments-We are grateful to Dr. J . K. Blasie for helpful discussion of his x-ray data. We also thank Drs. J . S. Leigh, B. Chance, H. Blum, and C. Kumar for stimulating discussion.
y (1.98)
REFERENCES 1. Erecinska, M., and Wilson, D. F. (1978) Arch. Biochem. Biophys.
188, 1-14 2. Palmer, G . (1978) in Mechanisms of Oxidizing Enzymes (Singer,
3. 4. 5. Cyt.
6.
gs -NO
7.
Both g,,
NormaltotheHemePlane
8. 9.
T. P., and Ondarga, R., eds) pp. 221-238, Elsevier/North-Holland Biomedical Press, Amsterdam Malmstrom, B. G . (1979) Biochim. Biophys. Acta 549,281-303 Azzi, A. (1980) Biochim. Biophys. Acta 594,231-252 Wikstrom, M., Krah, K., andSaraste, M. (1981) Cytochrome Oxidase, A Synthesis, Academic Press, New York Leigh, J. S., Jr., Wilson, D. F., Owen,C. S., and King, T. E. (1974) Arch. Biochem. Biophys. 160,476-486 Wilson, D. F., Erecinska, M., and Owen,C. S. (1976) Arch. Biochem. Biophys. 175,160-172 Malmstrom, B. G. (1973) Q.Rev. Biophys. 6,389-431 Linne, B., and Vanngird, T. (1978) Biochim. Biophys. Acta 501,
449-457 10. Frey, T. G., Chan, S. H. P., and Schatz, G. (1978) J . Biol. Chem. 253,4389-4395 between cytochrome a a n d a8 in cytochrome oxidase in the 11. Hatefi, Y. (1978) Methods Enzymol. 53,40-47 membrane. 12. Blokzijl-Homan, M. F. J., and Van Gelder, B. F. (1971) Biochim. Bwphys. Acta234,493-498 unchanged; while exchange and dipolar terms can cancel at 13. Dutton, L. (1978) Methods Enzymol. 54,411-435 some orientations, they mustadd at others. Thus, considering 14. Blum, H., and Ohnishi, T. (1980) Biochim. Biophys. Acta621,9both the effects of exchange coupling and of possible g axis 18 orientation errors, the a"'-aJ"-NO distance should lie between 15. Stevens, T. H., Brudwig, G.W., Bocian, D. F., and Chan, S. J. (1979) Proc. Natl. Acad.Sci. U. S. A . 76, 3320-3324 12 and 16 A. 16. Brudvig, T. H., Stevens, T. H., and Chan, S. I. (1980) Biochemistry Our picture of the overall geometry of the u " ' - u ~ ~ - N O 19,5275-5285 system, which is illustrated in Fig. 6, is easily summarized in 17. Yoshida, S., Hori, H., and Orii, Y. (1980) J. Biochem. (Tokyo) 88, intuitive terms. If the a"l-a~ll-NO vector were along the mem1623-1627 brane normal, the observed relaxation could have been gen- 18. Boelens, R., Rademaker, H., Pel, R., and Wever, R. (1982) Biochim. Biophys. Acta679,84-94 erated at distances of up toapproximately 25 A. However, this is inconsistent with the results of the x-ray diffraction data, 19. Hyde, J. S., and Rao, K. V. S. (1978) J. Mugn. Reson. 29,509-516 and in fact if the U ' * ' - ~ ~ ~ -vector N O were less than 30" from 20. Bloembergen, N. (1949) Physica 15, 386-426 21. Abragam, A. (1955) Physiol. Rev. 98, 1729-1735 the membrane normal we would expect to see splitting and/ 22. Barlow, C., and Erecinska, M. (1979) FEBS Lett. 98, 9-12 or line broadening. 23. Erecinska, M., Wilson, D. F., and Blasie, J . K. (1978) Biochim. At angles of over 60",the x-ray data are inconsistent with Biophys. Acta 501,53-62 center-to-center distances of less than 16 8, and suggests 24. Blum, H., Harmon, H. J., Leigh, J. S.,Salerno, J. C., and Chance, B. (1978) Biochim. Biophys. Acta 502, 1-10 distances of 20 8, or more. This does not compare well with power saturation data. Between 30" and 60" the x-ray and 25. Blasie, J. K., Pachence, J. M., Tavormina, A., Dutton, P. L., Stamatoff, J., Eisenberger, P., and Brown., G. (1982) . . Biochim. EPR results agree reasonably well. Biophys. Acta 679, 1881197 Weknow from previous work that both heme planes are 26. Coffman, R. E., and Buettner, G. R. (1979) J. Phys. Chem. 83, perpendicular to the plane of the membrane. The missing 2392-2400
FIG. 6 . Schematic presentation of the spatial relationship
