Partitional and Motional Properties of a Spin-Labeled Dauuomycin in Lipid Bilayers. An ESR Study Abbas Pea&k, Vida Pezeshk, Jerzy Wojas, and Witold K. Subczynski AP, VP.

Department of Chemistry, Moorhead State University, Moorhead, Minnesota.-JW, WKS. Department of Biophysics, Institute of Molecular Biology, Jagieilonian University, Krakow, Poland. - WKS. rational Biom~ica~ ESR Center, Medical College of Wisconsin, Milwaukee, W&con&

ABSTRACT The partition coefficient of a spin-labeled daunomycin (DAU-SL) in dimyristoylphosphatidylcholine membrane has been determined using the electron spin resonance (RSR) method. The experiment was carried out as a function of temperature between S’C and 35°C. giving partition coefficients between 2 and 6 without abrupt change at the phase transition. The thermodynamic pammeters on transferring the DAU-SL from the aqueous phase to the lipid bilayer were also calculated. The calculated values ate: AH = 6.11 kcal/mol and AS = 23 cd/K mol. The partitioning of the DAU-SL aud its motion in the membrane were investigated in a wide range of pH (4-10.3). The data show that pH haa no effect on partitioning of the DAU-SL which suggest that the drug exists in the uncharged form in the bilayer.

Adriamycin and daunomycin are closely related anthracycline antibiotics which are in clinical use for the bunt of a wide spectrum of human cancers [l-3]. Unfortunately, there is a risk of severe cardiotoxicity arising from multiple factors, namely oxidative stress [4] and inhibition of Na+ /Ca*’ ion exchange in the heart sarcolemma [S]. From the initial discovery of the antitumor activity of these drugs, efforts have been made to reduce their toxicity and tbe harmful side effects when used medicinally. Tbe toxicity associated with this class of antibiotics is thought to result from the conversion of the drugs into toxic free radicals following their

Address reprint requests to: Dr. Abbas Rex&k, Department of Chemistry, Moorhead StateUniversity, Moorhead, Minnesota 56563. Journal of ~nogonic Biochemistry.46.67-76 (1992)

67 0 1992 Elsevier Science Publishing Co.. Inc., 655 Avenue of the Americas, NY, NY 10010 0162-0134/92/$5.00

68

A.

metabolism. It has been suggested that the quinone moiety of the drug can participate in redox reactions by accepting electrons from microsomal and flavoprotein resulting in the formation of the semiquinone and hydroxyl free radicals [6-91. Thus, it appears that these oxygen-derived toxic radicals, particularly the OH ’ radical, are the triggering agents in cardiotoxicity and lipid peroxidation. It has also been shown that the iron-anthracyciine complexes cause a more extensive physiological damage (i.e., lipid peroxidation and DNA damage) than ~~ycline alone [lo-121. Formation of active oxygen species has been proposed to account for such damage. Recently, it was shown that the introduction of nitroxide free radical moiety into the structure of some cytostatic anticancer agents results in considerably decreased toxicity of parent drugs [ 13- 171. Emanuel et al. have shown that the spin-ladle derivative of daunomycin had lower cardiotoxicity than the drug itself and that it did not inhibit hemopoiesis at the ma~mum tolerated doses [ 13, 141. In addition to the pharmacological advantages, the spin-labeled drugs can be followed using ESR spectroscopy [ 181. The present study was un&~en to destine the interaction of DAU-SL with model membranes in order to better understand its action and transport into cellular targets.

MATERIALS AND METHODS Materials A spin-labeled derivative of ~unomy~in [Scheme l] (DAU-SL) was prepared from daunomycin and 4-amino TEMPO by the method analogous to the published procedure [19]. Dim~stoylphosphatidylcholine (DMPC) was purchased from Sigma (St. Louis, MO.). Paraffin oil (light) was obtained from MCB Manufacturing Chemists, Inc. The following buffers (prepared in deionized water) were used. 0.2 M acetic acid for pH 4-6, 0.2 M phosphoric acid for pH 7-8, and 0.1 M boric acid for pH 9-10.3. For all rn~ure~n~ except that for pH dependence, a phosphate buffer with pH 7.0 was used. Sample Preparations Liposomes used in this work were multil~e~ar dispersions of lipid. A mixture of lipid (1.0 x 10m5 mol) and DAU-SL (0.5 x 10W7mol) in chloroform was dried with a stream of nitrogen and further dried under a reduced pressure (0.1 mm Hg) for at least 12 hr. Buffer (0.1 ml) was added to dried lipid at 45’C and vortexed vigorously. The lipid dispersion was centrifuged briefly at 4’C and the loose pellet (20% w/w)’ was used for ESR measurements. The sample was placed in a capillary (0.7 mm i.d.) made of gas permeable methylpentene polymer called TPX [20]. This

PARTITIONAL AND MOTIONAL PROPERTIES

69

plastic is permeable to nitrogen, oxygen, and carbon dioxide and is substantially impermeable to water. A TPX sample tube was placed inside the ESR dewar insert and equilibrated with nitrogen gas that was used for temperature control. The sample was thoroughly deoxygenated to obtain the correct ESR line shape. ESR Measurements ESR spectra were obtained with a Varian E-3 or a Varian E-109 X-Band spectrometer with Varian temperature control accessories and an E-231 Varian multi-purpose cavity (rectangular T Mode). A magnetic field modulation amplitude of 0.5 G was used, with a microwave power of 5 mW. Partition Coefficients The partition coefficient K, the ratio of the concentration of the DAU-SL within the DMPC bilayers to that in the aqueous phase, was determined at four temperatures: 4, 21, 25, and 35°C. For these measurements, lipid dispersion containing 3.5 mg DMPC, 0.3 x lo-’ mols DAU-SL and 33 PL of buffer was equilibrated 30 min at a given temperature (4, 21, 25, or 35 “C) and then centrifuged at the same temperature. The intensity of the ESR signal of DAU-SL in supematant was measured and compared with that from a buffered standard (0.3 x lo-’ mol DAU-SL dissolved in 33 PL of pure buffer). For calculation of K, it was assumed that the densities of water and lecithin are the same [21, 221, giving a value of 9.6% of the sample volume occupied by the membranes. The average concentration of the DAU-SL within the whole bilayer was calculated and used.

RESULTS AND DISCUSSION 1. Localization and Motion of DAU-SL in Liposome Suspension At room temperature, the DAU-SL is soluble in water, chloroform, and paraffin oil, giving rise to ESR signals of fast tumbling spin label with average hypertine coupling constants, aiso, of 17.0 G, 15.8 G, and 15.3 G, respectively. It shows that DAU-SL is soluble in polar and hydrophobic solvents and that ESR spectra are sensitive to the polarity of the solvent. The ESR spectrum of DAU-SL in DMPC suspension consists of two components, an immobilized component resulting from the DAU-SL being bound to the membrane, and the fast rotating component resulting from the DAU-SL being in the aqueous phase (Fig. l(A)). Figure l(B) shows the isotropic ESR spectrum of the supematant which resembles the aqueous component (ais0 = 17.0 G) of the mixture shown in Figure l(A). The pure lipid component of the DAU-SL spectrum was obtained by the addition of a paramagnetic line broadening agent, potassium ferricyanide (K,Fe(CN),)‘, to the buffer during the preparation of liposome bilayers. The ferricyanide broadens the signal arising from the aqueous component inside and outside the liposomes due to Heisenberg exchange and dipole-dipole interaction [25]. The ESR spectrum of the resulting membrane bound DAU-SL is shown in Figure l(C). As expected, the broadening agent has strongly

‘Potassium ferricyanide is known as a very stable complex with the formation constant of 104’ [23], while the formation constant for Fe(DAU), complex is 10” [24]. Thus it seems unlikely that a complex is formed between iron and DAU-SL in the presence of K,Fe(CN),.

70

A. Pezeshk et al.

A

B

C

FIGURE 1. BSR spectra of DAU-SL. (A) in suspensionof DMPC &osomes containing 29% lipid w/w; (B) in suspernatant after centrititgation of the liposornes; (C) in suspension of DMPC liposomes (20% lipid w/w) in the presence of 0.5 M K,Fe(CNj,. All spectra were messured at 26°C. Maximum splitting of the membrane bound (hydrophobic) component, as well as peak heights of polar (P) and hydrophobic 0 components are shown.

decreased the polar component of the spectrum and has little effect on the lipid component. This clearly shows that the immobilized component is due to the DAU-SL being bound to the membrane in such a way that the spin-label moiety is in the hydrocarbon-hydrophobic region protected from broadening by the polar paramagnetic probe K,Fe(CN),. A small drug-to-lipid mole ratio ( C 0.002) was maintained throughout the study to reduce the degree of self-association. Similar anthracycline-to-lipid mole ratios (< 0.002-0.005) have been used in light absorption and fluorescence measurements [27, 281. The maximum splitting parameter of the membrane bound component (see Fig. l(A)) can be used to determine the mobility of the DAU-SL’ in the bilayer. This parameter was obtained from the ESR spectra of the DAU-SL in the wide range of temperatures up to 30°C (Fig. 2). At higher temperatures, because of overlapping of the polar and the immobilized components, the broadening agent, K,Fe(CN),, was used to distinguish the low-field peak. Figure 3 shows the maximum splitting parameter as a function of temperature. The data reveals that the’motion of the DAU-SL within the DMPC bilayer is sensitive to temperature, showing clear abrupt changes at the main phase transition of DMPC (23.6”C). 2. Membrane/Water

Partition Coefficient of DAU-SL

The DAU-SL spectral parameter f = H/(H+ P), defined by analogy with the TEMPO spectral parameter [26], is roughly proportional to the fraction of the DAU-SL being in the DMPC bilayers. In this equation, P is the polar peak height and H is the hydrophobic peak height of the ESR spectrum of the DAU-SL as shown in Figure l(A). The DAU-SL spectral parameter is found to decrease when decreasing the temperature with abrupt changing at the main phase transition (Fig. 4). The abrupt change at the main phase transition can be the result of changes in partitioning of the DAU-SL and/or the result of a change in the ESR line width of the lipid component. A plot of the polar P and the hydrophobic H components as a function of temperature is presented in Figure 4, showing no change in the slope of P across the phase transition temperature. The results indicate that the abrupt change of the DAU-SL spectral parameter at the main phase transition is due to the abrupt line broadening of the membrane bound component, and not the abrupt redistribution of the DAU-SL.

PmTITIONAL

AND MGTIONAL PROPERTIES

71

PIGURE 2. ESR spectra of DAUSL in D~~~at~t~ perames. The outer whlgs were also magnified by mmding at 10 times higher receiver ~~~~~er~~. The DAU-SL spectral pammeter can easily be obtain& from ESR spectra but it gives only qualitative information about the distribution of DAU-SL between membrane and water. The more meaningful pammeter is the DAU-SL partition coefficient K = DAU-SL (~~)]/[DAU-SL (water)] (see Materials and Methods), which was obtai& for DMPC liposomes. A plot of log K as a function of inverse temperature is shown in Figure 5. No abrupt change of K values or the slope at the main phase transition is observed. This somewhat sling result is also con&med bythedataob&inedfortheDAU-SLspectralpamme@rasdiscussedabove. Onthis basis, it is assumed that tbe best linear fit for temp&raaue &on 4-35’C will be a good ~~~~ to obtain disc ~ which will ckacteke transport of DAU-SL from water to the lipid bilayer. The partition coetkient K is relakd to the partial molar stawLvd free energy change (AC) on transferring DAU-SL from water to the lipid bilayer, by K = expf - AG/RT) .

)

(1)

FIGW 3. Maxkmun splitting of DAU-SL in DMPC snmnw plotted as a function of tempmture. Measure35 and 45”C, were obtakd in ~ntsath&htempetatuns, the presence of 0.5 M of K,R(CN), to reduce the intensity oftheEsRsignalintheaqueousphase.Thepresenoeof K,Fe(CN), had no effect on the ESR signal at lower tnmpmtaues. The vertical broken line shows the main transition temperamm.

72

A. Pezeshk et al.

1’

1

x

P

“I

”

~

H

L-+-c 10

20

30

5d

40

TEMPERATURE

(“C)

.

FIGURE 4. DAU-SL spectml parameter f in DMPC suspension plated as a fnnction of temperature. The DAU-SL spectral parameter is defined as f = H/(H + P) (see Fig. l(A)). Also, changing of amplitude for polar (P) and hydrophobic @I) component in the spectrum is shown. The vertical broken lie indicates the main transition temperature.

AH, and the entropy change, AS, on ~sfe~g DAU-SL from water to the lipid phase are related to AG by The partial enthalpy change,

AG=AH-TAS.

(2)

Making the cable scion that AH and AS are independent of temperature, all three thermodynamic quantities can be determined. The calculated values kcaljmol; and at 4°C are: K = 1.8, at 35°C are: K = 5.5, AG= -1.04 AG = - 0.32 kcal/mol. For all temperature regions (4-35”(Z), AH = 6.11 kcal/mol and AS = 23 cal/K mol. Similar values of AG, AH, and AS for partitioning of small spin labels between water and fluid phase lipid bilayer were obtained by Dix et al. [29, 301. Polar groups in DAU-SL are expected to decrease the K value since they lower the free energy of the solute in water via hydrogen bond formation. On the other hand, hydrocarbon hydrophobic groups tend to increase the K value by increasing the free energy of the solute in water and decreasing it in the membrane because of van der Waals forces between the solute and the hydrocarbon chains of the membrane. The presence of both polar and nonpolar groups in DAU-SL results in a small variance of K as a function of temperature and fluidity of the bilayer. According to Eq. (2), both AH and AS affect AG (and K) in opposite directions, since AH and AS are large and positive, AG is rather small, giving rise to a small partition coefficient of DAU-SL between membrane and water. The thermodynamic parameters calculated in this work are different from those reported for unlabeled adriamycin or daunomycin obtained in small unilamellar liposomes using fluorescence spectroscopy [27, 311. The partition coefficient for DAU-SL is much smaller than for an unlabeled drug and increases with temperature, while it decreases with temperature for au unlabeled drug. The difference in partition coefficient is due to the structural changing created by spin labeling of the drug

TEMPERATURE

1000/T

f”C)

(K-‘)

FXGUIUE5. The partition coefficient K of the DAU-SL in DMPC bilayers as a function of l/T. The vertical broken line shows the main phase transition temperamre.

PARTITIONAL AND MOTIONAL PROPERTIES

73

molecule. It is also possible that using different membrane systems, i.e., multilamellar liposomes in this work versus small unilamellar liposomes, could contribute to different partitioning. 3. DAU-SL Transport Across DMPC Bilayer It is known that the negative logarithm of the dissociation constant for the membrane dissolved component, so called pK,, is different from pK value of the same group in the bulk solution [32]. According to Chaimovich et al. [33], pK, corresponds to the pH in the aqueous phase when the charged and the uncharged species are equal. Shift of the pK value of one unit was reported for tetracaine in the presence of 0.8 % w/w egg yolk phosphatidylcholine membranes [34], and as large as 1.2 and 3.4 units for 5doxylstearic acid and 16-doxylstearic acid in phosphatidylcholine membranes, respectively [35, 361. The pK value for amino sugar groups of adriamycin in aqueous solution was estimated at 8.2 [37]. A pK, of 7.6 was obtained for the adriamycin amino group by studying its efflux from erythrocytes into a drug free medium [38]. To better understand the transport of DAU-SL into membrane, we investigated the influence of pH on its motion within the bilayer and its partitioning into the bilayer. Our results as shown in Figure 6 exhibit neither changing of the DAU-SL spectral parameter nor the maximum splitting parameter in ESR spectra for pH ranging from 4 to 10.3. These results were obtained for samples containing a large amount of lipid (20% w/w), which can induce a large shift of pH, to lower values and results in the formation of the uncharged form of the DAU-SL within the bilayer. (We consider zwitterion as a charged form.) To acquire a qualitative picture, the transport of the DAU-SL from the multilamellar DMPC liposomes was studied with DAU-SL both inside and outside the liposomes. To the above suspension, 10 mM sodium ascot-bate (pH 7.0) was added, and the kinetic of the chemical reduction of the DAU-SL was followed using its ESR signal (Fig. 7). The experiment was carried out at 25 “C. About one min after the addition of the ascorbate, the polar component was sharply reduced to about 10% of its original height. The 10% is approximately the volume of the water which is trapped inside the multilamellar liposomes for a sample containing 20% lipids w/w and is not accessible to ascorbate. Moreover, a decrease in the peak heights of both the hydrophobic and the polar components were observed with the same kinetics which indicates that the equilibrium between the lipid phase and the aqueous phase is fast. The slowest process is probably the crossing of the DAU-SL through the bilayers by the “Q-flop” mechanism [39]. Our data indicate that the reduction is fast with a half-life of about 4 min. This reduction time includes the redistribution of the DAU-SL between water and the bilayer and the time of flip-flop across the bilayers . The mechanism of action of the anthracycline antibiotics is fairly complicated and

FIGURE 6. Maximum splitting of the DAU-SL PH

and the DAU-SL spectral parameter for DMPC suspension plotted as a function of pH at 25 “C.

74

A. Pezeshk et al.

HGURE

7.

Kinetic of DALI-SL reduction in DMI’C

suspension after addition of 10 mM Na ascot-bate at 25’C. The average concentration of DAU-SL in the sample was 0.5 ntx

(min)

mM.

includes several stages, i.e., the interaction with the cytoplasmic and cell plasma membranes [40-421, and the intercalation and binding with the DNA of the cell [24]. It is known that most important cellular targets for these antibiotics are nucleic acids [24, 421. In order for a drug to reach these target sites, it has to diffise through aqueous extra cellular and polar hnracelhtlar environment, and penetrates through the hydrophobic bilayer of cell membranes. To overcome these two barriers, the drug must possess neither highly hydrophilic nor highly hydrophobic character. This is in agreement with the concept of Sosnovsky [ 15, 17, 43, 441 stating that moderate partition coefficients between membranes and water (or octanol and water) correlate with higher activity of the drug. DAU-SL fulfills this condition, giving partition coefficient K of about 4-6 for the fluid phase membranes, and about 2-4 for the gel-phase membranes. The values obtained in this work are much smaller than those reported for the unlabeled anthracyclines [27,3 11. This directly correlates with a larger biological activity of the spin-labeled daunomycin [ 13, 14, 161. It is known that anthracyclines penetrate the plasma membrane by passive diffusion of the uncharged form of the drug molecule through the lipid domains of the biomembrane [38,45]. Our results show that this mechanism can be effective for DAU-SL and would work in a wide range of pH because of shifting the pK, to lower values. Although the exact mechanism by which an anthracycline transverses the bilayer is not known, we suggest that the limiting process for DAU-SL is its “‘flip-flop” (or jump) across hydrophobic-hydrocarbon central part of bilayer. This is based on our kinetic measurements and on the fact that anthracycline possesses a large number of polar groups that may cause the molecule to anchor to the membrane surface. As stated earlier, the presence of nitroxide moiety in DAU-SL decreases the toxicity of daunomycin. It is connected with chemical activities of nitroxides which could act as antioxidants [46,47] or could mimic the superoxide dismutase activity [48,49]. We, therefore, believe that spin labeling of drugs is a promising direction in antitumor chemotherapy. Thk research was supported by NIH Grant No. RI5 GM-42006 (A.P.) and by the Nationul Biomedical ESR Center through NIH Grant No. RR 01008.

REVERENCES

1. P. Arcamov, in Doxorubicin Anticancer Antibiotics, Academic, New York, 1981. 2. F. Arcamone, Med. Res. Rev. 4, 153 (1984). 3. L. Gianni, B. J. Corden, and C. E. Myers, Rev. Biochem. Toxicol. 5, 1 (1983).

PARTITIONAL AND MOTIONAL PROPERTIES 75

4. R. D. Olson, R. C. Bocrth, J. G. Gerber, ad A. S. Nies, Lift Sci. 29, 1393 (1981). 5. P. Caror& P. Vii, and E. Cbafoli, FEBS L&t. 130,184 (1981). 6. N. R. Bachur, S. L. Gordon, M. V. Gee, and H. Kon, Proc. Nod. Acud. Sci. U.S. A. 76,954 (1979). 7. M. S. Ce&er, Biochem. Biophys. Res. Commun. 89, 1231 (1979). 8. 1. Goodman and P. Ho&stein, Biochem. Biophys. Res. Cbmmun. 77,797 (1977). 9. A. Someya ami N. Tanaka, J. Antibiot. 32, 839 (1979). 10. C. E. Myers, L. Gianni, C. B. Simone, R. Klecker, ad R. Greene, Biochemistv 21, 1707 (1982). 11. W. ELiot, L. Giad, and C. Myers, Biochemistry 23,928 (1984). 12, J. L. Zwcir, J. Biol. Chem. 259, 6056 (1984). 13. N. M. Emnnuel, N. P. Konovalova, R. F. Diihkovskaya, and L. K. Den&ova, Antibiotiki 27, 811 (1982). 14. N. M. Emanuel, N. P. Konovalova, and R. P. Diatchkovskaya, Neopiusma 32, 285 (1983). 15. T. Klenncr, M. R. Rerger, G. Sosnovsky, N. U. Rao, P. Wiier, ad D. Schmahl,

Amh. Geschwu&f~h. 59(4), 251 (1989). 16. 0. A. Shiryaeva, N. A. Semenova, L. A. Sibeldina, S. A. Goncharova, N. P. Konovnlova, Eksp. Onkoi. 11(4), 70 (1989). 17. G. Sosnov&y and N. U. Rao, J. Phurm. Sci. 79(4), 369 (1990).

and

18. N. P. K~~ov~, R. F. ~~ovs~y~, G. V. K&u&k& L. M. Volkova, V. N. Varfolomeev, L. S. Dombrovsky, and A. B. Shapiro, NeopZasmo 35(2), 185 (1988). 19. E. J. Rauckman, M. R. Rosen, and W. Wii IIord, Org. Prep. und Proc. Int. 9, 53

(19n). 20. J. S. Hyde and W. K. Subczynski, in Biological Magnetic Resonance, Vol. 8; Spin

21.

22. 23. 24.

25. 26. 27. 28. 29. 30. 31.

Labeling: Theory and Applications, L. J. Berliner and J. Reuben, Eds., Plenum Press, New York, 1989, pp. 399-425. J. P. Nagle and D. A. Win, Biophys. J. 23, 159 (lW8). W. Km& Chem. Phys. Lipids 28, 337 (1981). La@ Handbookof Chemistry, J. A. Dean, Ed., McGraw Hill Book Company, New York, 1%7. H. Beraldo, A. Gamieduillerot, L. Tosi, and P. Lavelle, Biochemistry 24,284 (1985). G. I. -in, Spin Labeling Methodr in MolecukzrBioIogy, ~~~-~~, New York, 1976. H. M. M&m&, in Spin Lubeling, 7’heory and Ap~~t~n, L. J. Be&i, Ed., Academic, New York, 1976, pp. 525-560. T. G. Burke and T. R. T&ton, Bi~~rn~t~ 24, 1768 (1985): A. m, P. L.-G. chOn8, Y. Bareaholz, and T. E. w, Cuncer Res. 46,594 (1986). J. A. Dix, J. M. Diamond, ad D. Kivelson. Proc. Natt. Acud. Sci. U.S. A. 71,474 (1974). J. A. Dix, D. Kivelson, and J. M. Diamod, J. Membrune Biol. 40, 315 (1978). T. G. Burke, M. Israel, R. !&shim& and J. H. Doroshow, Biochim. Biophys. Actu

982, 123 (1989). 32. A. G. Lee, Biochim. Biophys. Actu 514,95 (1978). 33. H. Chaimovich, R. M. V. Aleixo, I. M. cuccovia, D. w, & P. H. ~uina, i,, &&&on BehoviOr Of ~hctmts-bmtical and Applied Aspects, E. J. F&&jr d K. L. Mind, ads., Vol. 2, Plenum PWS, NewYork, 1982, pp. 949-974. 34. S. Sdueiex, W. A. FrezzaUi Jr., P. S. Aranjo, H. Chaimovich, aad I. M. Cuccovia,

B&him. Biophys. Acta 769,231 (1984). 35. A. Saason, M. Ptak, J. L. R@aud, and C. M. Gary-E&o, Chem. Phys. Lip&& 17, 435 (1975).

76

A. Pezeshk et at.

36. A. Kusumi, W. K. Subczynski, aud J. S. Hyde, Fed. Proc. 41, 1394 (1982). 37. R. J. Sturgeon and S. G. Schulman, .I. Pharm. Sci. 46, 958 (1977). 38. M. Dalmark, Stand. J. Clin. Lab. invest. 41,633 (1981). 39. R. D. Komberg aud H. M. McCoMel, Biochemir;try10, 1111 (1971). 40. E. ~~ and 3. M. Ruys~~~, Res. Commun. Chem. Pathoi. Pharmocol.

42, I49 (1983). 41. T. R. Tritton and G. Yee, Science 217,248 (1982). 42. 2. A. Tokes, K. E. Rogers, and A. Rembaum, Proc. Natl. Acad. Sci. U.S. A. 79, 2026 (1982). 43. P. Z. Gutierrez, M. Konieczuy, and G. Sosnovsky, 2. ~aturfo~ch 36(6), 1612 (1981). 44. G. Sosnovsky, Pure and Applied Chem. 62, 289 (1990). 45. M. Dalmark aud H. H. Storm, J. Gen. Physiol. 78, 349 (1981). 46. M. Hicks and J. M. Gebicki, Arch. Biochem. Biophys. 210, 56 (1981). 47. J. T. Weil, J. Van der Veen, and H. S. Olcott, Nature 219, 168 (1%8). 48. A. Samuni, M. C. K&ma, P. Riesz, E. Finkelsteiu, and A. Russo, J. Biol. Chem. 263, 17921 (1988). 49. J. B. Mitchell, A. Samuni, M. C. Krishna, W. G. DeGrti, M. S. Ahn, U. Samuui, and A. Russo, Biochemistty 29, 2802 (1990).

Received August 13, 1991; accepted October 0, 1991