J. CHEM. SOC. FARADAY TRANS., 1994, 90(18), 2643-2648 EPR Spin-labelling and Spin-trapping Study of Proteins in Reverse Micelles Graham S. Timmins, Michael J. Davies and Bruce C. Gilbert* Department of Chemistry, University of York, York, UK YO1 5DD Horia Caldararu* Romanian Academy, Institute of Physical Chemistry, Splaiul lndependentei 202,77208 Bucharest, Romania EPR spectroscopy has been used to study the motions of several spin-labelled and spin-trapped proteins (a-chymotrypsin, cytochrome c and myoglobin) enclosed within reverse micelles formed by sodium bis-(2-ethyl- hexyl) sulfosuccinate (AOT) in isooctane. In several cases the spectra obtained from the encapsulated protein are significantly different from those observed in bulk solution. The motions of the labelled proteins, inferred from the anisotropy of the EPR spectra (A values), vary with the amount of solubilized water (W,) and hence the physical size of the water-pool in the reverse micelles; it is suggested that the level of solvation and the solvent structure in the water-pool of the reverse micelle cause the observed changes in motion.These results support the previously postulated ' water-shell ' model of proteins contained in reverse micelles. The study of proteins, especially enzymes, encapsulated within reverse micelles (RM) has attracted increasing atten- tion, because of their use as models of biological membranes and their potential applications in biotechnology. *-' Despite this work, there remains considerable doubt as to how the RM's structural and chemical properties affect, and are affected by, intra-micellar proteins, and how incorporation within RMs may cause changes in enzyme stability, kinetics and substrate specificity.For example, it would be of particu- lar interest to determine whether the water present inside the RM (the water-pool) possesses different physico-chemical parameters (such as viscosity and relative permittivity) from bulk water, and if so, whether these changes may modify protein dynamics, conformation and activity. These questions have been addressed in several theoretical models of protein- containing RMs.'.' A variety of techniques have been used to study proteins in RMs including ultracentrifugation, circular dichroism, fluo- rescence and NMR spectroscopy, dynamic light scattering and small-angle X-ray and neutron ~cattering.'~~~'EPR spectroscopy, although extensively used for studies of protein dynamics in bulk solution (because of the sensitivity of EPR spectra to alterations in molecular motion), has been scarcely utilised thus far to study proteins within RMs.' '-14 As a result of the development of specific spin-labels which can be attached to particular sites of a pr~tein,'~and our recent studies in which stable free radicals have been generated on proteins through radical-induced damage and subsequent spin-trapping,' we have examined whether the specificity of EPR spectroscopy for radical species and its sensitivity to their molecular motions can provide information on the nature of the water-pool and its influence upon proteins con- tained within RMs.Recent studies by Marzola and co-workers of spin-labelled chymotrypsin and human serum albumin encapsulated within RMs formed by sodium bis(2-ethylhexyl) sul-fosuccinate (AOT) in is~octane,'~.'~ showed that such an approach is realistic: changes in EPR spectral parameters provided information on the conformation and dynamics of proteins in bulk water and within RMs. The current study has aimed to characterize further the interactions between RMs and a number of proteins solvated in AOT-isooctane reverse micelles using both spin-labelled proteins (a-chy- motrypsin spin-labelled at two specific amino acid side-chain sites, Met-192 and Ser-195, located at the surface of the protein)" and spin-trapped proteins formed as a result of radical-induced damage' (with these generated either before incorporation into the RM or within the RM).Investigation of the motion of these proteins with attached free radicals in both bulk solution and within RMs with different sized water-pools, from small RMs (W, = 3) to large RMs (W, = 40),where the molar ratio W, = [H,O]/[AOT], has allowed us to study how changes in water-pool size, and subsequent solvation parameters, affect the molecular dynamics of the labelled proteins. Results and Discussion Spin-trapped Proteins Spin-trapping in Bulk Solution Reaction of HO' (generated by the Fe2+/H,02 couple) with cytochrome c in bulk water in the presence of the spin-trap 3,5-dibromo-4-nitrosobenzene-sulfonic acid (DBNBS) resulted in an EPR spectrum of a partially immobilized spin- adduct formed via the reaction of the hydroxyl radical with the protein and subsequent trapping of the protein radical by the spin-trap DBNBS, as described previously'6 [Fig. l(a)]; these species have been shown to be stable for several hours.16 The broad, anisotropic, nature of the observed spec- trum is consistent with that of a slowly tumbling, partially immobilized, radical-adduct with some local freedom of motion around the aminooxy radical centre.Information regarding the molecular motion of the observed species (and similarly for spin-labelled proteins as described below) may be readily obtained from the splitting between the outermost features of the spectrum (denoted as 2AII; this value is char- acteristic of the extent of immobilization of the species, with increasing motion tending to decrease A II until isotropic spectra, indicative of rapid protein tumbling and/or rapid molecular motion of the aminooxy group, are obtained).Incorporation of these pre-formed cytochrome c-DBNBS radical-adducts within RMs of varying W, (between 33 and 73, gave spectra similar to those obtained in bulk water [Fig. l(b)-(e)]. At higher W,, All could be easily measured, whereas at lower W, (< 12) this was not possible owing to the poor signal-to-noise ratios (the amount of trapped species cannot be readily increased in these cases, since its addition J. CHEM.SOC. FARADAY TRANS., 1994, VOL. 90 -Fig. 1 EPR spectra of the cytochrome c-DBNBS radical-adduct formed upon attack of cytochrome c by the hydroxyl radical: (a)in water (pH 3.5); (b)-(e) the adduct obtained in (a)injected into AOT reverse micelles of W, = 33, 15, 12, 7.5, respectively. The measure- ment of 2AII is indicated in (a). in solution to the AOT-isooctane mixture determines the W, value) and underlying baseline drift due to a component of the iron species. For micelles with W, = 33 and 15, the All value (2.80 mT) is somewhat smaller than that found in bulk water (2.91 mT). This unexpected decrease in All in this range suggests that incorporation of the spin-adduct into the reverse micelle results in an increase in the molecular motion of the aminooxy group.This may be due to protein confor- mational changes, resulting in increased local freedom of motion about the aminooxy group, thereby reducing the All. value: interaction between the surface lysine residues of the protein (which are positively charged and basic in nature) with the (negatively charged) head groups of the AOT mol- ecules in the RMs might be responsible for the postulated conformational change. This type of interaction is known to direct the binding of cytochrome c to the membrane-bound cytochrome c oxidase complex in mitochondria.’ Chymotrypsin-DBNBS radical-adducts produced in a similar manner in bulk water gave rise to EPR spectra [Fig. 2(a)] that appear to contain at least two components, one of which is partially immobilised and similar to that obtained with cytochrome c (with All 73.03 mT) and one which gives rise to a much more isotropic spectrum. This suggests that there are several sites of radical formation upon the protein and hence several possible sites of spin-adduct formation on the chymotrypsin molecule, with differences in the nature of these species and hence local aminooxy radical mobilities.This is not unexpected since the HO’ radicals which generate the initial species are highly reactive, and will be generated across the protein surface in an approximately random manner (the iron-EDTA complex has a stability constant” -\\ 7-‘-1 1 mT-Fig. 2 EPR spectra of the chymotrypsin-DBNBS radical-adduct formed upon attack of chymotrypsin by the hydroxyl radical: (a) in water (pH 6); (b)-(e) the adduct obtained in (a) injected into AOT reverse micelles of W, = 40,20, 6, 3, respectively.The measurement of 2AI,is indicated in (a). of 1.6 x 1014 and hence this complex is free in solution and the iron is not bound to specific sites on the protein); differ- ences in the surface ‘topography’ could therefore result in differing freedom of motions of the DBNBS adduct. Incorporation of these radical-adducts into AOT RMs of varying W, resulted in the observation of the spectra shown in Fig. 2(b)-(e). As observed with cytochrome c, the value of All for the species giving the more anisotropic spectrum is smaller (2.87 mT) in large water-pools (W, = 40 and 20) than the value in bulk water (3.03 mT); however, on incorporation into RMs with smaller water-pools (W, = 10 and 6) an increase in All to 2.91 mT and 3.02 mT, respectively, is observed, though these values do not exceed those obtained in bulk solution.The reason for the decrease in A and hence the increase in mobility, upon incorporation into RMs of larger W, could again involve interactions of basic surface residues of the protein with AOT head groups resulting in conformational changes in the protein as postulated for cytochrome c, although it is known that cytochrome c and chymotrypsin appear differentially associated with the water-pool inter- face.” The increase in All with decreasing size of the water- pool (W,) can probably be assigned to an increase of the microviscosity of the water in its core (for a fuller discussion, see later); alternatively, conformational changes of the protein structure might also be brought about by the size of the water-pool within the RM, with the protein being ‘squeezed’ or altered in conformation at lower W, values. J.CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 Spin-trapping within Reverse Micelles Attempts to trap protein-derived radicals from cytochrome c or chymotrypsin subjected to HO' attack in the presence of DBNBS within the water-pool of the RM, resulted in the detection of isotropic spectra from two radical species which were also observed in the absence of the protein. This sug- gests that HO' preferentially attacks the hydrocarbon chains of the surfactant, producing two adduct species, one of which gives rise to a triplet spectrum (with aN= 1.36 mT), due to the trapping of a tertiary radical, and the other to a triplet of doublets (aN= 1.40 mT; aH = 0.80 mT), due to the trapping of a secondary radical; attempts to form protein radical-adducts preferentially with either protein by manipulation of the reagent concentrations failed.The presence of either protein in this system did, however, result in an increase in the rate of decay of the adduct species, implying that the adduct may interact with the protein, possibly through one- electron oxidation or reduction of the radical centre by protein-bound species (which could either be thiol groups or reactive species produced on the proteins ; radical-damaged proteins are known to contain such species2').To minimise radical attack on the surfactant chains by HO', myoglobin (a haem-containing protein) encapsulated within RMs was reacted with H20, and DBNBS in the absence of added iron; the observed spin-adduct arises from the trapping of a radical species on the protein resulting from the reaction of H,02 at the haem centre followed by an intramolecular electron-transfer reaction, without the forma- tion of H0'.'6 Fig. 3 shows the EPR spectra of myoglobin- DBNBS adducts, generated in this manner, in RMs of varying W,. The spectral shape of these adducts formed in RMs of W, = 40 and the All value (2.64 mT) are almost iden- \\ i I tical to those found in bulk water (2.61 mT).I6 However, at lower W, values an additional isotropic species was obtained ; this could be due to either the interaction of the protein with the AOT head groups, which may result in alteration of the protein structure, as observed with a-chymotrypsin and cyto- chrome c, or transfer of the damage from the protein to the AOT molecules and subsequent trapping of these radicals. No variation in the All of the anisotropic species with changes in W, was observed for myoglobin (in contrast to a-chymotrypsin and cytochrome c); this is believed to be due to the position of the spin-adduct on the protein with, in this case, the aminooxy group buried sufficiently deeply within the protein structure to be unaffected by changes in the local freedom at the protein surface owing to differences in water- pool size, except when the protein surface itself may be dis- rupted at very low W, values.Spin-labelled Proteins Reaction of a-chymotrypsin with the stable aminooxy radical 1 in bulk solution followed by purification, as described in the Experimental section, resulted in the formation of the spin-labelled protein with the spin-label attached at the Met-192 residue. The EPR spectrum of this material in bulk water, which is similar to that previously reported2' (see Fig. 4), corresponds to the region of fast motion and is indicative of moderately rapid rotation of the spin-label, which being at an exposed surface site, is only slightly motionally restricted by the protein structure.21*22 Incorporation of this labelled protein into RMs of varying W, resulted in alterations in the observed spectra [Fig.4(a)-(c)].Compared with the spectra in bulk water, these suggest increasing restriction of motion Fig. 3 EPR spectra of equine myoglobin-DBNBS radical-adducts formed by reaction with H202 in the absence of Fe": (a)in a blank experiment; (b)-(d) within the AOT reverse micelles of W, = 40, 15 and 10, respectively. Experimental settings were the same for all Fig. 4 EPR spectra of a-chymotrypsin spin-labelled with 1 in: (a)-samples except field modulation width in (b)and (b')is 0.32 mT and (c)AOT reverse micelles of W, = 3, 6 and 20, respectively; (d)buffer, 0.5 mT, respectively. pH 6.Experimental settings were identical for all samples. 2646 with decreasing W, , indicated by their increased linewidth (and hence greater z,). At high W, values (> 20) the linewidth remains constant never reaching the breadth observed in bulk solution. 0 II 0 F--P--OCH&H3II I H3C, ,N+-(CH.&CH3 Br-N H CC H*Br 0 I I H3C I000I I I 0-0-0-1 2 3 Spin-labelling the same protein with a second stable aminooxy radical 2, which is known to react at the nearby Ser-195 residue, resulted in a partially immobilized species in bulk solution, though in this case a weak isotropic signal is also observed in the EPR spectrum; this signal is believed to be due to the free spin-label formed by hydrolysis of the phosphate ester linking the spin-label to the protein, as observed previo~sly,~~.~~ and in accord with this suggestion its signal intensity increased with time after preparation of the labelled protein (data not shown).Previous crystallo-graphic and other studies have shown that the Ser-195 residue (to which the aminooxy radical is attached) lies in the active site of the enzyme in a shallow hydrophobic depression on the surface of the pr~tein;~~*~~ the partial immobilization of the aminooxy radical on binding to the protein, as indi- cated by its EPR spectrum, is believed to be due to its accom- modation within this po~ket.'~,'~Incorporation of this second labelled protein (immediately after its preparation) into RMs of varying W, gave the spectra shown in Fig.5(a)-(d); these spectra indicate varying degrees of mobility of the aminooxy radical, with lower W, micelles resulting in greater immobilization. In the case of RMs of W, b 10, an additional isotropic component was observed; since an increase in W, resulted in a greater proportion of this isotropic species, and the proportion of this species at fixed W, (except W, = 3) increased with time, this isotropic species is again attributed to the hydrolysed spin-label. The All values observed for this labelled protein in RMs of varying W, are summarised in Fig. 6; note that no further decrease in All was observed with W, > 20, and that the observed All value (2.91 mT) for these micelles indicates a greater degree of spin-label immobil- ization than in bulk water.The observed dependence of the spectral parameters on W, is most likely to derive from changes in the microviscosity of the solvent surrounding the protein (which alters the rate of local molecular motion of the spin label) or changes in polarity (which can change the nitrogen hyperfine splitting).? Studies of the microviscosity and polarity within RMs identi- cal to those used in the present study (but in the absence of protein) with the charged spin probe 3 give rotational corre- lation times (z,) and nitrogen hyperfine splittings (aN) of 7 Contributions to changes in spectral parameters from overall micellar and/or protein tumbling may be ruled out because: (i) non- protein-containing RMs of W, 2 5 have a hydrodynamic radius of 30 A'** this being above the threshold value of 20 A above which the contribution to anisotropicity by the rate of micellar tumbling is outside the range detectable by X-band EPR measurement^;^^*^^ incorporation of proteins will increase the RM size yet further.(ii) The differences in EPR parameters of chymotrypsin in bulk water labelled with 1 and 2 indicate that it is local spin-label motions, and not overall protein tumbling, that are significant in determining spectra in this system. J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 7 1, ,' I Fig. 5 EPR spectra of a-chymotrypsin spin-labelled with 2 in AOT reverse micelles: (a) and (a') Wo= 3, (6) and (6') Wo= 6, (c) and (c') Wo= 10, (d)and (d') Wo = 20.Experimental settings were the same for all samples except field modulation width in (a) 0.4 mT, (a') 0.63 mT, (b) 0.125 mT, (6') 0.32 mT, (c) 0.125 mT, (c') 0.4 mT, (d) 0.125 mT, (d')0.32 mT. 0.7 x lo-'' s and 1.67 mT, respectively, for bulk water, and 15.5 x lo-'' s and 1.51 mT, respectively, for RMs with W, = 3.5*27 Although this positively charged aminooxy radical may be anchored at the AOT/water interface (by elec- trostatic interaction), the differences in z, values indicate that there are differences in the microviscosity of the various water-pools and also of bulk water. When the dependence of z, of 3 and AI, of the Ser-195 labelled a-chymotrypsin are plotted against W, (Fig.6) it can be seen that the two are quite similar, with smaller z, values found at higher W, ;since changes in z, have been associated with the microviscosity of water within the RM,25327 a similar explanation would seem 3.3 17 3.2 I-3.1 --T 3.0 -0 10 20 30 40 wo Fig. 6 Variation of A,, of cr-chymotrypsin spin-labelled with 2 (0) and z, of spin probe 3 (m) in AOT reverse micelles as a function of WO J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 appropriate for the observed changes in All for the spin- labelled proteins in this study. These data are in accord with those obtained by Marzola et ~1.'~with spin-labelled human serum albumin, where a weakly immobilized species (believed to be due to attachment to a surface lysine residue28) and a more strongly immobilized species underwent similar changes in mobility in RMs when W, was changed from 2.2 to 7.2; this behaviour has been attributed to an increased depth of solvating water molecules around the protein as W, increases.In order to obtain further evidence to support the sugges- tion that it is the microviscosity within RMs which modu- lates the motion of the spin-label with the labelled proteins, their EPR spectra were recorded in buffer solutions contain- ing varying amounts of glycerol to alter their viscosity (from 1 :1 to 1 :4 w/w, whose relative viscosity values range from ca. 6 to 60, re~pectively).~~ The spectra observed with these mixtures (Fig. 7) are similar to those observed with RMs of varying W, (cf:Fig.4and 5), with increasing amounts of glyc- erol giving increasingly anisotropic spectra (with larger A II values). Attempts to use the EPR spin-probe technique to detect any structural changes in the RMs themselves brought about by encapsulation of the proteins failed (using spin-probe 3 or 5-doxylstearic acid [2-(3-carboxypropyl)-4,4-dimethyl-2-tridecyl-3-oxazolidin-l-yloxyl]},filled RMs only represent a low proportion of the total RMs (owing to the limited solu- bility of the proteins in the added water). Thus it was not possible to distinguish between RMs that did or did not contain protein;30 similar results have been obtained with other method^.^^'^,^' Conclusions Two models for the solubilization of proteins within RMs have been developed; the 'water-shell' model in which it is proposed that protein solubilization gives rise to larger Fig.7 EPR spectra of buffer solution containing a-chymotrypsin spin-labelled with 1 in buffer-glycerol mixtures (w/w): (a) 1:4, (b) 1 :1 and similarly, a-chymotrypsin spin-labelled with 2 (c) 1:3 and (d) 1 :1. Experimental settings were the same for all samples. ~~~1-3,32,33and an alternative model which suggests that no increase in RM size occurs.5,6*34,35 In the water-shell model, which is believed to be particularly suited to hydro- philic proteins and the one-protein-per micelle proposal,'Y2 it is supposed that the protein is surrounded by a shell of water with.in the RM.The dependence of the aminooxy radical motion for the spin-labelled a-chymotrypsin on the microviscosity/polarity and hence W, (at least for W, d 15-20) would seem to support the water-shell model in this case. This hypothesis is also in agreement with the observed dependence of the rate of hydrolysis of a-chymotrypsin labelled with 2 upon W,, with the very slow rate of hydro- lysis in RMs with W, = 3 indicating that at these values most of the water molecules are essentially bound to the protein and/or inner-RM surfaces. It has also been shown that the enzymic stability of a-chymotrypsin in AOT RMs is enhanced at low W, ,presumably for similar reasons.36 The decrease in the microviscosity of the water-pool as W, is increased is believed to result from the increased depth of the solvating water layer at higher W, causing a decrease in the short-range order of the water layer induced by its inter- actions (e.g.electrostatic) with the protein and AOT surfaces.For RMs with W, 2 20, where the protein radius (ca. 22 A) is much smaller than the radii of the water-pools (between 35 and 60 A, for W, = 20 and 40, respectively), the lack of observed changes in mobility of the aminooxy group attached to a-chymotrypsin for both labels with increasing W, probably results from little (or no) change in the micro- viscosity of the water-pool with water layer thickness, i.e. there is a threshold value above which little change occurs; it has previously been reported that little change in the struc- ture of the micelles occurs over this threshold W, 8.37.38 Experimental All chemicals, which were of the highest commercially avail- able quality, were obtained from either Sigma or Aldrich, and were used without further purification ; deionized water was used in the preparation of all solutions.The AOT-isooctane (<0.005% water) solution consisted of 0.1 mol dm-3 AOT in isooctane (W, = 0). Reverse micelles of varying W, values, containing spin-labelled or spin-trapped proteins were obtained by the injection method and gently stirred until completely transparent.',2 All sample flasks and EPR sample tubes were sealed to prevent changes in W, caused by evapo- ration; aqueous solutions were made up in 0.01 mol dm-3 ammonium acetate buffer, pH 6, unless otherwise stated.All the concentrations of added materials refer to those in the water-pool of the RM only, and are final concentrations. a-Chymotrypsin was spin-labelled at the Met-192 residue with 4-(2-bromoacetamido)-2,2,6,6-tetramethylpiperidin-l-y1-oxyl 1.' Separation of unreacted spin-label from labelled protein (total volume <2 cm3) was achieved by chromatog- raphy on Sephadex G-25 columns (100 mm x 10 mm) in ammonium acetate buffer and the protein fraction re-chromatographed (usually twice) until it contained no free spin-label, as ascertained by the absence of an isotropic EPR signal; the samples were then frozen in aliquots at -20°C until used. Protein concentrations were determined by the Biuret method using a kit obtained from Sigma.The Ser-195 residue of a-chymotrypsin was spin-labelled with 4-(ethoxy- fluorophosphinyloxy) -2,2,6,6- tetramethylpiperidin- 1 -yl -oxyl 2," purification and storage was as described above. The spin-labelled protein solution was subsequently injected in AOT-isooctane (W, = 0) solution to give RMs of differing W, values. Final protein concentrations within the water-pool of the RMs were between 5 x lo6 and 7 x lo-' mol dm-3. 2648 J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 Spin-trapped proteins, produced by radical damage, 13 P. Marzola, C. Pinzino and C. A. Veracini, Langmuir, 1991, 7, were either produced by reaction (at 293 K) of the protein (cytochrome c, 0.5 mmol dmP3; cc-chymotrypsin 1 mmol dm -3, with equimolar concentrations of FeSO,-thylene- diaminetetraacetic acid (EDTA) and H20, in the presence of the spin-trap 3,5-dibromo-4-nitrosobenzene sulfonic acid (2 mmol dm-3; DBNBS) in bulk solution; or they were pro- 14 15 16 238.P. Marzola, C. Forte, C. Pinzino and C. A. Veracini, FEBS Lett., 1991,289, 29. J. D. Morrisett, in Spin Labelling, Theory and Applications, ed. L. J. Berliner, Academic Press, New York, 1976, p. 274. M. J. Davies, B. C. Gilbert and R. M. Haywood, Free Radical Res. Commun., 1991, 15, 111. duced within RMs by injection of all components (as above) into the AOT-isooctane mixture with the H202 added last. In the case of radical-adducts produced from myoglobin (0.8 mmol dm-3) a similar procedure was adopted except that the FeSO, was omitted.EPR spectra were recorded at 291 K on a JEOL JES-RElX spectrometer with 100 kHz field modulation using X-band frequencies. For the reverse micelle experiments round silica sample tubes (internal diameter 4 mm) were used, whilst for aqueous samples, standard aqueous flat-cells were employed. Hyperfine coupling constants and A values were 17 18 19 20 21 22 V. M. Darley-Usmar, R. A. Capaldi, S. Takamiya, F. Millet, M. T. Wilson, F. Malatesta and P. Sarti, in Mitochondria, A Practi- cal Approach, ed. V. M. Darley-Usmar, D. Rickwood and M. T. Wilson, IRL Press, Oxford, 1987, p. 113. L. G. Sillen, E. Hogfeldt, A. E. Martell and R. M. Smith, in Stability Constants Supplement No. I, Special Publication No. 25, The Chemical Society, London, 1971.M. P. Pileni, T. Zemb and C. Petit, Chem. Phys. Lett., 1985, 118, 414. R. T. Dean, S. Gieseg and M. J. Davies, TZBS, 1993, 18,437. D. J. Kosman, J. Mol. Biol., 1972,67, 247. B. P. Siegler, D. M. Blow, B. W. Matthews and R. Henderson, J. Mol. Bid, 1968, 35, 143. measured by comparison with Fremy’s salt (a, 1.30mT). 23 J. D. Morrisett and C. A. Broomfield, J. Am. Chem. Soc., 1971, 93, 7297. We are grateful to the EC for supporting this work by the award of a fellowship to H.C., the SERC and AICR for the provision of EPR spectrometers and the YCRC for fellow- 24 25 J. D. Morrisett and C. A. Broomfield, J. Biol. Chem., 1972, 247, 7224. G. Haering, P. L. Luisi and H. Hauser, J. Phys. Chem., 1988, 92, 3574. ships to G.S.T. and M.J.D. 26 D.D. Lasic and H. Hauser, J. Phys. Chem., 1985,89,2648. 27 H. Caldararu, A. Caragheorgheopol, M. Dimonie, D. Donescu, I. Dragutan and N. Marinescu, J. Phys. Chem., 1992, %, 7109. References 28 C. N. Cornell, R. Chang and L. J. Kaplan, Arch. Biochem. 1 2 3 4 5 6 7 P. L. Luisi and L. J. Magid, CRC Crit. Rev. Biochem., 1986, 20, 409. P. L. Luisi, M. Giomini, M. P. Pileni and B. H. Robinson, Biochim. Biophys. Acta, 1988,947,209. P. L. Luisi, Angew. Chem., Znt. Ed. Engl., 1985,24,439. N. Pfammatter, M. Famiglietti, A. Hochkoppler and P. L. Luisi, in Biomolecules in Organic Solvents, ed. A. Gomez Puyou, CRC Press, London, 1992, p. 26 1. Y. L. Khmelnitski, A. V. Kabanov, N. L. Klyachko, A. V. Leva- shov and K. Martinek, in Structure and Reactivity in Reversed Micelles, ed.M. P. Pileni, Elsevier, Amsterdam, 1989, p. 230. K. Martinek, A. V. Levashov, N. Klyachko, Y. L. Khmelnitski and Y. V. Berezin, Eur. J. Biochem., 1986, 155,453. G. G. Ramos, D. A. Fernandez-Velasco, M. T. de Gomez-Puyou and A. Gomez-Puyuo, in Biomolecules in Organic Solvents, ed. A. Gomez-Puyou, CRC Press, London, 1992, p. 113. 29 30 31 32 33 34 35 Biophys., 1981, 209, 1. A. V. Wolf, M. G. Brown and P. G. Prentiss, in CRC Handbook of Chemistry and Physics, ed. R. C. Weast, CRC Press, Florida, 63rd edn., 1982, p. D239. H. Caldararu, M. J. Davies, B. C. Gilbert and G. S. Timmins, unpublished results. B. H. Robinson, C. Toprakcioglu and J. C. Dore, J. Chem. Soc., Faraday Trans. I, 1984,80, 13. F. J. Bonner, R. Wolf and P. L. Luisi, J. Solid Phase Biochem., 1980,5, 255. C. Grandi, R. E. Smith and P. L. Luisi, J. Bid. Chem., 1981,256, 237. Y. L. Khmelnitski, A. V. Levashov, N. L. Klyachko, V. Y. Chernyak and K. Martinek, Biokhimiya, 1982,47,86. A. V. Levashov, Y. L. Khmelnitski, N. L. Klyachko, V. Y. Chernyak and K. Martinek, J. Colloid Interface Sci., 1982, 88, 444. 8 9 10 11 B. M. D. Verhaert, R. Hilhorst, A. J. W. G. Visser and C. Veeger, in Biomolecules in Organic Solvents, ed. A. Gomez-Puyou, CRC Press, London, 1992, p. 133. C. Vos, C. Laane and A. J. W. G. Visser, Photochem. Photobiol., 1987,45, 863. M. P. Pileni, in Structure and Reactivity in Reversed Micelles, ed. M. P. Pileni, Elsevier, Amsterdam, 1989, p. 44. 0. V. Belonogova, G. I. Likhtenstein, A. V. Levashov, Y. L. 36 37 38 S. Barbaric and P. L. Luisi, J. Am. Chem. SOC., 1981,103,4239. P. D. I. Fletcher, A. M. Howe, J. C. Dore, B. H. Robinson, N. M. Perrins and C. Toprakcioglu, in Surfactants in Solution, ed. K. L. Mittal and B. Lindman, Plenum Press, New York, 1985, vol. 3, p. 1745. P. D. I. Fletcher, B. H. Robinson and J. Tabony, J. Chem. SOC., Faraday Trans. I, 1986,82,2311. Kmelnitski, N. L. Klyachko and K. Martinek, Biokhimiya, 1983, Paper 4/01779J; Received 24th March, 1994 48,379. 12 L. V. Belovolova, R. M. Davydov and V. P. Timopheyev, Zh. Fiz. Khim., 1983,57, 2765.