Proc. Natl. Acad. Sci. USA Vol. 73, No. 10, pp. 3558-3561, October 1976 Biophysics
Erythrocyte membranes undergo cooperative, pH-sensitive state transitions in the physiological temperature range: Evidence from Raman spectroscopy (protein-lipid boundary/amino acid spectra/methyl and methylene vibrations)
SURENDRA P. VERMA AND DONALD F. H. WALLACH* Tufts-New England Medical Center, Therapeutic Radiology Department, Radiobiology Division, 171 Harrison Avenue, Boston, Massachusetts 02111
Communicated by J. L. Oncley, August 6, 1976
lecithin-melittin complexes (24), and probably for the sharp shift in the spectra of thymocyte plasma membranes at +230 (13) as well as the two separate scattering discontinuities of erythrocyte ghosts, below 00 and at +170 (17). It appears, however, that the protein-lipid interface might give rise to the pH-sensitive, partially reversible thermotropic transition reported by paramagnetic quenching of tryptophan fluorescence in erythrocyte membranes at high physiological temperatures (25). We now show how laser-Raman spectroscopy can further elucidate the responses of erythrocyte membranes to variations of temperature and pH in physiological ranges.
ABSTRACT We have examined the Raman scattering from erythrocyte ghosts at 2700 to 3000 cm-l (CH-stretching region). Plots of the intensity (I) of the 2930 cm-1 band relative to the intensity of the thermally stable 2850 cm-' band, i.e., the [I2930/I2850] ratio, as a function of temperature reveal a sharp discontinuity, which at pH 7.4 has a lower limit of 380 and is irreversible above 420. [I2930/I25so1 is stable between pH 7.0 and pH 7.4, but increases or decreases sharply below pH 7.0 or above pH 7.5, respectively. Reduction of pH to 6.5 lowers the transition temperature by about 16°, and a shift to pH 6.0 drops the transition range to 0 to 70. The above effects of temperature and pH on Raman scattering closely correspond to those detected by studies on the interaction of membrane protein fluorophores and lipid-soluble fluorescence quenchers [Bieri, V. & Wallach, D. F. H. (1975) Biochim. Biophys. Acta 406, 415-4231. Taken together, these results suggest that the transitions represent concerted processes, involving hydrophobic amino acid residues and lipid chains at apolar protein-lipid boundaries. In 1967, Changeux et al. (1) presented a model for biomembranes in which ordered arrays of protein-lipid subunits, protomers, form "two-dimensional crystalline lattices." The model focuses on cooperative phase transitions in such systems. It "explicitly assumes that the transition of the protomer is highly cooperative with respect to its constitutive amino acid or phospho-lipid elements," with the unique aspects of membrane cooperativity deriving from state changes involving the "statistical-mechanical ensemble of protomers." That biomembranes may contain extended, two-dimensional lipid-protein lattices is evident from recent experimentation on hepatic "gap-junctions" (2) and the purple membranes of Halobacterium halobium (3). That membrane phospholipids can undergo thermally or ionically induced, cooperative phase transitions is also well established (see review in ref. 4), as is the fact that the function of many membrane proteins may vary profoundly with the state of membrane lipid (see review in ref. 5). Finally, that membrane proteins can influence the structure and composition of their apolar environment is now well documented (e.g., refs. 6-9). However, the dynamics of lipidprotein interfaces in biomembranes have proven elusive to date. We have used Raman spectroscopy to explore the interfaces between hydrocarbon residues of membrane proteins and lipids. Aqueous dispersions both of biomembrane (10-17) and of phospholipids (18-25) yield well-resolved Raman scattering spectra, whose CH-stretching regions are prominent and undergo marked band modifications at certain, defined temperatures (13, 17, 20-24). With pure lipids, the thermotropic discontinuities in Raman scattering can be identified as order >=± disorder transitions of hydrocarbon chains. This is almost certainly true also for the thermally shifted scattering steps of *
EXPERIMENTAL We used reagents of analytical grade and solvents of spectrochemical purity. Egg lecithin and other phosphatides were purchased from Lipid Products (South Nutfield, Great Britain). Erythrocyte ghosts were prepared (11, 17), suspended in 5 mM phosphate at stated pH to a concentration of 10 mg of protein per ml, transferred to thermoregulated capillaries, and their spectra recorded as in refs. 11, 13, 15, and 17 with a Ramalog 4 spectrometer (Spex Industries, Metuchen, N.J.) interfaced to an Interdata computer (model 70), using the 488 nm Ar+ laser line for excitation at 400,um slit width. Temperature equilibration was, as in ref. 17, for 20 min with the laser beam occluded and 5 min with the beam impinging on the sample, before scanning from 2750 cm-' to 3050 cm-1 (2-4 scans; not more than 300 s/scan). The averaged, smoothed (least-squares; ref. 17) spectra were plotted for each temperature. The data points or ranges presented derive from at least four separate membrane preparations.
RESULTS Variation of the CH-stretching region with temperature above 220 At 220 the CH-stretching region is characterized by three strong bands that lie at -2930 cm-', -2880 cm-', and 2850 cm-' (Fig. 1). As the temperature is raised from -220 to above 420, at pH 7.4, the -2930 cm-' band becomes very prominent, relative to the temperature-insensitive feature at 2850 cm-, and the -2880 cm-1 peak becomes obscured. A plot of the relevant intensity ratio, [I293/I2o/o], versus temperature is given in Fig. 2 for pH 7.4. [I2930/I28501 stays stable at -1.2 + 0.02 between set temperatures 200 and 350 and then shifts abruptly to -1.41 between set temperatures 360 and 450. This transition is reversible only up to a set temperature of 390. When we correct for the local heating due to the laser beam, we obtain a lower temperature limit of 380 for the discontinuity and an
To whom correspondence should be addressed.
3558
Biophysics:
Verma and Waflach
0
Proc. Natl. Acad. Sci. USA 73 (1976)
0 CY)
CY) 0) (%4
3559
1 .5
1.4
6.0 420C
1.3
1.2
pH 7.0
.22°C
20
30 40 TEMPERATURE (0C)
50
FIG. 2. Variation of [I2930/I2s,501 with temperature. Points were obtained with three separate membrane preparations. Conditions were as for data in Fig. 1.
Effect of pH on high-temperature membrane-
thermotropism jor pH 8.0 FIG. 1. Laser-Raman spectra, in the CH-stretching region, of erythrocyte ghosts at two temperatures (pH 7.4) and three pH values (220, set temperature). Membrane concentration was 10 mg of protein per ml. Excitation wavelength was 488 nm; slit width, 400 Am. Power at sample was 400 mW. The traces represent spectra from single membrane preparations. The variation between samples is given in Figs. 2-4 for [IV293/I2io, but also applies to the features between 2880 and 2900 cm-'.
irreversibility-temperature of 420.t The changes in [Ibso/I2&5o] that accompany the gel ± liquid-crystal transitions of egg-, dimyristoyl-, and dipalmitoyl-lecithins are 0.25,0.20, and 0.24, respectively. With lecithin:cholesterol (1:1, mol:mol), the change in ratio does not exceed 0.1. Variation of the CH-stretching region with pH at constant temperature
Fig. 1 shows the CH-stretching region at pH 6.0, 7.0, and 8.0 (220), while Fig. 3 gives a plot of [Iwao/I2o] versus pH. The spectra do not change significantly in the pH range 7.0-7.5, where [I2930/12bso] remains constant at 1.20 + 0.02. Between pH 7.5 and 8.5 [I29ao/I2&5o] drops discontinuously to about 1 and the CH-stretching region develops a pattern seen at # crystal transitions arise from altered interactions, amplified by Fermi resonance, between CH-stretching fundamentals and HCHdeformation overtones. Experimentation using differential thermal calorimetry and enzyme inactivation (29) suggests that erythrocyte membrane proteins undergo a thermotropic state transition with a lower limit near 40°. In addition, proton magnetic resonance spectroscopy (30) shows that methyl side chains of erythrocyte membrane proteins become more mobile as the temperature is raised above about 350. Our own studies (25) on the paramagnetic quenching of membrane tryptophan fluorescence by nitroxide analogs of stearic acid have focused on lipid-protein interactions at physiological and near physiological temperatures and pH values, rather than on overall protein behavior as studied in refs. 29 and 30. Our experiments indicate that the accessibility of the protein fluorophores to the lipid probes increases abruptly above 370, in a process that is reversible below 420, and shifts to lower temperatures upon pH reduction, concurrently diminishing in intensity. The data also document that the accessibility of the fluorophores is minimal at pH 7.0-7.2, increasing sharply at lower and higher pH. The strong correlation of the paramagnetic quenching results with the Raman data presented herein indicates that the two approaches describe different aspects of the same process. Indeed, a quantitative analysis of the Raman data suggests that the pH-sensitive, thermotropic change in the CH-stretching region represents a cooperative change of state, involving apolar amino acid residues, as well as lipid acyl chains: First, 60% of the CH3 groups in erythrocyte membranes are those of amino acid residues (27), which produce strong scattering near 2930
cm-1 (Table 1). Second, erythrocyte membranes give a step of 0.16 in [I2930/I2850] at -8° (17) and another step of not less than 0.22 above 380, adding up to 0.38, whereas the maximum shift we observe with the gel liquid-crystal transitions of diverse phosphatides is about 0.25. Barring a highly divergent behavior by membrane phosphatides, the magnitude of the overall change of [I2930/I285o1 from -200 to +50° implies a contribution by protein CH3 groups. This contribution is assigned to the high-temperature step reported here because (a) this is pH sensitive and only partially reversible; (b) other data (29, 30) also reveal no protein thermotropism in erythrocyte membranes below 350; and (c) fluorescence quenching (25) reveals clear modification of protein-lipid interaction above 350t That the magnitude of the low-temperature step is only 0.16 (17), i.e., less than observed with phospholipids undergoing gel i=± liquid-crystal transitions, suggests that the high-temperature step includes a lipid contribution. This proposal is consistent with paramagnetic quenching data (25), which shows increased accessibility of membrane protein fluorophores to lipoidal probes. The pH susceptibility of the thermosensitive [I2Wso/I2&so] step may relate to the reversible thermotropic transitions observed in some soluble globular proteins at pH values below their isoelectric points (31). However, the large response in the pH range 7.0-6.0 is unusual. One explanation for this might be that, because of the low dielectric constant of the membrane core, small changes in charge might influence the structure of proteins in membranes much more than that of proteins dissolved in aqueous solvents. The same argument applies to the pH sensitivity of [I2930/I2s50], or paramagnetic fluorescence quenching (25), at constant temperature. Taken together, present and previous (25, 29, 30) data lead us to hypothesize that: (a) The apolar residues of penetrating proteins, together with the hydrocarbon chains of phospholipids immediately apposed to these protein residues, constitute a separate phase. (b) The high-temperature, pH-sensitive thermotropic step in [I2930/I285o] represents a cooperative state transition in this phase, involving a concerted rearrangement of acyl-chain configuration and protein residue orientation. Our concept of a separate protein-lipid phase capable of a cooperative change in state takes into account recent nuclear magnetic resonance data documenting the mobility of apolar residues in globular proteins (reviewed in ref. 32) and in -
t Cholesterol CH3 groups are unlikely to be involved because of the weak CH-stretching signals of this sterol, its low CH3 contributions, and its tendency to reduce the magnitude of thermotropic shifts in the CH-stretching region (cf. ref. 24).
Biophysics:
Proc. Natl. Acad. Sci. USA 73 (1976)
Verma and Wallach
3561
Table 1. Raman frequencies in the CH-stretching region of amino acids bearing CH3 residues Alanine
3003 (9.5) 2982(6.5) 2960 (5.7) 2952 (10) 2930 (8.3)
2885 (3.6)
Valine
Leucine
Isoleucine
Methionine
2985 (8.9) 2975 (5.4) 2960 (10)
2985 (6.1)
2987 (4.6)
2925 (4.6)
2967 (10) 2958 (10) 2932 (10)
2960 (10) 2948 (7.9) 2930 (5.6)
2888 (1.9) 2979 (2.3) 2970 (1.9)
2895 (8.6) 2882 (6.5) 2878 (5.8)
2901 (10) 2896 (9.8) 2868 (8.0)
2882 (8.8)
2935 (1.7) 2925 (6.0) 2909 (10)
Threonine 3018 (7.9) 2992 (10) 2973 (8.3)
2936 (9.8)
2873 (6.5) 2870 (8.9)
Frequencies are in cm-1. Spectra were obtained on crystalline solids with 50 Am slits. Figures in parentheses give relative intensities.
erythrocyte membranes at temperatures above approximately 400. Our proposal is also in full accord with the model proposed by Changeux et al. (1) and Williams' treatment of lipid-protein interactions (32). This work was supported by Grant CA 13061 from the USPHS and PRA78 from the American Cancer Society to (D.F.H.W.). We thank Dr. R. Mikkelsen for his constructive criticism. 1. Changeux, J.-P., Thiery, J., Tung, Y. & Kittel, C. (1967) Proc. Natl. Acad. Sci. USA 57,335-341. 2. Goodenough, D. A. & Stoeckenius, W. (1972) J. Cell Biol. 54, 646-656. 3. Henderson, R. & Unwin, P. N. T. (1975) Nature 257,28-32. 4. Wallach, D. F. H. (1975) Membrane Molecular Biology of Neoplastic Cells (Elsevier Publishing Co., Amsterdam), pp. 147-153, 188-189, 218, 220. 5. Raison, J. K. (1973) J. Bioenerg. 4, 285-309. 6. Jost, P. C., Griffith, 0. H., Capaldi R. A. & Vanderkooi, G. (1973) Proc. Nati. Acad. Sci. USA 70,480-484. 7. Stier, A. & Sackmann, E. (1973) Biochim. Biophys. Acta 311, 400-408. 8. Wallach, D. F. H., Verma, S. P., Weidekamm, E. & Bieri, V. (1974) Biochim. Biophys. Acta 356,68-81. 9. Warren, G. B., Houslay, M. D., Metcalfe, J. C. & Birdsall, N. J. M. (1975) Nature 255,684-687. 10. Larsson, K. & Rand, P. (1973) Biochim. Biophys. Acta 326, 245-255. 11. Wallach, D. F. H. & Verma, S. P. (1975) Biochim. Biophys. Acta 382,542-551. 12. Lippert, J. L., Gorzyca, L. E. & Meiklejohn, G. (1975) Biochim. Biophys. Acta 382,51-57. 13. Verma, S. P., Wallach, D. F. H. & Schmidt-Ullrich, R. (1975) Biochim. Biophys. Acta 394, 633-665. 14. Schmidt-Ullrich, R., Verma, S. P. & Wallach, D. F. H. (1976) Biochim. Biophys. Acta 426, 477-488.
15. Schmidt-Ullrich, R., Verma, S. P. & Wallach, D. F. H. (1975) Biochem. Biophys. Res. Commun. 67,1062-1069. 16. Milano, F. P., Yeh, Y., Baskin, R. J. & Harney, R. C. (1976) Biochim. Biophys. Acta 419,243-250. 17. Verma, S. P. & Wallach, D. F. H. (1976) Biochim. Biophys. Acta
436,307-318. 18. Lippert J. L. & Peticolas, W. L. (1971) Proc. Natl. Acad. Sci. USA 68, 1571-1576. 19. Lippert, J. L. & Peticolas, W. L. (1972) Biochim. Biophys. Acta 282,8-17. 20. Brown, K. G., Peticolas, W. L. & Brown, E. (1973) Biochem. Biophys. Res. Commun. 56,358-364. 21. Spiker, R. C., Jr. & Levin, I. W. (1975) Biochim. Biophys. Acta
388,361-373. 22. Mendelsohn, R., Sunder, S. & Bernstein, H. J. (1976) Biochim. Biophys. Acta 419,563-569. 23. Faiman, R. & Long, D. A. (1975) J. Raman Spectrosc. 3, 379385. 24. Verma, S. P. & Wallach, D. F. H. (1976) Biochim. Biophys. Acta
426,616-623. 25. Bieri, V. & Wallach, D. F. H. (1975) Biochim. Biophys. Acta 406, 415-423. 26. Wallach, D. F. H. (1975) Membrane Molecular Biology of Neoplastic Cells (Elsevier Publishing Co., Amsterdam), pp.
15-33. 27. Wallach, D. F. H. & Winzler, R. (1974) Evolving Strategies and Tactics in Membrane Research (Springer-Verlag, New York), pp. 223-226. 28. Verma, S. P. & Wallach, D. F. H. (1976) Biochim. Biophys. Acta, in press. 29. Jackson, W. M., Kostyla, J., Nordin, J. H. & Brandts, J. F. (1973) Biochemistry 12, 3662-3667. 30. Sheetz, M. P. & Chen, S. T. (1972) Biochemistry 11, 548-555. 31. Tanford, C. (1968) Adv. Protein Chem. 23, 121-282. 32. Williams, R. J. P. (1975) Biochim. Biophys. Acta 416, 237286.
Erythrocyte membranes undergo cooperative, pH-sensitive state transitions in the physiological temperature range: Evidence from Raman spectroscopy (protein-lipid boundary/amino acid spectra/methyl and methylene vibrations)
SURENDRA P. VERMA AND DONALD F. H. WALLACH* Tufts-New England Medical Center, Therapeutic Radiology Department, Radiobiology Division, 171 Harrison Avenue, Boston, Massachusetts 02111
Communicated by J. L. Oncley, August 6, 1976
lecithin-melittin complexes (24), and probably for the sharp shift in the spectra of thymocyte plasma membranes at +230 (13) as well as the two separate scattering discontinuities of erythrocyte ghosts, below 00 and at +170 (17). It appears, however, that the protein-lipid interface might give rise to the pH-sensitive, partially reversible thermotropic transition reported by paramagnetic quenching of tryptophan fluorescence in erythrocyte membranes at high physiological temperatures (25). We now show how laser-Raman spectroscopy can further elucidate the responses of erythrocyte membranes to variations of temperature and pH in physiological ranges.
ABSTRACT We have examined the Raman scattering from erythrocyte ghosts at 2700 to 3000 cm-l (CH-stretching region). Plots of the intensity (I) of the 2930 cm-1 band relative to the intensity of the thermally stable 2850 cm-' band, i.e., the [I2930/I2850] ratio, as a function of temperature reveal a sharp discontinuity, which at pH 7.4 has a lower limit of 380 and is irreversible above 420. [I2930/I25so1 is stable between pH 7.0 and pH 7.4, but increases or decreases sharply below pH 7.0 or above pH 7.5, respectively. Reduction of pH to 6.5 lowers the transition temperature by about 16°, and a shift to pH 6.0 drops the transition range to 0 to 70. The above effects of temperature and pH on Raman scattering closely correspond to those detected by studies on the interaction of membrane protein fluorophores and lipid-soluble fluorescence quenchers [Bieri, V. & Wallach, D. F. H. (1975) Biochim. Biophys. Acta 406, 415-4231. Taken together, these results suggest that the transitions represent concerted processes, involving hydrophobic amino acid residues and lipid chains at apolar protein-lipid boundaries. In 1967, Changeux et al. (1) presented a model for biomembranes in which ordered arrays of protein-lipid subunits, protomers, form "two-dimensional crystalline lattices." The model focuses on cooperative phase transitions in such systems. It "explicitly assumes that the transition of the protomer is highly cooperative with respect to its constitutive amino acid or phospho-lipid elements," with the unique aspects of membrane cooperativity deriving from state changes involving the "statistical-mechanical ensemble of protomers." That biomembranes may contain extended, two-dimensional lipid-protein lattices is evident from recent experimentation on hepatic "gap-junctions" (2) and the purple membranes of Halobacterium halobium (3). That membrane phospholipids can undergo thermally or ionically induced, cooperative phase transitions is also well established (see review in ref. 4), as is the fact that the function of many membrane proteins may vary profoundly with the state of membrane lipid (see review in ref. 5). Finally, that membrane proteins can influence the structure and composition of their apolar environment is now well documented (e.g., refs. 6-9). However, the dynamics of lipidprotein interfaces in biomembranes have proven elusive to date. We have used Raman spectroscopy to explore the interfaces between hydrocarbon residues of membrane proteins and lipids. Aqueous dispersions both of biomembrane (10-17) and of phospholipids (18-25) yield well-resolved Raman scattering spectra, whose CH-stretching regions are prominent and undergo marked band modifications at certain, defined temperatures (13, 17, 20-24). With pure lipids, the thermotropic discontinuities in Raman scattering can be identified as order >=± disorder transitions of hydrocarbon chains. This is almost certainly true also for the thermally shifted scattering steps of *
EXPERIMENTAL We used reagents of analytical grade and solvents of spectrochemical purity. Egg lecithin and other phosphatides were purchased from Lipid Products (South Nutfield, Great Britain). Erythrocyte ghosts were prepared (11, 17), suspended in 5 mM phosphate at stated pH to a concentration of 10 mg of protein per ml, transferred to thermoregulated capillaries, and their spectra recorded as in refs. 11, 13, 15, and 17 with a Ramalog 4 spectrometer (Spex Industries, Metuchen, N.J.) interfaced to an Interdata computer (model 70), using the 488 nm Ar+ laser line for excitation at 400,um slit width. Temperature equilibration was, as in ref. 17, for 20 min with the laser beam occluded and 5 min with the beam impinging on the sample, before scanning from 2750 cm-' to 3050 cm-1 (2-4 scans; not more than 300 s/scan). The averaged, smoothed (least-squares; ref. 17) spectra were plotted for each temperature. The data points or ranges presented derive from at least four separate membrane preparations.
RESULTS Variation of the CH-stretching region with temperature above 220 At 220 the CH-stretching region is characterized by three strong bands that lie at -2930 cm-', -2880 cm-', and 2850 cm-' (Fig. 1). As the temperature is raised from -220 to above 420, at pH 7.4, the -2930 cm-' band becomes very prominent, relative to the temperature-insensitive feature at 2850 cm-, and the -2880 cm-1 peak becomes obscured. A plot of the relevant intensity ratio, [I293/I2o/o], versus temperature is given in Fig. 2 for pH 7.4. [I2930/I28501 stays stable at -1.2 + 0.02 between set temperatures 200 and 350 and then shifts abruptly to -1.41 between set temperatures 360 and 450. This transition is reversible only up to a set temperature of 390. When we correct for the local heating due to the laser beam, we obtain a lower temperature limit of 380 for the discontinuity and an
To whom correspondence should be addressed.
3558
Biophysics:
Verma and Waflach
0
Proc. Natl. Acad. Sci. USA 73 (1976)
0 CY)
CY) 0) (%4
3559
1 .5
1.4
6.0 420C
1.3
1.2
pH 7.0
.22°C
20
30 40 TEMPERATURE (0C)
50
FIG. 2. Variation of [I2930/I2s,501 with temperature. Points were obtained with three separate membrane preparations. Conditions were as for data in Fig. 1.
Effect of pH on high-temperature membrane-
thermotropism jor pH 8.0 FIG. 1. Laser-Raman spectra, in the CH-stretching region, of erythrocyte ghosts at two temperatures (pH 7.4) and three pH values (220, set temperature). Membrane concentration was 10 mg of protein per ml. Excitation wavelength was 488 nm; slit width, 400 Am. Power at sample was 400 mW. The traces represent spectra from single membrane preparations. The variation between samples is given in Figs. 2-4 for [IV293/I2io, but also applies to the features between 2880 and 2900 cm-'.
irreversibility-temperature of 420.t The changes in [Ibso/I2&5o] that accompany the gel ± liquid-crystal transitions of egg-, dimyristoyl-, and dipalmitoyl-lecithins are 0.25,0.20, and 0.24, respectively. With lecithin:cholesterol (1:1, mol:mol), the change in ratio does not exceed 0.1. Variation of the CH-stretching region with pH at constant temperature
Fig. 1 shows the CH-stretching region at pH 6.0, 7.0, and 8.0 (220), while Fig. 3 gives a plot of [Iwao/I2o] versus pH. The spectra do not change significantly in the pH range 7.0-7.5, where [I2930/12bso] remains constant at 1.20 + 0.02. Between pH 7.5 and 8.5 [I29ao/I2&5o] drops discontinuously to about 1 and the CH-stretching region develops a pattern seen at # crystal transitions arise from altered interactions, amplified by Fermi resonance, between CH-stretching fundamentals and HCHdeformation overtones. Experimentation using differential thermal calorimetry and enzyme inactivation (29) suggests that erythrocyte membrane proteins undergo a thermotropic state transition with a lower limit near 40°. In addition, proton magnetic resonance spectroscopy (30) shows that methyl side chains of erythrocyte membrane proteins become more mobile as the temperature is raised above about 350. Our own studies (25) on the paramagnetic quenching of membrane tryptophan fluorescence by nitroxide analogs of stearic acid have focused on lipid-protein interactions at physiological and near physiological temperatures and pH values, rather than on overall protein behavior as studied in refs. 29 and 30. Our experiments indicate that the accessibility of the protein fluorophores to the lipid probes increases abruptly above 370, in a process that is reversible below 420, and shifts to lower temperatures upon pH reduction, concurrently diminishing in intensity. The data also document that the accessibility of the fluorophores is minimal at pH 7.0-7.2, increasing sharply at lower and higher pH. The strong correlation of the paramagnetic quenching results with the Raman data presented herein indicates that the two approaches describe different aspects of the same process. Indeed, a quantitative analysis of the Raman data suggests that the pH-sensitive, thermotropic change in the CH-stretching region represents a cooperative change of state, involving apolar amino acid residues, as well as lipid acyl chains: First, 60% of the CH3 groups in erythrocyte membranes are those of amino acid residues (27), which produce strong scattering near 2930
cm-1 (Table 1). Second, erythrocyte membranes give a step of 0.16 in [I2930/I2850] at -8° (17) and another step of not less than 0.22 above 380, adding up to 0.38, whereas the maximum shift we observe with the gel liquid-crystal transitions of diverse phosphatides is about 0.25. Barring a highly divergent behavior by membrane phosphatides, the magnitude of the overall change of [I2930/I285o1 from -200 to +50° implies a contribution by protein CH3 groups. This contribution is assigned to the high-temperature step reported here because (a) this is pH sensitive and only partially reversible; (b) other data (29, 30) also reveal no protein thermotropism in erythrocyte membranes below 350; and (c) fluorescence quenching (25) reveals clear modification of protein-lipid interaction above 350t That the magnitude of the low-temperature step is only 0.16 (17), i.e., less than observed with phospholipids undergoing gel i=± liquid-crystal transitions, suggests that the high-temperature step includes a lipid contribution. This proposal is consistent with paramagnetic quenching data (25), which shows increased accessibility of membrane protein fluorophores to lipoidal probes. The pH susceptibility of the thermosensitive [I2Wso/I2&so] step may relate to the reversible thermotropic transitions observed in some soluble globular proteins at pH values below their isoelectric points (31). However, the large response in the pH range 7.0-6.0 is unusual. One explanation for this might be that, because of the low dielectric constant of the membrane core, small changes in charge might influence the structure of proteins in membranes much more than that of proteins dissolved in aqueous solvents. The same argument applies to the pH sensitivity of [I2930/I2s50], or paramagnetic fluorescence quenching (25), at constant temperature. Taken together, present and previous (25, 29, 30) data lead us to hypothesize that: (a) The apolar residues of penetrating proteins, together with the hydrocarbon chains of phospholipids immediately apposed to these protein residues, constitute a separate phase. (b) The high-temperature, pH-sensitive thermotropic step in [I2930/I285o] represents a cooperative state transition in this phase, involving a concerted rearrangement of acyl-chain configuration and protein residue orientation. Our concept of a separate protein-lipid phase capable of a cooperative change in state takes into account recent nuclear magnetic resonance data documenting the mobility of apolar residues in globular proteins (reviewed in ref. 32) and in -
t Cholesterol CH3 groups are unlikely to be involved because of the weak CH-stretching signals of this sterol, its low CH3 contributions, and its tendency to reduce the magnitude of thermotropic shifts in the CH-stretching region (cf. ref. 24).
Biophysics:
Proc. Natl. Acad. Sci. USA 73 (1976)
Verma and Wallach
3561
Table 1. Raman frequencies in the CH-stretching region of amino acids bearing CH3 residues Alanine
3003 (9.5) 2982(6.5) 2960 (5.7) 2952 (10) 2930 (8.3)
2885 (3.6)
Valine
Leucine
Isoleucine
Methionine
2985 (8.9) 2975 (5.4) 2960 (10)
2985 (6.1)
2987 (4.6)
2925 (4.6)
2967 (10) 2958 (10) 2932 (10)
2960 (10) 2948 (7.9) 2930 (5.6)
2888 (1.9) 2979 (2.3) 2970 (1.9)
2895 (8.6) 2882 (6.5) 2878 (5.8)
2901 (10) 2896 (9.8) 2868 (8.0)
2882 (8.8)
2935 (1.7) 2925 (6.0) 2909 (10)
Threonine 3018 (7.9) 2992 (10) 2973 (8.3)
2936 (9.8)
2873 (6.5) 2870 (8.9)
Frequencies are in cm-1. Spectra were obtained on crystalline solids with 50 Am slits. Figures in parentheses give relative intensities.
erythrocyte membranes at temperatures above approximately 400. Our proposal is also in full accord with the model proposed by Changeux et al. (1) and Williams' treatment of lipid-protein interactions (32). This work was supported by Grant CA 13061 from the USPHS and PRA78 from the American Cancer Society to (D.F.H.W.). We thank Dr. R. Mikkelsen for his constructive criticism. 1. Changeux, J.-P., Thiery, J., Tung, Y. & Kittel, C. (1967) Proc. Natl. Acad. Sci. USA 57,335-341. 2. Goodenough, D. A. & Stoeckenius, W. (1972) J. Cell Biol. 54, 646-656. 3. Henderson, R. & Unwin, P. N. T. (1975) Nature 257,28-32. 4. Wallach, D. F. H. (1975) Membrane Molecular Biology of Neoplastic Cells (Elsevier Publishing Co., Amsterdam), pp. 147-153, 188-189, 218, 220. 5. Raison, J. K. (1973) J. Bioenerg. 4, 285-309. 6. Jost, P. C., Griffith, 0. H., Capaldi R. A. & Vanderkooi, G. (1973) Proc. Nati. Acad. Sci. USA 70,480-484. 7. Stier, A. & Sackmann, E. (1973) Biochim. Biophys. Acta 311, 400-408. 8. Wallach, D. F. H., Verma, S. P., Weidekamm, E. & Bieri, V. (1974) Biochim. Biophys. Acta 356,68-81. 9. Warren, G. B., Houslay, M. D., Metcalfe, J. C. & Birdsall, N. J. M. (1975) Nature 255,684-687. 10. Larsson, K. & Rand, P. (1973) Biochim. Biophys. Acta 326, 245-255. 11. Wallach, D. F. H. & Verma, S. P. (1975) Biochim. Biophys. Acta 382,542-551. 12. Lippert, J. L., Gorzyca, L. E. & Meiklejohn, G. (1975) Biochim. Biophys. Acta 382,51-57. 13. Verma, S. P., Wallach, D. F. H. & Schmidt-Ullrich, R. (1975) Biochim. Biophys. Acta 394, 633-665. 14. Schmidt-Ullrich, R., Verma, S. P. & Wallach, D. F. H. (1976) Biochim. Biophys. Acta 426, 477-488.
15. Schmidt-Ullrich, R., Verma, S. P. & Wallach, D. F. H. (1975) Biochem. Biophys. Res. Commun. 67,1062-1069. 16. Milano, F. P., Yeh, Y., Baskin, R. J. & Harney, R. C. (1976) Biochim. Biophys. Acta 419,243-250. 17. Verma, S. P. & Wallach, D. F. H. (1976) Biochim. Biophys. Acta
436,307-318. 18. Lippert J. L. & Peticolas, W. L. (1971) Proc. Natl. Acad. Sci. USA 68, 1571-1576. 19. Lippert, J. L. & Peticolas, W. L. (1972) Biochim. Biophys. Acta 282,8-17. 20. Brown, K. G., Peticolas, W. L. & Brown, E. (1973) Biochem. Biophys. Res. Commun. 56,358-364. 21. Spiker, R. C., Jr. & Levin, I. W. (1975) Biochim. Biophys. Acta
388,361-373. 22. Mendelsohn, R., Sunder, S. & Bernstein, H. J. (1976) Biochim. Biophys. Acta 419,563-569. 23. Faiman, R. & Long, D. A. (1975) J. Raman Spectrosc. 3, 379385. 24. Verma, S. P. & Wallach, D. F. H. (1976) Biochim. Biophys. Acta
426,616-623. 25. Bieri, V. & Wallach, D. F. H. (1975) Biochim. Biophys. Acta 406, 415-423. 26. Wallach, D. F. H. (1975) Membrane Molecular Biology of Neoplastic Cells (Elsevier Publishing Co., Amsterdam), pp.
15-33. 27. Wallach, D. F. H. & Winzler, R. (1974) Evolving Strategies and Tactics in Membrane Research (Springer-Verlag, New York), pp. 223-226. 28. Verma, S. P. & Wallach, D. F. H. (1976) Biochim. Biophys. Acta, in press. 29. Jackson, W. M., Kostyla, J., Nordin, J. H. & Brandts, J. F. (1973) Biochemistry 12, 3662-3667. 30. Sheetz, M. P. & Chen, S. T. (1972) Biochemistry 11, 548-555. 31. Tanford, C. (1968) Adv. Protein Chem. 23, 121-282. 32. Williams, R. J. P. (1975) Biochim. Biophys. Acta 416, 237286.
