Вы находитесь на странице: 1из 22

Bioscience Reports, Vol. 6, No.

3, 1986

Review

Infrared Spectroscopic Studies of Biomembranes and Model Membranes


David C. Lee and Dennis Chapman
Received

KEY WORDS: IR; membranes; protein structure; lipid phase transitions; lipid-protein interactions INTRODUCTION The use of vibrational spectroscopy to probe the structure and dynamics of biological molecules is a rapidly expanding area of research. Its origins, however, lie in the early studies of simple, organic molecules by infrared and Raman techniques. The comparison of vibrational spectra with molecular structure provided organic chemists with correlation charts which are the basis of sample analysis today. A standard text describing this approach is that of Bellamy (1975). Analysis of biological molecules is a more recent development and excellent reviews are those of Susi (1969), Fraser and MacRae (1973), Fawcett and Long (1973), and Thomas and Kyogoku (1977). The application of infrared and Raman techniques in membrane research has been reviewed by Wallach et al. (1979), Fringeli and Gunthard (1981), Amey and Chapman (1983), Levin (1984), and Casal and Mantsch (1984). The amount of structural and functional information available to the IR spectroscopist who is interested in biological systems is expanding rapidly with the application of increasingly sophisticated methods for data acquisition and analysis and new techniques in sample handling. These techniques include the detection and assignment of minor spectral components using derivative and deconvolution calculations, the study of biochemical reactions using kinetic IR spectroscopy and the use of cylindrical internal reflection cells for easy analysis of aqueous samples. Initially, the major problem in the study of biological molecules was the absorption of liquid water over much of the IR spectrum. This severely limited the analysis of samples in their natural state and necessitated the use of high concentrations, low pathlength cells (less than 50 gm) or deuterium oxide as a solvent. The advent of microprocessor-controlled spectroscopy has permitted the subtraction Department of Biochemistryand Chemistry, Royal Free Hospital School of Medicine (University of London), RowlandHill Street, London NW3 2PF.

235
0144-8463/86/0300-0235 $05.00/0 9 1986 Plenum Publishing Corporation

236

Lee and Chapman

of background water absorptions from dilute samples (Cameron et al., 1979; Chapman et al., 1980) thereby enabling accurate assessment of native structures. This approach has revolutionized the application of IR spectroscopy to the study of membrane systems in recent years. Vibrational spectroscopy has several advantages for membrane studies. Firstly, variations in frequency, linewidth, and intensity are sensitive to structural transitions of both lipid and protein components. Secondly, the vibrations of individual groups provide structural information on highly localized regions of the bilayer. Thus, C--H stretching absorptions of the lipid acyl chains are readily distinguished from the carbonyl stretchings of the interracial region and the phosphate stretchings of the polar headgroup. Of particular importance in the study of lipid-protein interactions is the non-perturbing nature of the technique. The addition of an external probe molecule is not required and the absorptions of the lipid and protein groupings reflect their genuine environments. Other techniques, such as ESR and fluorescence, are limited by the perturbations which the added reporter groups may induce. Finally, the time-scale of the molecular vibrations is of the order of 1013 s- 1 which ideally complements the ESR and NMR timescales of 108 s- a and 105 s- 1.

INSTRUMENTATION The application of the traditional, dispersive IR instrument in the biological field expanded with the widespread use of minicomputers. Computer aquisition of IR data allows the operator to store spectra in undegraded form. Spectral subtractions, enhancements and expansions may then be carried out at a later date (Chapman et al., 1980). Following the advent of the fast Fourier transform algorithm, spectrometers based on the Michelson interferometer are now widely used (Fourier transform infrared). An FT-IR spectrometer is comprised of two parts: an optical bench containing an interferometer and a computer which controls all aspects of spectral scanning and analysis. The interferogram of a scan, or the sum of many such scans, is converted by means of a fast Fourier transform into the conventional form of transmittance (or absorbance) versus wavenumber spectrum. A short description of the mathematical treatments involved is that given by Griffiths (1980). The principal advantage of FT-IR compared with dispersive IR is that the former is able to obtain spectra of higher signal to noise ratio in a given scanning time. This arises through the multiplex or Fellgett's advantage, the detector examines all of the scanned spectrum almost simultaneously, and the Jacquinot advantage, the high optical throughput of the interferometer owing to the absence of slits. A further advantage of FT-IR is the much greater abscissa (wavenumber) accuracy which is achieved via laser referencing.

Kinetic Infrared Spectroscopy


Kinetic IR spectroscopy (KIS) is a new technique which has been used to investigate the molecular basis of biological trigger processes. The method was developed by Kreutz and co-workers and a recent review which gives a brief

IR Studies of Membranes

237

description of the apparatus and its applications is that of Kreutz et al. (1984). These applications include the visual transduction process where the absorption of light induces molecular changes in the chromophore, retinal, which are transferred to the protein (Siebert and Mantele, 1980). The light-induced reactions of bacteriorhodopsin in the purple membrane of Halobacterium halobiurn have also been studied (Siebert et al., 1982; Engelhard et al., 1985). In both cases the molecular changes involved take place on a time-scale too fast for conventional IR methods. In addition, the extent of band overlap prevents the resolution of the band-shifts in the normal static IR spectrum. The methods used for sample preparation are determined by the requirement for both high sample concentration and the use of hydrated systems. A thin-film is formed on the IR crystal by dehydration from aqueous solvent and the film rehydrated with either H20 or ZH20 by evaporation from saturated salt solution. The results obtained by this technique, particularly with respect to bacteriorhodopsin, will be reviewed later. Although these workers note that many biological samples will not tolerate drying, little attention has been paid to the possibility of irreversible structural transitions occurring in these samples.

THE APPLICATION OF IR SPECTROSCOPY TO MEMBRANE SYSTEMS


Restrictions

The complexity of most biological molecules leads to difficulties in the interpretation of their IR spectra. This arises from the possibility of 3n-6 normal vibrations (where n is the number of atoms); a number which may be increased by overtones, combination tones, and band splitting. Often, unambiguous assignments are impossible because of the extent of band overlap and the variety of molecular motions. Biological membranes are in a liquid crystalline state at ambient temperatures. The lack of order in the hydrocarbon chains leads to a broadening of the IR absorptions and a reduction in polarisation. As mentioned in the introduction, liquid water absorbs strongly over much of the mid-infrared which has meant that many IR studies of biological materials have been performed using dry samples. The advent of microprocessor-controlled IR has allowed us to study membranes in their natural aqueous state, although the use of low pathlength cells is required to prevent total loss of transmission. This in itself may lead to errors in the estimation of absorption coefficients, and the use of high sample concentrations may induce molecular association. An alternative approach is the use of 2 H 2 0 , whose absorptions do not obscure the amide I and II bands of proteins, as a solvent. However, under these conditions variations in H--ZH exchange, which is both temperature- and conformationdependent may give rise to ambiguous band shifts. A further problem which may be experienced is the absorption of atmospheric water vapour which overlaps with the amide I and amide II bands of proteins and may

238

Lee and Chapman

confuse conformationat assignments, as has recently been noted by us (Lee et aI., 1985b). To combat this, the spectrometer should be thoroughly purged with dry gas (air or nitrogen) or evacuated during scanning.

Phospholipids
Studies of the IR spectra of phospholipids have provided much information on the structure of the acyl chain, interracial and head group regions of membrane bilayers (Casal and Mantsch, 1984). Comprehensive tables describing the vibrational absorptions of phospholipids are included in the reviews by Fringeli and Gunthard (1981) and Amey and Chapman (1983). The types of normal vibration for a CH 2 group as part of an extended polymethylene chain are illustrated in Fig. 1.

Asymmetric stretching

Symmetric

stretching
Scissoring

W~igging

Twisting

Rocking
Fig. 1. Typesof normal vibration. The infraredactive vibrations of a CHz group.

IR Studiesof Membranes

239

The Acyl Chain Region


An early study of films of phosphatidylcholine by Chapman et al. 0967) provided assignments of the C--H stretching and bending modes, the carbonyl band, and a variety of phosphate vibrations. These results showed that the spectra, particularly in the region below 1400cm -1, were remarkably dependent on temperature and the method of sample preparation. Considerable fine structure was observed in spectra recorded of samples after dehydration or at low temperatures. The band progression due t o C H 2 wagging modes and the C H 2 rocking band near 720 cm- 1 were found to be particularly sensitive. For anhydrous samples, the single 720 cm-1 band present at room temperature split into several components at - 186~ This transition coincides with a change in acyl chain packing from hexagonal to orthorhombic according to Xray diffraction of the same samples. This type of study was extended by Fookson and Wallach (1978) to include films of phosphatidylethanolamine and mixed films of phosphatidylcholine/phosphatidylethanolamine (1:1). Phosphatidylethanolamine showed splitting of the CH2 rocking fundamental even at room temperature indicating a orthorhombic rather than hexagonal chain packing. On monohydration, these samples showed a complete absence of fine structure because of the decreased intermolecular interactions which occur as water is inserted into the headgroup lattice. The sharp main endothermic phase transition of aqueous phospholipid bilayers results in pronounced alterations in the methylene band parameters (Asher and Levin, 1977; Cameron and Mantsch, 1978; Cortijo and Chapman, 1981). The band maximum frequencies of the CH 2 asymmetric and symmetric stretching bands are sensitive to the static order of the acyl chains. The introduction of an increased proportion of 9auche conformers above the phase transition causes a shift in these bands to higher frequencies (see Fig. 2). The width of the IR absorptions is determined by rotational, translational and/or collisional effects (Casal et al., 1980). Thus, the C H 2 bandwidths are sensitive to the degree of motional freedom of the CH 2 groups. They are sensitive, therefore, to the phase transition, but they also reflect changes which do not result in an alteration in the proportion of 9auche conformers, such as the extent of librational or torsional motion. A study of the C H 2 scissoring mode in hydrated bilayers of dipalmitoylphosphatidylcholine (DPPC) in the gel phase has shown that this band is sensitive to the orthorhombic-like to hexagonal packing transition which occurs as the temperature is increased towards the pretransition (Cameron et al., 1980a). The hexagonal phase gives a single band at 1468 cm -~ whereas the low-temperature orthorhombic phase gives two bands near 1475 cm- 1 and 1465 cm- 1. In a separate study (Cameron et al., 1980b), these workers demonstrated that changes in the gelphase spectrum of DPPC, particularly those associated with the pretransition near 37~ reflect alterations in interchain interactions and the packing of the fully extended all-trans acyl chains. Above 38~ the spectra reflect the hexagonal chain packing with a high degree of molecular motion. Below 36~ the spectra are sensitive to a reduction in axial motion and a gradual introduction of orthorhombic packing. As part of their extensive studies of lipid phase behaviour by FT-IR, Mantsch et al. (1982, 1983) have also studied the C--H modes of aqueous dispersions of

240

Lee and Chapman

2924

2923

2922
u

m 2921
Z

< 2920

2919

2918

J
I I

I0

20

30
(~

4O

TEMPERATURE

Fig. 2. The variation of the CH 2 asymmetric stretching frequency of dimyristoyl phosphatidylcholine bilayers with temperature. Band frequencies were measured from difference spectra generated by the subtraction of the spectrum of the buffer at each temperature. The mid-point of the phase transition occurs at 23 ~ in agreement with calorimetric measurements.

phosphatidylsulphocholines and phosphatidylethanolamines. On passing through the main phase transition the CH2 symmetric stretching band of dipalmitoylphosphatidylethanolamine (DPPE) bilayers shows a larger increase in frequency compared with the corresponding mode for DPPC bilayers (Casal and Mantsch, 1983). This suggests that more gauche bonds are introduced at the phase transition for DPPE than DPPC. Phosphatidylethanolamines with unsaturated acyl chains display a reversible liquidcrystalline to a non-lamellar inverted hexagonal (Hn) phase transition on further heating above the phase transition temperature Tm(Luzatti et al., 1968). The transition involves a further increase in the gauche concentration of the liquid-crystalline phase as revealed by the temperature-dependence of the CH z symmetric stretching frequency of egg yolk PE (Mantsch et al., 1981).

IR Studiesof Membranes

241

The C--H modes have also been used to investigate the phase behaviour of bovine brain phosphatidylserine (PS) and its interaction with calcium ions (Dluhy et al., 1983a). As for PC, an increase in the CH 2 symmetric stretching frequency coincides with the main phase transition. On addition of Ca 2+, the acyl chains are further ordered below T,, and a small monotonic increase in gauche conformers occurs with temperature--the phase transition being abolished.

The Interfacial and Headgroup Regions

The structure of the interfacial region of lipid assemblies may be examined via the ester group vibrations. The most intense of these bands are the C-~-O stretching frequencies between 1750 cm- 1 and 1700 cm- 1 Two absorption bands in this region are associated with the two ester groupings in diacyl lipids. The sn-1 carbonyl gives rise to a band near 1740 cm- 1 and a band near 1725 cm- 1 is associated with the carbonyl at the sn-2 position (Bush et al., 1980; Mushayakarara and Levin, 1980; Levin et al., 1982). This splitting arises, in part, because of the conformational inequivalence about the C1-C 2 bonds of the sn-1 and sn-2 chains which adopt trans and gauche conformations, respectively, and through possible differences in the extent of hydration. Difference IR spectra of hydrated samples rlormally reveal only a single broad C ~ O band contour. The application of either spectral deconvolution (Casal and Mantsch, 1984) or second-derivative (Lee et al., 1985b) calculations usuaUy reveals two or more components. The midpoint of the broad C ~ O stretching band shifts by 2 cm-1 to higher frequencies at the pretransition temperature; at the main transition there is a shift back to lower frequencies by 4 cm- 1 (Casal and Mantsch, 1984). These changes arise from alterations in the relative intensity rather than shifts in the sn-1 and sn-2 components. The shift in the C ~ O bands at the pretransition temperature is a reflection of a decrease in the angle of tilt of the acyl chains with respect to the bilayer normal. The lamellar liquid-crystalline to inverted hexagonal phase transition of unsaturated phosphatidylethanolamines results in an increase in the C---~O stretching band frequency such that it is approximately the same as in the gel phase (Mantsch et al., 1981). The phosphate moiety of the headgroup gives rise to several strong vibrations in the infrared. Asymmetric and symmetric stretching modes for the PO 2 group are found near 1250 cm - * and 1085 cm- ~, respectively (Casal and Mantsch, 1984; Arrondo et al., 1984). Weaker single bond P r O stretching modes are found in the region 900800cm-* (Casal and Mantsch, 1984). A shoulder near 1060cm -~ on the PO2 symmetric stretching band in the spectrum of DPPC is attributed to a R - - O - - P - - O - - R ' stretching mode (Arrondo et al., 1984). Akutsu et al. (1975) studied ordered films of DPPE by polarized IR spectroscopy. They found that the moments of the ~ O stretching and the PO 2 and C---C--N + stretching modes are oriented almost parallel with each other and deviate by less than 20~ from the plane of the film. The frequencies of the PO~- stretching =lodes are sensitive to the state of hydration of phospholipid bilayers. Dehydration results in band-shifts towards higher wavenumbers (Arrondo et al., 1984). These workers also

242

Lee and Chapman

found that these band-frequencies were insensitive to the main phase transition of DPPC.

Phospholipid-Cholesterol Interactions
There have been several IR studies on the effect of the incorporation of cholesterol on the static order of the acyl chains of aqueous phospholipids (Asher and Levin, 1977; Umemura et al., 1980; Cortijo and Chapman, 1981 ; Cortijo et al., 1982). These studies have shown that cholesterol causes an increase in the proportion of gauche conformers below the phase transition temperature (T,,) and a decrease in these conformers above Tm compared with a pure lipid bilayer. This is manifested as an increase in both the CH2 asymmetric and symmetric stretching frequencies below Tm and a decrease in these parameters above Tm(Cortijo and Chapman, 1981 ; Cortijo et al., 1982). The width of the phospholipid phase transition, as detected by 1R, is increased by the presence of cholesterol in the bilayer but the midpoint of the transition is unaffected (Asher and Levin, 1977). At very high cholesterol concentrations (e.g., 1.5:1 phospholipid: cholesterol molar ratio) almost no change in the relative proportions of trans and gauche conformers occurs with temperature (Cortijo et al., 1982). These observations are in accord with the results obtained by calorimetric, NMR and ESR techniques. There is little IR evidence concerning the perturbation of the interfacial and headgroup regions which may be exerted by cholesterol. A combined IR and Raman investigation of the C = O stretching modes of DPPC (Bush et al., 1980) demonstrated that no hydrogen bonding occurs between the fatty acyl carbonyl and the 3/~-OH group of cholesterol in anhydrous samples. In hydrated samples, cholesterol reduces the conformational inequivalence between the sn-1 and sn-2 carbonyls by perturbing the latter. This was revealed as a decrease in the relative intensity in the 1720 cm- 1 region. A study of the PO~- stretching modes and the asymmetric N(CH3)3 stretching mode of fully hydrated DPPC showed no variation in frequency on the introduction of cholesterol into the bilayers (Umemura et al., 1980). This suggests that there is little interaction between cholesterol and the phosphorylcholine group.

Lipid-Protein Interactions
IR spectroscopy is an ideal method for examining the perturbations of the lipid bilayer which are introduced by the presence of integral polypeptides and proteins. It is a technique which operates on a time scale complementary to the widely used NMR and ESR approaches and which does not require the addition of a reporter group to the system under study. Perturbations to the hydrophobic acyl chain region may be studied via the C--H asymmetric and symmetric stretching frequencies. These parameters are sensitive to the static order (trans/oauche isomerisation) of the acyl chains. Cortijo et al. (1982) have reported the effects of incorporation of the intrinsic proteins Ca 2+-ATPase and bacteriorhodopsin, and the intrinsic polypeptide gramicidin A on the acyl chain order of dimyristoylphosphatidylcholine (DMPC) and DPPC bilayers. They found that below Tm, these molecules behaved in a similar manner to cholesterol; that is they caused an increase in the proportion of gauche isomers. The presence of the intrinsic

IR Studies of Membranes

243

molecules within the bilayer prevents the acyl chains from attaining the all-trans conformation at low temperatures. Above T,,, the perturbation of the acyl chains differs from that observed with cholesterol. At high lipid:protein molar ratios (e.g., D M P C / C a 2+-ATPase of 150:1) a reduction in the proportion of 9auche conformers with respect to the pure lipid bilayer was observed. However, when the concentration of the intrinsic protein or polypeptide was increased, this effect was removed and the static order of the acyl chains was essentially the same as in the pure lipid system. This IR study also provided further evidence for a reduction in the cooperativity of the gel to liquid-crystalline phase transition in the presence of these intrinsic molecules. At high polypeptide/protein concentrations a reduction in T,, was observed. A qualification of this type of study must be made because of the possibility of protein absorptions in the spectral region studied. The presence of high intrinsic protein concentrations within the lipid bilayer structure introduces considerable amino acid side-chain contribution to the C - - H bands. More recently, we have investigated lipid-protein interactions using diperdeuteriomyristoylphosphatidylcholine, an analogue of D M P C in which the acyl chain hydrogen atoms are completely substituted by deuterium atoms (Lee et al., 1984). Gramicidin A, alamethicin and bacteriorhodopsin, as well as cholesterol, were reconstituted in bilayers formed from this lipid, enabling an investigation of the C-2H stretching frequencies without interference from amino acid side-chains at high protein concentrations. Band frequency was used as a measure of the static order of the lipid chains, while bandwidths gave information on the librational motion of the chains. It was shown that above the lipid phase transition temperature, low concentrations of gramicidin A and alamethicin caused a small ordering of the lipid chains while bacteriorhodopsin had no effect. At high concentrations each intrinsic molecule caused a disordering of the lipid chains above Tm. Bacteriorhodopsin had no effect, at either concentration studied, on the rate of acyl chain motion above Tm. Below T,,, each intrinsic molecule caused a disordering of the chains and an increase in chain motion compared to the pure lipid bilayer. We present, in Fig. 3, F T - I R spectra of 2-d27 DMPC, an analogue of D M P C in which the acyl chain hydrogens of the sn-2 chain have been replaced by deuterium atoms but the sn-1 chain remains unmodified. This allows us to monitor separately the introduction of oauche conformers in each chain via measurement of the C - - H and C - - 2 H stretching frequencies. At 8~ well below the phase transition temperature, the CH 2 asymmetric and symmetric stretching bands are found at 2920crn -1 and 2851 c m - 1 and the corresponding C2H2modes at 2194 c m - ~ and 2090 c m - ~. Above the phase transition (i.e., at 36~ the CH z bands are found at 2924 cm - t and 2854 c m - ~ and the C2H2bands are at 2196 c m - ~ and 2095 c m - ~. Shifts in these bands may be used to examine specific perturbations to the sn-1 and sn-2 chains induced by the presence of cholesterol, simple polypeptides and membrane proteins in phospholipid bilayers. Mantsch and co-workers have investigated the effects of the human erythrocyte membrane protein glycophorin on the IR spectra of D M P C bilayers (Mendelsohn et al., 1981, 1982; Dluhy et al., 1983b). There was no IR evidence for an immobilized lipid (annular lipid) component in these mixtures. The effects on the static order of the acyl chains were essentially the same as those observed by Cortijo et al. (1982) tbr Ca 2+-

244

Lee and Chapman

Cb u~

3000

2750

2500 WAVENUMBER

2250 (r "I )

2000

FT-IR spectra of 2-d27 dimyristoylphosphatidylcholine(DMPC) in water after subtraction of the background water spectrum recorded at 8~ (solid line)and 36~ (brokenline).Spectrawererecordedusing a Perkin-Elmer model 1750 spectrometerat 4 cm-1 resolution and apodised with a medium Norton-Beer function.
Fig. 3.

ATPase and bacteriorhodopsin. These workers also studied the temperaturedependence of the bandwidth of the C H 2 symmetric stretching band. The presence of glycophorin caused an increase in chain motions at all temperatures studied. A subsequent study investigated the ternary system DPPC/DMPC-d54/glycophorin (Dluhy et al., 1983b). Deuteration of the acyl chains of D M P C permitted the observation of the interaction of the protein with each lipid species via measurement of the C - - H and C--ZH stretching frequencies. The perturbing effects of the protein were more pronounced on the D P P C component than on the DMPC-d54. Preferential partitioning of Ca2+-ATPase in mixtures of perdeuterated and non-deuterated phospholipids in different physical states has also been shown by F T - I R (Jaworsky and Mendelsohn, 1985). This type of study was also extended to the ternary system PS/DPPC-d62/ glycophorin where it was shown that glycophorin preferentially interacts with the PS component (Mendelsohn et al., t984b). Furthermore, a specific perturbation of the interfacial region of PS was deduced from Fourier deconvoluted spectra in the lipid C = O stretching region. The presence ofglycophorin caused an increase in the relative intensity of the 1742 cm-1 component compared to the pure lipid system above T m. This suggests that a higher proportion of the acyl chains adopt a trans conformation

IR Studies of Membranes

245

about C 1 - C 2, In addition, a new component was observed at 1714cm - t which suggest that a proportion of the sn-2 carbonyls are in a more polar environment than in free PS, possibly involved in a weak hydrogen bond. Our second-derivative analysis also permits an investigation of lipid-protein interactions at the interfacial region of the bilayer (Lee et al., 1985b). The secondderivative spectra reveal the sn-1 and sn-2 carbonyls near 1740cm-1 and 1726 c m - i , respectively. Measurement of t he band intensity ratio 11 ~26/I 1740 (Table i) indicates that the insertion of integral proteins into a lipid bilayer causes a conformational change at the interfacial region below the phase transition temperature (23~ but has little effect above this temperature. The presence of Ca 2 + -ATPase or bacteriorhodpsin reduces the conformational inequivalence of the chains such that a proportion of the originally bent sn-2 chains are constrained in a conformation similar to that of the sn-1 chain, leading to a reduction in band intensity at 1726cm -1 (Table 1). As a consequence the perturbed sn-2 carbonyls are displaced toward the centre of the bilayer. The frequencies of the bands arising from the phosphate and choline groups of D M P C have been measured in the presence of glycophorin (Mendelsohn et al., 1981). N o significant variations with temperature or protein concentration were observed.

TaMe 1. The effects of CaZ+-ATPase and bacteriorhodopsin on the intensity ratio of the carbonyl stretching bands of DMPC
11726/11740 DMPC/Ca2 +-ATPase 245:1 0.31 0.43 0.55 0.70 0.81 D MPC/bacteriorhodopsin 135:1 0.32 --0.83 --

Temperature/~ 10 20 25 35 45

DM PC 0.52 0.45 0.50 0.79 0.80

Data were obtained from the second-derivativespectra presented,in part, by Lee et al. (1985b). The negative absorbance intensities at 1726cm-1 and 1740cm-1 were measured at each temperature. Membrane Proteins

Polypeptides and Conformational Assignments


Information on the conformation of polypeptides and proteins may be derived from their IR spectra (Krimm, 1962; Susi, 1969; Fraser and MacRae, 1973; Fawcett and Long, 1973; Thomas and Kyogoku, 1977). The assignment of the characteristic absorptions (the so-called "amide" bands) has been assisted by the analysis of model compounds such as N-methyl acetamide and is summarised in these reviews. For many years there was a great deal of controversy concerning the molecular description of these vibrations and this is summarised by Susi (1969). It is now generally accepted that the amide I to amide VII vibrations cannot be described adequately by any single displacement coordinate. It should be noted that these amide bands are not the only

246

Lee and Chapman

absorptions observed. The IR spectra of polypeptides and proteins also include contributions from the amino acid side-chains. The amide I and II bands are sensitive to the secondary structure of polypeptides and proteins and are the most frequently used modes in conformational analysis. The fundamental theory has been described by Miyazawa (1960). Some early applications of this approach are those of Miyazawa and Blout (1961), Krimm (1962), Susi et al. (1967), Timasheff and Susi (1966), and Timasheff et aL (1967). These and other studies, both experimental and theoretical (Jakes and Krimm, 1971; Abe and Krimm, 1972; Bandekar and Krimm, 1980; Krimm and Bandekar, 1980; Dwivedi and Krimm, 1984; Chirgadze and Nevskaya, 1976a,b; Nevskaya and Chirgadze, 1976; Chirgadze et al., 1973; Chirgadze and Brazhnikov, 1974; Kawai and Fasman, 1978) have enabled confident correlations between amide I and II band frequencies and secondary conformation. These assignments are presented in Table 2.

Table 2. Characteristic amide I and II frequencies for various polypeptide conformations Anide I (cm -1) Conformation e-Helix Random coil Anti-parallel Chain fl-sheet Parallel chain /3-Sheet H20 1652 (s) 1646 (w) 1656 1632 (s) 1690 (w) 1630 (s) 1645 (w) 2H20 1650 (s) 1644 (w) 1643 1632 (s) 1675 (w) 1632 (s) 1648 (w) 1546 (s) 1516 (w) 1520 1530 (s) 1510 (w) 1530 (s) 1550 (w) Amide II (cm-1)

s: strong. w: weak. Suspension of soluble proteins in 2H20 shifts the amide II band (principally N - - H bending) to frequencies near 1450cm -~ (amide II').

An extensive IR study of the conformation of the amphipathic polypeptide melittin has been made (Lavialle et at., 1982). In aqueous solution melittin exists in a monomer-tetramer equilibrium which is dependent on both concentration and ionic strength(Faucon et al., 1979). Analysis of the amide I and II band frequencies in H 2 0 and 2HzO revealed that at melittin concentrations in which the tetrameric form is dominant, the conformation is mainly e-helical at high temperature (31~ and a mixture of fi- and e-helical forms at low temperature (8~ (Lavialle et al., 1982). The relative intensity of the amide II and amide I bands (A~I/A1)was high (1.8-1.1) for the tetrameric form in water compared to the monomeric form (0.75). Melittin inserts into phospholipid bilayers in its e-helical form (Dawson et at., 1978) and in these samples amide I and II band maxima of 1653 cm- 1 and 1545 cm- x, respectively, were obtained. This conformation was unaffected by the gel to liquid-crystalline phase transition of the bilayers. The structure of the polypeptide antibiotic, gramicidin A has also been investigated by IR spectroscopy in solution and after insertion into phospholipid bilayers (Cortijo et al., 1982; Sychev and Ivanov, 1982; Buchet et al., 1985). Gramicidin

IR Studiesof Membranes

247

A is a linear polypeptide antibiotic that renders biological membranes and lipid bilayers passively permeable to small monovalent cations by forming a channel structure (Chappell and Crofts, 1965; Hladky and Hadon, 1970, 1972). It consists of fifteen alternating D- and L-amino acids (Sarges and Witkop, 1965). The structure of the polypeptide in lipid bilayers is a matter of some dispute and is dependent on temperature, lipid :protein ratio and lipid class (Sychev and Ivanov, 1982). Although an N-terminal to N-terminal linked dimeric structure has been proposed on the basis of NMR evidence (Weinstein et a l., 1979), IR and CD data are consistent with the presence of an anti-parallel /~-sheet structure in DMPC liposomes at lipid:protein ratios between 350:1 and 35:1 (Sychev and Ivanov, 1982). In DPPC bilayers an anti-paralM double-helical structure (Veatch et al., 1974) is found at high lipid:protein ratios (Sychev and Ivanov, 1982).
Globular Proteins

A question remains concerning the extent to which the conformational assignments obtained from relatively simple polypeptides and summarised in Table 2 can be used to describe the structure of large globular proteins. The first detailed study of the IR spectra of globular proteins in HaO was performed by Koenig and Tabb (1980) using digital subtraction of the aqueous background. The difference IR spectra of haemoglobin, bovine serum albumin, ribonuclease, fl-lactoglobulin, and e-caesin were examined in the region 2000-800 cm- 1. The amide I frequency was shown to be the same in the aqueous and cast film samples. The amide II band was shifted to higher frequencies on dissolution in water as a result of hydrogen bonding to N--H. Although hydrogen bonding had no effect on the amide I maximum there was a shift in the distribution of the amide I frequencies to lower values which is indicative of increased hydrogen bonding on dissolution (Susi, 1972). A sharpening of the amide I band of those proteins with some e-helical conformation occurred upon dissolution (Koenig and Tabb, 1980). This reduction in bandwidth was interpreted as being due to ordering of e-helical segments in aqueous solution and suggests that e-helical proteins are not in their native state in the solid phase. By correlation of the difference spectra with the known secondary structures of the proteins studied, Koenig and Tabb were able to distinguish e-helical, random coil and fl-sheet contributions by a combination of amide I maximum frequency and band intensity in the amide III region. The amide II peak maxima of-these proteins did not permit a distinction between the various conformations. However, it was noted that the more ordered proteins (haemoglobin and bovine serum albumin) gave a smaller shift due to hydrogen bonding on dissolution than the less ordered proteins (ribonuclease,/% lactoglobulin and a-casein). In principle, a globular protein containing several types of substructure will give several amide I and II maxima. However, the large half-widths of these components prevents their resolution. Least squares optimisation and Fourier deconvolution procedures have been used for the analysis of overlapping bands (Fraser and Suzuki, 1966, 1969; Ruegg et al., 1975; Kauppinen et al., 1981; Jap et al., 1983; Mendelsohn et al., 1984a). The usefulness of least squares calculations is limited by the requirement for extensive input information. Typically, the number and peak frequencies of the

248

Lee and Chapman

component bands are required, together with their lineshapes, half-widths, and a linear baseline. An alternative method for assessing the number and position of the component peaks is derivative spectroscopy (Maddams and Southon, 1982). Secondderivative spectra of water-soluble proteins have recently been obtained (Susi and Byler, 1983). These authors were able to resolve peaks associated with a-helical,/%sheet, and fl-turn conformations together with the vibrations of some amino acid side-chains. More recently, different denatured forms of water-soluble proteins were resolved using resolution enhancement and second-derivative calculations (Purcell and Susi, 1984). While other workers use Fourier self-deconvolution as a means of resolution enhancement (Dev et al., 1984; Yang et al., 1985), our preferred approach is the use of second- (Lee et al., 1985a,b) and fourth-derivatives. Information on the absorption of amino acid side-chains in the 1800-1500 cmregion, i.e., overlapping the amide I and II bands, has been available for some time (Bendit, 1967; Chirgadze et al., 1975). The vibrations of asparagine, glutamine, aspartic, and glutamic acids, arginine and tyrosine may contribute to the spectra of proteins in this region. With the application of second-derivative methods the resolution of these bands has become possible (Susi and Byler, 1983; Lee et al., 1985b) thereby providing a probe into specific regions of globular proteins.
Detailed Studies on Bacteriorhodopsin and Ca 2 +-ATPase The structure of bacteriorhodopsin in the purple membrane of Halobacterium halobium has been studied by IR spectroscopy (Rothschild and Clark, 1979b; Cortijo et al., 1982; Krimm and Dwivedi, 1982; Lee et al., 1985). These studies have confirmed the existence of a predominantly a-helical structure for this protein. The observed amide I frequency of 1660-1662 cm - ~is unusually high for e-helices and has been attributed to "distorted" a-helices (Rothschild and Clark, 1979b) or eii helices (Krimm and Dwivedi, 1982; Dwivedi and Krimm, 1984). In the en helix the N - H bonds are tilted inward toward the helix axis whereas in the cq structure they are essentially parallel to the axis. Polarised IR spectroscopy has been applied to determine the angle of tilt of the ehelices in dry, oriented specimens (Rothschild and Clark, 1979a). In agreement with the electron microscope data, the dichroism of the amide A, I and I1 bands indicated an average spatial orientation for the helices of less than 26~ from the membrane normal. On reconstitution of bacteriorhodopsin in bilayers of D MPC, the major component of the amide I, revealed by second-derivative analysis, shifted by 5 cm -1 to lower frequency, suggesting a conformational change in the a-helices (Lee et al., 1985). This structure was unaltered by the change in membrane fluidity which is characteristic of the phase transition. We present, in Fig. 4, difference (a), second-derivative (b) and fourth-derivative (c) spectra of an aqueous suspension of purple membrane at 20~ after subtraction of the absorption of the water. The amide I maximum is found at 1660 cm- 1 and may be attributed to distorted or e~-helices. A predominantly e-helical structure is also indicated by the amide II maximum at 1545 cm -~, in agreement with the electron diffraction data of Henderson and Unwin (1975). However, an amide I band shoulder at 1630-1640cm -1 suggests the presence of some /~-structure. This absorption is rendered more observable by the second- and fourth-derivative treatments and shown to have a maximum at 1635 cm-1. The appearance of a band at 1684 cm-1 suggests

IR Studies of Membranes
I I I I I / L L I I I L I I L

249

,o

,o

6-

on

; i~-

i 1888

, IT-~) !688 1580 148~

W A V E N U M B E R

(cm -I)

Fig. 4. FT-IR spectra of purple membrane in water at 20~ Spectra were recorded on a Perkin-Elmer model 1750spectrometerat 2 cm - ~resolution and apodized with a medium Beer-Norton function. (a) Difference spectrum generated after subtraction of the spectrumof water, (b) second-derivativeof (a), (c) fourth-derivative of (a). Minima in the second-derivative correspond to maxima in the difference spectrum which also correspond to maxima in the fourth-derivative.

that a proportion of this e-structure is anti-parallel pleated sheet. Bands at 1618 c m - 1 and 1517cm -1 may be tentatively assigned to tyrosine side-chain absorptions (Chirgadze et al., 1975). Bands in the region 1469-1439 c m - 1 are principally due to C - - H deformation modes from the lipid acyl chains. The structure of the sarcoplasmic reticulum (SR) of rabbit striated muscle has also been investigated by IR spectroscopy (Cortijo et al., 1982; Mendelsohn et al., 1984; Lee et al., 1985b; Arrondo et al., 1985). Examination of the amide I and II maxima indicated that the major protein present in this membrane, the Caa+-ATPase, contains unordered as well as e-helical conformation (Cortijo et al., 1982; Lee et aL, 1985b). Fourier deconvolution (Mendelsohn et al., 1984) and second-derivative analysis (Lee et al., 1985b) of the amide I region revealed a minor amount of/3-sheet structure. Temperature-dependent changes in the amide II region revealed by Fourier selfdeconvolution may have been due to a reduction in the proportion of e-helical structure as the temperature was increased although the influence of increased 1 H 2 H exchange was not eliminated (Mendelsohn et al., 1984). Our second-derivative (Lee et al., 1985) and fourth-derivative (Lee, D. C. and Chapman, D., unpublished) analysis of

250
i r i ~ E L L l i r r l E i E L L L

Lee and Chapman

b
Z < r,,-" O er~

I~

181-3~

170~3

! ~1~0

1500

~4 0 0

WAVENUHBER (cm-11
Fig. 5.

FT-IR spectra of sarcop]asmic reticu]um in 10 mM Hepes, 50 mM

sucrose, 1 M KC1,pH 7.4 at 20~ Spectra wererecorded using a Perkin-Elmer model 1750 spectrometer at 2 cm-~ resolution and apodized with a medium Beer-Norton function. (a) Differencespectrum generated after the subtraction of the spectrum of the aqueous buffer, (b) second-derivativeof (a), (c) fourthderivative of (a). the amide I and II bands has revealed a remarkable stability to temperature changes over the range 8 ~ to 50~ Figure 5 shows difference (a), second-derivative (b) and fourth-derivative (c) F T 4 R spectra of sarcoplasmic reticulum membranes at 20~ after subtraction of the absorption of the aqueous buffer. Three main absorption bands are seen in the difference spectrum, the amide I and amide II bands from the protein at 1655 c m - a and 1547 c m - 1, respectively, and a C = O stretching band from the lipid at 1737 c m - 1. The frequency of the amide I maximum of 1655 c m - i may be assigned to the presence of a predominantly c~-helical structure for Ca2+-ATPase. However analysis of the second-derivative reveals the presence of fl-sheet structure, with absorptions at 1632 c m - a and 1680-1690 c m - t , and which has been predicted from the primary sequence (Allen et al., 1980; MacLennan et aI., 1985). Further analysis using the fourth-derivative reveals components in the amide I region which may be assigned to c~-helical structure (1657cm -1 and 1643cm-1), fl-sheet structure (1681 c m - i and 1630 c m - 1), and fl-turns (1690 cm t) (Susi and Byler, 1983; Lee et al., 1985). The band at 1531 cm -1 in the amide II region may also be due to fl-structure. Lipid-protein interactions in the isolated membranes have been examined via the temperature-dependence of the CH 2 symmetric stretching frequency. No cooperative

IR Studies of Membranes

251

phase transition was observed in these membranes; instead increasing temperatures resulted in a continuous increase in the disorder of the acyl chains (Mendelsohn et al., 1984). FT-IR difference spectra have been applied to the study of conformational changes in both bacteriorhodopsin and its chromophore, retinal (Rothschild et al., 1981). By comparison with resonance Raman spectra which are only sensitive to the absorptions of the chromophone, a negative peak at 1530 cm- 1 and a positive peak at 1567cm -1 (both C~---C stretching) were assigned to the loss of bRs7 o and the production of M412 on illumination. In the context of these studies it should be noted that, in contrast to resonance Raman, infrared light does not drive the photocycle (Rothschild et at., 1981). In a separate F T q R difference study, Rothschild and Mar rero (1982) confirmed that bR57 o contains a protonated Schiff base, the linkage between retinal and a lysine residue of bacteriorhodopsin. Furthermore, the transition from bRs 7o to the K intermediate, the only light-driven step of the photocycle, involved the movement of the Schiffbase proton away from a counterion. A simple model for proton translocation was proposed from the IR data. The K intermediate was stabilised by recording its spectrum at 77 K during illumination with 500 nm light. The assignment of the C = N stretch of bR~7o and K63 o to absorption bands at 1639cm -1 and 1609 cm-1 respectively, was recently confirmed by a series of 1H-ZH and 12C-13C isotopic substitutions in the chromophore (Rothschild et al., 1984). The results of this study supported the earlier model of an environmental change for the Schiff base during the bR570 to K transition. FT-IR difference spectroscopy has also provided evidence for deprotonation of the Schiff base (Bagley et al., 1982), and protonation of a protein carboxytate (aspartate or glutamate) (Rothschild et al., 1981), on formation of the M intermediate. Kinetic IR spectroscopy has been used in studies of the photocycle of bacteriorhodopsin (Kreutz et al., 1984; Siebert et al., 1982). The bRsv 0 to M412 difference spectrum was calculated by taking the amplitudes of the kinetic signals at 300/~s. The generated spectra were consistent with a deprotonation of the Schiff base on formation of M412 (Kreutz et al., 1984). This was deduced from a band shift from 1640 cm- 1 [C~___N(H+)_] to 1620 cm- 1 [C=-N--] and a reduction in the absorption intensity of the C~---C stretching vibration at 1525 cm-1. The importance of isotopic labelling was emphasized for the correct identification of chromophore bands (see also Rothschild et al., 1984). Changes in the protein which occur as part of the photocycle may be identified by comparison of the KIS signals with kinetic visible absorption changes (Kreutz et al., 1984). Those KIS signals which display non-chromophore kinetics are likely to be due to changes in the protein. Correct assignment may be ensured by chemical or isotopic labelling of specific amino acids and proteolytic modification of the protein. The KIS rise-time of a signal at 1765 cm -1 in H20 and 1755 cm -~ in 2H20 was shown to correspond closely to the rise-time of the M4a 2 intermediate in the same samples (Siebert et al., 1982). These changes were assigned to the protonation of an internal carboxylic acid residue on deprotonation of the Schiff base. The rise of a second band at 1755 cm-~ or 1745 cm-1 was assigned to the protonation of a second carboxylate and was postulated to be involved in the mechanism for pumping of a second proton. It is interesting to note that the changes observed by these authors using

252

Lee and Chapman

the KIS technique, agree well with the static difference spectra obtained by Rothschild and co-workers who also described the protonation of carboxylate residues on deprotonation of the Schiff base (see above). Very recently the static and time-resolved IR difference methods have been combined in a single study of the protonation states of internal carboxylates in purple membranes which had been labelled with 13C.aspartic acid (Engelhard et al., 1985). These workers showed that four aspartate residues are involved on protonation-deprotonation changes as part of the photocycle.

Studies on Other Proteins

FT IR has also been applied to the study of photoreceptor membranes (Rothschild et al., 1980a). The amide I and II maxima indicated a predominantly ~helical structure for the protein, rhodopsin, and a polarised IR study has shown that the average helix orientation is perpendicular to the membrane plane (Rothschild et al., 1980b). More recently both static (Rothschild et al., 1983; Siebert et al., 1983; Bagley et al., 1985) and kinetic (Siebert and Mantele, 1980) difference techniques were used in attempts to outline some of the molecular changes in rhodopsin which may occur as part of the visual process. In addition to the systems described in detail above, useful IR data have also been obtained from erythrocyte membranes (Maddy and Malcolm, 1965; Graham and Wallach, 1971; Chapman et al., 1977; Bennun et al., 1985), mitochondrial membranes (Graham and Wallach, 1969), Ehrlich ascites carcinoma membranes (Wallach and Zahler, 1968), purified axon membranes (Papakostidis et al., 1972), and clathrin-coated vesicles (Vincent et al., 1984; Steer et al., 1984). IR spectroscopy has been used to monitor lipid-protein interactions in whole biomembranes. Intact and deproteinated plasma membranes ofAcholeplasma laidlawii were studied after the biosynthetic incorporation of DPPC-d6z (Casal et al., 1980). Plots of the temperature dependence of the C2H2 symmetric stretching frequency indicated a broad phase transition, centred at the growth temperature of the organism, which was interpreted as a regulatory mechanism for cell growth. The effect of the endogenous proteins is to stabilise the acyl chains in their preferred conformation at a given temperature. Thus, below the phase transition, the intact membranes have a higher proportion of trans conformers in their acyl chains than the deproteinated membranes. However, above the phase transition there is an increase in the proportion of 9auche conformers. A reduction in the bandwidth also suggests a reduction in the mobility of the acyl chains in the intact membrane. This implies that the presence of the endogenous proteins reduces the rate of acyl chain interconversion between trans and gauche conformers. In subsequent studies (Cameron et al., 1983, 1985), these workers investigated the thermotropic behaviour of Acholeplasma laidlawii grown on a fatty acid-depleted medium supplemented with perdeuterated fatty acids and an inhibitor of endogenous fatty acid synthesis. Analysis of the C-2H symmetric stretching band frequency revealed that the isolated membranes are predominantly in the gel state at the growth temperature. However, 95-100~ of the lipids in the live cell membranes are in the liquid-crystalline phase at 37~ (Cameron et al., 1985).

IR Studies of Membranes CONCLUSIONS

253

Infrared spectroscopy is clearly a powerful technique for the study of membrane lipids, membrane proteins, and lipid-protein interactions. The recent expansion of research activity in this field has been a direct consequence of developments in IR instrumentation. Digital recording and analysis via microcomputer, Fourier transform methods for data acquisition and kinetic analysis of IR absorptions have been important developments. Furthermore, the application of resolution enhancement and second-derivative calculations permits the identification of absorbances which were previously lost beneath stronger bands of large width. An extensive collection of data on the structure of membrane phospholipids is now available. Further progress in this field is likely to occur in studies of more unusual membrane lipids. A particularly interesting example is the analysis of cerebrosides, an important constituent of brain cell membranes, where the potential for inter- and intramolecular hydrogen bonding may be investigated via the amide absorptions. Studies of the vibrational absorptions of the headgroup and interracial regions will continue to provide important information on lipid-protein interactions. The use of deuterated lipids has been shown to be useful in the investigations of the specificity of lipid-protein interactions in multi-component systems. By specifically deuterating phospholipids at selected sites in the acyl chains, perturbations to localized regions of the hydrophobic core may be studied. The study of membrane protein structure by IR spectroscopy will continue to be an active area of research. The elucidation of the functional activity of membrane proteins at the molecular level will proceed via a range of spectroscopic, biochemical and crystallographic approaches. IR spectroscopy has the advantage of being able to examine native, aqueous systems. The detection of specific amino acid absorptions via F T and second-derivative methods holds promise for the investigation of specific regions of complex membrane proteins and may eventually provide data on the molecular basis of transport, receptor and enzyme mechanisms.

ACKNOWLEDGEMENTS We wish to thank Dr A. A. Durrani for synthesising 2-d27 D M P C and the Wellcome Trust and the Science and Engineering Research Council for financial support.

REFERENCES Abe, Y., and Krimm, S. (1972). Biopolymers 11:1817-1839. Akutsu, H., Kyogoku, Y., Nakahara, H., and Fukuda, K. (1975). Chem. Phys. Lipids 15:222-242. Allen, G., Trinnaman, B. J., and Green, N. M. (1980). Biochem. J. 187:591-616. Amey,R. L., and Chapman, D. (1983).In Biomembrane Structure and Function (D. Chapman, Ed.), Topics in Molecular and Structural Biology, Vol. 4, MacMillan, London, pp. 199-256. Arrondo, J. L. R., Goni, F. M., and Macarulla, J. M. (1984). Biochim. Biophys. Acta 794:165-168.

254

Lee and Chapman

Arrondo, J. L. R., Urbaneja, M. A., Goni, F. M., Macarulla, J. M., and Sarzala, G. (1985). Biochem. Biophys. Res. Commun. 128 : 1159-1163. Asher, J. M., and Levin, I. W. (1977). Biochim. Biophys. Acta 468:63-72. Bagley, K. A., Balogh-Nair, V., Crotean, A. A., Dollinger, G., Ebrey, T. G., Eisenstein, L., Kyung Hong, M., Nakamishi, K., and Vittitow, J. (1985). Biochemistry 24:6055-6071. Bagley, K., Dollinger, G., Eisenstein, L., Singh, A. K., and Zimanyi, L. (1982). Proc. Natl. Acad. Sci. USA 79:4972-4976. Bandekar, J., and Krimm, S. (1980). Biopolymers 19:31-36. Bellamy, L. J. (1975). The Infrared Spectra of Complex Molecules, Methuen, London. Bendit, E. G. (1967). Biopolymers 5:525-533. Bennun, A., Needle, M. A., and De Bari, V. A. (1985). Bioehem. Soc. Trans. 13:127-128. Buchet, R., Sandorfy, C., Trapane, T. L., and Urry, D. W. (1985). Biochim. Biophys. Acta 821:8-16. Bush, S. F., Levin, H., and Levin, I. W. (1980). Chem. Phys. Lipids 27:101-111. Cameron, D. G., Casal, H. L., Gudgin, E. F., and Mantsch, H. H. (1980a). Biochim. Biophys. Acta 596:463467. Cameron, D. G., Casal, H. L., and Mantsch, H. H. (1979). J. Biochem. Biophys. Methods 1:21-36. Cameron, D. G., Casal, H. L., and Mantsch, H. H. (1980b). Biochemistry 19:3665-3672. Cameron, D. G., and Mantsch, H. H. (1978). Biochem. Biophys. Res. Comm. 83:886-892. Cameron, D. G., Martin, A., and Mantsch, H. H. (1983). Science 219:180-182. Cameron, D. G., Martin, A., Moffat, D. J., and Mantsch, H. H. (1985). Biochemistry 24:4355-4359. Casal, H. L., Cameron, D. G., Smith, I. C. P., and Mantsch, H. H. (1980). Biochemistry 19:444451. Casal, H. L., and Mantsch, H. H. (1983). Biochim. Biophys. Acta 735:387-396. Casal, H. L., and Mantsch, H. H. (1984). Biochim. Biophys. Acta 779:381-401. Chapman, D., Cornell, B. A., Eliasz, A. W., and Perry, A. (1977). J. Mol. Biol. 113:517-538. Chapman, D., Gomez-Fernandez, J. C., Goni, F. M., and Barnard, M. (1980). J. Biochem. Biophys. Methods 2:315-323. Chapman, D., Williams, R. M., and Ladbrooke, B. D. (1967). Chem. Phys. Lipids 1:445-475. Chappell, J. B., and Crofts, A. R. (1965). Biochem. J. 95:393-402. Chirgadze, Yu. N., and Brazhnikov, E. V. (1974). Biopolymers 13:1701-1712. Chirgadze, Yu. N., Fedorov, O. V., and Trushina, N. P. (1975). Biopolymers 14:679-694. Chirgadze, Yu. N., and Nevskaya, N. A. (1976a). Biopolymers 15:607-625. Chirgadze, Yu. N., and Nevskaya, N. A. (1976b). Biopolymers 15:627-636. Chirgadze, Yu. N., Shestopalov, B. V., and Venyaminov, S. Yu. (1973). Biopolymers 12:1337-1351. Cortijo, M., Alonso, A., Gomez-Fernandez, J. C., and Chapman, D. (1982). J. Mol. Biol. 157:597-618. Cortijo, M., and Chapman, D. (1981). FEBS Lett. 131:245-248. Dev, S. B., Rha, C. K., and Walder, F. (1984). J. Biomol. Struct. and Dynamics 2:431-442. Dluhy, R. A., Cameron, D. G., Mantsch, H. H., and Mendelsohn, R. (1983a). Biochemistry 22:6318-6325. Dluhy, R. A., Mendelsohn, R., Casal, H. L., and Mantsch, H. H. (1983b). Biochemistry 22:1170-1177. Dwivedi, A. M., and Krimm, S. (1984). Biopolymers 23:923-943. Engelhard, M., Gerwert, K., Hess, B., Kreutz, W., and Siebert, F. (1985). Biochemistry 24:400-407. Faucon, J. F., Dufourcq, J., and Lussan, C. (1979). FEBS Left. 102:187-190. Fawcett, V., and Long, D. A. (1973). In Molecular Spectroscopy 1:352-445, Chem. Soe., London. Fookson, J. E., and Wallach, D. F. H. (1978). Arch. Biochem. Biophys. 189:195-204. Fraser, R. D. B., and MacRae, T. P. (1973). "Conformations in Fibrous Proteins and Related Synthetic Polypeptides", Chapter 5, Academic Press, New York. Fraser, R. D. B., and Suzuki, E. (1966). Anal. Chem. 38:1770-1773. Fraser, R. D. B., and Suzuki, E. (I969). Anal. Chem. 41:37-39. Fringeli, U. P., and Gunthard, Hs. H. (1981). In: Membrane Spectroscopy (E. Grell, Ed.), Mol. Biol. Biochem. Biophys. 31:270-332, Springer-Verlag, New York. Graham, J. M., and Wallach, D. F. H. (1969). Biochim. Biophys. Acta 193:225 227. Graham, J. M., and Wallach, D. F. H. (1971). Bioehim. Biophys. Acta 241:180-194. Griffiths, P. R. (1980). In Analytical Applications of FT-IR to Molecular and Biological Systems (J. R. Durig, Ed.), Reidel, Holland, pp. 11-24. Henderson, R., and Unwin, P. N. T. (1975). Nature 257:28-32. Hladky, S. B., and Haydon, D. A. (1970). Nature 225:451-453. Hladky, S. B., and Haydon, D. A. (1972). Biochim. Biophys. Acta 274:294-312. Jakes, J., and Krimm, S. (1971). Spectrochim. Acta 27A :19-34. Jap, B. K., Maestre, M. F., Hayward, S. B., and Glaeser, R. M. (1983). Biophys. J. 43:81-89. Jaworsky, M., and Mendelsohn, R. (1985). Biochemistry 24:3422-3428. Kauppinen, J. K., Moffat, D. J., Mantsch, H. H., and Cameron, D. G. (1981). Anal. Chem. 53:1454-1457. Kawai, M., and Fasman, G. 0978). J. Am. Chem. Soc. 100:3630-3632.

IR Studies of Membranes

255

Koenig, J. L., and Tabb, D. L. (1980). In Analytical Applications of F T - I R to Molecular and Biological Systems (J. R. Durig, Ed.), Reidel, Holland, pp. 241-255. Kreutz, W., Siebert, F., and Hofmann, K. P. (1984). In Biological Membranes, Vol. 5 (D. Chapman, Ed.), Academic Press, London, pp. 241-277. Krimm, S. (1962). J. Mol. Biol. 4:528 540. Krimm, S., and Bandekar, J. (1980). Biopolymers 19:1-29. Krimm, S., and Dwivedi, A. M. (1982). Science 216:407-408. Lavialle, F., Adams, R. G., and Levin, I. W. (1982). Biochemistry 21:2305-2312. Lee, D. C., Durrani, A. A., and Chapman, D. (1984). Biochim. Biophys. Acta 769:49-56. Lee, D. C., Elliott, D. A., Baldwin, S. A., and Chapman, D. (1985a). Biochem. Soc. Trans. 13:684q585. Lee, D. C., Hayward, J. A., Restall, C. J., and Chapman, D. (1985b). Biochemistry 24:4364-4373. Levin, I. W. (1984). In Advances in lnfrared and Raman Spectroscopy, Vol. 11 (R. J. H. Clark and R. E. Hester, Eds.), Wiley Heyden, Chichester, pp. 1-48. Levin, I. W., Mushayakarara, E., and Bittman, R. (1982). J. Raman Spear. 13:231-234. MacLennan, D. H., Brandl, C. J., Korczak, B., and Green, N. M. (1985). Nature 316:696-700. Maddams, W. F., and Southon, M. J. (1982). Spectrochim. Acta 38A :459-466. Maddy, A. H., and Malcolm, B. R. (1965). Science 150:1616-1618. Mantseh, H. H., Cameron, D. G., Tremblay, P. A., and Kates, M. (1982). Biochim. Biophys. Acta 689:63-72. Mantsch, H. H., Hsi, S. C., Butler, K. W., and Cameron, D. G. (1983). Biochim. Biophys. Acta 728:325 330. Mantsch, H. H., Martin, A., and Cameron, D. G. (1981). Biochemistry 20:3138-3145. Mendelsohn, R., Anderle, G., Jaworsky, M., Mantsch, H. H., and Dluhy, R. A. (1984a). Biochim. Biophys. Acta 775 : 215-224. Mendelsohn, R., Dluhy, R. A., Cameron, D. G., and Mantsch, H. H. (1982). Biophys. J. 37:83-84. Mendelsohn, R., Dluhy, R. A., Crawford, T., and Mantsch, H. H. (1984b). Biochemistry 23:1498-1504. Mendelsohn, R., Dluhy, R., Taraschi, T., Cameron, D. G., and Mantsch, H. H. (1981). Biochemistry 20:66996706. Miyazawa, T. (1960). J. Chem. Phys. 32:1647-1652. Miyazawa, T., and Blout, E. R, (1961). J. Am. Chem. Soc. 83:712-719. Miyazawa, T., Masuda, Y., and Fukushima, K. (1962). J. Polymer Sci. 62 :$62-$64. Mushayakarara, E., and Levin, I. W. (1980). J. Phys. Chem. 86:2324-2327. Nevskaya, N. A., and Chirgardze, Yu. N. (1976). Biopolymers 15:637-648. Papakostidis, G., Zundel, G., and Mehl, E. (1972). Biochim. Biophys. Acta 288:277-281. Pimentel, G. C., and McClellan, A. L. (1960). The Hydrogen Bond, Freeman, San Francisco. Purcell, J. M., and Susi, H. (1984). J. Biochem. Biophys. Methods 9:193 199. Rothschild, K. J., Cantore, W. A., and Marrero, H. (1983). Science 219:1333-1335. Rothschild, K. J., and Clark, N. A. (1979). Biophys. J. 25:473-488. Rothschild, K. J., and Clark, N. A. (1979). Science 204:311-312. Rothschild, K. J., de Grip, W. J., and Sanches, R. (1980). Biochim. Biophys. Acta 596:338-351. Rothschild, K. J., and Marrero, H. (1982). Proc. Natl. Acad. Sci. USA 79:4045-4049. Rothschild, K. J., Roepe, P., Lugtenburg, J., and Pardoen, J. A. (1984). Biochemistry 23:6103-6109. Rothschild, K. J., Sanches, R., Hsiao, T. L., and Clark, N. A. (1980). Biophys. J. 31:53-64. Rothschild, K. J., Zagaeski, M., and Cantore, W. A. (1981). Biochem. Biophys. Res. Comm. 103:483-489. Ruegg, M., Metzger, V., and Susi, H. (1975). Biopolymers 14:1465-1471. Sarges, R., and Witkop, B. (1965). J. Am. Chem. Soc. 87:2011-2020. Siebert, F., and Mantele, W. (1980). Biophys. Struct. Mech. 6:147-164. Siebert, F., Mantele, W., and Gerwert, K. (1983). Eur. J. Biochem. 136:119-127. Siebert, F., Mantele, W.,and Kreutz, W. (1982). FEBS Lett. 141:82-87. Steer, C. J., Vincent, J. S., and Levin, I. W. (1984). J. Biol. Chem. 259:8052-8055. Susi, H. (1969). In Structure and Stability of Biological Macromolecules (S. N. Timasheffand G. D. Fasman, Eds.), Decker, New York, pp. 575-663. Susi, H. (1972). Meth. EnzymoI. 26"381-391. Susi, H., and Byler, D. M. (1983). Biochem. Biophys. Res. Comm. 115:391-397. Susi, H., Timasheff, S. N., and Stevens, L. (1967). J. Biol. Chem. 242:5460-5466. Sychev, S. V., and Ivanov, V. T. (1982). In Membranes and Transport, Vol. 2 (A. N. Martonosi, Ed.), Plenum, New York, pp. 301-307. Thomas, G. J., and Kyogoku, Y. (1977). In lnfrared and Raman Spectroscopy, Part C, Practical Spectroscopy, Vol. 1 (E. G. Braine and J. G. Grasselli, Eds.), Marcel Dekker Inc., N.Y. Timasheff, S. N., and Susi, H. (1966). J. Biol. Chem. 241:249-251. Timasheff, S. N., Susi, H., and Stevens, L. (1967). J. Biol. Chem. 242:5467-5473. Umemura, J., Cameron, D. G., and Mantsch, H. H. (1980). Biochim. Biophys. Acta 602:32-44. Veateh, W. R., Fossel, E. T., and Blout, E. R. (1974). Biochemistry 13:5249-5256.

256

Lee and Chapman

Vincent, J. S., Steer, C. J., and Levin, I. W. (1984). Biochemistry 23:625-631. Wallach, D. F. H., Verma, S. P., and Fookson, J. (1979). Biochim. Biophys. Acta 559:153 208. Wallach, D. F. H., and Zahler, P. H. (1968). Biochim. Biophys. Acta 150:186-193. Weinstein, S., Wallace, B. A., Blout, E. R., Morrow, J. S., and Veatch, W. (1979). Proc. Natl. Acad. Sci. USA 76:42304234. Yang, W.-J., Griffiths, P. R., Byler, D. M., and Susi, H. (1985). Appl. Spectrosc, 39:282-287.

Вам также может понравиться