Influence of the surface potential on the purple membrane structure and activity

ARCHIVES

Vol.

OF BIOCHEMISTRY

284, No. 1, January,

AND

BIOPHYSICS

pp. 1-8,

1991

Influence of the Surface Potential Membrane Structure and Activity Marie-Emmanuel

Riviere,’

Bernard

Arrio,

Robert

on the Purple Pansu,*

and Jean Faure*

C.N.R.S. URA 1116, Bio&erg&ique Membranaire, B&. 433; and T.N.R.S. des Rayonnements, B&t. 350; Universite’ de Paris-Sud, 91405 Orsay, France

Received

January

16, 1990, and in revised

form

August

16, 1990

The role of the divalent cations in the purple membrane is generally understood as the release mechanism of the blue form appearance. The reconstitution by cation addition leads to the recovery of the initial spectral properties. Numerous data are available in the literature on this matter but they are scattered, so that synthetic understanding is not easy. The role of divalent cations was studied through spectrophotometric titrations and electrophoretic mobility measurements, i.e., { potential valuations. Thus, correlations between the bacteriorhodopsin (bR) state and the whole membrane in equilibrium with a definite medium could be made. Deionization was not a fully reversible process. The absence of cations affect neither the rate of the M4i2 formation nor its lifetime but the yield of M&bR was 50% lower. The number of protons involved in the blue to purple transition of both membranes was different and the reconstitution did not erase this difference. It was observed that the number of protons dissociated upon cation addition corresponded approximately to the number of positive charges removed by deionization. Electrophoretic mobility titrations showed large differences between the membranes, illustrating the influence of the surface charge density on the pK of the transition. Taking advantage of the reversible light adaptation process, the reciprocal influence of the charge density of the membrane surface and the retinal state in bR was shown. Specificity of the divalent cations was questioned by a direct substitution of them by imidazol, which left the membrane intact. The partial reversibility of the deionization, the decrease of the Md12 yield, the differences in the titratable protons, and the nonstrict specificity toward divalent cations suggested that another unknown factor could be removed from the membrane. 0 1991 Academic Press, Inc.

r To whom correspondence should be addressed inergetique Membranaire, Institut de Biochimie, de Paris-Sud, 91405 Orsay Cedex, France. 0003-9861/91$3.00 Copyright 0 1991 by Academic Press, All rights of reproduction in any form

Inc. reserved.

URA 75, Physicochimie

at URA 1116, BioBat. 433, Universite

Bacteriorhodopsin (bR)2 is the only protein present in the purple membrane patches (PM) of Halobacterium halobium. This protein binds a retinal moiety through a Schiff base with the lysine 216 residue. The trimers of bR, with specific polar lipids, form an hexagonal lattice in the plane of the membrane. Bacteriorhodopsin functions as a light driven proton pump: when illuminated, the protein undergoes a well-characterized photocycle (1) accompanied by the vectorial proton transport (2). It has been recently shown that PM contains three or four divalent cations/bR (3, 4) which are essential for the stabilization of the purple form toward pH variations. The removal of these cations causes the spectral transition from purple to blue. The blue form is characterized by an absorbance maximum at 602 nm, a low divalent cations content, less than O.l/bR, a disorganized bR lattice (5), and a lower stability (6-8). This last characteristic is illustrated by a higher sensitivity toward bleaching agents (bright light, electric fields, sonication). The main disorder is the absence of the Md12intermediate during the photocycle of bR in the blue membrane (9). The bound divalent cations are also assumedto monitor the pK of the color-controlling residues of bR via the modification of the surface potential (10-13). Addition of cations to the blue form restores the purple color immediately; some reports also describe the color recovery by addition of small hydrophobic-charged or neutralmolecules to the blue membrane (12, 14-16). However, this reversibility is much debated since data from microcalorimetry (6), electrooptical properties (8), and reconstitution from deionized or deionized and bleached mem’ Abbreviations used: bR, bacteriorhodopsin; PM, native purple membrane; BM, native membrane in the blue form (acidified membrane); dePM, deionized membrane in the purple form; deBM, deionized membrane in the blue form; rePM, deionized membrane reconstituted by cation addition, in the purple form; rn2, halflife; nH+, number of protons implied in the spectral transition; bR,,, bR present in the purple form of the membrane (PM); bh2, bR present in the blue form of the membrane (BM). 1

2

RIVI&RE

ET

AL.

branes (7) point out the differences between the original and the cation-depleted membranes. To emphasize the relationship between the cation content and the structure of normal and depleted membranes, we studied pH and cation titrations on the retinal absorbance, which is a local indicator of the protein conformation, and on the surface charge density. These two parameters are sensitive to the whole surface modifications and consequently to the ionic equilibrium changes. An application of this relation between the chromophore environment and the surface was evidenced by the light adaptation process. Formation rates, lifetimes, and yields of the Mdlz intermediate were examined, particularly on the depleted membranes at neutral pH. From these titrations, the partial reversibility of the cation depletion was illustrated and the consequences on the Md12 yield were revealed.

mostated. The M,,, and bRETO rate constants were estimated as the average of 10 flashes on three samples. The Mb12 maximal absorbances were estimated for various laser intensities. The absorbance at 412 nm reflects the quantity of M,12 formed after the flash, i.e., the number of bR molecules that have undergone a photocycle. The maximal absorbance was reached for a laser intensity of 30 mJ and remained constant until 120 mJ. Thus, all the experiments were run with 90 mJ intensity to ensure maximal activation of the bR without any photobleaching of the sample. In addition, under these conditions, the maximal absorbance at 412 nm varied linearly with the bR concentration in the sample. The Md12 concentration was evaluated by extrapolation of its absorbance at t = 0 (e412 = 23,000 M-‘cm-’ (19)).

MATERIALS

RESULTS

AND

METHODS

Electrophoretic

Mobility

Measurements

A laser Doppler velocimetry (LDV) technique was used to determine the
Membrane Preparations

Spectrophotometric Titrations

PM was isolated from H. h&b&m strain S.9 as described (17). Deionized BM was prepared either by a cation exchange column (Chelex 100 mesh from Bio-Rad) according to Kimura et al. (4) or by an EDTA treatment according to Chang et al. (5). All samples were prepared by washing PM extensively with bid&tilled water or 1 mM imidazol, pH 7, when specified, PM concentration was set to 10 pM bR using an t568 = 62,700 M-’ cm-l and the dark-adapted state was checked before experiments. The deionization by Chelex was run by passing the equivalent of 4 ml of PM on 1 ml of stacked Chelex equilibrated in bidistilled water or 1 mM imidazol, pH 7. Reconstitution of deBM was carried out in CaC12 40 pM analytical grade at 20°C for 20 h; the sample was then equilibrated at 10, 20, or 40°C for pH titration.

PM PM was titrated in the pH range from 7 to 2, at different temperatures from 10 to 40°C in water. It is well known that this pH variation is accompanied by an absorption shift from 560 nm (PM) to 602 nm (BM), with an isobestic point around 575 nm. This titration was fully reversible between pH 7 and 3, but below pH 3 it became partially irreversible since aggregation of the membrane suspension occurred. In the range of full reversibility, the purple to blue transition was analyzed in terms of a macroscopic equilibrium driven by the pH of the suspension,

Spectrophotometric Titrations pH titrations. The spectrophotometric titrations were run on a Perkin-Elmer spectrophotometer 555, on 2 ml of membrane suspension, by addition of small amounts (~5 ~1) of HCl or NaOH in order to produce successive 0.10 pH variations. The pH measurements were carried out in the sample compartment with a Radiometer GK2421C combined electrode, within 30 s to avoid KC1 diffusion from the electrode. The spectrophotometer sample compartment was thermostated. The titrations are represented by the variations of the normalized absorbance at 620 nm (A,,,) and by the logarithmic plot of the A,,,/(1 - As*,,) ratio. Calcium binding. The CaC& binding to deBM was carried out under the same conditions as the titration curves, with addition of small amounts of CaC12, 10 PM-1 mM, into 2 ml of 10 PM deBM, 5 min incubation, and pH measurement along the binding process. Experimental data were analyzed by a multilinear regression program. The Ca2+ content of each sample was measured by atomic absorption spectrophotometry using a Perkin-Elmer apparatus (Model 360).

Time-Resolved Experiments A nanosecond laser photolysis apparatus (18) was used for the study of bR photocycle kinetics. The light-induced absorbance changes were initiated by a neodynium YAG pulsed laser Quantel providing flashes of 3 ns duration at 530 nm. The kinetics were recorded at 412 nm or at 570 nm on light-adapted samples; the sample compartment was ther-

PM+H+%BM

PI

Since the bR molecules, intermixed with 14 to 16 lipids per protein, form the PM patches (21), the whole membrane titrations reflect the retinal-protein absorbance variations: bRs6,,+ nH+ G H,bRsoz

121

The dissociation constant for Eq. [2] is

K = [H+l”~bR,,ol/W,o,l~ and the corresponding pK can be written PK

= npH

+ logA620/(As20(max)-A620)

where AC2,,is the absorbance at 620 nm of the blue form. The pK (near 3 at 20°C) was calculated as well from the titration curves at 620 nm (half transition, Fig. la) as from the logarithmic plot of the AezOratio (Fig. lb).

PURPLE

MEMBRANE

STRUCTURE

AND

SURFACE

3

POTENTIAL TABLE

Characteristics Membrane

I

of the pH Titrations of the Different Samples at Three Temperatures PM imidazol

dePM imidazol

Ca’+-rePM HZ0

2.2 5.0

7.5 2.45

6.8 2.95

3.2 3.75

19

5.2 3.25

2.4 5.0

8 3.2

7 3.35

8 3.3

20

40°c nH+ PK

4.8 3.7

3.2 5.4

FIG. 1. Comparative pH titrations of PM (:), Ca*+-rePM (0), and dePM (X) at 20°C: (a) normalized absorbance variation at 620 nm (A,,,) and (b) logarithmic plot of the A,,,/( 1 - A,& ratio.

Ca’+/bR

2.5

0.15

The slope of the logarithmic plot is proportional to the nH+ value: 5 H+/bR were involved in the transition of native PM at 20°C. This value is different from the low nH+ = 1.7 determined by Mowery et al. (22). However, their value may reflect a particular equilibrium between bR molecules and H+ for PM immobilized in polyacrylamide gels, as reported for some other enzymes (23, 24). The pK varied from 2.5 at 10°C to 3.6 at 40°C (Table I). The number of sites involved in the transition remained constant, around 5 t 0.5. Thus, heating did not induce the exposure of additional groups.

nH+ values, 2.5 for dePM and 5 for PM containing 2.5 Ca’+/bR, reinforced this assumption. The pK of the dePM transition was constant between 10 and 20°C (Fig. 2); there was only a slight variation of this pK, between 20 and 40°C as shown in Table I. The nH+ value (2.5) obtained from dePM was constant over the temperature range studied. It appeared that the number of protons titrated during the spectral transition was characteristic of the type of membrane (2.5 for deionized and 5 for native membrane).

PH

PH

Sample

PM HZ0

dePM Hz0

10°C nH+ PK

5.5 2.5

20°C nH+ PK

Note. The total

Reconstituted Deionized

BM

The deBM obtained by ion exchange chromatography (Chelex) presented an isoionic pH near 4 for 10 PM bR. So, the cation depletion of PM consisted of two processes: the cation release itself and a pH variation in agreement with a new equilibrium between ionizable groups and water. As expected, we observed that the isoionic pH varied with bR concentration. Therefore, the titration of deBM between pH 3 and 7 had to be performed with two identical samples: one acidified, the other alcalinized from the initial pH of the suspension. The transition was fully reversible. dePM + H’

% deBM

Ca2+ content

Ca2+-Kd PM

50

0.4

0.15

is expressed

10

in Ca2+ ions per bR molecule.

PM

Calcium binding. The cation binding was studied in the same experimental conditions as the pH titrations. The reconstitution at the isoionic point, pH 4 (10 PM bR),

131

The titration curves obtained from the deionized membrane showed that the purple to blue transition occurred with a higher pK (pK = 5) than in the native membrane. This observation was independent from the deionization process used (Chelex or EDTA). Thus, the presence of divalent cations on the membrane lowered the pK of the spectral transition, suggesting a competition between protons and cations for the protonation sites (11). The

0z 2.5 3 35 1 45 5 5.6 6 1.5 PH

2 z.5 3 3.6 1 4.5 5

0

PH

FIG. 2. Comparative pH titrations of PM (*), Ca*+-rePM (0), and dePM (X) at 1O’C: (a) normalized absorbance variation at 620 nm (A,,) and (b) logarithmic plot of the AcZO/(l - Asgo) ratio.

4

RIVIfiRE

ET

AL.

was chosen in order to check the reversibility in the same conditions and to avoid the contribution of other cations added to alcalinize the suspension (25). Addition of small amounts of Ca2+ to deBM, at its isoionic pH, shifted the maximum absorption from 602 to 560 nm and moderately acidified the suspension, according to the following equilibrium: deBM + Ca2+ % rePM

+ nH+

[41

The term rePM is used to question the reality of the reversibility of the deionization process. As shown in Fig. 3, the binding of Ca2+ on deBM is cooperative as it was described by Zubov et al. (26). The Kd values presented in Table I are in good agreement with the value determined for the high affinity site of deBM for divalent cations (27). The Kd was not greatly sensitive to temperature. It was observed that calcium additions slightly acidified the suspension. At the end of the titration, the pH variation corresponded to the dissociation of 4 to 6 H+/bR. Sodium ions had a lower affinity for deBM (Kd = 4 mM). Thus, after addition of NaOH to obtain dePM at pH 7 for kinetic studies, the total amount of sodium was far from the Kd. rePM titration. If the pK of the purple to blue transition is considered a criterion of the reversibility, the data presented in Table I leads to the following observations: the reconstitution was not achieved at 10°C (Fig. 2); at 2O”C, the reconstitution seemed to be complete (Fig. 1); at a higher temperature, the membrane aggregated in the presence of 2 calcium ions per bR. Now, if we consider the nH+ as an additional criterion of the reversibility, it appears that, at every temperature, the difference was significant between PM and rePM. From these observations, it is clear that the depletion affected the total nH+ value. The so-called reconstituted membrane, obtained by incubation for 20 h in the presence of calcium ions, was quite different from the native PM. Indeed, the binding of Ca2+ ions led, at 20°C only, to the recovery of the initial absorption properties of the retinal. Moreover, the increase of nH+ and the temperature sensitivity clearly evidence the absence of reversibility of the deionization. PM and deBM in Imidazol

Buffer

Deionization of PM in water by ion exchange on Chelex imposes an acid pH on the sample, and consequently the obligation of acid and base additions to obtain the titration curve. To get rid of this constraint, we tried to obtain dePM at a pH value far from the pK, i.e., by using a low concentration pH 7 imidazol buffer. After washing PM in 1 mM imidazol, we observed that the pK of the purple to blue transition was unmodified (Table I). After deionization on Chelex equilibrated with 1 mM pH 7 imidazol,

[CaCW

W)

FIG. 3. Spectrophotometric titrations of the Ca2* PM at 10°C (O), 20°C (*), and 40°C (X).

binding on deionized

the imid-dePM presented a pK approximately identical to the imid-PM. These observations were completed by Ca2+ titrations which brought to the fore the facts that the washing of PM with imidazol buffer reduced the Ca2+ content (0.4 Ca2+/bR) and that ion exchange chromatography lowered this content again (0.25 Ca2+/bR, Table I). These data are in agreement with a substitution of the Ca2+ ions by imidazol. This substitution occurred mainly during the washing and it was achieved after ion exchange chromatography. Unlike the substitution of divalent cations by protons, imidazol binding did not modify the transition of the membrane from purple to blue. Time-Resolved Deionization

Experiments

on the Photocycle

Effect

Since we suspected an irreversible modification of the membrane upon deionization, it was necessary to examine the kinetic properties of the photocycle. It is well known that the M4i2 intermediate does not appear in the course of the photocycle of the blue form of the native or deionized membrane (3, 9, 28, 29). However, few data arise from the literature dealing with the photocycle of the purple state of deionized membrane. An analysis of the lifetimes of the M4i2 intermediate and the recovery of bR5s8 (turnover measurements) were carried out at different pH and temperatures. The lifetimes and yields (M4i2/bR) for light-adapted PM, dePM, and rePM, are presented in Table II. On each type of mempresented a monoexponenbrane, the M4i2 intermediate tial decay (Fig. 4). The typical M4i2 risetime and lifetime were 100 ns and 8 ms, respectively. These values are in agreement with previous data (30) and they were identical on dePM. These observations suggested that PM and dePM have the same photocycle.

PURPLE TABLE Rate

MEMBRANE

STRUCTURE

AND

SURFACE

II

120T

and Yields of the bR Photocycle in PM, and Ca’+-rePM at 2OT and pH 6 in Water

Constants

dePM,

Sample

PM

dePM

0.60

0.31 86 7.6 8.9

Yield [M4121/[hRl M,,, formation 71,2 (nsec) W12 decay r1/2 (msec) bRhTO turnover 71,2 (msec)

94 7.4 7.8

5

POTENTIAL

loo I

(b)

Ca’+-rePM 0.46

100 7.1 8.2

Usually, the proton pumping efficiency is estimated from the photostationary level of M4i2 on PM sheets (31) or by the maximal H+ gradient established in PM proteoliposomes (32). In our experimental conditions and at any temperature assayed, the rise and decay rates of the M 412intermediate were identical on both native and deionized membranes. So, we consider that the analysis of the maximum absorbance increase at 412 nm, induced by a flash, might reveal the proton pumping efficiency of bR. However, the yields, presented in Table II, indicate that dePM is less efficient than native PM. The deionization reduced by a factor 2 the active fraction of the bR molecules at pH 6. This yield variation could not be a consequence of a different light-adapted state of both membranes since we used irradiation conditions leading to their light adaptation (data not shown). Moreover, it has been reported that these membranes contain the same ratio of all-truns retinal (33). The M4i2 yield was not recovered by calcium ion addition. Again, as in the titration experiments, it appeared that deionization was not completely reversible by simple cation addition.

PR FIG. 5. bR yields

PH

Dependence of the Md12 lifetime (c) on pH, in PM (*) and dePM

PR (a), risetime (0).

(b), and

M&

As shown in Fig. 5a, the Md12decay slows down below pH 3.5 for PM and below pH 5 for dePM, at 20°C. The increased Md12 lifetime at acidic pH may reveal the inhibition of the M4i2 protonation. This phenomenon can be explained by the stabilized protonated form of the proton donor, presumably an aspartate residue, according to Henderson et al. (34). In the sameway, we observed faster rates of M4i2 formation, i.e., shorter risetimes, which may be interpreted also by the higher proton affinity of the proton acceptor (see Fig. 5b). As observed on risetimes and lifetimes, the M4i2 yield was greatly sensitive to pH, particularly near the pK of the purple to blue transition of PM and dePM (Fig. 5~). Temperature Effect

pH Effect According to the previous observations, a pH dependence of the rate constants of the photocycle was expected.

The WI2 rise and decay times were faster when the temperature increased from 15 to 30°C at pH 6 and occurred similarly for native and deionized membranes. The

0.10

E c N G

0.08

0.06

5 0 :

0.04

," L 0

0.02

," a

0.00 0

5

10

15

20

Time

FIG.

4.

Typical

record

of the M 412 decay

30

25

35

40

45

50

(msec)

and semi-logarithmic

plot showing

a monoexponential

relation.

6

RIVItiRE 600

T

ET

with the removal of cations from PM. This potential became identical when both membranes were in the blue state. While dePM presented a linear relationship between the < potential and the conformation of bR, an important variation occured on PM, during the purple to blue transition. Thus, part of the cations bound to the PM surface may contribute to lowering the pK of the transition. Consequently, some groups, at the interface, were assumed to be involved in the color control of the membrane. Since we demonstrated that the removal of divalent cations was not completely reversible, it was necessary to show data illustrating the control of the ionization of the membrane surface by the bR conformation, especially in reversible conditions. Light adaptation is a fully reversible process on PM which is characterized by a typical red shift of the retinal absorbance from 558 to 568 nm (37). Again, we observed a different electrophoretic mobility for these two states of bR (Fig. 7) and this was fully reversible for several light-dark cycles.

9

560 4 0.2

---A: 0.3

0.4

Electrophoretic FIG. 6. Electrophoretic dePM (0) at various pH.

mobility

i 0.6

0.5

Mobility and X,,, variations

of PM

AL.

(*) and

dependence of the kinetic parameters of the photocycle on temperature did not reveal any major difference between the native and the deionized form of the membrane.

DISCUSSION Throughout the past decade, numerous reports have arisen dealing with the role of divalent cations on PM structure, in terms of stability, conformational modifications, and retinal Schiff’s base site configuration. The main conclusions have been obtained either on deBM in the blue form or on rePM. However, the exact requirements of PM for cations are still discussed and we attempted to correlate our observations with the data scattered in the literature. Therefore, we examined the equilibrium relationships between ionic species(H+, Ca2+, Na’, imidazol) and PM or dePM at the molecular level (retinal absorbance). The effect of these equilibria on the whole membrane surface was characterized by the j- potential measurements. Moreover, apart from this structural aspect of the ionic interactions, there was a lack of data concerning the photocycle of bR in the purple deionized state compared to the original.

Surface Potential The role of the cations may be highlighted by the electrophoretic mobility measurements using laser doppler velocimetry. This mobility is proportional to the { potential, which is closely related to the ionic environment of the membrane surface (35, 36). As expected, the electrophoretic mobility of PM and dePM was pH dependent and there was a relationship between the surface state and the protein conformation inside the membrane. At each pH, bR may be characterized by its maximum absorption wavelength inside the membrane, i.e., local environment, and the membrane surface by the electrophoretic mobility (Fig. 6). The { potential was higher on dePM than on PM in accordance

b

Mobility

FIG. 7. LDV adapted bR.

spectra

of: (a) dark-adapted

bR,

(b) light-adapted

bR kept

(a.~.)

for 4 h in the dark,

i.e., during

dark

adaptation,

and

(c) light-

PURPLE

MEMBRANE

STRUCTURE

The removal of divalent cations entailed a different membrane characterized by a higher pK of the purple to blue transition. This variation of the pK is reversible at 20°C but the number of protons titrated after reconstitution remained higher. This first observation indicated that the removal of divalent cations was not a reversible process. Evidence for a loss of membrane stability in the absence of cations (5, 6, 38) is now generally admitted. We also observed that cation removal induced a lack of membrane cohesion toward low concentrations of detergents (data not shown). So, the color of the membrane could not be considered as an absolute criterion of reversibility. The pK variation ( ApH = 2) was interpreted as a consequence of the increase of the surface potential. The predominant role of the surface potential is illustrated in Fig. 6. The presence of divalent cations decreased the surface potential and lowered the pK; the deionization had an inverse effect. In other words, when the number of negative charges diminished the pK became acidic and vice versa. This correlation can be expected whenever a surface modification is carried out, in the present report or in the literature, by cation (3, 4) or cationic molecule addition (12, 14, 15), by chemical modification (39-41), by deionization or anionic probe inclusion (12) or by lipid substitution (42-45). The assumption that the surface potential controls the local proton concentration and consequently the retinal absorption was supported by the charge density variation related to retinal isomerization. Indeed, discrete reversible modifications of the bR conformation occurring during the light adaptation affected the surface and the < potential. Thus, the surface charge density, depending on the divalent cations, and the bR conformation are related to the global electrostatic energy, endowing the membrane with a typical transition PK. Moreover, evidence of the cation competition with the protons for sites monitoring the color was brought by the acidification following upon divalent cations addition to a deionized membrane at its isoionic pH. This pH variation corresponded to the dissociation of 4 to 6 H+/bR, a value comparable to the nH+ observed on the dePM titration. The importance of these particular sites was still revealed by the substitution of calcium ions by imidazol. Unlike the deionization, the easy replacement of Ca2+ by imidazol at neutral pH preserved the characteristics of the native membrane. The affinity of small organic molecules for PM have already been reported (14, 15, 46). Our observations illustrated a direct displacement of Ca2+ by an organic molecule. According to the ligand (H+, Ca’+, imidazol), the temperature dependence on the pK was different, showing the complexity of the interactions able to maintain the membrane cohesion. Thus, the specificity of the divalent cations remains questionable since they

AND

SURFACE

7

POTENTIAL

were easily exchanged with imidazol and other organic molecules. All these observations were supplemented by measurements of the bR photocycle. At neutral pH, deionization did not modify the rate constants of the photocycle giving matter of thinking that Ca2+ ions were not directly involved in this process. However, the deionization induced a 50% MhX2 yield decrease and the original photochemical efficiency was partly restored by cation addition. Our results are in agreement with the low level of M4r2 reported by Edgerton et al. (47), on PM obtained by NaOH addition on an acidic BM sample. The formation of the blue form, either by acidification or by deionization, induced an irreversible structural modification of the membrane which cannot be simply reversed by cation addition or alkalinization. The lower efficiency of dePM, at neutral pH, would be a backward effect of the absence of divalent cations and the higher surface potential on the membrane cohesion. This structural change, explainable from a model like the bent membrane suggested by Czege (48), would reduce the positive cooperativity between the bR molecules (49, 50). The irreversibility of the deionization would be a simple consequence of the disorder introduced in the lattice, nevertheless, an other organic molecule would be eliminated during the deionization, since the organic substitutes have a marked hydrophobic character. The divalent cations would exert a dual control on the purple membrane, namely on the cohesion of the cristalline lattice (51) involving lipid protein interactions (50, 52, 53) and on the stability of the purple form towards pH variations, for this reason improving the Md12 yield, doubtless in conjunction with another factor. ACKNOWLEDGMENTS The work on the purple membrane was initiated by Professor L. Packer in our laboratory and followed by a cheerful cooperation with his team and particularly with Dr. E. Hrabeta-Robinson. We are greatly indebted to Dr. J. Belloni and the whole Laboratoire de Physico-Chimie des Rayonnements, URA 75 from C.N.R.S. for hospitality during this work. We thank Mr. J.-F. Delouis for teaching us how to use the Yag laser and assisting during set up operations.

REFERENCES 1. Stoeckenius, W., Lozier, R. H., and Bogomolni, Biophys. Acta 505, 215-278.

R. A. (1978) B&him.

2. For reviews see (1982) Methods in Enzymology (Packer, L., Ed.), Vol. 88, Academic Press. 3. Chang, C.-H., Chen, J.-G., Govindjee, R., and Ebrey, T. G. (1985) Proc. Natl. Acad. Sci. USA 82, 396-400. 4. Kimura, Y., Ikegami, A., and Stoeckenius, Photobiol. 40, 641-646. 5. Heyn, M. P., Dudda, C., Otto, H., Seiff, Biochemistry 28,9166-9172.

W. (1984) F., and Wallat,

Photo&em. I. (1989)

6. Cladera, J., Galisteo, M. L., Duiiach, M., Mateo, P. L., and Padros, E. (1988) Biochim. Biophys. Acta 943, 148-156. 7. Chang, C.-H., Jonas, R., Govindjee, R., and Ebrey, T. G. (1988) Photochem. Photobiol. 47, 261-265.

8

RIMtiRE

8. Tsuji, K., and Hess, B. (1987) Eur. Biophys. J. 15, 231-236. 9. Kobayashi, T., Ohtani, H., Iwai, J.-I., Ikegami, A., and Uchiki, H. (1983) FEBS Lett. 162, 197-200. 10. Chang, C.-H., Jonas, R., Melchiore, S., Govindjee, R., and Ebrey, T. G. (1986) Biophys. J. 49, 731-739. 11. Dufiach, M., Padros, E., Seigneuret, J. Biol. Chem. 263,7555-7559. 12. Kamo, N., Yoshimoto, M., Kobatake, Biophys. Acta 904, 179-186.

M., and Rigaud,

J.-L.

Y., and Itoh, S. (1987)

(1988) Biochim.

13. Gerwert, K., Ganter, U. M., Siebert, F., and Hess, B. (1987) FEBS Z&t. 213, 39-44. 14. Packer, L., and Hrabeta Robinson, E. (1987) in Biophysical Studies of Retinal Proteins, Proceedings of a Conference in the Memory of Laura Eisenstein, pp. 101-109, Univ. Illinois Press. 15. Dencher, N. A., Burghaus, P. A., and Grzesiak, (Ovchinikov, Y., Ed.), VNU Science Press. 16. Der, A., Toth-Boconadi, 229, 313-316.

S. (1987)

R., and Keszthelyi,

L. (1988)

R. A., and Stoeckenius,

Lett.

W. (1975)

Biophys.

K., Eds.),

P., and Combes, D. (1988) in Methods in Enzymology K., Ed.), Vol. 137, pp. 584-598, Academic Press, San

25. Ariki, M., and Lanyi, J. K. (1986) J. Biol. Chem. 261,8167-8174. 26. Zubov, B., Tsuji, K., and Hess, B. (1986) FEBS L&t. 200,226-230.

28. Tsuji,

K., and Rosenheck,

K. (1979)

J.-L., FEBS

and Padros, Lett.

Acad.

T. C., Ismail, K. Z., and El-Sayed, Sci. USA 84, 4094-4098.

M. A. (1987)

30. Ottolenghi, M. (1980) Ado. Photochem. 12,97-200. 31. Kamo, N., Racanelli, T., and Packer, L. (1982) in Methods zymology (Packer, L., Ed.), Vol. 88, pp. 356-359, Academic San Diego. 32. Govindjee, R., Ebrey, 30, 231-242.

T. G., and Crofts,

A. R. (1980)

Proc.

in EnPress,

Biophys.

J.

33. Pande, C., Callender, R. H., Chang, C.-H., and Evrey, T. G. (1985) Photochem. Photobiol. 42, 549-552. 34. Henderson, R., Baldwin, J. M., Ceska, T. A., Zemlin, F., Beckmann, E., and Downing, K. H. (1990) J. Mol. Biol. 213,899-929. 35. Gouy,

G. (1910)

36. McLaughlin,

J. Phys.

S. (1977)

9, 457-468. Curr.

Top. Membr.

Trans.

9, 71-144.

38. Dmiach, M., Padros, E., Muga, A., and Arrondo, J.-L. (1989) Biochemistry 28,8940-8945. A., Takeushi, Y., and Yoshizawa, T. (1982) Biochemistry 39. Maeda, 2 1,4479-4483.

23. Watanabe, I. (1987) in Methods in Enzymology (Mosbach, Vol. 136, pp. 523-530, Academic Press, San Diego.

Rigaud,

Natl.

FEBS

21. Kates, M., Kushwaha, S. C., and Sprott, G. D. (1982) in Methods in Enzymology (Packer, L., Ed.), Vol. 88, pp. 98-111, Academic Press, San Diego. P. C., Lozier, R. H., Chae, Q., Tseng, Y.-W., Taylor, M., 22. Mowery, and Stoeckenius, W. (1979) Biochemistry 18,4100-4107.

M., Seigneuret, M., 27. Dufiach, Biochemistry 26, 1179-1186.

29. Corcoran,

37. Ohno, K., Takeushi, Acta 462,575-582.

20. Riviere, M.-E., Johannin, G., Garnet, D., Molitor, V., Peschek, G. A., and Arrio, B. (1988) in Methods in Enzymology (Packer, L., and Glazer, A. N., Eds.), Vol. 167, pp. 691-700, Academic Press, San Diego.

24. Monsan, (Mosbach, Diego.

‘. AL.

in Proteins

17. Oesterhelt, D., and Stoeckenius, W. (1974) in Methods in Enzymology (Fleischer, S., and Packer, L., Eds.), Vol. 31, pp. 6677678, Academic Press, San Diego. J. A., Castella, M., Faure, J., Vanderauwera, P., and de 18. Delaire, Schryver, F. C. (1984) Nouu. J. Chim. 8, 231-237. 19. Lazier, R. H., Bogomolni, J. 15,955-962.

ET

g&368-372.

E. (1987)

40. Szundi,

Y., and Yoshida,

I., and Stoeckenius,

M. (1977)

W. (1988)

Biochim.

Biophys.

J. 54,

227-232.

41. Hrabeta, E., Robinson, E., and Packer, L. (1984) Biochem. Res. Commun. 122, 1110-1116. K., Hamanaka, T., Mitsiu, M., and Kito, Y. (1984) 42. Hiraki, Biophys. Acta 777, 232-240. 43. Ihara,

K., and Mukohata,

44. Lind, C., Hojeberg, 256,8298-8315. 45. Czbg6,

J. (1987)

Y. (1988)

FEBS

B., and Khorana, Acta Biochim.

Lett.

Biophys. Biochem.

240,148-152.

H. G. (1981)

H. 22,

Biophys.

J. Biol.

Chem.

463-478.

46. Riviere, M.-E., Grand, D., and Arrio, B. (1989) phys. 277, 130-136. 47. Edgerton, M. E., Moore, T. A., and Greenwood, J. 189,413-420.

Arch.

Biochem.

C. (1980)

Bio-

Biochem.

48. Czkgb, J. (1987) Acta Biochim. H. 22, 479-483. R., and Hess, B. (1982) in Methods in Enzymology 49. Korenstein, (Packer, L., Ed.), Vol. 88, pp. 1933201, Academic Press, San Diego. T. (1988) Biophys. J. 54, 197-204. 50. Marinetti, S., Yang, 51. Tristam-Nagle, Biophys. Acta 854,58-66. 52. Hrabeta-Robinson, Biochem. Biophys.

C.-P.,

E., Semadeni, 269.476-484.

53. Roux, M., Seigneuret, 27, 7009-7015.

M.,

and Nagle,

J. F. (1986)

M., and Packer,

and Rigaud,

J.-L.

(1988)

Biochim.

L. (1989)

Arch.

Biochemistry