SI. FI L phosphoruses) were obtained from Sigma. Phenol red was from. Fisher. Neutral red was from Eastman Kodak and was recrys- tallized from 95% ethanol.
Proc. Nati Acad. Sci. USA Vol. 78, No. 12, pp. 7422-7425, December 1981
Biochemistry
Method for studying kinetics of light-induced transport across membranes (membrane transport/phase spectrophotometry/bacteriorhodopsin/chiloroplast thylakoids)
T. G. DEWEY AND GORDON G. HAMMES Department of Chemistry, Cornell University, Ithaca, New York 14853
Contributed by Gordon G. Hammes, August 28, 1981
ABSTRACT A technique based on phase spectrophotometry is described for studying the rates of elementary processes associated with the light-driven transport of ions and molecules across membranes. The light-induced pumping of protons by bacteriorhodopsin reconstituted into phospholipid vesicles and by chloroplast thylakoids has been studied to illustrate the potential of this technique. The exciting light is modulated by a mechanical chopper over the frequency range 5 Hz to 2 kHz. The internal pH of the membrane vesicles is modulated at the same frequency as the exciting light but differs in phase because of the finite rate of proton pumping. Measurement of this phase difference or of the frequency dispersion of the amplitude of the internal pH modulation is accomplished by use of a lock-in amplifier. The results can be interpreted in terms of relaxation times characterizing the chemical steps in proton pumping. The shortest relaxation time that can be measured is about 50 ,usec, although the time resolution could be easily extended by use of faster light chopping techniques. At pH 8.0, two relaxation processes are associated with proton pumping by bacteriorhodopsin reconstituted into phospholipid vesicles; the relaxation times are 2 and 28 msec. Two relaxation processes also were found with chloroplast thylakoids at pH 7.8, with relaxation times of 2 and 16 msec. The former can be associated with photosystem II and the latter, with photosystem I.
The transport of ions and molecules across membranes is a critical aspect of many physiological processes. To develop a molecular mechanism for membrane transport, kinetic information must be obtained about the individual (elementary) steps in the transport process. Unfortunately, relatively few methods are available for studying such reactions. Stopped-flow and stoppedflow quench techniques have been used to study transport in vesicular systems (1, 2), and flash photolysis can be used to investigate light-driven transport in photosynthetic systems. (3). Although these methods have many advantages, they also have a serious disadvantage. Because of the small internal volume of most membrane systems, the species being transported is difficult to buffer. As a result, during ion transport the chemosmotic potential usually is continuously varying. This makes the interpretation of kinetic parameters difficult. Ideally the transport fluxes should be studied at a constant chemical potential for both the external and internal phases of the membrane system. In this paper, a periodic perturbation technique is presented for measuring light-driven ion fluxes under steady-state conditions such that the states of the external and internal phases are well defined. This technique has been used to characterize proton pumping in chloroplast thylakoids and in phospholipid vesicles containing bacteriorhodopsin.
With this technique, the actinic light driving the protonpumping photosystem is periodically modulated. This results in the establishment of a transmembrane pH gradient that has a small modulation about its steady-state value. The absorbance of an internal pH indicator is monitored with a second beam of light. This absorbance will be modulated at the same frequency as the actinic light but will be phase shifted due to the finite response time of the photosystem. This. phase lag can be determined by the use of phase-sensitive detection. By varying the modulating frequency, the chemical relaxation times for the system can be determined. This is accomplished by determining the phase lag directly or from the dependence of the signal amplitude on the modulation frequency (4). This application of phase spectrophotometry is similar to the techniques used to measure fluorescent lifetimes or short-lived intermediates produced by photolysis (5, 6). An important feature of this method is that proton translocation rates are measured about a constant pH gradient. This steady state is reached when the light-driven inward flux is balanced by the passive leak rate. To achieve large pH gradients and modulated amplitudes, the external solution is strongly buffered, and the internal solution is weakly buffered. This is accomplished by using a membrane-impermeant buffer, such as polyphosphate, in the external solution. The pH gradient can then be varied by changing the actinic light intensity, and the external pH can be easily controlled with the buffer. Thus, the relaxation times characterizing proton transport can be determined at a constant, well-defined, chemosmotic potential.
EXPERIMENTAL Phase Spectrophotometer. A diagram of the spectrophotometer is shown in Fig. 1. The actinic light was a Kodak 600 H slide projector equipped with a ELH 300-W lamp. To avoid 60-cycle ripple and to increase the stability of this source, the projector was rewired to run the lamp off a Kepco SM-75.dc power supply. This actinic light was filtered with a Corning CS 2-63 filter for chloroplast thylakoids and a Corning CS 3-69 filter for bacteriorhodopsin vesicles. In both cases, a Corning CS 1-75 UV-IR filter was also present. Light intensity was varied by using Ditric neutral density filters. The light was chopped by using a Bentham 218 optical chopper from Ithaco. The control unit for this mechanical chopper allowed the frequency to be continuously varied from 5 Hz to 2 kHz. The chopper generated a squarewave signal that was used as a reference signal for the phasesensitive detection. The probe beam, which is at right angle to the actinic light, used a 100-W GE Quartzline tungsten lamp powered by a Kepco ATE55-lOM dc power supply. A Bausch The publication costs ofthis article were defrayed in part by page charge and Lomb 250-cm monochromator with a UV-Vis grating was payment. This article must therefore be hereby marked "advertiseused to select the appropriate wavelength for this beam. Spatial ment" in accordance with 18 U. S. C. §1734 solely to indicate this fact. 7422
Proc. Nad Acad. Sci. USA 78 (1981)
Biochemistry: Dewey and Hammes
7423
ofthe phototube. filterwas in front placing a Ditric narrow at 435 nm monitored pH indicator of the bandpass The absorbance for phenol red and at either 520 or 435 nm for neutral red. The signal from the photomultiplier was amplified by using an Ithaco 393 lock-in amplifier. This lock-in amplifier allowed the simulM taneous measurement of the signal amplitude and of the phase relative to the chopped beam. All experiments were carried out at room temperature, about 230C. Materials and Methods. Octyl glucoside (octyl -O-glucowas purchased from Calbiochem; valinomycin, bopyranoside) SC Lvine serum albumin, carbonylcyanide p-trifluoromethoxyphen* J ylhydrazone, and sodium phosphate glass (chain length of five phosphoruses) were obtained from Sigma. Phenol red was from SI FI L Fisher. Neutral red was from Eastman Kodak and was recrystallized from 95% ethanol. The asolectin (95% purified soybean phospholipid) was obtained from Associated Concentrates. Pyocyanine, methyl viologen, dimethylquinone, dibromothymoC quinone, and durohydroquinone were gifts from Richard F2 Reference McCarty. Duroquinone was prepared by sodium borohydride reduction ofdurohydroquinone in a 1:1 ethanol/ethylene glycol PM \/PM solution. The solution was stabilized and excess borohydride was decomposed by acidification with HCl. All other reagents T LOC K-~I N J were of high grade commercial quality; solutions were made AMP with deionized distilled water. Cell suspensions of Halobacterium halobium S-9 were a gift from Russell MacDonald. Bacteriorhodopsin was purified from these suspensions using a sucrose step gradient (7). The conFIG. 1. Diagram of the phase spectrophotometer. S , actinic of bacteriorhodopsin was determined by the absorcentration Ditric narrow source; S2, probe source; F1, Corning cutoff filters; F2, bance at 560 nm, assuming an extinction coefficient of 54,600 bandpass filter; L, 7.6-cm (3 in.) focal length lens; C, mechanical chopM'1cm'1 and a molecular weight of 26,000. Chloroplast thyper which generates a reference signal; M, monochromator; SC, sample lakoids were prepared from market spinach by the method of cell; PM, photomultiplier whose signal goes to the lock-in amplifier. McCarty and Racker (8). Chlorophyll concentrations were determined from the absorbance at 663 and 645 nm (9). overlap of the two beams in the sample cell was necessary to Phospholipid vesicles were prepared by sonicating asolectin maximize the signal. (40 mg/ml) to clarity in 0.15 M KCl/8 mM Na N-trisThe sample cell was a standard 1-cm quartz cuvette with four (hydroxymethyl)methylglycine, pH 8.0/2 mM phenol red. optical faces. The absorbance of the sample was detected with Bacteriorhodopsin (0.5-1.0 mg/ml) was reconstituted into the a EMI 9635 end-on phototube. Scattered light was filtered by S2
010
10
812 0~~~~~ 1 1
1 1 11 111
111
1I
0~ ~~~~~~~~~~~~~~O.
2
0.1
0.2
0.5
1.0
2.0
5.0
10
20
50
w, rad/s x 10-2 FIG. 2. Plot of the observed amplitude, AO., versus the chopping frequency, w, for bacteriorhodopsin reconstituted into asolectin vesicles at pH 8.0 in 0.15 M KC1/8 mM Na N-tris(hydroxymethyl)methylglycine/20 mM sodium polyphosphate. Phenol red was the pH indicator. The bacteriorhodopsin concentration was 0.2 mg/ml, and the asolectin concentration was 5 mg/ml. The line is the nonlinear least-squares fit to Eq. 1 with n = 2.
7424
Biochemistry: Dewey and Hammes
Proc. Natl. Acad. Sci. USA 78 (1981)
Table 1. Lifetimes and amplitudes for proton pumping
Reconstituted bacteriorhodopsin Full light Half light Chloroplast thylakoids Photosystem I and II Photosystem I
72) msec
A,/(A, + A2)
A2/(A1 + A2)
AV
2.0 2.4
28 29
0.24 0.36
0.76 0.64
3160 1140
1.7 -
16 16
0.25 1.0
0.75 -
130 -
vesicles (20 mg of asolectin per ml) by incubation on ice for 10 min in the presence of 1.25% octyl glucoside. To remove the octyl glucoside, the vesicles were diluted 1:25 with buffer and centrifuged for 30 min at 40,000 rpm in a Beckman 60 Ti rotor. Phenol red was included in the dilution buffer to prevent loss of the dye from the inside of the vesicles. The pellet was taken up in buffer to give an asolectin concentration of =40 mg/ml. The vesicles were then passed through a Sephadex G-50 column to remove excess pH indicator from the external solution and were diluted into 20 mM sodium phosphate glass/0. 15 M KCl/ 8 mM Na N-tris(hydroxymethyl)methylglycine, pH 8.0. This procedure allowed the external solution to be strongly buffered and the internal solution to be weakly buffered. The final concentration of asolectin used in the experiments was =5 mg/ ml. Selective introduction of a pH indicator into the internal volume ofchloroplast thylakoids is difficult. Following the technique of Auslander and Junge (3), a permeant pH indicator,. neutral red, was used, and the external solution was buffered with either bovine serum albumin or sodium polyphosphate glass, neither ofwhich crosses the thylakoid membrane. Similar results were obtained with both buffers. 1A
A1 + A2,
T1) msec
RESULTS AND DISCUSSION The amplitudes ofthe modulated absorbance changes ofthe pH indicator, Aobs, are related to the frequency of the modulation, w, and the relaxation times for the n chemical processes, ri, by the relationship n
Aobs
0o
W2
Ti2)[/2[1]
i=l
in which Ai are the amplitudes of the individual relaxation processes (4). The observed signal amplitude is shown as a function of the modulation frequency for phospholipid vesicles containing bacteriorhodopsin in Fig. 2. These data cannot be described by a single relaxation process but conform well to the assumption of two relaxation processes (n = 2). The best fit parameters are given in Table 1, and the curve in Fig. 2 has been calculated with these parameters and Eq. 1. The parameters obtained with the actinic light intensity reduced by a factor of 2 also are given in Table 1. The relaxation times are unchanged, but the relative amplitudes of the two processes are somewhat altered. The absolute amplitudes naturally are smaller. The addition of valinomycin (1 ,uM), which eliminates the membrane potential, had
11II111II11
I
Ai/(l +
=>
1 I I I11 I I
I
I
I
II I
O
12 [
0
0 00
10
0000
CMK0
8 r-
0
000
x
00
0
61
0
¢
0
o
4 o 0s 0,. 2
I 0.1
0.2
1 IT 1 t1 1 11I
0.5
1 I I111 III
2.0
1.0 w,
rad/s
5.0 x
10
I 20
l I II
50
10-2
FIG. 3. Plot of the observed amplitude, Ab., versus the chopping frequency, w, for chloroplast thylakoids at pH 7.8 in 20 mM KC1/ 10 mM MgCl2/ 10 mM NaN-tris(hydroxymethyl)methylglycine/20 mM sodium polyphosphate/5 uM neutral red/i15 M pyocyanine. The chlorophyll concentration was 5pg/ml. The line is the nonlinear least-squares fit to Eq. 1 with n = 2.
Biochemistry: Dewey and Hammes no effect on the observed amplitudes. Because bacteriorhodopsin reconstituted into phospholipid vesicles does not develop large membrane potentials while pumping protons, * valinomycin would not be expected to alter the observed results. The uncoupler carbonylcyanide p-trifluoromethoxyphenylhydrazone (100 ALM) decreased the amplitude by about 80%. Kinetic studies of the rate of formation of the M412 intermediate in the bacteriorhodopsin photocycle indicate biphasic behavior on a time scale similar to the observed relaxation times (10). This intermediate is postulated to be involved in the proton release step of the pumping mechanism (11), and our experiment may be monitoring this step. Multiple relaxation processes also could arise from the inhomogeneous behavior of bacteriorhodopsin in the reconstituted system. Further experiments are required to better define the kinetic processes observed. The results obtained with the chloroplast thylakoids in the presence of pyocyanine are shown in Fig. 3. Under the conditions used, both photosystems I and II contribute to the proton pumping (12). A low concentration of the pH indicator (5 ,uM) was used to avoid uncoupling the system. Consequently, the signal amplitude is much smaller than observed with reconstituted bacteriorhodopsin. The data were analyzed according to Eq. 1. Two relaxation processes again were required to fit the data; the best fit parameters are given in Table 1, and the curve in Fig. 3 has been calculated with these parameters and Eq. 1. Photosystem I was examined separately by substituting duroquinone (0.5 mM) and methyl viologen (0.1 mM) for pyocyanine. Only a single relaxation process was observed. In this case, the relaxation time can be calculated directly from the phase lag angle, 4, according to the relationship tan 4 = W. The relaxation time calculated from this relationship is independent of frequency and is given in Table 1. Experiments designed to measure only the properties of photosystem II by using 1 /iM dibromothymoquinone, 0.5 mM dimethylquinone, and 1 mM potassium ferricyanide were only partially successful *
Casadio, R. & Dencher, N.A., European Molecular Biology Organization Workshop on Halophilic Microorganisms, Island of Ischia, Italy, July 1981.
Proc. NatL Acad. Sci. USA 78 (1981)
7425
because of the instability of the system, but the characteristic relaxation time clearly was less than 10 msec. Therefore, the slower relaxation process is associated with photosystem I and the faster process, with photosystem II. These results are in general agreement with those obtained previously by using flash photolysis (3). The results presented here demonstrate the utility of this type of phase spectrophotometry. This method can be used for any light-driven transport system, providing a suitable indicator can be found. Even if a light-driven pump is not directly coupled to membrane transport, the internal concentration might be modulated by chromophoric species that dissociate in the excited state. The time scale accessible to this technique can be greatly extended by using opto-acoustic modulators instead ofmechanical chopping of the actinic light. Ultimately, the time resolution of the systems studied here is limited by the rates of protolytic reactions of the indicator; such reactions usually have characteristic time constants of a few microseconds. The technique described here has the potential for becoming a versatile method of studying the transport of ions across memblanes in chemosmotic systems. We are grateful to Prof. Richard McCarty for helpful discussions. This work was supported by Grant GM 13292 from the National Institutes of Health. 1. 2. 3. 4.
Biegel, C. M. & Gould, J. M. (1981) Biochemistry 20, 3474-3479. Cash, D. J. & Hess, G. P. (1981) AnaL Biochem. 112, 39-51. Auslander, W. & Junge, W. (1975) FEBS Lett. 59, 310-315. Eigen, M. & de Maeyer, L. (1974) in Techniques of Chemistry, ed. Hammes, G. G. (Wiley Interscience, New York), Vol. 6, pp.
63-146. 5. Kaye, W. & West, D. (1967) in Fluorescence-Theory, Instrumentation and Practice, ed. Guilbault, G. G. (Dekker, New York), pp. 255-273. 6. Slifkin, M. A. & Walmsley, R. H. (1970)J. Phys. E 3, 160-162. 7. Becher, B. M. & Cassim, J. Y. (1975) Prep. Biochem. 5, 161-178. 8. McCarty, R. E. & Racker, E. (1967) J. Biol. Chem. 242, 3435-3439. 9. Arnon, D. I. (1949) Plant Physiot 24, 1-15. 10. Eisenbach, M., Bakker, E. P., Korenstein, R. & Caplan, S. R. (1976) FEBS Lett. 71, 228-232. 11. Lewis, A., Spoonhower, J., Bogomolni, R. A., Lozier, R. H. & Stoeckenius, W. (1974) Proc. Natl Acad. Sci. USA 71, 4462-4466. 12. Trebst, A. (1974) Annu. Rev. Plant Physiol 25, 423-458.
