Biophysics Infrared evidence that the Schiff base of bacteriorhodopsin is protonated : bR 570 and K intermediates ( purple membrane / resonance

It is possible, by using Fourier-transform infrared (FTIR) difference spectroscopy, to detect the conformational changes occurring in both the protein and the chromophore of bacteriorhodopsin during the photocycle. In contrast to Raman spectroscopy, a laser is unnecessary and hence the problem of a perturbing probe beam is eliminated. Furthermore, the relatively high signal-to-noise ratio obtainable with FTIR enables measurements to be made in minutes over a large spectral range. In the study reported in this paper, we used this method to examine the state ofprotonation ofthe retinylidene Schiff base in light-adapted bR570 and in K, the first intermediate in the photocycle. Resonance Raman spectroscopy provides evidence that bR570 is protonated, but these results have been questioned on the basis of theoretical and experimental grounds. FTIR difference spectral changes in the bR570-to-K transition clearly indicate that bR570 contains a protonated Schiff base. In contrast, the K intermediate displays a Schiff base that is altered but still is associated to some degree with a proton. Because the low-temperature FTIR difference spectrum of bR570 and K is similar to the recently reported low-temperature resonance Raman spectra of bR570 and K [Braiman, M. & Mathies, R. (1982) Proc. NatL Acad, Sci USA 79, 403407], we can assign most, but not all, vibrational changes in the bR570-to-K transition to the chromophore. These results are consistent with a simple model of the first step in the photocycle which involves a movement of the Schiff base proton away from a counterion. Understanding how ions move across biological membranes remains a key goal in biology. The membrane protein bacteriorhodopsin (bR) from the purple membrane of Halobacterium halobium (1) offers a unique system for studying light-driven active transport because both the amino acid sequence (2) and the three-dimensional structure to <7 A is known (3). In addition, the retinal chromophore ofbR offers a means to detect distinct steps in the proton transport by measuring changes in absorption of visible light. In order to elucidate the bR proton pump mechanism fully it will be necessary to understand the molecular changes occurring during each step ofthe photocycle. One part of bR that has been accessible to study is the C=N Schiff base which links the retinal chromophore to a lysine residue of the protein (4-6). Resonance Raman spectroscopy, which selectively probes the vibrations of the chromophore, indicates that a deprotonation of the Schiff base occurs by the M412 intermediates (7, 8). Because proton release from the membrane lags slightly behind production of M412 (9), movement of the Schiff base proton is likely to play a key role in proton transport. However, the existence of a protonated Schiff base has been questioned on the basis of theoretical and experimental work (10-13). For example, the possibility has been raised that the exciting light used in resonance Raman studies produces a spectrum that does not reflect the ground state configuration of the chromophore (10). Furthermore, an NMR study (13) on the closely related protein rhodopsin suggests that the Schiff base is deprotonated, in contrast to resonance Raman results. Hence, it is important to develop a new approach to address this question as well as a general technique to probe specific changes occurring in bR during the photocycle. It recently has been demonstrated that chromophore and protein vibrational changes can be detected by Fourier-transform infrared (FTIR) difference spectroscopy (14-16). There are several advantages to this approach. (i) In contrast to resonance Raman spectroscopy, infrared absorption measurements can be made on a sample that is not appreciably driven through the photocycle. This eliminates the necessity ofusing rapid flow and double-beam methods (17) which cope with the effects of the probe beam on the photocycle of bR. Furthermore, the absence of an intense probe beam eliminates the possibility of altering ground states or producing photoproducts. In addition, fluorescence which can hamper Raman measurements, particularly of K (18), does not affect infrared absorption measurements. (ii) Because infrared probes all the modes of purple membrane (19), not just the chromophore vibrations as in resonance Raman spectroscopy, it is possible to examine the conformational changes of the protein during the photocycle. (iii) The intrinsic high signal-to-noise ratio of FTIR spectroscopy allows spectra to be recorded in minutes, compared to hours in conventional Raman measurements. (iv) The direction of bonds can be determined with polarized IR radiation-for example, as demonstrated for the peptide groups in the a-helices of bR (19). In this paper, we report on FTIR difference measurements made at 77 K, a temperature at which the decay of the K intermediate of bR is blocked (20). This intermediate is the first photoproduct of bR with a rise time of <11 psec at room temperature (21). Its absorption is red-shifted to near 630 nm and is related in this respect to the bathorhodopsin primary photoproduct of the visual pigment rhodopsin which is also redshifted (22). There is considerable interest in elucidating the molecular changes that occur during the bR570-to-K transition, the only light-driven step in the photocycle. It is likely that key events occur at this step in both the transduction of the photon energy and the movement of a proton through the membrane. Resonance Raman studies on K (23-25) indicate that the retinal conformation is different from that of bR570. However, the reported spectra all differ in several important features including the C=C and C=N stretching modes. Our FTIR difference data help to resolve this controversy and offer confirming evidence that bR570 contains a protonated Schiff base. Of equal importance, we find that the Schiffbase in K undergoes a major Abbreviations: bR, bacteriorhodopsin; FTIR, Fourier-transform infiared. 4045 The publication costs ofthis article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U. S. C. §1734 solely to indicate this fact. 4046 Biophysics: Rothschild and Marrero alteration which is consistent with an energy storage and proton "switch" mechanism. We also detect vibrational changes which may be associated with peptide groups in the protein. MATERIALS AND METHODS Sample Preparation. Purple membrane was purified from strain S9 (26). Thin films with OD 0.3-0.5 at 570 nm were deposited on 1-cm-diameter AgCl windows (Fisher) by the method of isopotential spin-drying (27, 28). On the basis of Ireeze-fracture electron microscopy and linear dichroism measurements (27), these films have been shown to contain a multilamellar array of membranes that are highly ordered relative to the film plane. Polarized FTIR measurements on these films reveal clearly the transverse orientation ofthe bR a-helices (19). FTIR Measurements. Films were sealed in a humidified cell formed by an "O" ring and a second AgCl window. Humidification was accomplished by exposing the film to H20 or 2H20 vapor at 100% relative humidity for 1 min. The sealed cell was mounted on the copper tail of a low-temperature cryostat (Janis Research, Stoneham, MA) equipped with a KBr and a Ge window (to filter out the spectrometer laser beam). The presence of H20 or 2H20 could be monitored during the experiment by the presence of a strong water vibration at 3,400 or 2,600 cm , respectively. Prior to cooling, the sample was light adapted for 10 min by exposure to a 600-W incandescent light filtered through a yellow Wratten filter and two glass IR filters. A fiber optic cable was used to guide the light into the spectrometer compartment. At 273 K the adapting light was shut off and the sample was cooled to 77 K in the dark. FTIR measurements were made with a Nicolet MX-1 spectrometer. Each interferogram consisted of the average of 196 scans; averaging took a total of 3 min of recording time. The Fourier transform was preformed with triangular apodization. The spectra were recorded with 2-cm-1 resolution and the difference spectra were smoothed by using a 17-point fit which resulted in an effective resolution of8 cm-1. However, the frequency of peaks varied little for each measurement, resulting in an overall variability of <2 cm-1 for peak frequencies. All measurements were repeated at least three times. Measurements of absorption of visible light were made with a Cary 219