- Email: [email protected]
Mechanism of Photosensory Adaptation inHalobacterium salinarium
JMB—MS 365 Cust. Ref. No. CAM 418/94
[SGML] J. Mol. Biol. (1995) 246, 493–499
Mechanism of Photosensory Adaptation in Halobacterium salinarium Wolfgang Marwan, Sergei I. Bibikov, Marco Montrone and Dieter Oesterhelt Max-Planck-Institut fu¨r Biochemie, 82152 Martinsried, Germany
Phototaxis in Halobacterium salinarium is the result of an interplay of sensory rhodopsin excitation and adaptation to the stimulus background. Adaptation to orange light, received by sensory rhodopsin I was probed by measuring the behavioral response of cells to a step-like decrease in intensity. Cells were able to adapt to an intensity range of more than four orders of magnitude. The data were analysed on the basis of theoretical fluence rate response relationships calculated from the photocycle kinetics of the complex of sensory rhodopsin I with its transducer HtrI. Independent of the stimulus background, the cellular response was shown to be a function of the absolute number of photoreceptor complex molecules turned over by the light stimulus. Receptor deactivation was identified as the underlying mechanism of adaptation and was sufficient to account for the experimental results. We suggest that reversible methylation of the transducer protein HtrI provides the chemical mechanism of sensory adaptation in H. salinarium and also explains the different sensitivity of the cells to orange and UV light. Keywords: sensory rhodopsin I; HtrI; phototaxis; methylation; receptor adaptation
Introduction Understanding information processing by cellular signal transduction systems is based on the identification of molecular components and the analysis of their functional properties in the intact organism. Quantitative evaluation of behavioral responses in terms of chemical kinetics provides a way to monitor the activity of signalling molecules in situ. Choosing the appropriate theory it is possible to resolve specifically the characteristics of a single molecular species by measuring the output at the final target. In principle, even behavioral responses of most complex organisms (like mammals) should be accessible to this type of analysis. In the field of bacterial chemotaxis, genetic studies have led to the identification of many, perhaps most, components of the signalling pathway and their sites of interaction in part can be assigned down to the level of single amino acids. At this stage it is of considerable Abbreviations used: BR, bacteriorhodopsin; SRI and SRII, sensory rhodopsin I and II; MCP, methyl-accepting chemotaxis protein; SR, sensory rhodopsin I in the orange light absorbing state; M, sensory rhodopsin I in the UV absorbing state; HR, halorhodopsin; PCR, polymerase chain reaction. 0022–2836/95/090493–07 $08.00/0
analytical interest to resolve the dynamics of their interaction in situ. The same applies to archaebacterial signal transduction. In recent years a transformation system for Halobacterium salinarium has been developed, allowing genetic manipulation of signalling molecules also in this system. However, quantitative evaluation of behavioral experiments on excitation and adaptation processes in bacteria and plants have produced mainly cybernetic models in which molecular events are difficult to assign (for a review, see Galland, 1989). In this work we correlate behavioral experiments on light adaptation to photoreceptor kinetics in H. salinarium. Using caged chemorepellents, eubacterial chemotaxis could be investigated in an analogous way. The swimming behavior of H. salinarium cells is controlled by light and chemotactic stimuli. Responses to light are mediated by three bacterial rhodopsins which function as photoreceptors. Bacteriorhodopsin (BR) in co-operation with a proton motive force sensor is active in the high intensity range and only under conditions where the cellular proton motive force depends on BR’s action (Bibikov et al., 1991, 1993). Sensory rhodopsin I and II (SRI and SRII) cover the low to moderate intensity range and allow the discrimination of orange, UV and blue light (for reviews see Oesterhelt & Marwan, 1993; Spudich, 1993; and references therein). 7 1995 Academic Press Limited
JMB—MS 365 494 SRI, like other bacterial rhodopsins, is embedded in the cytoplasmic membrane by seven membrane-spanning a-helices (Blanck et al., 1989). In intact cells, it forms a stable, photochemically active complex with its transducer HtrI (Krah et al., 1994). HtrI is composed of two membranespanning helices and a large cytoplasmic domain that contains a stretch of 64% identity to the signalling domain of eubacterial methyl-accepting chemotaxis proteins (MCPs; Yao & Spudich, 1992; Ferrando-May et al., 1993). This molecule is thought to transduce light excitation of SRI to the motility-controlling signalling pathway since its deletion results in a loss of SRI-dependent photoresponses (Ferrando-May et al., 1993). By analogy with eubacterial chemotaxis, stimulus-induced reversible protein methylation (presumably of HtrI and its homologues) is thought to be the biochemical mechanism of sensory adaptation in halobacteria (Schimz, 1981, 1882; Bibikov et al., 1982; Spudich et al., 1988, 1989; Alam et al., 1989; Hildebrand & Schimz, 1990; Alam & Hazelbauer, 1991; Nordmann et al., 1994). Interplay between photoreceptor excitation and adaptation in tuning cellular behavior is similar in all three bacterial rhodopsins and explained here taking SRI as an example. In the absence of stimuli, halobacteria spontaneously switch the sense of flagellar rotation. These stochastically occurring events together with Brownian motion of the cell body produce a random walk of individual cells and an even distribution of the cell population in the suspending medium (Marwan et al., 1991). When exposed to a sudden increase in orange light intensity, spontaneous motor switching is suppressed for tens of seconds. Subsequently the cell resumes spontaneous switching with the same probability as before the stimulus. This so-called adaptation, guarantees that continuation of stimulating illumination no longer affects the behavior of a cell (i.e. the stimulus is subjectively quenched). When cells adapted to orange background light are exposed to a step down in intensity, motor switching is induced (deadaptation) and the cells reverse the swimming direction before they adapt to the new, and lower intensity (Hildebrand & Dencher, 1975; Spudich & Stoeckenius, 1979; Hildebrand & Schimz, 1985). Transient prolongation and shortening of single runs by increasing or decreasing light intensity respectively, accumulates the population in orange light (Hildebrand & Dencher, 1975; Stoeckenius et al., 1988). Responses to near UV light (received by SRI373 ) and blue light (received by SRII480 ) are inversed in sign and hence repel the cells from places with radiation of these qualities. Besides the fact that halobacteria adapt to light, nothing is known about the physiology of adaptation. Here we analyse the response of lightadapted cells to a step down in orange light intensity and identify receptor deactivation as the underlying mechanism of sensory adaptation.
Photosensory Adaptation in Halobacterium
Results Theoretical considerations Upon photoexcitation, the dark adapted sensory rhodopsin I molecule (SR) proceeds through a sequence of spectroscopically distinct photocycle intermediates before returning to the initial state. Except for the metastable SRI373 (for the sake of simplicity in the following called M), these intermediates all have very short life times (Bogomolni & Spudich, 1982). As a consequence, M accumulates in orange light and the photostationary concentration of all other intermediates can be neglected since it is very low compared to the concentration of SR and M. Based on the two-state model of the photocycle F 9 SR 9 M, k
the fraction of sensory rhodopsin I molecules converted to M under photostationary conditions is: [M] I = , [SR]t k +I sF
(1)
(for details see Marwan & Oesterhelt, 1990). In this equation, I is the intensity of the orange light, s is the absorption cross section of SR, F is the quantum yield and [SR]t is the total concentration of the photochemically active complex of SRI with HtrI. When [M]/[SR]t is plotted versus the logarithm of the light intensity, a sigmoidal curve is obtained (Figure 1). Changing the values for k, s, F, the curve is shifted in parallel on the abscissa but its shape is invariant. This invariance facilitates the analysis conducted below because the values for these parameters are not required a priori. Although linear transformation of equation (1) is possible, we have chosen a semilogarithmic plot since it is the preferred mode of displaying dose response relationships in the physiological literature. Light-induced formation of M (change in [M] > 0) in dark-adapted cells transiently suppresses spontaneous switching. When the light intensity remains constant (D[M]/Dt = 0) the cells resume spontaneous switching with the same frequency as observed in dark-adapted cells (Hildebrand & Schimz, 1985). In principle, three different mechanisms could mediate this phenomenon called adaptation: (1) the reversalsuppressing activity of the sensory rhodopsin I–HtrI complex could be switched off; (2) the sensitivity of the final target (the flagellar motor switch) or a preceeding step could be adjusted to a permanently altered steady state concentration of transmitter and/or (3) the degree of amplification (gain) within the signal cascade i.e. number of transmitter molecules formed per activated receptor could be changed depending on the number of receptors which are in the signalling state in the adapted cell. Following the nomenclature suggested by Galland
JMB—MS 365 495
Photosensory Adaptation in Halobacterium
(1989), model (1) is called adaptation and models (2) and (3) are called habituation. An appropriate method to quantify the adapted state of a cell is to measure the response to deadaptation (Galland, 1989). In praxi this is achieved by reduction of the initial stimulus, here the initial light intensity to which the cell had been adapted to a new and constant level. In Halobacterium this step down stimulation changes the photoequilibrium of SR and M (Figure 1). Adaptation by receptor deactivation or habituation predicts alternative fluence rate dependencies that are calculated as follows. According to model (1) the light-adapted cell would sense a change in the absolute number of sensory rhodopsin I molecules in the M state, independent of the initial concentration of M. The change in photostationary concentration of M upon a step in light intensity, from an initial value down to a certain fraction of this value, is plotted as a function of the initial intensity in Figure 2. Because of the logarithmic plot, the shape of the curves and their relative position are invariant for changing k/sF. Habituation (i.e. adaptation of the targets or change in gain) by models (2) and (3) would predict that the response to a step in light intensity depends on the steady state concentration of M. In other words activation of one receptor molecule would be less effective if many
receptors were already activated. The response would then be a function of D[M]/[M]o , where D[M] results from the jump in intensity and [M]o is the concentration of M before the step. This relationship is a molecular equivalent of Weber’s law, the empirical description of the fact that a constant response is obtained for a given ratio of stimulus and background (DI/Io = const.; Galland, 1989). Since Weber’s law holds for many sensory systems, we had to take it into account as an alternative to model (1). In Figure 3, D[M]/[M]o is plotted as a function of the initial intensity for different relative step sizes. Again, the shape of the curves is independent from k/sF. Stimulus response curves To assure that the light stimulation regime exclusively excites SR, we selected and used strain M417, a mutant which is highly motile and light sensitive but lacks the three other bacterial rhodopsins (see Materials and Methods). Cells were adapted to orange light of constant initial intensity and stimulated by inserting a neutral density filter into the beam. This step down caused a certain percentage of the cells to reverse their swimming direction which was recorded by a computer-assisted motion analysis system. The fraction of cells reversing depended on the size of the step and on the initial intensity to which the cells
Figure 1. Step down stimulation regime and resulting change in the photostationary concentration of the M intermediate of sensory rhodopsin I. A, Halobacterium cells are adapted to orange light of an initial intensity Io and then stimulated by a sudden drop in intensity down to a new constant level. B, The photostationary concentration of the M intermediate relative to the total cellular concentration of SRI, as obtained by irradiation with monochromatic, orange light, is plotted as a function of log(IsF). The data are calculated for k = 0.36 s−1, s = 2.1 × 10−20 m2 and F = 0.18. Using a logarithmic scale for the abscissa, the shape of the curve is invariant for any values of k, s, F that determine the photocycling rate of SR (compare eqn (1)). The change in the photostationary concentration of M in response to a step down stimulus is indicated by arrows. The initial light intensity is given in photons m−2 s−1.
JMB—MS 365 496
Figure 2. Normalized change in the photostationary concentration of M obtained by a step in orange light intensity down to a defined level. The change caused by inserting a neutral density filter of a given transmittance into the beam was calculated and plotted as a function of the initial intensity. For each curve the transmittance of the corresponding neutral density filter is indicated in percent. In the semilogarithmic plot, the shape of the curves and their relative position is invariant for any value of k. The initial light intensity is given in photons m−2 s−1.
were adapted before stimulation. For each relative step size the response displayed a parabolic dependence on the initial intensity. The maximal response increased with the relative step size applied (Figure 4). The data show that H. salinarium can adapt to light over an intensity range of up to four orders of magnitude. Qualitative comparison of these stimulus response relationships with the curves calculated for the alternative models of adaptation, clearly suggests
Figure 3. Relative change in the photostationary concentration of M obtained by a step in orange light intensity down to a defined level. The relative change in M was calculated as the number of sensory rhodopsin molecules removed from the M state by the step down divided by the number present at the initial intensity. Each curve corresponds to changes obtained by using a neutral density filter of the indicated transmittance.
Photosensory Adaptation in Halobacterium
Figure 4. Response of light-adapted cells to a step down in light intensity. Cells were adapted to an initial intensity of orange light which was then reduced to a certain percentage of this initial value. The fraction of cells responding to this stimulus was then plotted versus the logarithm of the initial intensity. The steps were down to 5% (W), 23% (R), 48% (Q) or 70% (R) of the initial intensity (in photons m−2 s−1). At least 100 cells were evaluated to give 1 data point. The rate of spontaneous reversals was 38%. The continuous lines represent the predicted response as calculated from the equation yielded by the straight line in Figure 5 and the data from Figure 2.
the cellular response to be a function of a change in the absolute number of activated receptor molecules (D[M]/[SR]t ; Figure 2), rather than depending on its relative change (D[M]/[M]o ; Figure 3). Obviously, the response was not simply proportional to D[M]/[SR]t because the shape of the stimulus response curves differed from those of Figure 2. To determine the actual relationship between the number of SRI molecules decaying from the M state and the response of the cells, the maximal change in D[M]/[SR]t for any given step size (as calculated in Figure 2) was plotted versus the corresponding maximal cellular response (obtained from Figure 4) and the result is shown in Figure 5. The empirically obtained dependence in turn was used to predict fluence rate response curves for any step size used in the experiments. The response was calculated taking the thermal decay rate of the M intermediate of the M·HtrIcomplex as measured in intact cells (k = 0.36 s−1; Krah et al., 1994); the result fits the experimental data assuming a quantum yield of 0.18 (Figure 4). For changing k, s, F, the curves would be shifted in parallel along the logarithmic abscissa whereas their shape and their relative position remains invariant. In an analogous way, the response to a step down in light intensity from a given value to zero (light off) was predicted. Light-adapted cells were stimulated
JMB—MS 365 Photosensory Adaptation in Halobacterium
Figure 5. Response of the cells as a function of normalized change in M as obtained from the maxima of the curves plotted in Figures 2 and 4.
by interrupting the actinic light beam through closing an electrically driven shutter. As expected, the experimental data fit the calculated response also in this case (Figure 6).
Discussion Adaptation is mediated by receptor deactivation We have provided evidence that light-adapted halobacterial cells respond to changes in the fraction of the M intermediate (D[M]/[SR]t ) of sensory rhodopsin I. The sensitivity to these changes was independent of the background light intensity and therefore independent of the steady state concen-
Figure 6. Response of Flx3 cells to a step down in light intensity to zero (off response, k = 0.16 s−1 ). The shape of this curve deviates from those of Figure 4 because here, the total amount of M decays in response to light off stimulation.
497 tration of M in the adapted cell. This independence characterizes receptor deactivation as the underlying mechanism of light adaptation. Neither regulation of the efficiency of signal amplification nor adjustment of the sensitivity of the target must be postulated to account for the experimental results. By spectroscopic measurements on intact cells we have shown that SRI forms a stable complex with HtrI (Krah et al., 1994) and that during transition of SR to the M state the complex does not dissociate. M thermally decays with k = 0.36 s−1. Taking this value, a quantum yield of 0.18 is calculated from the stimulus response relationships performed in this study. However, the accuracy of the value for the quantum yield directly depends on the accuracy of k. Since k was shown to slightly depend on the membrane potential (Manor et al., 1988), it is sensitive to the conditions under which the cells are kept.
Reversible methylation of the SRI·HtrI-complex as a molecular mechanism for adaptation Several observations suggest that light adaptation in Halobacterium is based on reversible methylation of HtrI. (1) Adaptation is mediated by switching off activated receptor molecules (shown here). (2) Sensory rhodopsin I excitation is transduced by HtrI, a protein which shares a high sequence homology with the functional domains of eubacterial MCPs (Yao & Spudich, 1992; Ferrando-May et al., 1993). (3) Adaptation and deadaptation to light in halobacteria are associated with a release of methanol from the cell (Alam et al., 1989; Spudich et al., 1989; Nordmann et al., 1994). (4) Inhibition of protein methylation by homocysteine reduces the decay rate of repellent and attractant signals in vivo (Hildebrand & Schimz, 1990). An adaptational mechanism that accounts for these observations is provided by the four state model shown in Figure 7. In the dark-adapted cell, the SRI·HtrI-complex is in its orange light absorbing form, not methylated and inactive. Upon absorption of orange light, the M intermediate is formed which generates an attractant signal i.e. motor switching is suppressed. The attractant signal is quenched by methylation of the complex. If the methylated complex thermally returns to the orange light absorbing intermediate, a repellent signal is generated and switching is induced. This signalling activity of the orange light absorbing complex is switched off by demethylation. The same model can explain the repellent response to UV light. The rate constant for the thermal decay of M as determined spectroscopically (Krah et al., 1994) is compatible with the value for the physiologically active receptor species found in the behavioral experiments. Hence, no dependence of the half-life on the methylation level of HtrI has to be postulated on the basis of these measurements. Because of the thermal decay of M, a continuous turnover of SRI molecules occurs under photostationary conditions
JMB—MS 365 498
Photosensory Adaptation in Halobacterium
Figure 7. Kinetic model for sensory adaptation by transducer methylation. Wavy arrows indicate photochemical reactions, straight arrows thermal reactions. As a minimal model, the scheme neither includes a possible regulation of the methylating or demethylating enzymes nor methylation of multiple sites.
and all four states should be populated. This means that at any time switching suppressing states and switching inducing states are formed which neutralize each other with respect to signalling activity. According to our model, the response to orange light step down depends on the net number of methylated M·HtrI molecules that decay into SR·Htr-CH3 . Quantitative behavioral analysis of specific HtrI mutants will provide further evidence for the role of methylation in sensory adaptation proposed in the model.
signalling molecules that is much higher in the case of UV stimulation than it can be reached by orange light step down (Figure 7). The alternative possibility that UV excitation of M produces a protein conformation that signals more efficiently cannot be excluded but is not necessary to account for the different sensitivity of the cells to UV and orange light step down.
Materials and Methods Mutant selection
Signalling efficiency of HtrI during orange and UV light sensing Upon step down in the intensity of orange light, the M·HtrI complex thermally decays with a rate that is proportional to the difference in the concentration of M under photostationary conditions before and after the step (d[M]/dt = − k([M]o − [M])). The difference in photoequilibria required to cause half maximal response was about one-third of the total population (Figure 4). Taking into account that a halobacterial cell contains about 4000 copies of SRI (Otomo et al., 1989), this turnover amounts to about 4000/3 = 1300 molecules. However, the average response time of the cells of 2s sets an upper practical limit of two half times, corresponding to the decay of 900 molecules that could contribute to the response. There are at least two effects that further reduce this value: (1) the rate of signal formation which is not infinitely fast but in the range of several hundred milliseconds to seconds (Marwan & Oesterhelt, 1990) and (2) receptor deactivation by methylation. The rate of deactivation is not known but might be fast as compared with signal formation. Both effects together might reduce the average number of contributing molecules down to the order of tens. In fact, it has been described that the response to orange light is much less sensitive than that to UV light (Hildebrand & Schimz, 1986). The more sensitive response to UV light emerges from the photochemical decay of M which is much faster than its thermal decay. This leads to a transient concentration of
Halobacterium salinarium strain M416 (BR−, SRII−; Lebert, 1987) was mutagenized with nitrosoguanidine and halorhodopsin-deficient colonies were selected as described (Spudich & Spudich, 1982). Highly motile cells were enriched by several cycles of growth on swarm plates. The finally selected strain M417 (BR−, HR−, SRII−, SRI+ ) was highly sensitive to orange and UV light but showed no response to blue light. PCR analysis of genomic DNA using the primers HR24 (5'-CGTAGTCGGCCACCGGGTATA) and HR13 (5'-TCGCGGTCGGCCAAGTCCAT) revealed an insert in the hop gene of 0.5 kb. Growth and preparation of cells Strain M417 was grown under standard conditions in the dark (Oesterhelt & Krippahl, 1983) for 4 days to obtain maximal SRI content and light sensitivity (Otomo et al., 1989). After growth, cells were diluted with motility medium containing a basal salt solution, 0.1% (w/v) arginine, 5 g/l oxoid peptone (code L37) and 10 mM HEPES (pH 7.0) to 5 × 108 cells per ml. Specimen preparation was as described (Marwan & Oesterhelt, 1990). After incubation for 24 hours in the dark, the measurements were performed on a thermostated stage at 37°C. Recording of behavioral responses The motion of the cells was recorded in non-actinic infrared observation light and raw video data were collected with a frequency of 1 frame per 100 ms using a Motion Analysis frame grabber (Motion Analysis, Santa Rosa, CA). Sequences of 6 s were recorded and the percentage of cells reversing was evaluated without smoothing the tracks. The microscopic setup and the
JMB—MS 365 Photosensory Adaptation in Halobacterium
motion analysis algorithms are described by Marwan & Oesterhelt (1990). Actinic light was emitted from a mercury lamp (Osram HBO, 100 W, AEG, Munich, Germany) and the 578 nm line was selected by heat blocking and cutoff filters (KG3, OG 570; Schott, Mainz, Germany). Cells were adapted to constant orange background light for 30 seconds and then exposed to a step down in light intensity. Data acquisition started one second before the stimulus was delivered. The step down was achieved by inserting an electrically driven neutral density filter into the actinic light beam.
References Alam, M. & Hazelbauer, G. L. (1991). Structural features of methyl-accepting taxis proteins conserved between archaebacteria and eubacteria revealed by antigenic cross reaction. J. Bacteriol. 173, 5837–5842. Alam, M., Lebert, M., Oesterhelt, D. & Hazelbauer, G. L. (1989). Methyl-accepting taxis proteins in Halobacterium halobium. EMBO J. 8, 631–639. Bibikov, S. I., Baryshev, V. A. & Glagolev, A. N. (1982). The roleofmethylationinthetaxisofHalobacteriumhalobium tolightandchemo-effectors.FEBSLetters,146,255–258. Bibikov, S. I., Grishanin, R. N., Marwan, W., Oesterhelt, D. & Skulachev, V. P. (1991). The proton pump bacteriorhodopsin is a photoreceptor for signal transduction in Halobacterium halobium. FEBS Letters 295, 223–226. Bibikov, S. I., Grishanin, R. N., Kaulen, A. D., Marwan, W., Oesterhelt, D. & Skulachev, V. P. (1993). Bacteriorhodopsinisinvolvedinhalobacterialphotoreception. Proc. Nat. Acad. Sci., U.S.A. 90, 9446–9450. Blanck, A., Oesterhelt, D., Ferrando, E., Schegk, E. S. & Lottspeich, F. (1989). Primary structure of sensory rhodopsinIaprokaryoticphotoreceptor.EMBOJ.8,3963–3971. Bogomolni, R. A. & Spudich, J. L. (1982). Identification of a third rhodopsin-like pigment in phototactic Halobacterium halobium. Proc. Nat. Acad. Sci., U.S.A. 79, 6250–6254. Ferrando-May, E., Krah, M., Marwan, W. & Oesterhelt, D. (1993). The methyl-accepting transducer protein HtrI is functionally associated with the photoreceptor sensory rhodopsin I in the archaeon Halobacterium salinarium. EMBO J. 12, 2999–3005. Galland, P. (1989). Photosensory adaptation in plants. Bot. Acta, 102, 11–20. Hildebrand, E. & Dencher, N. (1975). Two photosystems controlling behavioral responses of Halobacterium halobium. Nature (London), 257, 46–48. Hildebrand, E. & Schimz, A. (1985). Behavioral pattern and its sensory control in Halobacterium halobium. In Sensing and Response in Microorganisms (Eisenbach, M. & Balaban, M., eds), pp. 129–142, Elsevier. Hildebrand, E. & Schimz, A. (1986). Integration of photosensory signals in Halobacterium halobium. J. Bacteriol. 167, 305–311. Hildebrand, E. & Schimz, A. (1990). The lifetime of photosensory signals in Halobacterium halobium and its dependence on protein methylation. Biochim. Biophys. Acta, 1052, 96–105. Krah,M.,Marwan,W.,Verme´glio,A.&Oesterhelt,D.(1994). Phototaxis of Halobacterium salinarium requires a signalling complex of sensory rhodopsin I and its
499 methyl-accepting transducer HtrI. EMBO J. 13, 2150–2155. Lebert, M. (1987). Isolierung und Charakterisierung von Verhaltensmutanten von Halobacterium halobium. Diploma thesis, Universita¨t Marburg. Manor, D., Hasselbacher, C. A. & Spudich, J. L. (1988). Membrane potential modulates photocycling rates of bacterial rhodopsins. Biochemistry, 27, 5843–5848. Marwan, W. & Oesterhelt, D. (1990). Quantitation of photochromism of sensory rhodopsin-I by computerized tracking of Halobacterium halobium cells. J. Mol. Biol. 215, 277–285. Marwan, W., Alam, M. & Oesterhelt, D. (1991). Rotation and switching of the flagellar motor assembly in Halobacterium halobium. J. Bacteriol. 173, 1971–1977. Nordmann, B., Lebert, M. R., Alam, M., Nitz, S., Kollmannsberger, H., Oesterhelt, D. & Hazelbauer, G. L. (1994). Identification of volatile forms of methyl groups released by Halobacterium salinarium. J. Biol. Chem. 269, 16449–16454. Oesterhelt, D. & Krippahl, G. (1983). Phototrophic growth of Halobacteria and its use for isolation of photosynthetically deficient mutants. Ann. Microbiol. 134, 137–150. Oesterhelt, D. & Marwan, W. (1993). Signal transduction in halobacteria.InTheBiochemistryofArchaea(Archaebacteria) (Kates, M., et al. eds), pp. 173–187, Elsevier Science Publishers B. V. Amsterdam. Otomo, J., Marwan, W., Oesterhelt, D., Desel, H. & Uhl, R. (1989). Biosynthesis of the two halobacterial light sensors P480 and SR and variation in gain of their signal transduction chains. J. Bacteriol. 171, 2155–2159. Schimz, A. (1981). Methylation of membrane proteins is involved in chemosensory and photosensory behaviour of Halobacterium halobium. FEBS Letters, 125, 205–207. Schimz, A. (1982). Localization of the methylation system involved in sensory behaviour of Halobacterium halobium and its dependence on calcium. FEBS Letters, 139, 283–286. Spudich, J. (1993). Color sensing in the archaea: a eukaryotic-like receptor coupled to a prokaryotic transducer. J. Bacteriol. 175, 7755–7761. Spudich, E. N. & Spudich, J. L. (1982). Control of transmembraneionfluxestoselecthalorhodopsin-deficient and other energy-transduction mutants of Halobacterium halobium. Proc. Nat. Acad. Sci., U.S.A. 79, 4308–4312. Spudich, J. L. & Stoeckenius, W. (1979). Photosensory and chemo-sensory behavior of Halobacterium halobium. Photobiochem. Photobiophys. 1, 43–53. Spudich, E. N., Hasselbacher, C. A. & Spudich, J. L. (1988). Methyl-accepting protein associated with bacterial sensory rhodopsin-I. J. Bacteriol. 170, 4280–4285. Spudich, E. N., Takahashi, T. & Spudich, J. L. (1989). Sensory rhodopsins I and II modulate a methylation/demethylation system in Halobacterium halobium phototaxis. Proc. Nat. Acad. Sci., U.S.A. 86, 7746–7750. Stoeckenius, W., Wolff, E. K. & Hess, B. (1988). A rapid population method for action spectra applied to Halobacterium halobium. J. Bacteriol. 170, 2790–2795. Yao, V. J. & Spudich, J. (1992). Primary structure of an archaebacterial transducer, a methyl-accepting protein associated with sensory rhodopsin I. Proc. Nat. Acad. Sci., U.S.A. 89, 11915–11919.
Edited by I. B. Holland (Received 8 September 1994; accepted 18 November 1994)
[SGML] J. Mol. Biol. (1995) 246, 493–499
Mechanism of Photosensory Adaptation in Halobacterium salinarium Wolfgang Marwan, Sergei I. Bibikov, Marco Montrone and Dieter Oesterhelt Max-Planck-Institut fu¨r Biochemie, 82152 Martinsried, Germany
Phototaxis in Halobacterium salinarium is the result of an interplay of sensory rhodopsin excitation and adaptation to the stimulus background. Adaptation to orange light, received by sensory rhodopsin I was probed by measuring the behavioral response of cells to a step-like decrease in intensity. Cells were able to adapt to an intensity range of more than four orders of magnitude. The data were analysed on the basis of theoretical fluence rate response relationships calculated from the photocycle kinetics of the complex of sensory rhodopsin I with its transducer HtrI. Independent of the stimulus background, the cellular response was shown to be a function of the absolute number of photoreceptor complex molecules turned over by the light stimulus. Receptor deactivation was identified as the underlying mechanism of adaptation and was sufficient to account for the experimental results. We suggest that reversible methylation of the transducer protein HtrI provides the chemical mechanism of sensory adaptation in H. salinarium and also explains the different sensitivity of the cells to orange and UV light. Keywords: sensory rhodopsin I; HtrI; phototaxis; methylation; receptor adaptation
Introduction Understanding information processing by cellular signal transduction systems is based on the identification of molecular components and the analysis of their functional properties in the intact organism. Quantitative evaluation of behavioral responses in terms of chemical kinetics provides a way to monitor the activity of signalling molecules in situ. Choosing the appropriate theory it is possible to resolve specifically the characteristics of a single molecular species by measuring the output at the final target. In principle, even behavioral responses of most complex organisms (like mammals) should be accessible to this type of analysis. In the field of bacterial chemotaxis, genetic studies have led to the identification of many, perhaps most, components of the signalling pathway and their sites of interaction in part can be assigned down to the level of single amino acids. At this stage it is of considerable Abbreviations used: BR, bacteriorhodopsin; SRI and SRII, sensory rhodopsin I and II; MCP, methyl-accepting chemotaxis protein; SR, sensory rhodopsin I in the orange light absorbing state; M, sensory rhodopsin I in the UV absorbing state; HR, halorhodopsin; PCR, polymerase chain reaction. 0022–2836/95/090493–07 $08.00/0
analytical interest to resolve the dynamics of their interaction in situ. The same applies to archaebacterial signal transduction. In recent years a transformation system for Halobacterium salinarium has been developed, allowing genetic manipulation of signalling molecules also in this system. However, quantitative evaluation of behavioral experiments on excitation and adaptation processes in bacteria and plants have produced mainly cybernetic models in which molecular events are difficult to assign (for a review, see Galland, 1989). In this work we correlate behavioral experiments on light adaptation to photoreceptor kinetics in H. salinarium. Using caged chemorepellents, eubacterial chemotaxis could be investigated in an analogous way. The swimming behavior of H. salinarium cells is controlled by light and chemotactic stimuli. Responses to light are mediated by three bacterial rhodopsins which function as photoreceptors. Bacteriorhodopsin (BR) in co-operation with a proton motive force sensor is active in the high intensity range and only under conditions where the cellular proton motive force depends on BR’s action (Bibikov et al., 1991, 1993). Sensory rhodopsin I and II (SRI and SRII) cover the low to moderate intensity range and allow the discrimination of orange, UV and blue light (for reviews see Oesterhelt & Marwan, 1993; Spudich, 1993; and references therein). 7 1995 Academic Press Limited
JMB—MS 365 494 SRI, like other bacterial rhodopsins, is embedded in the cytoplasmic membrane by seven membrane-spanning a-helices (Blanck et al., 1989). In intact cells, it forms a stable, photochemically active complex with its transducer HtrI (Krah et al., 1994). HtrI is composed of two membranespanning helices and a large cytoplasmic domain that contains a stretch of 64% identity to the signalling domain of eubacterial methyl-accepting chemotaxis proteins (MCPs; Yao & Spudich, 1992; Ferrando-May et al., 1993). This molecule is thought to transduce light excitation of SRI to the motility-controlling signalling pathway since its deletion results in a loss of SRI-dependent photoresponses (Ferrando-May et al., 1993). By analogy with eubacterial chemotaxis, stimulus-induced reversible protein methylation (presumably of HtrI and its homologues) is thought to be the biochemical mechanism of sensory adaptation in halobacteria (Schimz, 1981, 1882; Bibikov et al., 1982; Spudich et al., 1988, 1989; Alam et al., 1989; Hildebrand & Schimz, 1990; Alam & Hazelbauer, 1991; Nordmann et al., 1994). Interplay between photoreceptor excitation and adaptation in tuning cellular behavior is similar in all three bacterial rhodopsins and explained here taking SRI as an example. In the absence of stimuli, halobacteria spontaneously switch the sense of flagellar rotation. These stochastically occurring events together with Brownian motion of the cell body produce a random walk of individual cells and an even distribution of the cell population in the suspending medium (Marwan et al., 1991). When exposed to a sudden increase in orange light intensity, spontaneous motor switching is suppressed for tens of seconds. Subsequently the cell resumes spontaneous switching with the same probability as before the stimulus. This so-called adaptation, guarantees that continuation of stimulating illumination no longer affects the behavior of a cell (i.e. the stimulus is subjectively quenched). When cells adapted to orange background light are exposed to a step down in intensity, motor switching is induced (deadaptation) and the cells reverse the swimming direction before they adapt to the new, and lower intensity (Hildebrand & Dencher, 1975; Spudich & Stoeckenius, 1979; Hildebrand & Schimz, 1985). Transient prolongation and shortening of single runs by increasing or decreasing light intensity respectively, accumulates the population in orange light (Hildebrand & Dencher, 1975; Stoeckenius et al., 1988). Responses to near UV light (received by SRI373 ) and blue light (received by SRII480 ) are inversed in sign and hence repel the cells from places with radiation of these qualities. Besides the fact that halobacteria adapt to light, nothing is known about the physiology of adaptation. Here we analyse the response of lightadapted cells to a step down in orange light intensity and identify receptor deactivation as the underlying mechanism of sensory adaptation.
Photosensory Adaptation in Halobacterium
Results Theoretical considerations Upon photoexcitation, the dark adapted sensory rhodopsin I molecule (SR) proceeds through a sequence of spectroscopically distinct photocycle intermediates before returning to the initial state. Except for the metastable SRI373 (for the sake of simplicity in the following called M), these intermediates all have very short life times (Bogomolni & Spudich, 1982). As a consequence, M accumulates in orange light and the photostationary concentration of all other intermediates can be neglected since it is very low compared to the concentration of SR and M. Based on the two-state model of the photocycle F 9 SR 9 M, k
the fraction of sensory rhodopsin I molecules converted to M under photostationary conditions is: [M] I = , [SR]t k +I sF
(1)
(for details see Marwan & Oesterhelt, 1990). In this equation, I is the intensity of the orange light, s is the absorption cross section of SR, F is the quantum yield and [SR]t is the total concentration of the photochemically active complex of SRI with HtrI. When [M]/[SR]t is plotted versus the logarithm of the light intensity, a sigmoidal curve is obtained (Figure 1). Changing the values for k, s, F, the curve is shifted in parallel on the abscissa but its shape is invariant. This invariance facilitates the analysis conducted below because the values for these parameters are not required a priori. Although linear transformation of equation (1) is possible, we have chosen a semilogarithmic plot since it is the preferred mode of displaying dose response relationships in the physiological literature. Light-induced formation of M (change in [M] > 0) in dark-adapted cells transiently suppresses spontaneous switching. When the light intensity remains constant (D[M]/Dt = 0) the cells resume spontaneous switching with the same frequency as observed in dark-adapted cells (Hildebrand & Schimz, 1985). In principle, three different mechanisms could mediate this phenomenon called adaptation: (1) the reversalsuppressing activity of the sensory rhodopsin I–HtrI complex could be switched off; (2) the sensitivity of the final target (the flagellar motor switch) or a preceeding step could be adjusted to a permanently altered steady state concentration of transmitter and/or (3) the degree of amplification (gain) within the signal cascade i.e. number of transmitter molecules formed per activated receptor could be changed depending on the number of receptors which are in the signalling state in the adapted cell. Following the nomenclature suggested by Galland
JMB—MS 365 495
Photosensory Adaptation in Halobacterium
(1989), model (1) is called adaptation and models (2) and (3) are called habituation. An appropriate method to quantify the adapted state of a cell is to measure the response to deadaptation (Galland, 1989). In praxi this is achieved by reduction of the initial stimulus, here the initial light intensity to which the cell had been adapted to a new and constant level. In Halobacterium this step down stimulation changes the photoequilibrium of SR and M (Figure 1). Adaptation by receptor deactivation or habituation predicts alternative fluence rate dependencies that are calculated as follows. According to model (1) the light-adapted cell would sense a change in the absolute number of sensory rhodopsin I molecules in the M state, independent of the initial concentration of M. The change in photostationary concentration of M upon a step in light intensity, from an initial value down to a certain fraction of this value, is plotted as a function of the initial intensity in Figure 2. Because of the logarithmic plot, the shape of the curves and their relative position are invariant for changing k/sF. Habituation (i.e. adaptation of the targets or change in gain) by models (2) and (3) would predict that the response to a step in light intensity depends on the steady state concentration of M. In other words activation of one receptor molecule would be less effective if many
receptors were already activated. The response would then be a function of D[M]/[M]o , where D[M] results from the jump in intensity and [M]o is the concentration of M before the step. This relationship is a molecular equivalent of Weber’s law, the empirical description of the fact that a constant response is obtained for a given ratio of stimulus and background (DI/Io = const.; Galland, 1989). Since Weber’s law holds for many sensory systems, we had to take it into account as an alternative to model (1). In Figure 3, D[M]/[M]o is plotted as a function of the initial intensity for different relative step sizes. Again, the shape of the curves is independent from k/sF. Stimulus response curves To assure that the light stimulation regime exclusively excites SR, we selected and used strain M417, a mutant which is highly motile and light sensitive but lacks the three other bacterial rhodopsins (see Materials and Methods). Cells were adapted to orange light of constant initial intensity and stimulated by inserting a neutral density filter into the beam. This step down caused a certain percentage of the cells to reverse their swimming direction which was recorded by a computer-assisted motion analysis system. The fraction of cells reversing depended on the size of the step and on the initial intensity to which the cells
Figure 1. Step down stimulation regime and resulting change in the photostationary concentration of the M intermediate of sensory rhodopsin I. A, Halobacterium cells are adapted to orange light of an initial intensity Io and then stimulated by a sudden drop in intensity down to a new constant level. B, The photostationary concentration of the M intermediate relative to the total cellular concentration of SRI, as obtained by irradiation with monochromatic, orange light, is plotted as a function of log(IsF). The data are calculated for k = 0.36 s−1, s = 2.1 × 10−20 m2 and F = 0.18. Using a logarithmic scale for the abscissa, the shape of the curve is invariant for any values of k, s, F that determine the photocycling rate of SR (compare eqn (1)). The change in the photostationary concentration of M in response to a step down stimulus is indicated by arrows. The initial light intensity is given in photons m−2 s−1.
JMB—MS 365 496
Figure 2. Normalized change in the photostationary concentration of M obtained by a step in orange light intensity down to a defined level. The change caused by inserting a neutral density filter of a given transmittance into the beam was calculated and plotted as a function of the initial intensity. For each curve the transmittance of the corresponding neutral density filter is indicated in percent. In the semilogarithmic plot, the shape of the curves and their relative position is invariant for any value of k. The initial light intensity is given in photons m−2 s−1.
were adapted before stimulation. For each relative step size the response displayed a parabolic dependence on the initial intensity. The maximal response increased with the relative step size applied (Figure 4). The data show that H. salinarium can adapt to light over an intensity range of up to four orders of magnitude. Qualitative comparison of these stimulus response relationships with the curves calculated for the alternative models of adaptation, clearly suggests
Figure 3. Relative change in the photostationary concentration of M obtained by a step in orange light intensity down to a defined level. The relative change in M was calculated as the number of sensory rhodopsin molecules removed from the M state by the step down divided by the number present at the initial intensity. Each curve corresponds to changes obtained by using a neutral density filter of the indicated transmittance.
Photosensory Adaptation in Halobacterium
Figure 4. Response of light-adapted cells to a step down in light intensity. Cells were adapted to an initial intensity of orange light which was then reduced to a certain percentage of this initial value. The fraction of cells responding to this stimulus was then plotted versus the logarithm of the initial intensity. The steps were down to 5% (W), 23% (R), 48% (Q) or 70% (R) of the initial intensity (in photons m−2 s−1). At least 100 cells were evaluated to give 1 data point. The rate of spontaneous reversals was 38%. The continuous lines represent the predicted response as calculated from the equation yielded by the straight line in Figure 5 and the data from Figure 2.
the cellular response to be a function of a change in the absolute number of activated receptor molecules (D[M]/[SR]t ; Figure 2), rather than depending on its relative change (D[M]/[M]o ; Figure 3). Obviously, the response was not simply proportional to D[M]/[SR]t because the shape of the stimulus response curves differed from those of Figure 2. To determine the actual relationship between the number of SRI molecules decaying from the M state and the response of the cells, the maximal change in D[M]/[SR]t for any given step size (as calculated in Figure 2) was plotted versus the corresponding maximal cellular response (obtained from Figure 4) and the result is shown in Figure 5. The empirically obtained dependence in turn was used to predict fluence rate response curves for any step size used in the experiments. The response was calculated taking the thermal decay rate of the M intermediate of the M·HtrIcomplex as measured in intact cells (k = 0.36 s−1; Krah et al., 1994); the result fits the experimental data assuming a quantum yield of 0.18 (Figure 4). For changing k, s, F, the curves would be shifted in parallel along the logarithmic abscissa whereas their shape and their relative position remains invariant. In an analogous way, the response to a step down in light intensity from a given value to zero (light off) was predicted. Light-adapted cells were stimulated
JMB—MS 365 Photosensory Adaptation in Halobacterium
Figure 5. Response of the cells as a function of normalized change in M as obtained from the maxima of the curves plotted in Figures 2 and 4.
by interrupting the actinic light beam through closing an electrically driven shutter. As expected, the experimental data fit the calculated response also in this case (Figure 6).
Discussion Adaptation is mediated by receptor deactivation We have provided evidence that light-adapted halobacterial cells respond to changes in the fraction of the M intermediate (D[M]/[SR]t ) of sensory rhodopsin I. The sensitivity to these changes was independent of the background light intensity and therefore independent of the steady state concen-
Figure 6. Response of Flx3 cells to a step down in light intensity to zero (off response, k = 0.16 s−1 ). The shape of this curve deviates from those of Figure 4 because here, the total amount of M decays in response to light off stimulation.
497 tration of M in the adapted cell. This independence characterizes receptor deactivation as the underlying mechanism of light adaptation. Neither regulation of the efficiency of signal amplification nor adjustment of the sensitivity of the target must be postulated to account for the experimental results. By spectroscopic measurements on intact cells we have shown that SRI forms a stable complex with HtrI (Krah et al., 1994) and that during transition of SR to the M state the complex does not dissociate. M thermally decays with k = 0.36 s−1. Taking this value, a quantum yield of 0.18 is calculated from the stimulus response relationships performed in this study. However, the accuracy of the value for the quantum yield directly depends on the accuracy of k. Since k was shown to slightly depend on the membrane potential (Manor et al., 1988), it is sensitive to the conditions under which the cells are kept.
Reversible methylation of the SRI·HtrI-complex as a molecular mechanism for adaptation Several observations suggest that light adaptation in Halobacterium is based on reversible methylation of HtrI. (1) Adaptation is mediated by switching off activated receptor molecules (shown here). (2) Sensory rhodopsin I excitation is transduced by HtrI, a protein which shares a high sequence homology with the functional domains of eubacterial MCPs (Yao & Spudich, 1992; Ferrando-May et al., 1993). (3) Adaptation and deadaptation to light in halobacteria are associated with a release of methanol from the cell (Alam et al., 1989; Spudich et al., 1989; Nordmann et al., 1994). (4) Inhibition of protein methylation by homocysteine reduces the decay rate of repellent and attractant signals in vivo (Hildebrand & Schimz, 1990). An adaptational mechanism that accounts for these observations is provided by the four state model shown in Figure 7. In the dark-adapted cell, the SRI·HtrI-complex is in its orange light absorbing form, not methylated and inactive. Upon absorption of orange light, the M intermediate is formed which generates an attractant signal i.e. motor switching is suppressed. The attractant signal is quenched by methylation of the complex. If the methylated complex thermally returns to the orange light absorbing intermediate, a repellent signal is generated and switching is induced. This signalling activity of the orange light absorbing complex is switched off by demethylation. The same model can explain the repellent response to UV light. The rate constant for the thermal decay of M as determined spectroscopically (Krah et al., 1994) is compatible with the value for the physiologically active receptor species found in the behavioral experiments. Hence, no dependence of the half-life on the methylation level of HtrI has to be postulated on the basis of these measurements. Because of the thermal decay of M, a continuous turnover of SRI molecules occurs under photostationary conditions
JMB—MS 365 498
Photosensory Adaptation in Halobacterium
Figure 7. Kinetic model for sensory adaptation by transducer methylation. Wavy arrows indicate photochemical reactions, straight arrows thermal reactions. As a minimal model, the scheme neither includes a possible regulation of the methylating or demethylating enzymes nor methylation of multiple sites.
and all four states should be populated. This means that at any time switching suppressing states and switching inducing states are formed which neutralize each other with respect to signalling activity. According to our model, the response to orange light step down depends on the net number of methylated M·HtrI molecules that decay into SR·Htr-CH3 . Quantitative behavioral analysis of specific HtrI mutants will provide further evidence for the role of methylation in sensory adaptation proposed in the model.
signalling molecules that is much higher in the case of UV stimulation than it can be reached by orange light step down (Figure 7). The alternative possibility that UV excitation of M produces a protein conformation that signals more efficiently cannot be excluded but is not necessary to account for the different sensitivity of the cells to UV and orange light step down.
Materials and Methods Mutant selection
Signalling efficiency of HtrI during orange and UV light sensing Upon step down in the intensity of orange light, the M·HtrI complex thermally decays with a rate that is proportional to the difference in the concentration of M under photostationary conditions before and after the step (d[M]/dt = − k([M]o − [M])). The difference in photoequilibria required to cause half maximal response was about one-third of the total population (Figure 4). Taking into account that a halobacterial cell contains about 4000 copies of SRI (Otomo et al., 1989), this turnover amounts to about 4000/3 = 1300 molecules. However, the average response time of the cells of 2s sets an upper practical limit of two half times, corresponding to the decay of 900 molecules that could contribute to the response. There are at least two effects that further reduce this value: (1) the rate of signal formation which is not infinitely fast but in the range of several hundred milliseconds to seconds (Marwan & Oesterhelt, 1990) and (2) receptor deactivation by methylation. The rate of deactivation is not known but might be fast as compared with signal formation. Both effects together might reduce the average number of contributing molecules down to the order of tens. In fact, it has been described that the response to orange light is much less sensitive than that to UV light (Hildebrand & Schimz, 1986). The more sensitive response to UV light emerges from the photochemical decay of M which is much faster than its thermal decay. This leads to a transient concentration of
Halobacterium salinarium strain M416 (BR−, SRII−; Lebert, 1987) was mutagenized with nitrosoguanidine and halorhodopsin-deficient colonies were selected as described (Spudich & Spudich, 1982). Highly motile cells were enriched by several cycles of growth on swarm plates. The finally selected strain M417 (BR−, HR−, SRII−, SRI+ ) was highly sensitive to orange and UV light but showed no response to blue light. PCR analysis of genomic DNA using the primers HR24 (5'-CGTAGTCGGCCACCGGGTATA) and HR13 (5'-TCGCGGTCGGCCAAGTCCAT) revealed an insert in the hop gene of 0.5 kb. Growth and preparation of cells Strain M417 was grown under standard conditions in the dark (Oesterhelt & Krippahl, 1983) for 4 days to obtain maximal SRI content and light sensitivity (Otomo et al., 1989). After growth, cells were diluted with motility medium containing a basal salt solution, 0.1% (w/v) arginine, 5 g/l oxoid peptone (code L37) and 10 mM HEPES (pH 7.0) to 5 × 108 cells per ml. Specimen preparation was as described (Marwan & Oesterhelt, 1990). After incubation for 24 hours in the dark, the measurements were performed on a thermostated stage at 37°C. Recording of behavioral responses The motion of the cells was recorded in non-actinic infrared observation light and raw video data were collected with a frequency of 1 frame per 100 ms using a Motion Analysis frame grabber (Motion Analysis, Santa Rosa, CA). Sequences of 6 s were recorded and the percentage of cells reversing was evaluated without smoothing the tracks. The microscopic setup and the
JMB—MS 365 Photosensory Adaptation in Halobacterium
motion analysis algorithms are described by Marwan & Oesterhelt (1990). Actinic light was emitted from a mercury lamp (Osram HBO, 100 W, AEG, Munich, Germany) and the 578 nm line was selected by heat blocking and cutoff filters (KG3, OG 570; Schott, Mainz, Germany). Cells were adapted to constant orange background light for 30 seconds and then exposed to a step down in light intensity. Data acquisition started one second before the stimulus was delivered. The step down was achieved by inserting an electrically driven neutral density filter into the actinic light beam.
References Alam, M. & Hazelbauer, G. L. (1991). Structural features of methyl-accepting taxis proteins conserved between archaebacteria and eubacteria revealed by antigenic cross reaction. J. Bacteriol. 173, 5837–5842. Alam, M., Lebert, M., Oesterhelt, D. & Hazelbauer, G. L. (1989). Methyl-accepting taxis proteins in Halobacterium halobium. EMBO J. 8, 631–639. Bibikov, S. I., Baryshev, V. A. & Glagolev, A. N. (1982). The roleofmethylationinthetaxisofHalobacteriumhalobium tolightandchemo-effectors.FEBSLetters,146,255–258. Bibikov, S. I., Grishanin, R. N., Marwan, W., Oesterhelt, D. & Skulachev, V. P. (1991). The proton pump bacteriorhodopsin is a photoreceptor for signal transduction in Halobacterium halobium. FEBS Letters 295, 223–226. Bibikov, S. I., Grishanin, R. N., Kaulen, A. D., Marwan, W., Oesterhelt, D. & Skulachev, V. P. (1993). Bacteriorhodopsinisinvolvedinhalobacterialphotoreception. Proc. Nat. Acad. Sci., U.S.A. 90, 9446–9450. Blanck, A., Oesterhelt, D., Ferrando, E., Schegk, E. S. & Lottspeich, F. (1989). Primary structure of sensory rhodopsinIaprokaryoticphotoreceptor.EMBOJ.8,3963–3971. Bogomolni, R. A. & Spudich, J. L. (1982). Identification of a third rhodopsin-like pigment in phototactic Halobacterium halobium. Proc. Nat. Acad. Sci., U.S.A. 79, 6250–6254. Ferrando-May, E., Krah, M., Marwan, W. & Oesterhelt, D. (1993). The methyl-accepting transducer protein HtrI is functionally associated with the photoreceptor sensory rhodopsin I in the archaeon Halobacterium salinarium. EMBO J. 12, 2999–3005. Galland, P. (1989). Photosensory adaptation in plants. Bot. Acta, 102, 11–20. Hildebrand, E. & Dencher, N. (1975). Two photosystems controlling behavioral responses of Halobacterium halobium. Nature (London), 257, 46–48. Hildebrand, E. & Schimz, A. (1985). Behavioral pattern and its sensory control in Halobacterium halobium. In Sensing and Response in Microorganisms (Eisenbach, M. & Balaban, M., eds), pp. 129–142, Elsevier. Hildebrand, E. & Schimz, A. (1986). Integration of photosensory signals in Halobacterium halobium. J. Bacteriol. 167, 305–311. Hildebrand, E. & Schimz, A. (1990). The lifetime of photosensory signals in Halobacterium halobium and its dependence on protein methylation. Biochim. Biophys. Acta, 1052, 96–105. Krah,M.,Marwan,W.,Verme´glio,A.&Oesterhelt,D.(1994). Phototaxis of Halobacterium salinarium requires a signalling complex of sensory rhodopsin I and its
499 methyl-accepting transducer HtrI. EMBO J. 13, 2150–2155. Lebert, M. (1987). Isolierung und Charakterisierung von Verhaltensmutanten von Halobacterium halobium. Diploma thesis, Universita¨t Marburg. Manor, D., Hasselbacher, C. A. & Spudich, J. L. (1988). Membrane potential modulates photocycling rates of bacterial rhodopsins. Biochemistry, 27, 5843–5848. Marwan, W. & Oesterhelt, D. (1990). Quantitation of photochromism of sensory rhodopsin-I by computerized tracking of Halobacterium halobium cells. J. Mol. Biol. 215, 277–285. Marwan, W., Alam, M. & Oesterhelt, D. (1991). Rotation and switching of the flagellar motor assembly in Halobacterium halobium. J. Bacteriol. 173, 1971–1977. Nordmann, B., Lebert, M. R., Alam, M., Nitz, S., Kollmannsberger, H., Oesterhelt, D. & Hazelbauer, G. L. (1994). Identification of volatile forms of methyl groups released by Halobacterium salinarium. J. Biol. Chem. 269, 16449–16454. Oesterhelt, D. & Krippahl, G. (1983). Phototrophic growth of Halobacteria and its use for isolation of photosynthetically deficient mutants. Ann. Microbiol. 134, 137–150. Oesterhelt, D. & Marwan, W. (1993). Signal transduction in halobacteria.InTheBiochemistryofArchaea(Archaebacteria) (Kates, M., et al. eds), pp. 173–187, Elsevier Science Publishers B. V. Amsterdam. Otomo, J., Marwan, W., Oesterhelt, D., Desel, H. & Uhl, R. (1989). Biosynthesis of the two halobacterial light sensors P480 and SR and variation in gain of their signal transduction chains. J. Bacteriol. 171, 2155–2159. Schimz, A. (1981). Methylation of membrane proteins is involved in chemosensory and photosensory behaviour of Halobacterium halobium. FEBS Letters, 125, 205–207. Schimz, A. (1982). Localization of the methylation system involved in sensory behaviour of Halobacterium halobium and its dependence on calcium. FEBS Letters, 139, 283–286. Spudich, J. (1993). Color sensing in the archaea: a eukaryotic-like receptor coupled to a prokaryotic transducer. J. Bacteriol. 175, 7755–7761. Spudich, E. N. & Spudich, J. L. (1982). Control of transmembraneionfluxestoselecthalorhodopsin-deficient and other energy-transduction mutants of Halobacterium halobium. Proc. Nat. Acad. Sci., U.S.A. 79, 4308–4312. Spudich, J. L. & Stoeckenius, W. (1979). Photosensory and chemo-sensory behavior of Halobacterium halobium. Photobiochem. Photobiophys. 1, 43–53. Spudich, E. N., Hasselbacher, C. A. & Spudich, J. L. (1988). Methyl-accepting protein associated with bacterial sensory rhodopsin-I. J. Bacteriol. 170, 4280–4285. Spudich, E. N., Takahashi, T. & Spudich, J. L. (1989). Sensory rhodopsins I and II modulate a methylation/demethylation system in Halobacterium halobium phototaxis. Proc. Nat. Acad. Sci., U.S.A. 86, 7746–7750. Stoeckenius, W., Wolff, E. K. & Hess, B. (1988). A rapid population method for action spectra applied to Halobacterium halobium. J. Bacteriol. 170, 2790–2795. Yao, V. J. & Spudich, J. (1992). Primary structure of an archaebacterial transducer, a methyl-accepting protein associated with sensory rhodopsin I. Proc. Nat. Acad. Sci., U.S.A. 89, 11915–11919.
Edited by I. B. Holland (Received 8 September 1994; accepted 18 November 1994)