 Wolfgang Marwan is Heisenberg-Fellow at
the Institut fur Biologie III, Albert-Ludwigs-Universitat, Schanzlestrab e 1, 79104
Freiburg, Germany. |
 Dieter Oesterhelt is Director of the
Max-Planck-Institut fur Biochemie, Am Klopferspitz 18A, 82152 Martinsried, Germany. |
Archaeal Vision and Bacterial Smelling
Two-component signalling and fumarate concentrations control the swimming behavior of several distinctive types of microorganisms
Wolfgang Marwan and Dieter Oesterhelt
Flagellated bacteria, whether pathogenic or not, escape harsh conditions by swimming away from them, and they also just as actively seek favorable sites. The flagellar apparatus is regulated by a well-tuned and robust sensory signalling network which enables such organisms to direct their movements while simultaneously monitoring a variety of environmental stimuli within seconds. When not stimulated by environmental signals, flagellated microbial cells perform a three-dimensional random walk by intermittently changing their swimming direction. However, when such a cell senses an attractant stimulus, it continues along a path to approach the stimulus source. And, in contrast, if it senses a repellent (negative) stimulus, the cell immediately reorients to explore a new direction. This stimulus-biased random process of reorientation is called taxis and may occur in response to nutrients, oxygen, light, pH, temperature, and other environmental stimuli.
During the past few years, investigators have come to realize that the molecular modules that control archaeal and bacterial swimming and related behaviors are similar, but not identical. Here we compare the biochemistry of archaeal vision, namely phototaxis, and bacterial smelling, or chemotaxis. These two environmental sensing mechanisms involve two-component signalling and metabolic signalling through fumarate. Indeed, archaeal and bacterial taxis rely on similar molecules and use common, albeit not identical, mechanisms that seem to be of ancient prokaryotic origin.
Despite striking similarities between these two systems, some important questions concerning the biochemistry of bacterial and archaeal sensing remain to be solved. Does metabolic signalling via fumarate as observed in Escherichia coli also occur in Halobacterium? Does stimulating the membrane chemosensory receptor proteins (MCPs) in E. coli also release fumarate? Does the trans-double bond of fumarate make it a key signalling molecule to indicate the metabolic state of such cells--not only for motility and taxis, but also for regulating transcription?
In fact, fumarate in vivo and in vitro directly and specifically inhibits transcription of the clcABD operon, encoding enzymes of a chloroaromatic biodegradative pathway in Pseudomonas putida. What sequence motif defines the fumarate-binding site at the switch complex of the flagellar motor and for the regulation of transcription? Are they related to one another? Fumarate-mediated signalling, but also the sensing of the proton motive force (PMF) and intracellular levels of arginine, suggest that the metabolic state of a prokaryotic cell regulates behavior through internal signals in addition to external sensory stimuli. Hence, investigators trying to develop quantitative models of bacterial behavior will need to account for this type of sensory integration.
Archaeal Rhodopsins: Light-Energy Converters and Sensory Photoreceptors
Archaeal rhodopsins are specialized photoreceptors that help certain extremely halophilic archaea, such as Halobacterium salinarum, to sense light and also to use light energy for growth. This microorganism is found in salt lakes, the Dead Sea, and in other saline habitats where the salt concentration comes close to saturation. When subject to low oxygen concentrations, a redox signalling pathway in this microbe is activated, inducing biosynthesis of bacteriorhodopsin, an outwardly directed proton pump that harnesses light energy via chemiosmotic coupling for growth. Besides bacteriorhodopsin, three other retinal proteins are expressed in such cells: halorhodopsin, a light-driven chloride pump that is used for salt uptake during growth, and two sensory rhodopsins (SRI and SRII) that mediate a simple form of color vision during phototaxis (Fig. 1A).
All four of these halobacterial rhodopsins consist of seven-helix-containing transmembrane proteins that are light sensitive through a retinal chromophore that becomes covalently linked through a Schiff base to a specific lysine of the polypeptide chain. The retinal is buried in the center of the protein. The protein structure helps to control both the color (wavelength) sensitivity and the photochemical reactivity of that bound chromophore. Two amino acids located adjacent to the Schiff base of the retinal chromophore optimize bacteriorhodopsin and halorhodopsin for ion pumping, efficiently guiding ions from the extracellular to the cytoplasmic side of the membrane or vice versa.
When the initially all-trans retinal within bacterial rhodopsin absorbs a photon, the chromophore photoisomerizes into its 13-cis form. This event acts as a molecular switch, triggering either bacteriorhodopsin or halorhodopsin to translocate ions or the sensors to form a signalling intermediate. Then the Schiff base loses a proton, shifting the absorption spectrum to the blue. This short wavelength-absorbing intermediate spontaneously returns to the initial state, and the retinal isomerizes back to all-trans while the Schiff base regains a proton (Fig. 1B). The protein catalyzes this reisomerization, which thermodynamically is driven by the energy of the photon that is stored in a ``tensed'' protein conformation. The ion pumps are fast, turning over within tens of milliseconds, whereas the sensors turn over more slowly, within hundreds of milliseconds, providing metastable signalling states.
Sensory rhodopsin I is photochromic, meaning it can be reversibly switched between two states by absorbing light at different wavelengths. Photoconversion of the initial state (SRI587) by orange light into SRI373 generates an attractant signal (Fig. 1B). SRI373 returns to its initial state in one of two ways--either spontaneously or by absorbing a second, UV photon. Photochemical reversion generates a repellent signalling intermediate. By this mechanism, SRI can distinguish between orange and UV light in generating antagonistic signals. SRII is a simpler blue light photoreceptor that generates a repellent signal when irradiated.
Halobacterial Transducer of Rhodopsin Has Two Functions
Photocycling makes the number of sensory rhodopsins in the UV-absorbing intermediate (SRI373) a function of the light intensity to which a cell containing it is exposed. Any change in light intensity changes the concentration of SRI373, which regulates the probability of switching the rotational direction of the flagellar apparatus.
Signal transduction occurs via another transmembrane protein, a halobacterial transducer of rhodopsin (HtrI or HtrII for SRI or SRII, respectively), that forms a stable complex with the corresponding sensory rhodopsin. However, because sensory rhodopsin and its transducer develop a tight physical interaction, light energy cannot drive them apart. Destroying the complex by removing the transducer or a cytoplasmic part of it genetically slows down the photocycling rate of SRI by two orders of magnitude due to the abolished functional interaction of the two molecules.
Based on its sequence and secondary structure, the transducer is homologous to the methyl-accepting chemotaxis proteins (MCPs) that serve as transmembrane chemoreceptors during eubacterial chemotaxis. However, there are some important structural differences between MCPs and halobacterial transducers. For instance, the large extracellular domain that binds the chemoligand in the MCPs is not found in the halobacterial transducer of rhodopsin. However, it contains an additional cytoplasmic domain that, when deleted, destroys the complex (Fig. 2).
Bacterial and archaeal transducers exhibit two antagonistic regulatory functions. First, they signal the two-component system (CheA/CheY) which regulates the rotational sense of flagellar motor rotation by reversibly methylating glutamate residues located in stretches called R1 and K1. This reversible methylation enables cells to adapt to sensory stimuli. The interplay of excitation and sensory adaptation guarantees that the output at the flagellar motor is related to the time-dependent change of the sensory input signal (dI/dt). According to quantitative analysis of the halobacterial photoresponse, adaptation occurs when the signalling activity of the transducers is adjusted, not by regulating the gain of the signalling pathway or the responsiveness of the flagellar motor to upstream components.
The archaea appear to combine functional elements in many different ways to generate transducer molecules. Besides the halobacterial transducers of rhodopsin, four other receptor types have been found among 13 different genes in H. salinarum. An oxygen receptor for aerotaxis is composed of six transmembrane helices with homology to the heme-binding sites of the eukaryotic cytochrome c oxidase and the MCP-homologous cytoplasmic domains for signalling and adaptation. BasT, the receptor that most closely resembles the E. coli MCP's, senses the concentration of five different amino acids. Another, CAR, is an intracellular soluble receptor that senses the cytoplasmic concentration of arginine and is also equipped with the MCP-like signalling modules (Fig. 2).
The function of other receptor types is not yet known, although one of them might be the PMF sensor. It generates a repellent signal when the cellular PMF suddenly drops, perhaps when certain starved cells are exposed to a step down in light intensity. Coupling via the membrane potential (D Y ) enables the light-driven proton pump bacteriorhodopsin to function as a high-irradiance photoreceptor for phototaxis, although it does not form a complex with any sensory transducer (Fig. 1).
Fumarate as a Switch Factor and Its Photosensory Control in Halobacterium
Some 10 years ago, when researchers realized that photoreception in halobacteria is mediated by sensory rhodopsins, no one understood the primary events in signal transduction involving halobacterial transducers. However, they did recognize that seven-helix transmembrane proteins serve as photoreceptors and appreciated that this structural motif is also typical for receptors in higher eukaryotes, such as in the [gb]-adrenergic receptor. Indeed, a new biochemistry of archaebacterial signalling was soon discovered. Without other means at their disposal, researchers followed a direct approach to identify signalling molecules in Halobacterium.
For example, at that time, we isolated a mutant without the ability to switch motor rotation either spontaneously or in response to light. The mutant could be cured by reversible permeabilization in the presence of extracts prepared from wild-type cells, and this effect was used as a bioassay for the ``switch factor.'' Even before we had isolated the switch factor, we found that its activity (as determined by the bioassay) is under sensory control by the photoreceptors. We learned this by stimulating halobacterial cells with blue light and rapidly quenching by osmotic lysis. After ultrafiltration, the switch factor activity in the low-molecular-weight fraction of the lysate drastically depended on whether the cells had been stimulated or not.
This result encouraged us to go for the molecular identity of the sensory-controlled switch factor. Soon we were disappointed to learn that the switch factor is not a protein but a low-molecular-weight molecule, converting a biochemical problem into a chemical one. Eventually, we determined by mass spectrometry that the switch factor is fumarate, and verified its role by testing fumarate in our bioassay. Next we analyzed the photosensory release, employing an enzymatic cycling assay sensitive enough to detect the low fumarate concentration in cell lysates. In fact, each photoactivated sensory rhodopsin II molecule catalyzed the release of 300 molecules of fumarate during its lifetime, amounting to as many as 60,000 molecules per second in a cell responding to a saturating light stimulus. Fumarate is released within 1 second from a membrane-associated or otherwise sequestered source of this small molecule. The halobacterial tricarboxylic acid (TCA) cycle is by far too slow to be the source for the release, since it produces only about 5,000 molecules in the same period.
Fumarate as Response Regulator and Metabolic Signal in E. coli
These results pointing to fumarate as a switch factor seemed unique to archaeal signal transduction until Rina Barak and Michael Eisenbach from the Weizmann Institute of Science in Rehovot, Israel, reported that it plays a similar role in cytoplasm-free envelopes of E. coli cells. Envelopes loaded with CheY and fumarate display flagellar motor switching, whereas CheY alone stimulates clockwise (CW) rotation but no switching. That experiment suggests three important points: (i) fumarate has switch factor activity in both archaea and bacteria, (ii) fumarate together with CheY acts at the level of the switch complex, and (iii) fumarate acts per se since it cannot be metabolized within a cytoplasm-free cell envelope.
The steady-state cytoplasmic concentration of fumarate in live, respiring E. coli cells is about 5,000 molecules per cell--a concentration that limits both switching and CW rotation. Deleting the gene encoding the enzyme fumarase to block the TCA cycle results in an eightfold increase of the steady-state level of fumarate. In turn, switching and CW rotation of the flagellar motor drastically increase even in unstimulated cells. The fumarate effect does not depend on whether CheY is phosphorylated or not, because it is seen even with a nonphosphorylatable mutant CheY.
Thus, CheY and fumarate appear to act like the two complementary keys needed to trigger the behavioral response of the cell (Fig. 3). But what about their mechanism of action? Is CheY a CW signal and fumarate a switching signal? To address this question one must ask how switching and CW rotation are correlated at different CheY or fumarate concentrations.
Halobacteria swim along straight lines by rotating the flagellar bundle in either direction. Sensory stimulation modulates the probability that a given cell will switch between forward and backward swimming (by CW and CCW flagellar rotation, respectively). Peritrichously flagellated bacteria, including E. coli, behave a little differently. An E. coli cell swims smoothly when most of its 10 or so peritrichously inserted flagella rotate in CCW direction and, due to their left-handed helicity, associate by forming a bundle that pushes the cell forward in a steady line. Upon switching from CCW to CW flagellar rotation, however, that bundle flies apart, allowing the cell to tumble and reorient until CCW rotation and smooth swimming resumes.
When wild-type E. coli cells are faced with a repellent stimulus (either by removal of a chemical attractant or addition of a repellent), there is CW rotation and, hence, tumbling. An individual flagellar motor responds to a stimulus of intermediate strength by oscillating between CCW and CW, spending about 50% of the time in either mode before adaptation resets it to prestimulus behavior. A strong stimulus increases the fraction of time spent by CW rotation and lowers switching frequency again. This results in a bell-shaped correlation when plotting switching frequency against the CW bias.
Stepwise overexpression of CheY in unstimulated E. coli cells gradually increases the time that flagellar motors spend in the CW mode and changes the switching frequency as predicted by the correlation curve. CheY overexpression can be functionally replaced by enhancing the cytoplasmic concentration of fumarate without changing the correlation of switching frequency and bias. Other lines of evidence suggest that CheY and fumarate cooperate (act additively) at the level of the switch complex. On the basis of these experiments, it makes no sense to discriminate between CW and switching signals.
In fact, fumarate may promote switching and CW rotation in absence of CheY. If cells deleted in CheY are cooled down to 5șC, their flagellar motors spontaneously switch but still show the fumarate effect. Using ``cool'' cells, Michael Eisenbach and coworkers determined that fumarate lowers the free energy difference between the states of the flagellar motor switch complex that mediate CW or CCW rotation. Hence, fumarate does not catalyze switching but facilitates CW rotation (as CheY-P does).
E. coli responds to Indole and Benzoate through Fumarate Signalling
Does fumarate regulate the behavior of E. coli? Because of its low steady-state concentration under aerobic conditions, fumarate is the limiting factor for switching and CW rotation in unstimulated E. coli cells. Therefore, metabolic fluctuations in the level of fumarate are expected to directly influence bacterial behavior.
This predicted behavior becomes obvious in mutant cells that cannot undergo phosphorylation-mediated chemotaxis. Nevertheless, even though such cells would be expected to be completely nonchemotactic, they respond to repellents such as indole and benzoate. Both these compounds are sensed through one of the methyl-accepting chemotaxis proteins, coupled to CheY via the sensor kinase CheA. Chemosensing of these compounds through the two-component system is only one order of magnitude more sensitive than the response observed in cells carrying the deletion.
In fact, both indole and benzoate inhibit E. coli fumarase in vitro and in vivo. In vivo inhibition increases along with the cytoplasmic fumarate concentration and, as expected, flagellar motor switching and CW rotation, the bacterial repellent response (Fig. 3). This metabolically mediated chemoresponse provides the basis for the chemotactic behavior observed in these cells. How efficient metabolic signalling is in relation to two-component signalling remains to be established.
SUGGESTED READING
Falke, J. J., R. B. Bass, S. L. Butler, S. A. Chervitz, and M. A. Danielson. 1997. The two-component signalling pathway of bacterial chemotaxis: a molecular view of signal transduction by receptors, kinases, and adaptation enzymes. Annu. Rev. Cell Dev. Biol. 13:457-512.
Hoff, W. D., K.-H. Jung, and J. L. Spudich. 1997. Molecular mechanism of photosignalling by archaeal sensory rhodopsins. Annu. Rev. Biophys. Biomol. Struct. 26:223-258.
Krah, M., W. Marwan, A. Vermeglio, and D. Oesterhelt. 1994. Phototaxis of Halobacterium salinarium requires a signalling complex of sensory rhodopsin I and its methyl-accepting transducer HtrI. EMBO J. 13:2150-2155.
Marwan, W., W. Schafer, and D. Oesterhelt. 1990. Signal transduction in Halobacterium depends on fumarate. EMBO J. 9:355-362.
Montrone, M., M. Eisenbach, D. Oesterhelt, and W. Marwan. 1998. Regulation of switching frequency and bias of the bacterial flagellar motor by CheY and fumarate. J. Bacteriol. 180:3375-3380.
Montrone, M., W. Marwan, H. Grunberg, S. Mu[gb]eleck, C. Starostzik, and D. Oesterhelt. 1993. Sensory rhodopsin-controlled release of the switch factor fumarate in Halobacterium salinarium. Mol. Microbiol. 10:1077-1085.
Montrone, M., D. Oesterhelt, and W. Marwan. 1996. Phosphorylation-independent bacterial chemoresponses correlate with changes in the cytoplasmic level of fumarate. J. Bacteriol. 178:6882-6887.
Prasad, K., S. R. Caplan, and M. Eisenbach. 1998. Fumarate modulates bacterial flagellar rotation by lowering the free energy difference between the clockwise and counterclockwise states of the motor. J. Mol. Biol. 280:821-828.
Email: webmaster@asmusa.org