![[Feedback]](/icons/banner/feedbackACT.gif)
![[For Subscribers]](/icons/banner/subscriberACT.gif)
![[Archive]](/icons/banner/archiveACT.gif)
![[Search]](/icons/banner/searchACT.gif)
![[Contents]](/icons/banner/tocACT.gif)
Microbiology and Molecular Biology Reviews, June 2004, p. 301-319, Vol. 68, No. 2
1092-2172/04/$08.00+0 DOI: 10.1128/MMBR.68.2.301-319.2004
Copyright © 2004, American Society for Microbiology. All Rights Reserved.
Diversity in Chemotaxis Mechanisms among the Bacteria and Archaea
Hendrik Szurmant and
George W. Ordal*
Department of Biochemistry, Colleges of Medicine and Liberal Arts and Sciences, University of Illinois, Urbana, Illinois 61801
The study of chemotaxis describes the cellular processes that control the movement of organisms toward favorable environments. In bacteria and archaea, motility is controlled by a two-component system involving a histidine kinase that senses the environment and a response regulator, a very common type of signal transduction in prokaryotes. Most insights into the processes involved have come from studies of Escherichia coli over the last three decades. However, in the last 10 years, with the sequencing of many prokaryotic genomes, it has become clear that E. coli represents a streamlined example of bacterial chemotaxis. While general features of excitation remain conserved among bacteria and archaea, specific features, such as adaptational processes and hydrolysis of the intracellular signal CheY-P, are quite diverse. The Bacillus subtilis chemotaxis system is considerably more complex and appears to be similar to the one that existed when the bacteria and archaea separated during evolution, so that understanding this mechanism should provide insight into the variety of mechanisms used today by the broad sweep of chemotactic bacteria and archaea. However, processes even beyond those used in E. coli and B. subtilis have been discovered in other organisms. This review emphasizes those used by B. subtilis and these other organisms but also gives an account of the mechanism in E. coli.
The discovery of microbes by Antonie van Leeuwenhoeik (59) was aided by their ability to swim, clearly indicating that they are living organisms. Not surprisingly the mechanism controlling this behavior has since then been studied extensively. Arguably, chemotaxis is the best understood of all signal transduction systems that control movement. While the motility apparatus differs among organisms, the general control mechanism is conserved throughout all bacteria and archaea. The centerpiece of this control mechanism is the "two-component" system in which phosphorylation of a response regulator reflects phosphorylation of a histidine autokinase that senses environmental parameters (117). This is the most common type of signal transduction system in bacteria and controls diverse processes such as gene expression, sporulation, and chemotaxis. In chemotaxis, events at the receptors control autophosphorylation of the CheA histidine kinase, and the phosphohistidine is the substrate for the response regulator CheY, which catalyzes the transfer of the phosphoryl group to a conserved aspartate (for a recent review, see reference 250). The resulting CheY-P can interact with the switch mechanism in the motor (42, 149, 186, 193, 234). This interaction causes a change in behavior, such as in direction or speed of rotation of flagella. Thus, for example, in Bacillus subtilis, binding of the attractant asparagine to the receptor McpB quickly increases the levels of CheA-P and CheY-P, as the excitation event, and produces increased counterclockwise (CCW) rotation of the flagella (265). The receptors undergo adaptation, a feature that allows the mechanism to reset so that bacteria can progress up concentration gradients of attractants or down concentration gradients of repellents (152). In general, the excitation process is highly conserved and there is considerable variety in the adaptation process (Fig. 1).
View larger version (22K):
[in this window]
[in a new window]
|
FIG. 1. General chemotaxis model. A schematic of the biochemical processes in the two-component chemotaxis pathway is shown. Hexagons represent response regulator domains. The universal components are in red; almost universal components are in orange; optional components are in yellow.
|
|
Besides these two core proteins, many other proteins contribute to making the process work. Chemotaxis proteins can be ordered into four groups—a signal recognition and transduction group, an excitation group, an adaptation group, and a signal removal group (to dephosphorylate CheY-P). The signal recognition and transduction group includes the receptors (9, 81, 118) and ligand binding proteins (4, 86), which are capable of binding effectors outside the cell; a few receptors, however, are cytoplasmic (92, 93, 229). The signal, i.e., changing concentrations of a chemical, is then transduced to the excitation proteins, CheA and CheY (34, 90). Adaptational proteins alter CheA activity to reset the system. This can be done either by influencing CheA activity directly or through the receptors. Lastly, the signal removal proteins ensure that CheY-P levels can be adjusted to prestimulus levels quickly (the roles of the chemtaxis proteins are summarized in Table 1).
All biochemical processes described here were first discovered in the enteric bacteria Escherichia coli and Salmonella enterica serovar Typhimurium. Fortunately, the enteric chemotaxis system turned out to be comparatively simple. Since then, chemotaxis in many diverse organisms has been studied and many, often complex, variations have been found. This review aims to summarize and compare the different chemotactic systems that have been studied to date. The main focus is not E. coli chemotaxis, since many review articles have dealt with its chemotaxis, but Bacillus subtilis, which is arguably the second best understood chemotactic bacterium. Other than cheZ, it possesses at least one copy of each chemotaxis protein found to date (although some fusions of chemotaxis proteins exist that are not found in B. subtilis).
As will become apparent, the chemotaxis mechanism in B. subtilis is probably close to that of the ancestral organism from which the bacteria and archaea descended, so that understanding this mechanism should provide considerable insights into mechanisms used in the diverse species of motile bacteria and archaea alive today. To appreciate the divergence in chemotaxis, we tried to include in this review some information about at least one representative of each phylum of bacteria and archaea in which a CheA homolog could be found, indicating the existence of a chemotaxis pathway. Of course, many organisms have not yet been studied in detail, and the available information is often based only on the genomic sequence of those organisms. We have made a special effort to include information about organisms whose chemotaxis mechanism appears to diverge from the E. coli paradigm. Other reviews, most of which emphasize the E. coli mechanism, include references 13, 39, 61, 224, and 227. The review by Berg (20) does justice to the rather considerable literature dating from the late 19th and early 20th centuries.
|
SIGNAL RECOGNITION AND TRANSDUCTION
|
|---|
Receptors
Understanding how receptors control the CheA kinase is at the heart of understanding chemotaxis. In E. coli, binding of attractant inhibits the CheA kinase (34), whereas in B. subtilis, binding of attractant stimulates the CheA kinase (64, 67). It is important to understand the structure of receptors and how binding of attractant (or repellent) changes the activity of the associated CheA kinase.
Classes of receptors.
The receptors are usually transmembrane proteins with an extramembrane sensing domain that binds attractant across the dimeric interface (253), two transmembrane (TM) regions (TM1, between the N terminus and the sensory region, and TM2, between the sensory region and the cytoplasmic regions), and several cytoplasmic regions. These include the signaling region, where the CheA kinase and the CheW coupling protein and analogs bind, and the methylation region, where methylation/demethylation of the receptors occurs to compensate for changes in CheA kinase activity caused by binding attractant (Fig. 2). The enzymes catalyzing these reactions, the CheR methyltransferase and the CheB methylesterase, are described below. Between the methylation region and the membrane is the HAMP (histidine kinase, adenylyl cyclase, methyl-accepting chemotaxis protein, and phosphatase) linker, which conveys the signal of attractant binding to the rest of the cytoplasmic region (see below). Based on crystal structures of the extramembrane N-terminal part of Tar, the aspartate receptor of E. coli, and of the cytoplasmic C-terminal part of Tsr, the serine receptor, virtually the entire receptor is thought to consist mainly of 
-helix (51, 108, 109).
View larger version (48K):
[in this window]
[in a new window]
|
FIG. 2. Schematic of the three classes of chemotaxis receptors. Shown is a representative dimer for each class of chemotaxis receptors. Insertion/deletion regions (indels) are shaded in dark gray. Stars indicate sites of methylation for (from left to right) E. coli Tar, M. xanthus FrzCD, and B. subtilis McpB.
|
|
Le Moual and Koshland (124) showed that the chemotaxis receptors in the bacteria and archaea fell into three classes, based on the presence or absence of two pairs of insertions (called "indels," for insertions/deletions) of 14 amino acids (four turns of the -helix). The class III receptors are likely to be the ancestral receptors (124). The locations of these pairs in B. subtilis McpB are illustrated in Fig. 2. They lie on the membrane-proximal side of the signaling regions and methylation regions. The class III receptors have both pairs, the class II receptors have only the pair between the signaling and methylation regions, and the class I receptors have neither. More recent analysis of sequences in the database (Fig. 3) gives some additional perspective. The original receptors are obviously the class III receptors, but they have undergone modifications early in the lines of descent. One modification, which presumably occurred after the gram-positive bacterial line had diverged from the original line and before the proteobacterial line had diverged, was the deletion of the first and fourth indels to produce the class II receptors. The class I receptors, which involve deletion of indels 2 and 3, may have arisen several times, once during formation of the -proteobacterial line (see Rhodobacter sphaeroides), once during formation of the 
-proteobacterial line (see Myxococcus xanthus), once during formation of the ß-proteobacterial line (see Ralstonia solanacearum), and once after the 
-line had been formed (see E. coli and Pseudomonas aeroginosa) (Fig. 3). Alternatively, and more probably, the class I receptors may have arisen fewer times and, early during the evolution of a particular line of descent, may have entered by gene transfer and displaced the original receptors. One obvious case of gene transfer is in Clostridium acetobutylicum, a gram-positive bacterium having 38 class III receptors (all the rest of the gram-positive bacterial receptors are class III) and one class I receptor (Fig. 3). The receptors in the spirochete Borrelia burgdorferi, here termed class O, are a special case: these receptors are missing indels 1 and 4 but are missing pairs of 21 amino acids where the loss of indels 2 and 3 would produce a loss of only a pair of 14 amino acids.
View larger version (30K):
[in this window]
[in a new window]
|
FIG. 3. Phylogenetic tree of chemotactic bacteria and archaea. The phylogenetic tree was generated from the 16S rRNA sequences by using the programs CLUSTALW and DRAWTREE. Included is information regarding the number and class of chemoreceptors for each respective organism. ND, not determined. Organisms, by phylum, are as follows: Archaea: Archaeoglobus fulgidus, Halobacterium salinarum, Methanosarcina mazei, and Pyrococcus abyssi; Thermotogales: Thermotoga maritima; spirochetes: Borrelia burgdorferi, Leptospira interrogans, and Treponema pallidum; cyanobacteria: Nostoc and Synechocystis spp.; gram-positive bacteria: Bacillus subtilis, Clostridium acetobutylicum, Listeria innocua, and Thermoanaerobacter tengcongensis; proteobacteria (-subgroup): Rhodobacter sphaeroides and Sinorizhobium meliloti; proteobacteria (ß-subgroup): Ralstonia solanacearum; proteobacteria (-subgroup): Myxococcus xanthus; proteobacteria ( -subgroup): Helicobacter pylori; proteobacteria (-subgroup): Escherichia coli, Pseudomonas aeroginosa, and Vibrio cholerae.
|
|
HAMP domain.
As implied above, the HAMP domain or linker is found in a number of different types of proteins. Although there is little sequence identity among HAMP domains, they generally have two segments of hydrophobic aminoacyl residues in a heptameric arrangement characteristic of amphipathic -helices joined by unstructured amino acids (11, 12). The -helices are probably in a coiled coil (206). The purpose of HAMP domains is generally to convey signals from input domains to output modules. Mutations in the E. coli receptor Tsr HAMP domain caused locked signal output (that is, persistently clockwise [CW] or CCW or in between, but switching rarely) (8).
Methylation of class I and III receptors.
Considerable work has been done on McpB from B. subtilis, which might be considered a prototype for the class III receptors. The first difference noticed between the class I and III receptors was that methanol was released in response to all stimuli in B. subtilis (111, 240) and Halobacterium salinarum (163, 216) whereas it was released from E. coli, the prototype organism for class I receptors, only after the application of negative stimuli; methanol evolution in this species was suppressed below background levels after the application of positive stimuli (105, 241, 242). Interestingly, methanol is released from the class I receptors of R. spheroides on addition of attractant (which, as in E. coli, inhibits CheA [196]) (145). There is no consequence on methanol formation of removing the attractant. However, as described below, this organism has multiple copies of chemotaxis genes. The principal methyltransferase appears to be CheR2; deletion of cheR1 causes methanol to be produced after both addition and removal of attractant (145). The related organism, Rhodospirillum centenum, however, behaved as might have been anticipated from the E. coli precedent: a reduction of light intensity, which would cause inhibition of the CheA, caused methanol formation, and increase of light intensity did not (97).
What is the reason for this difference in methanol formation? In E. coli, methylation of receptors increases CheA kinase activity, an adaptational mechanism to compensate for the decreased activity caused by attractant, and it does not appear to matter which sites become methylated (106, 197, 238). By contrast, each of the sites in McpB, the one class III receptor that has been characterized in some detail, appear to have a different function. Glu630 is demethylated both after addition and removal of attractant (265) (provided that this site is in the methylated form). When it is changed to Asp630, which cannot be methylated (198), the resulting mutant has a low prestimulus bias; when attractant is added, it adapts only to the higher wild-type bias, and when attractant is removed, it hardly adapts at all (265). Glu637 seems crucial for causing adaptation to attractants, and only on removal of attractants does methanol evolve from that site (provided that this site is in the methylated form). Glu371 (encoded as Gln371 but deamidated (probably by CheD [see below]) does not play a role during adaptation to addition or removal of attractant but may in some way be involved in being a regulator of methylation, since an mcpB Q371A mutant shows poor taxis to high concentrations of asparagine, similar to a cheB mutant (see below) (M. A. Zimmer and G. W. Ordal, unpublished data). Methanol arises from Glu371 only on addition of attractant (provided that this site is in the methylated form). Thus, in B. subtilis, each of the positions from which methanol arises has a different function, and these functions, in the case of McpB (as stated above), cause methanol production from Gln(Glu)371 and Glu630 on addition of attractant and from Glu630 and Glu637 on removal (provided that these sites are in the methylated form). The extent to which this rule about particular sites having particular functions applies generally is not known; further experiments are required, and one obvious organism with which to explore this question would be R. spheroides (see previous paragraph). In any case, what had seemed until now the general rule that class III receptors produce methanol in response to all stimuli and class I receptors do so only in response to negative stimuli does not appear to be true (the case of R. spheroides contradicts this rule). No work of this type has been done for any class II receptors.
Structure of receptors.
The E. coli receptors are stable dimers (153) arranged as trimers of dimers (9). The B. subtilis receptors are also similarly arranged, and, in fact, the trimers of dimers themselves interact near the outer part of the cytoplasmic membrane (46), as predicted by Kim et al. (110). X-ray structures of the extramembrane ligand binding domain of the E. coli receptor Tar (194) and the cytoplasmic domain of the E. coli receptor Tsr (109) are available, as well as of the ligand binding domain of the soluble, cytoplasmic B. subtilis receptor HemAT, in the presence and absence of the natural ligand O2. The dimer that binds O2 (the form that stimulates the CheA kinase) is very symmetrical, and the dimer that is free of O2 shows a distinct conformational change in the Tyr70 of one of the two subunits of the dimer (254, 255). As described in detail below, the CheR methyltransferase and the CheB methylesterase appear to be especially active on receptors that have just bound or released attractant and before the compensating methylation changes have occurred that would help bring about adaptation. The structural basis of this conformation of increased susceptibility is still unknown (45).
Effect of attractant on structure.
One longstanding question has been how attractant induces the receptors to change CheA kinase activity. In E. coli, attractant causes diminished CheA activity (35). Based on evidence from work on Cys-substituted receptors, whose cross-linking is accelerated by oxidant, there appears to be "downward piston" movement of TM2 of one monomer of the receptor dimer (50, 94, 123). Based on experiments where the nitroxide spin label was attached to Cys residues in Cys-substituted receptors, the extent of movement was deduced to be small, only about 1 Å (171). The effect of repellent on these receptors and the effect of attractant or repellent on class III receptors, such as McpB of B. subtilis, where attractant activates rather than inhibits the receptor (67), is unknown.
Maturation by CheD.
CheD is a deamidase that deamidates particular glutamines in the B. subtilis receptors (122), a function carried out by CheB in E. coli. Most chemotactic bacteria and archaea carry cheD. Therefore CheD is probably the ancestral mechanism of glutamine deamidation. The role of this function is not yet fully understood. However without CheD the receptors are undermethylated and activate the kinase poorly, indicating that deamidation is necessary for activation of the receptors (67, 182).
Binding proteins.
Although in most instances chemoeffectors are thought to interact directly with the receptors, in some particular instances, specifically dedicated binding proteins bind the chemoeffector and the complex then binds the receptor. Thus, in E. coli, galactose binds to the galactose binding protein (86), ribose binds to the ribose binding protein (4), and the complex binds to the receptor Trg (83, 87). Maltose binds to the maltose binding protein (85), and the complex binds to the aspartate chemoreceptor Tar (176). It is suspected that the B. subtilis receptor McpC binds attractants indirectly, via binding proteins, since it mediates taxis to all amino acids except asparagine, some (such as proline and alanine) at very low concentrations (159, 168); however, no mutants in any such putative binding proteins have been identified.
Oxygen sensing.
The oxygen sensor in B. subtilis is HemAT, which is homologous to myoglobin (92, 93). It is similar to the repellent oxygen sensor, also termed HemAT, in the archeon H. salinarum. It is a soluble receptor, having no transmembrane region, and hence senses the internal oxygen concentration. H. salinarum has another receptor for oxygen as an attractant. However, this receptor is homologous to cytochrome oxidase of mitochondria (44) and has six membrane-spanning regions and may be a heme protein that also senses oxygen directly. Conversely, Aer, the oxygen sensor of E. coli, binds flavin adenine dinucleotide (FAD) (21, 22, 177), and the signal caused by changing oxygen concentrations is probably mediated by changes in the level of reduction/oxidation of this FAD (21, 177, 235, 236), a process involving a PAS domain in the receptor (178). Pseudomonas putida would appear to use the same mechanism (161). Tsr in E. coli also mediates oxygen taxis, perhaps by sensing changes in the proton motive force across the cytoplasmic membrane (177, 204) but certainly not by binding oxygen directly (Fig. 4).
View larger version (48K):
[in this window]
[in a new window]
|
FIG. 4. Aerotaxis receptors. Shown is a schematic of the four known types of aerotaxis receptors. Indirect aerotaxis defines receptors that detect oxygen levels by the proton motive force or redox state of the cell. Direct aerotaxis defines receptors that detect levels directly by interacting with oxygen. Black diamonds represent the indicated receptor cofactors.
|
|
Many other organisms perform aerotaxis (reviewed in reference 236). Azospirillum brasilense, an -proteobacterium, accumulates in an oxygen gradient at 3 to 5 µM. At both lower and high oxygen tensions, the proton motive force is lower, so that it is assumed that both positive and negative aerotaxis, which causes accmulation of bacteria at the optimum oxygen tension, is due to sensing of changes in the proton motive force (262). R. sphaeroides also accumulates at an optimum oxygen concentration. It shows negative aerotaxis due to interaction between the chemotaxis machinery and the Prr system, which monitors electron flow through the alternative high-affinity cytochrome oxidase, cbb3, and positive aerotaxis by interaction with another, unknown sensor, both of which may operate through one of the chemotaxis kinases of the cell, CheA2 (180). Some species, for example Sinorhizobium meliloti, do not seem to respond to oxygen in the same way that E. coli and B. subtilis do (by modulating the frequency of CCW versus CW rotation) but, rather, change their swimming speed in response to oxygen gradients (263).
Phosphotransferase System
The phosphotransferase system (PTS) helps mediate taxis to a number of sugars and sugar alcohols in B. subtilis (68, 121, 169). Transport is required, but metabolism is not. Unlike taxis to PTS sugars in E. coli, which does not require a specific receptor and works even when the methylation system is inactivated by mutation (162) (but works poorly unless some receptor is present [136]), chemotaxis to PTS substrates in B. subtilis requires the C-terminal part of McpC (Fig. 5). In these experiments, chimeras between the asparagine receptor McpB and the proline receptor McpC revealed that the N-terminal, extramembrane part of the receptors mediated amino acid taxis, as expected, but only the C-terminal part of McpC could mediate taxis to PTS substrates and, in particular, the methylation region appeared to be involved. The data were best interpreted by a model in which unphosphorylated enzyme I interacted with McpC to bring about increased CheA activity and adaptation occurred through the normal means (121). Indeed, methanol was produced on addition of glucose, a sign that CheB was stimulated to help bring about adaptation (239). In E. coli, since unphosphorylated enzyme I interacts with CheA (135), it is suspected that interaction of CheA with unphosphorylated enzyme I inhibits CheA. Large changes in the levels of unphosphorylated enzyme I compared with phosphorylated enzyme I occur during chemotactic excitation (136). It would seem likely that the requirement for receptor found by Lux et al. (136) might be due to the inherent low activity of CheA in the absence of receptors (34) rather than the interaction with a specific receptor, as found for B. subtilis.
View larger version (33K):
[in this window]
[in a new window]
|
FIG. 5. PTS in chemotaxis. The two known chemotaxis pathways for PTS sugars are shown. Transport of PTS sugars increases the concentration of unphosphorylated enzyme I (EI) that can either directly interact and inhibit CheA (receptor-independent system), as is the case in E. coli, or indirectly stimulate CheA through the receptors (receptor-dependent system), as thought to be the case in B. subtilis. The letters A, B, and C represent components of an ABC transporter. For better understanding, PTS proteins are in yellow, B. subtilis chemotaxis proteins are in red, and E. coli chemotaxis proteins are in green.
|
|
With the exeption of Mycoplasma gliding motility, it appears that bacterial and archaeal motility is universely controlled by the two-component system of the CheA kinase and the CheY response regulator.
CheA Kinase
The central enzyme that mediates input, usually as sensed by the receptors, and creates an appropriate signal for the motor is the CheA kinase. Attractants inhibit it in E. coli (34), S. meliloti (192), and R. spheroides (196) and stimulate it in B. subtilis (67). As described in "CheY response regulator" (below), it is likely that for the archaea and the spirochetes, positive stimuli (for instance, chemoattractants or attractant light), decrease CheA activity. Thus, B. subtilis would appear to be the exception to this rather incompletely verified hypothesis that positive stimuli decrease CheA activity.
On activation, CheA becomes phosphorylated on a particular histidine residue (His46 in B. subtilis and His48 in E. coli [258]). CheA has five domains (described in detail in reference 23), and this His residue is located within the first, P1 or Hpt, domain (156). The fourth (P4) domain is where ATP binds and catalysis occurs. The third (P3) domain is the dimerization domain, important since CheA is a dimer that transphosphorylates (232) (the P4 region of one monomer phosphorylates a His residue in the P1 domain of the other monomer). The P5 domain is where CheA contacts the receptors and the coupling protein, CheW. The P2 domain is where CheY and CheB, which receive phosphoryl groups from CheA-P, dock (23, 89, 233). The exact mechanism of CheA autophosphorylation is not yet known; however, several conserved regions within the P4 domain—the N-box, G1-box, F-box, G2 box, and GT-block—are essential for catalysis in E. coli and are thought to be involved in positioning of ATP into the active site (Fig. 6) (91).
View larger version (12K):
[in this window]
[in a new window]
|
FIG. 6. Schematic of CheA and CheY. Shown are the histidine kinase CheA and the general response regulator CheY. For CheA, the five domains are labeled P1 through P5. The phosphoreceiving histidine in P1 is highlighted in gray, and conserved regions within domain P4 that are thought to play an active role in catalysis are also highlighted in gray. For CheY, conserved residues that participate in catalysis are positioned as indicated.
|
|
CheA-P from B. subtilis differs from its E. coli counterpart in being of considerably lower energy (Keq = 1.2 x 104 instead of 1 in the reaction CheA + ATP 
CheA-P + ADP) (66). When E. coli becomes somewhat deenergized, it becomes smooth swimming (107), since CheA cannot be phosphorylated, and thus the bacterium has a larger "diffusion constant" so that it will leave the local environment by rapid translational movement. However, when B. subtilis becomes somewhat deenergized, the chemotaxis system still functions. Thus, the bacteria do not become tumbly (the condition in the absence of CheY-P [27]) and thus unable to move away from the unfavorable environment but instead use chemotaxis to depart, a much more effective process than unregulated smooth swimming. In terms of phylogeny, CheA from B. subtilis clusters with CheAs from archaea and spirochetes, apart from CheAs of the proteobacteria (2).
Coupling to receptors.
CheA is coupled to the receptors via CheW. CheW is present in all bacteria that have chemotaxis or phototaxis receptors. A second protein, CheV, which is a CheW-CheY fusion protein, is also capable of coupling the kinase to the receptors and is described below. CheV is present in many eubacteria, including S. enterica serovar Typhimurium, a close relative of E. coli, but not in E. coli and not in any archaea. In E. coli, CheW is required for activation of CheA but not for its inhibition (7, 35). CheW is homologous to the P5 or regulator domain of CheA (23).
CheY Response Regulator
The primary response regulator that governs the direction of flagellar rotation is CheY-P (16, 25, 90) and, as implied above, it causes CCW flagellar rotation in B. subtilis and CW flagellar rotation in E. coli. It catalizes its own phosphorylation on a conserved aspartate residue by using CheA-P as a substrate. This process is thought to involve several conserved residues—two aspartates, which position an essential Mg2+ ion, a lysine, and a threonine (Fig. 6) (132, 244, 260). In flagellated bacteria, CheY-P interacts with FliM, shown for E. coli (248) and B. subtilis (26, 234). In the spirochete Treponema denticola, mutation of cheA blocked chemotaxis and caused the bacteria to have few reversals of motion (137), implying that CheY-P causes reversals of motion. To achieve this, it probably binds to FliM, which is present in the spirochetes. The archeon H. salinarum similarly showed no reversals of motion when cheA or cheY was deleted and, indeed, showed preferential forward swimming (chronic CW rotation of the flagella at each end of the archeon) (184). Thus, in this organism, CheY-P also causes reversals of motion and is required for CCW rotation of the flagella. (Note: these are right-handed flagella [5], not the more common left-handed flagella [139] found in E. coli and B. subtilis, and the flagella themselves are more similar to type IV pili [52] than to the flagella of bacteria.) There are no FliM homologs in the archaea, and the site of the interaction of CheY-P to control the direction of flagellar rotation is unknown. The implication of these findings is that the flagella of Halobacterium and spirochetes have a default direction of rotation in the absence of CheY-P, as do the flagella in the peritrichous bacteria like E. coli and B. subtilis, and that CheY-P not only facilitates rotation in the opposite direction but also facilitates switching between the two directions. How this might occur is mentioned below (see "CheC dephosphorylating and adaptational protein").
Signal Amplification
Binding of one or two receptors by attractant can lead to a behavioral response in B. subtilis (113) and E. coli (195). Using photoreleased aspartate, Jasuja et al. (96) found that nanomolar asparate (1.2 µM KD) could evoke a response and that the response times were proportional to changes in receptor occupany near the threshold, irrespective of prior occupancy. (Therefore, adaptation is complete.) In experiments with E. coli, using fluorescence resonance energy transfer to measure CheY-P levels (rather than CCW/CW rotation, which is a complex function of CheY-P levels), Sourjik and Berg (210) found that the cheB mutant was very insensitive to attractant compared to both the wild type and a cheR mutant. Using photoreleased asparate, Kim et al. (108) also found that the cheB mutant was far less sensitive than was the wild type. Both groups found that absence of CheZ, which catalyzes the dephosphorylation of CheY-P, had little effect on amplification; therefore, accelerated loss of CheY-P is not the cause of signal amplification; it must be sought in signal generation. How CheB might be involved in this is described in "CheB methylesterase" (below). Besides this, it seems likely that organization of the receptors into a lattice could lead to amplification; in this arrangement, judicious methylation of receptors to increase CheA activity (after reduction of activity from attractant [in E. coli]) would allow this amplification to occur over a broad range of attractant concentrations (201, 202). However, it would appear that this alleged lattice quickly forms and disappears, according to circumstances (see "Localization of chemotaxis proteins," below).
To sense ever higher concentrations of attractant and to move toward favorable environments, chemotaxis systems have to be able to adapt to existing stimuli. Additionally, the nature of bacterial motion requires the ability to recognize when the bacterium is moving in the wrong direction, i.e., away from higher attractant concentrations. To do that, a "memory" is required that is able to indicate whether higher or lower concentrations are being reached (120). This is achieved by the adaptational mechanisms. The methylation system of CheR and CheB is the only adaptational mechanism in E. coli that has been studied, although another, undescribed mechanism may exist (162, 228). However, other organisms, e.g., B. subtilis, have at least partly characterized adaptational systems, namely, CheV and CheC (Fig. 7), in addition to the mechanism involving CheR and CheB (99, 112).
View larger version (26K):
[in this window]
[in a new window]
|
FIG. 7. Adaptation systems. A flowchart of the possible means of adaptation is shown. The almost universal methylation-dependent adaptation system is shown on the left. The less highly conserved methylation-independent pathways are on the right. We speculate here that CheV might directly influence CheA activity following phosphorylation, as shown, while the means of CheC activation and adaptational action are not yet understood. Black diamonds represent a chemoeffector. Adaptational proteins are hatched; excitatory proteins are shaded.
|
|
Methylation
CheR methyltransferase.
CheR methyltransferase transfers methyl groups from S-adenosylmethionine to particular glutamate residues (102, 164, 238) on the receptors (82, 164), with production of S-adenosylhomocysteine (47, 48, 214). In B. subtilis, it is required for adaptation to repellents. In its absence, B. subtilis is very tumbly (with predominantly CW rotation of the flagella) (116). In E. coli, it is required for adaptation to attractants (76, 119) and binds both to a flexible tether at the C-terminal end of the receptor and to the methylation region of the same or a nearby receptor and so moves through a receptor cluster in a hand-over-hand fashion (58, 127, 252). This property, as might be expected, allows CheR bound to one receptor to methylate another (125, 126, 129), and receptors lacking the C-terminal binding site are poorly methylated unless receptors containing it are present in the same cell (19). The properties of E. coli CheR and B. subtilis CheR are likely to be similar since receptors of either organism can be methylated by CheR from either organism (47). CheR is present in virtually all bacteria and archaea showing chemotaxis, since it is important for adaptation; the exception is Helicobacter pylori, which has CheV, whose phosphorlation, as described below, is known to help bring about adaptation in B. subtilis (99).
CheB methylesterase.
Unlike CheR, CheB is usually regulated; it has an N-terminal response regulator domain, subject to phosphorylation, and a C-terminal enzymatic domain (57, 90, 134, 221). The E. coli enzyme is 10-fold more active when phosphorylated (134). The B. subtilis enzyme can satisfy the requirement for methylesterase for chemotaxis in E. coli (115), and, in vitro, both enzymes were able to demethylate both B. subtilis and E. coli receptors (160). One interesting possible difference is that CheB deamidates particular E. coli receptors, after which the site can be methylated and demethylated (102). In the instances where B. subtilis receptors are deamidated (for instances where the reaction has been characterized), this reaction is catalyzed by CheD, not by CheB (122). Perhaps, then, it is not surprising that the B. subtilis CheB appears to be more closely related to CheB in the archaea, which have CheD and class III receptors, than to CheB in most bacteria, some of which lack CheD and, in the case of the proteobacteria, have class I and class II receptors (2).
The isolated enzymatic domain of CheB catalyzes receptor demethylation in B. subtilis (45) and in E. coli (134), although not as effectively as does the phosphorylated whole enzyme (10, 57). However, in B. subtilis, a truncated cheB encoding the enzymatic domain complements a null cheB mutant, and this strain (null cheB mutant having truncated cheB on a plasmid) releases enhanced levels of methanol on both addition and removal of attractant. This result implies that the demethylation reaction involves primarily the existence of a suitable substrate. The value of phosphorylation of CheB would then be to increase the rate of receptor demethylation and thus speed up adaptation and to minimize unnecessary receptor demethylation, since loss of a methyl group is equivalent to hydrolyzing 11 to 14 ATP molecules to ADP and Pi (45, 225). Interestingly, CheB from Campylobacter jejuni lacks a response regulator domain (142). It is assumed that the time during which enhanced methanol formation occurs is the time between addition or removal of attractant and the resulting compensating (to bring about adaptation) methylation events on the receptor. This particular susceptible conformation does not require the coupling proteins CheW or CheV for events after the addition of attractant but does require them for events after the removal of attractant (45). The fact that the receptors that have bound attractant are more susceptible to methylesterase was shown in vitro many years ago (160).
The methylation system, involving CheR and CheB, is important for chemotaxis to high concentrations of attractant and only peripherally for chemotaxis to low concentrations of attractant in B. subtilis (113, 115). The reason is thought to be that at low concentrations of attractant, signal amplification occurs since binding attractant to a receptor activates neighboring receptors (203). At high concentrations of attractant, without CheB to generate charge-charge repulsion, there are no available free receptors that can activate the kinase on binding attractant (113). At low concentrations of attractant, methylation-independent systems suffice to bring about adaptation, such as the CheV system (see below). Thus, it may not be surprising that the amount of methanol evolved increases exponentially with receptor occupancy by attractant (111).
CheB may play another role as well. Under certain conditions, it appears to be required for a response to removal of attractants (113), and the first characterized cheB (then termed cheL) mutant (OI1130) was tumbly and unresponsive to stimuli (167). The implication is that this mutant CheB prevents normal functioning of the receptor-CheA complex. No similar mutants have been characterized in other organisms, such as E. coli, but absence of CheB results in an E. coli strain that is far less sensitive to the addition of attractant than is the wild type (108, 210), as described in more detail in "Signal amplification" (above). As stated above, attractants inhibit CheA in E. coli; this experiment implies that CheB is needed to amplify the process leading to low CheA-P and thus CheY-P levels (18). One way of achieving this is to inactivate the CheA associated with a receptor complex and have CheB molecules, which undergo very rapid autodephosphorylation (90), diffuse to neighboring complexes and inactivate them.
Methylation-Independent Adaptation
CheV adaptational and coupling protein.
CheV has two domains, a N-terminal domain homologous to CheW and a response regulator C-terminal domain (63), and can substitute for CheW in coupling receptors to CheA in B. subtilis (181). Insight into its function has come from experiments using a mutant in which the phosphorylated aspartatyl residue was replaced with an alanyl residue (CheV D235A strain) and the whole response regulator domain was deleted. Both mutant strains showed poor adaptation to the addition of attractants, a result implying that the purpose of CheV phosphorylation is to bring about adaptation (99). Interestingly, mutants lacking cheV altogether did adapt normally in the tethered-cell assay. Thus, it would appear that the conformation of the coupling ("CheW") domain of CheV is such as to strongly favor receptors bound with attractant in the conformation to activate CheA, since other adaptation systems like the methylation system are unable to restore the prestimulus bias. Since the CheV D235A strain adapts poorly, it would seem that adaptation requires phosphorylation of D235, probably so that the regulator domain can interact with the coupling domain to affect the conformation of the coupling domain and allow the attractant-bound receptors to reassume their prestimulus conformation.
CheV may be the only adaptation system in H. pylori, since CheR and CheB are absent. However, there are three CheVs, of which only CheV1 appeared to be required for chemotaxis, and none could substitute for CheW (172).
CheC dephosphorylating and adaptational protein.
As mentioned below CheC has CheY-P hydrolyzing activity (234a). However it is hard to explain the tethered cell phenotype of a B. subtilis 
cheC mutant, other than by assuming that it also plays a role in adaptation. While the prestimulus rotational bias of cheC is approximately that of the wild type, cells do not adapt to the addition of attractant (112, 182). This can be explained by the presence of persistently elevated CheA-P levels. CheC was shown to bind to McpB and CheA and so might either directly or indirectly influence CheA activity (112).
In addition, mutants with mutations in cheC have a lower frequency of switching the direction of rotation (from CCW to CW and from CW to CCW), implying that CheC lowers the energy of transition of switching (thus, the wild type, which has CheC, has a higher switching frequency than does the cheC mutant). Mutants with mutations in cheB have the opposite phenotype, an increased frequency of switching (190). It is hard to imagine that CheB binds to the switch but not so far-fetched to imagine that CheC does, since it is homologous to most of the FliM and FliY proteins (two of the three proteins comprising the switch). The state of two proteins being homologous does not necessarily imply that they bind each other; however, such binding does occur, for instance between CheA and CheW, the P5 domain of CheA being homologous to CheW (see above). These results can be accounted for by assuming that CheC has minimal affinity for overmethylated receptors so that in a cheB mutant, there would be more CheC bound at the switch. Very interestingly, a cheB mutant of H. salinarum shows increased frequency of reversals, with no effect on the ratio of CW and CCW rotation of the flagella, compared with the wild type (184). Similarly, the cheB mutant of B. subtilis has a normal bias (i.e., the same ratio of CW and CCW rotation of the flagella as in the wild type). Thus, it is not hard to imagine that reversal frequency in H. salinarum and B. subtilis is controlled by the same mechanism, namely, the amount of CheC bound at the switch. Presumably, in H. salinarum, the mechanism by which CheY-P produced by repellents or repellent light would cause increased reversals would involve inducing increased CheC binding at the switch. A similar situation may exist for the spirochetes, which undergo reversals of motion, except that there the CheC homolog is CheX, which is smaller than CheC (137).
CheC is not the only substance, however, that affects the switching frequency. Fumarate also promotes increased switching frequency in E. coli (17, 155, 175) and also increases the probability of CW rotation (154) by binding at the flagellar switch (175). As a central metabolite, fumarate would not be expected to be a chemotaxis signal whose concentration changes on a timescale of seconds, as does CheY-P, but might, instead, somehow be a barometer of the metabolic state of the cell. However, the way in which it would facilitate cell survival by reducing bias and switching frequency when present at low concentration and increasing bias and switching frequency when present at high concentration is, at this point, unknown.
One of the unique challenges faced by chemotaxis systems is the necessity for quick responses (on a timescale of seconds) to ever changing environments. This is in constrast to most other two-component signal transduction systems that control gene expression and act over minutes to hours. To cope with this problem, the half-lives of CheY-Ps, are brief, shorter than 1 min (38, 66). CheY is thought to actively catalyze autodephosphorylation, a process involving several conserved residues (two aspartates, a lysine, and a threonine) and a Mg2+ ion (133, 205, 222). However, the half-lives still appear to be too long. To further speed the signal removal, the enteric organisms as well as some other - and ß-proteobacteria express cheZ. The protein further destabilizes CheY-P (257). In these organisms, CheZ is essential for chemotaxis. It was puzzling that most chemotactic bacteria and archaea do not carry a cheZ gene. Based on data available for B. subtilis, it is now apparent that in many if not most of these organisms a com-bination of CheC, FliY, and/or possibly CheX perform this function (234, 234a). These three proteins are homologous but have no sequence similarity to CheZ (Fig. 8). A third mechanism of signal removal has been suggested for S. meliloti and other -proteobacteria where an alternative CheY is thought to act as a phosphate sink and possibly support signal removal (212). The methods of signal removal for different bacteria are summarized in Fig. 9. Interestingly, some organisms do not express a cheZ or cheC homologue or an alternative cheY. Other means of signal removal in these organisms could include the presence of a response regulator domain fused to other chemotaxis components acting as phosphate sink.
View larger version (13K):
[in this window]
[in a new window]
|
FIG. 8. Schematic of CheY-P-hydrolyzing proteins. For CheZ, the C-terminal CheY-P binding region is shown in black and the area including what is thought to be the active site is shaded in gray. For FliY, the CheY-P binding site is shown in black. For FliY, CheC, and CheX, conserved regions are in gray, with highly conserved residues positioned as indicated.
|
|
View larger version (39K):
[in this window]
[in a new window]
|
FIG. 9. Means of CheY-P hydrolysis in chemotactic organisms. The tree was generated as described for Fig. 3. Chemotactic organisms that encode a CheC homolog are highlighted in light gray; dark gray represents organisms that encode a CheZ; black represents organisms that encode an alternative CheY that acts as a phosphate sink; white represents organisms with no known mechanism of CheY-P hydrolysis. V. cholerae encodes both CheZ and CheC homologs and is therefore indicated by dark and light gray squares.
|
|
CheZ Phosphatase
CheZ is found exclusivly in the ß- and -proteobacteria (see the legend to Fig. 3 for type species). From its limited spread among bacteria it can be concluded that it evolved relatively late, and that the original chemotactic organism had other means of signal removal.
As part of the E. coli chemotaxis system, CheZ is by far the best studied CheY-P phosphatase and the only one for which an X-ray structure is available (257). CheZ is localized to the receptor complexes, in the enteric bacteria, via CheA-short (a form of CheA that lacks part of the N-terminal sequence, including the site of phosphorylation [251]) (209). However, the meaning of this localization is not yet clear, since a mutant not capable of making CheA-short does not have a chemotactic phenotype (150). At one time, it was thought that enhancement of CheZ activity might be the means by which CheY-P could be rapidly hydrolyzed following addition of attractant to generate the excitatory signal. However, CheZ does not appear to play any excitatory role (108, 210). Studies of fragments of the protein identified the C terminus of CheZ to be the CheY-P binding domain (29). However, insight into the mechanism of CheZ action remained elusive until the X-ray structure of CheZ in complex with activated CheY was solved recently. Based on this structure, it has been proposed that CheZ residue Gln147 is actively involved in increasing the rate of CheY-P hydrolysis by positioning and activating a water molecule in the active site of CheY-P (257).
CheC/FliY/CheX Phosphatase
Most chemotactic bacteria and archaea do not encode a CheZ homolog and so must cope with the problem of fast signal removal in some other way. A recent study found that the flagellar switch protein FliY in B. subtilis is able to increase the rate of CheY-P hydrolysis. The C-terminal region of FliY is homologous to E. coli FliN. The N-terminal domain is homologous to two other chemotaxis proteins, CheC and CheX (234). While FliY is exclusive to gram-positive bacteria, some spirochetes, and Thermatoga, CheC and/or CheX can be found in almost all phyla of chemotactic organisms, including some proteobacteria. Indeed, B. subtilis CheC shares the ability of FliY to hydrolyze CheY-P (234a). Therefore, conserved residues between CheC and FliY are possibly involved in the chemistry. Six residues—Asp39, Glu43, Asn46, Ser136, Glu140, and Asn143 (following B. subtilis FliY numbering)—are conserved among these proteins, and any could play a similar role to residue Gln147 in E. coli CheZ. Interestingly, the latter three residues are also conserved among CheX proteins, which appear to be truncated versions of CheC. Most chemotactic bacteria and all chemotactic archaea have a CheC homolog. Whether the mechanism of dephosphorylation of CheY-P is similar to that of CheZ will ultimately be shown only by obtaining an X-ray refraction structure of any of these proteins in complex with CheY.
Phosphate Sink
Work on the -proteobacteria S. meliloti and R. spheroides suggests a third mechanism of signal removal (174, 212). In each of these organisms, deletion of at least two CheY homologs causes a defect in chemotaxis. In S. meliloti, CheY2 is the main response regulator that interacts with the flagellar switch and causes the reversal of flagellar rotation. CheY1 does not interact with the switch, although, like CheY2, it is rapidly phosphorylated by CheA-P. CheY2, however, is capable of transferring its phosphoryl group back to CheA and subsequently to CheY1, so that CheY1 may act as a phosphate sink (212). In R. spheroides this mechanism is more complex, since this organism contains six cheY genes; some CheY proteins are thought to act as phosphate sinks (174). Since many chemotactic organisms have more than one cheY, one can imagine that the phosphate sink mechanism may be widespread. In addition, some organisms do not encode any of the known signal-removing proteins and, we speculate, may use response regulator domains fused to other chemotaxis proteins. However, no data suggesting this have yet been reported.
|
MULTIPLE COPIES OF CHEMOTAXIS GENES
|
|---|
As mentioned above, the E. coli chemotaxis system is simple in comparison to most other chemotaxis systems. This is because there is only one copy for each chemotaxis protein. B. subtilis already proves more complex since partially redundant proteins like CheW and CheV or like CheC and FliY make phenotypes less severe and conclusions about them less obvious (99, 181, 234). However, still more complex systems can be found. Some organisms contain multiple sets of chemotaxis genes, some of which may have functions other than controlling motility. P. aeruginosa has five clusters of chemotaxis genes (62, 230). Two of these clusters (I and V) are required for chemotaxis (100, 146). Another (IV) is required for chemotaxis by twitching motility (55, 101), which involves extension and retraction of type IV pili (207). This type of movement is thought to facilitate movement across surfaces and formation of biofilms (170); it usually involves rafts of cells rather than individual cells (147). It seems reasonable that the apparent redundancy of chemotaxis genes in this organism is due to genes within a cluster being devoted to a particular function, such as chemotaxis involving flagella or twitching motility, and is not actual redundancy.
Another organism with multiple copies of chemotaxis-type genes is M. xanthus. M. xanthus has two types of motility, A-motility and S-motility (98, 200). S-motility is homologous to twitching motility in P. aeruginosa and involves extension and retraction of type IV pili (231). M. xanthus has nine clusters of chemotaxis-type genes (14), of which the Frz genes mediate chemotaxis by controling reversals of (gliding) cells, the Dif genes are involved in fibril formation (necessary for S-motility), and the Che4 cluster is also involved in S-motility. However, another set, the Che3 cluster, affects the entry of M. xanthus into the developmental program to produce spores, and the output would appear to be the response regulator protein CrdA (whose cognate histidine kinase appears to be CheA3), predicted to be the transcriptional activator for 
54-dependent promoters (114). Thus, in this case, what must have originally been a chemotaxis-type set of genes controlling motility evolved into a set of genes controlling transcriptional activation. The main difference between ordinary transcriptional activation and this type might be that the latter would undergo adaptation so that the time derivative of the input signal, rather than the magnitude of the input signal itself, would control transcription. Such an arrangement might provide for sensitivity to changes over many orders of magnitude, as in the case for bacterial chemotaxis (41, 54). A detailed account of the issues involved may be found in reference 14.
The -proteobacterium R. spheroides has three sets of chemotaxis genes. cheOp1 contains cheY1, cheA1, cheW1, cheR1, and cheY2. cheOp2 contains cheY3, cheA2, cheW2, cheW3, cheR2, cheB1, and tlpC. cheOp3 contains cheA4, cheR3, cheB2, cheW4, slp, tlpT, cheY6, and cheA3. Besides these chemotaxis genes, there is one encoding a fusion protein of CheBRA and 13 encoding receptors, including 4 cytoplasmic receptors (lacking a membrane-spanning region) and cheY4 (www.jgi.doe.gov/JGI microbial/html/rhodobacter). Deletion of cheA2 prevents aerotaxis, phototaxis, and chemotaxis, but deletion of cheA1 has little effect. Deletion of cheW2 has a much bigger effect on localization (see below) of cheA2 at the poles of the cell than does deletion of cheW3, which marginally affects cheA2 localization. CheA2 causes the phosphorylation of CheY4, and CheY3 facilitates signal temination, possibly acting as a phosphate sink (196). CheA1, with its cognate response regulator CheY5, mediates a repellent ("inverted") response (196). The function of the genes in cheOp3 is unknown. In this organism, the response to negative stimuli is to stop (rather than rotate the single polar flagellum CW, as does E. coli for its peritrichous flagella). On stopping, the flagellum goes from helical to coiled (15). This transition, coupled with rotational Brownian motion, reorients the bacterium (173) so that the next smooth swim will take a new direction.
|
LOCALIZATION OF CHEMOTAXIS PROTEINS
|
|---|
Polar localization of chemotaxis proteins was first explored in Caulobacter, a natural organism with which to investigate polarity since it undergoes differentiation in which a stalk cell produces a swarmer cell with a single polar flagellum that is made shortly before cell division in every generation (249). Later, the flagellum is discarded and is replaced by a new stalk (60). The receptor is located at the pole (6). This expected finding led to an unexpected one, namely, that the receptors of E. coli are also located at the poles of the cell (140). This finding has led to a considerable body of research that has documented that chemotaxis receptors generally are clustered, usually at the pole but, for cytoplasmic receptors as in R. spheroides, at an apparently random place in the cytoplasm (84, 143, 246). The B. subtilis asparagine receptor, McpB (and presumably the other receptors spanning the membrane) is also located at the poles of cells (113). However, the significance of clustering for signal amplification is uncertain, since it was unaltered in E. coli strains lacking CheR or CheB (138), but strains lacking CheB are very impaired in sensitivity to attractants, although strains lacking CheR are still very sensitive (108, 210). Moreover, addition of a multivalent ligand that can bind two receptors simultaneously greatly increases the sensitivity of heterologous receptors to their ligands, and this sensitivity is diminished when other heterologous receptors are deleted (71, 72). Thus, it would appear that clustering of receptors may facilitate taxis since receptors are close to each other but active signaling must require a particular arrangement of the receptors, a goal that is hard to achieve when they are fully methylated.
|
CONCLUSIONS AND PROSPECTS
|
|---|
One great achievement in our understanding of bacterial chemotaxis in the 1970s was the discovery of the methylation system as foundational for bringing about adaptation to stimuli. In the 1980s and early 1990s came the discovery of the two-component system involving phosphotransfer as mediating excitation. Now, during the past decade, there has been a growing appreciation of the diversity of chemotactic mechanisms used in the broad sweep of bacteria and archaea. In this review we have emphasized this diversity. We have acknowledged that many of the principles have been worked out in the E. coli-S. enterica chemotaxis system and that great progress in elucidating that system is still occurring. However, it has become clear that the E. coli system is streamlined and lacks or has significantly modified some basic features of the primordial mechanism that existed when the bacteria and archaea separated during evolution. It seems that many of these features exist in the B. subtilis mechanism, and the elucidation of this mechanism has, accordingly, been one of the features of this review. The processes used to restore behavioral conditions to their prestimulus conditions have changed the most during the streamlining that has led to the E. coli mechanism. However, evolution in other organisms has not stood still, and a lot of changes in other directions have occurred since the primordial mechanism was widely used; we have tried to do justice to these. A lot of information, however, is still at the genome-sequencing level, and more behavioral, genetic, and biochemical work is needed on these organisms. Some of the most interesting and unanticipated advances are occurring in research on organisms that have multiple copies of chemotaxis genes and those that have employed, for controlling development, proteins that once had a chemotaxis function. These new areas, as well as the dynamics of receptor-receptor interactions in bringing about extreme sensitivity to the slightest changes in attractant concentrations over many orders of magnitude, are promising areas of future investigation.
Understanding the structural changes that underlie this remarkable capability is a major challenge; however, we believe that the talented cadre of investigators are up to meeting this challenge. All of these new investigations should serve to make the study of bacterial chemotaxis as exciting during the next 30 years as it has been during the past 30.
This work was supported by National Institutes of Health grant RO1 GM54365.
* Corresponding author. Mailing address: Department of Biochemistry, 190 Med. Sci. Building, University of Illinois, Urbana, IL 61801. Phone: (217) 333-9098. Fax: (217) 333-8868. E-mail: ordal{at}uiuc.edu. 
| 1. | Adler, J., G. L. Hazelbauer, and M. M. Dahl. 1973. Chemotaxis toward sugars in Escherichia coli. J. Bacteriol. 115:824-847.[Abstract/Free Full Text] |
| 2. | Aizawa, S. I., Zhulin, I. B., L. Marquez-Magana, and G. W. Ordal. 2002. Chemotaxis and motility, p. 437-452. In A. L. Sonenshein, R. Losick, and J. A. Hoch (ed.), Bacillus subtilis and its closest relatives: from genes to cells. ASM Press, Washington, D.C. |
| 3. | Aksamit, R., and D. E. Koshland, Jr. 1972. A ribose binding protein of Salmonella typhimurium. Biochem. Biophys. Res. Commun. 48:1348-1353.[CrossRef][Medline] |
| 4. | Aksamit, R. R., and D. E. Koshland, Jr. 1974. Identification of the ribose binding protein as the receptor for ribose chemotaxis in Salmonella typhimurium. Biochemistry 13:4473-4478.[CrossRef][Medline] |
| 5. | Alam, M., M. Claviez, D. Oesterhelt, and M. Kessel. 1984. Flagella and motility behaviour of square bacteria. EMBO J. 3:2899-2903.[Medline] |
| 6. | Alley, M. R., J. R. Maddock, and L. Shapiro. 1992. Polar localization of a bacterial chemoreceptor. Genes Dev. 6:825-836.[Abstract] |
| 7. | Ames, P., and J. S. Parkinson. 1994. Constitutively signaling fragments of Tsr, the Escherichia coli serine chemoreceptor. J. Bacteriol. 176:6340-6348.[Abstract/Free Full Text] |
| 8. | Ames, P., and J. S. Parkinson. 1988. Transmembrane signaling by bacterial chemoreceptors: E. coli transducers with locked signal output. Cell 55:817-826.[CrossRef][Medline] |
| 9. | Ames, P., C. A. Studdert, R. H. Reiser, and J. S. Parkinson. 2002. Collaborative signaling by mixed chemoreceptor teams in Escherichia coli. Proc. Natl. Acad. Sci. USA 99:7060-7065.[Abstract/Free Full Text] |
| 10. | Anand, G. S., P. N. Goudreau, and A. M. Stock. 1998. Activation of methylesterase CheB: evidence of a dual role for the regulatory domain. Biochemistry 37:14038-14047.[CrossRef][Medline] |
| 11. | Appleman, J. A., L. L. Chen, and V. Stewart. 2003. Probing conservation of HAMP linker structure and signal transduction mechanism through analysis of hybrid sensor kinases. J. Bacteriol. 185:4872-4882.[Abstract/Free Full Text] |
| 12. | Appleman, J. A., and V. Stewart. 2003. Mutational analysis of a conserved signal-transducing element: the HAMP linker of the Escherichia coli nitrate sensor NarX. J. Bacteriol. 185:89-97.[Abstract/Free Full Text] |
| 13. | Armitage, J. P. 1999. Bacterial tactic responses. Adv. Microb. Physiol. 41:229-289.[Medline] |
| 14. | Armitage, J. P. 2003. Taxing questions in development. Trends Microbiol. 11:239-242.[CrossRef][Medline] |
| 15. | Armitage, J. P., and R. M. Macnab. 1987. Unidirectional, intermittent rotation of the flagellum of Rhodobacter sphaeroides. J. Bacteriol. 169:514-518.[Abstract/Free Full Text] |
| 16. | Barak, R., and M. Eisenbach. 1992. Correlation between phosphorylation of the chemotaxis protein CheY and its activity at the flagellar motor. Biochemistry 31:1821-1826.[CrossRef][Medline] |
| 17. | Barak, R., I. Giebel, and M. Eisenbach. 1996. The specificity of fumarate as a switching factor of the bacterial flagellar motor. Mol. Microbiol. 19:139-144. |
Top
Next
