Chemotaxis
Chapter
73
JEFFRY B. STOCK and MICHAEL G. SURETTE
Most wild-type strains of Escherichia coli and Salmonella typhimurium (official designation, Salmonella enterica serovar Typhimurium) can synthesize flagella that allow them to move through their medium. The flagellar apparatus, the regulation of its synthesis, and its mechanism of action are discussed elsewhere in this volume (see chapter 10). In this chapter we discuss how cells control flagellar activity so that they move toward favorable environmental conditions. The mechanism involves receptors within the cell envelope that bind stimulatory ligands and undergo conformational changes that regulate the activities of a network of signal transduction proteins within the cytoplasm (for recent reviews, see references 31, 52, 68, 169, and 228).
The flagellar motor apparatus fluctuates between two states: one generates clockwise rotation of the flagellar filament and the other generates counterclockwise rotation (see chapter 10). Counterclockwise rotation is associated with running, smooth swimming motility, where the filaments coalesce into a bundle and push the bacterium along relatively linear trajectories at speeds of up to 40 μm/s (117). Clockwise rotation is associated with tumbling motility, where the filaments work in an uncoordinated manner to jerk the bacterium about without any net velocity. The signal transduction network modulates the frequency of transition between these states. In a constant environment a cell typically moves in a random walk of runs of approximately 1 s punctuated by tumbles of 0.1 s (18). When a cell detects that it is swimming toward increasing attractant or decreasing repellent concentrations, it tends to tumble less frequently, thereby biasing its random walk in the preferred direction (18).
In both E. coli and S. typhimurium there are six cytoplasmic signal transduction proteins, the products of the Che genes: cheA, cheB, cheR, cheW, cheY, and cheZ (17, 46, 196). The che genes from S. typhimurium are fully competent to restore chemotaxis to complement corresponding che mutations in E. coli and vice versa (48). In both organisms these genes are organized in two adjacent operons, mocha and meche (Fig. 1). In addition to these components, each cell has a family of transmembrane proteins with receptor functions (35, 36, 53, 102, 182, 272). These proteins are regulated by methyl esterification at a conserved set of glutamate residues, and they have therefore been termed methyl-accepting chemotaxis proteins, or MCPs (205). Each MCP mediates responses to a specific set of attractant and repellent stimuli.
Biochemical functions have been ascribed to each of the Che proteins as well as the MCPs (Table 1). Each MCP is composed of an extracytoplasmic sensory domain that binds stimulatory ligands connected via a membrane spanning sequence to an intracellular signaling domain (8, 101, 257). The MCPs signal through formation of a ternary complex with the CheW and CheA proteins (29, 62, 128, 156, 188). CheA is a protein kinase that autophosphorylates itself at one of its own histidine residues, His-48 (74). The phosphoryl group is then passed to an aspartate residue, Asp-57, in the CheY protein (75, 184, 269). The phosphorylated form of CheY binds to a flagellar motor component, FliM, to cause tumbly swimming behavior (260). The CheZ protein acts to dephosphorylate CheY (75). Attractant binding to the sensory domain of the MCPs inhibits the kinase activity of the ternary MCP-CheW-CheA complex, thereby decreasing the rate of CheY phosphorylation to favor smooth swimming (29, 156). Kinase activity is also controlled by methylation and demethylation of MCP glutamate residues (28, 156). CheR is an S-adenosylmethionine (AdoMet)-dependent methyltransferase (208) and CheB is a methylesterase (226).
Table 1Chemotaxis proteins and their biochemical activities |
E. coli and S. typhimurium sense levels of attractant chemicals in the periplasm through their binding to receptor proteins. Stimuli and their associated sensors are outlined in Table 2. Some members of the periplasmic binding protein family function as chemotaxis receptors. In E. coli these include the maltose (37, 65), ribose (69), galactose (67), and dipeptide binding proteins (130). In S. typhimurium the periplasmic binding proteins for ribose (5) and galactose (54) play similar roles. The periplasmic binding proteins that function as chemotaxis receptors, as well as all other members of the periplasmic binding protein family, function to deliver their bound ligands to transport systems in the cytoplasmic membrane (for a recent general review of binding protein-dependent transport systems, see reference 77. Those few binding proteins that function in chemotaxis seem to play this role as an adjunct to their primary activity in transport. Thus, for instance, their expression is controlled coordinately with the expression of the corresponding transport system and not with the expression of the fla, mot, and che genes, whose primary functions are in chemotaxis and motility.
Table 2Signal transduction pathways mediating attractant and repellent responses |
The periplasmic binding proteins that function in chemotaxis interact with MCPs (96, 130, 132). MCPs also function independently of periplasmic binding proteins to bind stimulatory ligands (39, 257). In contrast to the periplasmic binding proteins, MCPs are specialized for chemotaxis. They are not known to serve any other function and their expression is coordinately regulated with the che genes.
E. coli and S. typhimurium also exhibit chemotaxis responses that are associated with metabolic processes. The two most studied examples are the responses mediated by the phosphotransferase system that functions in the phosphoenolpyruvate-dependent transport and phosphorylation of a number of sugars, including glucose, and responses mediated by the electron transport system that functions in respiration (see chapters 17 and 75). Thus, the primary glucose chemotaxis receptors are the phosphotransferase enzyme IIs operating in conjunction with cytoplasmic phosphotransferase components, and the primary receptors for oxygen chemotaxis are cytochrome oxidases working in conjunction with components of the electron transport chain. In both of these cases, it is the rate of uptake of the stimulatory ligand that appears to operate to control motility, rather than simply its binding to a receptor protein.
The periplasmic binding proteins are soluble monomeric proteins ranging in molecular weight from 23,000 to 52,000 (77). They are generally composed of two domains, each having a doubly wound α/β fold. The ligand binding site is formed between the two domains from loops at the C-terminal edge of the parallel β-sheets. Ligand binding induces a conformational change whereby the domains come together over the substrate like a Venus flytrap (133, 183). This closed conformational state is thought to be the form that binds to the periplasmic domains of the transmembrane chemotaxis proteins.
There are four MCPs in E. coli K-12. Two, Tar and Tap, are encoded within the meche operon (Fig. 1 and reference 102), and two, Tsr and Trg, are encoded elsewhere in the genome. A fifth MCP homolog has recently been identified by sequence analysis of the E. coli genome (GenBank accession no. U28379), but its function has not been determined. Five MCPs have been identified in S. typhimurium LT2, variants of Tar, Tsr, and Trg and two distinct MCPs termed Tip and Tcp. The S. typhimurium meche operon differs from meche in E. coli in that there is no gene corresponding to tap. The MCPs are all very closely related in sequence (Fig. 2). Each has a short N-terminal extension within the cytoplasm, followed by a hydrophobic transmembrane sequence, TM1; an extracytoplasmic ligand binding domain; a second hydrophobic transmembrane sequence, TM2; a juxtamembrane domain, or linker region; a methylated helix, MH1; a signaling domain; a second methylated helix, MH2; and a variable C-terminal domain. Cross-linking studies have established that S. typhimurium Tar exists in membranes primarily as a homodimer (141). This conclusion has been confirmed by a hydrodynamic analysis of the purified protein in detergent (141).
Sensory Domain of the Aspartate Receptor.
Tar functions to sense aspartate. The purified Tar protein from S. typhimurium exhibits negative cooperativity in aspartate binding, with Kds for the first and second aspartates bound per dimer being 0.10 and 2.0 μM (20). Only one aspartate binding site per dimer has been detected for E. coli Tar, with a Kd of 1.0 μM (20). The aspartate binding properties of the S. typhimurium and E. coli proteins are unaffected by solubilization in octylglucoside (20). Moreover, a genetically engineered fragment of the S. typhimurium aspartate receptor consisting solely of the periplasmic ligand binding domain has the same affinity for aspartate as does the intact Tar protein (143).
The structure of the sensory domain of the S. typhimurium Tar protein has been determined by X-ray crystallographic methods in the presence and absence of aspartate (140, 275). The fragment crystallizes as a homodimer of two four-helix bundles (Fig. 3). Aspartate binds at the dimer interface at one of two nonoverlapping symmetric binding sites. Binding at one site causes the other site to become constrained, and this accounts for the negative cooperativity between the sites that has been observed with the purified Tar protein in membranes or detergent (20) and with the isolated sensory domain (143). The relative positioning of helices within the homodimer is almost unperturbed by aspartate binding, however, and no definitive conclusions can be made concerning the ligand-induced conformational changes that lead to transmembrane signaling (93). The X-ray analysis was performed in high LiSO4, and a sulfate anion is located within the aspartate binding site with salt linkages to two arginine residues that are involved in binding the aspartate α-carboxylate anion (Fig. 4). It is therefore possible that the crystal structure obtained in the absence of aspartate retains a conformation that is similar to that of the aspartate-bound form of the protein.
The structure of the sensory domain of S. typhimurium Tar confirms and extends genetic studies designed to assess the roles of specific residues in the E. coli protein (60, 100, 148, 189, 268). E. coli Tar, unlike S. typhimurium Tar (43), binds the liganded conformation of the maltose binding protein (95, 100, 132, 177, 204). E. coli tar mutants selected for loss of responsiveness to aspartate but not maltose have missense substitutions within a set of arginine codons, Arg-64, Arg-69, and Arg-73 (148, 268). In the structure these residues are directly involved in aspartate binding. Conversely, E. coli tar mutants that are responsive to aspartate and not maltose have missense substitutions that lead to alterations in residues that are in the loops between helices 1 and 2 and between helices 3 and 4 on the external surface of the Tar protein (60, 189). This is consistent with a site of interaction between Tar and the maltose binding protein that was predicted on the basis of simulated docking computations using the X-ray crystal structures of the two proteins (234, 235). This analysis, which involved the generation of a model for E. coli Tar based on the structure of its S. typhimurium counterpart, indicates a surface of interaction involving loops from different Tar monomers within the Tar homodimer. In the modeled complex one aspartate binding site is occluded while the other remains accessible to solvent. The two symmetrically opposed sites within the Tar dimer where the binding protein could interact are mutually exclusive because of steric overlap. These results provide a structural basis for beginning to understand the observation that aspartate and maltose responses are additive and independent despite the fact that both are mediated by Tar (147).
Although results from disulfide cross-linking studies indicate that S. typhimurium Tar in detergents, mixed micelles, or phospholipid bilayers is predominantly a dimer, there is a rapid exchange of subunits, with half times for equilibration of a few minutes (141). Concentrations of aspartate of 1 to 10 μM dramatically inhibit this exchange, as would be expected for aspartate binding at the dimer interface. Results with aspartate binding to the isolated sensory domain also indicate an effect on dimer formation consistent with binding at the dimer interface (143). In the absence of aspartate, sensory domain monomers associate with a Kd of 0.5 to 5 μM. At concentrations of aspartate in the micromolar range, the Kd for dimer dissociation is decreased by at least a factor of 100. Finally, the effect of aspartate binding at one site on the symmetrically opposed site is clearly evident in the structure with aspartate bound, which shows half-the-sites occupancy. The unoccupied site within the dimer is clearly constrained because of a slight conformational change both within each monomer and at the interface between monomers that readily accounts for the negative cooperativity in aspartate binding (20).
Sensory Domains of Tsr and Trg.
The structure of the sensory domain of Tar provides a basis for understanding the sensory domains of other members of the chemoreceptor family. The Tsr sensory domain has been modeled from the Tar structure (82), and the predicted serine binding pocket obtained from this exercise involves residues that have been identified by mutational studies to be essential for serine sensing (70, 110). Like the E. coli Tar protein, E. coli Tsr exhibits half-the-sites reactivity, with one serine residue binding per dimer (113). Reported values for the Kd for serine range from 5 to 29 μM (39, 113).
Genetic studies have been performed with trg, where mutants were targeted to a region of the gene encoding residues 69 to 88 that corresponds to residues 59 to 78 within the ligand binding site in Tar (270). Of nine different substitution mutations at seven sites that exhibited a phenotype that appeared to mimic ligand occupancy, all were at the predicted dimer interface between helices 1 and 1'. These receptor variants could still respond somewhat to ribose and galactose, however. Other mutations caused insensitivity to ribose and galactose, and these occurred in codons that encode residues at the distal end of helix 1 corresponding to the region of Tar that is thought to interact with the maltose binding protein.
Transmembrane Domain.
The transmembrane domain of Tar has also been extensively investigated (Fig. 5). This region contains four α-helices: two transmembrane α-helices, TM1 and TM2, from each monomer. On the basis of the structural studies of S. typhimurium Tar, TM1 and TM2 are thought to be contiguous α-helices with the first and last helices of the sensory domain, helices 1 and 4 (140). The relative positioning of the transmembrane helices has been investigated by cysteine cross-linking studies with S. typhimurium (124) and E. coli (165) Tar. Neither protein contains any cysteine residues. These were introduced by site-directed mutagenesis at various positions within the transmembrane sequences, and rates of cross-linking within Tar dimers were examined. The results are consistent with a loose four-helix bundle arrangement, with the TM1s being closely associated along one face and the TM2s symmetrically opposed to one another and peripheral to the central axis defined by the TM1s. Aspartate binding increases the rate of TM1-TM1' cross-linking with cysteine residues at positions 18 and 19, suggesting that changes in this interface may be associated with transmembrane signaling (124). Cysteine cross-linking studies predict similar arrangements of TM1 and TM2 for the Trg receptor (66).
In another study (83), site-directed mutagenesis was used to introduce Phe residues in positions 199 to 208 within TM2 of the E. coli Tar receptor. Only the Ile-204–Phe mutation caused a significant loss in Tar function. Other bulky hydrophobic residues substituted at this position, Leu, Tyr, and Trp, had a similar effect, whereas Val and Ala substitutions did not. From these results the authors suggest that Ile-204 lies on the interface of TM2 and the other transmembrane helices. The results clearly show that other positions in TM2 are relatively insensitive to the effects of mutations. This helps explain why the TM2 sequence is so poorly conserved between different MCPs.
Revertants and second-site suppressors have been selected in E. coli tar, starting with a mutation in TM1, Ala-19–Lys, that cannot mediate responses to aspartate (161). Mutation of Lys-19 in the mutant to Gln, Ile, or Thr could all restore chemotactic function, as could a second-site Val-17–Glu mutation. Several second-site suppressor mutations also mapped to TM2, Trp-192–Arg, Ala-198–Glu, Val-201–Glu, and Val-202–Leu. However, most suppressors mapped between Thr-294 and Thr-303 within the linker domain between TM2 and MH1. These findings support the view that the transmembrane helices provide a function that is relatively insensitive to substitutions at specific residues.
The results from cross-linking and mutagenesis studies are consistent with a transmembrane signaling mechanism whereby ligand binding between the sensory domains induces a movement of TM2s with respect to the central axis provided by the relatively fixed interface between TM1s (Fig. 6).
Methyl-Accepting Domains.
The portion of Tar distal to TM2 is located within the cytosol. This C-terminal region may be proteolytically cleaved from the membrane-associated ligand binding and transmembrane domains, or it may be produced as a stable independently folded protein from genetically engineered plasmid expression vectors. Circular dichroism spectroscopy on the isolated cytoplasmic portion of Tar indicates that it is highly α-helical (146), and hydrodynamic measurements are consistent with a rod-shaped monomeric structure (116). Electron microscopic studies of the intact receptor in multilamellar vesicles suggest a more globular cytoplasmic region (14). The hydrodynamic estimates are more reliable than the electron-microscopic results, which rely on several questionable assumptions. The possibility cannot be discounted, however, that the structure of the soluble cytoplasmic region differs significantly from that of the intact protein.
Analysis of the sequences of the MCPs with a program that determines the propensity for formation of α-helical coiled coils indicates two coiled coil sequences within the cytoplasmic region (123). These domains, designated methylated helices MH1 and MH2 (Fig. 7), contain the glutamate and glutamine residues that are methyl esterified by the AdoMet-dependent methyltransferase CheR and demethylated and deamidated by the methylesterase/amidase CheB. In the three MCPs that have been analyzed, Tar, Tsr, and Trg, MH1 contains three or four sites of glutamate modification (85, 86, 87, 91, 157, 176, 246, 247). Three of these sites occur every 7th residue over a 21-residue sequence, presumably along one face of the MH1 α-helix. Trg and Tsr contain a fourth site in this region (87, 157, 176). In Trg this site has been identified with a glutamate residue that is immediately proximal to the site that corresponds to the second methylated glutamate in Tar (157). In Tar and Trg there only appears to be one site of methylation in MH2. In Tsr there is a second site nine residues distal to the site shared by all three MCPs.
Some of the glutamyl residues that are subject to methylation are encoded as glutamine residues. The methylesterase that catalyzes the demethylation reaction also functions as an amidase to deamidate glutamine residues at these positions and produce glutamate residues that are then subject to methyl esterification by CheR (179, 192).
Signaling Domain.
The region between MH1 and MH2, termed the signaling domain, is the most highly conserved sequence between different MCPs (Fig. 8). Even in an organism as distantly related as the archaebacterium Halobacterium salinarium, MCPs have been identified with signaling domains that are ∼35% identical to a central portion of the signaling domains of the MCPs in E. coli and S. typhimurium (274). Many E. coli mutants have been isolated with missense substitutions in this domain (7, 8, 150). These are frequently dominant, with different alleles causing cells to exhibit either a predominantly smooth swimming or tumbly phenotype.
Genetic studies have implicated the signaling domain as a site of interaction with CheW in that second-site suppressors of CheW missense mutations map to this region of the E. coli tsr gene (115). This conclusion has been confirmed by the finding that the signaling domain may be produced as an independently folded protein that binds CheW (M. G. Surette, unpublished result).
A variety of fragments of the cytoplasmic regions of the E. coli tar and tsr genes have been engineered into plasmid expression vectors, and the effects of their products on swimming behavior have been examined (9, 150, 160). Fragments that include the region that encodes the signaling domain frequently inhibit chemotaxis and cause either a predominantly smooth swimming or a predominantly tumbly phenotype.
Fragments of the cytoplasmic regions of Tar and Tsr, both from the wild type and from mutants that cause dominant smooth or tumbly phenotypes, have also been isolated and characterized in vitro (9, 84, 116). The wild-type Tar "c-fragment" (extending from Met-257 to the C terminus) and c-fragments with tumble-inducing point mutations behaved as extremely elongated monomeric proteins (116). In contrast, Tar c-fragments that contained mutations that cause smooth swimming tended to form dimers or higher-order oligomers under physiological conditions.
The effect of similar fragments of the E. coli Tsr protein on CheA kinase activity has also been examined. At concentrations that are over 10-fold higher than the concentrations of receptor in wild-type cells, the fragments cause a marked stimulation or inhibition of CheA kinase activity that correlates with their effects when overproduced in intact cells (9). These effects are highly cooperative and increase with preincubation for over an hour. It seems likely that some sort of aggregation phenomenon is involved which may or may not be relevant to the mechanism of function of the intact Tsr protein.
Linker Region, N Terminus, and C Terminus.
Linker region. The sequence between TM2 and MH1, termed the juxtamembrane or linker region, is relatively variable in different MCPs. In a study of E. coli Tar, where mutations were selected that compensated for a mutation in TM1, most fell in this region (161). Several mutations of E. coli Tsr that bias or lock the receptor output are also localized to this region (7, 8). Presumably, the linker plays a critical role in transducing the conformational change induced by ligand binding. Sequence analyses of other bacterial receptors that function to regulate histidine kinase activities in response to environmental signals, NarX (41) and EnvZ (15), indicate homology in the linker region, suggesting a conserved mechanism of transmembrane signaling. A common mechanism seems likely in view of the fact that receptor hybrids between the E. coli Tar or Trg signaling domain and the EnvZ kinase domain (15, 253) produce chimeras that can regulate porin expression (the function of EnvZ signaling) in response to aspartate or ribose and galactose.
N terminus. The N terminus of the MCPs constitutes a small cytoplasmic extension of TM1 consisting of 6 residues in Tar, Tsr, and Tap, 5 residues in Tcp, and 16 residues in Trg. The sequence proximal to TM1 is somewhat conserved, including a completely conserved Ile residue at position 5 of Tar and an overall basic charge (102). The results of cysteine cross-linking studies on the N-terminal cytoplasmic extension of E. coli Tar indicate that this region is helical, with the face containing residue 4 from each subunit being in close contact. Aspartate binding enhances the rate of cross-linking of a Cys residue at this position. The receptor is still responsive to aspartate binding, as measured by increased methylation rates (233). The N terminus of the S. typhimurium Tar protein contains a formylmethionine residue (142).
C terminus. The C-terminal regions of the MCPs beyond the end of MH2 show considerable variability in both sequence and length. Trg has the shortest C-terminal extension beyond MH1 (26 residues shorter than Tar), followed by Tap, Tcp, and Tsr (16, 7, and 4 residues shorter than Tar, respectively). A truncated S. typhimurium tar encoding a protein with 35 residues deleted from its C terminus assembled normally into the membrane and bound aspartate but could not function in aspartate chemotaxis (182). This truncated receptor was capable of responding to aspartate by causing a smooth swimming bias but was unable to adapt to the presence of attractant. The protein was not effectively methylated by the CheR methyltransferase.
The CheA protein mediates the transfer of information from MCPs to the cytoplasm. A crucial step in this process is the transfer of a γ-phosphoryl group from MgATP to the imidazole side chain of a histidine residue, His-48, near the N terminus of the CheA protein (74, 221), as follows: ATP + His-48–CheA ↔ ADP + P∼His-48–CheA. This reaction is catalyzed by the CheA protein and occurs autocatalytically in the absence of any additional components (76, 269). CheA forms a ternary complex with MCPs and an 18,000-molecular-weight soluble protein, CheW, and when bound within this complex, the histidine kinase activity of CheA is regulated both positively and negatively by sensory inputs that impinge on the MCPs (28, 29, 62, 156).
Histidine Kinase Domain.
CheA is a member of the two-component histidine kinase superfamily that mediates a number of different signal transduction processes in microorganisms and plants (214). Almost 100 different histidine protein kinases have now been sequenced (for reviews, see references 78, 170, and 230). All of these proteins share a characteristic kinase domain. An analysis of sequence similarities between CheA and other members of the family indicates that this domain is located between residues 260 and 510 within the CheA protein (Fig. 9). Mutations carried by CheA that map to this central region are generally deficient in their ability to autophosphorylate (158). The kinase domain appears to be an independently folded unit that functions to bind MgATP and catalyze the phosphorylation of histidine residues. The domain contains several highly conserved sequence motifs which have been termed the N, D, F, and G boxes (232). Presumably, these motifs are arranged in the tertiary structure of the protein to form a nucleotide binding surface within the active site. The G box is a glycine-rich sequence reminiscent of, but distinct from, the glycine-rich sequences found in many kinases and other nucleotide binding proteins (186, 187, 245). The D box contains a conserved DXG motif. In serine, threonine, and tyrosine protein kinases, the aspartate β-carboxylate in a highly conserved DXG sequence functions to coordinate Mg(II) and the backbone amide of the glycine H bonds to the backbone carbonyl of the aspartate to hold it in position (245).
Although lysine residues are commonly involved in kinase function, there are no conserved lysine residues in histidine kinase domains. One might suppose that an essential lysine residue might not have been detected because it was flanked by sequences that are relatively divergent, but there are several examples of histidine kinases that have no lysine residues within the entire first subdomain leading up to the N box, within the regions between the N and D boxes, within the region between the D and G boxes, or in the entire C terminus beyond the G box. The lack of a conserved catalytic lysine residue distinguishes the histidine protein kinase superfamily from other kinases that have been characterized. In terms of function, histidine protein kinases are closely related to serine, threonine, and tyrosine protein kinases, but even a cursory examination of the conserved sequences in the two proteins clearly shows that they have different structures and catalytic mechanisms. Although both kinases have G-rich sequences, in the serine, threonine, and tyrosine kinases this motif is located near the N terminus, whereas in histidine kinases the G box is near the C terminus. The only other common motif, the DXG sequence, is located about 20 residues to the N terminus of the G-rich region in histidine kinases, whereas it is C terminal to the G-rich region in serine, threonine, and tyrosine kinases in a region associated with the substrate binding subdomain.
Serine, threonine, and tyrosine kinases also differ from histidine kinases in that the product is a phosphoester rather than a phosphoramidate. Phosphoramidates such as protein phosphohistidine have a large standard free energy of hydrolysis (259), and this appears to be the case for phospho-CheA, which has a phosphotransfer potential of 1 to 3 kcal (1 cal = 4.184 J) per mol greater than that of MgATP (241, 269). The forward reaction in vivo is presumably driven by a high ratio of ATP to ADP and by rapid transfer of the phosphoryl group from the phosphohistidine to a response regulator protein. In contrast, the phosphotransfer potential of serine and threonine phosphoesters is several kilocalories per mole lower than that of MgATP, so unlike histidine phosphorylation, ATP-dependent phosphorylation at serine and threonine residues is essentially an irreversible reaction (195). Phosphotyrosines have a phosphotransfer potential that is closer to MgATP (58, 79), but these phosphorylations also appear to be thermodynamically favored.
CheAS and the Site of Histidine Phosphorylation.
The site of histidine phosphorylation in CheA has been shown by peptide mapping to be His-48 (74). There is no evidence for any other phosphorylated residue, and substitution mutagenesis of His-48 completely abolishes autophosphorylation activity (74). The stability of the phosphohistidine group in CheA is precisely what would be predicted for a 3-phosphohistidine as opposed to a 1-phosphohistidine, which is significantly more labile (269). The identification of a 3-phosphohistidine has been confirmed by phosphorus nuclear magnetic resonance (NMR) analysis of the intact phosphorylated CheA protein (G. S. Lukat and J. B. Stock, unpublished observations). A report of cochromatography of a phosphohistidine residue isolated from CheA with 1-phosphohistidine involved a procedure that would not have resolved the 1- and 3-phosphohistidine isomers (74).
The cheA gene from both E. coli (94, 201) and S. typhimurium (214; E. G. Ninfa, Ph.D. dissertation, Princeton University, Princeton, N.J., 1992) has an alternative translational start site at Met-98 that produces a short variant called CheAS. CheAS lacks a sequence of almost 100 residues that contains the site of histidine phosphorylation. This region of CheA constitutes the major portion of a distinct, independently folding N-terminal domain that serves as a substrate for phosphotransfer from the kinase domain (145, 239). A genetically engineered N-terminal fragment of CheA consisting of the first 134 residues of the E. coli protein can be phosphorylated by an engineered fragment encoding residues 260 to 537 that contains the kinase domain (239). This domain, here termed the phosphotransfer domain, appears to be delineated by a protease-sensitive region approximately 134 residues from the N terminus at the beginning of a sequence of more than 20 residues that is poorly conserved between the E. coli and S. typhimurium cheA genes and is presumably a linker, L1, between the phosphotransfer domain and the remainder of the CheA protein (145, 239).
CheY Binding.
CheA binds CheY with a Kd of 1 to 2 μM (188; J. Li, R. V. Swanson, M. J. Simon, and R. M. Weis, Biochemistry, in press). The site of binding does not seem to involve the phosphotransfer domain, however, since CheAS binds CheY as well as does the full-length CheA protein (239). A genetically engineered fragment of CheA consisting of the first 233 residues of the protein also binds CheY (239). The complex of CheA1–233 and CheY elutes during molecular sieve chromatography at a position consistent with a stoichiometry of 1:1. CheA contains a second short protease-sensitive region extending from residues 230 to 260 in E. coli that is poorly conserved between E. coli and S. typhimurium and is presumed to be a linker, L2, between the CheY binding domain and more distal portions of the protein (145, 239).
CheA Dimerization.
Ultracentrifugation studies with the purified CheA protein indicate that at micromolar concentrations it is an elongated homodimer (61). Because of its nonspherical shape, CheA elutes during molecular sieve chromatography with an apparent molecular weight of 240,000 (214). The CheA autophosphorylation reaction can occur in trans, with the kinase domain from one subunit catalyzing the phosphorylation of His-48 in the second subunit. Thus, CheAS or a full-length CheA with a His-48–Gln mutation can complement a cheA mutant with a defective kinase domain (238, 267). At submicromolar concentrations of CheA, the kinetics of autophosphorylation exhibit a second-order dependence on CheA concentration that is consistent with an obligatory bimolecular reaction (M. G. Surette et al., submitted for publication). From these studies it is apparent that the Kd for CheA dissociation is approximately 0.2 to 0.4 μM. Most studies of kinase activity have been performed at significantly higher concentrations, and under these conditions the reaction is first order in CheA (61, 75). From the kinetics of CheA autophosphorylation (Ninfa, Ph.D. dissertation) and subunit exchange studies (238, 267), it is clear that CheA dimerization is a much more dynamic process than had originally been appreciated.
Which portions of the CheA protein are responsible for dimerization? One possibility is that the interaction between the kinase and phosphoaccepting domains causes dimerization. After all, it is clear that these domains must interact for phosphorylation at His-48 to occur in trans. If the kinase and phosphotransfer domains of each CheA monomer within a dimer bind to one another at independent and nonoverlapping sites, then the minimum Kd expected for the dimer would be the square of the Kd of a single phosphotransfer domain-kinase domain complex. Conversely, the minimum Kd for dissociation of the complex between isolated domains would be the square root of the Kd for the CheA dimer. Kinetic data from an analysis of the autophosphorylation activities of mixtures of various mutants and fragments of CheA are inconsistent with this idea, since CheAS participates in dimer formation as well as CheA (267). A mutant CheA with a defective kinase domain can form a heterodimer with CheAS just as well as with the wild type since a 1:1 mixture of these two proteins exhibits about 25% the rate of autophosphorylation as an equivalent concentration of wild-type CheA.
In an analysis of the family of histidine kinases that mediate signal transduction in bacteria, Parkinson and Kofoid (170) have argued that the conserved kinase domain is actually composed of two distinct subdomains. All of the most highly conserved residues, the N, D, F, and G boxes, are located within the putative C-proximal subdomain, whereas the N-proximal subdomain is relatively poorly conserved, especially in CheA. Recent studies indicate an effect of mutations in this subdomain on the kinase activity of an associated wild-type CheA monomer as though the site of dimerization involved this region (240). This makes sense in terms of the structure of most other histidine kinases where phosphorylation is also in trans (155, 166, 273) and the site of histidine phosphorylation is located at the N-terminal border of the N-proximal subdomain (153, 178).
CheW Binding.
CheW is a monomeric 18,000-molecular-weight soluble protein (61). The protein tends to dimerize and form higher-order aggregates when pure preparations are frozen and thawed or stored at –20°C (216). CheW binds to CheA with a Kd of approximately 15 μM and a 1:1 stoichiometry (61). Under these conditions, ultracentrifugation studies clearly showed that CheA was a dimer, with each subunit binding CheW independently.
The C-terminal portion of the CheA sequence that follows the kinase domain is thought to be a distinct region that interacts with CheW and perhaps with the receptor as well. The E. coli CheA protein is 654 residues long. The kinase domain extends at least to Gly-498 (214). Nonsense mutations introduced at various positions beyond this point produced C-terminally truncated CheA variants that were isolated and characterized for autophosphorylation ability (32). A mutant protein that terminates at Thr-500 lacks kinase activity, whereas mutants terminating at Pro-537 or at any of several sites distal to this residue appear to have normal kinase activity. Regardless of their kinase activity, however, all mutations that produced truncated CheA proteins with 33 or more residues missing from their C termini were unable to support chemotaxis in vivo. In reconstitution experiments, the kinase activities of CheA variants that terminated between Thr-537 and Ser-621 were not subject to activation or inhibition by Tar-enriched membranes and CheW.
Kinetics in the Absence of Other Che Proteins.
The autophosphorylation reaction catalyzed by pure CheA is relatively complex and involves a number of intermediate reactions. The essential features are kinase dimerization, nucleotide binding, and histidine phosphorylation. There are potentially six forms of each monomer that need to be considered: kinase alone, kinase with ATP bound, kinase with ADP bound, phosphorylated kinase, phosphorylated kinase with ATP bound, and phosphorylated kinase with ADP bound. Allowing all combinations of these within a dimer gives an additional 21 forms, for a total of 27. By using different kinetic methods, one can eliminate many of these to make an analysis experimentally tractable.
The simplest interaction to examine is CheA dimerization. Experiments performed with pure CheA indicate a rate of subunit exchange with half times on the order of minutes (238). Kinetic studies indicate that the Kd for dissociation of CheA homodimers is 0.2 to 0.4 μM (Surette et al., submitted). Nucleotide binding, phosphorylation, CheY binding, and CheW binding do not appear to affect CheA dimerization.
Measurements of the kinetics of autophosphorylation of CheA indicate Kms for MgATP(K ATP) of 0.2 to 0.3 mM, consistent with results obtained from direct measurements of MgATP binding, using the method of Hummel and Dreyer (241). The time course of CheA phosphorylation fits a simple exponential, with a stoichiometry approaching one phosphoryl group per monomer (75, 241). The MgATP-binding site of CheA also appears to function independently of dimerization. From these findings, if it is assumed that CheA autophosphorylation only occurs within the CheA dimer, CheA2, a simplified kinetic model for the autophosphorylation reaction can be formulated:
According to this model, the initial rate of autophosphorylation under conditions where CheA is predominantly in a dimeric state can be expressed in terms of a Michaelis-Menten equation where the initial velocities of the reaction correspond to k cat (CheAT), where (CheAT) corresponds to the total concentration of kinase active sites, which is simply twice the concentration of dimer. Alternatively, under conditions where there is a significant concentration of monomer, such as occurs at submicromolar concentrations of CheA, the effective kinase concentration would still be twice the dimer concentration or:
where Kd is the dissociation constant of the kinase dimer. Under these conditions the initial rate of phosphorylation at saturating concentrations of MgATP equals 2k cat(CheA2). When autophosphorylation kinetics are analyzed using this approach, the apparent turnover numbers obtained (i.e., k cat) are extremely low, typically about 10/min.
Pure CheA catalyzes an exchange of phosphoryl groups between ATP and ADP (30, 156, 241; Surette et al., submitted). This exchange reaction can be measured by assaying the ATP-dependent formation of 3H-ATP from 3H-ADP. According to the model outlined above, this reaction involves intermediate phosphorylation of CheA within the dimer:
CheA2 + ATP ↔ CheA2 · ATP ↔ P∼CheA2 · ADP ↔ P∼CheA2 + ADP
P∼CheA2 + A*DP ↔ P∼CheA2 · A*DP ↔ CheA2 · A*TP—> CheA2 + A*TP
The reaction has been measured as a function of ATP and ADP concentrations (Ninfa, Ph.D. dissertation). The results provide estimates of k cat as well as nucleotide binding constants for the phosphorylated and dephosphorylated forms of the enzyme. Values obtained for MgATP binding and phosphotransfer to histidine by this method are similar to those derived from measurements of initial rates of autophosphorylation. The values obtained for MgADP binding to phospho-CheA are similar to those for MgATP and MgADP binding to dephospho-CheA, but MgATP bound with almost threefold less affinity to phospho-CheA than to the dephospho form. Apparently, the phosphorylated histidine residue interferes with the positioning of the ATP γ-phosphoryl group at the active site. Similar kinetic constants have been obtained in experiments where MgADP binding to dephospho-CheA was directly measured using the Hummel-Dreyer method (241). Measurements of rates of ATP/ADP exchange at submicromolar CheA concentrations are consistent with a Kd for CheA dimer dissociation of 0.2 to 0.4 μM (Surette et al., submitted).
The nucleotide binding properties of CheA indicate equal or higher affinities of dephosphorylated CheA for ADP compared with ATP (241; Surette et al., submitted). This means that under the saturating nucleotide concentrations one would expect in vivo, 3 mM ATP and 0.25 mM ADP (24), the rate of autophosphorylation, V, should be sensitive to the ratio of ATP to ADP in accord with the following relationship: V = V max{(ATP)KADP/[(ATP)KADP + (ADP)KATP]}, where KATP and KADP correspond to the dissociation constants of the corresponding nucleotides for the dephosphorylated kinase. Thus, for the binding constants estimated for CheA, a 50% decrease in the ATP concentration coupled to a twofold increase in ADP would be expected to cause about a 20% decrease in the rate of histidine phosphorylation.
Sensory Regulation by MCPs.
CheA together with CheW binds to MCP signaling domains to form ternary complexes of CheA, CheW, and MCP (29, 62, 156). CheW appears to have independent binding sites for CheA and the MCPs and thereby functions to bring these proteins together (62, 156, 238). The stoichiometry of components within the ternary complex is 2:2:2 (62). The rate of CheA autophosphorylation within the ternary complex can be more than 100-fold greater than the rate of autophosphorylation of CheA alone. Neither CheW alone nor Tar alone has a significant effect on CheA activity, however. All three proteins must be together for activation to occur (29, 156).
Evidence for the formation of complexes between MCPs, CheW, and CheA has also been obtained by immunoelectron microscopy and immunofluorescence light microscopy of E. coli cells (128). The complex is frequently found in clusters at one or both poles. These clusters are seen with antibodies directed toward MCPs, CheW, or CheA, but not with antibodies directed against CheY or CheZ. In mutant strains that lack MCPs, CheW, or CheA, the polar distribution and clustering of the remaining components are dramatically reduced.
Aspartate binding to the sensory domain of Tar causes a decrease in the rate of CheA autophosphorylation to levels comparable to those obtained with the pure CheA protein (28, 29, 156). The aspartate concentration dependence for this inhibition fits what would be expected if (i) aspartate binding to Tar within a ternary complex inhibits the autophosphorylation activity of CheA within that complex and (ii) aspartate binds with approximately the same affinity to Tar within a complex as it does to purified Tar or the isolated sensory domain of Tar (29, 228).
The effects of mutations in the E. coli Tar signaling domain (150) that cause a dominant smooth swimming or tumbly phenotype have been examined (29). A Tar mutant that causes smooth swimming, Ser-325–Leu, failed to stimulate CheA activity, and a mutation that caused tumbly behavior, Val-346–Met, activated CheA to about the same extent as did wild-type Tar, but this effect was not blocked by aspartate. It has been proposed that MCP signaling follows from an attractant-induced dimerization of sensory domains that acts to bring two signaling domains into close approximation to one another (102, 116). Since CheA autophosphorylation involves an intersubunit mechanism, one might expect that in a ternary complex MCP dimerization would cause the activation of CheA; but the opposite seems to be the case. Aspartate binding stabilizes the Tar dimer (141), and mutations in the signaling domain that cause predominantly smooth behavior, presumably because they inhibit CheA autophosphorylation, favor formation of oligomers of the C-terminal region of Tar (116). Moreover, the degree of activation of CheA autophosphorylation within the ternary complex is 100-fold greater than could be accounted for by simply a dimerization mechanism. Recent results indicate that association with MCP and CheW may cause a conformational change in CheA that activates CheA independent of dimerization (266).
Phosphoryl groups incorporated into CheA by its autophosphorylation are rapidly transferred to the response regulator protein CheY (75, 269), as follows: P∼CheA + CheY → CheA + P∼CheY. The dephosphorylated form of CheY binds tightly to the CheY binding domain of CheA (188, 239), and phosphorylation causes its release (188). Phospho-CheY binds to the flagellar protein FliM, where it presumably induces a tumbly response (260). CheY has been purified as a 14,000-molecular-weight globular monomeric species (135, 215). The protein is homologous to the conserved regulatory domain of the response regulator superfamily (Fig. 10), all of whose members function to regulate effector outputs in response to sensory inputs from cognate members of the histidine protein kinase superfamily (215, 230, 254).
The structure of CheY has been determined by X-ray crystallographic methods in several laboratories under a variety of conditions (16, 218, 219, 256). There are some minor discrepancies between structures, but most of these differences are now understood. The first solution of the CheY structure (219) provided a relatively low-resolution (0.25-nm) model that contained some inaccuracies, most notably around Pro-110, which is preceded by a cis rather than a trans peptide bond as originally supposed. The 0.17-nm resolution structure obtained under almost identical conditions corrected these inaccuracies (256). Discrepancies between recent structures of CheY with Mg(II) bound (cf. references 16 and 218) are probably due to the binding of a second molecule of Mg(II) that may occur at the extremely high (0.2 M) concentrations of Mg(II) used in the latter study. The results obtained using solution NMR methods under more physiological conditions support this interpretation (149).
All structures of CheY clearly show a doubly wound α/β protein with five α-helices surrounding a five-stranded parallel β-sheet (Fig. 11). The α and β secondary structure elements alternate in the sequence, starting with β1αA and ending with β5αE. The topology of folding is β2β1β3β4β5, with the A and E helices on one side of the sheet and the B, C, and D helices on the other. The sequences of other members of the response regulator superfamily clearly conform with this structure in that all members of the family have hydrophobic residues at positions corresponding to the hydrophobic core of CheY made up of β1, β3, β4, and the helical surfaces with which they interact (231, 254).
The site of phosphorylation in CheY (Fig. 12) is located at Asp-57 in the loop between β3 and αC (184). Mutation of Asp-57 produces a CheY protein that cannot be phosphorylated and is unable to mediate chemotaxis responses (34, 184). This residue is completely conserved in all members of the response regulator superfamily, and adjacent loops connecting β1 to αA and β4 to αD contain residues that are also highly conserved (254). There is generally at least one aspartate residue within the β1-αA loop at a position corresponding to Asp-13 in CheY. In most response regulators this residue is flanked by one or more additional acidic residues, most commonly by an aspartate residue at the position corresponding to Asp-12 in CheY. With only a few exceptions, a serine or threonine residue is located within the β4-αD loop at a position corresponding to Thr-87 in CheY. There is only one other residue that is conserved to the same degree as those in the loops adjacent to the phosphorylated aspartate: a lysine residue within the loop connecting β4 to αD corresponding to Lys-109 in CheY. In CheY the side chain of this lysine extends toward Asp-57 so that the lysine ε-amino group and the aspartate β-carboxylate are in close proximity. Thus, all of the most highly conserved residues in response regulators cluster together around the site of phosphorylation.
The roles of these highly conserved residues in CheY function have been examined both biochemically and genetically. The carboxylate side chains in the acidic residues in the loop between β1 and αA, Asp-12 and Asp-13, act together with the phosphoaccepting β-carboxylate of Asp-57 to help coordinate Mg(II) (16, 119, 121, 152, 218). The binding constant of CheY for Mg(II) is 0.5 to 1.0 mM (121, 152). Metal binding is required for the phosphotransfer reaction (121). Model studies with small-molecule phosphoramidate donors and carboxylate acceptors indicate that divalent metals facilitate the formation of acyl phosphates by shielding the charges on the anions and forming a template for the trigonal bipyramide pentavalent phosphate transition state (72). The geometry of the active site of CheY with Mg(II) bound accommodates such a transition state for phosphotransfer from phosphoimidazole to Asp-57 (218). All that is required is rotation around the β-carbon of Asp-57 to position its β-carboxylate oxygen in proper orientation for nucleophilic attack of an incoming phosphorus with in-line displacement of the imidazole leaving group.
Thus, phosphotransfer from the phosphohistidine in CheA to Asp-57 in CheY appears to be in large part catalyzed by CheY itself. This supposition is confirmed by the observation that small-molecule phosphodonors such as phosphoramidate, acetyl phosphate, and carbamoyl phosphate can act in place of CheA to donate phosphoryl groups to CheY (120). The affinity of CheY for these donors is low, and phosphotransfer rates at millimolar concentrations are limited by the rate of binding of the donor rather than by the rate of phosphotransfer. The reported apparent Kms of approximately 1 mM correspond to the concentrations of phosphodonor required to maintain 50% of CheY phosphorylated under steady-state conditions of phosphorylation and dephosphorylation. The phosphodonor activity of acetyl phosphate seems to be physiologically relevant since its concentrations inside the cell vary from micromolar to millimolar depending on conditions of growth (136, 137, 175).
In considering the phosphotransfer chemistry, attention has focused on the role of the Mg(II) ion in facilitating formation of the transition state. One would expect an additional contribution from a proton donor at the active site since transfer from a 3-phosphohistidine requires protonation at the N1 atom (17). The requirement for protonation at this position explains why phosphohistidine groups are relatively stable in base as opposed to acid. A proton donor would also be expected to facilitate the transfer reaction from acyl phosphate donors such as acetyl phosphate through general acid catalysis. The obvious candidate for this function is the ε-amino group of Lys-109. Thr-87 in the β4-αD loop might also be expected to participate in this aspect of the phosphotransfer chemistry. Thr-87 and Lys-109 are not absolutely required for phosphotransfer (119, 254), but at least in the case of Lys-109, the rate is considerably reduced by conservative substitutions at this position (119).
Thr-87 and Lys-109 are required for the phosphorylation-induced conformational switch that leads to a response since mutations that introduce conservative substitutions at either of these positions prevent CheY from modulating motility (119, 254). Presumably, the mutant proteins cannot undergo a phosphorylation-induced conformational change and therefore cannot interact effectively with the motor. Lys-109 substitutions also block interactions of phospho-CheY with CheZ (21, 119).
Phosphorylation of CheY induces a change in its structure that allows it to bind to the FliM protein at the flagellar motor (260). There is a very close correlation between conditions that are known to cause an increase or decrease in CheY phosphorylation and tumbly or smooth swimming behavior (38, 168). Phosphorylation per se is not sufficient to allow CheY activity, however, since CheY mutants such as Lys-109–Arg can be phosphorylated but cannot bind to FliM or cause cells to tumble (119, 261). Moreover, a CheY mutant has been identified that causes tumbling but cannot be phosphorylated. This phenotype is produced by substitution of the conserved aspartate residue in the β1-αA loop, Asp-13, by a lysine or arginine residue (33, 34). The tumbly bias caused by the Asp-13–Lys/Arg mutation is not affected either by deletion of CheA or by introduction of a second mutation at the phosphoaccepting aspartate, Asp-57–Ala or Asp-57–Glu (33). The structure of Asp-13–Lys CheY has been determined by X-ray crystallographic methods, but no significant conformational differences could be detected from the structure of the wild-type CheY (254) and it has not been possible to show binding of Asp-13–Lys CheY to FliM (261). These negative results have been explained by assuming that Asp-13–Lys can achieve an active conformation in the absence of phosphorylation, but is not locked in the active conformational state, retains an inactive conformation within the crystals used to determine its structure, and has a relatively low affinity for FliM compared with the wild-type phosphorylated CheY protein.
Mutations of cheY have been selected as suppressors of Che– mutants with defects in fliM and other components of the flagellar motor-switch apparatus and vice versa (129, 172, 180, 203, 271). These mutations often involve substitutions for residues located on the face of CheY that is made by β5, αD, and αE (180, 203). From these results, it has been postulated that this is the region of CheY that physically interacts with FliM. One could take a contrary position, however, arguing that the mutations act by causing relatively small changes in CheY structure that perturb its interaction with CheA or CheZ to alter rates of phosphorylation or dephosphorylation and thereby rectify the disruptions caused by mutations in flagellar proteins that raise or lower levels of phospho-CheY required to produce a motor response. Mutations in residues that directly participate in binding to the motor might be expected to have more dramatic consequences that would lead to null rather than suppressor phenotypes. The latter idea is supported by the fact that suppressor mutations in cheZ are obtained as frequently as suppressors in cheY (171, 203). The cheZ mutants are thought to function by raising levels of phospho-CheY, thereby compensating for mutations in the switch that lower its sensitivity to phospho-CheY. Moreover, a detailed analysis of patterns of suppressor mutagenesis in the switch proteins shows that suppression is not allele specific (203). In fact, null mutations in CheZ can be partially suppressed by mutations in motor components.
To further assess the nature of the conformational changes that activate CheY, 4-fluorophenylalanine has been introduced into the protein in place of phenylalanine, and 19F NMR has been used to detect which of these substituted residues undergoes a change in its environment upon phosphorylation (33, 50). The results indicate that phenylalanine residues at the active site, within the hydrophobic core, and near the face of CheY opposite the active site are perturbed by phosphorylation. A phenylalanine residue, Phe-111, in the region that had been thought on the basis of suppressor mutagenesis to interact with the flagellar switch, is relatively unperturbed. Phe-111 contacts Lys-109. From these results it seems likely that phosphorylation induces a global restructuring of CheY that results in significant perturbations at the opposite side of the molecule. CheY variants with activating mutations such as Asp-13–Lys were also examined using the 19F NMR technique. The results confirm the finding from X-ray studies that the global structure is unaffected. In these mutant proteins, however, phenylalanine residues such as Phe-111 near the active site are significantly perturbed compared to wild type. It seems likely that these mutations interfere with the stability of this region of the protein, thereby facilitating the transition to an active conformation.
A recent NMR analysis of perturbations in backbone 15N-1H resonances that are associated with CheY phosphorylation strongly supports the conclusion that the protein undergoes a global conformational change (118). Residues that are significantly shifted include Met-17, Val-21, Asn-23, Gly-39, Asn-59, Met-60, Met-63, Asp-64, Gly-65, Leu-66, Glu-67, Leu-68, Leu-69, Met-85, Val-86, Thr-87, Ala-88, Asn-94, Val-107, Lys-109, Thr-112, Ala-113, Ala-114, and Asn-121. Some of these correspond to residues identified by suppressor mutagenesis to be important to the regulation of levels of CheY phosphorylation and/or to the regulation of the affinity of phospho-CheY for the motor-switch. These include Thr-112 (180) and Ala-113 (203). But there are clearly many changes outside this region. The conformational change associated with CheY phosphorylation has been shown to regulate its interactions with at least three distinct proteins, CheA (188), CheZ (21), and FliM (260). Given the global nature of the conformational change, one cannot assume that only one part of the surface of CheY is involved in all of these protein-protein contacts.
There has been considerable speculation concerning a salt bridge between Lys-109 and Asp-57 (68). Phosphorylation at Asp-57 would clearly break this interaction, and it has been posited that this is a trigger that induces the subsequent conformational change leading to regulator activation (256). It has recently been shown, however, that Mg(II) coordinates with Asp-57 and thereby disrupts its interaction with Lys-109 independent of phosphorylation (149, 218). Thus, in vivo, where CheY would be expected to be in its predominantly metal-bound state, the salt bridge probably has no relevance to either the phosphotransfer chemistry or the induction of a conformational transition. It is not even clear that the salt bridge forms in solution under physiological conditions since mutation of Lys-109 does not affect the binding of Mg(II) (121). The X-ray crystal structure of CheY where the salt bridge was observed was determined in the presence of high concentrations of ammonium sulfate (256). In this environment a sulfate ion is positioned near the active site and an ammonium ion is located within the acid pocket. CheY does not bind Mg(II) under these conditions (218). The absence of an Asp-57–Lys-109 interaction under physiological conditions would explain why Mg(II) binding does not cause a significant change in the NMR signal from a fluorophenylalanine residue at position 111 (50).
At physiological temperatures and pH, a phosphoaspartate group is relatively unstable compared with other types of phosphorylated residues (49, 98). Phospho-CheY denatured in sodium dodecyl sulfate exhibits a half-life of several hours at neutral pH and ambient temperatures, which is what one would expect from results obtained from studies of rates of hydrolysis of small-molecule acyl phosphates (221). Under nondenaturing conditions, however, the rate of phospho-CheY hydrolysis is much faster, with half-times for hydrolysis of only a few seconds (75). This reaction does not depend on the presence of CheA or any other protein (73, 120). The rate is dramatically enhanced by CheZ, which may function by causing a conformational change in phospho-CheY that stimulates its autophosphatase activity rather than by directly participating in the phosphoaspartate hydrolysis reaction.
Mg(II) is required for the dephosphorylation reaction whether or not CheZ is present (121). Studies with small molecules have established that Mg(II) catalyzes the hydrolysis of small-molecule acyl phosphates such as acetyl phosphate with rates of hydrolysis in the presence of molar concentrations of metal that are comparable to the autophosphatase rates obtained with phospho-CheY at millimolar concentrations of metal (75, 98). The transition state for dephosphorylation would be expected to be essentially the same as the transition state for phosphorylation, with water positioned in place of the histidinyl leaving group. It seems likely that the acidic function that protonates the histidine side chain at the N1 position to facilitate phosphorylation functions as a base to enhance the nucleophilic attack by water in the dephosphorylation reaction. In CheY, for instance, mutation of Lys-109 to arginine causes a substantial decrease in rates of both phosphorylation and dephosphorylation (119). CheZ could work by forcing a realignment of groups at the active site to facilitate this chemistry.
Several lines of evidence have established that CheZ functions in vivo to control the level of CheY phosphorylation by regulating its rate of dephosphorylation. The reaction has been demonstrated under defined conditions with purified components (75). To a first approximation the reaction is not affected by CheA or any other chemotaxis component (73, 120). It has also been established that phospho-CheY binds to the CheZ protein and that on dephosphorylation CheY is released from CheZ (21). CheZ does not facilitate the dephosphorylation of Lys-109–Arg, presumably because this protein cannot achieve an active, tumble-promoting conformation (119). The phosphorylated mutant protein does not bind to CheZ (21).
The purified CheZ protein elutes during molecular sieve chromatography at a position predicted for a globular protein of 115,000 molecular weight (220). The monomer molecular weight predicted from the cheZ gene is 24,000, so the purified protein is either very elongated or a multimer or both. A more highly polymerized form of CheZ was also isolated, eluting with an apparent molecular weight of more than 500,000 (220). It has been reported that CheZ forms a complex with CheAS, but not with full-length CheA (134). By immunoprecipitation with either anti-CheA or anti-CheZ antibodies, the ratio of subunits within the complex was estimated to be 10 to 30 CheZs per CheAS.
No evidence has been reported indicating regulation of phospho-CheY dephosphorylation by sensory inputs. Recent results suggest that increases in intracellular Ca(II) cause cells to tumble (250), and studies of the effects of different metals on phospho-CheY dephosphorylation indicate that, even in the presence of Mg(II), Ca(II) causes a substantial inhibition (121).
The MCPs that mediate chemotaxis responses in bacteria are reversibly methyl esterified at specific glutamate residues within α-helical regions termed MH1 and MH2 that flank the signaling domain (Fig. 2 and 7). To date, there is no other example in biology of carboxylmethylation at glutamate residues. Besides the MCPs, there are two enzymes that constitute the chemotaxis methylation system: a methyltransferase encoded by the cheR gene that catalyzes the transfer of methyl groups from the methyl donor AdoMet to substrate glutamate residues in the MCPs (208), and a methylesterase encoded by the cheB gene that catalyzes the hydrolytic cleavage of these methylesters (226). Several sites of glutamate methylation within the MCPs are encoded as glutamine residues in the corresponding genes (Fig. 7). The CheB protein functions as an amidase to cleave these amides and thereby produce glutamyl side chains that are subject to methylation by CheR (179, 192). Thus, the following three reactions are involved in the MCP methylation system (where AdoHcy is S-adenosylhomocysteine):
The level of MCP methylation clearly functions to regulate the activity of the signaling domain. This was established in early studies of the methylation system, where good correlations were observed between methylation and swimming behavior and it was determined that increased levels of methylation favored tumbly behavior and decreased levels favored smooth swimming (for a review of the early work on methylation, see reference 205). More recently it has been shown with purified components that increased levels of methylation or amidation of MCPs within ternary complexes with CheW and CheA increase the rate of CheA autophosphorylation and lowered levels inhibit this reaction (28, 156).
Enzymology.
The CheR protein has been purified as a 30,000-molecular-weight soluble monomeric species that catalyzes the transfer of a methyl group from AdoMet to glutamate residues in the MCPs (200). The Km for AdoMet in this reaction is 10 to 20 μM (197, 200), which corresponds well with the value of 10 to 30 μM that has been estimated from equilibrium dialysis studies to be the Kd for AdoMet binding to the pure CheR protein (27, 197). The product of the reaction, S-adenosylhomocysteine, is a competitive inhibitor with a Ki that is roughly equivalent to the Km for AdoMet (197). From the kinetics of methylation, the Km for Tar was estimated to be approximately 2 μM. Direct estimates of binding to Tar gave a value for the Kd of 4 μM. Moreover, the product of the reaction, methylated Tar, appears to act as a competitive inhibitor as though methylation did not affect CheR binding. The data also showed that AdoMet and Tar bind independently to CheR. The roles of cysteine residues in CheR function have also been investigated. The S. typhimurium enzyme is inhibited approximately 70% by sulfhydryl reagents (237). It has two cysteine residues, Cys-31 and Cys-229. Substitution of Cys-229 by Ser has no effect on activity or sensitivity to sulfhydryl reagents, whereas substitution of Cys-31 by Ser lowers the activity of the enzyme to approximately 20% of the wild type. The Cys-31–Ser mutant is no longer sensitive to sulfhydryl reagents. Since AdoMet protects the wild-type enzyme from the effects of sulfhydryl reagents, it seems likely that Cys-31 is near the AdoMet binding site. Cys-31 is clearly not essential for catalysis since mutation of the residue only causes an 80% reduction in activity, and Cys-31 is not conserved in E. coli which has a Ser residue at this position. The E. coli gene encodes a Cys-7 residue, which in S. typhimurium is a Ser residue (151).
Specificity for Sites of MCP Methylation.
There are four to five specific glutamate residues in each MCP that can function as substrates for methylation by CheR (Fig. 7). In Tar, three of these residues are located in MH1, sites 1, 2, and 3, and one is located in MH2, site 4. Trg has a fourth site just preceding site 2, which we term 2'. An analysis of sequences flanking these sites indicates a consensus, A(S)XXEEXA(T/S)AA(T/S), where E is the site of methylation, residues indicated in parentheses are acceptable varieties at the preceding position, and X denotes a variable residue (157, 248). Sites 1 and 4 in Tar deviate slightly from this consensus, with the conserved Ala residue three residues distal to site 1 being replaced by Ser and the conserved Glu residue preceding site 4 being replaced by Gln. Studies of rates of Tar methylation indicate that site 3 is methylated most rapidly, site 2 is methylated at about 30% of this value, and the two sites that deviate from the consensus, sites 1 and 4, are only methylated at about 2% of this rate (191, 248). However, a mutated Tar where site 3 was altered so that the corresponding Ala residue is substituted by Thr and the following Thr residue is substituted by Ala, Ala–Thr-312/313–Thr–Ala, is methylated at this position almost as well as the wild type (249). Moreover, from an analysis of patterns of Trg methylation, there appears to be a substantial preference for sites 2', 3, and 4, which would not be predicted from the proposed consensus sequence since at position 2' the conserved Glu residue preceding the methyl-accepting Glu residue is replaced by Ile (157).
Sites that are encoded as Gln rather than Glu must be deaminated by CheB prior to methylation. In addition, the 2' site in Trg does not appear to be methylated unless the Gln residue at site 2 is deamidated first (157). Presumably, since methylation generally mimics amidation, methylation at site 2 also precludes subsequent methylation at site 2'. Conversely, in Tsr there is an additional site, site 3', that can be methylated only when the Gln residue at site 3 remains amidated (176). Thus, methylation at site 3 may render site 3' susceptible to methylation. Although amidation, and by extension methylation, at sites on adjacent residues affects rates of modification, rates of methylation at more distant sites do not appear to be affected (157, 176, 248).
The CheB protein is composed of two distinct domains (199, 215): an N-terminal regulatory domain of approximately 120 amino acids, that is homologous to the phosphoaccepting domains of the response regulator superfamily including CheY, and a C-terminal catalytic domain that catalyzes demethylation and deamidation of the MCPs (Fig. 13). The domains are connected by a protease-sensitive linker sequence of 10 to 20 residues that is poorly conserved between E. coli and S. typhimurium (151, 199). Removal of the regulatory domain by either proteolytic cleavage or genetic deletion produces an isolated catalytic domain that is over 10-fold more active in demethylating MCP glutamyl methylesters than is the intact protein (122, 199). The full-length CheB protein and the catalytic domain, CheBC, have been isolated as monomeric species. They elute during molecular sieve chromatography as globular proteins with apparent molecular weights of 37,000 and 21,000, in reasonable agreement with values that would be predicted from their sequences (199).
Catalytic Domain.
The methylesterase activity of CheB is dramatically inhibited by sulfhydryl reagents, and it was proposed that the enzyme is a thiol esterase (198). It was subsequently demonstrated, however, that both Cys codons could be mutated to Ala without adversely affecting CheB activity (103), and further analysis indicated that the enzyme was a serine hydrolase with the primary active-site nucleophile being Ser-164 (103). Mutation of this residue caused a complete loss of esterase activity.
The structure of CheBC has recently been solved to 0.18-nm resolution by standard X-ray crystallographic methods (261a). The domain is essentially a doubly wound α/β structure with seven α/β strands (Fig. 14). An active-site catalytic triad involving Ser-164, His-190, and Asp-286 can clearly be discerned within this structure (Fig. 15). A His-190–Tyr mutation has previously been shown to inactivate the esterase (212).
Regulatory Domain.
The regulatory domain of CheB accepts phosphoryl groups from phospho-CheA, just like its homolog CheY (75, 122, 221). The rates of transfer appear to be comparable. As with CheY, the phosphorylated N-terminal phosphotransfer domain of CheA can function in place of the full-length phospho-CheA protein (74). Moreover, when CheA is complexed to CheW and MCP, phosphorylation of CheB is regulated by sensory inputs in the same way that the phosphorylation of CheY is regulated (156). The autophosphatase activity of phospho-CheB is significantly greater than that of phospho-CheY (cf. references 119 and 210), but this activity is not affected by CheZ or any other known protein (75).
Phosphorylation of the regulatory domain of CheB causes over a 10-fold activation of the associated catalytic domain (122, 221). This finding with purified components provides an explanation as to why repellent-stimulated rates of MCP demethylation in vivo are blocked by mutations in cheA, cheW, and the region of cheB that encodes the N-terminal regulatory domain (122, 206, 211, 212). In a detailed genetic analysis, it was shown that CheB residues are required for CheB activation that, by analogy with corresponding residues in CheY, should be required for phosphorylation. These include Asp-11 and Asp-56, which correspond to Asp-12 and Asp-57 in CheY.
In an extension of this work (210), a series of mutations in the regulatory domain were isolated and their effects on esterase activity in vitro were characterized with and without phosphorylation. In the unphosphorylated state three variants had higher esterase activity than did the wild type: Asp-11–Lys, Glu-58–Lys, and Glu-91–Lys. These mutations could be functioning to favor the activated conformational state of the regulatory domain. This is especially true for Asp-11–Lys since the corresponding mutation in CheY, Asp-13–Lys, seems to favor the active conformation (33, 34). Several other mutations in the regulatory domain caused an inhibition of CheB activity in the absence of phosphorylation (210). These included Arg-42–His, Arg-73–His, and Lys-107–Arg. These all retained some ability to be activated by phosphorylation, but generally not to the same extent as the wild type. Finally, an ATPase assay was used to estimate the kinetics of CheB-mediated phosphotransfer reactions (210). The data were expressed in terms of k 2, the second-order rate constant for phosphotransfer from phospho-CheA to CheB, and k 3, the first-order rate constant for phospho-CheB hydrolysis. The results indicated that Arg-73–His and Asp-91–Lys mutations had no effect on rates of phosphotransfer from phospho-CheA (k 2 = 25 μM–1 min–1) whereas Arg-42–His and Lys-107–Arg caused about a 5-fold reduction, Glu-58–Lys caused a 25-fold reduction, and Asp-10–Asn and Asp-11–Lys reduced rates to undetectable levels. The Lys-107–Arg and Glu-58–Lys mutations caused reductions in rates of dephosphorylation of 50 and 80%, respectively. The wild-type value for k 3 was 42 min–1. By using essentially the same assay, comparable values obtained for CheY are as follows: for the wild-type protein, k 2 = 12 μM–1 min–1 and k 3 = 2.2 min–1; for the Lys-109–Arg mutant, k 2 is reduced sevenfold and k 3 is reduced sixfold (119). The larger rate constants for CheB are partly due to the fact that the CheB assays were performed at 35°C while the CheY assays were performed at 24°C.
The ability of CheB to accept phosphoryl groups from small-molecule phosphodonors has also been investigated. Phosphoramidate can phosphorylate and activate CheB, as it can all other members of the response regulator superfamily that have been examined (55, 120, 136). Unlike most of these proteins, however, CheB is not phosphorylated by acetyl phosphate or carbamoyl phosphate (120).
Effects on CheA Autophosphorylation.
Changes in MCP methylation/amidation affect the rate of CheA autophosphorylation within MCP-CheW-CheA complexes (28, 156). High levels of methylation/amidation cause the rate of phosphorylation of the associated CheA protein to be activated over 100-fold compared with its activity alone. In contrast, low levels of Tar methylation inhibit the rate of autophosphorylation of the associated CheA protein to values as low as those obtained with CheA alone. In either of these extreme states of methylation, aspartate has little effect on CheA activity within a complex with Tar and CheW (28). Under these conditions, CheA appears to be locked in either an inactive or an active conformational state. It is only at intermediate levels of methylation/amidation that aspartate inhibits CheA activity. The role of glutamate modification in the regulation of MCP function can be understood by assuming that the MCP signaling complex is essentially a two-state system, with a tumble-inducing T state where CheA is active and a run-inducing R state where CheA is inactive (11, 225). According to this view, methylation stabilizes the T state and demethylation stabilizes the R state, whereas attractant stimuli stabilize the R state and repellent stimuli stabilize the T state. Thus, the phosphorylation output of an MCP-CheW-CheA complex is controlled by a balance between the effects of stimulatory ligands on the MCP sensory domain and the degree of modification of glutamate side chains in MH1 and MH2.
The molecular mechanism by which methylation controls the conformational state of the MCPs probably relates to coiled coil interactions between MH1 and MH2 domains within the homodimer (Fig. 16). The glutamyl modification reactions, deamidation, methylation, and demethylation, would be expected to affect the net negative charge density at the coiled coil interface (228). For instance, the electrostatic repulsion between glutamate side chains might cause helices to rotate so as to maximize the distance between these charged groups, and methylation could act to lower the charge density and might therefore allow the helices to rotate with respect to one another into a state where the methylated faces are in closer approximation. These ideas fit well with the sorts of models that are being proposed for the effects of stimulatory ligands on sensory domain conformation and transmembrane signaling, where inter- and/or intrasubunit conformational transitions in the sensory domains at the external surface of the cytoplasmic membrane are seen to induce changes in the relative positioning of signaling domains between subunits at the membrane-cytosol interface (93, 125, 164).
Effects on Aspartate and Serine Binding.
Given the counterbalance between the effects of increased methylation and the binding of attractants on CheA activity within the ternary complex, one might expect that methylation, by stabilizing the T conformational state, would oppose the binding of attractants, which stabilize the R state.
To test this possibility, the effect of MCP methylation and amidation on ligand binding has been examined. Considerable evidence indicates that amidation is, to a first approximation, equivalent to methylation in its effects on MCP function (28, 222, 223, 224). In the absence of CheB and CheR, Tar sites 1 and 3 are glutamine residues and sites 2 and 4 are glutamate residues (termed the QEQE form of Tar). This form of Tar is essentially equivalent to the 50% methylated form.
Yonekawa and Hayashi (276) measured serine and aspartate binding to cells from mutants that were defective in CheR (EEEE), CheB (QEmQEm), or CheR and CheB (QEQE). The results indicated that the highly modified Tsr from the CheB mutant membranes could not bind serine. The authors concluded that methylation dramatically lowers the affinity of Tsr for serine. In parallel experiments where aspartate binding was measured, high levels of Tar methylation caused about a 10-fold decrease in affinity.
In a later study, levels of modification were manipulated by using site-directed mutagenesis to introduce glutamine residues at sites 2 and 4 or by introducing glutamate residues at sites 1 and 3. When this was done, the resulting all-E and all-Q forms of S. typhimurium Tar bound aspartate, with Kds of 0.78 and 1.2 μM, respectively (51).
In another study a combination of methylation and site-directed mutagenesis was used to produce membranes with E. coli Tar modified to different levels and the effects on aspartate binding were determined (28). The results indicated Kds of 3.5 μM for EEEE, 7.5 μM for EQEQ, 30 μM for QEmQEm (methylated glutamate residues at sites 2 and 4), and 18 μM for EmEmEmEm. In a separate experiment, CheA and CheW were added to these different forms of Tar to form a ternary complex, and the steady-state level of phosphorylation of CheY was measured as an indication of CheA autophosphorylation rates. Finally, the concentration of aspartate required to cause a 50% decrease in steady-state levels of phospho-CheY was determined. The results were the following: 0.7 μM for the EEEE and QEQE forms of Tar, 120 μM for the QEmQEm form, and >50 mM for the EmEmEmEm form. Thus, methylation caused only about a 10-fold decrease in aspartate affinity, whereas it caused much more than a 100-fold decrease in sensitivity to the effects of aspartate on kinase activity. From this, the authors concluded that methylation-induced changes in ligand affinity cannot be responsible for the desensitizing effects of methylation on MCP signaling. In this study, however, aspartate binding was measured in the absence of CheW and CheA. This may be almost equivalent to measuring binding to the isolated sensory domain. Without having to do the work of converting CheA from an active to an inactive conformation, Tar signaling domains may be relatively free to change their conformation in response to ligand binding to the sensory domain so that any changes in Tar conformation induced by methylation might have relatively little effect on ligand binding. The seemingly contradictory results obtained by Yonekawa and Hayashi (276) may stem from the fact that the MCPs in their membrane preparations may have contained substantial levels of associated CheW and CheA, whereas in later work with genetically engineered Tar variants no ternary complex would have been present.
Levels of MCP methylation are controlled by the rate of methylation and demethylation at each methyl-accepting site. There are at least three distinct mechanisms that are known to regulate these rates in vivo: (i) the methylation rate is modified by changes in the intracellular concentration of AdoMet, (ii) the demethylation rate is modulated by changes in the level of phosphorylation of CheB, and (iii) methylation and demethylation rates are modulated by changes in MCP conformation that alter their ability to serve as substrates for CheR or CheB. Mechanisms i and ii, which operate at the level of enzyme activity, would act globally on methylation and demethylation of all MCPs, whereas mechanism iii, which operates at the level of the MCP substrates, acts differentially for each class of MCPs.
Regulation of CheR Activity.
There is no evidence that CheR is regulated by covalent modification or by any allosteric mechanism. The protein is by far the least abundant of all the Che proteins, with intracellular levels of at most 200 molecules per cell, or ∼0.2 μM (200). Kinetic studies indicate a very low turnover number for the enzyme (about 6 min–1), so that were CheR to be operating at maximal velocity in a cell, there could be at most approximately 20 methylation events per s. This is similar to the maximal steady-state rates of methanol production of approximately 50 s–1 that have been observed with intact cells under conditions where CheB is expressed at over 10-fold the wild-type levels (122). In wild-type S. typhimurium at normal levels of CheB, steady-state rates of methanol production are about half the maximal value, and substantially lower steady-state turnover rates are observed under conditions where cells were nutritionally stressed and probably have lowered intracellular concentrations of AdoMet (122). The observation that steady-state rates of methylation and demethylation approach the maximal rates possible for the methylating enzyme indicates that in vivo CheR may be saturated by MCP substrates. This provides one explanation for the finding that steady-state rates of methylation are nearly constant, independent of the level of methylation (88, 251).
The intracellular pool of AdoMet is maintained at approximately 100 μM under optimal nutritive conditions, but when the rate of AdoMet synthesis declines due to ATP depletion or lowered intracellular methionine, AdoMet can drop to submicromolar levels within minutes (13, 27). Since the Km of CheR for AdoMet is 10 to 20 μM (197), conditions that significantly lower intracellular AdoMet levels would be expected to have a dramatic effect on rates of MCP methylation. The discovery of a role for methylation in chemotaxis stemmed from the finding that cells starved for methionine were deficient in chemotaxis (3, 10, 13, 97).
Regulation of CheB Activity.
CheB is regulated by the CheA-dependent phosphorylation of its regulatory domain, which causes over a 10-fold increase in its esterase activity (75, 122, 212). This accounts for the global regulation of esterase activity by repellent and attractant stimuli originally postulated by Kehry et al. to explain the general inhibitory effect of a dominant tsr mutation that causes a smooth swimming phenotype (89). The idea was confirmed by experiments showing how the effects of multiple stimuli were integrated to give a single demethylation output (90). These results paralleled studies of mutants which showed an essential role for CheA and CheW and the regulatory domain of CheB in the global modulation of CheB activity (122, 206, 211, 212). The fact that mutations in CheA cause about a 50% reduction in steady-state levels of methanol production indicates that CheB is continuously maintained in a partially phosphorylated state, even in the absence of repellent stimulation (122, 206).
In vitro studies of CheB phosphorylation indicate that phosphoryl groups in CheA are partitioned between CheY and CheB with approximately equal efficiency (75, 156). The results suggest that the level of phospho-CheB, and as a consequence its methylesterase activity, may reflect the global autophosphorylation activity of CheA. Since MCP demethylation causes a dramatic inhibition of MCP-associated CheA activity, CheB phosphorylation provides a negative feedback mechanism to maintain CheA activity at an intermediate value.
Effects of MCP Conformation on Methylation.
Attractants cause increases in methylation and repellents cause decreases that are specific to the MCP that interacts with the stimulatory ligand. For instance, saturating concentrations of aspartate induce a 50 to 100% increase in methylation of Tar, but not Tsr, whereas saturating concentrations of serine cause a 50 to 100% increase in methylation of Tsr, but not Tar (205). The increased steady-state levels of methylation induced by ligands such as aspartate and serine remain constant as long as the ligand is present. Thus, for any fixed stimulus intensity, there is a corresponding steady-state level of MCP methylation.
A particularly well-documented example of this is provided by the E. coli Tar protein in experiments where the effects of aspartate and maltose were examined, both separately and in combination (147). When either of these attractants was added separately, there was an increase in the level of Tar methylation, with most of the additional methyl groups being incorporated at sites 2 and 3. Addition of aspartate and maltose together had an additive effect, again with most of the increase at sites 2 and 3. Thus, different attractant stimuli that are sensed by the same MCP appear to effect increases in methylation at the same sites, and when the stimuli are added together, their effects are essentially additive.
Attractant stimuli such as serine (for Tsr) or aspartate and maltose (for Tar) cause changes in the levels of methylation of their corresponding MCPs by increasing their rates of methylation and at the same time decreasing their rates of demethylation. Conditions such as attractant binding and low levels of MCP methylation that tend to inhibit CheA activity in ternary complexes with CheW make the MCPs much better substrates for the methyltransferase and much poorer substrates for the methylesterase (205). These effects do not depend on CheA or CheW, however, since they can be observed with purified MCPs and CheR or CheB (257). This type of CheB regulation occurs independently of the CheB regulatory domain so that, in the absence of CheA and CheW, demethylation of MCPs by CheBC is inhibited by attractants to the same extent as is demethylation catalyzed by the full-length CheB protein (26).
Studies of the turnover of methyl groups in intact cells showed that attractants effectively cause some residues to become inaccessible to CheB. In the absence of attractants, approximately 50% of the MCP methyl groups are continuously being added and removed over a period of several minutes while the remaining 50% are stable for more than 1 h (227). Addition of saturating concentrations of aspartate and serine causes roughly a doubling in the total level of methylation. Most of the methyl groups that had been constantly turning over were now stable for more than 1 h, and most of the newly added groups turned over within minutes. Thus, at the elevated levels of methylation induced by attractant addition, approximately 50% of the methyl groups turned over rapidly while 50% did not. On attractant removal the situation reverted to what had occurred in the absence of attractants, with the groups that were newly added in response to attractants being the ones that were released when attractants were withdrawn (207, 227). Thus, there is essentially an ordered addition and removal of methyl groups, with the last on being the first off.
The apparent preferential ordering of residues for methylation and demethylation (207, 227) seems to be contrary to the observation that attractants cause a 2- to 10-fold increase in the rate of methylation at all sites in Tar (248). The specificity of CheB for different sites has not been determined, however.
All of the results obtained from studies of MCP methylation are consistent with a two-state model for MCP conformation, where the CheA-inactivating, run-inducing R state is a substrate for methylation by CheR and the CheA-activating, tumble-inducing T state is a substrate for demethylation by CheB. Since increases in methylation stabilize the T state and decreases in methylation stabilize the R state, one can view the methylation system as a mechanism to fuel a dynamic equilibrium between receptor conformations at the expense of converting AdoMet to S-adenosylhomocysteine and methanol. This equilibrium would be expected to maintain the MCPs in a balanced conformation. There are eight potential sites of glutamate modification in each Tar homodimer. In a constant environment, once a steady-state level of methylation has been achieved, addition or removal of one or two methyl groups would be expected to cause a significant shift in conformation to the T or R state. The probability that a particular methyl group is added or removed depends on the relative rate constant for methylation or demethylation at that site. The apparent ordered addition and removal of groups is a reflection of differential rates at different sites. These different rate constants coupled to the methylation-driven R/T conformational transitions would be expected to cause methyl group addition and removal to be focused on only one or two glutamate residues.
According to this view, when a cell is exposed to attractant, the corresponding MCP undergoes a conformational transition to the R state and this transition inhibits CheA autophosphorylation and causes the cell to smooth swim. Concomitant increases in the level of methylation of the MCP cause its conformation to shift back to the T state until a new steady-state level of methylation has been achieved and prestimulus CheA activity has been restored despite the continued presence of stimulus (11, 63, 99, 205, 225).
In attempting to understand the mechanism of sensory motor regulation in chemotaxis, it is essential to know how steady-state swimming behavior is produced. All Che mutants that were selected as defective in chemotaxis exhibit aberrant steady-state swimming behavior: they are locked in an extreme, either constantly running or constantly tumbling (167). These behaviors may be understood from the known effects of Che mutations on phosphorylation of CheY, assuming that phospho-CheY binding to the flagellar switch, i.e., FliM, causes tumbly behavior (38, 168). Thus, strains deficient in CheY, CheA, CheW, all MCPs, or any combination of these components tend to run incessantly without ever tumbling (42, 114, 167). The CheY requirement follows from the postulated central role of phospho-CheY in generating tumbles (75, 269), the requirement for CheA follows from its role as a phosphodonor for CheY (75, 269), and the requirement for CheW follows from its role in activating CheA autophosphorylation (29, 62, 156). Tumbling can be restored in the absence of CheA and/or CheW by elevating the intracellular level of acetyl phosphate (45, 264), by overproducing an alternate histidine protein kinase with cross specificity for CheY (154), by overproducing CheY (40, 213), or by mutating CheZ (114, 265). In addition, there are tumble-inducing CheY mutants that function independently of CheA (33, 34), presumably by assuming an active conformational state in the absence of phosphorylation. The effect of a CheY deficiency on steady-state behavior can only be genetically counteracted by tumble-inducing mutations of flagellar motor components (47, 263).
It has been found that overproduction of CheA (42, 213), CheW (42, 114, 185, 213), or MCPs (114, 213) causes cells to become predominantly smooth swimming, the phenotype associated with deletion of any of these components. These seemingly paradoxical results have been taken as indications that there must be a run-inducing signal from the MCPs, perhaps a phospho-CheY phosphatase activity (38, 114, 185). The findings can be explained, however, in a straightforward way that does not require the formulation of a new mechanism for phospho-CheY regulation. Each situation may be considered separately as follows.
(i) CheA overproduction. Overproduction of CheA would be expected to yield a molar excess of CheA that is not associated with CheW and MCPs and which would compete for CheY binding to active CheA in ternary complexes. Since the rate of autophosphorylation of free CheA is very small compared with that of CheA in the complex, the overall rate of CheY phosphorylation will be inhibited and the cells will tend to smooth swim most of the time.
(ii) CheW overproduction. If CheW has independent binding sites for MCPs and CheA, a molar excess of CheW would be expected to bind independently to MCP and CheA and thereby block the trimeric interaction that is required to form a ternary complex (62, 156, 238).
(iii) MCP overproduction. The effect of MCP overproduction could stem from the low level of methyltransferase and the inability of this enzyme to maintain a large pool of MCP protein at a sufficiently high level of methylation to maintain active ternary complexes with CheW and CheA. Alternatively, overexpression of the MCPs may effectively titrate out CheW such that there would be insufficient CheW to form ternary complexes in a 2:2:2 stoichiometry (CheW would be in 1:2 complexes with the MCPs). CheW can form 1:1 complexes with monomeric signaling domain fragments (M. G. Surette and J. B. Stock, unpublished results).
Strains with mutations in cheZ exhibit an incessantly tumbly steady-state swimming behavior (168), as would be expected from the known enzymatic function of CheZ in accelerating the dephosphorylation of phospho-CheY (75, 120). As expected from its function in phospho-CheY dephosphorylation, CheZ antagonizes the tumbly phenotype associated with CheY overproduction (104). Double mutants that lack CheZ and CheW and/or all MCPs exhibit a random steady-state swimming behavior as long as CheA is present (114), as though the low rate of CheA autophosphorylation in the absence of MCPs and/or CheW is sufficient to maintain an intermediate level of CheY phosphorylation under conditions where CheY dephosphorylation is not stimulated by CheZ. As would be expected from its phosphatase activity, overproduction of CheZ causes cells to become predominantly smooth swimming (213).
In interpreting the literature on the behavioral consequences of CheZ mutations, it should be noted that the S. typhimurium cheZ221 allele, originally isolated in a mutant strain termed ST171 (12, 258), which has been the subject of intense physiological scrutiny (13, 39, 48, 59, 181, 202, 208, 209, 226), is really a missense allele of cheB that produces a defective CheB protein that retains some activity in vivo (106, 220).
CheR mutants exhibit a smooth swimming phenotype (208, 229), and CheB mutants have a tumbly phenotype (228, 277). Overproduction of these proteins causes the opposite effect of deficiency (213): excess CheR makes cells tumble and excess CheB makes them run. These effects are those that would be expected from the effect of MCP methylation on CheA autophosphorylation within ternary complexes of CheA, CheW, and MCPs (28, 156). Double mutants that are deficient in both CheR and CheB exhibit a random, essentially wild-type, steady-state swimming behavior that becomes biased toward tumbles or runs if cells are exposed to repellents or attractants (22, 190, 222, 223). This result may be understood in terms of the approximate equivalence of glutaminyl and glutamyl γ-methylesters (224). Mutants that lack CheR but retain CheB have MCPs that are essentially in the EEEE (all glutamate) form, mutants that lack CheB but retain CheR have MCPs that are essentially in the QEmQEm (all modified) form, and mutants that lack both CheR and CheB have MCPs in the QEQE (half modified) form, which is close to the level of MCP modification in wild-type cells.
These findings are all consistent with the view that the steady-state pattern of runs and tumbles is controlled by the steady-state level of phospho-CheY, in accord with the hypothesis that phospho-CheY binding to the FliM protein at the motor causes the cell to tumble. The flagellar motor is a large macromolecular assembly composed of a filament of flagellin attached to a basal structure surrounded by Mot proteins (126; see chapter 10). The Fli motor-switch proteins are peripherally associated with this structure (56, 57, 92). The entire motor assembly is in dynamic equilibrium between two states, a counterclockwise (CCW) run state and a clockwise (CW) tumbly state (107), as follows: run (CCW) ⇔ tumble (CW).
The run and tumble states of the motor are not in thermal equilibrium, considerable energy is constantly passing through the Che system, and this energy is used, in part, to determine the probabilities of run-to-tumble and tumble-to-run transitions. ATP is used to produce phospho-CheY, and in the absence of phospho-CheY, the run state is highly favored. According to this view, the probability of a motor being in the tumbly state, Pt, at any given time may be described in terms of the following equation: Pt = (P∼CheY)n/[(P∼CheY)n + KYn], where n is the number of phospho-CheY monomers that must bind to the motor to cause a tumble (i.e., induce a run-to-tumble transition), and KY is the dissociation constant for the phospho-CheY–FliM complex. As long as (P∼CheY) is maintained close to the threshold value defined by KY, the motor output should alternate between tumbles and runs.
This model predicts that, if phospho-CheY is maintained at a constant value, cells should exhibit a constant probability of a tumble at all times during a run. If this is true, then the duration of running intervals should be exponentially distributed. This was observed when the run intervals of individual E. coli cells were analyzed using a tracking microscope (18). There were a very large number of very short runs with a duration of <0.08 to 0.16 s, much fewer with a duration of 0.016 to 0.024 s, fewer with a duration of 0.024 to 0.032 s, etc., with the number in each successive interval dropping exponentially. This is what one would predict if the probability of a tumble were constant throughout a run. One can also address this question by continuously monitoring the direction of flagellar rotation of cells tethered through a single flagellum to a glass slide coated with antiflagellar antibodies and following the sense of rotation of the cell body as it attempts to turn the anchored flagellum. The results essentially confirm the findings obtained in the tracking experiments (23, 105). The probability of CCW-to-CW and CW-to-CCW transitions is random. Thus, under steady-state conditions, the level of phospho-CheY is probably maintained at a constant value and the steady-state tumbling frequency is determined by its concentration.
Results obtained in experiments where swimming behavior was measured as a function of intracellular CheY concentration indicate that n for CheY binding to the motor-switch is 5.5 ± 1.9 and that the threshold value for CheY is approximately 10 μM (105). In this study, CheY was expressed from a regulated promoter in an E. coli strain that was deleted for a portion of the Meche operon extending from CheA to CheZ. These findings are difficult to interpret because it was not possible to evaluate the level of CheY phosphorylation under the conditions used. A comparison of the data obtained in this study with the swimming behavior of wild-type E. coli suggests that the levels of CheY phosphorylation at intracellular concentrations of CheY near the putative threshold concentration of 10 μM were similar to the levels of phosphorylation of CheY in wild-type cells, where the level of CheY is also approximately 10 μM. It is clear that in wild-type cells CheY is not predominantly phosphorylated since conditions that increase rates of phosphorylation cause dramatic increases in tumbly behavior.
Cells migrate toward favorable environmental conditions by biasing the random walk of runs punctuated by tumbles (18). If a cell senses that it is running toward an attractant, it suppresses its tendency to tumble and thereby extends its run length. The resulting bias leads to a net migration in the desired direction. Thus, the output of the chemotaxis signal transduction system is essentially a go, no-go decision to continue running along a given course or to tumble and change direction.
How does a bacterium decide whether or not to change direction? This is tantamount to asking how a bacterium decides that it is moving in the right direction, e.g., toward higher concentrations of an attractant such as aspartate. One could envisage the following two possible sensing mechanisms.
(i) Spatial sensing. A swimming cell could constantly monitor the concentration of aspartate at its head and at its tail. When (Asp)head > (Asp)tail , the cell would keep going; when (Asp)head < (Asp)tail , the cell would change direction.
(ii) Temporal sensing. A swimming cell could measure the rate of change in aspartate concentration [d(Asp)/dt]; if d(Asp)/dt > 0, the cell would continue on course.
It has been argued that the length of a bacterium is too small and it moves too quickly to allow cells to use spatial sensing to detect the types of gradients that elicit chemotaxis responses (19). Moreover, bacteria respond to sudden changes in aspartate concentration under conditions where the concentration of aspartate is spatially homogeneous (127). From these results, it is generally assumed that chemotactic responses are mediated by a temporal, rather than a spatial, sensing mechanism.
Sensory-response mechanisms have often been considered in terms of two types of processes, excitation and adaptation (99, 205). This concept as it applies to chemotaxis is best understood within the context of a specific experimental paradigm: the measurement of swimming responses to stepwise additions or removal of stimuli (127). Cells swimming randomly in a homogeneous environment are suddenly transferred to a different homogeneous environment with an altered concentration of one defined component, the stimulus. They respond to this abrupt transition by either running, an attractant response, or tumbling, a repellent response. Despite the continued presence of the stimulus, the altered behavior generally persists for at most only a few minutes, after which time the cells return to their prestimulus pattern of intermittent runs and tumbles. Within the context of this type of experiment, the initial response to the stimulus has been termed excitation and the recovery to prestimulus behavior has been termed adaptation.
It is clear that under these conditions the rate-limiting step for recovery from attractant stimuli is the time required for increased levels of receptor methylation (99, 205, 224). During chemotaxis, cells respond and adapt to stimuli within a 1-s time frame. However, to obtain easily measurable recovery times, on the order of minutes, stepwise stimulation experiments have generally involved stimuli that are at least a 100-fold greater than those that cells actually encounter as they migrate up gradients of attractant chemicals. Slow and relatively insensitive aspects of the chemosensory mechanism that do not function to regulate responses to small stimuli can play a dominant role in responses to large stepwise changes. The net steady-state evolution of methanol in wild-type cells under optimal nutritive conditions approaches only about 30 molecules per cell per s (88, 122), and 10-fold-lower values have been routinely observed. During the few seconds that a bacterium swimming in an attractant gradient must decide whether to run or tumble, <1% of its ∼10,000 receptors undergo the addition or removal of a methyl group. Methylation is not required for these types of chemotaxis responses (222). Rather, methylation appears to be a mechanism that functions on a time scale of minutes to maintain the receptors in a delicate balance, poised to respond to any perturbations in their environment.
The rate of demethylation in response to large stepwise additions of repellent or removal of attractants is much faster than the rate of methylation in response to attractant addition. In fact, MCPs are demethylated faster than the cell adapts (64). Some other process besides MCP demethylation is the rate-limiting step in this recovery process. This conclusion has been substantiated by the observation that cheB mutants that completely lack the demethylating enzyme still exhibit almost the same kinetics of adaptation to repellent stimuli as do wild-type cells (223). A model was developed to explain this response, where it was assumed that a tumble-producing species, T, was initially bound to another component, M. It was proposed that repellent stimuli caused the release of T in an active, tumble-inducing form, T*. T* then decayed back to an inactive state, T, that could then reassociate with M and go through another round of activation. This model fits experimental results obtained with signaling components in vitro, where T corresponds to CheY, M corresponds to the ternary complex of CheA-CheW-MCP, and T* corresponds to phospho-CheY (see references 188 and 223). Thus, the demethylation-independent recovery from stepwise repellent stimuli may simply be a consequence of the mechanism of CheY phosphorylation and dephosphorylation coupled to the high affinity of CheY for the CheA-CheW-MCP complex.
Levels of regulation within the bacterial chemotaxis system may be characterized in terms of the time frame in which they operate. Ligand binding and some protein conformational transitions occur within milliseconds. Changes in protein phosphorylation, dephosphorylation, and demethylation occur on a time scale of ∼0.1 s. Beyond these processes are the much slower modulations associated with methylation, stable protein-protein interactions, irreversible posttranslational modifications, translation, and transcription. In E. coli and S. typhimurium, these types of regulatory mechanisms exert their effects in a time frame of minutes. Most research on bacterial chemotaxis has focused on more rapid events. During chemotaxis, cells are continuously exposed to fluctuating concentrations of attractants and repellents. Slow processes such as methylation and altered levels of gene expression provide adaptive mechanisms that maintain the system in an optimally sensitive state. Faster processes such as phosphorylation and dephosphorylation represent a somewhat different type of response coupling, where, depending on the context, a process can play either an excitatory or an adaptive role.
The signal transduction networks that E. coli and S. typhimurium use to control their motility are beginning to be understood at a molecular level. Receptor proteins at the cell surface continuously monitor the chemical composition of the environment and use this information to control the reversible phosphorylation of proteins within the cytoplasm. Phosphorylation of a response regulator protein, CheY, controls the sense of rotation of the flagellar motor. Changes in levels of methylation of a family of transmembrane proteins, the MCPs, provide a negative feedback mechanism that functions to maintain the level of phospho-CheY at an intermediate value under a wide variety of environmental conditions. Phosphorylation and methylation are controlled both by external signals from the receptors and by the overall metabolic state of the cell. A complex interplay between excitatory responses, counteracting adaptive mechanisms, and chance determines whether a bacterium continues on course or changes its direction of motion.
We thank the following for providing experimental results prior to publication and for their thoughtful comments: Howard Berg, Robert Bourret, Frederick Dahlquist, Michael Eisenbach, Daniel Koshland, Stanislas Leibler, Mikhail Levit, Robert Macnab, Austin Newton, Alexander Ninfa, Elizabeth Ninfa, John Parkinson, Thomas Silhavy, Melvin Simon, Richard Stewart, Ann Stock, and Ann West.
References
1. Adler, J., and W. Epstein. 1974. Phosphotransferase system enzymes as chemoreceptors for certain sugars in Escherichia coli chemotaxis. Proc. Natl. Acad. Sci. USA 71:2895–2899.
2. Adler, J., G. L. Hazelbauer, and M. M. Dahl. 1973. Chemotaxis toward sugars in Escherichia coli. J. Bacteriol. 115:824–847.
3. Adler, J., and B. Templeton. 1967. The effect of environmental conditions on the motility of Escherichia coli. J. Gen. Microbiol. 46:175–184
4. Aksamit, R. R., and D. E. Koshland, Jr. 1972. A ribose binding protein of Salmonella typhimurium. Biochem. Biophy. Res. Commun. 48:1348–1353.
5. 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.
6. Alley, M. R. K., J. R. Maddock, and L. Shapiro. 1992. Polar localization of a bacterial chemoreceptor. Genes Dev. 6:825–836.
7. Ames, P., J. Chen, C. Wolff, and J. S. Parkinson. 1988. Structure-function studies of bacterial chemosensors. Cold Spring Harbor Symp. Quant. Biol. 53:59–66.
8. Ames, P., and J. S. Parkinson. 1988. Transmembrane signaling by bacterial chemoreceptors: E. coli transducers with locked signal output. Cell 55:817–826.
9. Ames, P., and J. S. Parkinson. 1994. Constitutively signaling fragments of Tsr, the Escherichia coli serine chemoreceptor. J. Bacteriol. 176:6340–6348.
10. Armstrong, J. B. 1972. Chemotaxis and methionine metabolism in Escherichia coli. Can. J. Microbiol. 18:1695–1701.
11. Asakura, S., and H. Honda. 1984. Two-state model for bacterial chemoreceptor proteins. J. Mol. Biol. 176:349–367.
12. Aswad, D. W., and D. E. Koshland, Jr. 1975. Isolation, characterization, and complementation of Salmonella typhimurium chemotaxis mutants. J. Mol. Biol. 97:225–235.
13. Aswad, D. W., and D. E. Koshland, Jr. 1975. Evidence for an S-adenosylmethionine requirement in the chemotactic behavior of Salmonella typhimurium. J. Mol. Biol. 97:207–223.
14. Barnakov, A. N., K. H. Downing, and G. L. Hazelbauer. 1994. Studies of the structural organization of a bacterial chemoreceptor by electron microscopy. J. Struct. Biol. 112:117–124.
15. Baumgartner, J. W., C. Kim, R. E. Brissette, M. Inouye, C. Park, and G. L. Hazelbauer. 1994. Transmembrane signalling by a hybrid protein: communication from the domain of chemoreceptor Trg that recognizes sugar-binding proteins to the kinase/phosphatase domain of osmosensor EnvZ. J. Bacteriol. 176:1157–1163.
16. Bellsolell, L., J. Prieto, L. Serrano, and M. Coll. 1994. Magnesium binding to the bacterial chemotaxis protein CheY results in large conformational changes involving its functional surface. J. Mol. Biol. 238:489–495.
17. Benkovic, S. J., and E. J. Sampson. 1971. Structure-reactivity correlation for the hydrolysis of phosphoramidate monoanions. J. Am. Chem. Soc. 93:4009–4016.
18. Berg, H. C., and D. A. Brown. 1972. Chemotaxis in Escherichia coli analysed by three-dimensional tracking. Nature (London) 239:500–504.
19. Berg, H. C., and E. M. Purcell. 1977. Physics of chemoreception. Biophys. J. 20:193–219.
20. Biemann, H.-P., and D. E. Koshland, Jr. 1994. Aspartate receptors of Escherichia coli and Salmonella typhimurium bind ligand with negative and half-of-the-sites cooperativity. Biochemistry 33:629–634.
21. Blat, Y., and M. Eisenbach. 1994. Phosphorylation-dependent binding of the chemotaxis signal molecule CheY to its phosphatase, CheZ. Biochemistry 33:902–906.
22. Block, S. M., J. E. Segall, and H. C. Berg. 1982. Impulse responses in bacterial chemotaxis. Cell 31:215–226.
23. Block, S. M., J. E. Segall, and H. C. Berg. 1983. Adaptation kinetics in bacterial chemotaxis. J. Bacteriol. 154:312–323.
24. Bochner, B. R., and B. N. Ames. 1982. Complete analysis of cellular nucleotides by two-dimensional thin layer chromatography. J. Biol. Chem. 257:9759–9769.
25. Bollinger, J., C. Park, S. Harayama, and G. L. Hazelbauer. 1984. Structure of the Trg protein: homologies with and differences from other sensory transducers of Escherichia coli. Proc. Natl. Acad. Sci. USA 81:3287–3291.
26. Borczuk, A., A. Staub, and J. Stock. 1986. Demethylation of bacterial chemoreceptors is inhibited by attractant stimuli in the complete absence of the regulatory domain of the demethylating enzyme. Biochem. Biophys. Res. Commun. 141:918–923.
27. Borczuk, A., A. Stock, and J. Stock. 1987. S-Adenosylmethionine may not be essential for signal transduction during bacterial chemotaxis. J. Bacteriol. 169:3295–3300.
28. Borkovich, K. A., L. A. Alex, and M. I. Simon. 1992. Attenuation of sensory signaling by covalent modification. Proc. Natl. Acad. Sci. USA 89:6756–6760.
29. Borkovich, K. A., N. Kaplan, J. F. Hess, and M. I. Simon. 1989. Transmembrane signal transduction in bacterial chemotaxis involves ligand-dependent activation of phosphate group transfer. Proc. Natl. Acad. Sci. USA 86:1208–1212.
30. Borkovich, K. A., and M. I. Simon. 1990. The dynamics of protein phosphorylation in bacterial chemotaxis. Cell 63:1339–1348.
31. Bourret, R. B., K. A. Borkovich, and M. I. Simon. 1991. Signal transduction pathways involving protein phosphorylation in prokaryotes. Annu. Rev. Biochem. 60:401–441.
32. Bourret, R. B., J. Davagnino, and M. I. Simon. 1993. The carboxy-terminal portion of the CheA kinase mediates regulation of autophosphorylation by transducer and CheW. J. Bacteriol. 175:2097–2101.
33. Bourret, R. B., S. K. Drake, S. A. Chervitz, M. I. Simon, and J. J. Falke. 1993. Activation of the phosphosignaling protein CheY. Analysis of activated mutants by 19F NMR and protein engineering. J. Biol. Chem. 268:13089–13096.
34. Bourret, R. B., J. F. Hess, and M. J. Simon. 1990. Conserved aspartate residues and phosphorylation in signal transduction by the chemotaxis protein CheY. Proc. Natl. Acad. Sci. USA 87:41–45.
35. Boyd, A., K. Kendal, and M. I. Simon. 1983. Structure of the serine chemoreceptor in Escherichia coli. Nature (London) 301:623–626.
36. Boyd, A., A. Krikos, N. Mutoh, and M. Simon. 1983. A family of homologous genes encoding sensory transducers in E. coli. p. 551–562. In H. Sund and C. Veeger (ed.), Mobility and Recognition in Cell Biology. Degruyter, Berlin.
37. Brass, J. M., and M. D. Manson. 1984. Reconstitution of maltose chemotaxis in Escherichia coli by addition of maltose-binding protein to calcium-treated cells of maltose regulon mutants. J. Bacteriol. 157:881–890.
38. Bray, D., R. B. Bourret, and M. I. Simon. 1993. Computer simulation of the phosphorylation cascade controlling bacterial chemotaxis. Mol. Biol. Cell 4:469–482.
39. Clarke, S., and D. E. Koshland, Jr. 1979. Membrane receptors for aspartate and serine in bacterial chemotaxis. J. Biol. Chem. 254:9695–9702.
40. Clegg, D. O., and D. E. Koshland, Jr. 1984. The role of a signaling protein in bacterial sensing: behavioral effects of increased gene expression. Proc. Natl. Acad. Sci. USA 81:5056–5060.
41. Collins, L. A., S. M. Egan, and V. Stewart. 1992. Mutational analysis reveals functional similarity between NARX, a nitrate sensor in Escherichia coli K-12, and the methyl-accepting chemotaxis proteins. J. Bacteriol. 174:3667–3675.
42. Conley, M. P., A. J. Wolfe, D. F. Blair, and H. C. Berg. 1989. Both CheA and CheW are required for reconstitution of chemotactic signaling in Escherichia coli. J. Bacteriol. 171:5190–5193.
43. Dahl, M. K., and M. D. Manson. 1985. Interspecific reconstitution of maltose transport and chemotaxis in Escherichia coli with maltose-binding protein from various enteric bacteria. J. Bacteriol. 164:1057–1063
44. Dahlquist, F. W., P. Lovely, and D. E. Koshland, Jr. 1972. Quantitative analysis of bacterial migration in chemotaxis. Nature (London) New Biol. 236:120–123.
45. Dailey, F. E., and H. C. Berg. 1993. Change in direction of flagellar rotation in Escherichia coli mediated by acetate kinase. J. Bacteriol. 175:3236–3239.
46. DeFranco, A. L., and D. E. Koshland, Jr. 1981. Molecular cloning of chemotaxis genes and overproduction of gene products in the bacterial sensing system. J. Bacteriol. 147:390–400.
47. DeFranco, A. L., and D. E. Koshland, Jr. 1982. Construction and behavior of strains with mutations in two chemotaxis genes. J. Bacteriol. 150:1297–1301.
48. DeFranco, A. L., J. S. Parkinson, and D. E. Koshland, Jr. 1979. Functional homology of chemotaxis genes in Escherichia coli and Salmonella typhimurium. J. Bacteriol. 139:107–114.
49. DiSabato, G., and W. P. Jencks. 1961. Mechanism and catalysis of reactions of acyl phosphates. I. Nucleophilic reactions. J. Am. Chem. Soc. 83:4393–4405.
50. Drake, S. K., R. B. Bourret, L. A. Luck, M. I. Simon, and J. J. Falke. 1993. Activation of the phosphosignaling protein CheY. J. Biol. Chem. 268:13081–13088.
51. Dunten, P., and D. E. Koshland, Jr. 1991. Tuning the responsiveness of a sensory receptor via covalent modification. J. Biol. Chem. 266:1491–1496.
52. Eisenbach, M. 1991. Signal transduction in bacterial chemotaxis. Mod. Cell Biol. 10:137–208.
53. Engström, P., D. Nowlin, J. Bollinger, N. Magnuson, and G. L. Hazelbauer. 1983. Limited homology between Trg and the other transducer proteins of Escherichia coli. J. Bacteriol. 156:1268–1274.
54. Fahnestock, M., and D. E. Koshland, Jr. 1979. Control of the receptor for galactose taxis in Salmonella typhimurium. J. Bacteriol. 137:758–763.
55. Feng, J., M. R. Atkinson, W. McCleary, J. B. Stock, B. L. Wanner, and A. J. Ninfa. 1992. Role of phosphorylated metabolic intermediates in the regulation of glutamine synthetase synthesis in Escherichia coli. J. Bacteriol. 174:6061–6070.
56. Francis, N. R., V. M. Irikura, S. Yamaguchi, D. J. DeRosier, and R. M. Macnab. 1992. Localization of the Salmonella typhimurium flagellar switch protein FliG to the cytoplasmic M-ring face of the basal body. Proc. Natl. Acad. Sci. USA 89:6304–6308.
57. Francis, N. R., G. E. Sosinsky, D. Thomas, and D. J. DeRosier. 1994. Isolation, characterization and structure of bacterial flagellar motors containing the switch complex. J. Mol. Biol. 235:1261–1270.
58. Fukami, Y., and F. Lipmann. 1983. Reversal of Rous sarcoma-specific immunoglobulin phosphorylation on tyrosine (ADP as phosphate acceptor) catalyzed by the src gene kinase. Proc. Natl. Acad. Sci. USA 80:1872–1876.
59. Galloway, R. J., and B. L. Taylor. 1990. Histidine starvation and adenosine 5'-triphosphate depletion in chemotaxis of Salmonella typhimurium. J. Bacteriol. 144:1068–1075.
60. Gardina, P., C. Conway, M. Kossman, and M. Manson. 1992. Aspartate and maltose-binding protein interact with adjacent sites in the Tar chemotactic signal transducer of Escherichia coli. J. Bacteriol. 174:1528–1536.
61. Gegner, J. A., and F. W. Dahlquist. 1991. Signal transduction in bacteria: CheW forms a reversible complex with the protein kinase CheA. Proc. Natl. Acad. Sci. USA 88:750–754.
62. Gegner, J. A., D. R. Graham, A. F. Roth, and F. W. Dahlquist. 1992. Assembly of an MCP receptor, CheW, and kinase CheA complex in the bacterial chemotaxis signal transduction pathway. Cell 18:975–982.
63. Goldbeter, A., and D. E. Koshland, Jr. 1982. Simple molecular model for sensing and adaptation based on receptor modification with application to bacterial chemotaxis. J. Mol. Biol. 161:395–416.
64. Goy, M. F., M. S. Springer, and J. Adler. 1977. Sensory transduction in Escherichia coli: role of a protein methylation reaction in sensory adaptation. Proc. Natl. Acad. Sci. USA 74:4964–4968.
65. Hazelbauer, G. L. 1975. Maltose chemoreceptor of Escherichia coli. J. Bacteriol. 122:206–214.
66. Hazelbauer, G. L. 1992. Bacterial chemoreceptors. Curr. Opin. Struct. Biol. 2:505–510.
67. Hazelbauer, G. L., and J. Adler. 1971. Role of the galactose binding protein in chemotaxis of Escherichia coli toward galactose. Nature (London) New Biol. 230:101–104.
68. Hazelbauer, G. L., H. C. Berg, and P. Matsumura. 1993. Bacterial motility and signal transduction. Cell 73:15–22.
69. Hazelbauer, G. L., and S. Harayama. 1979. Mutants in transmission of chemotactic signals from two independent receptors of E. coli. Cell 16:617–625.
70. Hedblom, M. L., and J. Adler. 1980. Genetic and biochemical properties of Escherichia coli mutants with defects in serine chemotaxis. J. Bacteriol. 144:1048–1060.
71. Hedblom, M. L., and J. Adler. 1983. Chemotactic response of Escherichia coli to chemically synthesized amino acids. J. Bacteriol. 155:1463–1466.
72. Herschlag, D., and W. P. Jencks. 1990. The effects of Mg2+, hydrogen bonding, and steric factors on rate and equilibrium constants for phosphoryl transfer between carboxylate ions and pyridines. J. Am. Chem. Soc. 112:1942–1950.
73. Hess, J. F., R. B. Bourret, K. Oosawa, P. Matsumura, and M. I. Simon. 1988. Protein phosphorylation and bacterial chemotaxis. Cold Spring Harbor Symp. Quant. Biol. 53:41–48.
74. Hess, J. F., R. B. Bourret, and M. I. Simon. 1988. Histidine phosphorylation and phosphoryl group transfer in bacterial chemotaxis. Nature (London) 336:139–143.
75. Hess, J. F., K. Oosawa, N. Kaplan, and M. I. Simon. 1988. Phosphorylation of three proteins in the signaling pathway of bacterial chemotaxis. Cell 53:79–87.
76. Hess, J. F., K. Oosawa, P. Matsumura, and M. I. Simon. 1987. Protein phosphorylation is involved in bacterial chemotaxis. Proc. Natl. Acad. Sci. USA 84:7609–7613.
77. Higgins, C. F. 1992. ABC transporters: from microorganisms to man. Annu. Rev. Cell Biol. 8:67–113.
78. Hoch, J. A., and T. J. Silhavy (ed.). 1995. Two-Component Signal Transduction. American Society for Microbiology, Washington D.C.
79. Hubler, L., G. N. Gill, and P. J. Bertics. 1989. Reversibility of the epidermal growth factor receptor self-phosphorylation reaction. Evidence for formation of a high energy phosphotyrosine bond. J. Biol. Chem. 264:1558–1564.
80. Imae, Y., K. Oosawa, T. Mizuno, M. Kihara, and R. M. Macnab. 1987. Phenol: a complex chemoeffector in bacterial chemotaxis. J. Bacteriol. 169:371–379.
81. Ingolia, T. D., and D. E. Koshland, Jr. 1979. Response to a metal ion-citrate complex in bacterial sensing. J. Bacteriol. 140:798–804.
82. Jeffery, C. J., and D. E. Koshland, Jr. 1993. Three-dimensional structural model of the serine receptor ligand-binding domain. Protein Sci. 2:559–566.
83. Jeffery, C. J., and D. E. Koshland, Jr. 1994. A single hydrophobic to hydrophobic substitution in the transmembrane domain impairs aspartate receptor function. Biochemistry 33:3457–3463.
84. Kaplan, N., and M. I. Simon. 1988. Purification and characterization of the wild-type and mutant carboxy-terminal domains of the Escherichia coli Tar receptor. J. Bacteriol. 170:5134–5140.
85. Kehry, M. R., M. W. Bond, M. W. Hunkapiller, and F. W. Dahlquist. 1983. Enzymatic deamidation of methyl-accepting chemotaxis proteins in Escherichia coli catalyzed by the cheB gene product. Proc. Natl. Acad. Sci. USA 80:3599–3603.
86. Kehry, M. R., and F. W. Dahlquist. 1982. Adaptation in bacterial chemotaxis: CheB-dependent modification permits additional methylations of sensory transducer proteins. Cell 29:761–772.
87. Kehry, M. R., and F. W. Dahlquist. 1982. The methyl-accepting chemotaxis proteins of Escherichia coli. Identification of the multiple methylation sites on methyl-accepting chemotaxis protein. J. Biol. Chem. 257:10378–10386.
88. Kehry, M. R., T. G. Doak, and F. W. Dahlquist. 1984. Stimulus-induced changes in methylesterase activity during chemotaxis in Escherichia coli. J. Biol. Chem. 259:11828–11835.
89. Kehry, M. R., T. G. Doak, and F. W. Dahlquist. 1985. Aberrant regulation of methylesterase activity in cheD chemotaxis mutants of Escherichia coli. J. Bacteriol. 161:105–112.
90. Kehry, M. R., T. G. Doak, and F. W. Dahlquist. 1985. Sensory adaptation in bacterial chemotaxis: regulation of demethylation. J. Bacteriol. 163:983–990.
91. Kehry, M. R., P. Engstrom, F. W. Dahlquist, and G. L. Hazelbauer. 1983. Multiple covalent modifications of Trg, a sensory transducer of Escherichia coli. J. Biol. Chem. 258:5050–5055.
92. Khan, I. H., T. S. Reese, and S. Khan. 1992. The cytoplasmic component of the bacterial flagellar motor. Proc. Natl. Acad. Sci. USA 89:5956–5960.
93. Kim, S.-H. 1994. "Frozen" dynamic dimer model for transmembrane signaling in bacterial chemotaxis receptors. Protein Sci. 3:159–165.
94. Kofoid, E. C., and J. S. Parkinson. 1991. Tandem translation starts in the cheA locus of Escherichia coli. J. Bacteriol. 173:2116–2119.
95. Koiwai, O., and H. Hayashi. 1979. Studies on bacterial chemotaxis. IV. Interaction of maltose receptor with a membrane-bound chemotaxis component. J. Biochem. 86:27–34.
96. Kondoh, H., C. B. Ball, and J. Adler. 1979. Identification of a methyl-accepting chemotaxis protein for the ribose and galactose chemoreceptors of Escherichia coli. Proc. Natl. Acad. Sci. USA 76:260–264.
97. Kort, E. N., M. F. Goy, S. H. Larsen, and J. Adler. 1975. Methylation of a membrane protein involved in bacterial chemotaxis. Proc. Natl. Acad. Sci. USA 72:3939–3943.
98. Koshland, D. E., Jr. 1951. Effect of catalysts on the hydrolysis of acetyl phosphate. Nucleophilic displacement mechanisms in enzymatic reactions. J. Am. Chem. Soc. 74:2286–2292.
99. Koshland, D. E., Jr. 1981. Biochemistry of sensing and adaptation in a simple bacterial system. Annu. Rev. Biochem. 50:765–782.
100. Kossmann, M., C. Wolff, and M. D. Manson. 1988. Maltose chemoreceptor of Escherichia coli: interaction of maltose-binding protein and the Tar signal transducer. J. Bacteriol. 170:4516–4521.
101. Krikos, A., M. P. Conley, A. Boyd, H. C. Berg, and M. I. Simon. 1985. Chimeric chemosensory transducers of Escherichia coli. Proc. Natl. Acad. Sci. USA 82:1326–1330.
102. Krikos, A., N. Mutoh, A. Boyd, and M. I. Simon. 1983. Sensory transducers of E. coli are composed of discrete structural and functional domains. Cell 33:615–622.
103. Krueger, J. K., J. Stock, and C. E. Schutt. 1992. Evidence that the methylesterase of bacterial chemotaxis may be a serine hydrolase. Biochim. Biophys. Acta 1119:322–326.
104. Kuo, S. C., and D. E. Koshland, Jr. 1987. Roles of cheY and cheZ gene products in controlling flagellar rotation in bacterial chemotaxis of Escherichia coli. J. Bacteriol. 169:1307–1314.
105. Kuo, S. C., and D. E. Koshland, Jr. 1989. Multiple kinetic states for the flagellar motor switch. J. Bacteriol. 171:6279–6287.
106. Kutsukake, K., and T. Iino. 1985. Refined genetic analysis of the region II che mutants in Salmonella typhimurium. Mol. Gen. Genet. 199:406–409.
107. Larsen, S. H., R. W. Reader, E. N. Kort, W.-W. Tso, and J. Adler. 1974. Change in direction of flagellar rotation is the basis of the chemotactic response in Escherichia coli. Nature (London) 249:74–77.
108. Laszlo, D. J., B. L. Fandrich, A. Sivaram, B. Chance, and B. L. Taylor. 1984. Cytochrome o as a terminal oxidase and receptor for aerotaxis in Salmonella typhimurium. J. Bacteriol. 159:663–667.
109. Laszlo, D. J., and B. L. Taylor. 1981. Aerotaxis in Salmonella typhimurium: role of electron transport. J. Bacteriol. 145:990–1001.
110. Lee, L., T. Mizuno, and Y. Imae. 1988. Thermosensing properties of Escherichia coli tsr mutants defective in serine chemoreception. J. Bacteriol. 170:4769–4774.
111. Lengeler, J. 1975. Mutations affecting transport of the hexitols d-mannitol, d-glucitol, and galactitol in Escherichia coli K-12: isolation and mapping. J. Bacteriol. 124:26–38.
112. Lengeler, J. W., and A. P. Vogler. 1989. Molecular mechanisms of bacterial chemotaxis towards PTS-carbohydrates. FEMS Microbiol. Rev. 63:81–92.
113. Lin, L.-N., J. Li, J. F. Brandts, and R. M. Weis. 1994. The serine receptor of bacterial chemotaxis exhibits half-site saturation for serine binding. Biochemistry 33:6564–6570.
114. Liu, J., and J. S. Parkinson. 1989. Role of CheW protein in coupling membrane receptors to the intracellular signaling system of bacterial chemotaxis. Proc. Natl. Acad. Sci. USA 86:8703–8707.
115. Liu, J., and J. S. Parkinson. 1991. Genetic evidence for interaction between the CheW and Tsr proteins during chemoreceptor signaling by Escherichia coli. J. Bacteriol. 173:4941–4951.
116. Long, D. G., and R. M. Weis. 1992. Oligomerization of the cytoplasmic fragment from the aspartate receptor of Escherichia coli. Biochemistry 31:9904–9911.
117. Lowe, G., M. Meister, and H. C. Berg. 1987. Rapid rotation of flagellar bundles in swimming bacteria. Nature (London) 325:637–640.
118. Lowry, D. F., A. F. Roth, P. B. Rupert, F. W. Dahlquist, F. J. Moy, P. J. Domaille, and P. Matsumura. 1994. Signal transduction in chemotaxis: a propagating conformation change upon phosphorylation of CheY. J. Biol. Chem. 269:26358–26362.
119. Lukat, G. S., B. H. Lee, J. M. Mottonen, A. M. Stock, and J. B. Stock. 1991. Roles of the highly conserved aspartate and lysine residues in the response regulator of bacterial chemotaxis. J. Biol. Chem. 266:8348–8354.
120. Lukat, G. S., W. R. McCleary, A. M. Stock, and J. B. Stock. 1992. Phosphorylation of bacterial response regulator proteins by low molecular weight phospho-donors. Proc. Natl. Acad. Sci. USA 89:718–722.
121. Lukat, G. S., A. M. Stock, and J. B. Stock. 1990. Divalent metal ion binding to the CheY protein and its significance to phosphotransfer in bacterial chemotaxis. Biochemistry 29:5436–5442.
122. Lupas, A., and J. Stock. 1989. Phosphorylation of an N-terminal regulatory domain activates the CheB methylesterase in bacterial chemotaxis. J. Biol. Chem. 264:17337–17342.
123. Lupas, A., M. VanDyke, and J. Stock. 1991. Predicting coiled coils from protein sequences. Science 252:1162–1164.
124. Lynch, B. A., and D. E. Koshland, Jr. 1991. Disulfide cross-linking studies of the transmembrane regions of the aspartate sensory receptor of Escherichia coli. Proc. Natl. Acad. Sci. USA 88:10402–10406.
125. Lynch, B. A., and D. E. Koshland, Jr. 1992. Structural similarities between the aspartate receptor of bacterial chemotaxis and the trp repressor of E. coli. FEBS Lett. 307:3–9.
126. Macnab, R. M. 1992. Genetics and biogenesis of bacterial flagella. Annu. Rev. Genet. 26:129–156.
127. Macnab, R. M., and D. E. Koshland, Jr. 1972. The gradient-sensing mechanism in bacterial chemotaxis. Proc. Natl. Acad. Sci. USA 69:2509–2512.
128. Maddock, J. R., and L. Shapiro. 1993. Polar location of the chemoreceptor complex in the Escherichia coli cell. Science 259:1717–1723.
129. Magariyama, Y., S. Yamaguchi, and S.-I. Aizawa. 1990. Genetic and behavioral analysis of flagellar switch mutants of Salmonella typhimurium. J. Bacteriol. 172:4359–4369.
130. Manson, M. D., V. Blank, G. Brade, and C. F. Higgins. 1986. Peptide chemotaxis in E. coli involves the Tap signal transducer and the dipeptide permease. Nature (London) 321:253–256.
131. Manson, M. D., P. J. Boos, P. J. Bassford, Jr., and B. A. Rasmussen. 1985. Dependence of maltose transport and chemotaxis on the amount of maltose-binding protein. J. Biol. Chem. 260:9727–9733.
132. Manson, M. D., and M. Kossmann. 1986. Mutations in tar suppress defects in maltose chemotaxis caused by specific malE mutations. J. Bacteriol. 165:34–40.
133. Mao, B., M. R. Pear, J. A. McCammon, and F. A. Quiocho. 1982. Hinge bending in L-arabinose-binding protein: the Venus flytrap model. J. Biol. Chem. 257:1131–1133.
134. Matsumura, P., S. Roman, K. Volz, and D. McNally. 1990. Signalling complexes in bacterial chemotaxis, p. 135–154. In J. P. Armitage and J. M. Lackie (ed.), Biology of the Chemotactic Response, symposium 46. Cambridge University Press, Cambridge.
135. Matsumura, P., J. J. Rydel, R. Linzmeier, and D. Vacante. 1984. Overexpression and sequence of the Escherichia coli cheY gene and biochemical activities of the CheY protein. J. Bacteriol. 160:36–41.
136. McCleary, W. R., and J. B. Stock. 1994. Acetyl phosphate and the activation of two-component response regulators. J. Biol. Chem. 269:31567–31572.
137. McCleary, W. R., J. B. Stock, and A. J. Ninfa. 1993. Is acetyl phosphate a global signal in Escherichia coli? J. Bacteriol. 175:2793–2798.
138. Melton, T., P. E. Hartman, J. P. Stratis, T. L. Lee, and A. T. Davis. 1978. Chemotaxis of Salmonella typhimurium to amino acids and some sugars. J. Bacteriol. 133:708–716.
139. Mesibov, R., and J. Adler. 1972. Chemotaxis toward amino acids in Escherichia coli. J. Bacteriol. 112:315–326.
140. Milburn, M. V., G. G. Prive, D. L. Milligan, W. G. Scott, J. Yeh, J. Jancarik, D. E. Koshland, Jr., and S.-H. Kim. 1991. Three-dimensional structures of the ligand-binding domain of the bacterial aspartate receptor with and without a ligand. Science 254:1342–1347.
141. Milligan, D. L., and D. E. Koshland, Jr. 1988. Site-directed cross-linking. Establishment of the dimeric structure of the aspartate receptor of bacterial chemotaxis. J. Biol. Chem. 263:6268–6275.
142. Milligan, D. L., and D. E. Koshland, Jr. 1990. The amino terminus of the aspartate chemoreceptor is formylmethionine. J. Biol. Chem. 265:4455–4460.
143. Milligan, D. L., and D. E. Koshland, Jr. 1993. Purification and characterization of the periplasmic domain of the aspartate receptor. J. Biol. Chem. 268:19991–19997.
144. Mizuno, T., N. Mutoh, S. M. Panasenko, and Y. Imae. 1986. Acquisition of maltose chemotaxis in Salmonella typhimurium by the introduction of the Escherichia coli chemosensory transducer gene. J. Bacteriol. 165:890–895.
145. Morrison, T. B., and J. S. Parkinson. 1994. Liberation of an interaction domain from the phosphotransfer region of CheA, a signaling kinase of Escherichia coli. Proc. Natl. Acad. Sci. USA 91:5485–5489.
146. Mowbray, S. L., D. L. Foster, and D. E. Koshland, Jr. 1985. Proteolytic fragments identified with domains of the aspartate receptor. J. Biol. Chem. 260:11711–11718.
147. Mowbray, S. L., and D. E. Koshland, Jr. 1987. Additive independent responses in a single receptor: aspartate and maltose stimuli on the Tar protein. Cell 50:171–180.
148. Mowbray, S. L., and D. E. Koshland, Jr. 1990. Mutations in the aspartate receptor of Escherichia coli which affect aspartate binding. J. Biol. Chem. 265:15638–15643.
149. Moy, F. J., D. F. Lowry, P. Matsumura, F. W. Dahlquist, J. E. Krywko, and P. J. Domaille. 1994. Assignments, secondary structure, global fold, and dynamics of chemotaxis Y protein using three- and four-dimensional heteronuclear (13C,15N) NMR spectroscopy. Biochemistry 33:10731–10742.
150. Mutoh, N., K. Oosawa, and M. I. Simon. 1986. Characterization of Escherichia coli chemotaxis receptor mutants with null phenotypes. J. Bacteriol. 164:992–998.
151. Mutoh, N., and M. I. Simon. 1986. Nucleotide sequence corresponding to five chemotaxis genes in Escherichia coli. J. Bacteriol. 165:161–166.
152. Needham, J. V., T. Y. Chen, and J. J. Falke. 1993. Novel ion specificity of a carboxylate cluster Mg(II) binding site: strong charge selectivity and weak size selectivity. Biochemistry 32:3363–3367.
153. Ninfa, A. J., and R. L. Bennett. 1991. Identification of the site of autophosphorylation of the bacterial protein kinase/phosphatase NRII. J. Biol. Chem. 266:6888–6893.
154. Ninfa, A. J., E. G. Ninfa, A. N. Lupas, A. Stock, B. Magasanik, and J. Stock. 1988. Crosstalk between bacterial chemotaxis signal transduction proteins and the regulators of transcription of the Ntr regulon: evidence that nitrogen assimilation and chemotaxis are controlled by a common phosphotransfer mechanism. Proc. Natl. Acad. Sci. USA 85:5492–5496.
155. Ninfa, E. G., M. R. Atkinson, E. S. Kamberov, and A. J. Ninfa. 1993. Mechanism of autophosphorylation of Escherichia coli nitrogen regulator II (NRII or NtrB): trans-phosphorylation between subunits. J. Bacteriol. 175:7024–7032.
156. Ninfa, E. G., A. Stock, S. Mowbray, and J. Stock. 1991. Reconstitution of the bacterial chemotaxis signal transduction system from purified components. J. Biol. Chem. 266:9764–9770.
157. Nowlin, D. M., J. Bollinger, and G. L. Hazelbauer. 1987. Sites of covalent modification in Trg, a sensory transducer of Escherichia coli. J. Biol. Chem. 262:6039–6045.
158. Oosawa, K., F. Hess, and M. I. Simon. 1988. Mutants defective in bacterial chemotaxis show modified protein phosphorylation. Cell 53:89–96.
159. Oosawa, K., and Y. Imae. 1983. Glycerol and ethylene glycol: members of a new class of repellents in Escherichia coli chemotaxis. J. Bacteriol. 154:104–112.
160. Oosawa, K., N. Mutoh, and M. I. Simon. 1988. Cloning of the C-terminal cytoplasmic fragment of the Tar protein and effects on chemotaxis of Escherichia coli. J. Bacteriol. 170:2521–2526.
161. Oosawa, K., and M. Simon. 1986. Analysis of mutations in the transmembrane region of the aspartate chemoreceptor in Escherichia coli. Proc. Natl. Acad. Sci. USA 83:6930–6934.
162. Ordal, G. W., and J. Adler. 1974. Isolation and complementation of mutants in galactose taxis and transport. J. Bacteriol. 117:509–516.
163. Ordal, G. W., and J. Adler. 1974. Properties of mutants in galactose taxis and transport. J. Bacteriol. 117:517–526.
164. Pakula, A., and M. Simon. 1992. Pivots or pistons? Nature (London) 355:496–498.
165. Pakula, A. A., and M. I. Simon. 1992. Determination of transmembrane protein structure by disulfide cross-linking: the Escherichia coli Tar receptor. Proc. Natl. Acad. Sci. USA 89:4144–4148.
166. Pan, S. Q., T. Charles, S. Jin, Z.-L. Wu, and E. W. Nester. 1993. Preformed dimeric state of the sensor protein VirA is involved in plant-Agrobacterium signal transduction. Proc. Natl. Acad. Sci. USA 90:9939–9943.
167. Parkinson, J. S. 1977. Behavioral genetics in bacteria. Annu. Rev. Genet. 11:397–414.
168. Parkinson, J. S. 1988. Protein phosphorylation in bacterial chemotaxis. Cell 53:1–2.
169. Parkinson, J. S. 1993. Signal transduction schemes of bacteria. Cell 73:857–871.
170. Parkinson, J. S., and E. C. Kofoid. 1992. Communication modules in bacterial signaling proteins. Annu. Rev. Genet. 26:71–112.
171. Parkinson, J. S., and S. R. Parker. 1979. Interaction of the cheC and cheZ gene products is required for chemotactic behavior in Escherichia coli. Proc. Natl. Acad. Sci. USA 76:2390–2394.
172. Parkinson, J. S., S. R. Parker, P. B. Talbert, and S. E. Houts. 1983. Interactions between chemotaxis genes and flagellar genes in Escherichia coli. J. Bacteriol. 155:265–274.
173. Postma, P. W., and J. W. Lengeler. 1985. Phosphoenolpyruvate:carbohydrate phosphotransferase system of bacteria. Microbiol. Rev. 49:232–269.
174. Postma, P. W., J. W. Lengeler, and G. R. Jacobson. 1993. Phosphoenolpyruvate:carbohydrate phosphotransferase system of bacteria. Microbiol. Rev. 57:534–594.
175. Pruss, B. M., and A. J. Wolfe. 1994. Regulation of acetyl phosphate synthesis and degradation, and the control of flagellar expression in Escherichia coli. Mol. Microbiol. 12:973–984.
176. Rice, M. S., and F. W. Dahlquist. 1991. Sites of deamidation and methylation in Tsr, a bacterial chemotaxis sensory transducer. J. Biol. Chem. 266:9746–9753.
177. Richarme, G. 1982. Interaction of the maltose-binding protein with membrane vesicles of Escherichia coli. J. Bacteriol. 149:662–667.
178. Roberts, D. L., D. W. Bennett, and S. A. Forst. 1993. Identification of the site of phosphorylation on the osmosenser, EnvZ, of Escherichia coli. J. Biol. Chem. 269:8728–8733.
179. Rollins, C., and F. W. Dahlquist. 1981. The methyl-accepting chemotaxis proteins of E. coli: a repellent-stimulated, covalent modification, distinct from methylation. Cell 25:333–340.
180. Roman, S. J., M. Meyers, K. Volz, and P. Matsumura. 1992. A chemotactic signaling surface on CheY defined by suppressors of flagellar switch mutations. J. Bacteriol. 174:6247–6255.
181. Rubik, B. A., and D. E. Koshland, Jr. 1978. Potentiation, desensitization, and inversion of response in bacterial sensing of chemical stimuli. Proc. Natl. Acad. Sci. USA 75:2820–2824.
182. Russo, A. F., and D. E. Koshland, Jr. 1983. Separation of signal transduction and adaptation functions of the aspartate receptor in bacterial sensing. Science 220:1016–1020.
183. Sack, J. S., M. A. Saper, and F. A. Quiocho. 1989. Periplasmic binding protein structure and function: refined X-ray structures of the leucine/isoleucine/valine binding protein and its complex with leucine. J. Mol. Biol. 206:171–191.
184. Sanders, D. A., B. L. Gillece-Castro, A. M. Stock, A. L. Burlingame, and D. E. Koshland, Jr. 1989. Identification of the site of phosphorylation of the chemotaxis response regulator protein, CheY. J. Biol. Chem. 264:21770–21778.
185. Sanders, D. A., B. Mendez, and D. E. Koshland, Jr. 1989. Role of the CheW protein in bacterial chemotaxis: overexpression is equivalent to absence. J. Bacteriol. 171:6271–6278.
186. Saraste, M., P. R. Sibbald, and A. Wittinghofer. 1990. The P-loop—a common motif in ATP- and GTP-binding proteins. Trends Biochem. Sci. 15:430–434.
187. Schulz, G. E. 1992. Binding of nucleotides by proteins. Curr. Opin. Struct. Biol. 2:61–67.
188. Schuster, S. C., R. V. Swanson, L. A. Alex, R. B. Bourret, and M. I. Simon. 1993. Assembly and function of a quaternary signal transduction complex monitored by surface plasmon resonance. Nature (London) 365:343–347.
189. Scott, W. G., D. L. Milligan, M. V. Milburn, G. G. Prive, J. Yeh, D. E. Koshland, Jr., and S.-H. Kim. 1993. Refined structures of the ligand-binding domain of the aspartate receptor from Salmonella typhimurium. J. Mol. Biol. 232:555–573.
190. Segall, J. E., M. D. Manson, and H. C. Berg. 1982. Signal processing times in bacterial chemotaxis. Nature (London) 296:855–857.
191. Shapiro, M. J., D. Panomitros, and D. E. Koshland, Jr. 1995. Interactions between the methylation sites on the Escherichia coli aspartate receptor mediated by the methyltransferase. J. Biol. Chem. 270:751–755.
192. Sherris, D., and J. S. Parkinson. 1981. Posttranslational processing of methyl-accepting chemotaxis proteins in Escherichia coli. Proc. Natl. Acad. Sci. USA 78:6051–6055.
193. Shioi, J., C. V. Dang, and B. L. Taylor. 1987. Oxygen as attractant and repellent in bacterial chemotaxis. J. Bacteriol. 169:3118–3123.
194. Shioi, J., and B. L. Taylor. 1984. Oxygen taxis and proton motive force in Salmonella typhimurium. J. Biol. Chem. 259:10983–10988.
195. Shizuta, Y., J. A. Beavo, P. J. Bechtel, F. Hofmann, and E. G. Krebs. 1975. Reversibility of adenosine 3':5'-monophosphate-dependent protein kinase reactions. J. Biol. Chem. 250:6891–6896.
196. Silverman, M., and M. Simon. 1977. Identification of polypeptides necessary for chemotaxis in Escherichia coli. J. Bacteriol. 130:1317–1325.
197. Simms, S., and K. Subbaramaiah. 1991. The kinetic mechanism of S-adenosyl-l-methionine:glutamylmethyltransferase from Salmonella typhimurium. J. Biol. Chem. 266:12741–12746.
198. Simms, S. A., E. W. Cornman, J. Mottonen, and J. Stock. 1987. Active site of the enzyme which demethylates receptors during bacterial chemotaxis. J. Biol. Chem. 262:29–31.
199. Simms, S. A., M. G. Keane, and J. Stock. 1985. Multiple forms of the CheB methylesterase in bacterial chemosensing. J. Biol. Chem. 260:10161–10168.
200. Simms, S. A., A. M. Stock, and J. B. Stock. 1987. Purification and characterization of the S-adenosylmethionine:glutamyl methyltransferase that modifies membrane chemoreceptor proteins in bacteria. J. Biol. Chem. 262:8537–8543.
201. Smith, R. A., and J. S. Parkinson. 1980. Overlapping genes at the cheA locus of Escherichia coli. Proc. Natl. Acad. Sci. USA 77:5370–5374.
202. Snyder, M. A., and D. E. Koshland, Jr. 1981. Identification of the esterase peptide and its interaction with the cheZ peptide in bacterial sensing. Biochimie 63:113–117.
203. Sockett, H., S. Yamaguchi, M. Kihara, V. M. Irikura, and R. M. Macnab. 1992. Molecular analysis of the flagellar switch protein FliM of Salmonella typhimurium. J. Bacteriol. 174:793–806.
204. Springer, M. S., M. F. Goy, and J. Adler. 1977. Sensory transduction in Escherichia coli: two complementary pathways of information processing that involve methylated proteins. Proc. Natl. Acad. Sci. USA 74:3312–3316.
205. Springer, M. S., M. F. Goy, and J. Adler. 1979. Protein methylation in behavioral control mechanisms and in signal transduction. Nature (London) 280:279–284.
206. Springer, M. S., and B. Zanolari. 1984. Sensory transduction in Escherichia coli: regulation of the demethylation rate by the CheA protein. Proc. Natl. Acad. Sci. USA 81:5061–5065.
207. Springer, M. S., B. Zanolari, and P. A. Pierzchala. 1982. Ordered methylation of the methyl-accepting chemotaxis proteins of Escherichia coli. J. Biol. Chem. 257:6861–6866.
208. Springer, W. R., and D. E. Koshland, Jr. 1977. Identification of a protein methyltransferase as the cheR gene product in the bacterial sensing system. Proc. Natl. Acad. Sci. USA 74:533–537.
209. Spudich, J. L., and D. E. Koshland, Jr. 1975. Quantitation of the sensory response in bacterial chemotaxis. Proc. Natl. Acad. Sci. USA 72:710–713.
210. Stewart, R. C. 1993. Activating and inhibitor mutations in the regulatory domain of CheB, the methylesterase in bacterial chemotaxis. J. Biol. Chem. 268:1921–1930.
211. Stewart, R. C., and F. W. Dahlquist. 1988. N-terminal half of CheB is involved in methylesterase response to negative chemotactic stimuli in Escherichia coli. J. Bacteriol. 170:5728–5738.
212. Stewart, R. C., A. F. Roth, and F. W. Dahlquist. 1990. Mutations that affect control of the methylesterase activity of CheB, a component of the chemotaxis adaptation system in Escherichia coli. J. Bacteriol. 172:3388–3399.
213. Stewart, R. C., C. B. Russell, A. F. Roth, and F. W. Dahlquist. 1988. Interaction of CheB with chemotaxis signal transduction components in Escherichia coli: modulation of the methylesterase activity and effects on cell swimming behavior. Cold Spring Harbor Symp. Quant. Biol. 53:27–40.
214. Stock, A., T. Chen, D. Welsh, and J. Stock. 1988. CheA protein, a central regulator of bacterial chemotaxis, belongs to a family of proteins that control gene expression in response to changing environmental conditions. Proc. Natl. Acad. Sci. USA 85:1403–1407.
215. Stock, A., D. E. Koshland, Jr., and J. Stock. 1985. Homologies between the Salmonella typhimurium CheY protein and proteins involved in the regulation of chemotaxis, membrane protein synthesis, and sporulation. Proc. Natl. Acad. Sci. USA 82:7989–7993.
216. Stock, A., J. Mottonen, T. Chen, and J. Stock. 1987. Identification of a possible nucleotide binding site in CheW, a protein required for sensory transduction in bacterial chemotaxis. J. Biol. Chem. 262:535–537.
217. Stock, A., E. Schaeffer, D. E. Koshland, Jr., and J. Stock. 1987. A second type of protein methylation reaction in bacterial chemotaxis. J. Biol. Chem. 262:8011–8014.
218. Stock, A. M., E. Martinez-Hackert, B. F. Rasmussen, A. H. West, J. B. Stock, D. Ringe, and G. A. Petsko. 1993. Structure of the Mg2+-bound form of CheY and mechanism of phosphoryl transfer in bacterial chemotaxis. Biochemistry 32:13375–13380.
219. Stock, A. M., J. M. Mottonen, J. B. Stock, and C. E. Schutt. 1989. Three-dimensional structure of CheY, the response regulator of bacterial chemotaxis. Nature (London) 337:745–749.
220. Stock, A. M., and J. B. Stock. 1987. Purification and characterization of the CheZ protein of bacterial chemotaxis. J. Bacteriol. 169:3301–3311.
221. Stock, A. M., D. C. Wylie, J. M. Mottonen, A. M. Lupas, E. G. Ninfa, A. J. Ninfa, C. E. Schutt, and J. B. Stock. 1988. Phopho-proteins involved in bacterial signal transduction. Cold Spring Harbor Symp. Quant. Biol. 53:49–57.
222. Stock, J., A. Borczuk, F. Chiou, and J. E. B. Burchenal. 1985. Compensatory mutations in receptor function: a reevaluation of the role of methylation in bacterial chemotaxis. Proc. Natl. Acad. Sci. USA 82:8364–8368.
223. Stock, J., G. Kersulis, and D. E. Koshland, Jr. 1985. Neither methylating nor demethylating enzymes are required for bacterial chemotaxis. Cell 42:683–690.
224. Stock, J., and A. Stock. 1987. What is the role of receptor methylation in bacterial chemotaxis? Trends Biochem. Sci. 12:371–375.
225. Stock, J. B. 1994. Adaptive responses in bacterial chemotaxis, p. 3–24. In D. R. Sibley and M. D. Houslay (ed.), Regulation of Cellular Signal Transduction Pathways by Desensitization and Amplification. Wiley, New York.
226. Stock, J. B., and D. E. Koshland, Jr. 1978. A protein methylesterase involved in bacterial sensing. Proc. Natl. Acad. Sci. USA 75:3659–3663.
227. Stock, J. B., and D. E. Koshland, Jr. 1981. Changing reactivity of receptor carboxyl groups during bacterial sensing. J. Biol. Chem. 256:10826–10833.
228. Stock, J. B., G. S. Lukat, and A. M. Stock. 1991. Bacterial chemotaxis and the molecular logic of intracellular signal transduction networks. Annu. Rev. Biophys. Biophys. Chem. 20:109–136.
229. Stock, J. B., A. M. Maderis, and D. E. Koshland, Jr. 1981. Bacterial chemotaxis in the absence of receptor carboxylmethylation. Cell 27:37–44.
230. Stock, J. B., A. J. Ninfa, and A. M. Stock. 1989. Protein phosphorylation and regulation of adaptive responses in bacteria. Microbiol. Rev. 53:450–490.
231. Stock, J. B., A. M. Stock, and J. M. Mottonen. 1990. Signal transduction in bacteria. Nature (London) 344:395–400.
232. Stock, J. B., M. G. Surette, M. Levit, and P. Park. 1995. Two-component signal transduction systems: structure-function relationships and mechanisms of catalysis, p. 25–51. In J. A. Hoch and T. J. Silhavy (ed.), Two-Component Signal Transduction. American Society for Microbiology, Washington, D.C.
233. Stoddard, B. L., J. D. Bui, and D. E. Koshland, Jr. 1992. Structure and dynamics of transmembrane signaling by the Escherichia coli aspartate receptor. Biochemistry 31:11978–11983.
234. Stoddard, B. L., and D. E. Koshland, Jr. 1992. Prediction of the structure of a receptor-protein complex using a binary docking method. Nature (London) 358:774–775.
235. Stoddard, B. L., and D. E. Koshland, Jr. 1993. Molecular recognition analyzed by docking simulations: the aspartate receptor and isocitrate dehydrogenase from Escherichia coli. Proc. Natl. Acad. Sci. USA 90:1146–1153.
236. Strange, P. G. 1976. Receptor interactions in a signalling system: competition between ribose receptor and galactose receptor in the chemotaxis response. Proc. Natl. Acd. Sci. USA 73:762–766.
237. Subbaramaiah, K., H. Charles, and S. Simms. 1991. Probing the role of cysteine residues in the CheR methyltransferase. J. Biol. Chem. 266:19023–19027.
238. Swanson, R. V., R. B. Bourret, and M. I. Simon. 1993. Intermolecular complementation of the kinase activity of CheA. Mol. Microbiol. 8:435–441.
239. Swanson, R. V., S. C. Schuster, and M. I. Simon. 1993. Expression of CheA fragments which define domains encoding kinase, phosphotransfer, and CheY binding activities. Biochemistry 32:7623–7629.
240. Tawa, P., and R. Stewart. 1994. Mutational activation of CheA, the protein kinase in the chemotaxis system of Escherichia coli. J. Bacteriol. 176:4210–4218.
241. Tawa, P., and R. C. Stewart. 1994. Kinetics of CheA autophosphorylation and dephosphorylation reactions. Biochemistry 33:7917–7924.
242. Taylor, B. L. 1983. How do bacteria find the optimal concentration of oxygen? Trends Biochem. Sci. 8:438–441.
243. Taylor, B. L. 1983. Role of proton motive force in bacterial sensory transduction. Annu. Rev. Microbiol. 37:551–573.
244. Taylor, B. L., and J. W. Lengeler. 1990. Transductive coupling by methylated transducing proteins and permeases of the phosphotransfer system in bacterial chemotaxis, p. 69–90. In Membrane Transport and Information Storage. Alan R. Liss, Inc., New York.
245. Taylor, S. S., D. R. Knighton, J. Zheng, L. F. Ten Eyck, and J. M. Sowadski. 1992. Structural framework for the protein kinase family. Annu. Rev. Cell Biol. 8:429–462.
246. Terwilliger, T. C., E. Bogonez, E. A. Wang, and D. E. Koshland, Jr. 1983. Sites of methyl esterification on the aspartate receptor involved in bacterial chemotaxis. J. Biol. Chem. 258:9608–9611.
247. Terwilliger, T. C., and D. E. Koshland, Jr. 1984. Sites of methyl esterification and deamidation on the aspartate receptor involved in chemotaxis. J. Biol. Chem. 259:7719–7725.
248. Terwilliger, T. C., J. Y. Wang, and D. E. Koshland, Jr. 1986. Kinetics of receptor modification. The multiply methylated aspartate receptors involved in bacterial chemotaxis. J. Biol. Chem. 261:10814–10820.
249. Terwilliger, T. C., J. Y. Wang, and D. E. Koshland, Jr. 1986. Surface structure recognized for covalent modification of the aspartate receptor in chemotaxis. Proc. Natl. Acad. Sci. USA 83:6707–6710.
250. Tisa, L. S., and J. Adler. 1992. Calcium ions are involved in Escherichia coli chemotaxis. Proc. Natl. Acad. Sci. USA 89:11804–11808.
251. Toews, M. L., M. F. Goy, M. S. Springer, and J. Adler. 1979. Attractants and repellents control demethylation of methylated chemotaxis proteins in Escherichia coli. Proc. Natl. Acad. Sci. USA 76:5544–5548.
252. Tso, W.-W., and J. Adler. 1974. Negative chemotaxis in Escherichia coli. J. Bacteriol. 118:560–576.
253. Utsumi, R., R. E. Brissette, A. Rampersaud, S. A. Forst, K. Oosawa, and M. Inouye. 1989. Activation of bacterial porin gene expression by a chimeric signal transducer in response to aspartate. Science 245:1246–1249.
254. Volz, K. 1993. Structural conservation in the CheY superfamily. Biochemistry 32:11741–11753.
255. Volz, K., J. Beman, and P. Matsumura. 1986. Crystallization and preliminary characterization of CheY, a chemotaxis control protein from Escherichia coli. J. Biol. Chem. 261:4723–4725.
256. Volz, K., and P. Matsumura. 1991. Crystal structure of Escherichia coli CheY refined at 1.7-Å resolution. J. Biol. Chem. 266:15511–15519.
257. Wang, E. A., and D. E. J. Koshland, Jr. 1980. Receptor structure in the bacterial sensing system. Proc. Natl. Acad. Sci. USA 77:7157–7161.
258. Warrick, H. M., B. L. Taylor, and D. E. Koshland, Jr. 1977. Chemotactic mechanism of Salmonella typhimurium: preliminary mapping and characterization of mutants. J. Bacteriol. 130:223–231.
259. Weigel, N., M. A. Kukuruzinska, A. Nakazawa, E. B. Waygood, and S. Roseman. 1982. Sugar transport by the bacterial phosphotransferase system. J. Biol. Chem. 257:14477–14491.
260. Welch, M., K. Oosawa, S.-I. Aizawa, and M. Eisenbach. 1993. Phosphorylation-dependent binding of a signal molecule to the flagellar switch of bacteria. Proc. Natl. Acad. Sci. USA 90:8787–8793.
261. Welch, M., K. Oosawa, S.-I. Aizawa, and M. Eisenbach. 1994. Effects of phosphorylation, Mg2+, and conformation of the chemotaxis protein CheY on its binding to the flagellar switch protein FliM. Biochemistry 33:10470–10476.
261a. West, A. H., E, Martinez-Hackert, and A. M. Stock. 1995. Crystal structure of the catalytic domain of the chemotaxis methylesterase, CheB. J. Mol. Biol. 250:276–290.
262. Willis, R. C., and C. E. Furlong. 1974. Purification and properties of a ribose-binding protein from Escherichia coli. J. Biol. Chem. 249:6926–6929.
263. Wolfe, A. J., and H. C. Berg. 1989. Migration of bacteria in semisolid agar. Proc. Natl. Acad. Sci. USA 86:6973–6977.
264. Wolfe, A. J., M. P. Conley, and H. C. Berg. 1988. Acetyladenylate plays a role in controlling the direction of flagellar rotation. Proc. Natl. Acad. Sci. USA 85:6711–6715.
265. Wolfe, A. J., M. P. Conley, T. J. Kramer, and H. C. Berg. 1987. Reconstitution of signaling in bacterial chemotaxis. J. Bacteriol. 169:1878–1885.
266. Wolfe, A. J., B. P. McNamara, and R. Stewart. 1994. The short form of CheA couples chemoreception to CheA phosphorylation. J. Bacteriol. 176:4483–4491.
267. Wolfe, A. J., and R. C. Stewart. 1993. The short form of the CheA protein restores kinase activity and chemotactic ability to kinase-deficient mutants. Proc. Natl. Acad. Sci. USA 90:1518–1522.
268. Wolff, C., and J. S. Parkinson. 1988. Aspartate taxis mutants of the Escherichia coli Tar chemoreceptor. J. Bacteriol. 170:4509–4515.
269. Wylie, D., A. Stock, C.-Y. Wong, and J. Stock. 1988. Sensory transduction in bacterial chemotaxis involves phosphotransfer between Che proteins. Biochem. Biophys. Res. Commun. 151:891–896.
270. Yaghmai, R., and G. L. Hazelbauer. 1992. Ligand occupancy mimicked by single residue substitutions in a receptor: transmembrane signaling induced by mutation. Proc. Natl. Acad. Sci. USA 89:7890–7894.
271. Yamaguchi, S., S.-I. Aizawa, M. Kihara, M. Isomura, C. J. Jones, and R. M. Macnab. 1986. Genetic evidence for a switching and energy-transducing complex in the flagellar motor of Salmonella typhimurium. J. Bacteriol. 168:1172–1179.
272. Yamamoto, K., and Y. Imae. 1993. Cloning and characterization of the Salmonella typhimurium-specific chemoreceptor Tcp for taxis to citrate and from phenol. Proc. Natl. Acad. Sci. USA 90:217–221.
273. Yang, Y., and M. Inouye. 1992. Intermolecular complementation between two defective mutant signal-transducing receptors. Proc. Natl. Acad. Sci. USA 88:11057–11061.
274. Yao, V. J., and J. L. Spudich. 1992. Primary structure of an archaebacterial transducer, a methyl-accepting protein associated with sensory rhodopsin I. Proc. Natl. Acad. Sci. USA 89:11915–11919.
275. Yeh, J. I., H. P. Biemann, J. Pandit, D. E. Koshland, Jr., and S.-H. Kim. 1993. The three-dimensional structure of the ligand-binding domain of a wild-type bacterial chemotaxis receptor. J. Biol. Chem. 268:9787–9792.
276. Yonekawa, H., and H. Hayashi. 1986. Desensitization by covalent modification of the chemoreceptor of Escherichia coli. FEBS Lett. 198:21–24.
277. Yonekawa, H., H. Hayashi, and J. S. Parkinson. 1983. Requirement of the cheB function for sensory adaptation in Escherichia coli. J. Bacteriol. 156:1228–1235.
278. Zukin, R. S., P. G. Strange, L. R. Heavey, and D. E. Koshland, Jr. 1977. Properties of the galactose binding protein of Salmonella typhimurium and Escherichia coli. Biochemistry 16:381–386.
Chapter73
