The Conserved Cytoplasmic Module of the Transmembrane Chemoreceptor Mcpc Mediates Carbohydrate Chemotaxis in Bacillus Subtilis

Molecular Microbiology (2003) 47(5), 1353–1366

The conserved cytoplasmic module of the transmembrane chemoreceptor McpC mediates carbohydrate chemotaxis in Bacillus subtilis

Christopher J. Kristich, George D. Glekas and change upon ligand binding (Falke and Hazelbauer, George W. Ordal* 2001). The receptors control a two-component system Department of Biochemistry, Colleges of Medicine and consisting of the histidine kinase CheA and its cognate Liberal Arts and Sciences, University of Illinois, Urbana, response regulator CheY. For a given chemotactic organ- IL 61801, USA. ism, the repertoire and specificities of its receptors determine the chemoeffectors to which that organism responds. Summary Comparative sequence analysis of over 120 transduc- Escherichia coli cells use two distinct sensory cir- ers revealed that they comprise a superfamily character- cuits during chemotaxis towards carbohydrates. One ized by a modular architecture (Zhulin, 2001). In circuit requires the phosphoenolpyruvate-dependent particular, the receptors possess two distinct structural phosphotransferase system (PTS) and is independent and functional modules: (i) a highly conserved C-terminal of any specific chemoreceptor, whereas the other multidomain module; and (ii) a variable N-terminal module uses a chemoreceptor-dependent sensory mecha- responsible for diversity within the superfamily (Fig. 1). nism analogous to that used during chemotaxis The N-terminal module is responsible for sensing environ- towards amino acids. Work on the carbohydrate mental stimuli; the C-terminal module transmits signals to chemotaxis sensory circuit of Bacillus subtilis the downstream chemotactic machinery and mediates reported in this article indicates that the B. subtilis adaptation to the imposed stimulus. Because of the mod- circuit is different from either of those used by E. coli. ular nature of the chemoreceptors, chimeric receptors can Our chemotactic analysis of B. subtilis strains be readily constructed by exchanging N-terminal modules expressing various chimeric chemoreceptors indi- from different receptors (Krikos et al., 1985; Feng et al., cates that the cytoplasmic, C-terminal module of the 1997; Weerasuriya et al., 1998). Such chimeras are func- chemoreceptor McpC acts as a sensory-input element tional and respond to environmental signals according to during carbohydrate chemotaxis. Our results also the specificity of the parental receptor from which the N- indicate that PTS-mediated carbohydrate transport, terminal module is derived. but not carbohydrate metabolism, is required for pro- The chemoreceptor C-terminal module can be divided duction of a chemotactic signal. We propose a model into three domains with distinct functions (Fig. 1) (Zhulin, in which PTS-transport-induced chemotactic signals 2001). The HAMP domain mediates the transmission of are transmitted to the C-terminal module of McpC for ligand-induced conformational signals between the N- control of chemotaxis towards PTS carbohydrates. terminal and C-terminal modules. The methylation domain (MD) contains the sites of methylesterification that are necessary for adaptation. Modification of these sites is Introduction catalysed by the CheR methyltransferase and the CheB Motile prokaryotes regulate chemotactic responses to methylesterase. Finally, the highly conserved domain environmental stimuli using a sensory circuit based on a (HCD, or signalling domain) is coupled by CheW to the two-component signal transduction system (Stock and chemotactic histidine kinase CheA. The HCD is thought Surette, 1996; Armitage, 1999). Chemotactic signalling is to be the location where conformational signals of the initiated by receptors (methyl-accepting chemotaxis pro- receptor are translated into modulation of CheA kinase teins or transducers) that undergo a conformational autophosphorylation activity. CheA activity controls phosphoryl flux to the response regulators CheY and CheB (Hess et al., 1988). Phospho- Blackwell Science, LtdOxford, UKMMIMolecular Microbiology 0950-382X Blackwell Publishing Ltd, 200347Original ArticleC. J. Kristich, G. D. Glekas and G. W. OrdalMcpC detects intracellular chemotactic signals rylated CheY interacts with a flagellar switch complex to Accepted 14 November, 2002. *For correspondence. E-mail g-ordal @uiuc.edu; Tel. (+1) 217 333 9098 or (+1) 217 333 0268; Fax influence the rotational state of the flagellar motor (Barak (+1) 217 333 8868. and Eisenbach, 1992; Welch et al., 1994) and, hence, the

1354 C. J. Kristich, G. D. Glekas and G. W. Ordal carbohydrate-specific EII’s in the membrane. The EII component phosphorylates the incoming carbohydrate during translocation. The rate limiting step in the phosphorelay is the PEP-dependent autophosphorylation of EI (Weigel et al., 1982; Chauvin et al., 1994); thus, enhanced carbohydrate transport is thought to generate an increased level of unphosphorylated EI, which is known to inhibit receptor-bound CheA autophosphorylation activity (Lux et al., 1995). Inhibition of CheA activity results in a swimming response by the cells (i.e. the desired behavioural response when encountering gradients of attractant). Chemotaxis towards carbohydrates that are not substrates of the PTS is mediated by a periplasmic-binding- protein-dependent sensory circuit in which substrate- loaded periplasmic carbohydrate-binding proteins interact with the N-terminal module of a transmembrane chemoreceptor to initiate a chemotactic signal (Stock and Surette, 1996). This interaction produces a conformational change in the chemoreceptor which is identical to that generated by the binding of small-molecule chemo-effectors, such as amino acids, to the receptor (Hughson and Hazelbauer, 1996; Zhang et al., 1999). Given that this circuit is chemoreceptor-dependent, it is not surprising that Fig. 1. Schematic of chemoreceptor modular organization. Domains comprising each of the two receptor modules are indicated at the chemoreceptor methylation changes are essential for right. The horizontal shaded bar represents the cytoplasmic mem- adaptation to stimulation by loaded periplasmic binding brane. The transmembrane domains and C-terminal modules are proteins (Hazelbauer et al., 1989). Furthermore, unlike predicted to be highly helical, whereas the structures of the extracellular domains of the B. subtilis receptors used in this study (McpB the PTS-dependent sensory pathway, transport of carbo- and McpC) are unknown. HCD, highly conserved domain; MD, meth- hydrates detected by this type of sensory circuit is ylation domain; MH1, methylated helix 1; SH1, signalling helix 1; TM1, not required for chemotaxis (Ordal and Adler, 1974; transmembrane helix 1. Hazelbauer, 1975). Non-PTS carbohydrates such as galactose, maltose and ribose are detected by this type swimming behaviour of the cell. Phosphorylated CheB of sensory circuit (Stock and Surette, 1996). For example, catalyses demethylation of the chemoreceptors in a neg- the maltose-loaded maltose binding protein (MalE) of E. ative feedback loop to assist adaptation (Stock and coli interacts with the periplasmic portion of the N-terminal Koshland, 1978). module of the aspartate receptor Tar to initiate a chemo- Many carbohydrates, including maltose, ribose and glu- tactic signal (Koiwai and Hayashi, 1979; Richarme, 1982; cose are attractants for Escherichia coli (Adler et al., Gardina et al., 1997). 1973). Chemotaxis towards a given carbohydrate by E. The sensory circuit controlling chemotaxis towards car- coli is controlled by one of two distinct carbohydrate sen- bohydrates by Bacillus subtilis has not been well charac- sory circuits. Substrates of the phosphoenolpyruvate- terized. Preliminary data suggests that the B. subtilis dependent phosphotransferase system (PTS) in E. coli carbohydrate chemotactic sensory circuit is different from are detected by a sensory circuit in which chemotactic either of the two characterized carbohydrate sensory cir- signalling is obligatorily coupled to transport by the PTS cuits of E. coli. For example, an initial survey of carbohy- (Lux et al., 1995; 1999). The PTS catalyses the transport drates as attractants for B. subtilis revealed that only those and phosphorylation of carbohydrate substrates at the carbohydrates that are substrates of the PTS are good expense of phosphoenolpyruvate (PEP) hydrolysis using chemoattractants; B. subtilis accumulated very poorly, or a combination of general and carbohydrate-specific pro- not at all, to non-PTS carbohydrates (Ordal et al., 1979). tein components (Postma et al., 1993, 1996). Carbohy- This correlation suggests that the PTS plays an important drate uptake is initiated when the general PTS functional role in carbohydrate chemotaxis. However, one component, Enyzme I (EI), autophosphorylates using specific chemoreceptor (McpC) is required for chemotaxis PEP as the phosphodonor. This phosphoryl group is towards all tested PTS carbohydrates (Garrity et al., transferred to the second general PTS component, 1998), indicating that B. subtilis employs a carbohydrate Hpr, which shuttles phosphoryl groups to various sensory circuit distinct from that of E. coli, which uses no

McpC detects intracellular chemotactic signals 1355 specific chemoreceptor to mediate responses to PTS tion towards all McpC-specific attractants, including pro- substrates. line and the PTS carbohydrates mannitol, glucose, Our results indicate that the cytoplasmic C-terminal fructose (Fig. 2), as well as trehalose and N-acetylglu- module of the transmembrane chemoreceptor McpC cosamine (not shown), indicating that McpC is sufficient serves as a chemosensory-input device during carbohy- for a chemotactic response in swarm plates. Expression drate chemotaxis. Furthermore, chemotactic signalling in of mcpC did not restore ring formation towards the McpB- response to carbohydrates by B. subtilis requires carbo- specific attractant asparagine (not shown here, but see hydrate transport via the PTS, but not metabolism of the below). Because McpC restored chemotaxis by the recep- imported carbohydrate. We propose a model for carbohy- torless host to five different PTS carbohydrates (each of drate chemotaxis in B. subtilis in which PTS-dependent which is expected to be transported by its own specific carbohydrate transport generates an intracellular chemo- EII), a common sensory mechanism is likely to operate to tactic signal that is transmitted to the C-terminal module control chemotaxis towards all PTS carbohydrates. Fur- of McpC. thermore, McpC serves as the chemoreceptor in this sensory pathway. As an independent measure of chemotaxis, we also Results evaluated the ability of these strains to respond to spatial chemical gradients, as determined by their ability to accu- McpC is the only chemoreceptor required for chemotaxis mulate in attractant-filled capillary tubes. In full agreement towards substrates of the PTS with the swarm assay results, expression of mcpC as the The observation that an mcpC mutant does not respond sole chemoreceptor in the D10mcp background restored chemotactically to PTS carbohydrates suggested that chemotaxis towards McpC-specific attractants, including McpC is, in fact, the receptor responsible for sensing proline and the PTS carbohydrate mannitol (see Supple- these carbohydrates (Garrity et al., 1998). However, alter- mentary material, Fig. S1). Therefore, two different assays native hypotheses in which McpC plays only an indirect indicate that McpC mediates chemotactic responses to role in the carbohydrate sensory pathway were not PTS carbohydrates independently of all other B. subtilis excluded. Therefore, we tested whether McpC is sufficient chemoreceptors. to mediate chemotaxis towards PTS stimuli. If McpC is the genuine receptor for PTS stimuli, it might The cytoplasmic module of McpC mediates chemotactic be expected to support chemotaxis towards PTS carbo- responses to substrates of the PTS hydrates in the absence of all of the other B. subtilis chemoreceptors. We expressed mcpC as the sole Because McpC is a member of the chemoreceptor super- chemoreceptor in a B. subtilis strain carrying null muta- family, we hypothesized that the McpC N-terminal module tions in the genes for all 10 known receptors (D10mcp) conferred ligand specificity to the receptor. To test this and assessed the ability of this strain to perform chemo- hypothesis, we took advantage of two observations: (i) taxis towards PTS carbohydrates using two independent fully functional chimeric receptors can be readily created methods of analysis. by joining N-terminal and C-terminal modules from In the soft-agar swarm assay, cells are inoculated at a different members of the chemoreceptor superfamily; and single point in soft-agar Petri plates consisting of a mini- (ii) B. subtilis McpB does not generate a chemotactic mal medium supplemented with the attractant to be signal in response to PTS carbohydrates. We created tested. As the cells metabolize the attractant in the sur- two reciprocal chimeric receptors by exchanging the N- rounding medium, they create an attractant gradient in the terminal modules of McpB and McpC. These chimeras agar. If the cells can perform chemotaxis towards the were constructed such that the N-terminal module and attractant, they follow this gradient away from the point of cytoplasmic HAMP domain of one receptor were joined inoculation, creating an expanding chemotactic ring of cell to the MD and HCD of the C-terminal module of the other density visible in the agar (see Fig. 2). Thus, the presence (Fig. 3). If PTS carbohydrates (or occupied sugar binding of a clearly defined ring is diagnostic of a chemotactic proteins) interacted directly with the N-terminal module of response. The results of swarm assays, presented in McpC to initiate a chemotactic signal, then the McpC347B Fig. 2, show that a chemotactic ring characteristic of the chimera would be expected to mediate chemotaxis wild-type strain was visible, whereas the mcpC mutant towards PTS carbohydrates as well as the amino acid lacked such a ring, indicating that the mcpC mutant can- proline. Conversely, the McpB354C chimera, which lacks not perform chemotaxis towards these attractants. The the McpC extracellular module, would not be expected to D10mcp host lacking all B. subtilis chemoreceptors also mediate a chemotactic response to carbohydrates or exhibited no chemotactic ring. Expression of mcpC in the proline but would be expected to mediate responses to D10mcp host restored wild-type chemotactic ring forma- asparagine.

1356 C. J. Kristich, G. D. Glekas and G. W. Ordal

Fig. 2. Swarm analysis of various B. subtilis strains. Soft-agar swarm assays were performed as described in Experimental procedures using swarm plates composed of minimal medium supplemented with: A, proline; B, mannitol; C, glucose; and D, fructose. Strains analysed are (according to their position): 12 o’clock, Wild-type (OI1085); 3 o'clock, mcpC::erm (OI3280); 6 o'clock, D10mcp expressing mcpC (OI3974); 9 o'clock, D10mcp (OI3545).

We independently expressed these chimeras as the consistent with those from swarm assays: each chimeric sole chemoreceptor in the D10mcp background and receptor mediated chemotaxis exclusively towards the assessed the ability of the resulting strains to perform amino acid attractant for which the N-terminal module was chemotaxis towards various attractants. Swarm analysis specific (Supplementary material, Fig. S2A and B). of these strains is presented in Fig. 4. For comparison, In sharp contrast to the results with amino acid attrac- D10mcp strains expressing the wild-type mcpB or mcpC tants, Fig. 4 indicates that the C-terminal module of McpC are included. It is apparent from the data that the N- confers specificity of this chemoreceptor towards PTS terminal module of a given receptor specifies the amino carbohydrates. The chimera containing the C-terminal acid attractant to which the receptor responds. For exam- module of McpC, McpB354C, is completely sufficient for ple, the McpC347B chimera supports chemotactic ring for- chemotactic ring formation in response to the PTS carbo- mation in proline-containing swarm plates (as does hydrates mannitol, glucose (Fig. 4C and D), fructose and

McpC), whereas the McpB354C chimera does not trehalose (not shown). Conversely, the reciprocal chimera

(Fig. 4A). Conversely, in asparagine-containing swarm containing the N-terminal module of McpC, McpC347B, is plates, the McpB354C chimera supports chemotactic ring incapable of supporting ring formation in response to any formation (as does McpB) whereas the McpC347B chimera PTS carbohydrates (Fig. 4C and D). The results of capil- does not (Fig. 4B). In addition to swarm assays, we per- lary assays towards PTS carbohydrates are consistent formed capillary assays to evaluate the function of these with the swarm analysis: McpB354C supported chemotaxis chimeric receptors. The capillary assay results are fully towards PTS carbohydrates, whereas McpC347B did not

McpC detects intracellular chemotactic signals 1357 mediate a detectable response (Supplementary material, esis, we constructed strains harbouring null mutations in Fig. S2C and D). Taken together, these results indicate the genes encoding EIImtl, EIIfru and EIItre (the EII’s respon- that the C-terminal module of McpC performs the sensory sible for transport of the PTS carbohydrates mannitol, input function in the carbohydrate chemotactic sensory fructose and trehalose, respectively) and evaluated circuit. chemotaxis of the resulting mutants (Fig. 5). The results The C-terminal module of an MCP contains three func- show that each individual EII is specifically required for tional domains (Fig. 1): the MD, the HCD and the HAMP chemotaxis towards its cognate carbohydrate but not domains. In the McpB354C chimera, only the MD and HCD towards other PTS carbohydrates, supporting the hypoth- domains of McpC are present. Because this chimera is esis that the PTS is involved in production of a carbohy- fully competent at mediating chemotaxis towards PTS drate chemotactic signal. carbohydrates, the McpC HAMP domain cannot have any Transport of fructose by the B. subtilis PTS represents role in the specific detection of PTS stimuli. To determine an unusual case. The genome of B. subtilis encodes two if only one of the two remaining McpC domains was acting EII’s capable of transporting fructose: the fructose-1- as the sensor for PTS stimuli, we constructed an addi- phosphate producing EIIfru transporter (encoded by the tional pair of chimeric receptors in which we exchanged fruA gene), and the fructose-6-phosphate producing EIIlev either the MD or the HCD of McpB354C with the corre- transporter (encoded by the levDEFG operon) (Steinmetz, sponding domain from McpB; we then assessed chemo- 1993). A mutation in fruA does not eliminate the ability to taxis mediated by the resulting chimeras. Our preliminary grow on fructose as a sole carbon source due to the results suggest that the MD of McpC is capable of con- transport of fructose via EIIlev. However, the fruA mutant ferring a PTS-specific response to the chimera (unpub- clearly does not perform chemotaxis towards fructose lished data), although chemotaxis mediated by this (Fig. 5). Analogous results were obtained using soft-agar chimera was substantially reduced relative to McpB354C. swarm assays: the fruA mutant exhibited a chemotactic At the present time we do not understand the basis for ring in mannitol and glucose-containing swarm plates, but this substantially reduced response. in fructose-containing swarm plates it did not produce such a ring despite its ability to grow (not shown). Thus, although both fructose-specific EII’s can transport fructose PTS Enzymes II are specifically required for chemotaxis for use as a growth substrate, some intrinsic difference towards their cognate carbohydrate exists in their ability to generate a chemotactic signal. Because carbohydrates that are chemoattractants for B. subtilis are also substrates of the PTS, we hypothesized Transport of PTS carbohydrates via the PTS is required that the PTS is involved in chemotaxis. To test this hypoth- for production of a chemotactic signal The results presented above suggest that carbohydrate transport by the PTS is required for chemotaxis towards PTS carbohydrates. In order to test whether transport itself is required for production of a chemotactic signal, we

T T T T T T T T M M M M M M M M analysed a mutant with a block in the PTS phosphotrans- 1 2 1 2 1 2 1 2 fer relay required for transport. This mutant harbours an internal deletion of the ptsH gene (encoding the general PTS phosphorelay protein Hpr) and cannot use most PTS carbohydrates as sole carbon sources. It can, however, M M M M M M M M H H H H H H H H 2 1 2 1 2 1 2 1 grow on glucose as a sole carbon source, albeit at a reduced rate, as a result of the presence of one or more

S S S S S S S S H H H H H H H H non-PTS glucose transporters (Paulsen et al., 1998). 2 1 2 1 2 1 2 1 The tethered-cell assay provides an independent method of measuring chemotactic signalling. If the chemotactic sensory circuit is intact, tethered cells produce CCW ﬂagellar rotation upon attractant stimulation. Fig. 3. Chimeric chemoreceptors used in this study. The horizontal The DptsH mutant was evaluated using the tethered-cell shaded bar represents the cytoplasmic membrane. Chimeric receptors were constructed by joining different domains of B. subtilis McpB assay for its ability to produce a chemotactic signal in and McpC as described in Experimental procedures. The chimeras response to two McpC-mediated attractants: (i) the amino were expressed in B. subtilis by integration into the amyE locus under acid proline, and (ii) the PTS carbohydrate glucose. As the control of their native promoters. All receptors were found to be stably expressed at levels similar to the wild-type McpB and McpC. Fig. 6 shows, the DptsH mutant was capable of generating Abbreviations are as in Fig. 1. CCW ﬂagellar rotation in response to the amino acid

Fig. 4. Swarm analysis of B. subtilis strains expressing chimeric receptors. Soft-agar swarm assays were performed as described in Experimental procedures using swarm plates composed of minimal medium supplemented with: A, proline; B, asparagine; C, mannitol; and D, glucose. Strains analysed are (according to their position): 12 o'clock, D10mcp expressing mcpC (OI3974); 3 o'clock, D10mcp expressing mcpB (OI3973); 6 o’clock, D10mcp expressing mcpC347B (OI3976); 9 o'clock, D10mcp expressing mcpB354C (OI3975). attractant proline, demonstrating that McpC is functional concentration of some metabolic product resulting from and that the remainder of the downstream chemotactic degradation of the imported sugar. An analogous mecha- sensory circuit is intact. However, the DptsH mutant did nism is known to be involved in control of catabolite gene not exhibit any glucose-induced CCW rotation (Fig. 6), regulation in B. subtilis (Stulke and Hillen, 2000). If such indicating that the carbohydrate-specific chemotactic sen- a mechanism were operative, mutational blocks prevent- sory circuit is not intact. As a control, we verified that the ing carbohydrate catabolism should eliminate chemotaxis wild-type B. subtilis exhibited CCW flagellar rotation in towards that carbohydrate. To test whether catabolism of response to glucose stimulation under these experimental the PTS carbohydrates was necessary for chemotaxis, we conditions (Fig. 6). Because the DptsH mutant cannot constructed strains harbouring mutations in two genes respond chemotactically to glucose, it appears that trans- required for degradation of different PTS carbohydrates port of carbohydrates via the PTS is necessary for pro- and assessed the chemotactic ability of the resulting duction of a chemotactic signal. mutants. The pgi gene product (phosphoglucose isomerase) catalyses the first step in glucose-6-phosphate catabo- Metabolism of PTS carbohydrates is not required for PTS lism, its conversion to fructose-6-phosphate (Kunst et al., chemotaxis 1997). We disrupted this monocistronic gene using an One possible mechanism by which B. subtilis might sense integrative plasmid. The resulting mutant exhibited a carbohydrate attractants is by detecting changes in the severe growth defect when grown in CAM using glucose

Fig. 5. Capillary assay analysis of B. subtilis strains harbouring mutations in genes encoding various EII transporters. Cultures were grown and induced with the carbohydrate to be tested as described in Experimental procedures and assayed for chemotaxis towards the corresponding carbohydrate. Filled bars in all panels represent wild-type (OI1085) response. A. mtlA::pAIN1 (OI3968; EIImtl). B. treP::pAIN2 (OI3969; EIItre). C. fruA::pAIN3 (OI3970; EIIfru). as the sole carbon source, but not when grown with fruc- chemotaxis in response to both glucose and mannitol, tose or glycerol as the sole carbon source, indicating that indicating that degradation of these carbohydrates to fruc- the mutational block was affecting the expected metabolic tose-1,6-bisphosphate is not required for the generation step. We evaluated the ability of this strain to perform of a chemotactic signal. chemotaxis using the capillary assay (Fig. 7A). Although the mutant exhibited a severe growth defect on glucose, Discussion it performed chemotaxis towards glucose nearly as well as the wild type, suggesting that catabolism of glucose-6- Chemotaxis towards PTS carbohydrates by B. subtilis is phosphate is not required to generate a chemotactic sig- controlled by a sensory circuit distinct from either of the nal in response to glucose. carbohydrate sensory pathways of E. coli. This study pro- This finding is consistent with the complementary result vides evidence that the PTS chemotactic sensory circuit obtained using the DptsH mutant. In the case of the DptsH of B. subtilis uses the cytoplasmic C-terminal module of mutant, glucose can be transported and catabolised using McpC to control chemotaxis towards PTS carbohydrates. a non-PTS pathway but no chemotactic signal is produced (see previous section). Therefore, it appears that catabolism of glucose per se is not sufficient to generate a chemotactic signal. In the case of the pgi mutant, glucose catabolism cannot occur but chemotaxis towards glucose persists. Both of these results support the hypothesis that catabolism of the PTS carbohydrate glucose is not involved in generation of the glucose chemotactic signal. The B. subtilis pfkA gene encodes phosphofructokinase, the glycolytic enzyme that phosphorylates fructose- 6-phosphate to produce fructose-1,6-bisphosphate (Kunst et al., 1997). This reaction comprises the first common catabolic step for both glucose-6-phosphate and mannitol- 1-phosphate. We disrupted pfkA using an integrative plasmid. The resulting mutant exhibited a severe growth defect in CAM containing glucose or mannitol as a sole carbon source, but not in CAM containing glycerol as the sole Fig. 6. Tethered-cell analysis of the B. subtilis DptsH mutant. Cultures carbon source, indicating that the mutational block was were grown and induced with 10 mM glucose as described in Exper- affecting the expected metabolic step. We evaluated the imental procedures. The downward arrow indicates the addition of attractant. Thin solid line, DptsH (OI3302) response to 10-2 M proline; ability of this strain to perform chemotaxis using the cap- thick solid line, DptsH mutant (OI3302) response to 10-4 M glucose; illary assay (Fig. 7B). The pfkA mutant exhibited normal dashed line, wild type (OI1085) response to 10-4 M glucose.

© 2003 Blackwell Publishing Ltd, Molecular Microbiology, 47, 1353–1366 1360 C. J. Kristich, G. D. Glekas and G. W. Ordal carbohydrate requires the presence of its corresponding EII (Fig. 5). These experiments have not excluded the possibility that carbohydrate transport is required for the expression of additional genes involved in PTS sensory transduction; however, the other results presented in this report suggest that EII-mediated carbohydrate transport is sufficient in itself to produce the chemotactic signal. Consistent with this proposal is the observation that the DptsH mutant cannot produce a chemotactic signal in response to glucose (Fig. 6) despite the fact that it can transport glucose using at least one non-PTS transporter (Paulsen et al., 1998). Therefore, transport of glucose per se is not sufficient to generate a chemotactic signal; transport must occur through the PTS machinery in order to generate a chemotactic signal. Furthermore, in at least one case, PTS transport is not sufficient to produce a chemotactic signal. Bacillus subtilis contains 2 EII’s for fructose uptake: (i) the fructose-1- phosphate producing EIIfru transporter (encoded by the fruA gene); and (ii) the fructose-6-phosphate producing EIIlev transporter (encoded by the levDEFG operon) (Steinmetz, 1993). A previous study found that a mutation in fruA that rendered it non-functional resulted in a strain which grew substantially more slowly than the wild type on fructose as the sole carbon source (Gay and Delobbe, 1977). Our preliminary results using 14C-fructose to measure the initial rates of fructose uptake indicate that the fruA mutant transports fructose at ~10% the rate of the Fig. 7. Capillary assay analysis of B. subtilis strains harbouring muta- wild type (unpublished data), suggesting that fructose tions in genes involved in metabolism of PTS carbohydrates. Cultures uptake via EIIlev occurs at a slower rate than fructose were grown and induced with the carbohydrate to be tested as uptake via EIIfru. Our results indicate that the fruA mutant described in Experimental procedures and assayed for chemotaxis towards the corresponding carbohydrate. Filled bars in all panels does not perform chemotaxis towards fructose (Fig. 5) represent wild-type (OI1085) response. despite its ability to transport fructose through EIIlev, sug- A. pgi::pAIN4 (OI3971; phosphoglucose isomerase). gesting that carbohydrate transport must be sufficiently B. pfkA::pAIN5 (OI3972, phosphofructokinase). rapid to generate the chemotactic signal. Thus, PTS chemotactic signalling in B. subtilis appears to be obliga- Although the C-terminal module (in particular, the methy- torily coupled to transport and phosphorylation of the car- lation domain) of the E. coli receptor Tar mediates tactic bohydrate by the PTS. This requirement is reminiscent of responses to fluctuations in temperature (Nara et al., the chemotactic mechanism mediating chemotaxis 1996; Nishiyama et al., 1997; 1999), and the HAMP towards PTS carbohydrates in E. coli. domains of E. coli Tar and Tsr are believed to control Our other data is consistent with this model. In partic- chemotactic responses to cytoplasmic pH changes ular, carbohydrate metabolism is neither sufficient nor (Umemura et al., 2002), to our knowledge this is the first required to generate a chemotactic signal. For instance, example in which the C-terminal module of a transmem- in the two cases just discussed (glucose metabolism in brane receptor mediates chemotactic responses to small- the DptsH mutant and fructose metabolism in the fruA molecule chemical stimuli. mutant), the carbohydrates can be transported and metabolized for growth but still do not produce a chemo- Model for PTS carbohydrate chemotaxis: origin of the tactic signal, indicating that metabolism is not sufficient to chemotactic signal generate such a signal. In two other cases, strains harbouring a mutational block in the catabolic pathway for a Because carbohydrates that are attractants for B. subtilis PTS carbohydrate were capable of performing chemotaxis are also substrates of the PTS, we hypothesized that the towards that carbohydrate (Fig. 7). Therefore, the PTS PTS is involved the chemotactic sensory circuit. Consis- chemotactic sensory circuit remains intact in these tent with this model, chemotaxis towards a given PTS mutants. Alternative models in which the chemotactic sig-

Fig. 8. Model for the PTS chemotactic sensory circuit of B. subtilis. See text for details. This model proposes that enhanced carbohydrate transport via a PTS EII (EIIMtl and its substrate, mannitol, shown here) transiently generates a

T T C M M pool of unphosphorylated EI in the cell (in prin- 1 2 ciple, it is equally likely that unphosphorylated Hpr is the active signalling molecule). Unphos- phorylated EI (or Hpr) interacts with the methylation domain of McpC, stabilizing the receptor in a kinase-activating conformation. Receptor- driven CheA autophosphorylation activity is

M M H H enhanced, resulting in enhanced production of 2 1 Hpr Hpr phosphorylated CheY, which interacts with the ﬂagellar switch to increase the probability of CCW ﬂagellar rotation to produce smooth

S S H H swimming. 2 1 EI EI W CheA

nal results from changes in metabolic intermediate levels of transport is required to produce a chemotactic signal. are therefore unlikely. These results are consistent with a Given the established mechanism of PTS chemotaxis in model which postulates that transport of the carbohydrate E. coli, in which transport-induced changes in the phos- via the PTS is both necessary and sufficient to generate phorylation state of EI alter its affect on CheA activity, it the chemotactic signal. is reasonable to propose that analogous transport- induced changes in the phosphorylation state of B. subtilis EI or Hpr alter their ability to bind to McpC. (iii) Metabolism Model for PTS carbohydrate chemotaxis: integration of the of the PTS carbohydrates is neither required nor sufficient carbohydrate signal into the chemotaxis pathway for chemotaxis. Therefore, none of the catabolic enzymes acting on the carbohydrates (or on any of their metabolic Our results indicate that the C-terminal module of McpC products) plays a role in chemotactic signalling. is responsible for conferring the PTS carbohydrate spec- One simple model to account for chemotaxis towards ificity to this receptor (Fig. 4). Given the structural con- PTS carbohydrates by B. subtilis is depicted in Fig. 8. straints in this chemoreceptor domain (Falke and Kim, According to this model, chemotactic signalling is initiated 2000), it is likely that the C-terminal module of McpC when a PTS carbohydrate is transported and phosphory- detects chemotactic signals by means of protein–protein lated by its corresponding EII. Enhanced transport and interactions. Although the putative interacting partner is phosphorylation of PTS carbohydrates causes a shift in not known, the simplest hypothesis is that one of the the distribution of phosphoryl groups on EI (or Hpr) to two general PTS components (EI or Hpr) possesses this generate an increased level of the unphosphorylated form activity. of one or both. The unphosphorylated form of EI or Hpr The argument supporting this hypothesis can be sum- is specifically capable of interacting with the cytoplasmic marized as follows. (i) McpC mediates chemotaxis module of McpC. This interaction stabilizes a kinase-acti- towards all tested PTS carbohydrates. Because EI and vating conformation of McpC, enhancing CheA activity Hpr are components of the transport pathway for all PTS and, in turn, resulting in production of CCW flagellar rota- carbohydrates, one of these proteins could participate in tion for smooth swimming using the remainder of the a mechanism to couple transport of all PTS carbohydrates downstream chemotactic sensory circuit. with McpC. (ii) Transport of carbohydrate must occur This study has not specifically investigated the mecha- through the PTS in order to generate a chemotactic sig- nisms underlying adaptation to PTS stimuli. However, nal. Furthermore, the data suggest that a sufficient rate given that the B. subtilis PTS responses are mediated by

© 2003 Blackwell Publishing Ltd, Molecular Microbiology, 47, 1353–1366 1362 C. J. Kristich, G. D. Glekas and G. W. Ordal a transmembrane chemoreceptor (McpC), it is likely that the C-terminal module could stabilize a tightly packed standard mechanisms, already in place for adaptation dur- conformation of the receptor helices (or a loosely packed ing chemotaxis towards amino acid stimuli (Zimmer et al., conformation, whichever is appropriate to activate the 2000; Karatan et al., 2001; Kirby et al., 2001), are also kinase), thus shifting the receptor to the kinase-activating used in response to PTS carbohydrates. Thus, no new form. This receptor conformation would then enhance adaptational mechanism need be postulated. CheA activity, resulting in the production of phosphorylated CheY and subsequently CCW flagellar rotation for smooth swimming. Model for PTS carbohydrate chemotaxis: McpC function In summary, our studies on the mechanism of PTS According to the current paradigm of receptor function, carbohydrate chemotaxis in B. subtilis have provided the environmental signals originating from the receptor N- basis for a unique model of chemotactic stimulus- terminal module are sent to the methylation domain of the response coupling in which the PTS transport system C-terminal module (Falke and Hazelbauer, 2001). The works in concert with the C-terminal module of the function of the methylation domain has been compared to chemoreceptor McpC to transduce carbohydrate chemo- that of the central processing unit (CPU) of a computer: tactic signals. These results form the foundation for future it receives sensory signals (from the N-terminal module detailed studies of this stimulus-response circuit. of the receptor) and adaptational signals (by means of receptor methylesterification) and integrates those signals Experimental procedures to provide the appropriate averaged output to the HCD such that CheA kinase activity is properly controlled Bacterial strains and plasmids (Trammell and Falke, 1999). For B. subtilis McpC, our Bacterial strains and plasmids used in this study are listed in results indicate that the C-terminal module, and perhaps Tables 1 and 2 respectively. The DNA manipulations were the methylation domain, possesses an additional function: performed according to standard protocols (Sambrook et al., it acts as a sensor for chemotactic stimuli. 1989). Because of the structural constraints inherent in the Construction of pAIN402, pAIN450, pAIN750 mcpC, and coiled-coil helical bundle of the C-terminal module, it pAIN750 mcpB: Plasmid pAIN750 mcpC was constructed by seems unlikely that this module is capable of directly PCR-amplifying the mcpC gene and its promoter from B. binding intracellular carbohydrates to initiate a chemotac- subtilis chromosomal DNA using Pfu polymerase (Strat- agene). The amplification product was cloned into pSK– using tic signal. Therefore, the chemotactic signal is probably primer-encoded EcoRI and NotI restriction sites (creating transmitted to the receptor via a protein–protein interac- pAIN401). A plasmid clone of mcpB in pSK was available tion. How might such an interaction regulate receptor (pAIN700) (Zimmer et al., 2000). For another investigation, a activity? BglII restriction site had been inserted into the region coding Although the conformational changes that occur in the for each of the receptor C-terminal modules to create plas- transmembrane region of chemoreceptors as a result of mids pAIN402 and pAIN450 (containing mcpC and mcpB, ligand binding have been extensively characterized (Falke respectively) using a previously described method (Zimmer et al., 2000). These nucleotide sequence alterations were and Hazelbauer, 2001), comparatively little is known phenotypically ‘silent’; that is, they did not result in any about the changes that occur in the C-terminal module to change of the amino acid sequence of the receptors. The propagate the chemotactic signal to the bound kinase. inserted BglII site was located in the receptor genes at the One model has been proposed in which the packing inter- codons corresponding to a cytoplasmic amino acid sequence actions of the helices within the coiled-coil bundle of the conserved in both receptors approximately 20 residues C- MD directly control the proximity of the two monomers of terminal to TM2 (residue 324 in McpB; residue 317 in McpC). bound CheA kinase (and therefore control the rate of Subsequently, the chemoreceptor coding sequences were cloned into the amyE integration plasmid pAIN750 using CheA autophosphorylation) (Trammell and Falke, 1999). EcoRI and NotI restriction sites. This model proposes that the packing interactions of the Construction of pAIN750 mcpB354C and pAIN750 coiled-coil helices in the MD are regulated by signals from mcpC347B: The numbering of these reciprocal chimeras is the N-terminal module and by modification of the methy- offset (354 versus 347) because the McpB N-terminal mod- lation sites on these helices. These various parameters ule is seven amino acids longer than the McpC N-terminal are then integrated to yield an ‘averaged’ level of packing module. Opposing primers were designed to prime DNA syn- between the helices, which is subsequently translated into thesis outwards from the desired chimera fusion point in each receptor (inverse PCR). A second set of overlapping primers an appropriate level of CheA activity (Trammell and Falke, was designed to prime synthesis from the unique ScaI site 1999). in the bla ampicillin-resistance gene of pSK. Polmerase chain In the context of such a model, it is easy to imagine how reaction amplification from pAIN450 using the complemen- a protein–protein interaction could influence receptor sig- tary sets of primers yielded two amplification products: each nalling activity. Binding of unphosphorylated EI (or Hpr) to contained roughly half of the mcpB gene in addition to a

Table 1. Strains used in this study. The resulting recombinant plasmids were transformed into B. subtilis with selection for EmR by Campbell-type recombina- Relevant genotype or tion in the gene of interest. Strain description Reference

OI1085 trpF7 hisH2 metC che+ (Ullah and Ordal, 1981) OI3280 OI1085 mcpC1::erm (Muller et al., 1997) Growth media and solutions OI3302 OI1085 DptsH ykwC::cat (Garrity et al., 1998) OI3968 OI1085 mtlA1::pAIN1 This work Tr yptone broth (TB) contained 1% tryptone and 0.5% NaCl. OI3969 OI1085 treP1::pAIN2 This work Capillary assay minimal medium (CAM) contained 50 mM

OI3970 OI1085 fruA1::pAIN3 This work potassium phosphate, pH 7, 1.2 mM MgCl2, 1 mM OI3971 OI1085 pgi-1::pAIN4 This work (NH4) SO , 0.14 mM CaCl , 0.01 mM MnCl , 0.01% tryptone, OI3972 OI1085 pfkA1::pAIN5 This work 2 4 2 2 0.005% NaCl, and 15 g ml-1 required amino acids (Ordal OI3545 D10mcp che+ (Hou et al., 2000) m OI3973 OI3545 amyE5720::mcpB This work and Goldman, 1975). Chemotaxis buffer (CB) contained OI3974 OI3545 amyE5720::mcpC This work 10 mM potassium phosphate pH 7, 0.14 mM CaCl2, 0.3 mM

OI3975 OI3545 amyE5720::mcpB354C This work (NH4)2SO4, 0.1 mM EDTA, 5 mM sodium lactate, and 0.05% OI3976 OI3545 amyE5720::mcpC347B This work glycerol (Ordal and Goldman, 1975). Spizizen's salt mix (1x)

contained 0.2% (NH4)2SO4, 1.4% K2HPO4, 0.6% KH2PO4,

0.1% sodium citrate, and 0.02% MgSO4 (Spizizen, 1958). Selections for antibiotic resistance were performed on media portion of the pSK backbone and terminated in blunt ends. containing antibiotics at the following concentrations (in Identical reactions were performed for mcpC using pAIN402 -1 R R mg ml ): spectinomycin (Sp ; 100) and erythromycin (Em ; as the template. After digestion with ScaI, the appropriate 0.4). fragments were ligated such that the N-terminal mcpB fragment was fused to the C-terminal mcpC fragment and vice- versa, creating pAIN450 mcpB354C and pAIN402 mcpC347B. Soft-agar swarm assay for chemotaxis These chimeric receptor-coding sequences were then cloned into pAIN750, as above. Minimal-proline swarm plates contained 0.25% agar, 0.1 ¥ Chemoreceptors were expressed in B. subtilis strains by Spizizen’s salt mix, 5 mg ml-1 required amino acids, 0.7 mM integration into the amyE locus, as described (Zimmer et al., sorbitol and 0.2 mM proline. Minimal-asparagine swarm 2000). All chimeric receptors were expressed at levels plates contained 0.3 mM asparagine in place of proline. Car- similar to wild-type McpC in B. subtilis, as determined by bohydrate-containing minimal swarm plates contained 0.25% immunoblot analysis using anti-McpB or anti-McpC antisera agar, 0.1 ¥ Spizizen's salt mix, 5 mg ml-1 required amino acids (not shown). and 0.1 mM of the indicated carbohydrate (mannitol, glucose, Gene disruptions with the integrative plasmid pHV501 fructose, trehalose, or N-acetylglucosamine). Strains to be were carried out by ﬁrst amplifying an internal segment of the analysed were cultivated on Tryptose Blood Agar Base gene to be disrupted using PCR with primers encoding a (TBAB) or TBAB plus antibiotic (as appropriate) overnight at HindIII site (N-terminal end of the encoded protein) and a 30∞C and placed at room temperature until inoculation of the BamHI site (C-terminal end). The ampliﬁed internal fragment swarm plate (8–11 h later) with sterile toothpicks. Plates were of approximately 250–300 bp was digested with HindIII/ incubated at 30∞C, 12–15 h for amino acid swarm plates and BamHI and cloned into the HindIII/BamHI sites of pHV501. 20–25 h for carbohydrate swarm plates. All experiments were

Table 2. Plasmids used in this study.

Plasmid Description Reference pHV501 integrative plasmid for B. subtilis; EmR, ApR Vagner et al. (1998) pSK E. coli cloning vector; ApR Stratagene pAIN1 pHV501 containing internal HindIII/BamHI fragment of the B. subtilis mtlA gene This work pAIN2 pHV501 containing internal HindIII/BamHI fragment of the B. subtilis treP gene This work pAIN3 pHV501 containing internal HindIII/BamHI fragment of the B. subtilis fruA gene This work pAIN4 pHV501 containing internal HindIII/BamHI fragment of the B. subtilis pgi gene This work pAIN5 pHV501 containing internal HindIII/BamHI fragment of the B. subtilis pfkA gene This work pAIN401 mcpC PCR product cloned into pSK using primer-encoded EcoRI/NotI restriction sites This work pAIN402 mcpC from pAIN401 containing silent mutations introducing a BglII site in the C-terminal module This work pAIN450 mcpB from pAIN700 containing silent mutations introducing a BglII site in the C-terminal module This work pAIN450mcpB354C pSK containing mcpB354C chimeric receptor This work pAIN402mcpC347B pSK containing mcpC347B chimeric receptor This work pAIN700 pSK containing mcpB with silent mutations that eliminate an internal EcoRI site Zimmer et al. (2000) pAIN750 B. subtilis amyE integration vector; ApR, SpR Kristich and Ordal (2002) pAIN750mcpB pAIN750 containing mcpB subcloned from pAIN450 This work pAIN750mcpC pAIN750 containing mcpC subcloned from pAIN402 This work pAIN750mcpB354CamyE integration vector containing mcpB354C This work pAIN750mcpC347BamyE integration vector containing mcpC347B This work

© 2003 Blackwell Publishing Ltd, Molecular Microbiology, 47, 1353–1366 1364 C. J. Kristich, G. D. Glekas and G. W. Ordal repeated a minimum of four independent times. Photographs Armitage, J.P. (1999) Bacterial tactic responses. Adv Microb of representative swarm plates were taken using dark-field Physiol 41: 229–289. illumination at the Biomedical Communications Center in the Barak, R., and Eisenbach, M. (1992) Correlation between College of Veterinary Medicine, UIUC. phosphorylation of the chemotaxis protein CheY and its activity at the flagellar motor. Biochemistry 31: 1821–1826. Chauvin, F., Brand, L., and Roseman, S. (1994) Sugar trans- Capillary assay for chemotaxis port by the bacterial phosphotransferase system. Charac- terization of the Escherichia coli enzyme I monomer/dimer The capillary assay was performed as previously described, transition kinetics by fluorescence anisotropy. J Biol Chem with slight modifications (Ordal and Goldman, 1975). Strains 269: 20270–20274. to be analysed were cultivated on TBAB or TBAB plus anti- Falke, J.J., and Kim, S.H. (2000) Structure of a conserved biotic overnight at 30∞C. Cells were suspended at receptor domain that regulates kinase activity: the cyto- A600 = 0.014 in CAM supplemented with 20 mM glycerol, plasmic domain of bacterial taxis receptors. Curr Opin incubated with agitation (250 r.p.m.) at 37∞C for 6 h, and Struct Biol 10: 462–469. harvested for the experiment. For assays of carbohydrate Falke, J.J., and Hazelbauer, G.L. (2001) Transmembrane chemotaxis, the carbohydrate being tested was added to signaling in bacterial chemoreceptors. Trends Biochem Sci cultures (for induction of the transport apparatus) 75 min 26: 257–265. before harvest of the cells. The bacteria from suspensions at Feng, X., Baumgartner, J.W., and Hazelbauer, G.L. (1997) -1 A600 = 0.001 in CB plus Cm (250 mg ml ) that accumulated High- and low-abundance chemoreceptors in Escherichia in attractant-filled 1 ml microcapillaries after 30 min were enu- coli: differential activities associated with closely related merated by plating on TB agar. Each strain was assayed in cytoplasmic domains. J Bacteriol 179: 6714–6720. triplicate, and all experiments were repeated a minimum of Gardina, P.J., Bormans, A.F., Hawkins, M.A., Meeker, J.W., two independent times. and Manson, M.D. (1997) Maltose-binding protein interacts simultaneously and asymmetrically with both subunits of the Tar chemoreceptor. Mol Microbiol 23: 1181–1191. Tethered-cell assay Garrity, L.F., Schiel, S.L., Merrill, R., Reizer, J., Saier, M.H. The tethered-cell assay was performed as previously Jr and Ordal, G.W. (1998) Unique regulation of carbohy- described, with slight modifications (Kirby et al., 1999). Cul- drate chemotaxis in Bacillus subtilis by the phosphoe- tures were grown in CAM supplemented with 1 mg ml-1 nolpyruvate-dependent phosphotransferase system and disodium succinate and 1 mg ml-1 monopotassium glutamate the methyl-accepting chemotaxis protein McpC. J Bacteriol as the carbon source for 8 h at 37∞C with agitation 180: 4475–4480. (250 r.p.m). Glucose (10 mM) was added 75 min before har- Gay, P., and Delobbe, A. (1977) Fructose transport in Bacil- vesting of the cells for the experiment. The modified CB lus subtilis. Eur J Biochem 79: 363–373. contained 10 mM potassium phosphate, pH 7, 0.14 mM Hazelbauer, G.L. (1975) Maltose chemoreceptor of Escheri- -1 chia coli. J Bacteriol 122: 206–214. CaCl2, 0.3 mM (NH4)2SO4, 0.1 mM EDTA, and 1 mg ml disodium succinate. Hazelbauer, G.L., Park, C., and Nowlin, D.M. (1989) Adap- tational ‘crosstalk’ and the crucial role of methylation in chemotactic migration by Escherichia coli. Proc Natl Acad Acknowledgements Sci USA 86: 1448–1452. Hess, J.F., Oosawa, K., Kaplan, N., and Simon, M.I. (1988) We are indebted to John Kirby for valuable discussions and Phosphorylation of three proteins in the signaling pathway critical commentary on the manuscript. This research was of bacterial chemotaxis. Cell 53: 79–87. supported by a grant from the National Institutes of Health to Hou, S., Larsen, R.W., Boudko, D., Riley, C.W., Karatan, E., GWO (GM54365). Zimmer, M., Ordal, G.W., and Alam, M. (2000) Myoglobin- like aerotaxis transducers in Archaea and Bacteria. Nature 403: 540–544. Supplementary material Hughson, A.G., and Hazelbauer, G.L. (1996) Detecting the conformational change of transmembrane signaling in a The following material is available from: http://www. bacterial chemoreceptor by measuring effects on disulfide blackwellpublishing.com/products/journals/suppmat/mole/ cross-linking in vivo. Proc Natl Acad Sci USA 93: 11546– mole3375/mmi3375sm.htm 11551. Fig. S1. Capillary assay analysis of various B. subtitlis Karatan, E., Saulmon, M.M., Bunn, M.W., and Ordal, G.W. strains. (2001) Phosphorylation of the response regulator CheV is Fig. S2. Capillary assay analysis of B. subtilis strains required for adaptation to attractants during Bacillus subti- expressing chimeric receptors. lis chemotaxis. J Biol Chem 276: 43618–43626. Kirby, J.R., Saulmon, M.M., Kristich, C.J., and Ordal, G.W. References (1999) CheY-dependent methylation of the asparagine receptor, McpB, during chemotaxis in Bacillus subtilis. J Adler, J., Hazelbauer, G.L., and Dahl, M.M. (1973) Chemot- Biol Chem 274: 11092–11100. axis toward sugars in Escherichia coli. J Bacteriol 115: Kirby, J.R., Kristich, C.J., Saulmon, M.M., Zimmer, M.A., 824–847. Garrity, L.F., Zhulin, I.B., and Ordal, G.W. (2001) CheC is

related to the family of flagellar switch proteins and acts Postma, P.W., Lengeler, J.W., and Jacobson, G.R. (1996) independently from CheD to control chemotaxis in Bacillus Phosphoenolpyruvate: carbohydrate phosphotransferase subtilis. Mol Microbiol 42: 573–585. systems. In Escherichia Coli and Salmonella. Cellular and Koiwai, O., and Hayashi, H. (1979) Studies on bacterial Molecular Biology. Neidhardt, F.C., Curtiss, R.I., Ingraham, chemotaxis. IV. Interaction of maltose receptor with a mem- J.L., Lin, E.C.C., Low, K.B., Magasanik, B., et al. (eds). brane-bound chemosensing component. J Biochem Washington, DC: American Society for Microbiology Press, (Tokyo) 86: 27–34. pp. 1149–1174. Krikos, A., Conley, M.P., Boyd, A., Berg, H.C., and Simon, Richarme, G. (1982) Interaction of the maltose-binding pro- M.I. (1985) Chimeric chemosensory transducers of tein with membrane vesicles of Escherichia coli. J Bacteriol Escherichia coli. Proc Natl Acad Sci USA 82: 1326–1330. 149: 662–667. Kristich, C.J., and Ordal, G.W. (2002) Bacillus subtilis CheD Sambrook, J., Fritsch, E.F., and Maniatis, T. (1989) Molecu- is a chemoreceptor modification enzyme required for lar Cloning. A Laboratory Manual. Cold Spring Harbor, NY: chemotaxis. J Biol Chem 277: 25356–25362. Cold Spring Harbor Laboratory Press. Kunst, F., Ogasawara, N., Moszer, I., Albertini, A.M., Alloni, Spizizen, J. (1958) Transformation of biochemically deficient G., Azevedo, V., et al. (1997) The complete genome strains of Bacillus subtilis by deoxyribonucleate. Proc Natl sequence of the Gram-positive bacterium Bacillus subtilis. Acad Sci USA 44: 1072–1078. Nature 390: 249–256. Steinmetz, M. (1993) Carbohydrate catabolism: pathways, Lux, R., Jahreis, K., Bettenbrock, K., Parkinson, J.S., and enzymes, genetic regulation, and evolution. In Bacillus Lengeler, J.W. (1995) Coupling the phosphotransferase Subtilis and Other Gram-Positive Bacteria. Biochemistry, system and the methyl-accepting chemotaxis protein- Physiology, and Molecular Genetics. Sonenshein, A.L., dependent chemotaxis signaling pathways of Escherichia Hoch, J.A., and Losick, R., (eds). Washington D.C: Amer- coli. Proc Natl Acad Sci USA 92: 11583–11587. ican Society for Microbiology Press, pp. 157–170. Lux, R., Munasinghe, V.R., Castellano, F., Lengeler, J.W., Stock, J.B., and Koshland, D.E. Jr. (1978) A protein methyl- Corrie, J.E., and Khan, S. (1999) Elucidation of a PTS- esterase involved in bacterial sensing. Proc Natl Acad Sci carbohydrate chemotactic signal pathway in Escherichia USA 75: 3659–3663. coli using a time-resolved behavioral assay. Mol Biol Cell Stock, J.B., and Surette, M.G. (1996) Chemotaxis. In Escher- 10: 1133–1146. ichia Coli and Salmonella. Cellular and Molecular Biology. Muller, J., Schiel, S., Ordal, G.W., and Saxild, H.H. (1997) Neidhardt, F.C., Curtiss, R.I., Ingraham, J.L., Lin, E.C.C., Functional and genetic characterization of mcpC, which Low, K.B., Magasanik, B., et al. (eds). Washington, DC: encodes a third methyl-accepting chemotaxis protein in American Society for Microbiology Press, pp. 1103–1129. Bacillus subtilis. Microbiology 143: 3231–3240. Stulke, J., and Hillen, W. (2000) Regulation of carbon Nara, T., Kawagishi, I., Nishiyama, S., Homma, M., and Imae, catabolism in Bacillus species. Annu Rev Microbiol 54: Y. (1996) Modulation of the thermosensing profile of the 849–880. Escherichia coli aspartate receptor tar by covalent modifi- Trammell, M.A., and Falke, J.J. (1999) Identification of a site cation of its methyl-accepting sites. J Biol Chem 271: critical for kinase regulation on the central processing unit 17932–17936. (CPU) helix of the aspartate receptor. Biochemistry 38: Nishiyama, S., Nara, T., Homma, M., Imae, Y., and Kawag- 329–336. ishi, I. (1997) Thermosensing properties of mutant aspar- Ullah, A.H., and Ordal, G.W. (1981) In vivo and in vitro tate chemoreceptors with methyl-accepting sites replaced chemotactic methylation in Bacillus subtilis. J Bacteriol singly or multiply by alanine. J Bacteriol 179: 6573–6580. 145: 958–965. Nishiyama, S.I., Umemura, T., Nara, T., Homma, M., and Umemura, T., Matsumoto, Y., Ohnishi, K., Homma, M., and Kawagishi, I. (1999) Conversion of a bacterial warm Kawagishi, I. (2002) Sensing of cytoplasmic pH by bacte- sensor to a cold sensor by methylation of a single residue rial chemoreceptors involves the linker region that con- in the presence of an attractant. Mol Microbiol 32: 357– nects the membrane-spanning and the signal-modulating 365. helices. J Biol Chem 277: 1593–1598. Ordal, G.W., and Adler, J. (1974) Properties of mutants in Vagner, V., Dervyn, E., and Ehrlich, S.D. (1998) A vector for galactose taxis and transport. J Bacteriol 117: 517–526. systematic gene inactivation in Bacillus subtilis. Microbiol- Ordal, G.W., and Goldman, D.J. (1975) Chemotaxis away ogy 144: 3097–3104. from uncouplers of oxidative phosphorylation in Bacillus Weerasuriya, S., Schneider, B.M., and Manson, M.D. (1998) subtilis. Science 189: 802–805. Chimeric chemoreceptors in Escherichia coli: signaling Ordal, G.W., Villani, D.P., and Rosendahl, M.S. (1979) properties of Tar-Tap and Tap-Tar hybrids. J Bacteriol 180: Chemotaxis towards sugars by Bacillus subtilis. J Gen 914–920. Microbiol 115: 167–172. Weigel, N., Kukuruzinska, M.A., Nakazawa, A., Waygood, Paulsen, I.T., Chauvaux, S., Choi, P., and Saier, M.H. Jr E.B., and Roseman, S. (1982) Sugar transport by the bac- (1998) Characterization of glucose-specific catabolite terial phosphotransferase system. Phosphoryl transfer repression-resistant mutants of Bacillus subtilis: identifica- reactions catalyzed by enzyme I of Salmonella typhimu- tion of a novel hexose: H+ symporter. J Bacteriol 180: 498– rium. J Biol Chem 257: 14477–14491. 504. Welch, M., Oosawa, K., Aizawa, S.I., and Eisenbach, M. Postma, P.W., Lengeler, J.W., and Jacobson, G.R. (1993) (1994) Effects of phosphorylation, Mg2+, and conformation Phosphoenolpyruvate: carbohydrate phosphotransferase of the chemotaxis protein CheY on its binding to the flagel- systems of bacteria. Microbiol Rev 57: 543–594. lar switch protein FliM. Biochemistry 33: 10470–10476.

Zhang, Y., Gardina, P.J., Kuebler, A.S., Kang, H.S., Christo- ers: from physiology to genomics and back. Adv Microb pher, J.A., and Manson, M.D. (1999) Model of maltose- Physiol 45: 157–198. binding protein/chemoreceptor complex supports intra- Zimmer, M.A., Tiu, J., Collins, M.A., and Ordal, G.W. (2000) subunit signaling mechanism. Proc Natl Acad Sci USA 96: Selective methylation changes on the Bacillus subtilis 939–944. chemotaxis receptor McpB promote adaptation. J Biol Zhulin, I.B. (2001) The superfamily of chemotaxis transduc- Chem 275: 24264–24272.