SEARCH

SEARCH BY CITATION

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Motile bacteria differ largely in the number of chemoreceptors
  5. Different signals – different cellular compartments of sensing – different chemoreceptor sensor domains
  6. Direct and indirect signal recognition at the chemoreceptor
  7. Size differences in the chemoreceptor signalling and adaptation domain
  8. Effect of signal recognition on CheA autophosphorylation and receptor methylation
  9. Additional signal transduction proteins
  10. Difference in CheR-mediated adaptation
  11. Paralogues of signal transduction proteins – parallel signalling pathways
  12. Differences in the motor mechanism
  13. Signal termination
  14. Acknowledgements
  15. References

Bacterial taxis is one of the most investigated signal transduction mechanisms. Studies of taxis have primarily used Escherichia coli and Salmonella as model organism. However, more recent studies of other bacterial species revealed a significant diversity in the chemotaxis mechanisms which are reviewed here. Differences include the genomic abundance, size and topology of chemoreceptors, the mode of signal binding, the presence of additional cytoplasmic signal transduction proteins or the motor mechanism. This diversity of chemotactic mechanisms is partly due to the diverse nature of input signals. However, the physiological reasons for the majority of differences in the taxis systems are poorly understood and its elucidation represents a major research need.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Motile bacteria differ largely in the number of chemoreceptors
  5. Different signals – different cellular compartments of sensing – different chemoreceptor sensor domains
  6. Direct and indirect signal recognition at the chemoreceptor
  7. Size differences in the chemoreceptor signalling and adaptation domain
  8. Effect of signal recognition on CheA autophosphorylation and receptor methylation
  9. Additional signal transduction proteins
  10. Difference in CheR-mediated adaptation
  11. Paralogues of signal transduction proteins – parallel signalling pathways
  12. Differences in the motor mechanism
  13. Signal termination
  14. Acknowledgements
  15. References

Bacteria are exposed to rapid changes in their environment and efficient bacterial adaptation is of crucial importance for survival. The primary mechanism by which bacteria sense environmental signals are through the action of two-component regulatory systems (TCS), which contain a sensor kinase and a response regulator as essential components. The binding of signals at the sensor kinase modulates its autophosphorylation activity and consequently its transphosphorylation activity towards the cognate response regulator. The prototypal response regulator is composed of a response regulator receiver domain, which contains the phosphoryl group accepting aspartate residue, and an output domain. Since the majority of response regulators contain a DNA-binding output domain, the final regulatory output is mainly at the level of transcriptional regulation (Galperin, 2010). Other response regulator types contain output domains that interact with RNA, small ligands and proteins or have a transporter activity. The corresponding regulatory activity occurs at the transcriptional, post-transcriptional or post-translational level. Due to the large diversity of signal molecules recognized, TCS are characterized by a significant diversity in their mechanisms of sensing and action that we have reviewed recently (Krell et al., 2010).

In addition to bacterial adaptation to environmental changes at transcriptional and post-transcriptional level, many bacteria have acquired motile capacities which permit bacterial movements towards or away from different environmental stimuli. This process is also based on TCS signalling and most of what we know in the field of taxis is due to the study E. coli and Salmonella. Taxis is achieved through the concerted action of the excitatory pathway and adaptational mechanism(s). The excitatory pathway is initiated by the recognition of the signal by methylaccepting chemotaxis proteins (MCPs), which are typically transmembrane proteins with a periplasmic ligand binding region and a cytosolic signalling and adaptation domain, which forms a long α-hairpin. The CheA autokinase and the coupling protein CheW bind to the MCP signalling domain close to the apex of the α-hairpin (Park et al., 2006). Any of these three components are necessary for the formation of this ultrastable signalling complex (Erbse and Falke, 2009) and the essential role of CheW in signalling is underlined by the fact that cheW null mutants are non-chemotactic (Szurmant and Ordal, 2004). Signal recognition by the MCP ligand binding regions produces a molecular stimulus that modulates CheA autophosphorylation and in turn transphosphorylation activity towards the response regulators CheY and CheB (Borkovich et al., 1989; Garrity and Ordal, 1997). Contrary to the prototypal response regulator, CheY is only composed of a receiver domain. When phosphorylated, CheY undergoes a conformational changewhich permits an interaction with the flagellar motor. This interaction changes the sense of the flagellar motor rotation from counterclockwise to clockwise, which causes a change from swimming to tumbling (Sowa and Berry, 2008). Compared with other response regulators, CheY was found to undergo a very rapid spontaneous dephosphorylation (in the range of a few seconds, Sanders et al., 1989). However, to achieve optimal taxis this rate of spontaneous dephosphorylation is too slow and it is the role of the CheZ phosphatase to catalyse CheY dephosphorylation.

Adaptational mechanisms are indispensable for taxis and correspond to a restoration of the prestimulus behaviour in the presence of the stimulus. The canonical adaptation mechanism consists in the methylation and demethylation of MCPs which is catalysed by the CheR methyltransferase and CheB methylesterase respectively (Fig. 1). Ligand binding at the MCP increases reversible methylation of 4–6 glutamate residues at the MCP signalling domain, which in turn modulates the capacity of the receptor to alter CheA autophosphorylation (Bren and Eisenbach, 2000). Receptor methylation was found to significantly alter the affinity of signal molecules for the MCP–CheA–CheW ternary complex (Li and Weis, 2000). In contrast to the CheR methyltransferase, the activity of the CheB methylesterase is largely modulated by the phosphorylation by CheA. The fact that phosphoryltransfer from CheA to CheY is faster than to CheB ensures that signalling occurs before the onset of adaptation. CheB-mediated demethylation sets back the signalling state of the receptor–CheA–CheW complex. There is a series of reviews which describe in detail the molecular processes and components that give rise to taxis (Wadhams and Armitage, 2004; Baker et al., 2006; Hazelbauer et al., 2008).

image

Figure 1. Schematic representation of proteins and processes in chemotactic signalling. Proteins found in E. coli are shown in red (or other colour). Proteins and processes shown in blue occur in other bacteria.

Download figure to PowerPoint

In the study of taxis, Escherichia coli and Salmonella have served as the primary model organisms. However, the study of taxis in other bacteria has shown that bacterial taxis as a whole is characterized by an enormous diversity present at many different levels. It was suggested that the chemotactic system in E. coli corresponds to a streamlined version resulting from the evolutionary loss of chemotaxis proteins still present in other organisms (Miller et al., 2009). In this mini-review we aim at summarizing the diversity of taxis in bacteria and archaea.

Motile bacteria differ largely in the number of chemoreceptors

  1. Top of page
  2. Summary
  3. Introduction
  4. Motile bacteria differ largely in the number of chemoreceptors
  5. Different signals – different cellular compartments of sensing – different chemoreceptor sensor domains
  6. Direct and indirect signal recognition at the chemoreceptor
  7. Size differences in the chemoreceptor signalling and adaptation domain
  8. Effect of signal recognition on CheA autophosphorylation and receptor methylation
  9. Additional signal transduction proteins
  10. Difference in CheR-mediated adaptation
  11. Paralogues of signal transduction proteins – parallel signalling pathways
  12. Differences in the motor mechanism
  13. Signal termination
  14. Acknowledgements
  15. References

Taxis has been observed in response to many different types of environmental stimuli such as metabolites and signalling molecules, temperature, light, salinity, oxygen, the magnetic field and pH. The specificity of a tactic response is determined by the MCPs. The study of chemotaxis has revealed that different bacteria respond chemotactically to different compounds. The four chemoreceptors of E. coli mediate chemotaxis primarily to some amino acids, dipeptides, sugars and pyrimidine (Hazelbauer et al., 2008; Liu and Parales, 2008). However, other bacteria were found to posses a chemotactic behaviour to many different types of compounds like mono- and bicyclic aromatic hydrocarbons, tricarboxylic acid cycle intermediates, inorganic phosphate, plant hormones (ethylene) or herbicides (Parales and Harwood, 2002; Kato et al., 2008; Liu and Parales, 2009; Lacal et al., 2010a).

The analysis of 238 completed bacterial genomes, which have at least one mcp gene, has revealed very large differences in the number of mcp genes per genome (Lacal et al., 2010b). Interestingly, there was only a weak correlation between the genome size and the number of mcp genes, but bacterial lifestyle was found to correlate with the abundance of mcp genes. Strict pathogens formed the group that had the lowest number of mcp genes, as exemplified by Bacillus anthracis str. Sterne and Yersinia pestis Nepal516 that possess a single mcp gene. The group of bacteria with most mcp genes is that characterized by a complex behaviour and by the capacity to maintain interaction with other living beings. Examples of bacteria with complex behaviour are Myxococcus xanthus (21 mcp genes) that undergoes cell differentiation and developmental changes through their life cycles or Magnetospirillum magnetotacticum (61 mcp genes), which uses ferric iron as the terminal electron acceptor and exerts magnetotaxis allowing migration along the geomagnetic field lines. Bacteria that maintain interactions are for example Rhizobium leguminosarum, a plant symbiont, Agrobacterium tumefaciens, which induces plant tumours or Bdelliovibrio bacteriovorus that parasites other Gram-negative bacteria. These bacteria have typically 20–60 mcp genes per genome. However, the cognate signal molecules are only known for a relatively small number of chemoreceptors and their identification represents a major research need in this field.

Different signals – different cellular compartments of sensing – different chemoreceptor sensor domains

  1. Top of page
  2. Summary
  3. Introduction
  4. Motile bacteria differ largely in the number of chemoreceptors
  5. Different signals – different cellular compartments of sensing – different chemoreceptor sensor domains
  6. Direct and indirect signal recognition at the chemoreceptor
  7. Size differences in the chemoreceptor signalling and adaptation domain
  8. Effect of signal recognition on CheA autophosphorylation and receptor methylation
  9. Additional signal transduction proteins
  10. Difference in CheR-mediated adaptation
  11. Paralogues of signal transduction proteins – parallel signalling pathways
  12. Differences in the motor mechanism
  13. Signal termination
  14. Acknowledgements
  15. References

Methylaccepting chemotaxis proteins are composed of a ligand binding region and a cytoplasmic methylaccepting signalling and adaptation domain. The different types of input signals have shaped MCPs in a way that they differ largely in receptor topology and the type of sensor domain (Lacal et al., 2010b). As shown in Fig. 2 six major topologies can be distinguished. MCPs can be either membrane bound or cytoplasmic and differ in the location of the ligand binding region, that is either present in the extracellular space or the cytosol. Interestingly, extracellular signal recognition appears to dominate in bacteria whereas archaea have a higher amount of cytoplasmic sensing receptors (Lacal et al., 2010b). Membrane bound MCP with a cytoplasmic ligand binding region is a topology typically seen in bacterial aerotaxis receptors (Taylor, 2007). Some MCPs are exclusively composed of a signalling domain and lack any apparent ligand binding region (Fig. 2). The mode of signal recognition for this receptor topology remains yet to be established.

image

Figure 2. Diversity of chemoreceptors. Shown are the different receptor topologies. LBR: ligand binding region, MA: methylaccepting domain. Examples for each topology are given. Relevant references to these examples are: McpS (Lacal et al., 2010a), TlpB (Croxen et al., 2006); Aer (Taylor, 2007), DmcB (Li et al., 1999), Car (Storch et al., 1999), HemAT (Hou et al., 2000), Tm14 (Pollard et al., 2009), DifA (Yang et al., 1998). No characterized examples are available for topology 1b. Q8PLA8 and Q32FM8 are MCPs from Xanthomonas axonopodis pv. citri and Shigella dysenteriae serotype 1, respectively, with a topology 1b.

Download figure to PowerPoint

The MCP ligand binding regions are characterized by an enormous diversity in type and sequence. An analysis of MCPs by SMART has revealed that 88% of MCPs have a ligand binding region that remains un-annotated (Lacal et al., 2010b). It has been suggested that this is partially due to the imprecision of existing domain models (Ulrich and Zhulin, 2005). MCPs were found to use primarily PAS, GAF, TarH, CACHE and CHASE domains for sensing and, in addition, genome analyses have revealed there are MCPs that harbour HNOB (haem no binding), NIT (nitrate and nitrite sensing) and PrpR-N sensor domains (Wuichet et al., 2007). Furthermore, there is evidence that novel, yet uncharacterized types of ligand binding regions may exist (Lacal et al., 2010b).

In addition, Lacal and colleagues (2010b) have shown that ligand binding regions can be classified into two baseline resolved clusters according to their size. Cluster I receptors have a ligand binding region between 120 and 210 amino acids whereas cluster II receptors have larger ligand binding regions of 220–299 amino acids. Since there is evidence that suggests that some cluster II ligand binding regions are composed of two cluster I ligand binding regions, it was hypothesized that a physiological reason for the existence of some cluster II receptors may lie in their capacity to recognize multiple signals (Lacal et al., 2010b).

Direct and indirect signal recognition at the chemoreceptor

  1. Top of page
  2. Summary
  3. Introduction
  4. Motile bacteria differ largely in the number of chemoreceptors
  5. Different signals – different cellular compartments of sensing – different chemoreceptor sensor domains
  6. Direct and indirect signal recognition at the chemoreceptor
  7. Size differences in the chemoreceptor signalling and adaptation domain
  8. Effect of signal recognition on CheA autophosphorylation and receptor methylation
  9. Additional signal transduction proteins
  10. Difference in CheR-mediated adaptation
  11. Paralogues of signal transduction proteins – parallel signalling pathways
  12. Differences in the motor mechanism
  13. Signal termination
  14. Acknowledgements
  15. References

Ligands either bind directly to the MCP or via an additional soluble ligand binding protein. Both binding modes have been studied in some detail for the Tar receptor, that is able to respond to two different compounds at the same time, namely aspartate and maltose. Aspartate binds directly to the periplasmic ligand binding domain (Yen et al., 1996) whereas maltose binds to the periplasmic maltose binding protein (MBP). MBP in its closed, ligand-bound form interacts with Tar (Brass and Manson, 1984). Most interestingly, the chemotactic response to a mixture of aspartate and maltose is almost identical to the sum of the individual responses to both compounds (Mowbray and Koshland, 1987). This implies that Tar can mediate simultaneously a response towards both ligands. This view is supported by modelling the MBP-Tar complex which indicates that aspartate and MBP binding trigger similar conformational changes (Zhang et al., 1999).

In many cases the mode of ligand recognition, either direct or via an additional protein, is unknown. One such example is the recognition of Ni2+ by Tar that mediates a chemorepellent response. It has been proposed that the periplasmic binding protein NikA binds Ni2+ (de Pina et al., 1995). However, recent studies show that this is not the case, which is consistent with the notion that Ni2+ might bind directly to Tar (Englert et al., 2010). To determine the mode of ligand binding, Lacal and colleagues (2010a) have conducted isothermal titration calorimetry studies (Krell, 2008) of the recombinant ligand binding region of the McpS chemoreceptor from Pseudomonas putida KT2440. Using this approach it was shown that McpS- ligand binding region recognizes directly six different Krebs cycle intermediates and butyrate. However, very close structural homologues or derivatives of Krebs cycle intermediates, such as maleate, aspartate, itaconate or tricarballylate, were entirely devoid of binding. This approach has lead to a precise definition of the ligand profile in vitro, which was subsequently verified by in vivo studies using the wild-type strain and a mutant deficient in mcpS.

Size differences in the chemoreceptor signalling and adaptation domain

  1. Top of page
  2. Summary
  3. Introduction
  4. Motile bacteria differ largely in the number of chemoreceptors
  5. Different signals – different cellular compartments of sensing – different chemoreceptor sensor domains
  6. Direct and indirect signal recognition at the chemoreceptor
  7. Size differences in the chemoreceptor signalling and adaptation domain
  8. Effect of signal recognition on CheA autophosphorylation and receptor methylation
  9. Additional signal transduction proteins
  10. Difference in CheR-mediated adaptation
  11. Paralogues of signal transduction proteins – parallel signalling pathways
  12. Differences in the motor mechanism
  13. Signal termination
  14. Acknowledgements
  15. References

Methylaccepting chemotaxis proteins contain a cytoplasmic signalling and adaptation domain (MA) that interacts with CheA/CheW and which is the target for CheB and CheR. Structural information available on this domain reveals that it forms a long, dimeric α-hairpin (Park et al., 2006). Apart from the above mentioned size differences of ligand binding regions, the analysis of the totality of available MCP sequences revealed that the length of the signalling domain varies largely (Alexander and Zhulin, 2007). This diversity permits a domain classification into seven different groups. Small variants of this domain show 24 heptad repeats (168 amino acids) whereas the upper limit is formed by domains comprising 64 heptad repeats (448 amino acids). Independently of the domain length, both arms of this α-hairpin appear to maintain symmetry in length and it was suggested that signalling (lower part of this domain) and adaptation (upper part of this domain) have co-evolved throughout the natural history of chemotaxis.

Cryo-tomography images of 13 distantly related bacteria have provided experimental proof for this bioinformatic study. In all bacteria analysed, chemoreceptors group into arrays which can be recognized in these images as thin, pillar-like densities extending from the inner membrane to a prominent base plate formed by receptor-bound CheA and CheW (Briegel et al., 2009). This strongly suggests that the chemoreceptor assembly into arrays is a feature common to all bacteria. However, the distance between the CheA and CheW containing base plate to the inner membrane varied significantly between organisms, but was constant within each species. The chemoreceptors of a bacterial species belong primarily to one of the seven different groups mentioned above. Briegel and colleagues (2009) were able to establish a correlation between the measured distance of the base plate to the inner membrane and the predicted length of the prominent signalling domain of a given bacterial species. These results are also consistent with the notion that the entire cytosolic part of chemoreceptors is α-helical. However, the physiological relevance for the length differences of the signalling domain remains to be established.

Effect of signal recognition on CheA autophosphorylation and receptor methylation

  1. Top of page
  2. Summary
  3. Introduction
  4. Motile bacteria differ largely in the number of chemoreceptors
  5. Different signals – different cellular compartments of sensing – different chemoreceptor sensor domains
  6. Direct and indirect signal recognition at the chemoreceptor
  7. Size differences in the chemoreceptor signalling and adaptation domain
  8. Effect of signal recognition on CheA autophosphorylation and receptor methylation
  9. Additional signal transduction proteins
  10. Difference in CheR-mediated adaptation
  11. Paralogues of signal transduction proteins – parallel signalling pathways
  12. Differences in the motor mechanism
  13. Signal termination
  14. Acknowledgements
  15. References

Signal perception at an MCP has two major molecular consequences, which are the modulation of CheA autophosphorylation and MCP methylation. However, the direction of these alterations appears to differ among bacteria. In E. coli, positive stimuli decrease CheA activity whereas in Bacillus subtilis an increase in kinase activity is observed in response to positive stimuli (Borkovich et al., 1989; Garrity and Ordal, 1997). Differences between both organisms are also observed in the direction of the stimulus-mediated alteration of methylation: in E. coli positive stimuli increase methylation at all sites whereas negative stimuli reduce methylation at all sites. In B. subtilis the impact of stimuli is different since positive stimuli increase methylation of one amino acid but decreases methylation at another amino acid (Rao and Ordal, 2009). The same heterogeneity is observed for negative stimuli (Zimmer et al., 2000). Further studies will establish whether there is a dominant mechanism in bacteria.

Additional signal transduction proteins

  1. Top of page
  2. Summary
  3. Introduction
  4. Motile bacteria differ largely in the number of chemoreceptors
  5. Different signals – different cellular compartments of sensing – different chemoreceptor sensor domains
  6. Direct and indirect signal recognition at the chemoreceptor
  7. Size differences in the chemoreceptor signalling and adaptation domain
  8. Effect of signal recognition on CheA autophosphorylation and receptor methylation
  9. Additional signal transduction proteins
  10. Difference in CheR-mediated adaptation
  11. Paralogues of signal transduction proteins – parallel signalling pathways
  12. Differences in the motor mechanism
  13. Signal termination
  14. Acknowledgements
  15. References

The participants in the cytoplasmic signal transduction mechanism in E. coli are CheA, CheY, the coupling protein CheW, the methyltransferase CheR, the methylesterase CheB and the CheY phosphatase CheZ (Fig. 1). The study of chemotactic signalling in other organisms has revealed the presence of additional signalling proteins. In B. subtilis there is evidence for additional CheY-P phosphatases (CheC, FliY and CheX) (Muff and Ordal, 2008), the receptor deamidase CheD (Muff and Ordal, 2007) and the CheW homologue CheV (Szurmant and Ordal, 2004).

The discovery of additional proteins is associated with additional regulatory processes. Apart from the reversible receptor methylation, there is evidence for two additional adaptation mechanisms in which CheC, CheD and CheV are involved (Rao et al., 2008). CheD is a receptor deamidase that converts conserved glutamine residues to glutamates (Kristich and Ordal, 2002). In addition, CheD binds to CheC causing a modulation of the CheC phosphatase activity towards CheY-P. This was proposed to represent an alternative adaptation mechanism (Muff and Ordal, 2007). CheV has two domains, an N-terminal CheW domain and a C-terminal response regulator receiver domain that is phosphorylated by CheA. It was proposed that a phosphorylation of the response regulator receiver domain by CheA produces a conformational change that inhibits CheA kinase activity, probably by disrupting the interaction between the attractant-bound receptor and the kinase, which would correspond to a third adaptation mechanism (Rao et al., 2008).

Difference in CheR-mediated adaptation

  1. Top of page
  2. Summary
  3. Introduction
  4. Motile bacteria differ largely in the number of chemoreceptors
  5. Different signals – different cellular compartments of sensing – different chemoreceptor sensor domains
  6. Direct and indirect signal recognition at the chemoreceptor
  7. Size differences in the chemoreceptor signalling and adaptation domain
  8. Effect of signal recognition on CheA autophosphorylation and receptor methylation
  9. Additional signal transduction proteins
  10. Difference in CheR-mediated adaptation
  11. Paralogues of signal transduction proteins – parallel signalling pathways
  12. Differences in the motor mechanism
  13. Signal termination
  14. Acknowledgements
  15. References

Differences also exist among the chemoreceptor binding sites for CheR. Some chemoreceptors contain a flexible, C-terminal extension, which contains at its end the NWETF pentapeptide, which was found to be a high-affinity (2 µM) binding site for CheR (Wu et al., 1996). A crystal structure of CheR in complex with the NWETF has been determined (Djordjevic and Stock, 1998). Binding of CheR to this site is thought to hold CheR in close proximity to all of the methylation sites (Wu et al., 1996).

Sequence analysis of MCP genes from sequenced genomes revealed that only around 10% of receptors contain a potential C-terminal pentapeptide (Perez and Stock, 2007). Chemoreceptor TM1143 of Thermotoga maritima is an example of an MCP lacking the pentapeptide. Perez and Stock (2007) showed that CheR binds only very weakly to the recombinant signalling domain of TM1143. However, methylation studies using both proteins resulted in a KM value of close to 1 µM for the recombinant signalling domain (Perez and Stock, 2007). Further studies are necessary to understand the physiological relevance for the existence of receptors with and without pentapeptide.

There is also evidence for signalling proteins that contain additional domains of which its role in signal transduction yet needs to be identified. One such example is the CheR homologue in Myxococcus xanthus (Scott et al., 2008) that contains three tetra trico-peptide repeats (TPRs). TPRs are found in prokaryotes and eukaryotes and its biological function consists in establishing protein–protein contacts (D'Andrea and Regan, 2003). Scott and colleagues (2008) show that the TPRs present at CheR regulate site-specific methylation. However, the molecular basis for the action of CheR-TPR fusions remains to be established. A blast analysis reveals that CheR homologues in a large number of bacterial species contain one or several TPR repeats (Scott et al., 2008).

Paralogues of signal transduction proteins – parallel signalling pathways

  1. Top of page
  2. Summary
  3. Introduction
  4. Motile bacteria differ largely in the number of chemoreceptors
  5. Different signals – different cellular compartments of sensing – different chemoreceptor sensor domains
  6. Direct and indirect signal recognition at the chemoreceptor
  7. Size differences in the chemoreceptor signalling and adaptation domain
  8. Effect of signal recognition on CheA autophosphorylation and receptor methylation
  9. Additional signal transduction proteins
  10. Difference in CheR-mediated adaptation
  11. Paralogues of signal transduction proteins – parallel signalling pathways
  12. Differences in the motor mechanism
  13. Signal termination
  14. Acknowledgements
  15. References

Escherichia coli has a single cytoplasmic signalling cascade. However, the genome analysis of motile bacteria reveals the presence of multiple copies of cytoplasmic signal transduction proteins (Hamer et al., 2010). These authors analysed 206 species encoding at least one homologue of each of the five core chemotaxis proteins CheA, CheB, CheR, CheW and CheY. Interestingly, 61 species encode more than one of all of these five proteins, indicating an existence of multiple signalling pathways.

Studies of Rhodobacter sphaeroides (Porter et al., 2008a) have provided proof for the existence of multiple signalling cascades. This bacterium has three major operons (cheOp1-3) encoding homologues of signalling proteins and in addition a cheBRA locus. In addition, evidence was presented which demonstrated the existence of two different flagellar systems, termed fla1 and fla2 (del Campo et al., 2007). The observations of these authors indicate that genes encoded by cheOp1 control the activity of the fla2 system whereas protein of cheOp2 and cheOp3 regulate fla1 activity. The transmembrane chemoreceptors localized at the cell poles were found to interact with proteins encoded by cheOp2 whereas the cytosolic chemoreceptors cluster with proteins encoded by cheOp3 (Wadhams et al., 2003). Therefore, cytoplasmic and membrane-bound chemoreceptors form two separate signalling complexes. This enables the bacterium to sense cytoplasmic and extracellular signals independently. However, the signalling pathways associated with both clusters are likely to interact, since loss of any cheOp2 or cheOp3 signalling protein abolishes chemotaxis and it has been suggested that signals from both signalling pathways are necessary to generate a chemotactic response (Porter et al., 2002). However, the molecular bases for this mutual dependency of both signalling pathways need to be established.

The annotation of presumed chemotaxis genes is done on the basis of sequence similarity with genes, which were experimentally shown to be associated with taxis. There is also evidence that not all annotated chemotaxis gene clusters are involved in taxis. Signalling cascades to mediate other cellular processes might have evolved that use signalling proteins typically found in chemotactic signalling. One such example is Myxococcus xanthus (Zusman et al., 2007), which was found to have eight gene clusters containing proteins typically associated with taxis. Some of these clusters are involved in taxis whereas others can be associated with developmental processes leading to the formation of fruiting bodies. Therefore, CheA-mediated signalling can be considered in a broader sense as chemosensory signalling. Some of these chemosensory signalling cascades are involved in taxis whereas other might control additional cellular processes.

Differences in the motor mechanism

  1. Top of page
  2. Summary
  3. Introduction
  4. Motile bacteria differ largely in the number of chemoreceptors
  5. Different signals – different cellular compartments of sensing – different chemoreceptor sensor domains
  6. Direct and indirect signal recognition at the chemoreceptor
  7. Size differences in the chemoreceptor signalling and adaptation domain
  8. Effect of signal recognition on CheA autophosphorylation and receptor methylation
  9. Additional signal transduction proteins
  10. Difference in CheR-mediated adaptation
  11. Paralogues of signal transduction proteins – parallel signalling pathways
  12. Differences in the motor mechanism
  13. Signal termination
  14. Acknowledgements
  15. References

The flagellar motor is a complex assembly of a multitude of different proteins and represents the target for chemotactic signalling. CheY-P is able to bind to this motor thereby altering its activity. The function of the motor in E. coli and Salmonella typhimurium is based on a switch between the counterclockwise rotation that propel the cell smoothly and clockwise rotation that leads to a change in swimming direction called a tumble (Sowa and Berry, 2008).

There is evidence that motors of other bacteria operate according to a different mechanism. Rhodobacter sphaeroides employs a unidirectional motor and reorientation of swimming cells appears to be the result of Brownian rotation during the stop periods (Armitage and Macnab, 1987). It was shown that the motor rotates counterclockwise at constant speed. Chemotaxis signalling does not control motor speed but CheY-P was found to stop the motor (Pilizota et al., 2009). Sinorhizobium meliloti uses also a unidirectional motor, but the reorientation of the swimming path is not achieved by modulating stop periods, but is mediated by a variation of the speed of flagellar rotation (Attmannspacher et al., 2005). An alternative mechanism is that of Vibrio alginolyticus, that pushes or pulls the cell depending on rotation direction (Sowa and Berry, 2008). The physiological reasons for these differences in motor mechanism remain unclear.

Further differences among flagellar motors exist at the level of the energizing compound. Bacterial motors are driven by electrochemical gradients. Most motors like those of E. coli, S. typhimurium, B. subtilis, R. sphaeroides and P. aeruginosa have H+-driven motors and activity is achieved by a flux of H+ across the cytoplasmic membrane (Terashima et al., 2008). In contrast, Vibrio alginolyticus uses a motor, which is driven by a Na+ gradient. In contrast to the above species V. alginolyticus is a marine bacterium and is therefore exposed to a high salt concentration. The presence of high salt concentrations was proposed to be the reason for the evolution of the Na+-driven motor (Yorimitsu and Homma, 2001).

In archaea the proteins involved in chemotactic signalling are homologous to those in bacteria. For example the best-studied cytosolic chemoreceptor is from the archeon Halobacterium salinarium, which was found to mediate chemotaxis specifically towards arginine (Storch et al., 1999). In contrast to the conserved signalling cascades in bacteria and archaea, the corresponding motors are entirely different and have to be considered as functional analogues. Proteins that compose the motor in bacteria and archaea do not share sequence homology (Ng et al., 2006) and can thus be considered as the result of convergent evolution. The lacking sequence homology between motor proteins of bacteria and archaea is also reflected in a different mechanism of action. As stated above, bacterial motors are primarily driven by a proton-motive force, whereas archaeal motors are energized either by ATP hydrolysis directly or by an ATP-dependent ion gradient, which is not coupled to a proton-motive force (Streif et al., 2008). In addition, in some respects archaeal flagella resemble more bacterial type IV pili than bacterial flagella (Cohen-Krausz and Trachtenberg, 2002).

Signal termination

  1. Top of page
  2. Summary
  3. Introduction
  4. Motile bacteria differ largely in the number of chemoreceptors
  5. Different signals – different cellular compartments of sensing – different chemoreceptor sensor domains
  6. Direct and indirect signal recognition at the chemoreceptor
  7. Size differences in the chemoreceptor signalling and adaptation domain
  8. Effect of signal recognition on CheA autophosphorylation and receptor methylation
  9. Additional signal transduction proteins
  10. Difference in CheR-mediated adaptation
  11. Paralogues of signal transduction proteins – parallel signalling pathways
  12. Differences in the motor mechanism
  13. Signal termination
  14. Acknowledgements
  15. References

The majority of response regulators are involved in the regulation of gene expression. In this context, the time that passes from the removal or reduction of the input signal to the termination of the regulatory response is not critical and determined by the intrinsic dephosphorylation of response regulators, which have typically half-lifes in the minute range (Goudreau et al., 1999). This is different in the context of taxis and changes in attractant/repellent concentration have to impact almost immediately on motor function, and a lag of response caused by the relatively slow dephosphorylation of CheY cannot be tolerated. Among different response regulators CheY has evolved to undergo rapid auto-dephosphorylation, nevertheless its estimated half-time of 6 s (Sanders et al., 1989) is not sufficient to guarantee an optimal response to changes in attractant concentrations. Therefore, different mechanisms have evolved to dephosphorylate CheY-P causing rapid signal termination.

As mentioned above CheY dephosphorylation in E. coli is achieved by the CheZ phosphatase. It has long been debated whether CheZ is an allosteric enhancer of the intrinsic dephosphorylation of CheY or whether it actively participates in the removal of the CheY dephosphorylation. The inspection of the three-dimensional co-crystal structure of the CheY/CheZ complex (Zhao et al., 2002) supports the latter hypothesis since CheZ residues were found in positions that indicate their direct role in dephosphorylation catalysis.

Apart from the CheZ phosphatase in E. coli, CheY dephosphorylation in a number of other species is accomplished by members of the so-called CheC type phosphatases, which can be subdivided into three families, namely CheC, FliY and CheX (Muff and Ordal, 2008). Members of these families differ in length and domain composition but share a conserved sequence pattern that defines the phosphatase active site. At the sequence level members of the CheC family are unrelated to CheZ and have evolved convergently to dephosphorylate the same target. CheZ is limited to proteobacteria but CheC family members are much more widespread and also present in Firmicutes and Thermotogales (Muff and Ordal, 2008). Some members of the CheC family are not encoded in chemotaxis gene clusters and are therefore thought to be involved in non-chemotactic signalling.

Apart from CheZ and CheC phosphatases, evidence for the involvement of another protein family in CheY dephosphorylation has been obtained. Rhodobacter sphaeroides lacks homologues of CheZ and CheC phosphatases and Porter and colleagues (2008b) have revealed that a CheA paralogue acts as phosphodonor as well as specific CheY phosphatase. A CheA mutant in which the domain responsible for the phosphatase activity was deleted did not support chemotaxis. The use of sensor kinase-mediated phosphatase activity is a mode of signal termination not only found in chemotactic signalling but also widespread among TCSs (Krell et al., 2010).

The mechanisms discussed above are all based on phosphatase activity towards CheY. In the context of signal termination an alternative strategy has evolved that is based on CheY homologues which function as phosphate sink. Sinorhizobium melilotti has two CheY homologues: CheY2 is the principal homologue involved in taxis whereas CheY1 has a role in signal termination. Sourjik and Schmitt (1996) have shown that phophorylgroups from surplus CheY2 can be transferred back to CheA and then subsequently to CheY1. Since phosphorylated CheY1 does not mediate taxis and since its concentration is well above that of CheA, CheY1 can be considered as a phosphate sink and its phosphorylation by CheA was shown to accelerate dephosphorylation of the functionally active CheY2. Studies of the chemotactic system in Helicobacter pylori (Jiménez-Pearson et al., 2005) and Myxococcus xanthus (Black et al., 2010) have also provided evidence for CheY homologues that function as phosphate sink.

The use of response regulator homologues as phosphate sinks appears to be a more widespread mechanism in TCSs signalling. For example the TCS FixL/FixJ controls the expression of nitrogen fixation genes in nitrogen-fixing bacteria and its activity is inhibited by the FixT protein. Data have been presented that indicate that FixT is a phosphate sink response regulator which competes with the cognate response regulator FixJ for phosphoryl groups at the FixL sensor kinase (Crosson et al., 2005).

The study of chemotactic signalling in a wide range of bacteria over the last two decades has revealed an enormous diversity at the level of the molecular mechanisms as well as the abundance of chemotactic signalling proteins in bacterial species. A part of these differences can be explained by differences in the nature of the corresponding input signal. However, the physiological reasons for the majority of differences in chemotactic signalling remains poorly understood and its elucidation represents a primary research need in this field.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Motile bacteria differ largely in the number of chemoreceptors
  5. Different signals – different cellular compartments of sensing – different chemoreceptor sensor domains
  6. Direct and indirect signal recognition at the chemoreceptor
  7. Size differences in the chemoreceptor signalling and adaptation domain
  8. Effect of signal recognition on CheA autophosphorylation and receptor methylation
  9. Additional signal transduction proteins
  10. Difference in CheR-mediated adaptation
  11. Paralogues of signal transduction proteins – parallel signalling pathways
  12. Differences in the motor mechanism
  13. Signal termination
  14. Acknowledgements
  15. References