Figure 1.
The Chemotaxis Pathways in E. coli and B. subtilis
(A) E. coli. (B) B. subtilis. Both organisms respond to extracellular signals by regulating the activity of the CheA histidine kinase. CheA is coupled to transmembrane receptors (MCP) by an adaptor protein CheW. Chemoattractants, by binding the receptor, inhibit CheA in E. coli (red line) (Borkovich et al. 1989) and stimulate CheA in B. subtilis (green line) (Garrity and Ordal 1997). CheA phosphorylates CheY. Phosphorylated CheY binds to the flagellar motor and increases the frequency of tumbles in E. coli (Cluzel et al. 2000) and runs in B. subtilis (Bischoff et al. 1993). Phosphorylated CheY is also predicted to inhibit the receptor complex in B. subtilis (dashed line). Both organisms tune the sensitivity of CheA to ligands by reversibly methylating the receptors using the CheR methytransferase and CheB methylesterase (Zimmer et al. 2000; Sourjik and Berg 2002b). Phosphorylation of CheB by CheA increases its methylesterase activity nearly 100-fold (Anand and Stock 2002). CheA activity is proportional to the degree of receptor methylation in E. coli. In B. subtilis, CheA activity depends on which residue is methylated, akin to a binary switch. E. coli possesses a phosphatase, CheZ, not present in B. subtilis, that enhances the rate of CheY dephosphorylation. B. subtilis possesses three chemotaxis proteins not found in E. coli: CheC, CheD, and CheV. CheC is a negative regulator of receptor methylation and homologous to the CheY-binding domain (P2) in CheA (Rosario et al. 1995; Rosario and Ordal 1996). CheD is a positive regulator of receptor methylation and also deamidates specific residues on the receptor (Kristich and Ordal 2002). CheV is a CheW-response regulator fusion. CheV is functionally redundant to CheW and is predicted to negatively regulate receptor activity (dashed line) (Rosario et al. 1994; Karatan et al. 2001).
Table 1.
Summary of Differences between E. coli and B. subtilis Chemotaxis
Figure 2.
Model for the Phosphorylation Cascade in B. subtilis
The model assumes that the receptor complex (receptor, CheA, CheC, CheD, and CheW) exists either in an active (TA) or inactive (TI) state. Active receptors stimulate CheA. CheA phosphorylates CheB, CheV, and CheY. Phosphorylated CheY (Yp) binds the receptor and increases the likelihood a receptor adopts an inactive conformation (thick red line). Phosphorylated CheY also binds the flagellar motor (M). The motor switch enhances the rate of CheY dephosphorylation (Szurmant et al. 2003). The model assumes that unphosphorylated CheV inhibits CheA by disrupting the receptor complex (thick blue line).
Figure 3.
Model for Selective Methylation in B. subtilis
The model assumes that the receptor dimers exist in six different methylation states. The different methylation states are denoted by the variable Tij, where the index i denotes the methylation state of residue 630 and j denotes the state of residue 637. For example, T20 denotes the concentration of dimers with both residues methylated at position 630 and none at position 637. For simplicity, the model assumes that at most two residues are methylated as additional states are superfluous. When receptors are methylated at residue 630, the signaling complex preferentially adopts an active conformation. When residue 637 is methylated, the signaling complex preferentially adopts an inactive conformation. When the dimers are partially methylated, the strength of activation or inhibition is attenuated. Selective methylation is coordinated by phosphorylated CheY (Yp). CheR methylates residue 637 when phosphorylated CheY is bound to the receptor and methylates residue 630 otherwise.
Figure 4.
Simulation of Adaptation in E. coli and B. subtilis
Attractant (10 μM) is added at 500 s and removed at 1,000 s.
(A) Timecourse simulation of phosphorylated CheY (left) and receptor methylation (right) in E. coli.
(B) Timecourse simulation of phosphorylated CheY (left) and receptor methylation (right) in B. subtilis. In both species, adaptation correlates with changes in receptor methylation.
Figure 5.
Graphical Illustration of Mechanism for Robust Adaptation
(A) Qualitative relationship among receptor activity, methylation, and demethylation in E. coli. The rate of demethylation is proportional to the number of active receptors, and the rate of methylation is inversely proportional to the number of active receptors. The system reaches steady state only when the two solid lines cross. As the rate of methylation decreases monotonically with receptor activity and the rate of demethylation increases monotonically with receptor activity, only one steady state is possible (A*) if the rates depend solely on receptor activity. The kinetic parameters change the slope of the curves, but not their monotonicity. Hence, adaptation is robust with respect to changes in the kinetic parameters. However, the point where they intersect does change with the parameters.
(B) Qualitative relationship between receptor activity and the differential rate of methylation in B. subtilis. The net rate of methylation at residue 630 decreases monotonically with receptor activity, and the net rate of methylation at residue 637 increases monotonically with receptor activity. By the same arguments, only one steady state (A*) is possible and, hence, adaptation is robust in B. subtilis.
Figure 6.
Sensitivity to Parameters in E. coli and B. subtilis
(A) E. coli. (B) B. subtilis. The top figures are plots of the steady-state concentration of phosphorylated CheY as a function of CheB and CheR concentrations. The bottom figures are plots of the adaptation time as a function of CheB and CheR concentrations. Adaptation time is defined as the length of time from the peak concentration in phosphorylated CheY (Yp) to within 5% of the steady-state concentration after the addition of attractant (10 μM). For all the concentrations considered, both models precisely adapt.
Figure 7.
Oscillations and Methylation-Independent Chemotaxis
(A) Timecourse simulation of cheBCDR strain in B. subtilis subject to the addition of attractants (100 μM) at 200 s and the removal at 500 s. Concentration of CheV was set at 8 nM.
(B) Timecourse simulation of the cheBCDR strain in B. subtilis subject to the addition of attractants (100 μM) at 200 s and the removal at 500 s, where the concentration of CheV is halved (4 nM).
Figure 8.
CheY Feedback Is Sufficient for Precise Adaptation
Timecourse simulation of model subject to the addition of attractants (10 μM) at 200 s and removal at 500 s. The model is described in Materials and Methods.
Table 2.
Distribution of Chemotaxis-Like Genes and Number of Paralogs for a Representative Set of Microbial Organisms