There is nothing like computer animation for visualizing complex cellular processes. This animation reveals the two-component signalling mechanism by which bacteria control flagellar directional switching in response to chemical stimuli (attractants or repellents). Directional switching from counterclockwise to clockwise rotation results in the flagellar bundle breaking apart, causing the bacterium to tumble in place, thereby changing direction.

Bear in mind though that this mechanism varies from bacterium to bacterium, and what is necessary in one system may be dispensable in another. Null mutants of specific chemotaxis genes can also cause different phenotypes in different bacterial species. A comparative study conducted in 2004 of chemotaxis systems in Escherichia coli and Bacillus subtilis reported:

E. coli and B. subtilis bias their motion towards favorable conditions with nearly identical behavior by adjusting the frequency of straight runs and reorienting tumbles. Both pathways share five orthologous proteins with apparently identical biochemistry. How these individual orthologs contribute to the overall function, however, is different, as illustrated when synonymous orthologs are deleted in each organism. Deletion of the CheY response regulator causes E. coli to run exclusively and B. subtilis to tumble exclusively. When the CheR methyltransferase is deleted in E. coli, the cells are incapable of tumbles and only run.

Likewise, when the CheB methylesterase is deleted, E. coli cells are incapable of runs and only tumble. In B. subtilis, cells still run and tumble when either CheB or CheR is deleted, though they no longer precisely adapt. Remarkably, both genes complement in the heterologous host. Deletion of the CheW adaptor protein in E. coli results in a run-only phenotype, whereas there is no change in phenotype for the synonymous deletion in B. subtilis. When the genes involved in regulating methylation are deleted (cheBR in E. coli and cheBCDR in B. subtilis), E. coli does not adapt, whereas B. subtilis either oscillates or partially adapts when exposed to attractants. These differences demonstrate that the pathways are different even though they involve homologous proteins, [internal citations omitted].

Be sure to check out my previous discussion of bacterial chemotaxis and signal transduction, here.