Protein rings are critical to the remarkable signaling properties of bacterial chemotaxis nanoarrays

Abstract

Bacteria build an extensive sensory nanoarray to collect information that guides their swimming. In this issue of Science Signaling, Piñas et al. demonstrate that a key element of these arrays that enhances chemotaxis responses are hexameric rings of CheW, one of two types of rings that couple the responses of core signaling units to achieve remarkable signaling properties such as single molecule detection.

Bacterial chemotaxis proteins enable bacteria to sense temporal gradients of various molecules and use this information to direct their swimming towards more favorable environments, for example towards nutrients such as amino acids and sugars. Although chemotaxis is one of the best-understood signaling systems in biology, the mechanistic basis for several impressive signaling properties remains elusive. How does Escherichia coli chemotaxis achieve the sensitivity to detect 1-2 molecules per cell and the dynamic range to respond to gradients over 5 orders of magnitude (1)? A key element of the answer is that sensing occurs within a remarkable membrane-bound nanoarray of proteins. Maddock and Shapiro (2) made the unexpected discovery that chemoreceptors are clustered at the poles of E. coli cells. Cryoelectron microscopy of arrays in bacteria combined with x-ray crystallographic models of protein components and complexes yielded structural models of the nanoarray (3), a hexagonal architecture that is conserved throughout bacteria and archaea. The transmembrane chemoreceptors and the intracellular kinase CheA have clear functions in the array: the receptors bind ligands in the periplasm and transmit this information to control the activity of the CheA associated with their cytoplasmic tips. The third protein component of the array, CheW, has been called an adapter, linker, or coupling protein, as its role in signaling is less clear. In this issue, a study from the Parkinson lab (Piñas et al.) combines clever biochemistry with in vivo measurements of signaling to demonstrate that CheW coupling of responses across this sensor array is critical to efficient chemotaxis (4).

Chemotaxis signaling arrays (Fig. 1) are organized as hexagonal rings of receptor trimers of dimers that are connected at their cytoplasmic tips by rings of alternating CheA and CheW molecules. The CheA dimer bridges two adjacent CheA-CheW rings, creating a hexagon of CheA-CheW rings surrounding a space that cannot contain CheA and was proposed to contain CheW (5). Thus far, such CheW rings have been experimentally observed only for in vitro preparations (6), leaving the open question of whether CheW rings occur in native arrays in bacteria (Fig. 1A). Piñas et al. (4) generated double Cys mutants of CheW designed to probe the interface between members of CheW rings suggested by structural models and screened these with in vivo disulfide crosslinking. They discovered several pairs that formed multimeric crosslinked products with a maximum of hexamers, consistent with the presence of closed rings of 6 CheW molecules in arrays in E. coli (Fig. 1B). They chose the Cys pair with the best-resolved multimer bands on gels as a reporter to further investigate the requirements for and consequences of CheW rings (4).

Fig. 1. Chemotaxis sensor array, viewed from inside the bacterium and perpendicular to the membrane.

(A) A core signaling unit (black square) contains a pair of receptors trimers of dimers (gray) bridged by a CheA dimer (yellow) along with a pair of CheW molecules (blue). Three core units are linked by hexagonal rings of alternating CheA and CheW, leaving an open space lined by receptor trimers of dimers that may contain a CheW ring (question mark). (B) Piñas et al. demonstrated that E. coli chemoreceptor arrays contain hexameric CheW rings that enhance signaling cooperativity and sensitivity by functionally coupling core units (arrows) through both the central CheW ring and the surrounding CheA-CheW rings. (C) Even more CheW rings formed in the absence of CheA, suggesting the formation of a simpler array composed of only receptors and CheW. The fact that CheW alone can support formation of the array lattice may be important for other bacteria that produce low amounts of CheA to form sparse-CheA arrays and mount cooperative responses (4).

At a structural level, Piñas et al. (4) definitively show that wild-type E. coli arrays contain CheW rings and demonstrate several effects of altering the array component proteins. CheW ring formation required the presence of receptor trimers of dimers, but did not require CheA. CheW ring formation was not required for the formation of native CheA-receptor and CheA-CheW contacts. Furthermore, they developed a clever means of manipulating the array content, by using a CheW~CheA fusion protein to assemble functional arrays in the absence of CheW rings. These experiments all suggested that functional arrays can form without CheW rings. These results also provide useful tools for controlling and simplifying the array architecture to form either arrays containing no CheW rings (Fig. 1A) or arrays containing no CheA-CheW rings (Fig. 1C). These will be useful for future structural and mechanistic studies.

The beauty of studying signaling in bacteria is that one can perturb the chemotaxis system and then measure in vivo effects on the entire signaling pathway, both the molecular responses to stimuli and the overall chemotaxis behavior. Sourjik and Berg developed a powerful in vivo assay that monitors fluorescence resonance energy transfer (FRET) between two chemotaxis proteins downstream in the phosphorylation cascade to measure both the sensitivity and cooperativity of the response to receptor ligands. They used this assay for the first in vivo measurements of the cooperativity of the chemotaxis response, with Hill coefficients as high as 10 at native protein concentrations (7). Parkinson and co-workers have frequently combined this FRET assay with mutagenesis and crosslinking studies to gain insight into the physiological relevance of various interactions in the nanoarrays. For example, using this approach Piñas et al. previously demonstrated that CheW contacts within the alternating CheA-CheW rings are critical to both polar receptor clustering and to the cooperativity of the response to receptor ligands (8).

In the current study, Parkinson and coworkers used the FRET assay of in vivo signaling to demonstrate that CheW rings improve chemotaxis by increasing either the cooperativity (for arrays with high kinase activity) or the sensitivity (for arrays with low kinase activity) of the response (4). These results are consistent with previously observed effects of CheW amounts that were explained using a Monod–Wyman–Changeux (MWC) allosteric model (7) and other mathematical modeling (reviewed in (9)).

It is now clear that CheW is essential for both structural and functional coupling within the chemotaxis nanoarray (Fig. 1B): Piñas et al. previously demonstrated that CheA-CheW contacts in the CheA-CheW ring enable cooperativity between core signaling units (8) and have now shown that CheW rings augment that cooperativity (4) to achieve efficient sensing in bacterial chemotaxis. Finally, signaling within the core unit (Fig. 1A) is widely thought to involve extensive changes in receptor dynamics, and there are hints that CheW may transmit the signal from the receptor to CheA, thus playing a coupling role within the core units as well. It remains to be seen whether functional coupling through CheW within and between core units is a simple consequence of the stabilization of the protein network (10) or involves active propagation of a conformational change.