Activated chemoreceptor arrays remain intact and hexagonally packed

Abstract

Bacterial chemoreceptors cluster into exquisitively sensitive, tunable, highly ordered, polar arrays. While these arrays serve as paradigms of cell signalling in general, it remains unclear what conformational changes transduce signals from the periplasmic tips, where attractants and repellents bind, to the cytoplasmic signalling domains. Conflicting reports support and contest the hypothesis that activation causes large changes in the packing arrangement of the arrays, up to and including their complete disassembly. Using electron cryotomography, here we show that in Caulobacter crescentus, chemoreceptor arrays in cells grown in different media and immediately after exposure to the attractant galactose all exhibit the same 12 nm hexagonal packing arrangement, array size and other structural parameters. ΔcheB and ΔcheR mutants mimicking attractant- or repellent-bound states prior to adaptation also show the same lattice structure. We conclude that signal transduction and amplification must be accomplished through only small, nanoscale conformational changes.

Figure 1

Twelve-nanometre hexagonal arrangement of receptors is preserved upon activation. Tomographic slices (left column), enlarged to highlight the organization of the chemoreceptor arrays (middle column), and corresponding power spectra (right column) revealing 12 nm hexagonal order seen in a wild-type swarmer cell grown in M2G minimal media (A–C), a wild-type cell frozen ~7 s after exposure to 10 mM galactose (D–F), a ΔcheR mutant cell mimicking a fully attractant-bound state prior to adaptation (G–I), and a ΔcheB mutant cell mimicking a fully repellent-bound state prior to adaptation (J–L). Bars (A, D, G, J): 100 nm, Bars (B, E, H, K): 25 nm. Power-spectra not to scale.

Figure 2

Swim assays. A. Wild-type C. crescentus was stabbed at the indicated locations (asterisks) into 0.126% agar M2G media plates without glucose. A filter paper soaked with glucose-free M2G media supplemented with 1 M galactose was placed in the centre of the plate. After 72 h at room temperature, the swim colonies showed a clear bias towards the galactose source, confirming that galactose is an attractant for this organism. B. Cultures of C. crescentus wild-type, ΔcheR and ΔcheB cells were spotted onto 0.25% agar plates made with glucose-free M2G media. A filter paper soaked with glucose-free M2G media supplemented with 1 M galactose was placed in the centre of the plate. After 7-day incubation at room temperature, both the ΔcheR and the ΔcheB mutants failed to form a swarm on the semisolid plate and did not exhibit any bias towards the galactose source, confirming their loss of chemotaxis.

Figure 3

Swimming behaviour of the ΔcheB and ΔcheR mutants. A histogram of tumble rate (defined as a change in direction by more than 45°) is shown for the wild-type and mutant strains (A) above sample trajectories of different cells (colours) in movies of swimming cells (B). The frequent changes in swimming direction (‘tumbles’) observed in the ΔcheB mutant mimic the behaviour of cells sensing repellent prior to adaptation, whereas the smooth trajectories of the ΔcheR strain mimic cells sensing attractant prior to adaptation. Trajectories from 80 ΔcheB, 61 ΔcheR and 72 wild-type cells were analysed for the histogram; the trajectories in (B) represent a subset of these.

Figure 4

Side views of subvolume averages of the mutant strains. Receptors are seen as a series of regularly spaced densities in the CheA/CheW base plate (white arrows) and extending towards the inner membrane for ~18 nm (brackets). The absence of densities in the remaining ~13 nm towards the inner membrane (IM), as well as in the periplasmic region, indicates a lack of regular order. This region presumably includes the two HAMP and linker regions characteristic of C. crescentus MCPs as well as the transmembrane and periplasmic domains. Receptor densities emerge at an acute angle to the base plate, as expected for trimers-of-dimers. Note that the quality of both averages is not identical due to slight quality differences in the tomograms and the different number of subvolumes averaged (100 ΔcheB and 50 ΔcheR subvolumes). Bar: 25 nm. Braces indicate the ordered region of the receptors.