Signal integration in chemoreceptor complexes

Abstract

Motile bacteria use large receptor arrays to detect chemical and physical stimuli in their environment, process this complex information, and accordingly bias their swimming in a direction they deem favorable. The chemoreceptor molecules form tripod-like trimers of receptor dimers through direct contacts between their cytoplasmic tips. A pair of trimers, together with a dedicated kinase enzyme, form a core signaling complex. Hundreds of core complexes network to form extended arrays. While considerable progress has been made in revealing the hierarchical structure of the array, the molecular properties underlying signal processing in these structures remain largely unclear. Here we analyzed the signaling properties of nonnetworked core complexes in live cells by following both conformational and kinase control responses to attractant stimuli and to output-biasing lesions at various locations in the receptor molecule. Contrary to the prevailing view that individual receptors are binary two-state devices, we demonstrate that conformational coupling between the ligand binding and the kinase-control receptor domains is, in fact, only moderate. In addition, we demonstrate communication between neighboring receptors through their trimer-contact domains that biases them to adopt similar signaling states. Taken together, these data suggest a view of signaling in receptor trimers that allows significant signal integration to occur within individual core complexes.

Keywords: cell signaling; chemotaxis; receptor array; signal integration.

Fig. 1.

The receptor core complex and two assays for monitoring its ligand response. (A) Schematic of the core complex, which consists of two receptor trimers of dimers, one CheA dimer, and two protomers of the CheW scaffolding protein. Binding of an attractant ligand to the periplasmic sensing domain of a receptor initiates conformational changes that propagate to the cytoplasmic tip region to modulate the CheA autokinase activity, which we followed with an in vivo CheY/CheZ FRET assay (Materials and Methods). The top–down cross-section view of the tip region of the core complex (on the Right) shows the interface 1 interactions (black circles) between CheW and the P5 domain of CheA that are critical to core complex assembly and operation. A second CheW-CheA.P5 interaction, interface 2 (red circles), networks core units into an extended signaling array. The CheW interface 2 lesion (CheW-X2) prevents array assembly and permits study of signaling events in nonnetworked core units. (B) Signaling architecture of the receptor dimer and the positions of locked-ON Tsr lesions used here. (C) The fluorescence anisotropy (homo-FRET) assay of signaling responses (Materials and Methods). The mYFP-tag is positioned at the C termini of the receptor protomers. Ligand binding induces conformational changes in the complex that modulate packing of the fluorescent tags and their homo-FRET interactions, leading to changes in the polarization (anisotropy) of their emitted light upon excitation with plane polarized light.

Fig. 2.

Signaling behaviors of mutant Tsr and Tar receptors. (A) Receptor-controlled kinase responses (Fig. 1A and Materials and Methods). (Left) Fluorescence-ratio traces (CheY-mCherry/CheZ-mYFP) of VF7 cells (CheW-X2) expressing Tsr (4Q) (Inset), with methyl-mimicking Q residues at adaptation sites 1 to 4, or Tsr (5Q) [Tsr (4Q) carrying the ON-shifting E502Q lesion]. Cells were challenged with serine (SER; 10 mM) and NaCN (3 mM) during the periods indicated by the horizontal bars. (Right) For each receptor variant, the kinase activity in the presence of ligand (10 mM serine for Tsr or 10 mM MeAsp for Tar) is plotted and normalized by the corresponding no-ligand activity. Error bars represent additional data shown in SI Appendix, Fig. S1. Receptor variants were tagged with mCFP, which does not interfere with the FRET measurements (see text). (B) Ligand-induced conformational (anisotropy) responses (Fig. 1C and Materials and Methods). (Left) Two examples of anisotropy (r) responses in VF7 cells carrying mYFP-tagged parental receptors or their mutant derivatives: Tsr-mYFP (4Q) vs. (5Q) (main plot) and Tsr-mYFP [QEQE] vs. A413G (Inset). Cells were challenged with serine (10 mM) during the periods indicated by the horizontal bars. To allow for direct comparison, the entire traces of the mutant receptors were shifted upward (SI Appendix, Fig. S2). The cells also carried an empty pKG110 vector to control for the pKG110 derivative that supplied the FRET reporter pair in (A). (Right) Summary of ligand-induced anisotropy responses. For each receptor mutant, the response amplitude was normalized by the corresponding response amplitude of the parental receptors. Error bars represent additional data shown in SI Appendix, Fig. S2.

Fig. 3.

Signaling interactions between Tar and Tsr receptors in mixed core complexes. (A) Dose-dependent anisotropy responses to serine measured in VF7 cells expressing Tsr-mYFP [QEEE] (1Q) (Left column) or Tsr-mYFP (4Q) (Right column), alone (Top row) or in combination with Tar (Middle row), or Tar plus 0.5 mM MeAsp (Bottom row). At least three and up to seven repetitions were done for each plot at different intermediate concentrations. Each dose–response plot was normalized to its maximal response, which depended on the mixing. Tsr expression was induced at 1.25 µM IPTG; Tar expression at 0.3 to 0.4 µM (4Q) or 0.75 µM (4E) NaSal. The thicker lines on each side are identical in all three plots (Top to Bottom) and serve as a reference. (B) Dose-dependent anisotropy responses measured as in (A). (Left) Tsr-mYFP/A413T (locked-OFF; see SI Appendix, Fig. S1) with (filled circles) or without (open circles) coexpressed Tar (4Q) receptors. (Right) Tsr-mYFP/A413G (locked-ON; see Fig. 2 and SI Appendix, Fig. S1) coexpressed with Tar (4E) with (filled circles) or without (open circles) MeAsp.

Fig. 4.

Transmission of Tar ligand-binding signals to neighboring Tsr receptors. (A) Anisotropy responses measured in VF7 cells expressing Tsr-mYFP (4E) induced at 1.25 µM IPTG, either alone (black traces) or coexpressed with Tar (4E) or Tar (4Q), induced at 0.35 µM NaSal (dark gray traces) or 0.75 µM NaSal (light gray traces). (B) Anisotropy responses measured in VF7 cells expressing Tsr-mYFP (4Q) induced at 1.25 µM IPTG, either alone (black traces) or coexpressed with Tar (4E) or Tar (4Q), induced with 0.75 µM NaSal (light gray traces). Estimated Tar expression levels are shown in SI Appendix, Fig. S4A. Ligand responses were measured to serine (1 mM), which binds to the Tsr-mYFP receptors, or to MeAsp (1 mM), which binds to the untagged Tar receptors, as labeled. The dependence of the Tar-mediated indirect (MeAsp) responses of Tsr-mYFP (arrows) on Tar induction is shown in SI Appendix, Fig. S4B. Mechanistic interpretations of the opposing indirect responses in panels A and B (upward vs. downward arrows) based on tip interaction effects are shown on the Right (see text).

Fig. 5.

The effect of locked-ON/OFF tip lesions on the indirect conformational responses in mixed complexes. Anisotropy responses measured in VF7 cells coexpressing Tsr-mYFP (induced at 1.25 µM IPTG) and untagged Tar (induced at 0.75 µM NaSal). Upper traces: Tsr-mYFP (QEQE) carried a locked-OFF (A413T) or locked-ON (A413G) tip lesion in combination with Tar (4E). Lower traces: Tsr-mYFP (4Q) in combination with Tar (4E) or Tar (QEQE), each bearing the locked-ON (A411G) tip lesion. The downward arrow indicates an indirect “inverted” conformational response. See SI Appendix, Fig. S5 for additional data.

Fig. 6.

Anisotropy and kinase-activity dose–responses in Tsr core complexes. Conformational (anisotropy) dose–responses were measured in VF7 cells expressing Tsr-mYFP in the (4E), (QEQE) and (4Q) modification states. Cells also harbored an empty pKG110 vector. Kinase activity (CheY-CheZ FRET) responses were measured in VF7 cells expressing Tsr-mCFP in (4E), (QEQE) and (4Q) modification states. Cells also expressed CheY-Z FRET pair from a pKG110 vector. The Upper-Right panel shows the relative response amplitudes at the three modification states. Thin lines define the range of potential fits at each modification state and the thick lines correspond to their average. The parameters for the average fits are (K_1/2/Hill coefficient): Anisotropy: 4E - 3.5/1.28, QEQE - 22.5/1.44, 4Q - 112.5/1.2; Kinase activity: 4E - 4.5/1.825, QEQE - 36/1.875, 4Q - 220/1.7. The Lower Right panel shows a comparison between the averaged kinase (thicker lines) and anisotropy (thinner lines) responses.

Fig. 7.

A revised model for signaling in chemoreceptor trimers. (A) The model proposes that ligand-binding induces a primary response in the upper portion of receptor dimers that propagates to the receptor tip to control kinase output (see text for details). The energy biases of the primary (ΔE_primary) and tip (ΔE_tip) responses are different and have a finite coupling energy (J_intra). In addition, neighboring receptor tips have a coupling energy (J_tip) that biases them to adopt similar signaling states. The ligand-dependent ON probabilities of the “primary” and “tip” regions, P_primary(L) and P_tip(L), respectively, were calculated as detailed in SI Appendix, SI Model section. The Tsr parameter values used here were derived from fitting the data (SI Appendix, Fig. S6A). (B) The inverted ON-probabilities 1-P_primary(L) (dashed lines) and 1-P_tip(L) (solid lines) are plotted for trimers with either QEQE or 4Q Tsr receptors. These plots demonstrate that the two regions can exhibit distinct response behaviors. (C) 1-P_primary(L) is plotted for Tsr (4E) trimers (green dotted lines) or similar trimers in which one receptor was replaced with an ON-biased Tar receptor that does not respond to serine, Tar (ΔE_primary) = −2.4 (green solid lines). 1-P_primary(L) is also plotted for Tsr (4Q) trimers (red dotted lines) or similar trimers in which one receptor was replaced with a ligand-bound (OFF biased) Tar receptor, mimicked by setting Tar (ΔE_primary) = 20 (red solid lines). These plots demonstrate that one receptor in a trimer can shift the responses of the other two receptors. (D) 1-P_primary(L) is plotted for trimers containing only Tsr(QEQE) receptors, either native (dotted blue line) or with locked-ON tip domain (solid blue line), mimicked by setting ΔE_tip = −20. The corresponding anisotropy dose–response measured with Tsr-mYFP/A413G is also shown (open circles), fitted (gray line) by assuming also a shift in ΔE_primary (−1.6) imposed by the locked-ON tip. The inset demonstrates a considerable shift in P_primary(L) upon ligand binding in receptors with a locked-ON tip region (ΔE_tip = −20).