Engineered socket study of signaling through a four-helix bundle: evidence for a yin-yang mechanism in the kinase control module of the aspartate receptor

Abstract

The chemoreceptors of Escherichia coli and Salmonella typhimurium form stable oligomers that associate with the coupling protein CheW and the histidine kinase CheA to form an ultrasensitive, ultrastable signaling lattice. Attractant binding to the periplasmic domain of a given receptor dimer triggers a transmembrane conformational change transmitted through the receptor to its cytoplasmic kinase control module, a long four-helix bundle that binds and regulates CheA kinase. The kinase control module comprises three functional regions: the adaptation region possessing the receptor adaptation sites, a coupling region that transmits signals between other regions, and the protein interaction region possessing contact sites for receptor oligomerization and for CheA-CheW binding. On the basis of the spatial clustering of known signal locking Cys substitutions and engineered disulfide bonds, this study develops the yin-yang hypothesis for signal transmission through the kinase control module. This hypothesis proposes that signals are transmitted through the four-helix bundle via changes in helix-helix packing and that the helix packing changes in the adaptation and protein interaction regions are tightly and antisymmetrically coupled. Specifically, strong helix packing in the adaptation region stabilizes the receptor on state, while strong helix packing in the protein interaction region stabilizes the off state. To test the yin-yang hypothesis, conserved sockets likely to strengthen specific helix-helix contacts via knob-in-hole packing interactions were identified in the adaptation, coupling, and protein interaction regions. For 32 sockets, the knob side chain was truncated to Ala to weaken the knob-in-hole packing and thereby destabilize the local helix-helix interaction provided by that socket. We term this approach a "knob truncation scan". Of the 32 knob truncations, 28 yielded stable receptors. Functional analysis of the signaling state of these receptors revealed seven lock-off knob truncations, all located in the adaptation region, that trap the receptor in its "off" signaling state (low kinase activity, high methylation activity). Also revealed were five lock-on knob truncations, all located in the protein interaction region, that trap the "on" state (high kinase activity, low methylation activity). These findings provide strong evidence that a yin-yang coupling mechanism generates concerted, antisymmetric helix-helix packing changes within the adaptation and protein interaction regions during receptor on-off switching. Conserved sockets that stabilize local helix-helix interactions play a central role in this mechanism: in the on state, sockets are formed in the adaptation region and disrupted in the protein interaction region, while the opposite is true in the off state.

Figure 1. Schematic core complex illustrating different functional regions of the receptor

Shown is a receptor trimer-of-dimers (individual subunits are unique colors) in complex with the dimeric histidine kinase CheA (blue) and the coupling protein CheW (cyan). This core complex is believed to represent the minimal structural unit required for receptor-regulated kinase activity (1, 51). The structural-functional regions investigated in the present study are italicized. Attractant binding to the periplasmic ligand binding site is believed to trigger different types of conformational changes in each region denoted by a different arrow color, ultimately transmitting a regulatory signal to the bound CheA kinase. In the transmembrane signaling module, the conformational signal is a piston displacement of the signaling helix (blue arrow). The signal conversion or HAMP module converts this piston displacement into a different conformational signal (green arrow). The kinase control module is an extended 4-helix bundle that carries information from HAMP to bound CheA and possesses three functional regions. The yin-yang hypothesis developed herein proposes that signal transmission through the regions of the 4-helix bundle involve changes of helix-helix packing, and that the polarities of these packing changes are opposite in the adaptation and protein interaction regions. The coupling region is essential for communication (grey arrow) of these helix packing changes between the adaptation and protein interaction regions. While attractant binding and conformational changes are indicated only in one receptor dimer, positive cooperativity between receptors may generate corresponding conformational changes in the other dimers.

Figure 2. A method to analyze the effects of local helix-helix interactions

Helix-helix packing in 4-helix bundles is stabilized by socket interactions in which a knob side chain inserts into a hole formed by 4 side chains on an adjacent helix (38, 39). Thus, to locally reduce the strength of helix-helix packing, in general one can truncate the side chain of a selected knob to Ala, thereby decreasing the knob-in-hole packing of a native, inter-helix socket. Shown is a representative socket (left) in which L415 is the knob that inserts into a hole comprised by I359′, I362′, I363′, and I366′ on an adjacent helix provided by the other subunit. Truncating the knob side chain from Leu to Ala, a mutation termed a knob truncation, significantly decreases the knob-hole contacts (right).

Figure 3. Effects of weakened sockets on receptor function in vitro

A) Effects of knob truncations (defined in Fig. 2) on receptor-regulated kinase activity in the reconstituted receptor-CheA-CheW signaling complex. Shown are CheA kinase activities for signaling complexes containing each knob truncation mutant in both the apo (filled bar) and attractant-occupied (1 mM Asp, open bar) states. All kinase activities are normalized to that of the signaling complex containing the apo wild type receptor. Notably, a high percentage of knob truncations in the adaptation region inhibit kinase activity, while a high percentage in the protein interaction region prevent normal attractant inhibition of kinase activity. B) Effects of knob truncations on receptor methylation rates in the reconstituted receptor-CheR complex. Shown are rates of adaptation site methyl esterification by CheR, both in the apo (filled bar) and attractant-occupied (1 mM Asp, open bar) states. All rates are normalized to that of the apo wild type receptor-CheR complex. A high percentage of knob truncations in the adaptation region lead to high methylation rates even in the absence of attractant, while in the protein interaction region the methlation rates are typically low in both the absence and presence of attractant. In addition to the 28 knob truncation mutants, each panel also includes the wild type (WT) receptor and a control receptor (N379F), which does not form the trimer-of-dimers (43) and thus is inactive in the kinase and live cell assays, but is methylated faster than wild type due to the increased accessibility of the adaptation sites to CheR. Knob truncations that are operationally defined (see text) as signal-locking, either lock-on (open star) or lock-off (filled star), are indicated by stars.

Figure 4. Locations of lock-on and lock-off modifications in the kinase control module

Shown is the kinase control module of a single receptor homodimer (1), with the two subunits distinguished by color (blue, gold). Each arrowhead indicates the position of a signal-locking modification (small spheres) in the homodimer, as identified by the anti-symmetric activity assay (see text). Lock-on modifications are operationally defined by high rates of receptor-regulated CheA kinase activity, even in the presence of attractant that normally turns the kinase off, as well as low CheR methylation activity regardless of the attractant concentration. Lock-off modifications are defined by low rates of receptor-regulated CheA kinase activity, and high rates of CheR methylation activity, regardless of the attractant concentration. A) The locations of lock-on disulfide bonds and lock-on Cys substitutions, first described in previous studies (–37) and reconfirmed here using more rigorous methods. The lock-on disulfides covalently stabilize helix-helix packing and are located predominantly in the adaptation region, while the lock-on Cys substitutions likely weaken helix-helix packing and are predominantly located in the protein interaction region. B) The locations of lock-on and lock-off knob truncations, which are clustered in the protein interaction and adaptation regions, respectively. These modifications illustrate the opposite effects that weaker knob-in-holes packing have on receptor signaling state in the two different regions.

Figure 5. Conceptual basis of the yin-yang hypothesis

While the atomic structural-mechanical basis of the yin-yang signaling is not yet clear, the conceptual basis of the model is straightforward. Attractant binding to the periplasmic ligand binding domain triggers a transmembrane conformational change that is converted by the HAMP domain into a different type of conformational signal (1). Our current working model proposes that HAMP on-off switching triggers a scissors-type displacement of its C-terminal HD2–HD2′ helices that directly couple to the N-terminal helices of the kinase control module. In the kinase-activating on-state, 4-helix bundle packing is more stable (sockets formed) in the adaptation region, but less stable (sockets broken, at least partially) in the protein interaction region. The reverse is true in the kinase-inhibiting off-state. Together with previous findings on helix dynamics in the adaptation region (23), the present findings indicate that helix packing changes in this region yield a local frozen-dynamic transition (47) as illustrated. In the protein interaction region, no study of helix dynamics has yet been carried out and it is not clear whether the loss of socket interactions (i) allows other stabilizing interactions to form, yielding a different stable conformation, or (ii) simply destabilizes the region, yielding a looser, more dynamic structure. The helix-helix packing in the coupling region is proposed to possess a significantly lower density of stabilizing sockets (22), and the location of this region indicates that it serves as a signal transmission element between the strongly coupled adaptation and protein interaction regions. The space-filling sidechains represent the four adaptation Glu residues of each subunit, which can also drive yin-yang signaling via an electrostatic mechanism (23).