Symmetric allosteric mechanism of hexameric Escherichia coli arginine repressor exploits competition between L-arginine ligands and resident arginine residues

Abstract

An elegantly simple and probably ancient molecular mechanism of allostery is described for the Escherichia coli arginine repressor ArgR, the master feedback regulator of transcription in L-arginine metabolism. Molecular dynamics simulations with ArgRC, the hexameric domain that binds L-arginine with negative cooperativity, reveal that conserved arginine and aspartate residues in each ligand-binding pocket promote rotational oscillation of apoArgRC trimers by engagement and release of hydrogen-bonded salt bridges. Binding of exogenous L-arginine displaces resident arginine residues and arrests oscillation, shifting the equilibrium quaternary ensemble and promoting motions that maintain the configurational entropy of the system. A single L-arg ligand is necessary and sufficient to arrest oscillation, and enables formation of a cooperative hydrogen-bond network at the subunit interface. The results are used to construct a free-energy reaction coordinate that accounts for the negative cooperativity and distinctive thermodynamic signature of L-arginine binding detected by calorimetry. The symmetry of the hexamer is maintained as each ligand binds, despite the conceptual asymmetry of partially-liganded states. The results thus offer the first opportunity to describe in structural and thermodynamic terms the symmetric relaxed state predicted by the concerted allostery model of Monod, Wyman, and Changeux, revealing that this state is achieved by exploiting the dynamics of the assembly and the distributed nature of its cohesive free energy. The ArgR example reveals that symmetry can be maintained even when binding sites fill sequentially due to negative cooperativity, which was not anticipated by the Monod, Wyman, and Changeux model. The molecular mechanism identified here neither specifies nor requires a pathway for transmission of the allosteric signal through the protein, and it suggests the possibility that binding of free amino acids was an early innovation in the evolution of allostery.

Figure 1. Structural features of ArgR.

A. Intact ArgR in complex with DNA. The protein is viewed down the three-fold axis, with central ArgRC domains forming two trimers stacked on top of each other, and six peripheral ArgRN domains docked with bent B-form DNA. Subunit colors are yellow, A; red, B; green, C; blue, D; cyan, E; magenta, F. The protein structure was prepared from PDB ID 1B4A (apoBstArgR) and DNA structure as described in from PDB ID 1J59 (cAMP receptor). B. ArgRC rotation. Overlay of average structures of the ArgRC domain hexamer from the equilibrated part of the simulations of holoArgRC (orange) and apoArgRC (blue). The overlay was prepared by C_α RMSD minimization using the bottom (ABC) trimer of each structure to show the conformational shift that occurs uniquely in apoArgRC, in which one trimer rotates clockwise relative to the other. Left, top view in approximately the same size and viewpoint as in panel A. Right, side view. Selected secondary structure elements and N- and C-termini are labeled for orientation. CPK spheres mark C_α atoms of the Gly103-Asp128 residues whose interatomic distances (dashed lines) are measured to quantify rotation. C, D. L-arg binding sites. The view is approximately that of panel B, right, but zoomed in on the trimer interface, showing four subunits in front with the two in back faded for clarity. The viewpoint is fixed in all panels by using C_α RMSD minimization to overlay the dark grey helix shown at lower left. Each panel shows one snapshot that resembles the mean state of each indicated simulation. C. EcArgRC. Left, apoArgRC. Right, apoArgRC +1arg. Arg110 forms a doubly-hydrogen bonded salt-bridge (dashed lines) with Asp128. Arg and Asp residues are identified by number and by subunit as defined in panel A. For clarity, filled circles indicate approximate locations of some residues, and some are unlabeled: His99 (yellow), Asp113 (purple), Gly103 (blue). L-arg ligand, cyan; dashed lines indicate some of the interactions made by the Cα substituents of L-arg, with others omitted for clarity. D. ApoMtArgRC. Two pairs of Arg-Asp salt bridges (dashed lines) permit rotation in both directions. Left, Arg133-Asp146 promotes clockwise rotation. Right, Arg118-Asp132 promotes counterclockwise rotation.

Figure 2. Correlated motions.

The covariance matrix is defined as described in Materials and Methods. Vector products representing the maximum extent of correlated motion (Ų) for each C_α pair of each hexamer are plotted. The color scale indicates the degree of correlation: red, positively correlated; blue, negatively correlated; white, uncorrelated. Each panel presents two covariance matrices fused along the diagonal to eliminate redundancy and facilitate comparison of the two simulations indicated above the panel; the simulation presented in the upper left half is indicated first, followed by a slash (/) representing the diagonal, followed by the simulation in the lower right half. The reference state for calculating the extent of C_α motion as described in Materials and Methods is the hexamer for panels A and D and the trimer composed of ABC monomers for panels B and C.

Figure 3. Species distribution.

Distances between Gly103 and Asp128 C_α atoms are averaged over the six atom pairs in the hexamer every 50 ps during the final 10 ns of each indicated simulation. Distances are normalized by subtracting the average distance in holoArgRC crystals or, for apoArgRC110A, the distance in the energy-minimized starting structure. Distances are grouped by size in bins of 0.1 Å.

Figure 4. Reaction coordinate.

Vertical lines mark coordinates for which values of conformational entropy, inter-trimer hydrogen bond number, and relative free energy are plotted based on inspection or calculation from the indicated simulation; second line from left marks the high-energy, binding-competent state based on inference as described in the text; grey zone in bottom panel indicates the rare conformer of apoArgRC. The three lines for the 1arg state represent (left to right) apoArgRC+1b, apoArgRC+1a, and the three holoArgRC-5 simulations. Top panel, conformational contributions to total system entropy (S, kJ/molK), calculated from each covariance matrix in which the reference state is the hexamer. Middle panel, average number of inter-trimer hydrogen bonds (filled circles) counted during each simulation (light blue, persistent bonds (present >50% of time); dark blue, total bond number not counting L-arg; green, total bond number counting L-arg; lines are only to guide the eye). Bottom panel, estimated relative free energy of the system. Estimates combine the values for entropy and hydrogen bond number given in the upper panels with the average per-ligand enthalpy and binding free energy values given in Table S1 (or where unavailable, the values calculated for holoArgRC). Estimated energy levels for ligands three, four, and five are obtained by linear interpolation between the levels for two and six bound L-arg ligands. The data do not uniquely constrain the energy level of the +2 state (dashed). HoloArgRC-6 simulations that did not equilibrate are inferred to approach the energy barrier from the right (red dashed line).

Figure 5. Motion gradient.

View of one ArgRC subunit, with the L-arg binding site at the center of the hexamer at upper left (asterisk) and the outside of the hexamer at the right. The color gradient represents the extent of motion (blue, lowest to red, highest) defined by the root-mean-square displacement of each atom from its average position. Thickness of lines represents distance from the viewer (thickest in front). Left, apoArgRC; right, apoArgRC+1a.