Mechanism of partial agonism at the GluR2 AMPA receptor: Measurements of lobe orientation in solution

Abstract

The mechanism by which the binding of a neurotransmitter to a receptor leads to channel opening is a central issue in molecular neurobiology. The structure of the agonist binding domain of ionotropic glutamate receptors has led to an improved understanding of the changes in structure that accompany agonist binding and have provided important clues about the link between these structural changes and channel activation and desensitization. However, because the binding domain has exhibited different structures under different crystallization conditions, understanding the structure in the absence of crystal packing is of considerable importance. The orientation of the two lobes of the binding domain in the presence of a full agonist, an antagonist, and several partial agonists was measured using NMR spectroscopy by employing residual dipolar couplings. For some partial agonists, the solution conformation differs from that observed in the crystal. A model of channel activation based on the results is discussed.

Figure 1

Overview of the refinement protocol. In the first step, the RDCs are used to refine the orientation of the NH bond vectors within each lobe independently to remove structural noise. This employs data from five alignment media (for each medium, two alignment frames, one for each lobe, are used). Although the fit of the measured RDCs to the calculated RDCs is improved dramatically, this results in only a subtle reorientation of the NH bond vectors as shown illustrated by the structures in the lower left. The pink structure corresponds to the glutamate-bound crystal structure (1ftj), with the amide protons colored red. The light blue structure is the structure refined by RDCs in the first step, with the amide protons colored blue. No global movements of the lobes are made in this step. Following local refinement, one alignment frame for each medium is used to reorient the lobes for consistency with the RDC measurements. The last step also uses one alignment frame per medium and allows both the global and local structure to be refined simultaneously. The result is shown in the structure on the bottom right (coloring is the same as the structures on the left). Glutamate is shown in the binding site as a space-filling model. Lobes 1 of the crystal structure and the RDC-refined structure are aligned. Bottom: The RDC-refined glutamate-bound structure fits within the electron density map of the crystal structure of the glutamate-bound form of GluR2 S1S2 (3dp6), both when Lobe 1 is superimposed on the crystal structure (left) and when Lobe 2 is superimposed (right).

Figure 2

Comparison of the IW-bound structures of GluR2-S1S2 aligned along Lobe 1, illustrating the differences in lobe orientation. The crystal structure in the presence of Zn (1my4; 49) is shown in yellow, the crystal structure in the absence of Zn (1mqg; 5) is shown in cyan, and the RDC-refined structure is shown in magenta.

Figure 3

The efficacy of willardiine derivatives and kainate as a function of the degree of lobe opening relative to the glutamate-bound structure. The efficacies are taken from Jin et al. (5) and Mankiewicz et al. (12) and are relative to the full agonist, glutamate. When a given agonist was tested in both studies, the average value was used and the standard error of the mean shown as the error bar. Lobe opening relative to the corresponding glutamate-bound structure (see Experimental Procedures) is shown as filled squares. Crystal structures in the absence of zinc (1, 5) are shown as open squares, and those in the presence of zinc (; Ahmed et al., unpublished) are shown as open circles.

Figure 4

Structures of the peptide flip region of GluR2-S1S2. (A) The structure of glutamate-bound GluR2-S1S2 (1ftj; 1) is shown with the amide nitrogen and proton highlighted in blue. The box shows the position of the peptide flip region. (B) The flipped and the unflipped version of the glutamate-bound structure. The amide nitrogen and proton are highlighted in blue. The hydrogen bonds formed across the lobe interface are indicated. In the case of D651 and Y450, the hydrogen bond is mediated by a water molecule, which is shown as a tan sphere. (C) Superimposion of the glutamate-bound structures shown in B with the kainate-bound structure (1fw0; 1) and two UBP277-bound structures (Ahmed et al., unpublished). The orientation of the amide bond for S652 and G653 differs for the kainate- and UBP277-bound structures from either the flipped or the unflipped conformation of the glutamate-bound structure (indicated by asterisks).

Figure 5

(A) Tryptophan side chain region of the ¹H/¹⁵N TROSY spectrum of GluR2 S1S2 bound to each of the compounds in the willardiine series. (B) Relative intensity of the W767 peak as a function of temperature for FW and IW. The peak height of W767 was normalized to the height of the tryptophan peak at 10.15 ppm.