Two protonation switches control rhodopsin activation in membranes

Abstract

Activation of the G protein-coupled receptor (GPCR) rhodopsin is initiated by light-induced isomerization of the retinal ligand, which triggers 2 protonation switches in the conformational transition to the active receptor state Meta II. The first switch involves disruption of an interhelical salt bridge by internal proton transfer from the retinal protonated Schiff base (PSB) to its counterion, Glu-113, in the transmembrane domain. The second switch consists of uptake of a proton from the solvent by Glu-134 of the conserved E(D)RY motif at the cytoplasmic terminus of helix 3, leading to pH-dependent receptor activation. By using a combination of UV-visible and FTIR spectroscopy, we study the activation mechanism of rhodopsin in different membrane environments and show that these 2 protonation switches become partially uncoupled at physiological temperature. This partial uncoupling leads to approximately 50% population of an entropy-stabilized Meta II state in which the interhelical PSB salt bridge is broken and activating helix movements have taken place but in which Glu-134 remains unprotonated. This partial activation is converted to full activation only by coupling to the pH-dependent protonation of Glu-134 from the solvent, which stabilizes the active receptor conformation by lowering its enthalpy. In a membrane environment, protonation of Glu-134 is therefore a thermodynamic rather than a structural prerequisite for activating helix movements. In light of the conservation of the E(D)RY motif in rhodopsin-like GPCRs, protonation of this carboxylate also may serve a similar function in signal transduction of other members of this receptor family.

Fig. 1.

Thermodynamic and structural model for rhodopsin activation in a membrane environment. (A) The structure of the dark state of rhodopsin [based on Protein Data Bank ID code 1GZM (25)] includes 2 protonation switches. The cytoplasmic network around Glu-134 in the ERY motif of H3 is involved in proton uptake from the solvent, and the interhelical salt bridge between the protonated retinal Schiff base (PSB) linked to H7 and Glu-113 on H3 in the transmembrane domain is disrupted by internal proton transfer during receptor activation. (B) Previous studies (14, 18, 19) have proposed a sequential reaction scheme of coupled equilibria of photoproduct states for rhodopsin activation in a detergent environment, involving PSB deprotonation in the transition to Meta II_a, activating helix movements in the transition to Meta II_b, and cytoplasmic proton uptake by Glu-134 in the transition to Meta II_bH⁺. (The terms inactive and active are used to describe the receptor conformations before and after helix movements, respectively, and do not necessarily specify activity toward G protein.) (C) This reaction scheme predicts a complex titration behavior of photoproducts in a membrane environment, with Meta I, Meta II_a, and Meta II_b forming an equilibrium at alkaline pH. This intrinsically pH-independent equilibrium is coupled to a pH-dependent equilibrium with Meta II_bH⁺ at lower pH, reflecting proton uptake by Glu-134 from the solvent. (D) Because Meta II_a and Meta II_b are populated only at higher temperature, this 4-state scheme reduces to the classical 2-state Meta I/Meta II_bH⁺ scheme at lower temperature.

Fig. 2.

UV–visible spectroscopic characterization of the PSB salt bridge in the Meta I/Meta II equilibrium in native disk membranes. (A and B) At 20 °C and pH 5.0, photolysis converts the dark state of rhodopsin (black) to the Meta II_bH⁺ photoproduct state with deprotonated Schiff base (red), whereas, at 10 °C and pH 9.5, the Meta I photoproduct with protonated Schiff base (blue) is formed. (C) Difference spectra for the photoproduct minus dark state obtained at 10 °C reveal the pH dependence of the Meta I/Meta II_bH⁺ equilibrium, which is completely on the side of Meta II_bH⁺ at pH 5.0 and of Meta I at pH 9.0 and 10.0, as evident from the complete lack of absorption at 380 nm. (D) At 30 °C, the equilibrium is not fully shifted to Meta I at alkaline pH; instead a 380-nm Meta II photoproduct contribution is observed at very alkaline pH, which becomes pH-independent at pH 9.0 and above. (E) Schiff base deprotonation (θ_UV–vis) and, hence, the fraction of Meta II were determined from UV–visible difference spectra. At 10 °C the titration curve follows a regular Henderson–Hasselbalch function, whereas, at higher temperature, the alkaline endpoint θ_UV–vis^alk does not reach zero because of increasing population of the Meta II_a and Meta II_b states.

Fig. 3.

FTIR spectroscopic characterization of activating conformational changes in the Meta I/Meta II equilibrium in native disk membranes. (A) Meta I and Meta II_bH⁺ minus dark state FTIR difference spectra were obtained from rhodopsin in native disk membranes at 10 °C at pH 9.5 and 20 °C at pH 5.0, respectively, under identical conditions as in Fig. 2 A and B. By using the conformationally sensitive range marked in green, they served as basis spectra for inactive and active conformations, respectively, to determine the pH dependence of activating conformational changes. (B) The FTIR-based titration curves of receptor conformation, θ_FTIR (filled circles and solid lines), reveal a similar deviation from classical Henderson–Hasselbalch behavior as the UV–visible data, with a temperature-dependent increase of the alkaline endpoint θ_FTIR^alk. The alkaline endpoints θ_UV–vis^alk of the UV–visible-based titration curves (open symbols, broken lines, reproduced from Fig. 2E) at 30 and 37 °C are slightly higher than θ_FTIR^alk, revealing the additional contribution of inactive Meta II_a. (C) Rhodopsin reconstituted into synthetic POPC membranes shows the same general activation scheme but with shifted equilibria because of altered properties of the lipid bilayer. Both in native disk membranes and synthetic POPC membranes, proton uptake by Glu-134 and, hence, the transition to protonated Meta II_bH⁺ is necessary for full receptor activation.

Fig. 4.

Meta II_b and Meta II_bH⁺ share a similar active state FTIR signature. FTIR difference spectra for the photoproduct minus dark state obtained for disk membranes at the alkaline endpoint reveal the increasing contribution of the active conformation of unprotonated Meta II_b to the Meta I/Meta II_a/Meta II_b photoproduct equilibrium at higher temperature. Synthetic Meta I/Meta II_bH⁺ minus dark state spectra (red) were calculated as a linear combination of the Meta I and Meta II_bH⁺ basis spectra shown in Fig. 3A by using the endpoint values θ_UV–vis^alk (neglecting the small contribution of Meta II_a). These synthetic spectra show a remarkable agreement with the experimental endpoint spectra, indicating similar FTIR signatures of Meta II_b and Meta II_bH⁺.