Rhodopsin photointermediates in two-dimensional crystals at physiological temperatures

Abstract

Bovine rhodopsin photointermediates formed in two-dimensional (2D) rhodopsin crystal suspensions were studied by measuring the time-dependent absorbance changes produced after excitation with 7 ns laser pulses at 15, 25, and 35 degrees C. The crystalline environment favored the Meta I(480) photointermediate, with its formation from Lumi beginning faster than it does in rhodopsin membrane suspensions at 35 degrees C and its decay to a 380 nm absorbing species being less complete than it is in the native membrane at all temperatures. Measurements performed at pH 5.5 in 2D crystals showed that the 380 nm absorbing product of Meta I(480) decay did not display the anomalous pH dependence characteristic of classical Meta II in the native disk membrane. Crystal suspensions bleached at 35 degrees C and quenched to 19 degrees C showed that a rapid equilibrium existed on the approximately 1 s time scale, which suggests that the unprotonated predecessor of Meta II in the native membrane environment (sometimes called MII(a)) forms in 2D rhodopsin crystals but that the non-Schiff base proton uptake completing classical Meta II formation is blocked there. Thus, the 380 nm absorbance arises from an on-pathway intermediate in GPCR activation and does not result from early Schiff base hydrolysis. Kinetic modeling of the time-resolved absorbance data of the 2D crystals was generally consistent with such a mechanism, but details of kinetic spectral changes and the fact that the residuals of exponential fits were not as good as are obtained for rhodopsin in the native membrane suggested the photoexcited samples were heterogeneous. Variable fractional bleach due to the random orientation of linearly dichroic crystals relative to the linearly polarized laser was explored as a cause of heterogeneity but was found unlikely to fully account for it. The fact that the 380 nm product of photoexcitation of rhodopsin 2D crystals is on the physiological pathway of receptor activation suggests that determination of its structure would be of interest.

Figure 1

Rhodopsin photointermediate scheme near physiological temperatures in membrane. Some of the above intermediates can be trapped after low-temperature photolysis, but BSI (whose equilibrated mixture with Batho is sometimes called BL) and Meta I₃₈₀ only build up appreciable concentrations near physiological temperatures (7). The scheme shown is a simplified version describing the principal processes affecting absorbance. The time constants are appropriate for membrane suspensions of rhodopsin. This general scheme also holds for detergent samples (such as lauryl maltoside) except that beginning with Lumi, all decays are significantly faster and subsequent equilibria are more forward shifted.

Figure 2

Electron micrograph of 2D crystals of rhodopsin. An overview picture of a sample obtained from a dialysis reconstitution experiment was taken at low magnification and with strong defocus (300 um) to obtain high contrast. Single layer membranes (s), collapsed tubular membranes (t), and protein aggregates (p) are visible. The single layers and the tubes are two-dimensional crystals of bovine rhodopsin, which are well characterised (8,12,13). The exact nature of the protein aggregates (p) is not known.

Figure 3

Absorbance difference spectra recorded at time delays ranging from 1 μs to 5.54 s after photoexcitation of rhodopsin 2D crystal suspensions. To eliminate absorbance changes due to rotational diffusion, data were collected using probe light that was linearly polarized at the magic angle (54.7°) relative to the linear polarization direction of the 7 ns, 477 nm excitation laser pulse. Measurement temperature is shown in each panel and samples were pH 7. Although as the temperature is increased there is a substantial increase in the amount of 380 nm absorbance observed on the seconds time scale, even at 35 °C approximately 2/3 of the PSB absorbance in 2D rhodopsin crystals remains at 5.54 s after excitation. Note that the peak of the 1 μs data near 460 nm arises from the 494 nm absorbing species, Lumi, with the peak being shifted because the data reports the difference between two species (Lumi and rhodopsin) having similar λ_max.

Figure 4

pH dependence of the time-resolved absorbance difference spectra in rhodopsin 2D crystal suspensions at 35 C. Significantly less of the 380 nm absorbing species is observed on the seconds time scale in pH 5.5 low salt MES buffer compared to pH 7.0 low salt TRIS buffer. If the 380 nm absorbing product in rhodopsin 2D crystals was the “classical” Meta II photointermediate originally characterized by Matthews et al. (19), it should display the so-called anomalous pH dependence i.e. opposite to that expected when the Schiff base linkage titrates due to the pH change of the buffer, contrary to what is seen above. The final product in rhodopsin 2D crystals displays the usual pH dependence expected when a short wavelength absorbing Schiff base partially protonates to form the longer wavelength absorbing PSB in going from pH 7.0 to pH 5.5.

Figure 5

Difference spectra resulting from the thermal quench of photoexcited mixtures from 35 °C to 19 °C on the seconds time scale. The solid line shows the absorbance difference between a sample of rhodopsin 2D crystal suspension that was photoexcited at 35 °C and recorded at that temperature ∼1 s later and the spectrum from a sample that was similarly photoexcited at 35 °C but quenched to 19 °C before recording the spectrum. The dashed line shows the difference absorbance spectrum obtained from samples of a sonicated membrane suspension of rhodopsin in an identical experiment. To estimate the percent thermal reversibility of the equilibrium formed on the ∼1 second time scale, these data need to be compared to the increase in signal recorded in constant temperature measurements over this range. Those data indicate that in membrane the equilibrium is 100% thermally reversible at ∼1 s and approximately 85% thermally reversible in rhodopsin 2D crystals. Note that the larger change seen in this experiment compared to the data shown in Figure 3 for 2D rhodopsin crystals results from the larger bleach achieved by the three sequential flashes used here (see text) compared to the single laser pulses used to produce the data in Figure 3.

Figure 6

Time-dependent absorbance changes (b-spectra) associated with the exponential processes that fit rhodopsin 2D crystal data. Points show the b-spectra associated with the time dependent spectral changes (two exponentials at 15 and 25 °C, and three exponentials at 35 °C) and with b₀, the time independent absorbance change due to excitation. Curves show the fit to the b-spectra obtained using the photointermediate compositions given in Table 1.

Figure 7

Normalized temporal concentration profiles obtained by fitting photointermediate spectra to the raw data from crystal suspensions. Concentration profiles of photointermediates are normally determined using the microscopic rate constants describing the mechanism that best fits the data. In the case of rhodopsin 2D crystals, no single mechanism could be found to fit the data, and these concentration profiles shown by the plotted points (Lumi, ▼; Meta I₄₈₀, ▲; 380 nm absorbers Meta I₃₈₀ and Meta II*(MII_a), ■) were determined by directly fitting the raw data to photointermediate spectra. As a consequence of the empirical nature of the fit, the 380 nm absorbing product cannot be differentiated into Meta I₃₈₀ and Meta II*(MII_a). The continuous change seen in these concentration curves, as opposed to the temporally localized changes expected when exponentials describe the underlying processes, documents the difficulty encountered in fitting a single mechanism with discrete exponentials. The continuous curves in the 35 °C panel show the concentration profiles produced by the more complex, heterogeneous mechanism described in the text.