Chromophore structure in lumirhodopsin and metarhodopsin I by time-resolved resonance Raman microchip spectroscopy

Abstract

Time-resolved resonance Raman microchip flow experiments have been performed on the lumirhodopsin (Lumi) and metarhodopsin I (Meta I) photointermediates of rhodopsin at room temperature to elucidate the structure of the chromophore in each species as well as changes in protein-chromophore interactions. Transient Raman spectra of Lumi and Meta I with delay times of 16 micros and 1 ms, respectively, are obtained by using a microprobe system to focus displaced pump and probe laser beams in a microfabricated flow channel and to detect the scattering. The fingerprint modes of both species are very similar and characteristic of an all-trans chromophore. Lumi exhibits a relatively normal hydrogen-out-of-plane (HOOP) doublet at 951/959 cm(-1), while Meta I has a single HOOP band at 957 cm(-1). These results suggest that the transitions from bathorhodopsin to Lumi and Meta I involve a relaxation of the chromophore to a more planar all-trans conformation and the elimination of the structural perturbation that uncouples the 11H and 12H wags in bathorhodopsin. Surprisingly, the protonated Schiff base C=N stretching mode in Lumi (1638 cm(-1)) is unusually low compared to those in rhodopsin and bathorhodopsin, and the C=ND stretching mode shifts down by only 7 cm(-1) in D2O buffer. This indicates that the Schiff base hydrogen bonding is dramatically weakened in the bathorhodopsin to Lumi transition. However, the C=N stretching mode in Meta I is found at 1654 cm(-1) and exhibits a normal deuteration-induced downshift of 24 cm(-1), identical to that of the all-trans protonated Schiff base. The structural relaxation of the chromophore-protein complex in the bathorhodopsin to Lumi transition thus appears to drive the Schiff base group out of its hydrogen-bonded environment near Glu113, and the hydrogen bonding recovers to a normal solvated PSB value but presumably a different hydrogen bond acceptor with the formation of Meta I.

FIGURE 1

Rhodopsin photobleaching sequence.

FIGURE 2

Time-resolved resonance Raman microchip apparatus. The two cylindrically focused beams are displaced along the flow direction. The beam separation and flow velocity determine the time delay. The pump beam was focused to a 5-10 μm × 100 μm spot, and the probe beam was 5 μm × 100 μm. The expanded chip view shows just the bottom etched wafer that makes up the bonded two-layer glass sandwich.

FIGURE 3

Time-resolved (16 μs delay) resonance Raman microchip spectra of lumirhodopsin: (A) pump-plus-probe spectrum and (B) probe-only spectrum. Spectra C-E are lumirhodopsin spectra obtained by subtracting the indicated fractions of the probe-only spectrum from the pump-plus-probe spectrum. The optimum subtraction coefficient was determined to be 0.58. The pump wavelength was 488 nm, and the probe wavelength was 514.5 nm. A typical Lumi spectrum required only 5 mL of sample.

FIGURE 4

Time-resolved (1 ms delay) resonance Raman microchip spectra of metarhodopsin I: (A) pump-plus-probe spectrum and (B) probe-only spectrum. Spectra C-E are metarhodopsin I spectra obtained by subtracting the indicated fractions of the probe-only spectrum from the pump-plus-probe spectrum. The optimum subtraction coefficient was determined to be 0.23. The pump and probe excitation wavelengths were 531 and 476.5 nm, respectively. A typical Meta I spectrum required only 2 mL of sample.

FIGURE 5

Time-resolved resonance Raman microchip spectra of lumirhodopsin in (A) H₂O and (B) D₂O and of metarhodopsin I in (C) H₂O and (D) D₂O.

FIGURE 6

Schiff base regions of the Raman spectra of lumirhodopsin in (A) H₂O and (B) D₂O and of metarhodopsin I in (C) H₂O and (D) D₂O. Lorentzian fits to the bands are indicated by the dashed lines. The indicated frequencies are those of the band components.

FIGURE 7

Frequencies of the C=NH and C=ND stretching modes of rhodopsin, bathorhodopsin, lumirhodopsin, and metarhodopsin I. The error bars indicate that the frequencies in our microchip experiments are accurate within ±1 cm^-1 for rhodopsin and ±2 cm^-1 for lumirhodopsin and metarhodopsin I.