Near-IR resonance Raman spectroscopy of archaerhodopsin 3: effects of transmembrane potential

Abstract

Archaerhodopsin 3 (AR3) is a light driven proton pump from Halorubrum sodomense that has been used as a genetically targetable neuronal silencer and an effective fluorescent sensor of transmembrane potential. Unlike the more extensively studied bacteriorhodopsin (BR) from Halobacterium salinarum, AR3 readily incorporates into the plasma membrane of both E. coli and mammalian cells. Here, we used near-IR resonance Raman confocal microscopy to study the effects of pH and membrane potential on the AR3 retinal chromophore structure. Measurements were performed both on AR3 reconstituted into E. coli polar lipids and in vivo in E. coli expressing AR3 in the absence and presence of a negative transmembrane potential. The retinal chromophore structure of AR3 is in an all-trans configuration almost identical to BR over the entire pH range from 3 to 11. Small changes are detected in the retinal ethylenic stretching frequency and Schiff Base (SB) hydrogen bonding strength relative to BR which may be related to a different water structure near the SB. In the case of the AR3 mutant D95N, at neutral pH an all-trans retinal O-like species (O(all-trans)) is found. At higher pH a second 13-cis retinal N-like species (N(13-cis)) is detected which is attributed to a slowly decaying intermediate in the red-light photocycle of D95N. However, the amount of N(13-cis) detected is less in E. coli cells but is restored upon addition of carbonyl cyanide m-chlorophenyl hydrazone (CCCP) or sonication, both of which dissipate the normal negative membrane potential. We postulate that these changes are due to the effect of membrane potential on the N(13-cis) to M(13-cis) levels accumulated in the D95N red-light photocycle and on a molecular level by the effects of the electric field on the protonation/deprotonation of the cytoplasmic accessible SB. This mechanism also provides a possible explanation for the observed fluorescence dependence of AR3 and other microbial rhodopsins on transmembrane potential.

Figure 1. Sequence of AR3 and predicted folding pattern in membrane based on earlier models of archaerhodopsins and other microbial rhodopsins

(see for example reference). Red highlighted residues are involved in the BR proton transport mechanism (see text). D95 in AR3 (yellow), is homologous to the D85 Schiff base proton acceptor in BR and was replaced by an asparagine to form mutant D95N described in this paper. M155 and C154 are highlighted in blue and orange, respectively.

Figure 2. RRS of the BR purple membrane and AR3 reconstituted into E. coli polar lipids recorded in H₂O and D₂O at pH 7

Data was recorded at room temperature using a 785-nm probe laser with 100 mW power (40mW measured at the sample). A background spectrum of the quartz capillary and buffer was subtracted from the sample. The spectra were scaled using the intensity of the ethylenic band at 1526 cm⁻¹. Additional details given in Materials and Methods. Small negative or positive peaks in spectra are denoted by “*” (see also other Figures) and are due to artifacts caused by cosmic radiation entering detector during measurement.

Figure 3. Visible absorption of AR3 reconstituted into E. coli lipids measured after dark adaptation (dashed line) and light adaptation (solid line) at room temperature

The mutant D95N recorded in dark is shown for comparison (dotted line) (see text). Spectra were baseline corrected as described in Materials and Methods. Tick marks on Y-axis for the light and dark adapted spectra are both in 100 mOD intervals and 50 mOD intervals for the D95N spectrum. All spectra were measured in pH7 buffer identical to that used for RRS measurements, see Materials and Methods.

Figure 4. RRS of AR3 reconstituted into E. coli polar lipids measured from pH 2 to 10

All data was recorded under same conditions as described in Materials and Methods. Small artifact (negative peaks) in spectra are indicated by “*”.

Figure 5. RRS of AR3 D95N reconstituted into E. coli polar lipids measured at various pHs ranging from 1.5 to 11

All spectra were recorded using 100 mW, 785nm excitation. In addition to the 1518 cm⁻¹ band seen at pH 7 arising from the red-shifted O^all-trans species a second band appears at 1527 cm⁻¹ at pH>7 indicative of N^13-cis.

Figure 6. Variation of RRS of AR3 D95N at pH 10 with power of 785-nm excitation at 100 mW and 1 mW intensity

The contributions from the N^13-cis species drops significantly in AR3 D95N relative to O^all-trans as based on drop in intensity of 1527 and 1185 cm⁻¹ bands.

Figure 7. Comparison of RRS of AR3 D95N in membrane fragments and in vivo in E. coli at pH 9 under various conditions

A. Membrane fragments reconstituted in E. coli polar lipids. B. In E. coli cells. C. In E. coli cells with CCCP added. D. In sonicated E. coli cells E. E. coli cell not expressing AR3 D95N. All spectra were recorded using 100 mW 785nm excitation. Solid vertical lines with arrows indicate assignment of major bands in AR3 RRS spectra that arise from non-resonant contribution present in non-expressing E. coli (lower trace).

Figure 8. Reproducibility of RRS of AR3 D95N at pH 9 recorded in vivo in E. coli under different conditions

Top set (red) not treated, middle set (blue) sonicated and bottom set (green) treated with CCCP (see Materials and Methods).

Figure 9. Model of AR3 D95N red-light photocycles adapted from Tittor et al. for BR

Absorption of a photon by O^all-trans (red-light photocycle) results in transition to N^13-cis and switch from extracellular (EC) to cytoplasmic (CP) accessible SB followed by establishment of an equilibrium between N^13-cis and M^13-cis (red arrows) which is postulated here to be voltage sensitive and favor M^13-cis for normal negative cell potential.