Reproducible and sustained regulation of Gαs signalling using a metazoan opsin as an optogenetic tool

Abstract

Originally developed to regulate neuronal excitability, optogenetics is increasingly also used to control other cellular processes with unprecedented spatiotemporal resolution. Optogenetic modulation of all major G-protein signalling pathways (Gq, Gi and Gs) has been achieved using variants of mammalian rod opsin. We show here that the light response driven by such rod opsin-based tools dissipates under repeated exposure, consistent with the known bleaching characteristics of this photopigment. We continue to show that replacing rod opsin with a bleach resistant opsin from Carybdea rastonii, the box jellyfish, (JellyOp) overcomes this limitation. Visible light induced high amplitude, reversible, and reproducible increases in cAMP in mammalian cells expressing JellyOp. While single flashes produced a brief cAMP spike, repeated stimulation could sustain elevated levels for 10s of minutes. JellyOp was more photosensitive than currently available optogenetic tools, responding to white light at irradiances ≥1 µW/cm(2). We conclude that JellyOp is a promising new tool for mimicking the activity of Gs-coupled G protein coupled receptors with fine spatiotemporal resolution.

Figure 1. OptoXR sequence alignment.

A number of structural variants on a published OptoXR chimera comprising elements of the human β2AR and human rhodopsin sequences were generated in an attempt to increase response amplitude/reproducibility. An amino acid alignment of these variants and, for comparison the human β2AR and rhodopsin (Genbank NM_000539.3, in red and NM_000024, in black), and JellyOp (Genbank AB435549, in blue) sequences are shown. Structural boundaries are based on bovine rhodopsin with putative cytoplasmic regions shaded in dark grey. The lysine residue in TM7, which forms a Schiff-base linkage with retinaldehyde chromophore, is highlighted in green. Note that the terminal 9 amino acids of rod opsin are included as an epitope tag (1D4) in all receptors used in this study (light grey shading). In addition to the published OptoXR in which the entire cytoplasmic surface of rod opsin is replaced by that of the β2AR (Rh1B2AR 1-t), variants in which either 1^st or 1^st and 2^nd intracellular loops from rod opsin were retained (Rh1B2AR 2-t and Rh1B2AR C3,t) in the hope of improving chimera stability were generated. Other variants on Rh1B2AR 1-t employed site directed mutagenesis of phosphorylation sites (highlighted in red) important for arrestin binding and receptor inactivation (Rh1B2AR 1-t phos mutant) or a fusion of the human Gαs subunit at the C-terminal tail in purple (Rh1B2AR 1-t::Gαs).

Figure 2. Validating the cAMP biosensor.

A HEK293 cell line expressing the Glosensor™ cAMP biosensor under a tetracycline inducible promoter (FLP-IN™ system; Invitrogen) was generated. The reporter was validated by determining the effect of known doses of forskolin on levels of cAMP, determined by ELISA (a) and Glosensor bioluminescence (b). Data show mean for 2 (ELISA) and 3 (luminescence) separate experiments each of which contained samples in triplicate. Fits show sigmoidal dose response curves of the form y = a + b/1+10^(c-x) where a = bottom, b = top-bottom and c =  LogEC50, and yield EC50 values of 51 and 21 µM for ELISA and luminescence assays respectively. (c) A comparison of cAMP concentration and RLU for each forskolin concentration was used to infer the relationship between these parameters. This could be fit by a first order polynomial (R² value of 0.999), suggesting that, under these conditions, cAMP concentration can therefore be estimated from RLU as follows: [cAMP] µM  = 4.785 + (0.0003971a) + (0.000005261â2) where a  =  RLU/µl.

Figure 3. Real time analysis of a standard OptoXR response.

Light induced changes in cAMP biosensor (Glosensor) luminescence in HEK293 cells transiently transfected with Rh1B2AR 1-t. Data for cells incubated with 9 cis retinaldehyde (black) or all trans retinaldehyde (grey) are shown; yellow arrows depict presentation of light flash. Data points show mean ± SEM n = 7. Inset shows immunocytochemical staining for the 1D4 epitope (in red, alexa 555 secondary antibody) of HEK293 cells expressing Rh1B2AR 1-t. DAPI shown in blue, scale bar  = 10 µm.

Figure 4. Real time analysis of OptoXR variant responses.

(a–d) Light induced changes in cAMP biosensor (Glosensor) luminescence in HEK293 cells transiently transfected with Rh1B2AR 2-t (a) Rh1B2AR 1-t phos mutant (b), Rh1B2AR 1-t::Gαs (c) or Rh1B2AR C3,t (d). Luminescence was measured for 1 second every minute and _ormalized to baseline dark levels. Yellow arrows depict presentation of light flash. (e) Quantification of peak response amplitude for the first light flash following incubation with 9 cis retinaldehyde reveals that only the Rh1B2AR 2-t chimera shows an improved response amplitude compared to the Rh1B2AR 1-t OptoXR (one-way ANOVA with Dunnett's post hoc comparison to Rh1B2AR 1-t p<0.01). (f) None of these chimera exhibited reduced dark activity (luminescence prior to light exposure, _ormalized to mock transfected control cells), and indeed this parameter was significantly increased in cells expressing Rh1B2AR 1-t::Gαs when preincubated with 9 cis retinaldehyde (one-way ANOVA with Dunnett's post hoc comparisons to mock transfections as control group, p<0.05). *p<0.05, **p<0.01. Data show mean ± SEM (n≥4) shown for cells preincubated overnight with either 9-cis retinaldehyde (black) or all trans retinaldehyde (grey).

Figure 5. JellyOp driven light responses.

(a) Light induced changes in cAMP biosensor luminescence (depicted _ormalized to baseline before light exposure) in HEK293 cells transiently transfected with JellyOp (n = 5). (b,c) The peak increase in luminescence following the first light flash was significantly greater for 9-cis pretreated cells expressing JellyOp than either the Rh1B2AR 1-t (1-t) or Rh1B2AR 2-t (2-t) chimera. It was also significantly greater for all-trans pretreated cells expressing JellyOp than Rh1B2AR 1-t. (d,e) Sustained responses were further enhanced, with luminescence over 5 minutes of light exposure (flashes once per minute) reaching a plateau significantly higher in JellyOp than either Rh1B2AR chimera. (f) Tracking changes in luminescence with higher temporal resolution (reading every 30s) in HEK293 cells stably transfected with JellyOp revealed high magnitude responses that could be sustained over 15 minutes of repeated stimulation (n = 4). Inset shows immunocytochemical staining for 1D4 epitope (red) in JellyOp expressing HEK293 cells. Nuclei stained blue with DAPI; scale bar 10 µm. (g) Cells stably expressing JellyOp show markedly repressed responses to a light flash when treated with 100 µM MDL2330A (adenylate cyclase inhibitor; dashed line), n = 1 (h) cAMP biosensor luminescence responses in HEK293 cells transiently transfected with JellyOp (continuous line) are abolished in the JellyOp lysine mutant (dashed line), n = 1 (i) Irradiance response curves for cAMP reporter activity in cells stably expressing JellyOp and induced with 10s white light (LED) pulses. RLU values are _ormalized to the peak response (n = 3). A-I show cells pre-incubated with either 9-cis (black) or all-trans (grey) retinaldehyde; mean ± SEM; yellow arrows depict timing of light flash. Data in b-e were analysed using one-way ANOVA with Dunnett's post hoc comparisons to JellyOp as control group, *p<0.05, **p<0.01; n≥4.

Figure 6. MAPK phosphorylation.

Immunohistochemical labelling of phosphorylated MAPK (grey-black stain produced with horseradish peroxidise and diaminobenzidine) in HEK293 cells preincubated with 9-cis retinaldehyde and exposed to either 2 minutes of white light or kept in the dark. (a) OptoXRs elicit a light dependent increase in MAPK phosphorylation. Photomicrographs for untransfected cells and cells expressing each of the OptoXRs were taken with the same exposure settings. Scale bar  = 50 µm. (b) Phosphorylated MAPK in HEK293 cells expressing JellyOp exposed to the dark or 15 mins of light. Cells labelled ‘+PTX’ were treated with 100ng/ml pertussis toxin (a Gi inhibitor); ‘+U73122’ with 10 µM U73122 (a Gq inhibitor); ‘+MDL2330A’ with 100 µM (adenylate cyclase inhibitor). MDL inhibits light-induced MAPK phosphorylation following both 2 mins (not shown) and 15 mins of light. The increase in phosphorylation therefore appears to be due to a Gs dependent pathway.