Optogenetic manipulation of cGMP in cells and animals by the tightly light-regulated guanylyl-cyclase opsin CyclOp

Abstract

Cyclic GMP (cGMP) signalling regulates multiple biological functions through activation of protein kinase G and cyclic nucleotide-gated (CNG) channels. In sensory neurons, cGMP permits signal modulation, amplification and encoding, before depolarization. Here we implement a guanylyl cyclase rhodopsin from Blastocladiella emersonii as a new optogenetic tool (BeCyclOp), enabling rapid light-triggered cGMP increase in heterologous cells (Xenopus oocytes, HEK293T cells) and in Caenorhabditis elegans. Among five different fungal CyclOps, exhibiting unusual eight transmembrane topologies and cytosolic N-termini, BeCyclOp is the superior optogenetic tool (light/dark activity ratio: 5,000; no cAMP production; turnover (20 °C) ∼17 cGMP s(-1)). Via co-expressed CNG channels (OLF in oocytes, TAX-2/4 in C. elegans muscle), BeCyclOp photoactivation induces a rapid conductance increase and depolarization at very low light intensities. In O2/CO2 sensory neurons of C. elegans, BeCyclOp activation evokes behavioural responses consistent with their normal sensory function. BeCyclOp therefore enables precise and rapid optogenetic manipulation of cGMP levels in cells and animals.

Figure 1. Assessing fungal CyclOps for light-regulated cGMP generation.

(a) Plasma membrane fluorescence of Xenopus oocytes expressing BeCyclOp, AmCyclOp1, 2 and 3, and CaCyclOp, as C-terminal YFP fusions, with and without 1 μM ATR, as indicated. (b) cGMP production of different fungal CyclOps (as indicated), fused to YFP, was assessed in extracts prepared from oocytes expressing these proteins, that were either kept in the dark (D), or after 2 min green light (L; 532 nm, 0.15 mW mm⁻² – BeCyclOp) or 0.5 mW mm⁻² (all other CyclOps). ATR was added, unless noted. N=2 experiments, mean of 5–6 oocytes each, error bars, s.d. (c) Light-induced cGMP production in membrane preparations from oocytes expressing various fungal CyclOps as YFP fusions (unless otherwise noted), or combinations thereof (see Tab. 1 for amounts of cRNAs injected). Membranes were kept in dark (D, black bars) or illuminated (L, green bars) with 532 nm light (0.5 mW mm⁻²), and parallel reactions were quenched after 1 or 5 min, and cGMP measured. Note the split left y axis; logarithmic scale in lower part. On the right y axis, and shown in red, is the L/D ratio of the cGMP production rates. Average of N=3 experiments; error bars, s.d. Statistically significant differences were determined by one-way analysis of variance: *P<0.05; **P<0.01; ***P<0.001. (d) HEK293T cells were seeded on coverslips and transfected with a plasmid encoding BeCyclOp::YFP. Cells were fixed and analysed by confocal microscopy. Scale bar, 10 μm. (e) cGMP and cAMP assay of HEK293T cells transfected with BeCyclOp under different conditions.+ATR: 1 μM ATR added;—ATR: no additional ATR. L, in light. D, in dark. Statistically significant differences determined by 1-way ANOVA: *P<0.05; ***P<0.001. n=6, error bars, s.d.

Figure 2. BeCyclOp photoactivation and cGMP production.

(a) BeCyclOp action spectrum. Dotted grey line indicates the relative dark activity, light intensities at different wavelengths were adjusted to ∼0.02 mW mm⁻². cGMP production per photon amount was calculated, normalized to the action spectrum peak. Mean of N=3 experiments, error bars, s.d. (b) Light (532 nm) intensity dependence of mean normalized cGMP production of BeCyclOp-containing membranes; K_0.5=0.055 mW mm⁻². N=3 experiments, error bars, s.d. (c) Kinetics of mean, normalized cGMP production of BeCyclOp-containing membranes, measured at indicated times after 20 ms illumination. Fitting a mono-exponential yields τ=320±20 ms. n=5 experiments; error bars, s.d. (d) Mean BeCyclOp cGMP production in the dark and with continuous illumination (0.5 mW mm⁻², 532 nm), measured at indicated times. N=3 experiments, error bars, s.d.

Figure 3. Cytosolic localization and regulation of cyclase activity by the BeCyclOp N terminus.

(a) BeCyclOp BiFC construct: aa 1–155 of YFP (YN) fused to C-terminal, and aa 155–239 of YFP (YC) fused to the N-terminal ends of BeCyclOp, respectively. (b) Oocyte membrane fluorescence images, resulting from expression of different BeCyclOp YFP/BiFC fusion constructs. BeCyclOpS refers to deletion of aa 1–90 in the N terminus. (c) cGMP production in Xenopus oocyte expressing different BeCyclOp constructs in dark (D) and after illumination (L; 532 nm, 0.2 mW mm⁻², 2 min). N=2 experiments, mean value of six oocytes each; error bars, s.d. Statistically significant differences determined by one-way analysis of variance: **P<0.01; ***P<0.001.

Figure 4. Light-induced muscle activation in C. elegans via BeCyclOp and a CNG channel.

(a) Co-expression of BeCyclOp and the C. elegans CNG channel consisting of TAX-2 and TAX-4 subunits (not normally expressed in muscle; indicated by yellow and orange ovals) generates a light-activated system for cell depolarization and muscle activation. (b) Fluorescent micrographs of the head region of an animal expressing BeCyclOp from a bicistronic construct with mCherry, in body wall muscles, and co-expressing TAX-2::GFP/TAX-4::GFP. Anterior is to the left. Scale bar, 20 μm. Membrane expression of TAX-2/4 is clearly visible as well as cytosolic mCherry. (c) In vivo cGMP and cAMP concentration assessed in crude extracts derived from whole animals expressing BeCyclOp in muscle cells. The animals were cultivated in the absence or presence of all-trans retinal (ATR) and were illuminated with green light (540–580 nm, 150 μW mm⁻², 15 min; L) or were kept in the dark (d) during extract preparation. N=6 experiments; Error bars, s.d. Statistically significant differences: one-way ANOVA (***P<0.001). (d–f) Body length measurements of animals before and following 100 ms light pulses of blue (450–490 nm, 0.9 mW mm⁻²; d,f) or green light (520–550 nm, 0.9 mW mm⁻²; e). Animals were raised in the absence or presence of ATR (orange or green curves, respectively). Only animals supplemented with ATR exhibit contractions. These are long lasting (mono-exponential fits are shown by dashed lines and the relaxation time constants are indicated), and can be repeatedly evoked (e). Mean normalized body length of the indicated number of animals; error bars, s.e.m. (g) Body length changes were induced only when BeCyclOp (‘C') was co-expressed with TAX-2/TAX-4 (‘T2/T4') and animals were raised in the presence of ATR. Illumination was either with blue light (blue graphs), or with green light (green graphs). Bars indicate 2 s illumination period.

Figure 5. Light dose-response correlation of BeCyclOp-evoked body contractions in C. elegans.

(a) Mean normalized body length chronograms of animals before, during and after 1 s green light stimuli (green bar) of the indicated intensities. The negative control −ATR was performed at 1.52 μW mm⁻² (n=10 each, error bars, s.e.m.). (b) Dose-response graph (linear fit of contraction and the log of the light intensities used) shows a negative correlation. Above 1.52 μW mm⁻², contraction reaches saturation.

Figure 6. BeCyclOp triggers O₂ sensory BAG neurons that intrinsically use cGMP signalling.

(a) Confocal image of expression of BeCyclOp::SL2::mCherry bicistronic construct in the BAG neuron pair in the head of C. elegans. Cell bodies, nerve ring processes and sensory dendrites reaching the nose are visible. Outline of the animal indicated by a dashed white line, anterior is up. Green fluorescence: GFP co-expression marker localized to intestinal cell nuclei. Scale bar, 20 μm. (b) Mean absolute locomotion speed of the indicated numbers of animals expressing BeCyclOp in BAG neurons, during repeated blue light exposure (470 nm, 70 μW mm⁻²) for the indicated periods (blue bars), with 30 s interstimulus intervals, as deduced from video analysis using a multiworm tracker. Animals in the absence (orange curves) or presence of ATR (green curves) in the culture media are compared. Speed increase in animals without ATR is due to photophobic behavior. N=5 (4 for no ATR) experiments with n=15–20 animals each; error bars, s.e.m. (c) As in b, but animals expressing ChR2 in BAG neurons were used instead. N=3 experiments with n=15–20 animals each; error bars, s.e.m. The insets show mean speeds, as a moving average, using a sliding window bin corresponding to 1 s for filtering the data. Black bars, 30 s.