Melanopsin-Encoded Response Properties of Intrinsically Photosensitive Retinal Ganglion Cells

Abstract

Melanopsin photopigment expressed in intrinsically photosensitive retinal ganglion cells (ipRGCs) plays a crucial role in the adaptation of mammals to their ambient light environment through both image-forming and non-image-forming visual responses. The ipRGCs are structurally and functionally distinct from classical rod/cone photoreceptors and have unique properties, including single-photon response, long response latency, photon integration over time, and slow deactivation. We discovered that amino acid sequence features of melanopsin protein contribute to the functional properties of the ipRGCs. Phosphorylation of a cluster of Ser/Thr residues in the C-terminal cytoplasmic region of melanopsin contributes to deactivation, which in turn determines response latency and threshold sensitivity of the ipRGCs. The poorly conserved region distal to the phosphorylation cluster inhibits phosphorylation's functional role, thereby constituting a unique delayed deactivation mechanism. Concerted action of both regions sustains responses to dim light, allows for the integration of light over time, and results in precise signal duration.

Figure 1. Photoactivated-melanopsin is phosphorylated at C-terminus Ser/Thr residues

(A) Schematic diagram of mammalian rhodopsin, Drosophila rhodopsin, and mammalian melanopsin. (B) Melanopsin residues 381–397 are important for desensitization upon light stimulation. Xenopus oocytes injected with mRNAs encoding mouse Opn4, β-arr1 (or β-arr2), Gq, and TrpC3 and perfused with 11-cis retinal showed light-activated membrane depolarization. After lights off, the photocurrent returned to baseline in oocytes expressing Opn4^WT and arrestin, but remained depolarized if arrestin was not co-injected. Sustained depolarizing photocurrents were also observed with C-terminal truncated versions of melanopsin (Opn4^380Δ). A version of Opn4 with 397 amino acids (Opn4^397Δ) exhibited arrestin-dependent deactivation. (C) Putative phosphorylation sites in mouse melanopsin between residues 380 and 397 are conserved across species. (C and D) WT or C-terminus truncations (380Δ or 364Δ) were expressed in HEK cells and metabolically labeled with P³². Western blotting revealed that full-length melanopsin was phosphorylated, but not the truncated versions. (E) Phosphopeptide digested products from light activated melanopsin separated by thin-layer chromatography showed phosphorylation primarily at Ser and to less extent at Thr sites.

Figure 2. Melanopsin phosphorylation contributes to response properties

(A) Mouse Opn4 mutants carrying Alanine (A) amino acid substitution at candidate Ser/Thr (S/T) phosphorylation sites. The amino acid sequence and coordinates for mouse melanopsin are shown on the top. (B) Deactivation of melanopsin photoresponse in CHO cells transduced with different mutant versions of Opn4. Average time (sec ± s.e.m.) for the peak response to return to 30% of maximal response. (C) Representative calcium responses for CHO cells expressing Opn4^WT or Opn4 variants. (D, E, F) Comparison of arbitrary clusters of S/T mutations. Representative traces (D), design (E), and average peak amplitude and time (sec ± s.e.m.) for the peak response to return to 30% of peak amplitude (F).

Figure 3. Alterations to the melanopsin C-terminus tail results in altered response properties

(A) Circadian wheel running activity of rd;Opn4^Cre/Cre mice intravitreally transduced with AAV expressing Opn4^WT or Opn4 variants. A few days after transduction (red arrows), functional expression of Opn4 mediates photoentrainment of daily activity to the light:dark cycle. rd;Opn4^−/− and rd;Opn4^cre/cre mice retinas transduced with melanopsin are recorded on MEA. (B) Proportion of recorded cells responding to 100-ms, or 10-sec stimulations. Data for 100 ms, 1 sec, 10 sec, and 60 sec are shown in (C). Light responses are shown for rd;Opn4 ^−/− retinas expressing WT (Opn4^WT, n=57) and altered melanopsins, including truncated (Opn4^380Δ, n=33; Opn4^397Δ, n=45), phospho-null mutations (Opn4^2A, n=103; Opn4^4A, n=100; Opn4^9A, n=125) and phosphomimetic mutations (Opn4^2D, n=61,). Average traces (D, E), response latency (F, G), and response duration (H, I) are shown. Recordings (MEA and PLR) in response to monochromatic light stimulations of increasing duration (480 nm, 5.10¹² photons/cm²·s, 100 ms, 1 sec, 10 sec, and 60 sec in MEA and 480 nm, 1.10¹⁴ photons/cm²·s, 1 sec, and 60 sec in PLR), average values ± s.e.m. (F, G, H, I) For detailed statistics see Supplemental Table 1.

Figure 4. Altered response properties translate in vivo

ipRGCs responses to light.(A): rd;Opn4 ^cre/cre mice conditionally infected with floxed Opn4^WT, n=54 or altered melanopsins; Opn4^397Δ n=38, Opn4^9A, n=36 and Opn4^9A397Δ n=29) are mirrored by pupillary constriction in animal transduced with the same type of mutants (B: PLR average traces for Opn4^WT, n=4; Opn4^397Δ, n=5; Opn4^9A, n=2; and Opn4^9A397Δ, n=6; and C: representative pictures of the different phases of PLR).

Figure 5. Response sensitivity to light is affected by alterations to the melanopsin C-terminus tail

rd;Opn4 ^cre/cre mice retinas conditionally expressing melanopsin variants are recorded on MEA. (A, B, and C) Average responses of ipRGCs to 1-min stimulations of increasing irradiance (Opn4^WT, n=35; Opn4^397Δ, n=47; Opn4^9A, n=42). (D and E) Dose response curves associated with response duration and the number of spikes. Data have been fit with the appropriate function (sigmoidal or linear). Correlation coefficients for duration and number of spikes are: 0.9 and 0.96 for Opn4^WT, 0.95 and 0.96 for Opn4^9A, and 0.94 and 0.91 for Opn4^397Δ. Recordings (MEA) were performed in response to monochromatic light stimulations of increasing irradiance (480 nm, 60 sec, 5.10¹⁰ photons/cm²·s to 5.10¹³ photons/cm²·s), average values ± s.e.m.

Figure 6

(A) Distribution of the persistence (delay between light off and return to baseline) through increasing irradiances. At all irradiances, Opn4^9A displays a large range of persistence durations while Opn4^397Δ transduced cells display very little persistence of the response irrespective of the stimulation irradiance (In italic, number of responses out of boundaries of the recording). (B) Deactivation speed (plateau/tonic response (last 10 sec before light off) to the baseline discharge rate divided by the delay between the light off and the return to baseline) after 1-min stimulations of increasing irradiance. Opn4^WT deactivation speed remained constant independent of the irradiance of the stimulation (Opn4^WT linear regression slope not different from 0, p=0.1046), whereas it increased for Opn4^397Δ (MANOVA, F=91.82 (2,8), p<0.001).

Figure 7. Proposed melanopsin deactivation model

Active melanopsin (Ma, and/or Extra-melanopsin E) is photoisomerized into activated melanopsin (M*), the signaling form, upon photon absorption. Phosphorylation of M* induces its binding by arrestins and subsequent deactivation (Mi, inactive melanopsin) and recycling/regeneration. Absorption of photon from another wavelength may photo regenerate directly M* into Ma/E forms.