Mixed Palettes of Melanopsin Phototransduction

Abstract

Animal photoreceptors divide into two fundamental classes, ciliary and rhabdomeric. Jiang and colleagues demonstrate that this boundary is disregarded by the intrinsically photosensitive retinal ganglion cells of mammals. These neurons draw from phototransduction mechanisms of both classes, enriching the signals that they produce to drive a diversity of visual functions.

Animals use light to recognize objects, guide actions, regulate physiology, and tune development. These processes are triggered by photoreceptors of two basic classes. Ciliary photoreceptors include the familiar rods and cones of vertebrates, not to mention cells in the avian pineal gland and reptilian third eye. Their rhabdomeric counterparts are exemplified by photoreceptors of invertebrates like Drosophila, which detect light with remarkable sensitivity, dynamic range, and speed. These classes diverge in the G-protein cascades that mediate phototransduction: a motif based on cyclic nucleotides signifies the former and phospholipase C (PLC) the latter (Yau and Hardie, 2009). Jiang and colleagues now present compelling evidence that this divide is bridged by the intrinsically photosensitive retinal ganglion cells (RGCs; Do and Yau, 2010). These mammalian neurons appear capable of using both phototransduction motifs.

To capture light, intrinsically photosensitive RGCs express a visual pigment called melanopsin. Melanopsin phototransduction causes the membrane voltage to depolarize (becoming more positive), evoking electrical spikes that are conveyed to dozens of brain areas. Intrinsically photosensitive RGCs are key mediators of functions that include regulation of the circadian clock, pupil response, sleep, and melatonin level. Ablating these cells spares visual perception but renders many aspects of physiology, such as the clock, completely insensitive to light. Conversely, when rods and cones are lost but intrinsically photosensitive RGCs remain, light regulates physiology even in the absence of visual awareness.

Intrinsically photosensitive RGCs are best understood in the mouse retina, where they are subdivided into several types (M1 – M5; Schmidt et al., 2011). Molecularly, the M1 type is the spitting image of a rhabdomeric photoreceptor. Its cascade uses G proteins of the q/11 family, PLCβ4, and TrpC6/TrpC7 ion channels; these are closely related to elements of Drosophila phototransduction. For the last decade, the other types have been assumed to be of the same mold because they also use melanopsin and depolarize to light.

Jiang at al. overturn this assumption to provide new insights into phototransduction. Their principal approach is to record the electrical responses generated by single intrinsically photosensitive RGCs of different types, using mice that lack genes for various signaling molecules. They find that light elevates cyclic nucleotides in M4s to gate ion channels—a decidedly ciliary event. Fittingly, early evidence of these particular channels (dubbed “HCN” for hyperpolarization-activated, cyclic-nucleotide-gated) was found in rods, though they do not mediate phototransduction there. Another surprise lies with the M2s. Unlike any other photoreceptor observed to date, these cells blend rhabdomeric and ciliary motifs. Intriguingly, it is speculated that the common ancestor of extant photoreceptors possessed a mixed nature (Yau and Hardie, 2009).

Impressive experiments are on display. For example, testing for an elevation of cyclic nucleotides within M4s is not trivial. Their small number (<0.01% of all retinal cells) makes biochemistry foreboding and their photosensitivity is incompatible with fluorescent indicators. Jiang et al. met this challenge cleverly, by expressing in M4s an ion channel that is opened specifically by cyclic nucleotides. They selected a channel that opens faster than the HCN channel, allowing its activity to be distinguished from the natural light response. Indeed, illumination revealed telltale signs of this exogenous channel and thus of a rise in cyclic nucleotides. Another example is the evaluation of TrpC channels, where the authors guide brute force with a steady hand. They generate a mouse line bearing twelve mutant alleles and a reporter transgene, which eliminates this channel family while making intrinsically photosensitive RGCs identifiable, then carefully examine the light responses that remain. This is an uncommon feat—one of many here that indicate the existence of unexpected flexibility in phototransduction.

The work demystifies early findings concerning melanopsin’s promiscuity. It is a rhabdomeric pigment that nevertheless activates the G protein of rods with efficiency (Newman et al., 2003), and produces photosensitivity in practically any cell type that is made to express it. This indiscriminate character underlies its usefulness in optogenetic applications (Lin et al., 2008) but has seemed uncanny. It now appears purposeful, allowing melanopsin to activate different cascades in native cells.

What lies ahead? Most pressing is to explain disparities with parallel work that reached a starkly different conclusion (Sonoda et al., 2018). Where Jiang et al. observed no effect on M4s of deleting G_q/11/14 or PLC, Sonoda et al. found that artificial activation or pharmacological antagonism of this pathway could mimic or partially block phototransduction, respectively. Where Jiang et al. measured activation of HCN channels, Sonoda et al. documented the closure of potassium channels. Do M4s use a ciliary motif to open a depolarizing channel or a rhabdomeric motif to close a hyperpolarizing one?

There are clues for reconciliation. One is that Jiang et al. experimented nearer to body temperature and employed a recording solution that was richer in Mg-ATP, which is permissive for HCN channel activity (DiFrancesco, 1993; Pian et al., 2006). Another is that pharmacological block of HCN channels spares a small fraction of the light response, leaving room for a contribution from other channels. Additional experiments are required, particularly to address disagreement about the molecules that operate between the visual pigment and the transduction channel. Answers may be found in disentangling the consequences of pharmacological approaches, favored in the experiments of Sonoda et al., from the molecular-genetic methods that predominate here.

In providing evidence that a single pigment can activate both ciliary and rhabdomeric phototransduction motifs, even within single cells, Jiang et al. break ground for many new explorations. What are the trade-offs of using one motif over the other, and has each intrinsically photosensitive RGC type chosen according to its role? Melanopsin is found in species ranging from lancelets through subterranean mole rats to humans—are its signaling pathways diversified to suit the needs of each? And if merging motifs is advantageous, have other photoreceptors kept this capacity or evolved it anew? With over a thousand opsins identified in the animal kingdom, and a profusion of photoreceptors awaiting characterization, the approach taken here provides an elegant template for future work.

Varieties of phototransduction cascades and ganglion-cell photoreceptors.

Phototransduction in the animal kingdom uses G protein signaling cascades. The ciliary motif that is commonly found in vertebrate photoreceptors, such as rods and cones, involves a modulation of cyclic nucleotides. The rhabdomeric motif—which is typical of the principal, image-forming photoreceptors of invertebrates—employs phospholipase C. The mammalian retina contains photoreceptors that wire directly into the brain to execute diverse functions. These intrinsically photosensitive retinal ganglion cells are heterogeneous. Unusually, they can be ciliary, rhabdomeric, or mixed in nature. These cells are drawn after Estevez et al. (2012).