Ciliary and rhabdomeric photoreceptor

Our results are consistent with the presence of a ratio-chromatic depth-gauge in the planktonic larvae of Platynereis dumerilii. The depth gauge is formed by the antagonistic interaction of two distinct behaviors, mediated by two distinct types of photoreceptor systems with different spectral sensitivities. The rPRCs mediate phototaxis to a broad range of wavelengths between UV and green light, while the cPRCs mediate UV-violet avoidance. Under UV-violet light, after approximately 30 s the UV-violet avoidance response overrides an initial phototactic response. At a set UV/blue ratio, the two opposing behaviors cancel each other out. Due to the differential attenuation of UV and blue light in seawater, this mechanism could allow the larvae to stay at a certain depth in the ocean.

While the function of the rPRCs in the larval and adult eyes has previously been described (Jékely et al., 2008; Randel et al., 2014), here we defined, both physiologically and behaviorally, the function of the cPRCs and unraveled how they interact with the rPRC system.

The cPRCs express a UV-absorbing opsin, have a high resting calcium level and react by hyperpolarization and subsequent depolarization to violet light (405 nm). The cPRCs mediate an UV-violet avoidance response characterized by downward swimming. The action spectrum of this behavior matches the absorption spectrum of c-opsin1 and the response is severely impaired in c-opsin1 knockouts, demonstrating a critical role for c-opsin1 in mediating it. The depolarizing response in the cPRCs shows a similar delayed onset as UV-violet avoidance, suggesting that the behavior is due to cPRC depolarization.

Intriguingly, the depolarization also occurs if the stimulus light is switched off during the hyperpolarizing phase. This implies that cPRC depolarization is not due to direct opsin activation but to a sustained cell-autonomous or non-autonomous signal.

The activation of cPRCs also induces complex differential responses in postsynaptic neurons. The RGW cells follow the activity of cPRCs under violet but not blue illumination, whereas two groups of sensory neurons respond to violet light only (SNearly) or to both violet and, to a lesser extent, blue light (SNlate). This implies that different postsynaptic mechanisms operate in the different cPRC targets. The high calcium level in the cPRCs suggests the possibility of the tonic release of a neurotransmitter (probably acetylcholine) in the dark, similar to the tonic release of glutamate by vertebrate photoreceptors (Heidelberger et al., 2005). Blue illumination (488 nm) also induces cPRC hyperpolarization, but no subsequent depolarization. This response may be mediated by another opsin expressed in the cPRCs. One candidate is c-opsin2 with a maximal absorption at 490 nm (Ayers et al., 2018). One surprising observation is that cPRC hyperpolarization induced by violet or blue light has different downstream consequences. This may be due to the engagement of different signaling cascades by the different opsins.

How cPRC activation and the postsynaptic responses lead to downward swimming during UV-avoidance is unclear. The cPRC circuit connects to the Ser-h1 serotonergic ciliomotor neurons (Verasztó et al., 2017), suggesting that the modulation of ciliary activity may contribute to UV avoidance. In addition, we have genetic evidence implicating neuroendocrine volume transmission in UV avoidance. The cPRCs and their direct postsynaptic targets are part of the neurosecretory brain center of the larva (Williams et al., 2017a), and knocking out one neuroendocrine signaling component leads to strong defects in UV avoidance (unpublished results).

It also remains to be elucidated how the UV response overrides phototaxis at the neuronal level. We found that the cPRC circuit connects to the circuit of the rhabdomeric eyes via the cholinergic RGW interneurons and their distinct downstream synaptic pathways. The RGW neurons provide inputs to the phototactic circuit at the level of the INsn, INpro and the vMN cells. The strongest and most direct connection between the cPRC and rPRC circuits is through synapses from the RGW cells to the INsn neurons. Cholinergic input from the RGW cells to these premotor interneurons may antagonize phototaxis. Other sensory inputs may also tune phototaxis via the RGW-INsn pathway since neuropeptides derived from sensory-neurosecretory cells can influence RGW neuron activity (Williams et al., 2017). The RGW neurons also connect to the putatively mechanosensory MS cells which strongly synapse on the vMNs. UV light may thus tune mechanosensation that may in turn modulate phototaxis.

One outstanding question in eye evolution is why did the ciliary and rhabdomeric PRC types originally evolve (Nilsson, 2013; 2009)? Our findings suggest that the two types may have evolved antagonistic functions early in evolution. Invertebrates then elaborated on the rhabdomeric, and vertebrates on the ciliary type (Arendt et al., 2004). According to one hypothesis, as the brain cPRCs were recruited for vision in the vertebrate lineage, the rPRCs evolved into the retinal ganglion cells (Arendt, 2008; Arendt, 2003). This scenario is supported by the expression of melanopsin, an r-opsin ortholog, in the intrinsically photosensitive retinal ganglion cells (Hattar et al., 2003; 2002; Koyanagi et al., 2005; Lucas et al., 2001; Panda et al., 2003). Although we need more comparative data to test this model, we hypothesize that the cell-type mosaic of the vertebrate retina may have originated from the hierarchical integration of distinct cPRC and rPRC circuits mediating antagonistic behaviors, as observed in Platynereis larvae.