Evolution of Vision / Evolution de la vision

Séminaire FSER organisé par Claude Desplan (University of New York, USA) et Walter Gehring (Biozentrum, Bâle, Suisse) du 14 au 19 octobre 2013

Participants

Detlev Arendt, Nansi Jo Colley, Tiffany Cook, Claude Desplan, Markus Friedrich, Walter Gehring, Takashi Gojobori, William Harris, Samer Hattar, Gaspar Jekely, Zbynek Kozmik, Michael F. Land, Dan-E. Nilsson, David C.Plachetzki, Ilaria Rebay, Sylvie Rétaux, Simon Sprecher, Jonathan Tang, Veronica van Heyningen, Shozo Yokoyama

—

Twenty experts from Europe, the USA and Japan met to explore the evolution of developmental, physiological and clinical aspects of vision. They were chosen as to represent a broad variety of fields and to induce interdisciplinary discussions. All participants participated to the entire meeting and contributed extensively to its interactive nature.

David Plachetzki (University of California – Davis, USA) began the opening session on early evolution by discussing his work on the origins of phototransduction. Analysis of the cnidarian hydra support the hypothesis that CNG-based ciliary phototransduction was present in the common ancestor of cnidarian and bilaterian at least 600 million years ago. The reason for opsin-mediated phototransduction in eyeless cnidarians was discussed, with cnidocyte discharge in hydra shown to be a behavioral output of photoreception. A phylogenetic timeframe for the origin of animal sensory systems was discussed, showing that cryptochromes appear to make up the oldest sensory visual modality in animals.

Gaspar Jekely (Max Planck Institute-Tubingen, Germany) presented his work on the annelid Platynereis dumerilii, which has a proto-eye and is an ideal model for studying the mechanism and origin of visually-guided behavior. The ciliated larva shows phototaxic response to light via ciliary beating, whereas the adult undergoes body bending. Cell ablation, behavior assay, modeling, and electron microscopy studies revealed mechanisms by which Platynereis convert light information into motor output at these stages; reciprocal inhibition was found as a cellular mechanism to enhance contrast in sensing light direction.

Dan E. Nilsson (Lund University, Sweden) ended the session with a scheme for stepwise evolution of complex eyes. He first described the diversity of light-sensing systems that exist across phyla. Evolution of increasingly complex eyes can be categorized into 4 sequential steps: nondirectional photoreception, directional photoreception, low-resolution vision and high-resolution vision. Physical considerations suggest that each transition step was associated with a corresponding key innovation including efficient photopigments, directionality through screening pigments, photoreceptor membrane stacking and focused optics. Evolution of photoreceptive capabilities has emerged independently in evolution. By the Cambrian evolution, eyes of all types of complexity had been established.

Zbynek Kozmik (Institute of Molecular Genetics, Czech Republic) opened the session on Pax6. Similar to vertebrates, the box jellyfish Tripedalia cystophora, contains camera, lens-containing eyes with ciliary photoreceptors. Photobehavior was found to depend on opsins via Gαs signaling, revealing a common phototransduction mechanism between the lens-containing eyes of Cnidaria and Bilateria. Discussion then shifted to the multiple roles of Pax6 in mouse retinal development. Conditional knockout experiments showed that Pax6 is important for lens formation early on, but later regulates retinal progenitor cell proliferation, differentiation, and pigmentation of the eye. Pax6 was proposed to regulate differentiation genes in simple eyes, and to have been co-opted to control development genes in complex eyes.

Sylvie Retaux (Institut Alfred Fessard, France) presented the loss of vision in cave dwelling Astyanax mexicanus, a river-dwelling fish with interfertile strains adapted to the surface or cave lifestyles. To address how the cavefish strains lost their eyes over evolution, Retaux discussed genetic, developmental biology and behavioral studies that showed how regression of the eye could be explained by a combination of positive selection, either directly or indirectly, and genetic drift. The unexpected preservation of cavefish eye during development may be due to developmental constraint selecting for non-eye related features.

Markus Friedrich (Wayne State University, USA) discussed the eye selector gene network in the flour beetle Tribolium. RNAi knockdown suggest a genetic redundancy between ey/toy/dac in adult, but not in larval eye development. Pax6 was found to be essential for initiation of embryonic head lobe and to play an early role in compartmentalization of anterior head. This suggests that Pax6 affects development of the visual system as an indirect consequence of its early role in regional patterning of the proencephalon. Further, since Pax6 influences development of head regions outside of the visual system, Friedrich proposed that Pax6 is a master regulator of anterior head lobe development, but competent factor for eye development.

The October 16th session on Development began with a talk by Ilaria Rebay (University of Chicago, USA). She discussed the general issue of how cell fates are accurately specified, using the example of the EGF receptor network, focusing on Yan that affects even-skipped expression during embryonic development in Drosophila. This case illustrates the importance of precision and robustness of bistable switch network for carrying out precise and accurate transcription repression of cell fate. Second, the issue of how cell fate is maintained as terminal differentiation proceed was discussed, using the case of Abelson kinase in the development of the fly eye. Photoreceptor cell fate appears to not stabilized even late into retinal development and needs to be actively maintained.

Veronica van Heyningen (MRC Human Genetics Unit, UK): discussed developmental eye anomalies and evolutionary insights derived from them. Transcription factors such as PAX-6, SOX2 and OTX2 are implicated in human eye diseases such as aniridia, micropthalmia and anophthalmia. Eye defects could be due to disruptions to gene expression patterns, gene autoregulation, or protein-protein interactions, and showed the importance of maintaining proper network topology to ensure developmental robustness. Evolution of Pax-6 genes in zebrafish was discussed to show paralog divergence by subfunctionalisation.

William Harris (Cambridge University, UK) began by discussing how retinal progenitor cells (RPC) generate the complete repertoire of cell types in the mature vertebrate retina. Using lineage-tracing assays in the zebrafish, Harris’s group followed the progression of clones generated by RPCs in vivo, providing a complete quantitative map of retina generation; this led to a model in which equipotent RPC stochastically generate the right proportion of retinal cell types. Then, the discussion turned to the issue of how the fish retina continues to grow in size; retinal stem cells in the ciliary marginal zone contribute to the additional cells over time, via predominately asymmetric cell divisions.

Detlev Arendt (EMBL, Germany) began the afternoon session on Behavior. The flight from light theory posits photoreception to blue light evolved in aquatic organisms to mediate escape from UV radiations during daytime, via vertical migration. Arendt presented data showing a link between circadian system and locomotion needed for diel vertical migration in the annelid Platynereis. Melatonin is present in the annelid and production increases at night, resulting in changes in ciliary beating of Platynereis larvae. The annelid photoreceptive center in the brain, which also produces melatonin, further expresses marker genes specific to vertebrate pineal and retina, altogether suggesting a common origin of circadian centers in Urbilaterian.

Nansi Colley (University of Wisconsin-Madison, USA): Protein trafficking defects are implicated in many retinal diseases. Colley discussed two genes in the rhodopsin trafficking pathway that, when disrupted, lead to retinal degeneration in the fruitfly. First, the xport chaperone gene is required for TRP and Rhodopsin targeting to the membrane. Second, the Golgi SNARE protein Gos28 was found to be essential for rhodopsin trafficking in the secretory pathway. As Xport and Gos28 act specifically on the Rh1 rhodopsin, it remains to be determined why they are specifically trafficked, and how other (color) Rhodopsins are trafficked.

Samer Hattar (Johns Hopkins University, USA) discussed the multiple functions of intrinsically photosensitive retinal ganglion cells (ipRGCs), retina cells that sense light via melanopsin, a rhabdomeric opsin, and regulate daily changes in physiology and behavior. First, ipRGCs provide separate non-image forming pathways into the brain to independently control pupillary light reflex and photo-entrainment. Second, light input conveyed by ipRGCs during development help set the circadian period in the animal. Third, ipRGCs directly participate in pupillary light reflex of the ipsilateral eye, via direct innervation of the iris muscle. Lastly, ipRGCs controls animal mood and learning, but via mechanisms independent of circadian clock function or sleep duration.

After a morning in which participants spent toured the area surrounding The Fondation des Treilles, Claude Desplan (New York University, USA) began the October 17th section on Patterning. He discussed how stochastic expression of the transcription factor Spineless directly influences the choice of color-sensing photoreceptor subtypes in the fly. Each spineless allele is controlled by global activation coupled with random repression via silencer elements. Allele decision is then communicated to the other allele via interchromosomal communication, resulting in robust frequency of expression. In contrast to color-sensing photoreceptors, polarized light-sensing photoreceptors are specified deterministically. Some insects have regular arrays of photoreceptor subtypes and regular spineless expression patterns; comparative approaches could thus show how stochastic/deterministic patterns evolved in insect eyes.

Tiffany Cook (University of Cincinnati, USA) discussed how the Drosophila cone (lens) cells play a role not only in secreting the lens but also have glial-like functions in eye development and during the physiological response. The decision to make either a photoreceptor or a cone cell is governed by antagonistic action of Notch and Ras signaling, whose downstream effectors are dPax2 and Prospero, respectively. dPax2 and Prospero have homologs in vertebrates that also participate in lens formation, suggesting an evolutionary conserved function for these genes. Cone cells have diverse roles mediated by Prospero, including patterning of the ommatidium, as well as functional support for photoreceptors. Cone cells may be analogous to Muller glial cells of the vertebrate retina, with similar gene expression profiles.

Michael Land (University of Sussex, UK) ended the day with a talk on the evolution of eye movement. To gain information about objects or to maintain gaze on an image during motion, many animals adopt a ‘saccade and fixate’ eye movement strategy. This strategy is used by vertebrates such as fish and humans, and by invertebrates including cephalopod molluscs, crabs and many insects. Saccade and fixation evolved to provide photoreceptors the necessary time to respond to light intensity changes; motion blur sets in if the rate of image movement is too fast. The ‘saccade and fixation’ strategy is not universal; some invertebrates, such as heteropod sea snails and diving beetle larvae, have eyes with retinas arranged in a long narrow strip. Scanning visual scenes at right angles to the strip thus retrieves information about the 2nd dimension.

Jonathan Tang (Harvard Medical School, USA) began the October 19th morning session on Function. He described an issue arising in the mouse community: thousands of transgenic GFP lines have been generated and characterized for GFP expression, but these tools are limited in use to labeling purposes. To address this, Tang built synthetic systems to co-opt GFP as a scaffold for controlling transcription or recombination, enabling transgenic GFP lines to be used as tools for manipulating genes in GFP-labeled cells. This technology may be used to study the many retinal cell types that are selectively labeled in GFP lines.

Shozo Yokoyama (Emory University, USA) discussed approaches for studying adaptive evolution of visual systems. Yokoyama argued that the molecular basis of spectral tuning in visual pigments is best studied by examining how different species have adapted to various light environments. Studies on the evolutionary mechanisms producing various dim-light and color vision have been conducted by manipulating present-day visual pigments. Yokoyama stressed the caveats with these studies and the need to manipulate engineered ancestral pigments. It is important to unambiguously establish the relationship between genotype and phenotype, to engineer and manipulate proper ancestral phenotypes, and to clarify the precise roles of adaptive mutations in shaping the chemical structures of proteins. To achieve this, quantum chemical methods can be used in the visual system.

Simon Sprecher (University of Fribourg, Switzerland) presented his work on vision in the Drosophila larvae, which is advantageous for its relatively simple visual system. Specification of photoreceptor subtypes fate in the larvae depends on a Senseless-mediated feedforward loop. The transition from larval to adult photoreceptors is mediated by ecdysone signaling, leading to loss of one photoreceptor subtype and subtype switch of the remaining to form the eyelet involved in circadian entrainment. In adult eyelets, Senseless acts with the transcription factor Hazy in specification of rhodopsin fate, highlighting its multiple roles in adult and larval eyes. Functionally, larvae can navigate and learn using light as a cue; genetic and connectomic dissection of these processes was discussed.

Takashi Gojobori (National Institute of Genetics, Japan) began the last session of the meeting, on evolution of vision from bacteria to humans. First, he showed how high-throughput expression analyses could be applied in the planaria, hydra and sea urchin to draw insights regarding the ancestral characteristics and evolution of the nervous system. Second, comparative transcriptomics was used to reveal that the octopus and human eye share a large number of expressed genes, but whether this is evidence of convergent or divergent evolution remains up for debate. Third, Gojobori discussed the possibility that the camera eye of the ocelloid bearing dinoflagellate evolved as a result of horizontal gene transfer from light-sensing bacteria. Lastly, recent attempts to characterize the diversity of marine microorganisms were presented.

Walter Gehring (University of Basel, Switzerland): closed the meeting by continuing the focus on evolution of vision from cyanobacteria to humans. Different eye types across phyla were once assumed to be of a polyphyletic origin, but Gehring’s group showed that the master control gene Pax6 could direct eye formation from mouse to flies. Genes such as Eya, Six and Dac have conserved roles in eye development as well, arguing for a monophyletic origin of the eye. Divergence of eye types from the prototypic eye possibly involved intercalation of genes between the master control gene (Pax6) and structural genes such as rhodopsin, via mechanisms such as gene duplication and enhancer fusion. Gehring speculated on the origin of photoreceptor cells by discussing an endosymbiotic mechanism by which metazoan could acquire their photosensitivity. Such a scheme involves incorporation of photosynthetic cyanobacteria by red algae, then taken up by dinoflagellate, then taken up by cnidarians.

Active discussions accompanied each presentation as well as over breaks, lunches and dinners. The meeting ended on the night of October 19th, with many hugs and handshakes given to finish off a terrific meeting at a breathtaking location.