Dissection of the subcellular eye of microorganisms called warnowiid dinoflagellates reveals that this structure is composed of elements of two cellular organelles — the plastid and the mitochondrion. See Letter p.204
The ancient Greek physician Galen described the key anatomical features of the eye1, including the retina, lens, cornea and iris. Yet arguably the first true understanding of how the vertebrate eye works came in the early seventeenth century, with mathematician Johannes Kepler's demonstration that vision occurs as an image projected on to the surface of the retina2. As such, an eye can be defined as a cornea and/or a lens that forms an aperture allowing light arising from a specific direction to pass on to a sensory surface that processes this signal into a chemical message. But animals were not the only organisms to evolve such systems — analogous structures and biochemical responses exist in cells of several eukaryotic microorganisms (cells that package most of their DNA in a nucleus), allowing these microbes to move in response to light3. On page 204 of this issue, Gavelis et al.4 describe the subcellular features that make up the eye-like structures of warnowiid dinoflagellates, which in anatomical terms are remarkably similar to vertebrate eyes.
Warnowiid dinoflagellates are unicellular plankton that have not been cultured in the laboratory, but that are known to possess a remarkably complex eye-like structure, called the ocelloid. Ocelloids consist of distinct components similar to key parts of vertebrate 'camera-type' eyes: a cornea, a lens (called a hyalosome) and a pigmented cup or retina-like structure. Gavelis et al. studied warnowiids isolated from marine waters in Japan and Canada, and demonstrate that the anatomy of ocelloids is built from reconfigured plastids and mitochondria (Fig. 1a). These are subcellular compartments seen in many eukaryotic groups that formed in the distant past through the intracellular incorporation of symbiotic bacteria; these organelles usually contain their own genomes and typically function in energy transformation.
Specifically, Gavelis and colleagues show that the retinal body of ocelloids arises from a membrane network derived from plastids, and that multiple mitochondria form a cornea-like surface across a lens structure. To test these microscopy-based observations, the authors microdissected the warnowiid retinal body and sequenced its DNA, which contained a much higher proportion of DNA of plastid origin than equivalent samples from the whole cell.
Although ocelloids are exceptionally complex, warnowiids are not the only microbial cells with eye-like subcellular structures. A diversity of eukaryotic microorganisms perceive light using different kinds of eyespots. One such structure is the eyespot of the green alga Chlamydomonas reinhardtii (Fig. 1b), a unicellular relative of land plants. This eyespot is located at the edge of the alga's plastid and is made up of lipid globules, rich in orange carotenoid pigments, that are stacked in compartments inside the plastid envelope. As such, this globule layer is thought to provide directionality and contrast by shielding and reflecting light from one side of the organism on to two light-sensitive proteins called type 1 rhodopsins that localize with this eyespot5,6,7. These two proteins have intrinsic light-gated cation-channel activity (and are therefore named channelrhodopsins) and have been demonstrated to act as photoreceptors that trigger movement in response to light5,6,7.
Cryptophyte algae such as Guillardia theta also build eyespot structures that are located in plastids8, and movement of these cells in response to light is mediated by the function of at least two type 1 rhodopsin proteins9, similar to Chlamydomonas. The alga Euglena gracilis also has an orange-red eyespot, although, in contrast to the previous examples, this structure is associated with the base of the flagellum3, the cells' swimming propeller. The photoreceptor in Euglena has been identified as a photoactivated adenylyl cyclase10 protein.
In yet another branch of the tree of life are the eyespot-like structures of the swimming spores of Blastocladiomycota fungi (Fig. 1c). These structures are lipid-filled vesicles called side-body complexes that are located close to the large mitochondrion of these fungal cells11. The side-body complex is overlaid with type 1 rhodopsin proteins. In Blastocladiella emersonii, the type 1 rhodopsin photosensor contains a guanylyl cyclase domain, which allows the protein to control the production of cyclic GMP (ref. 12), a key chemical messenger in vertebrate vision. Recent work13 on warnowiid ocelloids has also suggested that messenger RNA encoding a type 1 rhodopsin is associated with the retinal body.
These examples demonstrate the wealth of subcellular structures and associated light-receptor proteins across diverse microbial groups. Indeed, all of these examples represent distinct evolutionary branches in separate major groups of eukaryotes3. Even the plastid-associated eyespots are unlikely to be the product of direct vertical evolution, because the Chlamydomonas plastid is derived from a primary endosymbiosis and assimilation of a cyanobacterium, whereas the Guillardia plastid is derived from a secondary endosymbiosis in which the plastid was acquired 'second-hand' by intracellular incorporation of a red alga14. Using gene sequences recovered from the warnowiid retinal body, Gavelis et al. investigated the ancestry of this organelle by building phylogenetic trees for the plastid-derived genes. Their analysis demonstrated that this modified plastid is also of secondary endosymbiotic origin from a red alga.
Although derived independently, there are common themes in the evolution of these eye-like structures. Many of them involve the reconfiguration of cellular membrane systems to produce an opaque body proximal to a sensory surface, a surface that in four of the five examples probably involves type 1 rhodopsins. Given the evolutionary derivation of these systems, this represents a complex case of convergent evolution, in which photo-responsive subcellular systems are built up separately from similar components to achieve similar functions. The ocelloid example is striking because it demonstrates a peak in subcellular complexity achieved through repurposing multiple components. Collectively, these findings show that evolution has stumbled on similar solutions to perceiving light time and time again.