Building Color Vision in the Fruitfly Eye

By Angela DePace

What can we learn about how things function simply by looking at how they are built? Often quite a lot, if we can filter out parts selectively and focus on one subsystem at a time. Think of looking at a building and seeing only the electrical wiring. The wiring is invisible in the whole structure, but even if you could peel away the walls, it would be buried in the complexity of other components. Yet seeing it in isolation could tell us about how power is routed to the building, how it is distributed, and even why lights on the fourth floor always flicker when you microwave your coffee.

Fluorescent microscopy allows biologists to do just that — focus on particular cell types and individual molecules in order to observe structures and formulate principles about how they are built. Dr. Claude Desplan of New York University studies the architecture and process of color vision, using fruitflies as a model. His fluorescent images have revealed remarkable similarities between the human and fruitfly visual systems. Both separate color from black and white (or motion) vision processing and distribute color photoreceptors randomly throughout the retina, but in constant proportion.

We asked Dr. Desplan about his image.

Why did you create this particular image of the entire visual system?

This image was taken by Franck Pichaud, who was a postdoc in the lab and is now a group leader at University College London. It only focuses on photoreceptors, ignoring all other cell types. The purpose was to focus on the primary cells of the visual system and to see how they connect (or “project”) to two different locations. A subset projects to the first “optic ganglion” called the lamina, whose purpose is to process motion detection (flies are extremely sensitive to motion). Another subset of photoreceptors goes through the lamina without stopping and projects to the second optic ganglion, the “medulla,” where color vision information is processed.

How was the image made?

The image was made by dissecting the entire visual system out of a fly head, fluorescently staining it with an antibody that only recognizes photoreceptors, then imaging it with a Zeiss LSM 510 confocal microscope with 3D reconstruction. In order to cover the entire system, six overlapping images were taken and juxtaposed using the tiling software from Zeiss.

The image is incredibly lovely; what also makes it scientifically successful?

The image beautifully communicates the separation between the motion detection and color vision neural processing pathways, as well as the crystalline organization of the system (called “retinotopic organization,” where the world is projected as an image onto the retina and further down to the optic ganglia). Its effectiveness comes from its relative simplicity as well as its highly ordered organization.

What feature of the visual system is revealed in the second image?

This “whole-mounted” retina was stained with antibodies against blue and green photopigments, and shows two things. One, these two photopigments are expressed in a mutually exclusive manner in two subclasses of photoreceptors. Two, it shows that there is no order in the distribution of the two classes of photoreceptor cells. There is no spatial pattern, only a relatively constant 30 to 70 percent ratio of blue to green photoreceptors. How do you create a random distribution like this? And how to do you maintain the ratio of cell types? Vision biologists have been interested in this image, as well as many physicists who are fascinated by the stochastic process.

What future tool or technology for visualizing data would significantly aid in your work?

Live imaging is what we need most: this means molecules that can be traced without damaging tissue, and microscopes that can penetrate deep into samples.

Are there any scientific disciplines that would benefit from each other’s approach to visually representing data?

The Howard Hughes Medical Institute has recently created a new research center, Janelia Farm, whose purpose is to bring together biologists, physicists, and informaticians to address ways to image the brain. It is at the forefront of neurobiology today.