Abstract
The Drosophila compound eye is made of 800 unit eyes (ommatidia) that each contains eight photoreceptors: Six (R1-R6) are involved in motion detection while two (R7 and R8) are required for color vision. Ommatidia belong to two subsets distributed stochastically throughout the retina in a 30:70 ratio: p-type ommatidia contain UV-sensitive Rh3 in photoreceptors R7 and Blue-Rh5 in R8 while y-type express another UV-Rh4 in R7 and green-Rh6 in R8. Comparison between the outputs of R7 and R8, and between p and y ommatidia allows flies to discriminate between colors.
I will describe the cascade of genes that specify the different subsets of photoreceptors through a series of fate restrictions and how this cascade is modified to define the various regions of the retina in Drosophila. In particular, spineless is expressed in a stochastic manner in R7 cells that express Rh4 (yR7). It allows the specification of the whole retina by specifying the y choice in R7 and allowing R7 to instruct R8 of its choice.
Processing of color information occurs in the medulla part of the optic lobe. It is formed by ∼40,000 neurons surrounding a neuropil where photoreceptors and medulla neurons interconnect. Associated with each set of R7/R8 projections, there are ∼800 ‘columns’, the functional units in the medulla. We are addressing how medulla cells process color information coming from R7 (sensitive to UV) and R8 (sensitive to blue or green) and send it to higher processing centers in the lobula complex and central brain to mediate color behavior.
We are silencing subsets of medulla neurons using specific Gal4 lines and testing the consequence for color discrimination. We have adapted to color vision the flight simulator originally designed by the Dickinson/Frye labs. In an operant paradigm, the fly is trained to associate color with a reward or punishment before being tested in the absence of the reward.