Switching from stochastic to deterministic patterning in fly retinas: mechanisms and behavioral significance
Developmental Neurobiology
Evolutionary Cell and Developmental Biology (Zoology)
Molecular Biology and Physiology of Neurons and Glial Cells
Final Report Abstract
Developmental programs rely on strictly controlled mechanisms that ensure highly reproducible outcomes and reduce variation from environmental noise. In many sensory systems, however, patterning processes often produce stochastic spatial distributions of sensory receptors. This is true for mammalian eye, as well as for Drosophila eye, the best-studied example of this phenomenon. In stark contrast, species of the Dolichopodidae (Doli) family produce a deterministic pattern consisting of alternating vertical stripes. Doli are visual predators that hunt for small insects on leaves or near the water surface. The set of highly specialized polarization sensitive photoreceptors (PRs) in Doli eyes filter out horizontally polarized light that is reflected by water surfaces or shiny leaves. This filter enhances contrast for possible prey detection. The function of the particular spatial arrangement in stripes, however, remains unknown. I constructed a 3D model of the Doli retina based on a detailed analysis of the morphology and the Rhodopsin (Rh) expression pattern across the retina. I found paired expression of Rh3 and Rh5 in one ommatidial subtype and Rh4 and Rh6 in a second subtype, as in Drosophila. In contrast to Drosophila, where only a small number of ommatidia have polarization-sensitive photoreceptors (PR), the morphology of Doli PRs suggests polarization sensitivity across the entire eye. The preference for either horizontally or vertically oriented polarized light changes in every other column along with the changing Rh expression. Notably, in the dorsal part of the retina an atypical ommatidial type is intermingled stochastically between the highly ordered ommatidia described above. This third ommatidial type resembles the socalled love spot type in its morphology and Rh expression. In many fly species this modified type is specialized in motion vision, which suggests a role in detecting and chasing potential mates. To understand the origin of the deterministic pattern, I examined whether the columns align with the developmental progress of the morphogenetic furrow (MF), a wave of differentiation that progresses across the eye imaginal disc from posterior to anterior. In Drosophila, the equator that divides the eye into two halves is orthogonal to the progressing wave. In Doli, however, the eye equator is much more dorsally located and on an oblique angle, thus not likely to serve as an appropriate reference for the MF that progresses in a dorsoventral band parallel to the stripes. Despite the drastic difference in patterning of the two fly eyes, the underlying regulatory machinery used in both systems is likely to be conserved. Indeed, key regulators responsible for the stochastic patterning in Drosophila are expressed in Doli in the same deterministic pattern that creates the retinal stripes. These regulators are best studied during development. An important breakthrough in this regard has been the progress in rearing the Doli larvae. Moreover, experiments with transgenic Drosophila showed that regulatory elements of key controllers are conserved. Future studies should focus on candidate signals from the MF that cause the formation of stripes in Doli in a pulse-like manner and why the Drosophila regulatory network does not respond to them.