From a random mosaic to a map
Nearly every insect eye distributes its color photoreceptor types at random. Mosquitoes do not. The Aedes aegypti retina is divided into fixed domains, each with its own photoreceptor composition: a dorsal rim area, a narrow ventral stripe, a dorsal-central field, and a ventral field. The developmental decision that is stochastic in flies and butterflies has been converted into a deterministic spatial pattern.
We are working out how. Our evidence points to regionalized expression of specific retina patterning factors that define four different territories. Because Culex and Anopheles mosquitoes have regionalized their retinas in different ways, comparing the three genera lets us separate the shared logic from lineage-specific detail. All three are reared in the lab.
The second half of the project asks what the map is for. Aedes aegypti finds hosts and egg-laying sites using vision as well as smell, and the ventral stripe sits where the horizon falls during flight. We are building transgenic lines that reshape regional identity, so that specific domains can be expanded, removed, or converted, and in collaboration with Jeff Riffell (University of Washington) the goal is to test how each region contributes to host-seeking and oviposition. Serial block-face electron microscopy of the stripe (with Kentaro Arikawa) and single-cell recordings of spectral and polarization sensitivity (with Gregor Belušič) tie the developmental map to its physiology, and we have joined the Aedes brain connectome consortium to follow the regions into the brain.
Aedes aegypti transmits dengue, Zika, chikungunya, and yellow fever. Understanding how its visual system is built, and how vision guides the behaviors that bring it to a host, is foundational knowledge for vector biology as well as a clean test of how a developmental mechanism can be rewritten.
The dorsal rim area
One region is shared by nearly every insect eye: a dorsal rim area (DRA), a strip of ommatidia along the top of the eye specialized for detecting the polarization of skylight and used for orientation. In Drosophila, the transcription factor Homothorax specifies DRA fate. With Mathias Wernet’s lab we found that this is not the ancestral arrangement: Homothorax is absent from dorsal rim photoreceptors in mosquitoes, butterflies, and honeybees, and acts there only in the higher flies (Roth et al., iScience 2026). The region itself is ancient, suggesting Homothorax may have been inserted into an older regulatory network as an intermediate step, much as Spineless sits on top of the ancestral eye ground plan. We are using our retina single-cell data from Aedes, butterflies, and crickets to identify the regulators that act in its place, and are testing the leading candidate by HCR in situ hybridization and CRISPR knockout. The larger question is how a gene regulatory network gains layers and becomes modular.
People: Julia Ainsworth (lead, retinal regionalization), Takami Sho (transgenics), Tiezheng Fan and Ke Gao (dorsal rim area), and undergraduate researchers on the connectome annotation. Collaborators: Jeff Riffell (University of Washington), Omar Akbari (UC San Diego), Mathias Wernet, Kentaro Arikawa, Gregor Belušič.
Key papers: Gao et al., PNAS 2025 (comparative eye ground plan, including Aedes); Roth et al., iScience 2026 (Homothorax and the dorsal rim area across insects, with the Wernet lab).
Funding: NIH NHGRI R01HG013634 (EDGE program) supports the comparative and developmental work.
