Development and function of Drosophila visual circuits

Chi-Hon Lee, MD, PhD, Head, Unit on Neuronal Connectivity
Shuying Gao, PhD, Postdoctoral Fellow
Songling Huang, PhD, Postdoctoral Fellow
Chun-Yuan Ting, PhD, Postdoctoral Fellow
Meiluen Yang, PhD, Postdoctoral Fellow
Moyi Li, BA, Biological Laboratory Technician

Using the Drosophila color-vision circuitry as a model, we are studying two neurobiology issues: (1) how neurons form complex interconnections during development and (2) how the assembled neural circuits function to guide animal behaviors. The fly retina contains three types of photoreceptors, R1-6, R7, and R8, each responding to a specific spectrum of light and connecting in a retinotopic fashion to a specific layer in the brain. Our studies focus on the connections made by the ultravioletsensitive R7 neurons. In a forward genetic screen, we identified Activin-signaling components by their requirement for establishing a precise retinotopic map of R7 termini. We further uncovered an extrinsic mechanism that is mediated by the protocadherin Flamingo and functions redundantly to the intrinsic mechanism by Activin signaling. To understand the structure and function of colorvision circuits, we are using molecular genetics and serial electronmicroscope (EM) reconstruction to determine the connection patterns of the first-order interneurons. Using quantitative behavioral assays and genetic methods to inactivate specific types of neurons, we are studying the function of the amacrine cell Dm8, which relays inputs from several R7s to projection neurons, thus integrating a large field of columnar elements.

Molecular mechanisms regulating synaptic target selection of R7 photoreceptor axons

Ting, Yang, Li, Lee

In both vertebrates and invertebrates, interneuronal connections are often organized into columns and layers that facilitate information processing and propagating. We use the Drosophila visual system as a model to study circuit assembly and focus on the mechanisms guiding R7 axons into specific layers and columns during development. In a large genetic screen based on R7-dependent behavior, we identified two loci, baboon and importin-alpha3, that are required for the establishment of a precise R7 retinotopic map. Baboon encodes a type I Activin receptor and importin-alpha3 a component of nuclear import machinery. Removal of Baboon or Importin-alpha3 in single R7s caused the cells’ axons to invade neighboring columns, indicating that Baboon and Importin-alpha3 are required cell-autonomously in R7s to restrict their growth cones to retinotopically appropriate columns. In addition, the synaptic boutons of baboon- or importin-alph3–mutant R7s appeared to be smaller and more irregular than those of the wild type, suggesting that these two gene products are involved in synaptogenesis. An examination of other known components of the Activin pathways, including the ligand Activin and the downstream transcription factor Smad2, revealed that the canonical Activin signaling pathway is required for restricting R7 growth cones to their retinotopically appropriate columns. Interestingly, Activin is functionally required in R7s, suggesting that it serves as an autocrine ligand; that is, it is secreted from and acts on R7 growth cones.

Several lines of evidence indicate that Importin-alph3 is a new component of the Activin signaling pathway. First, Smad2 and Importin-alph3 form a physical complex in the growth cones and axons. Second, nuclear accumulation of Smad2 depends on Importin-alpha3. Most important, these observations raise the intriguing possibility that Importin-alpha3 plays a role in the retrograde axonal transport of Smad2. A recent proposal posited a similar role for Importins in transporting transcription factors from axons/dendrites to nuclei, suggesting that Importins’ transport function is conserved in both flies and vertebrates. In summary, our results support a novel model for Activin signaling in R7s, by which autocrine Activin activates Baboon on R7 growth cones, resulting in the phosphorylation of the downstream transcription factor Smad2, which, together with Importinalpha3, is transported from the growth cones back to the nuclei to regulate transcription. This model is further supported by our observation that blocking retrograde axonal transport in R7s pheno-copies importin-alpha3/baboon phenotypes.

The removal of Importin-alpha3 or Baboon resulted in incomplete penetrance of R7 phenotypes, with only 12 to 30 percent of mutant R7 axons invading their neighboring columns. The outcome suggests an additional mechanism that functions redundantly to the Activin signaling pathway in restricting R7 growth cones to their retinotopically appropriate columns. To test whether repulsive interactions among neighboring R7s play a role in restricting R7 termini to appropriate columns, we genetically ablated most of the R7s and examined the targeting of the remaining R7s. We found that wild-type R7 axons form normal synaptic boutons in retinotopically correct columns even in a largely empty R7 terminal field. By contrast, removing neighboring R7s greatly increased the tendency of importin-alpha3– or baboon-mutant R7s to invade adjacent columns. The results suggested that importin-alpha3– and baboon-mutant R7 remained responsive to repulsion by neighboring R7s and that the repulsive interactions accounted for the mutations’ incomplete phenotypic penetrance. To determine the molecular nature of the R7-R7 interactions, we examined several candidate genes whose products are known to mediate repulsive interactions. Among them, we identified the gene encoding the protocadherin Flamingo. Removing Flamingo alone in single R7s did not cause any obvious phenotype, although the invasiveness of baboon mutant R7s was greatly enhanced by the removal of Flamingo in the neighboring R7s. In summary, at least two redundant mechanisms restrict R7 termini to the correct columns: (1) an intrinsic mechanism mediated by autocrine Activin signaling and (2) an extrinsic mechanism involving Flamingo-mediated repulsion among R7s.

Ting C-Y, Herman T, Yonekura S, Gao S, Wang J, Zipursky SL, Lee C-H. R7axons in the Drosophila visual system are restricted to retinotopic targetsboth by mutual repulsion and by transduction of a TGF-beta signal into thenucleus. Neuron 2007;56:793-806.
Ting C-Y, Lee C-H. Visual circuit development. Curr Opin Neurobiol 2007;17:65-72.
Yonekura S, Xu L, Ting C-Y, Lee C-H. Adhesive but not signaling activity of Drosophila N-cadherin is essential for photoreceptor target selection. Dev Biol 2007;304:759-770.

Mapping color-vision circuits

Gao, Huang, Ting, Lee; in collaboration with Meinertzhagen, Wang

To understand how color information is processed in the fly brain, we set out to (1) identify the neurons constituting the chromatic circuits and their synaptic connections and (2) determine the neurons’ functions by using various behavioral paradigms. To achieve the first objective, we identified the first-order interneurons in the medulla ganglion, which receive direct synaptic inputs from the chromatic channels R7 and R8. We exploited the knowledge that the photoreceptor neurons are histaminergic and therefore assumed that their synaptic target must express histamine receptors, which are histamine-gated chloride channels (encoded by the ort and HisCl1 genes). Using an ort-Gal4 driver, we identified the first-order interneurons, which include two projection neurons, Tm5 and Tm9, as well as an amacrine neuron, Dm8. Using serial EM reconstruction, we determined the synaptic circuits of these neurons and found that (1) the amacrine neuron Dm8 receives inputs from several R7s and relays the information to the projection neurons and (2) the projection neurons Tm5 and Tm9 receive direct synaptic inputs from the chromatic channels’ UV-sensing R7s and from the blue/green-sensing R8s, respectively, as well as hitherto unsuspected indirect input from the achromatic channel via L3 (R1–R6). Thus, the projection neurons integrate both chromatic and achromatic channels and relay the information to the higher visual center, the lobula. The integration of several channels at the level of first- or second-order interneurons in flies is similar to that observed in the primate system of color opponency, suggesting a convergent color-coding solution to a common problem in both visual systems.

To determine the function of the first-order interneurons for the second objective, we used a behavioral assay called the spectral preference test. This behavioral paradigm tests a few flies for their innate preference for UV light over green light, a behavior that has been known to depend on the UV-sensing R7s. We found that mutant flies lacking Ort, but not those lacking the other histamine receptor HisCl1, exhibited an aberrant preference for green light. The ort phenotype could be phenocopied by inactivating Ort-expressing neurons, indicating that Ort-expressing neurons are required for UV preference. To determine the specific neurons that mediate UV preference, we subdivided the Ort-expressing neurons into several categories based on their use of neurotransmitters. Using a combinatorial gene expression system called the split-Gal4 system, we found that the Tm5 projection neurons and the Dm8 amacrine neurons are glutaminergic while the Tm9 projection neurons are cholinergic. By restoring Ort expression in specific neuron subsets, we found that the glutaminergic Tm5 and Dm8 neurons are sufficient to drive UV preference, whereas the cholinergic Ort-expression neurons, including Tm9, are sufficient to drive green preference. In addition, inactivating glutaminergic Ort-expressing neurons abolished UV preference, indicating that they are required for UV preference. The results suggest that Tm5 and Dm8 relay the information from the UV-sensing R7s to the lobula and that Tm9 relays blue/green-sensing R8s to the lobula. We further refined the Gal4 expression pattern by dividing the ort promoter into three highly conserved regions: C1, C2, and C3. We found that the glutaminergic Ort-C2 subset, which includes primarily Dm8 neurons, is both required and sufficient for conferring a fly’s preference for UV. Similar to the vertebrate amacrine cells, which link bipolar to retina ganglion cells, Drosophila Dm8 neurons receive their main synaptic input from R7s and provide input to the projection neuron Tm5. The importance of these connections is underscored by the finding that the indirect connections, from R7 to Dm8 and then to Tm5, are both necessary and sufficient to drive UV preferences. We propose that Dm8 sacrifices spatial resolution for sensitivity by relaying signals from many R7s to smallfield projection neurons. We note that Ort-expressing neurons do not include any Dm8-like widefield neurons for R8s and that restoring activity in Tm9 projection neurons is sufficient to confer stronger green preference in ort mutants. Thus, we speculate that Dm8 evolved uniquely to meet the ecological need to detect dim UV in the background of ample visible light.

Gao S, Takemura S-Y, Ting C-Y Huang S, Lu Z, Luan H, Rister J, Thum AS, YangM, Hong S-T, Wang JW, Odenwald WF, White BH, Meinertzhagen IA, Lee C-H. Theneural substrate of spectral preference in Drosophila. Neuron 2008;60:328-342.
Rister J, Pauls D, Schnell B, Ting C-Y, Lee C-H, Sinakevitch I, Strausfeld NJ,Ito K, Heisenberg M. Dissection of the peripheral motion channel in the visualsystem of Drosophila melanogaster. Neuron 2007;27:11132-11139.
Root CM,Masuyama K, Green DS, Enell LE, Nassel DR, Lee C-H, Wang JW. A presynaptic gaincontrol mechanism fine-tunes olfactory behavior. Neuron 2008;59:311-321.

Collaborators

Ian Meinertzhagen, PhD, DSc, Dalhousie University, Halifax, Canada
Jing Wang, PhD, University of California San Diego, La Jolla, CA
Benjamin White, PhD, Laboratory of Molecular Biology, NIMH, Bethesda, MD

For further information, contact leechih@mail.nih.gov.