DEVELOPMENT AND FUNCTION OF DROSOPHILA VISUAL CIRCUITS
, 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
Molecular mechanisms regulating synaptic target selection of R7 photoreceptor axons
In the Drosophila visual system, each UV-responsive R7 photoreceptor axon projects within a single column to a specific layer of the optic lobe. In a genetic screen based on R7-dependent behavior, we identified the type I Activin receptor Baboon and the nuclear import component Importin-alpha3, which are required for column-specific targeting of R7 axons. In the past year, we focused on understanding how Activin signaling and nuclear import regulate R7-column–specific targeting.
Importin-alpha3– or baboon-mutant R7s have a novel phenotype: their axons target to the correct layer but extend into neighboring wild-type R7 columns, indicating that the local retinotopic map is defective. Baboon encodes an Activin-receptor serine/threonine kinase. In the so-called “canonical” pathway, Baboon—in response to Activin—phosphorylates the transcription factor Smad2, which then translocates into the nucleus to regulate transcription. We found that the canonical Activin signaling components are largely conserved in R7s. Using in situ hybridization, we found that, during the second target-selection stage, Activin is expressed in R7 and R8, but not in R1–6, neurons. With Activin mutants not available, we used dominant-negative and RNA-interference approaches to disrupt Activin function in specific tissues. We found that expressing a dominant negative form (DN) or RNAi of Activin in R7s resulted in R7 retinotopy defects that resembled those of baboon mutants. By contrast, use of various medulla drivers to express either Activin(DN) or Activin RNAi in the medulla neurons did not significantly affect R7 connection patterns. Together, the data suggest that Activin functions as an autocrine effector: Activin is both secreted by and exerts its effects on R7s. Similarly, removing the downstream transcription factor Smad2 disrupts R7 columnar specificity.
Figure 3.2
A schematic model shows how the R7 terminals are restricted to single columns by two partially redundant pathways: (1) an unknown signal (left-right arrows) that mediates repulsive interactions between adjacent R7 growth cones and (2) an Activin signal that regulates intrinsic R7 growth cone motility by the following mechanism: Activin, secreted from the R7 growth cone, activates its receptor Baboon in R7, resulting in phosphorylation of the transcription factor, Smad2; the phosphorylated Smad2, in complexes with Importin-alpha3, then shuttles into the nucleus to reduce growth cone motility via transcriptional regulation of yet-to-be identified target genes.
Given the similarity between Smad2 and importin-alpha3 mutant phenotypes and their known biological functions, we examined whether their two gene products interact physically. We found that Smad2 co-immunoprecipitated with Importin-alpha3 in vivo, indicating that Smad2 and Importin-alpha3 form a physical complex. To determine the subcellular localization of the Smad2/Importin-alpha3 complexes, we established a primary R-cell culture system and performed immunohistochemistry. We found that endogenous Importin-alpha3 concentrated in the growth cones as well as in vesicle-like structures along the axons. Interestingly, Smad2 staining largely overlapped with that of anti–Importin-a3, suggesting that Importin-alpha3 and Smad2 co-localize in axons and growth cones. Because of the known role of Importins in nuclear import, we sought to determine whether nuclear entry of Smad2 depends on importin-alpha3. While Smad2 accumulated in wild-type R7s, Smad2’s nuclear accumulation is greatly reduced in the importin-alpha3 or baboon mutant R7s. Together, the data indicate that Smad2 accumulation in R7 nuclei depends on Importin-alpha3 and Baboon and that Importin-alpha3 is a component of the Activin signaling pathway. In addition, the observation that Importin-alpha3 and Smad2 form complexes in axons and growth cones raises the intriguing possibility that Importin-alpha3 plays a role in the retrograde axonal transport of Smad2.
A consequence of removing Importin-alpha3 or Baboon is that only about 20 percent of the R7 terminals invade their neighbors. The incomplete penetrance of the phenotype suggests the existence of an additional mechanism that functions redundantly to the Activin signaling pathway. A previous study by Ashley and Katz suggested that competitive/repulsive interactions between adjacent R7s might play a role to restrict R7 terminals to single columns. To determine whether importin-alpha3– and baboon-mutant R7 are still subject to repulsion by their neighbors, we tested the effect of removing the R7s adjacent to importin-alpha3– or baboon-mutant R7s. To that end, we employed a temperature-sensitive allele of sevenless, v1, which removed most R7s at a non-permissive temperature. We found that wild-type R7 axons form normal synaptic boutons in retinotopically correct columns, even in a largely empty R7 terminal field, indicating that disrupting mutual repulsion is not sufficient to affect R7 columnar specificity. By contrast, removing neighboring R7s greatly increased the tendency of importin-alpha3– or baboon-mutant R7s to invade adjacent targets. The results suggest that importin-alpha3– and baboon-mutant R7s are still responsive to repulsion by neighboring R7s, accounting for the incomplete penetrance of their phenotypes. Thus, at least two redundant mechanisms restrict R7 terminals to the correct columns: (1) an extrinsic mechanism mediated by mutual repulsion among R7s and (2) an intrinsic mechanism mediated by Activin signaling.
Pramatarova A, Ochalski PG, Lee C-H, Howell BW. Mouse disabled 1 regulates the nuclear position of neurons in a Drosophila eye model. Mol Cell Biol 2006;26:1510-7.
Ting C-Y, Lee C-H. Visual circuit development. Curr Opin Neurobiol 2007;17:65-72.
Ting CY, Yonekura S, Chung P, Hsu SN, Robertson HM, Chiba A, Lee CH. Drosophila N-cadherin functions in the first stage of the two-stage layer-selection process of R7 photoreceptor afferents. Development 2005;132:953-63.
Yonekura S, Lei X, Ting C-Y, Lee C-H. Adhesive but not signaling activity of Drosophila N-cadherin is essential for photoreceptor target selection. Dev Biol 2007;304:759-70.
Yonekura S, Ting C-Y, Neves G, Hung K, Hsu S, Chiba A, Chess A, Lee C-H. The variable transmembrane domain of Drosophila N-cadherin regulates adhesive activity. Mol Cell Biol 2006;20:6598-608.
Mapping color-vision circuits
To determine the mechanism of visual information processing in Drosophila, we initiated a project to map the connection patterns and function of visual circuits. Given the vast complexity of neuron subtypes and their interconnections in the visual system of Drosophila, we adopted a “divide and conquer” strategy. First, we identified key neuron subtypes and their potential synaptic partners based on their unique axonal and dendritic patterns. Second, we subdivided the neurons based on their molecular properties, especially their use of neurotransmitters and receptors. Finally, we determined their functions by inactivating or restoring their synaptic function and examining the behavioral consequences. We chose to focus on the first-order interneurons in the medulla because (1) they process well-defined visual information directly from R7/R8 photoreceptors, (2) they are likely to be involved in color vision, and (3) they are most relevant to our developmental studies described above.
We reasoned that, because Drosophila photoreceptor neurons are histaminergic neurons, the first-order interneurons must express the histamine receptor Ort in order to respond to the photoreceptors. We identified the ort promoter region and generated an ort-Gal4 driver. The ort promoter–Gal4 driver labels subsets of lamina and medulla neurons, which are likely to be the direct synaptic targets of R1–6 and R7/8, respectively. In an electron-microscopic (EM) study, we confirmed that the ort+ medulla neurons indeed form synaptic connections with R7 and R8 axonal termini. Furthermore, expressing the ort gene under the control of ort-Gal4 completely rescued the electrophysiology and behavioral defects of the ort mutants, indicating that the ort-Gal4 construct faithfully recapitulated the endogenous ort expression pattern.
We performed single-cell analyses to identify the neuron subtypes that express Ort. Based on approximately 300 single-cell clones, we found that the ort-Gal4 labeled three types of lamina neurons (L1–3) as well as eight subtypes of medulla neurons. In particular, our observation that the ort-Gal4 driver labeled L1 through L3 is in accord with a previous EM reconstruction study and validates our approach. We chose to focus on Tm5, Tm9, and Tm20 because (1) they were labeled with high frequency and therefore are likely present in every medulla column, and (2) they extend dendritic processes in M6 (Tm5) and M3 (Tm5/9/20) and therefore might receive input from R7 and R8 (or L3) directly.
The ort-Gal4 driver allows us to manipulate the first-order interneurons as a group, but not as individual subtypes. We therefore employed the split-Gal4 system to restrict the Gal4 activity to different subclasses that use distinct neurotransmitters. Specifically, we combined the ort promoter with promoters of specific neurotransmitters. We found that the combination of ort-AD-Zip and Chat-Gal4-DBD-Zip (which is expressed in Ort+ cholinergic neurons) labeled two known subsets of Ort+ neurons, including Tm9 and Tm20. In addition, we generated an AD-Zip enhancer trap vector and performed a P-element swap to substitute Ok371vGlut enhancer trap, which labels glutaminergic neurons. The combinatorial driver, composed of OK371-AD-Zip and ort-DBD-Zip, labeled a different subset of Ort+ neurons, including a subclass of Tm5. The results indicated that the split-Gal4 system is suitable for manipulating distinct subsets of Ort+ neurons.
To determine the function of each neuron subtype in color vision, we manipulated the activity of single or combinations of Ort+ neuron subtypes and tested whether specific Ort+ subtypes are “required” or “sufficient” for color-discriminating behavior. We previously found that flies have an innate preference for shorter wavelength (UV>>blue>green). In forced two-choice tests under a high green/UV light intensity ratio (1:150), over 85 percent of wild-type flies still preferred UV over green light while ort mutant flies chose the two light sources with a similar frequency. Conversely, mutant flies bearing ort→Shits1 lost their preference for UV at a non-permissive but not permissive temperature. The results indicated that Ort+ neurons are functionally required for flies’ innate preference of UV. Our preliminary results revealed that the loss of innate preference for UV in ort mutants could be rescued by using the combinatorial driver of OK371 and ort driving the expression of the Ort gene, suggesting that the Tm5 subtype mediates innate UV preference, presumably by relaying signals from the UV-sensitive R7s to the higher optic ganglion lobula.
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 visual system of Drosophila melanogaster. Neuron 2007;56:155-70.
COLLABORATORS
Tory Herman, PhD, University of Oregon, Eugene, OR
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.
