Driver lines for studying associative learning in Drosophila

Yichun Shuai, Megan Sammons, Gabriella R Sterne, Karen L Hibbard, He Yang, Ching-Po Yang, Claire Managan, Igor Siwanowicz, Tzumin Lee, Gerald M. Rubin, Glenn Turner, Yoshinori Aso

TL;DR

Researchers have released a massive new toolkit of over 800 site-specific genetic "driver lines" for Drosophila, providing precise access to the neural circuits governing associative learning. By bridging the gap between high-resolution electron microscopy (EM) maps and functional light microscopy (LM), this resource allows scientists to manipulate specific neurons—including those for "reward" (sugar) and "punishment" (pain)—to see how the fly brain actually learns.

What Problem Does This Paper Address?

In the world of neuroscience, Drosophila melanogaster is a superstar because we have a nearly complete "wiring diagram" of its brain (the connectome) thanks to electron microscopy. We know the Mushroom Body (MB) is the brain's learning center, where odors (conditioned stimuli) are linked to rewards or punishments (unconditioned stimuli).

However, there was a major bottleneck: having a map isn't the same as having the keys. Researchers lacked specific genetic "drivers" (split-GAL4 lines) to target the specific interneurons that feed information into or carry instructions out of the Mushroom Body. Without these, it was impossible to selectively activate or silence these neurons to prove what they actually do during behavior.

Methods

The authors utilized the intersectional split-GAL4 method. Think of this as a genetic "AND" gate: a neuron only expresses a marker if two different genomic enhancers are active simultaneously.

  1. Screening: They screened over 4,000 combinations of genetic "hemidrivers" (AD and DBD).
  2. Anatomical Verification: They used high-resolution confocal microscopy and Multi-Color Flip-Out (MCFO)—a technique that stochastically labels individual neurons in different colors—to confirm the exact morphology of the cells being targeted.
  3. LM-to-EM Matching: They manually matched the light microscopy (LM) images of these genetic lines to the 3D skeletons of neurons in the Hemibrain EM connectome. This allows a researcher to look at a cell in the connectome and immediately find the genetic tool needed to study it in a living fly.
  4. Behavioral Validation: They used optogenetics (using light to trigger neurons) in a circular arena to test if specific lines could effectively substitute for real-world rewards (sugar) or punishments (heat/pain).

Datasets

The study leverages and contributes to several massive open-science resources:

  • Janelia FlyLight: A database containing over 28,000 confocal images from this collection.
  • Hemibrain Connectome (v1.2.1): The primary EM dataset used for neuronal matching.
  • neuPrint: The platform used for connectivity analysis.
  • Supplementary Files: The paper includes a comprehensive "lookup table" mapping genetic lines to specific EM BodyIDs.

Key Insights and Ideas

  • The "Clean" Reward Driver: The researchers identified a specific line (SS87269) that targets sugar-sensing neurons in the labellum and legs without the "noise" of other sensory neurons. This is now the gold-standard tool for "reward substitution" experiments.
  • Nociceptive Pathways: They identified a driver (SS01159) for ascending neurons that carry "pain" signals from the body to the brain, triggering robust avoidance behaviors.
  • Neural Individuality: Perhaps the most surprising finding was the lack of stereotypy in certain neurons. While most fly neurons are identical from animal to animal, the authors found that MBON08 and MBON09 exhibit striking asymmetry. One animal might have two of one type, while another has a mix, or the neurons might project to different hemispheres entirely. This suggests that even in "simple" insects, developmental randomness creates individual brain variations.

Key Results

  • The Collection: 828 useful split-GAL4 and split-LexA lines were identified.
  • Coverage: These lines cover approximately 300 distinct cell types.
  • MB Saturation: The collection achieves 87% coverage for non-projection neurons within the Mushroom Body.
  • Connectivity: They identified drivers for 30 major interneuron cell types that have massive synaptic inputs (over 100 synapses) from MB output neurons or to dopamine neurons.
  • Functional Success: 22 out of 34 tested split-GAL4 lines were successfully converted to split-LexA, allowing researchers to manipulate two different cell types simultaneously in the same fly using two different genetic systems.

Limitations and Future Work

While the coverage of neurons inside the Mushroom Body is nearly complete, the authors note that they only cover about 10% of the interneurons upstream and downstream of the MB. The fly brain is vast, and many "dark" neurons still lack specific drivers.

Future work will likely involve using these lines in "closed-loop" behavioral experiments—where the fly's own actions trigger the activation of these neurons—to further decode the logic of reinforcement learning. Additionally, the discovery of MBON08/09 asymmetry opens a new door into studying how individual brain differences affect learning performance.

Summary updated

Abstract

The mushroom body (MB) is the center for associative learning in insects. In Drosophila, intersectional split-GAL4 drivers and electron microscopy (EM) connectomes have laid the foundation for precise interrogation of the MB neural circuits. However, investigation of many cell types upstream and downstream of the MB has been hindered due to lack of specific driver lines. Here we describe a new collection of over 800 split-GAL4 and split-LexA drivers that cover approximately 300 cell types, including sugar sensory neurons, putative nociceptive ascending neurons, olfactory and thermo-/hygro-sensory projection neurons, interneurons connected with the MB-extrinsic neurons, and various other cell types. We characterized activation phenotypes for a subset of these lines and identified the sugar sensory neuron line most suitable for reward substitution. Leveraging the thousands of confocal microscopy images associated with the collection, we analyzed neuronal morphological stereotypy and discovered that one set of mushroom body output neurons, MBON08/MBON09, exhibits striking individuality and asymmetry across animals. In conjunction with the EM connectome maps, the driver lines reported here offer a powerful resource for functional dissection of neural circuits for associative learning in adult Drosophila.
BibTeX
@inproceedings{shuai2025driver,
  author = {Yichun Shuai and Megan Sammons and Gabriella R Sterne and Karen L Hibbard and He Yang and Ching-Po Yang and Claire Managan and Igor Siwanowicz and Tzumin Lee and Gerald M. Rubin and Glenn Turner and Yoshinori Aso},
  title = {{Driver lines for studying associative learning in Drosophila}},
  booktitle = {eLife},
  year = {2025},
  doi = {10.7554/elife.94168.4},
}
Figure