A searchable image resource of Drosophila GAL4 driver expression patterns with single neuron resolution

Abstract

Precise, repeatable genetic access to specific neurons via GAL4/UAS and related methods is a key advantage of Drosophila neuroscience. Neuronal targeting is typically documented using light microscopy of full GAL4 expression patterns, which generally lack the single-cell resolution required for reliable cell type identification. Here, we use stochastic GAL4 labeling with the MultiColor FlpOut approach to generate cellular resolution confocal images at large scale. We are releasing aligned images of 74,000 such adult central nervous systems. An anticipated use of this resource is to bridge the gap between neurons identified by electron or light microscopy. Identifying individual neurons that make up each GAL4 expression pattern improves the prediction of split-GAL4 combinations targeting particular neurons. To this end, we have made the images searchable on the NeuronBridge website. We demonstrate the potential of NeuronBridge to rapidly and effectively identify neuron matches based on morphology across imaging modalities and datasets.

Keywords: D. melanogaster; GAL4; MultiColor FlpOut; NeuronBridge; neuron search; neuroscience; split-GAL4.

Figure 1.. Generation 1 MultiColor FlpOut (MCFO) and electron microscopy (EM)/light microscopy (LM) comparison overview.

(A) Overall GAL4 expression pattern of a driver line containing a cell type of interest, shown as a color depth maximum intensity projection (MIP) (Otsuna et al., 2018). Original images are from published datasets (Jenett et al., 2012). (B1) Example MCFO labeled cells from the driver line in (A). MCFO labeling reveals a prominent descending neuron. (B2) An additional MCFO labeled cell of the same type but from a different line. The color depth MIPs in B1 and B2 represent data from one of the three MCFO markers, so color changes indicate changes in the z-dimension rather than differential MCFO labeling. (C1, C2) Matching EM reconstructions for the cell type. Both panels show reconstructions from the right-side Hemibrain; the lower panel is mirrored to facilitate comparison to the LM data. (D) PatchPerPixMatch (PPPM) overlay of MCFO from (B1) and EM reconstruction from (C2). (E) Split-GAL4 made from split hemidrivers derived from GAL4 lines in A and B. Driver lines used are R56H09 (A, B1), R23E11 (B2), and SS01588 (E). Hemibrain body IDs are 571346836 (C1) and 1786496543 (C2). All scale bars, 50 µm.

Figure 1—figure supplement 1.. Generation 1 MultiColor FlpOut (MCFO) expression density categories.

(A) Two example brain maximum intensity projections (MIPs) are shown for each expression density category, except Category 5, where a single brain is shown both as an MIP and a single confocal slice through its center. Qualitative categorization was manually performed on a line level based on 2D MIPs of MCFO and full central nervous system (CNS) expression patterns. Category 1 lines contained no visible neurons or only commonly repeated ones. Categories 2–4 labeled identifiable neurons with increasing density. Category 5 lines had such dense expression that the immunohistochemical labeling approach failed to fully label the center of the brain. Category 1 and 5 lines were generally excluded from imaging and the collection as a whole. Scale bar, 50 µm. (B) The frequency distribution of lines within the different expression density categories are shown. Sample size is all 4919 lines considered for inclusion in either phase of the 40× pipeline. 95% of lines were within the desired range.

Figure 2.. Phase 1 and 2 overview and labeling examples.

R14E12-GAL4 in attP2 crossed to (A) pJFRC2-10XUAS-IVS-mCD8::GFP, (B) R57C10-Flp MCFO, or (C) hs-Flp MCFO. Adult central nervous system (CNS) maximum intensity projections (MIPs) are shown, with neuropil reference in gray and neuronal signal in green (A) or full MultiColor FlpOut (MCFO) colors (B, C). Multiple examples are shown for B, C. Scale bars, 50 µm. (D) Glia are seen with VT008658-GAL4 in attP2 crossed to (D1) pJFRC2-10XUAS-IVS-mCD8::GFP and (D3) hs-Flp MCFO, but not (D2) R57C10-Flp MCFO. (E) Kenyon cell labeling is not seen with R86H02-GAL4 in attP2 crossed to (E1) pJFRC2-10XUAS-IVS-mCD8::GFP or (E2) R57C10-Flp MCFO, but is seen when crossed to (E3) hs-Flp MCFO. (F) Kenyon cell labeling is seen with R91B01-GAL4 in attP2 crossed to (F1) pJFRC2-10XUAS-IVS-mCD8::GFP and (F3) hs-Flp MCFO, but is not seen when crossed to (F2) R57C10-Flp MCFO. (G) An ascending neuron (sparse T) is commonly seen with many Gen1 GAL4 lines crossed to different reporters. VT010592-GAL4 in attP2 crossed to R57C10-Flp MCFO is shown as an example. A single neuron channel plus reference are shown for clarity. The inset shows a lateral (y-axis) MIP of the brain. All scale bars, 50 µm.

Figure 3.. Electron microscopy (EM)/light microscopy (LM) search for split-GAL4 creation.

Neuron search techniques allow for the identification of Gen1 MultiColor FlpOut (MCFO) images containing an EM body of interest. The corresponding Gen1 GAL4 lines should label the same neuron with other upstream activation sequence (UAS) reporters, as should split-GAL4 hemidrivers constructed with the same enhancer fragment. The two hemidrivers can then be combined into a split-GAL4 with the aim of generating a driver that specifically targets that neuron. An example is shown of the anticipated search process, from a neuron identified via EM to the creation of a split-GAL4 driver. As in Figure 1, NeuronBridge displays color depth maximum intensity projections (MIPs) of single MCFO markers rather than the full MCFO image, so color changes indicate depth rather than different neurons. NeuronBridge result order was reformatted for display purposes. The example shown includes FlyEM Hemibrain body ID 733036127 (Scheffer et al., 2020), Generation 1 GAL4 lines R17C11-GAL4, R52G04-GAL4, and split-GAL4 MB310C (MBON07) (Jenett et al., 2012; Aso et al., 2014b).

Figure 4.. Forward analysis: direct evaluation of Color Depth Maximum intensity projection (CDM) and PatchPerPixMatch (PPPM) search results.

EM bodies were searched for in Phase 1 40× Gen1 MultiColor FlpOut (MCFO) light images using CDM and PPPM approaches. Search results were qualitatively evaluated by an anatomical expert for the presence of the sought neuron. Most results were scored based on color depth maximum intensity projection (MIP) images. Full image stacks were used to score about 20% of samples, including the majority of samples scored as containing the sought neuron. The cumulative number of correct matches found is plotted against the depth of searching for CDM (green) and PPPM (magenta). (A) Average results for each search approach are plotted in bold on top of individual results. (B) Cell type LC18 (Hemibrain body 1722342048) search result evaluation. (C) Cell type CT1 (Hemibrain body 1311993208) search result evaluation.

Figure 4—figure supplement 1.. Forward analysis individual plots for Color Depth Maximum intensity projection (CDM) and PatchPerPixMatch (PPPM).

(A–J) Individual CDM and PPPM results for the indicated cell types. (K) All cell types composited with partial lobula and lobula plate. Includes duplicated images from Figure 4. Electron microscopy (EM) images are from https://neuprint.janelia.org (Clements et al., 2020; Scheffer et al., 2020). Figure 4—source data 1.