a, Whole-brain GFP expression driven by the split-Gal4 line VT065306-AD ∩ VT029306-DBD (green), which labels FC2 neurons, and anti-Bruchpilot neuropil stain (magenta). b, Each panel shows a maximum z-projection at a different depth of the anterior-posterior axis. Top: The number of GFP positive somas, roughly 70 to 100, is comparable to the 88 FC2 neurons identified in the hemibrain12. Middle: fan-shaped body. Bottom: crepine. Each FC2 neuron projects unilaterally to the crepine, a symmetric structure that flanks the central complex and is situated dorsal to the lateral accessory lobes. c, Multicolor flip-out of a single FC2 neuron (left) and several FC2 neurons (right) labeled by VT065306-AD ∩ VT029306-DBD. The innervation pattern in the fan-shaped body is consistent with the FC2B or FC2C subtypes. While the GFP expression in this line suggests that it is selective for crepine projecting neurons with FC2-like anatomy, it is possible that there are some non-FC2 central complex neurons labeled by the line as well. d, Whole-brain GFP expression in the 57C10-AD ∩ VT037220-DBD split-Gal4 line (used for LAL imaging and silencing experiments), which labels PFL3 neurons. e, Top: protocerebral bridge. The white asterisk highlights a glomerulus lacking clear PFL3 signal, indicating that the line does not target all 24 PFL3 cells. The yellow asterisk shows a glomerulus innervated by a non-PFL3 neuron (likely a PEG neuron), since PFL3 neurons do not innervate the outer two glomeruli in the bridge. Middle: fan-shaped body. Bottom: lateral accessory lobes. White arrows highlight PFL3 expression in the left and right LAL. Yellow arrows mark non-PFL3 expression, which we excluded from our regions of interest for imaging analysis. f-g, Same as panels d-e but for VT000355-AD ∩ VT037220-DBD split-Gal4 line (used for patch-clamp and LAL-stimulation experiments). This line also stochastically labels PEG neurons. This was not a concern for either our patch-clamp (see Extended Data Fig. 6) or our LAL-stimulation experiments, since PEG neurons do not innervate the LAL. h-i, Same as panels d-e but for 27E08-AD ∩ VT037220-DBD split-Gal4 line (used for silencing experiments). Whereas this line drives significant GFP expression outside the central complex, including in the mushroom body (panel i, bottom), TNT expression driven by this same line appeared to be sparse outside the central complex (see Extended Data Fig. 11j).
Source Data