The APL neuron innervating the mushroom body neuropil is GABAergic. (a) The GH146-Gal4 driver promotes expression of membranelocalized mCD8-GFP in the antennal lobes and the mushroom body lobes (outlined by dashed line for the right hemisphere). (b) Higher magnification image of the area marked by a square in (a) illustrating a complex punctate and reticular pattern of the innervation. (c, d) One example of a single cell clone showing the morphology of the APL neuron, as viewed from two different focal planes. The primary process of the APL neuron bifurcates (empty arrow head) prior to reaching the mushroom body neuropil into a branch that enters the vertical lobes (arrow) as shown in (c) and a branch that enters the mushroom body calyx as shown in (d). (e-g) Anti-GFP (e), anti-GABA (f) staining, and the merge (g) of GH146-Gal4 driving mCD8-GFP expression. (h-j) Higher magnification images of the areas marked by squares in (e-g) showing the cell body of the APL neuron. The only process observed from the APL cell body is the one that exits with a trajectory towards the mushroom body neuropil. Scale bars represent 20 μm in a, c, e, h and 10 μm in b.

Reducing GABA synthesis in the APL neuron enhances olfactory learning. (a) The flies carrying the GH146-Gal4 driver and the UAS-Gad-RNAi transgene exhibited an enhanced performance index (P.I.) after olfactory conditioning using 1, 3, or 12 electric shock pulses presented within a 1 min exposure to the conditioned odor. The P.I. of the Gad knock down group was not significantly higher than one control group, GH146/+, after 12 shocks of training. This was due to a ceiling effect as the P.I.s approached 1.0. (b) Expression pattern of mCD8-GFP driven by GH146-Gal4 alone (top row) or by the GH146-Gal4, Cha-Gal80 combined driver (bottom row). Both front and back views of the brains are shown, together with higher magnification images of the areas marked by squares in both views, illustrating the mushroom body lobes or the mushroom body calyx and the lateral horn (LH). The APL neuron cell bodies are marked by the arrows in the first panel of row 1. The punctate arc of fluorescence that is posterior to the vertical lobes shown in the second panel of row 1 (marked by the empty arrow head) is from an antennal lobe projection neuron previously characterized10,34. Note the loss of APL and mushroom body neuropil fluorescence by the introduction of Cha-Gal80. (c) Knock down of Gad by the combined GH146-Gal4, Cha-Gal80 driver failed to enhance learning, while knock down by GH146-Gal4 alone reproduced the enhanced learning shown in (a). n = 6 for each group under each condition. Means ± s.e.m.; *P<0.05; **P<0.01 (Fisher's PLSD). Scale bars represent 20 μm.

The APL neuron forms a memory trace of reduced calcium response for the conditioned odor after training. (a) Flies were exposed to 1 min odor simultaneously with 12 pulses of electric shock (Trained) or the same odor and shock but separated by 15 s (Control). The G-CaMP response for the odor was probed 3 min before (Pre), 30 s after (Post1), and 5 min after (Post2) the 1 min odor presentation. Each probe test lasted for 3 s. (b, c) Quantitative summary of group data and representative pseudo-color images of the groups treated as in (a) using 3-octanol (b) or methylcyclohexanol (c). Planned comparisons were made between each group at different time points, and between the two groups at each time point, with all statistically significant differences labeled. (d) Flies were differentially trained using one odor paired with shock and the responses to both odors used in the assay were tested at 3-4 min before (Pre) or 4.5-5.5 min after (Post) training. Tests to the alternative odors were separated by 1 min. (e, f) Quantitative summary and representative pseudo-color images of groups treated in (d) using 3-octanol (e) or methylcyclohexanol (f) as the conditioned odor. For (b) and (c), n = 7 in each group. Means ± s.e.m.; *P<0.05; **P<0.01 (paired Student's t-tests for within group, unpaired Student's t-tests for between group). For (e) and (f), n = 9 in each group. Means ± s.e.m.; **P<0.01 (paired Student's t-test). Scale bars represent 20 μm.

Tissue-specific knock down of GABA synthesis by driving a Gad-RNAi transgene. (a) Anti-GFP (left column), anti-GABA (middle column) staining, and merge of the two (right column) for brains of flies carrying the GH146-Gal4 driver alone (top two rows) or GH146-Gal4 driving an RNAi construct against glutamic acid decarboxylase (Gad) (bottom two rows). Rows 2 and 4 show higher magnification images of the APL cell body areas marked by squares highlighted in rows 1 and 3, respectively. The cell bodies of APL neurons are outlined by dashed lines in rows 2 and 4. Arrowheads indicate the cell bodies of nearby smaller GABAergic neurons that are not identified by the GH146-Gal4 driver. Some of the red fluorescence shown within the areas marked by squares in row 3 is not observable in row 4 because it is outside of the thinner average projection images of row 4 that encompass only the thickness of the APL cell body. (b) Anti-GABA staining of the ellipsoid body (dashed oval) of flies carrying the GH146-Gal4 driver alone (left) or the GH146-Gal4 driving Gad-RNAi (right). (c) Quantification of anti-GABA staining of the APL cell body, n = 16 samples per group. (d) Quantification of anti-GABA staining of the ellipsoid body, n = 8 samples per group. Means ± s.e.m.; ***P<0.001 (Student's t-test). Although not indicated on the figure, all flies depicted here carried one copy of UAS-spH (a pH sensitive GFP) on the second chromosome for visualizing the APL cell body and its processes. Scale bars represent 20 μm.

Calcium responses and neurotransmitter release of the APL neuron during odor or electric shock stimulation. Optical recordings were made from the APL innervation in the horizontal lobes of the mushroom bodies (circled by dashed lines). (a-c) Representative pseudo-color images taken at the peak response (average of 5 frames) and percent fluorescence change of group data across time during the presentation of 3-octanol (a), methylcyclohexanol (b) or electric shock (c), detected as calcium responses with G-CaMP driven by the GH146-Gal4 driver. (d, e) Representative pseudo-color images taken at the peak response (average of five frames) and percent fluorescence changes of group data across time during the presentation of 3-octanol (d) or electric shock (e), detected as neurotransmitter release using synapto-pHluorin (spH) driven by the GH146-Gal4 driver. The red bar under x-axis indicates the stimulation period. n = 5-8 trials. Means ± s.e.m. Scale bars represent 20 μm.