Document S1. Four Figures

Current Biology, Volume 20
Supplemental Information
The Shaping of Male Courtship Posture
by Lateralized Gustatory Inputs
to Male-Specific Interneurons
Masayuki Koganezawa, Daisuke Haba, Takashi Matsuo, and Daisuke Yamamoto
Figure S1. Effects of Inactivation of Gr32a-Expressing Neurons or Foreleg Amputation on Male
Courtship Behavior, Related to Figure 1
(A) Mating performance of males with inactivated Gr32a-expressing cells in a 8 mm mating chamber.
Mating success (left-hand side graph), latency to copulation (middle graph) and courtship index
(right-hand graph) were compared between the control IMP-TNT-expressing males (open bars) and
TNT-expressing males (filled bars). No statistical differences were detected between the two fly groups
in any parameters examined (p > 0.05 in the χ2 test for the mating success rate or u-test for the latency
to copulation and CI). The number of pairs observed was: 25 for w; UAS-IMP-TNT/+; Gr32a-Gal4/+,
32 for w; UAS-TNT/+; Gr32a-Gal4/+ in the experiment to quantify the mating success rate and CI; 20
for w; UAS-IMP-TNT/+; Gr32a-Gal4/+, 25 for w; UAS-TNT/+; Gr32a-Gal4/+ in the experiment to
quantify the latency to copulation.
(B) The effect of inactivating Gr32a-expressing neurons by shiK44A on wing extension of courting
males. The number of wing extension epochs (left-side graph), and the simultaneous wing extension
index (SWEI; right-side graph) were compared among the males of the indicated genotypes: w;;
Gr32a-Gal4/+ (n=22), w; UAS-shiK44A/+; UAS-shiK44A/+ (n=19) and w; UAS-shiK44A/+;
Gr32a-Gal4/UAS-shiK44A (n=23). Following Kruskal-Wallis ANOVA, statistical differences among the
data sets were evaluated by the Bonferroni/Dunn test (***p < 0.001).
(C) The effect of foreleg tarsus amputation on the transition pattern of wing usage during male
courtship. The amputation of a tarsus of the left foreleg diminished the usage frequency of the left wing
and that of the right foreleg diminished the usage frequency of the right wing (**p < 0.01, ***p <
0.001 by Wilcoxon test). The occurrence of simultaneous extension of two wings was drastically
increased by the unilateral tarsal amputation (***p < 0.001 by u-test). The numbers of flies examined
were 17 (intact), 27 (left foreleg amputation) and 29 (right foreleg amputation).
Figure S2. Analysis of Gr66a- and Gr68a-Expressing Neurons, Related to Figure 3
(A–D) The localization of Gr66a-expressing neurons in the foreleg and the effects of inactivation of
these neurons on wing extension during male courtship. Gr66a-expressing neurons were labeled with
Gr66a-I-GFP (green, A) and Gr32a-expressing neurons with UAS-lacZ as driven by Gr32a-Gal4
(magenta, B). GFP expression was detected with an anti-GFP antibody and visualized with
Alexa488-conjugated goat anti-rabbit IgG. The lacZ product was detected with an anti-β-galactosidase
antibody and visualized with Alexa546-conjugated goat anti-mouse IgG. (C) represents a merged image
of (A) and (B). Scale bar represents 100 µm. The genotype of the flies was w; UAS-lacZ/+;
Gr66a-I-GFP/Gr32a-Gal4. Gr66a-I-GFP was a gift from K. Scott.
(D) The number of wing extension epochs (left-side graph), and the simultaneous wing extension index
(SWEI; right-side graph) were compared among the males of the indicated genotypes: w; UAS-TNT/+
(n=18), w; UAS-IMP-TNT /+ (n=17), w; UAS-TNT/G66a-Gal4 (n=25) and w;
UAS-IMP-TNT/Gr66a-Gal4 (n=23). Following Kruskal-Wallis ANOVA, statistical differences among
the data sets were evaluated by the Bonferroni/Dunn test (***p < 0.001).
(E–I) Localization and projection of Gr68a-expressing neurons in males.
(E–G) The male-specific Gr68a-expressing neurons (indicated with arrowheads) are present in the
tarsus of the foreleg (E) but not in the midleg (F) or hindleg (G). In addition, Gr68a-expressing cells
are observed in the second antennal segment, wing base, and coxa-trochanteral segments of all legs
(white arrows in H). Green deposits observed in every joint of the midleg and hindleg are artifacts
unrelated to GFP.
(I) The central projection of Gr68a-expressing neurons. Gr68a-expressing neurons ramify extensively
in the thoracic neuropil in each segment, and the terminals of ascending fibers are located in the
anterior mechanosensory region of the brain (yellow arrowheads). The genotype of flies was w;
UAS-mCD8::GFP/+; Gr68a-Gal4/+.
Scale bars represent 100 µm for (E)–(G), 1 mm for (H), and 200 µm for (I).
Figure S3. Sexual Differences of mAL Neurites in Relation to the Projection of Gr32a-Expressing
Neurons, Related to Figure 4
(A–F) The effect of male-to-female sexual transformation of mAL neurons by fru mutations on the
connection with Gr32a-expessing neurons in the male brain. The left-side (27-37 µm in depth from the
anterior surface) panel illustrates the ventral brain and the right-side panel the dorsal brain (43-53 µm
in depth from the anterior surface). Shown are the brains of a male (A and D), female (B and E), and
“feminized” fru mutant male (C and F) double-stained with the nc82 monoclonal antibody (magenta;
A1–F1) and anti-GFP antibody (green; A3–F3). The nc82 monoclonal antibody intensely stains axon
terminals of Gr32a-expressing neurons (white arrows in D1, E1 and F1). (A2)–(F2) are merged images
for each raw. The genotypes of the flies were y hs-flp; FRT G13 UAS-mCD8:: GFP; fruNP21/TM6b (A,
B, D, and E) and y hs-flp; FRT G13 UAS-mCD8:: GFP/+; fru1/ fruNP21 (C and F). (A4)–(F4) represent
schematic drawings of the location of mAL neurites (green) and Gr32a-expressing axons (magenta).
Male-type simple dendritic branches are indicated with yellow arrows and female-type bifurcating
dendritic branches are indicated with white arrowheads. The feminized mAL dendrites appear in the
optical sections between 27-37 µm in depth from the anterior surface of the brain (indicated with white
arrowheads), whereas the axon terminal of Gr32a-expressing neurons is observed in the sections
between 43-53 µm in depth (indicated with white arrows). No overlap of mAL neuron dendrites and
the axon terminal of Gr32a-expressing neurons is discernible in the feminized male brain. Scale bar
represents 50 µm.
(G–N) Sex differences in the neural connections involving mAL and Gr32a-expressing neurons. The
number of mAL neuron somata (circled in G and K) and the mAL neurite positions in the
suboesophageal ganglion (white arrowheads in H and L) are different between the male (G and H) and
female (K and L). The genotype of the flies examined was UAS-syt-HA; FRT G13
UAS-mCD8::GFP/+; fruNP21/Gr32a-Gal4. fru-expressing neurons were visualized with anti-GFP
(green) in the brains counterstained with the nc82 monoclonal antibody (blue) that highlights neuropils
(G and K). Axons of Gr32a-expressing neurons were also labeled with anti-GFP in (H and L). Anti-HA
visualized the localization of synaptotagmin-HA (magenta), a presynaptic marker in the male (I) and
female (M) brain when observed at the depth of ca. 40–60 µm from the anterior brain surface. (J) and
(N) are merged images of (H) and (I) and (L) and (M), respectively. The mAL neurites in the
suboesophageal ganglion are distributed with putative presynaptic sites only in males (yellow
arrowheads in I). Scale bar represents 50 µm.
(O–V) Sex differences in presynaptic sites of mAL neurons. MARCM clones of mAL neurons were
visualized with anti-GFP (green) in the male (O and P) and female (S and T). Anti-HA visualized the
localization of synaptotagmin-HA (magenta) in the male (Q) and female (U). The genotype of the flies
examined was UAS-syt-HA; FRT G13 UAS-mCD8::GFP/FRT G13 tub-Gal80; fruNP21/MKRS hs-flp.
The mAL neurites in the superior lateral protocerebrum are major presynaptic sites in both male and
female (white arrows in Q and U). The neurites in the suboesophageal ganglion are distributed with
putative presynaptic sites only in males (yellow arrowheads in Q). The region of mAL neurites coming
into contact with the Gr32a-expressing neurons has no signal for synaptotagmin-HA, indicating that
this part of mAL neurites is likely to be a postsynaptic site (white arrowheads in P and R). Scale bar
represents 50 µm.
Figure S4. Analysis of the Structure and Function of fru-Expressing Neurons, Related to Figure 5
(A) Altered wing usages during courtship in fru mutant males. The total number of wing extensions
is decreased in fru2 mutant males, reflecting a reduced level of courtship (left-side graph). The
simultaneous wing extension during courtship is significantly increased in fru mutant males (right-side
graph). The numbers of flies examined were 17 (CS) and 11 (fru2). Statistical differences between the
data sets were evaluated by the u-test (**p < 0.01, ***p < 0.001).
(B) Inhibition of male courtship activities toward females by forced shiK44A expression with the aid of
fruNP21. The courtship index was significantly decreased when shiK44A was expressed in most of
fru-expressing neurons in the male brain (***p < 0.001 by u test). The genotypes of the flies examined
are y hs-flp; FRT G13 UAS-mCD8::GFP; fruNP21/TM6b (n=18) and y hs-flp; FRT G13
UAS-mCD8::GFP/CyO; fruNP21/UAS-shiK44A (n=18), respectively. This result verifies the use of shiK44A
to block synaptic transmission locally in MARCM clones.
(C–K) mAL neurons are GABAergic. Anterior (C-E) and posterior (F-H) views of a male brain doubly
stained with anti-GFP (C, F, and I) and anti-GABA (D, G, and J) antibodies. Merged images of left-side
and middle panels are shown in the right-hand panel (E, H, and K). Anti-GFP labeling reflects fru
reporter expression. The genotype of flies was y hs-flp; FRT G13 UAS-mCD8:: GFP; fruNP21/TM6b.
(I–K) An mAL clone doubly stained with the anti-GFP (I) and anti-GABA (J) antibodies. (K) represents
a merged image of (I) and (J). The genotype of fly was y hs-flp; FRT G13 tub-Gal80/ FRT G13
UAS-mCD8:: GFP; fruNP21/+. Neurites of the mAL clone (arrows) and nonclonal mAL (arrowheads)
were stained positive with the anti-GABA antibody when observed at the depth of 30-45 µm from the
anterior brain surface.
Scale bars represent 100 µm for (C)–(H) and 50 µm for (I)–(K).
(L–O) Bilateral mAL clones induced by MARCM. Confocal images of a single preparation rotated
every 30° are shown in (L) to (O). The fly genotype was y hs-flp; FRT G13 tub-Gal80 FRT G13
UAS-mCD8::GFP; fruNP21/Gr32a-I-GFP. The fly carried Gr32a-I-GFP in addition to fruNP21 and
UAS-mCD8::GFP, and all these reporters gave green labeling by anti-GFP antibody. The ganglion is
counterstained with nc82 monoclonal antibody (magenta). mAL cell bodies are circled. Note the
superposition of neurites of left and right mAL neurons in the suboesophageal ganglion (indicated with
arrows).