Supporting Information

Files in this Data Supplement:

Fig. 5. Tissue feminization caused by the C309 UAS-tra^F transgene combination. Flies were anesthetized with light ether then photographed with a camera attached to a dissecting microscope. All XY flies carrying one copy of each transgene exhibited external appearance as exemplified here, including lack of sex combs (Lower, side view); no XY/C309/UAS-tra^F flies, out of »500 that were generated, presented a wild-type (WT) male appearance. However, variability was observed with regard to anomalous appearance of external genitalia and of abdominal pigmentation on the surface of these doubly-transgenic chromosomal males. Dissections of eight XY/C309/tra^F flies and scrutiny of internal reproductive organs revealed that all such flies contained female seminal receptacles (but no spermathacae) and testes, although the latter appeared to be smaller than normal (n = 6 for observations of WT male structures). Dissections of the XY/C309/tra^F flies’ testes showed that seven of eight individuals contained motile sperm, although the number of such gametes appeared reduced compared with what was observed in dissections of WT male controls.

Fig. 6. Expression of the C309 transgene during development. Embryos were collected from agar-including medium to which parents laying C309/UAS-egfp eggs were added overnight; approximately half-developed embryos were then removed from the medium’s surface by using small amounts of water and a brush; these specimens were mounted on a microscope slide with glycerol and scanned on the confocal (see below) with respect to the GFP channel only, because male-specific fruitless expression does not commence until a late larval stage (1). Pupae carrying C309 and UAS-egfp, were collected by placing third-instar larval males onto moist filter paper then leaving them to develop halfway into the pupal period. Such 2-day-old pupae were processed for FRU^M/GFP coexpression as follows (same for adult specimens, see Figs. 3 and 4): Animals were dissected by using fine forceps to remove the proboscis and rip off most of one eye, along with removal of all thoracic appendages and gentle tearing open of the ventral thorax to facilitate penetration of fixative into the head capsule and thoracic box. Such application was effected by placing a partially dissected specimen in an Eppendorf tube containing 1 ml of 4% paraformaldehyde (diluted from 16% PFA solution, obtained from Electron Microscopy Sciences, Hatfield, PA) first on ice for 5 min, then on a shaker at room temperature (RT) for 10-15 min. The specimens were then rinsed twice in 1´ PBS (137 mM NaCl/2.7 mM KCl/10 mM phosphate buffer, pH 7.4), followed by dissecting them in 1´ PBS to remove completely the CNS. The prefix decreased the chance that CNS removal would result in tearing apart a given brain or ventral nerve cord (VNC); however, the former and latter major CNS regions sometimes got separated from one another during a given dissection (followed by processing of such a brain and VNC separately). An isolated specimen was placed in a porous basket (Costar, Corning, NY, Transwell polycarbonate membrane: 6.5-mm diameter, 8.0-mm pore size), which was then inserted into a well (Costar, 24-well plate) containing 1´ PBS. All subsequent steps were performed by transferring the tissue-containing basket from one well to another, militating against damage to the specimen. A given brain and/or VNC was further fixed in 4% PFA, for 20 min at RT, with gentle shaking. The specimen was then washed three times in 0.2 M phosphate buffer, then three times in TNT (0.1 M Tris·HCl/0.3 M NaCl, pH 7.4/0.5% Triton X-100), blocked in 4% normal donkey serum for 2 h at RT, then incubated in primary antibody overnight at 10-15°C, with gentle shaking. This anti-FRU^M, produced in rat and first reported by Lee et al. (1), was used at 1:200 dilution. The next day, the tissue was washed six times in TNT for 20 min each, followed by application of secondary antibody (Cy5-conjugated anti-rat IgG from donkey, Jackson ImmunoResearch, West Grove, PA), at 1:200 dilution for 2-3 h at RT in the dark, with gentle shaking. The tissue was then washed three times for 20 min each in TNT, followed by three times in 0.1 M phosphate buffer, then placed in 2% n-propyl gallate at 4°C for ~1-14 days before viewing the tissue (mounted in 2% n-propyl gallate) in a confocal microscope (see Materials and Methods). For this, a given CNS specimen was scanned at ´20-100 magnification (depending on the CNS cluster to be analyzed) by taking 1- to 2-mm sections of a given FRU^M-containing neuronal group; only the Cy5 channel was used. An all-sections Z series was projected onto the Leica’s screen; using a transparent grid placed over this screen, the border of the cluster was outlined with a pen to define the boundaries of the intra-brain or intra-VNC tissue to be analyzed. The cells within a given cluster-encompassing border were counted for each consecutive section within the right or bilaterally symmetrical left half of the brain or VNC region, until such a hemi-cluster was completely analyzed. Two investigators performed independent counts for a given CNS cluster; these values were averaged to specify per-cluster values for a given hemi-brain or hemi-VNC region within that specimen. The subsequently computed interspecimen averages (for WT male adults: Table 2) came into play in the analyses of doubly labeled specimens. For them, neurons coexpressing C309-driven GFP and FRU^M were quantified by obtaining a series of 1- to 2-mm optical sections from a given transgene-containing brain or VNC. After scanning the green and red channels (simultaneously), a Z series of all sections was projected to create a 3D image, with both channels merged to show the GFP/FRU^M overlap. For each section, the number of such white neurons was counted, within a given FRU^M CNS neuronal cluster exhibiting coexpression (see below and Table 2), using a transparency placed in front of the screen. Each white cellular signal was marked on the transparency. Numbers of 1- to 2-mm sections scrutinized varied among clusters, such that the total extent of neural material analyzed ranged from 50 to 120 mm (depending on the CNS region). The "overlap averages" were computed by taking the total numbers of white cells from the left plus right side of the cluster-defined CNS region within a given specimen, then dividing that value by the basic FRU^M cell-number for that cluster (see Materials and Methods). For data at the end of this legend, this denominator was computed by multiplying by 2 the relevant hemi-brain or hemi-VNC value from singly-labeled WT pupae in (as tabulated in ref. 1). (A) C309-driven GFP in a UAS-egfp-containing embryo: lateral aspect at ´20, anterior end toward bottom right corner, facing top right corner. (B) A C309/UAS-egfp embryo, ventral aspect at ´25. The embryonic signal pattern is broadly distributed throughout the ventral side and appears to involve generic peripheral structures, likely including PNS neurons. (C) Two-day-old male pupal brain plus VNC, carrying C309/UAS-egfp, imaged for GFP read-out and staining mediated by anti-FRU^M; anterior view at ´22; green, GFP-expressing cells; magenta, FRU^M immunoreactivity; white, coexpressing cells. (D) Anterior view of male pupal brain at ´40 (same transgenic type), depicting GFP/FRU^M overlap within clusters 2, 5, and 7 (see Table 2). (E) Posterior view of male pupal brain (same magnification as in C), showing C309/FRU^M coexpression within clusters 13 and 14. (F) VNC of male pupa at ´50, showing overlap within clusters 16 and 17. Scoring separate pupal CNS specimens (n values in parentheses) for extents of GFP/FRU^M overlap led to the following percentages (compare adult values in Table 2): cluster 2: 23 ± 4 (7), 5: 12 ± 2 (8); 7: 20 ± 2 (7), 8: 18 ± 2 (5), 13: 49 ± 3 (9), 14: 23 ± 2 (10), 16: 10 ± 2 (2), 17: 13 ± 2 (5), 18: 7 ± 2 (5), 20: 13 ± 1 (5).

1. Lee, G., Foss, M., Goodwin, S. F., Carlo, T., Taylor, B. J. & Hall, J. C. (2000) J. Neurobiol. 43, 404-426.

Table 3. Courtship song sounds produced by transgenic males

A subset of the recordings that led to the data in Table 1 and Fig. 1 was used to analyze the results of sounds produced by various male types when they were "wing-extending" at a female and potentially vibrating that appendage. The audio track from a given recording was "logged" for song pulses then analyzed for the following song parameters (all tabulated ± SEM): interpulse interval (IPI in msec), cycles per pulse (CPP), intrapulse frequency (IPF in Hz), and pulse amplitude (AMP in arbitrary dimensionless units). The lifesongx software applied (see Materials and Methods) also counted total number of pulses generated by a given male, leading to data in the rightmost column. Means for the five song values are tabulated ± SEM; n designates numbers of songs analyzed for a given male type. For 15 C309/UAS-shi^TS males recorded at 30°C, only seven of the files were logged for pulse characteristics, because the remaining eight of this doubly transgenic type generated no sounds at all at the high-temperature. To analyze interactions between GENOTYPE (GENO) and TEMPERATURE (TEMP), a two-way ANOVA was performed on each of the song parameters mentioned above for all males that carried UAS-shi^TS. For IPI, there was only a TEMP effect (P < 0.001) and no GENO or interaction effect (P values = 0.70 and 0.41, respectively). For AMP, there was a significant TEMP effect (P < 0.03), no GENO effect (P = 0.18), and a significant interaction between GENO ´ TEMP (P < 0.001). Subsequent pairwise comparisons revealed that pulses produced by C309/UAS-shi^TS males at 30°C had lower amplitudes compared with sounds produced at 25°C (P < 2.5 × 10^-4); whereas AMP for UAS-shi^TS males did not change with temperature (P = 0.34). For CPP, no GENO, TEMP, or interaction effects were detected with regard to songs of the doubly and singly transgenic males carrying shi^TS. For intrapulse frequencies, there was no GENO or interaction effect (P ³ 0.22); but a there was a significant TEMP effect for the IPF data (P = 0.01): males carrying either C309/UAS-shi^TS or UAS-shi^TS (alone) gave lower IPFs at 30°C. Analysis of the pulses/min data revealed no GENO or TEMP effect, but there was a significant interaction between the two (P = 0.01). Subsequent comparisons showed that C309/shi^TS males had a dramatic reduction in rate of pulse production when recorded at 30°C compared with 25°C (P < 0.002); in contrast, UAS-shi^TS control males sang with the same vigor at both temperatures (P = 0.60).

Table 4. Mating

Times (± SEM) between moment of pairing and start of copulation were determined for individual males each paired with a wild-type (WT) virgin female (n_T, total numbers of males tested for a given genotype). For pairs that did not mate within a given 1-h observation period, see rightmost column. For the high-temperature (Temp) tests involving control (WT) males or those carrying UAS-shi^TS, such flies were heat-shocked at 30°C for 30 min before testing at that elevated temperature. The "25" temperature entered for males carrying UAS-fru^MIR refers to doubly and singly transgenic males reared and tested in that condition; a separate group of C309/UAS-fru^MIR males was reared at 29°C then tested at 25°C. For males that mated (n_M), a one-way ANOVA on log-transformed latencies, with GENOTYPE (GENO) as the main effect, revealed significant differences among groups when tested at 25°C (F_[4,89] = 6.01, P = 0.0003). Subsequent pairwise comparisons revealed that C309/UAS-shi^TS males required equivalent times to initiate mating compared to the behavior of singly-transgenic UAS-shi^TS or WT controls, despite the nominal 2-fold latency difference. For the 25°C data, C309/UAS-fru^MIR males were no different in time to mating compared to the corresponding UAS-fru^MIR control; however, the presence of fru^MIR (whether or not C309-driven) caused longer latencies compared with values for WT males (P < 0.05). A two-way ANOVA, with GENO and TEMP as the main effects, disclosed no interaction component between the two factors (F_[5,115] = 1.58, P = 0.65).

Table 5. Attempted copulation

Individual males of the indicated genotypes were heat-shocked at 30° then paired at that temperature with one WT virgin female each. % Court, percentage of fly pairs for which any courtship occurred; % Bend, percentage of males that courted and also exhibited any abdominal bending directed at the female’s genitalia; % Mate, percentage of courting pairs that copulated.