Flies were reared on a standard corn meal, yeast, glucose agar medium at 24.5°C under a 12 h:12 h light:dark cycle. The fmn (fumin) mutant was described previously (Kume et al., 2005). TH-GAL4 was a gift from J. Hirsh. 104Y was a gift from D. Armstrong. NP lines were obtained from the Kyoto Drosophila Genetic Resource Center, To remove possible modifiers and allow comparisons in a common genetic background, we outcrossed all the alleles into the w¹¹¹⁸ or fmn background over at least five consecutive generations. As controls we used w¹¹¹⁸ or fmn flies. The transgenic RNA interference (RNAi) line for DAT (Transformant ID: 106961), UAS-Dicer-2 flies (Stock numbers: 60008) and the w¹¹¹⁸ (60000) which is the genetic background of the RNAi line were obtained from the Vienna Drosophila RNAi Center (VDRC), Vienna, Austria (Dietzl et al., 2007).

Flies were placed individually in glass tubes (length, 65 mm; inside diameter, 3 mm) containing 1% agar and 5% sucrose at 24.5°C. Male flies were used for sleep analysis. They were entrained for at least 3 d to LD conditions before being transferred to constant dark (DD) conditions. Locomotor activity was monitored by recording infrared beam crossings by individual flies in 1 min bins using the Drosophila activity monitoring system (Trikinetics, Waltham, MA, USA). Based on previous reports (Hendricks et al., 2000; Shaw et al., 2000), sleep was defined as periods of inactivity lasting 5 min or longer. For the pharmacological manipulations, 3-iodo-d-tyrosine and RU486 were obtained from Sigma-Aldrich. Flies were transferred to vehicle (1% agar + 5% sucrose) with 3 mM 3-iodo-d-tyrosine, 0.5 mM RU486. Daily sleep time was calculated with software written in Excel or R 2.11.1.

The efficiency of dDAT RNAi were examined by qPCR. Total RNA was extracted from 10 heads by using RNeasy Micro Kit (Qiagen) according to the manufacturer’s instructions. cDNA was synthesized from the total RNA using oligo (dT)₂₀ primer and ReverTra Ace reverse transcriptase (Toyobo). The cDNA was used for qPCR using THUNDERBIRD SYBR qPCR Mix (Toyobo). The primers were 5′-AACAATAGCA TCAGCGACGA-3′ and 5′- CAGGTTATGG CACCCTTACG-3′ for dDAT, 5′- TGGTACGACA ACGAGTTTGG-3′ and 5′- TTTCAGGCCG TTTCTGAAGT-3′ for GAPDH2. Before qPCR, all the primer sets were confirmed to yield one major band corresponding to each targeted mRNA on the gel when used in conventional RT-PCR. Each expression level was first normalized to GAPDH2. Then, the values were normalized to the average of independent control sample.

Brains of adult flies were dissected in cold PBS and fixed in 4% paraformaldehyde in PBS for 20 min at room temperature. After three 20 min washes in 0.3% Triton X-100 in PBS, samples were blocked and penetrated in 5% normal sheep serum (NSS) for 30 min at room temperature. Samples were then incubated in a primary antibody solution in 5% NSS at 4°C overnight. The following primary antibodies were used in this study: chicken anti-GFP (1:1000; Abcam), rabbit anti-TH (1:25 Pel-Freez), rat anti-DAT (1:1000; Millipore) and rabbit anti-dDA1 (1:1250: gift from Dr. Fred W. Wolf). After three 20 min washes in 0.3% Triton X-100 in PBS, brains were incubated at 4°C overnight in a secondary-antibody solution in 5% NSS. The following secondary antibodies were used: Alexa Fluor 568 goat anti-rat IgG (1:200; Invitrogen), Alexa Fluor 488 goat anti-chicken IgG (1:200; Invitrogen) and Alexa Fluor 568 goat anti-rabbit IgG (1:200; Invitrogen). After three 20 min washes in PBS, brains were mounted using PermaFluor (Thermo Scientific) between two coverslips separated by electrical tape of ~200 μm thickness, so that the brain sample was not flattened. Immunolabeled adult brains were imaged under a Leica TCS SP2 AOBS confocal microscope using 20x magnification. Z-stack images were scanned at 2 μm section intervals with a resolution of 1024 × 1024 pixels.

For quantification of dDA1 signal, a confocal slice image containing the holizontal lobes of the mushroom body was selected from a stack. The average pixel intensity values in the horizontal lobes and the adjacent control region (aimpr: anterior inferior medial protocerebrum) was measured by NIH ImageJ. The signal of the dDA1 in horizontal lobes was normalized by the control region of an identical confocal slice.

Olfactory conditioning with two odors (4-methylcyclohexanol and 3-octanol) was performed as described previously (Tully and Quinn, 1985). Briefly, approximately 100 flies were placed in a training chamber where they were exposed to odors and electric shocks. One of the aversive odors, OCT (3-octanol, Sigma) or MCH (4-methylcyclohexanol, Sigma), was paired with 12 pulses of electrical shocks (60 V DC) for 1 min, whereas the other was not. Testing was performed for 2 min after the training by placing flies at a choice point in a T-maze between the two odors for 2 min. If no flies moved to either arm of the T-maze or if a small number of flies were trapped in the middle compartment, it indicated that the flies did not choose either odor. A performance index (PI) was calculated so that a 50:50 distribution (no memory) yielded a PI of 0 and a 0:100 distribution away from the shock-paired odor yielded a PI of 100. Individual PIs were always the average of two experiments where the shock-paired odor was alternated.

The significance level in each experiment was set to 5%. Comparisons were made by using either Student’s t-test (for two groups) or analysis of one-way ANOVA with Tukey-Kramer HSD post hoc test (for more than two groups). Bars and error bars represent means and SEM, respectively. Data were analyzed as described in the figure legends using R 2.11.1.

As previously reported (Kume et al., 2005), the dDAT mutant fumin shows a short sleep phenotype (Figures 1A,B). To further investigate behavioral effects of dDAT expression, we examined the sleep phenotype of dDAT knockdown flies. Using GAL4/UAS system (Brand and Perrimon, 1993), we expressed dsRNA to induce RNAi in a cell specific manner. As shown in Figures 1A,B, flies with pan-neuronal knockdown of dDAT using Elav-GAL4 displayed a significant sleep reduction compared with control flies. The total daily sleep of dDAT RNAi flies decreased to about half that of control flies (Figure 1C). Next, we examined the effect of dDAT RNAi in specific cells or various brain regions using a series of GAL4 drivers. We performed this GAL4 driver screening in constant darkness since light has buffering effect against wake promoting effect of dopamine (Shang et al., 2011). It has been reported that dDAT gene expression is restricted to dopaminergic neurons (Porzgen et al., 2001). Consistent with the previous reports, dDAT knockdown in glial cells resulted in no significant sleep decrease. Flies subjected to dDAT knockdown in the dopamine neurons using Ddc-GAL4 showed a significantly short sleep time compared with control. On the other hand, dDAT knockdown in dopaminergic neurons using TH-GAL4 resulted in no significant sleep reduction. Ddc-GAL4 labels dopamine neurons including PAM cluster neurons, whereas this cluster neurons are sparcely labeled by the TH-GAL4 (Liu et al., 2012a). To check the efficiency of RNAi, we quantified dDAT mRNA level using qPCR. As shown in Figure 1D, pan-neuronal knockdown of dDAT using Elav-GAL4 decreased dDAT mRNA level to one-fifth of control. dDAT knockdown in dopaminergic neurons using TH-GAL4 or Ddc-GAL4 both showed decreased dDAT mRNA level to one-fourth of control, which is consistent with the previous report that showed dDAT expression is restricted to dopaminergic neurons. These results suggest that GAL4 expression pattern rather than GAL4 expression level contribute to the sleep phenotype. To further investigate the effect of dDAT knockdown in a subset of dopaminergic neurons, we crossed with HL5-, HL7-, HL9-GAL4 drivers. These drivers have GAL4 expression in PAM clusters whereas other clusters such as PPL1 neurons are sparcely labeled (Claridge-Chang et al., 2009; Figure 1E). Despite dense expression in PAM clusters, dDAT knockdown using these drivers showed no significant sleep reduction. We further knocked down dDAT in mushroom body and fan-shaped body. These neuropiles receive projection of dopaminergic neurons and contribute to arousal effect of dopamine (Andretic et al., 2008; Liu et al., 2012b; Ueno et al., 2012c). Consistent with Portzgen et al. that showed restricted dDAT expression in dopamine neurons, dDAT knockdown in these postsynaptic neurons has no effect on sleep phenotype. Taken these results together, we hypothesize that dDAT expression in a subset of dopamine neurons is sufficient for the reuptake of dopamine from synaptic cleft then shut down postsynaptic dopamine signaling for normal sleep.

To further investigate the effect of dDAT expression on fly sleep, we rescued dDAT in fumin mutant. Using GAL4 drivers which labels a subset of dopamine neurons, we expressed dDAT in dopaminergic neurons in fumin mutant. Although we could not observe significant sleep reduction with dDAT knockdown using TH-GAL4 (Figure 1C), fumin with dDAT expression in TH-GAL4 labeled neurons showed significant increase of sleep (Figure 2A). As described above, dopaminergic neurons except for PAM cluster neurons are labeled by TH-GAL4 line. Other GAL4 drivers which labels a subset of dopaminergic neurons, HL5- and HL7-GAL4, could also rescue the short sleep phenotype of fumin by cell specific dDAT expression. HL5-GAL4 has dense expression in PAM cluster neurons whereas PPL1 and PPM3 cluster neurons which are one of the wake promoting dopaminergic neurons are not labeled by this GAL4 driver. HL7-GAL4 also labeles PAM cluster neurons but PPL1 and PPM3 clusters are sparcely labeled (Kong et al., 2010). Based on the discrepancy between dDAT knockdown and dDAT rescue experiments, we consider that dDAT expression in a subset of dopamine neurons is sufficient to rescue the short sleep phenotype of fumin mutant. On the other hand, fumin with dDAT expression by HL9-GAL4 showed no significant increase of sleep. It might be due to its low expression level since we observed lower expression in HL9-GAL4 compared with HL5- and HL7-GAL4 drivers.

Since we could rescue the short sleep phenotype of fumin by dDAT expression in different subset of dopaminergic neurons which were insensitive to dDAT knockdown, we further investigated the effect of ectopic dDAT expression in non-dopaminergic cells. We presumed if dDAT expression in a subset of dopamine neurons is sufficient for the shut down of postsynaptic dopamine signaling, it is possible to control dopamine signaling by dDAT expression in postsynaptic neurons or extrasynaptic cells such as glia.

In order to express dDAT in various regions of the brain in a fumin background, we outcrossed the GAL4 drivers to the fumin mutant. Next, we examined the effect of dDAT expression on the short sleep phenotype of the fmn mutant. We found that fmn flies with dDAT expression in non-dopaminergic neurons showed a significant sleep increase (Figure 2A). Mushroom body (Joiner et al., 2006; Pitman et al., 2006) and fan-shaped body (Donlea et al., 2011) receive dopaminergic projection and regulates sleep and arousal. Although we observed no significant changes in sleep by dDAT knockdown in these postsynaptic structure (Figure 1C), dDAT expression in these neuropile rescued short sleep of fumin mutant. In addition, dDAT expression in glial cells also rescued short sleep in fumin. To confirm that the drivers have no GAL4 expression in dopamine neurons, we expressed GFP using the GAL4/UAS system and then counter stained with anti-TH antibody. As the merged signal patterns shown in Figure 2B, TH-GAL4 drives GFP expression in most, but not all, dopamine neurons. Conversely, we could not detect any merged signals when expressing GFP using the other drivers, which suggests that these drivers only expressed dDAT in non-dopaminergic neurons. Next, we analyzed the localization of ectopically expressed dDAT using immunohistochemistry. Although we could not detect endogeneous dDAT expression due to high background with the antibody, ectopic expression of dDAT in the fmn mutant using the 30Y-GAL4 driver, which has strong expression in the mushroom body, showed dDAT immunostaining in all lobes and calyx of the mushroom body (Figure 2C). These lobes and calyx are composed of axon bundles and dendrites of Kenyon cells, thus ectopically expressed dDAT appears to be transported and localized to the synaptic area.

To further investigate if the effect of dDAT expression in non-dopaminergic neurons is due to developmental changes or not, we expressed dDAT in postsynaptic neurons or glial cells using spatiotemporal expression system, TARGET system (McGuire et al., 2003). In this system, dDAT expression is suppressed at 19°C by the temperature sensitive GAL4 inhibitor GAL80^ts which is under the control of tubulin promotor. After shifting to 31°C, GAL80^ts was inactivated and dDAT expression will be induced. All the flies were raised at 19°C to adulthood. At the permissive temperature, progeny from the experimental crosses showed similar sleep patterns to control flies. The temperature shift to 31°C significantly increased sleep time of flies which has temporal dDAT expression in mushroom body and glial cells (Figure 3). These results suggest that dDAT expression in non-dopaminergic neurons, such as the mushroom body or glial cells, can rescue the short sleep phenotype in fmn without developmental changes.

The D1-like receptor dDA1 is a dopamine receptor expressed in the mushroom body and in central complex structures such as the ellipsoid body and fan-shaped body. dDA1 expression in the mushroom body mediates olfactory memory (Kim et al., 2007). As shown in Figure 4A, fmn mutants demonstrated decreased dDA1 expression in the mushroom body. Quantification of dDA1 expression based on image analysis of immunohistochemistry results (Knapek et al., 2011) also confirmed reduced dDA1 expression in the mushroom body (Figure 4B). Administration of a dopamine synthesis inhibitor (tyrosine hydroxylase inhibitor) rescued the short sleep phenotype of fmn mutants (Figure 4C) and increased dDA1 expression in the mushroom body (Figures 4D,E). Conditional rescue of dDAT in all neurons using the GeneSwitch system (Osterwalder et al., 2001) increased dDA1 expression in the mushroom body (Figure 4F). Without administration of RU486, there was also significant leakage of dDAT expression in the mushroom body as previously pointed out (Poirier et al., 2008). To investigate whether the decreased dDA1 expression was the result of short sleep or increased dopamine signaling, we checked the dDA1 expression in another sleep mutant. Both calcineurin knockout and knockdown result in the short sleep phenotype (Nakai et al., 2011; Tomita et al., 2011), but dDA1 expression in the calcineurin mutants was not reduced (Figure 4G). Taken these results together, increased dopamine signaling, rather than short sleep in fmn mutants, downregulates dDA1 expression in the mushroom body. Thus dDA1 expression level is applicable as an indicator of the dopamine signaling intensity.

To examine the effect of dDAT expression on dopamine signaling, we verified the dDA1 level in fmn with dDAT expression. As shown in Figure 5, dDAT expression in a subset of dopaminergic neurons with TH-GAL4 rescued dDA1 expression in fmn. In addition, dDAT expression in the mushroom body (121Y-GAL4, 11Y-GAL4), glial cells (dEAAT1-GAL4) and pars intercerebralis (c767-GAL4) increased dDA1 expression in the mushroom body, indicating decrease of dopamine signaling. These results are consistent with the rescued sleep phenotype in fmn by ectopic dDAT expression (Figure 2A). On the other hand, npf-GAL4 driven dDAT expression in fmn flies could not rescue the short sleep phenotype (Figure 2A) and showed no significant changes in dDA1 expression in the mushroom body (Figure 5). These results indicate that the ectopic expression of dDAT suppresses the increased dopamine signaling in the fmn mutant.

Next, we investigated the effect of dDAT expression on aversive olfactory memory formation. Dopamine signaling mediates associative olfactory memory via dDA1 receptor in mushroom body (Kim et al., 2007). As previously reported (Zhang et al., 2008), fmn mutants have an impaired aversive olfactory memory due to excessive dopaminergic signaling (Figure 6A). dDAT expression in glial cells (using Repo-GAL4) as well as dopamine neurons (using TH-GAL4) promoted memory formation in fmn mutants although we could not find statistical significance. Surprisingly, however, when some GAL4 drivers which are expressed in the mushroom body (such as 30Y-GAL4 and OK107-GAL4) were used to drive dDAT, these flies did not demonstrate aversive olfactory memory (Figure 6A). 30Y-GAL4 and OK107-GAL4 have broad and strong GAL4 expression in the mushroom body (Aso et al., 2009), which suggest that post-synaptic strong dDAT expression abolished dopamine signaling after activation of dopamine neurons with electric punishment. Another possibility is that strong expression of dDAT in the mushroom body caused a developmental defect in the Kenyon cells since flies with chemical ablation of mushroom body are unable to perform in a classical conditioning paradigm (de Belle and Heisenberg, 1994). To exclude this possibility, we checked the memory performance of flies with temporal dDAT expression using Elav-GeneSwitch. Without administration of RU486, there was significant leakage of dDAT expression which may have affected memory performance (Figures 6B,C), since there was a slight improvement in the PI even in flies fed with vehicle (compare control of Figure 6A and vehicle of Figure 6B). However, administration of RU486 results in the impairment of memory performance in a dose dependent manner. Furthermore, the flies fed with 500 μM of RU486 did not demonstrate associative memory (Figure 6B). These results suggest that strong expression of dDAT at the post-synaptic site results in abolition of dopamine signaling after activation of dopamine neurons.