Dopamine Signalling in Mushroom Bodies Regulates Temperature-Preference Behaviour in Drosophila

The ability to respond to environmental temperature variation is essential for survival in animals. Flies show robust temperature-preference behaviour (TPB) to find optimal temperatures. Recently, we have shown that Drosophila mushroom body (MB) functions as a center controlling TPB. However, neuromodulators that control the TPB in MB remain unknown. To identify the functions of dopamine in TPB, we have conducted various genetic studies in Drosophila. Inhibition of dopamine biosynthesis by genetic mutations or treatment with chemical inhibitors caused flies to prefer temperatures colder than normal. We also found that dopaminergic neurons are involved in TPB regulation, as the targeted inactivation of dopaminergic neurons by expression of a potassium channel (Kir2.1) induced flies with the loss of cold avoidance. Consistently, the mutant flies for dopamine receptor gene (DopR) also showed a cold temperature preference, which was rescued by MB–specific expression of DopR. Based on these results, we concluded that dopamine in MB is a key component in the homeostatic temperature control of Drosophila. The current findings will provide important bases to understand the logic of thermosensation and temperature preference decision in Drosophila.


Introduction
To counteract the harmful effects of environmental temperature fluctuation, animals use homeostatic thermoregulatory mechanisms and/or behavioral responses to maintain optimal body temperature. Mammals regulate their body temperatures with various physiological responses such as change of metabolic rates, vasomotor control, and sweating. Due to the large surface area-tovolume ratio of its body, Drosophila maintains body temperature by heat exchange with the surrounding environment. In a genetically controlled process, flies track to an environmental temperature as close as possible to their desired body temperature [1][2][3]. The behavioural responses in Drosophila can be recorded easily by examining TPB on a thermal gradient plate [1,4,5]. Wild type fruit flies show a strong response to external temperature and robust TPB peaking at 24uC on the temperature gradient [4].
To identify the molecular components and neural circuits regulating thermosensation and body temperature regulation, various studies such as calcium imaging, genetic screens, and electrophysiology have been done in fruit flies, C. elegans, mice, and cultured cells [6][7][8][9][10][11][12]. From these, various thermo-transient receptor potential (TRP) channels and cyclic nucleotide-gated channels have been found to be molecular sensors of temperature. The perceived temperature information generated in neurons bearing these sensor molecules is transmitted to a higher order brain center. Here, the sensed temperature is compared to a set point temperature to determine whether it is too low or too high [2]. In vertebrate, the hypothalamic preoptic area is known to be the higher order brain center interpreting the temperature information [13].
In Drosophila, the temperature that is instinctively preferred plays a role equivalent to the set point temperature in mammals. Three thermo-TRP channels have been reported as molecular sensors of hot or warm temperature in Drosophila. Painless, a thermo-TRP channel, is important for avoidance of hazardous heat above 40uC [14]. Pyrexia, another thermo-TRP channel, is activated around 40uC and acts to prevent paralysis during high temperature stress [15]. dTRPA1 is activated above 27uC and essential for avoidance of warm temperature, and the loss-offunction mutants of dTRPA1 prefer warm temperature [16]. In addition to the discovery of these thermo-TRP channels, various genetic studies to understand the brain center, and the responsible neuromodulators and neural circuits of TPB in Drosophila have been actively pursued [1,5,16].
The adult Drosophila brain includes about 100,000 neurons whose projections cluster in neuropil structures to make two symmetric hemispheres [17]. The central complex and the mushroom body (MB) are two prominent structures in the central brain. The central complex is located centrally between the two hemispheres and consists of four interconnected neuropils: the ellipsoid body, the fan-shaped body, the protocerebral-bridge, and the nodule [18,19]. The central complex is involved in behaviours such as locomotion, visual flight, courtship, and olfactory learning task [20,21].
However, MB is involved in behaviours like sleep [22], decision making [23], and learinig [24]. Recently we showed that MB acts as a higher order brain center regulating TPB in Drosophila [1]. MB is made up of approximately 2,500 Kenyon cells in each hemisphere, which send out dendrites to the calyx [24]. Kenyon cells are grouped into one of three neuron classes, whose axons form morphologically and functionally distinct lobes of the ab, a9b9 and c lobes. For TPB, the ab lobes are necessary while the c lobe is dispensable [1]. We also showed that the cAMP-PKA signalling pathway is the key regulator working inside of MB to determine the temperature that Drosophila prefers [1].
Fifteen clusters of six-hundred dopaminergic neurons are distributed in the Drosophila brain [25,26]. The dopaminergic neurons are broadly and stereotypically localized, and show particularly dense projections into MB [27]. For example, protocerebral posterior lateral 1 cluster which innervates MB is reported as routes for reinforcement and retrieval in memory formation [28,29]. Similar to mammals, dopamine receptors are divided into two major subfamilies in Drosophila. Two D1-like receptors, DopR and DpoR2, and a D2-like receptor, D2R, have been identified [30]. Interestingly, DopR is highly expressed in MB and the central complex [31], and mediates learning [32,33], caffeine-induced wakefulness [34], arousal [35], and ethanolstimulated locomotion [36]. DopR2 is expressed in MB and is involved in larval aversive olfactory learning [37,38]. D2R is expressed in some discrete cells in central nervous system and regulates locomotor activity [30]. However, it has remained unknown that dopamine is involved in TPB regulation in Drosophila.
To uncover the relationship between the dopamine system and TPB, we performed various genetic studies in Drosophila. Dopamine synthesis and dopaminergic neurons are critical in TPB regulation. In addition, we showed that DopR is essential to control TPB in the ab lobes of MB. These results strongly suggested that dopamine is an important neuromodulator for TPB in Drosophila.

Dopamine synthesis is required for TPB regulation
Our previous results have shown that cAMP signalling in MB modulates TPB [1]. Since dopamine is an important modulator of cAMP [39,40] and dopaminergic neurons have a strong innervation profile in MB [27], we suspected that the dopamine system can modulate TPB in Drosophila.
The enzymatic reaction of tyrosine hydroxylase (TH), which converts tyrosine to L-DOPA ( Figure 1A), is the rate limiting step in dopamine biosynthesis [25]. TH is encoded by pale locus in Drosophila. It is also known that loss of one copy of the TH gene can cause severe reduction of dopamine levels in the animal brain [25]. This led us to test a pale mutant for TPB as the first step to explore the involvement of dopamine signalling in the regulation of TPB. In the TPB assay, the experimental procedure was described previously [1]. The Avoidance Index against Low Temperatures (AI Low ) and Avoidance Index against High Temperatures (AI High ) were calculated using the formulas shown in Figure 1B [1,6]. The pale 4 /+ heterozygote showed abnormal TPB and the animals spread over a wide range of temperature colder than 24uC ( Figure 2A). Their AI Low was less than 34% of the wild type control w 1118 , while AI High was normal ( Figure 2A and Table S1).
To reduce the biosynthesis of dopamine further, we fed pale 4 /+ flies with a TH inhibitor, a-methyl-p-tyrosine methyl ester (AMPT), for 4 days. In the TPB assay, these flies showed a more severe defect in their ability to avoid low temperature ( Figure 2A). The flies spread almost randomly over all temperature ranges below 24uC (Figure 2A). The AI Low of the flies was below 0 ( Figure 2A and Table S1).
On the other hand, the reduced AI Low of pale 4 /+ flies was ameliorated by introducing a genomic DNA fragment of the pale locus into the pale 4 /+ flies ( Figure 2B). The genomic DNA fragment of the pale locus was reported to rescue pale flies from lethality [41]. The AI Low of pale 4 /+ flies with the pale genomic DNA was 91% of wild type ( Figure 2B and Table S1). This confirmed that the abnormal TPB of the pale 4 /+ flies was indeed resulted from a defect in the TH gene. These results strongly

Author Summary
Temperature affects almost all aspects of animal development and physiological processes. The dependence of the body temperature of small insects on ambient temperature and other heat sources makes it plausible that neuronal mechanisms for sensing temperature and behavioral responses for maintaining body temperature in a permissive range must exist. By using the fruit fly model system and previously settled paradigms of temperaturepreference test, we find that dopamine regulates temperature-preference behaviours. Wild-type flies show a strong temperature preference for 25uC, but inhibition of dopamine biosynthesis by genetic mutations or treatment with chemical inhibitors causes animals to prefer temperatures colder than normal. We also show that dopaminergic neurons are involved in the regulation of temperaturepreference behaviours and that dopamine signalling in mushroom body neurons plays a critical role in regulating the behaviours. These results suggest that dopamine is a key component in the homeostatic temperature control of fruit flies.
suggested that dopamine is a key neurotransmitter controlling fly temperature preference.
To convince ourselves further that dopamine is involved in TPB regulation, we examined a dopa decarboxylase (DDC) mutant. DDC converts L-DOPA to dopamine ( Figure 1A). As shown in Figure 2C, flies homozygous for a DDC hypomorphic allele Ddc DE1 showed reduced cold avoidance. The AI Low of Ddc DE1 / Ddc DE1 flies was 26% of wild type ( Figure 2C and Table S1).
To examine whether further reduction of DDC activity worsens the ability of flies to avoid cold temperature, we fed the Ddc DE1 / Ddc DE1 flies with 3-hydroxy benzyl hydrazine (HBH). HBH is a potent DDC inhibitor [42]. As expected, the DDC mutant flies fed HBH showed almost complete loss of cold avoidance in the TPB assay ( Figure 2C). Some flies even moved to the coldest region of the temperature gradient plate ( Figure 2C). The AI Low of these flies was 20.23 ( Figure 2C and Table S1).
These data suggested that dopamine biosynthesis is critical for flies to avoid cold temperature. This was supported further by over expression experiments. When TH or DDC was over expressed separately in dopaminergic neurons, the flies showed normal TPB (Figure 3A, 3B and Table S2). However, when TH and DDC were over expressed together in dopaminergic neurons, the flies became thermophilic or hypersensitive to cold temperature ( Figure 3A and 3C). Their AI Low was higher than that of wild type while their AI High was only 39% of wild type ( Figure 3C and Table S2).

Dopaminergic neurons are required for TPB regulation
To gain more evidence of TPB control by dopamine, we examined whether the normal function of dopaminergic neurons is essential for normal TPB. We over expressed tetanus toxin (TNT) and potassium channel Kir2.1 in dopaminergic neurons with TH-Gal4 and dopamine transporter (DAT)-Gal4. TNT only blocks the neural activity requiring n-Syb-dependent synaptic vesicle fusion [43], while Kir2.1 hyperpolarizes all neurons and suppresses action potential firing [43,44]. Unfortunately, Kir2.1 when combined with TH-Gal4 or DAT-Gal4 killed flies before they hatched. In contrast, TH-Gal4.TNT and DAT-Gal4.TNT flies were viable and both spread over temperature ranges broader than wild type, mostly over the areas colder than 25uC ( Figure 4A, 4B and Table S3).
To exclude the possibility of developmental defects induced by TNT and Kir2.1, we combined UAS-TNT and UAS-Kir2.1 with the Temporal and Regional Gene Expression Targeting system (TARGET) [45], which employs temperature-sensitive Gal80 protein (Gal80 ts ) under the control of tubulin promoter (tub-Gal80 ts ). In this system, the transcriptional activator of Gal4 is suppressed at 18uC but active at 32uC. As expected, TH-Gal4.TNT flies with tub-Gal80 ts showed normal TPB when they were raised continually at 18uC ( Figure 5A and 5C). However, they unexpectedly showed normal TPB even after exposure to 32uC for 16 hr to induce TNT before the TPB assay ( Figure 5B and 5D). We obtained similar results when we used DAT-Gal4 instead of TH-Gal4 (data not shown). It might be possible that temporal induction of TNT for 16 hr in the adult stage does not accumulate enough TNT to inactivate dopaminergic neurons sufficiently, while life-time induction of TNT with TH-Gal4 or DAT-Gal4 without Gal80 ts does accumulate sufficient TNT. It is also known that TNT might not block synaptic vesicle release strongly enough to block all neuronal activities [43,46]. On the other hand, it cannot be excluded that the abnormal TPB of TH-Gal4.TNT and DAT-Gal4.TNT flies shown in Figure 4A and 4B was resulted from developmental defects induced by TNT.
From the above results, we concluded that defective dopamine biosynthesis or inactivation of dopaminergic neurons causes the fly has abnormal TPB. These results strongly suggested that dopamine signalling controls TPB in Drosophila.

DopR is important for TPB
We next examined whether the activities of dopamine receptors are required for TPB. Drosophila has three subtypes of dopamine receptor; DopR, DopR2, and D2R [30]. dumb 3 is a hypomorphic allele of DopR, in which most MB expression of DopR was removed by a P element insertion [33]. dumb 3 /dumb 3 homozygote flies exhibited reduced ability to avoid cold temperature ( Figure 6A). Their AI Low was lower than 50% of wild type ( Figure 6A and Table S5). Consistently, flies having one copy of a deficiency chromosome Df(3R)red-P52, in which DopR is deleted, preferred temperatures colder than normal ( Figure 6B). The axis of distribution profile of flies over the temperature gradient was shifted from 24uC to 21uC ( Figure 6B). More convincingly, when this deficiency chromosome was combined with dumb 3 chromosome, resulting in the genotype of dumb 3 /Df(3R)red-P52, the flies lost cold avoidance furthermore ( Figure 6C). Their AI Low was 23% of wild type ( Figure 6C and Table S5).
There are no available mutant alleles for DopR2 and D2R. Therefore, we employed dsRNAi technology with and without UAS-Dcr2, which is known to enhance the effect of dsRNAi in neurons [47]. However, none of the RNAi transgenes reduced the transcript levels of these two receptors by more than 40% and the flies had normal TPB. Accordingly, we could not ascertain whether DopR2 or D2R is required for TPB control.

DopR regulates TPB in MB
Taken together, our results shown above demonstrated that dopamine signalling is a key factor controlling TPB in Drosophila. This finding raised the question as to which brain parts or neurons are the targets of dopamine signalling in TPB control. A clue was obtained from the results that we reported recently, which showed that MB is the brain center controlling TPB [1]. Importantly, DopR and DopR2 are expressed in MB [31]. However, they are also expressed in other Drosophila nerves, including the central complex and several neuro-secretory cells in the adult brain [31]. We confirmed that DopR is highly enriched in MB and DopR is also expressed in the central complex ellipsoid body, fan-shaped body, and nodules as previously reported ( Figure 7A and [31]). In dumb 3 /dumb 3 homozygote flies, DopR gene expression was strongly reduced in overall MB and the central complex ( Figure 7B).
To examine if MB is the location where DopR regulates TPB, we generated flies in which DopR expression is maintained in MB. This was achieved by ectopic expression of the UAS-DopR transgene with c309 or MB247 in the DopR mutant background of dumb 3 /dumb 3 ( Figure 7C and 7D). c309 and MB247 drive gene expression specifically in MB.
As shown in Figure 7C, dumb 3 /dumb 3 flies spread widely over the temperature gradient lower than 25uC. In contrast, the artificial restoration of DopR levels with c309.UAS-DopR transgene enabled dumb 3 /dumb 3 to gather sharply around the 25uC area ( Figure 7C). In other words, the flies displayed almost completely restored normal TPB. The AI Low of the flies was 99% of wild type ( Figure 7C and Table S6). Consistently, c309.DopR;dumb 3 /dumb 3 flies also showed rescued TPB compared with c309/+;dumb 3 /dumb 3 and UAS-DopR/+;dumb 3 /dumb 3 flies ( Figure S1A). This means that the abnormal TPB of dumb 3 /dumb 3 homozygote flies was caused by reduced expression of DopR in the region of c309 expression and that MB may be the location where DopR regulate TPB. However, we could not completely exclude the possibility that the central complex participated in the TPB regulation of DopR as the expression of c309 was observed in other regions of the brain including the central complex ( Figure S2A and [48]).
To confirm the specific role of MB, we next tested MB247 which strongly induces gene expression in the ab and c lobes with very low background expression ( Figure S2B and [49]). When we used MB247 instead of c309 ( Figure 7D, Figure S1B, and Table  S6), MB247. DopR;dumb 3 /dumb 3 flies also regained preference for 25uC, which is the normal preferred body temperature ( Figure 7D and Figure S1B). They showed the AI Low and AI High values comparable to those of wild type flies ( Figure 7D and Table S6). This means that the restoration of DopR levels only in MB enabled dumb 3 /dumb 3 to behave like wild type flies. These data indicated that dopamine signalling in MB is critical for TPB.
DopR expression in the ab lobes of MB is critical in TPB As described above, both c309 and MB247could restore the TPB of dumb 3 /dumb 3 with UAS-DopR ( Figure 7C and 7D). c309 induces gene expression strongly in the ab and c lobes, but weakly in the a9b9 lobes ( Figure S2A and [49]). MB247 induces gene expression in the ab and c lobes but not in the a9b9 lobes ( Figure  S2B and [49]). To investigate which lobe of MB is relevant to the DopR-mediated regulation of TPB, we used two more MB-specific Gal4 lines. c739.DopR;dumb 3 /dumb 3 showed restored TPB and avoidance index ( Figure 9A and Table S8). c739 is expressed in the ab lobes ( Figure S2C and [50]). However, 1471. DopR did not rescue the phenotype of dumb 3 /dumb 3 . The flies spread widely over the low and intermediate regions like dumb 3 /dumb 3 and AI Low was not restored to the level of wild type control ( Figure 9B and Table  S8). This 1471 induces gene expression in the c lobe ( Figure S2D and [51]). Taken together, the results suggested that the ab lobes are relevant to DopR-mediated regulation of TPB and the c lobe is dispensable.

DopR in the central complex is not involved in TPB
Finally, we examined whether the dopamine signalling in the central complex is required for normal TPB using two more Gal4 lines. c161 is expressed in the ellipsoid body of the central complex ( Figure S2E and [18]). The ectopic expression of UAS-DopR by c161 in dumb 3 /dumb 3 did not rescue the defective TPB. c161.DopR;dumb 3 /dumb 3 flies spread over the areas colder than 25uC like dumb 3 /dumb 3 and the lowered AI Low of dumb 3 /dumb 3 was not restored ( Figure 10A and Table S9). OK348 is expressed in the fan-shaped body ( Figure S2F and [52]). OK348.DopR;dumb 3 / dumb 3 flies also showed loss of cold avoidance in the TPB assay and AI Low was low like dumb 3 /dumb 3 ( Figure 10B and Table S9). These implicated that the ellipsoid body and the fan-shaped body are not relevant to DopR-mediated regulation of TPB.

Discussion
Monoamine neurotransmitters such as dopamine, norepinephrine, and serotonin have been identified as important factors in thermoregulation in mammals [53][54][55][56]. However, the exact functions of these monoamines in body temperature control are not clearly understood due to limited experimental methods in mammalian systems. Fortunately, our current study clearly revealed that dopamine plays a critical role in regulating TPB of Drosophila. Dopamine-deficient mutants and dopaminergic neuroninactivated flies lost cold avoidance significantly (Figure 2, Figure 4, and Figure 5). However, the flies overexpressing TH and DDC together in dopaminergic neurons became thermophilic or hypersensitive to cold temperature (Figure 3). This suggests that dopamine levels are critical to regulate TPB; high dopamine levels lead to warm temperature preference and low dopamine levels lead to cool temperature preference.
In this study, we have shown evidence to support that the dopamine signaling in MB regulates TPB. We also demonstrated previously that cAMP signaling in MB modulates TPB; flies with low levels of cAMP prefer cold and the flies with high levels of cAMP prefer warm temperature [1]. In addition, there have been  reports that dopamine regulates cAMP levels [39] and that dopamine stimulation of Drosophila brain leads to activation of PKA specifically in the a lobe [40]. Therefore, we propose that when a distinctive subset of dopaminergic neurons which innervate the TPB center in MB is activated, cAMP levels and PKA activity can be increased in the center. Activated PKA may phosphorylate and regulate various targets including ion channels and also induce gene expression of unidentified TPB regulatory genes with CREB binding sites. These PKA-dependent gene expression and posttranslational regulation would change the activities of the TPB center in MB, and fly can determine specific temperature preference or avoidance behaviours. Consistently, our unpublished data strongly suggest that a specific gene group induced by PKA-CREB signalling in the brain is critical for Drosophila TPB.
Previously, it was reported that MB is important for both cold and warm temperature avoidance and preference [1]. Interestingly, all the flies with defective dopamine signalling showed a significant loss in cold avoidance, but some of them also showed loss of warm avoidance (Figure 2, Figure 4, Figure 5, Figure 6). However, the loss of warm avoidance was not severe as MBdefective flies and dTRPA1-deficient flies [1,16]. We do not think that only dopamine regulates TPB in MB sufficiently. It is possible that other neuromodulators and/or multiple innervated neurons may be involved in modulation of the MB process to regulate TPB. The additive or synergistic effect of dopamine and these modulators may regulate cAMP levels in MB and TPB with broader temperature ranges as shown in MB-defective mutants or cAMP signaling-disrupted flies [1].
It is also possible that dopaminergic neurons are involved in transmitting temperature information to MB. It was reported that TRP and TRPL are required for avoidance of cold temperature in larvae [57], and their cold avoidance phenotypes resemble the TPB of dopamine signalling-deficient flies. Understanding whether the neuronal pathway for the cold avoidance of TRP and TRPL is connected to the dopaminergic pathway of MB will be helpful to understand the regulatory mechanism of TPB. In another report, the flies with surgically removed third antennal segment and aristae showed loss in cold avoidance [16]. Therefore, it is also possible that dopaminergic neurons are involved in transmitting the sensory signals to MB from sensory tissues like the third antennal segment and aristae.
Although three subtypes of dopamine receptors are known in Drosophila, we found that DopR in MB is involved in TPB regulation while DopR in the ellipsoid body and the fan-shaped body is not essential for TPB control. This is consistent with the previous study that these substructures of the central complex are not involved in TPB control [1]. According to the previous results from others and our current data, DopR regulates multiple Drosophila behaviours via distinct neural circuits. DopR in the ellipsoid body is required to promote ethanol-stimulated locomotion and to regulate exogenous arousal negatively, while DopR in PDF-expressing circadian pacemaker cells is needed to regulate endogenous arousal positively [35,36]. Moreover, DopR in MB is required in learning and memory [32]. In addition, all parts of MB may not work for TPB regulation; DopR in the ab lobes of MB is relevant but this in the c lobe is dispensable, which is also consistent with the previous report [1].
In conclusion, our results strongly suggest that dopamine controls TPB and body temperature in Drosophila. We believe that our findings provide important clues to understand the molecular mechanism of Drosophila TPB. Although further studies are required, we cautiously suggest that our fruit fly system can provide a highly useful model to further understand the physiological roles of dopamine in animal body temperature regulation.

Fly strains and transgenic lines
Fly stocks were raised on standard cornmeal food at 25uC and 40%-50% relative humidity. Canton-S and w 1118 flies all showed a strong temperature preference for 24-25uC regardless of the temperatures at which they were reared in every TPB tests [1,4].   [41]. DAT-Gal4 was generated as reported previously [58]. The UAS-DopR construct was generated as follows: a 1.8 Kb fragment including DopR coding sequences was obtained from RT-PCR of fly adult head mRNA with the primers tcgctgaaaagagggaagcaa and cagtaggtagagggctggg; the fragment was cloned into pGEM-T Easy vector (Promega) and sequenced, and the fragment was transferred into NotI and XbaI sites of pUAST vector. We produced the UAS-DopR transgenic flies with standard microinjection methods.
TPB assay TPB was performed as described previously [1]. Files were incubated under 12 hr:12 hr dark:light cycle for 4 days before TPB assays. A linear temperature gradient from 15 to 45uC was established along the aluminium block. All TPB assays were done in a room with 3565% relative humidity. The AI Low and AI High were calculated as shown in Figure 1B. A mixed population of both sexes was tested for TPB. The number of flies was recorded and processed using Microsoft Excel. The AI Low or AI High of each tested fly line was compared with the AI Low or AI High of the control fly lines and then t-tests were performed. *, p,0.05; **, p,0.01.

Drug treatments
Two to three days old flies were raised on drug-containing food for 4 additional days and tested. Melted pre-made standard dextrose media were blended with drugs or the same amount of distilled water, their solvent. As the final drug concentrations, 20 mM a-methyl-p-tyrosine methyl ester (AMPT) (Sigma) and 20 mM 3-hydroxy benzyl hydrazine (HBH) (Sigma) were used. The standard dextrose media contain dextrose 70 g, yeast flake 50 g, cornmeal 35.3 g, agar 5 g, tegosept 7.3 ml and propionic acid 4.7 ml in 1 L total volume.

Immunohistochemistry
To confirm various brain-specific Gal4 fly lines (c309, MB247, c739, 1471, c161 and OK348), they were combined with UAS-LacZ. Adult brains were removed from the head capsules and fixed in 4% paraformaldehyde in phosphate buffered saline (PBS) for 30 min, and rinsed in PBST (PBS containing 0.5% triton X-100). Brains were incubated with anti-b-galactosidase (1:1000; Promega) and then incubated with the appropriate red fluorescent labelled secondary antibody (Jackson Laboratories).