The nuclear receptor DHR3/Hr46 is required in the blood brain barrier of mature males for courtship

The blood brain barrier (BBB) forms a stringent barrier that protects the brain from components in the circulation that could interfere with neuronal function. At the same time, the BBB enables selective transport of critical nutrients and other chemicals to the brain. Many of these processes are still poorly understood. Beyond these functions, another recently recognized function is even less characterized, specifically the role of the BBB in modulating behavior by affecting neuronal function in a sex dependent manner. Notably, signaling in the adult Drosophila BBB is required for normal male courtship behavior. Courtship regulation also relies on male-specific molecules in the BBB. Our previous studies have demonstrated that adult feminization of these cells in males significantly lowered courtship. Here, we conducted microarray analysis of BBB cells isolated from males and females. Findings revealed that these cells contain male- and female-enriched transcripts, respectively. Among these transcripts, nuclear receptor DHR3/Hr46 was identified as a male-enriched BBB transcript. DHR3/Hr46 is best known for its essential roles in the ecdysone response during development and metamorphosis. In this study, we demonstrate that DR3/Hr46 is specifically required in the BBB cells of mature males for courtship behavior. The protein is localized in the nuclei of sub-perineurial glial cells (SPG), indicating that it might act as a transcriptional regulator. These data provide a catalogue of sexually dimorphic BBB transcripts and demonstrate a physiological adult role for the nuclear receptor DH3/Hr46 in the regulation of male courtship, a novel function that is independent of its developmental role. Author summary The blood brain barrier very tightly regulates which molecules can enter the brain. This is an important protection for the brain, however, it also complicates communication between molecules in the circulating fluid and the brain. In fly courtship, for example, circulating male-specific products are crucially required for normal courtship. But the neuronal circuits that ultimately control the behavior are inside the brain, separated from these molecules by the blood brain barrier. The mechanisms of this communication are not known. Here we show that the blood brain barrier itself contains sex-specific RNAs and we show that one of them, a nuclear receptor called DHR3, is required in adult males for normal courtship. These findings promise new insight into the communication between blood brain barrier and the brain.

specifically the role of the BBB in modulating behavior by affecting neuronal 23 function in a sex dependent manner. Notably, signaling in the adult Drosophila BBB 24 is required for normal male courtship behavior. Courtship regulation also relies on 25 male-specific molecules in the BBB. Our previous studies have demonstrated that 26 adult feminization of these cells in males significantly lowered courtship. Here, we 27 conducted microarray analysis of BBB cells isolated from males and females. 28 Findings revealed that these cells contain male-and female-enriched transcripts, 29 respectively. Among these transcripts, nuclear receptor DHR3/Hr46 was identified 30 as a male-enriched BBB transcript. DHR3/Hr46 is best known for its essential roles 31 in the ecdysone response during development and metamorphosis. In this study, we 32 demonstrate that DR3/Hr46 is specifically required in the BBB cells of mature males 33 for courtship behavior. The protein is localized in the nuclei of sub-perineurial glial 34 cells (SPG), indicating that it might act as a transcriptional regulator. These data 35 provide a catalogue of sexually dimorphic BBB transcripts and demonstrate a 36 physiological adult role for the nuclear receptor DH3/Hr46 in the regulation of male 37 courtship, a novel function that is independent of its developmental role. 38

Introduction 53
It is well established that the two layers of glial cells that tightly surround the 54 nervous system form the Drosophila blood brain barrier (BBB)(1). Flies have a non-55 vascular open circulatory system that distributes the hemolymph. The BBB forms 56 the tight exclusion barrier that is essential to protecting neurons from hemolymph 57 components that could interfere with neuronal function (2, 3). At the same time, the 58 barrier needs to allow selective uptake of nutrients and other molecules needed for 59 brain function. The Drosophila blood brain barrier (BBB) surrounds the brain like a 60 tight cap. It consists of two layers of glial cells. The outer perineurial glia cells (PG 61 cells) are thought to function as a barrier for large-molecular weight molecules. The 62 inner layer, the subperineurial glia (SPG), is adjacent to the neuronal cell bodies and 63 contains the tight junctions that form the physical barrier ( Fig 1A). It has been 64 shown in a number of genetic and functional studies that the barriers in flies and 65 vertebrates share not only structure and function, but also many homologous 66 proteins that ensure their function, as shown in (4). A recent microarray study of 67 isolated BBB cells has expanded on these earlier findings and shown that besides 68 the characteristic barrier proteins, fly and mouse BBB cells share a large number of 69 conserved proteins (5). That study has also provided the first detailed "inventory" of 70 these cells in Drosophila. While the barrier and selective uptake functions of the BBB 71 are its most obvious essential function, evidence is starting to accumulate that other 72 physiological processes in BBB cells are contributing to brain function. For example, 73 the G-protein-coupled receptor moody is specifically expressed in the subperineurial 74 glial cells (SPG)(6, 7). While the absence of both moody isoforms leads to a leaky 75 BBB (6, 7), mutants with only one of the isoforms have intact barriers, but 76 behavioral defects in their response to cocaine and alcohol (6). In addition, moody, 77 in a function independent of its function in barrier integrity, is also required in BBB 78 cells for normal male courtship (8). That active signaling processes in BBB cells 79 regulate neuronal output was further indicated by the finding that BBB-specific 80 reductions in the G protein Galpha(o) cause courtship defects, while leaving the 81 barrier integrity intact (8). It has been found that the circulating hemolymph 82 contains male-specific factors from the fat body that are needed to ensure normal 83 courtship (9). It is not clear how these factors interact with the male brain circuits 84 that regulate the behavior. Here we examine whether the BBB expresses sex-85 specific transcripts that might be part of this communication. This would be in 86 agreement with the finding that feminization of the BBB by expression of the 87 feminizing TRA protein specifically in the BBB of adult males results in a significant 88 reduction in male courtship (8). In these experiments, the tightness of the barrier 89 was unaffected, suggesting that specific male transcripts are physiologically 90 participating in courtship control. The identity of these factors and their function is 91 unknown. Here we identify sex-preferentially expressed transcripts in the BBB of 92 males and females and demonstrate that the nuclear receptor DHR3/Hr46, best 93 known for its roles in larval development (10)(11)(12), is physiologically required in the 94 BBB of adult mature males to ensure normal male courtship behavior. 95

97
A microarray screen identifies male-and female-enriched transcripts in the BBB 98 In our previous experiments, there was a strong reduction in male courtship when 99 we conditionally feminized adult BBB cells (8). This suggests that feminization 100 disrupts male-specific transcripts that are physiologically required for normal 101 mating behavior. In order to identify these transcripts, we isolated BBB cells from 102 males and females and characterized their transcripts. The Gal4/UAS system was 103 used to mark these cells (13). We expressed the fluorescent protein DsRed in the 104 nuclei of SPG cells, using the moody-Gal4 driver that drives expression in SPG cells 105 (SPG-Gal4) (6). As seen in Fig 1B, the large nuclei of the SPG cells were specifically 106 marked. We dissected fly brains and manually removed and collected fluorescent 107 cells. Cells from approximately 50 flies were pooled for each biological replicate, 108 and the RNA of three biological replicas from separate crosses was prepared for 109 each sex. The RNA was subsequently used for microarray analysis by GenUs 110 Biosystems (http://www.genusbiosystems.com/). The results confirmed the 111 presence of sex-preferentially expressed transcripts in the BBB of males and 112 females, respectively. 284 transcripts were identified that were enriched > 2 fold in 113 either males and females (Figures 2C,D). Of those, 112 were male-preferentially 114 expressed (S1 Table). As expected, the male-specifically expressed rox RNAs that are 115 required for dosage compensation were highly specific to males. Furthermore, sex-116 specific dsx transcripts were identified because male and female transcripts use 117 different polyA-sites and can thus be identified by microarray (14). An analysis of 118 the GeneOntology of the enriched transcripts is shown in Table S2. Sex specific 119 categories such as dosage compensation and sex determination are well 120 represented, further confirming that the isolated cells are sexually determined. 121 About half of the genes fall into one of these categories. The rest of the genes could 122 not be assigned to a specific category. In addition to identifying sex-preferentially 123 expressed RNAS, the experiment also provided an inventory of RNAs present in the 124 BBB cells. The vast majority of BBB transcripts is equally expressed in males and 125 females. Among them, as predicted for SPG cells, were RNAs that are characteristic 126 of BBB cells (4, 5): RNAs for the junction proteins sinu and neurexin, for example, 127 and the previously characterized SPG transcripts for moody and Mdr65. The most 128 likely contaminating cells from the dissections would be fat body cells which are in 129 close proximity to the BBB, and neuronal cells. We found very small amounts of the 130 fat body transcript Lsp-2, or of the neuronal marker elav. They were not 131 preferentially present in males or females, indicating that low amounts of these cells 132 are unlikely to affect the identification of sex-specific transcripts in the BBB. 133

134
The nuclear receptor DH3/Hr46 is required in the BBB for courtship 135 One of the male-enriched BBB RNAs is the transcript for the nuclear receptor 136 DHR3/Hr46 (Fig 1E). DHR3/Hr46 is an orphan nuclear receptor that is most related 137 to the mammalian ROR receptor (Retinoic acid related orphan receptor). 138 DHR3/Hr46 is a well-described transcriptional regulator of larval developmental 139 processes in response to ecdysone, but no adult functions have been described so 140 far. To examine whether DHR3/Hr46 is required in the BBB for courtship we 141 conditionally expressed two different DHR3/Hr46-RNAi constructs specifically in the 142 BBB of mature adult males and examined their courtship (Fig 2). Male courtship in 143 Drosophila melanogaster consists of well-defined stereotyped behavioral steps that 144 can easily be quantified in a courtship index (CI) (15)(16)(17). The CI is calculated as the 145 fraction of time the male spends displaying any element of courtship behavior 146 (orienting, following, wing extension, licking, attempted copulation, copulation) 147 within a 10 minute observation period (18). We used the Gal4/Gal80 ts system to 148 restrict knockdown to mature males (19). Two different BBB-Gal4 drivers were 149 used to direct expression, the previously described SPG-Gal4, and a SPG-cell-specific 150 Mdr65-Gal4 driver that was generated in our lab ( Fig 2D) of the steps of courtship, but they did so with lower probability. To eliminate 161 general sickness of the males as a cause for the reduced courtship, we performed a 162 short-term activity assay (22) and found no activity defects in the knockdown flies 163 ( Fig 2C). We conclude that DHR3/Hr46 is required in the BBB of mature males for 164 normal courtship behavior. SPG BBB cells are glial cells. To confirm the glial 165 requirement for DHR3/Hr46 we used the glia-specific driver repo-Gal4 to drive UAS-166 DHR3/Hr46-RNAi in adult males and observed equally reduced courtship (p<0.001) 167 ( Fig 3A). As expected, when we expressed DHR3/Hr46-RNAi in the BBB with Mdr-168 Gal4 in the presence of a glial-expressed Gal80 blocker (repo-Gal80) (23) we 169 observed a reversal of the courtship defects. Together our findings demonstrate that 170 DHR3/Hr46 is needed in the glial SPG cells for normal courtship. 171 To assess whether DHR3/Hr46 knockdown affects the integrity of the BBB, we 172 tested the tightness of the BBB by injecting 10kD Texas-Red (TR)-marked Dextran. 173 It is well documented that in wildtype animals TR-Dextran will be kept out of the 174 brain and accumulate at the BBB, whereas a leaky BBB would allow entry into the 175 brain (6). As shown in Fig 3E,  To examine the intracellular distribution of DHR3/Hr46, we used antibodies 183 generated by the Thummel lab (11) to study the protein distribution in SPG cells of 184 mature animals (Fig 4). Indy-GFP was used as a BBB marker; it is expressed in both 185 PG and SPG cells (5 pupariation, but strong activation was observed again in late pupae. We show here 254 activation of the reporter construct in the BBB of mature adult males. In these 255 experiments, the reporter construct is conditionally induced by a heat pulse in 256 mature males. Thus, the observed activation reflects a "snapshot" of the presence of 257 the putative ligand at that time. The observed activity coincides with the time when 258 knockdown of DHR3 causes a reduction in courtship. 259 DH3/Hr46 is best known for its essential role in development as an ecdysone 260 effector. It is activated by ecdysone and is a part of an activation cascade in response 261 to ecdysone. It induces another nuclear receptor, Ftz-F1, among numerous other 262 genes. Eventually, it acts as a negative feedback regulator to turn off ecdysone-263 receptor signaling (10-12, 30, 31  The courtship assay and activity assay were performed as previously described 313 (34). In short, males were placed in a plexiglass "mating wheel" (diameter 0.8 cm), 314 together with a 2-4 hrs old Canton-S virgin female. The courtship index was 315 calculated as the fraction of time the male spent displaying any element of courtship 316 behavior (orienting, following, wing extension, licking, attempted copulation, 317 copulation) within a 10-minute observation period (18). Short-term activity assays 318 were performed as previously described (22). Individual males were placed into the 319 ''mating wheel'' containing a filter paper with a single line dividing the chamber in 320 half. After 2-3 minutes of acclimation time, the number of times the male crossed 321 the center line within the three-minute observation time was counted. 322 Each graph represents sets of control and experimental genotypes that were grown, 323 collected, aged and tested in parallel. In each behavioral session, equal numbers of 324 all genotypes were tested. 325 326

Microarray analysis 327
To isolate blood-brain barrier cells, flies bearing the SPG-Gal4 driver were crossed 328 to flies carrying the fluorescent reporter transgene, UAS-DsRed. This resulted in the 329 expression of the visible fluorescent marker DsRed to mark the nuclei of BBB cells. 330 Prior to the experiment, both the driver SPG-Gal4 and the UAS-DsRed lines were 331 outcrossed with a Cantonized w 1118 strain for 10 generation. The flies were grown in 332 a 25 o C incubator under a 12hrs light/12hrs dark cycle. Eclosing males and females 333 were collected and kept in separate groups of 10-15 flies of the same sex under the 334 same conditions for 4 days and then dissected between ZT 5 and ZT 7 to control for 335 levels of cycling transcripts. Equal numbers of males and females originating from 336 the same culture were dissected in each sitting. The brains were dissected in ice-337 cold 1 X PBS. The dissected brains were then transferred to a new petri dish 338 containing ice-cold 1X PBS within half an hour. Carefully, under the fluorescent 339 microscope, individual and/or groups of blood-brain barrier cells marked with 340 DsRed were isolated manually by using Dumont # 5 SF superfine forceps (Fine 341 Science Tools, Inc). The cells were then immediately transferred to a frozen droplet 342 of Trizol reagent on dry ice and stored in -80°C until further processing. Cells were 343 isolated from at least 50 brains for each genotype. The approximate total number of 344 cells isolated per brain varied from ~60-120. The forceps were cleaned with 345 RNAZap when moving from one genotype to the other. 346 The isolated BBB cells of male and female flies were provided to GenUs Biosystems 347 (http://www.genusbiosystems.com/) for microarray analysis. A total of 3 biological 348 replicates for males and females were submitted. Cells were lysed in TRI reagent 349 (Ambion) and Total RNA was isolated using phenol/chloroform extraction followed 350 by purification over RNeasy spin columns (Qiagen). The concentration and purity of 351   Gal4 initiates transcription at UAS-dsRed(nuclear). dsRed can be seen expressed in 478 the characteristic large nuclei of SPG cells (red). Indy-GFP expression is used as a 479 BBB marker (green). Blue: DNA staining (DAPI, blue). 480 481 S1