Inactivation of hypocretin receptor-2 signaling in dopaminergic neurons induces hyperarousal and enhanced cognition but impaired inhibitory control

Hypocretin/Orexin (HCRT/OX) and dopamine (DA) are both key effectors of salience processing, reward and stress-related behaviors and motivational states, yet their respective roles and interactions are poorly delineated. We inactivated HCRT-to-DA connectivity by genetic disruption of Hypocretin receptor-1 (Hcrtr1), Hypocretin receptor-2 (Hcrtr2), or both receptors (Hcrtr1&2) in DA neurons and analyzed the consequences on vigilance states, brain oscillations and cognitive performance in freely behaving mice. Unexpectedly, loss of Hcrtr2, but not Hcrtr1 or Hcrtr1&2, induced a dramatic increase in theta (7–11 Hz) electroencephalographic (EEG) activity in both wakefulness and rapid-eye-movement sleep (REMS). DAHcrtr2-deficient mice spent more time in an active (or theta activity-enriched) substate of wakefulness, and exhibited prolonged REMS. Additionally, both wake and REMS displayed enhanced theta-gamma phase-amplitude coupling. The baseline waking EEG of DAHcrtr2-deficient mice exhibited diminished infra-theta, but increased theta power, two hallmarks of EEG hyperarousal, that were however uncoupled from locomotor activity. Upon exposure to novel, either rewarding or stress-inducing environments, DAHcrtr2-deficient mice featured more pronounced waking theta and fast-gamma (52–80 Hz) EEG activity surges compared to littermate controls, further suggesting increased alertness. Cognitive performance was evaluated in an operant conditioning paradigm, which revealed that DAHcrtr2-ablated mice manifest faster task acquisition and higher choice accuracy under increasingly demanding task contingencies. However, the mice concurrently displayed maladaptive patterns of reward-seeking, with behavioral indices of enhanced impulsivity and compulsivity. None of the EEG changes observed in DAHcrtr2-deficient mice were seen in DAHcrtr1-ablated mice, which tended to show opposite EEG phenotypes. Our findings establish a clear genetically-defined link between monosynaptic HCRT-to-DA neurotransmission and theta oscillations, with a differential and novel role of HCRTR2 in theta-gamma cross-frequency coupling, attentional processes, and executive functions, relevant to disorders including narcolepsy, attention-deficit/hyperactivity disorder, and Parkinson’s disease.

(a) Schematic representation of homologous recombination between the Hcrtr1 genomic locus and the targeting vector.Two loxP sites (red triangles) were inserted to flank the first coding exon (exon 3) and exon 4, which together encode the N-terminal first 126 aa of HCRTR1.The neomycin resistance gene used for selection in embryonic stem cells (neo; flanked by two FRT sites shown as orange triangles) was deleted using FLP recombinase, creating the Hcrtr1 flox allele.In Dat-IRES-Creexpressing cells, CRE excises the inter-loxP segment (1.1 kb), creating the Hcrtr1 del-Gfp allele.The ATG translation start codon of Hcrtr1 is replaced by the ATG of Gfp, and the Hcrtr1 gene promoter now drives expression of Gfp (green rectangle), instead of Hcrtr1.Gfp reading frame is followed by a polyadenylation site (pA) to terminate transcription and prevent downstream expression.Boxes  (a) Putative HA neurons were characterized by testing the electrophysiological response to consecutive 25 pA current steps, lasting 500 ms, ranging from -100 to +325 pA.Typical responses to -100 pA and -100 pA stimulations show, respectively, the classical Ih-dependent sag on hyperpolarization, and the low-frequency late firing.(b) Relation between injected current and firing frequency for putative histaminergic neurons from C57BL/6J, Hcrtr2 flox/flox and Hcrtr2 del/del mice, respectively).Data points are average firing frequencies calculated from a coherent ensemble of neurons and plotted as a function of injected current.Passive membrane properties are coherent with those reported by Haas and Reiner 1 (STable 1).(c) Post-recording morphological reconstruction of a biocytin-loaded TMN neuron from a Hcrtr2 flox/flox mouse, confirming the large diameter (20-30 µm) and multipolar shape of TMN HA neurons.Left: Maximal intensity projection of a 40x Z-stack of a 300 μm slice processed for IHF against biocytin and DAPI.Right: Projection of the 3D reconstruction.Scale bar 50 μm.(d-g) AP features of C57BL/6J and Hcrtr2 flox/flox neurons are also in line with prior reports, but Hcrtr2 del/del cells displayed a reduced mean and more variable AP amplitude (d, P=0.011), an increased mean and more variable AP half-width (e, P=0.013), a reduced AP rise slope (f, P=0.010) and an increased AP decay slope (g, P=0.043).These slightly slower AP dynamics suggest that maturation of TMN HA cells in Hcrtr2 del/del mice may be partly compromised, pointing to a potential role of HCRTR2 in the developmental specification of the HA neuronal phenotype.In fact, AP amplitude and duration are known to respectively increase and decrease during differentiation of cell types such as Cajal-Retzius cortical layer I neurons 2 and CA1 pyramidal neurons 3,4 .This hypothesis warrants further investigation, to decipher the potential role of HCRTR2 in neuronal maturation.(a) Schematic of experimental design.Mice were provided with a preferred nesting material (a Nestlet TM , see Methods) at dark-onset (ZT12) of Day 6 of a 9-day experimental timeline.On Day 9, mice were removed from the nest they had built and transferred to a fresh cage at ZT3, i.e., during an otherwise major sleeping period.The mice movements were monitored by infrared beam breaks, and locomotor activity (counts/h) is plotted across time during baseline (0-24 h, average of 2 baseline days), 6-h sleep deprivation (24-30h), and recovery (30-48h).The pronounced increase in theta dominated wakefulness of DA OxR2-KO mice is not coupled to increased locomotor activity.DA OxR1-KO mice showed higher activity during one timepoint in SD, but not during baseline or recovery.DA OxR1&2-KO showed higher activity during the first half of baseline dark phase, but not during SD or recovery (two-way ANOVA; DA OxR1-KO : SD: genotype effect F(1,23)=6.792,P=0.009; DA OxR1&2-KO : baseline: genotype effect F(1,23)=38.381, P<0.001, genotypeXtime interaction: F(23,524)=2.646,P<0.001; with Bonferroni post-hoc test, *P<0.05).n=9 mice per group, except n=10 for DA OxR1&2-CT .Hcrtr2 exon 1 and intron 1 were subcloned and used to construct the targeting vector.The targeting vector was designed to insert a 5'loxP site 70 bp downstream of the transcription start site (TSS)

Genotype
and 37 bp upstream of the ATG initiation codon within Exon 1.The 3'loxP site is inserted ~1 kb within Intron 1 followed by a promoter-less GFP coding sequence and a rabbit polyadenylation signal.This is designed so that, after Cre/loxP-recombination, the Hcrtr2 gene promoter drives expression of Gfp instead of Hcrtr2, thus creating an allele acting both as KO and GFP-reporter.The targeting vector was linearized and electroporated into IC1 cells (C57BL/6NTac embryonic stem (ES) cell line, ingenious targeting laboratory, Ronkonkoma, NY, USA).Single colonies were screened by Southern blotting using probes 5′ and 3′ external to the targeting vector and a Gfp internal probe.Two clones demonstrating correct recombination events were injected into BALB/cAnNHsd blastocysts.The FRT-flanked neo cassette was excised by mating the resulting chimeras with Tg(ACT-FLPe) 9205Dym mice.The resulting neo-excised conditional KO allele, Hcrtr2 tm1.1Ava (MGI-ID:5637402), is referred to as Hcrtr2 flox .
Immunohistofluorescence and confocal microcopy.Mice (~16-week-old of either sex) were deeply anesthetized with sodium pentobarbital (100 mg/kg, i.p.), and transcardially perfused with 4% paraformaldehyde (pH 7.4).Brains were quickly removed, post-fixed in the same fixative for 2 h at 4°C and immersed successively in 15% (1 h) and 30% sucrose (o/n) at 4°C, frozen and stored at −80°C until sectioning.For each mouse, 48X20 um-thick coronal sections from the midbrain (bregma Surgery and EEG recordings.Surgical implantation and EEG recordings were as described 5,6 . Briefly, after implantation of EEG/EMG electrodes, the mice were allowed one week for recovery and one week for habituation to the recording setup.EEG/EMG signals were acquired using EMBLA TM system at 200 Hz sampling rate and scored using the Somnologica-3 TM software (Medcare).All EEG data were acquired in 10-15-week-old males (27-31 g), housed individually after electrode implantation.Locomotor activity was monitored using infrared sensors and ClockLab software.
Animals were exposed sequentially to four behavioral contexts (Baseline-SD-Nest-Cage change 7 . After two baseline recording days, a 6-h sleep deprivation (SD) was initiated at light-onset (ZT0) on day 3.Following SD, mice were left undisturbed for 2 days.At dark-onset (ZT12) of day 6, a square of packed shreddable cotton (Nestlet TM , Ancare, Bellmore, NY, CatNr.14010) was introduced into the cage.Mice were left undisturbed until next light-onset with assessment of nest morphology.Mice were left undisturbed for another 2 days.On ZT3 of day 9, mice were transferred from their home cage where the nest had been built to a fresh cage.Latency to NREMS-onset was defined as the time until the first ≥2 min-long NREMS episode.Animals showing epileptic-like seizures or preepileptic-like spikes were excluded from analysis (~10% of mice).
Vigilance state analysis.We scored vigilance states manually, blind to the experimental conditions, in 4-s epochs by concurrent evaluation of EEG and EMG signals in the Somnologica-3 TM software (Medcare) using established criteria 6 .Scripts were developed to quantify wakefulness, NREMS, REMS episode number, duration and vigilance state fragmentation as previously described 5,6,8 .
TDW analysis was performed as described in Vassalli and Franken 6 .REMS latency was calculated as time from sleep-onset (first NREMS episode ≥12 s) to the first REMS episode following SD.

EEG power spectral density and time-course analysis. EEG signals were subjected to discrete
Fourier transform with non-overlapping 4-s hamming windows to determine power spectral densities (PSD) across 0.25-80 Hz with 0.25 Hz frequency-bin resolution 5,6 .For each vigilance state and time interval, PSDs of all artifact-free, same-state-flanked 4-s epochs were averaged to generate mean PSDs for each animal.To account for differences among animals in absolute EEG power, power density in each frequency bin and for each state was expressed as percentage of a baseline reference value, calculated for each mouse across 2 baseline days by summation of the power across 0.75-40 Hz frequency bins in all 3 behavioral states.This reference value was weighted so that for each animal each state contributed equally to the total EEG power 9  Detection of phasic REMS events.We detected phasic REMS events, which are transient increase in theta power and frequency during REMS, as described previously 10,11 .Briefly, we first bandpassfiltered EEG signal between 4 and 12 Hz using finite impulse response filters with an order equal to three cycles of the low cutoff frequency, and detected the individual theta peaks from the filtered signal.We then smoothed the interpeak interval time-series using an 11-sample moving average window, and selected the smoothed interpeak intervals shorter than the 10 th percentile as candidate phasic REMS events.The candidate events with the following criteria were considered as phasic REMS: (1) minimum event duration of 900 ms; (2) minimum smoothed interpeak interval shorter than 5 th percentile; (3) mean amplitude of theta peaks larger than mean amplitude of theta peaks across all REMS.
Theta-gamma cross-frequency coupling.We used the modulation index (MI) to measure thetagamma phase-amplitude coupling 12,13 .Using finite impulse response filters with an order equal to three cycles of the low-cutoff frequency, we bandpass-filtered EEG signals into theta (7-11 Hz) and fast-gamma (52-80 Hz) in both forward and reverse directions to eliminate phase distortion.We then estimated instantaneous phase of theta and the envelope of fast-gamma using the Hilbert transform.Albans, VT, USA), each enclosed in a wooden cubicle.Chambers featured an exhaust fan for ventilation, serving as white noise emitter for sound attenuation, a steel grid floor, three 20-mm horizontally-spaced nosepoking holes on one wall, and a reward receptacle with a liquid dipper and a 0.01-ml-cup on the opposite wall.Lights could be illuminated inside each nosepoking port, the reward receptacle, and on the chamber ceiling.Mice were trained to self-administer a 0.2% saccharine liquid reward (Sigma-Aldrich).A nosepoke in the "active" port illuminated the aperture and activated delivery of a 0.01 ml reward, which remained available for 3 s once head entry in the liquid dipper was detected.Supplementary entries into the active port in the absence of head entry above the liquid dipper and entries in inactive ports were recorded but had no consequence.preceded illumination of one of the ports, but the cue light stayed illuminated for only 2 s.Any response before cue light illumination is a premature response, without programmatic consequences.Omissions were recorded when no nosepoking occurred during the allowed 5 s after beginning of illumination of one the apertures.To advance to the "Test phase", the mice had to earn ≥30 rewards during the 30-min session.Mice underwent 2 sessions per day.Usually one day, in some cases two days, were necessary to reach next-stage criteria.The Test phase was as "Stage 4" except that premature responses during the 5-s delay were followed by a 5-s time-out period with illumination of the ceiling light.A perseverative response was defined as repetitive nosepoke in the liquid dispenser after reward consumption and served as measure of compulsive behavior.
Premature, correct, incorrect, omission, and perseverative responses were recorded.Sessions of the 3 last days in "Test phase" were averaged for each mouse to establish baseline performance.
Attention and motivation probe.The conditions in the attention probe were the same as during the "Test phase", except that the cue light duration was 3 s (easier than in the former test phase), then 2 s (as in test phase) and finally 1 s, making the task contingencies progressively more difficult.
Nosepoking before cue light illumination had no consequence, hence no premature response was recorded.Only a correct response gives access to the reward, nosepoking in the 2 unlit apertures triggers an incorrect response and no reward.During the motivation probe sessions, the mice' motivation to poke for rewards was tested without the attentional constraints above.One and the same hole was active throughout the test.Mice were first tested under a fixed ratio of 1, where one nosepoke in the active port provides one reward.The next day mice were submitted to a progressiveratio schedule of reinforcement, which consisted of a systematic within-session increase in number of responses required to earn one reward.The number of active nosepokes required increased from reward to the next according to the progression sequence: response ratio (rounded to nearest integer) = (5e rewardx0.2)-5 14 .Hence, the progressive-ratio schedule followed: In some experiments, 0.15% biocytin was added for post-recording staining and morphological

Figure S1 :
Figure S1: Generation of mice with dopamine neuron-specific Hcrtr1 gene disruption.

Figure S2 :
Figure S2: CRE immunostaining demonstrates the efficiency and specificity of the Dat-IRES-Cre allele to target dopamine neurons.

Figure S4 :
Figure S4: Vigilance states quantification in three HCRT-to-DA mutant mouse models.

Figure S5 :
Figure S5: Quantification of NREM and REM sleep recovery following sleep deprivation in dopaminergic Hcrtr2-ablated mice.

Figure S6 :
Figure S6: Upregulated EEG beta band activity in wakefulness of dopaminergic Hcrtr1&2 doubly-ablated mice during sleep deprivation.

Figure S7 :
Figure S7: Dopaminergic Hcrtr2-ablated mice exhibit profound increases in waking theta and fast-gamma activity in both rewarding and stressful environments.
(b,c) Time-frequency heatmaps show the spectral dynamics of wakefulness from (b) the time when nesting material was added at dark-onset (ZT12) of Day-6 until the next light-onset (ZT24), and (c) from nest removal/cage transfer until the next sleep-onset.In the left heatmaps of both b and c, colors encode the average EEG power in each 0.25-Hz frequency bin, and across each time bin, expressed relative to their mean values in the last 4 h (ZT8-12) of baseline (BSL) light phase.Right heatmaps represent the differential dynamics of the waking EEG of KO and CT mice, where the power values of KO mice are subtracted from their controls (KO-CT).(d) EEG PSD profile of wakefulness from nest removal and cage transfer to the next sleep-onset.Insets show magnification of spectra across 0.75-15 Hz.Red lines indicate significant differences.DA OxR2-KO mice exhibit higher theta power across 7.75-9.75Hz, while DA OxR1&2-KO mice show decreased power across 8.5-10 Hz, relative to controls (two-way ANOVA; DA OxR2-KO : genotypeXfrequency interaction F(280,3653)=1.243,P=0.005; DA OxR1&2-KO : genotype effect F(1,286)=6.517, P=0.011; with Tukey post-hoc test, P<0.05).DA OxR1 (n=7:9 KO:CT), DA OxR2 (n=8:7 KO:CT), DA OxR1&2 (n=7:8 KO:CT).
. To analyze how specific spectral components of the waking EEG evolve across time, EEG power density within delta (1-4 Hz), inter-delta/theta (4-7 Hz), theta(7-11 Hz), beta (15-30 Hz) and fast-gamma (52-80 Hz) bands were first normalized by the mean power density within that frequency range during wakefulness of the last 4 h of the two baseline light phase (ZT8-12; time of minimal sleep pressure), then averaged across each time interval, and each individual mouse of each genotype, and plotted across time.The number of time intervals were adjusted according to the prevalence of wakefulness, i.e., 6 during baseline light phase, 12 in dark phase, 8 during the 6-h SD, and 4 in recovery light phase.To generate time-frequency heatmaps of EEG power during wakefulness, power density in each 0.25 Hz frequency bin was as above first expressed relative to its average value during wakefulness of baseline light phase ZT8-12, and then averaged across each time interval, and individual mouse of each genotype.Time intervals are the same as for time-course analyses of the power in each frequency band, as indicated above.
Theta phase was discretized into 18 equal bins (N=18, each 20°) and the average value of fastgamma envelope within each bin was calculated.The resulting phase-amplitude histogram (P) was compared with a uniform distribution (U) using the Kullback-Leibler distance,   (, ) =∑ () * [()/()]  =1, and normalized by log(N) to obtain the modulation index, MI = DKL / log(N).For each animal, we filtered continuous 48 h baseline recordings into the theta and fastgamma bands and estimated the Hilbert transform, and then concatenated episodes of each state to calculate the MI during dark and light phases.To explore possible coupling patterns between different pairs of low and high-frequency bands, we used comodulogram analysis12 .We considered 16 frequency bands for phase (1-18 Hz, 1-Hz increments, 2-Hz bandwidth), and 14 frequency bands for amplitude (15-90 Hz, 5-Hz increments, 10-Hz bandwidth).MI values were then calculated for all these pairs to obtain the comodulogram graph.Three-choice serial reaction time task (3-CSRTT).Adult males and control littermates (5-7 months-old) were group-housed (3-5 animals/cage) under a 12-h reversed light/dark cycle (lightonset at 8:30 p.m.) at a constant temperature (22°C) and had ad libitum access to water.The 3-CSRTT was conducted in mouse operant chambers (15cmx15cmx13.5 cm, Med Associates, St

The 3 -
CSRTT comprised four training stages and a test phase: Stage 1: Mice learned to perform a head entry into the reward magazine to receive 0.01 ml saccharine.Mice were considered to have reached training criteria once they reached 50 rewards in 30 min.Stage 2: When the mouse head entered the reward magazine, a cue light stimulus appeared randomly in one of the three nosepoking ports and stayed lit until a nosepoke in this hole occurred, upon which the light turned off, the dipper was activated and the reward made available (even if the mouse additionally nosepoked in any of the two other unlit apertures).Three seconds after the reward was made available, the dipper deactivated, and a light cue appeared once again randomly in one of the three ports.This went on until the mouse earned 40 rewards during a 30-min training.Stage 3: similar conditions as "Stage 2", except that a 5-s delay separated beginning of a trial by magazine head entry and illumination of one of the ports.A correct response is a nosepoke in the illuminated port, which triggers a reward, and an incorrect response is a nosepoke in any other unlit port, which triggers illumination of the ceiling light for 5 s and no reward.After each correct or incorrect response, a new trial begins with head entry into the reward magazine.This went on until mice earned 50 rewards during the 30-min training.Stage 4: after trial start by nosepoking in the reward magazine, a 5-s-delay, like in Stage 3,

1 .
reconstruction.Traces were low-pass Bessel filtered at 2 kHz and digitized at 10 kHz, with pClamp9/Digidata 1322A (Molecular Devices).The resting membrane potential (Vrest) was measured in open circuit mode, soon after obtaining the whole-cell configuration.No correction was applied for liquid junction potentials.Orexin B (OXB), [Ala11,D-Leu15]-Orexin B (OXB-Ala,Leu), and (2S)-1-(3,4-dihydro-6,7-dimethoxy-2(1H)-isoquinolinyl)-3,3-dimethyl-2-[(4-pyridinylmethyl)amino]-1-butanone hydrochloride (TCS-OX2-29) were from Tocris Bioscience (Bristol, UK), and dissolved in distilled water-based stock solutions, aliquoted and stored at -20°C until usage.Analysis of patch-clamp data.Passive membrane properties, action potential (AP) features and firing responses to agonists were analyzed off-line using Clampfit 9.2 (Molecular Devices) and OriginPro 2019 (OriginLab Corporation, Northampton, MA, USA).All AP features were analyzed during the first step of current injection able to trigger multiple APs; individual AP parameters weremeasured on the first AP of the induced train.Spike width was calculated at half-amplitude, spike amplitude was computed as the difference between AP threshold and peak.Adaptation was measured as the ratio between the fourth and the first spike interval.Spike intervals were measured between consecutive peaks.After-hyperpolarization was the difference between the AP threshold and the most negative membrane potential (V m) reached on repolarization.The triggering depolarization slope was the difference between the most negative Vm reached on repolarization and the following AP threshold, divided by the relative time.The time constants of membrane charging and discharging were estimated by mono-exponential fit to the passive response to -100 pA injection.For the Ih-related sag, τ was derived from mono-exponential fit to the current decay from the negative peak to the end of the 0.5 s current pulse (-50 pA).Drugs were perfused in the bath and their effects on cell firing were measured for 2 min after reaching the maximal effect, which in the case of HCRTR2 agonists usually occurred ~30 s after the administration ended.Only one neuron was sampled in each slice, to avoid uncontrolled long-term effects of neuromodulators (e.g. on receptor desensitization).Putative histaminergic neurons were identified based on their typical electrophysiological properties, such as slow depolarization induced by step current injection, ~2 Hz spontaneous firing, late-spiking profile, Ih-related sag upon hyperpolarizing pulses, Vrest ~-50 mV and cell capacitance ~20 pF.The passive membrane properties, including τ's of membrane charging, discharging and Ih-related sag, as well as the main AP features were consistent across genotypes and with those previously reported Statistics.Animals of all genotypes were randomly distributed for sleep recording or behavioral sessions, and investigators involved in sleep scoring or data acquisition were blinded as to the animals' genotype.Statistical analyses were performed using SigmaPlot12.0or GraphPad prism8.4.2.Data normality and variance homogeneity were respectively verified using the Shapiro-Wilk and the F test as well as by Q-Q plot inspection.Statistical significance of comparisons was determined using independent t-test, or two-way ANOVA, with P, F, t and df values reported in Figure legends.Significant ANOVA analyses were followed by Bonferroni or Tukey multiple comparison post-hoc tests.Values in the text are reported as mean ± standard error mean (SEM) unless reported otherwise.The number of experiments refers to the number of animals, or neurons of different brain slices.Illustration software.Fig. 1b, Fig. 2a, Fig. S1b and Fig.S7a were created with the help of BioRender.com.All Figures were prepared using Adobe Illustrator CC (Adobe).

Table S3 : Comparison of waking theta peak frequency (TPF, Hz) between conditional knockout mice and their respective controls in different behavioral contexts
. TDW, Theta-Dominated Wakefulness.CC, Cage Change.n=9 mice per group, except n=10 for DA OxR1&2-CT .

DA OxR1-KO vs DA OxR1-CT DA OxR2-KO vs DA OxR2-CT DA OxR1&2-KO vs DA OxR1&2-CT
Mice for analysis were produced by intercrossing Hcrtr1 or 2 flox/flox mice, with one parent heterozygous for Dat IRES-Cre (Slc6a3 tm1.1(cre)Bkmn ).This cross generates two offspring groups: Hcrtr1 or 2 flox/flox , Dat +/IRES-Cre (KO), and Hcrtr1 or 2 flox/flox , Dat +/+ (CT) mice.Note that each line has its own genetic control group.Mice are in a mixed C57BL/6NTacXC57BL/6J background.Dat +/IRES-Cre mice show Cre-mediated reporter expression at E17, suggesting Hcrtr1 and Hcrtr2 gene inactivation occurs in late gestation in our mice.To rule out effects of Dat-IRES-Cre, we recorded mice segregating solely the Dat-IRES-Cre allele (Dat +/IRES-Cre and Dat +/+ ).None of the phenotypes we describe in DA OxR1 , DA OxR2 or DA OxR1&2 -ablated mice were observed.In contrast, theta power of Dat +/IRES-Cre tended to be lower in wakefulness and REMS compared to Dat +/+ mice.Mice were housed with food and water ad libitum under 12:12 light:dark cycle (lights-on at 08:00 AM, Zeitgeber Time ZT0).Animals were randomly distributed for sleep recording sessions and investigators performing SD and sleep scoring were blind to the animals' genotypes.Hcrtr2 gene targeting.Two contiguous NcoI fragments (8.1 kb and 2.1 kb) from PAC clone RP23-392M3 of a C57BL/6J genomic library (BACPAC Resources Center, Oakland, CA, USA) containing