Circadian and photic regulation of clock and clock-controlled proteins

Circadian and photic regulation of clockand clock-controlled proteins in the suprachiasmatic nuclei of calorie-restricted miceJorge Mendoza,Paul Pe´vet and Etienne ChalletInstitut de Neurosciences Cellulaires et Inte´gratives,De´partement de Neurobiologie des Rythmes,UMR7168⁄LC2,CNRS et Universite´Louis Pasteur,5rue Blaise Pascal,67084Strasbourg cedex,FranceKeywords:calorie restriction,circadian rhythms,clock proteins,photic resettingAbstractIn mammals,behavioural and physiological rhythms as well as clock gene expression in the central suprachiasmatic clock(SCN)are phase-shifted by a timed calorie restriction(T-CR;animals receiving at midday66%of their daily food intake).The molecular mechanism of SCN depends on feedback loops involving clock genes and their protein products.To understand how T-CR mediates its synchronizing effects,we examined the rhythmic expression of three clock proteins,PERIOD(PER)1,2and CLOCK,and one clock-controlled protein(i.e.vasopressin;AVP)in the SCN of mice either fed ad libitum(AL)or with T-CR.Moreover,we evaluated expression of these proteins in the SCN of AL and T-CR mice following a1-h light pulse.The results indicate that,while PER1and AVP rhythms were phase-advanced in T-CR mice,the PER2rhythm showed an increased amplitude.CLOCK was expressed constitutively in AL mice while in T-CR it was significantly reduced,especially after feeding time.A light pulse produced a delayed increase in PER1and a larger increase in PER2expression in the SCN of T-CR mice than in AL animals.In addition,light exposure triggered an increase in AVP-ir cells in both AL and T-CR mice,and also of CLOCK expression but in T-CR mice only.The circadian changes in clock and clock-controlled proteins and their acute responses to light in the SCN of T-CR mice demonstrate that metabolic cues induced by a calorie restriction modulate the translational regulation of the SCN clock.IntroductionIn mammals,the circadian timing system is hierarchical,with the main clock located in the suprachiasmatic nuclei(SCN)of the hypothalamus and subordinate oscillators in some cerebral structures and peripheral tissues(Hastings et al.,2003).The clockwork in the SCN operates in a self-sustained fashion,involving interacting positive and negative transcriptional and post-translational feedback loops(Reppert& Weaver,2001).In the positive loop,the transcription factors CLOCK and BMAL1drive rhythmic transcription of Period(Per1–3)and Cryptochrome genes(Cry1–2;Lowrey&Takahashi,2004).PERs and CRYs form heterodimers that translocate to the nucleus,where CRYs act as negative regulators by interacting with CLOCK and⁄or BMAL1, forming a negative feedback loop(Kume et al.,1999;Shearman et al., 2000).The nuclear orphan receptor genes Rev-erb a and Ror a are activated by the CLOCK:BMAL1heterodimer to produce proteins that have opposing actions on Bmal1transcription(Preitner et al.,2002; Sato et al.,2004).Transcription of clock-controlled genes such as arginine vasopressin(Avp)is also regulated by CLOCK:BMAL1(Jin et al.,1999).Light is the most potent cue for the entrainment of the SCN(Daan&Aschoff,2001).Retinal ganglion cells project directly to the SCN via the retinohypothalamic tract(Hannibal,2002).Light resetting of the mammalian SCN clock involves acute induction of immediate–early genes(Kornhauser et al.,1990)and Per genes (Albrecht et al.,1997;Shigeyoshi et al.,1997).On the other hand, temporally restricted feeding(i.e.limited duration of daily food access) is a strong synchronizer for rhythmic gene expression in peripheral oscillators(Damiola et al.,2000;Stokkan et al.,2001).Under constant darkness conditions(but not in a light–dark cycle),temporally restricted feeding or a palatable diet in addition to regular food ad libitum are able to entrain the SCN clock(Mistlberger,1994;Holmes& Mistlberger,2000;Castillo et al.,2004;Caldelas et al.,2005;Lamont et al.,2005;Mendoza et al.,2005a).In contrast to temporally restricted feeding paradigms,which do not reduce calorie intake and which allow enough food to be eaten daily,a calorie restriction is known to be responsible for many beneficial physiological changes(slowing of ageing,extended lifespan,delayed onset of major age-related diseases) in various species(Bordone&Guarente,2005).In rats and mice under a light–dark cycle,a timed calorie restriction(i.e.66%of the animal’s daily food intake given each day at the same time),but not temporally restricted feeding,leads to phase advances of behavioural and physiological circadian rhythms,whatever the feeding time over the daily cycle(Challet et al.,1997,1998).Moreover,the temporal organization of the SCN clockwork and its behavioural and clock-gene responses to light are affected by calorie restriction(Mendoza et al., 2005b),suggesting that these effects are related to calorie reduction. Clock mechanisms depend on the regulation not only of clock gene mRNA but also of their protein products.In order to characterize possible SCN translational modifications induced by a timed calorie restriction,we thus investigated the rhythmic expression of clock proteins and their photic responses in the SCN of mice.Correspondence:Dr Jorge Mendoza,as above.E-mail:jmendoza@neurochem.u-strasbg.frReceived1March2007,revised12April2007,accepted10May2007European Journal of Neuroscience,Vol.25,pp.3691–3701,2007doi:10.1111/j.1460-9568.2007.05626.x ªThe Authors(2007).Journal CompilationªFederation of European Neuroscience Societies and Blackwell Publishing LtdMaterials and methodsAnimals and housingAdult male C3H mice(Charles River,Lyon,France)werefirst housed in groups offive with food and water available ad libitum.After 15days of habituation,mice were then housed singly in cages equipped with a running wheel(10cm diameter)in light-proof, ventilated chambers(23±1°C)under a12–12h light–dark cycle (LD;light,200lux;darkness,0lux).Under LD conditions,zeitgeber time(ZT)was defined relative to lights-on(ZT-0)and lights-off (ZT-12).All experiments were performed in accordance with the Principles of Laboratory Animal Care(National Institutes of Health publication86–23,revised1985)and the French national laws. Experimental designIn thefirst experiment,after3weeks of basal conditions,mice were randomly divided into two groups:a group of mice fed ad libitum (AL)with a mean spontaneous food intake of4.6±0.04g and a timed calorie-restricted group(T-CR)which was given66%(i.e.3.0±0.02g)of the daily food intake at ZT-6.After3weeks of calorie restriction,T-CR mice as well as the AL control mice were killed and brains were removed for immunohistochemistry.Sampling was performed at six temporal points with4-h intervals(ZT-0,-4,-8, -12,-16and-20)over a24-h cycle(n¼4per group and temporal point).Because sampling started at ZT-0,animals killed at ZT-0and ZT-4could not be fed at ZT-6on the day of killing in contrast to animals killed at other ZT points.In the second experiment,for the analysis of clock proteins during light resetting we used other groups of mice under the same light(LD 12:12)and feeding conditions(AL vs.T-CR)as in the previous experiment.At the end of3weeks under each feeding condition, animals were exposed to a1-h light pulse(200lux)at ZT-16and were subsequently killed in darkness at two different times(i.e.4or8h after the onset of the light pulse)at ZT-20and ZT-0(n¼6T-CR and n¼4AL for each time point).According to the molecular phase response curve to light in mice under T-CR(Mendoza et al.,2005b), ZT-16is close to the phase when light leads to the most marked change(reduced photic induction)of Per1mRNA levels in the SCN, and the only time point studied when Per2mRNA levels were affected (increased photic induction)by T-CR compared to AL conditions. According to the behavioural phase response curve to light,ZT-16is close to the time when light led to similarly large phase delays in the locomotor activity rhythm in both AL and T-CR mice(Mendoza et al., 2005b).This similar resetting response contrasts with the altered light-induced phase shifts at other nocturnal ZT points in T-CR,when Per2 expression is not triggered by light.The lack of differences between AL and T-CR mice in behavioural phase delays induced by light at ZT-15(Mendoza et al.,2005b)may be specifically due to the overexpression of Per2by light(as a compensation for reduced photic induction of Per1?).Therefore,although not evaluated,we did not necessarily expect alteration in behavioural phase shifts after light exposure at ZT-16between AL and T-CR mice. ImmunohistochemistryAt the selected intervals,animals were killed with an isoflurane overdose and perfused transcardially with50mL of0.9%saline followed by50mL of cold4%paraformaldehyde(PAF)in0.1m phosphate buffer(pH7.4).Brains were removed,postfixed overnight in4%PAF at4°C and transferred to a cryoprotectant buffered sucrose solution(30%)for72h at4°C.Brains were then frozen in isopentane at)60°C and stored at)80°C.Four series of30-l m coronal cryosections through the SCN were prepared on a cryostat at)19°C and collected in0.1m phosphate-buffered saline.Free-floating sections were washed in cold50-m m Tris-buffered saline(TBS; pH7.4;Sigma,St Louis,USA)and incubated in a solution of3%of H2O2(Sigma)in TBS for30min at room temperature.Brain sections were then rinsed in TBS,and incubated for2h in a blocking solution in3%normal horse or goat serum(NHS or NGS)and5%of bovine serum albumin in TBS with0.1%Triton X-100(0.1%TBS-X), followed by an incubation in the primary antibody(in0.1%TBS-X plus NHS or NGS)for48h at4°C.We used a goat polyclonal anti-PER1antibody(1:2000,raised against an epitope mapping the N-terminus of human Per1,SC-7724;Santa Cruz Biotechnology, Santa Cruz,CA,USA),a rabbit polyclonal anti-PER2(1:1000, affinity-purified raised against an epitope mapping the C-terminus of mouse Per2;Alpha Diagnostic International,San Antonio,TX,USA), a goat polyclonal anti-CLOCK(1:1000,raised against an epitope mapping the N-terminus of clock of mouse origin,SC-6927;Santa Cruz Biotechnology),and a rabbit anti-arginine vasopressin(A VP; Truus,1:20000,a gift from Dr Ruud Buijs,Netherlands Institute for Brain Research,Amsterdam,the Netherlands).Following incubation in the primary antibody,sections were rinsed in TBS-X and incubated for2h at4°C with a biotinylated antirabbit IgG made in goat,or antigoat IgG made in horse(Vector Laboratories,Burlingame,CA, USA),diluted1:500with0.4%TBS-X.Following incubation with secondary antibody,sections were rinsed in TBS-X and incubated for 1h at room temperature with an avidin–biotin–peroxidase complex (Vectastain Standard Elite ABC Kit;Vector).Following incubation with the ABC reagents,sections were rinsed twice for10min in TBS, and incubated with0.025%3,3¢-diaminobenzidine(Sigma)with 0.01%H2O2in50-m m Tris buffer.After thisfinal incubation,sections were rinsed with TBS,wet-mounted onto gel-coated slides,dehydra-ted through a series of alcohols,soaked in xylene,and coverslipped. The specificity of antibodies was established in mouse brain by preadsorption control experiments.Antibody binding to antigen was blocked by adding the PER1or CLOCK peptides(PER1,sc-7724-P; CLOCK,sc-6927-P;1mg⁄mL,both diluted1:100;Santa Cruz Biotechnology),the PER2peptide(Mouse PER2blocking peptide, Alpha Diagnostic International;1mg⁄mL,diluted1:100)and the vasopressin peptide(H-1780,1mg⁄mL,diluted1:10;Bachem, Bubendorf,Switzerland)to the primary incubation solution.Addition of each peptide prevented respective immunostaining in the SCN. Cell and protein quantificationSections were viewed using a monitoring CCD video camera with Viewfinder Lite software coupled to a microscope(Leica).For the image analysis we used NIH ImageJ software.Positive cells(cells were considered positive when the signal was>3·the background value)were counted in the whole central SCN(two or three sections per animal),based upon the mouse brain atlas(from bregma )0.46mm;Paxinos&Franklin,2001).To estimate the number of stained nuclei,a standard-sized oval encompassing the whole medial SCN was placed over one SCN,and the number of cells within the oval was obtained.The number of immunoreactive(ir)nuclei was expressed as the mean number of immunopositive cells per section.In addition,because not only A VP-positive cells but also A VPfibres are present in the SCN,and these may affect cell counting,in a second analysis expression of A VP was quantified by measuring staining density.The optical density(OD)of A VP-ir cells in the SCN was3692J.Mendoza et al.ªThe Authors(2007).Journal CompilationªFederation of European Neuroscience Societies and Blackwell Publishing LtdEuropean Journal of Neuroscience,25,3691–3701measured in the same sections as for the cell number analysis.The OD was expressed in arbitrary units corresponding to grey levels.To calculate the OD,the background intensity of staining was subtracted from the intensity of staining in the middle SCN.The background intensity was measured in an area devoid of A VP cell bodies or fibres (lateral hypothalamic area)in the same coronal section as the SCN analysis.Statistical analysisThe data from all experiments were evaluated using one-,two-or three-way anova followed by an LSD post hoc test,corrected for multiple comparisons.A value of P <0.05was considered statistically significant.Statistical analysis was performed using the statistical package Statistica version 4.5(StatSoft,1993).ResultsBody weight was reduced significantly ( 20%)in T-CR mice during the first week of food restriction and remained stable up to the end of the experiment (Fig.1).Repeated-measures anova showed significant differences between groups and time (Group,F 1,47¼45.2,P <0.001;Time,F 5,235¼70.5,P <0.001).Such a body weight loss is fully reversible after 1week of food ad libitum (Challet et al .,1998;Mendoza et al .,2005b).Circadian and clock-controlled protein expression in calorie-restricted micePER1expression was examined through the SCN in animals under AL and T-CR conditions.One-way anova showed significant diurnal expression of PER1expression in both AL and T-CR mice (Table 1).Although PER1was not significantly different between groups (Group effect,F 1,37¼1.38,NS),there was a trend for a reduced expression of PER1in T-CR animals compared to AL animals (Fig.2).Furthermore,there was a significant difference for the factor time and for thegroup ·time interaction (Time effect,F 5,37¼274.1,P <0.01;interaction,F 5,37¼13.8,P <0.01).PER1acrophase that occurred at ZT-12in AL mice appeared to be phase-advanced to ZT-8in T-CR animals in keeping with changes of Per mRNA patterns that we previously reported (Mendoza et al .,2005b).As illustrated in Fig.2,the abundance of PER2-ir nuclei in the SCN revealed a pronounced daily rhythm under both AL and T-CR conditions (Time effect,F 5,37¼196.3,P <0.01;Table 1).The acrophase of the cycle under both feeding conditions occurred after the light–dark transition,slightly later than that of PER1.The daily peak in PER2expression was modulated by T-CR,leading to an increased amplitude compared to AL conditions (Fig.2).This amplitude difference was reflected in a significant difference between groups (Group effect,F 1,37¼18.4,P <0.01)and a significant interaction of group ·time (interaction,F 5,37¼4.2,P <0.01).Expression of CLOCK-ir was high throughout the SCN at all ZT in AL mice (Fig.3).In contrast,CLOCK-ir expression in the SCN of T-CR mice was slightly rhythmic,with decreased levels at ZT-8(Fig.3;Time effect,F 5,37¼2.6,P <0.05;Table 1).The total number of CLOCK-ir nuclei in AL mice was larger than that in T-CR mice (Group effect,F 1,37¼9.3,P <0.01),indicating that the constitutive level of CLOCK in the SCN cells can be changed by feeding conditions.In both AL and T-CR mice,a rhythm of A VP-positive cells was found essentially in the dorsal region of the SCN (Fig.3;Time effect,F 5,37¼4.4,P <0.002;Table 1).The anova does not show a difference between groups (Group effect,F 1,37¼2.1,NS),but there was a significant difference in the group ·time interaction (interac-tion,F 5,37¼5.5,P <0.001),revealing a clear phase advance of the rhythm in T-CR mice (Fig.3).Of note,a small population of A VP-ir cells was detected in the ventrolateral region of the middle SCN (see arrow in Fig.3).In addition,these results were checked with anova on the OD data;this showed a clear difference between groups,time and the interaction group ·time (Group effect,F 1,37¼6.2,P <0.01;Time effect,F 5,37¼9.1,P <0.001;interaction:F 5,37¼12.6,P <0.001),confirming a phase advance of the A VP rhythm by the T-CR (data not shown).Responses of circadian proteins to light in calorie-restricted miceIn both AL and T-CR animals,1h of light exposure at ZT-16led to a marked increase in PER1-ir in the medial SCN (Fig.4;F 1,28¼260.9,P <0.001)The effect of light was significant 4and 8h later (P <0.05).In T-CR mice,light induction of PER1intheTable 1.Statistical analysis of the diurnal expression of clock and clock-controlled proteins in the SCN of mice under ad libitum and T-CRF -valueP -value Ad libitum (F 5,18)PER-1302.60.001*PER-271.50.001*CLOCK 0.40.8A VP 7.00.001*T-CR (F 5,19)PER-192.00.001*PER-2131.50.001*CLOCK 3.80.01*A VP3.30.02**P <0.05,one-way anova .Timed calorie restriction alters the SCN clock 3693ªThe Authors (2007).Journal Compilation ªFederation of European Neuroscience Societies and Blackwell Publishing Ltd European Journal of Neuroscience ,25,3691–3701SCN was lower than in AL mice at ZT-20but higher at ZT-0(Fig.4).In addition,in an analysis for the rostral and caudal SCN regions the difference was significant only at ZT-0(P <0.001;data not shown).Light exposure increased the number of PER2-ir cells in the SCN of both AL and T-CR mice at ZT-20(Fig.5;F 1,28¼96.3,P <0.001).In T-CR mice,the light-increased number of PER2-ir cells was higher than in AL animals at ZT-20(Fig.5;P <0.05).The number of PER2-ir cells 8h after light exposure (i.e.ZT-0)was lower than at ZT-20in the SCN from both AL and T-CR mice (P <0.001),although the values were still higher than those in dark control animals.The same effect was present in rostral and caudal SCN (rostral,F 1,28¼153.2,P <0.001;caudal,F 1,28¼107.5,P <0.001;data not shown).Incidentally,we observed that PER2-ir in response to light was expressed in the SCN regions overlapping those cells expressing A VP (including a small ventrolateral region;see arrow in Fig.5).Fig .2.Rhythmic expression of PER1and PER2in the SCN of AL and T-CR mice.(A)Photomicrographs,and (B)mean (±SEM)number of PER1-and PER2-ir nuclei in the SCN from mice under a 12-h LD cycle and AL (black symbols)or T-CR (white symbols)at different ZTs (ZT-0,lights on;ZT-12,lights off).Black and white top bars indicate the LD cycle.Vertical arrow indicates the mealtime.*P <0.05.Scale bar,200l m.3694J.Mendoza et al.ªThe Authors (2007).Journal Compilation ªFederation of European Neuroscience Societies and Blackwell Publishing LtdEuropean Journal of Neuroscience ,25,3691–3701The CLOCK-ir cell number in the SCN was increased4h after the light pulse(Fig.6;F1,28¼15.8,P<0.001).Moreover,this increase depended on feeding conditions.The number of CLOCK-ir cells in the SCN at ZT-20was significantly increased in T-CR mice(P<0.001). In contrast,the small increase in CLOCK induction at ZT-0in the SCN of AL mice exposed to light was not significant(P¼0.08).In the rostral and caudal SCN no differences following either light treatment or feeding conditions were present(data not shown).A VP-ir cell number in the medial SCN was increased8h after light stimulation(Fig.7;F1,26¼22.2,P<0.001)in both AL and T-CR animals.No difference related to feeding conditions was found (F1,26¼0.9,NS).The A VP-ir increase in response to lightwasevident in the dorsomedial SCN,including the small set of cells in the ventrolateral region which also expressed PER2in response to light stimulation (see arrow in Fig.7).DiscussionIn mice,restricted feeding without calorie restriction may affect the behavioural phase angle of photic entrainment (Holmes &Mistlberger,2000;Sharma et al .,2000),while no clear effect of a timed normocaloric feeding was reported on clock gene expression in the SCN (Mendoza et al .,2005b).This difference on the effects at behavioural and molecular levels of these studies could reside in the feeding paradigms used.In our previous experiment (Mendoza et al .,2005b),100%of the daily food intake of the mice was given each day at the same time (i.e.normocaloric feeding)while in the behavioural study reported by Holmes &Mistlberger (2000)mice were under a temporally restricted feeding with 3h of daily food access.It will be interesting in further studies to investigate clock gene expression in mice exposed to this temporal restriction of food access.By contrast,timed calorie restriction (T-CR)modifies entrainment of the SCN to light,at both behavioural and molecular levels,as reflected by transcriptional alteration of the SCN clockwork (Challet et al .,1998;Mendoza et al .,2005b).Here we confirmed and extended our previous data by showing that T-CR also affects translational regulation of SCN clockwork and its light resetting,thus demonstrating that metabolic cues induced by a calorie restriction interact with synchronization tolight.Fig .4.Light exposure stimulated PER1translation in the mouse SCN.(Top)Representative photomicrographs show coronal sections through the middle SCN of mice kept in an LD cycle,under AL or T-CR conditions,and exposed to a 1-h light pulse at ZT-16.Mice were killed 4or 8h after the beginning of the light pulse or dark control exposure (at ZT-20and ZT-0,respectively);OC,optic chiasm;3V ,third ventricle.(Bottom)Semi-quantitative analysis of PER1-ir cells in the mouse SCN 4and 8h after 1h light exposure at ZT-16(white bars)or after exposure to darkness (black bars)in mice under AL or T-CR conditions.Data are presented as the mean ±SEM.Statistical differences between groups were determined by three-way anova followed by an LSD post hoc test.*P <0.05for differences between light conditions;#P <0.05between feeding conditions.Scale bar,200lm.Fig . 5.Light exposure triggered PER2translation in the mouse SCN.(Top)Representative photomicrographs show coronal sections through the SCN at central level of mice kept in an LD cycle,under AL or T-CR conditions,and exposed to a 1-h light pulse at ZT-16.Mice were killed 4or 8h after the beginning of the light pulse or control dark exposure (at ZT-20and ZT-0,respectively).Arrow shows the small group of cells in the A VP ventrolateral region of SCN expressing PER2.(Bottom)Semi-quantitative analysis of PER2-ir cells in the mouse SCN 4and 8h after 1h light exposure at ZT-16(white bars)or after exposure to darkness (black bars)in mice under AL or T-CR conditions.*P <0.05for differences between light conditions;#P <0.05between feeding conditions.For details see Fig.4.3696J.Mendoza et al.ªThe Authors (2007).Journal Compilation ªFederation of European Neuroscience Societies and Blackwell Publishing LtdEuropean Journal of Neuroscience ,25,3691–3701Altered daily pattern of circadian and clock-controlled proteins during T-CRIn AL mice,rhythmic expression of PER1and PER2has overlapping distribution within the SCN,in accordance with previous findings (Field et al .,2000;LeSauter et al .,2003;Maywood et al .,2003).PER1rhythm was phase-advanced in T-CR mice in comparison with control-fed animals.Moreover,the amplitude of the rhythm was modified by feeding conditions,showing a reduction in the cell number in almost the whole SCN from T-CR mice.Thus,there was a global effect of T-CR on the SCN clockwork without any apparent region-specific response.Because a normocaloric feeding does not change the daily oscillation of Per1expression in the SCN (Mendoza et al .,2005b),this suggests that food intake per se does not affect clock gene expression in the SCN.Therefore,it is unlikely that PER1changes found in the SCN of T-CR mice after feeding time (ZT-8and ZT-12)are due to the fact that mice sampled at these times were fed on the day of killing.If no food at all was provided for the groups of mice sampled from ZT-8to ZT-20,we would not have been able to avoid any possible effect of a longer duration of fasting (ranging from 26to 38h of fasting)compared to the daily fasting conditions (i.e.<24h)under T-CR conditions.The phase and amplitude effects observed at both mRNA and protein levels indicate that the timing and amplitude of PER1oscillation entrained to a LD cycle are changed by T-CR.In T-CR mice,consistent with the lack of shifting effect of T-CR on mRNA Per2expression (Mendoza et al .,2005b),daily expression of PER2does not differ between AL and T-CR.However,the amplitude of PER2oscillation was higher in T-CR mice than in AL mice.The mechanism leading to this PER2up-regulation is not clear.A recent paper raises the possibility that PER1represses light-inducedPER2Fig . 6.Light exposure increased CLOCK translation in the mouse SCN of T-CR mice.(Top)Representative photomicrographs show coronal sections through the SCN of mice kept in an LD cycle,under AL or T-CR conditions,and exposed to a 1-h light pulse at ZT-16.The light pulse induced a small but significant increase in CLOCK-ir in the SCN 4h after exposure in T-CR mice.(Bottom)Semi-quantitative analysis of CLOCK-ir cells in the mouse SCN 4and 8h after a 1-h light exposure at ZT-16(white bars)or after exposure to darkness (black bars)in mice under AL or T-CR conditions.*P <0.05for differences between light conditions.For details see Fig.4.Fig .7.Light exposure increased A VP translation in the mouse SCN of both AL and T-CR mice.(Top)Representative photomicrographs show coronal sections through the medial SCN of mice kept in an LD cycle,under AL or T-CR conditions,and exposed to a 1-h light pulse at ZT-16.Arrow shows the small group of cells in the ventrolateral region of SCN expressing A VP.The light pulse induced a significant increase in A VP-ir in the SCN 8h after exposure in both AL and T-CR mice.(Bottom)Semi-quantitative analysis of A VP-ir cells in the mouse SCN after a 1-h light exposure at ZT-16(white bars)or after exposure to darkness (black bars)in mice under AL or T-CR conditions.*P <0.05for differences between light conditions.For details see Fig.4.Timed calorie restriction alters the SCN clock 3697ªThe Authors (2007).Journal Compilation ªFederation of European Neuroscience Societies and Blackwell Publishing Ltd European Journal of Neuroscience ,25,3691–3701and attenuates phase-resetting by light(Masubuchi et al.,2005).The apparent lower amplitude of PER1expression could then partly be involved in the up-regulation of PER2expression.Moreover,acute photic responses of PER2further support this view,because its levels were increased in T-CR mice compared to AL animals.Nevertheless, even if light exposure leads to increased responses of both PER1and PER2expression,the difference in the entrainment patterns suggests that the two proteins respond to T-CR in different fashions. Furthermore,SCN resetting by T-CR has been reported to be altered in both Per1-and Per2-mutant mice(Feillet et al.,2006).Whereas Per1-mutant mice entrained to a T-CR do not show phase advances of the locomotor activity rhythm after transfer to constant darkness, Per2-mutant mice submitted to T-CR show instead larger phase advances than wild-type mice(Feillet et al.,2006).Taken together, these data and our results suggest that PER1and PER2are implicated in the phase-shifting effects of T-CR.The cellular redox state can modulate binding of CLOCK:BMAL1 (heterodimers)to DNA(Rutter et al.,2002).The circadian PER translation could then be modified by T-CR,because a chronic hypocaloric diet may affect the cellular metabolic state of the SCN clock. CLOCK is a nuclear protein expressed constitutively in the SCN of AL mice,contrasting with the rhythmic expression of PER1and PER2 (Field et al.,2000;Maywood et al.,2003;present study).It has already been proposed that CLOCK participates in resetting by T-CR because heterozygous Clock⁄+mice do not show phase advances of the circadian rhythm of locomotor activity induced by T-CR(Challet et al.,2000).Here we show that CLOCK expression was affected by T-CR,showing a general reduction in the number of immunopositive cells over the daily cycle in the SCN,with a more marked reduction 2h after mealtime.In hamsters,a short photoperiod has been shown to induce a circadian rhythm of Clock mRNA in the SCN with low levels from the beginning to the middle of light phase,suggesting that Clock is implicated in the SCN adaptation to photoperiodic changes(Tournier et al.,2003).Under natural short photoperiods,there is a reduction in food availability that is associated with many physiological changes (e.g.decreased body temperature and metabolic state,and weight loss;Malan,1978;Pe´vet,1987;Goldman,2001).Moreover,torpor produces changes in SCN clock genes in Siberian hamsters(Herwig et al.,2006)and its occurrence can be increased by food restriction (Paul et al.,2004).In mice,which are not considered a photoperiodic species,changes in body temperature(Feillet et al.,2006)and body weight as well as modifications of CLOCK expression in the SCN under T-CR conditions(present study)may reflect a short photo-period-like hypometabolic condition in this species.This hypothesis is supported by the fact that the nocturnal peak in pineal melatonin in T-CR mice is not only phase-advanced but is also longer(i.e.short photoperiod-like)than that in AL mice,in spite of similar LD (12–12h)conditions(Mendoza et al.,2005b).In nocturnal rodents fed ad libitum,A VP mRNA and protein oscillations in the SCN have been reported to peak at the middle and the end of the day,respectively(Dardente et al.,2004;Van der Zee et al.,2005).In the present study,we found that daily levels of A VP show a phase advance in T-CR mice compared to AL mice,in full accordance with previous data on A VP mRNA(Mendoza et al., 2005b).Moreover,we noted a reduction in the daily amplitude of A VP immunoreactivity during T-CR.Consistent with thisfinding,a daytime decrease in A VP expression has already been reported in rats under calorie restriction at both mRNA and protein levels(Andrade et al., 2004).In addition,the A VP release pattern from SCN is reduced in food-restricted rats(Kalsbeek et al.,2000).Taken together,these results show that food restriction,coupled or not with calorie restriction,decreases SCN A VP synthesis and release,indicating damped oscillations of this circadian output.Altered light induction of circadian proteins during T-CRLight resetting is associated with an up-regulation of Per mRNA in the SCN during the night(Albrecht et al.,1997;Shigeyoshi et al.,1997; Dardente et al.,2002;Yan&Silver,2002;Hamada et al.,2004).In mice fed ad libitum,a phase-delaying light pulse has been shown to increase PER1and PER2levels within4–9h(Field et al.,2000; Harmar et al.,2002;Yan&Silver,2004),with the number of positive cells similar to the present data(e.g.Harmar et al.,2002).In response to1-h light exposure,expression of PER1and PER2 changed depending on the feeding condition and on the time after light pulse.In T-CR animals,light-induced PER1expression was delayed, but it still occurred in the ventral SCN(i.e.in retinorecipient, vasoactive intestinal polypeptide-and gastrin-releasing peptide-con-taining cells;Dardente et al.,2002;Yan&Okamura,2002;Yan& Silver,2002;Karatsoreos et al.,2004),as observed in AL mice.This regional response is in accordance with previous studies in AL animals (von Gall et al.,2003;Yan&Silver,2004).Moreover,in T-CR mice we found that PER1expression was increased8h after light exposure whereas it was lower4h earlier.We have previously reported a reduced induction of Per1transcription in the SCN of T-CR mice1h after light exposure(Mendoza et al.,2005b).The present time course at the translational level supports the interpretation that light induction of Per1mRNA response could actually be delayed instead of being simply reduced.Moreover,because PER1oscillation in the SCN clock is advanced at least by1h in T-CR mice(present study),photic stimulation may have occurred at a slightly different circadian time. This suggests that the different PER1induction by light between AL and T-CR mice might be due to a different SCN phase at the moment of stimulation.This point,however,does not explain the delay of PER1 induction in T-CR mice that we attribute to the effects of calorie restriction.Moreover,according to the behavioural phase response curve to light,both AL and T-CR mice at ZT-16were still in the phase-delaying zone of light resetting(Mendoza et al.,2005b).For PER2expression at ZT-20,the responses to light were higher in the SCN of T-CR mice than AL animals.This specific light response profile of PER2is consistent with our data on mRNA(Mendoza et al.,2005b). Similar to previous results obtained in mice exposed to light at circadian time16(Yan&Silver,2004),PER2expression was preferentially triggered in the dorsal part(i.e.A VP cells).We also noticed light-induced PER2expression in the small ventrolateral region expressing A VP. Light exposure has been already shown to increase Clock mRNA levels in the rat SCN,suggesting that Clock is somehow involved in photic signal transduction of the circadian clock(Abe et al.,1999). However,we could not confirm this result at the translational level because no significant light-induced CLOCK expression was clearly detected in AL mice.By contrast,there was a light-induced increase in CLOCK cells in the SCN of T-CR mice,indicating that T-CR may change the molecular mechanisms underlying photic synchronization. Heterozygous Clock⁄+mice have high-amplitude phase resetting responses to6-h light pulses,an effect mediated by either reduced pacemaker amplitude or altered light inputs to the SCN clock (Vitaterna et al.,2006).Interestingly,Clock–⁄–mice do not show expected phase delays after light in the early night,but they show much larger late-night,light-induced phase advances than wild-type mice,suggesting that CLOCK may indeed have a role in the light input pathway or in SCN light sensitivity(DeBruyne et al.,2006). After a light pulse,the number of A VP-ir cells was enhanced in the medial SCN,and this increase was more marked(and significant)8h3698J.Mendoza et al.ªThe Authors(2007).Journal CompilationªFederation of European Neuroscience Societies and Blackwell Publishing LtdEuropean Journal of Neuroscience,25,3691–3701。