Daily variation of CNS gene expression in nocturnal vs. diurnal rodents and in the developing rat brain

Molecular Brain Research 48 Ž1997. 73–86

Research report

Daily variation of CNS gene expression in nocturnal vs. diurnal rodents and in the developing rat brain Bruce F. O’Hara ) , Fiona L. Watson, Rozi Andretic, Steven W. Wiler, Kimberly A. Young, Louise Bitting, H. Craig Heller, Thomas S. Kilduff Center for Sleep and Circadian Neurobiology, Departments of Biological Sciences and Psychiatry and BehaÕioral Sciences, Stanford UniÕersity, Stanford, CA 94305, USA Accepted 28 January 1997

Abstract Expression of c-fos has been shown to vary throughout the brain over the course of the 24-h day w17x. The magnitude of these changes appear to be similar in a light : dark ŽLD. cycle or in constant dark ŽDD.. To further examine whether the diurnal and circadian changes in c-fos and other immediate-early gene ŽIEG. expression in brain are related to waking behaviors such as locomotor activity, we conducted three experiments using Northern analysis. First, we compared IEG expression in nocturnal vs. diurnally active species. Second, we investigated IEG expression in a hibernating species during its active and inactive phases. Third, we examined the development of IEG expression in the young post-natal rat. As a comparison to results obtained in extra-SCN brain regions, we also examined IEG and vasopressin expression in the SCN itself across the circadian cycle. Animals maintained under a 12 : 12-h LD cycle were sacrificed in the morning Ž10:00–11:00 h, ZT2–ZT3. or night Ž22:00–23:00 h, ZT14–ZT15. or at the corresponding circadian times ŽCT. when kept in DD. Rats sacrificed in the morning always showed lower c-fos expression than at night in all brain areas examined while the reverse pattern was seen in squirrels under both LD and DD conditions, suggesting a direct correlation between c-fos message and activity. The cerebellum displayed the greatest magnitude change between morning and night Žoften reaching 10-fold.. Among other IEGs examined, the expression of NGFI-A and junB are similar to c-fos, but of lesser magnitude, whereas c-jun appears to be invariant in the rat but is increased during the active phase in squirrels. During the hibernation season, squirrels have lower levels of c-fos consistent with their low levels of activity even during their euthermic interbout periods. c-fos expression in the cerebellum and rest of brain of 1-week-old rats sacrificed at ZT3 and ZT15 showed low levels at both timepoints whereas 2- and 3-week-old animals had higher levels at night as do adults. Among other IEGs, junB and NGFI-A again were similar to c-fos while c-jun and junD were more constant. Our observations support the idea of a diurnal rhythm of IEG expression in the CNS that is related to waking behaviors. Among IEGs, c-fos exhibits the greatest daily variation in expression. Keywords: Immediate-early gene; c-fos; Vasopressin; PolyA tail; Circadian rhythm

1. Introduction The role of gene expression in diurnal and circadian rhythms has been actively studied in several model systems, including Neurospora, Drosophila, Arabidopsis, Acetabularia, Gonyaulaux, the Bulla eye, the vertebrate pineal gland and the mammalian suprachiasmatic nucleus ŽSCN., and, in many cases, transcription appears to be central to the circadian clock w14,40,53x. In mammals, the SCN is the dominant circadian pacemaker w24x and, within the SCN,

)

Corresponding author. Fax: q1 Ž415. 725-5356.

several mRNAs and proteins show circadian variation Žvasopressin, somatostatin. or changes under a light : dark ŽLD. cycle Žvasoactive intestinal peptide.. These changes have generally been measured at steady-state levels w22,24,54x but nonetheless are suggestive that, at least in some cases, gene transcription is under circadian control. In addition to undergoing modest daily changes in protein levels w8,25x, c-fos and some other immediate-early genes ŽIEGs. are dramatically induced by light pulses given during the subjective night but not during the subjective day, indicating circadian gating of gene expression w2,15,26,52x. Although the SCN appears to be the dominant circadian

0169-328Xr97r$17.00 Copyright q 1997 Elsevier Science B.V. All rights reserved. PII S 0 1 6 9 - 3 2 8 X Ž 9 7 . 0 0 0 8 4 - 3

74

B.F. O’Hara et al.r Molecular Brain Research 48 (1997) 73–86

pacemaker in mammals, most behavioral and physiological outputs of the clock, such as the sleeprwake cycle, are controlled largely by brain regions outside the SCN. Therefore, circadian and diurnal changes in gene expression in brain regions outside the SCN are also of interest. Relatively little data are available on changes in gene expression across the day in mammals for brain regions other than the SCN and pineal gland, although modest changes have been reported for a number of mRNAs, such as those coding for some of the serotonin receptors w21x which also undergo changes in ligand binding w60x along with changes in serotonin levels and metabolites w37,58x. Recently, c-fos and other IEGs have been found to undergo diurnal variations in multiple brain areas at both mRNA and protein levels. In some cases, these changes have been shown to persist under constant conditions, suggesting that they may be true circadian rhythms w17x. Other studies indicate that sleep, wake andror the activity of the animal influence the steady-state level of IEG expression w17,29x. Sleep deprivation, in particular, appears to increase IEG expression and recovery sleep appears to decrease expression w18,36,48x. Interestingly, unilateral lesions of the locus coeruleus ŽLC. dramatically reduce IEG levels in the hemisphere without LC input but are normal in the other hemisphere w10x. To further address whether such changes in IEG expression in brain are related to diurnal variations in behavior and the development of these behaviors, we conducted three experiments. First, we compared IEG expression in nocturnal vs. diurnally active species. Second, we investigated IEG expression in a hibernating species during its active and inactive phases. Third, we examined the development of IEG expression in the young post-natal rat. As a comparison to results obtained in extra-SCN brain regions, we also examined IEG expression in the SCN itself across the circadian cycle. Our results indicate that diurnal periods of higher activity strongly influence the expression of IEGs in multiple brain regions. Among IEGs, c-fos levels are most closely associated with periods of higher activity.

nocturnal rat to the diurnal ground squirrel Ž Spermophilus lateralis .. Sprague–Dawley rats were maintained under a 12 : 12-h LD cycle at 228C whereas ground squirrels Ž Spermophilus lateralis . were maintained at 78C Žlights on at 08:00 h and off at 20:00 h.. Animals of both species were sacrificed at ZT2 ŽZeitgeber time 2, i.e. 2 h after light onset. or ZT14 for the dayrnight comparison. To assess whether the diurnal changes in IEG expression in the squirrel persist under constant conditions as has been established for the rat w17x, some squirrels were placed into DD Žconstant darkness. 2 days prior to sacrifice at the corresponding circadian times, CT2 and CT14, based on the prior LD cycle. All animals were sacrificed by decapitation and the brains were dissected into cerebellum, cerebral cortex, hypothalamus, pons, medulla, thalamus, hippocampus, striatum and midbrain, frozen on powdered dry ice and stored at y708C. Squirrels kept in DD and those at ZT14 were sacrificed and their brains were removed under red light to avoid the possible effects of a light pulse. 2.1.2. Expt. 2: comparison of IEG expression in a hibernating species between its actiÕe and hibernating phases Ground squirrels Ž Spermophilus lateralis . were implanted with abdominal telethermometers ŽMini-Mitter, Sunriver, OR. and maintained at an ambient temperature of 78C on a constant 12 : 12-h LD cycle Žlights on at 08:00 h and off at 20:00 h.. Body temperature was monitored to determine whether each squirrel was in its hibernation season Žmarked by body temperatures near ambient levels punctuated with periodic arousals every 3–7 days when their body temperature returns to 378C for a period of usually 15–20 h w41x.. Squirrels designated as winter euthermic or interbout were determined to be in the hibernation season but were sacrificed at ZT2 during the 15–20 h interbout period when body temperatures was f 378C. Squirrels designated as hibernating had body temperatures of 88C and were also sacrificed at ZT2. Squirrels designated as summer-active had maintained a 378C temperature for at least 2 weeks. These squirrels were sacrificed either at ZT2 or ZT14 Ž2 h after dark onset.. Brain regions were handled as described in Expt. 1.

2. Materials and methods 2.1. Animals 2.1.1. Expt. 1: comparison of IEG expression in nocturnal Õs. diurnal mammals As a first test of the hypothesis that IEG expression was related to periods of higher activity, we compared the

2.1.3. Expt. 3: deÕelopment of gene expression in the young post-natal rat To determine whether diurnal changes in IEG expression was related to development of circadian activity periods, pregnant Sprague–Dawley and Long–Evans rats were maintained under 12:12-h LD cycle as above. The introduction of males and subsequent conception was staggered

Fig. 1. A: expression of c-fos, c-jun and jun-B in the cerebellum of adult rats and squirrels determined by Northern analysis. f 30 m g of total RNA was loaded in each lane in all experiments examining extra-SCN brain regions. The lanes labeled as L Žlight. and D Ždark. refer to animals sacrificed at ZT2 X X and ZT14 of rats sacrificed in an LD cycle; the lanes labeled as L and D refer to animals sacrificed at the corresponding CTs ŽCT2 and CT14. under DD conditions. b-Actin serves as a loading control and is used as denominator in the ratios presented in the graphs on the right side. B: expression of c-fos, c-jun and NGFI-A in the cerebellum of four additional rats sacrificed at ZT2 and ZT14. RNA from a single individual was loaded in each lane in Figs. 1–3, 5 and 7.

B.F. O’Hara et al.r Molecular Brain Research 48 (1997) 73–86

75

76

B.F. O’Hara et al.r Molecular Brain Research 48 (1997) 73–86

by f 1 week, such that on 1 experimental day three Sprague–Dawley litters were available for the experiment: one at post-natal day 8 ŽPD8., one at PD15 and one at PD22. For the Long–Evans strain, we also obtained three litters, one at PD8, one at PD17 and one at PD23. Half of each litter was sacrificed at ZT3 and the other half at ZT15. For the Sprague–Dawley litters, the brains were dissected into cerebellum and the rest of brain. In the Long–Evans pups, the SCN was removed prior to any further dissection. The ZT3 Sprague–Dawley litters were dissected prior to the litters sacrificed at ZT15. For the Long–Evans pups, this sequence was reversed to control for possible stress effects on the second set of pups Žwhich had been disturbed 12 h previously..

2.1.4. Expt. 4: assessment of circadian changes in gene expression within the suprachiasmatic nucleus To examine SCN gene expression across the circadian cycle, 136 adult Sprague–Dawley rats were placed into DD 2 days prior to sacrifice. At each of eight timepoints ŽCT1, CT4, CT7, CT10, CT13, CT16, CT19 and CT22., 17 rats were sacrificed and the SCN were dissected as previously described w35x. Other brain regions were handled as described in Expt. 1. 2.2. RNA isolation and analysis RNA was isolated from all brain regions using the guanidiniumrCsCl method w9x, fractionated on 1.2% for-

Fig. 2. Northern analysis of c-fos expression in five different brain regions of ground squirrels kept under 12 : 12-h LD cycle. Squirrels presented in the lanes labeled I and H were sacrificed in different phases of the hibemation cycle during winter; squirrels presented in the lanes labeled D and L were sacrificed in different phases of the light cycle during the summer Žactive. phase. I, interbout euthermic phase ŽT b s 378C. sacrificed at ZT2; H, deep hibernation phase ŽT b s 7–88C. sacrificed at ZT2; D, sacrificed at ZT14; L, sacrificed at ZT2. n s 2 for each condition, however, the L vs. D squirrels in Fig. 1A are essentially the same as those here in Fig. 2.

B.F. O’Hara et al.r Molecular Brain Research 48 (1997) 73–86

maldehyderagarose gels and transferred to Nytran membranes ŽSchleicher and Schuell.. RNA was visualized by ethidium bromide staining and cross-linked by UV irradiation. Following pre-hybridization, membranes were hybridized at 428C in 5 = SSC, 50% formamide, 50 mM sodium phosphate ŽpH 6.8., 1% SDS, 1 mM EDTA, 2.5 = Denhardt’s, 200 mgrml herring sperm DNA and 5 = 10 6 cpmrml of radiolabeled random-primed cDNA probe w16x. After mild washes in 1 = SSC, membranes were washed 2 = for 30 min at 588C in 0.4 = SSC and 0.5% SDS. Filters were then exposed to Kodak XAR5 film for 1–12 days and the band intensity quantitated using a MCID image analysis system. Filters were subsequently stripped to remove old probe and re-hybridized with other probes. Comparisons between samples were made relative to b-actin mRNA for adult samples or junD mRNA levels for developmental comparisons. b-Actin mRNA levels decline during development w33x while junD mRNA levels were found to remain constant under our conditions. Both probes served to control for unequal loading or transfer. 2.3. cDNA probes For hybridization to c-fos message, a 2.1-kb EcoRI fragment of rat c-fos was used w12x. To detect c-jun and

77

junB, 2.7- and 1.8-kb EcoRI fragments of the respective mouse clones were utilized w42,43x. For NGFI-A, a 450-bp PÕuIIrEcoRI fragment of a rat cDNA was used w30x. Human b-actin w20x and mouse junD w44x cDNA were used as control probes. To examine vasopressin gene expression, the probe AVPexCc containing only exon C of the rat was obtained from Dr. Thomas G. Sherman ŽUniversity of Pittsburgh..

3. Results 3.1. Expt. 1: comparison of IEG expression in nocturnal Õs. diurnal mammals The expression of c-fos and other IEGs in the cerebellum of rats and squirrels is shown in Fig. 1A,B. As shown by previous investigators w17x, the nocturnal rat has higher levels of c-fos at night ŽFig. 1A,1B.. On the other hand, diurnal summer-active ground squirrels have high levels of c-fos during the early daytime Ž CT2rZT2 vs. CT14rZT14.. In both species, higher levels of c-fos correspond to periods of high locomotor activity. Other IEGs show dayrnight changes that are less clear. In the squirrel,

Fig. 3. Northern analysis of five IEGs Ž NGFI-A, c-fos, c-jun, jun-B and jun-D . in the cerebellum of the developing Sprague–Dawley rat. The rats in the lanes labeled L were sacrificed at ZT3; the rats in the lanes labeled D were sacrificed at ZT15.

78

B.F. O’Hara et al.r Molecular Brain Research 48 (1997) 73–86

c-jun appears to be somewhat higher in the day, however, there is almost no dayrnight difference in the six rats in Fig. 1A,B nor in others not shown Žunpublished data.. NGFI-A ŽFig. 1B. and junB ŽFig. 1A. appear to be more similar to c-fos but with smaller dayrnight changes. Other brain regions examined show similar patterns but with smaller changes than the cerebellum Žunpublished observations.. 3.2. Expt. 2: comparison of IEG expression in a hibernating species between its actiÕe and hibernating phases Levels of c-fos mRNA were also compared in the brains of winter euthermicrinterbout ŽZT2., hibernating ŽZT2., summer-active night ŽZT14. and summer-active day ŽZT2. squirrels ŽFig. 2.. In all brain regions examined, the highest c-fos levels were observed in summer-active day squirrels, with the cerebellum and cortex exhibiting the greatest changes. These data support the idea that c-fos expression in the brain is correlated with periods of higher activity.

3.3. Expt. 3: deÕelopment of gene expression in the young post-natal rat Fig. 3 presents the results of Northern analysis of five IEGs in the cerebellum of the developing Sprague–Dawley rat. A trend toward increasing IEG expression with age was evident for all IEGs except jun-D which remained constant or even decreased slightly with age. The expression of jun-D was much more constant over the age range studied than, for example, expression of b-actin. Consequently, jun-D was used as a reference or denominator in all numerical analyses of gene expression presented below. The appearance of a junB doublet in Fig. 3 and in Fig. 7 Žbut not in Fig. 1. we believe is artifactual. It is probably due to the variable migration of the 18S rRNA which has a complex secondary structure that persists to varying degrees even under denaturing conditions. When it comigrates precisely with the junB mRNA, it appears to cause band splitting. We have never observed this doublet when using polyA RNA Žunpublished observations.. Fig. 4 presents the results of the densitometric analyses

Fig. 4. Densitometric analyses for the autoradiographs presented in Fig. 3. The expression of jun-D in Fig. 3 was used as denominator to calculate the ratios presented in this figure since junD expression appears to be constitutive. P, post-natal day.

B.F. O’Hara et al.r Molecular Brain Research 48 (1997) 73–86

79

Fig. 5. Northern analysis of the expression of the IEGs NGFI-A, c-fos and c-jun in the cerebellum of the developing Long–Evans rat. L, D and PD as in Fig. 3. Do to the deterioration of this filter, junD was probed on a replicate Northern not shown.

for the autoradiographs illustrated in Fig. 3. Analysis of variance revealed a significant change in expression of all IEGs with age Ž P F 0.0003.. Post-hoc tests ŽFisher PLSD. indicated that all age groups differed significantly from one another. Significant time-of-day effects were also evident for both c-fos Ž P - 0.0003; F s 24.594; df s 1,12.

and jun-B Ž P - 0.01; F s 8.087; df s 1,12. as well as a significant interaction of age by time-of-day Žc-fos: P 0.02; F s 6.114; df s 2,12; jun-B: P - 0.006; F s 8.127; df s 2,12.. There was also a trend toward increased NGFI-A expression by time-of-day which did not quite reach statistical significance Ž P - 0.07..

Fig. 6. Densitometric analyses corresponding to the data presented in Fig. 5. JunD was used as denominator as in Fig. 4 to calculate the ratios presented in this figure. P, post-natal day.

80

B.F. O’Hara et al.r Molecular Brain Research 48 (1997) 73–86

Fig. 5 presents the results of Northern analysis similar to that of Fig. 3 but from the cerebellum of the developing Long–Evans rat. As in Fig. 3, there appears to be a trend toward age-related increase in IEG expression and the results of the corresponding densitometric analyses are presented in Fig. 6. ANOVA confirmed the trend toward age-related increase in IEG expression. For c-fos, post-hoc tests revealed that all age comparisons were significantly different. For all other IEGs, post-hoc tests revealed that all age comparisons were significantly different except the PD17–PD23 comparison Žin contrast to the Sprague–Dawley rats.. This observation may reflect the older ages used in the Long–Evans rats than the Sprague Dawleys ŽPD17 v. PD15; PD23 v. PD22.. For c-fos, in addition to the age-related effect, ANOVA revealed significant time-ofday Ž P - 0.0001; F s 109.725; df s 1,12. effect and a significant interaction of age by time Ž P - 0.0001; F s 31.329; df s 2,12.. ANOVA also revealed significant time-of-day Ž P - 0.01. effect and a significant interaction of age by time-of-day Ž P - 0.002. for both c-jun bands. Curiously, however, the mean daytime values for c-jun exceeded the mean nighttime values by PD23, in contrast to the Sprague–Dawley rats.

Fig. 7 presents the expression of IEGs in the remainder of the brain from the experiment presented in Fig. 5. Again, the trend toward increasing IEG expression with age is evident for all IEGs except jun-D and the results of the corresponding densitometric analyses are presented in Fig. 8. ANOVA confirmed the age-related change for all IEGs except jun-D Ž P - 0.006.. Significant time-of-day effects were evident for c-fos Ž P - 0.0001; F s 64.379; df s 1,12., NGFI-A Ž P - 0.0001; F s 37.374; df s 1,12. and jun-B Ž P - 0.004; F s 13.217; df s 1,12. and a significant interaction of age by time-of-day was evident for c-fos Ž P - 0.0005; F s 15.285; df s 2,12. and NGFI-A Ž P - 0.009; F s 7.324; df s 2,12.. Among these latter three IEGs, the mean nighttime values exceeded the mean daytime values by PD17. 3.4. Expt. 4: circadian changes in gene expression within the suprachiasmatic nucleus Arginine vasopressin ŽAVP. mRNA in the SCN has been shown to undergo dramatic changes in polyA tail length resulting in an additional deadenylated mRNA size at night in some studies w39x but not in others w4x. Our

Fig. 7. Expression of the IEGs NGFI-A, c-fos, c-jun, jun-B and jun-D in the remainder of the brain Žminus the cerebellum and SCN. from the Long–Evans rats presented in Fig. 5. L, D and PD as in Figs. 3 and 5.

B.F. O’Hara et al.r Molecular Brain Research 48 (1997) 73–86

results in adult rats are consistent with the existence of this circadian and diurnal variation ŽFig. 9A.. We did not find dramatic changes in abundance of AVP mRNA throughout the circadian day, however, no statistical analysis is possible since each CT sample is a single pool of SCNs. We then investigated the RNA from the PD17 and PD23 SCNs of the Long–Evans rats used in Expt. 3 and again found the diurnal size variation, although not as clearly as in adults ŽFig. 9B.. The SCN dissections at the PD17 and PD23 timepoints were not as precise as our adult dissections which may account for the smearing since only the SCN appears to produce a completely deadenylated AVP mRNA w6,7,39x. PD8 SCN samples did not work in these experiments. However, in other experiments, we have seen the two band smear at inappropriate times in PD6 SCN at ZT9 but not adult SCN at ZT9 Žunpublished observations.. In addition, we have never observed the doublet at any timepoints in the mouse SCN. Using the same RNAs as in Fig. 9, we examined c-fos expression across the circadian cycle in the adult rat SCN ŽFig. 10.. Unlike the cerebellum and the rest of brain, the SCN appears to have higher c-fos expression during the subjective day, although statistical analyses cannot be undertaken due to the pooling of the tissue samples. In comparison to the nocturnal increase in c-fos mRNA ob-

81

served in the cerebellum ŽFig. 1., the elevation in the SCN during the day is quite modest.

4. Discussion The purpose of our studies was to further examine whether the diurnal and circadian changes in IEG expression in brain initially described by others w17x may be related to locomotor activity. The experiments presented here provide correlational evidence that is consistent with such a relationship. Among IEGs, c-fos appears to undergo the greatest daily variation, with c-fos expression in the SCN apparently being an exception Žsee below.. Our experiments do not allow us to address whether the variation in IEG expression in brain regions outside the SCN is directly driven by the SCN or is a secondary consequence of behavior Žsuch as locomotor activity. whose timing is controlled by the SCN but we favor the latter possibility. The examination of c-fos and expression of other IEGs in the brain is of interest for at least two reasons. Increased IEG expression is often related to increased neuronal activity w3,11,32x, making IEGs a good general marker of brain ‘activity’ regardless of their actual function. In this regard, c-fos is probably the best IEG to examine since it

Fig. 8. Densitometric analyses corresponding to the data presented in Fig. 7 using jun-D as denominator. P, post-natal day.

82

B.F. O’Hara et al.r Molecular Brain Research 48 (1997) 73–86

Fig. 9. A: expression of AVP mRNA in the suprachiasmatic nucleus ŽSCN. of adult Sprague–Dawley rats. In each lane, 4 m g total RNA has been loaded representing a portion of a pooled RNA sample from 17 rats sacrificed under DD at each of the 8 CTs indicated. b-Actin serves as a loading control. B: expression of AVP mRNA in the SCN of adult and young post-natal Long–Evans rats. L s ZT3 and Ds ZT15 for both the adult and post-natal rats ŽPD17 and PD23..

has been the most extensively characterized with respect to changes in neuronal activity w32,50x. Our data showing that c-fos mRNA levels are higher at night in nocturnal rats is, thus, consistent with the fact that neural activity is gener-

ally higher when animals are awake, at least as compared to the periods of slow-wave sleep which comprise 80% of total sleep time w23,51x. Our data are also consistent with previous work examining IEG expression in several brain regions of the rat throughout the 24-h period w17x. We have extended this work by showing that diurnal ground squirrels have higher levels of c-fos expression during the day in contrast to the nocturnal increase observed in rats. Furthermore, during the interbout intervals between hibernation periods which are characterized by relatively low levels of activity, squirrels have low levels of c-fos mRNA as compared to summer animals. Winter squirrels display little activity between their bouts of hibernation and spend a large percentage of these periods in slow-wave sleep, before their body temperature drops for the next hibernation cycle w57x. The observations reported here are consistent with a link between IEG expression and neuronal activity as well as locomotor activity. In fact, the basal changes in c-fos levels may correspond quite closely with the actual firing rates in neurons. Variation in c-fos mRNA is on the order of 5–10-fold Žsee Figs. 1 and 2 and w17x. and, despite substantial variation throughout the brain, the average firing rate of neurons is also probably f 5–10-fold higher in wake vs. slow-wave sleep. For example, EEG recordings reflecting thalamocortical activity have large amounts of low frequency delta waves during slow-wave sleep Ž0.5–4 Hz. while waking periods are dominated by much higher frequency waves w23,51x. In addition to their role as functional markers, IEG expression is of interest because of the role of IEGs as transcription factors. Since IEGs are often the first genes activated by various stimuli, they have been referred to as ‘master switch’ genes w32x. Therefore, it is important to determine which IEGs increase, decrease or remain constant in relation to behavioral or physiological events. IEGs interact at the protein level to form dimers Žreferred to as AP-1. and it appears that the specific combination confers functional specificity w32,45x, presumably resulting

Fig. 10. Expression of c-fos mRNA in the adult SCN across the circadian cycle using the same RNAs as shown in Fig. 9A. b-Actin serves as a loading control and is used as denominator in the ratios presented in the graph on the right side.

B.F. O’Hara et al.r Molecular Brain Research 48 (1997) 73–86

in the activation of specific ‘target’ genes. For example, in some instances, c-fos expression increases many-fold while c-jun remains essentially unchanged w29,34x. Our results in adult rats indicate that c-jun expression remains relatively constant between the two timepoints examined and is consistent with Menegazzi et al. w29x who found no significant changes in c-jun expression throughout the 24-h period. Earlier in development, we observed no diurnal change in c-jun mRNA in one experiment ŽFig. 3. but a reverse pattern in the second experiment ŽFig. 5. which is in clear contrast to c-fos expression. It is interesting to note that seizure increases c-jun expression in the rat w32x while the more modest and subtle increase in neuronal activity during wake produces no dectectable increase in the rat but does perhaps cause an elevation in the squirrel ŽFig. 1.. From our experiments, it appears that the junB gene is the most responsive IEG in the JUN family, followed by c-jun. JunD mRNA remains almost completely constant in all of our conditions Žage, day vs. night., consistent with other reports of junD w44x in addition to many unpublished studies Žthis is based on our comparisons of junD levels with the amount of RNA loaded.. Therefore, JUND and C-JUN proteins may be present in the majority of AP-1 binding during sleep and periods of inactivity while junB becomes more important during periods of higher activity, assuming that changes in mRNA are reflected as changes in protein levels. Since c-fos is low during periods of inactivity, it is not clear what Fos protein might be dominant during the day in rats. FosB appears to be even less abundant during the day than c-fos mRNA Žunpublished observations. and we did not examine Fra-1 or Fra-2. It may be that during periods of inactivity, AP-1 proteins are predominantly JUN-JUN dimers which have weaker binding and enhancerrpromoter effects w45x while at night there is a shift to c-fos-containing AP-1 proteins. In the SCN, JUND has been shown to be a constitutive component of AP-1 while complexes of JUNB and FOS required photic stimulation w55x. If c-fos expression is related to locomotor activity, the ontogeny of diurnal variation in this gene expression is of interest. The circadian clock located in the SCN appears to be functioning by E19 Žembryonic day 19. in rats w38x and the LD cycle can entrain at least the pineal melatonin rhythm by PD8, if not sooner w13x. Clear locomotor activity rhythms, however, are not present until at least PD15 in pups reared in isolation but may emerge earlier in development with maternal influence w1,56x. Our data showing the appearance of diurnal changes in c-fos expression at 2 and 3 weeks of age, but not at 1 week of age, is consistent with the development of activity rhythms. However, it is possible that expression of c-fos is simply too low in the cerebellum and rest of brain for us to detect diurnal variation at PD8. Our results and those of others show that c-fos substantially increases its expression between the 1st and 2nd post-natal week in the cerebellum and to a lesser extent in other brain regions w19,27x. NGFI-A Žalso known

83

as zif268 and krox24. also appears to increase in the first few post-natal weeks in both the cerebellum and rest of brain, consistent with earlier reports in the rat and cat w28,59x. Among the JUN family members during development, junB appears similar to c-fos, c-jun increases less dramatically and junD is almost entirely invariant. These developmental changes in IEG expression during development seem to typify each member of the JUN family: in most situations of which we are aware, junB expression generally parallels c-fos, junD is invariant and c-jun either is invariant or increases along with c-fos but to a lesser degree. This was true for our dayrnight differences and in our previous studies examining sleep deprivation w34x. The most unusual pattern we observed was in Fig. 5 where c-jun expression is reversed from c-fos; however, in the rest of brain from the same animals ŽFig. 7., the more typical pattern is observed. In general, the changes in expression observed in c-fos, junB and NGFI-A in the three different comparisons conducted here – Ž1. rats vs. squirrels in the day and night; Ž2. differing seasonal and hibernation phases in squirrels; and Ž3. day vs. night across post-natal development in rats – are all correlated with locomotor andror neuronal activity. Although stress and many other factors have also been shown to alter IEG expression w7,46,47x, there is no obvious difference in stress under our conditions. Indeed, it is possible that studies of stress may be inducing IEGs by altering arousal state and locomotor activity. Even if animals are restrained, many of the same pathways are likely to be activated. We have observed apparent stress effects in some of our studies, however. For example, we have observed large variations in c-fos expression at ZT8 Žunpublished data.. Part of this variation is likely due to an individual animal’s recent sleep and activity history, as documented by Grassi-Zucconi et al. w18x, and our consistent findings in rats of low c-fos mRNA at ZT2rZT3 and high c-fos at ZT14rZT15 may reflect the high levels of sleep in the early morning and high locomotor activity in the early night among animals. However, in some instances, we have also seen very high c-fos and other IEG mRNA levels following stressful events, such as cage changing and previous animal sacrifice in the same room Žunpublished data.. The rat SCN may be one brain region in which the interrelationship between IEG expression, locomotor activity and neuronal activity may be dissociated. The SCN has its peak firing rate and metabolism at CT6 Žor ZT6. in both nocturnal and diurnal species, even in the isolated SCN slice w31x. Thus, if IEG expression is related to neuronal activity rather than locomotion, we might expect higher levels of c-fos expression in the day rather than night in this structure, the opposite of most brain regions in the rat. One attempt to distinguish neuronal from locomotor activity was made by examining c-fos in the SCN. The results presented in Fig. 10 show slightly elevated c-fos expression in the SCN during the day which is consistent

84

B.F. O’Hara et al.r Molecular Brain Research 48 (1997) 73–86

with earlier reports at the protein level w8,25x. Since the SCN is not involved directly in locomotor activity, it is not surprising that c-fos would not always be correlated with locomotor activity. It seems likely that neuronal activity is most closely linked with c-fos expression but that locomotor activity may, in many instances, be the driving force for increased neuronal activity, especially in structures, such as the cerebellum. An independent measure that the SCN RNA examined in Fig. 10 was indeed under circadian regulation is the variation in polyA tail length in the AVP mRNA ŽFig. 9.. This variation represents another level of regulation beyond controlling transcriptional rate and was first observed by Robinson et al. w39x and later by Carter and Murphy w5x who also reported that transcriptional rate of this gene decreases at night w6x. The SCN produces a completely deadenylated form of AVP at night with a size of 530 bp in addition to a longer adenylated form of 740 bp. During the day, only the 740-bp size is observed which is also the only form present in other hypothalamic nuclei at both day and night under normal conditions w5,6,39x. Aside from an independent check that our SCN RNA was under circadian regulation, we also wanted to replicate the previous findings since there has been some inconsistencies in the published literature as to both polyA tail length and quantitative changes in steady-state AVP mRNA levels. Cagampang et al. w4x report a sharp peak of AVP mRNA levels at ZT8 and CT8 and a trough at ZT20 and CT20 but report that they do not consistently observe the shorter AVP mRNA at night. This quantitative change throughout the day has been reported as early as E19, providing evidence of SCN clock function before birth w38x. In contrast, Robinson et al. w39x found no significant variation in mRNA levels throughout the day despite a very clear rhythm of polyA tail length. Carter and Murphy w5,6x find both a quantitative and polyA change in AVP mRNA, however, they often observe only the short mRNA at night Žas opposed to both bands.. Our results are essentially identical to Robinson et al. w39x with a clear shift from a single band during the day to two bands at night. We observed no clear change in mRNA levels in our aroundthe-clock study ŽFig. 9A. but each timepoint represents only a single pooled sample from 17 rats. In other ‘two timepoint’ comparisons, we have seen f 2-fold greater levels of AVP mRNA in the day vs. night ŽFig. 9B adults and unpublished observations.. We have extended the above findings by showing the polyA tail length variation in younger rats ŽFig. 9B. but have seen considerable variation in rats during the 1st post-natal week and have not observed the smaller band in mice Žunpublished observations.. Deadenylation is also critical in regulating the level of c-fos mRNA since this is the first step in its rapid decay w49x. It may be of interest to note that in one of our experiments ŽFig. 3, c-fos panel., we observe an apparent size fluctuation in the six PD22 lanes, with the higher band being present at night. This size difference is even more

apparent on shorter autoradiographic exposures than that illustrated in Fig. 3. However, we have not seen this size variation in any other samples of young or adults rats. The expression and regulation of c-fos and AVP in the SCN reflect the unique role of this structure as the dominant circadian pacemaker in mammals. Our data from the cerebellum and rest of brain, however, suggest important changes in gene expression throughout the day in extraSCN regions. These changes appear to be correlated with daily activity patterns. Each IEG studied here responds differently to these presumed changes in activity. Given the role of IEGs in AP-1 formation and transcriptional regulation, it seems likely that variations in gene expression related to changes in daily activity are quite common.

Acknowledgements We thank Dr. Norman ŽBud. Ruby for assistance and advice in the experiments using squirrels and Ms. May Chin for help with the manuscript. We would also like to thank Drs. Daniel Nathans, Tom Curran, Jeffrey Milbrandt, Elaine Fuchs and Thomas Sherman for providing cDNAs. This work was supported by NIH Grants P50 HD29732, P01 AG11084 and K21 DA00187.

References w1x V.N. Anderson, G.K. Smith, Effects of feeding and light cycles on activity rhythms of maternally isolated rat pups, Physiol. BehaÕ. 39 Ž1987. 169–181. w2x N. Aronin, S.M. Sagar, F.R. Sharp, W.J. Schwartz, Light regulates expression of a Fos-related protein in rat suprachiasmatic nuclei, Proc. Natl. Acad. Sci. USA 87 Ž1990. 5959–5962. w3x D. Bartel, M. Sheng, L.F. Lau, M.E. Greenberg, Growth factors and membrane depolarization activate distinct programs of early response gene expression, Genes DeÕ. 3 Ž1989. 304–313. w4x F.R. Cagampang, J. Yang, Y. Nakayama, C. Fukuhara, S.T. Inouye, Circadian variation of arginine-vasopressin messenger RNA in the rat suprachiasmatic nucleus, Mol. Brain Res. 24 Ž1994. 179–184. w5x D.A. Carter, D. Murphy, Diurnal rhythm of vasopressin mRNA species in the rat suprachiasmatic nucleus: independence of neuroendocrine modulation and maintenance in explant culture, Mol. Brain Res. 6 Ž1989. 233–239. w6x D.A. Carter, D. Murphy, Nuclear mechanisms mediate rhythmic changes in vasopressin mRNA expression in the rat suprachiasmatic nucleus, Mol. Brain Res. 12 Ž1992. 315–321. w7x S. Ceccatelli, M.J. Villar, M. Goldstein, T. Hokfelt, Expression of ¨ c-fos immunoreactivity in transmitter-characterized neurons after stress, Proc. Natl. Acad. Sci. USA 86 Ž1989. 9569–9573. w8x I. Chambille, S. Doyle, J. Serviere, Photic induction and circadian expression of Fos-like protein. Immunohistochemical study in the retina and suprachiasmatic nuclei of hamster, Brain Res. 612 Ž1993. 138–150. w9x J.M. Chirgwin, A.E. Przybyla, R.J. MacDonald, W.J. Rutter, Isolation of biologically active ribonucleic acid from sources enriched in ribonuclease, Biochemistry 18 Ž1979. 5294–5299. w10x C. Cirelli, M. Pompeiano, G. Tononi, Neuronal gene expression in the waking state: a role for the locus coeruleus, Science 274 Ž1996. 1211–1215.

B.F. O’Hara et al.r Molecular Brain Research 48 (1997) 73–86 w11x A.J. Cole, D.W. Saffen, J.M. Baraban, P.F. Worley, Rapid increase of an immediate early gene messenger RNA in hippocampal neurons by synaptic NMDA receptor activation, Nature 340 Ž1989. 474–476. w12x T. Curran, M.B. Gordan, K.L. Rubino, L.C. Sambucetti, Isolation and characterization of the c-fosŽrat. cDNA and analysis of posttranslational modification in vitro, Oncogene 2 Ž1987. 79–84. w13x M.J. Duncan, M.J. Banister, S.M. Reppert, Developmental appearance of light-dark entrainment in the rat, Brain Res. 369 Ž1986. 326–330. w14x J.C. Dunlap, Closely watched clocks: molecular analysis of circadian rhythms in Neurospora and Drosophila, Trends Genet. 6 Ž1990. 159–165. w15x D.J. Earnest, M. Iadarola, H.H. Yeh, J.A. Olschowka, Photic regulation of c-fos expression in neural components governing the entrainment of circadian rhythms, Exp. Neurol. 109 Ž1990. 353–361. w16x A. Feinberg, B. Vogelstein, A technique for radiolabeling DNA restriction endonuclease fragments to high specific activity, Anal. Biochem. 132 Ž1982. 6–9. w17x G. Grassi-Zucconi, M. Menegazzi, A.C. De Prati, A. Bassetti, P. Montagnese, P. Mandile, C. Cosi, M. Bentivoglio, c-fos mRNA is spontaneously induced in the rat brain during the activity period of the circadian cycle, Eur. J. Neurosci. 5 Ž1993. 1071–1078. w18x G. Grassi-Zucconi, A. Giuditta, P. Mandile, S. Chen, S. Vescia, M. Bentivoglio, c-fos spontaneous expression during wakefulness is reversed during sleep in neuronal subsets of the rat cortex, J. Physiol. 88 Ž1994. 91–93. w19x R.M. Gubits, J.L. Hazelton, R. Simantov, Variations in c-fos gene expression during rat brain development, Mol. Brain Res. 3 Ž1988. 197–202. w20x I. Hanukoglu, N. Tanese, E. Fuchs, Complementary DNA sequence of a human cytoplasmic actin: interspecies divergence of 3X non-coding regions, J. Mol. Biol. 163 Ž1983. 673–678. w21x M.C. Holmes, K.L. French, J.R. Seckl, Modulation of serotonin and corticosteroid receptor gene expression in the rat hippocampus with circadian rhythm and stress, Mol. Brain Res. 28 Ž1995. 186–192. w22x S.T. Inouye, S. Shibata, Neurochemical organization of circadian rhythm in the suprachiasmatic nucleus, Neurosci Res. 20 Ž1994. 109–130. w23x B.E. Jones, Basic mechanisms of sleep-wake states. In M.H. Kryger, T. Roth and W.C. Dement ŽEds.., Principles and Practice of Sleep Medicine, W.B. Saunders, Philadelphia, PA, 1989, pp. 121-138. w24x D.C. Klein, R.Y. Moore and S.M. Reppert ŽEds.., Suprachiasmatic Nucleus: The Mind’s Clock, Oxford University Press, New York, NY, 1991, 467 pp. w25x J. Kononen, J. Koistinaho, H. Alho, Circadian rhythm in c-fos-like immunoreactivity in the rat brain, Neurosci. Lett. 120 Ž1990. 105– 108. w26x J.M. Kornhauser, D.E. Nelson, K.E. Mayo, J.S. Takahashi, Photic and circadian regulation of c-fos gene expression in the hamster suprachiasmatic nucleus, Neuron 5 Ž1990. 127–134. w27x B.E. Kosofsky, L.M. Genova, S.E. Hyman, Postnatal age defines specificity of immediate early gene induction by cocaine in developing rat brain, J. Comp. Neurol. 351 Ž1995. 27–40. w28x M.A. McCormack, K.M. Rosen, L. Villa-Komaroff, G.D. Mower, Changes in immediate early gene expression during postnatal development of cat cortex and cerebellum, Mol. Brain Res. 12 Ž1992. 215–223. w29x M. Menegazzi, A.C. De Prati, G. Grassi-Zucconi, Differential expression pattern of c-jun in the rat brain during the 24-h cycle, Neurosci. Lett. 182 Ž1994. 295–298. w30x J. Milbrandt, A nerve growth factor-induced gene encodes a possible transcriptional regulatory factor, Science 238 Ž1987. 797–799. w31x J.D. Miller, On the nature of the circadian clock in mammals, Am. J. Physiol. 264 Ž1993. R821–R832. w32x J.I. Morgan, T. Curran, Stimulus-transcription coupling in the nervous system: involvement of the inducible proto-oncogenes fos and jun, Annu. ReÕ. Neurosci. 14 Ž1991. 421–451.

85

w33x B.F. O’Hara, S. Fisher, M.L. Oster-Granite, J.D. Gearhart, R.H. Reeves, Developmental expression of the amyloid precursor protein, growth-associated protein 43, and somatostatin in normal and trisomy 16 mice, DeÕ. Brain Res. 49 Ž1989. 300–304. w34x B.F. O’Hara, K.A. Young, F.L. Watson, H.C. Heller, T.S. Kilduff, Immediate early gene expression in brain during sleep deprivation: preliminary observations, Sleep 16 Ž1993. 1–7. w35x B.F. O’Hara, R. Andretic, H.C. Heller, D.B. Carter, T. Kilduff, GABA A , GABA B , and NMDA receptor subunit expression in the suprachiasmatic nucleus and other brain regions, Mol. Brain Res. 28 Ž1995. 239–250. w36x M. Pompeiano, C. Cirelli, G. Tononi, Immediate-early genes in spontaneous wakefulness and sleep: expression of c-fos and NGFI-A mRNA and protein, J. Sleep Res. 3 Ž1994. 65–81. w37x L. Poncet, L. Denoroy, M. Jouvet, Daily variations in in vivo tryptophan hydroxylation and in the contents of serotonin and 5-hydroxyindoleacetic acid in discrete brain aeas of the rat, J. Neural Transm. 92 Ž1993. 137–150. w38x S.M. Reppert, G.R. Uhl, Vasopressin messenger ribonucleic acid in supraoptic and suprachiasmatic nuclei: appearance and circadian regulation during development, Endrocrinology 120 Ž1987. 2483– 2487. w39x B.C. Robinson, D.M. Frim, W.J. Schwartz, J.A. Majzoub, Vasopressin mRNA in the suprachiasmatic nucleus: daily regulation of polyadenylate tail length, Science 241 Ž1988. 342–344. w40x M. Rosbash, Molecular control of circadian rhythms, Curr. Opin. Genet. DeÕ. 5 Ž1995. 662–668. w41x N.F. Ruby, J. Dark, H.C. Heller and I. Zucker, Ablation of suprachiasmatic nucleus alters timing of hibernation in ground squirrels, Proc. Natl. Acad. Sci. USA Žin press.. w42x K. Ryder, D. Nathans, Induction of protooncogene c-jun by serun growth factors, Proc. Natl. Acad. Sci. USA 85 Ž1988. 8464–8467. w43x K. Ryder, L.F. Lau, D. Nathans, A gene activated by growth factors is related to the oncogene v-jun, Proc. Natl. Acad. Sci. USA 85 Ž1988. 1487–1491. w44x K. Ryder, A. Lanahan, E. Perez-Albuerne, D. Nathans, JunD: a third member of the Jun gene family, Proc. Natl. Acad. Sci. USA 86 Ž1989. 1500–1503. w45x R.-P. Ryseck, R. Bravo, c-jun, JUN B, and JUN D differ in their binding affinities to AP-1 and CRE consensus sequences: effect of FOS proteins, Oncogene 6 Ž1991. 533–542. w46x S.S. Schreiber, G. Tocco, T.J. Shors, R.F. Thompson, Activation of immediate early genes after acute stress, NeuroReport 2 Ž1991. 17–20. w47x F.R. Sharp, S.M. Sagar, K. Hicks, D. Lowenstein, K. Hisanaga, c-fos mRNA, Fos, and Fos-related antigen induction by hypertonic saline and stress, J. Neurosci. 11 Ž1991. 2321–2331. w48x J.E. Sherin, P. Shiromani, J. Morgan, C.B. Saper, Recovery sleep in Fos-LacZ mice leads to a rapid decline of the elevated Fos and b-gal levels which result from forced waking, Soc. Neurosci. Abstr. 21 Ž1995. 1679. w49x A.-B. Shyu, J.G. Belasco, M.E. Greenberg, Two distinct destabilizing elements in the c-fos message trigger deadenylation as a first step in rapid mRNA decay, Genes DeÕ. 5 Ž1991. 221–231. w50x R.J. Smeyne, K. Schilling, L. Robertson, D. Luk, J. Oberdick, T. Curran, J.I. Morgan, Fos-lacZ transgenic mice: mapping sites of gene induction in the central nervous system, Neuron 8 Ž1992. 13–23. w51x M. Steriade, Brain electrical activity and sensory processing during waking and sleep states. In M.H. Kryger, T. Roth and W.C. Dement ŽEds.., Principles and Practice of Sleep Medicine, W.B. Saunders, Philadelphia, PA, 1989, pp. 86-120. w52x E.L. Sutin, T.S. Kilduff, Circadian and light-induced expression of immediate early gene mRNAs in the rat suprachiasmatic nucleus, Mol. Brain. Res. 15 Ž1992. 281–290. w53x J.S. Takahashi, Circadian-clock regulation of gene expression, Curr. Opin. Genet. DeÕ. 3 Ž1993. 301–309.

86

B.F. O’Hara et al.r Molecular Brain Research 48 (1997) 73–86

w54x J. Takeuchi, H. Nagasaki, K. Shinohara, S.I.T. Inouye, A circadian rhythm of somatostatin messenger RNA levels, but not of vasoactive intestinal polypeptiderpeptide histidine isoleucine messenger RNA levels in rat suprachiasmatic nucleus, Mol. Cell Neurosci. 3 Ž1992. 29–35. w55x J. Takeuchi, W. Shannon, N. Aronin, W.J. Schwartz, Compositional changes of AP-1 DNA-binding proteins are regulated by light in a mammalian circadian clock, Neuron 11 Ž1993. 825–836. w56x M.H. Teicher, L.E. Flaum, Ontogeny of ultradian and nocturnal activity rhythms in the isolated albino rat, DeÕ. Psychobiol. 12 Ž1979. 441–454. w57x L. Trachsel, D.M. Edgar, H.C. Heller, Are ground squirrels sleep

deprived during hibernation?, Am. J. Physiol. 260 Ž1991. R1123– R1129. w58x L. Voog, T. Eriksson, Diurnal rhythms in rat brain large neutral amino acids ŽLNAAs., monoamines and monoamine metabolites, J. Neural Transm. 87 Ž3. Ž1992. 215–224. w59x M.A. Watson, J. Milbrandt, Expression of the nerve growth factorregulated NGFI-A and NGFI-B genes in the developing rat, DeÕelopment 110 Ž1990. 173–183. w60x N. Weiner, H.W. Clement, D. Gemsa, W. Wesemann, Circadian and seasonal rhythms of 5-HT receptor subtypes, membrane anisotropy and 5-HT release in hippocampus and cortex of the rat, Neurochem. Int. 21 Ž1992. 7–14.