758 Research Paper

Long-term monitoring of circadian rhythms in c-fos geneexpression from suprachiasmatic nucleus culturesMichael E. Geusz*, Colin Fletcher†, Gene D. Block*, Martin Straume*, Neal G. Copeland†, Nancy A. Jenkins†, Steve A. Kay‡ and Richard N. Day*

Background: The AP-1 family of transcription factors has been implicated inthe control of the expression of many genes in response to environmentalsignals. Previous studies have provided temporal profiles for c-fos expressionby taking measurements from many animals at several points in time, but thesestudies provide limited information about dynamic changes in expression. Here,we have devised a method of continuously measuring c-fos expression.

Results: A transgenic mouse line expressing the human c-fos promoter linkedto the firefly luciferase reporter gene (fos/luc) was generated to continuouslymonitor c-fos gene expression. A second transgenic mouse line expressingluciferase under the control of the cytomegalovirus promoter (CMV/luc) servedas a control. Luminescence originating from identifiable brain regions wasimaged from fos/luc brain slice cultures. Expression of the fos/luc transgeneaccurately reflected transcriptional responses of the endogenous c-fos gene.Dynamic changes in fos/luc expression in suprachiasmatic nuclei (SCN) explantcultures were monitored continuously, and luminescence showed almost24 hour rhythms lasting up to five circadian cycles. In contrast, bioluminescencemonitored from CMV/luc SCN explant cultures was not rhythmic.

Conclusion: The fos/luc transgenic mouse will be useful for long-term, non-invasive monitoring of c-fos transcriptional responses to the changing cellularenvironment. Circadian rhythms in c-fos expression can be monitored non-invasively in real time from the SCN, clearly demonstrating that c-fostranscription is regulated by the circadian clock.

BackgroundCells respond to developmental or environmental cues bychanging their patterns of gene expression. The AP-1family of transcription factors, composed of the Fos-related and Jun-related proteins, has been implicated as a‘master switch’ in modifying gene expression in responseto a changing environment. The c-fos gene is expressed inall tissues and, depending upon the cell type, can beinduced by a variety of stimuli including growth factors,endocrine hormones, neurotransmitters, stress or lightinput. The regulatory elements of the c-fos gene promoterprovide the transcriptional responses to all of these signalsand are highly conserved between species [1]. The tempo-ral and spatial distribution of c-fos expression in the centralnervous system has been intensely investigated using arange of experimental techniques [2].

A light-induced diurnal rhythm in c-fos expression in thesuprachiasmatic nuclei (SCN) of the hypothalamus haspreviously been characterized and is particularly importantto the studies presented here. The SCN are critical to cir-cadian-rhythm generation [3–5] and the rhythmic expres-sion of c-fos in the SCN provides a potentially useful

marker for the circadian pacemaker [6–12]. The AP-1 tran-scription factors have been implicated in the entrainmentof the SCN to daily light cycles [13], although theirprecise role as a molecular signal in the circadian time-keeping mechanism remains to be determined [14,15].Techniques currently used to assay c-fos expression in theSCN only give information about the degree and grossspatial pattern of expression. A technique that allows con-tinual monitoring of c-fos gene expression in living cellswould provide a unique perspective on transcriptionalresponses to the changing cellular environment. More-over, the ability to monitor a dynamic pattern of transcrip-tion over long periods of time would be particularlyrelevant to the study of the relatively slow cellular oscilla-tions that are the hallmark of circadian rhythms.

A transgenic mouse model, the fos–lacZ mouse, was shownto report accurately both constitutive and induced c-fosexpression [16,17]. As with immunohistochemistry or insitu hybridization, however, the lacZ reporter gene pro-vides limited information regarding dynamic changes in c-fos gene expression. In contrast, the bioluminescentfirefly luciferase protein has proven to be a useful reporter

Addresses: *NSF Center for Biological Timing,University of Virginia, Gilmer Hall, Charlottesville,Virginia 22903, USA. †Mammalian GeneticsLaboratory, Building 539, ABL-Basic ResearchProgram, NCI-Frederick Cancer Research andDevelopment Center, Frederick, Maryland 21702,USA. ‡Department of Cell Biology, The ScrippsResearch Institute, 10550 North Torrey Pines Road,La Jolla, California 92037, USA.

Correspondence: Michael GeuszE-mail: meg7u@virginia.edu

Received: 3 April 1997Revised: 27 June 1997Accepted: 29 July 1997

Published: 8 September 1997

Current Biology 1997, 7:758–766http://biomednet.com/elecref/0960982200700758

© Current Biology Ltd ISSN 0960-9822

protein for monitoring the dynamics of gene activity inliving cells [18,19]. Luminescence from luciferaseexpressed in transgenic plants, Drosophila, zebrafish andmammalian cells in culture has been used to monitor real-time dynamic changes in gene transcription from theliving organism [20–23].

Here, we report the characterization of a transgenic mousethat expresses the firefly luciferase gene under the controlof the human c-fos promoter (fos/luc). This animal modelprovides a unique system for monitoring changes in c-fospromoter activity in living tissue explant or dispersed cellcultures. A second transgenic mouse line was generated toserve as an important control for the fos/luc studies: inthese mice, luciferase transcription is driven by thecytomegalovirus (CMV) promoter (CMV/luc). Using dis-persed cell and tissue cultures derived from the mice, wepresent the first clear demonstration in vitro of a circadianrhythm in c-fos transcriptional activity.

ResultsAnalysis of the distribution of bioluminescence fromfos/luc and CMV/luc mouse tissuesFour independent founder strains were derived and char-acterized for luciferase expression for both the fos/luc andCMV/luc transgenic mice. Southern blot analysis was usedto confirm single-site integration and transmission of thetransgenes to the offspring. The expression of theluciferase transgene in protein extracts prepared from a

variety of different tissues from the fos/luc and CMV/lucmice was determined by luminometry. For the CMV/lucmice, luciferase expression was detected in all tissuesexamined, with levels ranging from a low of 100 Turnerlight units/mg protein (TLU/mg) in spleen to over7052 TLU/mg in heart. In contrast, luciferase activity wassubstantially lower in tissues from the fos/luc mice, withlevels ranging from 1 to 5 TLU/mg in kidney, liver, lung,spleen, heart and various regions of the brain. Testis andskin were found to have the highest luciferase expressionlevels, exceeding 100 TLU/mg.

We used imaging of bioluminescence from living brain-slice cultures prepared from either the fos/luc or CMV/luctransgenic mice to examine the spatial distribution of

Research Paper Circadian rhythm in c-fos expression Geusz et al. 759

Figure 1

Distribution of luciferase bioluminescence in coronal brain slicecultures prepared from fos/luc and CMV/luc adult mice. (a) Upper leftpanel: bioluminescence detected from anterior brain fos/luc slicecultures 4 h after dissection using a 30 min camera integration. Thepseudocolor overlay of the bright-field images shows relativebioluminescence as a scale from low indicated by blue to highindicated by red and is the same for all images. Upper right panel:bioluminescence detected using a 30 min camera integration from thesame preparations shown in the upper left panel after 22 h in culture.Note that the luminescence signal was highest in discrete, paired brainregions near the ventricles. Lower left panel: bioluminescencedetected from brain slice cultures prepared from the anterior brain ofCMV/luc mice. Luminescence was detected using 30 min integration1 h after dissection. Lower right panel: a different group of CMV/lucbrain slice cultures were imaged with 1 min camera integration 44 hafter dissection. Bioluminescence in CMV/luc cultures was notassociated with particular brain regions. (b) Bioluminescence from aneonatal fos/luc mouse anterior brain slice maintained in culture for1 week. The luminescence signals were detected using an optic fiber-coupled camera and a 10 min integration time. Luminescence wasobserved from the thalamus, hypothalamus, and hippocampus. (c) Bioluminescence from a brain slice culture prepared from a 20 dayold fos/luc mouse. The slice was maintained in culture for 2 daysbefore monitoring of bioluminescence. Luminescence was detectedoriginating from the region of the hypothalamus, including thesuprachiasmatic nucleus (SCN). A 30 min camera integration wasused and the pseudocolor overlay of the bright-field image indicatesrelative signal intensity. This culture stained with thionin is shown onthe right for comparison.

Fos/luciferase

CMV/luciferase

4 h 22 h

Fos/luciferase

Fos/luciferase Bright-field

SCN SCN

Hippocampus

Thalamus

Paraventricular nucleusSCN

Supraoptic nucleus

1 h 44 h

(a)

(b)

(c)

luciferase expression (Figure 1a, upper panels). By4–6 hours after dissection, the slice cultures from fos/lucmice showed diffuse bioluminescence over the entire sliceculture, which might be caused by a general induction ofc-fos transcription in response to dissection. Certain pairedbrain regions showed particularly high bioluminescence inslice cultures 22 hours after dissection (Figure 1a, upperpanels). Slice cultures from CMV/luc mice showeddetectable bioluminescence within 1 hour of dissection,consistent with the significantly higher level expression ofthe CMV/luc transgene compared to fos/luc. Biolumines-cence was more uniformly distributed than that in fos/luccultures and was not clearly associated with specific brainregions (Figure 1a, lower panels). Imaging of biolumines-cence from coronal slice cultures prepared from neonatalfos/luc mice and maintained in culture for one weekrevealed that luciferase activity was present in identifiablebrain regions: the septum, hippocampus, thalamus andhypothalamus (Figure 1b). Lower level signals wereobserved from suprachiasmatic, supraoptic and paraven-tricular nuclei of the hypothalamus, whereas little or nosignal was detected from the cerebral cortex. Lumines-cence was also observed from the hypothalamus, arcuateand hippocampus of slice cultures prepared from a 20 dayold fos/luc mouse after 2 days in culture (Figure 1c).

Induction of the fos/luc transgeneThe region encompassing –405 to +42 bp of the human c-fos promoter was used in these studies. This region ofthe promoter includes the identified sis-inducibleelement, the serum response and AP-1 composite ele-ments, and the consensus cAMP response element. Thesesequences are highly conserved between human andmouse [1]. In the brain, induction of c-fos promoter activ-ity by a variety of stimuli is both rapid and transient. Wenext used dispersed hypothalamic cell cultures from fos/lucmice to assess the response of the transgene to severalstimuli known to induce c-fos transcription. The fos/luchypothalamic cell cultures were maintained in serum-freemedium and bioluminescence was monitored continu-ously as described in Materials and methods. The changein bioluminescence was determined over a 5 hour periodfollowing the indicated treatment (Figure 2). When theculture medium was exchanged with medium containingthe convulsant pentylenetetrazole (PTZ) at 1 mM or45 mM KCl to trigger c-fos transcription, bioluminescencesignificantly increased by 5-fold and 20-fold, respectively,compared to the control cell culture (p < 0.05, Figure 2).Likewise, medium containing 10% newborn calf serum(NCS) resulted in approximately 50-fold induction of bio-luminescence (p < 0.01, Figure 2).

To demonstrate that expression of the transgene mirroredthat of the endogenous c-fos gene, we examined changes inmRNA levels for both c-fos and luciferase in response toserum, using dispersed brain cell cultures from fos/luc mice.

Total RNA was prepared from the brain cell cultures main-tained under serum-free conditions prior to a 25 minutepulse of 10% serum (0 minutes) or at 30 minutes or180 minutes after the beginning of the serum pulse. RNAblot hybridization was performed using uniformly labeledcRNA complementary to mouse c-fos exon 4, a portion ofthe firefly luciferase coding sequence, or the mouse glycer-aldehyde-3-phosphate dehydrogenase (GAPDH) tran-script. The results shown in Figure 3a demonstrated aprofound induction of the c-fos transcript 30 minutes afterthe addition of serum. As would be expected for thisimmediate-early gene, the mRNA levels returned tocontrol levels by 180 minutes after the addition of serum.Similarly, the luciferase transgene mRNA was induced at30 minutes and the levels of this transcript returned to nearcontrol levels at 180 minutes after serum treatment. Aninternal control, the mouse GAPDH mRNA, did not showthis pattern of rapid and transient response to serum butwas constitutively expressed. Taken together, these resultsdemonstrate that the luciferase transgene accuratelymirrors the responses that are characteristic of endogenousc-fos gene transcriptional activity.

We then used continuous photon counting to examine thetemporal characteristics of luciferase activity as a reporter

760 Current Biology, Vol 7 No 10

Figure 2

Induction of fos/luc transgene expression by stimuli known to activatec-fos gene transcription. Dispersed hypothalamic cell culturesprepared from fos/luc mice were maintained in serum-free medium at25°C. The indicated treatments were added to the culture medium;control cultures received medium alone. Medium was exchanged withmedium containing 1 mM pentylenetetrazole (PTZ), 45 mM KCl (K+) or10% newborn calf serum, and bioluminescence was monitoredcontinuously. The average change in the bioluminescence signal(± s.e.m.) 5 h after addition of the indicated treatment is plotted. Thenumber of cultures measured is given below the figure.

3000

2500

Cha

nge

in p

hoto

n co

unts

(cou

nts/

min

)

2000

1500

1000

500

0Mediumcontroln = 14

PTZ1 mMn = 8

K+45 mMn = 7

Serum10%n = 8

of the serum response for dispersed brain cell cultures.Some of the photon counting studies described below wereperformed using explant cultures maintained at 25°C. Astemperature was expected to have an effect on the expres-sion and turnover of the luciferase reporter protein, wecompared the time course of the expression of the trans-gene in response to serum for cultures maintained at either25°C or 36°C. The results shown in Figure 3b demonstrateinduction of the bioluminescence signal in response to a25 minute pulse of 10% serum. The time to peak biolumi-nescence after serum treatment was significantly faster forcultures maintained at 36°C than those maintained at 25°C(4.65 ± 0.22 hours, n = 4; and 12.28 ± 1.86 hours, n = 7,respectively, Figure 3b). Furthermore, the time to halfdecay in the peak bioluminescence signal was also signifi-cantly more rapid for cultures maintained at 36°C thanthose maintained at 25°C (5.5 ± 0.50 hours, n = 4; and12.35 ± 2.27 hours, n = 4, respectively, Figure 3b). Theinduction of Fos protein has previously been reported tofollow closely that of the transcript, peaking at30–45 minutes, and the protein half-life was determined tobe about 2 hours [24]. The half-life of the firefly luciferaseprotein, when expressed in mammalian cells in culturemaintained at 37°C, was reported to be similar to that ofthe Fos protein [25]. Our results show that the biolumines-cence signal from the fos/luc brain cell cultures maintainedat 36°C has slower induction and decay kinetics than mightbe predicted for the endogenous Fos protein. These

results, however, clearly demonstrated the transient natureof the bioluminescent response from these living cell cul-tures, illustrating the potential use of this preparation formonitoring changes in c-fos transcriptional activity overlong periods of time.

Circadian rhythms in fos/luc bioluminescenceThe diurnal rhythm in c-fos mRNA and Fos protein inSCN has been characterized [6–12], and Smeyne et al. [17]have demonstrated that β-galacosidase activity can beinduced in the SCN of fos–lacZ transgenic mice by expo-sure to a light pulse in the middle of the night. To examinethe ability of the fos/luc transgenic animal to monitor rhyth-mic changes in c-fos transcriptional activity over time, weapplied continuous photon counting to living fos/luc SCNexplant cultures. To improve the signal-to-noise ratio byreducing the dark counts of the photomultiplier tube, wemonitored the luminescence of the SCN explant culture at25°C. The results in Figure 4 demonstrate that circadianbioluminescence rhythms could be detected from bothneonatal (Figure 4a,b) and adult (Figure 4c,d) SCNexplant cultures prepared from fos/luc mice. Biolumines-cence from these cultures was recorded for as long as20 days, and showed almost 24 hour luminescence rhythmsthat persisted for up to five cycles (Figure 4a–d). One ofthe neonatal explant cultures that was not obviously rhyth-mic showed an increased luminescence signal and theemergence of a clear circadian luminescence rhythm after

Research Paper Circadian rhythm in c-fos expression Geusz et al. 761

Figure 3

c-fos Luciferase GADPH

0 30 180

2.2 kb

(a) (b) (c)

1.4 kb

0 30 180 0 30 180 Time to peak Time to half decayTime (h)

14

12

10

8

6

4

2

025ºC 36ºC 25ºC 36ºC

Tim

e (h

)

2000

1500

1000

Per

cent

age

chan

ge in

pho

tons

/min

500

0

0 10 20

10% serum25 min

30 40

25ºC

36ºC

50

Comparison of the induction of fos/luc transgene expression to that ofthe endogenous c-fos gene and analysis of the temporalcharacteristics of luciferase expression. (a) Northern blot analysis ofthe serum induction of endogenous c-fos mRNA, luciferase mRNA andGADPH mRNA in dispersed brain cell cultures prepared from fos/lucmice. The fos/luc brain cell cultures were maintained in serum-freemedium for 5 days prior to treatment with NCS. Total RNA wasprepared from control cell cultures before addition of serum, and at30 min and 180 min after a 25 min pulse of 10% serum. The RNA wasfractionated by denaturing agarose gel electrophoresis, and aftertransfer to nylon membranes, the blots were hybridized with uniformlylabeled cRNA probes complementary to mouse c-fos exon 4, theluciferase coding sequence or mouse GADPH. The 2.2 kb c-fos

mRNA, 3 kb luciferase transcript and 1.4 kb mRNA for the GADPHhousekeeping gene were detected. (b) Serum induction of luciferaseexpression in dispersed brain cell cultures from fos/luc mice.Bioluminescence signals were monitored continuously from fos/lucbrain cell cultures maintained at 25°C or 36°C. After 14 h, bothcultures received a 25 min treatment with 10% NCS. The change inthe bioluminescence signal over time is plotted. (c) Time to peakluminescence and decay of the luminescent signal. Both the time tothe peak luminescence after addition of serum and the time to halfdecay of the signal were assessed for cultures maintained at bothtemperatures. The results are plotted as time to peak signal and timeto half decay of the peak signal (± s.e.m.).

the addition of forskolin, a known inducer of c-fos genetranscription (Figure 4b). A circadian luminescencerhythm also appeared in adult SCN explants andemerged as early as 5 days after dissection (Figure 4c,d).Rhythms were detected in four of seven adult SCNexplant recordings. We applied a fast Fourier transformnon-linear least squares routine (FFT-NLLS) [26,27] toobtain period estimates for the luminescence rhythmsrecorded from fos/luc SCN explants (Figure 5). The vari-ance-weighted average period for both neonatal and adultfos/luc explant SCN cultures was 24.32 ± 0.38 hours (errorof the mean, n = 8).

These results suggest that circadian clock-regulated c-fostranscription gives rise to the rhythmic expression of theluciferase reporter gene and the resulting circadian

luminescence rhythm. It is possible, however, that thesteps in the luciferase enzymatic reaction were the targetof some activity driven by the circadian pacemaker, suchas rhythmic changes in metabolism. To control for possi-ble effects of the SCN pacemaker on luciferase itself,SCN explant cultures were made from CMV/luc mice andbioluminescence was recorded for up to 3 weeks (n = 7).These cultures produced a much higher signal than thefos/luc explants and showed no evidence of circadianrhythmicity. The signal remained high for several daysand then declined during the rest of the recording period(Figure 6a). As an additional control, we monitoredbioluminescence from fos/luc mouse brain explant cul-tures prepared from anterior hypothalamus and nodetectable rhythms in the luminescence signal wereobserved (Figure 6b).

762 Current Biology, Vol 7 No 10

Figure 4

Circadian rhythms in bioluminescencemonitored continuously from fos/luc SCNexplant cultures. (a) A rhythm from a neonatalexplant after 1 week in culture with 10%serum. (b) A neonatal explant showsincreased bioluminescence and an obviouscircadian rhythm after addition of 50 µMforskolin (arrow). (c,d) Circadian rhythms inbioluminescence in SCN explant cultures fromadult fos/luc mice maintained in definedmedium.

0 1 2 3 4 5230

240

250

260

270

280

290

300

310(c) (d)

Pho

ton

coun

ts/h

(x 1

03)

Time (days) Time (days)

0 1 2 3 4 50.005

0.010

0.015

0.020

Time (days)

0 1 2 3

0.04

0.05

(a) (b)

Time (days)

Turn

er li

ght u

nits

0 1 2 3 4 5 6166

167

168

169

170

171

172

173

Turn

er li

ght u

nits

Pho

ton

coun

ts/h

(x 1

03)

Research Paper Circadian rhythm in c-fos expression Geusz et al. 763

DiscussionWe have demonstrated that the fos/luc transgenic mousemodel is useful for long-term, non-invasive monitoring ofc-fos transcriptional activity from dispersed cell culturesor tissue explants. Using RNA blot hybridization, we

demonstrated that both the endogenous c-fos gene tran-script and the c-fos promoter-driven luciferase gene tran-script were rapidly and transiently induced by serum.These results showed that both the induction of fos/luctransgene expression and the turnover of the resultingluciferase mRNA transcript accurately reflected those ofendogenous c-fos. We have shown that expression of theluciferase reporter protein could be detected in particularbrain regions by imaging of bioluminescence from fos/lucmouse brain slice cultures, whereas imaging from brainslice cultures prepared from the CMV/luc transgenic miceshowed only a diffuse pattern of bioluminescence. Thepattern of c-fos expression in the brain has been used as ageneral indicator of neural activity [2,16,28,29], as a bio-marker for excitotoxin and shellfish neurotoxin activity[28,30] and even as a marker for wakefulness [31]. Ourstudies indicate that the fos/luc transgene is useful as adynamic indicator of c-fos transcriptional responses inspecific brain regions, and possibly also in other tissues.

In order to characterize the responses of the transgene, weused several different stimuli that are known to induce c-fos promoter activity. By monitoring bioluminescence ofhypothalamic cell cultures from fos/luc mice, we showedinduction of reporter protein activity in response to serum,depolarizing potassium levels, or the convulsant PTZ.These results demonstrated transcriptional responses ofthe transgene to a range of signal transduction pathwaysknown to induce c-fos gene transcription [28,29,32–34].Moreover, the response to the γ-aminobutyric acid(GABA) receptor antagonist, PTZ, indicated that func-tional GABAergic synapses, characteristically found withinthe SCN [34], were present in the hypothalamic cell cul-tures and capable of conferring a signal to the transgene.

Figure 5

Period estimates of SCN explant rhythms relative to age of thecultures. Cultures prepared from neonates, juveniles and adults areplotted relative to their time in culture plus their age when dissected.Period estimates were obtained using FFT-NLLS routine as describedin Materials and methods. Error bars are 95% confidence limits of theperiod estimates.

0 5020

25

30

Per

iod

(h)

35

40

100Age (days)

150 200

Juveniles oradultsNeonates

Figure 6

Bioluminescence monitored from controlexplant cultures. Continuous monitoring ofbioluminescence from (a) SCN explantsprepared from CMV/luc mice or (b) explantsfrom the anterior hypothalamus of adultfos/luc mice which did not include the SCN.In (a), CMV/luc SCN cultures produced highsignals of luminescence that then slowlydecayed over 2 weeks. Recordings are shownfrom two independent cultures. In (b), anteriorhypothalamic explants outside of the SCNfrom fos/luc mice showed slowly decayingsignals over 1 week of continuous monitoring.

0 1 2 3 4 5 6 7 8

158

160

162

164

166

168

Pho

ton

coun

ts/h

(x 1

03)

Pho

ton

coun

ts/h

(x 1

03)

Time (days)Time (days)

0 3 6 9 12 150

6000

12000

18000

24000

(a) (b)

764 Current Biology, Vol 7 No 10

Together, our results demonstrate that induction of trans-gene expression by specific signal transduction pathwayscan be monitored non-invasively from the living fos/luc cellcultures using the luciferase reporter protein.

The feature that distinguishes luciferase from otherreporter proteins that are commonly used to assess geneactivity is its rapid turnover, which allows the luciferaseprotein to accurately report rapid changes in gene expres-sion [19]. Thompson et al. [25] reported that the half-lifeof luciferase in a stably transfected human hepatoblastomacell line maintained at 37°C was approximately 3 hours,which is similar to the 2 hour half-life reported for the Fosprotein [24]. Using continuous monitoring of biolumines-cence from dispersed fos/luc mouse brain cell culturesmaintained at 36°C, we demonstrated that the peak lumi-nescence occurred approximately 4.6 hours after seruminduction, and decay of the luminescent signal to half ofthe peak response occurred over 5.5 hours. While usingthis method to monitor rhythms in c-fos transcriptionalactivity from SCN explant cultures, we observed that theluminescent signal from some of the SCN explant culturesincreased continuously for the duration of the recording(Figure 4). It is possible that the half-life of luciferase inthese preparations, monitored at 25°C, was sufficientlylong to prevent a return to baseline before the initiation ofthe next cycle, resulting in a continuous increase in thesignal. This trend, however, did not interfere with thedetection of the rhythmic expression of the transgene.Furthermore, analysis of the data by the NLLS routine[26,27] showed that this gradual change did not influencethe period estimates. Although bioluminescence from thefos/luc cell cultures has a profile of induction and decay ofsignal that was slower than that expected for the endoge-nous Fos protein, our results verified the transient natureof the luminescent signal from these living cell culturesand revealed that this preparation would be useful formonitoring changes in c-fos transcriptional activity overlong time periods.

Several investigators have used luciferase as a non-invasivereporter of dynamic changes in gene expression over longtime periods from living transgenic organisms and in cellsin culture [20–23]. Studies from Brandes et al. [21] provedthe advantages of this non-invasive approach over conven-tional biochemical techniques for obtaining high time-res-olution profiles of changes in the expression of theDrosophila clock gene period from living transgenic flies.We continuously monitored luminescence to assess thetemporal regulation of c-fos transcriptional activity in theSCN of fos/luc mice. It has been established that changesin c-fos mRNA and Fos immunoreactivity in SCN occurduring the circadian cycle, and that light induction of c-fosexpression depends on the phase of the pacemaker [6–11].However, direct evidence for a circadian rhythm in c-fosgene expression that persists under constant conditions is

based upon a limited number of time points obtainedusing conventional biochemical assays [35–38]. We presenthere the first high-time resolution measurements of c-fostranscriptional activity from several SCN explant culturesfrom fos/luc mice and demonstrate clear circadian rhythmsin the bioluminescence signal. These rhythms were notdue to an influence of the circadian pacemaker on theluciferase protein or its enzymatic activity because biolu-minescence signals from CMV/luc SCN cultures were notrhythmic. Furthermore, we found no evidence of circadianrhythmicity in brain explant cultures from fos/luc micetaken from outside the SCN, again suggesting that theSCN pacemaker itself was responsible for the biolumines-cence rhythm.

We have generated a transgenic mouse model that will beuseful for long-term monitoring in vitro of c-fos promotertranscriptional responses to the changing cellular environ-ment. Our results demonstrate that this unique animalmodel should be useful for the investigation of circadianpacemaker control of c-fos transcription in the SCN.Recent studies have demonstrated long-term, non-invasiverecording of circadian rhythms in spontaneous electricalactivity from neonatal rat SCN cell cultures [39] and mouseSCN explants [40] using microelectrode arrays. This tech-nology could possibly be combined with continuous moni-toring of luminescence rhythms from fos/luc SCN culturesto determine the temporal relationship between rhythmicc-fos transcription and spontaneous electrical activity. Ourstudies have been limited to the analysis of the spatial andtemporal expression of the fos/luc transgene in specificbrain regions. As the fos/luc transgene is expressed in alltissues, however, every cell type from these animals hasthe potential to be used as a biosensor for signal transduc-tion pathways that act upon the c-fos promoter.

Materials and methodsGeneration of transgenic miceA 447 bp fragment containing the human c-fos promoter (from –405 to+42 bp) was excised from the pFC700 plasmid [41] (kindly providedby Joseph Nevins, Duke University) and inserted into the pLucIAVvector (R. Day, unpublished observations) upstream of the luciferasereporter gene. This vector encodes a mutant luciferase protein thatlocalizes to the cytosol, preventing peroxisomal loading associated withthe native form of the protein that may interfere with cellular processes[42]. A 3750 bp DNA fragment containing the human c-fos promoter,luciferase reporter gene, SV40 small-t intron and polyadenylation signalwas excised and purified. The DNA fragment was microinjected intothe male pronuclei of fertilized mouse eggs (B6C3HF1 × B6). A similarstrategy was used for generation of mice transgenic for the CMV pro-moter linked to luciferase [43]. Confirmation of single-site integration ofthe transgenes into the mouse genome was assessed by Southern blotanalysis of DNA isolated from tail clips. Subsequent analysis of expression of the luciferase protein in mice that had integrated thetransgene was confirmed by analysis of luciferase activity in the skinfrom tail clips as described below. This approach was used to monitorpassage of the transgene to offspring. Of four original fos/luc founderlines, the line showing highest expression was used for all subsequentbreeding. Homozygous CMV/luc and heterozygous fos/luc mice wereused for bioluminescence recordings from cell and tissue cultures.

Assaying luciferase activity in tissue lysatesThe distribution of luciferase expression in fos/luc and CMV/luc trans-genic animals was determined by analysis of luciferase activity usingthe method described by Brasier et al. [44]. Tissues were rapidlyremoved, minced and immediately frozen. The frozen tissue fragmentswere thawed in lysis buffer containing 25 mM glycylglycine (free basepH 7.8), 15 mM MgSO4, 4 mM EGTA, 1 mM dithiothreitol and 1%Triton X-100. After trituration, the samples were centrifuged for 5 minat 13,000 × g and the supernatant was recovered. For analysis ofluciferase activity, each sample was diluted in 380 ml 25 mM glycyl-glycine buffer (pH 7.8) containing 15 mM MgSO4, 4 mM EGTA, 15 mMKPO4, 1 mM dithiothreitol and 2 mM ATP. Light emission from thesamples was measured using a Turner model 20e luminometer follow-ing the addition of 100 µl D-luciferin (monopotassium salt). Totalprotein was determined using the Bio-Rad protein assay.

Dispersed cell culturesMice were maintained on a cycle of 12 h light/12 h dark to entrain thecircadian system. All animal care protocols were approved by the Com-mittee on Animal Care and Use at the University of Virginia. After decap-itation of 4–6 day-old mice and removal of brains, the SCN region wasdissected with iridectomy scissors. The SCN and a small amount of sur-rounding hypothalamic tissue were placed in chilled Hank’s bufferedsalts solution with 10 mM Hepes, 100 U/ml penicillin and 100 mg/mlstreptomycin (HBSS). The tissue harvested from a litter of mice wasdigested for 30–60 min in 1 ml papain with DNase (Worthington) at37°C, triturated with glass fire-polished pipettes, and added to anexcess of culture medium consisting of Dulbecco’s modified eaglemedium (DMEM) without glutamate, with 10% NCS, B27 supplements(Gibco BRL) and antibiotics; this medium had previously been equili-brated with 5% CO2. The cells were centrifuged at 300 × g for 5 min,and resuspended in medium. Aliquots of 20 µl were placed on glasscoverslips coated with poly-D-lysine in 35 mm culture dishes and thecells were allowed to attach for 10 min before adding 2 ml medium. Typ-ically, one litter produced 15–20 12 mm cover slips. The dishes weremaintained in a 5% CO2 incubator for 5–21 days before recording. Halfof the medium in the dish was exchanged every 3–4 days.

Northern analysisDispersed whole-brain cell cultures were prepared from neonatal fos/lucmice as described above and used to inoculate 100 mm culture dishes.The cultures were maintained in serum-free medium for 5 days beforethe start of the experiment. Two 100 mm cell cultures were used for thecontrol (0 min) and the remaining dishes received a 25 min treatmentwith 10% NCS. Total RNA was extracted at the indicated time pointsusing the Total RNA isolation kit (Ambion) and integrity of the RNA wasassessed by electrophoresis and detection of 28S and 18S ribosomalRNAs with ethidium bromide. For determination of endogenous mousec-fos mRNA and luciferase mRNA transcribed from the transgene,10 µg total RNA for each time point was loaded on a denaturingagarose gel. For determination of mouse GAPDH mRNA levels as aninternal standard, 5 µg total RNA was loaded in adjacent lanes on thesame gel. The RNA was fractionated by electrophoresis and transferredto nylon membrane (Zeta-probe, Bio-Rad) by capillary diffusion. Themembranes were cut between lanes and, after prehybridization, themembranes were hybridized with 1 × 106 cpm/ml antisense 32P-labeledcRNA probes complementary to mouse c-fos exon 4, mouse GAPDH(pTri-c-fos/exon 4-mouse and pTri-GAPDH, respectively; Ambion), orthe last 413 bp of the firefly luciferase coding sequence (pGEM Luc,Promega) at 65°C overnight. The membranes were washed as recom-mended by Ambion and exposed to film at –70°C.

SCN explant culturesAdult and juvenile mice were decapitated after anesthetizing them withsodium pentobarbital injected I.P. The brain was removed and placedin chilled HBSS. The paired SCN were cut from 400 µm coronal brainslices made with a Vibroslicer (World Precision Instruments) andplaced in a thin film of serum-free Hepes-buffered DMEM in whichsodium bicarbonate was reduced to 0.35 mg/ml and 10 mM Hepes,

B27 supplements and 1 mM luciferin were added. This medium, whichjust covered the tissue, promoted attachment and thinning of the cul-tures. SCN explants from neonatal (3–6 day old) mice were also cutfrom 400 µm coronal slices. Before recording bioluminescence, theseexplants were maintained for at least 1 week in a 5% CO2 incubator at37°C with a thin layer of the cell culture medium, changing the mediumevery 2–3 days. The neonatal SCN explant cultures were given 10%serum at all times.

Bioluminescence recordingBioluminescence from cell cultures and brain slice cultures was moni-tored in three different ways. Localization of luciferase expression toregions in brain slice cultures was determined by imaging biolumines-cence with an intensified charge-coupled device (ICCD) camera(C2400-97) and an Argus-50 photon-counting image processor fromHamamatsu Photonic Systems. Photon counts were integrated overtimes between 1 min and 1 h. Cultures were maintained in a thin layer ofHepes DMEM medium with 1 mM luciferin at room temperature (22°C) ina humidified environment. Some neonatal brain slices were maintained ina CO2 incubator with medium used for dispersed cell cultures for5–7 days before imaging. Brain slices were imaged from above with a50/1.2 Nikon camera lens. Alternatively, brain-slice cultures were placedon top of an optic fiber bundle directly coupled to the ICCD camera. Tomonitor evoked bioluminescence responses, cell cultures were main-tained in 1 ml Hepes–DMEM in 23 ml glass scintillation vials that allowedfor exchange of medium under sterile conditions. The vial was fitted witha rubber stopper containing a spinal syringe needle (19 gauge, 3.5 inch)reaching the bottom of the vial and a second needle fitted with a sterilefilter (0.2 µm). Tubing for medium exchange extended from the vial to theoutside of the incubator. Explant cultures were sealed tightly in glassscintillation vials or plastic cover dishes and maintained as static culturesin 0.5 ml Hepes–DMEM during the recording period to avoid any move-ment of the tissue. Recordings were made with a Turner luminometerinterfaced to a computer, using 30 sec integrations obtained everyminute from samples maintained at 22 ± 1°C. Bioluminescence was alsomeasured with photon-counting modules (HC-135, Hamamatsu)selected with dark counts below 20 counts/sec (CPS). The modules andcultures were enclosed in a light-tight chamber inside a temperature-con-trolled incubator at 25 ± 0.1°C and interfaced to IBM PC-type computersfor continuous data acquisition. Photon counts were integrated over1 min intervals. The cell culture luminescent signal was allowed to reacha stable level prior to the addition of treatments; this typically took approx-imately 2 days. Pharmacological treatments were prepared in stock solu-tions and mixed with Hepes–DMEM to the appropriate concentrationbefore being added to the cultures. PTZ (Sigma) was dissolved in water,and forskolin (Sigma) was premixed with dimethylsulfoxide (DMSO) togive a final DMSO concentration of 0.2% (v/v).

Data analysisThe last several days of bioluminescence recordings from SCN explantcultures were used in estimates of period. The earlier portion of therecord in which there was no obvious circadian rhythm by eye was notused. The data were detrended and FFT-NLLS was used to obtainperiod estimates. This statistical analysis has been described else-where [26] and consists of a Fourier analysis to provide initial guessesof component frequencies that are then used in the NLLS analysis toprovide a best fit of the data with a composite of cosine waves. Jointconfidence limits that are approximate, non-linear and asymmetric werecalculated for all model parameters at 95% confidence probability afterconvergence of the NLLS routine [27].

AcknowledgementsWe would like to thank Diana Berry for breeding and maintaining the trans-genic mouse colonies, Andrew Millar and Jeff Plautz for their help with biolumi-nescence imaging, and Margaret Kawecki for help with northern blot analysis.This work was supported by the National Science Foundation Center for Bio-logical Timing (DIR 8920162), the National Institute of Mental Health(MH51573) and the National Cancer Institute DHHS, under contract withABL. This work was presented in part at the Society for NeuroscienceMeeting, 1996 (abstract 22:151).

Research Paper Circadian rhythm in c-fos expression Geusz et al. 765

References1. Verma IM, Sassone-Corsi P: Proto-oncogene fos: complex but

versatile regulation. Cell 1987, 51:513-514.2. Morgan JI, Cohen DR, Hempstead JL, Curran T: Mapping patterns of

c-fos expression in the central nervous system after seizure.Science 1987, 237:192-197.

3. Moore RY, Eichler VB: Loss of a circadian adrenal corticosteronerhythm following suprachiasmatic lesions in the rat. Brain Res1972, 42:201-206.

4. Stephan FK, Zucker I: Circadian rhythms in drinking behavior andlocomotor activity of rats are eliminated by hypothalamic lesions.Proc Nat Acad Sci USA 1972, 69:1583-1586.

5. Ralph MR, Foster RG, Davis FC, Menaker M: Transplantedsuprachiasmatic nucleus determines circadian period. Science1990, 247:975-978.

6. Rea MA: Light increases Fos-related protein immunoreactivity inthe rat suprachiasmatic nuclei. Brain Res Bull 1989, 23:577-581.

7. Aronin N, Sagar SM, Sharp FR, Schwartz WJ: Light regulatesexpression of a Fos-related protein in rat suprachiasmatic nuclei.Proc Nat Acad Sci USA 1990, 87:5959-5962.

8. Rusak B, Robertson HA, Wisden W, Hunt SP: Light pulses that shiftrhythms induce gene expression in the suprachiasmatic nucleus.Science 1990, 248:1237-1240.

9. Earnest DJ, Iadarola M, Hermes HY, Olschowka JA: Photic regulationof c-fos expression in neural components governing theentrainment of circadian rhythms. Exp Neurol 1990, 109:353-361.

10. Kornhauser JM, Nelson D, Mayo K, Takahashi J: Photic and circadianregulation of c-fos induction in the hamster suprachiasmaticnucleus. Neuron 1990, 5:127-134.

11. Colwell CS, Foster RG: Photic regulation of Fos-likeimmunoreactivity in the suprachiasmatic nucleus of the mouse. JComp Neurol 1992, 324:135-142.

12. Wollnik F, Brysch W, Uhlmann E, Gillardon F, Bravo R, ZimmermannM, et al.: Block of c-Fos and JunB expression by antisenseoligonucleotides inhibits light-induced phase shifts of themammalian circadian clock. Eur J Neurosci 1995, 7:388-393.

13. Kornhauser JM, Nelson DE, Mayo KE, Takahashi JS: Regulation ofjun-B messenger RNA and AP-1 activity by light and a circadianclock. Science 1992, 255:1581-1584.

14. Honrado GI, Johnson RS, Golombek DA, Spiegelman BM,Papaioannou VE, Ralph MR: The circadian system of c-fos deficientmice. J Comp Physiol 1996, 178:563-570.

15. Rea MA, Michel AM, Lutton LM: Is fos expression necessary andsufficient to mediate light-induced phase advances of thesuprachiasmatic circadian oscillator? J Biol Rhythms 1993,8(suppl):59-64.

16. Smeyne RJ, Curran T, Morgan JI: Temporal and spatial expressionof a fos–lacZ transgene in the developing nervous system. MolBrain Res 1992, 16:158-162.

17. Smeyne RJ, Schilling K, Robertson L, Luk D, Oberdick J, Curran T,Morgan JI: Fos–lacZ transgenic mice: mapping sites of geneinduction in the central nervous system. Neuron 1992, 8:13-23.

18. Bronstein I, Fortin J, Stanley PE, Stewart GS, Kricka LJ:Chemiluminescent and bioluminescent reporter gene assays.Anal Biochem 1994 219:169-181.

19. Wood KV: Marker proteins for gene expression. Curr OpinBiotechnol 1995, 6:50-58.

20. Millar AJ, Straume M, Chory J, Chua NH, Kay SA: The regulation ofcircadian period by phototransduction pathways in Arabidopsis.Science 1995, 267:1163-1166.

21. Brandes C, Plautz JD, Stanewsky R, Jamison CF, Straume M, WoodKV, et al.: Novel features of Drosophila period transcriptionrevealed by real-time luciferase reporting. Neuron 1996, 16:687-692.

22. Mayerhofer R, Araki K, Szalay AA: Monitoring of spatial expressionof firefly luciferase in transformed zebrafish. J BioluminChemilumin 1995, 10:271-275.

23. Rutter GA, White MRH, Tavaré JM: Involvement of MAP kinase ininsulin signalling revealed by non-invasive imaging of luciferasegene expression in single living cells. Curr Biol 1995, 5:890-899.

24. Muller R, Bravo R, Burckhardt J, Curran T: Induction of c-fos geneand protein by growth factors precedes activation of c-myc.Nature 1984, 312:716-720.

25. Thompson JF, Hayes LS, Lloyd DB: Modulation of firefly luciferasestability and impact on studies of gene regulation. Gene 1993,103:171-177.

26. Plautz JD, Straume M, Stanewsky R, Jamison CF, Brandes C, Dowse

HB, et al.: Quantitative analysis of Drosophila period genetranscription in living animals. J Biol Rhythms 1997, 12:204-217.

27. Straume M, Frasier-Cadoret SG, Johnson ML: Least squaresanalysis of fluorescence data. In Topics in FluorescenceSpectroscopy, Vol. 2: Principles. Edited by Lakowicz JR. New York:Plenum Press; 1991:177-240.

28. Morgan JI, Curran T: Stimulus–transcription coupling in thenervous system: involvement of the inducible proto-oncogenesfos and jun. Annu Rev Neurosci 1991, 14:421-451.

29. Hisanaga K, Sagar AM, Sharp FR: c-fos induction in culturedcortical neurons and astrocytes via multiple signaling pathways.Prog Brain Res 1992, 94:189-195.

30. Peng YG, Taylor TB, Finch RE, Switzer RC, Ransdell JS:Neuroexcitatory and neurotoxic actions of the amnesic shellfishpoison, domoic acid. NeuroReport 1994, 5:981-985.

31. Cirelli C, Pompeiano M, Tononi G: Neuronal gene expression in thewaking state: a role for the locus coeruleus. Science 1996,274:1211-1215.

32. Greenberg ME, Thompson MA, Sheng M: Calcium regulation ofimmediate early gene transcription. J Physiol 1992, 86:99-108.

33. Treisman R: The serum response element. TIBS 1993, 17:423-426.34. Moore RY, Speh, JC: GABA is the principal neurotransmitter of the

circadian system. Neurosci Lett 1993, 150:112-116.35. Chambille I, Doyle S, Serviere J: Photic induction and circadian

expression of Fos-like protein. Immunohistochemical study in theretina and suprachiasmatic nuclei of hamster. Brain Res 1993,612:138-150.

36. Earnest DJ, Ouyang S, Olschowka JA: Rhythmic expression of Fos-related proteins within the rat suprachiasmatic nucleus duringconstant retinal illumination. Neurosc Lett 1992, 140:19-24.

37. Sutin EL, Kilduff TS: Circadian and light-induced expression ofimmediate early gene mRNAs in the rat suprachiasmatic nucleus.Brain Res Mol Brain Res 1991, 15:281-290.

38. Prosser RA, Macdonald ES, Heller HC: c-fos mRNA in thesuprachiasmatic nuclei in vitro shows a circadian rhythm andresponds to a serotonergic agonist. Mol Brain Res 1994, 25:151-156.

39. Welsh DK, Logothetis DE, Meister M, Reppert SM: Individualneurons dissociated from rat suprachiasmatic nucleus expressindependently phased circadian firing rhythms. Neuron 1995,14:697-706.

40. Herzog ED, Geusz ME, Khalsa SBS, Straume M, Block GD:Circadian rhythms in mouse suprachiasmatic nucleus explant onmulti-microelectrode plates. Brain Res 1997, 757:285-290.

41. Fisch TM, Prywes R, Roeder RG: c-fos sequence necessary forbasal expression and induction by epidermal growth factor, 12-O-tetradecanoyl phorbol-13-acetate, and the calcium ionophore. MolCell Biol 1987, 7:3490-3502.

42. Sherf BA, Wood KV: Firefly luciferase engineered for improvedgenetic reporting. Promega Notes 1994, 49:14-21.

43. Boshart M, Weber F, Jahn G, Dorsch-Hasler K, Fleckenstein B,Schaffner W: A very strong enhancer is located upstream of animmediate early gene of human cytomegalovirus. Cell 1985,41:521-530.

44. Brasier AR, Tate JE, Habener JF: Optimized use of the fireflyluciferase assay as a reporter gene in mammalian cell lines.Biotechniques 1989, 7:1116-1122.

766 Current Biology, Vol 7 No 10

Because Current Biology operates a ‘Continuous PublicationSystem’ for Research Papers, this paper has been publishedon the internet before being printed. The paper can beaccessed from http://biomednet.com/cbiology/cub — forfurther information, see the explanation on the contents page.