The p42/44 Mitogen-activated Protein Kinase Pathway Couples Photic Input to Circadian Clock Entrainment*

In mammals, the suprachiasmatic nuclei (SCN) of the hypothalamus function as the major biological clock. SCN-dependent rhythms of physiology and behavior are regulated by changes in the environmental light cycle. Currently, the second messenger signaling events that couple photic input to clock entrainment have yet to be well characterized. Recent work has revealed that photic stimulation during the night triggers rapid activation of the p42/44 mitogen activated protein kinase (MAPK) pathway in the SCN. The MAPK signal transduction pathway is a potent regulator of numerous classes of transcription factors and has been shown to play a role in certain forms of neuronal plasticity. These observations led us to examine the role of the MAPK pathway in clock entrainment. Here we report that pharmacological disruption of light-induced MAPK pathway activation in the SCN uncouples photic input from clock entrainment, as assessed by locomotor activity phase. In the absence of photic stimulation, transient disruption of MAPK signaling in the SCN did not alter clock-timing properties. We also report that signaling via the Ca2+/calmodulin kinase pathway functions upstream of the MAPK pathway, coupling light to activation of the MAPK pathway. Together these results delineate key intracellular signaling events that underlie light-induced clock entrainment.

In mammals, the suprachiasmatic nuclei (SCN) of the hypothalamus function as the major biological clock. SCN-dependent rhythms of physiology and behavior are regulated by changes in the environmental light cycle. Currently, the second messenger signaling events that couple photic input to clock entrainment have yet to be well characterized. Recent work has revealed that photic stimulation during the night triggers rapid activation of the p42/44 mitogen activated protein kinase (MAPK) pathway in the SCN. The MAPK signal transduction pathway is a potent regulator of numerous classes of transcription factors and has been shown to play a role in certain forms of neuronal plasticity. These observations led us to examine the role of the MAPK pathway in clock entrainment. Here we report that pharmacological disruption of light-induced MAPK pathway activation in the SCN uncouples photic input from clock entrainment, as assessed by locomotor activity phase. In the absence of photic stimulation, transient disruption of MAPK signaling in the SCN did not alter clock-timing properties. We also report that signaling via the Ca 2؉ /calmodulin kinase pathway functions upstream of the MAPK pathway, coupling light to activation of the MAPK pathway. Together these results delineate key intracellular signaling events that underlie light-induced clock entrainment.
Within the suprachiasmatic nuclei (SCN) 1 resides an endogenous oscillator that functions as the master biological clock. The biological rhythm generated by the SCN regulates, with near 24-h periodicity, a wide array of cellular, physiological, and behavioral processes (1,2). Importantly, the SCN rhythm generator can be entrained by a number of external stimuli, of which light is the most potent. This ability to entrain the clock to photic cues allows animals to adjust their biological rhythms to changes in the external environment.
Recent work has revealed that photic stimulation affects clock timing in part by triggering rapid transcriptional activation in the SCN (3). In turn, these newly transcribed gene products are postulated to regulate the clock by resetting a transcription/translation feedback loop that generates the rhythm. Although many postsynaptic events including glutamate receptor activation and increased intracellular calcium have been shown to couple light to entrainment of the circadian clock (4 -6), there is still significant debate regarding the second messenger signaling events that trigger rapid transcriptional activation in the SCN. Along these lines, several reports have shown that the infusion of broad spectrum calcium/calmodulin kinase (CaMK) inhibitors into the SCN attenuates light-induced phase shifting of the circadian clock (7-10). Likewise, a role for nitric oxide and protein kinase G has been implicated in light-induced resetting of the clock (11)(12)(13)(14)(15). More recently, attention has turned to the p42/44 mitogen protein kinase (MAPK) pathway as a potential signaling intermediate coupling light to clock entrainment.
Interest in this pathway results, in part, from recent studies showing that brief exposure to light during the subjective night, but not during the subjective day, triggers rapid and robust activation of the MAPK pathway in the SCN (16). Furthermore, light-induced activation of ERK colocalizes with and regulates the activation state of the cAMP response elementbinding protein (16), a transcription factor proposed to be a key intermediate coupling photic stimulation to the rapid expression of clock genes (17). Additionally, in NIH-3T3 fibroblasts disruption of TPA (12-O-tetradecanoylphorbol-13-acetate)-induced activation of the MAPK pathway blocks circadian gene oscillations (18). Thus, the MAPK pathway fulfills several criteria likely to be essential for light entrainment of the clock: it is rapidly activated by light, its activation is restricted to the subjective night, and the pathway couples to transcription factor activation. These observations raise the possibility that the MAPK pathway plays a central role in the set of signaling events that couple photic input to clock entrainment. Here we report that pharmacological disruption of light-induced MAPK pathway activation in the SCN uncouples photic input from clock entrainment. The observations presented here reveal that the MAPK cascade functions as an SCN clock input pathway.
spinal fluid carries the infusate to the third ventricle, the location of the SCN. In a second group of animals the coordinates (posterior, 0.3 mm from bregma; lateral, 0.0 mm from the midline; and dorsoventral, Ϫ4.75 mm from dura) were used to place cannula tips in the third ventricle. Cannulae were held in place with dental cement. A 30-gauge stylus was secured in the cannula to ensure patentcy. After surgery animals were housed individually and allowed to recover for 2 weeks. A stainless steel injector needle (30 gauge) extending 500 m from the tip of the guide cannulae was used to infuse either vehicle (Me 2 SO) or U0126 (10 nm/l, Cell Signaling Research) at a rate of 0.40 l/min. A total volume of 3 l was delivered. Because of its photo-lability, care was taken to ensure minimal exposure of U0126 to light. The efficacy of another MEK inhibitor, SL-327 (50 g/l), was also examined. SL-327 was dissolved in Me 2 SO and infused as described above. SL327 was kindly provided FIG. 1. Light induces MAPK pathway activation. A, cannulated mice were infused with the MEK inhibitor U0126 (10 nmol/l, A3) or with drug vehicle (A1, A2). Relative to control mice (no light, A1), light (50 lux, 10 min; ZT 15) triggered robust ERK activation (A2). Infusion of U0126 45 min before photic stimulation (A3) depressed light-induced ERK activation. B, similar results were observed by Western analysis of excised SCN. Light-induced ERK activation (pERK) was blocked by infusion of U0126. The blot was then stripped and probed for total ERK expression. To assess ERK activation, the erk-2 phosphorylation level was divided by the total erk-2 level for each lane. These data are expressed as -fold stimulation relative to the erk-2 activation level under control conditions (no light, no U0126), which was set equal to one. C, robust light-induced ERK activation is specific to the SCN; modest induction was observed in tissue isolated from the lateral hypothalamus (lateral hypo) and piriform cortex (CTX). Membranes were also probed for total ERK expression. To quantify ERK activation levels, the erk-2 phosphorylation level was divided by the erk-2 level for each lane and expressed as -fold stimulation relative to the control (no light) ERK activation level for the corresponding brain region, which was set equal to one. Each experiment was repeated a minimum of three times. 3v, third ventricle; oc, optic chiasm. by Dr. James Trzaskos (DuPont Merck Pharmaceutical Co., Wilmington, DE). Mice were restrained by hand during insertion of the injector needle and allowed to roam freely during infusion. The infusion needle was maintained in the guide cannula for an additional 30 s after the infusion stopped. Infusions were performed under dim light (15 watt, Ͻ1 lux at cage level) using a red safelight (Kodak filter, series 2). Because the results obtained using lateral and third ventricle infusions were not statistically different, data from the two groups were combined. Initially another MEK inhibitor, PD 98059, was to be used in this study. However, because of its exceedingly limited solubility, we were not able to utilize a ventricular infusion technique to effectively deliver it to the SCN. KN 62 (10 nm/l, Biomol) was diluted in Me 2 SO and infused using the techniques described above.
Circadian Activity Protocol-Cannulated mice were individually housed and entrained to a 12-h:12-h light/dark (LD) cycle for 14 days before being transferred to dark/dark (DD). Luminescence was provided by fluorescent white light (ϳ100 lux at mid-cage level). During the DD period animals were under conditions of total darkness, except for a weekly food and water replenishment and a change of bedding. Cage maintenance occurred at varying times during the subjective night. During cage maintenance, mice were exposed to a dim red light. Circadian physiology was monitored by recording locomotor activity via a 15-cm diameter running wheel. Closures of a microswitch attached to the wheel were recorded automatically to a PC running Vital View (Minimitter Corp., Bend, OR) data acquisition software.

FIG. 2. MEK inhibition in vivo.
To track the diffusion pattern of U0126 from the ventricle into the brain, mice were infused with the MEK inhibitor (10 nmol/l) and then, 45 min later, intraperitoneally injected with kainate (30 mg/kg). Mice also were exposed to light (50 lux, 10 min; ZT 16). By infusing the MEK inhibitor and then stimulating MAPK pathway activation, we were able to track U0126 diffusion. Using immunohistochemical analysis of ERK activation, we found that U0126 did not diffuse broadly from the infusion site (B, dashed line denotes the approximate area of MAPK pathway inhibition; the arrow identifies the location of guide cannula/infusion needle tip). The vehicleinfused control animal (A, intraperitoneally injection with kainate, light exposure) exhibited high levels of ERK activation in periventricular regions including the SCN. Bar, 500 m.

FIG. 3. Disruption of light-induced MAPK pathway activation uncouples light from circadian clock entrainment.
A, representative double-plotted actogram from a mouse initially entrained to a standard LD cycle was transferred to DD and then infused with drug vehicle (Me 2 SO) 45 min before light (50 lux, 10 min) exposure at CT 15 (asterisk). 11 days later the same animal was infused with U0126 (10 nmol/l) and exposed to light (asterisk). B, in a second representative trace the order of drug/vehicle infusions was reversed, but the effect was the same as in A. C, acute disruption of the MAPK pathway during the subjective night (asterisk) did not phase shift the clock. D, mean Ϯ S.E. of the phase-delaying effects of light at CT 15. Numbers above bars denote sample sizes for each condition. **, p Ͻ 0.0001; two-tailed Student's t test.
Mice were organized into four groups: U0126-infused, no light (control); vehicle-infused, no light (control); U0126-infused, light-treated; and vehicle-infused, light-treated. For the two light-treated groups, 45 min after infusions animals were exposed to white light (50 lux) for 10 min and then returned to darkness. Mice were exposed to light during early subjective night, circadian time 15. Control groups not exposed to light were handled in a manner similar to the light-treated mice. Each group was rotated through at least two of the four stimulus paradigms. Mice were maintained under DD throughout the assay period; a minimum of 9 days separated each infusion. The Ohio State University Animal Welfare Committee approved all animal handling and experimental procedures.
Tissue Collection-To examine MAPK pathway activation, mice were sacrificed via cervical dislocation, and their brains were removed rapidly under red light. Brains were immersed in chilled, oxygenated physiological saline and then cut into 500-m coronal sections with a vibratome. Tissue utilized for immunohistochemical examination was fixed in a 5% formaldehyde/phosphate-buffered saline (PBS) solution for 4 -6 h at room temperature, cryoprotected with 30% sucrose for 12 h, and finally cut into thin (40 m) sections with a freezing microtome. Tissue sections destined for Western analysis were frozen immediately on dry ice, and the brain regions of interest were isolated and stored at Ϫ80°C.
Immunohistochemistry-To determine the ERK activation level, free floating sections were initially blocked with 10% bovine serum albumin/1% goat serum in PBS containing 0.1% Triton X-100 (PBST) and 1 mM NaF. Brain sections were immunolabeled (overnight at 4°C) with an affinity-purified rabbit polyclonal antibody (1:500 final dilution, Cell Signaling Research) that detects the dually phosphorylated form of ERK. The tissue was washed six times (5 min/wash) with PBST and then was incubated with an Alexa-488-conjugated secondary antibody directed against rabbit IgG (4 g/ml final concentration, Molecular Probes). ERK activation levels were also examined using the HRP-ABC technique (Vector Laboratories), following the instructions of the manufacturer. Nickel-intensified diaminobenzidine was used to visualize the signal. Images were captured using a 16-bit digital camera, and data were quantified using Metamorph image analysis software (Universal Imaging). To quantitate fluorescent signal intensity, a coronalcentral SCN image was captured with a 10ϫ objective, and a 140 (x-axis) ϫ 180 (y-axis)-pixel oval was placed over the digitized SCN, and intensity was measured. For each section, the SCN signal intensity was normalized by subtracting the immunofluorescent signal from the hypothalamic area immediately lateral to the SCN. Data from three consecutive SCN sections were averaged for each animal analyzed.
Western Blotting-Isolated brain regions were resuspended and sonicated (15 s) in 50 l of protease inhibitor buffer (50 mM ␤-glycerophosphate, 1.5 mM EGTA, 0.1 mM Na 3 VO 4 , 1 mM dithiothreitol, 10 g/ml aprotinin, 2 g/ml pepstatin, 10 g/ml leupeptin, and 1 mM phenylmethylsulfonyl fluoride, pH 7.4). 50 l of 6ϫ sample buffer then was added, and the samples were heated to 90°C for 10 min. 30 l of extract was electrophoresed through a 10% SDS-PAGE gel. Following transfer to polyvinylidene difluoride membranes (Immobilon-P, Millipore) and blocking with 10% powdered milk, samples were incubated (4°C, overnight) in PBST with primary mouse monoclonal antibody against pERK (1:10,000 final dilution; Sigma) followed by goat anti-mouse IgG alkaline phosphatase-conjugated secondary antibody (1:2,000, PerkinElmer Life Sciences). Immunoreactivity was developed using the Western-star alkaline phosphatase detection system (Tropix). Membranes then were probed for total ERK expression using a goat anti-ERK polyclonal antibody (1:1,000 final dilution, Santa Cruz Biotechnology) followed by a donkey anti-goat IgG antibody conjugated to horseradish peroxidase. The signal was visualized using Renaissance chemiluminescent horseradish peroxidase substrate (PerkinElmer Life Sciences). Between each antibody treatment, membranes were washed a minimum of six times (5 min/wash) in PBST with 5% milk. Scion Image analysis software was used to quantitate band intensity. Band intensity for phosphorylated erk-2 was normalized to total erk-2 levels for the corresponding lane. Each experiment was repeated a minimum of three times.
Assessment of Light-induced Phase Shifts-The linear regression method described by Daan and Pittendrigh (19) was used to assess light-induced phase shifts. Specifically, the difference in activity onset before and after the day of light exposure was determined by a least squares method, in which a line calculating the activity onset for a period of at least 6 days preceding light treatment was determined. This line was extended to project when activity onset should occur during the period after light exposure. A second regression line was generated to determine activity onset after light administration. Days 3-8 following light treatment were used to generate this line. The difference in the projected versus the actual activity onset after light treatment was the phase shift. Group data are expressed as mean phase shift Ϯ S.E. Significance was assessed using the two-tailed Student's t test.

RESULTS
To confirm that photic stimulation elicits MAPK pathway activation in the SCN, animals were exposed to light (10 min, 50 lux) during the early night (3 h after lights off; zeitgeber time 15, ZT 15) and then immediately sacrificed. SCN tissue A representative double-plotted actogram from a dark-adapted mouse initially infused with drug vehicle (Me 2 SO) 45 min before light (50 lux, 10 min) exposure at CT 15 (asterisk). 13 days later the animal was infused with SL-327 (50 g/l) and exposed to light (asterisk). In the absence of photic stimulation, the infusion of SL-327 (asterisk) did not dramatically alter clock phase. Please refer to the legend to Fig. 3D for quantification of the effects of SL-327. was processed using immunofluorescent staining techniques for the activated (i.e. dually phosphorylated, Thr 202 /Tyr 204 ) form of erk-1 and erk-2 (here, collectively referred to as P-ERK). This dual phosphorylation event is essential for ERK enzymatic activity and is a marker for MAPK pathway activation. Relative to control animals not exposed to light, photic stimulation triggered robust activation of the MAPK cascade in the SCN (Fig. 1A). P-ERK was observed from the ventral to dorsal regions of the SCN. Western blotting was also used to examine the activation of ERK (Fig. 1, B and C). Light triggered the expression of the catalytically active form of ERK in the SCN; modest levels of activation were observed in tissue isolated from the lateral hypothalamus and piriform cortex.
Collectively, these data reveal that photic stimulation triggers robust SCN-specific activation of the MAPK pathway.
To disrupt MAPK signaling in the SCN we employed a ventricular infusion technique. Mice were infused with the specific MEK inhibitor U0126 (10 nmol/l) (20) 45 min prior to photic stimulation. To determine whether U0126 administration disrupted MAPK pathway activation infused animals were light flashed and sacrificed, and SCN-containing brain sections were processed for expression of activated ERK (Fig. 1, A and B). Infusion of the MEK inhibitor significantly attenuated (typically Ͼ90%) light-induced ERK activation in the SCN. Similar results were obtained with a second MEK inhibitor, SL-327 (data not shown).
For the behavioral studies described below, it was imperative to determine the diffusion pattern of U0126 in the brain. Thus, to track its penetration into periventricular brain regions, mice were infused with the MEK inhibitor and then, 45 min later, intraperitoneally injected with kainate (to trigger MAPK pathway activation throughout the brain) and exposed to light (15 min, 50 lux; ZT 16). A region lacking ERK activation centered around the infusion needle delineates the area of U0126 diffusion and inhibition. Using this technique, we found that U0126 did not broadly diffuse (Fig. 2B) and that inhibition was limited to within ϳ400 m of the ventricle. Importantly, relative to the control animals ( Fig. 2A), U0126 infusion blocked ERK activation within the SCN. U0126-mediated inhibition of MAPK pathway activation was transient, lasting for ϳ90 min (data not shown). These data reveal that U0126 infusion is localized to the periventricular brain regions. Thus, the area of inhibition is not so great as to raise considerable doubt about the site of action.
Wheel-running activity was used as a circadian output to address the contribution of the MAPK pathway to light entrainment of the clock. Animals were initially entrained to an LD cycle for 14 days and then released into total darkness (DD). In the absence of external timing cues, overt circadian rhythms are governed by the SCN clock. After 9 -12 days in DD animals were infused with either drug vehicle or U0126 at CT 14.25 and exposed to light 45 min later (50 lux, 10 min). A marked phase delay in the onset of wheel-running activity was elicited by light exposure after vehicle infusion (Fig. 3, A, B,  and D). In striking contrast, U0126 infusion blunted this lightinduced phase shift, indicating that disruption of signaling via the MAPK pathway uncouples light from clock entrainment. U0126 infusion reduced the light-induced phase shift by Ͼ90% in 60% of mice (n ϭ 20). In the absence of photic stimulation, the acute disruption of MAPK signaling did not influence clock timing (Fig. 3, C and D). In a preliminary round of experimentation (data not shown) we found that the infusion of lower concentrations (Ͻ2 nmol/l) of U0126 that partially blocked (40 -60%) light-induced ERK activation did not dramatically alter the phase shifting effects of light.
To corroborate the effects of U0126, we tested the efficacy of another MEK inhibitor, SL-327. Similar to the experiments described above, infusion of SL-327 (50 g/l) significantly attenuated light-induced phase shifting of the clock (Figs. 3D and 4). As with U0126, the infusion of SL-327 in the absence of photic stimulation did not significantly alter clock phase. Collectively these data identify a role for the MAPK pathway as a signaling intermediate coupling light to clock entrainment.
Several studies have reported that CaMK signaling couples light to clock entrainment (7-10), raising the possibility that the MAPK and CaMK pathways work in parallel to affect downstream targets. Another possibility is that these pathways are also in series, with the CaMK pathway triggering MAPK cascade activation (21-23). To assess whether CaMKs trigger MAPK activation in the SCN, mice were infused with the broad acting CaMK inhibitor KN 62 at ZT 14.25. 45 min after infusion, mice were exposed to light and sacrificed, and SCN-containing sections were processed for ERK activation. Relative to vehicle-infused animals, the infusion of the CaMK inhibitor significantly attenuated light-induced ERK activation (Fig. 5). In the absence of light, the low basal expression of activated ERK was not significantly affected by CaMK inhibition. Together, these data suggest that CaMK signaling functions as an upstream regulator of the MAPK pathway in the SCN.

DISCUSSION
Signaling via the MAPK pathway in the central nervous system has recently become a topic of intense interests. For example, a role for the MAPK pathway has been identified in such diverse central nervous system processes as addiction, programmed cell death, and learning and memory (24 -26). Along these lines, stimulus protocols that trigger hippocampal long-term potentiation also activate the MAPK pathway (27). More importantly, disruption of signaling via the MAPK cascade uncouples stimulation from transcriptionally dependent forms of long-term potentiation formation, indicating a causal link between the MAPK pathway and stimulus-dependent strengthening of synaptic efficacy (28,29) Perhaps the most dramatic example of MAPK pathway activation in the central nervous system is observed in the SCN following photic stimulation (16). Interest in this pathway is derived in part from its rapid activation following photic stimulation, its phase-dependent activation, and its ability to potently regulate the activation state of numerous classes of transcription factors. These observations led us to examine the contribution of signaling via the MAPK pathway to light-induced clock entrainment. Here we show that disruption of light-induced MAPK pathway activation in the SCN blocks photic input from clock entrainment.
In a large percentage of animals, infusion of U0126 led to a dramatic (Ͼ90%) reduction in light-induced phase shifting, suggesting that the MAPK pathway plays a central role in the light-entrainment process. However, a second group of animals exhibited varying amounts of inhibition (20 -90%) of light-induced phase shifting of the circadian clock. One interpretation of this result is that infusion did not lead to effective MAPK pathway inhibition. Indeed, we have observed that partial inhibition (40 -60%) of light-induced ERK activation did not dramatically affect the ability of light to entrain the clock. Given that the MAPK pathway signal is amplified both spatially and temporally downstream of MEK (the target of U0126 and SL-327), partial MEK inhibition may have only marginal effects on cellular targets of the MAPK pathway. This difference in the efficacy of inhibition may explain the discrepancy between our work and that of Yokota et al. (7).
It should be noted that U0126 is a recently developed inhibitor that specifically antagonizes the activity of both MEK 1 and MEK 2 without detectable effects on other kinases such as PKA, PKC, Raf, ERK, MSK I, and JNK (20,30). Likewise, the U0126 analog SL-327 is specific to MEK 1 and MEK 2 and does not have measurable effects on PKA, PKC, and CaMK II (20,31). Furthermore, based on a number of studies that have utilized an infusion protocol to disrupt MAPK signaling in the central nervous system, it appears that the MAPK pathway specifically affects transcriptionally dependent forms of neuronal plasticity and does not alter such fundamental neurophysiological processes as sensory, motivational, or motor abilities (32,33). Even the intraperitoneal injection of SL327 did not affect sensory perception or arousal (34). Importantly, no long-term residual effects of MAPK pathway disruption were observed in any of these studies.
The activation state of the MAPK pathway varies as a function of circadian time in the SCN, showing maximal stimulation during the subjective day and low levels of activity during the early to mid-subjective night (16). This observation raised the possibility that the infusion of U0126 may alter clock phase and therefore preclude us from determining the sight of drug action. However, we found that the acute disruption of MAPK activation did not alter circadian clock timing, indicating that the effects on light-induced phase shifts resulted from inhibition of the photic input signal. The absence of an effect of U0126 infusion on clock phase might result from the transient nature of U0126 inhibition and the relatively low basal level of MAPK pathway activation during the time of the infusion (CT 14.25).
A number of studies have implicated a variety of kinase pathways in the light entrainment process. For example, disruption of nitric oxide synthetase activity blocks phase shifting (14). Likewise, a role for the CaMKs has also been identified (7)(8)(9)(10). Interestingly, in several model systems, signaling via nitric oxide and CaMKs have been shown to function as upstream regulators of the MAPK pathway (22,35). With respect to CaMK signaling, our data support this model. We observed that infusion of the CaMK inhibitor KN 62 attenuated lightinduced ERK activation. Similar results have been found in hippocampal neurons, where depolarization-induced ERK activation is attenuated by disruption of CaMK activation (21). One possible mechanism by which CaMK signaling leads to MAPK pathway activation involves CaMK II inhibition of the GTPase-activating protein p135 SynGAP (36). It is important to note that we are not ruling out other mechanisms by which CaMK signaling affects the clock. Indeed, it is probable that CaMK IV plays a significant role in coupling light to transcription factor activation in the SCN. Thus, there are likely to be a number of essential signaling pathways that function in an orchestrated manner to couple light to clock entrainment.
Signaling via the MAPK pathway has become an area of investigation for circadian biologists working in both invertebrate and vertebrate model systems (37)(38)(39)(40). In particular, MAPK signaling has been shown to function as an output pathway from the clock (37) and may play a role in the core clock timing mechanism (41). The observations described here are unique in that they reveal a role for the MAPK cascade as a signaling intermediate coupling photic input to circadian clock entrainment.