Circadian regulation of GABAAreceptor function by CKIε-CKIδin the rat suprachiasmatic nucleiKe Ning1,2, Lei Li1,2, Mingxia Liao1,2, Baosong Liu1,2,John G Mielke1, Yonghong Chen1,2, Ying Duan1,2,Youssef H El-Hayek1,2 & Qi Wan1,2

The type A GABA receptors are thought to mediatesynchronization of clock cell activity within thesuprachiasmatic nuclei (SCN). Here we report that caseinkinases I ε and δ (CKIε and CKIδ), the crucial clock regulators,form a complex with GABAA receptors and inhibit thereceptors’ function within the SCN according to a circadianrhythm. These results indicate that circadian variation of thekinase-receptor association may mediate regulation of GABAAreceptor function by CKIε-CKIδ in the SCN.

The master mammalian circadian clock in the SCN comprises multi-ple circadian clock cells1,2. γ-Aminobutyric acid (GABA), the principalneurotransmitter of the SCN3, can induce phase shifts and synchro-nization in SCN clock cells, and these effects are mediated by GABAAreceptors4. Because of the crucial involvement of the serine/threonineprotein kinase CKIε-CKIδ in the regulation of key clock proteins inthe transcriptional-translational feedback loops of the clockwork2,5–9

and the important role of protein kinases in the modulation of GABAAreceptor function10,11, we investigated the role of CKIε-CKIδ in mod-ulating GABAA receptor function in SCN clock cells.

Immunocytochemical staining showed that CKIε-CKIδ andGABAA receptors were mainly found colocalized in neurons located inthe dorsomedial region of the rat SCN (Fig. 1a–d). (The use of animalswas approved by the Animal Care Committee of University HealthNetwork.) We carried out coimmunoprecipitation assays12,13 to exam-ine whether CKIε-CKIδ physically associates with GABAA receptors,using microdissected SCN tissue collected at zeitgeber time (ZT) 2–4(see Supplementary Methods online) in a 12 h:12 h light-dark cycle.Immunoprecipitation with an antibody to CKIε or CKIδ resulted incoprecipitation of GABAA receptor β2/3 subunits that was significantlyhigher during the light phase (as assessed at ZT2–4) than during thedark phase (at ZT14–16) (Fig. 1e,f and Supplementary Fig. 1 online).In control studies, the initial abundance of solubilized proteins(inputs) or of directly immunoprecipitated proteins in parallel witheach coimmunoprecipitation did not change significantly in thecourse of the light-dark cycle (Fig. 1e,f). We also found that there wasno rhythmic variation for the CKIε–GABAA receptor association in

the hippocampus during the light-dark cycle and that CKIε did notform a complex with N-methyl-D-aspartate receptors in the SCN (seeSupplementary Fig. 2 online). These results are suggestive of a specificinteraction between CKIε-CKIδ complexes and GABAA receptors inthe SCN. Our study further demonstrated a circadian associationbetween CKIε-CKIδ and GABAA receptors in the SCN of rats exposedto constant darkness, with the peak association at circadian time (CT)6–7 (Fig. 1g,h and Supplementary Methods online), confirming thatthe CKIε-CKIδ–GABAA receptor interactions are circadian in nature.

1Division of Cellular & Molecular Biology, Toronto Western Research Institute, University Health Network, 399 Bathurst Street, Toronto, Ontario M5T 2S8, Canada.2Departments of Physiology and Surgery, University of Toronto, Toronto, Ontario M5S 1A8, Canada. Correspondence should be addressed to Q.W.(qi.wan@utoronto.ca).

Published online 18 April 2004; doi:10.1038/nn1236

B R I E F COM M U N I C AT I O N S

NATURE NEUROSCIENCE VOLUME 7 | NUMBER 5 | MAY 2004 489

Figure 1 Rhythmic variation of the physical association between GABAAreceptors and CKIε-CKIδ in the SCN. (a–d) GABAA receptors and CKIε-CKIδare colocalized in neurons located in the dorsomedial region of the SCN. III,third ventricle; OC, optic chiasm. d, higher magnification of box in c. Scalebars,100 µm for lower- and 30 µm for higher-magnification images. (e,f) CKIε and CKIδ form a complex with GABAA receptor β2/3 subunits duringthe light phase (ZT2–4); during the dark phase (ZT14–16), the kinase-receptor complex is significantly less abundant. Bar graphs summarize therhythmic variation of the kinase-receptor complexes (means ± s.d. of fourdeterminations; Student’s t-test, *P < 0.05). (g,h) Under constant darkness,the CKIε-CKIδ–GABAA receptor complexes are significantly less abundantduring CT18–19 than during CT6–7 and CT10–11. Line graphs summarizethe circadian variation of the kinase-receptor complexes (means ± s.d. of fourdeterminations; ANOVA, *P < 0.05).

©20

04 N

atur

e P

ublis

hing

Gro

up

http

://w

ww

.nat

ure.

com

/nat

uren

euro

scie

nce

B R I E F COM M U N I C AT I O N S

To determine whether the association between CKIε-CKIδ andGABAA receptors has functional consequences, we recorded GABAAreceptor–mediated whole-cell currents of neurons in rat SCN slicesat ZT2–4 (see Supplementary Methods online)13,14. GABA-inducedinward currents were blocked by the GABAA receptor antagonistbicuculline (20 µM; data not shown), confirming that they weregated through the GABAA receptor Cl– channel15. Bath applicationof the specific CKIε-CKIδ inhibitor IC261 increased peak currentamplitudes (1,076 ± 112 pA for IC261 treatment versus 669 ± 86 pAfor control; n = 18, Student’s t-test, P < 0.05; Fig. 2a), and the effectwas dose dependent (Fig. 2b). In addition, we found that IC261increased the slope of the current-voltage curve without affectingreversal potential (Fig. 2c). Together, these results suggest that for-mation of the kinase-receptor complex during the light phase maylead to the CKIε-CKIδ inhibition of GABAA receptor function inSCN neurons.

If an adequate association between CKIε-CKIδ and GABAA recep-tors were required for the modulation of GABAA receptors by CKIε-CKIδ during the early light phase (ZT2–4), the observed decrease inthe kinase-receptor association during the early dark phase(ZT14–16) would lead to the relief of GABAA receptor inhibition. Tovalidate this hypothesis, we recorded GABAA receptor–mediatedwhole-cell currents in SCN neurons at ZT14–16. As expected, wefound that the CKIε-CKIδ inhibitor IC261 did not exert significanteffects on GABAA receptor–mediated whole-cell currents (control,731 ± 79 pA; 1 µM IC261, 726 ± 63 pA; 2 µM IC261, 767 ± 88 pA;5 µM IC261, 792 ± 88 pA; n = 16, ANOVA, P > 0.05; Fig. 2d). Thus,rhythmic variation of the complex formed by CKIε-CKIδ andGABAA receptors may be crucial in the control of GABAA receptoractivity, with receptor function restricted during the light phase andrecovered during the dark phase (669 ± 86 pA of control GABAAreceptor currents for light phase versus 731 ± 79 pA for dark phase asdescribed above; Student’s t-test, P < 0.05).

We further examined the functional effects of the associationbetween CKIε-CKIδ and GABAA receptors at multiple time points

in the constant-darkness cycle by whole-cell patch-clamp recording.In agreement with the observed circadian variation of CKIε-CKIδ–GABAA receptor complexes in the SCN (Fig. 1g,h), the CKIε-CKIδ inhibitor IC261 (2 µM) significantly increased the peakGABAA receptor currents at CT6–7 and CT 10–11 (Fig. 2e), suggest-ing a circadian regulation of GABAA receptor function by CKIε-CKIδ in SCN neurons.

Our study provides evidence suggesting that the circadian interac-tion between GABAA receptors and CKIε-CKIδ in the SCN may represent an intracellular mechanism for the regulation of synchro-nization processes of clock cells.

Note: Supplementary information is available on the Nature Neuroscience website.

ACKNOWLEDGMENTSThis work was supported by a grant to Q.W. from Canadian Institutes of HealthResearch (CIHR). Q.W. is a CIHR New Investigator.

COMPETING INTERESTS STATEMENTThe authors declare that they have no competing financial interests.

Received 26 January; accepted 5 March 2004Published online at http://www.nature.com/natureneuroscience/

1. Welsh, D.K., Logothetis, D.E., Meister, M. & Reppert, S.M. Neuron 14, 697–706(1995).

2. Reppert, S.M. & Weaver, D.R. Nature 418, 935–941 (2002).3. Moore, R.Y. & Speh, J.C. Neurosci. Lett. 150, 112–116 (1993).4. Liu, C. & Reppert, S.M. Neuron 25, 123–128 (2000).5. Lowrey, P.L. et al. Science 288, 483–492 (2000).6. Lowrey, P.L. & Takahashi, J.S. Annu. Rev. Genet. 34, 533–562 (2000).7. Toh, K.L. et al. Science 291, 1040–1043 (2001).8. Eide, E.J., Vielhaber, E.L., Hinz, W.A. & Virshup, D.M. J. Biol. Chem. 277,

17248–17254 (2002).9. Eide, E.J. & Virshup, D.M. Chronobiol. Int. 18, 389–398 (2001).10. Swope, S.L., Moss, S.J., Raymond, L.A. & Huganir, R.L. Adv. Second Messenger

Phosphoprotein Res. 33, 49–78 (1999).11. Brandon, N., Jovanovic, J. & Moss, S. Pharmacol. Ther. 94, 113–122 (2002).12. Liu, F. et al. Nature 403, 274–280 (2000).13. Wan, Q. et al. J. Neurosci. 17, 5062–5069 (1997).14. Wan, Q. et al. Nat. Neurosci. 2, 401–403 (1999).15. Macdonald, R.L. & Olsen, R.W. Annu. Rev. Neurosci. 17, 569–602 (1994).

490 VOLUME 7 | NUMBER 5 | MAY 2004 NATURE NEUROSCIENCE

Figure 2 Rhythmic regulation of GABAA receptor function by CKIε-CKIδ inSCN neurons. (a) Response of GABAA receptor–mediated whole-cellcurrents to the CKIε-CKIδ inhibitor IC261 during light phase (ZT2–4). (b) Dose-dependent effects of IC261 on GABAA receptor currents duringlight phase (ZT2–4). Peak amplitudes of GABAA receptor currents arenormalized to the average of control recordings (1 µM IC261, 127 ± 21 pA;2 µM IC261, 168 ± 19 pA; 5 µM IC261, 179 ± 25 pA; ANOVA, n = 12, *P < 0.05). (c) A current-voltage (I-V) relationship curve shows that IC261(2 µM) potentiates GABAA receptor currents without changing the reversalpotential in SCN neurons at ZT2–4. (d) Response of GABAAreceptor–mediated whole-cell currents to IC261 (as in a) during dark phase(ZT14–16). (e) Under constant darkness, IC261 (2 µM) significantlyinhibits GABAA receptor currents during CT6–7 and CT10–11 (Student’s t-test, n = 10 for each time point, *P < 0.05). Notably, basal levels ofGABAA receptor currents (the control values) are significantly higher duringCT18–19 and CT22–23 (ANOVA, #P < 0.05).

©20

04 N

atur

e P

ublis

hing

Gro

up

http

://w

ww

.nat

ure.

com

/nat

uren

euro

scie

nce