[1] 12w Forebrain-specific constitutively active CaMKKa transgenic mice show deficits in hippocampus-dependent long-term memory

[1] 170w The Ca 2+ /calmodulin (CaM) kinase cascade is activated by Ca 2+ influx through the voltage-dependent Ca 2+ channels and the NMDA receptor. CaM kinase kinase (CaMKK), the most upstream kinase of the CaM kinase cascade, phosphorylates and activates both CaM kinase I (CaMKI) and CaMKIV, resulting in activation of cyclic AMP-responsive element binding protein (CREB)-dependent gene transcription. Using transgenic techniques, we created mutant mice in which a constitutively active form of CaMKK1, the autoinhibitory domain truncated protein, is over-expressed specifically in the forebrain. In these mice, although performance was normal in basal activity and short-term memory, specific impairments were shown in hippocampus-dependent long-term memory after training in spatial memory tasks and after contextual fear conditioning. In cultured neurons of these mice, phosphorylation of CaMKI was significantly increased in basal states, whereas the activity range of CaMKI phosphorylation by brain-derived neurotrophic factor (BDNF) and KCl stimulation was significantly diminished in mutant mice. Our results define a critical role for CaMKKa in synaptic plasticity and the retention of hippocampus-dependent longterm memory.

[1] 230w In neurons, the Ca 2+ /calmodulin kinase (CaMK) cascade transduces Ca 2+ signaling into gene transcription by activating the transcription factor cyclic AMP-responsive element binding protein (CREB) (Corcoran & Means, 2001). CaM kinase kinase (CaMKK) is an upstream kinase of the CaM kinase cascade that specifically phosphorylates and activates CaM kinase I (CaMKI) and CaMKIV (Soderling, 1999). Other CaMKK substrates have been found in previous studies, such as Protein kinase B (PKB) and 5 0 adenosine monophosphateactivated protein kinase (AMPK) (Anderson et al., 2008;Hawley et al., 1995;Yano, Tokumitsu, & Soderling, 1998). PKB protects cells from apoptosis via BAD phosphorylation. AMPK contributes to the regulation of energy balance and/or regulates autophagy and proteasome function (Høyer-Hansen et al., 2007;Viana et al., 2008). The CaM kinase cascade also plays an important role in neuronal development (Ribar et al., 2000;Wayman et al., 2004Wayman et al., , 2006)), as well as long-term memory and synaptic plasticity (Kang et al., 2001;Liu, Lyons, Mamounas, & Thompson, 2004;Mizuno, Ris, Sánchez-Capelo, Godaux, & Giese, 2006;Mizuno et al., 2007;Peters et al., 2003, andBlaeser et al., 2006). Most CaMKI isoforms have nuclear export signals and are excluded from the nucleus (Stedman, Uboha, Stedman, Nairn, & Picciotto, 2004), but these isoforms can indirectly activate CREB-dependent transcription (Kimura et al., 2002;Wayman et al., 2006). CaMKIV is predominantly a nuclear enzyme, and phosphorylates and activates CREB (Bito, Deisseroth, & Tsien, 1996;Matthews et al., 1994).

[2] 94w In mammals, CRE-dependent transcription is required for longterm memory (Barco, Pittenger, & Kandel, 2003). There is also increased phosphorylation of CREB following stimuli that produce memory-related long-term potentiation (LTP) (Bito et al., 1996) as well as after training on hippocampus-dependent tasks (Taubenfeld, Wiig, Bear, & Alberini, 1999). In a genetic model, Barco showed that restricted and regulated expression in mice of VP16-CREB, a constitutively active form of CREB, in hippocampal CA1 neurons lowers the threshold for eliciting a persistent late phase of LTP (L-LTP) in the Schaffer collateral pathway (Barco, Alarcon, & Kandel, 2002).

[3] 77w The downstream of the CaM kinase cascade, including CREB and AMPK, is involved in neurodegenerative and neuropsychiatric disorders, such as Huntington's disease and Rubinstein-Taybi Syndrome (Petrij et al., 1995). Activation of the CaM kinase cascade has the potential to improve learning and memory under normal conditions or in these disorders. Additionally, CaM kinase cascade activation leads to stress resistance in skeletal muscles via induced expression of peroxisome proliferator-activated receptor gamma coactivator 1 (PGC-1) (Wu et al., 2002).

[4] 127w In the present study, to create activated states of the CaM kinase cascade, we generated transgenic (tg) mice that express a constitutively active form of the CaMKK (CaMKKc) that has its autoinhibitory domain truncated. CaMKKc is expressed from the Ca 2+ /calmodulindependent protein kinase IIa (CaMKIIa) promoter, which drives expression postnatally in neurons within the hippocampus, striatum, amygdala, and cortex (Kojima et al., 1997;Mayford et al., 1996). Unexpectedly, CaMKKc tg mice exhibited impairments in hippocampus-dependent long-term memory after training in spatial memory tasks and after contextual fear conditioning, and had deficits in hippocampal synaptic plasticity. Our analysis of CaMKKc tg mice suggests that the CaM kinase cascade has both positive and negative roles in synaptic plasticity and memory storage, and provides a paradoxical theory for memory storage.

[1] 60w We generated transgenic mice that expressed a truncated form of CaMKK at a.a. 433, a CaMKKc. The transgene construction is shown in Fig. 1A. CaMKK, like other CaM kinases, contains autoinhibitory and CaM-binding domains following its catalytic domain (Tokumitsu, Enslen, & Soderling, 1995). CaMKKc, following removal of autoinhibitory and CaM-binding domains of CaMKK, is constitutively active (Matsushita & Nairn, 1998).

[2] 86w CaMKKc was EGFP tagged and expressed from the CaMKIIa promoter to restrict transgene expression spatially to the forebrain (Mayford et al., 1996). Three independent tg mouse lines expressing CaMKKc were generated. The three lines had low (lines 1) or medium (line 6, 10) levels of transgene expression as determined by northern blot. Phosphorylation of CaMKI was analyzed in all three lines, and line 6 exhibited the highest CaMKI phosphorylation level in the cortical neurons (data not shown), so we focused on line 6 for further analysis.

[3] 88w In Fig. 1B, northern blot analysis using a mouse CaMKK1 cDNA probe shows that the CaMKKc tg transcript (lower band) is expressed in the cortex and hippocampus, but there is little expression in the cerebellum in CaMKKc tg mice. In the hippocampus and cortex, the transcripts are expressed at similar levels to the endogenous CaMKK1 transcript (upper band). Expression of the transcripts is restricted to the forebrain, and no signals were detected in the heart, liver, kidney, skeletal muscle, or testis using an EGFP cDNA probe (Fig. 1C).

[4] 157w CaMKKc tg in mutant mice was designed to be active constitutively, and it was expected that increased phosphorylation of CaMKK substrates, such as CaMKI and CaMKIV, would be observed in basal conditions. To examine the effect of CaMKKc tg on CaMKI phosphorylation, we prepared cortical neurons from wild type (wt) or tg embryos, and analyzed the levels of phosphorylated CaMKI in basal conditions, or in response to brain-derived neurotrophic factor (BDNF) and KCl stimulation, by immunoblotting (Fig. 2A). CaM-KKc tg mouse neurons showed increased phosphorylation of CaMKIa in basal conditions (p < 0.01). BDNF stimulation significantly induced phosphorylation of CaMKIa in wt neurons compared to the control (p < 0.001). In contrast, in tg neurons, BDNF slightly induced phosphorylation, but this was not significant compared to the control (p = 0.380). Similarly, KCl stimulation significantly induced phosphorylation of CaMKIa in wt neurons (p < 0.001), but this induction was not significant in tg neurons (p = 0.058).

[5] 112w It has been reported that the CaM kinase cascade is also involved in stress resistance via induced expression of peroxisome proliferator-activated receptor gamma coactivator 1a (PGC-1a) (Wu et al., 2002). It has been shown that increased PGC-1a levels protect cultured neural cells from oxidative stressor-mediated death (St-Pierre et al., 2006). Oxidative stress contributes to aging, which is the greatest risk factor for neurodegenerative diseases such as Alzheimer's and Parkinson's diseases (Lin & Beal, 2006). Thus, we tested whether neuronal cells from CaMKKc tg mice showed stress resistance (Fig. 2B). Cortical neurons from CaMKKc tg mice were significantly protected against death caused by H 2 O 2 compared to neurons from wt mice.

[6] 47w Behavioral analysis was conducted as a behavioral test battery for mutant mice (Miyakawa et al., 2003;Takao & Miyakawa, 2006), as summarized in Supplementary Table 1. CaMKKc tg mice (n = 20) and their littermates (n = 20) were generated by mating hemizygous CaMKKc tg and wt mice.

[7] 87w Upon gross inspection, CaMKKc tg mice had normal fur, whiskers, and posture, and were indistinguishable from wt mice. Body weight and body temperature were similar to wt mice, and there were no significant differences in performance in grip-strength and wirehang tests (Supplementary Fig. S1A-D). In the rotarod and beam tests of motor coordination, performances were similar in both CaM-KKc tg and wt mice (Supplementary Fig. S1E-G). The raw data of behavioral tests, which are not described in this paper, are disclosed in the gene-brain-phenotyping database (http://www.mouse-phe- notype.org/).

[8] 106w In the open-field tests, there were no significant differences in total distance traveled, vertical activity, and stereotypic counts between wt and CaMKKc tg mice. The center time was significantly decreased in CaMKKc tg compared to wt mice (Fig. 3C; p < 0.05). Generally, CaMKKc tg mice tended to prefer the side of the field. We also performed a light-dark transition test (Supplementary Fig. S2A and B), elevated plus-maze test (Supplementary Fig. S2C-F), social interaction test in a novel environment (Supplementary Fig. S2G-K), and a Porsolt forced swim test (Supplementary Fig. S2L and M), none of which showed any significant differences between wt and CaMKKc tg mice.

[9] 31w In the hot-plate test (Supplementary Fig. S3A) to observe nociception, and the prepulse inhibition/startle response test (Supplementary Fig. S3B and C), there were no differences between wt and CaMKKc tg mice.

[10] 123w Next, we assessed spatial learning and memory in CaMKKc tg mice. In the Barnes circular maze, there were no differences in latency to the hidden platform or number of errors throughout all trials between CaMKKc tg and wt mice (Fig. 4A and B). Both CaMKKc tg and wt mice were able to remember the target 24 h after the final trial (Fig. 4C). However, we assessed the maintenance of memory 2 weeks later and found that there was a significant difference between the two groups (Fig. 4D; p < 0.05). Wt mice still remembered the correct target well, whereas CaMKKc tg mice did not. These results indicate that in CaMKKc tg mice, the expression of constitutively active CaMKKa impairs retention of spatial memory.

[11] 199w Contextual and cued fear conditioning are two forms of associative learning that induce effective memory for either the context or the cue after a single training session (LeDoux, 2000). Memory for the conditioned stimulus was measured as freezing, an absence of visible movement, when presented with the conditioned context or cue in 24 h and in a 14-day retention test. CaMKKc tg and wt mice showed a similar level of freezing before and after the shock on the training day (Fig. 5A). When re-exposed to the conditioned context 24 h later, CaMKKc tg mice showed a slight decrease in freezing compared to the wt mice (Fig. 5B). To assess long-term memory for contextual fear, the mice were tested in the conditioned context 14 days after conditioning. CaMKKc tg mice showed a significant decrease in freezing 14 days after conditioning compared to wt mice (Fig. 5C; p < 0.01). In the cued fear conditioning experiment, no differences were observed in the 24-h and 14-day retention tests, performed in a novel context, between groups either before or after the cue (Fig. 5B and C). These results indicate that CaMKKc tg mice have severe deficits in long-term memory of contextual conditioned fear.

[12] 136w Long-term potentiation (LTP), the activity-dependent change in the strength of neuronal connections, is a form of long-lasting synaptic plasticity that has been proposed as a cellular mechanism of learning and memory. To determine whether CaMKKc tg mice exhibit deficits in synaptic plasticity, we studied LTP at the synapses from Schaffer collateral fibers onto CA1 pyramidal cells in hippocampal slices of these mice. After recording stable baseline fEPSP responses for 30 min, LTP was induced by 1-s, 100-Hz trains. Mean responses for hippocampal slices from wt mice (n = 5) and CaMKKc tg mice (n = 11) show that this form of LTP is significantly reduced in CaMKKc tg mice (Supplementary Fig. S4; p < 0.05). These results suggest that constitutively active CaMKK expression followed by downstream substrate activation leads to a reduced response of synaptic plasticity.

[1] 86w We have shown that CaMKKc leads to impairment in the retention of long-term spatial and contextual memory and hippocampal synaptic plasticity. In CaMKKc tg mice, we observed slightly reduced center times in an open-field test. In learning and memory tasks, we found that CaMKKc tg mice exhibit long-term memory deficits in the Barnes circular maze and contextual fear conditioning. Additionally, hippocampal LTP was impaired in CaMKKc tg mice. These results suggest that constitutive activation of the CaM kinase cascade has inhibitory effects on learning and memory.

[2] 201w We generated CaMKKc tg mice that express a truncated form of CaMKKa which appears to act in a constitutively active fashion. Indeed, we found that phosphorylated CaMKIa was significantly increased in cultured hippocampal neurons of CaMKKc tg mice. The phosphorylation of CaMKI by wt CaMKK is dependent on Ca 2+ /CaM in two distinct ways. This enzyme must bind to Ca 2+ /CaM to be active, and cannot phosphorylate CaMKI unless Ca 2+ /CaM is bound to the substrate enzyme (Haribabu et al., 1995;Tokumitsu & Soderling, 1996). However, CaMKKc, in which the autoinhibitory domain was truncated, is active without binding to Ca 2+ /CaM (Matsushita & Nairn, 1998). It has been reported that purified CaMKKc protein increases phosphorylation of CaMKI by 8-fold in a Ca 2+ /CaM-independent manner (Tokumitsu, Muramatsu, Ikura, & Kobayashi, 2000). In the present study, we showed that phosphorylation of CaMKIa increased in cultured neurons of CaMKKc tg mice compared to wt mice, in the absence of stimulus. In contrast, stimulation with BDNF or KCl did not induce further phosphorylation of CaMKIa in tg mice. It is likely that the overall increase in CaMKK activity precluded the increase of CaMKIa phosphorylation in response to KCl or BDNF stimulation.

[3] 132w CaMKK is an upstream kinase of the CaMK cascade that specifically phosphorylates and activates CaMKI and CaMKIV (Soderling, 1999). Recent molecular and cellular studies indicate a role for CaMKK/CaMKI in the early stage of LTP (E-LTP) via crosstalk with MEK/Erk (Schmitt, Guire, Saneyoshi, & Soderling, 2005). Hippocampal LTP induced by theta-burst stimulation is suppressed $50% by inhibition of MEK/Erk (U0126) or CaMKK (STO-609), and the two effects are occlusive, suggesting that they utilize a common pathway (Guire, Oh, Soderling, & Derkach, 2008). CaMKIV is predominantly a nuclear enzyme, and phosphorylates and activates CREB (Bito et al., 1996;Matthews et al., 1994). Studies of CaMKIV knockout mice revealed impairments in the induction of hippocampal LTP, and long-term fear memory, suggesting deficits in CREB-dependent transcription (Bourtchuladze et al., 1994;Ho et al., 2000;Wei et al., 2002).

[4] 381w Knockout studies have also implicated roles for CaMKK in plasticity and learning and memory. Deletion of CaMKKa results in deficits in fear conditioning (Blaeser et al., 2006;Mizuno et al., 2006). Deletion of CaMKKb results in the loss of long-term, but not shortterm memory for social transmission of food preferences (Peters et al., 2003). Accordingly, this mutation also affects spatial learning and memory. The effects of these CaMKK deletions could result from either suppressed CaMKIV-mediated phosphorylation of CREB/CBP or via crosstalk of CaMKI with the MEK/Erk pathway and subsequent regulation of CREB-dependent transcription or other downstream targets of Erk (Wayman, Lee, Tokumitsu, Silva, & Soderling, 2008). In the present study, we found that CaMKKc tg mice exhibited selective impairments in hippocampus-dependent long-term memory after training in spatial memory tasks and after contextual fear conditioning. Short-term memory 24 h after training was normal in the Barnes circular maze. Tone testing with fear conditioning indicated that the CaMKKc tg mice could acquire and express normal fear memory. These data suggest that the effect of CaMKKc on spatial memory and contextual fear conditioning is not on the acquisition phase of the memory, but on the retention phase. In the contextual fear conditioning test, CaMKKc tg mice showed a slight decrease in freezing at 24-h retention (p = 0.063). It is reasonable to conclude that contextual fear memory was gradually reduced, since a mild loss of memory was observed at 24 h, but a more significant impairment appeared at the 14 day-retention test. This phenotype somewhat resembles the phenotypes of transgenic mice carrying a dominant-negative form of CaMKIV (dnCaM-KIV) (Kang et al., 2001). These transgenic mice show selective deficits in long-term memory of contextual fear conditioning, but the acquisition phase of the memory is normal. This is in contrast to previous reports showing that CaMKK is implicated in and facilitates long-term memory via CREB-dependent transcription. CaM-KKc tg is constitutively activated without Ca 2+ /CaM binding. It may be that a constitutively active CaM kinase cascade in CaMKKc tg mice leads to impairment of synaptic plasticity via fixed phosphorylation status of CaMKKa substrates. Supporting this idea, we found that phosphorylated CaMKIa was significantly increased in cultured neurons of tg mice with no stimulus, though BDNF or KCl did not induce further phosphorylation of CaMKIa in tg mice.

[5] 88w A previous report showed that constitutive activation of CREB during learning causes spatial memory deficits (Viosca et al., 2009). In that report, it was discussed that chronic activation of CREB may lead to an increase of noise, therefore reducing instead of enhancing the signal-to-noise ratio. It has also been reported that transgenic mice carrying a constitutively active form of CaMKIIa show impaired spatial learning (Bach, Hawkins, Osman, Kandel, & Mayford, 1995). These reports suggest that accurate regulation of CaM kinases and the CREB pathway is required for learning.

[6] 505w Previous studies using knockout mice support the view that CaM-KK activation and substrate phosphorylation are required for the induction of LTP (Peters et al., 2003;Wayman et al., 2008). Studies using CaMKIV transgenic mice showed that upregulation of CaMKIV enhances LTP and improves memory formation (Fukushima et al., 2008). It is thought that activation of the CaM kinase cascade will lead to facilitation of hippocampal LTP. In contrast with this idea, CaMKKc tg mice showed significant reduction in hippocampal LTP induced by 100-Hz stimulation, followed by impaired learning. In the present study, phosphorylated CaMKIa was increased in cultured neurons of tg mice with no stimulus because of constitutively activated CaMKKa. However, in BDNF or KCl stimulated conditions, a significant induction of CaMKIa phosphorylation was not observed compared to control mice. We suspect that these weak responses of substrates to such stimulus failed to activate CREB-dependent transcription and prevent excessive neuronal activity in tg mice. Thus, hippocampal LTP was reduced in CaMKKc tg mice, and this might led to insufficient memory formation, followed by the deficits in the retention of long-term memory. However, we have not constructed detail analysis of electrophysiology in tg mouse such as input-output curves or measured paired-pulse facilitation. Further analysis is required to reveal the synaptic function in tg mice. How does specific impairment of long-term memory result from CaMKKc tg expression? In previous work using mutant mice, CaMKKa knockout (KO) mice showed deficits in long-term, but not short-term, memory after contextual fear conditioning (Blaeser et al., 2006). These KO mice exhibited impaired activation of the downstream kinase CaMKIV and CREB upon fear conditioning. Activation of CREB and transcription via this transcription factor are necessary for longterm memory formation after contextual fear conditioning (Athos, Impey, Pineda, Chen, & Storm, 2002;Bourtchuladze et al., 1994). Disruption of CREB function might lead to specific deficits in long-term memory in these KO mice and in our tg mice. However, it is possible that molecules involved in neurotransmitter systems were activated in CaMKKc tg mice, such as synapsin I, which is thought to be a CaM-KI substrate. Proteome and microarray analysis might be required for CaMKKc tg mice to reveal these phenomena. Moreover, other signaling pathways activated by the CaMK cascade cannot be ruled out. CaMKKa has been demonstrated to act as a PKB kinase that phosphorylates PKB on the regulatory threonine 308 residue, leading to its activation (Yano et al., 1998). CaMKKa has also been shown to phosphorylate and activate AMPK (Birnbaum, 2005;Hong, Momcilovic, & Carlson, 2005;Hurley et al., 2005). The present study suggests that moderate levels of CaMKKa activity are necessary for normal signal transduction, followed by synaptic plasticity. Which other molecules are involved in deficits in long-term memory caused by CaMKKc tg expression remains to be investigated. Though these tg mice showed oxidative stress resistance, it is possible that activation of the CaMK cascade may mitigate neuronal death in neurodegenerative diseases, such as Alzheimer's and Parkinson's diseases. Our results demonstrate a role for CaMKK in hippocampal synaptic plasticity and the retention of hippocampus-dependent long-term memory.

[1] 134w Total RNA was prepared from mouse tissues using TRIzol reagent (Invitrogen). For northern blotting, 20-40 lg of total RNA was loaded on a 0.9% denaturing formaldehyde agarose gel and transferred onto a Hybond-N+ membrane (GE Healthcare, USA). The 360-bp fragment of mCaMKK1 cDNA and the 436-bp fragment of EGFP were amplified and labeled with DIG using a PCR DIG Probe Synthesis kit (Roche), and were used as probes. Forward (f) and reverse (r) PCR primers for DIG labeling were as follows: mCaMKK1, f-5 0 -ATGGAGAGTGGCCCAGCCGTC-3 0 and r-5 0 -CTCTG TGTCTGAGATGGCCA-3 0 ; EGFP, f-5 0 -TAAACGGCCACAAGTTCAGC-3 0 and r-5 0 -ATCTTGAAGTTCACCTTGATGCCG-3 0 . The membrane was then probed and washed with stringency buffer. Alkaline Phosphatase-Conjugated anti-digoxigenin antibody and a chemiluminescent detection system were used to visualize with a bio-imaging system (Quantity One, Bio-Rad).

[2] 181w The cortexes from 16 to 18 embryonic day fetal mice carrying either wt or tg genotypes were treated with 0.125% trypsin (Gibco) and 0.004% DNAse-I (Sigma) at 37 °C for 15 min and mechanically dissociated. Neurons were plated on poly-L-lysine (Sigma)coated plastic-bottomed 12-well culture plates (cell density %450,000-500,000/well). Cells were maintained in culture medium with Neurobasal medium, 100 lg/ml penicillin-streptomycin, and 2% B-27 supplement (Gibco). Cultures were maintained at 37 °C in a 95% air/5% CO 2 humidified incubator. On the second or third day after plating, and every 7 days thereafter, half of the culture medium was replaced with fresh medium. At 9-12 days after plating, neurons were washed once with Krebs-ringer HEPES (KRH) which contained 128 mM NaCl, 5 mM KCl, 1 mM NaHPO 4 , 1.2 mM MgSO 4 , 2.7 mM CaCl 2 , 10 mM glucose, and 20 mM HEPES (pH 7.4). After incubation for 1 h in KRH, cells were incubated at 37 °C for 15 min without (control) or with 100 ng/ml BDNF or 30 mM KCl. After incubation, total cellular lysate was extracted.

[3] 95w After treatment of cultured neurons, adherent cells were scraped with RIPA buffer (50 mM Tris-HCl, pH 7.5, 4 mM EGTA, 150 mM NaCl, 50 mM NaF, 1 mM Na 3 VO 4 , 30 mM Na 4 P 2 O 7 Á10 H 2 O, 10 mM EDTA, 100 mM b-glycerophosphate disodium salt hydrate, 1% triton X-100, 50 lg/ml aprotinin, 50 lg/ml leupeptin, 1 mM PMSF, 1 mM DTT). After sonication, the insoluble materials were removed by centrifugation at 15,000 rpm for 15 min, and pellet aliquots (3 ll) were used to determine protein content.

[4] 95w The supernatants were then mixed with Laemmli's sample buffer (0.38 M Tris-HCl, pH 6.8, 12% SDS, 30% b-mercaptoethanol, 10% glycerol, 0.05% bromophenol blue) and boiled for 4 min. Samples containing 20-30 lg of total protein were loaded and subjected to SDS-PAGE and then transferred onto nitrocellulose membranes using an iBlot gel transfer system (Invitrogen). The following antibodies were used at the following dilutions: anti-phosphorylated CaMKI (Tokumitsu et al., 2004), 1:1000; anti-b-actin antibody (Chemicon), 1:2000. The membranes were subsequently incubated with HRP-conjugated secondary antibody (Chemicon, 1:2000 dilution) and the immunoreactive proteins were visualized by standard protocol.

[5] 17w Behavioral analysis was done as described previously (Miyakawa et al., 2003;Takao et al., 2008;Yamasaki et al., 2008).

[6] 128w The background strain used to generate tg mice was C57BL/6 J. All CaMKKc tg mice were heterozygotes from parents of a heterozygous tg  pure C57BL/6 J wt mouse, and all wt mice were their littermates. All behavioral tests were carried out with male mice that were at least 10 weeks old at the start of the testing. Mice were housed in a room with a 12-h light/dark cycle (lights on at 07:00) with access to food and water ad libitum. Behavioral testing was performed between 08:00 and 18:00. After each test, the apparatus was cleaned with super hypochlorous water to prevent a bias due to olfactory cues. All behavioral testing procedures were approved by the Animal Care and Use Committee of Kyoto University Graduate School of Medicine.

[7] 58w To assess spontaneous locomotor activity, mice were placed into the center of an open-field apparatus (40 Â 40 Â 30 cm; Accuscan Instruments, Columbus, OH) (Yamasaki et al., 2008). The chamber was illuminated (100 lux). Motor activity parameters (distance traveled, number of vertical and stereotypic movements, and center time) were then monitored and recorded over a 120min period.

[1] 115w The full-length wt cDNA for mouse CaMKK1 was obtained from C57BL/6J mouse cerebellum by RT-PCR. Amino acids 1-433 (CaMKK433) were cloned into EGFP-N1 using EcoRI and ApaI sites. An EcoRI-NotI fragment of EGFP-CaMKK433 was then inserted into the EcoRV site of the vector, pNN265 (provided by N. Kojima, Gunma University), which contains the untranslated leader with a hybrid intron (Choi, Huang, Gorman, & Jaenisch, 1991) and SV40 polyA sequences. To add the CaMKIIa promoter to this construct, EGFP-CaM-KK433 with an untranslated leader and SV40 sequences was excised using NotI and inserted into the pBluescript-based vector, pNN279 (provided by N. Kojima), into which the 7.9-kb fragment, including the CaMKIIa promoter (Mayford et al., 1996), was inserted.

[2] 58w The transgene was excised from the vector using SalI and was gel purified. Transgenic mice were generated by microinjecting the transgene into the pronuclei of fertilized eggs collected from C57BL/6J mice; the injected eggs were then transferred into the oviducts of pseudopregnant females. Founders were analyzed by Southern blotting of tail DNA using a mouse CaMKK1 cDNA probe.

[3] 266w This test was performed as described previously (Takao et al., 2008) at 16-20 weeks old. The test was conducted on ''dry land,'' a white circular surface, 1.0 m in diameter, with 12 holes equally spaced around the perimeter (O'hara & Co., Tokyo). The circular open field was elevated 75 cm from the floor. A black Plexiglas escape box (17 Â 13 Â 7 cm), lined with paper cage bedding, was located under one of the holes. The hole above the escape box represented the target, analogous to the hidden platform in the Morris task. The location of the target was consistent for a given mouse, but was randomized across mice. The maze was rotated daily, with the spatial location of the target unchanged with respect to the distal visual room cues, to prevent a bias based on olfactory or proximal cues within the maze. Two or three trials per day were conducted for seven successive days. On day 8, a probe trial was conducted without the escape box, to confirm that this spatial task was acquired based on navigation using distal environment room cues. One trial was conducted immediately after the probe test and another probe trial was conducted 14 days later. The amount of time that mice took to enter the target hole (latency to 1st), the number of errors to reach the target hole, and the time spent around each hole were recorded by video tracking software (Image BM). Fourteen days after the last training trial, a probe trial was conducted to evaluate memory retention, and the time spent around each hole was recorded.

[4] 266w This test was performed as described previously (Miyakawa et al., 2003) at 19-23 weeks old. On the conditioning day, each mouse was placed in a test chamber (26 Â 34 Â 29 cm) inside a sound-attenuated room and was allowed to explore freely for 2 min. A 55-dB white noise, which served as the conditioned stimulus (CS), was presented for 30 s, followed by a mild (2 s, 0.3 mA) footshock, which served as the unconditioned stimulus (US). Two more CS-US pairings were presented with a 2-min inter-stimulus interval. One day after conditioning, context testing was conducted in the same chamber. Cued testing with an altered context was conducted after conditioning using a triangular box (35 Â 35 Â 40 cm) made of white opaque Plexiglas, which was located in a different room. Data acquisition, control of stimuli (i.e., tones and shocks), and data analysis were performed automatically, using Image FZ software. Images were captured at 1 frame per second. For each pair of successive frames, the amount of area (pixels) by which the mouse moved was measured. When this area was below a certain threshold (i.e., 20 pixels), the behavior was judged as 'freezing'. When the amount of area equaled or exceeded the threshold, the behavior was considered as 'non-freezing'. The optimal threshold (amount of pixels) to judge freezing was determined by adjusting it to the amount of freezing measured by human observation. 'Freezing' that lasted less than the defined time threshold (i.e., 2 s) was not included in the analysis. To evaluate memory retention, context and cued testing were conducted 14 days after conditioning.

[5] 100w All procedures involving mice were performed in compliance with the National Institutes of Health guidelines and were approved by the Animal Care and Use Committee of Mitsubishi Kagaku Institute of Life Sciences and the Kyoto University Graduate School of Medicine. Behavioral data were obtained automatically by applications based on the public domain NIH Image program and Image J program and modified for each test by Tsuyoshi Miyakawa (available through Ohara & Co.). Statistical analysis was conducted using StatView (SAS Institute). Data were analyzed by two-way ANOVA, or two-way repeated measures ANOVA. Values in graphs are expressed as means ± SEM.