RCAN1 (DSCR1 or Adapt78)* stimulates expression of GSK-3b Gennady Ermak, Cathryn D. Harris, Denis Battocchio and Kelvin J. A. Davies Ethel Percy Andrus Gerontology Center, and Division of Molecular & Computational Biology, The University of Southern California, Los Angeles, CA, USA The RCAN1 (regulator of calcineurin) gene was pre- viously known as DSCR1 or Adapt78. DSCR1 or Adapt78 was discovered by two independent laborator- ies as a chromosome 21 Down syndrome critical region (DSCR1) gene [1], and as a gene transiently induced during cellular adaptation (Adapt78) to oxida- tive stress [2]. The yeast homolog was named RCN1 [3] and the Drosophila homolog was named NEBULA [4]. Several groups simultaneously reported that DSCR1 or Adapt78 protein can specifically bind to, and down-regulate the activity of calcineurin [3,5–7]. Some groups therefore named this protein calcipressin 1, while others called it calcineurin binding protein (CBP 1) [7] or myocyte-enriched calcineurin interacting protein (MCIP1) [6]. In yeast, the RCAN1 protein was called Rcn1p [3], and in Drosophila, this protein was named sarah [8]. The new name RCAN1 (regulator of calcineurin) has recently been accepted by the HUGO Gene Nomenclature Committee for the protein prod- uct of the RCAN1 gene. RCAN1 is thought to play a role in number of physiological processes. It has been associated with Keywords Adapt78; calcineurin; DSCR1; GSK-3b; RCAN1 Correspondence K. J. A. Davies, Andrus Gerontology Center, University of Southern California, 3715 McClintock Avenue, Los Angeles, CA 90089–0191, USA Fax: +1 213 7406462 Tel: +1 213 7408959 E-mail: kelvin@usc.edu *The new name RCAN1 (regulator of calcineurin) has recently been accepted by the HUGO Gene Nomenclature Committee and the Mouse Genomic Nomenclature Committee (MGNC) for the gene previously known as DSCR1 or Adapt78. Similarly, RCAN1 is the new name for its protein product, which was previously called calcipressin1 or MCIP1. (Received 24 January 2006, revised 1 March 2006, accepted 7 March 2006) doi:10.1111/j.1742-4658.2006.05217.x The RCAN1 protein (previously called calcipressin 1 or MCIP1) binds to calcineurin, a serine ⁄ threonine phosphatase (PP2B), and inhibits its activ- ity. Here we demonstrate that regulated overexpression of an RCAN1 transgene (this gene was previously called DSCR1 or Adapt78) also stimu- lates expression of the GSK-3b kinase, which can antagonize the action of calcineurin. We also show that GSK-3b is regulated by RCAN1 at a post- transcriptional level. In humans, high RCAN1 expression is found in the brain, where at least two mRNA isoforms have been reported. Therefore, we further investigated expression of the various RCAN1 isoforms, result- ing from differential splicing and alternative promotors in human brain. We detected at least three distinct RCAN1s: RCAN1-1 Short at 31 kDa (RCAN1-1S), RCAN1-1 Long at 38 kDa (RCAN1-1 L), and RCAN1-4. Furthermore, the levels of RCAN1-1S, but not RCAN1-1 L or RCAN1-4 correlated with the levels of GSK-3b. This suggests that RCAN1-1S might induce production of GSK-3b in vivo. While RCAN1s can regulate cal- cineurin and GSK-3b, it has also been shown that calcineurin and GSK-3b can regulate RCAN1s. Here we propose a new model (incorporating all these findings) in which cells maintain an equilibrium between RCAN1s, calcineurin, and GSK-3b. Abbreviations CBP 1, calcineurin binding protein; GSK, glycogen synthase kinase; MCIP1, myocyte-enriched calcineurin interacting protein; NFATs, nuclear factors of activated T-cells; PP2B, protein phosphatase 2B. 2100 FEBS Journal 273 (2006) 2100–2109 ª 2006 The Authors Journal compilation ª 2006 FEBS Down’s syndrome and Alzheimer’s disease [5,9]. It has also been demonstrated to inhibit cardiac hypertrophy [10] and to attenuate angiogenesis and cancer [11]. Exactly how RCAN1 may contribute to human pathologies, however, is a subject of intense continuing study. So far, the only established function of the RCAN1 protein is the inhibition and regulation of calcineurin, which is protein phosphatase 2B (PP2B). Calcineurin, however, works together with (or against) kinases such as glycogen synthase kinase-3b (GSK-3b) to regulate protein phosphorylation ⁄ dephosphoryla- tion, and GSK-3b can antagonize calcineurin. There- fore, here we have investigated whether GSK-3b can also be regulated by RCAN1. One of the major cell types expressing RCAN1 in the human body is neurons [9], suggesting important roles for the protein in brain function. The RCAN1 (DSCR1, Adapt78) gene consists of seven exons, four of which (exons 1–4) can be alternatively spliced to produce a number of different mRNA isoforms. Since exons 5, 6, and 7 are likely to be common to all mRNA isoforms, for nomenclature simplification, it was proposed to assign them (and the proteins they encode) numbers which correspond to the first exons they contain [12]. Expression of RCAN1 mRNA in human brain was previously investigated [9], but not expression of the protein isoforms. Therefore, we have now identified which RCAN1s are expressed in human brain, and we have questioned which isoforms (if any) may regulate GSK-3b. It has recently been demonstrated, both in vitro and in vivo, that RCAN1 can be phosphorylated [13–16]. Phosphorylation may alter RCAN1’s ability to inhibit calcineurin, as well as RCAN1 degradation rate [14–16]. Therefore, phosphorylation may represent a major mechanism for regulating RCAN1. Remarkably, RCAN1 can be phosphorylated by GSK-3b and dephosphorylated by calcineurin [13,16] and we now demonstrate that RCAN1 can also, reversibly, regulate both enzymes. Results Transgenic RCAN1 stimulates expression of GSK-3b in vitro To test whether RCAN1 might regulate GSK-3b we developed a tet-off regulated gene expression system, in which RCAN1 expression can be regulated by doxy- cycline [12,17]. To keep the transgene ‘silent’, cells were continuously grown in the presence of doxycycline, and inducible RCAN1 transgene expression was achieved by doxycycline withdrawal from the cell culture med- ium. RCAN1 was maximally overexpressed about 6 h after doxycycline withdrawal from the cell medium, but its levels then declined over the next 48 h (Fig. 1A). Remarkably, GSK-3 protein levels followed the levels of RCAN1; they increased following RCAN1 overex- pression and declined at 48 h Fig. 1A,B. It seems that both GSK-3b and GSK-3a levels were elevated. However, the level of GSK-3b was increased much more significantly ( 2.3-fold after 24 h of RCAN1 overexpression) than was the level of GSK-3a. GSK-3 activity is regulated by phosphorylation. Par- ticularly, its activity is inhibited by phosphorylation at the S9 position [18]. Therefore, we also tested whether the phosphorylated form of GSK-3 (pGSK-3) was ele- vated following RCAN1 overexpression. We found no significant increase in pGSK-3 levels until 48 h after doxycycline withdrawal. These results indicate that RCAN1 stimulates production of active GSK-3, which then phosphorylates the tau protein and promotes accumulation of hyperphosphorylated tau. It seems, however, that cells might have a feedback mechanism that controls GSK-3 activity and, after 48 h, GSK-3 was increasingly phosphorylated (Fig. 1C). As hypothesized, the levels of phosphorylated tau cor- related with RCAN1 (Fig. 1A). Tau phosphorylation increased during RCAN1 overexpression (Fig. 1D), whereas the levels of total tau protein were unchanged. When combined, these results further strengthen our proposal that RCAN1 can induce the active form of GSK-3. It is interesting that the levels of GSK-3b were modulated. GSK-3b is closely associated with pTau of tangle-bearing neurons [19], and it is specifically the GSK-3b isoform that is increased in pretangle neurons [20]. GSK-3b levels did not perfectly correlate with pTau levels; the highest pTau levels were observed about 6 h after RCAN1 overexpression, and although GSK-3b levels were increased almost two-fold at this time point, they did not peak until 24 h (Fig. 1). This might be due to competition between calcineurin and GSK-3b, or to activation of other phosphatases after the 6-h time point. It is also possible that other kinases, that phosphorylate the tau protein, are also activated following RCAN1 overexpression but, unlike GSK-3b, they are activated maximally at the 6 h time point. GSK-3b appears to be regulated by RCAN1 at a post-transcriptional level We next tested whether RCAN1 induces transcription of the GSK-3b gene or if it regulates GSK-3b expres- sion at a post-transcriptional level. RCAN1 was overex- pressed as described in Fig. 1 and cells were split into two portions: one portion was used for western blot G. Ermak et al. RCAN1 stimulates GSK-3b FEBS Journal 273 (2006) 2100–2109 ª 2006 The Authors Journal compilation ª 2006 FEBS 2101 analysis and the other for northern blot analysis. Levels of GSK-3b and RCAN1 mRNA were evaluated by nor- thern blot analysis (Fig. 2). To verify that each sample was equally loaded, the membrane was probed with GAPDH (a housekeeping gene) probe. Results revealed that RCAN1 mRNA was elevated as early as 3 h and was clearly overexpressed at each time point: 6, 24, and 48 h. GSK-3b mRNA levels, however, remained unchanged at all time points. These results are also confirmed by our previous data obtained using micro- array analysis of genes that are regulated by RCAN1 [17]. In these previous experiments we did not observe any GSK-3b mRNA transcription level changes follow- ing RCAN1 down-regulation. Altogether these results indicate that RCAN1 does not affect the levels of tran- scription of GSK-3b, but it rather regulates production of the GSK-3b protein at a post-transcriptional level. This might be due to either accelerated GSK-3b trans- lation and ⁄ or slower GSK-3b degradation; the exact mechanism(s) remain(s) to be addressed. It is interesting that RCAN1 mRNA levels were not auto-down-regulated at any time point, while levels of the RCAN1 protein were slightly down-regulated 24 h after overexpression, and then returned to basal levels in 48 h (Fig. 1A, top panel). These results indicate that RCAN1 expression might be feed-back regulated at a Fig. 1. RCAN1 stimulates expression of GSK-3. (A) RCAN1 was overexpressed in PC-12 cells as described in Experimental procedures. Equal amounts of total protein from each sample were loaded and tubulin detection was used to control loading levels. Please note that the antibody that recognizes GSK-3 binds to both its unphosphorylated and its phoshorylated forms. The amount of phosphorylated GSK-3, how- ever, is much lower than the unphosphorylated form and therefore is not detected on the blots. Only the antibody that specifically recogni- zes phosphorylated GSK-3 produced clear signals. (B) X-ray films were quantified using IPLAB software (Scanalytics) and signals were adjusted according to the loading. The amount of the loaded protein was controlled using b-tubulin detection. GSK-3 levels are expressed in arbitrary units (AU) relative to b-tubulin. GSK-3b protein levels were elevated 2.3-fold after 24 h of RCAN1 overexpression, whereas GSK3 protein levels were increased by only 50%. Both elevations were statistically significant at the P £ 0.05 level, as evaluated by the student’s t-test (one population). (C) Protein levels were quantified as described in (B). pGSK-3 levels are expressed in arbitrary units (AU), relative to b-tubulin. (D) Protein levels were quantified as described in (B). pTau levels are expressed in arbitrary units (AU) relative to b-tubulin. In (B), (C), and (D) all bars represent mean values ± standard errors, of three experiments. RCAN1 stimulates GSK-3b G. Ermak et al. 2102 FEBS Journal 273 (2006) 2100–2109 ª 2006 The Authors Journal compilation ª 2006 FEBS post-transcriptional level, most likely, by accelerated degradation of RCAN1. At least three RCAN1s are expressed in human brain Western blot analysis of RCAN1 expression in human brain revealed that at least three different RCAN1 pro- teins are expressed (Fig. 3). The RCAN1 gene consists of seven exons that can be alternately spliced to produce a large number of isoforms [12]. Two mRNA isoforms were previously shown to be expressed in adult human brain: isoform 1 (encoded by exons 1, 5, 6, and 7) and isoform 4 (encoded by exons 4, 5, 6, and 7). To determine the structure of each of these RCAN1s we have developed three different antibodies: an antibody that recognizes RCAN1 encoded by exon 1 (RCAN1-1), an antibody that recognizes RCAN1 encoded by exon 4 (RCAN1-4), and an antibody that recognizes all poten- tial RCAN1 protein isoforms (directed against the invariant exon 7). Surprisingly, we found that not one but two RCAN1-1 proteins are expressed in human brain (Fig. 3A,C): RCAN1-1 ‘Long’ at 38 kDa (RCAN1-1 L) and RCAN1-1 ‘Short’ at 31 kDa (RCAN1-1S). Using the complete sequence of chromo- some 21, we analyzed for RCAN1 potential transcrip- tion initiation sites and translation codons. Computer Fig. 2. GSK-3b appears to be regulated by RCAN1 at a post-tran- scriptional level. (A) RCAN1 mRNA was overexpressed as des- cribed in Experimental procedures. Equal amounts of total RNA from each sample were loaded and GAPDH detection was used as a loading control. GSK-3b mRNA was detected using a labeled oligonucleotide probe, as described in Experimental procedures. Oligonucleotides used to prepare labeled probes, and reverse com- plimentary oligonucleotides, were bound to membranes to control the specificity of hybridization. Reverse complementary oligonucleo- tides produced a strong hybridization signal, while the original oligo- nucleotides did not produce a signal (not shown). (B) Membranes were scanned and the hybridization signal measured using IMAGE- QUANT software (Molecular Dynamics). Each signal was recalculated according to the amount of RNA actually loaded on the gels. The amount of RNA loaded was measured using hybridization with a GAPDH probe. GSK-3b levels are expressed in arbitrary units (AU) relative to GAPDH, and bars represent mean values of three experi- ments ± standard errors. GSK-3b mRNA levels were not signifi- cantly changed after RCAN1 overexpression. Fig. 3. RCAN1 expression in human brain. Examples of western blots analyzing RCAN1 expression in either the cerebral cortex (‘cortex’), or the hippocampus (‘hc’). (A) The antibody designed against isoform 1 recognizes RCAN1s of about 31 kDa (RCAN1-1S) and 38 kDa (RCAN1-1L). (B) The antibody designed against isoform 4 recognizes an RCAN1 protein of about 70 kDa (RCAN1-4). (C) The antibody designed against all possible RCAN1 isoforms (containing the invariant exon 7) recognizes three proteins with molecular weights of approximately 70, 38 and 31 kDa: RCAN1-4, RCAN1-1 L, and RCAN1-1S, respectively. G. Ermak et al. RCAN1 stimulates GSK-3b FEBS Journal 273 (2006) 2100–2109 ª 2006 The Authors Journal compilation ª 2006 FEBS 2103 analysis revealed that exon 1 has an alternative initi- ation translation codon, located further upstream than previously thought, indicating that, indeed, two RCAN1-1 isoforms can be expressed. This possibility was also confirmed by our experiments, in which RCAN1 was overexpressed using the tet-off system (please see above). The construct used to overexpress RCAN1 in these experiments carried a DNA fragment which encodes only RCAN1-1S, and this fragment was translated only into a 31-kDa protein (Fig. 1). Unexpectedly, RCAN1-4 was detected as a 70 kDa protein in human brain (Fig. 3A–C). Based on the sequence of the RCAN1 gene, the size of the 1–4 iso- form was expected to be roughly equal to RCAN1-1S, which is about 31 kDa. This raises questions about the specificity of our antibodies; however, all three RCAN1 antibodies, including the RCAN1-4 antibody, were carefully tested as shown in Fig. 4. The RCAN1- 4 antibody always showed one clear band. In addition, we further tested our antibodies in substrate competi- tion experiments, in which the peptide used to develop each antibody was added to the antibody solutions used for western blots. In each case, the addition of the correct peptide significantly reduced the appropri- ate signal levels (not shown). It seems that different RCAN1 isoforms can be pro- duced depending on the cell type, the model system used, and the particular conditions employed. We have tested RCAN1 expression in several cell lines and found that some of them strongly express RCAN1s of various sizes, while others express very low levels of RCAN1s (e.g. Fig. 5). In most cases, RCAN1-1L appears to be the predominant isoform expressed. GSK-3b correlates with the levels of RCAN1-1S, but not RCAN1-1L or RCAN1-4 in vivo We next tried to address whether, similar to our in vitro results, RCAN1 might influence GSK-3b expression in vivo. We analyzed 27 randomly chosen human brain samples from various brain regions: cer- ebral cortex, cerebellum and hippocampus (Fig. 6). Interestingly, GSK-3b levels in all samples was tightly correlated ( r ¼ 0.7, P < 0.001) with the expression of only one RCAN1 isoform, RCAN1-1S, but not RCAN1-1L or RCAN1-4. This correlation was inde- pendent of the brain region analyzed. RCAN1-1S was the isoform overexpressed in the experiments described in Fig. 1 above. These results therefore suggest that, similar to our in vitro observa- tions (Fig. 1), GSK-3b may also be induced by expres- sion of RCAN1-1S in human brain in vivo. This study was not designed to address whether induction of RCAN1 can actually cause Alzheimer’s disease. The brain samples examined in our studies were all from patients  80 years old, some of whom suffered from Alzheimer’s disease and some of whom did not. Alzhei- mer’s disease, however, is clearly an age-related disor- der. Almost half of all humans suffer some degree of Alzheimer’s disease by this age, and a large proportion of our population might have developed the disease without displaying overt symptoms [21]. Therefore, older patients might have increased expression of RCAN1 and GSK-3b even if they are not yet diag- nosed with Alzheimer’s disease. It has actually been shown that GSK-3b is induced in pretangle neurons [20]. For these reasons, a completely different approach will now be needed to address whether induction of RCAN1 can actually cause Alzheimer’s disease. Discussion Previously it has been shown that RCAN1 can down- regulate the activity of calcineurin, the serine-threonine phosphatase PP2B [3,5–7]. Here we demonstrate that, in addition, RCAN1 can also induce GSK-3b, a kinase that antagonizes calcineurin. Calcineurin and GSK-3b work in opposition to regulate the phosphorylation ⁄ Fig. 4. Production of the RCAN1-4 antibody. Rabbits were injected with exon 4 peptide, which is described in Experimental procedures. Serum from ‘immunized’ rabbits was then used for western blots. This serum revealed a new band that was absent in ‘preimmunized’ ani- mals. Next, the antibody was affinity purified from the serum of immunized animals using columns with the covalently bound exon 4 peptide that was used for immunization. ‘Purified’ (bound to the column) material detected clear bands that are similar to those observed using ‘immunized’ serum. Such bands were absent in ‘flow-through’ (unbound to the columns) material. RCAN1 stimulates GSK-3b G. Ermak et al. 2104 FEBS Journal 273 (2006) 2100–2109 ª 2006 The Authors Journal compilation ª 2006 FEBS dephosphorylation of proteins and, thus, to regulate their activities. Because RCAN1 inhibits calcineurin, its overexpression will lead to protein phosphorylation and, as RCAN1 induces GSK-3b, its overexpression should lead to further protein phosphorylation. Thus, in both cases, phosphorylated forms of proteins will be produced. The relationship between RCAN1 and cal- cineurin is, however, extremely complex. It has been shown, for example, that RCAN1 can inhibit, activate, or have no effect on calcineurin, depending on its con- centration and its phosphorylation status [13,14,16]. Therefore, since almost any change in calcineurin activity could be interpreted in any convenient (or triv- ial) way, it was not measured in our study. A number of vital proteins are regulated by GSK-3b and calcineurin. One of these is tau, a microtubule- associated protein that is important for cytoskeleton stabilization and cell growth. Accumulation of hyper- phosphorylated tau can lead to the formation of neurofibrillary tangles, and to neurodegeneration asso- ciated with tauopathies. Hyperphosphorylated forms of tau are also involved in pathologies such as Down’s syndrome and Alzheimer’s disease. Importantly, we observed accumulation of phosphorylated tau protein following overexpression of RCAN1 in our experi- ments. RCAN1 is a stress-inducible gene, and neurodegener- ative diseases are associated with stresses such as chro- nic inflammation. For example, it has been shown that mechanical stress and traumatic brain injury are robust factors for Alzheimer’s disease [22–25]. Interestingly, RCAN1 expression has been shown to be inducible by mechanical stimulus [26], thus providing a possible link between injury and Alzheimer’s disease. We have also observed RCAN1 overexpression following mechanical injury in cell culture models (not shown). Of course, any event causing inflammation would be expected to induce RCAN1 expression, since the gene was identified as an A B Fig. 6. Levels of GSK-3b in human brain correlate with RCAN1-1S. (A) Sample western blots showing GSK-3b and RCAN1 expression. Equal amounts of total protein from each brain sample were loa- ded. Three RCAN1 proteins are expressed in brain, but the levels of GSK-3b only correlate with RCAN1-1S (see samples 4 and 5). A ponceau stained membrane is shown to demonstrate loading lev- els. (B) summary of western blot analyses. Twenty-seven samples originating from different brain areas of 18 patients were analyzed. X-ray films were quantified using IPLAB software (Scanalytics, Fairfax, VA). GSK-3b and RCAN1-1S protein levels are expressed in arbitrary units (AU). The linear regression line shown revealed a correlation coefficient (r) of 0.7, P < 0.001. Fig. 5. Expression of RCAN1 in other cell lines. Examples of west- ern blots demonstrate RCAN1 expression in two cell culture mod- els. Our RCAN1 common antibody recognizing exon 7 that is common to all isoforms was used for the detection. (A) Detection of RCAN1 protein in human fibroblast cell line WI-38 VA-13. At least four RCAN1 isoforms (25, 38, 70, and  190 kDa) were detected. (B) Detection of RCAN1 protein in human neuroblastoma cell line SK-N-MC. Only one weak 38 kDa band was detected. G. Ermak et al. RCAN1 stimulates GSK-3b FEBS Journal 273 (2006) 2100–2109 ª 2006 The Authors Journal compilation ª 2006 FEBS 2105 oxidative stress adaptive factor [2]. Thus, it is possible that chronic stress may contribute to neurodegeneration through induction of RCAN1, which in turn leads to inhibition of calcineurin and activation of GSK-3b, both resulting in tau protein hyperphosphorylation. Besides regulation of tau, both calcineurin and GSK-3b are key regulators of nuclear factors of activa- ted T-cells (NFATs), which are important players in a number of physiological processes. The best character- ized role of NFATs is in T-cell activation. When NFATs are dephosphorylated (due to the induced activity of calcineurin and ⁄ or reduced activity of GSK- 3b) they are transported from the cytoplasm into the nucleus, and activate T-cells. Therefore, overexpression of RCAN1 could lead to phosphorylation of NFATs and T-cell inactivation. NFATs are also important players in musculoskeletal development, cardiovascular development and functions [27], and neuronal develop- ment [28]. Thus, NFATs and RCAN1 may contribute to the pathologies observed in Down’s syndrome, such as mental retardation, heart defects, skeletal muscle abnormalities, and immune system deficiencies. Indeed, it has been previously shown that activated GSK-3b phosphorylates NFAT proteins and suppresses cardiac hypertrophy [29]. On the other hand, it was also shown that overexpression of RCAN1 can inhibit car- diac hypertrophy [10], which is consistent with our hypothesis that overexpression of RCAN1 can activate GSK-3b, promoting NFATs phosphorylation and sup- pression of cardiac hypertrophy. GSK-3b has a number of substrates and is a multi- functional kinase. Besides Alzheimer’s disease and car- diac development, its function is closely linked to type II diabetes, various forms of cancer and other human pathologies [30]. Our finding that RCAN1 can modu- late the levels of this enzyme uncovers a possible new link between stress and human disease. Similarly, cal- cineurin is involved in many vital pathways [27,31] and inhibition of this phosphatase by RCAN1 may be expected to have serious repercussions. Thus, our results raise the possibility that RCAN1 may play important roles in a variety of chronic stress, or chro- nic inflammation-related diseases including (but not limited to) Alzheimer’s disease, Down’s syndrome, and other neurodegenerative disorders. It would seem that RCAN1 overexpression can lead to dramatic physiological changes, and our results suggest that cells cope with RCAN1 overexpression quickly and efficiently. In our experiments, cells recog- nized high levels of RCAN1-1S and down-regulated its forced expression to basal levels within 48 h (Figs 1 and 2). We wish to know how cells do this. Current data suggest that RCAN1 regulation is accomplished by phosphorylation. While RCAN1 can regulate calcineu- rin and GSK-3b, calcineurin and GSK-3b can also regulate the phosphorylation of RCAN1 [13,16,29]. It has been demonstrated that phosphorylation of RCAN1s increases their degradation rates [15]. There- fore, both activation of GSK-3b and inactivation of calcineurin should lead to the same results: increased RCAN1 phosphorylation and faster degradation (Fig. 7). Thus, overexpression of RCAN1 will eventu- ally lead to RCAN1 phosphorylation and removal. It appears that cells maintain an equilibrium between RCAN1, calcineurin, and GSK-3b, and serious prob- lems may arise when this balance is distorted. This model, however, remains to be tested. Experimental procedures Doxycycline-regulated RCAN1 expression system An RCAN1 DNA fragment corresponding to the RCAN1– 1S protein was produced by long and accurate (LA) RT-PCR as previously described [12]. The fragment was inserted into a pTRE vector from Clontech Laboratories, Inc. (Paolo Alto, CA) PC-12 tet-off cells from Clontech Laboratories, Inc., stably transfected with ‘regulator plas- mid’, were next transfected with the pTRE carrying RCAN1 fragment to finally produce a double-stable cell line in which the RCAN1 transgene could be turned off by doxycycline addition. All procedures were performed as described in the tet-off gene expression user manual from Clontech Laboratories, Inc. Fig. 7. RCAN1–calcineurin–GSK-3b equilibrium. Multiple stresses induce expression of RCAN1. RCAN1 down-regulates the activity of calcineurin and activates GSK-3b. Both activation of GSK-3b and inactivation of calcineurin lead to the same results: increased RCAN1 phosphorylation and faster proteolytic degradation. When RCAN1 levels are reduced to normal, calcineurin and GSK-3b activ- ities also return to normal levels. RCAN1 stimulates GSK-3b G. Ermak et al. 2106 FEBS Journal 273 (2006) 2100–2109 ª 2006 The Authors Journal compilation ª 2006 FEBS PC-12 cell culture PC-12 cells were cultivated in media containing 85% Dul- becco’s modified Eagle’s medium (DMEM), 10% horse serum, 5% antibiotic free fetal bovine serum (Clontech Laboratories, Inc.), 2 mml-glutamine, 100 UÆmL )1 penicil- lin, and 100 lgÆmL )1 streptomycin sulfate. PC-12 tet-off cells, stably transfected with ‘regulator plasmid’, were main- tained in media additionally containing 50 lgÆmL )1 G418. The double-stable cell line carrying both the ‘regulator plas- mid’ and the RCAN1 transgene was maintained in media additionally containing 50 lgÆmL )1 G418 and 50 lgÆmL )1 hygromycin. Cells were cultivated in a humidified 10% CO 2 atmosphere at 37°C. Expression of the RCAN1 transgene was turned off by application of 0.1 lgÆmL )1 doxycycline to the media. Post mortem human brain tissue Brain tissue ⁄ CSF used in this project was provided by the Alzheimer’s Disease Research Center, University of South- ern California. Samples were frozen at )70°C until use. Every sample was accompanied by a full medical history, and by detailed neuropathology summaries that included microscopic exams. Protocols using frozen samples of human brain were approved by Institutional Review Board. RCAN1 antibodies The following three RCAN1 antibodies were designed using the predicted open reading frame sequence of RCAN1: (1) an antibody that recognizes RCAN1 encoded by exon 1 (RCAN1-1) was developed using a hypothetical peptide encoded by the MEEVDLQDLPSAT portion of exon 1; (2) an antibody that recognizes RCAN1 encoded by exon 4 (RCAN1-4) was developed using a hypothetical peptide encoded by the VANSDIFSESETR portion of exon 4; and (3) an antibody that recognizes all potential RCAN1 pro- tein isoforms was developed using a hypothetical peptide encoded by the KIIQTRRPEYTPIHLS portion of the exon 7. We added a cysteine to the N-terminus of these peptides and conjugated them to keyhole limpet hemocyanin. The peptides were then used to immunize rabbits and polyclonal antibodies were raised commercially (ProSci, Inc., Poway, CA). Serum from the immunized rabbits was affinity puri- fied on columns with the covalently attached RCAN1 pep- tide used for immunization. Western blot analysis Western blot analysis was performed following standard protocols, using an ECL-detection system from Amersham Biosciences (Piscataway, NJ). Final dilution of the RCAN1 antibodies for western blot analysis ranged between 1 : 500 and 1 : 1000. Phosphorylated GSK-3 protein (pGSK-3) was detected using an antibody that specifically recognizes the GSK-3 protein, phosphorylated at serine 9. Phosphorylated tau protein (pTau) was detected using an antibody that specifically recognizes the tau protein, phosphorylated at threonine 231 (pT[231]). It has been demonstrated that tau is phosphorylated at threonine 231 by GSK-3 and it can be dephosphorylated by calcineurin [32]. Anti-tau antibody was obtained from Upstate Biotechno- logy (catalog #05–348). Anti-tau [pT[231]]-phosphospecific antibody was from Biosource International (catalog #44– 746) (Camarillo, CA). Anti-GSK-3 antibody was from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA) Anti- GSK-3 [pS[9]] phosphospecific antibody was from Santa Cruz Biotechnology, Inc. Tubulin antibody was from Santa Cruz Biotechnology, Inc. Commercial antibodies were dilu- ted as suggested by the manufacturers. RNA isolation Total RNA was extracted using the TRIzol reagent (Life Technologies, Inc.). RNA concentration was quantified spectrophotometrically, and relative content was further confirmed on ethidium bromide-stained gels. Integrity of the RNA was estimated by agarose gel electrophoresis; only RNA samples displaying discrete 28S and 18S RNA bands were used in our experiments. Northern blot analysis Samples containing 30 lg of total RNA were subjected to electrophoresis through 1% agarose formaldehyde gels, blotted to nylon membranes (Intergen, Purchase, NY) with HETS (Tel-Test Inc., Friendswood, TX), and cross-linked by ultraviolet radiation. Three probes were used in this study: RCAN1, GAPDH, and GSK-3b. RCAN1 and GAP- DH were synthesized by PCR, and to detect GSK-3b we used synthetic oligonucleotides. Therefore, hybridization conditions for these probes were different. For RCAN1 and GAPDH detection, the membranes were prehybridized for 4 h and hybridized for 15 h in Hybrizol I (Oncor) at 42°C. After hybridization they were washed with 2 · NaCl ⁄ Cit plus 0.1% SDS at room temperature for 1 and 10 min, then with 0.1 · NaCl ⁄ Cit plus 0.1% SDS at 60°C for 10 and 30 min. For hybridization with GSK-3b, we used ULTRA- hyb–Oligo solution, Ambion Inc. (Austin, TX) followed by protocol from manufacturer. The membranes were exposed, developed, and scanned using the phosphorimager system (Molecular Dynamics, Sunnyvale, CA). To rehybridize RNA blots, hybridized and labeled probes were removed by washing the membranes in a solution of 0.1 · NaCl ⁄ Cit, 0.1% SDS, and 10 mm Tris ⁄ HCl (pH 7.0) at 90°C for 10 min. To quantify the level of DSCR1 (Adapt78) mRNA expression the membranes were scanned G. Ermak et al. RCAN1 stimulates GSK-3b FEBS Journal 273 (2006) 2100–2109 ª 2006 The Authors Journal compilation ª 2006 FEBS 2107 and the hybridization signal measured using imagequant software (Molecular Dynamics). Each signal was recalcu- lated according to the amount of RNA actually loaded on the gels. The amount of RNA loaded was controlled by hybridization with a GAPDH probe. RCAN1 and GAPDH probes were labeled using [ 32 P]dCTP[aP] and the High Prime system (Boehringer Mannheim, Indianapolis, IN). A fully cloned hamster Adapt78 fragment was used to prepare DSCR1 (Adapt78) probes; and a PCR-fragment consisting of exons 7 and 8 was used to prepare GAPDH probes as we previously des- cribed [9, 17]. To prepare GSK-3b probe we synthesized the following oligonucleotide: 5¢-CTGTGGCCTGTCAGG ACCCTGTCCAGGAGT-3¢. The reverse complementary oligonucleotide was used as a negative control for hybridization. This oligonucleotide was labeled using T4 polynucleotide kinase from Invitrogen Inc. (Chicago, IL) following a protocol from the manufacturer. Acknowledgements Tissues for this study were obtained from the Alzhei- mer’s Disease Research Center Neuropathology Core, University of Southern California School of Medicine, Los Angeles, CA, USA. This work was supported by PHS grant # AG16256 to K.J.A.D and G.E. from the NIH ⁄ NIA. References 1 Fuentes JJ, Pritchard MA, Planas AM, Bosch A, Ferrer I & Estivill X (1995) A new human gene from the Down syndrome critical region encodes a proline-rich protein highly expressed in fetal brain and heart. Human Mol Genet 4, 1935–1944. 2 Crawford DR, Leahy KP, Abramova N, Lan L, Wang Y & Davies KJ (1997) Hamster adapt78 mRNA is a Down syndrome critical region homologue that is inducible by oxidative stress. Arch Biochem Biophysics 342, 6–12. 3 Kingsbury TJ & Cunningham KW (2000) A conserved family of calcineurin regulators. Genes Dev 14, 1595– 1604. 4 Chang KT, Shi YJ & Min KT (2003) The Drosophila homolog of Down’s syndrome critical region 1 gene regulates learning: implications for mental retardation. Proc Natl Acad Sci USA 100, 15794–15799. 5 Fuentes JJ, Genesca L, Kingsbury TJ, Cunningham KW, Perez-Riba M, Estivill X & de la Luna S (2000) DSCR1, overexpressed in Down syndrome, is an inhibi- tor of calcineurin-mediated signaling pathways. Human Mol Genet 9, 1681–1690. 6 Rothermel B, Vega RB, Yang J, Wu H, Bassel-Duby R & Williams RS (2000) A protein encoded within the Down syndrome critical region is enriched in striated muscles and inhibits calcineurin signaling. J Biol Chem 275, 8719–8725. 7 Gorlach J, Fox DS, Cutler NS, Cox GM, Perfect JR & Heitman J (2000) Identification and characterization of a highly conserved calcineurin binding protein, CBP1 ⁄ calcipressin, in Cryptococcus neoformans. EMBO J 19, 3618–3629. 8 Ejima A, Tsuda M, Takeo S, Ishii K, Matsuo T & Aigaki T (2004) Expression level of sarah, a homolog of DSCR1, is critical for ovulation and female courtship behavior in Drosophila melanogaster. Genetics 168, 2077–2087. 9 Ermak G, Morgan TE & Davies KJ (2001) Chronic overexpression of the calcineurin inhibitory gene DSCR1 (Adapt78) is associated with Alzheimer’s dis- ease. J Biol Chem 276, 38787–38794. 10 Rothermel BA, McKinsey TA, Vega RB, Nicol RL, Mammen P, Yang J, Antos CL, Shelton JM, Bassel-Duby R, Olson EN et al. (2001) Myocyte-enriched calcineurin- interacting protein, MCIP1, inhibits cardiac hypertrophy in vivo. Proc Natl Acad Sci USA 98, 3328–3333. 11 Minami T, Horiuchi K, Miura M, Abid MR, Takabe W, Noguchi N, Kohro T, Ge X, Aburatani H, Hamakubo T et al. (2004) Vascular endothelial growth factor- and thrombin-induced termination factor, Down syndrome critical region-1, attenuates endothelial cell proliferation and angiogenesis. J Biol Chem 279, 50537–50554. 12 Ermak G, Harris CD & Davies KJ (2002) The DSCR1 (Adapt78) isoform 1 protein calcipressin 1 inhibits calci- neurin and protects against acute calcium-mediated stress damage, including transient oxidative stress. Faseb J 16, 814–824. 13 Vega RB, Yang J, Rothermel BA, Bassel-Duby R & Williams RS (2002) Multiple domains of MCIP1 contri- bute to inhibition of calcineurin activity. J Biol Chem 277, 30401–30407. 14 Lin HY, Michtalik HJ, Zhang S, Andersen TT, Van Riper DA, Davies KK, Ermak G, Petti LM, Nachod S, Narayan AV et al. (2003) Oxidative and calcium stress regulate DSCR1 (Adapt78 ⁄ MCIP1) protein. Free Radic Biol Med 35, 528–539. 15 Genesca L, Aubareda A, Fuentes JJ, Estivill X, De La Luna S & Perez-Riba M (2003) Phosphorylation of cal- cipressin 1 increases its ability to inhibit calcineurin and decreases calcipressin half-life. Biochem J 374, 567–575. 16 Hilioti Z, Gallagher DA, Low-Nam ST, Ramaswamy P, Gajer P, Kingsbury TJ, Birchwood CJ, Levchenko A & Cunningham KW (2004) GSK-3 kinases enhance calci- neurin signaling by phosphorylation of RCNs. Genes Dev 18, 35–47. 17 Ermak G, Cheadle C, Becker KG, Harris CD & Davies KJ (2004) DSCR1 (Adapt78) modulates expression of SOD1. FASEB J 18, 62–69. 18 Cross DA, Alessi DR, Cohen P, Andjelkovich M & Hemmings BA (1995) Inhibition of glycogen synthase RCAN1 stimulates GSK-3b G. Ermak et al. 2108 FEBS Journal 273 (2006) 2100–2109 ª 2006 The Authors Journal compilation ª 2006 FEBS kinase-3 by insulin mediated by protein kinase B. Nature 378, 785–789. 19 Yamaguchi H, Ishiguro K, Uchida T, Takashima A, Lemere CA & Imahori K (1996) Preferential labeling of Alzheimer neurofibrillary tangles with antisera for tau protein kinase (TPK) I ⁄ glycogen synthase kinase-3 beta and cyclin-dependent kinase 5, a component of TPK II. Acta Neuropathol (Berlin) 92, 232–241. 20 Pei JJ, Braak E, Braak H, Grundke-Iqbal I, Iqbal K, Winblad B & Cowburn RF (1999) Distribution of active glycogen synthase kinase 3beta (GSK-3beta) in brains staged for Alzheimer disease neurofibrillary changes. J Neuropathol Exp Neurol 58, 1010–1019. 21 Evans DA, Funkenstein HH, Albert MS, Scherr PA, Cook NR, Chown MJ, Hebert LE, Hennekens CH & Taylor JO (1989) Prevalence of Alzheimer’s disease in a community population of older persons: higher than previously reported. JAMA 262, 2551–2556. 22 Heyman A, Wilkinson WE, Stafford JA, Helms MJ, Sig- mon AH & Weinberg T (1984) Alzheimer’s disease: a study of epidemiological aspects. Ann Neurol 15, 335–341. 23 Mortimer JA, French LR, Hutton JT & Schuman LM (1985) Head injury as a risk factor for Alzheimer’s dis- ease. Neurology 35, 264–267. 24 Guo Z, Cupples LA, Kurz A, Auerbach SH, Volicer L, Chui H, Green RC, Sadovnick AD, Duara R, DeCarli C et al. (2000) Head injury and the risk of AD in the MIRAGE study. Neurology 54, 1316–1323. 25 Plassman BL, Havlik RJ, Steffens DC, Helms MJ, Newman TN, Drosdick D, Phillips C, Gau BA, Welsh- Bohmer KA, Burke JR et al. (2000) Documented head injury in early adulthood and risk of Alzheimer’s disease and other dementias. Neurology 55, 1158–1166. 26 Wang Y, De Keulenaer GW, Weinberg EO, Muangman S, Gualberto A, Landschulz KT, Turi TG, Thompson JF & Lee RT (2002) Direct biomechanical induction of endogenous calcineurin inhibitor Down Syndrome Criti- cal Region-1 in cardiac myocytes. Am J Physiol Heart Circ Physiol 283, H533–H539. 27 Crabtree GR & Olson EN (2002) NFAT signaling: choreographing the social lives of cells. Cell 109 Suppl., S67–S79. 28 Groth RD & Mermelstein PG (2003) Brain-derived neu- rotrophic factor activation of NFAT (nuclear factor of activated T-cells)-dependent transcription: a role for the transcription factor NFATc4 in neurotrophin-mediated gene expression. J Neurosci 23, 8125–8134. 29 Antos CL, McKinsey TA, Frey N, Kutschke W, McAnally J, Shelton JM, Richardson JA, Hill JA & Olson EN (2002) Activated glycogen synthase-3 beta suppresses cardiac hypertrophy in vivo. Proc Natl Acad Sci USA 99, 907–912. 30 Doble BW & Woodgett JR (2003) GSK-3: tricks of the trade for a multi-tasking kinase. J Cell Sci 116, 1175– 1186. 31 Rusnak F & Mertz P (2000) Calcineurin: form and function. Physiol Rev 80, 1483–1521. 32 Billingsley ML & Kincaid RL (1997) Regulated phos- phorylation and dephosphorylation of tau protein: effects on microtubule interaction, intracellular traf- ficking and neurodegeneration. Biochem J 323, 577– 591. G. Ermak et al. RCAN1 stimulates GSK-3b FEBS Journal 273 (2006) 2100–2109 ª 2006 The Authors Journal compilation ª 2006 FEBS 2109 . kDa (RCAN1- 1S), RCAN1- 1 Long at 38 kDa (RCAN1- 1 L), and RCAN1- 4. Furthermore, the levels of RCAN1- 1S, but not RCAN1- 1 L or RCAN1- 4 correlated with the levels of GSK-3b. This suggests that RCAN1- 1S might induce. tightly correlated ( r ¼ 0.7, P < 0.001) with the expression of only one RCAN1 isoform, RCAN1- 1S, but not RCAN1- 1L or RCAN1- 4. This correlation was inde- pendent of the brain region analyzed. RCAN1- 1S. designed against isoform 1 recognizes RCAN1s of about 31 kDa (RCAN1- 1S) and 38 kDa (RCAN1- 1L). (B) The antibody designed against isoform 4 recognizes an RCAN1 protein of about 70 kDa (RCAN1- 4). (C) The