A Calmodulin-Binding Site on Cyclin E Mediates Ca2+-Sensitive G1/S Transitions in Vascular Smooth Muscle Cells
http://www.100md.com
Jaehyun Choi, Andrew Chiang, Nicolas Tau
参见附件。
the Heart & Stroke Richard Lewar Centre of Excellence in Cardiovascular Research and the Faculty of Medicine (J.C., A.C., A.P., M.H.)
Departments of Medicine, Physiology
Laboratory Medicine and Pathobiology (J.C., R.G., M.H.)
Department of Pharmaceutical Sciences, Leslie Dan Faculty of Pharmacy (N.T.)
McLaughlin Centre for Molecular Medicine (M.H.), University of Toronto, Ontario
Division of Cell and Molecular Biology (M.H.), Toronto General Hospital Research Institute, Ontario, Canada.
Abstract
Calcium transients are known to control several transition points in the eukaryotic cell cycle. For example, we have previously shown that a coordinate elevation in the intracellular free calcium ion concentration is required for G1- to S-phase cell cycle progression in vascular smooth muscle cells (VSMC). However, the molecular basis for this Ca2+ sensitivity was not known. Using buffers with differing [Ca2+], we found that the kinase activity of mouse and human cyclin E/CDK2, but not other G1/S-associated cell cycle complexes, was responsive to physiological changes in [Ca2+]. We next determined that this Ca2+-responsive kinase activity was dependent on a direct interaction between calmodulin (CaM), one of the major Ca2+-signal transducers of eukaryotic cells, and cyclin E. Pharmacological inhibition of CaM abrogated the Ca2+ sensitivity of cyclin E/CDK2 and retarded mouse VSMC proliferation by causing G1 arrest. We next defined the presence of a highly conserved 22 amino acid N-terminal CaM-binding motif in mammalian cyclin E genes (dissociation constant, 1.5±0.1 μmol/L) and showed its essential role in mediating Ca2+-sensitive kinase activity of cyclin E/CDK2. Mutant human cyclin E protein, lacking this CaM-binding motif, was incapable of binding CaM or responding to [Ca2+]. Taken together, these findings reveal CaM-dependent cyclin E/CDK2 activity as a mediator of the known Ca2+ sensitivity of the G1/S transition of VSMC.
Key Words: calcium calmodulin cell cycle cell cycle progression CDK2 cyclin E vascular smooth muscle cells
Introduction
Vascular smooth muscle cells (VSMC) are the end-effector organs in vasculoproliferative diseases such as atherosclerosis, restenosis, and hypertension.1 Elucidating the molecular mechanisms underlying the growth of VSMC may help to design novel therapies aimed at preventing and treating these conditions.
Many studies have implicated Ca2+ as a regulator of distinct "checkpoints" in the cell cycle of eukaryotes.2 Mammalian cells appear most sensitive to the depletion of extracellular Ca2+ at 2 points, in early G1 and near the G1/S boundary.3 In several cell types, DNA synthesis could be completely inhibited when extracellular Ca2+ was chelated by EGTA or when the influx of extracellular Ca2+ was blocked by cobalt.4,5 Specifically for VSMC, it has been reported that depletion of Ca2+ stores in the G1 phase of VSMC results in a profound G1 arrest that is not overcome until internal Ca2+ stores are replenished.6,7 We also found, in cultured VSMC, that intracellular Ca2+ concentrations ([Ca2+]i) exhibit significant increases as a function of cell cycle, such that resting free [Ca2+]i vary from 75 nmol/L at G0 to 150 nmol/L at G1/S, whereas sarcoendoplasmic reticulum Ca2+ ATPase (SERCA)-maintained releasable [Ca2+]i vary from 100 nmol/L at G0 to 500 nmol/L at G1/S.8–10 Importantly, preventing increases in resting and releasable [Ca2+]i inhibited G1- to S-phase transitions.8,9 Similarly, SERCA inactivation by thapsigargin has been shown to lengthen the G1 phase of HEK 293 cells.11
Calmodulin (CaM), a small acidic protein of 16.7 kDa, is an important transducer of Ca2+ signals in eukaryotic cells. It is recruited by a large number of proteins, collectively termed CaM-binding proteins (CaMBPs), through which Ca2+ sensitivity is expressed in a variety of cell biological functions.12 The role of CaM as a regulator of cell cycle progression has also been well established.2 Early indications came from experiments in which the mitotic cycle was arrested by anti-CaM drugs added to proliferating cells or to cells reentering cell cycle under the influence of mitogens.13 In addition, it was reported that monoclonal antibodies against CaM inhibited the synthesis of DNA in permeabilized cells.14 Moreover, progression through G1 and mitosis exit was observed to be sensitive to changes in the intracellular concentration of CaM.15 Although others have suggested a role for CaM in the G1/S transition through the activation or expression of proteins involved in DNA replication,16,17 and the phosphorylation of the retinoblastoma protein (Rb),18 the molecular and cell biological mechanisms through which Ca2+ and or CaM exerted these effects remained obscure.
Thus, the primary objectives of our study were to identify which cell cycle proteins expressed in VSMC were responsive to changes in [Ca2+], and whether their Ca2+ sensitivity was mediated by CaM, and how. Given that (1) cyclin E/CDK2 is an enzyme complex that is both rate limiting and essential for S phase entry19,20 and (2) cyclin E synthesis begins in G1 and reaches a peak at the G1/S transition,21,22 we hypothesized that the cyclin E/CDK2 complex may mediate Ca2+ sensitivity of the G1/S cell cycle transition of VSMC. The current study reports that this is the case and that specific CaM-cyclin E interactions underlie this.
Materials and Methods
Materials and Cell Culture
Purified active cyclin E/CDK2 and cyclin A/CDK2 were obtained from Upstate Biotechnology (Lake Placid, NY). CDK2 inhibitor II, KN-93, and cyclosporin A were obtained from Calbiochem (La Jolla, Calif). Calmidazolium, dithiothreitol (DTT), aprotinin, leupeptin, and other chemicals were purchased from Sigma (St Louis, Mo). Synthetic peptides were purchased from GenScript Corp (Piscataway, NJ).
The isolation and culture of primary mouse aortic VSMC and the characterization of MOVAS, a mouse VSMC line, have been described.10 MOVAS were grown in DMEM (GIBCO/BRL, Gaithersburg, Md) supplemented with 10% FBS (Hyclone, Logan, Utah) and 1% penicillin-streptomycin (GIBCO/BRL). For serum starvation, cells were grown to 60% to 70% confluence, washed twice with PBS, and cultured in starvation medium (DMEM without FBS) for 48 hours to achieve G0 arrest. Fresh DMEM with 10% FBS was used to initiate reentry into cell cycle, and cells were incubated for 16 hours to allow them to approach the G1/S transition or for 20 hours for S-phase synchronization. Primary mouse aortic SMC were grown in media supplemented with platelet-derived growth factor (PDGF) (50 ng/mL). G0 synchronization in these cells was achieved with 72 hours of 0.25% FBS-supplemented DMEM. Progression to G1/S was brought about by stimulation with 10% FBS and 50 ng/mL PDGF for 24 hours.
Immunoprecipitation and Kinase Assays
SMC (5x106) were harvested and resuspended with 2 mL of lysis buffer (50 mmol/L Tris [pH 7.4], 250 mmol/L NaCl, 5 mmol/L EDTA, 0.1% NP-40, 0.5 mmol/L DTT, 0.1 mmol/L Na3VO4, 5 μg/mL leupeptin, 5 μg/mL aprotinin, 2 mmol/L PMSF, and 10% glycerol). After homogenization, cell lysates were clarified by centrifugation at 12x103 rpm, 4°C for 30 minutes, and the supernatant was collected. An aliquot was taken for protein concentration determination using the BCA protein assay kit (Sigma). Cell extracts (200 μg) were then incubated for 2 hours at 4°C with saturating concentrations of polyclonal antibodies (Abs). Immune complexes were collected by incubation with GammaBind G Sepharose resin (Amersham Pharmacia, Piscataway, NJ) for 1 hour at 4°C. The beads were then washed three times with washing buffer (50 mmol/L Tris [pH 7.4] and 1 mmol/L DTT).
For the in vitro kinase assays, 20 μL of kinase reaction mixture (20 mmol/L Tris [pH 7.4], 5 mmol/L MgCl2, 2.5 mmol/L MnCl2, 1 mmol/L DTT, 10 μg of C terminal of human Rb protein or 4 μg of histone–H1 [both Upstate Biotechnology] as substrates, 20 μmol/L ATP, and 2.4 μCi [- 32P]ATP [Amersham Pharmacia]) was added to CDK2-, cyclin E–, or cyclin A–immunoprecipitated (IPd) complexes after removing the washing buffer from GammaBind G Sepharose resin. Based on the total calcium content of kinase reaction mixtures (as determined by inductively coupled plasma atomic emission spectrometry), amounts of EGTA required to yield final concentrations of free Ca2+ ion (0, 100 or 500 nmol/L) were calculated by Winmol/LaxC (http://www.stanford.edu/cpatton/winmaxc2.html). After 30 minutes of incubation at 37°C, reactions were stopped by adding 20 μL of 2x sodium dodecyl sulfate (SDS) loading buffer and heating at 100°C for 5 minutes. Labeled proteins were resolved by 16% SDS-PAGE. Phosphorylated Rb or histone–H1 bands were visualized by autoradiography and quantified in a Scintillation Counter LS6500 (Beckman Coulter, Fullerton, Calif) after being cut from the gel.
Western Blot
Between 20 to 30 μg of cell extracts were resolved by 12% SDS-PAGE and transferred to nitrocellulose membranes (Sigma, 0.2 μmol/L pore size). Blots were blocked with 5% nonfat dry milk in TBS-T (10 mmol/L Tris [pH 8.0], 150 mmol/L NaCl, and 0.05% Tween-20) overnight at 4°C and then incubated for 3 hours at 25°C in TBS-T plus 3% nonfat dry milk containing primary Abs, including anti-cyclin E and -cyclin A (Upstate Biotechnology), -phospho-CDK2 at Thr196 (Cell Signaling Technology, Danvers, MA), -CaM, -His probe, and -actin (Santa Cruz Biotechnology, Santa Cruz, Calif). Protein bands were detected with ECL Reagents (Amersham Pharmacia), with horseradish peroxidase–conjugated secondary goat anti-rabbit IgG (Santa Cruz Biotechnology). Quantification of band intensity was performed using Quantity One (Bio-Rad, Hercules, Calif).
CaM-Binding Analysis
Purified cyclin E/CDK2 and cyclin A/CDK2 (100 ng) (Upstate Biotechnology) complexes were diluted with Ca2+-binding buffer (50 mmol/L Tris [pH 7.5], 100 mmol/L NaCl, and 2 mmol/L CaCl2) into a final 20 μL and incubated with an equal volume of preequilibrated CaM–Sepharose 4B resin (Amersham Pharmacia) for 2 hours at 4°C. After centrifugation, the supernatant containing unbound proteins was removed and the resin was washed twice with the buffer used in the binding step. After removing buffer completely from the resin, SDS loading buffer with DTT was added to the dried resin to collect all proteins bound to CaM. The proteins were resolved by 12% SDS-PAGE and visualized by silver staining to determine the amount bound to resin-immobilized CaM.
Confocal Immunofluorescence
Methods and representative images used for analysis of the extent of CaM and cyclin E colocalization are shown in the online data supplement available at http://circres.ahajournals.org.
Flow Cytometric Analysis
Samples were prepared as previously described with minor modifications.8 Cells were counted in a flow cytometer (FACScan; BD Biosciences, Franklin Lakes, NJ), and G0/G1, S, and G2/M cell percentages were calculated with Cell Quest software (BD Biosciences).
Fluorescence Titration
The methods used to determine the dissociation constant (Kd) for peptide–protein binding are detailed in the online data supplement.
Transfection
pIRESpuro3 plasmid vectors (Clontech, Palo Alto, Calif) harboring wild-type (wt) and N-terminal–deleted mutant (N) cyclin E were obtained by PCR amplification of a pECE human cyclin E plasmid (a generous gift of Dr Paul Hamel, University of Toronto) with oligonucleotides 5'-ccggaattcatgaaggaggacggcg-3' (forward wt cyclin E), 5'-ccggaattcatgccagatgaagaaatgg-3' (forward N cyclin E), and 5'-ccggaattctcagtggtggtggtggtggtgcgccatttccggccc-3' (reverse for both). DNA sequences of each construct were confirmed by sequencing. Transfections were performed with the Lipofectamine 2000 reagent (Invitrogen, Carlsbad, Calif) according to the instructions of the manufacturer. Starting 48 hours posttransfection, MOVAS were cultured in the presence of 5 μg/mL puromycin for 2 weeks.
Statistics
ANOVA, Student t test, and coefficients of variance tests were used as appropriate. Analyses were performed on SPSS version 13.0 (Chicago, Ill).
Results
Cyclin E/CDK2 Activity at the G1/S Transition of VSMC Is Ca2+ Sensitive
Kinase activities of cyclin E–IPd complexes from G0- and G1/S-synchronized MOVAS were analyzed in different [Ca2+]. We observed, at 0 nmol/L [Ca2+], that histone–H1 kinase activity at G1/S was greater than at G0 (2.4-fold; Figure 1A). These data suggested a G1/S-associated increase in cyclin E/CDK2 activity, which has been reported previously.23 Our first novel finding was that an increase in reaction buffer [Ca2+] from 0 to 500 nmol/L consistently increased histone–H1 kinase–defined activity of cyclin E–IPd complexes from G1/S- (2.6-fold) but not G0-synchronized VSMC (Figure 1A). Identical results were obtained when the experiment was performed using CDK2-IPd complexes (G1/S: 2.7-fold; Figure 1B). Experiments with CDK2 inhibitor II showed that the histone–H1 kinase–defined activity of cyclin E immunoprecipitates (IPs) was CDK2-dependent (Figure 1C). Of note, Rb kinase-defined activity of cyclin E IPs was also Ca2+ responsive only at G1/S (Figure 1D). In addition, histone–H1 kinase activity of cyclin E–IPd complexes from primary mouse aortic VSMC exhibited the same Ca2+ responsiveness as those from MOVAS (G1/S: 2.3-fold; Figure 1E). As a next step, we determined that this enhancement was specific to an increased [Ca2+] and not other divalent cations (Figure 1F).
Unlike cyclin E/CDK2, cyclin D/CDK4 and cyclin A/CDK2 exhibited different properties. Neither cyclin D–IPd complexes from early G1 phase MOVAS nor cyclin A–IPd complexes from S-phase MOVAS (Figure 1G) showed Ca2+ responsive increases in histone–H1 kinase activity. In addition, purified cyclin A/CDK2 complexes did not exhibit the Ca2+ sensitivity observed with purified cyclin E/CDK2 (Figure 1H). Of interest, the ability of Ca2+ to enhance histone–H1 kinase activity of purified cyclin E/CDK2 complexes (1.7-fold; Figure 1H) was less than that observed in IPs of cell extracts (>2.5-fold; Figure 1A through 1F), suggesting perhaps that factors other than those present in "purified" complexes participate in their Ca2+ sensitivity.
Taken together, the above data showed that the activity of cyclin E/CDK2 on 2 important substrates (Rb and histone–H1) involved in the G1/S transition is Ca2+ sensitive within the physiological range of [Ca2+]i known to accompany this stage of the cell cycle in VSMC.8–10 These data implicated cyclin E/CDK2 as a potential mediator of the known Ca2+ sensitivity of the G1/S transition of both primary mouse VSMC and the MOVAS cell line.
CaM and, to a Limited Extent, Calcineurin, but Not CaM-Activated Kinases, Transduce Ca2+-Sensitive Cyclin E/CDK2 Activity in VSMC
We next explored what mediated the Ca2+ sensitivity of cyclin E/CDK2. A potential role for CaM, among the major Ca2+ signal–transducing proteins in eukaryotes, was examined first. When added to cell extracts, the potent CaM inhibitor calmidazolium inhibited cyclin E/CDK2 activity in a dose-dependent manner (Figure 2A), suggesting a role for CaM in regulating cyclin E/CDK2 activity. As known downstream mediators of CaM-regulated cell cycle progression in various cell types, the potential roles of multifunctional CaM-activated kinases (CaMKs) and calcineurin were tested next. KN-93, an inhibitor of several CaMKs,24 did not affect H1 kinase–defined cyclin E/CDK2 activity (Figure 2B). Only high concentrations of cyclosporin A (50 μmol/L), an inhibitor of calcineurin,25 exhibited at most 20% inhibition of cyclin E/CDK2 activity (Figure 2C).
To test directly whether CaM and calcineurin are necessary for Ca2+ sensitivity of cyclin E/CDK2 activity, immunodepletion experiments were performed. To deplete CaM or calcineurin from cell extracts, immunoprecipitation (IP) was first performed with excess amounts of anti-CaM or anti-calcineurin Abs, followed by a second IP with anti-cyclin E Ab on CaM- or calcineurin-depleted cell extracts. CaM-depleted cell extracts did not exhibit any Ca2+-mediated stimulation of cyclin E/CDK2 activity, suggesting that CaM is essential for Ca2+ sensitivity of cyclin E/CDK2 activity (Figure 2D). By contrast, calcineurin-depleted cell extracts still exhibited some Ca2+ sensitivity (Figure 2D).
To determine whether CaM has a direct effect on cyclin E/CDK2 activity, purified CaM (Sigma; purity >95%) was titrated with a fixed amount of purified human cyclin E/CDK2. CaM increased cyclin E/CDK2 activity in a dose-dependent manner only in the presence of Ca2+, an effect that was abolished by calmidazolium and W-13 (another CaM antagonist) (Figure 2E), suggesting that CaM has a direct stimulatory effect on cyclin E/CDK2 activity. W-12, an analog of W-13 with much lower CaM affinity, did not mimic the effect of W-13 (Figure 2E). Together, the above data suggested that the stimulatory effect of CaM on cyclin E/CDK2 activity was mediated primarily by a direct Ca2+-dependent mechanism involving CaM and cyclin E/CDK2, rather than through CaMKs or calcineurin.
Interaction Between CaM and Cyclin E Is Essential for Ca2+-Sensitive Kinase Activity of Cyclin E/CDK2
To explore further the mechanisms underlying CaM-dependent Ca2+-sensitive cell cycle kinase activities in VSMC, we used a CaM-binding assay. Equal amounts of purified cyclin E/CDK2 and cyclin A/CDK2 were loaded on a CaM–Sepharose column, and silver-stained blots of bound and unbound fractions revealed binding between CaM and cyclin E but not between CaM and cyclin A (Figure 3A). With varying buffer [Ca2+], these experiments revealed that the interaction between CaM and cyclin E occurs in a Ca2+-dependent manner (data not shown). These results were entirely consistent with the observed difference in Ca2+-sensitivities between cyclin E/CDK2 and cyclin A/CDK2 (Figure 1H). Moreover, these data further supported a direct interaction between CaM and cyclin E, not CDK2, as an essential requirement for Ca2+ sensitivity of cyclin E/CDK2.
Co-IP analysis between CaM and cyclin E revealed that the amount of CaM bound to cyclin E increased at G1/S (Figure 3B; 1.68-fold). The absence of any CaM in IPs that used a nonspecific Ab (anti-actin) supported further the specificity of the interaction between CaM and cyclin E. In addition, subcellular localizations of CaM and cyclin E were visualized in MOVAS by fluorescence confocal microscopy, and the proportion of cells showing strong colocalization signals were calculated in both G0- and G1/S- synchronized populations (n=1000 cells each). This analysis suggested a 62% increase in the number of nuclei with strong CaM and cyclin E colocalization at G1/S versus G0 (378 versus 233; P<0.001; online data supplement), a result mirroring closely the observed increase (68%) in co-IP experiments (Figure 3B). This may be related to increased expression levels of CaM at G1/S (Figure 3B), a finding also reported by others,13,26 and/or the higher [Ca2+]i of G1/S-, as compared with G0-synchronized cells.8–10
Inhibition of CaM Retards the Proliferation of VSMC
Having demonstrated a putative molecular mechanism mediating Ca2+-sensitive G1/S transitions, we sought to directly test its physiological significance. Flow-assisted cell sorting analysis (FACS) of DNA content was performed to determine whether inhibition of CaM retards the cell cycle progression of VSMC. Whereas control experiments showed that 24.4% of primary mouse aortic SMC had entered S phase after 24 hours of serum stimulation, only 10.8% of cells treated with calmidazolium, and 12.6% of cells treated with W-13, had done so (P<0.001; Figure 4A). A similar level of cell cycle inhibition was also observed in MOVAS (online data supplement).
Western blot analysis in primary VSMC showed that the interaction between CaM and cyclin E was severely inhibited by calmidazolium (Figure 4B, lanes 5 and 6), whereas the amounts of cyclin E and CaM itself were not (Figure 4B, lanes 2 and 3). This result indicated that the observed decrease in cyclin E-CaM interaction is not simply attributable to the inhibition of cell cycle progression. Importantly, the level of Thr160-phosphorylated CDK2 (the active form of CDK2) was inhibited by calmidazolium treatment (Figure 4B, lanes 2 and 3).
Together, these data suggested (1) that CaM facilitates G1- to S-phase transitions in VSMC via binding to cyclin E and (2) that this interaction may play a role in activating CDK2.
Identification of the CaM-Binding Motif on Cyclin E
To provide a structural basis for our functional model, we next sought to identify the CaM-binding region on cyclin E. A CaM-target database (http://calcium.uhnres.utoronto.ca/ctdb)27 predicted a CaM-binding motif on human cyclin E with successively high probability scores (Figure 5A). This consisted of 22 amino acids near the N terminus of human cyclin E (amino acids 5 to 26), with 5 hydrophobic and 5 basic residues predicted to be critical for a hydrophobic interaction between CaM and cyclin E. We termed this motif the "calmodulin-binding sequence" (CBS) of cyclin E and noted that mouse and rat cyclin E genes also harbor this conserved sequence, whereas the Drosophila and Caenorhabditis elegans homologs do not (Figure 5B), a finding that may restrict the physiological relevance of this domain to mammals. Consistent with the results of our kinase activity assays, no similar CBS was revealed in alignment and sequence analyses of all cyclins D and A.
To confirm the ability of the CBS of cyclin E to bind CaM, we performed tyrosine residue–based fluorescence titration experiments using purified peptides in vitro. Figure 5C shows a representative binding profile of the CBS peptide to CaM. The observed decrease in the relative intensity indicates changes in the environment of any (or all) of the 3 tyrosine residues (2 in CaM and 1 in the CBS peptide) associated with the binding event. The fit of the fluorescence profile (see online data supplement for details) yielded a CBS-CaM dissociation constant (Kd, 1.5±0.1x10–6 mol/L) consistent with other CaM-binding proteins such as caldesmon (Kd, 0.23x10–6 mol/L) and calponin (Kd, 1.3x10–6 mol/L).28
N Cyclin E Neither Binds CaM nor Exhibits Ca2+ Sensitivity
To investigate further the structural basis and functional importance of CaM–cyclin E interactions, expression vectors harboring either His-tagged full length (wt) or His-tagged N mutant forms of human cyclin E (lacking amino acids 1 to 26) were stably transfected into MOVAS (Figure 6A). Despite equal levels of His-tagged wt and N cyclin E protein expression (Figure 6B, left panel), co-IP analyses on G1/S-synchronized cell populations indicated that only wt cyclin E could bind CaM (Figure 6B, right panel). Moreover, histone–H1 kinase assays performed on G1/S-synchronized cells revealed that despite an equivalent level of activity at 0 nmol/L [Ca2+] to wt cyclin E/CDK2, N cyclin E/CDK2 complexes did not exhibit any appreciable Ca2+-sensitive enhancement of function. By contrast, IPd wt human cyclin E exhibited an identical level of Ca2+ sensitivity as endogenous mouse cyclin E in vector only transfected cells (2.5-fold; Figure 6C).
Discussion
This study was focused on identifying the molecular mechanism(s) underlying the poorly understood Ca2+ sensitivity of the G1 to S-phase cell cycle transition of VSMC. The results presented here implicate cyclin E/CDK2 in this process and show that a Ca2+-dependent binding of CaM to cyclin E at G1/S is critical for G1/S- and Ca2+-specific enhancement of cyclin E/CDK2 activity. We also defined a CaM-binding motif in the human cyclin E gene and demonstrated its essential role in mediating Ca2+-sensitive CDK2 activity.
Although the expression of CDK2 in T lymphocytes was noted to be depressed by the CaM antagonist W-13,29 our work is the first to show that CaM increases CDK2 activity via binding to cyclin E. This conclusion was supported by various analyses using several independent approaches including Western blot, co-IP, immunohistochemistry, peptide binding, and structure–function studies with mutants. To investigate further, fluorescence spectroscopy was used to calculate Kd between the CBS peptide and CaM. The finding that N cyclin E was unable to bind to CaM (Figure 6B) strongly suggested that this CaM-binding motif in human cyclin E1 (amino acids 5 to 26) is the only area involved in CaM–cyclin E interactions. As such, the determined Kd obtained from CaM and the CBS peptide likely reflects the true dissociation constant between CaM and cyclin E. If the fluorescent residues examined were far from the binding site or were not involved in a binding-mediated conformational change, no changes in fluorescence would have been observed.30 Rather, our experiments clearly support a structural relationship between the CBS and CaM with a Kd consistent with other CaMBPs such as G protein–coupled receptor kinase 2,31 caldesmon, calponin,28 spectrin,32 and CaM-dependent adenylate cyclases.33,34
Although previous reports have suggested a role for CaM in cell cycle,2 specific molecular pathways through which CaM might exert these effects were not known. Kahl and Means implicated CaM in the regulation of cyclin D1/CDK4 through their discovery that CaMKI, and not CaMKII, was the KN-93–sensitive CaMK responsible for KN-93–induced G1 arrest in WI-38 fibroblasts.35 They reported also that CaM activates cyclin D1 protein synthesis in fibroblasts via calcineurin.36 Previously, Taules and colleagues had implicated CaM in the nuclear entry of cyclin D/CDK4 and identified a relevant interaction between CaM and the CDK4 inhibitor p21Cip1.37,38 In contrast to the work in other cell types, we did not find KN-93–sensitive CaMKs to be involved in mediating Ca2+/CaM sensitivity of cyclin E/CDK2 complexes and demonstrated a very modest effect of cyclosporin A and calcineurin immunodepletion. Rather, we have identified a physiologically relevant role for CaM at a later and arguably more critical cell cycle checkpoint in VSMC and are the first to directly demonstrate Ca2+ sensitivity of its target. Indeed, it is interesting to note that although the studies of Kahl and Means and Taules and colleagues focused on the role of CaM in regulating the early-to-mid G1 factors cyclin D/CDK4, we have examined the late G1- and S-phase factors cyclin E/CDK2. Collectively, the key emerging concept from these studies is that CaM mediates Ca2+ sensitivities of both the early and late G1- to S-phase checkpoints through its interactions with cyclin D/CDK4/p21cip1 and cyclin E/CDK2/p27kip1, respectively.
Although CaM-binding to cyclin E clearly increased the levels of Thr160-phosphorylated (ie, active) CDK2 in extracts from G1/S-synchronized VSMC (Figure 4B), we have yet to determine how precisely this occurs. Possible explanations include Ca2+/CaM-cyclin E- or Ca2+/CaM-dependent augmentation of (1) another kinase involved in CDK2 activation (namely CDK-activating kinase [CAK], reviewed by Morgan39) or (2) the release and/or degradation of the CDK2 inhibitor p27kip1 from the otherwise mute cyclin E/CDK2/p27kip1 complex. As preliminary data from our laboratory indicate that the binding of CaM to cyclin E does not cause the release of p27kip1 from cyclin E/CDK2 complex (data not shown), we suspect that the binding of CaM to cyclin E changes the conformation of the cyclin E/CDK2 complex, such that CDK2 can be more easily phosphorylated by CAK. Activation mechanisms of CaMBPs by CaM strongly support the concept that CaM activates its target proteins by altering their conformation.40 Further studies will be required to define precisely the mechanisms underlying this observation and the potential involvement of CAK in the Ca2+/CaM-cyclin E-dependent activation of CDK2.
Acknowledgments
J.C. was supported in part by doctoral student stipend awards from the Ontario Graduate Scholarship and Canadian Institute of Health Research (CIHR). M.H. was the recipient of a Clinician Scientist Award from the CIHR and currently holds a Career Investigator Award from the Heart & Stroke Foundation of Ontario (CI5503 to M.H.). This work was supported by a CIHR operating grant (MOP14648 to M.H.). We thank Drs Paul Hamel and Brenda Andrews (University of Toronto) for thoughtful discussions. We also acknowledge Dr Tigran Chalikian (University of Toronto) for supporting the fluorescence titration work.
Footnotes
Original received November 23, 2005; revision received March 15, 2006; accepted April 6, 2006.
References
Ross R. The pathogenesis of atherosclerosis: a perspective for the 1990s. Nature. 1993; 362: 801–809. [Order article via Infotrieve]
Kahl CR, Means AR. Regulation of cell cycle progression by calcium/calmodulin-dependent pathways. Endocr Rev. 2003; 24: 719–736.
Takuwa N, Zhou W, Takuwa Y. Calcium, calmodulin and cell cycle progression. Cell Signal. 1995; 7: 93–104. [Order article via Infotrieve]
Kojima I, Matsunaga H, Kurokawa K, Ogata E, Nishimoto I. Calcium influx: an intracellular message of the mitogenic action of insulin-like growth factor-I. J Biol Chem. 1988; 263: 16561–16567.
Tomono M, Toyoshima K, Ito M, Amano H. Calcineurin is essential for DNA synthesis in Swiss 3T3 fibroblasts. Biochem J. 1996; 317 (pt 3): 675–680. [Order article via Infotrieve]
Short AD, Bian J, Ghosh TK, Waldron RT, Rybak SL, Gill DL. Intracellular Ca2+ pool content is linked to control of cell growth. Proc Natl Acad Sci U S A. 1993; 90: 4986–4990.
Ghosh TK, Bian JH, Short AD, Rybak SL, Gill DL. Persistent intracellular calcium pool depletion by thapsigargin and its influence on cell growth. J Biol Chem. 1991; 266: 24690–24697.
Husain M, Bein K, Jiang L, Alper SL, Simons M, Rosenberg RD. c-Myb-dependent cell cycle progression and Ca2+ storage in cultured vascular smooth muscle cells. Circ Res. 1997; 80: 617–626.
Husain M, Jiang L, See V, Bein K, Simons M, Alper SL, Rosenberg RD. Regulation of vascular smooth muscle cell proliferation by plasma membrane Ca(2+)-ATPase. Am J Physiol. 1997; 272: C1947–C1959. [Order article via Infotrieve]
Afroze T, Yang LL, Wang C, Gros R, Kalair W, Hoque AN, Mungrue IN, Zhu Z, Husain M. Calcineurin-independent regulation of plasma membrane Ca2+ ATPase-4 in the vascular smooth muscle cell cycle. Am J Physiol Cell Physiol. 2003; 285: C88–C95.
Simon VR, Moran MF. SERCA activity is required for timely progression through G1/S. Cell Prolif. 2001; 34: 15–30. [Order article via Infotrieve]
van Eldik LJ, Watterson DM. Calmodulin & Signal Transduction. New York: Academic Press; 1998.
Chafouleas JG, Bolton WE, Hidaka H, Boyd AE 3rd, Means AR. Calmodulin and the cell cycle: involvement in regulation of cell-cycle progression. Cell. 1982; 28: 41–50. [Order article via Infotrieve]
Reddy GP, Reed WC, Sheehan E, Sacks DB. Calmodulin-specific monoclonal antibodies inhibit DNA replication in mammalian cells. Biochemistry. 1992; 31: 10426–10430. [Order article via Infotrieve]
Rasmussen CD, Means AR. Calmodulin is required for cell-cycle progression during G1 and mitosis. EMBO J. 1989; 8: 73–82. [Order article via Infotrieve]
Lopez-Girona A, Bachs O, Agell N. Calmodulin is involved in the induction of DNA polymerases alpha and delta activities in normal rat kidney cells activated to proliferate. Biochem Biophys Res Commun. 1995; 217: 566–574. [Order article via Infotrieve]
Lopez-Girona A, Bosch M, Bachs O, Agell N. Addition of calmodulin antagonists to NRK cells during G1 inhibits proliferating cell nuclear antigen expression. Cell Calcium. 1995; 18: 30–40. [Order article via Infotrieve]
Takuwa N, Zhou W, Kumada M, Takuwa Y. Ca(2+)-dependent stimulation of retinoblastoma gene product phosphorylation and p34cdc2 kinase activation in serum-stimulated human fibroblasts. J Biol Chem. 1993; 268: 138–145.
Dulic V, Lees E, Reed SI. Association of human cyclin E with a periodic G1-S phase protein kinase. Science. 1992; 257: 1958–1961.
Lundberg AS, Weinberg RA. Functional inactivation of the retinoblastoma protein requires sequential modification by at least two distinct cyclin-cdk complexes. Mol Cell Biol. 1998; 18: 753–761.
Clurman BE, Sheaff RJ, Thress K, Groudine M, Roberts JM. Turnover of cyclin E by the ubiquitin-proteasome pathway is regulated by cdk2 binding and cyclin phosphorylation. Genes Dev. 1996; 10: 1979–1990.
Lew DJ, Dulic V, Reed SI. Isolation of three novel human cyclins by rescue of G1 cyclin (Cln) function in yeast. Cell. 1991; 66: 1197–1206. [Order article via Infotrieve]
Koff A, Giordano A, Desai D, Yamashita K, Harper JW, Elledge S, Nishimoto T, Morgan DO, Franza BR, Roberts JM. Formation and activation of a cyclin E-cdk2 complex during the G1 phase of the human cell cycle. Science. 1992; 257: 1689–1694.
Hook SS, Means AR. Ca(2+)/CaM-dependent kinases: from activation to function. Annu Rev Pharmacol Toxicol. 2001; 41: 471–505. [Order article via Infotrieve]
Liu J, Farmer JD Jr, Lane WS, Friedman J, Weissman I, Schreiber SL. Calcineurin is a common target of cyclophilin-cyclosporin A and FKBP-FK506 complexes. Cell. 1991; 66: 807–815. [Order article via Infotrieve]
Chafouleas JG, Lagace L, Bolton WE, Boyd AE 3rd, Means AR. Changes in calmodulin and its mRNA accompany reentry of quiescent (G0) cells into the cell cycle. Cell. 1984; 36: 73–81. [Order article via Infotrieve]
Yap KL, Kim J, Truong K, Sherman M, Yuan T, Ikura M. Calmodulin target database. J Struct Funct Genomics. 2000; 1: 8–14. [Order article via Infotrieve]
Medvedeva MV, Kolobova EA, Wang P, Gusev NB. Interaction of proteolytic fragments of calmodulin with caldesmon and calponin. Biochem J. 1996; 315 (pt 3): 1021–1026. [Order article via Infotrieve]
Colomer J, Lopez-Girona A, Agell N, Bachs O. Calmodulin regulates the expression of cdks, cyclins and replicative enzymes during proliferative activation of human T lymphocytes. Biochem Biophys Res Commun. 1994; 200: 306–312. [Order article via Infotrieve]
Lakowicz JR. Principles of Fluorescence Spectroscopy. 2nd ed. New York: Kluwer Academic/Plenum publishers; 1999.
Pronin AN, Satpaev DK, Slepak VZ, Benovic JL. Regulation of G protein-coupled receptor kinases by calmodulin and localization of the calmodulin binding domain. J Biol Chem. 1997; 272: 18273–18280.
Bjork J, Lundberg S, Backman L. Characterization of the binding of calmodulin to nonerythroid spectrin. Eur J Cell Biol. 1995; 66: 200–204. [Order article via Infotrieve]
Cali JJ, Parekh RS, Krupinski J. Splice variants of type VIII adenylyl cyclase. Differences in glycosylation and regulation by Ca2+/calmodulin. J Biol Chem. 1996; 271: 1089–1095.
Choi EJ, Wong ST, Hinds TR, Storm DR. Calcium and muscarinic agonist stimulation of type I adenylylcyclase in whole cells. J Biol Chem. 1992; 267: 12440–12442.
Kahl CR, Means AR. Regulation of cyclin D1/Cdk4 complexes by calcium/calmodulin-dependent protein kinase I. J Biol Chem. 2004; 279: 15411–15419.
Kahl CR, Means AR. Calcineurin regulates cyclin D1 accumulation in growth-stimulated fibroblasts. Mol Biol Cell. 2004; 15: 1833–1842.
Taules M, Rius E, Talaya D, Lopez-Girona A, Bachs O, Agell N. Calmodulin is essential for cyclin-dependent kinase 4 (Cdk4) activity and nuclear accumulation of cyclin D1-Cdk4 during G1. J Biol Chem. 1998; 273: 33279–33286.
Taules M, Rodriguez-Vilarrupla A, Rius E, Estanyol JM, Casanovas O, Sacks DB, Perez-Paya E, Bachs O, Agell N. Calmodulin binds to p21(Cip1) and is involved in the regulation of its nuclear localization. J Biol Chem. 1999; 274: 24445–24448.
Morgan DO. Principles of CDK regulation. Nature. 1995; 374: 131–134. [Order article via Infotrieve]
Hoeflich KP, Ikura M. Calmodulin in action: diversity in target recognition and activation mechanisms. Cell. 2002; 108: 739–742. [Order article via Infotrieve]
the Heart & Stroke Richard Lewar Centre of Excellence in Cardiovascular Research and the Faculty of Medicine (J.C., A.C., A.P., M.H.)
Departments of Medicine, Physiology
Laboratory Medicine and Pathobiology (J.C., R.G., M.H.)
Department of Pharmaceutical Sciences, Leslie Dan Faculty of Pharmacy (N.T.)
McLaughlin Centre for Molecular Medicine (M.H.), University of Toronto, Ontario
Division of Cell and Molecular Biology (M.H.), Toronto General Hospital Research Institute, Ontario, Canada.
Abstract
Calcium transients are known to control several transition points in the eukaryotic cell cycle. For example, we have previously shown that a coordinate elevation in the intracellular free calcium ion concentration is required for G1- to S-phase cell cycle progression in vascular smooth muscle cells (VSMC). However, the molecular basis for this Ca2+ sensitivity was not known. Using buffers with differing [Ca2+], we found that the kinase activity of mouse and human cyclin E/CDK2, but not other G1/S-associated cell cycle complexes, was responsive to physiological changes in [Ca2+]. We next determined that this Ca2+-responsive kinase activity was dependent on a direct interaction between calmodulin (CaM), one of the major Ca2+-signal transducers of eukaryotic cells, and cyclin E. Pharmacological inhibition of CaM abrogated the Ca2+ sensitivity of cyclin E/CDK2 and retarded mouse VSMC proliferation by causing G1 arrest. We next defined the presence of a highly conserved 22 amino acid N-terminal CaM-binding motif in mammalian cyclin E genes (dissociation constant, 1.5±0.1 μmol/L) and showed its essential role in mediating Ca2+-sensitive kinase activity of cyclin E/CDK2. Mutant human cyclin E protein, lacking this CaM-binding motif, was incapable of binding CaM or responding to [Ca2+]. Taken together, these findings reveal CaM-dependent cyclin E/CDK2 activity as a mediator of the known Ca2+ sensitivity of the G1/S transition of VSMC.
Key Words: calcium calmodulin cell cycle cell cycle progression CDK2 cyclin E vascular smooth muscle cells
Introduction
Vascular smooth muscle cells (VSMC) are the end-effector organs in vasculoproliferative diseases such as atherosclerosis, restenosis, and hypertension.1 Elucidating the molecular mechanisms underlying the growth of VSMC may help to design novel therapies aimed at preventing and treating these conditions.
Many studies have implicated Ca2+ as a regulator of distinct "checkpoints" in the cell cycle of eukaryotes.2 Mammalian cells appear most sensitive to the depletion of extracellular Ca2+ at 2 points, in early G1 and near the G1/S boundary.3 In several cell types, DNA synthesis could be completely inhibited when extracellular Ca2+ was chelated by EGTA or when the influx of extracellular Ca2+ was blocked by cobalt.4,5 Specifically for VSMC, it has been reported that depletion of Ca2+ stores in the G1 phase of VSMC results in a profound G1 arrest that is not overcome until internal Ca2+ stores are replenished.6,7 We also found, in cultured VSMC, that intracellular Ca2+ concentrations ([Ca2+]i) exhibit significant increases as a function of cell cycle, such that resting free [Ca2+]i vary from 75 nmol/L at G0 to 150 nmol/L at G1/S, whereas sarcoendoplasmic reticulum Ca2+ ATPase (SERCA)-maintained releasable [Ca2+]i vary from 100 nmol/L at G0 to 500 nmol/L at G1/S.8–10 Importantly, preventing increases in resting and releasable [Ca2+]i inhibited G1- to S-phase transitions.8,9 Similarly, SERCA inactivation by thapsigargin has been shown to lengthen the G1 phase of HEK 293 cells.11
Calmodulin (CaM), a small acidic protein of 16.7 kDa, is an important transducer of Ca2+ signals in eukaryotic cells. It is recruited by a large number of proteins, collectively termed CaM-binding proteins (CaMBPs), through which Ca2+ sensitivity is expressed in a variety of cell biological functions.12 The role of CaM as a regulator of cell cycle progression has also been well established.2 Early indications came from experiments in which the mitotic cycle was arrested by anti-CaM drugs added to proliferating cells or to cells reentering cell cycle under the influence of mitogens.13 In addition, it was reported that monoclonal antibodies against CaM inhibited the synthesis of DNA in permeabilized cells.14 Moreover, progression through G1 and mitosis exit was observed to be sensitive to changes in the intracellular concentration of CaM.15 Although others have suggested a role for CaM in the G1/S transition through the activation or expression of proteins involved in DNA replication,16,17 and the phosphorylation of the retinoblastoma protein (Rb),18 the molecular and cell biological mechanisms through which Ca2+ and or CaM exerted these effects remained obscure.
Thus, the primary objectives of our study were to identify which cell cycle proteins expressed in VSMC were responsive to changes in [Ca2+], and whether their Ca2+ sensitivity was mediated by CaM, and how. Given that (1) cyclin E/CDK2 is an enzyme complex that is both rate limiting and essential for S phase entry19,20 and (2) cyclin E synthesis begins in G1 and reaches a peak at the G1/S transition,21,22 we hypothesized that the cyclin E/CDK2 complex may mediate Ca2+ sensitivity of the G1/S cell cycle transition of VSMC. The current study reports that this is the case and that specific CaM-cyclin E interactions underlie this.
Materials and Methods
Materials and Cell Culture
Purified active cyclin E/CDK2 and cyclin A/CDK2 were obtained from Upstate Biotechnology (Lake Placid, NY). CDK2 inhibitor II, KN-93, and cyclosporin A were obtained from Calbiochem (La Jolla, Calif). Calmidazolium, dithiothreitol (DTT), aprotinin, leupeptin, and other chemicals were purchased from Sigma (St Louis, Mo). Synthetic peptides were purchased from GenScript Corp (Piscataway, NJ).
The isolation and culture of primary mouse aortic VSMC and the characterization of MOVAS, a mouse VSMC line, have been described.10 MOVAS were grown in DMEM (GIBCO/BRL, Gaithersburg, Md) supplemented with 10% FBS (Hyclone, Logan, Utah) and 1% penicillin-streptomycin (GIBCO/BRL). For serum starvation, cells were grown to 60% to 70% confluence, washed twice with PBS, and cultured in starvation medium (DMEM without FBS) for 48 hours to achieve G0 arrest. Fresh DMEM with 10% FBS was used to initiate reentry into cell cycle, and cells were incubated for 16 hours to allow them to approach the G1/S transition or for 20 hours for S-phase synchronization. Primary mouse aortic SMC were grown in media supplemented with platelet-derived growth factor (PDGF) (50 ng/mL). G0 synchronization in these cells was achieved with 72 hours of 0.25% FBS-supplemented DMEM. Progression to G1/S was brought about by stimulation with 10% FBS and 50 ng/mL PDGF for 24 hours.
Immunoprecipitation and Kinase Assays
SMC (5x106) were harvested and resuspended with 2 mL of lysis buffer (50 mmol/L Tris [pH 7.4], 250 mmol/L NaCl, 5 mmol/L EDTA, 0.1% NP-40, 0.5 mmol/L DTT, 0.1 mmol/L Na3VO4, 5 μg/mL leupeptin, 5 μg/mL aprotinin, 2 mmol/L PMSF, and 10% glycerol). After homogenization, cell lysates were clarified by centrifugation at 12x103 rpm, 4°C for 30 minutes, and the supernatant was collected. An aliquot was taken for protein concentration determination using the BCA protein assay kit (Sigma). Cell extracts (200 μg) were then incubated for 2 hours at 4°C with saturating concentrations of polyclonal antibodies (Abs). Immune complexes were collected by incubation with GammaBind G Sepharose resin (Amersham Pharmacia, Piscataway, NJ) for 1 hour at 4°C. The beads were then washed three times with washing buffer (50 mmol/L Tris [pH 7.4] and 1 mmol/L DTT).
For the in vitro kinase assays, 20 μL of kinase reaction mixture (20 mmol/L Tris [pH 7.4], 5 mmol/L MgCl2, 2.5 mmol/L MnCl2, 1 mmol/L DTT, 10 μg of C terminal of human Rb protein or 4 μg of histone–H1 [both Upstate Biotechnology] as substrates, 20 μmol/L ATP, and 2.4 μCi [- 32P]ATP [Amersham Pharmacia]) was added to CDK2-, cyclin E–, or cyclin A–immunoprecipitated (IPd) complexes after removing the washing buffer from GammaBind G Sepharose resin. Based on the total calcium content of kinase reaction mixtures (as determined by inductively coupled plasma atomic emission spectrometry), amounts of EGTA required to yield final concentrations of free Ca2+ ion (0, 100 or 500 nmol/L) were calculated by Winmol/LaxC (http://www.stanford.edu/cpatton/winmaxc2.html). After 30 minutes of incubation at 37°C, reactions were stopped by adding 20 μL of 2x sodium dodecyl sulfate (SDS) loading buffer and heating at 100°C for 5 minutes. Labeled proteins were resolved by 16% SDS-PAGE. Phosphorylated Rb or histone–H1 bands were visualized by autoradiography and quantified in a Scintillation Counter LS6500 (Beckman Coulter, Fullerton, Calif) after being cut from the gel.
Western Blot
Between 20 to 30 μg of cell extracts were resolved by 12% SDS-PAGE and transferred to nitrocellulose membranes (Sigma, 0.2 μmol/L pore size). Blots were blocked with 5% nonfat dry milk in TBS-T (10 mmol/L Tris [pH 8.0], 150 mmol/L NaCl, and 0.05% Tween-20) overnight at 4°C and then incubated for 3 hours at 25°C in TBS-T plus 3% nonfat dry milk containing primary Abs, including anti-cyclin E and -cyclin A (Upstate Biotechnology), -phospho-CDK2 at Thr196 (Cell Signaling Technology, Danvers, MA), -CaM, -His probe, and -actin (Santa Cruz Biotechnology, Santa Cruz, Calif). Protein bands were detected with ECL Reagents (Amersham Pharmacia), with horseradish peroxidase–conjugated secondary goat anti-rabbit IgG (Santa Cruz Biotechnology). Quantification of band intensity was performed using Quantity One (Bio-Rad, Hercules, Calif).
CaM-Binding Analysis
Purified cyclin E/CDK2 and cyclin A/CDK2 (100 ng) (Upstate Biotechnology) complexes were diluted with Ca2+-binding buffer (50 mmol/L Tris [pH 7.5], 100 mmol/L NaCl, and 2 mmol/L CaCl2) into a final 20 μL and incubated with an equal volume of preequilibrated CaM–Sepharose 4B resin (Amersham Pharmacia) for 2 hours at 4°C. After centrifugation, the supernatant containing unbound proteins was removed and the resin was washed twice with the buffer used in the binding step. After removing buffer completely from the resin, SDS loading buffer with DTT was added to the dried resin to collect all proteins bound to CaM. The proteins were resolved by 12% SDS-PAGE and visualized by silver staining to determine the amount bound to resin-immobilized CaM.
Confocal Immunofluorescence
Methods and representative images used for analysis of the extent of CaM and cyclin E colocalization are shown in the online data supplement available at http://circres.ahajournals.org.
Flow Cytometric Analysis
Samples were prepared as previously described with minor modifications.8 Cells were counted in a flow cytometer (FACScan; BD Biosciences, Franklin Lakes, NJ), and G0/G1, S, and G2/M cell percentages were calculated with Cell Quest software (BD Biosciences).
Fluorescence Titration
The methods used to determine the dissociation constant (Kd) for peptide–protein binding are detailed in the online data supplement.
Transfection
pIRESpuro3 plasmid vectors (Clontech, Palo Alto, Calif) harboring wild-type (wt) and N-terminal–deleted mutant (N) cyclin E were obtained by PCR amplification of a pECE human cyclin E plasmid (a generous gift of Dr Paul Hamel, University of Toronto) with oligonucleotides 5'-ccggaattcatgaaggaggacggcg-3' (forward wt cyclin E), 5'-ccggaattcatgccagatgaagaaatgg-3' (forward N cyclin E), and 5'-ccggaattctcagtggtggtggtggtggtgcgccatttccggccc-3' (reverse for both). DNA sequences of each construct were confirmed by sequencing. Transfections were performed with the Lipofectamine 2000 reagent (Invitrogen, Carlsbad, Calif) according to the instructions of the manufacturer. Starting 48 hours posttransfection, MOVAS were cultured in the presence of 5 μg/mL puromycin for 2 weeks.
Statistics
ANOVA, Student t test, and coefficients of variance tests were used as appropriate. Analyses were performed on SPSS version 13.0 (Chicago, Ill).
Results
Cyclin E/CDK2 Activity at the G1/S Transition of VSMC Is Ca2+ Sensitive
Kinase activities of cyclin E–IPd complexes from G0- and G1/S-synchronized MOVAS were analyzed in different [Ca2+]. We observed, at 0 nmol/L [Ca2+], that histone–H1 kinase activity at G1/S was greater than at G0 (2.4-fold; Figure 1A). These data suggested a G1/S-associated increase in cyclin E/CDK2 activity, which has been reported previously.23 Our first novel finding was that an increase in reaction buffer [Ca2+] from 0 to 500 nmol/L consistently increased histone–H1 kinase–defined activity of cyclin E–IPd complexes from G1/S- (2.6-fold) but not G0-synchronized VSMC (Figure 1A). Identical results were obtained when the experiment was performed using CDK2-IPd complexes (G1/S: 2.7-fold; Figure 1B). Experiments with CDK2 inhibitor II showed that the histone–H1 kinase–defined activity of cyclin E immunoprecipitates (IPs) was CDK2-dependent (Figure 1C). Of note, Rb kinase-defined activity of cyclin E IPs was also Ca2+ responsive only at G1/S (Figure 1D). In addition, histone–H1 kinase activity of cyclin E–IPd complexes from primary mouse aortic VSMC exhibited the same Ca2+ responsiveness as those from MOVAS (G1/S: 2.3-fold; Figure 1E). As a next step, we determined that this enhancement was specific to an increased [Ca2+] and not other divalent cations (Figure 1F).
Unlike cyclin E/CDK2, cyclin D/CDK4 and cyclin A/CDK2 exhibited different properties. Neither cyclin D–IPd complexes from early G1 phase MOVAS nor cyclin A–IPd complexes from S-phase MOVAS (Figure 1G) showed Ca2+ responsive increases in histone–H1 kinase activity. In addition, purified cyclin A/CDK2 complexes did not exhibit the Ca2+ sensitivity observed with purified cyclin E/CDK2 (Figure 1H). Of interest, the ability of Ca2+ to enhance histone–H1 kinase activity of purified cyclin E/CDK2 complexes (1.7-fold; Figure 1H) was less than that observed in IPs of cell extracts (>2.5-fold; Figure 1A through 1F), suggesting perhaps that factors other than those present in "purified" complexes participate in their Ca2+ sensitivity.
Taken together, the above data showed that the activity of cyclin E/CDK2 on 2 important substrates (Rb and histone–H1) involved in the G1/S transition is Ca2+ sensitive within the physiological range of [Ca2+]i known to accompany this stage of the cell cycle in VSMC.8–10 These data implicated cyclin E/CDK2 as a potential mediator of the known Ca2+ sensitivity of the G1/S transition of both primary mouse VSMC and the MOVAS cell line.
CaM and, to a Limited Extent, Calcineurin, but Not CaM-Activated Kinases, Transduce Ca2+-Sensitive Cyclin E/CDK2 Activity in VSMC
We next explored what mediated the Ca2+ sensitivity of cyclin E/CDK2. A potential role for CaM, among the major Ca2+ signal–transducing proteins in eukaryotes, was examined first. When added to cell extracts, the potent CaM inhibitor calmidazolium inhibited cyclin E/CDK2 activity in a dose-dependent manner (Figure 2A), suggesting a role for CaM in regulating cyclin E/CDK2 activity. As known downstream mediators of CaM-regulated cell cycle progression in various cell types, the potential roles of multifunctional CaM-activated kinases (CaMKs) and calcineurin were tested next. KN-93, an inhibitor of several CaMKs,24 did not affect H1 kinase–defined cyclin E/CDK2 activity (Figure 2B). Only high concentrations of cyclosporin A (50 μmol/L), an inhibitor of calcineurin,25 exhibited at most 20% inhibition of cyclin E/CDK2 activity (Figure 2C).
To test directly whether CaM and calcineurin are necessary for Ca2+ sensitivity of cyclin E/CDK2 activity, immunodepletion experiments were performed. To deplete CaM or calcineurin from cell extracts, immunoprecipitation (IP) was first performed with excess amounts of anti-CaM or anti-calcineurin Abs, followed by a second IP with anti-cyclin E Ab on CaM- or calcineurin-depleted cell extracts. CaM-depleted cell extracts did not exhibit any Ca2+-mediated stimulation of cyclin E/CDK2 activity, suggesting that CaM is essential for Ca2+ sensitivity of cyclin E/CDK2 activity (Figure 2D). By contrast, calcineurin-depleted cell extracts still exhibited some Ca2+ sensitivity (Figure 2D).
To determine whether CaM has a direct effect on cyclin E/CDK2 activity, purified CaM (Sigma; purity >95%) was titrated with a fixed amount of purified human cyclin E/CDK2. CaM increased cyclin E/CDK2 activity in a dose-dependent manner only in the presence of Ca2+, an effect that was abolished by calmidazolium and W-13 (another CaM antagonist) (Figure 2E), suggesting that CaM has a direct stimulatory effect on cyclin E/CDK2 activity. W-12, an analog of W-13 with much lower CaM affinity, did not mimic the effect of W-13 (Figure 2E). Together, the above data suggested that the stimulatory effect of CaM on cyclin E/CDK2 activity was mediated primarily by a direct Ca2+-dependent mechanism involving CaM and cyclin E/CDK2, rather than through CaMKs or calcineurin.
Interaction Between CaM and Cyclin E Is Essential for Ca2+-Sensitive Kinase Activity of Cyclin E/CDK2
To explore further the mechanisms underlying CaM-dependent Ca2+-sensitive cell cycle kinase activities in VSMC, we used a CaM-binding assay. Equal amounts of purified cyclin E/CDK2 and cyclin A/CDK2 were loaded on a CaM–Sepharose column, and silver-stained blots of bound and unbound fractions revealed binding between CaM and cyclin E but not between CaM and cyclin A (Figure 3A). With varying buffer [Ca2+], these experiments revealed that the interaction between CaM and cyclin E occurs in a Ca2+-dependent manner (data not shown). These results were entirely consistent with the observed difference in Ca2+-sensitivities between cyclin E/CDK2 and cyclin A/CDK2 (Figure 1H). Moreover, these data further supported a direct interaction between CaM and cyclin E, not CDK2, as an essential requirement for Ca2+ sensitivity of cyclin E/CDK2.
Co-IP analysis between CaM and cyclin E revealed that the amount of CaM bound to cyclin E increased at G1/S (Figure 3B; 1.68-fold). The absence of any CaM in IPs that used a nonspecific Ab (anti-actin) supported further the specificity of the interaction between CaM and cyclin E. In addition, subcellular localizations of CaM and cyclin E were visualized in MOVAS by fluorescence confocal microscopy, and the proportion of cells showing strong colocalization signals were calculated in both G0- and G1/S- synchronized populations (n=1000 cells each). This analysis suggested a 62% increase in the number of nuclei with strong CaM and cyclin E colocalization at G1/S versus G0 (378 versus 233; P<0.001; online data supplement), a result mirroring closely the observed increase (68%) in co-IP experiments (Figure 3B). This may be related to increased expression levels of CaM at G1/S (Figure 3B), a finding also reported by others,13,26 and/or the higher [Ca2+]i of G1/S-, as compared with G0-synchronized cells.8–10
Inhibition of CaM Retards the Proliferation of VSMC
Having demonstrated a putative molecular mechanism mediating Ca2+-sensitive G1/S transitions, we sought to directly test its physiological significance. Flow-assisted cell sorting analysis (FACS) of DNA content was performed to determine whether inhibition of CaM retards the cell cycle progression of VSMC. Whereas control experiments showed that 24.4% of primary mouse aortic SMC had entered S phase after 24 hours of serum stimulation, only 10.8% of cells treated with calmidazolium, and 12.6% of cells treated with W-13, had done so (P<0.001; Figure 4A). A similar level of cell cycle inhibition was also observed in MOVAS (online data supplement).
Western blot analysis in primary VSMC showed that the interaction between CaM and cyclin E was severely inhibited by calmidazolium (Figure 4B, lanes 5 and 6), whereas the amounts of cyclin E and CaM itself were not (Figure 4B, lanes 2 and 3). This result indicated that the observed decrease in cyclin E-CaM interaction is not simply attributable to the inhibition of cell cycle progression. Importantly, the level of Thr160-phosphorylated CDK2 (the active form of CDK2) was inhibited by calmidazolium treatment (Figure 4B, lanes 2 and 3).
Together, these data suggested (1) that CaM facilitates G1- to S-phase transitions in VSMC via binding to cyclin E and (2) that this interaction may play a role in activating CDK2.
Identification of the CaM-Binding Motif on Cyclin E
To provide a structural basis for our functional model, we next sought to identify the CaM-binding region on cyclin E. A CaM-target database (http://calcium.uhnres.utoronto.ca/ctdb)27 predicted a CaM-binding motif on human cyclin E with successively high probability scores (Figure 5A). This consisted of 22 amino acids near the N terminus of human cyclin E (amino acids 5 to 26), with 5 hydrophobic and 5 basic residues predicted to be critical for a hydrophobic interaction between CaM and cyclin E. We termed this motif the "calmodulin-binding sequence" (CBS) of cyclin E and noted that mouse and rat cyclin E genes also harbor this conserved sequence, whereas the Drosophila and Caenorhabditis elegans homologs do not (Figure 5B), a finding that may restrict the physiological relevance of this domain to mammals. Consistent with the results of our kinase activity assays, no similar CBS was revealed in alignment and sequence analyses of all cyclins D and A.
To confirm the ability of the CBS of cyclin E to bind CaM, we performed tyrosine residue–based fluorescence titration experiments using purified peptides in vitro. Figure 5C shows a representative binding profile of the CBS peptide to CaM. The observed decrease in the relative intensity indicates changes in the environment of any (or all) of the 3 tyrosine residues (2 in CaM and 1 in the CBS peptide) associated with the binding event. The fit of the fluorescence profile (see online data supplement for details) yielded a CBS-CaM dissociation constant (Kd, 1.5±0.1x10–6 mol/L) consistent with other CaM-binding proteins such as caldesmon (Kd, 0.23x10–6 mol/L) and calponin (Kd, 1.3x10–6 mol/L).28
N Cyclin E Neither Binds CaM nor Exhibits Ca2+ Sensitivity
To investigate further the structural basis and functional importance of CaM–cyclin E interactions, expression vectors harboring either His-tagged full length (wt) or His-tagged N mutant forms of human cyclin E (lacking amino acids 1 to 26) were stably transfected into MOVAS (Figure 6A). Despite equal levels of His-tagged wt and N cyclin E protein expression (Figure 6B, left panel), co-IP analyses on G1/S-synchronized cell populations indicated that only wt cyclin E could bind CaM (Figure 6B, right panel). Moreover, histone–H1 kinase assays performed on G1/S-synchronized cells revealed that despite an equivalent level of activity at 0 nmol/L [Ca2+] to wt cyclin E/CDK2, N cyclin E/CDK2 complexes did not exhibit any appreciable Ca2+-sensitive enhancement of function. By contrast, IPd wt human cyclin E exhibited an identical level of Ca2+ sensitivity as endogenous mouse cyclin E in vector only transfected cells (2.5-fold; Figure 6C).
Discussion
This study was focused on identifying the molecular mechanism(s) underlying the poorly understood Ca2+ sensitivity of the G1 to S-phase cell cycle transition of VSMC. The results presented here implicate cyclin E/CDK2 in this process and show that a Ca2+-dependent binding of CaM to cyclin E at G1/S is critical for G1/S- and Ca2+-specific enhancement of cyclin E/CDK2 activity. We also defined a CaM-binding motif in the human cyclin E gene and demonstrated its essential role in mediating Ca2+-sensitive CDK2 activity.
Although the expression of CDK2 in T lymphocytes was noted to be depressed by the CaM antagonist W-13,29 our work is the first to show that CaM increases CDK2 activity via binding to cyclin E. This conclusion was supported by various analyses using several independent approaches including Western blot, co-IP, immunohistochemistry, peptide binding, and structure–function studies with mutants. To investigate further, fluorescence spectroscopy was used to calculate Kd between the CBS peptide and CaM. The finding that N cyclin E was unable to bind to CaM (Figure 6B) strongly suggested that this CaM-binding motif in human cyclin E1 (amino acids 5 to 26) is the only area involved in CaM–cyclin E interactions. As such, the determined Kd obtained from CaM and the CBS peptide likely reflects the true dissociation constant between CaM and cyclin E. If the fluorescent residues examined were far from the binding site or were not involved in a binding-mediated conformational change, no changes in fluorescence would have been observed.30 Rather, our experiments clearly support a structural relationship between the CBS and CaM with a Kd consistent with other CaMBPs such as G protein–coupled receptor kinase 2,31 caldesmon, calponin,28 spectrin,32 and CaM-dependent adenylate cyclases.33,34
Although previous reports have suggested a role for CaM in cell cycle,2 specific molecular pathways through which CaM might exert these effects were not known. Kahl and Means implicated CaM in the regulation of cyclin D1/CDK4 through their discovery that CaMKI, and not CaMKII, was the KN-93–sensitive CaMK responsible for KN-93–induced G1 arrest in WI-38 fibroblasts.35 They reported also that CaM activates cyclin D1 protein synthesis in fibroblasts via calcineurin.36 Previously, Taules and colleagues had implicated CaM in the nuclear entry of cyclin D/CDK4 and identified a relevant interaction between CaM and the CDK4 inhibitor p21Cip1.37,38 In contrast to the work in other cell types, we did not find KN-93–sensitive CaMKs to be involved in mediating Ca2+/CaM sensitivity of cyclin E/CDK2 complexes and demonstrated a very modest effect of cyclosporin A and calcineurin immunodepletion. Rather, we have identified a physiologically relevant role for CaM at a later and arguably more critical cell cycle checkpoint in VSMC and are the first to directly demonstrate Ca2+ sensitivity of its target. Indeed, it is interesting to note that although the studies of Kahl and Means and Taules and colleagues focused on the role of CaM in regulating the early-to-mid G1 factors cyclin D/CDK4, we have examined the late G1- and S-phase factors cyclin E/CDK2. Collectively, the key emerging concept from these studies is that CaM mediates Ca2+ sensitivities of both the early and late G1- to S-phase checkpoints through its interactions with cyclin D/CDK4/p21cip1 and cyclin E/CDK2/p27kip1, respectively.
Although CaM-binding to cyclin E clearly increased the levels of Thr160-phosphorylated (ie, active) CDK2 in extracts from G1/S-synchronized VSMC (Figure 4B), we have yet to determine how precisely this occurs. Possible explanations include Ca2+/CaM-cyclin E- or Ca2+/CaM-dependent augmentation of (1) another kinase involved in CDK2 activation (namely CDK-activating kinase [CAK], reviewed by Morgan39) or (2) the release and/or degradation of the CDK2 inhibitor p27kip1 from the otherwise mute cyclin E/CDK2/p27kip1 complex. As preliminary data from our laboratory indicate that the binding of CaM to cyclin E does not cause the release of p27kip1 from cyclin E/CDK2 complex (data not shown), we suspect that the binding of CaM to cyclin E changes the conformation of the cyclin E/CDK2 complex, such that CDK2 can be more easily phosphorylated by CAK. Activation mechanisms of CaMBPs by CaM strongly support the concept that CaM activates its target proteins by altering their conformation.40 Further studies will be required to define precisely the mechanisms underlying this observation and the potential involvement of CAK in the Ca2+/CaM-cyclin E-dependent activation of CDK2.
Acknowledgments
J.C. was supported in part by doctoral student stipend awards from the Ontario Graduate Scholarship and Canadian Institute of Health Research (CIHR). M.H. was the recipient of a Clinician Scientist Award from the CIHR and currently holds a Career Investigator Award from the Heart & Stroke Foundation of Ontario (CI5503 to M.H.). This work was supported by a CIHR operating grant (MOP14648 to M.H.). We thank Drs Paul Hamel and Brenda Andrews (University of Toronto) for thoughtful discussions. We also acknowledge Dr Tigran Chalikian (University of Toronto) for supporting the fluorescence titration work.
Footnotes
Original received November 23, 2005; revision received March 15, 2006; accepted April 6, 2006.
References
Ross R. The pathogenesis of atherosclerosis: a perspective for the 1990s. Nature. 1993; 362: 801–809. [Order article via Infotrieve]
Kahl CR, Means AR. Regulation of cell cycle progression by calcium/calmodulin-dependent pathways. Endocr Rev. 2003; 24: 719–736.
Takuwa N, Zhou W, Takuwa Y. Calcium, calmodulin and cell cycle progression. Cell Signal. 1995; 7: 93–104. [Order article via Infotrieve]
Kojima I, Matsunaga H, Kurokawa K, Ogata E, Nishimoto I. Calcium influx: an intracellular message of the mitogenic action of insulin-like growth factor-I. J Biol Chem. 1988; 263: 16561–16567.
Tomono M, Toyoshima K, Ito M, Amano H. Calcineurin is essential for DNA synthesis in Swiss 3T3 fibroblasts. Biochem J. 1996; 317 (pt 3): 675–680. [Order article via Infotrieve]
Short AD, Bian J, Ghosh TK, Waldron RT, Rybak SL, Gill DL. Intracellular Ca2+ pool content is linked to control of cell growth. Proc Natl Acad Sci U S A. 1993; 90: 4986–4990.
Ghosh TK, Bian JH, Short AD, Rybak SL, Gill DL. Persistent intracellular calcium pool depletion by thapsigargin and its influence on cell growth. J Biol Chem. 1991; 266: 24690–24697.
Husain M, Bein K, Jiang L, Alper SL, Simons M, Rosenberg RD. c-Myb-dependent cell cycle progression and Ca2+ storage in cultured vascular smooth muscle cells. Circ Res. 1997; 80: 617–626.
Husain M, Jiang L, See V, Bein K, Simons M, Alper SL, Rosenberg RD. Regulation of vascular smooth muscle cell proliferation by plasma membrane Ca(2+)-ATPase. Am J Physiol. 1997; 272: C1947–C1959. [Order article via Infotrieve]
Afroze T, Yang LL, Wang C, Gros R, Kalair W, Hoque AN, Mungrue IN, Zhu Z, Husain M. Calcineurin-independent regulation of plasma membrane Ca2+ ATPase-4 in the vascular smooth muscle cell cycle. Am J Physiol Cell Physiol. 2003; 285: C88–C95.
Simon VR, Moran MF. SERCA activity is required for timely progression through G1/S. Cell Prolif. 2001; 34: 15–30. [Order article via Infotrieve]
van Eldik LJ, Watterson DM. Calmodulin & Signal Transduction. New York: Academic Press; 1998.
Chafouleas JG, Bolton WE, Hidaka H, Boyd AE 3rd, Means AR. Calmodulin and the cell cycle: involvement in regulation of cell-cycle progression. Cell. 1982; 28: 41–50. [Order article via Infotrieve]
Reddy GP, Reed WC, Sheehan E, Sacks DB. Calmodulin-specific monoclonal antibodies inhibit DNA replication in mammalian cells. Biochemistry. 1992; 31: 10426–10430. [Order article via Infotrieve]
Rasmussen CD, Means AR. Calmodulin is required for cell-cycle progression during G1 and mitosis. EMBO J. 1989; 8: 73–82. [Order article via Infotrieve]
Lopez-Girona A, Bachs O, Agell N. Calmodulin is involved in the induction of DNA polymerases alpha and delta activities in normal rat kidney cells activated to proliferate. Biochem Biophys Res Commun. 1995; 217: 566–574. [Order article via Infotrieve]
Lopez-Girona A, Bosch M, Bachs O, Agell N. Addition of calmodulin antagonists to NRK cells during G1 inhibits proliferating cell nuclear antigen expression. Cell Calcium. 1995; 18: 30–40. [Order article via Infotrieve]
Takuwa N, Zhou W, Kumada M, Takuwa Y. Ca(2+)-dependent stimulation of retinoblastoma gene product phosphorylation and p34cdc2 kinase activation in serum-stimulated human fibroblasts. J Biol Chem. 1993; 268: 138–145.
Dulic V, Lees E, Reed SI. Association of human cyclin E with a periodic G1-S phase protein kinase. Science. 1992; 257: 1958–1961.
Lundberg AS, Weinberg RA. Functional inactivation of the retinoblastoma protein requires sequential modification by at least two distinct cyclin-cdk complexes. Mol Cell Biol. 1998; 18: 753–761.
Clurman BE, Sheaff RJ, Thress K, Groudine M, Roberts JM. Turnover of cyclin E by the ubiquitin-proteasome pathway is regulated by cdk2 binding and cyclin phosphorylation. Genes Dev. 1996; 10: 1979–1990.
Lew DJ, Dulic V, Reed SI. Isolation of three novel human cyclins by rescue of G1 cyclin (Cln) function in yeast. Cell. 1991; 66: 1197–1206. [Order article via Infotrieve]
Koff A, Giordano A, Desai D, Yamashita K, Harper JW, Elledge S, Nishimoto T, Morgan DO, Franza BR, Roberts JM. Formation and activation of a cyclin E-cdk2 complex during the G1 phase of the human cell cycle. Science. 1992; 257: 1689–1694.
Hook SS, Means AR. Ca(2+)/CaM-dependent kinases: from activation to function. Annu Rev Pharmacol Toxicol. 2001; 41: 471–505. [Order article via Infotrieve]
Liu J, Farmer JD Jr, Lane WS, Friedman J, Weissman I, Schreiber SL. Calcineurin is a common target of cyclophilin-cyclosporin A and FKBP-FK506 complexes. Cell. 1991; 66: 807–815. [Order article via Infotrieve]
Chafouleas JG, Lagace L, Bolton WE, Boyd AE 3rd, Means AR. Changes in calmodulin and its mRNA accompany reentry of quiescent (G0) cells into the cell cycle. Cell. 1984; 36: 73–81. [Order article via Infotrieve]
Yap KL, Kim J, Truong K, Sherman M, Yuan T, Ikura M. Calmodulin target database. J Struct Funct Genomics. 2000; 1: 8–14. [Order article via Infotrieve]
Medvedeva MV, Kolobova EA, Wang P, Gusev NB. Interaction of proteolytic fragments of calmodulin with caldesmon and calponin. Biochem J. 1996; 315 (pt 3): 1021–1026. [Order article via Infotrieve]
Colomer J, Lopez-Girona A, Agell N, Bachs O. Calmodulin regulates the expression of cdks, cyclins and replicative enzymes during proliferative activation of human T lymphocytes. Biochem Biophys Res Commun. 1994; 200: 306–312. [Order article via Infotrieve]
Lakowicz JR. Principles of Fluorescence Spectroscopy. 2nd ed. New York: Kluwer Academic/Plenum publishers; 1999.
Pronin AN, Satpaev DK, Slepak VZ, Benovic JL. Regulation of G protein-coupled receptor kinases by calmodulin and localization of the calmodulin binding domain. J Biol Chem. 1997; 272: 18273–18280.
Bjork J, Lundberg S, Backman L. Characterization of the binding of calmodulin to nonerythroid spectrin. Eur J Cell Biol. 1995; 66: 200–204. [Order article via Infotrieve]
Cali JJ, Parekh RS, Krupinski J. Splice variants of type VIII adenylyl cyclase. Differences in glycosylation and regulation by Ca2+/calmodulin. J Biol Chem. 1996; 271: 1089–1095.
Choi EJ, Wong ST, Hinds TR, Storm DR. Calcium and muscarinic agonist stimulation of type I adenylylcyclase in whole cells. J Biol Chem. 1992; 267: 12440–12442.
Kahl CR, Means AR. Regulation of cyclin D1/Cdk4 complexes by calcium/calmodulin-dependent protein kinase I. J Biol Chem. 2004; 279: 15411–15419.
Kahl CR, Means AR. Calcineurin regulates cyclin D1 accumulation in growth-stimulated fibroblasts. Mol Biol Cell. 2004; 15: 1833–1842.
Taules M, Rius E, Talaya D, Lopez-Girona A, Bachs O, Agell N. Calmodulin is essential for cyclin-dependent kinase 4 (Cdk4) activity and nuclear accumulation of cyclin D1-Cdk4 during G1. J Biol Chem. 1998; 273: 33279–33286.
Taules M, Rodriguez-Vilarrupla A, Rius E, Estanyol JM, Casanovas O, Sacks DB, Perez-Paya E, Bachs O, Agell N. Calmodulin binds to p21(Cip1) and is involved in the regulation of its nuclear localization. J Biol Chem. 1999; 274: 24445–24448.
Morgan DO. Principles of CDK regulation. Nature. 1995; 374: 131–134. [Order article via Infotrieve]
Hoeflich KP, Ikura M. Calmodulin in action: diversity in target recognition and activation mechanisms. Cell. 2002; 108: 739–742. [Order article via Infotrieve]
您现在查看是摘要介绍页，详见ORG附件。