Id3 Is a Novel Regulator of P27kip1 Mrna in Early G1 Phase and Is

Home , EGR3, ETS1, JUNB, MAX (gene), NFATC2

Oncogene (2007) 26, 5772–5783 & 2007 Nature Publishing Group All rights reserved 0950-9232/07 $30.00 www.nature.com/onc ORIGINAL ARTICLE Id3 is a novel regulator of p27kip1 mRNA in earlyG1 phase and is required for cell-cycle progression

A-A Chassot1,4, L Turchi1,4, T Virolle1, G Fitsialos1, M Batoz2, M Deckert2, V Dulic3, G Meneguzzi1, RBusca` 1 and G Ponzio1

1INSERM U634; Faculte´ de Me´decine, Universite´ Nice Sophia Antipolis, Nice cedex, France; 2INSERM U576, Hoˆpital de L’Archet, Universite´ Nice Sophia Antipolis, Nice, France and 3CRBM-CNRS FRE 2593, Montpellier, France

P27kip is a keyinhibitoryprotein of the cell-cycle Introduction progression, which is rapidlydownregulated in earlyG1 phase bya post-translational mechanism involving the Most adult tissues are constituted by differentiated proteosomal degradation. In this study, using a wounding cells with reduced proliferative activity. Renewal and/or model that induces cell-cycle entry of human dermal fibro- regeneration of these tissues is supplied by quiescent blasts, we demonstrate that p27mRNA is downregulated cells sensitive to mitogenic signals. During the skin when cells progress into the G1 phase, and then it returns wound healing process, the proliferation of quiescent to its basal level when cells approach the S phase. By cells can be stimulated by signals such as serum afflux using a quantitative polymerase chain reaction screening and/or contact inhibition failure. we identified inhibitors of differentiation (Id3), a bHLH ‘Inhibitors of differentiation’ (Id) proteins are early transcriptional repressor, as a candidate mediator accoun- genes that operate as regulators of cell fate determi- ting for p27 mRNA decrease. Id3 silencing, using an small nation. They play a crucial role in the coordinated interfering RNA approach, reversed the injurymediated regulation of gene expression during cell growth, p27 downregulation demonstrating that Id3 is involved in differentiation and tumorigenesis (Norton, 2000; the transcriptional repression of p27. Reporter gene Benezra et al., 2001). The involvement of Ids was also experiments and a chromatin immunoprecipitation assay recently shown in cellular senescence (Alani et al., 2001; showed that Id3 likelyexerts its repressive action through Ohtani et al., 2001; Zheng et al., 2004) and in the fate of ELK1 inhibition. Byinhibiting earlyp27 downregulation, specialized cells such as lymphocytes, vascular endo- Id3 depletion blocked (i) the G1-phase progression as thelial cells and neurons (Lyden et al., 1999; Benezra assessed bythe inhibition of pRb phosphorylationand et al., 2001; Engel and Murre, 2001). p130 degradation and (ii) the G1/S transition as observed Today, four members of the Id protein family (Id1– bythe inhibition of cyclinA induction, demonstrating that Id4) have been identified. They directly associate with p27 mRNA decrease is required for cell proliferation. and regulate the activity of several families of transcrip- Apart from its effect on the earlyp27 diminution, Id3 tional regulators (Yates et al., 1999; Lasorella et al., appears also involved in the control of the steady-state 2000; Norton, 2000; Roberts et al., 2001). In mammals, level of p27 at the G1/S boundary. In conclusion, this Id proteins exert their biological effects mainly through studyidentifies a novel mechanism of p27 regulation the interaction with bHLH transcription factors by which besides p27 protein degradation also implicates a blocking their DNA-binding activity. Among the Id3 transcriptional mechanism mediated byId3. partners, the members of the E2A family of transcrip- Oncogene (2007) 26, 5772–5783; doi:10.1038/sj.onc.1210386; tion factors (E12, E47) are the most frequently described published online 2 April 2007 (Massari and Murre, 2000; Norton, 2000). However, they are not the exclusive partners of Id3 as it has also Keywords: Ids; cell-cycle; p27mRNA; wound healing been shown that Id3 also interacts with MyoD, Ets, Pax or TCFs Ets-domain transcription factors (Yates et al., 1999; Roberts et al., 2001; Trausch-Azar et al., 2004). Ids proteins lack a basic DNA-binding domain and readily associate with transcription factors of the bHLH family such as E proteins to inhibit their DNA-binding function. Recent data have demonstrated that Ids activity Correspondence: G Ponzio, INSERM U634, Faculte´ de Me´ decine, is inhibited by cell-cycle-dependent kinase activities Universite´ Nice Sophia Antipolis, 28 avenue de Valombrose, 06107 such as cyclin E- and A-CDK2 (Deed et al., 1997). In Nice cedex 02, France. addition, it has also been shown that E2A and Ids E-mail: [email protected] 4 proteins participate to the transcriptional regulation These two authors equally contributed to this work. cip1 Received 13 February 2006; revised 23 January 2007; accepted 1 February of the cyclin-dependent kinase inhibitors (CKIs) p21 2007; published online 2 April 2007 (Id1) (Prabhu et al., 1997; Takahashi et al., 2004) and Id3 regulates p27mRNA in early G1 A-A Chassot et al 5773 16ink4a (Id1) (Zheng et al., 2004) in different biological to date has been poorly studied compared to the p27 models including senescence. Recent studies have shown protein fate. that besides p21 and p16, Id1 and Id3 are also involved To perform this analysis, hyperconfluent contact inhi- in the regulation of p27 protein, notably in Xenopus bited (CI) normal HDF were injured to stimulate neural crest progenitors (Kee and Bronner-Fraser, 2005) cell-cycle entry and next p27 mRNA was measured by and in Epstein–Barr virus infected cells (Everly et al., real-time quantitative PCR, before wounding and 1, 3, 2004); however, the molecular basis of this inhibition 6, 15 and 24 h after injury. In parallel we also analysed has never been elucidated. the expression of p27 protein by Western blot. To The CKI p27kip1 is a key regulator of cell-cycle assess the synchronous cell-cycle entry, we measured commitment and progression. P27 is downregulated the expression of cyclin A as a marker of the G1-S when cells are stimulated with mitotic agents and it transition. is induced to play a key role in mediating G1 arrest We observed that p27 mRNA was rapidly down- in response to mitogen starvation, cell confluence or regulated in response to injury reaching its minimal transforming growth factor b (Sherr and Roberts, 1995). levels about 3 h after injury (Figure 1a). About 24 h It is well established that p27 expression is mainly post- later, p27 transcript reached again its basal levels translationally regulated by the ubiquitin-proteasome (Figure 1a) when cells entered in S phase as confir- pathway (Pagano et al., 1995; Montagnoli et al., 1999; med by the measure of cyclin A mRNA expression Malek et al., 2001; Kamura et al., 2004). However, some studies demonstrate that p27 might also be transcrip- tionally regulated (Servant et al., 2000; Sakakibara et al., 2005). Nevertheless, at present, the mechanisms 100 involved in such regulation remain unidentified. Dermal fibroblast proliferation is a major feature of cutaneous wound healing (Martin, 1997), which is altered in several pathologies such as keloids (Calderon 50 Cyclin A

et al., 1996) and chronic wounds, which can evoluate (% of max) p27 toward cutaneous carcinoma (Loots et al., 1999; Chraibi Cyc A & p27 mRNA et al., 2004). To date the mechanisms underlying cell 0 proliferation in the cutaneous wound healing context 02030 remain poorly understood. Using a suitable model to Hours after injury study in vitro wound healing (Turchi et al., 2002), we have explored the molecular mechanisms responsible b p27 for the cell-cycle entry in wounded normal human dermal fibroblasts (HDF). We show that this process Conf 3 6 15 24 depends, at least in part, on the rapid downregulation of Hours after wounding p27 mRNA. Using a small interfering RNA (siRNA) approach, we identify Id3 as a mediator of this effect. c 1 The silencing of Id3 abolishes the early downregulation 0.8 of p27 and inhibits the cell progression in G1 as well as the G1/S transition. 0.6 0.4 0.2 Results p27 mRNA level (A.U.) 0 Wounding regulates p27 at the mRNA level and this N WNW requires the synthesis of a mediator protein We have developed an original device that creates calibrated long size injuries within confluent cell cultures. In Figure 1 Effect of HDF injury on p27 expression: confluent HDF were stimulated to grow by mechanical injury as described in classical conditions, this device wounds about 50% of Materials and methods and harvested at the indicated times. (a) the cell monolayer allowing the detection of a wide Total RNAs were isolated then p27 and cyclin A mRNA were spectrum of molecular events (Turchi et al., 2002). Using analysed by real-time PCRusing the procedure described in our wounding system, hyperconfluent dermal fibroblasts ‘Materials and methods’. The presented experiment corresponds to are synchronously stimulated to enter the cell-cycle a typical one chosen among six different experiments. (b) Cells were lysed and p27 protein levels were measured by Western blot. The (unpublished data). Using this system we have investi- presented experiment was chosen among four independent ones (c) gated the molecular mechanisms involved in the initial Effect of cycloheximide on the downregulation of p27mRNA in step of the cell-cycle entry in HDF in response to injury. wounded HDF. Confluent HDF were incubated 30 min in the We have focused our attention on p27kip1, an inhibitor absence (À) or in the presence of 5 mM cycloheximide (CHX) before injury. Wounded (W) and control (N) cells were harvested 3 h after of cyclin-dependent kinases (Cdk) that plays a critical stimulation, total RNA were extracted and p27 mRNA was role in the regulation of cell proliferation. We have measured by real-time Q-PCR. This experiment is representative of precisely analysed the p27 mRNA regulation, which up three independent experiments.

Oncogene Id3 regulates p27mRNA in early G1 A-A Chassot et al 5774 Table 1 List of the transcription factors upregulated 1 h after HDF wounding Symbol LocusLink Molecular function Ratio W/NW

FOSB 2354 Subtilase activity; transcription factor activity; transcription factor binding 46737237 FOS 2353 DNA binding; speciﬁc RNA polymerase II transcription factor activity 46.575.6 KLF2 10365 Transcription factor activity; transcriptional activator activity 3075.3 NFATc2 4773 Transcription factor activity 9.272.1 EGR1 1958 Transcription factor activity 7.971.7 EGR2 1959 DNA-binding 6.871.5 EGR3 1960 transcription factor activity 5.570.9 Id3 3399 DNA-binding; transcription corepressor activity 5.471.1 JUNB 3726 Transcription factor activity; transcription coactivator activity; transcription corepressor activity; 4.670.6 REL 5966 transcription factor activity 4.371.2 JUN 3725 RNA polymerase II transcription factor activity; transcription factor activity 3.370.5

Abbreviations: NW, non-wounded cells; W, wounded. The first column (symbol) indicates the official symbol of the genes. Column 2 (LocusLink) indicates the ‘GeneID’ attributed to the genes in the LocusLink database. Column 3 shows the known molecular functions of the genes determined using Ease program (Hosack et al., 2003). Column 4 represents the ratio between the gene expression in W versus confluent NW cells. The presented data are from a typical experiment chosen among three individual ones.

(Figure 1a). P27 protein was regulated following the FCS-depleted HDF and strongly stimulated 1 h after same profile. It decreased between 0 and 6 h after injury FCS addition. The same stimulation occurred after and increased again around 15 h to reach its maximal HDF replating (Figure 2b). A clear downregulation of levels 24 h after (Figure 1b). p27 mRNA was also observed in these two conditions To elucidate the mechanisms responsible for p27 (Figure 2c), suggesting that Id3 also regulates p27 in mRNA downregulation, we performed the same wound- other situations than wound healing. ing experiment in the presence of cycloheximide that impairs the protein synthesis. P27 mRNA decrease was Induction of Id3 in response to cell detachment completely inhibited when cells were preincubated with In an attempt to find a functional signal accounting for cycloheximide before wounding (Figure 1c), indicating Id3 stimulation in response to injury, we focused our that this regulation likely depends on the synthesis of attention on cell-to-cell and cell to extra cellular matrix a mediator protein stimulated in response to injury. detachment, which constitute relevant immediate events inherent to the wounding process. Confluent HDF were detached by mild trypsinization, kept for 15, 30, 45 and Id3, is a candidate regulator of p27kip1 transcription 60 min in suspension at 371C, then Id3 protein was To characterize further this intermediary molecule, measured by Western blot. The expression of Id3 we screened the expression of several transcription increased between 30 min and 1 h after cell detachment factors in response to injury, using a Q-PCRapproach (Figure 2d). Id3 upregulation was completely inhibited described in ‘Material and methods’. Some of these when cells were treated with actinomycin D before genes were clearly upregulated by more than threefold trypsinization (Figure 2e), suggesting that the upregu- (see Table 1) in our experimental conditions. In this lation of Id3 itself resulted from a transcriptional context, Id3, a transcriptional repressor of bHLH regulatory mechanism activated by injury-induced cell transcription factors, appeared as a possible candidate detachment. to mediate the p27 mRNA diminution, as it was clearly stimulated at this time point. Indeed, Id family members (Id1–4) have been shown to regulate cell proliferation in Id3 is involved in injury-stimulated p27 downregulation other biological systems, notably by controlling the To investigate the potential role of Id3 in p27 down- expression of p21cip1 and p16ink4B, two other cell-cycle regulation, we generated a specific siRNA to deplete inhibitors (Prabhu et al., 1997; Zheng et al., 2004). We cells of Id3 expression. Fluorescence microscopy after analysed the profile of Id3 mRNA variation in response transfection of Cy3-labelled Id3 demonstrated that this to injury and we observed that Id3 was regulated in a siRNA was efficiently incorporated in confluent HDF biphasic manner with a maximal stimulation 1 h after (60% of the cells) (not shown) and clearly inhibited Id3 injury and a second burst 24 h after injury (Figure 2a). mRNA expression (Figure 3a), resulting in a decrease Western blot analysis showed that, like mRNA, Id3 of the Id3 protein (60% inhibition) (Figure 3, panel b, protein regulation was also biphasic. Id3 was weakly left and right panels). The efficiency of the Id3-siRNA expressed in confluent non-wounded cells, transiently was maintained during at least 24 h as shown in increased 1 h after injury and raised again after 24 h Figure 4a (upper graph). (Figure 2a). Next, the effect of Id3 silencing on p27 regulation Id3 mRNA expression was also studied in two other was assessed both at the mRNA and protein levels situations classically known to induce cell-cycle entry, (Figure 4a (lower panel) and b), using respectively real- namely serum (fetal calf serum (FCS)) stimulation time PCRand Western blot analysis. We observed that and cell replating (Figure 2b). Id3 mRNA was low in the p27 downregulation associated with the cell-cycle

Oncogene Id3 regulates p27mRNA in early G1 A-A Chassot et al 5775

8 Id3 mRNA c 1.2 4 1 p27m RNA (A.U.) 0.8 Id3 mRNA level 0 0.6 Conf1 3 6 24 0.4 Hours after wounding 0.2 P27 mRNA level(AU)

0 Id3 Conf W Replated -FCS + FCS

Conf1 3 6 24 d Id3 Hours after wounding b Conf 15 30 45 60 min 7 Id3mRNA e 6 5 Id3 4 Conf 1 3 6 Conf 1 3 6 hours 3

Id3 mRNA Ti Ti 2 (fold stimulation) 1 + Actino D 0 ConfW Replated -FCS + FCS Figure 2 (a) Kinetics of Id3 expression measured by real-time PCR(upper panel) and western blot (lower panel) in wounded HDF. The presented data are from typical experiment chosen among three independent ones. Measure of Id3 (b) and p27 (c) mRNA expression in (Conf) confluent HDF; (W) wounded HDF, (replated) confluent HDF trypsinized and replated at low density; (ÀFCS) HDF deprived of FCS for 36 h; ( þ FCS) depleted HDF stimulated with 20% serum. In each experimental condition, cells were harvested 1 h after stimulation and total RNA were isolated. The Id3 mRNA and p27 mRNAs levels expression were measured by real- time PCR. (d) Kinetics of Id3 expression after cell detachment by mild trypsinization. Confluent HDF (conf) were detached by Ti and maintained in suspension at 371C. After 15, 30, 45 and 60 min the cells were lysed and Id3 expression was analysed by Western blotting using a specific anti-Id3 antibody. (e) Cells were incubated in the absence or in the presence of 5 mg/ml actinomycin D (Actino D) and submitted to mild trypsinization (Ti) as described in d. Cells were harvested after 1, 3 and 6 h, lysed and Id3 was analysed by Western blotting. The data presented in d and e are from typical experiments chosen among three independent ones.

entry was reduced by about 50% when HDF were To gain supplementary information concerning the treated with Id3-siRNA. This effect was apparent 1, 3 putative Id3 partner involved in p27 transcriptional and 6 h after injury (Figure 4a). Interestingly, 24 h regulation, we measured p27 promoter activity in after wounding, Id3 depletion further increased p27 NIH3T3 cells, in the presence of constructs encoding mRNA over the basal level, suggesting that Id3 not only two possible candidates: (i) the E2A family members controls the early diminution of p27 mRNA but also E12 and E47 and (ii) elk1 a TCF Ets-domain transcrip- triggers a signal responsible for the return of p27 to its tion factor (Yates et al., 1999; Norton, 2000). A strong basal level. promoter activity was detected in NIH3T3 when the The Western blot analysis of p27 protein expression single p27-luciferase promoter was expressed (Figure 4c, (Figure 4b, lower panel) showed that Id3 silencing pcDNA3). Its cotransfection with plasmids encoding (Figure 4b, upper panel) also reduced (by about 50%) both E12 and E47 (Figure 4c, E12 þ E47) did not further the downregulation of p27 protein occurring in HDF increase this basal activity. In contrast, a vector enco- committed into the cell-cycle after injury. ding ELK1 significantly increased the basal transcrip- To better dissect the mechanisms of p27 mRNA regu- tional activity of p27 promoter (Figure 4c, ELK1), lation by Id3, we used a fragment of the p27 promoter indicating that ELK1 might be a key element in the cloned upstream the luciferase reporter gene. NIH3T3 control of p27 promoter activity. To verify the presence cells were transfected with a p27 promoter construct, of Elk1-occupied sites within this promoter, we used together with a plasmid expressing either the wild-type a chromatin immunoprecipitation (ChIP) approach. Id3 or a non-phosphorylable Id3-variant (Id3-Ala5), Chromatin complexes were immunoprecipitated from which exerts a stronger repressor activity. (Deed et al., confluent HDF extracts using an anti-ELK1 antibody. 1997). We detected a high basal activity of p27 promoter Then a PCRwas performed with different primer pairs in NIH3T3 cells, which was inhibited by the cotransfec- to amplify the ELK1-bound p27 promoter region. tion with the Id3 expression vector. This inhibition was As shown in Figure 4d, PCRamplification revealed enhanced when cells were transfected with Id3-Ala5, a 478 bp band corresponding to the À689/À211 region indicating that Id3 efficiently inhibits the transcriptional of p27 promoter (lane 3). This band was specific as it activity of p27 promoter (Figure 4c). was neither detected when the immunoprecipitation was

Oncogene Id3 regulates p27mRNA in early G1 A-A Chassot et al 5776 Id3-siRNA The Id3-dependent p27 downregulation is critical for 1.4 Ct-siRNA cell-cycle entry 1.2 We next investigated the physiological impact of Id3- 1 dependent p27 downregulation on cell-cycle progres- 0.8 sion. In this aim, we assessed the effect of Id3 silencing 0.6 on the cell-cycle commitment by analysing the effect of Id3-siRNA on two critical phases of the cell-cycle: 0.4 (i) before the G1-restriction point (in early-mid-G1) Id3 mRNA level(A.U.) 0.2 namely when the pocket proteins are initially phos- 0 phorylated by cdk4/6; and (ii) after the restriction point NW W at the G1/S transition when cells are irreversibly b 5000 engaged into the cell-cycle. In this aim cells were treated with Id3-siRNA or with control-siRNA and harvested 4000 AB 10 and 24 h after injury. 3000 To gain information concerning the effect of siRNA Id3 2000 on early-mid-G1 events, we ﬁrst analysed the effect of Id3-siRNAs on p27 downregulation, cyclin D1 stimula- NW W W 1000

Id3 protein level(AU) tion as well as on pRb phosphorylation and p130 Ct-siRNA Id3-siRNA 0 NW W W degradation (the two major physiological substrates of Id3-siRNA - + - cyclin D/Cdk4/6 kinases). These measures were carried CtsiRNA - - + out 10 h after injury (early-mid-G). We observed that Figure 3 (a) Effect of Id3 SiRNAs on Id3 mRNA expression. Id3 silencing impaired by 50% the p27 downregulation Confluent HDF were transfected with control or Id3 siRNA and also decreased the phosphorylation of pRb as well (100 nM), then wounded (W) or not (NW). Cells were harvested as the degradation of p130 (Figure 5a). This inhibition after 1 h (when Id3 level was maximum) and Id3 mRNA level was did not result from an inhibition of cyclin D1 stimu- measured by real-time PCR. The presented data are from a typical lation, as the injury-stimulated expression of cyclin experiment chosen among three independent ones. (b) Analysis of Id3 SiRNA effect on Id3 protein expression. HDF were treated like D1 was not altered after Id3 silencing compared to in panel a (1 h after wounding) and Id3 was measured by Western Ct-siRNA (Figure 5a). Moreover and in contrast with a blot (left part). The upper panel (AB) represents the loading control previous study (Prabhu et al., 1997), in our wounding (amido-black stained polyvinylidene fluoride membrane). The right model Id3 did not control the expression of p21, another part represents the quantification of the presented Western blot using the ImageJ software available on the web (http://rsb.info. cell-cycle inhibitor (Figure 5a). nih.gov/ij/). Cyclin D1 immunoprecipitation experiments showed that, 10 h after injury, the amount of p27 bound to cyclin D1-CDK4/6 complexes was higher in lysates of performed with a non-relevant antibody (lane 1) nor NHF transfected with id3-siRNAs than in control with the single sepharose beads (lane 2). By analysing conditions (Figure 5b), indicating that the excess the p27 promoter sequence (Minami et al., 1997), we of p27 resulting from the reduced downregulation of noted the presence of an ELK1-type consensus motif p27 in Id3-siRNA treated cells, likely associated with (CAccGCCATATTGG) in the À453/À439 region loca- cyclin D1/CDK4/6 complexes (Figure 5b). ted within the 478 bp region identified by ChIP. We Then we examined the effect of Id3 silencing 24 h after therefore mutated this site and transfected NIH3T3 cells injury namely during the G1/S transition and during with this new mutant construct. We observed that the the S-phase progression (Figure 5c, d and e). By using suppression of the elk1 site abolished more than 70% of real-time PCR(Figure 5c) and Western blot experiments the basal activity of p27 promoter indicating that this (Figure 5c), we found that, 24 h after wounding, the consensus sequence is pivotal for the regulation of the expression of cyclin A was inhibited by more than 40% promoter activity (Figure 4e, pcDNA3/mut-p27 prom). when cells were wounded in the presence of Id3-siRNA, Furthermore, the cotransfection of p27 promoter with a indicating that the Id3-dependent inhibition of p27 plasmid encoding a dominant-negative (dn) variant of blocked the G1/S transition. The inhibition of S-phase Elk1 strongly inhibited the promoter activity (Figure 4e, entry was verified by showing that Id3-siRNA inhibited dn-ELK1/wt-p27 prom). Then we investigated whether by 45% the DNA synthesis, visualized by BrdU Id3 could be a partner of ELK1 that might antagonize incorporation experiments (Figure 5e). its stimulatory effect on the promoter of p27. In this Taken together our data show, for the first time, that aim, we cotransfected the p27 promoter together with Id3-mediated p27 transcriptional regulation is a critical ELK1, and in the presence or absence of a plasmid event controlling the cell-cycle progression. The abroga- encoding Id3. We observed that the elk1 stimulation of tion of p27 mRNA downregulation is sufficient to the p27 promoter activity (Figure 4e, Elk1/wt-p27 prom) impair the cell-cycle entry and G1-phase progression. was totally inhibited by the overexpression of Id3 To further dissect this point and to assess whether Id3 (Figure 4e, Elk1 þ Id3/wt-p27 prom). Moreover Id3 controlled the cell-cycle by targeting p27 expression, we also inhibited the basal activity of the promoter. Taken tested the effect of Id3-siRNA on the proliferation of together these data indicate that ElK1 is a p27 promoter early passage ‘primary’ mouse embryonic fibroblasts activator which activity is counterparted by Id3. (MEFs) containing a homozygous targeted deletion of

Oncogene Id3 regulates p27mRNA in early G1 A-A Chassot et al 5777

Id3 mRNA 6

Id3-siRNA 4 Ct-siRNA

2 Id3 mRNA level (A.U.) 0

3 p27 mRNA

1 p27 mRNA level (A.U.)

0 conf 1h 3h 6h 24h

Hours post wounding

b NW 1243 Hrs after wounding

Id3

p27

Ct-siRNA + ++ Id3-siRNA +++

c e d 20 )

4 -3 18 16 ChIP assay

) 0.7kb

-4 0.5kb 8 2 0.4kb 6 1342 4 (A.U.x10 2 p27 promoter activity

promoter activity (A.U.x10 0 0 pcDNA3 Elk1 dn-Elk1 Elk1+Id3 pcDNA3 pcDNA Id3 Id3-Ala5 E12+E47 Elk1

wt-p27prom mut-p27prom

Figure 4 Effect of Id3 siRNA on (a) Id3 and p27 mRNA and (b) Id3 and p27 protein expression in response to injury. Confluent cells were transfected with Id3-siRNA and Ct-siRNA (100 nM) then left intact (NW) or wounded (W). At the indicated times after injury the levels of Id3 and p27 mRNA and protein were measured respectively by real time PCR (a) and western blot (b). (c) Effect of Id3, E2A and Elk1 on p27 promoter activity. A p27kip1 promoter construct was co-transfected as described in ‘Materials and methods’ in the presence of (i) pcDNA3 vector (pcDNA), (ii) a pcDNA3 vector encoding wild-type Id3 (Id3), (iii) a pcDNA3 vector encoding a variant of Id3 in which the Serine 5 was replaced by alanine (Id3-Ala5) (iv) PJ3omega vectors encoding E12 (E12) and E47 (E47) added together (E12 þ E47) and (v) a pcDNA vectors encoding wt-ELK1 (ELK1). The luciferase activity was normalized by b-galactosidase activity. (d) ChIPs were performed on extracts from confluent HDF cells as described in ‘Materials and methods’. Immuno- precipitation was performed using a non-relevant antibody (1) sepharose beads only (2) and a specific anti-ELK1 antibody (3) and primers spanning the p27kip1 promoter region were used for the PCRamplification. Lane 4 represents the oligonucleotidic fragment obtained after PCRamplification of the total genomic DNA before immunoprecipitation (input). ( e) P27 promoter activity in the presence of plasmids encoding wild type (wt-ELK1) or a dominant-negative variant of ELK1 (dn-ELK1). P27 promoter activity measured after cotransfection of plasmids encoding wt-ELK1 (200 ng) and Id3 (200 ng) (ELK1 þ Id3). The pcDNA3 histograms represent the basal activity of p27 promoter (wt-p27prom) and of p27 promoter mutated on the À453/À439 ELK1 site (mut-p27 prom) (hatched histogram). The data in (c, d and e) are from a typical experiment selected among two individual ones. the p27 gene (passage 2–4) (Nakayama et al., 1996) that p27À/À MEFs were effectively devoid of p27. and on their heterozygous counterparts (p27 þ /À). As Growing cells were then transfected with Id3-siRNAs shown in Figure 6a a real time PCRanalysis showed and Ct-siRNAs and Id3 and p27 expression were

Oncogene Id3 regulates p27mRNA in early G1 A-A Chassot et al 5778 12 Id3-siRNA

p27 Ct-siRNA 8

pRb 4 Stimulation p130 CycA mRNA Fold

0 NW W Cyc D1 d CyA p21 Id3 Ct Id3 Ct Id3 Ct Id3 Ct siRNA siRNA siRNA siRNA Bleo siRNA siRNA siRNA siRNA NW W NW W 10hrs e Ct-siRNA Id3-siRNA 40 29% b 30 14% p27 20 IP cycD1 Id3 Ct Id3 Ct 10 siRNA siRNA siRNA siRNA 0 W 10Hrs W 24 hrs % of Brdu positve cells Ct Id3 siRNA siRNA

Figure 5 Effect of p27 downregulation reversion by Id3-siRNA on the cell-cycle. Confluent HDF were transfected with control Ct-siRNA or Id3-siRNA as described in ‘Materials and methods’ and left intact (NW) or wounded (W). (a) Cells were harvested 10 h after injury, lysed then pRb (pRB) phosphorylation, p130 (p130), p27kip1 (p27), cyclin D1 (Cyc D1), and p21cip1 (p21) expressions were analysed by western blot. The arrows on pRb Western blot represent phosphorylated (upper arrow) and nonphosphorylated (lower arrow) forms of the protein. The lane ‘Bleo’ represented a condition where intact HDF were incubated with bleomycin (10 mg/ ml) to stimulate p21. (b) Cells were harvested at the indicated times and the level of p27 associated with CycD1 was analysed by Western blot after immunoprecipitation using an anti-CycD1 antibody as described in ‘Materials and methods’. (c) and (d) Cells treated with Ct-siRNA and Id3-siRNA were harvested 24 h after wounding then the levels of cyclin A mRNA and protein were respectively measured by real-time PCR( c) or western blot (d). (e) Effect of Id3-siRNA on S-phase entry measured by nuclear incorporation of BrdU and visualized by immunostaining 24 h after wounding with anti-BrdU monoclonal antibody. Confluents HDF were transfected with Ct or Id3-siRNA and wounded. Cells were fixed 24 h after injury (BrdU was added 6 h before cell fixation). The white arrows indicate BrdU-positive cells. The white line represents the ‘wound edges’. The nuclei are visualized by Hoechst 3358 (St Louis, MO, USA) staining. The shown images represent typical and representative fields. The histogram on the right represents the percentage of Brdu-positive nuclei versus total nuclei (Hoechst) obtained from the analysis of 12 independent fields resulting from two independent experiments (six fields per experiment).

analysed in parallel in p27À/À and p27 þ /À MEFs after studies indicate that, in cycling cells, p27 is degraded 36 h. during G1 and G1/S phases by mechanisms involving Data presented in Figure 6b show that Id3 was the ubiquitin-proteasome pathway (Pagano et al., 1995; expressed at equivalent levels in p27À/À and p27 þ /À Montagnoli et al., 1999; Malek et al., 2001; Kamura cells and was efficiently inhibited (60% inhibition) after et al., 2004). In this study, using a suitable wound- 36 h of transfection with Id3-siRNA. P27 was expressed healing model that stimulates the synchronous cell-cycle at low levels in cycling p27 þ /À MEF and its expression entry of HDF upon injury, we clearly show that p27 increased significantly in Id3-siRNA-treated cells, decrease is transcriptional regulated and, as a novelty, supporting the idea that in MEF, like in HDF, p27 we demonstrate that the transcriptional repressor Id3 regulation also involves Id3. We finally observed that plays a pivotal role in this process. However, although Id3 silencing inhibited about 50% the proliferation this transcriptional regulation appears crucial, it does of cycling of p27 þ /À MEF as assessed by the incor- not exclude the relevance of p27 protein degradation poration of BrdU (Figure 6c), but did not affect the in cell-cycle commitment. Indeed we observed that, in proliferation of p27À/À MEFs, indicating that the our system, the inhibition of p27 protein degradation regulation of p27 by Id3 was a key event in the cascade by proteasome inhibitors also prevented the cell-cycle of events controlling cell growth. (not shown). Together these observations suggest that the Id3-dependent transcriptional regulation and the proteasome post-translational degradation of p27 Discussion cooperate to allow an efficient protein p27 downregulation and the cell-cycle entry. Decreased p27 expression is one of the earlier events Id2/Id3 have been reported to regulate p27 in very observed when quiescent confluent cells are stimulated specific contexts (Everly et al., 2004; Kee and Bronner- to proliferate in the presence of mitogenic stimuli Fraser, 2005); however, their effect on p27 transcrip- (Malek et al., 2001; Kamura et al., 2004). Most of the tional control during cell proliferation had never been

Oncogene Id3 regulates p27mRNA in early G1 A-A Chassot et al 5779 5 b p27-/- p27+/- Id3 2.5 (A.U) p27

p27 mRNA level + + Id3-siRNA + + Ct-siRNA 0 p27-/- p27+/-

c Hoechst BrdU siRNAs

p27+/- 50 Id3 Id3-siRNA 40 Ct-siRNA 30

10 Ct % Brdu positive cells 0 p27-/- p27-/- p27+/-

Id3

Figure 6 Effect of Id3 silencing on p27-deficient MEFs proliferation. (a) Verification by real-time quantitative PCRof the absence of p27 in p27À/À MEFs. Total RNA of growing p27À/À and p27 þ /À MEFs were extracted and p27 was quantified by Q-PCR. The values on the vertical axis represent the relative amount of p27 in p27À/À and p27 þ /À.(b) Effect of Id3-siRNA on Id3 and p27 expression in p27À/À and p27 þ /À MEFs. Cells were transfected with Id3-siRNA ( þ ) or Ct-siRNA ( þ ) as described in ‘Material and method’; then p27 and Id3 were analysed after 36 h by Western blot. (c) Effect of Id3 silencing on BrdU incorporation. p27À/À and p27 þ /À MEFs were transfected with Id3-siRNA (Id3) and control-siRNA (Ct). After 24 h BrdU (10 mg/ml) was added and cells were harvested 12 h later. Cells were fixed and BrdU-positive cells BrdU were identified by immunofluorescence as indicated in ‘Material and methods’ (the presented image show one representative field out of 12). The nuclei are visualized by Hoechst 3358 staining (Hoechst). The data presented in (Figure 6b and c) are from one experiment chosen among two independent ones. The histograms on the right represents the percentage of Brdu-positive nuclei versus total nuclei (Hoechst) obtained after quantification of 12 independent fields (two distinct experiments, six fields per experiment) in p27À/À and p27 þ /Àrespectively. investigated. In our model, by using a siRNA approach TCF activity has been reported previously in the case we show that Id3 efficiently participates in the inhibition of the c-fos promoter (Yates et al., 1999). This inacti- of p27 by inactivating the transcriptional activity vation participates to the inhibition of c-fos transcrip- of its promoter thereby leading to a downregulation of tion after its classical early and transient stimulation by its mRNA and protein expression. It is worthnoting serum. Besides p27, Ids family members also participate that in our wounding model, Id3 has no effect on p21 to the transcriptional inhibition of p21cip1, another expression indicating that in this physiological situation, CKI. However, in contrast with the case of p27, this p27 is the major cyclin/Cdk inhibitor targeted by Id3. inhibitory mechanism involves E2A family members Under physiological conditions, Ids are transcrip- (Prabhu et al., 1997), indicating that although p21 and tional repressors of several transcription factors such as p27 exert similar effects in the control of the cell-cycle, E2A, Ets1/2, SAPs, Pax(2-5), TCFs, myoD, (Yates the pathways that regulate their expression are different. et al., 1999; Trausch-Azar et al., 2004; Roberts et al., Our study demonstrates that Id3-dependent p27 2001) and E2A family members are their most frequent downregulation is really critical in the control of cell partners (Norton, 2000). In our biological model, E2A proliferation in wounded cells, because its inhibition family members do not participate to the regulation of prevents the cell-cycle commitment and progression. p27 promoter and therefore are unlikely the partners of Chronologically, the first observed effect of Id3 silencing Id3 in this process. In contrast, reporter gene experi- (apart from its p27 regulation) was the inhibition of the ments and ChIP assays have shown that Id3 exerts its phosphorylation and degradation of the ‘pocket pro- inhibitory effect on p27 promoter by antagonizing the teins’ pRb and p130. In cycling cells, the phosphoryla- action of Elk1, a TCF Ets-domain transcription factor. tion of the ‘pocket proteins’ is a prerequisite for cyclin E The involvement of Id3 in the negative regulation of and A induction and G1/S transition (Sherr, 1995,

Oncogene Id3 regulates p27mRNA in early G1 A-A Chassot et al 5780 2002). In agreement with this notion, here we show that the widely studied genes involved in p27 degradation, the inhibition of the ‘pocket proteins’ by Id3-siRNAs is other genes responsible for the transcriptional control of associated with an inhibition of the G1/S transition as p27, like Id3, might play a key role in cell proliferation verified by the inhibition of cyclin A stimulation and and can be the target of genetic alterations leading to DNA synthesis. deregulated proliferation putatively responsible for The inhibition of ‘pocket proteins’ phosphorylation/ tumour development. degradation in early-mid-G1 likely results from a reduced cyclin D1-associated kinase activity as demonstrated by the weaker degradation of p130, a highly Materials and methods specific physiological substrate of cyclin D1-kinase, Cell culture observed in Id3-siRNA-treated HDF (Grana et al., NIH 3T3 cell line obtained from American Type Culture 1998). This reduced enzymatic activity is not a Collection were cultured in Dulbecco’s modified Eagle’s consequence of a decrease in cyclin D1 expression nor medium (DMEM) supplemented with fetal calf serum, 4- of an effect of Id3-siRNA on p16(ink4A) (not shown), but (2-hydroxyethyl)-1- piperazineethanesulfonic acid (HEPES) could be owing to a higher association of p27 with (20 mM), glutamine (2 mM), penicillin (1000 U/ml), and strep- cyclin-associated kinase in Id3-siRNA treated cells. The tomycin (1000 U/ml). Normal Dermal Fibroblasts (HDF) were effects of p27 on cyclinD1-CDK4/6 activities are still isolated from healthy neonatal foreskin according to Rhein- unresolved. Some studies evoke that p27Kip1 and wald and Green, (1975) and cultured in DMEM supplemented p21Cip1 function as obligate assembly factors for D with 10% bovine calf serum, HEPES (20 mM), glutamine cyclin-cdk4 complexes (Cheng et al., 1999) and that D (2 mM), penicillin (1000 U/ml), and streptomycin (1000 U/ml). P27-deficient (p27À/À) MEF and their p27 þ /À heterozygote cyclin-cdk4 complexes containing a single p27Kip1 or counterparts are primary MEFs isolated from p27 knockdown p21Cip1, are active (Blain et al., 1997; LaBaer et al., mice embryo (Nakayama et al., 1996) and cultured at low 1997). Other studies argue that D cyclin-cdk4 complexes passage (2–4). Cells were grown into DMEM supplemented containing p27Kip1 or p21Cip1 are always inactive with 10% bovine calf serum HEPES (20 mM), glutamine (Bagui et al., 2003). The inhibition of p130 degradation (2 mM), penicillin (1000 U/ml), streptomycin (1000 U/ml). that we observed in Id3-siRNA-treated cells seems to Confluent HDF were stimulated to proliferate by calibrated indicate that, in our biological context, the increased scratches (injuries) using a device that allow wounding up to association of p27 with cyclin D1-associated kinase 40% of the cell population and cells were harvested at the reduces its enzymatic activity. indicated times. For the replating experiments, confluent cells Taken together our data show that the Id3-mediated were trypsinized and diluted 1 : 4 in fresh medium. For serum stimulation experiment, low-density cell cultures were depleted p27 decrease is a crucial event for the cell-cycle 36 h in DMEM supplemented with 0.1% bovine serum progression and S-phase entry. The specificity of Id3 albumin, HEPES (20 mM), glutamine (2 mM), penicillin toward p27 is supported by experiments carried out in (1000 U/ml), streptomycin (1000 U/ml) before addition of p27-deficient primary MEFs and showing that, in the serum (20% final). absence of p27, Id3-siRNA does not anymore exert its antiproliferative effect. Reagents and antibodies After its down-regulation in early G1, the level of p27 DMEM, trypsin, EDTA, HEPES, glutamine, penicillin, mRNA steadily increased as soon as cells approached and streptomycin were purchased from LifeTechnologies the G1–S transition. We observed that the amplitude of (Carlsbad, CA, USA). Bovine calf serum was obtained from this second wave of p27 increase was significantly higher Hyclone. bovine serum albumin, Hoechst 33258, actinomycin when Id3 was silenced, indicating that, under normal D and cycloheximide were purchased from Sigma (St Louis, conditions, Id3 (besides its immediate function on p27 MO, USA). Anti-Id3 (C20), anti-p21 (C19), anti-p27 (C19) transcription) is also involved in the transcriptional and anti-p130 (C20) antibodies were obtained from Santa- Cruz (tebu-bio, Le Perrey en Yvelines, France). Monoclonal machinery controlling the steady-state level of p27 at anti-cyclin D1 (DCS-6) antibody was from Transduction the G1/S boundary. It is worthnoting that the levels of Laboratories (Lexington, KY, USA); monoclonal anti-cyclin Id3 measured 15–24 h after injury are lower than the A was from Novocastra and monoclonal anti-BrdU antibody ones measured immediately after injury, which probably was from Roche/Amersham (Les Ulis, France). Horseradish explains why, in cells approaching the G1/S transition, peroxidase-conjugated and FITC-conjugated antibodies IgG Id3 does not inhibit p27 mRNA expression but only were purchased from Dako (Glostrup, Denmark). modulates its protein levels. To conclude, this work reinforces the role of Id3 in Plasmids p27 regulation and consequently in the cell-cycle The promoter region of p27Kip1 was described in Minami et al. progression. Id3 is involved in the control of the early (1997) and was a generous gift of Dr Sakai. The pcDNA3 degradation phase of p27 mRNA but also participates constructs encoding wild type Id3 and serine to alanine 5 to the control of p27 mRNA stability at the G1/S mutated Id3 (Id3-Ala5) was generated as described previously boundary. P27 is not mutated in transformed cells but it (Deed et al., 1997) and provide by Dr JD Norton. The PJ3omega plasmids encoding E12 and E47 were a gift of Dr R presents low reduced expression in tumours (Okabe Stein and were previously described by Sharma (Sharma et al., et al., 2001; Kudo et al., 2005). Recent evidence suggests 1997). The pcDNA3 plasmids encoding wt-ELK1 and a that Skp2, the specific recognition factor for p27 ubiqui- dominant-negative variant (ELK1-dn) were kindly provided tination, is endowed of oncogenic properties (Bloom by Dr Robert Hipskind and were described elsewhere and Pagano, 2003). Our study shows that, apart from (Janknecht et al., 1993).

Oncogene Id3 regulates p27mRNA in early G1 A-A Chassot et al 5781 The mutation of the Elk1 site in p27 promoter was perfor- 10 min treatment with HCl (2 N). Cells were extensively washed med using the ‘quick change direct mutagenesis’ (Stratagene, with PBS, then incubated 10 min at room temperature with La Jolla, CA, USA) according the manufacturer instruction. PBS þ 0.1% Tween, washed five times with PBS and finally The primer used for mutagenesis was CGTCAGCCTCCCTT blocked for 30 min at room temperature in PBS þ 0.1% BSA. CGGATCCGAATTCGCCACTAAAAAAAGG. Cells were incubated with anti-BrdU antibody (1/10001) for 2 h at room temperature, washed extensively with PBS and Transfection and luciferase assay incubated with anti-mouse FITC-conjugated antibodies (1/ NIH 3T3 murine fibroblasts were seeded in 48-wells plates, at a 501) for 1 h at room temperature. Hoechst (1/10001) was added density of 25000 cells/mm2, and were co-transfected using the for 10 min, cells were washed in PBS in distilled water and calcium phosphate technique with 0.1 mg of the plasmids mounted in Mowiol (Calbiochem, Meudon, France). In the containing wt or ELK1-mutated p27Kip1 promoter in the case of p27-/- and p27 þ /- MEFs, BrdU (10 mg/ml) was added presence of 0.3 mg of each of the different expression vectors 24 h after siRNA transfection and cells were maintained in encoding Id3, Id3-Ala5 but also E12, E47, wt-ELK1 and culture for 12 additional hours. Cells were fixed in 3.7% dn-ELK1. Cells were harvested 24 h after transfection and formaldehyde and incubated with anti-BrdU antibodies as luciferase activity assay was performed accordingly to the described above. BrdU-positive HDF and MEF were counted manufacturer’s instruction (Promega, Madison, WI, USA). using the ImageJ Software (http://rsb.info.nih.gov/ij). Transfection efficiency was normalized using a pCMV-bGal. ChIP assay RNA extraction and real-time quantitative PCR Confluent HDF were treated with 1% formaldehyde for 5 min Total RNA was isolated from HDF according to Rezzonico at 371C and for 30 min at 41C. Next, cells were harvested et al. (1995). mRNA expression level was quantified by real- and suspended in the lysis buffer (10 mM Hepes, 1.5 mM time one-step PCRperformed in an ABI PRISM5700 MgCl2,10mM KCl, 0.5 mM DTT, 0.1% NP-40, and protease Sequence Detector System (Applied Biosystems, Foster City, inhibitors). After 10 min incubation on ice, cells were CA, USA) using the SYBRGreen detection protocol as out- centrifuged (5000 r.p.m., 5 min at 41C) to pellet the nuclei. lined by the manufacturer. Gene-specific primers were Nuclei were then suspended in the nuclei lysis buffer (50 mM designed using the Primer Express software (Applied Bio- Tris–HCl, pH 8.1, 10 mM ethylenediaminetetraacetic acid, 1% systems). The primer sequences can be obtained upon request. sodium dodecyl sulphate and protease inhibitors) and incu- Relative expression level of target genes was normalized for bated on ice for 10 min. The sheared chromatin was then RNA concentrations with two different house-keeping genes immunoprecipitated using (or not) a specific polyclonal anti- (GAPDH and a-actin). mRNA values are expressed as arbi- 7 ELK1 antibody (1/200), (Santa Cruz I120). After immuno- trary units and represent standard deviation (s.d.) of tripli- precipitation, the crosslink was reverted by heat treatment cates, and are representative of two independent experiments. (671C overnight) and proteinase K digestion. The gen omic captured fragments were then recovered by phenol–chloro- Small RNA interference experiments (siRNA) form extraction. The captured p27kip1 promoter fragments Confluent HDF were transfected for 24 h with 100 nM siRNA, were identified by PCRanalysis using specific primers using the LIPOFECTAMINE reagent (Invitrogen, Carlsbad, spanning the promoter. Thirty cycles of PCRwere performed CA, USA) according to the manufacturer procedure then were and the amplified products were analysed on a 2% agarose gel. submitted to injury. In the case of p27-/- and p27 þ /- primary The oligonucleotides to amplify p27kip1 promoter region were MEFs, the cells (50% of confluence) were transfected as 50-GTTTGTTGGCAGCAGTACC-30 (forward) and 50-CCG described above, and were left in culture for 12 additional CTGATCAAATGGACTGG-30 (reverse). hours after siRNA transfection. Id3-siRNA sequence was described previously (Kowanetz et al., 2004) and control- siRNA (Ct-siRNA) targeted the firefly luciferase mRNA Acknowledgements (Busca et al., 2005). We are grateful to Drs JD Norton for generous gifts of the Protein extracts and Immunoblotting analysis expression vectors encoding wt-Id3 and Id3-ala5, to Dr T The HDF lysates preparation and Western blot procedures Sakai for the plasmid containing p27 promoter, to Dr RStein were described previously (Turchi et al., 2002). The antibody for the E12 and E47 constructs, to Dr C Sardet and JM concentrations used in Western blot experiments were: anti- Blanchard for the p27-null MEFs and to Dr RHipskind for Id3 (1/2001), anti-p21 (1/2001), anti-p27 (1/2001), anti-pRB (1/ kindly providing the elk1 constructs. AA Chassot is recipient 5001), anti-p130 (1/2001), anti-cyclin D1 (1/2501) or anti-cyclin of a fellowship from the INSERM/Re´ gion PACA and from la A (1/2501). Ligue Nationale contre le Cancer (LNCC). L Turchi is recipient of a fellowship of the Centre Hospitalier Universi- BrdU-labelling experiments taire (CHU) de Nice. This work was supported by grants of the Confluent siRNA transfected HDF were wounded. After 18 h, Institut National de la Sante´ et la Recherche Me´ dicale BrdU (10 mg/ml) was added and cells were collected 6 h later (INSERM), L’association pour la Recherche contre le Cancer, (24 h after injury), cells were fixed in 3.7% paraformaldehyde (ARC grant N1 5965 and 4500) and of the ‘Ministe` re de for 20 min at room temperature, rinsed in phosphate-buffered l’Education Nationale et de la Recherche’ (MENR) (ACI ACI: saline (PBS) and the chromatin was rendered accessible by a TS/0220045).

References

Alani RM, Young AZ, Shifﬂett CB. (2001). Id1 regulation of Bagui TK, Mohapatra S, Haura E, Pledger WJ. (2003). cellular senescence through transcriptional repression of P27Kip1 and p21Cip1 are not required for the formation of p16/Ink4a. Proc Natl Acad Sci USA 98: 7812–7816. active D cyclin-cdk4 complexes. Mol Cell Biol 23: 7285–7290.

Oncogene Id3 regulates p27mRNA in early G1 A-A Chassot et al 5782 Benezra R, Rafii S, Lyden D. (2001). Id proteins and mellitus) show disturbed proliferation. Arch Dermatol Res angiogenesis. Oncogene 20: 8288–8289. 291: 93–99. Blain SW, Montalvo E, Massague J. (1997). Differential Lyden D, Young AZ, Zagzag D, Yan W, Gerald W, O’Reilly R interaction of the cyclin-dependent kinase (Cdk) inhibitor et al. (1999). Id1 and Id3 are required for neurogenesis, p27Kip1 with cyclin A-Cdk2 and cyclin D2-Cdk4. J Biol angiogenesis and vascularization of tumor xenografts. Nature Chem 272: 25863–25872. 401: 670–677. Bloom J, Pagano M. (2003). Deregulated degradation of the Malek NP, Sundberg H, McGrew S, Nakayama K, Kyriakides cdk inhibitor p27 and malignant transformation. Semin TR, Roberts JM. (2001). A mouse knock-in model exposes Cancer Biol 13: 41–47. sequential proteolytic pathways that regulate p27Kip1 in G1 Busca R, Berra E, Gaggioli C, Khaled M, Bille K, Marchetti B and S phase. Nature 413: 323–327. et al. (2005). Hypoxia-inducible factor 1{alpha} is a new Martin P. (1997). Wound healing – aiming for perfect skin target of microphthalmia-associat. J Cell Biol 170: 49–59. regeneration. Science 276: 75–81. Calderon M, Lawrence WT, Banes AJ. (1996). Increased Massari ME, Murre C. (2000). Helix-loop-helix proteins: proliferation in keloid fibroblasts wounded in vitro. J Surg regulators of transcription in eucaryotic organisms. Mol Cell Res 61: 343–347. Biol 20: 429–440. Cheng M, Olivier P, Diehl JA, Fero M, Roussel MF, Roberts Minami S, Ohtani-Fujita N, Igata E, Tamaki T, Sakai T. JM et al. (1999). The p21(Cip1) and p27(Kip1) CDK (1997). Molecular cloning and characterization of the human ’inhibitors’ are essential activators of cyclin D-dependent p27Kip1 gene promoter. FEBS Lett 411: 1–6. kinases in murine fibroblasts. EMBO J 18: 1571–1583. Montagnoli A, Fiore F, Eytan E, Carrano AC, Draetta GF, Chraibi H, Dereure O, Teot L, Guillot B. (2004). The Hershko A et al. (1999). Ubiquitination of p27 is regulated diagnosis and treatment of carcinomas occurring at the sites by Cdk-dependent phosphorylation and trimeric complex of chronic pressure ulcers. J Wound Care 13: 447–448. formation. Genes Dev 13: 1181–1189. Deed RW, Hara E, Atherton GT, Peters G, Norton JD. Nakayama K, Ishida N, Shirane M, Inomata A, Inoue T, (1997). Regulation of Id3 cell-cycle function by CDK2- Shishido N et al. (1996). Mice lacking p27(Kip1) display dependent phosphorylation. Mol Cell Biol 17: 6815–6821. increased body size, multiple organ hyperplasia, retinal Engel I, Murre C. (2001). The function of E- and Id proteins in dysplasia, and pituitary tumors. Cell 85: 707–720. lymphocyte development. Nat Rev Immunol 1: 193–199. Norton JD. (2000). ID helix-loop-helix proteins in cell growth, Everly Jr DN, Mainou BA, Raab-Traub N. (2004). Induction differentiation and tumorigenesis. J Cell Sci 113: 3897–3905. of Id1 and Id3 by latent membrane protein 1 of Epstein- Ohtani N, Zebedee Z, Huot TJG, Stinson JA, Sugimoto M, Barr virus and regulation of p27/Kip and cyclin-dependent Ohashi Y et al. (2001). Opposing effects of Ets and Id kinase 2 in rodent fibroblast transformation. J Virol 78: proteins on p16ink4 expression during cellular senescence. 13470–13478. Nature 409: 1067–1070. Grana X, Garriga J, Mayol X. (1998). Role of the Okabe M, Inagaki H, Murase T, Inoue M, Nagai N, Eimoto T. retinoblastoma protein family, pRB, p107 and p130 in the (2001). Prognostic significance of p27 and Ki-67 expression in negative control of cell growth. Oncogene 17: 3365–3383. mucoepidermoid carcinoma of the intraoral minor salivary Hosack DA, Dennis Jr G, Sherman BT, Lane HC, Lempicki RA. gland. Mod Pathol 14: 1008–1014. (2003). Identifying biological themes within lists of genes Pagano M, Tam SW, Theodoras AM, Beer-Romero P, Del Sal with EASE. Genome Biol 4: 11. G, Chau V et al. (1995). Role of the ubiquitin-proteasome Janknecht R, Ernst WH, Pingoud V, Nordheim A. (1993). pathway in regulating abundance of the cyclin-dependent Activation of ternary complex factor Elk-1 by MAP kinases. kinase inhibitor p27. Science 269: 682–685. EMBO J 12: 5097–5104. Prabhu S, Ignatova A, Park S, Sun XH. (1997). Regulation of Kamura T, Hara T, Matsumoto M, Ishida N, Okumura F, the expression of cyclin-dependent kinase inhibitor p21 by Hatakeyama S et al. (2004). Cytoplasmic ubiquitin ligase E2A and Id proteins. Mol Cell Biol 17: 5888–5896. KPC regulates proteolysis of p27(Kip1) at G1 phase. Nat Rezzonico R, Ponzio G, Loubat A, Lallemand D, Proudfoot A, Cell Biol 6: 1229–1235. Rossi B. (1995). Two distinct signalling pathways are involved Kee Y, Bronner-Fraser M. (2005). To proliferate or to die: role in the control of the biphasic junB transcription induced of Id3 in cell-cycle progression and survival of neural crest by interleukin-6 in the B cell hybridoma 7TD1. JBiolChem progenitors. Genes Dev 19: 744–755. 270: 1261–1268. Kowanetz M, Valcourt U, Bergstro¨ m R, Heldin CH, Rheinwald JG, Green H. (1975). Serial cultivation of strains of Moustakas A. (2004). Id2 and Id3 define the potency of human epidermal keratinocytes: the formation of keratiniz- cell proliferation and differentiation responses to transforming colonies from single cells. Cell 6: 331–343. ing growth factor beta and bone morphogenic protein. Roberts EC, Deed RW, Inoue T, Norton JD, Sharrocks AD. Mol Cell Biol 24: 4241–4254. (2001). Id helix-loop-helix proteins antagonize pax tran- Kudo Y, Kitajima S, Ogawa I, Miyauchi M, Takata T. (2005). scription factor activity by inhibiting DNA binding. Mol Downregulation of Cdk inhibitor p27 in oral squamous cell Cell Biol 21: 524–533. carcinoma. Oral Oncol 41: 105–116. Sakakibara K, Kubota K, Worku B, Ryer EJ, Miller JP, Koff A LaBaer J, Garrett MD, Stevenson LF, Slingerland JM, et al. (2005). PDGF-BB regulates p27 expression through Sandhu C, Chou HS et al. (1997). New functional activities ERK-dependent RNA turn-over in vascular smooth muscle for the p21 family of CDK inhibitors. Genes Dev 11: cells. J Biol Chemi 280: 25470–25477. 847–862. Servant JM, Coulombe P, Turgeon B, Meloche S. (2000). Lasorella A, Noseda M, Beyna M, Iavarone A. (2000). Id2 is a Differential regulation of p27kip1 expression by mitogenic retinoblastoma protein target and mediates signalling by and hypertrophic factors: involvement of transcriptional and Myc oncoproteins. Nature 407: 592–598. posttranscriptional mechanisms. J Cell Biol 148: 543–556. Loots MA, Lamme EN, Mekkes JR, Bos JD, Middelkoop E. Sharma A, Henderson E, Gamer L, Zhuang Y, Stein R. (1999). Cultured fibroblasts from chronic diabetic wounds (1997). Analysis of the role of E2A-encoded proteins in on the lower extremity (non-insulin-dependent diabetes insulin gene transcription. Mol Endocrinol 11: 1608–1617.

Oncogene Id3 regulates p27mRNA in early G1 A-A Chassot et al 5783 Sherr CJ. (1995). D-type cyclins. Trends Biochem Sci 20: 187–190. Turchi L, Chassot AA, Rezzonico R, Yeow K, Loubat A, Sherr CJ. (2002). D1 in G2 Mammalian G1 cyclins and cell- Ferrua B et al. (2002). Dynamic characterization of the cycle progression. Cell-Cycle 1: 36–38. molecular events during in vitro epidermal wound healing. Sherr CJ, Roberts JM. (1995). Inhibitors of mammalian G1 J Invest Dermatol 119: 56–63. cyclin-dependent kinases. Genes Dev 9: 1149–1163. Yates PR, Atherton GT, Deed RW, Norton JD, Sharrocks AD. Takahashi E, Funato N, Higashihori N, Hata Y, Gridley T, (1999). Id helix-loop-helix proteins inhibit nucleoprotein Nakamura M. (2004). Snail regulates p21 waf/cip1 expres- complex formation by the TCF ETS-domain transcription sion in cooperation with E2A and Twist. Biochem Biophys factors. EMBO J 18: 968–976. Res Commun 325: 1136–1144. Zheng W, Wang H, Xue L, Zhang Z, Tong T. (2004). Trausch-Azar JS, Lingbeck J, Ciechanover A, Schwartz AL. Regulation of cellular senescence and p16ink4 expression by (2004). Ubiquitin-Proteasome-mediated degradation of Id1 Id1 and E47 protein in human diploid ﬁbroblast. J Biol is modulated by MyoD. J Biol Chem 279: 32614–32619. Chem 279: 31524–31532.

Oncogene