Biochimica et Biophysica Acta 1809 (2011) 316326

Contents lists available at ScienceDirect

Biochimica et Biophysica Acta

j ourna l homepage: www.e lsev ie r.com/ locate /bbagrm

The nuclear exporter, Crm1, is regulated by NFY and Sp1 in cancer cells and repressedby p53 in response to DNA damage

Pauline J. van der Watt, Virna D. Leaner Division of Medical Biochemistry, Institute of Infectious Disease and Molecular Medicine, Faculty of Health Sciences, University of Cape Town, South Africa

Corresponding author at: Division ofMedical BiochemUniversity of Cape Town, Observatory, Cape Town, 7925,6250; fax: +27 21 406 6061.

E-mail address: Virna.Leaner@uct.ac.za (V.D. Leaner

1874-9399/$ see front matter 2011 Elsevier B.V. Adoi:10.1016/j.bbagrm.2011.05.017

a b s t r a c t

a r t i c l e i n f o

Article history:Received 11 January 2011Received in revised form 20 May 2011Accepted 27 May 2011Available online 13 June 2011

Keywords:Crm1NFYSp1p53Cancer

The nuclear exporter protein, Crm1, plays a key role in normal cell functioning, mediating the nucleo-cytoplasmic transport of cargo proteins. Elevated Crm1 expression has recently been identified in varioustumours; however, themechanisms driving its expression have not been investigated to date. In this study weidentified the Crm1 promoter and factors associated with its elevated expression and with its repressionunder conditions of DNA damage. The 1405 to +99 Crm1 promoter region was found to be significantlymore active in cancer and transformed cells compared to normal, and the 175 to +99 region identified asresponsible for the differential activity. Mutation of two CCAAT boxes and a GC box within this regionsignificantly diminished Crm1 promoter activity and ChIP analysis revealed binding of NFY and Sp1 to thesesites, with increased binding in transformed and cancer cells. In addition, p53 was found to repress Crm1promoter activity, after induction with doxorubicin, with p53 siRNA blocking the effect. Crm1 promoterconstructs withmutated CCAAT boxes were significantly less responsive to p53 repression, and in vivo bindingof NFY to the CCAAT boxes was diminished upon p53 binding, suggesting that p53 mediates repression of theCrm1 promoter via interfering with NFY. This was confirmed using NFY knock-down cells, in which Crm1promoter activity was significantly less responsive to p53. In vitro EMSAs revealed that NFY and p53 bind theCCAAT boxes as a single complex under conditions of DNA damage. In summary, this study is a first to analyseCrm1 promoter regulation and reveals NFY and Sp1 as contributors to Crm1 overexpression in cancer. Inaddition, this study reveals that Crm1 transcription is inhibited by DNA damage and that the mechanism ofinhibition involves p53 interfering with NFY function.

istry, Faculty ofHealth Sciences,South Africa. Tel.: +27 21 406

).

ll rights reserved.

2011 Elsevier B.V. All rights reserved.

1. Introduction

Crm1 (the Chromosome region maintenance 1 protein orExportin 1) is a member of the Karyopherin protein family andthe major nuclear export receptor in the cell [1]. It mediates thenuclear export of cargo proteins and certain RNAs from the nucleusinto the cytoplasm, across the nuclear pore complex (NPC), and thusfacilitates protein and RNA subcellular localisation. Crm1 recognisescargo proteins that carry a nuclear export signal (NES) [2], typicallycontaining three to four critically spaced leucines (LX(13)LX(23)LXJ,L: leucine, X: spacer, J: leucine, valine or isoleucine) [3,4]. Such cargoproteins include various transcription factors [5], cell cycle proteins[6,7], and signalling proteins [8], which require timely translocationacross the nuclear envelope. Many of the integral processes in the cellthus rely on Crm1 expression and function [9].

Recent studies have reported that the expression of Crm1 isaltered in cancer. Crm1 protein levels are elevated in cervical cancer

[10], ovarian cancer [11], osteosarcoma [12], glioma [13] andpancreatic cancer [14], with high levels of Crm1 being found toassociate with poor patient survival [1114]. Moreover, Crm1expression has been found to be elevated in transformed fibroblastscompared to normal fibroblasts, suggesting that the increasedexpression of Crm1 is a general feature of the transformed phenotype[10]. The high expression of Crm1 in cancer and transformed cells isfunctionally relevant, as the inhibition of its expression results in celldeath via apoptosis, while inhibition of its expression in normal cellsdoes not [10]. This implicates Crm1 as a potential anti-cancer drugtarget, and currently studies are underway aiming to develop effectiveinhibitors of Crm1 [15].

The increased expression of Crm1 protein in cancer derives fromincreased transcription of Crm1mRNA [10], suggesting that a transcrip-tional regulatory mechanism exists for differential Crm1 expression innormal and cancer cells, possibly at the level of promoter control.However, little is known about the factors that regulate the Crm1promoter. The aim of this study was therefore to investigate Crm1promoter activity and to identify cis- and trans-elements necessary forhigh Crm1 expression in cancer cells. We report that NFY/CBP, Sp1 andp53 transcription factors bind the Crm1 promoter and play an importantrole in Crm1 promoter regulation in cancer and transformed cells.

http://dx.doi.org/10.1016/j.bbagrm.2011.05.017mailto:Virna.Leaner@uct.ac.zahttp://dx.doi.org/10.1016/j.bbagrm.2011.05.017http://www.sciencedirect.com/science/journal/18749399

317P.J. van der Watt, V.D. Leaner / Biochimica et Biophysica Acta 1809 (2011) 316326

2. Materials and methods

2.1. Cell culture

WI38 normal lung fibroblasts, SVWI38 transformed WI38 fibro-blasts, and CaSki cervical cancer cells were obtained from theAmerican Type Culture Collection (ATCC) (Rockville, MD, USA). Cellswere maintained in Dulbecco's Modified Eagle's Medium (DMEM)supplemented with penicillin (100 U/ml), streptomycin (100 g/ml)and 10% Fetal Bovine Serum (FBS) (Gibco, Paisley, Scotland). All cellswere cultured at 37 C in a humidified atmosphere of 5% CO2.

2.2. Protein harvest and Western blot analysis

Cells in culture were grown to 80% confluency and lysed on ice inRIPA buffer (10 mM TrisCl, pH 7.4, 150 mM NaCl, 1% deoxycholate,0.1 % SDS, 1 % Triton X-100, 1 Complete Protease Inhibitor Cocktail(Roche, Basel, Switzerland)). Western blot analyses were performedusing the rabbit anti-Crm1 (H-300) (sc-5595, Santa Cruz Biotechnology,SantaCruz, CA,USA), rabbit anti--tubulin (H-235) (sc-9104, SantaCruzBiotechnology), rabbit anti-NFYA (H-209) (sc-10779, Santa CruzBiotechnology), rabbit anti-Sp1 (H-225) (sc-14027, Santa Cruz Biotech-nology),mouse anti-p65 (F-6) (sc-8008, Santa CruzBiotechnology), andmouse anti-p53 (M7001, DakoCytomation, Glostrup, Denmark)antibodies.

2.3. Construction of the Crm1 (1405 to +99) promoter construct

Primers were designed for amplification of the region from1405to +99 (with the transcription start site designated +1) of the Crm1gene (GenBank Accession Number: NC_000002). Primer sequenceswere: Crm1 F 5-AGGCTAGCGCGTTTTTCTTTATTGGAG-GG-3 (NheIsite is underlined) and Crm1 R 5-AGAAGCTTCTGTTGCTCTTGCT-GATGCT-3 (HindIII site is highlighted in bold). PCR was performedusing normal blood DNA as template and the high fidelity Expand PlusDNA Polymerase (Roche). Promoter PCR products were subclonedinto the vector pGEM-T Easy (Promega, Madison, WI, USA) andexcised with NheI and HindIII restriction enzymes for subcloning intothe luciferase reporter plasmid pGL3-Basic (Promega). Clonedpromoter constructs were verified by sequencing using the UCTHuman Genetics Sequencing Unit.

2.4. Generation of Crm1 promoter deletion constructs

Promoter deletion constructs were generated using either restric-tion sites common to both the cloned promoter fragment and thepGL3-Basicmultiple cloning site (SacI for the394 to+99 construct;NheI for the82 to+99 construct), or gene-specific forward primers(5-AGGCTAGCTCTCCGTCACCGAAGAGG-3 for the 500 to +99construct; 5-AGGCTAGCGTGAGGCGGGATTGACTG-3 for the 175to +99 construct; NheI restriction site is underlined) and the Crm1 Rprimer.

2.5. Luciferase assays

100 ng of each promoter construct was transfected into 30000cells/well of a 24-well plate, using 0.3 l Transfectin (Bio-Rad,Hercules, CA, USA). To normalise for transfection efficiency the cellswere co-transfected with 10 ng of the pRL-TK plasmid (Promega) thatencodes Renilla luciferase. Total cell lysates were prepared from cells24 h post-transfection using Passive Lysis Buffer (Promega) and fireflyluciferase activity was assayed using the Dual Luciferase Kit(Promega). Luminescence was monitored using the Glomax 96microplate luminometer (Promega).

2.6. Site-directed mutagenesis of potential transcription factorbinding sites

Mutations in potential transcription factor binding sites in the Crm1(175 to +99) promoter construct were prepared by site-directedmutagenesis, using mutagenic primers. Primer sequences were:

CCAATbox1mut: 5-GACTGGCGCTCGGGCtAgcGAGAGATAGGCT-CATG-3 and

5-CATGAGCCTATCTCTCgcTaGCCCGAGCGCCAGTC-3,YYmut: 5-CAATGAGAGATAGGCTagcGGCGCGGCTGTAGCTG-3

and5-CAGCTACAGCCGCGCCgctAGCCTATCTCTCATTG-3,CCAATbox2mut: 5-CTGTAGCTGGAACTGACgAATtcGAGGGCGGG-

CTGGGGG-3 and 5-CCCCCAGCCCGCCCTCgaATTcGTCAGTTCCAGCTA-CAG-3,

Sp1site1mut: 5-GGCGGGCTGGGGGCtagcGCCCAGGTTCCAAAC-3and

5-GTTTGGAACCTGGGCgctaGCCCCCAGCCCGCC-3,Sp1site2mut: 5-GC-TGGAACTGACCAATgctAGcGCGGGCTGGGGGC-

GGAG-3 and5-CT CCGCCCCCAGCCCGCgCTagcATTGGTCAGTTCCAGC-3,GCboxmut: 5-GTTTGAGGCACTAGctcGAGGGGGAGAAGCGG-3

and5-CCGCTTCTCCCCCTCgagCTAGTGCCTC-AAAC-3

(mutated bases are indicated in lowercase; NheI restriction sites areunderlined; EcoRI restrictions sites are in bold; XhoI restriction sitesare double underlined). Following PCR amplification, 10 U DpnI(Promega) was added and digestion carried out at 37 C for 90 min,before transformation into JM109 highly competent cells (Promega).Restriction digestion analysis was used to identify clones carrying therespective mutations.

2.7. siRNA transfection

For the effect of siRNAonCrm1promoter activity, 40 000 CaSki cells/well of a 24-well plate were transfected with 30 nM control siRNA-A(sc-37007, Santa Cruz Biotechnology), NFYA siRNA (sc-29947), Sp1siRNA (sc-29487) or p53 siRNA (sc-29435), and 24 h later transfectedwith 50 ng Crm1 (175 to +99) promoter construct and 5 ng pRL-TKplasmid. Luciferase activity was measured 24 h later and expressedrelative to Renilla luciferase in the same extract.

2.8. Chromatin immunoprecipitation (ChIP) assay

Cells were grown to approximately 90 % confluency and proteinDNA complexes cross-linked with 1% formaldehyde for 10 min,followed by the addition of 0.125 M Glycine, pH 2.5. Cells wereharvested, lysed in lysis buffer (1% SDS, 5 mMEDTA, 50 mMTrisCl, pH8.1, 1 Complete Protease Inhibitor (Roche)), and sonicated to lengthsof between400 and1000 bp. Cell lysateswerediluted in dilutionbuffer(1 % Triton X-100, 2 mM EDTA, 150 mM NaCl, 20 mM TrisCl, pH 8.1,1 Complete Protease Inhibitor) and then precleared with protein-Aagarose beads (Merck, NJ, USA) for 2 h. Beads had been previouslyblocked in 100 g/ml salmon sperm DNA and 5% BSA. Chromatin wasincubated with 2 g antibody (-NFYA (H-209 X, sc-10779 X, SantaCruz Biotechnology); -Sp1 (H-225, sc-14027, Santa Cruz Biotechnol-ogy); or no antibody negative control) at 4 C overnight. Protein-Aagarose beads were added for a further 2 h at 4 C, and immunocom-plexes bound by the beads recovered by centrifugation and washedtwice sequentially in TSE I (0.1% SDS, 1% Triton X-100, 2 mM EDTA,20 mM TrisCl, pH 8.1, 150 mM NaCl), TSE II (0.1% SDS, 1% Triton X-100, 2 mM EDTA, 20 mM TrisCl, pH 8.1, 500 mM NaCl), Buffer III(0.25 M LiCl, 1% NP-40, 1% Sodium Deoxycholate, 1 mM EDTA, 10 mMTrisCl, pH8.1) andTE, pH7.4. Boundmaterialwas elutedusingelution

318 P.J. van der Watt, V.D. Leaner / Biochimica et Biophysica Acta 1809 (2011) 316326

buffer (1% SDS, 0.1 M NaHCO3) at room temperature for 10 min, andthe input and eluted samples heated at 65 C overnight to reverse theformaldehyde cross-links. DNA was purified and used for real-timePCR, using primers designed to span the respective transcription factorbinding sites (NFY F: 5-GTGAGGCGG-GATTGACTG-3; NFY R: 5-TGCTGATGCT-GTAGCTCCC-3; Sp1 F: 5-CCATGGGCTGA-AAACTCAAC-3; Sp1 R: 5-GACTGCAGCA-GCAAAGACTG-3). These primers ampli-fied a 264 bp and 200 bp product, respectively.

2.9. Doxorubicin treatment

Doxorubicin (Adriamycin) was obtained from Sigma (St Louis, MO,USA). Cells were treated with 0.2 g/ml or 2 g/ml doxorubicin for allexperiments. For the analysis of the effect of doxorubicin on p53protein, cells were treated with doxorubicin for 8 h, while foranalysing effects on Crm1 protein levels, cells were treated withdoxorubicin for 48 h (due to the long half-life of Crm1). For analysingthe effects of doxorubicin on the Crm1 promoter, cells weretransfected with the promoter constructs and 8 h later treated withdoxorubicin for 18 h. For ChIP analysis of p53 binding to the Crm1promoter, CaSki cells were treated with 0.2 g/ml doxorubicin for18 h. For p53 or NFYA knock-down analysis, CaSki cells weretransfected with control siRNA-A (sc-37007, Santa Cruz Biotechnol-ogy) or p53 siRNA (sc-29435) or NFYA siRNA (sc-29947) 24 h prior todoxorubicin treatment.

2.10. EMSAs

Electrophoretic mobility shift assays were performed using 32P-labeled oligonucleotides (21-mers) containing either CCAAT box 1 orCCAAT box 2 and nuclear protein harvested from CaSki cells treatedwith 0.2 g/ml doxorubicin for 18 h. 5 g protein was used forexperiments and incubated with 2 g poly(dIdC). For supershiftexperiments, 2 g -NFYA (H-209) (sc-10779, Santa Cruz Biotech-nology) or -p53 (M7001, DakoCytomation) antibodies were used.For oligo competitions, cold oligos (wildtype and mutant) wereincluded at approximately 80 molar excess.

2.11. Statistical analysis

Experiments were performed in triplicate and are represented asthe meanstandard error of the mean (SEM). For group comparisonsthe Student's t-test was used, where two-tailed paired tests were usedand a p-value of less than 0.05 considered statistically significant.

3. Results

3.1. Differential Crm1 promoter activity in normal, transformed andcancer cells derives from the 175 to +99 region of the Crm1 promoter

Based on our previous findings showing elevated Crm1 expressionin cervical cancer and transformed cells compared to normal [10], weinvestigated Crm1 protein expression in a normal fibroblast cell line,WI38, an SV40-transformed fibroblast cell line, SVWI38, and a cervicalcancer cell line, CaSki. Western blot analysis revealed elevated Crm1expression in SVWI38 and CaSki cells, compared to normal WI38 cells(Fig. 1A).

To determine transcriptional regulatory mechanisms that associ-ate with elevated Crm1 expression in cancer and transformed cells, anapproximate 1.4 Kb region upstream of the transcription start site ofthe Crm1 gene, together with approximately 100 bp of the 5untranslated region (downstream from the transcription start site),was cloned into the pGL3-Basic vector for promoter analysis. The1405 to+99 Crm1 construct showed significantly higher activity inthe transformed SVWI38 and CaSki cancer cells compared to thenormal WI38 cells (Fig. 1B, pb0.05)).

In order to investigate the region responsible for the differentialexpression of Crm1 in normal, transformed and cancer cells, a series ofdeletion constructs of the Crm1 promoter were generated and assayedfor activity in WI38, SVWI38 and CaSki cells. Results showed that allconstructs displayed significantly more activity in SVWI38 and CaSkicells compared toWI38 cells (Fig. 1C). Sequential deletion of the1405to 175 region of the Crm1 promoter did not significantly alterpromoter activity, however, deletion of the 175 to 82 regionsignificantly diminished Crm1 promoter activity in SVWI38 and CaSkicells (by approximately five-fold) (Fig. 1C, pb0.05). In WI38 cells,however, deletion of this region had little effect on promoter activity.This suggests that important functional elements, necessary for highCrm1 promoter activity in transformed and cancer cells, are likely toreside in the 175 to 82 region of the Crm1 promoter. Elements inthe 82 to +99 region are also likely important for high Crm1expression in transformed and cancer cells, as this construct itself wassignificantlymore active in SVWI38 and CaSki cells compared to normalWI38 cells (pb0.05, Fig. 1C).

3.2. Identification and characterisation of transcription factors bindingwithin the 175 to +99 Crm1 promoter region

As the 175 to +99 region of the Crm1 promoter was identifiedas important for Crm1 promoter activity in transformed and cancercells, a bioinformatic analysis of this region was performed usingMatInspector [16] and Conreal (conserved regulatory elementsanchored alignment algorithm) [17] software programmes, in orderto identify putative transcription factor binding sites. The 175 to+99 region of the promoter was found to contain a number ofconsensus binding motifs, of which the CCAAT boxes (containing theflanking nucleotides corresponding to the binding site for NuclearFactor Y, NFY [18]), Sp1 sites, YinYang site, and NF B site wereinvestigated, since these proteins have been previously linked withproliferation and cancer [1922]. These binding sites were highlyconserved across species (Fig. 2A).

Firstly, the functionality of the NFB site was investigated. Crm1promoter activity (175 to +99) was analysed in a cancer cell linestably transfected with short hairpin-forming RNA (shRNA) designedto inhibit the NFB subunit p65, and compared to that in a control cellline. No change in Crm1 promoter activity was observed, nor wasthere any change in Crm1 protein expression (Supplementary Fig. 1).These results suggest that NFB does not have any significant role inthe regulation of Crm1 expression.

To determine the functionality of the putative NFY (also knownas CCAAT-binding protein, CBP), YinYang and Sp1 bindingelements, these sites were mutated in the 175 to +99 promoterconstruct and the mutated constructs transfected into CaSki cells.Results showed that mutation of both CCAAT boxes, at 150 to142 and at107 to99, and the GC box/Sp1 binding site, at +24to +36, significantly reduced Crm1 promoter activity (Fig. 2B,pb0.05). No change in promoter activity was observed when theputative Sp1 sites at 101 to 92 and 86 to 77 and theputative YinYang site at 133 to 126 were mutated, suggestingthat these sites do not contribute to Crm1 expression. Similarobservations were made using the SVWI38 cell line (SupplementaryFig. 2).

The effect of the CCAAT box and GC box mutations on Crm1promoter activity was next compared in normal WI38, transformedSVWI38 and CaSki cancer cells. Interestingly, mutation of bothCCAAT boxes and the GC box had a more profound impact on Crm1promoter activity in the transformed and cancer cells, compared tothe normal cells, suggesting that while these sites are functional innormal cells, activity of the respective transcription factors that bindthe sites appears to be increased in transformed and cancer cells(Fig. 2C).

Crm1

-tubulin

A B

C-1405 +99

-500

-394

-175

+99

+99

+99

-82

+1

+99

0 5 10

Crm1 Promoter Activity

15

WI38

SVWI38

CaSki

luc

luc

luc

luc

luc

0

2

4

6

8

10

12

14

WI38 SVWI38 CaSki

Crm

1 (-

1405

/+99

) p

rom

ote

r ac

tivi

ty

MI3

8

SV

WI3

8

Cas

ki

Fig. 1. Differential Crm1 expression in normal, transformed and cancer cells derives from the 175 to +99 region of the Crm1 promoter. A. Western blot analysis of Crm1 proteinlevels in normal WI38, transformed SVWI38, and cervical cancer CaSki cells reveals increased expression in the cancer and transformed cells. B. 1405 to +99 Crm1 promoteractivity was measured in cell lysates prepared from transfected WI38, SVWI38 and CaSki cells, and was significantly higher in the transformed and cancer cells compared to thenormal cells (*pb0.05). TK-Renilla-Luc was used as a control for transfection efficiency and luciferase activity is expressed relative to Renilla luciferase. C. Luciferase reporterconstructs containing deleted fragments of the human Crm1 promoter were transiently transfected intoWI38, SVWI38, and CaSki cells. Luciferase activity was significantly higher inthe SVWI38 cells and CaSki cells compared to normalWI38 cells for all constructs tested. Deletion of the175 to82 region of the Crm1 promoter resulted in a significant decreasein promoter activity in SVWI38 and CaSki cells (*pb0.05), while having only a minor effect in normal cells. Results shown are the meanSD of experiments performed in triplicateand repeated at least two times.

319P.J. van der Watt, V.D. Leaner / Biochimica et Biophysica Acta 1809 (2011) 316326

3.3. The NFY transcriptional complex binds the Crm1 promoter CCAATbox(es), while Sp1 binds the GC box

Since the CCAAT and GC boxes were identified as functionalelements within the Crm1 promoter, it was next investigated whetherthe transcription factors that generally associate with these sites,namely NFY/CBP [23] and Sp1 [24], played a role in Crm1 promoterregulation. siRNAs were employed to silence NFY and Sp1 expressionand effects on Crm1 promoter activity measured. NFY is a hetero-trimeric complex that comprises NFYA, B and C subunits [23], andsince all subunits are required for NFY function, an NFYA siRNA waschosen to inhibit NFY. Inhibition of both NFYA and Sp1 resulted in asignificant decrease in Crm1 promoter activity, confirming a role forthese proteins in the regulation of Crm1 expression (Fig. 3A, pb0.01).

In vivo binding of NFY and Sp1 to their respective sites was nextinvestigated. Chromatin Immunoprecipitation (ChIP) assays were per-formed using chromatin prepared from WI38, SVWI38 and CaSki cells,and after immunoprecipitation with an -NFYA or -Sp1 antibody, PCRamplification was performed using primers designed to span either bothCCAAT boxes (in the case of NFY) or the GC box alone (in the case of Sp1).A positive bandof the correct sizewas detected in both SVWI38 andCaSkicell lines (Fig. 3B, C), indicating binding of NFY and Sp1 to the Crm1promoter in these cell lines. Binding was, however, undetectable in the

normalfibroblast cell line,WI38. It is likely thatNFYandSp1dobind to theCrm1 promoter in normal WI38 cells, however, to a lesser extent than intransformed and cancer cells, and hence binding is virtually undetectable.In agreementwith this, endogenous NFY and Sp1 protein expressionwasmeasured inWI38, SVWI38 and CaSki cell lysates and increased NFY andSp1 expression detected in the transformed and cancer cells compared tonormal (Fig. 3D, E). In order to confirm that the increased expression ofNFYandSp1proteins in SVWI38andCaSki cellswasnot a characteristic ofall transcription factors, expressionof an independent transcription factor,NF B p65, which we had previously excluded from being a regulator ofCrm1 promoter activity, was analysed and found to be unchanged inWI38, SVWI38 and CaSki cell lines (data not shown).

Taken together, our data provides evidence for the activation of Crm1expression in cancer and transformed cells by NFY (binding in thepromoter regions150 to142 and107 to99) and Sp1 (binding inthe promoter region +24 to +36), and that elevated Crm1 expressionassociates with an increase in NFY and Sp1 expression in cancer cells.

3.4. p53 is a negative regulator of Crm1 expression

Interestingly, a common feature of promoters containing morethan one CCAAT box is the conservation of the distance between theCCAAT boxes; that is, 32 or 33 bp. While this spacing of CCAAT boxes

CaSki

luc-175 +99

lucXlucXlucXlucXlucXlucX

CCAAT box 1

Yin-Yang

CCAAT box 2

Sp1 site 1

Sp1 site 2

GC box

+1

Relative luciferase activity

B

0

0.2

0.4

0.6

0.8

1

1.2

CaSki SVWI38 WI38

Rel

ativ

e lu

cife

rase

act

ivity

Wildtype

mutated CCAAT box 1

mutated CCAAT box 2

mutated GC box

C

CCAAT box -175 Yin-Yang CCAAT box Sp1-82

NFB Sp1/GC box

+99

+1

ASp1

0 0.2 0.4 0.6 0.8 1 1.2

Fig. 2. Identification of functional CCAAT and GC boxes within the 175 to +99 region. A. Multalin sequence alignment showing homology between the human, mouse, cow, dog,and chicken Crm1 promoters in the 175 to +99 region. Putative transcription factor binding sites were identified using MatInspector and Conreal software programmes and areindicated in boxes. Bases that are conserved across species are indicated in red. B. Site-directed mutagenesis was performed using mutagenic primers, with the Crm1 175 to +99promoter construct as a template. Mutagenesis was performed to eliminate the potential CCAAT box 1, YinYang site, CCAAT box 2, Sp1 site 1, Sp1 site 2 and GC box (in order of theirappearance along the Crm1 promoter). Wildtype and mutated constructs were transfected into CaSki cells and luciferase activity determined. Significant changes in luciferaseactivity were observed when the CCAAT boxes and GC box were mutated (*pb0.05). C. Promoter activity of wildtype and mutated CCAAT and GC box constructs was analysed andcompared in CaSki, SVWI38 and WI38 cells. Results shown are the meanSD of experiments performed in quadruplicate.

320 P.J. van der Watt, V.D. Leaner / Biochimica et Biophysica Acta 1809 (2011) 316326

C

WI3

8

SV

WI3

8

CaS

ki

Sp1

-tubulin

E

WI3

8

SV

WI3

8

CaS

ki

D

NFY A

-tubulin

NFYAAb

NoAb

InputDNA

Watercontrol

WI38 SVWI38 CaSkiB

NFYAAb

NoAb

InputDNA

Watercontrol

NFYAAb

NoAb

InputDNA

Watercontrol

Sp1Ab

NoAb

InputDNA

Watercontrol

Sp1Ab

NoAb

InputDNA

Watercontrol

Sp1Ab

NoAb

InputDNA

Watercontrol

WI38 SVWI38 CaSki

0.0

5.0

10.0

15.0

20.0

25.0

30.0

ctl siRNA NFY siRNA Sp1 siRNA

Crm

1 (-

175

to +

99)

pro

mo

ter

acti

vity

A

Fig. 3. NFY and Sp1 are necessary for Crm1 promoter activity and show increased binding to the Crm1 promoter in transformed and cancer cells. A. CaSki cells were transfected with30 nM control, NFYA or Sp1 siRNA and Crm1 promoter activity measured. siRNA knock-down of NFYA and Sp1 transcription factors resulted in a significant decrease in Crm1promoter activity in CaSki cells (*pb0.01). B, C. ChIP analysis of NFYA (B) and Sp1 (C) binding to the CCAAT boxes and GC box in the Crm1 promoter, respectively. Cells were treatedwith formaldehyde to cross-link endogenous proteins and DNA and immunoprecipitations performed. ChIP PCR products were electrophoresed on a 2% agarose gel and positiveamplification of a 264 bp (B) or 200 bp (C) band, for NFY and SP1 respectively, was detected in SVWI38 and CaSki cell lines after immunoprecipitation with an-NFYA (B) or Sp1 (C)antibody, revealing binding of NFYA to the CCAAT boxes and Sp1 to the GC box in the Crm1 promoter. No binding was detected in normal WI38 cells. For input DNA, PCR wasperformed using sonicated chromatin DNA before immunoprecipitation. The negative control lane carries DNA immunoprecipitated without antibody, while the water blank servesas the negative control for the PCR. D, E. Protein was harvested from WI38, SVWI38 and CaSki cells and NFYA and Sp1 protein expression analysed by Western blot analysis. BothNFYA and Sp1 protein expression was elevated in SVWI38 and CaSki cells compared to normal WI38 cells. -tubulin was used as a control for protein loading.

321P.J. van der Watt, V.D. Leaner / Biochimica et Biophysica Acta 1809 (2011) 316326

is important for p300-mediated co-activation, it has also beenreported to facilitate promoter repression via p53 [25]. The distancebetween the CCAAT boxes in the Crm1 promoter was 33 bp, leadingus to investigate a potential role for p53 in regulating the Crm1promoter. Moreover, p53 is known to repress a variety of G2/M-associated genes, and Crm1 mRNA is known to accumulate in thisstage of the cell cycle [26].

To investigate the involvement of p53 in the regulation of Crm1promoter activity, CaSki cervical cancer cells, which contain wild-type p53 maintained at low levels due to enhanced degradationmediated by the HPV16 oncoprotein E6 [27], were treated with

doxorubicin (adriamycin): a potent DNA-damaging agent andknown inducer of p53 activity. In accordance with literature, p53was found to accumulate soon after doxorubicin treatment in aconcentration-dependent manner (Fig. 4A). Crm1 protein levels, onthe other hand, were significantly reduced after p53 induction withdoxorubicin (Fig. 4B). To determine whether the reduction in Crm1protein levels was due to altered Crm1 promoter regulation, CaSkicells were transfected with the 175 to +99 Crm1 promoterconstruct and exposed to doxorubicin for 18 h. As shown in Fig. 4C,doxorubicin treatment reduced Crm1 promoter activity in a dose-dependent manner, with 2 g/ml doxorubicin down-regulating

0 0.2 1 g/ml doxorubicin 0 0.2 1 g/ml doxorubicin

p53

-tubulin

Crm1

-tubulin

C

A B

0.0

0.2

0.4

0.6

0.8

1.0

1.2

CaSki SVWI38 WI38

Crm

1 p

rom

ote

r ac

tivi

ty

0 g/ml doxorubicin

0.2 g/ml doxorubicin

2 g/ml doxorubicin

2 2

Fig. 4. Doxorubicin treatment results in an induction of p53 and concomitant repression of Crm1 promoter activity. A. CaSki cells were treated with 0.2, 1 or 2 g/ml doxorubicin andprotein harvested for Western blot analysis 8 h (A) or 48 h (B) after treatment. p53 protein levels were increased (A) while Crm1 protein levels decreased (B) after doxorubicintreatment. C. Crm1 promoter activity was measured after an 18 h treatment with doxorubicin. Significant decreases in Crm1 promoter activity were observed in both CaSki andSVWI38 cell lines after treatment (*pb0.05), while in WI38 cells Crm1 promoter activity was only decreased marginally.

322 P.J. van der Watt, V.D. Leaner / Biochimica et Biophysica Acta 1809 (2011) 316326

Crm1 promoter activity by 90% in CaSki cells. Crm1 promoteractivity in SVWI38 cells was similarly repressed upon doxorubicintreatment, while Crm1 promoter activity in the normal WI38 cellline was not responsive (Fig. 4C). This is likely explained by thealready low Crm1 promoter activity in untreated WI38 cells due tolow levels of NFY and Sp1 in this cell line (Fig. 3D and E).

Given that the Crm1 promoter was repressed in response todoxorubicin treatment and the concomitant induction of p53, wenext determined whether the repression was specifically via p53.Firstly, p53 was inhibited in CaSki cells using p53 siRNA and Crm1promoter activity measured. No change in Crm1 promoter activitywas observed after p53 inhibition, suggesting that p53 does notrepress the Crm1 promoter in the absence of repair stress (Fig. 5A).However, to determine whether p53 represses the Crm1 promoterin response to DNA damage, hence accounting for the inhibition ofCrm1 promoter activity seen in Fig. 3, CaSki cells were transfectedwith control or p53 siRNA and then treated with doxorubicin 24hours later. In cells transfected with control siRNA, Crm1 promoteractivity was significantly repressed on treatment with doxorubicin,while in cells transfected with the p53 siRNA first, the decrease inCrm1 promoter activity was less marked with promoter activity ofdoxorubicin-treated cells not significantly different from that ofcells grown in the absence of doxorubicin (Fig. 5B). Western blotanalysis revealed that p53 induction was blocked in cells transfectedwith p53 siRNA, followed by doxorubicin treatment (Fig. 5C).Moreover, inhibiting p53 induction in doxorubicin-treated cellsresulted in Crm1 protein levels maintained at high levels (Fig. 5Dand E). These findings suggest that p53 plays a role in negativelyregulating Crm1 expression in response to DNA damage, asinhibiting p53 expression in doxorubicin-treated cells results inelevated Crm1 levels.

3.5. p53 regulation of the Crm1 promoter occurs via the CCAAT boxes

Having determined that p53 associates with repression of the Crm1promoter activity in response to doxorubicin treatment, we sought tounderstand the mechanism of repression and identify the p53-responsive region within the Crm1 promoter. It was anticipated thatthe response would likely occur via the two critically-spaced CCAATboxes, as no consensus p53 binding sites were identified in the Crm1promoter. ChIP analysis was therefore performed to investigate p53binding to the Crm1 promoter in the region of the two CCAAT boxes.Untreated and doxorubicin-treated cells were fixed with formaldehydeto cross-link protein-DNA complexes and p53-bound DNA complexeswere captured with an -p53 antibody. No binding of p53 to the Crm1promoter (in the region175 to+89) was detected in untreated cells,while p53 binding was strongly detected after treatment withdoxorubicin, in the region of the two CCAAT boxes (Fig. 6A). It wasnext investigated whether p53 binding affected the binding of NFY tothe CCAAT boxes. ChIP analysis was therefore performed to analyse NFYbinding to the Crm1 promoter in response to doxorubicin treatment.Interestingly, the binding of NFY to the Crm1promoterwas significantlyreduced on treatment with doxorubicin, suggesting that the binding ofp53 may interfere with NFY binding (Fig. 6A).

To confirm the finding that p53-mediated repression of the Crm1promoter occurs via interference with NFY, a Crm1 (175 to +99)promoter construct containing mutations in both CCAAT boxes wasgenerated and exposed to doxorubicin treatment. This constructcontained significantly less promoter activity than the wildtypeconstruct, and then the constructs containing mutations in eitherCCAAT box 1 or CCAAT box 2 (Fig. 6B). When exposed to doxorubicin, itwas determined that while the wildtype 175 to +99 construct wasrepressed approximately 15-fold, the Crm1 promoter construct with

D

Crm1

-tubulin-tubulin

doxorubicin

CControl siRNA p53 siRNA

p53

doxorubicinControl siRNA p53 siRNA

0

0.2

0.4

0.6

0.8

1

1.2

1.4

Rel

ativ

e C

rm1

(-17

5 to

+99

) p

rom

ote

r ac

tivi

ty

ctl siRNA

p53 siRNA

- + doxorubicin

E

0

0.2

0.4

0.6

0.8

1

1.2

1.4

Den

sito

met

ric

qu

anti

fica

tio

n o

fC

rm1

rela

tive

to

-t

ub

ulin

doxorubicin

Control siRNA p53 siRNA

B

0

5

10

15

20

25

30

35

ctl siRNA p53 siRNA

Crm

1 (-

175/

+99)

pro

mo

ter

acti

vity

A

Fig. 5. p53 associates with promoter repression in response to doxorubicin treatment. A. Crm1 (175 to+99) promoter activity was measured after p53 inhibition using 30 nM p53siRNA. No change in Crm1 promoter activity was observed. B. Cells were transfected with 30 nM control or p53 siRNA 24 h prior to 0.2 g/ml doxorubicin treatment, and Crm1(175 to +99) promoter activity was measured. Crm1 promoter activity in control siRNA-transfected cells was significantly reduced on treatment with doxorubicin (*pb0.05),however, a less marked effect was observed in p53 siRNA-transfected cells. C. Control and p53 siRNA-transfected cells were treated with doxorubicin prior to protein harvest andWestern blot analysis. Western blot analysis revealed an accumulation of p53 in the control siRNA-transfected cells, but not the p53-siRNA-transfected cells, after doxorubicintreatment. D. Conversely, Crm1 protein levels were reduced in the control siRNA-transfected cells, and to a much lesser extent in the p53 siRNA-transfected cells, after doxorubicintreatment. -tubulin was used as a control for protein loading. E. Densitometric quantification of the Western blots in C and D, showing Crm1 protein levels relative to -tubulin.

323P.J. van der Watt, V.D. Leaner / Biochimica et Biophysica Acta 1809 (2011) 316326

mutations in both CCAAT boxes was substantially less responsive(Fig. 6C). This suggests that functional CCAAT boxes are necessary forp53-mediated repression of the Crm1 promoter, and suggests that p53cannot repress Crm1 in the absence of NFY. To confirm this, NFY wasdepleted in CaSki cells using siRNA, and cells were subjected todoxorubicin treatment. As expected, the repression of Crm1 promoteractivitywas significantly less inNFYknock-down cells compared to cellstransfected with control siRNA (Fig. 6D). The Crm1 promoter was,however, still somewhat repressed in the NFY knock-down cells likelydue incomplete NFY protein knock-down.

Finally, to analyse the binding of NFY and p53 to the Crm1 promoterafter DNA damage, EMSAs were performed using nuclear proteinharvested from doxorubicin-treated cells and labeled probes corre-sponding to the individual CCAAT boxes. Results show that p53 binds

both CCAAT boxes in a single complex with NFY (Fig. 6E). This wasrevealedby the addition of antibodies againstNFYAandp53,whichbothresulted in diminished binding of the marked complex (Fig. 6E). It istherefore apparent that p53 does not competewith NFY for binding butrather binds NFY present on the CCAAT boxes of the Crm1 promoter,likely resulting in inhibition of NFY transcriptional activation. Additionof appropriate cold oligonucleotides (wildtype and mutant) confirmedthe specificity of the binding, with the wildtype oligonucleotidecompeting for binding and the mutant oligonucleotide not (Fig. 6F).

4. Discussion

This study is the first to our knowledge that describes upstreamcis-elements and transcription factors that are important for Crm1

p53 Ab No Ab Input DNA Water control- + - + - +dox

A

0

2

4

6

8

10

12

14

16

18

20

WT CCAAT1+2mut

Fo

ld r

epre

ssio

n

0 g/ml doxorubicin

0.2 g/ml doxorubicin

2 g/ml doxorubicin

C

B

D

0

2

4

6

8

10

12

14

ctl siRNA NFY siRNAF

old

rep

ress

ion

0

0.2

0.4

0.6

0.8

1

1.2

WT

CCAA

T1m

ut

CCAA

T2m

ut

CCAA

T1+2

mut

Rel

ativ

e C

rm1

(-17

5/+9

9)p

rom

ote

r ac

tivi

ty

free probe

NFYA Abp53 Ab

Cold WT oligoCold mutant oligo

NFY/p53

+ -- +

+ -- +

--

+ -- +

+ -- +

--

CCAAT box1probe

CCAAT box2probe

E

--

--

CCAAT box1 probe

CCAAT box2 probe

F

NFYA Ab No Ab Input DNA Water control- + - + - +dox

Fig. 6. p53 mediates repression of the Crm1 promoter via interference with NFY. A. ChIP analysis showing p53 and NFY binding to the Crm1 promoter after treatment of CaSki cellswith 0.2 g/ml doxorubicin. Cells were treated, followed by cross-linking of chromatin and pull-down with an -p53 or -NFYA antibody. PCR amplification of pulled-downchromatin revealed a strong recruitment of p53 to the Crm1 promoter in response to doxorubicin treatment, while NFY binding appeared diminished. B. Luciferase assay showingpromoter activity of wildtype and mutated Crm1 (175 to +99) constructs (*pb0.05). C. Promoter activity of wildtype or mutated Crm1 (175 to +99) constructs was measuredin CaSki cells following doxorubicin treatment. While a significant repression of wildtype Crm1 promoter activity was observed in response to doxorubicin (*pb0.05), mutation ofboth CCAAT boxes strongly reduced this effect. Results shown are the meanSD of experiments performed in triplicate and repeated at least two independent times. D. CaSki cellswere transfected with 30 nM control or NFY siRNA and treated with doxorubicin. The repression of Crm1 promoter activity in response to doxorubicin was significantly less after NFYinhibition, compared to control cells (*pb0.05). E, F. EMSA showing protein binding to labelled Crm1 promoter oligonucleotides containing the individual CCAAT boxes, where(E) both NFYA and p53 antibodies resulted in diminished protein binding, and (F) cold wildtype and mutant oligonucleotides confirmed the specificity of the binding.

324 P.J. van der Watt, V.D. Leaner / Biochimica et Biophysica Acta 1809 (2011) 316326

expression and that drive its high expression in cancer andtransformed cells. Using deletion and mutation analysis we identifiedthe 175 to +99 region of the Crm1 promoter as responsible for thedifferential expression of Crm1 in normal and cancer cells, andidentified within this region two CCAAT boxes and a GC box asimportant regulatory elements. Moreover, we identified binding of

NFY/CBP and Sp1 transcription factors to these respective sites intransformed and cancer cells. We propose that the elevated levels ofNFY and Sp1 proteins in transformed and cancer cells may be in partresponsible for overexpression of Crm1 in these cells. Furthermore,our study is a first to describe a role for p53 in repressing the Crm1promoter in response to DNA damage; a finding which could also

325P.J. van der Watt, V.D. Leaner / Biochimica et Biophysica Acta 1809 (2011) 316326

contribute to the high levels of Crm1 in cancer cells, since most cancercells either contain low levels of p53 protein (through enhanced p53degradation) or harbour p53 mutations [28].

There is increasing evidence implicating NFY as a key player incancer development. NFY binds the CCAAT box which is present inabout 30% of eukaryotic promoters [23]. It is reported that promotersthat lack a TATA box rely greatly on the CCAAT box for transcriptionalactivation [18], andwe show that the Crm1 promoter lacks a TATA boxbut contains two functional CCAAT boxes. The presence of more thanone CCAAT box in the promoter of a gene is not uncommon, especiallyin the promoters of G2/M-associated genes; genes that are reported torely greatly on NFY for activation [25,29,30]. Interestingly, Crm1mRNA is reported to reach a peak at G2/M [26] and Crm1 is known toplay a key role in the control of mitosis [31], likely contributing to itsreliance on NFY for transcriptional activation.

Literature reports that NFY plays an important role in proliferation[21] and the cell cycle [32], and its own activity is regulated in a cellcycle-dependent manner [33,34]. It is believed that NFY is a majorcontributor to the high expression of many cell cycle-regulated genesin cancer cells [35,36] and Niida et al. found that the NFY motif issignificantly enriched in genes overexpressed in breast cancer, with itbeing one of the principal regulatory motifs driving breast cancermalignant progression [37]. Furthermore, Pang et al. found that NFYbinding to the thymidine kinase promoter was at levels 5- and 15-foldgreater in SV-40 virus-transformed human cells and HeLa cells,compared to normal cells, respectively, and that the binding activity ofNFY changes from a cell cycle-dependent one in normal cells to aconstitutive one in transformed cells and tumour cells [38]. Theysuggest that this is due to a loss of post-translational control during/after cell transformation, as the half-life of NFY in transformed cellswas five-fold higher than in normal cells [38]. In this study we haveshown that NFY expression is elevated in cervical cancer andtransformed human fibroblast cells, resulting in increased binding ofNFY to the Crm1 promoter, and thus NFY is very likely a keycontributor to Crm1 overexpression in cancer.

As well as demonstrating a role for NFY, we identified Sp1 as animportant regulator of Crm1 expression in cancer. Interestingly, in anin silico study, Quan et al. predicted that the Sp1 binding site is one ofthe most common transcription factor binding sites in the promotersof karyopherin family genes, of which Crm1 is a member [39]. Sp1regulates a variety of biological functions that are critical to tumourdevelopment and progression [19,24], with its transcriptional activitymodulated in a cell cycle-dependent manner [40], and its target genesincluding many regulators of the cell cycle [4144]. It plays animportant role in growth regulation, as introduction of a dominantnegative Sp1 into HeLa cervical cancer cells significantly inhibited cellproliferation [45]. We identified increased Sp1 protein expression intransformed fibroblasts, SVWI38, and cervical cancer cells, CaSki,compared to normal fibroblasts, WI38. Sp1 expression has previouslybeen shown to be elevated in transformed cells [46], where itsincreased expression was found to result in increased activation of itstarget genes [46]. Moreover, elevated Sp1 expression has beenassociated with several cancers, including gastric cancer [47],pancreatic cancer [48], and skin cancer [49]. We show that inhibitionof Sp1 expression in cancer cells leads to diminished Crm1 promoteractivity, suggesting that it is, in part, the high expression of Sp1 incancer cells that maintains high Crm1 expression.

We also identified a role for the transcription factor p53 in theregulation of the Crm1 promoter. p53 can activate or repress thetranscription of target genes. While activation generally occursthrough binding of p53 to its consensus binding site in the promoterregion of target genes, for example, p21, GADD45 and 14-3-3, it canrepress promoters that lack a consensus p53 binding site. This canoccur via several mechanisms; either through its interferencewith thefunctions of other DNA-binding transcriptional activators, or throughits interference with components of the basal transcription machin-

ery, or through its recruiting chromatin-modifying factors such ashistone deacetylases (HDAC) to promoter regions, thereby reducingpromoter accessibility to the transcriptional machinery [50]. p53target genes repressed via these mechanisms include cyclin B1 [51],cdc2 [52], BRCA2 [53], survivin [54], hsp70 [55], amongst others. Inthe case of p53 interfering with other transcriptional activators, it isreported that p53 can interact specifically with NFY and Sp1 proteins,hampering the NFY- and Sp1-dependent recruitment of the generaltranscriptional machinery to the promoter region of target genes[25,51,52,54,55]. In this study we provide evidence that p53 repressesthe Crm1 promoter, and that repression occurs predominantly viainterference with NFY. Our results suggest that p53 can interfere withNFY-regulated Crm1 expression under conditions of DNA damage bydirectly interfering with NFY binding and in this way result inpromoter repression. Our results also suggest that p53 may complexwith NFY and in this way mediate promoter repression. There isevidence in the literature that under conditions of DNA damage, thebinding of p53 to certain promoters results in chromatin remodellingwhich associates with transcriptional repression [56]. We speculate inthe case of the Crm1 promoter that p53 possibly promotes chromatinremodelling; only when the promoter is in an active state (and boundby NFY) hence the dependence on NFY for repression, and that thisresults in a loss of NFY binding to the promoter as observed in ChIPanalysis. Further work, however, is required to investigate the exactmechanism of repression by p53.

While we have shown a role for p53 in negatively regulating theCrm1 promoter, it is also well understood that Crm1 plays a role in thenegative regulation of p53 activity by mediating its export from thenucleus to the cytoplasm [57]. Hence it appears that Crm1 and p53could share a complex regulatory loop, whereby p53 represses Crm1promoter activity in response to DNA damage, and the low levels ofCrm1 that result could possibly lead to a decreased rate of export ofp53 into the cytoplasm, resulting in increased p53 function. This couldbe one of the reasons why during tumourigenesis the cell mustabrogate p53 function and/or upregulate Crm1 transcription in orderto evade p53 and allow for uncontrolled proliferation.

Supplementarymaterials related to this article can be found onlineat doi:10.1016/j.bbagrm.2011.05.017.

Funding

Thisworkwas supported by grants from theUniversity of Cape Town,the Carnegie Corporation of NewYork, CANSA, and theMedical ResearchCouncil of South Africa.

References

[1] B. Ossareh-Nazari, F. Bachelerie, C. Dargemont, Evidence for a role of CRM1 insignal-mediated nuclear protein export, Science 278 (1997) 141.

[2] M. Fukuda, S. Asano, T. Nakamura, M. Adachi, M. Yoshida, M. Yanagida, E. Nishida,CRM1 is responsible for intracellular transport mediated by the nuclear exportsignal, Nature 390 (1997) 308.

[3] H.P. Bogerd, R.A. Fridell, R.E. Benson, J. Hua, B.R. Cullen, Protein sequencerequirements for function of the human T-cell leukemia virus type 1 Rex nuclearexport signal delineated by a novel in vivo randomization-selection assay, Mol.Cell Biol. 16 (1996) 4207.

[4] J.R. Davis, M. Kakar, C.S. Lim, Controlling protein compartmentalization toovercome disease, Pharm. Res. 24 (2007) 17.

[5] A. Komeili, E.K. O'Shea, Nuclear transport and transcription, Curr. Opin. Cell Biol.12 (2000) 355.

[6] S. Gaubatz, J.A. Lees, G.J. Lindeman, D.M. Livingston, E2F4 is exported from thenucleus in a CRM1-dependent manner, Mol. Cell Biol. 21 (2001) 1384.

[7] J. Yang, E.S. Bardes, J.D. Moore, J. Brennan, M.A. Powers, S. Kornbluth, Control ofcyclin B1 localization through regulated binding of the nuclear export factorCRM1, Genes Dev. 12 (1998) 2131.

[8] D.A. Dean, Nuclear transport: an emerging opportunity for drug targeting, Adv.Drug Deliv. Rev. 55 (2003) 699.

[9] N. Mosammaparast, L.F. Pemberton, Karyopherins: from nuclear-transportmediators to nuclear-function regulators, Trends Cell Biol. 14 (2004) 547.

[10] P.J. van derWatt, C.P. Maske, D.T. Hendricks, M.I. Parker, L. Denny, D. Govender, M.J. Birrer, V.D. Leaner, The karyopherin proteins, Crm1 and Karyopherin beta1, are

http://doi:10.1016/j.bbagrm.2011.05.017

326 P.J. van der Watt, V.D. Leaner / Biochimica et Biophysica Acta 1809 (2011) 316326

overexpressed in cervical cancer and are critical for cancer cell survival andproliferation, Int. J. Cancer 124 (2009) 1829.

[11] A. Noske, W. Weichert, S. Niesporek, A. Roske, A.C. Buckendahl, I. Koch, J. Sehouli,M. Dietel, C. Denkert, Expression of the nuclear export protein chromosomalregion maintenance/exportin 1/Xpo1 is a prognostic factor in human ovariancancer, Cancer 112 (2008) 1733.

[12] Y. Yao, Y. Dong, F. Lin, H. Zhao, Z. Shen, P. Chen, Y.J. Sun, L.N. Tang, S.E. Zheng, Theexpression of CRM1 is associated with prognosis in human osteosarcoma, Oncol.Rep. 21 (2009) 229.

[13] A. Shen, Y. Wang, Y. Zhao, L. Zou, L. Sun, C. Cheng, Expression of CRM1 in humangliomas and its significance in p27 expression and clinical prognosis, Neurosur-gery 65 (2009) 153.

[14] W.Y. Huang, L. Yue, W.S. Qiu, L.W. Wang, X.H. Zhou, Y.J. Sun, Prognostic value ofCRM1 in pancreas cancer, Clin. Invest Med 32 (2009) E315.

[15] S.C. Mutka, W.Q. Yang, S.D. Dong, S.L. Ward, D.A. Craig, P.B. Timmermans, S. Murli,Identification of nuclear export inhibitors with potent anticancer activity in vivo,Cancer Res. 69 (2009) 510.

[16] K. Quandt, K. Frech, H. Karas, E. Wingender, T. Werner, MatInd and MatInspector:new fast and versatile tools for detection of consensus matches in nucleotidesequence data, Nucleic Acids Res. 23 (1995) 4878.

[17] E. Berezikov, V. Guryev, R.H. Plasterk, E. Cuppen, CONREAL: conserved regulatoryelements anchored alignment algorithm for identification of transcription factorbinding sites by phylogenetic footprinting, Genome Res. 14 (2004) 170.

[18] R. Mantovani, A survey of 178 NF-Y binding CCAAT boxes, Nucleic Acids Res. 26(1998) 1135.

[19] A.R. Black, J.D. Black, J. zizkhan-Clifford, Sp1 and kruppel-like factor family oftranscription factors in cell growth regulation and cancer, J. Cell Physiol 188(2001) 143.

[20] S. Gordon, G. Akopyan, H. Garban, B. Bonavida, Transcription factor YY1: structure,function, and therapeutic implications in cancer biology, Oncogene 25 (2006)1125.

[21] Q. Hu, S.N. Maity, Stable expression of a dominant negative mutant of CCAATbinding factor/NF-Y in mouse fibroblast cells resulting in retardation of cellgrowth and inhibition of transcription of various cellular genes, J. Biol. Chem. 275(2000) 4435.

[22] Q. Li, Y.Y. Yu, Z.G. Zhu, Y.B. Ji, Y. Zhang, B.Y. Liu, X.H. Chen, Y.Z. Lin, Effect of NF-kappaB constitutive activation on proliferation and apoptosis of gastric cancer celllines, Eur. Surg. Res. 37 (2005) 105.

[23] R. Mantovani, The molecular biology of the CCAAT-binding factor NF-Y, Gene 239(1999) 15.

[24] S. Safe, M. Abdelrahim, Sp transcription factor family and its role in cancer, Eur. J.Cancer 41 (2005) 2438.

[25] C. Imbriano, A. Gurtner, F. Cocchiarella, A.S. Di, V. Basile, M. Gostissa, M. Dobbelstein,S.G. Del, G. Piaggio, R. Mantovani, Direct p53 transcriptional repression: in vivoanalysis of CCAAT-containing G2/M promoters, Mol. Cell Biol. 25 (2005) 3737.

[26] N. Kudo, S. Khochbin, K. Nishi, K. Kitano, M. Yanagida, M. Yoshida, S. Horinouchi,Molecular cloning and cell cycle-dependent expression of mammalian CRM1, aprotein involved in nuclear export of proteins, J. Biol. Chem. 272 (1997) 29742.

[27] S. Hietanen, S. Lain, E. Krausz, C. Blattner, D.P. Lane, Activation of p53 in cervicalcarcinoma cells by small molecules, Proc. Natl. Acad. Sci. U. S. A 97 (2000) 8501.

[28] M. Oren, Decision making by p53: life, death and cancer, Cell Death Differ. 10(2003) 431.

[29] V. Salsi, G. Caretti, M. Wasner, W. Reinhard, U. Haugwitz, K. Engeland, R. Mantovani,Interactions between p300 and multiple NF-Y trimers govern cyclin B2 promoterfunction, J. Biol. Chem. 278 (2003) 6642.

[30] R. Elkon, C. Linhart, R. Sharan, R. Shamir, Y. Shiloh, Genome-wide in silicoidentification of transcriptional regulators controlling the cell cycle in humancells, Genome Res. 13 (2003) 773.

[31] A. Arnaoutov, Y. Azuma, K. Ribbeck, J. Joseph, Y. Boyarchuk, T. Karpova, J. McNally,M. Dasso, Crm1 is a mitotic effector of Ran-GTP in somatic cells, Nat. Cell Biol. 7(2005) 626.

[32] G. Caretti, V. Salsi, C. Vecchi, C. Imbriano, R. Mantovani, Dynamic recruitment ofNF-Y and histone acetyltransferases on cell-cycle promoters, J. Biol. Chem. 278(2003) 30435.

[33] F. Bolognese, M. Wasner, C.L. Dohna, A. Gurtner, A. Ronchi, H. Muller, I. Manni,J. Mossner, G. Piaggio, R. Mantovani, K. Engeland, The cyclin B2 promoterdepends on NF-Y, a trimer whose CCAAT-binding activity is cell-cycleregulated, Oncogene 18 (1999) 1845.

[34] I. Manni, G. Caretti, S. Artuso, A. Gurtner, V. Emiliozzi, A. Sacchi, R. Mantovani,G. Piaggio, Posttranslational regulation of NF-YA modulates NF-Y transcrip-tional activity, Mol. Biol. Cell 19 (2008) 5203.

[35] S.H. Park, G.R. Yu,W.H. Kim,W.S. Moon, J.H. Kim, D.G. Kim, NF-Y-dependent cyclinB2 expression in colorectal adenocarcinoma, Clin. Cancer Res. 13 (2007) 858.

[36] K.V. Sitwala, K. Adams, D.M. Markovitz, YY1 and NF-Y binding sites regulate thetranscriptional activity of the dek and dek-can promoter, Oncogene 21 (2002)8862.

[37] A. Niida, A.D. Smith, S. Imoto, S. Tsutsumi, H. Aburatani, M.Q. Zhang, T. Akiyama,Integrative bioinformatics analysis of transcriptional regulatory programs inbreast cancer cells, BMC. Bioinformatics 9 (2008) 404.

[38] J.H. Pang, L.F. Good, K.Y. Chen, The age-dependent binding of CBP/tk, a CCAATbinding protein, is deregulated in transformed and immortalized mammaliancells but absent in premature aging cells, Exp. Gerontol. 31 (1996) 97.

[39] Y. Quan, Z.L. Ji, X. Wang, A.M. Tartakoff, T. Tao, Evolutionary and transcriptionalanalysis of karyopherin beta superfamily proteins, Mol. Cell Proteomics. 7 (2008)1254.

[40] B.P. Fojas, N.K. De, P.Du Collins, J. zizkhan-Clifford, M. Mudryj, Cyclin A-CDKphosphorylates Sp1 and enhances Sp1-mediated transcription, EMBO J. 20 (2001)5737.

[41] E.J. Cram, B.D. Liu, L.F. Bjeldanes, G.L. Firestone, Indole-3-carbinol inhibits CDK6expression in human MCF-7 breast cancer cells by disrupting Sp1 transcriptionfactor interactions with a composite element in the CDK6 gene promoter, J. Biol.Chem. 276 (2001) 22332.

[42] I. Eto, Molecular cloning and sequence analysis of the promoter region of mousecyclin D1 gene: implication in phorbol ester-induced tumour promotion, CellProlif. 33 (2000) 167.

[43] A. Martino, J.H. Holmes, J.D. Lord, J.J. Moon, B.H. Nelson, Stat5 and Sp1 regulatetranscription of the cyclin D2 gene in response to IL-2, J. Immunol. 166 (2001)1723.

[44] M. Paskind, C. Johnston, P.M. Epstein, J. Timm, D. Wickramasinghe, E. Belanger,L. Rodman, D. Magada, J. Voss, Structure and promoter activity of the mouseCDC25A gene, Mamm. Genome 11 (2000) 1063.

[45] F. Chen, F. Zhang, J. Rao, G.P. Studzinski, Ectopic expression of truncated Sp1transcription factor prolongs the S phase and reduces the growth rate, AnticancerRes. 20 (2000) 661.

[46] J.D. Saffer, S.P. Jackson, S.J. Thurston, SV40 stimulates expression of the transactingfactor Sp1 at the mRNA level, Genes Dev. 4 (1990) 659.

[47] J.C. Yao, L. Wang, D. Wei, W. Gong, M. Hassan, T.T. Wu, P. Mansfield, J. Ajani, K. Xie,Association between expression of transcription factor Sp1 and increased vascularendothelial growth factor expression, advanced stage, and poor survival inpatients with resected gastric cancer, Clin. Cancer Res. 10 (2004) 4109.

[48] N.Y. Jiang, B.A. Woda, B.F. Banner, G.F. Whalen, K.A. Dresser, D. Lu, Sp1, a newbiomarker that identifies a subset of aggressive pancreatic ductal adenocarcino-ma, Cancer Epidemiol. Biomarkers Prev. 17 (2008) 1648.

[49] A.P. Kumar, A.P. Butler, Enhanced Sp1 DNA-binding activity in murinekeratinocyte cell lines and epidermal tumors, Cancer Lett. 137 (1999) 159.

[50] J. Ho, S. Benchimol, Transcriptional repression mediated by the p53 tumoursuppressor, Cell Death Differ. 10 (2003) 404.

[51] S.A. Innocente, J.M. Lee, p53 is a NF-Y- and p21-independent, Sp1-dependentrepressor of cyclin B1 transcription, FEBS Lett. 579 (2005) 1001.

[52] J. Yun, H.D. Chae, H.E. Choy, J. Chung, H.S. Yoo, M.H. Han, D.Y. Shin, p53 negativelyregulates cdc2 transcription via the CCAAT-binding NF-Y transcription factor, J.Biol. Chem. 274 (1999) 29677.

[53] K. Wu, S.W. Jiang, F.J. Couch, p53 mediates repression of the BRCA2 promoter anddown-regulation of BRCA2 mRNA and protein levels in response to DNA damage,J. Biol. Chem. 278 (2003) 15652.

[54] P.O. Esteve, H.G. Chin, S. Pradhan, Molecular mechanisms of transactivation anddoxorubicin-mediated repression of survivin gene in cancer cells, J. Biol. Chem.282 (2007) 2615.

[55] S.N. Agoff, J. Hou, D.I. Linzer, B. Wu, Regulation of the human hsp70 promoter byp53, Science 259 (1993) 84.

[56] T. Banerjee, S. Nath, S. Roychoudhury, DNA damage induced p53 downregulatesCdc20 by direct binding to its promoter causing chromatin remodeling, NucleicAcids Res. 37 (2009) 2688.

[57] J.M. Stommel, N.D. Marchenko, G.S. Jimenez, U.M. Moll, T.J. Hope, G.M. Wahl, Aleucine-rich nuclear export signal in the p53 tetramerization domain: regulation ofsubcellular localization and p53 activity by NES masking, EMBO J. 18 (1999) 1660.

The nuclear exporter, Crm1, is regulated by NFY and Sp1 in cancer cells and repressed by p53 in response to DNA damage1. Introduction2. Materials and methods2.1. Cell culture2.2. Protein harvest and Western blot analysis2.3. Construction of the Crm1 (1405 to +99) promoter construct2.4. Generation of Crm1 promoter deletion constructs2.5. Luciferase assays2.6. Site-directed mutagenesis of potential transcription factor binding sites2.7. siRNA transfection2.8. Chromatin immunoprecipitation (ChIP) assay2.9. Doxorubicin treatment2.10. EMSAs2.11. Statistical analysis

3. Results3.1. Differential Crm1 promoter activity in normal, transformed andcancer cells derives from the 175 to +99 region of the Crm1 promoter3.2. Identification and characterisation of transcription factors binding within the 175 to +99 Crm1 promoter region3.3. The NFY transcriptional complex binds the Crm1 promoter CCAAT box(es), while Sp1 binds the GC box3.4. p53 is a negative regulator of Crm1 expression3.5. p53 regulation of the Crm1 promoter occurs via the CCAAT boxes

4. DiscussionFundingReferences