Regulation of cell growth-dependent expression of mammalian CDC6 gene by the cell cycle transcription factor E2F

Regulation of cell growth-dependent expression of mammalian CDC6 geneby the cell cycle transcription factor E2F

Kiyoshi Ohtani1, Atsumi Tsujimoto1, Masa-aki Ikeda2 and Masataka Nakamura1

1Human Gene Sciences Center, 2Department of Developmental Biology, Graduate School of Dentistry, Tokyo Medical and DentalUniversity, Tokyo, 113-8510, Japan

CDC6 of Saccharomyces cerevisiae regulates the DNAreplication initiation through the origin recognitioncomplex (ORC). Identi®cation of a human homolog ofthe CDC6 gene (HsCdc6) suggests a universal role of thegene product in DNA replication. Expression of HsCdc6is growth-regulated. We investigated the molecular basisof growth-regulated expression of mammalian Cdc6. Thepromoter activity of isolated HsCdc6 upstream regionwas activated at late G1 and G1/S boundary in the cellcycle of rat embryonic ®broblast REF52 cells by theaddition of serum. The isolated promoter was activatedby exogenous expression of E2F without serum stimula-tion. However a mutant promoter lacking the E2Frecognition sites failed to respond to serum stimulationand exogenous expression of E2F. Expression ofendogenous Cdc6 was induced by exogenous expressionof E2F. Therefore, we concluded that the growth-regulated expression of mammalian Cdc6 was mediatedby E2F. Moreover, we demonstrated that exogenousoverexpression of either HsCdc6 or HsOrc1 failed toinduce DNA synthesis unlike overexpression of E2F1,even though E2F1 induced both Cdc6 and Orc1,suggesting that E2F may regulate the expression ofanother gene(s), besides Cdc6 and Orc1, required forinduction of cellular DNA synthesis in mammalian cells.

Keywords: mammalian CDC6; E2F; promoter; cellcycle; DNA synthesis

Introduction

DNA replication in mammalian cells is a complexprocess and highly ordered during the progression ofthe cell cycle. The replication process requires a seriesof components which are prepared at G1 phase. Thesecomponents may be induced transcriptionally inexpression, activated and/or repressed in function in astrictly time-dependent manner. Some of theseassemble at G1 while others dissociate to proceed inother stages of the cell cycle. The initiation of DNAreplication is a crucial step and is regulated by subtlemechanisms, although such mechanisms remain to beelucidated in mammalian cells.Most of our understanding of the molecular

mechanisms responsible for the initiation of DNAreplication is based on genetic and biochemical studies

of the yeast system. A set of molecules that control theinitiation of DNA replication has been identi®ed.These include six proteins consisting a multisubunitcomplex known as the origin recognition complex(ORC) (Bell and Stillman, 1992; Di�ey and Cocker,1992), which binds to autonomously replicating originsof DNA. The binding between ORC and thereplication origin is observed throughout the cell cycle(Aparicio et al., 1997). Thus, regulation of initiation ofDNA replication is thought to be determined bymolecules that functionally or physically interact withORC. Several molecules capable of interacting withORC have been identi®ed; possible candidates areCdc6, licensing factors Mcms, a Dbf4/Cdc7 kinasecomplex and Cdc45 (Dowell et al., 1994; Hardy, 1997;Liang et al., 1995; Zou et al., 1997). Among these,Cdc6 has been shown to determine the rate ofinitiation through interaction with ORC (Liang et al.,1995). In addition, Cdc6 is necessary for loading ofMcm proteins onto the DNA replication origincomplex (Donovan et al., 1997; Tanaka et al., 1997).cdc18, a Schizosaccharomyces pombe (S. pombe)homolog of CDC6, induces re-replication of DNA,when overexpressed (Muzi-Falconi and Kelly, 1995;Nishitani and Nurse, 1995). These results indicate thatCDC6/cdc18 is an important determinant of theinitiation of DNA replication in yeasts.Expression of CDC6 is cell cycle-regulated. In

rapidly replicating cells, it is expressed at late mitosisand the expression is regulated by Mcm1 (McInerny etal., 1997). In nutrition limited state of yeast cells,CDC6 is expressed at G1/S boundary via activation byMBF (Piatti et al., 1995). cdc18 of S. pombe expressedat G1/S boundary is regulated by cdc10/sct1 (Kelly etal., 1993). MBF and cdc10/sct1 are similar to themammalian transcription factor E2F in that theirtargets are genes involved in DNA replication andactivation of these transcription factors are regulatedby G1 cyclins and cyclin-dependent kinases (Cdks).Thus, although there is little, if any, homology amongE2F, MBF and cdc10/sct1, these transcription factorscan be regarded as functional homologs.Transcriptional activity of E2F is regulated by G1

cyclins/Cdks through phosphorylation of Rb familyproteins. Growth stimulation activates G1 cyclins/Cdksactivity, resulting in expression of a class of genesrequired for DNA replication such as DNA polymer-ase a, dihydrofolate reductase (DHFR) and proliferat-ing cell nuclear antigen (La Thangue, 1994; Nevins,1992). Therefore, the basic control mechanism of G1/Stransition is conserved between yeast and mammals.Overexpression of E2F1 through E2F4 of the E2Ffamily has been shown to induce cellular DNAsynthesis in serum-starved quiescent ®broblasts (De-

Correspondence: M NakamuraReceived 2 April 1998; revised 5 May 1998; accepted 5 May 1998

Oncogene (1998) 17, 1777 ± 1785 1998 Stockton Press All rights reserved 0950 ± 9232/98 $12.00

http://www.stockton-press.co.uk/onc

Gregori et al., 1997; Johnson et al., 1993; Qin et al.,1994; Shan and Lee, 1994). Thus, molecules criticallyrequired for the initiation of DNA replication areprobably targets of transactivation by E2F. However,most of the genes identi®ed so far as E2F targetsencode molecules associated with enzymatic machineryfor DNA synthesis. It is less likely that these geneproducts determine the initiation of DNA replicationin mammalian cells.Identi®cation of molecules that critically regulate the

initiation of DNA replication in mammalian cells iscentral to the study of the molecular mechanismlinking cell cycle and DNA replication. In thiscontext, it is noteworthy that the expression of CDC6and DBF4, interacting with ORC, is cell cycle-regulated in yeast (Chapman and Johnston, 1989;Piatti et al., 1995; Zwerschke et al., 1994). Interest-ingly, the expression of a human ORC1 homolog(HsOrc1) is growth-regulated in an E2F-dependentmanner (Ohtani et al., 1996), unlike the expression ofyeast ORC1 which is independent of the cell cycle.These molecules are potential candidates for determin-ing the initiation of DNA replication. In this regard,the recent identi®cation of a human homolog ofCDC6, p62CDC6 (HsCdc6), prompted us to examinewhether the expression of HsCdc6 is regulated by E2F,considering cell cycle-dependent expression at G1/Sboundary and the presence of putative E2F consensussites in the 5' ¯anking sequence (Williams et al., 1997).We demonstrate here that serum stimulation-

regulated activity of HsCdc6 promoter is primarilycontrolled by E2F through binding to the E2Frecognition sites of the HsCdc6 promoter. In addi-tion, our results also suggest that HsCdc6 as well asHsOrc1, both of which are targets of E2F, are notsu�cient for the initiation of DNA replication.

Results

Regulation of expression of Cdc6 gene in cellproliferation

Expression of HsCdc6 mRNA is induced by serumstimulation in WI38 human ®broblasts, but not ingrowing HeLa cells (Williams et al., 1997). In order tocon®rm that the expression of mammalian Cdc6mRNA is growth regulated, we examined theexpression of Cdc6 mRNA after serum stimulationof quiescent REF52 cells. Quiescence of REF52 cellswas induced by culturing for 48 h in a mediumcontaining 0.1% fetal calf serum and stimulation ofcells by addition of serum to re-enter the cell cycle.RNA was extracted from cells harvested at quiescenceand various time intervals after the addition of serum.Poly(A)+ RNA was puri®ed and assayed for the levelsof Cdc6 mRNA by Northern blotting. Serum-starvedquiescent REF52 cells showed very low but asigni®cant level of mRNA for Cdc6 compared withasynchronously growing cells (Figure 1). In contrast,4 h after serum stimulation, the level of Cdc6 mRNAwas little, if any. A profound increase in the mRNAlevel was observed 8 h after stimulation, and themaximum level occurred at 12 h. Furthermore, apersistently high level of Cdc6 mRNA was presentat least until 28 h after stimulation. DNA synthesis

determined by [3H]thymidine uptake started 16 h afterserum stimulation (data not shown). Accumulation ofCdc6 mRNA seemed to parallel proceeding to G1/Sboundary of the cell cycle progression upon theaddition of serum, indicating the cell growthregulation of the expression of mammalian Cdc6gene. These results are similar to observationspreviously reported in WI38 human ®broblasts(Williams et al., 1997).

Isolation and function of HsCdc6 promoter

In order to elucidate the molecular basis for the cellgrowth-regulated expression of mammalian Cdc6 gene,the next series of experiments aimed at identifying thetranscriptional regulatory sequence. Screening of ahuman genomic DNA library (66105) with a 50 meroligonucleotide corresponding to the 5' end of HsCdc6cDNA identi®ed two positive clones. A 9.1-kb XhoIfragment and 4.6-kb XhoI fragment were separatelyisolated from the two positive clones. Sequenceanalysis showed that the 9.1-kb fragment encom-passed 5' untranslated sequence and 5' ¯ankingsequence of HsCdc6 and this fragment contained the4.6-kb XhoI fragment. No typical TATA-relatedsequences were seen and there were two putativeE2F recognition sites (nt 743 to 736 and nt 78 to71) adjacent to the transcriptional start site (Figure2), as previously reported (Williams et al., 1997).There was no Sp1-like binding site unlike many of theE2F site-containing promoters (Karlseder et al., 1996;Lin et al., 1996). The 5' end of the cDNA reported isnumbered as+1.We introduced these upstream sequences of HsCdc6

cDNA in front of the luciferase (Luc) gene in a reporterplasmid to investigate whether the 5' ¯anking regionpromotes the HsCdc6 gene in response to growthstimulation. Three HsCdc6-Luc reporter plasmidscontaining di�erent lengths of the 5' ¯anking se-quences, pHsCdc6-Luc(79000), pHsCdc6-Luc(74500)and pHsCdc6-Luc(7570), were transfected intoREF52 cells along with pCMV-b-gal as an internalcontrol. After transfection, the serum level in themedium was reduced to 0.1% to render the cellsquiescent for 48 h. Serum was then added to stimulate

Figure 1 Cell growth-dependent expression of Cdc6 mRNA.REF52 cells were serum starved and re-stimulated with 20% FCS.Cells were harvested at indicated time intervals after addition ofserum and poly(A)+ RNA prepared from 660 mg of total RNAwas separated by agarose gel electrophoresis and transferred tonylon membrane. The blot was probed separately with HsCdc6and GAPDH cDNAs. Lane A shows the results with RNA fromasynchronously cultured cells

E2F-mediated expression of mammalian CDC6 geneK Ohtani et al

1778

re-entry into the cell cycle. The promoter activity wasassayed in serum-starved and serum-stimulated REF52cells. Extracts from cells harvested at indicated timeintervals were assayed for luciferase activities. In cellsat early G1, the activity of HsCdc6 promoter of eachof three Luc reporter plasmids was as low as that inquiescent cells (Figure 3). As the cell cycle progressedthrough late G1 and G1/S boundary, luciferaseactivities increased markedly, consistent with thepattern of endogenous Cdc6 mRNA expression(Figure 1). These results indicate that activation ofthe HsCdc6 promoter is induced at this stage of the cellcycle progression. Luciferase activity gradually declinedafter 20 h of stimulation with serum. Essentially, thethree constructs, pHsCdc6-Luc(79000), pHsCdc6-Luc(74500) and pHsCdc6-Luc(7570), showed simi-lar kinetics of activation of the HsCdc6 promoter inresponse to serum stimulation, albeit, with a minordi�erence; pHsCdc6-Luc (7570) responded moremarkedly than pHsCdc6-Luc(79000) and pHsCdc6-Luc(74500), indicating that the essential regulatoryregion in response to serum stimulation is between7570 and+98.

E2F-dependent regulation of HsCdc6 promoter

Based on the results showing the presence of twoputative E2F recognition sequences within the HsCdc6promoter, we then examined whether the E2Frecognition sequences are primary regulatory elementsresponsive to growth stimulation. To test thispossibility, the response of the isolated HsCdc6promoter region to E2F was examined in a transientco-transfection assay. For this purpose, pHsCdc6-Luc(7570) was transfected into REF52 cells togetherwith a plasmid expressing E2F1, E2F2 or E2F3.Following transfection, the cells were cultured in amedium containing 0.1% FCS for 48 h, and assayedfor luciferase activity. Exogenous co-expression ofE2F1 and E2F2 caused ®ve- and eight-fold increasesin HsCdc6 promoter activity, respectively, and stimula-tion to a lesser extent was observed by E2F3 co-expression (Figure 4A). These results indicate that E2Fplays a critical role in the regulation of HsCdc6promoter activity.To con®rm the important role of E2F recognition

sequences in the expression of HsCdc6 gene, thepromoter activity of a 670-bp fragment (7570to+98) mutated in both putative E2F recognitionsequences was compared with the wild type using theluciferase reporter assay. The wild type [pHsCdc6-Luc(7570)] and mutant promoter-linked luciferasereporter [pHsCdc6-Luc(7E2F)] constructs were intro-duced into REF52 cells along with the E2F1 expressionvector. Quiescence of transfected cells was induced byculturing in medium with 0.1% FCS for 48 h, andluciferase activity in cell extracts was determined. Asexpected, under serum starvation, E2F1 expressioninduced 4.8-fold activation of the wild type promoter,however the mutant promoter exhibited a di�erentresponse to E2F. The profoundly high activity of themutant promoter was seen without E2F1. The level ofmutant promoter activity was only slightly elevated byco-expression of E2F1, but the di�erence between with

Figure 2 Nucleotide sequence in the 5' ¯anking region of theHsCdc6 gene. The+1 position was selected arbitrarily accordingto the 5' end of the HsCdc6 cDNA. Two separate putative E2Frecognition sites are marked by frames

Figure 3 Cell growth-dependent induction of HsCdc6 promoteractivity. REF52 cells were transfected with 5 mg of pHsCdc6-Luc(79000), pHsCdc6-Luc(74500) or pHsCdc6-Luc(7570). ACMV-b-galactosidase vector (2 mg) was simultaneously intro-duced as an internal control. Transfected cells were brought toquiescence and then stimulated with serum. Cell extracts wereprepared from cells harvested at indicated time intervals.Luciferase and b-galactosidase activities were measured andactivity of the former was normalized to that of b-galactosidase


1779

and without E2F was not signi®cant (Figure 4B). Thus,we speculated that with regard to the regulation ofexpression of the HsCdc6 gene, E2F acts as a negativeregulator in quiescent G0/G1 cells.

Binding of E2F to HsCdc6 promoter E2F sites

To gain insight into the molecular mechanism ofregulation by the HsCdc6 promoter E2F, we investi-gated the ability of E2F recognition sequences in thepromoter to bind E2F by gel mobility shift assay. TheHsCdc6 E2F recognition sequence oligonucleotides 1and 2 were able to generate complexes with partiallypuri®ed HeLa cell extract. Complex formation wascompeted away with the same oligonucleotides, but notwith mutant forms of HsCdc6 promoter E2Foligonucleotides (Figure 5A).Furthermore, complex formation between HeLa cell

extract and typical E2F binding site of the DHFRpromoter was inhibited by the wild type oligonucleo-tides of E2F recognition sequences 1 and 2 of theHsCdc6 promoter, but not by mutant forms ofoligonucleotides (Figure 5B). These results indicatethat the E2F recognition sequences in the HsCdc6promoter are authentic E2F binding sites.

E2F-dependent activation of HsCdc6 promoter throughcell cycle

Based on our results of growth-dependent control ofCdc6 gene expression, as shown by the Northern blotassay (Figure 1), we investigated the role of E2F sitesin HsCdc6 gene promoter in serum-stimulated growthprogression. The wild type and mutant promoter-

luciferase reporter constructs were transfected intoREF52 cells. Quiescence of transfected cells wasinduced by culturing in 0.1% FCS for 48 h followedby stimulation with serum. Cell extracts were preparedat various time intervals after serum stimulation andassayed for luciferase activities. The basal activity ofthe wild type HsCdc6 promoter markedly increased atG1/S boundary (Figure 6) consistent with the resultsshown in Figure 3. In contrast, the mutant promoterexhibited completely di�erent kinetics relative to thoseof the wild type promoter, abolishing cell growth-dependent control of promoter activity (Figure 6). Theactivity of the mutant promoter in quiescence and earlyG1 stage was profoundly elevated. This was consistentwith the results of co-transfection experiments shownin Figure 4. These results demonstrate that the activityof the HsCdc6 promoter is regulated by E2F whichprimarily suppresses its expression in G0/G1 as hasbeen reported for other E2F responsive genes such asB-myb, E2F1 and HsOrc1 (Johnson et al., 1994; Lamand Watson, 1993; Ohtani et al., 1996).

Induction of endogenous Cdc6 gene expression by E2F1

Based on our analysis, it seemed likely that E2F iscentral in regulating the expression of HsCdc6 gene.The next series of experiments were performed tocon®rm these results by examining the expression ofendogenous Cdc6 gene in response to E2F in quiescentstate. REF52 cells were rendered quiescent by serumstarvation then infected with a recombinant adenoviruswhich expresses E2F1. Cells were harvested at 21 hafter infection and RNA was extracted for Northernblot analysis (Figure 7). The level of Cdc6 mRNA

Figure 4 Activation of the HsCdc6 promoter by E2F. (A) Activation by members of the E2F family. REF52 cells were transfectedwith 5 mg of pHsCdc6-Luc(7570) and 0.2 mg of CMV promoter-driven expression vectors for either E2F1, E2F2 or E2F3. Aparental CMV vector, pcDNA1, was used as a control. pCMV-b-gal (0.2 mg) was co-transfected as an internal control. Transfectedcells were cultured in medium with 0.1% FCS for 48 h. Luciferase and b-galactosidase activities of cell extracts were measured.Luciferase activity was normalized to b-galactosidase activity, and is plotted for each co-transfected CMV expression vector. (B)E2F site-dependent regulation of the HsCdc6 promoter. REF52 cells were transfected with 5 mg of pHsCdc6-Luc(7570) (WT) ormutant vector pHsCdc6-Luc(-E2F) (Mut), in which both putative E2F recognition sites were mutated, along with E2F1 vector orpcDNA1


1780

increased 4.5-folds after infection with Ad-E2F1 viruscompared with control virus infected cells. Theseresults indicate that E2F plays an important role inactivation of the Cdc6 promoter not only in a transienttransfection assay but also endogenous Cdc6 geneexpression.

E�ects of overexpression of HsCdc6 and HsOrc1 oncellular DNA synthesis

Overexpression of E2F1 has been shown to initiatecellular DNA synthesis in serum starved REF52 cells(Johnson et al., 1993). HsCdc6 is an E2F-regulatedgene and, furthermore, CDC6 determines the frequencyof ®ring of DNA replication origins in yeast. Thus, it ispossible that enforced expression of HsCdc6 inducesthe initiation of DNA replication in mammalian cellsas is the case for overexpression of cdc18 in S. pombe.This notion was examined by cellular DNA synthesisin serum starved REF52 cells overexpressing HsCdc6.REF52 cells were transfected with the HsCdc6expression vector along with a b-galactosidase expres-sion vector and BrdU incorporation was measured inserum starved conditions. As shown in Figure 8,expression of E2F1 induced BrdU incorporation in

40% of cells positive for b-galactosidase expression.This value, similar to the results of Johnson et al.(1993), was signi®cantly higher than that of cellstransfected with a control vector. In contrast,transfection with the HsCdc6 expression vector failedto induce signi®cant changes in BrdU incorporationrelative to the background, even with as much as 10times the plasmid DNA, in which the HsCdc6 gene wasdriven by the same CMV promoter as E2F1 expressionvector. These results indicate that the HsCdc6expression alone is not su�cient for induction ofDNA replication in REF52 cells.In view of the fact that the expression of HsOrc1 is

regulated by E2F in a growth stimulation mannersimilar to HsCdc6 (Ohtani et al., 1996), we alsoexamined whether transfection with HsOrc1 expres-sion vector could induce cellular DNA synthesis in asimilar fashion. Introduction of HsOrc1 expressionplasmid DNA at 0.5 mg and 5.0 mg did not alter thenumber of BrdU positive cells relative to the back-ground, demonstrating that HsOrc1 alone is notsu�cient for initiation of DNA replication. Collec-tively, neither Cdc6 nor Orc1 is the sole E2F targetgene for the induction of cellular DNA synthesis inREF52 cells.

Figure 5 Binding of E2F to the HsCdc6 promoter. (A) Speci®city of E2F binding to the HsCdc6 promoter. Gel mobility shift assays wereperformed with partially puri®ed HeLa cell extract and end-labeled double stranded oligonucleotides containing the putative E2F recognitionsite 1 (lanes 1 ± 4) and putative E2F recognition site 2 (lanes 5 ± 8) from the HsCdc6 promoter. Competition assays were performed by adding100-molar excess of the unlabeled oligonucleotides of wild and mutant forms to gel shift reaction mixtures. (B) Competitive activity in E2Fbinding. Gel mobility shift assays were performed with partially puri®ed HeLa cell extract. An end-labeled DNA fragment which containstwo overlapping E2F recognition sites of the DHFR promoter was used as a probe. Competition experiments were performed by adding 100-fold molar excess of unlabeled oligonucleotides to gel shift reaction mixtures. Competitor oligonucleotides included the putative E2Frecognition site 1 from the HsCdc6 promoter (lane 11) and its mutant form (lane 12), putative E2F recognition site 2 (lane 13) and itsmutant form (lane 14). Lanes 9 and 10 contain no cell extract, and cell extract with no competitors, respectively


1781

Discussion

The major ®nding of the present study is that thegrowth-regulated expression of mammalian Cdc6 ismediated by E2F.The 570-bp region upstream of HsCdc6 encompasses

the transcriptional regulatory region of the HsCdc6gene, and the kinetics of promoter activity in response

to serum stimulation, as determined by the reporterassay with the isolated fragment, were similar to thoseof the endogenous promoter assayed by Northern blothybridization (Figure 3 vs Figure 1). The fragmentcontains two E2F binding sites which are close to eachother and adjacent to the RNA start site. It is ofinterest to note that most E2F binding sites in genesthat are primarily regulated in expression by E2F arelocated near RNA start sites in a multiple copy form intranscriptional regulatory regions.Disruption of E2F binding sites in the HsCdc6

promoter abolished the ability to regulate transcriptionof the gene in response to growth stimulation, clearlydemonstrating that E2F is responsible for regulation ofexpression of the HsCdc6 gene. This conclusion wassupported by the ®nding that, in quiescent cells, theenforced expression of E2F was able to relieverepression of the HsCdc6 gene promoter. E2F on theHsCdc6 promoter seems to be suppressive in quiescentcells, because a mutant HsCdc6 promoter disruptedwith the E2F binding sites exhibited promoter activitywithout serum stimulation as high as that of the wildtype promoter in cells at G1/S boundary. This isconsistent with previous observations of E2F-depen-dent regulation with genes for B-myb, E2F1 andHsOrc1 (Johnson et al., 1994; Lam and Watson,1993; Ohtani et al., 1996). Previous studies showedthat quiescent REF52 cells have E2F in a complexform preferentially associated with p130 in contrast tocycling REF52 cells having E2F complexed with pRband p107 (Smith et al., 1996), suggesting that the E2F/p130 complex functions as a machinery that silencesthe HsCdc6 promoter. Overexpression of E2F could

Figure 6 E2F binding site-dependent control of the HsCdc6promoter in response to cell growth. REF52 cells were transfectedwith 5 mg of either the reporter plasmid pHsCdc6-Luc (7570) orE2F site mutant vector pHsCdc6-Luc(7E2F), along with 0.2 mgof pCMV-b-gal. Cells were brought to quiescence and thenstimulated with serum for the indicated time intervals. At eachtime interval, cells were harvested, extracts were prepared, andluciferase and b-galactosidase activities were measured. Luciferaseactivity was normalized to b-galactosidase activity

Figure 7 Endogenous Cdc6 gene expression by E2F1. REF52cells were brought into quiescence by serum starvation for 48 h.Cells were ®rst infected with either Ad-E2F1 or Ad-Con thencultured with 0.1% FCS for 21 h. Poly(A)+ RNA was extractedfrom 3 mg of total RNA and then separated by agarose gelelectrophoresis. The RNAs were transferred to a nylon membraneand probed with either HsCdc6 or GAPDH cDNA

Figure 8 E�ects of overexpression of HsCdc6 or HsOrc1 oncellular DNA synthesis. Serum starved quiescent REF52 cellswere transfected with either pCMV-HsCdc6 or pCMV-HsOrc1along with pCMV-b-gal. Transfected cells were cultured inDMEM with 0.1% FCS for 48 h and BrdU was added for 24 hbefore harvest. Cells were ®xed, stained with anti-b-galactosidaseand anti-BrdU antibodies with ¯uorescence, and examined fortheir staining under a ¯uorescence microscope. Each barrepresents the average of four independent examinations. Morethan 200 cells were counted in each sample. Data are mean+s.d.


1782

yield E2F molecules free from the complex with Rbfamily members, resulting in relief from the suppres-sion.The mechanism of initiation of DNA replication in

mammalian cells has been investigated by identifyingcounterparts homologous to yeast components in-volved in the regulation of initiation of DNAreplication. Successful identi®cation of human homo-logs of ORC1 and 2, CDC6 and 7, and MCM2-7(Fujita et al., 1996; Gavin et al., 1995; Hu et al., 1993;Musahl et al., 1995; Sato et al., 1997; Thommes et al.,1992; Todorov et al., 1995; Tsuruga et al., 1997b;Williams et al., 1997) suggests that the main stream ofthe regulatory mechanism involved in initiation ofDNA replication is conserved between humans andyeasts. Indeed, the mechanisms that regulate theexpression of CDC6 gene are similar in humans andyeasts. As shown in the present results as well as thosefrom other laboratories (Williams et al., 1997), theHsCdc6 gene is transcribed at G1/S boundary in anE2F-regulatory manner. S. cerevisiae under nutrition-limited conditions similarly expressed CDC6 mRNA atG1/S boundary in an MBF-dependent manner. MBF isthought to be functionally equivalent to E2F inmammalian cells. However certain di�erences inregulation of the expression of homologous genesexist between both species. Even in S. cerevisiae, theCDC6 gene in rapidly replicating cells is expressed atlate mitosis under regulation by Mcm1 but not byMBF (McInerny et al., 1997). Another example is theORC gene. Expression of ORC mRNA and protein inS. cerevisiae is independent of the cell cycle; incontrast, HsOrc1 mRNA expression in human fibro-blasts is greatly induced upon growth stimulation(Ohtani et al., 1996). Thus, it is suggested thatHsOrc1 is involved in regulation of the initiation ofDNA replication in response to growth stimulation inmammalian cells. Expression of Mcm genes is alsodi�erentially regulated in yeasts and humans. In yeasts,the expression of Mcms is independent of the cell cycle.However, in human, a set of the MCM gene family hasbeen shown to be induced upon growth stimulation(Tsuruga et al., 1997a). Preliminary results from ourlaboratory indicate that the expression of humanhomologs of MCM5 and MCM6 genes is regulatedby E2F (unpublished data).The growth-dependent expression of a group of

genes regulatory to DNA replication is reasonable,considering that most cells are quiescent in adults anda small population of cells are cycling or induced toproliferate in certain conditions such as injury andimmune stimulation. In this regard, more genesinvolved in the regulation of initiation of DNAreplication might have become growth-responsive inmammals. Moreover, this regulation is achieved by theaction of E2F transcription factor. These observationsunderscore the importance of E2F not only in cell cycleregulation but also in cell growth in general. Thus, it isimportant to know whether the expression of thesegenes is cell cycle-regulated in cycling cells, and, if so,expression of which of these genes is important forregulation of the initiation of DNA replication. Atpresent, it is unclear which molecule among themachinery interacting with ORC is critical for theinitiation of DNA replication in mammalian cells. Themechanism of assembly of these gene products induced

by E2F may be critical for such initiation process, andthe assembly may even depend on a critical amount ofeach molecule. Alternatively, the fully functional ORCcomplex including Orcl and Cdc6 may be activated byother yet unknown mechanism(s) which may be alsoregulated by molecules a�ected by E2F in theirexpression. In this regard, it is noteworthy thatexpression of cyclin E, which is necessary for entryinto S phase, is regulated by E2F (Geng et al., 1996;Ohtani et al., 1995). Cyclin E-dependent kinase activitycould be required for activation of the functional ORCcomplex by phosphorylation of some components inthe complex.In S. cerevisiae, the CDC6 gene product determines

the rate of ®ring of origins (Liang et al., 1995). On theother hand, in S. pombe, overexpression of CDC6homolog, cdc18, has been shown to induce re-replication of DNA (Nishitani and Nurse, 1995).Thus, it seems that CDC6/cdc18 is the main regulatorof initiation of DNA replication in yeasts. In ourexperimental system, overexpression of either HsOrc1or HsCdc6 alone did not induce DNA synthesis unlikeoverexpression of E2F1 in REF52 cells. Thus, it isclear that Orc1 or Cdc6 alone is not the sole E2Ftarget gene in inducing DNA synthesis in REF52 cells.A series of experiments in yeast suggested that bothOrc1 and Cdc6 may be critical targets of E2F ininducing DNA replication in REF52 cells. Consideredtogether, the expressions of Orc1 and Cdc6 may benecessary but not su�cient for induction of initiationof DNA synthesis in REF52 cells, suggesting thatother molecules may be necessary for the initiation ofDNA replication. If this is true, E2F overexpressionmust induce the expression of those molecules inaddition to Orc1 and Cdc6. Among such molecules,members of the MCM family and a mammalianCDC7 homolog are potential candidates. Further-more, as yet an unidenti®ed mammalian homolog ofyeast DBF4 may also be involved in initiation ofDNA replication, considering the presence of mam-malian CDC7 homolog and the cell cycle-dependentexpression of DBF4 by MBF in yeast. Thus, acombination of those molecules may bypass therequirement of normal regulation for the initiationof DNA replication.It is noteworthy that overexpression of E2F induces

cellular DNA synthesis in ®broblasts with kineticssimilar to those of serum stimulation (DeGregori et al.,1995; Kowalik et al., 1995). Considering that E2Factivity is tightly involved in G1 cyclin /Cdk activation(Geng et al., 1996; Ohtani et al., 1995), it is possiblethat overexpression of E2F induces a set of genes thatare activated by serum stimulation and evokes acascade of reactions as in serum stimulation. Induc-tion of Orc1 and Cdc6 could be one set of such genesin this scenario.

Materials and methods

Cell culture

A rat embryonic ®broblast cell line REF52 was maintainedin Dulbecco modi®ed Eagle medium (DMEM) containing10% FCS. For experiments of promoter activity assay andexpression of endogenous Cdc6 mRNA, REF52 cells were


1783

serum starved in DMEM containing 0.1% FCS for 48 hand stimulated by the addition of 20% FCS for indicatedtime intervals.

Isolation of HsCdc6 promoter sequences

A human placenta genomic DNA library (66105 clones,Clontech) was screened with a 50 mer oligonucleotidecorresponding to the 5' end of the published HsCdc6cDNA sequence (Williams et al., 1997). Labeling of theprobe with DIG-UTP, hybridization and detection ofpositive signals were carried out using DIG Oligonucleo-tide 3'-End Labeling Kit and DIG Luminescent DetectionKit (Boehringer Mannheim, Germany) according to theprotocol recommended by the manufacturer with thefollowing modi®cations. Hybridization was performedwith 45% formamide at 408C and washing was performedin a solution with 0.56SSC and 0.1% SDS at 558C. Twopositive clones were ®nally isolated. Sequencing wasperformed by the primer walking method using a DNAsequencing kit and 310 genetic analyzer (Perkin Elmer)according to the protocol recommended by the manufac-turer.

Construction of plasmids

The 9.1-kbp XhoI fragment (nt 79000 to+134) and 4.6-kbp XhoI fragment (nt 74500 to+134) were isolated fromthe two clones using XhoI sites in the 5' untranslatedregion of HsCdc6 cDNA and in the multiple cloning sitesof the phage vector, and subcloned into the XhoI site of aLuc reporter plasmid pGL2basic (Promega) to makepHsCdc6-Luc(79000) and pHsCdc6-Luc(74500), respec-tively. A 670-bp BamHI fragment (nt 7570 to+98),common for both clones, was subcloned into the BglII siteof pGL2basic to make pHsCdc6-Luc(7570). Mutationwas introduced into the putative E2F recognition sites ofthe HsCdc6 promoter region in the pHsCdc6-Luc(7570)construct by site-directed mutagenesis which changedTTTGGCGG (nt 743 to 736) to TTTGATGG andTTTGGCGC (nt 78 to 71) to TTTGATGC, generatingpHsCdc6-Luc(-E2F). The E2F1, E2F2, and E2F3 expres-sion plasmids, as well as b-galactosidase expression vector,pCMV-b-gal, have been described previously (Johnson etal., 1994). To prepare HsCdc6 and HsOrc1 expressionvectors, HsCdc6 cDNA (nt 61 to 1917) and HsOrc1cDNA (nt 84 to 2963) ampli®ed by PCR using Pfu DNApolymerase (Stratagene) from 293 cell cDNA were ®rstcloned into the EcoRV site of pBluescript SK7 generatingpBSK-HsCdc6 and pBSK-HsOrc1, respectively. Subse-quently, the HsCdc6 cDNA and HsOrc1 cDNA were cutout and subcloned into an expression vector pcDNA3(Invitrogen) to make pCMV-HsCdc6 and pCMV-HsOrc1,respectively. Sequencing of both cDNAs found anucleotide substitution (A to G) at nt 154 in the 5'untranslated region of the HsCdc6 cDNA and C to Achange at nt 1965 of the HsOrc1 cDNA which changed anamino acid at 582 from His to Gln which is not conservedbetween species.

Transfection assay

Transfection of REF52 cells and luciferase assay wereperformed as described previously (Johnson et al., 1993).Expression plasmids and reporter plasmids were introducedinto REF52 cells by the calcium-phosphate method. b-Galactosidase expression vector pCMV-b-gal was co-transfected as an internal control to monitor transfectione�ciency. Luciferase activities were normalized to b-galactosidase activities. All assays were performed at leastthree times in duplicates and representative data arepresented.

Gel mobility shift assay

Double-stranded oligonucleotides containing each of thetwo putative E2F recognition sites (nt 743 to 736 and78 to 71) in the 5' ¯anking region of HsCdc6 andmutants of these sequences were used for the assay.Sequences are: wild type-1: 5'-CCGGGCTTTGGC-GGGAGGTGGG-3'; mutant-1: 5'-CCGGGCTTTGATG-GGAGGTGGG-3'; wild type-2: 5'-TTCGGATTTGGCG-CGAGCGCGG-3'; mutant-2: 5'-TTCGGATTTGATGC-GAGCGCGG-3'.

These double-stranded oligonucleotides were also used ascompetitors at 100-fold molar excess. A typical E2F site fromthe DHFR promoter (Ikeda et al., 1996) was also used as aprobe. Gel mobility shift assay was performed as describedpreviously (Yee et al., 1989). In brief, radio-labeled probeswere incubated with partially puri®ed HeLa nuclear extract(Ikeda and Nevins, 1993), in the presence or absence ofunlabeled competitor oligonucleotides for 20 min at roomtemperature, followed by eletrophoresis in a 5% polyacryla-mide gel in TBE (50 mM Tris-borate/1 mM EGTA) contain-ing 5% glycerol.

Northern (RNA) blot assay

Total RNA extraction and poly (A)+ RNA puri®cationwere carried out using Isogen (Nippon Gene) and PolyATract (Promega), respectively, according to the protocolrecommended by the manufacturer. Gel electrophoresis,transfer to nylon membrane, and hybridization wereperformed as described previously (Johnson et al.,1994). The HsCdc6 probe was a 1856-bp SmaI-SalIfragment cut out of pBSK-HsCdc6. The GAPDH cDNAwas used as a control probe. Radioactivity of the signalswas measured with an image analyzer BAS 1500 (FujiFilm).

Infection with recombinant adenoviruses

The recombinant adenovirus for expression of E2F1, Ad-E2F1, and control virus, Ad-Con (previously Ad-CMV),were described previously (Schwarz et al., 1995). Infectionof REF52 cells with Ad-E2F1 or Ad-Con was performed asdescribed previously (Schwarz et al., 1995). Brie¯y,quiescent REF52 cells cultured in DMEM containing0.1% FCS for 48 h were infected with Ad-E2F1 or Ad-Con at multiples of 300 plaque forming units per cell in2 ml per 150 mm plate for 1 h at 378C. Cells were furthercultured in DMEM containing 0.1% FCS for 21 h andharvested for RNA isolation.

DNA synthesis induction assay

The assay was performed using the method describedpreviously by Johnson et al. (1993) with minor modifica-tions. Brie¯y, REF52 cells were transfected by the calciumphosphate coprecipitation method with 5 mg RSV-b-galand 500 ng or 5 mg of either E2F1, HsCdc6 or HsOrc1expression vector. pcDNA1 was used as a control vectorwhile pBluescript SK7 was used as a carrier DNA toprepare a total DNA amount of 20 mg. Cells were exposedto DNA precipitate for 16 h, washed and then cultured inDMEM containing 0.1% FCS for 48 h. BrdU was addedto a ®nal concentration of 50 mM and the cells were furtherincubated for 24 h. The cells were ®xed with 4% para-formaldehyde in PBS for 10 min at room temperature andthen dehydrated with methanol/acetone (1 : 1) for 5 min atroom temperature. b-Galactosidase was stained with arabbit polyclonal antibody (1 : 500 in 1% BSA/PBS)(Cappel) and FITC-conjugated goat anti-rabbit IgG(1 : 500 in 1% BSA/PBS) (Cappel). Cells were ®xed againwith 4% paraformaldehyde and then treated with 2 N HClfor 30 min at room temperature. They were subsequently


1784

stained with an anti-BrdU monoclonal antibody (1 : 50 in1% BSA/PBS) (Boehringer-Mannheim) and rhodamine-conjugated goat anti-mouse IgG (1 : 500 in 1% BSA/PBS)(Cappel). The number of BrdU positive cells in b-galactosidase positive cells was counted under a fluores-cence microscope equipped with appropriate ®lters.Experiments were repeated four times and in eachexperiment, more than 200 b-galactosidase positive cellswere counted in each sample. Results were presented as amean of four experiments.

Nucleotide sequence accession number

The nucleotide sequence data reported in this paper havebeen given a sequence of DDBJ, EMBL and GenBankaccession number of AB010492.

AcknowledgementsThis work was supported in part by the Grant-in-Aid forGeneral Scienti®c Research and Cancer Research from theMinistry of Education, Science, Sports and Culture ofJapan, in part by a grant for CREST (Core Research forEvolutional Science and Technology) of the Japan Scienceand Technology Corporation (JST) and in part by thegrant from the Kanehara Foundation. We thank YMatsumura for excellent technical assistance.

References

Aparicio OM, Weinstein DM and Bell SP. (1997). Cell, 91,59 ± 69.

Bell SP and Stillman B. (1992). Nature, 357, 128 ± 134.Chapman JW and Johnston LH. (1989). Exp. Cell Res., 180,419 ± 428.

DeGregori J, Kowalik T and Nevins JR. (1995). Mol. Cell.Biol., 15, 4215 ± 4224.

DeGregori J, Leone G, Miron A, Jakoi L and Nevins JR.(1997). Proc. Natl. Acad. Sci. USA, 94, 7245 ± 7250.

Di�ey JF and Cocker JH. (1992). Nature, 357, 169 ± 172.Donovan S, Harwood J, Drury LS and Di�ey JF. (1997).

Proc. Natl. Acad. Sci. USA, 94, 5611 ± 5616.Dowell SJ, Romanowski P and Di�ey JF. (1994). Science,265, 1243 ± 1246.

Fujita M, Kiyono T, Hayashi Y and Ishibashi M. (1996). J.Biol. Chem., 271, 4349 ± 4354.

Gavin KA, Hidaka M and Stillman B. (1995). Science, 270,1667 ± 1671.

Geng Y, Eaton EN, Picon M, Roberts JM, Lundberg AS,Gi�ord A, Sardet C and Weinberg RA. (1996). Oncogene,12, 1173 ± 1180.

Hardy CF. (1997). Gene, 187, 239 ± 246.Hu B, Burkhart R, Schulte D, Musahl C and Knippers R.(1993). Nucleic Acids Res., 21, 5289 ± 5293.

Ikeda MA, Jakoi L and Nevins JR. (1996). Proc. Natl. Acad.Sci. USA, 93, 3215 ± 3220.

Ikeda MA and Nevins JR. (1993).Mol. Cell. Biol., 13, 7029 ±7035.

Johnson DG, Ohtani K and Nevins JR. (1994). Genes & Dev.,8, 1514 ± 1525.

Johnson DG, Schwarz JK, Cress WD and Nevins JR. (1993).Nature, 365, 349 ± 352.

Karlseder J, Rotheneder H and Wintersberger E. (1996).Mol. Cell. Biol., 16, 1659 ± 1667.

Kelly TJ, Martin GS, Forsburg SL, Stephen RJ, Russo Aand Nurse P. (1993). Cell, 74, 371 ± 382.

Kowalik TF, DeGregori J, Schwarz JK and Nevins JR.(1995). J. Virol., 69, 2491 ± 2500.

La Thangue NB. (1994). Trends Biochem. Sci., 19, 108 ± 114.Lam EW and Watson RJ. (1993). EMBO J., 12, 2705 ± 2713.Liang C, Weinreich M and Stillman B. (1995). Cell, 81, 667 ±676.

Lin SY, Black AR, Kostic D, Pajovic S, Hoover CN andAzizkhan JC. (1996). Mol. Cell. Biol., 16, 1668 ± 1675.

McInerny CJ, Partridge JF, Mikesell GE, Creemer DP andBreeden LL. (1997). Genes & Dev., 11, 1277 ± 1288.

Musahl C, Schulte D, Burkhart R and Knippers R. (1995).Eur. J. Biochem., 230, 1096 ± 1101.

Muzi-Falconi M and Kelly TJ. (1995). Proc. Natl. Acad. Sci.USA, 92, 12475 ± 12479.

Nevins JR. (1992). Science, 258, 424 ± 429.Nishitani H and Nurse P. (1995). Cell, 83, 397 ± 405.Ohtani K, DeGregori J, Leone G, Herendeen DR, Kelly TJand Nevins JR. (1996). Mol. Cell. Biol., 16, 6977 ± 6984.

Ohtani K, DeGregori J and Nevins JR. (1995). Proc. Natl.Acad. Sci. USA, 92, 12146 ± 12150.

Piatti S, Lengauer C and Nasmyth K. (1995). EMBO J., 14,3788 ± 3799.

Qin XQ, Livingston DM, Kaelin Jr WG and Adams PD.(1994). Proc. Natl. Acad. Sci. USA, 91, 10918 ± 10922.

Sato N, Arai K and Masai H. (1997). EMBO J., 16, 4340 ±4351.

Schwarz JK, Bassing CH, Kovesdi I, Datto MB, Blazing M,George S, Wang XF and Nevins JR. (1995). Proc. Natl.Acad. Sci. USA, 92, 483 ± 487.

Shan B and Lee WH. (1994).Mol. Cell. Biol., 14, 8166 ± 8173.Smith EJ, Leone G, DeGregori J, Jakoi L and Nevins JR.(1996). Mol. Cell. Biol., 16, 6965 ± 6976.

Tanaka T, Knapp D and Nasmyth K. (1997). Cell, 90, 649 ±660.

Thommes P, Fett R, Schray B, Burkhart R, Barnes M,Kennedy C, Brown NC and Knippers R. (1992). NucleicAcids Res., 20, 1069 ± 1074.

Todorov IT, Attaran A and Kearsey SE. (1995). J. Cell.Biol., 129, 1433 ± 1445.

Tsuruga H, Yabuta N, Hashizume K, Ikeda M, Endo Y andNojima H. (1997a). Biochem. Biophys. Res. Commun., 236,118 ± 125.

Tsuruga H, Yabuta N, Hosoya S, Tamura K, Endo Y andNojima H. (1997b). Genes Cells, 2, 381 ± 399.

Williams RS, Shohet RV and Stillman B. (1997). Proc. Natl.Acad. Sci. USA, 94, 142 ± 147.

Yee AS, Raychaudhuri P, Jakoi L and Nevins JR. (1989).Mol. Cell. Biol., 9, 578 ± 585.

Zou L, Mitchell J and Stillman B. (1997).Mol. Cell. Biol., 17,553 ± 563.

Zwerschke W, Rottjakob HW and Kuntzel H. (1994). J. Biol.Chem., 269, 23351 ± 23356.


1785

Documents

Regulation of cell growth-dependent expression of mammalian CDC6 gene by the cell cycle transcription factor E2F