The mdm-2 gene is induced in response to UV light in a p53-dependent manner

Proc. Natl.Acad. Sci. USA

Vol. 90, pp. 11623-11627,December 1993 CellBiology

The mdm-2 gene

is induced

in response to

UV

light

in a

p53-dependent manner

(cellcycle/transcription)

MARY ELLEN

PERRY, JACQUES PIETTE*,

JEROME A.

ZAWADZKI,

DIANE

HARVEYt,

ANDARNOLD J.

LEVINEt

Department of Molecular Biology, Princeton University, Princeton, NJ 08544-1014

ContributedbyArnold J. Levine, August31, 1993

ABSTRACT Irradiation ofmammalian cellswithUVlight results in a dose-dependent accumulation of the p53 tumor-suppressor gene product that is evident within 2 hr. UV treatmentcauses a dramatic increase inp53-specific transcrip-tional transactivation

activity

and an increase in expression of thep53-responsive gene mdm-2. UV-stimulated mdm-2 expres-sion is notdirectycorrelated with the level of p53 protein in a cell becausemdm-2inductionisdelayed at highUVdoses even though p53 levels rise almost immediately. Cells lacking p53 protein do not respond to UV by increasing their expression of mdm-2. The delayed induction of mdm-2 at high UV doses suggests that, inadditionto p53 proteinlevels,other factors contribute to theregulation of mdm-2 expressionfollowing UV treatment.The time of induction of mdm-2 incellstreated with UV light correlates with recovery ofnormal rates of DNA synthesis, presumably after DNA repair. These data indicate a possible role for mdm-2 in cell cycle progression.

Mammalian cells respond to irradiation with UV light by transiently decreasingboth RNA and DNAsynthesis and by inducing expression of several genes whose products are thought to haveprotective effects against DNA damage (1). Theregulation of these UV response genes appears to be mediated by several transcription factors which function after UV exposure (2). The p53 tumor-suppressor gene productisatranscription factor(3, 4) that also appears to be involved in the response to UVlight.Thep53 proteinlevels increase due to thestabilizationof this protein in both murine (5, 6)and human(7, 8)cellstreated with UVlight. Although therole ofp53in the response to UVlighthas not beenfully characterized,this tumor-suppressorproteinhasbeen shown to act as a cell cycle checkpoint in the response to y irradiation(9, 10). fyirradiation induces bothaG1and a G2 phase-specific cell cycle block, andexpression of wild-type

p53

isnecessaryfor theG1butnottheG2block. Thespecific DNA-binding activity ofp53 isincreased after yirradiation and aDNAdamage-inducible, growth arrest-specific gene, GADD45,hasbeenshown tocontainap53responseelement (10).

Characterization of the role of

p53

in the response to

'y

irradiationled to thehypothesisthatp53actsas acell cycle checkpoint, causing a delay in the G1 phase of the cycle during which damage is thought to be repaired (9, 10). It seemedlikelythat

p53

mightalso act as acheckpointin the responseof cellstoUV exposure. Inaddition,Zhanetal.(8) haverecentlyshown that

p53

transcriptionaltransactivation activity is increased in human cells exposed to UV

light.

Possible targetsfor

p53

transcriptionaltransactivation activ-ity in the UV response are the GADD45 gene, which was

isolatedas aUV response gene(11),and themdm-2 gene.

p53

and mdm-2 appeartoformafeedback controlloop:while

p53

Thepublicationcostsofthis articleweredefrayedin partbypagecharge

payment. Thisarticlemusttherefore beherebymarked"advertisement"

in accordance with 18U.S.C.§1734solelytoindicate this fact.

caninduce mdm-2transcription (12, 13),highlevelsofmdm-2 protein inhibit p53 transcriptional transactivation activity (13, 14). Theresults reportedheredemonstrate thattreatment ofmurinecellswith UVlight resultsinasuppression ofDNA synthesisthatis independent ofthe presenceofp53. Incells expressing functional wild-type p53, the mdm-2 gene is induced following a lag time that increases with UV dose. Thisenhanced expression of mdm-2 precedes the appearance ofcellsentering S phase at normal rates of DNA synthesis. These data are consistent with the idea that mdm-2 may reversethe possible negative regulatory effects of p53 in the G1phaseof thecell cycleand/orpromotethe entryof cells into S phase after DNA damage is presumably repaired.

MATERIALS

AND

METHODS

Cell Culture. Cells were grown in Dulbecco's modified Eagle's medium supplemented with 10% heat-inactivated fetal bovine serum in an atmosphere of 10%Co2.

UV Light Treatment. Cells were treated with UV light essentially asdescribed (5).The medium was removed and thecells were irradiated with an 8-Wgermicidal lamp deliv-ering 1.9

J'm-2 sec-1

at a distance of 40 cm. The dose was quantitated with a J-225 short-wave UV meter (Ultraviolet Products, SanGabriel, CA).

FlowCytometry.Cells were labeled for 30 min with 10 ,uM BrdUrdatvarious timesfollowing UV irradiation. The cells were trypsinized,

fixed

in 70% ethanol (-20°C), and later stainedwith anti-BrdUrd antibodies conjugated to fluores-ceinisothiocyanate (Becton Dickinson) essentiallyas recom-mendedbythemanufacturer. Sampleswereanalyzedwithan EPICS 753 cell sorter (Coulter) with an MDADS II data system. An argon laser(488 nm, 500 mW) was used for the

measurementofthelogarithmoffluorescein fluorescenceat 515-530nmand linearpropidium iodide fluorescenceat>610

nm.

Western Analysis. Following incubation of 3 mg of each lysatewith the monoclonalantibodiesPAb419 (specific for simian virus 40largeTantigen)orPAb421(specific forp53)

or with anti-mdm-2 serum, the immunoprecipitates were

separated by

SDS/polyacrylamide

gelelectrophoresis along with high molecular weight prestained markers (Bethesda ResearchLaboratories) and blotted onto Hybond ECL ni-trocellulose(Amersham). The membrane was sliced atthe 68-kDa prestained marker and the half of the membrane containingthe higher molecular weight proteins was incu-bated with the anti-mdm-2 serumwhile the other halfwas

incubated withPAb421. Antibody bindingwasdetected with Abbreviations: CAT, chloramphenicol acetyltransferase; MCK,

muscle-specificcreatine kinase.

*Present address: Centre National de la Recherche Scientifique, France.

tPresentaddress:RockefellerUniversity, 1230YorkAvenue,New York,NY10021.

*To

whomreprintrequests should beaddressed. 11623

125I-protein

A (NEN), andquantitation was on aMolecular Dynamics Phosphorlmager.

Chloramphenicol

Acetyltransferase

(CAT) Assays. The transfection andUV treatmentprotocol is essentially that of Devary et al. (2). A total of 10 ug of DNA (5

t.g

ofCAT constructand 5pg of salmonspermcarrierDNA)wasused, and CAT assays were as described(15). Datawere quanti-tated on aPhosphorImagerandnumbersquoted in the text

are the average oftwo duplicates. The total conversion of

[14C]chloramphenicol

to acetylatedproductwas<40%oin all reactions.

RESULTS

Wild-Typep53 Protein Increases inaDose-Dependent

Man-ner inResponse to UV Irradiation.Maltzman andCzyzyk (5) showed thatp53 levelsin

BALB/C3T3

cellsrose withinthe first2 hrafterUVtreatment.Twodifferentimmortalmurine cell lines (C127 and 12.1) expressing wild-type p53

protein

(16, 17)were treatedwithincreasing doses ofUVlightand analyzed for changes in p53 levels. Both cell lines showed increasing p53protein with increasing doses

(Fig.

1anddata notshown).At thehighest dose used (24.7

J/m2,

3% colony-forming ability, Fig. 1) C127 cells exhibiteda4-fold increase inthe incorporation of

[35S]methionine

into the p53

protein

immediatelyfollowing UVtreatment. The greatestincrease inp53expressionin thefirst 2 hr after UVtreatmentisincells sustainingthemostdamage.Fritscheetal.(6)haveindicated

apossible conformation change inthep53

protein

stabilized A 419

p53

UV Dose (J/m 2) 421 *:~~~~~~~~~~~~~~~~~~~~~~~~~~~... .. .. .. ... ... -O O 1.9 3.8 7.6 13.3 19.0 24.7 B 1000 'a) V1) 40~ 1 00 1 0 1 0 10 20 UV Dose

(J/m2

)

FIG. 1. Increaseinp53as afunction of UV dose andinhibition

ofcolonyformationability. (A)C127 cellswereirradiatedwithaUV

lampatthe indicated dosesand thenimmediatelylabeled for2 hr with

[35S]methionine. Normalizedaliquots of labeledcellextractswere

incubated with eitherPAb421, ap53-specificmonoclonalantibody,

orPAb419, anegativecontrol. (B) C127cellswereplatedatalow

density, irradiated with various doses of UV light, incubated in

growthmedium for 1week,andscored forcolonyformation.

in response to some types of DNA

damage.

Monoclonal

antibody PAb246,

which

recognizes

the

wild-type

murine

p53

protein

(18), reactedwith

virtually

all the

p53

presentinthese cells before and after UV treatment.

Antibody PAb240,

whichrecognizesamutantor

denaturation-specific epitope,

didnot reactafter UV treatment, andso noevidence forsuch a conformational

change

of

p53

protein

could be demon-strated in this system.

mdm-2 Protein Increases in Response to UV Irradiation. Sincethe mdm-2 geneis

regulated

by

the

p53 protein (12, 13)

and themdm-2

protein

is knowntobindtothep53

protein

and inhibitp53

transcriptional

transactivation

(13, 14),

itwasof interest to determine whether mdm-2 levels increased in responseto UVtreatment. The

steady-state

levels of both

p53

and

mdm-2,

aswellasthe

proportion

of each boundin

complex

withone

another,

weremeasured

by

Western blot

analysis.

C127 cellsweretreated witheitheralow

(3.8

J/m2,

94%

colony-forming ability)

or a

high (19 J/m2,

16%

colony-forming ability)

UVdose.The cellswereharvestedatvarious times afterUV treatment,

lysed,

andincubated with eithera

p53-specific antibody

(PAb421)

or an antiserum

against

mdm-2. The

immunoprecipitates

were

electrophoresed

and blottedontoa

membrane,

thetophalfof whichwasincubated withthe anti-mdm-2serumandthebottom half withPAb421. WhenC127 cellsweretreated withthe lowdoseof

UV,

the levelof

p53 protein

rose within 1.5

hr,

and mdm-2

protein

levels increased and

peaked by

2.5 hr

following

treatment

(Fig. 2A).

Both

p53

and mdm-2

proteins

increased about3-to

4-foldin responsetothe lowdose ofUV

(Fig. 2A). However,

the

proportion

of

p53

bound to mdm-2

(20-30%)

did not

change

significantly

with

time,

andat notimewas a

majority

ofthe

p53

boundtomdm-2.

When C127cells weretreated with the

high

doseof

UV,

p53

levelsrose more

dramatically:

atthe lowdose ofUV

(3.8

J/m2),

the maximalincreasein

p53

levelswasabout

3-fold,

whereasatthe

high

dose

(19

J/m2),

p53

levels increased

by

asmuchas60-foldat8hr. The

steady-state

levels

(Fig. 2B)

reflecta

larger

increasein

p53

than theresults with

labeling

with

[35S]methionine

(Fig. 1),

consistent withthe demonstra-tion

by

Maltzman and

Czyzyk

(5)

that the

p53

protein

is stabilized

by

UVtreatment. Theincreasein mdm-2

protein

wasalsogreateratthe

high

doseand

peaked

at13-fold10 hr after UVtreatment

(Fig. 2B). Thus,

the doseofUVaffects both the

timing

andthe levelof induction of mdm-2.

Com-parison

ofthe levelsof

p53

and mdm-2at4 hraftertreatment

clearly

showsthat theinduction ofmdm-2

expression

atthe

high

dose

lags

behind theincreasein

p53

levels.

Although p53

levelsare

higher

1.5hr aftertreatmentwitha

high

dose than after treatment with a low

dose,

the induction ofmdm-2

expression

is

delayed

afterthe

high

dose.

The Western blots

(Fig. 2)

quantitate

the

steady-state

levelsof

p53

and mdm-2 in cells treatedwithUV

light.

Similar

experiments

following

the

incorporation

of

[35S]methionine

into

p53

andmdm-2

proteins

in these cells demonstrateda

lag

inincreased

expression

ofmdm-2

protein following

stabili-zation of

p53

at

high

UVdoses. At

high

UV

doses,

p53

levels rose

rapidly (Fig. 1A),

but mdm-2

protein

levelsincreased

only

aftera

significant delay

(data

not

shown).

mdm-2mRNA Is Induced

by

UV

Light

ina

p53-Dependent

Manner. Northern blot

analysis

ofRNA from C127 cells treated with low and

high

doses of UV and harvested at various times after treatment showed a

good

correlation between increased mdm-2 message and

protein

levels

(data

not

shown).

At a low doseofUV

light (3.8

J/m2),

mdm-2 mRNAincreasedand

peaked

2hraftertreatment, whichwas

thetime atwhich the amountof

protein

also wasmaximal

(Fig. 2A).

At

higher

UV

doses,

the

peak

increase in mdm-2

protein

was

delayed (Fig.

2B)

and the

subsequent

increase in

mdm-2

protein

levelswas

accompanied by

increased mdm-2

mRNA

(data

not

shown).

This UV-stimulated increase in 421

Proc. Natl.Acad. Sci. USA 90(1993) 11625 A ImmunoprecipitatingAb: 419 421 amdm2 7 Blotting Ab: |amdm2 ..~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~.... 421 no no 1 2.5 4 6 10 no 1 2.5 4 6 10 UV UV B UV Hours following treatment Immunoprecipitating Ab: 419 421 amdm2

ml1

-i A Blotting Ab: amdm2 , . ... ... ...:.i:S.:.... ...."''...S'1'' ... i BSt ,~~~~~~~~~~~~~~~~~~~~~~~~~~~~...i *,...:.-;... T= IMI~ ~ ~~~421 no no 1. 5 4 6 8 10 24 no 1.5 4 6 8 10 24 UV UV UV

Hours following treatment

FIG.2. Levels ofp53and mdm-2changeas afunction of dose. (A)C127cellsweretreated with a UV doseof 3.8J/m2,harvestedatthe indicatedtimesfollowing treatment,andincubated witheitherPAb419(negative control),PAb421(specific forp53),or mdm-2antiserum (right sideofblot).Theimmunoprecipitates wereelectrophoresedin apolyacrylamidegelandblotted onto amembrane,thetophalfofwhich was incubated with themdm-2 antiserum while the bottomhalfwasincubated withPAW21.(B)C127 cells wereexposedto 19J/m2andanalyzed asinA.

both mdm-2 message andproteinwas notobserved in cells lacking functionalp53. Neither (10)3 cells (16), whichlack p53 protein, nor (12)1 cells containing T antigen, which expresswild-typep53 proteinin the presence of the simian virus40largeTantigen, induceexpression ofmdm-2 follow-ing UV treatment. Therefore, the induction ofmdm-2

ex-pression appears tobe dependentupon the transcriptional transactivation activity of

p53

protein.

p53

Transcriptional Activity

in

Ceils

Treatedwith UVLight. Next, thetranscriptional transactivation

activity

of endoge-nousmurine

p53

wasmeasured aftertreatmentofC127cells with UV light. A construct containing two copies of the

consensus

p53

DNA binding site from the murine muscle-specificcreatine kinase

(MCK)

promoterlinkedto areporter

constructcontainingtheCAT gene(15)wasinduced 39-fold aftertreatmentofC127cells with 40

J/m2

(Fig.

3A)whereas

asimilarconstructlackingtheMCK sequences

(1634CAT)

was notinduced more than2-fold. Sinceexpression of the mdm-2 gene is regulated by

p53

through a

p53-binding

element in thefirst intron of the gene(13), three fragments fromthefirst intron of the mdm-2gene were tested for UV inducibility. Constructs containing the entire intron (COSX1CAT)or an85-bp fragmentincludingthetwo imper-fect p53 consensus binding sites (BP100CAT) (13) were

induced3.3- and8.4-foldbyaUVdoseof40J/m2 (Fig.3B). Athird construct, containing400bp from theintron exclud-ing thep53bindingsites(HXO.SCAT),was not UVinducible (<10% increase in CAT activity) (Fig. 3B). Although the 85-bpfragmentconferred UVinducibilityonthe CATgene in C127 cells, it did not do so in (10)3 cells, which lackp53 protein (<2-fold induction at 40 J/m2; data not shown). Because this UV response element is also ap53response element(13)and it isnotinduced in cellslacking p53 protein, theseexperimentsindicate that mdm-2 isap53-inducibleUV response gene.

mdm-2 Expression Precedes Normal EntryofCells into S Phase.Undersomecircumstances,overexpression ofthep53 proteincausescellsto arrestin theG1phaseof the cellcycle Cell

Biology: Perry

etal.

50-2 CAT 1634 CAT

10 0 .0

_._{*_n} ...

M.

M-0 0 3.8 3.8 40 40 0 0 3.8 3.8 40 40

COSX1 CAT BPI00CAT HXO.5 CAT

.. .>4.It . 0 4 0

0 0 3.8 3.8 40 40 0 0 3.8 3.8 40

FIG.3. p53transcriptional transactivation activity is increased inresponsetoUVlight.(A) C127cellsweretransfectedwithaDNAconstruct

containing the CATreportergenefusedtoaminimalpromoter(1634CAT)ortothep53responseelementfrom the murineMCKpromoter

(50-2CAT).Twenty-four hours after transfection,cellsweretreated withahighor alowdoseof UVlight and after another 24hr,thecells were

harvested andlysed. Equalamountsofcellextractswereassayed for CAT activity. (B) Three fragments from the first intron of the mdm-2gene weretested for UVinducibility in C127 cells by the protocol described in A; COSX1CAT contains the entire intron, BP100CATcontains 85

bp that includetwoimperfectp53consensusbinding sites, and HXO.5CATcontains 400 bp from the first intronexcludingthep53bindingsites. (9, 19). While the function of mdm-2 is not understood,

overexpression oftheproteinhasbeen showntoinhibitp53 transcriptional transactivation activityintransient

transfec-tionexperiments (13, 14). Ithasbeenproposedthat

mdm-2-mediated inhibition of p53 transcriptional transactivation activitymayleadtoprogression ofcellsfromtheG1phaseto theS phase ofthecycle (20). Kastanetal.(10)have shown

that, after yirradiation,p53blocks cells in G1. It ispossible

that mdm-2 functions to overcome this G1 block. To see

whetherasimilar blockoccursinUV-treated cells, cellswere

labeled with thethymidine analogue BrdUrd and analyzedfor

No uv 4hr 0 0 c 0 0 0 - .-0 ie E z

BrdUrdcontentbyfluorescence-activatedcytometryusinga

fluorescein-conjugated monoclonal antibody specific for Brd-Urd. Cellswerealso stained withpropidium iodide, which

reflects the total amount of DNA in each cell, whereas

BrdUrd incorporation reflects nucleotide incorporation or

DNAsynthesisineachcell.Flowcytometryprofiles of C127

cells treated withamoderatelyhighUV dose of 14J/m2 (Fig.

4) showed aclear suppression in BrdUrd incorporation in

early,mid,andlate Sphase by4hr afterUVtreatment.The percentage ofUV-treated cells in G1, S, and G2 remained fairly constant (propidium iodide). Six to 8 hr after UV

Shr 12hr

q6 rA

24hr

I

PI fluorescence

FIG.4. Expressionofmdm-2precedestheresumptionof normal rates ofDNAsynthesisfollowingUVtreatment.C127 cellstreatedwith

aUVdoseof14J/m2werelabeled with 10,uMBrdUrd for 30minattheindicated timesfollowingtreatment.Cellswereharvested,processed

forfluorescence-activatedcytometry,and stainedwithanti-BrdUrdconjugatedtofluoresceinisothiocyanate (FITC) and counterstainedwith

propidiumiodide(PI).For each timepoint,the BrdUrdprofileis shownabovethepropidiumiodideprofile. Fifty thousandcellswereanalyzed

for each timepoint.

A UV Dose: (J/M2) B UVDose: (J/M2) .A

Proc. Natl. Acad. Sci. USA 90

(1993)

11627 irradiation, mdm-2 expression was maximal (data not

shown), and at 8 hr there was a pronounced increase in the incorporation of BrdUrd into cells with a

Gl-phase

contentof DNA(cells in early S phase) as seen by a recovery ofhigh BrdUrd fluorescence (Fig. 4). In fact, there was a greater percentage of cells in early S phase 8 hr after UV treatment than in untreated asynchronous cells. There was also a shoulder on the

G,

peak of cells stained for DNA content (propidium iodide), indicating an increase in the number of cells in early to mid S phase at 8 hr. This trend continuedas

thiswaveof cells proceeded through S phase (Fig. 4). Thus, during the 6- to 8-hr lag before the resumption of normal rates of DNA synthesis, either there must be an accumulation of cells in G1 which synchronously enter S phase 8 hr following UV treatment or, alternatively, the progressionofcells from G1 to S phase must be accelerated by some signals provided prior to 8 hr following treatment. The waveof cells moving through S phase with normal rates ofDNA synthesis might well have resulted from p53 and mdm-2interactions, since (10)3 cells, which lack p53 protein, also exhibited a decrease in DNA synthesis following UV irradiation but failed to recover from this suppression with a wave ofcells incorporating normal levels of BrdUrd (Fig. 4B).The(10)3 cells also failed to induce comparable levels of mdm-2 message and protein after UV irradiation (data not shown).

To substantiate the correlation between recovery of nor-mal S-phase DNA synthesis with expression of mdm-2 pro-tein, C127 cellsweretreated with a UV dose of 3.8 J/m2. At this low dose, mdm-2 levels peak at 2-2.5 hr following treatment(Fig. 2A). Asignificant increase in cells with a

G,

DNAcontent incorporating a large amount of BrdUrd ap-peared2hrafterUV treatment (data not shown), so mdm-2 expressionprecedes the resumption of a normal rate of DNA synthesisfollowing UV treatment with either a high or low UVdose. Athighdoses, enhanced expression of mdm-2 and a return to normal S-phase levels of DNA synthesis were delayed when compared with the response at lower UV doses.

DISCUSSION

Irradiation of mammalian cells with UV light results in a transient inhibition of DNA and RNA synthesis with a coordinate induction of expression of several immediate early genes such as c-jun and c-fos that in turn appear to regulate expression of other UV response genes (1). It has beenproposedthat this transient decrease in DNA and RNA synthesis following UV treatment reflects the time during which the UV-induced damage is repaired before the cells continue through the cycle (1, 21). This decrease could be brought about byacheckpoint control mechanism(21), and thep53tumor-suppressor geneproduct, which accumulates inresponse to DNAdamage (5-7, 10), acts as acell cycle checkpointin the response to yirradiation (9).

Thewild-typep53 proteinplays a role inblocking progres-sion through the G1 phaseof thecellcycle afteryirradiation, resulting in a complete inhibition of DNA synthesis in S-phase cells following treatment(9, 10). Incontrast, after UVirradiation,there is a decreased rate of DNA synthesis both in cells containing wild-type p53 protein and in cells lacking

p53

protein. p53 levels increase within 1-2 hrafter UV irradiation and continue toaccumulate when cells are treatedwithahigh doseof UV. Thehighlevels ofp53 protein, upto60-foldnormallevels,mightbeexpectedtoblockcells fromprogressingoutof G1 into S phase (19, 22), but there is no directevidence foraG1 blockintheseexperiments. The

transcriptional transactivation activity of p53 isdramatically increased following UV treatment, and one target of p53 function in the UV response is the mdm-2 gene, which contains ap53 response element in the first intron (13). At low doses of UV irradiation, theresponsiveness ofthis element is notapparent in a transient transfection assay (Fig. 3B), but theinduction of mdm-2 occurs rapidly (see Fig. 2A) and is dependent on the presence ofp53, since it does not occur in cells expressing simian virus 40 large T antigen (data not shown) or in cells lackingp53 protein. At higher doses, the synthesis of mdm-2 mRNA and protein are delayed several hours after the stimulation of high p53 protein levels in response to irradiation. Thus, at high UV doses other factors, inaddition top53 protein, are required for mdm-2 induction and these factors are rate-limiting for mdm-2 expression. Presumably, mdm-2 synthesis is delayed to permit repair synthesis to occur before normal rates of scheduled DNA synthesis can resume, and this delay is lengthened with increased amounts of DNA damage.

We aregrateful to Dr. T.Silhavyforthe useof hisUV light source and to Dr. J. Momand for stimulating discussions duringthe early stages of this work. We thank Dr. J. Chenforadvice on Western blotting and Drs. C. Finlay and X. Wuforhelpfuland supportive discussions. We acknowledge Dr. X. Wu forcritically readingthe manuscript.

1. Holbrook, N. J. & Fornace, A. J., Jr. (1991) New Biol. 3, 825-833.

2. Devary, Y., Gottlieb, R. A., Lau, L. F. & Karin, M. (1991) Mol. Cell. Biol. 11, 2804-2811.

3. Fields, S. & Jang, S. K. (1990)Science 249, 1046-1049. 4. Raycroft, L., Wu, H. & Lozano, G. (1990) Science 249,

1049-1051.

5. Maltzman, W. &Czyzyk, L. (1984) Mol. Cell. Biol. 4, 1689-1694.

6. Fritsche, M., Haessler, C. & Brander, G.(1993) Oncogene 8, 307-318.

7. Hall, P. A., McKee, P. H., Menage, H. duP., Dover, R. & Lane, D. P. (1993) Oncogene 8, 203-207.

8. Zhan, Q., Carrier, F. & Fornace,A. J., Jr. (1993)Mol. Cell. Biol. 13, 4242-4250.

9. Kuerbitz, S. J., Plunkett, B. S., Walsh, V. W. & Kastan, M. B.(1992) Proc.Natl. Acad. Sci. USA 89, 7491-7495. 10. Kastan, M. B., Zhan, Q.,El-Deiry, W. S.,Carrier,F.,Jacks,

T.,Walsh, W. V., Plunkett, B. S., Vogelstein, B. &Fornace, A. J., Jr. (1992) Cell 71,587-597.

11. Fornace,A. J., Jr.,Alamo,I., Jr., &Hollander,M. C.(1988) Proc.Natl.Acad. Sci. USA 85,8800-8804.

12. Barak, Y., Juven, T.,Haffner,R. &Oren, M. (1993)EMBOJ. 12,461-468.

13. Wu, X., Bayle, J. H., Olson, D. & Levine, A. J. (1993) Genes Dev. 7,1126-1132.

14. Momand, J., Zambetti, G. P., Olson, D. C., George, D. & Levine, A. J. (1992)Cell69, 1237-1245.

15. Zambetti, G. P., Bargonetti, J., Walker, K., Prives, C. & Levine, A. J. (1992) Genes Dev. 6, 1143-1152.

16. Harvey, D. & Levine, A. J. (1991) Genes Dev. 5,2375-2385. 17. Sherley, J. L. (1991) J. Biol. Chem. 266,24815-24828. 18. Gannon, J. V., Greaves, R., Iggo, R. & Lane, D. P. (1990)

EMBO J.9, 1595-1602.

19. Martinez, J.,Georgoff, I., Martinez, J. & Levine, A. J. (1991) Genes Dev. 5, 151-159.

20. Olson, D., Marechal, V.,Momand,J., Chen, J.,Romocki, C. &Levine, A. J. (1993) Oncogene8,2353-2360.

21. Brown, M., Garvik, B., Hartwell, L., Kadyk, L., Seeley, T. & Weinert,T.(1992) in The Cell Cycle, eds. Beach, D.,Stiliman, B. &Watson, J. D. (ColdSpringHarbor Lab. Press,Plainview, NY), Vol. 56, pp. 359-365.

22. Michalovitz, D., Halevy, 0. & Oren, M. (1990) Cell 62, 671-680.