Vol. 1, 535-542, November 1990 Cell Growth & Differentiation 535

Received 5/2 1/90., To whiini r(’quests for reprints should be addressed.

Expression of Nuclear Retinoic Acid Receptors inWild-type and Mutant Embryonal CarcinomaPCC4.azalR Cells

C. Nervi, T. M. Vollberg, J. F. Grii,po, D. A. Lucas,M. D. George, M. I. Sherman, K. �hudo, andA. M. Jetten1

Cell Biology Group, Laboratory of Pulmonary Pathobiology, NationalInstitute of Environmental Health Sciences, ResearchTriangle Park, North Carolina 27709 [C. N., T. M. V., M. D. G., A. M. J.J;Department of Toxicology and Pathology [J. F. G., D. A. L.] andDepartment of Oncology and Virology [M. I. SI, Roche ResearchCenter, Nutley, New Jersey 071 10; and Faculty of PharmaceuticalSciences, University of Tokyo, Tokyo, Japan [K. S.J

Abstract

Retinoic ack� (RA) induces differentiation of murineembryonal :arcinoma PCC4.azalR cells. In this study,the expres�ion of nuclear retinoic acid receptors (RARs)in �‘CC4.iza1R cells is examined. Analyses of [3H]RA-labeled nuclear extracts prepared from PCC4.azalR

ells by size-exclusion high-performance liquidchromatography demonstrated the presence of aspecific RA-binding activity that migrated with amolecular weight of approximately 50,000. More than95% of this binding activity was associated with thenuclear fraction. In contrast to cytosolic retinoic acid-binding protein, the RARs bound RA analogues of theCh-series very effectively. Northern blot analyses oftotal RNA with complementary DNA probes specificfor RARor, RARfl, and RAR#{176}yshowed that PCC4.azalRcells contain predominantly transcripts encoding RARcrand RAR”y; RARfI transcripts were undetectable.Treatment of PCC4.azalR cells with RA increased thelevels of RARI9 mRNA in a dose- and time-dependentmanner. The RA concentration for half-maximuminduction of RARfl mRNA was 1 n�i. An increase inRAR$ mRNA was detectable as early as 2 h after theaddition of RA. This increase was not abrogated bycycloheximide, suggesting that protein synthesis is notrequired for this response. The ability of severalretinoids to increase RAR�9 mRNA levels in PCC4.azalRcells correlated well with their binding affinity to theRARs but not with their binding affinity to cytosolicretinoic acid-binding protein. Two mutant cell lines,PCC4(RA)-1 and (RA)-2, which do not undergodifferentiation after RA treatment, contained levels ofRAR-binding activity very similar to those of theparental cells. PCC4(RA)-1 and PCC4(RA)-2 cellsexpressed RARa and RAR”y transcripts identical in sizeto that of PCC4.azalR cells. However, PCC4(RA)-1cells expressed only one RAR”y transcript of 3.1kilobases. The loss of the expression of the 3.3-kilobaseRAR’y transcript in PCC4(RA)-1 cells may be causallyrelated to their inability to differentiate in response to

RA. RA treatment increased the levels of RAR$ mRNAin both PCC4(RA)-1 and (RA)-2 cells, indicating thatthese mutant cell lines have not totally lost theirresponsiveness to RA.

Introduction

Embryonal carcinoma cells are tumorigenic, multipoten-tial stem cells that are closely related to totipotent em-bryonic stem cells (1, 2). A variety of different chemicalagents have been identified that can induce differentia-tion of these cells in culture (3). Retinoids, the naturallyoccurring and synthetic analogues of vitamin A, areamong the most effective inducers of differentiation ofseveral embryonal carcinoma cell lines (3-7). Our labo-ratories have been concentrating on the regulation ofdifferentiation in embryonal carcinoma PCC4.azal R cells(7, 8). These cells can undergo a multistage process ofdifferentiation (reviewed in Ref. 9). RA2 treatment ofPCC.azal R cells causes their differentiation into mesen-chymal stem cells, which can then further differentiateinto preadipocytes after treatment with 5-azacytidine ina manner similar to that reported for 1OT1/2 cells (10).At confluence, and in the presence of dexamethasoneand insulin, these cells terminally differentiate into adi-pocytes as indicated by the synthesis of fat cell-specificmarkers (9, 11). The differentiation of PCC.azalR cellsinto mesenchymal cells is accompanied by changes incellular morphology and the induction of several bio-chemical and molecular markers. For example, cells startto express increased levels of epidermal growth factorreceptor (1 2), the extracellular matrix proteins fibronectinand collagen I and the protooncogenes int-2 and c-los(9, 1 3). Moreover, in contrast to embryonal carcinomacells, the differentiated cells become permissive to viralgene expression (14, 15). At least some of the changesinvolved in this differentiation process appear to beregulated at the transcriptional level.3

The molecular mechanism by which retinoids inducedifferentiation in embryonal carcinoma cells has yet tobe established. Recently, the existence of several nuclearretinoic acid receptors, designated RARa, RARI3, andRAR’y, has been demonstrated (16-20). These receptorsare members of the steroid hormone/thyroid hormonesuperfamily and act as RA-dependent transcriptional fac-tors. It is likely that at least some of the alterations ingene expression during differentiation of PCC4.azalRcells are controlled by these nuclear RA receptors.

2 The abbreviations used are: RA, retinoic acid; RAR, retinoic acid recep-

tor; CRABP, cytosolic retinoic acid-binding protein; HPLC, high-perform-ance liquid chromatography; kb, kilobase(s); bp, base pair(s); poly(A)’,polyadenylated; GPDH, glyceraldehyde-3-phosphate dehydrogenase;

SDS, sodium dodecyl sulfate; SSC, standard saline citrate; cDNA, com-plementary DNA.

3 T. M. Vollberg. C. Nervi, J. F. Grippo, and A. M. Ietten, unpublishedobservations.

�o 15 20 25 30 35 40 45

Retention time ( mm

Fig. 1. HPLC analysis of [3H]RA binding in cytosolic and nuclear extractsprepared from embryonal carcinoma PCC4.azal R cells. Cytosolic (A) andnuclear (B and C) extracts were incubated with 5 nxi [3H]RA for 18 h at

4#{176}Cin the absence (#{149})or the presence of a 200-fold molar excess ofunlabeled RA (0) or Ch55 (A). The samples were fractionated over aSuperose 12 HR 10/30 size exclusion column at 4#{176}C(A and B) or at room

temperature (C), and radioactivity in each fraction was determined as

described in #{176}Materials and Methods.#{176} Solid arrows, position of RAR;open arrows, position of CRABP.

536 Retinoic Acid Receptors in Embryonal Carcinoma Cells

In this study, we examine the expression ofthe nuclearRA receptors RARa, RAR/3, and RAR�y in embryonal car-cinoma PCC4.azal R cells in relation to the induction ofdifferentiation and in two mutant cell lines, PCC4(RA)-1and PCC4(RA)-2, which are unable to differentiate afterthe addition of RA (21, 22).

ResultsAnalysis of [3H]RA Binding. Cytosolic and nuclear ex-tracts prepared from PCC4.azal R cells were incubatedwith 5 nM [3H]RA in the absence or presence of unla-beled RA or its analogue Ch55 (1 tiM), and the binding of[3H]RA was analyzed via size exclusion HPLC (Fig. 1).The HPLC profiles of the cytosolic extracts showed onesingle major peak of specific binding activity that elutedat 32 mm (Fig. 1A) and corresponded to a molecularweight of approximately 16,000, consistent with CRABP.This radioactivity was displaced by unlabeled RA but notby unlabeled Ch55, in agreement with the previouslyreported binding specificity of CRABP (23). The HPLCprofile of the nuclear extracts showed five peaks ofradioactivity (Fig. 18). The three minor peaks eluting at16, 41, and 46 mm represented nonspecific bindingactivity since there was no competition in the presenceof a 200-fold excess of unlabeled retinoids, whereas thetwo major peaks eluting at 32 and 27 mm representedspecific binding. Unlabeled RA, but not Ch55, was ableto compete with [3H]RA binding in the peak eluting at32 mm, the same position as CRABP. This binding activityprobably represents CRABP. Most of the CRABP activityin the nuclear extract appears to be due to a contami-nation of the nuclear preparation with the cytosolic frac-tion rather than reflecting a true association of CRABPwith nuclei since two more washings of the nuclei re-duced CRABP levels further to 2-5% of total CRABP butcaused only a slight decrease in the RAR levels (data notshown). Both unlabeled RA and Ch55 abolished thebinding of [3H]RA in the peak eluting at 27 mm. Theelution time of these proteins corresponds to a molecularweight of approximately 50,000. We have reported pre-viously that the RARs elute in this fraction and can bindCh55 (24); therefore, this RA binding probably representsRAR-binding activity. This RAR receptor activity was notdetectable in the cytosolic extract, suggesting that thisreceptor is predominantly associated with the nucleus.When HPLC analysis was carried out at room tempera-ture instead of at 4#{176}C,[3H]RA binding to CRABP wastotally abolished, whereas the binding to RAR was onlyreduced by about 50% (Fig. 1C).

RAR mRNA Expression. The expression of the RARa,RAR�3, and RAR’y genes in PCC4.azalR cells was exam-med by Northern blot analyses using total RNA isolatedfrom these cells and 32P-Iabeled inserts of hRARaO,hRARI3O, and mRAR’yO as probes (Fig. 2). The hRARaprobe hybridized to two RNA species of approximately2.6 and 3.5 kb, whereas the mRAR’y probe hybridized totwo RNAs of 3.3 and 3.1 kb, of which the latter transcriptwas predominantly expressed in PCC4.azal R cells. Tak-ing into account the specific activity and the length ofthe radiolabeled probes and the length of the exposureof the autoradiogram, it was calculated that PCC4.azal Rcells contained about 4-fold higher levels of RAR’ymRNAs than RARa mRNAs. Transcripts hybridizing tothe hRAR$ probe were undetectable.

We next analyzed the effect of RA on RAR expression.

‘0

‘C

0)C)

I.-.0

E0.

V

RA (1 �LM) induced a small (about 3-fold) increase in thelevels of RARa mRNA over a 48-h time period (Fig. 2).RAR#{176}ymRNA levels remained relatively unchanged dur-ing the first 24 h of RA treatment but decreased at longertimes of exposure to RA. RAR-y mRNA expression wasreduced by 60% 48 h after the addition of RA (Fig. 2).No significant changes in the levels of glyceraldehyde-3-phosphate dehydrogenase mRNA were observed afterthe addition of RA. Addition of 1 �zM RA to PCC4.azal Rcells induced two RAR$ mRNAs, 3.2 and 2.9 kb in size.The larger transcript was the most prominent. This in-duction occurred in a time-dependent manner (Fig. 2).An increase in RAR/3 mRNA could be observed within 2h of RA addition. The induction of RARL� mRNA wasdependent on the RA concentration (Fig. 3). The RA

0 2 4 8 16 24 48hr

Kb- 4.4- 28S

Retinoic Acid (M)

0 iO12i01 � 10�10 ��10 8 io � io

�RAR(3

RARf� 1� �

A

B

E

E‘C

E

0

- 2.4

- 18S

- 4.4- 28S

-24

I � ._‘ \oriI)iri) 111(r) .11)115 ‘.1’. ei Iii i’\I)rl’iOii iii hi’ Ii1i( l,’ar rc’tini)i(

.ii iii rr.i (1)thr� k’\k’�, k’\k’i, ,iiul R.\R.-� � iiiltrt,itt’(l tiriti’ (1 hi intl R\tr.itid l’( ( -liii IR i ills. ( (‘II’.ss iii tri’,ited s� tb 1 psi Ri\ for hi’

iiili .itiil liii, iiit’rs.iI�. I oi,iI R”�\’ � 1() pg) i’(il,it(’(l riini l�((�4.,i,,i 1 R

( II’. ss,ii’ ire tl�i1,ltl(1, lhtti’l ci Ns tr,in, .iiid hs l)ridi/ed ii 1P-l,il’lt’il

1)1)l)� t�r R’,k’, k’�\k’.�, .inI k’\k’-, .i� (l,’’.i ril),’(f in .\1.iti’ri,il’. md islet))( )d’.. Ill )t” 55 ‘ri ‘,triI)I)(’i( � hs l)ri(ll/e(l t ) ‘�l’- I,il)(’It(l � �\l )-2t1. ss l)i(Iii’!’u flih’-. gis i (‘U.iI(l(’ll\ ii,’- tpho�jih.it. (l,’I)\ ilrigi’il.i’a’ (( .11)1 I I. ,\titor,i

(Ill )l.�1,iti)’. 55 i’ll’ “�t�’ i’.t-il s’, itli k’ill),ii’l(ii)i. ‘.1 11’)’))’ it -‘() C icr -1 il,is�k’�\k’�, lit, � 1.i�’. �R�\I5’.� lht. ..i (1.155 (ls’��\R’, llot, it))1 I (l,i\ (l’l)t I

0 -12 -11 -10 -9 -8 -7 ‘6

( i’ll ( ,r� )�‘. th .�s, I )itIt’rt’nti.iti in 537

RARa

RARy #{149}....#{149}.�

- 18S

- 4.4- 28S

- 2.4

GPDH � . . 14 18S

( ()11( (nt rat on for 1111t-nla\i Ill u ri� sti n�ulat iOIt ‘tV,�1S aI)out 1

riM. (\( lOhe”.Iflhlde did HOt aI)olisll the RA-induced in-

creds(’ In Rt\R11 niRNA, indicating that the iIldtJ(tiOIl ot

RARd riiRNA l(’�’#{128}’lSI)v RA �viS riOt (IeJ)en(letlt on protein

S\’flth(SlS ( Fig. 4). In t,1( I, PCC4.azi 1 R cells treated si-

fllllII,Ifl(’Ot.iSl’, ��ith ( V( lOh(’).ilTlide sticl RA contained con-SISt(’fltl\ higher I(’t,/(�lS Ut RAR�3 riiRNA than cells treated

with RA (Il()I1(’. Nuclear run-on exl)erinlerlts ‘,vere carriedOUt to (let(’rnhIrle \vheth(’r the (‘nh,lncecl expression oftll(’ RAR3 grle was tiue to increaserl transcription. Inthree Ifl(l(’I)eIl(I(Ilt (‘xI)(’rirll(’rlts, ho (letectable transcrip-tIOfl (1) the RAR1f g’r�’ ( otild lx observed in either

untr(’,lt(’d ( (‘(IS or Cells tre�1te(l br 5 or 24 Ii �vith 1 �i,vi

RA ((hlta 001 ShO\Vfl). Flowever, the sanie exp(’rinl(’nts

sho�vt’d all ifl( r(’as(’ In the rat(’ ol transcription of the

GPDH

log [Retinoic acid], M

II).,’ 1 ln(lli( hurl iii Rt\k’� i’\liri’,’iiiil Ii) l’((.l .1/i I R i i’lI� ,i� ,t liii iii

(it th� R\ (itoh i’ntr,itein. liit,il k’”.\ ss.i’. Is()I,lt,il ruin ( ‘II, tr,’,ii,’il liii

(l,)s.,, ssith thi’ Iil(li( ,itt’il ( (‘iii ii)ti,itiiin 0) R\. \i)It(’i(’iii hurt .iri.ils ‘.t’’.

l)(’rtirnl(’d using 11) cii o1 mlii R”...-\ is ili’s( rili,il iii h� lig,’nul I ) 1 i�.

,\.,ititnr,i(lIi)grdni�: H, (l(’il’iiiinhitri( s ,iililiilt. il ,iiiti)r,i(lIii�r,iiii��

homeobox gene IIox-1.6, another gene induced rapidlyafter RA addition. � Examination of the stability of theRAR� mRNA is shown in Fig. 5. Densitonietric analysesof the RNA stability indicated that the rate ot the (le-crease in RAR�1 mRNA in cells treated for 48 h with 1 pM

RA was very similar to that in cells treated (or 5 Ii withRA (Fig. 5B). Since the level of RAR� niRNA increased

about 5-fold between 5 and 48 h of RA treatment,

differences in stability of this niRNA cannot a((ount (orthe total increase in RAR�3 niRNA expression. The half-life of the RAR/.� mRNA in untreated cells could not hedetermined since the RAR�3 niRNA was undetectable.

Structure-Activity Relationship. The abi lity of several

RA analogues to bind to nuclear RAR receptors iii

PCC4.azal R cells was analyzed and compared with their

binding to CRABP and their ability to iH( reas&’ RAR�

mRNA levels. A strong qualitative correlation was foundbetween the ability of the analogues to ( ompete with

RAR�3

106M RA

2gg/ml Cycloheximide

A

RAR�3

HOURS TREATED

5 48

- + - +

0.5 4.0 0.5 4.0 0.5 4.0 0.5 4.00 1.5 0 1.5 0 1.5 0 1.5

“p

B

z

Eso

0a

EI)

1.0

0.8

0.6

0.4

0.2

of

‘-S

2 40

Actinomycin D Treatment (h)

fig. 5. Effect of RA on the stability of RAR� mRNA. Cells were treated

first with 1 MM RA for 5 or 48 h and then incubated further in the presenceof 250 ng/ml actinomycin D. RA-treated and control cultures werecollected for RNA isolation at 0, 0.5, 1 .5, and 4.0 h after the addition of

actinomycin D. A, poly)A(� RNA was isolated, and 5 gg were analyzedvia Northern analysis. B, a slot blot was prepared using: 2.5, 0.5, and 0.1

� of each RNA sample and then hybridized to a ‘2P-labeled probe forRAR4. The blot was then stripped and hybridized with “P-labeled pGAD-28. Autoradiograms of slot blot and Northern blot were scanned with aBio-Rad 620 densitometer, and the level of remaining RAR4 RNA relativeto GDPH was calculated. Statistical evaluation was performed utilizing atwo-way analysis of variance followed by Tukey’s test using the SASprograni (SAS Institute, Cary, NC). #{149},5 h RA treatment; 0, 48 h RA

treatment. #{176},P < 0.05 versus same time point generated for the 48-h RA-treated samples. Bars, SEM.

538 Retinoic Acid Receptors in Embryonal Carcinoma Cells

HOURS TREATED

4 7

- + - + - + - +- - + + - - + +

106M RA

250mg/mi Actinomycin 0(fIR)

GPDHGPDH

fig. 4. Effect of cycloheximide 01) the induction of RARII mRNA.PCC4.azalR cells were treated with or without RA (1 MM) and/or cyclo-

heximide (2 �g/mI( for 4 and 7 h. Total RNA isolated from these cells wasanalyzed via Northern t)IOts using a “P-labeled insert of RAR4O as aprobe. One h of ( ycloheximide (2 �g/ml( treatn�’nt inhil)ited proteinsynthesis in PCC4.azalR ( s’lls by greater than 90% and RNA synthesis by

less than 20%. Cy loheximide inhibited RNA synthesis less than 40% at

4 h ot treatment. Cydoheximide did not cause any cytotoxicity in theC (‘115 during the tinie of tre,itment.

[3H]RA for binding to the RAR present in PCC4.azalRcells and their capacity to induce differentiation andincrease RAR/3 mRNA levels (Fig. 6). The analogues Ch55,Ch80, Ro-13-7410, and Ro-40-6055 were effective in-ducers of RAR/3 mRNA levels in PCC4.azal R cells andthe best competitors for [3H]RA binding to the RARreceptors (Fig. 6). The analogues Ch20, Ch30, and Ch40caused minimal induction of RAR1� mRNA. In agreementwith previous studies (23), the analogues of the Ch-seriesdid not bind to CRABP. Retinol was not very active inPCC4.azalR cells and did not bind effectively to eitherRAR or CRABP (Fig. 6). The potency of these retinoids toincrease RARI� mRNA and to bind to RARs correlatedwell with their ability to induce differentiation inPCC4.azal R cells (data not shown).

Mutant Cell Lines. The binding of [3H]RA was alsoanalyzed with cytosolic and nuclear extracts of two RA-nonresponsive mutant cell lines, PCC4(RA)-1 andPCC4(RA)-2 (21). Cytosolic extracts of PCC4(RA)-1 cellscontained CRABP at levels about 30% of that of parentalPCC4.azal R cells (Fig. 7A). Cytosolic extracts ofPCC4(RA)-2 cells did not contain any CRABP (Fig. 7B) inagreement with previous findings (21). The HPLC profileof nuclear extracts from PCC4(RA)-1 and PCC4(RA)-2cells (Fig. 7, C and D) were very similar to that fromPCC4.azal R cells except that CRABP was absent in thenuclear extract profile of PCC4(RA)-2 cells. The amountof total [3H]RA-binding activity to RAR receptor(s) wasreproducibly very similar in the three cell lines. Northernblot analyses (Fig. 8A) showed that PCC4(RA)-2 cellscontained two RARa and two RAR�y transcripts of similarsize as those in parental cells. PCC4(RA)-1 cells containedtwo RARa transcripts of 2.6 and 3.5 kb but expressedonly the smaller 3.1-kb RAR-y transcript and apparentlyhad lost expression of the 3.3-kb transcript. The level ofRARa expression in the three cell lines was very similar.RA induced an almost 2-fold increase in RARa mRNAlevels in PCC4(RA)-1 cells and a 4-fold increase inPCC4(RA)-2 cells, compared to a 3-fold stimulation in the

parental cells. RAR�y mRNA levels were 2-3-fold lowerin the mutant cell lines as compared to PCC4.azal R cells.

Although PCC4(RA)-1 and PCC4(RA)-2 cells do notundergo morphological differentiation after the additionof RA, these cells have not totally lost their responsive-ness to RA. As Fig. 8 indicates, similar to PCC4.azal Rcells, RA treatment increased the level of RARL� mRNAin both PCC4(RA)-1 and PCC4(RA)-2 cells. The relativeincrease was most pronounced in PCC4(RA)-2 cells.

Discussion

Previous studies have shown that RA induces differentia-tion of PCC4.azal R cells into mesenchymal stem cells (5,

9). This differentiation is characterized by morphological,biochemical, and molecular changes. During the induc-tion of this pathway of differentiation, the levels of severalmRNAs are increased, including the homeobox genes

VC

2 100

I‘I,

C0

0)

E0

0

6 A

b2

a)r�o 0

02-.,.

EV

RAR/3

0

RARCRABP �

I ig. #{244}. (�i)r)l1),irlsi in (it tI)t’ I)ln(lii)g capa it� ut sever,il r’tir�iids 1(1 RAR.113(1 CRAIIP with their aI)ilit5 to indli(e RAR4I nRNA. To exan#{236}ine the( iirllF)(’tlti(iI) iii rs’tiniiiiis t ir the I)ln(llng (ii I ‘F I I RA to thi’ (Rt\BP ,iI)(l theill)) tar Rt\ Ill i’ptiir iii P( C4.,iLa 1 R � (‘us, s’xtr,it Is sv’re iI) uh,ted svith �

nM I ‘ I I I Ri\ in t he .1 1)5(1)) i’ or F)r(’s()( I’ 0 ,l 200- 1)11(1 I�i iI.ir (‘�(t’S5 (it I li(’in(ll( ,iti’(l O’til)iiid. The [ ‘I IIRt\ I)In(ling 55.15 .iIl,)l’,/(’(l .is (l(’s(rilx’d in the

k’gt’ni I t � F ig. 1 . L f)J)( ‘r jianel. I he )(‘rcent.)gi’ (1 iI)lpt’titii in iii [ II IRA

l)in(liI)g 55 .15 1 ,iI( uI,iti’il ii))) )liitt(’(l ( snl,�1 !),)rs Rt\R ( i)n)petitlon: 11.10 (ndhits, (� Rr\IIP ( i itii1n’tit ii ii) ). .\ i�ddI pini’I, the (‘tt(’( I (if I li(’ S.ifll(’ retiniiitls

( 1(1 i)M ) i ii) I l)i’ iI)(fli( ti( ii) ( )t RARd (‘\f)r(’ssi(in 55,15 (l(’t(’rnhir)(’(l by Nortl�’rn1)10) ,in.ils ss using 3(1 pg (ii t(itdl RNA. .-\n,ils sis iii (;PDI I niRNA (‘sort’s

“ii ))1 55 ,iS US(’(l ,i� .i ( iintr( ii � lots er f).iu) ‘I.

O

� -�

‘-3 10 20 30 40 50 10 20 30 40 50

Retention time ( mm

I ig. 7. H PLC analysis of [ ‘H I RA l)in(ling in ( ytosolic md i�i tsr (‘xtra( Is

prepared from PCC4(RA)-1 and (RA(-2 ( ells. Cytosoli (-\ ,ind lfl intlnuclear (C and U) extra( Is were ,tnalyzed as des( rihed in the l(’gen(l ii

Fig. 1 . A and C, extracts Iron) PCC4( RA)- 1 ; B and I ). (‘xtra( Is triim

PCC4( RA(-2. #{149},[3HIRA only; 0, [ ‘1-1 1RA an(l 200-told (‘x( ess ut unl,iht’lt’dRA; A, [‘H]RA and 200-told excess ot Ch55. �“nIvl ,irrnws, )i)sitiiin ofRAR. Open arrows, position ot CRABP.

Cell Crowth S Differentiation 539

GPDH #{149}.Ss#{149}�#{149}s..

ilox-1.3 and -1.6, the oncogene int-2, epidermal growthfactor receptor, fibronectin, and collagen I (13).� We areinterested in understanding the mechanisms by whichRA induces differentiation and regulates the expressionof the various genes in PCC4.azal R cells.

Previous observations have revealed that PCC4.azalRcells contain relatively high levels of the specific cytosolicRA-binding protein CRABP (7, 2 1 , 22). In initial studiesexaniining structure-activity relationships, a rather good

correlation was found between the efficacy with whichretinoids bind to CRABP and their ability to induce dif-ferentiation ill PCC4.azal R cells (7). These experimentssuggested that CRABP was involved in this induction of

differentiation by retinoids. The isolation of RA-nonre-spOnsiVe mutant (elI lines containing low levels of CRABPsupported the concept that CRABP was essential in theniechanisni of action of RA (2 1 ). Later studies usingretinoids of the Ch-series showed that these analogueswere very effective inducers of differentiation in em-bryonal carcinoma F9 cells but were Llnable to bind toCRABP (23). The study presented here extends these

findings to PCC4.aza 1 R cells. These observations appearto indicate that CRABP may not be essential for inductionof (lifferentiation in enibryonal carcinonia cells by reti-

4

1k ______

C

noids. However, CRABP might still particil)ate in con-trolling gene expression in an indirect nianner l)y regu-

lating the concentration of retinoids in the cell (25), eitherthrough binding to the retinoid and/or by facilitating

retinoid metabolism (26).Recently, several RARs have been identified that func-

(ion as RA-dependent transcriptional factors ( 16- 20).PCC4.azal R cells contain RA receptors of a molecularweight of approximately 50,000 which are associated

almost solely (more than 95%) with the nucleus. Thesereceptors probably represent the RARs. Analysis of tll(’

structure-activity relationship of several retinoids showsa good correlation between their binding to RARs andtheir ability to induce RAR�3 niRNA and differentiation.

These results suggest that the induction of differentiationin PCC4.azalR cells by RA is niediated by RARs.

PCC4.azalR cells contain transcripts encoding RAR�and RAR-y, whereas RAR�1 transcripts were undetectal)l(’.PCC4.azalR cells expressed two RARS and two RAR-ymRNA species. RA induced a 3-fold increase in RARo

and, after 48 h of treatment, decreased RAR’� levels by60%. A decrease in RAR’y mRNA has also been observedin embryonal carcinoma F9 cells after treatnient with RA

and cyclic AMP analogues (27). Different isotorms of eachRAR have been reported which differ in their NH..-terminal A-region and are generated by alternative splic-ing (20, 28). Conceivably, the two transcripts expressed

by PCC4.azalR cells represent RAR�-1 and RAR-y-2.RAR/3 mRNA levels were induced after RA treatment ina time- and dose-dependent manner. Recently, RA-in-duced increases in RAR� mRNA were reported in othercell systems (27, 29, 30). Such an increase appeared tobe regulated at least in part via a transcriptionally con-trolled niechanism (27, 29). An RA-responsive (‘lementhas been identified in the RAR�3 gene (31), supporting

the concept that this gene is directly controlled by anRA:RAR complex. Our stu(lies suggest that the increasein RAR�1 mRNA in PCC4.azal R cells occurs iridepend-

A BPc��

PA) �RA�2�

0 4 24 0 4 .�4

RAR(,

RAR.) � �

RAR�

GPDH �

C

� A RAP

z

�

.‘ ‘Is

[a’,

II)� I) ( ))l)p.ir)s)))l 0) h’ ‘\pr,’ssIo)) 0) R\R. k’\R3, intl R-\R-, niR\-\

l,,\’l’. n F’( ( -liii] k’, I’( ( .1�(�’\: 1 � .)))(l l’( ( -tlk’\L2 ( i’ll’,. tot,il R’s-\ ) (I)

pt�)� ls0l,iti’(l 1(1)11 lflltri’,it,’(l ( (‘(Is .l))(l I)))))) ( �‘((� tri’,it’il or 4 ,)))(l 2-4 I)55 ith 1 � k� .\ ss ‘rt’ rn tSH),it,1l, tr.in-ttri’d (1 \\ r,i)i. 10)1 hs Irliied I))

�..t’ l,il),�l,-)l )r()I)t’� (II k’\k’, Ri #{149}\k’.�(�‘‘),f�’.. , ,i))il ( 01)1 I. ��\\ S,i)))1)li’’,

))) I))) ‘.I( (1 ( It 1),- 0)11tH)) ( i’ll litii’s � ,‘r, ‘ i ‘(i’( (1 )J)h( ri ‘soil I)) .i(lJ,ii ont .ini’s

)( I the’ si N .\ s.i)))))li’’. rI 0) h, Ri \ tiiiii ‘ ( I lti1S�’ ‘In iss ii In t Ii). 2 . Nnrthi’rn

.1)1.11’, SI’ of )lltlt.l)lt intl ).iri’))t.ul k’N\ ss,i� I ,irried nit .11 hi’ 5,10))’ tititi’.

\, ,llit(ll.i(lU)gI.l)1l’. ot “-.i>rth,’rn hInt .in.ils si’’ it tiit,il Fs’”-.\ lsnl,it(’(l triitit

)i))trt’.it�’(I .11(1 (5i�\ t i, ‘,it,’l iiut.nt I I II’, /� hill 1)1(111 1)41,11)15 tr, I))) I ig’.. 2

.1)1(1 tc� � “� 11)111(1 ssith ., bi,>�k’.iil �2() ll,’rlsltlnli’ti’r, .iti,l hi’ r,iti,, It

(�‘��\R ni1�’”-..\ 1) (,l�l )( I � � 1)l()ttl’(l.

540 Ri’tuion r\i ii R’ (‘htors I)) I )))I)r\ I ))).il ( ,ir( )))o)1),i ( ills

.- PAR

� H�

(‘ntl\� ot prot(’ri synthesis. Nuclear run-oti eXF)erinieflts

�v(’r(’ IIi(Oi1(IUSiV(’ Si1i((’ ‘�V(’ tail(’d to (((‘((ct synthesis

ol RARd niRNt\ in either untr(’,st(’(I or RA-treatedPCC4.aia 1 R � (‘1k, J)rol)ahlv (Itie to �‘ers’ lo�v rates ot

trans( ription of this gene in these ( ells. A siniilar obser-

v,lt I( )n has l)e’tl ni,ide in F9 eni l)ry( )fidl ( dI( itiOlita ((‘[IS

I 2 7 I. F \,fI1ilIi,ft l( )Ii of I lie SI,ll)ilitv of I he RAR1I ni RNAshowed no increase in the half-life of the RAR� mRNAwith longer tiriies of RA exposure, suggesting that the

IIi( reast’ in RAR/� niRNA in PCC4.azal R cells was prol)-

ably not due to enhanced niRNA stability.

The pr(’s(’n( (‘ of flit.iIti[)I(’ RARs in PCC4.,sza 1 R ((‘[Israises the (iU(’stii)Ii of their roles. Recent stiffiR’s l)y[s[�’s’t Ii t’( ,1/. ( �2 ) hav(’ r(’v(’ak’d that, in F9 (‘nil)ryonal( ar( iIi( riia ( t’IIs, I Ii(’ ovs’rt’xprt’ssion of a V(’( br contain-

rig ,i trurl( dt(’(I RARo ([)NA \�‘hi(h lacks the ligand-l)iridlrig (britain ahrogat(’s the In(Iu( tion of several, hutriot all, RA-iridu 1)10 genes. These ol)servations Sugg(’St

that (‘,H Ii of the RARs niav ( ontrol s[R’(ific s(’ts of genes

,mnl(I that hoth Rt\Ro (IS �\(‘II as RARd and/or RAR-� niavh(’ lIiV( )IV(’d in regulating genie e\pressiori during differ-

(‘rltI,ttIOn 01 F�) ((‘115. RARd niav play a role in ( ontrolling

I I1( I ranis� ri )tI( )fi of certain g(’n(’s at a Iat(’r stage in the

(JIItcfl’ntlatlon )ro((’ss. This is ( ertainlv in agreeniienit withthe WI(IeIy held vie�’ that differentiation occurs as a

( ,iS( ade of transcriptional arid translational events. Insu( Ii 1i scheme, then, following RA tr(’atnient, RARx arid!

01 RAR-y would m()(Iulate th(’ expression of a set of genes

including other transcription-regulatory factors, such asRAR/1 and certain homeobox genes, which would in turnalter the expression of other groups of genies.

We have examined the levels of RAR mRNA in twomutant embryonal carcinoma cell lines, PCC4(RA)-1 andPCC4(RA)-2, which fail to differentiate in response to RA.However, these cell lines have not totally lost their re-sponsiveness to RA, since RAR�3 mRNA is still inducedafter RA addition. In view of the likely role of RARs inmediating RA-induced differentiation in the parentalPCC4.azal R cells, the inability of the mutant cell lines todifferentiate could be related to a defect in RAR expres-sion. We have found that, in contrast to parental cellswhich contain two RARy mRNAs of 3.1 and 3.3 kb,PCC4(RA)-1 cells express only one RAR-y mRNA species

of 3.1 kb and appear to have lost the expression of the3.3-kb mRNA. The lack of expression of the 3.3-kb RAR-ymRNA, which may represent RARy-1, in PCC4(RA)-1cells may be related to the inability of these cells toundergo differentiation. If this is true, one would expectthat the introduction of this RAR-y into PCC4(RA)-1 cellswould enable the cells to undergo differentiation afterRA addition. Experiments are under way to test thishypothesis. No changes in RAR expression were notice-able in PCC4(RA)-2 cells. These cells might produce adefective RAR protein that is unable to undergo dimeri-zation or to interact correctly with DNA and/or otherregulatory proteins. Alternatively, the altered responsive-ness in PCC4(RA)-2 cells could be unrelated to a defectin the expression or function of one of the RARs hut

could be related to altered expression or function ofanother gene, possibly a target gene for RA activation orsuppression. Paradoxically, this cell line has a greatlyreduced level of CRABP, which appears not to be re-quired in induction of differentiation by at least someretinoids (23). It is conceivable that reduced CRABPlevels are not causative but indicative of this differentia-tion-defective phenotype; e.g. , CRABP might be underthe control of one or more RARs; if so, reduced expres-sion of CRABP might be due to a defective RAR species.Recently, Wei et a!. (33) have shown that in several

embryonal carcinoma cells, CRABP is up-regulated byRA, supporting the concept that CRABP is either directlyor indirectly under the control of RARs. Further charac-terization of RAR/CRABP expression in PCC4.azal R cellsand analyses of the changes in gene expression (luringdifferentiation of these cells will be necessary to deter-mine the role of each RAR isoform and CRABP in thisdifferentiation process.

Materials and Methods

Cell Culture and Differentiation. Characterization of theembryonal carcinonia cell line PCC4.azalR and its mu-(ant derivatives, PCC4(RA)-1 and PCC4(RA)-2, has beendescribed previously (21). The cells were grown in RPMI1640 medium supplemented with 10% fetal calf serum,penicillin (100 units/mI), arid streptomycin (100 pg/mI).All-trans-RA, RA analogues Ro 13-7410 and Ro 40-6055,and all-trans-retinol were synthesized by Hoffniann-LaRoche. The analogues of the Ch-series were synthesizedby K. Shudo (34). Medium, serum, and antibiotics were

purchased from GIBCO, Grand Island, NY. AII-rans-[H]RA (50 Ci/niniiol) was obtained froni NEN-Duponit, Bos-toni, MA. Induction of differentiation hy retinoids was

quantitated liy the aggregate assay as descril)e(l prey-

Cell Growth & Differentiation 541

ously (5). The retinoids were dissolved in dimethyl sulf-oxide. Controls received solvent only (0. 1% final concen-tration).

Preparation of Cytosolic and Nuclear Extracts. Thepreparation of nuclear and cytosolic embryonal carci-noma cell extracts has been described previously (24,35). Briefly, approximately 5 X iO� cells were washedtwice in phosphate-buffered salt solution containing 2mM EDTA, trypsinized, and collected by centrifugation.The cells were then washed gently at 4#{176}Cin 10 ml PTGbuffer consisting of 5 mM sodium phosphate (pH 7.4), 10mM thioglycerol, 10% glycerol, and the following pro-tease inhibitors: phenylmethylsulfonyl fluoride (1 mM),aprotinin (10 units/mI), and leupeptin (10 units/mI). Cellswere homogenized in 5 ml PTG buffer with a Douncehomogenizer (pestle B, 60-80 strokes). The homogenatewas centrifuged at 4#{176}Cfor 30 mm at 1,000 X g. Thepellet (2 ml), containing the loosely packed nuclei, waswashed twice with the same buffer. To reduce the con-tamination of the nuclear preparation with cytosol, nucleimay be washed an additional two times. The supernatantfractions were combined and centrifuged for 30 mm at130,000 x g at 4#{176}C.The resulting supernatant was des-ignated as cytosolic extract. The nuclear pellet was ex-tracted in 8 ml TTGK buffer containing 10 mrvt Tris-HCI(pH 8.5), 1.5 mM EDTA, 10 m�i thioglycerol, 10% glyc-erol, 0.8 M KCI, and the same protease inhibitors as inthe PTG buffer. The suspension was incubated for 1 hon ice with repeated resuspension every 10 mm and thencentrifuged at 130,000 x g for 30 mm. The resultingsupernatant is referred to as the nuclear extract.

Retinoic Acid Binding Assay. In routine assays, cyto-solic or nuclear extracts (0.3 ml) were added to 1.5 mlEppendorf tubes containing [‘HIRA to give a final con-centration of 5 nM. Nonspecific binding was measuredin the presence of a 200-fold excess of unlabeled RA.After 3 h incubation at 4#{176}C,extracts were transferred to

Eppendorf tubes containing dextran-charcoal pellets oh-tamed after centrifugation of 30 �l of a charcoal-dextransuspension (3% acid-washed Norit A, 0.3% dextran C in10 mM Tris-HCI, pH 7.4, and 0.02% sodium azide). Afterresuspension of the pellet, the mixture was incubated for10 mm at 4�C and then centrifuged for 15 mm at 15,000x g. The supernatant was analyzed for flH}RA-hindingactivity by size exclusion HPLC (24). HPLC analysis wasperformed at 4 or 25#{176}Cwith a Gilson liquid chromatog-raphy system. Extracts (200 pI) were fractionated over aSuperose 12 HR 10/30 size-exclusion column (Pharma-cia) at a flow rate of 0.5 mI/mm using PTG buffer con-taming 0.4 M KCI as eluent. Radioactivity was monitoredwith a Flo-One/Beta A-200 radioflow detector (Radio-matic Instruments, Tampa, FL). In certain instances, frac-(ions (0.5 ml) were collected, and radioactivity was de-termined with a scintillation counter. Binding data werecalculated as the area under the curve using the trape-zoidal rule or using the software provided by RadiomaticInstruments.

Plasmid DNA. The RAR probes, hRARaO, hRARI3O,and mRAR-yO were obtained from P. Chambon, Stras-bourg, France. hRARxO contains the full length codingsequence (1770 bp) for the human RARa protein clonedinto the EcoRl site of the expression vector PSG5 (16,18). hRARi�0 contains the human RAR(3 open readingframe sequence cloned into the EcoRl and BamHl sitesof the expression vector PSG5 (18). The clone mRAR-yOcontains a 1 .9-kb fragment of the murine RAR”y sequence

cloned into the EcoRl site of PSG5 (36). For Northernblot analyses, the full length hRARx sequence and hRAR�3and mRAR”y inserts were used as probes. The recombi-nant cDNA clone pGAD-28, which contains a 1261-hpcDNA insert of chicken glyceraldehyde-3-phosphate de-hydrogenase (37), was digested with PsO, and an approx-imately 1 1 20-hp fragment containing the coding regionplus some 5’- and 3’-untranslated sequences was iso-lated and used as a probe. The cDNA probes werelabeled with [a-32P]dCTP (3000 Ci/mmol; Amersham

Corp., Arlington Heights, IL) by random priming usingthe kit and the protocols supplied by Boehringer Mann-heim Biochemicals (Indianapolis, IN).

RNA Isolation and Northern Analysis. Total RNA from3-5 x 10” embryonal carcinoma cells was isolated by theguanidinium isothiocyanate method (38). Poly(A)� RNAwas selected by oligo(dT)-cellulose chromatography (39).Thirty pg total RNA or 5 pg poly(A)� RNA were fraction-ated by electrophoresis on a 1 .2% agarose/0.66 M form-aldehyde slab gel in the presence of ethidium bromideand transferred to Nytran (Schleicher and Schuell, Keene,NH) membranes on a Vacublot apparatus (AmericanBionetics, Hayward, CA) according to the manufacturer’ssuggestion. Transferred RNA was cross-linked to the blotby UV irradiation (40, 41). Blots were prehybridized in50% formamide, 1% SDS, 5x saline-sodium phosphate-EDTA buffer, lx Denhardt’s solution, and 100 pg/mIdenatured salmon sperm DNA at 42#{176}C.Hybridizationswere carried out with fl2PJdCTP-labeled RARaO, RAR/30,mRAR�y, and pGAD-28 probes (10” dpm/mI, specificactivity, 5-7 x l0� dpm/pg DNA) overnight at 42#{176}Cusingthe prehybridization buffer described above. Blots werewashed twice in 2x SSC-0.l% SDS for 15 mm at 25#{176}Cand once in 0.5x SSC-0.l% SDS for 30 mm at 55#{176}C.Blotsprobed for RARa and RAR’y were washed additionally in0.lx SSC-0.l% SDS for 15 mm at 65#{176}C.These hyhridi-

zation conditions do not allow any cross-hybridizationbetween the RAR probes.

In Vitro Transcription. Nuclei were isolated and ‘�P-labeled RNA transcripts were prepared by a modificationof the procedure of Greenberg and Ziff (42) as described

previously (43).

Acknowledgments

We thank Dr. P. Chambon for providing the RARa, RARIi, and RAR’�plasmids. We thank Drs. N. Saunders, S. Bernaki, and K. Korach for theircomments on the manuscript.

References

1. Sherman, M. I., and Solter, D. (eds.(. Teratomas and [)ifferentiation.New York: Academic Press, 1975.

2. Papaioannou, V. E., and Rossant, J. Effects of the embryonic environ-ment on proliferation and differentiation of embryonal carcinoma cells.

CancerSurv., 2: 165-183, 1983.

3. Jetten, A. M. Induction of differentiation of eml)ryonal carcinoma ( ells

by retinoids. In: M. I. Sherman ed.), Retinoids and Cell Differentiation,pp. 105-136. Boca Raton, FL: CRC Press, 1986.

4. Strickland, S., and Mahdavi, V. The induction of differentiation in

teratocarcinoma stem cells by retinoic acid. Cell, 1 5: 393 403, 1978.

5. Jetten, A. M., letten, M. E. R., and Sherman, M. I. Stimulation ofdifferentiation of several murine embryonal carcinoma (elI lines by reti-noic acid. Exp. Cell Res., 124: 381-391, 1979.

6. Jones-Villeneuve, E., McBurney, M., Rogers, K., and Kalnins, V. Reti-notc acid induces embryonal carcinoma cells to differentiate mm neurons

and glial cells. I. Cell Biol., 94: 253-262, 1982.

7. Jetten, A. M., and letten, M. E. R. Possible role of retinoic a (1 bindingprotein in retinoid stimulation of embryonal carcinon)a ( (‘II difterenria.

542 Retinoic Acid Receptors in Embryonal Carcinoma Cells

tion. Nature (Lond.), 278: 180-182, 1979.

8. Sherman, M. I., Paternoster, M. L., and Taketo, M. Effects of arotinoids

upon murine embryonal carcinoma cells. Cancer Res., 43: 4283-4290,1983.

9. Jetten, A. M. Regulation of gene expression by retinoic acid: embryonalcarcinoma cell differentiation. In: P. B. Fisher led.), Mechanisms of Dif-ferentiation, Vol. 1, pp. 49-74. Boca Raton, FL: CRC Press, 1990.

10. Taylor, S. M., and Jones, P. A. Multiple new phenotypes induced in

1OT1/2 and 3T3 cells treated with 5-azacytidine. I. Cell. Physiol., 1 1 1:187-194, 1979.

11. Chapman, A. B., Knight, D. M., Dieckmann, B. S., and Ringold, G.M. Analysis of gene expression during differentiation of adipogenic cellsin culture and hormonal control of the developmental program. J. Biol.

Chem., 259: 15548-15555, 1984.

1 2. Jetten, A. M. Retinoids specifically enhance the number of epidermalgrowth factor receptors. Nature (Lond.), 284: 626-629, 1980.

1 3. Smith, R., Peters, G., and Dickson, C. Multiple RNA’s expressed fromthe int-2 gene in mouse embryonal carcinoma cell lines encode a protein

with homology to fibroblast growth factors. EMBO J., 7: 1013-1022,1988.

14. Segal, S., and Khoury, G. Differentiation as a requirement for simianvirus 40 gene expression in F9 embryonal carcinoma cells. Proc. NatI.Acad. Sci. USA, 78: 5754-5758, 1979.

15. Loh, T. P., Sievert, L. L., and Scott, R. �V. Proviral sequences that

restrict retroviral expression in mouse embryonal carcinoma cells. Mol.

Cell. Biol., 7: 3775-3784, 1987.

16. Petkovich, M., Brand, N., Krust, A., and Chambon, P. A humanretinoic acid receptor which belongs to the family of nuclear receptors.

Nature (Lond.), 330: 444-450, 1987.

17. Giguere, V., Ong, E. S., Segui, P., and Evans, R. M. Identification ofa receptor for the morphogen retinoic acid. Nature (Lond.), 330: 624-

629, 1987.

18. Brand, N., Petkovich, M., Krust, A., Chambon, P., de The, H.,Marchio, A., Tiollais, P., and Dejean, A. Identification of a second retinoicacid receptor. Nature )Lond.), 332: 850-853, 1988.

19. Benbrook, D., Lernhardt, E., and Pfahl, M. A new retinoic acid

receptor identified from a hepatocellular carcinoma. Nature (Lond.), 333:669-671, 1988.

20. Krust, A., Kastner, P., Petkovich, M., Zelent, A., and Chambon, P. A

third human retinoic acid receptor, hRAR--y. Proc. NatI. Acad. Sci. USA,86: 5310-5314, 1989.

21. Schindler, I., Matthaei, K. I., and Sherman, M. I. Isolation and char-

acterization of mouse mutant carcinoma cells which fail to differentiatein response to retinoic acid. Proc. NatI. Acad. Sci. USA, 78: 1077-1081,1981.

22. McCue, P. A., Matthaei, K. I., Taketo, M., and Sherman, M. I.Differentiation-defective mutants of mouse embryonal carcinoma cells:response to hexamethylenebisacetamide and retinoic acid. Dev. Biol.,96:416-426, 1983.

23. Jetten, A. M., Anderson, K., Deas, M. A., Kagechika, H., Lotan, R.,Rearick, J. I., and Shudo, K. New benzoic acid derivatives with retinoid

activity: lack of direct correlation between biological activity and bindingto cellular retinoic acid binding protein. Cancer Res., 47: 3523-3527,1987.

24. Nervi, c., Grippo, J. F., Sherman, M. I., George, M. D., and Jetten,A. M. Identification and characterization of nuclear retinoic acid-binding

activity in human myeloblastic leukemia HL-60 cells. Proc. NatI. Acad.

Sci. USA, 86: 5854-5858, 1989.

25. Maden, M., Ong, D. E., and Chytil, F. Spatial distribution of cellular

protein binding to retinoic acid in the chick limb bud. Nature (Lond.),335: 733-735, 1988.

26. Gubler, M. L., and Sherman, M. I. Metabolism of retinoids by em-bryonal carcinoma cells. J. Biol. Chem., 260: 9552-9558, 1985.

27. Hu, L., and Gudas, L. J. Cyclic AMP analogs and retinoic acidinfluence the expression of retinoic acid receptors a, �, and ‘ymRNAs inF9 teratocarcinoma cells. Mol. Cell. Biol., 10: 391-396, 1990.

28. Kastner, P., Krust, A., Mendelsohn, C., Gamier, J. M., Zelent, A.,Leroy, P., and Chambon, P. Murine forms of retinoic acid receptor �y with

specific patterns of expression. Proc. NatI. Acad. Sci. USA, 87: 2700-2704, 1990.

29. de The, H., Marchio, A., Tiollais, P., and Dejean, A. Differentialexpression and ligand regulation of the retinoic acid receptor a and figenes. EMBO J., 8: 429-433, 1989.

30. Song, S., and Siu, C-H. Retinoic acid regulation of the expression ofretinoic acid receptors in wild-type and mutant embryonal carcinomacells. FEBS Lett., 256: 51-54, 1989.

31. de The, H., Vivanco-Ruiz, M., Tiollais, P., Stunnenberg, H., andDejean, A. Identification of a retinoic acid responsive element in theretinoic acid receptor j3 gene. Nature (Lond.), 343: 177-180, 1990.

32. Espeseth, A. S., Murphy, S. P., and Linney, E. Retinoic acid receptor

expression vector inhibits differentiation of F9 embryonal carcinoma cells.Genes & Dev., 3: 1647-1656, 1989.

33. Wei, L-N., Blaner, W. S., DeWitt Goodman, S., and Nguyen-Huu,

M. Regulation of the cellular retinoic acid binding proteins and theirmessenger ribonucleic acids during P1 9 embryonal carcinoma cell differ-

entiation induced by retinoic acid. Mol. Endocrinol., 3: 454-463, 1989.

34. Hashimoto, Y., Kagechika, H., Kawachi, E., and Shudo, K. Specificuptake of retinoids into human promyelocytic leukemia cells HL6O by

retinoid-specific binding protein: possibly the true retinoid receptor. Jpn.

J. Cancer Res., 79: 473-483, 1988.

35. Jetten, A. M., Grippo, J. F., and Nervi, C. Isolation and bindingcharacteristics of nuclear retinoic acid receptors. Methods Enzymol.,1990, in press.

36. Zelent, A., Krust, A., Petkovich, M., Kastner, P. H., and Chambon, P.Cloning of murine a and $ retinoic acid receptors and a novel receptor

_ypredominantly expressed in skin. Nature (Lond.), 339: 714-717, 1989.

37. Dugaiczyk, A. D., Haron, I. A., Stone, E. M., Dennison, 0. E.,Rothblum, K. N., and Schwartz, R. J. Cloning and sequencing of aribonucleic acid copy of glyceraldehyde-3-phosphate dehydrogenasemessenger ribonucleic acid isolated from chicken muscle. Biochemistry,

22: 1605-1613, 1983.

38. Maniatis, T., Fritsch, E. F., and Sambrook, J. Molecular Cloning: A

Laboratory Manual. Cold Spring Harbor, NY: Cold Spring Harbor Labo-ratory, 1979.

39. Aviv, H., and Leder, P. Purification of biologically active globinmessenger RNA by chromatography on oligothymidylic acid-cellulose.Proc. NatI. Acad. Sci. USA, 69: 1408-1412, 1972.

40. Davies, L. G., Dibner, M. D., and Battey, J. F. Basic Methods inMolecular Biology, pp. 143, New York: Elsevier, 1986.

41. Church, G. M., and Gilbert, W. Genomic sequencing. Proc. NatI.Acad. Sci. USA, 81: 1991-1995, 1984.

42. Greenberg, M. E., and Ziff, E. B. Stimulation of 3T3 cells inducestranscription of the c-los proto-oncogene. Nature (Lond.), 3 1 1: 433-438,1984.

43. Hosler, B. A., LaRosa, G. I., Grippo, J. F., and Gudas, L. I. Expression

of Rex-i, a gene containing zinc finger motifs, is rapidly reduced byretinoic acid in F9 teratocarcinoma cells. Mol. Cell. Biol., 9: 5623-5629,1989.