J. Biol. Chem.-1992-Sadowski-4722-31

8/8/2019 J. Biol. Chem.-1992-Sadowski-4722-31

1/10

THE OURNALF BIOLOGICALHEMISTRY0 1992 by The American Society for Biochemistry and Molecular Biology, Inc. Vol. 266, No. 7, Issue of March 5, pp. 4722-4731,1992Printed in U. .A.

Gene Expression During 3T3-L1 Adipocyte DifferentiationCHARACTERIZATION OF INITIALRESPONSESTO THE INDUCING AGENTS AND CHANGES DURINGCOMMITMENT TO DIFFERENTIATION*

(Received for publication, September 26, 1991)

Henry B. Sadowski$$, Thomas T. WheelerQ, nd Donald A. YoungnFrom the E. H enry K eutma nn aboratory, Endocrine-M etabolism Un it, Departments f Medicine, Biochemistry,ann Radiat ionBiology and Biophysics, and the Enviro nmen tal H ealthSciences Center, Universityof Rochester Medical Center,Rochester, New-York 14642

The mouse 3T3-Ll fibroblastic cell line rapidly dif-ferentiates to anadipocyte phenotype when post-con-fluent cells are treated for 48 h in fetal calf serum-containing medium supplemented with 1 M dexameth-asone (D),0.5 mM methylisobutylxanthine (M) and 10pg/ml insulin (I).D and act synergistically to ommitthe cells to differentiate 4-48 h after initiat ing trea t-ment, and this is blocked by the phorbol ester, 12 -0 -tetradecanoylphorbol- 13-acetate. In order to identifycellular proteinsnvolved in the differentiationrocesswe analyzed differentiating 3T3-Ll cells using two-dimensional electrophoresis on large format gels. Weobserved changes in over 300 proteins during differ-entiation (over 100within 5 h of initiating differentia-tion) and many of these are also changed at the levelof mRNA(by analysis of in vitro translation products).About 75% of the initial changes were maximally in-duced by treatment with a combination of M and I,while no more than 10 proteins and their correspond-ing mRNAs were maximally induced by D within 3.5h. Another 10 proteins were synergistically regulatedby the combination of all three agents (DMI) within3.5 h. Additional species were induced at later times.Five of these were synergistically induced by treat-ments that lead to differentiation, wereirst expressedat elevated levels during commitment and remainedelevated in fully differentiated adipocytes. One or moreof these proteins could well have a functional role inthe commitment to and/or expression of the adipocytedifferentiation program.

The study of myotube differentiation from the pluripotentstem cell line, C3HlOT1/2, has led to thediscovery of severalnovel protein factors, myoD, myogenin, herculin and myf-5,which seem to play a role in myotube differentiation (1).Although several cell culture models of adipocyte differentia-tion exist (e.g. the 3T3-L1, 3T3-F442A, and TA1 cell lines)the factors involved in determination to the adipocyte lineage,commitment to terminal differentiation, and evolution andmaintenance of the adipocyte phenotype are unknown. 3T3-* This work was supported by National Institutes of Health GrantsCA47650 and DK16177 and by Council for Tobacco Research U.S.A.Grant 1774. The costs of publication of th is article were defrayed inpart by the payment of page charges. This article must therefore behereby marked advertisement in accordance with 18U.S.C. Section1734 solely o indicate this fact.$ Present address: Cold Spring Harbor Laboratory, Cold SpringHarbor, N Y 11724.5 These two authors contributedequally, and theirorder should beregarded as interchangeable.V To whom correspondence should be addressed.

L1 preadipocytes are fibroblastic cells determined to theadi-pocyte lineage (2). Consistent, rapid (4-6 days), and uniform(85-95% of cells) conversion to adipocytes is obtained whenpost-confluent 3T3-Ll cells are treated for 48 h in fetal calfserum (FCS)l-containing medium supplemented with the glu-cocorticoid, dexamethasone (D); the cyclic AMP-phosphodi-esterase inhibitor,methylisobutylxanthine (M), and igh con-centrations of insulin (I), followed by a period of 3-6 days ofincubation in agent-free medium (3). Fully differentiated 3T3-L1 adipocytes possess most of the morphological,biochemical,and hormonal response characteristics of adipocytes in uiuo(2,4-6). Furthermore, they form a fat pad when injected intonude mice (7).The cDNAs of several adipocyte-specificproteins have beencloned (8-11).Analysis of the genomic structure andpromotersequences of these genes has resulted in the identification ofcis-regulatory sequences which maymediate their differentia-tion-dependent expression. Differentiation-dependent bind-ing of nuclear factors to some of these elements has beendemonstrated, and in several cases the promoter elementsconfer differentiation-dependent expression on a eportergene in cell culture (12-15). The promoters of several adipo-cyte-specific genes contain abinding site for the transcriptionfactor C/EBP, which is itself induced during adipocyte differ-entiation (16-18). It has recently been suggested that C/EBPmay also promote withdrawal from the cellcycle duringcommitment. However, expression of C/EBP alone is notsufficient to promote differentiation of 3T3-Ll cells in theabsence of inducing agents (19). Thus, he elucidation of theprotein factors involved in the expression of adipocyte-spe-cificgenes has so far not resulted in the identification ofmaster regulatory proteins, whose expression is alone suf-ficient to trigger commitment and/or to drive adipocyte dif-ferentiation.As an alternative approach to identifying proteins involvedin the differentiation process we used large format two-di-mensional electrophoresis to follow the synthesis of over 4000metabolically labeled proteins and 2000 in vitro translationproducts of cellular mRNA during differentiation. We soughtto identify proteins and mRNAs that are rapidly induced orrepressed in response to the nducing agents, as well as those

The abbreviations used are: FCS, fetal calf serum; D, dexameth-asone; M, 1-methyl-3-isobutylxanthine;, insulin; TPA, 12-0-tetra-decanoylphorbol-13-acetate;/EBP, CCAAT enhancer-binding pro-tein; DAP, differentiation-associated protein; PBS, phosphate-buff-ered saline; DMEM, DulbeccosmodifiedEaglesmedium; SDS,sodium dodecyl sulfate; Hepes, 4-(2-hydroxyethyl)-l-piperazineeth-anesulfonic acid; FCS, fetal calf serum.For simplicity, the word induction or repression is used hereto specify a change in abundance of a gene product without meaningto imply the mechanism underlying the change.4722


2/10

Gene Expressionuringdipocyteifferentiation 4723induced during commitment to differentiation and before theaccumulation of triglyceride. Addition of the inducing agentsresults in rapid changes in synthesis of more than 100 pro-teins, most of which are also changed at th e evel of mRNA(by analysis of in vitro translation products). Of these, ap-proximately 75% are maximally induced by treatment withM I while no more than 10 are maximally induced by D. Anadditional 10 are synergistically regulated by the colribinationof D, M, and I, which synergistically induces differentiation.One or more of these rapidly-induced proteins may mediatethe initial effects of the inducing agents in promoting terminaldifferentiation. Other proteins display a delayed synergisticinduction that still precedes the evolution of the adipocytephenotype. Analysis of these proteins has highlighted fivemRNA/protein species, which we call differentiation-associ-ated proteins (DAPs), that are candidate regulatory proteinsin theadipose differentiation program.

EXPERIMENTAL PROCEDURES

In Vitro Translation-Poly(A+) RNA (2.5 pg/BO-pl reaction) wastranslated using nuclease-treated rabbit reticulocyte lysate (PromegaBiotec, Madison WI) asdescribed by the manufacturer except that 1unit/ml RNAsin was included in the reaction mixture and the RNAwas preincubated for 10 min in 4 plof3.5 mM methylmercurichydroxide at 22 "C. After translat ion in the presence of 1 mCi/ml[35S]methionine (specific activity > 1000 Ci/mmol, Du Pont-NewEngland Nuclear), reaction mixtures were centrifuged for 15 min at100,000 X g in an airfuge (Beckman) a t 4 "C to sediment polysomes(23).35Sncorporated into in vitro translation products was measuredby scintillation counting after ncubation of a portion of the reactionmixture in alkaline solution, followed by richloroacetic acid precipi-tation and collection of the precipitate on glass fiber filters. Theremainder of the reaction mixture supernatant was lyophilized anddissolved in 0.1 ml of two-dimensional electrophoresis sample buffer(9.5 M urea, 2% (w/v) Nonidet P-40, 2% (w/v) ampholines, and 5%(v/v) 2-mercaptoethanol). Up to one-third of the reaction was sub-jected to two-dimensional electrophoresis.Metabolic Labeling-Cells were rinsed twice in methionine- andcystine-free medium and then incubated in medium containing 10%of the normal levels of these amino acids and combinations of theinducing agents. 300 .&i/ml [35S]methionine (Tran3'S-label, ICN,

Cell Culture and Differentiation-The 3T3-L1 preadipocytes specific&&ity 1000 Ci/mmol)'was added at various times thereafter.(kindly provided by Dr . M. D. Lane, Johns Hopkins University) were After incubation with 35S (normally 3 h) the cells were rinsed twicegrown in high glucose Dulbecco,s modified Eagle,s medium (DMEM) in ice-cold DMEM, and 150 11 of two-dimensional electrophoresiscontaining bovine calf serum (HyClone) with the medium sample buffer (9.5 M urea, 2% (w/v) Nonidet P-40,2% (w/v) ampho-changed every second day. Adipocyte differentiation was induced as Two-dimensional Electrophoresis-Procedures for the two-dimen-follOws' At days post-confluence (day the medium was changed sional separation of proteins on large format gels have been describedfor DMEM with lo% bovine Serum (FCS) in detail (24, 25). Briefly, %-labeled proteins were first separated bylines, and 5% (v/v) 2-mercaptoethanol) was added to lyse the cells.

(HyClone), 1 M dexamethasone (DL 0.5 m~ me th~ lisobu t~l xan thine isoelectric focusing in 3.3-mm x 30-cm rod gels containing 1.6% (w /(M),and 10 d m 1 nsulin (1). Thereafter the mediumwas replaced ampholytes, p~ 5-8 (pharmacia LKB Biotechnology I ~ ~ . )nd(Sigma) was diluted from a stock solution of16.2 mM in dimethyl the top of 32 x 36 x 0.075-cm, exponential polyacrylamidemedium.Assessment of Differentiation-Cell monolayers were rinsed twice Autoradiograms were carefully visually, and where ap-with PBS and fixed for at 22 " ith PBS containing 3'7% propriate the peak density of selected spots was quantified using aformaldehyde. Cells were stained for neutral lipid with Oil Red 0 microcomputer-~ased ensitometer (26). Densities were normalized

at 48-h intervals for DMEM supplemented with lo%CS TPA 0.4% ampholfles, pH 3.5-10. First dimension gels were attached to(stored at -70 O C) to a concentration lo o nM In slab gels. After electrophoresis, fixed, hi ed were sub-

jected to autoradiography for 6 X lo 7cpm loaded X days exposed.

(Sigma). After removal of stain, the m o n o b r s were rinsed briefly for differences in sensitivity ofdetection by scanning several invariantwith 60% isopropanol and several timeswith water. The Oil Red spots on each gel. Quantitative differences are expressed as the foldsta in was eluted with isopropanol and he absorbance of the extract change in peak density with he control value.measured at 510 nm (20).For determination of the percentage of lipid-containing cells, the RNA were to electrophoresis in 1.2% (w/v) agarose con-orthern Blot Analysis-Equivalent amounts (1 pg) of poly(A+)monolayers were tVPsinized and suspended cells were counted in taining 2.2 M formaldehyde (27) and transferred to nylon membranetrypan blue-containing l r d h m using a hemacytometer. A1iquotsOf (Duralon, Stratagene) by overnight capillary transfer in 10 X SSC (1the tmsiniz ed ells were replated, and after 2 hhe wells were insed x ssc is 0.15NaC1,0.015 sodium citrate). RNA was fixed totwice with PBS and the cells were fixed and stained as above. 500 the membrane by uv cross-linking (Stratalinker, Stratagene). Thecells/well were counted and scored for the absence or presence of Oil membrane was prehybridized for 1h at 42 oc in a solution containingRed 0-stainable lipid droplets. No significant difference in the plating 50% (v/v) deionized formamide, 2 x s s c , 1% w/v) SDS, 0.1% w/efficiency of adipocytes versus non-differentiated cells was observed. v) bovine Serum albumin, 0.1% o ~ y v ~ n y ~ p y r r o ~ ~ ~ o n e ,.1% i ~ ~ l l 4 0 0 ,DNA Synthesis-Triplicate cultures were labeled for 8 h in 2 pCi/ 50 mM sodium phosphate, p~ 6.5, and 150 pg/ml sonicated and

were fixed in 1 M citric acid, transferred to microcentrifuge tubes solution at 42 ec with 32p-labeled c-myc probe (human third exon;and left overnight in10% (w/v) trichloroacetic acid. Precipitated onco r) labeled by the random hexamer priming methodto a specificmaterial was collected on glass fiber filters, washed with ice-cold activity of greater han 109 cpm/pg DNA (28). The membrane wasethanol, and subjected to scintillation spectrometry. In some experi- washed in 1% (w/v) SDS in 2 X ssc at 65 "c for 60 min. ~ f t ~ments an alternative procedure was followed. After incubation with autoradiography, the membrane was reprobed with 32p-labeled 8-['Hlthymidine, the monolayers were rinsed twice with ice-cold PBS tubulin (human 40-mer, oncogeneciences) labeled by the 5' end-and subjected to three 10-min incubations with cold 10% trichloroa- labeling method.cetic acid, followed by three rinses with ethanol, and air-dried. The Subcellular Frationation-Cells were metabolically labeled withremaining material was solubilized in 0.3 M NaOH, neutralized with [35SImethionine or 4 h,1h after initiating treatmentwith FcS alone,HCl and subjected to scintillation spectrometry. DMI, DMI+ TPA, or 6 days after initiating DM1 treatment (fullyR N A Isolation-cells (2 15-cm diameter dishes/condition) were differentiated adipocytes) and subjected to the subcellular fractiona-guanidinium isothiocyanate, 0.5% (w/v) Sodium N-laUryl sarcosine, 1 mM EDTA, 3 mM MgC12,as well as protease (1pg/ml leupeptin, 11% v/v) 2-mercaptoethano1, and 25 mM trisodium citrate, PH 7.0. pg/ml aprotinin, 0.7 pg/ml pepstatin, and phenylmethylsulfonyl flu-Aftershearing the DNA, the RNA was pelleted through a Cs cl oride added to 1 mM before use) and phosphatase (20 mM sodiumcushion (21). RNA 0.4-1.0mg)was subjected to two successive fluoride, 1 mM sodium orthovanadate, and 5 mM sodium pyrophos-affinity chromatography purifications using a 50-mg column of phate) inhibitors, scraped off the plate, and lysed by 10 passes througholigo(dT)-cellulose (Type 3, Collaborative Research). The method of a 20-gauge needle. Nuclei and cell debris were pelleted by centrifu-Aviv and Leder (22) was modified as follows; RNA was bound to the gation at 1000 X g for 5 min, and the supernatantwas centrifuged atcolumn and washed in a solution containing 10 mM Tris.HC1, pH 100,000 X g for 60 min. The high speed supernatant (cytosolic7.5, 2 mM EDTA, 0.5% SDS, and 1 M LiC1, the bound RNA was fraction) was removed, and thepellet (crude membrane fraction) wasrinsed in a solution containing 10mM Tris. HCl, pH 7.5,2mM EDTA, briefly rinsed, then dissolved in two-dimensional electrophoresis sam-and 0.1 M LiC1, and the poly(A+)RNA was eluted with the above ple buffer. The low speed pellet was washed twice with a solutionsolution but containing no LiC1. The RNA concentration of a portion containing 10 mM Tris. HC1, pH 7.4, and 10 mM NaCl either with orof the eluate was determined by absorbance at 260 nm. without 0.5% (w/v) Nonidet P-40 to remove attached membrane

ml [3HlmethYlthYmidine (5.0 Ci/mmol, Amersham COrP.). The cells denatured salmon sperm DNA and incubated overnight in the Same

briefly rinsed in ice-cold PBS and ysed in a solution containing 4 M tion procedure. Cells were rinsed in 10 mM Hepes, pH 7.4, containing


3/10

4724 Genexpressionuringdipocyteifferentiation

B80

60

40

20

0C D I T r ea t men tD l M M I D M1 D M I + T P A

FIG. 1. Promotion of 3T3-Ll adipocyte differentiation: ef-fects of the inducing agents. A , phase contrast microscopy (mag-nification 140 X ) of monolayers of 3T3-Ll cells 6 daysafter headdition of agents. 2-day post-confluent cells were treated for 2 dayswitheither 10% FCS-containing medium alone (C), or with the

Time After Initiating Treatment (h )

rB ro o ,

0.24 h 23-48 h 48-72 hLabelling Period

C 3.5h 51.5 hC MI DM1 DM I+T C MI DM1 DMI+T

F I G . 2. DNA synthesis and c -myc expression during adipo-cyte differentiation. A , cells were diffe renti ated as escribed abovein th e presence ( D M I )or absence ( M I )of dexamethasone. Quadru-plicate dishes of cells were pulse-labeled for 8 h with ["Hlthymidinea t various points during differentiat ion. B , triplicate wells of cellswere incubated with ["Hlthymidine for 24 h a t various points duringdifferentiat ion. The incorporationf tritium into richloroacetic acid-precipitable m aterial was determined. Results represent he meanincorporation (cpm) f tandard error. C, cells were lysed 3.5 h and51.5 h (3.5 h after removal of induc ing age nts and the addition offreshedium) afternit iat ingreatm ent. 2 pg of twicepoly(A+)selected RNA was subjected to denaturing agarose gel elec-tropho resis, transferre d o nylon mem bran e, hybridized to '"P-labeledc-myc and @-tub ulin probes, and visualized by autoradiography asdescribed under "Experimental Procedures."proteins, and he washedpellet was dissolved in two-dimensionalelectrophoresis sample buffer (crude nuclear fraction). High saltextracts of crude nuclei were prepared according to the metho d ofDignam (29). Whole cell lysate, cytosolic fractions, an d high saltnuclear extracts were dialyzed overnight against water, then lyophi-lized an d redissolved in two-dimensional lectrophoresis amplebuffer.

RESULTST h e Synergistic Induction of Differentiation by Dexam eth-asone and Insulin and Inhibition by TPA-As an initial step

addit ion of various com binations of dexamethasone ( D ) , methyliso-buty lxanthine ( M ) , insulin ( I ) , an dT PA ( 5 " ) as described under"Exp erime ntal Procedures." Th ereafter , all cells received fresh FCS-containing medium every second ay un ti l staining. , he percentageof cells th at had differentiated a t 4 or 6 days after the addition ofvarious agents was quantified by replat ing cells at lower densit iesafter brief trypsinization, fixation, Oil Red 0 staining, and countingof individual cell as described un der "Exp erime ntal Procedures."


4/10

Genexpress ionuringdipocyteifferentiation 4725

Time after nitiating treatment (h)

0-12 12.24 24-36 3 6 -4 8 z 0-48 0 24-48Treatment with Dexamethasone (h)

FIG.3. Dexamethasone and TPA both influence commit-ment during the period 24-48 h after initiating differentia-tion. A , dexamethasone was added a t various times after initiatingtreatment of 2-day post-confluent cells with MI (open circles). TPAwas added at various times after initiating treatment with M1 (solidcircles). The medium was exchanged for10%FCS-containing mediumafter 48 h. The exten t of differentiation n duplicate cultures percondition (compared with cells treated with DM1 for the full 2-dayperiod) was estimated 6 days after initiating treatment by quantita-tion of Oil Red 0 staining as described under Experimental Proce-dures. B , dexamethasone was added to MI-t reat ed cells for 12-hperiods at he indicated times (0, no Dadded) followed by thre ewashes and replacement with MI-containing medium. In other re-spects cells were subjected to the standard di fferentiation protocol.Control treatments received DM1 for the full 48 h, either with orwithout washout and readdition of DM1 every 1 2 h. The extent ofdifferentiation in triplicate cultures per condit ion was estimated 6days after initiating treatment as described above. Results are themeans (5% of DM1 control) k S.E. 0-48 represents he summedvalue of the four treatments (0-12, 12-24, 24-36, and 36-48 h) .in identifying proteins involved in the differ entiatio nrocess,we determined the efficacy of combinations of the differen-t ia t ing agents and the phorbol ester, TPA, tonduce or inhibi ttriglyceride accumu lation. Lipid acc umu lation was f irst de-tected 72 h after ini t ia t ing treatmen t w ith ei ther dexam etha-sone and insulin (DI)or dexamethasone, m ethylisobutylxan-thine, and insulin (DM I), and differentia t ion was completeby 8 days (Fig.1) .The extent f differentia tion was quantifiedby determ ining the p ercen tage f adipocytes for each culturecondit ion 4 and 6 days after ini t ia t ing treatment. At least70% of the cells in DM I- and DI-tr eated cultu res, but lessth an 0.6% of t he cells t reated with FCS alone or with MI,contained numerous Oil Red 0-stainable l ipid droplets andhad the characterist ic roundeddipocyte morphology. N eitherD nor I alone induced more th an 20% of th e cells to differ-entiate (Fig. 1B). Consistent with the results of others (30,31), co-addition of TPAwith DM1 results n a dramaticinhibition of differentiation (9% conversion). These resultsreveal a synergism between D an d I to comm it cells to adi-pocyte differentiation which is blocked by T PA .

Com mitment to Adipocyte Differentiation-As a resu lt oftreatment with DM I, the cells undergo DNA synthesis anddivision followed by arres t in a uniq ue GI s ta te re la ted toterminal differentia t ion (32,33). Inur hands, the eak DNAsynthesis occurs a t abo ut 24 h and com mitte d cells havewithdraw n from th e cell cycle by 48 h (Fig. 2A). Analysis ofthe num ber of cells p er dish during differentiation (resultsnot shown) also su ppo rts the conclusion th at by 48 h theDMI- , but not the MI- t rea tedells, have withdraw n from th ecell cycle afte r an init ial rou nd of cell division. Additionalexperiments showed that TPA blunts the ini t ia l mitogenicresponses seen in response to DM1 and also inh ibi ts with-drawal from th e cell cycle after 48 h (upon ad dition of freshFCS-containing medium), as measuredy DNA synthesis andc-myc expression (Fig. 2, B a nd C ) .When the addit ion of dexamethasone is delayed unti l 36 hafter adding MI, theres a dramatic decrease in the proport ionof cells that undergo adipocyte differentiation. Similarly, theability of T PA to block d ifferentiation is also dram aticallyreduced when adde d after 36 h of DM1 trea tm ent (Fig. 3A) .Interestingly, whenTP A is added a f te rDM1 removalat 48 h,it does not block exp ression of the diffe rentiatio n programnor the enzymes of l ipid sy nthesis and i t actual ly increaseslater lipid accum ulation. These results suggest tha t D andTP A influence the comm itment ecision before36 h, althoughit is a lso possible th at a greater than 12-h exposure to th eseagen ts is required or their effect. T o addr ess his atterpossibility as well as o furth er deline ate whenD acts oinfluence com mitm ent, we also performed washo ut experi-men ts span ning the 48 -h treatm ent period. In this case MI-created cells were exposed to D for 12-h intervals ollowed byextensive washing in MI-containingedium. T h e triglycerideaccumulated was quantified 8 days after ini t ia t ing treatment(Fig. 3B). Since any ell that differentia tes accumu latesipidto about the same extent , thealues obtained are p roportion alto the num ber of differentiated cells per plate. Mo st of thecells differentia ted when D was presen t durin g the econd 24h of MI treatm ent. Moreover, the s um matio n of numbers ofdifferentiated cells over each of the 12-h perio dss very similarto the num ber of differentiated cells obtained with a full 48-h treatmen t with DMI, suggesting that independent subsetsof the cell population differentiate during each 1 2-h period.One interpretat ion of these results is that D acts during aparticular phase of the cell cycle to promote differentiation,most likely afterS-phase ,and ha tT P Aantagonizes heactions of D and/or I during the same t ime period result ingin commitment.Initial Responses to the Inducing Agentswithin 3.5 h-Havingestablished hat D andM I synergize topromotecom mi tm ent to adipocyte differentiation, we used large for-mat two-dimensiona lgel electro pho resis to dentify metabol-ically radiolabeled prot eins and in vitro translation prod uctsof cellular m RNA s whose synthesis or abund ance is synergis-tically altered by these agents. In particular, we wished tosingle out c andi dates for regulatory pro teins in the differen..tiation process. I nitial metabolic labeling experim ents indi-cated that the com bination f D, M, and I a l ters the relat iverat e of synth esis of well over 100 polypeptides with in 5 h(compared with fresh FCS alone). Furthermore, for m any ofthese initial inductions , the identity between the m etaboli-cal ly labeled pro teins and the prim ary in vitro translat ionproducts of cellular mRNA was readily established by com-parison of auto radio gram s and in some cases was confirmedby Cleveland peptide mapping (data no t show n), d emons trat-ing that the y are regulated at the level of mRNA (Fig. 4).Approximately 75% of these gene productsare maximally


5/10

4726 Gene Expression during Adipoc yteifferentiationInvivo

C

FIG. 4. Rapid modulation of geneexpression by the inducing agents.Similar regions of large format two-di-mensional gels loadedwith proteins met-abolically labeled 2-5 h after initiatingtreatment ( left) andhe in uitro trans- Dlation products of cellular mRNAs fromcells lysed 3.5 h after nitiating treat-ment (right). 2-day post-confluent 3T3-L1 cells were treated with 10%FCS me-dium containing either no further addi-tions (C), dexamethasone (D) , methyli-sobutylxanthine and insulin (MI),r allthree agents ( D M I ) . Species indicatedby arrowheads are regulated by MI andare not significantly influenced by D.Arrowed species are regulated by D andare either inhibited or not affected byMI. Circled proteinsare regulated byDM1 in a synergistic fashion. Not all of MIthe indicated species are equally detectedor regulated at both the level of proteinsynthesis and mRNA, and not all of thedifferences have been indicated.

Invitro

"' ** ... i" 0 . 3-

- . c

responsive to MI (D has no effect), most ikely reflecting theconsequence of mitogenic stimulation seen with MI treat -ment. On theother hand,no more than 10 proteinsaremaximally altered by D alone. These include two proteins(glutamine synthase and "glucocortin") that are induced inevery glucocorticoid-responsive cell type we have examined(34, 35). A number of proteins show more complex patternsof regulation. For example, the induction with D of mRNAs1 and 3 are inhibited with the coaddition of MI, while 10proteins are synergistically regulated by the combination ofD and MI. The induction or repression of these 10,as well assome of the D- and MI-induced proteins, was quantified. The10 synergistically regulated proteins as well as 4 D-inducedand 8 MI-induced proteins are listed in Table I, and thei r

locations on wo-dimensional gels are shown in Fig. 5. The 10synergistically induced proteins are of particular interest aspossible mediators of the effects of the inducing agents sincethese agents have synergistic effects on promotion of differ-entiation. These proteins were further characterized by fol-lowing the kinetics of induction a t th e level of mRNA byanalyzing in uitro translation products of mRNA isolatedduring the course of differentiation. Most of these rapidlyinduced mRNAs are only transiently elevated in response tothe inducing agents (for examples see Fig. 6), and furtheranalysis revealed either they are not synergistically inducedby D and I (which also promotes efficient differentiation), orthey are notnhibited by TPA (which inhibits differentiation);thus, the expression of no one of these proteins is predictive


6/10

Genexpressionuringdipocyteifferentiation 4727TABLE

Som e rapidly regulated proteins/m RNA sProtein no. Molecular Subcellular A/Pmass PI location" ratio'

Dexamethasone-regulated1

3 (glutamine synthase)'4 (glucocortin)5678910

2c.d

MI-regulated

111 2DM1 synergies13 (PCNA?)14151617'1819202 122

kDa42394217373532313150181636363532312828823930

7.2 Nuc/cyto6.6 Mem7.0 Cyto/mem4.0 Cytolnuc6.9 Mem6.8 Cyto6.2 Nuc7.3 Cyto6.2 Cyto6.3 Nuc/cyto5.1 Cyto/mem6.6 Cyto5.2 Cytolnuc6.3 Cyto/mem7.2 Cyto5.2 Mem7.1 Cyto/nuc6.8 Cyto6.4 Cyto6.5 Nuc4.9 Mem5.9 Nuc

3.2101.5110.71.7

4.60.80.20.70.21.01.53.60.011.21.01.55.90.50.80.4ND'3.6

Assignment of subcellular locations is based on relative enrich-ment above levels in whole cells and are given in rank order. nuc,nucleus; cyto, cytoplasm; mem, membrane.Fold difference in abundance of the in vitro translation productin fully differentiated adipocytes over untreated preadipocytes and isthe mean value of three independent determinations.e These proteins are also phenotypic markers.dProtein 2 is not detected by metabolic labeling in whole celllysates of preadipocytes but is detected in membrane fractions andin either whole cell lysates or membrane fractions from adipocytes.e Not detected in adipocytes.

TABLE1Differentiation-associated proteinsandphenotypic markers

Protein no. Molecular Subcellular A/Pmass PI locations ratiobDifferentiation-associated

DAP 1DAP 2DAP 3DAP 4DAP 5Ph 1Ph 2Ph 3Ph 4Ph 5 (GPDH?)dPh 6 (aP2?)Ph 7Ph 8

proteins

Phenotypic markers

kDa

53 6.9 Cyto/nuc/mem 10.851 5.7 Cyto/nuc 6.448.7 Mem4.339.8 ? 16.637.6 Nuc/cyto D e novo'29 6.9uc51 6.4 Cyto

D e novo58.742.3 Nuc 5.137.1 Cyto/nuc 7.3

14 6.5 Cyto/mem 21.436 6.4 Cyto39.4 Nuc/mem D e n o vo56.0

34.3 cy to 139.7

Refer to Table I. Definition of abbreviations is given in Table I.D e novo, not detected in preadipocytes.GPDH, glycerol-3-phosphate dehydrogenase.

* Refer t o Table I.

of the biological outcome. However, this does not rule out theinvolvement of these nitial nductions n commitment todifferentiation since it is possible that one or more maymediate the initial effects of one of the inducing agents.Later Responses to the Differentiation Agents Concurrent

with Commitment (24-48 h)-We also wished to identify theearliest proteins/mRNAs whose synthesis is altered concur-rent with or as a esult of comnlitment and thus re candidatesfor proteins involved in commitment, or that drive differen-tiat ion in committed cells. Such proteins should be (i) syner-gistically induced by DM1 and DI compared with D, I or MIalone, (ii) expressed at elevated levels before the appearanceof lipid droplets and the large group of later-evolving pheno-typic marker proteins, and (iii) should also be inhibited byTPA, since TPA blocks commitment. For this purpose wefurther analyzed the in vitro translation products of mRNAisolated a t various points during the differentiation process.Differentiation results in significant (>3-fold) changes in over200 in vitro translation productsby 96 h. Most (all but22) ofthese proteins were first altered between 51.5 and 96 h aftertreating the ells with DM1 or DI when the cellular phenotypeis in transition and emained at elevated levels in fully differ-entiated adipocytes; thus, we class these as phenotypic mark-ers of differentiation (Fig. 7). This group includes glutaminesynthase, non-neuronal enolase, and phosphoglyceromutase(identified either by immunoprecipitation or peptide micro-sequencing). Three others that are altered in differentiatedcells have been tentatively identified as aP2 (also known asgene 422 and atty acid-binding protein), glycerol-3-phos-phate dehydrogenase and proliferating cell nuclear antigen,bn he lattercase by comparison with published gels on whichthis protein has been identified (36,37) (see Fig. 5 and TableI). Of the 22 changes that were not phenotypic markers, theregulation of most did not correlate with commitment todifferentiation in that they were either not synergisticallyregulated by D and I in the absence of M, or the synergismwas not blocked by TPA. On the other hand, the regulationof five in vitro translation products and their correspondingmetabolically labeled proteins does predict the differentiationoutcome within the first 51 h; thus, they are candidates forproteins which may function in the differentiation process.These species are synergistically induced byDM1 and DItreatments before the appearance of lipid droplets and thephenotypic marker proteins, and this ynergism is blocked byTPA. The locations of these species, which we name differ-entiation-associated proteins (DAPs), are shown in Fig.5,and they are listed in Table 11.The DAP mRNAs do not all follow a single pattern ofregulation (see Fig. 8). DAP-1 and DAP-4 mRNAs display abiphasic response to DM1 and DI. DAP-1 is rapidly andsynergistically induced within 3.5 h by DM1or DI treatments(Fig. 8 A ) . Although this synergism is not inhibited by TPAat the level of mRNA, TPA inhibits the amount of proteinsynthesized during this time (results not shown). After 43 h,DAP-1 mRNA levels again increase in DMI-treated cells until96 h (10-fold induction)and remain at this level in fullydifferentiated adipocytes. In DI-treated cells the increase inDAP-1 occurs more slowly,consistent with the less synchron-ous differentiation induced by this treatment. In contrast, theinitial induction of DAP-4 mRNA by DM1 is inhibited byTPA (Fig. 8B) lthough the magnitude of the induction issomewhat variable at 3.5 h. DAP-4 reaches a peak 12-foldinduction by 18 h(resultsnot shown), after which evelsabruptly decrease and return o control levels by 24 h. There-after the level rises again by 51.5 h in DMI- and DI-treatedcells and reaches maximal levels (25-fold) by 96 h.In contrast to DAPs 1 and 4, the synergistic induction ofDAPs 2, 3, and 5 do not follow a biphasic pattern (Fig. 8, C-E ) . For these proteins the induction with DM1 and DI treat-ments begins with commitment (after 24 h), increases through


7/10

4728 Gene Expression during Adipocyte Differentiation7.0 6.5 PI 6.0.5.0.5

" Lr*z

Jt*

" 8a

99

67

43

(Da

26

1814

FIG.5 . Large format two-dimensional electrophoresis of in vitro translation products of poly(A+)mRNA isolated f r o m 3 T3 - L l adipocytes. Poly (A+ ) R NA isolated from cells 4 days after treating with DM1was translated in vitro in the prese nce f [:"S]methionine. The labeled in uitro trans lation produ cts were subjectedto large format two-dimensional electrophoresis and visualized by au toradiography. The locationsof the in vitrotranslation products l isted in Tables I and I1 are indicated.96 h, an d elevated levels are main tained inully differentiatedadipocytes.

DISCUSSIONMany features of our results are supportive of the resultsof others . The data are consis ten t wi th cur ren t ideas tha tcommitment toadipocyte differentiation involves synergisticactions of inducing agents (horm ones and growth factors)th at result in w ithdrawal from theell cycle and subseq uentlyth e induc tion of adipocyte-specific genes 32,33). t is unlikelythat the dramat ic changes inene expression th at occur laterin differentiation epresentdirectactions of the ndu cingagents on these genes, since the continued presence of t heinducing agents is not required to maintain the differentiatedsta te . Thus, i t seems tha t commitment to , and expressionf,the adipocyte phenotype is controlled by persistent changesin th e expression of certain gene products brought about byth e actions of the nducin g agents . In his paper we havedetermined the t ime 24-48 h after init iating differentiation)

during which the interactionbetween insulin and dexameth-asone results in commitment. We thendescribe five protei nsthat are candidate differentiation factors (DAPs 1-5) due totheir synergistic regulation during the timehen commitmentoccurs. The regulatio n of these proteins are amon g th earli-est-described markers f the differentia tion utcome. Wehavealso identified a num ber of protein and mRNA nductionsth at ar e ynergistically induced by D n combinationwith MI ,as well as some tha t repr esen t responses to th e individualinducing agents prioro the commitmenteriod. One or moreof these may also serve to media te thendividual or combinedeffects of the inducing agents on commitm ent.With the differentia tion protocol we used, addition of thecombination of fresh FCS, M, I , and D to post-confluentquiescent cells results in a round of DNA synthesis, (peakingaround 24 h) an d ell division. Th is is followed by withdrawalfrom the cell cycle by 51 h, and triglyceride accumulation isevident by 72-94 h. The high level of insulin necessary toinduce differentiation and additional bservations by us and


8/10

Genexpressionuringdipocyteifferentiation 4729

EAMI+TPAe 2000%!50Lm: 00

0 5E

00 1 2 2 4 36 4 8 B O 7 2 8 4 9 6

Time (h)

80

0 5E

00 1 2 2 4 36 4 8 B O 7 2 8 4 9 6Time (h)

Time (h)'40

-g% a 00L

>m 20mac0p 100LL

00 1 2 2 4 3 6 4 8 6 0 7 2 8 4 9 6Time (h)

FIG. 6. The kinetics of expression of some synergisticallyregulated proteins during differentiation.%day post-confluent3T3-Ll cells were treated for up t o 48 h a s indicated. Cultures werelysed at the indicated times a nd poly(A+ ) RNA was isolated, trans-lated in uitro, and the labeled translation products subjected to largeformat two-dimensional electrophoresis and visualized by autoradi-ography. The abundance of protein spots on matched exposures wasdetermined by densitometric scanning and the fold change (versusFCS control for each time point) determined.others, including the functional substitution of 6 nM, insulin-like growth factor-1 for insulin (data not shown), indicateth at insulin influences adipocyte differentiation through in-teractions with insulin-like growth factor-1 receptors (20, 38,39). We also find that initially D seems to potentiate theinitial mitogenic response elicited by M, I, and FCS. However,it is clear that post-confluent cell division alone is not suffi-cient to commit the cells. A subsequent withdrawal from thecell cycle also seems to be necessary since treatments thatpromote continued growth and cell division block commit-ment rather than promote differentiation (Refs. 33, 40, and41 and our data with TPA). Interestingly, the initial and thelater effects of D on MI-stimulatedDNA synthesis andc-mycexpression are opposite in sense. This may result from inter-actions of D with cell cycle-dependent orother evolvingcellular processes.

Platelet-derived growth factor and basic fibroblast growthfactor also inhibit adipocyte differentiation (Ref. 41 and data

o k r m r * c m m r . . . .0 1 2 2 4 36 4 8 6 0 7 2 8 4 9 6Time (h)

- . . . . . . . . . J1 2 2 4 3 6 4 8 6 0 7 2 8 4 9 6

Time (h)FIG. 7. The kinetics f expression of some phenotypic mark-ers during differentiation. 3T3-Ll cells were treated, and RNAwas isolated and analyzed as described in Fig. 6.

not shown). This suggests that inhibition by these growthfactors and TPA is through protein kinase C activation, asopposed to itsdown-regulation, since these peptide hormonesactivate but do not down-regulate protein kinase C (42-44).The presence of D (we find that 10 nM is sufficient) iscritical for efficient commitment, but the g1ucocorticoid;reg-ulated factors that mediate this action are unknown. Onepossibility is that D influences commitment by inhibitingprostaglandin production (45). Glucocorticoids inhibit pros-taglandin production in several cell types and tissues (46) andthe major prostaglandin species produced in 3T3-Ll preadi-pocytes (prostaglandins E2 and 12) are inhibitory for differ-entiation (6, 47, 48).Furthermore, indomethacin, which in-hibits prostaglandin H synthase (cyclooxygenase),enhancesadipocyte differentiation in 3T3-Ll cells (49, 50). In supportof this hypothesis, we find that D rapidly inhibits the synthe-sis of this enzyme in hese cells and blocks much of theincrease in synthesis seen in response to MI treatment (data

not shown). However, the concentration of indomethacinrequired for a significant effect on differentiation is muchgreater than that needed for complete inhibition of prosta-glandin H synthase activity (51); thus it appears likely thatthe effect of D on prostaglandin synthesis does not accountfor its entire influence on differentiation.Although the effects of M are most likely mediated throughits inhibition of soluble cyclic AMP phosphodiesterase (52),the role tha t it plays in initiating differentiation is unclear.The results presented in Fig. 1 show no dramatic enhance-ment of DI-induced differentiation by M, but in otherexper-iments M clearly increases the rapidity and uniformity ofdifferentiation when added together with D and I. The pro-moter regions of several adipocyte-specific genes containH. B. Sadowski, T. T. Wheeler, and D. A. Young, unpublishedresults.


9/10

4730 Gene Expression during Adipocyte Differentiation

FIG.8. The kinetics of expressionof he mRNAs of the differentiation-associated proteins. 3T3 -Ll cells weretreated, and RNA was isolated and ana-lyzed as described in Fig. 6.

/ I0 4 . . . I . . . , . . , . . . , . . . , . . . , . . . , . . . ,0 124 36 4 0 6 0 7 2 84 96

nm a th1Time (h) I..,

0 1 24 3 6 4 8 60 7 2 0 4 9 6Time (h)

cyclic AMP-responsive elements (14, 53), and M enhancesth e expression of some adipocyte phenotypic marker proteinsduring differentiation (see Fig. 7). However, the enhancementof these proteins above levels achieved by DI alone does notdepend on the continued presence of M after 48 h, when allthe inducing agentsare removed, suggesting that M alsoinfluences commitment to differentiation,perhaps by en-hancing the expression of one or more proteins involved inthe commitment process. An additional complicating factoris the suppressing effect that M has on the low extent ofdifferentiation in response to insulin alone (see Fig. 1).

Several of the DAPs (which are induced concurrent withcommitment) have characteristics consistentwith them func-tioning during the differentiation process. Of the over 4000proteins and 2000 in vitro translation products we surveyed,th e five DAPs represent the earliest predictors of the adipo-cyte phenotype. DAPs 1, 2, 4, and 5 are present in nuclei orhigh salt extractsof nuclei, while DAPs 2,3, and 5 are elevatedto near maximal levels in the 24-48-h period in which com-mitment occurs. However, we cannot rule out the possibilitythat these proteins may simply be very rapidly evolving phe-notypic markers with no functional role in controlling thedifferentiation process. Inaddition to he DAPs we have

1 2 2 4 36 400 7 2 8 4 9 6Time (h)

Time (n)

detected a number of rapid protein inductions that maymediate the initial effects of the inducing agents. Theseinclude about 75 MI- and 10D-mediated inductions as wellas 10 proteins and mRNAs tha t are synergistically regulatedby the combination of D and MI. The molecular cloning andfunctional characterization of the DAPs or of these rapidlyinduced proteins may shed light on the commitment process.Acknowledgments-We thank Ann Colasurdo, Michael Battaglia,and Steven Beck for excellent technical assistance, Dr. R. Slemmonfor peptide microsequencing, and Dr. M. D. Lane for kindly providingus with 3T 3-Ll cells.

REFERENCES1. Weintraub, H., Davis, R., Tapscott, S., Thayer, M., Krause, M.,Benezra, R., Blackwell, T. K., Turner, D., Rupp, R., Hollenberg,S., Zhuang, Y., and Lasser, A. (1991) Science 251, 761-7662. Green, H., and Kehinde, 0. (1976) Cell 7, 05-1133. Reed, B. C. , and Lane, M. D. (1980)Adu Enzymol. 18,97-1174. Mackall, J. C. , and Lane, M. D. (1977) Biochem. Biophys. Res.5. Kuri-Harcuch, W., and Green, H. (1977) J.Bwl. Chem. 252,6. Rubin, C. S., Hirsch, A ., Fu ng, C ., and Rosen, 0. M. (1978)J.7. Green, H., and Kehinde, 0. (1979)J . Cell. Physiol. 101, 169-172

Commun. 79, 720-7252158-2160Biol. Chem. 253,7570-7578


10/10

Genexpressionuringdipocyteifferentiation 47318. Spiegelman, B. M., Frank, M., and Green, H. (1983) J . Biol. 29. Dignam, J. D. , Lebovitz, R. M ., and Roeder, R. G. (1983) Nucleic9. Bernlohr, D. A,, Angus, C. W., Lane, M. D., Bolanowski, M. A., 30. Shimizu, y., Shimizu, N., Fujiki, H., and SWimura, T. (1983)Chem. 258,10083-10089 Acids Res. 11 , 1475-1489and Kelley, T. J. J. (1984) Proc. N atl. Acad. Sci. U. S . A. 8 1 , Cancer Res. 43,4974-49795468-5472 31. Gamou, S., and Shimizu, N. (1986) Cell Struct.unc. 1 1, 21-3010, Chapman, A. B., Knight, D. M., Dieckmann, B. S. , and Ringold, 32. Scott, R. E., Florine, D. L., Willie, J. J. J.7 and Yun, K. (1982)

11. Smith, P. J., Wise, L. S., Berkowitz, R., Wan, C., and Rubin, C. 33 . Frefiag, s.O. (1988) Mol . Bioi. 8, 614-162412. Hunt, C.R., Ro, J. H.-S., Dobson, D. E., Min, H. Y., and

G. M. (1984) J . Biol. Chem. 25 9 , 15548-15555S. (1988) J . Biol. Ch em. 263,9402-9408

Proc. Natl. Acad. Sci. U. S. A . 79, 845-84934. Colbert, R. A., and Young,D. A. (1988) Proc. Natl. Acad. Sci.U . S . A . 83 , 72-76spiegelman, B. M. (1986) proc.~ ~ t l ,~ ~ d ,ei, u,s,A. 83, 35. Colbert, R.A., and Young, D. A. (1986) Endocrinology 119,3786-3790 2598-260513. Distel, R. J., Ro, H.-S., Rosen, B. S., Groves, D. L., and Spiegel-

14. Wilkison, W. O., Min, H. Y., Claffey, K. P., Satterberg, B. L.,15. Yang, V. W., Christy, R. J., Cook, J. S., Kelly, T. J., and Lane,16. Christy, R. J., Yang, V. W., Ntambi, J. M., Geiman, D. E., Cell Res. 152,368-377

36. Prelich, G., Kostura, M., Marshak, D. R., Mathews, M. B., andman, B. M. (1987) Cell 49,835-844 37. Bravo, R., Frank, R., Blundell, P. A., and Macdonald-Bravo, H.and SPiegelman*B. M. (1ggo) J . Chem. 2659 477-482 38. Steinberg, M. M., and Brownstein, B.L. (1982) J. Cell. Physiol.M. D. (1989) proc. Natl. ACad.SCi. u.s. A . 8 6 , 3629-3633 39. Amri, E.-Z., Grimaldi, P., Negrel, R., and Ailhaud, G. (1984)Erp .Landschultz, W. H., Friedman, A. D., Nakabeppu, Y., Kelly, 40. Narve, M., and Ringold, G. M. (1988) J . Cell Biol. 10 7 , 279-286T. J., and Lane, M. D. (1989) Genes Deu. 3, 323-1335 41. Navre, M., and Ringold, G. M. (1989) J . Cell Biol. 10 9, 1857-Acad. Sci. U. S. A . 87,251-255 42. Blackshear, P. J., Wen, L., Glynn, B. P., and Witters, L. A. (1986)Spiegelman, B. M. (1989) Mol. Cell. Biol. 9,5331-5339 43. Coughlin, S. R.,Lee, W. M. F., Williams, P. W., Giels, G. M.,Science 25 1 , 288-292 44. Tsuda, T., Kaibucbi, K., Kawahara, Y., Fukuzaki, H., and Takai,

Stillman, B. (1987) Nature 326,471-475(1987) Nature 326, 515-517S ~ p p l . ,37-50

17. Kaestner, K. H., Christy, R. J. , and Lane, R. J. (1990) Proc. Natl. 186318. Herrera, R., Ro, H. S.,Robinson, G. S.,Xantbopoulos, K. G., and J. Biol. Chem. 261,1459-126919. Umek, R.M., Friedman, A. D., and McKnight, S. L. (1991) and Williams, L. T. (1985) Cell 4 3 , 243-25120. Green, H., and Kehinde, 0. (1975) Cell 5 , 19-27 Y. (1985) FEB S Le t t. 19 1, 205-210

W. J. (1979) Biochemistry 18 , 5294-5299 Biol. Chem. 253,7579-75831408

21. Chirpin, J. M., Pryzbyla, A. E., MacDonald, R. J., andRutter, 45. Rosen, 0. M.7 Smith, c. J.9 c. , and Rubin, c. s. (1978) J .22. Aviv, H., and Leder, P. (1972) Proc. Natl. Acad. Sci. U. S. A. 69 , 46. Flower, R. J. (1988) Br . J. PharmacoL 949 987-101547. Hopkins, N. K., and Gorman, R. R. (1981) Biochim.Biophys.23' Colbert' R' A'' and Young' D' A. (1986) J' Bioi' 261' 48. Hyman, B. T., toll, L. L., and Spector, A.A. (1982) Biochim.24' D' 'Oris, B' p '9 E. v't and Colbertl R' A' 49. Williams, I. H., and Polakis, S. E. (1977) Biochem.iophys. Res.

Acta 663,457-466Biophys. Acta 7 1 3 , 375-385Co mmu n . 77 , 175-186

14733-14739(1983) Methods Enzymol. 9 1 , 190-214484 4696

26. Levenson, R. M.y Maytin, E. v.* nd D. A. (1986) 51. Knight, D. M., Chapman, A. B., M. N., Drinkwater, L., Bruno, J.Biochem. 15 8, 294-30127. Sambrook, J., Frits& E. F., and Maniatis, T. (1989) Af&cuhr 52. Elks, M. L., and Manganiello, V. C. (1985) J. Cell. Physio l. 124,Cloning-A Laboratory Manual, 2nd Ed., Cold Spring HarborLaboratory, Cold Spring Harbor, NY28. Feinberg, A. P., and Vogelstein, B. (1983) Anal. Biochem. 132 , 53. Ntambi, J. M., Buhrow, S . A. , Kaestner, K. H., Christy, R. J.,Sibley, E., Kelly, T. J., Jr. and Lane, M.D. (1986) J . Biol.

25. Voris, B. p. , and Young, D. A. (l980) Anal . lo 4, 478- 50. Cbang, T. H., and Polakis, S,E. (1978)J . Bioi, C k m , 253,4693-J., and Ringold, G. M. (1987)Mol. Endocrinol. 1 , 36-43191-198

6-13 Ch em. 263,17291-17300

J. Biol. Chem.-1992-Sadowski-4722-31

Documents

Transcript of J. Biol. Chem.-1992-Sadowski-4722-31