Vol. 9, 327-336, April 1998 Cell Growth & Differentiation 327

Human Annexin I Is Highly Expressed during theDifferentiation of the Epithelial Cell Line A 549:

Involvement of Nuclear Factor Interleukin 6in Phorbol Ester Induction of Annexin �

EgIe Solito,2 Catherine de Coupade, Luca Parente,Rodenck J. Flower, and Fran#{231}oiseRusso-ManeInstitut Cochin de G#{233}n#{233}tiqueMol#{233}culaire,Institut National de Ia Santeet de Ia Recherche M&licale Unite 332, 75014 Paris, France [E. S.,C. d. C., F. R-M.]; Istituto di Farmacologia e Farmacognosia, Universlt#{224}di Palermo, 90134 Palermo, Italy [L P.]; and The William HarveyResearch Institute, St. Bartholomew’s and The Royal London School ofMedicine and Dentistry, London EC1M 6B0, England [R. J. F.]

AbstractThe role of annexin I (Ax 1) in cell differentiation wasstudied in the A 549 epithelial cell line, a human lungadenocarcinoma line, that responds to phorbol estersand glucocorticoids by induction of differentiatedproperties. Ax I has also been reported to be involvedin the control of cell proliferation. We report that Ax Isynthesis occurs upon phorbol 12-myristate 13-acetate(PMA) treatment of A 549 cells and its appearance iscorrelated with the presence of dipeptidyl peptidase IV,or CD26, a marker of epithelial cell differentiation. Inaddition, using transfection experiments and site-directed mutagenesis with the Ax I promoter coupledto a reporter gene, we report that a unique region ofthe Ax I promoter confers the response of the reportergene to PMA and dexamethasone. This response toPMA and/or dexamethasone involves the induction ofthe synthesis and/or the activity of trans/cis-activatingtranscriptional factors. Furthermore, we havedelineated the mechanism of the transcriptionalactivation of Ax I by PMA and the involvement of aspecific transcription factor, nuclear factor interleukin

6 (C/EBP 13).

Introduction

Annexins (also known as lipocortins) are a family of structur-

ally related proteins that exhibit Ca2�-dependent binding toanionic phospholipids. They have been identified in many

Received 6/23/97; revised 12/30/97; accepted 2/2/98.The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby markedadvertisement in accordance with 18 U.S.C. Section 1734 solely to mdi-cate this fact.1 This work was supported by the Association pour Ia Recherche sur IaPolyarthrite. E. S. is a Research Fellow of the CEE (Biotechnology Pro-gram Grant ERB 4001GT920002)and of lnstitut National do Ia Sante et deIa Recherche M&llcale. R. J. F. is a Principal Research Fellow of theWellcome Trust.2 To whom requests for reprints should be addressed, at lnstitut Coct�n doG#{233}n�tique Molecu�re, Institut National do Ia Sante at do Ia RechercheM#{233}dicaleUnite 332, 22 Rue M#{233}chain,75014 Paris, France. Phone: 33-1-40516450; Fax: 33-1-40517749; E-mail: sohto@icgm.cochin.lnserm.fr.

organisms, ranging from mammals to molds and even plants.Many annexins are abundant intracellular proteins, some-times comprising more than 2% of total cellular proteinmass. Clear roles for annexins have not yet been demon-strated, although a wide range of biological functions has

been proposed, including anti-inflammatory and anticoagu-lant activities. The involvement of some annexins in the ag-gregation and fusion of membranes and in endocytosis and

exocytosis is well documented (1) but not yet demonstrated

in vivo. Some annexins are expressed in a growth-dependent

manner and are targets for cellular kinases in viva, suggest-ing that they may be involved in differentiation and mitogen-

esis (2).Ax 1 ,� the first member of this family, has been reported to

be involved both in differentiation and in inflammation. In-deed, Ax I has been reported to be the mediator of theanti-inflammatory action of glucocorticosteroids. This effect

of GCs on Ax 1 formation requires that cells be initiallydifferentiated so that GCs can induce Ax 1 formation (3),raising the possibility that Ax I was involved in the differen-

tiation process. in parallel, it was shown that Ax 1 wasabundant in a restricted number of differentiated cell types inadult organs, whereas it was apparently totally absent fromundifferentiated tissues (2). The presence of Ax 1 was alsoobserved during the differentiation of neopiastic astrocytes,ependymocytes, and Schwann cells (4). A large number ofreports have shown that Ax 1 was induced during the differ-entiation of the U-937 and HL 60 cells toward the macro-phage lineage (5-7). Violette et a!. (8) reported that Ax 1 wasinvolved in the GC-induced terminal differentiation of a hu-

man cell squamous carcinoma and correlated this effect tothe extracellular presence of Ax 1 . These results were furthersupported by those of Kang et a!. (9), who demonstrated thatPMA-induced differentiation of the U-937 cell line was cor-

related with the translocation of Ax I to the membrane. Ax 1has also been found to be involved in the control of differ-

entiation and/or proliferation of many different types of epi-thelial cells: human epidermal keratinocytes (1 0, 1 1), culturedbronchial epithelial cells (1 2), and hepatocytes during hepa-toceliular carcinoma (1 3). Taken together, these data argue

for a role of Ax 1 in the control of cell proliferation and/ordifferentiation in cells of the macrophage lineage, as well asin cells of epithelial origin.

3 The abbreviations used are: Ax 1 , annexin 1; GC, glucocorticoid; PMA,phorbol 12-myristate 13-acetate; DPP-IV, dipeptidyl peptidase IV; DEX,dexamethasone-21 -phosphate; NFIL-6, nuclear factor Interleukin 6;C/EBP, CCMTenhancer binding protein; FACS, flow cytometric analysis;GRE, glucocorticoid response element; WT, wild-type; GA, GC receptor,EMSA, electrophoretic mobility shift assay.

A

kDa

1 2 3 4 5

B

kDa

37 �

C

Cytosol

Membrane

PMA Treated

___5

- FITC Ab

--- CD26Ab

100 10’ 102 i03 i04

Intensity of fluorescence

- FITC Ab

‘I �,�

� �

A --- CD26AbJ\‘I .4

I\ /k4� � j ,�

Ue�’ ii

ss �

100 10’ 102 � tO4

Intensity of fluorescence

328 Annexin 1 in Epithelial Cell Differentiation

37 �

32 �

*�� � �,

1 2 3 4

Untreated

Fig. 1 . A and B, immunoprecipitation analysis of Ax 1 . As reported in “Materials and Methods,” cell membrane-bound protein was EDTA-extracted fromA 549 cells that were either untreated (control; Lanes 1, c) or treated with PMA (Lanes 3), DEX (Lanes 4, dex), or PMA and DEX (Lanes 5, PMA + dex),immunoprecipitated, and then separated by SDS-PAGE. Cytosolic Ax 1 was also immunoprecipitated using anti-Ax 1 monoclonal antibody, separated bySDS-PAGE and immunoblotted with a polyclonal antibody Ax 1 . Human recombinant Ax 1 was used as marker (Lanes 2). A, Western blot of cytosol extracts.B, Western blot of cell membrane-associated Ax 1 . Top, stimuli are listed. C, representative flow cytometric histograms of DPP-IV (CD26)-FITC binding byA 549 cells. Left, untreated cells incubated with the antimouse FITC second antibody in the absence or in presence of the CD26 antibody; right, PMA-treatedcells similarly incubated. Shifted histogram, CD26-positive cells.

The A 549 cell line is an epithelial cell line derived from a

human lung adenocarcinoma. Epidermal growth factor stim-

ulates its proliferation and dexamethasone induces its

growth arrest. Ax 1 has been reported to be involved in the

growth arrest of this cell line (14, 15). in this model, the

extracellular presence of Ax 1 seems to also be related to the

role on cell differentiation. This epithelial cell line appeared,

therefore, to be a good model for studying the effect of PMA

and GC on cell differentiation, while evaluating Ax 1 at the

promoter level. Using this cell line, we have shown that Ax 1

synthesis occurs after PMA treatment and is correlated with

the presence of DPP-IV (or CD26), a marker of epithelial celldifferentiation. In addition, we have performed transfection

experiments with the Ax 1 promoter coupled to a reporter

gene. We now report that a unique region of the Ax 1 pro-

moter confers the response of the reporter gene to PMA and

DEX and that this response to PMA and/or DEX requires theinduction of the synthesis and/or the activity of trans/cis-

AP.1 NF.1 NF.1

I� INTRON

TCF AP.1 GREA

B

5’ A B C +20 y

Xho � Hindifi.610 .295 .180 .106 +270

I lx

Sx

E

C

I dex

I I’MA

a

pSV4O.Luc pAxl(880).Luc WT

Cell Growth & Differentiation 329

Fig. 2. The effect of dexamethasone and PMA on the transcription of the Ax 1 gene in differentiated A 549 cells. A, schematic diagram of Ax 1 -Lucconstruct and the consensus sequences for the protein-binding domains in Ax 1 promoter gene. The sequences termed A, B, and C, spanning - 1 80 to- 1 06 bp and covered by specific oligonucleotides, were further analyzed for nuclear binding proteins. The diagram shows the potential response elementsbased on sequence similarities to consensus response elements. Distances are given as nucleotide positions relative to the transcriptional start site (+ 1).B, A 549 cells were transiently transfected with the different plasmids (as reported in “Materials and Methods”) together with the cytomegalovirus promoterfused to the Lac Z gene plasmid, used as an internal control. Following transfection, the cells were incubated for 24 h with either no stimulus (control) or1 0 nM PMA, and for a further 1 6 h with 1 jiM DEX. Cells were then harvested, lysed, and assayed for both luciferase and j3-galactosidase activities. Columns,relative luciferase activity (percentage increase over unstirnulated cells) obtained by dividing the normalized luciferase activity by the unstimulated reportervector (control), mean of three independent experiments performed in triplicate; bars, SE.

activating transcriptional factors. Furthermore, using gel shiftassays and site-directed mutagenesis, we have delineatedthe mechanism of the transcriptional activation of Ax 1 by

PMA and the involvement of a specific transcription factor,NFIL-6 (C/EBP f3).

ResultsExpression of Ax I after PMA and DEX Treatment. Fig. 1shows the expression of Ax 1 protein in A 549 cells and itsinduction by different stimuli, both in the cytosol (Fig. 1A) andbound to the external membrane (Fig. 1B), as measured byimmunoprecipitation. in the cytosol (Fig. 1A), the protein is

constitutively abundant in untreated cells (Lane 1). PMA(Lane 3) and DEX (Lane 4) treatments up-regulate the syn-

thesis of the protein. The combined stimuli of PMA and DEX

(Lane 5) also show a strong induction of the protein. Thestimulation of the low molecular weight forms of the proteinis striking, although we worked in the presence of proteaseinhibitors, but it is known that Ax 1 exists in three different

forms, one of which can be phosphorylated. Fig. lB shows

membrane-associated Ax 1 . In unstimulated cells (control,

Lane 1) or cells stimulated with PMA alone (Lane 2), mem-

brane-associated Ax 1 is not detected. Treatment with DEX

alone (Lane 4) or with PMA and DEX (Lane 5) translocates theprotein to the membrane and processes it to the externalsurface of the cell. it is important to note that we have to takeinto account the total amount of the protein (cytosol and

membrane) after each treatment to correctly understand the

importance of the induction.

Differentiation Effect of PMA on A 549 Cells. Because it

is well documented that the expression of DPP-lV (or CD26)is correlated with the state of epithelial cell differentiation, thedifferentiating effect of PMA treatment on these cells of

epithelial origin was evaluated by FAGS using anti-GD26antibody as a marker. Fig. 1C illustrates A 549 cells, which,

after PMA treatment, become positive for GD26 when ananti-GD26 antibody is used. Dexamethasone, alone or added

after PMA treatment, did not modify either the constitutive orPMA-regulated expression of GD26 (data not shown).

Expression of the Cloned Ax I Promoter in A 549 Cells.As described in “Materials and Methods,” a region of 880 bp(-61 0 to +270 bp) of the Ax 1 promoter (Fig. 2A) was cloned

-610 AP-I

5.

A

B

+20

:� : : I�’

�8555555555521

Relative Activity

(% of SV 40 promoter)

0 20 40 60 80 100 120

M4

MS

M6

M7

Luciferase Index (% Increase of Control)

C

0 500 1000 1500

HIi

WT

Ml

M2

M3

M4

MS

M6

M7

. dcx� PMAD PMA-dex

330 Annexin 1 in Epithelial Cell Differentiation

pSV4O.Luc

pAx I (880).Luc ‘NT

p Ax I (555).Luc Ml

p Ax I (450).Luc M2

pAx 1 (376).Luc M3

pAx 1 (630).Luc M4

pAx I (318)-Luc MS

pAi 1 (200)-Luc M6

pAx 1 (l26).Luc M7

pAx 1 (76)-Luc M8

pbasic-Luc

NF.1 + +20 TCF AP-l GRE +270

w.� .-298 .180 -106 .56

JIi

:Ii

�eaaaaeaaesss�sessJ3’ WT

wasea�saeew�assnsss�aJ3’ Ml

eseaaeseesenaessaensJ M2

���asssa6J3#{149} M3

Fig. 3. Promoter activity of the 5-flanking region of the Ax 1 gene. Promoter activity was evaluated using fusion genes comprised of sequentially deleted5-flanking regions (described in “Materials and Methods”), the region downstream of the transcriptional start (+1) site of the Ax 1 gene, and the luciferasereporter gene. A, sequential deletions from the 5’ end and the 3’ sequence downstream from the transcriptional start site are shown. B, basal levels ofluciferase expression are shown as percentage activity of the expression of the positive control, pSv4O-Luc, a fusion gene of the SV4O promoter, and aluciferase reporter gene (oSV4O-Luc). A promoteriess luciferase plasmid was used as a negative control (obasic-Luc). Numbers in parentheses, length inbp of the cloned segment linked to the luciferase reporter gene. Values for luciferase activity were adjusted by transfection efficiency, as determined by�3-galactosidase expression by CMV-Lac Z plasmid. Columns, mean of three independent experiments performed in triplicate; bars, SE. C, stimulatedluciferase activity of A 549 cells transfected, as reported in “Materials and Methods,” with the WT or the different deletion mutants of Ax 1 promoter andstimulated with PMA, DEX, or the combination. The luciferase index was expressed as percentage increase above that for unstimulated cells. Columns,mean of three different representative experiments performed in triplicate; bars, SE.

into an expression vector reporting the luciferase gene

(pGL2-basic-Luc). This construct was tested by transient

transfection in A 549 cells. Transcriptional activity, measured

as luciferase activity, was compared to that of the controlplasmid (pSV4O-Luc). A 549 cells showed a marked re-

sponse to both steroid and PMA treatments with 3- and6-fold increases, respectively, in luciferase activity (Fig. 2B).

The combined stimuli (PMA and DEX) resulted in an 1 1-fold

increase of luciferase activity (Fig. 2B). The cells containingthe control piasmid (pSV4O-Luc) responded only to the PMA

stimulation, as expected.

5’ and 3’ Deletions of the Ax I Gene Promoter. A com-puter search for canonical consensus sequences of potential

regulatory elements is shown in Fig. 3. By sequence analysis

twojun (APi) consensus sequences were located at -425

and + 1 77 bp, respectively, the latter lying not far from a

simple GRE consensus sequence bearing only a hexamer

half-site. Upstream of the consensus for Jun (+1 77 bp), we

identified a consensus for the P62 TCF involved in the re-

sponse to serum factors. Two CCAAT sequences (nuclearfactor-i) were identified at -77 and -313 bp. To analyze the

promoter elements contained in this DNA sequence, eight

deletion mutants were constructed as described (Fig. 3). The

basal expression was analyzed by transient transfection in A549 cells (Fig. 3, A and B). The WI promoter containing the

entire cloned region showed a basal activity overlapping that

of the control construct containing the SV4O promoter. Onthe other hand, the basal expression of the mutants Mi

(-298to +270 bp), M2 (-i8Oto +270 bp), and M3 (-106 to+270 bp; Fig. 3A), in the presence of the first intron, was

Cell Growth & Differentiation 331

Table 1 Effect of GA occupancy on Ax 1 gene expression

Cells transiently transfected with the wild-type plasmid pAx 1 (880)-LucWlwere incubated, after PMA(24 h)treatment, with DEX(1 pi,i) and/or thesteroid analogue RU-486 (1 �M) for a further 16 h, after which luciferaseactivity was measured. The results are presented as a luciferase index (%increase over control) for each treatment. All data were also normalized tothe values of cells treated with PMA in the presence of RU-486.

pAx 1 (880)-Luc WT�(luciferase index)

DEX 248±96PMA 539 ± 90PMA+DEX 1069±16DEX-RU-486 112 ± 24PMA + DEX + RU-486 280 ± 70

significantly reduced. However, the results obtained with M5,M6, and M7 (-298 to +20 bp, -180 to +20 bp, and -106to +20 bp, respectively; Fig. 3B) mutants indicate that thisreduction was partially abrogated in the absence of the first

intron, except for M8, which has almost totally lost its tran-scriptional activity.

To verify if these promoter mutants responded to PMA andDEX, A 549 cells were transiently transfected either with the�wr promoter or with each individual the mutant and stimu-lated as described above. The results are shown in Fig. 3C.Treatment with DEX, PMA, or PMA and DEX stimulated lu-ciferase activity to a similar extent in the presence of the WTor Ml mutant. The M2 mutant showed a reduction of ap-proximately 20-50% in terms of luciferase activity after PMAand PMA plus DEX treatment. in contrast, a marked reduc-tion of the response to all of the different stimulations wasobserved in the M3 mutant. These results suggest that theregion spanning - 1 80 to -106 bp is important for DEX and

PMA responsiveness in the presence of the first intron. Theluciferase activity of the M4 mutant was reduced comparedto the � whereas the luciferase activity of MS was very

similar to that of the WI, suggesting that the region -610 to-298 bp contains negative cis-elements that may be acti-vated in the absence of the first intron. The mutant M6 had a

similar profile as the mutant M2. Finally, the very low lucif-erase activity of mutant M7 confirms the importance of theregion from -180 to -106 bp.

Effect of RU-486 on the Induction of Ax I Promoter. Todetermine whether GC effects required occupancy and/oractivation of the GR, we used the steroid analogue RU-486,which is a potent antagonist of GCs and which binds withhigh affinity to the GR. A 549 cells were transfected with thewr promoter of Ax 1 and treated with different stimuli. Table1 shows that RU-486 (1 p.M) inhibited DEX-mediated lucifer-ase induction by 45%. RU-486 treatment did not affect re-sponses induced by PMA (data not shown). However, RU-486 inhibited the combined action of DEX and PMA by 47%.We conclude that the GO effect requires GR occupancy (atleast with 50% efficiency) under these experimental condi-tions.

EMSA Using Oligonucleotides Designed According tothe Sequences of the Ax I Promoter. According to theresults obtained from the reporter gene assay, we performedgel retardation assays on the region spanning -1 80 to -106

bp of the Ax 1 promoter gene to identify potential trans/cistranscriptional factors involved. Three different oiigonucieo-

tides of 30 bp were synthesized (oligonucleotides A, B, andC) that covered the entire region. Nuclear proteins fromunstimulated or stimulated A 549 cells (PMA, DEX, and PMAplus DEX) were incubated with the different labeled oligonu-cleotides A, B, and C. Because we did not find any inducedcomplex after the treatment with either oligonucleotide A orC and because we observed no effect of DEX alone and nodifference between PMA and DEX plus PMA, we concen-trated our work on the sequence covered by oligonucleotideB and on PMA. The results are shown in Fig. 4. The incuba-tion of cells with PMA revealed the appearance of a complex(denoted by Ill in Fig. 4A) that was not present in the un-

stimulated cells (Fig. 4A, compare Lanes 3 and 2, respec-tively). The complexes designated I and II that were al-ready present in unstimulated cells were strongly inducedby the PMA treatment (Fig. 4A). Preincubation of the ex-

tracts with an excess of cold oligonucleotide B resulted in

the loss or strong reduction of all of the three complexesas compared to the control or to preincubation with anirrelevant oligonucleotide (Lane 6 versus Lanes 3 and 4).Taken together, these results show that, in the presence ofPMA, transcriptional factors that bind to a specific regionof the Ax 1 promoter spanning a 30-bp region (- 1 60 to

-130) are induced. A computer analysis of this regionshowed a sequence that is NFIL-6-like; examination of thelarger region spanning - 1 80 to - 1 06 bp of the Ax 1

promoter revealed 60% homology to the regulatory ele-ments of the haptogiobin promoter region, which is regu-lated by NFIL-6 (C/EBP f3). Consequently, we decided toanalyze the possibility that NFIL-6 could also participate inthe regulation of the PMA-induced transcriptional activa-

tion of Ax 1 . Fig. 4A (Lane 5) demonstrates that, afterincubation of nuclear extracts with an oligonucleotide de-

signed to bind NFIL-6, the complex II was partly revertedand the complex Ill was totally reverted.

The Ax I Oligonucleotide B Binds Complex C/EBP 13.To identify the nuclear protein(s) that could specificallybind to the region covered by oligonucleotide B, we incu-bated the nuclear extract with different antibodies: onedirected against C/EBP 13 (c-19), which does not cross-react with C/EBP a, C/EBP �, or C reactive protein 1 , andan irrelevant antibody (Ab 14). As shown in Fig. 4B (Lane

4 compared to Lanes 2 and 3), the antibody c-i 9, whichspecifically recognized C/EBP f3, was able to partly revertthe complex I and Ill, whereas an irrelevant antibody did

not (Lane 3). The intensity of the signal measured using thePhosphorlmager was reduced by 15 and 30%, respec-tively (Lane 4 compared to Lane 2). In addition, the pres-ence of a new labeled band (denoted with an asterisk) wasnoted on the same gel and could be due to a partialsupershift of the labeled complex ill. The addition of in-creasing amounts of anti-G/EBP p antibody failed to su-

pershift the entire complex detected in the nuclear ex-tracts from PMA-treated cells (data not shown), indicatingthe likely presence of an additional protein(s). These re-suits demonstrated that C/EBP f3 is an important compo-nent that contributed to the formation of the DNA-protein

12

-�

332 Annexin 1 in Epithelial Cell Differentiation

A complex between the consensus NFIL-6 sequence pres-Treatment ent in the Ax i promoter and nuclear extracts from PMA-

treated A 549 cells.Site-directed Mutagenesis and C/EBP trans-Activa-

F _ tion. To further support the data obtained with the EMSA

� analysis of the involvement of the NFIL-6 (C/EBP p) binding

. site within the Ax i promoter in the response to PMA, tran-

sient transfections were performed either with a luciferasegene driven by 880 bp of 5’ Ax i promoter [pAx 1(880)-Luc

I � WI] or with a luciferase reported construct [pAx 1(880)-Luc

II � NFIL-6 mut] that contained mutations within the NFIL-6-m binding site that had been identified by computer analysis

(see “Materials and Methods”).Two separate preparations of the mutated constructs were

. used. In unstimulated A 549 cells, luciferase activity wasmeasured in the cell extracts (Fig. 5) 18 h after transfectionwith either pAx i(880)-Luc WTor pAx 1 (880)-Luc NFIL-6 mut

(colonies 1 and 2). PMA treatment of pAx 1(880)-Luc WT-

transfected cells resulted in a 7-fold increase in luciferaseactivity, whereas cells transfected with pAx i(880)-LucNFIL-6 colonies 1 and 2 had only a i .5-fold increase. WhenA 549 cells were cotransfected with the expression constructcontaining the C/EBP 13cDNA (pMSV-GEBP/13), the reporteractivity of the pAx i(880)-Luc WT was increased 6-fold.Gotransfection of pAx 1(880)-Luc NFIL-6 mut resulted only ina 0.4-3-fold increase in activity.

PMA treatment of the A 549 cells transfected with pAx

1 (880)-Luc Wi- and pMSV-G/EBP �3 increased the luciferase

B activity by i 9-fold, whereas PMA treatment of A 549 cells4�s� transfected with pAx 1(880)-Luc NF!L-6 mut and pMSV-G/

Antibodies C9 EBP f3 increased luciferase activity only 0.3- and 3-fold,respectively (Fig. 5). Gotransfection with pMSV-G/EBP 13plus

PMA treatment resulted in a greater activation of the Ax i WTpromoter than that observed when PMA was omitted. Trans-fection with pMSV-G/EBP p demonstrates that PMA treat-

I ment resulted in an increased activity of the Ax 1 promoterII WT and that this increase can be mediated in part by a

m C/EBP f3 (NFIL-6) factor through the element previously iden-

tified by EMSA.

DiscussionHere, we show that PMA-induced differentiation of the A 549epithelial cell line, as measured by expression of DDP-iV, isassociated with an increased expression of Ax i . The addi-tion of the GC dexamethasone after incubation of A 549 cellswith PMA augments both the expression of Ax 1 and its

translocation to the cellular membrane, whereas dexametha-sone does not further induce expression of DPP-lV. These

labeled Ax 1 oligonucleotide B; Lane 4, 200-fold molar excess of nonspe-cific competitor oligonucleotide; Lane 5, 200-fold molar excess of oligo-

1 nucleotide NFIL-6; Lane 6, 50-fold molarexcess of cold oligonucleotide B.B, antibody analysis. One �g of antibodies directed against C/EBP fJ(c-19) or Ab 14 nonspecific antibody was added to the nuclear extract (see

Fig. 4. EMSA. A, nuclear protein extracts (1 0 �g) were made from un- “Materials and Methods”). Lane 1 , unstimulated cells incubated with thestimulated A 549 cells or from cells treated with PMA (10 n�+i, 24 h) and Ab c-19; Lane 2, nuclear extract from PMA-treated cells; Lane 3, PMAprocessed as described in “Materials and Methods.” Roman numerals, nuclear cell extract incubated with the antibody Ab 1 4; Lane 4, PMAthe different retarded bands formed. Lane 1, probe alone; Lane 2, nuclear nuclear cell extract in presence of the specific Ab C/EBP (c-19). Arrow,extracts from untreated cells incubated with labeled Ax 1 oligonucleotide partly reverted complexes are indicated. The results shown are one rep-B; Lane 3, nuclear extracts from PMA stimulated cells incubated with resentative of three independent experiments.

2000

II� 1000

0 i’�4-�’ (880) .�w’r

D � (880)4cc NFIL-6 met (col.1)� pAz 1 (880)4cc NFIL-6 met (coL2)

L�t� 6*

PMA(lOnM) + �+ +

pMSV.C/EBP� - - - + + +

Cell Growth & Differentiation 333

Fig. 5. C/EBP trans activation ofthe Ax1 promoter. A 549 cells were transientlytransfocted with p Ax 1(880) Luc-WT� orwith two different Isolated colonIes of thepAx 1(880) Luc NFIL-6mut. MSV-C/EBP13 vector was cotransfected into A 549cells. The cells were further treated withPMA (40 h). The luciferase index wasexpressed in luciferase activity abovethat of unstimulated cells. Columns,mean of three different experiments per-formed in triplicate; bars, SE.

+ + +

+ + +

results prompted us to analyze the effects of PMA and DEXon Ax 1 gene expression at the promoter level. Using severalapproaches, including transfection Into the A 549 celi8 of theAx 1 promoter coupled to a reporter gene and EMSA of thenuclear extracts, in control and treated cells, we report thatAx i transcription, already present at a high constitutivelevel, is augmented by PMA and DEX. Using deletion mu-tants and site-directed mutations of the Ax 1 promoter, weidentified a region spanning -180 to -106 bp of the Ax ipromoter that was responsible for the effect both of PMA andof DEX. In addition, we have identified NFIL-6 (C/EBP 13) as

the transcription factor that Is probably involved In the PMA

effect.The data reported here further confirm and extend the

involvement of Ax 1 in proliferation/differentiation that is al-ready reported in various tissues and models. These datashed some new light on the role of DEX in relation to differ-

entiation. Indeed, whereas DEX increases the expression ofAx 1 , alone and when combined with PMA, it has no effect onCD26 expression, suggesting that DEX does not up-regulatethe latter gene. The role of DEX on expression ofAx 1 protein,as already reported (3, i6), is not only to increase the intra-cellular pool of the protein but also to permit its translocatlonto the extracellular leaflet of the membrane. The functions ofthe augmented intracellular Ax 1 in the differentiation proc-ess are still unknown. However, it should be noted that PMAas well as DEX augments the low molecular weight forms ofthe protein. The identification of the low molecular weightforms is not clear and could reflect cleavage of the NH2-terminal end of Ax 1 or of additional, tyrosine-phosphory-lated forms. The extracellular form of Ax 1 is the only highmolecular weight form.

Our understanding of the gene regulation of Ax 1 is stillvery limited. The most extensive study thus far was per-

formed on the regulation of the two Ax i genes of thepigeon, cp 35 and cp 37 (i7-i9). The cp 37 gene isconstitutively expressed and is closely related to the mam-malian Ax 1 gene, whereas the cp 35 gene is only ex-

pressed in response to prolactin. Analysis of the mecha-

nism of prolactin induction of cp 35 gene suggests thattyrosine phosphorylation and subsequent activation of at

least one member of the STAT family of transcriptionfactors could lead to direct transcriptional activation of cp35, These data suggest that, in birds, two genes for Ax 1may have evolved, one giving rise to a constitutive proteinand the other giving rise to a prolactin-inducible proteininvolved in differentiation, In contrast, the single Ax 1 genein mammals Implies a potentially complex differential reg-ulatlon of a single protein product: constitutively ox-pressed, similar to the avian cp 37, but specifically up-regulated, similar to the inducible avian cp 35.

The data we have obtained with the human Ax I pro-

moter, indeed, demonstrate that there exist at least twosystems regulating transcription of Ax 1 : one constitutive

and the other inducible, Our data on the constitutive reg-ulation of Ax 1 promoter activity demonstrate that the firstintron, as well as the last 40 bp at the 5’ end of thepromoter, are important for the process of transcriptional

initiation, This high constitutive level of expression In vitrocould be explained by the presence of serum factors.Using DEX and PMA, we have found that the same gene isalso inducible. DEX alone in this cell line had a low distinctstimulatory effect, whereas PMA alone had a more pro-nounced effect, and the two agents together had a greaterthan additive effect in induction of total Ax I , as well as in

its subceliular localization.Although our results show that promoter activation by

DEX decreases when it is In the presence of the GAantagonist RU-486, the DEX regulation/response ofthe AxI promoter is independent of the GRE present in the firstintron because deletion of this GRE does not change itsresponse to any stimuli (data not shown). The consensussequence present in the first intron is not palindromic andhas been termed a “simple GRE”; It gives a slow responseto the endogenous steroids present in the serum. There-fore, the Ax 1 promoter could be considered a slow re-

334 Annexin 1 in Epithelial Cell Differentiation

sponder to DEX treatment. A number of DEX-induciblegenes, e.g. , phosphoenolypyruvate carboxykinase, albu-mm, and ai -acid glycoprotein, have been reported to

require protein synthesis for DEX induction. This type ofsteroid response is designated as a secondary GO re-sponse, i.e. , one that depends on protein synthesis. In-deed, the effect of DEX on Ax 1 promoter activation wasdiminished by actinomycin D (data not shown). Taken

together, these data indicate that DEX-mediated regula-tion of Ax 1 is a secondary response involving induction ofproteins. Transfection experiments using different deletionmutants showed that the response to DEX, similar to the

PMA response, was under the control of the region span-ning - 1 80 to - 1 06 bp of the Ax 1 promoter. However,

when EMSAs were performed using labeled oiigonucleo-tides spanning this region, no complex was found afterDEX treatment of the cells. These results suggest that theproteins induced by DEX participate in the promoter reg-uiation in a manner that has not yet been elucidated.

PMA is a strong inducer of the Ax 1 gene. Using deletionmutants and transfection experiments, we show that thefull-length promoter responds to PMA; we have mapped aminimal region of 80 bp that is important for the PMA re-

sponsiveness. This region, divided into three different sub-regions of 30 bp each and analyzed by EMSA, confirms the

mutagenesis data and indicates the involvement of induciblefactors that bind DNA or other protein(s) already present. Wedo not yet know whether these factor(s) are discrete ormultiple proteins or even proteolytic fragments of a single

protein that share similar DNA-binding domains. The patternof the complexes formed in PMA-treated cells is highly re-producible and specific because, under the same conditions,control cell extracts did not contain the same complexes. Inaddition, our data suggest that the transcriptional factorC/EBP 13 binds the specific region covered by oiigonucleo-tide B after PMA treatment. The site-directed mutagenesisdata in the region identified by EMSA and C/EBP trans-

activation of the Ax 1 promoter support our gel shift data on

involvement of a C/EBP �3 factor in up-regulation of the Ax 1

by PMA.Recently, it was reported that Ax 1 is involved in embryonic

palate development (20). Its striking immunolocalization tothe palate epithelium suggests that it plays a role in signalingdifferentiation of palatal epithelial cells.

The literature indicates that C/EBP (NFIL-6) and othermembers of this transcriptional factor family are involved inthe differentiation of several cellular systems, including bron-

chial and lung epithelial cells (21). NFIL-6 is involved in phor-bol ester induction of p-glycoprotein in U-937 cells (22), anddifferent C/EBP isoforms are active in the differentiation ofadipocytes (23).

The conclusion that expression of Ax 1 may be correlatedwith cell differentiation of is further supported by the finding

that certain cells in the central nervous system were onlyexpressed Ax 1 after they had stopped dividing and hadundergone terminal differentiation (24). Ax 1 overexpressionwas also found in well-differentiated hepatoceilular carci-noma tissue (1 3). This last report suggests that Ax 1 over-expression may be one of several factors that contribute to

malignant transformation, progression, and differentiation inhepatocellular carcinoma.

In parallel with the results reported here, we have shown,in A 549 cells, that Ax 1 synthesis and secretion are stimu-lated by interleukin 6 and that the transcription factor impii-cated in this regulation was also NFIL-6 (25). However, thepattern of the induced complexes was different, suggesting

that, although NFIL-6 participated in both cases, other pro-teins of the complexes may be different.

In conclusion, our data indicate that, although Ax 1 may beexpressed at a high constitutive level, in some cells, itsexpression is further regulated by differentiating agents aswell as by the inflammatory cytokine interieukin 6. An anti-inflammatory role of Ax 1 has been well documented, but arole for Ax 1 in differentiation, which now appears highlyprobable, is not yet well understood. Therefore, differentia-tion of A 549 cells may offer a good model for studying thehuman Ax 1 functions in differentiation and may shed light

upon the roles of this gene both in normal development andcarcinogenesis.

Materials and MethodsCell Culture and FACS. The human adenocarcinoma cell line A 549(kindly provided by Dr. Jamie Croxtall, The William Harvey ResearchInstitute, London, England) was cultured in monolayers In DMEMIF-12

(Life Technologies, Inc.) supplemented with 10% FCS (Life Technologies,Inc.). DEX, PMA, ONPG, ATP, and lucifenn were from Sigma Chemical Co.RU 486 was from Roussel-Uclaf (Paris, France). Cells were plated at 5 x105/plate in 10-mm plates the day after they were treated (or not) for 40 hwith 10 nM PMA. To assess the state of differentiation induced by thistreatment, CD26 expression (DPP-IV) was measured. Briefly, cells werefixed in 2% paraformaldehyde and incubated for 30 mm at 4#{176}C,thenwashed in 25 mM HEPES supplemented with I m�+i CaCI2 and MgCl2 andsaponin (0.025% from Saponaria species). Nonspecific binding was

blocked with human lgG (1 mg/mI) and incubation with a monoclonal

antibody against CD26 (Valbiotech)was performed as reported(16). FACSwas performed on a Epics-Elite (Coultronics) Flow Cytometer equippedwith an Argon ion laser beam operating at 488 nm using 15 mW of powerto excite the FITC and the propidium iodide. Logarithmic fluorescence

histograms (256 channels) were obtained from approximately 5000 viablecells for each sample. With the Elite 4.02 Data Analysis System fromCoultronics, mean channel number fluorescence was used to assess

differences in fluorescence intensity.Immunoprecipitation and Western Blot Analysis. A 549 cells were

treated with or without 10 np+iPMA for 24 h, followed by a further 16-hincubation with or without 1 p.M DEX. After this incubation, the proteinbound to the membrane was EDTA-extracted from the membrane, andthe cells, after being washed in PBS supplemented with 1 m�+i EDTAand a cocktail of protease inhibitors (Boehringor), were scraped In the

same buffer without EDTA. After three cycles of freeze-thawing, cellswere centrifuged to eliminate nuclei and broken membranes for pre-paring cytosol. Both the cytosol and the EDTA supematant were thenprocessed for immunoprecipitation as described. Immunoprecipitatedproteins were separated by 1 0% SDS-PAGE and transferred to a

nitrocellulose membrane (Bio-Aad). Immunostaining was performedusing a polyclonal antibody directed against the entire protein, and theimmunoreactive bands were detected using enhanced chemilumines-cence (Amersham).

DNA Transfectlons and Luciferase Assay. DNA transfections wereperformed by the calcium phosphate precipitation technique. Typically, atransfection experiment included 4 �g of reporter plasmid and 0.8 �g of

CMV-Lac Z plasmid used for controlling transfection efficiency (26). Cellswere harvested 48 h after transfection, and the luciferase activity wasmeasured using a Berthold Luminometer (LB 9501), as described by DoWet et aL (27).

Plasmid Constructions and Mutagenesis. An 880-bp fragmentcontaining the first nontranslated exon and 250 bp of the first intron

Cell Growth & Differentiation 335

19. Xu, V., and Horseman, N. D. Nuclear proteins and prolactin-induced

1CP35 gene transcription. Mol. Endocrinol., 6: 375-383, 1992.

was amplified by PCR from a AL28p1 1 phage genomic library (Ref. 28;a kind gift from Dr. Jeff Browning, Biogen, Cambridge, MA). ArtificialSail and HindlIl sites were introduced into the primers RPS5 and RPH3to facilitate cloning. Ollgonucleotides with engineered restriction sites

at 5’ and 3’ were used as primers for the PCRs used to obtain all of themutants. Eight different mutants, spanning the regions -610 to +270bp and -61 0 to +20 bp were constructed (see Fig. 3, A and B). The

mutants were designated by numbers corresponding to the nucleotidelength and the letter M followed by a number from 1 to 8, whereas thenonmutated construct is designated WT. The Ml mutant, pAx 1(568)-

Luc, contains the region spanning -298 to +270 bp; the M2 mutant,pAx 1(450)-Luc, contains the region spanning -180 to +270 bp; the

M3 mutant, pAx 1(376)-Luc, contains the region spanning -106 to+270 bp; the M4 mutant, pAx 1(630)-Luc, contains the region spanning-610 to +20 bp; the MS mutant, pAx 1(318)-Luc, contains the region-298 to +20 bp; the M6 mutant, pAx 1(200)-Luc contains the region-180 to +2Obp; M7 mutant pAx 1 (126)-Luc, contains the region

spanning -106 to +20 bp; the M8 mutant, pAx 1(76)-Luc, contains theregion from -56 to +20 bp. The PCR-amplified 880-bp segments fromthe Ax 1 promoter region and the mutants obtained were sequenced bythe Sanger method using the United States Biochemical Sequenase

kit. The promoter region and all of the mutants at the 5’ end were

cloned in a commercial mammalian vector containing the luciferasegene (pGL2-basic-Luc; Promega).

The WT pAxl(880)-Luc construct contained the sequence -180 5’-mGTMTGCCAG1TGM1TGGmAG-3’, whereas the mutant pAx

1(880)-Luc NFIL-6 mut construct contained the sequence -180 5-iTT-

GTAATGCCAGCCGMiTGGmAG-3’, where the underlined bases were

mutated by site-directed mutagenesis according to Kunkel eta!. (29). The

C/EBP 13expression vector (pMSV-CIEBP �) was a generous gift from Dr.Steven L McKnight (Carnegie Institution of Washington, Baltimore, MD;Refs. 23 and 30-32).

Nuclear Proteins Extracts and EMSA. Nuclear proteins for all of thegel shift experiments were prepared according to Schrelbereta!. (33). Gelretardation assays were performed as described previously (34). Briefly,

cell nuclear extracts (10 pg), as reported previously (10 n�+i PMA, 1 �.u+i

DEX, or PMA and DEX combined) were incubated with cold synthetic

oligonucleotides or the competitors (50 and 200-fold molar excess, re-

spectively) for 15 mm at room temperature, and then the 32P-labeled

double-stranded synthesized DNA oligonucleotide probe (0.1 pmol,100,000 cpm) was added, and samples were further Incubated for 20 mm

at room temperature. As a control, labeled oligonucleotides were tested

for intrinsic gel shift activity by incubation without nuclear proteins. The

oligonucleotides used for gel shift assay were synthesized and purified byGenset SA (Paris, France). The probe covering the sequences -298 to-106 bp was obtained by annealing the following oligonucleotides: oh-

gonucleotide A, S’-GGCCATAAAAGTCCAAAATMAGAGGGAGGMA-

AATCAGCCC’ITGGAAGCGG-3’; ohigonucleotide B, S’-GGCCTCCmG-TAATGCCAGiTGAAiTGGmAGTGAGTAGTAG1T-3’; ohigonucleotide

C, 5’-CCGGTGAGTA G1TIGCMGAC1TGTITCACmG-3’; and oligo-nucleotide NFIL-6, 5’-ACACCCGGTCACGTGGCCTACACC-3. The sam-pIes were mixed with a sample buffer containing 60 m� KCl, 4 m�

Tris-HCI, 12 mM HEPES, 1 mp.i MgCI2, 1 m� DTT, 1 m� phenylmethylsul-

fonyl fluoride, 1 mg/mI BSA, 1 mg/mI poly(dl-dC), 10% glycerol, and 1%Nonidet P-40 and run on a nondenaturating 6% pohyacrylamide gel in

0.5% Tris-borate buffer. After electrophoresis, the gel was fixed with 10%acetic acid, dried, and autoradiographed. Anti-C/EBP 13antibody (c-19)was supplied by Santa Cruz Biotechnology. Antibodies were added to the

nuclear proteins together with the labeled ohigonucleotide and further

incubated for 1 h at 4#{176}Cbefore gel loading and electrophoresis.

AcknowledgmentsWe thank Brigitte Canton for the technical assistance and BabetteWeksher (Cornell Medical College, New York, NV) and Anita Lewit-

Bentley (Laboratoire por L’Utihisation du Rayonnement Electromagne-tique, Orsay, Paris) for critical reading of this manuscript. We also

thank Dr. J. L. Browning (Biogen, Cambridge, MA) for providing theAL28p11 phage genomic library, Dr. J. D. Croxtahl (The William HarveyResearch Institute, London, England) for providing the A 549 cell line,

and Dr. Steven L. McKnight (Carnegie Institution of Washington,

Baltimore, MD) for providing the mammalian expression vector for

C/EBP 13.

References1 . Gerke, V. Annexins and membrane traffic. In: B. A. Seaton (ed), An-nexins: Molecular Structure to Cellular Function, pp. 67-75. Boston:

Springer-Verlag, 1996.

2. Fava, A. A., McKanna, J., and Cohen, S. Upocortin 1 (p35) is abundant

in a restricted number of differentiated cell types in adult organs. J. Cell.Physiol., 141: 284-293, 1989.

3. Sohito, E., Aaugei, G., Mehli, M., and Parente, L Dexamethasone in-

duces the expression of the mRNA of lipocortin 1 and 2 and the releaseof hipocortmn 1 and 5 in differentiated but not undifferentiated U-937 cells.FEBS Left., 291: 238-244, 1991.

4. Johnson, M., Kamso-Pratt, J., Pepinsky, A., and Whetsell, W. J. U-pocortin-1 immunoreactivity in central and peripheral nervous system ghial

tumors. Hum. Pathol., 20: 772-776, 1989.

5. Isacke, C., Undberg, A., and Hunter, T. Synthesis of p36 and p35 isincreased when U-937 cells differentiate in culture but expression is not

inducible by glucocorticoids. Mol. Cell. Biol., 9: 232-240, 1989.

6. Williams, F., Mroczkowski, B., Cohen, S., and Kraft, A. Differentiation ofHL-60 is associated with an increase in the 35-kDa protein lipocortin I. J.Cell. Physiol., 137: 402-410, 1988.

7. Leung, M-F., Un, T. S., and Sartorelhi, A. C. Changes in actin and

actin-binding proteins during the differentiation of HL-60 leukemia cells.Cancer Aes., 52: 3063-3066, 1992.

8. Violette, S. M., King, I., Browning, J. L, Pepinsky, R. B., Wallner, B. P.,and Sartorelhi, A. C. Role of hipocortin 1 in the glucocorticoid induction ofthe terminal differentiation of a human squamous carcinoma. J. Cell.Physiol., 142: 70-72, 1990.

9. Kang, S. A., Cho, Y. J., Moon, H-B., and No, D. S. Translocation ofhipocortin (annexin) 1 to membrane of U937 cells induced by phorbol

ester, but not by dexamethasone. Br. J. Pharmacol., 1 17: 1780-1784,1996.

10. Ma, A., and Ozers, L Annexin I and II show differences in subcellularlocalization and differentiation-related changes in human epidermal kern-tinocytes. Arch. Dermatol. Res., 288: 596-603, 1996.

11 . Sato-Matsumura, K., Koizumi, H., Matsumura, T., Ohkawara, A.,Takasu, T., Furuta, V., Sawa, H., and Nagashima, K. Localization of

annexin I (hlpocortin I, p35) mRNA in normal and diseased human skin byin situ hybridization. Arch. Dermatol. Aes., 288: 565-569, 1996.

12. Vishwanatha, J., Muns, G., Beckmann, J., Davis, A., and Rubinstein,I. Differential expression of annexins I and II in bovine bronchial epithehiahcells. Am. J. Respir. Cell Mol. Biol., 12: 280-286, 1995.

13. Masaki, T., Tokuda, M., Ohnishi, M., Watanabe, S., Fujimura, T.,Miyamoto, K., ltano, T., Matsui, H., A,ima, K., Shiral, M., Maeba, T.,Sogawa, K., Konishi, A., Taniguchi, K., Hatanaka, Y., Hatase, 0., and

Nishioka, M. Enhanced expression of the protein kinase substrate annexinin human hepatocehlular carcinoma. Hepatology, 24: 72-81 , 1996.

14. Croxtall, J. D., and Flower, A. J. Lipocortin 1 mediates the inhibitionby dexamethasone of the induction by endotoxin of nitric oxide synthasein the rat. Proc. Nath. Acad. Sci. USA, 89: 3571-3575, 1992.

15. Croxtall, J. D., Choudhury, 0., Tokumoto, H., and Flower, A. J. U-pocortin-1 and the control of arachidonic acid release in cell signalling.Biochem. Pharmacol., 50: 465-474, 1995.

16. Solito, E., Nuti, S., and Parente, L Dexamethasone-induced translo-cation of hipocortin (annexin) 1 to the cell membrane of U-937 cells. Br. J.Pharmacol., 1 12: 347-348, 1994.

17. Horseman, N. D. Annexin 1 genes: diversity and regulation. In: S. E.Moss (ed), The Annexins, pp. 23-34. London: Portland Press Ltd., 1992.

18. Goa, V., and Horseman, N. D. Structural and functional divergences

of the columbiad annexin-1 encoding cp 37 and cp 35 genes. Gene, 143:

179-186, 1994.

Annexin 1 In Epithehial Cell Differentiation

20. Chepnik, K. P., Shipman-Appasamy, P, Ahn, N., and Goldowltz, D,Developmental regulation of various annexins In the embryonic palate ofthe mouse, J. QranIgf�cleJ Qer�et. Dev. 8101., 15: 171-.181, 1995,

21 . Plotkewiti, V., Enerbach, 8., and Hedin, L Expression and hormonal

regulation of th� OCMT enhancer binding protein-a during differentiationof rat ovarian fgihicles. Endocrinology, 133; 23�7-�2333, 1993,

22. Combates, N., Kwon, P., Rzepka, A., and Cohen, D involvement ofthetranscription factor NFIL-6 In phorbolesterinduotlon of P-glycoprotiinin U937 cells. Cell Growth Differ., 8: 213-219, 1997.

23. Cao, Z., Umek, A. M., and McKnight, S. L Aegulated expression ofthree C/EBP Isoforms during adipose conversion of 3T3-L1 cells, GenesDev., 5: 1538-1882, 1991.

24. McKanna, J. A., and Cohen, S. The EGF receptor kinase substratap35 in the floor plate ofthe embryonic rat CNS. Science (Washington DO)’243: 1477-1479, 1989.

25. Solito, F, de Coupade, C., Pasente, L, Flower, H �J., and Russo’Marls,F. IL-S stimulates snnexln 1 expressk�n and trsnslecatk�n and suggests anew biological role sa class Uacute phase prutein, Qytoidno, in press, 1998,26. Cherrlngton, 4. M., and Mocaraki, �. 8� Human cytomegalovirus eltransactivates the a promoter-enhancer via an 18-base’pair repeat ole��ment. J. Viroi,, 65: 1435-1440, 1989.27. Ps Wet, J. M., Wood, K. V., Os Luca, M. Helsinki, 0, �,, arid Sub-remani, S. Firefly luclferaae gene: structure and expression in mammaliancells. Mol. Cell. Blol., 7: 725-737, 1987.

28 Kovaclc, R. T,, ‘flzard, R., Cate, A. L, Frey, A. Z., and Wallner, B. P.Correlation of gene and protein structure of rat and human Ilpocortin I.

Biochemistry, 30: 9015-4021, 1991.

29. Kunksl, T. A,, Roberts, J. D., and Zakour, R. A. Rapid and efficientsft.-sp.cmc mutagsnesis without phenOtypIC selection. Methods Enzy-mol,, 154: 387-382, 1987.

30. Gravis, B., Eissnman, R., and McKnlght, S. L DelIneation of tran-soriptional control signals within the Moloney murlne sarcoma vIrus longterminal repeat. Mol, Cell. Blol., 5: 1948-1958, 1985.

31 , Augsreau, P., and Wasylyk� B� The MLV and SV 40 enhancers have asimilar pattern oftranscrlption.l activatIon. Nucleic Acids Res., 12: 8801-

8818, 1984,

32, FrIedman, A., Landschuiz, W., and McKnlght, S. L CcMT/enhancer

binding protein activates the promoter of the serum albumin gene in

cultured hepatoma cells, Genes Day., 3: 1314-1322, 1989,

33, Sohrelber, E,, Matthlas, P., Muller, M., and Schaffner, W, Rapid do-tection of octamir binding proteins with “mInI-extracts,” prepared from asmall number of cells. Nucleic Acids Ri.., 17: 6419, 1989.

34, Wegenka, U. M., Buschmann, J., Luttlcken, C., Heinrich, P. C., andHorn, F, Acute-phase response factor, a nuclear factor bIndIng to acute-

phasi response elements, Is rapidly actIvated by lnte.leukln-6 at post-

translation level, Mol, Cell. Blol., 13: 276-288, 1993,