Localization of the Inducible Enhancer in the Mouse Interleukin-5 Gene That Is Responsive to T-cell Receptor Stimulation

By Peter F. Bourke, Barbara H. van Leeuwen, Hugh D. Campbell, and Ian G. Young

Transcriptional regulation of the interleukin-5 (IL-5) gene in T lymphocytes appears to be of central importance in the control of the eosinophilia characteristic of allergic re- sponses and certain parasite infections. Previous studies of IL-5 gene regulation have been hampered by the lack of a transfection assay, which detects the antigen-responsive en- hancer in the IL-5 promoter. Here we show that stable trans- fection of the Th2 clone D10.G4.1 and the T lymphoma EL4.23 with chloramphenicol acetyltransferase reporter gene constructs carrying the region to -3859 gives inducible expression with the known regulatory characteristics of the endogenous IL-5 gene. To facilitate detailed analysis of the promoter region, 3.9 kb of DNA sequence immediately up stream of the start of transcription was determined and the

hTERLEUKIN-5 (L-5) is one of a group of cytokines pro- duced by activated T lymphocytes that are involved in the

regulation of immune and inflammatory responses. L-5 ap- pears to play a central role in the regulation of the eosinophilia, which is characteristic of allergic diseases, such as asthma and certain parasite infections.’ L-5 is also active in regulating the growth and differentiation of murine B cells?4

IL-5 is produced by the Th2 subgroup of murine CD4+ T cells when stimulated by antigen or mitogens such as Conca- navalin A (Cod) , but relatively little is known about the mechanisms governing its regulation. Unlike some other T- cell-produced cytokines, such as L-3, IL-4, and granulocyte- macrophage colony stimulating factor (GM-CSF), the expres- sion of IL-5 is resistant to the immunosuppressive drug cyclosporin A5 suggesting that the antigen-responsive en- hancer in the IL-5 gene may be linked to the T-cell receptor via a different signalling pathway. Recent work in our labora- tory has shown that IL-5 gene expression can be selectively induced in the Th2 cell clone D10G4.1 by treatment with phorbol 12-myristate 13-acetate (PMA) or by stimulation of the IL-1 receptor or CD45, without induction of the IL-3, L- 4, or GM-CSF genes.6 Thus, although the IL-5 gene is part of a cytokine gene cluster on mouse chromosome 11 with L- 3, L-4, and GM-CSF,’ the regulation of expression of the L-5 gene must involve gene-specific mechanisms. The use of nuclear run-on transcription assays has established that the IL-5 gene in D10G4.1 cells or the T-cell lymphoma line EL4.23 is regulated primarily at the level of

The lack of a transfection assay, which measures inducible expression of the IL-5 promoter in response to T-cell activa- tion signals, has greatly hampered studies of IL-5 gene regu- lation. In the present work, we show that stable transfection assays with the T-cell clone, D10.G4.1 and the T-cell lymphoma line, EL4.23 can be used to study the transcrip- tional regulation of the IL-5 promoter. We have used this assay to localize the inducible promoter/enhancer in the IL- 5 gene and have demonstrated the involvement of a CTF/ NF1 site and a TCATTT-containing element in the regula- tory mechanism.

I

MATERIALS AND METHODS

Cell lines. The T-cell clone, D10.G4.1 was obtained from Dr T. Mosmann, DNAX Research Institute of Molecular and Cellular

Blood, Vol 85, No 8 (April 15). 1995: pp 2069-2077

minimum upstream region required for inducible expression was further localized, by stable transfection studies in EL4.23 cells, to the region up to -1016. A CTFINF1 site in the up- stream enhancer at -940 to -928 was shown to be required for regulated inducible expression. Mutation of this se- quence motif abolished inducibility and also prevented bind- ing of the sequence to a nuclear protein(s). A TCATTT-con- taining element in the proximal promoter region was also demonstrated to be essential for inducible expression of the IL-5 gene, similar to the role of this conserved element in the transcriptional regulation of the granulocyte-macro- phage colony-stimulating factor (GM-CSF) and IL-4 genes. 0 1995 by The American Society of Hematology.

Biology, Palo Alto, CA. EL4.23 is a subline of the murine lympho- sarcoma EL4 and was obtained from Dr C. Sanderson, National Institute for Medical Research, London, UK.

Sequencing. A IO-kb HindIII-Hind111 mouse (BALBk) DNA fragment containing the IL-5 gene was cloned into pSV2neo (pSV2neo-lOkbmIL-5G). The 3884-bp HindIII-Eco47III fragment (Fig 1A) was sonicated and subcloned into M13.9 Sequence of M13 subclones and alkali-denatured double-stranded plasmid DNA” was obtained by the chain termination method.” Both strands were com- pletely sequenced. DNA sequences were assembled and analyzed using the programs described by Staden.l2.’’ The L-5 gene 5’ flank- ing sequence was compared with the GenBank nucleotide sequence database and databases of known motifs to identify conserved se- quence elements. All L-SCAT mutants were sequenced by cycle sequencing using polymerase chain reaction (PCR).

Constructs for expression studies. The 10-kb HindIII-Hind111 IL-5 gene fragment (-3859 to +6181) was cloned into the promot- erless vector pSW3 (Promega, Madison, WI) after insertion of an EcoRV linker into the BstEII site in the 3’ untranslated region (creat- ing pSP73.rnIL-SG(EcoRV)) to distinguish expression from that of the endogenous gene. The 6.5-kb Xba I-Hind111 fragment (-312 to +6181) from pSP73.mIL-SG(EcoRV) was cloned into pCEXV3I4 to create pCEXV3.mIL-SG(EcoRV).

The IL-5 gene fragment encompassing -547 to +74 was obtained by PCR using primers corresponding to the known sequence of the gene15 plus additional sequence to generate a HindIII site at its 5’ end and an EcoRI site at its 3‘ end to facilitate cloning. This fragment was cloned into M13mp19 and its sequence confirmed. The blunt- ended 572-bp HindIII-Eco47III fragment was then cloned into the blunt-ended HindIII site of pSVOCAT“ yielding a construct, which has 547-bp of IL-5 genomic sequence 5’ of the transcriptional start

From the Division of Biochemistry and Molecular Biology, John Curtin School of Medical Research, The Australian National Univer- sity, Canberra, Australia.

Submitted June 13, 1994; accepted December 2, 1994. Address reprint requests to Ian G. Young, PhD, Division of Bio-

chemistry and Molecular Biology, John Curtin School of Medical Research, The Australian National University, GPO Box 334, Can- berra, ACT 2601, Australia.

The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. section 1734 solely to indicate this fact. 0 1995 by The American Society of Hematology. oooS-497//95/8508-0030$3.00/0

2069

2070 BOURKE ET AL

A 1 kb

-3859 -547

I I I I I I I 1 I I I - H SG A A A X T

0 B F PT, 0

0

0 0 +l

/

I I t, I I

C X N C E A

B

CAGGGAGAAA CATTCAGGCT CCTTCCAGCT TGATTCCTCT GCATCTTGAA ACTAGAGAGT GCTGAGGGAC TGAATTGTTC

TGATTGAACA ATGAGAAGAG CACGACTGGA ACCCGGTGCA GGGGAGCATG CCTGACACAG TGCTTGAGAG GCTGAGGCTG

AGGCAGGAGC ATCAGAAGTT CATAGACAGT TTGGGCCACA TAGTCAAGAC TGTGTCCTAA TAAACCCACC ACAAATGAAA

GACAGAAGGA AAAGGAGAAG AAATGCGAAC CTGCTGTGTG TGACAAGATA CAGAGAGAGC CAGGAGTCCT GTTGTULCCC

TGGAGCACCC CTCATACACA CAGTAAGCTG CTTGAAAAGT GGGTCAAGGC CAAGCCACCA AAGCAGTACA GGAAGGAGCC

TCTGCCCTGC ACTTTTTGTC ATAGGAACCA AGAAGCATCC TCTTATCTCC ACTCAGTCTG GGTTGGCCTT TTAGTAACAG

ACAGGATGCT GATCAAAAAC CTACATCATG CAACTCAACT CTGTCAGACC AGCTGCTTTA ATAAATGTCT TTCCCAGAGA

GTTTTCAATT CTTTTTTTCT TTTCTTTTTT TAAGCACAGA TAAGTCTAGC TACCGCCAAT ATAGAAAAAA AAACACCCCC

TTTCCCTTTT CTCTTTGGCA GTTTGGAGCA GGGTCTTGTG TGTTCCAGAT TGAACTCACC TTTGCTGTAC ACACAAAAAA

GCCAAAGATG CTGTCACCTC TTGATCCTCC TGACTGTACC TCCCACATCT GCTGGTGTGT ACCACCACAC CTAGTAAGAT

ATTCTCAACA TTTATGTATT TTAGCCTAAC CCTGTTGGAG GTATACATTT GAATACATTT TTTCTCACTT TATCAGGAAT

TGAGTTTAAC ACATATTAAA GCAGGTGTGG GGCAGGGAGG GGGGGATAAA AAAGAAGGTG CTCAAGAAAA GCCGATCACG

CTCCCAAGAG TGTGAGCATG GGCGTCTCTA GAGAGATCCG CCATATATGC ACAACTTTTA AAGAGAAATT CAATAACCAG

AATGGAGTGT AAATGTGGAT CAAAGTTGTA GAAACATTCT TTTATGTTAT AGAAAATGCT TTTTAAGCAG GGGTGSGGGT

CAAGATGTTA ACTATTATTA AAGAGCAAAA AAAAAAAAAT GCATTTTGTT TGAAGACCCA GGGCACTGGA A A C C C T E C K l

T X G G A C T C GCCTTTATTA GGTGTCCTCT ATCTGATTGT TAGCAATTAT TCATTTCCTC AGAGAGAG-TTGCTT

GRE

C T F f NF1

K T CU

t(lE T a m T I T I

GGGGATTCGG CCCTGCTCTG CGCTCTTCCT TTGCTGAAGG CCAGCGCT

(CAP) site and 25-bp 3‘ of the CAP site (-547IL-5CAT). The largest IL-SCAT construct contained -3859 to +25 of IL-5 genomic sequence relative to the CAP site (-3859IL-5CAT) and was simi- larly created using the 3884-bp HindIII-Em47111 fragment from pSV2neo-lOkbmIL-5G.

A number of 5’ truncation mutants were obtained by restriction endonuclease digestion and religation of these two constructs. The 5‘ end of the IL-5 sequence in -547IL-5CAT has an Nhe I site, which was generated on ligation of the two blunt-ended HindIII ends. Double digests were performed with Nhe I and four other restriction enzymes unique to the -547-bp of IL-5 sequence. The enzymes employed were Acc I creating -458IL-5CAT, XbaI creat- ing -312IL-5CAT, Nsi I creating -140IL-5CAT. and EcoNI creat- ing -88IL-5CAT. Digestion of -3859IL-5CAT with Nhe I removed a fragment of 2546-bp in length and yielded - 1313IL-SCAT. Nhe YSph I, and Nhe bTthl11I double digests were used to create - 1170IL-5CAT and - 1089IL-5CAT, respectively. These restriction enzyme sites are shown in Fig 1A. Exonuclease 111 was used to generate the remaining truncation constructs from -3859IL-5CAT using the F’romega Erase-a-Base system.” -3859IL-5CAT was di- gested with Sac I to prevent digestion of the vector with exonuclease 111, and Age I1 (-3656) or Bgl I1 (-1483) before treatment with exonuclease I11 to generate the deletion mutants. Each construct was sequenced to determine the 5’ IL-5 terminus using a 17- mer corresponding to sequence 5’ of the HindIII cloning site of pSVOCAT.

The CTFVNF1 site at -940 to -928 and the T C A m site at -49 to -44 were mutated by oligonucleotide-directed in vitro mutagene- sis in the - 1170IL-5CAT construct. Briefly, a 1199-bp Sph I-

-1220

-1140

-1060

-980

-900

-820

-740

-660

-580

-500

-420

-340

-260

-180

-100

-20

t29

Fig 1. Map and sequence of the 5’ untranslated region of the mouse IL-5 gene. (A) Restriction enzyme map of the 5’ flanking sequence of the IL-5 gene. Re- striction enzyme sites are Accl (A), Bglll (B), €CONI (Cl, Ecd7111 (E), Nhd (F), Agd (G), Hindlll (HI, NsA (NI. SpM (PI, Sad (SI. Tthll l l (TI, Xbd (X). (B) Nucleo- tide sequence of the IL-5 gene from -1299 to +29. Underlined are the TATA box, the T C A l T motif (IL-5/GM-CSF conserved region), the CKl and CK2 se- quences, possible glucocorticoid response elements (GRE), and the potential binding sites for the CTF/NFl and OCT transcrip- tion factors. The * marks the start of transcription.

Eco47III fragment of the IL-5 gene 5’ sequence was directionally subcloned into Sph I-Sma I cut phagemid pTZ1SU (Bio-Rad, Syd- ney, Australia). Mutagenesis was performed in Escherichia ( E coli) strain U1032 (dut-, ung-)’* as described previously.’’ The resulting double stranded heteroduplex phagemids were transformed into E coli strain MC1061.1 (dut+, ungC).’O Mutagenized fragments were then isolated by Sac I-Hind111 digestion, blunt-ended and cloned back into the blunt-ended HindIII cloning site of pSVOCAT. These plasmids were confirmed by sequencing. Supercoiled CsCI-purified plasmid DNA was used in transient transfection experiments. For stable transfection, L-SCAT construct DNA was linearized with Bgl I, while pMClneo PolyA DNA” was linearized with Sca I. DNA was then extracted with phenol-chloroform and ethanol precipitated before electroporation.

Reverse transcriptase-polymerase chain reaction. RNA was prepared from cells according to the method of Chomczynski and Sacchi” and cDNA prepared by conventional methods. The primers used to detect expression from pSP73.mIL-5G(EcoRV) and pCEXV- 3.mIL-5G(EcoRV) were PB19 and PB26. The 5‘ primer PB19 (ATT- GACCGCCAAAAAGAGAAGTG) spans exons 3 and 4 in the IL- 5 gene sequence and only detects mRNA. The first 15 bases of the 3’ primer PB26 (GTTCAGGGTCACAGATATCGTG) correspond to the sequence of the EcoRV linker. These primers amplify a frag- ment of 587-bp and do not amplify endogenously expressed IL-5 mRNA. PCR reactions were analyzed on agarose gels.

Cell culture. EL4.23 and DIO.G4.1 cells were maintained in RPM1 1640 (GIBCO, Ontario, Canada) medium, 10% fetal calf se- rum, with the following additives: 10 mmol/L NaC1; 1 mmol/L sodium pyruvate; 62 mUL monothioglycerol; 100 U/L benzylpeni-

TRANSCRIPTIONAL REGULATION OF THE IL-5 GENE 207 1

cillin (Commonwealth Serum Laboratories [CSL], Melbourne, Aus- tralia); 100 m g L streptomycin (Sigma, St Louis, MO); 2 mmolL L-glutamine (CSL) at 37T , 5% CO,, and 95% relative humidity. The RPM1 1640 medium for the T-cell clone D10.G4.1 also con- tained 0.05 mmoK p-mercaptoethanol (Eastman Kodak, Rochester, NY) and IL-2. The L 2 was either IO units/mL of human IL-2 (Amersham, Buckinghamshire, UK) or conditioned medium from a cell line transfected with a mouse IL-2 construct.23 DIO.G4.1 cell growth was induced by allogeneic stimulation every 10 to 14 days with irradiated spleen cells from 6 to 8 week old C57B1/6 female mice.24 D10.G4.1 cells were treated with one or more of the follow- ing as indicated: 5 pg/mL ConA, 10 nglmL PMA, 3D3 antibody” at 1 in 1,OOO dilution, 0.1 pg/mL dexamethasone or 1 pg/mL cyclosporin A. EL4.23 cells were stimulated with 10 ng/mL PMA where indicated.

Transient transfection. Cells were resuspended (at 1 X lo6 to 2 X 106/mL) in fresh medium the day before electroporation. Transfec- tion was performed only if greater than 90% cells were alive. A total of 2 X IO’ cells in complete medium were mixed with 30 pg of supercoiled IL-SCAT construct DNA in a total volume of 400 pL and electroporated using a Bio-Rad Gene Pulser and Bio-Rad Capacitance Extender. Optimum parameters were 250 V at 500 mi- crofarad (pF) for EL4.23 and 300 V at 500 pF for DIO.G4.1. The cells were immediately transferred to 20 mL of fresh medium and allowed to recover for 16 hours. The transfected cells were divided in half and PMA was added to one half for EL4.23, or ConA for DlO.G4.1. The other half of the cells was retained as the untreated control. Cells were harvested after 12 to 16 hours for chlorampheni- col acetyltransferase (CAT) analysis. The number of viable cells was routinely checked at the time of harvest.

Srable transfection. D10.G4.1 cells were electroporated 9 to 11 days after allogeneic stimulation. The day before transfection, DIO.G4.1 cells were resuspended at 1 X IO6 cells/mL, and EL4.23 cells were resuspended at 5 X IO5 cells/mL in fresh medium. Trans- fection was performed only if greater than 90% cells were alive. A total of 2 X IO’ cells were mixed with 30 pg of linearized IL-SCAT construct and 3 pg of linearized pMC 1 neo PoIyA in 400 pL complete medium. Electrical parameters for electroporation of each cell line were the same as those for transient transfection. Immediately after electroporation, cells were transferred to 30 mL of fresh medium. D10.G4.1 cells were allogeneically stimulated the day after electro- poration. Selection with the neomycin analogue G418 (400 pg/mL, GIBCO) was commenced after a further 2-day incubation. EL4.23 cells were incubated for 2 days after electroporation before com- mencing G418 selection (800 pg/mL). Selection with G418 was continued for 10 days, at which time all mock (no DNA) electropo- rated control cells were dead.

Selected polyclonal EL4.23 cells were then grown for 2 days without G418 before stimulation. The number of individual clones in the polyclonal populations was estimated from the growth rate of the transformed cell lines. The doubling time of transformed (3418- resistant EL4.23 cells was 15.5 hours and was not significantly af- fected by the presence of G418. By extrapolating, from cell numbers at the time of PMA stimulation, back to the day of electroporation, the number of original transfectants (and, therefore, the number of integration events after transfection) was estimated and found to vary from 350 to 1,500 for the different transfection experiments. This range is a minimum estimate of the base complexity of the polyclonal population because it assumes that the transfected cells grow immediately at their normal rate.

Selected polyclonal D10.G4.1 cells required 3 to 4 weeks growth before sufficient cells were available for the first stimulation and then a further 2 weeks growth for the duplicate stimulation. The estimation of the number of individual clones in the polycbnal populations was more difficult for these cells, as the doubling time

of the cells decreased to 18 hours after allogeneic stimulation and then increased to 48 hours over the next 12 days. However, by basing the estimation on the overall growth between allogeneic stim- ulations (60- to 100-fold), it was possible to estimate that the mini- mum of number of transfectants was at least as many as estimated for EL4.23 cells.

For CAT assay of the stably transfected polyclonal cells, 1 X IO’ cells were stimulated and harvested as for transiently transfected cells.

CAT assay. After decanting the cells, the tissue culture flask was rinsed once with IO mL of phosphate-buffered saline (PBS) containing 5 mmoVL EDTA, which was pooled with the culture sample. Cells were washed with PBS and then resuspended in 150 pL of 0.25 moVL Tris-HCI, pH 8.0. Samples were sonicated on ice, incubated at 65°C for 10 minutes and immediately centrifuged at 4°C for 10 minutes. The supernatant was assayed for CAT activity using the butyryl coenzyme Miquid scintillation counting (LSC) CAT a~say.2~

Gel-rerardation assay. Nuclear extracts were prepared by a modification of the method of Miskimins et aIz6 from unstimulated EL4.23 cells and cells stimulated with IO n&mL PMA for 6 hours. The cells were washed twice in PBS and resuspended at 3.5 X IO’ cells/mL in lysis buffer (10 mmoK Hepes pH 8.0,50 mmoK NaCI, 5 mmoYL MgClz, 0.5 m o m sucrose, 0.1 mmoVL EDTA, 0.5% Triton X-100, 1 mmol/L D”, 0.1 mmoVL phenylmethylsulfonyl fluoride [PMSFI). The nuclei were pelleted at 500g for 1 minute, and resuspended in the lysis buffer, on ice, at 7 X IO’ celldmL. Spermidine was added to 5 mmol/L and NaCl to 0.5 moK. The suspension was incubated on ice for 30 to 60 minutes and then centrifuged at 200,OOOg for 15 minutes. The supernatant was care- fully removed and dialyzed against IO mmoK HEPES, pH 8.0, 50 mmol/L NaCl, 1 mmol/L MgCIz, 1 mmoVL DTT, 50% glycerol at 4°C for 18 hours. The nuclear extract was then aliquotted, immedi- ately frozen in a dry ice-ethanol bath, and stored at -70°C. The protein concentration of the extract was determined using the Bio- Rad Protein Assay kit.

Complementary oligonucleotides were synthesized on an Applied Biosystems (Foster City, CA) Model 380A DNA Synthesizer and purified by high performance liquid chromatography (HPLC). The sequence of the oligonucleotides included AATT at the 5‘ ends to generate single-stranded ends after annealing and to enable labeling by the fill-in method. Oligonucleotide duplex CTF-I is two comple- mentary 37-mers containing the IL-5 gene sequence from -946 to -915 (Fig IB); oligonucleotide duplex CTF-2 is the same sequence except for the mutations at the CTFNF1 site. Oligonucleotide duplex RS is two complementary 38-mers of random sequence (5’-AAT- TCCCTTGTCTGAGATATACACTAGCTTITAGCCCG-3’). Pairs of oligonucleotides were annealed by mixing 6 pmol of each in Klenow buffer (50 mmolR. Tris-HC1 pH 7.2, IO mmol/L MgSO,, 0.1 mmoVL D m ) containing 50 pg/mL bovine serum albumin (BSA). The annealed oligonucleotides were labelled with [CY-~~P]- dATP by filling in the single-stranded end regions.

The gel retardation assay was carried out according to the method of Singh et al.” Briefly, 10 pg of nuclear extract was added to the labelled DNA in binding buffer (IO mmoVL Tris-HC1 pH 7.5, 50 mmoVL NaCI, 1 mmom DTT, 1 mmoYL EDTA, 0.05% Nonidet P-40, 5% glycerol) in a total volume of 25 pL and incubated at room temperature for 30 minutes with 15 fmol of labelled annealed oligonucleotide duplex. Poly(d1-dC).poly(dI-dC) ( l .2 pg, Phar- macia) was included as nonspecific DNA. After binding, 5 pL of the reaction mixture was loaded onto a polyacrylamide gel (6%; acrylamide:bis, 55:l) in l X TAE buffer (6.7 mmol/L Tris-HCI pH 7.5, 3.3 mmoVL sodium acetate, 1 mmoVL EDTA). The gel was preelectrophoresed in 1 X TAE for 30 minutes at 11 Vlcm, and after

2072

sample loading was electrophoresed at IS V/cm for 60 minutes at room temperature. The gel was dried and then autoradiographed.

GenRank durubuse submission. Nucleotide sequence of the S' flanking region of the mouse IL-S gene will be submitted to the GenBank database.

RESULTS

Determination of S'Jlrmking sequence of mouse IL-5 gene. A sequence of 6727 bp covering the mouse IL-5 gene was determined previously in this 1ab0ratory.l~ This sequence included 547 bp of the 5' flanking sequence of the IL-5 gene. Because upstream enhancers involved in transcriptional reg- ulation may be several kilobases away from their proximal promoter elements, we determined an additional 5' flanking sequence of the mouse IL-5 gene to -3859. The additional sequence information was used to facilitate the generation of reporter constructs carrying different lengths of 5' flanking region for attempted localization of the IL-5 enhancer re- sponsive to T-cell receptor stimulation. Database searching of the new sequence showed the presence of murine B 1 and B2 Alu-like elements at -2538 to -2403 (plus strand) and -1.530 to - 1374 (minus strand), respectively. Because the functional studies described below located the promoteden- hancer to the region to - 1016, a more detailed analysis of this region of upstream sequence was performed for potential transcription factor binding sites. A few sequence elements were identified in this region of the gene, which closely matched consensus binding sites of known transcription fac- tors. These include a CTF/NFI site2* at -940/-928 (TGGGTCAAGGCCA, consensus TGGCANNNTGCCA), an Octamer binding site29 at -453/-446 (ATTTGAAT, con- sensus ATGCAAAT), two possible glucocorticoid response elements, which may be involved in the dexamethasone re- pression of IL-5 gene expression,' and the conserved cyto- kine sequences CKI and CK23" (Fig IB). The TCATTT motif, which is conserved in the IL-5 and GM-CSF proximal promoter regions,Is is also shown in Fig IB.

Transient expression assays using reporter genes. Two cell lines, which show inducible expression of the IL-5 gene, were used to test expression of the reporter gene constructs. D10.G4.1'4 is a Th2 murine T-cell clone that requires T-cell receptor stimulation for growth and, of the cell lines used, is the most closely related to normal T lymphocytes. EL4.23 is a murine T-cell lymphoma line. Regulation of the IL-5 gene in each cell type is predominantly transcripti~nal.~~~ IL-5 mRNA expression is induced by ConA, PMA, or the antiidiotype specific monoclonal antibody 3D3 in D10.G4.1 cells." and by PMA in the case of EL4.23 cells.'

A series of constructs carrying different lengths of the 5' flanking region of the IL-5 gene linked to the CAT reporter gene was prepared and electroporated into the two cell lines, and assays for transient expression of CAT were performed. Optimum conditions for electroporation were determined for each cell line using control plasmids such as pRSVCAT.'" Using these conditions, the positive control pRSVCAT gave good levels of expression in each cell line, but no inducible expression was seen with any of the reporter constructs (data not shown). This result was confirmed in three separate ex- periments. Doubling the amount of DNA used for the

BOURKE ET AL

A Dl 0.G4.1

EL4.23

0 unstimulated stimulated

Fig 2. CAT activity of the IL-SCAT constructs after stable transfec- tion into D10.G4.1 cells or EL423 cells. DNA was stably transfected into D10.G4.1 cells (A) or EL423 cells (B). D10.G4.1 cells were stimu- lated with 5 pg/mL ConA and EL4.23 cells with 10 ng/mL PMA. The CAT activity of each construct is expressed relative to the activity of pRSVCAT in stimulated cells. The CAT activity is the average of two stimulations of the polyclonal stable cell lines, and the bars indicate the variation between these stimulations.

-38591L-SCAT construct to 60 ,&transfection did not af- fect the result.

There are a few possible explanations why the -38591L- SCAT construct was not expressed after transient transfec- tion into cell lines, which inducibly express IL-5.h.X One possible explanation is that the required regulatory elements are not present in the construct. To test if the necessary sequences are present in an intron or the 3' untranslated region of the gene, whole gene constructs were prepared. The IO-kb HindIII-Hind111 gene fragment (including 5' se- quence to -3859) was modified by insertion of a linker into the BstEII site in the 3' untranslated region of the IL-5 gene and cloned into the promoterless vector pSP73 (Promega), so that the expression of the gene construct could be measured independently of that of the endogenous gene by RT-PCR.

TRANSCRIPTIONAL REGULATION OF THE IL-5 GENE 2073

Fig 3. CAT activity of -38591L-5CAT after stable trans- fection into D10.G4.1 cells. Cells were treated with various com- binations of 5 pg/mL ConA, 3D3 antibody, 0.1 pglmL dexameth- asone, 1 pg/mL cyclosporin A, and 10 ng/mL PMA, as indicated.

1.2

.- b 1.0 .- - > 0 c 0.0 4 0 %!

a a, 0.4

0.6 .- - c m

0.2

0.0 -38591L5CAT

A control plasmid was also constructed, which carried the modified 1L-S gene (including S' sequence to -312) in pCEXV3, so that expression of the IL-S gene was under control of the SV4O early promoter. No inducible expression was detected after transient transfection of pSP73.mIL- SG(EcoRV) into EL4.23 cells. As expected, constitutive ex- pression of the IL-S gene was detected when the pCEXV3.- mIL-SG(EcoRV) control plasmid was transiently transfected into EL4.23 cells. The lack of inducible expression with pSP73.mILSG(EcoRV) was confirmed in four separate ex- periments.

The failure of the IO-kb gene fragment to give inducible expression in transient assays suggested that either the neces- sary regulatory sequences were located further away from the gene, or that they do not function in the absence of the normal chromatin environment. The latter possibility was tested by stably incorporating the CAT reporter constructs into the host genome.

Demonstration of the inducible IL-S protnoter/enltnncer using stable transfection assays in DIO.G4. I cells. Stable transfection assays were performed on the different reporter gene constructs using linearized plasmid DNA cotransfected with pMC1neo PolyA. The latter plasmid carries the tleo ge'ne under the control of the herpes simplex virus thymidine kinase promoter allowing stable transfectants to be selected using (3418. Polyclonal lines consisting of a minimum of three hundred independent transfectants were generated to minimize any influence of the site of integration and copy number on expression. Copy number and expression levels in stable transfectants are often uncorrelated,3'.33 so i t is not possible to completely standardize experiments of this type. Our criterion for demonstrating the presence of the IL-S enhancer was consistent high level inducible expression (minimum of IO-fold induction) in independent experiments.

When the IL-SCAT reporter constructs previously used in transient assays were stably transfected into DIOG4.I cells. the constructs carrying up to -S47 showed varying levels of constitutive expression but the -38S9IL-SCAT construct was highly inducible (Fig 2A). Other detailed studies in this laboratory have shown that the IL-S gene is expressed in DlO.Ci4.1 cells in response to stimulation with ConA or the antiidiotype specific monoclonal antibody 3D3.'.* In contrast

0 unstimulated

ConA

ed dexamethasone

303 antibody + dexamethasone

0 cyclosporin A 3D3 antibody + cyclosporin A

PMA

PMA + dexamethasone

Con A + dexamethasone

3D3 antibody

to the IL-3, IL-4, and GM-CSF genes, the IL-S gene is also stimulated in these cells by PMA, and its expression is resistant to inhibition by cyclosporin A." The expression of the -38S91L-SCAT construct in DlO.Ci4.I cells was also shown to have these characteristics (Fig 3).

IL-S gene expression in DIOG4.I cells is also strongly inhibited by the antiinflammatory drug dexamethasone: just a s i t is in human peripheral blood mononuclear cells." Ex- pression of the -38591L-SCAT construct induced by either ConA, 3D3 antibody, or PMA was strongly inhibited by dexamethasone (Fig 3). However, the constitutive expression shown by the shorter constructs. such as -5471L-SCAT, was not significantly affected by dexamethasone (data not shown).

Therefore, the upstream region of the mouse IL-S gene to -3859 carries an inducible enhancer responsive to T-cell receptor stimulation and confers all of the known regulatory properties of the endogenous gene.

Loccdixrtiorl of the irlrl~rcihle IL-S prr)tno,er/enhnrIL.er L I S -

iug EIA.23 cells. Comparative studies of the mechanisms regulating the IL-S gene and several other lymphokine genes in DIOG4.I cells and EL4.23 cells".x have shown that tran- scriptional regulation is predominant in both cells and that the regulatory mechanisms have very similar characteristics. EL4.23 cells were used for more detailed localization of the enhancer because these cells are much easier to cultivate and to transfect than the T-cell clone D1 OG4. I . Stable trans- fection assays showed that the -38S91L-SCAT construct also gave inducible expression in EL4.23 cells, whereas the expression of constructs carrying up to -S47 showed no significant inducibility (Fig 2B). Thus, in this respect, the reporter constructs showed analogous expression character- istics in EL4.23 and DlO.Ci4.I cells, although the magnitude of expression levels relative to pRSVCAT was generally greater for the shorter constructs in EL4.23 than in D IO.G4. I cells.

A number of S' deletion constructs were prepared using existing restriction endonuclease sites or exonuclease 111 de- letion as described in Materials and Methods. Stable trans- fection assays were performed with the aim of establishing the minimum upstream region required for inducible expres- sion. Representative data on the localization of the enhancer

2074 BOURKE ET AL

B. 1 .o

p 1 .o 2 V 'g 0.8 .- I

.-

c 0.8 a 5 0.6 2 0.6 0 In

2 0.4 3 0.4

0.2 0.2

0.0 0.0

m

> .- - a:

- a: al

-3502 -3461 -3144 -1540 -1089 -547 -3859 -1016 -912 -853 -808 -749 -547

3.0 1 C.

.- 0 2.5 5 .- - 0 2.0 l- 4:

1.5

5 1.0

0.5

0.0

.- I m

a:

-3859 -101 6 -980 -965 -952 -929

1 .o

'5 0.8

2 0.6 0

;ij 0.4

a:

h

.- W

I U m

.- F - al

0.2

0.0

0 unstimulated PMA stimulated

-3859 -980 -965 -864 -770 -726 -547

are given in Fig 4. There was some variation in the absolute level of CAT expression obtained for each construct in dif- ferent experiments, as would be expected for random poly- clonal populations of stable transfectants. However, the -38S91L-SCAT construct consistently gave a high level of inducible expression with inducibility greater than IO-fold, indicating the presence of the IL-S enhancer. The -38S91L- SCAT construct was used to generate four independent poly- clonal lines, which showed inducibility levels of 41, 90, I S , and 82-fold respectively (Fig 2B; Fig 4B, C, and D). In contrast, four independent polyclonal lines generated with the -5471L-SCAT construct all showed no inducibility (Fig 2B; Fig 4A, B, and D). Although the level of constitutive expression of the -5471L-SCAT construct was relatively high in the experiment described in Fig 2B, this was not a consistent finding. In general, the expression of this construct was much lower than the inducible expression of -38S91L- SCAT.

Initial localization experiments narrowed the functional enhancer to the region up to - 1089 (Fig 4A). Further experi- ments reduced the S' boundary of the enhancer to - 1016 (Fig 4B and C). In a number of separate electroporations, the polyclonal lines containing the - 10161L-SCAT construct gave comparable levels of inducible expression to the -38591L-SCAT construct. In general, reporter gene con- structs whose S' ends ranged from - 1016 to -929 gave progressively lower levels of inducible expression (Fig 4). Polyclonal lines containing from -S47 to -88 were not inducible (Figs 2B and 4). Overall, the S' deletion experi- ments suggest that the region - 1016 to -929 carries one or more elements important in the induction of IL-S gene

Fig 4. CAT activity of IL-SCAT constructs after stable transfec- tion into EL423 cells. The CAT activity of each construct is ex- pressed relative to the activity of PMA-stimulated -350211-5CAT (A) or -38591L-5CAT (B, C, and D). Cells were stimulated with 10 ng/mL PMA. The CAT activity is the average of two stimulations of the polyclonal stable cell lines, and the bars indicate the varia- tion between these two experi- ments.

expression and that the IL-S promoter/enhancer is contained within the region to - 1016.

Potential regulaton elements in the distcrl IL-5 enhrrncer region. The region from - 1016 to -929 and surrounding sequences were examined for the presence of sequences cor- responding to the binding sites of known mammalian tran- scription factors. A consensus site for CTF/NFI (-940 to -928) was found within the region essential for high level inducible expression. DNA-binding experiments with nu- clear extracts prepared from EL4.23 cells showed the pres- ence of a nuclear protein(s) binding to this region (Fig S ) . Mutations known to prevent CTF/NFI binding': were intro- duced into the - 1170IL-SCAT construct (Fig 6A) and Fig SB also shows that mutation of this site prevents binding of nuclear protein(s) to this region of the mouse IL-S gene.

Stable transfections of these mutated - I I701L-SCAT con- structs were performed in EL4.23 cells and compared with the parent construct - 1 1701L-SCAT. Mutation of the CTF/ NFI binding site had the very interesting effect of converting the - 1 I701L-SCAT stable transfectants to constitutive ex- pression (Fig 6B). It would appear, therefore, that the CTF/ NFI site is involved in the inducible expression of the IL-S gene in EL4.23 cells. Although the S' deletion experiments suggest the possibility that other binding sites may be present in the region - 101 6 to -929, we have not yet been able to demonstrate any further DNA binding in gel retardation experiments.

Role of n consenwl TCATlT-contnining element in the proximal pmn~oter region. A TCATTT-containing ele- ment extending from -61 to -41 of the mouse IL-5 gene is highly conserved in the human TL-S gene and the mouse

TRANSCRIPTIONAL REGULATION OF THE IL-5 GENE

1 2 3 4 5 6 7 8

Fig 5. Binding of nuclear protein(s) t o the CTF/NFl site in the IL- 5 gene. Gel retardation assay using labelled CTF-l duplex DNA and nuclear extracts from unstimulated EL423 cells (lanes 2 t o 5) or cells stimulated with 10 ng/mL PMA (lanes 6 t o 8). Binding was performed in the absence of competitor DNA (lanes2 and 61, and in the presence of 100-fold unlabelled CTF-1 (lanes 3 and 7). CTF-2 (lanes 4 and 8). or RS (lane 5). Lane 1 contains no nuclear extract. CTF-1 contains the sequence -946 t o -915 of the IL-5 gene, CTF-2 contains this sequence mutated at the CTF/NFl site, and RS is an oligonucleotide duplex of random sequence (see Materials and Methods and Fig 6A for nucleo- tide sequences). DNA-protein bound complexes and free unbound DNA are indicated.

and human GM-CSF genes.'" Evidence supporting the role of this element in transcriptional regulation of the GM-CSF gene has been previously reported.'6 To test whether this element was involved in inducible expression of the IL-5 gene in EL4.23 cells, three nucleotide changes were made

Fig 6. Mutation of the CTF/NFl site in -11701L- %AT. (A) Nucleotide sequence of -11701L-5CAT and mutagenized -11701L-5CAT constructs. (B) CAT ac- tivity of -11701L-5CAT and mutagenized -11701L- 5CAT constructs transfected into EL4.23 cells and stimulated with 10 ng/mL PMA. The CAT activity of each construct is expressed relative t o the activity of -11701L-5CAT in stimulated cells. The CAT activity is the average of five separately transfected poly- clonal stable cell lines before and after stimulation, and the bars indicate the variation between these transfected populations.

2075

to the TCATTT motif present in the - 1 170IL-5CAT reporter gene construct. The base substitutions used (Fig 6A) have previously been shown to abolish inducible expression of human GM-CSF reporter gene constructs in transient expres- sion assays.s5

Comparison of the - 117OIL-5CAT with the same con- struct containing the mutant TCATTT-containing element using stable transfections in EL4.23 cells showed that muta- tion of the TCATTT motif abolished inducible expression of the promoter (Fig 6B).

DISCUSSION

Other studies in this laboratory have shown that the con- trol of IL-5 gene expression in the T-cell clone D10.Ci4.1 and the T-cell lymphoma EL4.23 is predominantly transcrip- tional.".' In the present work, we have established a stable transfection assay, using either of these two cell lines, which measures IL-5 promoter activity, and we have used this assay to localize the mouse IL-5 gene transcriptional enhancer, which is responsive to stimulation of the T-cell receptor. The proximal region of the promoter up to -547 was shown to be capable of directing constitutive expression in stable expression assays in the T-cell clone D10.G4.1. However, the effect of inclusion of additional upstream sequence to -3859 resulted in high levels of inducibility. Reporter genes carrying 3884 bases of 5' sequence (-3859 to +25), when stably transfected into D10.G4.1 cells, showed all of the known regulatory characteristics of the endogenous IL-5

Induction could be achieved with ConA, PMA, or the T-cell receptor idiotype-specific monoclonal antibody 3D3. Induction by these agents was inhibited strongly by

B l 3 T T 0 unstimulated EL4 B PMA stimulated EL4

CTFlNFl TCATlT -11701L-5CAT mutation in mutation in

-1 1701L-5CAT -1 1701L-5CAT

2076 BOURKE ET AL

dexamethasone, but was relatively resistant to cyclosporin A.

Analogous results to those obtained with the T-cell clone D10.G4.1 were also obtained with the T-cell lymphoma EL4.23. The -3859IL-5CAT construct gave inducible ex- pression in response to PMA stimulation, the same as for the endogenous gene,8 but only when stable transfection assays were used. Once again, the proximal region of the promoter up to -547 directed constitutive expression. In EL4.23 cells, maximal inducible expression was dependent on the presence of the region - 1016 to -929. A consensus binding site for CTF/NFl was located at -940 to -928 within the critical region for inducible expression. The lower levels of inducibility of constructs -965 to -929 compared with - 1016IL-5CAT may be due to a requirement for flank- ing sequences for cooperative interaction between transcrip- tion factors. Gel retardation assays provided evidence for the presence of a nuclear protein(s), which was capable of binding to synthetic duplex DNA corresponding to the se- quence of this region. Mutation of the binding site for CTFI NF1 in a reporter gene construct (- 1170IL-5CAT) resulted in loss of inducible expression, suggesting involvement of CTF/NFl in the regulatory mechanism. Although CTF/NFl has more commonly been reported to function in the proxi- mal regions of gene promoters, it also has been shown to be active at upstream enhancer sites.".3x Our results clearly show that the CTF/NFI element at -940 to -928 is part of a distal enhancer involved in inducible expression of the IL- 5 gene and that induction of the gene may involve displace- ment or modification of CTF/NFl.

Mutation of a conserved TCATTT-containing element in the proximal region of the promoter showed that this element was also involved in inducible expression of the IL-5 gene in EL4.23 cells. This supports analogous findings for the role of this conserved region in the inducible transcription of the human and mouse GM-CSF gene.36.39 Furthermore, a recent study has demonstrated the functional importance of the analogous element in the mouse IL-4 pr~moter.~" Re- sponsiveness of both the GM-CSF and IL-4 promoters to T- cell activation signals appears to require other elements in addition to the intact TCATTT-containing LIS is also the case with the IL-5 gene.

Our failure to detect the inducible mouse IL-5 gene en- hancer in transient transfection assays was an unexpected finding. It does not appear to be due to low transfection efficiency because transient transfection of a whole gene construct driven by the SV40 early promoter gave detectable expression, and constitutive expression of deleted IL-5 pro- moter constructs was observed. One possible explanation is that the antigen-responsive enhancer is active only in the highly structured chromatin environment of the chromo- some, as has been established for several elements of the globin locus control region?' In a very recent study of IL- 5 gene regulation using a different subline of EL4 and a reporter construct carrying 1200 bp of the promoter region, inducible promoter activity was detected in transient assays providing a combination of PMA and substances which ele- vate levels of intracellular CAMP was used for induction.''

The present work has provided several significant ad-

vances in approaching the study of IL-5 gene regulation. First, stable transfection of target cells was found to be nec- essary to obtain high level, inducible expression, which mir- rored the inducible expression of the endogenous gene in every respect we were able to test. Second, using stable transfection, we were able to obtain inducible reporter gene expression in the T-cell clone D10.G4.1. This is a valuable advance, because D10.G4.1 has the relevant surface mole- cules and signalling pathways for a more normal induction of cytokine genes than is possible in transformed T-cell lines. Nevertheless, the T lymphoma EL4.23 cells have been useful in the dissection of the regulatory pathway.

Future work will need to define any other elements neces- sary for regulated expression and address in detail the coop- eration between distal and proximal elements in the induc- tion of transcription. Extension of these studies to the human IL-5 gene may provide opportunities for the development of drugs that inhibit induction of IL-5 and are useful in treating allergic disease. Definition of the normal regulation mecha- nisms also may allow investigation of possible abnormal IL- 5 regulation in disorders involving eosinophilia, such as the hypereosinophilic syndrome.44

ACKNOWLEDGMENT

We thank David Tremethick for many helpful discussions, and Margaret Sneddon and David Mann for assistance with the gel retar- dation assay, CAT assay, and in vitro mutagenesis.

REFERENCES 1. Sanderson CJ: Interleukin-S, eosinophils, and disease. Blood

79301 , 1992 2. Sanderson CJ, O'Garra A, Warren DJ, Klaus GGB: Eosinophil

differentiation factor also has B-cell growth factor activity: Proposed name interleukin 4. Proc Natl Acad Sci USA 83:437, 1986

3 . Tominaga A, Takaki S, Koyama N. Kat0 S, Matsumoto R, Migita M, Hitoshi Y, Hosoya Y, Yamauchi S, Kanai Y, Miyazaki J I , Usuku G, Yamamura KI, Takatsu K: Transgenic mice expressing a B cell growth and differentiation factor gene (interleukin 5 ) de- velop eosinophilia and autoantibody production. J Exp Med 173:429, 1991

4. Vaux DL, Lalor PA, Cory S, Johnson GR: In vivo expression of interleukin S induces an eosinophilia and expanded Ly- 1 B lineage populations. Int Immunol 2:965, 1990

S. Wang Y, Campbell HD, Young 1G: Sex hormones and dexa- methasone modulate interleukin-S gene expression in T lympho- cytes. J Steroid Biochem Mol Biol 44:203, 1993

6. Naora H, Akin JG, Young IG: TCR-dependent and -indepen- dent signaling mechanisms differentially regulate lymphokine gene expression in the murine T helper clone D10.G4.1. J Immunol 1525691, 1994

7. Lee JS, Campbell HD, Kozak CA, Young IG: The 1L-4 and IL-S genes are closely linked and are part of a cytokine gene cluster on mouse chromosome 1 1 . Somat Cell Mol Genet 15: 143, 1989

8. Naora H, Young IG: Mechanisms regulating the mRNA levels of interleukin-S and two other coordinately expressed lymphokines in the murine T lymphoma EL4.23. Blood 83:3620, 1994

9. Deininger PL: Random subcloning of sonicated DNA: Appli- cation to shotgun DNA sequence analysis. Anal Biochem 129:216, I983

I O . Chen EY, Seeburg PH: Supercoil sequencing: A fast and simple method for sequencing plasmid DNA. DNA 4:165, 1985

TRANSCRIPTIONAL REGULATION OF THE IL-5 GENE 2077

11. Sanger F, Nicklen S, Coulson AR: DNA sequencing with chain-terminating inhibitors. Proc Natl Acad Sci USA 745463, 1977

12. Staden R: Automation of the computer handling of gel reading data produced by the shot-gun method of DNA sequencing. Nucleic Acids Res 10:4731, 1982

13. Staden R: A computer program to enter DNA gel reading data into a computer. Nucleic Acids Res 12:499, 1984

14. Miller J, Malek TR, Leonard WJ, Greene WC, Shevach EM, Germain RN: Nucleotide sequence and expression of a mouse in- terleukin 2 receptor cDNA. J Immunol 134:4212, 1985

15. Campbell HD, Sanderson CJ, Wang Y, Hort Y, Martinson ME, Tucker WQJ, Stellwagen A, Strath M, Young IG: Isolation, structure and expression of cDNA and genomic clones for murine eosinophil differentiation factor. Eur J Biochem 174:345, 1988

16. Gorman CM, Moffat LFB, Howard BH: Recombinant ge- nomes which express chloramphenicol acetyltransferase in mamma- lian cells. Mol Cell Biol 2:1044, 1982

17. Henikoff S: Unidirectional digestion with exonuclease 111 cre- ates targeted breakpoints for DNA sequencing. Gene 28:351, 1984

18. Kunkel TA: Rapid and efficient site-specific mutagenesis without phenotypic selection. Proc Natl Acad Sci USA 82:488, 1985

19. Kunkel TA: Oligonucleotide-directed mutagenesis without phenotypic selection, in Ausubel F, Brent R, Kingston RE, Moore DD, Seidman JG, Smith JA, Struhls K (eds): Current Protocols in Molecular Biology. New York, NY, Wiley and Sons, 1987, p 8.1. I

20. Rice CM, Franke CA, Strauss JH, Hruby DE: Expression of Sindbis virus structural proteins via recombinant Vaccinia virus: Synthesis, processing and incorporation into mature Sindbis virus. J Virol 56:227, 1985

21. Thomas KR, Capecchi MR: Site-directed mutagenesis by gene targeting in mouse embryo-derived stem cells. Cell 51:503, 1987

22. Chomczynski P, Sacchi N: Single step method of RNA isola- tion by acid guanidinium thiocyanate-phenol-chloroform extraction. Anal Biochem 162:156, 1987

23. Karasuyama H, Melchers F: Establishment of mouse cell lines which constitutively secrete large quantities of Interleukin 2, 3, 4 or 5 , using modified cDNA expression vectors. Eur J Immunol 18:97, 1988

24. Kaye J, Porcelli S, Tite J, Jones B, Janeway CA: Both a monoclonal antibody and antisera specific for determinants unique to individual cloned helper T cell lines can substitute for antigen and antigen-presenting cells in the activation of T cells. J Exp Med 158:836, 1983

25. Seed B, Sheen J-Y: A simple phase-extraction assay for chlor- amphenicol acetyltransferase activity. Gene 67:27 1, 1988

26. Miskimins WK, Roberts MP, McClelland A, Ruddle FH: Use of a protein-blotting procedure and a specific DNA probe to identify nuclear proteins that recognize the promoter region of the transferrin receptor gene. Proc Natl Acad Sci USA 82:6741, 1985

27. Singh H, Sen R, Baltimore D, Sharp PA: A nuclear factor that binds to a conserved sequence motif in transcriptional control elements of immunoglobulin genes. Nature 319:154, 1986

28. Borgmeyer U, Nowock J, Sippel AE: The TGGCA-binding protein: Aeukaryotic nuclear protein recognising a symmetrical se- quence on double-stranded linear DNA. Nucleic Acid Res 12:4295, 1984

29. Staudt LM, Singh H, Sen R, Wirth T, Sharp PA, Baltimore D: A lymphoid-specific protein binding to the octamer motif of immunoglobulin genes. Nature 323:640, 1986

30. Shannon MF, Gamble JR, Vadas MA: Nuclear proteins inter- acting with the promoter region of the human granulocytdmacro- phage colony-stimulating factor gene. Proc Natl Acad Sci USA 85:674, 1988

3 1. Novak TJ, Rothenberg EV: Differential transient and long- term expression of DNA sequences introduced into T-lymphocyte lines. DNA 5:439, 1986

32. Yoshimura FK, Chaffin K Different activities of viral en- hancer elements before and after stable integration of transfected DNAs. Mol Cell Biol 7: 1296, 1987

33. Wilson C, Bellen HJ, Gehring WJ: Position effects on eukary- otic gene expression. Annu Rev Cell Biol 6:679, 1990

34. Rolfe FG, Hughes JM, Armour CL, Sewell WA: Inhibition of interleukin-5 gene expression by dexamethasone. Immunology 77:494, 1992

35. Jones K A , Kadonoga JT, Rosenfeld PJ, Kelly TJ, Tjian R: A cellular DNA-binding protein that activates eukaryotic transcription and DNA replication. Cell 48:79, 1987

36. Nimer S, Fraser J, Richards J, Lynch M, Gasson J: The re- peated sequence CATT(A/T) is required for granulocyte-macro- phage colony-stimulating factor promoter activity. Mol Cell Biol 10:6084, 1990

37. Johnson MR, Bruzdzinski CJ, Winograd SS, Gelehrter TH: Regulatory sequences and protein-binding sites involved in the ex- pression of the rat plasminogen activator inhibitor-l gene. J Biol Chem 267:12202, 1992

38. Seipel K, Georgiev 0, Schaffner W: Different activation do- mains stimulate transcription from remote (‘enhancer’) and proximal (‘promoter’) positions. EMBO J 11:4961, 1992

39. Miyatake S, Shlomai J, Arai K, Arai N: Characterization of the mouse granulocyte-macrophage colony-stimulating factor (GM- CSF) gene promoter: Nuclear factors that interact with an element shared by three lymphokine genes- those of GM-CSF, interleukin- 4 (IL-4) and IL-5. Mol Cell Biol 115894, 1991

40. Szabo SJ, Gold IS, Murphy TL, Murphy KM: Identification of cis-acting regulatory elements controlling interleukin-4 gene ex- pression in T cells: Roles for NF-Y and NF-ATC. Mol Cell Biol 13:4793, 1993

41. Abe E, de Waal Malefyt R, Matsuda I, Arai K, Arai N: An 1 l-base-pair DNA sequence motif apparently unique to the human interleukin 4 gene confers responsiveness to T-cell activation signals. Proc Natl Acad Sci USA 89:2864, 1992

42. Philipsen S, Talbot D, Fraser P, Grosveld F: The beta-globin dominant control region: Hypersensitive site 2. EMBO J 9:2159, 1990

43. Lee HJ, Koyano-Nakagawa N, Naito Y, Nishida J, Arai N, Arai K, Yokota T: CAMP activates the IL-5 promoter synergistically with phorbol ester through the signaling pathway involving protein kinase A in mouse thymoma line EL-4. J Immunol 151:6135, 1993

44. Owen WF, Rothenberg ME, Peterson J, Weller PF, Silberstein D, Sheffer AL, Stevens RL, Soberman RJ, Austen KF: Interleukin 5 and phenotypically altered eosinophils in the blood of patients with the idiopathic hypereosinophilic syndrome. J Exp Med 170:343, 1989