Vol.

176,

May

15, 1991

No.

3, 1991

BIOCHEMICAL

AND

BIOPHYSICAL

RESEARCH

COMMUNICATIONS Pages

1326-1332

IDENTIFICATION OF TWO DISTINCT NUCLEAR FACTORS WITH DNA BINDING ACTIVITY WITHIN THE GLUCOCORTICOID REGULATORY REGION OF THE RAT ALPHA-l-ACID GLYCOPROTEIN PROMOTER D. DI Lorenzo ' Institute 2

of

l,P.Williams

2 and G Ringold

Chemistry,School of Medicine,University Brescia,Italy

Institute of Cancer and Developmental Research,S3-1,3401 Hillview Avenue,Palo

Received

March

25,

2 of

Biology,Syntex Alto,CA 94304

1991

Summary:Efficient glucocorticoid induction of alpha-l-acid glycoprotein (AGP) mRNA in rat hepatoma cells HTC (JZ-1) requires the activity of one or more preexisting and labile proteins acting cooperatively with the glucocorticoid receptor.Inhibiting protein synthesis markedly diminishes the glucocorticoid induction of rat AGP mRNA without affecting the inducibility of other glucocorticoid inducible genes such as the mouse mammary tumor virus (MMTV) or tyrosine amino transferase (TAT).The sequences responsible for conferring glucocorticoid inducibility to the rat AGP gene have localized on the AGP between -121 promoter nucleotides and -42.A typical glucocorticoid regulatory element (GRE) is found between residues -121 and -105 and downstream of this are the sequences (-90 to -42) responsible for the cycloheximide inhibition of the hormonal induction (lO).Using mobility shift assays we have characterized the binding of two proteins or complexes of proteins to this promoter region (-90 to -64).Our data show that the binding of these factors (called ANF-1 and ANF-2) to the DNA is highly specific,and is not directly affected by cycloheximide.Furthermore a second binding site for ANF-2 has been localized in the AGP regulatory region to a sequence that overlaps the GRE. 0 1991Academic Press,Inc.

Glucocorticoids can either positively (l-3) or negatively (4-5) regulate expression of a variety of genes.Binding of the hormone-receptor complex to specific DNA sequences in the 5'flanking regions of these genes appears to be a critical step in the signal transducing mechanism for steroid hormones.Although binding of the hormone-receptor complex is a necessary event,in many cases it is not sufficient for proper regulation of transcription.The presence of binding sites for other nuclear transcription factors and correct spacing between the receptor and these nuclear factor binding sites also seem to play a critical role (6-7).The underlying mechanisms to explain the 0006-291X/91 Copyright All rights

$1.50

0 1991 by Academic Press. Itzc. of reproduction in any form reserved.

1326

Vol.

176,

No.

cooperativity

3, 1991

BIOCHEMICAL

between

the

AND

BIOPHYSICAL

glucocorticoid

RESEARCH

receptor

COMMUNICATIONS

and

other

transcription factors is still unknown. Various studies have shown that blocking proteins synthesis leads to a marked inhibition of steroid hormone induction of genes such as : alpha-l-acid glycoprotein (AGP) (8).1n contrast the induction of many promoters including tyrosine aminotransferase (TAT) (9) is not abrogated by similar treatment.In the case of AGP,we have recently provided evidence that the diminution in glucocorticoid inducibility observed in cycloheximide-treated cells is due to the decay of a preexisting and labile protein that presumably acts in concert with the glucocorticoid receptor.The region of the promoter responsible for both cycloheximide sensitivity and glucocorticoid inducibility has been mapped between -90 and -42 (10) .Here we report the existence of two different proteins or protein complexes (as detected by gel shift assays) that bind to the AGP promoter in the region -121 to -65.0ne of these proteins (or complexes) binds to two different sites within this region one of which overlaps the GRE.

MATERIALS

AND METHODS

Enzymes and Chemicals :The reagents were purchased from Boehringer Mannheim and Sigma.Radiolabeled reagents were purchased from New England Nuclear.The oligo deoxynucleotides were synthesized in the Institute of Bio-organic Chemistry,Syntex,using an Applied Biosystem Synthesizer. Cell Culture : The rat HTC hepatoma cell line JZ-1 (8),Hela,HepG2 and Mouse L-cells were grown in Dulbecco's Modified Eagle Medium (M.A.Bioproducts) supplemented with 10% calf serum (Hyclone). Preparation of nuclear extracts and gel retardation assay : Nuclear extracts were prepared form cell lines essentially as described by Denison et Al.(ll).Gel retardation assay was performed as described (12).After purification by polyacrylamide gel electrophoresis synthetic oligodeoxynucleotides were labeled at their 5' ends using T4 polynucleotide kinase and 32 P y-ATP. Nuclear extract (5-10 pg of protein depending on the cell line) was mixed with 12 ~1 of 25 mM Hepes,pH 7.5,1 mM EDTA , 1 mM dithiothreitol,lO% glycerol (HEDG buffer) and 2 ~1 (2 pg) of Poly d(I-C).The mixture was preincubated for 10 minutes at room temperature-The labeled fragment ( 32P) (10-20.000 cpm,0.05 - 0.1 ng in 1 pl of buffer) was added with and without a loo-fold molar excess of unlabeled oligonucleotide and the incubation was minutes room continued for an additional 20 at temperature.Protein-DNA complexes were run on a 4.0% nondenaturing polyacrylamide gel electrophoresis with recirculation,using 0.1 M Tris,pH 8.0,o.j M Na Acetate,0.05 M buffer for 2 hours at 110 volts EDTA as running (11 volts/cm) ,constant current and autoradiographed at -7O'C. 1327

Vol.

176,

No.

BIOCHEMICAL

3, 1991

-100

AND

BIOPHYSICAL

-80

RESEARCH

-70

-80

UA

COMMUNICATIONS

HE I

l-R~~~lWTGCACTCTCCAGCCAC

CClmT~ -80

-80

-40

-10

-20

20

+1

FTTTT

Figure 1 . The regions of digestion called X,Y and AGP-GRE (-120 to -105),the -104 overlapping part of and HB (-81 to -72 within

the AGP promoter protected by DNAse 1 Z are shown.Other elements are : the two homologous sequences HA (-113 to the GRE and the X footprinted region) the Z footprinted region). RESULTS

The results presented in Figure 2 demonstrate that two distinct proteins (or complexes) bind to the oligonucleotide OL1,representing the AGP promoter region -90 to -64 (Y and Z domains) (Figure l).The competition experiments with the wildtype and the mutated oligonucleotides (Table l),indicate that the two bands seen in the mobility shift assays arise from the binding of two different proteins or protein complexes with high specific affinities for the Y and Z domains.Somewhat to our surprise,oligonucleotide OL-16,corresponding to the region -120 to -90 (Figure 3,panel C) was able to compete efficiently with OL-1 (-90 to -64) for the binding of ANF-2 (Fig.2J.A careful analysis of the two sequences C-120 to -90 and -90 to -64) revealed a region of similarity (8 nucleotides out of 10 ) between position -113 to -104 (sequence called HA),within the domain GRE-X and position -81 to -72 (sequence called HB) within the domain Z (see Figure l).The results shown in Figure 3 support the contention that the same protein binds to the two homologous sequences,HA and HB. Overlap of the Glucocorticoid Regulatory Element and the ANF-2 Binding Site : The sequence HA (see Figure 1) contains half of Table

1 BINDING

1 2 3 4

(-90)

-ANF-l-

ANF-2

(-64)

A;

+ -

CTGGTGRGATTGTGCCACAGCTCTGCA .... ..GTG ..................

.............. ...... TTT .....

OF FACTORS ANF-2

+ +

+ -

..GTG ........ ..GTG ........

Oligonucleotides corresponding to the wild and type mutated sequence of the AGP promoter bases -94 to -64 used in the gel shift and competition assays.The plus indicates this

oligonucleotide competed as well as wild binding. The minus indicates a significantly for protein binding. 1328

type

for lower

ANF-1 or competition

2

Vol.

176,

No.

3, 1991

BIOCHEMICAL

AND

BIOPHYSICAL

PANEL L&&3d Template

RESEARCH

COMMUNICATIONS

B

OL-1

ANF-2 Labeled Template CM Competitors

OL-1

OL-1 P

1,

9

p

P

k

i

b

ANF-1

Figure 2 . Results of mobility shift assays with HTC(JZ-1) nuclear proteins.ANF-1 and ANF-2 indkcate the bands corresponding to the binding of nuclear factors to the region shown in figure 1 as Y and 2,respectively.A) In lane l,ANF-1 and ANF-2 binding to OL-1 is shown;this binding is competed by an excess of the same oligonucleotide OL-1 (lane 2).0L-2,mutated in Y,does not compete for the binding of ANF-l,(lane 3);lane 4,OL3,mutated in Z,does not compete for the binding of ANF-2 (lane 4);and OL-4,mutated in both Y and Z,does not compete for either factor. Figure 3 . Shown are competition experiments for binding to ANF1 and ANF-2 between the sequence -90 to -64 (oligonucleotide OL1 containing the Y and Z domains) and the sequence -120 to -90 (oligonucleotide OL-16 containing the binding site for the glucocorticoid receptor and the x domain).OL-17 and OL-18 contain the sequence HA (see figure 1) and an HA mutated sequence,respectively.Panel A,lane 1 shows the binding of ANF-1 and ANF-2 to template.Lanes 2 and 3 show that the oligonucleotides OL-16 and OL-17 compete efficiently for the binding of ANF-2 to OL-l.When the last three nucleotides of HA are mutated in oligonucleotide OL-18 the competition is abolished.Panel B,the oligonucleotide OL-12 containing a single point mutation (from A to G),in the sequence within HB (that resemble a half palindromic GRE) competes efficiently for the binding of ANF-2 to the template OL-l.Panel C shows the region the binding of the rat AGP promoter (-120 to -64) containing sites for the glucocorticoid receptor,ANF-1 and ANF-2.Below are shown the oligonucleotide OL-16 (-90 to -64) and the BGL II and oligonucleotide OL-17 (-113 to -104 plus flanking Hind III restriction sites).OL-18 is identical to OL-17 but with three nucleotides changed. the palindromic GRE of the AGP promoter plus three nucleotides immediately downstream (TTGTGCAAGA).HB contains a nucleotide sequence that strongly resembles the half GRE of the sequence in HA (TTGTGCCACA).To determine whether ANF-2 shares specific DNA binding characteristics with the glucocorticoid receptor we 1329

Vol.

176,

No.

3, 1991

BIOCHEMICAL

AND

BIOPHYSICAL

RESEARCH

COMMUNlCATlONS

;EL 9ETARDATlOh OF ANF I AND ANF 2 FRO!>1 NUCLEAR CXTRAClS OF CYCLOHEXMDE TREATEC HTC ,JZ i! CELLS

ANF-1 -i

Figure 4 . Mobility shift assays with OL-1 were performed using nuclear extract obtained from cells treated with cycloheximide (10 pg/ml) for 0,8,and 16 hours.The labeled DNA (OL-1) was incubated with each of the extracts using the same amount of total protein (5 j.lg). tested a variant of oligonucleotide,OL-1 (-90 to -64) mutated at the G residue in position -79 (Fig.3,panel B,OL-12).The equivalent guanine in a GRE has previously been shown to be essential for the binding of the glucocorticoid receptor (10,13).As shown in figure 3 this mutation does not affect the binding of ANF-2 since OL-12 is able to effectively compete with the labeled OL-l.We surmise that ANF-2 is not the glucocorticoid receptor nor does its binding to DNA require contact with the same nucleotides. ANF-1 and ANF-2 are stable in cycloheximide-treated cells : Since ANF-1 and ANF-2 bind to the region of the promoter responsible for cycloheximide inhibition of glucocorticoid stimulated AGP transcription we considered whether one or both of these proteins might be the labile factor responsible for this phenomenon.We therefore prepared nuclear extracts from JZ-1 cells treated with cycloheximide for various times.As seen in Figure 4 the intensity of the bands corresponding to both ANF-1 and ANF-2 did not change over a period of sixteen hours of cycloheximide treatment.From previous experiments we know that this length of time is sufficient for cycloheximide to dramatically inhibit the dexamethasone induction of the rat AGP gene in JZ-1 cells (14).We conclude that neither ANF-1 or ANF-2 represent the labile factor whose activity is required for efficient glucocorticoid induction of AGP. DISCUSSION The region of the rat to -42 from the start

AGP promoter comprising site of transcription 1330

is

nucleotides able to

-120 direct

Vol.

176,

No.

3, 1991

BIOCHEMICAL

AND

BIOPHYSICAL

RESEARCH

COMMUNICATIONS

the glucocorticoid inducibility (15) and the cycloheximide sensitivity (10) of the rat AGP promoter in the JZ-1 cell line.Here we have identified two nuclear factors,which we call ANF-1 and ANF-2 that bind within this region (-90 to -64).The data indicates that the binding of factors ANF-1 and ANF-2 to the DNA does not change after treatment of cells with cycloheximide for a length of time that is sufficient to strongly inhibit the hormone induction of the AGP gene.We therefore presume that blockade of protein synthesis affects other proteins whose activity is necessary for transcriptional activation of the AGP promoter.For instance it has recently been shown that some transcription factors,even if capable of binding to specific DNA sequences,may require phosphorylation or dephosphorylation to be fully activated (16,17) .Glycosylation may also lead to the activation of transcription factors bound to DNA (18).Moreover as indicated by others (19,20,21) proteinprotein interactions between a DNA-bound protein and an adaptor factor may be a necessary event for transcriptional multiple pathways that lead to the activation activation.Thus, or induction enzymes such as protein of kinases,phosphatases,glycosidases and transcription adaptor proteins may be target elements for the cycloheximide inhibition of transcription factor activity.Whether ANF-1 or ANF-2 are subject to modification or contacted by another labile protein remains to be determined. The regulatory unit of the AGP promoter contains a repeated element,HA (-113 to -104) and HB (-81 to -72),that serves as a binding site for ANF-2.It is intriguing that HA contains half of the GRE palindrome (-121 to -106) to which the glucocorticoid receptor binds (lO).Perhaps when ANF-2 is bound to HA,GR binding would be precluded.However,mutation of a critical G residue that is essential for GR binding does not eliminate ANF-2 binding.We speculate that ANF-2 and the GR may bind therefore simultaneously to this sequence by contacting opposite sides of that efficient the DNA helix.If so,it is conceivable glucocorticoid mediated induction of the AGP promoter is dependent upon specific protein-protein interactions between the GR and ANF-2.Purification of ANF-2 will be necessary to rigorously test this hypothesis.

REFERENCES 1. 2.

Ringold, G.M. (1985).Ann.Rev.Pharmacol.Toxicol.25, Yamamoto, K.R. (1985) .Ann.Rev.Genet. 19, 209-252. 1331

529-

566.

Vol.

3.

4. 5. 6. I. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21.

176,

No.

3, 1991

BIOCHEMICAL

AND

BIOPHYSICAL

RESEARCH

COMMUNICATIONS

Jantzen, H.M.,Strahle, U.,Glass, B.,Stewart, F.,Schmidt W.,Boshart, M.,Miksiek, R.,and Shutz, G.(1987).Cell 49,29-38. Slater P.E.,Beato, M.,Baxter, J.D., and Akerblom, I.E., Mellon, P.L. (1988) .Science 241,350-353. Sakai, D.D.,Helms, S.,Carlsted-Duke, J.,Gustafsson, and Yamamoto,K. J.A., Rottman, F.R., (1988) . Genes and Development 2,1144-1154. R.,Muller ,M.,Otsuka-Murakami, H., and Renkawitz, R. Schule, (1988) . Nature 332,87-90. Danesh, U.,Gloss, B.,Schmid, W.,Schutz, G.,Schule ,R., and Renkawitz, R. (1987) EMBO J.6,625-630 Vannice, J.L.,Ringold ,G.M.,McLean, J.W., and Taylor, J.M. (1983) .DNA 2,205-212. Peterkofsky, B., and Tomkins ,G.M. (1968) . Proc Acad Nat1 Sci 60,222-228 Klein, E.S .,Di Lorenzo, D.,Posseckert, G.,Beato, M., and Ringold, G. (1988).Mol Endocrinology 12,1343-1351. Denison,M.S.,Fischer,J.M.,and Whitlock,J.P.,Jr.(l988).Proc Nat1 Acad Sci 85,2528-2532. Singh,H.,Sen,R.,Baltimore,D., and Sharp,P.(1986).Nature 319:154-158. Scheidereit, C.,Westphal, H.M.,Carlson, C.,Bossahard, H.,Beato, M. (1986) .DNA 5,383-391 Klein, E.S.,Reinke, R.,Feigelson, P., and Ringold, G. (1987) J.Biol.Chem 262,520-523 Baumann ,H., and Maquat, L.E. (1986) . Mol Cel Biol 6,25512561 Christian,R., andBernard,L.(lggO).NAR 8,2125-2131. Lamph,W.,Dwarki,R.,Ofir,R.,Montminy,M.,and Verma,I. (1990). Proc Nat1 Acad Sci 87,4320-4324. Jackson, S.P., and Tjian, R. (1988) .Cell 55,125-133. Berk, A.J. (1986) . Ann Rev Genet 20,45-55. Sadowsky, I.Triezenberg,J.M.S. and Ptashne,M.(1988) .Nature 335,563-564. Mitchell,P., and Tjian,R.(1989).Science 245,371-378.

1332