Identification of regulatory sequences in the promoter of the PDGF B-chain gene in malignant mesothelioma cell lines

B i o c h i ~mie~a et Biophysica A~ta ELSEVIER

Biochimica et BiophysicaActa 1317 (1996) 223-232

Identification of regulatory sequences in the promoter of the PDGF B-chain gene in malignant mesothelioma cell lines Jan-Bas Prins a,*, Anthonie W. Langerak a Ron P.H. Dirks b, Carin A.J. Van der Linden-Van Beurden a, Petronella A.J.M. De Laat a, Henri P.J. Bloemers b, Marjan A. Versnel a a Departmentof Immunology, Erasmus UniversityRotterdam,P.O. Box 1738, Rotterdam, The Netherlands b Departmentof Biochemistry, UniversityofNijmegen, Nijmegen,The Netherlands Received 16 September 1996; accepted 9 October 1996

Abstract Platelet-derived growth factor (PDGF) B-chain mRNA is readily detectable in malignant mesothelioma (MM) cell lines, but not in normal mesothelial (NM) cell lines. The high affinity receptor for PDGF B-chain dimers, the PDGF/3-receptor, is expressed in MM cell lines. NM cell lines predominantly express the PDGF a-receptor. Coexpression of the PDGF/3-receptor and its ligand may lead to an autocrine growth stimulating loop in the malignant cell type. In nuclear run off experiments, PDGF B-chain mRNA was detectable in MM cells only, indicating an increased level of transcription in this cell type. The proximal promoter of the PDGF B-chain gene contains DNaseI hypersensitive (DH) sites and mediates reporter gene activation in both normal and malignant cells. Nuclear proteins, extracted from both cell types, interact with DNA sequences within the proximal promoter around bp - 6 4 to - 6 1 relative to the transcription start site. Electrophoretic mobility shift assays (EMSAs) indicate that these factors are more abundantly present in the malignant than in the normal cell type. A DH site around - 9 . 9 kb was found in both cell types. When tested in CAT assays, this region exerted a stimulatory effect on transcription in malignant cells. The elevated level of transcription of the PDGF B-chain gene in malignant cells may well be the result of interaction of regulatory sites in the proximal promoter and an enhancing element located at - 9.9 kb from the transcription start site.

Keywords: Platelet-derived growth factor B-chain; Malignant mesothelioma; Cell line; Transcription regulation; Promoter

1. Introduction Platelet-derived growth factor (PDGF) is a potent mitogen and chemoattractant for cells of mesenchymal origin. It is composed of two subunit polypeptides: the A- and the B-chains. Homodimeric (AA and BB) as well as heterodimeric (AB) forms of PDGF can be formed [18,19,40]. The polypeptides are encoded by distinct genes, which show a high level of structural similarity [2]. The PDGF B-chain is encoded by the c-sis protooncogene [48]. The three isoforms of PDGF (AA, AB, and BB) interact with

Abbreviations: CAT, chloramphenicol acetyl transferase; CHX, cycloheximide; DH, DNaseI hypersensitive; EGF, epidermal growth factor; EMSA, electrophoretic mobility shift assay; HC, hydrocortisone; MM, malignant mesothelioma; NM, normal mesothelium; PDGF, platelet-derived growth factor * Corresponding author. Fax: + 31 10 4367601.

cell surface receptors, which function as noncovalent dimers. The a-receptor subunit binds all three isoforms of PDGF with equal affinity, whereas the /3 subunit binds just the B-chain with high affinity [4,5,20,38]. Based on expression of PDGF chains and receptors in platelets, endothelial cells, smooth muscle cells and placental trophoblasts, PDGF is thought to function in wound healing and developmental processes among others [34]. In addition, the PDGF A- and B-chain genes have been shown to be expressed, either individually or together, in various tumors, like malignant mesothelioma (MM), and tumor derived cell lines [28,30,33,49]. MM is a tumor of mesodermal origin most often found in the pleura [47]. Our MM cell lines express PDGF B-chain mRNA, while normal mesothelial (NM) cells do not express this protooncogene [26,45]. Also, most MM cell lines express PDGF /3-receptor transcripts, while NM cells predominantly express PDGF a-receptor m R N A [46]. Immunofluorescence

0925-4439/96/$15.00 Copyright © 1996 Elsevier Science B.V. All rights reserved. PII S0925-4439(96)00060-9

224

J.-B. Prins et al. / Biochimica et BiophysicaActa 1317 (1996) 223-232

staining with antibodies directed against PDGF and the PDGF receptors of normal and malignant cells originating from either cell lines or primary material, largely confirmed the mRNA expression pattern [27]. Taken together, these findings are suggestive of an autocrine growth stimulating loop existent in malignant mesothelioma. Direct evidence for the autocrine growth stimulating potential of PDGF comes from studies with astrocytoma and glioblastoma cell lines in which dominant-negative PDGF mutants and PDGF neutralizing antibodies were shown to revert the transformed phenotype of the cell lines [39,43]. The elevated level of PDGF B-chain mRNA in MM cell lines prompted us to study the regulation of expression of this protooncogene. Recently, proximal promoter elements of the PDGF B-chain gene [21,22,24,25,31] as well as several regions containing regulatory elements outside the minimal promoter region have been identified in other cell types [6,7,11]. We demonstrate that the elevated level of PDGF B-chain expression in MM cells is caused by enhanced transcription. Enhanced binding of nuclear factoffs) to sequences around bp - 6 4 to - 6 0 bp in MM cells may contribute to this elevated expression, as well as a DNaseI hypersensitive site at - 9 . 9 kb, which exerts a stimulatory effect on transcription.

2. Materials and methods 2.1. Cell lines and growth conditions

Experiments were performed using human MM cell lines (Mero-25, -41 and -82) and NM cell lines (NM-5, -20 and -21) [45]. All cell lines were cultured in Ham's F10 medium (Gibco/BRL, Life Technologies, Paisley, UK) with 15% fetal calf serum (FCS) under 5% CO 2 at 37°C. Epidermal growth factor (EGF; Collaborative Research, Lexington, MA, USA; 10 ng/ml) and hydrocortisone (HC; Sigma, St. Louis, MO, USA; 0.4 /xg/ml) were added to the culture medium of NM cells. In some experiments cells were exposed to cycloheximide (CHX; Sigma; 10 /xg/ml medium) for 2 h. 2.2. Probes

1.7 kb BamHI PDGF B-chain [16] and 0.7 kb EcoRIPstI GAPDH [1] fragments were used as probes in Northern blot analysis. Nuclear run off assays were performed with the 2.0 kb XhoI PDGF B-chain fragment from pSM1 [23] and the 1.25 kb PstI fl-actin fragment from pAct [9]. Both these probes were subcloned in pUC18 and subsequently spotted onto nitrocellulose filters prior to hybridization. Subcloned sequences of human genomic PDGF B-chain clones were used for mapping of DH sites (see also Fig. 2). PR3 is the 0.61 kb SmaI-EcoRI fragment of pAO56, PR7 the 0.35 kb PstI-HindlII fragment of pAO121, PR12 the 0.9 kb HindlII-KpnI fragment of

pAO56, PR13 the 0.46 kb SmaI-HindIII fragment of pAO149, and PR16 the 0.38 kb PstI-PvulI fragment of pAO121, pAO56, pAO149, and pAO78 contain 11 and 5 kb EcoRI and 8 kb BamHI fragments of cosmid clone ALLW- 1283-C1 21, respectively. Cosmid ALLW- 1283-C1 21, pAO68 and pAOI21 were described elsewhere [41,42]. DNA probes DHI-IWT ( - 1 0 3 / - 2 3 ) , DHI-IMUTA ( - 1 0 3 / - 23), DHI-IWT-5' ( - 1 0 3 / - 54) and DH11WT-3' ( - 7 3 / - 23) were used in electrophoretic mobility shift assays (EMSAs). These probes were obtained by PCR with linker primers (BamHI or EcoRI) using cloned promoter sequences as templates. The amplified products were cloned into pUC 19 after restriction enzyme digestion and gelpurification. These constructs were checked by sequencing. DHI- 1MUTA is the mutant form of DH1- lWT in which the sequence - 6 4 TCTC - 6 1 has been changed into - 6 5 ATATC - 6 1 [8]. End-labeling of the probes was performed as described earlier [44]. Enzyme combinations were either B a m H I / E c o R I or EcoRI/BamHI. 2.3. Reporter gene constructs for CAT analysis

pSV2CAT and pSuperCAT (a promoterless CAT construc0 were used as positive and negative controls for CAT analysis, respectively, pSis-1758/+43CAT, pSis4 2 5 / + 43CAT, and pSis-112/+ 43CAT have been described by others [6,8]. pSis-425/+ 43CAT was used to construct unidirectional deletion mutants: pSis-65/+ 43CAT, pSis-64/+43CAT, pSis-60/+43CAT, pSis4 4 / + 43CAT, and pSis-36/+ 43CAT, using exonuclease III (Promega, Madison, WI, USA) [8]. p S i s - l l 2 / + 18mutaCAT is a site-directed mutant form of pSis-112/+ 18CAT in which the sequence - 6 4 TCTC - 6 1 has been changed into - 6 5 ATATC - 6 1 [8]. All these constructs were checked by sequencing. All other reporter constructs were made by subcloning PDGF B-chain genomic fragments in sense (s) or antisense (a) orientation in the SmaI-site, upstream of the PDGF B-chain promoter in pSisl(s)fAT ( = p S i s - l l 2 / + 43CAT) (Fig. 3, [6,7]). PDGF B-chain-CAT fusion constructs 2(s), 3(s), and 4(s) contain the 170 bp (+302 to +472) Hinfl-XhoI, 297 bp (+472 to +761) XhoI-BamHI, and 449 bp (+761 to +1210) BamHI-BstEII fragments of pAO121, respectively; 5(s) and 6(s) the 0.7 kb PstI and 3.4 kb KpnI fragments of pAO78, respectively; 10(s/a), 12(s/a), 13(s), and 14 (s/a) the 6 kb HindlII-BamHI, 0.6 kb BstUI, 0.7 kb SstI-NarI, and 1.5 kb PzulI-BamHI fragments of pAO56, respectively. 2.4. Northern and Southern blot analysis

DNA isolation and Southern blot analysis (Hybond-N membranes, Amersham Int., Buckinghamshire, UK) were performed according to standard procedures. RNA isolation and Northern blotting were performed as described elsewhere [26].

225

.L-B. Prins et al. / Biochimica et Biophysica Acta 1317 (1996) 223-232 2.5. Nuclear run off assay

Isolation of nuclei, in vitro labeling of nuclear RNA and hybridization protocol have been described elsewhere [26]. 2.6. Mapping o f DNaseI hypersensitive (DH) sites

Mapping of DH sites was performed largely according to a previously described protocol [6]. Nuclei were isolated from either 6 X 108 NM or 6 X 1 0 8 MM cells. After DNaseI (Gibco/BRL, Life Technologies, Paisley, UK) treatment and restriction enzyme (Pharmacia, Uppsala, Sweden) digestion, electrophoresis and blotting, hybridization (5 x SSPE, 0.1% SDS, 0.1% Ficoll, 0.1% BSA, 0.1% polyvinylpyrrolidone and 1 0 0 / x g / m l salmon sperm DNA) was performed at 65°C, overnight. Filters were washed in 1 X SSPE, 0.1% SDS at 65°C. Filters were exposed to Kodak X-OMAT film (Eastman Kodak Co., Rochester, NY, USA) with intensifying screens at -80°C. 2.7. In vivo footprinting

DMS treatment of subconfluent NM or MM cell cultures, DNA isolation and piperidine cleavage of in vitro or in vivo modified methylated guanine residues were all performed as described earlier [8]. Subsequently, genomic footprinting was performed by means of ligation-mediated PCR (LM-PCR) [12] using PDGF B-chain promoterspecific primer sets. PDGF B-chain primers ls (5'-CATGGACTGAAGGGTTGCTC-3'; - 1 9 0 / - 1 7 1 ) , 2s (5'CTCTCAGAGACCCCCTAAGCGCCCC-3'; - 1 6 8 / 144), and 3s (5'-AGACCCCCTAAGCGCCCCGCCCTGG-3'; - 1 6 1 / - 137) were used for lower strand analysis. Primers la (5'-CGCAAAGTATCTCTATCTAGGGAA-3'; + 1 3 1 / + 108), 2a ( 5 ' - T A G G G A A T GAAAAATGGGCGCTGGC3'; + 1 1 4 / + 90), and 3a (5'GGAATGAAAAATGGGCGCTGGCGGC-3'; + 111 / + 86) were used for upper strand analysis. First strand synthesis (primers l s and la), PCR amplification (primers 2s and 2a), and labeling reactions ([y-32P]ATP-labeled primers 3s and 3a) were all performed as described, using Vent DNA polymerase (Thermococcus litoralis DNA polymerase; New England Biolabs, Beverly, MA, USA) [8]. Mixtures were separated on 6% polyacrylamide sequencing gels (Bio-Rad Laboratories, Hercules, CA, USA) in 1 X TBE (90 mM Tris, 90 mM Boric acid, 25 mM EDTA). Gels were dried and exposed to Kodak X-OMAT film (Eastman Kodak Co.) with intensifying screens at - 8 0 ° C for 1 wk. 2.8. Transfections and CAT assays

1.8-2.0 x 10 6 NM or 1.8-2.0 X 10 6 MM cells were seeded in 10 cm-dishes and cultured in the appropriate culture medium at 37°C under 5% CO 2 for 24 h. Cells from subconfluent cultures were transfected with 1 /zg

pCHI10 (/3-galactosidase expression vector; Phannacia) and 9 /zg reporter construct in combination with 30 /~g Lipofectin (Gibco/BRL, Life Technologies Ltd., Paisley, UK) in 3 ml serum-free OptiMEM culture medium (Gibco/BRL, Life Technologies Ltd., Paisley, UK) [10]. Two 10 cm-dishes with cells were used for each transfection with reporter construct. After 16 h of culture, 3 ml culture medium containing 30% FCS (supplemented with EGF and HC for NM cells) were added, resulting in a final concentration of 15% FCS. All cells were harvested 48 h later and cell lysates were prepared [35]. /3-galactosidase activity was used as a means to correct for the variation in transfection efficiencies of the various reporter constructs. Amounts of protein corresponding to equal /3-galactosidase activity were used in CAT analysis [15]. After 1 h incubation at 37°C, an additional volume of 20 /xl 4 mM acetyl coenzyme A was added and the incubation was allowed to continue for another hour. Samples were extracted with ethyl acetate and analyzed on silica gel TLC plates (J.T. Baker, Phillipsburg, NJ, USA). CAT activity was quantified by means of a PhosphorImager (Molecular Dynamics, Sunnyvale, CA, USA). In each experiment, CAT activity of the promoterless pSuperCAT construct was subtracted from the values obtained for each construct. After correction, CAT activities were expressed as percentage of the activity of a reference construct. Hence, the mean and standard deviation were calculated based on log-transformed values. TM

2.9. Electrophoretic mobility shift assays (EMSAs)

Nuclear extracts were prepared from either 2.5 X 10 6 NM or 2.5 X 10 6 MM cells according to an earlier described method [37] with minor modifications [44]. Usually, 2 - 4 /zl of nuclear extract contained 4 /xg protein as determined by the method of Bradford [3]. Standard EMSA reactions contained 4 /xg of nuclear extract in a final reaction volume of 20 /~1. In titration experiments, either 0.5, 1, 2 or 4 / z g were used. Reaction mixtures contained: nuclear extract, 10 mM HEPES pH 7.9, 60 mM KC1, 4% ( v / v ) glycerol, 5 mM MgC12, 1 mM CaC12, 1 mM DTT, and non-specific, unlabeled competitor DNA (poly(dI:dC).(dI:dC) (Pharmacia)). The optimal poly(dI:dC).(dI:dC) concentration of 0.4 /xg per 20 /zl reaction volume was determined in titration experiments. After the addition of 1 /zl (7500 cpm//xl) labeled DNA, binding reactions were incubated on ice for 30 min. Reactions were separated on a 4% polyacrylamide gel (Bio-Rad laboratories, Hercules, CA, USA) in 0.25 X TBE (22.5 mM Tris, 22.5 mM Boric acid, 6.25 mM EDTA) at 150 V for 1.5 h. Gels were fixed in 10% M e O H / 10% glacial acetic acid for 15 min, dried on a slab gel dryer and exposed to Kodak X-OMAT film (Eastman Kodak Co.) with intensifying screens at - 8 0 ° C overnight. To define the protein nature of the retarded bands, crude nuclear extracts were heated for 5 min at 90°C or

226

J.-B. Prins et al. / Biochimica et Biophysica Acta 1317 (1996) 223-232

3.2. Mapping of DH sites in the PDGF B-chain gene extended promoter

!

z

actin

PDGF

DNase I hypersensitivity of a genomic DNA region is thought to correlate with accessibility of that region to trans-acting factors [17]. Nuclei from NM-5, Mero-25, and -82, were digested with DNaseI. Subsequently, genomic DNA was isolated and digested with EcoRI or HindlII. The positions of DH sites were determined after gel electrophoresis, blotting and probe hybridization (Figs. 2 and

-

B -

3). DH +0.4, +0.6, and + 1.0 were detected after hybridization of the 22 kb EcoRI and the 5.2 kb HindlI fragments with PR16 and PR7, respectively. These DH sites were present in both NM and MM cell lines. Two additional DH sites, DH + 1.9 and +4.0, were identified within the first intron of the PDGF B-chain gene in MM cell lines, only. The proximal promoter region was found to be DNase I hypersensitive in both NM and MM cell lines. No DH sites were detected in other intron and exon sequences or immediately upstream of the proximal promoter. Examination of the partly overlapping 9.2 kb EcoRI and 8.1 kb HindlII fragments with probes PR3 and PR13/PR12, respectively, resulted in the detection of DH sites far upstream from the transcription start site. DH sites - 8 . 6 and - 9 . 9 were found in both NM and MM cell lines. An additional DH site, DH - 7 . 3 , could only be detected in Mero-25 and -82.

Fig. 1. Nuclear run off analysis. 32P-labeled nuclear RNA of NM-5 and Mero-25 were hybridized to nitrocellulose filters spotted with pUC18, pUC18 + actin insert, and pUC18 + PDGF B-chain insert.

treated with 20 /zg of Proteinase K (Merck, Darmstadt, Germany) at 37°C for 15 min. In competition experiments with sequence specific or random probes, nuclear extracts were first incubated on ice for 10 min with either 0, 100 or 200 ng of unlabeled competitor DNA. Subsequently, 7500 cpm of labeled probe were added and the incubation allowed to proceed on ice for another 30 min. Reactions were separated as described above.

3. Results 3.1. Nuclear run off analysis

3.3. Mapping of the proximal promoter area

Nuclear run off assays were performed to determine the levels of PDGF B-chain nuclear mRNA in NM and MM cell lines (Fig. 1). No nuclear PDGF B-chain mRNA was detectable in NM cells. Neither could PDGF B-chain mRNA be detected after culturing in the presence of the protein synthesis inhibitor cycloheximide (CHX). In contrast, PDGF A-chain transcripts and transcripts of the constitutively transcribed /3-actin gene were readily detectable [26]. In the MM cell lines, nuclear PDGF B-chain transcripts were clearly present.

-1'6

'

-42

'

-8

'

-~

'

PR3

The PDGF B-chain gene proximal promoter region was analyzed by in vivo DMS footprinting. This technique allows the identification of sequences bound by protein factors. The guanine (G) residues - 1 and - 6 1 in the lower strand were found to be hypermethylated in Mero-25, -41, and -82, compared to the corresponding residues in in vitro DMS-treated naked mesothelial cell DNA (Fig. 4). No such hypermethylated residues were detectable in NM cells.

1

2

•

•

;

'

~,

'

;

'

PR16

H

E

H

E

H

H

H

I

L

I

I

I

I

I

I

PR13

2'o'

2'.'

28

'kb

PR9

E

PR12

7 .....

.....

E

E

I

J

PR7

Fig. 2. Partial restriction map and probes used for mapping of DH sites in the extended promoter region of the PDGF B-chain gene. Exons are indicated as numbered boxes. Lower panel: numbered boxes indicate the position of genomic probes. E = EcoRI; H = HindIII.

J.-B. Prins et al. / Biochimica et Biophysica Acta 1317 (1996) 223-232

227

2

1 b

i

-16

I

.ll2

NM-5

f

f

Mero-82

~'

t

10 12 13 14

(s/a) (s/a) (s) (s/a)

I

l

-8

I

I

-4

I

t

I

0

I

I

I

kb

8

4

tiff t

tt~'t

t

t

m m

1 2 3 4 5 6

(s/a) (s) (s) (s) -(s) -(s)

+1 +1 +1 +1

(s) (s) (s) (s)

+ +1 +1 +1 +1 +1

(s) (s) (s) (s) (s)

Fig. 3. Localization of DH sites within, and upstream of the proximal promoter of the PDGF B-chain gene. Exons are indicated as numbered boxes. Arrows indicate DH sites. Fragment l ( s / a ) corresponds to the - 1 1 2 / + 43 'minimal' promoter area. Bars indicate fragments cloned in sense (s) a n d / o r antisense (a) orientation 5' of the l(s) fragment in CAT reporter gene constructs.

to

3.4. Functional analysis of the proximal promoter area ~, The activity of the PDGF B-chain proximal promoter area in NM as well as MM cell lines was analyzed in transfection studies with chloramphenicol acetyl transferase (CAT) reporter gene constructs. Comparable basal promoter activity was obtained with a 1758 bp fragment immediately upstream from the transcription start site in NM and MM cells. Most of the activity was retained when the 1758 bp fragment was narrowed down to 112 or even 65 bp upstream of the transcription start site. The activity of the 112 bp fragment was much reduced when used in the antisense orientation. We tested the importance of the hypermethylated G-61 in a series of transfection experiments with deletion constructs ranging from bp - 6 4 to bp - 3 6 . A reduction in activity was observed between bp - 6 4 and - 6 0 in the MM cell lines (Fig. 5A). The site directed mutant, - 1 1 2 / + 18mutaCAT ( - 65 ATATC - 6 1 instead of - 6 4 TCTC - 6 1 ) was constructed to analyze this sequence in more detail. In the MM cell lines, a significant reduction of activity was obtained with this construct (Fig. 5B). A similar, but not significant effect, was seen in NM cells (Fig. 5B).

+

G-1

"

Z

i._

6

6 x_

m

m

~

+

- G-1

TATA I

G-60 G-61 -u-

- G-60

-L_ G-61

3.5. Electrophoretic mobility shift assays (EMSAs) EMSAs were used to analyze the binding of nuclear protein factors to the proximal promoter. DNA constructs, 50 to 80 bp in length, were end-labeled and tested for their in vitro binding capacity of nuclear proteins. A bandshift pattern consisting of two major bands was obtained with

Fig. 4. In vivo footprint analysis of the human PDGF B-chain gene promoter. Cells were treated with DMS in vivo, DNA was cleaved by piperidine, and G residues were sequenced. Numbers indicate positions of G residues in the lower strand relative to the transcription initiation site in NM-21, Mero-25, and Mero-41. + = in vitro treated naked DNA.

J.-B. Prins et a L / Biochimica et Biophysica Acta 1317 (1996) 223-232

228 10

A

B

"5 t~

T -64

-60

-44

Mero-82

-36

wt

muta

NM

wt

muta

Mero-25

l wt

muta

Mero-41

wt

muta

Mero-82

Fig. 5. Relative CAT activities of NM and MM cell lines transfected with various PDGF B-chain gene promoter/CAT fusion constructs. CAT activities are expressed as percentage conversion after correction for the transfection efficiency. CAT activity from the promoterless pSuperCAT construct was subtracted from each value. Transfection efficiencies were determined by measuring /3-galactosidase activity obtained from the co-transfected expression vector pCH110. Corrected activities are plotted on a logarithmic scale relative to p S i s - 6 4 / + 43CAT ( - 64; A) or p S i s - 1 1 2 / + 18CAT (wt; B), which have a set value of 1. The calculated means of data from 2 - 4 independent transfection experiments are plotted. SD are indicated by error bars.

construct D H l - l W T ( - 1 0 3 / 24) and nuclear extracts from both NM and MM cells. The formation of the bandshift pattern was inhibited after the nuclear extracts were treated with proteinase K or when incubated at 90°C for 5 min. indicating the protein nature of the factors. From titration experiments with increasing amounts of nuclear extract of either NM or MM cell lines, it becomes apparent, that less MM nuclear extract is needed to generate a bandshift pattern than NM nuclear extract (Fig. 6A). Also, higher amounts of unlabeled sequence specific competitor DNA were needed to inhibit the formation of the bandshift pattern of nuclear extracts from MM cells than to inhibit the bandshift pattern of nuclear extracts from NM cells (Fig. 6B). These data suggest, that the nuclear factors responsible for the bandshift are more abundantly present in MM than in NM cells. When D H I - I M U T A ( - 6 5 ATATC - 6 1 instead of - 6 4 TCTC - 6 1 ) was used, the in vitro binding potential of the construct was markedly reduced (Fig. 6A). The bandshift pattern of DH11WT(-103/-24) was retained when probe D H I - I W T 5'(-103/-54) was used. Only the faster moving band was obtained with D H l - l W T - 3 ' ( - 7 3 / - 2 3 ) and nuclear extracts of both NM and MM cells (Fig. 6C).

3.6. Functional analysis of other DH sites All other DH sites were analyzed for their ability to enhance or reduce basal PDGF B-chain gene promoter-induced CAT activity in NM and MM cells. To this end, CAT reporter gene constructs were prepared by cloning sequences representing DH sites upstream and downstream of the 112 bp B-chain 'minimal' promoter [ = construct l(s)] (Fig. 3). These constructs were transfected into the appropriate cell lines. DH sites, DH + 1.9 and + 4.0, could only be detected in MM cells. Construct 5s contained DH + 1.9 and construct 6s contained DH + 4.0. Neither of these constructs had an effect on the level of CAT activity. The integral activity of all DH sites in the - 1 2 / - 6 kb region was analyzed in transfection experiments with the 10(s) and 10(a) constructs. These constructs induced a reduction of activity in NM cells (Fig. 7A). In MM cell lines, construct 10a induced no or a minor increase in basal activity (Fig. 7B and 7C). Construct 10s had no obvious effect in these cell lines. A minor silencing effect was obtained with the 12(s) fragment in NM cells (Fig. 7A). This fragment contains

Fig. 6. EMSA experiments with nuclear extracts of MM or NM cell lines and end-labeled probes. Increasing amounts of nuclear extract of either NM-21 or Mero-25: 0.5, 1, 2, 4 /xg, were incubated with 7500 cpm of D H I - I W T or DHI-IMUTA (A). EMSA experiment with 4 p,g of nuclear extract of either NM-21 or Mero-25 and 7500 epm of D H I - I W T and either 0, 100 or 200 ng of unlabeled, specific competitor DNA (B). EMSA experiment with 4 g g of nuclear extract of either NM-21 or Mero-25 and 7500 cpm of either DHI-1WT, DHI-IWT-5' or DHI-lWT-3' (C). * indicate specifically retarded bands.

J.-B. Prins et al. / Biochimica et Biophysica Acta 1317 (1996) 223-232

229

A NM-21 DHI-1 WI" 0.5 1

2

II

Mero-25

DHI-1 MUTA

4 0.5 1

B NM-21

0

0 0 T-

0 0 0,1

Mero-25

i,

2

4

I,

DHI-1 WT 0.5 1

2

0 0 ',"-

0 0 0~1

DHI-1 MUTA

4 0.5 1

2

4

C t

0

I,

NM-21

a

D

Mero-25

ii

a

D

J.-B. Prins et al. / Biochimica et Biophysica Acta 1317 (1996) 223-232

230

10

100

A

I

o

0 >

T

2

T ~

10

.

i

i ls

lOs

lOa

NM

12s

12a

ls

lOs

lOa

12s

Mero-25

12a

ls

lOs

lOa

12s

12a

Mero-82

Fig. 7. Relative CAT activities of NM (A) and MM (B and C) cell lines transfected with various PDGF B-chain gene promoter/CAT fusion constructs. After correction for transfection efficiencies, CAT activities were determined as percentage conversion. Transfection efficiencies were determined by measuring fl-galactosidase activity obtained from the co-transfected expression vector pCH110. CAT activity of the promoterless pSuperCAT construct was subtracted from each value. Corrected activities are plotted on a logarithmic scale relative to pSis-112/+ 43CAT [l(s)], which has a set value of 1. The calculated means of data from 2-4 independent transfection experiments are plotted. SD are indicated by error bars. ND, not determined.

DH - 9 . 9 and exerts its silencing effect in both the sense (s) and antisense (a) orientation. In Mero-25, and -82, the 12(a) construct resulted in a 2- to 3-fold enhancement of basal activity. The enhancing effect of the 12(s) fragment, however, could only be demonstrated in Mero-82 (Fig. 7C). Neither in NM nor in MM cells did 13(s) and 14(s/a), which contain DH - 8 . 6 and DH - 7 . 3 , respectively, affect basal promoter activity.

4. Discussion Earlier studies showed, that PDGF B-chain m R N A expression was clearly detectable in human MM cell lines, but not in N M cell lines [14,45]. The elevated PDGF B-chain expression in MM cell lines could not be attributed to rearrangements or amplifications of chromosome 22 [45]. Therefore, nuclear run off experiments were conducted to investigate PDGF B-chain m R N A expression at the transcriptional level. Nuclear PDGF B-chain transcripts were detectable in MM cells, while no such transcripts were found in NM cells. PDGF A-chain transcripts, however, were readily detectable in these cell lines [26]. Even after CHX treatment no PDGF B-chain m R N A was detectable in NM cells. Hence, the inability to detect PDGF B-transcripts among total R N A of NM cells cannot

be attributed to messenger instability. Similarly, in glioblastoma and bladder carcinoma cell lines as well as in human umbilical vein endothelial cells CHX treatment did not affect the detection of transcripts [13,32]. In contrast, CHX clearly enhanced steady-state PDGF B-chain m R N A levels in TPA treated HL-60 cells and in resting and LPSand TPA-treated human monocytes [29,36]. The PDGF B-chain m R N A half-life was determined in actinomycin D-treated MM cell lines (data not shown). Its value of 2 to 3 h is similar to earlier reported half-lives in glioblastoma, bladder carcinoma, PC3, HeLa and TPA-treated K562 ceils [2,32]. The elevated steady-state PDGF B-chain m R N A level in MM cell lines, therefore, can be attributed to an increased transcription rate. The extended promoter region of the PDGF B-chain gene was studied to identify structures responsible for the elevated level of transcription in MM cells. The proximal promoter region was found to be DNaseI hypersensitive in NM as well as in MM cell lines. Moreover, the 1758 bp immediately upstream from the transcription start site could direct transcription of the C A T reporter gene in both cell types. Most of the activity was retained when the promoter area was narrowed down to - 1 1 2 or - 6 5 bp relative to the transcription start site [42]. A slight decrease in activity was detected when the promoter area was reduced by another five nucleotides. The elevated level of transcription in MM cells can not be explained by a mutational

J.-B. Prins et al. / Biochimica et Biophysica Acta 1317 (1996) 223-232

event in the 'minimal' promoter area. Sequencing of bp - 393 to + 140 revealed a nucleotide sequence in MM cell lines identical to that in NM cell lines (data not shown). Site directed mutagenesis of the area between bp - 64 and - 6 1 , however, resulted in a clear inhibition of transcriptional activity in MM cells. Only a minor inhibitory effect was observed in NM cells. Therefore, this area was analyzed for its capacity to bind nuclear factors in EMSA experiments. Specific binding of nuclear factors to a 80 bp construct, D H I - I W T ( - 1 0 3 / 24), was demonstrated in both MM and NM cells. However, the concentration of factors was higher in extracts from MM cells than in extracts from NM cells as demonstrated in competition experiments with unlabeled probe DNA and in titration experiments with increasing amounts of nuclear extract (Fig. 6A and B). The lower concentration of factors in the NM cells may explain the results of the footprint experiments, in which factors binding to this region could only be demonstrated in MM cells. In both cell types, binding in EMSA experiments was much reduced using a construct with the mutated sequence, D H l - l M U T A ( - 1 0 3 / 24) (Fig. 6A). In undifferentiated K562 cells and HeLa cells, the - 6 4 / - 61 region is involved in activation of PDGF B-chain gene transcription [8]. Adjacent to this region the core binding site for the ubiquitously expressed transcription factor SP1 (CCACCC) is located at bp - 6 1 to - 5 6 and an overlapping binding site for Egr-1, which interacts with SP1 in endothelial cells [24,25]. The bandshift pattern of D H I - I W T ( - 1 0 3 / - 24) was retained when a 3'-truncated mutant, D H I - I W T - 5 ' ( - 1 0 3 / - 54), was used. With the 5'-truncated mutant, D H l - l W T - 3 ' ( - 7 3 / - 2 4 ) only the faster moving band was obtained (Fig. 6C). These results can be explained through the interaction of nuclear factors with regulatory elements, which have been reported to be present in the 30 bp stretch from bp - 1 0 0 to - 7 0 relative to the transcription start site [22,24,31]. From our experiments it is clear that there are no qualitative differences between NM and MM cells in the binding capacity of nuclear factors to this region, but that rather quantitative differences discriminate MM from NM cells. Binding of factors, therefore, cannot explain the enhanced transcription of the PDGF B-chain gene in MM cells. Methylation of promoter sequences or inaccessibility of the promoter due to nucleosome phasing are not expected to play an important regulatory role either as determined by Southern blot analysis with methylation sensitive restriction enzymes (data not shown) and by mapping of DNaseI hypersensitive sites. DH sites were also found in other regions of the PDGF B-chain promoter. No enhancing nor repressing effects on transcription were exerted by DH - 8 . 6 or DH - 7 . 3 . Neither was transactivating potential detectable for first exon and first intron sequences in MM cell lines (data not shown). This is in agreement with other reports on cervix and prostate carcinoma cell lines [6,7]. A transcription enhancing effect was exerted by construct 12a in MM

231

cells. This construct contains DH - 9 , 9 . In the sense orientation, the enhancing effect could only be demonstrated in Mero-82. When transfected into NM cells, both 12a and 12s had a rather transcription inhibitory effect. We have shown that the difference in PDGF B-chain mRNA expression between NM and MM cells is mainly determined at the transcriptional level. In both cell types, the proximal promoter appears to be nucleosome free and confers basal levels of transcription. Nuclear factors, which may interact with sequences in the proximal promoter, are present in both NM and MM cells. However, their concentrations seem to be higher in MM cells. Interaction of nuclear factors with an enhancer element at position - 9 . 9 kb in MM cells and silencer regions in NM cells in concert with factors binding to the 'minimal' promoter determine the differential level of PDGF B-chain gene transcription in NM and MM cell lines.

Acknowledgements We would like to thank prof. dr. R. Benner and prof. dr. A. Hagemeijer for continuous support, dr. M. Verschuren for setting up the EMSA technology, Mrs. E. Franken for technical assistance, Mr. T. van Os for preparing the figures and Ms. P.C. Assems for secretarial assistance. The work described in this paper was supported by a grant from the Dutch Cancer Society.

References [1] Benham, F.J., Hodgkinson, S. and Davies, K.E. (1984) EMBO J. 3, 2635-2640. [2] Betsholtz, C., Johnsson, A., Heldin, C.H., Westermark, B., Lind, P., Urdea, M.S., Eddy, R., Shows, T.B., Philpott, K., Mellor, A.L., Knott, T.J. and Scott, J. (1986) Nature 320, 695-699. [3] Bradford, M.M. (1976) Anal. Biochem. 72, 248-54. [4] Claesson-Welsh, L., Eriksson, A., Moren, A., Severinsson, L., Ek, B., Ostman, A., Betsholtz, C. and Heldin, C.H. (1988) Mol. Cell. Biol. 8, 3476-3486. [5] Claesson-Welsh, L., Eriksson, A., Westermark, B. and Heldin, C.H. (1989) Proc. Natl. Acad. Sci. USA 86, 4917-4921. [6] Dirks, R.P., Jansen, H.J., Gerritsma, J., Onnekink, C. and Bloemers, H.P. (1993) Eur. J. Biochem. 211,509-519. [7] Dirks, R.P.0 Jansen, H.J., Onnekink, C., De Jonge, R.J. and Bloemers, H.P. (1993) Eur. J. Biochem. 216, 487-495. [8] Dirks, R.P., Jansen, H.J., Van Gerven, B., Onnekink, C. and Bloemers, H.P. (1995)Nucl. Acids Res. 23, 1119-1126. [9] Dodemont, H.J., Soriano, P., Quax, W.J., Ramaekers, F., Lenstra, J.A., Groenen, M.A., Bernardi, G. and Bloemendal, H. (1982) EMBO J. 1, 167-171. [10] Feigner, P.L., Gadek, T.R., Holm, M., Roman, R., Chan, H.W., Wenz, M., Northrop, J.P., Ringold, G.M. and Danielsen, M. (1987) Proc. Natl. Acad. Sci. USA 84, 7413-7417. [11] Franklin, G.C., Donovan, M., Adam, G.I., Holmgren, L., PfeiferOhlsson, S. and Ohlsson, R. (1991) EMBO J. 10, 1365-1373. [12] Garrity, P.A. and Wold, B.J. (1992) Proc. Natl. Acad. Sci. USA 89, 1021-1025. [13] Gay, C.G. and Winkles, J.A. (1991) J. Cell. Physiol. 147, 121-127.

232

J.-B. Prins et al. / Biochimica et Biophysica Acta 1317 (1996) 223-232

[14] Gerwin, B.I., Lechner, J.F., Reddel, R.R., Roberts, A.B., Robbins, K.C., Gabrielson, E.W. and Harris, C.C. (1987) Cancer Res. 47, 6180-6184. [15] Gorman, C.M., Moffat, L.F. and Howard, B.H. (1982) Mol. Cell. Biol. 2, 1044-1051. [16] Groffen, J., Heisterkamp, N., Stephenson, J.R., Van Kessel, A.G., De Klein, A., Grosveld, G. and Bootsma, D. (1983) J. Exp. Med. 158, 9-15. [17] Gross, D.S. and Garrard, W.T. (1988) Annu. Rev. Biochem. 57, 159-197.

[18] Hammacher, A., Hellman, U., Johnsson, A., Ostman, A., Gunnarsson, K., Westermark, B., Wasteson, A. and Heldin, C.H. (1988) J. Biol. Chem. 263, 16493-16498. [19] Heldin, C.H., Johnsson, A., Wennergren, S., Wernstedt, C., Betsholtz, C. and Westermark, B. (1986) Nature 319, 511-514. [20] Heldin, C.H., Ernlund, A., Rorsman, C. and Ronnstrand, L. (1989) J. Biol. Chem. 264, 8905-8912. [21] Jin, H.M., Brady, M.L. and Fahl, W.E. (1993) Proc. Natl. Acad. Sci. USA 90, 7563-7567. [22] Jin, H.M., Robinson, D.F., Liang, Y. and Fahl, W.E. (1994) J. Biol. Chem. 269, 28648-28654. [23] Josephs, S.F., Ratner, L., Clarke, M.F., Westin, E.H., Reitz, M.S. and Wong-Staal, F. (1984) Science 225, 636-639. [24] Khachigian, L.M., Fries, J.W., Benz, M.W., Bonthron, D.T. and Collins, T. (1994) J. Biol. Chem. 269, 22647-22656. [25] Khachigian, L.M., Lindner, V., Williams, A.J. and Collins, T. (1996) Science 271, 1427-1431. [26] Langerak, A.W., Dirks, R.P. and Versnel, M.A. (1992) Eur. J. Biochem. 208, 589-596. [27] Langerak, A.W., De Laat, P.A.J.M., Van der Linden - Van Beurden, C.A.J., Delahaye, M., Van der Kwast, T.H., Hoogsteden, H.C., Benner, R. and Versnel, M.A. (1996) J. Pathol. 178, 151-160. [28] Leveen, P., Claesson-Welsh, L., Heldin, C.H., Westermark, B. and Betsholtz, C. (1990) Int. J. Cancer 46, 1066-1070. [29] Nagaoka, I., Someya, A., Iwabuchi, K. and Yamashita, T. (1991) Comp. Biochem. Physiol. b 100, 313-319. [30] Nister, M., Libermann, T.A., Betsholtz, C., Pettersson, M., Claesson-Welsh, L., Heldin, C.H., Schlessinger, J. and Westermark, B. (1988) Cancer Res. 48, 3910-3918. [31] Pech, M., Rao, C.D., Robbins, K.C. and Aaronson, S.A. (1989) Mol. Cell. Biol. 9, 396-405. [32] Press, R.D., Samols, D. and Goldthwait, D.A. (1988) Biochemistry 27, 5736-5741.

[33] Raines, E.W., Bowen-Pope, D.F. and Ross, R. (1990) Handbook in Experimental Pharmacology 95, 173-262. [34] Ross, R., Bowen-Pope, D.F. and Raines, E.W. (1990) Philos. Trans. R. Soc. Lond. B. Biol. Sci. 327, 155-169. [35] Sambrook, J., Fritsch, E.F. and Maniatis, T. (1989) Molecular cloning. A Laboratory Manual (2nd ed.), Cold Spring Harbor, NY, USA. [36] Sariban, E. and Kufe, D. (1988) Cancer Res. 48, 4498-4502. [37] Schreiber, E., Matthias, P., Muller, M.M. and Schaffner, W. (1989) Nucl. Acids Res. 17, 6419. [38] Seifert, R.A., Hart, C.E., Phillips, P.E., Forstrom, J.W., Ross, R., Murray, M.J. and Bowen-Pope, D.F. (1989) J. Biol. Chem. 264, 8771-8778. [39] Shamah, S.M., Stiles, C.D. and Guha, A. (1993) Mol. Cell. Biol. 13, 7203-7212. [40] Stroobant, P. and Waterfield, M.D. (1984) EMBO J. 3, 2963-2967. [41] Van den Ouweland, A.M., Breuer, M.L., Steenbergh, P.H., Schalken, J.A., Bloemers, H.P. and Van de Ven, W.J. (1985) Biochim. Biophys. Acta 825, 140-147. [42] Van den Ouweland, A.M., Roebroek, A.J., Schalkeu, J.A., Claesen, C.A., Bloemers, H.P. and Van de Ven, W.J. (1986) Nucl. Acids Res. 14, 765-778. [43] Vassbotn, F.S., Ostman, A., Langeland, N., Holmsen, H., Westermark, B., Heldin, C.H. and Nister, M. (1994) J. Cell. Physiol. 158, 381-389. [44] Verschuren, M.C.M., Van Bergen, C.J.L., Van Gastel - Mol, E.J., Bogers, A.J.J.C. and Van Dongen, J.J.M. (1996) J. Immunol. 156, 3806-3814. [45] Versnel, M.A., Hagemeijer, A., Bouts, M.J., Van der Kwast, T.H. and Hoogsteden, H.C. (1988) Oncogene 2, 601-605. [46] Versnel, M.A., Claesson-Welsh, L., Hammacher, A., Bouts, M.J., Van der Kwast, T.H., Eriksson, A., Willemsen, R., Weima, S.M., Hoogsteden, H.C., Hagemeijer, A. and Heldin, C.H. (1991) Oncogene 6, 2005-2011. [47] Wagner, J.C., Sleggs, C.A. and Marchand, P. (1960) Br. J. Ind. Med. 17, 260-271. [48] Waterfield, M.D., Scrace, G.T., Whittle, N., Stroobant, P., Johnsson, A., Wastesou, A., Westermark, B., Heldin, C., Huang, J.S. and Deuel, T.F. (1983) Nature 304, 35-39. [49] Westermark, B. and Heldin, C.H. (1991) Cancer Res. 51, 5087-5092.