Characterization of the promoter for vascular cell adhesion molecule-1 ...

Vol. 267, No. 23, Issue of August 15, pp. 16323-16329.1992 Printed in 11.S. A

THEJOURNAL OF BIOLOGICAL CHEMISTRY 8 1992 by The American Society for Biochemistry and Molecular Biology, Inc.

Characterization of the Promoter for Vascular Cell Adhesion Molecule- 1 (VCAM-1)* (Received for publication, February 27,1992)

Michael F. IademarcoSB, Jay J.McQuillan$, Glenn D. Roseny, and Douglas C. Dean11 From the Deoartments of Internal Medicine and Cell Biology and Physiology, Washington University School of Medicine, St. Louis, Mksouri 631 10

sionmolecule-1(VCAM-1)’(alsoknown asINCAM-110), which is a ligand forthe integrinreceptor very late activation antigen-4 (VLA-4) (8). VLA-4 is found on the surface of T cells and monocytes where it mediates theadhesion of these cells tocytokine-stimulatedendothelium via VCAM-1 on endothelial cells (9). Two other members of this class are ICAM-1 andICAM-X. Like VCAM-1, both of these proteins are cytokine-inducible and they bind to an integrinreceptor, LFA-1, on immune cells (3, 10). ICAM-1 also binds to the integrin MAC-1(11).ICAM-2 is anothermember of this class. It also binds to LFA-1, but, unlike the other members, it is expressed constitutively on endothelialcells (7). The second class is known as selectins and consists of the lectin-likeproteins L-selectin (Mel-14),P-selectin(GMP140), and E-selectin (ELAM-1) that bind carbohydrate receptors on immune cells (2, 12-20). As with VCAM-1, ICAM-1, andICAM-X,theselectinsE-selectinandP-selectinare responsive to cytokines. Therefore, inflammatorycells can be recruited to the endothelium through pathways involving two seemingly unrelatedsets of interactions: members of the immunoglobulin gene superfamily interacting with integrins and selectins binding to carbohydrates. There is evidence that receptors for some of these endothelial cell ligands may signal the nucleus during receptor-ligand binding. For example, the interaction of VLA-4 or LFA-1 on T-cells with VCAM-1 or ICAM-1 on endothelialcells, respectively, induces T-cell antigen receptor-dependent activation of CD4+ T lymphocytes (21, 22), suggesting that the function of these receptor-ligand pairs extends beyond cell-cell interactions to include nuclear signaling. Endothelial cell surfaceligands may also be involved in The recruitment of immune cells to the endothelium is a complex process involving the interactionof ligands or coun- metastasis of tumor cells. Ithas been proposed that the ter receptors on the surface of endothelial cells with receptors interaction of tumor cells with activated endothelium is a on immune cells (1-4). The endothelial cell ligands can be mechanism that facilitates themigration of circulating tumor divided into two classes. The first class consists of proteins cells into tissues (23, 24). VLA-4 is not present on melanothat are members of the immunoglobulin gene superfamily cytes, osteoclasts, and osteoblasts, but its expression is actiosteosarcoma cell lines, which allows (5-7). These ligands are recognized by integrin receptors on vated on melanoma and these tumor cells to adhere to VCAM-1 on cytokine-stimuimmune cells. One protein in this class is vascular cell adhelated endothelialcells (23-26). These interactions arespecific * This work was supported by National Institutes of Health Grants for VCAM-1 and are not mediated by other endothelial cellHL29594 and HL43418 and a Career Investigator Award from the surface ligands (24). However, other endothelial cell surface American Lung Association (to D. C. D.). The costs of publication of ligands may be involved in metastasis of different types of this article were defrayed in part by the payment of page charges. tumor cells. For example, E-selectin has been shown to meThis articlemusttherefore be hereby marked“aduertisement” in diate the binding of colon carcinoma cell lines to activated accordance with 18 U.S.C. Section 1734 solely to indicate this fact. endothelial cells (24),and small cell lung cancers express $ These two authors contributed equally to this article. LFA-1and MAC-1 (27), suggesting thatthey may utilize Supported by NationalInstitutes of HealthTrainingGrant

Vascular cell adhesion molecule-1 (VCAM-1) was first identified as a protein that appears on the surface of endothelial cellsafter exposure to inflammatory cytokines. Through interaction with its integrincounter receptor VLA-4, VCAM-1 mediates cell-cell interactions important for immune function. We have cloned and begun characterization of the promoter for the VCAM-1 gene. In a series of transfection assays into human umbilical vein endothelial cells (HUVECs), we find that silencers between positions -1.641 kilobases and -288 base pairs restrict promoter activity, and that treatment with tumor necrosis factor-a overcomes this inhibition and activates the promoter through two NFKB sites located at positions -77 and -63 base pairs of the VCAM-1 gene. This responsiveness appears cell-specific since constructs containing the VCAM-1 NFKB sites are not responsive to tumor necrosis factor a in the T-cell line Jurkat. The two VCAM-1 NFKB sites,which differ slightly in their sequence, form distinct complexes in gel retardation assays,suggesting that they interact withdifferent NFKB-site binding proteins. The distribution of these proteins could then control activity of the NFKB sites. We conclude that the pattern of VCAM-1 expression in HUVECs is controlled by a combination of these silencers and NFKB sites.

HL07317. 7 Supported by National Institutes of HealthGrant HL02428 Clinical Investigator Award. 11 To whom correspondence and reprint requestsshould be addressed Box 8052, Washington University School of Medicine, 660 S. Euclid, St. Louis, MO63110. Tel.: 314-362-8989; Fax: 314-3628987.

The abbreviations used are: VCAM-1, vascular cell adhesion molecule-1; TNF, tumor necrosis factor-a; CAT, chloramphenicol acetyltransferase; HUVEC(s), humanumbilical vein endothelial cells; VLA-4, very late activation antigen-4; Hepes, 4-(2-hydroxyethyl)-lpiperazineethanesulfonic acid bp, base pair(s); kb, kilobase(s); PCR, polymerase chain reaction.

16323

VCAM-1 Promoter Activity

16324

pg of protein, and reactions were for 16 h. Acetylation of [“C] ICAM-1 to adhere to endothelialcells. chloramphenicol was quantified by thin layer chromatography folEven though selectins (E-selectin and P-selectin) and members of the immunoglobulin gene superfamily (VCAM-1 and lowed by scintillation counting. RNA Analysis-RNA was isolated by the guanidinium isothiocyaICAM-1) have distinct structures and interact with different nate/CsCl method, and primer extension and S1 nuclease mapping receptors, their pattern of expression on endothelial cells is assays were done as described previously (32). Poly(A+) RNA was similar. Therefore, a common mechanism may control expres- isolated by affinity chromatography on oligo(dT)-cellulose (34). Screening of a Genomic Library and DNA Characterization-A sion of these different genes. An understanding of this mechanism should provide further insight into the role of these human lung fibroblast genomic library in XFix (Stratagene) was screened with a fragment of VCAM-1 cDNA (5) between positions ligands on endothelial cells. +13 and +113 bp (see Fig. 4). ThiscDNA was constructed by reverse In addition to endothelial cells, VCAM-1 is also expressed transcriptase/polymerase chain reaction (PCR) of VCAM-1 mRNA. on lymphoid dendritic cells, on stromal fibroblasts in bone Five pg ofpoly(A+)RNA derived from HUVECs that had been treated marrow, in Peyer’s patch in the intestine, and on some tissue for 16hwith10 ng/ml TNF was annealed to aprimer (5’macrophages (28-30). In the bonemarrow, VCAM-1 mediates TTGCTGTCGAGATGAGAAA ATA-3’) located at position +113 bp the interaction between hematopoietic cells, which express of the VCAM-1 gene, and cDNA was synthesized with reverse transcriptase (32).This cDNA was then amplified by PCR using the same VLA-4, and stromal fibroblasts (29, 30). This interaction is 3’ primer along with a 5’ primer (5’-CGGGCCTCACTGGCrequired for maturation of the hematopoieticcells. In lymph- TTCAGGA-3’) located at position +13 bp. The resulting PCR prodoid tissue, VCAM-1 mediates the interaction of maturing B- uct was gel-purified, labeled with 32P, and used as a probe to screen cells with dendritic cells at germinal centers (28). In contrast the human genomic library as described (32). Southern blot analysis was done using the 32P-labeled VCAM-1 to an endothelialcell where its expression is dependent upon inflammatory cytokines, VCAM-1 is expressed constitutively cDNA fragment derived from reverse transcriptase/PCR as a probe (32). Hybridization was done at 42 “C in 4 X SSPE (1X SSPE is 0.18 at these other sites. M NaCl, 10 mM NaH2P04,1 mM EDTA, pH 7.4), 10 X Denhardt’s (1 We began examining VCAM-1 as both a model for how X Denhardt’s is 0.02% polyvinylpyrrolidone, 0.02% Ficoll, and 0.02% expression of endothelial cell surface ligands is controlled and bovine serum albumin), 0.1% sodium dodecyl sulfate, and 0.4 mg/ml a model of a gene that is subject toalternate regulatory sheared, denatured salmon sperm DNA. Blots were washed a t 55 “C mechanisms in differenttissues. As a first step in determining in 0.2 X SSPE. DNA Sequencing-The 5’ end and flanking region of the VCAMhow VCAM-1 expression is controlled at the molecular level, 1 gene was sequenced using the dideoxynucleotide method as dewe have cloned and begun an analysisof the promoter for the scribed previously (32) with synthetic oligonucleotide primers made VCAM-1 gene. Here we show in a seriesof transfection assays on an Applied Biosystems 391 DNA synthesizer.Both strands of that silencers restrictbasal VCAM-1 promoter activity in DNA were completely sequenced. Plasmid Construction-To create 2.18OVCAMCAT, the VCAM-1 human umbilical vein endothelial cells (HUVECs), and that theinflammatorycytokinetumor necrosis factor-a (TNF) genomic clone XVCAM-1-1 was digested with BglI (position +22 bp) can overcome theinhibitory effect of these silencers and and EcoRI (position -2.18 kb), and the resultingfragments were with T4 DNA polymerase and cloned into the EcoRV activate the promoter. This activation requires two adjacent blunt-ended site of the promoterless CAT expression vector SKCAT-Pst/Bam NFKBsites, which formdifferent complexes withnuclear (32). In this vector, the 5’ end of the CAT gene is cloned into the proteins. The activity of these sites appears be to cell-specific, PstI site of Bluescript SK, whereas the 3’ end of the gene is cloned and we suggest that the combination of these sites could be into the EamHI site. In 2.18OVCAMCAT, the orientation of the VCAM-1 gene fragment is such that the endcontaining the EglI site responsible for their cell-specific activity. MATERIALSANDMETHODS

Cell Culture and DNA Transfection-HUVECs were obtained from the American Type Culture Collection and were maintained in Roswell Park Memorial Institute media (RPMI) (16401, 20% fetal calf serum,endothelial growth supplement(Sigma) at 50 pg/ml, and heparin at 100 pg/ml (endothelial cell media). The T-cell line Jurkat was maintained in RPMI 1640 containing 10% calf serum supplemented with iron. HUVECs on 10-cm plates were switched to Dulbecco’s modified Eagle’s media containing 10% fetal bovine serum and transfected with 10 pg of reporter plasmid along with 10 pg of the parent vector Bluescript SK (Stratagene Inc.) by the calcium phosphate method (31).Fourhaftertransfection, cells were rinsed with phosphatebuffered saline, and endothelial cell media were replaced. Jurkat cells were transfected by electroporation using a BTX Transfector 300 (BTX Inc., San Diego, CA) as described (32). Briefly, approximately 1 X lo7 Jurkat cells were suspended in 0.1 ml of RPMI 1640 media containing10 mM Hepes, 70 mM NaC1,2.5 mM KCl, 0.35 mM NazHP04,3.0 mM dextrose, 1.25 mg/ml salmon sperm DNA, and 30 pg of plasmid DNA and subjected to electroporation at 250 V and 950 pF. Two pg of pRSV/L, which contains the Rous sarcoma virus long terminal repeat fused to the firefly luciferase gene (33), were cotransfected into both HUVECs and Jurkats as an internal control, and equal amounts of luciferase activity were added to each CAT reaction. Luciferase assays were done using a bioluminometer (Analytical Bioluminescence Laboratory) as described (33). Experiments were also done in the absence of pRSV/L to ensure that there was no promoter competition between pRSV/L and the VCAM-1 constructs. Thirty-six h after transfection, protein extracts were made, and chloramphenicolacetyltransferase (CAT) activity was determined (32). Approximately 20 pgof protein from HUVECS were used for CAT assays, and reactions were allowed to proceed for 3 h. CAT assays with extracts from Jurkat cells contained approximately 200

at position +22 bp is adjacent to the 5’ end of the CAT gene. The orientation of the insert was determined by DNA sequencing. 288VCAMCAT was constructed by digesting XVCAM-1-1 with EglI, blunt-ending the site with T4 DNA polymerase, digesting with HindIII (position -288 bp), and cloning the resulting VCAM-1 promoter fragment into SKCAT-Pst/Bam that had been digested with EcoRV (blunt-ended BglI site) and HindIII. The orientation of the insert was determined by DNA sequencing. 1.641VCAMCAT was constructed by first digesting 2.18OVCAMCAT with PstI (the 5’ end of the CAT gene) and EamHI (position -1.641 bp of the VCAM-1 gene). The resulting VCAM-1 promoter fragment was then cloned into SKCAT-Xho that had been digested with PstI and BamHI.SKCAT-Xho was constructed by digesting SKCAT-Pst/Bam with PstI (5’ end of CATgene) and EamHI (3’ end of CAT gene), blunt-ending the CAT gene fragment with T4 DNA polymerase, and cloning it into Bluescript SK digested with XhoI. In SKCAT-Xho, the 5’ end of the CAT gene is adjacent to the PstI site in the vector (determined by DNA sequencing). 933VCAMCAT was constructed by digesting 2.18OVCAMCAT with PstI (5’ end of the CAT gene) and StuI (position -933 bp of the VCAM-1 gene) and cloning the resulting VCAM promoter fragment into SKCAT-Pst/Bam thathad been digested with PstI and EcoRV. 32VCAMCAT, 68VCAMCAT, and 68VCAMCAT-M were constructed using synthetic oligonucleotides. To create 32VCAMCAT, complementary oligonucleotides corresponding to theregion between positions -32 and +12 bp were synthesized suchthat, after annealing, a HindIII site is present on the 5’ end and a PstI site is present on the 3’ end. The double-stranded oligonucleotide was then cloned into SKCAT-Pst/Bam that had been digested with PstI and HindIII. 68VCAMCAT was constructed in the same fashion. Synthetic oligonucleotides extending from position -68 to +12 bp were annealed and cloned into SKCAT-Pst/Bam that had been digested with PstI and HindIII.68VCAMCAT-M is identical with 68VCAMCAT except the NFKB site at position -63 bp 5’-GGGATTTCCC-3’ in 68VCAMCAT is changed to 5’-CTCATCTCCC-3’ in 68VCAMCAT-

VCAM-1 Promoter Activity M. This mutation is known to disrupt the binding of nuclear protein (35). The VCAM-1 gene insert in each of the above plasmids was sequenced. 224VCAMCAT, lZOVCAMCAT, and 85VCAMCAT were constructed using PCR. Ten ofng288VCAMCAT that had been digested with HindIII were used as a template in each PCR reaction along with acommon 3’ primer (5”GCGTCTGCAGATACCGCGGAGTGAGGT-3’) located a t position +12 bp.The italicized sequence corresponds to a PstI site that wasused to clone resulting PCR fragments into CATvectors. The 5’ PCR primersfor 224VCAMCAT, lZOVCAMCAT, and 85VCAMCAT are 5“ACATAAGCTTGATG AGGAAAAGCCTGT-3’, 5”ACATAAGCTTGGCTGGGTGTCTG TTAAAC-3’, and 5’-AGTGGAAGCTTCTGCCCTGGGTTTCCCC T - 3 ‘ , respectively. The italicized sequence in each oligonucleotide corresponds to a HindIII site that was used to clone resulting PCR fragmentsintoCAT vectors. ProductsfromPCRreactions were digested with PstI and HindIII and cloned into SKCAT-PstlBarn that had been digested with the same enzymes. The VCAM-1 fragment in eachof the resulting constructswas completely sequenced to ensure that no mutations occurred during PCR. Complementary double-stranded oligonucleotides containing the NFKB site at position -77 bp (5”AGCTTGCCCTGGGTTTCCCCTTGAAGA-3’, the coding strand is shown and italicized portions are from the VCAM-1 promoter) were synthesized such that, after annealing, HindIII sites are present on each end. The annealed oligonucleotides were then cloned into pTACAT (36),a minimal promoter construct containing the SV40 early gene TATA equivalent driving theCAT gene, thathad beendigestedwith HindIIItocreate pTA(-77)CAT. The HindIII site islocated immediately upstream of theTATA box. Likewise, oligonucleotides containingbothNFKB sites (positions -77 and -63 bp) (5”AGCTTGCCCTGG GTTTCCCCTTGAAGGGATTTCCCTCCGCCA-3‘, the coding strand is shown and italicized portions arefrom the VCAM-1 promoter) were cloned into pTACAT to create pTA(-77/-63)CAT. In both pTA(-77)CAT and pTA(-77/-63)CAT, the insert is in the same orientation as in the VCAM-1 gene; orientation was determined by DNA sequencing. Gel Retardation Assays-Nuclear protein extracts were prepared from HUVECs that were either untreated or which had been stimulated for 2 h with 10 ng/ml TNF as described (37). Double-stranded synthetic oligonucleotides were labeledon their5‘ ends with:12Pusing polynucleotide kinase and used as probes in gel retardation assays as described(31). Samples were subjected toelectrophoresison 4% polyacrylamide gels at approximately 4 V/cm a t room temperature. Gels were dried andexposed to x-rayfilm with an intensifying screen a t -70 “C. Probes for sites at positions -77 and -63 bp, and -77 bp alone were described above in construction of pTACAT vectors. The probe for the site at position-63 is described above in the construction of 68VCAMCAT.

16325 D

-63

FIG. 1. Stimulation of VCAM-1 mRNA in HUVECs by TNF and IL-1. A primer extension assay was used to quantify VCAM-1 mRNA levels in response to TNF and IL-1. Cells were treated with 10 ng/ml T N F or IL-1 for 16 h, and RNA was isolated. A primer (5’ACCTGTGTGTGCCTGGGAGGG TATTCAGCT-3’) located 50 bp from the 5’ end of the known VCAM-1 cDNA was used in theassays along with 20 pg of RNA (see “Materials and Methods”). Extension products were separated on an 8% sequencing gel; the size of the extension product and the size of the standards (std. ) is given in nucleotides.

genomic DNA was hybridized to the same VCAM-1 cDNA fragment (Fig. 2B). Bands of similar mobility hybridized to the digests of XVCAM-1-1 and genomic DNA, indicating that the clone is not rearranged and that it is representative of genomic DNA. A single band is observed with each enzyme, suggesting that a single VCAM-1gene is present.A restriction enzyme map of the area surrounding the 5‘ end of the VCAM1 gene is shown in Fig. 2C. Identification of the 5’ End of the VCAM-1 Gene-A combination of S1 nuclease and primer extension analysis was used to define the 5‘ end of the VCAM-1 gene. A single band that mapped to the same location was observed with each assay, indicating that VCAM-1 gene transcription initiates RESULTS from a single site designated+I in Fig. 3 (results not shown). Effect of TNF and IL-1 on Expression of VCAM-1 mRNAAdditionally, the same site was identified in Fig. 1 using a It has been demonstrated that VCAM-1 expression is stimu- different primer for primer extension. lated by TNF and IL-1 in HUVECs and that the activation Sequence of the First Exon and 5’ Flanking Region of the by T N F occurs at the level of VCAM-1 mRNA (5). Here we VCAM-1 Gene-The sequence of the 5’ flanking region and compare the effect of IL-1 and TNF on VCAM-1 mRNA first exon of VCAM-1 is shown in Fig. 3. The gene contains levels. The results demonstrate that TNF increases the level 120 bp of 5‘ untranslated region that is not interrupted by an of VCAM-1 mRNA more effectively than IL-1 (Fig. 1). intron and the firstexon is 184 bp in length. A TATA box is Isolation of a VCAM-1 Genomic Clone-In order to under- located a t position -29 bp, and consensus binding sites for stand the molecular details of how VCAM-1expression is transcription factors are present throughout the 5’ flanking controlled, we began studies to isolate the promoter for the region. VCAM-1 gene. The primer extension assay in Fig. 1 demonPotential binding sites for NFKB are found at positions -63 strates that the5’ end of the VCAM-1 gene is only approxi- and -77 bp), and a potential binding site for AP-1 is located mately 13 nucleotidesbeyond the 5’ end of the published a t position -495 bp. Both of these elements have been shown cDNA sequence (5) (Fig. 1). to be responsive to TNF in other promoters (38-45). A100-bp fragment of VCAM-1cDNA (located between Consensus binding sites for the ets family of proto-oncopositions +13 and +113 bp) was used as a probe to screen a genes (46) are found a t positions -221, -981, and -1033 bp genomic library forclones containing the5’ end of the VCAM- of the VCAM-1 gene. Members of this family are important 1 gene. At least three differentgenomic clones were obtained. in cell cycle control and are thought to play a role in controlA Southern blot of restriction enzyme digests of one clone, ling tissue and developmental specific expression (47-51). XVCAM-1-1, is shown in Fig. 2A. The blot was hybridized to GATA boxes are found at positions -259 and -245bp. the samecDNA fragment used to screen thegenomic library. Theseelementscanbind afamily of zinc-finger nuclear A Southernblot of restriction enzymedigests of human proteinsthat differ intheirtissue specificity: GATA-1 is

Activity VCAM-1 Promoter

16326

important in red cell differentiation (52), GATA-2 is important for expression of endothelin-1 and, thus, mayplay a role in endothelial cell differentiation (53, 54), and GATA-3 isTcell specific (55).

21 2 -

9 46 6-

-

. )

. ""...I.

,,

21 2 9 4-

6 6-

4 4-

4 4-

23. 2 00

0 56-

2 32 0-

C. @

fiPJ+&-JJ.~& -6 k l l PSt I

1-23>

B I1

FIG. 2. Southern blot analysis of the VCAM-1 gene. A , Southern blot of restriction enzyme digests of XVCAM-1-1, a genomic clone that contains the 5' end of the VCAM-1 gene. The probe for the blot was a cDNA fragment between positions +13 and +113 bp of the VCAM-1 gene (see Fig. 3). B , Southern blot of human genomic DNA. Fifteen pgof human genomic DNA were digested with the indicated restriction enzyme, separated by agarose gel electrophoresis, and Southern blotted. The blot was hybridized to theprobe described in panel A . Note that the size of the restriction fragments match those obtained with XVCAM-1-1 in panel A (the Hind111 fragment ran off the endof the gel), indicating that theclone has not undergone rearrangement. C, a restriction enzyme map of the 5' end and flanking region of the VCAM-1 gene. This map was derived from Southern blots of XVCAM-1-1 that had been digested with the indicated enzymes and hybridized to thesame probe as in panels A and B. -2180

-2060 -1940 -1820 -1700 -1 580

-1460 -1340

FIG. 3. Sequence of the 5' endand flanking region of human the VCAM-1 gene. The site of transcriptional initiation is indicated by +I and an arrow. Boxed sequences correspond to potential regulatory elements and are discussed in the text.

-1220

-lloo -860 -740

-620 -500 -380 -260

-1 40

-20

+lo1 +208

Three potential octamer binding sitesfound are at positions -729, -1180, and -1554 bp. Octamer binding sites arefound in a number of different enhancers and theybind a family of homeobox proteins that are differentially expressed during developmentand in adulttissues (56, 57). While octamer binding sites arenormally thought to mediate transcriptional activation, they can alsobe targets of repression (58). Finally, there is a consensus binding site for TEF-1 (position -719 bp), a tissue-specific enhancer factor that binds to the GT-IIC, Sph-I, and Sph-I1 enhansons in the SV40 enhancer (59). While our studies were being completed, the sequence and chromosomal localization of the VCAM-1 gene was reported (60). In this study, the sequence of the first 520 bp of 5' flanking sequence was reported. Comparison of the two sequences in the region of overlap revealed complete identity. However, there was a discrepancy in the identification of the 5' end of the gene. Using a combination of primer extension and S1 nuclease protection assays, we find that transcription initiates at a G residue 5 bp 3' of the A residue suggested previously to be the 5' end of the gene. The VCAM-1 Promoter Is TNF-responsive and Cell-specific-To determine if the 5' flanking region of the VCAM-1 gene functions as a promoter, a 2.18-kb fragment of VCAM1 5' flanking sequence was fused to the CAT reporter gene, andtheresultingconstruct (2.18OVCAMCAT) was transfected into HUVECs. Basal expression was low, but significant activation occurred in the presence of T N F (Fig. 4A). Additionally, transcription from 2.18OVCAMCAT initiated from the same site as in the endogenous VCAM-1 gene (data not shown). These results suggest that the first 2.18 kb of VCAM-1 gene 5' flanking region is an active promoter and thatitsactivity reflects the pattern of expression of the endogenous VCAM-1 gene in HUVECs. As a control, 2.180VCAMCAT was transfected into theTcell line Jurkat,which is responsive t o T N F(61) butdoes not express VCAM-1 (data not shown). The promoter showed little basal activity in Jurkat cells and was unresponsive to T N F (Fig. 4, right). Silencers RestrictVCAM-1 Promoter Activity in theAbsence

VCAM-1 Promoter Activity

16327

expression of 224VCAMCAT. However, expression of 224VCAMCAT is still activated by TNF, indicating that the TNF GATA boxes arenotimportant for T N F responsiveness. I + I I I Subsequent deletions to positions -120 (120VCAMCAT) and -85 bp (85VCAMCAT) also failed to eliminateT N F responsiveness. However, responsiveness was lost on deletion to position -68 bp (68VCAMCAT). This deletion removes an NFKB site atposition -77 bp but leaves a second NFKB site a t position -63 bp. These results suggest that the upstream NFKB site is requiredfor T N F responsiveness but do not preclude the possibility that thedownstream site plays a role. FIG. 4. The VCAM-1 5’ flanking region is an active pro- The most active VCAM-1 promoter construct in HUVECs, moter that is TNF-responsive and tissue-specific. Left, trans- 288VCAMCAT, was not active in Jurkat cells and it did not fection of 2.18OVCAMCAT, which contains 2.18 kb of the VCAM-1 respond to treatmentwith T N F (data notshown). Since NFKB gene 5’ flanking region fused to the CATgene, into HUVECs. Right, sites from other promoters are responsive to TNF in Jurkat transfection of 2.18OVCAMCAT into Jurkatcells. Percent conversion of [‘4C]chloramphenicol to acetylated formswas determined by scin- cells (39-42), tissue-specific factors could be involved in the response of the VCAM-1 promoter to TNF. tillation counting. Where indicated, cells were treated with 10 ng/ml Two NFKB Sites AreRequired for TNF ResponsivenessT N F for 16 h. RSVCAT (right lane in each panel) contains the Rous sarcoma virus long terminal repeat driving the CATgene. T o determine if the NFKB sites at positions -63 and -77 bp were able to convey T N F responsiveness on a heterologous promoter, theywere cloned into pTACAT, which contains the SV40 early gene minimal promoter (containsonly the TATA equivalent) fused to the CAT gene (36). In transfectionassays into HUVECs, these NFKB sites significantly stimulated basal expression and also were sufficient to mediate T N F responsiveness (Fig. 6). We show in Fig. 5 that deletion of the NFKB site a t position -77 bp eliminated T N F responsiveness, indicating that this site isrequired and that the site at position -63 bp is not sufficientfor the effect. The NFKB site at position -77 bp was then cloned into pTACAT to determine if it alonecould mediate T N F responsiveness. This construct was not TNF-responsive; therefore, we conclude that both FIG. 5 . Effect of 5’ deletions on VCAM-1 promoter activity. NFKB sites arerequired for the effect (Fig. 6). Constructs containing different amounts of VCAM-1 gene 5’ flanking NFKBSites at Positions -77 and -63 bp Form Different region were transfected into HUVECs as in Fig. 4. Numbers in the constructnameindicatethelength of VCAM-1gene 5’ flanking Protein Complexes-Binding of nuclear proteins to the NFKB region. Constructs are described under “Materials and Methods.” In sites was examined using gel retardation assays. Three com68VCAMCAT-M, theNFKBsiteat position -63 bp is mutated. plexes of different mobility were observed when both sites SKCAT-PstlBam is the parent vector and does not contain a pro- were present in theprobe (Fig. 7). Competitibn with an excess moter. Relative CAT activity was determined by first separating the of the site atposition -77 bp selectively inhibited formation acetylated and unacetylated forms of [’4C]chloramphenicol using thin layer chromatography (as in Fig. 4) and then quantifying the forms of complex 3, whereas competition with the site at position by scintillation counting. Results are an average of duplicate assays -63 bp selectively inhibited formation of complexes 1 and 2. HUVEC

”

Jurkat

”

from two separate experiments.

of Cytokines-VCAM-1 is not expressed onHUVECsin theabsence of inflammatory cytokines, and, accordingly, promoter activity is not above background in transfection assays (Fig. 5). Deletion of VCAM-1 5’ flanking region from position -1.641 kb (1.641VCAMCAT) to position -933 (933VCAMCAT) stimulated promoter activity (Fig. 5), indicatingthat asilencer that normally suppressespromoter activity is removed with this deletion. A subsequent deletion to position -288 bp (288VCAMCAT) further stimulated promoter activity, implying that this deletion removes an additional silencer. Treatment with T N F overcomes the inhibitory effect of these silencers and activates theVCAM-1 promoter (Fig. 5). However, the silencers do blunt the response to TNF. Compare the effect of TNF on expression of 2.180VCAMCAT (contains silencers) and 288VCAMCAT (lacks silencers) in Fig. 5. The Effect of 5’ Deletions on TNF Responsiveness of the VCAM-1 Promoter-Deletion of VCAM-1 5’ flanking region from position -288 bp to position -224 bp (224VCAMCAT) decreased promoter activity(Fig. 5). GATA boxes a t positions -259 and -254 bp are the only apparent elements in this region (Fig. 3), implying that these sites are important for

FIG. 6. Both the NFKBsite at position -77 bp and the site at position -63 bp are required for TNF responsiveness in HUVECs. Both NFKB sites (positions -77 and -63 bp) or only a single site (position -77 bp) were cloned into the minimal promoter construct pTACAT (36), which contains only the SV40 early gene TATA equivalent driving the CAT gene, to determine the effect of these sites on a heterologous minimal promoter. The resulting constructs, pTA(-77/-63)CAT and pTA(-77)CAT, and the parentvector were transfected into HUVECs, and CATactivity was determined as in Fig. 5.

Activity VCAM-1 Promoter

16328 Probe

-77/.63

proteins could be a target for constitutive activation of the VCAM-1 promoter in transformed cells where cell cycling is aberrant. TNF-responsiveness of the VCAM-1promoter is dependent upon two adjacent NFKBsites at positions -77 and -63 bp. Thesiteat position -77 bp is similar toanNFKB site identified recently inthe c-myc promoter (63).The c-myc site was shown to be serum-responsive and seems to be important in activation of gene expression during the G,/G, transition. The site at position -63 bp is similar in sequence to an NFKB site in the E-selectin promoter (35). It was shown that this E-selectin promoter site binds nuclear proteins in response to TNF and it competes efficiently with other NFKB sites for FIG. 7. NFKBsites atpositions -77 and -63 bp form differbinding of NFKB-like proteins. However, another group dement complexes with nuclear proteins. Probes containing either both NFKBsites or individual sites were usedin gel retardation assays onstrated that thissite, like the site at position -63 bp in the with nuclear protein extracts prepared from HUVECs that were either VCAM-1 promoter, is necessary but not sufficient for TNF untreated (-) or that had been treated for 2 h with 10 ng/ml TNF responsiveness of the E-selectin promoter (43). (+). Approximately 2 ng of probe and 20 pg of nuclear extract were Interestingly, the NFKBsites from the VCAM-1 promoter used in each assay. In competition assays, 50 ng of the indicated are notTNF-responsive in theT-cell line Jurkat even though unlabeled competitor was included. The numbers 1-3 denote specific complexes, whereas N S denotes a nonspecific complex. The nonspe- NFKBsites from other promoters are known to be responsive cific complex was not seen with the probe containing only the position to TNF in this cell line (61), suggesting that the VCAM-1 NFKB sites are cell-specific. Since a family of proteins are -63-bp site. known to interactwith NFKBsites (39, 64-66), it is conceivConsistent with these competition assays, the site atposition able that the NFKB sites in the VCAM-1 promoter bind -77 bp only formed complex 3, whereas the site a t position selectively to cell-specific proteins. In support of this possi-63 bp formed predominantly complexes 1 and 2. Therefore, bility, we show that the two VCAM-1 NFKBsites appear to the two NFKB sites appear to interactwith different subsets interact with distinct nuclear proteins, indicating that different NFKBsites can discriminate between NFKBsite binding of NFKB sitebinding proteins. proteins. Since both NFKBsites are required for TNF responsiveness and the two sites seem to interact with different DISCUSSION nuclear proteins, it is also possible that the specific combiAs a firststep indiscerning the molecular events that nation of proteins interacting with the sitescould be responcontrol VCAM-1 expression, we have cloned and begun a sible for their cell-specific activity. characterization of the promoter for the VCAM-1 gene. Here Finally, the VCAM-1 promoter contains consensus binding we show that theactivity of the VCAM-1promoter determines sites for GATA binding proteins, ets, and octamer binding the pattern of VCAM-1 expression and that this activity is proteins, which are all criticalfor development. Furthermore, dependent upon the interplay between silencer elements and GATA-1 and ets-1 maybe important indifferentiation of cytokine-dependent enhancers. endothelium (51,53). The presence of binding sites for develSilencers located between positions -1.641 kb and -288 bp opmentally specific proteins in the VCAM-1 promoter could in the VCAM-1 promoter are responsible, at least in part,for indicate a role for VCAM-1 in development. In support of preventing expression of VCAM-1 in HUVECs in theabsence this possibility, we have found that VCAM-1 is expressed in of inflammatory cytokines. There areat least two silencers in a developmentally specific fashion in skeletal muscle where this region: one located between positions -1.641 and -933 it, along with its counter receptor VLA-4, appears to play a bp and a second between positions -933 and -288 bp. Inter- role in muscle differentiation (67). Continuing studies with estingly, these regions contain consensus binding sites for the VCAM-1 promoter should provide further insight into the octamer binding proteins (positions -1554, -1180, and -729 mechanisms that control the intriguing pattern of VCAM-1 bp). In two of these sites (positions -1554 and -1180 bp), expression in different tissues. there is a substitution of an A for the G in the consensus REFERENCES sequence (ATTTGCAT to ATTTACAT), which results in an 1. Albelda, S. M. (1991)Am. J. Respir. Cell Mol. Biol. 4, 195-203 element identical withthat bound by MAT a2, a yeast hom2. Bevilacqua, M., Butcher, E., Furie, B., Furie, B., Gallatin, M., Gimbrone, M., Harlan, J., Kishimoto, K., Lasky, L., McEver, R., Paulson, J., Rosen, eobox protein that functions as a silencer (62). Additional S., Seed, B., Siegelman, M., Springer, T., Stoolman,L., Tedder, T., Varki, studies are currently under way to more precisely localize and A., Wagner, D., Weissman, I., and Zimmerman, G. (1991)Cell 67,233 3. Shimizu, Y., Newman, W., Gopal, T. V., Horgan, K. J., Graber, N., Beall, characterize the silencer elements in theVCAM-1 promoter. L. D., van Seventer, G. A., and Shaw, S. (1991)J. Cell Biol. 113, 1203In contrast to endothelial cells, VCAM-1 is constitutively 1212 expressed on lymphoid dendritic cells, bone marrow fibro4. Springer, T. A., and Lasky, L. A. (1991)Nature 349,196-197 5. Osborn, L., Hession, C., Tizard, R., Vassallo, C., Luhowskyj, S., Chi-Rosso, blasts, and certain tissue macrophages (28-30). It is possible G., and Lobb, R. (1989)Cell 59,1203-1211 that constitutive expression occurs because silencer elements 6. Staunton, D. E., Marlin, S. D., Stratowa, C., Dustin, M. L., and Springer, T. A. (1988)Cell 52,925-933 are not active in these cell types. In addition to these normal 7. Staunton, D. E., Dustin, M. L., and Springer, T.A. (1989)Nature 339, cells, VCAM-1 is expressed constitutively on endothelial cell 61-64 8. Elices, M. J., Osborn, L., Takada, Y., Crouse, C., Luhowskyj, S., Hemler, lines (29), suggesting that its expression is activated during M. E., and Lobb, R. R. (1990)Cell 60,577-584 9. Hemler. M. E., Elices. M. J.. Parker,. C... and Takada, Y.(1990)Immunol. transformation. A transformation-dependent loss of silencer Reu. 114,45-65 activity could cause constitutive expression of VCAM-1 in 10. Dustin, M. L., and Springer,T. A. (1988)J. Cell Biol. 107, 321-331 S., Staunton, D. E., de Fougerolles, A. R., Stacker, S. A,, A., these cell lines. However, it is conceivable that transforma- 11. Diamond, M. S., Garcia-Aguilar, J., Hibbs, M. L., and Springer, T. A. (1990)J. Cell Biol. tion-dependentactivation of VCAM-1 involves elements 111,3129-3139 S., and other than the silencers. The VCAM-1 promoter contains AP- 12. Bevilacqua, M. P.,Pober, J. S., Mendrick, D. L., Cotran, R. S., Gimbrone, J. (1987)Proc. Natl. Acad. Sei. U.S. A. 84, 9238-9242 1,NFKB, and etssites which are all known to be involved in 13. Bevilacqua, M.M.P., Stengelin, S., Gimbrone, M. J., and Seed, B. (1989) cell cycle control. Therefore, a binding site for one of these Science 243, 1160-1165

r-"%

Activity VCAM-1 Promoter 14. Geng, J. G., Bevilacqua, M. P., Moore, K. L., McIntyre, T. M., Prescott, S. M., Kim, J. M., Bliss, G. A,, Zimmerman, G. A., andMcEver, R. P. (1990) Nature 3 4 3 , 757-760 15. Johnston, G. I., Cook, R. G., and McEver, R. P. (1989) Cell 56,1033-1044 16. Lawrence, M. B., and Springer, T. A. (1991) Cell 65,859-873 17. Moore, K. L., Varki, A,, and McEver, R. P. (1991) J. Cell Biol. 1 1 2 , 491499 18. Phillips,M. L., Nudelman, E., Gaeta, F. C., Perez,M.,Singhal, A. K., Hakomori, S., and Paulson, J. C. (1990) Science 2 5 0 , 1130-1132 19. Spertini, O., Kansas, G. S., Munro, J. M., Griffin, J. D., and Tedder, T. F. (1991) Nature 3 4 9 , 691-694 20. Walz. G.. Aruffo. A.. Kolanus. W.. Bevilacaua. M.. and Seed. B. (1990) Science 2 5 0 , 1132-1135 21. Van Sevente, G. A,, Shimizu, Y., Horgan, K. J., and Shaw, S. (1990) J. Immunol. 144.4579-4586 22. Damle, N. K., and Aruffo, A. (1991) Proc. Natl. Acad. Sci. U. S. A. 8 8 , 6403-6407 23. Rice, G. E., and Bevilacqua, M. P. (1989) Science 2 4 6 , 1303-1306 24. Taichman, D. B., Cybulsky, M. I., Djaffar, I., Longenecker, B. M., Teixido, J., Rice, G. E., and Aruffo, A. (1991) Cell Regul. 2 , 347-355 25. Albelda, S. M., Mette, S. A,, Elder, D. E., Stewart, R., Darnjanovich, L., Herlyn, M., and Buck,C. A. (1990) Cancer Res. 50,6757-6764 26. Horton, A. N., and Davies, J. (1989) J. Bone Miner. Res. 4,803-808 27. Feldman, L. E., Shin, K. C., Natale, R. B., and Todd, R.(1991) Cancer Res. 5 1 , 1065-1070 28. Freedman, A. S., Munro, J. M., Rice, G. E., Bevilacqua, M. P., Morimoto, C., McIntyre, B. W., Rhynhart, K., Pober, J. S., and Nadler, L. M. (1991) Science 249,1030-1033 29. Miyake, K., Medina, K., Ishihara, K., Kimoto, M., Auerbach, R., and Kincade, P. W. (1991) J. Cell Biol. 1 1 4 , 557-565 30. Ryan, D. H., Nuccie, B. L., Abboud, C. N., and Winslow, J. M. (1991) J. Clin. Inuest. 8 8 , 995-1004 31. Bowlus. C. L.. McQuillan, J. J.. and Dean, D. C. (1991) J. Biol. Chem. 266, 1122-1127 32. Rosen, G. D., Birkenmeier, T. M., and Dean, D. C. (1991) Proc. Natl. Acad. Sci. U. S. A. 88,4094-4098 33. de Wet, J. R., Wood, K. V., DeLuca, M., Helinski, D. R., and Subramani, S. (1987) C."" ~ l l Riol. . 7. 725-737 ~" , Mol. . ~, ". 34. Aviv, H., and Leder, P. (1972) Proc. Natl. Acad. Sci. U. S . A . 6 9 , 14081412 35. Montgomery, K. F., Oshorn, L., Hession, C., Tizard, R., Goff, D., Vassallo, C., Tarr, P. I., Bomsztyk, K., Lobb, R., Harlan, J. M., and Pohlman, T. H. (1991) Proc. Natl. Acad. Sci. U. S. A. 8 8 , 6523-6527 36. Weintraub, S. J., and Dean, D. C. (1992) Mol. Cell. Biol. 1 2 , 512-517 37. Dignam, J. D., Lebowitz, R. M., and Roeder, R. G. (1983) Nucleic Acids Res. 1 1 , 1475-1489 38. Sen, R., and Baltimore,D. (1986) Cell 4 6 , 705-716 39. Israel, A., LeBail, O., Hatat, D., Piette, J., Kieran, M., Logeat, F., Wallach, D., Fellous, M., and Kourilsky, P. (1989) EMBO J. 8 , 3793-3800 40. Lowenthal, J., Ballard, D., Bogerd, N., Bohnlein,E., and Greene, W. (1989) J . Immunol. 142, 3121-3128 ~I

~~~

""

I

16329

41. Osborn, L., Kunkel, S., and Nabel, G. (1989) Proc. Natl. Acad. Sci. U. S. A. 85. 1482-1486 42. Sen,R:, and Baltimore, D. (1986) Cell 47,921-928 43. Whelan, J., Ghersa, P., Hooft,V. H. R., Gray, J., Chandra, G., Talabot, F., and DeLamarter, J. F. (1991) Nucleic Acids Res. 1 9 , 2645-2653 44. Brenner, D. A,, O'Hara, M., Angel, P., Chojkier, M., and Karin, M. (1989) Nature 337,661-663 45. Dixit, V. M., Marks, R. M., Sarma, V., and Prochownik, E. V. (1989) J . Biol. Chem. 264.16905-16909 46. Watson, D. K., Ascione, R., and Papas, T. S. (1990) Crit. Reu. Oncog. 1, 409-436 47. Bhat, N. K., Fisher, R.J., Fujiwara,S., Ascione, R., and Papas,T. S. (1987) Proc. Natl. Acad. SCL.U. S. A . 8 4 , 3161-3165 48. Bhat, N. K., Thompson, C. B., Lindsten, T., June, C. H., Fujiwara, S., Koizumi, S., Fisher, R. J., and Papas, T. S. (1990) Proc. Natl. Acad. Sci. U. S. A . 87. 1792-2797 49. Chen, Z. Q.,B Izdett, L ,A ,:! Seth, A,, K,, Lautenberger, J. A,, and Pappas, T. S. (1990) science m u , I 416-1418 50. Seth, A. K., Watson, D. K., BIair, D. G., and Papas, T. S. (1989) Proc. Natl. Acad. Sci. U. S. A 86,7833-7837 51. Vandenbunder, B., Pardanaud, L., Jafferedo, T., Mirabel, M. A,, and Stehelin, D.; (1989) Deuelo ment 107,265-274 52. Whitelaw, E.. rsai, S. F., d g b e n , P., and Orkin, S. H. (1990) Mol. Cell. Biol. 10I.,6k96-6finfi 53. Lee, M . E ., Temizer, D. H., Clifford, J. A,, and Quertermous, T. (1991) J. Rir 11.. Chern. 266,16188-16192 ._ 54. Wilscm, D. B., Dorfman, D. M., and Orkin, S. H. (1990) Mol. Cell. Biol. 1 0 , iA-ARfi? 4at. 55. Joulin, V., Bories, D., Eleouet, J. F., Labastie, M. C., Chretien, S., Mattei, M. G., and Romeo, P. H. (1991) EMBO J. 10,1809-1816 56. Kemler, I., and Schaffner, W.(1990) FASEB J . 4 , 1444-1449 57. Pierani, A., Heguy, A., Fujii, H., and Roeder, R. 1G.(1990) Mol. Cell. Biol. Y II V - Y

1 "

I

1 I

""

I

____

-1 n. - ,w - -n.4.--~-m - -5

58. Lenardo, M. J., Staudt, L., Robbins, P., Kuang, A,, Mulligan, R. C., and Baltimore, D. (1989) Science 2 4 3 , 544-546 59. Xiao, J. H., Davidson, I., Matthes, H., Gamier, J. M., and Chambon, P. (1991) Cell 65,551-568 60. Cybulsky, M.I., Frles, J. W. U., Williams, A. J., Sultan, P., Eddy,R., Byers, M. Shows T., Gimbrone M. A., and Collins, T. (1991) Proc. Natl. Acad. Sei. U. S. 8 8 , 7859-7t363 61. Nabel, G., and Baltimore,D. (1987) Nature 3 2 6 , 711-713 62. Porter, S. D., and Smith, M. (1986) Nature 3 2 0 , 766-768 63. Baldwin, A. B., Azikhan, J. C., Jensen, D. E., Beg, A. A,, and Coodly, L. R. (1991) Mol. Cell. Bid. 1 1 , 4943-4951 64. Ghosh, S., Gifford, A,, Riviere, L., Tempst, P., Noland, G., and Baltimore, D. (1990) Cell 6 2 , 1019-1029 65. Kieren, M., Blank, V., Logeat, F., Vondekerckhove, J., Lottspeich, F., LeBail, O., Urban, M., Kourllsky, P., Baeurele, P., and Israel, A. (1990) Cell 6 2 , 1007-1018 66. Nolan, G., Ghosh, S., Liou, H.-C., Tempst, P., and Baltimore, D. (1991) Cell 64,961-969 67. Rosen, G. D., Sanes, J. R., La Chance, R., Cunningham, J. M., Roman, J., and Dean, D. C. (1992) Cell, in press

A.