The Cell Adhesion Molecule CD31 Is Phosphorylated after Cell ...

15 downloads 0 Views 4MB Size Report
adhesion molecule 1 (9) or endothelial cell adhesion molecule. (13), the cellular distribution of CD31 is not restricted to these two cell types. By indirect ...
Vol. 267, No.8, Ieaue of March 15, pp. 5243-5249,1992 Printed in U.S.A.

THEJOURNAL OF BIOLOGICAL CHEMISTRY

0 1992 by The American Society for Biochemistry and Molecular Biology, Inc.

The Cell Adhesion MoleculeCD31 Is Phosphorylated after Cell Activation DOWN-REGULATION OF CD31 IN ACTIVATED T LYMPHOCYTES* (Received for publication, June 6, 1991)

James L. ZehnderS, Keiji Hirai, Margaret Shatsky, John L. McGregorQ,Lee J. Levittq, and Lawrence L. K. Leungll From the Division of Hematology, Stanford University School of Medicine, Stanford, California 94305-5112 and the glnstitut National de la Sante et de la Recherche Medicale U331, Lyon, France

We report the independent cloning of the cDNA for CD31, a recently described cell adhesion molecule of the immunoglobulin gene superfamily presenton platelets, granulocytes, monocytes, lymphocytes, and endothelial cells. Northern analysis revealed three major mRNA transcripts inJurkat (a human T cell line) and KS62 and HEL (leukemia cell lines) cells with an additional 5.3-kilobase transcript seen in culturedhuman umbilical vein endothelial cells. Following T cell activation, CD3 1mRNA was down-regulatedby Northern analysis, and decreased CD31 protein expression was confirmed by immunoblots. The down-regulation of CD31 was partially mediated by decreased transcription as demonstrated by nuclear run-on studies. CD31 became rapidly phosphorylated inplatelets, Jurkat cells, and endothelial cells after cell activation. We were unable to demonstrate the presence of a phosphotyrosine in CD31 using monoclonal and polyclonal phosphotyrosine antibodies. In addition, CD31 phosphorylation in platelets was induced by phorbol ester and was blocked by staurosporin, a protein kinase C inhibitor, suggesting that CD31phosphorylation is mediated by protein kinase C andinvolves serine and/or threonine residues. The phosphorylation of CD31 following cell activation may modulate its cellular adhesiveness, and the down-regulation of its expression may serve to impart target specificity and to localize effector lymphocytes to areasof inflammation.

Cell adhesion molecules mediate important and diverse intercellular events such as organ and tissuedevelopment, the immune response to foreign antigens, lymphocyte homing and trafficking, vascular hemostasis and thrombosis, and tumor metastasis (1, 2). At least three distinct groups of adhesion molecules are presenton hematopoietic cells and thevascular endothelium: the integrin and lectin-epidermal growth factor-

* This work was supported in part by a grant from the Giannini Foundation (to L.L. K. L.)and by National Institutes of Health Grant lROl-HL35774 (to. L. J. L.). Production of LYP21 was supported by Association pour la Recherche Contre le Cancer (Subvention 6586) and the Ligue Nationales Francaise Contre le Cancer. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. $ Recipient of Postdoctoral Fellowship Grant 89-66 from the American Heart Association, California Affiliate. ll Recipient of Research Career Development Award KO4 HL0221301 from the National Institutes of Health. )I Established Investigator of the American Heart Association.

complement-cell adhesion molecule (also known as selectin) families and the immunoglobulin gene superfamily (1, 3,4). CD31 is a member of the immunoglobulin gene superfamily. Members of the superfamily are functionally diverse and include the major histocompatibility complex class I and I1 molecules, T cell receptor and several other T cell antigens, intercellular adhesion molecules 1and 2, neural cell adhesion molecule, carcinoembryonic antigen,and platelet-derived growth factor receptor (4,5). Thesemolecules share acommon ancestral gene, participate in cell-cell recognition reactions, and are structurally similar in having a variable number of extracellular immunoglobulin-like folds resembling immunoglobulin variable region or constant regions. Some of the molecules of the superfamily are used by viruses to infect cells; human immunodeficiency virus utilizes the CD4 receptor (6), whereas the majority of rhinoviruses are tropic to ICAM-1 (7,8). Newman et al. (9) recently reported the cloning of the cDNA for CD31 from ahuman endothelial (EC)l cDNA library, subsequently confirmed by others (10, 11).Analysis of the translatedsequence revealed a 738-amino acid protein with six immunoglobulin-like extracellular domains, a 19amino acid putative transmembranedomain, and a118-amino acid cytoplasmic tail (9). There are multiple potential N linked glycosylation sites, and the protein is substantially glycosylated (11, 12). Whereas CD31 has been termed platelet-endothelial cell adhesion molecule 1 (9) orendothelial cell adhesion molecule (13), the cellular distribution of CD31 is not restricted to these two cell types. By indirect immunofluorescence, CD31 is detected in peripheral blood monocytes, granulocytes, and a subpopulation of lymphocytes (11, 14) in addition to plateletsand EC. CD31 is localized to points of intercellular contact in confluent endothelial cells (9, 13, 15). Polyclonal antibodies directed against CD31 prevented EC from developing intercellular contacts and achieving confluence (13). More recently, mouse L cells transfected with CD31 cDNA demonstrated calcium-dependent aggregation (16), lending support to thehypothesis that CD31 functions as anintercellular adhesion molecule. In the process of characterizing a new monoclonal antibody raised against activated platelets, we have independently cloned the CD31 cDNA from a HUVEC cDNA library. We The abbreviations used are: EC, endothelial cell(s); HUVEC, human umbilical vein endothelial cell(s); LFA-1, lymphocyte-function-associated antigen 1; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis; bp, base pair(s); PHA, phytohemagglutinin; PMA, phorbol 12-myristate 13-acetate; HEPES, 4-(2-hydroxyethy1)-1-piperazineethanesulfonicacid.

5243

5244

CD31 Phosphorylation with

confirm the presence of CD31 in a variety of leukemia cell lines and show that CD31 is present on human peripheral blood T cells.We now report two new findings. CD31 is rapidly phosphorylated following cell activation in platelets, EC,and Jurkat cells; and the expression of CD31 is downregulated following T cell activation. We have also assessed the molecular mechanisms responsible for CD31 down-regulation in activated T cells.

Cell Activation

immobilized by UV cross-linking (Stratalinker, Stratagene)and then hybridized with a CD31 probe labeled by the random hexamer method (37). The membrane was washed first at room temperature with 2 X SSC and 0.1% SDS and then under high stringency conditions (0.1 X SSC and 0.1% SDS at 65 "C for 30 min). Following autoradiography, the CD31 probe was stripped from the membrane and then rehybridized with a random hexamer-labeled p-actin probe (a gift of Dr. Linda Boxer (StanfordUniversity)) and washed as described above. Human genomic Southern blots were a gift of Dr. Sabine Kohler (Stanford University). Purified human genomicDNAwas digested with EcoRI, HindIII, BglII, and BarnHI (Promega Biotec); EXPERIMENTALPROCEDURES the fragments were separated on a 0.8% agarose gel, transferred to a Cells-HEL, K562, C32, and Jurkat cells were obtained from the nylon membrane (Genatran) (20), and fixed by UV cross-linking. The American Type Culture Collection and grown in RPMI 1640 medium, blots were probed with a random hexamer-labeled full-length CD31 10% fetal calf serum, 2 mM glutamine. HUVEC were isolated from cDNA probe and washed under high stringency conditions as deumbilical cords (17) and grown in Medium 199 with 10% fetal calf scribed above. serum, 2 mM glutamine, and 0.1% endothelial mitogen (Biomedical RNA Dot Blot Analysis-Total RNA (10 pg) was denatured in a Technologies, Inc.). Confluent cells between passages 2 and 4 were formaldehyde sample buffer by heating at 65 "C for10 min. The used for studies. Human peripheral blood T cells were prepared by samples were then applied to a nitrocellulose membrane using a first isolating peripheral blood mononuclear cells by Ficoll-Hypaque vacuum dot blot apparatus (Bio-Rad). The RNA was fixed, hybriddensity gradient centrifugation and then using double plastic adher- ized, and washed as described above for Northern analysis. ence to remove monocytes, followed by application to a nylon wool Nuclear Run-on Assay-Lymphocyte nuclei were isolated after 3 h column (200L, Robbins Scientific) and 30-40% discontinuous Percoll of stimulation by resuspension in hypotonic buffer (10 mM Tris, 10 gradient centrifugation (18).Purity was >95% as measured by flow mM NaCl, 5 mM MgClz, pH 7.4), followedby lysis with 0.5% Nonidet cytometry with Leul and Leu4 monoclonal antibodies. Cells were P-40. The nuclei were harvested by centrifugation, washed twice in maintained in RPMI 1640 medium with 10% fetal calf serum and 2 hypotonic buffer, suspended in nuclear storage buffer (50 mM Tris, 5 mM glutamine. For studies of CD31 after cell activation, cells were mM MgCl,,0.1 mM EDTA, 40% glycerol, pH 8.3), and stored at treated with purified phytohemagglutinin (PHA-P, Wellcome) at 10 -70 "C for 1-7 days. Nuclei (40 X lo6) were thawed and incubated pg/106 cells.Cell concentration was 1 X lo6 cells/ml, except for for 30 min a t 30 "C in reaction buffer (25 mM Tris, pH8.0) containing phosphoric acid labeling experiments, where the concentration was 275 pCi of [a-"PIUTP. RNA was isolated and hybridized for 3 days 3 X lo6 cells/ml. to Nytran filters containing 5 pg of various dot-blotted immobilized Monoclonal Antibodies-LYP21 was obtained by immunizing mice nucleic acids (2300-bp CD31 cDNA, 870-bp NcoI granulocyte-macwith Lubrol-PX-solubilized platelet glycoproteins isolated by wheat rophage colony-stimulating factor coding strand, 404-bp EcoRI 8germ agglutinin-Sepharose 4B affinity chromatography and Mono Q actin coding strand, 1500-bp pBR322 plasmid DNA, yeast tRNA). anion-exchange chromatography.' Pooled culture supernatants were After hybridization, the blots were washed to a final stringency of 0.1 used for studies with LYP21. CD31 monoclonal antibodies 7E8 and X SSPE at 65 "C for 30 min, 0.1% SDS and then washed for 15 min 8B6 were gifts of Drs. Karel Nieuwenhuis, Marcel Metzelaar, and Jan at 23 "C with RNase (1 pg/ml) in 2 X SSC and rinsed in 1 X SSPE, Sixma (University Hospital, Utrecht, The Netherlands). Rabbit pol- 0.1% SDS. yclonal anti-phosphotyrosine antiserum was a gift of Dr. James Farrel Zrnrnunoblots-Cells were lysed in Triton X-100 lysis buffer (10 (University of Wisconsin). Monoclonal anti-phosphotyrosine anti- mM Tris-HC1, pH 7.4,0.15 M NaC1, and 1%(v/v) Triton X-100) bodies PY20 and PY69were purchased from ICN Biochemicals. containing protease inhibitors (1mM phenylmethylsulfonyl fluoride, OKT3 was purchased from Ortho Pharmaceutical. Control antibodies 0.01 mM aprotinin, 1mM EDTA) a t 4 "C for 30 min. The lysates were used were 1F5 (anti-CD20, provided by Dr. Robert Negrin (Stanford centrifuged at 12,000 X g at 4 "C for 15 min, and the supernatant University)) and Leul (CD5, purchased from Becton-Dickinson). containing solubilizedproteins was stored at -70 "C until use. Protein Cloning of CD31 cDNA-Plaques (2 X lo6) from oligo(dT)-primed concentrations were measured using a Coomassie Blue assay (Pierce HUVEC Xgtll cDNA libraries (obtained from Dr. J. Evan Sadler Chemical Co.). Cell lysate (100-200 pg) was applied to each lane of (Washington University) and Dr. David Ginsberg (University of 7.5% SDS-polyacrylamide gels. After SDS-PAGE, the proteins were Michigan)) were screened with monoclonal antibody LYP21. Fifteen electrophoretically transferred to a polyvinylidone membrane (Imclones were identified at tertiary screening. These phage were isolated mobilon, Millipore). After blocking the nonspecific binding sites by and amplified, and phage inserts were obtained by EcoRI digestion, soaking in Tris-buffered saline containing 0.05% Tween 20 and 1% subcloned into the M13mp18 vector, and sequenced bidirectionally bovine serum albumin for 2 h, the membranes were incubated with by the dideoxy chain termination method (36) using Sequenase DNA LYP21 supernatants at1:500 dilution and with control CD20 monopolymerase (Pharmacia LKBBiotechnology Inc.). Unfortunately, the clonal antibody 1F5 at 5 pg/ml for 1 h. Following extensive washing, cDNA had two EcoRI sitesin the open reading frame, yielding the blot was developed by sequential addition of a goat anti-mouse incomplete clones after EcoRI digestion. To obtain the 5'-end of the antibody conjugated to alkaline phosphatase (1:5000 dilution, Cappel) cDNA, a sense primer with a 5'-SpeI restriction site was synthesized and nitro blue tetrazolium/5-bromo-4-chloro-3-indoyl phosphate that annealed just upstream of the X EcoRI site (5"GGAC(Promega) substrates. TAGTGGATTGGTGGCGACGACTC-3'). Using this in combinaCell-surface Zodination-Confluent EC monolayers were detached tion with downstream antisense primers 3' to the EcoRI site and by collagenase treatment andwashed with phosphate-buffered saline. recombinant phage as template, the 5'-region of the CD31 cDNA was Jurkat cells were pelleted and washed with phosphate-buffered saline. amplified by the polymerase chain reaction, subcloned into M13, and Cells were surface-labeled with lz5I by the lactoperoxidase method sequenced. Three phage isolates contained the entire open reading (Enzymobeads, Bio-Rad) according to the manufacturer's directions frame. These were amplified and subcloned into a phagemid and lysed with Triton X-100 lysis buffer containing protease inhibi(pBluescript SK-, Stratagene). All polymerase chain reaction se- tors as described above. quence data were verified by duplicate amplifications or X insert [3'P]PhosphoricAcid Labeling-Jurkat and human peripheral sequence. blood T cells were washed, resuspended in phosphate-free RPMI 1640 Northern and Southern Blots-Total cellular RNA wasisolated by medium with 5% dialyzed fetal calf serum, and incubated at 37 "C for the guanidinium thiocyanate method (19). The RNA was denatured 18 h. The cells were resuspended into two 10-ml aliquots at a concenby heating in loading buffer (20 mM borate, pH 8.3, 0.2 mM EDTA, tration of 3 X lo6 cells/ml in phosphate-free RPMI 1640 medium 50% formamide, 10% glycerol, and 6% formaldehyde, used at a 4:l with 10% dialyzed fetal calf serum, and 100 pCi/ml ~rtho[~'P]phosratio with RNA samples) for 10 min at 65 "C. The samples (5-10 pg phoric acid (Du Pont-New England Nuclear) was added. Following of total RNA/lane) were separated on a 0.8% agarose gel in 20 mM incubation for 2 h at 37 "C, one aliquot was stimulated with PHA (10 borate, pH 8.3, 0.2 mM EDTA, and 3% formaldehyde; the gel was pg/106 cells) for 5 min and then lysed in Triton X-100 lysis buffer rinsed in water to remove formaldehyde and then transferred to a with protease inhibitors as described above with the addition of 50 nylon membrane (Genatran, Plasco) with 10 X SSC by capillary mM NaF and 2 mM vanadate to inhibit phosphatase activity. Resting action for 16 h. The membrane was rinsed in 4 X SSC; the RNA was Jurkat and peripheral blood T cells were lysed in the same buffer. For [3'P]phosphoric acid labeling of platelets, washed platelets were ' S. Parmentier and J. L. McGregor, manuscript in preparation. resuspended at a concentration of2.5 X 108/ml in phosphate-free

CD31 Phosphorylation with

Cell Activation

5245

buffer (12 mM NaHC03, 10 mM HEPES, 138 mM NaCI, 5.5 mM nylation signal was found after sequencing 1.1 kilobases of glucose, 2.9 mM KCl, pH 7.4) with 100 pCi/ml ~rtho[~*P]phosphoric the 3"untranslated region. acid and incubated at 37 "C for 30 min. Activated platelets were Northernand Southern Analyses of CD31-Northern prepared by stimulating platelets in the above phosphate-free buffer containing 1 mM CaCI2 with 3 units/ml human thrombin (a gift of analysis (Fig. 2) revealed three major transcripts in EC and cells of approximately 3.7,3.4, and 3.0 Dr. Marc Shuman (University of California, San Francisco)) or 1 X HEL and Jurkat 10' M histamine (Sigma) for 3 min at 37 "C. For experiments with kilobases, respectively. K562 cells were weakly positive. Of the proteinkinaseinhibitorstaurosporin(Sigma),platelets were interest, EC hadan additional 5.3-kilobase band not found in preincubated for 10 min with 3 p~ staurosporin, followed by stimu- theother cell types. Human genomic Southern blotting lation with thrombin or histamine as described above. For experishowed a simple pattern consistent with a single copy gene ments with EC, confluent monolayers were incubated for 18 h in phosphate-free RPMI 1640 medium and then incubatedwith 100 (data not shown). CD31 Expression in Resting and Activated T Cells-Because pCi/ml ortho["P]phosphate for 1 h at 37 "C. Cells were stimulated with histamine (1 X lo5 M ) or human thrombin (3 units/ml) for 5 of our finding of CD31 expression in the Jurkat leukemia T min at 37 "C or with recombinant human tumor necrosis factor (10 cell line, we examined human peripheral blood T cells for nM, Genentech) for 4 h at 37 "C. Resting and activated cells were CD31 expression by dot blots and Northern hybridization and lysed in theTriton X-100 (lysis buffer)containingprotease (Fig. 3). CD31 expression was consistently present in resting phosphatase inhibitors as described above. T cells with a Northern hybridization pattern similar human Zmmunoprecipitation-'2sI-Radiolabeled and [32P]phosphoricacidlabeled cells were lysed as described above. Cell lysates were pre- to other cell types. We next examined the effect of T cell cleared with protein A- or protein G-Sepharose (Pharmacia LKB activation upon CD31 message expression. PHA stimulation Biotechnology Inc.) and incubated with 1.25 pg of LPY2l or control resulted in a significant decrease in CD31 message expression monoclonal antibody 1F5 for 1 h at room temperature. Protein A- or at 16 h compared with resting cells, whereas actin expression protein G-Sepharose was then added; and following extensive washing, the bound complexes were eluted by boiling in SDS sample remained unchanged (Fig. 3). The decrease in CD31RNA expression was accompanied by a decrease in CD31 protein buffer, followed by SDS-PAGE and autoradiography. RESULTS

Identification of LYP21 Antigen as CD31-The LYP2l monoclonal antibody immunoprecipitated and immunoblotted and 130-kDa protein from platelets and a 130-kDa major protein as well as a 110-kDa minor protein from EC (Fig. 1). A similar lower molecular mass band in EC has been previously observed with the Hec7 CD31 monoclonal antibody (21). LYP21 was used to screen a HUVEC X g t l l library. Fifteen overlapping clones were identified and sequenced, revealing a 207-bp 5"untranslated region, an open reading frame of 2214 bp, and a 3"untranslated region of 1172 bp. Three clones encoded the entire open reading frame, which upon translation predicted a 738-amino acid protein. Comparison with the National Biomedical Research Foundation Data Base revealed a 21%homology to carcinoembryonic antigen at theamino acid level and established the molecule as being a member of the immunoglobulin gene superfamily. At the time this workwas completed, Newman et al. (9) published the cloning of CD31, which they termed plateletendothelial cell adhesion molecule 1. Comparison of our sequence with the platelet-endothelial cell adhesion molecule 1 sequence revealed that they were identical except for four bases, which resulted in four different amino acids: A425T, resulting in asparagine 115 becoming isoleucine; A532G, resulting in asparagine 151 becoming aspartic acid; A1206G, resulting inisoleucine 375 becoming methionine; and A1253T, resulting in aspartic acid 391 becoming valine. No polyade-

expression in PHA-treated cells compared to resting cells at 16 and 65 h as shown by immunoblotting with LYP2l (Fig. 4). CD31 was present in the nonionic detergent-soluble fraction in both resting and activated T cells and was not detect-

kb 4.4

CD3I

-

2.37-

2.37 -

1.35

ACTIN 1.35

-

FIG. 2. Northern hybridization of total RNA using CD31 and &actin probes. Total RNA was isolated from EC and K562, HEL, Jurkat, and C32 cells. Each RNA (10 pg) was separated on a 0.8% agarose gel and transferred to a nylon membrane. Random hexamer-labeled CD31 and @-actin probes were used to probe the same membrane. kb, kilobases.

A

% '

Go

d VG kb

4.4

A

P EC E7 C

P21

-

c

kDa 110 " "

1.35 -

80 -

2.37-

33 50

e

ACTIN

-

FIG. 1. A, immunoblot of platelets (P),EC, and fusion peptide ( E 7 ) with LYP21. The E7 clone was identified in the screening of the HUVEC library with LYP2l and plaque-purified. Cell lysates were immunoblotted as described under "Experimental Procedures." CD20 Monoclonal antibody 1F5 was used as control (C). B, immunoprecipitation of surface-iodinated EC lysates with LYP2l (P21) and control (C) antibodies.

1.35

m

u

-

FIG. 3. CD31 RNA expression in resting and activated T

cells. A, dot blot; B, Northern blot. Human peripheral blood T cells were isolated, and totalRNA was prepared. RNA (10 pg) was applied to a nitrocellulose membrane ( A ) or separated on an agarose gel and transferred to a nylon membrane ( B ) .Labeled CD31 or 0-actin probes were used to probe the same membrane. RESTING, resting T cells after 16 h of incubation; PHA, T cells stimulated with PHA for 16 h; C32, C32 melanoma cell control. kb, kilobases.

CD31 Phosphorylation with Cell Activation

5246

CONTROL

P21 I TO

T16 T65'16 R PHA

-

R

Resting Activated

1

T65T65 '16 R PHA R

T65 '0 PHA

CD3I

PHA

p actin tRNA

-

PBR

COOMASSIE BLUE

TO T16 7 6 R PHA w w w

GM-CSF

T65 TsS R

PHA

FIG. 5. Nuclear run-on studies of CD31 transcription in T cells. Membranes were prepared with CD31, 0-actin, yeast-tRNA, pBR322, and granulocyte-macrophage colony-stimulating factor (GM-CSF) target DNA as described under"Experimental Procedures.'' The membranes on the left were hybridized to nuclear RNA from resting T cells, and thoseon the right were hybridized to nuclear RNA from PHA-activated T cells. A

FIG.4. Immunoblot of resting and activated T cells with CD31 monoclonal antibody LYP21. Human peripheral blood T cells were incubated in medium alone ( R ) or stimulated with PHA (10 pg/106 cells) for 16 or 65 h. At the endof the specified incubation times, cells were lysed, and SDS-PAGE and immunoblotting were performed. CD20 monoclonal antibody 1F5 was used as control. To refers to resting T cells prior to stimulation. Another gel loaded in a n identical manner was stained with Coomassie Blue to show that the amount of protein loaded per lane was comparable.

CD31

CONTROL

kDa 205

-

116 -

80

-

50 -

-

able in the insoluble cytoskeletal fraction in eithercase (data B CD31 not shown). Thus, the decrease in protein levels after cell kDa 0 3 5 10 20 30 activation is not due tocytoskeletal association and removal 205 fromtheTriton-solublefraction,butis aconsequence of 116decreased CD31 message expression. 80 50 Mechanism of CD31 Down-regulation after T Cell Activation-Decreased levels of CD31 mRNA following T cell activation could result from a decreased rate of transcription or FIG. 6. Phosphorylation of CD31 in platelets with thrombin diminished stabilizationof CD31 transcripts,which appear to be constitutively expressed in T cells. To assess possible post- activation. A, washed platelets were incubated in phosphate-free bufferwith ~rtho[~*P]phosphate and stimulated with thrombin (3 transcriptional regulation of CD31 transcripts, unstimulated units/ml) for 3 min. The cells were then lysed and immunoprecipiand PHA-activatedT cells were cultured withactinomycin D tated either with a mixture of CD31 monoclonal antibodies (LYP21, (5 pglml)for 5-120 min. This amount of actinomycinD 8B6, and 7E8) or with control monoclonal antibody 1F5. B, time suppressed >95% of transcription in T cells as assessed by course of 32Pincorporation into CD31 in thrombin-stimulated plate[14C]uridineincorporation. RNAwas harvested and examined lets (in minutes). R, resting platelets; Thr, platelets stimulated with for CD31 mRNA levels by Northern analysis. The tllPof CD31 thrombin 3 units/ml for 3 min. mRNA in resting T cellswas30 min and wasminimally reduced in PHA-activated T cells (20 min) (data notshown). bin stimulation, with a time course lasting for at least 30 min PHA stimulation had no effect on the tIl2 of T cell 0-actin (Fig. 6). In cultured confluent HUVEC, there was base-line mRNA after >8 h of actinomycin D exposure. phosphorylation of CD31 in the resting state.Upon stimulaTranscriptional regulation of T cell CD31 expression was tion with either thrombinof histamine, significant enhancenext assessed using an invitro nuclear run-on assay. CD31 is ment of CD31phosphorylation was observed. In contrast, constitutively transcribed in resting T cells, whereas PHA- stimulation of EC with tumor necrosis factor did not result stimulated T cells have a 50% reduction in CD31 transcript in appreciably increased phosphorylation (Fig. 7). Similarly, (Fig. 5). Control hybridizations showed no significant differ- 0 3 1 was not phosphorylated in the resting Jurkatcells, but ence in 0-actin transcripts in resting and activated T cells, became rapidly phosphorylated after stimulationwith phytowhereasgranulocyte-macrophage colony-stimulatingfactor hemagglutinin (Fig. 8). transcription increased -2-fold in activated T cells. Nuclear Further Characterization of CD31 Phosphorylation-BeRNAs from both resting and activated T cells failed to hy- cause of the homology between the CD31 cytoplasmic domain bridize to either pBR322 plasmid DNA or yeast tRNA (Fig. and the platelet-derived growth factor receptor (9), which 5). Mean incorporations of [32P]dUTP into resting or acti- autophosphorylates a cytoplasmictyrosine, the possibility vated T cells were nearly equal. that CD31 phosphorylation involved tyrosine phosphorylaPhosphorylation of CD31 after Cell Activation-CD31 was tion was raised. However, there are three lines of evidence not involve rapidly phosphorylated after cell activation in three separate that suggest that CD31phosphorylationdoes cell types. In platelets, CD31 was not phosphorylated in the tyrosine phosphorylation. First, we examined restingand resting state, butbecame rapidly phosphorylated after throm- activated Jurkatcells for evidence of tyrosine phosphorylation

---

-

CD31 Phosphorylation with Cell Actiuation A kDa

CD31

'

116 EO -

205

R

'

Thr HIS2

-

CONTROL R

-

Thr His2

-.%

A

5247

+

B

PMA

+

-

PMA

52 b

D

n o

"

-

I -

50 -

R

kDa

I I C

"

TNF

CD31

n n CD31

c

CD31

C

-

I

u

c

CD31

205 116 80 50 -

-.

I

FIG. 7. Phosphorylation of CD31 in resting and activated cultured endothelial cells. A, cultured HUVEC were labeled with '"P and either were notstimulated ( R ) or were stimulated with thrombin (Thr) or histamine ( H i s t ) for 5 min. The cells were then lysed and immunoprecipitated with CD31 or control monoclonal antibodies as described for Fig. 6. B, HUVEC were treated with tumor necrosis factor ( T N F )for 4 h and then processed as described for A .

FIG. 9. Phosphorylation of CD31 in platelets treated with PMA in presence or absence of staurosporin. A, immunoprecipitation of "P-labeled washed platelets treated with PMA with CD31 or control ( C )monoclonal antibodies; B, immunoprecipitation of ">Plabeled CD31 from platelets treated with PMA in the presence (+) or absence (-) of staurosporin ( S ) . PHA

"0

l4

PHA + S

' TO

l4 ' 8 '

Jurkat Restlng

I

Actlvated

I I

1

FIG. 10. Immunoblot of CD31 in T cells treated with PMA in presence or absence of staurosporin. Human peripheral blood

T cells were activated by PHA (10 qg/106 cells) in the presence or absence of staurosporin ( S ) (3 PM). At the end of the specified incubation times, cells were lysed, equal amounts of cell lysates were separated by SDS-PAGE, and immunoblotting was performed using LYP21. Torefers to resting T cells prior to activation. FIG.8. Phosphorylation of CD31 in resting and activated Jurkat cells. Jurkat cells were labeled with '*P and stimulated with

the data strongly suggest that CD31 phosphorylation is mediated by protein kinase C and involves serine and/or threoPHA (10 pg/106 cells) for 5 min and then processed for immunoprecipitation with CD31 monoclonal antibodies,control monoclonal nine residues. Effect of Phosphorylation on Cell-surface Expression of CD31 antibody 1F5, or monoclonal antibody anti-phosphotyrosine PY20. after Cell Activation-To explore thepotentialfunctional The phosphorylated CD31 (*) did not co-migrate with a tyrosinephosphorylated protein (arrowhead) on SDS-PAGE. consequences of CD31 phosphorylation, the cell-surface expression of CD31 on T cells after cell activation with PHA was examinedin thepresence or absence of staurosporin (Fig. using two monoclonal anti-phosphotyrosine antibodies. Several phosphotyrosine-containing proteins were identified by 10). Decreased cell-surface expression of CD31 was noticed after 4 h of T cell activation. However, pretreatment of cells immunoprecipitation after Jurkat activation with PHA, in withstaurosporinpreventedtherapid down-regulation of agreementwith published data (22). However, thephosCD31, suggesting that protein kinase C-mediated phosphorylphorylated CD31 didnot co-migrate with themajor phosphoation results in increased turnover of CD31 protein. Whether tyrosinated proteins on SDS-PAGE (Fig. 8). A similar negathis is due to phosphorylation of CD31 itself and/or a protein tive result was obtained when the LYP2l immunoprecipitates from activated platelets and Jurkat cells were immunoblotted kinase C effect on another protein that altersCD31 protein using a monospecific polyclonal anti-phosphotyrosine anti- stability remains tobe determined. body (data not shown).Second, CD31phosphorylationin DISCUSSION platelets was rapidly induced by treatment with phorbol 12We have independently cloned the cDNA for CD31, a newly myristate 13-acetate (PMA),which is known to activate protein kinaseCin platelets (23) (Fig. 9A). Enhanced CD31 described surfacemembrane glycoprotein of the immunoglobphosphorylation in PMA-stimulated EC was also observed ulin gene superfamily, using a monoclonal antibody raised (data not shown). The involvement of protein kinase C was against activated platelets. Our data show minor differences further substantiated by the use of the relatively specific from the published sequence (9-11) and may represent polyprotein kinaseC inhibitorstaurosporin (24). Staurosporin morphisms or reverse transcription errors duringlibrary con, has been struction. We show that CD31 is present on Jurkatleukemia was used a t a very low concentration (3p ~ )which shown to maximally inhibitthethrombin-inducedphosT cells and human peripheral blood T cells in addition to phorylation of P47, a known substrate of protein kinase C in other myeloid cells, platelets, and EC. There are conflicting platelets, while not affecting the phosphorylation of many reports on theexpression of CD31 in T cell lines (26). It is of other proteins, indicating its relative specificity (25). Stauros- note thatmost of the cell lines reported not toexpress CD31 porin completely inhibitedthe PMA-inducedCD31 phos- were thymomas or thymic lymphomas representing an early phorylation in activated platelets (Fig. 9B). Taken together, stage of thymic differentiation,whereas Jurkat cells are more

5248

CD31 Phosphorylation with Cell Activation

differentiated. In preliminary studies, we have been unable to detect CD31 in human thymocytes, suggesting that acquisition of CD31may be alatestep in T cell development. Southern analysis of human genomic DNAwith CD31 probes revealed a restriction pattern consistent with CD31 being a single copy gene rather than a member of a large family of related genes such as carcinoembryonic antigen (27). There are three major RNA transcripts in most cell types, but an additional 5.3-kilobase transcript in cultured HUVEC was observed (Fig. 2). The significance of this is unknown, but it has been recently reported that alternate forms of vascular cell adhesion molecule 1, another member of the immunoglobulin superfamily, can be generated by differential splicing in HUVEC (28). CD31 is not phosphorylated in the basal resting state in platelets andJurkat cells, and theconstitutive phosphorylated state in HUVEC may be related to the culture condition. In all three cell types, CD31 is rapidly phosphorylated after cell activation (Figs. 6-8). Although the cytoplasmic domain of CD31 is homologous to the platelet-derived growth factor receptor, which has tyrosine kinase activity and autophosphorylates a cytoplasmic tyrosine, we were unable to demonstrate the presence of a phosphotyrosine using several monoclonal and polyclonal anti-phosphotyrosine antibodies (Fig. 8). In addition, the induction of CD31 phosphorylation by PMA and the inhibition by staurosporin (Fig. 9) strongly suggest that phosphorylation is mediated by protein kinase C and involves serine and/or threonine residues. Rigorous proof of this awaits analysis of CD31 phosphoamino acids. Recent studies suggest that CD31 functions as a cell adhesion molecule (16). Our data suggest that phosphorylation of CD31 may be functionally important. The CD31 protein halflife is decreased after PHA activation of T cells. This effect is abrogated in the presence of the protein kinase C inhibitor staurosporin (Fig. lo), suggesting that protein kinase C activation and subsequent CD31 phosphorylation result in faster CD31 turnover. Phosphorylation may also cause a conformational change in CD31 that modulates its cellular adhesiveness, ashas been reported for several other adhesion molecules. Platelet activation confers an adhesive phenotype that is closely associated with phosphorylation and activationdependent conformational changes in surface adhesion molecules such as glycoproteins IIIa and Ib, both of which are serine/threonine-phosphorylated(24, 29). Similarly, the affinity of LFA-1 for its ligand, intercellular adhesion molecule 1,becomes markedly increased in a transient mannerfollowing T cell activation and is postulated to be a mechanism whereby T cell antigen receptor activation regulates lymphocyte adhesion and de-adhesion (30). It has recently been reported that LFA-1 is phosphorylated by protein kinase C (31,32). It has also been demonstrated that transientadhesion is markedly augmented in the presence of PMA. These observations suggest that protein kinase C-mediated phosphorylation is important in LFA-1 adhesion (29,33). Our data demonstratethat CD31 is down-regulated after T cell activation at the mRNA and protein levels (Figs. 2 and 3). The decreased expression is mediated by decreased CD31 transcription as well as a decrease in the CD31 protein halflife (Figs. 5 and 10).In contrast, Ohto et al. (14) reported that CD31 expression by indirect immunofluorescence increased after activation of peripheral blood mononuclear cells by phytohemagglutinin in one sample tested. The reason for these disparate findings is unclear. Of note, down-regulation of CD31 surface expression after granulocyte activation by met-Leu-Phe has recently been demonstrated (11). Several adhesion molecules exhibit an activation-depend-

ent increase or decrease in cell-surface expression. CD28 expression is increased following lymphocyte activation (34), and the C D l l . CD18 (LFA-1-Mac 1)complexes are up-regulated following neutrophil activation. On the other hand, murine GP100MEL.14, a selectin family adhesion molecule present on neutrophils, is down-regulated following neutrophil activation with PMA (35). It is interesting that the rapid down-regulation of GP100MEL"4 is mediated by shedding into the extracellular milieu; in contrast, CD31 appears to be down-regulated through decreased mRNA transcription. Since CD31 is known to be localized to areas of cell-cell contact on confluent EC, CD31 may play a role in lymphocyte and neutrophil trafficking across endothelial barriers. Downregulation of CD31 and other adhesion molecules following cell activation may occur as a consequence of differential expression of cell adhesion molecules that serve to impart target specificity and to localize effector leukocytes to areas of inflammation. Acknowledgments-We thank Dr. Sophie Parmentier (Institut National de la Santk et de la Recherche Medical&U331, Lyon, France) for her efforts in producing the LYP2l monoclonal antibody; Drs. Karel Nieuwenhuis and Marcel Metzelaar for providing monoclonal antibodies 7E8 and 8B6; Dr. J. Evan Sadler and Dr. David Ginsberg for endothelial cellcDNA libraries; Deborah Grantand Diana Thompson for technical assistance; and Drs. Sabine Kohler, Naomi Galili, and JayneDanska for helpful discussions. REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22.

Albelda, S. M., and Buck, C. A. (1990) FASEB J. 4, 2868-2880 Springer, T. A. (1990) Nature 346,425-434 Ruoslahti, E. (1991) J. Clin. Invest. 87, 1-5 Williams, A. F., and Barclay, A.N. (1988) Annu. Reu. Immunol. 6,381-405 Hunkerpiller, T.,and Hood, L. (1989) Ada Imrnunol. 4 4 , 1-63 Maddon, P. J., Dalgeish, A. G., McDougal, J. S., Clapham, P. R., Weiss, R. A., and Axel, R. (1986) Cell 47,333-348 Greve, J. M., Davis, G., Meyer, A. M., Forte, C. P., Yost, S. C., Marlor, C. W., Kamarck, M. E., and McClelland, A. (1989) Cell 66,839-847 Staunton, D. E., Merluzzi, V. J., Rothlein, R., Barton, R., Marlin, S. D., and Springer, T.(1989) Cell 66,849-853 Newman, P. J., Berndt, M. C., Gorski, J., White, G. C., 11, Lyman, S., Paddock, C., and Muller, W.A. (1990) Science 2 4 7 , 12191222 Simmons, D. L., Walker, C., Power, C., and Pigott, R. (1990) J. Exp. Med. 171,2147-2152 Stockinger, H., Gadd, S. J., Eher, R., Majdic O., Schreiber, W., Kasinrerk, W., Strass, B., Schnabl, E., and Knapp, W. (1990) J. Immunol. 146,3889-3897 Goyert, S. M., Ferrero, E. M., Seremetis, S. V., Winchester, R. J., Silver, J., and Mattison, A.C. (1986) J. Immunol. 137, 3909-3914 Albelda, S. M., Oliver, P. D., Romer, L. H., and Buck, C.A. (1990) J. Cell Bwl. 110,1227-1237 Ohto, H., Maeda, H., Shibata, Y., Chen, R. F., Ozaki, Y., Higashihara, M., Takeuchi, A., and Tohyama, H. (1985) Blood 6 6 , 873-881 van Mourik, J. A., Leeksma, 0. C., Reinders, J. H., de Groot, P. G., and Zandbergen-Spaargaren, J . (1985) J. Biol. Chem. 260, 11300-11306 Albelda, S. M., Muller, W. A., Buck, C.A., and Newan, P. J . (1991) J. Cell Biol. 114, 1059-1068 Jaffe, E. A., Nachman, R.L., Becker, C.G., and Minick, C. R. (1973) J. Clin. Invest. 62, 2745-2756 Levitt, L. J., Kipps, T. J., Engleman, E. G., and Greenberg, P. L. (1985) Blood 66,663-679 Chomczynski, P., and Sacchi, N. (1987) Anal. Biochem. 1 6 2 , 156-159 Southern, E. M. (1975) J. Mol. Biol. 98,503-517 Muller, W. A,, Ratti, C. M., McDonnell, S. L., and Cohn, Z. A. (1989) J.Exp. Med. 170,399-414 June, C. H., Fletcher, M. C., Ledbetter, J. A., and Samelson, L. E. (1990) J. Immunol. 1 4 4 , 1591-1599

CD31 P h o ~ p ~ ~ ~with t i Cell o nA ~ t i v ~ t i o n 23. Sano, K., Takai, Y., Yamanishi, J., and Nishizuka, Y. (1983) J. Biol. Chem. 258,2010-2013 24. Tamaoki, T., Nomoto, H., Takahashi, I., Kato, Y., Morimoto, M., and Tomita, F. (1986) Biochem. Biophys. Res. Commun. 1 3 5 , 397-402 25. Parise, L. V., Criss, A. B., Nannizzi, L., and Wardell, M. R. (1990) Blood 76,2363-2368 26. Lyons, A. B., Cooper, S. J., Cole, S. R., and Ashman, L. K. (1988) Pathology 20.137-146 27. Zimmerman, W., Weber, B., Ortlieb, B., Rudert, F., Schempp, W., Fiebig, H. H., Shively, J. E., von Kliest S., and Thompson, J. A. (1988) Cancer Res. 48,2443-2550 28. Hession, C., Tizard, R., Vassallo, C., Schiffer, S. B.,Goff,D., Moy, P., Chi-Rosso, G., Luhowskyi, S., Lobb, R., and Osborn, L. (1991) J. Bioi. Chem. 266,6682-6685 29. Wardell, M. R., Reynolds, C. C., Berndt, M. C., Wallace, R. W.,

5249

and Fox, J. E. B. (1989) J. Biol. Chem. 2 6 4 , 15656-15661 30. Dustin, M. L., and Springer, T. (1989) Nature 341,619-624 31. Hara, T., and Fu, S. M. (1986)in Leucocyte Typing I1 (Reinhez, E. L., ed) Vol. 3, pp. 77-84, Springer-Verlag New York Inc., New York 32. Chatifa, T. A,, Geha, S., and Arnaout, M. A. (1989) J. Cell Biol. 109,3435-3444 33. Hibbs. M. L.. Xu.H.. Stacker. S. A.. and Smineer. T. (1991) . Sciince 25i,16il-i613 34. Linslev. P. S.. Clark. E. A.. and Ledbetter. J. A. (1990) . . Proc. Nati. Acad. Sci. U. S. A. 87, 5031-5035 35. Kishimoto, T. K., Jutila, M. A., Berg, E. L., and Butcher, E. C. (1989) Science 2 4 5 , 1238-1241 36. Sanger, F., Nicklen, S., and Coulson, A. R. (1977) Proc. Natl. A d . Sei. U. S. A. 7 4 , 5463-5467 37. Feinberg, A. P., and Vogelstein, B. (1983) Anal. Biochem. 1 3 2 , 6-13

-

'

I

.

I