THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1992 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 267, No. 31, Issue of November 5, pp. 22190-22197,1992 Printed in U, S. A.

Interleukin- 1-inducible Genes in Endothelial Cells

(Received for publication, June 4, 1992)

Ferruccio BreviarioS, Elisabetta M. d’AnielloS, Josee Golay$$, Giuseppe PeriS, Barbara BottazziS, Amos Bairochll, Salvatore Sacconell**, Rosalia MarzellaSS, Valentina PredazziSS, Mariano Rocchi#$, Giuliano Della Vallell, Elisabetta DejanaS, Albert0 Mantovanil, and Martino Introns#@ From the #Istituto Ricerche Farmacologiche “Mario Negri,” via Eritrea 62, 20157 Milano, Italy, the llh4edical Biochemistry Department, Centre Medical Universitaire, Geneva, Switzerland, the [(Dipartimento di Genetica e Microbiologia “A. Buzzati- Traverso,” Universita di Pavia, via Abbiategrasso 207, 27100 Pavia, Italy, and the $$Istituto di Genetica, Via Amendola 165/A, 70126 Bari, Italy

Differential screening of a cDNA library constructed from human umbilical vein endothelial cells exposed for 1 h to interleukin-1s (IL-1s) has led to the identi- fication of a novel gene (PTX3) related to pentaxins (C-reactive protein and serum amyloid P component in man), a subclass of acute phase proteins. Sequencing of the full-length cDNA clone and RNase mapping revealed that the PTX3 transcript is 1861 base pairs long and has a unique transcription start site. The predicted protein sequence of 381 amino acids is highly similar to pentaxins in its COOH-terminal half where it also contains a typical 8-amino acid “pentaxin sig- nature’’ sequence. The NH2-terminal half of PTX3 shows no similarity to any known protein sequence and initiates with a putative signal peptide indicating that PTX3 is secreted. The genome of PTX3 is orga- nized into three exons. Interestingly, the region of homology between PTX3 and pentaxins corresponds to the third PTX3 exon. The PTX3 gene has been local- ized on human chromosome 3 band q25 by Southern blots of somatic cell hybrids and by in situ hybridiza- tion. The PTX3 mRNA is induced in endothelial, he- patic, and fibroblastic cells by IL-10 and tumor necro- sis factor a but not by IL-6 and interferon-y. PTX3 may represent a novel marker of inflammatory reac- tions, particularly those involving the vessel wall.

C-reactive protein (CRP)’ and serum amyloid P component (SAP) are plasma proteins that show a high degree of sequence similarity and belong to the pentaxin or pentraxin family,’ so

*This work was supported in part by Minister0 Sanith-Istituto Superiore Sanita (National AIDS Project) and Consiglio Nazionale delle Ricerche, Rome (“Progetto Finalizzato Ingegneria Genetica,” “Bioingegneria e Biostrumentazione,” and “Progetto A.C.R.O.”) and by generous contributions from the Italian Association for Cancer Research (AIRC), by the project “Telethon,” and the Istituto Gaslini, Genova (“Ricerca Finalizzata Gaslini Sud”). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

§ Supported by the Foundation Valenti, Milano, Italy. ** Supported by the Italian Ph.D. program.

To whom correspondence should be addressed. The abbreviations used are: CRP, C-reactive protein; SAP, serum

amyloid P component; HUVEC, human umbilical vein endothelial cells; TNF, tumor necrosis factor; bp, base pair(s); PIPES, 1,4- piperazinediethanesulfonic acid IL, interleukin.

Based on the Greek derivation (penta, five), we share the convic- tion of Pepys and Baltz (1) that the term pentaxin rather than pentraxin should be used.

named because of their pentameric structure (1). They belong to the vast group of proteins known as acute phase response proteins, although only CRP is elevated in the circulation during inflammation in man (Z), whereas SAP is the only acute phase reactant in the mouse (3,4).

CRP and SAP have been known for several years, and the former is widely used in the clinic to monitor different forms of infection, inflammation, or tissue damage, but their precise biological function is to a large extent still undefined (1). Interestingly, their protein sequence and genomic organiza- tion are well conserved from horseshoe crab to man (5-8), and all known pentaxins present a common structural motif of eight amino acids referred to as the “pentaxin family signature” (9).

In recent years, it has become clear that many events associated with the “acute phase response” are the conse- quence of the interaction of cytokines, in particular interleu- kin-6 (IL-6), interleukin-1 (IL-1), and tumor necrosis factor a (TNFa), with several target organs, most importantly with the liver (10-12). These mediators, IL-6 in particular, appear to directly regulate the production of CRP and SAP by acting on their 5’ regulatory sequences (13-16).

Vascular endothelial cells have emerged in recent years as playing an important and active role in inflammatory reac- tions, immunity, and thrombosis (17-19). This has raised considerable interest in the identification of genes that are expressed in inflammatory endothelium and that may serve as markers for vascular involvement in disease. In this per- spective, we have searched for new genes that are induced rapidly following exposure of human umbilical vein endothe- lial cells (HUVEC) to IL-16. A classical strategy of differential screening of cDNA libraries was chosen, which allowed the isolation of a novel pentaxin-related gene, which we called PTX3.

MATERIALS AND METHODS

Cell Sources and Culture Conditions-HUVEC were obtained and cultured as described previously (20) and used at the third to seventh passage. The hepatoma cell line HEPSB was cultured in Dulbecco’s modified minimal essential medium supplemented with 15% fetal calf serum (Gibco, Paisley, Scotland). The fibrosarcoma cell line 8387 was cultured in minimal essential medium and 10% fetal calf serum.

Stimulation with Cytokines-The cells were grown to confluence in 75-cm2 flasks, washed, and then incubated in 7 ml of endotoxin- free RPMI 1640 (Gibco) supplemented with 5% fetal calf serum, and 20 pg/ml polymyxin B sulfate (Sigma) was then added with or without the indicated cytokines. Human recombinant IL-lP (courtesy of Dr. A. Tagliabue, Sclavo, Siena, Italy) was used at 20 ng/ml final concen- tration, and human recombinant TNFa (BASF/Knoll, Ludwighafen,

22190

by guest on February 15, 2019http://w

ww

.jbc.org/D

ownloaded from

New CRP and SAP-related IL-l-inducible Gene 22191

Germany) was used at 500 units/ml. Human recombinant IL-6 was a kind gift of Dr. S. Gillis (Immunex Corp., Seattle, WA) and was used at 50 units of Cess/ml. At these concentrations, IL-10 and TNFa were able to induce biological effects on HUVEC such as neutrophil adhesion, procoagulant activity, and prostacyclin production (21).

RNA Extraction and Northern Blot Analysis-RNA was extracted and purified from HUVEC as previously described (22). 10 pg of total cellular RNA was run in standard formaldehyde-agarose gels, blotted onto nitrocellulose membranes (Schleicher and Schuell, Dassel, Ger- many), and fixed under vacuum at 80 "C for 2 h. Probes, labeling, and hybridizations were performed as described previously (22).

cDNA Libraries-Total RNA was isolated from HUVEC cultured for 1 h in the presence or absence of both IL-lp and cycloheximide at 10 pg/ml. Poly(A+) RNA was further purified by affinity chroma- tography on oligo(dT)-cellulose (23). A cDNA library was constructed in the X-ZAP11 vector (Stratagene, La Jolla, CA) as described (23).

Differential Screening-The cDNA probes were obtained as de- scribed (24); the library was plated at approximately l plaque/cm2 and transferred onto nitrocellulose membranes in duplicates by stand- ard procedures (23). The membranes were hybridized at 42 "C for 48 h in 0.2 M Tris-HC1, pH 7.6, 50% deionized formamide, 5 X SSC, 1 X Denhardt's solution, 10% dextran sulfate, 0.1% SDS, 100 pg/ml denatured salmon sperm DNA, and 1 X lo7 cpm/100 cm2 cDNA probe. Membranes were washed at 50-60 "C in 0.2 X SSC, 0.1% SDS. Plaques showing a stronger hybridization signal with the cDNA probes obtained from IL-lp-stimulated relative to untreated HUVEC were picked. These were rescreened three times differentially to obtain single clones. Phage inserts were rescued in the Bluescript vector and sequenced with the dideoxynucleotide chain termination method (25).

Genomic Library Screening-The genomic library was a kind gift of Dr. A. Neri (Ospedale Maggiore I.R.C.C.S., Milano, Italy). It consisted of human placental DNA partially digested with Sau3Alp and cloned into the BamHI site of the EMBL4 vector. 1.2 X lo6 plaques were screened with the 32P-labeled PTX3/A and E inserts.

In Vitro Transcription and Translation-Following digestion with BamHI and KpnI, the complete insert of the PTX3/D clone was subcloned into the pGem3 vector downstream from the SP6 promoter. The recombinant plasmid was cut with KpnI and transcribed with SP6 RNA polymerase according to the manufacturer's protocol (Pro- mega). Following digestion with RQ1 DNase (Promega), phenol chlo- roform extraction, and ethanol precipitation, one-tenth of the product was run in a standard formaldehyde-agarose gel to check for the quantity and size of the transcript obtained. Approximately 1 pg of in vitro RNA was used for the translation reaction in wheat germ extract in the presence of "S-labeled methionine according to the manufacturer's protocol (Promega). The final products were run in a standard 10% SDS-polyacrylamide gel, treated with Enhance inten- sifier (Du Pont-New England Nuclear) for 1 h, dried, and autoradi- ographed.

RNase Protection Assay-A 310-bp PstI fragment overlapping the 5' end of the PTXB first exon was subcloned in pGem3. It was linearized with HindIII, and a 32P-labeled RNA probe was generated with T7 RNA polymerase and [32P]CTP according to the manufac- turer's instructions (Promega). After digestion of the DNA template with DNase, the probe was precipitated in ethanol. 15 pg of total RNA or yeast tRNA control were hybridized with the 5 X lo5 cpm of probe in 86% deionized formamide, 0.4 M NaCl, 1 mM EDTA, 40 mM PIPES, pH 6.7. The hybridization was carried out for 16 h at 65 "C. The hybridization mix was then diluted 10 times into a buffer con- taining 10 mM Tris-HC1, pH 7.5, 5 mM EDTA, 0.3 M NaC1, and 40 pg/ml RNase A (Sigma) and 2 pg/ml RNase T1 (Boehringer Mann- heim) were added. Incubation was done for 5 min at room tempera- ture. After inactivation of the enzymes and addition of 10 pg of Escherichia coli tRNA, the hybrids were extracted once in phenol chloroform, precipitated, and then loaded onto a urea/polyacrylamide sequencing gel.

Southern Blots-10 pg of digested human placental DNA was blotted onto Hybond-N membranes (Amersham Corp.) in 2 X SSC. The membranes were heated to 80 "C for 2 h. Hybridization with a "P-labeled probe was carried out overnight at 65 "C in 5 X SSC, 0.5% SDS, 5% dextran sulfate, 5 X Denhardt's solution, and 100 yg/ml denatured salmon sperm DNA.

Peptide Synthesis and Specific Antiserum Production-The peptide GTWNSEEGLT was synthesized on solid phase with the 430 A peptide synthesizer (Applied Biosystems), using Fmoc (9-fluorenyl- methoxycarbonyl) as protective agent of the amino residue and 1- hydroxybenzotriazole and dicyclohexylcarbodiimide as activating

agents of the carboxyl residue. The synthetic peptide included the sequence GGC at its carboxyl terminus to allow conjugation with hemocyanin through cross-linking of the SH group of the peptide with the NH2 groups of hemocyanin with maleimidobenzoic acid N- hydroxysuccinimide.

10 pg of keyhole limpet hemocyanin-PTX3 peptide conjugate were emulsified in complete Freund's adjuvant and injected subcutaneously into a rabbit at multiple sites. Anti-PTX3 peptide antibodies were titrated by standard enzyme-linked immunosorbent assay using un- conjugated peptide as the first layer.

Cell Labeling and Immunoprecipitation-Confluent HUVEC were stimulated with TNF in the presence of 250 pCi of [%]methionine (Amersham Corp.) for 20 h at 37 "C. Aliquots of supernatant were preincubated for 1 h at 4 "C with protein G-Sepharose (at 10%). Anti- PTXB antiserum was adsorbed to 500 ~1 of protein G-Sepharose for 1 h at room temperature. The IgG-protein G complex was washed three times in wash buffer (0.01 M Tris-HC1, pH 7.4, 0.15 M NaCl, 0.5% Nonidet P-40) and then incubated with the precleared super- natants for 2 h at 4 "C. The immunoprecipitates were washed three times in wash buffer, resuspended in reducing sample buffer, boiled, and loaded onto a 10% SDS-polyacrylamide gel. The gels were treated with Enhance solution, dried, and autoradiographed.

Somatic Cell Hybrids-The somatic cell hybrids were obtained as already described (26). Human lymphocytes or fibroblast cells were fused with the HPRT-Chinese hamster ovary cell lines YH-21 (27) or RJK 88 (28) following a published protocol (29). The hybrids HY.86D13 and RJ.3PT1, used for subregion mapping, retained the region 3q21+3qter and 3p14"*qter, respectively, as described (30, 31).

In Situ Hybridization-Metaphase chromosome spreads were ob- tained from phytohemagglutinin-stimulated lymphocyte cultures of normal male individuals using standard cytogenetic methods. XP1 and XP2 phage DNAs were labeled by nick translation with biotin- 16-dUTP, purified through a Sephadex G-50 spun column, and co- precipitated with sonicated salmon sperm DNA and yeast tRNA, both at a concentration in 50-fold excess relative to the probe. Probes were hybridized to metaphase samples at a concentration of 5-10 ng/ pl in the presence of a 500-fold excess of unlabeled sonicated human DNA as competitor to suppress the signals of the repetitive sequences present in the probe. Hybridization and detection with fluorescein isothiocyanate-conjugated avidin were performed as described previ- ously (32). Metaphase spreads were stained with propidium iodide in antifade medium and photographed. The same metaphase spreads were subsequently G-banded with Wright's stain and rephotographed. Mapping was carried out by overlap image.

Computer Methods-The sequences were screened against the EMBL Data Bank (33) using the FASTA program (34) as imple- mented on the EMBnet computer resource (35). Screening of the SWISS-PROT protein sequence data bank (36) and detailed sequence analysis were carried out with the PC/Gene sequence analysis pack- age (IntelliGenetics).

RESULTS

Isolation of cDNA Clone+" cDNA library was constructed in the X-ZAP11 vector from poly(A+) mRNA purified from HUVEC stimulated with IL-16 for 1 h in the presence of cycloheximide. More than 4000 recombinant clones were screened for IL-16-inducible genes by differential hybridiza- tion. 38 clones repeatedly gave a stronger hybridization signal with the probe from IL-16-stimulated cells compared with untreated cells, and they were partially sequenced. 36 clones corresponded to genes known to be induced by IL-16 in endothelial cells, that is interleukin-8 (IL-8) (16 clones), endothelial leukocyte adhesion molecule 1 (7 clones), Gro-a (6 clones), Gro-6 (5 clones), and plasminogen activator inhib- itor (2 clones) (17, 18). The remaining two clones contained an identical sequence of approximately 900 base pairs which did not correspond to any known sequence present in the EMBL Data Bank. These are schematically represented in Fig. 1A as PTX3/A. PTX3/A contains an open reading frame starting at its 5' end (EcoRI site) and a TAA stop codon followed by a long AT-rich 3"untranslated region (70% AT) and a polyadenylation signal. The 3'-untranslated portion contains, in addition, two consensus sequences for mRNA

by guest on February 15, 2019http://w

ww

.jbc.org/D

ownloaded from

22192 New CRP and SAP-related IL-1 -inducible Gene

PTXBIA

PTX3IB

PTXYC PTX3lD

PTXBlE

FIG. 1. A, restriction map of PTXB cDNA. The positions of the ATG and TAA codons are indicated. E, EcoRI; H, HindIII; P, PvuII; S, Sad; 5'2, SucII. The arrows represent the extent and direction of sequencing of the longest PTX3/E clone. All of the different PTX3 clones obtained, labeled PTX3/A to -/E, are shown in the lower part. B , complete nucleotide and amino acid sequence of PTXB cDNA. The nucleotides are numbered from 1 to 1837 on the left side. The longest open reading frame and its translated protein sequence are shown. The amino acids are numbered from 1 to 381 on the right side. The potential signal peptide of 17 amino acids starting with the first methionine is underlined with a thin line. The protein putative instability region (PEST) is underlined with dots. The dotted line underscores the three amino acids that constitute a potential N- glycosylation site. The double line underlines the eight amino acids that constitute the "pentaxin family signature." Two RNA instability consensus sequences (ATTTA) are present in the 3"untranslated region and are underlined with a thick line. The polyadenylation signal is underlined three times. The triangle at the center of the sequence indicates the cysteine residue where the homology with the pentaxins begins. The positions of the three exons (ezl -3) relative to the cDNA sequence are indicated with arrows. These sequence data are available from EMBL/GenBank/DDBJ/PIR/SWISSPROT under accession no. X63613 for the nucleotide sequence and S18947 (PIR) and P26022 (SWISSPROT) for the amino acid sequence.

instability as shown in Fig. 1B. Characterization of the Complete PTX3 cDNA-To clone

the full-length cDNA, two other cDNA libraries from HUVEC were screened with PTX3/A as probe. Four additional clones were found (PTX3/B-E) and are schematically represented in Fig. lA. The longest (PTX3/E) contains an insert of 1837

base pairs (bp), which is quite close to the size of the mature message detected in Northern blots and estimated to be of approximately 1.8 kilobase pairs (see below). The different clones were sequenced in both orientations as schematized by arrows in Fig. lA. The complete nucleotide sequence of PTX3/E is shown in Fig. 1B and consists of a 68-bp 5'- untranslated region, an open reading frame of 1143 bp with a stop codon at position 1211, and a polyadenylation signal at position 1802.

Screening the complete nucleotide sequence against the EMBL Data Bank revealed that the region spanning from nucleotide 1 to 894 (the EcoRI site) is almost completely colinear (98% homology) with a sequence (accession no. M31166) corresponding to a partial cDNA sequence of a TNF- inducible gene, called TSG-l4a, isolated from human fibro- blasts (37). The M31166 sequence is three nucleotides longer at its 5' end compared with PTX3/E and is different at 17 positions within the region of alignment with the PTX3 sequence (it has nine nucleotides missing, two additional, and six altered relative to the PTXS sequence).

Analysis of the Predicted PTX3 Protein Sequence-The PTX3/E cDNA clone starts with an open reading frame. The most 5' methionine residue (at nucleotide position 68; Fig. 1B) is likely to be the one that is effectively used since it fits the Kozak consensus sequence (38) quite well. Furthermore, this methionine is immediately followed by a typical signal peptide sequence (underlined in Fig. 1B) as predicted accord- ing to the method of von Heijne (39). The predicted cleavage site for the signal peptide lies between the alanine and glu- tamic acid at positions 17 and 18, respectively. This putative signal peptide shows a 9 out of 12 amino acid identity with that of murine and human tyrosinases (SWISS-PROT acces- sion nos. P11344 and P14679), a further element suggesting that it is indeed functional.

The predicted protein sequence initiating with the first methionine is 381 amino acids long and was analyzed against the SWISS-PROT protein sequence data bank using the PC/ Gene package. A significant alignment was found between the COOH-terminal portion of the PTXB sequence, from the cysteine at position 179 (marked with a triangle in Fig. 1B) to the COOH-terminus, and the nine cloned members of the pentaxin gene family. Fig. 2A shows the alignment of the predicted protein sequence of PTX3 with the two human members of the pentaxin family, namely CRP and SAP 35 of the 208 amino acids of the consensus region are identical (17%) and 118 are conserved (57%) in all three proteins. The alignment for PTXS with the nine cloned pentaxins originat- ing from the mouse, rabbit, hamster, and horseshoe crab, in addition to man, shows an overall identity of 22 amino acids over 259 in consensus length (9%) and a similarity of 55 amino acids (21%) (data not shown). In particular, there are two cysteines at amino acid positions 210 and 271 of PTX3 (Fig. 2A) that are conserved in all members of the family and are thought to play a crucial role in determining the secondary structure of the molecule (1, 7). PTX3, on the other hand, completely diverges from all other pentaxins in its NH2- terminus and is longer. The degree of similarity between the COOH-terminal pentaxin-like domain of PTX3 and the var- ious known members of the pentaxin family was analyzed using the CLUSTAL program (40). The results, as shown by the dendrogram in Fig. 2B, suggest that PTX3 belongs to a fourth subgroup of this gene family, the first three containing the rodent and human CRPs, the SAP proteins, and the horseshoe crab CRPs, respectively.

The predicted PTX3 amino acid sequence was also screened against the PROSITE protein pattern data base (91, which

by guest on February 15, 2019http://w

ww

.jbc.org/D

ownloaded from

New CRP and SAP-related IL-I-inducible Gene 22193

FIG. 2. A , alignment and degree of conservation between PTX3 and the hu- man CRP and SAP proteins. The full protein sequences of human CRP and SAP are shown, whereas only the region of PTX3 showing similarity to the pen- taxins is indicated, from the cysteine at positions 179 to the COOH terminus. The dots represent conservation; the as- terisks identify identity. The "pentaxin family signature" is doubly underlined. The two cysteines conserved among all pentaxins are indicated by triangles. B , dendrogram of all cloned members of the pentaxin family of proteins indicating the degree of evolutionary relatedness between PTX3 and the other pentaxins. CRP$HUMAN, human CRP; CRP$- MOUSE, mouse CRP; CRP$RABBZT, rabbit CRP; CRPl$LZMPO, Atlantic horseshoe crab CRP1.1; CRP3$LIMPO, Atlantic horseshoe crab CRP3.3; CRP4$- LIMPO, Atlantic horseshoe crab CRP1.4; FP$CRZMZ, hamster female protein (SAP); SAP$HUMAN, human SAP; SAP$MOUSE, mouse SAP.

A

CRP SAP PTX3

CRP SAP P T X 3

CRP SAP PTX3

1 MEK-LLCFLVLTSLSHAFGQTDMSRMrVFPKESDTSYVSLKAPLTKPLKAFTVCLHFYT 1 MNKPLLWISVLTSLLEAFAHTDLSCKVFVFPRESVTDHVXLITPLEKPLPNFTLCFIVIYS

V

179 """""""_"""" CETAILFP~RSKKiFCSVHPV~MRLESFSAcIWV~ . .... =*. . .. . .....*, *. .

ELSSTRGYSIFSYATKR~NEILIFYS-KDIGYSFTVCGSEILFEVPEVTVAPVHICTSW DLS--RAYSLFSYN~RDNELLVYKE-RVGEYSLYICRHKVTPKVIEKFPhPVHICVSW

V

TDV~KTI-LFSYGTKRNPYEIQLYLSYQSIVNVGCEPNKLVAEAllVSLCRWT~ . .. . *==.*. , *. .. . . . . .. .. .*.* .*

ES~IVERIVffiK-PRVIU(SLKKGY~GAWSIIWEQDS--FGGNFEGSQSLVGDIG

I/liEEGLTSU)VWCELAATWEIUTCHIVPeGtILQICOEKNCCCVGGCFDET~FSGRLT ESSSCIAEIWIWCT-PLVKKG~OGYN~~KIV~QEQDS--YGGKFDRS~SFVCEIG

.*..*....e..*. . * .. 8 . * . . .***... * * *. . .. 9 .. CRP NVWI(*DFVLSPDEIWTI--YiCGPFSPNVLNWRALKYEVQCEVFT)(PPLYP SAP DLmWDSVLPPENILSA--YOGTPLDANILD~WYEIRCYVIIKPLVWV PTX3 GFNIYDSVLSNEEIRETCCA~SCHIRGNIVCffiVTEI9PHCC----AOYVS . . ** * e . ..= . . ... * . . . e . . . . = .

I

I

revealed the presence in PTX3 of the eight amino acids (H- X-C-X-S/T-W-X-S) that constitute the "pentaxin family sig- nature." These amino acids are doubly underlined in Figs. 1B and 2 A . On the other hand, the NHs-terminal portion of the predicted PTX3 protein, down to the cysteine residue at position 179 (Fig. lB) , is not significantly related to any of the currently known protein sequences.

Computer analysis showed, in addition the presence of a potential N-linked glycosylation site at amino acid position 220 (Fig. 1B) and a PEST consensus sequence within the NHs-terminal portion of the predicted protein (Fig. 1B, posi- tions 40-52), a motif generally associated with protein insta- bility (41).

Expression of PTX3 mRNA"PTX3 gene expression in different cell types was studied by Northern blot analysis (Fig. 3). In HUVEC, low but appreciable levels of PTX3 expression can be detected in unstimulated cells and PTXS mRNA is strongly induced by IL-lB and TNFa but not by IL-6. Induc- tion is already visible at 1 h with a peak between 2 and 6 h (Fig. 3A). The appearance of the message is evident also when IL-1P is added in the presence of cycloheximide (data not shown), as expected since the library was constructed follow- ing exposure of HUVEC to both IL-lP and cycloheximide. This suggests that PTXS is an "early" response gene to IL- I@. Interestingly, the same pattern of induction in response to IL-lP and TNFa, but not to IL-6, is maintained in the

CRPSHUMAH

CRPSRABBIT

CRPSMOLJSE

FPSCRIMI

SAPSMOUSE

SAPSHUMAH

CRPISLIMPO

CRP3SLIMPO

CRP4SLIMPO

P T X 3

5 ? K O

215

116

274 117

1 1 5 174 334

224 2 2 3 3 8 1

human hepatoma cell line HEP3B (Fig. 3B), which is an IL- 1- and IL-6-responsive cell line (14). Again, the PTX3 mes- sage appears with rapid and transient kinetics. Finally, also in the human fibrosarcoma cell line 8387, both IL-lP and TNFa are able to induce the expression of PTXS mRNA (Fig. 3C). In this case, however, the TNFa signal seems to be more effective than IL-18 in inducing PTX3.

PTX3 Genomic Organization-To determine the genomic structure of PTX3, we screened a human placental genomic library with the 32P-labeled PTX3/A and PTX3/E cDNAs. Two clones (AP1 and AP2) were isolated that contained inserts of 12.5 and 16.8 kilobase pairs, respectively (Fig. 4A). Restric- tion enzyme analysis combined with hybridization to several PTX3 cDNA fragments together with the subcloning and sequencing of exon-containing fragments (pP1-pP4) allowed us to determine the complete genomic map of PTX3 (Fig. 4A). The PTX3 genome is made of three exons, the first extending to nucleotide 197, the second covering 198-599, and the third extending from nucleotide 600 to the 3' end of the cDNA (Fig. 1B). Interestingly, the third exon of PTX3 matches exactly the region of homology of the predicted PTX3 protein with the other members of the pentaxin family (Figs. 1B and 2 A ) , whereas the first two exons are divergent.

To confirm the PTX3 genomic structure, human placental DNA was digested with EcoRI, BarnHI, Hind111 and Sac1 and analyzed by Southern blotting. The results are shown in Fig.

by guest on February 15, 2019http://w

ww

.jbc.org/D

ownloaded from

22194 New CRP and SAP-related IL-1-inducible Gene @ A B C A B C O A B C D A B C O

185- - 0

0 A B C A B C A B C

XIS-

... :. ,.,

FIG. 3. PTX3 gene expression in cytokine-stimulated cells. HUVEC (A), the hepatoma cell line HEP3B ( B ) , or the fibrosarcoma cell line 8387 (C) were cultured in the absence (lanes A ) or presence of human recombinant IL-lP (20 ng/ml) (lanes B ) , TNFa (500 units/ ml) (lanes C), or IL-6 (50 units/ml) (lanes D). Total RNA was extracted at the indicated times and analyzed by Northern blotting. In the lower part of each panel, the photograph of the ethidium bromide-stained nitrocellulose filters after RNA transfer is shown.

4B and fit the restriction map shown in Fig. 4A. Transcription Initiation Site-To determine the exact

PTX3 transcription initiation site, a 310-bp PstI fragment, spanning the 5' boundary of the PTX3 first exon, was used to generate an RNA probe (Fig. 5B) for RNase protection analysis. The results show a most abundant protected frag- ment of 134 bases, indicating a unique transcription initiation site (Fig. 5A). The other two higher molecular weight bands detected were nonspecific products since they were also pres- ent in the negative control (Fig. 5A, lane C). The sequence of the protected fragment is shown in Fig. 5C, and the portion overlapping with the longest cDNA clone (PTX3/E) is under- lined. These data reveal that the transcript contains 24 addi- tional nucleotides at its 5' end relative to the PTX3/E clone and completes the sequence analysis of the PTX3 transcript. The exact size of the PTX3 transcript is therefore 1861 nucleotides. The ATG boxed in Fig. 5C and located a t position 92 of the complete PTXS transcript is thus the most 5' ATG in frame with the longest PTXS open reading frame.

RNase mapping was also performed with RNA extracted from HUVEC stimulated with IL-10. The results show that the same transcriptional start site is used in activated as in untreated cells. Furthermore, the signal is clearly increased a t 30 min and even more a t 6 h (Fig. 5A), thus confirming and extending the Northern blot data (Fig. 3A). Finally, the RNase protection assay revealed the lack of induction of PTX3 transcript by human interferon-y (Fig. 5A).

PTX3 Protein Synthesis and Secretion-To verify that the first methionine was that which effectively initiates protein synthesis, the longest cDNA clone (PTX3/E, Fig. lA) was transcribed and translated in uitro. The most abundant prod- uct was a 42-kDa protein that was absent from the negative control lane (data not shown), in total agreement with the predicted molecular mass of a protein starting with the most 5' methionine.

With the view of demonstrating that the predicted PTX3

B , expt 1 expt 2 . E B H S

4.4

25 2.0

5.1

2: 21

FIG. 4. A, cloning and restriction map of the PTX3 genome. Two separate X clones were isolated and are indicated with thick lines labeled XPI and "2. Their approximate length is indicated in paren- theses. Fragments of the XP1 and XP2 inserts were subcloned in the pGem3 plasmid and are indicated with thin lines labeled PPI-pP4. The X and plasmid clone inserts are positioned in relation to the complete restriction map of the PTX3 genome. The boxes represent the three exons. The portions of the boxes shown in black represent the translated part of the exons, whereas the dashed regions represent the untranslated sequences. The two regions of the PTX3 genome that contain the three exons were completely sequenced and their restriction is shown in more detail in the lower part. The arrows indicate the lengths and positions of the sequences performed. E, EcoRI; H , HindIII; P, PuuII; Ps, PstI; S, Sad ; S2, SacII; X, XbaI. None of the PTX3 clones were cut by BamHI. B, Southern blot analysis of PTX3. Human placental DNA was digested with the EcoRI ( E ) , BamHI ( B ) , HindIII ( H ) , or Sac1 ( S ) enzymes and analyzed by Southern blotting using the longest PTX3 cDNA clone (PTXB/E) as probe.

protein is indeed secreted from stimulated HUVEC, an anti- serum was raised in a rabbit against a synthetic decapeptide (GTWNSEEGLT) extending from amino acid position 272 to 281 (Fig. 1B). This serum was able to bind the unconjugated peptide in an enzyme-linked immunosorbent assay and spe- cifically recognized the bacterially made PTX3 protein in Western blots (data not shown). Furthermore, this antiserum immunoprecipitated a 46-kDa protein from the supernatant of TNFa-stimulated HUVEC (data not shown), suggesting that PTX3 is glycosylated like the SAP of several animal species (1, 7). Indeed, a potential N-linked glycosylation site is present a t amino acid position 220 in the predicted PTX3 sequence (Fig. 1B).

Chromosomal Localization-To determine the chromosomal localization of the PTX3, Southern analysis was performed on DNA extracted from the human-hamster somatic hybrids shown in Table I (26). The DNA was cut with BamHI, and the blots were hybridized with the "'P-labeled PTX3/D cDNA as probe. The human genomic DNA and four hybrid clones showed a positive signal, whereas the hamster DNA did not show any cross-hybridization. All four positive clones carried the human chromosome 3 giving a 100% concordance between a positive PTX3 signal and the presence of chromosome 3 (Table I). To confirm and extend these data, Southern analy- sis was carried out on three more hybrids, HY.86D13 and FLJ.3PT2, which retain the region 3q21-qter and 3p14-qter,

by guest on February 15, 2019http://w

ww

.jbc.org/D

ownloaded from

New CRP and SAP-related IL-1-inducible Gene 22195

respectively, and HY.137RT9, which retains the entire human chromosome 3 only (30,31). All three hybrids showed a human PTX3-specific, positive band (data not shown). These results show that PTX3 is localized on human chromosome 3q21- qter.

To determine precisely the localization of PTX3 on human chromosome 3, in situ hybridization was carried out using both XP1 and XP2 phage DNAs as probes (Fig. 4A). Analysis

134 I

P'&

C

G C C l C l C n C i C r C r C l C l C C 1 C C G C I C l r r C I C ~ G C 1 C n C r ' 5 " 5 n ~ ~

~ ~ G t C ~ 5 ' 0 ~ r " n G a a C ' r ' ( ; C b r C ~ ~ ~ ~ ~ ~

FIG. 5. A, RNase protection assay. Total RNA was isolated from unstimulated HUVEC (-) or cells treated with IL-1@ (20 ng/ml) or interferon-y (1000 units/ml) for the indicated times. 15 pg of total RNA or yeast tRNA as control (lane C) were hybridized with the n2P- labeled probe indicated in B, digested with RNase, and run in a polyacrylamide gel along with a known sequence ladder as molecular weight marker. B, schematic representation of the probe used relative to the position of the first PTX3 exon. C, complete sequence of the protected fragment. The portion of the protected sequence in common with the longest cDNA clone (PTX3/E) is underlined. The first ATG codon in frame with the long open reading frame is boxed.

'.'OtC"'C-~j,~!'O"C 'C't'GGTC' CtrG '3.:

of 32 metaphases hybridized with XP1 DNA revealed a total of 124 fluorescent signals. Only signals in contact with chro- matids were scored. 68 (55%) were localized on chromosome 3 at the q24-q26.1 region, with a peak (45 spots corresponding to 36% of the total) on band q25 (Fig. 6). In 31% of the metaphases analyzed, specific signals on this region were present on both chromosome 3s, and 41% of all chromosome 3s showed signals on both chromatids. Similar results were obtained with the XP2 phage DNA probe (data not shown). No other significant signal cluster was detected on any other chromosome. These results are in agreement with the analysis of somatic cell hybrids described above and permit localization of the PTX3 gene on the q25 band of human chromosome 3.

DISCUSSION

During the differential screening of a cDNA library con- structed from IL-I@-stimulated HUVEC, we have isolated a new gene, which we called PTX3, that shows homology in amino acid sequence to the pentaxin family of genes. We have characterized the complete PTX3 transcript, studied the reg- ulation of its expression in cytokine-stimulated cells, and immunoprecipitated the protein from stimulated HUVEC. In addition, we show that the PTX3 genome is made of three exons and is localized on human chromosome 3 band q25.

The isolation of a nearly full-length cDNA clone together with the data from RNase mapping and sequencing of the first exon for PTXB has shown that the PTXB transcript is 1861 nucleotides long and has a unique transcription initia- tion site. The transcript starts with a 1143-nucleotide-long open reading frame. Several lines of evidence (including the presence of a consensus Kozak sequence and a typical signal peptide sequence as well as in vitro translation) demonstrate that the first methionine, a t nucleotide 92 of the complete transcript, is the site of translation initiation, leading to the synthesis of a 381-amino acid-long protein.

Computer analysis of the PTX3 protein sequence has re- vealed a significant homology between the COOH-terminal, the 203-amino acid-long portion of PTX3, and the members of the pentaxin family of genes, which include the CRP and SAP of several animal species. PTX3 also contains the typical

TABLE I Synteny analysis for pTX3

The percent concordance is calculated by dividing the number of concordant hybrids by the total number of hybrids analyzed. Human chromosome present (+) or absent (-)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 X Hybrids

Positive HY.95A1 HY.95S Y. 173.5CT3 Y.XY.BF6 RJ.3PT2

HY.19.16T3D HY.22AZAl HY.31.24E HY.6OA HY.70B2 HY.75E1 HY.94A HY.95B HY.137J HY.166T4 RJ.387.51T5 RJ.387.58 YC2Tl

Negative

" + - + " " + + " + - - - - - - - - + - + + " " " " _ + """ + + - + + - + + - + - + " + + - + + - - + - - + + + " + + + + + - - + + + - " + + - + " + + " + " " " " " " " " " "

Concordance (%) 66 77 100 77 55 55 61 55 61 77 77 44 55 50 55 66 66 44 66 55 61 72 33

by guest on February 15, 2019http://w

ww

.jbc.org/D

ownloaded from

22196 New CRP and SAP-related IL-1-inducible Gene

FIG. 6. In situ hybridization of PTX3 DNA to human chro- mosome band 3q25. a, b, and c show partial metaphases hybridized with biotin-labeled XP1 phage DNA. Hybridization was detected with fluorescein isothiocyanate-conjugated avidin (white spots), and the chromosomes were counterstained with propidium iodide. A , B, and C show the same partial metaphases presented in a, b, and c, respec- tively, after G-banding with Wright’s stain. Specific hybridization signals on band 3q25 are indicated by arrows. On the right, the idiogram of the human G-banded chromosome 3 showing the distri- bution of the hybridization signals is shown. Each large clot represents five hybridization signals.

pentaxin family signature sequence (H-X-C-X-S/T-W-X-S) and the two cysteines conserved in all known pentaxins from different species, which form a disulfide bridge in SAP and CRP (1, 7).

Several lines of evidence, however, suggest that PTX3 belongs to a different pentaxin subgroup than the human SAP and CRP. The first is indicated by the degree of homology between these various proteins and by the fact that PTXB is much longer than SAP and CRP at its NH2 terminus (178 amino acids in PTX3 relative to 22 and 23 in CRP and SAP, respectively). The second derives from the comparison of the genomic organization of PTX3 relative to that of CRP and SAP. The latter two are made of two exons (5,6,8), whereas PTX3 has three. Interestingly, the third exon corresponds exactly to the region of homology with the pentaxins, sug- gesting an early evolutionary divergence of this exon. Third, the PTX3 gene is localized on human chromosome 3 (q25), whereas those of SAP and CRP are on chromosome 1 (42, 43). It is perhaps intriguing that the gene for another acute phase protein, ceruloplasmin, is localized on chromosome 3 band q21-q25 (44).

The NH2-terminal, 178-amino acid-long portion of PTXB did not show homology to any known protein sequence.

Particular features of this region include a highly hydro- philic profile and the presence of a putative “PEST sequence, indicating that the PTX3 protein is likely to be relatively unstable.

Our finding that PTX3 mRNA expression can be induced by cytokines is another characteristic that PTX3 shares with SAP and CRP. A number of differences in the regulation of these genes are worth noting, however. PTX3 expression is strongly induced in HUVEC and hepatocytic and fibroblastic cells, whereas CRP and SAP expression is confined to hepa- tocytes (45, 46) and has once been reported in blood mono- nuclear cells (47). Furthermore, PTXB mRNA is strongly induced by both IL-lD and TNFa but not by IL-6 or inter-

feron-?, whereas the major and direct inducer of SAP and CRP is IL-6 (13-16).

Vascular endothelial cells have long been viewed as a pas- sive lining of blood vessels endowed with negative properties such as being non-thrombogenic. In recent years, however, endothelial cells have emerged as active participants of in- flammatory, immunological, and thrombotic reactions (17- 19). In particular, upon exposure to inflammatory mediators such as IL-1 and TNFa, endothelial cells undergo profound alterations in gene expression and function (18). The recog- nition of the central role played by vascular endothelium in a variety of pathophysiological conditions has stimulated the search for novel genes associated with endothelial activation (Refs. 48-52 and the present study). Among these, the PTX3 gene reported here has characteristics that make it attractive as a potential marker for inflammatory conditions. PTXB is clearly related to CRP and SAP, two classical diagnostic indicators of the acute phase response (3, 4). Yet, several features distinguish PTX3 from classical pentaxins, including its sequence, chromosomal localization, gene organization, pattern of expression in different cell types, and specificity of induction by cytokines. Thus, PTX3 may complement the classical pentaxins and represent a novel indicator of tissue reactions, particularly of those involving the vessel wall.

Acknouledgments-We thank Dr. E. Monzani, G. Calabrese, and M. Salmona for the synthesis, purification, and conjugation of the peptide, Dr. L. Lombardi for useful discussions, and V. De Ceglie and F. De Ceglie for the photographic work.

REFERENCES 1. Pepys, M. B., and Baltz, M. L. (1983) Adu. Zmmunol. 34,141-211 2. Pepys, M. B., Baltz, M., Gomer, K., Davies, A. J. S., and Doenhoff, M.

3. Tillett, W. S., and Francis, T., Jr. (1930) J. Exp. Med. 52, 561-571 4. Abernethy, T. J., and Avery, 0. T. (1941) J. Exp. Med. 73,,173-182 5. Woo, P., Korenberg, J. R., and Whitehead, A. S. (1985) J. B~ol. Chem. 260,

6. Lei, K. J., Liu, T., Zon, G., Soravia, E., Liu, T. Y., and Goldman, N. D.

7. Nguyen, N. Y., Suzuki, A., Boykins, R. A., and Liu, T. Y. (1986) J. Biol.

8. Ohnishi, S., Maeda, S., Shimada, K., and Arao, T. (1986) J. Biochem.

9. Bairoch, A. (1991) Nucleic Acids Res. 19,2241-2245

(1979) Nature 278, 259-261

13384-13388

(1985) J. Biol. Chem. 260, 13377-13383

Chem. 261,10456-10465

(Tokyo) 100,849-858

10. Kushner, I., Ganapathi, M., and Schultz, D. (1988) Ann. N. Y. Acad. Sci.

11. Ciliberto, G. (1989) in Acute Phose Proteins and the Acute Phase Response (Pepys, M., ed) pp. 29-46, Springer Verlag, Heidelberg, Germany

12. Yap, S. H., Moshage, H. J., Hazenber , B. P. C., Roelofs, H. M. J., Bijzet, J.! Limburg, P. C., Aarden, L. A., a n t van Rijswijk, M. H. (1991) Biochim.

13. Li, S. P., Liu, T. Y., and Goldman, N. D. (1990) J. Biol. Chem. 265,4136- B~ophy~ . Acta 1091,405-408

557, 19-30

4142 14. Ganter. U.. Arcone. R.. Toniatti. C.. Morrone. G.. and Ciliberto, G. (1989)

15.

16.

17. 18.

19. 20.

21.

22.

23.

24. 25.

26.

27.

28.

29. 30.

EMBO J. 8,3773-3779 457-465

. . . .

Majello, B., Arcone, R., Toniatti, C., and Ciliberto, G. (1990) EMBO J. 9,

Toniatti, C., Demartis, A., Monaci, P., Nicosia, A., and Ciliberto, G. (1990)

Mantovani, A., and Dejana, E. (1989) Immunol. Today 10,370-375 Mantovani, A., Bussolino, F., and Dejana, E. (1992) FASEB J. 6, 2597-

Pober, J., and Cotran, R. S. (1990) Physiol. Reu. 70,427-451 Dejana, E., Colella, S., Conforti, M. G., Abbadini, M., Gaboli, M., and

Dejana, E., Breviario, F., Erroi, A., Bussolino, F., Mussoni, L., Gramse, M., Marchisio, P. C. (1988) J. Cell Biol. 107, 1215-1223

Pintucci. G.. Casali. B.. Dinarello. C. A.. van Damme. J.. and Mantovani.

EMBO J. 9,4467-4475

2599

A 1 ~ ( 1 9 8 7 j Bbod 69; 69g-699

M. (1991) Blood 77.149-158

I~ I , .

Golay, J., Capucci, A., Arsura, M., Castellano, M., Rizzo, V., and Introna,

Sambrook, J., Fritsch,‘E. F., and Maniatis, T. (1989) Moleculnr,Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory, Cold Sprlng Harbor,

Berger, S. L., and Kimmel, A. R. (1987) Methods Enzymol. 152,324-325 NY

Tabor, S., and Richardson, C. C. (1987) Proc. Natl. Acad. Sci. U. S. A. 84,

Rocchi, M., Roncuzzi, L., Santamaria, R., Archidiacono, N., Dente, L., and

Rosentraus, M., and Chasin, L. A. (1975) Proc. Natl. Acad. Sci. U. S. A.

Fuscoe. J. C.. Fenwick. R. G.. Ledbetter. D. H., and Caskev. C. T. (1983)

4767-4771

Romeo, G. (1986) Hum. Genet. 74,30-33

72,493-497

MoL’Cell. Biol. 3, 1086-1096 Davidson, R. L. (1976) Somatic Cell Genet. 2, 165-176 Rocchi. M.. Colantuoni. V.. Covone. A,, Faronio, R., and Romeo, G. (1989)

Sombtic Cell Mol. Genet. 15, 185-190

by guest on February 15, 2019http://w

ww

.jbc.org/D

ownloaded from

31.

32.

33. 34.

35.

New CRP and SAP-related IL-I-inducible Gene 22197 Ottone, F., Giussani, F., Gaudi, S., Malcovati, M., Tenchini, M., and Rocchi,

M. (1991) C to enet CellGenet Montanaro, 7, Easamassimi, A:, DUrso, M., Yoon, J. Y., Freije, W.,

Schlessinger, D., Muenke, M., Nussbaum, R. L., Saccone, S., Mau eri, S., Santoro, A. M., Motta, S., and Della Valle, G. (1991) Am. J. d m . Genet. 4 8 , 183-194

Stoehr, P. J. and Cameron G. N. (1991) Nucleic Acids Res. 19,2227-2230 Pearson, W.'R., and Lipmin, D. J. (1988) Proc. Natl. Acad. Sci. U. S. A.

Stoehr, P. J., and Omond, R. (1991) in Genome Analysis: From Se uence to 85, 2444-2448

Functron (Collins, J., and Driesel, A. J., eds) p. 177, Huethig B u d Verlag, Heidelberg Germany

36. Bairoch, A., s'nd Boeckmann, B. (1991) Nucleic Acids Res. 19, 2247-2249 37. Lee, T. H., Lee, W. L., Ziff, E. B., and Vilcek, J. (1990) Mol. Cell. Biol. 10 ,

1982-19RR 38. Kozak. M. (1989) J. Cell Biol. 108,229-241

- - . - - . - -

39. von Heijne.G. (i986) Nucleic Acids Res. 14,4683-4690 40. Hi gins, D.' G., and Sharp, P. M. (1989) Comput. Appl. Biosci. 5,151-153 41. Ro%inson, E. A Yoshimura. T Leonard, E. J. Tanaka, S., Griffin, P. R.,

Shabanowitz? J.. Hunt. D. F.'. and Auoella. E. (1989) Proc. Natl. Acad. Sci. U. S. A . 86,' 1850-1854 '

_.

42. Mantzouranis, E. C., Dowton, S. B., Whitehead, A. S., Edge, M. D., Bruns, G. A. P., and Colten, H. R. (1985) J. Biol. Chem. 260,7752-7756

43. Whitehead, A. S., Bruns, G. A. P., Markham, A. F., Colten, H. R., and Woods, D. E. (1983) Science 221,69-71

44. Royle, N. J., Irwin, D. M., Koschinsky, M. L., MacGillivray, R. T., and Hamerton, J. L. (1987) Somatic Cell Mol. Genet. 13, 285-292

45. Kushner, I., and Feldmann, G. (1978) J. Exp. Med. 148,466-477

47. Murphy, T. M., Baum, L. L., and Beaman, K. D. (1991) J. Exp. Med. 173 , 46. Zahedi, K., and Whitehead, A. S. (1989) J. Immunol. 143,2880-2886

48. Hla, T., and Maciag, T. (1990) J. Biol. Chem. 265,9308-9313 49. Dixit, V. M., Green, S., Sarma, V., Holzman, L. B., Wolf, F. W., ORourke,

K., Ward, P. A,, Prochownik, E. V., and Marks, R. M. (1990) J. Biol. C h m . 265,2973-2978

50. Ouiuari. A. W.. Jr.. Bowski. M. S.. and Dixit, V. M. (1990) J. Biol. Chem.

495-498

-266,'14705-14708 - '

5830-5838 51. Holzman, L. B., Marks, R. M., and Dixit, V. M. (1990) Mol. Cell. Biol. 10,

52. Wolf, F. W., Marks, R. M., Sarma, V., Byers, M. G., Katz, R. W., Shows, T. B., and Dixit, V. M. (1992) J. Biol. Chem. 267 , 1317-1326

by guest on February 15, 2019http://w

ww

.jbc.org/D

ownloaded from

Marzella, V Predazzi and M RocchiF Breviario, E M d'Aniello, J Golay, G Peri, B Bottazzi, A Bairoch, S Saccone, R

C-reactive protein and serum amyloid P component.Interleukin-1-inducible genes in endothelial cells. Cloning of a new gene related to

1992, 267:22190-22197.J. Biol. Chem. 

  http://www.jbc.org/content/267/31/22190Access the most updated version of this article at

 Alerts:

  When a correction for this article is posted• 

When this article is cited• 

to choose from all of JBC's e-mail alertsClick here

  http://www.jbc.org/content/267/31/22190.full.html#ref-list-1

This article cites 0 references, 0 of which can be accessed free at

by guest on February 15, 2019http://w

ww

.jbc.org/D

ownloaded from