Molecular Cloning and Expression of Two Distinct cDNA-encoding

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

Vol. 267, No. 7, Issue of March 5 , pp. 4870-4877,1992 Printed in U.S.A.

Molecular Cloningand Expressionof Two Distinct cDNA-encoding Heparan Sulfate Proteoglycan Core Proteins from a Rat Endothelial Cell Line* (Received for publication, August 23, 1991)

Tetsuhito KojimaSQ,Nicholas W.ShworakSV, and RobertD. RosenbergSOll From the $Department of Biology, Massachusetts Institute of Technology, Cambridge,Massachusetts 02139 and the §Department of Medicine, Harvard Medical School,Beth Israel Hospital, Boston, Massachusetts 02215

The cloned rat fat pad endothelial cell (RFP-EC) line (HSPGlaCt) andanticoagulantly inactive heparan sulfate prosynthesizesanticoagulantlyactiveheparansulfate teoglycans (HSPGL“BCt)were isolated from a rat fat padendoproteoglycans (HSPGeCt)and anticoagulantly inactive thelial cell (RFP-EC) line as outlined in the previous article heparansulfateproteoglycans(HSPGnaCt),both of (Kojima et al., 1992). HSPGBCt possess mainly anticoagulantly which exhibit 25-, 30-, and 50-kDa core proteins of active heparan sulfate chains (HSSCt)whereas HSPG‘”BCt posextremely similar structure. The primary sequences of sess mainly anticoagulantly inactive heparan sulfate chains internal peptides obtained from H S P P t core proteins (HSinact). These observations suggest that HSPGBCt and and theNH2-terminal sequence analyses of the 25-kDa HSPG“t core proteins might have different structures which component from the HSPGinaCtcore proteins demon- couldbe involved in establishing the monosaccharide sestrate that the 30-kDa component is a previously uni- quence of covalently linked GAG chains. However, both sets dentified species, designated as ryudocan, with the 25-of core proteins exhibit molecular masses of 50-, 30-, and 25kDa component representing a proteolytic degradation kDa, respectively. The structures ofHSPGBCt and H S P P t were examined product, while the BO-kDa component is the rat homolog of syndecan (Saunders, S. Jalkanen, M., O’Far- by cleaving the two sets of core proteins with trypsin or rell, S., and Bernfield, M. (1989) J. Cell Biol. 108, endopeptidase Glu-C, the resultant peptides were separated 1547-1556). Specific oligonucleotide probeswere ob- by reverse-phase HPLC, and the chromatographic profiles tained for ryudocan and syndecan by polymerase chain revealed similar peptide fragments. Based upon the above reaction, and the correspondingcDNAs were isolated data, it remained unclear whether the RFP-EC line synthefrom a RFP-EC library. The cDNAs encode type I sizes a single class of HSPGs with partial degradation to integral membrane proteins of 202 and 313 amino several different molecular weight forms or multiple classes acids, respectively, which have homologous transmem- of HSPGs and whether the structural microheterogeneity of in determining the brane and intracellular domains but very distinct ex- the core protein(s) couldbeinvolved monosaccharide sequence of the linked GAG chains. tracellular regions. In particular, ryudocan exhibits In this paper, we have cloned two different cDNAs which onlythreepotential glycosaminoglycan attachment encode the three components within the proteoglycan prepasites within the extracellular region while syndecan has five glycosaminoglycan attachment sites within theration, determined their nucleotide sequences, and demonsame domain. Both species are expressed in RFP-EC strated expression in endothelial cells and smooth muscle lines, primary rat aorticsmooth muscle cells and pri- cells. These data allow us to define a new core protein of mary rat skinfibroblast cells. The levels of ryudocan importance to theblood vessel walland to suggest a new role and syndecan mRNA were measured by quantitative for a previously defined core protein. polymerase chain reaction in primary microvascular endothelial cells and closely associated non-endothelial EXPERIMENTALPROCEDURES cells isolated by cell sorting. Ryudocan and syndecan Materials-Materials were obtained from the same sources as mRNAs were abundantly expressed in both popula- indicated in the preceding paper (Kojima et al., 1992) except as noted tions representing about0.1-0.5% of mRNA. below.

The anticoagulantly active heparan sulfate proteoglycans

* This work wassupported in part by National Institutes of Health Grant HL41484. 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. The nucleotide sequencefs)reported in this paper has been submitted to the GenBankTM/EMBL Data Bank withaccessionnumber(s) M81785 and M81786. ll Recipient of an Medical Research Council Centennial Fellowship. 11 To whom correspondence and reprint requests should be addressed: Massachusetts Institute of Technology, E25-229, 77 Massachusetts Ave., Cambridge, MA 02139.

Cell Culture-The primary RFP-EC, primary rat skin fibroblast cells (RFC), and primary rat aortic smooth muscle cells (RASMC) were isolated, cultured, and utilized at low passage as previously described (Marcum and Rosenberg, 1985; Marcum et al., 1986; Castellot et al., 1982; Madri etal., 1983). The RFP-ECline was originally isolated as outlined above, cloned from single cells, and then spontaneously transformed into a stable cell line which was extensively characterized in prior publications (de Agostini et al., 1990; Kojima

’ The abbreviations used are: HSPG, heparin sulfate proteoglycan; DiI-Ac-LDL, ~,l’-dioctadecyl-1-1-3,3,3’,3’-tetramethyl-indo-carbocyanine perchlorate-labeled acetylated low density lipoprotein; GAG, glycosaminoglycan; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; kb, kilobase; PAGE, polyacrylamide gel electrophoresis; bp, base pair(s); r, rat; RASMC, rat aortic smooth muscle cell; RFP-EC, rat fat pad endothelial cell; RFC, rat skin fibroblast cell; SDS, sodium dodecyl sulfate; PCR, polymerase chain reaction; HPLC, high performance liquid chromatography.

4870

CloningExpression and

of Two Distinct Core Proteins

4871

et al., 1992). The RFP-EC line was subcloned by seeding trypsinized TABLEI single cells into 96-well culture plates a t a concentration of 0.5 cells/ Peptide and primersequences for PCR cloning well. Single colonies were selected and then expanded in wells of A: NHz-terminal sequences of trypticor endopeptidase Glu-Cincreasing size. digested peptides derived from HSRG'""et core proteins of the RFPPeptide Purification and Sequencing-HSPWt core proteins were EC line and utilized for PCR primer constructs (S, sense primer; A, purified from the RFP-EC line, samples of20-40pg were digested antisense primer). The -1 position of peptide V46-1 was designated with either trypsin or endopeptidase Glu-C (V8) (Boehringer Mann- as E or Dbecause of the specificity of endopeptidase Glu-C. Peptides heim), andthen chromatographed ona microbore reverse-phase V41 and T12 contain overlapping sequence (see Fig. 3). B: degenerate HPLC column as described in the accompanying paper (Kojima et PCR primers containing a 5' EcoRI or XbaI restriction site with a al., 1992). The fractions were monitored for absorbance at 214 and "GC clamp." 280 nm, peptide peaks were individually collected, and analyzed with A. Peptide a model 470A Protein Sequenator (Applied Biosystems, Foster City, CA). The isolation and sequencing of peptides were carried out twice 1s 1A in order to confirm the structural data. Purified core proteins (4 pg) 1. V46-1 ( E /D)NAQPGIRvPsEPKELE from HSPG'"""were also subjected to 12% SDS-PAGE, separated 2s 2A proteins were transferred to Immobilon-N membranes (Millipore, 2. T25 YFXGEEEDAGGLEQDSDFEL Bedford, MA), and theNH,-terminal sequence of the 25-kDa species 3s 3A was determined (Matsudaira, 1987).The latterprocess was monitored 3. V41/T12 TSCENTAVAGVEPDLRNQSPV~~~~ by including tracer-iodinated core proteins of HSPG'""". B. Primer Degeneracy CommonMolecularBiological Techniques-Total RNA was isoEcoRI lated by the procedure of Chomczynski and Sacchi (1987). Plasmid 1s 5' - GGGGAATTCGANAAYGCNCARCCNGG - 3l e x256 DNAs were prepared by the alkaline lysis method (Maniatis et al., Xba I 1982) and sequenced by the dideoxynucleotide chain-termination method (Sanger et al., 1977)with Sequenase (United States Biochem1A 5"GGGTCTAGATCNARYTCYTTNGGYTC-3' X256 ical Corp., Cleveland, OH) using both dGTP and dITPaccording to Xba I the manufacturer's protocols. Unless stated otherwise, hybridizations 2s 5"GGGTCTAGAGCNYTNCCNGAYGAYGA-3' x512 were performed at 65 "C in 6 X SSC, 5 X Denhardt's solution, 1% EcoRI X512 2A 5' -GGGGAATTCARYTCRAARTCNSh'RTC - 3' SDS, and 100 pg/ml herring DNA with washing at 65 "C in 0.2 X SSC with 0.1% SDS. Xba I 3 s 5' - GGGTCTAGAGGNGARAAYACNGCNGT-3' X256 Isolation ofPoly(A)+RNA and First Strand cDNA SynthesisEcoRI Poly(A)+ RNA was purified from total RNA by two rounds of chroX256 3A 5"GGGGAATTCCCNGTNGCNCCYTCRTC-3' matography on oligo(dT)-cellulose spun columns (Pharmacia LKB Biotechnology Inc.). First strand cDNA synthesis was performed as a (R, A or G; Y, T or C; W, A or T; S, C or G; and N, A or C or G described by Moremen (1989) using 5 pg of poly(A)+ RNA primed or T). with oligo(dT)12.18 (Pharmacia LKBBiotechnology Inc.) and Moloney murine leukemia virus reverse transcriptase (BethesdaResearch Labfollowing changes. Each reaction contained 3 pg of RFP-EC poly(A)+ oratories). PCR Cloning-Degenerate PCR primers (Table I), containing a 5' RNA and 100fmol of 32Pend-labeled primer. The primers were EcoRI or XbaI restriction site with a "GC clamp," were synthesized dTCCTGCTCAAGGCCCCCAGCGTCTTC for ryudocan, and dACwith a 380B DNA synthesizer (Applied Biosystems). PCR amplifi- TGCCAGATTCCTTCCTCAGCGCC (Pl) or dGTCTTCAGGAGGcations were performed with a GeneAmp Kit (Perkin Elmer Cetus) CACATTTGCGGTG (P2) for syndecan. Primer and template were using the following modifications to the manufacturer's protocol. annealed for 2 h at 74 "C for syndecan or 2 h at 81 "C for ryudocan. Reactions contained 100 ng of RFP-EC first strand cDNA with 10- Primed DNA was extended at the annealing temperature with 10 20 pmol of primers and were subjected to 35 cycles of amplification units of avian reverse transcriptase (Molecular Genetics Resources, which included denaturation (1min, 94 "C), annealing(1 min, 35 "C), Tampa, FL). and extension (1 min, 72 "C) followed by a final extension at 72 "C I n Vitro Transcription and Translation-Synthetic capped ryudofor 10 rnin. Reaction products were resolved by ethidium bromide- can mRNA was generated from the vector pNWS96 which contains agarose gel electrophoresis (3% NuSieve agarose, FMC BioProducts, a reconstituted full-length ryudocan cDNA (positions 1-2452 plus Rockland, ME; 1%agarose, Bethesda Research Laboratories). PCR approximately 130 bpof poly(A) tail) inserted 5' to 3' between BamHI fragments of the anticipated size were isolated by 10% PAGE, purified and EcoRI of pBluescript SK-. In vitro transcriptions were performed on Nensorb 20 columns (Du Pont-New England Nuclear), digested with T3 polymerase in the presence of m7G(5')ppp(5')G (New Engwith EcoRI and/or XbaI, and subcloned into pUC19. land Biolabs) as previously described (Sambrook et al., 1989) using Preparation and Screening of a cDNA Library-Using the manu- 2.5 pg of Sal1 or Hind111 linearized pNWS96 to respectively generate facturer's recommended conditions, a X ZAPII (Stratagene, La Jolla, a 2674-bp full-length transcript or a 679-bp transcript which termiCA) library of RFP-EC cDNA constructed as described previously nates 7 bp past the coding region. In vitro translations, 25-pl reac(Doi et al., 1987) was prepared and used to infect Escherichia coli tions, were performed with Promega's nuclease-treated rabbit reticXL1-Blue. Approximately lo6 plaques were transferred to Colony/ ulocyte lysate and canine pancreatic microsomal membranes accordPlaque Screen (Du Pont-New England Nuclear) and screened with ing to the manufacturer's specifications using 0.5pgof synthetic ''2P-labeled probes specific for ryudocan (random primed TV215 oli- mRNA and 4 pCi of ~-[3,4,5-~H]leucine (158 Ci/mmol, 1 mCi/ml, Du gonucleotide; hybridization at 42 "C with 48% formamide), or syn- Pont-New England Nuclear). Translation products were separated decan (end-labeled s216-44 oligonucleotide; hybridization and wash- by 12% SDS-PAGE (Kojima et al., 1992) and visualized by fluoroging at 50 "C). Independent clones from each screening were plaque- raphy using Enhance (Du Pont-NewEngland Nuclear). purified and excised in viuo to pBluescript SK- plasmids by super Northern Blot Analysis-Total RNA (30 pg) from RASMC, RFC, infection with R408 helper phage. and the RFP-EC line was separated by 1.2% formaldehyde-agarose Characterization of cDNAClones-For each independent cDNA gel electrophoresis (Davis et al., 1986), transferred to a Zeta-Probe clone, the entire length of both strands was enzymatically sequenced membrane (Bio-Rad), and then cross-linked using a UV Stratalinker as outlined above. The majority of the sequence was acquired with 1800 (Stratagene). Membranes were hybridized for 16 h to random nested deletion subclones prepared by using exonuclease III/Sl nu- primed 2.4-kb fragments excised from r-Ryud4 and r-Synd4, respecclease (Erase-a-Base System, Promega, Madison, WI) while synthetic tively, or a 778-bp PstI-XbaI fragment of human glyceraldehyde-3oligonucleotides, based on the preliminary sequence, were employed phosphate dehydrogenase (GAPDH) cDNA (Tso et al., 1985). to obtain the remainder of the structure. The replication error of Measurement of Ryudocan and Syndecan mRNA in Rat Epididymal avian reverse transcriptase used during cDNA construction was elim- Fat Pad Endothelial Cells-The primary RFP-EC were isolated as inated by comparing the sequences of independently isolated clones. outlined above and then expanded in culture for 4 days as previously Analysis of the 5' End of HSPG mRNAs-The primer extension described (Marcum and Rosenberg, 1985; Madri et al., 1983). Cells analyses were performed as previously described' except for the were then incubated for 16 h at 37 "C under standard cell culture conditions with 2 pg/ml of l,l'-dioctadecyl-1-1-3,3,3',3'-tetramee N.W. Shworak, T. O'Connor, C. Po, N. C. W.Wong, and L. thyl-indo-carbocyanine perchlorate-labeled acetylated low density Gedamu, manuscript in preparation lipoprotein (DiI-Ac-LDL, a generous gift fromM. Krieger, MIT,

4872

Cloning and

Expression of Two Distinct Core Proteins

Cambridge, MA) as described by Voyta et al. (1984). Cells were harvested by trypsinization as single cell suspensions and then subjected to fluorescence-activated cell sorting (Parks and Herzenberg, 1984). The uptake of DiI-Ac-LDL by endothelial cells was monitored at 514 nmand cell sorting was carried out on a FACStar Plus (Becton Dickinson, Mountain View,CA) at 2000 cells/s to collect DiI-AcLDL positive cells (0.12 X lo6) and DiI-Ac-LDL negative cells (2.9 X 106). Total RNA was isolated from each cell type, and competitive PCR was employed to determine the levels of ryudocan mRNA as well as syndecan mRNA (Gilliland et al., 1990). The primers

dGAAGACGCTGGGGGCCTTGAGanddTCTGAGGGGACACGGATGCCA were used to generate a 167-bp PCR fragment (position 166-332)from ryudocan cDNA and a 462-bp fragment from the competitor plasmid which bore the amplified region containing a295bp insertion. The primers dCAACAGGGTATGGACTATCTGTand dCTCACACAGGCTCTTCCAATGwere used to generate a 158-bp PCR fragment (position 1895-2052) from syndecan cDNA and a 106bp fragment from the competitor plasmid which bore the amplified region containing a 52-bp deletion. PCR conditions were as outlined for cloning except for total volume (50 pl), annealing temperature (60 "C), extension time (30 s), and cycle number (40). Each reaction mixture contained first strand cDNA, derived from 10 ng of total RNA, and increasing amounts of competitor cDNA. Analyses and Comparisons of cDNA Seqwnces-Sequence analyses were performed with the University of Wisconsin Genetics Computer Group sequence analysis software package. Sequence comparison searches were performed on the GenBank, GNEW, EMBL, and NBRF databases. RESULTS ANDDISCUSSION

Peptide Sequencing-The information necessary for the molecular cloning of HSPG'""'' was obtained by sequencing peptides from the 50-, 30-, and 25-kDa core proteins as well as the amino terminus of the 25-kDa core protein. These studies established the structures of nine peptides with two overlapping sets of tryptic and endopeptidase Glu-C peptides (V46-2 with T27-2,V41 with T12) (see Fig. 3). The two overlapping sets of peptides probably originate from the 50kDa core protein since theyare extremely homologous to mouse syndecan, a previously described heparan sulfate core protein which displays the same apparent size on SDS-PAGE (Weitzhandler et al., 1988). However, until completion of the structure of the corresponding rat cDNA (see below), it remained possible that the encoded core protein was homologous to syndecan only in the restricted region from which the overlapping sets of peptides were derived. We presume that the remaining five peptides are mainly derived from the 25kDa component which is designated as ryudocan (the maintenance of blood fluidity, ryiidousei for fluidity in Japanese, is probably a major function of this proteoglycan) since they lack homology with any known heparan sulfate core protein. The 30-kDa core protein which is presentin only small amounts appears to represent native ryudocan that has not been proteolytically cleaved (see below). It is also noteworthy that theT27-1 peptide overlaps with sequence from the amino terminus of ryudocan. Unfortunately, the amounts of available HSPG"" were insufficient to carry out similar studies. Generation of Probes for Library Screening-To obtain a rat ryudocan-specific probe, we synthesized degenerate sense and antisense primers based on the sequences of T25 and V46-1 (Table I). PCR was performed with the two sets of primers on an RFP-EC cDNA template, and the expected products of 74 and 68 bp, respectively, as well asother nonspecific bands were visualized by agarose gel electrophoresis with ethidium bromide staining. The two desired fragments were purified from the gel, cloned into pUC19, and then confirmed by nucleotide sequencing. Unfortunately, the two probes obtained, as described above, lacked sufficient specificity for library screening. Therefore, a longer probe was

produced by performing PCR with various combinations of the two sets of sense and antisense primers on the same cDNA template. A combination of the T25 sense primer and the V46-1 antisense primergenerated a 215-bp product which encoded not only the T25 and V46-1 peptides but also the T35 peptide. The ryudocan-specific probe is designated as TV215. A rat syndecan-specific probe was produced by synthesizing degenerate sense and antisense PCR primers based on the sequence of the V41/T12 peptide (Table I). PCR was performed on the RFP-EC cDNA template, and the expected 92-bp product was isolated, cloned, and confirmed as outlined above. The syndecan-specific oligonucleotide probe is designated as S216-44. Isolation and Characterization of Ryudocan and Syndecan cDNAs-We constructed a RFP-EC cDNA library and employed the TV215 and S216-44 probes to screen about lo6 plaques. This approach allowed us to isolate five independent overlapping ryudocan cDNAs and six independent overlapping syndecan cDNAs. Sequencing of the various cDNA clones produced identical structures except for relatively rare errors in the function of reverse transcriptase (Preston et al., 1988; Bebenek et al., 1989). Fig. 1 shows the structures of the two sets of cDNA clones and outlines the sequencing strategies. Figs. 2 and 3 depict the primary structures of the rat ryudocan and syndecan cDNAs, respectively. The ryudocan and syndecan composite cDNAs span 2.4 kb and contain 3' poly(A) tails. The sizes of the full-length ryudocan and syndecan mRNAs A

r-Rymdl r-Ryudl r-llyud4 r.Ryud5 r-Ryudll

B

r-Symdl I

r-Symdl

I

r-Symd3 r-Synd4

I

I

I

r.SyadS r-S~md6

t

I

I

FIG. 1. Restriction mapandsequencingstrategyforrat ryudocan and syndecan cDNA clones. The cross-hatched box in the restriction map of each cDNA indicates a coding region for rat ryudocan (A) or rat syndecan ( B ) core protein, respectively. The arrows summarize the length and direction of sequencing of various portions of the cDNAs. The size and orientationof the various cDNA clones are provided next to a restriction map of the cDNAs.

4873

Cloning and Expression of Two Distinct Core Proteins 1

TGTGCTGTTGGAACCATGGCGCCTGTCTGTCTGCCTGTTTTGCGCCGCTGCCTGTTGCTGTTGCTCCTCGGAGGTTTCCCCGTCGCCCCAGGCWLGTCG M

91

P

V

C

L

F

R

E

T

E

core protsln

underlines indicate segments of the predicted protein structure which match sequence information obtained by aminoterminal analyses of the 25-kDa core protein and tryptic or endopeptidase Glu-C-digestedfragments isolated by reverse-phase HPLC (T16, T25, T27-1, T35, and V46-1). The arroul (t)points to the predicted cleavage site of the signal peptide. The double underline shows the putative transmembrane domain. The potential glycosaminoglycanattachment sites are indicated by the open diaThe polyadenylation signal monds (0). sequence AATAAA is dot-underlined.

L

L

V

T

D

P

O

~

L

L

E

T21-1

E

O

D

S

D

F

E

L

S

G

S

0 271

P

L

L

L

L

G

G

F

P

V

A

P

G

E L 25kd

t

25

G R Y T25 0

F

S

G

A

L

P

~

~

E

~ 55A

G

D

L

D

D

T

E

E

P

R

0

I

P

P

E

V

T

S

P

85

TTGGTGCCACTAWLTAACCCATCCCCGAG~TGCCCA~CT~~~~TCCGTGTCCCCTCAWL~CCAAGGAACT~~~AAWLGAATWLGGTC V P L O N H I P E N A O P G I R V P S ~ P K F . l . E F , N E V 115 V46-1

451

541

T16

ATTCCCAAAAGGGTCCCCTCCGACGTGGGGWLTGACWLTGTGTCCAACAAAGTGTCCATGTC~GCACTTCCCAGGGCAG~CATTTTT I P K R V P S D V G D D D V S N K V S M S S T S Q G S N I F

145

GAAAGAACTGAGGTCTTGGCAGCTCTGATTGTGTGGGCGGCGTAGTGGGCATCCTCTTCGCCGTTTTCCT~TCCTGCTGCTGGTGTACCGC E R T E V L A A L I V G G V V G I L F A V F L I L L L V Y R

175

ATGAAGAAGAAGWLTGAAGGULGTTACGACTTACGACTTGGGCAAG~CCCATCTAC~GCCCCCACCAACGAGTTCTACGCATGAAGCTTC M

631 721 811 901 991 1081 1171 1261 1351 1441 1531 1621 1111 1801 1891 1981 2071 2161 2251 2341 2431

G

T35

L

361

G

CTTGAGCAGGACTCTGACTTTGAGCTGTCGGGTTCCGGAGATCTAGATWLCACGWLGGAGCCCAGGACCTTCCCTGAGGTWLTTTCACCC L

FIG. 2. Sequence of rat ryudocan cDNA clones and corresponding coreprotein sequence. The single

A

ATTCWLWLWLCTWLGGTCATAGACCCCCAGWLCCTCCTGGAAGGCAWLTACTTCTCTGWLGCCCTCCCGGACWLTGAAWLCGCTGGffiGC J

181

A

K

K

K

D

E

G

S

Y

D

L

G

K

K

P

I

Y

K

K

A

P

T

N

E

F

Y

A

S

t

o

p

202

TTCCCATGAGTGCTGCTTGGACTTCATGGGWLWLGGAGTTWLGGATTGTGGA~GTffiACATTGGCCTGTAGAGAGAGGGCACCTTAATACTWL C T T G T A T C T C C A T C T C T G G T C A C C T T T C T G G T G T C A G R A G

WLWLWLTGGGATGGGGTGCGWLGGWLGTGTCCGTGTGTGTGT~GTGTGTGTGTGTGTGTGTGTGTGTGTGTGTCTGTCTGTCTGTCTWL GCTGGAACCCGCTGCACCTTWLWLTGTGTTCACCCGAGTACTTCCTCACACTACAGGGTCTCTGTGGTGTATCTCGGGGCATTCTAGGCT CAGTWLCTTTTGAAATTCAACCTTTTTTTTTTTTTTTTAAATCCAGGGAGGGTGGGACTGAAGTGCTGACAGCTCATGCTGAAGTACACT TGTAGAAGATTTGTAAAATGTAAGGTTTTTTTTTTTTTTTTTTTAATGGTCCATTCCTTCATGGGAGCGTGTGCCCTGG~TGAGAGCGTG GGWLTGCACAWLTGTTCTTTCTAGAACATATTCGTT~AACAGCTAACTTTGTGTTTTCATGGTTTTTTATGTTTTGTTTTGTTTTTTTG AAAATGAGAGAAWLGCTGWLWLWLTGATTTTTATWLTTTTTTTTTGTTTTGTTTTTTACTAT~ATAGCTTCAGACGGGGCTGCTTTTCT CTACCTTTCTGTCTTTACTGTTTCCCACTATTTTTTTTTTTTTAATGTTCTGT~TCTTGTTTTTWLCCCTffiCCCTTTCTGAAGTTGCTT T A T C T T A A A A A G T A G C T A C A G T G T T C T A G C A G A T T C C A G A

CCTTCCTCffiTTCTAGGAGAATAGATRAATATATTTTTTTAGWLTATAGAATWLTACTACAGGTCTCA~TTffiCTGGGTGGCTGCGTWL G T G A G T T T T C G T G C G G C T G A G T A A G C T G T T C T A G T G A C T G A G G G

~AlrrTOArTCTAAlrrTrACT~T~CCTTCTGCC~~r~GCTCTGCCCAGGAGCCTCTGGGTTTGCTTTCTCCAGGCCTGTTCTCTAWLTGCACGAT

were ascertained by determining the transcriptional start sites of both messages via primer extension using poly(A)+ RNA. For the ryudocan transcript, a 194-bp extension product was obtained which maps the initiation site to 8 bp upstream of the 5’ limit of the cDNA (Fig. 4). For the syndecan transcript, two primers wereused and each generated two extension products that terminate 28 and 59 bp upstream of the cDNA 5’ limit. We presume that the smaller extension product is caused by premature termination although it is possible that two unique start sites exist. These data indicate that ryudocan and syndecan composite cDNAs are nearly full-length. The ryudocan cDNA of 2452 bp exhibits its first ATG at position 16 and is followed by an open reading frame of 606 bp. The 3’-nontranslated region of 1830 bp possesses a polyadenylation signal (AATAAA) which is 20 bp upstream of the poly(A) tail (Fig. 2). The syndecan cDNA of 2396 bp exhibits its first ATG at position 224 and is followedby an open reading frame of 939 bp. The 3’-nontranslated region of 1233 bp contains aAATAAA signal which is 29 bp upstream of the poly(A) tail (Fig. 3). The two cDNAs are similar in that the region flanking the first ATG conforms to the consensus initiation sequence (Kozak, 1989) and theopen reading frame terminates with TGA. The two cDNAs are different in that the 5’ and coding regions of syndecan are longer than ryudocan while the 3‘ region of syndecan is shorter than ryudocan. Predicted Protein Structures of Ryudocan and SyndecanThe rat ryudocan cDNA encodes a novel 202 amino acid protein with a predicted molecular mass of 21,948 daltons. The deduced primary structure contains regions which correspond to five sequenced peptides (T16, T25, T27-1, T35, and V46-1) as well as theNH,-terminal of the 25-kDa HSPG core protein (Fig. 2). Ryudocan is a previously unidentified protein as indicated by extensive data bank searching. The rat syndecan cDNA encodes a known 313 amino acid protein with a predicted molecular mass of 33,168 daltons. The de-

duced primary structure contains regions which correspond to four sequenced peptides (T12, T27-2, V41, and V46-2, Fig. 3). Rat syndecan as compared to its mouse, hamster, and humancounterpartsis 90.4,86.6, and 76.7%homologous, respectively. Ryudocan and syndecan exhibit the four classical domains of a type I integral membrane protein (Figs. 2 and 3). The NH2 terminus ishydrophobic and presumably functions as a signal peptide for directing the two core proteins to the cell surface. The first domain terminates in a potential signal peptidase cleavage site which occurs between residues 22 and 23 for syndecan and between residues 23 and 24 for ryudocan as predicted by the method of von Heijne (1986). However, NH2-terminal analysis of ryudocan suggests that the actual cleavage site may lie between residues 24 and 25. The second domain is the most divergent region of the two core proteins and spans 126 residues in ryudocan and 232 residues in syndecan. This acidic extracellular domain contains three putative GAG attachment sites for ryudocan and five putative GAG attachment sites for syndecan. Ryudocan and syndecan exhibit a single SGSG consensus GAG attachment sitesat positions 65 and 45, respectively. The remaining ryudocan GAG attachment site at position 44 possesses the atypical sequence YFSGA, whereas the remaining syndecan GAG attachment sites at positions 37, 209, and 219 conform to the (E/D)GSG(E/D) consensus sequence which occurs in other proteoglycans (Zimmermann and Ruoslahti, 1989).3It is possible that the unusual ryudocan GAG attachment site serves to specify or limit the structure of the linked complex Peptide sequence through position 44 of ryudocan exhibited no serine residue whereas peptide sequence through position 219 of syndecan exhibited a serine residue at relatively high levels. These data suggest that position 44 of ryudocan represents a glycosylation site but that position 219 of syndecan may not be substituted with a GAG chain.

4874

Cloning and Expression of Two Distinct Core Proteins 1 91 181

GCGGGCTAGAGGAGACAGAGCCTGGCGCTGAGGAAG~TCTGGCAGTCGAGAGCTWLCTCCAGCCWLCG~CCTGCAGCCCTTGCTCC GGACAGCAGCGCGCCGGGCAGGAGTCCGGGATAGCGAAGCGCAGAG~TCAGCAWLGCCTGCCCGWLGCTCCGTG~CCGGCAACTCC GACCCACATAGCCCGCCGAGCTCCCGCCGCCGGTCCGGGCAGCATGAGACGTGCGGCGCTCTGGCTTTGGCT~GCGCGCTGGCGCTGCG M

R

R

A

A

L

L

W

L

C

A

L

A

L

R

16

211

CCTGCAGCCTGCCCTCCCGClWLTTGTCACCGCAAATGTCACCG~TG~CCTCCTGAAGACCAAGATGGCTCTGGGGACGA~CAGACAACTTCTCTGG L Q P A L P Q I V T A N V P P E D P D G S G D D S D N F S G 46

361

CTCAGGCACAGGTGCTTTGCCAGATATWLCTTTTGTCACGGCAGACACC~CCACTT~GGATGTGTGGCTCCTGACAGCTACACCCAC 76 S G T G A L P D M T L S R Q T P S T W K D V W L L T A T P T 0

451

AGCTCCAGAACCCACCAGCAGGGATACCWLGGCCACCCTCAC~CTATCCTGCCGGCTGGAGA~GCCTGAGGAGGGAGAGCCCG~GC A P E P T S R D T E A T L T S I L P A G E K P E E G E P V A

f

FIG. 3. Sequence of rat syndecan cDNA clones and corresponding core protein sequence. The single underlines indicate the segments of the predicted protein structure which match sequence information obtained by amino-terminal analyses of tryptic or endopeptidase Glu-C-digested fragments isolated by reverse-phase HPLC (T12, T27-2, V41, and V46-2). The arrow (f) points to the predicted cleavage site of the signal peptide. The double underline shows the putative transmembrane domain. The potential glycosaminoglycan attachment sites and N-glycosylation sites are indicated by the open diamonds (0) and the closed diamonds (e),respectively. The polyadenylation signal sequence AATAAA is dot-underlined.

W

54 1

121

o

V

E

A

E

P

D

F

T

A

R

D

K

E

K

E

A

T

T

R

P

R

E

T

T

Q

L

P

V

T

166

GACCTTGGCTCCCACAGCACCCGGCCAACCTGACCATCAGCCTCULAGTG~GAGGATGGAGGCACTTCTGT~TC~GAGGTTGTGGA T L A P T A P G Q P D H Q P P S V E D G G T S V I K E V V E

196

.

GWLTGRAACTACCAATCAGCTTCCTGCAGGAGAGGGCTCTGGAGAA~GACTTCACCTTT~CATCTGGGGAGAACACAGCTGTGGC O

E

T

T

H

O

L

P

A

G

E

G

S

G

E

Q

D

F

T

F

0 901

136

CCAACAGGCCTCAACAGCAGCCAGAGCCACCACGGCCCAGGCATCTGTCACGTCTCATCCCCAC~GGATGTGCAACCTGGCCTCCACGA Q Q A S T A A R A T T A Q A S V T S H P H G D V Q P G L H E

T21-2 V46-2

811

106

CCACGTUIlAAGCAGAGCCTGACTTCACTGCTCGG~ULAGGA~GGAGGCCACCACCAGGCCTAGGGAGACCACACAGCTCCCAGTCAC

H 631

*

0

E

~

S

G

E

N

T

A

V 226 ~

Val0

TGGCGTCGAGCCTGACCTTCGGAATCAGTCCCCAGTG~TGAAGGAGCCACAGGTGCTTCTCAGGGCCTTTT~ACAG~G~GTGCT G V R P O L R U O S P V D E G ~ T G A S Q G L L D R K E 256 ~ *T12

991

1081

GGGAGGTGTCATTGCTGGAGGCCTGGTGGGCCTCATCTTTGCTGTGTGCCTGGTGGCTTTCATGCTATACCGGATGAAGAAGAAGGACGA G G V I A G G L V G L I F A V C L V A F M L Y R M K K K D E

286

AGGCAGTTACTCCTTGWLGGAGCCCAAACAAGCCAATGGCGGTGCCTACCAGRAACCCACCAAGCAGGAAGAGTTCTACGCCTGATGGGG G

S

Y

S

L

E

E

P

K

Q

A

N

G

G

A

Y

Q

K

P

T

K

Q

E

E

F

Y

A

s

t

o

p

313

1171 1261 1351 1441 1531 1621 1111 1801 1891 1981 2071 2161 2251 2341

carbohydrate chain. Syndecan exhibits two potential N-glycosylation sites and possesses a dibasic sequence prior to the start of the third domain which may serve as a proteolytic cleavage site for release of the extracellular domain from the cell surface (Weitzhandler et al., 1988). In contrast, ryudocan exhibits no potential N-glycosylation sites andpossesses only a single basic residue close tothe cell membrane which suggests that its extracellular domain could not be liberated from the cell surface by proteolytic enzymes with specificity similar to thatinvoked for syndecan. The third domain is comprised of 25 hydrophobic residues for both core proteins and presumably spans the cell membrane. It possesses a single cysteine residue in the case of syndecan, but not ryudocan, which might allow interaction with other membrane proteins. The fourth domain is an intracellular region which contains 28 residues for ryudocan and 34 residues for syndecan. The first 7 residues are highly charged and could act as a stop transfersignal for the transmembrane region (Sabatini et al., 1982).The combined transmembrane/intracellular domains are extremely homologous between rat ryudocan, rat syndecan, and human fibroblast HSPG core protein (Marynen et al., 1989) with conservation at the positions of all 4tyrosine groups (Fig. 5). Theseresidues could serve as sitesof phosphorylation which might modulate the functions of the respective HSPGs. Determination of Ryudocan Core Protein Size-A discrepancy exists in that the cDNA sequences predict that mature (NH, terminus processed) ryudocan and syndecan core pro-

teins should be 19,528 and 30,619 Da, respectively, whereas the apparent molecular masses of the RFP-EC HSPG core proteins were 25-, 30-, and 50-kDa (Kojima et al., 1992). We conclude thatthelatter species is syndecan since it has previously been demonstrated that homologous mature mouse syndecan core migrates on SDS-PAGE with anapparent molecular mass of 50-kDa (Weitzhandler et al., 1988). The 25- and 30-kDa species could be accounted for if ryudocan core protein, like syndecan core protein, migrates aberrantly on SDS-PAGE. To test this hypothesis, full-length 2.7-kb ryudocan mRNA was synthesized in vitro and translatedwith rabbit reticulocyte lysate (as described under “Experimental Procedures”). Cotranslational NH2-terminal processing was accomplished by including caninepancreatic microsomal membranes, which provide signal peptidase activity (Walter and Blobel, 1983). The analysis of the translation products by SDS-PAGE revealed that the unprocessed core protein displayed an apparent molecular mass of 33-kDa while inclusion of the microsomal membranes resulted ina 30-kDa product (Fig. 6). These products must have been encoded by the anticipated coding region as identical in vitro translation results were obtained with a synthetic679-bptranscript which terminates just 7 bp downstream of the coding region. The above data demonstrate that the amino terminus processed ryudocan core protein migrates aberrantly on SDS-PAGE with an apparent molecular mass of 30-kDa. It is of interest to note that core proteinpreparationscontaina variable amount of the 25- and 30-kDa components, with the 25-kDa

Cloning and Expression of Two Distinct Core Proteins

B

A FIG.4. Primer extension analyses of rat ryudocan and syndecan mRNAs. A, poly(A)+ RNA served as template for primer extension reactions using one primer specific for ryudocan mRNA (lane I ) or two different primers for syndecan mRNA ( l a n e 2, primer P2; lane 3, primer PI). Size markers were ‘”P-end-labeled Sau 3A-digested fragments of pGEM3 ( l a n e M).Solid arrows (4) indicate extension products (a, ryudocan; b and c, syndecan with primer P2; d and e,: syndecan with primer Pl). B, schematic representation of experimental design and results. Boxes indicate primers for ryudocan (solid) and syndecan (P2,open; P1, hatched), respectively. Arrows (1)and numbers show the 3’ end positions of these primers in each cDNA. The positions of the ATGs at the translational initiation site of each cDNA are designated. Extension products are indicated as single lines with each primer (box) (a, ryudocan; b and c, syndecan with primer P2; d and e, syndecan with primer Pl). Dotted lines show the distance between the 5‘ ends of the extension products and thatof each cDNA.

M 1 2 3 M bp 809

607

rat Ryudocan

--

-

AlIj

t:

286 341

258

4875

“ 1

r.Ryud4 cDNA

a: 194bp 8bp

-

rat Syndecan 322

48

4

10s

1

AIF

4

r-Synd3 cDNA

Y

b 35Obp

IY

c: 381bp

-

d 76bp

hr: 107bp 91

-

78 75

-

I----{ 59bp 1-4 28bp

250 L D R I C E P E G S Y S L E E P K Q A N G G A Y Q K P T K Q E E F Y A * 313 ::l.ll l:::l.ll:ll::lll I: : : : I I I I I I I I I I I . I : I::. 1.1::. :Ill1 145 FERTE EGSYDLG..KKPI...YKKAPT.NEFYA* 202 I.IlIIIII:I.III:I:III:IIIIIIIIII:IIIIIIIIII :II. I . I I I I .IIII

t-S@ t-Ryud

BumESPGc

336 FKRTE

EGSYDLGE.RKPSSAAYQKAPT.KEFYA* 397

Q QQQ

m ma

m

QQ

m

Q Q ~ QOQQQQQ

m

0 0

OmQQ

FIG.5. Homologous transmembrane and cytoplasmic domains of rat ryudocan, rat syndecan, and human fibroblast HSPG. The underlines indicate the transmembrane domains. Identical amino acids are designated by (I), and like amino acids are designated by (:) or (.) as defined by the University of Wisconsin Genetics Computer Group sequence analysis software. The conserved positions of the three core proteins are The amino acid sequence identity in this region is 55.2% between rat ryudocan and syndecan and indicated by @. 81.1%between rat ryudocan and human fibroblast HSPG core protein. microsomal membranes:

-

k h

+

.I

2oo 925 69 46

14.4

FIG.6. Determination of ryudocan core protein size. In vitro synthesized 2.7-kb ryudocan mRNA wastranslated with rabbit reticulocyte lysate without (-) and with (+) canine pancreatic microsomal membranes, as described under “Experimental Procedures.” About 4000 acid precipitable cpm were analyzed by SDS-PAGE and fluorography. The molecular masses of I4C-methylatedprotein size markers (Amersham Corp.) are displayed. species predominating in large scale fractionations and the 30-kDa species predominating in small scale fractionations whererapidprocessingispossible. The NH2-terminal sequence of the 25-kDa component corresponds to that present immediately after the signal sequence of the ryudocan cDNA.

These results indicate that the 30-kDa component represents native ryudocan whereas the 25-kDa component constitutes a carboxyl terminus cleaved formof ryudocan. Expression of Ryudocan and Syndecan-We determined by Northern blot analyses the extent of ryudocan and syndecan expression in RASMC, RFC, and the RFP-EC line. Fig. 7 shows that the level of syndecan mRNA is somewhat greater than that of ryudocan mRNA in the RFP-EC line whereas the level of ryudocan mRNA is somewhat greater than that of syndecan mRNA in RASMC. The quantitation of bands by Betascope 603 BlotAnalyzer (Betagen, Waltham, MA) after normalizationto GAPDH and subtraction of background reveals that the levels of ryudocanmRNA and syndecan mRNA in the RFP-EC line are approximately equal, whereas the levels of ryudocan mRNA and syndecan mRNA in RASMC are approximately 28 and 13% of that in the RFPEC line, respectively. Ryudocanand syndecan are expressed in both cell types as transcripts of 2.6 kb which indicate that each mature messagepossesses a 200-bppoly(A) tail. An additional syndecan mRNA of 3.4 kb is also visualized which probably results from an alternative downstream polyadenylation signal as noted for hamster syndecan (Kiefer et aL, 1990). We also note that that the levels of ryudocan mRNA and syndecan mRNA in RFC are each approximately 13%of that in the RFP-EC line (data not shown). Previously, syn-

Cloning and Expression of Two Distinct Core Proteins

4876

A

A

M

l

2

3

4

5

6

=8J

M

1 I

QP

.I419 511

. 396 . 214 '

28s-

'

28s-

18s-

b,

18s-

. 7 5 8 65

B

FIG. 7. Northern blot analysis for ryudocan and syndecan mRNAs. Total RNA (30 pg) from RASMC and the RFP-EC line was separated by1.2% formaldehyde-agarose gel electrophoresis, transferred to a Zeta-Probe membrane, and hybridized for 16 h to "'P-labeled 2.4-kb fragments excised from r-Ryud4 ( A ) or r-Synd4 ( B ) or 778-bp fragment of human GAPDH cDNA ( C ) as outlined under "Experimental Procedures." A-C represent exposures of 15,24, and 4 h, respectively.

decan expression was considered to be limited to epithelial tissues (Hayashi et aL, 1987);thus, the detection of syndecan expression in endothelial and smooth muscle cells is significant. We ascertained whether the RFP-EC line uniformly expresses ryudocan and syndecan as well as HS"" and HS'""C'. To this end, RFP-EC were reclonedby limiting dilution, and eight independent subclones were isolated as outlined under "Experimental Procedures." Northern analyses demonstrated that all subclonesexpress ryudocan and syndecan transcripts with the same relative abundance observed in the original cell population (data notshown). The subclones were also labeled for 24 h with 35S, extensively proteolyzed with Pronase and Papain, heparan sulfate chains were isolated by ion exchange chromatography, and thenfractionated with ATIII/concanavalin A-Sepharose to ascertain the relative amounts of HS""' and HSinaCt as outlined in the preceding paper (Kojima et aL, 1992). These data show that the levels of HS""' averaged 9.9 f 2.0% of total HS (S.E., n = 6). Thus, the RFP-EC line is composed of a single clone of cells with regard to expression of ryudocan and syndecan as well as the synthesis ofHS""' and HSinac'. To demonstrate that ryudocan and syndecan are expressed in primary RFP-EC as well as in the RFP-EC line, a population of microvascular endothelial cells wasisolated from rat epididymal fat pads, expanded in culture for 4 days, and then incubated with DiI-Ac-LDL whichis preferentially internalized by endothelial cells(Voyta,1984). The DiI-Ac-LDL positive and negative cells wereisolated by cell sorting, total RNA was obtained from both populations, and the levels of ryudocan as well as syndecan mRNAs were quantitated by competitive PCR as outlined under "Experimental Procedures." This approach involves coamplification of the target cDNA with varying concentrations of a competitor cDNA which requiresthe same primers but generates a PCR product of different size. The results of a typical experiment (Fig. 8) show that coamplification of DiI-Ac-LDL positive cell cDNA or DiI-Ac-LDL negative cell cDNA,each derived from 10 ng of total RNA, with 1.75 x lo6 copies of competitor cDNA generates ryudocan and competitor PCR products of approximately equal intensity. The above data allow us to calculate that ryudocan mRNAconstitutes about 0.5% ofmicrovascular endothelial cell mRNA (DiI-Ac-LDL positive cells) and microvascular non-endothelial cell mRNA (DiI-Ac-LDL negative cells) based upon an assumed 100% efficiency of first strand cDNA synthesis and mRNA levels equivalent to 5% of total RNA. We carried out additional experiments which confirmed the above estimate of abundance for ryudocan mRNA and demonstrated that syndecan mRNA constitutes

M

I

2

3

4

5

6

M

QP

. ,419 .

51 7

396 2'1 7 5 8 65

FIG.8. Competitive PCR for measurement of rat ryudocan mRNA level in the primary rat microvascular cells. The DiIAc-LDL positive and negative cells were isolated by cell sorting, total RNA was obtained from both populations, and thelevels of ryudocan mRNA were quantitated by competitive PCR as outlined under "Experimental Procedures." PCRproducts were analyzed by ethidium bromide/agarose gel electrophoresis (2.3% NuSieve agarose and 0.7% agarose). First strandcDNA derived from 10 ng of total RNA of each population (A,: DiI-Ac-LDL positive cell; B, DiI-Ac-LDL negative cell) Versus various concentration of competitor cDNA (lane 1 , 1.75 X lo7 molecules; lane 2, 8.75 X lo6 molecules; lane 3, 1.75 X lo6 molecules; lane 4, 8.75 X lo5 molecules; lane 5, 1.75 X lo5 molecules) or 10 ng of each total RNA without competitor (lane 6 )was used as template for PCR amplification. 167-bp PCR fragments were generated from ryudocan cDNA and 462-bp fragments from the competitor cDNA, respectively. Molecular size makers (lane M ) are HinfI-digested fragments of pUC19.

about 0.1% of DiI-Ac-LDL positive cell mRNA and 0.3% of DiI-Ac-LDLnegativecellmRNA,respectively (data not shown). These data suggest that ryudocan and syndecan are expressed at relatively high levelsin primary RFP-EC aswell as in smooth muscle cells,fibroblasts, or adipocytes which are associated with epididymal fat pad endothelial cells. The levels of ryudocan and syndecan mRNAs in RFP-EC line were also quantitated by the above methodology which revealed that these species constitute about 0.3 and 0.4% of RFP-EC line mRNA,respectively. Thus, the ratio of the relative levels of ryudocan and syndecan mRNAs are virtually identical to the molar ratio of25 30 kDa to 50-kDa core proteins isolated from the same cell line (3:4 uersus 6:7) (Kojima et al., 1992). In the preceding paper (Kojima et aL, 1992), we demonstrated that HSPG"' and HSPG'""" core proteins exhibit the same molecular sizes and possess remarkably similar sets of trypsin or endopeptidase Glu-C-generated peptides as judged by reverse-phase HPLC. These datasuggest that the primary structures of HSPG"'' and HSPG'"Be' core proteins are identical and that the cellular events which direct syntheses of H p t and HSinaet involve post-translational signals. To examine this issue, we utilized peptide sequence data from the abundant HSPG'""' core proteins to clone the two cDNAs described above.It proved impossibleto obtain similar structural information for the HSPG"' core proteins because of the extremely small amounts available. However,the sequencing of 5-10 independent ryudocan and syndecan cDNA clones failed to uncover variant core protein transcripts which might serve as unique templates for synthesizing HS""' (data not shown). We also note that syndecan and ryudocancDNAs are encoded by single genes rather than by families of related genes with somewhat different nucleotide sequences that might generate variant sets of core proteins with different GAG chain subtypes. On the one hand, chromosomal mapping studies show that murine syndecan is encoded by a single

+

Cloning andExpression of Two Distinct Core Proteins functional gene (Oettinger et aL, 1991). On the other hand, ryudocan is also present as a single copy gene. To show that this is the case, we have carried out Southern blottinganalysis with EcoRI or HindIII digested RFP-EC genomic DNA. An extracellular domain AluI generated probe (positions 205472) detected a single 4.5-kb EcoRI fragment and asingle 10kb HindIII fragment. An intracelluar domain AluI-HindIII generated probe (positions 473-624) detected a single 0.8-kb EcoRI fragment andthe same 10-kb H i d 1 1 fragment (results not shown). Thisfurtherstrengthenstheassertionthat HSPGBCt and HSPG'"BCtcore proteins are identical. Despite the above evidence, it remains possible that minor differences in the primary structures of the two core proteins generated by alternative splicing or post-translational modifications of the two core proteins could beinvolved in directing syntheses of H P ' and HSinact. This hypothesis is currently being tested by stably expressing epitope-tagged ryudocan in rodent cells which synthesize H P t and HS"', and then examining the structuresof the covalently linked GAG chains. In summary, we have cloned two different rat cDNAs that encode heparan sulfate core proteins from a RFP-EC line and have demonstrated expression of both messages in primary rat microvascular endothelial cells and smooth muscle cells. The newly defined core protein ryudocan is likely to play a critical role in maintaining blood fluidity whereas the previously identified core protein syndecan now appears to be involved in an important but unsuspected physiologic function. Acknowledgments-We thank Dr. James A. Marcum for his helpful suggestions and a generous gift of RFC RNA, Dr. Richard F. Cook and Dr. William S. Lane for peptide sequencing and amino acid analyses, Stewart Conner for cell sorting of vascular cells, and Lynne D. Butler for her excellent technical help. REFERENCES Bebenek, K., Abbotts, J., Roberts, J. D., Wilson, S. H., and Kunkel, T. A. (1989)J. Bid. Chem. 264, 16948-16956 Castellot, J. J., Jr., Favreau, L. V., Karnovsky, M. J., and Rosenberg R. D. (1982)J. Biol. Chem. 257, 11256-11260 Chomczynski, P., and Sacchi, N. (1987)Anal. Biochem. 162, 156159 Davis, L. G., Dibner, M. D., and Battey, J. F. (1986)Basic Methods in Molecular Biology, pp. 143-146,Elsevier Science Publishing Co., Inc. New York de Agostini, A. I., Watkins, S. C., Slayter, H. S., Youssoufian, H., and

4877

Rosenberg, R. D. (1990)J. Cell Bwl. 111, 1293-1304 Doi, T., Greenberg, S. M., and Rosenberg, R.D. (1987)Mol.Cell. Bwl. 7,898-904 Gilliland, G., Perrin, S., and Bunn, H. F. (1990)in PCR Protocols: A Guide to Methods and Applications (Innis, M. A., Gelfond, D. H., Sninsky, J. I., and White, T. J., eds) pp. 60-69, Academic Press, New York Hayashi, K., Hayashi, M., Jalkanen, M., Firestone, J. H., Trelstad, R.L., and Bernfeild, M. (1987)Histochem. Cytochem. 35, 10791088 Kiefer, M. C., Stephans, J. C., Crawford, K., Okino, K., and Barr, P. J. (1990)Proc. Natl. Acad. Sci. U. S. A. 87, 6985-6989 Kojima, T., Leone, C.W., Marchildon, G.A., Marcum, J. A., and Rosenberg, R. D. (1992)J. Bwl. Chem. 267,4859-4869 Kozak, M. (1989)J. Cell Biol. 108,229-241 Madri, J. A., and Williams, S. K. (1983)J. Cell Bwl. 97, 153-156 Maniatis, T., Fritsch, E. F., and Sambrook, J. (1982) Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY Marcum, J. A., and Rosenberg, R. D. (1985)Biochem. Biophys. Res. Commun. 126,365-372 Marcum, J. A., Conway, E. M., Youssoufian, H., and Rosenberg, R. D.(1986)Exp. Cell Res. 166,253-258 Matsudaira, P. (1987)J. Biol. Chem. 262,10035-10038 Marynen, P., Zhang, J., Cassiman, J-J., Van den Berghe, H., and David, G. (1989)J. Biol. Chem. 264,7017-7024 Moremen, K. W. (1989)Proc. Natl. Acud. Sci. U. S. A. 86,5276-5280 Oettinger, H. F., Streeter, H., Lose, E., Copeland, N. G., Gilbert, D. J., Justice, M. J., Jenkins, N. A., Mohandas, T., and Bernfeild, M. (1991)Genomics 11,334-338 Parks, D. R., and Herzenberg, L. A. (1984)Methods Enzymol. 108, 197-241 Preston, B. D., Poiesz, B. J., and Loeb, L.A. (1988)Science 242, 1168-1171 Sabatini, D. D., Kreibich, G., Morimoto, T., and Adesnik, M. (1982) J. Cell Bwl. 92,l-22 Sambrook, J., Fritsch, E.F., and Maniatis, T. (1989)Molecular Cloning:A Laboratory Manual, Second Edition, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY Sanger, F., Nicklen, S., and Coulson, A. R. (1977)Proc. Natl. Acud. Sci. U. S. A, 74, 5463-5467 Saunders, S., Jalkanen, M., OFarrell, S., and Bernfield, M. (1989)J. Cell Biol. 108, 1547-1556 Voyta, J. C., Netland, P. A., Via, D. P., and Zetter, B. R. (1984)J. Cell Biol. 99,81a von Heijne, G. (1986)Nucleic Acids Res. 14,4683-4690 Walter, P., and Blobel, G. (1983)Methods Enzymol. 96,84-93 Weitzhandler, M., Streeter, H. B., Henzel, W. J., and Bernfield, M. (1988)J. Biol. Chem. 263, 6949-6952 Tso, J. Y., Sun, X., Kao, T., Reece, K. S., and Wu, R. (1985)Nucleic Acids Res. 13,2485-2502 Zimmermann, D. R., and Ruoslahti, E. (1989)EMBO J. 8, 29752981