Proc. Nati. Acad. Sci. USAVol. 86, pp. 9891-9895, December 1989Cell Biology

MUC18, a marker of tumor progression in human melanoma, showssequence similarity to the neural cell adhesion molecules of theimmunoglobulin superfamilyJURGEN M. LEHMANN*, GERT RIETHMULLER, AND JUDITH P. JOHNSONtInstitute of Immunology, University of Munich, Goethestrasse 31, D-8000 Munich 2, Federal Republic of Germany

Communicated by Hans J. MuIller-Eberhard, August 31, 1989 (received for review July 18, 1989)

ABSTRACT The MUC18 antigen is an integral membraneglycoprotein of 113 kDa whose expression on primary humanmelanomas correlates with poor prognosis and the develop-ment of metastatic disease. MUC18 is expressed only sporad-ically in benign melanocytic nevi and thin primary melanomasthat have a low probability of metastasizing. However, withincreasing tumor thickness, MUC18 expression becomes morefrequent and it is found on 80% of advanced primary tumorsand metastases. MUC18-encoding cDNA clones were obtainedby screening a human melanoma phage A expression librarywith monoclonal antibodies produced against the denaturedantigen. The deduced sequence of 603 amino acids consists ofa signal peptide, five immunoglobulin-like domains, a trans-membrane region, and a short cytoplasmic tail. The highestsequence similarity is with a group of nervous system celladhesion molecules, which includes neural cell adhesion mol-ecule (N-CAM). The close structural relationship with thesemolecules suggests that MUC18 may also be a developmentallyregulated cell adhesion molecule.

Malignant cells are not inherently metastatic but appear todevelop this capacity during tumor progression as theyacquire new characteristics through a multistep process ofselection from continuously arising variants (1). Analysis ofin situ phenotypic changes that correlate with the progressionofcutaneous malignant melanoma has led to the identificationof molecules that may contribute to the metastatic potentialof this tumor (2-5). Two of these molecules are HLA-DR andintracellular adhesion molecule 1, which are not found on thintumors that have a good prognosis but begin to be expressedon primary tumors .1 mm thick and appear to be coordi-nately expressed (6). A third melanoma antigen whoseexpression has been shown to correlate with tumor progres-sion is the 113-kDa cell surface glycoprotein MUC18 (7).Although the expression of MUC18 is independent of HLA-DR and intracellular adhesion molecule 1, it is also notexpressed on benign melanocytic lesions or thin tumors witha good prognosis. Here, we report the cDNA cloning andsequencing of MUC18 from melanoma cells.f The derivedamino acid sequence indicates that MUC18 is a member ofthe immunoglobulin superfamily and shows the greatestsequence similarity to a group of neural cell adhesion mole-cules expressed during organogenesis. On the basis of thissequence similarity, we speculate that MUC18 may also bedevelopmentally regulated and mediate intercellular adhe-sion. Since the expression of MUC18 in normal adult tissuesappears limited to vascular smooth muscle (7), expression ofthis molecule by melanoma cells may contribute to theirinteraction with elements of the vascular system, an essentialstep in the dissemination of tumor cells.

MATERIALS AND METHODSAntibodies and Cells. Monoclonal antibody (mAb) MUC18

(IgG2a) has been described (7) and mAb 96.5 (anti-melanotransferrin, ref. 8) was kindly provided by I. Hell-strom (Oncogen, Seattle), and the myeloma protein UPC10(IgG2a) was purchased from Sigma. Cell lines were obtainedfrom the American Type Culture Collection, through ex-change, or were established in our laboratory.Immunohistochemistry and Flow Cytometry. Indirect

immunoperoxidase staining on frozen tissues was performedas described (7). Tissues were regarded as positive when-5% of the tumor cells were stained. The following numbersof primary melanomas were examined: 17 at <0.75 mm and15 at 0.76-1.5 mm, 15 at 1.51-3.0 mm, and 15 at .3.0 mm.For flow cytometry, cells were incubated with mAb followedby fluorescein isothiocyanate-conjugated second antibodyand analyzed using a FACScan (Becton Dickinson).

Selection of the Antibodies for Western Blot Analysis. mAbsMUCBA18.1-18.5 (all IgGl) were produced from a (C57BL/6x BALB/c)F1 mouse injected i.p. with 100 mg of lipid A (9)followed 14 days later by heat-denatured immunoprecipitatedMUC18 in 0.1% SDS. Three injections of immunoprecipitatesfrom 3 x 109 Mel-JuSo cells with Freund's complete adjuvant,Freund's incomplete adjuvant, and Bordetella pertussis adju-vant, respectively, were spaced over 6 weeks. The spleen cellswere fused with P3x63Ag8.63 myeloma cells and culturesupernatants were screened on MUC18 glycoprotein-loadednitrocellulose strips. Immunoprecipitates from Mel-JuSo cells(5 x 108 cells per cm of nitrocellulose) were separated bySDS/PAGE under reducing conditions and electrophoreti-cally transferred to nitrocellulose membranes (BA85; Schlei-cher & Schull). The nitrocellulose strips were incubated withthe mAb followed by peroxidase-conjugated rabbit antiserumto mouse immunoglobulins (Dakopatts, Copenhagen) as de-scribed (4).

Isoelectric Focusing and Peptide Mapping. Cells were la-beled using lactoperoxidase-glucose oxidase-catalyzed sur-face iodination (10) and immunoprecipitates were prepared asdescribed (7). Isoelectric focusing was performed essentiallyas described by Neefjes et al. (11). Peptide mapping by partialproteolysis in gel slices (12) was performed using 100 ng ofStaphylococcus aureus V8 protease (Sigma).cDNA Libraries and DNA Sequencing. cDNA expression

libraries were constructed in the Agtll and AZAP vectors(Strategene). cDNA was prepared from Mel-JuSo cells(cDNA synthesis system; BRL) and size-selected [-1200base pairs (bp)] by agarose gel electrophoresis. Fusion pro-teins were induced, blotted on nitrocellulose filters, and

Abbreviation: mAb, monoclonal antibody.*Present address: La Jolla Cancer Research Center, 10901 NorthTorrey Pines Road, La Jolla, CA 92037.tTo whom reprint requests should be addressed.MThe sequence reported in this paper has been deposited in theGenBank data base (accession no. M28882).

9891

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

Nov

embe

r 23

, 202

0

9892 Cell Biology: Lehmann et al.

80 -

60-

0

0)a)40 -

20 -

A

0

.0

E

a)<0.75 0.76-1.5 1.51-3.0 >3.0

Tumor thickness, mm

FIG. 1. Reaction pattern ofmAb MUC18 with primary melanomain situ. Open bars, percentage of lesions that contained -5% stainedtumor cells; solid bars, 5-year mortality rate in each group (17).

screened with mAbs MUCBA18.1-18.5 (13). Reactiveplaques were detected using peroxidase-coupled second an-

tibody. For sequencing, suitable restriction fragments weresubcloned in pUC18 and double-stranded cDNA was se-

quenced using the dideoxynucleotide chain-terminationmethod (14) with Sequenase (United States Biochemical).Northern and Southern Blot Analysis. RNA was denatured

in formaldehyde, separated on a 1% agarose/formaldehydegel (15), and transferred to Hybond-N (Amersham). The1200-bp Sst I-Pst I cDNA fragment labeled with [a-32P]dCTP(Amersham) by random-priming was used as probe. Mem-branes were hybridized overnight at 650C in 6x SSC/5xDenhardt's solution/0.5% SDS/salmon sperm DNA (20 pzg/ml). (1 x SSC = 0.15M NaCl/0.015 M sodium citrate, pH 7.0;lx Denhardt's solution = 0.02% polyvinylpyrrolidone/0.02% Ficoll/0.02% bovine serum albumin.) The filters werewashed at 650C in 3x SSC/0.1% SDS, lx SSC/0.1% SDS,0.3x SSC/0.1% SDS, and 0.lx SSC/0.1% SDS. DNA was

digested with restriction endonucleases, electrophoresed,

A B1 2 3 4 5 6

71

61

.. IIS IeiF IAm MeUuSo

SKMeI 1

.,AA375

if100 10' 102 4 100 10'

Relative fluorescence

B

4.4-3.3-

2.4-

1 2 3 4 5 6

i.l

C

1 2 1 2 3

L 92.5-

68-.

46

-1

FIG. 2. Biochemical comparison of the molecules recognized bymAbs MUC18 and MUCBA18.1-18.5. (A) Immunoblots. The blottedMUC18 immunoprecipitates were probed with mAbs MUCBA18.1(lane 1), MUCBA18.2 (lane 2), MUCBA18.3 (lane 3), MUCBA18.4(lane 4), MUCBA18.5 (lane 5), and UPC10 (lane 6). (B) Isoelectricfocusing banding pattern of the MUC18 immunoprecipitate (lane 1)and the MUCBA18.1 immunoprecipitate (lane 2). Immunoprecipitateswere prepared from Mel-JuSo cells labeled by surface iodination. ThepH gradient is indicated on the left. (C) Peptide mapping. MUC18immunoprecipitate (lane 1), the MUCBA18.1 immunoprecipitate (lane2), and, as control, the melanotransferrin immunoprecipitate (p97,lane 3) were partially digested with 100 ng of Staphylococcus aureus

V8 protease. Immunoprecipitates were prepared from Mel-JuSo cellslabeled by surface iodination. The locations of molecular mass stan-dards are indicated on the left in kDa.

FIG. 3. Flow cytometric analysis of the MUC18 glycoproteinexpression and Northern blot analysis. (A) Immunofluorescencestaining of the melanoma cell lines Mel-JuSo, SK-Mel-28, SK-Mel-1,SK-Mel-25, A375, and B16F1 by mAb MUC18 (dashed lines) andisotype control UPC10 (dotted lines). (B) Total RNA (20 pug) from themelanoma cell lines Mel-JuSo (lane 1), SK-Mel-18 (lane 2), SK-Mel-1(lane 3), SK-Mel-25 (lane 4), A375 (lane 5), and B16F1 (lane 6) waselectrophoresed, blotted on nylon membranes, and hybridized withthe 32P-labeled MUC18 cDNA probe.

transferred to Hybond-N nylon membranes, and hybridizedas described above.

RESULTS AND DISCUSSIONCorrelation of MUC18 Expression with the Development of

Human Metastatic Melanoma. mAb MUC18 produced againsthuman malignant melanoma was selected because it does notreact with benign melanocytic lesions. In contrast 45% of the62 primary melanomas and 71% of the 31 metastatic mela-nomas examined expressed the MUC18 antigen (7). Thevertical thickness of primary melanomas (the Breslow index)is directly correlated with the probability of metastatic dis-ease (16). Division of the primary melanomas into four evenly

| ~~~~SKMel28

1,.*_,l: e

6..a . 0a

Proc. Natl. Acad. Sci. USA 86 (1989)

113--l.i 0 0 a m

Dow

nloa

ded

by g

uest

on

Nov

embe

r 23

, 202

0

Proc. Natl. Acad. Sci. USA 86 (1989) 9893

T TK PSI I 11

TP PsT P PPs

IllTPT

I

o 500 1000 15a00 2000 2500 3000 3

CDNA probe

3'

zapy FIG. 4. Structure of the MUC18 glycoprotein- drop mRNA and relationship between the cDNA clones

(A), sequencing strategy (B), and localization ofdrop4 1300-bp Sst I-Pst I MUC18 cDNA probe used fordrop 4.2 Northern and Southern blot analysis (C). Threedrop4.7 cDNA clones, zapyl, dropi, and drop4, were iso-

lated by using the antibodies used in Western blotdro°p 4.11 analysis as probes and the clones drop4.2, drop4.7,

and drop4.11 were isolated by using a 32P-labeled1300 MUC18 cDNA probe derived from the 5' end of- clone drop4. The divergent nucleotide sequence of

drop4.7 in the 3' region is indicated by dashed line.The 1809-bp open reading frame is indicated by theopen box. Various restriction sites were used togenerate subclones for sequencing. P, Pvu 11; T,Taq 1; Ps, Pst I; K, Kpn I; S, Sst 1.

matched groups according to their vertical thickness revealsthat expression of MUC18 shows a linear correlation withthis parameter (Fig. 1). The association ofMUC18 expressionwith tumor thickness suggests that the appearance of thismolecule reflects changes in the tumor cells that may con-tribute to the development of metastatic capacity.

Characterization of the Antibodies for Western Blot Anal-ysis. To isolate MUC18-encoding cDNA clones, five mAbs(MUCBA18.1-18.5) reactive with the denatured MUC18glycoprotein were produced (Fig. 2A). These mAbs precip-itated a single protein of 113 kDa from surface-labeledMel-JuSo cells (data not shown). In addition, MUC18 andMUCBA18.1 immunoprecipitates had identical isoelectricpoints and the same pattern of microheterogeneity (Fig. 2B)as well as identical peptide maps (Fig. 2C), indicating thatboth antibodies recognize the same molecule.

Isolation of MUC18 cDNA Clones and Identification ofMUC18 mRNA. Three immunopositive cross-hybridizingclones were isolated from two human melanoma cDNA A

expression libraries. In addition to reactivity with mAbs

A1 2 3 4 5

B1 2 34 5

-5-

I

V

2/

FIG. 5. Southern blot analysis. (A) Genomic DNA from Mel-JuSocells was digested with Taq I, Pvu II, Pst I, EcoRl, and Bgl II (lanes1-5, respectively). (B) Genomic DNA from peripheral blood lym-phocytes of a normal donor was digested with EcoRI (lane 1).Genomic DNA from Mel-Ho cells (lanes 2 and 4) and the autologousB-lymphoblastoid cell line (lanes 3 and 5) were digested with Pvu II

(lanes 2 and 3) or Taq I (lanes 4 and 5) and hybridized to the32P-labeled MUC18 cDNA probe.

MUCBA18.1, -18.2, -18.3, and -18.5, the 3.3-kilobase mRNAdetected by these clones demonstrates an expression patternin human melanoma cell lines that mirrors exactly the bindingof the MUC18 mAb (Fig. 3). High levels of MUC18 antigenand the 3.3-kilobase mRNA were seen in Mel-JuSo andSk-Mel-28 cells, whereas lower levels were seen in Sk-Mel-1cells and no products were detected in Sk-Mel-25 and A375cells. The detection of a 3.3-kilobase mRNA in the mousemelanoma cell line B16F1 suggests that the MUC18 gene maybe conserved across species, although the particular epitopesdefined by the MUC18 antibodies are absent.

Inspection of the cDNA sequence revealed that it was

incomplete in the 5' region. Therefore, the melanoma librarywas screened with a cDNA probe derived from the 5' end ofclone drop4. Three additional clones were identified and thesix overlapping clones correspond to the size of the observedmRNA. The relationship of the various cDNA clones isschematically depicted in Fig. 4A.

Analysis of Mel-JuSo DNA digested with various restric-tion endonucleases showed a pattern consistent with a single-copy gene, as only a single hybridization fragment wasobtained (Fig. 5A). Comparison of the restriction patterns inmelanoma and autologous B-lymphoblastoid cell lines fromdifferent patients provides no indication that the gene isrearranged in melanomas or polymorphic in the population(Fig. SB).

Nucleotide and Amino Acid Sequence. The MUC18 cDNAsequence (Fig. 6A) reveals a 7-bp 5' untranslated region, a

single open reading frame of 1809 bp, and a 1487-bp 3'untranslated region, which includes the consensus polyade-nylylation signal (AATAAA) at position +3264 followed 15bases later by a poly(A) tail.

In contrast to this sequence, the cDNA clone drop4.7contains an unrelated 92-bp 3' untranslated region that beginsat position +2847 (Fig. 6B) and lacks the consensus polya-denylylation signal. If not a cloning artifact this cDNA clonemay indicate that the expression of the MUC18 molecule isregulated by the production of mRNAs with different 3'untranslated regions (18).The translation initiation site was assigned to the first ATG

codon of the single open reading frame at position + 1, whichis surrounded by a nucleotide sequence corresponding to theconsensus sequence proposed for putative initiation codons(GGAAGCATGG) (19,20). Hydrophobicity analysis (21) sug-gests that this sequence is succeeded by a 28-residue signalpeptide sequence, which may be cleaved between alanine-28and glutamic acid-29 (22). The open reading frame ends withthe termination codon TAA at position + 1810. Thus, themature MUC18 molecule appears to consist of 575 amino acid

A

s.

B

C

Cell Biology: Lehmann et al.

Dow

nloa

ded

by g

uest

on

Nov

embe

r 23

, 202

0

9894 Cell Biology: Lehmann et al. Proc. Natl. Acad. Sci. USA 86 (1989)

A GCT-7 GGGAAGC ATG GGG CTT CCC AGG CTG GTC TGC GCC TTC TTG CTC GCC GCC TGC TGC TGC TGT CCT CGC GTC GCG GGT GTG CCC GGA GAG GT

Met GlY Leu Pro Ara Leu Val CYs Ala Phe Lou Leu Ala Ala Cys CYS CYS CY. Pro Ara Val Ala GlY Val Pro Gly Glu Ala 28

85 GAG CAG CCT GCG CCT GAG CTG GTG GAG GTG GAA GTG GGC AGC ACA GCC CTT CTG AAG TGC GGC CTC TCC CAG TCC CM GGC MC CTC AGCGlu Gln Pro Ala Pro Glu Lou Val Glu Val Glu Val Gly Ser Thr Ala Lou Lou Lys Cys Gly Leu Ser Gln Ser Gln Gly Asn Leu Ser 58

175 CAT GTC GAC TGG TTT TCT GTC CAC AAG GAG AAG CGG ACG TCA TCT TCC GTG TGC GCC AGG GCC AGG GCC AGA GCG MC CTG GGG AGT ACGHis Val Asp Trp Phe Ser Val His Lys Glu Lys Ar8 Thr Ser Ser Ser Val Cys Ala Arg Ala Arg Ala Ar8 Ala Asn Lou Gly Ser Thr 88

265 AGC MG CGG CTC AGC CTC CAG GAC AGA GGG GCT ACT CTG GCC CTG ACT CAA GTC ACC CCC CM GAC GAG CGC ATC TTC TTG TGC CAG GGCSer Lys Arg Leu Ser Lou Gln Asp Ar8 Gly Ala Thr Leu Ala Lou Thr Gln Val Thr Pro Gln Asp Glu Are Ile Phe Leu Cys Gln Gly 118

355 MG CGC CTC GGT CCC AGG AGT ACC GCA TCC AGC TCC GCG TCT ACA MG CTC CGG ATG CCA MC ATC CAG GTC AAC CCC CTG GGC ATC CCTLys Are Lou Gly Pro Are Ser Thr Ala Ser Ser Ser Ala Ser Thr Lys Leu Are Met Pro Asn Ile Gln Val Asn Pro Lou Gly Ile Pro 148

445 GTG MC AGT MG GAG CCT GAG GAG GTC GCT ACC TGT GTA GGG AGG MC GGG TAC CCC ATT CCT CM GTC ATC TGG TAC MG MT GGC CGGVal Asn Ser Lys Glu Pro Glu Glu Val Ala Thr Cys Val Gly Are Asn Gly Tyr Pro Ile Pro Gln Val Ile Trp Tyr Lys Asn Gly Arg 178

535 CCT CTG MG GAG GAG MG MC CGG GTC CAC ATT CAG TCG TCC CAG ACT GTG GAG TCG AGT GGT TTG TAC ACC TTG CAG AGT ATT CTG MGPro Lou Lys Glu Glu Lys Asn Are Val His Ile Gln Ser Ser Gln Thr Val Glu Ser Ser Gly Lou Tyr Thr Lou Gln Ser Ile Lou Lys 208

625 GCA CAG CTG GTT MA GM GAC MA GAT GCC CAG TTT TAC TGT GAG CTC MC TAC CGG CTG CCC AGT GGG MC CAC ATG MG GAG TCC AGGAla Gln Lou Val Lys Glu Asp Lys Asp Ala Gln Phe Tyr Cys Glu Lou Asn Tyr Are Leu Pro Ser Gly Asn His Met Lys Glu Ser Are 238

715 GM GTC ACC GTC CCT GTT TTC TAC CCG ACA GM AAA GTG TGG CTG GM GTG GAG CCC GTG GGA ATG CTG MG GM GGG GAC CGC GTG GAAGlu Val Thr Val Pro Val Phe Tyr Pro Thr Glu Lys Val Trp Leu Glu Val Glu Pro Val Gly Met Lou Lys Glu Gly Asp Are Val Glu 268

805 ATC AGG TGT TTG OCT GAT GGC AAC CCT CCA CCA CAC TTC AGC ATC AGC MG CAG MC CCC AGC ACC AGG GAG GCA GAG GM GAG ACA ACCIle Are Cys Lou Ala Asp Gly Asn Pro Pro Pro His Phe Ser Ile Ser Lys Gln AIn Pro Ser Thr Are Glu Ala Glu Glu Glu Thr Thr 298

895 MC GAC MC GGG GTC CTG GTG CTG GAG CCT GCC CGG MG GM CAC AGT GGG CGC TAT GM TGT CAG GCC TGG MC TTG GAC ACC ATG ATAAsn Asp Asn Gly Val Lou Val Lou Glu Pro Ala Are Lys Glu His Ser Gly Are Tyr Glu Cys Gln Ala Trp Asn Lou Asp Thr Met Ile 328

985 TCG CTG CTG AGT GM CCA CAG GM CTA CTG GTG MC TAT GTG TCT GAC GTC CGA GTG AGT CCC GCA GCC CCT GAG AGA CAG GM GGC AGCSer Lou Lou Ser Glu Pro Gln Glu Lou Leu Val Asn Tyr Val Ser Asp Val Are Val Ser Pro Ala Ala Pro Glu Are Gln Glu Gly Ser 358

1075 AGC CTC ACC CTG ACC TGT GAG GCA GAG AGT AGC CAG GAC CTC GAG TTC CAG TGG CTG AGA GM GAG ACA GAC CAG GTG CTG GM AGG GGGSer Lou Thr Leu Thr Cys Glu Ala Glu Ser Ser Gln Asp Lou Glu Phe Gln Trp Leu Are Glu Glu Thr Asp Gln Val Leu Glu Ar8 Gly 388

1165 CCT GTG CTT CAG TTG CAT GAC CTG AM CGG GAG GCA GGA GGC GGC TAT CGC TGC GTG GCG TCT GTG CCC AGC ATA CCC GGC CTG MC CGCPro Val Lou Gln Lou His Asp Lou Lys Are Glu Ala Gly Gly Gly Tyr Ar8 Cys Val Ala Ser Val Pro Ser Ile Pro Gly Lou Ajg Arg 418

1255 ACA CAG CTG GTC MG CTG GCC ATT TTT GGC CCC CCT TGG ATG GCA TTC MG GAG AGG MG GTG TGG GTG MA GAG MT ATG GTG TTG MTThr Gln Leu Val Lys Leu Ala Ile Phe Gly Pro Pro Trp Met Ala Phe Lys Glu Ar8 Lys Val Trp Val Lys Glu Asn Met Val Lou Axi 448

1345 CTG TCT TGT GM GCG TCA GGG CAC CCC CGG CCC ACC ATC TCC TGG MC GTC MC GGC ACG GCA AGT GAA CM GAC CM GAT CCA CAG CGALou Ser Cys Glu Ala Ser Gly His Pro Are Pro Thr Ile Ser Trp Asn Val Aim Gly Thr Ala Ser Glu Gln Asp Gln Asp Pro Gln Are 478

1435 GTC CTG AGC ACC CTG MT GTC CTC GTG ACC CCG GAG CTG TTG GAG ACA GGT GTT GM TGC ACG WCC TCC MC GAC CTG GGC AM MC ACCVal Lou Ser Thr Leu Asn Val Leu Val Thr Pro Glu Lou Leu Glu Thr Gly Val Glu Cys Thr Ala Ser Asn Asp Lou Gly Lys Ajn Thr 508

1525 AGC ATC CTC TTC CTG GAG CTG GTC MT TTA ACC ACC CTC ACA CCA GAC TCC MC ACA ACC ACT GGC CTC AGC ACT TCC ACT WCC AGT CCTSer Ile Lou Phe Lou Glu Lou Val An Lou Thr Thr Lou Thr Pro Asp Ser Thr Thr Thr Gly Lou Ser Thr Ser Thr Ala Ser Pro 538

1615 CAT ACC AGA WCC MCC AC ACC TCC ACA GAG AGA MG CTG CCG GAG CCG GAG AGC CGG GOC GTG GTC ATC GTG OCT GTG ATT GTG TGC ATCHis Thr Are Ale An Ser Thr Ser Thr Glu Are Lys Lou Pro Glu Pro Glu Ser ArB GlY Val Val Ile Val Ala Val Ile Vol Cvs Ile 568

1705 CTG GTC CTG GCG GTG CTG GOC GCT GTC CTC TAT TTC CTC TAT MG MGG GO MG OCT WCC GTG CAG GCG CTC AGG GM GCA GGA GAT CACLeu Val Lou Ala Val Leu Gly Ala Val Lou Tyr Phe Leu Tyr Lys Lys Gly Lys Ala Ala Val Gln Ala Lou Are Glu Ala Gly Asp His 598

1795 OCT WCC CCC GTC TCG TM GACCGMCTTGTAGTTGMGTTMGTCAGATMGCTCCCAGMGAGATGGGCCTCCTGCAGGCAGCAGCGGTGACMGAGGGCTCCGGGAGACCAAla Ala Pro Val Ser END 603

1908 GGGAGAGMMTACATCGATCTGAGGCATTAGCCCCGMTCACTTCAGCTCCCTTCCCTGCCTGGACCATTCCCAGTCCCTGTCACTCTTCTCTCAGCCMAGCTCAAAGGGACTAGA2027 GAGMGCCTCCTGTCCCCTCGCCTGCACACCCCCTTTCAGAGGGCCACTGGGTTAGGACCTGAGvACCTCACTTGGCCCTGCMGGCCCGCTTTTCAGGGACCAGTCCACCACCATCT21486 CCTCCACGTTGAGTGAAGCTCATCCCMGCMGGAGCCCCAGTCTCCCGAGCGGGTAGGAGAGTTTCTTGCAGMCGTGTTTTTTCTTTACACACATTATGCTGTMMTACGTCGTCC2265 TGCAGAGTGAGTGGGTAGCCTCTCTGAGCTGGTTTCCTGCCCCMMGGCTGGCATTCCACCATCCAGGTGACCACTGMGTGAGGACACACCGGAGCCAGGCGCTGCTCATGT2384 TGMGTGCGCTGTTCACACCCGCTCCGGAGAGCACCCCAGCAGCATCCAGMGCAGCTGCAGTGCAAGCTTGCATGCCTGCGTGTTGCTGCACCACCCTCCTGTCTGCCTCTTCAMGT2503 CTCCTGTGACATTTTTTCTTTGGTCAGAGGCCAGGACTGTGTCATTCCTTAAAGATACGTGCCGGGGCCAGGTGTGGCTCACGCCTGTATCCCAGACTTTGGGAGGCCGAGGCGGC2622 GGATCACAAGTCAGACGAGACCATCCTGGCTACACGGTGAACCCTGTCTCTACTAA TAC AAAATTAGCTAGGCGTAGTGGTTGGCACCTATAGTCC AGCTACTCGGA2741 AGGCTGAGCAGGAGATGGTATGATCCAGGAGGTGGAGCTTGCAGTGAGCCGAGACCGTGCACTGACTCCAGCCTGGGCA CACAGGAGACTCC AAAAAAAAT2860 CGTGCTCGTAGAGTGGCTCTGTTTCGAGTCAGGTGMTTAGCCTCMTCCCCGTGTTCACTTGTCCCATAGCCCTCTTGATvGATCACGTAMMCTGMAGGCAGCGGGGAGCAGA2979 CAAAGATGAGGTCTACACTGTCCTTCATGGGGATTAAACTATGGTTATATTAGACCAACTTCTACAACCAGCTCAGGACCCACCCTAGAGGGCCCAATGAGAGATGGTAC3098 TTAGGGATGGCMMCGSGCCTGGCTAGAGTTCGGGTGTGTGTGTCTGTCTGTGTATGATACATATGTGTGTATATATGGTTTTGTCAGGTGTGTMMTTTGCMMTTGTTTCCTTT3217 ATATATGTATGTATATATATATATGAMATATATATATATATGAMMTAAACTTMTTGTCCCAGAAAMAAMAAA

B2847 CCGGCCGGTTOGCGGGCCCTCGGACCCTCAGAGAGGCGAGGGTTCGAGGGCACGAGTTCGAGGCCMCCTGGTCCACATGGGTTGAAAAAAA

FIG. 6. Nucleotide and deduced amino acid sequence of the MUC18 glycoprotein. (A) Nucleotides are numbered at left and amino acidsare at right of the sequence, starting with the initiation codon and the corresponding methionine, respectively. The hydrophobic putative signalpeptide and transmembrane sequences are underlined. The potential signal peptide cleavage site between Ala-28 and Glu-29 is marked with anarrow. Nine potential asparagine-linked glycosylation sites are indicated by asterisks. The consensus polyadenylylation site AATAAA atposition +3264 is marked with a bold line. (B) Divergent 3' nucleotide sequence ofcDNA clone drop4.7, which begins at position +2847 (markedwith an arrow in A).

residues with a calculated molecular weight of 63,195. This is extracellular part of MUC18 is constructed from five do-in excellent agreement with the observed size of65 kDa for the mains, all homologous to each other and to immunoglobulin-deglycosylated MUC18 (glycopeptidase F digestion, data not like domains (Fig. 7). Domains I and II of the MUC18shown). The deduced amino acid sequence includes a second molecule belong to the V-set sequences and the domainsstretch of 24 hydrophobic residues characteristic for a mem- III-V belong to the C2-set sequences. Each of the domains isbrane-spanning region and eight sites for putative asparagine- approximately 100 amino acids long and contains two appro-linked glycosylation (Asn-Xaa-Ser/Thr) that are located in the priately spaced cysteines as well as conserved residuesextracellular region of the molecule. Due to proline in the flanking the cysteines. The observed change in the apparentAsn-Pro-Ser sequence at positions 286-288, this site will not molecular mass from 113 kDa under reducing conditions tobe used for glycosylation (23). 100 kDa under nonreducing conditions in SDS/PAGE (7)MUC18 Is a Member of the Immunoglobulin Superfamily. strongly suggests the presence of intrachain disulfide bonds

The MUC18 molecule fulfills all criteria proposed for mem- in the MUC18 molecule. The algorithim of Chou and Fasmanbership in the immunoglobulin superfamily (24). The entire (25) to predict protein secondary structure indicates predom-

Dow

nloa

ded

by g

uest

on

Nov

embe

r 23

, 202

0

Proc. Natl. Acad. Sci. USA 86 (1989) 9895

........... S . . . . . . . . . . . . . . . . . . . . ........ s .. . . . . .

MUC18 I 37 lV G lSTIL L[lQ G L S 8 V Dlir S V 34aa L AgT QF T Pt- Q R I F L C Q O - K R LFGII 153 PZt Z - r tc V G G YM[ I P Q V IYl11K@G1 29aa L K A Q LLJK D It A7 Q r Y c * L - YlIf LIII 263 D R P L SIS[ Q 16aa L V IL - P A R K C Q IL D T

IV 356 Z C SSJm C ASSQ D L S - F Q L F FZ 10: L Q LI1 - D L K R 9 A 0O A1 C VAS P S IV 443 1f1VLEr3~ -[8RmT I S 1 V1r~115aa 11 V L -|V T P - rLV- G V F T A SGD L[

N-CAl III 209 L S Q LACGDAD-FGZ-F T T[3Djor 20aa I I K KnVD K - SD I|A KIY A|IV 304 1-D Q IS L IC z I !~!S IIT wK T S T 24& JTtELK I Q Y - T DA G F T V TI S IT IE

UAG III 253 [G SS[MSL G i-P[M L LT WlER Dg 12aa [D E E V T P A iBD- AMJ LIA iFilA YWlIV 339 ZGTlT I L S T I LLTI FnE3 K llaa LELP AV T P -EDDG YVCVIA zIEQYT

B C z F

FIG. 7. Sequence homology between five repeated domains in the MUC18 molecule and relationship to two immunoglobulin-like domainsof neural cell adhesion molecule (N-CAM) and myelin-associated glycoprotein (MAG). Sequences were visually aligned to show maximuminternal homology. Boxed amino acids correspond to residues shared by three or more sequences. The middle portion of each domain wasomitted for simplicity, and the number of amino acids not shown is indicated. The position of the putative disulfide bond is indicated byS . . . S. The location of the known 13-strands in immunoglobulin domains is marked below the aligned sequences with dashed lines. Thesingle-letter amino acid sequence is used.

inantly }-strands in each ofthe MUC18 domains, fulfilling theprediction for an immunoglobulin domain.A search of the Max Planck and Martinsried Institute

Protein Sequence Data Bank (MIPSZ; release 9) using theFASTP program (26) detected no significant similarities out-side the immunoglobulin superfamily. Among the immuno-globulin-related molecules, MUC18 shows the greatest sim-ilarity to the cell adhesion molecules of the nervous systemneural cell adhesion molecule (N-CAM) (27), Li (28), amal-gam (29), fasciclin 11 (30), contactin (31), and myelin-associated glycoprotein (MAG) (32) and to human carcino-embryonic antigen (CEA) (33), which has also been shown tomediate cell adhesion (34) (Table 1). The immunoglobulin-related molecules are thought to have been derived by geneduplication and divergence from an archetypal gene. Withinthe superfamily, some molecules show greater than averagesimilarity suggesting a more recent common origin. One suchgroup consists of the neural cell adhesion molecules thatshare functional as well as structural properties. The closestructural relationship of MUC18 to these adhesion mole-cules predicts that MUC18, like these molecules, may also bedevelopmentally regulated and involved in intercellular ad-hesion. Melanocytes are derivatives of the neural crest andMUC18 expression by melanomas may reflect a stage in theirontogeny. If the MUC18 molecule is indeed involved inadhesive interactions, its expression may allow melanomacells to interact with cellular elements of the vascular system(where MUC18 is also expressed) thereby enhancing hemato-geneous tumor spread.

We thank F. Lottspeich, S. Modrow, and F. Pfeiffer for computerconsultations, and E. W. Breitbart and E. B. Brocker for providingmelanoma specimens. This work was supported by a grant from theDeutsche Krebshilfe, Mildred Scheel Stiftung, Bonn, F.R.G.

Table 1. Proteins with sequence similarity to MUC18 glycoproteinOverlap

Identity, amino acids, Optimized SDMolecule % no. score value

N-CAM 25.5 133 133 22.1CEA 20.2 297 160 19.7Li 22.6 288 172 17.2Amalgam 26.0 169 133 15.4Fasciclin II 16.2 229 132 14.7Contactin 22.2 117 79 13.3MAG 21.8 362 130 11.9

Comparisons were performed using the FASTP and RELATE pro-grams (26). The optimized score is obtained when one takes deletionsand insertions into account (that is, gaps). It is the usual output fromthe FASTP program and is discussed in ref. 26. SD values greater than10 are considered to indicate significant sequence homology. N-CAM, neural cell adhesion molecule; CEA, carcinoembryonic anti-gen; MAG, myelin-associated glycoprotein.

1. Fidler, I. J. & Hart, I. R. (1982) Science 217, 998-1003.2. Brocker, E. B., Suter, L., Bruggen, J., Ruiter, D. J., Macher, E. &

Sorg, C. (1985) Int. J. Cancer 36, 29-35.3. Holzmann, B., Brocker, E. B., Lehmann, J. M., Ruiter, D. J.,

Sorg, C., Riethmuller, G. & Johnson, J. P. (1987) Int. J. Cancer 39,466-471.

4. Johnson, J. P., Stade, B. G., Hupke, U., Holzmann, B. & Rieth-muller, G. (1988) Immunobiology 178, 275-284.

5. Johnson, J. P., Stade, B. G., Holzmann, B., Schwable, W. &Riethmuller, G. (1989) Proc. Natl. Acad. Sci. USA 86, 641-644.

6. Brocker, E. B., Zwaldo, G., Holzmann, B., Macher, E. & Sorg, E.(1988) Int. J. Cancer 41, 562-567.

7. Lehmann, J. M., Holzmann, B., Breitbart, E. W., Schiegelow, P.,Riethmuller, G. & Johnson, J. P. (1987) Cancer Res. 47, 841-845.

8. Woodbury, R. G., Brown, J. P., Yeh, M. Y., Hellstrom, I. &Hellstrom, K. E. (1980) Proc. Natl. Acad. Sci. USA 77, 2183-2187.

9. Haeffner-Cavaillon, N., Cavaillon, J. M. & Szab6, L. (1985) inHandbook of Endotoxin, ed. Berry, L. J. (Elsevier Science,Amsterdam), Vol. 3, pp. 1-24.

10. Pink, J. R. L. & Ziegler, A. (1979) in Research Methods in Immu-nology, eds. Lefkovitis, I. & Pernis, B. (Academic, New York), pp.169-180.

11. Neefjes, J. J., Doxiadis, H., Stam, N. J., Beckers, C. J. & Ploegh,H. L. (1986) Immunogenetics 23, 164-171.

12. Cleveland, D. W., Fischer, S. G., Kirschner, M. W. & Laemmli,U. K. (1977) J. Biol. Chem. 252, 1102-1106.

13. Young, R. A. & Davis, R. W. (1983) Proc. Natl. Acad. Sci. USA 80,1194-1198.

14. Sanger, F., Nicklen, S. & Coulson, A. R. (1977) Proc. Natl. Acad.Sci. USA 74, 5463-5467.

15. Maniatis, T., Fritsch, E. F. & Sambrook, J. (1982) MolecularCloning: A Laboratory Manual (Cold Spring Harbor Lab., ColdSpring Harbor, NY).

16. Breslow, A. (1970) Ann. Surg. 172, 902-908.17. Sober, A. J. (1984) in Cutaneous Melanoma and Precursor Lesions,

eds. Ruiter, D. J., Welvaar, K. & Ferrone, S. (Nijhoff, Boston), pp.126-142.

18. Shaw, G. & Kamen, R. (1986) Cell 46, 659-667.19. Kozak, M. (1989) J. Cell Biol. 108, 229-240.20. Chou, P. Y. & Fasman, G. D. (1974) Biochemistry 13, 211-245.21. Kyte, J. & Doolittle, R. F. (1982) J. Mol. Biol. 157, 105-132.22. Heijne, G. (1986) Nucleic Acids Res. 14, 4683-4690.23. Bause, E. & Hettkamp, H. (1979) FEBS Lett. 108, 341-344.24. Williams, A. F. & Barclay, A. N. (1988) Annu. Rev. Immunol. 6,

381-405.25. Chou, P. Y. & Fasman, G. D. (1974) Biochemistry 13, 211-245.26. Lipman, D. J. & Pearson, W. R. (1985) Science 227, 1435-1439.27. Cunningham, B. A., Hemperly, J. J., Murray, B. A., Prediger,

E. A., Brackenbury, R. & Edelman, G. M. (1987) Science 236,799-806.

28. Moos, M., Tacke, R., Scherer, H., Teplow, D., Fruh, K. &Schachner, M. (1988) Nature (London) 334, 701-703.

29. Seeger, M. A., Haffley, L. & Kaufman, T. C. (1988) Cell 55,589-600.

30. Harrelson, A. L. & Goodman, C. S. (1988) Science 242, 700-708.31. Ranscht, B. & Dours, M. T. (1988) J. Cell Biol. 107, 1561-1573.32. Salzer, J. L., Holmes, W. P. & Colman, D. R. (1987) J. Cell Biol.

104, 957-965.33. Oikawa, S., Imajo, S., Noguchi, T., Kosaki, G. & Nakazato, H.

(1987) Biochem. Biophys. Res. Commun. 144, 634-642.34. Benchimol, S., Fuks, A., Jothy, S., Beauchemin, N., Shirota, K. &

Stanners, C. P. (1989) Cell 57, 327-334.

Cell Biology: Lehmann et al.

Dow

nloa

ded

by g

uest

on

Nov

embe

r 23

, 202

0