structure and chromosomal location of the human gene ... · lage. the structure and chromosomal...

THE JOURNAL OF BIOLOGICAL CHEMISTRY (0 1990 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 265, No. 32, Issue of November 15, PP. 19624-19631,199O Printed in U.S. A.

Structure and Chromosomal Location of the Human Gene Encoding Cartilage Matrix Protein*

(Received for publication, July 5, 1990)

Robert N. Jenkins$§, Sherri L. Osborne-Lawrence$§, Andrea K. Sinclair$$, Roger L. Eddy, Jr.ll, Mary G. Byersn, Thomas B. Showsll, Allan D. Duby$$)) From the *Harold C. Simmons Arthritis Research Center, SDepartment of Internal Medicine, University of Texas Southwestern Medical Center, Dallas, Texas 75235 and the llDepartment of Human Genetics, Roswell Park Memorial Institute, Buffalo, New York 14263

Cartilage matrix protein (CMP) is a major component of the extracellular matrix of nonarticular cartilage. The structure and chromosomal location of the human gene encoding CMP was determined by molecular cloning analysis. We used a partial chicken CMP cDNA probe to isolate three overlapping human genomic clones. From one of these clones, a probe containing 2 human CMP exons was isolated and used to map the gene to chromosome 1~35 and to screen a human retina cDNA library. Two overlapping cDNA clones were isolated. The predicted protein sequence of 496 amino acids includes a 22-residue signal peptide and a 474-residue mature protein of M, 51,344. The human CMP gene and polypeptide are strikingly similar to the chicken CMP gene and polypeptide. Human CMP is 79% identical to chicken CMP and contains two homologous domains separated by an epidermal growth factor-like domain. One potential N-glycosylation site is conserved between the two species. The human CMP gene spans 12 kilobase pairs with 8 exons and 7 introns which are similar in size to those of the chicken CMP gene. Both RNA splice junctions of intron G in the human and chicken CMP genes are noncon- forming to the consensus splice sequences. This suggests that the CMP gene utilizes a new RNA splicing mechanism.

The extracellular matrix (ECM)’ is an insoluble network of protein and carbohydrate. The ECM of cartilage comprises 95% of its weight; the major components are Type II collagen and aggregates of the major proteoglycan of cartilage (1, 2). The biochemical and functional properties of many of the noncollagenous macromolecules of cartilage are incompletely understood (3). Cartilage matrix protein (CMP) is a molecule of unknown function found in the ECM of cartilage. It nor- mally exists as a homotrimer of M, 148,000 with three sub-

* This work was supported by United States Public Health Service Grant GM2040454 (to T. B. S.). The costs of Dublication of this article were defrayed in part by .the payment of page charges. This article must therefore be hereby marked “aduertisement” in accord- ance with 18 U.S.C. Section 1734 solely to indicate this fact.

The nucleotide sequence(s) reported in this paper has been submitted to the GenBankTM/EMBL Data Bank with accession number(s) JO5666 and 505667.

11 Pew Scholar in the Biomedical Sciences. To whom correspond- ence should be addressed: Dept. of Internal Medicine, Universitv of Texas Southwestern Medical-Ctr., 5323 Harry Hines Blvd., Dallas, TX 75235-8884. Tel.: 214-688-8772.

’ The abbreviations used are: ECM, extracellular matrix; CMP, cartilage matrix protein; SDS, sodium dodecyl sulfate; bp, base pair; kb, kilobase pair.

units linked by disulfide bonds (4). The distribution of CMP has been studied by radioimmunoassay in bovine tissues where it was detected in tracheal, nasal septal, xiphisternal, auricular, and epiphysial cartilage, but not in articular cartilage or extracts of the intervertebral disc (5). The largest quantities of CMP were found in tracheal cartilage where it comprised up to 5% of the wet weight of aged tissue (6). Cartilage matrix protein has also been detected by immunocytochemistry in embryonic chick eyes (7). Antiserum raised in rabbits against bovine CMP detected large amounts of CMP in human trachea, but CMP was not detectable in articular cartilage from human knees (8).

The structure of chick CMP has been deduced by molecular cloning analysis (9, 10). It consists of two homologous repeated domains separated by a domain homologous to epider- ma1 growth factor. The functional role of the epidermal growth factor-like domain has not been determined. The repeated domains, termed CMPl and CMP2, are homologous to segments of von Willebrand factor, complement factors B and C2, the 01 chains of members of the integrin family, and Type VI collagen (10). Two of these three homologous domains in von Willebrand factor have the ability to bind collagen (11, 12). VLA-2, a member of the integrin family, is thought to be a collagen receptor. The (Y subunit of VLA-2 contains a domain homologous to CMPl and CMP2 (13). The ability of CMP to bind collagen has not yet been demon- strated.

We report here the cloning from a retina cDNA library of cDNAs encoding human CMP, and the structure and chromosomal localization (1~35) of the human CMP gene. The deduced protein sequence of human CMP is 79% identical to the chicken sequence. There is striking similarity between the organization of the human and chicken CMP genes. Not only is there conservation of the structure of the exons, but intronic sizes are also conserved between the two species. Furthermore, one of the introns of the human gene displays the same novel RNA splice junctions that are found in the chicken CMP gene.

EXPERIMENTAL PROCEDURES

Isolation of Human CMP Genomic and cDNA Clones-A 770-base pair (bn) Dartial cDNA probe, pCMP1. encoding the entire CMP2 homain, the carboxyl-terminal iomain,‘and approximately 50 bp of the 3’-untranslated region of the chick CMP gene was a generous gift of Dr. P. Goetinck (9). This probe was used to screen two human genomic libraries cloned into X bacteriophage vectors (14, 15). Filters were hybridized at 40 “C overnight in 50% formamide, 5 X SSC (16), 10 mM Tris, pH 7.6,l X Denhardt’s (16), 0.1% SDS, and 10% dextran sulfate with heat-denatured, nick-translated probe (-2 X 10” dpm/ 150-mm nitrocellulose filter). Filters were washed nonstringently in 0.2 x SSC, 0.1% SDS at 45 “C and exposed overnight to film. Three overlapping clones, CM(C)l, CM(M)P, and CM(M)3, were isolated

19624

by guest on March 8, 2020

http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

Human Gene Encoding Cartilage Matrix Protein 19625

and mapped with restriction enzymes (Fig. 1A). DNA sequence analysis confirmed the identification of two human exons (exons 5 and 6). From one of the human genomic clones, CM(C)l, a P-kilobase pair (kb) PstI-PstI fragment, denoted HCMP5/6, containing exons 5 and 6 was isolated and used as a probe. The 5’ PstI site of HCMP5/6 is derived from the left arm of the EMBL3 bacteriophage vector. A human retina cDNA library, kindly provided by Dr. Jeremy Nathans (Howard Hughes Medical Institute, Baltimore, MD), was screened with HCMP5/6 using the conditions described above, except the final wash was performed under stringent conditions (0.1 X SSC, 0.1% SDS, at 66 “C). Two overlapping cDNAs, pHCMP1 and pHCMP2, were isolated (Fig. 1B).

DNA Analysis and Sequencing-Bacteriophage and plasmid DNA were isolated by standard procedures (16). Restriction enzyme diges- tions were done per the manufacturer’s instructions. A 7-kb Xba-Xba fragment of CM(M)P, a 6-kb KpnI-KpnI fragment of CM(C)l, and 3.1-, 3.5-, and 9-kb EcoRI-EcoRI fragments of CM(M)3 were subcloned into pBR322 or pUC18 plasmids for further analysis. The 5’ KpnI site and 1.5 kb of CM(C)1 are derived from the EMBL3 bacteriophage vector. DNA fragments were subcloned into the M13- derived vectors MP18 and MPlS and sequenced by the chain-termination method (17). The coding regions of the genomic clones and the cDNA clones were sequenced at least once on both strands, except as noted in the figure legends.

Computer Analysis of DNA and Protein Sequences-The Beckman Microgenie (18) program was used for routine DNA sequence analysis and for generation of the deduced protein sequences. The University of Wisconsin Genetics Computer Group programs Wordsearch, FASTA, and TFASTA were used for searching data banks for homologous protein and DNA sequences (19).

Hybrid Cell Lines-A panel of hybrid cell lines derived from 19 unrelated human cell lines and 4 mouse cell lines was used to determine the chromosomal location of the CMP gene. Descriptions of the origin and construction of these cell lines are published elsewhere (20-22).

Southern Blot Analysis-High molecular weight genomic DNA was isolated from human, mouse, and hybrid cells by standard techniques (16). For each restriction endonuclease, 10 pg of DNA was digested to completion and fragments were separated by electrophoresis in 0.8% agarose gels. Nitrocellulose or Zetabind (AMF CUNO) filters containing the transferred DNA were hybridized under stringent conditions to the “‘P-labeled human CMP probe HCMP5/6, as described above. Prior to autoradiography the filters were washed under stringent conditions, as described above.

In Situ Hybridization-These techniques are described in detail elsewhere (23). Briefly, phytohemagglutinin-stimulated, human male leukocytes were synchronized in the cell cycle using bromodeoxyuri- dine. Radiolabeled HCMP5/6 probe was used for in situ hybridization prior to chromosome staining with Hoechst 33258/Giemsa.

RESULTS

Isolation of Human Genomic and cDNA Clones-A 770-bp cDNA fragment encoding the carboxyl-terminal portion of chick CMP (9) was used as a probe to screen two genomic libraries (14, 15) for the human CMP gene. Two overlapping clones, CM(M)2 and CM(M)3, were isolated from one genomic library (14), and a single clone, CM(C)l, was isolated from the other (15) (Fig. lA). A 2.0-kb fragment of CM(C)l, HCMP5/6, which includes human exons 5 and 6 was used as a probe (Fig. lA). Because it has been suggested that CMP binds collagen (9), we reasoned that the CMP gene might be expressed in tissues other than cartilage where Type II collagen is found. Accordingly, we screened a human, retina cDNA library because the retina and the vitreous, which is elaborated by the retina, are known to contain Type II collagen (24). Two overlapping cDNA clones, pHCMP1 and pHCMP2, that hybridized to HCMP5/6 were isolated. While our experiments were in progress, data showing expression of CMP by embryonic chick eyes were published (7).

Analysis of the cDNA Clones-Fig. 1B depicts the restriction enzyme map of the human CMP cDNA clones. The cDNA sequence and the deduced protein sequence are presented in Fig. 2. pHCMP1 is a 2.76-kb cDNA encoding 460 bp of coding

A. Human CMP Gene

5’ 3’ 1 * 3 4 5 6 78 3’“T

A 0 C DEFG

B. Human CMP cDNA

w oncw1

I IIIII H G HG MH 6 G

t 0.5kb 1

FIG. 1. Human CMP genomic (A) and cDNA clones (B). The genomic clones CM(C)l, CM(M)2, and CM(M)3 and the cDNA clones pHCMP1 and pHCMP2 are depicted as bars. The 5’-3’ orientation of the clones is indicated. The exon-intron structure of the gene as determined by DNA sequence analysis is shown (A). The position of the HCM5/6 probe is indicated by a dotted bar. The positions of exons 3-8 and the 3’.untranslated region (3’UT) in pHCMP1 and pHCMP2 are depicted in a bar above the cDNAs. Selected restriction enzyme sites are indicated as: X, XbnI; R, EcoRI; B, BamHI; S, SalI; H, HindIII; M, SmaI; G, BglII. The scale of the two panels is different as shown.

region and 2.3 kb of 3’-untranslated region. A consensus polyadenylation signal (25), AATAAA, is located at the 3’ end of this clone (position 3206-3211, Fig. 2). Although pHCM1 does not have a poly(A) tail, examination of the genomic sequence (see below) reveals a GT-rich segment downstream of the AATAAA motif, suggesting that it is likely to be a functional polyadenylation site.

pHCMP2 is a 1.37-kb cDNA encoding 1020 bp of coding region and 350 bp of 3’untranslated region. It overlaps the 5’ end of pHCM1 by 800 bp. pHCMP2 ends in a 17-bp poly(A) tail and contains an acceptable polyadenylation signal (25), AATATA (position 1338-1343, Fig. 2). There are several nucleotide polymorphisms in the 3’-untranslated region of these two cDNA clones (Fig. 2). Together the two cDNA clones encode the carboxyl-terminal 340 amino acids of human CMP and 2.3 kb of 3’-untranslated region. Comparison to the predicted protein sequence of the chick CMP precursor indicates that the human cDNA encodes part of CMPl, the EGF-like domain, CMP2, and the carboxyl-terminal domain. Because the chick CMP gene encodes a precursor of 493 amino acids including a 23-amino acid signal peptide, it is apparent that pHCMP1 and pHCMP2 do not comprise the entire coding region.

Identification of Exons 1 and 2 of the Human CMP Gene- pHCMP1 and pHCMP2 were used as probes to find their corresponding exons in the genomic clones. Comparison of the genomic and cDNA sequences identified the 5 introns and 6 exons at the 3’ end of the gene. Although our cDNA clones were not full length, we were able to identify the 5’ portion of the coding region in the genomic clones by taking advantage of the high degree of organizational similarity between the human and chick CMP genes.

The location and boundaries of exons 1 and 2 and introns A and B were deduced as follows. We noted that the sizes of


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

19626 Human Gene Encoding Cartilage Matrix Protein

FIG. 2. Nucleotide and deduced amino acid sequence of the human CMP cDNAs. Overlapping cDNA clones pHCMP1 and pHCMP2 were merged to give the depicted nucleotide sequence. Nucleotide numbering is shown to the right of the sequence. The one-letter amino acid code is used. The relationship of the cDNA sequence to the intron-exon organization of the CMP gene is indicated. The CTT codon (position 403-405), which differs from the genomic clones, is indicated by an asterisk. Polyadenylation signal sequences are underlined. The polyadenylation site of pHCMP2 is indicated by A. Sequence polymorphisms: at position 805-807, (GAA) in pHCMP1 uersuS (GAG) in pHCM2 represents a silent polymorphism; position 1065 (G) in pHCMP1 ucrsus (A) in pHCMP2; and beginning at position 1204, the number of (TG) repeats varied from 11 to 13.

TABLE I Comparison of the intron/enon organization of the human and chicken CMP genes

Species EXOll Length 5’ splice” Intron Length* 3’ splice” Intron Amino acid phase’ interrupted

Human Chicken

Human Chicken

Human Chicken

Human Chicken

Human Chicken

Human Chicken

Human Chicken

Human Chicken

bp Id 94 1 88

2 347 2 347

3 223 3 223

4 126 4 126

5 417 5 414

6 153 6 153

7 81 7 81

@ 47 8 47

CAGAGgtgagg A CAGAGgtaagt A

GCAAGgtacgt B ATAAGgtaaga B

CTGCGgtgcgc C CTGTGgtaagt C

CAATGgtcagt D CAGTGgtatg D

CCTCGgtaggt E TTTAGgtaggt E

TGTGGgtgagt F CGTTGgtgaga F

GAAACatatcc G GAAATatatcc G

b 1647 2720

2473 4840

1617 2270

464 280

585 1190

829 570

644 2940

tgccccactcccagGCC cctcctctgtgcagGCA

cccgcgtgtcgcagGTG CctcctctgtgcagGTG

tattgtccccgcagTGG actttttcttgcagTGG

cctgtcctccacagTCT ttattttgcttcagCTT

gttcctcctggcagGCT CttattgttcacagGCT

CttgtcccctgcagAGG actctggtccacagAGG

aactctgagtccacTGG taactctcactcacTGG

Consensus splice sequencese AGgtiagt :zttfzffEfncagG

I I

I I

G (32) G (30)

K/V (147/148) K/V (145/146)

v (222) v (220)

V (264) A (262)

G (403) G (400)

E (454) E (451)

L (481) L (478)

’ Exon sequences are in capital letters; intron sequences are in lower case. b Lengths of the chicken CMP gene introns were determined by electron microscopy (10); human lengths were

determined by DNA sequencing. ’ Phase 0 introns lie between two codons; phase I introns lie between the first and second nucleotides; and

phase II lie between the second and third nucleotides of a codon (40). d The lengths of the 5’- and 3’-untranslated regions in exons 1 and 8 are not included. ’ Shapiro and Senapathy (33).

introns C-F were similar to those of the chicken gene (Table at both ends of the deduced boundaries of human introns A I). Reasoning that introns A and B at the 5’ end of the human and B (Table I). The DNA and the predicted protein se- CMP gene might also be similar in size to the chicken introns, quences of the human CMP introns and exons are presented we sequenced the 5’ end of the human genomic DNA in areas in Fig. 3. where the chicken CMP gene structure predicted the remain- There are two potential translation initiation sites 27 bp ing human exons would be found. This strategy was successful apart (positions 490 and 517, Fig. 3). We suspect that the in identifying DNA sequences (exons 1 and 2) whose deduced proximal ATG is the initiator codon because the 5’-proximal protein sequences are highly homologous to the chick CMP ATG triplet served as the initiator codon in >90% of the protein sequence. RNA consensus splice signals are present mRNAs examined by Kozak (26). Furthermore, the presence


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


FIG. 3. Nucleotide and deduced amino acid sequence of human genomic clones encoding CMP. Over- lapping sequences from CM(C)l, CM(M)P, and CM(M)3 clones were merged to give the complete nucleotide sequence. Where the sequences from the different clones differ, the CM(M)2 sequence is depicted. The one-letter amino acid code is used. Numbering of the nucleotide sequence begins arbitrarily. The exons were sequenced on both strands with the following exceptions: Exon 3, 6 bp of exon 3 and the adjoining 5’ end of intron C were sequenced on one strand twice; Exon 7, 20 bp of exon 7 and the adjoining 5’ end of intron G were sequenced three times on one strand. The putative CAAT and TATA sites are underlined. The putative polyadenylation signals in the 3’-untranslated region are underlined. The downstream GT-rich sequences are double underlined. The polyadenylation sites of the cDNA clones are indicated by “-1” for pHCM1 and “-2” for pHCM2. The TTT codon (position 8280), which differs from the cDNA sequence, is indicated by an asterisk. Silent polymorphisms: in exon 5 (position 846998471), CM(M)2 = (ACT) uersus CM(C)1 = (ACC); in exon 6 (position 9267-9269) CM(M)2 = (GAG) versus CM(C)1 = (GAA) (the same polymorphism is present in the cDNA clones at position 807, see Fig. 2 legend); in intron E at position 9042 CM(M)2 = (T) aemus CM(C)1 = (G) (not shown); and in the 3’untranslated region at 11,093 CM(M)2 = (T) versus CM(C)1 = (C). Additional differences between the genomic and cDNA sequence were noted at positions 11,093, 11,253, 11,788, 11,956, 12,118, 12,222, 12,241, 12,469, 12,490, 12,569, 12,585, 12,795, 12,798, 12,818, 12,906, 12,979, and 13,129.

of a purine nucleotide 3 bases upstream of the initiator codon is the most conspicuous conserved feature of initiation sequences (26), and the proximal ATG has the purine nucleotide adenylic acid (A) in this position, whereas the distal ATG does not.

Deduced Amino Acid Sequence of Human CMP-The protein sequence deduced from the cDNA and genomic clones is presented in Fig. 4. The entire gene is predicted to encode a precursor protein of 496 residues. To determine the probable beginning of the mature polypeptide, we examined the sequence of the amino-terminal portion of the precursor protein for characteristics of a signal peptide and the likely, proteo- lytic cleavage site. The algorithm of von Heijne (27), applied by computer program (28), indicated that more than one

acceptable signal peptidase processing site is present. The sites producing the highest scores were after residue 20 (alanine) and after residue 22 (cysteine). The former site would leave a cysteine (at position 22) in mature human CMP that is not present in mature chicken CMP. Therefore, we suspect that the cleavage site occurs after residue 22. This analysis suggests that the signal peptide consists of the first 22 amino acids, and mature human CMP begins with serine at position 23 and contains 474 residues with a calculated M, of 51,344.

There is a potential N-glycosylation site (NAT) (29) in CMPl at residues 76-78; such a site is found in a similar location of chicken CMP. There are two domains, CMPl and CMP2, of approximately 190 amino acids, encoded by exons 2 and 3 and exons 5 and 6, respectively. The protein sequences


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

19628 Human Gene Encoding Cartilage Matrix Protein

MRVLSGTSLMLCSLLLLLOALC5PGLAPQ~RGHLCRTRPiDLVFVVDSS~SVRPVEFEK~KVFLSQVIESLDVGP~RVGHVNYASTVKoEFSLRAHYf 100

W\ALL~AVR~IQPLSTGTHiGLAI~FAIT~AFGDAEGGR~RSPDISKVViVVTDGRP~D~VQDVSARAR~SGVELFAIG~GSVOKATLR~IASEP~DEH~ 200

DYVESYSVIEKLSRKFqEAFCVVSDLCATGDHDCEQVCIS 300

IVDTLDVSD~LAQVGLV~Y~SSVRQEFPL~RFHTKKDIKkAVRNMSYME~GTMTGAALK~LIDNSFTVS~GARPGA~KV~IVFTDGRS~~YINDAAKK~ 400

DLGFKMFAVGVGNAVEDELREIASEPVAEHYFYTADFKTiNPIGKKLBKKICVEEDPCACESLVKFQAKVEGLLqALTRI;LEAVSKRLAiLENTVV 496

FIG. 4. Deduced amino acid sequence of human CMP. Residues are numbered consecutively from the first methionine residue. The mature polypeptide is deduced to begin at the serine (*) at position 23. The site of potential N-glycosylation at position 76-78 is underlined. The leucine/phenylalanine polymorphism at position 291 is shown.

FIG. 5. Domain structure of hu- LLW” P RI 9 PLrTGTHTGLAmF*I TI AF G 0 * EG G R 5 ROP D I 5 K” ” l”“TLx R F-xl 5 “Lm”SARAR as G ” E L FAIGVG 5 man and chicken CMP. Amino acid clcll00000001101 llcln LLPA” 9 RI E PLSTG7HTOLAW.4I SR AF s 0 T EG A R L RSP N I N K” h I”“TDG R mrl 0 “PDVSARIR q4 G I E I FA,G”G I

sequences were aligned by visual inspec- rluclclu “D KA TLRQl*SEP a D E H”DY”ESYS”IEKL SR KFQEP.FC” “D Mm TwqIASEP L 0 D H”DY”ESYS”IEYL TH WqEAFC”

tion. Identical residues are enclosed in boxes. The signal peptidase cleavage site III. EGF-LIKE DOMAIN for the human precursor protein is not “SDLCATGDHDCEQ ” -21s s PGSY * CAC n EGFTLN 5 C-XT‘ N known. The rationale for ending the sig- HRN Iuclclln CHICK “SDLCRTGOHDCEQ I CIS T PGS” K CAC K EGFTLN N DGKTC 5

nal peptide sequence as shown is provided under “Results.” IV. C; ?A”

” CS‘G ‘ OS s A T DL”FLIMjS~S”RPENFEL”~~~, %K I\ CSGO 5 GS - A L DL”FLroGsrs”RPfHFEL”nlFI -ICI 0cl0-cl I cl1 5 91” DT L 0 “5 D I: L I\Q”GL”Q”SSS”lqEFPLG R F HT KKDIKAA” RN H s YMEIGTMTG N 91” ES L E “S E I( 9 b.Q”GL”Q”SSS”R9EFPLG 9 F KN KKDIWA” tx H a “HEIG,MTG

V. CARBOXY-TERMINAL DOMAIN

of CMPl and CMPZ are 42% identical. At each end of CMPl and CMP2 there is a cysteine available for intradomain disulfide bonding. CMPl and CMP2 are separated by a 41- amino acid, EGF-like region, encoded by exon 4. The carboxyl-terminal domain is encoded by exons 7 and 8 and contains 2 cysteine residues separated by a single amino acid.

Comparison to Chicken CMP and Other Proteins-The amino acid sequences of the mature polypeptides of human and chicken CMP are 79% identical (Fig. 5). The locations of the 14 Cys residues are completely conserved between species. Human and chicken CMPl domains are 79% identical, while the CMP2 domains are 84% identical. The EGF-like region is 85% identical to the corresponding region of the chicken protein. The carboxyl-terminal domains of human and chicken CMP are 61% identical.

A computer search of the GenBank, EMBL, and NBRF data banks for sequences homologous to the CMPl and CMP2 domains produced the three tandem repeats of human von Willebrand factor, the LY subunits of human LFA-1, human p150,95, human and murine MAC-l, the 01~ chain of chicken Type VI collagen, and human complement factor B.

According to the Type A and B classification scheme of Doolittle (31), CMP EGF-like domains are Type B and lack the residues thought to be required for Ca2+ binding activity that are found in many Type A EGF-like domains. There are no published sequences (except chicken CMP) with signifi- cant homology to the carboxyl-terminal domain.

Analysis of the Human CMP Gene-The human CMP gene spans approximately 12,000 bp. There are 8 exons (numbered l-8) separated by 7 introns (A-G) (Fig. lA, 3). The nomen- clature used is the same as for the chick gene (10). The exon- intron structures of the human and chicken CMP genes are compared in Table I. In every case the human and chicken exons encode virtually the same number of amino acid resi-

dues. The human introns, with the exception of intron G, are 50-200% the size of the corresponding chicken introns.

The RNA splice junctions are homologous to the consensus

sequences (33), AGgtiagt, at the 5’ end of the introns and, tt ttftft &;cccccncagG, at the 3’ end of the introns with the exception of intron G. At the 5’ end of human intron G the dinucleotide “at” is found instead of the consensus dinucleotide “gt” and at the 3’ end “ac” is found instead of “ag.” Exceptions to the “gt-ag” rule are very rare; the identical substitutions are found in the RNA splice sites for intron G of the chicken CMP gene (Table I) (10). Interestingly, both species share 13 identical residues (tctccttaactct) 7 bp upstream of 3’ splice site for intron G.

3’-Untranslated Region-Searching the 3’-untranslated region for acceptable polyadenylation signals revealed seven possible sites (Fig. 3). In view of evidence that transcription termination in eukaryotes depends on “GT-rich” sequences about 15-40 bp downstream of the polyadenylation signal and within 30 bp of the actual site of polyadenylation (34), we examined the genomic sequence for GT-rich modules downstream of these seven sites. There is a GT-rich sequence downstream from the putative polyadenylation signal (AA- TATA, position 11,233-11,238) used by the cDNA clone pHCMP2. Another GT-rich module is found 18 bp downstream from the AATAAA sequence (position 13,204-13,209), the putative polyadenylation signal for pHCMP1. One other potential polyadenylation signal at position 11,322-11,327 has a GT-rich element downstream. This sequence (ATTAAA) is identical to the polyadenylation signal used by the chicken CMP gene.

The 3’-untranslated region of the chicken CMP gene is approximately 1.5 kb long. No extended ranges of sequence


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


identity were found in the 3’-untranslated region between the chicken (400 bp are available for comparison) and human CMP genes.

Promoter Region--We examined the human DNA sequence 5’ from the putative initiation codon for features similar to those noted in the same region of the chicken gene. This region is likely to be the promoter if the human gene, like the chicken gene, lacks introns within the 5’-untranslated sequence (10). Most eukaryotic promoters have a TATA box (35). The human CMP gene contains the sequence TTGATA (position 426-431), 65 bp upstream from the first of two potential initiation codons (see above). The chicken CMP gene has a similar sequence, TTAATA, 69 bp upstream from its initiation codon. Neither sequence conforms to the consensus sequence

TATAAAA TT

The GGCCAGGCT sequence (position 278-286) may rep- resent a CAAT box as it is identical in 7 of 9 bp to the consensus sequence (36).

GG$CAATCT

The CAAT-like sequence (10) in the chicken CMP gene (ATGCAATAT) is identical to the consensus sequence in 5 of 9 bp. Four characteristic binding sites (GGGCGG) for the DNA binding protein SPl (37) are found in the putative human promoter region (positions 290-295, 331-336, 357- 362, 411-416); the 5’ pair is in the opposite orientation from the 3’ pair.

Examining the human and chicken promoters for homologous sequences revealed the following shared motifs:

CTTCT-$~GCAA~GAGCC$TTGTGG (position 229-255)

and

CCACTeCAC (position 444-455).

The sequence CCCAGCCC and ,its 6-base inverted repeat GGCTGG were noted (10) to be prominent in the chicken gene promoter as was CCCATCCC. In the human gene the sequence GGGCTGGG is present twice (positions 94-101, 344-351), while its 6-base inverted repeat CCAGCC is present once (position 162-167). The shorter sequence GGCTGG is also present twice (positions B-13,407-412). Conservation of these sequences between evolutionary, distant species implies their possible involvement in binding to regulatory proteins.

Polymorphisms-A few polymorphisms were detected among the sequences of the cDNA and genomic clones. In exon 5, TTT (position 8280-8282) in CM(C)1 and CM(M)2 encodes a phenylalanine residue (Fig. 3), whereas in the cDNA clone, pHCM5, CTT (position 403-405) encodes a leucine residue (Fig. 2). Two silent polymorphisms are described in the legend of Fig. 3.

In the 3’-untranslated region there were differences in the number of repeated (TG) dinucleotides (beginning at positions 1206, Fig. 2 and 11,067, Fig. 3) in different sequencing reactions of the cDNA and genomic clones. Presumably, these discrepancies are due to inaccurate replication of this region during its cloning in plasmids or M13-derived vectors. In 2290 bp of overlapping 3’-untranslated region, there were 17 differences between the genomic and cDNA sequences which are described in the legend of Fig. 3.

Chromosomal Location-Southern blot analysis of somatic cell hybrids and in situ chromosomal hybridization were used to map the human gene for CMP. The results of Southern blot analysis of genomic DNA from 45 mouse-human hybrid cell lines (Fig. 6) probed with HCM5/6 are shown in Table II. The discordancy scores localize the human CMP gene to chromosome 1. The hybrids containing translocations (Table II legend) localize the gene to the lpl2-pter region. One can also see from Fig. 6 that on HindIII-digested human DNA there is evidence of only one gene hybridizing to this probe (two bands are detected because HCM5/6 has an internal Hind111 site). Further Southern blot analysis of human genomic DNA with the HCMP5/6 probe under nonstringent conditions where homologous human genes would be detected provided evidence for only one human CMP gene (data not shown).

In situ hybridization was used to define the subchromoso- ma1 location of the CMP gene. One hundred normal metaphase chromosomes were analyzed by hybridization with HCM5/6. Of 234 grains counted, 56 (24%) were associated with chromosome 1 (Fig. 7). 27% of the grains on chromosome 1 and 6.4% of the total grains were found at 1~35. These data indicate that the 1~35 region encodes the human CMP gene.

DISCUSSION

In this paper we report the cloning of the gene encoding human CMP and its chromosomal location. Partial cDNA clones were isolated from a retina cDNA library. Overlapping X clones were isolated from two human genomic libraries. The human gene encompasses approximately 12 kb and is com- posed of 8 exons and 7 introns, with a 2.3-kb 3’-untranslated region. The deduced protein sequence of human CMP contains 496 residues, has one potential N-glycosylation site, and is highly homologous to chicken CMP. We found one protein sequence polymorphism (position 291) where the genomic clones encode a phenylalanine residue, while the cDNA clone pHCMP2 encodes a leucine. The physiological significance of this alteration remains to be determined.

The isolation of cDNA clones encoding CMP from a retina library indicates that the human CMP polypeptide may be expressed in the retina and tissues other than nonarticular cartilage. In the embryonic chick eye, CMP was detected by immunocytochemistry in the cornea, sclera, choroid, lens capsule, lens epithelium, and lens fibers, but not in retina or pigment epithelium (7). If, in fact, CMP has a collagen binding function, it may be that the tissue distribution of CMP will

+ - MH

I

WC. 6. Southern blot of somatic cell hybrids. A representative blot is shown. 10 pg of DNA prepared from the somatic cell hybrids, XER-7 (lane +) and XTR-3BSAgB (lane -), murine cells (RAG, lane M), and human cells (GM00131, lane H) were digested with HindIII, size-fractionated on a 0.8% gel, transferred to nitrocellulose, and hybridized to the HCM5/6 probe. The HCM5/6 probe hybridized to a 7.7-kb murine restriction fragment while it hybridized to 5.9- and 3.0-kb human restriction fragments. Somatic hybrid cells, XER, were scored as positive while somatic hybrid cells, XTR-3BSAgB, were scored as negative. Lane 1, hybrid positive for chromosome 5; lane 2, hybrid negative for chromosome 5; lane 3, mouse DNA; lane 4, human DNA.


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

Human Gene Encoding Cartilage Matrix Protein

TABLE II Concordancy analysis of somatic cell hybrid panel

The data are compiled from 45 hybrids involving 19 unrelated human cell lines and 4 mouse cell lines (20, 52). The hybrids were characterized by karyotypic analysis and by mapped enzyme markers, and partly by mapped DNA probes (20, 53, 54). The DNA probe, HCM5/6, was hybridized to Southern blots containing BamHI-digested DNA from the human-mouse hybrids. The scoring for HCM5/ 6 was determined by the presence (+) or absence (-) of human hands in the hybrids on the blots. Concordant hybrids have either retained or lost the human bands together with a specific human chromosome. Discordant hybrids have either retained the human bands, but not a specific chromosome or the reverse. Percent discordancy indicates the degree of discordant segregation for a marker and a chromosome. A 0% discordancy is the basis for chromosome assignment. HCM5/6 mapped to human chromosome 1. The following hybrids helped localize HCM5/6: JWR-22H (+) with the 2/l translocation: Zpter - 2q37::lp21 + lpter; JWR-26C (-) with the l/2 translocation: 1~21 + lqter::2q37 + Bqter; XOL-6 (+) with the l/X translocation: lpter + lq12::Xq26 --) Xqter; XOL-9 (-) with the X/l translocation: Xpter + Xq26::lq12 -+ lqter. From the hybrid data the gene for human CMP localizes to the ~12 - pter region of chromosome 1.

Hybridization of probe/presence of chromosome

Chromosome Concordant Discordant % discordancy

+/+ -/- +/- -/+

1 11 29 0 0 0 2 a 22 4 10 32 3 11 16 2 13 36 4 9 21 4 11 33 5 12 19 1 13 31 6 8 23 5 9 31 7 8 20 4 12 36 a 9 la 4 14 40 9 4 24 9 7 36

10 13 13 0 ia 41 11 a 23 4 9 30 12 11 14 2 18 44 13 7 20 6 12 40 14 11 14 2 ia 44 15 9 21 4 11 33 16 a 25 5 7 27 17 13 a 0 23 52 18 10 20 3 12 33 19 a 28 5 4 20 20 10 13 3 19 49 21 10 6 3 26 64 22 9 22 4 a 28 X 7 9 3 19 58

parallel that of a particular type of collagen, e.g. Type II collagen.

Application of algorithms designed to predict the signal peptidase processing site produced at least two possible sites. Beginning the mature human polypeptide with alanine at position 27 would match the alanine residue at the amino terminus of chick CMP. But this would place the proline at position 23 in the -3 position of the leader segment. Proline residues are rarely seen in positions -3, -2, and -1 of the leader segment and position +1 of the mature polypeptide. Interestingly, beginning the mature polypeptide at leucine at position 26 would produce a human polypeptide the same length as chick CMP. Furthermore, this would place serine and glycine at positions -3 and -1, respectively, of the signal peptide and proline at position 2 of the mature polypeptide. Although these are very common residues at these positions, this cleavage site would put a proline a position -2. Sequenc- ing of human CMP polypeptides will resolve this issue and add to our understanding of signal peptidase processing sites.

Integrin cytoadhesins, such as the vitronectin receptor and platelet gpIIb/IIIa, recognize sites on their ECM ligands that

I --

P31 ..@.@

221 I-

21

22 23 24 25

41

42 43 AA

1 FIG. 7. Distribution of labeled sites on chromosome 1. Sum-

mary of the analysis of 100 normal human metaphase cells from phytohemagglutinin-stimulated peripheral blood lymphocytes that were hybridized to the HCM5/6 probe. Each dot represents one labeled site observed in the corresponding band. 27% (15/56) of the labeled sites on chromosome 1 were located at 1~35; this cluster represented 6.4% (15/234) of all labeled sites.

include the tripeptide RGD (38). Human and chicken CMP do not contain this sequence, but both domains of both molecules contain the reverse sequence DGR (position 154, Fig. 4) in the same location. There is evidence that the DGR sequence may be involved in some mechanism of fibroblast adhesion to ECM glycoproteins. Yamada et al. (39) showed that the peptide SDGR inhibited spreading of fibroblasts on fibronectin, vitronectin, laminin, and collagen substrates. Whether the DGR sequence in the CMP domains is involved in its interaction with other ECM components or with cell surface ECM receptors remains to be determined.

The structure of the human CMP gene is quite similar to that of the chicken gene both in the exon/intron organization and the size of the introns (Table I). All of the human and corresponding chicken introns are in the same phase, i.e. they interrupt codons in the same place (40), and in all cases the DNA sequences at the splice sites were quite similar. Inter- estingly, as in the chicken gene, neither splice site of intron G conforms to the consensus sequences (33) for nuclear mRNA splice junctions (reviewed in Ref. 41). Furthermore, intron G does not share the characteristics found in Group I (reviewed in Ref. 42) or group II (reviewed in Ref. 43) mito- chondrial introns. Whatever mechanism is used to splice out intron G is likely to be shared by chicken and human.

Comparison of the human and chicken genes in the 3’-


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


untranslated region revealed no extended ranges of homology. Interestingly, other ECM proteins, such as cartilage link protein, show large regions of homology in the 3’-untranslated regions of homologous human and chicken genes (44). This suggests that regulatory regions in the 3’-untranslated region of the CMP gene differ between humans and chickens.

Neither pHCMP1 or pHCMP2 used the polyadenylation signal ATTAAA found in chicken CMP cDNA and human al(I) collagen cDNA (45). However, this sequence is present in the 3’-untranslated region and has a downstream GT-rich segment. Thus it is possible that additional cDNA clones would reveal this to be a functional polyadenylation signal in the human.

The poly(TG) direct repeat found in the 3’-untranslated region is apparently one of the estimated lo5 copies of poly(TG) sequences distributed throughout the human ge- nome (46, 47). Some of these sequences have been shown to play a role in the regulation of gene expression through formation of Z-DNA structures (48). The potential role of these DNA sequences in the mRNA stability and translational efficiency of the human CMP gene can now be addressed with the cDNA clones.

Mouse-human hybrid cell lines and in situ hybridization were used to map the human CMP gene to 1~35. Other loci in the region of chromosome 1 near the CMP gene include GALE (UDP-galactose-4-epimerase), FUCAl (fucosidase, ai- ~-1, tissue), RH (Rhesus blood group), C8A (complement component 8, cu-polypeptide), C8B (complement component 8, P-polypeptide), SC (Scianna blood group), APNH (anti- porter, Na+/H+), HMG17 (high mobility group protein 17), EBVSl (Epstein-Barr virus insertion site l), TD0211 (tryp- tophan-2,3-dioxygenase-like l), LCK (lymphocyte-specific protein tyrosine kinase), and GLUT1 (glucose transport 1) (reviewed in Ref. 49). These loci have no obvious functional relationship to CMP, and there are no disorders linked to the short arm of chromosome 1 which are likely to be related to CMP (49).

Of note is that ALPL (alkaline phosphatase from liver/ bone/kidney) mapped to lp36.1-~34. Human placental alkaline phosphatase (PLAP) was noted by Tsonis et al. (50) to be homologous to chick CMP. Placental and liver/bone/ kidney alkaline phosphatases are products of homologous genes and are 52% identical at the amino acid level (51). Over a stretch of 40 amino acids human CMPl and CMP2 have 19-23% identity with human PLAP and ALPL (data not shown). Thus, it is possible that CMP is evolutionarily related to the nearby ALPL gene. It will be interesting to determine the precise distance between the human CMP and ALPL genes, and whether other homologous genes are encoded nearby.

Acknowledgments-We thank Dr. Paul F. Goetinck for his gift of the chicken CMP probe. We thank Drs. Richard A. Padgett and David W. Russell for useful discussion.

REFERENCES 1. Heinegird, D., and Pa&son, M. (1987) Methods Enzymol. 145, 336-363 2. Hascall~ V. C. (1988) ISIAt1a.s Sci. Biochem. 1, 189-198 3. Heinegard, D., and Oldberg, A. (1989) FASEB J. 3, 2042-2051

4. Paulsson, M., and Heinegird, D. (1981) Biochem. J. 197,367-375 5. Pa&son, M., and Heinegird, D. (1982) Biochem. J. 207, 207-213 6. Paulsson, M., Inerot, S., and HeinegBrd, D. (1984) Biochem J. 221, 623-

630 7. Tsonis, P. A., and Goetinck, P. F. (1988) Exp, Eye Res. 46, 753-764 8. Saxne, T., and Heinegkrd, D. (1989) Arthritis Rheum. 32, 1080-1086 9. Argraves, W. S., Deik, F., Sparks, K. J., Kiss, I., and Goetinck, P. F. (1987)

Proc. N&l. Acad. Sci. (1. S. A. 84, 464-468 10. Kiss, I., De6k, F., Holloway, R. G., Jr., Delius, H., Mebust, K. A., Frimber-

ger, E., Argraves,. W. S., Tsonis, P. A., Winterbottom, N., and Goetnick, P. F. (1989) J. Bml. Chem. 264,8126-8134

11. Roth, G. J., Titani, K., Hoyer, L. W., and Hickey, M. J. (1986) Biochemistry 25,8357-8361

12. Paret], F. I., Niiya, K., McPherson, J. M., and Ruggeri, 2. M. (1987) J. Biol. Chem. 262, 13835-13841

13. Takada, Y., and Hemler, M. E. (1989) J. Cell Biol. 109,397-407 14. Mania&, T., Hardison, R. C., Lacy, E., Lauer, J., O’Connell, C., Quon, D.,

Sim, G. K., and Efstratidis, A. (1978) Cell 15, 687-701 15. Duby, A. D., and Seidman, J. G. (1986) Proc. Natl. Acad. Sci. U. S. A. 83,

4890-4894 16. Sambrook, J., Fritsch, E. F., and Maniatis, T. (1989) Molecular Cloning: A

Laboratory Manual, pp. 1.22-1.24, 2.82-2.84, 9.49, 12.13, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY

17. Sanger, F., Nicklen, S., and Co&on, A. R. (1977) Proc. Natl. Acad. Sci. U. S. A. 74,5463-5467

18. Queen, C., and Korn, L. J. (1984) ~u$eic Acids Res. 12,581-599 19. De3we~;;;J., Haeberli, P., and Snuthws, 0. (1984) Nucleic Acids Res. 12,

20. Shows, T. B., Sakaguchi, A. Y., and Naylor, S. L. (1982) Hun. Genet. 12, 341-452

21. Owerbach, D., Bell, G. I., Rutter, W. J., Brown, J. A., and Shows, T. B. (1981) Diabetes 30,267-270

22. Nakai, H., Byers, M. G., Shows, T. B., and Taggart, R. T. (1986) Cytogenet. CeUGenet.43,215-217

23. Zabel, B. U., Naylor, S. L., Sakaguchi, A. Y., Bell, G. I., and Shows, T. B. (1983) Proc. N&l. Acad. Sci. U. S. A. 80,6932-6936

24. Linsenmayer, T. F., and Little, C. D. (1978) Proc. Natl. Acad. Sci. U. S. A. 75,3235-3239

26. Kozak, M. (1987) Nucleic Acids Res. 15,8125-8148 27. van Heijne, G. (1986) Nucleic Acids Res. 14, 4683-4690 28. Folz, R. J., and Gordon, J. I. (1987) Biochem. Biophys. Res. Commun. 146,

870-877 29. Hughes, R. C. (1973) Prog. Biophys. Mol. Biol. 26, 189-268 30. Steiner, D. F., Quinn, P. S., Chan, S. J., Marsh, J., and Tager, H. S. (1980)

Ann. N. Y. Acad. Sci. 243, 1-16 31. Doolittle, R. F., Feng, D. F., and Johnson, M. S. (1984) Nature 307, 558-

560 32. Rees, D. J. G., Jones, I. M., Handford, P. A., Walter, S. J., Esnouf, M. P.,

Smith, K. J., and Brownlee, G. G. (1988) EMBO J. 7, 2053-2061 33. Shapiro, M. B., and Senapathy, P. (1987) Nucleic Acids Res. 15,7155-7174 34. Birnstiel, M. L., Busslinger, M., and Strub, K. (1985) Cell 41, 349-359 35. Man&as, T., Goodhourn, S., and Fischer, J. A. (1987) Science 236, 1237-

1245 36. Corden, J., Wasylyk, B., Buchwalder, A., Sassone-Corsi, P., Kedinger, C.,

and Chambon, P. (1980) Science 209, 1406-1414 37. Dynan, W. S., and Tijan, R. (1983) Cell 36, 79-87 38. Ruoslahti, E., and Pierschbacher, M. D. (1987) Science 238,491-497 39. Yamada, K. M., and Kennedy, D. W. (1987) J. Cell. Physiol. 130,21-28 40. Patthy, L. (1987) FEBS Lett 214, 1-7 41. Padgett, R. A., Grabowski, P. J., Konarska, M. M., Seiler, S., and Sharp,

P. A. (1986) Annu. Reu. Eiochem. 55, 1119-1150 42. Cech, T. R. (1988) Gene (Amst.) 73,259-271 43. Michel, F., Umesono, K., and Ozeki, H. (1989) Gene (Amst.) 82, 5-30 44. Osborne-Lawrence, S. L., Sinclair, A. K., Hicks, R. C., Lacey, S. W., Eddy,

R. L., Byers, M. G., Shows, T. B., and Duby, A. D. (1990) Genomics, in press

45. Myers, J. C., Dickson, W. A., de Wet, W. J., Bernard, M. P., Chu, M.-L., Di Liberto, M., Pepe, G., Sangiorgi, F. O., and Ramirez, F. (1983) J. Biol. Chem. 258, 10128-10135

46. Hamada, H., and Kakunaga, T. (1982) Nature 298,396-398 47. Hamada, H., Petrino, M. G., and Kakunaga, T. (1982) Proc. N&l. Acad.

Sci. I/. S. A. 79,6465-6469 48. Nordheim, A., and Rich, A. (1983) Nature 43,674-678 49. Bruns, G. A. P., and Sherman, S. L. (1989) Cytogenet. Cell Genet. 51, 67-

90 50. Tsonis, P. A., Argraves, W. S., and Millin, J. L. (1988) Biochem. J. 254,

623-624 51. Weiss, M. J., Henthorn, P. S., Lafferty, M. A., Slaughter, C., Raducha, M.,

and Harris, H. (1986) Proc. N&l. Acad. Sci. U. S. A. 83, 7182-7186 52. Shows, T., Eddy, R., Haley, L., Byers, M., Fujita, T., Matsui, H., and

Tanuguchi, T. (1984) Somat. Cell Mol. Genet. 10,315-318 53. Shows, T. B., Brown, J. A., Haley, L. L., Byers, M. G., Eddy, R. L., Cooper,

E. S., and Goggin, A. P. (1978) Cytogenet. Cell Genet. 21., 99-104 54. Shows, T. B. (1983) in Zsozymes: Current Topics in Biologrcal and Medical

Research (Rattazzi, M. C., Scandalios, J. G., and Whitt, G. S., eds) Vol. 10, pp. 323-339, Alan R. Liss, New York


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

Shows and A D DubyR N Jenkins, S L Osborne-Lawrence, A K Sinclair, R L Eddy, Jr, M G Byers, T B

protein.Structure and chromosomal location of the human gene encoding cartilage matrix

1990, 265:19624-19631.J. Biol. Chem.

http://www.jbc.org/content/265/32/19624Access 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/265/32/19624.full.html#ref-list-1

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


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/content/265/32/19624

http://www.jbc.org/cgi/alerts?alertType=citedby&addAlert=cited_by&cited_by_criteria_resid=jbc;265/32/19624&saveAlert=no&return-type=article&return_url=http://www.jbc.org/content/265/32/19624

http://www.jbc.org/cgi/alerts?alertType=correction&addAlert=correction&correction_criteria_value=265/32/19624&saveAlert=no&return-type=article&return_url=http://www.jbc.org/content/265/32/19624

http://www.jbc.org/cgi/alerts/etoc

http://www.jbc.org/content/265/32/19624.full.html#ref-list-1

http://www.jbc.org/

structure and chromosomal location of the human gene ... · lage. the structure and chromosomal...

Documents