THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1991 by The American Society for Biochemistry and Molecular Biology, Inc

Vol , 266, No. 7, Issue of March 5, pp. 4508-4513,1991 Printed in U. S. A.

The cul(VII1) Collagen Gene Is Homologous to the cul(X) Collagen Gene and Contains a Large Exon Encoding the Entire Triple Helical and Carboxyl- terminal Non-triple helical Domains of the cul(VII1) Polypeptide*

(Received for publication, October 1, 1990)

Noriko YamaguchiS, Richard Maynelj, and Yoshifumi NinomiyaSV From the $Department of Anatomy and Cellular Biology, Harvard Medical School, Boston, Massachusetts 02115-6092 and the §Department of Cell Biology, University of Alabama at Birmingham, UAB Station, Birmingham, Alabama 35294

We recently cloned and sequenced al(VII1) collagen cDNAs and demonstrated that type VI11 collagen is a short-chain collagen that contains both triple helical and carboxyl-terminal non-triple helical domains sim- ilar to those of type X collagen (Yamaguchi, N., Benya, P., van der Rest, M., and Ninomiya, Y. (1989) J. Biol. Chem. 264, 16022-16029). We report here on the structural organization of the gene encoding the rabbit al(VII1) collagen chain. The al(VII1) gene contains four exons, whose sizes are 69, 120, 331, and 2278 base pairs. The first and second exons encode only 5’- untranslated sequences, whereas the third exon codes for a very short (3 nucleotides) stretch of 5”untrans- lated sequence, the signal peptide, and almost the en- tire amino-terminal non-triple helical (NC2) domain (109% codons). Interestingly, the last exon encodes the rest of the translated region, including 7% codons of the NC2 domains, the complete triple helical domain (COLl, 454 amino acid residues), the entire carboxyl- terminal non-triple helical domain (NC1, 173 amino acid residues), and the 3”untranslated region. This exon-intron structure is in stark contrast to the multi- exon structure of the fibrillar collagen (types I, 11,111, V, and XI) genes, but it is remarkably similar to that of the type X collagen gene (LuValle, P., Ninomiya, Y., Rosenblum, N. D., and Olsen, B. R. (1988) J. Biol. Chem. 263, 18278-18385). The data suggest that the al(VII1) and the a l (X) genes belong to the same sub- class within the collagen family and that they arose from a common evolutionary precursor.

Type VI11 collagen was initially detected in biosynthetic studies of bovine aortic (1) and rabbit cornea ( 2 ) endothelial cells. Subsequently its presence in culture media of various endothelial cell lines was demonstrated, and it was detected as a biosynthetic product of both normal and malignant cell lines of nonendothelial origin (3-5). Recent immunofluores- cent studies using monoclonal (6,7) and polyclonal antibodies (8,9) indicated that type VI11 collagen is localized in special-

* This work was supported by National Institutes of Health Re- search Grants EY07334 (to Y. N.), AR32984 and AR30481 (to R. M.). 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 sequence(s) reported in this paper has been submitted to the GenBankTM/EMBL Data Bank with accession number(s) M58483.

!I To whom correspondence should be addressed.

ized extracellular matrices such as the cornea stroma, Desce- met’s membrane, sclera, choroid, optic nerve sheath and its septum, dura mater of the spinal cord, periosteum, and peri- chondrium. In addition, Kittelberger et al. (8) recently re- ported the presence of the collagen in the subendothelial region of most blood vessels.

In spite of this wide tissue distribution, the molecular structure of type VI11 collagen has been controversial. Re- cently, we demonstrated by cDNA cloning that one of the chains of type VI11 collagen is a relatively short polypeptide (744 amino acid residues) (10). The chain was given the designation al(VII1). Of great interest was the finding that the length of the triple helical domain and the number and relative locations of the 8 imperfections found within the al(VII1) chain are strikingly similar to those of chicken d ( X ) collagen which is expressed exclusively in hypertrophic chon- drocytes. Furthermore, the primary structure of the al(VII1) chain is remarkably similar to that of the al(X) chain. The structure of the chicken type X gene is unique among mem- bers of the collagen supergene family in that it contains only three exons, one of which is large (2137 bp)’ and encodes the entire triple helical domain (11, 12).

These findings prompted us to analyze the structural or- ganization of the gene for the al(VII1) chain. In this paper we show that the al(VII1) collagen gene contains four exons. The fourth exon is large (2278 bp) and codes for the entire triple helical domain (COLl), the carboxyl-terminal non- triple helical domain (NCl), and the 3”untranslated region. Therefore, it seems likely that the d(VII1) and the d ( X ) genes are homologous and belong to the same subclass within the collagen gene family.

MATERIALS AND METHODS

Cornea Endothelial Cell Culture and RNA Preparation-Cornea endothelial cells were isolated from young rabbit eyes and cultured as described (13, 14).

Total cellular RNA was isolated from confluent endothelial cell cultures with the guanidinium thiocyanate method (15). Oligo(dT)- cellulose chromatography was used to isolate poly(A)+ RNA. The mRNA activity was assayed by cell-free translation using a reticulo- cyte lysate (Amersham Corp.) as described (11).

Rotary Shadowing Electron Microscopy-Confluent cornea endo- thelial cells were incubated in serum-free Dulbecco’s modified Eagle’s medium for 16 h. Medium proteins were precipitated by 30% ammo- nium sulfate. The precipitate was dissolved in 0.2 M ammonium bicarbonate containing protease inhibitors. Rotary shadowing was performed as described previously (16).

cDNA Synthesis and Cloning-cDNA was synthesized with an

The abbreviations used are: bp, base pair; kb, kilobases; SDS, sodium dodecyl sulfate.

4508

by guest on February 10, 2018http://w

ww

.jbc.org/D

ownloaded from

The al(VIII) Collagen Gene 4509 oligo(dT) primer and poly(A)+ RNA isolated from rabbit cornea endothelial cells using the Amersham cDNA synthesis ki t (17). Dou- ble-stranded cDNA was blunt-ended with T4 DNA polymerase (Amersham), methylated, and ligated into X g t l O arms (Promega Biotech) via EcoRI linkers (New England Biolabs). Recombinant DNA was packaged and plated in the C600 strain (18). The cDNA library thus prepared was screened using a fragment of the cDNA pCE1230 as probe in the presence of N-laurylsarcosine (19).

Isolation of Rabbit Genomic DNA and Construction of Genomic Libraries-Splenocytes were collected from several rabbit spleens. High molecular weight cellular DNA was purified from splenocytes by a proteinase K/phenol extraction method as described (20). South- ern blotting was performed as described (21). Subgenomic libraries were constructed according to the procedure described by Maniatis et al. (22). Based on preliminary results of Southern blotting 2.0-, 4.7-, and 9.6-kb fractions of EcoRI-digested rabbit genomic DNA were separated by agarose gel electrophoresis and recovered by elec- trophoretic elution. Two fractions (2.0 and 4.7 kb) were ligated into the X g t l O arms digested with EcoRI (Amersham). The 9.6-kb fraction was ligated into the XgtWES-EcoRI vector (BRL). A commercial packaging extract (Promega) was used. Efficiencies of 3 X lo7 and IO9 plaque-forming units/ml were obtained from X-WES and X g t l O libraries, respectively.

Screening of Genomic Libraries-The two libraries thus prepared were screened with the insert of the cDNA NK1 as probe using standard hybridization (5 X SSC, 0.1% N-laurylsarcosine at 65 "C, overnight) and washing (3 X SSC, 0.1% N-laurylsarcosine at 65 "C for 30 min twice; (1 x SSC, 0.15 M NaCI, 0.015 M sodium citrate, pH 7.0) conditions (19). Two clones GY1 and GY4 (Fig. 3) were isolated from these two libraries and identified as gene fragments encoding parts of the al(VII1) chain (see "Results"). To isolate overlapping clones coding for the entire al(VII1) chain, we screened a genomic library (Cat. TL1003d) purchased from Clontech using the inserts of pCE1230 and GY4 (2 kb) as probes. GY5 and GY6 clones were isolated with the GY4 probe, whereas GY2 and GY3 were isolated with the insert of pCE1230. The most 5' gene clone (GY7) was isolated by screening the Clontech library using a 5' Pst-AluI (60 bp) restriction fragment of pCE1230. For this screening we used a stand- ard formamide protocol for hybridization (50% formamide, 0.1% SDS, 5 X Denhardt's (0.02% Ficoll, 0.02% polyvinylpyrrolidone, 0.02% bovine serum) solution, 5 X SSPE (1 X SSPE, 0.15 M NaCI, 10 mM sodium phosphate, 1 mM EDTA, pH 7.4); and 100 pg/ml salmon sperm DNA at 42 "C overnight) and washing (2 X SSC and 0.05% SDS at room temperature for 1 h, and 1 X SSC and 0.1% SDS at 68 "C for 1 h twice). Isolation and purification of phage DNA was performed according to standard methods (22). Restriction enzyme fragments of isolated clones were subcloned into pBluescript vectors. Nucleotide sequencing analysis was performed using the dideoxy chain termination technique (23) either on double-stranded DNA in pBluescript vectors (24) or in the single-stranded M13 vectors mp18 or mp19 (25).

Primer Extension Analysis-Primer extension analysis was used to find out where the 5' end of al(VII1) mRNA is. Two neighboring restriction fragments NlaIV-AluI (40 nucleotides) and AluI-PuuII (147 nucleotides) located close to the 5' end within the pCE1230 insert were labeled at their 5' ends by ["PIdATP and T4 polynucle- otide kinase, denatured, and annealed to 10 pg of total RNA prepared from cultured rabbit cornea endothelial cells (26). The two primers were extended with reverse transcriptase and the products were analyzed on 5% sequencing gels containing urea. 4x174 DNA di- gested with HaeIII was used for size markers.

RESULTS

Rotary Shadowing-Biosynthetic products of rabbit cornea endothelial cells in culture were examined by electron micro- scope after rotary shadowing. Fig. 1 shows electron micro- graphs of short molecules which were the major collagenous molecules observed in the preparation. The molecules had a rod-like structure with a prominent knob a t one end. A second smaller knob was sometimes observed at the other end. The central collagenous region was 132.5 k 5 nm ( n = 90) in length.

Genomic Southern Blot Analysis Using the al(VIII) cDNA Insert as Probe-cDNA sequence analysis demonstrated that the primary structure of the al(VII1) chain is strikingly

FIG. 1. Rotary shadowing images of type VI11 collagen mol- ecules. Medium proteins from cultured cornea endothelial cells were precipitated by ammonium sulfate and used for rotary shadowing analysis (see "Materials and Methods"). The major component con- sists of a 132.5 f 5-nm long rod-shaped structure with a large knob at one end and a smaller knob at the other end. Bar = 100 nm.

A B SOUMERN BLOT -" - ORlGH

al(VQ ~ 1 4 2 % "'AA %r+,,CooH

v dm. 9.4

. 23.1 kb

cDNAs 5' , NE1230 ! E N K 1 1 . 6.7

PA I 3' . 4.4 E

PROBE 1 3 . 2.3

I 1 4

E I 2.0

b 1367

3 b 2232 W 0 9 1.1

PROBE: 1 2 3

FIG. 2. Genomic Southern blot analysis using various re- gions of a cDNA coding for the al(VII1) chain. A, relative locations of al(VII1) coding regions, two cDNAs pCE1230, and NK1, and the probes 1, 2, and 3 used for genomic Southern blot analysis. Locations of restriction sites are indicated by capital letters of A (AluI), E (EcoRI), and P (PstI). B, an EcoRI digest of rabbit genomic DNA (20 pgllane) was applied onto 1% agarose gel, electrophoresed at 25 V overnight, and blotted onto nitrocellulose. EcoRI digestion was performed with 3 units of enzyme/pg of DNA overnight at 37 "C. The three blots were hybridized to three different '*P-labeled probes covering various regions of the cDNA pCE1230. Three bands of 9.6, 4.7, and 2.0 kb were recognized by the probes.

similar to that of the al(X) chain (10). The structure of the chicken type X gene is unique among collagen genes in that it is small in size (only 5 kb) and contains only three exons, one of which is large (2137 bp) and encodes the entire triple helical domain COLI (11, 12). This prompted us to analyze the structure of the al(VII1) gene. As shown in Fig. 2, South- ern blot analysis of rabbit genomic DNA completely digested with EcoRI demonstrated that the insert of the al(VII1) cDNA pCE1230 hybridized to three bands of 9.6, 4.7, and 2 kb, depending on the region used as probe. This indicates that the al(VII1) gene is longer than the type X gene.

Isolation of Genomic Clones-A X g t l O library containing 4.7-kb EcoRI fragments of rabbit genomic DNA (see "Mate- rials and Methods") was first screened with the whole insert of NK1 as probe. Screening of 3.8 x lo4 plaques led to the isolation of GY1 (see Fig. 3). Nucleotide sequence analysis of GY 1 demonstrated that it contains a large exon which encodes half of the COLI domain and the entire NC1 domain of al(VII1) chains, suggesting that the type VI11 gene structure is similar to that of type X. To obtain overlapping clones, we screened a genomic EMBL3 library using the inset of pCE1230 as probe, and isolated two overlapping clones, GY2 (16 kb) and GY3 (16 kb). Nucleotide sequence analysis of a HindIII-EcoRI (1.3 kb) fragment of GY3 located in the middle

by guest on February 10, 2018http://w

ww

.jbc.org/D

ownloaded from

4510 The a1 (VIII) Collagen Gene

3‘

.... . ’ I ’ m R N A

FIG. 3. Diagram showing the structure of the crl(VII1) gene and the relative locations of genomic DNA clones GY1-GY7. The al(VII1) gene contains 4 exons separated by 3 introns. Clones GY6 and 7 do not overlap, neither do clones GY2 and 5, indicating that introns 1 and 2 are large, while intron 3 is 4.7 kb. Exons 1 and 2 code for the 5”untranslated region, whereas the translated portion of al(VII1) mRNA is encoded by exons 3 and 4. Exon 4 is a large exon and encodes part of the NC2 domain (only 7% codons), the entire triple helical (COL1, hatched box) and 3’ non-triple helical (NC1) domains, and the 3”untranslated region. For the sake of comparison, the structure of the chicken type X gene is drawn in the left bottom inset on the same scale. Note the striking similarity between the two genes as well as in the domain structure of their protein products.

of its insert demonstrated that the large exon continues 859 bases further in the 5’ direction. Consequently, this large exon in the al(VII1) gene contains 2278 bp of DNA and codes for 7% codons of the NC2 domain, the whole COLl (454 amino acid residues) domain, the NC1 (173 amino acid resi- dues) region, and the 3”untranslated region. Upstream of the large exon, a second exon was found in GY3 and GY2, separated by 4700 bp from the large exon. The second exon encodes almost the entire NC2 domain of 109% codons.

Since a “gene walking” approach was unsuccessful in the isolation of further upstream fragments of the al(VII1) gene, we isolated a 2-kb EcoRI fragment (GY4) from the X g t l O library (see “Materials and Methods”) and used the insert of GY4 as probe to screen the EMBL3 library. This led to the isolation of GY5 and GY6. DNA sequence analysis demon- strated that the 2-kb clone GY4 contained one small exon of 120 bp which encodes two-thirds of the 5‘-untranslated re- gion. Finally, we used the most 5’ region (Pst-AluI fragment, 60 bp in length) of the pCE1230 insert as probe for screening the EMBL library, resulting in the isolation of the 16-kb-long clone, GY7.

It is now possible to compare the hybridized EcoRI frag- ments (9.6, 4.7, and 2.0 kb in Fig. 2) with various probes to EcoRI fragments (in Fig. 3) generated from the genomic clones GY1-GY7. The fragment of 9.6 kb size-hybridized with probe 1 is located most 5’ in the map in Fig. 3. Part of this fragment lies within GY7, and part of it lies outside the 5’ end of GY7. Probe 2 hybridized with two bands of 2.0 and 9.6 kb. The 2.0-kb fragment is GY4 and the 2.0-kb EcoRI frag- ments contain exon 2 within GY5 and GY6. The 9.6-kb band recognized by probe 2 consists of two different EcoRI frag- ments of the same size. One fragment must be the 9.6-kb fragment that is located at the 5‘ end of GY7 and is recognized also by probe 1. A second 9.6-kb EcoRI fragment recognized by probe 2, but not by probe 1 (compare intensities and thickness of 9.6-kb bands in lanes 1 and 2 of Fig. 2B) must extend from the EcoRI site in GY2 to the EcoRI site in the middle of GY3. This fragment contains exon 3 and part of exon 4. Finally, the 4.7-kb fragment hybridized to probe 3 corresponds to the GY1 insert and the 4.7-kb fragment in

GY3. This fragment contains the rest of exon 4. The relative alignment of the 7 genomic clones containing

the rabbit al(VII1) collagen gene is shown in Fig. 3. Exons are numbered from the 5’ end. The clones cover approxi- mately 61 kb including 3 kb upstream and 5 kb downstream of the gene. However, there are gaps between GY7 and GY6 and between GY5 and GY2, both in intron sequences. There- fore, we estimate the entire length of the rabbit al(V1II) gene to be at least 53 kb.

Exon Sizes-The al(VII1) gene contains only 4 exons. This is in contrast to the multi-exon structure of fibrillar collagen genes (27), but quite similar to the al(X) collagen gene (12). The sizes of the exons are 69, 120, 331, and 2278 bp. The exon/intron junctions are shown in Table I. The first and second exons encode only the 5‘-untranslated region, whereas the third exon codes for part (3 nucleotides) of the 5’-untrans- lated region, the signal peptide, and almost the entire NC2 domain (total 109% codons). Interestingly, the fourth exon codes for part (7243 codons) of the NC2 domain, the whole triple helical domain (COL1, 454 amino acid residues), the whole carboxyl-terminal non-triple helical domain (NC1, 173 amino acid residues), and the 3”untranslated region. The chicken type X collagen gene (12) has a similar structure with 3 exons, one of which encodes a portion (4243 codons) of the NC2 domain, the entire triple helical domain (COL1, 460 amino acid residues), the entire carboxyl-terminal non-triple helical domain (NC1, 162 amino acid residues), and the 3’- untranslated region.

Sequences determined at the boundaries between exons and introns follow the consensus rules with a minor exception (Table I). The introns all contain AG and GT consensus sequences at the 3’ and 5‘ splice junctions except for intron 2 (CA instead of GT). An acceptor splice junction is found to split the codon (GAA) of the glutamic acid residue at the end of exon 3 and the beginning of exon 4. A similar split codon for a glycine residue has been found in the type X collagen gene (12).

Characterization of the 5‘ Region of the al(VIII) Collagen mRNA-To find out whether the cDNA pCE1230 extended to the 5‘ end of the al(VII1) mRNA, primer extension analysis was performed. Two restriction fragments, NlaIV-AluI (40 nucleotides) and AluI-PuuII (147 nucleotides), were labeled at their 5’ ends, denatured, and annealed to endothelial cell RNA. The two primers were extended by reverse transcriptase and the products were analyzed on polyacrylamide gels. AS shown in Fig. 4, primer 1 was extended to a band at 74 nucleotides, whereas primer 2 was extended to 219 nucleo- tides. The result indicates that the 5’ end of al(VII1) mRNA is 12 nucleotides upstream of the 5’ end of the cDNA pCE1230. The transcription start site is therefore probably the A at nucleotide number 1 in Fig. 5, making exon 1 69 bp long.

The 5’-flanking region of the gene contains an ATAAAA box at -31 to -26 upstream of exon 1. A GC-rich (72% GC) sequence of 186 nucleotides containing two CCCGCC boxes is located upstream of the ATAAAA box. Two other inverted sequences (GGCGGG) are found at -540 and -520 within another GC-rich region (-497 to -600). These GC boxes are potential sites for the binding of transcription factor s p l (28).

Three Different mRNA Species Are Due to Three Alternative Poly(A) Signals”pCE1230 showed strong hybridization to an mRNA species of about 2800 nucleotides and weaker hybrid- ization to two larger mRNAs that migrate around the 28 s rRNA marker. The sizes of the two larger mRNA species are estimated to be 4200 and 3600 nucleotides. To examine the possibility that the three different mRNAs are due to different

by guest on February 10, 2018http://w

ww

.jbc.org/D

ownloaded from

The al(VIII) Collagen Gene 4511 TABLE I

Exon-intron boundaries of the gene coding for al(VIII) chain The numbers above nucleotides are from the cDNA pCE1230 sequence (Ref. 10). Exons 3 and 4 contain the

translated region startine with the underlined ATG at 181-183. Exon No. Exon-intron boundaries

1 57

1

2

3

4

Consensus sequence

ATCAGAGCCCTTCCC " " _ CCTTACGTGCCCAAG 58 177

gtaagaccacaca tg

c t t c t c t a c c t t c a g AAGCTGTTCTGGAAG CGGCCAAGTCTAAAT c a t t c a c t c c g c g g t 178

" " _ 508

t c t t t c c t t c c a c a g GTGCGGCCGTGCCA " " _ CCCAAGAAAGGCAAAG g t a c a c c a t c c t c c t M A V P P K K G K

509

t c c t c c t c c t c t c a g AAATACCATTAGCC (E) I P L A

" " _ cag AG g t a a g t

A 6

e p 4 NE1230 p

7, i..... -1 219 ................. . 7 i E P RineR

(147 m) b

- 194

118

72

FIG. 4. Primer extension analysis using fragments of al(VII1) cDNA. The relative location of exon 1 (thick closed boz), the 5' region of the mRNA, the cDNA pCE1230, and primers 1 and 2 used for primer extension are shown in A. Primer 1 (40 nucleotides long, a NluIV(N)-AluI(A) fragment) and primer 2 (147 nucleotides long, an AluI(A)-PuuII(P,) fragment) were annealed to mRNA and extended by reverse transcriptase. The products were analyzed on a 5% sequencing gel in denaturing conditions (shown in B ) . Primer 1 and primer 2 were extended to 74 nucleotides (lam I ) and 219 nucleotides ( l a n e 2), respectively. Considering the relative location of the two primers, the sizes of the two extended products correspond to the same 5' end point. This indicates that the 5' end of the cDNA pCE1230 is 15 nucleotides downstream of the 5' al(VII1) mRNA, and predicts that exon 1 is 69 bp long. The P shown at the 5' end of pCE1230 and the two vertical bars next to the P indicate a PstI site and stretch of G. for cloning. Upstream of exon 1, the position of an ATAAAA box is indicated by a small horizontal bar. The first and last lanes (indicated by S) contain HaeIII fragments of @X174 DNA as size markers. The sizes of the markers in nucleotides are indicated a t the right-hand side of the figure. The heavy bands at the bottom in lune I and above and below the 118-nucleotide marker in lune 2 represent excess primers in single-stranded and double-stranded form. nt, nucleotide.

poly(A) signals as found in COLlA2 (29) or in the al(1X) gene (30), we took two different approaches. First, the nucleo- tide sequence of the 5' half (EcoRI-HindIII) of the genomic fragment GY1 (see Fig. 6) was determined. Two potential poly(A) signals both TATAAA (1415 and 1894 bp away from the EcoRI site, respectively), were found in this 2.3-kb se- quence. The possible sizes of the putative two mRNA species would be larger than 2794 and 3273 nucleotides, depending on the length of the poly(A) segment. These two sizes are in good agreement with the calculated sizes of the two mRNA

species at 2800 and 3600 nucleotides. Because there are no additional potential poly(A) signals within the sequence be- tween the EcoRI and the HindIII sites in GY1, a third signal could be beyond the HindIII site, indicating that the third mRNA species could be larger than 3621 nucleotides. Second, we prepared three different hybridization probes (probes 1,2, and 3 in Fig. 6) from GY1 and used them for Northern blot analysis. As shown in Fig. 6B, probe 1 located most 5' hybrid- izes to all three species of mRNA, whereas only the two larger mRNA species were detected by probe 2 and none hybridized to probe 3. This result is consistent with the sequencing analysis of the gene, and suggests that a third poly(A) signal is located within the sequence between the HindIII and PuuII sites in GY1. Taking all these results into account, we con- clude that the three different mRNA species are due to different poly(A) signals.

DISCUSSION

The Translation Product of the al(VIII) Gene Has a Struc- ture Similar to That of the al(X) Gene-In a previous paper (lo), we demonstrated that the putative translation product deduced from al(VII1) cDNA sequences has a structure sim- ilar to that of the d ( X ) collagen chain (11). We also dem- onstrated that the al(VII1) collagen chain is 60 kDa in size and suggested that the product is a major constituent of the hexagonal lattice structure of Descemet's membrane (10). The predicted length of a triple helical domain with 454 amino acid residues is 129.8 nm based on a value of 0.286 nmlresidue for collagen triple helices (31). In the present report, we observed biosynthetic products of rabbit endothelial cells by electron microscopy. A thin rod-like structure with tiny knobs at both ends is seen in Fig. 1. The length of the rod-like structure of the molecule is 132.5 f 5 nm ( n = 90), consistent with the prediction from the cDNA analysis. The size differ- ences of the two knobs at the ends can be explained by the size differences of the amino (96 residues without signal peptide) and the carboxyl (173 residues) non-triple helical domains.

We conclude therefore that the molecules seen in Fig. 1 represent type VI11 collagen molecules. Interestingly, type X collagen synthesized by hypertrophic chondrocytes also ap- pear as rods with a length of 132 nm and a knob at one end (16). The striking similarity between the primary structures of a l (X) and al(VII1) chains (10,l l) is therefore reflected in

by guest on February 10, 2018http://w

ww

.jbc.org/D

ownloaded from

4512 The al(VIII) Collagen Gene -810 -860 -850 -840 -830 -820

GAT CCT CTC T G A ATG TlT AM GAC ACA CAG AGC AAG AGC TOT GGC ACC AGG ACG TlT

-8 iO -800 -190 -180 -110 -160

TCC AAT TGT TAG AGA AGG AGA GCT GTP T M ACT CCG AM ACA ACC AGA GGG CAG AGA

-750 -740 -130 -120 -110 -100

GCC Toc AAT CCT 1\00 TPA CAT AAG CAG TOT CCG AGG AGC GAG GCT GGC TCC GAG GTP

-690 -680 -610 -660 -650

CTT TPC TOC AAG TOT CCA TCA GCG CAG ATC CTC CCC AAC CTL' GGC TOG ATP TAC CTG

-640 -630 -620 -610 -600 -590

AAT GCC GCC GCC GGC O K CCA AGA GGG GAC AGC OCA CAC TAC AGG CAG CCC CAG GAG

- 5 p -510 -560 -550 -5f0 -530

GTG GTC CCG OTC CAG CGA GCG GCC GGC CCG OTO Too AGG A C C m T o c K C AGC

-5;a -510 -500 -490 -480

GGG TbG CGG @G CTG CAG CAC CAG TCG GTC CCC Tpo GCT CCA GGT GCT GCT Tl!C CCA

-470 -460 -450 -440 -430 -420

GCT CCC TCT CCT CGC CTG GCT TAG GGA CAT GCA GCT CTC CTG AAC COT TAT TPA OCT

-410 -400 - 3 y -380 -310 -360

TCT W CTG GCA GAA GAC ACC TOT AGT,CAT M G AGG CAC TCC ACC AGA TCC CAA GGG

-350 -340 -330 -320 -310

TCA TGC AAG GGA GAA AM AM GAC AAT PAT GCA ATP CCC CCT TCT Tpo CTG GCA TAT

-300 -290 -280 -210 -260 -250

TCG TCT CCA CTC TOT C K CTL' Toc TCC CCC ACC m CTC CAG CTG Toc TCT AGA TCC

-2:o -230 -220 -210 -200 -190

CTG TCC CTA TlT CTC ATP. TTC TCT Tpo CCG GCT &C CGC &C GCT CCD GAT K C ACT

-180 -170 -160 -150 -1fQ -130

CCT TCC TGC GCT G M ACC CCC TCC TCC CTC CCG CGG GCT ACC GGT CTG CCC TCT CGC

-120 -1p -100 -90 -80

GCC K C CCA GAG AAG AGT GTC CGG CTC CTC CCG GGG CTC CCT o(cC CGC a C TCC TPC

-10 -60 -50 -40 -30 -20

CCT GCC CAC CTA CTA T G A GGA GCG ACT CCT CCC ACC TGC GCA TAA M G CCA AGT GCA

-10 1 io 20

TGT TAC CCG GCG CGG ATC AGA GCC CTP CCC CGG TCC TCT CCG TGG GAG CCC GCG AGC

Exon 1 l a 80 ? fa

CTC TCT GCC CGA CCT TAC GTG CCC AAG GTA AGA CCA CAC ATG Toc GGC AG

FIG. 5. Nucleotide sequence of the 5' region of the al(VII1) collagen gene. The sequence shown includes the promoter region, the first exon (indicated by boxed area), and part of the first intron. Numbering starts at the 5' end of exon 1. Two CCCGCC boxes and an AATAAA box are indicated by boxes and by an underlined se- quence, respectively.

a similar molecular structure as visualized by electron mi- croscopy.

One h r g e Exon Encodes the Entire Triple Helical Domain of the al(VZZZ) Chin-The results presented here provide the first description of the exon structure of the d(VII1) gene (Fig. 3). Of special interest is that exon 4 codes for the entire triple helical domain. In fibrillar collagen (type I, 11, 111, V, and XI) genes, the coding region for the triple helical domain is split into many exons (for example, 44 exons in a2(I) gene (32)) of characteristic sizes (frequently 54 or 108 bp (27)). Other collagen genes such as types IX and XI1 (33), IV, and VI have many exons as well, although their exon structure is different from that of fibrillar collagen genes. Only the type X collagen gene (11,17) is similar to the al(VII1) gene in that it contains a large exon (2137 bp) that encodes the entire triple helix (COL1) and carboxyl non-triple helical domain (NC1) (Fig. 3). As seen in Fig. 7, the relative locations of

A

cDNAs 0CE1230 *

0 Nathem Mot

. 1 kb . I

1 2-3

FIG. 6. Three mRNA species are due to differential utiliza- tion of poly(A) sites. A, diagram showing the relative locations within the 3' region of the al(VII1) gene of exon 4, four cDNA clones (pCE1230, NK1, NK25, arid NK40), an EcoRI gene fragment (GYl), and three probes (probes 1-3) derived from different parts of the GY1 fragment. B, probes 1-3 were used for Northern blot analysis. Probe 1 hybridized to three different mRNA species of 4200,3600, and 2900 nucleotides, whereas only the upper two mRNA species of 4200 and 3600 nucleotides were recognized by probe 2. No mRNA species hybridized to probe 3.

acu 12 3 4

FIG. 7. Diagram showing the relative locations of introns in al(X) and (rl(VII1) mRNAs. The translated regions are indi- cated by boxed regions, whereas the 5'- and 3"untranslated regions are shown by thin lines. The central triple helical domains (filled with dots) are flanked by NC1 and NC2 non-triple helical domains (black areas). Intron locations are indicated by vertical lines with closed triangles at the ends. Exon sizes are shown by nucleotide numbers. The numbers within parentheses are the number of amino acid residues within each domain. Note that the relative locations of introns are similar in the al(X) and the al(VII1) genes except for the presence of the first intron in the al(VII1) gene.

introns are similar in the a l (X) and al(VII1) genes except for the presence of an additional intron in the al(VII1) gene. This intron is located in the 5"untranslated region in the al(VII1) gene.

The lack of introns within the region that codes for the triple helical domain in both the al(VII1) and al(X) genes leads us to place these two genes in the same subclass within the collagen gene superfamily. We suggest that the designa- tion short-chain collagen be used for this subclass. Type VI collagen was initially designated a short-chain collagen be- cause of its short triple helical region; however, subsequent molecular cloning demonstrated that type VI chains are not really short, despite their short triple helical domains (34). Moreover, the exon structure of type VI genes is entirely different from that of the al(VII1) and al(X) genes. There- fore, type VI collagen does not belong in the short-chain collagen group. We have previously included type IX collagen among the short-chain collagens (35), but since the type IX genes are large multi-exon genes, they clearly do not belong in the same subclass as type VI11 and X. Therefore, the term short-chain collagen genes should be restricted to genes that are homologous to the a l (X) and al(VII1) genes.

Acknowledgments-We are grateful to Dr. Bjorn R. Olsen for his encouraging discussions and support throughout the course of this work. We would like to thank David W. Wright for his technical assistance and Masha Rodin for her excellent secretarial assistance.

by guest on February 10, 2018http://w

ww

.jbc.org/D

ownloaded from

The a1 (VIII) Collagen Gene 4513 REFERENCES

1. Sage, H., Pritzl, P., and Bornstein, P. (1980) Biochemistry 1 9 ,

2. Benya, P. D. (1980) Renal Physiol. 3,30-35 3. Alitalo, K., Bornstein, P., Vaheri, A., and Sage, H. (1983) J. Biol.

4. Sage, H., Balian, G., Vogel, A. M., and Bornstein, P. (1984) Lab. Inuest. 50, 219-231

5 . Sage, H., and Bornstein, P. (1987) in Structure and Function of Collagen Types (Mayne, P., and Burgeson, R., eds) pp. 173- 194, Academic Press, London

6. Sawada, H., Konomi, H., and Hirosawa, K. (1990) J. Cell Biol.

7. Kapoor, R., Sakai, L. Y., Funk, S., Roux, E., Bornstein, P., and

8. Kittelberger, R., Davis, P. F., Flynn, D. W., and Greenhill, N. S.

9. Kapoor, R., Bornstein, P., and Sage, E. H. (1986) Biochemistry

10. Yamaguchi, N., Benya, P., van der Rest, M., and Ninomiya, Y. (1989) J. Biol. Chem. 264, 16022-16029

11. Ninomiya, Y., Gordon, M., van der Rest, M., Schmid, T., Linsen- mayer, T., and Olsen, B. R. (1986) J. Biol. Chem. 2 6 1 , 5041- 5050

12. LuValle, P., Ninomiya, Y., Rosenblum, N. D., and Olsen, B. R. (1988) J. Biol. Chem. 263, 18378-18385

13. Kay, E. P., Smith, R. E., and Nimni, M. E. (1982) J. Biol. Chem. 257 , 7116-7121

14. Kay, E. P., Smith, R. E., and Nimni, M. E. (1982) J. Biol. Chem.

15. Chirgwin, J. M., Przybyla, A. E., MacDonald, R. J., and Rutter,

16. Schmid, T. M., Mayne, R., Bruns, R. R., and Linsenmayer, T. F.

17. Gubler, V., and Hoffman, B. J. (1983) Gene (Amst.) 25,263-269 18. Huynh, T. V., Young, R. A,, and Davis, R. W. (1985) in DNA

Cloning: A Practial Approach (Glover, D. M., ed) Vol. I, pp. 49- 78, IRL Press, Oxford

5747-5755

Chem. 258,2653-2661

110,219-227

Sage, E. H. (1988) J. Cell Biol. 107 , 721-730

(1990) Connect. Tissue Res. 24,303-318

25,3930-3937

257,7116-7121

W. J. (1979) Biochemistry 18 , 5294-5299

(1984) J. Ultrastruct. Res. 86, 1286-1291

19. Overbeek, P. A., Merlino, G., Peters, N. K., Cohn, V. H., Moore, G. P., and Klinsmith, L. J. (1981) Biochem. Biophys. Acta 656,

20. Perbal, B. (1984) A Practical Guide to Molecular Biology, Wiley

21. Thomas, P. S. (1980) Proc. Natl. Acad. Sci. U. S. A. 7 7 , 5201- Interscience Publications, New York

5205 22. Maniatis, T., Fritsch, E. F., and Sambrook, J. (1982) Molecular

Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY

23. Sanger, F., Nicklen, S., and Coulson, A. R. (1977) Proc. Natl. Acad. Sci. U. S. A. 74,5463-5467

24. Hattori, M., and Sakaki, Y. (1986) Anal. Biochem. 162, 232-238 25. Norrander, J., Kempe, T., and Messing, J. (1983) Gene (Amst.)

26. Kohno, K., Sullivan, M., and Yamada, Y. (1985) J. Biol. Chem.

27. Vuorio, E., and de Crombrugghe, B. (1990) Annu. Reu. Biochem.

28. Kadonaga, J. T., Jones, K. A., and Tjian, R. (1986) Trends Biochem. Sci. 11 , 20-23

29. Myers, J . C., Dickson, L. A., de Wet, W. J., Bernard, M. P., Chu, M.-L., DiLiberto, M., Pepe, G., Sangiorgi, F. O., and Ramirez,

195-205

26,101-105

260,4441-4447

59,837-872

F. (1983) J. BWl. Chem. 258,10128-10135 30. Vasios, G. W. (1986) Ph.D. Thesis, Rutgers Medical School 31. Rich, A., and Crick, F. H. C. (1961) J. Mol. Biol. 3 , 483-506 32. Boedtker, H., Finer, M., and Aho, S. (1985). Ann. N. Y. Acad.

33. Ninomiya, Y., Castagnola, P., Gerecke, D., Gordon, M. K., Ja- cenko, O., LuValle, P., McCarthy, M., Muragaki, y., Nihimura, I., Oh, S., Rosenblum, N., Sato, N., Sugrue, S., Taylor, R., Vasios, G., Yamaguchi, N., and Olsen, B. R. (1990) in Extra- cellular Matrix Genes (Sandell, L., and Boyd, C., eds) pp. 79- 114, Academic Press, Orlando, FL

34. Chu, M.-L., Mann, K., Deutzmann, R., Pribula-Conway, D., Hsu- Chen, C.-C., Bernard, M. P., and Timpl, R. (1987) Eur. J. Biochem. 168 , 309-317

35. Olsen, B. R., Ninomiya, Y., Lozano, G., Konomi, H., Gordon, M., Green, G., Parsons, J., Seyer, J., Thompson, H., and Vasios, G. (1985) Ann. N. Y. Acad. Sci. U. S. A. 460 , 141-153

Sci. 460,85-116

by guest on February 10, 2018http://w

ww

.jbc.org/D

ownloaded from

N Yamaguchi, R Mayne and Y Ninomiyanon-triple helical domains of the alpha 1 (VIII) polypeptide.

and contains a large exon encoding the entire triple helical and carboxyl-terminal The alpha 1 (VIII) collagen gene is homologous to the alpha 1 (X) collagen gene

1991, 266:4508-4513.J. Biol. Chem. 

  http://www.jbc.org/content/266/7/4508Access 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/266/7/4508.full.html#ref-list-1

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

by guest on February 10, 2018http://w

ww

.jbc.org/D

ownloaded from