Embryonic chicken cornea and cartilage synthesize type IX collagen ...

Proc. Natl. Acad. Sci. USA Vol. 85, pp. 7496-7500, October 1988 Biochemistry

Embryonic chicken cornea and cartilage synthesize type IX collagen molecules with different amino-terminal domains (extracellular matrix)

KATHY K. SVOBODA*, ICHIRO NISHIMURA, STEPHEN P. SUGRUE, YOSHIFUMI NINOMIYA,

AND BJ$RN R. OLSENt

Department of Anatomy and Cellular Biology, Harvard Medical School, Boston, MA 02115

Communicated by Elizabeth D. Hay, July 1, 1988

ABSTRACT We have analyzed embryonic chicken cornea for the presence of type IX collagen mRNA and protein. Using RNA transfer blot analysis, we demonstrate that al(IX) and a2(IX) mRNAs are expressed by corneal epithelial cells at the time that the primary stromal components are synthesized. The levels of the mRNAs decrease with increasing developmental age and are barely detectable at day 11 of development. In contrast, type IX collagen protein is detectable by immunefluorescence at days 5 and 6 and undetectable by day 8. Using probes specific for al(IX) and a2(IX) mRNAs, we demonstrate that the size of a2(IX) mRNA is the same in cornea as in chondrocytes, the major source of type IX collagen. However, the al(IX) mRNA is about 700 nucleotides shorter in the cornea than in cartilage because the corneal form of the mRNA does not contain the 5' region that encodes the non-triple-helical amino-terminal globular domain of cartilage type IX collagen. Therefore, corneal type IX collagen must lack this domain. This structural modulation of an extracellular matrix protein is likely to contribute to the functional differences between cartilage matrix and the early corneal stroma, both of which are rich in type II collagen.

type lI collagen

& we IX collagen

---

are members of a class of vertebrate proteins that are associated with cross-striated collagen fibrils in extracellular matrices (1-4). Substantial structural information based on DNA and peptide sequence analysis is available for cartilage type IX collagen and it serves therefore as prototype for the group. Type IX molecules are distinctly different from those of fibril-forming collagens in that they contain three triple-helical domains interrupted by non-triple-helical regions (2). Of the three genetically distinct subunits of type IX collagen, al(IX) contains a large amino-terminal globular domain (NC4) of almost 250 amino acid residues (5), while a second subunit, a2(IX), contains a chondroitin sulfate chain covalently linked to the non-triple-helical region that separates the aminoterminal and central triple-helical domains (6-11). This nontriple-helical domain (NC3) represents a flexible hinge region that allows type IX collagen molecules, located in a periodic manner along type II-containing cartilage collagen fibrils, to project their amino-terminal domains out from the surface of the fibrils into the surrounding matrix (12-14) (Fig. 1). These findings suggest that the amino-terminal non-triplehelical domain of type IX collagen represents a site of interaction between collagen fibrils and the proteoglycan-rich matrix in cartilage. We recently determined the complete primary structure of this domain (5) and find that it is a relatively basic protein (estimated pI is 9.7). This supports the hypothesis that the domain could be involved in ionic interactions with negatively charged glycosaminoglycans.

FIG. 1. Diagram showing periodic cross-striated collagen fibrils in cartilage with the location of type II and IX collagen molecules relative to the overlap (gray) and gap (white) zones of the fibrils. The type II molecules extend over almost 4.5 periods and are represented by straight rods. Type IX molecules have three triple-helical domains separated by non-triple-helical domains. Due to the flexible hinge at one of these non-triple-helical domains (NC3), the amino-terminal triple-helical domain and the amino-terminal globular domain (NC4) project out from the surface of the fibrils. In one of the fibrils a crosslink is shown between the amino telopeptide of a type II molecule and a sequence close to the NC3 domain of a type IX molecule. For simplicity, the glycosaminoglycan side chain that is attached to the NC3 domain is not indicated. Also, to better show the periodic cross-striated nature of the fibrils, their diameters are somewhat exaggerated.

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

*Present address: Department of Anatomy, Boston University School of Medicine, Boston, MA 02118. tTo whom reprint requests should be addressed.

Types IX and XII collagens

In the present study we have examined a non-cartilage tissue rich in type II collagen, the chicken primary corneal stroma, for the presence of type IX collagen. This primary stroma is the product of corneal epithelial cells from the third to the sixth day of development. It consists of striated fibrils containing a mixture of type I and type II collagens (15, 16), which are arranged in an orthogonal pattern (17), and does not contain the large aggregating proteoglycan found in cartilage. We reasoned that since the primary stroma contains type II collagen, type IX collagen may also be present. We demonstrate here that al(IX) and a2(IX) mRNAs and type IX collagen are indeed expressed by corneal epithelial cells at the time that the primary stromal components are synthesized. We show that while the size of a2(IX) mRNA is

7496

Biochemistry: Svoboda et al. identical in cornea and in cartilage, the al(IX) mRNA is significantly smaller in cornea than in cartilage because several hundred nucleotides are absent at the 5' end of the miRNA encoding the major portion of the amino-terminal globular domain. Our results indicate that tissue-specific variations at the 5' end of al(IX) mRNAs lead to the synthesis of type IX molecules with different amino-terminal domains.

MATERIALS AND METHODS Preparation of RNA. RNA was extracted from whole corneas of 6-, 8-, and 11-day-old chicken embryos and from sternal cartilage of 17-day-old embryos by using guanidine hydrochloride (18) or guanidine thiocyanate (19). Poly(A)+ RNA was separated from rRNA by oligo(dT)-cellulose chromatography. Corneal epithelia were isolated with the basal lamina intact by using Dispase II (20). RNA extracts were prepared (21) and the RNA was bound to nitrocellulose filters in a dot blot apparatus (Schleicher & Schuell). The blots were-hybridized with probes specific for al(IX) and a2(IX) mRNAs (see below). RNA Transfer Blot Analysis. For gel blot analysis, poly(A)+ RNA or total RNA was electrophoresed in 1% agarose gels in the presence of formamide/formaldehyde and transferred to nitrocellulose (BA-85, Schleicher & Schuell) (22). Hybridization probes were nick-translated restriction fragments of the cDNAs pYN1738 [al(IX)] (1), pYN1731 [a2(IX)] (23), pYN2142 [al(II)] (24), and IN321 [al(IX)] (5). As control probe for detection of an mRNA of the same size in cartilage and cornea extracts, we used cDNA specific for chicken f3-actin (25). Immunofluorescence. For immunolocalization of type IX collagen in cornea, we used two monoclonal antibodies (2C2 and 4D6, provided by R. Mayne, University of Alabama at Birmingham) against chicken type IX collagen. The epitopes recognized by the antibodies are located within different parts of the triple-helical domains (11, 14). Cryostat sections of corneas obtained from 5-, 6-, 8-, and 11-day-old chicken embryos were digested with testicular hyaluronidase (3.3 mg/mil in phosphate buffer, pH 6.0) for 0.5 hr at 37°C. Incubations with primary antibody were for 1 hr at room temperature. After three 5-min washes in phosphatebuffered saline, the secondary antibody (fluorescein-conjugated goat anti-mouse IgG, Boehringer Mannheim) was applied (1:1000 dilution) for 1 hr.

RESULTS AND DISCUSSION Temporal Expression of Type II and IX Collagen mRNA in Chicken Cornea. RNA extracted from whole corneas of chicken embryos incubated for 6, 8, or 11 days and from sternal cartilage of 17-day-old embryos was used for transfer blot analysis. Early embryonic cornea was found to contain both type II and type IX mRNAs (Fig. 2). The levels of the mRNAs clearly decreased with increasing developmental age. This change in mRNA levels was especially striking for type IX mRNAs, which were barely detectable at day 11. In situ hybridization and immunohistochemistry with probes and antibodies specific for type II procollagen have shown that corneal epithelial cells represent the major source of type II collagen in the embryonic chicken cornea (26). Therefore, it seemed reasonable to assume that the epithelial cells would also be the major source of type IX collagen mRNA. To prove that the epithelium was indeed the source of type IX RNA, epithelium was isolated from 6-day-old embryos and RNA extracted and examined by slot blot analysis. The results (data not shown) showed that epithelial cells contain RNA sequences that hybridize to the type IX cDNA probes. This

Proc. Natl. Acad. Sci. USA 85 (1988)

s 6 8 11

s 6 8 11

7497

s 6 8 11

e27s-y 18S-

0(111XJ

. We

0(211XI

0(11111

FIG. 2. Transfer blot analysis of poly(A)+ RNA extracted from 17-day-old embryonic chicken sternal cartilage (lanes s) and total RNA from 6-, 8-, and 11-day-old embryonic chicken cornea. The blots were hybridized to cDNAs specific for al(IX), a2(IX), and al(II) collagen chains. The positions of 27S and 18S rRNA markers are indicated on the left. One hundred nanograms of poly(A) + RNA was loaded in each of the cartilage lanes and 5 j/g of total RNA was loaded in each of the cornea lanes. The blots hybridized to the type IX collagen probes were exposed for 41 hr, whereas the blot with the al(II) collagen probe was exposed for 5 hr.

does not exclude, of course, expression of type IX genes in other ocular cells, such as corneal endothelial cells. Corneal al(IX) mRNA Is Smaller Than Cartilage al(IX) mRNA. Transfer blots with sternal cartilage RNA show that al(IX) mRNA migrates as two bands in agarose gels at about the position of 27S rRNA (Fig. 2). The difference in size between the two bands is about 700 nucleotides. Corneal al(IX) mRNA also migrates as two bands, a major upper band and a minor band about 700 nucleotides shorter (Fig. 3A). However, the corneal bands migrate faster than the cartilage bands, with the uppermost corneal band comigrating with the lower cartilage band. In contrast, the a2(IX) mRNAs appear to be the same size in cartilage and cornea (Fig. 2). The Smaller Size of acl(IX) mRNA in Cornea Is Not Due to the Use of Alternative Polyadenylylation Signals. The nucleotide sequence of the 3' untranslated portion of the cartilagederived, al(IX)-specific cDNA pYN1738 shows the presence of multiple canonical polyadenylylation signals (AATAAA) (1). These signals are clustered in two areas of the untranslated portion of the mRNA (Fig. 3C). Three signals are located in the 5' half of the untranslated sequence, at nucleotide positions 246, 270, and 440 (counted from the first base of the translation stop codon). A second group is formed by signals at positions 867 and 924, with the signal at 924 utilized by the mRNA copied in pYN1738 (1). Since the two groups of polyadenylylation signals are about 600 nucleotides apart, it seemed reasonable to suspect that they provide the explanation for the two bands of al(IX) mRNA seen on transfer blots of cartilage RNA. This is in fact proven by the blot analysis shown in Fig. 3B. The probe was a restriction fragment spanning the region between an Xba I site in the middle of the 3' untranslated portion (Fig. 3C) and the 3' end of the cDNA pYN1738. This probe does not hybridize to the mRNAs migrating in the lower band of cartilage and corneal al(IX) mRNA. However, the probe hybridizes strongly to the major band of both cartilage and corneal mRNA. This demonstrates that the upper, major band of corneal al(IX) mRNA must utilize polyadenylylation signals that are located downstream (3') of the Xba I site. If' the smaller size of corneal mRNAs as seen on transfer blots is not due to utilization of one of the upstream polyadenylylation signals (at position 246, 270, or 440), corneal and cartilage al(IX) mRNAs must be different at their 5' ends or in some internal region. Corneal and Cartilage al(IX) mRNAs Are Different in Their 5' Regions. Since al(IX) chains participate in the formation

7498

Proc. Natl. Acad. Sci. USA 85

Biochemistry: Svoboda et al.

SC

SC 1

R

60

ATG AAA GGG CCA TGA M K G P

p a

250 AATAAA

870 AATAAA

CAG CTGPvu 11

270

440

AATAAA

AATAAATTCTAGAXbal

925

AATAAA

950

AA-An Pvu

A

(1988)

B

11

C

FIG. 3. Transfer blot analysis of poly(A)+ RNA extracted from 17-day-old embryonic chicken sterna (lanes s) and 6-day-old embryonic chicken cornea (lanes c). (A) The blot was hybridized with the nick-translated insert of pYN1738, a ct)NA that encodes the al(IX) collagen chain in cartilage. The probe hybridizes to two bands of mRNA, about 700 nucleotides apart, in both sternal cartilage and corneal RNA. (B) The blot was hybridized to a nick-translated Xba I-Pvu II restriction fragment from the 3' untranslated portion of the insert of pYN1738 (see C). This fragment hybridizes only to the upper band of sternal cartilage and corneal al(IX) RNA. This suggests that the lower band of hybridization in both tissues is due to utilization of polyadenylylation signals located upstream (5') of the Xba I site. (C) Multiple polyadenylylation signals within the 3' untranslated portion of al(IX) mRNA, as deduced from the sequence of the cDNA pYN1738. The last four amino acid residues of the al(IX) chain (M, methionine; K, lysine; G, glycine; P, proline) are shown below the nucleotide sequence. The 3' untranslated region is 947 nucleotides long, starting at the first base of the translation stop codon. One cluster ofthree polyadenylylation signals is separated from a second cluster of two signals by several hundred nucleotides. Also indicated are an Xba I site between the two clusters of polyadenylylation signals and two Pvu II sites. The most 3' of the Pvu II sites lies in the vector downstream of the 3' end of the insert of pYN1738. For the transfer blot in B a fragment between the Xba I site and this 3' Pvu II site was used as probe.

of triple-helical molecules with two other gene products, a2(IX) and a3(IX), internal differences in the RNAs are unlikely. Therefore, we compared the 5' regions of the corneal and cartilage cl(IX) mRNAs by transfer blot analysis (Fig. 4). In this experiment, corneal and cartilage mRNAs were hybridized to two nick-translated restriction fragments. One of the fragments, covering the triple-helical domains of al(IX), hybridized to both cartilage and corneal mRNA. The same result was seen when different parts of this fragment were used separately as probes (data not shown). The second probe, covering the 5' untranslated region and the aminoterminal half of the amino-terminal globular domain of cartilage al(IX) chains, hybridized to the cartilage transcripts but not the corneal transcripts. This suggests that a major portion of the 5' region of al(IX) mRNA, encoding the non-triple-helical domain NC4 in cartilage, is absent in cornea. Obviously, until a cDNA covering the 5' end of corneal al(IX) mRNA is isolated and sequenced, we do not know the precise difference between cartilage and corneal forms of al(IX). However, since cartilage al(IX) mRNA contains about 150 nucleotides of 5' untranslated sequence and 780 nucleotides that code for the amino-terminal globular domain (5), the difference of about 700 nucleotides in length between cartilage and cornea al(IX) mRNAs suggests that almost the complete coding sequence of the globular domain is absent from the corneal transcript. Type IX Collagen Is D)etected in Corneal Primary Stroma. To determine whether type IX collagen mRNAs are translated into protein in the cornea, we used two monoclonal antibodies against chicken type IX collagen for immunolocalization of the molecule. The results indicate that both epitopes are present in the 5-day-old embryonic primary stroma (Fig. 5) but disappear and are undetectable by 8 days of development. This is in contrast to the mRNA levels, which remain high even at day 8 of development (Fig. 2). Whether the decrease in antibody staining is actually due to disappearance of type IX protein or is due to masking of the epitopes recognized by the monoclonal antibodies is currently unknown. In an independent study, Fitch et al. (27) reported a similar finding and pointed out the interesting

correlation between the disappearance of immunodetectable type IX and the subsequent swelling of the avian corneal matrix followed by migration of periocular mesenchymal S C S C 6I

;I

27s-

4~ rl

18s-E IN 321 pYN1738

4

3

2

1

FIG. 4. Transfer blot analysis of total RNA extracted from 17-day-old embryonic chicken sternal cartilage (lanes s) and 6-dayold embryonic chicken cornea (lanes c). The positions of 27S and 18S rRNA markers are indicated. Below the blots is a diagram showing the three triple-helical (lines) and four non-triple-helical domains (boxes, numbered 1-4) of the cartilage al(IX) collagen chain and the relative positions of two restriction fragments, derived from the cDNAs IN321 and pYN1738, that were used as probes. When a fragment derived from IN321 and encoding halfof the amino-terminal globular domain NC4 (labeled 4 in diagram) was used as probe, only the sternal cartilage RNA was positive. In contrast, when a restriction fragment from pYN1738 and encoding the central portion of the al(IX) chain was used as probe, both sternal cartilage and corneal RNAs were positive. Five micrograms of RNA was loaded in each of the cartilage lanes; 15 pg of RNA was loaded in each of the corneal lanes. The cartilage blots were exposed to x-ray film for 24 hr; the cornea blots were exposed for 7 days.

Biochemistry: Svoboda et al.

Proc. Natl. Acad. Sci. USA 85 (1988)

7499

FIG. 5. Immunofluorescence micrograph of the anterior eye segment of a 5-day-old chicken embryo. The unfixed frozen section was incubated with monoclonal antibody 2C2, directed against type IX collagen. Note the intense reactivity in the primary corneal stroma between the corneal epithelium and lens and in the area delineated by epithelium, lens, and the lip of the optic cup. (Bar = 50 ,um.)

cells into the matrix. In fact, these authors presented as a working hypothesis the idea that type IX collagen in the embryonic primary stroma prevents swelling of the matrix by forming crossbridges between collagen fibrils and possibly other components such as proteoglycans (27). They further suggested (27) that rapid enzymatic cleavage of type IX collagen would eliminate the force opposing hyaluronateinduced osmotic swelling, allowing the stroma to expand. The results presented here suggest that type IX collagen molecules are different in cornea and hyaline cartilage, in that the amino-terminal globular domain typical of cartilage is absent in the eye tissue. Given the strategic location of this domain along collagen fibrils in cartilage (14), the difference suggests that the function of type IX collagen in cornea is different from what it is in the matrix of sternal cartilage in 17-day-old embryos. It is, of course, possible that a short form of al(IX) chains is expressed by chondrocytes at other times during chondrogenesis in the chicken embryo or in restricted locations within 17-day-old sternal cartilage. Since the major, upper band of the corneal al(IX) mRNA comigrates with the lower band of the cartilage form, it would be difficult to detect the presence of small amounts of the short form of the al(IX) mRNA in the cartilage RNA extracts. Cloning of cDNA specific for the 5' region of the short form of the al(IX) mRNA will be necessary before these possibilities can be addressed experimentally. The finding of a tissue, the embryonic chicken cornea, that expresses mostly the short form of al(IX) collagen, as reported here, should make isolation of such a cDNA clone possible. We thank Drs. E. D. Hay and M. K. Gordon for helpful discussions; J. Cotton, M. Dews, and J. Parsons for technical assistance; and E. A. McIsaac for typing the manuscript. This work was supported in part by Research Grants AR36820, AR38960, EY05711, and HD00143 from the National Institutes of Health. 1. Ninomiya, Y. & Olsen, B. R. (1984) Proc. Natl. Acad. Sci. USA 81, 3014-3018. 2. van der Rest, M., Mayne R., Ninomiya, Y., Seidah, N. G., Chretien, M. & Olsen, B. R. (1985) J. Biol. Chem. 260, 220225. 3. Gordon, M. K., Gerecke, D. R. & Olsen, B. R. (1987) Proc.

Natl. Acad. Sci. USA 84, 6040-6044. 4. Dublet, B. & van der Rest, M. (1987) J. Biol. Chem. 262,1772417727. 5. Vasios, G., Nishimura, I., Konomi, H., van der Rest, M., Ninomiya, Y. & Olsen, B. R. (1988) J. Biol. Chem. 263, 23242329. 6. Noro, A., Kimata, K., Oike, Y., Shinomura, T., Maeda, N., Yano, S., Takahashi, N. & Suzuki, S. (1983) J. Biol. Chem. 258, 9323-9331. 7. Vaughan, L., Winterhalter, K. H. & Bruckner, P. (1985) J. Biol. Chem. 260, 4758-4763. 8. Bruckner, P., Vaughan, L. & Winterhalter, K. H. (1985) Proc. Nat!. Acad. Sci. USA 82, 2608-2612. 9. Konomi, H., Seyer, J. M., Ninomiya, Y. & Olsen, B. R. (1986) J. Biol. Chem. 261, 6742-6746. 10. McCormick, D., van der Rest, M., Goodship, J., Lozano, G., Ninomiya, Y. & Olsen, B. R. (1987) Proc. Nat!. Acad. Sci. USA 84, 4044-4048. 11. Irwin, M. H. & Mayne, R. (1986) J. Biol. Chem. 261, 1628116283. 12. Eyre, D. R., Apon, S., Wu, J. J., Ericsson, L. H. & Walsh, K. A. (1987) FEBS Lett. 220, 337-341. 13. van der Rest, M. & Mayne, R. (1988) J. Biol. Chem. 263, 16151618. 14. Vaughan, L., Mendler, M., Huber, S., Bruckner, P., Winterhalter, K., Irwin, M. I. & Mayne, R. (1988) J. Cell Biol. 106, 991-997. 15. Linsenmayer, T. F., Smith, G. N. & Hay, E. D. (1977) Proc. Nat!. Acad. Sci. USA 74, 39-43. 16. Hendrix, M. J. C., Hay, E. D., von der Mark, K. & Linsenmayer, T. F. (1982) Invest. Ophthalmol. Visual Sci. 22, 359375. 17. Hay, E. D. & Revel, J.-P. (1969) Monographs in Developmental Biology (Karger, Basel), Vol. 1, pp. 1-144. 18. Adams, S. L., Sobel, M. E., Howard, B. H., Olden, K., Yamada, K. M., deCrombrugghe, B. & Pastan, I. (1977) Proc. Natl. Acad. Sci. USA 74, 3399-3403. 19. Chirgwin, J. M., Przybyla, A. E., MacDonald, R. J. & Rutter, W. J. (1979) Biochemistry 18, 5294-5299. 20. Svoboda, K. K. & Hay, E. D. (1987) Dev. Biol. 123, 455-469. 21. White, B. A., Lufkin, T., Preston, G. M. & Bancroft, C. (1986) Methods Enzymol. 124, 269-278. 22. Thomas, P. S. (1980) Proc. Nat!. Acad. Sci. USA 77, 52015205.

7500

Biochemistry: Svoboda et al.

23. Ninomiya, Y., van der Rest, M., Mayne, R., Lozano, G. & Olsen, B. R. (1985) Biochemistry 24, 4223-4229. 24. Ninomiya, Y., Showalter, A. M., van derRest, M., Seidah, N. G.,

Chretien, M. & Olsen, B. R. (1984) Biochemistry 23, 617-624.

25. Cleveland, D. W., Lopata, M. A., MacDonald, R. J., Cowan,

Proc. Nati. Acad. Sci. USA 85 (i988) N. J., Rutter, W. J. & Kirschner, M. W. (1980) Cell 20, 95-105. 26. Hayashi, M., Ninomiya, Y., Hayashi, K., Linsenmayer, T. F.,

Olsen, B. R. & Trelstad, R. L. (1988) Development 103, 27-36. 27. Fitch, J. M., Mentzer, A., Mayne, R. & Linsenmayer, T. F. (1988) Dev. Biol. 128, 396-405.