Extracellular Matrix-Associated Protein Sc1 Is Not Essential for Mouse ...

12 downloads 0 Views 821KB Size Report
PETER J. MCKINNON,1* SUSAN K. MCLAUGHLIN,2 MANUELA KAPSETAKI,1 ..... Gilmour, D. T., G. J. Lyon, M. B. Carlton, J. R. Sanes, J. M. Cunningham,.
MOLECULAR AND CELLULAR BIOLOGY, Jan. 2000, p. 656–660 0270-7306/00/$04.00⫹0 Copyright © 2000, American Society for Microbiology. All Rights Reserved.

Vol. 20, No. 2

Extracellular Matrix-Associated Protein Sc1 Is Not Essential for Mouse Development PETER J. MCKINNON,1* SUSAN K. MCLAUGHLIN,2 MANUELA KAPSETAKI,1 3 AND ROBERT F. MARGOLSKEE Department of Genetics, St. Jude Children’s Research Hospital, Memphis, Tennessee 381051; Department of Neurobiology and Behavior, SUNY at Stony Brook, Stony Brook, New York2; and Howard Hughes Medical Institute and Department of Physiology and Biophysics, Mount Sinai Medical Center, New York, New York3 Received 23 July 1999/Returned for modification 13 September 1999/Accepted 11 October 1999

Sc1 is an extracellular matrix-associated protein whose function is unknown. During early embryonic development, Sc1 is widely expressed, and from embryonic day 12 (E12), Sc1 is expressed primarily in the developing nervous system. This switch in Sc1 expression at E12 suggests an importance for nervous-system development. To gain insight into Sc1 function, we used gene targeting to inactivate mouse Sc1. The Sc1-null mice showed no obvious deficits in any organs. These mice were born at the expected ratios, were fertile, and had no obvious histological abnormalities, and their long-term survival did not differ from littermate controls. Therefore, the function of Sc1 during development is not critical or, in its absence, is subserved by another protein. certain endothelial cell lines (2, 3); the follistatin-like module present in Sparc (and Sc1) may be responsible for this cytokine regulation (10, 25). The direct modulatory effect of Sparc on other growth factors is unknown but may be important, since Sparc can also influence basic fibroblast growth factor activity towards endothelial cells (31). Both Sc1 and Sparc possess a strong antiadhesive activity toward attachment of endothelial cells to different substrates in vitro (7, 13). A requirement of SPARC for metastasis and tumorigenicity of human melanoma cells has also been reported (14, 22, 29). This is due to its ability to augment adhesive and invasive capacities and implies that it has a function as a tumor suppressor. While the significance of the above to Sc1 function is unclear, matrix remodeling and selective modulation of growth factors are clearly features of development. To delineate the biological role of Sc1, we used homologous recombination to inactivate mouse Sc1. Although Sc1 is abundantly expressed during development and in the adult mouse, inactivation of this gene had no obvious consequences for development or function of the mouse.

Molecules of the extracellular matrix (ECM) show definite patterns of spatial and temporal regulation and are critical for a wide range of events during development. For example, development of the nervous system requires ECM components for axonal target finding, neural crest guidance, and normal tissue morphogenesis (33, 36, 37). While the ECM is a critical modulator of spatial and temporal information during development, there are many matrix components whose function are unknown. One such protein is Sc1. Sc1 is abundantly expressed in the adult nervous system and heart (11, 17, 21). Sc1 is also found in the high endothelial venules of the immune system, where it has been suggested to function in lymphocyte extravasation (6). Sc1 is a member of a gene family that includes Sparc, testican, Qr1, agrin, and follistatin (1, 8, 25, 35). The similarity of Sc1 to members of this family, which varies from 30 to 65% amino acid identity, is by virtue of a 230-amino-acid stretch at the carboxyl terminus of the molecule, suggesting functional conservation of this region. However, Sc1 also contains a unique amino terminus (⬃400 amino acids), which shows no similarity to any protein sequence in the current gene databases. The region of Sc1 encoding the carboxy terminus contains a similar exon structure to that of Sparc, suggesting that these two genes evolved from a common ancestral gene and may have related functions (16). Sparc is highly expressed in bone (hence its alternative name, osteonectin), as well as a number of other tissues (12, 20, 23, 32). Sparc has been implicated in angiogenesis (27, 30), and both Sc1 and Sparc are associated with astrocytes in the adult rodent brain (17, 19). Because astrocytes are required for maintenance of the blood-brain barrier and a nutrient supply interface between capillary blood supply and neurons (34), the presence of Sc1 and Sparc in astrocytes supports a role in angiogenesis. The upregulation of Sc1 and Sparc after neural injury (4, 17) may also suggest a role in tissue remodeling or repair. Sparc is able to modulate platelet-derived growth factor binding to its receptor, and this affects cell cycle progression in

MATERIALS AND METHODS Northern blotting and RPA. RNase protection analysis (RPA) was performed as previously described (18). RPA of Sc1 during development was done with pSC1-423 (17). For analysis of Sc1 expression in the Sc1-null animals, PCR was used to obtain Sc1 cDNA spanning exons 2 to 4 that was cloned into the BamHI site of pBSII (Stratagene). The primers used were CGGGATCCAGCCACCT CTCCGCACA and CGGGATCCACATAGGAAGTGGACAC (BamHI sites are underlined). The antisense 32P-labelled Sc1 probe was obtained with T7 RNA polymerase by using SalI-linearized plasmid DNA. The actin probe was as described previously (18). Sc1 and Actin RPAs were done simultaneously with an RPAII kit (Ambion) to standardize input RNA levels. Northern blots were performed on tissue RNA [2 ␮g of poly(A) RNA] that had been separated on formaldehyde-containing 1% agarose gels and transferred to nylon membranes (Clontech). The probes used for Northern analysis were the full-length mouse Sc1 gene (GenBank accession no. U64827) originally isolated from a mouse brain cDNA library (16). The Sc1 probe was generated by using [32P]dCTP and hybridized to filters overnight at 50 in 50% formamide–7% sodium dodecyl sulfate (SDS)–2⫻ SSPE (1⫻ SSPE is 0.8 M NaCl, 10 mM NaH2PO4, and 1 mM EDTA [pH 7.7])–0.5% Blotto–200 ␮g of salmon sperm DNA per ml. The blots were washed twice in 0.2⫻ SSC (1⫻ SDS is 0.15 M NaCl plus 0.015 M sodium citrate)–0.1% SDS at 65°C for 20 min and then exposed to Kodak Xomat-AR film at ⫺80°C. The membranes were then stripped in 0.5% SDS at 95°C for 5 min and probed with a human actin fragment (Clontech). The actin control was labeled, hybridized, and washed under the conditions described above.

* Corresponding author. Mailing address: Department of Genetics, St. Jude Children’s Research Hospital, 332 N. Lauderdale, Memphis, TN 38105. Phone: (901) 495 2700. Fax: (901) 526 2907. E-mail: peter [email protected]. 656

VOL. 20, 2000

Sc1 KNOCKOUT MICE

657

FIG. 1. Analysis of Sc1 experssion. (A) RPA was used to assess murine Sc1 expression from E8 to E14. Actin was used as an internal standard for mRNA quantitation. The sizes of the protected products are 427 bp for Sc1 and 120 bp for Actin. (B) Northern blot analysis of poly(A)⫹ RNA from a number of mouse tissues shows a single 2.7-kb band for Sc1. The lower panel results from reprobing the blot with an Actin probe to normalize the RNA quantity. (C) While Sc1 is widely expressed at E11, it is absent from the developing CNS (asterisk). However, by E12, Sc1 is expressed in the developing CNS (iii). Panels i and iii were hybridized with antisense Sc1 probe and panel ii was hybridized with sense Sc1 probe to demonstrate the specificity of the antisense signal.

In situ hybridization. Paraffin-embedded tissue sections were obtained from Novagen (Madison, Wis.) and prepared for in situ analysis as specified by the manufacturer. 33P-antisense transcript was generated by SP6 polymerase (Promega) with SalI-linearized pSC1-423 template (17), and the sense transcript was generated with T7 polymerase (Promega) with NotI-linearized cDNA. In situ hybridization was performed as described in reference 15, with the following modifications. Cryosections were digested with 0.001% proteinase K in 0.1 M Tris-HCl–50 mM EDTA (pH 8.0) for 10 min. The sections were washed briefly in diethylpyrocarbonate-treated water and then incubated in 0.1 M triethanolamine (pH 8.0) for 2 min. They were acetylated with 0.25% acetic anhydride–0.1 M triethanolamine for 10 min, rinsed in 2⫻ SSC, and dehydrated through a stepped ethanol series to 100% ethanol. They were then hybridized overnight at 55°C in a humidified environment in 100 ␮l of 0.6 M NaCl–10 mM Tris–0.02% Ficoll–0.02% bovine serum albumin–0.02% polyvinylpyrrolidone–1 mM EDTA–10% dextran sulfate (Pharmacia)–0.05% yeast RNA–0.05% herring sperm DNA–0.005% yeast tRNA–0.1% SDS–50% formamide, containing 106 cpm of 33P-labeled riboprobe. The sections were rinsed in 4⫻ SSC–50% formamide followed and washed in 2⫻ SSC, and then nonspecifically bound probe was removed with RNase A (100 ␮g/ml in 0.5 M NaCl–10 mM Tris–1 mM EDTA [pH 7.4]) for 30 min at 37°C. Finally, the sections were washed in 2⫻ SSC at 60°C for 1 h and then in 0.2⫻ SSC at 65°C for 2 h and dehydrated in ethanol. They were dipped in NTB2 emulsion (Kodak), exposed at 4°C for 2 weeks, and developed as recommended by Kodak. They were counterstained with hematoxylin and eosin. Gene targeting. Mouse Sc1 cDNA was used as a probe to obtain a 14-kb Sc1 genomic fragment from a mouse 129Svj genomic library in ␭Fix II (Stratagene). This 14-kb fragment contained exons 2 through 8 and was cloned into pBlue-

script II as a NotI fragment. XbaI digestion was used to remove exon 2 (which contains the initiator methionine and secretion signal sequence) and, the remaining was cloned into pPGKNeo/TK as a 8-kb HindIII-SalI fragment (long homology region). The short homology region was generated from genomic DNA by PCR with primers CATAAGAATGCGGCCGCCACACCAAGTCTG AATGCCTCAA and CATAAGAATGCGGCCGCCCCTAGATAATTTCAC AAGGACTAGC (NotI sites are underlined), and the product was digested with NotI and cloned into the NotI site of PGKNeo/TK to generate PGKNeo/TK-Sc1. This construct was linearized with SalI and electroporated into W9.5 embryonic stem (ES) cells. Targeted ES cells were identified by Southern blot analysis of HindIII-digested ES genomic DNA with an Sc1 NcoI cDNA fragment that spanned exons 10 and 11. HindIII digestion generated a 13-kb fragment, due to an introduced HindIII site in the mutant allele, that was readily distinguishable from the endogenous 20kb Sc1 genomic HindIII fragment. Immunohistochemistry of Sc1-null tissues was performed as previously described (17), and Western blot analysis was done with anti-Sc1 (17) by the method of Herzog et al. (9).

RESULTS AND DISCUSSION To gauge the likely consequences of Sc1 inactivation, Sc1 expression was examined during mouse development. RPA showed high levels of Sc1 expression from embryonic day 8 (E8) (the earliest point examined) through E14 (Fig. 1A). Sc1 was also expressed at high levels in the adult mouse brain,

658

MCKINNON ET AL.

MOL. CELL. BIOL.

FIG. 2. Gene targeting of Sc1 in the mouse. (A) Sc1 was inactivated by replacing exon 2 (which contains the initiator methionine and secretion signal sequence) with a neomycin selection cassette driven by the PGK promoter. X, XbaI sites; Xh, XhoI sites. (B) Homologous recombination introduces an additional HindIII site into the Sc1 locus. This results in a 13-kb fragment derived from the mutant Sc1 allele after HindIII digestion of genomic DNA and Southern blot analysis with an NcoI cDNA fragment that encompasses exons 10 and 11. The HindIII sites listed in panel A are only a guide for Southern analysis, since other HindIII sites exist in the first intron of Sc1. The Southern blot shown is representative of the results obtained with mice derived from mating Sc1 heterozygotes. Lanes: 1 (asterisk), Sc1⫺/⫺; 6, Sc1⫹/⫺; 2 to 5 and 7, wild type. The 6-kb band is a 3⬘ HindIII fragment containing exon 11 that also hybridizes to the probe (see panel A) (C) RPA shows that Sc1 mRNA is present in the wild-type and Sc1 heterozygotes but not in the Sc1⫺/⫺ mice. Sc1 is the upper protected product, and Actin is the lower product and is included as an internal standard. (D) Western blot analysis of proteins obtained from the cerebellum of an 8-week-old mouse shows that Sc1 is present in wild-type mice (⫹/⫹) but is absent from homozygous mutant animals (⫺/⫺). (E) Immunohistochemical localization of Sc1 in the adult cortex shows strong staining for Sc1 in the wild-type animals but an absence of staining in the Sc1⫺/⫺ brain. The arrow indicates similar regions in each panel.

heart, lung, and muscle (Fig. 1B). We used in situ hybridization to determine the spatial distribution of Sc1 during mouse development. Embryo sections probed with 33P-labeled antisense Sc1 at E11 and E12 (Fig. 1C) showed widespread expression. At E11, Sc1 was widely expressed, particularly in the mesenchyme, and was largely absent from the neuroepithelium (Fig. 1C, panel i). However, between E11 and E12, a change in the spatial distribution of Sc1 occurred. At E12, Sc1 expression was apparent in the nervous-system structures, particularly the differentiating fields of the spinal cord (Fig. 1C, panel iii), and by E15, Sc1 transcripts were absent from the mesenchyme and found exclusively throughout the nervous system (results not shown). The striking specification of expression that occurs from E12 and the high levels in adult brains suggested that Sc1

was important in the developing central nervous system (CNS). Therefore, we proceeded to inactivate Sc1. To inactivate Sc1, we replaced exon 2 of the murine Sc1 genomic locus with a Neor selection cassette to create an outof-frame disruption (Fig. 2A). This replacement also removed the secretion signal sequence, abolishing any possibility of generating a truncated version of Sc1 that could be secreted. Targeting of ES cells occurred at a frequency of approximately 1/20, and two of these targeted lines were used for the creation of chimeras to generate Sc1 heterozygous mice. Sc1 heterozygotes were used to generate Sc1-null mice at the expected frequency of 1/4 (Fig. 2B). We confirmed that Sc1 expression was disrupted by using a RPA probe that encompassed Sc1 exons 2 to 4. Sc1 protection products were identified in RNA

Sc1 KNOCKOUT MICE

VOL. 20, 2000

obtained from wild-type and heterozygous, but not homozygous, Sc1⫺/⫺ adult cerebellum (Fig. 2C). We also confirmed that Sc1 protein was absent in the Sc1⫺/⫺ mice by using Western blot analysis (Fig. 2D) and immunohistochemistry (Fig. 2E) with Sc1 antisera (17). Both inbred (129svj) and outbred (crossed to C57BL/6) Sc1⫺/⫺ lines were fertile and had no obvious defects, and their long-term survival was indistinguishable from that of their wild-type littermates. Survival was monitored for up to 1 year for the 129svj line and longer than 18 months for the outbred Sc1⫺/⫺ line. Histological analysis of the mice showed no gross anatomical defects in any organs (data not shown). Additionally, histological analysis of the nervous system at various ages up to 6 months showed no differences compared to controls. Immunohistochemical studies with a variety of markers including glial fibrillary acidic protein, neurofilament, and calbindin also failed to reveal any differences between Sc1-null and controls at 2 or 8 months of age. Since Sc1 is upregulated following neural injury (17, 19), we examined reactive astrocytosis in the Sc1-null mice following focal mechanical trauma. While pronounced astrocytosis following the injury was observed at 24 and 48 h after trauma, as shown by glial fibrillary acidic protein immunohistochemical detection, no differences were seen between Sc1-null mice and littermate controls (data not shown). Inactivation of mouse Sparc leads to cataract formation (5, 24). However, no cataracts were found in the Sc1null mice. We also found no differences in the levels of Sparc expression in a variety of Sc1⫺/⫺ tissues by RPA (data not shown). It was surprising that inactivation of a gene so abundantly expressed throughout development had no apparent consequence for the development of the mouse. While other Sc1related genes may be able to substitute for Sc1 function, a large portion of this protein is unique and has no obvious similarity to presently known genes. The amino-terminal region of Sc1 has been conserved during evolution and therefore is likely to be important for its function. Of course, this region of the protein may have a nonessential function for development. Inactivation of many ECM components produced pronounced developmental defects (38, 40). However, many other ECMrelated genes have been inactivated, with little obvious consequence for the organism (26, 28, 41), while others have very subtle defects (39). The spectrum of phenotypes from the various knockout mice highlights the functional diversification of ECM components. It is likely that the function of genes such as Sc1 will be revealed by the generation of multigene knockout mice. ACKNOWLEDGMENTS We thank Colin Stewart for providing ES cells, Gwen Wong for expert advice on generation of the knockout mice, Galya Vassileva for help with microinjection, Robert Wurtzberger for synthesis of oligonucleotides and DNA sequencing, and Suzanne Baker for comments on the manuscript. These studies were supported in part by Cancer Center CORE grant NIH P30 CA 21765-19 (P.J.M.) and by the American Lebanese and Syrian Associated Charities (ALSAC) of St. Jude Children’s Research Hospital. R.F.M. is an Associate Investigator of the Howard Hughes Medical Institute. REFERENCES 1. Alliel, P. M., J. P. Perin, P. Jolles, and F. J. Bonnet. 1993. Testican, a multidomain testicular proteoglycan resembling modulators of cell social behaviour. Eur. J. Biochem. 214:347–350. 2. Funk, S. E., and E. H. Sage. 1991. The Ca2⫹-binding glycoprotein SPARC modulates cell cycle progression in bovine aortic endothelial cells. Proc. Natl. Acad. Sci. USA 88:2648–2652. 3. Funk, S. E., and E. H. Sage. 1993. Differential effects of SPARC and cationic

4. 5.

6. 7. 8. 9. 10. 11. 12. 13. 14.

15. 16. 17. 18. 19.

20. 21. 22. 23. 24. 25. 26.

27.

28. 29.

659

SPARC peptides on DNA synthesis by endothelial cells and fibroblasts. J. Cell. Physiol. 154:53–63. Gillen, C., M. Gleichmann, P. Spreyer, and H. W. Muller. 1995. Differentially expressed genes after peripheral nerve injury. J. Neurosci. Res. 42:159– 171. Gilmour, D. T., G. J. Lyon, M. B. Carlton, J. R. Sanes, J. M. Cunningham, J. R. Anderson, B. L. Hogan, M. J. Evans, and W. H. Colledge. 1998. Mice deficient for the secreted glycoprotein SPARC/osteonectin/BM40 develop normally but show severe age-onset cataract formation and disruption of the lens. EMBO J. 17:1860–1870. Girard, J.-P., and T. A. Springer. 1995. Cloning from purified high endothelial venule cells of hevin, a close relative of the antiadhesive extracellular matrix protein SPARC. Immunity 2:113–123. Girard, J. P., and T. A. Springer. 1996. Modulation of endothelial cell adhesion by hevin, an acidic protein associated with high endothelial venules. J. Biol. Chem. 271:4511–4517. Guermah, M., P. Crisanti, D. Laugier, P. Dezelee, L. Bidou, B. Pessac, and G. Calothy. 1991. Transcription of a quail gene expressed in embryonic cells is shut off sharply at hatching. Proc. Natl. Acad. Sci. USA 88:4503–4507. Herzog, K. H., M. J. Chong, M. Kapsetaki, J. I. Morgan, and P. J. McKinnon. 1998. Requirement for Atm in ionizing radiation-induced cell death in the developing central nervous system. Science 280:1089–1091. Hohenester, E., P. Maurer, and R. Timpl. 1997. Crystal structure of a pair of follistatin-like and EF-hand calcium-binding domains in BM-40. EMBO J. 16:3778–86. Johnston, I. G., T. Paladino, J. W. Gurd, and I. R. Brown. 1990. Molecular cloning of SC1: a putative brain extracellular matrix glycoprotein showing partial similarity to osteonectin/BM40/SPARC. Neuron 2:165–176. Lane, T. F., and E. H. Sage. 1994. The biology of SPARC, a protein that modulates cell matrix interactions. FASEB J. 8:163–173. Lane, T. F., and E. H. Sage. 1990. Functional mapping of SPARC: peptides from two distinct Ca⫹(⫹)-binding sites modulate cell shape. J. Cell Biol. 111:3065–3076. Ledda, M. F., S. Adris, A. I. Bravo, C. Kairiyama, L. Bover, Y. Chernajovsky, J. Mordoh, and O. L. Podhajcer. 1997. Suppression of SPARC expression by antisense RNA abrogates the tumorigenicity of human melanoma cells. Nat. Med. 3:171–176. Lugo, D. I., J. L. Roberts, and J. E. Pintar. 1989. Analysis of proopiomelanocortin gene expression during prenatal development of the rat pituitary gland. Mol. Endocrinol. 3:1313–1324. McKinnon, P. J., M. Kapsetaki, and R. F. Margolskee. 1996. The exon structure of the mouse Sc1 gene is very similar to the mouse Sparc gene. Genome Res. 6:1077–1083. McKinnon, P. J., and R. F. Margolskee. 1996. SC1: A marker for astrocytes in the adult rodent brain is upregulated during reactive astrocytosis. Brain Res. 709:27–36. McLaughlin, S. K., P. J. McKinnon, and P. J. Margolskee. 1992. Gustducin is a taste cell specific G protein closely related to the transducins. Nature 357:563–569. Mendis, D. B., G. O. Ivy, and I. R. Brown. 1996. SC1, a brain extracellular matrix glycoprotein related to SPARC and follistatin, is expressed by rat cerebellar astrocytes following injury and during development. Brain Res. 730:95–106. Mendis, D. B., L. Malaval, and I. R. Brown. 1995. SPARC, an extracellular matrix glycoprotein containing the follistatin module, is expressed by astrocytes in synaptic enriched regions of the adult brain. Brain Res. 676:69–79. Mendis, D. B., S. Shahin, J. W. Gurd, and I. R. Brown. 1996. SC1, a SPARC-related glycoprotein, exhibits features of an ECM component in the developing and adult brain. Brain Res. 713:53–63. Mok, S. C., W. Y. Chan, K. K. Wong, M. G. Muto, and R. S. Berkowitz. 1996. SPARC, an extracellular matrix protein with tumor-suppressing activity in human ovarian epithelial cells. Oncogene 12:1895–1901. Nomura, S., A. J. Wills, D. R. Edwards, J. K. Heath, and B. L. Hogan. 1989. Expression of genes for non-collagenous proteins during embryonic bone formation. Connect. Tissue Res. 21:31–35. Norose, K., J. I. Clark, N. A. Syed, A. Basu, E. Heber-Katz, E. H. Sage, and C. C. Howe. 1998. SPARC deficiency leads to early-onset cataractogenesis. Investig. Ophthalmol. Visual Sci. 39:2674–2680. Patthy, L., and K. Nikolics. 1993. Functions of agrin and agrin-related proteins. Trends Neurosci. 16:76–81. Rosati, R., G. S. Horan, G. J. Pinero, S. Garofalo, D. R. Keene, W. A. Horton, E. Vuorio, B. de Crombrugghe, and R. R. Behringer. 1994. Normal long bone growth and development in type X collagen-null mice. Nat. Genet. 8:129– 135. Rosenblatt, S., J. A. Bassuk, C. E. Alpers, E. H. Sage, R. Timpl, and K. T. Preissner. 1997. Differential modulation of cell adhesion by interaction between adhesive and counter-adhesive proteins: characterization of the binding of vitronectin to osteonectin (BM40, SPARC). Biochem. J. 324:311–319. Saga, Y., T. Yagi, Y. Ikawa, T. Sakakura, and S. Aizawa. 1992. Mice develop normally without tenascin. Genes Dev. 6:1821–1831. Sage, E. H. 1997. Terms of attachment: SPARC and tumorigenesis. Nat. Med. 3:144–146.

660

MCKINNON ET AL.

30. Sage, E. H., and R. B. Vernon. 1994. Regulation of angiogenesis by extracellular matrix: the growth and the glue. J. Hypertens. Suppl. 12:S145–S152. 31. Shiba, H., S. Nakamura, M. Shirakawa, K. Nakanishi, H. Okamoto, H. Satakeda, M. Noshiro, K. Kamihagi, M. Katayama, and Y. Kato. 1995. Effects of basic fibroblast growth factor on proliferation, the expression of osteonectin (SPARC) and alkaline phosphatase, and calcification in cultures of human pulp cells. Dev. Biol. 170:457–466. 32. Soderling, J. A., M. J. Reed, A. Corsa, and E. H. Sage. 1997. Cloning and expression of murine SC1, a gene product homologous to SPARC. J. Histochem. Cytochem. 45:823–835. 33. Tessier-Lavigne, M., and C. S. Goodman. 1996. The molecular biology of axon guidance. Science 274:1123–1133. 34. Tsacopoulos, M., and P. J. Magistretti. 1996. Metabolic coupling between glia and neurons. J. Neurosci. 16:877–885. 35. Vannahme, C., S. Schubel, M. Herud, S. Gosling, H. Hulsmann, M. Paulsson, U. Hartmann, and P. Maurer. 1999. Molecular cloning of testican-2: defining a novel calcium-binding proteoglycan family expressed in brain. J. Neurochem. 73:12–20. 36. Varela-Echavarria, A., and S. Guthrie. 1997. Molecules making waves in

MOL. CELL. BIOL. axon guidance. Genes Dev. 11:545–557. 37. Walsh, F. S., and P. Doherty. 1997. Neural cell adhesion molecules of the immunoglobulin superfamily: role in axon growth and guidance. Annu. Rev. Cell. Dev. Biol. 13:425–456. 38. Watanabe, H., and Y. Yamada. 1999. Mice lacking link protein develop dwarfism and craniofacial abnormalities. Nat. Genet. 21:225–229. 39. Weber, P., U. Bartsch, M. N. Rasband, R. Czaniera, Y. Lang, H. Bluethmann, R. U. Margolis, S. R. Levinson, P. Shrager, D. Montag, and M. Schachner. 1999. Mice deficient for tenascin-R display alterations of the extracellular matrix and decreased axonal conduction velocities in the CNS. J. Neurosci. 19:4245–4262. 40. Xu, T., P. Bianco, L. W. Fisher, G. Longenecker, E. Smith, S. Goldstein, J. Bonadio, A. Boskey, A. M. Heegaard, B. Sommer, K. Satomura, P. Dominguez, C. Zhao, A. B. Kulkarni, P. G. Robey, and M. F. Young. 1998. Targeted disruption of the biglycan gene leads to an osteoporosis-like phenotype in mice. Nat. Genet. 20:78–82. 41. Zheng, X., T. L. Saunders, S. A. Camper, L. C. Samuelson, and D. Ginsburg. 1995. Vitronectin is not essential for normal mammalian development and fertility. Proc. Natl. Acad. Sci. USA 92:12426–12430.