Biochemical and structural studies of tenascin/hexabrachion proteins

Tenascin is a large, disulfide-bonded glycoprotein of the extracellular matrix. The predominant form of tenascin observed by electron microscopy is a six-armed oligomer, termed a hexabrachion. We have determined the molecular mass of the native human hexabrachion to be 1.9 x 10(6) Da by sedimentation equilibrium analysis and by electrophoresis on non-reducing agarose gels. On reducing polyacrylamide gel electrophoresis (SDS-PAGE), human tenascin showed a single prominent band at 320 kDa and minor bands of 220 and 230 kDa. The molecular weight of the native human hexabrachion is thus consistent with a disulfide-bonded hexamer of the 320 kDa subunits. Upon treatment with neuraminidase, the apparent molecular weights of all human and chicken tenascin subunits on reducing SDS-PAGE were decreased by about 10 kDa. Prolonged incubation with alpha-mannosidase, however, caused no apparent change in the apparent molecular weight of tenascin subunits. Sedimentation in a cesium chloride gradient gave a higher buoyant density for human tenascin than for fibronectin, suggesting that it has a higher degree of glycosylation. The far-UV circular dichroism spectrum indicates a predominance of beta-structure and a lack of collagen-like or alpha-helical structure. When human hexabrachions were reduced and acetylated, the resulting fragments were single arms which sedimented at 6 S in glycerol gradients and migrated at 320 kDa on non-reducing gels. Treatment of tenascin with trypsin and alpha-chymotrypsin also produced large fragments which were fractionated by gradient sedimentation and analyzed by non-reducing SDS-PAGE and electron microscopy. We present a structural model for the assembly of the observed fragments into the elaborate native hexabrachion.     

Two contrary functions of tenascin: dissection of the active sites by recombinant tenascin fragments

A structural and functional model of tenascin was elaborated using recombinant parts of three alternatively spliced tenascin variants and anti-tenascin monoclonal antibodies. The fusion proteins were compared with intact tenascin for their functions and by electron microscopy. A strong cell binding site was localized within 104 amino acids. This fragment also contains the epitope of the monoclonal antibody anti-Tn68, which inhibits cell attachment to tenascin and binds near the tips of the six arms of tenascin. In contrast, constructs containing the 13 1/2 EGF-like repeats of tenascin showed an antiadhesive effect. The coexistence of the two contrary signals on the same molecule might be responsible for the versatile features of tenascin.     

Tenascin variants: differential binding to fibronectin and distinct distribution in cell cultures and tissues.

In the chicken, three tenascin variants have been characterized that are generated by alternative splicing of 3 of its 11 fibronectin type III repeats. Using monoclonal antibodies that react with common regions versus extra repeats of tenascin, we could distinguish and separate tenascin variants and investigate their interaction with fibronectin using multiple experimental procedures. Interestingly, in all assays used the smallest tenascin variant bound more strongly to fibronectin than the larger ones. These biochemical data were paralleled by the observation that in chick embryo fibroblast cultures only the smallest form of tenascin could be detected in the fibronectin-rich extracellular matrix network laid down by the cells. Furthermore, each tissue present in adult chicken gizzard contained a distinct set of tenascin variants. Those tissues particularly rich in extracellular matrix, such as the tendon, contained the smallest tenascin only. Intermediate-sized tenascin was present in smooth muscle, whereas the largest form was exclusively detectable underneath the epithelial lining of the villi. Thus it appears that cell type-specific forms of tenascin exist that are appropriate for the functional requirements of the respective extracellular matrices.     

Proteolytic cleavage of pyridoxal kinase into two structural domains

Chymotryptic digestion of sheep brain pyridoxal kinase, a dimer of identical subunits each of 40 kDa, yields 2 fragments of 24 and 16 kDa with concomitant loss of catalytic activity. These fragments were separated by chromatographic techniques and analyzed for interaction with the ATP analogue, trinitrophenyl-ATP, using fluorescence spectroscopy. The absorption and fluorescence properties of trinitrophenyl-ATP bound to the fragment of 24 kDa (emission maximum, 540 nm, emission anisotropy at 460 nm, 0.30, and fluorescence lifetime, gamma = lns) are indistinguishable from those of the ATP analogue bound to the native enzyme. The fragment of 16 kDa does not bind trinitrophenyl-ATP. The results are consistent with the hypothesis that monomeric pyridoxal kinase is folded into 2 domains connected by a single polypeptide chain sensitive to proteolytic cleavage.     

The antigenic structure of the acetylcholine receptor from Torpedo californica

The immunological structure of the acetylcholine receptor (AChR) from the electric organ of Torpedo californica was studied using a large number of monoclonal antibodies which were initially selected for their abilities to bind to intact AChRs. The monoclonal antibodies were tested for their ability to bind to denatured AChR subunits labeled with 125I. Antibodies derived from rats immunized with individual denatured subunits or a mixture of subunits of Torpedo AChR reacted well in the assay. A much smaller proportion of antibodies derived from rats immunized with native Torpedo AChR or native AChR from Electrophorus electricus electric organ, bovine muscle, or human muscle reacted with denatured subunits of Torpedo AChR. Many monoclonal antibodies reacted with more than one subunit, but they always reacted best with the subunit used for immunization. Those monoclonal antibodies that bound to intact subunits were mapped more precisely by their ability to bind characteristic fragments of each subunit generated by proteolysis with Staphylococcal V8 protease. These fragments were analyzed by SDS polyacrylamide gel electrophoresis, and monoclonal antibodies that precipitated the same fragment pattern were placed in groups. By this method, we define a minimum of 28 determinants on Torpedo AChR.     

Identification of a cytoplasmic region of the Torpedo nicotinic acetylcholine receptor alpha-subunit by epitope mapping.

Analysis of the binding of monoclonal antibodies (mAbs) by Torpedo nicotinic acetylcholine receptor (AChR) has demonstrated that a region of the alpha-subunit between alpha-156 and alpha-179 is exposed on the cytoplasmic surface of the nicotinic post-synaptic membrane. A panel of mAbs was produced that recognized sodium dodecyl sulfate-denatured subunits of the Torpedo AChR. Antibodies recognizing alpha-subunit were distinguished in terms of their ability to bind alpha-subunit fragments generated by Staphylococcus aureus V8 protease: an 18-kDa fragment beginning at Val-46, a 20-kDa fragment beginning at Ser-173/Ser-162, and a 10 kDa fragment beginning at Asn-339. Three mAbs, selected for binding to each of the V8-protease alpha-subunit fragments, respectively, were characterized in detail. The location of epitopes recognized by both anti-V8-18 and anti-V8-20 mAbs was determined to be within alpha-156 to alpha-179 by isolation of small immunoreactive peptides from proteolytic digests of the alpha-subunit, while the mAb reactive to V8-10 was bound to an epitope within alpha-339 to alpha-386. Quantitative evaluation of binding of the anti-V8-18 and anti-V8-20 mAbs to overlapping synthetic peptides corresponding to alpha-147 to alpha-179 localized the epitopes to distinct portions of this region. Further screening of the panel of mAbs using these synthetic peptides revealed three additional mAbs that bind in this region. The mAbs that bound the three distinct V8-protease alpha-subunit fragments were shown to bind to native AChR by indirect immunofluorescence on frozen sections of Torpedo electric organ. Binding to the native AChR was to the cytoplasmic surface of the AChR since the mAbs could bind to AChR in native vesicles, in which the AChR is oriented right-side-out, only after permeabilization of the vesicles by alkaline treatment or after scrambling of the orientation of the AChR by solubilization and reconstitution into liposomes. The location of the mAb-binding sites at the cytoplasmic surface of the AChR was visualized directly by freeze-etch immunoelectron microscopy. The identification of alpha-156 and alpha-179 as containing a cytoplasmic exposed sequence implies the existence of two non-hydrophobic transmembrane sequences between the site of N-glycosylation (Asn-141) and Cys-192, a site alkylated by the cholinergic affinity labels.     

The nuclear matrix phosphoprotein p255 associates with splicing complexes as part of the [U4/U6.U5] tri-snRNP particle.

The monoclonal antibody CC3 recognizes a phosphorylated epitope present on an interphase protein of 255 kDa. Previous work has shown that p255 is localized mainly to nuclear speckles and remains associated with the nuclear matrix scaffold following extraction with non-ionic detergents, nucleases and high salt. The association of p255 with splicing complexes is suggested by the finding that mAb CC3 can inhibit in vitro splicing and immunoprecipitate pre-messenger RNA and splicing products. Small nuclear RNA immunoprecipitation assays show that p255 is a component of the U5 small nuclear ribonucleoprotein (snRNP) and the [U4/U6.U5] tri-snRNP complex. In RNase protection assays, mAb CC3 immunoprecipitates fragments containing branch site and 3' splice site sequences. As predicted for a [U4/U6.U5]-associated component, the recovery of the branch site-protected fragment requires binding of U2 snRNP and is inhibited by EDTA. p255 may correspond to the previously identified p220 protein, the mammalian analogue of the yeast PRP8 protein. Our results suggest that changes in the phosphorylation of p255 may be part of control mechanisms that interface splicing activity with nuclear organization.     

Tenascin interferes with fibronectin action

Primary chick embryo fibroblasts attach to a tenascin substrate, but remain rounded and do not spread out. The proportion between tenascin and fibronectin in mixtures used to coat the substrate determines the shape of the cells. Tenascin inhibits integrin-mediated chick fibroblast attachment to fibronectin, laminin, and the GRGDS peptide. Rat fibroblast attachment to fibronectin, but not to laminin, is inhibited by tenascin. A monoclonal antibody against tenascin, as well as its Fab fragments, is able to neutralize the inhibitory activity on cell attachment and is therefore assumed to mask the cell-binding site of tenascin. On electron micrographs showing this monoclonal antibody bound to tenascin, its epitope can be localized to the terminal knob at the distal ends of the tenascin arms.     

The structure of a juvenile hormone-binding lipophorin from the hemolymph of Diploptera punctata

The high molecular weight, high affinity juvenile hormone binding protein from the hemolymph of Diploptera punctata was identified as a lipophorin by gradient KBr ultracentrifugation and SDS gradient PAGE. This juvenile hormone binding lipophorin (JHBL) was composed of two subunits, apolipoprotein I (230 kDa mol. wt) and apolipoprotein II (80 kDa mol. wt). The density of the native protein was 1.15 g/ml. Photoaffinity labeling using the JH analog [³H]EFDA demonstrated that the JH binding site resides on apolipoprotein I. The amino acid composition of both native lipophorin and its two subunits was determined and the N-terminal sequence of the 80 kDa apolipoprotein described for 19 of the first 21 amino acids. This sequence did not have similarity to any known protein. The N-terminus of the 230 kDa apolipoprotein was blocked. The specificity of a monoclonal antibody to purified native JHBL was also demonstrated. We show that the monoclonal antibody was specific to the 230 kDa subunit and did not recognize the 80 kDa apolipoprotein.     

Purification and partial characterization of Xenopus laevis tenascin from the XTC cell line.

We report here the purification of tenascin, an extracellular matrix molecule involved in the control of morphogenesis, from the conditioned medium of the Xenopus XTC cell line. Tenascin was purified by affinity chromatography on a column of the monoclonal antibody mAb TnM1; the molecule eluted from this column has a relative molecular mass of 210 kDa after reduction. Electrophoretic analysis under non-reducing conditions shows that the purified components are oligomeric disulfide-linked complexes which barely enter a 4% polyacrylamide gel. Upon rotary shadowing these molecules appear to possess a central globular domain to which pairs or triplets of arms are attached. Polyclonal antibodies have been raised against purified Xenopus tenascin. They recognise specifically the antigen on Western blots of XTC conditioned medium and adult brain, by immunofluorescence, these antibodies reveal large amounts of tenascin in the secretory vesicles as well as in the extracellular matrix of XTC cells. In the Xenopus tadpole, they stain the developing cartilage, the basal lamina of skin epidermis, myotendinous ligaments and restricted regions of the central nervous system.     

Structural analysis of rod GTP-binding protein, Gt. Limited proteolytic digestion pattern of Gt with four proteases defines monoclonal antibody epitope.

The epitope of monoclonal antibody (mAb 4A), which recognizes the alpha subunit of the rod G protein, Gt, has been suggested to be both at the carboxyl terminus (Deretic, D., and Hamm, H.E. (1987) J. Biol. Chem. 262, 10839-10847) and the amino terminus (Navon, S.E., and Fung, B.K.-K. (1988) J. Biol. Chem. 263, 489-496) of the molecule. To characterize further the mAb 4A binding site on alpha t and to resolve the discrepancy between these results limited proteolytic digestion of Gt or alpha t using four proteases with different substrate specificities has been performed. Endoproteinase Arg-C, which cleaves the peptide bond at the carboxylic side of arginine residues, cleaved the majority of alpha t into two fragments of 34 and 5 kDa. The alpha t 34-kDa fragment in the holoprotein, but not alpha t-guanosine 5'-O-(3-thiotriphosphate), was converted further to a 23-kDa fragment. A small fraction of alpha t-GDP was cleaved into 23- and 15-kDa fragments. Endoproteinase Lys-C, which selectively cleaves at lysine residues, progressively removed 17 and then 8 residues from the amino terminus, forming 38- and 36-kDa fragments. Staphylococcus aureus V8 protease is known to remove 21 amino acid residues from the amino-terminal region of alpha t, with the formation of a 38-kDa fragment. L-1-Tosylamido-2-phenylethyl chloromethyl ketone-treated trypsin cleaved alpha t progressively into fragments of known amino acid sequences (38, then 32 and 5, then 21 and 12 kDa) and a transient 34 kDa fragment. The binding of mAb 4A to proteolytic fragments was analyzed by Western blot and immunoprecipitation. The major fragments recognized by mAb 4A on Western blots were the 34- and 23-kDa fragments obtained by endoproteinase Arg-C and tryptic digestion. Under conditions that allowed sequencing of the 15- and 5-kDa fragments neither the 34- nor the 23-kDa fragments could be sequenced by Edman degradation, indicating that they contained a blocked amino terminus. The smallest fragment that retained mAb 4A binding was the 23-kDa fragment containing Met1 to Arg204. Thus the main portion of the mAb 4A antigenic site was located within this fragment, indicating that the carboxyl-terminal residues from Lys205 to Phe350 were not required for recognition by the antibody. Additionally, the antibody did not bind the 38- and 36-kDa or other fragments containing the carboxyl terminus, showing that the amino-terminal residues from Met1 to Lys17 were essential for antibody binding to alpha t.     

Ail protein binds ninth type III fibronectin repeat (9FNIII) within central 120-kDa region of fibronectin to facilitate cell binding by Yersinia pestis

The Yersinia pestis adhesin molecule Ail interacts with the extracellular matrix protein fibronectin (Fn) on host cells to facilitate efficient delivery of cytotoxic Yop proteins, a process essential for plague virulence. A number of bacterial pathogens are known to bind to the N-terminal region of Fn, comprising type I Fn (FNI) repeats. Using proteolytically generated Fn fragments and purified recombinant Fn fragments, we demonstrated that Ail binds the centrally located 120-kDa fragment containing type III Fn (FNIII) repeats. A panel of monoclonal antibodies (mAbs) that recognize specific epitopes within the 120-kDa fragment demonstrated that mAb binding to (9)FNIII blocks Ail-mediated bacterial binding to Fn. Epitopes of three mAbs that blocked Ail binding to Fn were mapped to a similar face of (9)FNIII. Antibodies directed against (9)FNIII also inhibited Ail-dependent cell binding activity, thus demonstrating the biological relevance of this Ail binding region on Fn. Bacteria expressing Ail on their surface could also bind a minimal fragment of Fn containing repeats (9-10)FNIII, and this binding was blocked by a mAb specific for (9)FNIII. These data demonstrate that Ail binds to (9)FNIII of Fn and presents Fn to host cells to facilitate cell binding and delivery of Yops (cytotoxins of Y. pestis), a novel interaction, distinct from other bacterial Fn-binding proteins.     

Molecular characterization of the minimal protease resistant tau unit of the Alzheimer's disease paired helical filament.

The Alzheimer's disease paired helical filament (PHF), after digestion with Pronase, retains its characteristic morphological features. We term this the protease resistant core PHF. A 12 kDa tau fragment can be released from the core as an essentially pure preparation. Sequence analysis of this fragment revealed six distinct N-termini beginning in the repeat region of tau. The precise C-terminus is unknown, but the fragment is approximately 100 residues long. A monoclonal antibody, mAb 423, which recognizes the core PHF and the 12 kDa tau fragment, does not recognize normal full-length tau. We describe cDNA synthesis and expression of candidate 12 kDa tau analogues which permit the mapping of the mAb 423 epitope. mAb 423 recognizes all and only those analogues which terminate at Glu391, which lies beyond the homology repeat region. Addition or removal of a single residue at the C-terminus abolishes immunoreactivity. Therefore, mAb 423, together with knowledge of the N-terminus, can be used to measure the precise extent of 12 kDa PHF core tau fragment which we term the minimal protease resistant tau unit of the core PHF. This unit is 93-95 residues long, which is equivalent to three repeats, but is 14-16 residues out of phase with respect to the maximum homology organization of the repeat region. mAb 423 labels isolated PHFs prior to Pronase digestion and intracellular granular and neurofibrillary degeneration in Alzheimer's disease tissues. The constraints which determine endogenous truncation at Glu391 appear to be characteristic of an assembled configuration of tau, either within the PHF or its precursor.     

Isolation and characterization of human fibroblast tenascin. An extracellular matrix glycoprotein of interest for developmental studies.

We have developed a biochemical method for purifying human tenascin from cultured fibroblasts or the culture medium. The method is a series of biochemical procedures including gel filtration, gelatin gel affinity chromatography and ion-exchange high performance liquid chromatography. The final preparation was identified as tenascin from its immunological cross-reactivity to antibody against chicken tenascin, strong hemagglutination activity which has been reported to be one of the biological functions of chicken tenascin, and from the electron microscopic study demonstrating a six-armed structure. Gel chromatography showed that intact human tenascin has an apparent molecular weight of over one million. Analysis of the purified tenascin with SDS-PAGE under reducing conditions demonstrated that tenascin consists of two kinds of subunits (250K and 190K). We established rat x mouse heterohybridoma cell lines which produce tenascin-specific antibodies. One monoclonal antibody (RCB1) was selected for immunohistochemical study and partially characterized. RCB1 bound native tenascin but not reduced and alkylated tenascin. Immunohistochemistry of normal and neoplastic tissues demonstrated that RCB1 bound the connective tissues surrounding the cancer nests and various normal tissues including interstitium of renal distal tubule, periosteum, endosteum, smooth muscles of digestive tract and media of arteries and arterioles.     

Alternative model for the internal structure of laminin

A monoclonal antibody to laminin, LMN-1, was generated by immunizing rats with laminin from the EHS tumor and fusing the rat spleen cells with mouse NS-1 myeloma cells. Laminin fragments were generated by proteolytic digestion with thrombin, thermolysin, and chymotrypsin. Monoclonal antibody binding fragments were identified by immunoblotting. Fragments which bound monoclonal antibody LMN-1 included a 440-kilodalton (kDa) chymotrypsin fragment and thermolysin fragments of 440 and 110 kDa. These fragments could also be generated from within a 600-kDa thrombin fragment. Digestion of the 440-kDa chymotrypsin fragment with thermolysin generated the 110-kDa antibody binding fragment and a 330-kDa nonbinding fragment. Immunoblotting was performed on extracts of PYS-2 cells and EHS cells using polyclonal and monoclonal antibodies to laminin. Polyclonal antibodies stained the intact 850-kDa complex and the 200- and 400-kDa subunits, while monoclonal LMN-1 stained only the 400-kDa subunit and the complete molecule. Rotary shadowing of monoclonal LMN-1 bound to laminin molecules indicated that the binding site was within the long arm of laminin. Changes in the model of the internal organization of the laminin molecule are proposed, based on the binding of LMN-1 to the 400-kDa subunit and specific proteolytic fragments. The locations of the major thrombin and chymotrypsin fragments in the model are rotated 180 degrees relative to the previously described model [Ott, U., Odermatt, E., Engel, J., Furthmayr, H., & Timpl, R. (1982) Eur. J. Biochem. 123, 63-72] to include part of the 400-kDa subunit of laminin.     

The high-affinity binding of laminin to cells. Assignation of a major cell-binding site to the long arm of laminin and of a latent cell-binding site to its short arms

The laminin proteolytic fragments 1 (derived from the intersection of the short arms of the cruciform laminin molecule) and 8 (derived from the laminin long arm) bind to distinct receptors on HT-1080 human fibrosarcoma cells; both fragments are shown here to inhibit the high-affinity binding of laminin to these cells. Inhibition of binding between fragment 8 and laminin was competitive, whereas that between fragment 1 and laminin was noncompetitive. This indicates that laminin and fragment 8 most probably share the same cellular receptors, whereas laminin and fragment 1 bind to distinct receptors, inhibition being due to steric hindrance. Surprisingly, fragment 1-4 (corresponding to the complete short arms of laminin) neither bound to HT-1080 cells nor inhibited the binding of laminin or fragment 1. After treatment of fragment 1-4 with pepsin, however, the smaller subfragment 1 was liberated, which could then bind to the cells, and so was shown to block the binding of laminin and fragment 1. We conclude that native laminin bound to HT-1080 cells via the fragment-8-binding site near the end of its long arm. Although these cells also have distinct receptors for the short arm fragment 1, this receptor-binding site was not used as it appeared to be latent within the native laminin molecule.     

The identification of epitopic sites in human alpha 1-proteinase inhibitor.

Human alpha 1-proteinase inhibitor (alpha 1-PI) yielded nine fragments on cleavage with CNBr. The amino acid sequences of these fragments were determined. Three of these CNBr-cleavage fragments, namely fragment I (residues 64-220), fragment II (residues 243-351) and fragment III (residues 1-63), were found to bind rabbit polyclonal antibodies against chemically oxidized alpha 1-PI and mouse polyclonal antibodies against native alpha 1-PI by the Bio-Dot method (enzyme-linked immunosorbent assay on nitrocellulose). These fragments, I, II and III, inhibited by 60%, 25% and 5% respectively the binding between alpha 1-PI and the rabbit antibodies. Fragments I, II and III were subjected to proteolytic digestion, and 15, ten and five peptides were obtained from these fragments respectively. Only four of these peptides showed binding to the mouse antibodies against native alpha 1-PI. These were residues 40-63, 79-86, 176-206 and 299-323. A panel of monoclonal antibodies was prepared by conventional hybridoma technology, with chemically oxidized alpha 1-PI as the antigen. The ability of the monoclonal antibodies to bind native alpha 1-PI and CNBr-cleavage fragments I-III was determined. The monoclonal antibodies fell into three categories. Most (over 90%) belonged to group I, which was capable of binding alpha 1-PI and only fragment I. Antibodies in groups II and III bound alpha 1-PI and either fragment II or fragment III respectively. The ability of the peptides derived from proteolytic digestion of fragments I, II and III to bind three monoclonal antibodies representing each of the three groups was determined. Among all the peptides tested, only one (residues 176-206) derived from fragment I showed binding to the antibodies from group I, one (residues 299-323) derived from fragment II showed binding to the antibodies from group II, and one (residues 40-63) from fragment III showed binding to the antibodies from group III. Each of these three peptides also inhibited the binding between alpha 1-PI and the corresponding monoclonal antibodies. From these data we concluded that at least four epitopic regions (residues 40-63, 79-86, 176-206 and 299-323) were present in alpha 1-PI. Specific monoclonal antibodies to three of these sites were obtained.     

Oligomerization and ligand-binding properties of the ectodomain of the alpha-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid receptor subunit GluRD.

The extracellular part of ionotropic glutamate receptor (iGluR) subunits can be divided into a conserved two-lobed ligand-binding domain ("S1S2") and an N-terminal approximately 400-residue segment of unknown function ("X domain") which shows high sequence variation among subunits. To investigate the structure and properties of the N-terminal domain, we have now produced affinity-tagged recombinant fragments which represent the X domain of the GluRD subunit of alpha-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid (AMPA)-selective glutamate receptors either alone or covalently linked to the ligand-binding domain ("XS1S2"). These fragments were expressed in insect cells as secreted soluble proteins and were recognized by a conformation-specific anti-GluRD monoclonal antibody. A hydrodynamic analysis of the purified fragments revealed them to be dimers, in contrast to the S1S2 ligand-binding domain which is monomeric. The X domain did not bind radiolabeled AMPA or glutamate nor did its presence affect the ligand binding properties of the S1S2 domain. Our findings demonstrate that the N-terminal domain of AMPA receptor can be expressed as a soluble polypeptide and suggest that subunit interactions in iGluR may involve the extracellular domains.     

A major oligomeric fibroblast proteoglycan identified as a novel large form of type-XII collagen.

Cultured chick embryo skin fibroblasts release a major component with a native molecular mass of about 1 MDa, which resolves into three polypeptide bands of about 300, 350 and 600 kDa upon reduction. We report here the purification of this oligomeric protein and show, by means of polyclonal and monoclonal antibodies, that its three polypeptide constituents are closely related. The 600-kDa polypeptide is likely to be a dimer of two smaller subunits which are cross-linked by non-reducible bonds. By electron microscopy, isolated oligomeric molecules exhibit a novel cruciform structure with a large central globular domain. One arm has the shape of a thin rod about 70 nm in length. The three other arms are thicker, longer (90 nm) and flexible, and carry a prominent double globule at their distal ends. Collagenase treatment of the oligomeric fibroblast protein yields two resistant fragments of about 270 kDa and 320 kDa. The intact 350-kDa and 600-kDa (but not the 300-kDa) polypeptides are chondroitinase sensitive and labeled by metabolic incorporation of [35S]sulfate; collagenase treatment does not remove any [35S] sulfate. Hence, the intact fibroblast protein has glycosaminoglycan chains attached to its non-collagenous domain. Three amino acid sequences obtained from chymotryptic fragments of the fibroblast protein correspond to sequences predicted for chick type-XII collagen from its full-length cDNA [Yamagata, M., Yamada, K. M., Yamada, S. S., Shinomura, T., Tanaka, H., Nishida, Y., Obara, M. & Kimata, K. (1991) J. Cell Biol. 115, 209-221]. However, the novel fibroblast protein described here differs significantly from previously isolated forms of type-XII collagen: its subunits are larger by one third, and it is a proteoglycan.