The Inclusion of the Type I11 Repeat ED-B in the Fibronectin Molecule Generates Conformational Modifications That Unmask a Cryptic Sequence*

We have previously reported an anti-fibronectin monoclonal antibody (mAb) (BC-1) which reacts with an ED-B-containing@-galactosidase-fibronectin fusion protein but not with an identical 8-galactosidase-fibro- nectin fusion protein in which the ED-B sequence is omitted. In further experiments aimed at localizing more precisely the epitope recognized by this mAb, we demonstrate that 1) the mAb BC-1 is indeed specific for ED-B-containing fibronectin (FN) molecules even though the epitope recognized by this mAb is localized on the type I11 homology repeat 7 (the one which pre- cedes the ED-B sequence) and 2) in fibronectin molecules lacking the ED-B sequence, this epitope is masked. We further demonstrate that, to mask the epitope recognized by the mAb BC-1, the presence of at least half of the FN type 111 homology repeat 9 is necessary. We also report the production of the mAb IST-6 which recognizes only FN molecules in which the ED-B sequence is lacking. These data clearly demonstrate that the presence of the ED-B sequence within FN molecules generates conformational modification in the central part of the molecules that unmasks pre- viously cryptic sequences and masks others.


IST-6 which recognizes only FN molecules in which
the ED-B sequence is lacking. These data clearly demonstrate that the presence of the ED-B sequence within FN molecules generates conformational modification in the central part of the molecules that unmasks previously cryptic sequences and masks others. Fibronectins (FNs)' are high molecular mass adhesive glycoproteins present in the extracellular matrix and in body fluids. These molecules are involved in different biological phenomena such as the establishment and maintenance of normal cell morphology, cell migration, hemostasis and thrombosis, wound healing, and oncogenic transformation (for reviews, see Alitalo and Vaheri (1982), Yamada (1983), Hynes (1985), Ruoslahti (1988), Hynes (1990), Owens et al. (1986)). FN polymorphism is due to alternative splicing patterns in three regions (IIICS, ED-A, and ED-B) of the single FN primary transcript (see Fig. 1) as well as to post-translational modifications. The alternative splicing of the FN pre-mRNA is regulated in a cell-, tissue-, and developmentally specific manner (see above-mentioned reviews and references therein). Furthermore, it has been demonstrated that the splicing pattern of FN pre-mRNA is deregulated in transformed cells and in malignancies (Castellani et al., 1986; * This study was partially supported by Associazione Italiana per la Ricerca sul Cancro (AIRC) funds and Consiglio Nazionale delle Ricerche (CNR), "Progetto finalizzato: applicazioni cliniche della ricerca oncologica." 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 abbreviations used are: FN, fibronectin; mAb, monoclonal antibody; PAGE, polyacrylamide gel electrophoresis. et Vartio et al., 1987;Zardi et al., 1987;Barone et al., 1987;Carnemolla et al., 1989;Oyama et al., 1989Oyama et al., , 1990Borsi et al., 1992). In fact, the FN isoforms containing the IIICS, ED-A, and ED-B sequences are expressed to a greater degree in transformed human cells and in tumor tissues than in their normal counterparts. In particular, the FN isoform containing the ED-B sequence, which, with some very rare exceptions, is undetectable in normal adult tissues, has a much greater expression in fetal and tumor tissues as well as during wound healing (Norton and Hynes, 1987;Schwarzbauer et al., 1987;Gutman and Kornblihtt, 1987;Carnemolla et al., 1989;ffrench-Constant et al., 1989;ffrench-Constant andHynes, 1989, Laitinen et al., 1991). The ED-B oncofetal domain, a complete type I11 homology repeat composed of 91 amino acids and coded for by a single exon, is the most conserved FN region with 100% and 96% homology with rat and chicken FN, respectively (Norton and Hynes, 1987);Zardi et al., 1987). This could indicate either a more recent evolution of the ED-B sequence, with less time to diverge, or a more stringent requirement due to some unknown function(s) performed by this sequence. While the alternative spliced sequence IIICS contains a cell type-specific cell binding site, the biological functions of the ED-A and ED-B are still a matter of speculation (Humphries et al., 1986).
We describe two monoclonal antibodies, BC-1 and IST-6, which are specific for FN isoforms containing and not containing the ED-B sequence, respectively. Given this specificity, we assumed that the epitope recognized by the mAb BC-1 was localized within the ED-B sequence (Carnemolla et al., 1989). However, we now demonstrate that this epitope is localized within the type I11 repeat 7 (which precedes the ED-B) and that it is cryptic in FN molecules lacking the ED-B, while it is unmasked in molecules containing this sequence.

MATERIALS AND METHODS
Cell Lines, Monoclonal Antibodies (mAbs), Radioimmunoassay, and Purification of F N and Its Proteolytic Fragments-The cultured normal human fibroblast cell line GM-5659 (from non-fetal skin) and the SV4O-transformed WI-38-VA cell line were purchased from NIGMS, Human Genetic Mutant Cell Repository (Camden, NJ) and from the American Tissue Type Culture Collection, respectively. The mAb IST-6 was prepared as described using splenocytes from mice immunized with FN purified from human plasma (Zardi et al., 1980). The mAb BC-1 has been described previously (Carnemolla et al., 1989). FNs were purified from human plasma and from the conditioned medium of the various cell lines using a modification (Zardi et al. (1980)) of the procedure of Engvall and Ruoslahti (1977). The radioimmunoassay experiments, using purified FNs or FN fragments, were carried out as reported by Zardi et al. (1980). The 120-kDa (containing the ED-B) and 110-kDa (lacking the ED-B) FN domain 4 isoforms (see Fig. 1) were purified from thermolysin digests as previously described (Zardi et al., 1987); Borsi et al., 1991).
cDNA Clones and Fusion Proteins-cDNA clones XF2 and XFGc, obtained as previously reported (Carnemolla et al., 1989), cover the type 111 homology repeats 7, 8, and half of the 9 (from residues 1138 to 1380) (Kornblihtt et al., 1985). However, the ED-B sequence was included in the XF2 clone while it was omitted in the XF6c clone. In order to facilitate the reading of the manuscript, we have here named these clones according to the type 111 repeats or fractions of type 111 repeats they contain: XF2 = XF7.B.8.912; XF6c = XF7.8.912 (see Fig.  lA). The cDNA clones XF7.8118 and XF7 were obtained from the XF7.8.912 FN insert by digestion with BspMl and PV2, respectively. The XF7.8118 cDNA clone includes the type 111 homology repeat 7 and few amino acids of the 8 (from residues 1138 to 1239), while the XF7 cDNA clone is made up of repeat 7 (from residues 1138 to 1234). The inserts of the clones XF7.8 and XF713.8 were obtained from polymerase chain reaction amplification of the XF7.8.912 using Xgtll primers (New England Biolabs) and appropriate oligonucleotides. XF7.8 included the type I11 homology repeats 7 and 8 (from residues 1138 to 1318) while the clone XF713.8 included one-third of the 7 and the 8 (from residues 1167 to 1318). PCR reactions were performed for 35 cycles (1 min at 94 'C, 1 min a t 48 "C, and 1 min at 68 "C) in a final volume of 100 pl containing 50 m M KCl, 10 mM Tris-C1, pH 8.3, 1.5 mM MgC12, 0.01% (w/v) gelatin, 200 mM concentration of each dNTP, 100 pmol of each oligonucleotide, 5 units of Tag DNA polymerase (Amplitaq, Perkin-Elmer Cetus), and 10 ng of purified XF7.8.912 DNA as template. All cloning and subcloning procedures were carried out, using the expression vector Xgtll phage, according to Sambrook et al. (1989), and each DNA insert was analyzed using Sequenase version 2.0 DNA sequencing kit (United States Biochemical Corp.). The 8-galactosidase-FN fusion proteins were prepared as previously described (Carnemolla et al., 1989), and their immunoenzymatic reaction with mAbs IST-6 and BC-1 was carried out using a ProtoBlot immunoscreening system kit purchased from Promega Biotec. SDS-PAGE and immunoblotting were carried out as previously reported (Carnemolla et al., 1989). taining FN is absent in plasma and present in a very low percentage in FN from cultured normal skin fibroblasts, while it is present in a high percentage in FN from SV40-transformed WI-38-VA cells (Borsi et al., 1992;Carnemolla et aZ., 1989). Fig. 1 shows the results obtained by radioimmunoassay experiments using the mAbs BC-1 and IST-6 on FNs from these three different sources. The mAb BC-1 does not react with plasma FN and barely reacts with FN from normal skin fibroblasts, while it reacts strongly with FN from the SV40transformed WI-38-VA cell line. On the contrary, the mAb IST-6 gives a strong reaction with FNs from plasma and normal skin fibroblasts while it shows a weaker reaction with FN from WI-38-VA cells. Fig. 2 shows the results obtained by immunoscreening using the monoclonals BC-1 and IST-6 on lysis plaques generated by the clones XF7.B.8.912 and XF7.8.912 (Fig. 2). The mAb BC-1 reacts with the fusion protein XF7.B.8.912 but not with AF7.8.912. On the contrary, the mAb IST-6 does not react with the fusion protein XF7.B.8.912 but does with the fusion proteins XF7.8.912. Identical results were obtained in immunoblotting experiments after SDS-PAGE using the XF7.B.8.912 and XF7.8.912 fusion proteins (Fig. 3)  indicate the molecular masses, in kDa, of the standards. domains 4 (Zardi et al., 1987) with the mAbs BC-1 and IST-6 (Fig. 4). These results demonstrate that BC-1 and IST-6 are specific for FN molecules containing and not containing the ED-B sequence, respectively.

RESULTS AND DISCUSSION
In immunoblotting experiments using different FN-&galactosidase fusion proteins, we have more precisely localized the epitope recognized by the mAb BC-1. Even though the experiments shown in Figs. 1-4 could suggest that this epitope is localized within the ED-B sequence, since it is specific only for ED-B-containing FN molecules or fragments, the results shown in Fig. 5 demonstrate that it is localized within the type I11 homology repeat 7 and that the presence of the type I11 homology repeat 9 makes this epitope cryptic, while the expression of the ED-B sequence unmasks it. In fact, a FN-P-galactosidase fusion protein identical with the hF7.8.912 but lacking the repeat 9 reacts strongly with the mAb BC-1. In addition, the P-galactosidase-FN fusion protein hF7 which contains only the repeat 7 also shows the same positive reaction with the mAb BC-1. The mAb BC-1 also reacted with a P-galactosidase-FN fusion protein which contains the repeats 7 and ED-B (data not shown) but did not show any reaction with P-galactosidase-FN fusion proteins containing, respectively, only the repeat ED-B or the repeat 9 (data not shown).
The epitope recognized by the mAb IST-6 was more problematic to localize; in fact, the minimum requirement for a positive reaction with the mAb IST-6 for a FN-@-galactosidase fusion protein was the simultaneous presence of the complete type I11 homology repeats 7 and 8 (Fig. 5).
The main observations described here are that 1) the mAb BC-1 is specific for ED-B-containing FN molecules even though the epitope recognized by this mAb is not localized within the ED-B sequence and 2) the mAb IST-6 is strictly specific for FN molecules lacking the ED-B sequence. Thus, these two mAbs represent useful reagents to study the distribution of different FN isoforms in different tissues. Furthermore, these data demonstrate that the presence of the ED-B sequence within the FN molecules induces conformational modification in the central part of the molecule which, in turn, leads to the unmasking of a previously cryptic sequence within the type I11 repeat 7 and to the masking of the epitope recognized by the mAb IST-6. The fact that these structures are detectable also in immunoblotting experiments, after electrophoresis in presence of SDS, indicates that strong interactions between different FN regions take place. It has been reported that the free sulfhydryl group present in the repeat 7 is cryptic when plasma FN is in solution, while it is unmasked in FN bound to solid phase substrates (Narasimhan et al., 1988;Narasimhan and Lai, 1991). However, the mAb BC-1 recognized an epitope in which this cysteine is not involved because, in FN molecules lacking the ED-B, it is cryptic when FN is in solution or when it is bound to solid phase substrates. Furthermore, in the 0-galactosidase-FN fusion protein hF7.8.912 (see Fig. 5), we have substituted the cysteine present in the repeat 7 with a serine; the reactivity of this mutant with mAbs BC-1 and IST-6 did not show any modification (data not shown). At present, we have no evidence of whether these conformational modifications may influence the biological functions of FN. However, it has been reported that the cell binding site GRGD needs two synergistic regions in order to express its activity, and these regions have been localized within type I11 homology repeats 8 and 9, respectively (Obara et al., 1988);Nagai et al., 1991). Further studies are needed in order to clarify the structural conformation of this crucial region of the FN molecule and the possible functional modifications induced by the expression of the ED-B sequence.