JBC Oz Biosciences

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Originally published In Press as doi:10.1074/jbc.M201238200 on March 25, 2002

J. Biol. Chem., Vol. 277, Issue 22, 20095-20103, May 31, 2002
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow All Versions of this Article:
277/22/20095    most recent
M201238200v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Chen, H.
Right arrow Articles by MacLeod, J. N.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Chen, H.
Right arrow Articles by MacLeod, J. N.
Social Bookmarking
Add to CiteULike   Add to Complore   Add to Connotea   Add to Del.icio.us   Add to Digg   Add to Reddit   Add to Technorati  
What's this?

Absence of the I-10 Protein Segment Mediates Restricted Dimerization of the Cartilage-specific Fibronectin Isoform*

Hao ChenDagger , Da-Nian GuDagger , Nancy Burton-Wurster§, and James N. MacLeodDagger

From the Dagger  Department of Biomedical Sciences and the § Department of Microbiology and Immunology, James A. Baker Institute for Animal Health, College of Veterinary Medicine, Cornell University, Ithaca, New York 14853

Received for publication, February 6, 2002

    ABSTRACT
TOP
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
REFERENCES

The cartilage-specific (V + C)- fibronectin isoform does not efficiently heterodimerize with other V-region splice variants of fibronectin. To understand better the structural elements that determine this restricted dimerization profile, a series of truncated fibronectin expression constructs with various internal deletions in the V, III-15, or I-10 segments were constructed and co-transfected into COS-7 cells with either the V+C+ or the (V + C)- isoform. SDS-PAGE and immunoblot analyses of the resulting conditioned media suggest that the I-10 segment must either be present in both monomeric subunits of fibronectin or absent from both subunits for efficient dimerization to occur. Further studies suggest that the I-10 segment specifically, not simply a balanced number of type I repeats at the carboxyl terminus of each monomeric subunit, plays an important role in determining different fibronectin dimerization patterns. Neither I-11 nor I-12 could be substituted for segment I-10 without significantly reducing the formation of heterodimers. Therefore, absence of segment I-10 explains why (V + C)- fibronectin is not found in heterodimeric configurations with other native V-region splice variants in cartilage. The unique dimerization pattern of (V + C)- fibronectin does not prevent matrix formation yet is consistent with this isoform having specialized properties in situ that are important for either the structural organization and biomechanical properties of cartilage or the regulation of a chondrocytic phenotype.

    INTRODUCTION
TOP
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
REFERENCES

Fibronectin (FN)1 is an extracellular glycoprotein that has important roles in cell adhesion and migration, cell differentiation and proliferation, cell morphology and cytoskeletal organization, tissue remodeling and wound repair, and cancer progression (1). It is expressed by cells primarily in an anti-parallel dimeric configuration, composed of two 200-250-kDa monomeric subunits that are linked together by a pair of disulfide bonds near their carboxyl termini (2-4). Dimers are formed in the endoplasmic reticulum and are subsequently transported through the Golgi and secreted (5). A portion of secreted, soluble FN is captured by the cell surface in a reversible and saturable manner (6-8), which is then irreversibly assembled into a fibrillar extracellular matrix that is insoluble in the detergent deoxycholate and consists of disulfide-stabilized multimers (8-11).

Alternative splicing of FN gene transcripts results in different protein isoforms. Four sites of alternative splicing have been reported, extra domain A, extra domain B, the variable (V or IIICS) region, and a cartilage-specific (C) region composed of nucleotides that encode protein segments III-15 and I-10. In adult canine and equine articular cartilage, 50-80% of the FN transcripts have an unique splicing pattern, designated (V + C)-, that deletes both the V and C regions (12). Dimerization patterns of the (V + C)- isoform were studied under native conditions within canine articular cartilage and experimentally in COS-7, NIH-3T3, and CHO-K1 cell cultures (13). In all systems, the (V + C)- isoform exists predominantly in a homodimeric configuration. Heterodimers composed of (V + C)- and the other V-region splice variants (V+C+ or V-C+) are either not observed or detected at only low levels. The homodimeric configuration of (V + C)- FN does not reflect the laws of random assortment (14, 15). By using isoform-specific constructs, it was shown that this pattern results from a problem with heterodimer formation involving the (V + C)- isoform, rather than secretion (13).

Different patterns of alternative splicing and dimerization have been shown to influence FN solubility and matrix assembly (1, 16-18). FN dimers containing the extra domain A and extra domain B domains are incorporated more efficiently into pre-existing matrices (18). In addition, Ichihara-Tanaka et al. (19) found that the segments III-15 and I-10 through I-12 are actively involved in FN matrix assembly, and deletion of even one of the type I modules reduces the matrix assembly activities. In the current study, we test the hypothesis that the restricted dimerization pattern of (V + C)- FN is related to the absence of III-15 and/or I-10 protein segments, and we study the matrix structure of this naturally occurring isoform.

    EXPERIMENTAL PROCEDURES
TOP
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
REFERENCES

Determinants of Heterodimerization

cDNA Constructs-- Deminectins (DNs) are amino-terminal truncations of rat FN extending from within segment III-8 to the carboxyl terminus (Fig. 1). Construction of DN1, DN2, and DN3 have been described previously, and they have been shown to model accurately the dimerization profiles of native full-length FN isoforms (13, 14, 20). A series of new DN constructs, containing various V, III-15, and I-10 segment deletions, were made by PCR-sequence overlap extension (PCR-SOEing) (21, 22). The procedure involved the generation of two PCR fragments. The first fragment extended from a 5' PstI site in the region encoding III-12 to a targeted junction site defined by the desired construct. The 5' end of the second fragment started from the same targeted junction site and extended to a 3' SalI site downstream of nucleotides encoding the carboxyl terminus. Oligonucleotide primers used to generate the two PCR fragments were designed with overlaps in the targeted junction region between 18 and 24 nucleotides in length (Table I). This overlap allowed fusion of the two fragments in a third PCR that used only the 5' (TGG TTC AGA CTG CAG TGA)- and 3' (TCT AGA GTC GAC CCG G)-flanking primers. The resulting fusion product was digested with PstI and SalI and substituted for the corresponding PstI/SalI fragment of DN1. The only exception was construction of DN14, which required two independent PCR-SOEing steps (DN14a and DN14b in Table I). Oligonucleotide primers were synthesized by Integrated DNA Technologies, Inc. (Coralville, IA), and PCR amplifications were carried out with Pfu DNA Polymerase (Stratagene, La Jolla, CA). Nucleotide identities of all new DNs were confirmed by bi-directional sequencing (Biotechnology Service, Cornell University, Ithaca, NY).

                              
View this table:
[in this window]
[in a new window]
 
Table I
Overlapping primers used for the construction of individual DN cDNAs by PCR-SOEing
A series of DN constructs, containing various V, III-15, and I-10 segment deletions, were made by PCR-SOEing. The procedure involved the generation of two PCR fragments. The first amplicon extended from a 5' PstI site in the region encoding III-12 to a targeted junction site defined by the desired construct. The 5' end of the second amplicon started from the same targeted junction site and extended to a 3' SalI site downstream of nucleotides encoding the carboxyl terminus. Oligonucleotide primers used to generate the two PCR amplicons were designed with overlaps in the targeted junction region (in bold). This overlap allowed fusion of the two cDNA fragments in a third PCR that used only the 5' (TGG TTC AGA CTG CAG TGA) and 3' (TCT AGA GTC GAC CCG G)-flanking primers containing the PstI and SalI restriction sites, respectively. The resulting fusion product was digested with PstI and SalI and substituted for the corresponding PstI/SalI fragment of DN1. The only exception was construction of DN14, which required two independent PCR-SOEing steps (DN14a and DN14b).

DN Expression and Dimerization-- Each DN construct, both singly and in combination with the other DNs, was transiently transfected into COS-7 cells (American Type Culture Collection number 1651-CRL, Manassas, VA) using Superfect transfection reagents following the protocol recommended by the manufacturer (Qiagen, Valencia, CA). The amount of DNA used for each co-transfection was adjusted empirically to achieve approximately equal expression levels. Twenty four hours after transfection, the cells were re-fed with Dulbecco's modified Eagle's medium, 10% fetal bovine serum. Conditioned media were collected another 24 h later, clarified by centrifugation (175 × g), stabilized with protease inhibitors (0.3 mM benzamidine, 20 mM EDTA, 10 mM N-ethylmaleimide, 0.4 mM phenylmethylsulfonyl fluoride, final concentrations), and frozen at -20 °C pending analysis of dimerization patterns.

Samples were subjected to electrophoresis and immunoblot analysis by heating at 95 °C for 5 min in the presence of 0.2% (w/v) SDS without beta -mercaptoethanol. Heterodimers, homodimers, and monomers of recombinant DNs were separated according to size on 6% SDS-polyacrylamide gels in Tris glycine buffer at pH 8.6 (23), transferred onto nitrocellulose membranes (24), and probed with a rat FN-specific monoclonal antibody IC-3 (25) (a gift from Dr. Jean Schwarzbauer at Princeton University), followed by either a peroxidase-linked (Sigma) or a 35S-labeled anti-mouse IgG antibody (Amersham Biosciences). The binding epitope of IC-3 has been localized to the cell binding region within III-8 to III-11,2 which is common to all FN isoforms. Relative binding affinities of IC-3 to the DN constructs are equivalent to anti-human FN polyclonal IgG (ICN Pharmaceuticals Inc., Aurora, OH; data not shown). Peroxidase activity was detected by chemiluminescence (ECL Western blotting detection system, Amersham Biosciences), and 35S decay events were measured by a PhosphorImager (Fuji Photo Film Co., Ltd., Japan). Band intensities on radiograph films or PhosphorImager plates were quantified using MacBas software (Fuji Photo Film Co., Ltd.). Quantitative data obtained with the two protein detection strategies (ECL versus 35S-labeled secondary antibody) yielded comparable results (data not shown). Therefore, for DNs with higher level of expression, band intensities were quantified with the 35S-labeled secondary antibody. The increased sensitivity of the ECL system was useful in samples where DN expression was reduced. Relative homodimer and heterodimer percentages were calculated with total band intensity of DN set at 100%. All quantitative data are presented as the mean ± S.D. of triplicate co-transfection experiments.

Expression of Recombinant FN isoforms and Matrix Structure

Expression and Purification of Recombinant FN Isoforms-- Construction of recombinant baculovirus containing the rat V+C+ construct has been described previously (26). Full-length (V + C)- FN cDNA was subcloned into the baculovirus vector PVL 1392 (BD PharMingen). Recombinant viruses were generated according to established procedures (27). To produce recombinant protein, 50% confluent T175 flasks of Sf21 cells that had been adapted for growth under serum-free conditions (gift from Dr. Ping Wang at Cornell University) were incubated with recombinant viruses for 2 h in Excell-400 medium (JRH Biosciences, Lenexa, KS). After re-feeding, conditioned media from infected cells were collected 48 h later and clarified by centrifugation (175 × g). The mixture of protease inhibitors (above) was added to inhibit proteolysis. Recombinant FN proteins were purified using affinity chromatography columns of gelatin-agarose (Amersham Biosciences), followed by heparin-agarose (Pierce) as described by Poulouin et al. (28). The concentration of purified FN was determined by enzyme-linked immunosorbent assay using a polyclonal (goat) antiserum to human FN, followed by peroxidase-conjugated rabbit anti-goat IgG (ICN Pharmaceuticals Inc.) as described previously (29).

Matrix Structure-- FN null mouse embryonic fibroblasts (30) were a gift of Dr. Deane Mosher (University of Wisconsin, Madison). The fibroblasts were plated into the wells (3 × 104 cells/well) of immunofluorescence slides (Polysciences, Warrington, PA) in Dulbecco's modified Eagle's medium supplemented with 10% FN-depleted fetal bovine serum (31), glutamine (0.29 mg/ml), and penicillin (100 units/ml)/streptomycin (100 µg/ml). After a 3-h period for cell attachment and spreading, media were removed, and the cells were re-fed with Dulbecco's modified Eagle's medium containing 30 µg/ml of either rat plasma FN (pFN, Sigma), recombinant V+C+ FN, or recombinant (V + C)- FN. The cellular assembly of a FN matrix was allowed to proceed for 24 h.

For immunofluorescence microscopy, cells were fixed with 3.5% paraformaldehyde for 15 min at 4 °C and incubated sequentially with a monoclonal antibody against rat FN (IC-3) in Tris-buffered saline containing 0.1% bovine serum albumin overnight at 4 °C and goat anti-mouse IgG conjugated to fluorescein isothiocyanate (Molecular Probe, Eugene, OR) overnight at 4 °C. After rinses, the slides were covered with prolong anti-fade mounting media (Molecular Probe). The pattern of fluorescence was assessed with an Olympus IX70 confocal microscope using an argon laser with 488 nm excitation and bandpass filters for collecting green fluorescence.

    RESULTS
TOP
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
REFERENCES

Segment I-10 Influences FN Heterodimerization Patterns-- Unlike the V+C+ and V-C+ FN isoforms that can heterodimerize efficiently with each other, the cartilage-specific (V + C)- FN isoform exists predominantly as homodimers within canine articular cartilage (13). This dimerization pattern of (V + C)- FN is retained in experimental cell culture models (13) and led to the hypothesis that absence of the III-15 and/or I-10 protein segments restricts heterodimerization with the other two V-region splice variants (V+C+ or V-C+). To test the hypothesis, an experimental system using truncated versions of FN, termed DNs, was utilized. In addition to DN1 representing the V+C+ isoform, DN2 representing the V-C+ isoform, and DN3 representing the (V + C)- isoform, five new DN constructs, DN4 to DN8, were made with various combinations of the V, III-15, and I-10 segments deleted (Fig. 1). Either DN1 or DN3 was transiently co-transfected into COS-7 cells with one of the other seven DNs. Homodimers, heterodimers, and monomers of expressed DNs in conditioned medium were separated by SDS-PAGE under non-reducing conditions and detected using immunoblots. Culture media from cells transfected with only single DNs were run in parallel to confirm the identity of homodimers.


View larger version (40K):
[in this window]
[in a new window]
  		Fig. 1.   Schematic representation of FN protein structure and derived truncated DN constructs. Monomeric FN protein structure is illustrated with types I, II, and III homologous segments depicted by rectangles, circles, and squares, respectively. Two cysteine residues near the carboxyl terminus necessary for dimerization are indicated by S. In the endoplasmic reticulum, two FN monomers align in an anti-parallel orientation to form homodimers or heterodimers. DNs contain the carboxyl-terminal half of FN starting from within protein segment III-8. DN1, DN2, and DN3 represent native splice variants of FN. The cartilage-specific (V + C)- FN isoform, represented by DN3, deletes not only nucleotides encoding the entire V-region but also the III-15 and I-10 segments. DN4 to DN8 represent other theoretical deletions of V, III-15, and I-10. DN9, DN10, DN13, and DN14 have deletions or duplications of the carboxyl-terminal type I repeats as schematically indicated.

As illustrated in Fig. 2, there are two dimerization patterns observed following co-transfection of DN1 with other DNs. They either have the potential to form a heterodimer or remain only as their corresponding homodimers. For example, DN1 does not form heterodimers with DN6 (Fig. 2, panel a). When they are co-transfected into COS-7 cells, only bands corresponding to DN1 and DN6 homodimers are formed (Fig. 2, lane 2). In contrast, DN1 heterodimerizes efficiently with DN7 (Fig. 2, panel b). Co-transfection of these two DNs results in a large amount of heterodimers (Fig. 2, lane 5) that have an intermediate size between DN1 (Fig. 2, lane 4) and DN7 homodimers (Fig. 2, lane 6). Since DN1, DN4, and DN5 are very similar in size, DN1/DN4 and DN1/DN5 heterodimers could not be resolved by 6% SDS-PAGE. Therefore, full-length rat V+C+ FN (FN1) was co-transfected with DN4 or DN5. FN1 heterodimerizes with DN5 (Fig. 2, panel c) but not DN4 (not shown). Fig. 2B summarizes the quantitative ratios of homodimer/heterodimer formation in each co-transfection combination with DN1 (or FN1). DN1 (or FN1) does not heterodimerize well with DN3, DN4, DN6, and DN8, whereas it heterodimerizes efficiently with DN2, DN5, and DN7.


View larger version (41K):
[in this window]
[in a new window]
  		Fig. 2.   DN1 (or FN1), representing the V+C+ isoform, displays variable heterodimerization potential with other DN constructs. A, conditioned culture media from COS-7 cells following transient transfection with the indicated cDNA construct(s) were analyzed by SDS-PAGE and immunoblotting. Panel a, DN1 does not form heterodimers with DN6. They remain as their corresponding homodimers following co-transfection (lane 2). Lanes 1 and 3 indicate DN1 and DN6 homodimers, respectively. Panel b, DN1 heterodimerizes efficiently with DN7. Co-transfection results in a large amount of DN1/DN7 heterodimers (lane 5). Lanes 4 and 6 illustrate DN1 and DN7 homodimers, respectively. Panel c, FN1 heterodimerizes with DN5. Co-transfection of FN1 and DN5 generates high levels of intermediate-sized FN1/DN5 heterodimers (lane 8). Lanes 7 and 9 indicate single transfections of FN1 and DN5, respectively. B, relative homodimer and heterodimer percentages (mean ± S.D.) determined from immunoblots by either direct quantification of chemiluminescence or decay events from 35S-labeled secondary antibody. DN1 heterodimerizes very inefficiently with DN3, DN6, and DN8, whereas it heterodimerizes efficiently with DN2 and DN7. Homodimers and/or heterodimers from co-transfection of DN1/DN4 and DN1/DN5 could not be resolved by 6% SDS-PAGE. Therefore, full-length V+C+ FN (FN1) was substituted and used for the DN4 and DN5 analyses.

A similar experimental strategy was used to analyze the heterodimerization potential of DN3. Again, two general dimerization patterns are observed (Fig. 3). For example, DN3 heterodimerizes with DN4 (Fig. 3, panel a) and with DN6 (Fig. 3, panel c). Co-transfection of DN3 with either DN4 or DN6 results in a high level of intermediate-sized heterodimers. In contrast, DN3 does not heterodimerize with DN5 (Fig. 3, panel b). Co-transfected cells express only DN3 and DN5 homodimers (Fig. 3, lane 5). Fig. 3B lists the relative percentages of DN3 homodimers, DN3/DNX heterodimers, and DNX homodimers. DN3 does not heterodimerize with DN1 and DN5 but heterodimerizes with the other DNs to varying degrees.


View larger version (41K):
[in this window]
[in a new window]
  		Fig. 3.   DN3, representing the (V + C)- isoform, displays variable heterodimerization potential with other DN constructs. A, conditioned culture media from COS-7 cells following transient transfection with the indicated cDNA construct(s) were analyzed by SDS-PAGE and immunoblotting. Panel a, DN3 heterodimerizes with DN4. Single transfection of DN3 (lane 1) or DN4 (lane 3) generates their corresponding homodimers. When these two DNs are co-transfected into COS-7 cells, high levels of intermediate-sized DN3/DN4 heterodimers are formed (lane 2). Panel b, DN3 does not heterodimerize with DN5. No DN3/DN5 heterodimers are formed in cells co-transfected with DN3 and DN5 (lane 5). They remain as DN3 homodimers (lane 4) and DN5 homodimers (lane 6). Panel c, DN3 forms heterodimers with DN6. Co-transfection of these two DNs results in the formation of a large amount of DN3/DN6 heterodimers (lane 8). Lanes 7 and 9 illustrate the homodimers formed following transfections with DN3 and DN6 alone. B, relative homodimer and heterodimer percentages (mean ± S.D.) were determined from immunoblots by either direct quantification of chemiluminescence or decay events from 35S-labeled secondary antibody. DN3 does not heterodimerize with DN1 and DN5 but demonstrates variable heterodimerization potential with the other DN constructs.

Comparative heterodimerization patterns of DN1 and DN3 are summarized in Table II. The percentages of DN1/DNX and DN3/DNX heterodimers are scored in deciles by asterisks (<= 10%, *; 11-20%, **; 21-30%, ***; 31-40%, ****; >= 41%, *****). The data indicate that isoforms that include segment I-10 heterodimerize preferentially with DN1, which also includes the I-10 segment. Similarly, isoforms that lack protein segment I-10 heterodimerize preferentially with DN3, which also lacks the I-10 segment. For each DNX, the level of DN1/DNX heterodimerization is significantly different (p < 0.05) than the level of DN3/DNX heterodimerization. The data also indicate that the presence or absence of protein segments V and III-15 does not influence substantially whether heterodimerization is favored with DN3 or DN1.

                              
View this table:
[in this window]
[in a new window]
 
Table II
Summary of the dimerization patterns of DN1 and DN3 with co-transfected DNs
The presence or absence of V, III-15, and I-10 protein segments in the various DN constructs is indicated by the + or - sign. Relative dimerization efficiencies (from Figs. 2B and 3B) were scored in deciles by asterisks as indicated. Percent heterodimerization of DNs: *, <= 10%; **, 11-20%; ***, 21-30%; ****, 31-40%; *****, >= 41%. The data demonstrate the following: 1) a DN that lacks protein segment I-10 heterodimerizes preferentially with DN3, which also lacks the I-10 segment; 2) a DN that includes the I-10 segment heterodimerizes preferentially with DN1 (or FN1), which also includes the I-10 segment; and 3) the presence or absence of protein segment V and III-15 does not influence substantially whether heterodimerization is favored with DN3 or DN1.

Importance of a Balanced Number of Type I Repeats at the Carboxyl Terminus-- Is there something specific about protein segment I-10 that regulates the heterodimerization patterns of DN1 and DN3, or is it simply necessary to have a balanced number of type I repeats at the carboxyl terminus? For example, it could be necessary to have the same number of carboxyl-terminal type I repeats on each monomeric FN subunit for efficient interchain association or as structural elements for proper alignment. To answer this question, DN9 and DN10, which delete I-11 and I-12 protein segments, respectively, were prepared (Figs. 1 and 4B). DN4, DN9, and DN10 all contain V and III-15, but each has a different combination of two type I repeats at the carboxyl terminus. These constructs were co-transfected individually with DN3 into COS-7 cells. The results are shown in Fig. 4. If a balanced number of type I repeats at the carboxyl terminus determines heterodimerization efficiency, then DN3 would be expected to form similar amounts of heterodimers with DN4, DN9, and DN10. Although DN3 does heterodimerize with all three (Fig. 4A), the levels of heterodimer formation are significantly different (Fig. 4B). DN3 heterodimerizes most efficiently with DN4, both of which lack protein segment I-10.


View larger version (32K):
[in this window]
[in a new window]
  		Fig. 4.   Absence of segment I-10, not simply two type I repeats at the carboxyl terminus, enhances heterodimerization efficiency with DN3. A, to determine whether heterodimerization patterns are influenced by protein segment I-10 specifically or simply a balanced number of type I repeats at the carboxyl terminus, DN3 was co-transfected into COS-7 cells with either DN4, DN9, or DN10. These three DNs all have the V and III-15 segments, but each has a different combination of two type I repeats at the carboxyl terminus. Panel a, co-transfection of DN3 and DN4. Lane 1, DN3 homodimers; lane 2, homodimers and DN3/DN4 heterodimers; lane 3, DN4 homodimers. Panel b, co-transfection of DN3 and DN9. Lane 4, DN3 homodimers; lane 5, homodimers and DN3/DN9 heterodimers; lane 6, DN9 homodimers. Panel c, co-transfection of DN3 and DN10. Lane 7, DN3 homodimers; lane 8, homodimers and DN3/DN10 heterodimers; lane 9, DN10 homodimers. B, the table indicates the presence (+) or absence (-) of carboxyl-terminal type I protein segments in DN3, DN4, DN9, and DN10, and relative levels of DN3/DNX heterodimer formation. Heterodimerization data are presented in percentages (mean ± S.D.) of triplicate experiments. Superscript letters symbolize differences at p < 0.05. DN3 heterodimerizes most efficiently with DN4, which also lacks protein segment I-10.

DN13 and DN14 were constructed to enable a parallel analysis with DN1. They are structurally similar to DN7, all three lacking V and III-15, but segment I-10 is removed and substituted with duplications of either I-11 (DN13) or I-12 (DN14) (Figs. 1 and 5B). As reported in Fig. 5, these changes greatly reduce the formation of DN1 heterodimers. Once again, efficient heterodimerization of different DN isoforms requires that both subunits have the same I-10 status. In this case, both DN1 and DN7 contain protein segment I-10 and heterodimerize most efficiently.


View larger version (32K):
[in this window]
[in a new window]
  		Fig. 5.   Presence of segment I-10, not simply three type I repeats at the carboxyl terminus, enhances heterodimerization efficiency with DN1. A, to confirm the role of segment I-10 in determining heterodimerization patterns, DN1 was co-transfected into COS-7 cells with either DN7, DN13, or DN14. Each of these lack the V and III-15 segments, but only DN7 retains segment I-10. In DN13 and DN14, segment I-10 has been removed and substituted with I-11 and I-12, respectively. Panel a, co-transfection of DN1 and DN7. Lane 1, DN1 homodimers; lane 2, homodimers and DN1/DN7 heterodimers; lane 3, DN7 homodimers. Panel b, co-transfection of DN1 and DN13. Lane 4, DN1 homodimers; lane 5, homodimers and DN1/DN13 heterodimers; lane 6, DN13 homodimers. Panel c, co-transfection of DN1 and DN14. Lane 7, DN1 homodimers; lane 8, DN1 and DN14 do not heterodimerize, they remain as homodimers; lane 9, DN14 homodimers. B, the table indicates the presence (+) or absence (-) of carboxyl-terminal type I segments in DN1, DN7, DN13, and DN14 and relative levels of DN1/DNX heterodimer formation. Heterodimerization data are presented in percentages (mean ± S.D.) of triplicate experiments. Superscript letters symbolize differences at p < 0.05. DN1 heterodimerizes most efficiently with DN7, which also contains segment I-10.

FN Matrix Structure-- The unique primary structure and restricted heterodimerization of (V + C)- FN raise the potential that this isoform may have different matrix assembly characteristics that influence the matrix organization and biomechanical properties of cartilage. To explore this possibility and compare FN matrix structure, pFN, recombinant V+C+ FN, and recombinant (V + C)- FN (30 µg/ml) were added to mouse embryonic fibroblasts that lack endogenous FN expression. Matrix assembly was allowed to proceed for 24 h. Immunofluorescence microscopy indicated that all three FN preparations are able to be assembled into a matrix and have similar linear arrays of fibrils (Fig. 6). Compared with pFN and V+C+ FN, the fibrils of the (V + C)- FN are subjectively less extensive (Fig. 6, A-C). At higher magnification, however, no differences in either the pattern or the thickness of fibrils are appreciated (Fig. 6, D-F).


View larger version (99K):
[in this window]
[in a new window]
  		Fig. 6.   Matrices of rat plasma FN, V+C+ FN, and (V + C)- FN. FN null fibroblasts were cultured in medium containing 30 µg/ml rat plasma FN (pFN, A and D), recombinant V+C+ FN (B and E), or recombinant (V + C)- FN (C and F). After 24 h, fibrillar FN matrices were detected for all FN isoforms by immunofluorescence microscopy using rat-specific monoclonal antibody IC3 and fluorescein isothiocyanate-conjugated goat anti-mouse IgG. Visual assessment at ×20 (A-C) suggests less extensive matrix formation with (V + C)- FN. However, under ×60 objective (D-F), no marked differences in either the pattern or thickness of fibrils between three FN preparations are appreciated. Bar equals 10 µm.


    DISCUSSION
TOP
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
REFERENCES

The cartilage-specific FN isoform, (V + C)-, is expressed predominantly in homodimeric configurations both within native cartilaginous tissues and in experimental cell culture systems (13). Heterodimers between the (V + C)- isoform and other V-region splice variants (V+C+ or V-C+) are either not observed or detected at only low levels. This is not due to a problem with secretion, rather it is that (V + C)- FN heterodimers simply are not produced efficiently by either chondrocytes or transfected cell lines. In this report, we tested the hypothesis that lack of the III-15 and/or I-10 protein segments in (V + C)- FN restricts its ability to heterodimerize with other V-region splice variants. By constructing truncated DNs with various deletions of the V, III-15, or I-10 segments and co-transfecting them with either DN1 (V+C+ DN) or DN3 ((V + C)- DN), we have shown that heterodimerization occurs with increased efficiency when the two monomeric subunits either both contain protein segment I-10 or both have segment I-10 deleted (data summarized in Table II). Levels of heterodimerization are reduced when one subunit contains I-10 and the other lacks I-10. Based on this relationship, the absence of protein segment I-10 in (V + C)- FN appears to explain why this isoform is not found in heterodimeric configurations with the other native V-region splice variants within cartilage.

Analyses of FN dimers by high performance liquid chromatography (2) and nuclear magnetic resonance spectroscopy (3) reveal that the two monomeric subunits are linked in an anti-parallel fashion. Two coiled segments at the carboxyl terminus, each composed of only four amino acids, may serve as local interchain recognition sites (4). Although not an absolute structural determinant of dimerization, two monomeric subunits that match in terms of the presence or absence of segment I-10 may be more efficient at achieving the alignment that is necessary to bring the carboxyl-terminal cysteine pairs and adjacent coiled segments into close proximity with an anti-parallel orientation prior to disulfide bond formation. Alternatively, conformational relationships mediated by segment 1-10 may allow efficient binding of protein-disulfide isomerase (32-35) or interactions with protein segment I-12 of FN that has been shown to have intrinsic protein-disulfide isomerase activity (36). The formation and rearrangement of intra- and intermolecular disulfide bonds mediated by protein-disulfide isomerase are critical to the stabilization of FN dimers and multimers. As with other enzymes, the catalytic capability of protein-disulfide isomerase is related to its ability to bind substrates and stabilize the intermediate enzyme-substrate complexes (37-40). In support of the current data, studies by other groups (19, 41) have also suggested that the three type I repeats at the carboxyl terminus may have important functional roles in dimerization related to interchain associations and disulfide-bond formation. Deletional mutants of FN lacking segments I-10 through I-12 or III-15 and I-10 through I-12 failed to heterodimerize with endogenous mouse FN when stably expressed in mouse L cells, although they retained the ability to form homodimers (19). Additionally, Sottile and Mosher (41) found that only 30% of FN deleting I-10 through I-12 segments was dimeric as compared with 60% for endogenous COS cell FN.

Considering the findings of these two papers and the data presented in Figs. 1 and 2 of the current study, an important issue is whether segment I-10 specifically plays an important role in FN dimerization. Another possibility is that I-10 is no more important than I-11 or I-12 in this context. Each monomeric subunit may only need to have the same number of type I repeats at the carboxyl terminus for efficient dimerization. Addressing this question using the V+C+ isoform as represented by DN1, segment I-10 appears to be critical. Deleting I-10 but maintaining a total of three type I repeats carboxyl terminus by duplicating either I-11 (DN13) or I-12 (DN14) reduced heterodimer formation with DN1 to near zero (Fig. 5). A similar relationship was observed with the (V + C)- isoform (Fig. 4), but the constraints were not absolute. DN3/DN9 and DN3/DN10 heterodimerization did occur, albeit at significantly lower levels. These data do not indicate that a balanced number of type I repeats at the carboxyl terminus of each FN monomer is not a variable that influences dimerization efficiency; we think it probably is a variable. Rather, segment I-10 appears to have a role over and above this issue by a structural or functional mechanism yet to be determined. Nuclear magnetic resonance spectroscopy of recombinant I-4 and I-5 indicates that these two amino-terminal type I repeats interact structurally. A tryptophan unique to I-4 interacts with an arginine residue in I-5 creating a hydrophobic interface that fixes these two segments in a constrained conformation (42). Similar inter-module interactions involving I-10 might be important for maintaining conformational relationships at the carboxyl terminus that are important for disulfide bonding and FN dimerization.

Different isoforms and dimerization patterns have been shown to influence FN solubility and matrix assembly (1, 16-18). Ichihara-Tanaka et al. (19) reported that the III-15 and I-10 through I-12 segments are actively involved in efficient incorporation of FN into the extracellular matrix. A deletion en bloc of these segments or the deletion of any single carboxyl-terminal type I repeat markedly impairs the matrix assembly of recombinant FN. In addition, Sottile and Mosher (41) found that matrix assembly is decreased ~60% in FNs that have segments I-10 through I-12 deleted. Our findings indicate that the native (V + C)- FN isoform can be assembled into an extensive fibrillar matrix even though segments III-15 and I-10 are absent (Fig. 6). Compared with pFN and recombinant V+C+ FN, the matrix distribution of recombinant (V + C)- FN is subjectively less extensive. However, no marked differences in either the pattern or thickness of FN fibrils are appreciated by light microscopy.

The molecular mechanism by which FN is assembled into a fibrillar matrix is an area of active investigation. Matrix assembly appears to be initiated by binding of FN to alpha 5beta 1 integrin (43-47) and to unknown protein/carbohydrate sites on the cell surface, called "matrix assembly sites" (6, 7, 48, 49). The binding of compact soluble FN to the cell surface exposes cryptic FN-FN interaction sites (50-53) and facilitates multimer formation (10, 11, 54-57). Release of FN from the cell surface to adjacent FN molecules regenerates cell surface-binding sites (9). Polymerization of FN involves not only the carboxyl-terminal heparin II binding domain (45) but also its interaction with structural elements at the amino terminus of the same or a different molecule (58-60). Epitope mapping of a monoclonal antibody specific for (V + C)- FN supports the presence of unique structural features in the heparin II-binding region of this isoform created by the novel juxtaposition of protein segments III-14 and I-11.3 The changes are also reflected functionally in terms of altered proteoglycan binding properties (61). Therefore, the unique structural features of the cartilage-specific (V + C)- FN isoform that change dimerization and the heparin II binding region clearly have the potential to influence matrix organization and contribute to the specialized biomechanical properties of cartilage that provide load distribution and facilitate joint motion.

    ACKNOWLEDGEMENTS

We are grateful to Dr. Jean Schwarzbauer (Princeton University) for the gifts of DN1, DN2, V+C+ FN constructs, and the IC-3 monoclonal antibody; Dr. Deane Mosher (University of Wisconsin, Madison) for the gift of the FN null mouse embryonic fibroblasts; and Dr. Ping Wang (Cornell University) for Sf21 cells. We appreciate the secretarial support of Dorothy Scorelle.

    FOOTNOTES

* This work was supported by National Institutes of Health Grant AR44340 and the Arthritis Foundation.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

To whom correspondence should be addressed. Tel.: 607-256-5645; Fax: 607-256-5608; E-mail: jnm1@cornell.edu.

Published, JBC Papers in Press, March 25, 2002, DOI 10.1074/jbc.M201238200

2 J. Schwarzbauer, personal communication.

3 N. Burton-Wurster, H. Chen, R. Gendelman, M. L. Jackson, L. F. Gagliardo, D.-N. Gu, R. R. Zelko, G. Lust, and J. N. MacLeod, submitted for publication.

    ABBREVIATIONS

The abbreviations used are: FN, fibronectin; DNs, deminectins; PCR-SOEing, polymerase chain reaction-sequence overlap extension; pFN, rat plasma FN; V-region, variable region of fibronectin alternative splicing.

    REFERENCES
TOP
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
REFERENCES

1. Hynes, R. O. (1990) Fibronectins , pp. 200-364, Springer-Verlag Inc., New York
2. Skorstengaard, K., Jensen, M. S., Sahl, P., Petersen, T. E., and Magnusson, S. (1986) Eur. J. Biochem. 161, 441-453[Medline] [Order article via Infotrieve]
3. An, S. S., Jimenez-Barbero, J., Petersen, T. E., and Llinas, M. (1992) Biochemistry 31, 9927-9933[CrossRef][Medline] [Order article via Infotrieve]
4. Kar, L., Lai, C. S., Wolff, C. E., Nettesheim, D., Sherman, S., and Johnson, M. E. (1993) J. Biol. Chem. 268, 8580-8589[Abstract/Free Full Text]
5. Choi, M. G., and Hynes, R. O. (1979) J. Biol. Chem. 254, 12050-12055[Abstract/Free Full Text]
6. McKeown-Longo, P. J., and Mosher, D. F. (1985) J. Cell Biol. 100, 364-374[Abstract/Free Full Text]
7. Quade, B. J., and McDonald, J. A. (1988) J. Biol. Chem. 263, 19602-19609[Abstract/Free Full Text]
8. McKeown-Longo, P. J., and Mosher, D. F. (1989) in Fibronectin (Mosher, D. F., ed) , pp. 163-179, Academic Press Inc., San Diego
9. Mosher, D. F., Sottile, J., Wu, C., and McDonald, J. A. (1992) Curr. Opin. Cell Biol. 4, 810-818[CrossRef][Medline] [Order article via Infotrieve]
10. Sechler, J. L., Corbett, S. A., Wenk, M. B., and Schwarzbauer, J. E. (1998) Ann. N. Y. Acad. Sci. 857, 143-154[CrossRef][Medline] [Order article via Infotrieve]
11. Schwarzbauer, J. E., and Sechler, J. L. (1999) Curr. Opin. Cell Biol. 11, 622-627[CrossRef][Medline] [Order article via Infotrieve]
12. MacLeod, J. N., Burton-Wurster, N., Gu, D. N., and Lust, G. (1996) J. Biol. Chem. 271, 18954-18960[Abstract/Free Full Text]
13. Burton-Wurster, N., Gendelman, R., Chen, H., Gu, D.-N., Tetreault, J. W., Lust, G., Schwarzbauer, J. E., and MacLeod, J. N. (1999) Biochem. J. 341, 555-561[Medline] [Order article via Infotrieve]
14. Schwarzbauer, J. E., Mulligan, R. C., and Hynes, R. O. (1987) Proc. Natl. Acad. Sci. U. S. A. 84, 754-758[Abstract/Free Full Text]
15. Schwarzbauer, J. E. (1991) Bioessays 13, 527-533[CrossRef][Medline] [Order article via Infotrieve]
16. Yamada, K. M., Schlesinger, D. H., Kennedy, D. W., and Pastan, I. (1977) Biochemistry 16, 5552-5559[CrossRef][Medline] [Order article via Infotrieve]
17. Alexander, S. S. J., Colonna, G., and Edelhoch, H. (1979) J. Biol. Chem. 254, 1501-1505[Abstract/Free Full Text]
18. Guan, J.-L., Trevithick, J. E., and Hynes, R. O. (1990) J. Cell Biol. 110, 833-847[Abstract/Free Full Text]
19. Ichihara-Tanaka, K., Titani, K., and Sekiguchi, K. (1995) J. Cell Sci. 108, 907-915[Abstract]
20. Schwarzbauer, J. E., Spencer, C. S., and Wilson, C. L. (1989) J. Cell Biol. 109, 3445-3453[Abstract/Free Full Text]
21. Ho, S. N., Hunt, H. D., Horton, R. M., Pullen, J. K., and Pease, L. R. (1989) Gene (Amst.) 77, 51-59[CrossRef][Medline] [Order article via Infotrieve]
22. Horton, R. M., Hunt, H. D., Ho, S. N., Pullen, J. K., and Pease, L. R. (1989) Gene (Amst.) 77, 61-68[CrossRef][Medline] [Order article via Infotrieve]
23. Laemmli, U. K. (1970) Nature 277, 680-685
24. Towbin, H., Staehlin, T., and Gordon, J. (1979) Proc. Natl. Acad. Sci. U. S. A. 76, 4350-4354[Abstract/Free Full Text]
25. Castle, A. M., Schwarzbauer, J. E., Wright, R. L., and Castle, J. D. (1995) J. Cell Sci. 108, 3827-3837[Abstract]
26. Sechler, J. L., Takada, Y., and Schwarzbauer, J. E. (1996) J. Cell Biol. 134, 573-583[Abstract/Free Full Text]
27. O'Reilly, D. R., Miller, L. K., and Luckow, V. A. (1994) Baculovirus Expression Vectors: A Laboratory Manual , pp. 124-138, Oxford University Press, New York
28. Poulouin, L., Gallet, O., Rouahi, M., and Imhoff, J. M. (1999) Protein Expression Purif. 17, 146-152[CrossRef][Medline] [Order article via Infotrieve]
29. Wurster, N. B., and Lust, G. (1982) Biochem. Biophys. Res. Commun. 109, 1094-1101[Medline] [Order article via Infotrieve]
30. Saoncella, S., Echtermeyer, F., Denhez, F., Nowlen, J. K., Mosher, D. F., Robinson, S. D., Hynes, R. O., and Goetinck, P. F. (1999) Proc. Natl. Acad. Sci. U. S. A. 96, 2805-2810[Abstract/Free Full Text]
31. Engavall, E., and Ruoslathi, E. (1977) Int. J. Cancer 20, 1-5[Medline] [Order article via Infotrieve]
32. Freedman, R. B. (1989) Cell 57, 1069-1072[CrossRef][Medline] [Order article via Infotrieve]
33. Creighton, T. E., Bagley, C. J., Cooper, L., Darby, N. J., Freedman, R. B., Kemmink, J., and Sheikh, A. (1993) J. Mol. Biol. 232, 1176-1196[CrossRef][Medline] [Order article via Infotrieve]
34. Darby, N. J., Freedman, R. B., and Creighton, T. E. (1994) Biochemistry 33, 7937-7947[CrossRef][Medline] [Order article via Infotrieve]
35. Creighton, T. E., Zapun, A., and Darby, N. J. (1995) Trends Biotechnol. 13, 18-22[CrossRef][Medline] [Order article via Infotrieve]
36. Langenbach, K. J., and Sottile, J. (1999) J. Biol. Chem. 274, 7032-7038[Abstract/Free Full Text]
37. Zapun, A., and Creighton, T. E. (1994) Biochemistry 33, 5202-5211[CrossRef][Medline] [Order article via Infotrieve]
38. Zapun, A., Missiakas, D., Raina, S., and Creighton, T. E. (1995) Biochemistry 34, 5075-5089[CrossRef][Medline] [Order article via Infotrieve]
39. Darby, N. J., and Creighton, T. E. (1995) Biochemistry 34, 3576-3587[CrossRef][Medline] [Order article via Infotrieve]
40. Darby, N. J., Penka, E., and Vincentelli, R. (1998) J. Mol. Biol. 276, 239-247[CrossRef][Medline] [Order article via Infotrieve]
41. Sottile, J., and Mosher, D. F. (1993) Biochemistry 32, 1641-1647[CrossRef][Medline] [Order article via Infotrieve]
42. Williams, M. J., Phan, I., Harvey, T. S., Rostagno, A., Gold, L. I., and Campbell, I. D. (1994) J. Mol. Biol. 235, 1302-1311[CrossRef][Medline] [Order article via Infotrieve]
43. Akiyama, S. K., Yamada, S. S., Chen, W. T., and Yamada, K. M. (1989) J. Cell Biol. 109, 863-875[Abstract/Free Full Text]
44. Fogerty, F. J., Akiyama, S. K., Yamada, K. M., and Mosher, D. F. (1990) J. Cell Biol. 111, 699-708[Abstract/Free Full Text]
45. Bultmann, H., Santas, A. J., and Peters, D. M. (1998) J. Biol. Chem. 273, 2601-2609[Abstract/Free Full Text]
46. Giancotti, F. G., and Ruoslahti, E. (1990) Cell 60, 849-859[CrossRef][Medline] [Order article via Infotrieve]
47. Wu, C., Bauer, J. S., Juliano, R. L., and McDonald, J. A. (1993) J. Biol. Chem. 268, 21883-21888[Abstract/Free Full Text]
48. Sottile, J., Schwarzbauer, J. E., Selegue, J., and Mosher, D. F. (1991) J. Biol. Chem. 266, 12840-12843[Abstract/Free Full Text]
49. Zheng, M., and Hakomori, S. (2000) Arch. Biochem. Biophys. 374, 93-99[CrossRef][Medline] [Order article via Infotrieve]
50. Erickson, H. P., and Carrell, N. A. (1983) J. Biol. Chem. 258, 14539-14544[Abstract/Free Full Text]
51. Bushuev, V. N., Metsis, M. L., Morozkin, A. D., Ruuge, E. K., Sepetov, N. F., and Koteliansky, V. E. (1985) FEBS Lett. 189, 276-280[CrossRef][Medline] [Order article via Infotrieve]
52. Rocco, M., Infusini, E., Daga, M. G., Gogioso, L., and Cuniberti, C. (1987) EMBO J. 6, 2343-2349[Medline] [Order article via Infotrieve]
53. Hynes, R. O. (1999) Proc. Natl. Acad. Sci. U. S. A. 96, 2588-2590[Free Full Text]
54. Wu, C., Keivens, V. M., O'Toole, T. E., McDonald, J. A., and Ginsberg, M. H. (1995) Cell 83, 715-724[CrossRef][Medline] [Order article via Infotrieve]
55. Ugarova, T. P., Zamarron, C., Veklich, Y., Bowditsch, R. D., Ginsberg, M. H., Weisel, J. W., and Plow, E. F. (1995) Biochemistry 34, 4457-4466[CrossRef][Medline] [Order article via Infotrieve]
56. Halliday, N. L., and Tomasek, J. J. (1995) Exp. Cell Res. 217, 109-117[CrossRef][Medline] [Order article via Infotrieve]
57. Zhong, C., Chrzanowska-Wodnicka, M., Brown, J., Shaub, A., Belkin, A. M., and Burridge, K. (1998) J. Cell Biol. 141, 539-551[Abstract/Free Full Text]
58. Homandberg, G. A., and Erickson, J. W. (1986) Biochemistry 25, 6917-6925[CrossRef][Medline] [Order article via Infotrieve]
59. Homandberg, G. A. (1987) Biopolymers 26, 2087-2098[CrossRef][Medline] [Order article via Infotrieve]
60. Woods, A., Johansson, S., and Hook, M. (1988) Exp. Cell Res. 177, 272-283[CrossRef][Medline] [Order article via Infotrieve]
61. Gendelman, R., Burton-Wurster, N., MacLeod, J. N., and Lust, G. (2001) Proceedings of the 47th Annual Meeting, San Francisco, February 25-28, 2001, Orthopaedic Research Society , Vol. 26 , p. 385, Orthopaedic Research Society

Copyright © 2002 by The American Society for Biochemistry and Molecular Biology, Inc.
Add to CiteULike CiteULike   Add to Complore Complore   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us   Add to Digg Digg   Add to Reddit Reddit   Add to Technorati Technorati    What's this?


This article has been cited by other articles:


Home page
 J. Biol. Chem.Home page
T. Kozaki, Y. Matsui, J. Gu, R. Nishiuchi, N. Sugiura, K. Kimata, K. Ozono, H. Yoshikawa, and K. Sekiguchi
Recombinant Expression and Characterization of a Novel Fibronectin Isoform Expressed in Cartilaginous Tissues
J. Biol. Chem., December 12, 2003; 278(50): 50546 - 50553.
[Abstract] [Full Text] [PDF]

Home page
 J. Biol. Chem.Home page
R. Gendelman, N. I. Burton-Wurster, J. N. MacLeod, and G. Lust
The Cartilage-specific Fibronectin Isoform Has a High Affinity Binding Site for the Small Proteoglycan Decorin
J. Biol. Chem., March 21, 2003; 278(13): 11175 - 11181.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow All Versions of this Article:
277/22/20095    most recent
M201238200v1