Understanding the elasticity of fibronectin fibrils: unfolding strengths of FN-III and GFP domains measured by single molecule force spectroscopy.

While it is well established that fibronectin (FN) matrix fibrils are elastic, the mechanism of fibril elasticity during extension is still debated. To investigate the molecular origin of FN fibril elasticity, we used single molecule force spectroscopy (SMFS) to determine the unfolding behavior of a recombinant FN-III protein construct that contained eight FN-III domains ((1-8)FN-III) and two green fluorescent protein (GFP) domains. FN-III domains were distinguished from GFP domains by their shorter unfolding lengths. The unfolding strengths of both domains were determined for a wide range of pulling rates (50 to 1,745 nm/s). We found that the mechanical stabilities of FN-III and GFP domains were very similar to each other over the entire range of pulling speeds. FN fibrils containing GFP remain brightly fluorescent, even when stretched, meaning that GFP domains remain largely folded. Since GFP and FN-III have equal unfolding strengths, this suggests that FN-III domains are not extensively unraveled in stretched FN fibrils. Our results thus favor an alternative model, which invokes a conformational change from a compact to an extended conformation, as the basis for FN fibril elasticity.     

SLLISWD Sequence in the 10FNIII Domain Initiates Fibronectin Fibrillogenesis

Fibronectin (FN) assembly into extracellular matrix is tightly regulated and essential to embryogenesis and wound healing. FN fibrillogenesis is initiated by cytoskeleton-derived tensional forces transmitted across transmembrane integrins onto RGD binding sequences within the tenth FN type III (10FNIII) domains. These forces unfold 10FNIII to expose cryptic FN assembly sites; however, a specific sequence has not been identified in 10FNIII. Our past steered molecular dynamics simulations modeling 10FNIII unfolding by force at its RGD loop predicted a mechanical intermediate with a solvent-exposed N terminus spanning the A and B β-strands. Here, we experimentally confirm that the predicted 23-residue cryptic peptide 1 (CP1) initiates FN multimerization, which is mediated by interactions with 10FNIII that expose hydrophobic surfaces that support 8-anilino-1-napthalenesulfonic acid binding. Localization of multimerization activity to the C terminus led to the discovery of a minimal 7-amino acid “multimerization sequence” (SLLISWD), which induces polymerization of FN and the clotting protein fibrinogen in addition to enhancing FN fibrillogenesis in fibroblasts. A point mutation at Trp-6 that reduces exposure of hydrophobic sites for 8-anilino-1-napthalenesulfonic acid binding and β-structure formation inhibits FN multimerization and prevents physiological cell-based FN assembly in culture. We propose a model for cell-mediated fibrillogenesis whereby cell traction force initiates a cascade of intermolecular exchange starting with the unfolding of 10FNIII to expose the multimerization sequence, which interacts with strand B of another 10FNIII domain via a Trp-mediated β-strand exchange to stabilize a partially unfolded intermediate that propagates FN self-assembly.     

Identification of novel and distinct binding sites within tenascin-C for soluble and fibrillar fibronectin

Interactions between fibronectin and tenascin-C within the extracellular matrix provide specific environmental cues that dictate tissue structure and cell function. The major binding site for fibronectin lies within the fibronectin type III-like repeats (TNfn) of tenascin-C. Here, we systematically screened TNfn domains for their ability to bind to both soluble and fibrillar fibronectin. All TNfn domains containing the TNfn3 module interact with soluble fibronectin. However, TNfn domains bind differentially to fibrillar fibronectin. This distinct binding pattern is dictated by the fibrillar conformation of FN. TNfn1-3, but not TNfn3-5, binds to immature fibronectin fibrils, and additional TNfn domains are required for binding to mature fibrils. Multiple binding sites for distinct regions of fibronectin exist within tenascin-C. TNfn domains comprise a binding site for the N-terminal 70-kDa domain of fibronectin that is freely available and a binding site for the central binding domain of fibronectin that is cryptic in full-length tenascin-C. The 70-kDa and central binding domain regions are key for fibronectin matrix assembly; accordingly, binding of several TNfn domains to these regions inhibits fibronectin fibrillogenesis. These data highlight the complexity of protein-protein binding, the importance of protein conformation on these interactions, and the implications for the physiological assembly of complex three-dimensional matrices.     

Domain unfolding plays a role in superfibronectin formation

Superfibronectin (sFN) is a fibronectin (FN) aggregate that is formed by mixing FN with anastellin, a fragment of the first type III domain of FN. However, the mechanism of this aggregation has not been clear. In this study, we found that anastellin co-precipitated with FN in a ratio of approximately 4:1, anastellin:FN monomer. The primary binding site for anastellin was in the segment (III)1-3, which bound three molecules of anastellin and was able to form a precipitate without the rest of the FN molecule. Anastellin binding to (III)3 caused a conformational change in that domain that exposed a cryptic thermolysin-sensitive site. An additional anastellin binds to (III)11, where it enhances thermolysin digestion of (III)11. An engineered disulfide bond in (III)3 inhibited both aggregation and protease digestion, suggesting that the stability of (III)3 is a key factor in sFN formation. We propose a three-step model for sFN formation: 1) FN-III domains spontaneously unfold and refold; 2) anastellin binds to an unfolded domain, preventing its refolding and leaving it with exposed hydrophobic surfaces and beta-sheet edges; and 3) these exposed elements bind to similar exposed elements on other molecules, leading to aggregation. The model is consistent with our observation that the kinetics of aggregation are first order, with a reaction time of 500-700 s. Similar mechanisms may contribute to the assembly of the native FN matrix.     

Ligation of the fibrin-binding domain by β-strand addition is sufficient for expansion of soluble fibronectin

How fibronectin (FN) converts from a compact plasma protein to a fibrillar component of extracellular matrix is not understood. "Functional upstream domain" (FUD), a polypeptide based on F1 adhesin of Streptococcus pyogenes, binds by anti-parallel β-strand addition to discontinuous sets of N-terminal FN type I modules, (2-5)FNI of the fibrin-binding domain and (8-9)FNI of the gelatin-binding domain. Such binding blocks assembly of FN. To learn whether ligation of (2-5)FNI, (8-9)FNI, or the two sets in combination is important for inhibition, we tested "high affinity downstream domain" (HADD), which binds by β-strand addition to the continuous set of FNI modules, (1-5)FNI, comprising the fibrin-binding domain. HADD and FUD were similarly active in blocking fibronectin assembly. Binding of HADD or FUD to soluble plasma FN exposed the epitope to monoclonal antibody mAbIII-10 in the tenth FN type III module ((10)FNIII) and caused expansion of FN as assessed by dynamic light scattering. Soluble N-terminal constructs truncated after (9)FNI or (3)FNIII competed better than soluble FN for binding of FUD or HADD to adsorbed FN, indicating that interactions involving type III modules more C-terminal than (3)FNIII limit β-strand addition to (1-5)FNI within intact soluble FN. Preincubation of FN with mAbIII-10 or heparin modestly increased binding to HADD or FUD. Thus, ligation of FNIII modules involved in binding of integrins and glycosaminoglycans, (10)FNIII and (12-14)FNIII, increases accessibility of (1-5)FNI. Allosteric loss of constraining interactions among (1-5)FNI, (10)FNIII, and (12-14)FNIII likely enables assembly of FN into extracellular fibrils.     

Fibronectin aggregation and assembly: the unfolding of the second fibronectin type III domain

The mechanism of fibronectin (FN) assembly and the self-association sites are still unclear and contradictory, although the N-terminal 70-kDa region ((I)1-9) is commonly accepted as one of the assembly sites. We previously found that (I)1-9 binds to superfibronectin, which is an artificial FN aggregate induced by anastellin. In the present study, we found that (I)1-9 bound to the aggregate formed by anastellin and a small FN fragment, (III)1-2. An engineered disulfide bond in (III)2, which stabilizes folding, inhibited aggregation, but a disulfide bond in (III)1 did not. A gelatin precipitation assay showed that (I)1-9 did not interact with anastellin, (III)1, (III)2, (III)1-2, or several (III)1-2 mutants including (III)1-2KADA. (In contrast to previous studies, we found that the (III)1-2KADA mutant was identical in conformation to wild-type (III)1-2.) Because (I)1-9 only bound to the aggregate and the unfolding of (III)2 played a role in aggregation, we generated a (III)2 domain that was destabilized by deletion of the G strand. This mutant bound (I)1-9 as shown by the gelatin precipitation assay and fluorescence resonance energy transfer analysis, and it inhibited FN matrix assembly when added to cell culture. Next, we introduced disulfide mutations into full-length FN. Three disulfide locks in (III)2, (III)3, and (III)11 were required to dramatically reduce anastellin-induced aggregation. When we tested the disulfide mutants in cell culture, only the disulfide bond in (III)2 reduced the FN matrix. These results suggest that the unfolding of (III)2 is one of the key factors for FN aggregation and assembly.     

Altered dynamics in Lck SH3 upon binding to the LBD1 domain of Herpesvirus saimiri Tip

The Tip protein from Herpesvirus saimiri interacts with the SH3 domain from the Src-family kinase Lck via a proline-containing sequence termed LBD1. Src-family kinase SH3 domains related to Lck have been shown to be dynamic in solution and partially unfold under physiological conditions. The rate of such partial unfolding is reduced by viral protein binding. To determine if the Lck SH3 domain displayed similar behavior, the domain was investigated with hydrogen exchange and mass spectrometry. Lck SH3 was found to be highly dynamic in solution. While other SH3 domains require as much as 10,000 sec to become totally deuterated, Lck SH3 became almost completely labeled within 200 sec. A partial unfolding event involving 8-10 residues was observed with a half-life of approximately 10 sec. Tip LBD1 binding did not cause gross structural changes in Lck SH3 but globally stabilized the domain and reduced the rate of partial unfolding by a factor of five. The region of partial unfolding in Lck SH3 was found to be similar to that identified for other SH3 domains that partially unfold. Although the sequence conservation between Lck SH3 and other closely related SH3 domains is high, the dynamics do not appear to be conserved.     

Folding-unfolding of FN-III domains in tenascin: an elastically coupled two-state system

In a single-molecule atomic force microscopy (AFM) experiment, the tenascin molecule is stretched by an external force causing an elongation which is due to the unfolding of the FN-III modules. The features of the force-extension curves depend on the pulling speed and show a saw-tooth pattern (lower speeds) or a smooth pattern (higher speeds). In any case, the unfolded domains are elastically coupled to the unfolded modules, acting as transmitters of the external force. In this communication, the folding-unfolding process of the FN-III domains in tenascin is studied using reaction rate theory and a simple two-state model. The main hypothesis of the study is that, at microscopic level, the force needed to unfold a domain and the unfolding rate (unfolding velocity) can mimic the macroscopic process of measurement by AFM. As the external force is applied, the probability of unfolding increases as dictated by the reaction rate theory. Within this context, a relationship between the unfolding force and the unfolding velocity is obtained. The latter relation will describe microscopically the process in a phenomenological fashion. Moreover, while relating the results of this study with other experimental (AFM measurements) and theoretical (Monte Carlo simulations) data, we found that the graph of unfolding force-unfolding velocity is similar to that of external force-pulling velocity. The refolding process can also be studied within this model and the results show similar trends. The latter suggests a generic and universal behavior of such kind of molecular domains at least in the light of the proposed model.     

The mechanical hierarchies of fibronectin observed with single-molecule AFM

Mechanically induced conformational changes in proteins such as fibronectin are thought to regulate the assembly of the extracellular matrix and underlie its elasticity and extensibility. Fibronectin contains a region of tandem repeats of up to 15 type III domains that play critical roles in cell binding and self-assembly. Here, we use single-molecule force spectroscopy to examine the mechanical properties of fibronectin (FN) and its individual FNIII domains. We found that fibronectin is highly extensible due to the unfolding of its FNIII domains. We found that the native FNIII region displays strong mechanical unfolding hierarchies requiring 80 pN of force to unfold the weakest domain and 200 pN for the most stable domain. In an effort to determine the identity of the weakest/strongest domain, we engineered polyproteins composed of an individual domain and measured their mechanical stability by single-protein atomic force microscopy (AFM) techniques. In contrast to chemical and thermal measurements of stability, we found that the tenth FNIII domain is mechanically the weakest and that the first and second FNIII domains are the strongest. Moreover, we found that the first FNIII domain can acquire multiple, partially folded conformations, and that their incidence is modulated strongly by its neighbor FNIII domain. The mechanical hierarchies of fibronectin demonstrated here may be important for the activation of fibrillogenesis and matrix assembly.     

The RGD motif in fibronectin is essential for development but dispensable for fibril assembly

Fibronectin (FN) is secreted as a disulfide-bonded FN dimer. Each subunit contains three types of repeating modules: FN-I, FN-II, and FN-III. The interactions of alpha5beta1 or alphav integrins with the RGD motif of FN-III repeat 10 (FN-III10) are considered an essential step in the assembly of FN fibrils. To test this hypothesis in vivo, we replaced the RGD motif with the inactive RGE in mice. FN-RGE homozygous embryos die at embryonic day 10 with shortened posterior trunk, absent tail bud-derived somites, and severe vascular defects resembling the phenotype of alpha5 integrin-deficient mice. Surprisingly, the absence of a functional RGD motif in FN did not compromise assembly of an FN matrix in mutant embryos or on mutant cells. Matrix assembly assays and solid-phase binding assays reveal that alphavbeta3 integrin assembles FN-RGE by binding an isoDGR motif in FN-I5, which is generated by the nonenzymatic rearrangement of asparagines (N) into an iso-aspartate (iso-D). Our findings demonstrate that FN contains a novel motif for integrin binding and fibril formation whose activity is controlled by amino acid modification.     

The compact conformation of fibronectin is determined by intramolecular ionic interactions.

Fibronectin exists in a compact or extended conformation, depending upon environmental pH and salt concentration. Using recombinant fragments expressed in bacteria and baculovirus, we determined the domains responsible for producing fibronectin's compact conformation. Our velocity and equilibrium sedimentation data show that FN2-14 (a protein containing FN-III domains 2 through 14) forms dimers in low salt. Experiments with smaller fragments indicates that the compact conformation is produced by binding of FN12-14 of one subunit to FN2-3 of the other subunit in the dimer. The binding is weakened at higher salt concentrations, implying an electrostatic interaction. Furthermore, segment FN7-14+A, which contains the alternatively spliced A domain between FN11 and 12, forms dimers, whereas FN7-14 without A does not. Segment FN12-14+A also forms dimers, but the isolated A domain does not. These data imply an association of domain A with FN12-14, and the presence of A may favor an open conformation by competing with FN2-3 for binding to FN12-14.     

Immuno-electron-microscopic study of fibronectin in gastrulating amphibian embryos

Summary The ultrastructural organization of fibronectin (FN) in early amphibian embryos (Ambystoma mexicanum, Pleurodeles waltlii) was studied with the use of antibodies directed against amphibian plasmatic FN. Scanning and transmission electron microscopy combined with immunogold labeling of FN revealed that the extracellular matrix that covers the inner surface of the ectodermal layer consists of FN-containing fibrils. During gastrulation, the mesodermal cells appear to be devoid of FN. These cells extend filopodia adhering to the FN-containing fibrils and are spreading along them. These findings suggest that FN may be involved in contact formation between mesodermal cells and the extracellular matrix that serves as a substratum for migration.     

Altered rate of fibronectin matrix assembly by deletion of the first type III repeats

The assembly of fibronectin (FN) into a fibrillar matrix is a complex stepwise process that involves binding to integrin receptors as well as interactions between FN molecules. To follow the progression of matrix formation and determine the stages during which specific domains function, we have developed cell lines that lack an endogenous FN matrix but will form fibrils when provided with exogenous FN. Recombinant FNs (recFN) containing deletions of either the RGD cell- binding sequence (RGD-) or the first type III repeats (FN delta III1-7) including the III1 FN binding site were generated with the baculovirus insect cell expression system. After addition to cells, recFN matrix assembly was monitored by indirect immunofluorescence and by insolubility in the detergent deoxycholate (DOC). In the absence of any native FN, FN delta III1-7 was assembled into fibrils and was converted into DOC-insoluble matrix. This process could be inhibited by the amino- terminal 70 kD fragment of FN, showing that FN delta III1-7 follows an assembly pathway similar to FN. The progression of FN delta III1-7 assembly differed from native FN in that the recFN became DOC-insoluble more quickly. In contrast, RGD- recFNs were not formed into fibrils except when added in combination with native FN. These results show that the RGD sequence is essential for the initiation step but fibrils can form independently of the III1-7 modules. The altered rate of FN delta III1-7 assembly suggests that one function of the missing repeats might be to modulate an early stage of matrix formation.     

Sequential unfolding of the two-domain protein Pseudomonas stutzeri cytochrome c(4)

P. stutzeri cytochrome c(4) is a di-haem protein, composed of two globular domains each with His-Met coordinated haem, and a hydrogen bond network between the domains. The domain foldings are highly symmetric but with specific differences including structural differences of ligand coordination, and different spin states of the oxidised haem groups. We have studied unfolding of oxidised P. stutzeri cyt c(4) induced thermally and by chemical denaturants. Horse heart cyt c was a reference molecule. Isothermal unfolding induced by guanidinium chloride and acid was followed by Soret, alpha/beta, and 701-nm band absorption, and by far-UV circular dichroism spectroscopy. Multifarious patterns emerge, but the two domains clearly unfold sequentially. One phase, assigned to unfolding of the N-terminal domain, proceeds at guanidinium concentrations up to approximately 1.0 M. This is followed by two overlapping phases at higher concentrations. The intermediate state maintains Fe-Met coordination, assigned to the C-terminal domain. Interdomain interaction is reflected in decreasing values of the cooperativity parameters. Differential scanning calorimetry shows a single peak, but two peaks appear when guanidinium chloride up to 0.4 M is present. This reflects different chemical action in chemical and thermal unfolding. Acid-induced unfolding kinetics was addressed by pH jumps using diode array stopped-flow techniques. Three kinetic phases in the 701 nm Fe-Met marker band, and four phases in the Soret and alpha/beta bands, spanning 4-1000 ms could be distinguished on pH jumps from 7.5 to the range 2.5-3.5. In this range of time and pH cyt c appears to unfold in no more than two phases. Spectral properties of the kinetic intermediates could be identified. Sequential domain unfolding, formation of high-spin states, and an intermediate state with Fe-Met coordination to a single haem group are features of the unfolding kinetics.     

Human phenotypically distinct TGFBI corneal dystrophies are linked to the stability of the fourth FAS1 domain of TGFBIp

Mutations in the human TGFBI gene encoding TGFBIp have been linked to protein deposits in the cornea leading to visual impairment. The protein consists of an N-terminal Cys-rich EMI domain and four consecutive fasciclin 1 (FAS1) domains. We have compared the stabilities of wild-type (WT) human TGFBIp and six mutants known to produce phenotypically distinct deposits in the cornea. Amino acid substitutions in the first FAS1 (FAS1-1) domain (R124H, R124L, and R124C) did not alter the stability. However, substitutions within the fourth FAS1 (FAS1-4) domain (A546T, R555Q, and R555W) affected the overall stability of intact TGFBIp revealing the following stability ranking R555W>WT>R555Q>A546T. Significantly, the stability ranking of the isolated FAS1-4 domains mirrored the behavior of the intact protein. In addition, it was linked to the aggregation propensity as the least stable mutant (A546T) forms amyloid fibrils while the more stable variants generate non-amyloid amorphous deposits in vivo. Significantly, the data suggested that both an increase and a decrease in the stability of FAS1-4 may unleash a disease mechanism. In contrast, amino acid substitutions in FAS1-1 did not affect the stability of the intact TGFBIp suggesting that molecular the mechanism of disease differs depending on the FAS1 domain carrying the mutation.     

Implications for Collagen Binding from the Crystallographic Structure of Fibronectin 6FnI1�C2FnII7FnI

Collagen and fibronectin (FN) are two abundant and essential components of the vertebrate extracellular matrix; they interact directly with cellular receptors and affect cell adhesion and migration. Past studies identified a FN fragment comprising six modules, ⁶FnI^1–2FnII^7–9FnI, and termed the gelatin binding domain (GBD) as responsible for collagen interaction. Recently, we showed that the GBD binds tightly to a specific site within type I collagen and determined the structure of domains ^8–9FnI in complex with a peptide from that site. Here, we present the crystallographic structure of domains ⁶FnI^1–2FnII⁷FnI, which form a compact, globular unit through interdomain interactions. Analysis of NMR titrations with single-stranded collagen peptides reveals a dominant collagen interaction surface on domains ²FnII and ⁷FnI; a similar surface appears involved in interactions with triple-helical peptides. Models of the complete GBD, based on the new structure and the ^8–9FnI·collagen complex show a continuous putative collagen binding surface. We explore the implications of this model using long collagen peptides and discuss our findings in the context of FN interactions with collagen fibrils.     

The Cysteine-rich Region of Type VII Collagen Is a Cystine Knot with a New Topology

Collagens are a group of extracellular matrix proteins with essential functions for skin integrity. Anchoring fibrils are made of type VII collagen (Col7) and link different skin layers together: the basal lamina and the underlying connective tissue. Col7 has a central collagenous domain and two noncollagenous domains located at the N and C terminus (NC1 and NC2), respectively. A cysteine-rich region of hitherto unknown function is located at the transition of the NC1 domain to the collagenous domain. A synthetic model peptide of this region was investigated by CD and NMR spectroscopy. The peptide folds into a collagen triple helix, and the cysteine residues form disulfide bridges between the different strands. The eight cystine knot topologies that are characterized by exclusively intermolecular disulfide bridges have been analyzed by molecular modeling. Two cystine knots are energetically preferred; however, all eight disulfide bridge arrangements are essentially possible. This novel cystine knot is present in type IX collagen, too. The conserved motif of the cystine knot is CX₃CP. The cystine knot is N-terminal to the collagen triple helix in both collagens and therefore probably impedes unfolding of the collagen triple helix from the N terminus.     

ECM Protein Nanofibers and Nanostructures Engineered Using Surface-initiated Assembly

The extracellular matrix (ECM) in tissues is synthesized and assembled by cells to form a 3D fibrillar, protein network with tightly regulated fiber diameter, composition and organization. In addition to providing structural support, the physical and chemical properties of the ECM play an important role in multiple cellular processes including adhesion, differentiation, and apoptosis. In vivo, the ECM is assembled by exposing cryptic self-assembly (fibrillogenesis) sites within proteins. This process varies for different proteins, but fibronectin (FN) fibrillogenesis is well-characterized and serves as a model system for cell-mediated ECM assembly. Specifically, cells use integrin receptors on the cell membrane to bind FN dimers and actomyosin-generated contractile forces to unfold and expose binding sites for assembly into insoluble fibers. This receptor-mediated process enables cells to assemble and organize the ECM from the cellular to tissue scales. Here, we present a method termed surface-initiated assembly (SIA), which recapitulates cell-mediated matrix assembly using protein-surface interactions to unfold ECM proteins and assemble them into insoluble fibers. First, ECM proteins are adsorbed onto a hydrophobic polydimethylsiloxane (PDMS) surface where they partially denature (unfold) and expose cryptic binding domains. The unfolded proteins are then transferred in well-defined micro- and nanopatterns through microcontact printing onto a thermally responsive poly(N-isopropylacrylamide) (PIPAAm) surface. Thermally-triggered dissolution of the PIPAAm leads to final assembly and release of insoluble ECM protein nanofibers and nanostructures with well-defined geometries. Complex architectures are possible by engineering defined patterns on the PDMS stamps used for microcontact printing. In addition to FN, the SIA process can be used with laminin, fibrinogen and collagens type I and IV to create multi-component ECM nanostructures. Thus, SIA can be used to engineer ECM protein-based materials with precise control over the protein composition, fiber geometry and scaffold architecture in order to recapitulate the structure and composition of the ECM in vivo.     

A novel fibronectin binding site required for fibronectin fibril growth during matrix assembly

Fibronectin (FN) assembly into a fibrillar extracellular matrix is a stepwise process requiring participation from multiple FN domains. Fibril formation is regulated in part by segments within the first seven type III repeats (III1-7). To define the specific function(s) of this region, recombinant FNs (recFNs) containing an overlapping set of deletions were tested for the ability to assemble into fibrils. Surprisingly, recFN lacking type III repeat III1 (FNDeltaIII1), which contains a cryptic FN binding site and has been suggested to be essential for fibril assembly, formed a matrix identical in all respects to a native FN matrix. Similarly, displacement of the cell binding domain in repeats III9-10 to a position close to the NH2-terminal assembly domain, as well as a large deletion spanning repeats III4-7, had no effect on assembly. In contrast, two deletions that included repeat III2, DeltaIII1-2 and DeltaIII2-5, caused significant reductions in fibril elongation, although binding of FN to the cell surface and initiation of assembly still proceeded. Using individual repeats in binding assays, we show that III2 but not III1 contains an FN binding site. Thus, these results pinpoint repeat III2 as an important module for FN-FN interactions during fibril growth.