Differential binding of factor XII and activated factor XII to soluble and immobilized fibronectin – localization of the Hep-1/Fib-1 binding site for activated factor XII Inger Schousboe1, Birthe T Nystrøm1 and Gert H Hansen2 Department of Biomedical Sciences, The Panum Institute, University of Copenhagen, Denmark Department of Cellular and Molecular Medicine, The Panum Institute, University of Copenhagen, Denmark Keywords association; factor XII; factor XIIa; fibronectin Correspondence I Schousboe, Department of Biomedical Sciences, Heart and Circulatory Research Section, The Panum Institute, University of Copenhagen, Blegdamsvej 3C, DK-2200 Copenhagen, Denmark Fax: +45 35367980 Tel: +45 35327800 E-mail: schousboe@imbg.ku.dk (Received May 2008, revised July 2008, accepted 18 August 2008) doi:10.1111/j.1742-4658.2008.06647.x Fibronectins (FNs) are dimeric glycoproteins that adopt a globular conformation when present in plasma and solution and an extended conformation in the extracellular matrix Factor XII (FXII) is a zymogen of the proteolytically active FXIIa that plays a role in thrombus stabilization by enhancing clot formation and in inflammation by enhancing bradykinin formation To investigate whether the extracellular matrix could play a role in these events, we have recently shown that FXIIa, but not FXII, binds to the extracellular matrix (ECM), and suggested that FN may be the target for the binding Immunofluorescence microscopy has in the present investigation confirmed that FXIIa added to the ECM colocalizes with FN deposited during growth of human umbilical vein endothelial cells The aim of the present study, therefore, was to further elucidate the interaction between FXIIa and FN by the use of a solid face binding assay This showed, like the binding to the ECM, that FXIIa, but not FXII, binds in a Zn2+-independent manner to immobilized FN The KD for the binding was 8.5 ± 0.9 nm (n = 3) The binding was specific for the immobilized FN, as the binding could not be inhibited by soluble FN Furthermore, soluble FN did not bind to immobilized FXIIa However, soluble FN could bind to FXII, and this binding inhibited the surface-induced autoactivation of FXII and subsequent binding of the generated FXIIa to immobilized FN The presence of FXII in an anti-FN immunoprecipitate of plasma indicated that some FXII in plasma circulates bound to FN The binding of FXIIa to FN was inhibited by gelatine and fibrin but not by heparin, indicating that FXIIa binds to immobilized FN through the type I repeat modules Accordingly, FXIIa was found to bind to immobilized fragments of FN containing the type I repeat modules in the N-terminal domain to which fibrin and gelatine bind Several studies have suggested that in the cardiovascular system, the interaction between the vessel wall and the contact activation system of blood coagulation, including factor XII (FXII), high molecular mass kininogen (HK) and prekallikrein, involves Zn2+dependent and receptor-mediated binding of FXII and HK Thus, investigations of FXII and HK binding to endothelial cells in the vascular wall mimicked by Abbreviations CTI, corn trypsin inhibitor; DS, dextran sulfate; ECM, extracellular matrix; Fib-1, the N-terminal fibrinogen binding domain on fibronectin; FN, fibronectin; FXII, factor XII; FXIIa, activated factor XII; Hep-1, the N-terminal heparin binding domain on fibronectin; Hep-2, the C-terminal heparin binding domain on fibronectin; HK, high molecular mass kininogen; HRP, horseradish peroxidase; HUVEC, human umbilical vein endothelial cell; OPD, o-phenylenediamine FEBS Journal 275 (2008) 5161–5172 ª 2008 The Authors Journal compilation ª 2008 FEBS 5161 Factor XII binding to fibronectin I Schousboe et al cultures of human umbilical vein endothelial cells (HUVECs) have shown that FXII and HK interact by multiprotein assembly [1–3] FXII is a precursor of the proteolytically active activated FXII (FXIIa) FXII and FXIIa bind equally well to a confluent layer of HUVECs [4] However, a recent investigation has shown that the binding might have been artefactual, and that FXII in the presence but not in the absence of a negatively charged surface bound rather to the extracellular matrix (ECM) generated during growth of HUVECs The presence of negatively charged surfaces appeared to serve two purposes: (a) it induced and enhanced the autoactivation of FXII, generating FXIIa; and (b) it abrogated nonspecific binding of FXIIa [5] The binding of FXIIa to the ECM showed several differences from the binding to HUVECs Thus, the binding to the ECM was: (a) specific for FXIIa; (b) Zn2+-independent; (c) not inhibited by HK; and (d) nonelectrostatic As proteolytic degradation of the ECM abrogated the binding of FXIIa, it was assumed that a matrix protein was the target for the binding Therefore, it was tentatively analyzed and found that FXIIa binds to fibronectin (FN) [5] FN is a dimeric high molecular mass glycoprotein that is found both as a circulating soluble molecule in the blood and as insoluble molecules forming elongated multimeric structures in the ECM [6,7] The monomer of the dimeric soluble molecule is a mosaic protein composed of modular subunits generating different domains [8], which harbor binding sites for glycosaminoglycans, collagen or gelatine, fibrin, and integrin receptors Some of these binding sites become available only in the multimeric, elongated forms in which internal sequences of amino acid residues become exposed [9,10] Several factors mediate the transition from the soluble to an elongated form, including adsorption of FN to plastic surfaces [11–14] As our previous studies have shown that the binding of FXIIa to the ECM could be due to binding of FXIIa to FN generated during growth of HUVECs [5], we here report on studies of the interactions of FXIIa with FN using a solid-phase binding assay in which either FN or FXIIa is immobilized Results FXII/FXIIa binding to FN The association of FXIIa with the ECM was assumed to take place through binding to FN Therefore, investigations were first performed to determine whether it could be shown that FXIIa associated with FN depos5162 A B Fig Colocalization of the ECM-bound FXII and FN HUVECs were grown to near confluence, and the generated ECM was exposed by detaching the cells with EDTA After washing, the ECM was incubated for h with 20 nM FXIIa The ECM was then washed again and incubated first with a mixture of goat anti-FXII IgG (1 : 100) and rabbit anti-FN IgG (1 : 100) for h, and second with a mixture of Alexa 594-conjugated donkey anti-(goat IgG) (1 : 800) and Alexa 488-conjugated goat anti-(mouse IgG) (1 : 800) (A) Red indicates the presence of FXIIa (B) Green indicates the presence of FN Bar: 20 lm ited on the surface of the culture dish after depletion of HUVECs by EDTA extraction Immunofluorescence clearly showed that FXIIa bound to the deposited FN (Fig 1) No FN was deposited on and no FXIIa bound to surfaces incubated with growth medium under the same conditions and for the same periods of time as the cells but in the absence of cells To obtain more information about this association, the interaction between FXIIa and FN was subsequently analyzed by measuring the binding of FXIIa to FN immobilized on a plastic surface Using a solid-phase binding assay, the binding of FXIIa to FN was visualized by reactions with an FEBS Journal 275 (2008) 5161–5172 ª 2008 The Authors Journal compilation ª 2008 FEBS I Schousboe et al Factor XII binding to fibronectin antibody against FXII and a horseradish peroxidase (HRP)-labeled secondary antibody Neither the antibody against FXII nor the secondary antibody was observed to bind to FN in the absence of FXIIa This excludes the possibility that the response was nonspecific and due to a direct interaction between the immobilized FN and the immunoglobulins, as previously noted [15] Furthermore, preincubation of FXIIa for h with a two-fold molar excess of the antibody against FXII prior to incubation with FN abolished the binding Surprisingly, the binding could not be inhibited if the immobilized FN had been preincubated with a polyclonal antibody against soluble FN (data not shown) This could be due to lack of recognition of the binding site on the immobilized FN for FXIIa, but it could also be due to a nonspecific interaction between FXIIa and the plastic surface However, very little FXIIa bound to wells devoid of FN (controls) Moreover, nonspecific binding is nonsaturable The binding of FXIIa to immobilized FN was saturable even at low concentrations of FXIIa This was demonstrated by analyzing the binding of varying concentrations of FXIIa At low concentrations of FXIIa, considerably more FXIIa bound to FN than to control Concentration of FXIIa/ absorbance units 3.5 FXIIa bound, absorbance units 2.5 70 wells At high concentrations of FXIIa, the binding to FN increased linearly with the concentration of FXIIa, and in parallel with the binding of FXIIa to control wells After subtraction of nonspecific binding from the total binding, saturated binding to immobilized FN was observed at FXIIa concentrations ‡ 20 nm (Fig 2) Linear transformation of the the binding isotherm (Fig insert) obtained in one of three independent experiments, each performed in triplicate, showed high-affinity binding, the KD of which was estimated to be 8.5 ± 0.9 nm, using all available data To determine whether the binding of FXIIa to immobilized FN was mediated through the N-terminal surface binding sequence in FXIIa, investigations were performed to determine whether the presence of negatively charged compounds such as sulfatides would affect the binding of FXIIa to FN This showed that sulfatides neither inhibited nor enhanced the binding to immobilized FN The apparently higher-affinity binding of FXIIa in the present experiment in the absence than in the presence of sulfatides was due to parallel higher nonspecific binding However, if FXIIa was exchanged with FXII, the presence of sulfatides induced binding of FXII, which in the absence of y = 1.1214x + 9.7607 R = 0.993 60 FN Control FN - Control 50 40 30 20 10 0 10 20 30 40 Concentration of FXIIa, nM 50 60 1.5 0.5 0 10 20 30 40 50 60 Concentration of FXIIa, nM Fig Concentration-dependent binding of FXIIa to FN The microtiter plate was coated overnight with FN (10 lgỈmL)1) and NaCl ⁄ Pi (control), respectively, and subsequently blocked with blocking buffer Then, it was incubated for h with increasing concentrations of FXIIa in blocking buffer The amount of bound FXIIa was determined by sequential incubation with goat anti-FXII IgG and HRP-conjugated rabbit anti(goat IgG) and visualized by reactions with OPD as described in Experimental procedures d, total amount of FXIIa bound to wells coated with FN; s, total amount of FXIIa bound to control wells (devoid of FN but ‘coated’ overnight with NaCl ⁄ Pi; , binding of FXIIa to FN, calculated as the difference between binding of FXIIa to the former and the latter Linear transformation of the results shown in the figure, which is representative of three experiments performed in triplicate, gave a KD of 8.7 nM Results are means ± SD (n = 3), shown by vertical bars when extending beyond the symbols FEBS Journal 275 (2008) 5161–5172 ª 2008 The Authors Journal compilation ª 2008 FEBS 5163 Factor XII binding to fibronectin I Schousboe et al FXIIa bound, absorbance units 2.5 1.5 0.5 FXII – sulfatide FXII + sulfatide FXII + sulfatide + CTI FXIIa – sulfatide FXIIa + sulfatide FXIIa + anti- Block buffer FXII antibody Fig The effect of sulfatide on the binding of FXIIa to immobilized FN The microtiter plate, coated overnight with FN (10 lgỈmL)1) and NaCl ⁄ Pi (control), respectively, was blocked with blocking buffer and incubated for h with FXII (20 nM) and FXIIa (20 nM) in the presence (+sulfatide) and absence ()sulfatide) of sulfatides (20 lgỈmL)1) To ensure that possible sulfatide-dependent binding of FXII could not be explained by autoactivation of FXII, incubation of FXII in the presence of sulfatides was additionally performed in the presence of CTI (10 lgỈmL)1) FN was also incubated for h with FXIIa, which had been preincubated for h with a twofold molar excess of goat anti-FXII IgG At the end of the incubation, the incubation mixtures were removed, and the microtiter plate was washed extensively Then, the microtiter plate was incubated sequentially with goat anti-FXII IgG, and HRP-conjugated rabbit anti-(goat IgG) in 1% skimmed milk, and the amount of bound FXIIa was visualized by reaction with OPD The combination of primary and secondary antibodies did not bind to either FN-coated or control wells in the absence of FXIIa ⁄ FXII + sulfatides, as indicated by the column showing the binding of blocking buffer The total amount of FXIIa bound to FN and control wells is indicated by gray and white, respectively Results are means ± SD (n = 3), shown by vertical bars sulfatides was negligible (Fig 3) The sulfatide-dependent binding of FXII was most likely due to a sulfatide-induced and sulfatide-enhanced autoactivation of FXII [16,17] Accordingly, the presence of corn trypsin inhibitor (CTI), which inhibits the activity of FXIIa, and thus the autoactivation of FXII, almost completely blocked the sulfatide-induced binding of FXII to FN As compared to FXIIa, a small amount of FXII bound to FN Binding of the activated form of FXII was shown by western blots of extracts of immobilized FN incubated with FXII in the presence of sulfatides (Fig 4) As FXII and FXIIa bind equally well to sulfatides [5], the lack of binding to immobilized FN of FXII and the lack of inhibition of FXIIa by sulfatides indicate that the binding is not brought about by the N-terminal surface-binding region in FXIIa To confirm this, it was investigated whether the binding of FXIIa to FN could be inhibited by the nine amino acid peptide YHKCTHKGR(39–47), containing the surfacebinding sequence [18] The presence of this peptide did not inhibit the binding of FXIIa to immobilized FN (data not shown) 5164 In plasma and in solution, FN adopts a compact soluble conformation in which the two subunits of the dimer are thought to be folded upon each other [7] Several studies have reported a change in the FN conformation upon binding to plastic [11–14], exposing a cryptic binding site by transition from the soluble to the immobilized form [19,20] To determine whether these conformational changes were of significance for the binding of FXIIa, subsequent investigations were performed to determine whether the presence of soluble FN could inhibit the binding to immobilized FN This was shown not to be the case The amount of FXIIa that bound to immobilized FN was the same regardless of the presence of soluble FN In contrast, the presence of soluble FN reduced the sulfatideinduced binding of FXII (P < 0.001) (Fig 5) However, as sulfatides had hardly any effect on the binding of FXIIa to immobilized FN, and FXII did not bind to immobilized FN in the absence of sulfatides (Fig 3), the inhibition could be due to an inhibition of the interaction between FXII and sulfatides To investigate this further, the solid-phase binding assay was turned around and the microtiter plate was coated FEBS Journal 275 (2008) 5161–5172 ª 2008 The Authors Journal compilation ª 2008 FEBS I Schousboe et al FXIIa 80 50 FXIIa bound, absorbance units FXII Factor XII binding to fibronectin 1.8 FN Control 1.6 * 1.4 1.2 0.8 ** 0.6 0.4 0.2 lf I– XI I– with FXII and FXIIa instead of FN Then, the binding of soluble FN to immobilized FXII and FXIIa was visualized by incubation with rabbit anti-(soluble FN) IgG as the primary antibody and HRP-conjugated swine anti-(rabbit IgG) as secondary antibody Figure shows that whereas almost no FN could bind to immobilized FXIIa, it could bind to FXII The presence of sulfatides increased only slightly the binding to both FXII and FXIIa Although it seemed most unlikely, these differences in the amount of bound FN could be due to differences in the amount of FXII and FXIIa coated on the plate This was found not to be the case, as the immunochemical response was analyzed and observed to be identical using goat I+ + gl I+ I FX rF a ul IIa ob F f ul lf N su XI s I FX Fig Western blot of extracts of bound protein after incubation of FXII on immobilized FN in the absence and presence of sulfatides The microtiter plate was coated overnight with FN (10 lgỈmL)1) and subsequently blocked with blocking buffer Then, it was incubated for h with 20 nM FXII in blocking buffer in the presence and absence of 20 lgỈmL)1 sulfatide After washing, the proteins bound to immobilized FN were extracted with SDS under reducing conditions (SDS containing dithiothreitol) and subjected to reduced SDS ⁄ PAGE and western blotting FXII, FXIIa and standard samples of molecular mass markers were run simultaneously Antibody-reacting bands were visualized by sequential incubation with goat anti-(human FXII) IgG (1 : 2500), HRP-conjugated rabbit anti(goat IgG) (1 : 2500) and SuperSignal West Femto Maximum Sensitivity Substrate FXII; FXIIa; Lane 1: proteins extract from control wells devoid of FN in which FXII had been incubated in the absence of sulfatides Lane 2: proteins extracted from immobilized FN after incubation with FXII in the absence of sulfatides Lane 3: proteins extracted from control wells in which FXII had been incubated in the presence of sulfatides Lane 4: proteins extracted from immobilized FN after incubation with FXII in the presence of sulfatides The positions of 50 kDa and 80 kDa proteins are indicated to the left The blot shows that only FXIIa binds to FN rF a ul ob F f ul lf N su + gl – FX s Ia XI – f ul su ar FN l bu o + gl s F Fig The effect of soluble FN on the binding of FXII and FXIIa to immobilized FN The microtiter plate was coated overnight with FN (10 lgỈmL)1) and NaCl ⁄ Pi, respectively Then, it was blocked with blocking buffer and incubated for h with FXII (20 nM) or FXIIa (20 nM) in blocking buffer in the absence ()sulf) and presence (+sulf) of sulfatides (20 lgỈmL)1) and in the absence and presence of soluble FN (10 lgỈmL)1), as indicated The amount of FXIIa bound to FN was measured by sequential incubation with goat anti-FXII IgG and HRP-conjugated rabbit anti-(goat IgG) as described in Experimental procedures The amounts of FXIIa bound to FN and control wells are indicated by gray and white, respectively Statistically significant differences in binding of FXIIa to FN coated on the microtiter plate when incubated in the absence and presence of soluble FN are indicated by asterisks (*not significant and **P < 0.001) Results are means ± SD (n = 3), shown by vertical bars anti-FXII IgG Moreover, to ensure that FXII had not been activated during the coating period, the wells were coated in the presence of CTI, which inhibits the activity of FXIIa and thus the conversion of FXII to FXIIa Furthermore, in order to prevent FXII from activation during the incubation with FN, CTI was added to the incubation mixture This did not affect the binding of FN (results not shown) Thus, these results clearly show that soluble FN interacts directly with FXII in the absence of sulfatides To determine whether this interaction also occurs in plasma, the presence of FXII was analyzed in immunoprecipitates of FN Plasma was immunoprecipitated with antibodies against FN and adsorbed to protein G–Sepharose, from which FXII was extracted The plasma was not preabsorbed to protein G–Sepharose, as binding of FXII to the Sepharose could disturb the equilibrium for the binding of FXII to FN Instead, the amount of FXII bound to protein G–Sepharose in the absence of antibodies against FN was simultaneously analyzed (Fig 7) A much greater amount of FXII could be FEBS Journal 275 (2008) 5161–5172 ª 2008 The Authors Journal compilation ª 2008 FEBS 5165 Factor XII binding to fibronectin I Schousboe et al Globular FN bound, absorbance units FXIIa FXII 2.5 1.5 0.5 FN FN + Sulf FN FN + Sulf Fig Binding of soluble FN to immobilized FXII and FXIIa The microtiter plate was coated overnight as indicated with FXII (20 nM) and FXIIa (20 nM), respectively, diluted in NaCl ⁄ Pi Then, the microtiter plate was blocked with blocking buffer and incubated for h with FN (10 lgỈmL)1) in blocking buffer or in blocking buffer containing sulfatides (+sulf; 20 lgỈmL)1) The amounts of FN bound to FXII and FXIIa, respectively, were determined by sequential incubation with rabbit anti-FN IgG, HRP-conjugated swine anti-(rabbit IgG) and OPD, as described in Experimental procedures Results are means ± SD (n = 3), shown by vertical bars Fig Western blots of FXII present in FN immunoprecipitates of plasma FN was isolated from plasma by immunoprecipitation with a rabbit antibody against FN and protein G–Sepharose The presence of FXII in the immunoprecipitate (lane 2) was analyzed by western blotting using goat anti-FXII IgG as primary antibodies and HRP-conjugated rabbit anti-(goat IgG) as secondary antibody To assure that the presence of FXII in the immunoprecipitate was not due to adsorption of FXII to the protein G–Sepharose, the amount of adsorbed FXII in the absence of the antibody against FN was analyzed simultaneously (lane 1) extracted from FN immunoprecipitates of plasma than from the plasma alone This indicates that FXII also forms a complex with FN in plasma Further characterization of FXIIa binding to immobilized FN The high-affinity interaction between FXIIa and immobilized FN and the lack of interference by soluble FN indicated that the binding site on FN for FXIIa may be buried in the compact soluble form of FN [7] FN has binding sites for a series of ligands 5166 such as glycosaminoglycans, collagens or gelatine, fibrin and integrins [21–27] Figure shows a sketch of FN and the localization of the different binding sites used in our attempt to identify the binding site for FXIIa Thus, concentration-dependent inhibition of FXIIa binding to immobilized FN was observed with gelatine and high concentrations of dextran sulfate (DS) but not with heparin (Fig 9) As shown in Fig 8, FN has two binding sites for heparin The Hep-1-binding site is a low-affinity binding site, and the Hep-2-binding site is a high-affinity binding site [21–23] If FXIIa bound to the C-terminal high-affinity Hep-2-binding site, it would have been expected that its interaction with immobilized FN would be inhibited by heparin Thus, the lack of inhibition by heparin indicated that FXIIa did not bind to the C-terminal high-affinity heparin-binding domain in FN (Hep-2) However, the inhibition by high concentrations of DS and gelatine may indicate that FXIIa binds to the N-terminal region of FN, including the low-affinity Hep-1-binding domain DS is a heparinlike molecule and may, as such, be assumed to bind to the heparin-binding sites on FN To investigate this further, the binding of FXIIa to commercially available proteolytic fragments of FN was analyzed Each of these fragments contains binding domains for heparin, gelatine and cells, respectively Surprisingly, the binding of FXIIa to these fragments showed that although heparin was unable to inhibit the binding of FXIIa to intact FN, FXIIa bound primarily to the 30 kDa low-affinity heparin-binding fragment (Hep-1), less to the 45 kDa gelatine-binding fragment, and not at all to the 120 kDa fragment containing the cell-binding domain (Fig 10) The amount of FXIIa that bound to the 30 kDa Hep-1binding fragment was similar to the amount of FXIIa bound to FN The N-terminal 30 kDa Hep-1binding domain has also been identified as a binding site for fibrinogen and fibrin [25,26] Further evidence for FXIIa binding to this domain was therefore provided, showing that the binding of FXIIa to immobilized FN was inhibited in a concentrationdependent manner by both fibrin generated by incubation of fibrinogen with thrombin and fibrinogen As compared to the inhibition by fibrin, however, an approximately 100-fold higher fibrinogen concentration was needed to yield an identical amount of inhibition (Fig 11) Discussion Although the presence in the blood of FXII has been known for more than 50 years, its physiological func- FEBS Journal 275 (2008) 5161–5172 ª 2008 The Authors Journal compilation ª 2008 FEBS I Schousboe et al Factor XII binding to fibronectin Fib-1/Hep-1 Fig Schematic diagram of the modular structure of the FN monomer The FN dimer is formed through interchain disulfide bonds at the C-terminus Each subunit consists of type I, type II and type III repeating modules Sets of repeats form domains of regions implicated in adhesion of different ligands The squares show the positions and the sizes of the different fragments Gelatine Cell Hep-2 Fib-2 COOH NH2 SS 30 kDa 45 kDa 120 kDa FXIIa binding Type II Type I Type III FXIIa bound, absorbance units 2.0 1.5 * * 1.0 * 0.5 0.0 Blo ck He bu pa ffe r He rin pa ,2 0µ Ge g·m rin lat ,4 0µ L –1 g·m Ge ine lat ,3 3µ L –1 DS ine g·m L –1 30 DS ,2 0µ ,3 ,4 µg g·m ·m L –1 0µ g·m L –1 L –1 Fig The effect of gelatine and heparin on binding of FXIIa to immobilized FN The microtiter plate was incubated overnight with FN (10 lgỈmL)1) and NaCl ⁄ Pi, respectively, and blocked with blocking buffer Then, it was incubated for h with FXIIa (20 nM) in blocking buffer containing heparin (20 and 40 lgỈmL)1), gelatine (33 and 330 lgỈmL)1) or DS (20 and 40 lgỈmL)1) The amount of bound FXIIa was determined by sequential incubation with goat anti-FXII IgG and HRP-conjugated rabbit anti-(goat IgG) and visualized by reactions with OPD as described in Experimental procedures Results are mean ± SD (n = 3), shown by vertical bars Statistically significant differences between FXIIa bound to FN incubated in the presence and in the absence of effectors are indicated by asterisks (*P < 0.001).The binding to control wells was less than 0.05 absorbance units tion is still not known For the past 15 years it has been assumed that its function is connected with Zn2+-dependent binding to a surface or a receptor The present study has demonstrated that in purified systems, activated FXII (FXIIa), but not its zymogen (FXII), binds with high affinity to immobilized FN The binding is independent of the presence of Zn2+, is not affected by the presence of a negatively charged surface represented by sulfatides, and is not inhibited by soluble FN Accordingly, soluble FN did not bind to immobilized FXIIa The binding of FXIIa to immobilized FN occurs through type I modules in the 30 kDa N-terminal heparin (Hep-1)-binding and fibrin (Fib-1)-binding domain of FN Immunohistochemical visualization of the interaction between FXIIa and FN deposited on the surface of the culture dish during days of growth of HUVECs clearly showed that FXIIa associated with FN left behind on the plastic surface after removal of the cells The visualization showed that FN had been deposited in a sparse and patchy manner, which may reflect the conditions under which the cells had been cultivated and subsequently removed by EDTA extraction Indeed, the majority of the deposited FN was attached to the cells and was thus removed during extraction of the cells Furthermore, experiments with cultures of arterial endothelial cells have shown that the amount of FN deposited on the surface of the cells varied dramatically when preconfluent, newly confluent and postconfluent cultures were analyzed Thus, whereas sparse patches of FN were generated in preconfluent and newly confluent cultures, a massive net FEBS Journal 275 (2008) 5161–5172 ª 2008 The Authors Journal compilation ª 2008 FEBS 5167 Factor XII binding to fibronectin I Schousboe et al 2.0 1.8 FXIIa bound, absorbance units 1.6 Absorbance units 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0 PBS buffer 30 kDa 45 kDa coating 120 kDa FN 1.0 0.5 0.0 100 200 300 400 500 600 700 Concentration of fibrinogen/fibrin, nM Fig 10 The binding of FXIIa to immobilized fragments of FN The microtiter plate was incubated overnight with the 30 kDa heparinand fibrin-binding fragment, the 45 kDa gelatine-binding fragment, the 120 kDa cell-binding fragment, and FN, respectively The fragments, as well as FN, were coated at a concentration of 10 lgỈmL)1 in NaCl ⁄ Pi The plate was then washed, blocked with blocking buffer, and incubated for h with FXIIa (20 nM) in blocking buffer The amount of bound FXIIa was determined by sequential incubation with goat anti-FXII IgG and HRP-conjugated rabbit anti(goat IgG) and visualized by reactions with OPD, as described in Experimental procedures Results are means ± SD (n = 3), shown by vertical bars of FN covering the entire surface of the cells was formed only in postconfluent cultures [28] It may be claimed that the deposited FN originates from the serum present in the cell culture medium However, the lack of appearance of deposited FN on culture dishes incubated with the medium using the same conditions and periods of time as in the presence of cells but in their absence showed that the deposited FN in the present investigation was generated by a cell-mediated process This process induces conformational changes in FN, exposing cryptic sites of importance for fibril generation and elongation [28–30] The high-affinity binding of FXIIa to the ECM with a KD of 12.8 nm [5] and the binding of FXIIa to the immobilized FN with a KD of 8.5 nm make it probable that FN, whether deposited during growth of HUVECs or coated on a plastic surface, constitutes a binding site for FXIIa Indeed, this binding site was found not to be present in soluble FN, as soluble FN was unable to inhibit the binding of FXIIa to immobilized FN Together with the observed lack of inhibition by an antibody against soluble FN, this suggests that the association between FXIIa and FN involves a cryptic site in FN Such a binding site has been shown to be also responsible for the interaction of FN with fibrinogen and fibrin [27] Hence, fibrinogen and fibrin inhibited the binding of FXIIa The binding of fibrinogen and fibrin has been mapped to type I modules of 5168 1.5 Fig 11 Fibrin inhibition of FXIIa binding to immobilized FN The microtiter plate was incubated overnight with FN (10 lgỈmL)1) and NaCl ⁄ Pi, respectively, and blocked with blocking buffer Meanwhile, 1.74 lM fibrinogen dissolved in blocking buffer was incubated overnight with 90 mmL)1 thrombin or blocking buffer at room temperature, and subsequently diluted with blocking buffer containing hirudin (100 mL)1) to give the indicated final concentrations of fibrinogen and fibrin after mixing with FXIIa (final concentration: 20 nM) The presence of hirudin did not affect the binding of FXIIa to FN, and the concentration of hirudin was sufficiently high to completely block the activity of thrombin The amounts of FXIIa bound to FN in the presence of fibrinogen ( ) and fibrin (d), and the amount of FXIIa bound to control wells (s), were determined by sequential incubation with goat anti-FXII IgG and HRP-conjugated rabbit anti-(goat IgG) and visualized by reactions with OPD, as described in Experimental procedures Results are mean ± SD (n = 3), shown by vertical bars when extending beyond the symbols FN present both N-terminally and C-terminally (Fig 8) Binding of FXIIa to the 30 kDa N-terminal fragment of FN indicates that FXIIa binds to FN through the type I modules in the cryptic N-terminal end of FN but does not exclude the possibility that FXIIa may also interact with the C-terminal Fib-2binding site The binding site in FXIIa is unknown, but lack of inhibition of the binding of FXIIa to FN by sulfatides and the surface-binding peptide of FXII strongly indicates that the binding does not involve the surfacebinding region in FXIIa [18] The lack of inhibition of FXIIa binding to immobilized FN by the surface-binding peptide strengthens the statement that FN is the target for the binding of FXIIa to the ECM, as this binding also could not be inhibited by the peptide [5] Thus, the affinities for FXIIa binding to ECM and to immobilized FN were the same, and neither one of the binding events could be inhibited by the surfacebinding peptide of FXIIa The binding to immobilized FN was specific for FXIIa, as FXII did not bind This indicates that the binding is of no physiological relevance for the FEBS Journal 275 (2008) 5161–5172 ª 2008 The Authors Journal compilation ª 2008 FEBS I Schousboe et al activation of FXII The binding of FXIIa to the same domain as fibrin and fibrinogen indicates, however, that FXIIa may interfere with fibril formation and elongation during fibrillogenesis and not with the binding of FN to its cellular receptors Further studies are needed to determine whether and how the binding of FXIIa to immobilized FN regulates these processes FXII was observed not to bind to immobilized FN, but soluble FN bound to immobilized FXII, and immunoprecipitates of plasma FN revealed the presence of FXII This indicates a role of FN in the activation and function of FXII The general concept of the function of FXII is connected to its binding to a surface This generates FXIIa, which circumstantially can cleave FXI and prekallikrein However, the mechanism of this activation in vivo has still not been elucidated Furthermore, the significance of FXIIa for the activation of FXI and prekallikrein in vivo has been questioned, as FXII deficiency is not associated with hemophilia In addition, FXI can be activated by thrombin [31], and prekallikrein by a prolylcarboxypeptidase [32] and the HSP90 protein [33] However, recent investigations have shown that FXII in vivo plays an important role in thrombus formation, being activated on the surface of activated platelets [34] to which FN binds [35] The mechanism for this activation is unknown, but although speculative, the present investigation may be of importance in understanding the impact of FXII in thrombus formation Thus, the binding of FXII to soluble FN may be of relevance for the activation of FXII on the surface of activated platelets, but this remains to be established Experimental procedures Materials FXII and thrombin were obtained as 50% glycerol solutions from Haematologic Technologies Inc (Essex Junction, VT, USA) and stored at )20 °C; FXII appeared as a single band with a molecular mass of 80 kDa in reduced SDS ⁄ PAGE (Fig 4) Lyophilized FXIIa was obtained from Enzyme Research Laboratories (Swansea, UK) FXIIa was dissolved in water as recommended by the company, and stored in aliquots in siliconized test tubes at )80 °C Siliconized test tubes were likewise used for subsequent dilutions of FXII and FXIIa, and excess dilutions were discarded Human plasma FN was from Gibco (Invitrogen, Carlsbad, CA, USA) CTI, hirudin, the N-terminal 29 kDa heparin-binding fragment and the 45 kDa gelatine-binding fragment were from Sigma Chemicals (St Louis, MO, USA) The 120 kDa cell-binding fragment was obtained Factor XII binding to fibronectin from Chemicon (AH Diagnostics, Aarhus, Denmark) YHKCTHKGR(39–47), the surface-binding region of FXII ⁄ FXIIa, was a gift from A H Schmaier (Case Western Reserve University, Cleveland, OH, USA) Fibrinogen from bovine serum was obtained lyophilized from citrate buffer (pH 7.4) It was purchased from Calbiochem (La Jolla, CA, USA) The concentration of fibrinogen in solution was determined at 280 nm absorbance using an extinction coefficient (E1%280 nm) of 15.1 Heparin [sodium salt; H3125; Grade from porcine intestinal mucosa (181 USP unitsỈmg)1)] was from Sigma Chemicals, and DS (sodium salt; Mr  500 000) was from Pharmacia Fine Chemicals (Uppsala, Sweden) All other chemicals were of the purest grade commercially available Affinity-purified goat anti-(human FXII) IgG (GAFXIIAP) was from Affinity Biologicals Inc (Hamilton, ON, Canada) Rabbit anti-FN IgG (ab 299) and monoclonal antibody to FN, (Fn-3, ab 18265), which reacts with human cellular fibronectin but not with plasma fibronectin, were from Abcam (Cambridge, UK) HRP-conjugated rabbit anti-(goat IgG) (P-0449), HRP-conjugated swine anti(rabbit IgG) (P-0399) and o-phenylenediamine (OPD) were from DAKOCytomation (Ejby, Denmark) Secondary Alexa 488 ⁄ 594-conjugated antibodies for immunofluorescence microscopy were from Invitrogen (Copenhagen, Denmark) Solid-phase binding assay The solid-phase binding assay was performed in 96-well maximum-binding polystyrene microtiter plates (NUNC, Roskilde, Denmark) The plates were coated with 150 lL per well of either 10 lgỈmL)1 FN or FN fragments in NaCl ⁄ Pi (0.1 m sodium phosphate, pH 7.4) and incubated overnight at °C Control wells were coated concurrently with NaCl ⁄ Pi This was followed by two washing cycles with Locke’s buffer (154 mm NaCl, 5.6 mm KCl, 3.6 mm NaHCO3, 2.3 mm CaCl2, 5.6 mm glucose, mm Hepes, pH 7.4), and unoccupied binding sites were blocked by incubation for a minimum of 30 at room temperature or overnight at °C with 200 lL per well of blocking buffer [0.35% (w ⁄ v) of essentially fatty acid-free BSA (A7030; Sigma Chemicals) dissolved in Locke’s buffer] The wells were then incubated for 60 with FXII or FXIIa added in a final volume of 100 lL in blocking buffer in the presence or absence of 20 lgỈmL)1 sulfatides Bound FXII antigens were measured following washing of the wells with washing buffer [Tween-20 0.05% v ⁄ v in NaCl ⁄ Tris (50 mm Tris, 0.15 mm NaCl, pH 8.0)] The wells were then incubated for h with goat anti-(human FXII) IgG, diluted : 2000 in 1% (w ⁄ v) skimmed milk in washing buffer, and for h with HRP-conjugated secondary antibodies diluted : 2500 in the skimmed milk solution Extensive washing with washing buffer was performed between each change of FEBS Journal 275 (2008) 5161–5172 ª 2008 The Authors Journal compilation ª 2008 FEBS 5169 Factor XII binding to fibronectin I Schousboe et al incubation conditions Finally, the plates were incubated for 10–30 with OPD, dissolved in water according to the manufacturer’s recommendations The peroxidase reaction was stopped by twofold dilution with 0.5 m H2SO4, and the relative amount of bound FXII antigen was determined as absorbance units at 490 nm All experiments were performed in triplicate and repeated at least twice To obtain estimates of affinity constants, the data were analyzed according to the isotherm A ẳ Amax [FXIIa]/(KD ỵ [FXIIa]) where [FXIIa] is the molar concentration of FXIIa, A is the absorbance of the oxidized HRP substrate, which is assumed to be proportional to the amount of FXIIa bound, and Amax represents the absorbance at saturating concentrations of FXIIa Alternatively, the microtiter plate was coated with 20 nm FXII or FXIIa in NaCl ⁄ Pi, and incubated with FN (10 lgỈmL)1) The amount of soluble FN bound to immobilized FXII or FXIIa was visualized by sequential incubation with rabbit anti-FN IgG and HRP-conjugated swine anti(rabbit IgG), both diluted : 2000 in 1% (w ⁄ v) skimmed milk in washing buffer, and OPD, as described above Immunoprecipitation Ten microliters of rabbit anti-FN IgG was added to one of two aliquots containing 200 lL of plasma, lL of hirudin (10 mL)1) and 0.4 lL of CTI (10 lgỈmL)1), and the mixtures were rotated overnight at °C Then 200 lL of a : slurry of protein G–Sepharose (Sigma-Aldrich, St Louis, MO, USA) was added to each aliquot, and the rotation was continued for another night Following centrifugation (1 min, 2000 g) and 10-fold washing of the precipitate with 0.5 mL of NaCl ⁄ Tris (10 mm Tris, mm EDTA, mm EGTA, 0.2 m NaCl, pH 7.4), the protein adsorbed to the protein G–Sepharose was extracted by boiling for 10 with 100 lL of SDS ⁄ glycerol ⁄ dithiothreitol according to the standard procedure for SDS ⁄ PAGE electrophoresis Immunofluorescence microscopy For immunofluorescence microscopy of FXIIa bound to the ECM, HUVECs were plated on eight chamber slides (Nalgene Nunc International Corp., Roskilde, Denmark) at a density of 104 cellsỈcm)2, and grown with a change of medium on the second day On day 4, the cells were detached by EDTA The ECM was incubated with 20 nm FXIIa in blocking buffer for h After the washing procedure described above for the solid-phase binding assay, the slides were incubated with antibodies The primary antibodies were a mixture of goat anti-FXII IgG (diluted : 100) and mouse anti-FN IgG (Fn-3) (diluted : 100) The secondary antibodies were a mixture of Alexa 594-conjugated donkey anti-(goat IgG) (diluted : 800) and Alexa 588-conjugated goat anti-(mouse IgG) (1 : 800) Finally, the slides were mounted in antifade medium (DAKOCytomation, Ejby, Denmark) and examined in a Leica DM 4000 B microscope equipped with a Leica DC 300 FX digital camera Specificity analyses of the antibodies showed no reaction of the secondary antibodies with the ECM incubated in the absence of the primary antibodies Sulfatide preparation Sulfatides extracted from bovine brain were from Sigma Chemicals Vesicles of sulfatides were prepared as previously described [5] Statistics The results are shown as means ± SD, and statistically significant differences were calculated using Student’s t-test SDS/PAGE and immunoblotting For western blot analysis, bound proteins were extensively washed with Locke’s buffer and then extracted with electrophoresis buffer containing 2% (w ⁄ v) SDS and 0.1 m dithiothreitol Aliquots of the extracts and FXII, FXIIa and molecular weight markers were run simultaneously Proteins were separated on 4–12% SDS ⁄ polyacrylamide gels, and transferred to poly(vinylidene difluoride) membranes according to standard procedures The membrane was then incubated for h with NaCl ⁄ Tris blocking buffer (50 mm Tris, 0.15 mm NaCl, pH 8.0, containing 0.1% v ⁄ v Tween20 and 0.1% w ⁄ v BSA) and probed with goat anti-FXII IgG (diluted : 5000) ⁄ HRP-conjugated rabbit anti-(goat IgG) (diluted : 5000) Dilutions of antibodies were per- 5170 formed in 1% nonfat skimmed milk in NaCl ⁄ Tris blocking buffer Detection was carried out using the chemiluminescence enhancer SuperSignal West Femto Maximum Sensitivity Substrate (Pierce Biotechnology, Rockford, IL, USA) as recommended by the manufacturer, and the results were monitored on a Las Chemiluminator Acknowledgements The work was supported by grants 2005-1-192 and 2006-1-0247 from the Carlsberg Foundation References Hasan AA, Cines DB, Ngaiza JR, Jaffe EA & Schmaier AH (1995) High-molecular-weight kininogen is exclusively membrane bound on endothelial cells to influence activation of vascular endothelium Blood 85, 3134– 3143 FEBS Journal 275 (2008) 5161–5172 ª 2008 The Authors Journal compilation ª 2008 FEBS I Schousboe et al Motta G, Rojkjaer R, Hasan AA, Cines DB & Schmaier AH (1998) High molecular weight kininogen regulates prekallikrein assembly and activation on endothelial cells: a novel mechanism for contact activation Blood 91, 516–528 Joseph K, Ghebrehiwet B & Kaplan AP (1999) Cytokeratin and gC1qR mediate high molecular weight kininogen binding to endothelial cells Clin Immunol 92, 246–255 Schousboe I (2001) Rapid and cooperative binding of factor XII to human umbilical vein endothelial cells Eur J Biochem 268, 3958–3963 Schousboe I (2006) Endothelial cells express a matrix protein which binds activated factor XII in a zincindependent manner Thromb Haemost 95, 312– 319 Erickson HP & Carrell NA (1983) Fibronectin in extended and compact conformations Electron microscopy and sedimentation analysis J Biol Chem 258, 14539–14544 Johnson KJ, Sage H, Briscoe G & Erickson HP (1999) The compact conformation of fibronectin is determined by intramolecular ionic interactions J Biol Chem 274, 15473–15479 Petersen TE, Thøgersen HC, Skorstengaard K, VibePedersen K, Sahl P, Sottrup-Jensen L & Magnusson S (1983) Partial primary structure of bovine plasma fibronectin: three types of internal homology Proc Natl Acad Sci USA 80, 137–141 Hocking DC, Sottile J & McKeown-Longo PJ (1994) Fibronectin’s III-1 module contains a conformationdependent binding site for the amino-terminal region of fibronectin J Biol Chem 269, 19183–19187 10 Oberhauser AF, Badilla-Fernandez C, Carrion-Vazquez M & Fernandez JM (2002) The mechanical hierarchies of fibronectin observed with single-molecule AFM J Mol Biol 319, 433–447 11 Narasimhan C & Lai CS (1989) Conformational changes of plasma fibronectin detected upon adsorption to solid substrates: a spin-label study Biochemistry 28, 5041–5046 12 Lhoest J, Detrai E, van den Bosch de Aguilar P & Bertrand P (1998) Fibronectin adsorption, conformation, and orientation on polystyrene substrates studied by radiolabeling, XPS, and ToF SIMS J Biomed Mater Res 41, 95–103 13 Baugh L & Vogel V (2004) Structural changes of fibronectin adsorbed to model surfaces probed by fluorescence resonance energy transfer J Biomed Mater Res 69A, 525–534 14 Mao Y & Schwarzbauer JE (2005) Fibronectin fibrillogenesis, a cell mediated matrix assembly process Matrix Biol 24, 389–399 15 Rostagno A, Williams M, Frangione B & Gold LI (1996) Biochemical analysis of the interaction of Factor XII binding to fibronectin 16 17 18 19 20 21 22 23 24 25 26 27 28 fibronectin with IgG and localization of the respective binding sites Mol Immunol 33, 561–572 Tans G, Rosing J & Griffin JH (1983) Sulfatide-dependent autoactivation of human blood coagulation factor XII (Hageman factor) J Biol Chem 258, 8215–8222 Rojkjaer R & Schousboe I (1997) The surface-dependent autoactivation mechanism of factor XII Eur J Biochem 243, 160–166 Citarella F, te Velthuis H, Helmer-Citterich M & Hack CE (2000) Identification of a putative binding site for negatively charged surfaces in the fibronectin type II domain of human factor XII – an immunochemical and homology modeling approach Thromb Haemost 84, 1057–1065 Zenhausen F, Adrian M & Descouts P (1993) Solution structure and direct imaging of fibronectin adsorption to solid surfaces by scanning force microscopy and cryo-electron microscopy J Electron Microsc (Tokyo) 42, 378–388 Ingham KC, Brew SA & Erickson HP (2004) Localization of a cryptic binding site for tenascin on fibronectin J Biol Chem 279, 28132–28135 Gold LI, Frangione B & Pearlstein E (1983) Biochemical and immunological characterization of three binding sites on human plasma fibronectin with different affinities for heparin Biochemistry 22, 4113–4119 Yamada KM, Kennedy DW, Kimata K & Pratt RM (1980) Characterization of fibronectin interactions with glycosaminoglycans and identification of active proteolytic fragments J Biol Chem 255, 6055–6063 Ingham KC, Brew SA & Atha DH (1990) Interaction of heparin with fibronectin and isolated fibronectin domains Biochem J 272, 605–611 Rostagno A, Williams MJ, Baron M, Campbell ID & Gold LI (1994) Further characterization of the NH2terminal fibrin-binding site on fibronectin J Biol Chem 269, 31938–31945 Matsuka YV, Medved LV, Brew SA & Ingham KC (1994) The NH2-terminal fibrin-binding site of fibronectin is formed by interacting fourth and fifth finger domains Studies with recombinant finger fragments expressed in Escherichia coli J Biol Chem 269, 9539– 9546 Pickford AR, Potts JR, Bright JR, Phan I & Campbell ID (1997) Solution structure of a type module from fibronectin: implications for the structure and function of the gelatine-binding domain Structure 5, 359–370 Makogonenko E, Tsurupa G, Ingham K & Medved L (2002) Interaction of fibrin(ogen) with fibronectin: further characterization and localization of the fibronectin-binding site Biochemistry 41, 7907–7913 Magnusson MK & Mosher DF (1998) Fibronectin: structure, assembly, and cardiovascular implications Arterioscler Thromb Vasc Biol 18, 1363–1370 FEBS Journal 275 (2008) 5161–5172 ª 2008 The Authors Journal compilation ª 2008 FEBS 5171 Factor XII binding to fibronectin I Schousboe et al 29 Schwarzbauer JE & Sechler JL (1999) Fibronectin fibrillogenesis: a paradigm for extracellular matrix assembly Curr Opin Cell Biol 11, 622–627 30 Kowalczyk AP, Tulloh RH & McKeown-Longo PJ (1990) Polarized fibronectin secretion and localized matrix assembly sites correlate with subendothelial matrix formation Blood 75, 2335–2342 31 Gailani D & Broze GJ Jr (1991) Factor XI activation in a revised model of blood coagulation Science 253, 909– 912 32 Shariat-Mada Z, Mahdi F & Schmaier AH (2002) Identification and characterization of prolylcarboxypeptidase 5172 as an endothelial cell prekallikrein activator J Biol Chem 277, 17962–17969 33 Joseph K, Tholanikunnel BG & Kaplan AP (2002) Activation of the bradykinin-forming cascade on endothelial cells: a role for heat shock protein 90 Int Immunopharmacol 2, 1851–1859 ´ 34 Renne T, Nieswandt B & Gailani D (2006) The intrinsic pathway of coagulation is essential for thrombus stability in mice Blood Cells Mol Dis 36, 148–151 35 Cho J & Mosher DF (2006) Characterization of fibronectin assembly by platelets adherent to adsorbed laminin-111 J Thromb Haemost 4, 943–951 FEBS Journal 275 (2008) 5161–5172 ª 2008 The Authors Journal compilation ª 2008 FEBS ... FXII Factor XII binding to fibronectin 1.8 FN Control 1.6 * 1.4 1.2 0.8 ** 0.6 0.4 0.2 lf I– XI I– with FXII and FXIIa instead of FN Then, the binding of soluble FN to immobilized FXII and FXIIa... binding to ECM and to immobilized FN were the same, and neither one of the binding events could be inhibited by the surfacebinding peptide of FXIIa The binding to immobilized FN was specific for. .. FXIIa and immobilized FN and the lack of interference by soluble FN indicated that the binding site on FN for FXIIa may be buried in the compact soluble form of FN [7] FN has binding sites for