BM-40) binds to fibrinogen fragments D and E, but not to native fibrinogen

Matrix Biology 25 (2006) 20 – 26 www.elsevier.com/locate/matbio

Secreted protein acidic and rich in cysteine (SPARC/osteonectin/BM-40) binds to fibrinogen fragments D and E, but not to native fibrinogen Hua Wang a, Gail Workman b, Shengfu Chen a, Thomas H. Barker b, Buddy D. Ratner a,*, ER Helene Sage b,*, Shaoyi Jiang a,* b

a Department of Chemical Engineering, University of Washington, Seattle, WA 98195, United States Hope Heart Program, Benaroya Research Institute at Virginia Mason, Seattle WA 98101, United States

Received 22 January 2005; received in revised form 18 July 2005; accepted 15 September 2005

Abstract Secreted protein acidic and rich in cysteine (SPARC/osteonectin/BM-40) is a matricellular protein that functions in wound healing. Fibrinogen is a plasma protein involved in many aspects of wound healing, such as inflammation, fibrosis and thrombosis. In this study, the binding of SPARC to both native and plasmin-cleaved fibrinogen under physiological conditions was examined by the use of a surface plasmon resonance (SPR) biosensor. We show that SPARC binds to plasmin-cleaved fibrinogen, but not to native fibrinogen. SPARC binds to both fibrinogen fragments D and E fg D and fg E with similar dissociation constants (8.67  10 8 M for Fg D and 1.61 10 7 M for Fg E). Results from endothelial cell proliferation assays show that the binding of SPARC to Fg E suppressed the inhibition of proliferation by SPARC, whereas the binding of SPARC to Fg D did not influence the activity of SPARC on the cell cycle. The interaction of SPARC with fibrinogen fragments D and E, which are produced as a result of proteolytic activation of fibrinolysis, reveals potential storage sites in provisional extracellular matrix for SPARC during the wound healing process and indicates a regulatory role of SPARC in fibrinolysis and angiogenesis. D 2005 Elsevier B.V./International Society of Matrix Biology. All rights reserved. Keywords: SPARC; Fibrinogen; SPR; Fibrinolysis; Wound healing

1. Introduction SPARC (secreted protein acidic and rich in cysteine, also known as osteonectin and BM-40) is a 32-kDa calcium-binding glycoprotein secreted by many types of cells, e.g., endothelial cells, fibroblasts, and platelets (Brekken and Sage, 2000). It belongs to a class of proteins termed ‘‘matricellular’’ that are involved in cell – extracellular matrix (ECM) interactions during tissue repair and differentiation (Bradshaw and Sage, Abbreviations: BSA, bovine serum albumin; DMEM, Dulbecco’s modified Eagle’s medium; ECM, extracellular matrix; FBS, fetal bovine serum; Fg D, fibrinogen fragment D; Fg E, fibrinogen fragment E; MMP, matrix metalloproteinase; SAM, self-assembled monolayer; SP, SPARC; SPARC, secreted protein acidic and rich in cysteine; SPR, surface plasmon resonance. * Corresponding authors. Sage is to be contacted at Tel.: +1 206 341 1314. Ratner, Tel.: +1 206 685 1005. Jiang, Tel.: +1 206 616 6509. E-mail addresses: [email protected] (B.D. Ratner), [email protected] (E.H. Sage), [email protected] (S. Jiang).

2001). Elevated expression of SPARC was observed at early time points after injury and during remodeling of ECM (Hunzelman et al., 1998). The principal functions of SPARC include disrupting cell – ECM interactions, inhibiting cell-cycle progression, and regulating the expression of a number of growth factors (Raines et al., 1992), ECM proteins (Kamihagi et al., 1994), and matrix metalloproteinases (MMPs) (Tremble et al., 1993). Given that many of these functions are requisite to wound repair, SPARC is believed to play a significant role in the control of wound healing, as confirmed by studies in vivo. For example, mice deficient in SPARC exhibit accelerated closure of both excisional and incisional dermal wounds (Puolakkainen et al., 2003). The precise biological functions of SPARC in wound healing remain to be elucidated. For example, little is known about the interaction of SPARC with fibrinogen, one of the most important proteins in wound healing that affects inflammation, fibrosis, and thrombosis (Drew et al., 2001). In the present study, we asked whether SPARC binds to

0945-053X/$ - see front matter D 2005 Elsevier B.V./International Society of Matrix Biology. All rights reserved. doi:10.1016/j.matbio.2005.09.004

H. Wang et al. / Matrix Biology 25 (2006) 20 – 26

fibrinogen, and whether proteolytic processing of fibrinogen by plasmin, which occurs in a developing wound, would affect the putative interaction. We used a surface plasmon resonance (SPR) biosensor to measure the binding of SPARC to native or plasmin-cleaved fibrinogen. In addition, we determined the binding affinities of SPARC for fibrinogen fragments D and E (FgD and FgE). SPR biosensors have been widely used for the identification of protein – protein interactions because they allow direct real-time and label-free monitoring of processes with high sensitivity (Homola et al., 2001). The SPR chips were functionalized with carboxylic acid terminated selfassembled monolayers of alkanethiolates (COOH – SAMs), to create a platform with precisely controlled surface chemistry and nano-scale structure (Wink et al., 1997). It should be pointed out that the control of non-specific binding is an important step to ensure the reliable identification of binding pairs. Thus preliminary studies were performed to select the optimal blocking agent. In addition to the control of nonspecific binding, we also used one of the dual channels as a reference to compensate for any drift caused by changes in temperature and flow conditions. Herein we report that SPARC bound specifically to Fg D and Fg E, whereas no interaction occurred with native, uncleaved fibrinogen. Specific binding of SPARC to plasmin-cleaved fibrinogen could regulate the wound healing process, not only by augmentation of SPARC levels in the local environment, but also by modification of the cellular functions of SPARC. For example, through its specific association with plasmin-cleaved fibrinogen, the conformation of SPARC, as well as the accessibility of the functional domains of SPARC to specific cellular receptors, could be altered. Several recent reports have indicated that SPARC inhibits the proliferation of bovine aortic endothelial (BAE) cells, with growth arrest occurring in the mid-G1 phase of the cell cycle (Lane and Sage, 1994). In this study, endothelial cell proliferation assays were performed to investigate whether the binding of SPARC to fibrinogen fragments influenced its general inhibitory effect on the cell cycle. We report that the binding of SPARC to Fg E, but not to Fg D, suppressed the inhibition of cell proliferation by SPARC. 2. Results and discussion 2.1. Cleavage of fibrinogen by plasmin Electrophoresis on 4– 20% SDS-polyacrylamide gradient gels under non-reducing conditions was performed to characterize both native and plasmin-cleaved fibrinogen (Fig. 1A). As shown in Fig. 1B (Walker and Nesheim, 1999), cleavage of the aC regions of fibrinogen by plasmin gave rise to fragment X; subsequent cleavage of X at the plasmin-sensitive site yielded one fragment Y and one fragment D. Further cleavage of fragment Y produced a second fragment D and fragment E. As can be seen from Fig. 1A, native fibrinogen (lane 1) migrated as one major band, whereas plasmin-cleaved fibrinogen (lane 2) showed several major bands corresponding to X, Y, D and E, respectively.

A

1

250 kDa 150

21

B

2 X Y

100 75

D

50

E

37 25 20 15

Name

Structure

kDa

Fg

340

X

260

Y

160

D

100

E

60

Fig. 1. SDS-polyacrylamide gel electrophoresis (non-reducing) of native (lane 1) and plasmin-cleaved (lane 2) fibrinogen (A), and the structures of native and plasmin-cleaved fibrinogen (B) (redrawn from Walker and Nesheim, 1999). Under the conditions used, native fibrinogen was thoroughly digested into a mixture of plasmin-cleaved products X, Y, D and E.

The absence of bands in the control sample (data not shown), containing only plasmin and a2-antiplasmin in amounts identical to those in the sample of plasmin-cleaved fibrinogen, indicated that all the bands in lane 2 were derived from fibrinogen. Thus, under the conditions used, native fibrinogen was thoroughly digested into a mixture of plasmin-cleaved products X, Y, D and E. 2.2. Control of non-specific binding Our preliminary studies showed that SPARC tends to adsorb non-specifically onto the COOH – SAMs covered SPR chips, which could be partly due to the relatively small size (< 35,000) and acidic nature (pI 4.7) of SPARC. Thus, the control of nonspecific binding is critical for the reliable identification of specific SPARC binding. In this study, we compared two commonly used blocking proteins, casein and bovine serum albumin (BSA), for blocking of non-specific adsorption of SPARC. Control experiments were performed following the procedure shown in Fig. 2, except that casein or BSA (instead of fibrinogen) was adsorbed onto the SPR chip in the first step. Thus, any SPARC binding detected in the subsequent step was attributed to non-specific binding. With BSA, non-specific binding of SPARC was significant (data not shown). With casein, however, there was minimal non-specific binding of SPARC, as shown in Fig. 3A and E. Therefore, casein was used as a blocking agent throughout this study. Understanding and control of non-specific adsorption of proteins are critical steps toward improving the accuracy of SPR binding assays. Since the two proteins commonly used as blocking agents are casein and BSA, our laboratory is currently performing systematic comparison of these two proteins as blocking agents on different SAMs using SPR. Preliminary results showed that casein blocks various proteins (e.g. fibrinogen, SPARC, and osteopontin) better on COOH – and NH2 – SAMs, whereas BSA blocks better on CH3 – SAMs. Generally, an optimal blocking agent should be chosen based on both the property of the adsorbed protein and the surface chemistry (Lai et al., 2004).

22

H. Wang et al. / Matrix Biology 25 (2006) 20 – 26

Wavelength

A

Sensing Channel glutaraldehyde

805 800 795 790 785 780 775 770 765 760

SPARC

blocking agent: (e.g., casein) Fg 0

20

40

60

80

100

120

140

Time

B

Reference Channel glutaraldehyde

805 800

buffer

Wavelength

795 790 785 780 blocking agent: (e.g., casein)

775 770 Fg

765 0

20

40

60

80

100

120

140

Time

Wavelength

C

Compensated Response

–1.6 –1.7 –1.8 –1.9 –2.0 –2.1 –2.2 –2.3 –2.4 –2.5 –2.6

with plasmin and a2-antiplasmin in amounts identical to those in plasmin-cleaved fibrinogen, was immediately adsorbed onto the SPR chip without incubation at 37 -C. There was minimal binding of SPARC on the surface covered with a freshly prepared mixture of plasmin, a2-antiplasmin, and native fibrinogen (data not shown). It is not surprising that the interaction between SPARC and fibrinogen is affected by the proteolysis of fibrinogen by plasmin, because enzymatic cleavage has been shown to regulate the exposure of biologically active sequences to provide important signals within the injury site (Davis et al., 2000). It has been reported that SPARC binds to plasminogen, enhances its activation, and anchors it at sites of tissue remodeling (Kelm et al., 1994). SPARC has also been shown to increase the expression of certain MMPs (Tremble et al., 1993). These studies suggested a role for SPARC in the regulation of fibrinolysis, with the hypothetical matrix-binding site of SPARC to be identified. That SPARC binds to plasmin-cleaved fibrinogen indicates that the latter could serve as a storage matrix for SPARC during tissue remodeling, as confirmed by the colocalization of SPARC and fibrin(ogen) degradation products in human carotid arteries from endarterectomy samples (Bini et al., 1999). Furthermore, the interaction of SPARC with plasmincleaved fibrinogen could indicate an important role for SPARC in a positive feedback loop that regulates fibrinolysis, as shown in Fig. 4. Increased levels of SPARC in the ECM, due to its binding to plasmin-cleaved fibrinogen, could in turn promote cleavage of fibrinogen, as a consequence of the activation of plasminogen and the increased expression of MMPs. 2.4. Binding of SPARC to fibrinogen fragments D and E

130

135

140

145

150

Time Fig. 2. A typical set of SPR response curves for binding assays. One of the dual channels was used as a reference channel, in which buffer instead of SPARC solution was applied in the final step. The compensated response (C) is obtained by subtracting the SPR response of the reference channel (B) from that of the sensing channel (A).

2.3. Binding of SPARC to native and plasmin-cleaved fibrinogen It can be seen from Fig. 3B and E that there is essentially no binding of SPARC to native fibrinogen, data indicating that native fibrinogen is not a storage matrix for SPARC. Also shown in Fig. 3C and E is the binding of SPARC to collagen I as a positive control, data consistent with previous reports (Maurer et al., 1997; Sasaki et al., 1997). We next asked whether SPARC could alternatively recognize fibrinogen digested by plasmin, that is, whether extracellular proteolytic processing, which occurs in a developing wound, could alter the interaction of fibrinogen with SPARC. Fig. 3D and E show that SPARC binds to plasmin-cleaved fibrinogen. To ensure that the binding of SPARC to plasmin-cleaved fibrinogen was due to the cleavage of fibrinogen by plasmin rather than the presence of plasmin/a2-antiplasmin, native fibrinogen, mixed

Since SPARC binds to fibrinogen only after proteolytic cleavage by plasmin, it appears that structural changes in fibrinogen expose biologically active cryptic sites. Fibrinogen is a dimeric molecule composed of three pairs of polypeptide chains covalently associated and organized into a central E domain and two spatially distant D domains connected by the putative coiled-coil region. The major components of plasmincleaved fibrinogen are fibrinogen fragments D, E, X and Y, with fragments X and Y as combinations of D and E (Fig. 1B). Thus, the binding of SPARC to purified Fg D and Fg E was also examined in an attempt to locate the SPARC binding sites on fibrinogen. Fig. 5 shows the concentration dependence of SPARC binding to Fg D and Fg E. The data were fit to Langmuir binding isotherms. Since the predicted local concentration of SPARC at a site of vessel damage due to platelet activation and release of a-granule constituents is ¨ 30 Ag/mL (Kelm et al., 1994), we chose a series of SPARC solutions with concentrations ranging from 5 to 30 Ag/mL. The obtained forward (k a) and reverse (k d) rate constants, as well as the equilibrium dissociation constant K d (calculated as K d = k d / k a) for the binding of SPARC to both Fg D and Fg E, are summarized in Table 1. While similar dissociation constants on the order of 10 7 M are found for both binding pairs, the binding of SPARC has higher k a and k d to Fg E than to Fg D. Thus, our data indicate that SPARC has binding sites on both D

H. Wang et al. / Matrix Biology 25 (2006) 20 – 26

B 0.9

0.9

0.8

0.8

0.7

0.7

Wavelength (nm)

Wavelength (nm)

A

0.6 0.5 0.4 0.3 0.2

0.6 0.5 0.4 0.3 0.2 0.1

0.1

0.0

0.0 0

400

800

1200

0

1600

Time (seconds) 0.9

0.8

0.8

0.7

0.7

Wavelength (nm)

Wavelength (nm)

800

1200

1600

D 0.9

0.6 0.5 0.4 0.3 0.2

0.6 0.5 0.4 0.3 0.2 0.1

0.1

0.0

0.0 0

400

800

1200

1600

0

Time (seconds)

Wavelength Shift (nm)

400

Time (seconds)

C

E

23

400

800

1200

1600

Time (seconds)

1.0 0.8 0.6 0.4 0.2 0.0 Casein

collagen I

fibrinogen

plasmincleaved fibrinogen

Fig. 3. Binding of SPARC to (A) casein, (B) fibrinogen, (C) collagen I, and (D) plasmin-cleaved fibrinogen. Results from 2 or 3 replicated experiments for each binding pair are summarized in (E), with the error bars representing standard deviations. SPARC binds to plasmin-cleaved fibrinogen, but not to native fibrinogen. The minimal binding of SPARC to casein as a negative control ensures the reliable identification of binding pairs. The binding of SPARC to collagen I is included as a positive control.

and E fragments of fibrinogen accessible only after proteolytic activation by plasmin. The precise SPARC binding sites on Fg D and Fg E remain to be determined. Previous studies have revealed that significant portions of the Aa and Bh chains are exposed on the native fibrinogen molecule, whereas most of the g chains are sequestered within the molecule (Gollwitzer et al., 1975). It has also been found that plasmin cleavage of fibrinogen leads to the exposure of g chain peptides located in both the C-terminus of the D domain and the N-terminus of the E domain, as well as a g95 – 264 sequence representing the central region of the g chain (Fair et al., 1981). Thus, the SPARC binding sites possibly lie within the g chain, which become accessible only after the disruption of the compact structure of fibrinogen by plasmin cleavage.

2.5. Effects of the binding of SPARC to fibrinogen fragments D and E on the inhibition of endothelial cell proliferation by SPARC Since SPARC has been shown to inhibit proliferation of BAE cells (Funk and Sage, 1991), we asked whether the SPARC Plasmin(ogen), MMPs

Fibrinogen

Fibrinolysis

Plasmin–cleaved fibrinogen

Fig. 4. A model depicting the participation of SPARC in a positive feedback loop that influences fibrinolysis.

24

H. Wang et al. / Matrix Biology 25 (2006) 20 – 26

Theoretical thymidine incorporation (% of control) 0.75

Wavelength (nm)

0.65 0.55 0.45

30µg/mL

0.35 20µg/mL 0.25 10µg/mL

0.15

5µg/mL

0.05 -0.05 0

500

1000

1500

2000

2500

Time (seconds)

Wavelength (nm)

B

80 60 40 20 0

0.75

+ -

+ +

+ -

0.65

Fg E

-

-

Fg

-

-

0.55 30µg/mL 0.45 20µg/mL

0.35

10µg/mL

0.25

5µg/mL

0.15 0.05 0

500

1000

1500

2000

2500

Time (seconds) Fig. 5. Binding of SPARC to (A) Fg D and (B) Fg E. Purified recombinant human SPARC was used at concentrations of 5 – 30 Ag/mL as shown. Smooth curves show the best fit of the data to 1 : 1 a binding isotherm.

binding of SPARC to fibrinogen fragments D and E modulated the inhibitory effect of SPARC. 3H-thymidine incorporation into cellular DNA in the presence of different reagents, represented as percentages of the thymidine incorporation by control cells in Dulbecco’s modified Eagle’s medium (DMEM) containing 2% fetal bovine serum (FBS), is shown in Fig. 6. Fibrinogen and its fragments D and E have been reported to bind to endothelial cells (Dejana et al., 1985). Fibrinogen and Fg E have been shown to promote and to inhibit cell migration, respectively (Bootle-Wilbraham et al., 2000), and Fg D increased endothelial monolayer permeability (Ge et al., 1992). To our knowledge, no effects of fibrinogen and its fragments on endothelial cell proliferation have been reported. There was no apparent effects of Fg D and Fg E on BAE cell proliferation. However, at the same molar concentration (0.06 AM), fibrinogen inhibited BAE cell proliferation by approximately 25%. This effect could be due to the inhibition of the adhesion of BAE cells to the substrate by fibrinogen via the recognition of its RGD site by avh3 integrin. SPARC also showed an inhibitory effect on BAE cell proliferation, as

Table 1 Binding of SPARC to fibrinogen fragments D and E measured by SPR biosensor

FgD FgE

100

SP Fg D

-0.05

Ligand

Thymidine incorporation (% of control)

A

k a (M

1

s

1

)

523.7 T 9.28 2617 T 19.52

k d (s

1

)

4.542 T 0.468  10 4.221 T 0.0546  10

K d (AM) 5 4

0.0867 0.161

+ -

+

-

-

+

-

-

+

-

-

+

-

-

+

Fig. 6. Effects of the interaction between SPARC and Fg D or Fg E on endothelial cell proliferation. 3H-thymidine incorporation of cells cultured in the presence of 0.06 AM SPARC, Fg D, Fg E, Fg alone, and their preincubated mixtures at 1 : 1 molar ratio are plotted as percentages of the thymidine incorporation of cells cultured in DMEM/2% fetal bovine serum (FBS) (closed bars). For cells cultured with preincubated protein mixtures, the theoretical percentages of thymidine incorporation (stippled bars) are calculated as the products of the percentages of thymidine incorporation by cells cultured with each component alone. Error bars represent standard deviations (n = 3) for each condition.

expected (Funk and Sage, 1991). We used a 10-fold lower concentration of recombinant human SPARC than the reported ED50 for BAE cells (reported as 20 Ag/mL, 0.6 AM by Funk and Sage, 1991), because an inhibition of 3H-thymidine incorporation of approximately 25% is not only significant but also allows for the detection of additional effects by other reagents. For cells cultured with preincubated mixtures of proteins, their theoretical percentages of thymidine incorporation (stippled bars), expected when the interactions between components have no effects on their individual functions, were calculated as the products of the percentages of thymidine incorporation by cells cultured with each component alone. It can be seen from Fig. 6 that proliferation of cells cultured in the presence of both SPARC and fibrinogen is similar to what is expected when SPARC and fibrinogen function independently, a reasonable result because SPARC and fibrinogen do not interact with each other. Although SPARC binds to both Fg D and Fg E, the binding of SPARC to Fg E significantly suppressed the inhibitory effect of SPARC on BAE cell proliferation, whereas the binding of SPARC to Fg D had little effect. Although similar dissociation constants were found for both binding pairs (Table 1), the k a and k d for the binding of SPARC to Fg E were higher than the corresponding values for Fg D, a measure that may account for the greater effect of the interaction between SPARC and Fg E on the inhibition of cell proliferation by SPARC. Further analysis of the interaction between SPARC and Fg D/Fg E, such as precise identification of the binding sites, is needed to elucidate the different effects of the two fibrinopeptides that we have reported here. In conclusion, the SPR biosensor has been used to analyze the interaction of SPARC with fibrinogen under different

H. Wang et al. / Matrix Biology 25 (2006) 20 – 26

conditions. Casein was used as a blocking agent to remove the non-specific adsorption of SPARC to the SPR chip functionalized with COOH – SAMs. SPARC binds to plasmincleaved fibrinogen, but not to native fibrinogen. SPARC is also shown to have binding sites with similar dissociation affinities on both the D and E fragments of fibrinogen. Our results indicate that plasmin-cleaved fibrinogen could be a matrix storage site for SPARC at loci of injury. SPARC might therefore participate in a positive feedback loop for the regulation of fibrinolysis through its stimulation of plasminogen and MMP activity. In addition to the elevation of SPARC in the local environment, the binding of SPARC to plasmincleaved fibrinogen could modify certain functions of SPARC. For example, the binding of SPARC to Fg E suppressed the inhibition of endothelial cell proliferation by SPARC, whereas the binding of SPARC to Fg D appeared not to affect the endothelial cell cycle. 3. Experimental procedures

25

followed by cross-linking with 0.05% glutaraldehyde for 5 min to avoid the subsequent replacement of these proteins during the experimental procedure (Chen et al., 2003). Uncoated sites were blocked with 10 mg/mL casein or BSA. Finally, SPARC (5– 30 Ag/mL) was applied for 20 min, followed by flushing with buffer for 15 min to remove non-specifically bound SPARC. A typical set of SPR response curves is shown in Fig. 2. One of the dual channels was used as a reference channel, in which buffer instead of SPARC solution was applied in the final step. Fig. 2C shows the compensated response of SPARC obtained by subtracting the response in the marked region of the reference channel (Fig. 2B) from that of the sensing channel (Fig. 2A). Using the software ClampXP Biosensor Data Analysis V. 3.50, developed by T. Morton and D. Myszka (Center for Bimolecular Interaction Analysis, University of Utah), we fitted data for the binding of SPARC to FgD and FgE (1 nm from the SPR sensor = 214.6 BIACORE RU at 760 nm) to the model for simple 1 : 1 interactions to obtain the forward (k a) and reverse (k d) rate constants. The equilibrium dissociation constant was calculated as K d = k d / k a.

3.1. Cleavage of fibrinogen by plasmin 3.4. BAE cell proliferation assay Purified human plasma fibrinogen was acquired from Enzyme Research Laboratories (South Bend, IN). Human plasma plasmin and a2-antiplasmin were from Calbiochem (San Diego, CA). Fibrinogen was digested in Tris-buffered saline at 37 -C by plasmin (6 U plasmin:1 Hmol fibrinogen) for 1.5 h. The reaction was stopped by incubation with a2antiplasmin (1.2 IU a2-antiplasmin:1 U plasmin) for another hour at 37 -C. The digestion products were examined by SDSPAGE on 4 – 20% polyacrylamide gradient gels under nonreducing conditions (Bio-Rad Laboratories, Hercules, CA). 3.2. SPR and surface functionalization The dual-channel SPR apparatus is an instrument built inhouse based on the Kretschmann configuration exploiting attenuated total reflection (Homola et al., 2001). The amount of adsorbed proteins was quantified by measuring the refractive index change-induced shift of the resonant wavelength. SPR chips were functionalized by immersion into a 1 mM ethanolic solution of 16-mercaptohexadecanoic acid [HS(CH2)15COOH, Sigma-Aldrich, St. Louis, MO] overnight, followed by sequential rinses with ethanol, ethanolic solution of acetic acid (10% v/v), and ethanol, and drying in a stream of nitrogen (Chen et al., 2003).

BAE cells were prepared from bovine aortas by our laboratory as previously described (Funk and Sage, 1991). All reagents were tested for endotoxin before addition to cells. 0.06 AM SPARC was incubated with Fg D, Fg E or fibrinogen in DMEM (Invitrogen, Carlsbad, CA) at a molar ratio of 1 : 1 overnight at 4 -C. SPARC, Fg D, Fg E and fibrinogen in amounts identical to those in the mixtures were also kept individually in DMEM overnight at 4 -C. Low-passage BAE cells were grown to confluence, allowed to quiesce for 1 day, and replated in Costar\ 24-well cell culture plates at 2.5  104 cells/well in 500 AL DMEM containing 2% FBS (Invitrogen, Carlsbad, CA). Cells were allowed to adhere for 30 min before the replacement of the original culture media with the preincubated media containing SPARC, Fg D, Fg E, fibrinogen alone, or their mixtures. 2% FBS was added to the preincubated media immediately before their addition to the cells. BAE cells in DMEM/2% FBS were included as controls. After incubation at 37 -C in 95% humidity and 7.5% CO2 for 16 –18 hours, cells were pulse-labeled with 0.55 ACi 3H-thymidine (PerkinElmer Life Sciences Inc., Boston, MA) for 4 hours. DNA precipitated in ice-cold 10% trichloroacetic acid was solubilized in 0.4 N NaOH and was subsequently measured in a liquid scintillation counter (Funk and Sage, 1991).

3.3. SPR binding assay Acknowledgements All SPR binding assays were performed in 25mM Trisbuffered saline with 125 mM NaCl and 1 mM CaCl2 (pH 7.2). Recombinant human SPARC was produced and purified as previously described (Bradshaw et al., 2000). Human plasma FgD and FgE were from Calbiochem (San Diego, CA). Casein and BSA were from Sigma-Aldrich (St. Louis, MO). First, 100 Hg/mL ligands (i.e., native or plasmin-cleaved fibrinogen, FgD or FgE) were physically adsorbed onto to the SPR chips,

This work was supported by NSF EEC-9529161 through the University of Washington Engineered Biomaterials (UWEB)-Engineering Research Center and NIH GM-40711. We thank Dr. Jiri Homola and Zuzana Manikova of Institute of Radio Engineering and Electronics, Czech Republic for the custom-built SPR sensor and their help with the analysis of data.

26

H. Wang et al. / Matrix Biology 25 (2006) 20 – 26

Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.matbio.2005.09.004. References Bini, A., Mann, K., Kudryk, B., Schoen, F., 1999. Noncollagenous bone matrix proteins, calcification, and thrombosis in carotid artery atherosclerosis. Arterioscler. Thromb. Vasc. Biol. 19, 1852 – 1861. Bootle-Wilbraham, C.A., Tazzyman, S., Marshall, J.M., Lewis, C.E., 2000. Fibrinogen E-fragment inhibits the migration and tubule formation of human dermal microvascular endothelial cells in vitro. Cancer Res. 60, 4919 – 4924. Bradshaw, D.A., Sage, E.H., 2001. SPARC, a matricellular protein that functions in cellular differentiation and tissue response to injury. J. Clin. Invest. 107, 1049 – 1054. Bradshaw, A., Bassuk, A.J., Francki, A., Sage, E.H., 2000. Expression and purification of recombinant human SPARC produced by baculovirus. Mol. Cell. Biol. Res. Commun. 3, 345 – 351. Brekken, A.R., Sage, E.H., 2000. SPARC, a matricellular protein: at the crossroads of cell-matrix. Matrix Biol. 19, 569 – 580. Chen, S., Liu, L., Zhou, J., Jiang, S., 2003. Controlling antibody orientation on charged self-assembled monolayers. Langmuir 19, 2859 – 2864. Davis, G.E., Bayless, K.J., Davis, M.J., Meininger, G.A., 2000. Regulation of tissue injury responses by the exposure of matricryptic sites within extracellular matrix molecules. Am. J. Pathol. 156, 1489 – 1498. Dejana, E., Languino, L.R., Polentarutti, N., Balconi, G., Rychewaert, J.J., Larrieu, M.J., Donati, M.B., Mantovani, A., Marguerie, G., 1985. Interaction between fibrinogen and cultured endothelial cells: induction of migration and specific binding. J. Clin. Invest. 74, 11 – 18. Drew, A.F., Liu, H., Davidson, J.M., Daugherty, C.C., Degen, J.L., 2001. Wound-healing defects in mice lacking fibrinogen. Blood 97, 3691 – 3698. Fair, D.S., Edgington, T.S., Plow, E.F., 1981. Immunochemical mapping of the conformation of human fibrinogen. J. Biol. Chem. 256, 8018 – 8023. Funk, E.S., Sage, E.H., 1991. SPARC modulates cell cycle progression in bovine aortic endothelial cells. Proc. Natl. Acad. Sci. U. S. A. 88, 2648 – 2652. Ge, M., Tang, G.X., Ryan, T.J., Malik, A.B., 1992. Fibrinogen degradation product fragment-D induces endothelial-cell detachment by activation of cell-mediated fibrinolysis. J. Clin. Invest. 90, 2508. Gollwitzer, R., Hahn, E., Nowack, H., Timpl, R., 1975. Immunochemistry of bovine fibrinogen. 1. Immunogenic activity and diversity of antigenic determinants of reduced and carboxymethylated alpha, beta-chains and gamma-chains. Immunochemistry 12, 893 – 897.

Homola, J., Lu, B.H., Nenninger, G.G., Dostalek, J., Yee, S.S., 2001. A novel multichannel surface plasmon resonance biosensor. Sens. Actuators, B 76, 403 – 410. Hunzelman, N., Hafner, M., Anders, S., Krieg, T., Nischt, R., 1998. BM-40 (osteonectin, SPARC) is expressed both in the epidermal and in the dermal compartment of adult skin. J. Invest. Dermatol. 110, 122 – 126. Kamihagi, K., Katayama, M., Ouchi, R., Kato, I., 1994. Osteonectin/SPARC regulates cellular secretion rates of fibronectin and laminin extracellular matrix protein. Biochem. Biophys. Res. Commun. 200, 423 – 428. Kelm, R.J., Swords, N.A., Orfeo, T., Mann, K.G., 1994. Osteonectin in matrix remodeling: a plasminogen – osteonectin – collagen complex. J. Biol. Chem. 269, 30147 – 30153. Lai, C., Taylor, A., Jiang, S., 2004. Comparison of bovine serum albumin and casein as blocking agents on different self-assembled monolayers surfaces using a surface plasmon resonance (SPR) sensor. J. Under Res. Bioeng. 4, 64 – 68. Lane, F.T., Sage, E.H., 1994. The biology of SPARC, a protein that modulates cell-matrix interactions. FASEB J. 8, 163 – 173. Maurer, P., Gohring, W., Sasaki, T., Mann, K., Timpl, R., Nischt, R., 1997. Recombinant and tissue-derived mouse BM-40 bind to several collagen types and have increased affinities after proteolytic activation. CMLS Cell. Mol. Life Sci. 53, 478 – 484. Puolakkainen, P., Bradshaw, D.A., Kyriakides, R.T., Reed, M., Brekken, R., Wight, T., Bornstein, P., Ratner, B., Sage, E.H., 2003. Compromised production of extracellular matrix in mice lacking secreted protein, acidic and rich in cysteine (SPARC) leads to a reduced foreign body reaction to implanted biomaterials. Am. J. Pathol. 162, 627 – 635. Raines, E.W., Lane, T.F., Iruelaarispe, M.L., Ross, R., Sage, E.N., 1992. The extracellular glycoprotein SPARC interacts with platelet-derived growth factor (PDGF)-AB and-BB and inhibits the binding of PDGF to its receptors. Proc. Natl. Acad. Sci. U. S. A. 89, 1281 – 1285. Sasaki, T., Gohring, W., Mann, K., Maurer, P., Hohenester, E., Knauper, V., Murphy, G., Timpl, R., 1997. Limited cleavage of extracellular matrix protein BM-40 by matrix metalloproteinase increases its affinity for collagens. J. Biol. Chem. 272, 9237 – 9243. Tremble, P.M., Lane, F.T., Sage, E.H., Werb, Z., 1993. SPARC, a secreted protein associated with morphologenesis and tissue remodeling, induces expression of metalloproteinases in fibroblasts though a novel extracellular matrix-dependent pathway. J. Cell Biol. 121, 1433 – 1444. Walker, B.J., Nesheim, E.M., 1999. The molecular weights, mass distribution, chain composition, and structure of soluble fibrin degradation products released from a fibrin clot perfused with plasmin. J. Biol. Chem. 274, 5201 – 5212. Wink, T., vanZuilen, S.J., Bult, A., vanBennekom, W.P., 1997. Self-assembled monolayers for biosensors. Analyst 122, R43 – R50.