Interactions between depolymerized fucosylated glycosaminoglycan and coagulation proteases or inhibitors

Thrombosis Research 146 (2016) 59–68

Contents lists available at ScienceDirect

Thrombosis Research journal homepage: www.elsevier.com/locate/thromres

Full Length Article

Interactions between depolymerized fucosylated glycosaminoglycan and coagulation proteases or inhibitors Chuang Xiao a,b, Wu Lian a,c, Lutan Zhou a,b, Na Gao a,b, Li Xu a,b, Jun Chen a,c, Mingyi Wu a,⁎, Wenlie Peng c,⁎, Jinhua Zhao a,⁎ a State Key Laboratory of Phytochemistry and Plant Resources in West China, Yunnan Key Laboratory of Natural Medicinal Chemistry, Kunming Institute of Botany, Chinese Academy of Sciences, Kunming 650201, China b University of Chinese Academy of Sciences, Beijing 100049, China c College of Life Sciences, Sun Yat-sen University, Guangzhou 510275, China

a r t i c l e

i n f o

Article history: Received 17 July 2016 Received in revised form 12 August 2016 Accepted 26 August 2016 Available online 28 August 2016 Keywords: Fucosylated glycosaminoglycan coagulation proteases biolayer interferometry binding kinetics structure-activity relationship

a b s t r a c t Fucosylated glycosaminoglycan (FG) is a structurally novel glycosaminoglycan derivative, and it has potent anticoagulant activity. Depolymerized FG (dFG) is a selective factor Xase (FXase, FIXa-FVIIIa complex) inhibitor and it has antithrombotic action without major bleeding risks. In this study, we report the effects of dFG-3 (Mw ~14 kDa) on the catalysis rates of factor IIa (FIIa), factor Xa (FXa) and factor IXa (FIXa) inhibition by antithrombin (AT), and the kinetic of the interactions between coagulation proteases or inhibitors and dFG-3 were also studied using biolayer interferometry (BLI) technology. We found that dFG-3 had much weaker catalysis activity of coagulation proteases inhibition by AT compared with heparin (UFH). The binding affinity of AT bound to dFG-3 was lower than UFH, and the UFH-AT interaction fitted well with biphasic-binding model while dFG-3-AT interaction was monophasic-binding, suggesting dFG-3 might not have allosteric activation effect on AT. The results are consistent with AT-independent inhibitory activities of dFG-3. dFG-3 could strongly bind to FIXa with much higher affinity than UFH, further explained the reason for its potent FXase inhibitory activity. Additionally, the binding ability of dFG-3 and FIXa decreased with decreasing molecular, and the fucose side chains and carboxyl groups of dFG-3 might be required for its high affinity binding with FIXa. Our data supports further the investigation of dFG-3 as a promising anticoagulant drug inhibiting the intrinsic FXase by binding to FIXa. © 2016 Elsevier Ltd. All rights reserved.

1. Introduction Venous thromboembolism is the third leading cause of cardiovascular-associated death, and anticoagulants are used to treat a wide variety of conditions that involve arterial or venous thrombosis [1]. Unfractionated heparin (UFH) and low molecular weight heparins (LMWHs) have been the clinical cornerstones of antithrombotic treatment and prophylaxis for the past 70 years, but the risk of hemorrhagic complications is still a major concern with their use [2,3].

Abbreviations: FG, fucosylated glycosaminoglycan; dFG, depolymerized fucosylated glycosaminoglycan; UFH, unfractionated heparin; LMWH, low molecular weight heparin; FXase, factor Xase; AT, antithrombin; FIIa, thrombin; FXa, factor Xa; FIXa, factor IXa; BLI, biolayer interferometry; HCII, heparin cofactor II; FVIII, factor VIII; Mw, molecular weight; Fpx, fondaparinux; FXIa, factor XIa; k1, pseudo-first order rate constant; k2, second-order rate constant. ⁎ Corresponding authors. E-mail addresses: [email protected] (M. Wu), [email protected] (W. Peng), [email protected] (J. Zhao).

http://dx.doi.org/10.1016/j.thromres.2016.08.027 0049-3848/© 2016 Elsevier Ltd. All rights reserved.

UFH has multiple potential anticoagulant mechanisms, mainly antithrombin (AT) dependent inhibition of coagulation proteases by conformational activation of serpin and template mechanisms [4,5]. In recent years, a series of studies has shown that inhibitors of the intrinsic coagulation pathway could prevent thrombosis with negligible bleeding risks [1,6–8]. Fucosylated glycosaminoglycan (FG) is a glycosaminoglycan derivate with chondroitin sulfate like backbones and fucose side chains from marine sea cucumber [9]. FG has potent anticoagulant and antithrombotic activities [9], and the mechanisms include: inhibition of factor Xa (FXa) generation by the intrinsic tenase complex, heparin cofactor II (HCII) dependent inhibition of thrombin (FIIa), AT dependent inhibition of FIIa, inhibition of factor VIII (FVIII) activation by FIIa [9–13]. Depolymerized FG (dFG) has been demonstrated to inhibit plasma thrombin generation primarily by reducing factor X activation [14], and a depolymerized FG has been found to exert an antithrombotic effect with less bleeding than UFH and LMWH in rats and dogs [15,16]. Recently, studies in our group also showed that depolymerized FG retained the anticoagulant activity and could inhibit venous thrombosis without causing side effects [13,17,18]. Thus, for the depolymerized FG

60

C. Xiao et al. / Thrombosis Research 146 (2016) 59–68

as a promising antithrombotic candidate drug, it is very valuable to study its anticoagulant mechanism related to its low bleeding tendency. In our previous study, we found that dFG had significantly reduced AT-dependent anti-FIIa activity, but its FXase inhibition activity and HCII-dependent anti-FIIa activity remain strong potency [13,19]. The anticoagulant mechanism of dFG was obviously different from that of UFH. Thus, further studies should be needed to elucidate the differences between dFG and UFH. In this study, we further investigated the interactions between coagulation proteases or inhibitors (AT, HCII) and a depolymerized FG (dFG-3, Mw ~14 kDa) and compared with UFH. We firstly reported the effect of dFG-3 on the second-order rate constants of proteases (FIIa, FXa, FIXa) inhibition by AT. And then, the kinetic and structural features of the interactions between coagulation proteases or inhibitors and dFG-3 were studied using biolayer interferometry (BLI) technology, and competitive binding assays were performed to elucidate the effects of molecular weight and chemical modifications on the interaction between dFG-3 and FIXa. 2. Methods 2.1. Materials UFH (212 USP U/mg) were purchased from Sigma (USA). LMWH (Enoxaparin, 0.4 mL × 4000 AXaIU) was from Sanofi-Aventis (France). Fondaparinux sodium (Fpx) was from GSK (UK). Human HCII, AT, thrombin, human factor IXa, human factor Xa, thrombin chromogenic substrate CS-01(38), factor IXa chromogenic substrate CS-51(09) and Heparin Anti-Xa kits were all from Hyphen Biomed (France). Human factor XIa was from Assaypro (USA). EZ-Link Amine-PEG3-Biotin and Zeba Spin desalting columns (N7 kDa) were purchased from Thermo Scientific (USA). SA and SSA biosensors were purchased from Fortebio (USA). All other chemicals were of reagent grade and obtained commercially. 2.2. Fucosylated glycosaminoglycan The native FG was extracted and purified from the body walls of the sea cucumber Thelenota ananas as previously described [20]. Depolymerized FGs (dFG-1, dFG-2, dFG-3, dFG-4, dFG-5, dFG-6) were obtained by controlled chemical depolymerization [19]. Carboxylreduced dFG-3 (dFG-a), carboxylic ethyl ester of dFG-3 (dFG-b), carboxylic benzyl ester of dFG-3 (dFG-c), carboxylic 1-butenyl ester of dFG-3 (dFG-d), partially deacetylated dFG-3 (dFG-e), and partially defucosylated FG (dFG-f) were developed as previously described [13]. 2.3. Effect of glycosaminoglycans on the rates of FIIa, FXa and FIXa inhibition by AT Reactions of AT and coagulation proteases (FIIa, FXa, FIXa) were measured in a discontinuous assay under pseudo-first order rate conditions at room temperature in TS/PEG buffer (0.02 M Tris-HCl, pH 7.4, 0.15 M NaCl, 0.1% PEG8000, 2 mM CaCl2) [21,22]. The concentrations of AT ranging from 20 to 8000 nM and were at least ten fold higher than that of protease (2 to 40 nM). AT and protease were incubated in a 20 μL reaction volume in a 96-well plate at various time intervals, then 80 μL chromogenic substrate (1 mg/mL, containing 1 mg/mL Polybrene) were added simultaneously. Additionally, 30% ethylene glycol was contained for the chromogenic substrate of FIXa. The absorbance at 405 nm was measured for 5–30 min to determine the rate of substrate cleavage (V). The pseudo-first order rate constant (k1) was determined by plotting ln (V/V0) versus time, where V0 is the rate of substrate cleavage by the protease in the absence of AT. The apparent second-order inhibition rate (k2) was determined by dividing k1 by the AT concentration. To determine the effect of glycosaminoglycans (dFG, UFH, Fpx) on the k2 of FIIa, FXa or FIXa inhibition by AT, various concentrations of

glycosaminoglycans were premixed with AT before incubation with protease. 2.4. Biotinylation of dFG-3 or UFH dFG-3 or UFH was biotinylated by using Amine-PEG3-biotin as previously described [23,24]. And the reaction mixture was desalted with the Zeba Spin desalting columns. Then the biotinylated compounds were immobilized onto the surface of SA or SSA biosensors. 2.5. BLI kinetic measurements of coagulation proteases or inhibitors binding to immobilized dFG-3 and UFH Increasing concentrations of coagulation proteases (FIIa, FIXa, FXIa) or inhibitors (AT, HCII) were allowed to interact with immobilized dFG3 and UFH. All interaction experiments were conducted at 30 °C in PBSB (0.05 M sodium phosphate at pH 7.2, 0.15 M NaCl, 0.1% BSA) or HEPES buffer (0.15 M NaCl, 20 mM HEPES, pH 7.4, 2 mM CaCl2, and 0.05% Tween 20) using an Octet Red 96 instrument (Fortebio, USA). Final volume for all the solutions was 200 μL. Assays were performed in black solid 96-well flat bottom plates with agitation set to 1000 r/min. A 600–900 s biosensor washing step was applied prior to the analysis of the association of the ligand on the biosensor to the analyte in solution for 600–1200 s. Finally the dissociation was followed for 900–1800 s. Dissociation wells were used only once to ensure buffer potency. After dissociation, the sensor surface was regenerated in 4 M NaCl or 2 M NaCl in HEPES buffer. Correction of any systematic baseline drift was done by subtracting the shift recorded for a sensor loaded with ligand but incubated with no analyte. 2.6. Solution competition BLI study To assess the relative ability of soluble UFH or dFG-3 to compete with the immobilized dFG-3 or UFH for binding to coagulation proteases or inhibitors, a competition binding assay was performed to determine their respective EC50 values. AT (1000 nM), HCII (500 nM) or FIXa (50 nM) was preincubated with increasing concentrations of UFH or dFG-3 prior to interact with immobilized dFG-3. After each run, dissociation and regeneration were performed as described above. To determine the structure-activity relationship of dFG-3-FIXa interaction, gradient concentrations of dFG-3 derivatives were preincubated with FIXa (50 nM) prior to interact with immobilized dFG-3. Association, dissociation and regeneration process were performed as described above. 2.7. Date analysis For BLI kinetic assays, data were analyzed using the Octet software version 7.0 and the binding curves were globally fitted using a 1:1 or 2:1 model [25]. For the competition assays, the response at the end of the association step for each compound concentration was plotted as the relative proportion of remaining free AT, HCII or FIXa, with response of AT, HCII or FIXa alone (no compounds) normalized to 1. The EC50 was determined by fitting the data to the following equation using the Origin 8.0 software (OriginLab, USA):

B¼

EC 50 n EC 50 n þ ½I n

ð1Þ

Where B represents the fractional specific binding, [I] represents the concentration of compounds used as a competitor, EC50 represents the concentration of compounds that causes a 50% reduction in the BLI response, and n represents the pseudo-Hill coefficient [26,27].

C. Xiao et al. / Thrombosis Research 146 (2016) 59–68

3. Results Effect of dFG-3 on the rates of FIIa, FXa and FIXa inhibition by AT. To investigate the effect of dFG-3 on the kinetics of FIIa, FXa and FIXa inhibition by AT, the rates of proteases inhibition by AT were determined under pseudo-first order conditions in the absence or presence of dFG-3. UFH and Fpx were used for comparison. The resulting second-order rate constants (k2) were calculated and plotted with the concentrations of glycosaminoglycans (Fig. 1), and the maximum k2 values in the detected concentration range were given in Table 1. The uncatalyzed rate of FIIa, FXa and FIXa inhibition by AT was 3.74 × 10 5 M− 1 min− 1, 3.37 × 105 M − 1 min− 1 and 3.25 × 103 M− 1 min− 1 respectively, consistent with previously reported data in literature [28]. The plots of the k2 of FIIa inhibition by AT versus UFH concentration was bell-shaped (Fig. 1A), which was a typical

A

9

10

dFG-3 UFH Fpx

-1

-1

k2 (M min )

8

10

7

10

6

10

5

10

-1

-1

k2 (M min )

B

C

10

9

10

8

10

7

10

6

10

5

-7

-8

-5

-6

10 10 GAG (M)

10

10

-4

10

dFG-3 UFH Fpx

0

10

-9

-7

-6

10

-7

-6

10

-8

10 10 GAG (M)

-8

10 10 GAG (M)

10

-5

10

-5

10

-4

9

10

8

10 -1

7

10

-1

k2 (M min )

-9

0 10

10

dFG-3 UFH Fpx

6

5

10

4

10

3

10

-9

0 10

10

-4

Fig. 1. Effects of dFG-3, UFH and Fpx on the second-order rate constants of AT and FIIa (A), FXa (B), and FIXa (C). Second-order rate constants (k2) were measured in a discontinuous assay under pseudo-first order rate conditions at room temperature in TS/PEG buffer. AT and protease were incubated in a 20 μL reaction volume at various time intervals in the presence of various concentrations of glycosaminoglycan. Chromogenic substrate was added to determine the residual enzyme activity. The apparent second-order inhibition rate (k2) was determined by dividing pseudo-first order rate constant (k1) by the AT concentration.

61

template mechanism of catalysis. Fpx had no effect on FIIa inhibition by AT in the detection concentration range, indicating that bridging effect made a predominate contribution to FIIa inhibition of UFH. In the presence of dFG-3, the k2 of FIIa inhibition by AT increased with the increasing concentration of dFG-3, while decreased in much higher dFG-3 concentration (~ 4 μM). The results suggest a template mechanism might also exist for dFG-3. Comparing the accelerating effect of k2, UFH increased FIIa inhibition by AT 17-fold more than dFG-3. The plots of the k2 of FXa inhibition by AT versus UFH or Fpx concentration reached a plateau (Fig. 1B), resulted from the conformational change effect of AT induced by UFH and Fpx. The maximum k2 of FXa inhibition by AT in the presence of UFH was about one order of magnitude higher than that of Fpx, in agreement with previous study [22,28]. The k2 of FXa inhibition by AT increased with the increasing concentration of dFG-3 and did not reach a plateau in the detected concentration range. Compared with UFH, the concentration of dFG-3 was 2 ~ 3 orders of magnitude higher to achieve the same k2 value. For FIXa inhibition by AT, the plot of k2 versus UFH concentration was also bell-shaped (Fig. 1C), indicating that a template mechanism of catalysis exist. The plot of k2 versus Fpx concentration yield a plateau, suggesting the conformational change effect of AT induced by Fpx. The maximum k2 value of UFH was ~ 64 higher than Fpx (Table 1). In contrast, the k2 of FIXa inhibition by AT increased with the increasing concentration of dFG-3 and did not reach a plateau at the concentration of 32 μM. Also, the concentration of dFG-3 was much higher than that of UFH to achieve the same k2 value. Interactions between dFG-3 and coagulation proteases or inhibitors using BLI measurements. In a previous study of our group, native FG had strong FXase inhibition activity and AT- or HCII-dependent anti-FIIa activity [13]. After depolymerization, dFG had significantly reduced AT-dependent anti-FIIa activity, but its FXase inhibition activity and HCII-dependent anti-FIIa activity remains [13]. To further understand the anticoagulant mechanisms of action of dFG, we studied the interactions of coagulation proteases or inhibitors with dFG-3 and UFH using BLI technology (Fig. 2). Our results showed both dFG-3 and UFH could bind coagulation proteases or inhibitors with high affinities. The binding kinetics and affinities of HCII for dFG-3 and UFH were similar, while the binding kinetics and/or affinities of AT, FIIa, FIXa, FXIa for dFG-3 and UFH were different. (See Table 2.) The UFH-AT interaction fitted well with biphasic-binding model while dFG-3-AT interaction was monophasic-binding, and the binding affinity of UFH-AT (KD1 = 5.15 × 10−8 M, KD2 = 2.72 × 10−8 M) was about 10 fold higher than that of dFG-3-AT (KD = 5.29 × 10− 7 M). The biphasic-binding model of UFH-AT interaction was consistent with the allosteric activation mechanisms of AT [5,29], and the KD values were comparative with the reported results in the literature detected by SPR [24]. The binding model of dFG-3-HCII and UFH-HCII were fitted well to 1:1 binding model, and their binding affinities were similar (KD = 8.20 × 10−8 M for dFG-3 versus KD = 6.91 × 10−8 M for UFH). The interactions of dFG-3/UFH and coagulation proteases (FIIa, FIXa, FXIa) were detected in HEPES buffer containing calcium. The binding affinities of FIIa, FIXa, FXIa for UFH were much weaker than the KD reported by Badellino using SPR method [30], this might be due to the different methods and buffers used. The interaction of UFH-FIIa fitted well with biphasic-binding model (KD1 = 5.85 × 10− 7 M, KD2 = 1.45 × 10−7 M), while dFG-3-IIa fitted well with monophasicbinding with a KD of 1.84 × 10−7 M. The dFG-3-FIXa and UFH-FIXa interactions were fitted well to 2:1 (Heterogeneous L) model, and the binding affinity of dFG-3-FIXa was much higher than that of UFH-FIXa (KD1 = 2.25 × 10− 8 M, KD2 = 5.79 × 10− 10 M for dFG-3 versus KD1 = 8.35 × 10−7 M, KD2 = 3.08 × 10−9 M for UFH). The interactions of dFG-3-FXIa and UFH-FXIa were fitted well to 2:1 binding model with strong binding affinities (KD1 = 1.82 × 10−8 M, KD2 = 4.99 × 10−9 M for dFG-3 versus KD1 = 4.21 × 10−8 M, KD2 = 1.19 × 10−8 M for UFH).

62

C. Xiao et al. / Thrombosis Research 146 (2016) 59–68

Table 1 Second-order rate constants of AT and coagulation proteases. a Second-order rate constant (k2, M−1·min−1)

FIIa FXa FIXa a

AT

dFG-3-AT

UFH-AT

Fpx-AT

3.74 ± 0.04 × 105 3.37 ± 0.10 × 105 3.25 ± 0.09 × 103

2.02 ± 0.05 × 107 9.42 ± 0.32 × 107 1.73 ± 0.05 × 106

3.48 ± 0.20 × 108 4.97 ± 0.15 × 108 7.32 ± 0.01 × 107

3.41 ± 0.08 × 105 3.79 ± 0.04 × 107 1.14 ± 0.04 × 106

Results are expressed as mean ± SE.

Solution competition assay of the interactions of AT, HCII, FIXa with dFG-3 or UFH. Our previous study showed that dFG-3 was a selective FXase inhibitor, different from UFH which had multiple activities on coagulation proteases [13]. We further compared the interactions of AT, HCII, FIXa with dFG-3 or UFH using a solution competition assay. Our results showed that UFH or dFG-3 in solution inhibited AT binding to immobilized dFG-3 on the sensor with a dose dependent fashion, and the calculated EC50 values were 3.8 nM for UFH and 8.4 nM for dFG3. UFH in solution also inhibited AT binding to immobilized UFH on the sensor with a dose dependent fashion (EC50 ~ 783 nM), but dFG-3 showed little inhibition to AT binding (EC50 above 4000 nM), suggesting that dFG-3 did not compete with UFH for binding to AT (Fig. 3). As shown in Fig. 4, solution competition BLI study also showed that UFH/dFG-3 in solution inhibit HCII binding to immobilized UFH/dFG-3 with a dose dependent fashion (EC50: 13.5 nM for UFH versus 15.8 nM for dFG-3, 28.8 nM for UFH versus 25.2 nM for dFG-3), suggesting that the binding site of UFH and dFG-3 on HCII were similar. Competitive binding results showed that UFH/dFG-3 could inhibit FIXa binding to immobilized UFH/dFG-3 in dose dependent manner (Fig. 5), but the concentration of dFG-3 which inhibited half FIXa binding was much lower than UFH (EC50: 286 nM for UFH versus 28 nM for dFG-3, 154 nM for UFH versus 26 nM for dFG-3), suggesting that dFG-3 and UFH may have similar binding site on FIXa and the binding affinity of dFG-3-FIXa was much higher than UFH-FIXa. Competition inhibition of dFG-3-FIXa interactions by FG with different Mw. It has demonstrated that FXase inhibition activity of dFG is attributed to its binding to FIXa heparin-binding site [14]. To study the structureactivity relationship of FIXa binding of FG and also the correlation between FXase inhibition and FIXa binding of FG, solution competition BLI study were performed to examine the inhibition of the binding of FIXa to immobilized dFG-3 by UFH, LMWH and different molecular weight of FG. EC50 (nM) values were measured to see the inhibition activity of these compounds to dFG-3-FIXa binding. The results were summarized in Table 3.

All these compounds could inhibit FIXa binding to immobilized dFG3 in a dose dependent fashion (Fig. 6). The compounds concentration were plotted versus the relative FIXa binding responses (Fig. 8A), and the EC50 (nM) values were calculated as described above (Table 3). The EC50 values increased with decreasing Mw, indicating that the ability of compounds binding to FIXa decreased with decreasing Mw. The EC50 (nM) values of FG, dFG-1, dFG-2, dFG-3, dFG-4, dFG-5 were lower than those of UFH and LMWH, suggesting that these compounds bound to FIXa with higher affinity relative to UFH and LMWH. In addition, the activity of each GAG compete with the binding of FIXa to immobilized dFG-3 was expressed as heparin U per mg using a parallel standard curve based on a heparin standard (212 U/mg). The results showed that when the Mw decreased from 65 kDa to 8.5 kDa (7.6-fold), the activities of FG only decreased about 2-fold, while the Mw decreased from 8.5 kDa to 5.0 kDa (1.7-fold), the acitvities of FG decreased markedly (~ 2-fold). The result indicating that the FG with Mw around 8.5 kDa (8 ~ 9 repeated trisaccharide units) represent the minimal FG to interact with FIXa with higher affinity. Competition inhibition of dFG-3-FIXa interactions by different modifications of dFG-3. Different modifications of dFG-3 were prepared as described in the methods. The results of competitive binding assays were shown in Fig. 7. In terms of EC50 (nM) values summarized in Table 4, the activity of dFG-c (carboxylic benzyl ester) was equivalent to that of dFG-3, while the activities of dFG-a (carboxyl-reduced), dFG-b (carboxylic ethyl ester), and dFG-d (carboxylic 1-butenyl ester) reduced 42-fold, 4.8fold, and 1.5-fold respectively. The partially defucosylated FG (FG-f) showed most significant effect on dFG-3-FIXa interaction since the EC50 (nM) value increased 49-fold. The activity of partially deacetylated dFG-3 (FG-e) decreased 2.6-fold, indicating that partially deacetylation had little impact on the binding of FIXa to dFG-3. 4. Discussion Our previous studies have shown that native FG has potent FXase inhibitory activity, HCII-dependent anti-IIa activity, and also AT-

Table 2 Kinetic constants of interactions between clotting proteases or inhibitors and dFG-3/UFH. a Protein AT

GAG UFH

FIIa

dFG-3 UFH dFG-3 UFH

FIXa

dFG-3 UFH

HCII

dFG-3 FXIa

UFH dFG-3

a

kon (M−1 s−1) 5

4

kon1 = 5.61 × 10 (±2.98 × 10 ); kon2 = 2.68 × 103 (±26.4) kon = 7.02 × 102 (±2.36) kon = 2.58 × 103 (±17.0) kon = 2.99 × 103 (±15.4) kon1 = 1.63 × 105 (±3.95 × 102); kon2 = 1.09 × 104 (±1.04 × 103) kon = 2.67 × 105 (±5.24 × 103) kon1 = 3.61 × 104 (±6.50 × 102); kon2 = 3.41 × 104 (±2.23 × 102) kon1 = 3.12 × 105 (±2.25 × 104); kon2 = 8.39 × 104 (±3.30 × 103) kon1 = 5.33 × 105 (±8.14 × 103); kon2 = 5.87 × 104 (±4.36 × 102) kon1 = 1.31 × 106 (±2.21 × 104); kon2 = 9.11 × 104 (±6.01 × 102)

The date with (±) in parentheses are the standard errors (SE) from the global fitting.

koff (s−1)

KD (M)

koff1 = 2.89 × 10−2 (±3.96 × 10−4); koff2 = 7.28 × 10−5 (±3.06 × 10−6) koff = 3.71 × 10−4 (±7.46 × 10−7) koff = 1.78 × 10−4 (±1.74 × 10−6) koff = 2.45 × 10−4 (±1.42 × 10−6) koff1 = 9.54 × 10−2 (±4.00 × 10−3); koff2 = 1.58 × 10−3 (±1.96 × 10−5) koff = 4.90 × 10−2 (±3.20 × 10−4) koff1 = 3.01 × 10−2 (±1.09 × 10−4); koff2 = 1.05 × 10−4 (±2.77 × 10−6) koff1 = 7.02 × 10−3 (±2.61 × 10−4); koff2 = 5.01 × 10−5 (±2.51 × 10−5) koff1 = 2.24 × 10−2 (±1.56 × 10−4); koff2 = 6.99 × 10−4 (±2.40 × 10−6) koff1 = 2.38 × 10−2 (±1.64 × 10−4); koff2 = 4.54 × 10−4 (±2.15 × 10−6)

KD1 = 5.15 × 10−8; KD2 = 2.72 × 10−8 KD = 5.29 × 10−7 KD = 6.91 × 10−8 KD = 8.20 × 10−8 KD1 = 5.85 × 10−7; KD2 = 1.45 × 10−7 KD = 1.84 × 10−7 KD1 = 8.35 × 10−7; KD2 = 3.08 × 10−9 KD1 = 2.25 × 10−8; KD2 = 5.79 × 10−10 KD1 = 4.21 × 10−8; KD2 = 1.19 × 10−8 KD1 = 1.82 × 10−8; KD2 = 4.99 × 10−9

C. Xiao et al. / Thrombosis Research 146 (2016) 59–68 Table 3 The EC50 values of FG/dFGs to inhibit FIXa binding to immobilized dFG-3. Compd.

Mw (kDa)

FXase a ng/mL

U/mg

nM

U/mg

FG dFG-1 dFG-2 dFG-3 dFG-4 dFG-5 dFG-6 LMWH UFH

65.8 42.6 23.4 14.0 12.7 8.55 4.96 ~4.50 ~18.0

13.3 12.1 11.5 11.9 11.4 13.1 30.3 52.7 18.2

290 319 336 324 338 295 127 73 212

4.7 ± 0.6 6.2 ± 0.4 16.6 ± 2.0 26.2 ± 1.3 34.0 ± 1.1 71.3 ± 3.0 230.7 ± 7.6 1558.3 ± 134.6 154.2 ± 11.7

1903 2228 1515 1604 1363 965 514 84 212

Competitive binding to FIXa

b

a Data from [13]. The concentration of each compound required to inhibit 50% of FXase activity. b The concentration of each compound required to inhibit 50% FIXa binding to immobilized dFG-3. Results are expressed as EC50 ± SE or as heparin U per mg using a parallel standard curve based on a heparin standard (212 U/mg).

dependent anti-IIa activity [13]. Depolymerized FG (dFG) retains its FXase inhibitory activity and HCII-dependent anti-IIa activity, but has reduced AT-dependent anti-IIa activity [13,31]. AT has a relatively high concentration (~ 2.3 μM) in human plasma [32], and it inhibits the activity of coagulation protease by forming a stable complex [5,33]. UFH primarily accelerates the inactivation of FIIa, FXa and FIXa by AT to exert its anticoagulant activity [22,28]. In this work, we studied the effect of dFG-3 (Mw ~ 14 kDa) on the catalysis rates of FIIa, FXa and FIXa inhibition by AT using UFH and Fpx for comparison. We found that the maximum acceleration effect of FIIa inhibition by AT of dFG-3 was ~ 17-fold less than that of UFH, and kinetic analyses suggest a template mechanism may exist for FIIa inhibition by AT of dFG-3. Recently, computer docking studies also suggest the possible formation of the ternary complex between FIIa, AT and FG [34]. For FXa and FIXa inhibition by AT, although dFG-3 could increase the second-order rate constants (k 2 ) with increasing concentration, the concentrations of dFG-3 were much higher than that of UFH to achieve the same k 2 value. These results further confirm the AT independent activity of dFG-3 in our recently published results [18]. To further demonstrate the different mechanisms of action of dFG-3 and UFH, the interactions between coagulation proteases or inhibitors and dFG-3/UFH were studied using BLI technology. The binding affinity of dFG-3-AT (KD = 5.29 × 10− 7 M) was about 10fold weaker than that of UFH-AT, and the dFG-3-AT interaction fitted well with monophasic-binding model, indicating that dFG-3 might

Table 4 The EC50 values of FG derivatives to inhibit FIXa binding to immobilized dFG-3. Compd.

dFG-3 dFG-a dFG-b dFG-c dFG-d dFG-e dFG-f

Chemical modification

Depolymerization Carboxyl reduction (100%) Ethyl ester (100%) Benzyl eater (60%) 1-butenyl ester (40%) N-deacetylation (49%) Defucosylation (64%)

Mw (kDa)

14.0 11.6 11.6 16.0 15.4 11.3 13.1

FXase a ng/mL

24.50 104.04 52.24 68.96 25.48 20.57 947.83

Competitive binding to FIXa b nM

U/mg

26.2 ± 1.3 1094.8 ± 95.5 125.2 ± 8.3 26.3 ± 1.9 40.1 ± 2.9 67.7 ± 6.1 1277.2 ± 22.2

1604 46 405 1398 953 769 35

a Data from [13]. The concentration of each compound required to inhibit 50% of FXase activity. b The concentration of each compound required to inhibit 50% FIXa binding to immobilized dFG-3. Results are expressed as EC50 ± SE or as heparin U per mg using a parallel standard curve based on a heparin standard (212 U/mg).

63

not produce conformational activation on AT. A previous study by Minamiguchi had not detected the binding of AT to depolymerized holothurian glycosaminoglycan (DHG) using a DHG-cellulofine column [35], possibly due to their weak binding affinity. Our results showed that dFG-3 and UFH have the same binding model and similar binding affinities with HCII, which was consistent with the HCII dependent anti-IIa activities of dFG-3 and UFH in our previous study [13]. However, FIIa, FIXa or FXIa bound to dFG-3 with relatively higher affinity than that of UFH, especially for the interaction of dFG-3-FIXa (K D 1 = 2.25 × 10 − 8 M, K D 2 = 5.79 × 10 − 10 M for dFG-3 versus KD1 = 8.35 × 10− 7 M, KD 2 = 3.08 × 10− 9 M for UFH), which was consistent with the stronger FXase inhibitory activity of dFG-3 (EC50: 0.85 nM) than UFH (EC50: 1.01 nM) [13]. Earlier studies have also shown that DHG had strong affinity for FIXa and FIIa by using DHG column [10,35]. The binding of FIXa, FXIa to dFG-3 both fitted well with biphasic-binding model which were similar to those of UFH, while the binding of FIIa to dFG-3 fitted well with monophasic-binding model and was different from that of UFH. Further studies were needed to elucidate the difference. We compared the interactions of AT, HCII, FIXa with dFG-3 or UFH by a solution competition assay (Fig. 2-4). dFG-3 did not compete with UFH for binding to AT, suggesting that dFG-3 and UFH have different binding sites on AT and this may explain why dFG-3 fails to induce a conformational change in AT. The EC50 values of soluble UFH to compete with immobilized dFG-3 or UFH to bind HCII were comparative with that of dFG-3, indicating that dFG-3 and UFH bound to HCII with similar affinity and they might have similar binding sites on HCII. Competitive binding assays also showed that dFG-3 and UFH had similar binding sites on FIXa, but the binding affinity of FIXa and dFG-3 was 6 ~ 10-fold higher than that of UFH, since the EC50 values of soluble UFH to compete with immobilized dFG-3 or UFH were 6 ~ 10-fold higher than that of dFG-3. Considering that the apparent affinity of protease binding to UFH will increase with chain length [36], we suspect that the intrinsic FIXa binding affinity of dFG-3 will be much higher than UFH since UFH has much longer chain length than dFG-3. Our results were in agreement with the studies of DHG which is thought to bind with FIXa heparin-binding exosite and interfere the binding of FVIIIa and FIXa [26]. To assess the ability of dFG, UFH, and LMWH to interact with FIXa, soluble GAGs to compete with the binding of factor IXa to immobilized dFG-3 were examined by a solution competition assay. GAGs inhibit the FIXa binding to immobilized dFG-3 in a dose depended manner, the EC50 values were calculated (see Materials and Methods) and were summarized in Table 3. The EC 50 (nM) values increase with the decreasing Mw, indicating that the ability of dFG binding to FIXa decreases with the decreasing Mw. The EC 50 (nM) values of dFG are all lower than that of LMWH, which is consistent with the previous report that DHG bound to FIXa with significantly higher affinity than LMWH [26]. Furthermore, our result indicating that the FG with Mw around 8.5 kDa represent the minimal FG to interact with FIXa with higher affinity, this result is consistent with the result observed in the abilities of FG inhibiting the FXase [13]. Nonspecific binding might also exist and contribute to the apparent affinity of FIXa and dFG with higher Mw, therefore, further studies will be needed to distinguish the “site specific” and “nonspecific” binding of FIXa to dFG using lower Mw dFG or dFG oligosaccharides. The binding ability of dFG-3 with FIXa was affected by chemical modifications of functional group of dFG-3. In terms of EC50 (heparin U) values, the partially defucosylated FG (FG-f) and carboxyl-reduced dFG-3 (FG-a) showed most significant effect on dFG-3-FIXa interaction, and the effect of carboxylic ethyl ester was also very notable, while other modifications had less effect on dFG-3-FIXa interaction, indicating that carboxyl groups and fucose side chains

64

C. Xiao et al. / Thrombosis Research 146 (2016) 59–68

4000 2000 1000 500 250 125

0.3

0.0 0

600

1200 1800 Time (s)

AT (nM)

0.3

4000 2000 1000 500 250 125

0.0 0

0.21

0.14

1000 500 250 125 62.5 31.3

0.07

0.15

0

0.10 0.05

600

1200 1800 Time (s)

0

2400

2.4

0.9

1.6 0.8 0.0 0

300 600 Time (s)

FIXa (nM) 400 200 100 50 25 12.5 900

G

Response (nm)

F 1.2

Response (nm)

E 3.2

600

1200 Time (s)

1800

FIXa (nM) 250 125 62.5 31.3 15.6 7.81

0.6 0.3 0.0 0

300 600 Time (s)

900

H 0.4

0.3 IIa (nM) 400 200 100 50 12.5

0.2 0.1 0.0 0

300 600 Time (s)

Response (nm)

0.4

0.8

0.4 0.2 0.0 0

300

600 900 Time (s)

XIa (nM) 100 50 25 12.5 6.25 1200

IIa (nM) 400 200 100 50 25 12.5

0.2 0.1 0.0 0

J

1.0

0.6

0.3

900

Response (nm)

Response (nm)

1000 500 250 125 62.5 31.3 15.6

0.00

0.00

Response (nm)

1200

HCII (nM)

HCII (nM)

I

600 Time (s)

D 0.20

C Response (nm)

0.6

2400

Response (nm)

Response (nm)

AT (nM)

0.6

Response (nm)

B 0.9

A 0.9

300 600 Time (s)

900

1.2 1.0 0.8 0.6 0.4 0.2 0.0 0

300

600 900 Time (s)

XIa (nM) 100 50 25 12.5 6.25 1200

Fig. 2. BLI sensorgram. Gradient concentrations of AT (4000 nM, 2000 nM, 1000 nM, 500 nM, 250 nM, 125 nM) interact with dFG-3 (A) and UFH (B). Gradient concentrations of HCII (1000 nM, 500 nM, 250 nM, 125 nM, 62.5 nM, 31.25 nM) interact with dFG-3 (C) and UFH (D). Gradient concentrations of FIXa (400 nM, 200 nM, 100 nM, 50 nM, 25 nM, 12.5 nM) interact with dFG-3 (E) and UFH (F). Gradient concentrations of FIIa (400 nM, 200 nM, 100 nM, 50 nM, 12.5 nM) interact with dFG-3 (G) and UFH (H). Gradient concentrations of FXIa (100 nM, 50 nM, 25 nM, 12.5 nM, 6.25 nM) interact with dFG-3 (I) and UFH (J).

C. Xiao et al. / Thrombosis Research 146 (2016) 59–68

B

A 0 2 4 8 16 32 128

UFH (nM)

0.12

0.06

0.24

0.00 0

600

1200 1800 Time (s)

0 0.5 1 2 4 8 32

0.18 0.12 0.06 0.00

2400

C 0.45

0

600

1200 1800 Time (s)

2400

UFH (nM) 0 62.5 250 500 1000 2000 4000

0.27 0.18 0.09 0.00 0

600 1200 Time (s)

Response (nm)

D 0.4

0.36 Response (nm)

0.30

dFG-3 (nM) Response (nm)

Response (nm)

0.18

65

dFG-3 (nM) 0 62.5 250 500 1000 2000 4000

0.3 0.2 0.1 0.0

1800

0

600 1200 Time (s)

1800

Fig. 3. Competition of AT binding to immobilized dFG-3 or UFH by soluble dFG-3 and UFH. Increasing concentrations of dFG-3 (A), UFH (B) were preincubated with 1000 nM AT prior to interact with immobilized dFG-3. Increasing concentrations of UFH (C), dFG-3 (D) were preincubated with 1000 nM AT prior to interact with immobilized UFH.

are required for the high affinity binding of FIXa to dFG-3. The results are consistent with our previous study that free carboxyl and fully fucosylated GlcA may account for potent anti-FXase activity [13,37]. FGs from various species of the sea cucumber have the same chondroitin sulfate like backbone and sultated fucosyl branching units linked to GlcA in a α1,3 manner [9,18]. The sulfation patterns of these branches vary accordingly with sea cucumber species [9].

We have detected the FXase inhibitory activities of FGs from different sea cucumber species in a previous research. These FGs and their depolymerized products have potent and similar antiFXase activity [13]. We suspect that these FGs/dFGs may also have strong binding affinities with FIXa according to our results. Further research is still needed to elucidate the effect of sulfation pattens of fucose side chains on the binding affinities between FGs/dFGs and FIXa.

B

A 0.24 0.18 0 4 8 16 32 128

0.12 0.06

UFH (nM)

0.12 Response (nm)

Response (nm)

dFG-3 (nM)

0 2 6 18 54 162

0.06

0.00

0.00 600

1200 1800 Time (s)

UFH (nM)

C

0

2400

600

1200 1800 Time (s)

2400

D

0.8

dFG-3 (nM)

Response (nm)

0

0.6

0 2 6 18 54 162 486

Response (nm)

0.6 0 2 6 18 54 162 486

0.4

0.2

0.4 0.2 0.0

0.0 0

600

1200 Time (s)

1800

0

600

1200 Time (s)

1800

Fig. 4. Competition of HCII binding to immobilized dFG-3 or UFH by soluble dFG-3 and UFH. Increasing concentrations of dFG-3 (A), UFH (B) were preincubated with 500 nM HCII prior to interact with immobilized dFG-3. Increasing concentrations of UFH (C), dFG-3 (D) were preincubated with 500 nM HCII prior to interact with immobilized UFH.

66

C. Xiao et al. / Thrombosis Research 146 (2016) 59–68

B 0.30 0 4 16 32 64 128 512

0.18 0.12 0.06 0.00 0

C

300 600 Time (s)

UFH (nM)

0.24

0 12 48 96 192 384 1536

0.18 0.12 0.06 0.00

900

0

300 600 Time (s)

900

D 0.30

0.30 UFH (nM) 0 40 80 160 320 640 1280

0.24 Response (nm)

Response (nm)

dFG-3 (nM)

0.24

0.18 0.12 0.06

Respone (nm)

Response (nm)

A 0.30

0.24

dFG-3 (nM) 0 4 16 32 64 128 512

0.18 0.12 0.06 0.00

0.00 0

300 600 Time (s)

0

900

300 600 Time (s)

900

Fig. 5. Competition of FIXa binding to immobilized dFG-3 or UFH by soluble dFG-3 and UFH. Increasing concentrations of dFG-3 (A), UFH (B) were preincubated with 50 nM FIXa prior to interact with immobilized dFG-3. Increasing concentrations of UFH (C), dFG-3 (D) were preincubated with 50 nM FIXa prior to interact with immobilized UFH.

0.24

0.24

0.24

0.06 0.00 0

300 600 Time (s)

(nM) 0 1.5 3 6 12 24 48

0.18 0.12 0.06 0.00 0

300 600 Time (s)

dFG-2

0.24 0.18 0.12 0.06 0.00 0

300 600 Time (s)

(nM) 0 4 16 32 64 128 512 900

0.12 0.06 0.00 0

300 600 Time (s)

(nM) 0 3 6 12 24 48 144 900

0.24

dFG-4

0.24 0.18 0.12 0.06 0.00 0

300 600 Time (s)

(nM) 0 4 16 32 64 128 512 900

dFG-5 0.18 Response (nm)

dFG-3

0.18

900

0.30

0.30 Response (nm)

dFG-1 Response (nm)

0.12

(nM) 0 1 3 4 8 16 64 900

Response (nm)

0.18

Response (nm)

Response (nm)

FG

0.12 0.06 0.00 0

300 600 Time (s)

(nM) 0 15 45 90 180 360 1080 900

0.24

0.18 0.12 0.06 0.00 0

300 600 Time (s)

(nM) 0 25 100 200 400 800 3200 900

0.30

UFH

0.24

(nM) 0 12 48 96 192 384 1536

0.18 0.12 0.06 0.00 0

300 600 Time (s)

900

LMWH Response (nm)

dFG-6 Response (nm)

Response (nm)

0.24

(nM) 0 200 400 800 1600 3200 6400

0.18 0.12 0.06 0.00 0

300 600 Time (s)

900

Fig. 6. Competition of FIXa binding to immobilized dFG-3 by soluble FG, dFG, UFH and LMWH. Increasing concentration of FG, dFG-1, dFG-2, dFG-3, dFG-4, dFG-5, dFG-6, UFH, LMWH were preincubated with 50 nM FIXa prior to interact with immobilized dFG-3.

C. Xiao et al. / Thrombosis Research 146 (2016) 59–68

0.5

0.28

dFG-a

0.07 0.00 0

300 600 Time (s)

Response (nm)

0.14

dFG-b

0.4

(nM) 0 150 450 900 1800 3600 10800

0.21 Response (nm)

67

(nM) 0 4 16 32 64 128 512

0.3 0.2 0.1 0.0 0

900

300 600 Time (s)

900

0.24

dFG-c (nM) 0 6 12 24 48 96 192

0.18 0.12 0.06

dFG-d

0.00 0

300 600 Time (s)

0.12 0.06 0.00

900

0

300 600 Time (s)

900

0.24

0.25

dFG-f

dFG-e

0.20 0.15 0.10 0.05 0.00 0

300 600 Time (s)

(nM) 0 150 450 1350 2700 4050 12150

0.18

(nM) 0 16 32 64 128 256 512

Response (nm)

Response (nm)

(nM) 0 12 24 36 48 96 384

0.18 Response (nm)

Response (nm)

0.24

0.12 0.06 0.00

900

0

300 600 Time (s)

900

Fig. 7. Competition of FIXa binding to immobilized dFG-3 by soluble dFG derivatives. Increasing concentration of dFG-a, dFG-b, dFG-c, dFG-d, dFG-e, dFG-f were preincubated with 50 nM FIXa prior to interact with immobilized dFG-3.

In conclusion, we have determined the effects of dFG-3 on the rates of FIIa, FXa and FIXa inhibition by AT and also its binding affinities with coagulation proteases or inhibitors, and these results further support dFG-3 as a selective FXase inhibitor and may help explain the lower bleeding tendency of dFG-3 compared with heparin. Competitive binding assay suggests that dFG-3 and heparin have similar binding site on FIXa. The binding ability of dFG-3 and FIXa decreased with decreasing molecular, and the fucose side chains and carboxyl groups of dFG-3 significantly affect its binding with FIXa.

The authors declare no conflicts of interest. Acknowledgements This work was supported in part by grants from the Yunnan Provincial Science and Technology Department in China (no. 2010CI116, 2012FB177 and 2013FA046), the National Natural Science Foundation of China (no. 81102372, 81673330 and 81373292), Outstanding

B

1.0

Fractional Factor IXa Binding

Fractional Factor IXa Binding

A

Conflict of interest statement

0.8 0.6 0.4 0.2 0.0 -1 10

FG dFG-1 dFG-2 dFG-3 dFG-4 dFG-5 dFG-6 UFH LMWH

10

0

10

1

10

2

10

Glycosaminoglycan (nM)

3

10

4

1.0 0.8 0.6 0.4 0.2 0.0 -1 10

dFG-3 dFG-a dFG-b dFG-c dFG-d dFG-e dFG-f UFH LMWH 10

0

10

1

10

2

10

3

10

4

Glycosaminoglycan (nM)

Fig. 8. Competition of FIXa binding to immobilized dFG-3 by soluble FG, dFG and its derivatives. The concentration of glycosaminoglycan was plotted versus the proportion of remaining free FIXa and fit by nonlinear regression to obtain the EC50 for competition by the glycosaminoglycan.

68

C. Xiao et al. / Thrombosis Research 146 (2016) 59–68

Technical Talent Foundation of the Chinese Academy of Sciences and West Light Foundation of the Chinese Academy of Sciences. References [1] N. Mackman, Triggers, targets and treatments for thrombosis, Nature 451 (2008) 914–918. [2] S. Schulman, R.J. Beyth, C. Kearon, M.N. Levine, Hemorrhagic complications of anticoagulant and thrombolytic treatment, Chest 133 (2008) 257S–298S. [3] J.W. Eikelboom, J.I. Weitz, New anticoagulants, Circulation 121 (2010) 1523–1532. [4] W. Li, D.J. Johnson, C.T. Esmon, J.A. Huntington, Structure of the antithrombinthrombin-heparin ternary complex reveals the antithrombotic mechanism of heparin, Nat. Struct. Mol. Biol. 11 (2004) 857–862. [5] S.T. Olson, B. Richard, G. Izaguirre, S. Schedin-Weiss, P.G. Gettins, Molecular mechanisms of antithrombin-heparin regulation of blood clotting proteinases. A paradigm for understanding proteinase regulation by serpin family protein proteinase inhibitors, Biochimie 92 (2010) 1587–1596. [6] R.W. Colman, Are hemostasis and thrombosis two sides of the same coin? J. Exp. Med. 203 (2006) 493–495. [7] T. Renne, Safe(r) anticoagulation, Blood 116 (2010) 4390–4391. [8] J.E. Geddings, N. Mackman, New players in haemostasis and thrombosis, Thromb. Haemost. 111 (2014) 570–574. [9] V.H. Pomin, Holothurian fucosylated chondroitin sulfate, Mar. Drugs 12 (2014) 232–254. [10] H. Nagase, K. Enjyoji, K. Minamiguchi, K.T. Kitazato, K. Kitazato, H. Saito, H. Kato, Depolymerized holothurian glycosaminoglycan with novel anticoagulant actions: Antithrombin III- and heparin cofactor II-independent inhibition of factor X activation by factor IXa-factor VIIIa complex and heparin cofactor II-dependent inhibition of thrombin, Blood 85 (1995) 1527–1534. [11] H. Nagase, K. Enjyoji, M. Shima, K. Kitazato, A. Yoshioka, H. Saito, H. Kato, Effect of depolymerized holothurian glycosaminoglycan (DHG) on the activation of factor VIII and factor V by thrombin, J. Biochem. 119 (1996) 63–69. [12] B.F. Glauser, M.S. Pereira, R.Q. Monteiro, P.A. Mourao, Serpin-independent anticoagulant activity of a fucosylated chondroitin sulfate, Thromb. Haemost. 100 (2008) 420–428. [13] M. Wu, D. Wen, N. Gao, C. Xiao, L. Yang, L. Xu, W. Lian, W. Peng, J. Jiang, J. Zhao, Anticoagulant and antithrombotic evaluation of native fucosylated chondroitin sulfates and their derivatives as selective inhibitors of intrinsic factor Xase, Eur. J. Med. Chem. 92 (2015) 257–269. [14] Y. Buyue, J.P. Sheehan, Fucosylated chondroitin sulfate inhibits plasma thrombin generation via targeting of the factor IXa heparin-binding exosite, Blood 114 (2009) 3092–3100. [15] K. Kitazato, K.T. Kitazato, H. Nagase, K. Minamiguchi, DHG, a new depolymerized holothurian glycosaminoglycan, exerts an antithrombotic effect with less bleeding than unfractionated or low molecular weight heparin, in rats, Thromb. Res. 84 (1996) 111–120. [16] K. Kitazato, K.T. Kitazato, E. Sasaki, K. Minamiguchi, H. Nagase, Prolonged bleeding time induced by anticoagulant glycosaminoglycans in dogs is associated with the inhibition of thrombin-induced platelet aggregation, Thromb. Res. 112 (2003) 83–91. [17] N. Gao, F. Lu, C. Xiao, L. Yang, J. Chen, K. Zhou, D. Wen, Z. Li, M. Wu, J. Jiang, G. Liu, J. Zhao, beta-Eliminative depolymerization of the fucosylated chondroitin sulfate and anticoagulant activities of resulting fragments, Carbohydr. Polym. 127 (2015) 427–437. [18] L. Zhao, M. Wu, C. Xiao, L. Yang, L. Zhou, N. Gao, Z. Li, J. Chen, J. Chen, J. Liu, H. Qin, J. Zhao, Discovery of an intrinsic tenase complex inhibitor: Pure nonasaccharide from fucosylated glycosaminoglycan, Proc. Natl. Acad. Sci. U. S. A. 112 (2015) 8284–8289.

[19] M. Wu, S. Xua, J. Zhao, H. Kang, H. Ding, Physicochemical characteristics and anticoagulant activities of low molecular weight fractions by free-radical depolymerization of a fucosylated chondroitin sulphate from sea cucumber Thelenata ananas, Food Chem. 122 (2010) 716–723. [20] M. Wu, S. Xu, J. Zhao, H. Kang, H. Ding, Preparation and characterization of molecular weight fractions of glycosaminoglycan from sea cucumber Thelenata ananas using free radical depolymerization, Carbohydr. Res. 345 (2010) 649–655. [21] S.T. Olson, I. Bjork, J.D. Shore, Kinetic characterization of heparin-catalyzed and uncatalyzed inhibition of blood coagulation proteinases by antithrombin, Methods Enzymol. 222 (1993) 525–559. [22] E.M. Wiebe, A.R. Stafford, J.C. Fredenburgh, J.I. Weitz, Mechanism of catalysis of inhibition of factor IXa by antithrombin in the presence of heparin or pentasaccharide, J. Biol. Chem. 278 (2003) 35767–35774. [23] B. Li, J. Suwan, J.G. Martin, F. Zhang, Z. Zhang, D. Hoppensteadt, M. Clark, J. Fareed, R.J. Linhardt, Oversulfated chondroitin sulfate interaction with heparin-binding proteins: new insights into adverse reactions from contaminated heparins, Biochem. Pharmacol. 78 (2009) 292–300. [24] S. Cochran, C.P. Li, V. Ferro, A surface plasmon resonance-based solution affinity assay for heparan sulfate-binding proteins, Glycoconj. J. 26 (2009) 577–587. [25] M. Figuera-Losada, P.V. LoGrasso, Enzyme kinetics and interaction studies for human JNK1β1 and substrates activating transcription factor 2 (ATF2) and c-Jun N-terminal kinase (c-Jun), J. Biol. Chem. 287 (2012) 13291–13302. [26] J.P. Sheehan, E.N. Walke, Depolymerized holothurian glycosaminoglycan and heparin inhibit the intrinsic tenase complex by a common antithrombin-independent mechanism, Blood 107 (2006) 3876–3882. [27] W. Lian, M. Wu, N. Huang, N. Gao, C. Xiao, Z. Li, Z. Zhang, Y. Zheng, W. Peng, J. Zhao, Anti-HIV-1 activity and structure-activity-relationship study of a fucosylated glycosaminoglycan from an echinoderm by targeting the conserved CD4 induced epitope, Biochim. Biophys. Acta 1830 (2013) 4681–4691. [28] S.T. Olson, R. Swanson, E. Raub-Segall, J. Bedsted, M. Sadri, M. Petitou, J.P. Herault, J.M. Herbert, I. Bjork, Accelerating ability of synthetic oligosaccharides on antithrombin inhibition of proteinases of the clotting and fibrinolytic systems - Comparison with heparin and low-molecular-weight heparin, Thromb. Haemost. 92 (2004) 929–939. [29] I. Capila, R.J. Linhardt, Heparin-protein interactions, Angew. Chem. Int. Ed. 41 (2002) 391–412. [30] K.O. Badellino, P.N. Walsh, Localization of a heparin binding site in the catalytic domain of factor XIa, Biochemistry 40 (2001) 7569–7580. [31] L. Zhao, S. Lai, R. Huang, M. Wu, N. Gao, L. Xu, H. Qin, W. Peng, J. Zhao, Structure and anticoagulant activity of fucosylated glycosaminoglycan degraded by deaminative cleavage, Carbohydr. Polym. 98 (2013) 1514–1523. [32] N.S. Quinsey, A.L. Greedy, S.P. Bottomley, J.C. Whisstock, R.N. Pike, Antithrombin: in control of coagulation, Int. J. Biochem. Cell Biol. 36 (2004) 386–389. [33] E. Holmer, K. Kurachi, G. Soderstrom, The molecular-weight dependence of the rateenhancing effect of heparin on the inhibition of thrombin, factor Xa, factor IXa, factor XIa, factor XIIa and kallikrein by antithrombin, Biochem. J. 193 (1981) 395–400. [34] A.G. Gerbst, N.E. Ustyuzhanina, N.E. Nifantiev, Computational study of the possible formation of the ternary complex between thrombin, antithrombin III and fucosylated chondroitin sulfates, Mendeleev Commun. 25 (2015) 420–421. [35] K. Minamiguchi, H. Nagase, K.T. Kitazato, K. Kitazato, Interaction of a new depolymerized holothurian glycosaminoglycan with proteins in human plasma, Thromb. Res. 83 (1996) 253–264. [36] S.T. Olson, H.R. Halvorson, I. Bjork, Quantitative characterization of the thrombinheparin interaction discrimination between specific and nonspecific binding models, J. Biol. Chem. 266 (1991) 6342–6352. [37] M. Wu, R. Huang, D. Wen, N. Gao, J. He, Z. Li, J. Zhao, Structure and effect of sulfated fucose branches on anticoagulant activity of the fucosylated chondroitin sulfate from sea cucumber Thelenata ananas, Carbohydr. Polym. 87 (2012) 862–868.