The Mechanism of Inhibition of Factor III (Thromboplastin) Activity by Apolipoprotein B-100
Abstract Factor III (thromboplastin) activity is inhibited by apoB-100, but the mechanism of inhibition is unknown. By examining the effect of purified apoB-100 on factor III activity, we showed that apoB-100 can inhibit factor III via a different mechanism from that caused by the tissue-factor pathway–inhibitor, which is mainly carried on the surface of lipoproteins. Although the presence of calcium ions and factors X and VII may enhance the rate of inhibition, they are not a prerequisite for the inhibition of factor III by apoB-100. In addition, by investigating the changes in the UV spectra of apoB-100 on interaction with factor III and factors X and VII and by assigning the shifts in absorption spectra to particular amino acids, we showed that these interactions involve negative and positive residues within these proteins. By following the rates of interactions between apoB-100 and either factors III, X, or VII, a two-step mechanism for the inhibition process involving factors X and VII was postulated. In this mechanism, the primary interaction of apoB-100 with factor III is followed by a rate-limiting step that can be accelerated by the presence of either factor X or VII and leads to the inhibition of factor III. Furthermore, a computer-based analysis of the sequences of factor III revealed a possible binding site for apoB-100.
- Received September 11, 1995.
- Accepted February 6, 1996.
The exposure of FIII upon injury to endothelium initiates the extrinsic pathway of coagulation and leads to formation of a blood clot on the injured surface.1 2 3 Following clot formation the procoagulant effect of FIII is restrained by a number of circulating inhibitors within the serum. Previously, the inhibitory action of LDL was thought to be derived solely from an associated protein known as TFPI (also known as lipoprotein-associated coagulation inhibitor), which requires the presence of activated clotting factors X and VII and calcium ions for its inhibitory effect.4 5 6
Recently, it was demonstrated that LDL7 8 9 10 and specifically the protein component of LDL (apoB-100) was capable of inhibiting the ability of FIII to initiate the extrinsic coagulation pathway.11 Studies with other FIII inhibitors, such as TFPI,4 5 6 apoA-II,12 and placental anticoagulant protein,13 have shown that certain conditions must be met for FIII-inhibitor complexes to form. These conditions include the presence of activated clotting factors X and VII and calcium ions as well as thermodynamic criteria.5 14 In addition, the inhibition of FIII activity by TFPI seems to occur at a faster rate than that by apoB-100.7 8 9 10 The aim of this work was to elucidate the mechanism of FIII inhibition by apoB-100 in the absence of TFPI.
Isolation and Reconstitution of ApoB-100 and FIII
ApoB-100 was isolated to homogeneity10 11 and reconstituted (1.9 μmol/L) in a microemulsion of soybean phosphatidylcholine (1 g/L) (Sigma Chemical Co) and dispersed in sodium deoxycholate (6 mmol/L) (Calbiochem Novabiochem).10 15 16 Commercial rabbit brain FIII (Sigma) and recombinant rabbit FIII (Gamidor Ltd) were used throughout this work so that any procoagulant activity measured would be standardized, as the activity of FIII varies among differing preparations. On the other hand, in some experiments, the presence of impurities made the use of crude FIII questionable. In these experiments, the assays were performed with rabbit brain FIII and recombinant rabbit FIII and then repeated with affinity-purified rabbit brain FIII. The purification of FIII from the crude rabbit brain extract was performed by using monoclonal antibodies to FIII (Calbiochem) coupled to p-nitrophenyl chloroformate–activated agarose (Sigma).17 The purified FIII exhibited a single band (47 kD) on 12% denaturing polyacrylamide gel electrophoresis. However, the yield of the purified protein was not sufficient for measurement of FIII activity and hence was used for spectroscopic measurements. The rFIII had the advantage of purity at measurable quantities but the disadvantage of incomplete glycosylation and posttranslational modifications, which may be necessary for ideal activity. The activity of FIII and the percentage inhibition were calculated according to an arbitrary scale.10 11
Measurement of Inhibition of FIII by ApoB-100
Assays to measure inhibition of FIII activity were performed as described below, and the results were interpreted.10 11 Samples (1 mL) of purified apoB-100 (0.95 μmol/L), reconstituted in soybean phosphatidylcholine (final concentration, 0.5 g/L) and dispersed in sodium deoxycholate (6 mmol/L), were incubated with equal volumes of rabbit brain FIII or rFIII (final concentration, 50×103 U/L) and CaCl2 solution (final concentration, 0.25 mmol/L) at 37°C either in the absence of additional factors or the presence of FX and FVII (final concentrations for both, 10 U/L) (Sigma). At these concentrations, FX and FVII had no significant effect on the assay method used. The study was repeated in the presence of FVII (10 U/L) only, FX (10 U/L) only, and finally at 4°C in the absence of other factors. Samples were removed from the incubation at intervals over a period of 180 minutes, and residual FIII activity was measured by means of the one-stage prothrombin time assay. Deoxycholate had no effect on FIII activity at this concentration.18 The percent inhibition in each case was calculated as the reduction in activity divided by the initial activity multiplied by 100.
Samples (1 mL) of purified apoB-100 (final concentration, 0.95 μmol/L), reconstituted in soybean phosphatidylcholine (final concentration, 0.5 g/L) and dispersed in sodium deoxycholate (6 mmol/L), were incubated with FX or FVII (final concentration, 10 U/L) together with the CaCl2 solution (final concentration, 0.25 mmol/L) at 37°C for 60 minutes before the addition of rabbit brain FIII or rFIII (final concentration, 50×103 U/L). The incubation was continued and samples were removed at intervals over a period of 180 minutes and assayed as described above. In addition, samples of FIII (1 mL) were preincubated at 37°C for 60 minutes with the same amounts of either FX or FVII before the addition of apoB-100 and monitored for any inhibition over 180 minutes.
Agarose Gel Electrophoresis of the FIII–ApoB-100 Complex
Samples of reconstituted apoB-100 (0.95 μmol/L) were incubated with rFIII at 37°C for 0, 10, 30, and 60 minutes. Samples of reconstituted apoB-100 and FIII were used as a reference. All samples and references were prestained with the Chromatophor protein stain (Promega Corp). The samples were then examined on a 50×50-mm glass-plate agarose gel (10 g/L) in 0.12 mol/L barbitone buffer, pH 8.6. Electrophoresis was performed at 1 V/mm until the bands were clearly separated.
UV Spectroscopy of the Interaction Between FIII and ApoB-100
The interaction of equal volumes of apoB-100 (0.95 μmol/L reconstituted in 0.5 g/L soybean phosphatidylcholine) and FIII (50×103 U/L) in the absence of other factors was monitored at a wavelength range of 230 to 400 nm. Experiments were performed with commercial rabbit brain FIII and repeated with affinity-purified rabbit brain FIII and recombinant rabbit FIII. Difference spectra at 20 and 40 minutes were obtained by zeroing the reading for the measured range immediately after the addition of the reactants. When using crude and purified FIII, the samples were constantly mixed to prevent sedimentation of the FIII. The difference spectra were then used to calculate the resultant second derivative spectra. Similar spectra of the interaction between apoB-100 (0.95 μmol/L reconstituted as before) and FVII (10 U/L) recorded at 15-, 90-, and 180-minute intervals and FX (10 U/L) recorded at 15-, 60-, and 120-minute intervals were obtained.
The spectra of rFIII and apoB-100 were recorded individually over a range of pH values (5.5 to 9.0). Aliquots of FIII were reconstituted (final concentration, 50×103 U/L) in 0.5 mL Tris-HCl (0.01 mol/L) adjusted at pH 5.5 to 9.0. The spectra (300 to 400 nm) were obtained on a Beckman DU-70 spectrophotometer, and the spectra were superimposed. Similarly, the spectra of apoB-100 (0.95 μmol/L), reconstituted as before in 0.5 mL Tris-HCl (0.01 mol/L) adjusted to pH 5.5 to 9.0, were obtained. These were compared with the spectra of monomeric and polymeric amino acids, histidine, arginine, lysine, tyrosine, and glutamic and aspartic acids to assign the various shifts in the spectra of particular residues.
The interactions of 1-mL samples of apoB-100 (0.95 μmol/L, reconstituted as before) with 1 mL of either rFIII or purified FIII (50×103 U/L) were followed individually by measuring the absorption at 270 nm until no further rise in absorption was detected. In addition, the interaction of apoB-100 and FIII was followed at 383 nm as described above. The assays were then repeated with 1 mL of 100×103 and then 200×103 U/L FIII. Similar experiments were performed with three concentrations of FVII (5, 10, and 20 U/L) or FX (5, 10, and 20 U/L) and followed individually by measuring the absorption at 270 nm until no further rise in absorption was detected.
Finally, any homology between the primary structure of FIII and the LDL receptor protein (obtained from GenBank) was investigated by using the Bestfit and Gap programs (Genetic Computing Group, University of Wisconsin).
Inhibition of FIII by ApoB-100
The isolation of the apoB-100 to purity ensured the absence of any clotting factors or other FIII inhibitors.11 On incubation with FIII, apoB-100 was capable of inhibiting the procoagulant activity of FIII at 37°C but not 4°C (Fig 1⇓). The inhibition of FIII by apoB-100 did not require the presence of clotting factors X or VII in either the activated or zymogen form, nor was it calcium dependent. However, the presence of either the activated or zymogen form of these factors, together with calcium ions, enhanced the rate of inhibition (Fig 1⇓). In the absence of calcium ions (not shown) all the inhibition assays had similar rates of increase to that in which no additional factors were added, indicating that the accelerating effect of either FVII or FX was dependent on calcium ions.
In the following experiments, apoB-100 was preincubated with either FX or FVII for 60 minutes before the addition of FIII and measuring. The inhibition was completely blocked for 30 minutes on preincubation of apoB-100 with FVII prior to the addition of FIII and monitoring. The delay was increased to 90 minutes when apoB-100 was preincubated with FX before the addition of FIII (Fig 2⇓). Next, FIII was preincubated with either FX or FVII for 60 minutes before the addition of apoB-100. Preincubation of FIII and FVII together, before adding apoB-100, partially protected the FIII activity, as indicated by a slower rate of inhibition. The preincubation of FIII with FX had no effect on the rate of inhibition following the addition of apoB-100.
Agarose Gel Electrophoresis of the FIII–ApoB-100 Complex
The samples containing reconstituted apoB-100, rFIII, and apoB-100–FIII that were incubated for 0, 10, 30, and 60 minutes were examined on an agarose gel. The reconstituted apoB-100 showed a fast-running band, whereas the FIII band exhibited a slower, more diffuse band, probably due to heterogenous association with lipids. Agarose gel electrophoresis of FIII–apoB-100, following a minimum of 30 minutes’ preincubation, indicated the formation of a complex with a mobility between that of FIII and apoB-100 (Fig 3⇓). With incubations of less than 30 minutes both the FIII and apoB-100 bands were visible, whereas after longer periods of incubation the FIII band disappeared and there was a retardation of the apoB-100 band, suggesting the formation of a complex between the two proteins.
UV Spectroscopy of the Interaction Between FIII and ApoB-100
The incubation of apoB-100 with FIII produced various shifts in the UV spectrum of these proteins. The difference spectra of the interaction of apoB-100 and the three preparations of FIII were used to calculate the resultant second derivative spectra after incubation for 20 or 40 minutes. The spectra from the interaction of apoB-100 with either crude, purified, or recombinant FIII produced essentially the same difference spectra: the spectra for interaction with rFIII are shown in Fig 4⇓. In the first 20 minutes, the shift in the wavelengths corresponding to the aromatic amino acids (241, 246, 257, and 285 nm) and cystine/cysteine (235, 250, and 323 nm) to longer wavelengths indicated the formation of a hydrophobic environment.19 20 After incubation for a further 20 minutes, the shifts in wavelengths arising from tryptophan residues (278 and 285 nm), alterations in aspartic and glutamic residues (>390 nm), and the polarization of ε-amino groups (383 and 318 nm) of lysine and arginine residues were detected.20 Measurements at wavelengths >400 nm were too ambiguous, possibly due to masking by the absorptions arising from the presence of lipids.
The interaction between apoB-100 and FVII (Fig 5⇓) produced bathochromic shifts in the wavelengths corresponding to the aromatic amino acids (240 to 290 nm). Furthermore, some shifts in the wavelengths corresponding to delocalization of free amino side chains of positively charged amino acids were observed (383 nm).20 However, all the shifts occurred much more slowly and simultaneously. The interaction between apoB-100 and FX (Fig 6⇓) produced bathochromic shifts in wavelengths corresponding to aromatic amino acids similar to those with FVII. Moreover, the shifts occurred at rates that were faster than those for FVII and were completed within 90 minutes.
The interaction of apoB-100 with either FIII, FX, or FVII produced shifts in the wavelengths corresponding to those of tyrosine residues (223 to 240 and 274.5 to 293.5 nm), which may have been due to ionization/deionization of the hydroxyl groups of these residues.20
Assignment of Absorption Wavelengths
With increasing pH, the spectrum of FIII exhibited wavelength shifts (>390 nm). This indicated the polarization of the carboxyl groups of glutamic and aspartic acid residues (Fig 7⇓), since spectra of aspartate and glutamate monomers and polymers exhibited similar shifts (not shown) as the pH was increased. Moreover, this corresponded to the shift observed at these wavelengths on incubation with apoB-100. Therefore, the shift in absorption at >390 nm was assigned to the involvement of the negatively charged residues within FIII. The increase in pH produced specific shifts in the absorption of apoB-100 at 318 and 383 nm (Fig 8⇓). The spectra of histidine, arginine, and lysine, either in monomer or polymer form, exhibited similar shifts (not shown) at lower pH that were assigned to a delocalization of the side-chain nitrogen at neutral pH values. The shifts were greatest in histidine and lowest in lysine residues, as expected. These shifts in absorption were also detected on interaction of apoB-100 with FIII.
Time-Course Study of the Interaction Between FIII and ApoB-100 by UV Spectroscopy
By following the absorption of the samples at 270 nm over a period of time and using a range of concentrations of the reactants, it was possible to deduce the relative rate of interaction of apoB-100 with FIII, FVII, and FX. The chosen wavelengths were those that produced the greatest differences in absorption when examining the spectra at 230 to 400 nm in all three samples tested. The interaction of apoB-100 with FIII at 270 nm occurred at a faster rate (t=3.5 minutes) than that of apoB-100 and FX (t=5 minutes) (Fig 9⇓). The interaction between apoB-100 and FVII was completed even more slowly (t=21 minutes). The apparent dissociation constant (Kd) values for the interaction of apoB-100 with FIII measured at 270 and 383 nm were 0.03 and 0.171, respectively; with FX and FVII, these values were 0.08 and 0.97 measured at 270 nm. On the other hand, the interaction between apoB-100 and FIII, followed by absorption at 383 nm (corresponding to derivatization of ε-amino groups), occurred at a slower rate (t=10 minutes) than that measured at 270 nm (corresponding to aromatic interactions). These results indicate that the primary interaction between apoB-100 and FIII occurs more rapidly than the interactions between apoB-100 and FX or FVII. However, the second interaction between apoB-100 and FIII, which leads to the inhibition of the latter, occurs at a slower rate than that between apoB-100 and FX. Therefore, the presence of FX, in particular, may influence the interaction of apoB-100 and FIII.
The computer-based analysis of the structures of FIII and the LDL receptor protein (obtained from GenBank) revealed regions of similar amino acids within the two proteins (Table⇓). The similarity (77.8%) and identity (44.5%) of the examined peptides were considerable. Furthermore, both peptide domains were predicted to have similar secondary structures and to be partially α-helical (Fig 10⇓).
The data presented here indicate that the interaction of FIII with apoB-100 occurs in at least two sites. Thus, while the first site involves the attainment of a more hydrophobic configuration or the formation of hydrophobic interactions19 20 and occurs within 20 minutes of incubation, the evidence suggests that the second interaction involves negatively charged residues and positive free amino groups20 and occurs over a longer period of time (up to 60 minutes). This coincides with the progressive inhibition of FIII activity by apoB-100, which reached a maximum after 60 minutes. The incubation of apoB-100 with either FX or FVII also produced various modifications to the UV spectra. By monitoring the interactions between apoB-100 and FIII, FX, or FVII at various concentrations of each coagulation factor, we found that apoB-100 has a higher affinity for FIII than for the other two clotting factors. Hence, these data suggest a model for the formation of an apoB-100–FIII clotting factor complex in which, initially, apoB-100 interacts with FIII. The subsequent binding of FX and/or FVII leads to a more rapid induction of a complex that mediates the inhibition of FIII activity (Fig 11⇓).
While the sequence of the protein component of FIII (residues 58 through 66) is rich in negative residues, the receptor-binding domain of apoB-100 is rich in positive and tryptophan residues.21 22 The peptide region corresponding to the LDL receptor protein (283 through 291) is predicted to be partly α-helical,23 24 and the domain within FIII (residues 58 through 66) is mainly α-helical.25 When the secondary structures of these two α-helical sequences were reconstructed, it was possible to identify similar motifs within them (Fig 10⇑). We propose, therefore, that this domain may be involved in the interaction and inhibition of FIII by apoB-100. We are currently investigating the possible function of this domain in the interaction with apoB-100.
From this study, we conclude that the interaction and inhibition of FIII by apoB-100 are two separate but interrelated processes. Furthermore, although the presence of clotting factors X and VII are not essential for this interaction and, under certain conditions, they may compete with FIII for the binding sites on apoB-100, they may act cooperatively in the formation of the inhibited complex under physiological conditions. The progress of the inhibition was found to be temperature-dependent, as no inhibition was observed at 4°C. Also, the interactions that lead to FIII inhibition seem be to be time dependent. This would explain the negative results obtained by Aviram and Presser,26 who did not observe any effect when they assessed FIII activity immediately after adding LDL to FIII. The blocking of the inhibition that resulted from preincubation of apoB-100 with either clotting factor X or VII reinforces the suggestion that FIII and these clotting factors are bound by apoB-100 in a definite order of priority if inhibition is to occur. Alternatively, the FIII binding site may be masked by FX or FVII, thereby preventing inhibition (Fig 11⇑).
The variable effects of the interaction between FIII and apoB-100 that result in either inhibition or enhancement of FIII activity under different conditions8 9 10 11 15 16 could arise from changes in the conformation of apoB-100 due to variations in the lipid composition of the lipoprotein (Reference 2727 and C. Ettelaie et al, unpublished data, 1995). This indicates a mechanism of interaction that is more complex than the simple adsorption of the FIII complex to the LDL. We propose that in vivo the action of FIII is suppressed by its interaction with the apoB-100 of LDL as well as other FIII inhibitors, eg, TFPI. This interaction occurs at two sites, one of which is likely to be within the receptor-binding domain on apoB-100.
In conclusion, this study indicates that apoB-100 within LDL is capable of inhibiting the procoagulant activity of FIII via a different and independent mechanism than that of TFPI and other FIII inhibitors. In addition, the rate at which the inhibition by apoB-100 occurs is slower than that by TFPI. Therefore, LDL may play an important role in the physiological functioning of FIII and the generation of thrombosis in disease conditions.
Selected Abbreviations and Acronyms
|FIII||=||factor III (thromboplastin)|
|rFIII||=||recombinant factor III|
|TPFI||=||tissue-factor pathway inhibitor|
We would like to acknowledge the support of the Wellcome Trust and the British Heart Foundation.
Nemerson Y, Pepke D. Tissue factor accelerates the activation of coagulation factor VII: the role of a bifunctional coagulation cofactor. Thromb Res. 1985;51:165-173.
Bajaj SP, Harmony JAK, Martinez-Carrion M, Castellino FJ. Human plasma lipoproteins as accelerators of prothrombin activation. J Biol Chem. 1976;251:5233-5236.
Nemerson Y. Tissue factor and haemostasis. Blood. 1988;71:1-8.
Broze GJ, Miletich JP. Characterization of the inhibition of tissue factor in serum. Blood. 1987;69:150-155.
Rao LVM, Rapaport SI. Studies of a mechanism inhibiting the initiation of the extrinsic pathway of coagulation. Blood. 1987;69:645-651.
Sandset PM, Warn-Cramer BJ, Rao LVM, Maki SL, Rapaport SI. Deletion of extrinsic pathway inhibitor [EPI] sensitizes rabbits to disseminated intravascular coagulation induced with tissue factor: evidence supporting a physiologic role for EPI as a natural anticoagulant. Proc Natl Acad Sci U S A. 1991;88:708-712.
Kondo S, Kisiel W. Evidence that plasma lipoproteins inhibit the factor VIIa-tissue factor complex by a different mechanism than extrinsic pathway inhibitor. Blood. 1987;70:1947-1954.
Howell RM, Ettelaie C. Inhibition of thromboplastin by serum fractions from various species. Biochem Soc Trans. 1989;17:576-577. Abstract.
Howell RM, Ettelaie C. The effect of barium sulphate adsorbable proteins on thromboplastin activity. Biochem Soc Trans. 1990;18:680-681. Abstract.
Carson SD. Tissue factor (coagulation factor III) inhibition by apolipoprotein A-II. J Biol Chem. 1987;262:718-721.
Sanders NL, Bajaj SP, Zivelin A, Rapaport SI. Inhibition of tissue factor/factor VIIa activity in plasma requires factor X and an additional plasma component. Blood. 1985;66:204-212.
Bruckdorfer KR, Ettelaie C. The use of LDL-resembling particle (LDL-r) as a model for the investigation of the effect of oxidation on the behaviour of LDL. Biochem Soc Trans. 1995;23:243. Abstract.
Ettelaie C, Howell RM, Bruckdorfer KR. Vitamin E protects the inhibitory effect of low density lipoprotein towards thromboplastin. 35th Meeting of the International Society of Lipid Biochemists; Aberdeen, Scotland; 1994. Abstract.
Wilcheck W, Miron T. Immobilization of enzymes and affinity ligands onto agarose via stable and uncharged carbamate linkages. Biochem Int. 1982;4:629-635.
Howell RM, Rezvan H. Circular dichroic, infrared and other studies on the protein component of pig brain thromboplastin. Biochem J. 1980;189:209-218.
Rao CNR. Ultraviolet and Visible Spectroscopy. 2nd ed. London, England: Butterworth; 1967.
DeLoof H, Rosseneu M, Yang CY, Li WH, Gotto AM, Chan L. Human apolipoprotein B: analysis of internal repeats and homology with other apolipoproteins. J Lipid Res. 1987;28:1455-1465.