An Anticoagulant RNA Aptamer That Inhibits Proteinase-Cofactor Interactions within Prothrombinase

Abstract

The interaction of factor Xa with factor Va on membranes to form prothrombinase profoundly increases the rate of the proteolytic conversion of prothrombin to thrombin. We present the characterization of an RNA aptamer (RNA_11F7t) selected from a combinatorial library based on its ability to bind factor Xa. We show that RNA_11F7t inhibits thrombin formation catalyzed by prothrombinase without obscuring the active site of Xa within the enzyme complex. Selective inhibition of protein substrate cleavage arises from the ability of the aptamer to bind to factor Xa and exclude interactions between the proteinase and cofactor within prothrombinase. Competition for enzyme complex assembly results from the binding of RNA_11F7t to factor Xa with nanomolar affinity in a Ca²⁺-dependent interaction. RNA_11F7t binds equivalently to the zymogen factor X as well as derivatives lacking γ-carboxyglutamic acid residues. We suggest that the ability of RNA_11F7t to compete for the Xa-Va interaction with surprisingly high affinity likely reflects a significant contribution from its ability to indirectly impact regions of Xa that participate in the proteinase-cofactor interaction. Thus, despite the complexity of the macromolecular interactions that underlie the assembly of prothrombinase, efficient inhibition of enzyme complex assembly and thrombin formation can be achieved by tight binding ligands that target factor Xa in a discrete manner.

Introduction

The assembly of the prothrombinase complex is essential for rapid thrombin formation after the initiation of blood coagulation (1, 2). This macromolecular complex assembles though reversible interactions between the serine proteinase, factor Xa, and its cofactor, factor Va, on membranes exposing aminophospholipids (1, 2). Although factor Xa itself can catalyze the conversion of prothrombin to thrombin, its assembly into prothrombinase yields a profound increase in catalytic efficiency for thrombin formation (1, 2). Comparable interactions of coagulation serine proteinases with their corresponding cofactor on membranes render equally spectacular enhancements in function in other proteolytic activation steps of blood coagulation (2, 3). Thus, membrane-dependent complex assembly and its targeted down-regulation represent endogenous strategies employed to achieve stringent biological control of product formation in the individual steps of the clotting cascade (2).

Because dysregulated blood coagulation leading to excessive clot formation is a principal contributor to human vascular disease and mortality, extensive efforts continue to be devoted toward developing specific inhibitors for the modulation of coagulation (4, 5). The conversion of prothrombin to thrombin, the only activation step that is not duplicated in the pathway toward thrombin formation, poses advantages as a potential target for inhibition. This assertion has been borne out by the success of inhibitors that target factor Xa for therapeutic gain (4, 5). However, the widely practiced approach of developing reversible inhibitors targeting the active site of factor Xa yields ligands with unexpected inhibitory properties with respect to the action of prothrombinase on its protein substrate (6, 7). This arises in part from the major role played by extended surfaces, distinct from the active site, in driving the affinity of the enzyme complex for prothrombin (7). In addition, reversible inhibition of factor Xa interferes with endogenous mechanisms of regulation including proteinase elimination by inhibitors such as antithrombin III (8).

As the incorporation of factor Xa into prothrombinase yields an ∼10⁵-fold increase in thrombin formation, disruption of the high affinity membrane-dependent interaction between factors Xa and Va is expected to yield a dramatic reduction in the rate of prothrombin activation. Accordingly, an inactive derivative of factor Xa that competes with active Xa for interactions within prothrombinase has proved an efficient inhibitor of clot formation (9). Furthermore, because the assembly of prothrombinase is largely unaffected by ligand binding to the active site of the proteinase, disruption of prothrombinase can be accomplished without abrogating active site function that is otherwise also required for the down-regulation of proteinase function by endogenous inhibitors (1, 2).

However, the membrane-dependent interaction between factors Xa and Va is stabilized by linked protein-protein and protein-membrane interactions within the complex (10). The protein-protein contacts are most likely spread over multiple sites as well as communicating domains in the interacting species (11–13). These constraints coupled with the likelihood that some of these contacts occur over shallow surfaces point to difficulties in achieving high affinity inhibition of prothrombinase assembly by small ligands that discretely target the interacting species in a site-specific manner. This suggestion is consistent with a long history documenting relatively inefficient inhibition of prothrombinase function by peptides corresponding to epitopes within Xa and Va considered responsible for mediating their interactions within prothrombinase (13–16).

The systematic evolution of ligands by exponential enrichment (SELEX)³ approach has proved useful for the development of unique nucleic acid probes and/or inhibitors that bind their target proteins with high affinity and specificity (17, 18). We present the characterization of an RNA aptamer (RNA_11f7t) developed by cyclical screening of a combinatorial library against factor Xa. We show that potent inhibition of thrombin formation catalyzed by prothrombinase unexpectedly results from the ability of RNA_11f7t to bind factor Xa with high affinity and inhibit its interaction with factor Va on the membrane surface.

EXPERIMENTAL PROCEDURES

Reagents

Small unilamellar phospholipid vesicles (PCPS) composed of 75% (w/w) hen egg l-α-phosphatidylcholine and 25% (w/w) porcine brain l-α-phosphatidylserine (Avanti Polar lipids) were prepared and quality-controlled as described (19). Peptidyl substrates methoxycarbonyl-d-cyclohexylglycyl- glycyl-l-arginine-p-nitroanilide (SpXa) and H-d-phenylalanyl-l-pipecolyl-l-arginine-p-nitroanilide (S2238) were from American Diagnostica and Chromogenix, respectively. Stock solutions (∼4 mm) were prepared in water, and concentrations were determined using E₃₄₂ = 8270 m⁻¹·cm⁻¹ (20). Oregon Green₄₈₈ maleimide, succinimidyl acetothioacetate (Invitrogen), l-glutamyl-l-glycinyl-l-arginine chloromethyl ketone (EGR-CH₂Cl, Calbiochem), chymotrypsin (Worthington), soy bean trypsin inhibitor-Sepharose (Sigma), and (4-amidinophenyl)-methanesulfonyl fluoride (APMSF, Sigma) were obtained from the indicated suppliers. Acetothioacetyl-EGR-CH₂Cl was prepared, purified, and characterized as described (21). Human plasma used for protein isolation was a generous gift of the Plasmapheresis unit of the Hospital of the University of Pennsylvania. Unless otherwise specified, fluorescence and activity measurements were performed in 20 mm Hepes, 150 mm NaCl, 5 mm CaCl₂, 0.1% (w/v) polyethylene glycol 8000, pH 7.5 (assay buffer) at 25 °C.

Proteins

Factor X, prothrombin, antithrombin III, factor IX, and factor V were isolated from human plasma using established procedures (22–24). Human factor X was freed of trace contaminants by chromatography on soy bean trypsin inhibitor-Sepharose and immunoaffinity chromatography using the Ca²⁺-dependent monoclonal antibody 4G3 (21, 25). Factor X and traces of Xa were depleted from the prothrombin and factor IX preparations using the same approach. Prethrombin 2 and thrombin were obtained by preparative proteolysis of prothrombin followed by isolation as detailed (26). Factors Xa and Va were prepared by proteolytic activation of factors X and V, respectively, and repurification as before (21). Factor Xa preparations contained approximately equimolar amounts of the α and β forms and yielded 1.1–1.2 mol of active sites/mol of protein in kinetic titrations with p-nitrophenol p′-guanidinobenzoate (27). Quality control of the factor Va preparations was performed by equilibrium binding measurements of the ability of factor Va to assemble into prothrombinase (below). Factor Xa was inactivated with APMSF and purified to yield Xa_i or inactivated with acetothioacetyl-EGR-CH₂Cl, reacted with Oregon Green₄₈₈ maleimide after thioester hydrolysis, and purified to yield OG₄₈₈-Xa (21). A derivative of factor Xa lacking the γ-carboxyglutamate domain was prepared by proteolysis using chymotrypsin pretreated with tosyl-lysine chloromethyl ketone and purified as described (28, 29). The product was inactivated with APMSF to yield ΔGla-Xa_i. Recombinant factor X containing an Ala substitution for the catalytic serine (X_S195A) was expressed in HEK293 cells and purified to separate fully γ-carboxylated species from the uncarboxylated species as before (30). The fraction devoid of γ-carboxyglutamic acid (desGla-X_S195A) was activated and purified by affinity chromatography with soy bean trypsin inhibitor-Sepharose to yield desGla-Xa_S195A (30). Described procedures were used to convert IX to IXaβ and repurify the product (24). Recombinant factor VIIa was from Novo-Nordisk (Denmark) and full-length recombinant factor VIII was obtained as a generous gift from Lisa Regan (Bayer, Berkeley, CA). Recombinant tissue factor (TF) was reconstituted into PCPS vesicles and quality controlled as described (31). Recombinant factor VIII was repurified by anion exchange chromatography as described (32), dialyzed into assay buffer supplemented with 0.02% (v/v) Tween 20, and stored at −80 °C. A recombinant version of tissue factor pathway inhibitor comprising residues 13–161 (TFPI_13–161) was produced in HEK293 cells and purified from conditioned medium. Protein concentrations were determined using the following molecular weights (M_r) and extinction coefficients (E₂₈₀, mg⁻¹·cm²): prothrombin, 72,000, 1.47 (26); prethrombin 2 and thrombin, 37,500, 1.89 (26); X, 56,500, 1.16 (33); Xa, Xa_i, Xa_S195A and desGla-Xa_S195A, 45,300, 1.16 (33); ΔGla-Xa_i, 38,000, 1.16; Va, 168,000, 1.74 (21, 34); IXaβ, 46,000, 1.32 (35); VIIa, 50,000, 1.39 (36); VIII, 264,700, 1.22 (37); TFPI_13–161, 17,400, 0.45 calculated from the primary sequence by the method of Gill and von Hippel (38).

Generation of RNA Aptamers

SELEX procedures were carried out as previously described (39–41). The sequence of the starting RNA combinatorial template was 5′-GGGAGAGAGGAAGAGGGATGGGN₄₀CATAACCCAGAGGTCGATAGTACTGGATCCCCCC-3′, where N₄₀ represents 40 nucleotides containing equimolar A, G, C, and U residues at each position. The 2′-fluoro derivatives of cytidine triphosphate and uridine triphosphate (TriLink Biotechnologies) were incorporated into RNA libraries by in vitro transcription as described (40). In vitro selection was performed by incubation of 5 nmol of RNA with 0.5 nmol of Xa in 20 mm Hepes, 0.05 m NaCl, 2 mm CaCl₂, 0.01% (w/v) bovine serum albumin, pH 7.4, followed by isolation of proteinase-bound RNA by filtration through a 0.45-μm nitrocellulose membrane (Schleicher and Schuell). Bound RNA was eluted using phenol:chloroform:isoamyl alcohol (25:24:1) and precipitated with ethanol. One-quarter of the precipitated RNA was amplified by reverse transcriptase PCR. The product was transcribed, and the resulting RNA was used in the next round of selection. As the rounds progressed, the concentration of Xa was decreased to increase the ratio of RNA to protein. After 11 rounds of selection, the products were digested with EcoR1 and BamH1 (New England Biolabs) and directionally cloned into pUC19 linearized with the same enzymes. Individual clones were sequenced, and clonal RNA transcripts were analyzed in filter binding assays. The lead molecule was systematically shortened to obtain a truncated version (RNA_11F7t) that retained binding activity. In parallel studies point mutations were introduced to yield RNA_MUT with a greatly reduced ability to bind Xa. Ribooligonucleotides were chemically synthesized by Dharmacon Research and supplied desalted after deprotection form 2′-hydroxylated purines. The aptamer RNA_11F7t corresponded to 5′- GAGAG(^2′FC)(^2′FC)(^2′FC)(^2′FC)AG(^2′FC)GA GA(^2′FU)AA(^2′FU)A(^2′FC)(^2′FU)(^2′FU)GG(^2′FC) (^2′FC)(^2′FC)(^2′FC)G(^2′FC)(^2′FU)(^2′FC)(^2′FU) (^2′FU)-idT, and RNA_MUT comprised 5′-GAGAG(^2′FC)(^2′FC)(^2′FC)(^2′FC)AG(^2′FC)GAGA(^2′FU)AA(^2′FU)A(^2′FC)(^2′FU)(^2′FU)G(^2′FU)A(^2′FC)(^2′FC)(^2′FC)G(^2′FC)(^2′FU)(^2′FC)(^2′FU)(^2′FU)-idT, where ^2′FC is 2′-flurocytosine, ^2′FU is 2′flurouracil, and idT denotes inverted deoxythymidine. RNA_MUT differs from RNA_11F7t by substitutions at positions 26 and 27. Aptamers were dissolved in assay buffer or dialyzed into assay buffer lacking polyethylene glycol and stored at −20 °C. Concentrations were determined using E₂₆₀ = 353,000 m⁻¹·cm⁻¹. The calculated formula weight (11,827) was confirmed by mass spectrometry performed at the Emory University Microchemical Facility. Aptamer preparations were renatured before each use by melting at 60 °C for 5 min followed by cooling to ambient temperature.

Coagulation Measurements

Clotting assays were performed using a model ST4 mechanical coagulometer (Diagnostica Stago). For measurements of the prothrombin time (PT), 50 μl of pooled normal human plasma (George King Bio-Medical) was incubated for 5 min at 37 °C with increasing concentrations of aptamer. Clotting was initiated by the addition of 100 μl of Simplastin (BioMerieux). For activated partial thromboplastin time (APTT) measurements, 50 μl of pooled normal human plasma was mixed with 50 μl of MDA platelin reagent (BioMerieux) and incubated with increasing concentrations of aptamer for 5 min at 37 °C. Clotting was initiated by the addition of 50 μl of 25 mm CaCl₂. Clotting results are presented as the ratio of clot times in the presence of aptamer to the clot time in buffer.

Protein/RNA Binding

Binding measurements were conducted with ³²P end-labeled RNA using purified coagulation proteins obtained from Hematologic Technologies as previously detailed (39). Proteins were serially diluted in 20 mm Hepes, 150 mm NaCl, 2 mm CaCl₂, 0.01% (w/v) bovine serum albumin containing a fixed and trace amount of end-labeled RNA. After incubation at 37 °C, reaction mixtures were filtered under vacuum with a Protran membrane (Schleicher and Schuell) positioned over a GeneScreen Plus nylon membrane (PerkinElmer Life Sciences) to adsorb protein-bound RNA and free RNA, respectively. Binding constants were estimated as previously described (39, 40).

Progress Curves for Prothrombin Cleavage

Reaction mixtures (300 μl) containing 1.4 μm prothrombin, 36 μm PCPS, 30 nm Va with or without 250 nm RNA_11F7t or RNA_mut at 25 °C were initiated with 0.2 nm Xa. Aliquots (10 μl), withdrawn at various times after initiation, were quenched by mixing with 90 μl of assay buffer lacking Ca²⁺ but containing 50 mm EDTA. Quenched samples were further diluted in the same buffer in wells of a 96-well plate, and initial rates of S2238 hydrolysis were determined by monitoring the change in absorbance at 405 nm after the addition of 100 μm peptidyl substrate using a Gemini kinetic plate reader (Molecular Devices). Initial rates were converted to concentrations of proteinase product(s) formed as a function of time from the linear dependence of initial rate on known concentrations of thrombin.

Initial Velocity Studies of Prethrombin 2 Cleavage

Initial velocity measurements of thrombin formation from prethrombin 2 were determined discontinuously using the strategy described above with the detailed considerations previously described (6). Trivial effects on initial rates were ruled out by the equivalence in rates of S2238 hydrolysis catalyzed by thrombin both in the absence or presence of the highest concentrations of aptamer used. In each case initial velocities were determined from the linear appearance of product with time in six serially quenched samples. For studies of the inhibition of prethrombin 2 cleavage by increasing concentrations of aptamer, reaction mixtures contained either 1.4 or 3.0 μm prethrombin 2, 50 μm PCPS, 40 nm Va, the indicated concentrations of aptamer and 5 nm Xa. Equivalent results were obtained when reactions were initiated with Xa or with prethrombin 2 after prolonged incubation of aptamer with the Xa/Va/PCPS mixture. For studies in which the concentration of factor Va was varied, reaction mixtures contained 1.4 μm prethrombin 2, 50 μm PCPS, the indicated concentrations of Va, 5 nm Xa, and aptamer concentrations fixed at 0, 0.05, 0.2, 1.0, and 2.0 μm.

Peptidyl Substrate Hydrolysis by Xa and Prothrombinase

Reaction mixtures (150 μl) containing 1 nm Xa or 1 nm Xa plus 50 μm PCPS with Va fixed at 0, 5, 15, and 50 nm and increasing concentrations of the indicated aptamer were initiated with SpXa (50 μl) to achieve a final concentration of 100 μm. Initial velocities were determined in a kinetic plate reader. Steady state constants were determined by initial velocity measurements either in the absence or presence of 250 nm RNA_11F7t by varying the concentration of SpXa to 240 μm (24 values) using either 1 nm Xa, 1 nm Xa plus 60 μm PCPS or 1 nm Xa, 60 μm PCPS and 15 nm Va.

Fluorescence Measurements

All steady state fluorescence measurements were performed at 25 °C using a PTI QuantaMaster fluorescence spectrophotometer (Photon Technology Inc.). Intensity measurements of the binding of aptamer to OG₄₈₈-Xa were performed in assay buffer supplemented with 0.02% (v/v) Tween 20, which prevented systematic and time-dependent decreases in fluorescence at low concentrations of aptamer, possibly because of adsorption artifacts. Ratiometric fluorescence intensity was recorded by averaging 30 readings over 30 s using λ_EX = 490 nm and λ_EM = 520 nm with a KV-500 (Schott Glass) long pass filter in the emission beam and polarizers oriented under “Magic Angle” conditions (42). Reaction mixtures (2.5 ml) in 1 × 1-cm stirred cuvettes typically contained 15–18 nm OG₄₈₈-Xa, and intensity was recorded after matched incremental additions of RNA_11F7t, RNA_MUT, or buffer. For measurements in the absence of Ca²⁺, the buffer contained 50 μm EDTA in the place of 5 mm Ca²⁺. Division of the signal observed in samples containing aptamer by the signal of the buffer sample yielded the normalized change in fluorescence in OG₄₈₈-Xa, corrected for dilution (F/F_o). For competitive binding measurements, reaction mixtures contained 18 nm OG₄₈₈-Xa and different fixed concentrations (between 15 and 80 nm) of RNA_11F7t matched with a sample containing the same concentration of RNA_MUT. Fluorescence intensity was recorded after equal incremental additions of competitor to both cuvettes. F/F_o was calculated by dividing the signal observed in the sample containing RNA_11F7t by that observed in the sample containing RNA_MUT.

Fluorescence anisotropy measurements of the membrane-dependent interaction of OG₄₈₈-Xa with factor Va were performed as previously detailed (21). For competition studies with aptamer, reaction mixtures (2.5 ml) in assay buffer contained 35 nm OG₄₈₈-Xa, 50 μm PCPS, and fixed aptamer concentrations of 0, 0.35, 2, and 5 μm. Fluorescence anisotropy and total intensity were measured after incremental additions of factor Va as previously described (21). All measurements were corrected for scattering using a parallel reaction mixture containing 35 nm Xa_i in place of OG₄₈₈-Xa. Small decreases in the anisotropy of OG₄₈₈-Xa in the presence of aptamer were accommodated by expressing the data as the change in anisotropy (Δr) from the sample lacking Va.

Sedimentation Equilibrium

Analytical ultracentrifugation was performed using an AN60-Ti rotor in a XL-I instrument (Beckman Coulter). Samples were dialyzed extensively into 20 mm Hepes, 0.15 m NaCl, 5 mm Ca²⁺, pH 7.4, and 100-μl aliquots were loaded into 6 sector cells with an epon centerpiece and sapphire windows. Centrifugation was at 20 °C using increasing rotor speeds of 15,000, 20,000, 25,000, and 30,000 rpm. Absorbance scans were collected at 260 and 280 nm between 20 and 22 h after the rotor had achieved speed, and an equilibrium distribution of species was verified by the equivalence of successive scans obtained after an additional 4 or 8 h had elapsed.

Right Angle Light Scattering

Measurements were performed in the fluorescence spectrophotometer at 25 C using λ_EX = λ_EM = 320 nm. Reaction mixtures (2.5 ml) contained assay buffer alone, 0.3 μm Xa_i, 0.6 μm RNA_11F7t, or 0.3 μm Xa_i plus 0.6 μm RNA_11F7t. Scattering intensity was measured by averaging readings for 30 s after incremental additions of PCPS (0–30 μm) to all four cells. Scattering changes associated with membrane binding were determined by subtracting the linear dependence of scattering intensity from the sample lacking Xa_i from the experimental sample.

Inhibition of Factor Xa by Macromolecular Inhibitors

The first order decay in Xa (10 nm) in the presence of antithrombin III (2.0 μm) was measured in assay buffer at 25 °C either in the absence or presence of 270 nm RNA_11F7t using approaches previously described (24). Inhibition studies with TFPI_13–161 were performed after prolonged incubation of proteinase with inhibitor as previously detailed (43). For these measurements, 0.4 nm Xa was incubated with increasing concentrations of TFPI_13–161 (0–4 nm, 24 concentrations) either without or with 254 nm RNA_11F7t for 1 h at room temperature before residual Xa activity was determined with SpXa.

Kinetics of Factor X Activation

Initial velocity studies of factor X activation by VIIa.TF were conducted using described procedures (31). Xa formation was determined using 100 nm factor X ([S] ≅ K_m), 0.1 nm VIIa, and 43 nm TF incorporated into PCPS membranes (43 nm TF, 50 μm PCPS) with increasing concentrations of RNA_11F7t. A similar strategy was employed in studies of Xa formation by IXa·VIIIa. Factor VIII (40 nm), rapidly activated by a 30-s incubation with 3 nm thrombin, was treated with 40 nm hirudin and further diluted to prepare a 2× enzyme solution containing 0.4 nm IXa, 100 μm PCPS, and 20 nm VIIIa. Xa formation was initiated by mixing with an equal volume of 80 nm factor X containing increasing concentrations of RNA_11F7t followed by discontinuous measurements of initial velocity. Measured steady state kinetic constants for Xa formation in the absence of aptamer under these conditions were K_m = 33.8 ± 1.4 nm and V/E = 1.65 ± 0.08 s⁻¹.

Activation of VIII

Factor VIII activation was initiated by the addition of 3 nm proteinase (thrombin or Xa) to reaction mixtures containing 30 μm PCPS and 50 nm VIII in assay buffer plus 0.02% Tween 20 with and without 256 nm RNA_11F7t. Aliquots were removed at the indicated times and quenched by mixing with an equal volume of 125 mm Tris, 50 mm EDTA, 120 mm dithiothreitol, 2% (w/v) SDS, 20% (v/v) glycerol, 0.02% (w/v) bromphenol blue, pH 6.8, heated at 90 °C, and subject to SDS-PAGE. Products were detected by Western blotting as previously detailed (22), using a monoclonal antibody directed toward the A2 domain (GMA-012, Green Mountain Antibodies).

Data Analysis

Data were analyzed according to the indicated equations by nonlinear least squares regression employing the Levenberg-Marquardt algorithm (44). Fitted parameters are listed ±95% confidence limits. The data illustrated are typically representative of two or more experiments performed at a similar level of detail. In some instances we have employed normalization approaches to permit the simultaneous display and analysis of multiple experiments performed with different preparations of the interacting species.

The change in OG₄₈₈-Xa fluorescence with increasing concentrations of aptamer was analyzed assuming the interaction of the varied ligand with equivalent and non-interacting sites using Equation 2 in Betz and Krishnaswamy (45) to yield fitted values for K_d, mol of aptamer bound/mol of OG₄₈₈-Xa at saturation, and the limits of the fluorescence change (F_o and F_max/F_o). Competitive binding measurements were analyzed with the same assumptions employing the Newton-Raphson solution of cubic Equation 17 in Olson et al. (46) to yield fitted values for K_d and stoichiometry for the binding of RNA_11F7t to OG₄₈₈-Xa and its binding to the competing species, F_max/F_o for the indicator reaction and F/F_o at infinite concentrations of competitor. Because of high parameter correlation, analysis of some data sets was done by fixing the stoichiometry for the binding of RNA_11F7t to OG₄₈₈-Xa to 2 and/or F/F_o at infinite competitor to 1.

Absorbance as a function of radial position acquired at two wavelengths and multiple speeds at sedimentation equilibrium were globally analyzed using SEDPHAT (47). Calculations were constrained by mass conservation, and the molecular weight was derived using ρ = 1.0046 g/cm³ and v̄ = 0.508 cm³/g (48). Confidence limits (67%) of the fits were calculated by Monte-Carlo analysis.

Activity and anisotropy measurements examining prothrombinase assembly in the absence or presence of fixed concentrations of aptamer were performed at saturating concentrations of PCPS, which allows for the determination of equilibrium binding parameters for the membrane-dependent interaction between Xa and Va as previously detailed (10). The dependence of the initial rate of thrombin formation on increasing concentrations of factor Va at different fixed concentrations of aptamer were analyzed by assuming mutually exclusive binding interactions in rapid equilibrium between Xa and factor Va versus Xa and aptamer. Analysis according to the cubic equation of reference (46) yielded fitted values for K_d for the membrane-dependent interaction between Xa and Va (K_dXa,Va), mol of Va bound/mol of Xa at saturation, K_d for aptamer binding to Xa (K_{dRNA, Xa}), mol of aptamer bound/mol of Xa at saturation, rate at zero Va, and rate at infinite Va.

Anisotropy measurements of the binding of increasing concentrations of Va to OG₄₈₈-Xa in the presence of different fixed concentrations of aptamer were analyzed with the same assumptions and an equivalent strategy. To accommodate bias introduced by differential increases in fluorescence intensity that accompany the binding of OG₄₈₈-Xa to Va (F_max/F_o = 1.38) or to aptamer (F_max/F_o = 1.76), the observed change in anisotropy at any given concentration of Va (Δr_obs) was related to the fractional saturation of OG₄₈₈-Xa with Va (f_b) by the expression analogous to that described by Lakowicz (42),

where Δr_max is the maximum change in anisotropy at saturating Va, and R is the ratio of intensities for OG₄₈₈-Xa saturated with Va and OG₄₈₈-Xa in the absence of Va. In the presence of fixed concentrations of aptamer, the intensity of OG₄₈₈-Xa uncomplexed with Va was higher than that at saturating Va. Analysis by these approaches yielded fitted values for K_d for the membrane-dependent interaction between OG₄₈₈-Xa and Va (K_dXa,Va), mol of Va bound/mol of OG₄₈₈-Xa at saturation, K_d for aptamer binding to OG₄₈₈-Xa (K_dRNA,Xa), mol of aptamer bound/mol of OG₄₈₈-Xa at saturation, and the maximum change in anisotropy (Δr_max) at infinite Va.

RESULTS

Specificity of RNA Binding

Screens for binding specificity of RNA_11F7t using filter binding assays revealed the ability of the aptamer to discriminate between the homologous proteinases of coagulation (Table 1). Binding with apparently nanomolar affinity was observed with factor Xa, whereas the other coagulation proteinases bound weakly with K_d > 3 μm (Table 1). Thus, RNA_11F7t exhibits ∼2700-fold selectivity for binding to Xa over the other coagulation proteinases.

TABLE 1.

Binding Specificity of RNA_11F7t

Apparent binding affinities were determined by filter binding using trace concentrations of RNA_11F7t, and increasing concentrations of the indicated proteinase. Values of K_d > 3 μm denote instances in which binding of RNA_11F7t to the protein remained low and was not saturated at concentrations as high as 3 μm.

Proteinase target	Kd
	nm ± S.D.
Factor Xa	1.1 ± 0.2
Thrombin	>3000
Factor VIIa	>3000
Factor IXa	>3000
Factor XIa	>3000
Activated protein C	>3000

Inhibition of Biological Function

RNA_11F7t was an effective inhibitor of clot formation in human plasma. Increasing concentrations of RNA_11F7t prolonged both the PT as well as the APTT (Fig. 1A). It follows that the binding of RNA_11F7t inhibits some or all functions of factor Xa that regulate clot formation in plasma. Specificity for inhibition is documented by the lack of effect of RNA_MUT on the clotting time in either assay format (Fig. 1A). The apparently larger impact of RNA_11F7t on the APTT relative to the PT suggests potentially complex effects of the aptamer in these plasma assays. Further insights into the ability of RNA_11F7t to inhibit coagulation were sought through studies in purified and reconstituted systems. Progress curves for the activation of prothrombin, present at its physiological concentration, by factor Xa saturably incorporated into prothrombinase revealed that 250 nm RNA_11F7t inhibited the initial rate of product formation by ∼85%, whereas RNA_MUT was without effect (Fig. 1B). These findings establish RNA_11F7t as an inhibitor of prothrombinase and indicate that its ability to prolong clot time is at least partly related to its ability to inhibit thrombin formation. However, interpretations are compromised by the fact that prothrombin is activated by two successive cleavage reactions with two possible pathways for thrombin formation (49). Because both thrombin and the intermediate meizothrombin are detected as products in this measurement, interpretation could be compromised by complex effects of the aptamer on the enzyme as well as pathway for product formation. Consequently, further characterization of the inhibitory properties of RNA_11F7t was pursued using prethrombin 2 as a substrate analog. Prethrombin 2 is cleaved by prothrombinase at a single site to yield thrombin with kinetic constants that can be interpreted in the traditional way without the added complexities associated with multiple enzyme-catalyzed reactions or rate-limiting effects of membrane-mediated substrate delivery to the enzyme (50). This reaction is established to mirror the features of the first cleavage reaction in prothrombin activation by prothrombinase (23).

FIGURE 1.

Aptamer-dependent inhibition of biological function. Panel A, shown is plasma clotting. The effects of increasing concentrations of RNA_11F7t (●, ○) or RNA_MUT (▴, ▵) on the APTT (open symbols) or the PT (closed symbols) of normal human plasma are illustrated. The dashed line denotes the increase in APTT seen with factor X-deficient plasma compared with normal human plasma (solid line). Symbols and error bars denote the mean ± S.D. from three measurements. Panel B, shown is prothrombin activation by prothrombinase. Progress curves for the formation of proteinase product were constructed using reaction mixtures containing 1.4 μm prothrombin, 36 μm PCPS, 30 nm Va, and 0.2 nm Xa with no addition (●), 250 nm RNA_MUT (○), or 250 nm RNA_11F7t (▴). Results with two independent preparations of RNA_11F7t are illustrated. The lines were arbitrarily drawn.

Inhibition of Prethrombin 2 Cleavage

Initial velocity studies of prethrombin 2 activation catalyzed by Xa assembled into prothrombinase with saturating concentrations of PCPS and Va confirmed the initial findings with prothrombin (Fig. 2). Increasing concentrations of RNA_11F7t reduced the initial velocity for thrombin formation, whereas no obvious decrease in rate was observed with RNA_MUT (Fig. 2). Because the two aptamers differ in composition and sequence by only two nucleotide bases, these findings again establish specificity in the ability of RNA_11F7t to inhibit prothrombinase function. It also appears improbable that this inhibitory effect arises from the polyanionic character of the aptamer or from some other common feature of ribooligonucleotides in general. Inhibition by RNA_11F7t was the same when reaction mixtures were initiated with Xa or after prolonged incubation of aptamer with preformed prothrombinase (Fig. 2). Thus, inhibition proceeds rapidly compared with the time scale of steady state initial velocity measurements. Furthermore, normalized inhibition by RNA_11F7t was not significantly affected by doubling the concentration of prethrombin 2 (Fig. 2). As the concentration of prethrombin 2 was well below the K_m (51), extent of inhibition normalized as V/Vo (Fig. 2) is expected to be most sensitive to substrate concentration if inhibition results from interference with substrate binding to prothrombinase. Therefore, RNA_11F7t is unlikely to compete for substrate binding to prothrombinase.

FIGURE 2.

Inhibition of prethrombin 2 activation. Initial velocities of thrombin formation were determined using increasing concentrations of RNA_11F7t (●, ○) or RNA_MUT (▴) in reaction mixtures containing 50 μm PCPS, 40 nm Va, 5 nm Xa, and either 1.4 μm (●, ▴) or 3.0 μm (○) prethrombin 2. The results of five different experiments are presented after normalization as v/v_o, where v_o is the initial rate observed in the absence of added RNA for each experiment.

Active Site Function

A hallmark of the unique strategy employed in the recognition of the protein substrate by prothrombinase is that although competitive inhibition requires ligands that block exosite-dependent tethering of the substrate to enzyme, those directed toward the active site of the catalyst uniformly act as classical non-competitive inhibitors (7). The latter possibility could readily explain the findings in Fig. 2. Initial velocity studies of peptidyl substrate hydrolysis were used to assess effects of RNA_11F7t on active site function of factor Xa in the absence or presence of the other constituents of prothrombinase (Fig. 3). Increasing concentrations of RNA_11F7t had no obvious effect on the rate of peptidyl substrate hydrolysis at ∼K_m concentrations of substrate when the catalyst was Xa alone or Xa saturated with PCPS (Fig. 3). In the absence of aptamer, initial rate was reduced by ∼50% when both PCPS and Va were present (Fig. 3). This decrease arises from the established ∼2-fold increase in K_m for SpXa that accompanies the incorporation of factor Xa into prothrombinase (21). Under these conditions, increasing concentrations of RNA_11F7t restored the initial rate to that observed in the absence of factor Va, whereas RNA_MUT had no such stimulatory effect (Fig. 3). The RNA_11F7t-dependent increase in rate varied with different fixed concentrations of Va (Fig. 3). Steady state kinetic constants confirmed that the modest increase in rate observed at 250 nm RNA_11F7t resulted from a correction of the larger K_m (205 ± 10 μm) observed in the presence of Va to a value (102 ± 6 μm) that was indistinguishable from that observed with Xa alone either in the presence or absence of PCPS. These results rule out the possibility that RNA_11F7t affects enzyme function by blocking or restricting substrate access to the active site of Xa or Xa within prothrombinase. The ability of RNA_11F7t to increase the rate in the presence of Va and PCPS and the peculiar dependence of this rate-enhancing effect on the fixed concentration of Va imply that the aptamer may interfere with the interaction between Xa and Va. However, definitive conclusions are precluded by the modest rate effects.

FIGURE 3.

Peptidyl substrate cleavage. Initial velocities for SpXa hydrolysis were determined with increasing concentrations of RNA_11F7t with 100 μm SpXa and 1 nm Xa (▴), 1 nm Xa, plus 50 μm PCPS (●) or 1 nm Xa, 50 μm PCPS, plus 5 nm Va (○), plus 15 nm Va (▵), or plus 50 nm Va (▿). Initial rates with increasing concentrations of RNA_MUT (▾) were determined using 1 nm Xa, 50 μm PCPS, and 50 nm Va. The lines were arbitrarily drawn.

Binding of Aptamer to Xa

Fluorescence measurements with OG₄₈₈-Xa revealed a large change in steady state intensity upon the addition of RNA_11F7t. Titration of OG₄₈₈-Xa with increasing concentrations of RNA_11F7t yielded a saturable increase in fluorescence with features of a tight binding interaction that saturated at ∼2 eq of added RNA_11F7t (Fig. 4). The titration curve was obviously sigmoid when performed in assay buffer with time-dependent decreases in intensity at low concentrations of aptamer (not shown). These features were greatly reduced when measurements were performed after preconditioning the cuvette with 0.2% (v/v) Tween 20 or when assay buffer contained 0.02% (v/v) Tween 20 (Fig. 4). We assume that inclusion of detergent resolves artifacts related to protein adsorption at low aptamer concentrations. However, we are uncertain if this manipulation has deliberately obscured intrinsic positive cooperativity in the binding interaction. Analysis assuming the interaction of RNA_11F7t with equivalent and non-interacting sites yielded a K_d ≅ 0.7 nm with ∼2 mol of aptamer bound bound/mol of OG₄₈₈-Xa at saturation (Fig. 4). No change in fluorescence intensity was observed when the buffer contained 50 μm EDTA in place of 5 mm Ca²⁺ or with increasing concentrations of RNA_MUT (Fig. 4). RNA_11F7t binding and/or the fluorescence change associated with the binding interaction requires the presence of Ca²⁺. The lack of fluorescence increase observed with increasing concentrations of RNA_MUT illustrates the specificity of the interaction inferred in this way. Reversibility of the inferred interaction was documented by the ability of increasing concentrations of Xa_i to effectively decrease fluorescence intensity of the preformed aptamer-OG₄₈₈-Xa complex to a value expected for OG₄₈₈-Xa in solution (Fig. 4, inset). Analysis of the competitive equilibria yielded parameters for the interaction of RNA_11F7t with Xa_i that were similar to those for its interaction with the fluorescent Xa derivative (Table 2). Therefore, the inferred binding of aptamer to OG₄₈₈-Xa is not affected in an obvious way by the presence of the bulky fluorophore or occlusion of the active site with the peptidyl chloromethyl ketone. Protein specificity was established by the inability of prothrombin to compete with OG₄₈₈-Xa for aptamer binding (Table 2). RNA_11F7t was found to bind equivalently to both X and Xa (Table 2), indicating the inability of the aptamer to distinguish between the zymogen and proteinase. Finally, despite the strong Ca²⁺ dependence of the binding of RNA_1F7t to OG₄₈₈-Xa, competitive binding studies with a proteolytic derivative of Xa_i lacking the γ-carboxyglutamic acid domain or recombinant derivatives of X and Xa lacking posttranslational modification with γ-carboxyglutamic revealed that affinity and stoichiometry for RNA_11F7t binding was independent of this region well known to participate in Ca²⁺ binding (Table 2).

FIGURE 4.

Binding of RNA species to factor Xa. Steady state fluorescence intensities were measured using reaction mixtures containing 18 nm OG₄₈₈-Xa titrated with increasing concentrations of RNA_11F7t (●) or RNA_MUT (○) in assay buffer containing 0.02% (v/v) Tween 20. Titrations with RNA_11F7t were also performed in the same buffer lacking Ca²⁺ but supplemented with 50 μm EDTA (▴). The line is drawn after analysis as described with the fitted values: K_d = 0.67 ± 0.06 nm, n = 2.13 ± 0.02 mol of RNA_11F7t/mol of OG₄₈₈-Xa at saturation, F_max/F_o = 1.76 ± 0.01, and F_o = 0.996 ± 0.003. Inset, competitive binding studies performed with 15 nm OG₄₈₈-Xa and increasing concentrations of Xa_i in the presence of 20 nm (▴), 40 nm (●), and 80 nm (○) RNA_11F7t. The lines are drawn after analysis according to competitive equilibria using the fitted parameters listed in Table 2.

TABLE 2.

Equilibrium constants for RNA_11F7t binding determined by fluorescence competition

Binding parameters were determined from competition studies similar to those illustrated in Fig. 4 inset. Fitted parameters are listed ±95% confidence limits.

Competitora	RNA11F7t/OG488-Xa interaction			RNA11F7t/competitor
	Kd	Mol RNA/mol OG488-Xa	Fmax/Fo	Kd	Mol RNA/Mol competitor
	nm			nm
Xai	1.26 ± 0.23	1.96 ± 0.04	1.76 ± 0.01	0.61 ± 0.10	2.08 ± 0.04
Prothrombin	NDb
X	0.79 ± 0.13	2.0 (F)c	1.75 ± 0.01	0.56 ± 0.08	1.89 ± 0.05
ΔGlaXai	0.42 ± 0.13	2.0 (F)	1.70 ± 0.01	0.35 ± 0.10	1.78 ± 0.05
desGlaXaS195A	1.12 ± 0.40	1.87 ± 0.04	1.75 ± 0.02	0.87 ± 0.28	2.05 ± 0.07
desGlaXS195A	1.56 ± 0.36	2.09 ± 0.05	1.78 ± 0.01	0.85 ± 0.19	1.88 ± 0.04

^a Fluorescence intensity was measured using increasing concentrations of the indicated competitor in the presence of OG₄₈₈-Xa and two or three different fixed concentrations of RNA_11F7t.

^b No competition detected.

^c Terms used as constants in the analysis are indicated as F.

Solution Molecular Weights

Analytical ultracentrifugation was used to assess dimerization of the aptamer in solution, which could provide an explanation for the stoichiometries consistently observed in the binding measurements. Concentration distributions of RNA_11F7t at sedimentation equilibrium could be adequately described by analysis according to a single sedimenting species with a molecular weight of 12,604 ± 109 (Fig. 5). Agreement with the formula weight of 11,827 suggests that the aptamer is predominantly monomeric even at μm concentrations. Although the modest discrepancy could reflect counterions bound to the oligonucleotide, uncertainty in v̄ most likely also contributes to this difference. Further analysis assuming a monomer-dimer equilibrium produced a small improvement in the fit (not shown), providing a lower limit estimate of 20 μm for the equilibrium dissociation constant for dimer formation. Thus, negligible concentrations of dimeric aptamer are expected to preexist at the concentrations used throughout this study. Additional efforts sought to independently establish the stoichiometry for the interaction between RNA_11F7t and desGlaXa_S195A by sedimentation equilibrium, employing a fixed concentration of desGlaXa_S195A (1.67 μm) and increasing concentrations of aptamer (0 to 3.2 μm, not shown). Global analysis of multiple data revealed two sedimenting species present in reciprocally varying amounts with increasing RNA_11F7t. Molecular weights of these species corresponded to 46,070 ± 390 for unligated desGlaXa_S195A and 67,550 ± 580, presumed to represent the aptamer-proteinase complex. Adequate fits of the sedimentation profiles did not require the inclusion of a significant amount of free aptamer. Although these findings are generally consistent with ∼2 mol of RNA_11F7t bound/mol of proteinase in a tight binding interaction, caution is suggested in interpretations. The molecular weight of the ligated species is heavily dependent on the partial specific volume of the product of the interaction of desGlaXa_S195A (v̄ = 0.721 ml/g) with aptamer (v̄ = 0.508 ml/g), which in turn was calculated assuming a stoichiometry of 2. Nevertheless, similar calculations assuming a stoichiometry of 1 yielded a larger molecular weight for the ligated species (73,700), which is even less consistent with 1 mol of aptamer bound/mol of desGlaXa_S195A at saturation.

FIGURE 5.

Sedimentation equilibrium analysis. Absorbance profiles were measured at 260 nm at sedimentation equilibrium after centrifugation of 1.62 μm RNA_11F7t in assay buffer lacking polyethylene glycol at 15,000 (▵), 20,000 (▴), 25,000 (○), and 30,000 (●) rpm at 20 °C. The lines are drawn after global analysis of scans at 260 and 280 nm with ρ = 1.0046 g/ml and v̄ = 0.508 ml/g (48) assuming a homogenous sedimenting species with molecular weight = 12,604 ± 109. Residuals to the fitted lines are shown in the upper panel. The root mean square deviation for this set was 0.0044.

Factor Va-dependent Inhibition of Prothrombinase Function

Evidence implicating a role for aptamer binding to Xa in affecting its interaction with factor Va was pursued by more detailed kinetic studies of prethrombin 2 cleavage. Initial rates determined with increasing concentrations of factor Va allow the indirect assessment of the membrane-dependent interaction between Xa and Va (21). Velocity curves were systematically shifted to the right at increasing fixed concentrations of RNA_11F7t, whereas RNA_MUT was without effect (Fig. 6). The data could be adequately described assuming competitive equilibria between the binding of aptamer to Xa and the binding of Xa to Va (Fig. 6). Inferred equilibrium binding parameters for the interaction of Xa with Va were in agreement with the findings of previous studies (21). Analysis assuming a stoichiometry of 2 mol of aptamer bound/mol of Xa at saturation yielded an equilibrium dissociation constant (∼3 nm) in tolerable agreement with the values determined in direct binding studies. The findings indicate that binding of RNA_11F7t to Xa competes with the ability of the proteinase to interact with factor Va within prothrombinase.

FIGURE 6.

Factor Va dependence of inhibition by RNA_11F7t. Initial velocities for thrombin formation were measured at increasing concentrations of Va in reaction mixtures containing 1.4 μm prethrombin 2, 50 μm PCPS, 5 nm Xa, and concentrations of RNA_11F7t fixed at 0 (●), 50 nm (▴), 200 nm (▵), 1.0 μm (▾), and 2.0 μm (▿). Initial velocities were also determined using 2.0 μm RNA_MUT as a control (○). The lines are drawn after analysis according to competitive equilibria with the fitted constants K_dXa,Va = 1.55 ± 0.3 nm, n = 1.19 ± 0.09 mol of Va/mol of Xa at saturation, K_dRNA,Xa = 3.14 ± 0.35 nm, n (assumed) = 2 mol of RNA_11F7t/mol of Xa at saturation, and rate at saturating Va = 83.0 ± 0.9 nm IIa formed/min.

Inhibition of Prothrombinase Assembly

Evidence for the ability of the aptamer to disrupt the assembly of prothrombinase was sought from equilibrium binding studies employing measurements of the fluorescence anisotropy of OG₄₈₈-Xa (21). In the absence of aptamer, increasing concentrations of Va produced a saturable increase in anisotropy (Fig. 7). This change reflects the binding of OG₄₈₈-Xa to factor Va on the membrane surface (21). Titration curves obtained at increasing concentrations of RNA_11F7t exhibited systematic shifts to the right, whereas RNA_MUT had no effect (Fig. 7). The data could be adequately described by the system of competitive equilibria involving mutually exclusive interactions between Va or RNA_11F7t with OG₄₈₈-Xa (Fig. 7). The fitted terms were consistent with previously established equilibrium parameters for the interaction between Xa and Va on the membrane surface (10, 21) as well as parameters for the interaction between aptamer and Xa established by multiple approaches in this study (Figs. 4 and 6, Tables 1 and 2). Because prothrombinase assembles through linked protein-protein and protein-membrane interactions (10), deleterious effects of RNA_11F7t on the binding of Xa to PCPS could also yield apparently competitive effects on prothrombinase assembly. Right angle light-scattering measurements of the binding of Xa to PCPS in the absence and presence of saturating concentrations of aptamer produced no evidence for a decrease in affinity for membrane binding by the Xa-aptamer complex (not shown). Indeed, the greater amplitude of the light-scattering signal observed in the presence of saturating concentrations of aptamer and PCPS was consistent with a larger species bound to membranes in the presence of RNA_11F7t. These findings exclude the potentially trivial possibility that interference with membrane binding after the interaction of RNA_11F7t with Xa accounts for the ability of the aptamer to inhibit the assembly of prothrombinase.

FIGURE 7.

Competitive inhibition of prothrombinase assembly. Fluorescence anisotropy was measured using 35 nm OG₄₈₈-Xa, 50 μm PCPS, and increasing concentrations of Va in the presence of fixed concentrations of RNA_11F7t corresponding to 0 (○), 0.35 μm (●), 2 μm (▵), and 5 μm (▴). The binding profile in the presence of 5 μm RNA_MUT (▾) is also illustrated as a control. The lines are drawn after analysis according to competitive equilibria using the fitted constants K_dXa,Va = 0.69 ± 0.2 nm, n = 1.08 ± 0.02 mol of Va/mol of OG₄₈₈-Xa at saturation, K_dRNA,Xa = 2.58 ± 1.2 nm, n = 2.2 ± 0.39 mol of RNA_11F7t/mol of OG₄₈₈-Xa at saturation, and Δr_max = 0.058 ± 0.004.

Other Functions of X and Xa

Surveys were undertaken to examine the effects of aptamer binding on other functions of factors X and Xa. The presence of ∼250 nm RNA_11F7t had a minor effect on the second order rate constant for the inhibition of factor Xa by antithrombin III (Table 3). This result is not out of the ordinary for a ligand that does not occlude access to the active site of the proteinase. In contrast, 250 nm RNA_11F7t greatly reduced the ability of the truncated two domain form of TFPI to inhibit Xa (Table 3). Thus, tight binding inhibition of Xa by this Kunitz inhibitor is mediated by interactions in addition to those at the active site of the proteinase. At K_m concentrations of factor X, RNA_11F7t had little effect on Xa formation by the VIIa·TF·PCPS complex but reduced Xa formation catalyzed by the IXa·VIIIa complex (Table 3). Initial velocity studies over a broader range of aptamer concentrations revealed little inhibition of Xa formation by the extrinsic Xase at concentrations as high as ∼1 μm RNA_11F7t, whereas Xa activation by the intrinsic Xase was reduced by ∼75% (Fig. 8). Initial velocity studies of Xa formation by IXa·VIIIa at increasing concentrations of substrate at different fixed concentrations of aptamer yielded complex curve shapes consistent with inhibition arising from aptamer binding to factor X, rendering it a substrate that is cleaved with a reduced V_max (not shown). Therefore, binding of RNA_11F7t differentially affects recognition of factor X by the two enzyme complexes responsible for its activation. Surprisingly, RNA_11F7t also inhibited VIII activation by factor Xa (Fig. 8, inset) without affecting cofactor activation by thrombin (not shown). Thus, features observed for the inhibition of the Xa-Va interaction are mirrored by effects on the interaction between Xa and TFPI as well as the interactions of X and Xa relevant to intrinsic Xase function.

TABLE 3.

Effects of RNA_11F7t on other functions of factors Xa and X

Measured reactiona	Parameter	−Aptamer	+RNA11F7tb
Xa + ATIII	kobs/I (m−1 s−1)	(1.78 ± 0.17) × 103	(1.54 ± 0.04) × 103
Xa + TFPI13–161	Ki* (nm)	0.35 ± 0.03	>600
X → Xa (VIIa.TF.PCPS)	Rate (nm Xa/min/nmE)	178	203
X → Xa (IXa.VIIa.PCPS)	Rate (nm Xa/min/nmE)	61	17.5

K_i*, overall equilibrium dissociation constant; E, enzyme.

^a The indicated reactions and corresponding parameters were determined as described under “Experimental Procedures.”

^b Results obtained using RNA_11F7t fixed at between 250 and 270 nm.

FIGURE 8.

Inhibition of X activation. Initial velocities for Xa formation were determined at increasing concentrations of RNA_11F7t for factor X activation catalyzed by the extrinsic Xase (●) or the intrinsic Xase (○). For measurements with extrinsic Xase, reaction mixtures contained 0.1 nm VIIa, 43 nm TF/50 μm PCPS, and 100 nm X (∼K_m). For measurements with intrinsic Xase, reaction mixtures contained 0.2 nm IXa, 10 nm VIIIa, 50 μm PCPS, and 40 nm X (∼K_m). Initial velocities were normalized as v/v_o to permit presentation in the same scale. Note the break in the ordinate between 0 and 0.25. Inset, shown is analysis of the activation of factor VIII (50 nm) with 3 nm Xa plus 50 μm PCPS in the absence (left panel) or the presence (right panel) of 250 nm RNA_11F7t determined by Western blotting with an anti-A2 antibody. In each panel, lanes 1-9 correspond to reaction times of 0, 20, 40, 60, 90, 120, 180, 300, and 480s.

DISCUSSION

The assembly of proteinase and cofactor into a membrane-bound enzyme complex profoundly enhances the rate of activation of the cognate zymogen substrate in several steps of coagulation (1, 2). These characteristics are exemplified by the activation of prothrombin by prothrombinase (1, 2). Our findings establish RNA_11F7t as a tight binding probe for factor Xa that competes with factor Va for the assembly of the proteinase into prothrombinase. Accordingly, the aptamer is effective at inhibiting thrombin formation without affecting access to the active site of factor Xa, impacting membrane binding or influencing substrate binding to the enzyme complex. Thus, RNA_11F7t is endowed with properties expected for a specific inhibitor of the protein-protein interaction between Xa and Va within prothrombinase.

Difficulties with incisive interference with the assembly of prothrombinase relate to the high affinity for the membrane-dependent interaction between Xa and Va (10), the obvious complexity of macromolecular interactions enjoying resurgent recognition (52), the contribution of linkage effects in the multicomponent interactions that stabilize prothrombinase (10) as well as the likelihood that interfering with a subset in a complex network of contacts may not yield efficient inhibition of protein-protein interactions relevant to function. Consequently, efficient competition for prothrombinase assembly is achieved by catalytically inactive derivatives of Xa that can participate in most or all aspects of the interactions of active Xa within prothrombinase (9). Inhibition of the Xa-Va interaction in the bovine system has also previously been documented with a monoclonal antibody directed toward the light chain of factor Va (53). Rather than purely arising from site-specific effects, inhibition in this case probably also reflects long-range steric phenomena associated with the binding of an immunoglobulin probe of roughly comparable size to the cofactor. In line with such reasoning, modest inhibitory effects, judged by affinity and/or extent, are observed with peptides presumed to represent regions that participate in interactions between proteinase and cofactor (15, 16). Our findings with RNA_11F7t are somewhat unexpected in that they illustrate competitive inhibition of prothrombinase assembly by a relatively small ligand that binds factor Xa with high affinity in an oligoribonucleotide sequence-specific interaction.

Analogous approaches have led to the development of inhibitory RNA aptamers targeting IXa, VIIa, and thrombin (39, 54, 55). These also bind their target proteinases with high affinity and interfere with complex macromolecular interactions necessary for function. Highly complementary interactions between RNA and protein, occluding ∼1200 Ų of solvent accessible surface, have been documented in the x-ray structure of thrombin bound to its inhibitory RNA aptamer (56). The ability of an otherwise modestly sized RNA aptamer to intimately engage extensive surfaces on the proteinase in this way may lie at the heart of the explanation for the repeated success of these reagents in inhibiting complex but diverse macromolecular interactions.

Mutually exclusive interactions between RNA_11F7t and factor Va for binding to Xa are most readily interpreted in terms of the ability of the aptamer to specifically occlude surfaces on the proteinase that participate in cofactor binding. Although such straightforward occlusive effects may well contribute to the inhibitory properties of RNA_11F7t, a role for alternative phenomena is implied by the finding that the aptamer binds equivalently to both X and Xa. As the zymogen, factor X, does not detectably bind factor Va (57), it follows that the binding of RNA_11F7t is probably not confined to structural motifs specific to the interaction between proteinase and cofactor. We offer the reasonable speculation that competitive inhibition of the Xa-Va interaction also reflects indirect or allosteric phenomena wherein RNA_11F7t distorts structural features in the proteinase that, in turn, deleteriously impact cofactor binding.

Complexity in the binding of RNA_11F7t to Xa is further implied by the peculiar stoichiometry and the strong Ca²⁺ dependence of the interaction. That the latter property is independent of the γ-carboxyglutamic acid-containing domain suggests a role for Ca²⁺ ligation at the site in the proteinase domain and/or in the first epidermal growth factor-like domain in modulating proteinase structure and impacting RNA_11F7t binding (58, 59). It is equally possible that the findings are unrelated to Ca²⁺ binding to either X or Xa but, rather, because of the stabilization of the inhibitory structure of the aptamer by bound Ca²⁺.

Despite the fact that RNA_11F7t is predominantly monomeric, all binding and functional studies are consistent with 2 mol of aptamer bound/mol of X or Xa at equivalent and apparently non-interacting sites. An alternative explanation is that there are two equally populated, non-interconverting and differentially folded structures of RNA_11F7t, only one of which is capable of binding Xa. In this case, stoichiometries based on total RNA_11F7t would be overestimated by 2-fold with a corresponding ∼2-fold overestimate in K_d. This possibility is suggested by the findings in the filter binding assay wherein 60–65% of radiolabeled aptamer was maximally retained as the bound species at saturating concentrations of Xa (not shown). In contrast, the analytical ultracentrifugation studies are more consistent with a tight binding stoichiometry of 2. However, the complexity of analyzing such interactions between components with vastly differing values of v̄ could adversely impact interpretations. We have been unable to convincingly distinguish between these possibilities either by affinity chromatography using immobilized Xa or by native gel electrophoresis. These points highlight the substantial difficulties inherent in the unambiguous assignment of stoichiometries that deviate significantly from 1 even when sound physical measurements are employed.

Finally, it is also possible that the Ca²⁺ dependence of binding and inhibition is somehow related to the observed stoichiometry of 2. It is possible that binding of RNA_11F7t to X and Xa could facilitate aptamer dimerization in a potentially Ca²⁺-dependent manner. This scenario could explain the inability to detect a singly ligated Xa species in sedimentation equilibrium studies or otherwise dissect contributions of the individual binding events to function. We have also been unable to test these and other possibilities because of repeated failures in UV-cross-linking attempts to covalently trap RNA_11F7t in complex with either X or Xa.

The conclusion that RNA_11F7t and Va represent mutually exclusive ligands for factor Xa is based on initial velocity studies as well as equilibrium binding measurements of prothrombinase assembly. In both cases the data were adequately described by competing equilibria in which Xa either binds to Va within prothrombinase or binds to the aptamer, with agreement between fitted parameters and directly measured terms for the individual component interactions. Despite the consistency between multiple approaches, it remains possible that the binding interactions are not strictly mutually exclusive. In this case our data indicate that the affinity of the RNA_11F7t·Xa complex for Va would be at least 100-fold lower than for Xa alone. There are no facile approaches to convincingly distinguish this possibility from the alternate case of competitive binding.

In addition to binding Xa and affecting its ability to interact with Va within prothrombinase, the interaction of RNA_11F7t selectively impacts other functions of factor X and Xa. These include the effects on the inhibition of Xa by TFPI and the functions of both X and Xa related to the intrinsic pathway. The dramatic effect of ∼250 nm RNA_11F7t in raising the overall equilibrium dissociation constant for inhibition by TFPI by ∼2000-fold is somewhat unexpected given the high affinity with which even the truncated inhibitor species binds factor Xa. We have not investigated this phenomenon in detail, and it remains possible that the unanticipated effects of the aptamer could contribute to the observations. In contrast, the effects of RNA_11F7t on X activation by the intrinsic Xase and the action of Xa on VIII are more in line with the effects observed on the assembly of prothrombinase. In the case of X activation by the IXa·VIIIa complex, aptamer binding to the zymogen transforms it into a poor substrate, likely with a 5–6-fold lower V_max, for the enzyme complex. Coupled with the lack of observable effects of the aptamer on X activation by the VIIa/TF complex, our findings point to fundamental differences in mechanism underlying the recognition of factor X as a substrate by the extrinsic versus intrinsic Xase complexes. There is some other experimental support for this idea (22, 60). Inhibition of VIII activation by Xa further suggests commonalities underlying the interaction of the proteinase with factor Va as a cofactor or with factor VIII as substrate. Together, the observations point to differential roles played by multiple types of extended surfaces in factor X and Xa in the different biological functions of the zymogen and proteinase.

A series of DNA and RNA aptamers have been described as specific inhibitors of the reactions of blood coagulation as well as the cell adhesive functions of von Willebrand factor with potential practical utility (39, 54, 55, 61, 62). The use of RNA aptamers as potential anticoagulants also benefits from the generalizable strategy developed for the preparation of a specific antidote that binds the aptamer and obviates its inhibitory properties (63). Additional considerations arise from the ability of RNA_11F7t to bind both zymogen and proteinase with similar affinity. This feature along with the Ca²⁺ dependence of its binding is a likely explanation for the high concentrations required to prolong clotting times after the incubation of the aptamer with citrated plasma. Nevertheless, therapeutic administration of RNA_11F7t would lead to its circulation in blood complexed with the precursor of the proteinase target. The analogous properties of nematode anticoagulant peptide C2 and experience with its therapeutic use establish the enzymologic and pharmacokinetic advantages of this feature (64). On the other hand, whereas some of the other effects of RNA_11F7t on the functions of X and Xa in the intrinsic pathway could further contribute to its anticoagulant properties, apparent abrogation of the inhibition of Xa by TFPI suggests a potential limitation in this regard. It also suggests that inhibition of coagulation by RNA_11F7t arises from the net effect of the aptamer on multiple steps in the pathways leading to thrombin formation. Perhaps this is responsible for the differential response observed in the two coagulation assay systems (Fig. 1A).

In summary, our studies establish RNA_11F7t as a tight binding ligand for factor Xa with the ability to compete for the interaction between proteinase and its cofactor. This property imbues the aptamer with the unique credential of representing a ligand, generated de novo by SELEX, that can efficiently block the assembly of the prothrombinase complex. Inhibition of complex assembly is achieved with high affinity despite the fact that the complex is stabilized by linked protein-protein and protein-membrane interactions as well as a high affinity interaction between proteinase and cofactor that likely involves extended surfaces. We suggest that this property likely reflects a significant contribution from more generalized allosteric effects of RNA_11F7t on factor Xa rather than solely resulting from the direct occlusion of specific interactions between proteinase and cofactor. Our findings document the feasibility of interfering with enzyme complex assembly as a viable strategy to specifically modulate the formation of thrombin as well as other proteinases in the analogous steps of coagulation.