Meizothrombin: Zymogen or Proteinase? Slow, Ligand-Dependent Equilibration Between Equally Populated Zymogen-Like and Proteinase-Like Forms Explains Its Selectively Anticoagulant Function

Abstract 533 Prothrombinase activates prothrombin by catalyzing its ordered proteolysis at Arg320 followed by cleavage at Arg271. Initial cleavage at Arg320 yields the proteinase meizothrombin (mIIa), which accumulates abundantly as an intermediate before its conversion to thrombin (IIa). Although mIIa is a proteinase, it only acts on a limited subset of substrates cleaved by IIa. mIIa is considered an anticoagulant proteinase because it functions efficiently in protein C activation while exhibiting poor clotting activity or reactivity towards antithrombin III. This limited substrate repertoire of mIIa has remained enigmatic and is generally considered to lie in the retention of covalent linkage to the fragment 1.2 (F12) domain allowing for membrane binding. Instead, our recent findings with IIa, illustrating its ligand dependent interconversions between zymogen-like and proteinase-like states, predict that covalent linkage of F12 to the proteinase domain in mIIa would impart it with zymogen-like properties. We produced stable and active mIIa (mIIaQQQ) using a recombinant prothrombin variant in which the bonds susceptible to autolysis were rendered uncleavable by substitution of Arg at 155, 271 and 284 with Gln. Fluorescence stopped flow studies were pursued with the probes dansyl arginine 3-ethyl piperidine amide (DAPA) or Nα-dansyl-(p-guanidino)-L-phenylalanine-piperidide (I-2581) to characterize the binding of ligands to the active site of mIIaQQQ and IIa. Binding to IIa was rapid and consistent with a rate limiting, bimolecular interaction between probe and the active site of the proteinase. In contrast, traces with mIIaQQQ were distinctly biphasic with ∼50% of the fluorescence change occurring on the millisecond timescale followed by a slow second phase (∼50%) that occurred over several seconds. Global fitting indicated that the findings were consistent with a pre-equilibrium between two forms of mIIa, one which binds the active site probe with μM affinity and a second that binds with nM affinity. The two forms interconvert with forward and reverse rate constants of ∼2 s−1. We surmise that these reflect zymogen-like and proteinase-like forms that are equally populated and interconvert slowly with each other in a ligand-dependent fashion. Accordingly, the distributions of the two forms could be altered by ligands established to affect the transition of IIa between zymogen-like and proteinase-like states. The equilibrium distribution was altered to favor the zymogen-like form by decreasing Na+ to 0 at constant ionic strength. In contrast, soluble thrombomodulin (sTM) drove the equilibrium towards the proteinase-like state in a manner consistent with a 1:1 interaction between mIIaQQQ and sTM. Surprisingly, the pre-equilibrium was heavily dependent on covalent linkage with fragment 1 (F1) or its structural integrity. Proteolytic removal of F1, chelation of Ca2+ with EDTA or elimination of 4-carboxyglutamic acid modifications had a profound effect on forcing the enzyme into the proteinase-like state. Thus, the equilibrium distribution of mIIa between zymogen-like and proteinase-like forms is affected by F1 and its Ca2+-stabilized conformation despite the fact that this domain is expected to be distant from the catalytic site. Our findings shed unexpected light into the mechanisms underlying the peculiar activity profile of mIIa relative to IIa. Its ability to interconvert slowly and reversibly between equally populated zymogen-like and proteinase-like states lies at the heart of its properties. By driving it to proteinase, thrombomodulin imparts full activity to mIIa allowing for efficient function in the anticoagulant pathway. In contrast, more weakly binding substrates, inhibitors or ligands will be less effective at perturbing the equilibrium thereby allowing mIIa to persist in blood with reduced activity towards procoagulant substrates. The F1 domain participates in an unexpected way in enforcing these unique features of mIIa. By virtue of its essential role in modulating the equilibrium distribution between zymogen-like and proteinase-like states, we document a new function for F1 in its role as a zymogenizer of mIIa. Disclosures: No relevant conflicts of interest to declare.

Targeting Btk in Lymphoma: PCI-32765 Inhibits Tumor Growth in Mouse Lymphoma Models and a Fluorescent Analog of PCI-32765 Is an Active-Site Probe That Enables Assessment of Btk Inhibition In Vivo.

There is increasing evidence indicating that B-cell-receptor (BCR) signaling is required for survival of non-Hodgkin’s lymphoma (NHL) cells. Bruton’s tyrosine kinase (Btk) is required for BCR signaling and mutations that inactivate human Btk cause X-linked-agammaglobulinemia, a B-cell immunodeficiency. Although Btk functions selectively in B cells, the Btk active site is structurally similar to the active site in several Src and Abl kinases and as a result, there have been few highly selective small molecule inhibitors of Btk. We have developed a series of covalent Btk inhibitors that target Cys-481 in Btk and this approach results in increased potency and selectivity over related kinases that lack a Cys residue at this position (ChemMedChem 15:58). PCI-32765 is a Cys-481 targeting Btk inhibitor that has been optimized for potency, selectivity and pharmacokinetics. In cellular assays, PCI-32765 inhibits BCR-stimulus induced phosphorylation of Phospholipase-C-gamma, a Btk substrate, as well as downstream phosphorylation of Erk (IC50 < 100 nM). In addition, PCI-32765 induces apoptosis and inhibits proliferation in a subset of NHL cell lines including DHL-4, DHL-6, WSU-DLCL2, OCI-Ly10 and DOHH2 (IC50s = 0.6–1.6uM). We have used RNAi knockdown in DOHH2 cells as an independent method to confirm that Btk is required for lymphoma cell proliferation. In vivo, orally dosed PCI-32765 (50mg/kg) inhibits growth of DOHH2 and WSU-DLCL2 xenografts. PCI-32765 also prevents disease progression in a mouse collagen-induced arthritis model (12.5mg/kg PO), indicating that other B cell lineage diseases are sensitive to Btk inhibition. In order to further characterize the selectivity and in vivo potency of PCI-32765, we have developed PCI-33380, an active-site probe consisting of a covalent Btk inhibitor linked to the fluorophore Bodipy-FL. PCI-33380 binds to Btk and can be detected by flow cytometry or by denaturing gel electrophoresis of cell lysates. In cell lysates, the probe labels a single predominant band of the same molecular weight as Btk and this band is absent in cells from xid mice. Labeling of this band is inhibited (IC50=10nM) by a brief pre-treatment of cells with PCI-32765, indicating that the probe can be used to assess occupancy of Btk by a covalent inhibitor. We have used the probe to quantitate the inhibition of Btk by PCI-32765 in vivo. A single oral dose of PCI-32765 (10mg/kg) delivered to mice leads to rapid and complete inhibition of Btk in spleen. In addition, a single oral dose of PCI-32765 fully inhibits Btk in xenograft tumors and peripheral blood cells and this inhibition is maintained for up to 24hr. The Btk probe provides pharmacodynamic measurements that may allow optimization of dosing and schedule for in vivo studies and we are currently adapting the probe assays for use in monitoring the inhibition of Btk in human clinical trials.

Ordered Cleavage of Prothrombin by Prothrombinase Results from Constrained Binding and Preferential Active Site Engagement by One of the Two Cleavage Sites in Prothrombin.

The conversion of prothrombin to thrombin is a paradigm for proteolytic activation reactions wherein product is produced following cleavage at more than one site in the substrate. Human prothrombinase catalyzes thrombin formation by sequential cleavage of prothrombin at Arg320 followed by cleavage at Arg271. The largely ordered activation arises because Arg320 in intact prothrombin is hydrolyzed by prothrombinase with a Vmax that is ~30-fold greater than that for cleavage at Arg271 while substrate affinity is unchanged. Paradoxically, this phenomenon has been proposed to arise from the constrained binding of prothrombin to prothrombinase. The alternate proposal is that two interconverting forms of the enzyme differentially recognize and cleave the two sites in prothrombin. We have investigated substrate binding using prothrombinase assembled with a catalytically inactive recombinant factor Xa (XaS195A) in which Ser195 was replaced with Ala and a series of recombinant variants of prothrombin. The prothrombin variants included wild type prothrombin (IIWT) containing both cleavage sites, IIQ271 with Arg271 replaced with Gln and a single cleavable site at Arg320, IIQ320 with Arg320 replaced with Gln and a single cleavable site at Arg271 and the uncleavable IIQQ containing both Gln substitutions. Titration of XaS195A assembled into prothrombinase with increasing concentrations of the fluorescent active site probe p-aminobenzamidine (PAB) produced a saturable increase in fluorescence confirming the expected ability of XaS195A to bind ligands, such as PAB, at the active site despite its lack of catalytic activity. The addition of a saturating concentration of IIWT to reaction mixtures containing PAB and prothrombinase assembled with XaS195A resulted in a decrease in PAB fluorescence arising from the engagement of the active site by the substrate and the associated displacement of PAB. Only a minor change in PAB fluorescence was observed with IIQQ even though this uncleavable derivative binds to prothrombinase with the same affinity as IIWT. Thus, features of the substrate-enzyme interaction that determine affinity are independent of active site engagement by the substrate. Saturating concentrations of IIQ271 displaced PAB from the active site of XaS195A within prothrombinase to the same extent observed with IIWT while essentially no fluorescence decrease was observed with IIQ320. It follows that although all prothrombin derivatives bind with similar affinity to prothrombinase, elements flanking Arg320 can readily engage the active site of XaS195A within the enzyme complex while those flanking the Arg271 site cannot. Our equilibrium binding measurements establish a primary role for active site binding in determining the perceived rate constant for catalysis without influencing substrate affinity. The differential action of prothrombinase on the two cleavage sites in prothrombin likely arises from constraints imposed by active site-independent interactions that drive substrate affinity and permit active site docking by the Arg320 site in the substrate but not the Arg271 site. These constraints in active site engagement explain asymmetry in the recognition of the two bonds in intact prothrombin and its ordered cleavage by prothrombinase.

Evaluation of the Specificity and Cytotoxicity of Three Proteasome Inhibitors.

The proteasome is a mutlicatalytic protease with three main catalytic activities - chymotrypsin-like (CT-L), trypsin-like (T-L) and peptidylglutamyl peptide hydrolising (PGPH). Proteasome inhibition is an emerging therapy for many cancers and is a novel treatment for multiple myeloma (MM). The CT-L activity, considered to be the rate-limiting step in protein degradation, is the primary target of many proteasome inhibitors. We have compared the specificity and potency of the novel proteasome inhibitor BzLLLCOCHO to the previously characterised inhibitors PS-341 (Velcade, bortezomib) and MG-132. Specific fluorogenic substrates were used to measure proteasome proteolytic activity in the presence and absence of the inhibitory compounds. An active site directed probe with a dansyl-sulfonamidohexanoyl hapten tag was used in conjuction with immunoblotting to determine the subunit specificity of the proteasome inhibitors (Nature Methods2005;2:357–362). MM cell lines (U266, OPM-2, KMS-11, KMS-18) were incubated with 10 μM BzLLLCOCHO, 5 nM PS-341 or 1 μM MG-132 for 24 hrs and proteasome activity was measured. Addition of BzLLLCOCHO reduced CT-L activity by 83 ± 13 % in the fluorogenic assay, and T-L and PGPH activities were reduced by 93 ± 6 % and 92 ± 2 % respectively. Immunoblot results revealed a similar pattern, the T-L and PGPH subunits were completely inhibited by BzLLLCOCHO and there was only weak labeling of the CT-L subunit with the active site probe. In contrast, treatment with PS-341 completely inhibited the CT-L and PGPH activities and incubation with MG-132 resulted in weak inhibition of the CT-L and PGPH activities, neither inihibitor significantly affected T-L activity. The ability of the different inihibitors to induce apoptosis in MM cell lines was then evaluated. All three inhibitors were demonstrated to act through both the caspase-8 and caspase-9 signalling pathways. Using Mitosensor™ and Hoescht/Propidium Iodide staining we found that MM cells were more sensitive to the induction of apoptosis by PS-341 and MG-132 than BzLLLCOCHO (U266 cells treated for 72 hrs with BzLLLCOCHO 51 % apoptosis, PS-341 79 % apoptosis and MG-132 84 % apoptosis). BzLLLCOCHO is a cell permeable and potent inhibitor of all three proteolytic activities of the proteasome. PS-341 and MG-132 inhibited only two of the three proteasome activities but were more efficient than BzLLLCOCHO at inducing apoptosis in MM cell lines. MG-132 is known to inhibit non proteasomal proteases such as Cathepsin B and Calpain 1 which may contribute to its potency. Further investigation on the effects of these inhibitors on gene and protein expression in the cell may lead to the development of more specific and targeted inhibitors.