A structural spotlight on extrinsic Xase complex Free

In this issue of Blood, Sedzro et al¹ report the cryoelectron microscopy (cryo-EM) structure of tissue factor (TF)/factor VIIa (FVIIa) bound to a substrate factor X (FX) mimetic on the membrane surface. This study represents a significant step forward in understanding the intricate interactions of the TF/FVIIa complex and the substrate on membrane surface, which are critical for the initiation and regulation of blood coagulation and is a follow-up to their recent molecular dynamic simulation studies.² 

The coagulation response to vascular damage is achieved by interconnected and regulated series of complex protein reactions leading to the generation of thrombin, and ultimately the formation of a blood clot. Interactions between coagulation serine proteases, their cofactors, Ca²⁺ ions and anionic membrane surfaces is of a central importance in the regulation of hemostasis.³ The TF-FVIIa complex (also known as the extrinsic Xase) serves as a good example of a cofactor-mediated modulation of a serine protease activity.⁴ This complex is the physiological initiator of the extrinsic pathway of blood coagulation and is tightly controlled by tissue factor pathway inhibitor (TFPI) to prevent excessive coagulation which can lead to thrombosis.⁴ 

The extrinsic Xase complex consists of a trypsin-like serine protease, FVIIa, and a cofactor, TF, which is an integral transmembrane protein, exposed on different cellular surfaces following injury.^3,4 TF recruits and facilitates the activation of the plasma zymogen FVII and acts as a nonenzymatic cofactor within the extrinsic Xase complex.⁴ TF/FVIIa assembled on phosphatidylserine-containing membranes activates the zymogen substrates, factor IX and X, 2 major enzymes responsible for the generation of thrombin. The preferred substrate, FX is cleaved by TF/FVIIa at a highly conserved site (Arg15-Ile16 in chymotrypsin numbering system) resulting in the release of the activation peptide and formation of FXa.⁵ FVIIa itself has insignificant activity toward FX, however, binding to the membrane-bound cofactor, TF enhances its catalytic activity by several orders of magnitude.^3,4 

The combination of mutagenesis, enzyme kinetics, and structural studies have provided important insights into the specific interaction sites of the TF/FVIIa complex. Although the crystal structures of soluble TF alone or in complex with FVIIa have been available since the mid-1990s,^6,7 a high-resolution crystal structure of the ternary complex in the presence of a phospholipid membrane has not been resolved until now.

Sezdro et al employed a spectrum of approaches to solve the cryo-EM structure of TF/FVIIa complex bound to a substrate mimetic on a membrane surface. This includes the use of a phospholipid membrane bilayer (nanodiscs), a FX surrogate (XK1), and a noninhibitory anti-TF Fab fragment (10H10) antibody. To create a stable complex, the authors used a FX mimetic, XK1, a hybrid protein, in which the FX protease domain was swapped with the first Kunitz-type (K1) domain of TFPI. They documented that K1-swapped hybrid FX interacts with membrane-bound TF/FVIIa similar to native FX, likely mirroring the physiological contacts within the TF/FVIIa/FX complex on membrane surface.

TF/FVIIa/XK1 complexes assembled on membrane nanodiscs bound to a 10H10 Fab were subsequentially used for structure determination using cryo-EM analysis. The resulting structure had a high enough resolution (3.7 Å) to permit modeling of the protein backbones. In agreement with a previous fluorescence resonance energy transfer study,⁸ TF/FVIIa (“mushroom shaped”) is positioned perpendicular to the nanodisc enabling optimal alignment of FVIIa active site. Earlier findings support a role of the epidermal growth factor–like (EGF) domains of FX in binding of TF/FVIIa.⁹ In contrast, the current crystal structure suggests that the light chain of XK1 shows a high degree of flexibility, adopting a “handle-shape,” with 1 point of contact between K1 domain of XK1 and the active site of FVIIa, and another between the γ-carboxyglutamic acid (GLA) domain of XK1 and TF’s substrate-binding exosite located near the membrane surface. Apart from a minor contact between EGF-1 domain of XK1 and TF, most of the EGF-1 and all EGF-2 domains do not bind TF/FVIIa. The FX and FVIIa GLA domains also interact with each other and the membrane surface. These findings support the concept that FX binds specifically with TF/FVIIa but not too tight to prevent FXa dissociation from the complex.

This study also revealed a previously unknown, membrane-dependent allosteric mechanism that mediates the function of TF exosite region in engaging FX as a substrate. The structure shows that a specific loop (4xSer) in TF partially blocks the TF exosite and must undergo a shift to expose a hidden FX GLA domain binding site to the membrane-bound TF/FVIIa complex. The functional importance of this loop is supported by the observations that mutations or deletions of loop’s residues markedly impact the FX activation rate.¹⁰ This membrane-dependent allosteric mechanism could also clarify the enigmatic phenomenon of TF encryption/decryption.

The crystal structure of membrane-bound TF/FVIIa/XK1 reported by Sedzro al significantly advances our understanding of a key regulatory step in the initiation of blood coagulation. Some limitations of this study could be related to the use of XK1 and not of the natural substrate, FX; nevertheless, it provides new insights into how the FX light chain engages the TF/FVIIa and unravels a novel membrane-dependent allosteric activation mechanism between TF and FVIIa, that contributes to the binding of FX Gla domain. This study opens new research avenues that are necessary to fully elucidate the roles of the TF/FVIIa in hemostasis and thrombosis.

Conflict-of-interest disclosure: The author declares no competing financial interests.