Complete abolishment of coagulant activity in monomeric disulfide-deficient tissue factor

To the editor:

The initiator of coagulation, tissue factor (TF), resides on the cell surface in an inactive (cryptic) state that binds factor VIIa with reduced affinity and cannot activate factor X, or in an active (decrypted), conformation.^1,2  Protein disulfide isomerase-dependent regulation of the TF allosteric Cys¹⁸⁶-Cys²⁰⁹ disulfide has been proposed as a key event in TF decryption, but this remains controversial.^3-7  Although elimination of the TF disulfide by Cys²⁰⁹-to-Ala conversion reduces affinity for VIIa and abolishes procoagulant activity, Kothari et al have questioned the validity of the disulfide switching model of TF decryption.⁸  They showed that mutation of either Cys-residue to Ser led to 10-fold lower cellular TF expression compared with wild-type TF, but coagulant activity normalized per TF molecule at supraphysiologic VIIa concentrations was indistinguishable between mutants and wild-type TF.

We reasoned that excessive membrane surface in cells expressing lower levels of TF mutants led to an overestimation of procoagulant activity. To exclude this variable, we created BHK cells expressing similar amounts of surface wild-type TF (TF^WT) or TF containing single or double Cys-to-Ala mutations. Analysis of 40 clones per mutant showed that Cys¹⁸⁶ mutation led to intracellular retention, as described before,¹  while a pair of TF^WT and TF^Cys209Ala clonal cell lines with similar surface expression was identified. TF^Cys209Ala expression was approximately 60% of TF^WT (Figure 1A), supported by surface-immunostaining with an antibody with similar reactivity for TF^WT and TF^Cys209Ala (Figure 1B). In contrast, mAb-5G9, which has reduced affinity for disulfide-mutated TF, only recognized surface TF^WT. Consistent with previous results in HUVEC,¹  BHK cells expressing TF^WT efficiently activated X, whereas TF^Cys209Ala-expressing BHK cells did not differ from untransfected BHK cells, even at supraphysiologic VIIa concentrations (Figure 1C).

We next addressed the possibility that the differences observed in Kothari's⁸  and our study were because of the specific replacements for these Cys-residues and generated BHK cells stably expressing TF containing Cys-to-Ser conversions. Surface biotinylation identified TF^Cys209Ser and TF^{Cys186/209Ser} clones with expression equal to the TF^WT clone. Interestingly, TF^Cys209Ser showed substantial surface-exposed dimers compared with TF^Cys209Ala, while TF^{Cys186/209Ser} did not form dimers, as expected from a TF mutant lacking unpaired cysteines (Figure 1D). While procoagulant activity of the TF^{Cys186/209Ser} clone did not differ from untransfected BHK cells, TF^Cys209Ser showed residual activity (Figure 1E). To prevent dimerization, TF^Cys209Ser-expressing BHK cells were incubated with N-ethyl-maleimide. Blocking free thiols impaired dimerization and also completely abolished TF^Cys209Ser activity, while TF^WT activity was up-regulated (Figure 1F), consistent with previously reported effects of NEM on flippase-induced PS exposure.⁹ 

These data with controlled surface expression of TF support previous conclusions that disulfide-mutated TF is impaired in adopting a procoagulant conformation. Of note, Kothari et al observed dimers when expressing TF mutated at both Cys-residues.⁸  Because our data demonstrate that dimerization can increase procoagulant activity of single-free-thiol mutants, their assay conditions may have favored TF dimerization by mechanisms unrelated to the mutated Cys-residues. In conjunction with increased membrane availability, residual dimer activity may have led to overestimation of TF-specific procoagulant activity. Thus, we consider general dismissal of the disulfide switching hypothesis as proposed by Kothari et al, unjustified.