Biology Reference
In-Depth Information
the amino acid residues most essential for interaction and hence for the biological
functions of the protein.
10.2.6 Comparison of the functional architecture of early mammalian
protein C with the putative ancestor of all vitamin K-dependent
factors
Six of the 13 variable regions (VR's) in the model of early mammalian protein C
were derived by reference to the crystal structure of factor Xa. The similarity in
VR loop length between early mammalian protein C and extant factor X is con-
sistent with both proteins sharing a considerable period of evolutionary history. It
would seem reasonable to suppose that factor X (an activator of prothrombin and
therefore a procoagulant) and protein C (an anticoagulant by virtue of its
inhibitory action in inactivating factors Va and VIIIa) have coevolved in such a
way as to ensure that thrombin generation is appropriately promoted yet ade-
quately limited in response to hemostatic challenge.
Doolittle (1993), Patthy (1990), and Krawczak
et al
. (1996) all concluded that
protein C and prothrombin probably emerged at a similar stage of vertebrate evo-
lution. In the primordial vertebrate hemostatic system, one role for protein C
could have been to act as part of a negative feedback control mechanism. Since
both the catalytic efficiency of thrombin and the efficiency of activation of pro-
thrombin were presumably undergoing a continual process of optimization at
that time, there would have been an increasing requirement for an efficient nega-
tive regulator of thrombin.
Any change of function from the common ancestor of the vitamin K-dependent
factors to early mammalian protein C would have necessarily required changes in
the active site of the protease, thereby altering its substrate specificity from fib-
rinogen to the emerging co-factors VIIIa and Va. The 'accelerated evolution' of
active sites is thought to have been important in the diversification of the sub-
strate specificity of serine proteases, subsequent to gene duplication (Creighton
and Derby, 1989; Ohta, 1994). From inspection of our models, changes in the
active site region of protein C appear to have involved the elimination of the aryl
binding site (with the exception of residue W215 which is highly conserved
between all serine proteases) and the chemotactic binding site (
Figure 10.4
). Both
of these modifications resulted from the loss of two pincer-like highly mobile
insertion loops, Leu59-Asn62 and (to a lesser extent) Leu144-Gly150, originally
present in the vitamin K-dependent factor ancestral protein. Since both loops are
required for the catalytic activity of extant thrombin (Stubbs and Bode, 1993), it
may be inferred that the early mammalian ancestor of protein C had already lost
its ability to cleave fibrinogen. Moreover, the absence of those residues which in
thrombin bind protein C [with the exception of Leu73, Arg75, and Asp186A
(Tsiang
et al
., 1995) and see below] implies that early mammalian protein C was
probably not auto-catalytic.
In the early mammalian protein C molecule, the distribution of residues analo-
gous to the fibrinogen-binding site of extant thrombin was very different from
that predicted in the putative vitamin K-dependent factor ancestral protein
(
Figure 10.4
). In terms of evolutionary conservation, the fibrinogen-binding patch
