A

FIGURE 2 (See color insert) (A) Computer-generated model (WebLab ViewerPro; Molecular Simulation Inc., San Diego, CA) of the human PF4 tetramer, based on the crystallographic coordinates. (B) AC dimer view of human PF4: the amino acid residues crucial for heparin binding are displayed. Abbreviation: PF4, platelet factor 4. Source: (A) from Zhang et al., 1994; (B), from Loscalzo et al., 1985; Mayo et al., 1995a.

lysine, arginine, and histidine residues that encircle the tetramer along a line perpendicular to the a-helices and are available for interaction with solvent. Modeling studies (Stuckey et al., 1992) support the possibility that a negatively charged heparin molecule, containing 18 saccharide residues (MW ~ 5.4 kDa), interacts with these positively charged residues spanning about half the tetramer. Mayo et al. (1995a,b) created a PF4 mutant (PF4-M2) in which the NH2-terminal 11 residues were replaced by eight residues from the homologous CXC chemokine IL-8 to produce a tetramer that binds heparin with the same avidity as native PF4 but is more nearly symmetrical around all three axes, facilitating nuclear magnetic resonance (NMR) structural analysis. Their data, contrary to PF4-heparin-binding models that center around COOH-terminal a-helix lysines, indicate that arginines 20, 22, and 49, and to a lesser extent, histidine 23, threonine 25, and lysine 46, are also important for heparin binding (Fig. 2b and Fig. 3). On the basis of these findings, it was speculated that heparin does not bind perpendicularly to the a-helices of the AB dimer as had been suggested (Stuckey et al., 1992) but instead reacts with the a-helix at an angle, interacting preferentially with PF4 along the AD dimer, where it would encounter arginine and other positively charged residues. In either model, it is plausible that binding of a linear polyanion of sufficient length and linear charge density to positively charged residues on the surface of PF4 could cause the structural rearrangement throughout the entire tetramer necessary for generation of HIT antibody epitopes.

On the basis of these reports, it is possible to propose a model of how heparin and other linear polyanions react with PF4 to produce configurational changes in the tetramer and create sites for HIT antibody binding. We suggest that linear polyanions, such as heparin, that carry appropriately spaced, strong negative charges interact with PF4 by binding to the ring of positive charges extending between the A and D or B and C subunits or both. The minimum length for a fully active polyanion is about 50 A, equivalent to six disaccharide subunits (12-mer), with each disaccharide measuring about 8.4 A in length (Visentin et al., 2001). Reconfiguration of the tetramer, resulting from binding of the polyanion, creates the neoepitope(s) for which HIT antibodies are specific.

0 0

Post a comment