Although both talin and kindlin-3 binding to the β3 cytoplasmic domain are required for agonist-induced αIIbβ3 activation in platelets, the biochemical basis for this dual requirement is not clear. Recent NMR and hydrogen-deuterium exchange studies of disulfide-stabilized complexes containing the full cytoplasmic domains of αIIb and β3 dispersed in detergent micelles or lipid bilayers revealed that the β3 cytoplasmic domain consists of three helices: a stable proximal helix contiguous with the transmembrane domain and two distal dynamic amphiphilic helices whose fluctuations allow interaction of the helices with lipid bilayers or cytoplasmic proteins. These results suggest a cooperative model for talin and kindlin-3 binding to β3 with the talin and kindlin-3 binding sites kinetically- and thermodynamically-linked. Whether there is a preferred temporal sequence for kindlin-3 versus talin binding to β3 during physiological αIIbβ3 activation in platelets is not known, but the greater mobility of the kindlin-3 binding site suggests it might have a kinetic advantage over talin, assuming both are present in appropriate forms for binding. Much is known about the structure and function of talin, but substantially less is known about kindlin-3 in part because it has not been possible to express the complete molecule in bacterial expression systems. To address this issue, we have examined kindlin-3 expression and function in human platelets. Two kindlin-3 isoforms have been identified, a long form (Mr ∼ 76 kDa; accession: NM_178443) and a short form (Mr ∼75 kDa accession: NM_031471), that differ by the presence of 4 residues (RIPR; residues 360–363) in the PH domain of the long isoform. Using highly purified platelet and leukocyte RNA and RT PCR, we found that kindlin-3 expressed in platelets and leukocytes consist almost entirely of the 663 amino acid short isoform. Kindlin-3 present in platelet extracts is functionally active and spontaneously binds to the β3 cytoplasmic domain in pull down assays. Moreover, because kindlin-3 binding is abrogated by the β3 mutations S752P and T759A, but not T759F, the in vitro binding of kindlin-3 does not appear to be phosphorylation dependent. Further, surface plasmon resonance spectroscopy suggests that the PH domain of kindlin-3 partially drives membrane binding in the presence of phospholipids. Treatment of washed human platelets with the PAR1-activating peptide TRAP-6 (SFLLRNP) resulted in the rapid incorporation of kindlin-3 present in the platelet cytosol into the platelet cytoskeleton. We also found that kindlin-3 is present in dense fractions when platelets extracts were fractionated on sucrose gradients. Likewise, immunofluorescent images of platelets adherent to fibrinogen and platelet immuno-electron microscopy detected the presence of kindlin-3 in vesicular structures. Finally, using out-dated human platelets as starting material, we purified kindlin-3 to near homogeneity by the sequential use of ion-exchange and gel-filtration chromatography. These results provide a foundation for understanding the unique role that kindlin-3 plays in regulating the activity of platelet αIIbβ3.
No relevant conflicts of interest to declare.
Asterisk with author names denotes non-ASH members.