Differential Binding of Plasminogen, Plasmin, and Angiostatin4.5 to Cell Surface β-Actin: Implications for Cancer-Mediated Angiogenesis.

Angiostatin4.5 (AS4.5), the naturally occurring human angiostatin isoform, is the product of plasmin autoproteolysis and consists of kringles 1 to 4 and ~85% of kringle 5. An important question related to tumor angiogenesis is the mechanism by which tumor cells convert plasminogen to angiostatin locally and yet the angiostatin acts systemically. We recently reported that cancer cell surface globular β-actin mediates plasmin autoproteolysis to AS4.5. We now characterize the binding of plasminogen, plasmin and AS4.5 to globular β-actin. Plasminogen and plasmin bind comparably to immobilized globular β-actin, with a KD of approximately 140 nM. The binding is inhibited by ε-aminocaproic acid (EACA), a lysine analog, indicating the requirement for a lysine-kringle domain interaction. Using a series of peptides derived from β-actin in competitive binding studies, we show that the actin domain necessary for binding of plasminogen is within “GDEAQSKRGILTLKY”, (amino acids 55 to 69), which includes lys61 and lys68. Substitution of lys61 or Lys68 with arginine results in the loss of the peptide’s ability to block plasminogen binding, indicating that lys61 and Lys68 are necessary for plasminogen binding. Other actin peptides, including peptides with other lysines, did not inhibit the plasminogen-actin interaction. While plasminogen and plasmin bind to actin efficiently, AS4.5 did not bind, even at high concentrations (500 nM). Plasminogen, plasmin, and AS4.5 all contain kringles 1–4, however, kringle 5 is truncated in AS4.5, suggesting that the kringle 5 is the domain by which plasminogen and plasmin bind to actin. We thus tested pure recombinant kringle 5 for its ability to specifically bind actin. Isolated kringle 5 was shown to bind to actin supporting the hypothesis that intact kringle 5 is necessary for plasminogen and plasmin to bind to cell surface β-actin, and the truncated kringle 5 in AS4.5 results in its release from β-actin. We propose the following paradigm: Plasminogen binds cell surface globular β-actin, where uPA mediates activation to plasmin which remains bound to actin. Actin then mediates plasmin autoproteolysis, with an intra-kringle 5 cleavage, resulting in AS4.5 formation. With the truncation of kringle 5, the AS4.5 dissociates from the surface actin and enters the systemic circulation. These data may help explain why AS4.5 is formed locally, on cancer cell surfaces, and yet acts systemically.