Optimization of plasma-based BioID identifies plasminogen as a ligand of ADAMTS13.

ADAMTS13, a disintegrin and metalloprotease with a thrombospondin type 1 motif, member 13, regulates the length of Von Willebrand factor (VWF) multimers and their platelet-binding activity. ADAMTS13 is constitutively secreted as an active protease and is not inhibited by circulating protease inhibitors. Therefore, the mechanisms that regulate ADAMTS13 protease activity are unknown. We performed an unbiased proteomics screen to identify ligands of ADAMTS13 by optimizing the application of BioID to plasma. Plasma BioID identified 5 plasma proteins significantly labeled by the ADAMTS13-birA* fusion, including VWF and plasminogen. Glu-plasminogen, Lys-plasminogen, mini-plasminogen, and apo(a) bound ADAMTS13 with high affinity, whereas micro-plasminogen did not. None of the plasminogen variants or apo(a) bound to a C-terminal truncation variant of ADAMTS13 (MDTCS). The binding of plasminogen to ADAMTS13 was attenuated by tranexamic acid or ε-aminocaproic acid, and tranexamic acid protected ADAMTS13 from plasmin degradation. These data demonstrate that plasminogen is an important ligand of ADAMTS13 in plasma by binding to the C-terminus of ADAMTS13. Plasmin proteolytically degrades ADAMTS13 in a lysine-dependent manner, which may contribute to its regulation. Adapting BioID to identify protein-interaction networks in plasma provides a powerful new tool to study protease regulation in the cardiovascular system.

Metalloprotease domain latency protects ADAMTS13 against broad-spectrum inhibitors of metalloproteases while maintaining activity toward VWF.

BACKGROUND: ADAMTS13 is a circulating metalloprotease that cleaves von Willebrand factor (VWF) in a shear-dependent manner. ADAMTS13 is secreted as an active protease but has a long half-life, suggesting that it is resistant to circulating protease inhibitors. These zymogen-like properties indicate that ADAMTS13 exists as a latent protease that is activated by its substrate. OBJECTIVES: To investigate the mechanism of ADAMTS13 latency and resistance to metalloprotease inhibitors. METHODS: Probe the active site of ADAMTS13 and variants using alpha-2 macroglobulin (A2M), tissue inhibitors of metalloproteases (TIMPs), and Marimastat. RESULTS: ADAMTS13 and C-terminal deletion mutants are not inhibited by A2M, TIMPs, or Marimastat, but cleave FRETS-VWF73, suggesting that the metalloprotease domain is latent in the absence of substrate. Within the metalloprotease domain, mutating the gatekeeper triad (R193, D217, D252) or substituting the calcium-binding (R180-R193) or the variable (G236-S263) loops with corresponding features from ADAMTS5 did not sensitize MDTCS to inhibition. However, substituting the calcium-binding loop and an extended variable loop (G236-S263) corresponding to the S1-S1' pockets with those from ADAMTS5, resulted in MDTCS-GVC5 inhibition by Marimastat, but not by A2M or TIMP3. Substituting the MD domains of ADAMTS5 into full-length ADAMTS13 resulted in a 50-fold reduction in activity compared with the substitution into MDTCS. However, both chimeras were susceptible to inhibition, suggesting that the closed conformation does not contribute to the latency of the metalloprotease domain. CONCLUSION: The metalloprotease domain protects ADAMTS13 from inhibitors and exists in a latent state that is partially maintained by loops flanking the S1 and S1' specificity pockets.