Functional phenotyping of human plasma using a 361-fluorogenic substrate biosensing microarray.

Abstract

A microarray presenting glycerol nanodroplets of fluorogenic peptide substrates was used as a biosensor for the detection of multiple enzyme activities within human plasma. Using 10 different plasma proteases (kallikrein, factor XIIa, factor XIa, factor IXa, factor VIIa, factor Xa, thrombin, activated protein C, uPA and plasmin) and a 361-compound fluorogenic substrate library (Ac-Ala-P3-P2-Arg-coumarin for P = all amino acids except Cys), a database was created for deconvoluting the relative activity of each individual enzyme signal in human plasma treated with various activators (calcium, kaolin, or uPA). Three separate deconvolution protocols were tested: searching for "optimal" sensing substrate sequences for a set of 5 enzymes and using these substrates to detect protease signals in plasma; ranking the "optimal" sensing substrates for 10 proteases using local error minimization, resulting in a set of substrates which were bundled via weighted averaging into a super-pixel that had biosensing properties not obtainable by any individual fluorogenic substrate; and treating each 361-element map measured for each plasma preparation as a weighted sum of the 10 maps obtained for the 10 purified enzymes using a global error minimization. The similarity of the results from these latter two protocols indicated that a small subset of <90 substrates contained the majority of biochemical information. The results were consistent with the state of the coagulation cascade expected when treated with the given activators. This method may allow development of future biosensors using minimal and non-specific markers. These substrates can be applied to real-time diagnostic biosensing of complex protease mixtures.