Supporting Information

Files in this Data Supplement:

Fig. 3. Synthesis of cathepsin probe AX6429. i. H₂, 10% Pd/C, TsOH, MeOH (quantitative yield) ii. EDAC, HOBt, TEA, DCM (58% yield) iii. 4 N HCl/dioxane then 5'-TAMRA-OSu, TEA, DMSO (52% yield).

Fig. 4. Evaluation of Anti-TAMRA affinity capture method. (A) Mouse liver-soluble proteome was labeled with the FP-PEG-TAMRA probe, and processed and digested with trypsin as described in Materials and Methods. Before affinity capture of the labeled peptides, a sample was removed for CE analysis (red trace). Following isolation of the TAMRA-labeled peptides, the eluate from the anti-TAMRA affinity resin was diluted 10-fold to normalize to the precapture volume and analyzed by CE (red trace). The yield of the affinity capture appears to be near 100%. (B) FP-PEG-TAMRA-labeled elastase (0.6 pmol) and butyrycholinesterase (1.5 pmol) were added to 10 mg of mouse liver-soluble proteome (10 mg/ml). The samples were processed and digested, and the TAMRA-labeled peptides isolated as described in Materials and Methods. LC-MS/MS analysis of the eluted material indicated that no significant contamination of unlabeled material was present.

Fig. 5. Distribution of SEQUEST search scores from FP-PEG-TAMRA-labeled proteomic samples. Several mouse proteomes were labeled with the FP-PEG-TAMRA probe and processed as described in Materials and Methods for LC-MS/MS analysis. The MS/MS data were searched against publicly available mouse databases by using the SEQUEST algorithm. The results were then sorted based on whether the peptide matched to the MS/MS spectrum contained a probe modified serine hydrolase active nucleophile (maroon bars) or any other peptide (light blue bars). The two sets of results clearly exhibited distinct distributions of Xcorr values (A), suggesting the method was successfully identifying the probe-modified peptides A modified score was developed that eliminated the charge state and mass dependence of the Xcorr. This new score, ActivXcorr (B), significantly improved the ability to distinguish correct search results from incorrect results.

Fig. 6. Mass and charge dependence of Xcorr and ActivXcorr. The SEQUEST search results derived from a panel of FP-PEG-TAMRA-labeled mouse proteomes was divided into two pool representing either active-site serine-modified peptides (blue) or other peptides (red). Plotting the Xcorr values of these results against the mass of the identified peptides reveals a strong charge dependence of the Xcorr, with larger peptides generally yielding higher scores for both pools of results (A). Similarly, peptides with higher charge states typically yielded higher Xcorr values (C). Analysis of the results using a mass and charge normalized Xcorr, called ActivXcorr, illustrates that this new score eliminates the mass (B) and charge (D) dependence of the Xcorr. Additionally, the distinction between the serine hydrolase active-site peptides, which are presumably correct, and all other results was significantly better by using the ActivXcorr (differences between vertical bars representing SD of each data set in C and D).

Fig. 7. CE peak assigment. Fractions were collected during LC-MS/MS analysis by using a split flow line connected to a fluorescence detector and fraction collector. The base peak mass spectrum shows numerous peptide species eluting (A, blue trace). Extracted ion current from the peptide identified as Kallikrein 9 (A, red trace), indicates that this ion is the primary contributor to the peak at 61 min. CE analysis of the fraction collected from 61 to 61.5 min indicates a single major peak (B, blue trace), which could be aligned with the overall CE spectrum from submaxillary (B, red trace).

Fig. 8. The structure of nafamostat.