Interest in using peptide molecules as therapeutic agents due to high selectivity and efficacy is increasing within the pharmaceutical industry. However, most peptide-derived drugs cannot be administered orally because of low bioavailability and instability in the gastrointestinal tract due to protease activity. Therefore, structural modifications peptides are required to improve their stability. For this purpose, several in-silico software tools have been developed such as PeptideCutter or PoPS, which aim to predict peptide cleavage sites for different proteases. Moreover, several databases exist where this information is collected and stored from public sources such as MEROPS and ExPASy ENZYME databases. These tools can help design a peptide drug with increased stability against proteolysis, though they are limited to natural amino acids or cannot process cyclic peptides, for example. We worked to develop a new methodology to analyze peptide structure and amide bond metabolic stability based on the peptide structure (linear/cyclic, natural/unnatural amino acids). This approach used liquid chromatography / high resolution, mass spectrometry to obtain the analytical data from in vitro incubations. We collected experimental data for a set (linear/cyclic, natural/unnatural amino acids) of fourteen peptide drugs and four substrate peptides incubated with different proteolytic media: trypsin, chymotrypsin, pepsin, pancreatic elastase, dipeptidyl peptidase-4 and neprilysin. Mass spectrometry data was analyzed to find metabolites and determine their structures, then all the results were stored in a chemically aware manner, which allows us to compute the peptide bond susceptibility by using a frequency analysis of the metabolic-liable bonds. In total 132 metabolites were found from the various in vitro conditions tested resulting in 77 distinct cleavage sites. The most frequent observed cleavage sites agreed with those reported in the literature. The main advantages of the developed approach are the abilities to elucidate metabolite structure of cyclic peptides and those containing unnatural amino acids, store processed information in a searchable format within a database leading to frequency analysis of the labile sites for the analyzed peptides. The presented algorithm may be useful to optimize peptide drug properties with regards to cleavage sites, stability, metabolism and degradation products in drug discovery.

Rationale: Liquid chromatography/mass spectrometry is an essential tool for efficient and reliable quantitative and qualitative analysis and underpins much of contemporary drug metabolism and pharmacokinetics. Data-independent acquisition methods such as MS E have reduced the potential to miss metabolites, but do not formally generate quadrupole-resolved product ion spectra. The addition of ion mobility separation to these approaches, for example, in High-Definition MS E (HDMS E ) has the potential to reduce the time needed to set up an experiment and maximize the chance that all metabolites present can be resolved and characterized.We compared High-Definition Data-Dependent Acquisition (HD-DDA), MS E and HDMS E approaches using automated software processing with Mass-MetaSite and WebMetabase. Methods:Metabolite identification was performed on incubations of glucagon-like peptide-1 (7-37) (GLP-1) and verapamil hydrochloride. The HD-DDA, MS E and HDMS E experiments were conducted on a Waters ACQUITY UPLC I-Class LC system with a VION IMS quadrupole time-of-flight (QTOF) mass spectrometer operating under UNIFI control. All acquired data were processed using MassMetaSite able to read data from UNIFI 1.9.4. WebMetabase was used to review the detected chromatographic peaks and the spectral data interpretations. Results:A comparison of outcomes obtained for MS E and HDMS E data demonstrated that the same structures were proposed for metabolites of both verapamil and GLP-1. The ratio of structurally matched to mismatched product ions found by MassMetaSite was slightly greater for HDMS E than for MS E , and HD-DDA, thus improving confidence in the structures proposed through the addition of ion mobility based data acquisitions . Conclusions: HDMS E data acquisition is an effective approach for the elucidation of metabolite structures for both small molecules and peptides, with excellent accuracy and quality, requiring minimal tailoring for the compound under investigation.

Peptide drugs have been used in the treatment of multiple pathologies. During peptide discovery, it is crucially important to be able to map the potential sites of cleavages of the proteases. This knowledge is used to later chemically modify the peptide drug to adapt it for the therapeutic use, making peptide stable against individual proteases or in complex medias. In some other cases it needed to make it specifically unstable for some proteases, as peptides could be used as a system to target delivery drugs on specific tissues or cells. The information about proteases, their sites of cleavages and substrates are widely spread across publications and collected in databases such as MEROPS. Therefore, it is possible to develop models to improve the understanding of the potential peptide drug proteolysis. We propose a new workflow to derive protease specificity rules and predict the potential scissile bonds in peptides for individual proteases. WebMetabase stores the information from experimental or external sources in a chemically aware database where each peptide and site of cleavage is represented as a sequence of structural blocks connected by amide bonds and characterized by its physicochemical properties described by Volsurf descriptors. Thus, this methodology could be applied in the case of non-standard amino acid. A frequency analysis can be performed in WebMetabase to discover the most frequent cleavage sites. These results were used to train several models using logistic regression, support vector machine and ensemble tree classifiers to map cleavage sites for several human proteases from four different families (serine, cysteine, aspartic and matrix metalloproteases). Finally, we compared the predictive performance of the developed models with other available public tools PROSPERous and SitePrediction.