Predicted multiple selected reaction monitoring to screen activated drug-mediated modifications on human serum albumin Fumio Osaki a,b , Takaaki Goto a , Seon Hwa Lee a , Tomoyuki Oe a,⇑ a Department of Bio-analytical Chemistry, Graduate School of Pharmaceutical Sciences, Tohoku University, Miyagi 980-8578, Japan b Analysis and Pharmacokinetics Research Labs, Astellas Pharma, Inc., Ibaraki 305-8585, Japan article info Article history: Received 13 November 2013 Accepted 12 December 2013 Available online 19 December 2013 Keywords: Human serum albumin Mass spectrometry Predicted multiple selected reaction monitoring Activated drug Chemical modification Screening abstract Metabolic activation of drugs frequently generates electrophilic products that may undergo covalent binding to biological macromolecules, such as proteins and DNA. The resulting covalent adducts are of considerable concern in drug discovery and development. Several strategies for assessing the potential risks of candidate drugs have been reported. Of these, glutathione trapping is the most commonly used method together with mass spectrometry. Furthermore, drug-mediated protein modifications have been studied using serum albumin and CYP enzymes to clarify target amino acids and mechanism-based inhibition, respectively. In this article, we introduce a practical way to screen drug-mediated protein modifications. The method, referred to as ‘‘predicted multiple selected reaction monitoring,’’ is based on the selected reaction monitoring (SRM) strategy, but targets all possible chemically modified tryptic peptides. The creation of SRM lists may require patience; however, this strategy could facilitate more sensitive screening compared with the common strategy of data-dependent product ion scanning. Ketoprofen-N-hydroxysuccinimidyl ester (equivalent to glucuronide) and N-acetyl-p-benzoquinone imine (NAPQI) were allowed to react with human serum albumin as a model experiment. Using this strategy, 11 ketoprofen-adduction sites (at Lys 137, 195, 199, 212, 351, 402, 432, 436, 525, 536, and 541 ) and 1 NAPQI-adduction site (at Cys 34 ) were easily identified. Ó 2013 Elsevier Inc. All rights reserved. Metabolic activation of drugs frequently generates electrophilic products, such as quinones [1,2] and epoxides [3]. These reactive metabolites may undergo covalent binding to biological macro- molecules, such as proteins and DNA, and the resulting covalent adducts are of considerable concern in drug discovery and develop- ment. For instance, protein haptenation causes allergic reactions [4], adduction on CYP causes mechanism-based inhibition (MBI) [5–7], and adduction on DNA can result in genotoxicity [8]. Such adverse drug reactions are rare, but may create potentially fatal problems [9,10]. Several strategies are used in the pharmaceutical industry to assess the potential risks of candidate drugs. A covalent binding assay using radiolabeled candidate drugs can be used for in vivo experiments [11]. However, this method requires the prep- aration of radioisotope-labeled drugs and does not provide struc- tural information. In vitro trapping assays are more practical and widely performed, especially at early stages. The glutathione (c- glutamylcysteinylglycine; GSH) trapping method is the most com- monly used method, together with liquid chromatography (LC)/ mass spectrometry (MS) [12]. Furthermore, in combination with a stable isotope-labeled GSH, [1,2- 13 C 2 , 15 N 1 ]Gly-labeled GSH, it has been reported to facilitate screening by monitoring of doublet peaks of light/heavy GSH adducts [13,14]. However, GSH is insuffi- cient for trapping various reactive metabolites, because the sulfhy- dryl group is its only reactive functional group. For instance, GSH adducts from aldehydes, acylglucuronides, iminium ions, and ni- troso derivatives are rarely formed or are not stable enough for fur- ther LC/MS analyses [12]. Therefore, the combined use of cyanide anion (hard nucleophile) together with GSH (soft nucleophile) has been also performed to trap various kinds of electrophiles [15]. As an alternative, a GSH analogue, c-glutamylcysteinyllysine, has been reported as a trapping reagent containing both hard and soft nucleophiles [16]. In addition, a synthetic peptide containing multiple nucleophilic amino acids, namely, Gly-D-Tyr-D-Pro-D- Cys-D-Pro-D-His-D-Pro, has also been reported and is well designed to resist proteolysis using D-amino acids and Pro [17]. As another type of approach, drug-mediated protein modifications have been studied using CYP enzymes to clarify MBI [5–7] and human serum albumin (HSA) to identify the target amino acids. Especially, HSA is widely used as a model protein as summarized in Table 1 [18–32]. HSA (585 amino acids; 66.5 kDa; t 1/2 = 19 days) is the most abun- dant plasma protein in blood and constitutes about 60% of the total proteins [33]. Since HSA circulates freely throughout the body, it is believed to be one of the main targets of active drug metabolites 0003-2697/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.ab.2013.12.016 ⇑ Corresponding author. Fax: +81 22 795 6816. E-mail address: t-oe@mail.pharm.tohoku.ac.jp (T. Oe). Analytical Biochemistry 449 (2014) 59–67 Contents lists available at ScienceDirect Analytical Biochemistry journal homepage: www.elsevier.com/locate/yabio