ABSTRACT: Excessive oxidative stress, which is associated with several diseases, can trigger oxidative protein modifications, including protein carbonylation, which refers to the oxidative formation of aldehyde or keto groups in proteins. The structural diversity of these non-enzymatic reactive carbonyl groups and their low abundance pose a significant challenge to their detection and quantitation. Here, we report a strategy to identify and quantify protein carbonylation in human serum samples from 39 patients diagnosed with rheumatoid arthritis and 29 healthy donors. Reactive carbonyl groups were derivatized with aldehyde reactive probe (ARP), digested with trypsin, enriched via the biotin in ARP by avidin affinity chromatography, and analyzed by RP-HPLC-ESI-IMS-MS/MS. Ion mobility spectrometry (IMS) was used in both data dependent and independent acquisition (DDA/DIA) modes. DDA was used to generate spectral libraries of ARP-derivatized carbonylated peptides (ARP-peptides), which were used for their peptide-centric detection in DIA data. This allowed the identification of 86 manually confirmed ARP-peptides that were consistently detected in the serum samples, with 93.8% of all peak areas having a signal-to-background ratio greater than 3. Of the 32 carbonylation sites identified, 28 belonged to human serum albumin, including residues Cys58, Lys214, Lys219, Lys223, Lys456, Lys543, Lys549, and Lys565, which were identified as hotspots with multiple modification types. To the best of our knowledge, six previously unreported species were identified at these hotspots based on IMS, DIA, ARP-reporter ions, and de-novo sequencing. The 86 identified carbonylated peptides were consistently quantified in serum quality controls samples with an intra-batch reproducibility of at least 75%, considering an area coefficient of variation (CV) of less than 20%. All studied modifications were present at similar levels in patient and control samples, suggesting basal disease-independent modification levels for the selected modification within the accuracy of the quantitation.
Juan Camilo Rojas Echeverri 
