Improved N- and O-Glycopeptide Identification using High-Field Asymmetric Waveform Ion Mobility Spectrometry (FAIMS)

Abstract

Mass spectrometry is the premier tool for identifying and quantifying site-specific protein glycosylation globally. Analysis of intact glycopeptides often requires an enrichment step, after which the samples remain highly complex and exhibit a broad dynamic range of abundance.Here, we evaluated the analytical benefits of high-field asymmetric waveform ion mobility spectrometry (FAIMS) coupled to nano-liquid chromatography mass spectrometry (nLC-MS) for analyses of intact glycopeptide devoid of any enrichment step. We compared the effects of compensation voltage on the transmission of N- and O-glycopeptides derived from heterogeneous protein mixtures using two FAIMS devices. We comprehensively demonstrate the performance characteristics of the FAIMS device for glycopeptide analysis and recommend optimal electrode temperature and compensation voltage (CV) settings for N- and O-glycopeptide analysis.Under optimal CV settings, FAIMS-assisted gas-phase fractionation in conjunction with chromatographic reverse phase separation resulted in a 31% increase in the detection of both N- and O-glycopeptide compared to control experiments without FAIMS. Overall, our results demonstrate that FAIMS provides an alternative means to access glycopeptides without any enrichment providing an unbiased global glycoproteome landscape. In addition, our work provides the framework to verify ‘difficult-to-identify’ glycopeptide features.

False-Positive Glycopeptide Identification via In-FAIMS Fragmentation

“…As mentioned in the introduction, these studies investigated the glycoproteome of three Burkholderia species, 23 TMT-labeled N-glycopeptides, 24 tryptic N-glycopeptides from α-1-acid glycoprotein, 25 and both N-and Oglycopeptides from depleted human serum. 26 In agreement with Alagesan et al, 26 we found that as sample complexity increased, the utility of FAIMS was more apparent. Further, we showed that when using CVs lower than −40 V, we identified shorter glycopeptides with a higher aliphatic index, similar to Izaham et al 23 Though, contrary to this study, we found that by using CVs higher than −40 V, the observed glycopeptides were significantly longer and with a smaller aliphatic index.…”

“…This sample had lower complexity when compared to the entire human proteome, which should allow more reliable search results while still providing a wider variety of peptide sequences. 30 In accordance with previous CV optimization in other studies, 20,21,23,25,26 we performed single CV experiments from −25 to −80 V in −5 V intervals, as well as a control experiment without FAIMS. To select settings for double and triple CV experiments, we compared glycopeptides identified from single CV experiments, where we found that 87% of identifications were found using CVs −40, −45, or −50 V. As such, a combination of these CVs were used for double and triple CV experiments (Figure S6).…”

“…To investigate whether these observations were unique to our instrumentation setup, we repeated the above analyses with data from depleted human serum used in a recent preprint by Alagesan et al 26 MS files were downloaded from PRIDE (identifier PXD038673), searched with Byonic, and analyzed in the manner described above. Our results demonstrated that on average, 39.7% of GSMs originated from IFF in −40 V, 44.6% in −45 V, and 47.3% in −50 V (Table S9, Figure S13).…”

“…Notably, the observed increase in IFF relative to our data could have been caused by the electrode temperature. While our experiments were performed with an inner electrode temperature of 70 °C, data collected by Alagesan et al 26 was acquired at 100 °C. The higher temperature could increase ion heating of glycopeptides as they are separated in the FAIMS source, thus leading to more instances of IFF.…”

“…23 Other studies showed FAIMS enhanced detection and confidence of glycopeptide assignments using tandem mass tag (TMT)-labeled Nglycopeptides, 24 N-glycopeptides from α-1-acid glycoprotein, 25 and both N-and O-glycopeptides from depleted human serum. 26 As previous studies primarily concentrated on N-glycoproteomics, we targeted our analysis to O-glycoproteins and mucin-domain glycoproteins to determine if FAIMS could improve the detection and separation of these often-overlooked species. In order to achieve this, we analyzed three samples consisting of a mucin-domain glycoprotein, a mixture of recombinant glycoproteins, and a platelet glyco-sheddome.…”