Introduction
Protein glycosylation is a conserved process observed across all domains of life which allows the augmentation of protein properties including solubility, enzymatic activity and protein stability [1-4]. While once thought to be absent in bacterial systems it is increasingly recognized that numerous diverse protein glycosylation pathways exist across bacterial species ranging from dedicated systems responsible for modifying single proteins, such as flagellin [5] and autotransporters [6], to those responsible for modifying hundreds of proteins known as general glycosylation systems [1, 2, 7]. Within Gram-negative species, multiple general glycosylation systems have now been identified with the PglO (also known as PglL) family of oligosaccharyltransferases emerging as a widespread class of O- linked glycosylation systems [8]. This family of enzymes is increasingly recognized to be functionally diverse with multiple variants identified that possess unique properties including glycan specificities and substrate ranges [9-11]. While multiple PglO enzymes have been experimentally confirmed to drive O‑ glycosylation in genera including Moraxella [9]Acinetobacter [11, 12], Burkholderia[13-15], Francisella[16], Pseudomonas[17], Ralstonia[18] and Neisseria[19-21] the glycoproteomes of these bacterial species are still poorly defined with MS based approaches increasingly used to characterize bacterial glycoproteomes.
The characterization of glycopeptides using liquid chromatography–mass spectrometry (LC-MS) has become an indispensable tool for glycoproteomics [22]. While isolation of bacterial glycopeptides using offline enrichment such as hydrophilic [11, 12, 15, 18, 23, 24], charge [20] or affinity [25, 26] chromatography approaches followed by LC-MS have been ubiquitously used for bacterial glycoproteomic studies, these approaches increase the complexity of experimental designs and, as highlighted within comparisons of conventional multidimensional fractionation approaches [27], suffer from significant sample loss. An alternative to offline fractionation techniques is the use of online separation approaches with Ion mobility spectrometry (IMS) based separation/fractionation emerging as a powerful technique to streamline the analysis of biomolecules [28]. IMS enables the separation of analytes based on a combination of charge and collisional cross-section making it uniquely suited for separating mixtures containing diverse analytes [29]. Previously it has been demonstrated that IMS allows the enrichment of populations of cross-linked peptides [30, 31], the fractionation of combinations of post-translational modifications (PTMs) to dissect PTM crosstalk [32] and even the ability to distinguish glycopeptides from non-glycosylated components based on their collisional cross-section alone [33]. Leveraging IMS, multiple groups have shown different IMS platforms are advantageous for glycoproteomic analysis including trapped ion mobility separation (TIMS) [34] as well as high-field asymmetric waveform ion mobility spectrometry (FAIMS) [35-38]. Using FAIMS - based glycopeptide enrichment, we previously demonstrated that unique glycopeptides within bacterial glycoproteomes inaccessible to hydrophilic enrichment chromatography could be identified, allowing the identification of novel glycosylation events without offline glycopeptide enrichment [35]. While we have used FAIMS - based glycopeptide enrichment to characterize both bacterial N - and O -linked glycopeptides [35, 39], it is still unclear if this approach is uniformly applicable to all bacterial glycosylation systems due to the inherent chemical diversity of bacterial glycans [40].
Within members of the Neisseria genus, O- linked glycosylation was first confirmed nearly 30 years ago with protein crystallography and MS-based analysis revealing the presence of glycosylation events within the major pilin subunit (PilE) ofNeisseria gonorrhoeae [41] andNeisseria meningitidis [42]. Future work expanded understanding of the genetics [19, 20, 43-45] and the biochemical processes [45-47] associated withO -linked glycosylation in Neisseria to reveal that diverse arrays of glycans can be utilized [20, 44] to glycosylate multiple extracytoplasmic proteins [48, 49]. Within N. meningitidis andN. gonorrhoeae, O- linked glycosylation has been shown to contribute to antigenic variation allowing the evasion of the humoral immune response [43, 50, 51] yet its conservation across Neisseria species suggests that glycosylation may play roles beyond this function. Within N. gonorrhoeae, arguably the best characterized Neisseriaglycoproteome to date, glycan serotyping [48] and immunoaffinity protein enrichment [26] coupled with LC-MS analysis have identified at least 19 glycoproteins [26] yet the true extent of theN. gonorrhoeae glycoproteome remains unclear. A critical barrier to glycoproteomic studies within N. gonorrhoeae has been the dependency on bespoke affinity reagents such as monoclonal antibodies to facilitate the identification or enrichment of glycoproteins prior to LC-MS analysis. Recently, we demonstrated that N. gonorrhoeaeglycopeptides could be readily identified from whole cell lysates using LC-MS allowing the assessment of the N. gonorrhoeae glycoproteome without the need for offline glycopeptide enrichment [52]. Notably, we also observed proteome and glycoproteomic changes in N. gonorrhoeae in response to the expression of an exogenous pglOO- oligosaccharyltransferase from the distantly related speciesNeisseria elongata [52]. Thus, these findings suggest alterations in pglO alleles withinN. gonorrhoeae may drive both proteome as well as glycoproteome alterations. However, a systematic assessment of how differentpglO alleles influence the proteome/glycoproteome of N. gonorrhoeae is required to confirm if this is a general phenomenon.
Within this work, we explore the use of FAIMS fractionation to expand our understanding of the glycoproteome of N. gonorrhoeae . Leveraging FAIMS based enrichment, we explore differences in glycosylation patterns across N. gonorrhoeae strains expressing different pglO allelic chimeras with unique substrate targeting activities identifying 44 unique glycoproteins. To further understand the proteomic changes driven by alterations in PglOs, we utilized Data-Independent Acquisition (DIA) [53, 54] revealing proteomic alterations across N. gonorrhoeae strains expressing different pglOalleles. These findings support the expression of different pglOalleles drive glycoproteome and proteome alterations withinNeisseria species as well as that the expression of a glycoprotein alone, even if known to be compatible with other PglO enzymes, may be insufficient to predict glycosylation efficiency with a given PglO. Combined, this work expands our understanding of theN. gonorrhoeae glycoproteome and supports that the distinct targeting activities of different pglO alleles appears to drive proteomic changes independent of the glycoproteins targeted for glycosylation.