Influence of Protein Glycosylation on Campylobacter fetus Physiology

Campylobacter fetus is commonly associated with venereal disease and abortions in cattle and sheep, and can also cause intestinal or systemic infections in humans that are immunocompromised, elderly, or exposed to infected livestock. It is also believed that C. fetus infection can result from the consumption or handling of contaminated food products, but C. fetus is rarely detected in food since isolation methods are not suited for its detection and the physiology of the organism makes culturing difficult. In the related species, Campylobacter jejuni, the ability to colonize the host has been linked to N-linked protein glycosylation with quantitative proteomics demonstrating that glycosylation is interconnected with cell physiology. Using label-free quantitative (LFQ) proteomics, we found more than 100 proteins significantly altered in expression in two C. fetus subsp. fetus protein glycosylation (pgl) mutants (pglX and pglJ) compared to the wild-type. Significant increases in the expression of the (NiFe)-hydrogenase HynABC, catalyzing H2-oxidation for energy harvesting, correlated with significantly increased levels of cellular nickel, improved growth in H2 and increased hydrogenase activity, suggesting that N-glycosylation in C. fetus is involved in regulating the HynABC hydrogenase and nickel homeostasis. To further elucidate the function of the C. fetus pgl pathway and its enzymes, heterologous expression in Escherichia coli followed by mutational and functional analyses revealed that PglX and PglY are novel glycosyltransferases involved in extending the C. fetus hexasaccharide beyond the conserved core, while PglJ and PglA have similar activities to their homologs in C. jejuni. In addition, the pgl mutants displayed decreased motility and ethidium bromide efflux and showed an increased sensitivity to antibiotics. This work not only provides insight into the unique protein N-glycosylation pathway of C. fetus, but also expands our knowledge on the influence of protein N-glycosylation on Campylobacter cell physiology.


Analysis of N-glycans produced by Cff-PglX and Cff-PglY in the heterologous E. coli system
To determine the N-glycan structures produced upon expression of Cff-pglX and pglY in the presence of defined Cj-pgl operon mutants (lacking pglH, pglI or pglHI) in E. coli, mass spectrometric analyses of CmeA glycopeptides was carried out (Supplementary MS data 2). We found that full length Cj-N-glycan was added to CmeA (93-ATFENASKDFNR-104) when the pglH mutation on the Cj-pgl operon was complemented with Cj-pglH (pglH mut + Cj-pglH, positive control), that (as expected) diNAcBac-GalNAc2 was produced in the absence of the complementation plasmid (pglH mut ) and that no glcyopeptides were observed in the absence of the Cj-pgl operon (negative control). Upon expression of Cff-pglX in ppgl-pglH::kan cells (pglH mut + Cff-pglX) we detected the unmodified tri-saccharide substrate (diNAcBac-HexNAc2) as well as tri-saccharide substrate modified with Hex, HexNAc-Hex2 or HexNAc-Hex. Similarly upon expression of pglY alone (pglH mut + Cff-pglY) we detected the formation of incomplete substrate (diNAcBac-HexNAc), unmodified substrate (diNAcBac-HexNAc2) and substrate with one or two hexoses added to the second HexNAc of the N-glycan. Expression of pglX and pglY (pglH mut + Cff-pglXY and pglH mut + Cff-pglX+Y) resulted in the addition of Hex, HexNAc-Hex and HexNAc-Hex2 (also observed upon expression of pglX alone). In addition we observed the addition of a single HexNAc as well as the addition of HexNAc2 or HexNAc2-Hex resulting in the formation of a diNAcBac-HexNAc4 penta-and a diNAcBac-HexNAc4-Hex hexasaccharide; the latter composition would be consistent with the formation of the minor form of the native Cff-N-glycan structure.
In the ppgl-pglI::kan control strain the absence of a complementation plasmid (sample pglI mut ) or expression of Cj-pglH, (pglI mut + Cj-pglH) led to the formation of diNAcBac-HexNAc5 (as expected) and also diNAcBac-HexNAc6. For the latter case it is worth mentioning that the addition of 4 HexNAc units by Cj-PglH was also observed by Ramirez and colleagues (in vitro and in the absence of PglI) and might be an artefact of the experimental conditions (Ramirez et al., 2018). Expression of pglX (pglI mut + Cff-pglX) not only resulted in the formation of the substrate diNAcBac-HexNAc5 (and diNAcBac-HexNAc6) but also in shorter diNAcBac-HexNAc3 and diNAcBac-HexNAc3-Hex Nglycans, the latter structure indicates the addition of a Hex residue to an incomplete Cj-N-glycan pglI structure by PglX. Expression of PglY alone (pglI mut + Cff-pglY) resulted in the formation of diNAcBac-HexNAc5, diNAcBac-HexNAc6 as well as in a shorter diNAcBac-HexNAc4 N-glycan variant. Expression of PglX and PglY (pglI mut + Cff-pgXY, pglI mut + Cff-pglX+Y) resulted in the formation of diNAcBac with 3, 4, 5, or 6 HexNAc residues attached, representing the identical Nglycans formed in the absence of a complementation plasmid (pglI mut ) and shorter versions thereof. In the pglHI mutant background (pglHI mut ) we predominantly found diNAcBac with 2, 3, 4, or 5 HexNAc residues upon expression of Cj-pglH (pglHI mut + Cj-pglH), PglX, PglY and PglX/PglY (samples pglHI mut + Cff-pglX, pglHI mut + Cff-pglY, pglHI mut + Cff-pglXY and pglHI mut + Cff-pglX+Y), however, in the absence of the complementation plasmid a diNAcBac-HexNAc2-Hex variant could be seen. This was unexpected since no additional pgl gene was present that could explain the addition of the Hex residue that might therefore be added by an E. coli GTase; however, upon expression of Cff-pglX (pglHI mut + Cff-pglX) we also detected a diNAcBac-HexNAc3-Hex N-glycan variant indicating that this hexose residue could also be a product of this Cff-GTase.

Digestion of CmeA
Isolated CmeA bands were processed as previously described (Shevchenko et al., 2006) with minor modifications. Briefly, CmeA-His6 was enriched by Ni-NTA gravity flow as described (Feldman et al., 2005) and subsequently separated by 12% PAGE followed by Coomassie straining. CmeA-His6 bands were excised and destained in a 50:50 solution of 50 mM NH4HCO3: 100% ethanol for 20 min at room temperature with shaking at 750 rpm. Destained bands were washed with 100% ethanol for 10 min at 750 rpm for dehydration and then rehydrated in 10 mM DTT in 50 mM NH4HCO3. Reduction was carried out for 60 min at 56°C with 750 rpm shaking. The reduction buffer was then removed, and the gel bands washed twice in 100% ethanol for 10 min to ensure the removal of DTT. Reduced ethanol washed samples were sequentially alkylated with 55 mM iodoacetamide in 50 mM NH4HCO3 in the dark for 45 min at RT. Alkylated samples were then washed with 2 rounds of 100% ethanol and then vacuum-dried. Alkylated samples were then rehydrated with 12 ng µl -1 trypsin (Promega, Madison WI) in 40 mM NH4HCO3 at 4°C for 1 hr. Excess trypsin was removed, gel pieces were covered in 40 mM NH4HCO3 and incubated overnight at 37°C. Peptides were concentrated and desalted using C18 stage tips (Ishihama et al., 2006;Rappsilber et al., 2007) and stored on tips at 4°C. Peptides were eluted in buffer B (0.1% formic acid, 80% MeCN) and dried before analysis by LC-MS.

Identification of glycosylated peptides using reversed phase LC-MS
Purified CmeA peptides prepared were resuspended in Buffer A* (2% acetonitrile, 0.1% TFA) and separated using a two-column chromatography set up composed of a PepMap100 C18 20 mm x 75 μm trap and a PepMap C18 500 mm x 75 μm analytical column (Thermo Fisher Scientific). Samples were concentrated onto the trap column at 5 μL/min for 5 minutes and infused into an Orbitrap Fusion™ Lumos™ Tribrid™ Mass Spectrometer (Thermo Fisher Scientific) at 300 nl/min via the analytical column using a Dionex Ultimate 3000 UPLC (Thermo Fisher Scientific). Then, 60 min gradients were run altering the buffer composition from 1% buffer B (0.1% formic acid, 80% MeCN) to 28% B over 35 min, then from 28% B to 40% B over 10 min, then from 40% B to 100% B over 2 min, the composition was held at 100% B for 3 min, and then dropped to 3% B over 5 min and held at 3% B for another 10 min. The Lumos™ Mass Spectrometer was operated in a data-dependent mode automatically switching between the acquisition of a single Orbitrap MS scan (120,000 resolution) every 3 seconds and Orbitrap HCD scans for each selected precursor (NCE 28, maximum fill time 100 ms, AGC 4*10 4 with a resolution of 15000). For MS/MS events observed to contain the HexNAc oxonium ion 204.0867 three additional scans were triggered; One Orbitrap EThcD MS-MS scan (NCE 25; maximum fill time 250 ms, AGC 2*10 5 with a resolution of 30000); One Orbitrap HCD MS-MS scan (stepped NCE of 25, 40 and 48, maximum fill time 250 ms, AGC 2*10 5 with a resolution of 30000) and a ion-trap CID scan (NCE 35, maximum fill time 50ms).

Mass spectrometry data analysis of CmeA glycosylation
Identification of CmeA glycoprotein samples was accomplished using MaxQuant (v1.6.3.4) (Cox and Mann, 2008). Searches were performed against the CmeA sequence (Uniprot sequence: Q0PBE3) with carbamidomethylation of cysteine set as a fixed modification. Searches were performed with trypsin cleavage specificity allowing 2 mis-cleavage events and the variable modifications of oxidation of methionine, the Campylobacter glycan mass (elemental composition C56H91N7O34; Asn) and acetylation of protein N-termini. The precursor mass tolerance was set to 20 parts-permillion (ppm) for the first search and 10 ppm for the main search, with a maximum false discovery rate (FDR) of 1.0% set for protein and peptide identifications. Glycopeptides of CmeA were identified by manually interrogating possible glycopeptide scans based on the presence of the diagnostic oxonium ion (204.09 m/z) of HexNAc. For the CmeA glycopeptide ATFENASKDFNR glycopeptide ion intensities for the + 2 and +3 charge states were manually extracted and used to assess relative glycoform abundance. Representative annotated spectra for each glycoform and charge stated identified by MS/MS is provided (Supplementary MS data 2). The pgl genes were inactivated by insertion of the kanamycin resistance gene (kan) into the chromosome of C. fetus subsp. fetus ATCC 27374. PCR products were introduced by electroporation and double-crossover events resulted in the formation of pglX-and pglJ-. Primer sets CS392/CS393 and CS394/395, were used to confirm the correct insertion of the antibiotic cassette. (B) Agarose gel with PCR products amplified to verify the insertion of the cassette. Genomic DNA from WT or the respective pgl mutant strain that was used as template is indicated above each lane. The primer set used for each PCR reaction is indicated below the gel.

-
Cj-CmeA-His 6 in E. coli ppgl-pglHI::kan(cat-) ppgl-pglH::kan(cat-) Sensititre™ was used to assess the minimum inhibitory concentration (MIC) of the indicated Cff strains. The data (in μg/ml -1 ) represent one assay done at the Athens Veterinary Diagnostic Facility. Restriction sites are underlined, ribosomal binding sites introduced by PCR are in italics, start codons are in bold, 5-prime phosphorylation is indicated by /5Phos/.