Salmonella enterica Serovar Typhi Lipopolysaccharide O-Antigen Modification Impact on Serum Resistance and Antibody Recognition

ABSTRACT Salmonella enterica serovar Typhi is a human-restricted Gram-negative bacterial pathogen responsible for causing an estimated 27 million cases of typhoid fever annually, leading to 217,000 deaths, and current vaccines do not offer full protection. The O-antigen side chain of the lipopolysaccharide is an immunodominant antigen, can define host-pathogen interactions, and is under consideration as a vaccine target for some Gram-negative species. The composition of the O-antigen can be modified by the activity of glycosyltransferase (gtr) operons acquired by horizontal gene transfer. Here we investigate the role of two gtr operons that we identified in the S. Typhi genome. Strains were engineered to express specific gtr operons. Full chemical analysis of the O-antigens of these strains identified gtr-dependent glucosylation and acetylation. The glucosylated form of the O-antigen mediated enhanced survival in human serum and decreased complement binding. A single nucleotide deviation from an epigenetic phase variation signature sequence rendered the expression of this glucosylating gtr operon uniform in the population. In contrast, the expression of the acetylating gtrC gene is controlled by epigenetic phase variation. Acetylation did not affect serum survival, but phase variation can be an immune evasion mechanism, and thus, this modification may contribute to persistence in a host. In murine immunization studies, both O-antigen modifications were generally immunodominant. Our results emphasize that natural O-antigen modifications should be taken into consideration when assessing responses to vaccines, especially O-antigen-based vaccines, and that the Salmonella gtr repertoire may confound the protective efficacy of broad-ranging Salmonella lipopolysaccharide conjugate vaccines.

The sample was permethylated, depolymerized, reduced, and acetylated, and the resultant partially methylated alditol acetates (PMAAs) were analyzed by gas chromatography-mass spectrometry (GC-MS) using a modification of the method described by York et al. (2) About 1 mg of each sample was suspended in 200 μl DMSO. The samples were permethylated using 400 μL a suspension of NaOH in DMSO (3). After stirring for 15 min, 50 μL CH3I was added, and the mixture was left for 45 min. The addition of NaOH and CH3I was repeated to ensure complete methylation of the polymer. After an additional 45 min, 2 mL water was added, and excess CH3I was removed by sparging with nitrogen gas. The mixture was extracted with CH2Cl2, and after washing the organic phase 3 times with water and drying it down with nitrogen gas, the permethylated material was hydrolyzed using 2 M trifluoroacetic acid (2 h in sealed tube at 121 °C), reduced with NaBD4, and acetylated using acetic anhydride/trifluoroacetic acid. The resulting PMAAs were analyzed on a Agilent 7890NA GC interfaced to a 5975BC MSD (mass selective detector, electron impact ionization mode). Separation was achieved on a 30 m Supelco 2380 bonded phase fused silica capillary column.
For mild acid hydrolysis, the samples were heated in 1% acetic acid at 100 °C for 1 h. The lipid fraction was removed by centrifugation at 10,000 g, and the supernatant was freeze-dried.
For de-O-acetylation, the samples were dissolved in water (10 mg/mL), and the solution was brought to pH 11 by addition of concentrated ammonia water. After allowing the reaction to proceed at room temperature for 18 h, the solution was dialyzed (2 kDa) against water and freeze-dried.

NMR Spectroscopy
For de-O-acetylated polysaccharide analysis, the samples were deuterium exchanged by lyophilization from D2O (99.9% D, Aldrich) and dissolved in 300 μL D2O (99.96% D, Cambridge Isotopes). One-dimensional proton and 2-D gCOSY, TOCSY, NOESY, gHSQC, and gHMBC spectra were obtained on a Varian Inova-600 MHz spectrometer at 70 °C using standard Varian pulse sequences. TOCSY and NOESY mixing times were 80 and 300 ms, respectively. Chemical shifts were measured relative to internal acetone (H=2.218 ppm, C=33.0 ppm) (4). To analyze native LPS, the samples were deuterium exchanged by dissolving in D2O and lyophilization and dissolved in 700 μL D2O (some material did not dissolve). 1-D proton and 2-D gCOSY, TOCSY, NOESY, HMQC (or gHSQC), and gHMBC (STy-Gluc and STy-FM only) spectra were obtained on a Varian Inova-600 MHz spectrometer at 70 °C using standard Varian pulse sequences. TOCSY and NOESY mixing times were 80 and 300 ms, respectively. Chemical shifts were measured relative to the residual HDO signal (=4.31 ppm at 70 °C), using a  value of 0.25144953 for 13 C.

Differences in glucosylation
We obtained the de-O-acetylated O-chains of the three LPS samples by mild acid hydrolysis to cleave off Lipid A and ammonium hydroxide saponification to remove O-acetyl groups.
We performed NMR in order to test whether the mutations entailed any differences in the monosaccharide sequence of the O-chain repeating unit. We were particularly interested to know if there was a difference in the DG (degree of glucosylation) on O-4 of the galactose residue. The 1-D proton spectra of the de-O-acetylated polysaccharides were almost identical in all three samples, but we did notice some differences in the intensity of some of the minor anomeric signals ( Figure S2, Peaks G, K, I, and M). In order to identify these peaks, we obtained 2-D COSY, TOCSY, NOESY, HSQC, and HMBC NMR spectra of the de-Oacetylated samples in order to assign most of the signals, hoping that this would allow us to measure the DG. We were able to identify the spin systems related to the presence or absence of the -Glc residue, most notably 3-linked and 3,4-linked -Gal, but also two versions each of 4--Rha and 2,3--Man (see Table S2). Two additional Tyv anomeric signals were observed as well. Unexpectedly, we also detected a -Gal residue that was 3-linked to another 4-Rha and may represent a defect in the polysaccharide chain. Using the connectivities obtained from the NOESY and HMBC spectra, we were able to group these Subunits I and II have been reported previously (5, 6), but Subunit III has not. With the completed assignments in hand, it seemed possible to quantify the 3-Gal and 3,4-Gal residues in order to obtain the DG in each sample, but the Gal residues were not resolved in either the 1-D proton or the 2-D HSQC spectra. Nevertheless, the Rha residue of each subunit was resolved from the others, and they could be used to measure the overall DG.
The intensities of the anomeric signals and the calculated DG are listed in Table S3. This analysis showed that BRD948 had a significantly lower DG (55.9%) than both STy-Gluc (72.8%) and STy-FM (71.2%).

Differences in acetylation
To