Muc5ac gastric mucin glycosylation is shaped by FUT2 activity and functionally impacts Helicobacter pylori binding

Abstract

The gastrointestinal tract is lined by a thick and complex layer of mucus that protects the mucosal epithelium from biochemical and mechanical aggressions. This mucus barrier confers protection against pathogens but also serves as a binding site that supports a sheltered niche of microbial adherence. The carcinogenic bacteria Helicobacter pylori colonize the stomach through binding to host glycans present in the glycocalyx of epithelial cells and extracellular mucus. The secreted MUC5AC mucin is the main component of the gastric mucus layer, and BabA-mediated binding of H. pylori to MUC5AC confers increased risk for overt disease. In this study we unraveled the O-glycosylation profile of Muc5ac from glycoengineered mice models lacking the FUT2 enzyme and therefore mimicking a non-secretor human phenotype. Our results demonstrated that the FUT2 determines the O-glycosylation pattern of Muc5ac, with Fut2 knock-out leading to a marked decrease in α1,2-fucosylated structures and increased expression of the terminal type 1 glycan structure Lewis-a. Importantly, for the first time, we structurally validated the expression of Lewis-a in murine gastric mucosa. Finally, we demonstrated that loss of mucin FUT2-mediated fucosylation impairs gastric mucosal binding of H. pylori BabA adhesin, which is a recognized feature of pathogenicity.

Figure 1. Muc5ac and Lewis b-terminated glycan structures are expressed in close molecular proximity in “secretor” wild-type mice.

(a) Schematic representation of the biosynthetic pathway leading to type 1 Le^a and Le^b antigens expression. Paraffin-embedded sections of wild-type and Fut2-null mice gastric mucosa, fixed with Carnoy’s solution, were used for: (b) immunofluorescence labeling with Le^a (7LE), Le^b (BG6) and Muc5ac (45M1) recognizing antibodies, images on the right present higher magnifications of the areas within the white rectangles; (c) double-immunofluorescence labeling with Muc5ac and Le^b recognizing antibodies, using fluorescent-labeled immunoglobulin subtype specific secondary antibodies; and (d) in situ Proximity Ligation Assay (PLA) for Muc5ac and Le^b. Each fluorescent red spot corresponds to a PLA signal and is indicative that Muc5ac and Le^b are in close proximity. FVBN wild-type and FVBN-LeB transgenic mice gastric mucosa tissue sections were also included in PLA analysis as experiment biological positive controls. All tissue sections were stained with DAPI for nucleus visualization.

Figure 2. Muc5ac from “non-secretor” Fut2-null mice displays enrichment in Lewis a glycan structures.

Wild-type and Fut2-null mice total gastric mucosa protein lysates were used for evaluation of (a) Muc5ac (45M1) and (b) Le^a (SPM279) expression by Western blotting. Labels (1) and (2) represent protein lysates from two independent mice samples. (c) G-sepharose beads coupled with Muc5ac recognizing antibody 45M1 were incubated with wild-type and Fut2-null mice gastric mucosa protein extracts. Proteins that bound to the beads were analyzed by Western blotting with an anti -Le^a antibody.

Figure 3. Lack of FUT2 results in altered Muc5ac glycosylation patterns.

(a) Slot-blot analysis of collected supernatant fractions (C1, C2) and pellet fraction, upon solubilization by DTT/alkylation (DTT/A) treatment, using the LUM5-1 and LUM6-3 antibodies for Muc5ac and Muc6, respectively. (b) Serial dilutions (800; 80; 40; 20; 10 and 5 μg/μl) of purified Muc5ac from wild-type and Fut2-null mice gastric samples were probed with Le^a (SPM279) and Le^b (BG6) recognizing antibodies. Muc5ac loading amounts were verified using the PM7 mixture of antibodies (Table 1). A negative control, without primary antibody incubation, was included for detection of background staining of non-specific secondary antibody binding. (c) Quantification of dot blot relative intensity. The optical density of each signal was measured in three independent dot-blot analyses. Densitometry average values with standard deviation are shown.

Figure 4. Structural analysis of the glycan structures carried by Muc5ac in wild-type and Fut2-null mice gastric mucosa.

MS/MS spectra of the pseudomolecular ion at m/z 919, isolated from gastric MUC5AC mucins from (a) wild-type and (b) Fut2-null mice, acquired in the positive ion mode. Nuclear magnetic resonance analysis of (c,e) wild-type and (d,f) Fut2-null mice for determination of the glycan structures present on the gastric Muc5ac mucin. Part of ¹H-¹H 2D-NMR COSY spectrum acquired on a 14.3Teslas spectrometer of total O-glycans isolated from gastric mucin of (c) wild-type and (d) Fut2-null mice Muc5ac mucin. (e) Extended part of the wild-type even COSY spectrum showing the chemical shifts zone of H6/H5 correlation from fucose residues. (f) Extended part of the Fut2-null mice NMR COSY spectrum shown in panel (d).

Figure 5. Functional impact of altered Muc5ac glycosylation on Helicobacter pylori binding.

(a) Binding of H. pylori reference strain 17875/Leb (BabA+/SabA−) to purified Muc5ac from wild-type and Fut2-null mice labeled with ¹²⁵I was tested by radio immuno assay (RIA). The y axis gives the percentage of bacteria that bound to radio labeled Muc5ac. E. coli was included in the assay to test for non-specific bacterial binding to Muc5ac. (b) H. pylori blood group antigen binding adhesin (BabA) histobinding to wild-type and Fut2-null mice gastric mucosa. Paraffin-embedded sections of wild-type and Fut2-null mice gastric mucosa, fixed with Carnoy’s solution, were incubated with native BabA adhesin isolated from 17875/Leb strain. BabA binding to the mice tissue sections was detected by immunofluorescent labeling with an anti-BabA specific secondary antibody. Sequential mice tissue sections were probed with BabA that was pre-incubated with its cognate ligand Le^b, to assess the specificity of BabA mucosal binding. Images on the right represent higher magnifications of the regions within the white rectangles.