Involvement of the Streptococcus mutans PgfE and GalE 4-epimerases in protein glycosylation, carbon metabolism, and cell division

Abstract

Streptococcus mutans is a key pathogen associated with dental caries and is often implicated in infective endocarditis. This organism forms robust biofilms on tooth surfaces and can use collagen-binding proteins (CBPs) to efficiently colonize collagenous substrates, including dentin and heart valves. One of the best characterized CBPs of S. mutans is Cnm, which contributes to adhesion and invasion of oral epithelial and heart endothelial cells. These virulence properties were subsequently linked to post-translational modification (PTM) of the Cnm threonine-rich repeat region by the Pgf glycosylation machinery, which consists of 4 enzymes: PgfS, PgfM1, PgfE, and PgfM2. Inactivation of the S. mutans pgf genes leads to decreased collagen binding, reduced invasion of human coronary artery endothelial cells, and attenuated virulence in the Galleria mellonella invertebrate model. The present study aimed to better understand Cnm glycosylation and characterize the predicted 4-epimerase, PgfE. Using a truncated Cnm variant containing only 2 threonine-rich repeats, mass spectrometric analysis revealed extensive glycosylation with HexNAc2. Compositional analysis, complemented with lectin blotting, identified the HexNAc2 moieties as GlcNAc and GalNAc. Comparison of PgfE with the other S. mutans 4-epimerase GalE through structural modeling, nuclear magnetic resonance, and capillary electrophoresis demonstrated that GalE is a UDP-Glc-4-epimerase, while PgfE is a GlcNAc-4-epimerase. While PgfE exclusively participates in protein O-glycosylation, we found that GalE affects galactose metabolism and cell division. This study further emphasizes the importance of O-linked protein glycosylation and carbohydrate metabolism in S. mutans and identifies the PTM modifications of the key CBP, Cnm.

Keywords: Streptococcus mutans; 4-epimerases; Cnm; protein O-glycosylation.

Fig. 1

Mass spectrometric analysis of Cnm. A) Native Cnm was purified from S. mutans OMZ175 and analyzed via MS LC–MS/MS following trypsin digestion. An exemplary MS/MS spectrum (HCD fragmentation) at m/z of 618.5594 (z = 4) for the indicated peptide is shown. B) A schematic of Cnm and the full amino acid sequence of the protein are shown. Cnm contains a signal sequence (SS), collagen binding domain and LPXTG cell wall anchor. The threonine-rich region contains 21 threonine-rich repeat regions, totaling 64 threonine residues. The truncated version of Cnm (tCnm) was designed to lack all but 2 threonine-rich repeat regions (deleted region is underlined). tCnm contains only 2 out of 21 repeats and was purified directly from S. mutans wildtype. C) The tCnm protein was enzymatically digested by trypsin and analyzed via LC–MS/MS. The glycan occupancy of 12 threonine residues in the T-rich region of tCnm is shown as percent occupancy with HexNAc2, which was manually determined by extracting ion abundances of each glycopeptide. D) An exemplary MS/MS spectrum (HCD fragmentation) at m/z of 1434.4 (z = 4) for the indicated peptide of tCnm is shown. Please see Fig. S3 for more details.

Fig. 2

Detection of GalNAc and GlcNAc residues in S. mutans whole cell lysates using lectin blots. Pure Cnm, as well as whole cell lysates of S. mutans, was prepared and separated by SDS-PAGE. Proteins were transferred onto a nitrocellulose membrane and probed with HRP-conjugated soybean agglutinin (SBA) or HRP-conjugated wheat germ agglutinin (WGA). The Coomassie-stained SDS-PAGE, the SBA lectin blot, and the WGA lectin blot are shown. In the lectin blots, the BLUEstain protein ladder (goldbio) and sample proteins were visualized using different fluorescent filters, resulting in their representation in color or black and white, but the ladder and the wildtype sample were run in adjacent wells on the same gel, as depicted.

Fig. 3

Sequences and structures of the “gatekeeper” region in UDP-4-epimerases. A) Amino acid sequences corresponding to the “gatekeeper” region; the residues with side chains closest to the 2-N-acetyl group of UDP-GalNAc are highlighted in bold font and the gatekeeping residue is shown in red following the work of Beerens et al. (2015a). B, C) Crystal structure of P. aeruginosa WbpP bound to UDP-GalNAc (PDB 1SB8); the red circle highlights the binding pocket for the GalNAc moiety and the gatekeeper region. Molecular models of (D) PgfE and (E) GalE from S. mutans were constructed by AlphaFold CoLab v2.1.0 (Jumper et al. 2021). F) Crystal structure of C. jejuni Gne (PDB 7K3P). An approximate model for the location of UDP-GalNAc (magenta) was determined by superimposing various structures onto the crystal structure of P. aeruginosa WbpP bound to UDP-GalNAc to evaluate potential binding interactions and steric clashes.

Fig. 4

The substrate specificity of the epimerases was analyzed using NMR and CE. A) PgfE and GalE from S. mutans, as well as Gne from C. jejuni, were cloned into expression vectors including C-terminal hexa-His tags. The proteins were expressed in E. coli and purified using a nickel-NTA column. The eluted epimerases used in this study are shown (Coomassie stain and anti-polyhistidine western blot). Molecular weights are indicated in kDa. B) The epimerases were expressed and purified from E. coli. Then, 25 ng/μL protein, 1 mM NAD+, and 300 μM UDP-sugar in PBS were incubated for up to 1 h at 37 °C. Regions of 1D proton NMR spectra show the anomeric protons for UDP-sugars. The top trace shows starting material and lower traces the results after addition of the indicated enzymes. Spectra showing the epimerase products correspond to less than 15 min reactions, whereas spectra without products were obtained after 60 min reactions. Peaks are labeled with SNFG symbols for Glc, Gal, GlcNAc, and GalNAc (Varki et al. 2015). C) For the analysis of the substrate specificity with CE, 250 ng/μL, 0.5 mM NAD+ and 100 μM UDP-sugar in Tris buffer were incubated for 1 h at 37 °C. All spectra shown in the same lane contain the same sugar substrate as indicated in the standards (left column), the substrate peak for each sample is additionally indicated by the labeling with SNFG glycan symbols.

Fig. 5

PgfE plays a role in protein glycosylation, while GalE is important for galactose metabolism in S. mutans. A) Whole-cell lysates of S. mutans strains were separated by SDS-PAGE, transferred onto a nitrocellulose membrane, and probed with an anti-Cnm antibody. B–D) Streptococcus mutans wildtype and indicated mutant strains were grown on BHI or CDM + 1% sugar as carbon source as indicated.

Fig. 6

The loss of GalE causes an increase in bacterial chain length and changes in surface structure. A) All S. mutans strains were grown on low glucose medium supplemented with 1% glucose or galactose as the main carbon source overnight. Bacteria were applied to microscopy slides, stained with crystal violet, and imaged at 100× magnification using an Axio A1 light microscope. (B) S. mutans wildtype, ΔpgfE and ΔgalE were grown in 1% galactose and visualized with TEM. Defects in septum formation are indicated by asterisks.