A Sugar Cloak of Invisibility

Harmful bacteria are constantly evolving new strategies to increase their pathogenicity, and now, with help from a bacteriophage, they have a new trick. Methicillin-resistant Staphylococcus aureus (MRSA) is a serious public health threat causing difficult-to-treat infections that are often fatal. A majority of humans possess antibodies against S. aureus as a result of its membership in our normal microbiota, with a large percentage targeting wall techoic acids (WTAs). In S. aureus, WTAs are anionic glycopolymers composed of repeating ribitol phosphate (RboP) units that are anchored to peptidoglycan and decorate the cell surface (Figure 1).¹ Specific hydroxyl groups of these repeating ribitol units are modified with D-alanine and N-acetylglucosamine (GlcNAc), which influence a range of important physiological phenomena, including virulence and antibiotic resistance (Figure 1). The increased pathogenicity of recently emerging healthcare-associated (HA)-MRSA and livestock-associated (LA)-MRSA has been proposed to arise from effective strategies for evading the host immune system; however, the mechanisms of evasion remain ill-defined. Whether these immune evasion strategies involve structural changes to dominant surface antigens like WTAs has been a tantalizing possibility. In this work, Gerlach et al. demonstrate that certain MRSA strains use a novel, prophage-encoded O-GlcNAc transferase to glycosylate WTAs at a different position on RboP, producing an isomer of extremely low immunogenicity and avoiding detection by anti-WTA antibodies.² In effect, MRSA has swapped its normal sugar coating for a new sugar cloak of invisibility.

Figure 1.

Structure of wall techoic acids (WTAs) from S. aureus highlighting the different attachment positions of GlcNAc by TarS and TarP glycosyltransferases.

Noting the importance of WTA modifications in modulating critical aspects of bacterial physiology, Gerlach et al. decided to search for homologues of WTA biosynthetic genes in S. aureus genomes that may afford altered WTA antigens with the capacity for immunoevasion. Surprisingly, this search revealed three S. aureus prophages that encode a putative WTA β-GlcNAc transferase (TarP) that is 27% identical to the β-GlcNAc transferase TarS, which was also present in the genomes. With a putative β-GlcNAc glycosyltransferase identified, Gerlach et al. moved to validate its function in several important interactions. When tarP was expressed in different WTA glycosylation-deficient S. aureus mutants, it was able to restore susceptibility to siphophage infection and full oxacillin resistance, both of which require WTA GlcNAcylation. Notably, an extensive analysis of 90 clinical isolates of MRSA CC5 and CC398 revealed a dramatic difference in podophage susceptibility between tarP-containing and tarP-lacking strains. Only tarP-containing strains were resistant to podophage, suggesting that TarP produces WTAs with an altered glycosylation pattern compared to that of TarS-modified WTAs. Structural analysis of purified WTAs by nuclear magnetic resonance (NMR) spectroscopy confirmed that TarP attaches GlcNAc to the RboP backbone via a β-linkage similar to TarS (Figure 1). However, TarP was shown to glycosylate RboP at the C3 hydroxyl compared to the C4 hydroxyl as mediated by TarS. Importantly, TarP appears to outperform TarS as a glycosyltransferase because WTAs from strain N315, which contains both TarP and TarS, were almost exclusively decorated with GlcNAc at the C3 position. Determination of the structure of TarP in complex with co-substrate UDP-GlcNAc and a synthetic WTA substrate mimic provided a structural basis for explaining the altered regioselectivity of TarP compared with that of TarS.

Given that the site of attachment of GlcNAc to WTAs was shown to have a profound effect on podophage susceptibility, Gerlach et al. hypothesized that human antibodies may also discriminate between these two isoforms and that this structural difference could lead to immune evasion in TarP-expressing MRSA strains. To investigate this hypothesis, human antibody preparations were tested for their ability to bind a panel of MRSA N315 strains. Compared to the glycosylation-deficient mutant N315ΔtarSΔtarP, strains expressing only TarS bound significantly higher levels of human IgG, indicating that C4 GlcNAcylated WTA is a prominent antigen for eliciting antibodies against S. aureus. On the other hand, strains expressing either TarP alone or TarP and TarS bound only slightly higher levels IgG compared to that of the glycosylation-deficient mutant. These results demonstrate that humans naturally produce only a small fraction of anti-S. aureus antibodies that are specific for TarP-modified WTA with GlcNAc at the C3 position, suggesting a mechanism for evading immune surveillance. In support of this hypothesis, deletion of tarP from N315 strains resulted in a significantly increased capacity of human neutrophils to phagocytose cells bound with anti-WTA antibodies. To further investigate this intriguing result, the authors directly tested the immunogenicity of purified C4 GlcNAcylated WTA (TarS-modified) and C3 GlcNAcylated WTA (TarP-modified) in mice. While administration of C4 GlcNAcylated WTA to mice produced a robust IgG antibody response as expected, negligible IgG levels were produced when C3 GlcNAcylated WTA was administered, further demonstrating that C3 GlcNAcylated WTA is much less immunogenic than the normal C4 isomer. Additionally, vaccination of mice with either C4 GlcNAcylated WTA or C3 GlcNAcylated WTA did not protect these mice from subsequent infection with a tarP-expressing N315 strain.

The striking investigations by Gerlach et al. provide strong evidence that some emerging MRSA strains use an alternative WTA glycosyltransferase to modify cell surface antigens, thus increasing their pathogenicity by subverting the adaptive host immune response. These tarP-expressing strains achieve immune evasion through the extremely low immunogenicity of their dominant C3 GlcNAcylated WTA antigen and by avoiding detection from antibodies generated against C4 GlcNAcylated WTA from previous S. aureus infections. It is currently unclear why C3 GlcNAcylated WTA is poorly immunogenic, and future work to elucidate this interesting conundrum may reveal new biology. It is possible that C3 GlcNAcylated WTA resembles a currently unknown self-antigen. Alternatively, very little is known about antigen presentation of glycopolymers in adaptive immunity, and the structural change introduced by TarP may prevent C3 GlcNAcylated WTA from engaging this pathway.³

The high percentage of tarP-encoding prophages in prevalent HA- and LA-MRSA CC5 and CC398 strains suggests an important fitness advantage and poses the risk of more widespread dissemination of this novel virulence factor that could complicate treatment and vaccine development. Unfortunately, all attempts to develop a vaccine against S. aureus have failed, using both protein and glycopolymer antigens.⁴ The low immunogenicity of C3 GlcNAcylated WTA has significant implications for any vaccine program using WTA antigen and may guide the development of more effective vaccine strategies. Importantly, the discovery and structural elucidation of TarP enable the development of selective TarP inhibitors to prevent S. aureus from altering the immunogenicity of dominant cell surface antigens.

Finally, the surprising finding that the alternative WTA glycosyltransferase TarP is prophage-encoded highlights a growing appreciation for phage in regulating the dynamics and function of microbe—microbe and host—microbe interactions within microbiotas.⁵ It is tempting to speculate the alternative glycosylation selectivity of TarP was selected to enable better S. aureus survival and in turn phage propagation. The functions of phage-encoded proteins are often extremely poorly annotated and understood; however, characterizing their functions will likely be critical to fully understanding the role of the human microbiota and host interactions with bacterial pathogens in health and disease. Who knows how many bespoke cloaks of invisibility await discovery.