Human evolution has been guided by the viruses that take refuge inside us; this ancient relationship is recorded in our genomes1. Most human viruses have been studied in the context of infection (pathogens) rather than colonization (commensals); however, we are beginning to appreciate the ways that the latter promotes fundamental aspects of our survival2,3. Regardless of its classification, the goal of each virus is universally the same: to transfer its genetic material into host cells and create more viruses. This seemingly straightforward goal is replete with challenges; namely, most viruses are not on the guestlist for admission to the exclusive party in our body. Undeterred, these con artists have evolved clever strategies to disguise themselves as friends and to forge their own invitations. A prime example — the ongoing SARS-CoV-2 pandemic — has illuminated how adept viruses can be at sweet-talking their way in.

All cells wear a carbohydrate coat. Glycans (glycoproteins, glycolipids and polysaccharides) make up this coat, radiating from the cell surface and facilitating communication and engagement with the surrounding environment. Although glycans are present across all domains of life, we are still adding to their already known functions in health and disease4. These enigmatic modifications are installed onto lipids or proteins by a suite of enzymes that act in concert to synthesize, modify and attach cognate glycans to biomolecules in and on our cells5. As cells divide and differentiate during human development, unique glycan signatures on cell surfaces help distinguish cell types from one another. Our bodies are trained to monitor these extracellular glycans and, importantly, to read glycan signatures to distinguish self from non-self. In short, if viruses want to gain entry to our bodies and stay there, they must adhere to our dress code.

Masters of disguise

Viruses can cleverly cloak their surfaces with a glycan coat that mimics the sugar chains found on host cells6. Enveloped viruses (for example, influenza A, measles, human immunodeficiency virus (HIV) and coronaviruses) are encapsulated in lipid envelopes decorated with densely glycosylated proteins. This capsule masks immunogenic viral components from the host immune system and facilitates viral attachment and entry into host cells (Fig. 1). Once inside, they hijack host protein-folding and glycosylation machinery to assemble new viruses disguised with host glycans. Non-enveloped, or naked, viruses (for example, hepatitis A, rotavirus and poliovirus) lack an attached lipid envelope, yet recent studies suggest that many package themselves into glycan-containing host membrane vesicles, or exosomes, to achieve the same effect7,8.

Fig. 1: Viruses and mammalian cells both express lectins.
figure 1

a, Viral lectins or adhesins (red) recognize mammalian glycans to facilitate viral entry. b, Lectins (blue) on host cells can recognize viral glycans, leading to clearance or viral uptake. The glycans are depicted using the standard code.

Full size image

Viruses are notorious for rapid acquisition of mutations and corresponding changes in virulence. As new strains of virus emerge, we are readily able to decode amino-acid mutations from sequencing data, yet the effects on viral glycosylation are far more challenging to predict. The glycoproteins embedded in enveloped viruses often serve as antigens to direct humoral responses. How does viral glycosylation change as viruses evolve and what are the consequences for host recognition and immunity? Changes in the location and composition of viral glycans may have dramatic effects on tissue tropism, immune activation and the efficacy of circulating antibodies and vaccination efforts. Thus, we need tools to address these questions. Methods for the analysis of viral glycosylation and its consequences are critical for understanding infection, immunity and immune evasion.

Analytical and computational approaches are essential components of glycan analysis (Fig. 2). Advances in mass spectrometry (MS) have been instrumental in characterizing the location and extent of viral glycosylation in emergent pathogens such as SARS-CoV-2, but certain caveats still remain9,10,11. To manufacture enough product for analysis while limiting exposure to infectious virions, MS analysis is routinely performed on recombinantly expressed viral proteins. As such, the cell type chosen for expression (for example, ovarian versus epithelial cells), as well as the protein expression system (bacterial, insect or mammalian), can result in dramatic differences in the observed glycosylation patterns on the viral protein under investigation. MS analyses can deconvolute glycan sequences and structures, but the connectivity of the sugars is inferred; that is, the data are based on assumptions about the type of glycan being analyzed (for example, N-glycosylation) and biosynthetic pathways present in the host12. In most cases, the method does not directly report on anomeric position and polysaccharide branching. Although these assumptions are generally warranted, they may break down in cells under stress, for example, during infection or in altered states of proteostasis13,14.

Fig. 2: Elucidating viral glycosylation.
figure 2

a, MS yields information regarding glycan structures on viral (or host) proteins. Glycosylation is determined through mass-to-charge ratio (m/z), and the connectivity is typically inferred from known host biosynthetic pathways. b, Glycan arrays help determine the carbohydrate binding specificity of rapidly evolving or strain-specific viral adhesins.

Full size image

However, MS studies do not only provide information on viral glycan identity, but also insight into how glycosylation may alter the conformational dynamics of glycoproteins. For example, MS identification of site-specific glycans on the SARS-CoV-2 spike protein helped guide molecular dynamic simulations to examine glycoprotein dynamics. These computational approaches, in combination with genetic and structural studies, afforded models that shed light on how glycan structures and dynamics contribute to host–virus interactions15,16,17. As these approaches evolve and improve, the field will be poised to predict how viruses and their variants employ glycan masks in cell attachment, infection or colonization. Looking forward, the development of complementary approaches that more rapidly fingerprint changes in viral glycosylation would be a welcomed addition to monitor emergent strains in real-time.

Getting past security: lectins as bouncers

The innate immune system is at the frontline of our response to microbial colonization and infection. A central component of innate immunity are lectins, or carbohydrate-recognition proteins, that help our bodies differentiate self from non-self (Fig. 1). Lectins involved in microbial recognition are characterized by their monosaccharide specificity and their structural architecture18. Most lectins oligomerize into higher-order structures. In this way, they achieve high avidity for the multivalent glycans found on host cells and viral envelopes.

Lectins are used by both the host and viruses. Viral lectins are used to adhere to host tissues, as in the case of influenza virus, which binds to sialic-acid residues that coat our airways19. Influenza is just one of the many viruses that encode lectins and adhesion proteins to latch onto glycans and glycoproteins on host cells and thereby force viral entry20,21,22. As another example, the proteoglycan heparan sulfate is a popular target for viral adhesion proteins, including those found on SARS-CoV-2, HIV, papillomaviruses, dengue viruses and hepatitis C, among others23,24,25. However, a comprehensive understanding of other host cell glycan targets for viruses is lacking. Clustered regularly interspaced short palindromic repeats (CRISPR)-based strategies have been used recently to reveal host factors that are involved in coronavirus infections26, but genetic modifications can have deleterious and pleiotropic phenotypes that may be challenging to parse.

Chemical biology approaches complement genetic efforts and provide powerful tools in glycomics research. A prime example is live cell proximity-based tagging, which offers an impressive route to elucidate the carbohydrate targets of glycan-binding proteins27,28,29. Applying this approach to viral adhesins could reveal a comprehensive portrait of adhesin–carbohydrate interactions. Similarly, synthesis and incorporation of non-natural sugars into the host glycome can help address this question. In this approach, synthetic monosaccharides functionalized with covalent capture and label transfer chemical handles are metabolically incorporated or appended via glycosyltransferases into the host extracellular glycocalyx30,31. Activation of these non-natural sugars in the presence of a binding partner facilitates capture of proteins that recognize these monosaccharides. One could easily envision how the development and application of chemical approaches could advance our understanding of viral pathogenesis.

Viral entry can also be mediated by viral glycans interacting with host lectins. Which lectins are targeted by viruses during infection? This question is central to viral pathogenesis. Lectins exist as secreted or cell-bound receptors, and their expression is tissue specific. Elucidating lectin–virus interactions can thus help reveal the tissue-specific tropism of human viruses. Of particular interest are viruses can that enter through one cell type and become infectious in another cell type. For example, the lectin DC-SIGN (dendritic cell-specific intercellular adhesion molecule-3-grabbing non-integrin) facilitates the uptake of HIV into dendritic cells. These antigen-presenting cells then introduce HIV to T cells, which are the major cell type for infection22. Similarly, SARS-CoV-2 begins its infection in the respiratory tract, yet pleiotropic effects have been observed in the neurological, cardiovascular and gastrointestinal systems. Elucidating the molecular underpinnings of these observations is an area of active investigation; we suspect changes in glycan recognition contribute to these phenotypes.

Customized camouflage: tailoring of viral glycans

Viruses can alter their glycan shield by perturbing host glycosylation and proteostasis pathways, and by encoding their own glycosidases and glycosyltransferases that prune and elaborate glycans. By modifying their sugary coat, viruses can tune their attachment to host lectins and, consequently, either promote infection (lectin-mediated uptake of viruses) or evade lectin detection and clearance. One way to explore viral glycosylation is lectin arrays32. Lectin arrays serve as excellent platforms to characterize altered glycosylation states during infection and reveal the human lectins that engage viruses33. In this approach, glycosylated biomolecules, cells or viruses are applied to an array of immobilized lectins with varying monosaccharide specificities32. Bound biomolecules are detected via fluorescence, thus revealing the lectins that attach to their glycans. Expansion of mammalian lectins on these arrays can help identify lectins that may target viruses for uptake or expulsion. Toward this goal, efforts to produce sufficient quantities of properly folded and functional recombinant human lectins should be prioritized.

Recently, lectin arrays revealed how perturbations to proteostasis and glycosylation are intimately linked14. Accordingly, many viruses have mechanisms to control unfolded protein responses in the cell, which gives insight into additional mechanisms by which viruses may alter their patterns of glycosylation to manipulate lectin–virus interactions and promote infection33. The relationship between proteostasis and glycosylation is supported by previous observations documenting changes in cellular glycosylation during infection13. Alterations to host glycosylation can affect cell signaling, adhesion and differentiation, among others. Aberrations in cellular glycosylation have been mapped in tumor biopsies using MS imaging — a technique that has revealed differences in the patterns of glycosylation in the context of cancer progression34. Application of this approach to viral infected tissues could therefore reveal changes in the mammalian glycome in the context of infection and colonization. With this information, we could gain insight into how viral infection leads to downstream effects in cell signaling and recognition (Fig. 2).

Moreover, recent advances in glycan visualization offer opportunities to track how cellular glycosylation patterns change over the course of infection. In one approach, non-natural sugars are metabolically incorporated into extracellular mammalian glycans35. These carbohydrates are functionalized with bio-orthogonal conjugation handles that facilitate visualization of cell glycans following a reaction with a ‘clickable’ fluorophore. More recently, biosynthetic incorporation of non-natural carbohydrates offers another glycan visualization approach that bypasses the metabolic pathways previously required for incorporation36. Although pioneered in bacterial systems, its expansion to mammalian glycomes could report on glycan dynamics during viral infection.

Many questions remain in viral pathogenesis and mutualism. For example, what are the consequences of mounting an immune response against viruses that display self-glycans? A correlation between viral infection and humoral autoimmune responses has been documented, but the contribution of glycans is unclear37,38. Glycan arrays may offer insight into the carbohydrate-binding specificity antibodies from convalescent patients. A recent study using glycan arrays revealed that patients convalescent for SARS-CoV-2 possessed antibodies capable of recognizing self-glycans, including gangliosides, N-glycans, N-acetyllactosamine (LacNAc) and sialyl LewisX (ref. 39). Autoimmune targeting of these epitopes may in part explain the symptoms observed in tissues distal to the primary infection site in SARS-CoV-2 patients. Furthermore, how do mutations in viral adhesins influence glycosylation and carbohydrate recognition? With regard to the former, changes in spike protein glycosylation influence viral neutralization40. What will be the glycan specificity for the next viral pathogen? Understanding the structure–function relationships in both host and viral glycan recognition is critical.

Conclusion

As the SARS-CoV-2 pandemic appears to relent in its severity, we reflect on the many unknowns left in its wake. Will new emerging variants, with altered glycan shields, be able to thwart global vaccination efforts? What are the short- and long-term effects of SARS-CoV-2 infection on host glycosylation, especially in tissues with low capacity for cellular renewal (for example, brain and heart tissue)? We need means to address these questions, including methods to assess alterations in the glycosylation of infected cells, to profile viral interactions with human lectins and image changes in viral glycosylation, and to evaluate the effects of viral glycosylation on antibody neutralization. New tools and research programs that address these needs will be critical to combat and manage the next sweet-talking virus that walks into the room.