Evolution of Glycan Diversity

Ajit Varki; Hudson H Freeze; Pascal Gagneux

NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Varki A, Cummings RD, Esko JD, et al., editors. Essentials of Glycobiology. 2nd edition. Cold Spring Harbor (NY): Cold Spring Harbor Laboratory Press; 2009.

Essentials of Glycobiology. 2nd edition.

Show details

< Prev Next >

Chapter 19Evolution of Glycan Diversity

Ajit Varki, Hudson H Freeze, and Pascal Gagneux.

This chapter provides a brief comparative overview of the patterns of glycosylation in various taxa of living organisms and discusses the complexity and diversity of these glycans from an evolutionary perspective. Because much of the currently available information concerns vertebrates, this chapter focuses on comparisons between the glycans of vertebrates and those of other taxa. The evolutionary processes that likely determine the generation of glycan diversity are briefly considered, including intrinsic host glycan-binding protein functions and interactions of hosts with extrinsic pathogens or symbionts.

RELATIVELY LITTLE IS KNOWN ABOUT GLYCAN DIVERSITY IN NATURE

The genetic code is essentially the same in all known living organisms, and several core functions such as gene transcription and energy generation tend to be conserved across various taxa. Although glycans are also found in all organisms, considerable diversity of their structure and expression exists in nature, both within and between evolutionary lineages. Partly because of the inherent difficulties in studying their structures, relatively little is known about the details of this glycan diversity, and there are few comprehensive data sets on this subject. For many taxa, essentially no information is available on their glycan profiles. Sufficient data are available, however, to indicate that there is no universal “glycan structure code” akin to the genetic code. Indeed, the glycans expressed by most free-living Eubacteria and Archaea (formerly grouped as prokaryotes) (see Chapter 20) have relatively little in common structurally with those of eukaryotes (an exception occurs when bacterial pathogens mimic eukaryotic host structures). On the other hand, most major glycan classes identified in animal cells seem to be represented in some related form among other eukaryotes, and sometimes in Archaea. Figure 19.1 outlines the major branches of the eukaryotic tree of life, with an emphasis on the protostome-deuterostome split and the phylogeny of “model” organisms, for which whole-genome sequence data have been generated. In contrast, far fewer organisms have been the subject of in-depth glycan structural analyses. The high levels of diversity encountered in the best-studied vertebrate species are a predictor of similar diversity in other groups of organisms. The existing information on the distribution of glycan types points to complicated patterns. On the one hand, glycan patterns can form “trends” and characterize entire phylogenetic lineages where one encounters further biochemical variation with subsets unique to certain sublineages. On the other hand, many glycans show rather discontinuous distribution across the tree of life and distantly related organisms can express surprisingly similar glycans.

FIGURE 19.1

Phylogenetic trees of the three domains of cellular life (upper panel) and of the multicellular Eukarya (lower panel). The universal tree of life (upper panel) is inferred from maximum likelihood analysis of 1620 homologous nucleotide positions of small-subunit (more...)

EVOLUTIONARY VARIATIONS IN GLYCANS

O-Glycans

Homologs of the UDP-GalNAc:polypeptide N-acetylgalactosaminyltransferases (ppGalNAcTs) that initiate synthesis of the most common O-glycan class in vertebrates have been found throughout the animal kingdom (see Chapter 9). Multiple isoforms of ppGalNAcTs exist in most species. The common core-1 Galβ1-3GalNAcα1-O-Ser/Thr structure of vertebrates is present in insects, where it also forms part of a mucin-like protective layer in the gut. In contrast, plants do not appear to have O-linked GalNAc. Instead, they express arabinose O-linked to hydroxyproline and galactose O-linked to serine and threonine (see Chapter 22). Far less is known about bacterial O-glycosylation, although it is clear that novel O-glycans can be found within bacterial “S” layers, for example, a Galβ1-O-Tyr core (see Chapter 20).

Glycosphingolipids

Glucosylceramide is found in both plants and animals (see Chapter 10). However, the commonest core structure of vertebrate glycosphingolipids (Galβ1-4Glc-Cer) is varied in other organisms, for example, Manβ1-4Glc-Cer and GlcNAcβ1-4Glc-Cer in certain invertebrates. Other variations are inositol-1-O-phosphorylceramide, for example, mannosyldiinositolphosphorylceramide, which is the most abundant sphingolipid of yeast, and GlcNAcα1-4GlcAα1-2-myo-inositol-1-O-phosphorylceramide, which is found in tobacco leaves. Galactosylceramide and its derivatives seem to be limited to the nervous system of the deuterostome lineage of “higher” animals (see Figure 19.1). In contrast, all protostome nerves contain mainly glucocerebrosides. An evolutionary trend is suggested: A transition from gluco- to galactocerebrosides corresponds with changes in the nervous system from loosely structured to highly structured myelin. With regard to the complex gangliosides of the deuterostome nervous system, some general trends are seen in comparing reptiles to fish to mammals: an increase in sialic acid content, a decrease in the complexity of ganglioside composition, and a decrease in “alkali-labile” molecules (bearing O-acetylated sialic acids). A general rule has also been suggested: the lower the environmental temperature, the more polar the composition of brain gangliosides. Thus, poikilothermic (cold-blooded) animals tend to express many polysialylated gangliosides in the brain.

N-Glycans

Perhaps the broadest base of evolutionary information concerns asparagine–N-linked glycans (see Chapter 8). All plants and animals studied to date seem to share the same early stages of the classic N-glycan processing pathway (see Chapter 8), including the generation and transfer of Glc₃Man₉GlcNAc₂ from a dolichol-linked precursor to asparagine residues on newly synthesized proteins. Such an extreme degree of conservation is understandable, given the critical role of this glycan structure in modulating the folding and maturation of newly synthesized glycoproteins in the endoplasmic reticulum (ER) (see Chapter 36). However, some parasitic protists can transfer truncated forms of otherwise similar lipid-linked oligosaccharides, sometimes even just the core GlcNAc₂ sequence. The intracellular localization of these structures also makes them less likely to be involved in rapid evolutionary arms races due to exploitation by parasites and pathogens (see below). The trimming and extension steps that occur thereafter along the vertebrate N-glycan processing pathway are recapitulated to varying extents in other eukaryotic taxa (Figure 19.2). Yeasts and vegetative slime molds do not appear to complete the trimming of mannose residues, and are thus unable to generate typical “complex-type” N-glycans. Yeast often further specialize their high-mannose glycans by extending them into large mannans. In contrast, developing slime molds trim down the high-mannose forms to some extent but then do not extend them. In insects, mannose trimming appears to be generally completed, as in mammals, down to a Man₃GlcNAc₂ structure. The subsequent addition of GlcNAc residues is frequently followed by removal of these residues by a very active β-hexosaminidase (see Chapter 24). Thus, the final structure found in insects often has only the three core mannose residues (Figure 19.2). Prior to the removal of the GlcNAc residues in insect cells, an α1-3-linked fucose unit is often added to the core GlcNAc residue (frequently in addition to the α1-6-linked fucose typically found on the core GlcNAc of vertebrate N-glycans). Plants follow a pathway similar to that in vertebrates in the initial stages, but then often add a bisecting β1-2-linked xylose residue on the β-linked mannose residue (see Chapter 22 and Figure 19.2). The latter structure is also present in some invertebrates, but it appears to be immunogenic in vertebrates. In keeping with the above findings, the early-processing α-mannosidases of the N-glycan pathway have a wide evolutionary distribution. In contrast, the endo-α-D-mannosidase processing enzyme that provides an “alternate deglucosylating pathway” for N-glycans (see Chapter 8) appears to be limited to members of the chordate phylum, with the exception of the Mollusca, where it was detected in three distinct classes. The absence of this enzyme in other invertebrates examined, as well as in yeasts, various protozoa, and higher plants, suggests that the need for an alternate deglucosylation route paralleled the development of complex N-glycans in higher animals.

FIGURE 19.2

Dominant pathways of N-glycan processing among different taxa. See text for discussion. Symbol Key: Image symbol_key_small.jpg

Overall, it is clear that the N-glycan pathway is evolutionarily ancient and found throughout the eukaryotes. However, limited information exists regarding the presence and distribution of most N-glycan pathway enzymes and genes in most animals and plants. Thus, there is still insufficient information to paint a clear picture of exactly how it has been diversified and specialized during evolution. It was once thought that only eukaryotes express N-glycans. However, it is now clear that Archaea express GalNAc-Asn and Glc-Asn linkages, and recent data indicate that a few bacteria such as Campylobacter jejuni express a very similar N-glycan pathway, but with a novel linkage unit, involving bacillosamine (2,4-diacetamido-2,4,6-trideoxy-D-glucose or BacAc₂) linked to asparagine.

Shared Outer Chains of Glycans

Outer terminal sequences are often shared among N- and O-glycans and glycosphingolipids (Chapter 13). The common outer-chain Galβ1-4GlcNAcβ1- (N-acetyllactosamine or “LacNAc”) structure of vertebrates (see Chapter 13) is also found in plants. Some plants even add outer-chain Fucα1-3 residues to the GlcNAc residues of LacNAc units, generating Lewis^x-like structures identical to those found in animal cells. In some taxa, such as mollusks, an outer GalNAcβ1-4GlcNAcβ1- structure (the so-called LacDiNAc or LDN unit) tends to dominate, in place of the typical LacNAc structure more commonly seen in vertebrates. The SO₄-4-GalNAcβ1-4GlcNAcβ1- terminal units of pituitary glycoprotein hormones (see Chapter 13) have been conserved throughout vertebrate evolution, suggesting that they are critical for biological activity. Controversy exists as to whether or not further extensions and terminations of N-glycan antennae typical of vertebrates occur in insect or plant cells (see discussion regarding sialic acids below). The genetic elimination of the common vertebrate terminal sequence Galα1-3Galβ1-4GlcNAcβ1- in Old World primates and variations in sialic acids are discussed below.

Sialic Acids

Sialic acids are prominently expressed at the outer termini of N-glycans, O-glycans, and glycosphingolipids of the deuterostome lineage of animals (see Figure 19.1 and Chapter 14). It was once thought that sialic acids were an evolutionary innovation unique to this lineage that originated during the Cambrian Expansion, and that all other reports of sialic acids in a few scattered taxa reflected lateral gene transfer and/or convergent evolution (i.e., independent evolution of sialic acid synthesis in these taxa). However, although such lateral transfer mechanisms exist and may explain the presence of sialic acids in some bacterial taxa, sialic acids are also reported in some fungi and mollusks. Together with evidence for a limited set of genes for sialic acid production and addition in some protostomes (e.g., in insects such as Drosophila), the situation is indicative of an earlier evolutionary origin for sialic acids. Caenorhabditis elegans, the free-living nematode, does not contain genes for synthesizing or metabolizing sialic acid. In addition, prior claims of the presence of sialic acids in plants are probably due to environmental contamination and/or incorrect identification of the chemically related sugar Kdo (3-deoxy-octulosonic acid). However, recent studies have found that the sialic acid biosynthetic genes of some insect and bacterial species share homology with those of vertebrates. Overall, it appears likely that sialic acids were an ancient invention derived from genes of the related pathway for Kdo synthesis. In this scenario, sialic acids were differentially exploited during evolution, becoming prominent only in the deuterostome lineage, while being abandoned or substantially reduced in complexity and/or biological importance in other animal and fungal taxa. Meanwhile, a variety of bacteria synthesize other sialic-acid-like molecules using a very similar biosynthetic pathway (see Chapter 14).

It is curious that the most sialic acid diversity tends to be found in invertebrate deuterostomes, such as echinoderms (sea urchins and starfish), and the simplest profiles are found in humans. Likewise, whereas complex substituted polysialic acids are found in echinoderms and fish eggs, simpler polysialic acids are found in humans. Thus, sialic acids are not subject to incremental sophistication along recently evolved lineages. Rather, they seem to have evolved in many possible directions, disappearing altogether, or complicating or simplifying their structures. Although there is a tendency for some types of sialic acids to be dominant in certain mammalian species (e.g., N-glycolylneuraminic acid [Neu5Gc] in pigs and 4-O-acetylated sialic acids in horses), careful investigation reveals the presence of lower quantities of such sialic acids in most other species. Polysialosyl groups and sialic acid O-acetylation in gangliosides seem to be particularly enriched in poikilothermic (cold-blooded) animals. An interesting finding is that humans are “knockout” primates for the enzyme CMP–Neu5Ac hydroxylase (CMAH), because they contain a mutated and inactive CMAH gene. Thus, unlike the closely related great apes, humans are deficient in expression of Neu5Gc acid (see Chapter 14). As chickens also make an immune reaction against Neu5Gc, it remains to be determined whether birds represent another lineage that lost this sialic acid.

Glycosaminoglycans

Structures thought to be typical of “higher” animal heparan and chondroitin sulfate chains have been found in many invertebrates, including insects (see Chapter 24) and mollusks. The most widely distributed and evolutionarily ancient class appears to be chondroitin chains, which are not always sulfated (e.g., in C. elegans) (see Chapter 23). The more highly sulfated and epimerized forms of heparin and dermatan sulfate tend to be found primarily in “higher” animal species of the deuterostome lineage. The same is true of hyaluronan. Echinoderms such as the sea cucumber make typical chondroitin chains, but some glucuronic acids have branches containing fucose sulfate. Simpler multicellular animals such as sponges can have novel glycosaminoglycans that include uronic acids, but they do not have the typical repeat units of chondroitin sulfate and heparan sulfate. Plants do not have typical animal glycosaminoglycans. Instead, they have acidic pectin polysaccharides, characterized by the presence of galacturonic acid and its methyl ester derivative (see Chapter 22). Bacteria have completely distinct polysaccharides (Chapter 20), although certain pathogenic strains can mimic mammalian glycosaminoglycan chains (see below).

Glycosylphosphatidylinositol Anchors

Glycosylphosphatidylinositol (GPI)-anchored proteins and lipids (see Chapter 11) that share the “core” motif Manα1-4GlcNα1-6-myo-inositol-1-P lipid are distributed ubiquitously in eukaryotes. In some species (e.g., yeasts and slime molds), the lipid tail can be a ceramide instead of a phosphatidylinositol (see Chapter 21). GPI-anchored lipids and proteins can constitute the major components of the highly variable outer membranes of some parasitic protozoans, such as Leishmania and Trypanosoma (see Chapter 40). GPI anchors are generally thought to be absent in prokaryotes. However, at least one archaeal organism has been reported to have a GPI-anchored protein.

Nuclear and Cytoplasmic Glycans

The O-β-GlcNAc modification commonly found on cytoplasmic and nuclear proteins (see Chapter 18) is widely expressed in “higher” animals and in plants. Conserved homologs of the O-GlcNAc transferase that is responsible for synthesizing this structure have been found in many eukaryotic taxa. No clear homolog is evident in the yeast genome. Although the structure has been claimed in Dictyostelium, this is actually a distinct α-linked O-GlcNAc. There is currently no evidence that bacteria or Archaea can generate this modification.

VIRUSES ACQUIRE GLYCOSYLATION FROM THEIR HOSTS

Viruses often carry minimalist genomes that typically do not direct glycosylation of their own glycoproteins but instead utilize host-cell machinery. Thus, the glycosylation of viruses reflects that of host cells from which they emerge. However, there are some exceptions to this rule, with some viruses and especially bacteriophages containing genes encoding unusual glycosyltransferases. For example, the chlorella virus generates a glycoprotein termed PBCV-1 that is modified by a “Fringe-type” glycosyltransferase encoded in the viral genome. Some of the phage viral glycosyltransferases can modify their surface antigens to change the serotype of their host bacteria or glycosylate their own DNA to block it from degradation by restriction enzymes. Baculoviruses also encode their own glycosyltransferases to glycosylate host ecdysteroids, allowing them to block molting of the insect host.

Utilization of host glycosylation machinery is particularly prominent in the case of enveloped viruses. Most viral envelope glycoproteins are glycosylated (mostly with N-glycans) during the passage of these proteins through the host Golgi apparatus. This glycosylation is typically quite extensive and appears to protect the virus from host immune reactions directed against the underlying viral polypeptide. In this regard, it has been suggested that the relatively common occurrence of the heterozygous state for congenital disorders of glycosylation in humans (see Chapter 42) may reflect selection for heterozygous individuals whose genomes interfere with viral replication by preventing complete glycosylation of proteins of invading viruses. In other instances, host lectins may be “hijacked” by the glycans on viral surface glycoproteins, aiding attachment and/or entry into target host cells.

VAST DIVERSITY IN BACTERIAL AND ARCHAEAL GLYCOSYLATION

Despite the enormous potential for structural diversity built into monosaccharides, a rather limited subset of all possible monosaccharides and their possible linkages and modifications are found in eukaryotic cells. Why one encounters only such a limited subset of the possible glycan structures is one of the puzzling questions of glycobiology. On the other hand, this limited subset has allowed extensive elucidation of the structure of eukaryotic glycans. In contrast, Bacteriae and Archaea have had several billion additional years to experiment with glycan variation. These organisms also have short generation times and are capable of exchanging genetic material across vast phylogenetic distances, via plasmid-mediated horizontal gene flow. They also inhabit a much wider range of ecological niches with innumerable physicochemical and biological conditions, ranging from the deep litho-and hydrosphere to the stratosphere. Thus, it should not come as a surprise that bacteria and Archaea express a much greater diversity in glycosylation, both in terms of range of monosaccharides that they utilize or synthesize and with regard to their types of linkages and modifications. Some discussion of such structures is presented in Chapter 20. However, much of the work to date has focused on the glycans of pathogens, and it is safe to say that we have barely scratched the surface of this diversity.

MOLECULAR MIMICRY OF HOST GLYCANS BY PATHOGENS

It is evident that great differences exist between the pathways generating the glycan structures of bacteria and those of vertebrates. Despite this, occasional microbial surface structures are found to be strikingly similar to those of mammalian cells. Interestingly, most examples of this type of “molecular mimicry” occur in pathogenic microorganisms, presumably adapting them for better survival in the host by avoiding, reducing, or manipulating host immunity. A few examples are listed in Table 19.1. The initial hope of scientists trying to clone vertebrate glycosyltransferases was that most of the responsible microbial genes arose from lateral gene transfer and that these would provide a backdoor approach to isolating the corresponding ones from eukaryotes. However, in most instances where full genetic information has become available, the evidence points toward convergent evolution rather than gene transfer as the dominant mechanism. For example, the genes involved in synthesizing sialic acids in bacteria seem to have been mainly derived from the preexisting bacterial pathways for the biosynthesis and transfer of Kdo, a bacterial sugar with a structural resemblance to sialic acids. Meanwhile, bacterial sialyltransferases bear little resemblance to those of eukaryotes, and the vast sequence differences between different bacterial sialyltransferases indicate that these have even been reinvented on several separate occasions. On the other hand, lateral gene transfer appears to have been quite common among the bacteria and Archaea themselves, facilitating rapid phylogenetic dissemination of such enzymatic “inventions.”

TABLE 19.1

Examples of molecular mimicry of animal glycans by pathogenic bacteria

INTERSPECIES AND INTRASPECIES DIFFERENCES IN GLYCOSYLATION

Why do closely related species differ with regard to the presence or absence of certain glycans? Does the same glycoprotein have the same type of glycosylation in different but related species? Relatively little data are available concerning these issues, but examples of both extreme conservation and extreme diversification can be found. A reasonable explanation is that conservation of glycan structure is only required when there are very specific functions for the glycans in question. In other instances, considerable drift in the details of glycan structure might be tolerated, as long as the underlying protein is able to carry out its primary functions (changes with no consequences for survival or reproduction, i.e., selectively neutral).

Even in the absence of important functions within an organism, glycans can have important roles in the mediation of interactions with symbionts and pathogens. The evolution of diversity and microheterogeneity (across tissues and cell types) in glycosylation could well be of value to the organisms in evading pathogens that use glycans as signposts for attachment and entry. Glycans can also have important roles in attracting the important symbiont microbial communities needed for gastrointestinal functions and in accommodating or restricting these to particular areas of the host.

It is also clear that there can be significant variation in glycosylation among members of the same species, particularly with regard to terminal glycan sequences. The classic example is that of the ABH(O) blood group system (see Chapter 13), a glycan-defined polymorphism found in all human populations, which has also persisted for tens of millions of years of primate evolution and has even been independently rederived in some instances. Somewhat surprisingly, despite its great clinical importance for blood transfusion, this polymorphism appears to cause no major differences to the intrinsic biology of individuals of the species (see Chapter 13). Like other blood groups, the ABO polymorphism is accompanied by the production of antibodies against the other variants. It has been suggested that these antibodies are protective, by causing complement-mediated lysis of enveloped viruses generated within other individuals who can express the target structure for the antibody. Thus, an enveloped virus generated in a B blood group individual might bear this structure and be susceptible to lysis upon contact with an A or O blood group individual, who would express anti-B antibodies. Recent experimental evidence is supportive of such a mechanism. However, this mechanism alone should strongly favor O individuals, as these form antibodies against the A and B variants and should lead to higher frequencies of O type than are observed.

Another possible explanation for interspecies diversity is the selection exerted by pathogens that recognize glycans as targets for attachment and entry into cells. This mechanism is likely operative in generating the diversity of sialic acid types and linkages (see above). However, it should result in selection of ABO subtypes and result in approximately even frequencies of each phenotype, not what is observed. Recent analyses have tried to combine the two mechanisms: the antibody-mediated protection from intracellular viruses and possible frequency-dependent protection from glycan-exploiting extracellular pathogens, such as Noroviruses and Plasmodium falciparum malaria. Modeling approaches have successfully generated observed frequencies of ABO by incorporating these two simultaneous selection pressures. It is fair to say that the evolutionary persistence of the ABO system needs further explanation.

Another unexplained phenomenon is the genetic inactivation in Old World primates of the ability to synthesize the otherwise very common terminal Galα1-3Galβ1-4GlcNAc-R structure. This glycan variation system is also associated with spontaneously appearing and persistently circulating antibodies against the missing glycan determinant, thus forming a kind of interspecies “blood group.” It has been proposed that this glycan difference is protective for the primate lineage which lost “α Gal” and has a high-titer circulating antibody, as it is now better protected against infection by viruses emanating from other mammals.

Regardless of the precise underlying purposes of these types of polymorphic systems, such intra- and interspecies diversity might also provide for “herd immunity,” a phenomenon whereby one glycan-variant-resistant individual can effectively protect other susceptible individuals by limiting the spread of a pathogen through the population. It is also important to emphasize that these proposed protective functions of glycan diversity are only apparent at the level of populations and not the individual. This complicates their study in model organisms, where the focus is classically on the individual.

Future studies will have to test precisely how much of interspecies and intraspecies glycan variation is directly driven by such host–pathogen interactions. Despite a lack of comprehensive studies of such phenomena, it is becoming clear that glycan variation forms an important determinant of host susceptibility and must be considered when trying to understand disease, especially epidemics or zoonotics involving different host species and their interactions, for example, influenza A (see Chapter 14).

“MODEL” ORGANISMS FOR STUDYING GLYCAN DIVERSITY

Details of glycan expression patterns in various popular “model” organisms can be found in other chapters in this volume. In recent years, there has been increasing definition of the structures of bacterial inner cell wall peptidoglycans and the outer membranes that are composed of lipooligosaccharides and lipopolysaccharides, particularly those of Escherichia coli (see Chapter 20). Chapter 21 provides some details about various genera and species of fungi and protists, including Saccharomyces cerevisiae and Dictyostelium discoideum. Pathogenic protists such as trypanosomes and leishmanial parasites express very high densities of surface GPI anchors and are discussed in Chapters 11 and 40. Chapter 23 presents an overview of the roundworm C. elegans, its development, glycan structures, and expression of glycosyltransferases. The functional insights derived from this organism cover all classes of glycans. For some details about glycobiology of Drosophila, see Chapter 24. Like C. elegans, this workhorse of genetics has made important functional contributions to the field during the last 5 years. It has been especially important in understanding how O-fucosylation modifies Notch signaling (Chapter 12) and how heparan sulfate proteoglycans determine morphogen and growth factor gradients. Chapter 22 discusses the glycans of plants, including those of the model organism, Arabidopsis. Various aspects of sea urchin glycobiology, including the acrosome reaction, egg/sperm interactions, and the role of proteoglycans and lectins, are covered in Chapter 25, which also discusses aspects of Xenopus glycobiology, including the synthesis of chitin oligosaccharides, the role of proteoglycans in determining left–right asymmetry, and lectins that help in fertilization and the innate immune system. The same chapter also discusses aspects of zebrafish glycobiology, including glycoproteins, proteoglycans, and lectins.

The recent discovery that rodents are the closest evolutionary cousins to primates has provided added justification for the use of rats and mice as model organisms to understand the mechanisms of human disease (see Chapter 25). Last but not least, Nobel Laureate Sydney Brenner has suggested that we now have enough information about humans to consider ourselves to be the “next model organism.” Indeed, there is an increasing tendency to focus tractable questions about glycans and their biology directly on humans and on naturally occurring human mutants. Most recently, there has also been interest in studying the great apes (our closest evolutionary cousins) and an independent realization of recent hominid evolution, particularly with regard to several differences in sialic acid biology (see Chapter 14).

WHY DO WIDELY EXPRESSED GLYCOSYLTRANSFERASES SOMETIMES HAVE LIMITED INTRINSIC FUNCTIONS?

Prior to the generation of glycosyltransferase-deficient mice, it was popular to suggest that every single glycan on every single cell type must have a critical intrinsic host function. Analysis of available gene disruption data indicates that this is not the case. For example, the ST6Gal-I α2-6 sialyltransferase is the main enzyme that produces Siaα2-6Galβ1-4GlcNAcβ1- termini on vertebrate glycans. Although this sequence serves as a specific ligand for the B-cell regulatory molecule CD22 (Siglec-2; see Chapter 32), it is also found on many other cell types, as well as on many soluble secreted glycoproteins. Furthermore, the ST6Gal-I mRNA varies markedly among cell types, and its transcription is regulated by several cell-type-specific promoters, which are in turn modulated by hormones and cytokines. Despite all these data suggesting very diverse and complex roles for this enzyme and its products, the prominent functional consequences of eliminating its expression in mice so far seem to be restricted to the B cell, with decreased signaling and proliferative responses and impaired antibody production (see Chapter 32). Few other obvious abnormalities have yet been found in organ structure and physiology, morphology, or behavior. If the specific intrinsic functions of the ST6Gal-I glycan product are in fact restricted to B cells, why does the organism express it in so many other locations? Even more puzzling, why up-regulate its expression so markedly in the liver and endothelium during a so-called “acute phase” inflammatory response? Could it be that scattered expression of this structure in other locations represents a “smoke-screen” effect, restricting intraorganismal spread of an invading pathogen? Could it be that heavily glycosylated nonnucleated cells like erythrocytes act as a “sink” to divert viral pathogens that need nucleated cells for replication? The answers to these questions must take into account the evolutionary selection pressures (both intrinsic and extrinsic recognition phenomena such as host–pathogen interactions and innate immune contributions) on glycosyltransferase products. Many of these effects may also not be apparent in inbred genetically modified mice living in hygienic vivaria but may rather require population studies of animals in a natural, pathogen-rich environment. It is also possible that other gene products are masking the phenotypes in these model systems, by compensating for the genetic loss. Furthermore, it is likely we have not looked hard enough at such genetically modified mice nor applied the right environmental pressures to elicit phenotypes.

EVOLUTIONARY FORCES DRIVING GLYCAN DIVERSIFICATION IN NATURE

There is too little information available today to allow a comprehensive exposition of the evolution of even the major classes of glycans. On the basis of the available data, it is reasonable to suggest that glycan diversification in complex multicellular organisms has been driven by evolutionary selection pressures of both intrinsic and extrinsic origin relative to the organism under study (see Chapter 6). It is reasonable to postulate that glycans are particularly susceptible to the “Red Queen” effect, in which host glycans must keep on changing in order to stay ahead of the pathogens, which have extremely rapid evolutionary rates because of short generation times, high mutation rates, and much horizontal gene transfer. Given the rapid evolution of extrinsic pathogens and their frequent use of glycans as targets for host recognition, it seems likely that a significant portion of the overall diversity in vertebrate cell-surface glycan structure reflects such pathogen-mediated selection processes. Meanwhile, even one critical intrinsic role of a glycan would disallow its elimination as a mechanism to evade pathogens. Thus, the glycan expression patterns of a given organism may represent a compromise between evading pathogens and preserving intrinsic functions.

More gene disruption studies in intact animals would be helpful to differentiate between these intrinsic and extrinsic glycan functions. More systematic comparative glycobiology could also contribute, by making predictions about intrinsic glycan function; that is, the consistent (conserved) expression of the same structure in the same cell type across several taxa would imply a critical intrinsic role. Such work might also help define the rate of glycan diversification during evolution, better define the relative roles of the intrinsic and extrinsic selective forces, and eventually lead to a better understanding of the functional significance of glycan diversification during evolution. The possibility that glycan diversification might even drive the process of speciation (via reproductive isolation) also needs to be considered.

Bookshelf