The Structural Diversity of Natural Glycosphingolipids (GSLs)

Abstract

Glycosphingolipids (GSLs) are a subclass of glycolipids made of a glycan and a ceramide that, in turn, is composed of a sphingoid base moiety and a fatty acyl group. GSLs represent the vast majority of glycolipids in eukaryotes, and as an essential component of the cell membrane, they play an important role in many biological and pathological processes. Therefore, they are useful targets for the development of novel diagnostic and therapeutic methods for human diseases. Since sphingosine was first described by J. L. Thudichum in 1884, several hundred GSL species, not including their diverse lipid forms that can further amplify the number of individual GSLs by many folds, have been isolated from natural sources and structurally characterized. This review tries to provide a comprehensive survey of the major GSL species, especially those with distinct glycan structures and modification patterns, and the ceramides with unique modifications of the lipid chains, that have been discovered to date. In particular, this review is focused on GSLs from eukaryotic species. This review has listed 251 GSL glycans with different linkages, 127 glycans with unique modifications, 46 sphingoids, and 43 fatty acyl groups. It should be helpful for scientists who are interested in GSLs, from isolation and structural analyses to chemical and enzymatic syntheses, as well as their biological studies and applications.

Graphical Abstract

1. Introduction

Glycosphingolipids (GSLs) are a subclass of glycolipids and the major glycolipids in the cell membrane of all eukaryotic species. In fact, the vast majority of glycolipids in vertebrates are GSLs.^[1] As depicted in Figure 1, GSL consists of a hydrophilic glycan and a lipophilic ceramide (Cer), which are coupled together by a glycosidic linkage. In turn, the ceramide moiety of GSL is composed of a sphingoid base, typically a sphingosine (an example is provided in Figure 1) and its derivatives in mammals, and a fatty acyl group that is attached to the 2-N-position of sphingosine. GSLs are associated with the cell membrane through imbedding their lipophilic tails into the lipid bilayer, leaving the glycans on the cell surface as plasma membrane outer leaflets.

Figure 1.

A typical GSL (e.g., LacCer) consists of a glycan and a ceramide, while the latter is composed of a fatty acyl group [e.g., stearic group (C18:0)] and a sphingoid [e.g., sphingosine (d18:1)] as illustrated.

GSLs are essential components of the cell membrane. They interact with the complementary molecules on other cells or with other membrane components on/in the same cell to modulate various functions, such as cell recognition, adhesion, division and apoptosis,^[2–4] signal transduction,^[5] etc. Consequently, GSLs are directly involved in many biological and pathological processes.^[6,7] For example, GSLs are found to play a critical role in brain development^[8] and immune responses,^[9,10] while abnormal GSL expression is closely related to many central nervous system-related diseases^[11–13] and oncodevelopments.^[14,15] GSLs have been discovered in species from prokaryotes to eukaryotes and from animals to plants, and there is a large body of literature concerning GSL research. However, this review is limited to the structural diversity of GSLs discovered in nature, especially those found in eukaryotes.

2. Structural Classification

The structures of GSLs are complex and diverse due to the possibility of dual variations in their glycan and ceramide moieties. The glycan structure of a GSL defines its specific species, while GSLs having the same glycan but different ceramide moieties are referred to as different lipid forms of the same GSL species. To date, over five hundred GSL species carrying different glycans have been identified.^[16,17] Some glycans comprise different types and numbers of sugar residues, whereas others contain the same sugar residues but have them in different linkage forms and positions or modified with various functional groups. On top of this structural diversity, each GSL species can have numerous lipid forms with variations in the chain length, branching pattern, and unsaturation degree for both the sphingoid and the fatty acyl moieties. For example, the fatty acyl group in ceramide can vary widely from C14 to C30.^[1]

Typically, GSLs are further classified into subgroups or series according to the different structures and physicochemical properties of their glycans or sometimes according to their different biological functions. Based on the physicochemical properties of their glycans, GSLs are divided into neutral and acidic GSLs. The former contains only neutral sugar units, but the latter contains at least one acidic sugar residue, most commonly N-acetylneuraminic acid (Neu5Ac, also widely known as sialic acid).^[17,18] Sialic acid-containing acidic GSLs are traditionally referred to as ganglioside, while GSLs containing sulfated glycans are known as sulfatide. GSLs carrying only a monosaccharide are called cerebroside, and its most common members are galactosylceramide (GalCer, galactocerebroside) and glucosylceramide (GlcCer, glucocerebroside) that have a D-galactose (Gal) and a D-glucose (Glc) residue β-linked to a ceramide, respectively. More recently, α-linked galactosylceramide and glucosylceramide are also discovered in mammalian natural killer T cells at very low concentrations.^[19] Based on their core carbohydrate structures, GSLs are classified into various series, as listed in Table 1,^[20] and each series can contain a number of subseries. For example, ganglio-series GSLs are further divided into asialo-, a-, b-, and c-series having 0, 1, 2, and 3 sialic acid residue(s) attached to the inner core Gal residue, respectively, while those with a sialic acid attached to the inner core N-acetylgalactosamine (GalNAc) residue are called α-series gangliosides.^[20] The subgroup and series names of GSLs are often correlated to their structures or biological origins. For example, cerebroside, which is the most abundant GSL in brains, comes from the Latin word “cerebrum”—meaning brain, and ganglioside is so named because of the abundance of this group of GSLs in “ganglion”. The structures of lacto-series GSLs can become very complex and diverse upon elongation or branching of the glycans with lactosamine (LacNAc) as a repeating unit.

Table 1.

The names of various GSL series and their core carbohydrate structures^[1,20]

^aAbbreviations (all sugars are in pyranose form except for those specifically defined): AEP: 2-aminoethyl-phosphonyl; Ara: L-arabinose; Cer: ceramide; Fuc: L-fucose; Gal: D-galactose; Galf: D-galactofuranose; GalNAc: N-acetyl-D-galactosamine; Glc: D-glucose; GlcA: D-glucouronic acid; GlcNAc: N-acetyl-D-glucosamine; GlcNH₂: D-glucosamine; KDN: 2-keto-3-deoxy-D-glycero-D-galactonononic acid; Man: D-mannose; Neu5Ac: N-acetylneuraminic acid; Neu5Gc: N-glycolylneuraminic acid; Neu5NH₂: neuraminic acid; NMAEP: 2-(N-methylamino)ethyl-phosphonyl; PC: phosphocoline; Rha: L-rhamnose; Sph: sphingosine; Xyl: D-xylose.

^bThe symbols used to represent the monosaccharides mentioned in this paper are:

According to the biosynthetic pathways for GSLs,^[20–22] the majority of GSLs should be derived from GalCer and GlcCer. However, only a small number of GalCer-based GSLs (within the gala-, neogala- and spirometo-series) are found because, typically, elongating the GalCer glycan is restricted. As a result, most mammalian GSLs are generated from GlcCer, and these GSLs belong to three main groups, including the ganglio-/isoganglio-series, the globo-/isogalobo-series, and the lacto-/neolacto-series. The structures of the core glycans for the ganglio- or isoganglio-series and the globo- or isoglobo-series differ only in the linkage site of the third sugar residue (^IIIGalNAc or ^IIIGal) to the conserved core ^IIGalβ1–4^IGlcβ1–1’Cer (LacCer) motif. It is the ^IIGal 4-O-position for ganglio- and globo-series but the ^IIGal 3-O-position for isoganglio- and isoglobo-series. The only structural difference between the core glycans of the lacto- and neolacto-series is the linkage site of their fourth sugar unit (^IVGal) to the third sugar unit [i.e., N-acetyl-D-glucosamine (^IIIGlcNAc)]. The core structures of some GSL series, for instance, the lactoganglio-series, are hybrids of different GSL core glycans. Besides these main groups, there are also other series of GlcCer-derived GSLs found in various organisms.

The conventional method to name GSLs, which was first coined by Svennerholm,^[23,24] is based on their series names and the number, type, linkage form and location of sugar residues in the glycan. For example, gangliosides are usually called (e.g.) GM1, GD1, GT1, GQ1 …, where G refers to ganglio-series, the second letter (M, D, T, and Q) refers to the number of sialic acid residues (i.e., mono, di, tri, and quad, respectively) in the glycans, and the number (1, 2, 3, etc.) refers to the order of gangliosides displaying on the thin-layer chromatography (TLC) plate (with R_f values: GM3 > GM2 > GM1).^[25,26] The latter property is correlated with the length and composition of a glycan. On the other hand, Gb3, Gb4, and Gb5 simply refer to globo-series GSLs containing a tri-, tetra-, and pentasaccharide glycan, respectively. However, many of the less common GLSs do not receive broadly accepted common names. IUPAC has recommended a systematic method to nomenclate GSLs (https://iupac.qmul.ac.uk/misc/glylp.html), but since these official names are rather long and complex they are not widely adopted for the relatively common GSLs.

3. Diversity of Glycans

A large number of GSL species (more than 500, which does not include different lipid forms for each GSL) have been characterized. Among them, some are ubiquitous and abundant across different organisms, whilst others are rare or only amply expressed in certain species and tissues or under specific pathological conditions. For example, many of the neutral GSLs are related to the blood type determinants, whereas gangliosides are especially enriched in the central nervous system. Usually, GSLs discovered in the low species of the tree of life are more diverse than those in high animals, such as mammals. Listed in Table 2 are the major GSL glycans identified so far, especially those having distinctive carbohydrate frameworks found in eukaryotes. More complex GSLs with similar glycan motifs formed through further elongation and branching of the characteristic glycan structures with repeating units, such as LacNAc, are not included herein; many of these complex GSLs are presented in the review by Yu and coworkers.^[27] GSLs with the same carbohydrate sequences but consisting of de-N-acetylated amino sugars [such as neuraminic acid (Neu5NH₂) and glucosamine (GlcNH₂) generated through de-N-acetylation of Neu5Ac and GlcNAc, respectively] and uniquely N-acylated amino sugars [such as N-glycolylneuraminic acid (Neu5Gc)] or having some sugar residues further modified with functional groups are not included in this table either, because these GSLs will be separately discussed later in this review. The same GSLs are usually found from multiple sources, such as in different tissues and animals or under different conditions, and typically, there are many studies and reports about the important GSLs. In this table, however, we have included only the references and/or the sources from which the GSL was first found or the conditions, such as diseases, under which the GSL is most abundant.

Table 2.

The structures of major GSL glycans discovered in nature

The structures listed in Table 2 clearly indicate that the GSL glycans are almost exclusively composed of the pyranose of several common monosaccharides, including Gal, GalNAc, Glc, GlcNAc, Fuc, Man, and sialic acids (both Neu5Ac and its analogs). Despite the rarity, GSLs consisting of other types of sugars do exist in nature. For example, in recent years, a number of GSLs containing deoxy sugars including xylose (Xyl)^{[184,186,187,189]} and rhamnose (Rha),^[224] furanose [e.g., galactofuranose (Galf)],^[221,222] and carbohydrate uronic acid [e.g., glucouronic acid (GlcA)]^{[186,198,200,223,224]} have been characterized. However, it seems that these special GSLs are limited to insects and aquatic species, and they are not common in mammals. It is also worth noting that fungi, aquatic animals, insects, and some other low species in the tree of life produce structurally more diverse and intriguing GSLs, many of which contain unique core glycans. For example, the biosynthesis of some GSLs in ascidians starts with GlcACer,^[223,224] which is rather rare.

4. Functionalization of Glycans

Many other GSLs have the same carbohydrate sequences as those depicted in Table 2 but their glycans are further modified with functional groups, such as sulfates and acetyl groups, or contain Neu5NH₂ and GlcNH₂ and alternatively N-acyl forms of Neu5NH₂, such as Neu5Gc. These modifications on the glycan can significantly affect the physicochemical and, thereby, biological properties of GSLs. For example, sulfates can readily dissociate under physiological conditions to result in negatively charged glycans. On the other hand, de-N-acylation results in free amino groups in the sugar residues, which can be easily protonated to bear positive charges. O-Acetylation and O-methylation can alter the polarity of glycans and disrupt the ability of sugar residues to form hydrogen bonds. Such changes in physicochemical properties can have a huge impact on the interaction of GSLs with other biomolecules. Consequently, these modifications are either functionally required, e.g., to mediate signal transduction, cell differentiation, organ development and so on, or are directly involved in various pathological processes,^[225,226] and hence these uniquely modified GSLs are useful targets for the development of new therapies.^[226]

Sulfation typically occurs to the 3-O-position of Gal and GalNAc residues and only occasionally to the 6-O-position of GlcNAc and Glc^[227–229] or 8-O-position of Neu5Ac. O-Acetylation can happen to multiple sites of the Gal and Glc residues in simple GSLs, e.g., GalCer and GlcCer, but in more complicated ones, acetylation is mainly found in gangliosides and occurs to the 9-O- and 7-O-positions of Neu5Ac or other sialic acids. Lactonization is a unique phenomenon for Neu5Ac but rather common for some gangliosides, such as GM3. This can significantly affect the biological properties of gangliosides, e.g., immune responses to the glycans,^[230,231] and thus ganglioside lactones have been actively exploited for therapeutic applications. Other interesting but less common modifications of GSL glycans include O-methylation^{[186,187,192,222,232–235]} and O-ketalation^[236–239] of some sugar residues, as well as attachment of a phosphonyl or phosphoryl group, e.g., 2-aminoethyl-phosphonyl (AEP)^{[198,200,240–242]} and 2-(N-methylamino)-ethyl phosphonyl (NMAEP)^[206] groups and phosphocoline (PC),^{[197,241,243]} to Gal, GlcNAc, and Man residues, usually at the 6-O-position. Neu5Gc-containing gangliosides are common in animals and occasionally in some human tissues.^[244] On the other hand, Neu5NH₂- and GlcNH₂-containing GSLs are rare but exist.^[245–248] Removal of the long fatty acyl chain in the ceramide moiety, typically by acid ceramidases under physiological conditions, gives rise to another group of unique GSLs that are known as lyso-GSLs. Lyso-GSL accumulation is often associated with human diseases, and lyso-GSLs are important disease markers.^[249] GSLs modified with sulfates, acetyl groups, and other functional groups and GSLs containing Neu5Gc, Neu5NH₂, and GlcNH₂ residues, as well as lyso-GSLs, are listed in Table 3.

Table 3.

The structures of GSLs containing lactonized, phosphorylated, sulfated, O-acetylated, O-ketalated, and O-methylated glycans or Neu5Gc, Neu5NH₂ and GlcNH₂ sugars, as well as lyso-GSLs

^aO-Ketalation: The ketal is formed between the sugar diol and the carbonyl group of a pyruvate moiety.

4. Diversity of Lipid Forms

The ceramide structure of GSLs can also be variable. Besides deleting the long fatty acyl chain to form lyso-GSLs as mentioned above, structurally diverse ceramides may attach to each GSL glycan in the same cell or in different tissues and animals and under different conditions to generate distinct lipid forms. In this context, both the sphingoid and the fatty acyl group may differ, e.g., having varied degrees of unsaturation and/or branching and substitution patterns, in addition to varied chain length.^[316–319] As the lipid structure can have a significant impact on the organizations of GSLs in the cell membrane and their interactions with corresponding receptors and/or other biomolecules, ^[320–323] changes in the GSL ceramide structure can have important biological consequences.^[324] For example, variations in GSL lipid forms are associated with Alzheimer’s and other central nervous system diseases.^[325–327]

Ceramide and sphingoid research itself is a broad and important subject, because ceramide is not only found in GSLs but also in free or protein-bound forms and within many other lipids. Therefore, it has shown diverse biological functions.^[328–331] In the meantime, novel ceramides and sphingoids are being constantly discovered and reported. Consequently, instead of covering all ceramides identified thus far, this review is only focused on the common ceramides found in GSLs and the recently discovered ceramides possessing unique structures, to demonstrate the significant structural diversity of lipids in GSLs. To find novel GSL ceramides, the most convenient method is to analyze GlcCer and GalCer, although different ceramides and lipid forms have also been identified in complex GSLs. This is because the modest structures of GlcCer and GalCer make their extraction and structural study simple. As the number of potential ceramides formed from various combinations of different sphingoids and fatty acyl groups can be huge, the structures of sphingoids and fatty acyl groups are separately presented in Table 4.

Table 4.

The structures of some key sphingoids and fatty acids that constitute the ceramides in GSLs

The length of the carbon chains of the sphingoid and fatty acyl moieties in GSLs can change in a large range,^[1,17,354] whereas their most common form contains 18 carbons. Sphingosine (d18:1) and its homologs are the most abundant sphingoids in mammals,^[355] although their hydrated and hydrogenated analogs, i.e., 4-hydroxysphigosines and dihydrosphingosines, are also abundant. Additional C=C double bonds regularly occur at different locations of the carbon chain to generate a variety of sphingedienes and sphingetrienes. Some other frequent and intriguing modifications of sphingosine, especially in the low species of the tree of life, include the addition of methyl or methylene groups to form branches or rings and oxidation of some carbons to generate hydroxyl or oxo groups at various positions of the carbon backbone. N-Methylation of sphingosine is a modification observed in the urine of Fabry patients.^[333] This seems to be unique and only present in lyso-GSLs, thus representing a promising biomarker for the disease. Similarly, the most common fatty acyl moieties in the ceramides of mammalian GSLs are stearic group (C18:0) and its homologs, e.g., palmitic (C16:0) and arachidic (C20:0) groups. However, the composition of acyl groups in various lengths changes significantly for different GSL species and during different developmental stages. For example, the proportion of very long chain acyl groups, such as C24:0 and C24:1, increases with maturation and in simple GSLs.^[356–361] The most frequent modifications of acyl groups are α-hydroxylation and desaturation at various locations in the carbon chain. Interestingly, substituting the carbonyl oxygen atom of the fatty acyl group with a sulfur atom happens to the α-galactosylceramide in the mouse invariant natural killer T (iNKT) cell.^[353] This triggered a significant change in the ability of this antigen to stimulate iNKT cells in vivo. In addition, β-glucosylation at the α-position of the fatty acid was also observed with GlcCer,^[219] but its biological relevance is not clear yet.

5. Conclusion

Clearly, the structures of natural GSLs are highly complex and diverse. Such complexity and diversity come from the variabilities of both the glycan and the ceramide moieties in GSLs, and are believed to be responsible for their broad and important biological and pathological functions. Concerning the metabolism and functions of GSLs, there are many reviews in the literature. Although this article tries to cover all major types of GSLs (251 species in Table 2 and 127 species in Table 3), as well as various potential variations in their lipid forms (46 sphingoids and 43 fatty acids in Table 4), to demonstrate their great diversity, there are still many GSLs and ceramides that are not included herein, especially the elongated GSLs with glycans of similar characteristics or epitopes and ceramides containing common homologs or not directly associated with GSLs. Furthermore, with the advancement in analytical technologies, especially various MS methods, novel GSLs and ceramides are discovered and reported continuously. This will definitely further increase the structural pool and diversity of natural GSLs. Although our current understanding on the most common GSLs with regard to their structures, functions, and relationships with some diseases has been expanding on a daily basis, there is still a lack of fundamental knowledge about the large majority of GSLs. One of the important hurdles in the field is the difficulty to access complex GSLs in sufficient quantity and purity for biological studies. Another hurdle in the field is associated with the difficulty to perform large-scale, comprehensive, and quantitative profiling of global GSLs (GSLcomes) in different tissues and species or generated under different conditions, such as normal physiological and abnormal pathological conditions. However, we are optimistic that with the presently rapid progresses in carbohydrate chemistry to improve the accessibility of various GSLs and in analytical chemistry to provide high-throughput methods for GSL analysis, future studies and understanding of GSLs will grow exponentially. This would eventually help the development of new diagnoses and therapies for diseases through targeting at GSLs.