1.1 Sulfatide structure

Sulfatide is a family of sulfated galactosylceramides, in which a residue of galactose is sulfated at position 3 and is connected through a β-glycosidic bond to the primary hydroxyl group of a N-acylated d-erythro-sphingosine base, typically with a C-18 backbone (Fig. 1). Sulfatide exhibits variation of its structure in the composition of the fatty acid residue that acylates the amino group of sphingosine. The alkyl or alkenyl residues are of different lengths and they can be also hydroxylated. This structural variability shows a cell type-specific pattern, which at least in part can be explained by the different expression of the enzymes involved in sulfatide biosynthesis. The nervonic acid (C24:1)-containing sulfatide is the most abundant in myelin, while significantly elevated levels of stearic acid (C18:0)-containing sulfatide are found in the cortical grey matter [5]. Mammalian kidney sulfatides are instead particularly rich in C22:0 acid, which is found in quantities more than 10 times higher than those in the brain. On the other hand, mass spectroscopic analyses of pancreatic sulfatide revealed a high proportion of short chain C-16 fatty acid and no hydroxylated form [6].

1.2 Biochemical pathways

The synthesis of sulfatide takes place in the endoplasmic reticulum (ER) and the Golgi apparatus. It starts in the ER with the addition of galactose from UDP-galactose to ceramide catalyzed by the UDP-galactose:ceramide galactosyltrasferase (CGT; EC 2.4.1.45) to give galactosylceramides (GalCer). GalCer is then transported to the Golgi apparatus where sulfatide is synthesized by regioselective sulfation catalyzed by cerebroside sulfotrasferase (CST, EC 2.8.2.11). It has been proven that these enzymes have no isozymes. The distribution of sulfatide is tissue-specific and its expression is mainly regulated by the tissue-specific expression of the CGT and CST genes [7].

The degradation of sulfatide occurs mainly in lysosomal compartments within cells where the sulfate group is hydrolyzed by arylsulfatase A (ASA; EC 3.1.6.8). This reaction requires the presence of a sphingolipid activator protein-1 (SAP-1), also called saposin B (SapB), which extracts sulfatide from membranes and makes it accessible to ASA.

1.3 Biological significance

Sulfatide is a biologically multifunctional glycolipid involved in the nervous system, diabetes, virus infection and in the immune system. As the major glycolipid component of the myelin sheath, where it is exclusively found on the extracellular leaflet of the cell membrane, it has a role in the development and maintenance of a compact myelin structure able to exert its functions [8]. The accumulation of sulfatide in the central and peripheral nervous system leads to the human autosomal recessive demyelinating neurodegerative disorder known as metachromatic leukodystrophy (MLD). MLD is a degenerative disease of the central nervous system with essentially motor manifestations and psychosis and is caused by deficiency of the lysosomal enzyme ASA or mutations in the gene encoding for SapB [9]. Furthermore, sulfatide is a major L-selectin ligand in kidney and this interaction plays a central role in monocyte infiltration into the kidney interstitium and hence in the induction of inflammatory response [10]. On another side, sulfatide is also expressed by pancreatic β-cells where it is localized in the same cellular compartment as insulin and is probably involved in processing, maintenance and secretion of insulin [6].

Sulfatide is also involved in human immunodeficiency virus (HIV-1) infection [11]. The entry of HIV-1 into the cell is a complex process mediated by the viral envelop glycoprotein gp120 which interacts mainly with the CD4 molecules at the surface of the cell. Nevertheless, gp120 is also able to bind other receptors including sulfatide. In particular, sulfatide acts as a secondary attachment site during infection, as an alternative binding receptor in the infection of CD4-negative cells of neural and intestinal tissues and participates actively in transmembrane signaling.

Studies on the activation of the immune system mediated by CD1 proteins (CD1s) showed that sulfatide is an antigen of CD1s. CD1 proteins are a family of antigen presenting molecules that after lipid and glycolipid recognition are able to activate the immune cascade in a way different and complementary to the one mediated by the proteins of the major histocompatibility complex (MHC). Sulfatide has been described as a promiscuous ligand of group I CD1 molecules that binds different CD1 isoforms stimulating specific T-cells [12].

Significant changes in sulfatide levels have also been observed in Alzheimer's as well as other diseases, suggesting a role in the pathogenesis of these disorders [13].

This stimulating biological background reveals the need for scientists to have access to quantities of sulfatides necessary for experimental studies. Most biological work has been made with sulfatides from natural sources, e.g., isolated from pig brain [14] or bovine brain [15], or semisynthetic sulfatides, some of them commercially available. This material has been used directly or after simple synthetic transformation, such as the removal of the acyl chain to lyso sulfatide and the re-esterification with the desired fatty acid [6,16] or enzymatic modification of the natural glycolipid [17]. Nevertheless, the material from natural origin can be sometimes contaminated with some isoforms and this has led to the development of synthetic strategies to obtain sulfatide by chemical synthesis in high purity and precise lot consistency.

The aim of this account is to show the synthetic work, correlated to the biological background, that has been described for the preparation of sulfated galactocerebrosides, with particular focus on the syntheses of sulfatide and of some significant related compounds modified on the galactose moiety.

2.1 The sulfation reaction

The presence of the sulfate in position 3 is the characterizing element which differentiates sulfatide from cerebroside (β-d-GalCer). Nevertheless, as mentioned above, the introduction of the sulfate is not difficult to accomplish. It can be easily achieved by reacting the target hydroxyl group with sulfur trioxide-pyridine or trimethylamine complex. This approach requires a protecting group strategy that allows during the synthesis the selective deprotection of the hydroxyl group to be sulfated. In the case of sulfatide, a great advantage, in terms of a significant reduction of the synthetic steps, was introduced by S. Flitsch who developed a method for the regioselective sulfation of glycosides [20]. The treatment of a deprotected galactoside with dibutyltin oxide allows the almost exclusive formation of the five membered cyclic dibutylstannylene acetal among the C-3 and C-4 cis hydroxyl groups. In this complex, the nucleophilicity of the equatorial 3-OH group is enhanced towards acylation, alkylation and silylation. In the case of galactocerebroside (Fig. 2), treatment of the dibutylstannylene acetal with the reactive sulfation reagent Me₃N·SO₃ gives almost exclusively the 3-O-sulfated derivative, being the 3,6-O-disulfate the only trace side-product and with no sulfation of the ceramide allylic hydroxyl group.

2.2 The glycosylation reaction

One of the most relevant aspects in glycoconjugate synthesis is the stereocontrol in the glycosylation. In the case of sulfatide, the formation of the 1,2-trans glycosidic bond can be easily controlled and mostly achieved taking advantage of neighboring group assistance by ester protecting group at C-2 of galactose, such as O-acetyl or O-benzoyl. The strategic choice relies mainly in the type of acceptor used in the glycosylation, either azidosphingosine, sphingosine or ceramide and in the selection of the proper galactosyl donor.

The majority of sufatide syntheses have been performed by using 3-O-benzoyl azidosphingosine 1 as the sphingoid acceptor. The “azidosphingosine glycosylation procedure” is generally preferred over glycosylation with ceramide because it provides higher yields. This is mainly due to the non-nucleophilic character of the azido group, which hence does not compete with the primary hydroxyl group towards glycosylation and to the higher nucleophilicity of the hydroxyl group of 1, which is no more involved in intramolecular hydrogen bond formation [18a]. Furthermore, azidosphingosine is an excellent precursor of ceramide, since reduction of the azido function to an amino group is usually easy and the subsequent N-acylation is in general high yielding. The azido group is commonly reduced by Staudinger reaction (usually Ph₃P or Bu₃P in THF/H₂O) or by treating the reaction mixture with hydrogen sulfide in pyridine.

After the seminal synthesis of Schmidt [21], many other preparation of natural (2S, 3R, 4E)-3-O-benzoyl azidosphingosine have been published [22]. These syntheses are all based on the chiral pool approach, which ensures the proper stereochemistry of at least one stereocenter, by starting from naturally occurring chiral substrates (Fig. 3).

Scheme 1 gives an overview of the experimental conditions exploited for the glycosidation of 3-O-benzoyl azidosphingosine 1 in the context of sulfatide synthesis. In one of the first examples, tetra-O-acetyl galactosyl bromide 2 was reacted with acceptor 1 under Koenigs-Knorr conditions. The reaction, catalyzed by mercuric cyanide, gave the β-galactoside 3, together with some α-anomer, in 76% and 9.5% yield, respectively [23]. However, glycosyl trichloroacetimidates are the most popular glycosyl donors for glycosidation of azidosphingosine. Thus, the peracetylated galactosyl trichloroacetimidate 4 was reacted with azidosphingosine 1 using BF₃·Et₂O as the promoter to give the β-glycosylated product 3 in 56%. Yields were slightly increased to 62% when the reaction was promoted with TMSOTf [24]. A significant increase of the yield was also observed for the acetylated donor 4 by promoting the reaction with tin(IV) chloride. The use of this catalyst, unconventional in trichloroacetimidate chemistry, allowed to recover 3 in 76% with only traces of the 1-O-acetylated azidosphingosine byproduct 1a [23].

In carbohydrate chemistry, protecting groups play an important role both in terms of stereoselectivity of glycoside bond formation (α or β) and of the yield of such reactions. In this regard, the pivaloyl group has been found to be a very useful protecting group especially at C-2 to avoid orthoester formation and yet to give β-glycosylation. In the case of glycosylation of sphingoid bases a lower migration of the C-2 acyl protecting group to give 1a has also been observed, thus resulting in a more high yielding glycosylation. In fact, when the reaction was carried out on the pivaloyl derivative 5, the β-glycosylated 6 was obtained in excellent 88% yield [22e].

Similar procedures were then followed for the conversion of compounds 3 and 6 to final sulfatides. Thus, after deprotection of the sugar moiety and reduction of the azido group, the fatty acid chains were introduced and the resulting compounds were selectively sulfated at O-3 of the sugar moiety by using the stannylene methodology to afford the desired sulfatides. The non-physiological sulfatide 7, with a non-natural C-17 fatty acid chain, obtained starting from 3, has been prepared as an internal standard for the quantitative determination of urine sulfatides by mass spectrometry. Large quantities of sulfatides can be detected in the urine of patients with metachromatic leukodystrophy, so the search of sulfatiduria and the quantification of sulfatides appears an useful test for the follow-up of new therapies [24]. The family of sulfatides 8–11 differing in the acyl chain length has instead been obtained from 6. In particular, the sulfatides containing a palmitic, a stearic, a behenic or a nervonic fatty acid, all constituents of the natural mixture, have been prepared and subjected to immunological evaluation which showed that the activation of sulfatide-specific and CD1a-restricted T-cell clones is influenced by the acid chain length [22e].

Palmitic-containing sulfatide 8 could also be prepared from 3 with a different strategy [23], which does not exploit the stannylene methodology (Scheme 2). In this case, the acetates were removed from galactose and position 2 and 6 were benzoylated after introduction of the 3,4-O-isopropylidene acetal to give 12. The azido group was then reduced and N-acylated and the isopropylidene deprotected to compound 13. Introduction of the acetate in the axial position 4 was then followed by sulfation of the free 3-hydroxyl group. Final basic hydrolysis of the protecting groups gave sulfatide 8.This is a typical example of a standard carbohydrate protection-deprotection strategy, which underlines the great synthetic advantage introduced by the methodology of Flitch [20]. It is worth noting that the preparation of compound 8 was done in a wider context together with the synthesis of other sulfatide related compounds. The purpose was to evaluate their activity as ligands for selectine receptors. In fact, native sulfatide significantly blocks activity in selectine-dependent inflammatory response. The study showed a binding selectivity for the different position isomers of sulfatide.

Again, in the context of studies as selectine ligand, sulfatide was also obtained by glycosylation of ceramide [25] (Scheme 3). The galactosyl trichloroacetimidate donor 15 was prepared with a levulinoyl protecting group at position 3, orthogonal to the benzoates. Compound 15 was coupled with ceramide 16 in the presence of molecular sieves AW-300 and BF₃·Et₂O. β-glycoside was obtained in 74% yield and the levulinoyl group was selectively deprotected to give 17. Sulfatide 9 was recovered after sulfation and ester hydrolysis.

More recently, a different glycosylation approach was used for the synthesis of the same sulfatide [26]. It is based on the glycosylation of sphingosine 18, where the amino group is doubly protected as tert-butoxycarbonyl and as N,O-isopropylidene with the secondary hydroxyl group (Scheme 4). The choice of this protecting group has been done in order to avoid the formation of the hydrogen bond between the electron pair of C-1 oxygen and the amide hydrogen of sphingosine, which in general reduces the reactivity of the sphingoid base towards glycosylation. Sphingosine 18 was glycosylated with peracetylated galactosylbromide 19, under the catalysis of Hg(II) salts. Galactosylsphingosine 20 was recovered in 62% yield. Higher yield (77%) and a cleaner reaction was obtained from the benzoylated donor 21. Starting from both 20 and 22 hydrolysis of the sphingosine protecting groups followed by N-acylation with stearic acid gave the deprotected galactocerebroside which was selectively sulfated to sulfatide 9.