Citation

Ning Wang, Sheng-Tao Li, Tian-Tian Lu, Hideki Nakanishi, Xiao-Dong Gao. Approaches towards the core pentasaccharide in N-linked glycans[J]. Chin. Chem. Lett., 2018, 29(1): 35-39  

Ning Wang, Sheng-Tao Li, Tian-Tian Lu, Hideki Nakanishi, Xiao-Dong Gao    

Containing glycans, glycoconjugates such as polysaccharides, glycoproteins, and glycolipids, are considered as important biological molecular species. Asparagine (N)-linked glycoproteins are intra- and extra-cellularly distributing proteins that play important roles in numerous biological processes, such as cell-cell interaction, infection, and immune response [1-8]. As shown in Fig. 1A, N-glycan assembly starts with the stepwise modification of the dolichyl-linked oligosaccharide (DLO) precursors on the endoplasmic reticulum (ER) membrane, thereby producing tetra-decasaccharides (three glucose (Glc), nine mannose (Man), and two N-acetylglucosamine (GlcNAc) residues) containing DLO analogs (i.e., Glc3Man9GlcNAc2-PP-Dol), which can be found in nearly all eukaryotic cells [9]. The introduction of tetradecasaccharides Glc3Man9GlcNAc2 (G3M9) to nascent proteins is initiated by oligosaccharyltransferase (OST) in a co-translational manner. The glycan is transferred to asparagine (Asn) residues embedded in the Asn-X-Thr/Ser (Ser, serine; Thr, threonine; X, any amino acid except proline) triad of the newly formed polypeptides [10, 11]. Subsequent sequential digestion by glucosidase Ⅰ (G-Ⅰ) and glucosidase Ⅱ (G-Ⅱ) removes the outer two glucose residues, thus yielding glycopolypeptides with monoglucosylated high-man-nose-type Glc1Man9GlcNAc2 (G1M9) glycan, which is the key glycoform related to glycoprotein folding [12-16]. After cleaving the last terminal glucose by G-Ⅱ and one mannose residue by ER mannosidase Ⅰ (ER Man-Ⅰ), trimmed glycoprotein is translocated to the Golgi apparatus to undergo processing and elaboration procedures (Fig. 1A). Various monosaccharides are added by glycosyltransferases, namely, N-acetylglucosaminyltransferase Ⅰ and Ⅱ (GnT-Ⅰ and GnT-Ⅱ) transfer GlcNAc, galactosyltransferase (GalT) can add galactose (Gal) and sialytransferase (SialT) transfers the sialic acid (Neu5Ac) moieties (Fig. 1B). Combined with Golgi mannosidases (Golgi Man-Ⅰ and Man-Ⅱ) which remove mannose residues to generate suitable substrates for glycosyltransferases, the N-glycan processing procedures lead to the formation of diverse glycan structures, which typically determine the destination of the mature glycoproteins [9].

Figure 1

Download: JPG larger image

Fig. 1. (A) Schematic of the biosynthesis of DLO precursors and folding of nascent N-linked glycoproteins in the ER. The tetradecasaccharide G3M9 is transferred to Asn residue by oligosaccharyltransferase (OST). G-Ⅰ: glucosidase Ⅰ, G-Ⅱ: glucosidase Ⅱ, ER Man-Ⅰ: ER mannosidases Ⅰ. (B) Schematic of glycans processing in the Golgi apparatus. GnT-Ⅰ: N-acetylglucosaminyltransferase Ⅰ, GnT-Ⅱ: N-acetylglucosaminyltransferase Ⅱ, GalT: galactosyltransferase, SialT: sialytransferase, Golgi Man-Ⅰ: Golgi mannosidases Ⅰ, Man-Ⅱ: mannosidase Ⅱ. (C) Three major classes of N-glycans which share the core structure pentasaccharide M3.

In general, glycoforms affect N-linked glycoprotein biological functions, including protein folding, stability, and intercellular traffic [8, 15, 17]. N-Linked glycans are classified into the following types according to their structures: High-mannose, hybrid, and complex (Fig. 1C). Aberrant glycan structures and functions (in DLOs and glycoproteins) are correlated with various diseases, such as cancer and congenital disorders of glycosylation, and thus attract the interest of medical researchers [12, 18-22]. All the N-glycan and DLO classes contain a common carbohydrate region specifically known as the pentasaccharide O-α-_D-mannopyranosyl-(l→3)-[O-α-_D-mannopyranosyl-(l→6)]-O-β-_D-mannopyranosyl-(1→4)-O-(2-acetamido-2-deoxy-β-_D-glucopyranosyl)-(l→4)-2-acetamide-2-deoxy-_D-glucopyranose (Man3GlcNAc2, M3, Fig. 1C). M3 is the structurally defined substrate of various glycosyltransferases and thus highly desired for the elucidation of N-glycan elaboration pathways. However, as the key intermediate of in vivo synthetic pathways, M3 is difficult to be obtained from naturally occurring resources and researchers are often compelled to use chemical methods for its preparation. In the past decades, the chemical synthesis of M3 was studied to overcome its synthetic challenges, especially the β-mannoside linkage.

This account summarizes typical chemical synthetic methods in solution and on solid phase, as well as recent efforts for the chemoenzymatic assembling of M3, which provide alternative approaches for obtaining the core saccharide. This study does not cover all approaches to obtain M3, nevertheless it is applicable for providing brief understanding of the advances in N-glycan construction.

M3 contains the crucial Man β(1-4) GlcNAc linkage, which mainly poses a considerable challenge to N-glycan chemical assembly procedures. Since the early 1980s [23], numerous efforts have been focused on the accomplishment and modification of strategies for M3 synthesis, leading to various methodologies to solve the problem of β-selectivity in mannopyranoside synthesis [24-28]. Crich and co-workers investigated a protocol for the direct coupling of aglycons to simple mannopyranosyl donors enriching β-mannopyranosides for primary glycosyl acceptors [29, 30]. This protocol has been exploited in numerous subsequent studies to establish N-glycans [31]. Meanwhile, intramolecular aglycon delivery (IAD) garners considerable attention and is a proven efficient technique to obtain β-mannoside. For instance, Ogawa and co-workers [32, 33] reported β-selective glycosylation mediated by a 4-methoxybenzylidene linker (Fig. 2A). However, although their approach resulted in the successful construction of core pentasaccharide M3 in N-glycan, it has not been widely applied. More recently, Fairbanks and co-workers [34, 35] investigated a novel IAD strategy mediated by a propargyl group, which was developed from the allyl IAD approach reported by the same group [36, 37]. In this strategy, the isomerization of the 2-O-progargyl in mannoside to an allene is followed by acetal formation with the 4-OH of the GlcNAc moiety in the acceptor. The formation of acetal leaded to the complete stereocontrolled glycosylation for pentasaccharide M3 generation (Fig. 2B).

Figure 2

	Download: JPG larger image
Fig. 2. (A) A 4-methoxybenzylidene group mediated IAD glycosylation process to form the β-mannoside [32]. (B) A propargyl group mediated IAD glycosylation process to form the β-mannoside [35].

Along with the solution phase synthesis, solid-phase oligosaccharide synthesis (SPOS) technique was comprehensively investigated in the past decades to access usable amount of N-glycans [38-40]. Incidentally, SPOS was not widely explored until several powerful solution-phase glycosylation methods were available. In general, acceptors are bound to the solid phase, allowing the excessive addition of donors to increase the yields of desired glycosidations. Notable works regarding this feature were reported by Seeberger and co-workers [41] and Schmidt and co-workers [42], who independently reported the solid-phase synthesis of core pentasaccharide M3. In 2003, Seeberger and co-workers described the first automated solid-phase synthesis of M3 backbones via stepwise assembling of mono-(2 and 4) and disaccharide (3) building blocks (Fig. 3) [41]. They used the Crich method to prepare β-mannosidic linkages via direct β-mannosylation and successfully obtained disaccharides. Several years later, Schmidt and co-workers successfully constructed a small library of 17 Nglycan structures comprising pentasaccharide core M3 structures [42]. Their SPOS strategy was based on a hydroxymethylbenzyl benzoate spacer linker, which is connected to the Merrifield resin as solid support. In their study, glycosylation was performed with the O-glycosyl trichloroacetimidates of glucosamine and mannose to extend the glycan chain. Different cleavable esters (such as benzoate, 9-fluorenylmethyloxycarbonate, and phenoxyacetate) were used for temporary protection of building blocks and the linker. By changing the donor imidates to N-phenyltrifluoroacetimidate, Fukase and co-workers [43] achieved highly stereoselective and efficient glycosylation on JandaJel resin to produce M3, which was successfully used for the synthesis of a complextype N-glycan.

Figure 3

	Download: JPG larger image
Fig. 3. Solid phase synthesis of pentasaccharide M3 via stepwise assembly on octenediol functionalized Merrifields resin (1) [41].

The chemical synthesis of M3 consists of numerous synthetic steps with complicated processes and stereo-selective glycosylation and thus offer poor economic efficiency. As an alternative option for pentasaccharide M3 preparation, semisynthesis via enzymatic trimming from homogeneous N-glycan resources was investigated. Homogeneous N-glycan resources mostly refer to sialylglycopeptide (SGP) obtained from egg yolks. This approach was first reported by Yamamoto and co-workers in 1997 [44]. In the widely applied chemoenzymatic synthesis of homogeneous glycoproteins [45-49], SGP is sometimes truncated to M3-containing substrate by removing nonreducing end saccharides with specific glycosidases (Fig. 4). Typically, the two terminal sialic acids in a complex-type glycan are cleaved by neuraminidase. The cleavage is followed by the removal of the two exposed galactose moieties by β-1, 4 galactosidase (β-Gal) and subsequently reachable GlcNAc moieties by N-acetyl-glucosaminidase (NAG). This strategy was performed by Shirai and co-workers [50], who used a chemoenzymatic approach to engineer monoclonal antibodies with homogeneous N-glycans, including M3. Except when SGP is digested, M3 is prepared in the N-glycan library originating from high-mannose type glycans, although the construction is usually complicated. In this instance, the high-mannose-type N-glycan precursor must be precisely designed to enable a comprehensively trimmable structure. Based on their previous work [51], Ito and co-workers reported the chemical synthesis of a defined orthogonal masked tridecasaccharide in 2015 [52, 53]. The nonreducing ends of the synthesized tridecasaccaride were systematically trimmed by glycosidases to yield 37 high-mannose-type glycans, including M3.

Figure 4

	Download: JPG larger image
Fig. 4. Typical gradual digestion procedures to prepare sialylglycopeptide (SGP) from egg yolk. β-Gal: β-1, 4 galactosidase, NAG: N-acetyl-glucosaminidase.

Until recently, the bottom-up construction of sufficient amount of M3 was only achieved through chemical synthesis, whereas chemoenzymatic strategies were always top-down. It was once concluded that generating a large amount of M3 through the in vitro reconstitution of the biosynthetic pathway of N-linked glycans is unrealistic. In the biosynthetic pathway, M3 containing DLO substrate Man3GlcNAc2-PP-Dol is stepwise generated by mannosyltransferase Alg1 and Alg2 on the cytoplasmic side of the ER (Fig. 1A). The two Alg proteins sequentially transfer three mannose residues from GDP-mannose to an N-acetyl-chitobiose linked to a long polyprenyl tail (GlcNAc2-PP-Dol). Alg1 then catalyzes the addition of the first mannose to produce trisaccharide Man1GlcNAc2 (M1), and Alg2 transfers two mannoses to M1 with α-1, 3 and α-1, 6 linkages to produce M3 (Fig. 1A) [54]. Although this biosynthesis pathway was confirmed in vivo and reproduced by several groups in vitro [55], it is not applicable for large-scale M3 production, largely due to the lack of practical methods for Alg1 and Alg2 purification.

However, several breakthroughs in the expression and purification of Alg1 and Alg2 were achieved in 2017. Gao and co-workers [56] first reported the successful expression of truncated Alg1 (without transmembrane domain, 35-449 aa) from Saccharomyces cerevisiae in Escherichia coli. The truncated Alg1 showed reasonable mannosyltransferase activity after purification. Alg1 is the essential β-1, 4 mannosyltransferase, which adds the first mannose onto GlcNAc2 by using GDP-mannose, thus the achievement of producible Alg1 leads to the formation of β-mannoside and address the most difficult issue in M3 generation. Soon after, Locher and co-workers reported a chemoenzymatic method for the effective synthesis of M3 and Man5GlcNAc2, another important N-glycan intermediate [57]. In their work, except for cytosolic Alg1 (S. cerevisiae, 33-349 aa), human Alg2 (full length) was expressed in human embryonic kidney (HEK293) cells and purified (Fig. 5A), as well as the truncated Alg11 (S. cerevisiae, 46-548 aa). They assembled pentasaccharide M3 by using β-1, 4 mannosyltransferase Alg1 and bifunctional enzyme Alg2 (Fig. 5B). They also revealed that the length of lipid tail in a DLO analog has no critical role in glycosylation reactions catalyzed by Alg1.

Figure 5

	Download: JPG larger image
Fig. 5. (A) Schematic structures of recombinant Alg proteins. The name, origin and length of each protein are indicated [57]. (B) In vitro chemoenzymatic synthesis of M3 which mimics the DLO biological pathway, by using purified Alg proteins [57].

At present, a sufficient amount of M3 can be stepwise assembled from chemically synthesized DLO analogs (lipidpyrophosphoryl-α-N, N'-diacetylchitobioside) with purified Alg1 and Alg2 proteins. The approach used simulates the assembly in in vivo pathway. This bottom-up chemoenzymatic synthesis serves as an alternative access to the core structure M3 in various glycoconjugates, providing useful substrates for mechanistic and structural analyses.

Undoubtedly, the core pentasaccharides M3 is the key molecule in both biological and synthetic pathways. Formation of β-mannoside, which is the major problem in the chemical synthesis of M3, has been comprehensively studied and well advanced in the past decades. Except for the works summarized herein, there are many other excellent studies on β-mannoside construction that we could not cite due to the lack of space [58]. The chemical strategy is not only the powerful method to access the sufficient amount of M3, but also leaves the possibility for further elongation and modification of the glycan via designed protecting groups. Although chemical synthesis of M3 has gained great success, researchers always attempt to develop the in vitro strategy using enzymes to mimic the biological pathway, which would inevitably be highly effective and complete selective. Finally, the goal was achieved that M3 could be constructed by chemoenzymatic synthesis catalyzed by purified Alg1 and Alg2 proteins. The significant progress will open the new prospects in the N-glycan assembly. Nevertheless, Alg2 protein is currently obtained as its full length form (with transmembrane domain) and expressed in human cell lines, which should be improved to become easily operable.

With M3 in hand, evaluation of the biological activity and substrate specificity of various enzymes could be performed, leading to a better understanding of the details and mechanism in N-glycan biosynthesis. Consisting with the reported study [57], the breakthrough in enzymatic synthesis towards M3 will also inspire exploiting novel methods for N-glycan constructions. It is evident from the on-going work that the optional assembling of N-glycans with glycosyltransferase will be realizable.

This work was supported by the National Natural Science Foundation of China (No. 21576118), Fundamental Research Funds for the Central Universities (Nos. JUSRP51629B, JUSRP11727) and Open Project Program of Key Laboratory of Carbohydrate Chemistry and Biotechnology (No. KLCCB-KF201604) and Program of Introducing Talents of Discipline to Universities (No. 111-2-06).