Chemical synthesis of homogeneous glycopeptides and glycoproteins

Authors

  • Yasuhiro Kajihara,

    Corresponding author
    1. International Graduate School of Arts and Sciences, Yokohama City University, 22-2 Seto; Kanazawaku, Yokohama, 236-0027 Japan
    2. Department of Chemistry, Faculty of Science, Osaka University, 1-1, Machikaneyama, Toyonaka, 560-0043 Japan
    • Telephone: +81-6-6850-5380. Fax: +81-6-6850-5382
    Search for more papers by this author
  • Naoki Yamamoto,

    1. International Graduate School of Arts and Sciences, Yokohama City University, 22-2 Seto; Kanazawaku, Yokohama, 236-0027 Japan
    Search for more papers by this author
  • Ryo Okamoto,

    1. International Graduate School of Arts and Sciences, Yokohama City University, 22-2 Seto; Kanazawaku, Yokohama, 236-0027 Japan
    Search for more papers by this author
  • Kiriko Hirano,

    1. International Graduate School of Arts and Sciences, Yokohama City University, 22-2 Seto; Kanazawaku, Yokohama, 236-0027 Japan
    Search for more papers by this author
  • Takefumi Murase

    1. International Graduate School of Arts and Sciences, Yokohama City University, 22-2 Seto; Kanazawaku, Yokohama, 236-0027 Japan
    Search for more papers by this author

Abstract

Oligosaccharides linked to proteins are known to play important roles in several biological events. However, oligosaccharides are heterogeneous, which has hindered detailed elucidation of oligosaccharide functions. In order to solve this problem, glycoproteins having homogeneous oligosaccharides have long been required. For this purpose, an efficient preparative method of complex-type oligosaccharides has been investigated from a natural source and this method was found to afford over 24 kinds of diverse complex-type oligosaccharides by use of chemical methods and branch-specific sequential glycosidase digestion. The sufficient amount of homogeneous complex type oligosaccharides obtained enabled us to examine the synthesis of homogeneous glycopeptides as well as glycoproteins by use of solid phase glycopeptide synthetic method and native chemical ligation. This review describes recent progress related to the efficient method of oligosaccharide preparation and synthesis of glycoproteins including bioactive erythropoietin. © 2010 The Japan Chemical Journal Forum and Wiley Periodicals, Inc. Chem Rec 10: 80–100; 2010: Published online in Wiley InterScience (www.interscience.wiley.com) DOI 10.1002/tcr.200900024

Introduction

Posttranslational modification (PTM) is essential for protein activation. The modification of the biological effects of protein is classified into several types, such as phosphorylation for signal transduction, ubiquitination for proteolysis, attachment of fatty acids for membrane anchoring, glycosylation for extending protein half-life, targeting, and cell-cell interactions.1 These PTMs of polypeptide chains occur as co- and/or posttranslational modifications, and the complexity of these PTM patterns have been given much attention, because PTM seems to alter the physical and chemical properties of protein such as the folding, conformational properties, stability, and function of the proteins in living systems.2

Among the PTM, glycosylation is a major form of protein modification and 50% of human proteins are reported to be glycosylated.3 In spite of intensive investigation, the functions of those oligosaccharides have not yet been completely elucidated. In particular, oligosaccharides linked to proteins display many diverse structures, such as dibranching, tribranching, and tetrabranching patterns, and their terminal sugar moieties are frequently heterogeneous. In addition, the oligosaccharides in proteins exhibit site-specific diversity.4 Some of these oligosaccharides linked to proteins exhibit extensive heterogeneity, while others at other sites of the same protein exhibit a homogeneous pattern. The oligosaccharides in glycoproteins are roughly classified into two types, O-linked and N-linked. In terms of the O-linked type, a small oligosaccharide attaches to the alcohol of serine or threonine by an N-acetyl-α-D-galactosaminyl linkage. The sequences of the small oligosaccharides are classified into 8 groups.5a In contrast, the N-linked type is a large and branched oligosaccharide, which is further divided into three types, i.e.: complex, hybrid, and high-mannose types (Figure 1).5 All oligosaccharides are linked to the nitrogen of asparagine by an N-glycosyl linkage. The biosynthesis of these N-linked oligosaccharides has been extensively studied. During the generation of the polypeptide chain in the endoplasmic reticulum (ER) based on the mRNA codon, highmannose-type oligosaccharide is incorporated into the asparagine residues in the NXS sequence as a co-translational modification. Recently, research on the quality control of glycoproteins in the ER has demonstrated that a chaperone, calnexin/calreticulin, monitors the folding process of proteins by recognizing high mannose-type oligosaccharides on its protein surface.6 Unfolded glycoproteins are then transported into the cytoplasm for digestion,7 while folded glycoproteins are transported into the Golgi apparatus. In the Golgi apparatus, highmannose-type oligosaccharides are converted into complex-type oligosaccharides or hybrid-type oligosaccharides by glycosidases and glycosyltransferases. Imperfection of the complex-type oligosaccharide synthesis in this process appears to be the cause of various diseases.8–10 However, because of the heterogeneity of complex-type oligosaccharides, it is still difficult to elucidate specifically how the complex-type oligosaccharide structure is involved in the trafficking, secretion, and bioactivities of glycoproteins in detail.1,11 In order to reveal the function of oligosaccharides, preparation of glycoproteins having homogeneous oligosaccharides would be needed, but a satisfactory method for doing this has not been established.

Figure 1.

Sructure of N-linked oligosaccharides.

Recently, glycoprotein expression systems for the purpose of investigation of oligosaccharide functions have advanced,12 but as yet the oligosaccharide patterns cannot be regulated precisely. Under these circumstances, a chemical synthetic method to precisely prepare the desired oligosaccharide patterns on proteins has been anticipated, as this would allow the study of oligosaccharide function.13

This review will describe recent advances in the chemical synthesis of glycopeptides as well as glycoproteins, and future aspects for understanding PTM through chemical approaches.

Preparation of Diverse Complex-Type Oligosaccharides

In order to synthesize glycopeptides and glycoproteins having complex type oligosaccharides, an appropriate amount of particular oligosaccharides is essential. The complex type oligosaccharides have been synthesized by chemical methods and many commendable synthetic methodologies have been reported. During these investigations, the formation of β-mannosyl and α-sialyl linkages, which are difficult to accomplish, is needed in order to obtain appropriate amounts of complex type oligosaccharides. Through such syntheses, multi antennary complex-type14–16 oligosaccharides 1 and high-mannose17 type oligosaccharides 3 have been synthesized (Figure 1), but the procedures are very time-consuming and require particularly involved techniques for handling the oligosaccharides.

In order to prepare diverse oligosaccharides and to solve the above mentioned difficulties, an efficient preparation method of diverse oligosaccharides has been examined by a strategy of isolation from a natural source.18 The homogeneous N-linked complex-type oligosaccharides 1 were prepared by peptidase digestion of oligosaccharyl peptide isolated from egg yolk (ca 30–40 mg/ egg),18 and then branch-specific stepwise enzymatic removal of sugar residues from a NeuAc-α-2,6-Gal-β-1,4-GlcNAc-β-1,2-Man-α structure was performed to obtain diverse oligosaccharides. For this strategy, selective acid hydrolysis of NeuAc is preferable for isolating each of the homogeneous monosialyloligosaccharides (such as 5 and 6, Figure 2). Isolation of these two products enables us to easily obtain several homogeneous oligosaccharides by subsequent asialobranch-specific exo-glycosidase digestion steps (N-acetyl-β-D-glucosaminidase, and α-D-mannosidase).

Figure 2.

Acid hydrolysis reaction of NeuAc from complex type oligosaccharide 1.

However, this simple strategy shown in Figure 2 has a drawback. Acid hydrolysis of NeuAc is not selective and affords various products. Although each oligosaccharide generated needs to be isolated for sequential glycosidase digestion, it was found to be difficult to purify by HPLC on a synthetic scale (ca. 50 mg scale). Therefore, oligosaccharides were modified with the hydrophobic protecting groups to increase the degree of interaction between the oligosaccharides and a reverse phase HPLC column in order to isolate these products. The asparagine residue of each oligosaccharides was protected by an Fmoc (9-fluorenylmethyl group) group and this protection enabled isolation of the oligosaccharides: the disialyloligosaccharide (4), two kinds of monosialyoligosaccharides (5 and 6), and the asialooligosaccharide (the over-reaction product, 7) (Figure 2). Although monosialyloligosaccharides (5 and 6) were still isolated as a mixture, subsequent galactosidase treatment enabled the isolation of each of the monosialyloligosaccharides 8 and 14 (step 3 in Scheme 1). Successful purification afforded the monosialyloligosaccharides 8 and 14, and subsequent branch-specific exoglycosidase digestion afforded many diverse oligosaccharides, as shown in Scheme 1. In terms of isolation of monosialyoligosaccharides 5 and 6 (Figure. 2), additional protection by a hydrophobic group was examined. Throughout extensive studies, selective benzyl esterification18 on the carboxylic acid of sialic acid was achieved, when the Cs2CO3 and BnBr conditions were used. This protection by a hydrophobic protecting group enabled separation of the monosialyloligosaccharides (5 and 6) from each other. According to this strategy, 24 kinds of diverse oligosaccharides were prepared.18 Selection of oligosaccharides from this small library enabled the tailor-made synthesis of glycopeptides and glycoproteins having the desired homogeneous oligosaccharide structure.

Scheme 1.

Preparation of diverse asparagin linked oligosaccharides by exo-glycosidase digestion.

The structure of these oligosaccharides were confirmed by comparison of reported NMR data18 and advanced NMR analysis method.19

Synthesis of Glycopeptides

Peptide chemistry has been developed by the use of solid-phase synthetic methods and a number of examples of commendable technologies are summarized in various different handbooks.20 Peptide synthetic methods are divided into two methods, the Boc method and the Fmoc method.20 Of these two methods, the Fmoc strategy has been used as the conventional method for glycopeptides, because the oligosaccharide chain is labile toward a strong acid such as HF, which have been used for detaching the peptide from the solid support. In terms of the formation of glycopeptides, N-(9-fluorenylmethoxycarbonyl)-N-(oligosaccharyl)-l-asparagine (Fmoc-Asn-(oligosaccharyl)-OH) was used for the solid-phase glycopeptide synthesis and the hydroxy groups of the oligosaccharide are generally protected by acetyl groups in order to avoid esterification of the sugar hydroxy groups during elongation of the peptide chain by the activated Fmoc-amino acids. In the case of N-linked large oligosaccharide 4, which can be prepared through isolation of oligosaccharyl peptide 20, peptidase digestion, and protection of asparagine,18 the protection of over 40 hydroxy groups is required. However, deprotection of a number of protecting groups such as acetyl groups from the synthesized glycopeptide may give rise to concerns regarding epimerization at the α-position of certain amino acids, owing to the treatment under basic conditions. Therefore, we examined the use of Fmoc-Asn-(oligosaccharyl)-OH 4 having unprotected hydroxy groups in the solid-phase glycopeptide synthesis.18 However, Fmoc-Asn-(sialyloligosaccharyl)-OH 4 has three carboxylic acids. In order for this Fmoc-Asn-(sialyloligosaccharyl)-OH 4 to form a peptide bond, selective esterification conditions of sialic acid were essential. In terms of this problem, as has been mentioned, Cs2CO3 and BnBr condition was fortunately found to afford sialyloligosaccharide 21 (Scheme 2).18 In addition, it was found that the esterified Fmoc-Asn-(sialyloligosaccharide)-OH 21 exhibited enough stability against hydrolysis of the sialoside linkage when exposed to 95% TFA for 3 h. The conditions are conventionally used for detaching peptide from the solid phase.20 It is known that a sialyl linkage to oligosaccharide, sialoside, is labile to acid treatment rather than other glycosyl bonds. According to these results, the effect of this ester was hypothesized to be that the proton of the carboxylic acid interacts with the sialoside oxygen as a 5-membered ring, while the esterified form of sialoside cannot afford protonation to the sialoside oxygen. This protonation from the carboxylic acid may work well as an acid catalyst.21 The stability of Fmoc-Asn-sialyloligosaccharide 21 allowed us to realize the solid phase synthesis of sialylglycopeptide 22 corresponding to a fragment of erythropoietin (Scheme 3).22

Scheme 2.

Reagents: a) Actinase-E, Tris-HCl buffer, NaN3, pH 7.5 (86%); b) 9-fluorenylmethyl N-succinimidyl carbonate (Fmoc-OSu), NaHCO3, Acetone, H2O (68%); c) Cs2CO3, H2O then BnBr, DMF (85%).

Scheme 3.

Solid-phase synthesis of sialylglycopeptide 22.

Throughout this glycopeptide synthesis, sugar hydroxy groups remained free. The solid-phase synthesis of the sialylglycopeptide 22 was performed on the poly(ethylene glycol)-poly(dimethylacrylamide) copolymer (PEGA)23 resin having an acid-labile linker, hydroxymethylphenoxyacetic acid (HMPA). After incorporation of the dipeptide onto the resin, Fmoc-Asn-(sialyloligosaccharyl)-OH 21 was incorporated by 2-(1H-9-azobenzotriazole-1-yl)-1,1,3,3-tetramethyluronium hexafluorophosphonate (HATU). Subsequent peptide elongation was performed by use of the conventional diisopropylcarbodiimide (DIPCDI) and 1-hydroxy-1H-benzotriazole (HOBt) condition. During this process, a low concentration of activated Fmoc-amino acid was used to avoid the esterification of sugar hydroxy groups by the activated Fmoc-amino acid. After construction of the desired glycopeptide, 95% TFA treatment and a short period of saponification successfully afforded the desired sialylglycopeptide 22.22

However, during the activation of the α-carboxylic acid of Fmoc-Asn-(sialyloligosaccharide)-OH 21, the formation of the unexpected aspartimide 23 occurred as a byproduct. This side reaction resulted in low coupling yields of Fmoc-Asn-(sialyloligosaccharyl)-OH 21 with the peptide-resin.24 As the result of investigation toward solving this problem, 3-(diethoxy-phosphoryloxy)-1,2,3-benzotriazin-4(3H)-one (DEPBT), a coupling reagent, was found to be the most suitable to minimize aspartimide formation. Considering that the aspartimide derivative 23 is inevitably generated during the coupling reaction of 21 with the peptide-resin even by the use of this reagent, each 2 equivalents of both Fmoc-Asn-(sialyloligosaccharyl)-OH 21 and diisopropylethylamine (DIPEA), and 3 equivalents of DEPBT were employed toward the peptide on the resin (HMPA-PEGA: 1.0 mmol scale). In terms of the undesired esterification of the sugar hydroxy groups, a concentration as low as 40 mM was found to suppress esterification even with the most reactive Fmoc-Gly-OH, while peptide elongation occurred smoothly. By employing these conditions, the synthetic strategy, as exemplified in Scheme 3, successfully afforded several other glycopeptide fragments 24–26 in good yields (Figure 3).24 The typical feature of this synthetic strategy in solid-phase glycopeptide synthesis is to use an Fmoc-Asn-(sialyloligosaccharyl)-OH 21 without protection of its hydroxy groups. This strategy shortens the processes required for removal of the protecting groups.

Figure 3.

Several glycopeptides synthesized by solid phase procedure.

Chemical Synthesis of Glycoprotein

In order to synthesize glycoproteins, the synthetic chemistry of both the proteins and glycopeptides can be combined and used, in addition to the protein folding and protein purification methods. For the synthesis of proteins, Dawson and Kent have developed native chemical ligation (NCL, Figure 4).25 This peptide coupling reaction caused a revolution in the synthetic strategy of proteins and became an essential reaction for protein synthesis. NCL occurs spontaneously between αthioester at the C-terminal in peptide segment-1 27 and the cysteine residue at the N-terminal of another peptide, segment-2 28, to afford native amide linkage 30 through a thioester exchange reaction 29. This reaction can be applied to the synthesis of large glycopeptides as well as glycoproteins. Although several groups have examined the synthesis of glycoproteins,26 the glycoprotein synthesized thus far do not have mature human complex type sialyloligosaccharides. Therefore, complex type sialyloligosaccharide isolated form egg yolk was examined for the synthesis of glycoproteins having human complex-type sialyloligosaccharides. For this challenging task, monocyte chemotactic protein-3 (MCP-3) was selected as the first target. The peptide sequence of MCP-3 is shown in Figure 5A.

Figure 4.

Native chemical ligation.

Figure 5.

Synthesis of glycosylated MCP-3 37.

The N-terminus is a pyroglutamate (Pyr) region and residues 6–8 encompass an NTS consensus sequence for N-glycosylation. In addition, two disulfide bonds are formed between Cys11-Cys36 and Cys12-Cys52 (Figure 5A).27,28 For the synthesis of glycosylated MCP-3 37, the synthetic scheme employed a twice-repetitive NCL strategy using three unprotected polypeptides (Figure 5B). By using the NCL strategy at the Cys residues, MCP-3 can be assembled from three peptides, segment-1 (1–10 residues) 31 (synthesized by the Fmoc method), segment-2 (11–35 residues) 32, and segment-3 (36–76 residues) 33 (Figure 5B). This strategy places the glycosylated Asn in the peptide segment-1 (31), which must be synthesized as a C-terminal peptide-αthioester. In addition, the N-terminal cysteine residue in segment-2 (32) must be protected to prevent undesired polymerization or intramolecular cyclization by NCL. In this synthesis, we used the 1,3-thiazolidine-4-carboxyl (Thz) group.29 Since segment-2 (32) and segment-3 (33) contain no posttranslational modifications, they were synthesized by standard Boc-solid phase peptide synthesis (SPPS).30

In terms of segment-1 31, both the Boc strategy and Fmoc strategies successfully afforded the desired sialylglycopeptide-αbenzyl thioester. Here, the Fmoc strategy for the synthesis of segment-1 as a sialylglycopeptide-αbenzyl thioester 31 will be described (Scheme 4). In this synthesis, a highly acid sensitive linker, 4-(4-hydroxymethyl-3-methoxyphenoxy)-butyric acid (HMPB) resin,31 was used. As shown in Scheme 4, the first amino acid residue, Thr, was introduced to the HMPB-PEGA resin using 1-mesitylenesulfonyl-3-nitro-1,2,4-triazole (MSNT). The peptide elongation was carried out using a DIPCDI/HOBt coupling method, except for the oligosaccharyl asparagine. Fmoc-Asn-(sialyloligosaccharyl)-OH 21 was coupled (>95%) with the peptide-resin using DEPBT/DIPEA. After elongation of the peptide, the resin was treated with acetic acid (AcOH) and trifluoroethanol (TFE) solution to release the side chain-protected glycopeptide 38 from the resin. Without further purification, αthioesterification with α-toluenethiol (benzyl mercaptan:BnSH) was performed at the C-terminus of the protected glycopeptide 38 using an optimized condition, benzotriazol-1-yl-oxy-trispyrrolidinophosphonium hexafluorophosphate (PyBOP) and DIPEA at –20° for 2 h. The low temperature condition can prevent epimerization at the C-terminal amino acids. This protocol afforded the desired glycopeptides-αthioester in good yield. The crude αthioesterification 39 was treated with 95% TFA, 2.5% triisopropylsilane (TIPS), and 2.5% water to remove the protecting groups of the side chains. Finally, HPLC purification afforded segment 1 as a sialylglycopeptide-αbenzyl thioester 31 in 43% yield.32

Scheme 4.

Synthesis of glycopeptide-thioester by Fmoc strategy.

Segment 2 and segment 3 were synthesized by manual Boc-SPPS. The first NCL (Figure 5B), which occurred between segment-2 (32) and segment-3 (33) to afford segment-4 (34), was performed in 0.1 m phosphate buffer (pH 7.6) containing 6 m guanidine hydrochloride (Gn-HCl), 1% (vol/vol), BnSH, and 1% (vol/vol) thiophenol (PhSH) (data not shown) and this mixture was stirred for 24 h. This ligation proceeded cleanly and afforded the desired ligation product. Without further purification, the conversion of Thz to Cys at the N-terminus of the ligation product was performed using methoxylamine hydrochloride to afford segment-4 (34). A second NCL was carried out using segment-1 (31) and segment-4 (34) (Figure 5B). Segment-1 (31) and segment-4 (34) were dissolved in 0.1 m phosphate buffer (pH 7.6) containing 6 m Gn-HCl, 1% (vol/vol) BnSH, and 1% (vol/vol) PhSH (final pH ∼6.5) and this mixture was stirred for 34 h. The second NCL reaction afforded the desired product 35, which was characterized by mass spectrometry and HPLC (Figure 6A and B). Oxidative folding was undertaken by diluting the ligation mixture (Figure 6B) with five volumes of 0.1 m tris (hydroxymethyl) aminomethane hydrochloride (Tris-HCl) buffer (pH 8.0), and bubbling air for 1 min. The reaction mixture was then incubated for 24 h at room temperature. HPLC and ESI mass spectrometry analysis of the reaction mixture revealed that the main peak, which was the desired product 36 (Figure 6C), eluted at an earlier time than that of the original unfolded form 35 (Figure 6B). It is known that the folded form elutes earlier than that of the unfolded form in the case of a chemokine (non-glycosylated protein).33 This clean conversion, monitored by HPLC, suggested that the folding reaction occurred smoothly and afforded the folded form 36 even in the presence of a large and highly hydrophilic oligosaccharide on the peptide backbone. Mass analysis indicated that glycoprotein 36 adopted two disulfide bonds. After HPLC purification of the folded form 36 (Figure 6D), it was treated with 50 mm NaOH solution for 10 minutes to remove the benzyl groups from the NeuAc residues to afford glycoprotein 37, which was characterized by ESI mass spectrometry and HPLC (Figure 6E and 6F).

Figure 6.

Synthesis of glycosylated MCP-3 37. RP-HPLC profiles and observed mass of final product. A) Ligation reaction mixture after 0 min, B) ligation reaction mixture after 48 hours, C) crude sample after folding, D) purified product after folding, E) purified glycosylated MCP-3 after saponification of benzyl group of NeuAc. HPLC elution condition: column, Cadenza CD-18 (3 mm, 4.6 × 75 mm) at a flow rate of 1.0 ml min−1, linear gradient of 18%–54% CH3CN containing 0.09% TFA in 0.1% TFA aq. over 15 min. F) ESI mass spectra of glycosylated MCP-3 37. The Na+ and K+ adducted ion peaks were also observed.

In order to confirm the conformational properties of the synthetic glycosylated MCP-3 37, we investigated the disulfide bond connectivity and measured the CD spectrum. To better understand the disulfide bond pattern, we employed a standard strategy using chymotrypsin digestion and subsequent HPLC and mass analysis of the resultant peptide fragments. This analysis clearly indicated that the synthetic glycosylated MCP-3 has 2 disulfide bonds between Cys11/12 and Cys36/52. However, a minor product (ca. 7% purity based on HPLC analysis) was found to have an undesired disulfide bond between Cys36 and Cys52.34 In addition, the CD spectrum of the glycosylated MCP-3 37 was measured. Synthetic glycosylated MCP-3 exhibited the expected CD spectrum, which is consistent with the single alpha helix and poorly formed beta sheet characteristic of these chemokines.34 In addition, an ELISA assay using a sandwich system clearly indicated that synthetic glycosylated MCP-3 37 has the two correct epitope structures.34 Taken together, these analyses support the correct folding of the protein with greater than 90% purity. These studies demonstrate the successful synthesis of glycosylated MCP-3 37 having an intact complex-type N-linked sialyloligosaccharide.34

Expanding the Scope of Native Chemical Ligation

NCL using cysteine residues has been successfully used for the synthesis of both glycoproteins and proteins. However, occasionally the cysteine residue is not properly located or does not exist in the target proteins. In order to improve this potential difficulty, a reduction method, which changes the sulfuhydryl group of cysteines to hydrogen atoms after NCL and the utilization of an auxiliary group having a sulfuhydryl group has been developed.35 In addition to the improved NCL, a new strategy to use a serine site for the ligation position was found. Serine residues are frequently found in the peptide backbone as well as in the consensus sequence (NXS), where an asparagine residue is generally incorporated along with the N-linked oligosaccharides. In order to use the serine site for NCL, a new ligation concept employing conversion from a cysteine residue to a serine residue after NCL was developed. For such a technique, it was necessary to explore concise reaction sequences. As a result, the CNBr cleavage method was found to be applicable at the methyl cysteine site, which was obtained by specific cysteine methylation.36 The strategy is shown in Figure 7. After NCL between 40 and 41, the conversion of the cysteine to serine was performed by a series of reactions, the S-methylation of the cysteine 43 with methyl 4-nitrobenzene sulfonate, intramolecular rearrangement by CNBr activation in 80% HCOOH solution (Figure 7 (4448)), and an O to N acyl shift. Activation of the S-methyl group by CNBr resulted in intramolecular attack by the neighboring carbonyl oxygen on the β-carbon of the methyl cysteine residue and generated an O-ester peptide intermediate 48. This intermediate is converted into the desired peptide through spontaneous O to N-acyl shift under a slightly basic condition (pH 7–8).37a This improved ligation method was applied to several peptide sequences as well as an N-linked glycopeptide 54, which was a fragment of erythropoietin (79 to 98 residues, Scheme 5). NCL, between a glycosyl hexapeptide-αthioester 50 having complex type N-linked asialooligosaccharides prepared by a reported method32 and a tetradecapeptide 51 having a cysteine residue, was performed by the conventional method. The glycosyl icosapeptide 52 thus obtained was then subjected to by S-selective methylation 53, followed by activation of the S-methyl cysteine residue by CNBr/HCOOH condition. Although random formylation of the sugar hydroxy groups during the CNBr reaction, the slightly basic condition used for the O to N-acyl shift (<pH 10) resulted in the concomitant removal of the O-formyl groups to successfully afford desired icosapeptide 54. This treatment successfully afforded the target glycosyl icosapeptide 54 having serine residues.37a The structure and purity of this glycopeptide was confirmed by comparison with the authentic sample synthesized by the same solid phase glycopeptide synthetic conditions as shown in Schemes 3 and 4. In addition, this strategy was improved to avoid random formylation by use of TFA/MeCN solution instead of formic acid solution and was used for the synthesis of acid sensitive sialyl-Tn glycopeptide synthesis of 60, after NCL between 55 and 56 (Scheme 6).37b

Figure 7.

Expanding the scope of native chemical ligation by use of CNBr activation method.

Scheme 5.

Synthesis of glycopeptide by use of CNBr activated native chemical ligation.

Scheme 6.

Synthesis of glycopeptide having sialylTN epitopes by use of CNBr activated native chemical ligation. a) 6 M Guanidine hydrochloride, 0.1 m sodium phosphate, 60 mM 4-mercaptophenylacetic acid, 20 mM tris(2-carboxyethyl)phosphine buffer (pH 7.2); b) methyl 4-nitrobenzenesulfsonate, 6 M guanidine hydrochloride, 0.25 M Tris-HCl 3.3 mM EDTA-2Na buffer (pH 8.6), CH3CN; c) 1. CNBr, 80% HCOOH, 2. TFA, NH4I, Me2S, 3. 5% hydrazine hydrate solution.

Applications using the Native Chemical Ligation

As mentioned above, oligosaccharides exhibit diverse structures even at the same glycosylation site of glycoproteins. This results in various glycoprotein isoforms, even though the protein sequence and its 3D structure are identical. These isoforms are called “glycoforms”. In order to elucidate the oligosaccharide function dependent on its structure, the synthesis of homogeneous individual glycoforms is essential. Toward solving this problem, a combinatorial strategy using the NCL method was demonstrated to prepare glycoforms systematically. A simple combinatorial strategy using two kinds of oligosaccharides (61 and 62) and two peptide segments (63 and 66) of a target peptide can afford several kinds of glycoconjugates with varying oligosaccharide structures and glycosylation positions. In terms of the incorporation of oligosaccharide into the peptide, the haloacetamide method can be used. This method employs a convenient coupling reaction between the thiol of the cysteine and the haloacetamide group at the reducing end of the oligosaccharide.38 Therefore, a strategy for this systematic synthesis of a small glycoconjugate library was examined. The target glycoconjugate was a glycopeptide fragment in AILIM/ICOS (Activation inducible lymphocyte immunomediately molecule/Inducible co-stimulator) on T-cells.39 The AILIM/ICOS extracellular domain has two N-linked oligosaccharides close to the signal accepting domain at the 89Asn and 110Asn positions. In order to investigate the relationship between AILIM/ICOS activity and the oligosaccharide structures, the synthesis of glycoconjugates (83Cys-135Leu) 6972 was examined with varying types of oligosaccharide structures at the 89 and 110 positions. In order to use NCL, the target glycoconjugate was divided into two segments (83Cys-109Cys 63 and 110Cys-135Leu 66). In terms of the oligosaccharides, we selected the acidic and neutral forms, complex type disialyloligosaccharide 61 and its asialoform 62, respectively, because certain protein-oligosaccharide binding events seem to be regulated by the topology of acidic functional groups in the oligosaccharide.40 As shown in Scheme 7, systematic sequential coupling strategy afforded the desired glycoconjugate library.41 This strategy may be useful for the predetermination of the oligosaccharide structures, which would be suitable for the synthesis of homogeneous glycoprotein.

Scheme 7.

Synthesis of ICOS glycopeptide fragments having diverse oligosaccharides.

Synthesis of Erythropoietin Analogue

Erythropoietin (EPO) is known to be a useful drug for the treatment of renal anemia. In this section, the synthesis of an EPO analogue having two complex type sialyloligosaccharides is described (Scheme 8). EPO consists of 166-amino acid residues and has three N-linked complex type oligosaccharides attached to the Asn24, 38, and 83 and O-linked oligosaccharides at the serine 126 position.42 It is known that the three complex type sialyloligosaccharides are mainly tri- or tetra-antennary type structures, and these sialyloligosaccharides enable EPO to increase its half-life in blood43 owing to interference with glomerular filtration or interaction with galactose-binding lectins. For the efficient synthesis of a glycosylated EPO analogue, a synthetic strategy was designed to prepare an entire glycosylated polypeptide chain by NCL between a synthetic glycopeptide-αthioester having complex type sialyloligosaccharides 74 and the N-terminal cysteine of a large polypeptide chain 76 prepared by expression in E. coli.44 Fortunately, EPO has a cysteine residue at the 33 position and thus this position was used as a ligation site for NCL.44b Because E. coli can not afford mammalian-type glycosylation of proteins, incorporation of an oligosaccharide at the natural position Asn83 was not possible, and therefore additional sialyloligosaccharides were incorporated at alternative suitable positions. In this research, two glycosylation sites at the natural 24 position and unnatural 30 position were selected, and Cys33 was used for the ligation site. In order to synthesize the glycopeptide-αthioester 74 (32 amino acid residues) containing two complex type sialyloligosaccharides at positions 24 and 30, the haloacetamide method was used as mentioned in the synthesis of a glycopeptide library (Scheme 7). As shown in Scheme 8, a peptide comprising the N-terminal 32 residues of EPO (shown in red) was synthesized by SPPS and the subsequent haloacetamide reaction was examined. EPO contains two disulfide bonds, between Cys7 and Cys161, and between Cys29 and Cys33. Therefore, the thiol groups of Cys7 and Cys29 were protected by acetamidomethyl (Acm) groups. In terms of Fmoc SPPS for the preparation of the peptide thioester, HMPB-PEGA resin was used and peptide elongation was performed by using 2-(1H-benzotriazole-1-yl)-1,1,3,3-tetramethylaminium hexafluorophosphate (HBTU), HOBt, and DIPEA. After construction of the 32-residue peptide, the fully protected peptide was detached from the HMPB linker by means of AcOH-TFE. The C-terminal α-carboxylic acid of this side-chain protected peptide was converted to the peptide-αthioester using PyBOP, DIPEA, and excess EtSH at −20 °C, as described in the synthesis of MCP-3 (Scheme 4). Toward this peptide-αthioester 73, a haloacetamide reaction was then performed by using bromoacetamidyl undecasaccharide 61. This protocol smoothly afforded the desired glycopeptides-αthioester 74. In terms of the polypeptide chain 76 comprising residue 33 to 166, the peptide was expressed as His-tag fusion peptide 75 in which an inserted methionine represents the His-tag junction.44 After expression, purification using a His-tag column afforded a homogeneous fusion peptide 75. To remove the His-tag, CNBr treatment toward methionine afforded the homogeneous polypeptide chain 76 having a cysteine at the N-terminus. Then, in order to obtain the whole glycosylated polypeptide chain 77, NCL reaction between glycopeptide 74 and polypeptide segment 76 was performed in a solution of 6 m Gn-HCl and 4-(carboxymethyl)thiophenol (MPAA) as the catalyst.45 After 15 h, this NCL reaction afforded EPO-polypeptide chain 77 in moderate yield. After purification, the glycosylated polypeptide 77 was analyzed by mass spectrometric analysis, which clearly verified that the NCL reaction had afforded the glycosylated polypeptide chain77. Subsequently, removal of the Acm groups using silver acetate afforded EPO analogue 78. This deprotection step was monitored by ESI mass spectra. Within 4 h, substrate 77 was converted into the desired product 78. Purification using a reversed phase HPLC column afforded pure glycosylated peptide 78. Finally, a protein folding experiment was performed by a dialysis protocol46 in order for this glycosylated polypeptide to adopt its ideal three-dimensional structure. This protocol initially employed 3 m Gn-HCl to denature glycosylated polypeptide chain 78, followed by a three step dilution process of Gn-HCl concentration in the presence of a cysteine-cystine redox system.46 This folding experiment successfully afforded glycosylated EPO 79.47 After purifying the synthetic sample of EPO 79, its folded structure was evaluated by several analyzing methods in solution.

Scheme 8.

Synthesis of erythropoietin analogue.

As shown in Figure 8, electrospray ionization mass spectrum (Figure 8a), CD spectrum of EPO analogue 79 (Figure 8b), and HPLC analysis in addition to SDS-PAGE (Figure 8c), suggest that glycosylated EPO 79 has the desired three-dimensional structure and is also consistent with the helical secondary structure of a correctly folded EPO.48 To confirm the disulfide bond positions in the folded EPO analogue 79, peptidase digestion and subsequent mass analysis of the resultant peptide fragments were performed. These analyses clearly indicated that two disulfide bonds were correctly formed by the folding process. The biological assay of the folded EPO analogue 79 was also evaluated based on a cell proliferation assay using TF-1 cells.49 Assays using synthetic EPO 79 showed a successful induction of TF-1 cell proliferation from 50 pg/ mL and that the degree of proliferation was almost the same as that of native EPO in vitro. These experiments suggest that the current synthetic procedure successfully afforded an homogenous erythropoietin analogue having two human complex-type sialyloligosaccharides by the combined use of chemical synthesis and protein expression in E. coli.47

Figure 8.

Analytical data of Erythropoietin analogue 79. (a) ESI-mass data (calcd; 22977.2, found; 22977.7) (b) CD spectrum. (c) HPLC and SDS-PAGE, lane*: marker; and **: folded EPO6 (d) Estimation of cell (TF-1 cell) proliferation.

Conclusion

Although glycoproteins having mature complex type oligosaccharides have been thought to be preparable only from cell expression systems, chemical synthesis has gradually emerged as a powerful method for synthesizing glycoproteins having homogeneous oligosaccharides at the desired position of the protein backbone. This technique will alter the concept behind the preparation of glycoproteins and open an avenue for exploring new organic synthetic targets based on glycoproteins. In addition, this new approach will be expected to afford homogeneous glycoprotein therapeutic agents and to help in understanding the function of oligosaccharide profiles from a chemical point of view. The demonstration described here and recent progress by other groups50–52 show that the chemical synthesis of bioactive homogeneous glycoproteins is now within reach of organic chemists.

Acknowledgements

YK would like to thank Mr. Yasutaka Tanabe, and Mr. Yasuhiro Suzuki for their collaboration in preparing the Asn-oligosaccharide library and glycoprotein syntheses. Dr. Philip E. Dawson (Scripps Research Institute) is thanked for collaboration in the MCP-3 synthesis. In particular, Dr. Dawson designed the new synthetic method of the glycopeptide-αthioester by using the Boc strategy. Dr. Derek Macmillan (University College of London), Dr. Katsunari Tezuka (Otsuka Chemical CO) and Prof. Dr. Takashi Tsuji (Tokyo university of Science) are thanked for collaboration in the EPO analogue synthesis. Dr. Macmillan kindly provided plasmid of EPO segment and discussed synthetic strategy. The financial support from the Japan Society for the Promotion of Science (Grant-in-Aid for Creative Scientific Research No. 17GS0420) is acknowledged and appreciated. YK sincerely thanks Dr. Shoichi Kusumoto (Director: Suntory Institute for Bioorganic Research, Osaka) and Dr. Yukishige Ito (RIKEN, Saitama) for generous support and encouragement.

Biographical Information

original image

Yasuhiro Kajihara received his PhD from Tokyo Institute of Technology in 1993 under the supervision of Prof. Hironobu Hashimoto. He spent for two years in Life Science Research Laboratory, Japan Tobacco Inc. as postdoctoral fellow. As a PhD student and post doctoral fellow, he studied the synthesis of glycosyltransferase inhibitors and synthetic method of sugar nucleotides. In 1995, he joined Yokohama City University (YCU) as Assistant Professor and was then promoted to Associate Professor in 2001 and Professor in 2007. In YCU, he developed the synthetic method of oligosaccharides as well as glycoproteins. In 2009, he moved to the Department of Chemistry, Osaka University as Professor.

Biographical Information

original image
original image
original image
original image

Naoki Yamamoto (left) received his PhD from YCU in 2006 and studied the synthesis of glycopeptides as well as glycoprotein, MCP3. Ryo Okamoto (second from left) received his PhD from YCU in 2009 and studied new sialylation and new ligation method for the synthesis of large glycopeptides. Takefumi Murase (third from left) received his PhD from YCU in 2010 and studied efficient synthetic method of glycopeptides library. Kiriko Hirano (right) received her PhD from YCU in 2010 and she achieved the synthesis of erythropoietin analogue having homogeneous human complex type sialyloligosaccharides.

Footnotes

  1. 1

    The Japan Chemical Journal Forum and Wiley Periodicals, Inc.

  2. 2

    The Japan Chemical Journal Forum and Wiley Periodicals, Inc.

  3. 3

    The Japan Chemical Journal Forum and Wiley Periodicals, Inc.

  4. 4

    The Japan Chemical Journal Forum and Wiley Periodicals, Inc.

  5. 5

    The Japan Chemical Journal Forum and Wiley Periodicals, Inc.

  6. 6

    The Japan Chemical Journal Forum and Wiley Periodicals, Inc.

  7. 7

    The Japan Chemical Journal Forum and Wiley Periodicals, Inc.

  8. 8

    The Japan Chemical Journal Forum and Wiley Periodicals, Inc.

Ancillary