Dirk Schmidt1, Joachim Thiem. 1. Faculty of Science, Department of Chemistry, University of Hamburg, Martin-Luther-King-Platz 6, D-20146 Hamburg, Germany.
Abstract
α ,2-3- and α ,2-6-sialylated lactosaminide precursor structures obtained by various enzymatic procedures could be used for glycosylations employing triflic acid/N-iodosuccinimide. Easily accessible selectively protected lactoside derivatives served as acceptor disaccharides to give the corresponding human milk pentasaccharides in good yields. These were characterized by spectroscopic means in the form of their peracetylated derivatives.
α ,2-3- and α ,2-6-sialylated lactosaminide precursor structures obtained by various enzymatic procedures could be used for glycosylations employing triflic acid/N-iodosuccinimide. Easily accessible selectively protected lactoside derivatives served as acceptor disaccharides to give the corresponding human milk pentasaccharides in good yields. These were characterized by spectroscopic means in the form of their peracetylated derivatives.
From an inspection of contemporary syntheses of biologically and medicinally relevant oligosaccharides, it is evident that the majority is performed either by classical chemical methods or exclusively by enzymatic procedures. Even although considerable progress has been reported during the recent decades, every synthesis of a complex heterooligosaccharide still represents a challenge. To arrive at an oligosaccharide structure with specific patterns of substitution and defined regio- and stereochemical layout, all the presently available procedures need to be checked for efficiency with respect to not only all the above points, but also the efforts required at the purification steps as well as the yields. Whilst both approaches can be employed advantageously in certain cases, in others this is certainly not so.For instance, for structures that contain glycosamino units, it has been demonstrated that these units can be introduced by classical methods in high yields and with good to excellent stereochemical control. In case of glucosylations and galactosylations, and in particular for β-galactosylations, classical and enzymatic methods are almost equal in terms of stereoselectivity and transfer efficiency. By contrast, for syntheses of sialylated structures, enzymatic procedures are still considerably superior to classical chemical sialylations with respect to both stereochemical outcome and preparative input.The use of both procedures in a synergistic mode should also be considered. One of the general approaches ideally suited in such cases is the block synthesis method. Moreover, in recent years a number of combined chemical and chemoenzymatic syntheses have been reported [1-4].As a proof of principle, we were interested to employ some trisaccharide building units previously obtained by enzymatic routes in block syntheses en route to interesting structures. To this end, two human milk pentasaccharides of prominent importance, sialyllacto-N-tetraose (1) and an epimer of sialyllacto-N-neotetraose (2) (Figure 1) were selected as target molecules. Both these pentasaccharides, Neu5Acα2-3Galβ1-3GlcNAcβ1-3Galβ1-4Glc (1) and Neu5Acα2-6Galβ1-4GlcNAcβ1-4Galβ1-4Glc (2), are dominant constituents of complex human milk oligosaccharides (Figure 1). They are considered to play a major role in immuno defense against bacterial and viral infections in the gastrointestinal tract of infants [5]. It is thought that they effectively inhibit bacterial adhesion to epithelial surfaces and so block the first stages of infection processes. Thus, these human milk oligosaccharides are considered as soluble receptor analogues of epithelial cell surfaces [6].
Figure 1
Structures of pentasaccharides 1 and 2.
Structures of pentasaccharides 1 and 2.
Results and Discussion
Previously, we reported the chemoenzymatic synthesis of the 3-sialylated lactosamine derivative 3 obtained by the enzymatic β-galactosylation of the 2-azidothioglucoside with p-nitrophenyl β-galactopyranoside and β-galactosidase (Bovin testes). The subsequent transsialylation was carried out with p-nitrophenyl sialoside (pNp-αNeu5Ac) and either sialidase from Salmonella typhimurium or from Newcastle disease virus [7]. Recently, a more effective higher yielding transfer has been reported in which sialylation with recombinant transsialidase (Trypanosoma cruzi) gave the trisaccharide 3 in 32% yield [8]. Treatment of 3 with methanol and acidic ion exchange resin led to the methyl ester (for the method cf. lit. [9]) which was then peracetylated to give trisaccharide 4 as the donor building block.For formation of the disaccharide acceptor 6, a straight-forward three-step standard reaction sequence was used [10]. Methyl β-lactoside was isopropylidenated at 3′,4′-position with dimethoxypropane and p-toluene sulfonic acid in DMF/acetone. Peracetylation (Ac2O/Py) and subsequent cleavage of the isopropylidene group with 80% acetic acid at 80 °C gave the diol acceptor 6. Since it is known that in galactopyranosyl structures the nucleophilicity of 3-OHconsiderably exceeds that of the 4-OH-group, further protecting group manipulations were not required.Glycosylation of 4 by 6 catalyzed by N-iodosuccinimide and trifluoromethane sulfonic acid (as introduced by van Boom et al. [11]) gave the β,1-3-linked pentasaccharide 7 in 61% yield. About 5% of the corresponding α,1-3-linked compound and ca. 7% of the bis (β,1-3- and β,1-4-) linked octasaccharide were observed as side products and separated by chromatography but these were not further characterized. Reduction of the 2′′′-azido to the 2′′′-amino group with nickel boride [12-13] followed by peracetylation gave 8 in 81% yield (Scheme 1). 1H NMR spectrum contained a doublet for H-1″ at δ 4.96 (J1″2″ = 8.0 Hz) and a down field shifted doublet for H-4′ at δ 5.37 (J3′4′ = 2.9 Hz).
Scheme 1
Preparation of pentasaccharide 8. 1) MeOH, acidic ion exchange resin; 2) Ac2O, pyridine; 3) 80% HOAc, 90 °C; 4) NIS, CF3SO3H, 61%; 5) NiCl2•6H2O, H3BO3, EtOH, then NaBH4, EtOH and acidic workup.
Preparation of pentasaccharide 8. 1) MeOH, acidic ion exchange resin; 2) Ac2O, pyridine; 3) 80% HOAc, 90 °C; 4) NIS, CF3SO3H, 61%; 5) NiCl2•6H2O, H3BO3, EtOH, then NaBH4, EtOH and acidic workup.A similar approach was employed for the synthesis of the protected epimer of sialyllacto-N-neotetraose 14. β-Galactosylation of 2-azidothioglucoside with p-nitrophenyl β-galactopyranoside and β-galactosidase (Bacillus circulans) gave the β,1-3-linked isolactosamine derivative. Further sialylation at position 3′-OH with pNp-αNeu5Ac and either sialidase from Vibrio cholerae or Clostridium perfringens afforded the α,2-6-sialylated trisaccharide 9 exclusively [7]. Later studies showed that 9 could be obtained in an enhanced yield of 32% by transsialylation with recombinant transsialidase (Trypanosama cruzi) [8]. Formation of the methyl ester and peracetylation led to the trisaccharidedonor building block 10.Synthesis of the disaccharide acceptor in this case started from methyl β-lactoside, which was transformed into its 4′,6′-benzylidene-protected derivative 11 in almost quantitative yield by transacetalization with benzaldehydedimethylacetal in acetonitrile under p-toluenesulfonic acidcatalysis. Subsequent peracetylation with acetic anhydride/pyridine, selective cleavage of the benzylidene group with 80% acetic acid at 90 °C and finally treatment with tert-butyldiphenylsilyl chloride and imidazole in DMF afforded the disaccharide 12 (cf. references [14-15]).After activation of the trisaccharidedonor 10 with N-iodosuccinimide and trifluoromethanesulfonic acid, the disaccharide acceptor unit 12 could be glycosylated to give the β,1-4-linked pentasaccharide derivative 13 in 53% yield. In addition, the corresponding α,1-4-linked pentasaccharide was obtained in 8% yield.Finally, the azido group was reduced by the nickel boride method with sodium borohydride, nickel chloride and boric acid [12-13]. During this step partial cleavage of the tert-butyldiphenylsilyl groups was also observed. Complete removal was achieved with trifluoroacetic acid in dichloromethane. For characterization purposes, peracetylation was carried out to give the completely protected pentasaccharide 14 in 67% yield (Scheme 2). As evident from a comparison of the 1H NMR data of 14 with the precursor tri- and disaccharide units 10 and 12, the novel characteristic doublet for the anomericH-1″ of the β-GlcNAc unit at δ 5.12 (J1″2″ = 8.2 Hz) as well as the downfield shift Δδ 0.15 of H-4′ to δ 4.14 compared to 12 were in accord with structure of the target pentasaccharide.
Scheme 2
Preparation of pentasaccharide 14. 1) MeOH, acidic ion exchange resin; 2) Ac2O, pyridine; 3) 80% HOAc, 90 °C; 4) TBDPSCl, imidazole, DMF; 5) NIS, CF3SO3H, 53%; 6) NiCl2•6H2O, H3BO3, EtOH, then NaBH4, EtOH, then acidic workup; 7) CF3CO2H, CH2Cl2.
Preparation of pentasaccharide 14. 1) MeOH, acidic ion exchange resin; 2) Ac2O, pyridine; 3) 80% HOAc, 90 °C; 4) TBDPSCl, imidazole, DMF; 5) NIS, CF3SO3H, 53%; 6) NiCl2•6H2O, H3BO3, EtOH, then NaBH4, EtOH, then acidic workup; 7) CF3CO2H, CH2Cl2.
Conclusion
In this contribution chemoenzymatically generated sialyl α,2-3- and sialyl α,2-6-glycosylated thiophenol 2-azido-lactose derivatives were employed as precursors for sialylated lactosaminidedonor substituents in triflic acid/N-iodosuccinimide glycosylations. With easily accessible selectively unprotected lactose acceptor glycosides the pentasaccharide structures sialyllacto-N-tetraose and the epimer of sialyllacto-N-neotetraosecould be obtained in good yields, and subsequently transformed into their peracetylated derivatives for structure elucidation. Thus, a combination of enzymatic and purely chemical procedures was shown to be advantageous in the preparation of complex oligosaccharides.
Experimental
For general methods cf. reference [16]. The NMR data for the saccharide rings in the pentasaccharides 7, 8, 13 and 14 are denoted according to the Roman numberals I-V from the reducing end, as depicted for compounds 8 and 14 (Figure 2):
Figure 2
Roman numbering of saccharide units in all pentasaccharides for NMR assignment.
Authors: Andrea Marchesi; Fabio Parmeggiani; João Louçano; Ashley P Mattey; Kun Huang; Tanistha Gupta; Mario Salwiczek; Sabine L Flitsch Journal: Angew Chem Int Ed Engl Date: 2020-10-02 Impact factor: 15.336