Edward S Hems1, Sergey A Nepogodiev1, Martin Rejzek1, Robert A Field2. 1. Department of Biological Chemistry, John Innes Centre, Norwich Research Park, Norwich, NR4 7UH, UK. 2. Department of Biological Chemistry, John Innes Centre, Norwich Research Park, Norwich, NR4 7UH, UK. Electronic address: rob.field@jic.ac.uk.
Abstract
A suite of glycosylated glycerol derivatives representing various fragments of the glycosylated ichthyotoxins called prymnesins were chemically synthesised. Glycerol was used to represent a small fragment of the prymnesin backbone, and was glycosylated at the 2° position with the sugars currently reported to be present on prymnesin toxins. Neighbouring group participation was utilised to synthesise 1,2-trans-glycosides. SnCl2-promoted glycosylation with furanosyl fluorides gave 1,2-cis-furanosides with moderate stereocontrol, whilst TMSOTf promoted glycosylation with a furanosyl imidate gave a 1,2-cis-furanoside with good stereocontrol. The chemical synthesis of two larger glyceryl diglycoside fragments of prymnesin-1, glycosylated with α-ʟ-arabinopyranose and α-ᴅ-ribofuranose, is also described. As the stereochemistry of the prymnesin backbones at this region is undefined, both the 2R- and 2S- glycerol isomers were synthesised. The separated diastereoisomers were distinguished by comparing NOESY NMR with computational models.
A suite of glycosylated glycerol derivatives representing various fragments of the glycosylated ichthyotoxins called prymnesins were chemically synthesised. Glycerol was used to represent a small fragment of the prymnesin backbone, and was glycosylated at the 2° position with the sugars currently reported to be present on prymnesin toxins. Neighbouring group participation was utilised to synthesise 1,2-trans-glycosides. SnCl2-promoted glycosylation with furanosyl fluorides gave 1,2-cis-furanosides with moderate stereocontrol, whilst TMSOTf promoted glycosylation with a furanosyl imidate gave a 1,2-cis-furanoside with good stereocontrol. The chemical synthesis of two larger glyceryl diglycoside fragments of prymnesin-1, glycosylated with α-ʟ-arabinopyranose and α-ᴅ-ribofuranose, is also described. As the stereochemistry of the prymnesin backbones at this region is undefined, both the 2R- and 2S- glycerol isomers were synthesised. The separated diastereoisomers were distinguished by comparing NOESY NMR with computational models.
Prymnesins-1 and -2 are potent ichthyotoxins produced by Prymnesium parvum, a flagellated alga which is widely distributed across brackish waters [1]. As a result, P. parvum has been attributed to large scale fish kills globally, resulting in detrimental ecological and economic impacts to affected areas [2]. There are currently 16 different prymnesins reported in literature [3,4]. According to the structure of their ladder-frame polyether backbones, prymnesins can be divided into three main types. The backbone of A-type prymnesins is composed of 91 carbon atoms, B-type of 85 and C-type of 83 carbon atoms. The high diversity of these natural products arises from different patterns of glycosylation combined with the various extent of backbone chlorination. Chemical structures of three prymnesins, prymnesin-1 and prymnesin-2 derived from the A-type backbone (Fig. 1) and Prymnesin-B1, a B-type backbone derivative, have been elucidated, although the absolute configuration of several stereocenters in these molecules remains unknown [[3], [4], [5], [6]].
Fig. 1
The reported structures of the A-type prymnesin ichthyotoxins [[3], [4], [5], [6]].
The reported structures of the A-type prymnesin ichthyotoxins [[3], [4], [5], [6]].In order to gain insight into the structure and properties of prymnesin toxins, we aimed to synthesise a library of glyceryl glycosides which can serve as model fragments of specific toxins relating to prymnesin-1 and prymnesin-2 (Fig. 2). A few glyceryl glycosides have been reported in natural products literature and prepared by chemical synthesis. Thus, 1,3-dihydroxypropan-2-yl α-ᴅ-glucopyranoside is an osmolyte used by cyanobacteria to combat the effects of salt-stress and drought [7]. Glyceryl glycosides have also found use as moisturising ingredients in cosmetics [8]. ‘Floridoside’ (1,3-dihydroxypropan-2-yl α-ᴅ-galactopyranoside) has been isolated from the red alga Mastocarpus stellatus [9] and found to be a potent activator of the classical complement pathway [9]. Although relevant to B-type prymnesins, floridoside's chemical synthesis and characterisation has been reported in the literature [10,11], and so will not be repeated in this study. The synthesis of α-ʟ-arabinopyranoside (2), which is also relevant to our study, was published in 1972 but with little characterisation data.
Fig. 2
Target glyceryl glycosides as fragments of prymnesin-1 (β-ᴅ-galactofuranoside 1, α-ʟ-arabinofuranoside 2 and α-ᴅ-ribofuranoside 3) [3] and prymnesin-2 (either α-ʟ-xylofuranoside 4 or β-ᴅ-arabinofuranoside 5) [3,4].
Target glyceryl glycosides as fragments of prymnesin-1 (β-ᴅ-galactofuranoside 1, α-ʟ-arabinofuranoside 2 and α-ᴅ-ribofuranoside 3) [3] and prymnesin-2 (either α-ʟ-xylofuranoside 4 or β-ᴅ-arabinofuranoside 5) [3,4].One of the problems associated with detection of the prymnesin toxins is that current methods rely on detection of the parent organism [12,13]. None of these methods can provide quantification of toxin levels in water samples. Furthermore, the link between amount of P. parvum in water and prymnesin concentration is not straight forward [1,13,14]. Thus, it has been noticed that in some cases blooms of P. parvum do not necessarily lead to fish kills [15]. Work in our group revealed a new mechanism of P. parvum toxicity, which results from mass cell lysis caused by the infection of alga with a newly discovered double stranded DNA megavirus, PpDNAV [16]. Therefore sensitive and quantitative techniques for direct detection for P. parvum toxins are highly desirable. As an example, antibody recognition has been developed for polyketide algal toxins such as okadaic acid and brevetoxins [17,18]. We decided to focus on a glycosylated polyol fragment of prymnesin-1 and synthesise diglycosides 6 and 7 (Fig. 3) incorporating α-ᴅ-ribofuranosyl and α-ʟ-arabinopyranosyl residues as possible candidates for antigens for anti-carbohydrate antibodies production. These fragments represent the region between C76-C78 on the prymnesin-1 backbone (Fig. 1). The design of 6 and 7 consists of glycerol as a polyol backbone mimic which is glycosylated at positions 1 and 2 and alkylated at position 3 with 3-aminopropyl linker. The linker is required for conjugation of 6 and 7 via cross-linking techniques to a protein carrier, with the aim to make carbohydrate antigens for antibody production [19,20]. As the exact stereochemistry of the glycosylated backbone region of prymnesin-1 is not yet known [3], both the (2R)- and (2S)- glycerol diastereoisomers 6 and 7 were required.
Fig. 3
Target diastereomeric diglycosides 6 and 7.
Target diastereomeric diglycosides 6 and 7.
Results and discussion
Synthesis of 1,2-trans-linked glyceryl glycoside fragments of prymnesin-1
Tin (IV) chloride-promoted glycosylation of 1,3-di-O-benzyl glycerol (9) with per-O-benzoyl-β-ᴅ-galactofuranose (8) [21,22] produced exclusively 1,2-trans-glycoside 10 (Scheme 1). After hydrogenation to give 11, the benzoyl protecting groups were removed using a 5:2:1 mixture of MeOH-H2O-Et3N leading to β-galactofuranoside 1. The configuration of the newly formed glycosidic linkage was confirmed by the appearance of H-1′ signal either as a singlet (δ 5.61 for 10) or a doublet (δ 4.99 with J1,2 1.7 Hz for 1) in the 1H NMR spectra. These J1,2 values are known to be characteristic for a β-galactofuranosidic linkage [22].
Synthesis of 1,3-dihydroxypropan-2-yl β-ᴅ-galactofuranoside (1). Reagents and conditions: (a) SnCl4, CH2Cl2, 39%; (b) H2, 10% Pd/C, EtOAc, Et3N, 64%. (c) MeOH-H2O-Et3N (5:2:1), 87%.The synthesis of 1,3-dihydroxypropan-2-yl α-ʟ-arabinopyranoside (2) has previously been described [23], but the product was only characterised by optical rotation. Silver carbonate-promoted glycosylation of alcohol 9 with bromide 13 prepared from per-O-benzoyl-β-ʟ-arabinopyranose (12) [24,25] gave arabinopyranoside 14 in 50% yield (Scheme 2). The presence of 2-O-benzoyl group in donor 13 ensured the formation of only the 1,2-trans α-anomer 14, and this was confirmed by NMR spectra which showed signals for the anomeric proton (doublet at 5.06 ppm with J1′,2′ 5.9 Hz) and anomeric carbon (100.4 ppm) consistent with α-configuration as followed from comparison with spectra of methyl 2,3,4-tri-O-benzoyl-α- and β-arabinopyranosides [26]. The glycerol benzyl ether protecting groups were removed by hydrogenation over a Pd/C catalyst, followed by NaOMe catalysed removal of the benzoates to give 1,3-dihydroxypropan-2-yl α-ʟ-arabinopyranoside (2) as a white powder. 1H NMR spectroscopy analysis of 2 showed a H-1′ signal at 4.35 ppm as a doublet with a larger J1′,2′ coupling value of 7.5 Hz, which confirmed the presence of the α-anomer.
Synthesis of 1,2-cis-linked glyceryl furanoside fragments
Synthesis of 1,2-cis-furanosides can prove challenging as it can be difficult to control the stereochemistry at the anomeric centre [27,28]. Unless indirect glycosylation approaches are used [[29], [30], [31]], application of non-participating protecting groups remains the main method of minimising the formation of 1,2-trans-glycofuranosylation products.For syntheses of 1,2-cis-furanosides 3–5, we resorted to Mukaiyama's general methodology employing SnCl2-promoted glycosylation with benzylated glycofuranosyl fluorides [32,33]. 2,3,5-Tri-O-benzylated glycofuranosyl fluorides 16–18 were synthesised in high yield by fluorination of the corresponding hemiacetals with DAST (for more details see Suppl. Information) [34,35]. For the glycosylation of 1,3-di-O-benzyl-glycerol (9) with β-d-ribofuranosyl fluoride 16, only the purified β-fluoride was used. As this resulted in a mixture of α- and β-linked products (19a & 19b), we decided that there was no benefit in removing the 1,2-cis fluoride. Therefore, in the case of ʟ-xylofuranosyl and ʟ-arabinofuranosyl glycosides 17 and 18, respectively, α/β-mixtures mixtures of the fluoride anomers were employed (Scheme 3).
Synthesis of prymnesin-1 fragments 3–5. Reagents and conditions: (a) 1,3-di-O-benzyl glycerol, 1 equiv. SnCl2, 1 equiv. TrClO4, MS 4 Å, Et2O; (b) H2, 10% Pd/C, MeOH-EtOH (5:1); (c) 1,3-di-O-benzyl glycerol, 1 equiv. SnCl2, 4 Å MS,Et2O; (d) H2, 10% Pd/C, MeOH-PrOH (9:1); (e) H2 10% Pd/C, MeOH-EtOAc (9:1).All glycosylations, which were promoted by one equivalent of SnCl2 and carried out in the presence of 4 Å MS in diethyl ether, resulted in the formation of anomeric mixtures of furanosides 19–21 in which the 1,2-cis-glycosides dominated. It has been previously noted that in Mukaiyama's synthesis of ribofuranosides both yields and stereoselectivity could be improved with the addition of triphenylmethyl perchlorate (TrClO4) [32]. In our hands, addition of TrClO4 to the reaction involving donor 16 had only a negligible effect on the reaction outcome, so this variation was not applied in glycosylation with 17 and 18.To establish the anomeric configuration of each component of mixtures of glycosylation products 19–21, individual glycosides were isolated and characterised by NMR spectroscopy. It is well known that α- and-β-glycofuranosides can be distinguished on the basis of the differences in values of their 3JH-1,H-2 coupling constants in 1H NMR spectra, which are in the range of 0–2 Hz for of 1,2-trans-furanosides and ∼4 Hz larger for 1,2-cis-furanosides [[36], [37], [38]]. In addition, resonances of anomeric carbons of 1,2-cis-furanosides in 13C NMR spectra appear 4–6 ppm upfield compared to corresponding signals of 1,2-trans-furanosides [37]. Thus, the stereochemistry of the glycosidic linkages in pairs of anomeric glycosides 19–21 was determined using these characteristic 1H and 13C NMR signals, as summarised in Table 1. Finally, per-O-benzylated 1,2-cis-furanosides structurally related to A-type prymnesin were deprotected by catalytic hydrogenolysis over Pd/C (Scheme 3). NMR analyses of resulting α-ᴅ-ribofuranoside 3, α-l-xylofuranoside 4 and β-ᴅ-arabinofuranoside 5 confirmed their anomeric configuration (Table 1).
Table 1
Characteristic data for anomeric signals in 1H and C NMR spectra of pentofuranosides 3–5, 19–21 and 27–28.
Compound
Configuration
δH-1 (ppm)
J1,2 (Hz)
δC-1 (ppm)
3
1,2-cis (α-d-ribo)
5.11
4.3
102.0
4
1,2-cis (α-l-xylo)
5.09
4.4
101.3
5
1,2-cis (β-d-arabino)
5.05
4.7
101.3
19a
1,2-cis (α-d-ribo)
5.38
4.3
101.5
19b
1,2-trans (β-d-ribo)
5.30
0
104.9
20a
1,2-cis (α-l-xylo)
5.35
4.3
99.9
20b
1,2-trans (β-l-xylo)
5.31
1.9
107.0
21a
1,2-cis (β-d-arabino)
5.32
4.4
100.8
21b
1,2-trans (α-d-arabino)
5.35
<1
106.1
27
1,2-cis (α-d-ribo)
5.21
4.2
101.5
28
1,2-cis (α-d-ribo)
5.23
4.4
101.4
Characteristic data for anomeric signals in 1H and C NMR spectra of pentofuranosides 3–5, 19–21 and 27–28.
Synthesis of diastereomeric diglycoside 6 and 7
We initially tried to synthesise the separate enantiomers of 23 by DMTST promoted glycosylation of optically pure glycidol with ethyl 2,3,4-tri-O-benzoyl-1-thio-β-l-arabinopyranoside [39]. Unfortunately the glycosylation was very low yielding and gave a complex mixture of products which proved challenging to separate. As such we changed our approach and for the synthesis of 1,2-O-glycosylated glycerol derivatives 6 and 7 we used allyl alcohol as a precursor. It was first arabinosylated, then oxidised to a diastereomeric mixture of epoxides which were then regioselectively ring-opened with 3-azidopropanol in the presence of Sc(OTf)3 (Scheme 4). This generated an acceptor for a final ribosylation step and introduced a functionalised linker at the same time (Scheme 4). Attempts to glycosylate allyl alcohol with per-O-benzoyl β-ʟ-arabinopyranose (12) in the presence of BF3.OEt2 gave a 3:1 mixture of α and β anomers, which proved very challenging to separate, in a combined 76% yield. To ensure that only the desired α-anomer was synthesised, Koenigs-Knorr glycosylation conditions with perbenzoylated ʟ-arabinopyranosyl bromide 13 were applied, leading to the target 1,2-trans-ʟ-arabinopyranoside 22 in 62% yield.
Synthesis of diastereomeric diglycosides 6 and 7. Reagents and conditions: (a) AllOH, BF3-OEt2, CH2Cl2, α/β 3:1; (b) HBr/AcOH, CH2Cl2, 85%; (c) AllOH, Ag2CO3, DCE, 4 Å MS, 62%; (d) mCPBA, DCE, reflux, 56%; (e) 3-azidopropanol, Sc(OTf)3, toluene, 40%; (f) TMSOTf, CH2Cl2, 4 Å MS, −78 °C, 15% (27), 10% (28); (g) i) H2, 10% Pd/C, EtOAc-MeOH 1:1. ii) MeOH/H2O/Et3N (5:2:1), 59% (6), 41% (7).The 1,2-trans configuration of 22 was confirmed by 1H NMR spectrum which showed the H-1 signal at 4.82 ppm as a doublet with a J1′,2′ coupling value of 6.1 Hz. Epoxidation of the alkene 22 was achieved using mCPBA [8] giving compound 23 as an inseparable 1:1 mixture of (R,S)-diastereomers, as judged by integration of the anomeric proton signals at 4.92 and 4.82 ppm. Using a combination of HSQCed and COSY 2D NMR spectra, it was possible to fully assign the 1H and 13C NMR spectra of the individual isomers. Since both (R)- and (S)-isomers of the prymnesin-1 fragments were targeted, the synthesis continued without diastereomer separation at this stage.The epoxide 23 was ring-opened with 3-azidopropanol in the presence of Sc(OTf)3, leading to installation of a linker with azido functional group, suitable for later reduction and conjugation of the glycoside fragment to a carrier protein for antibody generation. The resulting alcohol 24 remained an 1:1 mixture of (R,S)-diastereomers which were inseparable, but again it was possible to fully assign their 1H and 13C NMR spectra using a combination of HSQCed and COSY 2D NMR. To effect 1,2-cis-ribofuranosylation of 24, we employed imidate 25 which has been used several times to good effect for stereoselective ribosylations using TMSOTf as a promotor and either DCM or DCE as a solvent [[40], [41], [42]]. The initial attempt to couple 24 to 25 was carried out in DCE at −30 °C and resulted in predominant formation of the β-riboside (α/β 1.0:5.8, as judged by 1H NMR). The stereoselectivity improved when the glycosylation was conducted in CH2Cl2 at −78 °C. Since acceptor 24 was a 1:1 (R,S)-mixture, the glycosylation resulted in four products. 1H NMR of the mixture of ribosylation products showed a slight excess of the desired α-ribosides, with an α/β ratio of 1.4:1. The target α-ribosides 27 and 28 were separated from the (R,S)-mixture of β-ribosides 26 using preparative TLC. The stereochemistry of the newly formed 1,2-cis-ribofuranosyl linkages in 27 and 28 was confirmed by 1H and 13C NMR, which showed typical values for 1,2-cis-furanosides (Table 1). To assign (R,S)-configurations in 27 and 28, we used a combination of NMR spectroscopy and modelling approaches. The lowest energy conformations for both isomers were calculated using MarvinSketch 15.1.19.0 Conformer Calculator Plugin (Drieding Force Field), and revealed a noticeable difference in the orientation of the benzyl protecting groups with respect to the azido-propyl linker (Fig. 4). In the computational model for the (S)-isomer 28, the phenyl ring of the 2-OBn group of the ribosyl residue is in close proximity to the azidopropyl group. In contrast, the model of (R)- isomer 27 showed the aromatic benzyl protecting groups clustered together and oriented away from the azidopropyl group. The conclusions arising from the computational models were reinforced by 2D NOESY NMR spectra of compound 28, which show a clear cross-peak between signals for the azidopropyl and the benzyl protecting groups. In contrast, no such cross-peaks can be identified in NOESY spectra of diastereoisomer 27. These observations allowed us to conclude that the chiral centres of the aglycones in compounds 27 and 28 have (R)- and (S)-configurations, respectively.
Fig. 4
3D Model representations of lowest energy calculations (MarvinSketch 15.1.19.0 Conformer Calculator Plugins) for diastereomers 27 and 28 and expanded 2D NOESY NMR spectra.
3D Model representations of lowest energy calculations (MarvinSketch 15.1.19.0 Conformer Calculator Plugins) for diastereomers 27 and 28 and expanded 2D NOESY NMR spectra.Finally, global deprotection of compounds 27 and 28 was achieved by first Pd/C-catalysed hydrogenolysis of benzyl groups with simultaneous reduction of the azide to amine, followed by de-O-benzoylation using a mixture of MeOH-H2O-Et3N (5:2:1). In this way, diastereomeric 2,3-di-O-glycosylated derivatives of glycerol, 6 and 7, equipped with an aminopropyl linker were synthesised in 59% and 41% yields, respectively.
Conclusions
A library of glyceryl glycoside fragments inspired by the structures of prymnesin toxins was prepared. Central to these syntheses were efforts to control glycosylation stereochemistry. The 1,2-trans-linked fragment 1,3-dihydroxypropan-2-yl β-ᴅ-galactofuranoside (1) was synthesised using SnCl4 promoted glycosylation with excellent stereocontrol, while 1,3-dihydroxypropan-2-yl α-ʟ-arabinopyranoside (2) was synthesised using Koenigs-Knorr methodology, again with excellent stereocontrol at the anomeric position [43]. 1,2-cis-Glyceryl glycoside fragments 3, 4 and 5 were successfully synthesised from the corresponding glycofuranosyl fluoride donors, but with only moderate stereoselectivity.Elaborating with respect to the stripped-down glyceryl glycoside fragments, two glyceryl diglycoside fragments inspired by prymnesin-1, 6 and 7, were also chemically synthesised. As the stereochemistry of the prymnesin backbone in this region has not been established, both the 2R- and 2S- diastereoisomers with respect to the glycerol backbone were synthesised. This was achieved by glycosylating 3-(3-azidopropoxy)-2-R/S-hydroxypropyl 2,3,4-tri-O-benzoyl-α-ʟ-arabinopyranoside (24) with 2′,3′,5′-tri-O-benzyl-β-ᴅ-ribofuranosyl (N-phenyl)-2,2,2-trifluoroacetimidate (25), which gave a mixture of α/β-ribosides in a ratio of 1.4:1. The α-ribosides were separable both from the β-ribosides (26) and also from each other, to yield the diastereomeric (2R)- and (2S)-3-(3-azidopropoxy)-2-[(2′,3′,5′-tri-O-benzyl-α-ᴅ-ribofuranosyl)oxy]propyl-2″,3″,4″-tri-O-benzoyl-α-ʟ-arabinopyranoside (27 and 28 respectively). The assignment of the stereochemistry at the 2° position of the glycerol backbone was achieved by comparing nOe NMR spectra with computational models of the lowest energy conformations of both isomers. Global deprotection of 27 and 28 yielded the desired (2R)-3-(3-aminopropoxy)-2-(α-ᴅ-ribofuranosyloxy)propyl α-ʟ-arabinopyranoside (6) and (2S)-3-(3-aminopropoxy)-2-(α-ᴅ-ribofuranosyloxy)propyl α-ʟ-arabinopyranoside (7), respectively, which are suitably functionalised for prospective bioconjugation studies.
Experimental
General
Reagents and anhydrous solvents were supplied by Sigma Aldrich, and were used without further purification. Analytical grade solvents were supplied by Fischer Scientific. Protected sugars which were not synthesised in the lab were supplied by Carbosynth. Glassware was oven-dried and purged with nitrogen immediately before use, and reactions requiring inert atmosphere were run under N2.Reactions were monitored by thin-layer chromatography (TLC) on aluminium-backed, pre-coated silica gel plates (Silica Gel 60 F254, Merk) with the indicated eluents, and the TLC plates were visualised under UV light (λ 254 nm) and charring by dipping in ethanol-sulfuric acid (95:5, v/v) followed by heating. Preparative TLC was run on Analtech preparative uniplates (silica gel 1000 μm, 20 × 20 cm) and flash column chromatography (FCC) was performed on a Biotage Horizon Isolera One using pre-packed SNAP ULTRA 25 μm silica gel cartridges.NMR spectra were recorded using a Bruker Ultrashield Plus 400 MHz spectrometer at 298 K and analysed using TopSpin software. Chemical shifts (δ) are reported in ppm with respect to internal tetramethylsilane or the residual HOD signal in D2O. NMR assignments were made with the aid of COSY and HSQCed experiments, with signals of glycosyl residues being distinguished from those of aglycon by numbers with prime and double prime as shown in Fig. 5. Note that compounds 23 and 24 are a mixture of R- and S-isomers; whilst we could not determine which isomer was R- and which was S-, we could determine which molecule the signals related to via 2D NMR. Therefore, all signals marked with an asterisk are for one diastereoisomer; those without an asterisk are the other.
Fig. 5
Numbering system used for assigning NMR spectra.
Numbering system used for assigning NMR spectra.Optical rotation values were measured using a Perkin Elmer® Model 341 Polarimeter at 20 °C at a wavelength of 589 nm unless otherwise noted. Infrared spectra were recorded using a Perkin Elmer® Spectrum BX and UV–vis spectra using a Varian 50 Bio spectrometer. Low resolution mass spectrometry was employed for monitoring reactions using an Advion Expression CMS spectrometer by direct injection or extraction from a TLC plate using an Advion Plate Express, with a 9:1 mixture of methanol-aqueous formic acid (0.1%) used as a mobile phase. For high resolution mass spectrometry (HRMS) 0.1% (w/v) samples were prepared by dissolving or diluting in methanol-formic acid (1:1) and infused into a Synapt G2-Si mass spectrometer (Waters, Manchester, UK) at 5–10 L min−1 using a Harvard Apparatus syringe pump. The mass spectrometer was controlled by MassLynx 4.1 software (Waters). It was operated in high resolution and positive ion mode and calibrated using sodium formate. The sample was analysed for 2 min with 1 s MS scan time over the range of m/z 50–1200 with 3.5 kV capillary voltage, 40 V cone voltage, 120 °C cone temperature. Leu-enkephalin peptide (1 ng μL−1, Waters) was infused at 10 L min−1 as a lock mass (m/z 556.2766) and measured every 10 s. Spectra were generated in MassLynx 4.1 by combining a number of scans, and peaks were centred using automatic peak detection with lock mass correction. For completeness, synthetic procedures to produce all known compounds used in this study are included as electronic Supplementary Information.
Sodium (20 mg) was added to dry methanol (18 mL) under N2. After the cessation of effervescence 1,3-dihydroxypropan-2-yl 2,3,4-tri-O-benzoyl-α-ʟ-arabinopyranoside (15) (180 mg, 0.3 mmol) was added in dry MeOH (8 mL) and the mixture was stirred overnight at room temperature. The reaction mixture was neutralised using Amberlite IR-120 (H+) resin, filtered and the solvent removed in vacuo. The crude mixture was dissolved in H2O and passed through Dowex® 1X2-400 (OH− form) resin (1 g) to give 2 (62 mg, 76%) as a colourless oil. [α]D + 6.8 (c 1.0, H2O) [Lit: [α]D27 + 5 (H2O)]; 1H NMR (400 MHz, D2O): 4.35 (d, J1′,2′ = 7.5 Hz, 1H, H-1′), 3.83–3.78 (m, 2H, H-4′,5′a), 3.77–3.73 (m, 1H, H-2), 3.65–3.62 (m, 6H, H-1,2,5′b,3′), 3.45 (dd, J1′,2′ = 7.5 Hz, J1′,2′ = 9.4 Hz, 1H, H-2′); 13C NMR (100.6 MHz, D2O): 103.4 (C1′), 81.1 (C2), 72.9 (C3′), 71.3 (C2′), 68.4 (C4′), 66.2 (C5′), 61.4 (C1), 61.0 (C3); HRMS (ESI+) m/z calc. for C8H16O7Na+ 247.0794 [M+Na]+ found 247.0798 [M+Na]+.
1,3-Bis(benzyloxy)propan-2-yl 2,3,5-tri-O-benzyl-α-ᴅ-ribofuranoside (19a) and 1,3-bis(benzyloxy)propan-2-yl 2,3,5-tri-O-benzyl-β-ᴅ-ribofuranoside (19b)
2,3,5-Tri-O-benzyl-β-ᴅ-ribofuranosyl fluoride (16) [35] (210 mg, 0.5 mmol), SnCl2 (95 mg, 0.5 mmol) and triphenylmethyl perchlorate (170 mg, 0.5 mmol) were dissolved into dry Et2O (5 mL) containing 4 Å MS (1.0 g) under N2. The reaction vessel was wrapped with aluminium foil to exclude light, cooled to −15 °C. 1,3-Di-O-benzyl glycerol (9) (100 μL, 0.40 mmol) was added in a single portion and the reaction mixture was stirred at −15 °C for 6 h before kept at 4 °C overnight. The reaction mixture was diluted with Et2O (50 mL), filtered, and the filtrate washed with saturated aqueous NH4Cl solution (3 × 10 mL). The organic layer was separated, dried over MgSO4 and filtered before the solvent was removed under reduced pressure. The anomers were separated by FCC to give 19a and 19b as colourless oils.The α-anomer 19a: Rf 0.10 (hexane/Et2O 7:3); [α]D +52.5 (c 1.0, CHCl3); 1H NMR (400 MHz, CDCl3): 7.32–7.18 (m, 25H, Ar), 5.38 (d, J1′,2′ = 4.3 Hz, 1H, H-1′), 4.73–4.38 (m, 10H, 5 × PhCH2), 4.26–4.24 (m, 1H, H-4′), 4.23–4.17 (m, 1H, H-2), 3.82 (dd, J2′,3′ = 7.1 Hz, J3′,4′ = 3.9 Hz, 1H, H-3′), 3.75 (dd, J1′,2′ = 4.3 Hz, J2′,3′ = 7.1 Hz, 1H, H-2′), 3.74–3.59 (m, 4H, H-1,3), 3.41 (dd, J4′,5′a = 3.9 Hz, 2J5′a,5′b = 10.6 Hz, 1H, H-5′a), 3.34 (dd, J4′,5′b = 4.2 Hz, 2J5′a,5′b = 10.6 Hz, 1H, H-5′b); 13C NMR (100.6 MHz, CDCl3): 138.6, 138.6, 138.4, 138.1, 138.0 (5 × Ar), 128.3, 128.3, 128.2, 128.0, 127.9, 127.6, 127.6, 127.6, 127.5, 127.5, 127.4 (Ar), 101.5 (C1′), 81.5 (C4′), 77.2 (C2′), 75.6 (C2), 75.5 (C3′), 73.4, 73.4, 73.3 72.2, 72.0 (5 × CH2Ph), 71.7 (C1), 70.7 (C3), 69.9 (C5′); HRMS (ESI+) m/z calc. for C43H46O7Na+ 697.3141 [M+Na]+ found 697.3129 [M+Na]+.The β-anomer 19b: Rf 0.20 (hexane/Et2O 7:3); [α]D + 42.4 (c 1.0, CHCl3); 1H NMR (400 MHz, CDCl3): 7.31–7.22 (m, 25H, Ar), 5.30 (s, 1H, H-1′), 4.65–4.39 (m, 10H, PhCH2), 4.36–4.30 (m, 1H, H-4′), 4.05–4.01 (m, 1H, H-2), 4.03 (dd, J2′,3′ = 4.8 Hz, J3′,4′ = 7.4 Hz, 1H, H-3′), 3.92 (d, J2′,3 = 4.8 Hz, 1H, H-2′), 3.65–3.42 (m, 6H, H-1,3,5′a,5′b); 13C NMR (100.6 MHz, CDCl3): 138.4, 138.3, 138.3, 138.0, 138.0, 128.4, 128.4, 128.3, 128.0, 127.8, 127.7, 127.6, 127.6, 127.5, 127.5 (Ar), 104.9 (C1′), 80.4 (C4′), 79.7 (C2′), 78.5 (C2), 74.8 (C3′), 73.4, 73.2, 73.0, 72.4, 72.1 (5 × PhCH2), 71.5 (C5′), 70.2 (C1), 70.2 (C3).
1,3-Dihydroxypropan-2-yl α-ᴅ-ribofuranoside (3)
1,3-Bis(benzyloxy)propan-2-yl 2,3,5-tri-O-benzyl-α-ᴅ-ribofuranoside (19a) (80 mg, 120 μmol) was dissolved in MeOH-EtOH (5:1, 12 mL) and palladium on activated charcoal (10% Pd basis) (50 mg) was added. The system was flushed with N2 ( × 3) followed by H2 ( × 3) and stirred overnight at room temperature. After the system had been flushed with N2 ( × 3) the catalyst was filtered through Celite® and the filter was washed with AcOH (50 mL). The organic washes were combined and the solvent was removed under reduced pressure to give 3 (20 mg, 75%); [α]D +68.2 (c 1.0, MeOH); 1H NMR (400 MHz, CD3OD): 5.11 (d, J1′,2′ = 4.3 Hz, 1H, H-1′) 4.01–3.89 (m, 3H, H-4′,2′,3′), 3.72–3.47 (m, 7H, H-2,5′a,5′b,1,3); 13C NMR (100.6 MHz, CD3OD): 102.0 (C1′), 85.9 (C4′), 79.0 (C2), 71.9 (C2′), 70.1 (C3′), 61.9 (C5′), 61.5 (C1), 60.8 (C3); HRMS (ESI+) m/z calc. for C8H16O7Na+ 247.0794 [M+Na]+ found 247.0798 [M+Na]+.
1,3-Bis(benzyloxy)propan-2-yl 2,3,5-tri-O-benzyl-α-ʟ-xylofuranoside (20a) and 1,3-bis(benzyloxy)propan-2-yl 2,3,5-tri-O-benzyl-β-ʟ-xylofuranoside (20b)
2,3,5-Tri-O-benzyl-α/β-ʟ-xylofuranosyl fluoride (17) (200 mg, 0.47 mmol) and 1,3-di-O-benzyl glycerol (10) (100 μL, 0.4 mmol) were dissolved into Et2O (5 mL) containing 4 Å MS (1.0 g). The suspension was cooled in an ice bath and SnCl2 (90 mg, 0.47 mmol) was added in a single portion to initiate the reaction. The reaction mixture was then kept overnight at 4 °C, filtered and the filtrate was concentrated in vacuo to give a crude product, which contained a mixture of α/β -glycosides in 1.0:0.8 ratio as judged by 1H NMR. Purification by FCC afforded 20a and 20b as colourless oils.The α-anomer 20a: colourless oil; yield 59 mg, (19%); Rf 0.48 (hexane/Et2O 7:3); [α]D −39.6 (c 1.0, DCM); 1H NMR (400 MHz, CDCl3): 7.31–7.22 (m, 25H, Ar), 5.35 (d, J1′,2′ = 4.3 Hz, 1H, H-1′), 4.71 (d, 2J = 11.9 Hz, 1H, PhCHH), 4.63 (d, 2J = 11.9 Hz, 1H, PhCHH), 4.57 (d, 2J = 12.1 Hz, 1H, PhCHH), 4.53–4.45 (m, 7H, PhCH2), 4.43–4.41 (m, 1H, H-4′), 4.31 (dd, J2′,3′ = 5.9 Hz, J3′,4′ = 7.0 Hz, 1H, H-3′), 4.20–4.15 (m, 1H, H-2), 3.97 (dd, J1′,2′ = 4.3 Hz, J2′,3′ = 5.9 Hz, 1H, H-2′), 3.74–3.54 (m, 6H, H-5′a,1,3,5′b); 13C NMR (100.6 MHz, CDCl3): 138.4, 138.4, 138.3, 138.3, 137.9, 128.4, 128.4, 128.3, 128.3, 128.0, 127.7, 127.7, 127.6, 127.6, 127.5, 127.5 (Ar), 99.9 (C1′), 83.8 (C2), 81.6 (C3′), 76.0 (C4′), 75.4 (C2), 73.4, 73.4, 73.3, 72.4, 71.8 (4 × PhCH2), 71.5 (C1), 70.4 (C3), 69.4 (C5′); HRMS (ESI+) m/z calc. for C43H46O7Na+ 697.3141 ([M+Na]+) found 697.3129 [M+Na]+.The β-anomer 20b: colourless oil; yield 32 mg(12%); Rf 0.54 (hexane/Et2O 7:3); [α]D +5.6 (c 1.0, DCM); 1H NMR (400 MHz, CDCl3): 7.32–7.23 (m, 25H, Ar), 5.31 (d, J1′,2’ = 1.9 Hz, 1H, H-1′), 4.58–4.40 (m, 11H, 5 × PhCH2, H-4′), 4.11–4.06 (m, 3H, H-2′,3′,2), 3.76 (dd, J4′,5a′ = 4.9 Hz, 2J5′a,5′b = 10.3 Hz, 1H, H-5′a), 3.71 (dd, J4′,5′b = 3.7 Hz, 2J5′a,5′b = 10.3 Hz, 1H, H-5′b), 3.71–3.55 (m, 4H, H-1,3); 13C NMR (100.6 MHz, CDCl3): 138.5, 138.4, 138.3, 138.0, 137.7, 128.4, 128.3, 128.3, 128.3, 127.8, 127.7, 127.7, 127.6, 127.6, 127.6, 127.5, 127.5, 127.5 (Ar), 107.0 (C1′), 86.8 (C2′), 82.0 (C3′), 79.9 (C4′), 75.7 (C2), 73.4, 73.3, 73.3, 72.0, 71.8 (5 × PhCH2), 70.7 (C1), 70.4 (C3), 69.8 (C5′);
1,3-Dihydroxypropan-2-yl α-ʟ-xylofuranoside (4)
1,3-Bis(benzyloxy)propan-2-yl 2,3,5-tri-O-benzyl-α-ʟ-xylofuranoside (20a) (100 mg, 145 μmol) was dissolved in a mixture of MeOH/n-PrOH (9:1) (20 mL) and palladium on activated charcoal (10% Pd basis) (50 mg) was added. The system was flushed with N2 ( × 3) followed by H2 ( × 3) and stirred overnight at room temperature. After the system had been flushed with N2 ( × 3) the catalyst was filtered through Celite® and the filter was washed with MeOH (20 mL). The organic washes were combined and the solvent was removed under reduced pressure to give 4 (16 mg, 45%) as a colourless oil; Rf 0.28 (DCM/MeOH 85:15); [α]D −139 (c 1.0 MeOH); 1H NMR (400 MHz, CD3OD): 5.09 (d, J1′2′ = 4.4 Hz, 1H, H-1′), 4.16–4.08 (m, 2H, H-3′,4′), 3.92 (dd, J1′2′ = 4.4 Hz, J2′,3′ = 4.4 Hz, 1H, H-2′), 3.65–3.51 (m, 7H, H-2,1,3,5′); 13C NMR (100.6 MHz, CD3OD): 101.3 (C1′), 80.0 (C2), 78.7 (C4′), 78.1 (C2′), 75.7 (C3′), 61.7 (C5′), 61.2 (C1), 61.0 (C3); (HRMS ESI+) m/z calc for C8H16NaO7+ 247.0794 ([M+Na]+) found 247.0785 [M+Na]+.
1,3-Bis(benzyloxy)propan-2-yl 2,3,5-tri-O-benzyl-β-ᴅ-arabinofuranoside (21a) and 1,3-bis(benzyloxy)propan-2-yl 2,3,5-tri-O-benzyl-α-ᴅ-arabinofuranoside (21b)
2,3,5-Tri-O-benzyl-α,β-ᴅ-arabinofuranosyl fluoride (18) (950 mg, 2.3 mmol) and 1,3-di-O-benzyl glycerol (9) (540 μL, 2.2 mmol) were dissolved into a suspension of 4 Å MS (1.0 g) in Et2O (7 mL). The suspension was cooled in an ice bath and SnCl2 (440 mg, 2.3 mmol) was added in a single portion to initiate the reaction. The reaction mixture was then kept in a refrigerator overnight at 4 °C, the solid materials were filtered off and the solvent removed in vacuo to give a mixture of anomers which had very close Rf. Three rounds of purification by FCC gave the β-anomer 21a (560 mg), α-anomer 21b (47 mg) and some remaining unseparated mixture of both.Compound 21a: colourless oil; Rf 0.46 (hexane/Et2O 6:4); [α]D −38.7 (c 1.0, CHCl3); 1H NMR (400 MHz, CDCl3): 7.32–7.24 (m, 25H, Ar), 5.32 (d, J1′,2′ = 4.4 Hz, 1H, H-1′), 4.71–4.36 (m, 10H, 5 × PhCH2), 4.13–4.06 (m, 3H, H-3′,4′,2), 4.03 (dd, J1′,2′ = 4.4 Hz, J2′,3′ = 7.0 Hz, 1H, H-2′), 3.71 (dd, J4′,5′ = 3.4 Hz, 2J5a′,5b′ = 10.3 Hz, 1H, H-5a′), 3.64–3.47 (m, 5H, H-5b′,1,3); 13C NMR (100.6 MHz, CDCl3): 138.4, 138.3, 138.3, 138.2, 137.9 (5 × Ar), 128.4, 128.4, 128.4, 128.3, 128.0, 127.8, 127.7, 127.7, 127.7, 127.6, 127.6, 127.6, (Ar), 100.8 (C1′), 83.7 (C2′), 83.1 (C3′), 80.1 (C4′), 75.6 (C2), 73.5, 73.3, 73.1 (3 × PhCH2), 72.6 (C1), 72.3, 71.8 (2 × PhCH2), 71.3 (C5′), 70.3 (C3); HRMS (ESI+) m/z calc. for C43H46O7 697.3136 [M+Na]+ found 697.3132 [M+Na]+.Compound 21b: colourless oil; Rf 0.46 (hexane/Et2O 6:4); [α]D + 26.8 (c 1.0, CHCl3); 1H NMR (400 MHz, CDCl3): 7.35–7.22 (m, 25H, Ar), 5.35 (bs, 1H, H-1′), 4.57–4.41 (m, 10H, 5 × PhCH2), 4.23 (dm, J3′,4′ = 7.3 Hz, 1H, H-4′), 4.15–4.09 (m, 1H, H-2), 4.08 (d, J2′,3′ = 3.4 Hz, 1H, H-2′), 3.92 (dd, J2′,3′ = 3.4 Hz, J3′,4′ = 7.3 Hz, 1H, H-3′), 3.70 (dd, J1,2 = 3.8 Hz, 2J1a,1b = 10.2 Hz, 1H, H-1a), 3.65–3.54 (m, 5H, H-1b,3,5′); 13C NMR (100.6 MHz, CDCl3): 138.4, 138.3, 138.2, 138.0, 137.7 (5 × Ar), 128.4, 128.3, 128.0, 127.8, 127.7, 127.7, 127.6, 127.6, 127.5 (Ar), 106.1 (C1′), 88.5 (C2′), 83.7 (C3′), 80.3 (C4′), 75.0 (C2), 73.4 73.3, 73.3, 72.1, 71.8 (5 × PhCH2), 70.8 (C1), 70.5 (C3), 69.6 (C5′); HRMS (ESI+) m/z calc. for C43H46O7 697.3136 [M+Na]+ found 697.3132 [M+Na]+.
1,3-Bis(benzyloxy)propan-2-yl 2,3,5-tri-O-benzyl-β-ᴅ-arabinofuranoside (21a) (560 mg, 0.8 mmol) was dissolved in a mixture of MeOH/EtOAc (9:1) (50 mL) and palladium on activated charcoal (10% Pd basis) (50 mg) was added. The system was flushed with N2 ( × 3) followed by H2 ( × 3) and stirred for 48 h at room temperature. After the system had been flushed with N2 ( × 3) the catalyst was filtered through Celite® and the filter was washed with MeOH (20 mL). The organic washes were combined and the solvent was removed under reduced pressure to give 5 (80 mg, 45%) as a colourless oil; Rf 0.45 (DCM/MeOH 85:15); [α]D −78.1 (c 1.0, MeOH); 1H NMR (400 MHz, CD3OD): 5.05 (d, J1′,2′ = 4.7 Hz, 1H, H-1′), 4.18 (dd, J2′,3′ = 8.0 Hz, J3′,4′ = 8.0 Hz, 1H, H-3′), 4.00 (dd, J1′,2′ = 4.7 Hz, J2′,3′ = 8.0 Hz, 1H, H-2′), 3.78–3.73 (m, 3H, H-1,4′), 3.72–3.68 (m, 1H, H-2), 3.66–3.61 (m, 4H, H-5′,3); 13C NMR (100.6 MHz, CD3OD): 101.3 (C1′), 82.5 (C2), 80.8 (C4′), 77.6 (C2′), 73.5 (C3′), 61.7 (C5′), 61.4 (C1,C3); HRMS (ESI+) m/z calc. for C8H16O7Na+ 247.0794 [M+Na]+ found 247.0788 [M+Na]+.
Authors: Gerbrand J van der Heden van Noort; Herman S Overkleeft; Gijsbert A van der Marel; Dmitri V Filippov Journal: Org Lett Date: 2011-05-11 Impact factor: 6.005
Authors: Hans A V Kistemaker; Gerbrand J van der Heden van Noort; Herman S Overkleeft; Gijsbert A van der Marel; Dmitri V Filippov Journal: Org Lett Date: 2013-04-24 Impact factor: 6.005
Authors: L Galluzzi; E Bertozzini; A Penna; F Perini; A Pigalarga; E Graneli; M Magnani Journal: Lett Appl Microbiol Date: 2007-12-13 Impact factor: 2.858
Authors: Jerome Naar; Andrea Bourdelais; Carmelo Tomas; Julia Kubanek; Philip L Whitney; Leanne Flewelling; Karen Steidinger; Johnny Lancaster; Daniel G Baden Journal: Environ Health Perspect Date: 2002-02 Impact factor: 9.031
Authors: Anthony Courtois; Christelle Simon-Colin; Claire Boisset; Christian Berthou; Eric Deslandes; Jean Guézennec; Anne Bordron Journal: Mar Drugs Date: 2008-07-10 Impact factor: 5.118