Notice: Wiley Online Library will be unavailable on Saturday 30th July 2016 from 08:00-11:00 BST / 03:00-06:00 EST / 15:00-18:00 SGT for essential maintenance. Apologies for the inconvenience.
A highly stereoselective method for the preparation of nitro- and aminocyclitols, using fructose-1,6-bisphosphate aldolase, which catalyzes the aldol reaction of dihydroxyacetone phosphate (DHAP) on hydroxynitrobutanals, is reported. The key part of the synthesis is based on a one-pot /two-enzyme process whereby three reactions take place; rabbit muscle aldolase (RAMA) catalyzed aldolization, phytase-catalyzed phosphate hydrolysis, and intramolecular spontaneous nitroaldolization. Two families of nitrocyclitols were obtained depending on the carbon configuration in the β position to the nitro group. Reduction of the latter afforded the aminocyclitols. Evaluation of the inhibition properties of the amines towards five commercially available glycosidases has shown selectivity for β-glucosidase and β-galactosidase.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Aminocyclitols, such as valiolamine and voglibose (Figure 1), form a class of compounds with interesting biological properties.1 As glycosidase inhibitors, they are potential tools to study the mechanisms of cellular interactions, biosynthesis of glycoproteins, catabolism of glycoconjugates, and mechanisms of digestion.2, 3 Thus, they are used as antibiotics,4 antidiabetics,5, 1c and antiviral agents.6 More recently, applications of some N-alkylated aminocyclitols as pharmacological chaperones for the treatment of lysosomal storage disorders were developed.7 In particular, N-octyl-4-epi-β-valienamine (NOEV; Figure 1), which can be useful for the treatment of human GM1-gangliosidosis.7d Consequently, these inhibitors have great medical potential and the design of such molecules, as they have been isolated from natural sources, has raised considerable interest.1
In this context, synthetic routes to these aminocyclitol analogues are still desired. Although chemical strategies have been widely developed,1b, d to our knowledge, chemoenzymatic methods to access aminocyclitols or nitrocyclitols have been scarcely used.1d, 8 As part of our studies on the synthesis of new glycosidase inhibitors,9 we have recently developed the first one-pot, straightforward fructose-1,6-bisphosphate aldolase-mediated synthesis of aminocyclohexitols.10 Aldolases, such as fructose-1,6-bisphosphate aldolase, in cascade reactions or one-pot processes have been studied by various groups. The requirement for a specific donor, dihydroxyacetone phosphate, has governed most of the work.11 Other studies have focused on the in situ formation of the acceptor aldehydes.12 More recently, aldol reactions catalyzed by fructose-6-phosphate aldolase have been associated with catalytic hydrogenation to afford iminocyclitols.13
In a strategy explored by our group (Scheme 1), the key to the synthesis is a stereoselective one-pot/two-enzyme process in which the cyclitol ring is built. Three reactions take place; aldolization catalyzed by fructose-1,6-bisphosphate aldolase (rabbit muscle aldolase, RAMA) introducing the 2S,3R configurations, phosphate hydrolysis catalyzed by a phosphatase, and an intramolecular Henry reaction14 (nitroaldolization). In the presence of the racemic 3-hydroxy-4-nitrobutanal, two aminocyclitol isomers 4 and 5 (1S,2S,3R,5S,6R and 1R,2S,3R,5R,6S, respectively) were obtained after reduction of the nitro group. During this remarkable one-pot cascade reaction, four stereocenters were created. Herein, we report on further application of this highly stereoselective one-pot/two-enzyme process using other hydroxynitrobutanals. A short elucidation of the sequence of reactions occurring in this one-pot process is proposed. Our related approach to the synthesis of new aminocycitols is outlined (Scheme 2).
We envisioned that the cyclitols 6 and 7 (4S, R=OH), 8 and 9 (4R, R=OH), 10 and 11 (4R and 4S, R=H) could be derived from the aldolase-catalyzed reaction of dihydroxyacetone phosphate (DHAP) on hydroxylated nitroaldehydes (12 and 13). The aldehydes in turn could be designed from the corresponding ketals 14 and 15. The compound 15 would arise from 14 through elimination to form a double bond, which is then reduced. Finally, it was envisaged that ketal 14 could be prepared from alcohol 16 through ozonolysis and a nitroaldolization with nitromethane. A lipase-catalyzed kinetic resolution of alcohol 16 would afford the nonracemic series of compounds.
Results and Discussion
We started with the exploration of the three-reaction sequence that occurs in the one-pot process. As the use of RAMA in organic synthesis usually involves an enzymatic phosphate hydrolysis with pH modification of the reaction medium, without isolation of the aldol phosphate intermediate, the intramolecular Henry reaction could occur either before or after the phosphate hydrolysis. To elucidate this sequence, we decided to isolate the phosphorylated compounds. For this purpose, we used an easily prepared nitroaldehyde 2 as a model.10b The key DHAP was prepared following our previously published method.15 The preparation of phosphates 17 and 18 is depicted in Scheme 3. These phosphates were precipitated as their barium salts after partial concentration of the reaction mixture and addition of BaCl2 and ethanol. After centrifugation, barium was exchanged with sodium by means of a cationic resin. The sodium salt was isolated with no further purification in 59 % yield. NMR studies have shown that the phosphates, also characterized by a coupled HPLC–mass spectrometry analysis, were cyclized. In conclusion, the Henry reaction occurred at neutral pH before adding the phytase. As in previous studies,10b the diastereomeric ratio was in favor of 17 (1:0.6 from the NMR spectrum).
We then focused on the synthesis of new aminocyclitols 6, 7, 8, 9, 10, and 11 (Scheme 2). We started with the synthesis of alkene 16 and its lipase-catalyzed kinetic resolution by transesterification with vinyl acetate as described by Chênevert et al.,16 with a slight modification: the source of lipase B from Candida antarctica was the immobilized form, commercialized as Novozyme 435, instead of the Chirazyme L-2. For both the alcohol (remaining substrate 16) and the ester (product 19; Scheme 4), the enantiomeric excesses were improved to 98 % (previously 95 %) and the enantioselectivity (E) reached a value greater than 200, compared to that with Chirazyme L-2 (E>100).16
To generate the aldehyde from the alkene group, we found that the ozonolysis performed on the esterified alkene 19 gave better results than that on the alcohol 16. Thus, (S)-16 was classically acetylated to give (S)-19 in 77 % yield. Ozonolysis of the two enantiomers 19 afforded the corresponding acetoaldehydes, which were directly used in the next step without purification. Nitroaldolization with nitromethane in the presence of NaOH afforded the nitrodiols (2S)-14 a and b and (2R)-14 a and b in 46 % and 48 % yields, respectively, over two steps (Scheme 4). They were isolated as a 1:1 mixture of diastereoisomers. No chiral induction was detected and the diastereoisomers were not separable by flash chromatography. Thus, the following reactions were performed on the mixtures.
For the synthesis of the nitrocyclitols 20, 21, 22, and 23 (Scheme 5), the ketal functions of (2S)-14 a and b and (2R)-14 a and b were hydrolyzed in water in the presence of an acidic resin (H+ form). The aldehydes were directly reacted with DHAP in the presence of the RAMA at room temperature and pH 7.5. After 24 h, the pH of the solution was decreased to 3.9 and phytase was added to catalyze the phosphate hydrolysis.
Nitrocyclitols 20 and 21 were synthesized from (2S)-14 a and b in 25 and 16 % yields, respectively (1:0.64 ratio), and nitrocyclitols 22 and 23 were synthesized from (2R)-14 a and b in 32 and 17 % yields, respectively (1:0.53 ratio). Based on the 3S stereoselectivity of the aldolase and the NMR studies (coupling constants and NOESY experiments), the conformation and stereochemistries of the nitrocyclitols were determined. As previously established,10b the configuration of the hydroxy group on C3 of the nitroaldehyde determined the stereoselectivity of the Henry reaction, affording two series of diastereoisomers: 1S,6R when C5 is R (20 (1S,2S,3R,4S,5R,6R) and 22 (1S,2S,3R, 4R,5R,6R)) and 1R,6S when C5 is S (21 (1R,2S,3R,4S,5S,6S) and 23 (1R,2S,3R,4R,5S,6S)).
We then proceeded to synthesize nitrocyclitols 25 and 26 (Scheme 6). The monohydroxylated ketal 15 was prepared from diols (2S)-14 a and b. For this purpose, the diols (2S)-14 a and b were esterified under classical conditions to afford a non-isolated alkene derivative. The latter was then directly reduced using NaBH4 in a methanolic solution in order to yield compound (2S)-24 in 70 % yield (over the two steps). Solvolysis of ester (2S)-24 led to the alcohol (2S)-15 in 77 % yield. The ketal group of 15 was then hydrolyzed in acidic conditions to generate the aldehyde, which was directly used in the next step. The enzyme catalyzed one-pot process described above for compounds 14 a and b was applied.
The expected nitrocyclitol (1S, 6R)-25 was formed in a 44 % yield from ketal (2S)-15. Based on the 3S stereoselectivity of the aldolase and the NMR studies (coupling constants and NOESY experiment), the conformation and stereochemistries of the nitrocyclitol were determined: 25 possessed the configuration 1S,2S,3R,4S,6R. Following the same procedure, from diols (2R)-14 a and b via the alcohol (2R)-15, the nitrocyclitol (1S,6R)-26 was isolated in a 48 % yield. The conformation and stereochemistry of nitrocyclitol 26 were determined to be 1S,2S,3R,4R,6R.
By means of the combined analysis of the stereochemistries described previously10 and these new results, a particular stereoselectivity was revealed for the intramolecular Henry reaction. Depending on the ketal used (i.e., (2R)- or (2S)-14 a and b and (2R)- or (2S)-15), the intramolecular Henry reaction afforded two diastereoisomers in variable ratios or one major diastereoisomer (Figure 2). In the absence of the hydroxy group in the β position to the nitro group (for example 25 or 26), the stereoisomer 1S, 6R is largely preponderant (the other isomer 1R, 6S was not detected). If we reasonably assume that the reaction was under kinetic control, we could propose the “cis-decalin-like” transition states to explain this stereoselectivity (Scheme 7). In these two chair-like structures, the nitro group and hydroxymethylphosphate are in the equatorial position (the unfavorable situation in which the groups would be axial was not envisaged). The transition state corresponding to the formation of A (Scheme 7) could be favored because the C2 and C3 hydroxy groups are in the equatorial position. However, the situation is different when a hydroxy group is present on C5. Two major diastereoisomers, 1S,6R and 1R,6S, were isolated depending on the C5 configuration, with the 1R,6S isomer always being disfavored. To illustrate this situation, compounds 27 and 2910b were selected (Scheme 8 and Scheme 9).
In the transition state proposed for the formation of compound 27, all the substituents (the hydroxy group in 5, the nitro group, and hydroxymethylphosphate) are in pseudo-equatorial positions. The chair obtained is 4C1. To explain the formation of compound 29 (Scheme 9), we could propose the “cis-decalin-like” transition state where the hydroxy group on C5, the nitro group, and hydroxymethylphosphate are still in pseudo-equatorial positions, whereas the hydroxy groups on C2 and C3 are in pseudoaxial positions and the chair is 4C1. Surprisingly, the epimer C of compound 27 was never detected. Thus, the transition state with the axial OH on C5 and the same chair conformations, as proposed for 29, was probably disfavored. As a consequence, the presence of a 1R,6S stereoisomer could be due to a stabilization of the transition state by a possible antiperiplanar overlap between the σ orbital of the C5OH bond and the σ* orbital of the new CC bond being formed. This stabilization would not occur during the formation of the C structure.
We were also intrigued by the variable ratios obtained for the two series of isomers: 1S,6R and 1R,6S. We expected 1:1 ratios, which correspond to the racemic mixtures at the β-carbon to the nitro group. However, none of the ratios corresponded to that value. The 1S,6R isomer was always found preponderant: 1:0.64 for 20/21, 1:0.53 for 22/23 and 1:0.6 for 17/18 (or 27/29). Thus, we decided to investigate the stability of the disfavored isomer (1R,6S) with nitrocyclitol 27 as a model compound. When (1R,6S)-27 was dissolved in 2 % NaHCO3 in water or heated to 50 °C in water, the isomer (1S,6R)-29 was quantitatively isolated. (Scheme 10)
The molecular models constructed for 27 and 29 (Figure 3) show that the substituents on C1 (hydroxymethyl), C6, and C5 are in the equatorial position, with differences remaining in the C2 and C3 hydroxy groups in axial for 27 and equatorial for 29, respectively. Thus, a difference of 9.3 kJ mol−1 for the free energy, calculated at ab initio level in the gas phase (+12.4 kJ mol−1 in aqueous phase), was found as expected in favor of 29.
All of these results suggested that the (1R,6S) isomers were less stable and, more importantly, could be isomerized through a formal twofold retroaldolization (Scheme 11). The structures in brackets are putative and the sequence of the two retroaldolizations remains unknown. To conclude, on the different ratios obtained during our experimental procedures, a slight and partial isomerization could occur, which would explain the variations always in favor of the more stable (1S,6R) isomer.
Finally, the aminocyclitols were prepared by reduction of the nitro group catalyzed by PtO2 under 169 319 Pa (50 psi) of hydrogen (Scheme 12). Aminocyclitols 6, 7, 8, 9, 10, and 11 were isolated in 74, 71, 60, 65, 80 and 80 % yields, respectively.
The new aminocyclitols were screened for inhibitor activity against five commercially available glycosidases (α- and β-glucosidase, α- and β-galactosidase, α-mannosidase) and biological evaluations were carried out as reported in the literature.9 The most relevant results are gathered in Table 1 together with the previously reported activity of compounds 4 and 5.10b The value given for compound 3 was measured more precisely for this study.
Table 1. Inhibitory activity of aminocyclitols; Ki values (in μM). Aminocyclitols are grouped according to their stereochemistry analogies: on the left are the compounds with (1S,6R) configurations (3, 4, 6, 8, 10, and 11) and on the right are the compounds with (1R,6S) configurations (5, 7, and 9).
[a] See Ref. 10b. [b] Competitive inhibitor. [c] Determined with one apparent KM. NI: no inhibition.
None of the compounds showed inhibitory activity at 1 mM concentration towards α-glucosidase from baker’s yeast and α-mannosidase from jack beans. They were found to be more effective against β-glucosidase from almonds, β-galactosidase from A. oryzae, and α-galactosidase from green coffee beans. If we consider the set of results, surprisingly, it must be noted that the best Ki and the best selectivity were obtained with compound 3 against β-galactosidase (Ki=40 μM), which is the less hydroxylated aminocyclitol. From the results for β-glucosidase, both families were active, with compounds 4 and 7 being the most active. Interestingly, for the (1S,6R) family, a C5 hydroxy group (5S) revealed to be important for the inhibition, as 4 was active but 3, 10, and 11 were not. However, the presence of an additional hydroxy group at C4 (R or S) had a negative effect as the inhibition decreased (6 and 8). Regarding the (1R,6S) family, a C4 hydroxy group was found to be important as 5, which does not possess this group, was inactive. Moreover, the S configuration is the most favorable, as 7 is more active than 9, which has the R configuration. The behavior towards α-galactosidase was different. The (1R,6S) compounds (7, 9) are particularly selective, as none of the (1S,6R) compounds (3, 4, 6, 8, 10, or 11) were active. The same discussion about the presence of the C4 hydroxy could be raised: (4S)-7 was more active than (4R)-9, and 5, with no C4 hydroxy, was inactive.
Compounds 3 and 7 from each family were active against β-galactosidase. Within the (1S,6R) family, comparison with compound 3 shows that the presence of one or two more hydroxy groups is always unfavorable as the inhibition largely decreases.
In conclusion, twelve new nitrocyclitols and aminocyclitols were synthesized following a highly stereoselective procedure. Combination of an aldolase and an intramolecular Henry reaction offered an effective methodology to access the polyfunctionalized cyclitols. We also showed that, in this one-pot/three-reaction process, the intramolecular cyclization occurs at an early stage, probably just after DHAP condensation. This cyclization, which is probably under kinetic control, affords access to two stereochemistries: (1S,6R) and (1R,6S). Once the nitrocyclitols are formed, the less stable isomer (1R,6S) could be isomerized in basic medium or thermally into the more stable form (1S,6R). This phenomenon could occur only if we envisage a double retronitroaldolization. Finally, some of the aminocyclitols were found to be quite active towards β-glucosidase, β-galactosidase, and α-galactosidase.
General: All of the reactions were monitored by TLC with Merck 60F-254 precoated silica (0.2 mm) on aluminum. Flash chromatography was performed with Merck Kieselgel 60 (40–63 μm); the solvent systems are given as v/v. NEt3 was distilled over CaH2. Melting points were measured with a Reichert microscope and are uncorrected. 1H NMR (400 MHz) and 13C NMR (100 MHz) spectra were recorded on a Bruker Avance 400 in CDCl3, CD3OD, or D2O (see indication). Chemical shifts (δ) are reported in ppm and coupling constants are given in Hz. IR spectra were recorded on a Perkin–Elmer FTIR Paragon 500. Optical rotations were measured on a JASCO DIP-370 polarimeter with a sodium (589 nm) lamp at 25 °C. High-resolution mass spectra (HRMS) were recorded by the Centre Régional de Mesures Physiques de l′Ouest, Rennes.
Kinetic resolution of alcohol 16: To a solution of alcohol 1612 (4.4 g, 33.8 mmol) in vinyl acetate (60 mL) was added Novozym 435 lipase (C. antarctica, 4.4 g) and the mixture was stirred at room temperature. The reaction course was monitored by GC and when the conversion reached 50 % (36 h), the reaction was quenched by filtration. The solvent was evaporated and the mixture was purified by flash column chromatography (FCC) with pentane/Et2O (6:4). Ester (R)-19 was isolated (2.556 g, 43 %, >98 % ee, =31 (c=1.29 CHCl3)) and alcohol (S)-16 was isolated (2.109 g, 46 % 98 % ee, =−46 (c=1.34 CHCl3)).
(S)-1,1-dimethoxybut-3-en-2-yl acetate 19: To a solution of alcohol (S)-16 (1.3 g, 9.84 mmol, 1 equivalent) in CH2Cl2 (25 mL) was added NEt3 (3.8 mL, 29.53 mmol, 3 equivalents) and Ac2O (2.78 mL, 29.53 mmol, 3 equivalents). The mixture was stirred at room temperature for 24 h and was then concentrated under vacuum. The crude mixture was directly purified by FCC (pentane/Et2O 7:3). (S)-19 was isolated (1.32 g, 77 %). =−30 (c=1.3 CHCl3).
(2S)-1,1-dimethoxy-4-nitrobutan-2,3-diols 14 a and b: A solution of alkene (S)-19 (1.3 g, 7.46 mmol, 1 equivalent) in CH2Cl2 (20 mL) was stirred at −78 °C. Ozone was then bubbled through the solution until the color became blue. After elimination of ozone excess under Ar flow, triphenylphosphine (3.2 g, 8.2 mmol, 1.1 equivalents) was added. The mixture was then allowed to reach room temperature, stirred for 12 h, and concentrated under vacuum. The crude mixture was treated twice by pentane (30 mL) and the crystalline triphenylphosphine oxide was removed by filtration. The crude aldehyde was then directly used in the next step after evaporation of the solvent.
A solution of nitronate was prepared at 0 °C by addition of 6 m NaOH (1.23 mL, 7.46 mmol, 1 equivalent) to nitromethane (403 μL, 7.46 mmol, 1 equivalent) in MeOH (6 mL). After 30 min, this solution was added dropwise at 0 °C to a solution of aldehyde (1 equivalent) in MeOH (6 mL). The solution was stirred at 0 °C for 45 min. The mixture was neutralized with AcOH (425 μL, 7.46 mmol, 1 equivalent), diluted with water (10 mL), and extracted with ether (3×60 mL). The combined organic phases were dried over MgSO4, filtered, and concentrated. The residue was purified by FCC (Cy/AcOEt 2:8), to afford diols (2S)-14 a and b as a slightly yellow oil (0.669 g, 46 %). Mixture of two diastereoisomers (50:50): Rf=0.5 (Cy/AcOEt, 2:8); =20 (c=1.05; CHCl3); IR (film): =3434, 1527, 1380, 1196, 1076 cm−1. 1H NMR (400 MHz, CDCl3): δ=4.68–4.51 (m, 2×2H+1 H); 4.49–4.44 (m, 2×1H+1 H); 3.68 (dd, J=6, 6 Hz, 1 H); 3.54 (d, J=6 Hz, 1 H); 3.48 (s, 4×3 H); 3.21 (s large, 1 H); 2.90 (s large, 1 H) 2.82 ppm (s large, 1 H); 13C NMR (100 MHz, CDCl3): δ=104.6–104.5; 78.5–78; 71.4–70.8; 69.3–67.9; 56.5–56.1–55.4–55.2 ppm. HRMS (LSIMS+) calcd for C6H13NO6 [M+Na]+: 218.0641; found: 218.0632.
(2R)-1,1-dimethoxy-4-nitrobutan-2,3-diols 14 a and b: The same procedure was applied to alkene (R)-19. Diols (2R)-14 a and b were obtained as an inseparable mixture of two diastereoisomers (48 %, 50:50). =−7 (c=1.1; CHCl3). Spectral data were identical to (2S)-14 a and b.
(2S)-1,1-dimethoxy-4-nitrobutan-2-yl acetate 24: To a solution of diol (2S)-14 a and b (300 mg, 1.53 mmol, 1 equivalent) in Et2O (8 mL) was added DMAP (7 mg, 0.06 mmol, 0.04 equivalents) and Ac2O (290 μL, 3.06 mmol, 2 equivalents). The mixture was stirred at room temperature for 2 h. Then MeOH (10 mL) was added and the solution was cooled to 0 °C before fractional addition of NaBH4 (290 mg, 7.65 mmol, 5 equivalents). The solution was stirred at room temperature for 3 h. The mixture was neutralized with saturated aqueous NH4Cl (10 mL) and extracted with Et2O (3×50 mL). The combined organic phases were dried over MgSO4, filtered, and concentrated under vacuum. The residue was purified by FCC (Cy/AcOEt 6:4), to afford (S)-24 as a clear oil (214 mg, 70 %, two steps): Rf=0.61 (Cy/AcOEt, 5:5); =−16 (c=1.1, CHCl3); IR (film): =2943, 1740, 1557, 1374–1237, 1080 cm−1; 1H NMR (400 MHz, CDCl3): δ=4.94 (ddd, 1 H, H2, J(H2,H3a)=J(H2,H3b)=7.5, J(H2,H1)=8 Hz); 4.39 (t, 2 H, H4, J(H4,H3)=7.2 Hz); 4.27 (d, 1 H, H1, J(H1,H2)=8 Hz); 3.37–3.34 (s, 6 H, OCH3); 2.38 (tdd, 1 H, H3a, J(H3a,H3b)=8.8, J(H3a,H2)=7.5, J(H3a,H4)=7.2 Hz); 2.28 (tdd, 1 H, H3b, J(H3b,H3a)=8.8, J(H3b,H2)=7.5, J(H3b,H4)=7.2 Hz); 2.02 ppm (s, 3 H, CH3); 13C NMR (100 MHz, CDCl3): δ=170.6 (CO); 104.2 (C1); 71.9 (C4); 69.8 (C2); 56.1–55.1 (OCH3); 26.5 (C3); 14.2 ppm (CH3). HRMS (LSIMS+) calcd for C6H13NO6 [M+Na]+: 244.0797; found: 244.0804.
(2S)-1,1-dimethoxy-4-nitrobutan-2-ol 15: To a solution of (S)-24 (214 mg, 1.38 mmol, 1 equivalent) in MeOH (10 mL) was added 1 n NaOH (1.65 mL, 1.65 mmol, 1.2 equivalents). The solution was stirred at room temperature for 45 min. The mixture was extracted with ether (3×50 mL). The combined organic phases were dried over MgSO4 and concentrated under vacuum. The residue was purified by FCC (Cy/AcOEt 5:5) to give (S)-15 as a clear oil (183 mg, 77 %): Rf=0.28 (Cy/AcOEt, 5:5); =−28 (c=0.7, CHCl3); IR (film): =3460, 1555, 1382, 1028 cm−1; 1H NMR (400 MHz, CDCl3): δ=4.50 (m, 2 H, H4); 4.11 (d, 1 H, H1, J(H1,H2)=8 Hz); 3.64 (m, 1 H, H2); 3.40–3.37 (s, 6 H, OCH3); 2.33 (m, 1 H, H3a); 1.96 ppm (m, 1 H, H3b); 13C NMR (100 MHz, CDCl3): δ=106.4 (C1); 72.3 (C4); 68.2 (C2); 55.4–55.3 (OCH3); 29.3 ppm (C3). HRMS (LSIMS+) calcd for C6H13NO6 [M+Na]+: 202.0691; found: 202.0701.
(2R)-1,1-dimethoxy-4-nitrobutan-2-yl acetate 24: The procedure described to obtain (S)-24 was applied to (2R)-14 a and b. The residue was purified by FCC (Cy/AcOEt 6:4) to afford (R)-24 as a clear oil (214 mg, 70 %). Rf=0.61 (Cy/AcOEt, 5:5); =16 (c=1, CHCl3). HRMS (LSIMS+) calcd for C6H13NO6 [M+Na]+: 244.0797; found: 244.0804. Spectral data were identical to (S)-24.
(2R)-1,1-dimethoxy-4-nitrobutan-2-ol 15: The procedure described to obtain (S)-15 was applied to (R)-24. The residue was purified by FCC (Cy/AcOEt 5:5) to afford (R)-15 as a clear oil (77 %). Rf=0.28 (Cy/AcOEt, 5/5); =22 (c=1, CHCl3). HRMS (LSIMS+) calcd for C6H13NO6 [M+Na]+: 202.0691; found: 202.0700. Spectral data were identical to (S)-15.
General Procedure for preparation of nitrocyclitol: The ketal (2R)-14 a and b, (2S)-14 a and b, (S)-15 or (R)-15 (1 mmol, 1.2 equivalents) was suspended in water (2 mL) and a cation-exchange resin (Dowex 50×8, H+ form, 1 g) was then added. The suspension was stirred at 45 °C for 2.5 h (quantitative by TLC). The resin was filtered off and rinsed with water and the pH was adjusted to 7.5 with 1 M NaOH. To this solution was added DHAP (0.83 mmol, 1 equivalent) followed by water (10 mL) and the pH was adjusted to 7.5 with 1 M NaOH. The mixture was bubbled with Ar and centrifuged aldolase (40 U) was then added. After stirring for 24 h at room temperature, the mixture was washed with AcOEt (3×20 mL). The pH of the water phase was adjusted to 3.9 with HCl (1 m) and phytase (56 U) was added. The resulting solution was stirred at room temperature for 24 h and was then concentrated under vacuum. The residue was purified by FCC with CH2Cl2/MeOH (85:15 then 80:20), to afford the nitrocyclitols 20 and 21, 22 and 23 or 25, and 26 as brown solids.
(1S,2S,3R,4S,5R,6R)-1-hydroxymethyl-6-nitrocyclohexan-1,2,3,4,5-pentaol 20 and. (1R,2S,3R,4S,5S,6S)-1-hydroxymethyl-6-nitrocyclohexan-1,2,3,4,5-pentaol 21: Compounds 20 and 21 were obtained from ketals (2S)-14 a and b as slightly yellow solids (63 mg, 25 %) and (39 mg, 16 %), respectively.
(1S,2S,3R,4R,5R,6R)-1-hydroxymethyl-6-nitrocyclohexane-1,2,3,4,5-pentaol 22 and (1R,2S,3R,4R,5S,6S)-1-hydroxymethyl-6-nitrocyclohexane-1,2,3,4,5-pentaol 23: Compounds 22 and 22 were obtained from ketals (2R)-14 a and b as slightly yellow solids (73 mg, 32 %) and (45 mg, 17 %) yields, respectively.
[(1S,2S,3R,5S,6R)-1,2,3,5-tetrahydroxy-6-nitrocyclohex-1-yl]methyl phosphate 17 and [(1R,2S,3R,5R,6S)-1,2,3,5-tetrahydroxy-6-nitrocyclohex-1-yl]methyl phosphate 18: The general method described above for nitrocyclitols was followed until the step of aldolase addition. The nitroaldehyde 210b (0.75 mmol, 133 mg, 1 equivalent) and DHAP (0.375 mmol, 0.5 equivalents) were used. After stirring for 24 h at room temperature, the volume of the solution was reduced by half under vacuum, and then BaCl2 (0.75 mmol, 2 equivalents) and EtOH (14 mL) were added. The suspension was kept at 4 °C for 5 h. The precipitate was centrifuged, washed with EtOH, and dried under vacuum. The barium salt was then suspended in water and a cation-exchange resin (Dowex 50Wx8, Na+ form) was added. Water was evaporated under vacuum. The mixture of phosphates 17 and 18 was obtained as a pale yellow solid (66 mg, 59 %, 1:0.6 ratio from NMR).
Isomerization of 27 into 29: Nitrocyclitol 2710b (30 mg) was treated with a 2 % solution of Na2CO3 (600 μL). The mixture was stirred at room temperature for 30 min and then concentrated under vacuum. The crude produce was purified by FCC with CH2Cl2/MeOH (90:10 then 80:20), to afford the nitrocyclitol 29 in a quantitative yield. NMR spectra were identical to those of compound 29.10b
General procedure for reduction of nitrocyclitols: To a solution of nitrocyclitol 20, 21, 22, 23, 25, or 26 (1 mmol) in MeOH/AcOH (95:5, 40 mL) was added PtO2 (40 mg). The mixture was submitted to 169 319 Pa (50 psi) of H2 in a Parr apparatus. After stirring for 48 h at room temperature, the catalyst was removed by ultrafiltration and washed with MeOH. The filtrate was concentrated under vacuum and the crude product was purified by cation-exchange chromatography (Dowex 50Wx8, 200–400 mesh, H+ from) and eluted with 1 M NH4OH. Compound 6, 7, 8, 9, 10, or 11, respectively, was obtained as a white solid.
Molecular Modeling: The diastereoisomers 27 and 29 were designed with the Sybyl17 package. For each of them, both the axial and equatorial conformations of the hydroxymethyl (substituents on C1) were built. One conformational analysis was performed by simulated annealing using the TRIPOS17 force field, Gasteiger–Hückel charges, and a dielectric constant set to 1.50 conformations were generated for each stereoisomer.
The global minima of 27 and 29 were then optimized at ab initio level using the Gaussian 0318 package. The free energies were obtained with DFT19 B3LYP/6-31G** and thermochemistry properties in the gas phase, and with a polarized continuum model (PCM)20 model for the aqueous phase.