The development of a one-step amination-cyclization cascade reaction for the synthesis of N-substituted iminosugars from iodo-pentoses and hexoses is reported. This novel methodology allows for the stereoselective conversion of easily accessible iodo-aldoses and iodo-ketoses into iminosugars in a single step, in highly efficient yields (63-95%), and in aqueous media. Furthermore, the use of functionalized amines allows for the synthesis of N-functionalized iminosugars without additional steps. To illustrate this methodology, a number of biologically important iminosugars were prepared, including 1-deoxynojirimycin, (3S,4R,5S,6R)-azepane-3,4,5,6-tetraol, and N-functionalized 1-deoxymannojirimycins.
The development of a one-step amination-cyclization cascade reaction for the synthesis of N-substituted iminosugars from iodo-pentoses and hexoses is reported. This novel methodology allows for the stereoselective conversion of easily accessible iodo-aldoses and iodo-ketoses into iminosugars in a single step, in highly efficient yields (63-95%), and in aqueous media. Furthermore, the use of functionalized amines allows for the synthesis of N-functionalized iminosugars without additional steps. To illustrate this methodology, a number of biologically important iminosugars were prepared, including 1-deoxynojirimycin, (3S,4R,5S,6R)-azepane-3,4,5,6-tetraol, and N-functionalized 1-deoxymannojirimycins.
Iminosugars are naturally occurring monosaccharide
analogues in
which the ring-oxygen is replaced by nitrogen. As carbohydrate mimics,
iminosugars have a variety of biological properties, which is predominantly
due to their ability to interact with the active site of glycosidases,
and to a lesser extent, glycosyltransferases.[1] For example, the piperidine 1-deoxymannojirimycin (DMJ, 1, Figure ) exhibits
promising mammalian α-fucosidase activity,[2] while the gluco-configured 1-deoxynojirimycin
(DNJ, 2) is an α-glucosidase inhibitor, and the
related N-alkylated derivatives Miglustat (3) and Miglitol (4) have found application in
the treatment of Gaucher’s disease[3,4] and
type II diabetes,[5] respectively. The seven-membered
iminosugar (3S,4R,5S,6R)-azepane-3,4,5,6-tetraol (5) inhibits
numerous glycosidases, including α- and β-galactosidase,
β-glucosidase, and α-fucosidase.[6] More recently, various N-functionalized piperidines have been found
to inhibit α-glucosidases that are crucial to protein synthesis
pathways exploited within virus-infected cells, thus allowing for
the potential host-directed treatment of viral infections such as
HIV,[7,8] hepatitis C,[9] and influenza.[10]
Figure 1
Representative Iminosugars.
Representative Iminosugars.The low natural abundance and growing pharmacological
application
of iminosugars have led to this class of compounds remaining a relevant
and interesting target for synthetic chemists. As such, there is a
large body of work pertaining to their syntheses.[11] In recent years, we have established efficient strategies
for the protecting-group-free synthesis of pyrrolidines;[12−14] however, applying this methodology to the synthesis of piperidines
proved more challenging.[15,16]To date, the
most efficient routes for the synthesis of piperidines
rely on either double reductive aminations[17−24] or the incorporation of a nitrogen atom into a ketose sugar followed
by intramolecular reductive amination using either catalytic hydrogenation
or a borohydride reagent. For example, in 1965, Paulsen and co-workers
reported the first synthesis of DNJ[25] and
demonstrated that intramolecular reductive amination of 6-amino-6-deoxy-l-sorbose (6a) proceeds to selectively provide
DNJ (2, Scheme ). DNJ (2) was isolated as the sole isomer, as
the intermediate imine is selectively attacked from the α-face.[17] Several years later, it was shown that 6-deoxy-6-benzyloxycarbonylamido-ketose 6b(26) or 6-deoxy-6-azido-ketose 6c, prepared either enzymatically[27] or chemically,[28] provided convenient
access to DNJ (2) using the same hydrogenation strategy,
again with exclusive formation of the gluco-isomer,
and others have subsequently developed routes for the synthesis of
polyhydroxypiperidines utilizing iminium ions as key reactive intermediates.[29]
Scheme 1
Strategies for the Synthesis of DNJ and
Related Iminosugars
Building on the aforementioned studies and our
experience with
iminosugar synthesis from ω-deoxy-ω-iodo-glycosides,[30,31] we envisioned that iodo-glycosides would be ideal starting materials
for a cascade reaction leading to the efficient syntheses of a variety
of iminosugars with minimal use of protecting groups (Scheme ). Key in our approach is the
avoidance of elaborate protecting-group strategies and functional
group interconversions to reach the imine intermediate. As such, imine
formation between the ketose carbonyl of I and amine II could be followed by an intramolecular displacement of
the iodide to give cyclic imine III, which could then
be stereoselectively reduced to form the desired iminosugar(IV). This overall process not only reduces the number of steps
required for iminosugar synthesis but also keeps the use of protecting
groups to a minimum making the route highly atom-economic. In addition,
this reaction would allow for the incorporation of N-substituents
in the same step.
Scheme 2
Proposed Syntheses of N-Substituted Iminosugars
Results and Discussion
To explore the proposed synthetic
route, we commenced our studies
using d-fructose (7) as the starting sugar (Scheme ). To this end, d-fructose (7) was converted into methyl glycoside 8 and subsequently iodinated to give methyl 6-deoxy-6-iodo-d-fructofuranoside (9) in two steps and 65% overall
yield.[23] Hydrolysis of methyl iodoglycoside 9 was then optimized, whereby exposure to 0.1 M HCl in H2O at 0.3 atm and 50 °C resulted in complete conversion
into 6-deoxy-6-iodo-d-fructose (10), as evidenced
by thin layer chromatography (TLC). While 6-deoxy-6-iodo-d-fructose (10) could be isolated, this compound was
sensitive to pH and decomposed rapidly in vacuo under
weakly acidic or basic conditions. Thus, a crude reaction mixture
of 10 in 0.1 M aqueous HCl was treated with excess AcONH4 (130 equiv) and aq. NH3 (90 equiv) followed by
NaCNBH3 (4 equiv), and the reaction mixture was stirred
at 80 °C for 18 h. After purification by Dowex-H+ resin
and silica gel flash column chromatography, DMJ (1) was
then isolated in 88% yield as the sole stereoisomer.
Scheme 3
Synthesis
of DMJ (1) from d-fructose (7)
Spurred on by these results, we then sought
to optimize the synthesis
of DMJ (1) by altering the reaction conditions for the
reductive amination. A large excess of ammonium salt is a requirement
for the effective reductive amination of aldehydes and aldoses;[32] however, the reductive amination of ketoses
does not suffer from the dimerization reactions observed in the synthesis
of primary amines.[24] Accordingly, the reductive
amination reaction of ketose 10 was repeated with aq.
NH3 in the absence of AcONH4, which led to DMJ
(1) being formed in 89% yield and moreover, facilitated
the purification process. Further optimization studies involving changes
to reaction concentration, temperature, and time then followed, which
ultimately allowed for DMJ (1) to be synthesized in 95%
yield from 6-deoxy-6-iodo-d-fructose (9) via
a two-step-one-pot reaction using aq. NH3 (90 equiv) and
NaCNBH3 (4 equiv) and stirring the reaction mixture at
r.t. for 18 h. Taken together, this resulted in a four-step (three-pot)
synthesis of DMJ (1) in 62% overall yield from readily
available d-fructose (6). Other notable syntheses
of DMJ include those published by Furneaux et al. (five steps, 25%
overall yield from d-fructose),[33] by Maier et al. (seven steps, 35% overall yield from 1,5-anhydro-d-fructose),[34] and several enzymatic
and chemo/enzymatic syntheses (2–5 steps, 9–44% yield).[35−37]To account for the stereoselectivity of the amination–cyclization
cascade reaction en route to the synthesis of DMJ (1),
it is postulated that exposure of iodide 10 to ammonia
leads to ring opening, imine formation, and subsequent intramolecular
iodide displacement at the 6-position to give cyclic iminium ion 11 (Scheme ). Iminium ion 11 then undergoes a stereoselective reduction,
whereby the lowest energy transition state for the reduction is suggested
to occur when the maximum number of substituents is in a pseudoequatorial
orientation,[38−40] which can be achieved when conformer 11 undergoes Si face reduction to give DMJ (1), a selectivity previously observed in Paulsen’s
first synthesis of DMJ through hydrogenation.[17]To determine whether the amination–cyclization cascade
reaction
could be successfully applied to the synthesis of other piperidines,
the synthesis of 1-deoxynojirimycin (DNJ, 2) from l-sorbose (12) was then attempted (Scheme ). First, l-sorbose
(12) was converted into methyl glycoside 13(41) and selectively iodinated at the 6-position
to afford methyl 6-deoxy-6-iodo-l-sorbofuranose (14) in 43% yield (over two steps). Sorbofuranoside 14 was
then subjected to the previously optimized two-step-one-pot amination–cyclization
cascade reaction to yield DNJ (2) selectively and in
95% yield. Once again, the preferential formation of DNJ could be
explained via the stereoselective reduction of the intermediate cyclic
imine so as to generate a transition state that places the maximum
number of substituents in pseudoequatorial orientations. Taken as
a whole, the total synthesis of DNJ was thus completed in four steps
(three-pot) and 41% overall yield from inexpensive and readily available l-sorbose. Other efficient strategies for the synthesis of DNJ
include those involving the use of enzymes (three steps, 55–65%
overall yield from glucose or protected derivatives thereof)[42−44] and a nonenzymatic route developed by Demailly et al. (four steps,
54% overall yield from l-sorbose).[20]
Scheme 4
Synthesis of DNJ (2) from l-sorbose (12)
Next, we attempted to synthesize l-1-deoxygalactonojirimycin
(l-DGJ) from d-tagatose using our amination–cyclization
cascade strategy. First synthesized in 1990,[45]l-DGJ has since been identified as an inhibitor and molecular
chaperone of galactosidases and galactosyl transferases and thus shows
much promise for the treatment of lysosomal storage disorders and
other protein deficiencies.[46−49] While our syntheses of DMJ (1) and DNJ
(2) involved Fischer glycosylation and the subsequent
installation of an iodide at the 6-position of d-fructose
and l-sorbose, respectively, we previously noted that Fischer
glycosylation and iodination of d-tagatose led to a complex
mixture of products.[22] Accordingly, d-tagatose (15) was protected with isopropylidene
groups to give 1,2:3,4-di-O-isopropylidene-d-tagatofuranose 16 in which the 6-position was subsequently
iodinated to give 6-iodo-tagatoside 17 in 78% yield over
two steps (Scheme ).[50] Acid-mediated deprotection of 17 under the agency of 0.15 M HCl at 50 °C and 0.3 atm
for 18 h in MeOH, rather than H2O, led to complete conversion
of 17 to 6-deoxy-6-iodo-d-tagatose, as evidenced
by TLC. The crude reaction mixture was then subjected to NH3 (aq.) and NaCNBH3, and the solution was stirred at room
temperature for a further 18 h. Workup and purification of the reaction
mixture then provided l-DGJ (18) in 86% yield
over the two steps. As with DMJ (1) and DNJ (2), formation of the major product l-DGJ (18) resulted from the stereoselective reduction of an intermediate
imine to place the maximum number of substituents in the pseudoequatorial
orientation. Thus, our total synthesis of l-DGJ was achieved
in four steps (three-pot) and 67% overall yield from d-tagatose.
The most comparable and efficient synthesis was reported by Jenkinson
et al. in 2011,[42] with l-DGJ being
synthesized in four steps and in 66% overall yield from d-tagatose.
Scheme 5
Synthesis of l-DGJ (18) from d-tagatose
(15)
Having demonstrated the versatility of the amination–cyclization
cascade reaction for the synthesis of a variety of piperidines, the
applicability of the route for the preparation of N-functionalized piperidines was then investigated. To this end, methyl
6-deoxy-6-iodo-d-fructofuranoside (9) was treated
with 0.1 M aqueous HCl, and the intermediate, 6-deoxy-6-iodo-fructose 10, was subjected to n-butylamine in the
presence of NaCNBH3 (4 equiv) at room temperature for 18
h (Entry 1, Table ). Following purification by Dowex-H+ ion exchange resin
and silica gel flash column chromatography, N-butyl-DMJ
(19) was isolated in 69% yield (over two steps). Here,
optimization studies revealed that the use of 10 equivalents of n-butylamine gave the best yield of N-butyl-DMJ
without unnecessarily complicating the purification process. Our four-step
synthesis of N-butyl-DMJ (19) from d-fructose is the highest yielding to date (45% overall yield),
with previous syntheses having been achieved via the alkylation of
DMJ,[51,52] the epimerization of N-butyl-DNJ,[53] or through the synthesis and modification of d-mannolactam.[54]
Table 1
Synthesis of N-Functionalized
DMJ Derivativesa,b
All reaction mixtures used NaCNBH3 (4 equiv) and were stirred at room temperature for 18 h followed
by concentration in vacuo and purification using
Dowex-H+ exchange resin and silica gel flash column chromatography.
Isolated yield calculated over
two
steps.
All reaction mixtures used NaCNBH3 (4 equiv) and were stirred at room temperature for 18 h followed
by concentration in vacuo and purification using
Dowex-H+ exchange resin and silica gel flash column chromatography.Isolated yield calculated over
two
steps.Next, the synthesis of N-methyl-DMJ
(20) was attempted (Entry 2, Table ). Given the low boiling point of methylamine
(−6
°C), a large excess of this reagent (26 equiv) was used as the
excess reagent could be readily removed by evaporation following completion
of the reaction. In this way, N-methyl-DMJ (20) was isolated in an excellent 87% yield following purification
using Dowex-H+ exchange resin and silica gel flash column
chromatography (56% overall yield from d-fructose). For the
synthesis of N-benzyl-DMJ (21) (Entry
3), N-phenethyl-DMJ (22) (Entry 4),
and N-(2-hydroxyethyl)-DMJ (23) (Entry
5), 10 equivalents of amine were required to prevent difficulties
in separating the residual amine from the desired products; however
in all instances, the reactions occurred smoothly to give 21, 22, and 23 in 63% yield, 64% yield, and
67% yield, respectively. As with N-butyl-DMJ (19), the total syntheses of N-methyl-DMJ
(20),[55−57]N-benzyl-DMJ (21),[58−60] and N-(2-hydroxyethyl)-DMJ (23)[43,44,51] reported herein represent the
shortest and highest yielding syntheses to date. The synthesis of N-phenethyl-DMJ (22) was hitherto unpublished.Having demonstrated the versatility of our amination–cyclization
cascade reaction when using ketose sugars as starting materials, we
then sought to extend this strategy to include aldoses. To this end, d-xylose (24) was subjected to Fischer glycosylation
and subsequent iodination using previously optimized conditions[12] to give methyl 6-deoxy-6-iodo-d-xylofuranoside
(25) in 60% yield over the two steps (Scheme ). Next, iodoxyloside 25 was refluxed in a 0.3 M HCl solution for an hour, after
which point the in situ generated 5-deoxy-5-iodo-d-xylose
was exposed to NH3 (aq.) and NaCNBH3 at room
temperature for 18 h. Following purification via Dowex-H+ resin and silica gel flash column chromatography, (3R,4r,5S)-piperidine-3,4,5-triol
(26) was isolated in an excellent 88% yield (53% overall
yield from d-xylose). Triol 26 belongs to a
class of iminosugars first synthesized in the mid-1960s,[61−64] with the shortest and highest yielding synthesis of 26 previously being achieved in five steps and 40% overall yield from
methyl 6-deoxy-6-bromo-α-d-glucopyranoside.[65]
Scheme 6
Synthesis of (3R,4r,5S)-Piperidine-3,4,5-triol (26) from d-xylose
(24)
Finally, we sought to extend our methodology
to the synthesis of
azepanes (Scheme ).
Because of the difficulty associated with the hydrolysis of methyl
aldohexosides,[66] our synthesis thus began
with isopropylidene protection and subsequent iodination of d-galactose (27) to give protected iodogalactoside 28 in 78% yield (two steps).[16] To
avoid the formation of methyl 6-deoxy-6-iodo-d-galactopyranose
in the subsequent deprotection step, isopropylidene deprotection was
initially achieved using a 9:1 mixture of AcOH:H2O under
reflux,[67] whereby full conversion to 6-deoxy-6-iodo-d-galactose (29) was observed via TLC analysis
after 2 days. Alternatively, treatment of protected iodogalactoside 28 with a 9:1 mixture of TFA/H2O at 50 °C
and 0.3 atm led to the formation of 6-deoxy-6-iodo-d-galactose
(29) within 30 min. Unlike the previously described iodo-sugars,
iodide 29 was stable in the presence of acid in vacuo (pH = 1) and thus could be isolated following CH2Cl2:H2O extraction and subsequent concentration in vacuo. Next, 6-deoxy-6-iodo-d-galactose (29) was exposed to NH3 (aq.) and NaCNBH3 at room temperature for 18 h. Purification of the crude product
by Dowex-H+ resin, requiring careful elution with 0.1%
aq. NH3, then afforded (3S,4R,5S,6R)-azepane-3,4,5,6-tetraol
(5) in 68% yield over two steps (55% overall yield from d-galactose). Comparably, Wong et al. achieved the most efficient
synthesis of azepane 5 to date in four steps and 63%
overall yield, also from d-galactose.[6]
Scheme 7
Synthesis of (3S,4R,5S,6R)-Azepane-3,4,5,6-tetraol (5) from d-Galactose (27)
Conclusions
In conclusion, we developed an amination–cyclization
cascade
reaction that has been successfully applied to a variety of readily
available ketose and aldose carbohydrate starting materials. In this
way, DMJ (1), DNJ (2), l-DGJ (18), (3R,4r,5S)-piperidine-3,4,5-triol (26), and (3S,4R,5S,6R)-azepane-3,4,5,6-tetraol
(5) were all prepared in four steps and in overall yields
that were comparable or higher than those previously reported. Key
in these syntheses was the formation of an appropriate iodoglycoside
intermediate, which was readily accessible in two steps from the corresponding
monosaccharide. In addition, the scope of the amination–cyclization
cascade reaction was further exemplified through the reaction of 6-deoxy-6-iodo-d-fructose with various amines thereby allowing for the first
reported synthesis of N-(2-phenyl)ethyl-DMJ (22) and the shortest and highest yielding syntheses of N-butyl-DMJ (19), N-methyl-DMJ
(20), N-benzyl-DMJ (21),
and N-(2-hydoxyl)ethyl-DMJ (23) to date.
Given the versatility of this synthetic methodology, it is thus anticipated
that it can be readily adapted to the synthesis of other iminosugars,
particularly other N-functionalized derivatives, without compromising
overall yields.
Experimental Section
Unless otherwise stated, all reactions
were performed under atmospheric
air. THF (Lab-Scan) was distilled from activated zinc prior to use.
MeOH (Pure Science), EtOH (absolute, Pure Science), AcOH (Lab Scam),
CH2Cl2 (LabServ), 30% aqueous NH3 (Fisher Science), isopropanol (BDH), d-fructose (Carbosynth), d-xylose (Carbosynth), d-galactose (Carbosynth), NaCNBH3 (Aldrich), imidazole (Aldrich), I2 (BDH), triphenyl
phosphine (Acros), AcONH4 (Aldrich), aqueous 35% HCl (Univar),
and 98% H2SO4 (Panreac), d-tagatose
(Carbosynth), NaOH (Pure Science), anhydrous CuIISO4 (Scientific & Chemical Supplies), TFA (Aldrich), aminodiphenylmethane
(Aldrich), n-butylamine (Aldrich), ethanolamine (BDH), benzylamine
(Aldrich), phenylethylamine (BDH), and aq. 40% methylamine (BDH) were
used as received. Drum petroleum ether and ethyl acetate were distilled
before use. Distilled H2O was generated using a Millipore
RiOs 8 purifier. Zn dust was activated by the careful addition of
conc. H2SO4 to Zn powder in the presence of
ethanol, and the solid was decanted, washed with ethanol, diethyl
ether, and then finally washed (and stored) in petroleum ether. All
solvents were removed by evaporation under reduced pressure (in vacuo).
Reactions were monitored by TLC analysis on Macherey-Nagel silica
gel-coated plastic sheets (0.20 mm, with fluorescent indicator UV254)
with detection by UV absorption (254 nm), by dipping in 10% H2SO4 in MeOH or 3% ninhydrin in EtOH followed by
charring at ∼150 °C. Column chromatography was performed
on Pure Science silica gel (40–63 μm). Dowex-H+ 50wx8-100 ion exchange resin was activated by 1 hour exposure to
1 M HCl. High-resolution mass spectra were recorded on a Waters Q-TOF
Premier Tandem Mass Spectrometer using positive electrospray ionization.
Optical rotations were recorded using a PerkinElmer 241 polarimeter
at the sodium D-line. Infrared (IR) spectra were recorded as thin
films using a Bruker Tensor 27 Fourier transform infrared spectrometer,
equipped with an attenuated total reflectance sampling accessory,
and are reported in wave numbers (cm–1). Nuclear
magnetic resonance (NMR) spectra were recorded at 20 °C in CDCl3 or D2O using either a Varian Unity-INOVA operating
at 300 MHz or a Varian Unity operating at 500 MHz. Chemical shifts
are given in ppm (δ) and are relative to chloroform or water,
and all given 13C spectra are proton decoupled. NMR peak
assignments were made using correlation spectroscopy, heteronuclear
single quantum coherence, and heteronuclear multiple bond correlation
experiments, and carbohydrate numbering has been employed where possible.
Methyl d-Fructofuranoside (8)
d-Fructose (7, 3.6 g, 20 mmol) and H2SO4 (1.0 mL, 18 mmol) were added to 200 mL of MeOH. After
the solution was stirred for 15 min, aq. NH3 (4 mL, 30%)
was added, and the reaction mixture was concentrated to ca. 50 mL
in vacuo, cooled over ice, filtered, and concentrated in vacuo. The remaining oil was purified by silica gel flash column chromatography
(EtOAc/MeOH, 99/1 to 95/5, v/v) to afford 8 in an anomeric
mixture (3.43 g, 87%). α-8, Rf = 0.57, β-8, Rf= 0.70 (EtOAc/iPrOH/H2O, 6/4/1, v/v/v). HRMS: m/z calcd for [C7H14O6 + Na]+: 217.0682, obsd.: 217.0690.
IR and NMR spectral data matched those previously reported in ref (68).
Methyl 6-Deoxy-6-iodo-d-fructofuranoside (9)
Methyl glycoside 8 (2.02 g, 10.5 mmol), PPh3 (4.12 g, 15.7 mmol), and imidazole (1.54 g, 20.9 mmol) were
dissolved in dry THF (84 mL) and brought to reflux. A solution of
I2 (3.99 g, 15.7 mmol) in THF (42 mL) was added dropwise
to the refluxing solution. The resulting solution was refluxed for
a further 10 mins, cooled to room temperature, filtered over celite
(washing with THF), and concentrated in vacuo. The remaining orange
oil was purified via silica gel flash column chromatography (Petroleum
ether/EtOAc, 4/1 to 1/2, v/v) and reverse-phase column chromatography
(H2O/MeOH, 100/0 to 9/1, v/v) to afford the desired product 9 (2.37 g, 75% yield) as a colorless oil. α-9, Rf = 0.34, 1H-NMR (500 MHz,
D2O) δ 4.16 (d, J3,4 =
2 Hz, 1H, H-3), 3.93–3.89 (m, 2H, H-4, H-5), 3.79 (d, J1a,1b = 12.5 Hz, 1H, H-1a), 3.68 (d, J1b,1a = 12.5 Hz, 1H, H-1b), 3.49 (dd, J6a,6b = 4.5 Hz, J6b,5 = 10.5 Hz, 1H, H-6a), 3.41–3.41 (m, 1H, H–6b), 3.32
(s, 3H, OMe). 13C-NMR (125 MHz, D2O) δ
108.1 (C-2), 81.7 (C-5), 80.7 (C-4), 80.4 (C-3), 57.7 (C-1), 48.2
(OMe), 5.2 (C-6). β-9, Rf = 0.29 (DCM/MeOH, 5/1, v/v). 1H-NMR (500 MHz, D2O) δ 4.20 (d, J3,4 = 8 Hz, 1H,
H-3), 4.06 (t, J4,3 = 8 Hz, 1H, H-4),
3.88–3.84 (m, 1H, H-5), 3.71 (d, J1a,1b = 12.5 Hz, H-1a), 3.66 (d, J1b,1a =
12.5 Hz, H-1b), 3.49 (dd, J6a,6b = 4.5
Hz, J6b,5 = 10.5 Hz, 1H, H-6a), 3.41–3.41
(m, 1H, H-6b), 3.36 (s, 3H, OMe). 13C-NMR (125 MHz, D2O) δ 103.7 (C-2), 80.1 (C-5), 78.6 (C-4), 77.0 (C-3),
59.4 (C-1), 49.3 (OMe), 6.9 (C-6). IR (film) 3350, 2895, 1462, 1039,
1031 cm–1. HRMS: m/z calcd. For
[C7H13IO5 + Na]+: 326.9699,
obsd.: 326.9704. Spectral data matched those previously reported in
ref (69).
6-Deoxy-6-iodo-d-fructofuranose (10)
Methyl 6-deoxy-6-iodo-d-fructofuranoside (9, 0.25 g, 0.82 mmol) was dissolved in 8.2 mL of a 0.15 M HCl solution
and stirred at room temperature until TLC confirmed full conversion
to 6-deoxy-6-iodo-d-fructofuranoside (ca. 3 days). The resulting
mixture was neutralized using NaHCO3, filtered over celite,
and concentrated in vacuo to give the desired product (0.054 g, 0.18
mmol, 23%), Rf = 0.25 (DCM/MeOH, 5/1,
v/v), 1H-NMR (500 MHz, D2O) δ 4.20 (d, J3,4 = 8.4 Hz, 1H, H-3), 4.13 (t, J4,3 = J4,5 = 7.7 Hz, 1H, H-4),
3.85 (m, 1H, H-5), 3.65 (d, J1a,1b = 12.3
Hz, 1H, H-1a), 3.60 (d, J1b,1a = 12.3
Hz, 1H, H-1b), 3.56 (dd, J6a,6b = 10.8
Hz, J6a,5 = 5.2 Hz, 1H, H-6a), 3.45 (dd, J6b,6a = 10.8 Hz, J6b,5 = 6.3 Hz, 1H, H-6b). 13C-NMR (125 MHz, D2O)
δ 101.5 (C-2), 79.5 (C-5), 78.4 (C-4), 75.5 (C-3), 62.7 (C-1),
7.3 (C-6). HRMS: m/z calcd. For [C6H11IO5 + H]+: 290.9724, obsd.: 290.9728.
Methyl l-Sorbofuranoside (13)
To a flask containing H2SO4 in MeOH (0.03 M,
400 mL) was added l-sorbose (12, 2.00 g, 11.1
mmol), and the reaction mixture was stirred at room temperature for
2 h, before the addition of aq. 35% NH3 (3 mL). The reaction
mixture was concentrated in vacuo to ca. 100 mL and filtered over
celite (cold MeOH wash), and the mother liquor was collected and concentrated.
The residue was purified using silica gel flash column chromatography
(EtOAc to EtOAc/MeOH, 9/1) to give an α,β mixture of methyl-d-sorbofuranoside (1.19 g, 55%), which was used as is for subsequent
reactions. α-13Rf =
0.32 (EtOAc/i-PrOH/H2O, 6/4/1, v/v/v). β-13Rf = 0.30 (EtOAc/i-PrOH/H2O 6/4/1, v/v/v). Spectral data matched those previously reported
in ref (33).
Methyl 6-Deoxy-6-iodo-l-sorbofuranoside (14)
To an α,β-mixture of methyl l-sorbofuranoside
(13, 0.55 g, 2.8 mmol) in THF (28 mL) were added PPh3 (1.85 g, 70 mmol) and imidazole (0.57 g, 84 mmol). The reaction
mixture was heated to 70 °C, and a solution of I2 (1.43
g, 56 mmol) in THF (14 mL) was added portion wise over 5 min. The
reaction mixture was stirred at 70 °C until TLC showed complete
conversion of the starting material to the desired product (ca. 7
h), after which time MeOH (15 mL) was added and the reaction mixture
was concentrated in vacuo. The resulting mixture was purified using
silica gel flash column chromatography (Petroleum ether/EtOAc 4/1
to 1/1, v/v) and HP20 (H2O to H2O/MeOH, 9/1,
v/v) to give an α,β-mixture (α:β = 1:5) of
methyl 6-deoxy-6-iodo-l-sorbofuranoside (0.67 g 78%).: α-14Rf = β-14Rf = 0.35 (DCM/MeOH, 5/1, v/v). IR (film)
3401, 2980, 2880, 1462, 1039, 1031 cm–1. β-141H-NMR (500 MHz, D2O) δ 4.41
(m, 2H, H-4, H-5), 4.36 (m, 1H, H-3), 3.76 (d, J1a,1b = 12.2 Hz, 1H, H-1a), 3.66 (d, J1b,1a = 12.2 Hz, 1H, H-1b), 3.35 (dd, J6a,6b = 9.3 Hz, J6a,5 = 5.4 Hz,
1H, H-6a), 3.29–3.23 (m, 1H, H-6b), 3.31 (s, 3H, OMe).13C-NMR (125 MHz, D2O) δ 108.2 (C-2), 80.6
(C-4), 75.2 (C-3), 71.5 (C-5), 58.5 (C-1), 48.8 (OMe), −0.4
(C-6). HRMS: m/z calcd. For [C7H13IO5 + Na]+: 326.9699, obsd.: 326.9694.
1,2:3,4-Di-O-isopropylidene-d-tagatofuranose
(16)
Anhydrous Cu(II)SO4 (4.17 g, 26 mmol) and d-tagatose (15, 1.17 g, 6.5 mmol) were added to a flask under an argon atmosphere.
To this flask, H2SO4 (36 mM) in acetone (distilled
and degassed, 22 mL) was added, and the resulting mixture was stirred
at room temperature for 18 h. The reaction mixture was quenched with
sodium carbonate, filtered over celite, concentrated, and purified
via silica gel flash column chromatography (Petroleum ether/EtOAc,
100/0 to 4/1, v/v) to give 16 as a colorless oil (1.47
g, 87% yield), Rf = 0.3 (petroleum ether/EtOAc,
1/1, v/v). Spectral data matched those previously reported in ref (42).
Diisopropylidene-protected
sugar 16 (2.31 g, 8.9 mmol), PPh3 (6.75 g,
25.8 mmol), and imidazole (1.81 g, 26.6 mmol) were added to freshly
distilled THF (89 mL), and the solution was brought to reflux. To
this, I2 (4.56 g, 18 mmol) in THF (44 mL) was added dropwise
over 1.5 h. The resulting mixture was refluxed for a further 12 h,
then quenched with methanol, and concentrated. The residue was subjected
to silica gel flash column chromatography (Petroleum ether/EtOAc,
100/0 to 4/1, v/v) to give iodide 17 as a white crystalline
solid (2.84 g, 87% yield), m.p. 41–42 °C, Rf = 0.8 (petroleum ether/EtOAc, 1/1, v/v). [α]D20 = + 46.1 (c =
1.1, CDCl3). IR (film) 2989, 2391, 1376, 1209, 1028, 851
cm–1. 1H-NMR (500 MHz, CDCl3) δ 4.82 (m, 1H, H-4), 4.63 (d, J3,4 = 5.5 Hz, 1H, H-3), 4.23 (dd, J1a,1b = 9.5 Hz, J1a,3 = 1 Hz, 1H, H-1a), 4.20
(dd, J1a,1b = 9.5 Hz, J1b,3 = 1 Hz, 1H, H-1b), 4.20 (m, 1H, H-5), 3.28 (m, 2H,
H-6a,b), 1.41 (s, 3H, H-8), 1.32 (s, 3H, H-9), 1.47 (s, 3H, H-11),
1.39 (s, 3H, H-12); 13C-NMR (125 MHz, CDCl3)
δ 112.9 (C-2), 111.8 (C-10), 111.8 (C-7), 85.4 (C-3), 79.8 (C-5),
79.7 (C-4), 69.3 (C-1), 26.4 (C-11), 26.4 (C-12), 26.0 (C-8), 25.0
(C-9), −0.9 (C-6). HRMS: m/z calcd. For [C12H20IO5 + H]+: 371.0350,
obsd.: 371.0347. Spectral data matched those previously reported in
ref (70).
General Amination–Cyclization Cascade Reaction Conditions
The methyl iodo-glycosides (1.0 mmol) were added to a solution
of aq. HCl (0.15 M, 10 mL) and stirred under reduced pressure (0.3
atm) at 50 °C until TLC analysis showed full conversion to the
corresponding iodo-ketofuranose. Following this, the appropriate amine
and NaCNBH3 (4 mmol) were added sequentially, and the solution
was stirred at room temperature for 18 h. The resulting mixture was
concentrated in vacuo and purified using Dowex-H+ and silica
gel flash column chromatography.
To a solution of HCl in MeOH (0.15 M, 12 mL) was added 1,2:3,4-di-O-isopropylidene-6-deoxy-6-iodo-d-tagatose (17, 0.45 g, 1.21 mmol), and the solution was put under reduced
pressure (0.3 atm) and heated in a 50 °C water bath until the
starting material had completely reacted (observed via TLC analysis,
2 h), following which H2O (12 mL) was added, the MeOH was
removed in vacuo, and the remaining solution was stirred at room temperature
for 18 h to give complete conversion to 6-deoxy-6-iodo-d-tagatoside
(observed via TLC analysis). Following this, 35% aq. NH3 (6.0 mL, 111 mmol) and NaCNBH3 (0.30 g, 4.84 mmol) were
added sequentially to the reaction mixture, which was stirred at room
temperature for a further 18 h. The resulting mixture was concentrated
in vacuo and purified using Dowex-H+ (5% aq. NH3) and silica gel flash column chromatography (DCM/EtOH/MeOH/ aq.
NH3, 20/2/2/1, v/v/v/v) to give 1-deoxygalactojirimycin
(0.17 g, 86%). Rf = 0.10 (DCM/EtOH/MeOH/
aq. NH3, 5/2/2/1, v/v/v/v). [α]D20 -10.1 (c = 0.96, H2O), lit.[42] [α]D20 -9.2 (c = 0.425, H2O); IR (film) 3380 cm–1; 1H-NMR (500
MHz, D2O): δ 4.19 (dd, J4,3 = 3.0 Hz, J4,5 = 1.4 Hz, 1H, H-4), 4.09
(ddd, J2,1b = 11.4 Hz, J2,3 = 9.6 Hz, J2,1a = 5.3
Hz, 1H, H-2), 3.90 (dd, J6a,6b = 12.3
Hz, J6a,5 = 4.9 Hz, 1H, H-6a), 3.82 (dd, J6b,6a = 12.2 Hz, J6b,5 = 8.8 Hz, 1H, H-6b), 3.66 (dd, J3,2 =
9.7 Hz, J3,4 = 3.0 Hz, 1H, H-3), 3.53
(dd, J1a,1b = 12.5 Hz, J1a,2 = 5.4 Hz, 1H, H-1a), 3.44 (ddd, J5,6b = 8.9 Hz, J5,6a = 4.8
Hz, J5,4 = 1.4 Hz, 1H, H-5), 2.90 (dd, J1b,1a = 12.5 Hz, J1b,2 = 11.5 Hz, 1H, H-1b); 13C-NMR (125 MHz, D2O): δ 72.8 (C-3), 66.8 (C-4), 64.6 (C-2), 60.0 (C-5), 59.0
(C-6), 46.0 (C-1). HRMS: m/z calcd. For [C6H13NO4 + H]+: 164.0917, obsd.: 164.0912.
Spectral data matched those previously reported in ref (42).
Methyl 5-Deoxy-5-iodo-α/β-d-xylofuranoside
(25)
To a solution of d-xylose (24, 4.16 g, 27.7 mmol) in MeOH (138 mL), AcCl (0.42 mL) was
added and the reaction mixture was stirred at room temperature for
24 h. The reaction mixture was neutralized by the addition of Dowex-OH–, filtered, and concentrated. The resulting oil was
purified by flash chromatography (MeOH/EtOAc, 1/9, v/v) to give the
pure methyl xylofuranosides. To a solution of the methyl xylofuranosides
(27.7 mmol) in dry THF (152 mL) under an argon atmosphere, PPh3 (10.9 g, 41.5 mmol) and imidazole (3.71 g, 55.4 mmol) were
added. I2 (10.4 g, 41.5 mmol) in dry THF (42 mL) was cannulated
into the reaction vessel. The reaction mixture was refluxed for 2
h, then cooled, filtered, and concentrated. The product was dissolved
in petroleum ether/EtOAc (3/1, v/v), filtered, and then purified by
reverse-phase HP20 (MeOH/H2O, 5/1, v/v) to give methyl
5-deoxy-5-iodo-α/β-d-xyloside (25) as a colorless oil (4.63 g, 61%). Rf = 0.65 (EtOAc/MeOH, 9/1, v/v); [α]D20 = −19.7 (c = 1.5, CHCl3); IR (film) 3446, 1216, 770 cm–1; α-25:1H-NMR (500 MHz, CDCl3): δ 5.06 (d, J1,2 = 4.4, 1H, H-1), 4.40 (m, 1H, H-4), 4.29
(dd, J2,3 = 3.3 Hz, J3,4 = 4.6 Hz, 1H, H-3), 4.17 (dd, J1,2 = 4.4 Hz, J2,3 = 3.3 Hz, 1H,
H-2), 3.51 (s, 3H, OMe), 3.31 (dd, J4,5a = 7.6 Hz, J5a,5b = 9.8 Hz, 1H, H-5a),
3.25 (dd, J4,5b = 6.1 Hz, J5a,5b = 9.8 Hz, 1H, H-5b); 13C-NMR (125 MHz,
CDCl3): δ 102.2 (C-1), 79.1 (C-4), 78.4 (C-2), 76.9
(C-3), 56.2 (OMe), 1.6 (C-5); β-25:1H-NMR (500 MHz, CDCl3): δ 4.93 (s, 1H, H-1), 4.60
(dt, J3,4 = 3.9 Hz, J4,5a = J4,5b = 7.7 Hz, 1H,
H-4), 4.28 (s, 1H, H-2), 4.14 (d, J3,4 = 3.9 Hz, 1H, H-3), 3.40 (s, 3H, OMe), 3.32 (d, J4,5 = 7.7 Hz, 2H, H-5a,b); 13C-NMR (125 MHz,
CDCl3): δ 109.0 (C-1), 83.7 (C-4), 79.7 (C-2), 76.0
(C-3), 55.5 (OMe), 1.9 (C-5); HRMS: m/z calcd. For
[C6H11O4I + Na]+: 296.9594,
obsd.: 296.9601.
(3R,4r,5S)-Piperidine-3,4,5-triol (26)
A solution of
methyl 5-deoxy-5-iodo-α/β-d-xyloside (25, 0.39 g, 1.42 mmol) in aqueous HCl (0.3 M, 14 mL) was refluxed until
full conversion of the starting material was observed via TLC (ca.
1 h). Following this, 35% aq. NH3 (7 mL, 130 mmol) and
NaCNBH3 (0.36 g, 5.68 mmol) were added sequentially to
the reaction mixture, which was stirred at room temperature for a
further 18 h. The resulting mixture was concentrated in vacuo and
purified using Dowex-H+ (5% aq. NH3) and silica
gel flash column chromatography (DCM/EtOH/MeOH/aq. NH3,
20/2/2/1, v/v/v/v) to give (3R,4r,5S)-piperidine-3,4,5-triol (0.16 g, 88%). Rf = 0.10 (DCM/EtOH/MeOH/aq. NH3 5/2/2/1
v/v/v/v); IR (film) 3420, 3350, 2902, 2887 cm–1; 1H-NMR (500 MHz, D2O): δ 3.64 (ddd, J2,1b = 10.4 Hz, J2,3 = 8.7 Hz, J2,1a = 4.8 Hz, 2H, H-2),
3.40 (t, J3,2 = 8.7 Hz, 1H, H-3), 3.30
(dd, J1a,1b = 12.7 Hz, J1a,2 = 4.8 Hz, 2H, H-1a), 2.71 (dd, J1b,1a = 12.7 Hz, J1b,2 = 10.4
Hz, 2H, H-1b); 13C-NMR (125 MHz, D2O): δ
76.1 (C-3), 68.5 (C-2), 47.4 (C-1). HRMS: m/z calcd.
For [C6H12NO3 + H]+: 135.0812,
obsd.: 135.0817. Spectral data matched those previously reported in
ref (74).
Authors: Sarah F Jenkinson; George W J Fleet; Robert J Nash; Yuriko Koike; Isao Adachi; Akihide Yoshihara; Kenji Morimoto; Ken Izumori; Atsushi Kato Journal: Org Lett Date: 2011-07-11 Impact factor: 6.005
Authors: Andreas J Steiner; Arnold E Stütz; Chris A Tarling; Stephen G Withers; Tanja M Wrodnigg Journal: Carbohydr Res Date: 2007-03-28 Impact factor: 2.104