The synthesis of a D-glucose-like piperidin-2-one: Isofagomine lactam

Article abstract of DOI:10.1071/CH03228

Benzyl 4-cyano-4-deoxy-α-D-arabinoside was converted into both its 2,3-di-O-acetyl and 2,3-di-O-(tert-butyldimethylsilyl) derivatives. The latter, by a process of hydrogenolysis, oxidation, and methanolysis, gave methyl 2,3-di-O-(tert-butyldimethylsilyl)-

Full text of DOI:10.1071/CH03228

Full Paper
CSIRO PUBLISHING
www.publish.csiro.au/journals/ajc
Aust. J. Chem. 2004, 57, 449–453
The Synthesis of a d-Glucose-like Piperidin-2-one: Isofagomine Lactam
James M. Macdonald A and Robert V. Stick A,B
A Chemistry, School of Biomedical and Chemical Sciences, M313, University of Western Australia,
Crawley WA 6009, Australia.
B Author to whom correspondence should be addressed (e-mail: rvs@chem.uwa.edu.au).
Benzyl 4-cyano-4-deoxy-α-d-arabinoside was converted into both its 2,3-di-O-acetyl and 2,3-di-O-(tert-
butyldimethylsilyl) derivatives. The latter, by a process of hydrogenolysis, oxidation, and methanolysis, gave methyl
2,3-di-O-(tert-butyldimethylsilyl)-4-cyano-4-deoxy-d-arabinonate. Reduction of this methyl ester with borane
dimethyl sulfide gave, after deprotection, isofagomine lactam.
Manuscript received: 8 September 2003.
Final version: 4 November 2003.
Introduction
placed in the pseudo-anomeric and C2 positions, respectively.
Williams et al. have synthesized a number of potent xylo-
biose (imino sugar analogue) inhibitors of the endo-xylanase
Cex from Cellulomonas fimi, including the d-xylosyl xylo-
1-deoxynojrimycin 3,[6] the d-xylosyl xylo-isofagomine 4,[6]
and the d-xylosyl xylo-isofagomine lactam 5[7] (Scheme 2).
Whilst not being quite as potent as the xylo-isofagomine 4,
the xylo-isofagomine lactam 5 shows remarkable inhibition
and binds almost twenty times more tightly than the xylo-
deoxynojirimycin 3. This result stands in contrast to the
inhibition of glycosidases observed for the d-glucono lac-
tam 1, which is a weaker inhibitor than 1-deoxynojirimycin
2. Williams et al. postulated that the remarkable potency of
5 is a consequence of the lactam’s binding in the active site
as its (protonated) tautomer 6 (Scheme 2).[7] This tautomer
approaches an ideal transition-state analogue, fulfilling sev-
eral requirements. The protonated nitrogen can bind tightly
to the catalytic amino acid carboxylate(s), the C N bond
flattens the ring to a conformation potentially more rep-
resentative of the normal transition state, and, unlike an
isofagomine derivative, a hydroxyl group at the pseudo-
C2 position is available for additional stabilization of a
transition-state-like intermediate. It has been shown that a
hydroxyl group at C2 of a normal substrate provides strong
Somewhat surprisingly, glycono-1,5-lactams are effective
inhibitors of various glycoside hydrolases.[1−4] For example,
the d-glucono lactam 1 (Scheme 1) shows good inhibition
against almond β-glucosidase (Ki 37 µM[1] and 51 µM[3]).
The reasons for these inhibitions are not clear but are gener-
ally ascribed to the (half-chair) shape of the molecules, pur-
portedly mimicking the shape of a transition-state involved in
the hydrolytic process.[3,4] Certainly, the glycono lactams are
weaker inhibitors of glycosidases than are the related imino
sugars, for example, 1-deoxynojirimycin 2, where recent
observations suggest that some of these basic amines bind as
their conjugate acids to an inactive form of the hydrolase.[5]
The inhibitory potency of the glycono lactams increases
markedly when the nitrogen atom and carbonyl moieties are
OH
NH
OH
NH
HO
HO
HO
HO
O
HO
OH
1
2
Scheme 1.
OH
O
OH
NH
OH
HO
O
HO
HO
O
HO
HO
HO
O
NH
3 Ki 5.8␮M
4 Ki 0.13␮M
OH
O
OH
O
HO
HO
HO
HO
HO
HO
O
NH
O
N
OH
O
5 Ki 0.34␮M
6
Scheme 2.
© CSIRO 2004
10.1071/CH03228
0004-9425/04/050449

450
J. M. Macdonald and R. V. Stick
interactive forces to active-site residues (particularly of the
Synthesis
catalytic nucleophile), in the order of 10–40 kJ mol−1 [8–10]
.
Apart from the synthetic work of Williams et al., the
d-galacto-isofagomine lactam 9 has been prepared by Bols
and co-workers.[12] We first announced a synthesis of
isofagomine lactam 8 in 2002[13] and now give full details
of this work; Bols and co-workers subsequently reported a
separate synthesis of 8 in 2003.[14]
For our synthesis of the lactam 8, we decided to apply the
retrosynthetic analysis outlined in Scheme 4, with the nitrile
10 being an advanced intermediate in our synthesis of
isofagomine itself.[15] However, we foresaw problems with
the somewhat labile isopropylidene acetal in 10 and chose to
replace it with more robust protecting groups.
The acetal 10 was carefully converted into the diol 11,
and thence the diacetate 12 (Scheme 5). Hydrogenolysis of
the benzyl glycoside then furnished the hemiacetal 13. The
oxidation of 13, utilizing the Swern procedure or pyridinium
dichromate, appeared to proceed well (by TLC analysis) but
isolation of the expected lactone 14 proved troublesome.
Therefore, the hemiacetal was first oxidized and the crude
product subjected to an acid-catalyzed methanolysis. Upon
(re)acetylation of this crude material, the desired acyclic
nitrile 15 was obtained in low yield. More disappointingly,
the nitrile 15 was inert to the normal conditions of catalytic
hydrogenation (PtO2 in methanol).
Borane dimethyl sulfide reportedly reduces nitriles more
rapidly than esters.[16,17] However, treatment of the nitrile
15 with the borane reagent gave an unrecognizable mess.
Bearing in mind that the nitrile 15 also contained various
ester moieties, and considering the unsatisfactory yield of
15, we decided to change the protecting groups for the diol
11, to silyl ethers.
X-ray crystallography, at a resolution of 2.0 Å, of the
enzyme (Cex) inhibitor 5 complex was thought largely to
confirm the hypothesis that the lactam 5 was indeed binding
as its tautomer 6.[7] However, more recent enzyme-inhibitor
X-ray crystallographic studies, conducted at higher resolu-
tion (1.0 Å) and employing a homologous enzyme to Cex
from Cellulomonas fimi, now refute this earlier assertion.[11]
It appears that the observed inhibition is a result of the binding
of the lactam 5 to an inactive form of the enzyme, where the
protonation states of the nucleophile and acid/base residues
are reversed.
Given the unexpected potency of the d-xylosyl xylo-
isofagomine lactam 5 as an inhibitor of glycoside hydrolases,
we wished to prepare the d-glucosyl isofagomine lactam
7 (Scheme 3). This paper reports our efforts towards just
isofagomine lactam 8, the necessary precursor for any
synthesis of 7.
HO
HO
OH
OH
OH
OH
O
O
HO
HO
HO
HO
HO
NH
NH
NH
OH
O
O
O
7
8
9
Scheme 3.
OH
PO
O
HO
HO
NC
NH
O
OP
O
The diol 11 was easily converted into the disilyl ether 16
(Scheme 6). Interestingly, analysis of the 1H NMR spectrum
8
of this disilyl ether showed one large value (10.8) for J4,5
,
4
indicating that the molecule existed solely in a C1 confor-
mation. Such was probably not the case for the molecules
11–13 where there may be contributions from the other chair
(1C4) conformation. This phenomenon has been observed
previously with di- or poly-silylated pyranoses in which
the sugars possess adjacent sterically encumbering silyloxy
groups.[18–21]
PO
OBn
O
O
O
NC
OH
O
NC
OP
10
Scheme 4. Retrosynthetic analysis. P is a protecting group.
HO
OBn
AcO
O
O
O
O
NC
NC
(a)
(b)
(c)
O
NC
OH
OBn
OAc
OBn
10
11
12
CO2Me
AcO
AcO
AcO
O
O
(d)
(e)
OAc
CN
NC
NC
OH
O
CH2OAc
OAc
OAc
13
14
15
Scheme 5. (a) PPTS, MeOH, H2O; (b)Ac2O, pyridine, DMAP; (c) Pd/C, H2,AcOH, H2O, THF; (d) PDC,
4-Å sieves, CH2Cl2; (e) PTSA, MeOH, and then Ac2O, pyridine, DMAP.

The Synthesis of a d-Glucose-like Piperidin-2-one
451
OAc
O
Hydrogenolysis of the disilyl ether 16 gave the hemiacetal
17, and a subsequent Swern oxidation gave the lactone 18 in a
modest yield. More satisfactory was a pyridinium dichromate
oxidation of 17, followed by transesterification of the crude
lactone 18, to give the acyclic methyl ester 19 in a pleasing
yield of 80% over the two steps.
AcO
AcO
NH
22
(a)
(b)
Next, the nitrile 19 was treated with borane dimethyl
sulfide in tetrahydrofuran at reflux (Scheme 7). The
‘borazine’ intermediate was decomposed upon the addition
of hydrochloric acid in methanol to give, presumably, the
hydrogen chloride salt 20 of the amine. This crude material
was then treated with base resin, hopefully to liberate the
free amine that could then perform the desired intramolecu-
lar attack upon the methyl ester.The 1H NMR spectrum of the
crude material indicated the presence of several compounds,
also difficult to separate by TLC, and, therefore, difficult to
purify by flash chromatography. Fortuitously, upon trituration
of this crude material with an ethyl acetate/petrol mixture, a
single product crystallized that was found to be the desired
lactam 21. The lactam 21 was thus obtained in a modest yield
of 34% from the nitrile 19.
OH
OH
ButMe2SiO
ButMe2SiO
HO
HO
(c)
NH
NH
O
O
21
8
Scheme 8. (a) H3O+, EtOH, then Ac2O pyridine; (b) NaOMe,
MeOH; (c) H3O+, EtOH.
OH
OH
OH
HO
HO
HO
NH
NH
23
24
Finally, the lactam 21 was treated with 1% hydrochloric
acid in ethanol (Scheme 8).[22] The crude product was acety-
lated under standard conditions, for ease of purification, to
give the triacetate 22 of isofagomine lactam in 75% yield.
Deacetylation of the triacetate gave the desired isofagomine
lactam 8 in 89% yield. It was later found more convenient to
purify the lactam 8 directly after removal of the silyl protect-
ing groups. In this manner, the yield of isofagomine lactam
8 was 86% from the disilyl ether 21.
Scheme 9.
1
also able to locate the NH resonance for 8 in the H NMR
spectrum, and record a value for the optical rotation. Bols’
inhibition studies revealed isofagomine lactam to be a poor
inhibitor of almond β-glucosidase (Ki 29 µM)[14] relative to
isofagomine 23 (Ki 0.11 µM; Scheme 9).[23] This disappoint-
ing result stands in contrast to the potent inhibition observed
for the two xylobiose-derived analogues 4 and 5 (Scheme 2)
for the xylanase (Cex) from Cellulomonas fimi,[6,7] and for
Our synthetic material matched well with the isofagomine
lactam 8 produced by Bols and co-workers;[14] we were
OSiMe2But
O
OSiMe2But
O
HO
O
(a)
(b)
NC
NC
NC
OH
ButMe2SiO
ButMe2SiO
OBn
OH
OBn
16
17
11
(c)
(d)
CO2Me
OSiMe2But
O
ButMe2SiO
NC
ButMe2SiO
OSiMe2But
O
CN
CH2OH
18
19
Scheme 6. (a) ButMe2SiCl, imidazole, DMF; (b) Pd/C, H2, AcOH, H2O, THF; (c) DMSO, (COCl)2, CH2Cl2,
and then Et3N; (d) PDC, 4-Å sieves, CH2Cl2, and then PTSA, MeOH.
CO2Me
CO2Me
OH
ButMe2SiO
ButMe2SiO
(a)
(b)
ButMe2SiO
ButMe2SiO
OSiMe2But
OSiMe2But
CH2NH3Cl
CH2OH
NH
CN
O
CH2OH
19
20
21
Scheme 7. (a) Me2SBH3, THF, and then H3O+, MeOH; (b) Amberlite IRA 400 (OH−).

452
J. M. Macdonald and R. V. Stick
the two galactose-derived analogues 24 and 9 (Scheme 3) for
the β-galactosidase from Aspergillus oryzae.[12,24]
triacetate 15 (37 mg, 17%) as a colourless oil, [α]D +33.2◦ (Found: C
49.7, H 5.7, N 4.2, m/z 316.1041. C13H17NO8 requires C 49.5, H 5.4,
N 4.4%, [M + H]+ 316.1032). δH (500 MHz) 2.08, 2.11, 2.33 (9H, 3×s,
CH3), 3.40 (dt, J3,4 8.7, J4,5≈J4,5 4.9, H4), 3.74 (s, OCH3), 4.22–4.28
(2H, m, H5), 5.38 (d, J2,3 2.3, H2), 5.61 (dd, H3). δC (125.8 MHz) 20.26,
20.31, 20.45 (3×C, CH3), 32.77 (C4), 52.93 (OCH3), 59.44 (C5), 67.52
(C3), 70.83 (C2), 115.82 (CN), 166.61, 169.09, 169.54, 170.05 (4C, C1,
CH3CO).
Experimental
General experimental procedures have been given pre-
viously.[25]
Benzyl 4-Cyano-4-deoxy-α-D-arabinoside 11
Pyridinium p-toluenesulfonate (25 mg) was added to benzyl 4-
cyano-4-deoxy-2,3-O-isopropylidene-α-d-arabinoside 10[15] (1.51 g) in
MeOH/H2O (2 : 1, 30 mL) and the mixture concentrated to 20 mL
by gentle distillation (1 atm, 5 h). The reaction mixture was treated
with resin (Amberlite IRA 400, OH−), filtered, and then concentrated
(co-evaporating with toluene) to afford a colourless solid. Flash chro-
matography (EtOAc/petrol, 2 : 3 and then 4 : 1) of this solid gave the diol
11 (1.08 g, 83%) as a colourless solid.A small portion was recrystallized
to give colourless micro-needles, mp 138–139◦C (CH2Cl2/petrol), [α]D
+12.3◦ (MeOH) (Found: C 63.0, H 6.1, N 5.7. C13H15NO4 requires C
62.6, H 6.1, N 5.7%). δH (500 MHz) 3.25 (dt, J3,4≈J4,5 3.9, J4,5 6.2,
H4), 3.69 (dd, J5,5 12.1, H5), 3.76 (dd, J1,2 4.9, J2,3 6.8, H2), 3.90 (dd,
H3), 4.17 (dd, H5), 4.51 (d, H1), 4.57, 4.87 (ABq, J 11.6, PhCH2), 7.32–
7.40 (m, Ph). δC (125.8 MHz) 32.74 (C4), 59.07 (C5), 68.95, 69.96 (C2,
C3), 70.72 (PhCH2), 100.26 (C1), 117.60 (CN), 128.28, 128.43, 128.70,
136.11 (Ph).
Benzyl 2,3-Di-O-(tert-butyldimethylsilyl)-4-cyano-4-deoxy-α-D-
arabinoside 16
The diol 11 (650 mg, 2.6 mmol), ButMe2SiCl (1.96 g, 13.0 mmol), and
imidazole (1.80 g, 26.0 mmol) were heated (100◦C, 5 days) in DMF
(10 mL). The mixture was concentrated and subjected to a normal
work-up (EtOAc). Flash chromatography (EtOAc/petrol, 3 : 97 and then
1 : 9) gave the disilyl ether 16 (1.03 g, 83%) as a colourless solid. A
small portion was recrystallized to give colourless prisms, mp 65–67◦C
(MeOH), [α]D +78.8◦ (Found: C 62.9, H 8.8, N 2.8, m/z 478.2795.
C25H43NO4Si2 requires C 62.8, H 9.1, N 2.9%, [M + H]+ 478.2809).
δH (500 MHz) 0.01, 0.04, 0.10, 0.16 (12H, 4×s, SiCH3), 0.86, 0.88
(18H, s, CCH3), 3.27 (ddd, J3,4 2.6, J4,5 4.0, 10.8, H4), 3.59–3.64 (m,
2H), 3.92–3.95 (m, 1H), 4.27 (t, J5,5 10.8, H5), 4.45, 4.71 (ABq, J 11.9,
PhCH2), 4.53 (br s, H1), 7.26–7.35 (m, Ph). δC (125.8 MHz) −5.01,
−4.99, −4.89, −4.44 (4×C, SiCH3), 17.86, 18.00 (2×C, CCH3), 25.60,
25.63 (6×C, CCH3), 30.98 (C4), 55.44 (C5), 68.90, 69.01 (C2,3), 69.62
(PhCH2), 99.06(C1), 118.56(CN), 127.68, 128.14, 128.25, 137.39(Ph).
Benzyl 2,3-Di-O-acetyl-4-cyano-4-deoxy-α-D-arabinoside 12
The diol 11 (580 mg, 2.33 mmol), Ac2O (1.75 mL, 18.6 mmol), and
4-dimethylaminopyridine (DMAP) (10 mg) were stirred overnight in
pyridine (10 mL). MeOH (4 mL) was added and, after a further 30 min,
the mixture was concentrated and then subjected to normal work-up
(EtOAc) to give a pale yellow oil. Flash chromatography (EtOAc/petrol,
3 : 7 and then 2 : 3) of this oil gave the diacetate 12 (750 mg, 97%) as a
colourless oil, [α]D +52.8◦ (Found: C 61.6, H 5.7, N 4.0, m/z 334.1307.
C17H19NO6 requires C 61.2, H 5.7, N 4.2%, [M + H]+ 334.1291). δH
(500 MHz) 2.08, 2.09 (2s, 6H, CH3), 3.39 (quintet, J3,4≈J4,5 3.8, J4,5
7.7, H4), 3.73 (dd, J5,5 11.7, H5), 4.26 (dd, H5), 4.53, 4.81 (ABq, J
12.0, PhCH2), 4.65 (d, J1,2 3.7, H1), 5.03 (dd, J2,3 5.8, H3), 5.08 (dd,
H2), 7.28–7.38 (m, Ph). δC (125.8 MHz) 20.57, 20.64 (2C, CH3), 30.34
(C4), 57.85 (C5), 66.57, 66.71 (C2, C3), 69.95 (PhCH2), 97.27 (C1),
116.15 (CN), 127.47, 127.92, 128.40, 136.72 (Ph), 168.84, 169.72 (2C,
CH3CO).
2,3-Di-O-(tert-butyldimethylsilyl)-4-cyano-4-deoxy-D-arabinose 17
The benzyl glycoside 16 (370 mg), Pd/C (140 mg of 10%), and AcOH
(100 µL) were stirred vigorously in THF (15 mL) under an atmosphere
of H2 (1 atm, 3 days). The mixture was filtered through Celite, wash-
ing with EtOAc, and the combined filtrate/washings were concentrated.
Flash chromatography (EtOAc/petrol, 1 : 24 and then 1 : 10) of the
residue gave the hemiacetal 17 (280 mg, 94%) as an oil, [α]D −9.1◦
(Found: C 56.0, H 9.5, N 3.5, m/z 388.2365. C18H37NO4Si2 requires
C 55.8, H 9.6, N 3.6%, [M + H]+ 388.2339). δC (75.5 MHz) −4.97,
−4.92, −4.89, −4.83, −4.57, −4.53 (4×C, SiCH3), 17.93 (2×C,
CCH3), 25.62, 25.65 (6×C, CCH3), 29.54, 30.48 (C4), 54.31, 60.65
(C5), 67.21, 69.83, 70.10, 70.35 (C2, C3), 91.42, 94.51 (C1), 117.92,
118.03 (CN).
2,3-Di-O-(tert-butyldimethylsilyl)-4-cyano-4-deoxy-D-
arabinono-1,5-lactone 18
2,3-Di-O-acetyl-4-cyano-4-deoxy-D-arabinose 13
DMSO (25 µL, 0.35 mmol) was added dropwise to (COCl)2 (30 µL,
0.32 mmol) in CH2Cl2 (4 mL) at −78◦C. After 3 min, the hemiacetal
17 (50 mg, 0.13 mmol) in CH2Cl2 (4 mL) was added dropwise and the
mixture stirred (30 min, −78◦C). Et3N (200 µL) was added and the
mixture was allowed to warm to room temperature. The mixture was
washed with H2O (twice), and brine, and then dried (MgSO4), filtered,
and concentrated. Flash chromatography (EtOAc/petrol, 3 : 97 and then
3 : 22) of the residue gave the lactone 18 (24 mg, 48%) as a colourless
solid. A small portion was recrystallized to give needles, mp 64–67◦C
(petrol), [α]D −47.3◦ (Found: C 56.4, H 9.0, N 3.4, m/z 386.2174.
C18H35NO4Si2 requires C 56.1, H 9.1, N 3.6%, [M + H]+ 386.2183).
δH (300 MHz) 0.13, 0.15, 0.16, 0.22 (12H, 3×s, SiCH3), 0.89, 0.92
(18H, s, CCH3), 3.59 (ddd, J3,4 1.9, J4,5 5.9, 11.2, H4), 4.04 (d, J2,3
3.8, H2), 4.16 (dd, H3), 4.49 (dd, J5,5 11.0, H5), 4.62 (t, H5). δC (75.5
MHz) −5.36, −5.07, −4.90, −4.78 (4×C, SiCH3), 17.91, 17.99 (2×C,
CCH3), 25.52 (6×C, CCH3), 28.29 (C4), 64.30 (C5), 70.54, 70.75 (C2,
C3), 116.75 (CN), 166.83 (C1).
The diacetate 12 (570 mg), Pd/C (110 mg of 10%), and aqueous AcOH
(0.2 mL of 9 M) were stirred vigorously in THF (20 mL) under an atmo-
sphere of H2 (1 atm, overnight).The mixture was filtered through Celite,
washing with EtOAc, and the combined filtrate/washings were concen-
trated. Flash chromatography (EtOAc/petrol, 3 : 7 and then 3 : 2) of the
residue gave the hemiacetal 13 (300 mg, 71%) as a colourless oil, [α]D
−176.5◦. δC (125.8 MHz) 20.66, 20.70, 20.72, 20.74 (2C, CH3), 32.42,
33.47 (C4), 58.33, 60.60 (C5), 65.59, 67.88, 68.98, 70.19 (C2, C3),
90.88, 95.35 (C1), 116.26, 116.73 (CN), 169.84, 169.88, 169.96, 170.32
(2C, CH3CO).
Methyl 2,3,5-Tri-O-acetyl-4-cyano-4-deoxy-D-arabinonate 15
The hemiacetal 13 (170 mg, 0.51 mmol), pyridinium dichromate
(960 mg, 2.6 mmol), and crushed molecular sieves (4 Å, 1.2 g) were
stirred overnight in CH2Cl2 (15 mL). The mixture was filtered through
Celite, washing with CH2Cl2, and the combined filtrate/washings were
washed with H2O (twice), dried (MgSO4), filtered, and concentrated.
The residue was taken up in MeOH (10 mL) and treated with p-
toluenesulfonic acid (5 mg). After 2 h the mixture was neutralized with
resin (Amberlite IRA 400, OH−), filtered, and then concentrated. The
residue was taken up in pyridine (2 mL) and treated withAc2O (250 µL,
2.6 mmol) for 2 h. MeOH (1 mL) was added and, after 30 min, the mix-
ture was concentrated and then subjected to normal work-up (EtOAc).
Flash chromatography (EtOAc/petrol, 4 : 7 and then 1 : 1) then gave the
Methyl 2,3-Di-O-(tert-butyldimethylsilyl)-4-cyano-4-
deoxy-D-arabinonate 19
The hemiacetal 17 (260 mg, 0.55 mmol), pyridinium dichromate (1.0 g,
2.8 mmol), and crushed molecular sieves (4 Å, 1.2 g) were stirred in
CH2Cl2 (15 mL) for 30 h. The mixture was filtered through Celite,
washing with CH2Cl2, and the combined filtrate/washings were washed

The Synthesis of a d-Glucose-like Piperidin-2-one
453
with H2O (twice), and brine, dried (MgSO4), filtered, and concen-
trated. The residue was taken up in MeOH (15 mL) and treated with
p-toluenesulfonic acid (5 mg). After 24 h the mixture was neutralized
with resin (Amberlite IRA 400, OH−), filtered, and then concentrated.
Flash chromatography (EtOAc/petrol, 1 : 19 and then 3 : 17) of the
residue gave the methyl ester 19 (230 mg, 80%) as a colourless oil, [α]D
+15.0◦ (Found: C 54.7, H 9.1, N 3.2, m/z 418.2429. C19H39NO5Si2
requires C 54.6, H 9.4, N 3.3%, [M + H]+ 418.2445). δH (500 MHz)
0.06, 0.10, 0.13 (12H, 3×s, SiCH3), 0.88, 0.93 (18H, s, CCH3), 2.42 (br
s, OH), 3.17 (dt, J3,4 7.4, J4,5≈J4,5 4.8, H4), 3.47 (s, OCH3), 3.85 (dd,
J5,5 11.4, H5), 3.92 (dd, H5), 4.31 (dd, J2,3 3.4, H3), 4.38 (d, H2). δC
(125.8 MHz) −5.31, −5.06, −4.86, −4.82 (4×C, SiCH3), 17.97, 18.21
(2×C, CCH3), 25.56, 25.65 (6×C, CCH3), 37.60 (C4), 51.96 (OCH3),
59.46 (C5), 71.59, 73.78 (C2, C3), 119.31 (CN), 171.46 (C1).
Method 2: A small piece of Na was added to the triacetate 22
(10 mg) in dry MeOH (3 mL). After 2 h the mixture was neutralized
withresin(AmberliteIR-120, H+), filtered, andthenconcentrated. Flash
chromatography (EtOAc/MeOH/H2O, 15 : 2 : 1 and then 7 : 2 : 1) of the
residue gave the lactam 8 (5.0 mg, 89%) as a colourless, amorphous
solid, consistent, by 1H (500 MHz) and 13C (125.8 MHz) NMR spectro-
scopy, with the material prepared above.
Acknowledgment
We thank the Australian Research Council for financial
assistance.
References
[1] T. Kajimoto, K. K.-C. Liu, R. L. Pederson, Z. Zhong,Y. Ichikawa,
J. A. Porco, C.-H. Wong, J. Am. Chem. Soc. 1991, 113, 6187.
[2] G. Pistia, R. I. Hollingsworth, Carbohydr. Res. 2000, 328, 467.
doi:10.1016/S0008-6215(00)00124-5
[3] Y. Nishimura, H. Adachi, T. Satoh, E. Shitara, H. Nakamura,
F. Kojima, T. Takeuchi, J. Org. Chem. 2000, 65, 4871.
doi:10.1021/JO000141J
[4] B. Ganem, G. Papandreou, J. Am. Chem. Soc. 1991, 113, 8984.
[5] A. Varrot, C. A. Tarling, J. M. Macdonald, R. V. Stick,
D. L. Zechel, S. G. Withers, G. J. Davies, J. Am. Chem.
Soc. 2003, 125, 7496. doi:10.1021/JA034917K
[6] S. J. Williams, R. Hoos, S. G. Withers, J. Am. Chem. Soc. 2000,
122, 2223. doi:10.1021/JA993805J
[7] S. J. Williams, V. Notenboom, J. Wicki, D. R. Rose, S. G. Withers,
J. Am. Chem. Soc. 2000, 122, 4229. doi:10.1021/JA0002870
[8] M. N. Namchuk, S. G. Withers, Biochemistry 1995, 34, 16194.
[9] A. White, D. Tull, K. Johns, S. G. Withers, D. R. Rose, Nature
Struct. Biol. 1996, 3, 149.
[10] V. Notenboom, C. Birsan, M. Nitz, D. R. Rose, R. A. J. Warren,
S. G. Withers, Nature Struct. Biol. 1998, 5, 812. doi:10.1038/
1852
(3S,4R,5R)-3,4-Di-(tert-butyldimethylsilyloxy)-5-
(hydroxymethyl)piperidin-2-one 21
The nitrile 19 (150 mg, 0.36 mmol) and Me2SBH3 (70 µL, 0.72 mmol)
were heated under reflux in THF (5 mL) for 1.5 h. The mixture was
allowed to cool, and was then brought to pH 5 by the cautious addi-
tion of hydrochloric acid in MeOH (0.1 M) and stirred (30 min). The
mixture was neutralized with resin (Amberlite IRA 400, OH−), fil-
tered, and then concentrated. The residue was subjected to normal
work-up (EtOAc) to afford an oil that crystallized. Recrystallization
gave the lactam 21 (50 mg, 36%) as colourless micro-needles, mp 202–
210◦C (EtOAc/petrol), [α]D −26.8◦. δH (500 MHz) 0.090, 0.097, 0.18,
0.20 (4×s, 12H, SiCH3), 0.88, 0.91 (2×s, 18H, CCH3), 2.12 (m, H5),
2.24 (br s, OH), 3.28 (dt, J5,6≈J6,NH 3.4, J6,6 12.0, H6), 3.65 (ddd,
J 1.7, 5.5, H6), 3.77–3.81 (2H, m), 3.85–3.87 (1H, m), 3.99 (1H, t,
J≈J 3.0), 5.49–5.54 (m, NH). δC (125.8 MHz) −5.29, −4.94, −4.75,
−4.70 (4×C, SiCH3), 17.93, 18.12 (2×C, CCH3), 25.67, 25.72 (6×C,
CCH3), 39.78 (C6), 41.09 (C5), 61.88 (CH2O), 72.36, 72.91 (C3, C4),
170.03 (C2). m/z (FAB) 390.2505 (C18H40NO4Si2 [M + H]+ requires
390.2496).
[11] T. Gloster, S. J. Williams, C. A. Tarling, S. Roberts, C. Dupont,
P. Jodoin, F. Shareck, S. G. Withers, G. J. Davies, Chem.
Commun. 2003, 944. doi:10.1039/B301829F
(3S,4R,5R)-3,4-Diacetoxy-5-(acetoxymethyl)piperidin-
2-one 22
[12] H. Sohoel, X. Liang, M. Bols, J. Chem. Soc., Perkin Trans. 1
2001, 1584. doi:10.1039/B103831C
[13] J. M. Macdonald, R. V. Stick, D. M. G. Tilbrook, 21st Int.
Carbohydrate Symposium 2002, 149. (Abstract.)
[14] V. H. Lillelund, H. Liu, X. Liang, H. Sohoel, M. Bols, Org.
Biomol. Chem. 2003, 1, 282. doi:10.1039/B208784G
[15] W. M. Best, J. M. Macdonald, B. W. Skelton, R. V. Stick,
D. M. G. Tilbrook, A. H. White, Can. J. Chem. 2002, 80, 857.
doi:10.1139/V02-060
[16] H. C. Brown, Y. M. Choi, S. Narasimhan, J. Org. Chem. 1982,
47, 3153.
[17] H. C. Brown, Y. M. Choi, S. Narasimhan, Synthesis 1981, 605.
doi:10.1016/0040-4039(95)02227-9
[18] T. Hosoya, Y. Ohashi, T. Matsumoto, K. Suzuki, Tetrahedron
Lett. 1996, 37, 663. doi:10.1016/0040-4039(95)02227-9
[19] H. Abe, S. Shuto, S. Tamura, A. Matsuda, Tetrahedron Lett.
2001, 42, 6159. doi:10.1016/S0040-4039(01)01233-3
[20] C. Walford, R. F. W. Jackson, N. H. Rees, W. Clegg, S. L. Heath,
Chem. Commun. 1997, 1855. doi:10.1039/A704534D
[21] H. Abe, S. Shuto, A. Matsuda, J. Am. Chem. Soc. 2001, 123,
11870. doi:10.1021/JA011321T
The lactam 21 (45 mg, 0.12 mmol) in hydrochloric acid (10 m)/EtOH
(1 : 99, 2 mL) was kept overnight. The mixture was concentrated
(co-evaporating with toluene), and then taken up in pyridine (2 mL)
and treated withAc2O (200 µL, 2.2 mmol) for 2 h. MeOH (0.5 mL) was
added and, after a further 30 min, the mixture was concentrated. Flash
chromatography (EtOAc/petrol, 4 : 1 followed by MeOH/CHCl3, 1 : 19)
of the residue gave the triacetate 22 (25 mg, 75%) as a colourless glass,
[α]D +6.2◦. δH (600 MHz) 2.06, 2.09, 2.13 (3×s, 9H, CH3), 2.42–2.44
(m, H5), 3.31–3.36 (m, H6), 3.43 (ddd, J 3.6, 5.0, 12.3, H6), 4.09 (1H, dd,
J5,H 3.5, JH,H 11.6, CH2O), 4.19 (1H, dd, J5,H 6.4, CH2O), 5.08 (d, J3,4
9.1, H3), 5.37 (dd, J4,5 10.7, H4), 6.09–6.14 (m, NH). δC (150.9 MHz)
20.61, 20.68 (3×C, CH3), 38.07 (C5), 41.07 (C6), 61.30 (CH2O), 69.21,
72.37 (C3, C4), 167.19, 169.81, 170.23, 170.55 (4×C, C=O). m/z (FAB)
288.1069 (C12H18NO7 [M + H]+ requires 288.1083).
(3S,4R,5R)-3,4-Dihydroxy-5-(hydroxymethyl)piperidin-
2-one (Isofagomine Lactam) 8
Method 1: The disilyl ether 21 (14 mg) in hydrochloric acid
(10 M)/EtOH (1 : 99, 2 mL) was heated to 50◦C for 1.5 h. The mixture
was concentrated (co-evaporating with toluene). Flash chromatography
(EtOAc/MeOH/H2O, 15 : 2 : 1 and then 7 : 2 : 1) of the residue gave the
lactam 8 (5.0 mg, 86%) as a colourless, amorphous solid, [α]D +11.0◦
(MeOH). δH (600 MHz, D2O) 2.12–2.18 (m, H5), 3.16 (ddd, J6,NH 0.8,
J5,6 10.3, J6,6 12.8, H6), 3.40 (dd, J5,6 5.8, H6), 3.67–3.72 (2H, m, H4,
CH2O), 3.82 (1H, dd, J5,H 3.8, JH,H 11.5, CH2O), 4.03 (dd, J3,NH 0.8,
J3,4 9.3, H3), 8.45 (s, NH).[14] δC (150.9 MHz, D2O) 41.15 (C5), 41.24
(C6), 60.89 (CH2O), 71.24 (C4), 73.39 (C3), 174.25 (C2).[14] m/z (EI)
[22] H. Wetter, K. Oertle, Tetrahedron Lett. 1985, 26, 5515.
doi:10.1016/S0040-4039(01)80874-1
[23] T. M. Jespersen, W. Dong, M. R. Sierks, T. Skrydstrup, I. Lundt,
M. Bols, Angew. Chem. Int. Ed. Engl. 1994, 33, 1778.
[24] Y. Ichikawa, Y. Igarashi, Tetrahedron Lett. 1995, 36, 4585.
doi:10.1016/0040-4039(95)00870-I
[25] J. C. McAuliffe, R. V. Stick, Aust. J. Chem. 1997, 50, 193.
doi:10.1071/CH96151
•
161.0694 (C6H11NO4 [M]+ requires 161.0688).