Iminosugars: Potential inhibitors of liver glycogen phosphorylase

Article abstract of DOI:10.1016/S0968-0896(00)00291-1

The first synthesis of the single isomers 3R, 4R, 5R); 3S,4S,5S); 3R,4R,5S) and 3S,4S, 5R) of 5-hydroxymethyl-piperidine-3,4-diol from Arecolin is reported, including the synthesis of a series of N-substituted derivatives of the 3R,4R,5R)-isomer Isofagomi

Full text of DOI:10.1016/S0968-0896(00)00291-1

Bioorganic & Medicinal Chemistry 9 (2001) 733±744
Iminosugars: Potential Inhibitors of Liver Glycogen Phosphorylase
Palle Jakobsen,a,Ã Jane M. Lundbeck,a Marit Kristiansen,a Jens Breinholt,b
Helle Demuth,b Jan Pawlas,a Maria P. Torres Candela,a Birgitte Andersen,c
Niels Westergaard,c Karsten Lundgrenc,y and Naoki Asanod
aMedicinal Chemistry Research, Novo Nordisk A/S, Health Care Discovery, Novo Nordisk Park, DK-2760 Maaloev, Denmark
bPharmaceutical Chemistry, Novo Nordisk A/S, Health Care Discovery, Novo Nordisk Park, DK-2760 Maaloev, Denmark
cDiabetes Biochemistry & Metabolism, Novo Nordisk A/S, Health Care Discovery, Novo Nordisk Park, DK-2760 Maaloev, Denmark
dHokuriku University, Faculty of Pharmaceutical Sciences, Kanazawa-920-11, Japan
Received 26 June 2000; accepted 31 October 2000
AbstractÐThe ®rst synthesis of the single isomers (3R,4R,5R); (3S,4S,5S); (3R,4R,5S) and (3S,4S,5R) of 5-hydroxymethyl-piper-
idine-3,4-diol from Arecolin is reported, including the synthesis of a series of N-substituted derivatives of the (3R,4R,5R)-isomer
(Isofagomine). The inhibitory e€ect of these isomers as well as of a series of N-substituted derivatives of the (3R,4R,5R)-isomer and
selected hydroxypiperidine analogues on liver glycogen phosphorylase (GP) showed that the (3R,4R,5R) con®guration was essential
for obtaining an inhibitory e€ect at submicromolar concentration. The results also showed that all three hydroxy groups should be
present and could not be substituted, nor were extra OH groups allowed if sub-micromolar inhibition should be obtained. Some
inhibitory e€ect was retained for N-substituted derivatives of Isofagomine; however, N-substitution always resulted in a loss of
activity compared to the parent compound, IC50 values ranging from 1 to 100 mM were obtained for simple alkyl, arylalkyl and
benzoylmethyl substituents. Furthermore, we found that it was not enough to assure inhibitory e€ect to have the (R,R,R) con®g-
uration. Fagomine, the (2R,3R,4R)-2-hydroxymethylpiperidine-3,4-diol analogue, showed an IC50 value of 200 mM compared to
0.7 mM for Isofagomine. In addition, Isofagomine was able to prevent basal and glucagon stimulated glycogen degradation in cul-
tured hepatocytes with IC50 values of 2±3mM. # 2001 Elsevier Science Ltd. All rights reserved.
Introduction
development of new anti-hyperglycemic agents. Since
the discovery8 of NN 42-1007 (Isofagomine) as a potent
liver glycogen phosphorylase (GP) inhibitor, we have
tested several other iminosugars for their ability to
inhibit the GP enzyme (Fig. 1).
Iminosugars have attracted considerable interest due to
their often e€ective and speci®c inhibition of various
glycosidases involved in the intestinal degradation of
carbohydrates.1À3 However, for the treatment of dia-
betes this e€ect is weak and associated with side e€ects. It
is generally recognised that the hepatic glucose output in
type 2 diabetes is elevated and thus signi®cantly con-
tributes to hyperglycemia.4À6 Glucose is produced by the
liver from two di€erent pathways: gluconeogeneis (de
novo synthesis of glucose) and from glycogenolysis
(breakdown of glycogen). However, it has been reported7
that a substantial portion of the glucose produced by
gluconeogenesis is cycled through the glycogen storage
before e‚ux from the liver. This makes inhibition of
the glycogenolysis a bene®cial target to attack in the
We here report on the SAR of a series of Isofagomine
analogues together with a new synthetic approach to
Isofagomine isomers starting from Arecolin.
Chemistry
Isofagomine has been prepared from d-lyxose in a
multistep synthesis giving an overall yield of 11%,1 and
from 1,6 dianhydro-beta-d-glucopyranose yielding
10%.9 Recently, a possible intermediate in the synthesis
of Isofagomine isomers was prepared from a piperidone
carboxylate derivative, and some isomers were prepared
in a synthesis sequence similar to the one reported
here.10,11 An Isofagomine N-substituted with an extra
glucose residue has been reported.12
ÃCorresponding author.
yPresent address: TransTech Pharma, 4170 Medenhall Oaks Pkwy, Suite
110, High Point, NC 27265, USA.
0968-0896/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(00)00291-1

734
P. Jakobsen et al. / Bioorg. Med. Chem. 9 (2001) 733±744
The reaction pathway from Arecolin to a mixture of
Isofagomine isomers is depicted in Scheme 1. Arecolin 1
was demethylated by means of 1-chloroethyl chloro-
formate13 followed by N-protection with Boc anhydride
giving 2.
Subsequent isomerisation using the method described
by Allan and Fong14 gave the isomerised product 3 in
good yield. Reduction by means of sodium borohydride
or lithium triethyl borohydride gave the unsaturated
carbinol 4a which was epoxidised by means of 3-
chloroperbenzoic acid to give 5 as an isomer mixture,
ring opening and deprotection proceeded by re¯ux in
KOH solution (10%) to give the isomeric mixture 6.
Figure 1.
In addition to the synthetic approach towards Iso-
fagomine, we have also investigated a solid phase
method for the preparation of N-alkylated derivatives of
Isofagomine.
In order to prepare the single isomers of 6 we used two
approaches; puri®cation of the isomer mixture 6 on chiral
plate followed by HPLC separation (Scheme 2), and
Scheme 1. Synthetic pathway from Arecolin to Isofagomine (compounds 4a, 5 and 6 were isomer mixtures): (i), (a) ClCOOCHClMe, toluene; (b)
MeOH; (c) (Boc)2O, Et3N, MeOH; (ii) LDA,THF, À78 ꢀC; (iii) LiEt3BH,THF; (iv) MCPBA, CH2Cl2, RT; (v) KOH (10%), re¯ux.
Scheme 2. Synthetic pathway to enantiomeric mixtures of Isofagomine: (a) P=H; (b) P=(À)-menthyloxyacetyl; (i) MCPBA, CH2Cl2, rt; (ii) (a)
KOH (10%) re¯ux; (b) puri®cation on chiral plate (Merchery Nagel), eluent PriOH/25% NH3 (3/1); (c) separation on HPLC, Merck LiChrosorb
NH2, eluent 70% CH3 CNaq; identi®cation by MS and NMR.

P. Jakobsen et al. / Bioorg. Med. Chem. 9 (2001) 733±744
735
Scheme 3. Synthetic pathway to four stereoisomers of Isofagomine: (i) Chiracel OD or chiralpack AS, eluent PriOH/heptane (1/99); (ii) MCPBA,
CH2Cl2, rt; (iii) KOH (10%).
separation of the enantiomers 4a by chiral HPLC fol-
lowed by further reaction as described above (Scheme 3).
In order to separate at an earlier step in the synthetic
sequence, the unsaturated carbinol 4a was separated in
its enantiomers on a chiral column and the following
reaction steps were performed on the pure enantiomers
leading to diastereomeric mixtures which were separated
by normal phase chromatography.
Separation of the isomer mixture of 6 as depicted in
Scheme 2 was attempted with both an unprotected OH
group and with a (À)-menthyloxyacetyl group, the latter
in an attempt to achieve a greater di€erence in the chro-
matographic parameters. The epoxy compounds 5a and
5b were isolated as isomer mixtures and the subsequent
step (ring opening and deprotection) resulted in a mixture
of compounds 7, 8, 9, and 10. This mixture was subse-
quently separated on chiral plate followed by HPLC
separation. The structure elucidation of the resulting
enantiomer mixtures 7,8 and 9,10 showed that Iso-
fagomine and its enantiomer (7,8) were the major products
from the reaction sequence epoxidation/ring opening of
4a, while the Isofagomine diastereomers (9,10) domi-
nated from 4b.
The yield in the separation step was found to 55%. No
attempt was carried out to increase the yield. The
enantiomer ratio was found to 1:3 the di€erence from
1:1 probably due to incomplete recovery, the ee of the
enantiomers was >98%. Intermediates 5c and 5d were
not isolated.
The product mixtures 7,9 and 8,10 were puri®ed on a
short silica-gel column using PriOH/NH3 aq as eluent,
and the mixture was isolated as the hydrochloride salt
for further separation by HPLC on a Merck LiChro-
sorb NH2 7 mm 25Â250 mm column. 7 was collected
between 6.75 and 7.5 min while 9 was eluted between 5
and 6 min. Their enantiomers 8 and 10 were collected
after 6.25±7 and 5±5.75 min, respectively; the enantio-
meric pair 7,8 thus being the isomer with the longest
retention time.
This di€erence is explainable from the directing e€ect of
the hydroxyl with respect the menthyloxyacetyl group. No
attempt was made to explore this directing e€ect in the
epoxidation and ring opening step further. However, the
NMR spectrum of 4a indicated a preferred conformation
in which one of the CH2OH methylene protons strongly
interacted with the hydrogens at the 2-methylene group.
Con®rmation of relative stereochemistry of 7,8 and 9,10
was carried out by a NMR analysis.
The identity of the actual enantiomeric mixture formed
was estimated by analysis of the NMR spectra, the 7,8
mixture being identical to Isofagomine, and the other
mixture elucidated to be the Isofagomine diastereomers
9,10.
Protons are denoted H-n (a or e) where n refers to the
position labels. The methylene protons at C-2 and C-6
are tagged with an `e' or an `a' denoting equatorially-
1
oriented and axially-oriented H-nuclei, respectively.

736
P. Jakobsen et al. / Bioorg. Med. Chem. 9 (2001) 733±744
The relative con®guration of the two enantiomers, 7 and
1
using standard N-alkylation methods like alkyl halo-
genid and base, or by reductive amination using alde-
hyde and cyanoborohydride (Scheme 4).
8, was con®rmed by the observed vicinal H-1H-coupl-
ing constants (3JHH). Thus, large coupling constants
attributable to trans-diaxial interaction between H-6a
and H-5 (J6a,5=12.0 Hz) and between H-3 and H-2a
(J3,2a=11.3 Hz) demonstrate that both H-3 and H-5
adapt axial positions. Furthermore, large coupling con-
An alternative route was using the method described for
Isofagomine9 replacing ammonia with RNH2.
stants between H-4 and both its neighbours, H-5 (J5,4
=
In order to prepare a large number of N-substituted
compounds for a SAR investigation, we modi®ed the
solid phase method developed by Brown et al.24
(Scheme 5).
10.5 Hz) and H-3 (J4,3=9.0 Hz), provide evidence for the
trans-relation to both H-3 and H-5 in accordance with
the depicted stereochemistry (7 and 8). The optical rota-
tion of 7 was in accordance with the literature value.9
Wang resin was reacted with vinyl chloroformate and
subsequently with Isofagomine. The thereby formed
solid phase bound Isofagomine was reacted with di€er-
ently activated alkyl- or aryl-bromides, forming the
quaternary ammonium compound. The desired N-sub-
stituted Isofagomine was liberated from the resin by
treatment with DIPEA. Yields varied considerably (20±
100%) and it was dicult to get rid of the DIPEA-salt
and all the DMF at ®rst, so HPLC puri®cation was
required in order to isolate the pure compounds. Proof
of identity and yields were obtained form LC/MS and
NMR measurements.
Similarly, the relative con®guration of the enantiomeric
1
pair, 9 and 10, was con®rmed by analysing the H±1H-
coupling pattern. A large trans-diaxial coupling between
H-6a and H-5 (J6a,5=12.7 Hz) proved that H-5 is in the
axial position, and a small coupling between H-5 and
H-4 (J5,4=2.9 Hz) places H-4 in the equatorial position.
Finally, small coupling constants between H-3 and both
H-2a and H-2e (J3,2a=2.3 and J3,2e=2.1 Hz) place H-3
equatorially, and conform with the expected stereo-
chemistry (9 and 10).
The calculated 1H data for the enantiomers were in accor-
dance with the experimental values (data not shown).
This method, although not optimised, opened for a
variety of N-substituted derivatives for the SAR inves-
tigation. A representative selection of the approximately
100 compounds prepared is presented in Tables 1 and 2.
A series of N-alkylated Isofagomine derivatives were pre-
pared by N-alkylation of 3-O benzylated Isofagomine9
Scheme 4. Synthesis of N-substituted Isofagomine derivatives: (i) R-I, K2CO3, acetone or RCHO, NaCNBH3, pH=6, MeOH; (ii) Pd/C in EtOH/
HCl (aq); Yields varying from 25 to 95%.
Scheme 5. Solid phase synthesis of N-substituted Isofagomine derivatives: (i) CH2Cl2, 7H, rt; (ii): DMF, 22 h, rt, wash; (iii) (a) R-Hal, DMF, 24 h, rt
(b) DIPEA, DMF, 24 h, rt.

P. Jakobsen et al. / Bioorg. Med. Chem. 9 (2001) 733±744
737
Table 1. Inhibition of pig liver glycogen phosphorylase (IC50 in mM)
for N-substituted 7 (Isofagomine)
Table 2. Inhibition of pig liver glycogen phosphorylase for ring-sub-
stituted N-benzoylmethyl Isofagomine derivatives (IC50 in mM)
Compound no.
N-substituent
IC50 (mM)
Compound no.
Phenyl substituent
IC50 (mM)
7
-H
-Me
-Bu
-Allyl
0.7
24
25
26
27
28
29
30
31
32
33
34
4-Br
4-H
3,4-diF
3-CH3,4-Cl
4-Ph
3-OCH3
4-OCH3
2,4-di-OCH3
4-N(CH2CH3)2
4-OCF3
3-F
40
7
8
27
70
8
10
11
12
13
14
15
16
17
18
19
20
21
22
23
15
20
4
-Propyn-3-yl
-1-Dodecyl
-Ac
-CH2CH2COOH
-Bn
-CH2CH2Ph
-NO2PhCH2CH2
-CH2CH2CH2-Ph
Cyclohexylprop-3-yl
-CH2CHˆCHPh
>200
8
60
15
46
75
16
1
>100
>100
80
6
2
8
Scheme 6. Inhibition of rat or pig liver glycogen phosphorylase (IC50 in mM) for selected Isofagomine analogues.

738
P. Jakobsen et al. / Bioorg. Med. Chem. 9 (2001) 733±744
Table 3. IC50 values (nM) for inhibition of glycogen phosphorylase
by Isofagominea
GPa
GPb
Rat liver
Pig liver
Rabbit muscle
697Æ85 (4)
773Æ9 (5)
718Æ39 (3)
n.d.b
n.d.
1215Æ180 (4)
aResults are means of the numbers in parentheses ÆSEM. The Hill coef-
®cients were calculated to be GPa rat liver: 1.10Æ0.05; GPa pig liver
1.07Æ0.03; GPa and GPb rabbit muscle: 1.07Æ0.14 and 1.13Æ0.06,
respectively. No statistical signi®cance among the groups was observed.
an.d.=not determined)
Biology
Data for inhibition of liver GP for a series of close
analogues of Isofagomines including some compounds
previously prepared or isolated from plant material are
summarised in Scheme 6.
The data depicted in Scheme 6 show a surprising special
structural requirements for inhibitory activity of this
type of iminosugars as only compounds 7 and 8 showed
inhibitory e€ect in the low mM range, and compound 7
was the only one giving an IC50 value below 1 mM.
Tables 1 and 2 show the inhibition data for N-substituted
Isofagomine derivatives. These data show that some
N-substitution is allowed. However, no N-substituted
derivatives synthesised exhibit a better inhibitory e€ect
on glycogen phosphorylase enzyme than Isofagomine.
Figure 2. E€ect of Isofagomine on basal and glucagon-stimulated
glycogenolysis. (A) Basal (*) and glucagon-induced (0.5 nM gluca-
gon) (&) glucose production were measured over 70 min (basal glu-
cose production at 0 mM Isofagomine corresponds to 100%). Results
are meansÆSEM (n=4). (B) After 70 min of glycogenolysis the glyco-
gen levels were determined: glycogen level after basal glycogenolysis
(*) and glycogen levels after glucagon-induced glycogenolysis (0.5 nM
glucagon) (&) (glycogen level before glycogenolysis corresponds to
100%). Results are meansÆSEM (n=3).
Compounds 14, 23, 31, and 32 prepared by parallel
synthesis on Wang-resin were tested as 50% solutions,
the other components being DIPEA salt and DMF as
seen from the NMR spectra.
The test value was corrected for the concentration
<100%. However, the DIPEA salt/DMF mixture was
shown not to inhibit the glycogen phosphorylase enzyme.
The inhibitory e€ect of the most potent compound Iso-
fagomine (7) was furthermore characterised on GP
obtained from various tissues (Table 3). Also, the ability
of Isofagomine to inhibit glycogenolysis in cultured
primary hepatocytes was investigated, with respect to
inhibition of glucose production and the ability to retain
glycogen levels. In hepatocytes, Isofagomine inhibited
both basal and glucagon-induced glucose production
dose-dependently with IC50 values of 3.0Æ0.4 and
2.0Æ0.5 mM, respectively (Fig. 2). In addition, 80% of
the glycogen level could be retained at the highest dose
of Isofagomine.
on that matter. The structural requirements of the
N-unsubstituted Isofagomine-like iminosugars for inhi-
bition of liver glycogen phosphorylase clearly show that
inhibitory e€ect in the low mM range can be achieved
only when the compounds contain the 3R,4R,5R- or the
3S,4S,5S-con®guration, as it appears from Scheme 6.
Furthermore, no extra OH group seems to be allowed
(compounds 47, 49 and 50), nor is it allowed to sub-
stitute on the OH groups (35). It is surprising that nei-
ther Fagomine (37) nor Deoxynojirimycin (49) were
found to be potent inhibitors of GP, bearing in mind
that the ®ve-ring analogue DAB was reported to show
strong inhibition of glycogen phosphorylase15,25 indi-
cating that structures with the three hydroxy groups
closer to the nitrogen are allowed, but obviously only
when the ring size is diminished
Conclusion
The present investigation presents the ®rst comparative
data for inhibition of liver glycogen phosphorylase for
six-membered iminosugars.
It is furthermore clear that all three OH groups should be
present, compounds 42, 43 and 44 showing no activity.
Substitution at other positions of the piperidine ring
seems not to be allowed.
The synthetic methodology described gives the ®rst
synthesis of the pure enantiomer of Isofagomine (8) and
add new data to the increasing set of literature reports

P. Jakobsen et al. / Bioorg. Med. Chem. 9 (2001) 733±744
739
This strict requirement might indicate that the com-
pounds bind in the catalytic site in the same manner as
the natural ligand, with the three hydroxy groups
mimicking the hydroxymethyl and the hydroxy groups in
the 3 and 4 positions in a glucose moiety. Substitution on
nitrogen always results in compounds with equal or less
inhibitory e€ect, indicating that N-substitution does not
destroy the interaction with the binding site.
stirred with 30g K2CO3/300mL H2O for 10min, extrac-
ted with DCM (3Â100 mL), and dried over MgSO4, eva-
poration gave crude methyl 1,2,5,6-tetrahydropyridine-
3-carboxylate (95%).23 Methyl 1,2,5,6-tetrahydropyri-
dine-3-carboxylate (20.44 g, 145 mmol) and triethylamine
(25 mL, 180 mmol) were dissolved in methanol (250mL),
and di-tert-butyl dicarbonate (65.48g, 300 mmol) was
added at rt. The mixture was heated at 60 ꢀC for 30 min,
evaporated to dryness and partitioned between water
and ethyl acetate. The organic layer was isolated, dried
over magnesium sulphate and evaporated to dryness in
vacuo. Puri®cation on silica-gel (eluent: DCM/methanol
(19:1)) gave 2 as a yellow oil (yield: 34.9g, 94%). 13C
NMR (CDCl3, ppm) d 166.20, 155.14, 138.30, 128.5,
80.37, 52.07, 42.96, 28.81, 25.91. 1H NMR (CDCl3, ppm)
d 7.08 (1H, m), 4.10 (2H, m), 3.75 (3H, s), 3.48 (2H, t),
2.32 (2H, m) 1.48 (9H, s).
The most potent compound inhibiting GP Isofagomine
(7), was also able to inhibit glycogen breakdown using
cultured hepatocytes as the functional model system.
Experimental
General procedures
Melting points (uncorrected) were measured on a Buchi
535 apparatus. 1H and 13C NMR spectra were recorded
on a Bruker Avance DPX 200 MHz, or a 300 MHz
apparatus, or a Bruker DRX400 or DRX600 instru-
ment equipped with 5mm selective inverse z-gradients
probe heads operating at room temperature. Chemical
shifts (d-values) are reported relative to TMS. Coupling
constants in Hz were directly measured from the 600 MHz
1H NMR spectrum and con®rmed by comparing experi-
mental with computer simulated multiplet patterns (data
not given).
Methyl 1-tert-butylcarbonyl-1,2,3,6-tetrahydropyridine-3-
carboxylate (3). Diisopropyl amine (7.64 mL) was dis-
solved in dry THF (50 mL), cooled to À78ꢀC in N2
atmosphere followed by slow addition of BuLi (1.9M,
30.50 mL). The temperature was allowed to rise to À10ꢀC
for 30 min. Dry THF (50 mL) was added to the reaction
mixture at À78 ꢀC, then 2 (7.0 g) was added and the
mixture stirred at À78 ꢀC for 5 min. The reaction mix-
ture was poured on a cooled NH4Cl solution (1 M,
100 mL). The aqueous phase was adjusted to pH 10 and
then extracted with Et2O (3Â100 mL). The organic
phase was dried over MgSO4, ®ltrated and evaporated,
1
Mass spectra were obtained on a Finnigan MAT TSQ
70 apparatus using a direct inlet system. The HPLC±MS
analyses were performed on a PE Sciex API 100 LC/MS
System using a WatersTM 3 mmÂ150 mm, 3.5 m, C-18
Symmetry column and positive ion spray with a ¯ow
rate at 20 mL/min. The column was eluted with a linear
gradient of 5±90% A, 85±0% B and 10% C for 15 min
at a ¯ow rate of 1 mL/min (solvent A= acetonitrile, sol-
vent B=water and solvent C=0.1% tri¯uoroacetic acid
in water). Preparative HPLC was performed on Gilson
HPLC equipped with 233XL combined injector/fraction
collector, 305 and 306 pumps each with 25 mL/min
pumping heads.
giving 3 (6.83 g, 97%). H NMR (200 MHz, CDCl3) d
5.86 (q, broad, 2H), 3.91 (q, broad, 2H), 3.72 (s, 3H),
3.52 (broad, 2H), 3.22 (broad,1H), 1.48 (s, 9H).
1-tert-Butoxycarbonyl-1,2,3,6-tetrahydropyridyl-3-meth-
anol (4a). 3 (2.35 g, 9.8 mmol) was dissolved in dry di-
ethyl ether (200 mL) under a nitrogen atmosphere.
Lithium borohydride (0.6g, 27.5 mmol) was slowly added
followed by addition of 1.0 M lithium triethylborohydride
in tetrahydrofuran (3 g, 28.3 mmol). The resulting mixture
was re¯uxed for 1 h, poured into 1 M sodium hydroxide
(100mL) and extracted with methylene chloride. The
organic layer was dried over magnesium sulphate, eva-
porated to dryness in vacuo and puri®ed on a silica-gel
column (eluent DCM/methanol (19/1)) to give 4a as a
colourless oil (Yield: 1.46 g, 70%) MS(EI): m/e 213
LC±MS analyses were obtained on a Hewlett Packard
1100 system, binary pump, degasser, injector, column
oven and DAD detector and MSD single quadropole
mass spectrometer. Microanalyses were performed by
Novo Nordisk Analytical Department.
1
(M+). H NMR (CDCl3, ppm) d 5.76 (H4 and H5, b,
2H), 3.86 (H6, b, 2H), 3.61 (H2, dd, 1H), 3.50 (CH2OH,
dd, 1H), 3.38 (H2 and CH2OH, b, 2H), 2.36 (H3, b,
1H), 1.48 (s, 9H). 13C NMR (CDCl3, ppm) d 155.9
(CO), 126.7, 126.8 (C4 and C5), 80.1 (C(Me)3), 63.3
(CH2OH), 45.1 and 44.4 (C6), 44.4 and 43.2 (C2), 38.6
(C3), 27.7 (Me3).
The syntheses of the following compounds were repor-
ted earlier: 37, 38, 39, 40, 41, 50 and 5316-20; 45, 46, 5121
and 36.22 Compounds 35, 47, 48 and 52 were received
from M. Bols, Aarhus University; compound 49 is
commercially available.
tert-Butyl 3-((1R,2S,5R)-2-isopropyl-5-methyl-1-cyclo-
hexyloxy)acetoxymethyl-1,2,3,6-tetra-hydropyridine-1-
carboxylate (4b). (À)-Menthoxyacetyl chloride (1.03g,
4.4 mmol) was slowly added to a solution of 1-tert-
butoxycarbonyl-1,2,3,6-tetrahydropyridyl-3-methanol (4a)
(0.59 g, 2.8 mmol) in DCM (50 mL) under a nitrogen
atmosphere. The mixture was re¯uxed for 4 h, evapo-
rated to dryness in vacuo and puri®ed on a silica-gel col-
umn (eluent: methyl tert-butylether/methylene chloride
Methyl 1-tert-butoxycarbonyl-1,2,5,6-tetrahydropyridine-
3-carboxylate (2). Arecoline (13.01 g) was dissolved in
dry toluene (400 mL), 0 ꢀC, in an inert atmosphere (N2).
a-Chloroethyl chloroformiate (14.00 g) was added and
the mixture re¯uxed for 3 h and then evaporated. The
residue was dissolved in MeOH (200 mL) and stirred at
RT overnight. Evaporation gave a yellow oil which was

740
P. Jakobsen et al. / Bioorg. Med. Chem. 9 (2001) 733±744
0
000
(1/20)) to give a diasteromeric mixture of 4b as a col-
ourless oil (yield: 0.54 g, 48%). A minor amount of the
two isomers was separated on a chiral column (chiracel;
eluent: PriOH/heptane (5/95)) (Rt 9.6 and 11.2 min). MS:
m/e 409 (M+).13C NMR (CDCl3, ppm) d 171.18, 155.34,
127.7 (m), 127.1 (m), 125.3 (m), 80.67, 80.59, 80.20, 66.29,
66.23, 65.33, 48.51, 43.4 (m), 42.9 (m), 40.35, 35.36, 34.81,
31.90, 28.82, 25.87, 23.68, 22.69, 21.38, 16.68.
J5,5 =3.4 Hz, J5,5 =6.4 Hz and J5,4=10.5 Hz), d 3.43
(H-4, J4,3=9.0 Hz), d 3.6 (H-3, J32a=11.3 Hz and
J3,2e=5.0 Hz), d 2.68 (H-2a, J2a,2e=12.5 Hz), d 3.37
(H-2e), d 3.82 (H-50, J5 ,5 =11.7 Hz), d 3.70 (H-5 ). H
NMR (D2O) 9 and 10 d 3.03 (H-6a, J6a,6e=12.7 Hz and
J6a,5=12.7 Hz), d 3.28 (H-6e, J6e,5=4.0 Hz), d 2.43 (H-5,
00
1
0
00
0
00
J5,5 =6.6 Hz, J5,5 =7.0 Hz and J5,4=2.9 Hz), d 4.02 (H-
4, J4,3=3.9 Hz, J4,2e=0.5 Hz), d 4.07 (H-3, J3,2a=2.1 Hz
and J3,2e=2.3 Hz), d 3.36 (H-2a, J2a,2e=13.7 Hz), d 3.27
00
(H-2e), d 3.73 (H-50, J5 ,5 =11.3 Hz), d 3.63 (H-5 )
0
00
1-tert-Butoxycarbonyl-1,2,3,6-tetrahydropyridyl-3-meth-
anol (4c and 4d). 4a (0.28 g crude) was separated in
enantiomers on a chiracel OD (250±20 mm) column using
2-propanol/heptane (1/99) as eluent at a ¯ow of 6 mL/
min. 4c (S) 1-tert-butoxycarbonyl-1,2,3,6-tetrahydro-
pyridyl-3-methanol (40 mg) was collected from 16±
23.5 min (ee >98%) and 4d (R) 1-tert-butoxycarbonyl-
1,2,3,6-tetrahydropyridyl-3-methanol (110 mg) was col-
lected from 26.5±35.5 min (ee >98%). The NMR data
were found in accordance with the ones reported for 4a.
The actual enantiomeric form was estimated from the data
of the ®nal products 4c leading to 8 and 10; 4d leading to 7
and 9.
tert-Butyl 3-((1R,2S,5R) 2-isopropyl-5-methyl-1-cyclo-
hexyloxy)acetoxymethyl-4,5-epoxypiperidine-1-carboxy-
late (5b). A solution of 3-chloroperoxybenzoic acid
(0.9 g, 5.2 mmol), tert-butyl 3-((1R,2S,5R)-2-isopropyl-
5-methyl-1-cyclohexyloxy)acetoxymethyl-1,2,3,6-tetra-
hydropyridine-1-carboxylate (0.5 g, 1.2 mmol) and methyl-
ene chloride (150 mL) was stirred at rt for 20 h. An extra
portion of 3-chloroperoxybenzoic acid (1.1 equiv) was
added, and the mixture was re¯uxed for 2 h. The reac-
tion mixture was extracted with 1 M sodium hydroxide
and water, dried over magnesium sulphate and evapo-
rated to dryness in vacuo. The residue was puri®ed on a
silica-gel column (eluent: heptane/ethyl acetate (5/1)) to
give 5b as an oil (yield: 0.34g, 66%). 13C NMR (CDCl3,
ppm) d 173.16, 157.08, 82.71, 82.61, 68.28, 66.27, 65.78,
54.43, 53.9, 52.99, 52.35, 50.54, 44.5 (broad), 42.38, 41.22,
36,82, 33.92, 32.13, 30.77, 27.93, 25.72, 25.1, 23.4, 18.72.
(3S,4S,5S)-Hydroxymethyl-3,4-piperidinediol (8) and (3R,
4R,5S)-hydroxymethyl-3,4-piperidinediol (10). MCPBA
(230 mg, 70%) was dissolved in DCM (20 mL) and
extracted with water (20 mL). The DCM phase was iso-
lated and dried over MgSO4. 4c (48.5 mg) dissolved in
DCM (10 mL) was added to the MCPBA solution, and
the mixture was stirred at rt overnight. Subsequently,
the reaction mixture was extracted with NaOH (10 mL),
the water layer was extracted with DCM (3Â10 mL), the
combined DCM layers were pooled and evaporated.
The residue was dissolved in KOH (10%), re¯uxed for
3 h and stirred at RT overnight. Evaporation gave a
crystalline mass which was puri®ed on a silica-gel col-
umn using 2-propanol/25% NH3 as eluent. The crude
product was dissolved in MeOH/H2O, acidi®ed with
HCl (1 M) to pH 1 and evaporated to dryness. The
resulting diastereomeric mixture was separated on a
Merck LiChrosorb NH2. A 7 mm, 25Â250 mm column.
Flow 25 mL/min, 70% aqueous acetonitrile, 0.25 min/
fraction. The samples were dissolved in 3 mL of mobile
phase and injected. Every third fraction was analysed by
FI±MS, EIC 148 m/z showed fraction containing `rele-
vant' molecular weight. The fractions containing the
relevant compounds were all analysed by LC±MS. Flow
1 mL/min, 70%, acetonitrile, 30% 1mM ammonium-ace-
tate, column Merck LiChrosorb NH2 5 mm, 4Â250 mm.
10 (4.7 mg) eluted between 5.0 and 5.75 min, and 8
(3 mg) eluted between 6.25 and 7 min. Both compounds
eluted from the column as their HCl-salts. [a]d for 8 was
found to be À19ꢀ ( MeOH,c=0.20, 20 ꢀC).9
tert-Butyl 3-hydroxymethyl-4,5-epoxypiperidine-1-car-
boxylate (5a). was prepared from 1-tert-butoxycar-
bonyl-1,2,3,6-tetrahydropyridyl-3-methanol 4a (40 mg)
as described for 5b. Isolated as an oil, yield 70%. LC±
MS showed a 100% ELS purity, however, no molecular
ion+1 was detected due to split o€ of the Boc-group by
1
the CF3 COOH in the eluent. H NMR (CDCl3, ppm)
3.6±4.0 (m, 4H), 3.35 (t, 1H), 3.1±3.3 (m, 3H), 2.23 (m,
1H), 1.48 (s, 9H).
5-Hydroxymethyl-3,4-piperidinediol (9,10). 5b (0.25 g,
0.59 mmol) and 10% aqueous potassium hydroxide
were re¯uxed for 3 h. The reaction mixture was evapo-
rated to dryness in vacuo and puri®ed on a silica-gel
column (Eluent: 2-propanol/25% NH3 aq (3:1)) to give
an oil (yield: 56 mg, 65%). Further puri®cation on chiral
TLC (Machery-Nagel 811058 plates; eluent: 2-propa-
nol/25% NH3 aq (3/1)) gave two isomers Rf 0.55 resp.
0.59. However they were not well separated and the
mixture of the two isomers was subsequently separated
on HPLC using a Merck Lichrosorb NH2 column, elu-
ent 70% CH 3CNaq. Only one fraction containing 9,10
was isolated and structure elucidated. The minor frac-
1
tion was not further identi®ed. H NMR was as found
for compound 9.
(3R,4R,5R) 5-Hydroxymethyl-3,4-piperidinediol (7) and
(3S,4S,5R) 5-hydroxymethyl-3,4-piperidinediol (9). 4d
(70 mg) was treated with MCPBA and KOH as descri-
bed above for 4c. The two diastereoisomers eluted as
their HCl salts between 5 and 6 min (9) (8.2 mg), resp.
6.75 and 7.5 min (7) (2.4 mg). [a]D for 7 was found to be
+17ꢀ (MeOH,c=0.16, 20 ꢀC).9 1H NMR (D2O) 7 and 8
d 2.73 (H-6a, J6a,6e=12.7 Hz and J6a,5=12.0 Hz), d 3.34
(H-6e, J6e,5=4.5 Hz and J6e,2e=1.7 Hz), d 1.85 (H-5,
MS: m/e 147 (M+). 13C NMR (D2O, ppm) d 71.52,
68.94, 65.64, 65.54, 60.59, 59.38, 47.03, 45.20, 44.57,
41.44, 40.98, 35.26.
5-Hydroxymethyl-3,4-piperidinediol (7,8) was prepared
analogously from 5a (0.65 g). Puri®cation on chiral TLC
(Machery-Nagel 811058 plates; eluent: 2-propanol/25%
NH3 aq (3/1)) gave two isomers Rf 0.52 resp. 0.58.

P. Jakobsen et al. / Bioorg. Med. Chem. 9 (2001) 733±744
741
Subsequent separation on HPLC using a Merck
Lichrosorb NH2 column, eluent 70% CH 3CNaq. Only
one fraction containing 7,8 was isolated and structure
elucidated. The minor fraction was not further identi-
(3R,4R,5R)-3-Benzyloxy-5-hydroxymethyl-1-methyl-
piperidin-4-ol (51) was prepared analogous to reported
for 54 from 4-O-benzyl-2-deoxy-2-C-hydroxymethyl-d-
xylo-pentodialdose (2.78 g), using methyl amine hydro-
chloride (1.16 g) and sodium cyanoborohydride (0.62 g) in
methanol (50 mL). Yield 1.15g (38%). MS (M+1=252.
Elemental analysis: calculated for C14H22NO3Cl, 2.9
H2O: C, 49.45%; H, 7.71; N, 4.12. Found: C: 49.21%;
1
®ed. H NMR was as found for compound 7.
Yield of 7,8: 0.35 g (51%) as an oil, LC±MS M+1=148.
1
1-Methyl-1,2,3,6-tetrahydro-3-pyridylmethanol (11). Di-
isopropyl amine (21.7 mL), was dissolved in dry THF
(50 mL), cooled to À78 ꢀC in N2 atmosphere followed
by addition of BuLi (2.2 M,70.80 mL). The temperature
was allowed to rise to À10 ꢀC for 30 min. Dry THF
(150 mL) was added to the reaction mixture at À78ꢀC,
then Arecoline (12.61 g) was added and the mixture
stirred at À78ꢀC for 5 min. The reaction mixture was
poured on a cooled saturated NH4Cl solution (1 M,
300 mL). The aqueous phase was adjusted to pH 10 and
then extracted with Et2O (3Â200 mL). The etheral layer
dried over MgSO4, ®ltrated and evaporated in vacuo,
giving methyl 1-methyl-1,2,3,6-tetrahydro-pyridine-3-
carboxylate (6.96 g, 55%).14 Lithium triethylborohy-
dride in tetrahydrofuran (1 M, 75 mL, 75 mmol) was
slowly added to a solution of methyl 1-methyl-1,2,3-6-
tetrahydropyridine-3-carboxylate (5 g, 31.8 mmol) in dry
tetrahydrofuran (200 mL). The mixture was re¯uxed for
3 h, cooled to rt and poured into 1 M aqueous hydro-
chloric acid (150 mL). The aqueous layer was extracted
with methylene chloride and pH was adjusted to 9 with
aqueous sodium hydroxide. The solution was extracted
with methylene chloride which was dried over magnesium
sulfate and evaporated to dryness in vacuo. The residue
was distilled (kugelrohr) (bp: 65±75 ꢀC/0.28mbar) to give
11 as an oil (yield: 1.9 g, 48%). 13C NMR (CDCl3, ppm)
127.28, 126.34, 66.94, 57.31, 54.83, 46.00, 37.58.
H, 7.49; N, 4.38. H NMR (CDCl3, ppm) d 1.8 (ddd,
2H), 1.9 (m, 1H), 2.3 (s, 1H), 2.75 (m, 1H), 3.1 (m, 1H),
3.2±3.8 (m, 6H), 4.6 (dd, 2H), 7.2±7.4 (m, 5H). 13C
NMR (CDCl3, ppm) d 42.07, 44.97, 55.59, 56.75, 63.11,
71.27, 75.06, 79.12, 126.94, 127.01, 127.65, 137.38.
(3R,4R,5R)-1-Butyl-5-hydroxymethyl-3,4-piperidinediol
(12). (3R,4R,5R)-3-Benzyloxy-5-hydroxymethyl-4-piperi-
dinol9 (87 mg, 0.37 mmol), 1-iodobutane (81 mg, 0.44
mmol, 50 mL), potassium carbonate (154 mg, 1.1 mmol)
and dry acetone (8 mL) were stirred for 24 h at 40 ꢀC
under a nitrogen atmosphere. The solvent was evapo-
rated in vacuo, and the residue puri®ed on a silica-gel
column (eluent: ethyl acetate/methanol/25% ammo-
nium hydroxide (6/1/1%)). This a€orded the free base
of (3R,4R,5R)-3-benzyloxy-1-butyl-5-hydroxymethyl-4-
piperidinol (56) (yield: 72 mg, 67%). 1H NMR (CD3
OD, ppm) d 7.4±7.2 (m, 5H), 4.70 (dd, 2H), 3.80 (dd,
1H), 3.54 (m, 1H), 3.34 (m, 1H), 3.25 (d, 1H), 3.08 (m,
2H), 2.38 (t, 2H), 1.85 (dd, 2H), 1.70 (m, 1H), 1.45 (m,
2H), 1.31 (m, 2H), 0.93 (t, 3H).
56 (105 mg, 0.4 mmol) was dissolved in ethanol (20 mL)
containing aqueous hydrochloric acid (4 N, 0.3 mL),
and palladium on carbon (10%, 30 mg) was added. The
mixture was hydrogenated at 1 atm H2-pressure for 3 h,
®ltered through Celite and concentrated in vacuo. Puri-
®cation on a silica-gel column (eluent: ethyl acetate:
methanol:25% ammonium hydroxide (4/1/1)) gave the
free base of 12 (yield: 53 mg, 74%) as a semicrystalline
(3R,4R,5R)-1-Benzyl-3-benzyloxy-5-hydroxymethylpiper-
idin-4-ol (54). 4-O-Benzyl-2-deoxy-2-C-hydroxymethyl-
d-glucopyranose9 (6.0 g, 21.1 mmol) was dissolved in
methanol (200 mL). Sodium periodate (21.4 g,
100 mmol) in water (200 mL) was added dropwise over
15 min. The mixture was stirred at 45±47 ꢀC for 3.5 h.
The precipitate was ®ltered o€, and the mixture was
concentrated and puri®ed by ¯ash chromatography on
silica-gel using ethyl acetate as eluent. The puri®ed pro-
duct of 4-O-benzyl-2-deoxy-2-C-hydroxymethyl-d-xylo-
pentodialdose was dissolved in methanol (50 mL) and
benzyl amine (4.36 g, 40.7 mmol), and sodium cyano-
borohydride (1.0 g, 15.9 mmol) was added. The mixture
was stirred at rt for 3 days. Subsequently, the mixture
was acidi®ed to pH=2 with concd hydrochloric acid
and evaporated to dryness. The residue was dissolved in
water (50 mL), pH was adjusted to 10 with aqueous
sodium hydroxide, and the solution was extracted with
methylene chloride. The organic phase was dried over
magnesium sulphate, evaporated and puri®ed on silica-
gel using ethyl acetate as eluent. This a€orded 54 as an
oil (yield: 25%). 1H NMR (CDCl3, ppm) d 1.6±1.85 (m,
3H), 2.8±3.5 (m, 7H), 3.7 (dd, 1H), 4.55 (dd, 2H), 4.8 (s,
2H), 7.1±7.3 (m. 10 H).13C NMR (CDCl3, ppm) d 44.07,
55.02, 55.92, 61.65, 62.46, 72.43, 74.05, 80.20, 127.14,
127.33, 127.55, 127.89, 128.22, 128.27, 129.42, 137.48,
139.01, 172.84.
ꢀ
1
compound. Mp 94±95 C; MS(SP) m/e 203 (M+). H
NMR (CD3OD, ppm) d 3.80 (dd, 1H), 3.51 (m, 2H),
3.06 (m, 3H), 2.43 (dd, 2H), 1.91 (dt, 2H), 1.75 (m, 1H),
1.53 (m, 2H), 1.33 (m, 2H), 0.92 (t, 3H).
In a similar way the following compounds were prepared.
((3R,4R,5R)-1-Dodecyl-5-hydroxymethyl-3,4-piperidine-
diol (15). (3R,4R,5R)-3-Benzyloxy-1-dodecyl-5-hydroxy-
methylpiperidin-4-ol (57) was prepared from (3R,4R,
5R)-3-benzyloxy-5-hydroxymethyl-4-piperidinol9 and
iodododecane. 1H NMR (CD3OD, ppm) 7.4±7.2 (m,
5H), 4.70 (dd, 2H), 3.80 (dd, 1H), 3.54 (m, 1H), 3.34 (m,
1H), 3.25 (d, 1H), 3.08 (m, 2H), 2.38 (t, 2H), 1.89 (dd,
2H), 1.78 (m, 1H), 1.5 (m, 2H), 1.3 (m, 18H), 0.9 (t, 3H).
Reduction of 57 a€orded 15. Mp 59±60 ꢀC; MS(SP) m/e
1
315 (M+). H NMR (CD3OD, ppm) d 3.80 (dd, 1H),
3.51 (m, 2H), 3.06 (m, 3H), 2.40 (dd, 2H), 1.91 (dt, 2H),
1.75 (m, 1H), 1.53 (m, 2H), 1.3 (m, 20H), 0.9 (t, 3H).
(3R,4R,5R)-5-Hydroxymethyl-1-(3-phenylpropyl)-piperi-
dine-3,4-diol (21). From (3R,4R,5R)-3-benzyloxy-5-hy-
droxymethyl-1-(3-phenylpropyl)-piperidin-4-ol. 1H N MR
(CD3OD, ppm) d 7,2 (m, 5H), 3,80 (dd, 1H), 3.50 (m,

742
P. Jakobsen et al. / Bioorg. Med. Chem. 9 (2001) 733±744
2H), 3.06 (m, 3H), 2.62 (t, 2H), 2.40 (m, 2H), 1.92±1.70
(m, 3H), 1.23 (m, 2H).
(Yield: 35mg, 78%). LC±MS m/e 264 (MH+). 1H N MR
(CD3OD, ppm) d 7.5 (d, 2H), 7.3 (m, 3H), 6.9 (d, 1H), 6.35
(dt, 1H), 3.9 (d, 2H), 3.8 (dd, 1H), 3.7 (m, 2H), 3.55 (m,
2H), 3.4 (t, 1H), 2.76±3.0 (m, 2H), 1.95 (m, 1H).
(3R,4R,5R)-1-(3-Cyclohexyl)propyl)-3,4-dihydroxy-5-hy-
droxymethylpiperidin (22). (3R,4R,5R)-3-Benzyloxy-5-
hydroxymethyl-4-piperidinol9 (0.397 g, 1.7 mmol) was
dissolved in dry methanol (5 mL) and HCl in methanol
(2 N) was added in excess. Evaporation to dryness in
vacuo a€orded the hydrochloride salt. The salt was dis-
solved in dry methanol (20mL) and sodium cyanoboro-
hydride (0.126g, 2.0 mmol) was added. HCl in methanol
(2N) was added dropwise to obtain pH=6 of the mixture
(2 dr). 3-Cyclohexylpropionaldehyde (0.306mL, 0.281 g,
2.0 mmol) was added, and the mixture was stirred under a
nitrogen atmosphere for 18 h at rt. HCl in methanol
(2 N) was added twice to maintain pH=6 of the solu-
tion. Concentration in vacuo and puri®cation of the
residue on a silica-gel column (eluent: ethyl acetate:
methanol/25% ammonium hydroxide (6/1/1)) a€orded
(3R,4R,5R)-3-benzyloxy-1-(3-(cyclohexyl)propyl)-5-hy-
droxymethylpiperidin-4-ol (yield: 0.58g, 96%). 1H N MR
(CDCl3, ppm) d 7.4±7.2 (m, 5H), 4.62 (dd, 2H), 3.68 (m,
3H), 3.52±3.35 (m, 2H), 3.16 (m, 1H), 2.82 (m, 1H), 2.32
(t, 2H), 1,90 (m, 1H), 1.87±1.59 (m, 5H), 1.44 (m, 2H),
1.15 (m, 7H), 0.88 (m, 2H).
The following compounds were prepared by use of the
same solid support technique, the crude products were
puri®ed on HPLC, the identity con®rmed by their LC±
1
MS data and H NMR spectra.
(3R,4R,5R)-1-Allyl-5-hydroxymethyl-3,4-piperidinediol
(13). LC±MS m/e 188 (M+1). 1H NMR (CD3OD, ppm)
d 5.9±6.1 (m, 1H), 5.5±5.7 (m, 2H), 3.3±3.9 (m, 8H), 2.7±
3.0 (m, 2H), 1.95 (m, 1H). 13C NMR (CD3OD, ppm)
127.99, 126.77, 72.70, 70.14, 60.71, 60.41, 55.93, 43.12.
(3R,4R,5R)-1-Benzyl-5-hydroxymethyl-3,4-piperidinediol
(18). LC±MS m/e 238 (M+1). 1H NMR (CD3OD,
ppm) d 7.5 (m, 5H), 4.24 (s, 2H), 3.6±3.9 (m, 3H), 3.3±
3.5 (m, 3H), 2.7±3.0 (m, 2H), 1.95 (m, 1H).
(3R,4R,5R)-1-(4-Bromophenacyl)-5-hydroxymethyl-3,4-
piperidinediol (24). LC±MS 345 (M+1).
(3R,4R,5R)-5-Hydroxymethyl-1-phenacyl-3,4-piperidine-
diol (25). LC±MS 266 (M+1).
(3R,4R,5R)-3-Benzyloxy-1-(3-(cyclohexyl)propyl)-5-hy-
droxymethylpiperidin-4-ol (0.53 g, 1.48 mmol) was dis-
solved in ethanol (20 mL) containing aqueous hydro-
chloric acid (4 N, 0.3 mL) and palladium on carbon
(10%, 50 mg) was added. The mixture was hydrogenated
at 1 atm H2-pressure for 3 h, ®ltred and concentrated in
vacuo. Puri®cation on a silica-gel column (eluent: ethy-
lacetate/methanol/25% ammonium hydroxide (6/1/1))
gave the free base of 22 (yield: 0.21 g, 46%) as crystals.
(3R,4R,5R)-1-(3,4-Di¯uorophenacyl)-5-hydroxymethyl-
3,4-piperidinediol (26). LC±MS 302 (M+1).
(3R,4R,5R)-5-Hydroxymethyl-1-(4-phenylphenacyl)-3,4-
piperidinediol (28). LC±MS 342 (M+1).
(3R,4R,5R)-5-Hydroxymethyl-1-(3-methoxyphenacyl)-
3,4-piperidinediol (29). LC±MS 296 (M+1).
Mp 107±108 ꢀ C. MS(FAB): m/z 272 (M+1). H N MR
1
(CD3OD, ppm) d 3.80 (dd, 1H), 3.50 (m 2H), 3.06 (m,
3H), 2.37 (m, 2H), 1.88±1.58 (m, 8H),1.52 (m, 2H), 1.2
(m, 6H), 0.90 (m, 2H).
(3R,4R,5R)-1-(4-Tri¯uoromethoxyphenacyl)-5-hydroxy-
methyl-3,4-piperidinediol (33). LC±MS 350 (M+1).
(3R,4R,5R)-1-(3-Fluorophenacyl)-5-hydroxymethyl-3,4-
piperidinediol (34). LC±MS 284 (M+1).
(3R,4R,5R)-5-Hydroxymethyl-1-(3-phenylallyl)-3,4-pi-
peridinediol (23). To a suspension of Wang±OH resin
(10 g, 9.2 mmol, Bachem, loading: 0.92 mmol/g) in
methylene chloride N,N-diisopropylethylamine (15 mL,
87.8 mmol) and acryloyl chloride (10 mL, 122.7 mmol)
were added. After stirring slowly for 7 h, the mixture
was ®ltered and the resin washed with methylene chlor-
ide (500mL) and methanol (500mL). Drying in vacuo
gave Wang resin±O±CO±CHˆCH2 (yield: 9.4 g). To a
suspension of this resin (4 g, 3.7 mmol) in DMF (30 mL) a
solution of 7 (0.9 g, 6.1 mmol) in DMF (5 mL) was added.
The mixture was shaken for 22h, ®ltered and the resin
washed with DMF (4Â25mL), DCM (4Â25mL) and
methanol (4Â25mL). The resin was dried in vacuo to give
approx. 4.1 g material. 3-Bromo-1-phenyl-1-propene
(180mg, 0.91mmol) was added to a suspension of this
resin (190 mg, 0.17 mmol) in DMF (5 mL). The mixture
was shaken for 24 h, ®ltered and the resin washed with
DMF (5Â5 mL), methylene chloride (3Â5 mL) and DMF
(2Â5 mL). N,N-Diisopropylethyl amine (0.125 mL, 0.73
mmol) and DMF (3 mL) were added to the resin. The
mixture was shaken for 24 h, ®ltered and the ®ltrate was
concentrated to dryness in vacuo giving 23 as an oil
(3R,4R,5R)-1-(4-Chloro-3-methylphenacyl)-5-hydroxy-
methyl-3,4-piperidinediol (27). LC±MS 314, 316 (M+1).
1H NMR (CD3OD, ppm) d 7.95 (d, 1H), 7.85 (dd, 1H),
7.6 (d, 1H), 5.0 (s, 2H), 2.9±3.9 (m, 8H), 2.5 (s, 3H), 2.1
(m, 1H).
(3R,4R,5R)-5-Hydroxymethyl-1-(2-(40-methoxyacetophe-
1
none))-3,4-piperidinediol (30). LC±MS 296 (M+1). H
NMR (CD3OD, ppm) d 8.0 (d, 2H), 7.07 (d, 2H), 4.7 (s,
2H), 3.9 (s, 3H), 2.7±3.85 (m, 8H), 2.05 (m, 1H).
(3R,4R,5R)-2,4-Dimethoxyphenacyl)-5-hydroxymethyl-
3,4-piperidinediol (31). LC±MS 326 (M+1).
(3R,4R,5R)-1-(4-Diethylaminophenacyl)-5-hydroxymeth-
yl-3,4-piperidinediol (32). 1H NMR (CD3OD, ppm) d
7.85 (d, 2H), 6.75 (d, 2H), 4.85 (s, 2H), 3.6±3.9 (m, 3H),
3.5 (q, 3H), 2.9±3.9 (m, 8H), 2.05 (m, 1H), 1.2 (t, 6H).
(3R,4R,5R)-5-Hydroxymethyl-1-phenethyl-3,4-piperidine-
diol (19). 7 (0.1 g), 2-phenyl-1-bromoethane (0.15 g), KI

P. Jakobsen et al. / Bioorg. Med. Chem. 9 (2001) 733±744
743
(0.14 g) and K2CO3 (0.19 g) were mixed in DMF
(20 mL) and stirred at rt for 10 days. NaOH (1 M,
50 mL) was added and the mixture extracted with DCM
(2Â50 mL), The aqueous phase was evaporated and
puri®ed twice on silica-gel (0.04±0.06 mm) using 2-pro-
panol/25% NH3 (3/1) as eluent. Yield 50 mg (29%) of
was added, the organic phase was seperated, dried
(Na2SO4) and evaporated yielding methyl 5-acetyloxy-
pyridine-3-carboxylate (58) (6.63 g). The crude product
was converted to the hydrochloride by use of HCl in
MeOH (25%), the resulting salt was dissolved in MeOH
(120 mL) and hydrogenated by use of PtO2 (0.75 g) as
catalyst. After shaking for 48 h the solution was ®ltered,
evaporated and puri®ed on a silica-gel column using
MeOH/DCM (1/9) as eluent resulting in a rather low
yield (1.5 g) of methyl 5-acetyloxypiperidine-3-carboxy-
late (59). This product was dissolved in THF (10 mL)
and added dropwise to a solution of LiAlH4 (0.33 g) in
THF (10 mL). Due to diculties in the isolation of the
product, the crude product was acetylated by means of
excess acetic anhydride/sodium acetate. The resulting
product was puri®ed on a silica-gel plate (MeOH/DCM
(1/9) as eluent) yielding 0.15mg of 1-acetyl-3-acetyloxy-
methyl-5-acetyloxypiperidine (60). This was O-deacetyl-
ated using MeONa (1% in MeoH) (0.15 mL) by stirring
at rt overnight. Yield of 1-acetyl-5-hydroxymethyl-
piperidine-3-ol (61) as an oil (0.35 g). MS m/e 173
an oil. LC±MS 252 (M+1). HPLC (MeCN65%, H
PO4 (pH=3), ¯ow 1 mL/min), rt 1.73 min, 98.5%.
3
(3R,4R,5R)-5-Hydroxymethyl-1-[2-(4-nitrophenyl)ethyl]-
piperidine-3,4-diol (20). Preparation from 7 (0.1 g) and
2-(4-nitrophenyl)-1-bromoethane (0.19 g) as described
for 19 gave 0.053 g (25%) of a yellow oil. Attempts to
make the hydrochloride gave a hygroscopic mass which
resulted in an oil; no further attempts were made to
make crystals.
(3R,4R,5R)-1-Acetyl-5-hydroxymethyl-3,4-piperidinediol
(16). (3R,4R,5R)-3-Benzyloxy-5-hydroxymethyl-4-piperi-
dinol (0.1 g, 0.4 mmol) was dissolved in pyridine (0.5
mL) at 0 ꢀC, and acetic anhydride (0.5 mL) was added
dropwise over 1 min. Stirring at rt for 20 h and evapora-
tion to dryness gave a crude product which was puri®ed
on silica-gel using ethanol as eluent (yield: 55%). The
puri®ed 3-benzyloxy-3-acetoxy-5-acetoxymethyl-1-acetyl-
piperidine (75 mg) was debenzylated in methanol at rt
using 10% Pd/C as catalyst. Filtration and evaporation
to dryness gave 50 mg of crude 4-acetoxy-5-acetoxy-
methyl-1-acetyl-3-piperidinol which was dissolved in dry
methanol (4 mL) containing 50 mL of a 1% sodium
methoxide in methanol. After stirring at rt for 4 days,
the solution was evaporated to dryness to give 16 (yield:
1
(M+); H NMR (MeOH) 4.42 (m, 1H), 3.82 (m, 1H),
3.5±3.0 (m, 4H), 2.68 (dd, 1H), 2.21 (dd, 1H), 2.01 (s,
3H), 1.87±2.09 (m, 1H), 1.6 (m, 1H). 61 (25 mg) was
deacetylated by stirring with HCl (6 N, 1.5 mL) at 80 ꢀC
for 2 days, the reaction mixture was ®ltered and evapo-
rated to dryness. Yield of 42 (10 mg, 40%). MS m/e 131
(M+). 13C NMR 64.7, 64.5, 50.1, 46.8, 36.6, 36.1.
cis- and trans-3-Hydroxymethyl-4-piperidinol (43). Ethyl
1-benzyl-4-oxopyridine3-carboxylate (15 g) was dis-
solved in dry MeOH (300 mL). NaBH4 (30 g) was added
to this solution over 20 min under N2, the mixture was
cooled during the addition, the cooling was removed
whereby the temperature rose to 80 ꢀC in 30 min. The
mixture was further stirred at rt overnight. H2O
(10 mL) was added and the solution evaporated to dry-
ness. EtAc (400 mL) was added and the mixture re¯uxed
for 20 min, ®ltered hot and evaporated to dryness
resulting in a crude oil (13.6 g). The mixture was pur-
i®ed on a silica-gel column (eluent: DCM/MeOH, 9/1).
The ®rst eluting compound was identi®ed as the trans
isomer of 1-benzyl-3-hydroxymethyl-4-piperidinol (62)
(2.6 g). 13C NMR (DMSO) 138.6, 128.7, 128.0, 126.8,
68.5, 62.1, 61.3, 55.0, 51.8±46.0, 38.4. 1H N MR
(DMSO) 7.2±7.35 (m, 5H), 4.55 (d, 1H), 3.62 (dd, 1H),
3.3±3.5 (b, 2H), 3.22 (d, b, 1H), 3.15 (m, 1H), 2.90 (d, b,
1H), 2.7 (m, 1H), 1.7 (m, 2H), 1.51 ( m, 1H), 1.40 (m,
1H). The last eluted was identi®ed as the cis isomer
(2.4 g) of 1-benzyl-3-hydroxymethyl-4-piperidinol (63)
1H NMR (DMSO) 4.35 (b, 1H), 3.78 (s, 1H), 3.4 (m,
3H), 3.31 (m, 1H), 2.5±2.3 (m, 3H), 1.68 (m, b, 1H), 1.58
(m, 2H). 13C NMR (DMSO) 138.5, 128.8, 128.7, 126.8,
64.1, 62.4, 61.3, 51.0, 48.2, 43.3, 32.6. 62 and 63 (1.0 g)
were debenzylated by means of Pd/C (10%) (0.25 g)
catalysed hydrogenation at atm. pressure in abs. ethanol
(100 mL) giving quantitative yield of crude product as
oil.
1
35 mg). H NMR (CD3OD, ppm) d 1.6 (m, 1H, H-5),
2.1 (s, 3H, O±Me), 2.4±2.7 (m, 1H, H-6ax), 2.8±3.1 (m,
1H, H-4), 3.2±3.4 (m, 2H, H-2), 3.5±3.7 (m, 1H, H-50),
3.7±4.1 (m, 2H, H-50+H-3), 4.4±4.6 (m, 1H, H-6equiv).
13C NMR (CD3OD, ppm) d 24.11, 24.16, 47.01, 48.17,
49.17, 50.14, 51.54, 51.78, 51.98, 54.81, 64.75, 64.84,
75.28, 75.88, 78.14, 78.36.
(3R,4R,5R)-3-(3,4-Dihydroxy-5-hydroxymethyl-piperidin-
1-yl)propionic acid (17). To a suspension of Wang
resin±O±CO±CHˆCH2 (2.0 g, 1.84 mmol) in DMF
(15 mL) a solution of (3R,4R,5R)-5-hydroxymethyl-3,4-
piperidinediol (0.8 g, 5.4 mmol) in DMF (3 mL) was
added. The mixture was shaken for 22 h, ®ltered and the
resin washed with DMF (2Â20 mL), methanol (3Â20mL),
DMF (2Â20 mL), methylene chloride (2Â20 mL) and
methanol (2Â20 mL). The resin was dried in vacuo to
give approx. 2.2 g material. To a suspension of this resin
(0.5 g, 0.46 mmol) in methylene chloride (5 mL) tri-
¯uoroacetic acid (5 mL) was added. The mixture was
shaken for 20 min, ®ltered and the ®ltrate was evapor-
ated to dryness in vacuo to give (3R,4R,5R)-3-(3,4-
dihydroxy-5-hydroxymethyl-piperidin-1-yl)propionic
acid as a yellow oil (yield: 40 mg, 40%). LC±MS m/e
1
220 (MH+). H NMR (CD3OD, ppm) d 3.3±3.9 (m,
8H), 2.8±3.1 (m, 4H), 1.95 (m, 1H).
cis-5-Hydroxymethyl-3-piperidinol (42). Methyl 5-hy-
droxypyridine-3-carboxylate (5.0 g), acetic anhydride
(4.95 mL), and sodium acetate (10 mg) were mixed in
DCM (20 mL). After stirring at rt for 6 h H2O (30 mL)
cis-3-Hydroxymethyl-4-piperidinol. 1H NMR (MeOH)
4.05 (dd, 1H), 3.4±3.7 (m, 2H), 2.97 (m, 1H), 2.8±2.7 (m,
3H), 1.82±1.2 (m, 3H). 13C NMR (MeOH) 66.6, 63.0,
44.4, 44.2, 41.5, 34.0.

744
P. Jakobsen et al. / Bioorg. Med. Chem. 9 (2001) 733±744
trans-3-Hydroxymethyl-4-piperidinol. 1H NMR (MeOH)
3.60 (dd, 1H), 3.4±3.15 (m, 2H), 3.0±2.65 (m, 2H), 2.45±
2.25 (m, 1H), 2.13 (t, 1H), 1.8±1.6 (m, b,1H), 1.45±1.12
(m, 2H). 13C NMR (MeOH) 70.6, 62.9, 48.4, 48.1, 45.5,
36.1.
References
1. Ichikawa, Y.; Igarashi, Y.; Ichikawa, M.; Suhara, Y. J.
Am. Chem. Soc. 1998, 120, 3007.
2. Schuster, M. Bioorg. Med. Chem. Lett. 1999, 9, 615.
3. Schneider, C.; Kazmeier, U. Eur. J. Org. Chem. 1998, 1155.
4. Firth, R. G.; Bell, P. M.; Hansen, H. M.; Rizza, R. A. J.
Clin. Invest. 1986, 77, 1525.
5. Ferrannini, E.; Simonson, D. C.; Katz, L. D.; Reichars, G.,
Jr.; Bevilacqua, S.; Barrett, E. J.; Olsson, M.; DeFronzo, R. A.
Metab., Clin. Exp 1988, 37, 79.
6. Mitrakou, A.; Kelley, D.; Veneman, T.; Jenssen, T.; Pang-
burn, T.; Reilly, J.; Gerich, J. Diabetes 1990, 39, 1381.
7. Hellerstein, M. K.; Neese, R. A.; Linfoot, P.; Christiansen,
M.; Turner, S.; Letscher, A. J. Clin. Invest. 1997, 100, 1305.
8. Lundgren, K.; Rassov, A.; Bols, M. Diabetes 1996, 45(Supp.
l2), 521.
trans-3,4-Piperidinediol (44). 1-Benzyl-1,2,3,6-tetrahy-
dropyridine (4.87 g) was slowly added to a cooled mix-
ture of acetic acid (31 mL) and H2O2 (7.5 mL, 30%).
The resulting mixture was heated for 5 h at 80 ꢀC. The
solution was evaporated to 1/5 volume and Na2CO3 was
added to pH 9. Satd NaCl (10 mL) was added and the
mixture extracted with DCMÂ3. The organic phases
were collected and evaporated to dryness. The resulting
crystalline mass was dissolved in EtOH (25 mL) and
acidi®ed with HCl (6 M, 4.7 mL). The mixture was
evaporated, redissolved in EtOH and precipitated with
ether.
9. Jespersen, T. M.; Bols, M.; Sierks, M. R.; Skrydstrup, T.
Tetrahedron 1994, 47, 13449.
10. Altenbach, H.-J.; Blanda, G. Tetrahedron Asymmetry
1998, 9, 1519.
11. Hansen, S. U.; Bols, M. J. Chem. Soc., Perkin Trans. 1
2000, 911.
12. Dong, W.; Jespersen, T. M.; Bols, M.; Skrydstrup, T.;
Sierks, M. R. Biochemistry 1996, 35, 2788.
13. Olofson, R. A.; Martz, J. T.; Senet, J.-P.; Piteau, M.;
Malfroot, T. J. Org. Chem. 1984, 49, 2081.
14. Allan, R. D.; Fong, J. Aust. J. Chem. 1983, 36, 601.
15. Andersen, B.; Rassov, A.; Westergaard, N.; Lundgren, K.
Biochem. J. 1999, 342, 545.
0.16 g of this crystalline mass was reduced by hydro-
genation at atm pressure using Pd/C (10%) as catalyst
in abs ethanol (20 mL), giving quantitative yield of 44.
Mp 128 ꢀC (HCl±salt). MS 117 (M+). 13C N MR
(MeOH) 67.7, 67.5, 45.7, 41.2, 27.1. H NMR (MeOH)
3.75 (m, 2H), 3.31 (dt, 2H), 3.3±2.92 (m, 2H), 2.2 (M,
1H), 1.8±1.6 (m, 1H).
1
16. Asano, N.; Tomioka, E.; Kizu, H.; Matsui, K. Carbohydr.
Res. 1994, 253, 235.
Biology
17. Asano, N.; Oseki, K.; Tomioka, E.; Kizu, H.; Matsui, K.
Carbohydr. Res. 1994, 259, 243.
18. Kato, A.; Asano, N.; Kizu, H.; Matsui, K.; Watson, A. A.;
Nash, R. J. J. Nat. Prod. 1997, 60, 312.
19. Asano, N.; Oseki, K.; Kizu, H.; Matsui, K. J. Med. Chem.
1994, 37, 3701.
20. Asano, N.; Kato, A.; Miyauchi, M.; Kizu, H.; Tomimori,
T.; Matsui, K.; Nash, R. J.; Molyneux, R. J. Eur. J. Biochem.
1997, 248, 296.
21. Streith, J.; Defoin, A. Synlett. 1996, 189.
22. Ichikawa, Y.; Igarashi, Y. Tetr. Lett. 1995, 36, 4585.
23. Olofson, R. A.; Martz, J. T.; Senet, J-P.; Piteau, M.;
Malfroot, T. J. Org. Chem. 1984, 49, 2081.
24. Brown, A. R.; Rees, D. C.; Rankovic, Z.; Morphy, J. R. J.
Am. Chem. Soc. 1997, 119, 3288.
Pig and rat liver GPa were semi-puri®ed as described in
Fosgerau et al.25 Rabbit muscle GPa and GPb were
obtained from Boehringer Manneheim. Enzyme activity
was measured in the direction of glycogenolysis essentially
as described by Bergmeyer.26
Rat hepatocytes were isolated and cultured as described
by Andersen et al.15 The e€ect of Isofagomine on glyco-
genolysis was measured in glucose-free medium in the
absence or presence of glucagon (0.5nM) as described
previously.15
25. Fosgerau, K.; Westergaard, N.; Quistor€, B.; Grunnet,
N.; Kristiansen, M.; Lundgren, K. Arch. Biochem. Biophys
2000, 380, 284.
26. Bergmeyer, H. U.; GrasslWalter, H.-E. In Methods of
Enzymatic Analysis; Bergmeyer H. U., Ed.; Verlag Chemie:
Weinheim, 1983; Vol. 2, pp 293±294.
Acknowledgements
The assistance of Esther Gammelgaard, Ulrik Jùrgen-
sen, Claus Frederiksen, and Bo R. Pedersen is gratefully
acknowledged.