Synthesis of uronic-Noeurostegine - A potent bacterial β-glucuronidase inhibitor

Article abstract of DOI:10.1039/c1ob06038d

Inhibition of β-glucuronidases has recently been shown to be useful in alleviating drug toxicity for common colon cancer chemotherapeutic CPT-11 (also called Irinotecan). We have prepared a new compound of the nortropane-type, uronic-Noeurostegine, and de

Full text of DOI:10.1039/c1ob06038d

View Article Online / Journal Homepage / Table of Contents for this issue
Organic &
Biomolecular
Chemistry
Dynamic Article Links
Cite this: Org. Biomol. Chem., 2011, 9, 7807
www.rsc.org/obc
PAPER
Synthesis of uronic-Noeurostegine – a potent bacterial b-glucuronidase
inhibitor†‡
Tina S. Rasmussen,a Heidi Koldsø,a,b Shinpei Nakagawa,c Atsushi Kato,c Birgit Schiøtta,b and
Henrik H. Jensen*a
Received 28th June 2011, Accepted 11th August 2011
DOI: 10.1039/c1ob06038d
Inhibition of b-glucuronidases has recently been shown to be useful in alleviating drug toxicity for
common colon cancer chemotherapeutic CPT-11 (also called Irinotecan). We have prepared a new
compound of the nortropane-type, uronic-Noeurostegine, and demonstrated that this is a competitive
and potent E. coli b-glucuronidase inhibitor, while inhibition of the mammalian b-glucuronidase from
bovine liver was found to be less significant. Although not intended, two other compounds having
N-ethyl and N-(4-hydroxybutyl) substituents were also prepared in this study due to the sluggish
debenzylation in the final step. The N-substituents are believed to come from reaction with the solvents
used being ethanol and THF, respectively. These compounds also inhibited the two b-glucuronidases
albeit to a lesser extent compared to the parent compound. Noeurostegine and the three
uronic-noeurostegines were additionally evaluated as inhibitors against a wide panel of glycosidases
with the former showing potent inhibition of rat intestinal lactase and trehalase, whereas the latter was
found to be inactive.
Introduction
The chemistry literature holds numerous examples of how struc-
tural inspiration from naturally occurring compounds has lead the
way to the design and synthesis of new biologically active species.
This indeed also holds true with regards to how iminosugars
like 1-deoxynojirimycin (1) and its analogues1 have led to more
potent synthetic variants like isofagomine (2),2 azafagomines
(3)3 and noeuromycin (4)4 (Fig. 1). The glycosidase inhibition
by compounds of these iminosugar/azasugar families are of
significant interest due to their potential use in treatment of
a variety of diseases like cancer, diabetes and viral infections
Fig. 1 Structures of various iminosugars/azasugars including the focus
including HIV.5
of this study, compound 7.
Recently, we drew inspiration from the naturally occurring,
stable nortropane hemiaminal calystegine B2 (5)6 to prepare a
it as a potentially valuable compound against Gaucher’s disease.8
hybrid of this with the compound noeuromycin (4). We christened
Here we describe the synthesis and glucuronidase inhibitory effects
from the first stable hemiaminal possessing a carboxylic acid
functionality (7) (Fig. 1).
this new stable and structurally intermediary compound noeu-
rostegine (6)7 and demonstrated its potent inhibition of almond
and Thermotoga maritima b-glucosidase and furthermore found
A potentially valuable strategy for alleviating harmful side
effects for the common colon cancer chemotherapeutic CPT-11
was demonstrated by Redinbo and co-workers in the journal
Science in 2010.9 Here, by studying E. coli b-glucuronidase the au-
thors found potent non-carbohydrate-like uncompetitive enzyme
inhibitors to be capable of diminishing the concentration of toxic
CPT-11 metabolites in the gastrointestinal tract. Based on this
newly established and important potential use of glucuronidase
inhibitors, we decided to prepare a glucuronic acid analogue of
noeurostegine and test its inhibition profile.
aDepartment of Chemistry, Aarhus University, Langelandsgade 140, 8000,
Aarhus C, Denmark. E-mail: hhj@chem.au.dk; Tel: +458942 3963
bCenter for Insoluble Protein Structures (inSPIN) and Interdisciplinary
Nanoscience Center (iNANO), Denmark
cDepartment of Hospital Pharmacy, University of Toyama, 930-01940, Japan
† This paper is dedicated to Professor Mikael Bols on the occasion of his
50th birthday
‡ Electronic supplementary information (ESI) available: Molecular dock-
ing calculations, glycosidase inhibition, Michaelis–Menten plots, com-
pound numbering and NMR spectra. See DOI: 10.1039/c1ob06038d
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 7807–7813 | 7807
©

View Article Online
Scheme 1 Synthesis of protected uronic-noeurostegine (14) from levoglucusan.
Celite and addition of new catalyst. The product from this reaction
turned out to be the N-ethyl congener of the desired product.
We speculate that this product could originate from palladium
catalysed dehydrogenation of ethanol to produce acetaldehyde,
which later underwent reductive amination to give N-ethyl uronic-
noeurostegine (15).11
To avoid this unwanted side reaction during hydrogenolysis
we decided to change the solvent to a THF/water mixture.
The reaction also proceeded sluggishly in this medium over 5
days but resulted in a product different from N-ethyl uronic-
noeurostegine (15) by TLC analysis. This however turned out
to be the N-4-hydroxybutyl alkylated nortropane 16.11 In this
case we speculate that a palladium catalysed dehydrogenation of
THF in the presence of water ultimately will lead to formation
of 4-hydroxybutanal, which then undergoes reductive amination
with the amine to give N-4-hydroxybutyl uronic-noeurostegine (16)
(Scheme 2).
Results and discussion
Similar to the synthesis of noeurostegine (6)7 we started with
levoglucosan as previously described. Using this known sequence
of reactions, we are at this time able to achieve 36% of 8 over eight
steps compared to the previous 28% (Scheme 1). The anhydrosugar
8 subsequently underwent a series of reactions also described
earlier to obtain azidocycloheptane 9 in 13% yield over 10 steps.7
Next, the primary benzyl ether was selectively acetolysed into
the corresponding acetate by treatment with ZnCl2 in acetic
anhydride/acetic acid in 74% yield to give 10. A di-deacylation
to remove both the acetyl and benzoyl groups was conducted in
methanol using sodium methoxide afforded 11 (85%), which was
later oxidised by the Jones reagent to the keto carboxylic acid 12.
To avoid working with a zwitter ionic compound until the last
step we decided to protect the carboxylic acid as a benzyl ester.
This was accomplished by reaction of 12 with benzyl bromide
in dry acetonitrile in the presence of cesium carbonate in 67%
yield from diol 11.10 The azide function of 13 was next reduced
by the Staudinger reaction in THF/water to spontaneously give
bicyclo[3.2.1]-compound 14 as indicated by the disappearance of
dC 208 ppm and appearance of dC 92.6 ppm by 13C NMR.
Until this point an old batch of Pearlman’s catalyst
(Pd(OH)2/C) had been used, which could be of weakened activity.
After using a freshly purchased batch of Pearlman’s catalyst for
the debenzylation of 14 in ethanol/ethyl acetate/water the desired
compound 7 was formed.
To obtain the target compound (7), only protecting group
removal remained. We have previously in our synthesis of noeu-
rostegine (6) reported the uneventful removal of benzyl protecting
groups by hydrogenolysis over Pd(OH)2/C,7 but in this project
we experienced significant difficulties in this final step. Heating
or increasing the H2-pressure did not seem to have a noteworthy
effect on starting material consumption. We assume this was due
to a minute impurity that has poisoned and thereby to some
extent inactivated the catalyst towards debenzylation. First the
reaction was carried out in ethanol/water and went on for 10
days with intermediary filtration of the reaction mixture through
The stability of uronic-noeurostegine was evaluated by leaving
it in either a neutral aqueous solution at ambient temperature
or in acetate buffer (100 mM, pH 4.6) at 37 ◦C for several
days. No change after this time was observed by NMR spec-
troscopy demonstrating again that these nortropane hemiaminal
compounds indeed are stable.
Inhibition studies
In line with the recent study by Redinbo and co-workers where
it was shown that E. coli and bovine liver b-glucuronidases are
7808 | Org. Biomol. Chem., 2011, 9, 7807–7813
This journal is
The Royal Society of Chemistry 2011
©

View Article Online
Scheme 2 Hydrogenolysis of 14 using aged and fresh catalyst.
markedly different and that selective inhibition of the former
could have great medicinal potential in cancer treatment, we
decided to evaluate the synthesised nortropanes 7,15,16 against
these enzymes.
We found that all three inhibitors showed similar and low
micromolar competitive inhibition (Ki) against the mammalian
enzyme from bovine liver and that the parent uronic-noeurostegine
(7) was slightly more potent (Ki 2.3 mM) than the N-alkylated
congeners (Ki 9.5 mM and 3.6 mM) (Table 1). These inhibitory
constants are similar to what has been established for the uronic-
azafagomine (19, Ki 1 mM, Fig. 2)3c and uronic-1-deoxynojirimycin
(18, Ki 6.5 mM, Table 1),12 but significantly less potent than
uronic-isofagomine (17, Ki 79 nM).2b This is despite the fact that
17 does not possess a hydroxyl group in what would be the 2-
position of the substrate, which is known to contribute greatly
to binding.13 The high potency of 17 over 18 is most likely due
to the fact that b-glycosidases generally prefer a basic N-atom in
what would be C-1 of the glycoside substrate.5a The noeurostegines
have, like isofagomines, azafagonines and noeuromycins, an N-
atom in the C-1 position, but require that the enzyme is capable of
accommodating the ethylene bridge, which seems to be possible in
this case.
The bacterial b-glucuronidase from E. coli was inhibited
competitively and very potently by uronic-noeurostegine (7) with
Ki of 60 nM whereas N-ethyl and N-4-hydroxybutyl uronic-
noeurostegines (15 and 16) demonstrated a lower degree of
inhibition with Ki-values of 1000 nM and 740 nM, respec-
tively (Table 1). This demonstrates that the ethylene bridge
of the nortropane skeleton can be accommodated by the en-
zymes (vide infra). Uronic-1-deoxynojirimycin (18) is known as
a weak inhibitor of this bacterial enzyme12 and thereby has
an inverse inhibition profile compared to uronic-noeurostegine
(7).
Fig. 2 Structure of uronic-azafagomine (19), glucaro-d-lactam (20)14 and
uncompetitive inhibitors 21 and 22.
Inhibition by the best compound (21, Fig. 2) found by Redinbo
and co-workers showed a Ki of 164 nM (uncompetitive) against E.
coli b-glucuronidase and no inhibition of the bovine liver enzyme.9
To obtain a more complete picture of the inhibition profile of
uronic-noeurostegines 7,15,16 were tested against a series of other
glycosidases. Parent noeurostegine (6), was included in this study.
No significant inhibition of any of the tested glycosidases was
found by the uronic-noeurostegines (7,15,16) at 1 mM concen-
tration illustrating the selectivity displayed by these compounds
(Table 2). In comparison, uronic-azafagomine (19) inhibits yeast
a-glucosidase with Ki 160 mM and almond b-glucosidase with Ki
7 mM.3c Noeurostegine (6) has previously been found not to inhibit
a-glucosidase from yeast, which we also found to be the case for
the rice and rat intestinal enzymes. Mild inhibition of Asp. niger
a-glucosidase was found with an IC50 of 240 mM.
b-Glucosidase from bovine liver (IC50 96 mM) and b-glucosidase
from human lysosome (IC50 12 mM), was inhibited by noeuroste-
gine, while a-galctosidase from human lysosome was inhibited
at a higher concentration (IC50 330 mM). Despite the glucose
configuration of noeurostegine we previously found it to be an
inhibitor of b-galactosidase from both Asp. oryzae and E. coli.
Table 1 Ki values in mM (—, not tested)
Compound
7
15
9.5
1.0
16
3.6
0.74
17
18
b-glucuronidase (bovine liver)a
b-glucuronidase (E. coli)d
2.3
0.060
0.079b
—
6.5c
400c
a pH 4.6, 37 ◦C. b Value from ref. 2b. c Value from ref. 12 d pH 6.8, 37 ◦C.
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 7807–7813 | 7809
©

View Article Online
Table 2 IC50 values in mM. NI: IC50 > 1 mM; (—, not tested)
Compound
a-glucosidase (yeast)
a-glucosidase (rice)
a-glucosidase (rat intestinal maltase)
a-glucosidase (Asp. niger)
b-glucosidase (bovine liver)
b-glucosidase (almonds)
b-glucosidase (human lysosome).
a-galactosidase (coffee beans)
a-galactosidase (human lysosome)
b-galactosidase (bovine liver)
b-galactosidase (rat intestinal lactase)
a-mannosidse (jack bean)
b-mannosidase (snail)
a-L-rhamnosidase (P. decumbens)
a-L-fucosidase (bovine kidney)
a-L-fucosidase (bovine epididymis)
trehalase (porcine kidney)
trehalase (rat intestine)
7
15
16
6
NI
NI
—
—
NI
NI
—
NI
—
NI
—
NI
NI
NI
NI
—
NI
—
NI
NI
NI
—
NI
NI
—
NIa
NI
NI
—
—
240
96
NI
NI
—
NI
—
NI
—
NI
NI
NI
NI
—
NI
NI
—
NI
—
NI
—
NI
NI
NI
NI
—
Ki 0.050a
12
Ki 2.5a
330
47
0.25
NI
670
470
—
NI
—
NI
—
NI
NI
—
NI
0.44
980
amyloglucosidase (Asp. niger)
a Ref. 7.
Here we additionally found 6 to be an inhibitor of the bovine liver
(IC50 47 mM) and rat intestinal enzymes, with the latter being in
the sub-micromolar region (IC50 250 nM).
No inhibition of jack bean a-mannosidase and bovine a-
fucosidase, whereas moderate inhibition of snail b-mannosidase
(IC50 670 mM), P. decumbens a-rhamnosidase (IC50 470 mM) and
Asp. niger amyloglucosidase (975 mM) was found.
Trehalase from rat intestine was inhibited potently with IC50 of
0.44 mM.
to the bacterial enzyme by molecular docking calculations. We
performed a re-docking of the co-crystallised glucaro-d-lactam
(20) as a reference, with the binding mode from the calculations
˚
overlapping perfectly (RMSD = 0.20 A for all ligand atoms) with
the one observed in the crystal structure. Additionally we were
able to dock the novel competitive inhibitor 7 into the binding
pocket in a position overlapping with the competitive inhibitor
20. The NH group of 7 is overlapping with the amide carbonyl of
glucaro-d-lactam resulting in the carboxylic acid being located in
the same pocket of the binding site.
Uronic-noeurostegine (7) was hence found to occupy the central
substrate binding site and have interactions with the two active
site carboxylates (Glu413 and Glu504) in addition to several other
polar interactions, see Fig. 3. Furthermore, the N-atom of uronic-
noeurostegine (7) occupies the position of the C-1 of the substrate
and thereby it binds in a ‘noeuromycin binding mode’ rather than
Molecular docking calculations
Redinbo and co-workers have succeeded in crystallising the E. coli
b-glucuronidase in the presence of both the competitive inhibitor
glucaro-d-lactam (20) and the uncompetitive inhibitor 22 (Fig. 2).
We decided to explore how uronic-noeuromycin (7) would bind
Fig. 3 Docking of 7 in E. coli b-glucuronidase (PDB : 3K4D) and the observed interactions between water and protein residues. The catalytic glutamates
form interactions with the inhibitor 7 in addition to several other polar interactions with charges and hydrophilic residues within the binding pocket. (A)
Schematic representation of the binding site interactions and (B) interactions within the central binding pocket. Uronic-noeurostegine (7) in cyan and
protein side chains in orange. The protein is coloured by secondary structure with a-helices in purple and b-strands in yellow.
7810 | Org. Biomol. Chem., 2011, 9, 7807–7813
This journal is
The Royal Society of Chemistry 2011
©

View Article Online
a ‘1-deoxynojirimycin binding mode’ in which the N-atom takes
the place of O-5 of the substrate. This binding mode has also been
established for calystegine B2 (5) in T. maritima b-glucosidase.15
The location of the uncompetitive inhibitor 22 was found
by Redinbo and co-workers to slightly overlap with the bind-
ing site for uronic-noeurostegine determined by our docking
calculations.
analytes are calculated and reported in Daltons for uncharged
species.
Melting points were measured on a Bu¨chi B-540 instrument
and are uncorrected. Optical rotation was measured on a PE-241
polarimeter and reported in units of deg cm2 g-1. Concentrations
are reported in g/100 mL.
(1R,2R,3R,4R,5R)-4-(Acetoxymethyl)-5-azido-1-O-benzoyl-2,
3-di-O-benzyl-cycloheptane (10). (1R,2R,3R,4R,5R)-5-Azido-1-
O-benzoyl-2,3-di-O-benzyl-4-C-benzyloxymethyl-cycloheptane
(9)7 (0.43 g, 0.73 mmol) was dissolved in acetic anhydride (10 mL)
and acetic acid (5 mL). ZnCl2 (1.57 g, 11.5 mmol, 15 eq) was added
and the reaction mixture was stirred at rt for 3 h. The reaction
mixture was poured onto a saturated aqueous solution of NaHCO3
(100 mL). The aqueous phase was extracted with CH2Cl2 (3 ¥
50 mL) before the combined organic layers were washed with H2O
(2 ¥ 100 mL), dried over MgSO4, filtered and concentrated. The
resulting residue was purified by flash column chromatography
(pentane/EtOAc 15 : 1 → 10 : 1) to give the desired product 10
(0.29 g, 74%) as colourless crystals. Mp 120.5–122.5 ◦C. [a]D20 40.6
Conclusion
We have prepared a new nortropane type b-glucuronidase in-
hibitor in 24 steps from levoglucosan and christened this uronic-
noeurostegine due to its resemblance to calystegine B2 and
noeuromycin. To the best of our knowledge, this is the first stable
hemiaminal amino acid. The final deprotection using an older
batch of Pearlman’s catalyst was found to be sluggish and yielded
the N-ethyl and N-4-hydroxybutyl congeners by reaction with
ethanol and tetrahydrofuran, respectively.
The compounds synthesised in this study were all established
to be potent inhibitors of both bovine liver and E. coli b-
glucuronidase with the parent compound uronic-noeurostegine
exerting the most powerful inhibition with a Ki of 60 nM
against the latter. Inhibition against the bacterial enzyme was
furthermore found to be more pronounced than against the
enzyme of mammalian origin with a selectivity factor of 38. Given
the recent promising results in colon cancer chemotherapy uronic-
noeurostegine or other compounds of the iminosugar/azasugar
family could be a useful medicinal candidate.
1
(c 1.0, CHCl3). Rf (pentane/EtOAc 10 : 1) 0.23. H NMR (400
MHz, CDCl3): dH (ppm) 7.93 (m, 2H, o-Bz), 7.50 (m, 1H, ArH),
7.35 (m, 2H, ArH), 7.23–7.17 (m, 8H, ArH), 7.11–7.09 (m, 2H,
ArH), 5.37 (m, 1H, H1), 4.69–4.63 (m, 3H,CH2Ph), 4.40–4.36 (m,
3H, OCH2Ph, H8), 4.32 (dd, 1H, J 11.2 Hz, J 2.8 Hz, H8¢), 3.96
(t, 1 H, J 6.4 Hz, H2), 3.66 (m, 1H, H5), 3.56 (dd, 1H, J 9.2 Hz,
J 6.4 Hz, H3), 2.09 (tt, 1H, J 9.2 Hz, J 2.8 Hz, H4), 2.00 (m, 4
H, H6,H6¢,H7,H7¢), 1.98 (s, 3H, OCOCH3). 13C NMR (100 MHz,
CDCl3): dC (ppm) 170.9 (CO), 165.9 (CO), 137.9, 133.2, 130.2,
129.7–127.8 (ArC), 82.5 (C2), 78.1 (C3), 74.6, 74.3 (CH2Ph), 73.9
(C1), 63.6 (C8), 60.6 (C5), 47.1 (C4), 28.4, 24.6 (C6, C7), 21.1
(OCOCH3). HRMS(ES+): calcd. for C31H33N3O6Na: 566.2267,
found 566.2271.
Noeurostegine synthesised in a previous study was here also
tested against a wide panel of glycosidases.7 In addition to our
previous findings that this compound is a potent inhibitor of b-
glucosidases from T. maritima and almonds we also found strong
inhibition of rat intestinal lactase and trehalase.
(1R,2R,3R,4R,5R)-5-Azido-2,3-di-O-benzyl-4-C -hydroxy-
methyl-cycloheptanol (11). NaOCH3 (0.079 g, 1.47 mmol, 2 eq)
was added to a solution of 10 (0.39 g, 0.72 mmol) in dry MeOH
(10 mL) under an atmosphere of nitrogen. The reaction mixture
was stirred overnight at rt before diluted with EtOAc (50 mL)
and the organic phase was washed with H2O (3 ¥ 50 mL), dried
over MgSO4, filtered and concentrated. The crude product was
purified by flash column chromatography (pentane/EtOAc 2 : 1)
to give the desired diol 11 (0.24 g, 85%). [a]2D0 -2.3 (c 1.0, CHCl3).
General experimental information
Organic synthesis
All reagents except otherwise stated were used as purchased with-
out further purification. Enzymes were purchased from Sigma.
Dried glassware from the oven (ca. 120 ◦C) was used for reac-
tions carried out under nitrogen or argon atmosphere. Solvents
were dried using MB-SPS Solvent Purification System. Flash
chromatography was performed with Merck silica 60 (230–400
mesh) as stationary phase. TLC was performed on silica-coated
aluminium plates (Merck 60 F254). TLC plates were first observed
in UV-light and then visualized with ceric sulfate/ammonium
molybdate in 10% H2SO4 stain. 1H NMR (400 MHz) and
13C NMR (100 MHz) were recorded at a Varian Mercury 400
spectrometer. CDCl3 (d 7.26 ppm (CHCl3) for proton and d 77.16
ppm for carbon resonances) and D2O (d 4.79 ppm for proton)
were used as internal references. Spectra were assigned based on
gCOSY, gHMQC and DEPT-135 experiments. MS spectra were
recorded at a Micromass LC-TOF instrument by using electro-
spray ionization (ESI). High resolution spectra were recorded
with one of the following compounds as internal standard: (Boc-L-
alanine: C8H15NO4Na: 212.0899; BzGlyPheOMe: C19H20N2O4Na:
363.1321; BocSer(OBn)SerLeuOMe: C25H39N3O8Na: 532.2635;
erythromycin: C37H67NO13Na: 756.4510) Masses of standards and
1
Rf (pentane/EtOAc 3 : 1) 0.24. H NMR (400 MHz, CDCl3): dH
(ppm) 7.38–7.29 (m, 10H, ArH), 4.82 (d, 1H, J 11.2 Hz, CH2Ph),
4.76 (d, 1 H, J 11.2 Hz, CH2Ph), 4.63 (d, 2H, J 11.2 Hz, CH2Ph),
3.83–3.75 (m, 2H, H1, H8), 3.72–3.65 (m, 3H, H2, H3, H8¢), 3.59
(m, 1H, H5), 3.19 (d, 1H, J 4.4 Hz, OH1), 2.33 (t, 1H, J 6.0
Hz, OH8), 2.09 (m, 1H, H6), 1.95 (m, 1H, H4), 1.89–1.75 (m,
3H, H6¢,H7,H7¢). 13C NMR (100 MHz, CDCl3): dC (ppm) 137.8,
137.4, 128.9–128.0 (ArC), 85.0, 79.8 (C2, C3), 74.8, 74.5 (CH2Ph),
72.2 (C1), 63.1 (C8), 61.1 (C5), 49.9 (C4), 27.7, 26.1 (C6, C7), 21.1
(OCOCH3). HRMS(ES+): calcd. for C22H27N3O4Na: 420.1899,
found 420.1898.
(1S,2R,3S,7R) Benzyl 7-azido-2,3-di-O-benzyl-4-oxocyclo-
heptanecarboxylate (13). The diol 11 (0.21 g, 0.52 mmol) was
dissolved in acetone (8 mL) and cooled to 0 ◦C. Jones reagent
(1.2 mL; Jones reagent: 2.67 g of CrO3 in concentrated sulfuric
acid (2.3 mL) and then diluted with H2O to 10 mL) was added in
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 7807–7813 | 7811
©

View Article Online
three portions over 30 min. The reaction mixture was stirred at 0 ◦C
N2. Pearlman’s catalyst (36 mg) was again added and the reaction
mixture stirred at rt for another 8 h under an atmosphere of H2
(balloon) before being filtered through Celite, concentrated to
give the desired product 7 (26 mg, 92%) as a clear oil. [a]2D0 14 (c
1.0, H2O). Rf (EtOAc/isopropanol/water 1 : 2 : 1) 0.29. 1H NMR
(400 MHz, D2O): dH (ppm) 3.85 (t, 1H, J 10.4 Hz, H3), 3.68
(m, 1H, H5), 3.54 (d, 1H, J 8.4 Hz, H2), 2.50 (m, 1H, H4), 2.07
(m, 2H, H6, H7), 1.73 (m, 1H, H6 or H7), 1.60 (m, 1H, H7 or
H6). 13C NMR (100 MHz, D2O): dC (ppm) 176.7 (COOH), 92.1
(C1), 76.8, 70.8 (C2, C3), 54.7, 54.6 (C4, C5), 27.5, 23.3 (C6, C7).
HRMS(ES+): calcd. for C8H14NO5: 204.0872, found 204.0870.
for 2 1 h. i-PrOH (25 mL) and H2O (60 mL) were then added before
2
the acetone was removed under reduced pressure and the aqueous
phase extracted with EtOAc (5 ¥ 60 mL) and the combined organic
layers dried (MgSO4), filtered and concentrated. The resulting
crude product (12) (0.20 g) was dissolved in dry CH3CN (8 mL)
and Cs2CO3 (249 mg, 0.76 mmol, 1.5 eq) and BnBr (0.20 mL,
1.50 mmol, 3 eq) were added and the reaction mixture heated to
60 ◦C for 15 min. Then the reaction mixture was cooled to rt and
diluted with EtOAc (50 mL). The organic phase was washed with
H2O (3 ¥ 50 mL), dried over MgSO4, filtered and concentrated.
The resulting oil underwent flash column chromatography in
pentane/EtOAc 20 : 1 → 10 : 1 which gave the desired benzyl ester
13 (174 mg, 67%) over two steps. [a]2D0 13.5 (c 1.0, CHCl3). Rf
(1R,2S,3R,4S,5R)-8-Ethyl-3,4,5-tri-hydroxy-8-azabicyclo
[3.2.1]octane-2-carboxylic acid (15). Treatment of 14 (47 mg, 0.1
mmol) in a similar manner as described for 7 using an old batch
of Pd(OH)2 resulted in formation of the unexpected compound
15 after stirring for 10 days. Repeatedly filtration through Celite,
concentration under reduced pressure and treatment of the result-
ing residue with another portion of Pd(OH)2 in EtOH/EtOAc
was conducted. The crude product was purified by flash column
chromatography (EtOAc/isopropanol/water 1 : 1 : 1) to give com-
pound 15 (11 mg, 49%). Rf (EtOAc/isopropanol/water 1 : 1 : 1)
0.41. 1H NMR (400 MHz, D2O): dH (ppm) 3.96 (m, 1H, H5), 3.90
(dd, 1H, J 10.4 Hz, J 8.8 Hz, H3), 3.75 (d, 1 H, J 8.8 Hz, H2),
3.27 (m, 1H, H9), 2.89 (m, 1H, H9¢), 2.66 (dd, 1H, J 10.4 Hz, J 3.2
Hz, H4), 2.28–2.10 (m, 2H, H6, H7), 1.89–1.75 (m, 2H, H6¢,H7¢),
1.27 (t, 3H, J 7.2 Hz, H10). 13C NMR (100 MHz, D2O): dC (ppm)
176.4 (COOH), 96.9 (C1), 74.9, 70.6, (C2, C3), 58.0, 51.5 (C4, C5),
39.4 (C9), 26.1, 21.1 (C6, C7), 10.3 (C10). HRMS(ES+): calcd. for
C10H17NO5Na: 254.1004, found 254.1003.
1
(pentane/EtOAc 10 : 1) 0.25. H NMR (400 MHz, CDCl3): dH
(ppm) 7.32–7.18 (m, 15H, ArH), 5.06 (s, 2H, CH2Ph), 4.64 (d,
1H, J 11.2 Hz, CH2Ph), 4.51 (d, 1H, J 11.6 Hz, CH2Ph), 4.47 (d,
1H, J 11.2 Hz, CH2Ph), 4.37 (d, 1H, J 11.6 Hz, CH2Ph), 4.21–4.17
(m, 2H, H2, H7), 4.07 (d, 1 H, J 6.8 Hz, H3), 2.89 (dd, 1 H, J 6.8
Hz, J 9.2 Hz, H1), 2.69 (m, 1H, H5), 2.51 (ddd, 1 H, J 3.6 Hz, J 7.2
Hz, J 15.2 Hz, H5¢), 2.28 (m, 1H, H6), 2.05 (m, 1H, H6¢). 13C NMR
(100 MHz, CDCl3): dC (ppm) 208.0 (CO), 170.8 (CO2Bn), 137.3,
136.6, 135.4, 128.6–127.9 (ArC), 86.8 (C3), 78.2 (C2), 73.5, 72.8
(CH2Ph), 67.2 (COOCH2Ph), 61.2 (C7), 54.0 (C1), 37.1 (C5), 27.7
(C6). HRMS(ES+): calcd. for C29H29N3O5Na: 522.2005, found
522.2000.
(1R,2S,3R,4S,5R) Benzyl 3,4-Di-O-benzyl-5-hydroxy-8-aza-
bicyclo[3.2.1]octane-2-carboxylate (14). PPh3 (0.21 g, 0.77 mmol,
2.5 eq) was added to a stirred solution of ketone 13 (0.16 g, 0.31
mmol) in THF (5 mL) and H2O (0.5 mL). The reaction mixture
was stirred at 40 ◦C for 4 h before it was concentrated and the crude
product purified by flash column chromatography (CH2Cl2/Et2O
4 : 1→ 2 : 1 → 1 : 1 → 0 : 1) to give the desired product 14 (0.11 g,
78%) as a colourless oil. [a]2D0 -5.0 (c 1.0, CHCl3). Rf (EtOAc) 0.53.
1H NMR (400 MHz, CDCl3): dH (ppm) 7.40–7.25 (m, 13H, ArH),
7.17–7.15 (m, 2H, ArH), 5.17 (d, 1H, J 12.0 HZ, CH2Ph), 5.10 (d,
1 H, J 12.0 Hz, CH2Ph), 5.02 (d, 1H, J 11.2 Hz, CH2Ph), 4.83 (d,
1H, J 10.8 Hz, CH2Ph), 4.79 (d, 1 H, J 11.2 Hz, CH2Ph), 4.65 (d,
1H, J 10.8 Hz, CH2Ph), 3.99 (dd, 1H, J 8.0 Hz, J 10.0 Hz, H3),
3.57 (dd, 1H, J 6.8 Hz, J 3.2 Hz, H1), 2.74 (dd, 1H, J 10.0 Hz, J 3.2
Hz, H2), 2.30 (m, 1H, H6), 1.96 (m, 1H, H7), 1.76 (m, 1H, H7¢),
1.52 (m, 1H, H6¢). 13C NMR (100 MHz, CDCl3): dC (ppm) 171.1
(CO2Bn), 139.0, 138.7, 135.7, 128.7–127.6 (ArC), 92.6 (C5), 87.5
(C4), 79.1 (C3), 75.5, 74.9, 66.8 (CH2Ph), 55.6 (C2), 54.1 (C1), 30.9
(C6), 24.3 (C7). HRMS(ES+): calcd. for C29H31NO5Na: 496.2100,
found 496.2095.
(1R,2S,3R,4S,5R)-3,4,5-tri-hydroxy-8-(4-hydroxybutyl)-8-aza-
bicyclo[3.2.1]octane-2-carboxylic acid (16). A solution of 14 (40
mg, 0.08 mmol) in 50% aqueous THF was treated with Pd(OH)2
(old batch) under an atmosphere of hydrogen. The reaction
mixture was stirred for 5 days where repeatedly filtration through
Celite, concentration under reduced pressure and treatment of the
resulting residue with another portion of Pd(OH)2 in THF/water
was conducted. The product isolated after flash column chro-
matography (EtOAc/isopropanol/water 1 : 2 : 1) was the unex-
pected compound 16 (14 mg, 58%). Rf (EtOAc/isopropanol/water
1
1 : 1 : 1) 0.44. H NMR (400 MHz, D2O): dH (ppm) 3.95 (m, 2H,
H3, H5), 3.80 (d, 1H, J 8.8 Hz, H2), 3.67 (t, 2H, J 6.4 Hz, H12),
3.17 (m, 1H, H9), 2.86 (m, 1H, H9¢), 2.75 (dd, 1H, J 12.0 Hz, J 4.0
Hz, H4), 2.30–2.15 (m, 2H, H6, H7), 1.90–1.63 (m, 6H, H6¢,H7¢,
H10, H10¢, H11, H11¢). 13C NMR (100 MHz, D2O): dC (ppm)
176.6 (COOH), 96.0 (C1), 74.5, 70.5, (C2, C3), 61.1 (C12), 58.3,
51.1 (C4, C5), 43.8 (C9), 29.0, 26.2, 22.3, 21.1 (C6,C7,C10,C11).
HRMS(ES+): calcd. for C12H22NO6: 276.1447, found 276.1449.
(1R,2S,3R,4S,5R)-3,4,5-tri-hydroxy-8-azabicyclo[3.2.1]octane-
2-carboxylic acid (7). The tribenzylated compound (14) (65 mg,
0.14 mmol) was dissolved in EtOH (4 mL) and EtOAc (2 mL)
and flushed with N2. Pearlman’s catalyst (30 mg) was added and
the reaction mixture was stirred at rt under an atmosphere of H2
(balloon) for 8 h before the reaction mixture was filtered through
Celite and concentrated. The resulting residue (ester deprotected)
was re-dissolved in EtOH (4 mL) and flushed with N2. Pearlman’s
catalyst (42 mg) was added and the reaction mixture stirred at
rt under an atmosphere of H2 (balloon) overnight. The mixture
was then filtered through Celite, concentrated and the residue
was dissolved in EtOH (4 mL) and H2O (1 mL) and flushed with
Inhibition studies
Inhibition constants (Ki’s) were determined by measuring initial
rates (<10% of substrate conversion) using p-nitrophenyl b-D-
1
glucuronide at seven concentrations ranging from 4 - 4 times
KM monitoring at 400 nm with A < 1 using either a Varian
Cary 100 Bio UV-vis Spectrophotometer or a PerkinElmer 2300
EnSpire Multilabel Plate Reader. Reduced volume Polystyren
cuvettes were used for spectrophotometer measurements while
PolySorb (flat bottom) Immuno 96 MicroWell Solid Plates from
7812 | Org. Biomol. Chem., 2011, 9, 7807–7813
This journal is
The Royal Society of Chemistry 2011
©

View Article Online
Nunc was used for plate reader measurements. Measurements
were conducted in 50 mM phosphate buffer (pH 6.8) at 37 ◦C
for 2 min for E. coli b-glucuronidase using a continuous assay.
Measurements of inhibition of bovine liver b-glucuronidase were
conducted in 50 mM sodium acetate buffer (pH 4.6) at 37 ◦C
for 2 min using a stopped assay. In this context it should be
mentioned that the enzyme wasn’t active at the same pH in
citrate buffer. Temperature equilibration was conducted prior to
enzyme addition and monitoring of reaction rates. K¢M and KM are
Michaelis–Menten constants with or without inhibitor present.
K¢M and KM were obtained from Hanes plot, which was also used
to ensure that inhibition was competitive. Ki values were calculated
as Ki = [I]/((K¢M/KM) - 1) having [I] ª Ki. For Hanes plots, see
the ESI.‡
Ichikawa, Y. Igarashi, M. Ichikawa and Y. Suhara, J. Am. Chem. Soc.,
1998, 120, 3007–3018.
3 (a) M. Bols, R. G. Hazell and I. B. Thomsen, Chem.–Eur. J., 1997, 3,
940–947; (b) H. H. Jensen and M. Bols, J. Chem. Soc., Perkin Trans. 1,
2001, 905–909; (c) H. H. Jensen, A. Jensen, R. G. Hazell and M. Bols,
J. Chem. Soc., Perkin Trans. 1, 2002, 1190–1198.
4 H. Liu, X. Liang, H. Søhoel, A. Bu¨low and M. Bols, J. Am. Chem.
Soc., 2001, 123, 5116–5117.
5 (a) V. H. Lillelund, H. H. Jensen, X. Liang and M. Bols, Chem. Rev.,
2002, 102, 515–553; (b) Iminosugars: From Synthesis to Therapeutic
Applications, ed. P. Compain and O. R. Martin, Wiley-VCH, 2007.
6 (a) A. Goldmann, M.-L. Milat, P.-H. Ducrot, J.-Y. Lallemand, M.
Maille, A. Lepingle, I. Charpin and D. Tepfer, Phytochemistry, 1990, 29,
2125–2127; (b) N. Asano, A. Kato, M. Miyauchi, H. Kizu, Y. Tomimori,
K. Matsui, R. J. Nash and R. J. Molyneux, Eur. J. Biochem., 1997, 248,
296–303; (c) B. Dra¨ger, Nat. Prod. Rep., 2004, 21, 211–223.
7 T. S. Rasmussen and H. H. Jensen, Org. Biomol. Chem., 2010, 8, 433–
441.
8 T. S. Rasmussen, S. Allman, G. Twigg, T. D. Butters and H. H. Jensen,
Bioorg. Med. Chem. Lett., 2011, 21, 1519–1522.
9 B. D. Wallace, H. Wang, K. T. Lane, J. E. Scott, J. Orans, J. S. Koo, M.
Venkatesh, C. Jobin, L.-A. Yeh, S. Mani and M. R. Redinbo, Science,
2010, 330, 831–835.
10 J. C. Lee, Y. S. Oh, S. H. Cho and J. D. Lee, Org. Prep. Proced. Int.,
1996, 28, 480–483.
11 P. D. Bailey, M. A. Beard, H. P. T. Dang, T. R. Phillips, R. A. Price and
J. H. Whittaker, Tetrahedron Lett., 2008, 49, 2150–2153.
12 Y. Yoshimura, C. Ohara, T. Imahori, Y. Saito, A. Kato, S. Miyauchi,
I. Adachi and H. Takahata, Bioorg. Med. Chem., 2008, 16, 8273–
8286.
13 D. L. Zechel and S. G. Withers, Acc. Chem. Res., 2000, 33, 11–18.
14 T. Niwa, T. Tsuruoka, S. Inouye, Y. Naito, T. Koeda and T. Niida, J.
Biochem., 1972, 72, 207–211.
15 T. M. Gloster, R. Madsen and G. J. Davies, ChemBioChem, 2006, 7,
738–742.
Acknowledgements
We thank Mr Mikkel T. Jensen for technical support. We are
furthermore grateful for financial support from the Faculty of
Natural Sciences, OChem Graduate School, Aarhus University,
The Danish National Science Research Foundation, The Lund-
beck Foundation, and DCSC.
References
1 Iminosugars as Glycosidase Inhibitors – Nojirimycin and Beyond, A. E.
Stu¨tz, ed., Wiley-VCH, 1999.
2 (a) T. M. Jespersen, W. Dong, M. R. Sierks, T. Skrydstrup, I. Lundt
and M. Bols, Angew. Chem., Int. Ed. Engl., 1994, 33, 1778–1779; (b) Y.
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 7807–7813 | 7813
©