Synthesis of 3, 5-diazabicyclo [5.1.0] octenes. A new platform to mimic glycosidase transition states

Article abstract of DOI:10.1016/j.tet.2010.05.064

All-cis 1-hydroxymethyl 2, 3 bis-aminomethyl cyclopropane was used to construct the first 3, 5- diazabicyclo [5.1.0]-3-octenes. This system has the interesting ability to exist in a conformation that resembles a snapshot of a glycoside hydrolysis reaction

Full text of DOI:10.1016/j.tet.2010.05.064

Tetrahedron 66 (2010) 5566e5572
Contents lists available at ScienceDirect
Tetrahedron
journal homepage: www.elsevier.com/locate/tet
Synthesis of 3,5-diazabicyclo [5.1.0] octenes. A new platform to mimic glycosidase
transition states
,
Fedra M. Leonik b, Ion Ghiviriga a, Nicole A. Horenstein a
*
a Department of Chemistry, Box 117200, University of Florida, Gainesville, FL 32611-7200, USA
b Adesis, Inc. 27 McCullough Drive, New Castle, DE 19720, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
All-cis 1-hydroxymethyl 2,3 bis-aminomethyl cyclopropane was used to construct the first 3,5-
diazabicyclo [5.1.0]-3-octenes. This system has the interesting ability to exist in a conformation that
resembles a snapshot of a glycoside hydrolysis reaction with respect to charge and geometric analogy to
an oxocarbenium ion, and the positioning of the departing aglycon. The cis-configured cyclopropane core
was synthesized by Cu-catalyzed intramolecular cyclopropanation of benzyl protected cis-2-butene-1,4-
diol diazoacetate ester. Serial functionalization to bis-aminomethyl cyclopropanes and subsequent
cyclization to amidines lead to the target bicyclic compounds in good overall yields. Several glycosidases
were surveyed for the inhibitory potential of these transition state analogs, and amongst them, selective
competitive inhibitors with micromolar Ki values were identified.
Received 16 March 2010
Received in revised form 15 May 2010
Accepted 17 May 2010
Available online 11 June 2010
Ó 2010 Elsevier Ltd. All rights reserved.
1. Introduction
a trajectory nearly orthogonal to the plane defining the carbocation
atoms (Fig. 1).
As an enzyme guides reacting substrate ensembles over the
reaction coordinate, it provides a dynamic complementarity to af-
ford low energy pathways leading to products.1,2 A significant
component to the rate acceleration has been ascribed to transition
state complementarity,3 whereby favorable binding interactions
between enzyme and the fleeting transition state provide a de-
crease in the activation barrier. As this concept evolved further, the
observed tight binding of stable chemical compounds that were
analogs of the transition state was rationalized on the basis of the
favorable complementarity between the transition state analog and
the enzyme.4 A number of potent inhibitors has been identified that
presumably operate in this way.5e7 This is never a simple matter;
an implicit design problem is that stable ground state molecules
only approximate the electronic and geometric configuration of
a transition state. For example, consider SN2-like (e.g., ANDN type)8
reactions, there is a degree of pentavalent character at the elec-
trophilic carbon that depends on nucleophile and leaving group
bond orders, and the timing, or symmetry of bond breaking and
formation. Likewise, in a dissassociative SN1-like e.g., [DNþAN]8
reaction that generates carbocationic transition states or in-
termediates, (or the corresponding capture of these species by
nucleophile in the reverse reaction) the leaving group departs in
R
R
ANDN
Nu
LG
Nu:
products
R
LG
R
R
R
R
R
[DN + AN]
(first step)
LG
intermediate
R
LG
R
R
R
Figure 1. Transition structures for hypothetical associative and disassociative
reactions. Note the non-standard geometries at the central carbon atom.
In both of these reaction types, the bonding between the
leaving group, nucleophile, and central carbon atom have non-
ground state angular geometries and partial bonds having lengths
greater than found for stable ground state molecules. Mimicry of
these structures using direct connectivity with first or second
row elements is difficult, if not impossible. We sought to develop
an approach to circumvent the geometric limitations imposed by
direct connectivity. Our initial focus is on glycosidase and
* Corresponding author. Tel.: þ1 352 392 9859; e-mail address: horen@chem.ufl.
edu (N.A. Horenstein).
0040-4020/$ e see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tet.2010.05.064

F.M. Leonik et al. / Tetrahedron 66 (2010) 5566e5572
5567
glycosyltransferase enzymes, responsible for creation of carbo-
hydrate structures of incredible diversity and complexity.9 The
creation and hydrolysis of the glycosidic bond are central to the
extremely broad biological functions found for carbohydrates,10
making glycosyl transfer enzymes important targets for in-
hibition.11 The reaction mechanisms for glycosidases vary in de-
tail with respect to acid/base chemistry, and the nature of
nucleophilic participation.12,13 However, with few exceptions14
a common mechanistic element is that as the glycosidic bond is
being cleaved, the sugar glycon develops oxocarbenium ion
character. This places the departing aglycon in a Burgi/Dunitz like
trajectory15,16 over the glycon ring, which is flattened and posi-
tively charged.17,18 Figure 2 shows a conceptual model for in-
hibitor design whereby the transition state is simulated with
a rigid bicyclic ring system used to position an aglycon analog
over the plane of a positively charged ring, analogous to the
flattened oxocarbenium ion like transition state of typical glyco-
hydrolase enzymes.
which to attach an analog of a leaving group aglycon. Earlier work
has explored geometric and electronic aspects of this design.22,23
The compounds we report here are an interesting counterpoint to
the calystegines, which are glycosidase inhibitors with a bridged
bicyclic nortropane skeleton.24 Presumably the bicyclic framework
of the calystegines serves to rigidify the ring and hydroxyl group
positions, whereas, the design presented in this work utilizes a bi-
cyclic diaza-framework to position the aglycon and provide delo-
calized charge mimicry.
2. Results and discussion
2.1. Synthesis of bicyclic amidines
We envisioned arriving at our target [5.1.0] diaza bicyclooctenes
by cyclization of bis-amino all-cis cyclopropanes. Scheme 1 pres-
ents the route for synthesis of the key intermediate bis-amino-
methyl cyclopropane 8.
aglycon
analog
aglycon
HO
O
HO
oxocarbenium
δ+
HO
analog
+
HO
glycosyl transfer transition state
stable equivalent
Figure 2. A mimic (right) for disassociative glycosyl transfer transition states (left).
With the exploration of this design strategy, we expand the
well-known ability of sugar analogs with basic nitrogen atoms to
inhibit glycosidases.7,19,20 Further, inclusion of the aglycon or a nu-
cleophile into the inhibitor structure with transition state geo-
metric constraint affords opportunities to increase binding affinity
and direct specificity. In the present study, we describe synthesis of
parent molecules that could serve to validate this concept and
provide the basis for synthesis of more elaborate compounds.
Figure 3 presents our targets, substituted 3,5-diazabicyclo-[5.1.0]
octanes, which to the best of our knowledge have not been repor-
ted. They are both geometric and charged analogs of glycosidase
transition states. The amidinium portion of the ring mimics the
charge and shape of a glycosyl oxocarbenium ion,21 because it is
flattened, positively charged, and has delocalization of charge. The
all-cis cyclopropyl ring system creates an oriented tether with
cis-2-Butene-1,4-diol 1 was converted to the monobenzylated
derivative 2 using standard methods.25 Subsequent reaction of the
free hydroxyl group with diketene gave the unsaturated
b-keto
ester 3 in 86% yield.26 Reaction of 3 with p-ABSA (p-acetamido
benzenesulfonyl azide) in Et3N gave the desired cis-diazoacetic
ester 4.26 A key step in this synthetic design was the intramolecular
cyclopropanation of 4 to give an all-cis trisubstituted cyclic system.
With intramolecular cycloaddition, only one fused bicyclic ring is
formed due to geometric constraints.27 Hence, starting from a cis-
alkene, the resulting fused cyclopropyl lactone 526 will have all
substituents in an all-cis relative configuration. Several cyclo-
propanation catalyst systems were explored, leading to identifica-
tion of copper triflate and Evan’s bis-oxazoline valinol ligand28 as
being optimal, producing (þ)-5 in 60% yield. The ee obtained for 5
was only 22%,26 but was fully acceptable for this work given that
our target and intermediates were meso compounds. The lactone
ring of 5 was reductively opened with LiAlH4 to give the meso diol 6
in 74% yield. Conversion to the bis-azide 7 was smoothly effected by
mesylation of 6 and displacement with azide in DMF at 60 ꢀC. The
diazide 7 was used without further purification for synthesis of
diamine 8 in (95% yield, two steps) by triphenylphosphine medi-
ated reduction. All-cis cyclopropanes related to compound 8 have
recently been discussed as potential tripodal ligands for Pd-based
catalysts, so compounds like 8 may be of further interest for crea-
tion of new Pd ligands.29
A-
AH
H
NO2
O
OH
O
O
NH
δ+
HO
HO
H
N
H
OH
O-
O-
O
O
Scheme 2 presents the conversion of 8 into different bicyclic
amidines 9e13. Initial attempts to synthesize compound 9 using
neat trimethyl orthoformate/HCl instead lead to bis-formamide 14
based on MS and 500 MHz NMR analyses. While cyclization with
1 equiv of trimethyl orthoformate proved to afford 9 in 46% yield,
we sought an alternate method to give improved yields. Reaction
Figure 3. Archetypical glycosidase transition state and bicyclic amidinium analog. On
the left is a model for an enzyme bound glycosidase transition state featuring oxo-
carbenium ion character and the leaving group orthogonal to the
p-system of the
oxocarbenium ion. The bicyclic amidine on the right places an analog of a phenolic
leaving group aglycon above the amidine plane that mimics the electronics and flat
geometry of an oxocarbenium ion.

5568
F.M. Leonik et al. / Tetrahedron 66 (2010) 5566e5572
O
O
Benzylbromide
NaEtO/THF
p-ABSA
CH3CN / Et3N
OBn
O
O
O
OH
OH
OH
OBn
NaAcO / THF
61%
CuOTf
N2
O
O
O
O
Evans Ligand
60 %
OBn
79 %
OBn
4
5
1
2
3
OBn
H
1) MsCl / Et3N
2) NaN3 / DMF
H
H
NH2
NH2
PPh3 / H2O / Δ
LiAlH4 / THF
79 %
OH
OH
N3
N3
BnO
O
BnO
BnO
92 %
H
70 %
O
H
H
5
8
6
7
Scheme 1. Synthesis of bis-aminomethyl cyclopropane 8.
NH2
OBn
OH
Ammonium
formate
Pd/C
H
O
H
NH2
NH2
EtOH
NH
H
BnO
H
H
88 %
H
N
N
NH
8
10
CH(OEt)3
HCl
40 %
9
NH2
OBn
O
OH
O
Ammonium
formate
Pd/C
OH
NH
H
HN
HN
H
H
H
EtOH
70 %
H
BnO
NH
OH
N
N
OH
H
11
14
12
O
Scheme 2. Synthetic route to the [5.1.0] diaza bicyclooctenes.
of 8 with the appropriate imidate ester hydrochloride30 in
refluxing absolute ethanol lead to bicyclic amidines 9 and 11 in
88% and 70% yields, respectively. The relative stereochemistry of
amidine 9 was analyzed by NOESY 2D NMR in order to confirm the
all-cis configuration. A strong cross peak was observed between
Ha (3.70 ppm) and Hb (3.93 ppm), confirming the stereochemistry
we anticipated based on intramolecular cyclopropanation of cis-
butene 4.
the disposition of the benzyl group with respect to it. The exo,exo
conformer shown in Figure 4 was the global minimum based on
a search through rotational itinerary of all CeC non ring bonds and
alternate ring puckers on the seven-membered portion of the
[5.1.0] system. The exo,exo designation refers to the exo-ring
pucker and the exo-disposition of the benzyl group with respect to
the fused ring system. The transition-state like endo,endo con-
former was a local minimum, estimated to be 2.2 kcal/mol higher
in energy than the exo,exo conformer. While it suffers from
a modestly higher angle strain term, 9-endo,endo does enjoy a fa-
vorable dipole/charge term that we ascribe to the non-bonding
electrons of the benzyl oxygen and the positive charge of the
amidinium moiety given their close proximity. We suggest that 9-
endo,endo is an energetically accessible conformation, relevant for
enzymatic binding. Effective enzymatic recognition does not re-
quire that the global minimum in solution be the bound species,
especially when energetically close local minima may bind with
interactions that offset a small solution penalty.31
OBn
Ha
Ha
Hb
Hb
NH
N
Hydroxymethyl amidine 10, and bis-hydroxymethyl amidine 12
were synthesized by catalytic transfer hydrogenation of com-
pounds 9 and 11, respectively, giving in both cases quantitative
conversion. The new amidines were found to be stable in aqueous
solution near neutral pH for at least 24 h, based on NMR analyses.
Another point about the bicyclic amidines is that although they are
racemic in their neutral form, they are protonated in aqueous
media (see enzymology below) and are therefore meso compounds
in water at neutral pH.
9-endo,endo
9-exo,exo
A molecular mechanics conformational analysis of compound
9 was performed to investigate the overall ring conformation and
Figure 4. The transition state like conformation and global minimum for compound 9
via molecular mechanics.

F.M. Leonik et al. / Tetrahedron 66 (2010) 5566e5572
5569
Table 2
Inhibition constants, Ki,
2.2. Inhibition of glycosidases
m
M, and Km/Ki for compounds 9, 11, 12 at pH 7.5
The bicyclic amidines 9e12 were screened against a small panel
- and -glycosidases to yield the data reported in Table 1. All of
Compound
b
-gal/A. orizae Ki; Km/Ki -glc/almond Ki; Km/Ki
b
of
a
b
9
11
12
nd
270ꢃ40; 18
150ꢃ30; 32
nd
760ꢃ90; 2.1
the glycosidases utilized in these experiments were retaining en-
zymes, and the appropriate p-nitrophenyl (PNP)-glycosides were
used as substrates. The inhibition screening data are presented as
percentage inhibition, as expressed by Eq. 1, where n0 is the initial
velocity in the absence of inhibitor, ni is the initial velocity in the
presence of inhibitor.
1500ꢃ100; 1.1
nd¼not determined
The estimated Ki for 11 is 150ꢃ30
competitive. Removing the hydroxymethyl group (compound 9,
Ki¼270ꢃ40 M) or the benzyl aglycon mimic (compound 12)
results in decreased inhibition.(Tables 1 and 2) The 2-fold higher Ki
mM, and the inhibition mode is
m
% inhibition ¼ ðv0 ꢁ viÞ=v0 ꢂ 100
(1)
Table 1
Inhibition screens for compounds 9e12
Compound
a
-Galactosidase
a
-Glucosidase
b-Galactosidase
b
-Galactosidase
b-Glucosidase
from green
from S. cerevisiae (%)
from A. orizae (%)
from E. coli (%)
from almonds (%)
coffee beans (%)
pH 7.5
pH 6
pH 7.5
pH 6
pH 7.5
pH 6
pH 7.5
pH 6
pH 7.5
pH 6
9
2
d
d
4
d
2
d
d
3
16
11
40
34
16
d
16
d
10
3
10
10
d
d
13
5
83
45
87
46
65
24
71
26
10
11
12
2
d
d
d
3
d
d
The data are expressed as % inhibition, as defined in the text. A dash (d) indicates no inhibition was observed.
In the experiments the concentration of substrate and inhibitor
were both set to the substrate Km, which would give a reasonably
sensitive probe for effective inhibition.
for 9 versus 11 for almond b-glucosidase suggests that the
hydroxymethyl group of 11 engages in an active site hydrogen
bond, the nature of which remains to be determined. Inhibition of
The screening data of Table 1 demonstrate that
and -glucosidase were not inhibited significantly by any of the
compounds. Table 1 also reveals that almond -glucosidase was
a
-galactosidase
almond b-glucosidase was facilitated when compounds included
a benzylic group analogous to the departing aglycon of the PNP
substrate as evidenced by the data for compound 11 versus 10 or 12.
a
b
inhibited to varying degree by all of the compounds, which is in
accord with its broad specificity.32 However, the data presented in
Table 1 also suggest that b-glucosidase inhibition is promoted by
inclusion of aglycon mimicry in the amidine inhibitors. Compari-
son of inhibition data for 9 versus 10, and 11 versus 12, shows
a halving of the apparent inhibition when the benzyl functionality
is absent. From the comparison between molecules 9 and 11, it was
observed that including a hydroxyl group doubled the potency of
the amidines TS analog, indicating a modest contribution to
binding.
One general trend seen in the data for almond b-glucosidase is
that the inhibition appears to be stronger at pH 7.5 than at pH 6.0.
This is unlikely to be a reflection of an unusually low amidine pKa,
because our titrametric estimate of the pKas for 9 and 11 placed
them over 10.0. It is possible that the pH behavior observed here
reflects a deprotonated state of the enzyme with better electro-
statics for binding the amidines at higher pH. A deeper in-
vestigation and interpretation of the pH behavior will be of interest,
but is beyond the scope of the present study. As most recently
discussed by Davies, analysis of glycosidase inhibitor binding is
exceptionally complex, impacted in varying ways by enthalpic and
entropic factors.33
Figure 5. Lineweaver/Burk plot for almond
substrate and compound 11 at concentrations of: :, 4 mM; -, 2 mM; ꢄ, 0 mM.
Ki¼150ꢃ30 M.
b-glucosidase with b-PNP-glucoside as
m
The trends observed in the inhibition data for the A. orizae
-galactosidase parallel those for the almond -glucosidase. In
particular, compound 11 is the best inhibitor, with a Ki of
M, and loss of the hydroxymethyl or benzylic function-
The three
b-glycosidases tested showed diverse responses to
b
b
compounds 9e12 under screening conditions, enabling identifica-
tion of enzyme/inhibitor combinations for characterization. Com-
pounds 11 and 12 were selected for further kinetic analysis with the
760ꢃ90
m
alities (Ki¼1,500ꢃ100
mM) results in poorer binding interactions.
b
-galactosidase from Aspergillus oryzae and compounds 9 and 11
were selected for analysis with almond -glucosidase. In those
b
experiments, initial velocity measurements were obtained over
a range of substrate and inhibitor concentrations, and the data were
found to be well-fit to the equation for competitive inhibition. The
estimated Ki values and Km/Ki ratios are presented in Table 2.
The double reciprocal plot for the inhibition kinetics of com-
3. Conclusions
We have described the first syntheses of diazabicyclo [5.1.0]
octenes with an all-cis configuration about the cyclopropyl ring and
an overall endo configuration for the system. This ring system has
been used to develop a prototype glycosidase inhibitor design in
pound 11 with b-glucosidase from almonds is presented in Figure 5.

5570
F.M. Leonik et al. / Tetrahedron 66 (2010) 5566e5572
which the planar amidine functionality serves as an electronic and
geometric mimic of a glycosyl oxocarbenium ion. The overall endo
configuration allows incorporation of leaving group (aglycon)
mimicry into the molecule, but importantly, allows it to assume
a position over the amidine group that mimics the hypervalent
shape and trajectory a departing aglycon enjoys relative to the
flattened oxocarbenium ion-like glycon. We have identified selec-
with brine (60 ml). The organic phase was dried over MgSO4 and
evaporated under vacuum. The crude oil was purified by flash
chromatography (silica, 6:1 petroleum ether/AcOEt) to give 3 in
60% yield (1.80 g). 1H NMR (CDCl3)
d
ppm 2.25 (s, 3H), 3.44 (s, 2H),
4.12 (d, J¼8 Hz, 2H), 4.51 (s, 2H), 4.68 (d, J¼8 Hz, 2H), 5.71 (m, 1H),
5.82 (m, 1H), 7.33 (m, 5H), 13C NMR (CDCl3)
(ppm) 200.7, 167.3,
d
138.4, 131.9, 128.9, 128.2, 126.4, 73.0, 66.1, 61.6, 50.4, 30.6.
tive inhibition of two beta glycosidases. The best Ki was 150
mM, for
compound 11 against almond -glucosidase, with a corresponding
b
4.1.3. Diazo-acetic acid 4-benzyloxy-but-2-enyl ester (4). The title
compound was synthesized using the reported procedure.26 A so-
lution of acetoacetate ester 3 (1.72 g, 6.56 mmol) and Et3N (1.15 ml,
8.47 mmol) in anhydrous acetonitrile (18 ml) was prepared at room
temperature. Then a solution of p-ABSA (2.05 g, 8.54 mmol) in
acetonitrile (18 ml) was added dropwise over a 30 min period. The
reaction mixture was stirred for 3 h, after which time a solution of
3 N LiOH was added and the mixture was left to stand for ap-
proximately 12 h. The reaction mixture was then extracted with
ether/AcOEt 2:1 (3ꢂ50 ml). The organic layers were combined and
extracted with brine (100 ml), dried over MgSO4, and concentrated
by rotary evaporation. The crude product was purified by flash
chromatography (silica, 10:1 petroleum ether/AcOEt) to give 4 as
Km/Ki ratio of 32. What is notable about the 32-fold binding ad-
vantage for competitive inhibitor 11 is that effective micromolar
inhibition has been obtained with an austere substitution pattern
lacking the hydroxylation, that is, found in the substrate, or
found in the 200 m
M competitive inhibitor gluconolactone.32 We
hypothesize that inclusion of an appropriate hydroxylation pattern
about the bicyclic ring system will result in opportunities to lower
Ki values and provide the platform for in-depth analysis of the
nature of inhibition and transition state analogy34 in these
compounds.
4. Experimental section
4.1. General
a yellow oil in a 86% yield (1.40 g). 1H NMR (CDCl3)
(d, J¼6 Hz, 2H), 4.49(s, 2H), 4.70 (d, J¼7 Hz, 2H), 4.71(br s, 1H), 5.70
(m, 1H), 5.78 (m, 1H), 7.31 (m, 5H), 13C NMR (CDCl3)
(ppm) 167.0,
d ppm 4.12
d
Solvents and reagents were purchased from Aldrich and Acros.
Glycosidase enzymes and nitrophenyl glycoside substrates were
purchased from Sigma. The organic solvents were dried overnight
138.4, 131.4, 128.9, 128.3, 127.0, 72.9, 66.1, 61.0, 46.7. IR: 2116s,
1687s. EI HRMS calcd for C13H14N2O3Na (MNaþ): 269.0897, found:
269.0901. The spectroscopic data were consistent with previous
data.25
ꢀ
over CaH2 or 4 A molecular sieves and freshly distilled before use.
NMR spectra were obtained at 300 MHz for 1H and 75 MHz for 13C,
Varian spectrometers (VXR, Gemini, Mercury). Chemical shifts
were referenced to internal tetramethylsilane, 0 ppm in appropri-
ate deuterated solvents. Mass spectra were obtained on a Finnigan
MAT 95Q spectrometer operated in FAB or EI modes. Infrared (IR)
spectra were obtained by deposition of CHCl3 solutions on NaCl
plates followed by evaporation of the solvent. Optical rotations
were recorded on a Perkin/Elmer 241 polarimeter at the sodium D
line. UV/vis spectra were obtained with Beckman DU 640 and
Agilent 8453 spectrometers. Constant temperature was maintained
with an Isotemp 210 from Fisher Scientific. Molecular mechanics
calculations utilized MM2 parameters as implemented within
Chem3D Ultra 6.0 (CambridgeSoft.com).
4.1.4. (1S,5R,6S)-6-(Benzyloxymethyl)-3-oxabicyclo[3.1.0]hexan-2-
one (5). Diazoester 4 (3.00 g, 12.2 mmol) was dissolved in dry
CH2Cl2 (131 ml) to make a 0.09 M solution of the starting material.
This solution was added from an addition funnel to a refluxing
suspension of CuOTf (61 mg, 0.25 mmol) and Evan’s bis-oxazoline
ligand28(prepared as described)35 (80 mg, 0.27 mmol) in 400 ml of
CH2Cl2 over a period of 18 h. When the addition was finished, the
solvent was removed under vacuum and the residue was purified
by flash chromatography (silica, 5:1 hexanes/AcOEt). The bicyclic
product was obtained as a colorless oil in 60% yield (1.60 g). 1H NMR
(CDCl3)
d
ppm 1.83 (m, 1H), 2.32(dd, J¼6, 9 Hz, 1H), 2.39 (dd, J¼6,
7 Hz, 1H), 3.42 (dd, J¼9, 11 Hz, 1H), 3.70 (dd, J¼6, 11 Hz, 1H), 4.24
(d, J¼10 Hz, 1H), 4.41 (dd, J¼6, 10 Hz, 1H), 4.50 (d, J¼12 Hz, 1H), 4.57
4.1.1. Z-4-Benzyloxy-but-2-en-1-ol (2). The title compound was
synthesized by the reported method.25 cis-2-Butene 1,4-diol
(56.7 ml, 690 mmol) was added to a stirred suspension of NaH (60%
dispersion in mineral oil, 5.5 g, 138 mmol) in THF (460 ml) that was
previously cooled to 0 ꢀC. Stirring was continued at room temper-
ature for 30 min. After this, benzyl bromide (14.0 ml, 118 mmol)
was added. The reaction mixture was stirred at room temperature
until starting material was consumed. The THF was removed by
rotary evaporation and the residue was dissolved in ether (400 ml).
The organic layer was washed with water (3ꢂ150 ml) and dried
over MgSO4. Fractional distillation of the crude product at 5 mm Hg
(bp 156e159 ꢀC; lit.25 970.25 ꢀC) gave the desired mono protected
(d, J¼12 Hz, 1H), 7.35 (m, 5H), 13C NMR (CDCl3)
d (ppm) 174.7, 138.2,
128.9, 128.3, 73.8, 66.8, 65.1, 22.9, 21.8. EI HRMS calcd for C13H14O3
20
(Mþ): 218.0943, found: 218.0944. [
a
]
þ5.3 (c 0.55, CHCl3);
D
ee¼22%.
4.1.5. ((1R,2S,3S)-3-(Benzyloxymethyl)cyclopropane-1,2-diyl)dime-
thanol (6). Lactone 5 (1.00 g, 4.59 mmol) was added dropwise to
a suspension of LiAlH4 (1.57 g, 41.3 mmol) in dry THF (46 ml) at
0 ꢀC. After 3 h, total consumption of starting material was ob-
served and the reaction mixture was cautiously quenched with
saturated aqueous NH4Cl. The reaction mixture was diluted with
additional THF and filtered through Celite. The organic layer was
separated and dried over MgSO4. After evaporation of solvent, the
reaction mixture was purified by flash chromatography (silica, 2:1
petroleum ether/AcOEt). Diol 6 (0.80 g) was obtained in 74% yield.
diol in 60% yield (12.9 g). 1H NMR (CDCl3)
d ppm 2.02 (br s, 2H), 4.06
(d, J¼6 Hz, 2H), 4.13 (d, J¼6 Hz, 2H), 4.50 (s, 2H), 5.75 (m, 2H), 7.32
(m, 5H).
1H NMR (CDCl3)
3.70 (m, 2H), 3.78 (m 2H), 4.54 (s, 2H), 7.34 (m, 5H), 13C NMR
d ppm 1.49 (m, 3H), 2.55 (br s, 2H), 3.63(m, 2H),
4.1.2. (Z)-4-Benzyloxy-but-2-enyl 3-oxobutanoate (3). The title
compound was synthesized according to published methodology.26
Mono protected diol 2 (2.00 g, 11.2 mmol) was dissolved in 8 ml of
anhydrous THF containing NaOAc (56 mg, 0.68 mmol). A solution of
diketene (1.0 ml, 12.6 mmol) in 4 ml of THF was added dropwise
over 1 h to the refluxing mixture. The mixture was refluxed for 1 h
until complete consumption of mono protected diol was observed.
The reaction mixture was taken up in ether (40 ml) and washed
(CDCl3)
d (ppm) 138.0, 128.9, 128.4, 128.3, 73.6, 66.6, 59.2, 22.2,
19.5. EI HRMS calcd for C13H19O3 (MþH)þ: 223.1334, found:
223.1330.
4.1.6. ((((1S,2R,3S)-2,3-Bis(azidomethyl)cyclopropyl)methoxy)-
methyl)benzene (7). A mixture of diol 6 (0.38 g, 1.7 mmol) and Et3N
(0.56 ml, 5.08 mmol) in CH2Cl2 (57 ml) was cooled to 0 ꢀC. Freshly

F.M. Leonik et al. / Tetrahedron 66 (2010) 5566e5572
5571
distilled methanesulfonyl chloride was added dropwise (0.40 ml,
5.1 mmol) and the resulting mixture was stirred at 0 ꢀC for 20 min,
at which time all of the starting material was consumed. The re-
action mixture was washed with water, then brine, and dried over
MgSO4. The crude dimesylated product was dissolved in DMF
(35 ml) and NaN3 was added. The reaction was stirred at 60 ꢀC for
4 h and the solvent was evaporated under reduced pressure. The
residue was dissolved in AcOEt (150 ml) and washed with water
(3ꢂ100 ml). The organic phase was dried over MgSO4 and the
solvent removed by rotary evaporation. The product was purified
by flash chromatography (silica, 20:1 petroleum ether/AcOEt)
more starting material was observed. The suspension was filtered
through a Celite pad, which was washed with five portions of
MeOH to maximize recovery of the product. The filtrate was con-
centrated by rotary evaporation. The products were found to be
sufficiently pure for use without further purification.
4.2.1. (1R
*
,7S
*
,8S ,Z)-3,5-Diazabicyclo[5.1.0]oct-3-en-8-ylmethanol
*
(10). This compound was obtained as described in the general
procedure for the deprotection of cyclic amidines as a yellowish
solid in quantitative yield. 1H NMR (CD3OD)
d ppm 1.29 (m,1H),1.70
(m, 2H), 3.48 (dd, J¼6, 10 Hz, 2H), 3.72 (d, J¼8 Hz, 2H), 3.92 (d,
giving a white solid in 73% yield (0.30 g). 1H NMR (CDCl3)
d
ppm
J¼12 Hz, 1H), 3.98 (d, J¼12 Hz, 1H), 7.49 (s, 1H), 13C NMR (CD3OD)
d
1.48 (m, 2H), 1.55 (m, 1H), 3.34 (dd, J¼7,13 Hz, 2H), 3.39 (dd,
(ppm) 154.4, 56.9, 40.5, 22.1, 17.6. ESI-FTICR (þ) calcd for C7H13ON2
J¼6,13 Hz, 2H), 3.56 (d, J¼8 Hz, 2H), 4.50 (s, 2H), 7.32 (m, 5H), 13C
(MþH)þ: 141.1022, found: 141.1024.
NMR (CDCl3)
d (ppm) 138.0, 128.7, 128.1, 73.4, 65.8, 47.6, 19.3, 18.4.
IR: 2096. EI HRMS calcd for C13H17ON6 (MþH)þ: 273.1464, found:
4.2.2. (1R
thanol (12). This compound was obtained as described above as
a white solid in quantitative yield. 1H NMR (CD3OD)
ppm 1.24 (m,
*
,7S
*
,8S
*
,Z)-3,5-Diazabicyclo[5.1.0]oct-3-ene-4,8-diyldime-
273.1388.
d
4.1.7. ((1R,2S,3S)-3-(Benzyloxymethyl)cyclopropane-1,2-diyl)dime-
thanamine (8). Triphenylphosphine (0.74 g, 2.84 mmol) was added
to a solution of diazide 7 (0.19 g, 0.71 mmol) in dry CH2Cl2 (7.1 ml).
The mixture was stirred at room temperature under argon for
24 h. Then, water (0.38 ml) was added and the mixture refluxed
for 3 h. After the solvent was evaporated, the mixture was purified
by flash chromatography (silica, 1:1 CH2Cl2/MeOH then 1:1:0.01
CH2Cl2/MeOH/NH4OH). The desired cyclopropyl diamine was
obtained as a white solid in 95% yield (0.20 g). 1H NMR (CDCl3)
1H),1.65 (m, 2H), 3.52 (dd, J¼6, 12 Hz, 2H), 3.74 (d, J¼8 Hz, 2H), 3.86
(dd, J¼11, 15 Hz, 2H), 4.14 (s, 2H), 13C NMR (CD3OD)
d (ppm) 166.7,
59.0, 56.9, 40.1, 21.7, 17.5. ESI-FTICR (þ) calcd for C8H15O2N2
(MþH)þ: 171.1128, found: 171.1138.
4.3. Inhibition studies
Initial velocities for enzyme-catalyzed reactions were de-
termined by monitoring absorption at 400 nm due to the appear-
ance of nitrophenol produced during hydrolysis of the
corresponding nitrophenyl hexopyranoside. The kinetic assays
were performed at 37 ꢀC at pH 6.0 and pH 7.5 using 50 mM Na/
d
ppm 1.17 (m, 2H), 1.30 (m, 1H), 1.81 (br s, 4H), 2.76 (br d, J¼7 Hz,
4H), 3.56 (d, J¼8 Hz, 2H), 4.50 (s, 2H), 7.31 (m, 5H), 13C NMR
(CDCl3)
d (ppm) 138.6, 128.9, 128.2, 128.1, 73.5, 66.8, 37.9, 22.9,
18.5. EI HRMS calcd for C13H21ON2 (MþH)þ: 221.1654, found:
citrate/phosphate buffers with the exception of the
from Escherichia coli whose buffer contained 10 mM of MgCl2.
The glycosidases used were -galactosidase (green coffee beans),
-galactosidase (E. coli), -galactosidase (A. orizae), -glucosidase
(Saccharomyces cerevisiae), and -glucosidase (almonds). The
substrate used for each enzyme were the following: p-nitrophenyl-
-galactopyranoside ( -galactosidase), p-nitrophenyl- -glucopyr-
anoside ( -glucosidase), p-nitrophenyl- -glucopyranoside ( -glu-
cosidase), and o-nitrophenyl- -galactopyranoside (ONPG)
b-galactosidase
221.1654.
a
4.1.8. (1R
*
,7S
*
,8S ,Z)-8-(Benzyloxymethyl)-3,5-diazabicyclo[5.1.0]-
*
b
b
a
oct-3-ene (9). Diamine 8 (0.50 g, 2.3 mmol) was dissolved in 20 ml
of EtOH. Then, ethyl formimidate hydrochloride (0.25 g, 2.3 mmol)
was added and the reaction mixture was stirred under reflux for
10 h. After the solvent was evaporated under vacuum, the residue
was purified by flash chromatography (silica, CH2Cl2/MeOH 10:1) to
give the amidine as a white solid in 88% yield (0.50 g). 1H NMR
b
a
a
a
a
b
b
b
(b-
galactosidase). Reaction mixtures were delivered to 1.0 ml quartz
cuvettes and allowed to equilibrate to temperature for 3 min prior
to addition of enzyme. The reactions were initiated by the addition
(CDCl3)
3.70 (d, J¼8 Hz, 2H), 3.93 (dd, J¼11, 16 Hz, 2H), 4.48 (s, 2H), 7.27
(m, 5H), 7.49 (s, 1H), 13C NMR (CDCl3)
(ppm) 154.4, 138.5, 128.4,
d
ppm 1.39 (m,1H),1.73 (m, 2H), 3.47 (ddd, J¼2,5,14 Hz, 2H),
d
of the following units and volumes of enzymes: 0.2 U of
tosidase (A. orizae) in 5 l for both pHs; 0.2 U of -galactosidase
(green coffee beans) in 10 l at pH 7.5, and 0.02 U in 15 l at pH 6.0;
0.1 U of -glucosidase (S. cerevisiae) in 5 l at pH 7.5 and 0.2 U in
10 l at pH 6.0; 0.3 U of -galactosidase (E. coli) in 5 l at pH 7.5 and
0.5 U in 10 l at pH 6.0; 0.1 U of -glucosidase (almonds) in 5 l at
pH 7.5 and 0.01 U in 5 l at pH 6.0. For the inhibition screening
b-galac-
127.9, 127.8, 72.8, 65.4, 40.6, 19.7, 17.7. ESI-FTICR (þ) calcd for
m
a
C14H19ON2 (MþH)þ: 231.1489, found: 231.1489.
m
m
a
m
4.1.9. ((1R
*
,7S
*
,8S ,Z)-8-(Benzyloxymethyl)-3,5-diazabicyclo[5.1.0]-
*
m
b
m
oct-3-en-4-yl)methanol (11). Diamine 8 (0.5 g, 2.5 mmol) was dis-
solved in 13 ml of EtOH. Then, ethyl 2-hydroxyacetimidate hydro-
chloride (0.26 g, 2.5 mmol) was added and the reaction mixture
was stirred under reflux for 12 h. After the solvent was evaporated
under vacuum, the residue was purified by flash chromatography
(silica, CH2Cl2/MeOH 15:1) to give the hydroxymethyl amidine as
m
b
m
m
assays, the substrate concentration used was around Km and the
amount of enzyme was adjusted to obtain less than 10% hydrolysis
of the substrate over 5e10 min reaction time courses. The kinetic
parameters (Km, Vmax, Ki) were determined using 5e8 substrate
concentrations between 0.05 mM and 30 mM, and three different
inhibitors concentrations ranging from 0.3 mM to 4 mM. Initial
velocity data were fit with non-linear regression directly to the
equation for competitive inhibition (Eq. 2) to obtain inhibition
constants.
a white solid in a 70% yield (0.50 g). 1H NMR (CD3OD)
d ppm 1.42
(m, 1H), 1.76 (m, 2H), 3.57 (dd, J¼6, 14 Hz, 2H), 3.77 (d, J¼8 Hz, 2H),
3.95 (dd, J¼11, 15 Hz, 2H), 4.19 (s, 2H), 4.53 (s, 2H), 7.33 (m, 5H), 13C
NMR (CDCl3)
d (ppm) 166.9, 138.5, 128.5, 128.0, 127.9, 72.9, 65.4,
59.1, 40.3, 19.4, 17.6. ESI-FTICR (þ) calcd for C15H21O2N2 (MþH)þ:
261.1598, found: 261.1575.
v ¼ Vmax ꢂ ½Sꢅ=ðKm ꢂ ð1 þ ½Iꢅ=KiÞ þ ½SꢅÞ
(2)
4.2. General procedure for the deprotection of cyclic
amidines
Acknowledgements
The amidine (1 mmol) is dissolved in 5 ml of freshly distilled
MeOH. Ammonium formate (9 mmol) and 0.37 g of 10% Pd/C were
added to the solution of amidine. The mixture was refluxed until no
N.A.H. wishes to thank the National Science Foundation for
support of initial phases of this work under MCB0091881. F.M.L.

5572
F.M. Leonik et al. / Tetrahedron 66 (2010) 5566e5572
18. Horenstein, B. A. In ACS Symp. Ser. #721, Truhlar, D., Kniskern, D. P., Morokuma, K.,
Eds.; American Chemical Society: Washington, DC, 1999; pp 411e423.
19. Lillelund, V. H.; Jensen, H. H.; Liang, X.; Bols, M. Chem. Rev. 2002, 102, 515e553.
20. Asano, N.; Nash, R. J.; Molyneux, R. J.; Fleet, G. W. J. Tetrahedron: Asymmetry
2000, 11, 1645e1680.
21. Papandreou, G.; Tong, M. K.; Ganem, B. J. Am. Chem. Soc. 1993, 115, 11682e11690.
22. Sun, H. B.; Yang, J. S.; Amaral, K. E.; Horenstein, B. A. Tetrahedron Lett. 2001, 42,
2451e2453.
wishes to acknowledge support via the Reugamer Fellowship. We
thank Ms Mirna El Khatib for help recording optical rotations.
Supplementary data
Supplementary data associated with this article can be found in
the online version at doi:10.1016/j.tet.2010.05.064.
23. Young, J. E. P.; Horenstein, N. A. Tetrahedron Lett. 2004, 45, 9505e9507.
€
24. Drager, B. Nat. Prod. Rep. 2004, 21, 211e223.
25. Schmidt, B.; Pohler, M.; Costisella, B. Tetrahedron 2002, 58, 7951e7958.
26. Collado, I.; Pedregal, C.; Bueno, A. B.; Marcos, A.; Gonzalez, R.; Blanco-Urgoiti, J.;
Perez-Castells, J.; Schoepp, D. D.; Wright, R. A.; Johnson, B. G.; Kingston, A. E.;
Moher, E. D.; Hoard, D. W.; Griffey, K.; Tizzano, J. P. J. Med. Chem. 2004, 47,
456e466.
27. Doyle, M. P.; Austin, R. E.; Bailey, A. S.; Dwyer, M. P.; Dyatkin, A. B.; Kalinin, A.
V.; Kwan, M. M.; Liras, S.; Oalmann, C. J.; Pieters, R. J.; Protopopova, M. N.; Raab,
C. E.; Roos, G. H. P.; Zhou, Q. L.; Martin, S. F. J. Am. Chem. Soc. 1995, 117,
5763e5775.
References and notes
1. Schramm, V. L. Arch. Biochem. Biophys. 2005, 433, 13e26.
2. Benkovic, S. J.; Hammes, G. G.; Hammes-Schiffer, S. Biochemistry 2008, 47,
3317e3321.
3. Pauling, L. Chem. Eng. News 1946, 24, 1375e1377.
4. Wolfenden, R. Nature 1969, 223, 704e705.
5. Schramm, V. L. J. Biol. Chem. 2007, 282, 28297e28300.
6. Hiratake, J. Chem. Rec. 2005, 5, 209e228.
7. Asano, N. Glycobiology 2003, 13, 93Re104R.
28. Evans, D. A.; Woerpel, K. A.; Hinman, M. M.; Faul, M. M. J. Am. Chem. Soc. 1991,
113, 726e728.
29. Schill, H.; de Meijere, A.; Yufit, D. S. Org. Lett. 2007, 9, 2617e2620.
30. Faust, J. A.; Mori, A.; Sahyun, M. J. Am. Chem. Soc. 1959, 81, 2214e2219.
31. Marcelo, F.; He, Y.; Yuzwa, S.; Nieto, L.; Jimenez-Barbero, J.; Sollogoub, M.;
Vocadlo, D.; Davies, G.; Bleriot, Y. J. Am. Chem. Soc. 2009, 131, 5390e5392.
32. Dale, M. P.; Ensley, H. E.; Kern, K.; Sastry, K. A. R.; Byers, L. D. Biochemistry 1985,
24, 3530e3539.
33. Gloster, T. M.; Meloncelli, P.; Stick, R. V.; Zechel, D.; Vasella, A.; Davies, G. J.
J. Am. Chem. Soc. 2007, 129, 2345e2354.
34. Gloster, T. M.; Davies, G. J. Org. Biomol. Chem. 2010, 8, 305e320.
35. Ebinger, A.; Heinz, T.; Umbricht, G.; Pfaltz, A. Tetrahedron 1998, 54,
10469e10480.
8. Guthrie, R. D.; Jencks, W. P. Acc. Chem. Res. 1989, 22, 343e349.
9. Demarco, M. L.; Woods, R. J. Glycobiology 2008, 18, 426e440.
10. Varki, A. Glycobiology 2003, 3, 97e130.
11. (a) de Melo, E. B.; Gomes, A. D.; Carvalho, I. Tetrahedron 2006, 62, 10277e10302;
(b) Asano, N. Cell. Mol. Life Sci. 2009, 66, 1479e1492.
12. Vasella, A.; Davies, G. J.; Bohm, M. Curr. Opin. Chem. Biol. 2002, 6, 619e629.
13. Rye, C. S.; Withers, S. G. Curr. Opin. Chem. Biol. 2000, 4, 573e580.
14. Yip, V. L. Y.; Withers, S. G. Biocat. Biotrans. 2006, 24, 167e176.
15. Burgi, H. B.; Dunitz, J. D.; Lehn, J. M.; Wipff, G. Tetrahedron 1974, 30, 1563e1572.
16. Burgi, H. B.; Dunitz, J. D. Acc. Chem. Res. 1983, 16, 153e161.
17. Barnes, J. A.; Williams, I. H. Chem. Commun. 1996, 193e194.