Alejandro Puet1, Gema Domínguez1, F Javier Cañada2,3, Javier Pérez-Castells1. 1. Department of Chemistry and Biochemistry, Facultad de Farmacia, Universidad San Pablo-CEU, CEU Universities, Urbanización Montepríncipe, 28668 Boadilla del Monte, Madrid, Spain. 2. Dep. Biología FisicoQuímica, CIB Margarita Salas, CSIC, C/Ramiro de Maetzu 9, 28040 Madrid, Spain. 3. CIBER de Enfermedades Respiratorias (CIBERES), Avda, Monforte de Lemos 3-5, 28029 Madrid, Spain.
Abstract
Synthesis of four iminosugars fused to a cyclopropane ring is described using l-serine as the chiral pool. The key steps are large-scale preparation of an α,β-unsaturated piperidinone followed by completely stereoselective sulfur ylide cyclopropanation. Stereochemistry of compounds has been studied by nuclear Overhauser effect spectroscopy (NOESY) experiments and 1H homonuclear decoupling to measure constant couplings. The activity of these compounds against different glycosidases has been evaluated. Although inhibition activity was low (compound 8a presents a (K i) of 1.18 mM against β-galactosidase from Escherichia coli), interestingly, we found that compounds 8a and 8b increase the activity of neuraminidase from Vibrio cholerae up to 100%.
Synthesis of four iminosugars fused to a cyclopropane ring is described using l-serine as the chiral pool. The key steps are large-scale preparation of an α,β-unsaturated piperidinone followed by completely stereoselective sulfur ylide cyclopropanation. Stereochemistry of compounds has been studied by nuclear Overhauser effect spectroscopy (NOESY) experiments and 1H homonuclear decoupling to measure constant couplings. The activity of these compounds against different glycosidases has been evaluated. Although inhibition activity was low (compound 8a presents a (K i) of 1.18 mM against β-galactosidase from Escherichia coli), interestingly, we found that compounds 8a and 8b increase the activity of neuraminidase from Vibrio cholerae up to 100%.
Iminosugars are azaheterocycles
with promising biological activities
such as glycosidase and glycosyltransferase inhibition and modulation.[1] Many iminosugars are natural or synthetic polyhydroxylated
piperidines, which can act as biomimetics of their corresponding pyranose
analogues. Some of the most important natural piperidineiminosugars
are nojirimycin (Figure , I) and its epimers, which, together with their deoxy
analogues have turned out to be the lead molecules for drug design.
Thus, stereochemical changes and functional group variation have led
to iminosugars that can modulate glycosidase enzymes, exhibiting immunosuppressive,
antiviral, or anti-inflammatory activities.[2] Bicyclic iminosugars, such as swainsonine (II), lentiginosine
(III), castanospermine (IV), and their derivatives
exhibit antitumor and immunosuppressive activities.[3]
Figure 1
Structures of nojirimycin (I), swainsonine (II), lentiginosine (III), castanospermine (IV), miglitol (V), migalastat (VI), and miglustat (VII).
Structures of nojirimycin (I), swainsonine (II), lentiginosine (III), castanospermine (IV), miglitol (V), migalastat (VI), and miglustat (VII).Several iminosugars, miglitol (Glyset, V),[4] migalastat (Galafold, VI),[5] and miglustat (Zavesca, VII)[6] are commercially available for the treatment
of type II diabetes and Fabry disease, and as the first oral treatment
for Gaucher disease, respectively. Several other competitive inhibitors
of glycosidases are being developed as new drugs and are in different
phases of clinical trials.The mechanism associated with glycosidase
activity modulation is
generally attributed to structural similarity to the oxacarbenium
ion-like transition-state, formed during the hydrolysis of carbohydrates.[7] These transition states present diverse conformational
pathways for different glycosidases,[8] making
selective inhibition possible. In this context, designing conformationally
restricted inhibitors seems to be an interesting approach. In addition,
adequate metabolic stability is needed, which may be achieved with
more rigid compounds. Recently, a study on α-mannanases showed
how the enzyme surface restricts the conformational landscape of the
substrate, rendering the B2,5 conformation the most stable
on-enzyme (Figure a).[9] In another study, a cyclopropane
containing a cyclophellitol analogue, was designed as a specific β-glucosidase
inhibitor for enzymes reacting through the 4H3 transition-state conformation (Figure b).[10]
Figure 2
Transition-state
conformations on-enzyme: (a) mannose B2,5 conformation
(VIII, reprinted with permission from
ref (9)) and (b) cyclophellitol
analogue and its 4H3 conformation (IX, ref (10), copyright
2017 American Chemical Society).
Transition-state
conformations on-enzyme: (a) mannose B2,5 conformation
(VIII, reprinted with permission from
ref (9)) and (b) cyclophellitol
analogue and its 4H3 conformation (IX, ref (10), copyright
2017 American Chemical Society).With these precedents, we expected that the introduction of a three-membered
ring annulated to a piperidine ring would render novel iminosugars
with a locked conformation that may be the starting point for finding
the therapeutic compounds (Figure ). The substituted cyclopropane moiety renders a fixed
conformation and allows many different configurations that could increase
selectivity to specific glycosidases.
Figure 3
Structure of the target compound.
Structure of the target compound.The development of efficient routes for the preparation
of iminosugars
has received much attention from the synthetic community.[11] Most of the methods use carbohydrates as the
chiral pool, which are transformed using reductive aminations[12] or other transformation strategies.[13] Alternatively, some asymmetric or biocatalysed
approaches have been used.[14] But, there
are fewer reports on approaches where amino acids are used as the
chiral pool for the synthesis of iminosugar derivatives.[15]Herein, we envisioned the preparation
of novel bicyclic iminosugars
that include the cyclopropane motif fused with piperidine starting
from the natural amino acid l-serine as the chiral pool.
The final compounds present five stereogenic centers, and the synthesis
involves the inversion of the configuration of the starting l-serine (S-configuration) into the C5 configuration
of d-carbohydrates (R).[16] Preliminary glycosidase inhibition evaluation is shown.
Results
and Discussion
Our first goal was the synthesis of α,β-unsaturated
ketone 5 in which the chiral center has R configuration. This configuration was selected as it corresponds
to C5 in natural sugars and iminosugars, which share the R configuration in that position. This compound has already been prepared
from d-serine and described.[17] In our case, we developed a synthesis approach using cheaper and
natural l-serine, as depicted in Scheme . From this intermediate, a cyclopropanation
reaction and further transformations resulted in a new family of piperidines
fused to cyclopropanes. l-Serine was esterified and protected
with Boc2O, and the resulting intermediate was further
protected and reduced to give desymmetrized alcohol 1 in which the configuration has changed from S to R in a few steps. This compound 1 was transformed
into oxazolidinone by reaction with a base followed by allylation
to give compound 2. Following the previously reported
methodology,[18]2 was de-protected
and oxidized into carboxylic acid 3. This was converted
into Weinreb amide 4, which was treated with vinylmagnesium
bromide and subjected to a ring-closing metathesis, RCM, using second-generation
Grubbs’ catalyst (Grubbs Catalyst M204), giving the starting
material 5.[19] This precursor
containing the piperidine core was obtained in 13% global yield after
11 steps. No racemization was observed during the synthesis.
Scheme 1
Synthesis
of Compound 5
The cyclopropanation reaction of 5 was performed using
sulfur ylide. Interestingly, only one reaction product was observed
and isolated in 70% yield. This product was designated as structure 6, as a result of NMR analysis (nuclear Overhauser effect
(NOE) and coupling constants) of compounds 7a–b (vide infra). Cyclopropanation occurred on the
same side of oxazolidinone (endo attack) and the subsequent ring-closing
step exclusively gave exo-cyclopropane. The ylide
mediated cyclopropanation is a stepwise reaction in which the formation
of the first C–C bond is the rate-determining step.[20] The attack of the ylide on 5 is
more favored from the opposite face to the nitrogen lone electron
pair as depicted in Scheme; therefore, it occurs through the same face of the oxazolidinone
ring (endo). This selectivity was observed previously in one unsaturated
γ-lactam cyclopropanation,[21] although
other precedents have described mixtures of endo and exo attacks.[22] Then, the stereoselectivity of cyclopropane
is determined in the second step. Studies reported by Aggarwal’s
group,[23] showed that two intermediate betaines A and B are formed in a 1:1 ratio after the nucleophilic
addition. The cyclization of betaine A is faster than
that of B. Moreover, B can epimerize to
give A before it closes the three-membered ring generally
leading to high diastereoselectivity of exo-cyclopropane.
Scheme 2
Stereoselectivity
of the Cyclopropanation Reaction of 5(24)
With the cyclopropane containing
compound 6 in hand,
the reduction of the ketone afforded a (3:2) mixture of the two diastereomers 7a–b (Scheme ). These alcohols were separated, characterized, and separately
transformed into the final products. The stereochemistry of 7a and 7b was determined using NOE experiments
and coupling constant values. Figure shows the main correlations observed for 7a and 7b that allowed assigning the relative configuration
of H6, H7, and H7a. The coupling constants between H7a, H7, and H6a
were determined using homonuclear decoupling experiments. Values are
shown in Figure ,
and the model agrees with the calculated angles for these couplings.
Scheme 3
Synthesis of Final Compounds from 5
Figure 4
NOE signals
and constant couplings in products 7a and 7b, respectively.[24]
NOE signals
and constant couplings in products 7a and 7b, respectively.[24]In continuation of the synthesis, treatment with trifluoroacetyl
(TFA) and further reaction with ethylendiamine gave products 8a and 8b, respectively, in excellent yields.
On the other hand, compounds 7a and 7b gave
different products on reacting with DIBAL-H, whereas 7b gave the expected alcohol 10 (62%), additionally, the
reaction of 7a caused the cleavage of the oxazolidinone
ring giving 9 in 58% yield. This behavior has been described
previously (Scheme ).[25]Final compounds were screened
for glycosidase inhibition activities
(α-glucosidase from Bacillus stearothermophilus, β-glucosidase from almonds, α-galactosidase from green
coffee beans, β-galactosidase from Escherichia
coli, α-mannosidase from Jack
beans, β-mannosidase from Helix
pomatia, and α-l-fucosidase from Homo sapiens) using p-nitrophenyl
monosaccharides as substrates.Carboxylic containing compounds 8a and 8b, could resemble the zwitterionic form
of oseltamivir and zanamivir,
well-known inhibitors of neuraminidase from Influenza virus[26] with carboxyl-amino and carboxyl-guanidine moieties,
respectively. Preliminary docking calculations using AutoDock[27] showed that 8a and 8b could fit in the binding site of neuraminidase. Thus they were also
evaluated as possible inhibitors of neuraminidase from Vibrio cholerae.The enzymatic activities were
calculated by measuring the absorbance
of the phenoxide released in the enzymatic reaction at 405 nm. The
compounds were initially screened at 1, 5, and 25 mM concentrations.
With compounds 8a and 9, inhibition over
50% was observed with selected enzymes at 5 mM; 8b and 10 did not show any significant inhibition of any glycosidase
(Chart ). Inhibition
constants (Ki) were estimated assuming
a competitive type inhibition in the cases of compound 8a against β-galactosidase (Ki =
1.18 mM) and 9 against β-glucosidase (Ki = 4.43 mM). These two compounds exhibit some selectivity
such that even at 25 mM no significant inhibition was observed against
other glycosidases.
Chart 1
Residual Glycosidase Activities in the Presence of
5 mM Synthesized
Compounds
However, the observed inhibition
was very weak compared to other
iminosugar-based glycosidase inhibition, for example the measured Ki for deoxinojirimycin (DNJ) is 0.44 μM
for α-glucosidase from B. stearothermophilus. The inhibition constant changes depending on the species that is
studied, even for the same glycosidase of other species.[28] Other iminosugars present great activity against
mannosidases.[29]On the other hand,
we found that compounds 8a and 8b, bearing
a carboxylate group, also did not show any inhibition
against neuraminidase but unexpectedly produced activation of the
enzyme; these two compounds increased neuraminidase activity up to
100%. The possibility that the compounds act as favorable transglycosylation
acceptors causing an increase of nitrophenol release was considered.
NMR experiments were performed continuously following the reaction,
but potential transient transglycosylation products could not be observed.
Further research to explain this behavior is needed. Interestingly
there are not many precedents on glycosidase activation by iminosugars.
Two reports have accounted for this activation behavior. Thus, up
to 70-fold activation of some of glycosidases was detected with multivalent
iminosugars.[30] In another study, thienopyrimidines
were found to activate certain glycosidases.[31] The activation mechanism could be explained by the stabilization
of the active structure of the enzyme by the introduction of a small
molecule adjacent/close to the substrate-binding site, locking the
reactive form. Alternatively, if the activation is of the allosteric
type occurring in a site different from the active site, it could
be interesting to check if the activators have any pharmacological
chaperone activity but avoid the temporal inhibition of the enzymatic
activity, unlike the aforementioned migalastat and other proposed
pharmacological chaperones that help maintaining the correct fold
of the protein although temporally blocking the active site of the
enzyme.
Conclusions
We described a multigram synthesis of an
α,β-unsaturated
ketone, which upon a stereoselective cyclopropanation reaction and
further transformations gave a novel series of bicyclic piperidine-based
iminosugars. The final products were studied against different glycosidases.
Inhibition in most cases was low, but interestingly, the activation
of neuraminidase was observed with products 8. Possible
explanations of this behavior, for example, allosteric activation,
enzyme stabilization, or transglycosylation acceptor activity can
be proposed. Current studies in our lab will provide insight into
these possible mechanisms, and their potential applications will be
explored/pursued.
Experimental Section
General Information
All chemicals were obtained from
Aldrich/Merck, VWR, Fluorochem, and ABCR. Thin-layer chromatography
(TLC) analyses were performed on Merck silica gel 60 F254 plates using
phosphomolybdic acid or anisaldehyde and heat for detection. Silica
gel NORMASIL 60 40–63 μm was used for flash chromatography.
NMR spectra were recorded on a Bruker spectrometer (400 MHz for 1H and 100 MHz for 13C). Chemical shifts are reported
in δ ppm referenced to CDCl3 (δ = 7.26 for 1H and 77.00 for 13C), CD3OD (δ
= 3.31 for 1H and 49.00 for 13C), or D2O (δ = 4.79 for 1H). Bidimensional spectra (heteronuclear
multiple quantum coherence (HMQC), heteronuclear multiple bond coherence
(HMBC), correlated spectroscopy (COSY), and nuclear Overhauser effect
spectroscopy (NOESY)) were recorded in order to carry out the assignment.
IR spectra were recorded on a Perkin-Elmer Spectrum 100. Specific
optical rotation was measured using a polarimeter Anton Parr MCP 100.
Melting points of solid compounds were determined using a Stuart Scientific
Melting Point Apparatus SMP3. The absorbance of p-nitrophenoxide released in the enzymatic reactions was measured
at 405 nm in a Perkin-Elmer Lamba 25.
Methyl (tert-Butoxycarbonyl)-l-serinate
Thionyl chloride (83
mL, 1.1 mol) was added to methanol (280 mL)
at 0 °C, then l-serine (60.00 g, 571 mmol) is added.
After 10 min, at 0 °C, the solution is heated at 65 °C for
2 h. The solvent is evaporated in vacuo, and 600
mL of AcOEt and a saturated solution of NaHCO3 (until basic
pH) are added. Di-tert-butyl dicarbonate (124.62
g, 0.571 mmol) in 265 mL of AcOEt is added. The reaction is stirred
overnight at room temperature. The aqueous layer is extracted with
AcOEt (2 × 300 mL). The combined organic layers are washed with
brine (200 mL), dried over MgSO4, and evaporated in vacuo. The crude product is filtered through a pad of
silica gel using Hex/AcOEt (9:1) to Hex/AcOEt (3:1) as eluents. A
colorless oil is obtained (101.6 g, 81% after two steps). 1H NMR (400 MHz, CDCl3) δ 5.46 (brs, 1H, NH), 4.39
(brs, 1H, CH), 3.99–3.89 (m, 2H, CH2O), 3.78 (s,
3H, OMe), 2.47–2.32 (m, 1H, OH), 1.45 (s, 9H, 3 × CH3). 13C NMR (100 MHz, CDCl3) δ
171.6, 155.9, 80.3, 63.3, 55.8, 52.8, 28.3 (3C). IR (NaCl): 3378,
2984, 2868, 1740, 1708 cm–1. [α]D25 (c 0.13 in dichloromethane (DCM)):
+4.14. Found: C, 49.1; H, 7.9%. Calc. for C9H17NO5: C, 49.3; H, 7.8%.
To a solution of methyl (tert-butoxycarbonyl)-l-serinate (101.4 g, 463 mmol) in 400 mL
of dimethylformamide (DMF) cooled to 0 °C is added imidazole
(37.8 g, 555 mmol) and 4-dimethylaminopyridine (DMAP; 5.6 g, 46 mmol).
After 10 min, tert-butyldimethylsilyl chloride (73.2
g, 486 mmol) is added. The reaction is stirred for 30 min at room
temperature. AcOEt (400 mL) is added, and the organic layer is washed
with water (3 × 1 L) and brine (400 mL), dried over MgSO4, and evaporated in vacuo. To a suspension
of NaBH4 (35.0 g, 926 mmol) and LiCl (39.3 g, 926 mmol)
in 800 mL of ethanol cooled to 0 °C, a solution of the crude
in 190 mL of ethanol is added slowly. The reaction is stirred at 0
°C for 10 min, at room temperature for 30 min, and at 50 °C
for 2.5 h. The reaction is cooled to 0 °C, and a saturated solution
of NH4Cl is added (until salts are dissolved, 450 mL).
The aqueous layer is extracted with AcOEt (3 × 350 mL). The combined
organic layers are washed with brine (200 mL), dried over MgSO4, and evaporated in vacuo. The reaction crude
is filtered through a pad of silica gel using Hex/AcOEt (19:1) to
Hex/AcOEt (1:1) as eluents. A colorless oil is obtained (123.2 g,
85% after two steps). 1H NMR (400 MHz, CDCl3) δ 5.14 (brs, 1H, NH), 3.86–3.66 (m, 5H, 2 × CH2O + CH), 2.70 (brs, 1H, OH), 1.45 (s, 9H, 3 × CH3), 0.90 (s, 9H, 3 × CH3), 0.08 (s, 6H, 2 ×
CH3Si). 13C NMR (100 MHz, CDCl3)
δ 156.1, 79.6, 64.0 (2C), 52.7, 28.5 (3C), 25.9 (3C), 18.3,
−5.5 (2C). IR (NaCl): 3371, 2952, 2861, 1707 cm–1. [α]D25 (c 0.28 in
DCM): +9.30. Found: C, 55.1, H, 10.4%. Calc. for C14H31NO4Si: C, 55.0; H, 10.2%.
To a suspension of NaH 60% w/w (18.9 g, 472 mmol) in 300 mL of
THF cooled to 0 °C is added a solution of 1 (123.2
g, 403 mmol) in 550 mL of THF. The reaction is stirred at 0 °C
for 15 min, at room temperature for 25 min, and at 40 °C for
2.5 h. The reaction is cooled down to 0 °C, and a saturated solution
of NH4Cl is added until all salts are dissolved (250 mL).
The aqueous layer is extracted with AcOEt (2 × 500 mL). The combined
organic layers are washed with brine (300 mL), dried over MgSO4, and evaporated in vacuo. A colorless wax
is obtained (78.9 g, 72%). 1H NMR (400 MHz, CDCl3) δ 6.27 (brs, 1H, NH), 4.42 (t, J = 8.6 Hz,
1H, CH2O), 4.18 (dd, J = 8.8, 4.8 Hz,
1H, CH2O), 3.94–3.88 (m, 1H, CH), 3.60 (d, J = 5.4 Hz, 2H, CH2OSi), 0.87 (s, 9H, 3 ×
CH3), 0.05 (s, 6H, 2 × CH3Si). 13C NMR (100 MHz, CDCl3) δ 160.2, 67.3, 64.8, 53.8,
25.9 (3C), 18.3, −5.4 (2C). IR (NaCl): 3315, 2959, 2848, 1745
cm–1. [α]D25 (c 0.32 in DCM): −15.94. Found: C, 52.2, H, 9.0%.
Calc. for C10H21NO3Si: C, 51.9; H,
9.2%.
To a suspension of NaH 60% w/w (16.4 g,
409.1 mmol) in 500 mL of THF at 0 °C, a solution of (R)-4-(((tert-butyldimethylsilyl)oxy)methyl)oxazolidin-2-one
(78.9 g, 340.9 mmol) in 500 mL of THF is added slowly. Allyl bromide
(29.5 mL, 340.9 mmol) is added and stirred for 15 min at 0 °C,
30 min at room temperature, and 2 h at 50 °C. A saturated solution
of NH4Cl is added until the salts are dissolved. The aqueous
layer is extracted with AcOEt (3 × 200 mL). The combined organic
layers are washed with brine (150 mL) and dried over MgSO4. The solvent is evaporated in vacuo, and the residue
is filtered through a pad of silica gel using Hex/AcOEt (19:1) to
Hex/AcOEt (1:1) as eluents. A yellow oil is obtained (64.8 g, 70%). 1H NMR (400 MHz, CDCl3) δ 5.84–5.74
(m, 1H, HC=), 5.27–5.21 (m, 2H, H2C=),
4.33 (t, J = 8.7 Hz, 1H, CH2O), 4.18–4.13
(m, 2H, CH2O + CH2N), 3.86–3.80 (m, 1H,
CH), 3.69–3.62 (m, 3H, CH2OSi + CH2N),
0.89 (s, 9H, 3 × CH3), 0.06 (s, 6H, 2 × CH3Si). 13C NMR (100 MHz, CDCl3) δ
158.4, 132.7, 118.5, 65.0, 62.2, 56.0, 45.3, 25.9 (3C), 18.3, −5.4
(2C). IR (NaCl): 3084, 2948, 2866, 1744 cm–1. [α]D25 (c 0.21 in CHCl3): −11.84. Found: C, 57.1, H, 9.1%. Calc. for C13H25NO3Si: C, 57.5; H,9.3%.
(S)-3-Allyl-4-(hydroxymethyl)oxazolidin-2-one
To
a solution of 2 (55.5 g, 204.6 mmol) in 220 mL
of THF is added TBAF·3H2O (58.8 g, 225.0 mmol). The
mixture is stirred for 30 min at room temperature. The solvent is
evaporated in vacuo and filtered through a pad of
silica gel using Hex/AcOEt (2:1) to Hex/AcOEt (1:2) as eluents. A
colorless oil is obtained (28.9 g, 90%). 1H NMR (400 MHz,
CDCl3) δ 5.86–5.76 (m, 1H, HC=), 5.30–5.24
(m, 2H, H2C=), 4.36 (t, J = 8.8
Hz, 1H, CH2O), 4.25 (dd, J = 8.7, 6.0
Hz, 1H, CH2O), 4.09 (ddt, J = 15.7, 5.3,
1.6 Hz, 1H, CH2N), 3.90–3.84 (m, 1H, CHN), 3.80–3.73
(m, 2H, CH2N + CH2OH), 3.65 (dd, J = 11.9, 3.3 Hz, 1H, CH2OH). 13C NMR (100 MHz,
CDCl3) δ 158.7, 132.5, 118.9, 64.7, 60.9, 56.3, 45.4.
IR (NaCl): 3427, 3048, 2975, 2851, 1753 cm–1. [α]D25 (0.25 in CHCl3): −44.28. Found:
C, 53.8; H, 7.3%. Calc. for C7H11NO3: C, 53.5; H, 7.1%.
A solution of (S)-3-allyl-4-(hydroxymethyl)oxazolidin-2-one
(28.9 g, 184.1 mmol) in 1 L of acetone is cooled to 0 °C, and
92 mL of Jones’ reagent is added slowly. The reaction is stirred
for 1.5 h at 0 °C. Isopropanol is added until the solution turns
blue. The mixture is filtered through a pad of celite. The solvent
is evaporated in vacuo and the residue is filtered
through a pad of silica gel using Hex/AcOEt (1:1) to AcOEt 100% as
eluents. A pale yellow oil is obtained (23.0 g, 73%). 1H NMR (400 MHz, CDCl3) δ 9.83 (brs, 1H, OH), 5.80–5.70
(m, 1H, HC=), 5.27–5.24 (m, 2H, H2C=),
4.52 (t, J = 9.4 Hz, 1H, CH2O), 4.42 (dd, J = 9.0, 4.5 Hz, 1H, CH2O), 4.36 (dd, J = 9.7, 4.6 Hz, 1H, CHN), 4.27 (dd, J =
15.4, 4.8 Hz, 1H, CH2N), 3.75 (dd, J =
15.4, 8.0 Hz, 1H, CH2N). 13C NMR (100 MHz, CDCl3) δ 172.3, 158.5, 131.0, 120.1, 65.1, 56.0, 46.0. IR
(NaCl): 3454, 2933, 2839, 1731 cm–1. [α]D25 (0.30 in CHCl3): +8.98. Found: C,
48.8; H, 5.4%. Calc. for C7H9NO4:
C, 49.1; H, 5.3%.
A solution of 3 (23.0 g, 134.4
mmol) in 500
mL of DCM is cooled to 0 °C. Diisopropylethylamine (DIPEA; 23.5
mL, 134.4 mmol), EDCI (25.8 g, 134.4 mmol), and N,O-dimethylhydroxylamine hydrochloride (13.1 g,
134.4 mmol) are added. The reaction is stirred for 2 h at 0 °C.
The solvent is evaporated in vacuo and filtered through
a pad of silica gel using Hex/AcOEt (1:4) as the eluent. A yellow
oil is obtained (25.3 g, 88%). 1H NMR (400 MHz, CDCl3) δ 5.83–5.73 (m, 1H, HC=), 5.25–5.21
(m, 2H, H2C=), 4.66 (dd, J = 9.8,
5.6 Hz, 1H, CHN), 4.50 (t, J = 9.3 Hz, 1H, CH2O), 4.32 (ddt, J = 15.4, 4.7, 1.7 Hz, 1H,
CH2N), 4.18 (dd, J = 8.8, 5.6 Hz, 1H,
CH2O), 3.69 (s, 3H, OCH3), 3.68 (dd, J = 15.4, 8.3 Hz, 1H, CH2N), 3.22 (s, 3H, NCH3). 13C NMR (100 MHz, CDCl3) δ
169.6, 158.0, 132.2, 119.3, 64.5, 61.7, 54.8, 45.9, 32.7. IR (NaCl):
3088, 2979, 2928, 1754, 1672 cm–1. [α]D25 (c 0.37 in DCM): +31.11. Found:
C, 50.4; H, 6.9%. Calc. for C9H14N2O4: C, 50.5; H, 6.6%.
(R)-4-Acryloyl-3-allyloxazolidin-2-one
To a solution of 4 (12.0 g, 56.0 mmol) in 270 mL of
THF cooled to −30 °C, 0.7 M vinylmagnesium bromide (200
mL) is added slowly, keeping the temperature below −25 °C.
When the addition is finished, the reaction is stirred for another
30 min at −30 °C. The reaction mixture is poured into
a mixture of 200 mL of HCl 10% and 100 mL of MeOH cooled in a bath
at −15 °C. This mixture is stirred for another 15 min.
The aqueous layer is extracted with AcOEt (3 × 150 mL). The combined
organic layers are washed with a solution of 1 M HCl (200 mL), with
a saturated solution of NaHCO3 (150 mL) and brine (150
mL), dried over MgSO4, and evaporated in vacuo. The crude is used without further purification. 1H NMR
(400 MHz, CDCl3) δ 6.50 (dd, J =
17.5, 10.4 Hz, 1H, =CHCO), 6.39 (d, J = 17.4
Hz, 1H, H2C = CHCO, trans), 6.01 (d, J = 10.4 Hz, 1H, H2C = CHCO, cis), 5.78–5.68
(m, 1H, HC=), 5.24–5.16 (m, 2H, H2C=),
4.59 (dd, J = 10.0, 5.2 Hz, 1H, CHN), 4.52 (t, J = 9.3 Hz, 1H, CH2O), 4.26 (dd, J = 15.3, 4.6 Hz, 1H, CH2N), 4.15 (dd, J = 8.6, 5.2 Hz, 1H, CH2O), 3.59 (dd, J = 15.3, 8.0 Hz, 1H, CH2N). 13C NMR (100 MHz,
CDCl3) δ 195.1, 157.7, 132.3, 131.8, 131.4, 120.1,
63.8, 60.5, 46.2.
To a
solution of 5 (2.4 g, 15.9 mmol) in 13 mL of DCM at 0
°C is added a solution of tert-butyl (tetrahydrothiophenylidene)acetate
(9.6 g, 47.6 mmol) in 207 mL of DCM slowly. The reaction is stirred
at room temperature for 30 min. Deionized water (20 mL) is added.
The aqueous layer is extracted with DCM (2 × 30 mL). The combined
organic phases are washed with brine (20 mL), dried over MgSO4, and evaporated in vacuo. The crude is purified
in silica gel using Hex/AcOEt (2:1) as the eluent. A yellow wax is
obtained (3.0 g, 70%). 1H NMR (400 MHz, CDCl3) δ 4.53 (t, J = 9.7 Hz, 1H, CH2O), 4.32 (dd, J = 9.3, 5.7 Hz, 1H, CH2O), 4.23 (d, J = 14.1 Hz, 1H, CH2N),
4.04 (dd, J = 10.3, 5.7 Hz, 1H, CHN), 3.53 (d, J = 14.0 Hz, 1H, CH2N), 2.39 (dd, J = 7.9, 4.2 Hz, 1H, CHCO), 2.21–2.18 (m, 1H, CHCH2N) 2.10 (t, J = 4.5 Hz, 1H, CHCO2), 1.44 (s, 9H, 3 × CH3). 13C NMR
(100 MHz, CDCl3) δ 200.0, 168.9, 156.9, 82.8, 64.2,
58.8, 37.2, 31.6, 28.1, 24.5, 22.7. IR (NaCl): 2975, 2863, 1748, 1736,
1719 cm–1. [α]D25 (c 0.11 in DCM): +38.02. Found: C, 58.0; H, 4.9%. Calc. for
C13H17NO5: C, 58.4; H, 4.6%.
tert-Butyl (5aR,6S,6aS,7S,7aR) and tert-butyl (5aR,6S,6aS,7R,7aR)-7-Hydroxy-3-oxohexahydro-1H,3H-cyclopropa[d]oxazolo[3,4-a]pyridine-6-carboxylate (7a and 7b)
To a solution of 6 (1.1 g, 4.1 mmol) in 35
mL of absolute ethanol at 0 °C is added NaBH4 (312
mg, 8.2 mmol). The reaction is stirred for 1 h at room temperature.
A solution of saturated NH4Cl (20 mL) and water (until
salts dissolve) is added. The aqueous phase is extracted with AcOEt
(3 × 60 mL). The combined organic layers are washed with brine
(50 mL), dried over MgSO4, and evaporated in vacuo. The crude contained a 3:2 mixture of isomers 7a/7b as determined by the integration of signals in the 1H NMR spectrum of the reaction crude. This mixture was separated
by silica gel chromatography using Hex/AcOEt (1:1) to Hex/AcOEt (1:2)
as eluents. A yellow wax is obtained for isomer 7a (463
mg, 42%). A yellow solid is obtained for isomer 7b (330
mg, 30%).Spectroscopic data for tert-butyl
(5aR,6S,6aS,7S,7aR)-7-hydroxy-3-oxohexahydro-1H,3H-cyclopropa[d]oxazolo[3,4-a]pyridine-6-carboxylate 7a: 1H
NMR (400 MHz, CDCl3) δ 4.52 (t, J = 8.6 Hz, 1H, CH2O), 4.12 (dd, J = 9.1,
4.8 Hz, 1H, CH2O), 4.00 (d, J = 13.6 Hz,
1H, CH2N), 3.80 (dd, J = 8.5, 4.5 Hz,
1H, CHOH), 3.40 (dd, J = 13.6, 4.1 Hz,
1H, CH2N), 3.33 (td, J = 8.3, 4.9 Hz,
1H, CHN), 2.41 (d, J = 5.0 Hz, 1H, OH), 1.76–1.70
(m, 1H, CHCH2N), 1.67 (dd, J = 9.1, 5.0 Hz, 1H, CHCHOH), 1.44 (s, 9H, 3 × CH3), 1.40 (t, J = 4.9 Hz, 1H, CHCO2). 13C NMR (100 MHz, CDCl3) δ 172.2,
157.5, 81.4, 69.0, 68.3, 56.8, 38.5, 28.2 (3C), 27.3, 24.5, 20.3.
IR (NaCl): 3361, 2975, 2863, 1748, 1736 cm–1. [α]D25 (c 0.14 in DCM): −2.20.
Found: C, 58.2; H, 6.8%. Calc. for C13H19NO5: C, 58.0; H, 7.1%.Spectroscopic data for tert-butyl (5aR,6S,6aS,7R,7aR)-7-hydroxy-3-oxohexahydro-1H,3H-cyclopropa[d]oxazolo[3,4-a]pyridine-6-carboxylate 7b: 1H
NMR (400 MHz,
CDCl3) δ 4.54 (dd, J 0 8.5, 6.4
Hz, 1H, CH2O), 4.29 (t, J = 8.8 Hz, 1H,
CH2O), 4.24–4.20 (m, 1H, CHOH), 4.00
(d, J = 13.4 Hz, 1H, CH2N), 3.61–3.55
(m, 1H, CHN), 3.29 (dd, J = 13.4, 4.0 Hz, 1H, CH2N), 2.11 (td, J = 8.4, 5.0 Hz, 1H, CHCHOH), 1.97 (d, J = 4.1 Hz, 1H, OH), 1.84–1.74
(m, 2H, 2 × CH cyclopropane), 1.44 (s, 9H, 3 × CH3). 13C NMR (100 MHz, CDCl3) δ 172.1,
157.8, 81.5, 63.1, 60.4, 55.7, 38.8, 28.3 (3C), 25.9, 21.2, 20.5.
IR (KBr): 3361, 2975, 2863, 1748, 1736 cm–1. [α]D25 (c 0.04 in DCM): −10.46.
Found: C, 58.3; H, 7.0%. Calc. for C13H19NO5: C, 58.0; H, 7.1%. Mp > 180.0 °C, dec.
To a solution
of 7b (117 mg, 0.4 mmol) in 3 mL of DCM at 0 °C
is added 1.80 mL of 1,2M DIBAL-H in toluene. The reaction is stirred
for 4 h at room temperature. Methanol is added (10 mL). The salts
are filtered and rinsed with methanol (2 × 10 mL). The solvent
is evaporated in vacuo. The crude is purified in
silica gel using MeCN as the eluent. A yellow wax is obtained (54
mg, 62%). 1H NMR (400 MHz, MeOD) δ 4.48 (dd, J = 8.5, 5.8 Hz, 1H, CH2O), 4.32 (t, J = 8.8 Hz, 1H, CH2O), 4.16 (dd, J = 8.1, 3.8 Hz, 1H, CHOH), 3.83 (d, J = 12.9 Hz, 1H, CH2N), 3.69 (ddd, J =
9.4, 5.8, 3.8 Hz, 1H, CHN), 3.54 (dd, J = 11.2, 6.4
Hz, 1H, HOCH2Ccyclopropane), 3.42 (dd, J = 11.3, 6.7 Hz, 1H, HOCH2Ccyclopropane),
3.36 (dd, 12.9, 4.7 Hz, 1H, CH2N), 1.42 (td, J = 8.4, 4.9 Hz, 1H, CHCHOH), 1.30–1.20 (m, 1H,
HOCH2CHcyclopropane), 1.19–1.09 (m,
1H, CHCH2N). 13C NMR (100 MHz,
MeOD) δ 160.4, 65.6, 65.0, 61.4, 57.2, 40.2, 21.9, 21.0, 16.5.
IR (NaCl): 3384, 2991, 2888, 1705 cm–1. [α]D25 (c 0.01 in MeOH): −35.3.
Found: C, 54.5; H, 6.5%. Calc. for C9H13NO4: C, 54.3; H, 6.6%.
General Procedure for Enzymatic
Reactions
Glycosidase
activities were assessed in 80 μL reaction volumes in Eppendorf
vials. Buffer composition and enzyme concentration were adjusted depending
on the enzyme assayed: 20 mM Na2HPO4 at pH 7.3
for β-glucosidase (3 μg/mL) and β-galactosidase
(1 μg/mL); 20 mM Na2HPO4 at pH 6.8 for
α-glucosidase (1 μg/mL) and α-galactosidase (20
μM); 20 mM NaH2PO4 at pH 5.5 for α-
and β-mannosidase (7 and 2 μM respectively); 0.1 M NaOAc
at pH 4.0 with 1 mg/mL of bovine serum albumin (BSA) for α-l-fucosidase (2 μM); and 50 mM NaOAc at pH 5.0 for neuraminidase
(6 μM). The inhibitors were tested at 1, 5, and 25 mM final
concentrations in the assays. Each enzyme mixture and inhibitor were
homogenized and preincubated for 10 min at 37 or 40 °C (α-l-fucosidase). Each reaction was initiated and brought to a
final volume of 80 μL, by addition of an aliquot of the corresponding p-nitrophenyl glycoside substrate to obtain the following
final concentrations in the reaction mixtures: p-nitrophenyl
α- and β-d-glucopyranoside (1 mM), p-nitrophenyl α- and β-d-galactopyranoside (0.5
mM), p-nitrophenyl α- and β-d-mannopyranoside (1 mM), p-nitrophenyl α-l-fucopyranoside (1 mM), or p-nitrophenyl neuraminic
acid (1 mM). After 10 min of incubation time at the same temperature,
each reaction was quenched with 400 μL of 1.0 M Na2CO3, and the absorbance at 405 nm was measured. Assays
were repeated twice and data were averaged.The residual activity
of each enzyme was calculated by the ratio of the absorbance measured
after 10 min of reaction in the presence and absence of synthesized
compounds. The equation used to calculate Ki was derived from Michaelis–Menten, where Vi is the absorbance measured in the absence of the synthesized
compounds; V is the absorbance when the compounds
were added to the enzymatic reaction; Km indicates the Michaelis–Menten constant for each enzyme;
and [I] is the concentration of the synthesized compounds (5 mM) and
[S] is the concentration of the substrate (eq ).
Authors: Graeme Horne; Francis X Wilson; Jon Tinsley; David H Williams; Richard Storer Journal: Drug Discov Today Date: 2010-09-21 Impact factor: 7.851
Authors: Lionel Moisan; Pierre Thuéry; Marc Nicolas; Eric Doris; Bernard Rousseau Journal: Angew Chem Int Ed Engl Date: 2006-08-11 Impact factor: 15.336
Authors: Steven S Henry; Molly D Brady; Dana L T Laird; J Craig Ruble; David L Varie; James A Monn Journal: Org Lett Date: 2012-05-21 Impact factor: 6.005
Authors: Rupert J Russell; Lesley F Haire; David J Stevens; Patrick J Collins; Yi Pu Lin; G Michael Blackburn; Alan J Hay; Steven J Gamblin; John J Skehel Journal: Nature Date: 2006-08-16 Impact factor: 49.962
Authors: Thomas J M Beenakker; Dennis P A Wander; Wendy A Offen; Marta Artola; Lluís Raich; Maria J Ferraz; Kah-Yee Li; Judith H P M Houben; Erwin R van Rijssel; Thomas Hansen; Gijsbert A van der Marel; Jeroen D C Codée; Johannes M F G Aerts; Carme Rovira; Gideon J Davies; Herman S Overkleeft Journal: J Am Chem Soc Date: 2017-05-05 Impact factor: 15.419