Preparation of both C5′ epimers of 5′- C -methyladenosine: Reagent control trumps substrate control

Article abstract of DOI:10.1021/jo500089t

Adenosine-derived ketone 5 was subjected to Noyori asymmetric transfer hydrogenation (ATH) using aqueous sodium formate as a stoichiometric reductant. Despite the well-known sensitivity of ATH to stereoelectronic effects from a contiguous stereogenic center, the 5′ stereochemistry was overwhelmingly controlled by the chirality of the catalyst. Both the (5′S,4′R) and the (5′R,4′R) diastereomers could be prepared selectively in good yields. An efficient three-step route that provides ketone 5 in 75% overall yield was developed.

Full text of DOI:10.1021/jo500089t

Note
pubs.acs.org/joc
Preparation of Both C5′ Epimers of 5′‑C‑Methyladenosine: Reagent
Control Trumps Substrate Control
Cavan M. Bligh, Luigi Anzalone, Young Chun Jung, Yuegang Zhang, and William A. Nugent*
Vertex Pharmaceuticals, Inc., 50 Northern Avenue, Boston, Massachusetts 02210, United States
S
* Supporting Information
ABSTRACT: Adenosine-derived ketone 5 was subjected to
Noyori asymmetric transfer hydrogenation (ATH) using
aqueous sodium formate as a stoichiometric reductant. Despite
the well-known sensitivity of ATH to stereoelectronic effects
from a contiguous stereogenic center, the 5′ stereochemistry
was overwhelmingly controlled by the chirality of the catalyst.
Both the (5′S,4′R) and the (5′R,4′R) diastereomers could be
prepared selectively in good yields. An efficient three-step route that provides ketone 5 in 75% overall yield was developed.
Scheme 1. Synthesis of Ketone 5 from Protected Adenosine
2
(5′S)-C-Methyladenosine (1a) and its (5′R) diastereomer 1b
have been of interest to medicinal chemists for over 50 years.1
These nucleosides and their phosphoric esters typically remain
functionally competent analogues of adenosine, ADP, and ATP
that allow systematic variation of their solution conformations.2
Consequently, they have long played an important role as
structural probes in molecular biology and enzymology.3
Existing routes1−5 to 1a and 1b require chromatography,
which complicates scale-up. In this paper, we report a novel
approach to 1a and 1b in which a protected adenosine
derivative is converted to methyl ketone 5 in three
straightforward steps6 and is then subjected to stereoselective
reduction.7
generate ketone 5 and allow addition of a second equivalent of
Grignard reagent.15
Our approach to ketone 5 is summarized in Scheme 1 (AdBz
= 6-N-benzoyladenine).8 TEMPO-catalyzed oxidation of the
readily available9 protected adenosine derivative 2 does not
stop at the aldehyde oxidation stage but proceeds directly to the
known10 carboxylic acid 3. This observation is extensively
precedented11 and presumably reflects the tendency of the
electron-deficient aldehyde intermediate to exist as a readily
oxidizable hydrate. Carboxylic acid 3 was conveniently isolated
as its sodium salt and was then converted to the known12
Weinreb amide 4 using T3P/pyridine13 as a coupling reagent.
The conversion of 4 to the known14 ketone 5 requires 2
equiv of methylmagnesium bromide: 1 equiv to deprotonate
the amide N−H and 1 equiv for the nucleophilic substitution of
the Weinreb amide. It proved advantageous to delay the
addition of the second equivalent of Grignard reagent, allowing
time for the deprotonation of the amide nitrogen to proceed to
completion. Presumably this protocol avoids premature
protonolysis of the Weinreb addition product, which would
Several approaches to the reduction of ketone 5 were
explored. As previously reported,14 reduction with sodium
borohydride proved unselective, producing a 2:1 mixture of
diastereomeric alcohols at the 5′ position. Attempted
Meerwein−Pondorf reduction with either aluminum isoprop-
oxide or indium isopropoxide16 instead resulted in removal of
the benzamide protecting group. Reduction with L-Selectride
proved unselective, while the attempted reduction using the
CBS catalyst17 resulted in partial decomposition of ketone 5.
Better results were obtained using lithium tri-tert-butox-
yaluminum hydride as the reductant. As shown in Scheme 2,
the enantioselectivityand even the predominant absolute
configuration of the productshowed an unusual temperature
dependence.18 The major product obtained at room temper-
Received: January 14, 2014
Published: March 18, 2014
© 2014 American Chemical Society
3238
dx.doi.org/10.1021/jo500089t | J. Org. Chem. 2014, 79, 3238−3243

The Journal of Organic Chemistry
Note
with a dynamic kinetic resolution (DKR) process, it provides
highly useful ATH/DKR reactions that allow control of the
absolute stereochemistry at contiguous stereogenic centers.27
Such reactions depend on the fact that ATH for one
enantiomer is significantly slower than interconversion between
the enantiomers. Since we wished to retain the biologically
relevant (4′R) stereochemistry, it was essential to avoid such a
DKR process.
Scheme 2. Effect of Temperature on the Reduction of
Ketone 5
A small (3 × 3) combinatorial study28 was undertaken to
identify a suitable combination of ligand and metal precursor
for the ATH of ketone 5. Under the conditions of this screen
(25 °C, 28:1 substrate/catalyst ratio, sodium formate as the
stoichiometric reductant, two-phase water/ethyl acetate solvent
system), the highest enantioselecitivity was observed using (p-
cymene)ruthenium(II) dichloride dimer as the metal precursor.
Using this precursor in combination with (S,S)-TsDPEN
predominantly afforded the (5′S) alcohol 6a (6a/6b = 94:6),
while using (R,R)-TsDPEN gave the reverse selectivity (6a/6b
= 8:92).
Diastereoselectivity in the reduction of ketone 5 was
expected to reflect both reagent control stemming from the
chirality of the catalyst and substrate control resulting from the
inherent chirality of the rigid furanose substrate. Thus, the
minor difference in stereoselectivity when the results for (S,S)-
TsDPEN and (R,R)-TsDPEN were compared was unexpected.
In contrast to the reduction by LiAlH(OtBu)3, substrate
control apparently had little impact on the stereochemical
course of ATH. Consistent with this proposal, reduction of
ketone 5 using the corresponding Ru catalyst bearing an achiral
(N-tosyl)ethylenediamine ligand afforded a 52:48 mixture of 6a
and 6b.
ature could be enriched to 98% diastereomeric purity after a
single crystallization from heptane/ethyl acetate.
In order to determine the absolute stereochemistry at C4′
and C5′ of the major product at room temperature (6a), X-ray
crystal structure analysis was carried out. The molecular
structure clearly shows that 6a has the (5′S,4′R) configuration
(see the Supporting Information).
We turned our attention to asymmetric transfer hydro-
genation (ATH) using Noyori’s N-toluenesulfonyl-1,2-diphe-
nylethylenediamine (TsDPEN) ligands.19 In recent years, the
replacement of formic acid/triethylamine with aqueous sodium
formate as a stoichiometric reductant has significantly improved
the operational simplicity of ATH reactions.20
The absolute configuration of the products obtained in the
reduction of ketone 5 is the same as that observed for the Ru-
TsDPEN-catalyzed ATH of aromatic substrates such as
acetophenone. For the ATH of acetophenone, several
theoretical studies have established that a key control element
is an attractive π-interaction between the aryl group of the
ketone and the electron-deficient metal arene moiety.29 By
analogy, we tentatively ascribe our results to an attractive
interaction between the negative electron density around the
lone electron pairs of the tetrahydrofuran oxygen atom and the
partial positive charge surrounding the η6 aromatic ring
(Scheme 4).30,31 Similar arguments have been put forth to
rationalize the ATH of other heteroatom-containing ketones.32
We employed the Ru−TsDPEN catalyst system for the
synthesis of both 6a and 6b. In the case of the (5′S) alcohol 6a,
complete conversion was observed in 24 h at room temperature
using a 1% catalyst loading. The resulting 95:5 mixture of
diastereomers could be freed of 6b and catalyst residues by
trituration in hot 60:40 ethyl acetate/heptane. Upon cooling,
the product was collected by filtration to afford 6a in 82% yield
with 99% diastereomeric purity.
Most of the published examples of ATH reactions using
TsDPEN-type ligands involve differentiation between saturated
and unsaturated ketone groups. The unsaturated substituent
has most often been an aromatic ring, but examples involving
alkene21 and alkyne22 substituents are also known. Moreover,
during the past decade examples have begun to appear in which
heteroatoms attached to the α-carbon of the ketone carbonyl
serve as effective control elements.23 While we are unaware of
examples in which a nonaromatic heterocycle serves in this
role,24 2-acetylfuran is an excellent substrate for ATH, and this
reaction has been utilized in organic synthesis.25
The presence of pre-existing chirality in ketone 5 was a
concern since ATH is usually26 sensitive to such chiral centers.
As illustrated in Scheme 3, when this sensitivity is combined
Scheme 3. Typical ATH/DKR Reaction
Similar reaction conditions were successfully applied to
prepare 6b. However, the trituration protocol described above
could not be applied to this lower-melting product.
Consequently, an alternative isolation was developed that
takes advantage of the tendency of 6b to reversibly form a
poorly soluble hydrate in the presence of bulk water. The crude
product was dissolved in 20 volumes of ethyl acetate and was
stirred with 10 volumes of water at room temperature. The
amorphous product was collected by filtration and after drying
afforded 6b in 72% yield with 98% diasteromeric purity.
3239
dx.doi.org/10.1021/jo500089t | J. Org. Chem. 2014, 79, 3238−3243

The Journal of Organic Chemistry
Note
with in situ-generated catalysts, while synthetic runs utilized
commercial samples of pregenerated catalysts.
Scheme 4. Proposed Model for the Absolute Stereochemistry
in ATH of Ketone 5
Synthetic Details. Preparation of Carboxylic Acid 3 as Its
Sodium Salt Monohydrate. A 3 L jacketed round-bottom flask was
equipped with an overhead stirrer, nitrogen inlet, and temperature
probe. The flask was charged with technical grade alcohol 2 (48.75 g,
92 wt %, 109.0 mmol) and dichloromethane (270 mL). A solution of
sodium hydrogen carbonate (12.81 g, 152.5 mmol) in water (240 mL)
was added, and the mixture was cooled to 5 °C with stirring. TEMPO
(3.476 g, 21.80 mmol) was added all at once, after which a 9.42 wt %
solution of sodium hypochlorite (206.7 g, 261.6 mmol) was added
dropwise over 2 h, keeping the internal temperature below 7 °C. After
about half of the bleach was added, solids began to separate from
solution. Additional dichoromethane (180 mL) was added, and the
temperature was allowed to rise to 15 °C, at which time the solids
were collected by filtration. The flask and filter cake were rinsed with
water (90 mL), and the solids were pulled dry for 1 h. The resultant
red/brown solids were suspended in dichloromethane (225 mL) and
stirred overnight. The resulting material was collected by filtration,
rinsed with dichloromethane (90 mL), and dried for 3 days in a
vacuum oven at 55 °C to afford the sodium salt of 3 (44.35 g, 87%
yield) as an off-white solid that was 98.7% pure by HPLC and
contained 1.0 molar equiv of water as determined by NMR analysis.
It proved convenient to use this sodium salt directly in the
subsequent Weinreb amide formation. However, in order to allow
comparison with literature data, the salt was converted to the free
carboxylic acid 3 for purposes of characterization. This was
accomplished by dissolving 2.00 g of the sodium salt in 20 mL of
water and acidifying resulting solution to pH 1 with 3 N HCl. The
precipitated solid was removed by filtration, washed with water, and
dried overnight in a 60 °C vacuum oven to afford the acid 3 (1.60 g,
87%) as a white solid. The product could be recrystallized from 98:2
In conclusion, ATH of ketone 5 using the appropriate
enantiomer of the Ru−TsDPEN catalyst allows straightforward
synthesis of either diastereomer of protected 5′-C-Me-
adenosine. Compounds 6a and 6b can be deprotected using
known protocols14 to afford the adenosine derivatives 1a and
1b, respectively. In addition, we have developed an efficient
three-step synthesis of the ketone 5 starting material from
protected adenosine 2 in 75% overall yield.
ethanol/water. Mp: 216.8−219.9 °C (lit.10b 208−209 °C). H NMR
1
It is perhaps not surprising that, under the mild conditions of
the ATH reaction using aqueous sodium formate, epimerization
at C4′ is not observed. Nevertheless, given the general
observation that ATH is sensitive to stereoelectronic effects
from contiguous stereogenic centers, and particularly in view of
the considerable steric bulk of the protected nucleoside in 5, we
were surprised to find that ATH proceeds readily with either
enantiomer of the catalyst. Finally, our successful use of an
ether oxygen atom as a control element provides another
example of the rapidly growing list of substrate classes that
undergo selective ATH reactions.
(400 MHz, DMSO-d6): δ 1.38 (s, 3H), 1.54 (s, 3H), 4.77 (d, 1H, J =
0.9 Hz), 5.58 (s, 2H), 6.47 (s, 1H), 7.55 (m, 2H), 7.65 (m, 1H), 8.04
(m, 2H), 8.59 (s, 1H), 8.69 (s, 1H), 11.18 (br s, 1H), 12.39 (br s, 1H).
13C NMR (100 MHz, DMSO-d6): δ 24.0, 26.4, 83.4, 83.8, 85.7, 89.8,
112.7, 125.3, 128.1, 128.41, 128.45, 132.4, 133.4, 144.0, 150.2, 151.3,
152.0, 165.5, 170.7. HRMS (ESI-TOF) m/z: [M + H]+ calcd for
+
C20H20N5O6 426.1408, found 426.1428. The spectroscopic data
match those previously reported.10b
Preparation of Weinreb Amide 4. A 1 L jacketed reactor was
equipped with an overhead stirrer, temperature probe, and nitrogen
inlet. The reactor was charged with the sodium salt monohydrate of
carboxylic acid 3 (20.0 g, 45.0 mmol), N,O-dimethylhydroxylamine
hydrochloride (5.04 g, 52 mmol), ethyl acetate (60 mL), and pyridine
(20 mL). The resultant thick slurry was cooled with stirring to 0 °C. A
50 wt % solution of T3P in ethyl acetate (56 mL, 94 mmol) was slowly
added by syringe over the course of 30 min, keeping the internal
temperature below 5 °C. Stirring was continued for an additional 2 h
at 0 °C, at which time the reaction was quenched by addition of 20%
aqueous citric acid (80 mL). The product was extracted into ethyl
acetate (200 mL then 100 mL), and the combined organic layers were
washed with half-saturated sodium bicarbonate (80 mL) and then
water (80 mL). The solvent was distilled at reduced pressure to afford
crude 4, which was twice azeotropically dried by addition of
dichloromethane (120 mL) followed by distillation at reduced
pressure. This afforded 4 (19.4 g, 92%) as a pale-yellow foam that
retained a small amount of ethyl acetate. 1H NMR (400 MHz,
CDCl3): δ 1.30 (s, 3H), 1.51 (s, 3H), 3.02 (s, 3H), 3.60 (s, 3H), 5.16
(m, 2H), 5.33 (m, 1H), 6.16 (m, 1H), 7.33 (m, 2H), 7.42 (m, 1H),
8.00 (m, 2H), 8.48 (s, 1H), 8.62 (s, 1H), 9.37 (br s, 1H). 13C NMR
(100 MHz, CDCl3): δ 25.3, 27.0, 32,3, 61.9, 83.2, 83.7, 85.1, 91.7,
114.2, 122.9, 127.9, 128.8, 132.7, 133.8, 142.3, 149.5, 151.8, 152.6,
164.7, 169.5. HRMS (ESI-TOF) m/z: [M + H]+ calcd for
EXPERIMENTAL SECTION
■
General Information. NMR spectra were determined at 25 °C at
a field strength of 400 MHz (1H spectra) or 100 MHz (13C spectra).
The chemical shifts are expressed in parts per million (δ) relative to
tetramethylsilane, and coupling constants (J) are given in hertz. The
following abbreviations are used: s = singlet, d = doublet, t = triplet, q
= quartet, br = broad. The starting material, 2′,3′-isopropylidene-6-N-
benzoyladenosine (2), was prepared by the procedure of van Delft and
co-workers9a or was purchased from commercial sources. Anhydrous
tetrahydrofuran and all of the synthetic reagents were commercial
materials and were used as received. Flash chromatography was
performed on 220−400 mesh silica following the standard procedure
of Still.33 The reactions reported below were run under an atmosphere
of dry nitrogen. Melting points were determined on an automated
apparatus employing digital image processing technology. The
following HPLC method was employed in all of the work reported
below: Agilent Poroshell SB-C18 column, 4.6 mm × 150 mm (2.7 μm
particle size); mobile phase A = water with 0.1% TFA, mobile phase B
= methanol with 0.1% TFA, gradient 30−70% B in 20 min, total 29
min; flow rate = 0.8 mL/min; 70:30 water/methanol as the diluent;
column temperature 40 °C; UV detection at 260 nm. Retention times
were as follows: carboxylic acid 3 = 14.8 min, Weinreb amide 4 = 16.5
min, ketone 5 = 15.4 min, alcohol 6a = 16.7 min, alcohol 6b = 16.9
min. The initial catalyst screen for the ATH reactions was carried out
+
1
C22H25N6O6 469.1830, found 469.1853. The H NMR data match
those previously reported.12
Preparation of Ketone 5. A 250 mL jacketed reactor was equipped
with a magnetic stirrer, a nitrogen inlet, and a temperature probe. The
reactor was charged with Weinreb amide 4 (18.5 g, 39.5 mmol) and
3240
dx.doi.org/10.1021/jo500089t | J. Org. Chem. 2014, 79, 3238−3243

The Journal of Organic Chemistry
Note
tetrahydrofuran (266 mL), and the stirred mixture was cooled to −5
°C. A 3.0 M solution of methylmagnesium chloride in THF (13.2 mL,
39.6 mmol) was added at a rate such that the internal temperature did
not rise above −2 °C, after which the mixture was stirred for an
additional 1 h at −18 °C. A second portion of 3.0 M
methylmagnesium chloride in THF (17.1 mL, 51.3 mmol) was
added at a rate such that the internal temperature did not rise above
−4 °C, and the mixture was again stirred for 1 h at −18 °C. The
reaction was carefully quenched (caution: gas evolution) by addition of
half-saturated aqueous ammonium chloride (260 mL), and the
product was extracted into ethyl acetate (185 mL and then 90 mL).
The combined organic layers were washed with water (185 mL) and
dried over magnesium sulfate. The solvent was removed on a rotary
evaporator initially at 50 °C and 300 mbar to promote azeotropic
distillation of water. Completion of the distillation at full vacuum
resulted in a gooey residue that retained considerable ethyl acetate.
MTBE (100 mL) was added, and the flask was spun on the rotary
evaporator in a 50 °C bath for 1 h. Distillation of the solvent at
reduced pressure then afforded ketone 5 (15.7 g, 94%) as a crisp white
mixture was heated to 80 °C for 15 min, and the still-heterogeneous
mixture was allowed to cool to room temperature with stirring. The
solids were collected by filtration and dried in a vacuum oven
overnight at 60 °C to afford 6a as an off-white solid. HPLC analysis
indicated that the 5′S/5′R diastereomeric ratio was 99:1. Mp 165.7−
166.6 °C. 1H NMR (400 MHz, DMSO-d6): δ 1.12 (d, 3H, J = 6 Hz),
1.35 (s, 3H), 1.58 (s, 3H), 3.83 (m, 1H), 4.09 (dd, 1H, J = 4, 3 Hz),
4.98 (dd, 1H, J = 6, 3 Hz), 5.16 (br s, 1H), 5.34 (dd, 1H, J = 6, 3 Hz),
6.30 (d, 1H, J = 3 Hz), 7.56 (m, 2H), 7.65 (m, 1H), 8.06 (m, 2H),
8.77 (s, 1H), 8.78 (s, 1H), 11.23 (s, 1H). 13C NMR (100 MHz,
CDCl3): δ 19.5, 25.3, 27.6, 68.4, 82.6, 82.8, 88.9, 93.5, 114.2, 124.3,
128.0, 128.8, 132.9, 133.5, 142.6, 150.3, 150.6, 152.3, 164.7. HRMS
+
(ESI-TOF) m/z: [M + H]+ calcd for C21H24N5O5 426.1772, found
426.1793.
2′,3′-O,O-Isopropylidene-6-N-benzoyl-(5′R)-5′-C-methyladeno-
sine (6b). A 100 mL round-bottom flask equipped with a septum-
covered side arm was charged with η6-(p-cymene)-(R,R)-N-toluene-
sulfonyl-1,2-diphenylethylenediamine(1−)ruthenium(II) chloride (15
mg, 0.024 mmol, 1%), and ketone 5 (1.00 g, 2.4 mmol), and the
system was flushed with nitrogen. A solution of sodium formate (6.8 g,
100 mmol) in water (40 mL) was added, followed by ethyl acetate (10
mL). The resulting two-phase mixture was stirred for 30 h at room
temperature, during which time some of the solid product separated
from solution. Sufficient dichloromethane (ca. 30 mL) to redissolve
the product was added, and the two homogeneous phases were
transferred to a separatory funnel. The organic layer was separated,
and the solvent was distilled using a rotary evaporator. The residue
(5′R/5′S = 92:8) was transferred to a straight-walled vessel and
dissolved in ethyl acetate (20 mL), after which water (10 mL) was
added. The mixture was stirred overnight at room temperature, during
which time a white solid separated and the colored catalyst residue
remained dissolved in the organic phase. The solid was collected by
filtration, washed with MTBE (2 × 10 mL), and finally dried at 15
Torr to afford 6b (723 mg, 72%) as an amorphous snow-white
powder. HPLC analysis indicated that the 5′R/5′S diastereomeric ratio
1
foam. H NMR (400 MHz, DMSO-d6): δ 1.38 (s, 3H), 1.56 (s, 3H),
1.86 (s, 3H), 4.78 (d, 1H, J = 2.3 Hz), 5.46 (d, 1H, J = 6.4 Hz), 5.56
(dd, 1H, J = 6.4, 2.3 Hz), 6.51 (s, 1H), 7.55 (m, 2H), 7.65 (m, 1H),
8.06 (m, 2H), 11.24 (br s, 1H). 13C NMR (100 MHz, CDCl3): δ 25.2,
26.0, 26.8, 82.7, 84.2, 91.4, 93.2, 123.3, 128.0, 128.8, 132.9, 133.5,
142.6, 149.9, 151.0, 152.5, 164.8, 205.0. HRMS (ESI-TOF) m/z: [M +
+
H]+ calcd for C21H22N5O5 424.1615, found 424.1633. The
spectroscopic data match those previously reported.14
In the original literature report on ketone 5,14 it was noted that this
compound epimerized to the extent of 80% upon attempted
chromatography on “silicic acid”. We did not observe any degradation
upon flash chromatography using ethyl acetate as the eluent. In fact,
flash chromatography was found to be a useful method for purifying
compound 5 and could remove minor impurities that otherwise would
make the downstream purification of alcohols 6a and 6b less efficient.
ATH Catalyst Screening. The screen was conducted as a small (3 ×
3) combinatorial study. Ligand inputs were (S,S)-TsDPEN, (R,R)-
TsDPEN, and (S,S,S)-CsDPEN.28 Metal precursor inputs were (p-
cymene)ruthenium(II) dichloride dimer, (pentamethylcyclopenta-
dienyl)rhodium(III) dichloride dimer, and (pentamethylcyclopenta-
dienyl)iridium(III) dichloride dimer. A set of borosilicate glass tubes
equipped with Teflon screw caps were taken into a nitrogen-filled
glovebox. Each tube was charged with transition metal precursor
(0.0090 mmol of the dimer, delivering 0.018 mg-atom of metal),
ligand (0.021 mmol), and deaerated water (10 mL). The resulting
mixture was stirred at 40 °C for 1 h to afford a homogeneous solution.
To each tube was added a solution of ketone 5 (211 mg, 0.500 mmol)
in ethyl acetate (2 mL) followed by sodium formate (1.70 g, 25
mmol), and the two-phase mixture was stirred under nitrogen for 24 h
at room temperature. The contents of each tube was transferred to a
separatory funnel with ethyl acetate (15 mL). The organic phase was
separated, and the aqueous phase was further washed with ethyl
acetate (5 mL). HPLC analysis of the combined organic phases was
used to determine the conversion and yield of alcohols 6a and 6b. In
all cases, the conversion was >99%. A tabulation of the observed
stereoselectivities is given in the Supporting Information.
1
was 98:2. H NMR (400 MHz, DMSO-d6): δ 1.04 (d, 3H, J = 6 Hz),
1.36 (s, 3H), 1.57 (s, 3H), 3.75 (m, 1H), 3.99 (dd, 1H, J = 5, 2 Hz),
5.10 (dd, 1H, J = 6, 2 Hz), 5.18 (br s, 1H), 5.42 (dd, 1H, J = 6, 3 Hz),
6.27 (d, 1H, J = 3 Hz), 7.56 (m, 2H), 7.65 (m, 1H), 8.06 (m, 2H),
8.69 (s, 1H), 8.77 (s, 1H), 11.23 (s, 1H). 13C NMR (100 MHz,
CDCl3): δ 18.1, 25.2, 27.5, 67.3, 79.4, 83.0, 89.7, 93.6, 114.5, 124.4,
127.9, 128.9, 132.9, 133.4, 142.5, 150.3, 150.5, 152.3, 164.6. HRMS
+
(ESI-TOF) m/z: [M + H]+ calcd for C21H24N5O5 426.1772, found
426.1778.
Reduction of Ketone 5 with Lithium Tri-tert-butoxyaluminum
Hydride. A three-neck round-bottom flask was equipped with a
temperature probe, an addition funnel, and a nitrogen inlet. To the
flask was added a solution of crude ketone 5 (10.0 g, 80 wt % purity,
18.9 mmol) in anhydrous THF (170 mL), and the system was flushed
with nitrogen. A 1.0 M solution of lithium tri-tert-butoxyaluminum
hydride in THF (47.2 mL, 47.2 mmol) was added dropwise over the
course of 5 min, during which time the internal temperature rose from
17.4 to 22.4 °C. The mixture was stirred for 1 h at room temperature
and was then subjected to an inverse quench by addition to 5%
aqueous oxalic acid (100 mL). The mixture was concentrated at 50 °C
on a rotary evaporator until most of the THF had distilled over. The
aqueous residue was extracted with ethyl acetate (100 mL), and the
organic phase was stirred for 30 min with aqueous sodium bicarbonate
(100 mL), after which the organic phase was dried over sodium sulfate.
In order to azeotrope off any remaining water, the solution was
distilled on a rotary evaporator at 50 °C bath temperature and 300
mbar pressure. Two additional 100 mL portions of ethyl acetate were
added and distilled off to further dry the residue (5′S/5′R = 88:12).
The crude product (9.86 g) was suspended in ethyl acetate (60 mL)
and heated to reflux to produce a clear solution. Heptane (35 mL) was
added, keeping the temperature above 60 °C. The heating mantle was
unplugged, and the mixture was allowed to slowly cool to room
temperature, during which time a white solid separated. The solid was
collected by filtration and dried at 60 °C for 3 h in a vacuum oven with
a nitrogen sweep to afford 6a (3.96 g, 49%) as an off-white solid.
2′,3′-O,O-Isopropylidene-6-N-benzoyl-(5′S)-5′-C-methyladeno-
sine (6a). A 100-mL round-bottom flask equipped with a septum-
covered side arm was charged with η6-(p-cymene)-(S,S)-N-toluene-
sulfonyl-1,2-diphenylethylenediamine(1−)ruthenium(II) chloride (15
mg, 0.024 mmol, 1%) and ketone 5 (1.00 g, 2.4 mmol), and the system
was flushed with nitrogen. A solution of sodium formate (6.8 g, 100
mmol) in water (40 mL) was added, followed by ethyl acetate (10
mL). The resulting two-phase mixture was stirred for 24 h at room
temperature, during which time some of the solid product separated
from solution. The solids were redissolved with additional ethyl acetate
(20 mL). The organic phase was separated, and the aqueous phase was
extracted with another 10 mL of ethyl acetate. The solvent was
removed from the combined organic layers at reduced pressure on a
rotary evaporator. The solid residue (5′S/5′R = 95:5) was suspended
in a mixture of 12 mL of ethyl acetate and 8 mL of heptane. The
3241
dx.doi.org/10.1021/jo500089t | J. Org. Chem. 2014, 79, 3238−3243

The Journal of Organic Chemistry
Note
HPLC analysis indicated that the 5′S/5′R diastereomeric ratio was
98:2. The spectroscopic properties were identical to those reported
above for alcohol 6a.
(15) For similar reasons, it is recommended to quench the reaction
by inverse addition of the reaction mixture to aqueous acid.
(16) Lee, J.; Ryu, T.; Park, S.; Lee, P. H. J. Org. Chem. 2012, 77,
4821−4825.
(17) Corey, E. J.; Bakshi, R. K.; Shibata, S. J. Am. Chem. Soc. 1987,
109, 5551−5553.
ASSOCIATED CONTENT
■
S
* Supporting Information
(18) In contrast, changing the reaction solvent from THF to diethyl
ether, toluene/THF, or dichloromethane/THF had essentially no
effect on the stereoselectivity.
(19) Hashiguchi, S.; Fujii, A.; Takehara, J.; Ikariya, T.; Noyori, R. J.
Am. Chem. Soc. 1995, 117, 7562−7563.
Tabulated results for the ATH catalyst screen, details of the
structure assignment for 6a and 6b, and H and 13C NMR
1
spectra for all compounds. This material is available free of
charge via the Internet at http://pubs.acs.org.
(20) (a) Wu, X.; Li, X.; Hems, W.; King, F.; Xiao, J. Org. Biomol.
Chem. 2004, 2, 1818−1824. (b) Wu, X.; Li, X.; Zanotti-Gerosa, A.;
Pettman, A.; Lin, J.; Mills, A. J.; Xiao, J. Chem.Eur. J. 2008, 14,
2209−2222. (c) For a review, see: Wu, X.; Xiao, J. Chem. Commun.
2007, 2449−2466.
AUTHOR INFORMATION
Corresponding Author
*E-mail: William_nugent@vrtx.com.
Notes
■
(21) (a) Soni, R.; Collinson, J.-M.; Clarkson, G. C.; Wills, M. Org.
Lett. 2011, 13, 4304−4307. (b) Zhang, Q.-Q.; Xie, J.-H.; Yang, X.-H.;
Xie, J.-B.; Zhou, Q.-L. Org. Lett. 2012, 14, 6158−6161.
(22) (a) Fang, Z.; Wills, M. J. Org. Chem. 2013, 78, 8594−8605.
(b) Arai, N.; Satoh, H.; Utsumi, N.; Murata, K.; Tsutsumi, K.;
Ohkuma, T. Org. Lett. 2013, 15, 3030−3033. (c) Matsumura, K.;
Hashiguchi, S.; Ikariya, T.; Noyori, R. J. Am. Chem. Soc. 1997, 119,
8738−8739.
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We are deeply grateful to Prof. Martin Wills (University of
Warwick) for helpful discussions. We also thank Mr. Eduard
Luss-Lusis for obtaining high-resolution mass spectra.
(23) (a) Sterk, D.; Stephan, M. S.; Mohar, B. Tetrahedron Lett. 2004,
45, 535−537. (b) Sterk, D.; Stephan, M.; Mohar, B. Org. Lett. 2006, 8,
REFERENCES
■
5935−5938. (c) Slungard, S. V.; Krakeli, T.-A.; Thvedt, T. H. K.;
̊
(1) (a) Reist, E. J.; Goodman, L.; Spencer, R. R.; Baker, B. R. J. Am.
Chem. Soc. 1958, 80, 3962−3966. (b) Reist, E. J.; Goodman, L.; Baker,
B. R. J. Am. Chem. Soc. 1958, 80, 5775−5779.
Fuglseth, E.; Sundby, E.; Hoff, B. H. Tetrahedron 2011, 67, 5642−
5650. (d) Perryman, M. S.; Harris, M. E.; Foster, J. L.; Joshi, A.;
Clarkson, G. J.; Fox, D. J. Chem. Commun. 2013, 49, 10022−10024.
(24) The reduction of α-phenoxyacetophenone using a Ru(II)−
TsDPEN catalyst proceeds with 84.4% ee; however, the absolute
stereochemistry of the reduction indicates that the benzene ring of the
acetophenone rather than the ether oxygen atom is the predominant
control element. See: Jolley, K. E.; Zanotti-Gerosa, A.; Hancock, F.;
Dyke, A.; Grainger, D. M.; Medlock, J. A.; Neddon, H. G.; Le Paih, J. J.
M.; Roseblade, S. J.; Seger, A.; Sivakumar, V.; Prokes, I.; Morris, D. J.;
Wills, M. Adv. Synth. Catal. 2012, 354, 2545−2555.
(2) (a) Mikhailov, S. N.; Meshkov, S. V.; Kuznetsov, D. A.; Lysov,
Yu. P.; Gorelik, E. Sh.; Formicheva, M. V.; Beigel’man, L. N.;
Padyukova, N. Sh. Bioorgan. Khim. (USSR) 1989, 15, 969−975.
(b) Follmann, H.; Gremels, G. Eur. J. Biochem. 1974, 47, 187−197.
(3) For seminal examples, see: (a) Hampton, A.; Harper, P. J.; Sasaki,
T. Biochemistry 1972, 11, 4736−4739. (b) Grant, A. J.; Lerner, L. M. J.
Med. Chem. 1979, 22, 1016−1018. (c) Bruns, R. F. Can. J. Physiol.
Pharmacol. 1980, 58, 673−691.
(4) Howgate, P.; Hampton, A. Carbohydr. Res. 1972, 21, 309−315.
(5) Ciuffreda, P.; Loseto, A.; Santaniello, E. Tetrahedron: Asymmetry
2004, 15, 203−206.
(25) (a) Guo, H.; O’Doherty, G. A. Angew. Chem., Int. Ed. 2007, 46,
5206−5208. (b) Croatt, M. P.; Carreira, E. M. Org. Lett. 2011, 13,
1390−1393.
(26) A Cbz-protected (S)-α-aminoalkynone was reported to undergo
enantioselective ATH with either enantiomer of TsDPEN with
retention of stereochemistry at the nitrogen-bearing carbon, providing
an example where the catalyst overrides the influence of a pre-existing
chiral center. Only a 4-fold rate difference between the matched and
mismatched cases was observed. See ref 22c.
(27) For recent examples, see: (a) Goodman, C. G.; Do, D. T.;
Johnson, J. S. Org. Lett. 2013, 15, 2446−2449. (b) Corbett, M. T.;
Johnson, J. S. J. Am. Chem. Soc. 2013, 135, 594−597. (c) Villacrez, M.;
Somfai, P. Tetrahedron Lett. 2013, 54, 5266−5268. (d) Liu, Z.; Schultz,
C. S.; Sherwood, C. A.; Krska, S.; Dormer, P. G.; Desmond, R.; Lee,
C.; Sherer, E. C.; Shpungin, J.; Cuff, J.; Xu, F. Tetrahedron Lett. 2011,
52, 1685−1688.
(28) For details, see the Experimental Section. In addition to (R,R)-
and (S,S)-TsDPEN, this study also examined camphor-derived (S,S,S)-
CsDPEN. See: Li, X.; Blacker, J.; Houson, I.; Wu, X.; Xiao, J. Synlett
2006, 1155−1160.
(29) (a) Yamakawa, M.; Ito, H.; Noyori, R. J. Am. Chem. Soc. 2000,
122, 1466−1478. (b) Yamakawa, M.; Yamada, I.; Noyori, R. Angew.
Chem., Int. Ed. 2001, 40, 2818−2821. (c) Noyori, R.; Yamakawa, M.;
Hashiguchi, S. J. Org. Chem. 2001, 66, 7931−7944. (d) For a review of
aromatic interactions as control elements in stereoselective organic
reactions, see: Krenske, E. H.; Houk, K. N. Acc. Chem. Res. 2013, 46,
979−989.
(30) In contrast, highly enantioselective reductions of saturated
ketones achieved with tethered TsDPEN-type ligands show a reversal
of absolute configuration from that seen with aromatic ketones,
implying that the usual attractive interaction has been replaced by a
(6) For a related strategy, see: McNulty, J.; Grunner, V.; Mao, J.
Tetrahedron Lett. 2001, 42, 5609−5612.
(7) For a related approach, see: Prasad, K. R.; Chandrakumar, A.
Synthesis 2006, 2159−2166.
(8) The sequence in Scheme 1 was scaled up to produce 3 kg of
ketone 5.
(9) Compound 2 is commercially available or it can be prepared
using a variety of published protocols. For example, see: (a) Chladek,
S.; Smrt, J. Collect. Czech. Chem. Commun. 1964, 29, 214−233.
(b) Jawalekar, A. M.; Meeuwenoord, N.; Cremers, J. G. O.; Overkleeft,
H. S.; van den Marel, G. A.; Rutjes, F. D. J. T.; van Delft, F. L. J. Org.
Chem. 2008, 73, 287−290.
(10) (a) Barton, D. H. R.; Gero, S. D.; Quiclet-Sire, B.; Samadi, M. J.
Chem. Soc., Perkin Trans. 1 1991, 981−985. (b) Epp, J. B.; Widlanski,
T. S. J. Org. Chem. 1999, 64, 293−295.
(11) For recent examples, see: (a) Bagmare, S.; Varada, M.; Banerjee,
A.; Kumar, V. A. Tetrahedron 2013, 69, 1210−1216. (b) Volpini, R.;
Marucci, G.; Buccioni, M.; Ben, D. D.; Lambertucci, C.; Lammi, C.;
Mishra, R. C.; Thomas, A.; Cristalli, G. ChemMedChem 2011, 6,
1074−1080. (c) Buchs, B.; Fieber, W.; Vigouroux, E.; Sreenivasachary,
N.; Lehn, J.-M.; Herrmann, A. Org. Biomol. Chem. 2011, 9, 2906−
2919.
(12) Hong, W. Guangzhou Hoagong 2012, 40, 52−53.
(13) T3P is a registered trademark for propylphosphonic anhydride.
For its use with pyridine as a base, see: Dunetz, J. R.; Xiang, Y.;
Baldwin, A.; Ringling, J. Org. Lett. 2011, 13, 5048−5051.
(14) Ketone 5 has been prepared previously by oxidizing the mixed
secondary alcohols. See: Ranganathan, R.; Jones, G. H.; Moffatt, J. G. J.
Org. Chem. 1974, 39, 290−298.
3242
dx.doi.org/10.1021/jo500089t | J. Org. Chem. 2014, 79, 3238−3243

The Journal of Organic Chemistry
Note
different mechanism based on steric repulsion. (a) Morris, D. J.;
Hayes, A. M.; Wills, M. J. Org. Chem. 2006, 71, 7035−7044.
(b) Cheung, F. K.; Lin, C.; Minissi, F.; Criville, A. L.; Graham, M. A.;
Fox, D. J.; Wills, M. Org. Lett. 2007, 9, 4659−4662.
(31) Although the position of the tetrahydrofuran ring strongly
suggests its involvement as the control element, we cannot exclude the
possible involvement of the adenine heterocycle or the acetonide
oxygen atoms.
(32) See refs 23c,d and 24.
(33) Still, W. C.; Kahn, M.; Mitra, A. J. Org. Chem. 1978, 43, 2923−
2955.
3243
dx.doi.org/10.1021/jo500089t | J. Org. Chem. 2014, 79, 3238−3243