A Scalable Synthesis of (R,R)-2,6-Dimethyldihydro-2H-pyran-4(3H)-one

Article abstract of DOI:10.1021/op500135x

A scalable synthesis of (R,R)-2,6-dimethyldihydro-2H-pyran-4(3H)-one is reported. Key to this strategy is the Ti(OiPr)₄-catalyzed Kulinkovich cyclopropanation of silyl protected (R)-ethyl 3-hydroxybutanoate, and subsequent oxidative fragmentati

Full text of DOI:10.1021/op500135x

Article
pubs.acs.org/OPRD
A Scalable Synthesis of (R,R)‑2,6-Dimethyldihydro‑2H‑pyran-
4(3H)‑one
Ian S. Young,* Matthew W. Haley, Annie Tam, Steven A. Tymonko,* Zhongmin Xu, Ronald L. Hanson,
and Animesh Goswami
Chemical Development, Bristol-Myers Squibb Company, One Squibb Drive, New Brunswick, New Jersey 08903, United States
S
* Supporting Information
ABSTRACT: A scalable synthesis of (R,R)-2,6-dimethyldihydro-2H-pyran-4(3H)-one is reported. Key to this strategy is the
Ti(OiPr)4-catalyzed Kulinkovich cyclopropanation of silyl protected (R)-ethyl 3-hydroxybutanoate, and subsequent oxidative
fragmentation of the cyclopropanol. The resulting vinyl ketone intermediate was then subjected to oxidative Heck cyclization to
form the enone substrate required for conjugate addition. A diastereoselective copper-catalyzed Grignard addition procedure was
implemented to install the requisite methyl group, with the inclusion of 1,3-bis(diphenylphosphino)propane and trimethylsilyl
chloride greatly increasing the robustness of this process.
Scheme 1. Route to enone 2 based on Astra Zeneca
publication
INTRODUCTION
■
Dimethyl pyranone 1 was a key intermediate required for the
synthesis of an asset within our portfolio. The chosen strategy
to access 1 relied upon the addition of methyl cuprate to enone
2 (Figure 1). Previous reports of this transformation utilizing
Figure 1. Retrosynthesis of dimethylpyranone 1.
rac-2 proceeded in 87% isolated yield, indicating a high degree
of trans-diastereoselectivity.1 To utilize this strategy for our
needs, an efficient, chiral synthesis of enone 2 was required. In
2010, Astra Zeneca (AZ) published their efforts toward a
scalable synthesis of the antipode of 2.2 Guided by their in-
depth investigations, we originally utilized their route to enone
2 to produce material for initial development work. On the
basis of the previously reported success of the cuprate addition,
it was anticipated that the conversion of 2 to 1 would be
straightforward.
Initially we repeated the AZ sequence (Scheme 1) to deliver
approximately 10 kg of 2. The reaction sequence was slightly
modified to serve our purpose, with the enantiomeric ester 3
serving as the starting material, and the Weinreb amide of 4/5
being utilized as the substrate for vinyl Grignard addition
instead of the dimethylamide. Overall, similar yields to those
reported by AZ were achieved (four-step yield was 14%, 11% at
AZ). Areas for improvement were noted during this campaign,
with the vinyl Grignard addition causing significant challenges.
We found that the outcome of this step was highly dependent
on the quality of the vinyl Grignard, and in addition, cryogenic
temperatures (−20 °C) were necessary. As a result, an
alternative route to vinyl ketone intermediate 6 that did not
require vinyl Grignard could help streamline the large-scale
production of 2. We also hoped to improve the robustness of
the oxidative Heck cyclization where the poor stability of both
intermediate 7 and product 2 led to modest yields. These
stability issues were compounded on scale where degradation
occurred during prolonged reaction holds under acidic
conditions or during the basic workup.
RESULTS AND DISCUSSION
■
Development of Kulinkovich Cyclopropanation Proc-
ess. We set out to evaluate alternative routes towards 2, with
the ultimate goal of identifying a more streamlined synthesis of
vinyl ketone 6. The AZ paper did a thorough job in laying the
groundwork for this endeavor, as many of the logical routes to
6 had already been investigated and could be excluded. In 2005,
Singh utilized 6 to prepare all the stereoisomers of
tarchonanthuslactone.3 Singh’s strategy to 6 was based on
Kulinkovich cyclopropanation and subsequent oxidative
fragmentation of an appropriately protected 3-hydroxybuta-
noate ester (Scheme 2). The reported yields for these
Special Issue: Non-precious Metal Catalysis
Received: April 22, 2014
© XXXX American Chemical Society
A
dx.doi.org/10.1021/op500135x | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
upon addition of triethylamine (2.0 equiv). This process was
scaled to produce high-quality 6 (>450 kg, yield 86−96%), with
aqueous extractions being the only means of purification during
the three-step process. This was a welcome outcome, as further
purification would require reduced-pressure distillation as all of
the intermediates in this sequence are oils. The overall three-
step yield (89% lab scale, 82% multikg scale) to vinyl ketone 6
from ester 3 compares favorably to the yields obtained using
the AZ route (27−48%), and has the additional advantages of
not requiring cryogenic (−20 °C) temperatures or vinyl-
magnesium bromide.
With a scalable route to 6 in place, we set out to further
optimize the oxidative cyclization. The reported yield for this
transformation in the AZ communication was 40%, but this was
a combined yield obtained in two separate operations: (1) the
initial product distillation, and (2) recovery of additional
product by back extraction of the aqueous followed by a second
distillation. This modest yield not only is a testament to the
difficulty of the transformation but is also due to the inherent
properties of 2 (volatility, water solubility, acid/base instability)
complicating the isolation. The sensitivity of 2 towards acid/
base led us to develop a process that did not require additional
acid for the removal of the TBS-group.
A 1985 report by Lipshutz includes the serendipitous result
that PdCl2(MeCN)2 can remove TBS-groups from alcohols.6 A
follow up report by Keay suggested that the effectiveness of this
deprotection is enhanced by the inclusion of a few equivalents
of water.7 Since this is the same catalyst utilized in our oxidative
cyclization, we were curious to see how the transformation
would proceed without additional acid. Performing the reaction
via a modified AZ procedure (5 equiv of water were used
instead of aq HCl) led to rapid removal of the TBS-group to
liberate the free-alcohol. This compound then efficiently
cyclized to form compound 2. In-process yields as high as
81% have been achieved for this acid-free transformation on
kilogram scale, but the yield for the product after workup was
only slightly improved (47%, Scheme 4). The discrepancy
Scheme 2. Literature route to vinyl ketone 6 based on
Kulinkovich cyclopropanation/oxidative fragmentation
transformations were very high (87% and 95% respectively),
although no procedures were reported. Intrigued by this result,
we began to evaluate this strategy with regards to scalability and
reproducibility. If this approach could be used on scale, it would
have the advantage of obviating the requirement for the vinyl
Grignard reagent, the inconsistent quality of which caused
significant problems during the previous scale-up campaign.
During initial lab-scale investigations (<250 g), alcohol 34
was protected as the TBS-silyl ether, and then a 2-MeTHF
solution of this compound was treated with stoichiometric
titanium(IV) isopropoxide (1.1 equiv) and ethyl magnesium
bromide (3.0 equiv). After complete consumption of the
starting material, quench of the reaction with aqueous
ammonium chloride led to a large amount of black precipitate
which was removed by filtration. This reaction performed as
reported, and after workup the product was isolated in very
high purity and yield (>90%) without further purification. To
lessen the burden of the filtration on scale, a catalytic version of
this transformation was developed (0.4 equiv Ti(OiPr)4, 2.40
equiv of EtMgBr) to minimize titanium-derived salt formation5
(Scheme 3). This procedure was scaled to produce 73.7 kg of 8
Scheme 3. Kilogram-scale conditions for conversion of ester
3 to vinyl ketone 6 on scale
Scheme 4. “Acid-free” oxidative Heck cyclization
(94.3% yield) without issue. While successful, we realized that it
would be optimal to remove the filtration of the titanium-
derived salts entirely. For the subsequent delivery, the process
was simplified through the introduction of a 20% aqueous citric
acid wash to adjust the pH to 2−3 which solubilized the solids.
This led to easily separable layers, with minimal impact on yield
(506 kg of 8 produced, 85.9−90.7% yield).
Development of Oxidative Fragmentation Process.
With a viable process to cyclopropanol 8 in hand, we began to
examine the oxidative fragmentation to produce vinyl ketone 6.
Again, the reaction performed as described in the initial
communication, and high-quality product could be isolated
without purification in >90% yield on lab scales (up to 250 g).
In practice the reaction is performed as a two-stage process.
First, cyclopropanol 8 is treated with NBS (1.0 equiv), added in
multiple portions to control the reaction exotherm; then the
intermediate β-bromoketone 9 is converted to vinyl ketone 6
between the in-process and workup yields is a direct result of
the physical properties of 2 and their impact during processing.
Significant loss (20−30%) was realized upon solvent swaps due
to the volatility of 2, and 10% was lost to the aqueous washes
due to water solubility. No attempt was made in this campaign
to recover this material as the improvement in yield brought on
by the Kulinkovich/oxidative fragmentation pathway to 6
allowed us to satisfy our delivery requirement even with this
loss. Performing the solvent swap with better vacuum control,
and recovery of the material lost to the aqueous stream would
likely increase the yield in subsequent campaigns. Running the
reaction on lab scales has led to isolation of 2 in yields ranging
from 65 to 72%. It should be noted that the modified oxidative
Heck procedure (no additional acid), produced high-quality
material after workup. The THF stream resulting from solvent
swap could be directly utilized in the methyl cuprate addition
without further purification.
B
dx.doi.org/10.1021/op500135x | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Although the acid-free oxidative Heck process offered the
potential for significant yield increase, separation of the product
from the benzoquinone/hydroquinone (BQ/HQ) by methods
other than product distillation remained a challenge. This was
hoped to be avoided, due to timeline constraints and the
volatile nature of the product leading to additional losses during
distillation. Lab-scale experiments demonstrated that BQ/HQ
could be effectively precipitated from DCM and removed by
filtration. Upon further evaluation of this protocol, reproduci-
bility issues were encountered, and in some cases >10 mol % of
BQ/HQ was present in the THF stream of 2 after workup. As
an alternative, the kilogram-scale batches utilized the non-
optimal process of washing the DCM stream of 2 with aq LiOH
(1.5 equiv with respect to HQ quantitation). Although minimal
degradation was encountered during the <2 h necessary to
perform this operation, we were aware of potential problems
imposed by the base sensitivity of enone 2. Despite the risk, the
LiOH protocol produced material (15.3 kg) of sufficiently high
quality which could be used in the next transformation without
additional purification.
To better understand the variability of the conjugate addition
process, we utilized in-process IR to monitor the consumption
of the enone on laboratory scale. A reverse addition protocol,
where the cuprate slurry was slowly added to a solution of the
enone at −20 °C, clearly suggested that the reaction was dose
controlled.8 This finding indicated that efficient mixing during
this process is essential, a requirement difficult to achieve even
on 500 g scale due to the density of the cuprate slurry. It should
be noted that, although process modifications such as the
inclusion of a stoichiometric equivalent of lithium bromide with
respect to copper could be used to produce a near-
homogeneous cuprate slurry,9 this protocol was found to be
extremely sensitive to moisture (during charging) and was not
robust. As our quantity requirements for 1 were increasing, we
knew the limitations associated with this process would hinder
throughput moving forward. We began simultaneously
evaluating the following two options: modification of the
cuprate addition protocol to arrive at a more robust and
scalable process, and an alternative strategy where an enzyme-
mediated reduction would allow for the methyl group to be
installed by methods that do not rely on conjugate addition.
Search for an Enzymatic Solution To Set Relative
Configuration of Methyl Groups. The literature is replete
with examples of metal-catalyzed hydrogenation of enone
substrates such as 13 producing the 2,6-cis-substituted product
(14).10 It was postulated that an enzymatic reduction might
override the inherent facial bias induced by the existing chiral
center, allowing us to access the trans-pyranone 1. Examples of
asymmetric enzymatic reductions of β-substituted cyclohex-
enones are known,11 but to the best of our knowledge, no
examples have been reported where the ring contains the
additional oxygen found in the pyranone structure (Scheme 6).
Methyl Cuprate Process Development. With a more
robust process to prepare 2, we next turned our attention to the
final transformation (methyl cuprate addition) which presented
significant challenges. The procedure utilized for the first
campaign is displayed in Scheme 5, and consisted of addition of
Scheme 5. Initial procedure used to produce 1, and major
byproducts that are generated in the process
Scheme 6. Proposed biocatalytic generation of 1
a solution of enone 2 to a slurry of methyl cuprate reagent. This
process presented many technical complications on scale, such
as extremely thick and difficult to stir slurries, poor mass
balance, variable yields, and the issues arising from the required
removal of 6.2 equiv of metal salts (4 equiv Li, 2.2 equiv Cu).
The major byproducts detected were the cis-diastereomer (10),
the compound resulting from ring-opening after initial cuprate
addition and subsequent addition of a second cuprate to
generate 11, and a mixture of aldol dimerization products (12)
(at least two visible by GC). The levels of the dimerization
products could be as high as 15% and thus were a major source
of yield loss. Performing the reaction at −30 °C controlled the
level of ring-opening product 11 (<5%) and the cis-
diastereomer (<2%). Due to the poor mass balance, variable
yields (45−70% before purification), and the equipment
limitations set forth by the requirement of a −30 °C reaction
temperature, this procedure was performed on a maximum
scale of 500 g. As a result, 18 batches were necessary to provide
the desired 5.0 kg of dimethyl pyranone 1.
The synthesis of the proposed enzymatic reduction substrate
15, has already been reported by Xian,12 but due to the
amounts that we would potentially require, we began evaluating
alternative syntheses.
Gouverneur has reported the gold-(I)-catalyzed oxy-Michael
cyclizations of β-hydroxy ynones,13 providing support that this
strategy could be a viable option to access enzymatic substrate
15. Although the yield for this reported transformation was low
(34%), unreacted starting material accounted for the remainder
of the material balance, so we were optimistic that we could
further improve the yields of this process with our substrate.
Addition of propynyl Grignard to TES-protected Weinreb
amide 17 (prepared from ethyl ester 16) led to cyclization
substrate 18 (Scheme 7). Initial Au(I)-catalyzed cyclization
attempts in DCM and CDCl3 provided inconsistent yields. It
was found that the presence of water (2 equiv) assisted in silyl
protecting group removal (similar to the oxidative Heck
C
dx.doi.org/10.1021/op500135x | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
only proceeded to ∼10% conversion. Taken together, these
results implicate the presence of the methyl group on the olefin
as the primary reason for the poor reactivity of 15. Initial efforts
to improve the reduction of 15 through directed evolution of
the relevant enzyme resulted in a ∼3-fold increase in
conversion without impacting the diastereoselectivity, which
could potentially be further improved with additional
optimization of the enzyme. This evolved enzyme also gave
100% reduction of 3,5-dimethyl cyclohex-2-enone (20).
Scheme 7. Synthesis of enzymatic reduction substrate 15
Process Change to a Copper-Catalyzed Grignard
Addition. On the basis of the difficulties with increasing the
conversion of the enzymatic transformation, and the delivery
timelines associated with the need of intermediate 1, we fully
invested our efforts into making the cuprate addition to enone
2 more robust. In an attempt to enhance the reproducibility
and yield of the cuprate addition for the upcoming campaigns,
we needed to address the issues encountered with the original
process; (1) difficult to stir reaction slurry, (2) cryogenic (−30
°C) conditions, (3) dimerization and ring-opening byproducts,
(4) 6 equiv of metal in the reaction, and (5) poorly
reproducible yields (45−70%). Changing the process to a
copper-catalyzed Grignard addition could, in theory, overcome
the problems encountered with the thick cuprate slurry and the
need to remove 6 equiv of metals from the reaction stream.
Inclusion of TMS-Cl in the reaction should also prevent the
aldol dimerization product as well as the byproduct from
double addition. A deprotection strategy, either as part of the
workup or as a second independent step, would have to be
included to remove the fairly labile trimethylsilyl group.
Evaluation of the copper-catalyzed Grignard addition in the
absence of TMS-Cl determined that CuCl (vs CuBr or CuI)
was the optimum Cu(I) source and could be used in levels as
low as 2%. Addition of MeMgCl (1.2 equiv) over 20−30 min to
a thin slurry of CuCl (2−5%) and enone 2 at −20 °C led to
results comparable to the stoichiometric cuprate process. Less
than 2% of the 1,2-addition product was observed; however, the
double addition product 11 and dimerization products 12 and
the cis-diastereomer 10 were still present at similar levels.
Performing the copper-catalyzed process in the presence of
TMS-Cl (1.5 equiv), led to a vast improvement with respect to
impurities 11 and 12, as trapping of the enolate as the silyl enol
ether prevented ring-opening and dimerization. It is known that
the inclusion of TMS-Cl in cuprate conjugate addition
reactions produces a more reactive cuprate reagent15 and can
also alter the diastereoselectivity16 (vs the non-TMS-Cl
process). Unfortunately, TMS-Cl had a detrimental impact
on the trans/cis ratio of our process (Scheme 9),17 which could
be offset by running the reaction at −40 °C to ensure the level
of the cis-diastereomer was below 2%. Running the reaction at
−5 °C produced the cis-diastereomer 10 in levels as high as
11%. The distillation of 1 is not an effective process for the
removal of 10, and as a result, the limit of the cis-diastereomer
was set at 2% based on the ability to remove this impurity in
the downstream chemistry. One of the goals of the redesigned
process was to allow the reaction to be run at the more
manageable temperature of −20 °C or above (vs −30 °C in the
previous campaign). The impact of TMS-Cl was counter-
productive with our desire to increase the reaction temperature,
but TMS-Cl was necessary to control the level of the other
impurities (11 and 12). We reasoned that if we could temper
the reactivity of the active copper species, it could be possible
to return to the high level of trans-selectivity previously
reaction, 6 to 2) which greatly increased the rate and yield of
this reaction. Using a solvent that is miscible with water
(acetonitrile) produced more consistent reaction rates, and
catalyst loadings as low as 0.2 mol % produced enone 15 in
95% yield. In addition to the excellent yield for this
transformation, this mode of closure has the advantage of not
requiring a co-oxidant such as benzoquinone, the removal of
which greatly complicated product isolation in the oxidative
Heck reaction to form 2.
With a route that would provide gram quantities of enone 15
in hand, we needed to evaluate the plausibility of the enzymatic
reduction before investing additional effort into developing a
scalable synthesis of this intermediate.14 After extensive
screening of commercially available ene-reductase enzymes,
the enzymatic reduction of the double bond would not proceed
further than 5% conversion (Scheme 8). Although this result
Scheme 8. Effect of β-methyl group on enzymatic reduction
was disappointing, it is noteworthy that the product that
formed was the desired 2,6-trans-dimethylpyranone (1), with
no indication of the 2,6-cis compound 10. To determine the
cause of the poor conversion, ene-reductase-catalyzed reduction
of substrate analogues was investigated. Pyrenone 2 showed up
to 80% conversion, indicating that oxygen could be tolerated
adjacent to the olefin. To the best of our knowledge, this is the
first report of ene-reductase-catalyzed reduction of a pyrenone.
No available ene-reductases were found to reduce 2,6-dimethyl-
4-pyrone, while 4-pyrone was reduced by many ene-reductase
enzymes. Also, while cyclohex-2-enone showed complete
reduction, reduction of 3,5-dimethyl cyclohex-2-enone (20)
D
dx.doi.org/10.1021/op500135x | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Scheme 9. Copper-catalyzed Grignard addition, and the effect of DPPP on diastereoselectivity
Scheme 10. Overall process to prepare dimethylpyranone 1; only distillation of the final product was required for purification
observed, and consequently allow operation at a higher reaction
temperature.
oil on lab scale (62%, 1.8% 10). Although there was little
reduction in the level of cis-diastereomer 10 during distillation,
the level in the isolated product was <2.0% and was adequate
for the project needs. Although the program that utilized this
intermediate was terminated prior to demonstration of this new
process on kilogram scale, our development experience and lab
results suggested this would have been a scalable process
capable of delivering kilogram quantities of 1 in good yield and
quality.
Chiral phosphine ligands have been extensively studied for
the enantioselective conjugate addition of carbon nucleophiles
to enones.18 Guided by an account by Feringa19 which reports
the copper-catalyzed asymmetric conjugate addition of
Grignard reagents to cyclic enones (>95% ee), we wondered
if we could use a bulky phosphine ligand to return the level of
facial selectivity we observed prior to TMS-Cl inclusion. We
assumed that the use of more exotic/expensive chiral
phosphine ligands would not be required, as the stereocenter
already present in the molecule should direct the addition.
Inclusion of the readily available bidentate ligand 1,3-
bis(diphenylphosphino)propane (DPPP) (7 mol % for 5 mol
% CuCl), had a significant impact on the diastereoselectivity of
the addition (Scheme 9). The reaction could now be run at
temperatures as high as −15 °C while maintaining cis-
diastereomer 10 below the required limit (<2%). Remarkably,
reactions conducted at 4 °C in the presence of DPPP produced
only 2.7% of 10. Removal of the DPPP from the product was
facilitated by its low solubility in MTBE, the workup solvent for
the reaction. The majority of the ligand was removed by
filtration prior to TMS−enol ether hydrolysis. Catalyst loadings
as low as 1 mol % CuCl and 2 mol % DPPP produced 1.9% of
the cis compound at −20 °C. The more economical triphenyl
phosphine (0.15 equiv for 0.05 equiv of CuCl) was also
examined. While the diastereomer 10 was formed in an
acceptable level of 1.9% at −20 °C, the reaction profile was not
as clean, and ligand removal proved to be more difficult. These
challenges led us to proceed with the DPPP reaction
conditions.
CONCLUSION
■
The Ti(OiPr)4 Kulinkovich cyclopropanation/oxidative frag-
mentation strategy to 6 increased the overall yield to this
intermediate from 27%−48% to 82%. This alternative route
offered the additional benefits wherein purification, cryogenic
conditions, and vinyl magnesium bromide were no longer
required. The oxidative Heck conversion of 6 to 2 was made
more robust by developing an acid-free cyclization process.
Although the demonstrated yield was only slightly improved
(47 vs 40%), the in-process yield was high (81%), and areas for
yield improvements were identified: (1) more controlled
concentrations (20−30% loss during solvent distillation) and
(2) additional back extractions to minimize loss to the aqueous
stream (10%). A copper-catalyzed Grignard conjugate addition
in the presence of trimethylsilyl chloride was demonstrated to
overcome the issues that were encountered with the previous
standard cuprate addition process. To obtain a sufficient level of
diastereoselectivity during the conjugate addition it was
necessary to increase the steric bulk of the active species via
the addition of DPPP. This led to the cis/trans selectivity being
less susceptible to temperature effects and allowed us to
effectively control the level of the undesired diastereomer and
impurities to produce high quality trans-dimethylpyranone (see
Scheme 10 for overall process).
Hydrolysis of the intermediate TMS-enol ether (22/23) was
straightforward. Washing the organic workup stream with 1 M
aqueous citric acid led to rapid conversion to 1/10. Purification
via careful distillation yielded dimethylpyranone 1 as a colorless
E
dx.doi.org/10.1021/op500135x | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
1.3 Hz, 1 H), 1.63 (ddd, J = 14.4, 3.8, 1.2 Hz, 1 H), 1.26 (d, J =
6.3 Hz, 3 H), 0.92 (s, 9 H), 0.84−0.78 (m, 1 H), 0.72−0.66 (m,
1 H), 0.47 (ddd, J = 10.8, 6.3, 5.2 Hz, 1 H), 0.38 (ddd, J = 10.8,
6.0, 4.6 Hz, 1 H), 0.14 (s, 3 H), 0.13 (s, 3 H) ppm; 13C NMR
(CDCl3, 100 MHz) 69.8, 54.9, 45.3, 25.7, 23.7, 17.8, 13.6, 11.6,
EXPERIMENTAL SECTION
■
General. Reactions were monitored by thin layer
chromatography (TLC) carried out on 0.25 mm E. Merck
silica plates (60F-254), using basic potassium permanganate
stain and heat as the visualizing agent. For reactions monitored
by GC, a Hewlett-Packard HP6890 with an auto injector was
utilized (ZB-1701 Column, 30 m × 0.32 mm × 1.0 μm; carrier
= helium at 2.0 mL/min; oven temp 65 °C for 1 min, 5 °C/min
to 220 °C, hold 1 min, 33 min run time; inlet (front) 200 °C,
split 20:1, gas saver 15 mL/min at 2 min; detection FID (rear)
280 °C, 30 mL/min hydrogen + 300 mL/min air). Reported
yields are for isolated materials or calculated solution yields and
are corrected for potency.
−4.3, −4.9 ppm; [α]25 = −33.4 (c = 1.0, CHCl3); HRMS
D
calcd for [M + H] C12H27O2Si = 231.1775, found = 231.1781.
Alternative Procedure for 8 That Avoids Filtration of
the Mg/Ti Salts. A THF solution (298.24 kg of 51.3 wt %) of
silylated 3 (153 kg, 623.8 mol, 1.0 equiv) and THF (1200 L)
were placed in a reactor and the contents cooled to between −5
and 0 °C. Titanium(IV) isopropoxide (86.87 kg, 305.6 mol,
0.49 equiv) was added, and the reactor contents were stirred for
10−20 min. Ethyl magnesium bromide (2.0 mol/L in 2-
MeTHF, 873.3 L, 1747 mol, 2.8 equiv) was added over ∼1 h,
keeping the temperature between −5 and 10 °C. The reaction
was then stirred at 0 to 5 °C until GC showed less than 1% of
silylated 3 with respect to 8 (2 h). NH4Cl (25 wt % aqueous,
900.1 kg) was added at a rate that maintained the temperature
between −5 and 10 °C (pH 8−9 after addition) and minimized
initial off-gassing. 20 wt % aqueous citric acid (1002 kg) was
added at a rate to maintain the temperature between −5 and 10
°C (pH 2−3 after addition), and the layers were mixed for 20
min. MTBE (291.5 kg) was added, and the reactor contents
were warmed to 10−15 °C. After mixing for 30 min, the layers
were separated. The aqueous was extracted with MTBE (590.1
kg) and the combined organics were washed with 7 wt %
aqueous NaHCO3 (610 kg, pH 7−8 after addition), and brine
(2 × 650 kg). The organic layer was then concentrated to ∼600
L under reduced pressure (temp = 25 °C). DCM (1501 kg)
was added, and the volume was reduced to 750 L under
reduced pressure (25 °C). This distillation was repeated two
more times, after which DCM (2800 kg) was added. The
stream was dried with anhydrous MgSO4 (22 kg), and after
removal of the solid by filtration, the solution was concentrated
to yield 886.8 kg of a 14.6 wt % solution of 8 (129.5 kg, 561.9
mol, 90.1% yield) in DCM. This stream could be used without
purification in the subsequent step.
(R)-5-((tert-Butyldimethylsilyl)oxy)hex-1-en-3-one (6).
DCM (983.7 kg) and a DCM stream (178.9 kg) containing
41.2 wt % cyclopropanol 8 (73.7 kg, 319.93 mol, 1.0 equiv)
were placed in a reactor. The solution was cooled to between
−5 and 5 °C, and N-bromosuccinimide (56.7 kg, 318.6 mol 1.0
equiv) was added in 5−10 kg portions in 15 min intervals (each
addition is slightly exothermic). After the final addition, the
reaction was stirred at −5 to 5 °C until complete by GC (<1%
of 8 vs 9), 1 h. Triethylamine (64.8 kg, 640.4 mol, 2.0 equiv)
was added at a rate of 40−60 kg/h to maintain the temperature
between −5 and 5 °C. The reaction was stirred at this
temperature for 2 h, and was complete by GC (<1.0% 9 vs 6).
The 0 °C reaction was washed with 16.3 wt % aqueous citric
acid (275.2 and 317.6 kg), and 7.9 wt % aqueous NaHCO3
(217.1 kg). Hydroquinone (7.4 g, 0.067 mol, 0.0002 equiv) was
added to increase product stability during solvent exchange.
The solvent was removed under reduced pressure until 250−
300 L remained. The concentrated stream was filtered through
a nutsche filter preloaded with silica gel (16.0 kg), and the cake
was washed with DCM (49.4 kg × 2). The filtrate and rinses
were combined, and hydroquinone was added (7.4 g, 0.067
mol, 0.0002 equiv). The stream was concentrated under
atmospheric pressure until 90−150 L remained, and then under
reduced pressure (∼100 Torr, <50 °C) for 2−3 h. This yielded
a stream (82.2 kg, 82.4 wt %) of 6 (67.73 kg, 296.7 mol, 92.3%)
(R)-1-(2-((tert-Butyldimethylsilyl)oxy)propyl)-
cyclopropanol (8). DCM (164.4 kg) was added to a reactor,
followed by methyl (R)-3-hydroxybutyrate (3) (43.1 kg, 364.8
mol, 1.0 equiv) and imidazole (29.7 kg, 436.3 mol, 1.2 equiv).
Once the solid dissolved, the solution was cooled to 0−10 °C,
and a solution of TBS-Cl (60.2 kg, 399.4 mol, 1.09 equiv) in
DCM (65.0 kg) was added at a rate of 60−90 kg/h. Once the
addition was complete, the reactor contents were warmed to
20−25 °C over 90 min and stirred for 5 h. The reaction was
complete by GC (<3% SM vs silylated 3), and water (215.1 kg)
was added. After mixing for 30 min, the layers were separated,
and the aqueous was extracted with DCM (200.5 kg). The
combined organics were washed with water (129 kg), and the
DCM solution of silylated 3 was dried with activated molecular
sieves (20.2 kg) for 3 h to a KF < 0.05%. The molecular sieves
were removed by filtration, and the cake was washed with
DCM (2 × 50 kg). The filtrate and rinses were combined. The
organic stream was concentrated under reduced pressure (<50
°C) until 100−150 L remained. THF (221.0 kg) was added,
and the mixture was concentrated under reduced pressure until
<0.5% DCM remained. THF (430 kg) was added, leading to a
16.6 wt % solution of silylated 3 in THF. The mixture was
cooled to 16 °C, and titanium(IV) isopropoxide (41.3 kg, 145.0
mol, 0.40 equiv) was added. After cooling to 0 °C, a THF
solution (889.1 kg) of 13.1 wt % ethyl magnesium bromide
(116.5 kg, 874.2 mol, 2.40 equiv) was added at a rate of 60−
150 kg/h (11 h total). After stirring an additional 2 h, GC
indicated reaction completion (<1% of silylated 3 vs 8). The
reaction was cooled to 0 °C, and a 20 wt % aqueous NH4Cl
solution (300.0 kg) was added slowly as an initial exotherm and
gas evolution was observed. After this portion was added, an
additional 20 wt % aqueous NH4Cl (544.4 kg) was added at a
rate (200−300 kg/h) to maintain the temperature below 20 °C
(2 h 20 min). The resulting slurry was stirred for 30 min and
then filtered with a centrifuge. The cake was washed with
MTBE (161.6 + 160.8 kg), with the cake being soaked for 1 h
before solvent removal. The filtrate and the rinses were
combined, stirred for 30 min, held for 30 min, and then
separated. The aqueous phase was extracted with MTBE (96.2
kg) and this wash combined with the other organic stream.
This produced 1524.4 kg of a 5.2 wt % solution of
cyclopropanol 8 (79.3 kg, 344.1 mol, 94.3%). The stream was
then solvent swapped to DCM (<0.5% MTBE) for use in the
next step, as subsequent purification was not required. A small
portion of this material was isolated via flash column
chromatography on silica gel (5% ethyl acetate in hexanes as
eluent) for characterization purposes. Colorless oil, Rf = 0.40
1
(10% ethyl acetate in hexanes); H NMR (CDCl3, 400 MHz)
4.29−4.21 (m, 1 H), 3.93 (br. s, 1 H), 1.82 (ddd, J = 14.4, 7.4,
F
dx.doi.org/10.1021/op500135x | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
that could be used in the next step without further purification.
Caution: Vinyl ketone 6 was identified as a skin sensitizer in a
LLNA assay (EC3 <1%). A small portion of this material was
isolated via distillation for characterization purposes. Colorless
isolated via distillation for characterization purposes. Colorless
oil, BP = 63−65 °C at 15 mmHg; H NMR (CDCl3, 400
1
MHz) 7.33 (d, J = 6.0 Hz, 1 H), 5.38 (dd, J = 6.0, 1.0 Hz, 1 H),
4.59−4.49 (m, 1 H), 2.45 (dd, J = 16.6, 12.6 Hz, 1 H), 2.42
(ddd, J = 16.6, 4.6, 1.0 Hz, 1 H), 1.44 (d, J = 6.6 Hz, 3 H) ppm;
13C NMR (CDCl3, 100 MHz) 192.4, 163.1, 106.6, 75.8, 43.2,
20.1 ppm; [α]25D = +205.7 (c = 1.0, CHCl3); HRMS calcd for
[M + H] C6H9O2 = 113.0597, found =113.0602.
1
oil, BP = 89−91 °C at 4−5 mmHg; H NMR (CDCl3, 400
MHz) 6.36 (dd, J = 17.5, 10.5 Hz, 1 H), 6.22 (d, J = 17.5 Hz, 1
H), 5.85 (d, J = 10.5 Hz, 1 H), 4.38−4.30 (ddd, J = 7.2, 6.0, 5.2
Hz, 1 H), 2.85 (dd, J = 14.8, 7.2 Hz, 1 H), 2.54 (dd, J = 14.8,
5.2 Hz, 1 H), 1.20 (d, J = 6.0 Hz, 3 H), 0.85 (s, 9 H), 0.06 (s, 3
H), 0.02 (s, 3 H) ppm; 13C NMR (CDCl3, 100 MHz) 199.6,
137.5, 128.5, 65.8, 49.1, 25.8, 24.2, 18.0, −4.6, −5.0 ppm;
(2R,6R)-2,6-Dimethyldihydro-2H-pyran-4(3H)-one (1)
via Stoichiometric Cuprate on 500 g Scale. Copper(I)
iodide (1.87 kg, 9.81 mol, 2.2 equiv) was slurried in dry
isopropyl ether (5.0 L) and cooled to 0 °C. A solution of
methyl lithium (3.0 M in diethoxymethane, 5.94 L, 17.83 mol,
4.0 equiv) was added to the CuI slurry, and the mixture was
stirred for 30 min at 0 °C. The bright-yellow slurry was cooled
to −20 to −30 °C, and crude 2 (500 g, 4.46 mol, 1.0 equiv) in
isopropyl ether (7.5 L) was added over 10 min. The reaction
was then stirred at −20 to −30 °C for 30 min, and then added
to 1.0 M aqueous HCl (3 L) that was previously cooled to 0
°C. After stirring for 20 min, the gray suspension was filtered,
and the filtrate was extracted with MTBE (2 × 3 L). The
combined organics were dried with MgSO4, filtered, and then
concentrated to yield crude 1 (394.4 g, 3.08 mol, 69% yield).
This material could be further purified by distillation. For
distillation conditions and characterization data, see copper-
catalyzed Grignard addition procedure that follows. As
mentioned in the text, this procedure produced variable yields
before distillation, ranging from 45 to 70%.
[α]25 = −33.3 (c = 1.0, CHCl3); HRMS calcd for [M + H]
D
C12H25O2Si = 229.1618, found = 229.1623.
(R)-2-Methyl-2H-pyran-4(3H)-one (2). Acetone (213.6
kg), purified water (26.0 kg, 1443 mol, 4.8 equiv),
benzoquinone (34.4 kg, 318.3 mol, 1.07 equiv), and bis-
(acetonitrile)dichloropalladium (0.99 kg, 3.82 mol, 0.013
equiv) were charged to a reactor. The reactor contents were
heated to 45−50 °C, and compound 6 (80 kg of 82.5 wt %
from above, 68.0 kg, 297.7 mol, 1.0 equiv) was added at a
constant rate over 2 h. The reaction was held at this
temperature until complete (<1.5% of 6 vs 2 by GC), typically
3−4 h. The reaction was cooled to 20 °C, and the mixture was
concentrated under reduced pressure (>160 Torr, 14−28 °C)
until 120−150 L remained. Water (266 kg) was added, the
reactor contents were cooled to 0−10 °C, and stirred at this
temperature for 2 h. The resulting slurry was filtered and the
cake washed with water (3 × 67 kg). The filtrate and all rinses
were combined. n-Hexane (44.0 kg) was added to the
combined filtrates, and the mixture was stirred at 15−25 °C
for 40 min. The layers were separated, and the aqueous was
washed further with n-hexane (44.0 + 43.9 kg). The combined
n-hexane washes were washed with water (2 × 33.3 kg), and
aqueous fractions were combined. Dichloromethane (267.1 kg)
and sodium chloride (96.2 kg) were added to the aqueous, and
the layers were mixed for 40 min, before being allowed to stand
for 40 min. The layers were separated, and the aqueous was
extracted with DCM (400.2 kg). Analysis of the organic layer
determined the level of hydroquinone (26.2 kg, 237.9 mol)
present. The organic was washed with a solution of lithium
hydroxide (14.9 kg, 355.1 mol, 1.5 equiv vs hydroquinone) in
water (297.9 kg) for 15 min, before separating the layers. The
organic was washed with a solution of sodium dihydrogen
phosphate (0.9 kg) and disodium hydrogen phosphate (2.5 kg)
in water (50 kg) to adjust the pH. The aqueous layer from this
wash was back extracted with DCM (198.9 kg), and this
organic wash was combined with the previous. The organic
phase was dried with 4 Å molecular sieve powder (10.0 kg) for
3−4 h, after which an additional 5 kg of molecular sieves was
added and stirring continued for 1−2 h (this was repeated one
additional time). At this point the KF of the solution was
<0.05%. The molecular sieves were removed by filtration. The
cake was washed with THF (2 × 66.5 kg), and these washes
were initially kept separate. The DCM filtrate was concentrated
at atmospheric pressure (<50 °C) until 50−90 L remained, and
then the pressure was reduced to 300 Torr until 30−60 L
remained. The THF from the cake washes was added, and the
distillation continued to a volume of 30 L. THF (99.8 kg) was
added, and the concentration continued until the level of DCM
and acetone were <0.5 vol %. This resulted in a 95.4 kg stream
containing 16.0 wt % of 2 (15.4 kg, 137.4 mol, 46% yield). This
material was of sufficient purity to be used in the next step
without further purification. A small portion of this material was
Ethyl (R)-3-((Triethylsilyl)oxy)butanoate Silylated 16.
Ethyl (R)-(−)-3-hydroxybutyrate (16) (70.0 g, 0.520 mol, 1.0
equiv) was dissolved in dichloromethane (840 mL) and cooled
to 0 °C. Imidazole (70.0 g, 1.03 mol, 2.0 equiv) was added and
stirred until completely dissolved. Chlorotriethylsilane (93.7
mL, 0.556 mol, 1.05 equiv) was slowly added to the mixture
which was then allowed to warm to 20 °C and stirred for 18 h.
Water (700 mL) was added, and the phases were separated.
The aqueous layer was washed with dichloromethane (300
mL). The combined organics were then washed with water
(200 mL) and brine (200 mL), dried over MgSO4, and
concentrated under reduced pressure to afford ethyl (R)-3-
((triethylsilyl)oxy)butanoate (silylated 16, 123.0 g, 499.2
mmol, 96% yield) as a clear, colorless oil. This material could
be used in the subsequent step without additional purification.
1
Rf = 0.62 (20% ethyl acetate in hexanes); H NMR (CDCl3,
500 MHz) 4.33−4.23 (m, 1 H), 4.17−4.06 (m, 2 H), 2.48 (dd,
J = 14.6, 7.3 Hz, 1 H), 2.36 (dd, J = 14.6, 5.6 Hz, 1H), 1.25 (t, J
= 7.2 Hz, 3 H), 1.21 (d, J = 6.1 Hz, 3 H), 0.94 (t, J = 8.0 Hz, 9
H), 0.59 (q, J = 7.9 Hz, 6 H) ppm; 13C NMR (CDCl3, 126
MHz) 171.6, 65.6, 60.2, 44.9, 23.9, 14.2, 6.7, 4.8 ppm; [α]25
=
D
−21.3 (c = 1.38, CHCl3); HRMS calcd for [M + H]
C12H27O3Si = 247.1724, found = 247.1734.
(R)-N-Methoxy-N-methyl-3-((triethylsilyl)oxy)-
butanamide (17). Silylated 16 (60.0 g, 244 mmol, 1.0 equiv)
was dissolved in tetrahydrofuran (0.60 L). N,O-dimethylhy-
droxylamine hydrochloride (36.8 g, 337 mmol, 1.55 equiv), was
added, and after complete dissolution the solution was cooled
to −31 °C. To this mixture was added isopropylmagnesium
chloride (2.0 mol/L in THF, 360 mL, 720 mmol, 2.96 equiv)
over 25 min to maintain the temperature below −15 °C. Once
addition was complete, the reaction was maintained at −20 °C
for 2 h followed by quenching with saturated aqueous NH4Cl
(400 mL). The solution was allowed to warm to 20 °C and
stirred overnight. The phases were separated, and the aqueous
G
dx.doi.org/10.1021/op500135x | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
layer was extracted with MTBE (2 × 100 mL). The combined
organics were dried over MgSO4. Removal of solvent under
reduced pressure afforded Weinreb amide 17 (60.0 g, 230
mmol, 94% yield) as a light-brown oil, which could be used
directly in the subsequent step. A small portion of this material
was purified by flash column chromatography on silica gel. Rf =
for 2 min, 15 °C/min to 200 °C, hold for 2 min (run time = 33
min). Detection was by FID, 270 °C 30:300 hydrogen:air mL/
min. Retention times with the enantiomer separation oven
program were 18.3 min for 1, 19.2 min for ent-1, 11.8 min for
the cis isomer 10.
(2R,6R)-2,6-Dimethyldihydro-2H-pyran-4(3H)-one (1)
via Copper-Catalyzed Grignard Addition. 1,3-Bis-
(diphenylphosphino)propane (DPPP, 1.96 g, 4.61 mmol, 0.07
equiv), copper(I) chloride (0.33 g, 3.3 mmol, 0.05 equiv), and
THF (110 mL) were added to a reactor. After stirring for 30
min, the inorganics had almost completely dissolved.
Compound 2 (7.39 g, 65.9 mmol, 1.0 equiv) dissolved in
THF (37 mL) was added to the Cu/ligand slurry. After stirring
for 1 h, the reaction was cooled to between −15 and −20 °C,
and chlorotrimethylsilane (12.9 mL, 99.0 mmol, 1.5 equiv) was
added over 5 min (slight exotherm of 2 °C). Methylmagnesium
chloride (26.4 mL of 3.0 M in THF, 79.2 mmol, 1.20 equiv)
was added over 1 h. Once the addition was complete, the
reaction was stirred for 10 min, and then transferred over 5 min
to a 5 °C mixture of MTBE (100 mL) and saturated aqueous
NaHCO3 (100 mL) with stirring. The mixture was allowed to
warm to 23 °C, and then stirred for 10 min. The layers were
separated, and the aqueous was washed with MTBE (25 mL).
The combined organics were washed with saturated aqueous
NH4Cl (50 mL) and then brine (50 mL). At this point there
was a significant amount of DPPP ligand precipitate. The
mixture was then stirred overnight to precipitate more ligand
and then filtered. The solution was concentrated to ∼80 mL
(25 °C, 160 Torr). Quantitation at this stage showed 9.49 g of
silylenol ether 22 (72%) and 0.675 g of compound 1, for a
combined solution yield of 80%. This stream was added to 1 M
aqueous citric acid (100 mL) and mixed for 1 h. The layers
were separated, saturated aqueous brine (5 mL) was added to
the aqueous layer, and it was back extracted with MTBE (2 ×
25 mL). The combined organics were washed with saturated
aqueous NaHCO3 (2 × 25 mL) and brine (25 mL). The
MTBE product stream was passed through a short plug of silica
(washing with MTBE). The solvent was removed under
reduced pressure (175 Torr), and then distillation (40 Torr, 89
°C) yielded compound 1 (5.12 g, 40.57 mmol, 62% yield) as a
colorless oil. The level of cis-diastereomer 10 was 1.8%
(identical to that before distillation). Colorless oil, bp = 89
°C at 40 mmHg; Rt (GC) = 11.83 min, (10.52 min for cis-
compound 10); 1H NMR (CDCl3, 400 MHz) 4.36−4.29 (m, 2
H), 2.55 (ddd, J = 14.2, 4.8, 1.5 Hz, 2 H), 2.24 (ddd, J = 14.2,
6.6, 1.5 Hz, 2 H), 1.27 (d, J = 6.6 Hz, 6 H) ppm; 13C NMR
(CDCl3, 100 MHz) 207.6, 68.2, 48.3, 20.6 ppm; [α]25D = +24.7
(c = 1.0, CHCl3); HRMS calcd for [M + H] C7H13O2 =
129.0910, found = 129.0914.
1
0.28 (20% ethyl acetate in hexanes); H NMR (CDCl3, 500
MHz) 4.41−4.32 (m, 1 H), 3.70 (s, 3 H), 3.18 (s, 3 H), 2.78
(dd, J = 14.2, 6.3 Hz, 1 H), 2.40 (dd, J = 14.8, 5.7 Hz, 1 H),
1.24 (d, J = 6.0 Hz, 3 H), 0.95 (t, J = 8.0 Hz, 9 H), 0.61 (q, J =
8.1 Hz, 6 H) ppm; 13C NMR (CDCl3, 126 MHz) 172.2, 65.6,
61.1, 41.7, 31.8, 24.1, 6.6, 4.6 ppm; [α]25 = −14.3 (c = 0.86,
D
CHCl3); HRMS calcd for [M + H] C12H28O3NSi = 262.1833,
found = 262.1839.
(R)-6-((Triethylsilyl)oxy)hept-2-yn-4-one (18). Weinreb
amide 17 (20.0 g, 76.5 mmol, 1.0 equiv) was dissolved in
MTBE (200 mL) and cooled to −40 °C. 1-Propynylmagnesium
bromide (0.5 mol/L in THF, 306 mL, 153 mmol, 2.0 equiv)
was added slowly to maintain the temperature below −10 °C.
The reaction mixture was then allowed to warm to −2 °C over
2.5 h and then was quenched with saturated aqueous NH4Cl
(300 mL). The phases were separated, and the aqueous layer
was extracted with MTBE (3 × 100 mL). The combined
organics were dried over MgSO4, and solvent was removed
under reduced pressure. The crude oil was then purified via
silica gel column chromatography (300 g column, 0−20%
EtOAc in hexanes gradient) to afford ynone 18 (18.4 g, 68.2
mmol, 89.2% yield) as a pale-yellow oil. Rf = 0.52 (20% ethyl
1
acetate in hexanes); H NMR (CDCl3, 500 MHz) 4.44−4.37
(m, 1 H), 2.77 (dd, J = 15.0, 7.1 Hz, 1 H), 2.58 (dd, J = 15.1,
5.7 Hz, 1 H), 2.03 (s, 3H), 1.22 (d, J = 6.3 Hz, 3 H), 0.96 (t, J =
7.9 Hz, 9 H), 0.61 (q, J = 7.9 Hz, 6 H) ppm; 13C NMR
(CDCl3, 126 MHz) 186.1, 90.1, 80.6, 65.0, 55.2, 23.9, 6.7, 4.7,
3.9 ppm; [α]25 = −19.5 (c = 1.50, CHCl3); HRMS calcd for
D
[M + H] C13H25O2Si = 241.1618, found = 241.1624.
(R)-2,6-Dimethyl-2H-pyran-4(3H)-one (15). Ynone 18
(5.0 g, 20.8 mmol, 1 equiv) was dissolved in MeCN (50 mL),
then AuCl (51 mg, 0.21 mmol, 0.01 equiv) and water (0.75 mL,
41.6 mmol, 2.0 equiv) were added. The reaction mixture was
stirred at 20 °C for 18 h and then transferred to a separatory
funnel and washed with hexanes (2 × 25 mL). The MeCN
layer was concentrated under reduced pressure. The resulting
oil contained solids and was diluted with a minimal amount of
dichloromethane, filtered through a plug of silica gel eluting
with dichloromethane, and concentrated to afford cyclic enone
15 (2.62 g, 20.0 mmol, 96% yield) as a light-yellow oil. This
compound has been reported previously (Bianchi, G.; Tava, A.
Agric. Biol. Chem. 1987, 51, 2001), and all spectroscopic data
was in agreement.
Procedure for Ene-Reductase-Mediated Reduction.
Ene-reductase (5 mg), glucose (12 mg), NADP (1 mg),
glucose dehydrogenase (0.1 mg, 7.65 U from Amano), and
enone (2 μL) in 1 mL 0.1 M potassium phosphate buffer pH 7
were incubated in microfuge tubes at 30 °C. Samples (0.5 mL)
were extracted with 1 mL ethyl acetate and centrifuged for 5
min, and the extract (1 μL injection) was analyzed by GC. The
chiral column was a BGB-174 30 m × 0.25 mm × 0.25 μm,
carrier was helium 1.4 mL/min, inlet 220 °C, split 20:1. For
screening for reductions the oven program was 40 °C for 2 min,
15 °C/min to 200 °C, hold for 2 min (run time = 14.67 min),
and the retention times for 10 and 1 were 11.3 and 9.5 min,
respectively. The oven program for separation of product
enantiomers was 85 °C for 2 min, 0.5 °C/min to 96 °C, hold
ASSOCIATED CONTENT
■
S
* Supporting Information
1H- and 13C NMR for compounds 1, 2, 6, 8, 17, 18, and
silylated 16 (listed as SI-2), as well as the in situ IR cuprate
addition studies. This material is available free of charge via the
Internet at http://pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Authors
*E-mail: ian.young@bms.com (I.S.Y.).
*E-mail: steven.tymonko@bms.com (S.A.T.).
H
dx.doi.org/10.1021/op500135x | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We thank Drs. David Kronenthal, Albert DelMonte, Rajendra
Deshpande, Susanne Kiau, and David Conlon for careful review
of this manuscript. Drs. Brent Karcher (GC development) and
David Ayers (HRMS) are acknowledged for contributions to
this research.
REFERENCES
(1) Reddy, D. S.; Velde, D. V.; Aube,
(2) Anderson, K. R.; Atkinson, S. L. G.; Fujiwara, T.; Giles, M. E.;
Matsumoto, T.; Merifield, E.; Singleton, J. T.; Saito, T.; Sotoguchi, T.;
Tornos, J. A.; Way, E. L. Org. Process Res. Dev. 2010, 14, 58.
(3) Baktharaman, S.; Selvakumar, S.; Singh, V. K. Tetrahedron Lett.
2005, 46, 7527.
■
́
J. J. Org. Chem. 2004, 69, 1716.
(4) For the initial lab-scale experiments, the ethyl ester (16) of 3 was
used. The methyl ester was easier to source on scale.
(5) As low as 0.1 equiv of Ti(OiPr)4 has been demonstrated to be
effective on lab scale, but for the scale-up campaign we utilized 0.40−
0.49 equiv to ensure minimal ethyl Grignard addition to the ester.
(6) Lipshutz, B. H.; Pollart, D.; Monforte, J.; Kotsuki, H. Tetrahedron
Lett. 1985, 26, 705.
(7) Wilson, N. S.; Keay, B. A. J. Org. Chem. 1996, 61, 2918.
(8) See Supporting Information for in-process IR studies.
(9) Lipshutz, B. H.; Sengupta, S. Organocopper Reagents:
Substitution, Conjugate Addition, Carbo/Metallocupration, and
Other Reactions. In Organic Reactions; Paquette, L., et al., Eds.; John
Wiley and Sons, Inc: New York, 1992; Vol. 41, pp 135−631.
(10) (a) Fuwa, H.; Mizunuma, K.; Matsukida, S.; Sasaki, M.
Tetrahedron 2011, 67, 4995. (b) Wipf, P.; Reeves, J. T. Chem.
Commun. 2002, 18, 2066. (c) Sato, M.; Kuroda, H.; Kaneko, C.;
Furuya, T. J. Chem. Soc., Chem. Commun. 1994, 6, 687.
(11) (a) Swiderska, M. A.; Stewart, J. D. J. Mol. Catal. B: Enzym.
2006, 42, 52. (b) Adalbjornsson, B. V.; Toogood, H. S.; Fryszkowska,
̈
A.; Pudney, C. R.; Jowitt, T. A.; Leys, D.; Scrutton, N. S.
ChemBioChem. 2010, 11, 197. (c) Yanto, Y.; Winkler, C. K.; Lohr,
S.; Hall, M.; Faber, K.; Bommarius, A. S. Org. Lett. 2011, 13, 2540.
(12) Wang, H.; Shuhler, B. J.; Xian, M. J. Org. Chem. 2007, 72, 4280.
(13) Schuler, M.; Silva, F.; Bobbio, C.; Tessier, A.; Gouverneur, V.
Angew. Chem., Int. Ed. 2008, 47, 7927.
(14) As anticipated, we were unable to override the inherent
substrate influence in the transition metal catalyzed reduction of 15 to
1. Extensive screening produced at best a 1:1 trans:cis mixture.
(15) (a) Lipshutz, B. H.; Dimock, S. H.; James, B. J. Am. Chem. Soc.
1993, 115, 9283. and references within (b) Bertz, S. H.; Chopra, A.;
Eriksson, M.; Ogle, C. A.; Seagle, P. Chem.Eur. J. 1999, 5, 2680 and
references within..
(16) (a) Horiguchi, Y.; Komatsu, M.; Kuwajima, I. Tetrahedron Lett.
1989, 30, 7087. (b) Jameleddine, K.; Yakdhan, K.; Jamil, K.; Bechir, B.
H.; Denis, G. Syn. Comm. 2002, 32, 2719.
(17) When the diastereomer level for 10 is presented, it was
determined after hydrolysis of the silylenol ether (22/23).
(18) (a) Sibi, M. P.; Manyem, S. Tetrahedron 2000, 56, 8033.
̀ ́
(b) Alexakis, A.; Backvall, J. E.; Krause, N.; Pamies, O.; Dieguez, M.
̈
Chem. Rev. 2008, 108, 2796. (c) Harutyunyan, S. R.; den Hartog, T.;
Geurts, K.; Minnaard, A. J.; Feringa, B. L. Chem. Rev. 2008, 108, 2824.
(19) Feringa, B. L.; Badorrey, R.; Pena, D.; Harutyunyan, S. R.;
̃
Minnaard, A. J. Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 5834.
I
dx.doi.org/10.1021/op500135x | Org. Process Res. Dev. XXXX, XXX, XXX−XXX