Improved large-scale synthesis of(R)-benzyl 4-hydroxyl-2-pentynoate from (R)-3-butyn-2-ol

Article abstract of DOI:10.1021/op010235a

A reliable one-pot process for the title compound has been developed for large-scale productions. The effect of reaction conditions such as temperature, order of addition, and amount of the lithiation reagent has been extensively investigated, giving rise to an optimal process with highest attainable yield.

Full text of DOI:10.1021/op010235a

Organic Process Research & Development 2002, 6, 308−310
Improved Large-Scale Synthesis of(R)-Benzyl 4-Hydroxyl-2-pentynoate from
(R)-3-Butyn-2-ol
Xiaoyong Fu, Jianguo Yin,* T. K. Thiruvengadam, Timothy L. McAllister, Chou-Hong Tann, and Cesar Colon
Synthetic Chemistry, Schering-Plough Research Institute, 1011 Morris AVenue, Union, New Jersey 07083, U.S.A.
Abstract:
Scheme 1
A reliable one-pot process for the title compound has been
developed for large-scale productions. The effect of reaction
conditions such as temperature, order of addition, and amount
of the lithiation reagent has been extensively investigated, giving
rise to an optimal process with highest attainable yield.
As a part of our drug development effort in synthesizing
himbacine analogues, the potential therapeutical agents for
Alzheimer’s disease,1 we needed to identify an efficient way
to produce (R)-benzyl 4-hydroxyl-2-pentynoate 1 in multi-
kilogram scale. A one-pot, four-step procedure from (R)-3-
,2
3
butyn-2-ol as shown in Scheme 1 was reported. This
procedure utilized trimethylsilyl (TMS) as the hydroxyl
protection group that could be removed during aqueous work
up. Lithiation of the TMS-protected butynol 3 in the presence
of hexamethyldisilazane (HMDS) with n-butyllithium (n-
BuLi), followed by the addition of benzyl chloroformate at
Scheme 2
-
25 to -30 °C. Subsequent hydrolysis provided the desired
4
product 1 in moderate yields after acidic workup. However,
this procedure was found to be not reproducible even on
relatively small scale. Furthermore, an attempt to scale-up
the process in the plant failed to produce any useful material,
but rather a complex mixture containing only about 20% of
the desired product 1. This situation prompted us to undertake
a more extensive study on the reaction. In this contribution
we present a detailed study that resulted in a process with
excellent yield on multikilogram-scale productions.
pathway. When the reaction mixture 6 was treated with
benzyl chloroformate, bis-carbobenzylation product 7 was
detected as the major impurity in the final product. To
determine the stability of 4, a solution of 4 was generated at
5
-
30 °C and aged at different temperatures for 5 or 6 h before
benzyl chloroformate treatment. As shown in entry 1 of Table
, compound 4 showed very high stability at -25 °C. Upon
1
the treatment of the aged solution with benzyl chloroformate,
6
only a trace amount of impurity 7 was observed in the final
Result and Discussion
product. In contrast, experiments showed that decomposition
of 4 became increasingly noticeable at elevated temperatures.
A 5% and a significant 45% of the impurity was observed
at -15 and 0 °C, respectively. These results suggested that
4 is stable in solution only at a temperature below -25 °C
and therefore the lithiation must be done at below this critical
temperature to avoid any decompositions.
Effect of Order of Addition on Reaction Yield. When
benzyl chloroformate was added to the solution of compound
4 (normal addition), the resulting intermediate 5 was partially
decomposed to some unidentified impurities during the
course of addition, presumably through nucleophilic attack
of the acetylide 4 on the ester group of 5. To avoid this
Stability of Lithium Acetylide 4. It is important to
determine the stability of lithium salt 4 at different temper-
atures. The lithiation reaction must be carried out below a
temperature that does not result in decomposition of the
acetylide 4. One of major decomposition products appeared
to be the cleavage of the TMS protecting group, leading to
the formation of bis-lithium salt 6 as shown in Scheme 2.
The mechanism for the formation of 6 is unknown, but
presumably it occurs through an intermolecular exchange
(
(
(
1) Malaska, M. J.; Fauq, A. H.; Kozikowski, A. P.; Aagaard, P. J.; Mckinney,
M. Bioorg. Med. Chem. Lett. 1995, 5, 61.
2) Chackalamannil, S.; Davies, R. J.; Asberom, T.; Doller, D.; Leone, D. J.
Am. Chem. Soc. 1996, 118, 9812.
3) (R)-3-Butyn-2-ol was resolved from its racemic mixture via reaction with
phthalic anhydride, followed by treatment with (R)-1-phenylethylamine.
The product was isolated as a THF solution. The concentration was
determined by GC analysis.
(5) An analytically pure sample was isolated by preparative TLC (silica gel,
1
ethyl acetate/hexane, 30/70). H NMR (CDCl
3
) δ 7.40 (br, 10 H), 5.45 (q,
1
3
3
1H), 5.30 (s, 2H), 5.25 (s, 2H), 1.60 (d, 3H). C NMR (CDCl ) δ 152.8,
151.7, 133.6, 133.5, 127.70, 127.68, 127.64, 127.61, 127.40, 124.8, 83.4,
69.2, 66.9, 62.4, 28.7, 19.3.
(
4) Lai, G.; Colon, C. Synth. Commun. 1999, 29, 3011. The configuration of
the starting material 2 and the product 1 was described as (S) in this paper,
despite the fact that the (R) isomer was the development candidate.
1
(6) The content of impurity 7 was determined by H NMR analysis based on
the integral ratio of signals at 5.45 (q, 1H) for 7 and 4.65 (q, 1H) for 1.
3
08
•
Vol. 6, No. 3, 2002 / Organic Process Research & Development
10.1021/op010235a CCC: $22.00 © 2002 American Chemical Society
Published on Web 04/04/2002

Table 1. Stability of 4
Table 3. Effects of the amount of BuLi
entry
T (°C)
time (h)
7 (%)
entry
BuLi (equiv)
1 (1%)
1
2
3
-25
-15
0
6
5
5
<1
5
45
1
2
3
0.97 to 0.99
0.87
1.12
85-90
75
64
Table 2. Effect of order of addition
the acidity of an acetylene is much higher than an amine
pK
of approximately 23 and 38, respectively).8
Effects of Overcharge and Undercharge of BuLi. It is
(
a
entry
order of addition
T (°C)
yield (%)
1
2
3
4
normal
reverse
normal
reverse
-78
-78
-30
-30
85
88
32
86
crucial to charge exactly or near exactly 1 equiv of n-BuLi
to maximize the reaction yield. Due to the variable moisture
level and accuracy of the concentration measured for both
the starting material and n-BuLi solutions, addition of 1 equiv
of n-BuLi by calculation often led to unsatisfactory results.
To prevent an overcharge of n-BuLi, a preferable 0.80 equiv
of n-BuLi was initially charged, and the extent of lithiation
was monitored thereafter. Small increments of n-BuLi, if
necessary, were charged until a near-complete lithiation was
undesired side reaction, a reverse addition was performed
by adding a solution of 4 to the solution of benzyl
chloroformate. Since the concentration of lithium acetylide
4
remained very low in the reaction mixture and it reacted
much faster with benzyl chloroformate than with 5, a much
cleaner reaction and significantly higher yield were achieved.
As shown in Table 2, the effect of the addition order on the
product yield was minimal at the low temperature (entries 1
and 2), but a striking difference was observed at the elevated
temperature (entries 3 and 4). These results also suggested
that under reverse-addition conditions, the reaction yield was
much less sensitive to the reaction temperature. Better, the
highly exothermic nature of the reaction made it very difficult
to maintain the reaction at low temperature, and reverse
addition appeared to be a practical solution when running
large-scale productions.
9
reached. Implementation of an appropriate sampling tech-
nique also played a critical role in determining the lithiation
progress. When the lithiation process was well controlled
and the resulting solution was mixed with benzyl chloro-
formate by reverse addition, the yield for product 1 was
typically between 85 and 90%, regardless of the reaction
1
0
scale. As shown in Table 3, on one hand an undercharge
of n-BuLi led to a lower yield because of lower conversion,
whereas on the other hand, an overcharge of n-BuLi could
cause even more damage. Addition of excess n-BuLi clearly
produced 7 as the major impurity through intermediate 6,
suggesting that n-BuLi attacked the TMS protecting group
when acetylene 3 was not present.
Compatibility of BuLi and TMS Group. It has been
previously reported that the treatment of 3 with n-BuLi led
4
to decomposition even at -78 °C due to the incompatibility
of n-BuLi and the TMS protecting group; n-BuLi could only
be used as the lithiation reagent when 0.1 equiv of HMDS
was present. The explanation was that n-BuLi deprotonated
HMDS to form lithium bis(trimethylsilyl)amide in situ that
acted as a deprotonation reagent, converting 3 to 4. Our
experiments indicated that these statements were probably
made on the basis of incomplete studies and that the observed
decomposition might have been complicated by other factors
such as overcharging of n-BuLi. Extensive experiments
showed that there was no apparent compatibility issue
between n-BuLi and the TMS group even if the lithiation
was done at a considerably high temperature of -25 °C. It
was not only unnecessary to add the extra 0.1 equiv of
HMDS before the n-BuLi charge, but also a reaction from a
solution of 3 in THF containing no HMDS or a trace amount
Conclusions
The reaction conditions that can affect this one-pot process
have been carefully examined and optimized. The study
clarified the compatibility issues between n-BuLi and the
TMS protecting group. The process has been scaled to
produce multikilogram product 1 in the plant with excellent
yield by employing the conditions presented in this report,
and product 1 has been used without further purification for
the synthesis of himbacine analogues.
Experimental Section
Benzyl chloroformate, racemic 3-butyn-2-ol, HMDS, and
n-BuLi were purchased from Aldrich. NMR spectra were
recorded on a Bruker 400 MHz spectrometer. HPLC analysis
was performed on Waters 2690 Alliance using the following
of HMDS provided a slightly better yield of the final product
7
(8) March, J. AdVanced Organic Chemistry, 4th ed.; John Wiley & Sons: 1992;
p 252.
1
(85-89%). Decomposition was observed only in experi-
ments where an excess amount of n-BuLi was charged. In
addition, it is generally unlikely that n-BuLi would depro-
tonate HMDS in preference to an acetylenic hydrogen since
(9) Sampling must be done completely under nitrogen because of the highly
hydroscopic nature of the reaction mixture. Our experiments showed that
3
quenching the reaction mixture into CD COOD inside the reaction flask
under nitrogen provided consistent results. The vial that contained CD
3
-
COOD was brought into the flask by tweezers. A few drops of reaction
mixture were withdrawn to a dry pipet and quenched into the NMR solvent
immediately. Precautions should be taken to keep the reaction mixture only
on the tip of the pipet. Inappropriate sampling often led to underestimating
the extent of the lithiation and unsatisfied results. A similar sampling
technique was successfully used in the plant
(
7) Compound 2 was converted to 3 quantitatively by refluxing an acidic
solution of 2 in THF (pH ≈ 3) with slight excess of HMDS. The resulting
solution contained small amounts of HMDS and ammonium salts as the
impurities. Compound 3 could be isolated as a THF solution by distillation.
The distillate contained either no HMDS or trace amount of HMDS. Both
distilled or nondistilled solution could be used for the next step. However,
controlled experiments showed that the reactions from distilled solutions
consistently gave slightly better yield.
(10) For production in the plant, the THF solution of the acetylide was transferred
to the reactor containing benzyl chloroformate solution through an insulated
stainless steel tube (diameter 0.25 in.).
Vol. 6, No. 3, 2002 / Organic Process Research & Development
•
309

conditions: column, Symmetry C18 (3.9 mm × 150 mm,
Waters); mobile phase, MeCN/H O (40:60); flow rate of 1.0
mL/min; and detector at UV 230 nm.
transferred slowly through a cannula to a solution of benzyl
chloroformate (128 mmol, 1.2 equiv) in 50 mL of THF at
2
1
0
-30 °C. The reaction mixture was stirred for 30 min and
quenched with 50 mL of 6 N H SO solution. The mixture
Preparation of 1. A solution containing 7.5 g (107 mmol)
of (R)-3-butyn-2-ol was reacted with HMDS by using the
procedure reported previously.4 Compound 3 was then
distilled out at 1 atm as a THF solution. Toluene (10 mL)
was then added to the distillation flask, and the mixture was
distilled out and mixed with the THF solution. The resulting
solution was cooled to about -30 °C, and an n-BuLi solution
in hexane (86 mmol, 0.80 equiv) was charged while
maintaining the reaction temperature between -25 and -30
2
4
was stirred for 1 h at room temperature. Toluene (50 mL)
was charged, and the organic phase was separated and treated
with 5% ammonium hydroxide solution to destroy the excess
benzyl chloroformate. The organic solution was washed with
water and dried on Na SO . Solvent was removed under
2 4
vacuum to give 30.6 g of red oil. The yield of 1 was
determined by HPLC as 18.6 g, 86% yield. An analytical
sample was isolated by preparative HPLC as pale-yellow
1
9
1
°
C. The reaction was monitored by H NMR, and additional
oil. H NMR (CDCl
3
) δ 7.40 (m, 1H), 5.23 (s, 2H), 4.65 (q,
13
amounts of n-BuLi were charged based on NMR analysis.
1H), 2.10 (br, 1H), 1.53 (d, 3H). C NMR (CDCl
134.6, 128.7, 128.6, 88.8, 75.6, 67.8, 58.0, 23.2.
3
) δ 153.2,
1
To analyze the extent of lithiation by H NMR, a few drops
3
of the reaction mixture were quenched into 1.0 mL of CD -
COOD. The reaction was assessed on the basis of the integral
ratio of the peaks at 2.2 ppm (doublet, the proton on acetylene
carbon) to the peaks at 4.3 ppm (multiple, the proton on
hydroxy carbon). A ratio of 0.2 to 1.0, for example, indicated
that 20% of 3 was still present and an additional amount of
BuLi (20% of the amount initially charged) should be added.
The same sampling procedure was repeated until the depro-
tonated acetylene was greater than 97%. The solution was
Acknowledgment
We thank Mr. Anthony Toto for his supportive efforts
during plant production and Dr. Evgeny Aronov for purifica-
tion of compound 1 by preparative HPLC.
Received for review October 25, 2001.
OP010235A
310
•
Vol. 6, No. 3, 2002 / Organic Process Research & Development