Enzyme-catalyzed hydrolysis of bicycloheptane and cyclobutene diesters to monoesters

Article abstract of DOI:10.1021/op400254c

Diacid formation is a major problem in the conventional chemical hydrolysis of a diester to a monoester. Enzyme-catalyzed hydrolysis of dimethyl bicyclo[2.2.1]heptane-1,4-dicarboxylate (1) by lipases from Candida antarctica and Burkholderia cepacia gave t

Full text of DOI:10.1021/op400254c

Article
pubs.acs.org/OPRD
Enzyme-Catalyzed Hydrolysis of Bicycloheptane and Cyclobutene
Diesters to Monoesters
Zhiwei Guo, Michael Kwok Y. Wong, Matthew R. Hickey, Bharat P. Patel, Xinhua Qian,
and Animesh Goswami*
Chemical Development, Bristol-Myers Squibb, New Brunswick, New Jersey 08903, United States
S
* Supporting Information
ABSTRACT: Diacid formation is a major problem in the conventional chemical hydrolysis of a diester to a monoester. Enzyme-
catalyzed hydrolysis of dimethyl bicyclo[2.2.1]heptane-1,4-dicarboxylate (1) by lipases from Candida antarctica and Burkholderia
cepacia gave the corresponding monoester 4-(methoxycarbonyl)bicyclo[2.2.1]heptane-1-carboxylic acid (2) in excellent yields
with negligible amounts of diacid 3. About 100 kg of monoester 2 was prepared in 78% yield by hydrolysis of diester 1 with a
commercially available lipase from B. cepacia. A more efficient process for the hydrolysis of 1 that give monoester 2 in 82% yield
was subsequently developed using significantly lower amounts of the commercially available immobilized lipase B from C.
antarctica. The commercially available immobilized lipase B from C. antarctica and porcine liver esterase were also efficient for the
selective hydrolysis of dimethyl cyclobut-1-ene-1,2-dicarboxylate (4) to the corresponding monoester 5 in yields of 78% and
87%, respectively.
hydrolytic activity showed lower selectivity and produced
higher levels of diacid 3 along with monoester 2. Lipases from
C. antarctica, Candida rugosa, and Burkholderia cepacia
(previously known as Pseudomonas cepacia) showed the highest
activities and selectivities. After a few small-scale experiments,
the lipase from B. cepacia (Amano Lipase PS-30) was found to
have the best combination of activity and selectivity. We had
several kilograms of this lipase in our stock and dedicated our
efforts to the quick development of a process for preparing
monoester 2 on the larger scale required for the development
of a prospective drug candidate.
Process Development Using the Lipase from B. cepacia,
Lipase PS-30. The effects of cosolvent, temperature, and pH on
Lipase PS-30-catalyzed hydrolysis of diester 1 were evaluated.
Seven commonly used cosolvents (acetonitrile, isopropanol,
tert-butanol, tetrahydrofuran, 2-methyltetrahydrofuran, dime-
thylformamide, and dimethyl sulfoxide) were evaluated at the
10% v/v level, but none of them showed any improvement;
consequently, a cosolvent was not used for scale-up. The effects
of temperature and the pH of the starting buffer were evaluated
on a 1 mL scale with a substrate concentration of 20 g/L at
various levels of enzyme loading. As expected, the reaction rate
increased with increasing temperature. The reaction was faster
when the starting buffer was at pH 7 compared with either pH
6 or pH 8. The pH dropped significantly to 4−5 in the small-
scale reactions without any pH control.
INTRODUCTION
■
A useful approach for desymmetrization of a symmetric diester
is its selective hydrolysis to the corresponding monoester.1,2
Diacid formation is a major problem in the conventional
chemical hydrolysis of a diester. In the case of the diester
dimethyl bicyclo[2.2.1]heptane-1,4-dicarboxylate (1),1−7 sig-
nificant levels of diacid were generated with potassium
hydroxide,1,4,6 sodium hydroxide,5,7 or barium hydroxide2 as
the base even at <100% conversion (Scheme 1). The
downstream isolation process for the separation of the
remaining diester 1, the desired monoester 2, and the undesired
diacid 3 required careful pH control and was tedious.
Monoester 2 is a building block of many potential therapeutic
candidates.5−7 Enzymatic hydrolysis is a powerful tool for the
selective hydrolysis of esters.8,9 Hydrolysis and desymmetriza-
tion of diesters to monoesters by porcine liver esterase,10−14
porcine pancreatic lipase,11,12 lipase from Psudomonas species,13
lipase from Candida antarctica,12,14,15 and other lipases16,17 have
been reported extensively in the literature. We previously
reported4 preliminary results of our work on the synthesis and
chemical and enzymatic hydrolysis of 1. This paper describes
detailed work on the selective enzyme-catalyzed hydrolysis of 1
and dimethyl cyclobut-1-ene-1,2-dicarboxylate (4) to their
corresponding monoesters 2 and 5, respectively.
RESULTS AND DISCUSSION
■
Hydrolysis of Dimethyl Bicyclo[2.2.1]heptane-1,4-
dicarboxylate (1) to Monoester 2. Enzyme Screening.
More than 100 hydrolytic enzymes were screened for the
hydrolysis of 1. The promising enzymes identified are listed in
Table 1. Many enzymes showed good hydrolytic activities and
gave monoester 2 as the major product. Most lipases that
showed hydrolytic activity also had good selectivity for
monoester 2 and produced undetectable or negligible amounts
of diacid 3. On the other hand, proteases that showed
The optimum conditions were developed by carrying out the
reaction on a larger scale at constant pH and temperature. A
constant pH of 7, a reaction temperature of 40 °C, a diester 1
input of 50 g/L, and a Lipase PS-30 concentration of 25 g/L
Special Issue: Biocatalysis 14
Received: September 16, 2013
Published: November 5, 2013
© 2013 American Chemical Society
774
dx.doi.org/10.1021/op400254c | Org. Process Res. Dev. 2014, 18, 774−780

Organic Process Research & Development
Article
Scheme 1. Hydrolysis of Dimethyl Bicyclo[2.2.1]heptane-1,4-dicarboxylate (1)
Table 1. Promising Enzymes for the Hydrolysis of Diester 1
amounts (%) at the end of the reaction
enzyme
source
supplier
Amano
diester 1
monoester 2
diacid 3
<0.5
Lipase PS-30
Lipase SL
Burkholderia cepacia
Burkholderia cepacia
Candida rugosa
3
3
97
97
98
98
36
32
27
31
3
Meito Sangyo
Meito Sangyo
Novozymes
Amano
<0.5
<0.5
1
Lipase OF
2
Novozym 435
Lipase M-10
Lipase AY-30
Protease P-6
Protease P4032
porcine liver esterase
Candida antarctica
Mucor javanicus
Candida rugosa
1
64
68
67
62
17
<0.5
<0.5
6
Amano
Aspergillus melleus
Aspergillus oryzae
porcine liver
Amano
Sigma
7
Sigma
80
(2:1 substrate to enzyme ratio) were found to be optimal for
the reaction.
and there are numerous reports in the literature on its use in
organic synthesis.19−21 Since Lipase PS-30 is no longer
commercially available, it would be interesting to see how the
new formulation Lipase PS-SD would behave in organic
synthesis examples reported in the past. The difference between
the activities of Lipase PS-30 and Lipase PS-SD could not be
due to possible differences in fillers commonly used in solid
enzyme formulations since two common fillers (diatomaceous
earth and dextrin) had no effect on their activities for the
hydrolysis of diester 1. Fortunately, the selectivity (monoester 2
to diacid 3 ratio) of Lipase PS-SD for the hydrolysis of diester 1
was high, similar to that of Lipase PS-30. We decided to quickly
optimize and develop a process to scale-up the hydrolysis of 1
with Lipase PS-SD to supply monoester 2 in time for
development studies, knowing fully that a large ratio of Lipase
PS-SD to substrate would be required and that a better enzyme
might be required in the long run.
Process Development Using Lipase PS-SD from B. cepacia.
Again, eight commonly used cosolvents were evaluated, and as
in the case of Lipase PS-30, there was no improvement; thus, a
cosolvent was not used for further Lipase PS-SD process
development. Lipase PS-SD-catalyzed hydrolysis of diester 1
was evaluated on a 1 mL scale with 20 mg of substrate and
various enzyme concentrations at different temperatures for 24
h. The best temperature was found to be 50 °C. However, there
was a sharp drop in activity above 50 °C. Hence, the
temperature was maintained between 45 and 50 °C for
optimum enzyme activity and was carefully controlled to avoid
overheating. With 20 g/L substrate and 30 g/L Lipase PS-SD at
50 °C, the conversion reached 99.5% in 24 h and produced a
negligible amount of diacid 3.
Three hydrolysis batches were carried out under the
optimum conditions using 251, 266, or 304 g of crude diester
1 in the batch. The diester 1 used in these experiments was
crude, with potencies of about 87%, and contained considerable
amounts of several unknown impurities originating from its
synthesis. The enzyme tolerated the impurities, and the
reaction reached 98% conversion in 24 h with negligible
formation of diacid 3 in all cases. Most of the impurities were
removed together with the remaining diester 1 during the
extraction with MTBE at pH 8. The aqueous phase was then
immediately acidified and extracted with MTBE to obtain
monoester 2 with an average “as is” yield of 78% and corrected
(for the potency of starting material) yield of 90% for the three
batches. The success of the enzymatic hydrolysis step with
crude diester 1 and the formation of pure monoester 2 in high
yield and high purity showed that the process is robust and
efficient.
Lipase from B. cepacia, Lipase PS-SD. After developing the
process for scale-up with Lipase PS-30 from B. cepacia, we
required a large supply of Lipase PS-30 to prepare large
quantities of monoester 2. However, we learned that Lipase PS-
30 was no longer commercially available and was replaced by
Lipase PS-SD obtained from the same source (B. cepacia) by its
supplier, Amano. Though both Lipase PS-30 and Lipase PS-SD
have the same activity (as per certificate of analysis from
Amano) of 30 000 units/g for the hydrolysis of the standard
substrate olive oil, Lipase PS-SD was found to have only one-
third the activity of Lipase PS-30 for the hydrolysis of diester 1.
The difference in activities for the standard substrate (olive oil)
and the specific compound 1 displayed by the two enzyme
formulations from the same source highlights an important
issue. The activity of the enzyme with the actual substrate
rather than a general compound is the key criterion for its use
in a biocatalytic synthetic step. Wells et al.18 also suggested the
activity of the enzyme versus the actual substrate as a key
specification of enzymes for the synthesis of active
pharmaceutical ingredients. Lipase PS-30 is a versatile enzyme,
Scale-Up of the Process Using Lipase PS-SD from B.
cepacia. The optimum conditions for the Lipase PS-SD-
catalyzed hydrolysis were developed, and the enzymatic
hydrolysis was carried out on increasing scale. A constant pH
of 7, a reaction temperature of 48° 2 °C, a diester 1 input of
50 g/L, and a Lipase PS-SD concentration of 75 g/L (2:3
substrate to enzyme ratio) were found to be optimal for the
reaction.
775
dx.doi.org/10.1021/op400254c | Org. Process Res. Dev. 2014, 18, 774−780

Organic Process Research & Development
Article
The hydrolysis was scaled up to a 405 g batch of crude
diester 1 using 600 g of Lipase PS-SD at 50 g/L substrate input
Interestingly, the adsorption of the substrate on the
immobilized enzyme beads became less pronounced as the
reaction temperature was increased from 40 to 60 °C. A
maximum operating temperature of 60 °C was recommended22
for optimum productivity with Novozym 435. The optimum
temperature for the hydrolysis of 1 was established to be 55 °C.
Diacid formation at 60 °C was only slightly higher than that at
55 °C, and thus, a slight excursion of temperature above 55 °C
should not be detrimental. The optimum pH for the reaction
was 7, and the pH was kept constant during the reaction by
addition of NaOH as needed. The optimum enzyme to
substrate ratio was 1:50, requiring only 1 g/L Novozym 435 to
hydrolyze 50 g/L diester 1. The workup was significantly easier
than the previous processes with Lipase PS-30 or Lipase PS-SD
with no emulsion issues.
Scale-Up of the Process Using Novozym 435. The
Novozym 435-catalyzed hydrolysis was scaled up with
increasing amounts of diester 1 up to a 500 g scale. The
diester 1 for the 500 g scale reaction was crude, with an HPLC
(UV 210 nm) AP of 41, and contained several impurities
originating from its synthesis. One impurity was identified as
2,6-di-tert-butyl-4-methylphenol (BHT) with an AP of 35.
There were other unknown impurities, the highest being the
one with an AP of 6, and several unknown impurities each with
AP as high as 2. BHT and probably other impurities have
stronger UV absorption and HPLC areas, causing their
amounts to be overestimated. The potency of the diester 1
was high (90%), and it was used as such for the enzyme
reaction without purification. The enzyme-catalyzed hydrolysis
of diester 1 (500 g) was carried out with 10 g of Novozym 435
under the optimum reaction conditions at 55 °C and pH 7 with
manual pH adjustment. The composition of the reaction
mixture at various times is shown in Figure 1. At 31 h, the
at 48
2 °C. The conversion reached 97.7% at 25 h with
negligible diacid 3 formation. After workup, 300 g of monoester
2 (79% “as is” yield, 90% yield after correction for the potency
of starting diester) was obtained.
The Lipase PS-SD-catalyzed hydrolysis process was then
scaled up to 80 kg of crude diester 1 input to obtain 58 kg of
monoester 2 in 78% “as is” yield with high purity (AP > 97;
diester 1, AP < 0.5; diacid 3, AP < 0.5). About 100 kg of total
monoester 2 was produced by the Lipase PS-SD-catalyzed
hydrolysis of diester 1 to support preparation of a drug
candidate. The optimum substrate to enzyme ratio for Lipase
PS-SD (2:3) is 3 times that of Lipase PS-30 (2:1) for hydrolysis
of diester 1; thus, Lipase PS-SD required 3 times the enzyme
loading compared with Lipase PS-30 for reaction completion in
the same amount of time.
Immobilized Lipase B from C. antarctica, Novozym 435.
Though the Lipase PS-SD enzyme-catalyzed hydrolysis
provided about 100 kg of monoester 2 for drug development
studies, we still desired a better enzyme that could catalyze the
hydrolysis at a much lower enzyme loading. Furthermore, some
emulsion was encountered during the workup of both the
Lipase PS-30 and Lipase PS-SD-catalyzed hydrolysis proce-
dures, and either Celite pad filtration or centrifugation had to
be employed to break up the emulsion. We thought that an
immobilized enzyme such as Novozym 435, where the protein
is immobilized on an insoluble carrier, might solve the emulsion
issue. Novozym 435, a product of Novozymes, is lipase B from
C. antarctica immobilized on a macroporous resin. During the
initial enzyme screening, Novozym 435 had been identified as a
promising enzyme on the basis of its activity, but it was not
selected because of the increased level of diacid 3 formation
compared with Lipase PS-30. It was later found (see below)
that the formation of diacid 3 by Novozym 435 could be
minimized by appropriately controlling the substrate to enzyme
ratio, reaction pH, and reaction temperature.
Process Development Using Novozym 435. The Novozym
435-catalyzed hydrolysis was evaluated on a 40 mg substrate
scale in 1 mL of sodium phosphate buffer (0.2 M, pH 7) at 40
°C for 24 h at various substrate to enzyme ratios. The enzyme
loading had a significant effect on the reaction rate and the
amount of diacid 3 formation. As expected, more diacid was
formed at high enzyme loading. The 1 mL scale experiments
were performed without pH adjustment during the reaction,
and the pH dropped to ∼5 as the hydrolysis progressed. The
desired monoester 2 was probably not ionized completely at
pH ∼5 and was likely to remain adsorbed on the nonpolar
acrylic beads of the immobilized enzyme, thus undergoing
further hydrolysis to diacid 3. This was probably the major
reason for the higher levels of diacid 3 seen during enzyme
screening. It was especially important to maintain the pH close
to 7 to minimize further hydrolysis of monoester 2 to diacid 3
when using an immobilized enzyme. Eight commonly used
cosolvents (10% v/v in buffer) were evaluated, and as with the
lipase PS-SD, the addition of cosolvent did not show any
benefit and was not pursued.
Figure 1. Enzymatic hydrolysis of diester 1 to monoester 2 by
immobilized lipase B from C. antarctica (Novozym 435).
monoester 2:diester 1:diacid 3 ratio was 97.9:1.5:0.6. Extractive
workup gave 370.7 g of monoester 2 (93% potency, 82% “as is”
yield, 91% yield corrected for the potency of the starting
diester). HPLC showed good purity, with APs of 86.1 for
monoester 2 and 0.3 for diacid 3, no detectable diester 1, no
detectable BHT, and a total AP of 13.6 for eight impurities, all
coming from the starting material. The crude monoester 2 was
successfully applied for the subsequent synthetic step where
these impurities were purged.
Reuse of Novozym 435. In order to demonstrate the
potential for recycling, Novozym 435 was collected by filtration
after one reaction. The collected Novozym 435 was sticky,
which may be due at least in part to the adsorption of
impurities present in the diester on the immobilized resin. The
Further optimization of the Novozym 435-catalyzed
hydrolysis of diester 1 was performed on a 1 g scale at a
substrate concentration 50 g/L. Various enzyme to substrate
ratios and reaction temperatures were evaluated. Diester 1
immediately adsorbed onto the immobilized enzyme beads and
remained as a large sticky mass in the reaction mixture.
776
dx.doi.org/10.1021/op400254c | Org. Process Res. Dev. 2014, 18, 774−780

Organic Process Research & Development
Article
Scheme 2. Hydrolysis of Dimethyl Cyclobut-1-ene-1,2-dicarboxylate (4)
recovered Novozym 435 could not be reused as such. It was
found that washing with MTBE gave a free-flowing material,
probably by removing the impurities from the resin surface.
After the MTBE wash, the recovered enzyme was dried in a
vacuum oven at room temperature overnight to remove
residual MTBE. The recovered enzyme (300 mg) was used
to hydrolyze 12 g of diester 1 under the same optimum
conditions. The conversion reached 93% in 54 h with a 0.9%
yield of diacid 3. The recovered enzyme showed the same
selectivity as the fresh enzyme but only about 60% of the
original activity. MTBE washing might have reduced the
enzyme activity. Other solvents for washing can be explored to
minimize loss of activity. However, reuse of enzyme requiring
such a low loading (enzyme to substrate ratio 1:50) is not a
significant concern in this case and was not pursued further.
Hydrolysis of Dimethyl Cyclobut-1-ene-1,2-dicarbox-
ylate (4) to Monoester 5. Like the bicycloheptane diester 1,
conventional chemical hydrolysis of dimethyl cyclcobut-1-ene-
1,2-dicarboxylate (4) to the monoester 5 (Scheme 2) was
difficult without significant hydrolysis to the diacid 6. Selective
methylation of diacid 6 by DBU and iodomethane was reported
to provide monoester 5 in 66% yield.23 Screening of enzymes
for the hydrolysis of 4 to monoester 5 (Scheme 2) identified
Novozym 435 as one of the best. Interestingly, porcine liver
esterase (PLE) was found to be even better than Novozym 435.
Both hydrolyses afforded almost no diacid 6. The optimum
reaction conditions with both enzymes were developed and
used to carry out larger-scale reactions. The hydrolysis of
diester 4 at 50 °C for 4 h catalyzed by Novozym 435 with an
enzyme to substrate ratio of 1:20 and a simple isolation method
provided monoester 5 in 78% yield with excellent quality
(monoester 5, AP = 99.3; remaining diester 4, AP = 0.7; diacid
6 not detected). Hydrolysis of diester 4 at 40 °C for 5.5 h
catalyzed by PLE with an enzyme to substrate ratio of 1:200
and a simple isolation method provided monoester 5 in 87%
yield with excellent quality (monoester 5, AP = 99.7; remaining
diester 4, AP = 0.3; diacid 6 not detected).
determined from the relative area percentages (APs) of the
HPLC peaks without any correction. HPLC area versus
concentration standard curves for authentic standards were
used to determine the potencies of various batches of diesters
and monoesters.
Two reversed-phase HPLC methods were used to monitor
hydrolysis of 1. Both employed a Waters XTerra RP-18 column
(3.5 μm, 150 mm × 4.6 mm) at ambient temperature with UV
detection at 210 nm with a flow rate of 1 mL/min of solvent A
(0.05% TFA in 80:20 water/methanol) and solvent B (0.05%
TFA in 80:20 acetonitrile/methanol), but they used different
gradients. HPLC method 1 was used for enzyme screening with
a gradient from 20% to 70% solvent B over 8 min. The
retention times were 3.4 min for diacid 3, 5.2 min for
monoester 2, and 7.0 min for diester 1. HPLC method 2 was
used to analyze the reaction progress and impurity profile with
a gradient from 0 to 100% solvent B over 20 min. The retention
times were 8.0 min for diacid 3, 10.1 min for monoester 2, and
12.0 min for diester 1.
HPLC method 3 was used to monitor the hydrolysis of 4
using a Waters XTerra RP-18 column (3.5 μm, 50 mm × 4.6
mm) at ambient temperature with UV detection at 220 nm
with a flow rate of 2 mL/min with solvent A (0.05% TFA in
5:95 acetonitrile/water) and solvent B (0.05% TFA in 95:5
acetonitrile/water), employing 0% B for 2 min followed by a
gradient from 0 to 100% B for 2 min. The retention times for
diester 4, monoester 5 and diacid 6 were 3.0, 1.5, and 0.6 min,
respectively.
Enzyme Screening for Hydrolysis of Diester 1 or 4 to
Monoester 2 or 5, respectively. Sodium phosphate buffer
(0.1 M, pH 7, 1 mL) was added to each well of multiwell plates,
with each well containing about 10 mg of enzyme. Each plate
was shaken in a Thermomixer R shaker at 600 rpm and 25 °C
for 30 min. A solution containing 4 mg of diester 1 in 20 μL of
DMSO was added to each well. The enzymatic hydrolysis was
conducted by shaking the plate in the same shaker for 24 h at
30 °C. To each well were added 50 μL of 1 M HCl and 1 mL of
methanol. After 5 min of mixing, the mixture was filtered
through a 0.2 μm filter and analyzed by HPLC.
A solution of 10 mg of diester (1 or 4) in 50 μL of DMSO
was used for some enzymes. For immobilized enzymes or
enzyme solutions that were not in any multiwell plate format,
the screening experiments were carried out in 4 mL vials using
the same buffer. The vial was placed in the multiwell plate, and
the plate was shaken in a Thermomixer R shaker at 30 °C for
24 h before analysis in the same way as before.
In summary, monoester 2 was prepared by hydrolysis of
dimethyl bicyclo[2.2.1]heptane-1,4-dicarboxylate (1) with a
lipase from B. cepacia and immobilized lipase B from C.
antarctica. The immobilized lipase B from C. antarctica and
porcine liver esterase were also efficient for the selective
hydrolysis of dimethyl cyclobut-1-ene-1,2-dicarboxylate (4) to
the corresponding monoester 5.
EXPERIMENTAL SECTION
■
Diester 1 was purchased from commercial suppliers and
prepared on a large scale as described in the literature.3−7
Diester 4 was purchased from commercial suppliers. Lipases
PS-30 and PS-SD from B. cepacia were purchased from Amano.
Immobilized Lipase B from C. antarctica (Novozym 435) was
purchased from Novozymes. PLE was purchased from Sigma.
Other commercial enzymes were purchased from their
suppliers.
Optimization of Lipase PS-30-Catalyzed Hydrolysis of
1 on a Small Scale: Effects of Cosolvent, Temperature,
and pH. The experiments were carried out in 4 mL vials. To
each vial were added diester 1 (20 mg), Lipase PS-30 (2, 10, or
20 mg), 1 mL of sodium phosphate buffer (0.1 M, pH 6, 7, or
8), and an organic solvent or water (100 μL). The vials were
placed on multiwell plates, and the plates were shaken in a
Thermomixer shaker at 600 rpm at different temperatures (30,
35, or 40 °C). After 24 or 65 h, 50 μL of 1 M HCl and 3 mL of
HPLC Methods. The conversion and the amounts of
substrates and products in the reaction mixture were
777
dx.doi.org/10.1021/op400254c | Org. Process Res. Dev. 2014, 18, 774−780

Organic Process Research & Development
Article
methanol were added to the whole reaction mixture, followed
by filtration through a 0.2 μm filter and analysis by HPLC.
Lipase PS-30-Catalyzed Hydrolysis of Diester 1. To a
500 mL jacketed reactor equipped with a pH meter and an
overhead stirrer were added 10 g of pure diester 1, 5 g of Lipase
PS-30, and 200 mL of sodium phosphate buffer (0.1 M, pH
7.0). The temperature was maintained at 40 1 °C using a
water circulator. The mixture was stirred at 300 rpm. The pH
dropped slowly and was adjusted back to 7.0 periodically by
addition of 5 M NaOH (every 2 h during the first 8 h, no
adjustment between 8 and 24 h, and then every hour from 24
to 31 h. Samples (200 μL) of the reaction mixture were taken
out with a wide-bore pipet tip at various times, mixed with 20
μL of 1 M HCl and 3 mL of MeOH, filtered, and analyzed by
HPLC. At 31 h, about 2.8% of diester 1 remained. The mixture
was cooled to about 23 °C, adjusted to pH 8.0 with 5 M
NaOH, and extracted with 300 mL of MTBE. The MTBE layer
containing the unreacted diester 1 was discarded. The aqueous
phase was immediately adjusted to pH 2 using 5 M H2SO4 and
then extracted twice with MTBE (300 mL for the first time and
200 mL for the second). The MTBE phase contained some
emulsion and was filtered through a thin pad of Celite to help
with phase separation. Removal of solvent from the combined
MTBE phases gave 8.63 g of monoester 2 as a white solid
(92.4% yield, AP = 99.7) along with an unknown impurity (AP
= 0.3), and neither diester 1 nor diacid 3 was detected.
Lipase PS-SD-Catalyzed Hydrolysis of Diester 1. To a 1
L jacketed reactor equipped with a pH meter and an overhead
stirrer were added 20 g of diester 1 (crude, AP = 60, 87%
potency), 30 g of Lipase PS-SD, and 400 mL of sodium
phosphate buffer (0.2 M, pH 7.0). The temperature was
maintained at 48 2 °C. The mixture was stirred at 300 rpm.
The pH dropped slowly and was adjusted back to 7.0
periodically by addition of 5 M NaOH. At 25 h, the conversion
reached 97.4% with 2.6% of the diester 1 remaining and a
negligible amount of diacid 3. The reaction mixture was cooled
to about 23 °C, adjusted to pH 8.0 with 5 M NaOH, and then
extracted with 300 mL of MTBE. Removal of solvent from this
MTBE extract gave 2.2 g (11% of the starting amount) of
material containing unreacted diester 1 (no monoester 2 or
acid 3) and other impurities present in the starting crude
diester. The aqueous phase was immediately acidified and
extracted with MTBE (2 × 400 mL). The MTBE phase
contained some emulsion and was filtered through a thin pad of
Celite to help with phase separation. The combined MTBE
phases were washed with 100 mL of water, concentrated to
dryness, and further dried in a vacuum oven at 30 °C overnight
to give 15.26 g of monoester 2 (AP = 95, 82% “as is” yield), and
neither the diester nor the diacid was detected. A total HPLC
AP of 5 was due to nine unknown impurities coming from the
starting material. The corrected yield for this batch calculated
from the “as is” yield divided by 0.87 (the potency of the crude
starting diester) was 94%.
the pH dropped slowly and was not adjusted. Samples were
taken periodically and analyzed in the same way as described
above. At 23 h, HPLC analysis showed that the reaction
mixture contained 0.4% diester 1, 98.9% monoester 2, and 0.7%
diacid 3.
Novozym 435-Catalyzed Hydrolysis of Diester 1. The
diester 1 had a potency of 90% and an HPLC (UV 210 nm) AP
of 41. There was a major impurity of BHT with AP = 35,
another unknown impurity with AP = 6, and several other
unknown impurities each with an AP as high as 2. To a 20 L
jacketed reactor equipped with an overhead stirrer, a pH meter,
and a water circulator were added 10 g of Novozym 435, 500 g
of crude diester 1, and 10 L of sodium phosphate buffer (0.2 M,
pH 7.0). The mixture was stirred at 150 rpm while being
gradually heated to and maintained at 55
2 °C. At the
beginning, Novozym 435 resins were floating on the top, and
they gradually sank into the buffer mixture in about 1 h. The
diester 1 appeared as a slurry suspension of particles at the start,
gradually became a sticky gel, and finally melted into fine
droplets. The pH of the reaction mixture dropped slowly with
the progression of hydrolysis and was manually adjusted back
to 7.0 with 5 M NaOH every 30 min during the first 4 h, every
hour during the following 4 h, and every 2 h until 12 h. After 12
h, the pH dropped slowly and was not adjusted until 23 h. The
pH was again adjusted back to 7.0 with 5 M NaOH at 23 and
28 h. Samples were taken periodically and analyzed as described
before. At 31 h, HPLC showed that the reaction mixture
contained 97.9% monoester 2, 1.5% diester 1, and 0.6% diacid
3. The stirring was stopped, and the reactor was cooled to 4 °C
for 16 h. This holding is not necessary if the workup can be
done immediately.
The mixture was warmed to 20 °C, adjusted to pH 8.0 with 5
M NaOH, and filtered through a Buchner-style M porosity
filter funnel. The immobilized enzyme residue was washed with
1 L of phosphate buffer (0.2 M, pH 8.0). The combined filtrate
and washing were twice extracted with 5 L of MTBE to remove
remaining unreacted diester 1 along with some impurities
coming from the starting material. The aqueous phase was
immediately adjusted to pH 2.5 with 5 M H2SO4 (467 mL) and
twice extracted with 5 L of MTBE. The acidic MTBE extracts
were combined, washed with 3.5 L of water, and concentrated
on a rotary evaporator to give a white solid, which was further
dried in a vacuum oven at 30 °C overnight to give 370.7 g of
monoester 2 (93% potency, 82% “as is” yield, 91% yield
corrected for the potency of the starting diester). HPLC
showed good quality of monoester 2 (AP = 86.1), a small
amount of diacid 3 (AP = 0.3), no detectable diester 1, no
detectable BHT, and eight impurities all coming from the
starting material (total AP of 13.6). The crude monoester 2 was
successfully applied in the next synthetic step, where these
impurities were purged.
Novozym 435-Catalyzed Hydrolysis of Diester 4. To a
2 L jacketed reactor equipped with an overhead stirrer, an
addition funnel, and a pH meter were added 3.0 g of Novozym
435 and 1200 mL of sodium phosphate buffer (0.2 M, pH 7.0).
The reaction temperature was maintained at 50 2 °C. After
the inside temperature rose to 50 °C, 60.0 g of diester 4 (353
mmol) was added. The mixture was stirred, and the pH was
maintained at 6.8−7.1 by addition of 1 M NaOH as necessary.
Samples of the reaction mixture (100 μL) were taken out
periodically using a wide-bore pipet tip, mixed with 20 μL of 1
M HCl and 10 mL of MeOH, filtered through a 0.2 μm filter,
and analyzed by HPLC. The amounts of diester 4, monoester
Optimization of Novozym 435-Catalyzed Hydrolysis
of Diester 1. Initial studies to explore the best reaction
conditions using Novozym 435 were done on a 1 mL scale in 4
mL vials, which were followed by a 1 g scale. The best
conditions developed are as follows. To the reactor were added
20 mg of Novozym 435, 1 g of diester 1, and 20 mL of sodium
phosphate buffer (0.2 M, pH 7.0). The mixture was stirred with
a magnetic stir bar. The temperature was maintained at 55
2
°C. The pH dropped slowly and was adjusted back to 7.0 using
aqueous 5 M NaOH each hour during the first 8 h. After 8 h,
778
dx.doi.org/10.1021/op400254c | Org. Process Res. Dev. 2014, 18, 774−780

Organic Process Research & Development
Article
5, and diacid 6 after 4 h were 2.7%, 88%, and 9.3%, respectively.
After 4 h the reaction was stopped, and the immobilized
enzyme was filtered out. The filtrate was extracted with EtOAc,
and the EtOAc layer containing the unreacted diester 4 was
discarded. The aqueous layer was acidified to pH 3.0 with 6 M
HCl (45 mL) and extracted with 900 mL of EtOAc. The pH of
the aqueous phase was pH 3.8 after the first extraction. To the
aqueous phase was added 150 g of NaCl, and the pH was
adjusted to 3 with 6 M HCl (15 mL) before a second extraction
with 600 mL of EtOAc. Again, the aqueous phase was adjusted
to pH 3.0 with 6 M HCl (10 mL) and then was extracted a
third time with 600 mL of EtOAc. The aqueous phase was
analyzed after each extraction to ensure complete extraction of
the monoester into the organic phase. Each successive
extraction showed reduced levels of monoester in the aqueous
phase. After the third extraction, the aqueous phase contained
only diacid with a very small amount of monoester and no
diester. The aqueous phase was discarded after the third
extraction.
aqueous layer, from which diacid 6 was recovered as described
below.
The combined EtOAc extracts were filtered through filter
paper. Removal of solvent on a rotary evaporator at 30 °C gave
a solid, which was transferred into a dish and dried in a vacuum
oven at 30 °C overnight to give 120.2 g of monoester 5 as white
solid (AP = 99.7, 87.3% yield).
The aqueous phase was further acidified to pH 0.1 with
concentrated HCl, and diacid 6 was extracted with MTBE.
Crystallization from MTBE solution gave 2.5 g of diacid 6 as a
white solid (AP = 97).
ASSOCIATED CONTENT
* Supporting Information
■
S
NMR spectral data of compounds 1−6 and Tables 2−4
containing detailed data on the hydrolysis of diester 1 with
Lipase PS-30, Lipase PS-SD, and Novozym 435 showing the
effects of temperature, pH, substrate and enzyme input, and
organic solvents. This material is available free of charge via the
Internet at http://pubs.acs.org.
The combined EtOAc extracts were filtered through filter
paper. Removal of solvent on a rotary evaporator at 40 °C gave
a liquid (62 g), to which 500 mL of heptane was added under
stirring. The product monoester first oiled out and then
solidified. The mixture was stirred overnight at room
temperature. The mixture was filtered through filter paper.
The solid was collected, set in a hood for 1 h, and dried in a
vacuum oven at 30 °C overnight to give 41.4 g of the first crop
of monoester 5 (75.2% yield, AP = 99.3) a small amount of
diester (AP = 0.7); the diacid was not detected. The filtrate was
stirred in an ice bath for 1 h and some white solid formed. The
solid was collected by filtration through filter paper, set in a
hood for 1 h, and dried in a vacuum oven at 30 °C overnight to
give the second crop of monoester 5 (1.5 g, 2.7% yield, AP =
99.6) with a small amount of diester (AP = 0.3); the diacid was
not detected.
AUTHOR INFORMATION
Corresponding Author
*Tel.: +1 732 227 6225. Fax: +1 732 227 3994. E-mail:
animesh.goswami@bms.com.
Notes
■
The authors declare no competing financial interest.
REFERENCES
■
(1) Wilcox, C. F.; Leung, C. J. Org. Chem. 1968, 33, 877−880.
(2) Adcock, W.; Gangodawila, H. J. Org. Chem. 1989, 54, 6040−
6047.
(3) Della, E. W.; Tsanaktsidis, J. Aust. J. Chem. 1986, 39, 2061−2066.
(4) Qian, X.; Ramirez, A.; Wong, M. K. Y.; Yamamoto, K.; Fox, R.;
Guo, Z.; Goswami, A.; Patel, B.; Cann, R. O.; Deshpande, R. P.
Presented at the 240th ACS National Meeting, Boston, MA, Aug 22−
26, 2010.
PLE-Catalyzed Hydrolysis of Diester 4. To a 5 L jacketed
reactor setup with an overhead stirrer, a pH meter, and a water
circulator was added 750 mg of PLE (41 units/mg) followed by
1.5 L of sodium phosphate buffer (0.2 M, pH 7.0). The
temperature was maintained at 40 °C using a circulating water
bath. The mixture was stirred for 20 min to dissolve the
enzyme. To the enzyme solution was added 150 g of solid
diester 4 (0.88 mol). The pH of the reaction mixture was
maintained at 7.0 by periodic addition of 2 M NaOH. Samples
(50 μL) of reaction mixture were taken out, mixed with 20 μL
of 1 M HCl and 10 mL of MeOH, filtered through a 0.2 μm
filter, and analyzed by HPLC to follow the conversion. At 5.5 h,
the amounts of diester 4, monoester 5, and diacid 6 were 0.5%,
94.4%, and 5.1%, respectively. The pH of the reaction mixture
was adjusted to 7.5 with 2 M NaOH (35 mL), and then the
mixture was extracted with 1 L of EtOAc. Some emulsion
formed in the organic layer, which was filtered through a pad of
Celite to help with phase separation. The EtOAc layer
contained the unreacted diester 4 and was discarded. The
aqueous layer was acidified to pH 3.0 with 6 M HCl (160 mL)
and extracted with 1.5 L of EtOAc. The aqueous layer showed a
pH of 3.8 and was readjusted to pH 3.0 with 6 M HCl and then
extracted second time with 1.5 L of EtOAc. The aqueous layer
showed a pH of 4.0 and was again adjusted to pH 3.0 with 6 M
HCl and then extracted a third time with 1 L of EtOAc. The
aqueous layer after three extractions contained mostly the
diacid 6. HPLC showed about 3% loss of monoester 5 in the
(5) Al Hussainy, R.; Verbeek, J.; van der Born, D.; Braker, A. H.;
Leyson, J. E.; Knol, R. J.; Booij, J.; Herscheid, J. D. M. J. Med. Chem.
2010, 54, 3480−3491.
(6) Peddi, S.; Patel, M. V.; Rohde, J. J. U.S. Pat. Appl. US2010/
0267738A1, Oct 21, 2010.
(7) Bennett, B. L.; Elsner, J.; Erdman, P.; Hilgraf, R.; Lebrun, L. A.;
Mccarrick, M.; Moghaddam, M. F.; Nagy, M. A.; Norris, S.; Paisner, D.
A.; Sloss, M.; Romanow, W. J.; Satoh, Y.; Tikhe, J.; Yoon, W. H.;
Delgrado, M. PCT Int. Pat. Appl. WO2012/145569A1, Oct 26, 2012.
(8) Bornscheuer, U. T.; Kazlauskas, R. J. Hydrolases in Organic
Synthesis: Regio- and Stereoselective Biotransformations, 2nd ed.; Wiley-
VCH: Weinheim, Germany, 2006; pp 61−184.
(9) Paravidino, M.; Bohm. P.; Groger, H.; Hanefeld, U. Enzyme
̈
Catalysis in Organic Synthesis, 3rd ed.; Drauz, K., Groger, H., May, O.,
̈
Eds.; Wiley-VCH: Weinheim, Germany, 2012; pp 249−362.
(10) Sabbioni, G.; Jones, J. B. J. Org. Chem. 1987, 52, 4565−4570.
(11) Hultin, P. G.; Mueseler, F.-J.; Jones, J. B. J. Org. Chem. 1991, 56,
5375−5380.
(12) Chenevert, R.; Ngatcha, B. T.; Rose, Y. S.; Goupil, D.
Tetrahedron: Asymmetry 1998, 9, 4325−4329.
(13) Van der Eycken, J.; Vandewalle, M.; Heinemann, G.; Laumen,
K.; Schneider, M. P.; Kredel, J.; Sauer, J. J. Chem. Soc., Chem Commun.
1989, 306−308.
(14) Torres, C.; Bernabe, M.; Otero, C. Enzyme Microb. Technol.
1999, 25, 753−761.
(15) Goswami, A.; Kissick, T. P. Org. Process Res. Dev. 2009, 13, 483−
488.
(16) Kashima, Y.; Liu, J.; Takenami, S.; Niwayama, S. Tetrahedron:
Asymmetry 2002, 13, 953−956.
779
dx.doi.org/10.1021/op400254c | Org. Process Res. Dev. 2014, 18, 774−780

Organic Process Research & Development
Article
(17) Cabrera, Z.; Palomo, J. M. Tetrahedron: Asymmetry 2011, 22,
2080−2084.
(18) Wells, A. S.; Finch, G. L.; Michels, P. C.; Wong, J. W. Org.
Process Res. Dev. 2012, 16, 1986−1993.
(19) Kazlauskas, R. J.; Weissfloch, A. N. E.; Rappaport, A. T.; Cuccia,
L. A. J. Org. Chem. 1991, 56, 2656−2665.
(20) Weissfloch, A. N. E.; Kazlauskas, R. J. J. Org. Chem. 1995, 60,
6959−6969.
(21) Mezzetti, A.; Keith, C.; Kazlauskas, R. J. Tetrahedron: Asymmetry.
2003, 14, 3917−3924.
(22) Novozym 435 Product Sheet from Novozymes A/S, February
2002.
(23) Mal, D. Synth. Commun. 1986, 16, 331−335.
780
dx.doi.org/10.1021/op400254c | Org. Process Res. Dev. 2014, 18, 774−780