Scale-up of a chemo-biocatalytic route to (2 R,4 R)- and (2 S,4 S)-monatin

Article abstract of DOI:10.1021/op1001947

Monatin, a natural sweetener, refers to a collection of four isomers of 2-((1H-indol-3-yl)methyl)-4-amino-2-hydroxypentanedioic acid. A chemo-biocatalytic approach to kilogram quantities of enantiopure 2S,4S-monatin and 2R,4R-monatin from indole is described. Key steps in the process include a (2 + 3) cycloaddition reaction followed by nickel-catalysed reduction to construct the monatin backbone, and a highly selective enzyme resolution of the 2S,4S- and 2R,4R-monatin diastereomeric pair to afford each enantiomer in 99% ee.

Full text of DOI:10.1021/op1001947

Organic Process Research & Development 2011, 15, 249–257
Scale-Up of a Chemo-Biocatalytic Route to (2R,4R)- and (2S,4S)-Monatin
Amanda L. Rousseau,*,† Subash R. Buddoo, Gregory E. R. Gordon, Sharon Beemadu, B. Godfrey Kupi, M. Jerry Lepuru,
Munaka C. Maumela, Arvesh Parsoo, Duncan M. Sibiya, and Dean Brady
CSIR Biosciences, PriVate Bag X2, Modderfontein, 1645, South Africa
Abstract:
Monatin, a natural sweetener, refers to a collection of four isomers
of 2-((1H-indol-3-yl)methyl)-4-amino-2-hydroxypentanedioic acid.
A chemo-biocatalytic approach to kilogram quantities of enan-
tiopure 2S,4S-monatin and 2R,4R-monatin from indole is de-
scribed. Key steps in the process include a (2 + 3) cycloaddition
reaction followed by nickel-catalysed reduction to construct the
monatin backbone, and a highly selective enzyme resolution of
the 2S,4S- and 2R,4R-monatin diastereomeric pair to afford each
enantiomer in 99% ee.
A number of synthetic routes to monatin isomers, both
stereoselective and racemic, have been reported.5 However,
these processes have been conducted at laboratory scale and
are not readily applied to the large-scale production of monatin.
In this contribution we describe a chemo-biocatalytic approach
for the production of kilogram quantities of enantiopure 2S,4S-
and 2R,4R-monatin.6,7
Results and Discussion
Introduction
The chemo-biocatalytic approach allows access to all four
isomers of monatin by the initial preparation of racemic monatin
in three steps from indole (Scheme 1). In this route, the
diastereomeric pairs (2R,4R/2S,4S-monatin and 2R,4S/2S,4R-
monatin, hereinafter referred to as RR/SS monatin and RS/SR
monatin respectively) are separated by crystallization, and each
pair may be subjected to enzymatic resolution after esterification
and lactonization. This affords, after subsequent deprotection
steps, the corresponding individual monatin enantiomers.The
route to racemic monatin developed by the CSIR and first
described in 1992,5i was modified for scale-up to kilogram scale
and to exclude reagents such as bromine and sodium/mercury
amalgam utilized in the original process. The enzymatic
resolution was developed by Altus Biologics,7a and here we
describe modifications to the process to allow for scale-up to
kilogram scale. Although the process described in this paper
has not been fully optimized, it represents the first application
of a synthetic route to monatin isomers on this scale.
Monatin is the coined name that describes a collection of
four isomers of 2-((1H-indol-3-yl)methyl)-4-amino-2-hydroxy-
pentanedioic acid, isolable from Schlerochitin ilicifolius, a spiny-
leaved hardwood shrub indigenous to South Africa. The
isolation of the 2S,4S isomer of monatin (1) from the plant and
the elucidation of its structure were reported 20 years ago,
together with the identification of this compound as a high-
intensity, natural sweetener.1 Originally, 2S,4S-monatin was
thought to be the only isomer isolated from the extracted roots
of S. ilicifolius, but it is now known that the plant produces a
mixture of all four monatin isomers.2 Of further interest is the
fact that the sweetness potential appears dependent upon the
isomer. For example, 2R,4R-monatin (2) and its salts (e.g., Na,
NH3, K) have been shown to elicit the sweetest taste, with a
relative sweetness ranging from 1000 to 2700 times that of
sucrose.3,4 By comparison, synthetic sweeteners such as aspar-
tame exhibit relative sweetness of only 180-200 times that of
sucrose. As a result, there may be commercial interest in the
natural product, monatin, which is present in very low quantities
in the bark of the root of S. ilicifolius (1.73 g isolated from 160
kg of root).1a
Acrylate derivative (3) was prepared from ethyl acrylate by
initial Baylis-Hillman8 reaction with paraformaldehyde to
(5) (a) Tamura, O.; Shiro, T.; Toyao, A.; Ishibashi, H. Chem. Commun.
2003, 2678–2679. (b) Tamura, O.; Shiro, T.; Ogasawara, M.; Toyao,
A.; Ishibashi, H. J. Org. Chem. 2005, 70, 4569–4577. (c) Nakamura,
K.; Baker, T. J.; Goodman, M. Org. Lett. 2000, 2, 2967–2970. (d)
Oliveira, D. J.; Coelho, F. Synth. Commun. 2000, 30, 2143–2159. (e)
Oliveira, D. J.; Coelho, F. Tetrahedron Lett. 2001, 42, 6793–6796.
(f) Abushanab, E.; Arumagam, S. U.S. Patent 5,994,559, 1999; (g)
Holzapfel, C. W. Synth. Commun. 1994, 24, 3197–3211. (h) Holzapfel,
C. W. Synth. Commun. 1993, 23, 2511–2526. (i) Olivier, J.; Bischof-
berger, K. U.S. Patent 5,128,482, 1992.
* To whom correspondence should be addressed. E-mail:
Amanda.Rousseau@wits.ac.za.
† Current address: Molecular Sciences Institute, School of Chemistry,
University of the Witwatersrand, PO Wits 2050, South Africa.
(1) (a) Vleggaar, R.; Ackerman, L. G. J.; Steyn, P. S. J. Chem. Soc., Perkin
Trans. 1 1992, 3095–3098. (b) van Wyk, P. J.; Ackerman, L. G. U.S.
Patent 4,975,298, 1990.
(6) Buddoo, S. R., Rousseau, A. L., Gordon, G. E. R. US Patent
Application US20090088577A1, 2009.
(2) Bassoli, A.; Borgonovo, G.; Busnelli, G.; Morini, G.; Drew, M. G. B.
Eur. J. Org. Chem. 2005, 1652–1658.
(7) (a) Buddoo, S. R., Rousseau, A. L., Brady, D., Lalonde, J., Yao, Y.,
Wang, Y. US Patent Application US20090087888A1, 2009; (b) Brady,
D., Steenkamp, L. H., Rousseau, A. L., Buddoo, S. R., Steenkamp,
P. A. US Patent Application US20090087829A1, 2009.
(3) Bassoli, A.; Borgonovo, G.; Busnelli, G.; Morini, G.; Merlini, L. Eur.
J. Org. Chem. 2005, 2518–2525.
(4) Amino, Y.; Yuzawa, K.; Mori. K.; Takemoto, T. WO/2003/045914,
2003.
(8) Basavaiah, D.; Rao, A. J.; Satyanarayana, T. Chem. ReV. 2003, 103,
811–892.
10.1021/op1001947  2011 American Chemical Society
Published on Web 12/02/2010
Vol. 15, No. 1, 2011 / Organic Process Research & Development
•
249

Scheme 1. Overall schematic approach to monatin enantiomersa
a Reagents and conditions: a) Indole, THF, methylmagnesium chloride, 0 °C, 30 min, then (3) in THF, 5 °C, 1 h, 49% (11 runs) [Average yields reported over
multiple runs, number of runs listed in parentheses after yield given]; b) CHCl3, Et3N, 30 °C, 2 h, 68% (11 runs); c) EtOH, KOH, sponge Ni (0.5 equiv), H2, 2 bar,
21 °C, 1.5 h, 98% conversion by HPLC; d) separation of RR/SS monatin from RS/SR monatin by crystallization, 76% isolated yield (5 runs); e) methanol, HCl(g), 25
°C, 1.5 h; f) EtOAc, aq NaHCO3, CbzCl, 25 °C, 1 h; g) p-TsOH (1%), toluene, 25 °C, 16 h, 60% over three steps from RR/SS monatin, (5 runs); h) ChiroCLEC-BL,
ACN, phosphate buffer (0.3 M, pH 7.5), 23 °C, 15-20 h 98% conversion of SS-9; i) 10% m/m Pd/C (5%), MeOH, Et3N, H2 (3 bar), 40 °C, 1 h; j) aq KOH, 25 °C,
30 min. Yield of SS-monatin (1) from SS-9: 52% (5 runs). For RR-product, order of deprotection steps i) and j) are reversed. Yield of RR-monatin (2) from RR-9: 47%
(5 runs).
Scheme 2. Preparation of ethyl 2-(bromomethyl)acrylatea
a Reagents and conditions: a) paraformaldehyde (30% m/m), H2O, DABCO (0.1 equiv, 4.5% m/m), NaOH (0.40% m/m), 85 °C, 30 min; b) aq HBr (48%), H2SO4
(35% m/m), 15-20 °C, 16 h, 34% over two steps.
afford the intermediate ethyl 2-(hydroxymethyl)acrylate (4),
followed by bromination to afford the desired 2-(bromomethyl)-
acrylate (3) (Scheme 2). This process, which was found to be
low yielding on laboratory scale (50-100 g scale, 31% average
yield over the two steps), proved somewhat problematic on
scale-up. However, with modifications of the laboratory pro-
cedure, a comparable yield of product was obtained on kilogram
scale (34% average yield over the two steps).
reaction mixtures on large scale. In order to address this
problem, the loading of DABCO catalyst was adjusted to 10
mol %, and triethylamine was excluded from the large-scale
reaction. In addition, the reaction was carried out in a closed
vessel instead of under reflux conditions to obtain a higher
operating temperature of 85 °C and to allow for better utilization
of ethyl acrylate (boiling point of 99 °C) and paraformaldehyde.
No pressure buildup was observed under these conditions. Using
this modified procedure, 70% conversion of ethyl acrylate was
observed within 30 min, with minimal byproduct formation on
this scale (<10%). The downstream processing of the reaction
mixture was modified to facilitate the isolation of (4), free of
unreacted ethyl acrylate and paraformaldehyde, and polymeric
byproduct. This was essential to ensure adequate performance
in the subsequent bromination reaction, and to minimise
decomposition of the final product (3) during distillation.
Isolation of (4) was therefore achieved by adjusting the pH of
the reaction mixture to 4.5 with aqueous sulfuric acid (20%
m/m), followed by extraction of ethyl acrylate and polymeric
byproduct from the aqueous layer with hexane. The intermediate
(4) was then isolated from the aqueous phase by extraction with
1,2-dichloroethane.
The bromination reaction to afford (3) employed on labora-
tory scale involved producing a concentrated aqueous hydro-
bromic acid solution by treatment of 48% aq HBr with hydrogen
bromide gas. In the laboratory process reported in 1992,5i
hydrogen bromide gas was generated from the action of bromine
on tetralin. We modified this process to avoid the use of bromine
or hydrogen bromide gas on large scale and utilised a method
whereby treatment of aqueous hydrobromic acid (48%) with
concentrated sulfuric acid (3.25 mol equiv) afforded a concen-
trated hydrobromic acid solution which performed well in this
On laboratory scale (50-100 g), optimized reactions to
prepare (4) were carried out under reflux with reactor temper-
ature at 80 °C and in the presence of 10 mol % DABCO
catalyst. This gave 60% conversion to the desired product
without significant byproduct formation (<10%), as observed
by GC analysis. The same results were obtained with 2%
DABCO in the presence of triethylamine (7 mol %), which
maintains the pH of the reaction mixture between 8 and 9,
limiting deactivation of the DABCO catalyst. The product was
isolated by extraction into ethyl acetate, followed by evaporation
to dryness. The yield of (4) in the large-scale Baylis-Hillman
reaction under these conditions (80 °C, 2 mol % DABCO, 7
mol % triethylamine) was found to be lower than that obtained
on the laboratory scale, with a significantly higher level of
polymeric byproduct formed in the process (approximately
30%). The down stream processing of the reaction mixture also
proved to be problematic, with poor phase separation during
the extraction of (4) into ethyl acetate, and loss of product to
the aqueous phase occurring as a result of the poor separation.
This in turn had a detrimental effect on the subsequent formation
of (3), produced in an overall yield of only 10% from ethyl
acrylate. Upon investigation, triethylamine was shown to
facilitate byproduct formation under the prolonged periods of
heating associated with the heating and cooling profiles of
250
•
Vol. 15, No. 1, 2011 / Organic Process Research & Development

reaction.9 On laboratory scale, an average overall yield of 31%
of (3) was achieved from ethyl acrylate over two steps. The
bromination of (4) to afford (3) on large scale was carried out
using this modified process. The product was extracted into
hexane, washed with aqueous sodium bicarbonate, and purified
by distillation (0.8 mmHg, 38-40 °C, reboiler temperature <80
°C) to afford a comparable yield of product on 50 kg scale
(34% average isolated yield over the two steps). To avoid
decomposition of the final product, the distillation was carried
out at <1 mmHg.
Formation of ethyl 2-((1H-indol-3-yl)methyl)acrylate (5) was
achieved by reaction of ethyl 2-(bromomethyl)acrylate (3) and
indolylmagnesium chloride, which was generated from indole
and methylmagnesium chloride in THF. In addition to the
formation of the desired product (5), other complex indole/
acrylate adducts were produced during the reaction and identi-
fied as sequential addition products resulting from reaction at
the indole nitrogen. A number of screening experiments were
therefore conducted on laboratory scale in order to determine
the effect of variables such as residual water present in the
solvent, order of reagent addition, temperature, stoichiometry,
and concentration of reagents on the reaction performance at
bench scale (15 L).
At this scale, the reaction performed equally well with
solvent that had been predried using sodium metal, and with
solvent used directly from the supplier10 without pretreatment.
The order of reagent addition also had no effect on the yield of
product and did not result in a decrease in byproduct formation.
When acrylate (3) was used as the limiting reagent in the
reaction, a decrease in byproduct formation was observed as
expected; however, this was accompanied by a concomitant
decrease in both indole conversion and product yield. Although
removal of unreacted indole from the crude reaction mixture
could be achieved by steam distillation, the reduction in product
yield and prospect of an additional process step prompted us
to use conditions which would allow for maximum indole
conversion (mole ratio indole/acrylate/Grignard ) 1.0:1.08:1.1).
The reaction was not adversely affected when carried out at
higher concentrations of indole and acrylate (3) (31% m/m and
44% m/m, respectively, compared with 18% m/m and 29% m/m
respectively), allowing for smaller reaction volumes. The
reaction was sensitive to temperature, with increased temper-
atures resulting in increased levels of byproduct formed.
Optimum yields of acrylate (5) on kilogram scale were
therefore obtained when a solution of indolylmagnesium
chloride in THF, generated at 0-10 °C over a period of 2 h,
was treated with a solution of 2-(bromomethyl)acrylate (3) in
THF, added over a period of 2 h at 0-5 °C, and left for a further
1 h at this temperature. On a laboratory scale, the product was
purified by silica gel column chromatography, and alternative
methods of purification were sought for scale-up. Although
conventional distillation proved to be successful on a laboratory
scale, this led to significant decomposition during extended
heating times and loss in yield at bench scale. We assessed an
alternative short path distillation of the mixture which was found
to minimise the degradation that had been observed using the
Scheme 3. Preparation of ethyl
2-chloro-2-(hydroxyimino)acetatea
a Reagents and conditions: a) 32% aq HCl, HCl(g), NaNO2 (2.0 equiv, in
H2O, 69% m/m), -5-0 °C, 4 h; (40% average yield over 10 runs).
conventional distillation process. Although this afforded the
acrylate (5) in lower purity (70-80% m/m by HPLC), losses
due to degradation were minimised, and the short path distil-
lation was chosen as the purification method of choice for this
intermediate. The acrylate (5) was isolated in an average yield
of 49% by this process, which was somewhat lower than that
obtained on laboratory scale (50 g, 61% yield) where column
chromatography was utilized for purification of the product. This
may be attributed to losses associated with the short path
distillation.
The (2 + 3) cycloaddition step to afford the isoxazoline
diester (6) from indolyl acrylate (5) and carbethoxyformonitrile
oxide (7) in the presence of triethylamine proved to be robust.5g,i
The reaction performance was not adversely affected when
indolyl acrylate (5) of lower purity (70-80%) was utilised as
the substrate. Lengthy chromatography of the indolyl acrylate
(5) was therefore avoided and the product isolated from short
path distillation was carried through to the cycloaddition step.
On a 2 kg scale, the cycloaddition reaction afforded isoxazoline
diester (6) in an average yield of 68%. The product was isolated
by crystallization, and recrystallized when necessary to achieve
the desired level of purity (>98%) for the subsequent hydro-
genation step.
The nitrile oxide (7) is formed in situ from ethyl 2-chloro-
2-(hydroxyimino)acetate (8) during the cycloaddition reaction.
The synthesis of (8), starting from glycine ethyl ester hydro-
chloride, was carried out at 0-5 °C with strict temperature
control on 5-10 kg scale (Scheme 3).11 The product was
isolated in an average yield of 40%, which was comparable
with the yields obtained on laboratory scale (200 g).12 The
byproduct formed during this reaction, ethyl chloroacetate, is a
known sensitizer, and continued exposure to this chemical
results in skin rashes and shortness of breath. The symptoms
disappear once contact with the chemical is eliminated. The
chemical has a low vapour pressure and even though protective
equipment is used the risk of exposure is very high. Strict
containment measures were put in place in order to minimise
exposure to ethyl chloroacetate during the preparation of (8).
Handling of the material was minimised, and all work was
carried out in large-scale fume hoods in a single work area with
access control. Effective PPE (including respirators) and a
designated wash-up area further reduced the level of exposure
to this sensitizer.
(11) Kozikowski, P.; Adamczyk, M. J. Org. Chem. 1983, 48, 366–372.
Ethyl 2-chloro-2-(hydroxyimino)acetate (8) is commercially available
in bulk from a number of suppliers; however, costs associated with
the purchase and shipping of (8) (US $400-700/kg) prompted us to
synthesise the material in-house.
(12) The yield of (8) on 150 g scale or less was found to be slightly higher,
at an average of 56%. However, all reactions carried out at a scale of
200-500 g afforded lower yields of (8), with an average yield of 43%
obtained on this scale.
(9) In Organic Syntheses, 2nd ed.; Gillman, H., Blatt, A. H., Eds., John
Wiley and Sons, Inc.: London, 1958; Collective Vol. 1.
(10) Merck, AR grade, containing 0.03% (m/m) water.
Vol. 15, No. 1, 2011 / Organic Process Research & Development
•
251

Scheme 4. Preparation of monatin derivatives for enzymatic resolutiona
a Reagents and conditions: a) methanol, HCl(g), 25 °C, 1.5 h; b) EtOAc, aq NaHCO3, CbzCl, 25 °C, 1 h; c) p-TsOH (1%), toluene, 25 °C, 16 h, 60% over three
steps from RR/SS monatin (5 runs).
Scheme 5. Enzymatic resolution of RR/SS monatin derivativesa
a Reagents and conditions: a) ChiroCLEC-BL, MeCN, phosphate buffer (0.3 M, pH 7.5) 22 °C, 15-20 h, 98% conversion of SS-9.
Nickel-catalysed hydrogenation of the isoxazoline diester (6)
was carried out on a 5 kg scale, and the two pairs of monatin
diastereomers were separated by crystallization. RR/SS monatin
is only partially water-soluble and was isolated by precipitation
from aqueous acetic acid to afford, after subsequent recrystal-
lization, RR/SS monatin in 76% yield at >98% m/m purity by
quantitative HPLC. By comparison, RS/SR monatin is highly
water-soluble, and was isolated by precipitation from ethanol.
As our interest lay in the RR/SS diastereomeric pair, the RS/SR
monatin formed during this process was retained, but not
processed further.
The N-protected monatin lactone (9) required for the
enzymatic resolution was prepared in three steps as shown in
Scheme 4. RR/SS monatin was treated with methanol saturated
with hydrogen chloride gas to afford a mixture of the dimethyl
ester of monatin (10) as well as the monatin lactone methyl
ester (11). On laboratory scale, (10) and (11) were formed in
an approximate ratio of 75:25 by HPLC analysis of the reaction
mixture, and complete conversion to the lactone methyl ester
(11) was facilitated by removal of methanol on a rotary
evaporator. This process was not feasible on >500 g scale, with
all attempts to facilitate complete conversion to (11) resulting
in either loss in yield or decomposition of the product, and an
alternative approach was adopted. The crude mixture of
dimethyl ester (10) and lactone (11) was therefore subjected to
N protection with benzylchloroformate under basic conditions
in a biphasic system. After complete reaction, the product
mixture was extracted into ethyl acetate and concentrated under
reduced pressure. The N-protected lactone/dimethyl ester mix-
ture was treated with catalytic p-toluenesulfonic acid (1%) in
toluene at room temperature, and the desired product precipi-
tated from the toluene reaction matrix affording pure product
in solid form. Using this procedure, RR/SS-9 was obtained in
an average yield of 60% from RR/SS-monatin on a 1.5 kg scale.
The enzymatic resolution (shown in Scheme 5) was carried
out on 1 kg of substrate (9) at 22 °C for 19 h and stirrer speeds
of 150-200 rpm, and the reaction progress was monitored by
HPLC. The enzyme used in this transformation, a protease
(subtilisin) from Bacillus lichenformis, was immobilized by the
CLEC method and designated ChiroCLEC-BL.13 This biocata-
lyst had previously demonstrated to be enantioselective for this
reaction on a laboratory scale.14
The resolution reaction was found to be highly selective over
temperatures ranging from 17 to 25 °C, with RR-9 left entirely
unreacted with only 2-5% residual SS-9. A drop in selectivity
was observed at higher temperatures (35-40 °C) on laboratory
scale with 10-15% of RR-9 consumed in the reaction before
50% conversion of RR/SS-9 had been reached. The enzyme was
recovered by filtration and could be recycled after washing with
a mixture of phosphate buffer and acetonitrile. On laboratory
scale, the enzyme was recycled twice with no significant loss
in selectivity or activity. After removal of the enzyme by
filtration, the unreacted RR-9 was isolated from the resolution
reaction mixture by extraction into dichloromethane. The
(13) (a) Margolin, A. L. Trends Biotechnol. 1996, 14, 223–230. (b) St.
Clair, N. L.; Navia, M. A. J. Am. Chem. Soc. 1992, 114, 7314–7316.
(14) Personal communication with J. Lalonde, Y. Yao, and Y. Wang of
Altus Biologics Inc., who selected the enzyme for this resolution, see
ref 7a. The ChiroCLEC-BL was supplied by Altus Biologics from a
single batch, and the resolution was run five times on this scale.
Although ChiroCLEC-BL is no longer commercially available on this
scale, the nonimmobilized form of the enzyme, alcalase, which is
commercially available, performed equally well in this transformation
on laboratory scale. For another example of the use of the ChiroCLEC-
BL in synthesis see: Ema, T.; Jittani, M.; Furuie, K.; Utaka, M.; Sakai,
T. J. Org. Chem. 2002, 67, 2144–2151.
252
•
Vol. 15, No. 1, 2011 / Organic Process Research & Development

Scheme 6. Deprotection sequence to afford SS-monatin (1)a
a Reagents and conditions: a) 10% m/m Pd/C (5%), MeOH, Et3N, H2 (3 bar), 40 °C, 1 h; b) aq KOH, 25 °C, 30 min, then AcOH.
Scheme 7. Deprotection sequence to afford RR-monatin (2)a
a Reagents and conditions: a) RR-9 in DCM, aq KOH, 25 °C, 30 min, then AcOH; b) 10% m/m Pd/C (5%), MeOH, Et3N, H2 (3 bar), 40 °C, 1 h.
dichloromethane extract was used without further processing
in the deprotection sequence.
the product was free of potassium acetate.15 This afforded SS-
monatin isolated as the monopotassium salt in an average yield
of 56% from SS-12 at >98% m/m purity (quantitative HPLC),
99.5% de and 99% ee.
Unfortunately, this deprotection sequence could not be
applied to RR-9 as lactam formation occurs readily during the
removal of the Cbz protecting group. The dichloromethane
extract from the resolution reaction containing RR-9 was
therefore treated with aqueous potassium hydroxide to facilitate
hydrolysis. The biphasic system limited epimerization at the
4-position during ester hydrolysis (Scheme 7).
The aqueous phase containing the intermediate RR-diacid
(14) was then treated with methanol and acetic acid to afford a
homogeneous solution, followed by catalytic reduction with
Pd/C in a hydrogen atmosphere to facilitate complete removal
of the Cbz protecting group. The reaction mixture was filtered
from catalyst and the filtrate concentrated to afford a mixture
of RR-monatin and potassium acetate. This mixture was treated
as described for the SS-product to afford RR-monatin as the
monopotassium salt in an average yield of 60% from RR-9 at
>99% m/m purity (quantitative HPLC), 97% de and 99% ee.
Isolation of the resolution product (12) was facilitated by
addition of dichloromethane to the remaining aqueous phase,
followed by acidification to pH 2.0 with 10% aq HCl. This
ensured the formation of a well-dispersed precipitate, allowing
for recovery and isolation of SS-12 by filtration.
Hydrolysis of the ester groups and removal of the N-Cbz
protecting group from the products isolated from the enzymatic
resolution affords the individual RR- and SS-monatin enanti-
omers. In initial studies on laboratory scale, epimerization at
the 4-position was observed under the basic conditions required
for ester and/or lactone hydrolysis of RR-9 and SS-12, owing
to the acidity of the proton R to the carbonyl group of the
lactone. This effect was exacerbated on larger scales (>250 g),
where the hydrolysis reaction was much slower and increased
epimerization (>15%) was observed.
In an attempt to minimize epimerization, hydrogenation of
SS-12 was carried out first, followed by hydrolysis as shown
in Scheme 6. Efficient mixing in the hydrogenation reaction is
essential to allow for optimal interaction between the gas
(hydrogen), solid (catalyst), and liquid phases (substrate in
solution). As SS-12 was not sufficiently soluble in methanol,
the solvent used for the hydrogenation reaction, the slurry that
formed was treated with triethylamine (0.2 equiv) to facilitate
dissolution of SS-12. Palladium supported on carbon was then
added and the mixture stirred in a hydrogen atmosphere for
3 h, after which time HPLC analysis indicated complete
reaction. The presence of triethylamine at this level was not
found to adversely affect the performance of the hydrogenation
reaction. The intermediate product (13) precipitates out of the
reaction matrix; thus, removal of the catalyst by filtration was
not feasible. The reaction mixture was therefore treated with
aqueous potassium hydroxide and left to stir at room temper-
ature for 1 h before removal of the catalyst by filtration. The
filtrate was acidified to pH 4.0 with glacial acetic acid, and the
mixture was concentrated to afford a mixture of SS-monatin
(1) and potassium acetate. This mixture was treated with ethanol
to dissolve the potassium acetate, and the remaining precipitate
was filtered by Nutsche pressure filtration and dried in a vacuum
oven. This process was repeated until analysis indicated that
Conclusions
We have prepared kilogram quantities of both 2S,4S-monatin
and 2R,4R-monatin in 99% ee using a chemo-biocatalytic
approach. No other synthetic route published has afforded
monatin enantiomers on this scale, and isolation of kilogram
quantities of monatin from the plant source is not feasible owing
to the low concentrations present in S. ilicifolius and the
scarceness of the plant. Key steps in the synthesis which proved
to be robust on scale-up were the (2 + 3) cycloaddition reaction
and the highly selective enzymatic resolution. This route
potentially allows access to all four monatin isomers in sizable
quantities.
(15) 1H NMR spectroscopy was used to monitor approximate levels of
potassium acetate with each subsequent ethanol wash by the presence
of a peak at δ 2.1 ppm. As the products (1) and (2) are isolated as the
monopotassium salt, the accurate potassium level (% m/m) was
determined by atomic absorption once the peak in the 1H NMR
spectrum was no longer present.
Vol. 15, No. 1, 2011 / Organic Process Research & Development
•
253

analysed for residual paraformaldehyde16 by adding a 2 g sample
of the extract to 100 mL of a 125 g/L solution of sodium sulfite.
Three drops of phenolphthalein were added and the mixture
titrated to a colourless end point with 0.5 N sulfuric acid. If no
paraformaldehyde was present, the solvent was removed at
40-60 °C under reduced pressure. If paraformaldehyde was
present, the organic extract was washed again with water, and
the titration repeated. The resultant residue containing ethyl
2-(hydroxymethyl)acrylate (27.0 kg, contaminated with DCE)
was analysed for moisture content.
Table 1
time (min)
% eluent A
% eluent B
0.00
30.0
35.0
45.0
90
10
90
90
10
90
10
10
Eluent A: 3.5 g of phosphoric acid diluted with 900 mL of nano pure water, pH
adjusted to 3 by addition of triethylamine. Volume made up to 1000 mL with nano
pure water, mobile phase then degassed and filtered. Eluent B: 100% acetonitrile.
The flow rate was set at 1.0 mL/min and UV detection at 280 nm.
Ethyl 2-(hydroxymethyl)acrylate (27.0 kg, 207 mol) was
charged to a 400 L glass lined reactor (ex Pfaudler), equipped
with a propeller type impeller (100 rpm). Aqueous HBr (48%,
3.25 equiv, 112 kg, 672 mol) was added, and the mixture was
cooled to 12 °C. Sulfuric acid (2.0 equiv, 39.0 kg, 398 mol)
was added at such a rate that the internal temperature did not
exceed 15 °C to avoid the formation of byproduct resulting from
polymerisation. After complete addition of the sulfuric acid,
the mixture was left to stir at room temperature (recorded as
varying between 22-26 °C) overnight. The reaction mixture
was extracted with hexane (3 × 33 kg), and the combined
hexane extracts were washed twice with a 5% aqueous sodium
bicarbonate solution (2 × 33 kg). The washed hexane extract
was then transferred to the distillation column and the solvent
removed at 40 °C under reduced pressure. The temperature was
increased to 50 °C to remove any water that might be present.
The residue containing ethyl 2-(bromomethyl)acrylate 3 was
then distilled (0.8 mmHg, 38-40 °C, reboiler temperature <80
°C to avoid decomposition of the final product) to afford the
desired product (16.4 kg, 34% yield) in >95% purity by GC
(retention time, GC method A, 4.62 min). 1H NMR (200 MHz;
CDCl3) δ 1.33 (t, J ) 7.2 Hz, 3H), 4.19 (s, 2H), 4.28 (q, J )
7.2 Hz, 2H), 5.95 (s, 1H), and 6.33 (s, 1H).
Ethyl 2-((1H-Indol-3-yl)methyl)acrylate (5). A solution of
indole (2.00 kg, 17.1 mol) in THF (3.88 kg) was charged to a
GR-15 glass reactor (ex Buchi) which was equipped with a
propeller-type impeller (100-300 rpm) and cooled to 0 °C
under an atmosphere of nitrogen. A solution of methylmagne-
sium chloride (1.1 equiv, 3.0 M in THF, 6.63 kg, 18.8 mol)
was added over 105 min, and with the slight exotherm, the
reactor temperature was maintained between 5 and 10 °C, and
then stirred at 150 rpm for a further 30 min under these
conditions. A solution of (3) (1.08 equiv, 3.54 kg, 18.4 mol) in
THF (3.57 kg) was then added over 2 h, while maintaining the
reaction temperature below 5 °C, and then stirred for an
additional 1 h at this temperature. The reaction mixture was
then quenched with water (7.5 kg) and heated to 20 °C in order
to dissolve salts. The aqueous and organic phases were
separated, and the aqueous phase was extracted twice with ethyl
acetate (10 kg). The ethyl acetate/THF phases were combined
and concentrated under reduced pressure at 40-90 °C. The
organic residue was passed through a short path distillation unit
(KD-5 SPD ex UIC) to facilitate removal of high-boiling
components at a feed temperature of 90 °C, and vacuum of
0.02 mmHg. The column temperature was maintained at
170-190 °C, and a flow rate of 140-300 mL/h ensured
Experimental Section
General. NMR spectra were run on either a Varian 200
MHz Gemini 2000 instrument, or on a 400 MHz Varian
INOVA instrument. A 20 L Buchi rotary evaporator was used
to carry out solvent recovery at bench scale, while solvent
recovery at larger scale was carried out using a 40 L glass batch
distillation apparatus. In all cases the term “reduced pressure”
refers to a pressure of 2 kPa.
GC Method A. DB-Wax 5 column. Program: initial
temperature 80 °C, initial time 2 min, rate 20 °C/min, final time
0 min, final temperature 240 °C.
HPLC Method A. Luna 5 µ C18 (2), 150 mm × 4.6 mm
I.D, at 25 °C. A linear gradient elution system was set up as
shown in Table 1.
HPLC Method B. Chiracel OD, 250 mm × 4.6 mm I.D at
25 °C, eluting with hexane/EtOH/TFA (70:30:0.1, prepared by
mixing 700 mL of hexane, 300 mL of ethanol, and 1 mL of
trifluoroacetic acid in a 1000 mL volumetric flack, degassing,
and filtering). Flow rate set at 1 mL/min, 220 nm UV detection.
HPLC Method C. Crownpak CR (+) AD, 15 cm × 4.0
mm at 35 °C, eluting with a 15% methanol in aq HClO4 solution
at pH1.5 (prepared by diluting 16.3 g of commercially available
70% perchloric acid with nano pure water to 1000 mL to give
a perchloric acid solution of pH 1. Further dilution of pH 1
perchloric acid solution (316 mL) with 1000 mL of nano pure
water affords a perchloric acid solution of pH 1.5. Eluent
prepared by mixing 850 mL of perchloric acid solution (pH
1.5) with 150 mL of methanol in a 1000 mL volumetric,
degassing, and filtering), flow rate set at 1 mL/min, 220 nm
UV detection.
Ethyl 2-(Bromomethyl)acrylate (3). Paraformaldehyde
(18.8 kg, 625 mol) dissolved in water (62.5 kg) was treated
with a 10% aqueous sodium hydroxide solution (250 g, 6.50
mol) and heated to 80 °C in a 100 L stainless steel reaction
vessel (with a pressure rating of 6 bar) equipped with a Rushton
turbine (150 rpm). To this preheated mixture were added
DABCO (0.10 equiv, 2.8 kg, 25 mol) and ethyl acrylate (25.0
kg, 250 mol). The contents were heated to 85 °C with stirring,
and the reaction progress was monitored by GC analysis
(retention time of product, GC method A, 3.93 min). After
complete conversion of starting material, the reaction mixture
was cooled to 25 °C and treated with aqueous sulfuric acid (4.4
kg of a 20% solution) to adjust the pH of the mixture to 4.5.
The reaction mixture was extracted with hexane (2 × 16.5 kg)
to facilitate removal of ethyl acrylate and some byproduct. The
remaining aqueous phase was then extracted with 1,2-dichlo-
roethane (2 × 20 kg). The combined organic extracts were
(16) Borgstrom, P.; Horsch, W. G. J. Am. Chem. Soc. 1923, 45, 1493–
1497.
254
•
Vol. 15, No. 1, 2011 / Organic Process Research & Development

recovery of product at >70% m/m purity (boiling point of (5)
47-52 °C at 0.02 mmHg). The distillate (2.38 kg, 76.8% m/m)
contained 1.83 kg of (5) (47% yield) as determined by
quantitative HPLC (HPLC Method A, retention time 17.06
min). A pure sample of (5), obtained by silica gel column
chromatography (20:80 ethyl acetate/hexane), was analysed by
1H and 13C NMR spectroscopy. 1H NMR (200 MHz, CDCl3)
δ 1.35 (t, J ) 7.2 Hz, 3H), 3.85 (s, 2H), 4.29 (q, J ) 7.2 Hz,
2H), 5.55 (s, 1H), 6.27 (s, 1H), 7.03 (d, J ) 2.6 Hz, 1H),
7.03-7.29 (m, 2H), 7.38 (dt, J ) 8.0 and 1.1 Hz, 1H),
7.59-7.63 (m, 1H), and 8.16 (br s, 1H). 13C NMR (50 MHz,
CDCl3) δ 14.15, 27.56, 60.64, 111.09, 112.72, 118.98, 119.21,
121.82, 122.78, 125.16, 127.25, 136.30, 139.78, and 167.37.
Ethyl 2-Chloro-2-(hydroxyimino)acetate (8). Glycine ethyl
ester hydrochloride (6.30 kg, 45.0 mol) was dissolved in
hydrochloric acid (32% m/v, 5.13 kg) in a glass-lined CR-26
reactor (ex Buchi), equipped with a retreat blade impeller
(100-300 rpm). The solution was cooled to 0 °C and hydrogen
chloride gas was bubbled into the reactor Via a sparging tube
for 1 h while maintaining the temperature below 5 °C. The
reaction mixture was then cooled to -5 °C, and a solution of
sodium nitrite (1.0 equiv, 3.11 kg, 45.0 mol) in water (4.5 kg)
was dosed into the reactor using a diaphragm pump, while
maintaining the temperature between -5 and 0 °C. This
operation was complete within an hour. A second portion of
gaseous HCl was sparged into the reactor over 1 h at 0-5 °C.
This was followed by a second portion of sodium nitrite (1.0
equiv, 3.11 kg, 45.0 mol) in water (4.5 kg) which was added
over one hour while maintaining the temperature between -5
and 0 °C. The reaction mixture was stirred for an additional
1 h and was then extracted with chloroform (3 × 2 kg) at
ambient temperature. Isolation of the product from the reaction
mixture by crystallisation resulted in significant losses in yield
on this scale. The organic extracts were therefore combined,
dried (MgSO4), and evaporated in Vacuo (40 °C). An oily
residue remained which crystallised upon cooling to ambient
temperature. The crystals were filtered, washed with hexane (2
kg) at 5 °C, and dried under vacuum to afford the product (8)
(2.53 kg, 36.9% yield) as a white solid. 1H NMR (200 MHz,
CDCl3) δ 1.34 (t, J ) 7.2 Hz, 3H), 4.36 (q, J ) 7.2 Hz, 2H),
and 10.06 (br s, 1H). 13C NMR (50 MHz, CDCl3) δ 13.89,
63.85, 132.77, and 158.67.
ethanol/water (3.2 kg, 80:20 v/v) to facilitate precipitation of
the product out of solution. The precipitate was filtered to afford
1.80 kg (67% yield) of desired product (6) at >98% m/m purity
1
by HPLC (HPLC Method A, retention time 14.90 min). H
NMR (400 MHz, CDCl3) δ 1.25 (t, J ) 7.4 Hz, 3H), 1.30, (t,
J ) 7.2 Hz, 3H), 3.20 (d, J ) 18.4 Hz, 1H), 3.43 (d, J ) 15.6
Hz, 1H), 3.51 (q, J ) 15.6 Hz, 1H), 3.61 (d, J ) 18.4 Hz, 1H),
4.19-4.29 (m, 4H), 7.12-7.20 (m, 3H), 7.36 (d, J ) 7.6 Hz,
1H), 7.61 (d, J ) 7.6 Hz, 1H), and 8.20 (br s, 1H). 13C NMR
(100 MHz, CDCl3) δ 13.95, 31.22, 40.36, 62.13, 62.33, 91.87,
107.86, 111.15, 118.61, 119.82, 122.15, 124.17, 127.79, 135.72,
151.55, 159.97, and 170.83.
2S,4S/2R,4R-Monatin Diastereomeric Pair. Potassium
hydroxide (4.0 equiv, 3.01 kg, 53.6 mol) was dissolved in
ethanol (37 kg) in a 75-L glass-lined reactor (ex Pfaudler),
equipped with a propeller-type impeller (100 rpm), and cooled
to 21 °C. To this was added isoxazoline diester (6) (4.60 kg,
13.4 mol) with stirring. Upon complete dissolution of the
substrate, sponge nickel catalyst (4.60 kg, A-7063 from AMC,
approximately 50% wet) was added to the reactor with stirring
(215 rpm) to minimise settling. The reactor was sealed and
purged with nitrogen before hydrogen was fed to the reactor to
a total pressure of 2 bar (g). The reaction was run at this pressure
for 30 min with the hydrogen supply line open. The conversion
after 30 min was 72% by HPLC analysis. Hydrogen was then
used to further pressurise the reactor to 3 bar (g) and the system
maintained for 25 min, after which 97% conversion of starting
material was obtained. The system was then pressurised to 5
bar (g) and run for 30 min to ensure complete conversion of
the isoxazoline diester. The incremental increase in pressure
was implemented to contain the exotherm of the hydrogenation
reaction, which resulted in temperatures >25 °C. After complete
reaction, the reactor was vented and flushed with nitrogen (1
bar (g)) for 5 min, prior to being pressurised to 1.5 bar (g) with
nitrogen. The material was filtered through a Nutsche filter with
nitrogen at this pressure. The catalyst bed was not filtered to
dryness owing to the pyrophoric nature of the catalyst. The
filtrate (46.1 kg) was then concentrated in a 40 L vessel
equipped with a steam coil, condenser, and receiving vessel.
The final mass of the concentrate was 10.8 kg. The concentrate
was treated with water in a 16 L glass-lined, stirred reactor at
20-25 °C and acidified to a pH of 4.5 with glacial acetic acid.
The RR/SS diastereomeric pair of monatin isomers precipitated
from this medium over a period of 12-15 h at 20 °C. The
slurry was drained and centrifuged to isolate the precipitate,
which is a 90:10 mixture of RR/SS-monatin and RS/SR-monatin.
To enrich this ratio, the mixture was dissolved in aqueous
ammonium hydroxide solution at pH 8.5 and then acidified with
glacial acetic acid to pH 4.5 to facilitate precipitation. The
reactor was again cooled to approximately 20 °C and maintained
for 2 h, after which the slurry was drained and centrifuged.
The procedure for the second precipitation was repeated twice
to obtain 1.54 kg of RR/SS monatin (79% of theoretical yield)
of purity >98% m/m by quantitative HPLC, and >96% de.
(HPLC Method A, retention time of RR/SS diastereomeric pair,
5.24 min). 1H NMR (400 MHz, D2O + NaOD) δ 1.51 (dd, J
) 14.4 and 10.8 Hz, 1H), 2.15 (dd, J ) 14.4 and 2.0 Hz, 1H),
2.81 (d, J ) 14.6 Hz, 1H), 2.99 (d, J ) 14.6 Hz, 1H), 3.07 (dd,
Diethyl 5-((1H-Indol-3-yl)methyl)-4,5-dihydroisoxazole-
3,5-dicarboxylate (6). Ethyl 2-((1H-indol-3-yl)methyl)acrylate
5 from short path distillation (2.38 kg, 76.8% m/m, containing
1.83 kg, 7.98 mol of 5) was dissolved in chloroform (4.3 kg),
and triethylamine (2.0 equiv, 1.62 kg, 16.0 mol) was added to
this solution with stirring in a glass-lined CR-26 reactor (ex
Buchi), equipped with a retreat blade impeller (100-300 rpm).
A solution of 2-chloro-2-(hydroxyimino)acetate (8) (1.8 equiv,
2.18 kg, 14.4 mol) in chloroform (8.7 kg) was added slowly to
the reaction mixture over a period of 45-70 min, with the
reaction temperature maintained below 30 °C. The reaction
progress was monitored by HPLC. After complete conversion
of starting material, the reaction mixture was quenched with
water (13 kg), and the phases were separated. The organic phase
was washed twice more with water (2 × 13 kg) and concen-
trated under vacuum, and the crude residue was treated with
Vol. 15, No. 1, 2011 / Organic Process Research & Development
•
255

J ) 10.8 and 2.0 Hz, 1H), 6.88-6.98 (m, 3H), 7.23 (d, J )
8.0 Hz, 1H) and 7.48 (d, J ) 7.6 Hz, 1H). 13C NMR (100 MHz,
D2O + NaOD) δ 35.25, 42.69, 54.68, 80.71, 109.49, 111.36,
118.78, 119.06, 121.28, 124.49, 127.78, 135.52, 181.38, and
182.30. RS/SR monatin can be isolated from the mother liquor
of the first precipitation by first concentrating this mixture under
reduced pressure, followed by addition of ethanol to the
concentrate. RS/SR monatin precipitates from this matrix, and
can be recrystallized from ethanol/water to afford RS/SR
monatin in >99% de (HPLC Method A, retention time of RS/
SR diastereomeric pair, 4.96 min). 1H NMR (400 MHz, D2O)
δ 1.99 (dd, J ) 15.0 and 9.7 Hz, 1H), 2.27 (d, J ) 15.0 Hz,
1H), 3.02 (s, 2H), 3.78 (d, J ) 9.7 Hz, 1H), 6.92-7.03 (m,
3H), 7.28 (d, J ) 8.0 Hz, 1H) and 7.53 (d, J ) 8.4 Hz, 1H).
13C NMR (100 MHz, D2O) δ 33.58, 38.65, 51.78, 78.01,
109.03, 111.42, 118.86, 119.03, 121.40, 124.56, 127.62, 135.58,
174.56, and 180.82.
(2R,4R)-Methyl 2-((1H-Indol-3-yl)methyl)-4-(benzyloxy-
carbonylamino)-5-oxotetrahydrofuran-2-carboxylate RR-9
and (2S,4S)-2-((1H-Indol-3-yl)methyl)-4-(benzyloxycarbo-
nylamino)-5-oxotetrahydrofuran-2-carboxylic Acid SS-12.
RR/SS-Lactone (9) (1.0 kg, 2.34 mol) was dissolved in MeCN
(6.0 kg) and added to phosphate buffer (0.3 M, pH 7.5, 24 L)
in a glass-lined reactor (CR-26 ex Buchi) fitted with an anchor
stirrer. A suspension of the enzyme (706 mL, ChiroCLEC-BL,
batch lot BLSD00414) was added to the reactor, and the resulting
mixture stirred at 22 °C for 19 h, with a stirrer speed of 155
rpm. After this time, approximately 98% of the SS-lactone had
been consumed as observed by HPLC analysis. DCM (6.1 kg)
was added to allow for ease of filtration (Whatman No.1,
Nutsche pressure filter pressurised with nitrogen) to recover the
enzyme, and the phases were separated. The aqueous phase was
extracted twice with DCM (2 × 5 kg), and the combined
organic phases, containing RR-9 (0.48 kg by quantitative HPLC
analysis, 96% of theoretical yield, containing 2% residual SS-
9) were concentrated to a volume of approximately 2.5 L under
reduced pressure. This concentrate was used without further
processing in the subsequent deprotection sequence to afford
RR-monatin (2). The aqueous phase from the resolution reaction
was treated with DCM (5.3 kg) and a 10% aq HCl solution
(2.88 kg) to adjust the pH to 2.0. SS-12 precipitates from this
mixture, and was isolated by filtration, which gave, after drying,
0.5 kg of the desired product as a white solid in 99% theoretical
yield, 100% m/m SS-12 by HPLC (HPLC Method B, retention
time of RR-9 11.47 min, retention time of SS-12 6.45 min).1H
NMR (400 MHz, CDCl3) δ 2.45 (dd, J ) 13.2 and 9.6 Hz,
1H), 2.76 (dd, J ) 13.2 and 9.6 Hz, 1H), 3.37 (d, J ) 14.8 Hz,
1H), 3.45 (d, J ) 18.8 Hz, 1H), 3.61 (dt, J ) 9.6 and 8.0 Hz,
1H), 5.02 (s, 2H), 6.25 (d, J ) 8.0 Hz, 1H), 7.08-7.17 (m,
3H), 7.27-7.39 (m, 7H), 7.64 (d, J ) 7.6 Hz, 1H), and 9.67
(br s, 1H). 13C NMR (100 MHz, CDCl3) δ 32.59, 35.93, 50.37,
66.78, 84.10, 107.10, 111.57, 118.59, 119.58, 121.85, 124.89,
127.61, 127.95, 128.04, 128.41, 136.09, 136.19, 155.82, 172.97,
and 174.04.
2S,4S-Monatin (1). SS-12 (440 g, 1.08 mol) was treated
with methanol (4.2 kg) in an 8 L Parr reactor, and triethylamine
(6.7 g, 0.067 mol) was added to facilitate dissolution, with
warming to 40 °C. The solution was added to a pressure vessel
and treated with 10% Pd/C (44 g of 50% wet paste). The
pressure vessel was sealed, evacuated, and purged with nitrogen
twice. The reactor was then pressurised to 3 bar with hydrogen,
and the resulting suspension was stirred (500 rpm) at 40 °C.
The reaction progress was monitored by HPLC. After 1 h,
complete consumption of starting material had been achieved;
the reactor was depressurised, purged with nitrogen gas, and
cooled to 25 °C. A solution of KOH (4.0 equiv, 242 g, 4.31
mol) in water (2.2 kg) was added to the reaction mixture with
stirring while maintaining the temperature at 25 °C. The reaction
progress was monitored by HPLC, and after 1 h the reaction
was complete. The catalyst was removed by filtration and the
reaction mixture acidified to pH 6.5 with glacial acetic acid.
The reaction mixture was concentrated under reduced pressure
to a volume of 600 mL, and the product (1) precipitated from
this solution. The precipitate isolated was washed with ethanol
(3 × 1 L) until free of potassium acetate and dried to afford
Methyl 2-((1H-Indol-3-yl)methyl)-4-(benzyloxycarbonyl-
amino)-5-oxotetrahydrofuran-2-carboxylate (9). A slurry of
RR/SS-monatin (1.47 kg, 5.03 mol) in MeOH (13 L) in a glass
lined CR-26 reactor (ex Buchi), equipped with a retreat blade
impeller (100-300 rpm) was saturated with HCl gas while
maintaining the temperature between 35 °C - 45 °C. The
reaction mixture was then stirred at room temperature for 2 h
after which HPLC analysis showed complete consumption of
starting material and the formation of two products: dimethyl
ester (10) and lactone methyl ester (11) (75:25 by HPLC
analysis). Solvent was removed under reduced pressure at 35
°C - 40 °C to afford an oil that was used without further
purification. The crude mixture of monatin esters (10) and (11)
was dissolved in EtOAc (5 L) and added to a saturated aq
NaHCO3 solution (7 L). The reaction mixture was cooled to 5
°C - 10 °C, and benzyl chloroformate (approximately 1.0 equiv,
720 mL) was added to the reactor with stirring over a period
of 30 min. The reaction mixture was stirred for an additional
1 h, after which HPLC analysis showed complete consumption
of starting material. The phases were separated and the aqueous
phase was extracted with EtOAc (5 L). The combined organic
phases were washed with water and dried (MgSO4). Removal
of solvent (8 L) under reduced pressure afforded a brown oil
which was dissolved in toluene (6 L), treated with p-TsOH (1
mol %, 9.6 g, 0.050 mol) and the reaction mixture stirred at
room temperature for 15 h. The white precipitate that formed
was filtered and dried to afford methyl 2-((1H-indol-3-yl)-
methyl)-4-(benzyloxycarbonylamino)-5-oxotetrahydrofuran-2-
carboxylate (9) (1.28 kg, 60% yield, 98% m/m purity by
HPLC). (HPLC Method B, retention time of RR-CBz lactone
1
11.47 min, retention time of SS-CBz lactone 8.88 min). H
NMR (400 MHz, CDCl3) δ 2.44 (dd, J ) 13.2 and 10.0 Hz,
1H), 2.84 (dd, J ) 13.2 and 10.0 Hz, 1H), 3.37 (d, J ) 15.0
Hz, 1H), 3.46 (d, J ) 15.0 Hz, 1H), 3.48-3.51 (m, 1H), 3.83
(s, 3H), 5.02 (s, 2H), 5.08 (br s, 1H), 7.14-7.38 (m, 9H), 7.64
(d, J ) 7.2 Hz, 1H) and 8.27 (br s, 1H). 13C NMR (100 MHz,
CDCl3) δ 32.60, 36.37, 50.41, 53.21, 67.30, 84.42 110.05,
111.44, 118.61, 120.32, 122.64, 124.61, 127.39, 128.11, 128.31,
128.55, 135.74, 135.88, 155.52, 171.49, and 173.46.
256
•
Vol. 15, No. 1, 2011 / Organic Process Research & Development

193 g of 2S,4S-monatin (1); 54% yield, 98% m/m purity by
HPLC, 99.5% de and 99% ee (HPLC Method C, retention time
of 1 26.44 min).1H NMR (400 MHz, D2O) δ 1.81 (dd, J )
15.2 and 12.0 Hz, 1H), 2.43 (dd, J ) 15.2 and 1.6 Hz, 1H),
2.84 (d, J ) 14.4 Hz, 1H), 3.05 (d, J ) 14.4 Hz, 1H), 3.38 (dd,
J ) 12.0 and 1.6 Hz, 1H), 6.89-6.99 (m, 3H), 7.24 (d, J )
8.0 Hz, 1H), and 7.48 (d, J ) 7.6 Hz, 1H). 13C NMR (100
MHz, D2O) δ 35.21, 38.04, 53.60, 80.14, 108.99, 111.57,
119.04, 119.20, 121.54, 124.86, 127.84, 135.77, 174.13, and
179.93.
2R,4R-Monatin (2). The solution of RR-9 in dichlo-
romethane (approx 2.5 L from the enzyme resolution extraction
containing 0.48 kg, 1.14 mol RR-9 by quantitative HPLC
analysis) was treated with a solution of KOH (4.0 equiv, 264 g,
4.72 mol) in water (2.4 kg) with stirring, while maintaining the
temperature at 25 °C. The reaction progress was monitored by
HPLC, and after 1 h the reaction was complete. After this time
the phases were separated, and the pH of the aqueous phase
was adjusted to 7.0 using glacial acetic acid (approximately 2
equiv, 140 g). Methanol (2.5 kg) was then added to the solution
and the resulting mixture transferred to an 8 L Parr reactor.
The mixture was treated with 10% Pd/C (48 g of 50% wet
paste), and the pressure vessel sealed, evacuated, and purged
with nitrogen twice, and then pressurised to 3 bar with hydrogen.
The resulting suspension was stirred at 25 °C and reaction
progress monitored by HPLC at half-hourly intervals. After 1 h,
all of the starting material had been converted to product, and
the reactor was depressurised and purged with nitrogen gas.
The catalyst was removed by filtration and the pH of the
solution adjusted to 6.5 with glacial acetic acid. The resulting
mixture was concentrated to a volume of approximately 600
mL and the product allowed to precipitate from this mixture.
The precipitate isolated was washed with ethanol (3 × 1 L) to
facilitate removal of potassium acetate and dried to afford 240 g
of 2R,4R-monatin (2); 64% yield, 99% m/m purity by HPLC,
97% de and 99% ee (HPLC Method C, retention time of 2
13.32 min).1H NMR (400 MHz, D2O) δ 1.85 (dd, J ) 15.2
and 12.0 Hz, 1H), 2.47 (dd, J ) 15.2 and 2.0 Hz, 1H), 2.87 (d,
J ) 14.4 Hz, 1H), 3.08 (d, J ) 14.4 Hz, 1H), 3.42 (dd, J )
12.0 and 2.0 Hz, 1H), 6.92-7.04 (m, 3H), 7.28 (d, J ) 8.4 Hz,
1H), and 7.52 (d, J ) 7.6 Hz, 1H). 13C NMR (100 MHz, D2O)
δ 35.12, 37.95, 53.47, 80.06, 108.80, 111.47, 118.95, 119.08,
121.45, 124.76, 127.70, 135.65, 174.05, and 179.82.
Acknowledgment
We thank Altus Biologics Inc. for developing the enzymatic
resolution on laboratory scale and for supplying the Chiro-
CLEC-BL for this process. We are grateful to V. R. Mnisi, C.
Kgaje and M. Zulu for providing analytical support. T. B.
Chokwe, G. Lethwane, S. Mahlatjie, W. Mabotja, A. Kekae,
and P. Khoza are thanked for their assistance during the pilot-
plant scale-up.
Received for review July 16, 2010.
OP1001947
Vol. 15, No. 1, 2011 / Organic Process Research & Development
•
257