Large-scale synthesis of a substituted d -phenylalanine using asymmetric hydrogenation

Article abstract of DOI:10.1021/op200129m

A synthetic route to an N-BOC d-phenylalanine pharmaceutical intermediate suitable for rapid scale-up to 150-kg scale was required. A seven-step route based on asymmetric hydrogenation of an N-acetyl dehydroamino-acid was developed. Starting with terephthalic dialdehyde, monoreduction of one aldehyde group, Erlenmeyer condensation, and ring-opening/O-deacetylation with methanol provided the 4-(hydroxymethyl)-substituted dehydrophenylalanine hydrogenation substrate. Asymmetric hydrogenation of this enamide using [((R,R)-Ethyl-DuPhos) Rh(COD)BF₄ proceeded in high enantiomeric excess. Subsequently, the cis-2,6-piperidyl group was introduced by mesylation/displacement, the BOC group was introduced, and acetyl and methyl ester groups were removed by basic hydrolysis. This route was used to manufacture 150 kg of the BOC amino acid 1.

Full text of DOI:10.1021/op200129m

ARTICLE
pubs.acs.org/OPRD
Large-Scale Synthesis of a Substituted D-Phenylalanine Using
Asymmetric Hydrogenation
Martin E. Fox,*,† Mark Jackson,† Graham Meek,† and Matthew Willets‡
†Chirotech Technology Centre, Dr. Reddy’s Laboratories EU Limited, Unit 410 Cambridge Science Park, Milton Road, Cambridge CB4
0PE, U.K.
‡Chirotech Technology Ltd., Dr. Reddy’s Laboratories EU Limited, Steanard Lane, Mirfield, West Yorkshire WF14 8HZ, U.K.
S Supporting Information
b
ABSTRACT: A synthetic route to an N-BOC D-phenylalanine pharmaceutical intermediate suitable for rapid scale-up to 150-kg
scale was required. A seven-step route based on asymmetric hydrogenation of an N-acetyl dehydroamino-acid was developed.
Starting with terephthalic dialdehyde, monoreduction of one aldehyde group, Erlenmeyer condensation, and ring-opening/O-
deacetylation with methanol provided the 4-(hydroxymethyl)-substituted dehydrophenylalanine hydrogenation substrate. Asym-
metric hydrogenation of this enamide using [((R,R)-Ethyl-DuPhos)Rh(COD)]BF4 proceeded in high enantiomeric excess.
Subsequently, the cis-2,6-piperidyl group was introduced by mesylation/displacement, the BOC group was introduced, and acetyl
and methyl ester groups were removed by basic hydrolysis. This route was used to manufacture 150 kg of the BOC amino acid 1.
’ INTRODUCTION
An unnatural D-phenylalanine intermediate in the form of the
N-BOC acid 1 was required as part of a development program for
a novel therapeutic agent (see Figure 1). A rapid scale-up to
150 kg was required, and thus, it was necessary to use an approach
to amino acid synthesis that we were confident from the outset
could be reliably used on this scale. Therefore, we were attracted
to asymmetric hydrogenation, which we expected could be used
with relatively little development and for which suitably active
Figure 1. Unnatural D-phenylalanine derivative 1.
and selective catalyst systems were readily available on the scale
required, as an approach to this amino acid derivative.1,2
the chiral pool routes were not attractive compared to asymmetric
synthesis, and the resolution-based routes were likely to be more
wasteful and expensive, while being no more easily scaled up than
asymmetric synthesis using asymmetric hydrogenation.
Potential dehydroamino acid hydrogenation substrates are
shown in Scheme 1. The piperidyl group could be introduced at a
late stage by coupling to a (4-(hydroxymethyl)phenyl)alanine
derivative 2, hence requiring a 4-hydroxymethyl or protected
4-hydroxymethyl substituted 2-acetamidocinnamate substrate 3.
Alternatively, the piperidyl group could be introduced before the
asymmetric hydrogenation, employing the more advanced sub-
strate 4. Placing the asymmetric hydrogenation step early in the
synthesis would minimise the issue of removal of residual metal
from the product. In addition, we anticipated that coupling of the
piperidyl group to 2 should be a relatively clean, high-yielding
reaction, and that synthesis, isolation, and purification of the
piperidyl-containing substrate 4 could be more difficult than the
simple hydroxymethyl enamide 3. Therefore, we chose to carry
out the asymmetric hydrogenation of a hydroxymethyl enamide
3 and introduce the piperidyl group at a later stage. The key
starting material for synthesis of the enamide by condensation
chemistry by either approach was 4-(hydroxymethyl)benzaldehyde
5. Syntheses of single isomer (4-(hydroxymethyl)phenyl)alanines
2 had previously been achieved from tyrosine by cyanation3 or
carbonylation,4 resolution using an amino acid acylase,5 from
4-(hydroxymethyl)phenyl hydantoin using a hydantoinase6 and
asymmetric phase-transfer catalysed alkylation.7 As a D-amino-acid,
’ RESULTS AND DISCUSSION
The route selection work carried out on laboratory scale is
described in three sections covering substrate synthesis, the
asymmetric hydrogenation step and downstream chemistry
and is followed by an account of the development on 100-g
scale and implementation of the chosen route in the pilot plant.
Synthesis of Enamides. 4-(Hydroxymethyl)benzaldehyde 5
was prepared by sodium borohydride reduction of readily
available, inexpensive terephthalic dialdehyde (Scheme 2).9 In
alcoholic solvents, a complex mixture of products was obtained,
but in THFÀwater, a roughly 70:30 mixture of 5 and diol 6 with
around 1% remaining dialdehyde was obtained in a clean
reaction. The product mixture was isolated by aqueous workup
Special Issue: Asymmetric Synthesis on Large Scale 2011
Received: May 13, 2011
Published: August 05, 2011
r
2011 American Chemical Society
1163
dx.doi.org/10.1021/op200129m Org. Process Res. Dev. 2011, 15, 1163–1171
|

Organic Process Research & Development
ARTICLE
Scheme 1. Retrosynthetic analysis
Scheme 2. Synthesis of enamides
and extraction with toluene, followed by precipitation of a
mixture of aldehyde 5 and diol 6 by addition of hexane. The
aldehyde 5 could be obtained nearly free of diol 6 by suspension
of the diol/aldehyde mixture in toluene and filtration of the
insoluble diol 6.
The most inexpensive method for synthesis of enamides from
aromatic aldehdyes is the Erlenmeyer condensation reaction,10
which has the advantage of providing these compounds exclu-
sively as the (Z)-geometric isomer. When 5 was subjected to
standard Erlenmeyer condensation conditions (N-acetylglycine,
acetic anhydride, sodium acetate, ethyl acetate, reflux), both
acetylation of the hydroxyl group and condensation took place to
give azlactone 9 in 57% yield. Better yields of around 64% in 96%
purity by GC were obtained if the reaction was carried out stepwise.
Thus, esterification of hydroxyaldehyde 5 to acetate 7 with acetic
anhydride was carried out first, then acetylglycine was added
to the reaction. Purification of the aldehyde 5 was undesirable
on large scale, and unnecessary; the crude approximately 3:1
hydroxyaldehyde 5/diol 6 mixture obtained in the sodium
borohydride reduction could be used directly in the Erlenmeyer
condensation. Noncrystalline diacetate 8 arising from diol 6 was
removed in the crystallization of azlactone 9, which crystallized
from the reaction on cooling to about 60 °C. After addition of
water, further azlactone 9 was precipitated, after which this was
isolated by filtration. On a 50-g scale, hydrolysis of the azlactone
9 to acid 3d was not a significant problem. Although the overall
yield from dialdehyde was modest at around 30%, this route
provided rapid assembly of the carbon framework of the amino
1164
dx.doi.org/10.1021/op200129m |Org. Process Res. Dev. 2011, 15, 1163–1171

Organic Process Research & Development
ARTICLE
Scheme 3. Introduction of piperidyl group to enamide 3b
Scheme 4. Asymmetric hydrogenation step.a
Table 1. Asymmetric hydrogenation of 3a with
[(Ligand)Rh(COD)]BF4.a
entry
ligand
conversion %b
ee (%)c
1
(R,R)-Me-DuPhos
(R,R)-Et-DuPhos
(S,S)-iPr-DuPhos
(R,R)-Me-FerroTANE
(R,R)-Et-FerroTANE
(R,R)-Me-BPE
>98
>98
>98
>98
>98
>98
>98
>98
>98
>98
>98
>98
0
91.3 (R)
96.2 (R)
93.6 (R)
85.9 (R)
80.1 (R)
71.6 (R)
98.2 (S)
42.9 (S)
4.0 (R)
2
3
4
5
6
7
(R,R)-Ph-BPE
a Reagents and conditions: see Tables 1 and 2.
8
(S)-PhanePhos
(S)-Tol-BINAP
(R)-(S)-JosiPhos
(S,S,R,R)-TangPhos
(R,R)-DIPAMP
(S,S)-BPPM
9
acid from inexpensive, readily available materials, and the chem-
istry appeared to be readily scalable. Enamides 3 differing in their
protection at the alcohol and carboxylate functionalies could
be obtained by employing different conditions in opening of
9. Thus, by reaction with methanol under neutral conditions
or by direct addition of methanol to the Erlenmeyer con-
densation reaction before product isolation, azlactone 9 was
opened in situ to methyl ester 3a. Although, this had the
attraction of allowing isolation of a more hydrolytically stable
material than azlactone 9, this reaction was found to be un-
reliable and often to give incomplete conversion; thus, isola-
tion of azlactone 9 was preferred for scale-up. Treatment of 9
with methanol and 2 mol % potassium carbonate at ambient
temperature both opened the azlactone and cleaved the O-acetyl
group to give the hydroxymethyl enamide 3b. By treatment with
methanol and stoichiometric sodium hydroxide, acid 3c was
obtained.
Introduction of the piperidyl group prior to hydrogenation by
conversion of enamide 3b to 4a was a further possibility outlined
in Scheme 1. Under the conditions described below for coupling
of the piperidyl group to the post-asymmetric hydrogenation
compound 2b, reaction of 3b was possible (Scheme 3) but lower
yielding, and the mesylate intermediate appeared to be less
stable. Therefore, we were satisfied that, of the possible orders
of steps shown in Scheme 1, our choice of carrying out the
asymmetric hydrogenation before introduction of the piperidyl
group was preferable.
10
11
12
13
14
15
16
15.3 (R)
69.3 (R)
74.2 (S)
À
23.3 (R)
28.3 (R)
92.5 (R)
(R)-MonoPhos2
(R,R)-Me-5-Fc
>98
>98
>98
(S,S)-tBuP*
a Reaction conditions 2 mmol substrate, s/c 250:1, MeOH, 25 °C, 7 bar
c
H2. b Determined by 1H NMR. Determined by Chiral HPLC, condi-
tions see Experimental.
Table 2. Asymmetric hydrogenation of substrates 3 with
[(DuPhos)Rh(COD)]BF4.a,b
reaction
catalystc
loading
time
ee
entry substrate
ligand
(min)
(%)
1
2
3
4
5
6
7
3a
3a
3b
3b
3b
3b
3c
(R,R)-Et-DuPhos
(R,R)-Et-DuPhos
(R,R)-Me-DuPhos
(R,R)-Me-DuPhos
(R,R)-Et-DuPhos
(R,R)-Et-DuPhos
(R,R)-Et-DuPhos
1000:1
2500:1
1000:1
2500:1
1000:1
2500:1
250:1
6
20
5
97.6
97.8
97.6
95.1
99.7
99.6
99.0
35
12
45
15
a Reaction conditions: 2 mmol substrate, s/c 1000À2500:1, MeOH,
Selection of Substrate and Catalyst. An initial catalyst
screen was carried out with 3a using a range of cationic rhodium
complexes of commercially available ligands at a relatively high
catalyst loading (Scheme 4, Table 1) using an Argonaut En-
deavor catalyst screening system. The highest enantioselectivity
was achieved with Ph-BPE,11 but high enantioselectivities
(>90%) were also achieved with Me-, Et-, and iPr-DuPhos.
However, Me- and Et-DuPhos were available at lower cost than
Ph-BPE and iPrDuPhos, and at the time, more readily available
on the scale required. Therefore, catalysts derived from these
40 °C, 7 bar H2. b Full conversion achieved for all entries
1
c
(determined by H NMR). Estimated from mass of substrate and
catalyst used.
ligands were preferred for scale-up and provided sufficiently high
enantioselectivities, and sufficiently low catalyst loadings could
be achieved.
Further small-scale studies with all three substrates were
carried out with the Me- and Et-DuPhos rhodium catalysts to
1165
dx.doi.org/10.1021/op200129m |Org. Process Res. Dev. 2011, 15, 1163–1171

Organic Process Research & Development
ARTICLE
Scheme 5. Downstream steps after asymmetric hydrogenation
determine the most favorable substrate/ligand combination for
scale-up (Table 2). The results were inconsistent, because the
effect of substrate purity on both reactivity and selectivity was not
initially recognised. Higher enantioselectivities than in the initial
screen were achieved with all three substrates. The acetoxy-
protected substrate 3a was more reactive than the hydroxy
substrate 3b, but higher enantiomeric excess was achieved with
3b. With the hydroxy substrate 3b, higher reactivity but lower
selectivity was achieved with Me- than with Et-DuPhos. Full
conversion was not achieved with either 3a or 3b at molar s/c >
2500:1. High selectivity was also achieved with hydroxy acid 3c,
but this required a higher catalyst loading than the other
substrates to achieve full conversion. When less pure samples
of 3a and 3b from pilot-plant batches were used, selectivities and
reaction rates were much lower. This dependence of enantio-
meric excess on substrate purity could also account for the
differences seen between the selectivities found in the initial
catalyst screen and these studies. The lower reactivity of 3c can
also be ascribed to the greater difficulty in isolation and purifica-
tion of this compound. Subsequently, in the asymmetric hydro-
genation of 3c with [((R,R)-Ethyl-DuPhos)Rh(COD)]BF4 on a
larger (10 g) scale using substrate purified by crystal digestion, a
catalyst loading of 5000:1 was achieved.
While all three substrates could potentially be used in the
synthesis of 1, the shortest overall synthetic scheme (Scheme 6)
employed hydroxyester 3b. In addition, the hydrogenation
product 2b crystallized readily, which was advantageous for
isolation on large scale at this stage of the synthesis. With the
highest enantioselectivity and highest rate with this substrate
being achieved with different catalysts, compromise was required
in selecting a catalyst for scale-up; however, with the need for
rapid scale-up, the higher selectivity with this substrate achieved
with Et-DuPhos was the most important criterion, so hydro-
genation of 3b with [((R,R)-Ethyl-DuPhos)Rh(COD)]BF4 was
selected for scale-up.
Reductive amination using either palladium on carbon and
hydrogen or sodium triacetoxyborohydride was sluggish, and
aldehyde reduction predominated; thus, we favored the alcohol
activation/alkylation approach. Alkylation of cis-2,6-dimethylpi-
peridine with mesylate 10 occurred under mild conditions.
Activation was carried out by mesylation of 2b in dichloro-
methane with methanesulfonic anhydride, after which in situ
reaction with cis-2,6-dimethylpiperidine to give N-acetyl piper-
idinyl amino acid methyl ester 11 proceeded cleanly at room
temperature. Methanesulfonyl anhydride rather than methane-
sulfonyl chloride was used to avoid formation of the unreactive
chloride, and the hindered base, ethyldiisopropylamine, was used
to avoid quaternisation of the base. The additional cost and lead-
time of methanesulfonic anhydride were offset by the higher
degree of confidence in scale-up with this reagent. The non-
crystalline 11 could be purified to remove nonbasic impurities by
extraction into mild aqueous acid such as citric acid, basification,
and re-extraction into dichloromethane. After introduction of the
piperidyl group, completion of the synthesis required adjustment
of the amino and carboxylate protecting groups. The N-BOC
group was introduced with di-tert-butyl dicarbonate and catalytic
4-(dimethylamino)pyridine in THF. Finally, both the N-acetyl
group and methyl ester of imide 1213 were cleaved with lithium
hydroxide, which could be achieved without racemisation pro-
vided the base was added slowly to control both the reaction
exotherm and pH. While the final product 1 is an amino acid, it
was sufficiently lipophilic to extract the neutral form into organic
solvents from weakly acidic solutions. Thus, isolation was
achieved by acidification to pH close to 5.5 and extraction into
dichloromethane. However, the presence of residual inorganics
made crystallization difficult, and it was necessary to remove
these by an ion-exchange resin treatment in aqueous solution.
Subsequent removal of water by distillation, addition of ethanol,
removal of ethanol by distillation, and final crystallization from
ethyl acetateÀMTBE gave 112 in an acceptable crystalline form.
The final crystallization did not provide a sufficient upgrade of
enantiomeric excess, so high enantiomeric excess in the asym-
metric hydrogenation and careful control of the temperature and
addition of base in the final deprotection reaction were essential
to obtaining 1 within the specification of >98.5% enantiomeric
excess, >98% purity. The overall yield from terephthalic dialde-
hyde was 15% in seven steps, but the nonselective monoreduction
Downstream Steps to Amino Acid 1. The remaining steps of
the synthesis from ((4-hydroxymethyl)phenyl)alanine 2b are
shown in Scheme 5. Introduction of the benzylic cis-2,6-di-
methylpiperidyl group could be achieved by oxidation of the
alcohol to the aldehyde and reductive amination12 or by activa-
tion as a leaving group and N-alkylation. We found the hindered
cis-2,6-dimethylpiperidine to be a generally unreactive amine.
1166
dx.doi.org/10.1021/op200129m |Org. Process Res. Dev. 2011, 15, 1163–1171

Organic Process Research & Development
ARTICLE
Scheme 6. Overall route used on 100 g to 30 kg scale
in the first step was responsible for a large proportion of the loss in
yield, which was about 45% from azlactone 9. The chemistry
appeared to be scalable, and we were confident of the ability of this
route to deliver the product on the 100 kg+ scale required for the
pilot-plant campaigns.
1.5 equiv of acetic anhydride, after which the solvent was removed
by distillation and the crude aldehyde (∼3:1 7/8) was drained
from the reactor. This allowed the solid reagents for the Erlen-
meyer condensation to be charged to an empty reactor. After
charging the ethyl acetate solvent and 2 equiv of acetic anhydride
and heating the reaction, aldehyde 7 was charged, thus avoiding
any potential runaway exotherm. Difficulty was encountered with
both solidification of the reaction on cool-down and hydrolysis of
azlactone 9 to 3d if water was added at too high a temperature.
Slow cooling to 20À25 °C followed by addition of water at this
temperature minimized both these problems. Isolation of azlac-
tone 9 at this stage gave sufficiently pure material to allow the
azlactone opening and asymmetric hydrogenation to be tele-
scoped while maintaining a catalyst loading of molar s/c
1500À2500:1. The azlactone opening was carried out using
sodium carbonate rather than potassium carbonate, and the
methanolic solution of enamide 3b was carried forward and used
directly in the asymmetric hydrogenation without isolation. It was
later recognized that isolation and purification of this intermediate
could be desirable to permit a lower catalyst loading to be used in
the asymmetric hydrogenation reaction. After the asymmetric
hydrogenation, MTBE antisolvent was added, and crystalline 2b
was isolated in 80% yield from azlactone 9.
Laboratory Scale-Up to 100 g. Scale-up from the small-scale
initial studies was carried out in three stages, laboratory synthesis
of a 100-g sample, followed by 30- and 150-kg campaigns
(Scheme 6). The 100-g sample preparation allowed development
in readiness for transfer to the pilot plant. In the reduction of
terephthalic dialdehyde to aldehyde 5 with sodium borohydride,
purification by precipitation of diol 6 was found to be unnecessary.
After workup of the reaction and distillation of most of the solvent
from the toluene/THF solution, the ∼3:1 aldehyde 5/diol 6
mixtureremained liquidand could beused directly inthe nextstep,
provided the solution was kept hot at about 70 °C. The method of
conversion of 4-(hydroxymethyl)benzaldehyde 5 to azlactone 9
used on small scale had been by reaction of the aldehyde with the
full 2.5 equiv of acetic anhydride required for both O-acetylation
and Erlenmeyer condensation, then addition of solid N-acetylgly-
cine to the reaction. This was not possible on the plant due to the
need for addition of a solid reagent to the reaction. Therefore, this
procedure was modified, and the acetylation was carried out with
1167
dx.doi.org/10.1021/op200129m |Org. Process Res. Dev. 2011, 15, 1163–1171

Organic Process Research & Development
ARTICLE
In the coupling with cis-2,6-dimethylpiperidine, the piperidine
was found to be suitable as both base and reagent, and the
coupling procedure could be simplified by carrying out the
reaction by addition of a 4-fold excess of cis-2,6-dimethylpiper-
idine to a solution of 2b and 1.6 equiv of methanesulfonic
anhydride in dichloromethane. While methanesulfonic anhy-
dride and 2,6-dimethylpiperidine do react, this is slower than
the alcohol mesylation reaction due to the hindered nature of 2,6-
dimethylpiperidine. However, this change had unforeseen con-
sequences as described below. Purification of 11 by extraction
into aqueous citric acid, basification, and re-extraction into
dichloromethane to remove nonbasic impurities such as cis-2,6-
dimethylpiperidine methanesulfonamide arising from the cou-
pling step was retained. No significant changes to the reactions in
the final stages of the synthesis were made, other than a reduction
in the relative quantities of solvents required and the previous
yields were maintained. A minor change to the final isolation was
required. After extraction of the product 1 into dichloromethane,
rather than evaporation to dryness, the solvent exchange back to
water was carried out by addition of water, after which resin was
charged. Finally, the dichloromethane was removed by distilla-
tion. The final drying of the amino acid 1 had previously proved
problematic. The product readily formed stable solvates with
common solvents which required heating to nearly 100 °C (close
to the melting point of the compound) to break down; this
tended to change the form of the product from a crystalline solid
to a foam. Careful control of the drying by carrying it out slowly at
70 °C eliminated this problem. Removal of ethanol was particu-
larly difficult, and a more careful choice of solvent for the final
isolation would be desirable.
Scale-Up to Pilot Plant. The process modified as described
for the 100 g sample preparation was used for the 30 kg pilot
campaign. For steps as far as the asymmetric hydrogenation, little
further modification was required. Two 30 kg asymmetric
hydrogenation reactions were carried out in this campaign. A
molar catalyst loading of s/c 3500:1 was used. The crystallization
of 2b was carried out using toluene, rather than MTBE as the
antisolvent. In the coupling of cis-2,6-dimethylpiperidine to 2b,
difficulty was encountered in achieving high conversion to 11. In
a separate observation, which proved to be connected, final
analysis of the product 1 showed one previously minor impurity
to be present at a much higher level of 0.52%, and at close to 10%
in the filtrates of the final crystallization. LC/MS analysis showed
a molecular weight of 364, lower than the mass of 1 by 14,
corresponding to a difference of one methyl group. A known
impurity in the cis-2,6-dimethylpiperidine was 2-methylpiperi-
dine, present at a level of 0.25%. Therefore, the corresponding
2-methylpiperidyl compound 13 (Figure 2) was a reasonable
candidate for this impurity. The much higher level of this
impurity in 1 compared to the level of 2-methylpiperidine in
the cis-2,6-dimethylpiperidine can be explained by the combina-
tion of use of a large excess of cis-2,6-dimethylpiperidine in the
coupling with 2b and higher reactivity of the less hindered
impurity, resulting in concentration of the impurity in the
product. The need for a large excess was partly due to reaction
between methanesulfonic anhydride and cis-2,6-dimethylpiper-
idine consuming a substantial quantity of both reagents when the
reagents were charged to the reaction in the manner described
above. While using a lower stoichiometry of 2,6-dimethylpiper-
idine and an external base (as in the initial route selection work)
could offer a solution to this process-related impurity issue,
insufficient time was available to investigate this option.
Figure 2. Process-related impurity 13.
It was recognized that the BOC protection step and final
deprotection and isolation, which required multiple solvent
exchanges, could be streamlined, but insufficient time was
available to develop a more concise alternative procedure.
For the 150 kg campaign, difficulties were encountered when
moving to a larger scale for the first two steps. The conversion in
the sodium borohydride reduction was lower due to inefficient
mixing in the reactor used. Variable levels of hydrolysis of the
azlactone 9 during isolation was again apparent. A short-term
remedy available was to dehydrate partially hydrolysed material
with acetic anhydride. These difficulties led to a lower yield over
the first two steps to about 19% from terephthalic dialdehyde.
The asymmetric hydrogenation reaction was not scaled up
further due to the limit placed by the size of the largest available
reactor, and material was processed in eight batches. The
mesylation/piperidine introduction was carried out in two
batches. The use of cis-2,6-dimethylpiperidine as both base and
reagent was maintained in this campaign, but a change was made
to the method of addition of the reagents. Instead of addition of
the full quantity of methanesulfonic anhydride at the beginning
of the reaction followed by the cis-2,6-dimethylpiperidine, the
methanesulfonic anhydride was added in four separate portions,
and after each, one-quarter of the 2,6-dimethylpiperidine re-
quired was added in four aliquots. This led to an improvement in
the conversion in this reaction due to a reduction in the side
reaction of the cis-2,6-dimethylpiperidine and methanesulfonic
anhydride, and an improvement in the purity of the final product.
With this improvement in purity, the final treatment with resin
was found to be unnecessary, and the product crystallized
directly. The BOC protection and final deprotection and isola-
tion was carried out as a single batch. The overall yield from
azlactone 9 was 85%. The difficulties encountered on the 150-kg
campaign were largely due to the need for rapid scale-up, and
with a relatively small amount of development these difficulties
could be reduced or eliminated. The key issues are the following:
achieving better control over the product quality in the tereptha-
lic dialdehyde reduction step, achieving a higher and consistent
yield in the azlactone formation, for which the former would be
critical, better control of the purity of enamide 3b, allowing a
consistently lower catalyst loading to be achieved in the asym-
metric hydrogenation step, a more robust procedure for the
piperidine introduction step in which the number of equivalents
of 2,6-dimethylpiperidine is reduced, for example by employing
an external base, and a more streamlined isolation process for the
final product 1 with careful attention to choice of solvent,
especially with regard to the difficulty of drying this material.
Overall, the seven-step asymmetric hydrogenation-based route
was well suited to rapid scale-up for the production of amino acid
derivative (1). The asymmetric hydrogenation step was readily
integrated into the overall synthetic route, and the material was
produced on time and in specification using this chemistry.
1168
dx.doi.org/10.1021/op200129m |Org. Process Res. Dev. 2011, 15, 1163–1171

Organic Process Research & Development
ARTICLE
’ EXPERIMENTAL SECTION
for 5 h to provide a clear solution. A portion of this solution
(50 mL, ∼5.60 g) was concentrated in vacuo. Methanol (11 mL)
and toluene (11 mL) were added, and the suspension was heated
to 80 °C to give a clear solution, which was slowly cooled to room
temperature before placing in an iceÀwater bath. Filtration
followed by washing with methanolÀtoluene (1:1, 5 mL) pro-
vided enamide 3a as a very pale, yellow solid (3.60 g, 64%). 1H
NMR (400 MHz, CDCl3): δ 7.45 (d, J = 7 Hz, 2H), 7.38À7.34
(m, 3H), 7.12 (br s, 1H), 5.10 (s, 2H), 3.85 (s, 3H), 2.13 (s, 3H),
and 2.11 (s, 3H). 13C NMR (100 MHz, CDCl3) 171.2, 169.6,
166.1, 137.5, 134.0, 132.1, 130.2, 128.5, 124.9, 66.1, 53.1, 23.7,
and 21.3.
General. Melting points were recorded using a Perkin-Elmer
DSC6 digital scanning calorimeter. Optical rotations were de-
termined using a Perkin-Elmer 341 polarimeter in units of 10À1
deg cm2 gÀ1 (c in g/100 mL). Proton and 13C NMR spectra were
recorded on a Bruker AV400 spectrometer using Me4Si or
residual CHCl3 as an internal reference. Mass spectra were
recorded using a ThermoFinnigan LCQ Advantage using APCI.
GC methods were run using a Perkin-Elmer Autosytem XL gas
chromatograph. HPLC methods were run using a Gilson mod-
ular system or a TSP modular system.
Preparation of 4-(Hydroxymethyl)benzaldehyde 5.9 To a
2-L three-neck flask, fitted with an overhead agitator, thermo-
meter and condenser, terephthalaldehyde (85.8 g, 0.640 mol)
and THF (500 mL) were charged to give a pale yellow solution.
Sodium borohydride (8.75 g, 0.231 mol) was dissolved in water
(35 mL), and the resultant solution was charged to the three neck
flask over 20 min maintaining the internal temperature below
30 °C. The reaction mixture was stirred for a further 2 h; reaction
was complete by GC (dialdehyde 0.5%, monoaldehyde 5 58.9%,
diol 6 36.0%). Water (500 mL) and toluene (500 mL) were
charged with stirring, and then the layers were separated. The
aqueous layer was further extracted with toluene, and then the
combined organics were backwashed with water. The organic
layer was distilled at ambient pressure until the internal tem-
perature reached 105 °C; approximately 600 mL of distillates
were collected. The mixture was cooled to below 30 °C, and then
hexane (500 mL) was charged before cooling to 0À5 °C in order
to precipitate the product from solution. The solid product was
isolated by vacuum filtration, and the solid cake was washed with
a 50/50 mixture of toluene and hexane (200 mL total) and then
dried at 40À45 °C under vacuum to give 69.0 g of 5 (∼53%).
The solid melted in the vacuum oven. The product purity by GC
was 66.2% with 27.9% of the diol 6 present. 1H NMR (400 MHz,
CDCl3): δ 10.00 (s, 1H), 7.88 (d, J = 8.0 Hz, 2H), 7.53 (d, J = 8.0
Hz, 2H), 4.80 (s, 2H), and 2.1 (br s, 1H). 13C NMR (100 MHz,
CDCl3): δ 192.7, 148.4, 135.9, 130.4, 127.0, and 64.8.
Preparation of Methyl (Z)-2-(Acetylamino)-3-[4-(hydroxy-
methyl)phenyl]propenoate 3b. Potassium carbonate (89 mg,
0.64 mmol) was suspended in methanol (12 mL). The azlactone
9 (3.40 g, 13.1 mmol) was added in portions over 45 min, then
the brown solution was stirred for another 1 h. The solution was
neutralized from pH 12.3 to pH 6À7 with Dowex 50WX-8-200
ion-exchange resin; then the resin was removed by filtration. The
solvent was evaporated, and the compound was purified by
recrystallization from ethyl acetate (about 20 mL) to give the
enamide 3b as a yellow, granular solid (3.03 g, 91.8%). 1H NMR
(400 MHz, acetone-d6): δ 8.7 (br s, 1H), 7.60 (d, J = 8.0 Hz, 2H),
7.39 (d, J = 8.0 Hz, 2H), 7.24 (s, 1H), 4.65 (s, 2H), 4.4 (br s, 1H),
3.75 (s, 3H), and 2.09 (s, 3H). 13C NMR (100 MHz, acetone-
d6): δ 170.2, 170.1, 166.9, 145.1, 133.7, 132.6, 131.0, 127.8,
+
127.6, 127.4, 64.6, 52.8, 23.2. m/z (APCI) 267, (M + NH4 ,
78%) and 250 (MH+, 100%).
Preparation of (Z)-2-(Acetylamino)-3-[4-(hydroxymethyl)-
phenyl]propenoic Acid 3c. Sodium hydroxide (16.9 g, 422 mmol)
was dissolved in methanol (200 mL). Azlactone 9 (36.5 g,
141 mmol) was added over 1 h, the dark-brown solution was
heated to reflux for 1 h, then allowed to cool to room temperature.
Water (200 mL) was added, then most of the solvent was
evaporated to about 150 mL. The solution was extracted with
MTBE (100 mL), then acidified from pH (13.5 to 2.5 with 2 M
sulfuric acid (about 100 mL). The suspension was stirred for 1 h,
then filtered. The collected solid was washed with water (4 Â
50 mL), then dried to give a green solid (23.5 g). This was
suspended in isopropyl alcohol (80 mL), heated to reflux for
30 min, then allowed to cool to room temperature, stirred for 2 h,
then filtered, washed with isopropyl alcohol (2 Â 30 mL), and
dried to give the acid 3c as a pale-green solid (21.8 g, 66%). 1H
NMR (400 MHz, DMSO-d6): δ 12.65 (br s, 1H), 9.46 (s, 1H),
7.57 (d, J = 8.2 Hz, 2H), 7.34 (d, J = 8.2 Hz, 2H), 7.15 (s, 1H), 5.27
(br s, 1H), 4.51 (s, 2H), and 1.98 (s, 3H); 1H NMR (100 MHz,
DMSO-d6): δ 169.5, 166.8, 144.2, 132.4, 131.6, 129.9, 127.1,
Preparation of 4-[(Z)-(2-Methyl-5-oxo-1,3-oxazol-4(5H)-yli-
dene)methyl]benzyl Acetate 9. To a 2-L three-neck flask,
fitted with an overhead agitator, thermometer, and condenser
were charged sodium acetate (59.2 g, 0.722 mol), crude 4-
(hydroxymethyl)benzaldehyde 2 (69.0 g, ∼0.33 mol), ethyl
acetate (100 mL) and acetic anhydride (120 mL, 1.27 mol).
The slurry was heated to reflux (104À106 °C) and maintained at
reflux for 15 min before cooling to below 40 °C. N-acetylglycine
(42.2 g, 0.360 mol) and ethyl acetate (100 mL) were charged,
and the resultant mixture was heated to reflux (93À95 °C) and
maintained for 18 h. The batch was cooled to below 60 °C at
which point it solidified; and water (400 mL) was added, and the
resultant solid was isolated by vacuum filtration, the slurry
washed in water (300 mL), then washed in MTBE (300 mL),
and then displacement washed with MTBE (100 mL). The
yellow solids were dried under vacuum at 45 °C to give 56.2 g of
+
126.8, 62.9, and 22.9. m/z (APCI) 253, (M + NH4 , 100%) and
250 (MH+, 73%).
Preparation of Methyl (2R)-2-(Acetylamino)-3-[4-(acetoxy-
methyl)phenyl]propanoate 2a. A glass linear was charged with
3a (500 mg, 2.01 mmol) and methanol (4 mL) and then was
secured in the Argonaut Endeavor. The manifold was flushed
with nitrogen, and the vessel was charged with nitrogen to a
pressure of ∼5.7 bar, the contents were stirred (250 rpm) for
5 min, and the vessel was vented. This charge/stir/vent cycle was
repeated four times. The vessel was heated to 40 °C while stirring
at 250 rpm for 20 min under 0.35 bar of nitrogen and then
vented. A Schlenk flask was charged with [((R,R)-Me-
DuPhos)Rh(COD)]BF4 (5.7 mg, 0.0094 mmol) and evacu-
ated/refilled with nitrogen three times prior to adding deoxyge-
nated methanol (11.6 mL). An aliquot of the catalyst solution
1
azlactone 9 (65%); purity by GC was 96.1%. H NMR (400
MHz, CDCl3): δ 8.08 (d, J = 7.8 Hz, 2H), 7.42 (d, J = 7.8 Hz,
2H), 7.13 (s, 1H), 5.14 (s, 2H), 2.41 (s, 3H), and 2.13 (s, 3H).
13C NMR (100 MHz, CDCl3): δ 171.2, 168.1, 166.7, 139.4,
133.4, 133.2, 132.7, 131.1, 128.7, 66.1, 21.4, and 16.1.
Preparation of Methyl (Z)-2-(Acetylamino)-3-[4-(acetoxy-
methyl)phenyl]propenoate) 3a. A suspension of azlactone 9
(10.0 g, 38.6 mmol) in methanol (100 mL) was heated at reflux
1169
dx.doi.org/10.1021/op200129m |Org. Process Res. Dev. 2011, 15, 1163–1171

Organic Process Research & Development
ARTICLE
(1 mL, 0.000804 mmol, s/c 2500) was added to the pressure
vessel. The manifold was flushed with hydrogen, and the vessel
was charged with hydrogen to ∼5.7 bar without stirring. This
charge/vent cycle was repeated twice. The vessel was charged to
∼5.7 of hydrogen, and an automated program which maintained
a constant hydrogen pressure of 100 psi, stirring at 1000 rpm,
and heating to 40 °C was executed for 16 h. The vessel was
vented, and the contents were diluted with methanol (7 mL) and
dichloromethane (3 mL) to obtain a clear solution. An aliquot
was concentrated in vacuo to provide 2a as a white solid. 1H NMR
and HPLC analysis indicated >98% conversion and 95.1% ee,
respectively. 1H NMR (400 MHz, CDCl3): δ 7.28 (d, J = 8 Hz,
2H), 7.08 (d, J = 2 Hz, 2H), 5.90 (d, J = 8 Hz, 1H), 5.08 (s, 2H),
4.92À4.87 (m, 1H), 3.74 (s, 3H), 3.17 (dd, J = 14, 6 Hz, 1H),
3.10 (dd, J = 14, 5 Hz, 1H), 2.11 (s, 3H), 2.00 (s, 3H).
Preparation of Methyl (2R)-2-acetylamino-3-(4-{[(2R,6S)-
2,6-dimethylpiperidin-1-yl]methyl}phenyl)propanoate 11. To
a 250-mL three-neck flask, fitted with an overhead agitator,
thermometer, and addition funnel, 2b (20.0 g), were charged
diisopropylethylamine (16.3 g, 126 mmol) and dichloromethane
(80 mL). Methanesulfonic anhydride (18.6 g, 107 mmol) was
dissolved in dichloromethane (70 mL) and charged to the
addition funnel, then charged to the batch over 20 min, main-
taining the temperature below 25 °C. The resultant orange
solution, containing O-mesylate 10 was stirred for 3 h, and then
cis-2,6-dimethylpiperidine (18.6 g, 164 mmol) was charged; the
mixture was stirred at ambient temperature for 5 h, and then
further methanesulfonic anhydride (4.5 g, 2.6 mmol) and cis-2,6-
dimethylpiperidine (9.0 g, 8.0 mmol) were charged. The dark-
orange solution was stirred overnight at ambient temperature.
The product was extracted into 20% w/w aqueous citric acid
(120 g), washed with dichloromethane (70 mL), then basified
with 50% aqueous potassium carbonate solution (100 g). The
product oil was extracted into dichloromethane (2 Â 100 mL)
and then distilled at 40 °C to give 18.5 g of 11, a pale-orange
Preparation of Methyl (2R)-2-(Acetylamino)-3-[4-(hydroxy-
methyl)phenyl]propanoate 2b. To a 50-mL Parr hydrogenation
vessel were charged 3b (1.25 g, 5.01 mmol) and methanol
(20 mL), and the vessel was sealed and purged. The catalyst
[((R,R)-Ethyl-DuPhos)Rh(COD)]BF4 (50 mg, 0.08 mmol) was
charged, and after resealing and purging the vessel the mixture
hydrogenated at 7.1 bar until uptake of hydrogen gas ceased. The
clear yellow solution was concentrated under vacuum at 40 °C to
give a suspension of a pale-yellow solid in methanol, toluene
(50 mL) was charged, and the remainder of the methanol was
removed at 40 °C. The solid was isolated by vacuum filtration,
washed with toluene (20 mL), and dried to 45 °C to give 1.1 g of
2b (88%). The enantiomeric excess of the product was greater
than 98%. 1H NMR (400 MHz, CDCl3): δ 7.30 (d, J = 8.0 Hz,
2H), 7.08 (d, J = 8.0 Hz, 2H), 5.88 (d, J = 6.8 Hz, 1H), 4.92À4.86
(m, 1H), 4.68 (s, 2H), 3.75 (s, 3H), 3.17 (dd, J = 13.6, 5.6 Hz,
1H), 3.10 (dd, J = 13.6, 5.4 Hz, 1H) and 2.00 (d, 3H); 1H NMR
(400 MHz, acetone-d6) δ ppm 7.40 (d, J = 6.4 Hz, 1H), 7.28 (d,
J = 7.6 Hz, 2H), 7.19 (d, J = 7.6 Hz, 2H), 4.71À4.65 (m, 1H),
4.60 (s, 2H), 3.10 (dd, J = 13.2, 5.2 Hz, 1H), 2.95 (dd, J = 13.2, 8.2
Hz, 1H), 2.9 (br s, 1H) and 1.88 (s, 3H). 13C NMR (100 MHz,
acetone-d6): δ 173.3, 170.5, 142.2, 136.9, 130.2, 127.8, 64.7, 54.9,
1
viscous oil (67%). Purity by HPLC was 93%. H NMR (400
MHz, CDCl3): δ 7.30 (d, J = 7.8 Hz, 2H), 7.00 (d, J = 7.8 Hz,
2H), 5.97 (d, J = 8.0 Hz, 1H), 4.89À4.84 (m, 1H), 3.77 (s, 2H),
3.71 (s, 3H), 3.14À3.04 (m, 2H), 2.46À2.44 (m, 2H), 1.98 (s,
3H), 1.67À1.62 (m, 1H), 1.57À1.55 (m, 2H), 1.32À1.28 (m,
3H), and 1.05 (d, J = 5.6 Hz, 6H). 13C NMR (100 MHz, CDCl3):
δ 172.6, 170.0, 141.3, 133.8, 129.2, 128.7, 57.7, 53.7, 53.5, 52.7,
37.9, 35.0, 24.7, 23.6, and 22.6. m/z (APCI) 347 (MH+, 100%).
Preparation of Methyl 2(R)-Acetyl-(tert-butoxycarbonyl)-
amino-3-(4-{[(2R,6S)-2,6-dimethylpiperidin-1-yl]-methyl}-
phenyl)propanoate 12. The N-acetyl-amino acid methyl ester
11 (6.07 g, 17.5 mmol) was dissolved in THF (15 mL) under
nitrogen. A solution of di-tert-butyl-dicarbonate (5.73 g, 26.3 mmol,
1.5 equiv) in THF (10 mL) was added, and then 4-dimethylami-
nopyridine (115 mg, 1.02 mmol, 5 mol %) was added. The solution
was heated at 40 °C under nitrogen for 16 h. Water (2.5 mL) was
added, and the solution was stirred for 30 min whilst allowing to
cool to room temperature (this treatment helps the following
partition). The mixture was partitioned between toluene (40 mL)
and water (25 mL). The aqueous phase was separated, and the
organic phase was washed with water (10 mL). The organic phase
was dried (Na2SO4), filtered, and evaporated to give mixed acetyl/
BOC imide 12 as a brown oil (7.73 g, 17.3 mmol, 99%). 1H NMR
(400 MHz, CDCl3): δ 7.26 (d, J = 7.8 Hz, 2H), 7.04 (d, J = 7.8 Hz,
2H), 5.47 (dd, J = 10.2, 6.0 Hz, 1H), 3.76 (s, 2H,), 3.73 (s, 3H),
3.40 (dd, J = 14.0, 6.0 Hz, 1H), 3.14 (dd, J = 14.0, 10.2 Hz, 1H),
2.45À2.42 (m, 2H), 2.28 (s, 3H), 1.64À1.62 (m, 1H), 1.57À1.54
(m, 2H), 1.43 (s, 9H), 1.31À1.26 (m, 3H), and 1.05 (d, J = 5.6 Hz,
6H). 13C NMR (100 MHz, CDCl3): δ 173.0, 171.3, 152.4, 140.6,
135.4, 129.4, 128.4, 84.2, 57.6, 57.5, 57.4, 53.7, 52.7, 35.7, 35.1,
35.1, 28.3, 26.7, 24.7, and 22.6. m/z (APCI) 447 (MH+, 100%)
and 347 (27).
+
52.6, 38.4, and 23.0. m/z (APCI) 269, (M + NH4 , 100%) 252
(MH+, 42%), and 234 (17).
Preparation of (2R)-2-(Acetylamino)-3-[4-(hydroxymethyl)-
phenyl]propanoic Acid 2c. The enamide-acid 3c (10.5 g, 44.6
mmol) was suspended in methanol (80 mL) in a 300-mL
pressure vessel linear. The vessel was assembled, then purged
with nitrogen (3 Â 3.6 bar); then a solution of [((R,R)-Et-
DuPhos)Rh COD]BF4 (59 mg, 0.09 mmol) in degassed methanol
(1.0 mL) was added. The vessel was purged with hydrogen (2 Â 3.6
bar), then charged to 7.1 bar with hydrogen. The reaction was
stirred for 120 min, repressurising to 7.1 bar psi at intervals. A total of
5.6 psi hydrogen was consumed. The pressure was released, then the
vessel was purged with nitrogen (3.6 bar). The solvent was
evaporated to give the (R)-N-acetyl (4-(hydroxymethyl)phenyl)-
alanine 2c as an orange foam (11.0 g, 104%); enantiomeric excess
>98%. 1H NMR (400 MHz, acetone-d6): δ 7.42 (d, 1H, J= 8.0 Hz),
7.28 (d, 2H, J = 8.0 Hz), 7.22 (d, 2H, J = 8.0 Hz), 4.74À4.68 (1H,
m), 4.60 (s, 2H), 3.19 (dd, J = 13.8, 5.4 Hz, 1H), 2.98 (dd, J = 13.8,
8.0 Hz, 1H,) and 1.90 (s, 3H). 1H NMR (400 MHz, DMSO-d6): δ
7.95 (d, J = 8.0 Hz, 1H,), 7.01 (d, J = 8.4 Hz, 2H,), 6.96 (d, J = 8.4
Hz, 2H), 4.25 (s, 2H), 4.19À4.13 (m, 1H), 2.81 (dd, J = 13.8, 5.0
Hz, 1H), 2.60 (dd, J = 13.8, 9.6 Hz, 1H), and 1.57 (s, 3H). 13C
NMR (100 MHz, DMSO-d6): δ 173.6, 169.5, 140.9, 136.4, 129.1,
Preparation of 2(R)-(tert-Butoxycarbonyl-amino)-3-(4-(2R,6S)-
2,6-dimethylpiperidine-1-yl)-methyl-phenyl)-propionic Acid
1.12. The mixed imide 12 (7.70 g, 17.2 mmol) was dissolved in
THF (25 mL). Methanol (8 mL) was added, and the solution
was cooled to 2 °C. A solution of LiOH H2O (1.45 g, 34.5 mmol,
3
2 equiv) in water (35 mL) was added over 75 min. [Temperature
was maintained at 2À5 °C and pH at start of LiOH addition =
10.2. This quickly rose to 13.7 with the first two drops of LiOH.]
Addition rate was adjusted such that pH remained below 13.7;
pH upon completion of addition = 13.3. [Care needs to be taken
+
126.7, 63.0, 54.0, 48.9, 36.9, and 22.7. m/z(APCI) 255, (M + NH4 ,
100%), 238, (MH+, 100%) and 220 (11).
1170
dx.doi.org/10.1021/op200129m |Org. Process Res. Dev. 2011, 15, 1163–1171

Organic Process Research & Development
ARTICLE
at the start of the addition since pH and temperature can rise
rapidly at this point.] The reaction was stirred for a further
75 min at 3À5 °C (final pH 13.05). TLC (silica plate, DCM/
MeOH, 9:1, visualised with KMnO4) showed reaction to be
complete. The mixture was acidified with 10% KHSO4 (33 mL)
[final pH 6.2, temperature rises from 5 to 15 °C, addition of
KHSO4 was then stopped since pH was changing very slowly and
pH paper indicated a pH of 5À6 had been reached]. The solution
was concentrated under reduced pressure (final pressure
50 mbar, 30 °C) to remove THF/MeOH. Some insoluble
material separated. This was removed by extraction with MTBE
(25 mL), and the MTBE phase was discarded. The aqueous
phase (pH = 5.4) was saturated with NaCl and extracted with
DCM (3 Â 25 mL). The combined extracts were dried
(Na2SO4) and filtered. A sample was removed for analysis (ee
97.8%). The bulk was concentrated under reduced pressure
(300 mbar, 30 °C). At the first sign of foaming, evaporation
was stopped (residual solution 17 g), and water (50 mL) was
added. Evaporation was continued (final conditions 70 mbar,
30 °C) to give a clear aqueous solution (traces of insoluble
particles, pH of solution = 3.5). Amberlite IRA-67 (Merck, cat.
no. 1.15959.0500, weakly basic ion-exchange resin, 2.25 g,
prewashed with water) was added in portions, and the solution
was stirred until pH 6.0 was reached (approximately 1 h). The
solution was filtered through a pad of Celite, and the Celite was
washed with water (10 mL). The solution was evaporated (final
conditions 50 mbar, 50 °C). The residue was evaporated from
ethanol (10 mL) and then from ethyl acetate (10 mL) (final
conditions 20 mbar, 40 °C). The residue was dissolved in ethyl
acetate (10 mL) (warming at 40 °C required). Heating was then
stopped, and MTBE (20 mL) was added whilst stirring the
solution. [After a few minutes precipitation was observed.
Usually on smaller scale it has been necessary to seed the
mixture.] The mixture was stirred at room temperature over-
night. The precipitate was filtered, washed with EtOAc/MTBE
(1:1, 2 Â 5 mL), and dried under high vacuum (warm water bath,
approximately 40À50 °C required) to give 1 as an off-white solid,
(4.90 g, 73%), ee 98.6%, mp 105À110 °C. [R]2D5 À53.9 (c = 1.0,
MeOH). 1H NMR (400 MHz, DMSO-d6): δ 7.22 (d, J = 8.2 Hz,
2H), 7.11(d, J = 8.2Hz, 2H), 6.63 and 6.21 (d, J = 7.6 Hzand d, J =
6.0 Hz, 1H), 4.00À3.94 (m, 1H), 3.67 (s, 2H), 3.00 (dd, J = 13.2,
4.4 Hz, 1H), 2.78 (dd, J = 13.2, 9.4 Hz), 2.55À2.35 (m, 2H),
1.63À1.48 (m, 3H), 1.31 (s, 9H), 1.27À1.14 (m, 3H), and 0.96
(d, J = 6.0 Hz, 6H). 13C NMR (100 MHz, DMSO-d6): δ 174.3 (s),
172.6 (s), 155.4 (s), 139.5 (s), 136.6 (s), 129.2 (d), 127.6 (d), 77.9
(s), 57.5(d), 56.0(d), 53.4(t), 37.0(t), 34.2 (t), 28.5(q), 23.9(t),
and 21.7 (q). m/z (APCI) 391 (MH+, 100%), 335 (6).
2c: column: Chirobiotic T (250 mm  4.6 mm, 5 μm); flow rate:
1 mL/min; detector: UV/vis 210 nm, mobile phase: methanol/
TEA/Acetic Acid (100:0.1:0.1); retention times: (S), 5.80 min;
(R), 10.78 min.
HPLC method for chiral analysis of 2(R)-(tert-butoxycarbonyl-
amino)-3-(4-(2R,6S)-2,6-dimethylpiperidine-1-yl)-methyl-phenyl)-
propionic acid 1: column: Chirobiotic T (250 mm  4.6 mm,
5 μm); flow rate: 0.8 mL minÀ1; detector: UV/vis 210 nm; mobile
phase: 0.1% triethylammonium acetate (pH 4.5)/methanol iso-
cratic (62:38); retention times: (S), 10.77 min; (R), 14.68 min.
’ ASSOCIATED CONTENT
S
Supporting Information. Chromatographic methods for
b
analysis of compounds 1, 2a, 2b, 2c, 3b, 5, 6, 7, 8, 9, 11, and 12.
Pilot-plant procedures for 30-kg manufacture of 1. H NMR
1
spectra of compounds 1, 2aÀc, 3aÀc, 9, 11, and 12. This
material is available free of charge via the Internet at http://
pubs.acs.org.
’ AUTHOR INFORMATION
Corresponding Author
*mfox@drreddys.com
’ ACKNOWLEDGMENT
We thank George Evans for assistance with mass spectra and
Colin Dewar, Thomas Mahoney and Brendan Mullen for
assistance with chiral analysis.
’ REFERENCES
(1) A partial description of this work can be found in Fox, M. E.;
Meek, G. A., WO 2007038116, 2007.
(2) For reviews see: (a) Lennon, I. C.; Moran, P. H. Curr. Opin. Drug
Discovery Dev. 2003, 6, 855–875. (b) Lennon, I. C.; Pilkington, C. J.
Synthesis 2003, 1639–1642. (c) Blaser, H. U.; Malan, C.; Pugin, B.;
Spindler, F.; Steiner, H.; Studer, M. Adv. Synth. Catal. 2003, 345,
103–151. (d) Ager, D. Curr. Opin. Drug Discovery Dev. 2002, 5, 892. (e)
de Vries, J. G. In Encyclopedia of Catalysis; Horvath, I. T., Ed.; John Wiley &
Sons: New York; Vol. 3, p 295. (f) Burk, M. J. Acc. Chem. Res. 2000,
33, 363–372.
(3) Herzner, H.; Kunz, H. Tetrahedron 2007, 63, 6423–6436.
(4) Morera, E.; Ortar, G.; Varani, A. Synth. Commun. 1998,
28, 4279–4285.
(5) Herzner, H.; Kunz, H. Tetrahedron 2007, 63, 6423–6436.
(6) Smith, S. C.; Sloane, N. H. Biochim. Biophys. Acta 1967,
148, 414–422.
(7) Nakazawa, M.; Hashimoto, A.; Nozaki, H. WO 2006101266 A1,
2006.
(8) Kim, M. H.; Lai, J. H.; Hangauer, D. G. Int. J. Pept. Protein Res.
1994, 44, 457–465.
(9) Mak, C. C.; Bampos, N.; Darling, S. L.; Montalti, M.; Prodi, L.;
Sanders, J. K. M. J. Org. Chem. 2001, 66, 4476–4486.
(10) For an informative discussion of methods of enamide
synthesis see: Panella, L.; Aleixandre, A. M.; Kruidhof, G. J.; Robertus,
J.; Feringa, B.; de Vries, J. G.; Minnaard, A. J. Org. Chem. 2006,
71, 2026–2036.
(11) Pilkington, C. J.; Zanotti-Gerosa, A. Org. Lett. 2003, 5, 1273–1275.
(12) Ferrari, B; Gougat, J; Muneaux, C; Muneaux, Y; Perreaut, P;
Planchenault, C. WO 9725315 A1, 1997.
HPLC method for chiral analysis of N-acetyl-4-(acetoxy-
methyl)-phenylalanine methyl ester 2a and N-acetyl-4-
(hydroxmethyl)-phenylalanine methyl ester 2b: column: Chir-
acel OD (250 mm  4.6 mm, 10 μm); flow rate: 1.0 mL minÀ1
;
mobile phase: heptane/propan-2-ol isocratic (80:20); detector:
UV/vis 254 nm; retention times: (R)-2a, 7.08 min; (S)-2a, 8.71
min; (R)-2b, 8.57; (S)-2b, 10.06 min.
HPLC method for chiral analysis of N-acetyl-4-(hydrox-
methyl)-phenylalanine methyl ester 2b: column: Chiracel OJ
(250 mm  4.6 mm); flow rate: 1.0 mL minÀ1; mobile phase:
heptane/propan-2-ol isocratic (85:15); detector: UV/vis
214 nm; temperature: 40 °C; retention times: 3b, 24.7 min;
(R)-2b, 15.4; (S)-2b, 18.2 min.
HPLC method for chiral analysis of methyl (2R)-2-
(13) Burk, M. J.; Allen, J. G. J. Org. Chem. 1997, 62, 7054–7057.
(acetylamino)-3-[4-(hydroxymethyl) phenyl]propanoic acid
1171
dx.doi.org/10.1021/op200129m |Org. Process Res. Dev. 2011, 15, 1163–1171