Selection and development of a route for cholesterol absorption inhibitor AZD4121

Article abstract of DOI:10.1021/op200314z

The development of a synthetic route to the cholesterol absorption inhibitor AZD4121 is presented. Key steps are a highly enantioselective CBS reduction, a stereospecific Staudinger reaction, an amine/lithium chloride mediated ester hydrolysis, and a resolution of a 50:50 diastereomeric mixture by recrystallization. The synthesis was accomplished in 10 linear steps, and the overall yield, when compared with the lead optimization (LO) route, was improved from 1% to 20%. All purifications of intermediates through preparative HPLC or silica gel chromatography were avoided. This was possible since many of the intermediates along the route could be used as such in the next step until an intermediate with suitable crystalline properties could be identified and purified through crystallization.

Full text of DOI:10.1021/op200314z

Article
pubs.acs.org/OPRD
Selection and Development of a Route for Cholesterol Absorption
Inhibitor AZD4121
Staffan Karlsson and J. Henrik Sorensen*
̈
Medicinal Chemistry, AstraZeneca R&D Molndal, S-431 83 Molndal, Sweden
̈
̈
S
* Supporting Information
ABSTRACT: The development of a synthetic route to the cholesterol absorption inhibitor AZD4121 is presented. Key steps are
a highly enantioselective CBS reduction, a stereospecific Staudinger reaction, an amine/lithium chloride mediated ester
hydrolysis, and a resolution of a 50:50 diastereomeric mixture by recrystallization. The synthesis was accomplished in 10 linear
steps, and the overall yield, when compared with the lead optimization (LO) route, was improved from 1% to 20%. All
purifications of intermediates through preparative HPLC or silica gel chromatography were avoided. This was possible since
many of the intermediates along the route could be used as such in the next step until an intermediate with suitable crystalline
properties could be identified and purified through crystallization.
thioglycolate furnished sulfide 4, which sequentially was
protected as its ketal and then hydrolyzed to compound 5
(Scheme 1B). For the ring closure to give the azetidinone, the
LO route was based on a condensation of chiral acylox-
azolidinone 6 with a Schiff’s base 3 (prepared according to
Scheme 1A) in the presence of TiCl4 and Ti(OiPr)4 followed
by cyclization to give the azetidinone scaffold 8 via intermediate
7. The azetidinone 8 could be obtained diastereomerically pure
after silica gel chromatography. After deprotection of the ketal
to ketone 9, the ester was hydrolyzed to the corresponding acid
10 followed by coupling with tert-butyl 2-aminoacetate to give
tert-butyl ester 11. Cleavage of the tert-butyl ester followed by
NaBH4 reduction gave acid 12 in a low diastereoselectivity,
which then was coupled with (R)-2-amino-3-cyclohexylpropa-
noic acid to give the acid 13 in low yield. This was initially
purified by preparative HPLC followed by freeze-drying to
obtain a solid with suitable properties. Later, it was found that
the acid 13 nicely crystallized as its tert-butylammonium salt,
which also turned out to be the preferred solid state form for
the API 1 (AZD4121). The synthesis was completed in 10
steps and substeps, of which the cyclization step of 6 to 8 did
not give reproducible results. The sequence also included
several low-yielding steps. This, in combination with several
challenging separations by flash chromatography or preparative
HPLC, made it clear that the sequence was not attractive for
scale up.4,5 Thus, in order to supply the project with the first
hundred grams of API material for toxicity studies, further
efforts to improve the overall yield and robustness of the route
were necessary.
INTRODUCTION
■
Hypercholesterolemia is a metabolic disorder that is associated
with several diseases, most notably atherosclerosis, which is
often followed by early death. Cholesterol absorption inhibitors
(CAI) is a class of compounds that have attracted attention for
the treatment of hypercholesterolemia. The mechanism of CAI
is based on blocking the uptake of cholesterol. The azetidinone
(β-lactam) structure is common to the class of CAIs to which
also ezetimibe1 (Figure 1) belongs. A feature of AZD4121
dissimilar to ezetimibe is that it has a low bioavailability, which
was thought to give a low systemic exposure and low risk of
side effects. Much synthetic experience has been gained from
the synthesis of ezetimibe and its analogues, and a number of
syntheses have been devised.2 Well known routes include
catalytic asymmetric synthesis,2a asymmetric induction by chiral
auxiliary,2b−d and chiral pool based chemistry.2e In the present
work, we have optimized a synthesis of a CAI, viz. AZD4121 1
(Figure 1).3
Unlike ezetimibe, AZD4121 incorporates a sulfur atom
which potentially can stabilize a vicinal carbocationic center, i.e.
the benzylic one bearing the hydroxyl group. Thus, we were not
surprised to hear from the lead optimization (LO) group that
epimerization of this center occurred when treated with strong
acids such as formic acid and HCl. Moreover, the sulphur atom
probably has a pKa lowering effect of the adjacent methine
proton. This increases the risk of epimerization of this position
when treated with strong bases. It was actually reported from
the LO group that bases such as LiOH and NaOH efficiently
effected this transformation, and consequently, mild conditions
had to be employed in every single operation when handling
this sensitive scaffold.
Second-Generation Synthesis. Retrosynthesis. At the
outset of our scaling up, we identified 14 as a strategic target.
The retrosynthesis of 14 is outlined in Figure 2. Our synthetic
strategy was based on the formation of the azetidinone scaffold
from two appropriately protected building blocks, A and B. We
hoped the Staudinger reaction between chiral building block A
RESULTS AND DISCUSSION
■
First-Generation Synthesis. The LO route utilized a
strategy similar to one of the published procedures for
ezetimibe.1b,c The route, shown in Scheme 1, was used for
producing the first milligrams of material for in vivo studies.
Reaction between 2-bromo-4′-fluoroacetophenone and ethyl
Received: November 3, 2011
Published: March 20, 2012
© 2012 American Chemical Society
586
dx.doi.org/10.1021/op200314z | Org. Process Res. Dev. 2012, 16, 586−594

Organic Process Research & Development
Article
Figure 1. Structures of AZD4121 and ezetimibe.
and imine B would furnish a diastereomeric mixture of
azetidinone 14. We believed that this mixture at some stage
of the synthesis could be resolved into one pure diastereomer.
Building block A should be easily accessible via coupling of C
with methyl thioglycolate. Chiral building blocks such as C are
easily obtained via asymmetric CBS-reduction of the corre-
sponding halo-ketone.6 Imine B can be obtained from
condensation of 4-fluoroaniline with aldehyde D.
Much experience had been gained during the LO-phase in
the synthesis of 14, and several issues had to be addressed. It
had been found that, upon formic acid catalyzed removal of a
tert-butyl ester of 14, 5−10% epimerization of the benzylic OH
position took place. Moreover, the sulfur substituted position of
the azetidinone was found to epimerize in the presence of
various bases, and furthermore, ring-opening of the azetidinone
could take place with nucleophilic bases. Another concern was
that the thioether function would probably interfere in catalytic
hydrogenations, e.g. debenzylation. Thus, our avenue of
protective groups was narrow in terms of conditions for
deprotection, and we concluded that protective groups were to
be avoided if possible.
Since it had been reported from the LO group that most of
their intermediates in the first generation synthesis were
noncrystalline and, thus, had to be purified through
chromatography, we were aware of the risk of ending up in
the same situation with our second generation route. In that
case and in order to avoid chromatography, our strategy was to
identify a route that included steps that were both high yielding
and which could give intermediates that were pure enough to
be processed in the next step without further purification. This
would enable us to telescope the intermediates until material
with suitable crystalline properties could be identified and
purified through crystallization.
The synthesis of the acid 14 proceeded through eight
intermediates, of which seven turned out to be oils or
amorphous solids, and thus, to avoid chromatography, we
had to use all these intermediates as such with no purifications
other than extractions.
necessary for the completion of building block 18. Among
many silyl protective groups available, both TES and TMS were
tried but rejected because of their lability in the later hydrolysis
step. The TBDMS was found to fulfill our requirements, and
standard protection conditions using TBDMS-chloride in DMF
in the presence of imidazole gave building block 17 as an oil
which was pure enough to be used as such in the next step.
Simple alkaline hydrolysis of the methyl ester 17 gave us the
chiral building block 18 in good yield over 4 steps. This viscous
acid was used as such in the next step without further
purification.
The Schiff base 3 was prepared from 4-hydroxybenzaldehyde
by alkylation of this with methyl bromoacetate to give aldehyde
2 followed by the condensation of this with 4-fluoroaniline
using azeotropic removal of water, according to Scheme 1A.
The pure Schiff base 3 was obtained in 62% overall yield over
the 2 steps after a crystallization from acetonitrile and
methanol.
Staudinger Reaction. With the two building blocks 18 and
3 (corresponding to building blocks A and B, Figure 2) in hand,
the ring closure was effected using 1-methyl-2-chloropyridi-
nium iodide to give azetidinone trans-19 (Scheme 2B).7 During
the optimization of the reaction, it was found that slow addition
of the acid 18 as a solution in DCM to a mixture of imine 3,
triethylamine, and 1-methyl-2-chloropyridinium iodide in DCM
gave superior results. The resulting product 19 was obtained as
an oil, a 50:50 diasteromeric mixture of only trans isomers.8
The crude product was of sufficient purity to be used as such in
the next step.
Hydrolysis and Deprotection. Hydrolysis of the methyl
ester in 19 was a major challenge. Standard conditions using
alkali hydroxide all resulted in ring-opening and/or epimeriza-
tion of the sulfur substituted position in the azetidinone. Other
conditions tried were reflux in toluene with magnesium iodide,
which, apart from a messy mixture, led to 35% formation of cis
isomers.9 Lewis acid catalysis, such as ferric(III) chloride in wet
solvents, did give hydrolysis to various extents but was not
possible to balance against side reactions on the benzylic OH
position. Mild hydrolysis using Me3SnOH10 was considered but
abandoned due to toxicity and waste issues. After much
experimentation, a novel hydrolysis method was developed
using a mixture of triethylamine and lithium chloride in wet
acetonitrile at 20 °C.11 After having completed the ester
hydrolysis, as little as 1−2% cis isomers was detected as a result
of epimerization at the sulfur substituted position of the
azetidinone. Removal of the TBDMS group was also a slightly
delicate problem. The benzylic position bearing the hydroxyl
group was found to epimerize quite readily under acidic
conditions. After some optimization work using various
additives, we finally found that mild, albeit slow, deprotection
Building Block A and B Synthesis. Scheme 2 outlines our
synthesis for securing the first 0.8 mol of building block 14. The
synthesis commenced on a 3 mol scale by the introduction of
the chiral benzylic alcohol in 15 via an asymmetric (R)-(+)-2-
methyloxazaborolidine catalyzed borane-dimethyl sulfide re-
duction6 of 2-bromo-4′-fluoroacetophenone. After successfully
having isolated 15 as an oil in >99% ee, the material was next
reacted with methyl thioglycolate in the presence of triethyl-
amine and cesium carbonate to give 16 as an oil. The
protection of the resulting benzylic alcohol had to be carefully
balanced, since mild deprotection conditions later on were
essential based on the previously mentioned known lability of
the molecule, yet withstand the ester hydrolysis which was
587
dx.doi.org/10.1021/op200314z | Org. Process Res. Dev. 2012, 16, 586−594

Organic Process Research & Development
Article
a
b
Scheme 1. (A) Preparation of Imine 3 and (B) First Generation Synthesis of AZD4121
a
b
Reaction conditions: (i) K2CO3, methyl bromoacetate, acetone, reflux; (ii) 4-fluoroaniline, toluene, reflux, 62% overall yield. Reaction conditions:
(i) K2CO3 (1 equiv), acetone, reflux (2 h), 94%; (ii) (a) 2,2-dimethyl-1,3-propanediol (7.5 equiv), p-TSA (cat.), benzene; (b) LiOH, H2O, THF, 20
°C, 19 h, 87%; (iii) DCC (1.1 equiv), DCM, 0 °C, (S)-4-phenyloxazolidin-2-one (1 equiv), 72 h, flash chrom. 74%; (iv) (a) Ti(OiPr)4 (0.25 equiv),
TiCl4 (1 equiv) DCM, 10 min; (b) 3 (2 equiv), −30 °C, 20 min; (c) DIPEA (2.5 equiv), −30 °C, 90 min; (d) iPrOH, −78 to 0 °C; (e) cryst.
(MeOH), 67%; (v) (a) N,O-bis(trimethylsilyl)acetamide (3 equiv), toluene, 90 °C, 1 h; (b) TBAF, 45 °C, 21 h, flash chrom. 52%; (vi) HCOOH/
DCM 25:75, 20 °C, 16 h, flash chrom. 43%; (vii) Et3N, LiCl, MeCN, H2O, 20 °C, 18 h, 56%; (viii) tert-butyl 2-aminoacetate·HCl (1.2 equiv),
TBTU (1.3 equiv), NMM (3.6 equiv), DCM 20 °C, 1 h, flash chrom. 87%; (ix) (a) HCOOH, 35 °C, 6 h; (b) NaBH4, MeOH, prep HPLC, 36%; (x)
(a) NMM, TBTU (1.2 equiv), DMF, 20 °C, 60 min; (b) H-(R)-cyclohexylalanine-OH, 20 °C, 60 min, prep. HPLC, 43%.
could be accomplished by using a mixture of acetic acid, water,
and lithium chloride at 25 °C to give compound 20.
was not possible to improve further.12 After isolation of free
acid 14, only attachment of the dipeptide 23 remained.
Peptide Coupling. With the inherent sensitivity of our
scaffold in mind, we sought a coupling with the dipeptide
without having to resort to further deprotection procedures.
The dipeptide was synthesized by uneventful peptide synthesis
as depicted in Scheme 3A.
Resolution. The resolution of the resulting diastereomeric
mixture of the foamlike carboxylic acid, 20, was tried as the
corresponding ammonium salt with various amines. With initial
slightly positive results using tert-butylamine, we then found
that the salt with adamantylamine had favorable properties.
After crystallization, a 40% yield of 21 over four steps was
obtained with a diastereomeric ratio of 92:8 that in our hands
The final coupling of dipeptide 23 with carboxylic acid 14 to
give the API AZD4121 is outlined in Scheme 3B:
588
dx.doi.org/10.1021/op200314z | Org. Process Res. Dev. 2012, 16, 586−594

Organic Process Research & Development
Article
Figure 2. Retrosynthesis of 14.
a
b
Scheme 2. (A) Preparation of Building Block 18 and (B) Second Generation Route to Compound 14
a
Reaction conditions: (i) (R)-(+)-2-Me-CBS-oxazaborolidine (0.05 equiv), BH3-DMS (0.6 equiv), THF, 0 °C, 99% ee; (ii) HSCH2COOMe, Et3N,
b
Cs2CO3, 0 °C, DMF; (iii) TBDMSCl, imidazole, DMF, 25 °C; (iv) NaOH, MeOH, 92% effective yield over 4 steps. Reaction conditions: (i) 2-
chloro-1-methylpyridinium iodide, Et3N, CH2Cl2, 23 °C; (ii) Et3N, LiCl, MeCN/water 95:5, 10−25 °C; (iii) LiCl, HOAc/water 10:1, 23 °C; (iv)
adamantylamine (0.7 equiv), acetone/MeOH 3:1, 40% over four steps; (v) 0.2 M HCl, EtOAc, 100%.
We wanted an active ester of 14 that was resistant enough
toward intermolecular attack of the free benzylic alcohol,
resulting in dimers but at the same time reactive toward attack
of the dipeptide. After having tried different13 active esters on
similar structures, our experience was that the 4-chlorophenyl
ester of 14 would be of suitable reactivity to fulfill the
requirements. TBTU mediated esterification between 4-
chlorophenol and 14 in the presence of triethylamine at −15
to 0 °C gave us the desired intermediate active ester 24 while
the competing intermolecular esterification could be minimized
589
dx.doi.org/10.1021/op200314z | Org. Process Res. Dev. 2012, 16, 586−594

Organic Process Research & Development
Article
a
b
Scheme 3. (A) Synthesis of Peptide Fragment 23 and (B) Activation and Coupling of 14 with Peptide 23
a
Reaction conditions: (i) thionyl chloride, methanol, 83%; (ii) N-Boc-Gly-OH, TBTU, NMM, DCM, 23 °C, 85%; (iii) NaOH, MeOH, 20 °C; (iv)
b
HCOOH, 40 °C, 71% (two steps). Reaction conditions: (i) 4-chlorophenol (2 equiv), Et3N (2.1 equiv), TBTU (1.3 equiv), DMF, −15 to 0 °C, 17
h; (ii) 23 (1.1 equiv), LiCl (15 equiv), 25−30 °C, 5 days; (iii) tert-butylamine (1.1 equiv), acetone, 35 °C, 2 days, 44% + 10% (three steps).
as a result of less reactivity of the benzylic OH as compared
with 4-chlorophenol.14 Even though isolation of the 4-
chlorophenyl ester was possible, the simplest procedure was
found to be direct addition of peptide 23 to the reaction
mixture. The dipeptide had a very low solubility in the reaction
mixture, leading to a slow reaction. With the aim of increasing
the rate of reaction, we investigated if additives such as lithium
salt could have a beneficial effect on the rate. Fortunately, we
found that concomitant addition of lithium chloride facilitated
dissolution of the dipeptide and also increased the speed of the
Figure 3. Main contaminant in the API.
reaction significantly. After workup, tert-butylamine was added
to the crude 13 in order to convert the final API to a crystalline
isolable salt. Residual 4-chlorophenol, which was found to
interfere with the crystallization, was removed by washing of
the salt with MTBE. The tert-butylammonium salt of AZD4121
was finally crystallized from acetone in 44% yield15 over three
steps to give a material with a diastereomeric ratio of 95:5.16
This material was acceptable for our initial toxicological studies.
Apart from the diastereomeric impurity, the main impurity is
the dehydrated product depicted in Figure 3 (1.3%, the
structure is suggested on the basis of the HRMS).17
CONCLUSION
■
We have shown a route to prepare the cholesterol absorption
inhibitor AZD4121 without the use of chromatographic
purifications. This was possible since the route included steps
that were both high yielding and furnished intermediates that
were pure enough to be telescoped into the next step. Only two
purification steps through crystallization were used in the 10
linear step sequence. The accomplishments also include the
following: (1) the use of very mild methods for ester hydrolysis
and TBDMS deprotection; (2) partial resolution of a
diastereomeric mixture of acids by use of adamantylamine;
and (3) coupling of the resulting acid with a dipeptide without
590
dx.doi.org/10.1021/op200314z | Org. Process Res. Dev. 2012, 16, 586−594

Organic Process Research & Development
Article
the use of protective groups. For final isolation, the API was
isolated as its tert-butylammonium salt.
(1 L) at a temperature of 0 °C. After 30 min a solution of the 2-
bromo-4′-fluoroacetophenone (662 g, 3.05 mol) in THF (2.5
L) was added over a 3 h period at a reaction temperature of +4
to +7 °C. Fifteen minutes after the addition was complete,
HPLC analysis of the mixture indicated full conversion.
Methanol (400 mL) was added over 10 min with a jacket
temperature set at 0 °C. Due care should be taken during the
quench, as the reaction is effervescent and exothermic! The
temperature was set to +10 °C, and after 45 min the solvent
was removed in vacuo at an inner temperature of 25 °C (the
jacket temperature was 35 °C, and the pressure was 120
mmHg) over a 3 h period to give a total volume of 1.7 L.
Diethyl ether (3 L) was added followed by the addition of
hydrochloric acid (1 M, 1 L) at 20 °C. A precipitate started to
form almost immidiately. The mixture was cooled to +5 °C and
then stirred for another 45 min. The precipitate was removed
by filtration, and the organic phase was separated and washed
once with brine (1.2 L). DMF (2 L) was added, and the ether
was removed in vacuo to give the desired compound as a
solution in DMF which was used as such in the next step. The
yield was assumed to be quantitative. The enantiomeric excess
(ee) was measured to be 99% by chiral HPLC analysis. A small
sample was concentrated to give an analytical sample. 1H NMR
(400 MHz, CDCl3) δ 2.73 (s, br, 1H), 3.50 (dd, 1H), 3.60 (dd,
1H), 4.90 (dd, 1H), 7.03−7.09 (m, 2H), 7.33−7.39 (m, 2H).
Methyl 2-[(2R)-2-(4-Fluorophenyl)-2-hydroxyethyl]-
thioacetate (16). To the crude DMF solution of (1R)-2-
bromo-1-(4-fluorophenyl)ethanol (15) from the previous step
+ 85 g (0.39 mol) of a smaller batch was added methyl
thioglycolate (370 g, 3.48 mol). The mixture was chilled to 0
°C and then the triethylamine (383 g, 3.78 mol) was added
over a 10 min period, while keeping the inner temperature at
approximately 0 °C. Very soon after the addition of
triethylamine was complete, a precipitate started to form.
EXPERIMENTAL SECTION
■
General. All materials were purchased from commercial
suppliers and used as such without further purification. All
reactions were performed under an atmosphere of nitrogen.
Reactions were performed in glass reactors equipped with an
overhead stirrer and connected to an external Huber aggregate.
1
IPCs were recorded either by HPLC or H NMR analysis of
the crude reaction mixture. Determination of the enantiomeric
purity of 15 was made by HPLC on a (R,R) Whelk-01 column
and using heptane/2-propanol (99/1) as eluent. The
diastereomeric purities of the intermediates were recorded
using 1H NMR integration. High resolution mass spectrometry
was performed on a TOF-MS instrument LCT (Waters), mass
precision 5 ppm. NMR measurement was performed using a
Varian Unity spectrometer. HPLC analyses were recorded
using a Waters LC 2000 instrument, column Kromasil C8, 7
μm, eluent MeCN/ammonium acetate buffer. LC/MS analyses
were recorded on a Waters ZMD, LC column x Terra MS C8
(Waters); detection with a HP 1100 MS-detector diode array.
Methyl 2-(4-Formylphenoxy)acetate (2).18 With the mantle
temperature set at 20 °C, 4-hydroxy benzaldehyde (200 g, 1.64
mol) was dissolved in acetone (0.8 L), and potassium carbonate
(278 g, 2.01 mol) was added followed by portionwise addition
of methyl bromoacetate (308 g, 2.01 mol). The mixture was
heated to reflux for 15 h and cooled to 20 °C, followed by
addition of water (1 L) and diethyl ether (1 L). The aqueous
phase (also containing a heavy precipitate, presumably
potassium hydrogen carbonate) was removed, and the organic
phase was washed with more water (2 × 150 mL). By
coevaporation with toluene, the water, acetone, excess methyl
bromoacetate, and ether were removed under reduced pressure,
yielding a white solid (318 g, 1.64 mol). The crude product is
used as such in the next step. The yield is assumed to be
quantitative. 1H NMR (400 MHz, CDCl3) δ 3.77 (s, 3H), 4.69,
(s, 2H), 6.97 (d, 2H), 7.80 (d, 2H), 9.85 (s, 1H). 13C NMR
(150.9 MHz, CDCl3) δ 52.5, 65.2, 115.0, 130.9, 132.1, 162.6,
168.6, 190.8. HRMS (ESI) m/e 193.0503 [(M − H)−, calcd for
C10H9O4 193.0501].
Methyl 2-[4-[(E)-(4-Fluorophenyl)iminomethyl]phenoxy]-
acetate (3). To a solution of the crude methyl 2-(4-
formylphenoxy)acetate 2 (318 g, 1.64 mol) from the previous
step in toluene (1.4 L) was added 4-fluoroaniline (182 g, 1.64
mol), and the mixture was heated to reflux. Water was
azeotropically removed using a Dean−Stark apparatus. When
the aldehyde had been consumed (about 1 h, monitored by
HPLC analysis), the toluene was distilled off under reduced
pressure. To the remaining solid was added acetonitrile (400
mL). Addition of methanol (1.2 L) resulted in precipitation of
the product, which was isolated by filtration. Yield 292 g (62%
over 2 steps). 1H NMR (400 MHz, d6-DMSO) δ 3.72 (s, 3H),
4.90 (s, 2H), 7.06, (d, 2H), 7.19−7.32 (m, 4H), 7.87 (d, 2H),
8.53 (s, 1H). 13C NMR (150.9 MHz, CDCl3) δ 52.5 (d), 65.3,
114.9, 115.9 (d, J = 23 Hz), 122.3 (d, J = 8 Hz), 130.2, 130.6,
148.3 (d, J = 3 Hz), 159.3 (d, J = 1 Hz), 161.0 (d, J = 244 Hz),
160.4, 169.0. MS: m/e 288 [M + H]+. HRMS (ESI) m/e
288.1046 [(M + H)+, calcd for C16H15FNO3 288.1036].
(1R)-2-Bromo-1-(4-fluorophenyl)ethanol (15). Borane di-
methylsulfide complex (135 g, 1.8 mol) was added over a 5 min
period to a solution of the (R)-(+)-2-methyl-CBS-oxazabor-
olidine (157 g of a 1 M solution in toluene, 0.165 mol) in THF
1
Stirring was continued at 0 °C for 15 h. H NMR analysis of
the mixture showed 13% unreacted starting bromide. Cesium
carbonate (61 g, 0.19 mol) was added, and the mixture was
then stirred at 20 °C for an additional 18 h. At this point, the
reaction was complete and ethyl acetate (2.5 L) and water (2
L) were added. After separation of the phases, the organic
phase (∼4 L) was washed with water (2 × 1 L). The pooled
aqueous phases were extracted with extracted with ethyl acetate
(2 × 1 L), and the combined organic extracts were washed with
water (2 × 1 L). All organic layers were combined, dried
(MgSO4), and concentrated to give an oil (930 g). The crude
product containing 11% DMF by weight was used as such in
1
the next step. H NMR (400 MHz, CDCl3) δ 2.81 (dd, 1H),
2.97 (dd, 1H), 3.25 (d, 1H), 3.29 (d, 1H), 3.73 (s, 3H), 4.78
(dd, 1H), 6.99−7.03 (m, 2H), 7.31−7.35 (m, 2H).
Methyl 2-[(2R)-2-[tert-Butyl(dimethyl)silyl]oxy-2-(4-
fluorophenyl)ethyl]thioacetate (17). With a mantle temper-
ature set at 20 °C, to the crude alcohol 16 (840 g, 3.44 mol,
assuming quantitative yield in the previous step) from the
previous step was added DMF (1.25 L) followed by the
addition of imidazole (585 g, 8.6 mol) and TBDMS-Cl (673 g,
4.5 mol). The reaction temperature rose from 18 to 27 °C. Full
conversion (as judged by HPLC analysis) had been obtained
after stirring over the weekend. The mixture was diluted with
MTBE (2.5 L) and then washed with water (2 L). The organic
phase was washed with water (1 L) and concentrated in vacuo
at 40 °C to give the crude product as a pale yellow oil (1250 g).
According to 1H NMR analysis, the crude product was
contaminated with TBDMS-OH as one major impurity (∼60
591
dx.doi.org/10.1021/op200314z | Org. Process Res. Dev. 2012, 16, 586−594

Organic Process Research & Development
Article
1
g calculated from H NMR), giving a corrected yield of 97%.
vigorously at 20 °C for 24 h. At this point the ester hydrolysis
was complete according to HPLC. EtOAc (3.5 L) and water
(3.5 L) were added. The mixture was stirred for 10 min, and
the phases were then separated. The organic phase was
concentrated in vacuo at 30 °C, toluene (2 L) was added, and
1
The mixture was used as such in the next step. H NMR (400
MHz, CDCl3) δ −0.11 (s, 3H), 0.06 (s, 3H), 0.87 (s, 9H), 2.82
(dd, 1H), 2.92 (dd, 1H), 3.12 (d, 1H), 3.18 (d, 1H), 3.71 (s,
3H), 4.80 (dd, 1H), 6.98−7.04 (m, 2H), 7.28−7.33 (m, 2H).
2-[(2R)-2-[tert-Butyl(dimethyl)silyl]oxy-2-(4-fluorophenyl)-
ethyl]thioacetic Acid (18). The crude ester 17 (1250 g) from
the previous step was dissolved in MeOH (1000 mL). The
mixture was cooled to 17 °C (the mantle temperature was set
at 0 °C). NaOH (180 g, 4.5 mol) in MeOH (1300 mL) was
added during 10 min. A 90% conversion to the desired product
had been obtained after 10 min of stirring. The mixture was
allowed to stir at 10 °C for 16 h, after which full conversion had
been obtained. The temperature of the reaction mixture was
increased to 30 °C, and most of the MeOH was distilled off
under reduced pressure until a total volume of ∼2 L was
achieved. Diethyl ether (7 L) and water (5 L) were added. After
separation of the phases, the aqueous phase was acidified to pH
1 using 2.5 M HCl (aqueous, 2 L). Diethyl ether (2 L) was
added. The organic phase was dried (MgSO4) and concen-
trated to give the desired acid as a pale yellow oil (1144 g). The
crude product was contaminated with TBDMS-OH as the main
impurity (∼10 mol %, 4% w/w), giving a corrected yield of
92%. The acid was pure enough to be used as such in the next
1
the solvent was removed in vacuo at 25 °C. H NMR analysis
of the mixture revealed that 1−2% cis-isomer was present in the
crude product. The 50:50 diastereomeric mixture was used as
such in the next step. An analytical sample of the acid could be
obtained through extraction with 1 M HCl/EtOAc followed by
1
concentration in vacuo. H NMR (500 MHz, DMSO-d6) δ
−0.18 and −0.15 (s, 3H), −0.04 and 0.03 (s, 3H), 0.77 and
0.81 (s, 9H), 2.83−3.01 (m, 2H), 4.27 and 4.32 (d, 1H), 4.66
(s, 2H), 4.89−4.99 (m, 1H), 5.00 and 5.09 (d, 1H), 6.91−7.40
(m, 12 H). HRMS (ESI) m/e 600.2062 [(M + H)+, calcd for
C31H36F2NO5SSi 600.2052]. To remove the TBDMS group
from the intermediate acid, AcOH (4.5 L) and water (450 mL)
were added. At a temperature of 10 °C, LiCl (1500 g, 35 mol)
was added portionwise over a period of 10 min. The mixture
was allowed to stir vigorously at 25 °C. After several days of
stirring, the conversion had reached 97.5% by HPLC.19
The temperature of the reaction mixture was lowered to 13
°C, and water (2 L) and MTBE (4 L) were added. The
temperature rose from 13 to 25 °C during this addition. The
organic phase was washed with water (3 × 2 L). The organic
layer was concentrated under reduced pressure at 26 °C and
then coevaporated with isopropanol (2 × 2.5 L), heptane (2.5
L), and finally 2-propanol/toluene (50%, 3 L). This was done
in order to remove most of the AcOH. The crude product 20
was obtained as a brown semicrystalline solid (a 50:50 mixture
of two diastereomeric trans-isomers). The crude product was
used as such in the next step. The material contained
approximately 30 mol % of AcOH.
1
step. H NMR (400 MHz, CDCl3) δ −0.11 (s, 3H), 0.06 (s,
3H), 0.88 (s, 9H), 2.85 (dd, 1H), 2.94 (dd, 1H), 3.17 (d, 1H),
3.23 (d, 1H), 4.83 (dd, 1H), 6.98−7.04 (m, 2H), 7.28−7.33
(m, 2H). MS: m/e 343 [M − H]−. HRMS (ESI) m/e 343.1189
[(M − H)−, calcd for C16H24FNO3SSi 343.1199].
Methyl 2-[4-[(2RS, 3RS)-3-[(2R)-2-[tert-Butyl(dimethyl)-
silyl]oxy-2-(4-fluorophenyl)ethyl]thio-1-(4-fluorophenyl)-4-
oxo-azetidin-2-yl]phenoxy]acetate (trans-19). Under an
atmosphere of nitrogen, a 10 L reactor was charged with the
imine 3 (963 g, 3.35 mol), CH2Cl2 (3 L), and triethylamine
(969 g, 9.58 mol) followed by the addition of 2-chloro-1-
methylpyridinium iodide (856 g, 3.35 mol) in one portion. The
mantle temperature was set at +23 °C, and the crude acid 18
(1100 g from above) dissolved in CH2Cl2 (3 L) was slowly
added via syringe pump at a rate of approximately 5 mL/min.
The mixture turned red-brown immediately after the addition
of acid 18 had started. The temperature of the reaction mixture
was maintained at 23 °C during the whole reaction. After
approximately 16 h, all the acid 18 had been added. HPLC
analysis of the mixture indicated full conversion of the starting
acid, 18. 2 M HCl (aqueous 3.5 L) was added. After separation,
the organic phase was extracted with NaHCO3 (aqueous
saturated, 3 L), brine (2.5 L), and water (2 L). The organic
phase was then concentrated to give the title compound (2018
g, brown viscous oil) as a 50:50 mixture of two diastereomeric
trans-isomers. The compound was pure enough to be used as
such in the next step. The main impurities were the aldehyde
1-Adamantylammonium; 2-[4-[(2R,3R)-1-(4-Fluorophen-
yl)-3-[(2R)-2-(4-fluorophenyl)-2-hydroxyethyl]thio-4-oxoaze-
tidin-2-yl]phenoxy]acetate (21). To the crude acid 20 above
was added acetone (3.9 L) and methanol (2.2 L). The solution
was heated to 27 °C, and adamantan-1-amine (200 g, 1.3 mol)
dissolved in methanol (400 mL) was added over 1 min to give a
solution with a temperature of 30 °C. Stirring was halted,
seeding crystals were added, and a temperature gradient was set
from 35 to 23 °C over 3 h with no stirring. The mixture was
allowed to stand for 18 h at 23 °C.12 A sample of the crystals
was taken which showed a diastereomeric ratio of 80:20. The
supernatant was allowed to drain off, and the crystals were
subsequently suspended in 40% methanol in acetone (4 L).
Stirring at 30 °C for 45 min was followed by filtration (without
previous cooling). The diastereomeric ratio was now 90:10.
The crystals were again suspended in 40% methanol in acetone
(2 L), and the mixture was stirred at 50 °C for 1.5 h. The warm
mixture was filtered, and the solid collected was dried overnight
under reduced pressure at 55 °C to give 377 g (0.78 mol, 40%
effective yield over three steps) of the chemically pure salt,
1
derived from the imine and 1-methylpyridone. H NMR (400
MHz, CDCl3) δ −0.13 and −0.10 (s, 3H), 0.01 and 0.08 (s,
3H), 0.84 and 0.88 (s, 9H), 2.85−3.13 (m, 2H), 3.83 (s, 3H),
3.84−3.86 (m, 1H), 4.58−4.63 (m, 1H), 4.66 (s, 2H), 4.83−
4.93 (m, 1H), 6.90−7.02 (m, 6H), 7.20−7.35 (m, 6H).
2-[4-[(2RS,3RS)-1-(4-Fluorophenyl)-3-[(2R)-2-(4-fluoro-
phenyl)-2-hydroxyethyl]thio-4-oxoazetidin-2-yl]phenoxy]-
acetic Acid (20). The crude ester 19 (1400 g) was dissolved in
acetonitrile (4.5 L) and triethylamine (1.5 L, 10.8 mol)
followed by the addition of water (260 mL). The mixture was
chilled to 13 °C, and then lithium chloride (500 g, 11.8 mol)
was added in one portion. The mixture was then stirred
1
diastereomeric ratio 92:8. H NMR (500 MHz, d6-DMSO) δ
1.50−1.64 (m, 6H), 1.72 (d, 6H), 2.00−2.04 (m, 3H), 2.88−
2.95 (m, 2H), 4.15 (s, 2H), 4.24 (d, 1H), 4.26 (d, 1H, cis-
isomer), 4.71 (t, 1H), 5.01 (d, 1H, cis-isomer), 5.03 (d, 1H),
6.79−6.82 (m, 2H), 7.09−7.17 (m, 4H), 7.22−7.36 (m, 6H).
13C NMR (150.9 MHz, DMSO-d6) δ 28.4, 35.2, 38.8, 40.2,
50.2, 58.4, 62.0, 67.4, 71.3, 114.7 (d, J = 14 Hz), 114.9, 116.0
(d, J = 22 Hz), 118.80 (d, J = 8 Hz), 127.2, 127.6, 128.0 (d, J =
8 Hz), 133.4 (d, J = 2 Hz), 140.4 (d, J = 3 Hz), 158.4 (d, J =
241 Hz), 159.3, 161.3 (d, J = 242 Hz), 163.8, 170.7.
592
dx.doi.org/10.1021/op200314z | Org. Process Res. Dev. 2012, 16, 586−594

Organic Process Research & Development
Article
2-[4-[(2R,3R)-1-(4-Fluorophenyl)-3-[(2R)-2-(4-fluorophen-
yl)-2-hydroxyethyl]thio-4-oxo-azetidin-2-yl]phenoxy]acetic
Acid (14). The “free acid” 14 was obtained through extraction
of the salt 21 with EtOAc and 0.2 M HCl and washing of the
organic layer with water followed by drying (MgSO4) and
concentration of the organic layer to a white solid
(diastereomeric ratio 92/8 of the two trans-isomers). 1H
NMR (600 MHz, d6-DMSO) δ 2.88−2.95 (m, 2H), 4.27 (d,
1H), 4.67 (s, 2H), 4.70−4.74 (m, 1H), 5.08 (d, 1H), 5.64−5.68
(m, 1H), 6.92−6.94 (m, 2H), 7.09−7.39 (m, 10H). HRMS
(ESI) m/e 484.1035 [(M − H)−, calcd for C25H20F2NO5S
484.1030].
6.98−7.54 (m, 13H), 8.34 (t, 1H). 13C NMR (150.9 MHz, d6-
DMSO) δ 25.8, 25.9, 26.2, 27.6, 32.5, 33.3, 33.7, 38.9, 40.0,
40.7, 41.9, 50.1, 52.0, 58.4, 61.9, 66.8, 71.3, 114.7 (d, J = 21.1
Hz), 115.2, 116.0 (d, J = 22.7 Hz), 118.8 (d, J = 8.1 Hz), 128.0
(d, J = 7.9 Hz), 128.9, 133.3 (d, J = 2.3 Hz), 140.4 (d, J = 2.8
Hz), 158.0, 159.2 (d, J = 379.8 Hz), 160.5 (d, J = 443.5 Hz),
163.7, 167.1, 167.7, 174.7. HRMS (ESI) m/e 696.2568 [(M +
H)+, calcd for C36H40F2N3O7S 696.2555].
Methyl (2R)-2-[[2-(tert-butoxycarbonylamino)acetyl]-
amino]-3-cyclohexylpropanoate (22). N-BOC-Glycine (84.5
g, 0.48 mol) and N-methylmorpholine (146 g, 1.44 mol) were
dissolved in CH2Cl2 (500 mL). TBTU (168 g, 0.52 mol) was
added, and the mixture was allowed to stir at 23 °C for 1 h.
(R)-Cyclohexylalanine methylester hydrochloride (107 g, 0.48
mol)20 was added, and the mixture was stirred for 1 h at 23 °C.
After a wash with water (300 mL), the organic layer was filtered
and then concentrated under reduced pressure. The residue
was triturated with heptane (500 mL), and the resulting
suspension was then stirred for 1 h with ice-cooling. Filtration
of the mixture and washing of the solid with heptane gave after
tert-Butylammonium; (2R)-3-Cyclohexyl-2-[[2-[[2-[4-
[(2R,3R)-1-(4-fluorophenyl)-3-[(2R)-2-(4-fluorophenyl)-2-
hydroxyethyl]thio-4-oxoazetidin-2-yl]phenoxy]acetyl]-
amino]acetyl]amino]propanoate (1) (AZD4121). The acid 14
(122 g, 0.25 mol), 4-chlorophenol (65 g, 0.35 mol), and
triethylamine (53 g, 0.53 mol) were dissolved in DMF (1 L).
The solution was cooled to −15 °C. TBTU (107 g, 0.33 mol)
was added in three portions over a 10 min period. After stirring
for 16 h at −15 °C, an HPLC analysis of the mixture revealed
that there was still 15% starting material left in the mixture.
Another portion of TBTU (10 g) was added. The temperature
was set to 0 °C, and after a further 1.25 h, more TBTU (9 g)
was added. After 1.5 h, full conversion to the intermediate p-
chlorophenyl ester 24 had been obtained. The dipeptide 23 (63
g, 0.28 mol) was added in one portion followed by the addition
of lithium chloride (160 g, 3.77 mol). The mantle temperature
was then set to 30 °C. After 4 days of stirring, full conversion to
the acid 13 had been obtained. The temperature was lowered to
16 °C, and ethyl acetate (550 mL) and water (550 mL) were
added. The aqueous phase was extracted with EtOAc (3 × 250
mL). The pooled organic phases were washed with 0.5 M
hydrochloric acid (0.5 L), brine (250 mL), and, last, water (250
mL). Drying (MgSO4), filtration through Celite, and removal
of the solvent in vacuo at 25 °C gave 280 g of crude 13 as a pale
brown oil (contaminated with 4-chlorophenol, tetramethylurea,
and DMF). This was dissolved in acetone (1.5 L). To this
solution was added tert-butylamine (18 g, 0.25 mol). The
solution was left for 24 h at 35 °C without any proper
precipitation observed. Therefore, most of the solvent was
removed in vacuo to leave a syrup which was washed with
MTBE (2 × 300 mL) to remove most of the DMF and 4-
chlorophenol. The washed syrup was partitioned between ethyl
acetate (500 mL) and 0.5 M hydrochloric acid (500 mL). The
aqueous phase had a pH of 2. The phases were separated, and
the ethyl acetate phase was washed with water (2 × 250 mL)
and brine (2 × 100 mL), followed by drying (MgSO4) and
concentration to give a viscous oil. This was dissolved in
acetone (1.2 L) and heated to 35 °C, and then tert-butylamine
(16.7 g, 0.22 mol) was added. The mixture was left without
stirring with a mantle temperature of 35 °C for 46 h. The
suspension formed was filtered, and the solid collected was
washed with acetone (500 mL), followed by drying of the salt
to give 1 (88 g, 0.11 mol, 44% yield from 14). The mother
liquor was purified through preparative HPLC followed by
crystallization as described above to give an additional 20 g (26
mmol, an additional 10%) of 1. Purity by HPLC (210 nm) =
1
drying 140 g (85%) of the desired product. H NMR (300
MHz, CDCl3) δ 0.82−1.85 (m, 13H), 1.45 (s, 9H), 3.72 (s,
3H), 3.77 (dd, 1H), 3.85 (dd, 1H), 4.59−4.69 (m, 1H), 5.12−
5.26 (m, 1H), 6.52 (d, 1H).
(2R)-2-[[2-Aminoacetyl]amino]-3-cyclohexylpropanoate
(23). The ester 22 (140 g, 0.41 mol) was dissolved in MeOH
(600 mL). NaOH (20 g, 0.5 mol) dissolved in water (50 mL)
was added, and the mixture was allowed to stir overnight, after
which full hydrolysis of the ester had been obtained. The
solution was neutralized with CO2(s) followed by concen-
tration of the mixture. Trace of water was azeotropically
removed with formic acid (50 mL). To the residue was added
formic acid (500 mL), and the mixture was allowed to stir at 40
°C for 17 h. The formic acid was removed under reduced
pressure followed by addition of water (500 mL). The pH of
the solution was adjusted to ∼5 using 25% NH3 solution. The
suspension formed was stirred at 23 °C for 6 h followed by
filtration of the mixture and washing of the solid with water
(100 mL). The wet solid (110 g) was suspended in hot (50 °C)
acetone (400 mL), and the mixture was then allowed to attain
23 °C followed by filtration of the mixture and washing of the
solid with acetone (200 mL). After drying under reduced
pressure, the title compound was obtained as a white solid
1
(67.3 g, 72%). H NMR (300 MHz, DMSO-d6) δ 1.30−2.35
(m, 13H), 4.44 (d, 1H), 4.51 (d, 1H), 5.03−5.10 (m, 1H),
12.02 (br s, 4H). 13C NMR (126 MHz, CD3COOD) δ 27.5,
27.7, 27.9, 33.7, 35.1, 35.6, 40.7, 42.7, 52.5, 168.8, 178.4. MS:
m/e 229 [M + H]+. HRMS (ESI) m/e 229.1557 [(M + H)+,
calcd for C11H21N2O3 229.1552].
ASSOCIATED CONTENT
■
S
* Supporting Information
1H NMR spectra (compounds 1, 2, 3, 14, 15, 16, 17, 18, 19,
21, 22, and 23), 13C NMR spectra (compounds 1, 21, and 23),
HPLC purity analysis (compound 1), LC-MS analysis (during
syntheses of 15, 19, and the ester hydrolysis product of 19),
chiral HPLC analysis (15), and HRMS (1, 2, 3, 14, 18, the
ester hydrolysis product of 19, 23, and the major contaminant
in 1). This material is available free of charge via the Internet at
http://pubs.acs.org.
1
96%, >95:5 diastereomeric purity. H NMR (500 MHz, d6-
DMSO) δ 0.70−0.87 (m, 2H), 1.03−1.15 (m, 3H), 1.20 (s,
9H), 1.25−1.38 (m, 2H), 1.47−1.77 (m, 6H), 2.92 (dd, 1H),
2.95 (dd, 1H), 3.70 (dd, 1H), 3.76 (dd, 1H), 3.92−3.99 (m,
1H), 4.29 (d, 1H), 4.53 (s, 2H), 4.72 (t, 1H), 5.07 (d, 1H),
593
dx.doi.org/10.1021/op200314z | Org. Process Res. Dev. 2012, 16, 586−594

Organic Process Research & Development
Article
(14) About 4% intermolecular dimerization through ester formation
was observed. Reaction between 4-chlorophenol and TBTU gave rise
to small amounts of 2-(4-chlorophenyl)-1,1,3,3-tetramethylisouronium
byproduct also.
AUTHOR INFORMATION
Corresponding Author
*E-mail: henrik.sorensen@astrazeneca.com.
■
(15) Purification of the mother liquor by preparative HPLC and
conversion to the tert-butylammonium salt followed by crystallization
raised the yield of the last three steps to 54%.
Notes
The authors declare no competing financial interest.
1
(16) Based on H NMR measurement in DMSO-d6.
ACKNOWLEDGMENTS
■
(17) The high resolution MS and fragmentation patterns agree with
the drawn structure within 5 ppm. See Supporting Information.
(18) Hoogendoorn, S.; Blom, A. E. M.; Willems, L. I.; van der Marel,
G. A.; Overkleeft, H. S. Org. Lett. 2011, 13 (20), 5656−5659.
(19) Halfway through the procedure it was noted that by applying an
intermediate vacuum pressure the reaction mixture became a clear
solution and the speed of deprotection was increased.
(20) Raynham, T. M.; Hammonds, T. R.; Gilliatt, J. H.; Charles, M.
D.; Pave, G. A.; Foxton, C. H.; Carr, J. L.; Mistry, N. S. U.S. Patent
2009/0247519 A1, Oct 1, 2009, p 142.
We thank our Separation Science Laboratory group for
recorded analytical data and for determination of the
enantiomeric purity of the chiral alcohol 15, Dr. Gustaf Hulthe
for mass spectrometric assistance, our NMR group for
assistance and recording of certain spectra, and Ingemar Starke
for valuable discussions.
REFERENCES
■
(1) (a) Rosenblum, S. B.; Huynh, T.; Afonso, A.; Davis, H. R.;
Yumibe, N.; Clader, J. W.; Burnett, D. A. J. Med. Chem. 1998, 41, 973−
980. (b) Burnett, D. A. Curr. Med. Chem. 2004, 11, 1873−1887.
(c) Castaner, R. M.; Sorbera, L. A.; Castaner, J. Drugs Future 2000, 25
̃
̃
(7), 679−85.
(2) (a) Wu, G.; Tormos, W. J. Org. Chem. 1997, 62, 6412−6414.
(b) McKittrick, B. A.; Ma, K.; Huie, K.; Yumibe, N.; Davis, H.; Clader,
J. W.; Czarniecki, M. J. Med. Chem. 1998, 41, 752−759.
(c) Thiruvengadam, T. K.; Tann, C.-H.; Lee, J.; McAllister, T.;
Sudhakar, A. U.S. Patent 5306817, Apr 26, 1994. (d) Michalak, M.;
Stodulski, M.; Stecko, S.; Mames, A.; Panfil, I.; Soluch, M.; Furman, B.;
Chmielewski, M. J. Org. Chem. 2011, 76, 6931−36. (e) Wu, G. G. Org.
Process Res. Dev. 2000, 4, 298−300.
(3) Alenfalk, S.; Dahlstrom, M.; Hunegnaw, F.; Karlsson, S.;
̈
Lemurell, M.; Lindqvist, A.-M.; Skjaret, T.; Starke, I. PCT Int. Appl.
̈
WO 2005061452 A1, July 7, 2005.
(4) The strategy was demonstrated to be practical on scale for
ezetimibe; see ref 5.
(5) Sasikala, C. H. V. A.; Padi, P. R.; Sunkara, V.; Ramayya, P.;
Dubey, P. K.; Uppala, V. B. R.; Praveen, C. Org. Process Res. Dev. 2009,
13, 907−910.
(6) (a) Kawanami, Y.; Murao, S.; Ohga, T.; Kobayashi, N.
Tetrahedron 2003, 59, 8411−8414. (b) Hett, R.; Stare, R.; Helquist,
P. Tetrahedron Lett. 1994, 35 (50), 9375−9378. (c) Corey, E. J.;
Shibata, S.; Bakshi, R. K. J. Org. Chem. 1988, 53, 2861−2863.
(7) (a) Amin, S. G.; Glazer, R. D.; Manhas, M. S. Synthesis 1979,
210−213. (b) Huang, H.; Iwasawa, N.; Mukaiyama, T. Chem. Lett.
1984, 8, 1465−66. (c) Georg, G. I.; Mashava, P. M.; Guan, X.
Tetrahedron Lett. 1991, 32 (5), 581−584. (d) Matsui, S.; Hashimoto,
Y.; Saigo, K. Synthesis 1998, 1161−1166. (e) Donati, D.; Morelli, C.;
Porcheddu, A.; Taddei, M. J. Org. Chem. 2004, 69, 9316−9318.
(8) The stereochemical outcome of this reaction has been subject to
several studies. See: (a) Hegedus, L. S.; Montgomery, J.; Narukawa, Y.;
Snustad, D. C. J. Am. Chem. Soc. 1991, 113 (15), 5784−5791.
(b) Cossio, F. P.; Ugalde, J. M.; Lopez, X.; Lecea, B.; Palomo, C. J. Am.
Chem. Soc. 1993, 115 (3), 995−1004. (c) Lop
́
ez, R.; Sordo, T. L.;
ez, J. J. Org. Chem. 1993, 58, 7036−7037.
Sordo, J. A.; Gonzal
́
(d) Arrieta, A.; Lecea, B; Cossío, F. P. J. Org. Chem. 1998, 63, 5869−
5876. (e) Liang, Y.; Jiao, L.; Zhang, S.; Yu, Z.-X.; Xu, J. J. Am. Chem.
Soc. 2009, 131 (4), 1542−1549.
(9) Martinez, A. G.; Barcina, J. O.; del Veccio, G. H.; Hanack, M.;
Subramanian, L. R. Tetrahedron Lett. 1991, 32 (42), 5931.
(10) Nicolaou, K. C.; Estrada, A. A.; Zak, M.; Lee, S. H.; Safina, B. S.
Angew. Chem., Int. Ed. 2005, 44 (9), 1378−82.
(11) The conditions for the hydrolysis were further optimized and
published. See: Mattsson, S.; Dahlstrom, M.; Karlsson, S. Tetrahedron
̈
Lett. 2007, 48 (14), 2497−2499.
(12) It was found that crystallization under nonstirring conditions
was paramount for a successful resolution of the diastereomers.
(13) Attempts were made using 3,4-dichlorophenyl-, 4-nitrophenyl-,
phenyl-, O-succinimide ester, cyanide catalyzed amide coupling of a
methyl ester, and TBTU coupling.
594
dx.doi.org/10.1021/op200314z | Org. Process Res. Dev. 2012, 16, 586−594