Synthesis of Akt inhibitor ipatasertib. part 2. Total synthesis and first kilogram scale-up

Article abstract of DOI:10.1021/op500270z

Herein, the first-generation process to manufacture Akt inhibitor Ipatasertib through a late-stage convergent coupling of two challenging chiral components on multikilogram scale is described. The first of the two key components is a trans-substituted cyc

Full text of DOI:10.1021/op500270z

Article
pubs.acs.org/OPRD
Synthesis of Akt Inhibitor Ipatasertib. Part 2. Total Synthesis and First
Kilogram Scale-up
Travis Remarchuk,*,† Frederic St-Jean,† Diane Carrera,† Scott Savage,† Herbert Yajima,† Brian Wong,†
Srinivasan Babu,† Alan Deese,† Jeffrey Stults,† Michael W. Dong,† David Askin,† Jonathan W. Lane,‡
and Keith L. Spencer‡,§
†Small Molecule Process Chemistry, Genentech, Inc., a member of the Roche Group, 1 DNA Way, South San Francisco, California
94080-4990, United States
‡Array BioPharma Inc., 3200 Walnut Street, Boulder, Colorado 80301, United States
S
* Supporting Information
ABSTRACT: Herein, the first-generation process to manufacture Akt inhibitor Ipatasertib through a late-stage convergent
coupling of two challenging chiral components on multikilogram scale is described. The first of the two key components is a
trans-substituted cyclopentylpyrimidine compound that contains both a methyl stereocenter, which is ultimately derived from the
enzymatic resolution of a simple triester starting material, and an adjacent hydroxyl group, which is installed through an
asymmetric reduction of the corresponding cyclopentylpyrimidine ketone substrate. A carbonylative esterification and
subsequent Dieckmann cyclization sequence was developed to forge the cyclopentane ring in the target. The second key chiral
component, a β2-amino acid, is produced using an asymmetric aminomethylation (Mannich) reaction. The two chiral
intermediates are then coupled in a three-stage endgame process to complete the assembly of Ipatasertib, which is isolated as a
stable mono-HCl salt.
In Part 1 of this series,2 the route scouting and early process
development efforts focused on the identification of a new
INTRODUCTION
■
Retrosynthetically, Ipatasertib is comprised of two chiral
components, the cyclopentylpyrimidine core trans-1 and β2-
racemic route for the synthesis of a challenging cyclo-
amino acid (S)-2 (Figure 1).1 There are several key challenges
pentylpyrimidine compound (R)-12 used in the preparation
of trans-1. In the new strategy, dihydroxypyrimidine rac-53 was
identified as a key intermediate that originated from a simple
triester starting material bearing the requisite methyl stereo-
center (rac-4). This intermediate set the stage for a downstream
Dieckmann cyclization and decarboxylation of the resulting β-
keto ester to afford ketone rac-12 in seven steps from rac-5.4
The focus of the work reported herein entails the
development of an asymmetric route to DHP (R)-5 from the
enzymatic resolution of precursor triester rac-4. The conversion
of intermediate (R)-5 into ketone (R)-12 on multikilogram
scale was subsequently achieved in high enantiopurity. From
the chiral ketone (R)-12, alcohol trans-1 was obtained via an
asymmetric Noyori transfer hydrogenation using a chiral
ruthenium catalyst.5
For preparation of the second chiral component, the β2-
amino acid (S)-2, an asymmetric aminomethylation (Mannich)
reaction using Evans’s chiral auxiliary was initially employed to
access the desired target.6 In the discovery synthesis, the
method was successful in terms of providing enantiomerically
enriched compound (S)-2. However, this chemistry was not
immediately amenable to scale-up due to the use of cryogenic
temperatures and silica gel chromatography for purification,
where the product was isolated as a syrup. Herein we describe
an improved process using the Evans’s chiral auxiliary approach
Figure 1. Retrosynthetic analysis for the synthesis of Ipatasertib.
to the synthesis and scale-up of Ipatasertib: (1) the source and
control of the chiral purity of the starting materials are critical;
(2) impurities formed during the coupling of the two chiral
starting materials must be controlled, and (3) poor thermal
stability of the free-base form of Ipatasertib leads to a β-
elimination degradation product, which necessitates develop-
ment of a stable mono-HCl salt form.
Received: August 20, 2014
© XXXX American Chemical Society
A
dx.doi.org/10.1021/op500270z | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
that enabled the production of kilogram-scale quantities of
crystalline and enantiomerically pure β2-amino acid (S)-2.
The convergent endgame synthesis of Ipatasertib com-
menced with the initial deprotection of the tert-butyloxycar-
bonyl (Boc)-protected trans-1 core followed by amide bond
formation with β2-amino acid (S)-2 to produce the N-Boc-
protected API ((S,R,R)-22). Although the use of the
guanidinium coupling reagent HBTU7 was shown to be
effective in providing the amide product without detectable
epimerization, the byproducts generated were challenging to
remove and required the development of alternative isolation
conditions. The final deprotection step to remove the N-Boc
group with HCl initially afforded Ipatasertib di-HCl, which was
the final form employed in early biological investigations.
However, hygroscopicity issues associated with Ipatasertib di-
HCl and the unstable nature of the free-base form of the API
led to the development of a stable mono-HCl salt form of
Ipatasertib that will be described herein.
acylation reaction to access triester (R)-4,9 whereas the second
approach involved a more practical enzymatic resolution of
triester rac-4 and is the approach described below.10
Exploration to access triester (R)-4 commenced with the
screening of various hydrolytic lipases with the goal of
identifying one that would lead to both high enantioselectivity
and high conversion.11 From the list of representative examples
shown in Table 1, the commercially available lipase from
Candida rugosa (entry 6) and lipase AYS “Amano” (entries 7
and 8) were shown to give (R)-4 in 92−99% ee after extended
reaction time (5−12 d). The enantioselectivity increased over
time with both of these enzymes and is best illustrated by entry
6. Increasing the enzyme loading from 20 to 30 wt% in the case
of the lipase AYS “Amano” reduced the resolution time from 7
days to 5 days (entry 8) without compromise to the
enantioselectivity.12 The isolated yields for both wild-type
lipases were modest (50%), but reproducible.
In a separate study, a mutant form of Chromobacterium lipase
(entry 9) was identified to be highly effective in the resolution
of rac-4. Specifically, treatment of rac-4 with 11 wt% of enzyme
at 0 °C gave triester (R)-4 in >99% ee and 79% yield after only
1 day, and this method was subsequently chosen for our first
production batch of DHP (R)-5. On scale-up of this process, a
mixture of triester (R)-4 (27.6 kg) and SC-YM-1 enzyme (2.8
kg, 10 wt%) in phosphate buffer (pH 7.5) was agitated at 0 °C
until the reaction was deemed complete (chiral purity >98%
ee) after ∼19 h (Scheme 1). The triester product was extracted
from the basic aqueous reaction mixture with MTBE and
concentrated under reduced pressure to give (R)-4 (12.9 kg,
99.8% ee) in 94% yield. Following the protocol established in
the preceding Part I of this series, the crude triester (R)-4 (12.9
kg) was used directly in the next step and reacted with
formamidine acetate and NaOMe/MeOH to forge the DHP
ring system. The product was precipitated from HCl in MeOH
to give DHP (R)-5 (8.7 kg, >99.9% ee, and >99 A%13) in 74%
yield.14
RESULTS AND DISCUSSION
■
1. Preparation of 2,6-Dihydroxypyrimidine (R)-5. Our
first objective was to secure an efficient and scalable route to
the substituted dihydroxypyrimidine (R)-5 that was previously
prepared only in racemic form through the condensation of
triester rac-48 with formamidine acetate in 87% yield on 500 g
scale.2 Our efforts focused on developing a scalable route to
enantiomerically pure triester (R)-4 (eq 1). Two approaches
were initially investigated to produce kilogram quantities of 2,6-
dihydroxypyrimidine (R)-5. The first approach originated from
the commercially available (R)-monomethyl 3-methylgluarate
(MGM) substrate that involved a challenging cryogenic
2. Preparation of Ketone (R)-12 from DHP (R)-5. The
seven-step synthesis of key intermediate ketone (R)-12 from
DHP (R)-5 is illustrated in Scheme 2, where several
a
Table 1. Lipase Screen for Resolution of Triester rac-4
b
temp
(°C)
ee of 4
(%)
entry
lipase
wt equiv
time (h)
yield
1
2
3
4
Candida antarctica (Novozym 435)
porcine pancreas
0.1 w/w
0.1 w/w
1.5 w/w
0.3 w/w
10−15
30−35
40−45
26−30
36
18
36
36
0
0
nd
nd
nd
nd
lipase PS “Amano” SD
8
protease Bacillus licheniformis (alcalase-2.4L,
Novozyme Co.)
−82
5
6
pig liver esterase (ICR-123, BioCatalytics/
Codexis)
0.02 w/w
0.1 w/w
10−15
40−45
24
20
nd
c
Candida rugosa
i. 40
ii. 110
iii. 284
170
i. 52
ii. 76
iii. 92
98
12.5 g (50% from 50 g of rac-4); 98.9 A% GC purity
c
7
8
9
lipase AYS “Amano”
0.2 w/w
0.3 w/w
0.11 w/w
30−35
30−35
0
2.5 g (50% from 10 g of rac-4); 99.3 A% GC purity
c
lipase AYS “Amano”
127
99
2.5 g (50% from 10 g of rac-4); 99.2 A% GC purity
Chromobacterium strain SC-YM-1
23
>99
4.49 g (79%)
a
The reactions were performed by adding lipase to a mixture of rac-4 in phosphate buffer (10 vol) and maintaining the pH 7.0−7.2 during the
reaction at the specified temperature (see eq 1). At the end of the reaction, the pH was adjusted to pH 3−4 with aqueous 2 N HCl, extracted with
b
MTBE, filtered, concentrated, and analyzed for A% purity by GC and chiral purity by HPLC. Enantiomeric excess (ee) was measured by HPLC:
Chiralpak IC column (4.6 × 150 mm, 5 μm); mobile phase = n-Hexane/EtOH (9:1); flow rate =1.0 mL/min; column temp = 23 °C; inj vol = 10
c
μL; detector wavelength = 210 nm; approx RRT of (S)-4 = 1.07. GC conditions: DB-5 column (30 m × 0.25 mm, 0.25 μm); flow rate = 3.0 mL/
min; inj vol = 0.5 μL; inj temp = 250 °C; detector temperature = 280 °C; split ratio = 1:50; oven program = 60 °C (5 min); ramp to 250 °C (40
min) and hold at 250 °C (7 min). nd = not determined.
B
dx.doi.org/10.1021/op500270z | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Scheme 1. Enzymatic Resolution Approach to DHP (R)-5
Scheme 2. First Kilogram Scale-up of Ketone (R)-12 from DHP (R)-5
modifications were made to the initial process to render the
sequence amenable to scale-up.15 Starting with 5.0 kg of DHP
(R)-5, the dichlorination reaction using POCl3 and 2,6-lutidine
was performed in toluene instead of acetonitrile. The higher
boiling point of toluene improved process safety on increased
scale since the POCl3 addition was exothermic, and the reagent
had to be added as rapidly as possible to maintain good stirring.
A second advantage of using toluene was that it allowed for a
more efficient extractive aqueous work-up.16 Using this
procedure, the solid dichloropyrimidine (DCP) product (R)-
6 (5.63 kg) was produced in excellent yield and purity (95%
yield, >98 A%). The subsequent SNAr reaction of (R)-6 with
Boc-piperazine was performed in MeOH at 50 °C in the
presence of diisopropylethylamine (DIPEA) as base, and the
crude product (R)-7 was isolated in 96% yield with 97 A%
purity. Higher reaction temperatures (>50 °C) would increase
the rate of the SNAr reaction but would also favor the formation
of a byproduct, arising from the double addition of N-Boc
piperazine to (R)-6, which was only observed in trace amounts
with the current conditions. The next step, involving methyl
ester hydrolysis with LiOH in THF/H2O, was uneventful when
performed at ambient temperature (23 °C); however, at higher
reaction temperatures (up to 55 °C), increased hydrolysis of
the chloropyrimidine was observed, resulting in production of
the corresponding hydroxypyrimidine carboxylic acid deriva-
tive. On scale-up, hydrolysis of methyl ester (R)-7 (8.63 kg)
was complete after stirring at room temperature for ∼15 h, and
the product carboxylic acid (R)-8 (6.88 kg, 80% yield) was
obtained in two crops after trituration of the isolated solids in
MTBE. Re-esterification to the benzyl ester (R)-9 was required
for efficient decarboxylation in the downstream chemistry, as
described in Part 1 of this series.17 The use of benzyl bromide
and Cs2CO3 in DMF gave quantitative conversion to benzyl
ester (R)-9, which was isolated as a viscous oil with >98 A%
purity.18
3. Carbonylation of Chloropyrimidine (R)-9. In order to
generate the Dieckmann diester substrate (R)-10, the carbon-
ylative esterification of the chloropyrimidine group of (R)-9
was performed using Pd(OAc)2/dppp and K2CO3 in 3:1 i-
PrOH/THF at ∼55 °C with 55 psi carbon monoxide.19 Early
findings revealed that thorough degassing to remove oxygen
and efficient stirring were the two key factors that affected the
yield of this reaction.20 In a representative experiment (1.2 kg
(R)-9, ∼50 psi CO, 50 °C), the in-process control (IPC) by
HPLC after 14 h indicated ∼75% conversion to diester (R)-10
with significant (∼40 psi) gas uptake. Following a single re-
pressurization of the reactor with CO (50 psi), the reaction
eventually proceeded to completion. The reaction was filtered
over a pad of silica gel to remove the Pd catalyst, oxidized
ligand, and inorganic salts, affording the crude diester (R)-10
(∼1.3 kg, 95 A%). Several batches of benzyl ester (R)-9 were
processed under identical reaction conditions to produce a total
of 9.3 kg of crude diester (R)-10 with 98 A% purity in near-
quantitative conversion.
4. Telescoped Dieckmann Cyclization and Decarbox-
ylation Reactions. The key step in the synthesis of ketone
(R)-12 is the Dieckmann cyclization of diester (R)-10, followed
by decarboxylation (eq 2). Although this two-step process was
C
dx.doi.org/10.1021/op500270z | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
In addition to the catalyst, the solvent also played an
important role in minimizing the formation of compound 1. On
the laboratory scale, use of THF in the Dieckmann cyclization
and subsequent use of MeOH for the decarboxylation (formic
acid, 10 wt% Pd/C) resulted in complete consumption of
intermediate (S)-11 after ∼4 h, where 18% of the over-reduced
product 1 was observed (Table 2, entry 7). When 2-MeTHF
was employed for both the Dieckmann cyclization and the
decarboxylation under otherwise similar reaction conditions,
the decarboxylation reaction was complete in ∼3 h, and the
crude product ketone (R)-12 was isolated in 73% yield
containing only 3% of over-reduced compound 1 (Table 2,
entry 8). Unfortunately, further scale-up (8 kg) of the two-step
reaction sequence in 2-MeTHF as solvent still resulted in 15%
compound 1 (Table 2, entry 9) in the crude ketone product
(R)-12 after isolation. The higher level of over-reduced product
formed on larger scale was attributed to the difficulty
experienced in terminating the hydrogenation after the
decarboxylation was deemed complete.24 Although we planned
to address the over-reduction issue before any future scale-up of
ketone (R)-12 using this route, the presence of the alcohol
mixture of 1 did not interfere with the next step (Noyori
asymmetric ketone reduction), and the crude ketone could be
used without further purification. Overall, starting with 5 kg of
DHP (R)-5, a total of 5 kg crude ketone (R)-12 (85 A%,
containing 14:1 cis/trans-1) was produced in 54% isolated yield.
5. Noyori Asymmetric Transfer Hydrogenation of
Ketone (R)-12. Reduction of the ketone group in (R)-12 in
the presence of a non-chiral catalyst is intrinsically cis-selective.
For example, as illustrated in Table 2 (entry 6), transfer
hydrogenation of (R)-12 with Pd/C gave 88% de in favor of cis-
1.25 As was first exemplified in the medicinal chemistry route,
to access the desired trans diastereomer, use of a chiral catalyst
is necessary (Scheme 3).26 In that report, transfer hydro-
genation of ketone (R)-12 with Noyori’s [(R,R)-TsDPEN−
Ru(p-cymene)Cl]27 catalyst gave trans-1 in up to 91% de.
These results were obtained using 0.01 equiv of catalyst loading
with 1.2 equiv of HCOOH and 1.0 equiv of of Et3N as the
source of hydrogen in DCM.28 Since our ketone substrate (R)-
12 obtained from the Dieckmann cyclization−decarboxylation
chemistry already contained over-reduced alcohol 1, we
decided to retain the Noyori catalyst system in order to satisfy
the tight timelines associated with this program.29 Purification
first demonstrated in the preparation of rac-12 and proceeded
in 91% yield with >99 A% purity, the original procedure was
found to be problematic upon scale-up. The Dieckmann
cyclization reaction, promoted by potassium tert-butoxide base,
proceeded uneventfully to afford the β-keto ester intermediate
(S)-11; however, decarboxylation of the benzyl ester group was
much more challenging.21 The main competing side reaction
was over-reduction of the ketone group in compound 12 to the
corresponding hydroxyl derivative 1, where the undesired cis-
diastereomer was favored as the major byproduct (up to ∼17:1
cis/trans).22 This undesired pathway proved to be difficult to
control as the scale of the reaction was increased. The over-
reduction product was less pronounced when Pd/alumina was
used as catalyst, and this reaction typically gave ≤5 A%
compound 1 (Table 2, entry 1). Unfortunately, the reduced
reactivity of this catalyst often led to incomplete decarbox-
ylation and low yields of ketone (R)-12. Conversely, the more
active Pd/C catalyst successfully facilitated complete decarbox-
ylation but resulted in more over-reduction product that
favored cis-1 (Table 2, entry 2).23 Various sources of hydrogen
were examined with the Pd/C catalyst system, including
ammonium formate, methyl cyclohexadiene, and formic acid,
the latter of which proved to be superior with respect to
product yield (Table 2, entries 2−4). For the development and
scale-up of this part of the telescoped reaction, the Pd/C
catalyst was used, as it was deemed more important to achieve
complete decarboxylation than to suppress over-reduction. This
preference was a result of the difficulties experienced in the
purification of trans-1 with unreacted starting material present.
However, under stress conditions, the hydrogenation could
result in complete over-reduction to the alcohol product, as
observed when using 50 wt% of 5% Pd/C catalyst over 30 h
(Table 2, entry 6). Thus, it was important to carefully monitor
the decarboxylation and stop the hydrogenation reaction as
soon as intermediate (S)-11 was consumed.
Table 2. Conditions for Decarboxylation of β-Keto Ester (S)-11 to Crude Ketone (R)-12 (See Eq 2)
scale
time
(h)
cis/trans-1 (A%) in crude
a
a
entry (R)-10 (g)
catalyst
source of H2
(yield); HPLC purity of (R)-12
(R)-12
1
2
20
17
2 wt% Pd/alumina
40 psi H2
72
6
5 g (40%); 91.6 A%
3.0 g (28%); 51.1 A%
4
10% Pd/C (10 wt%)
5% ammonium formate in i-PrOH
(5 vol)
11
3
17
10% Pd/C (10 wt%)
10 equiv of methylcyclohexadiene in
i-PrOH (5 vol)
5
3.4 g (32%); 99.8 A%
nd
4
5
6
17
25
50
10% Pd/C (10 wt%)
5% Pd/C (20 wt%)
5% Pd/C (50 wt%)
5% HCOOH in i-PrOH (5 vol)
5% HCOOH in MeOH (5 vol)
5% HCOOH in MeOH (5 vol)
4
2
4.4 g (41%); 99.8 A%
7.7 g (49%); 99.5 A%
none isolated
nd
nd
30
20 g of 1 (63%);
94:6 cis/trans
7
8
9
470
10% Pd/C (10 wt%)
10% Pd/C (15 wt%)
10% Pd/C (15 wt%)
5% HCOOH in MeOH (5 vol)
5% HCOOH in 2-MeTHF (5 vol)
5% HCOOH in 2-MeTHF (5 vol)
4
3
4
360 g crude; 54.6 A%
18% (17:1 cis/trans)
3% (2:1 cis/trans)
15% (14:1 cis/trans)
b
110
50 g crude (70%); 97.6 A%
5040 g crude (84%); 85.3 A%
b
8125
a
Measured by HPLC: ACE-3-C18, 4.6 × 150 mm, 3 μm; mobile phase A = 20 mM ammonium formate, pH 3.7; mobile phase B = MeOH; Gradient
= 30−80% B (0−15 min), 80−90% B (15−16 min), 90−30% B (16−16.1 min), 30% B (16.1−20 min); flow rate = 0.8 mL/min; column temp = 30
b
°C; inj vol = 10 μL; detector wavelength = 280 nm; approximate RRT of cis-1 = 0.94 and RRT of trans-1 = 0.97. nd = not determined. In these
experiments, 2-MeTHF was used in place of THF for the Dieckmann cyclization.
D
dx.doi.org/10.1021/op500270z | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Scheme 3. Asymmetric Ketone Reduction, Derivatization, and Hydrolysis to trans-1
Scheme 4. Scale-up of the Chiral Auxiliary Route to β2-Amino Acid (S)-2
of trans-1 following the asymmetric transfer hydrogenation was
implemented in order to reduce the levels of cis-1 diastereomer
to meet the desired specification (≤2 A%) for this intermediate.
In the production effort, the transfer hydrogenation of crude
ketone (R)-12 containing 15% of 1 (5.04 kg, 85 A%) was
divided into three batches of ∼1.7 kg each. A DCM solution of
the crude ketone mixture was charged with 1 mol% [(R,R)-
MsDPEN−Ru(p-cymene)Cl] catalyst [S/C = 263:1] and
HCO2H/Et3N (1.2:1.0)30 and resulted in complete carbonyl
reduction after 20 h at room temperature. For each batch, crude
1 (1.7 kg) was isolated by concentration under reduced
pressure to an oil−solid matrix as a mixture of trans/cis
diastereomers (76% de) by HPLC (Scheme 3).
The purification of the crude trans-1 mixture ultimately
required separation of the two diastereomers; however,
crystallization was not efficient when trans-1 was <95% de.31
As a result, the hydroxyl group of 1 was acylated in order to
provide a less polar compound that could be separated
efficiently using silica gel chromatography. Additional research
revealed that the pivalate ester derivative afforded the best
separation of the corresponding diastereomers in comparison
to the p-nitrobenzoate and acetate derivatives.32 Therefore, the
crude alcohol mixture was initially derivatized with pivaloyl
chloride, after which the diastereomers were separated by silica
gel chromatography to provide a total of 5.37 kg of ester trans-
14 with 98 A% purity.33 Hydrolysis of the ester in trans-14 gave
a total of 3.8 kg trans-1 (98.9% de, >99.9% ee, 98.2 A%).34 In
summary, trans-1 was obtained in 48% overall yield from 5.0 kg
DHP (R)-5 in nine chemical steps, providing the first of two
starting materials for the convergent synthesis of Ipatasertib.
6. Preparation of β2-Amino Acid (S)-2. The scale-up
process to access β2-amino acid (S)-2 using Evans’s chiral
auxiliary is illustrated in Scheme 4.35 For preparation of acyl
oxazolidinone derivative (R)-17,36 the carboxylic acid 15 was
activated via the mixed anhydride method with pivaloyl
chloride. High conversion of product was achieved by the
portionwise addition of the acid chloride reagent to a mixture
of the carboxylic acid, chiral auxiliary, and Et3N followed by
heating to reflux. In the pilot plant, the coupling of 4-
chlorophenylacetic acid 15 (17.4 kg) and (R)-4-benzyloxazo-
lidin-2-one 16 proceeded to ∼92% conversion, and product
(R)-17 (17.4 kg) was isolated in 62% yield with 92.5 A% purity
by HPLC. The subsequent asymmetric aminoalkylation
reaction was performed using a protocol similar to that
reported by Evans and co-workers, excepting modifications to
the reaction temperature and concentration.37 Specifically,
good conversion and diastereoselectivity were achieved at a
higher reaction temperature (−20 vs −78 °C) and increased
concentration (21 vs 30 vol), effectively improving the
convenience of the process. In the process, a DCM solution
of acyl auxiliary (R)-17 (17.4 kg) was initially treated with 1 M
TiCl4 in toluene at −20 °C, resulting in the bright red-orange
titanium complex typically reported. Addition of DIPEA then
led to the characteristic dark purple titanium enolate species,
E
dx.doi.org/10.1021/op500270z | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Scheme 5. Telescoped One-Pot Synthesis of Boc-API (S,R,R)-22
which was subsequently treated with N-Boc aminal 18.
Importantly, the slow addition of 18 and maintaining the
internal temperature ≤ −20 °C were essential for good
diastereoselectivity, and in-process HPLC analysis indicated
complete consumption of (R)-17 after 4 h.38 Following
aqueous work-up and solvent exchange into THF, the
alkylation product (S,R)-19 was produced in >20:1 dr.
Hydrolysis of the auxiliary group was achieved by addition of
lithium hydrogen peroxide, and the reaction was complete after
∼3 h. After acidic aqueous work-up followed by extraction of
the product into toluene, a solution of crude-2 with ∼66 A%
purity was obtained.39
concentration to dryness via rotary evaporation.42 Unfortu-
nately, the solid product was extremely hygroscopic and
difficult to handle outside of an inert environment, and as a
result we focused our development efforts on a telescoped
process to the Boc-API (S,R,R)-22 where piperazine (R,R)-21
was not isolated. By using HCl in EtOH in the deprotection
step, the concentrated mixture at the end of the reaction was
found to be acceptable to carry forward into the coupling
reaction. The same coupling protocol that we initially
developed for the two-step reaction sequence involving
HBTU, DIPEA, and DCM was shown to be equally effective
in the one-pot process. Using the HBTU reagent resulted in
complete conversion in the coupling reaction with no
detectable epimerization of the β-amino acid moiety, thus
other coupling reagents were not thoroughly investigated at this
stage of the project.
On scale-up of the one-pot process to Boc-API (S,R,R)-22,
deprotection of trans-1 (3.49 kg) using 3 equiv of 2.5 M HCl in
EtOH and toluene as co-solvent resulted in complete
consumption of starting material after ∼1 h at 60 °C (Scheme
5). Distillation and a co-evaporation with toluene afforded a
suspension of (R,R)-21 di-HCl that was then diluted with
DCM and charged with excess DIPEA to give dissolution of the
salt mixture. Next, addition of a DCM solution of β-amino acid
(S)-2 (1 equiv) followed by the HBTU reagent at room
temperature resulted in complete consumption of starting
material (S)-2 after ∼3 h.
Attempts to purify crude (S)-2 through the extractive
removal of the oxazolidinone auxiliary and other reaction
byproducts from the basic aqueous media resulted in significant
product loss. Alternatively, a purification strategy relying on the
Boc-deprotection of the amino group in crude (S)-2 to form a
highly aqueous-soluble amino acid sodium salt and subsequent
extraction of the organic impurities was successful. A solution of
crude (S)-2 was treated with 12 N HCl to give the
corresponding deprotected product (S)-20. After toluene
extraction, the acidic aqueous layer was adjusted to pH 9
with aqueous NaOH resulting in the precipitation of product.
The filtered solids were dried to give (S)-20-Na (7.65 kg, 60%
yield, 99.5 A%, 97.8% ee). Since the product contained 1.1% of
(R)-2, a recrystallization was required in order to avoid
formation of diastereomers in the downstream chemistry
during the coupling with trans-1. Recrystallization of (S)-20-
Na from 2-propanol provided the product (S)-20-Na (6.85 kg,
The aqueous work-up and purification of the (S,R,R)-22
intermediate was critical in controlling the final purity of
Ipatasertib, which was not crystalline in either the free-base or
various salt forms. Unexpectedly, the most difficult impurities
to remove were found to originate from the HBTU coupling
reagent itself that produced stoichiometric amounts of
tetramethylurea (TMU) and hexafluorophosphate anion
99.7% ee, and 99.9 A%) in 56% overall yield from (R)-17.40
A
final Boc-reprotection of (S)-20-Na followed by precipitation
from MTBE/heptane with continuous sparging of nitrogen
resulted in 99% isolated yield of β2-amino acid (S)-2.41 The
overall process produced a total of 8.9 kg of (S)-2 in >99.9% ee
and >99.9 A% purity as a free-flowing solid that would be used
in the kilogram-scale synthesis of Ipatasertib.
7. First-Generation Process Route to Ipatasertib
Mono-HCl. With trans-1 and β2-amino acid (S)-2 in hand, it
remained to first couple these key chiral components to
produce the penultimate Boc-API (S,R,R)-22, and then to
convert this compound into Ipatasertib mono-HCl. In the first
step of the process, removal of the N-Boc protecting group in
trans-1 could be accomplished using either aqueous HCl, or
HCl in EtOH along with toluene as co-solvent at elevated
temperatures. Both reaction conditions required ≥3 equiv of
the HCl in order to obtain complete deprotection to the
piperazine (R,R)-21. In our early process chemistry batches, the
piperazine (R,R)-21 di-HCl salt was generated using 12 N HCl
and was isolated as a solid by co-evaporation with toluene and
43
−
(PF6 ). Washing the organic reaction mixture sequentially
with saturated aqueous NaHCO3 and saturated aqueous NH4Cl
effectively removed the HOBt from the organic phase;
however, these washes were ineffective in removing the TMU
and the PF6 anion.44 We found that replacing DCM with
isopropyl acetate (IPAc) and washing with copious amounts of
water allowed for complete removal of TMU from the crude
Boc-API (S,R,R)-22. The PF6 anion in turn required the use of
a basic aqueous solution, and dilute aqueous NH4OH was
found to be the most efficient for coordinating and extracting
the anion.45 Having successfully removed the HBTU and PF6-
related impurities, the crystallization of crude Boc-API (S,R,R)-
22 could be performed; however, due to the high solubility of
(S,R,R)-22 in IPAc, an efficient crystallization from this solvent
could not be developed. Alternatively, a seeded crystallization
F
dx.doi.org/10.1021/op500270z | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Scheme 6. Preparation of Ipatasertib Mono-HCl Salt from Boc-API (S,R,R)-22
from MTBE/heptane gave the best results where 5.06 kg of the
amide product was produced in 81% yield with 99.6 A% purity.
What remained at this stage was cleavage of the N-Boc group
in (S,R,R)-22 and subsequent formation of Ipatasertib mono-
HCl. In our early-stage development work, we found that the
Boc-group deprotection also required ≥3 equiv of HCl (similar
to the deprotection of trans-1) and resulted in Ipatasertib di-
HCl upon isolation. This behavior is attributed to the two basic
sites in the molecule as illustrated by the two pKa values shown
in Scheme 6. Since this di-HCl salt was hygroscopic, non-
crystalline, and difficult to isolate, it was not suitable for
formulation development. In contrast, we also attempted to
isolate the less-hygroscopic free-base form of the API by way of
concentration of the DCM solution under reduced pressure
followed by drying at elevated temperatures (≤75 °C) to
remove residual solvents. In this approach, we unexpectedly
observed the formation of a major impurity that was identified
as β-elimination product 23.46 Noteworthy, both the Boc-API
(S,R,R)-22 and Ipatasertib di-HCl compounds did not produce
the β-elimination impurity 23 when dried at ∼100 °C for >24
h, which strongly suggested that the isopropyl amino group
must be protected or protonated in order to prevent thermal
degradation.47
safely dried at temperatures ≤80 °C over several days without
risk of degradation.
On scale-up of the salt formation, the deprotection of Boc-
API (S,R,R)-22 (4.69 kg) in toluene with 12 N HCl was
complete after ∼1 h. The reaction mixture was diluted with
water, and the toluene layer containing any unreacted starting
material and other organic impurities was removed. The
aqueous layer was then basified to pH ≥ 12 with aqueous
NaOH in order to extract the Ipatasertib free-base with DCM.
After treatment of the DCM solution with charcoal and
SiliaMetS thiol to remove colored impurities and trace heavy
metals, filtration over Celite resulted in a 99% assay yield of
Ipatasertib free-base. Next, the DCM solution was charged with
0.96 equiv of 12 N HCl, after which azeotropic distillation
followed by solvent exchange to EtOAc produced a suspension
of Ipatasertib mono-HCl. Filtration and drying of the solids in a
filter dryer at ≤70 °C for ∼48 h resulted in Ipatasertib mono-
HCl (3.23 kg, 99.5 A%) in 80% yield. The relative and absolute
stereochemistry of Ipatasertib mono-HCl were successfully
confirmed by single-crystal X-ray analysis (Figure 2).
Our salt selection strategy initially aimed to take advantage of
the higher pKa of the basic isopropylamino group (pKa of
conjugate acid ∼9.5) that was absent in impurity 23.48 This
would provide a method to remove the elimination impurity 23
in the event that quantities were formed during the free-base
step. In a laboratory demonstration, the free-base API
(containing ∼1 A% of 23) dissolved in DCM, was treated
with ≤0.95 equiv of 12 N HCl resulting in a homogeneous
biphasic solution. Azeotropic distillation with DCM followed
by co-evaporation with EtOAc resulted in white slurry that was
easily filtered and was characterized as Ipatasertib mono-HCl.
This protocol resulted in complete purging of impurity 23 and
provided a thermally stable mono-HCl salt form that could be
Figure 2. X-ray single-crystal structure of Ipatasertib mono-HCl.
G
dx.doi.org/10.1021/op500270z | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
the aqueous layer was extracted with MTBE (14 kg) and then
filtered. The aqueous was extracted once more with MTBE (14
kg), and then the organics were combined, washed with 5%
aqueous NaHCO3 (14 kg), and concentrated under reduced
pressure to give triester (R)-4 (12.9 kg, 94% yield, 99.8% ee) as
a yellow oil. The crude triester product was used directly in the
next step.
CONCLUSION
■
Herein we described the first successful multikilogram-scale
convergent synthesis of the Akt inhibitor Ipatasertib mono-
HCl. Preparation of a challenging cyclopentyl pyrimidine core
trans-1 was accomplished in 10 chemical steps from chiral DHP
(R)-5. A Dieckmann cyclization−decarboxylation sequence was
developed to forge the cyclopentyl ketone ring system of (R)-
12, which was followed by an asymmetric transfer hydro-
genation with Noyori’s Ru-diamine catalyst to produce 3.8 kg
of trans-1 with 98.9% de in 48% overall yield. Synthesis of the
β2-amino acid (S)-2 utilized the asymmetric Mannich reaction,
where Evans’s chiral oxazolidinone auxiliary was employed to
set the relevant stereochemistry with >20:1 diastereoselectivity.
This subsequently provided our first scale-up batch of β2-amino
acid (S)-2 (8.9 kg) with >99% ee. Following the coupling of the
two key components, we identified a crucial process to convert
the Boc-API (S,R,R)-22 intermediate into the thermally stable
mono-HCl salt form. The new process led to the first successful
production of cGMP Ipatasertib mono-HCl (3.23 kg) in high
purity, providing drug substance to support the Phase 1 clinical
demand.
(R)-Methyl 3-(4,6-Dihydroxypyrimidin-5-yl)butanoate
(DHP), 5. A reactor was charged with MeOH (51.6 kg) and
formamidine acetate (6.1 kg, 58.7 mol, 1.05 equiv) and cooled
to 0 °C with agitation. A 28% NaOMe/MeOH solution (32.2
kg, 167 mol, 3.9 equiv) was added over 1.5 h and then agitated
for an additional 1 h. A solution of triester (R)-4 (12.9 kg, 55.6
mol, 1.0 equiv) in MeOH (12.9 kg) was charged and the
reaction mixture warmed to 25 °C. After 10 h, the reaction
mixture was cooled to 0 °C, neutralized with aqueous HCl, and
then concentrated under reduced pressure. The residue was
diluted with MeOH, washed with MTBE, and acidified with
aqueous HCl. The mixture was warmed to 60 °C, agitated for 1
h, cooled to 0 °C, and then filtered, and the solids were washed
with H2O to give DHP (R)-5 (8.7 kg, 74% yield, 99.6 A%,
99.8% ee (220 nm) and 99.9% ee (254 nm)) as an off-white
solid. HRMS calcd for C9H12N2O4 [M+H]+ 213.087; found
213.0869. See Supporting Information S2 for the HPLC
chromatograms of DHP (R)-5 in Figure S2−2.6.
EXPERIMENTAL SECTION
■
NMR measurements were carried out on a Bruker Avance 3,
500 MHz spectrometer equipped with a 5 mm, BBO, Z-
gradient probe with TMS internal standard. Data for H NMR
(R)-Methyl 3-(4,6-Dichloropyrimidin-5-yl)butanoate
(DCP), 6.3 A 50 gallon reactor was charged with DHP (R)-5
(5.0 kg, 23.6 mol [99.8% ee, 99.6 A%; KF = 0.2% H2O]),
toluene (20 L), and 2,6-lutidine (2.7 L, 23.6 mol, 1 equiv) with
agitation at room temperature. The mixture (white slurry) was
warmed to 50 °C, and phosphorous oxychloride (POCl3) (4.85
L, 53.02 mol) was added slowly over 1 h. The first quarter of
POCl3 added resulted in a gum-like residue and was exothermic
(50 to 85 °C) for the first half of the addition. Upon compete
addition of POCl3, a clear brown solution was formed, and
good agitation was possible. The reaction temperature was
maintained at 70 °C for 24 h. IPC of the reaction mixture by
TLC analysis (1:1 EtOAc/heptane) [Rf (R)-6) = 0.6, Rf (R)-5)
= 0.1] indicated complete consumption of starting material.
The reaction mixture was cooled to 0 °C and slowly charged
with 20% aqueous NaOH (5.5 kg/28 L H2O) [Caution:
Exothermic!] until a final pH ∼5.5 was achieved. The internal
temperature during the addition was maintained at ≤30 °C.
EtOAc (13 L, 2.5 vol) was charged to the mixture, which was
then stirred for 30 min and filtered over Celite, and then the
filter was rinsed with EtOAc (2 × 5 L). The organic layer was
removed, and the aqueous layer was extracted with EtOAc (3 ×
12 L). The combined organics were washed with saturated
aqueous NH4Cl (3 × 12 L) and brine (12 L), dried (MgSO4),
filtered through a glass fiber filter, and then concentrated under
reduced pressure to give crude DCP (R)-6 (5.63 kg, ∼95 wt%,
98.7 A%) as a brown oil. HRMS calcd for C9H10Cl2N2O2 [M
+H]+ 249.0192; found: 249.0190. Crystals formed upon
standing overnight, and the crude product (contains residual
2,6-lutidine) was used directly in the next step. Note:
approximately 0.4 A% of the monochlorination adduct was
observed in the isolated product as determined by LCMS (M
1
are recorded as follows: chemical shift (ppm), multiplicity (s,
singlet; bs, broad singlet; d, doublet; t, triplet; q, quartet; h,
heptet; dd, doublet of doublet; m, multiplet), integration,
coupling constant (Hz). Data for 13C NMR are reported in
terms of chemical shift (ppm). Trace metal analysis was
performed by Intertek Pharmaceutical services using inductively
coupled plasma atomic emission spectroscopy (ICP-AES).
Melting points were measured by differential scanning
calorimetry (DSC). All reactions were performed under an
inert atmosphere. Unless otherwise noted, all chemicals were
commercially available and used as received. SiliaMetS thiol
(40−63 um, 60 Å) was purchased from Silicycle, and activated
carbon was Darco G60, 100 mesh. Palladium acetate, 99.95%,
was purchased from Kadai Technology Limited; N-Boc-
piperazine was prepared at SAI Advantium; 1,3-bis-
(diphenylphosphino)propane (dppp), 98%, was purchased
from Strem; 4-chlorophenylacetic acid was purchased from
Aldrich; and (R)-4-benzyl-2-oxazolidinone auxiliary, 99.0%
[102029-44-7], was purchased from Jiangxi Kingnord Industrial
Ltd. The chiral catalysts [(R,R)-MsDPEN−RuCl(p-cymene)],
and [(R,R)-TsDPEN−Ru(p-cymene)Cl] were purchased from
Johnson Matthey (Catalysis and Chiral Technologies). O-
Benzotriazole-N,N,N′,N′-tetramethyluronium hexafluorophos-
phate (HBTU) was purchased from Oakwood Products, Inc.
Analytical methods are found in the Supporting Information
S1.
(R)-Trimethyl 2-Methylpropane-1,1,3-tricarboxylate,
4. An aqueous solution of Na2HPO4 (4.5 kg in 184 kg H2O)
with pH adjusted to 7.5 was charged with triester rac-4 (27.6
kg, 119 mol) and Cromobacterium strain SC-YM-1 (2.8 kg, 10
wt%) at 0 °C. The reaction mixture was agitated at this
temperature and the pH maintained between 7.0−7.6 until the
reaction was deemed complete (∼19 h) by HPLC (chiral
purity: (R)-4 >98% ee). The reaction mixture was charged with
MTBE (55 kg) and Radiolite filter aid (0.8 kg), agitated, and
then filtered. The organic layer of the filtrate was removed, and
1
+H = 231 amu). See Supporting Information S2 for the H
NMR spectrum of DCP (R)-6 (Figure S2−3.1) and HPLC
chromatogram in Figure S2−3.2.
(R)-tert-Butyl 4-(6-Chloro-5-(4-methoxy-4-oxobutan-
2-yl)pyrimidin-4-yl)piperazine-1-carboxylate, 7. A 50
gallon reactor was charged with Boc-piperazine (4.63 kg,
H
dx.doi.org/10.1021/op500270z | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
24.84 mol, 1.1 equiv) and MeOH (20 L. 3.6 vol) with agitation
at room temperature until the solids dissolved. A solution of
DCP (R)-6 (5.63 kg, 22.58 mol, 1.0 equiv) in MeOH (15 L, 2.7
vol) and then DIPEA (4.34 L, 24.84 mol, 1.1 equiv) were
charged to reaction mixture and warmed to 50 °C. Analysis by
TLC (1:1 EtOAc/heptane) [Rf (R)-7 = 0.4, Rf (R)-6 = 0.6]
indicated complete consumption of starting material after 4 h.
The reaction mixture was cooled to room temperature and
concentrated under reduced pressure to a light brown residue.
The residue was dissolved in EtOAc (50 L), washed with
saturated NH4Cl (2 × 20 L) and brine (20 L), dried (Mg2SO4),
filtered, and concentrated to give crude (R)-7 (8.68 kg, 96%
yield, 97.4 A%) as a brown viscous oil. HRMS calcd for
C18H27ClN4O4 [M+H]+ 399.1793; found 399.1784. The crude
product was used directly in the next step.
(R)-3-(4-(4-(tert-Butoxycarbonyl)piperazin-1-yl)-6-
chloropyrimidin-5-yl)butanoic Acid, 8. A 50 gallon reactor
was charged with a solution of (R)-7 (8.63 kg, 21.62 mol) in
THF (36.5 L) with agitation at room temperature. A solution
of LiOH·H2O (2.72 kg, 64.87 mol, 3 equiv) dissolved in H2O
(19 L) was charged and the mixture agitated at room
temperature overnight. Analysis by TLC (1:1 EtOAc/heptane)
[Rf (R)-8 = 0.1, Rf (R)-7 = 0.4] indicated complete
consumption of starting material after 15 h. The reaction
mixture was cooled to 0 °C and then slowly charged with 6 N
HCl (11.5 L, 1.3 vol) over 1 h to achieve pH 2−3 while
maintaining the internal temperature ≤20 °C. The mixture was
charged with EtOAc (35 L) and stirred vigorously for 30 min,
and then the layers were separated. The aqueous layer was
extracted with EtOAc (3 × 35 L), and the organics were
combined, washed with saturated NH4Cl (2 × 45 L) and brine
(40 L), dried over Mg2SO4, filtered, and then concentrated to a
light pink solid. The solids were slurried in MTBE (15 L),
agitated for 3−4 h at room temperature, and filtered, and the
wet cake was washed with MTBE (2 × 5 L). The solids were
dried under vacuum at 30 °C to give carboxylic acid (R)-8 (1st
crop: 6.0 kg, 72.1% yield, 98.6 A%).
A second crop of product was obtained by concentration of
the filtrates under reduced pressure to a yellow residue followed
by addition of a solution of 10:1 MTBE/DCM (2.2 L) with
agitation at room temperature overnight. The precipitated
solids were filtered and washed with MTBE (2 × 0.5 L), the
filtrate was concentrated again, and the process was repeated
using 10:1 MTBE/DCM (1.1 L) and MTBE wash (3 × 0.4 L).
The solids were combined and dried under vacuum at 30 °C to
give a second crop of carboxylic acid (R)-8 (2nd crop: 0.88 kg,
10.5% yield, 96.6 A%); HRMS calcd for C17H25ClN2O2 [M
+H]+ 385.1637; found: 385.1631. See Supporting Information
S2 for the 1H NMR spectrum (CDCl3) of carboxylic acid (R)-8
(1st crop) in Figure S2−3.5 and the HPLC chromatogram
(270 nm) of the second crop in Figure S2−3.6.
(R)-tert-Butyl 4-(5-(4-(Benzyloxy)-4-oxobutan-2-yl)-6-
chloropyrimidin-4-yl)piperazine-1-carboxylate, 9. A 22
L, three-neck round-bottom flask equipped with a mechanical
stirrer was charged with carboxylic acid (R)-8 (2.75 kg, 7.15
mol, 1.0 equiv), DMF (13.0 L), and benzyl bromide (1.28 kg,
7.50 mol, 1.05 equiv) with agitation at room temperature.49
The solution was charged with Cs2CO3 (2.45 kg, 7.50 mol, 1.05
equiv) in three portions at a rate to maintain the internal
temperature <40 °C (exotherm observed from 20 to 35 °C).
Analysis by TLC (1:1 EtOAc/heptane) [Rf (R)-9 = 0.4, Rf (R)-
8 = 0.2] indicated complete consumption of starting material
after 15 h. The mixture was filtered over a pad of Celite 545 (1
kg) and washed with EtOAc (3 × 2 L). The filtrate was diluted
with EtOAc (12 L) and washed with saturated aqueous NH4Cl
(8 L), and the aqueous part was then extracted with EtOAc (3
× 6 L). The combined organics were washed with saturated
aqueous NH4Cl (2 × 8 L) and brine (10 L), dried over
Mg2SO4, filtered, and concentrated under reduced pressure to
give benzyl ester (R)-9 (3.42 kg, 98.6 A%, >theoretical yield) as
a viscous brown oil. Four batches of compound (R)-8 (7.27 kg
total) were processed to give a total of 9.06 kg benzyl ester (R)-
9. See Supporting Information S2 for a representative 1H NMR
spectrum (CDCl3) of benzyl ester (R)-9 in Figure S2−3.7, and
HPLC chromatogram (270 nm) in Figure S2−3.8.
(R)-Isopropyl 5-(4-(Benzyloxy)-4-oxobutan-2-yl)-6-(4-
(tert-butoxycarbonyl)piperazin-1-yl)pyrimidine-4-carb-
oxylate, 10. [Caution: Carbon monoxide gas (CO) is poisonous!
A CO detector must be used and the reaction performed in a well
ventilated hood. All equipment and connections were leak-tested
with nitrogen gas and soap solution before use.] A 5-gallon
autoclave was charged with crude benzyl ester (R)-9 (1.71 kg,
3.60 mol), THF (4.0 L), and 2-propanol (6.0 L) with agitation
at room temperature. The solution was sparged with nitrogen,
and the vessel was pressurized (25 psi) five times with heating
to 35−40 °C. The mixture was then charged with a slurry of
Pd(OAc)2 (0.081 kg, 0.36 mol, 0.10 equiv) and dppp (0.163 kg,
0.40 mol, 0.11 equiv) in 2-propanol (2.0 L) and warmed to
40−45 °C with good agitation. The autoclave was pressurized
and evacuated with 2 × 25 psi N2, stirred for 20 min 40−45 °C
to dissolve the rust-red solids, and then charged with a slurry of
K2CO3, 325 mesh (0.299 kg, 2.16 mol, 0.6 equiv) in 2-propanol
(2.0 L). The autoclave was pressurized, evacuated with 5 × 25
psi N2 and carbon monoxide (CO) gas 5 × 40 psi, and then
pressurized to 55 psi. The reaction mixture was agitated at 50
°C and maintained at 55 psi CO for ≥50 h. The reaction was
analyzed by TLC and HPLC and sampled by slow evacuation
of CO gas and sparging with N2 5 × 25 psi. Analysis by TLC
(1:1 EtOAc/heptane) [Rf (R)-10 = 0.5, Rf (R)-9 = 0.4] and
HPLC [Rf (R)-10) = 12.10 min, Rf (R)-9 = 12.67 min]
indicated complete consumption of starting material after 50 h.
The reaction mixture was cooled to room temperature, charged
with silica gel (0.4 kg) and Celite 545 (0.4 kg), and then
agitated for 2 h. The mixture was filtered over silica gel (1.0 kg)
and washed with EtOAc (4−5 L) until product was no longer
observed in the filtrate. The filtrates were concentrated under
reduced pressure (bath temp ∼45 °C) to give diester (R)-10
(1.9 kg, >100% yield; contains residual solvents and dppp
ligand) as a viscous dark brown oil. A total of six batches of
benzyl ester (R)-9 were processed to produce 10 kg crude
diester (R)-10 (see Supporting Information for the lot
summary information).
For the purification of diester (R)-10, six batches of diester
(R)-10 were combined, dissolved in a total of 18 L MTBE, and
then slurried with silica gel (3 kg) with agitation for 1 h. This
slurry was charged in a 2-L sintered funnel and rinsed with
EtOAc. The mixture was filtered through a pad of silica gel and
washed with MTBE (∼10 L). The filtrate was concentrated
under reduced pressure to give pure diester (R)-10 (9.3 kg,
92.8%, 97.7 A%) as viscous dark yellow oil. Note: Plug filtration
over a pad of silica gel removes dppp[O] that can otherwise
interfere with the ruthenium catalyzed asymmetric ketone
1
reduction. See Supporting Information S2 for the H NMR
spectrum (CDCl3) of diester (R)-10 in Figure S2−3.9, and
HPLC chromatogram (270 nm) in Figure S2−3.10.
I
dx.doi.org/10.1021/op500270z | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
(R)-tert-Butyl 4-(5-Methyl-7-oxo-6,7-dihydro-5H-
cyclopenta[d]pyrimidin-4-yl)piperazine-1-carboxylate,
12. A reactor was charged with diester (R)-10 (8.13 kg, 15.4
mol), 2-MeTHF (82 L, 10 vol), sparged with nitrogen (20
min), and cooled to 0 °C with agitation. Under an inert
atmosphere, potassium tert-butoxide (1.91 kg, 17.0 mol, 1.1
equiv) was charged slowly (exothermic) in two portions to the
solution of the diester while maintaining the internal temper-
ature ≤5 °C. IPC by TLC analysis (1:1 EtOAc/heptane) [Rf
(S)-11 = 0.1, Rf (R)-10 = 0.5] indicated complete consumption
of starting material after 30 min. The reaction mixture was
cooled to −5 to 0 °C, and formic acid, 98% (1.1 kg, 23.1 mol,
1.5 equiv), was charged at a rate to maintain internal
temperature <5 °C and obtain a final pH 6−7, resulting in a
yellow slurry. A separate vessel was charged with 10 wt% Pd/C
[50% wet] (1.2 kg, 15 wt%) and 5% formic acid (2.1 L) in 2-
MeTHF (20 L), which was then transferred to the reaction
mixture of intermediate (S)-11 at a rate to maintain the internal
temperature <10 °C. The vessel was rinsed with 2-MeTHF (20
L), and then the reaction mixture was slowly warmed to ∼20
°C over 1 h. The reaction was monitored by TLC analysis (9:1
EtOAc/MeOH) [Rf (R)-10 = 0.4, Rf (S)-11 = 0.1, Rf (R)-12 =
0.3], which indicated complete consumption of starting
material (11) and formation of the over-reduced alcohol 1.
The reaction mixture was slowly quenched with saturated
aqueous NaHCO3 (60 L) at a rate to maintain minimum
foaming. The reaction mixture was filtered over Celite 545 (4
kg) and rinsed with 2-MeTHF (50 L). The filtrate was
extracted with EtOAc (50 L), and the organics were washed
with saturated aqueous NaHCO3 (2 × 60 L). The combined
aqueous part (∼120 L) was extracted with EtOAc (100 L), and
the organics were combined and then washed with brine (100
L). The agitated organics were charged with charcoal (3 kg, 35
wt%) and Mg2SO4 (7.5 kg, 90 wt%). The mixture was filtered
and the solution concentrated under reduced pressure at <45
°C to a minimum stir volume. The mixture was charged with
identical conditions to give a total of ∼5.1 kg of crude-1
containing an 88:12 trans/cis mixture of diastereomers.51
Derivatization of Crude-1 and Purification of Pivalate
Ester trans-14. A 50 L reactor containing a solution of crude-1
(2.5 kg, 7.5 mol) in CH2Cl2 (25 L) was charged with pivaloyl
chloride (0.99 kg, 8.22 mol, 1.1 equiv) followed by DIPEA
(1.16 kg, 8.97 mol, 1.2 equiv) at room temperature. [Caution:
Exothermic!] The reaction was monitored by TLC analysis (9:1
EtOAc/MeOH) Rf (trans-14) = 0.8, Rf (trans-1) = 0.2,
indicated complete consumption of starting material (trans-1)
after 2 h at room temperature. The reaction mixture was
quenched with saturated aqueous NaHCO3 (12 L), and the
layers were separated. The aqueous was extracted with CH2Cl2
(2 × 2 L), and then the combined organics were washed with
saturated aqueous NH4Cl (12 L), dried (Mg2SO4), filtered, and
concentrated under reduced pressure to give crude-14 (∼3.1
kg) as a dark brown oil-solid. An additional reaction of
equivalent size was performed to give a combined total of ∼6.2
kg of crude-14 with 75:25 trans/cis by HPLC. Purification by
silica gel plug filtration was performed by dissolving 2.5 kg of
crude-14 in 85:15 heptane/EtOAc (5 L) and eluting with the
same ratio of solvent over a bed of silica gel (35 kg, ∼14 g/g
14). The trans-14 isomer eluted first; the separation was
monitored by TLC (1:1 EtOAc/MeOH) [Rf trans-14 = 0.5, Rf
cis-14 = 0.4]. The heptane/EtOAc solution (∼800 L) of
product was concentrated under reduced pressure to give trans-
14 (1.24 kg, 98.1 A%) as a viscous dark yellow oil. Two
additional batches of crude-14 (2.4 and 2.3 kg) were purified in
a similar process to give a total of 3.74 kg trans-14 (>98 A%, see
Supporting Information S2, Figures S2−3.15−17). An addi-
tional second crop of 1.63 kg of trans-14 with ∼95 A% purity
was also obtained.
Ester Hydrolysis to Hydroxyl trans-1. A 50 L reactor
containing a solution of trans-14 (3.74 kg, 8.94 mol) in THF
(22.5 L) was charged with an aqueous solution of LiOH−H2O
(1.13 kg, 26.81 mol, 3.0 equiv in 11.3 L of H2O) with agitation
at room temperature. The hydrolysis was monitored by HPLC
[Rt trans-1 = 5.48 min, Rt trans-14 = 9.43 min] and deemed
complete after 20 h. The reaction mixture was charged with
brine (20 L) and EtOAc (20 L) and stirred for 20 min before
separation of the organic layer. The aqueous layer was extracted
with EtOAc (2 × 10 L), and then the combined organics were
treated with charcoal (0.37 kg) and agitated for 30 min at room
temperature. The mixture was treated with MgSO4 (0.37 kg),
filtered, and concentrated under reduced pressure to give trans-
1 (2.70 kg, 90.4% yield with 98.6 A% and 99.4:0.6 trans/cis by
chiral HPLC). A second batch performed under identical
reaction conditions gave trans-1 (1.12 kg, 94.1% yield with 96.1
A% and 99.2/0.8 trans/cis by chiral HPLC). The two lots of
pure trans-1 were combined to give 3.80 kg (98.2 A%, 98.8% de
and >99.9% ee by chiral HPLC): 1H NMR (600 MHz, DMSO-
d6) 8.46 (s, 1H), 5.39 (s, 1H), 4.85 (t, J = 6.7 Hz, 1H), 3.67 (m,
J = 3.4 Hz, 2H), 3.55 (m, J = 3.5 Hz, 2H), 3.52 (m, J = 3.2 Hz,
2H), 3.46 (m, J = 5.5 Hz, 2H), 3.39 (m, J = 5.6 Hz, 2H), 1.96
50
CH2Cl2 (25 L), filtered, and concentrated to give crude-(R)-
12 (5.04 kg, 85 A% containing approximately 14:1 trans/cis-1
by HPLC) as a light brown semi-solid. See Supporting
1
Information S2 for the H NMR spectrum (CDCl3) of crude
ketone (R)-12 in Figure S2−3.11, HPLC chromatogram (280
1
nm) in Figure S2−3.12, and for the H and 13C NMR spectra
(CDCl3) of β-keto ester intermediate (S)-11 in Figure S2−
3.22A and Figure S2−3.22B, respectively.
Preparation of tert-Butyl 4-((5R,7R)-7-Hydroxy-5-
methyl-6,7-dihydro-5H-cyclopenta[d]pyrimidin-4-yl)-
piperazine-1-carboxylate, Crude-1. A 50 L reactor was
charged with crude (R)-12 (1.43 kg, 4.31 mol by assay, 1.68 kg
crude weight) and DCM (25 L), and then the agitated solution
was sparged with nitrogen ∼30 min. Triethylamine (0.60 L,
4.33 mol, 1.0 equiv) was added in one portion, followed by
slow addition of formic acid (0.91 L, 5.05 mol, 1.2 equiv) over
15 min. [Caution: Addition of formic acid is exothermic and
fumes!] The reaction mixture was sparged with nitrogen,
charged with [(R,R)-MsDPEN−RuCl(p-cymene)] catalyst 13
(19.2 g, 0.04 mol, 0.01 equiv; [S/C = 263:1]), and then
agitated at room temperature overnight. The reaction was
monitored by HPLC [Rt (trans-1) = 13.72 min, Rt (cis-1) =
14.06 min, Rt ((R)-12) = 14.58 min] and found to be complete
after consumption of starting material (20 h). The reaction
mixture was concentrated under reduced pressure at ≤45 °C to
give crude-1 (1.7 kg) as a dark brown oil−solid. Two additional
batches of crude ketone (R)-12 (1.43 kg) were processed under
(m, J = 4.6 Hz, 2H), 1.42 (s, 9H), 1.09 (d, J = 7.0 Hz, 3H); 13
C
NMR (150 MHz, DMSO-d6) 172.0, 159.6, 156.4, 153.8, 120.8,
79.0, 72.0, 45.0 (2C), 43.6−42.5 (2C), 40.9, 34.4, 28.8 (3C),
19.8. HRMS calcd for C17H24N4O3 [M+H]+ 335.2078; found
335.2078. Trace metal analysis by ICP-AES showed Pd and Ru
<2 ppm. See Supporting Information S2 for the NMR spectra
of trans-1: (1H) in Figure S2−3.18A and (13C) in Figure S2−
3.18B, and the HPLC chromatograms in Figure S2−3.19 (254
nm, chiral) and Figure S2−3.20 (280 nm). See Supporting
J
dx.doi.org/10.1021/op500270z | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Information S1 and S2 for the characterization and NMR
spectral data of cis-1 (1H) in Figure S2−3.21 and (13C) in
Figure S2−3.21B.
washed with 25 wt% aqueous NH4Cl (36 kg) and 15% aqueous
NaCl (36 kg), dried over Na2SO4 (10 kg), and filtered into a
clean 800 L reactor. The organics are distilled to a minimum
working volume while maintaining the internal temperature
≤35 °C. Tetrahydrofuran (166 kg) is charged and distillation
continued until half the volume of THF added is distilled. The
THF solution was cooled to 20 °C and assayed to give 25 wt%
of (S,R)-19 in THF, indicating 28.8 kg and 109% (solution)
yield. A 100 mL sample was removed, concentrated to dryness,
and analyzed by HPLC to gave 88.4 A% purity for (S,R)-19,
containing 1.2% of the R,R-diastereomer [Method 2.11]. The
structure was consistent with the reported literature.5 The
crude product was used directly in the next step.
(R)-4-Benzyl-3-(2-(4-chlorophenyl)acetyl)oxazolidin-
2-one, 17.52 An 800 L reactor was charged with 4-
chlorophenylacetic acid 15 (17.4 kg, 102.0 mol, 1.2 equiv),
(R)-benzyl-2-oxazolidinone 1653 (15.0 kg, 84.7 mol, 1.0 equiv),
and toluene (181 kg, 14 vol), and the stirred suspension was
cooled to 15 °C. Triethylamine (34 kg, 336 mol, 4 equiv) was
charged to the mixture, and the solids were allowed to dissolve.
Pivaloyl chloride (13.0 kg, 108 mol, 1.28 equiv) was added
slowly at a rate to maintain the temperature below 30 °C. The
reaction mixture was heated to reflux and stirred overnight.
Analysis by HPLC showed (R)-16 (3.3 min) = 8.5%, (R)-17
(6.9 min) = 91.5 A%. The mixture was cooled to ∼20 °C, water
(120 kg) was charged to the reactor, and the mixture was
stirred for 10 min. After 10 min, the layers were separated and
the organics washed with 4% aqueous NaHCO3 (120 kg) and
10% aqueous NH4Cl (45 kg), dried over Na2SO4 (5 kg), and
filtered into a clean 200 L reactor. The organics were
concentrated via distillation to a minimum working volume,
and then 2-propanol (82 kg) was charged to the reactor,
followed by another distillation to a minimum working volume
(maintained internal temperature ≤45 °C). Additional 2-
propanol (35 kg) was added, distillation to a minimum working
volume resulted in a suspension that was filtered, and the filter
cake was rinsed with 2-propanol (2 × 2 kg).54 The wet cake
was dried under vacuum maintaining internal temperature ≤50
°C until a constant weight was achieved, resulting in (R)-17
(17.40 kg, 92.5 A% [Method 2.11]) in 62% yield as an off-white
solid. The product was of sufficient purity to carry forward to
the next step. The structural data of (R)-17 is consistent with
that reported in the literature.1a
Crude (S)-3-(tert-Butoxycarbonyl(isopropyl)amino)-2-
(4-chlorophenyl)propanoic Acid, Crude-(S)-2. An 800 L
reactor was charged with LiOH·H2O (5.6 kg, 133.5 mol, 2.5
equiv) and water (185 kg, 7 vol) with agitation at room
temperature until the solids dissolved. Tetrahydrofuran (221
kg) was then charged into the reactor followed by the slow
addition of 30% H2O2 (13.5 kg, 105.4, 2.0 equiv) such that the
internal temperature does not exceed 30 °C. The contents were
cooled to ∼0 °C, and then the THF solution of crude (S,R)-19
(26.4 kg, 52.7 mol, 1.0 equiv, 115.1 kg total weight) was
charged slowly so that the internal temperature is maintained
between 0 and −10 °C. The mixture was then agitated at ∼0
°C for 3 h and assayed by HPLC for conversion of crude (S,R)-
19 to crude-2 (reaction deemed complete when crude (S,R)-19
≤2 A% HPLC). HPLC sample preparation: A 10 mL sample
was removed and quenched into 12.5 wt% aqueous Na2SO3 (5
mL) and EtOAc (5 mL). Upon completion, the reactor is
slowly charged with 12.5 wt% aqueous Na2SO3 (55 kg)
maintaining internal reaction temperature <10 °C. After stirring
90 min, the pH of the reaction mixture was recorded (pH 14 by
litmus). A solution of 27 wt% aqueous KHSO4 (70 kg) was
slowly charged was charged slowly to the reactor while
maintaining internal temperature <15 °C until the pH was
between 2 and 3. Toluene (114 kg) was charged to the reactor,
the mixture was stirred at ∼20 °C for 15 min, and then the
layers were allowed to separate (1 h). The layers were
separated, and the aqueous layer was extracted with toluene (2
× 79 kg). The combined organics were dried (Na2SO4), filtered
into a clean reactor, and distilled to a minimum working
volume (internal temperature maintained ≤45 °C). Toluene
(342 kg) was charged to the reactor and then distilled to a
minimum working volume (internal temperature maintained
≤45 °C). The process was repeated with additional charge of
toluene (87.2 kg). A 100 mL sample was removed and
concentrated to dryness to afford 23.3 g. The bulk toluene
solution was assayed to be 13.4 wt% crude (S)-2 and ∼66 A%
purity [Method 2.11]. Note: Residual chiral auxiliary (R)-16 is
still present but is removed in the next step.
Preparation of tert-Butyl (S)-3-((R)-4-Benzyl-2-oxoox-
azolidin-3-yl)-2-(4-chlorophenyl)-3-oxopropyl-
(isopropyl)carbamate, 19.55 A dry 800 L reactor was
charged with (R)-17 (17.4 kg, 52.7 mol, 1.0 equiv) and CH2Cl2
(332 kg, 19 vol) with agitation at room temperature. The
contents were cooled to ≤ −20 °C, and a solution of 1 M TiCl4
in toluene (54.0 kg, 54.0 mol, 1.05 equiv) was charged slowly
charged slowly in order to maintain an internal temperature ≤
10 °C. Note: This step is exothermic, and the solution will turn
orange to orange-red during the charge. The mixture was
agitated at ≤ −20 °C for ∼2 h, and then DIPEA (7.5 kg, 57.7
mol, 1.1 equiv) is slowly charged over ∼2 h period. The
temperature during this charge is maintained ≤ −10 °C, and
the solution should turn dark purple. The contents were
agitated at ≤ −20 °C for ∼1.5 h, and then a solution of Boc-
aminal 1856 in CH2Cl2 (13.9 kg, 1.3 equiv of 18 (for
preparation, see Supporting Information S1), 50.6 kg total
weight including DCM) was added at such a at such a rate in
order to maintain an internal temperature ≤ 10 °C (addition
time ∼1 h). The reaction mixture was stirred at ≤ −20 °C for 4
h and then assayed by HPLC for conversion of (R)-17 to
product (S, R)-19 (reaction complete when (R)-17 ≤10 A% by
HPLC). A 10 mL sample was removed and quenched into 15
wt% NH4Cl (5 mL), and the bottom DCM layer analyzed.
Upon completion, the reaction mixture was warmed to 0 °C,
and then 25% aqueous NH4Cl (36 kg) was charged maintaining
internal reaction temperature <10 °C (emulsion observed).
After agitation for 90 min, water (36 kg) was charged to
solubilize the solids, and then the layers were allowed to
separate. The aqueous layer is removed, and the organics are
(S)-2-(4-Chlorophenyl)-3-(isopropylamino)propanoic
Acid Sodium Salt, (S)-20-Na. A toluene solution of crude
(S)-2 (18.0 kg, 52.7 mol, 1.0 equiv, 173.1 kg total from the
previous step) was charged into a 800 L reactor with agitation
at room temperature. Concentrated HCl (16.0 kg, 158 mol, 3
equiv) was slowly pumped into the reactor with good agitation,
and the pump was rinsed with H2O (2 kg). The reaction
mixture was warmed to 30 °C, and after 6 h IPC by HPLC
indicated the Boc group deprotection was complete (crude (S)-
2 ≤1.0 A% HPLC). Note: For sampling, a 10 mL aliquot was
removed from the reactor containing two layers, and both
layers were analyzed by HPLC for starting material. The
K
dx.doi.org/10.1021/op500270z | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
agitation was stopped, and the layers were separated. The
aqueous layer (pH 0 by litmus) containing the product was
extracted with toluene (3 × 80 kg), and then the aqueous layer
was charged back into the reactor, and EtOAc (132 kg) was
charged. The biphasic mixture was agitated at ∼0 °C for 20
min, and then aqueous 25 wt% NaOH (23.2 kg) was slowly
charged to the reactor until pH 9 (litmus) was obtained. The
internal temperature was maintained ≤10 °C during the
hydroxide addition and resulted in precipitation of product 20-
Na. The mixture was stirred at ∼5 °C for 1 h and then filtered
into a Nutsche filter, and the wet cake was washed with EtOAc
(2 × 8 L). The solids were dried in a vacuum (45 °C, 13 h) to
give sodium carboxylate 20-Na (7.65 kg, 99.5 A%) in 60%
yield. The chiral purity measured 98.9:1.1 S/R = 97.8% ee by
HPLC [Method 2.10].
For crystallization, 2-propanol (180 kg, 24 vol) was charged
to the reactor containing crude 20-Na (7.65 kg), and the
mixture was heated to reflux (IT = 79 °C) for 1 h. The solution
was allowed to cool to 20 °C over ∼3 h and maintained
agitation for an additional 4 h. The solids were filtered, and the
filter cake was washed with 2-propanol (40 kg) and then dried
in a vacuum oven at 45 °C (internal) to constant weight (∼10
h), giving sodium carboxylate (S)-20-Na (6.85 kg), in 90%
yield from crude 20-Na, with 99.7% ee and 99.9 A% [Methods
2.10 and 2.11] as white solid. The product was used directly in
the next step.
(S)-3-(tert-Butoxycarbonyl(isopropyl)amino)-2-(4-
chlorophenyl)propanoic Acid, β2-Amino Acid (S)-2. A
HDPE container was charged with tetramethylammonium
hydroxide (Me4N+OH−) (9.7 kg, 53.5 mol, 2.1 equiv) and H2O
(20 kg) and agitated at room temperature until a solution was
obtained. A separate HDPE container was charged with di-tert-
butyl dicarbonate (Boc2O) (9.8 kg, 44.9 mol, 1.7 equiv) and
acetonitrile (11 kg, 1.6 vol), and the mixture was agitated until a
solution was observed. The tetramethylammonium hydroxide
solution was slowly charged into a solution of sodium
carboxylate (S)-20-Na (6.85 kg, 25.8 mol, 1.0 equiv) and
acetonitrile (150 L, 22 vol), maintaining the internal temper-
ature ≤20 °C during the addition. The reaction mixture was
cooled to ∼15 °C, and then the di-tert-butyl dicarbonate
solution was charged over ∼1 h, maintaining the internal
temperature ≤20 °C. The reaction mixture was warmed to ∼25
°C for 4 h and then sampled. In-process analysis indicated the
Boc protection was complete (∼1.4 A% HPLC (S)-20-Na
remaining). The contents were distilled (IT ≤35 °C) until a
minimum working volume was achieved. The reactor was
charged with 2% NaOH (68 kg) and MTBE (40 kg), and the
mixture was agitated for 15 min. The layers were separated, and
the basic aqueous (pH 11) was extracted again with MTBE (60
kg). The aqueous was acidified slowly with 27 wt% aqueous
KHSO4 (47 kg) until pH 3 was obtained. [Caution: This step is
exothermic and IT should be maintained <25 °C.] The agitation
was stopped, and the layers were allowed to separate over 1 h,
and the upper layer was removed. MTBE (35 kg) was charged
to the reactor containing the aqueous layer and agitated for 15
min. The MTBE layer was removed, dried (Na2SO4), filtered,
and concentrated via rotary evaporation (bath temp. = 40 °C)
until ∼90% of the solvent was removed. Heptane (5 L) was
charged and the solution distilled via rotary evaporation (bath
temp. = 40 °C) until ∼85% of the solvent was removed. The
solution was then sparged with nitrogen at 38 °C for >20 min
resulting in crystallization of the product. The suspension was
further concentrated to dryness and then the solids oven-dried
(≤50 °C) to constant weight to give β2-amino acid (S)-2 (8.85
kg, >99% ee and >99 A% [Methods 2.10 and 2.11]) in 99%
yield. See Supporting Information S2 for the NMR spectra of
β2-amino acid (S)-2: (1H) Figure S2−5.1 and (13C) Figure S2−
5.2.
tert-Butyl (S)-2-(4-Chlorophenyl)-3-(4-((5R,7R)-7-hy-
droxy-5-methyl-6,7-dihydro-5H-cyclopenta[d]-
pyrimidin-4-yl)piperazin-1-yl)-3-oxopropyl(isopropyl)-
carbamate, 22. A 100 L reactor was charged with trans-1
(3.49 kg, 10.4 mol, 1.0 equiv), toluene (16 L, 5 vol), and 2.5 M
HCl in EtOH (12.5 L, 31.2 mol, 3.0 equiv) with agitation. The
reaction mixture was warmed to 60 °C, and after 1 h, HPLC
analysis showed no starting material remained. The reaction
mixture was co-evaporated with toluene (9.5 L) and distilled to
∼22 L. The biphasic mixture of crude-21 was cooled to room
temperature, charged with CH2Cl2 (16 L) and then DIPEA
(9.0 L, 5.0 equiv), and agitated for ∼30 min. A solution of β2-
amino acid (S)-2 (3.57 kg, 10.5 mol, 1.0 equiv) and CH2Cl2
(14 L) was charged to the batch containing (R,R)-21, and the
reactor was rinsed with CH2Cl2 (3.5 L). The reaction mixture
was charged with HBTU (4.49 kg, 11.85 mol, 1.13 equiv) in
two portions at room temperature and monitored by HPLC.
Analysis of the reaction mixture after 3 h showed amino acid
(S)-2 = 0.53 A% by HPLC and met specification ((S)-2 ≤ 1%).
The reaction mixture was charged with 7.4 wt% aqueous
NaHCO3 (20 L) with good agitation, the aqueous layer was
removed, and then solvent exchanged from CH2Cl2 to IPAc
(20 L). The IPAc solution was washed with aqueous 0.6 N
NH4OH (2 × 21 L), 25 wt% aqueous NH4Cl (20 L), and then
H2O (2 × 20 L). A final wash of the organics with H2O (20 L)
in combination with 25 wt% aqueous NH4Cl (3 L) provided a
clear phase separation. Analysis of the organic layer by GC for
tetramethylurea (TMU) content indicated below specification
of ≤0.15 wt%. The batch was charged with a slurry of charcoal
(0.7 kg, 20 wt%) in IPAc (6.5 L) and SiliaMetS thiol (1 kg, 20
wt%) and then warmed to 47 °C for 18 h. The mixture was
cooled to ambient temperature and filtered over Celite (3 kg).
The filtered solids were rinsed with IPAc (19 L), and the filtrate
was concentrated then solvent exchanged from IPAc into
MTBE (∼25 L). The MTBE mixture was heated to an internal
temperature of 50 °C for complete dissolution of solids and
then slowly cooled to 40 °C over 2.3 h. A slurry of seed
material Boc-API (S,R,R)-22 (12 g, 0.5 wt%) in heptane (1 L)
was added and the mixture cooled to 36 °C over 30 min.
Additional heptane (5.9 L, 1.7 vol) was slowly charged, and the
slurry was cooled to 25 °C and then aged overnight. The
crystallized product was filtered and rinsed with heptane (6 L),
and then the filter cake was dried at ≤55 °C under reduced
pressure to give Boc-API (S,R,R)-22 (4.69 kg, 81% yield, 99.6 A
% [Method 2.2]; ruthenium content by IPC-AES = 11 ppm])
as a white solid. Analysis by CAD-HPLC showed PF6 anion =
0.35% [Method 2.3]. The filtrate from the crystallization
contained an additional ∼820 g (S,R,R)-22 by weight assay.
HRMS calcd for C29H40ClN5O4 [M+H]+ 558.2841; found:
558.2838. Spectral data are consistent with those previously
reported.1a
(S)-2-(4-Chlorophenyl)-1-(4-((5R,7R)-7-hydroxy-5-
methyl-6,7-dihydro-5H-cyclopenta[d]pyrimidin-4-yl)-
piperazin-1-yl)-3-(isopropylamino)propan-1-one, Ipata-
sertib Mono-HCl. A 100 L reactor containing an agitated
solution of Boc-API (S,R,R)-22 (4.57 kg, 8.19 mol) and toluene
(20 L) was slowly charged with 12 N HCl (2.1 L, 24.6 mol, 3
equiv) at room temperature. The mixture was heated to 50 °C
L
dx.doi.org/10.1021/op500270z | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
and the reaction monitored by HPLC for consumption of
(S,R,R)-22 (not detected after 1 h). The reaction mixture was
then diluted with H2O (12 L) and agitated, and the layers
allowed to separate. The lower aqueous layer containing
Ipatasertib di-HCl was removed, and the organic layer was
washed with H2O (4 L). The aqueous extracts were combined
and distilled under reduced pressure in order to remove
residual toluene (∼2 L). The aqueous solution was cooled to
20 °C, and CH2Cl2 (14 L) was added. To the biphasic mixture
was slowly added 5 N NaOH (6.3 L) until pH >12 with good
agitation. The layers were allowed to separate, and the bottom
organic layer was removed. The basic aqueous layer was
extracted again with CH2Cl2 (16 L), and the combined organics
were washed with saturated aqueous NaHCO3 (15 kg)
followed by brine (19 kg). The organics were dried over
anhydrous Na2SO4 (2.79 kg) and then filtered. A slurry of
charcoal (0.91 kg, 20 wt%) in CH2Cl2 (1 L) was added to the
filtrate followed by SiliaMetS thiol (1 kg, 20 wt%) and the
mixture maintained at 22 °C for ∼4 h. The mixture was filtered
through a pad of Celite (3.4 kg) and rinsed with CH2Cl2 (12 L)
until no product was observed eluting from the filter. HPLC
weight assay of the filtrate indicated 3.74 kg (8.17 mol) of
Ipatasertib free-base. The free-base solution at 19 °C was slowly
charged with 12 N HCl (0.65 L, 0.96 equiv), and then the
mixture was distilled under reduced pressure to a minimum
working volume (∼10 L). Additional CH2Cl2 (10 L) was
charged to the reactor, and the solution was distilled until water
was no longer observed in the distillate. To the DCM solution
was slowly charged EtOAc (33 L). The mixture was distilled to
a minimum working volume (∼20 L), charged with EtOAc (20
L), and then distilled. EtOAc (15 L) was added to the mixture,
agitated 4 h, and then filtered. The filter cake was rinsed with
EtOAc (12 L), and the solids were dried at ≤70 °C with
nitrogen purge for ∼48 h. The product was discharged from the
filter dryer to give Ipatasertib mono-HCl (3.23 kg, 80% yield)
as an off-white solid. Analytical results: 99.7 A% [0.26% S,R,S-
diastereomer observed)]; impurity 23 (M399) was not
detected (<0.02 A%) [Method 2.2]; ruthenium content by
IPC-AES = 5 ppm; analysis for PF6 anion by CAD-HPLC
resulted in not detected [Method 2.3]; residual solvent = 0.4%
EtOAc; ion chromatography (IC) = 8.5% chloride (1.14 salt
equivalent); DSC = 141 °C; FTIR (neat) 3269 (br OH),
2,6-dihydroxypyrimidine (R)-5 from (R)-MGM. Spectral and
chromatographic data for the key compounds are found in S2.
This material is available free of charge via the Internet at
http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
*E-mail: travisr@gene.com.
■
Present Address
§K.L.S.: OPX Biotechnologies Inc., 2425 55th St., Suite 100,
810 Boulder, CO 80301, USA
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
The authors thank the following people for their contributions
to this project: Nik Chetwyn, Derrick Yazzie, Christine Gu,
Tina Nguyen, Ila Patel, Sarah Stowers, Kelly Louie, Venkata
Ravuri, and Bob Garcia, Jr., for analytical support; Beckee
Rogers and Greg Sowell for process chemistry support; Josef
Bencsik and Nick Kallen at Array Biopharma Inc.; and external
collaborators on this project, including Cambridge Major
Laboratories Inc. and Asymchem. We thank Professors Barry
M. Trost, Scott Denmark, Larry E. Overman, and E. J. Corey
and Drs. Trevor Laird and James Zeller for insightful discussion
and suggestions during the early development stages of this
project. Special thanks to Remy Angelaud for editorial
suggestions with this manuscript.
ABBREVIATIONS
■
API: Active Pharmaceutical Ingredient
cGMP: Current Good Manufacturing Practice
Boc: tert-butoxycarbonyl
DCM: dichloromethane
Dppp: 1,3-bis(diphenylphosphino)propane
TsDPEN: (1R,2R)-N-(p-toluenesulfonyl)-1,2-diphenyl-1,2-
ethanediamine
MsDPEN: (1R,2R)-N-(methanesulfonyl)-1,2-diphenyl-1,2-
ethanediamine
REFERENCES
■
1
2961−2865 (N−H stretch), 1637 (CO stretch); H NMR
(1) (a) Blake, J. F.; Xu, R.; Bencsik, J. R.; Xiao, D.; Kallan, N. C.;
Schlachter, S.; Mitchell, I. S.; Spencer, K. L.; Banka, A. L.; Wallace, E.
M.; Gloor, S. L.; Martinson, M.; Woessner, R. D.; Vigers, G. P. A.;
Brandhuber, B. J.; Liang, J.; Safina, B. S.; Li, J.; Zhang, B.; Chabot, C.;
Do, S.; Lee, L.; Oeh, J.; Sampath, D.; Lee, B. B.; Lin, K.; Liederer, B.
M.; Skelton, N. J. J. Med. Chem. 2012, 55, 8110−8127. (b) Bencsik, J.
R.; Xiao, D.; Blake, J. F.; Kallan, N. C.; Mitchell, I. S.; Spencer, K. L.;
Xu, R.; Gloor, S. L.; Martinson, M.; Risom, T.; Woessner, R. D.;
Dizon, F.; Wu, W.-I.; Vigers, G. P. A.; Brandhuber, B. J.; Skelton, N. J.;
Prior, W. W.; Murray, L. J. Bioorg. Med. Chem. Lett. 2010, 20, 7037−
7041.
(600 MHz, DMSO-d6) 9.39 (s, 1H), 8.64 (s, 1H), 8.49 (s, 1H),
7.49 (q, J = 2.9 Hz, 2H), 7.41 (q, J = 2.9 Hz, 2H), 5.58 (s, 1H),
4.91 (t, J = 6.9 Hz, 1H), 4.78 (dd, J = 8.9, 4.5 Hz, 1H), 3.81 (m,
J = 3.3 Hz, 1H), 3.68 (m, J = 3.3 Hz, 1H), 3.67 (m, J = 3.1 Hz,
1H), 3.65 (m, J = 3.2 Hz, 1H), 3.63 (m, J = 3.6 Hz, 1H), 3.59
(m, J = 4.3 Hz, 1H), 3.51 (m, J = 3.5 Hz, 1H), 3.46 (m, J = 3.5
Hz, 1H), 3.36 (m, J = 3.2 Hz, 1H), 3.30 (m, J = 5.7 Hz, 1H),
3.21 (m, J = 3.4 Hz, 1H), 2.98 (m, J = 5.8 Hz, 1H), 1.97 (m, J =
4.8 Hz, 2H), 1.26 (d, J = 6.6 Hz, 3H), 1.25 (d, J = 7.0 Hz, 3H);
13C NMR (150 MHz, DMSO-d6) 170.2, 168.2, 159.4, 155.2,
135.3, 132.5, 129.7 (2C), 129.1 (2C), 120.8, 71.7, 50.4, 47.0,
44.8, 44.5, 44.1, 41.4, 40.8, 34.5, 19.8, 18.4, 18.1; HRMS calcd
for C24H32ClN5O2 457.2245; found [M+H]+ 458.2306.
(2) Lane, J.; Spencer, K. L.; Shakya, S. R.; Kallan, N. C.; Stengel, P. J.;
Remarchuk, T. Org. Process Res. Dev. 2014, DOI: 10.1021/op500271w.
(3) In 2011, a patent appeared several years after our initial work for
the preparation of DHP rac-5: Beight, D. W.; Burkholder, T. P.;
Clayton, J. R.; Eggen, M.; Henry, K. J. Jr.; Johns, D. M.; Parthasarathy,
S.; Pei, H.; Rempala, M. E.; Sawyer, J. S. AKT Inhibitors. Patent
WO2011050016(A1).
(4) Schaefer, J. P.; Bloomfield, J. J. The Dieckmann Condensation
(Including the Thorpe-Ziegler Condensation). Org. React. 2011, 15,
1−203.
(5) Fujii, A.; Hashiguchi, S.; Uematsu, N.; Ikariya, T.; Noyori, R. J.
Am. Chem. Soc. 1996, 118, 2521−2522.
ASSOCIATED CONTENT
* Supporting Information
■
S
Supporting Information S1 describes the analytical methods
used during the preparation of cyclopentylpyrimidine trans-1,
β2-amino acid (S)-2, and Ipatasertib; Supporting Information
S1 also provides an alternative procedure for the preparation of
M
dx.doi.org/10.1021/op500270z | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
(6) (a) See ref 1 for the discovery synthesis and alternative
approaches to the preparation of β-amino acid (S)-2. (b) For a more
recent and state-of-the-art synthesis of β-amino acid (S)-2, see:
Remarchuk, T.; Babu, S.; Stults, J.; Zanotti-Gerosa, A.; Roseblade, S.;
Yang, S.; Huang, P.; Sha, C.; Wang, Y. Org. Process Res. Dev. 2014, 18,
135−141.
reference describing the dianion formation of 1-indanone: Trost, B.
M.; Latimer, L. H. J. Org. Chem. 1977, 42, 3212−3214. Further
mechanistic studies for the epimerization β-keto ester 11 are
warranted.
(23) We considered the strategy of converting ketone (R)-12
selectively to cis-1 and then inverting the hydroxyl group via a
Mitsunobu reaction to obtain trans-1; however, this was not practical
since it required two extra steps and purification would still be needed
to separate the diastereomers.
(24) (a) Use of a stronger base such as aqueous NaOH could result
in decomposition of ketone (R)-12 via cross-aldol-type condensation.
(b) Not quenching the formic acid at the end of this reaction also
resulted in cleavage of the Boc group.
(25) The highest cis diastereoselectivity observed with heterogeneous
hydrogenation was from using 5% Pd/C Type A405038 catalyst
(Johnson Matthey, 5 bar H2, 40 °C in EtOAc), which gave >99%
conversion to cis-1 with 96% de after 16 h.
(26) We acknowledge Arch Pharmalabs Ltd., India, for trying the
TarB-X catalyst that gave ∼85% yield of ∼1:1 cis/trans-1: Suri, J. T.;
Vu, T.; Hernandez, A.; Congdon, J.; Singaram, B. Tetrahedron Lett.
2002, 43, 3649−3652. The DIP-Cl (B-chloro-diisocamphenyl borane)
catalyst was unreactive toward ketone reduction: Chandrasekharan, J.;
Ramachandran, P. V.; Brown, H. C. J. Org. Chem. 1985, 50, 5446−
5448.
(27) (a) Noyori, R.; Hashiguchi, S. Acc. Chem. Res. 1997, 30, 97−102.
(b) Fujii, A.; Hashiguchi, S.; Uematsu, N.; Ikariya, T.; Noyori, R. J. Am.
Chem. Soc. 1996, 118, 2521−2522. (c) Magano, J.; Dunetz, J. R. Org.
Process Res. Dev. 2012, 16, 1156−1184.
(28) For examples of transfer hydrogenation of ketones with
HCOOH/Et3N mixture as the source of hydrogen, see: (a) Zhang, J.;
Blazecka, P. G.; Bruendl, M. M.; Huang, Y. J. Org. Chem. 2009, 74,
1411−1414. (b) Miyagi, M.; Takehara, J.; Collet, S.; Okano, K. Org.
Process. Res. Dev. 2000, 4, 346−348.
(29) The [(R,R)MsDPEN−Ru(p-cymene)Cl] catalyst was used in
our first kilogram scale-up campaign for the reduction of ketone (R)-
12 since it was readily available and gave good activity and selectivity
(∼90% de) compared to the [(R,R)TsDPEN−Ru(p-cymene)Cl]
catalyst (≤92% de) in a screening study starting with pure ketone
(R)-12.
(7) Marder, O.; Albericio, F. Industrial application of coupling
reagents in peptides. Chem. Oggi 2003, 21, 35−40.
(8) Preparation of triester rac-4 from sodium methylmalonate and
́
methylcrotonate has been reported: Alvarez, E.; Cuvigny, T.; Herve du
Penhoat, C.; Julia, M. Tetrahedron 1988, 44, 119−126.
(9) The experimental details to prepare triester (R)-4 from (R)-
monomethyl methylglutarate (MGM) can be found in the Supporting
Information S1.
(10) Although DHP rac-5 was also a viable substrate for enzymatic
resolution, a preliminary screen of various commercial lipases was not
promising, and this approach was not pursued. Similarly, the classical
resolution of the carboxylic acid derivative of DHP rac-5 with (R)- or
(S)-α-methylbenzylamine resulted in only poor selectivity (highest
chiral ratio observed was ∼60:40).
(11) In a similar approach, the enzymatic resolution of dimethyl 3-
methylpentanedioate, the precursor to triester rac-4, to produce (R)-3-
methylglutaric acid monomethyl ester (MGM) was reported by
Alvarez and co-workers in ref 8.
(12) Increasing the enzyme loading of lipase AYS “Amano” further to
50 wt% had a deleterious effect and resulted in enantiomeric ratio of
4.1:95.9 after 250 h (10 d).
(13) A% indicates area% purity by HPLC unless otherwise noted
throughout the manuscript.
(14) See Supporting Information S1 for the preparation of DHP (R)-
5 using triester (R)-4 from this route.
(15) See ref 2 for the initial proof-of-principle synthesis of ketone rac-
12.
(16) Similar good results for the conversion of DHP 5 to DCP 6
were obtained using Et3N instead of 2,6-lutidine, and although 1,2-
dichloroethane could also be substituted for toluene, the reaction
times were extended from ∼24 to 48 h for complete conversion.
(17) Other common ester derivatives of (R)-9 prepared did not lead
to efficient decarboxylation following the Dieckmann cyclization
reaction; see ref 2.
(18) A major side product from the combination of benzyl chloride
and K2CO3 for re-esterification was the lactone product, which was not
observed using the current conditions. See Supporting Information S1,
section 4.4, for the proposed structure.
(30) Using a 1:1 mixture of HCCOH/Et3N versus a 5:2 mixture (as
originally reported by Noyori in ref 27b) generally gave ∼10% higher
de of trans-1 product with the [(R,R)TsDPEN−Ru(p-cymene)Cl]
catalyst.
(31) Crystallization of trans-1 was not optimized at this stage of the
project but was achieved from several single solvents, such as EtOAc,
IPAc, or MTBE. Alternatively, the cis/trans-1 diastereomers could be
separated without derivatization using preparative HPLC (column,
CHIRALPAK IC, 20 μm, 11 × 25 cm; mobile phase, MeOH/0.05%
DMEA; flow rate, 570 mL/min; column temp, 25 °C; UV detection,
320 nm). We acknowledge Chiral Technologies Inc. for identifying
these separation conditions.
(32) The p-nitrobenzoate ester of 1 was used for separation of
diastereomers by silica gel chromatography in the medicinal chemistry
synthesis (see ref 1a).
(33) Attempted crystallization of the pivalate ester trans-14, as well as
the acetate and p-nitrobenzoate derivatives, was not successful from
CPME, DCM/petroleum ether, EtOAc/hexanes, i-PrOH, or acetone
solvents.
(34) Trace metal analysis of trans-1 using ICP-AES indicated that
both ruthenium and palladium were efficiently removed in the
purification process and both were ≤2 ppm.
(35) (a) The authors acknowledge IRIX Pharmaceuticals for the
pilot plant scale-up of β2-amino acid (S)-2. (b) For an example of the
use of Evans’s chiral auxiliary on kilogram scale see: Slade, J.; Parker,
D.; Girgis, M.; Mueller, M.; Vivelo, J.; Liu, H.; Bajwa, J.; Chen, G.-P.;
Carosi, J.; Lee, P.; Chaudhary, A.; Wambser, D.; Prasad, K.; Bracken,
K.; Dean, K.; Boehnke, H.; Repic, O.; Blacklock, T. J. Org. Process. Res.
Dev. 2006, 10, 78−93.
(19) The origin of this work is described in Part 1 of this series (ref
2).
(20) (a) Higher reaction temperatures led to SNAr of the
chloropyrimidine with the 2-propanol to give the aryl isopropyl
ether. In addition, poor stirring of the heterogeneous reaction mixture
resulted in a significant amount of the same byproduct. Formation of
the isopropyl ester byproduct from transesterification was not an issue
using the optimized reaction conditions. (b) For a good discussion of
the influence of the reaction variables such as CO pressure and effect
of base on the carbonylation reaction, see: Barnard, C. F. Org. Process
Res. Dev. 2008, 12, 566−574.
(21) β-Keto ester intermediate (S)-11 has been isolated by
acidification with glacial AcOH (or aqueous HCl) to pH 5−6,
followed by aqueous work-up and extraction with EtOAc. The
concentrated organics resulted in a solid product that was slurried in
MTBE/petroleum ether (1:2) to afford white solids after filtration.
From one experiment, 119 g of diester (R)-10 gave 84 g of β-keto
ester (S)-11 in 80% isolated yield. This isolation procedure was not
favored due to the loss of yield on work-up, and it was speculated that
the product was sensitive to oxidation at the enolic position with
exposure to air.
(22) We also observed epimerization of the methyl stereocenter
during this two-step process and speculate that it can occur through
formation of a dianion intermediate during the Dieckmann cyclization
step with KOtBu. The authors thank Dr. Lee Latimer for a related
N
dx.doi.org/10.1021/op500270z | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
(36) (a) The procedure for the thermal reaction conditions was first
provided to us by Josef Bencsik at Array Biopharma; see also: Prashad,
M.; Kim, H.-Y.; Har, D.; Repic, O.; Blacklock, T. J. Tetrahedron Lett.
1998, 39, 9369−9372. (b) For a reference on acylation using n-BuLi
at low temperature, see: Hoekstra, M. S.; Sobieray, D. M.; Schwindt,
M. A.; Mulhern, T. A.; Grote, T. M.; Huckabee, B. K.; Hendrickson, V.
S.; Franklin, L. C.; Granger, E. J.; Karrick, G. L. Org. Process. Res. Dev.
1997, 1, 26−38.
(52) For a previous report of (R)-17, see: Prashad, M.; Kim, H.-Y.;
Har, D.; Repic, O.; Blacklock, T. J. Tetrahedron Lett. 1998, 39, 9369−
9372.
(53) The chiral auxiliary (R)-16 had a specific rotation [α]18D +63.2°
(c = 1) and melting point = 87.6−88.0 °C.
(54) Ethanol was used previously in place of 2-propanol to crystallize
the product but requires more solvent to azeotrope the toluene, and
the isolated yields are lower due to the greater solubility of (R)-17 in
EtOH.
(55) Beck, A. K.; Sebesta, R.; Seebach, D. Org. Synth. 2008, 85, 295−
306.
(56) Azeotropic removal of water from Boc-aminal 18 via distillation
from toluene was critical in order for the asymmetric Mannich reaction
to proceed to completion.
(37) Evans, D. A.; Urpi, F.; Somers, T. C.; Clark, J. S.; Bilodeau, M.
T. J. Am. Chem. Soc. 1990, 112, 8215−8216.
(38) If the reaction mixture was warmed to 0 °C before the reaction
was complete, Boc group deprotection was observed and resulted in a
lower yield of the reaction.
(39) Evans, D. A.; Britton, T. C.; Ellman, J. A. Tetrahedron Lett. 1987,
28, 6141−6144.
(40) Recrystallization of carboxylic acid (S)-20 from acetonitrile was
also successful in increasing the enantiopurity of the product from 97%
ee to >99% ee; however, the product was more hygroscopic than the
(S)-20-Na crystallized from 2-propanol and made the filtration more
difficult.
(41) Although a coupling reaction between the unprotected
carboxylic acid (S)-20 and deprotected trans-1 was possible, a Boc
group was required for purification of the penultimate (S,R,R)-22 via
crystallization.
(42) Extraction of the piperazine (R,R)-21 product from an acidic or
basic aqueous phase was not possible due to the high aqueous
solubility.
(43) (a) Hexafluorophosphate (PF6) anion is not listed as a
pharmaceutically acceptable salt: Haynes, D. A.; Jones, W.;
Motherwell, W. D. S. J. Pharm. Sci. 2005, 94, 2111−2120. (b) See
also: Wigman, L.; Remarchuk, T.; Gomez, S. R.; Kumar, A.; Dong, M.
W.; Medley, C. D.; Chetwyn, N. P. Byproducts of Commonly Used
Coupling Reagents: Origin, Toxicological Evaluation and Methods for
Determination. Am. Pharm. Rev. 2014, 17, No. 1.
1
(44) The presence of TMU and PF6 anion was detected using H
and 19F NMR spectroscopy of a concentrated aliquot of the organic
phase during the extraction. See Supporting Information S1 for the
analytical details.
(45) Other aqueous solutions were also found effective in removing
the PF6 anion, including 10% aq LiOH and 10% aq Na2CO3.
(46) See Supporting Information S1 for the preparation of Ipatasertib
free-base and the structural determination of β-elimination product 23
by NMR spectroscopy. A second minor impurity, M416 (RRT =
0.80), with mass [M+H]+ = 416 amu, was also formed in ∼0.2% by
HPLC.
(47) This result was surprising since, in general, the leaving group
ability of a dialkyl amino group in a β-elimination reaction is poor:
Marshall, D. R.; Thomas, P. J.; Stirling, C. J. M. J. Chem. Soc., Chem.
Commun. 1975, 23, 940−941. Investigation into the mechanism of β-
elimination of Ipatasertib and various related analogues is ongoing.
(48) For a similar strategy used to prepare a mono-HCl salt of a
dibasic API, see: Caine, D. M.; Paternoster, I. L.; Sedehizadeh, S.;
Shapland, D. P. Org. Process. Res. Dev. 2012, 16, 518−523.
(49) A major side product obtained using the combination of benzyl
chloride and K2CO3 was the six-membered-ring lactone that was
observed up to 10 A% (by LCMS). Lactone formation was not
observed using the current conditions.
(50) The crude ketone 12 was co-evaporated with DCM to remove
residual EtOAc and 2-MeTHF prior to the subsequent asymmetric
transfer hydrogenation step since the presence of these solvents
resulted in a lower %de of trans-1.
(51) In a later campaign, ketone (R)-12 (11.3 kg) that was free of
over-reduced product was reduced under similar reaction conditions
except using 1 mol% of [(R,R)-TsDPEN−Ru(p-cymene)Cl] catalyst
and gave a 96:4 trans/cis-1 mixture in 93% assay yield. After aqueous
work-up and solvent exchange into MeOH, scavenging with 50 wt%
SiliaMetS thiol at 55 °C for 18 h resulted in 64 ppm Ru by ICP-AES.
O
dx.doi.org/10.1021/op500270z | Org. Process Res. Dev. XXXX, XXX, XXX−XXX