Enantioselective Synthesis of a γ-Secretase Modulator via Vinylogous Dynamic Kinetic Resolution

Article abstract of DOI:10.1021/acs.joc.8b01734

Two efficient asymmetric routes to γ-secretase modulator BMS-932481, under investigation for Alzheimer's disease, have been developed. The key step for the first route involves a challenging enantioselective hydrogenation of an unfunctionalized trisubstituted alkene to establish the benzylic stereocenter, representing a very rare case of achieving high selectivity on a complex substrate. The second route demonstrates the first example of a vinylogous dynamic kinetic resolution (VDKR) ketone reduction, where the carbonyl and the racemizable stereocenter are not contiguous, but are conjugated through a pyrimidine ring. Not only did this transformation require both catalyst and substrate control to correctly establish the two stereocenters, but it also necessitated that the nonadjacent benzylic center of the ketone substrate be more acidic than that of the alcohol product to make the process dynamic. DFT computations aided the design of this novel VDKR pathway by reliably predicting the relative acidities of the intermediates involved.

Full text of DOI:10.1021/acs.joc.8b01734

Article
pubs.acs.org/joc
Cite This: J. Org. Chem. XXXX, XXX, XXX−XXX
Enantioselective Synthesis of a γ‑Secretase Modulator via
Vinylogous Dynamic Kinetic Resolution
Neil A. Strotman,* Antonio Ramirez, Eric M. Simmons, Omid Soltani, Andrew T. Parsons, Yu Fan,
James R. Sawyer, Thorsten Rosner, Jacob M. Janey, Kristy Tran, Jun Li, Thomas E. La Cruz,
Charles Pathirana, Alicia T. Ng, and Joerg Deerberg
Chemical and Synthetic Development, Bristol-Myers Squibb, 1 Squibb Drive, New Brunswick, New Jersey 08903, United States
S
* Supporting Information
ABSTRACT: Two efficient asymmetric routes to γ-secretase modulator BMS-932481, under investigation for Alzheimer’s
disease, have been developed. The key step for the first route involves a challenging enantioselective hydrogenation of an
unfunctionalized trisubstituted alkene to establish the benzylic stereocenter, representing a very rare case of achieving high
selectivity on a complex substrate. The second route demonstrates the first example of a vinylogous dynamic kinetic resolution
(VDKR) ketone reduction, where the carbonyl and the racemizable stereocenter are not contiguous, but are conjugated through
a pyrimidine ring. Not only did this transformation require both catalyst and substrate control to correctly establish the two
stereocenters, but it also necessitated that the nonadjacent benzylic center of the ketone substrate be more acidic than that of
the alcohol product to make the process dynamic. DFT computations aided the design of this novel VDKR pathway by reliably
predicting the relative acidities of the intermediates involved.
hydrogenation of trisubstituted cyclopentenes C, and (c) a
dynamic kinetic resolution of β-aryl ketones D. This work
enabled the development of robust, sustainable, and scalable
approach toward (S)-1.
INTRODUCTION
■
γ-Secretase modulator BMS-932481 ((S)-1) is a drug
candidate for the treatment of Alzheimer’s disease, a
degenerative brain disorder that represents the most common
cause of dementia (Scheme 1).1,2 The highly nitrogenated
compound is composed of a cyclopenta[d]pyrimidine core
(A), a bis-benzylic stereocenter in the (S) configuration, and
two different amine substituents at the C2 and C4 positions of
the pyrimidine ring. The initial process from our discovery
team involved a synthesis of racemic product 1 followed by
chromatographic separation on a chiral phase, and was capable
of delivering small quantities of (S)-1 in high enantiopurity.3
However, a sustainable, long-term solution would be required
to support the high demand of (S)-1 anticipated for
commercialization.
To efficiently synthesize (S)-1, we envisioned a convergent
strategy that established the C7 stereocenter of the cyclopenta-
[d]pyrimidine core A and selectively connected the triazol-1-
yl-phenylamino group with fragment A through the C2−N
bond in a later stage (Scheme 1). Herein we describe three
approaches to the preparation of cyclopenta[d]pyrimidine core
A, which explored (a) an enantioselective intramolecular
hydroacylation of unactivated alkene B, (b) an enantioselective
RESULTS AND DISCUSSION
■
Synthesis of (S)-1 by Enantioselective Hydroacyla-
tion. Although (S)-1 could be obtained in high enantiomeric
purity using the chromatographic separation approach, focus
was shifted to evaluating asymmetric options which could
provide a streamlined route with improved yields and greater
synthetic efficiency. Accordingly, a new route was devised that
exploited the use of commercially available trichloropyrimidine
34 as a highly functionalized precursor (Scheme 2). Adaptation
of an established procedure5 facilitated the regioselective
installation of the methylamine substituent at C4 to prepare
methylamino pyrimidine 4. Suzuki coupling of 4 with α-vinyl
boronate 56 afforded pyrimidin-5-carbaldehyde derivative 6.
The key transformation in this sequence was envisioned as an
asymmetric hydroacylation7 of 6 to construct the C2−C6 bond
Received: July 9, 2018
© XXXX American Chemical Society
A
DOI: 10.1021/acs.joc.8b01734
J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
Scheme 1. Retrosynthetic Analysis for (S)-1
and install the C7 stereogenic center of ketone 7. Reduction of
7 to intermediate 8 would be followed by the substitution at
the chloropyrimidine moiety to form (S)-1.8
Scheme 2. Potential Route to (S)-1 through
Enantioselective Hydroacylation
A promising precedent for the enantioselective cyclization
was reported by Morehead and co-workers using cationic
Rh(I)BINAP catalysts for the hydroacylation of 2-vinyl
benzaldehydes.9 Inspired by this work, we conducted an
extensive screen of chiral Rh(I) complexes for the hydro-
acylation of 6 (Table 1).10 Unfortunately, using analogous
conditions to those reported by Morehead gave poor
conversions and selectivities (entries 1 and 2). On the basis
of the beneficial effect of Brønsted acid additives observed for
related transformations,11 HBF4 was explored as an additive
and was found to greatly improve the rate of hydroacylation,
but still gave racemic 7 (entries 3 and 4).12,13 Interestingly, the
2,4-dimethoxy analog 9 underwent hydroacylation without the
Table 1. Enantioselective Hydroacylation of 6 and 9
a
b
b
entry
substrate
ligand
time (h)
additive (mol %)
conv. (%)
ee (%)
0
1
2
3
4
5
6
7
8
6
6
6
6
9
9
9
9
R-BINAP
R-TolBINAP
R-BINAP
R-TolBINAP
R-BINAP
R-TolBINAP
R-Xyl-SDP
R-DuanPhos
21
21
21
21
16
16
16
16
18
0
HBF4
HBF4
78
60
85
100
90
99
0
3
3
95
94
95
a
b
Substrate (0.1 M in 1,2-dichloroethane), ([Rh(nbd)2]BF4 (18 mol %), and ligand (20 mol %) at 80 °C, 16−24 h. Determined by HPLC analysis
of the crude reaction mixture.
B
DOI: 10.1021/acs.joc.8b01734
J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
requirement of acidic additives to afford cyclopenta[d]-
pyrimidin-5-one 10 with excellent yields and enantioselectiv-
ities (e.g., entries 6−8) revealing that the steric and electronic
properties of the substituents on the pyrimidine ring play a
critical role in the success of the annulative process. A two-step
reduction of ketone 10 afforded cyclopenta[d]pyrimidine 11
with limited epimerization (Scheme 3). However, under the
trisubstituted alkene, if reduced enantioselectively, could lead
to the desired cyclopenta[d]pyrimidine core (S)-8.
Synthesis of (S)-1 by Enantioselective Hydrogena-
tion. While introducing the C7 stereocenter via an
enantioselective hydrogenation of 14 was conceptually feasible,
highly selective asymmetric hydrogenations of minimally
functionalized cyclic alkenes remain an unmet challenge in
asymmetric catalysis.18 The search for an effective enantiose-
lective hydrogenation of cyclopentene 14 started with a series
of high-throughput screening experiments using Ir and Rh
catalysts (Table 3). Although Ir catalysts have found the most
success in the hydrogenation of unfunctionalized alkenes,19 the
chiral Ir catalysts examined for the hydrogenation of 14 gave
poor conversion and enantioselectivity (entry 1).20 In contrast,
chiral Rh catalysts such as catASium M complex Rh-1 provided
higher conversion, albeit with modest enantioselectivity (entry
2).21 This promising reactivity prompted the study of a series
of Rh(I) catalysts generated in situ from [Rh(nbd)2]BF4 and
chiral bisphosphine ligands.22 Among the numerous ligands
examined, bulky, electron-rich axially chiral bisphosphine
ligands such as BIPHEP and SEGPHOS offered the most
encouraging results (entries 3−8). In particular, a combination
of [Rh(nbd)2]BF4 with DTBM-SEGPHOS-R (L6) gave (S)-8
in 82% conversion and 76% ee (entry 8). Along with the
desired (S)-8, significant amounts of dechlorinated byproduct
15 were also observed.23
With a functional catalyst system in hand (i.e., Rh/L6),
further optimization focused on minimizing the formation of
byproduct 15, as well as lowering the Rh loading and
improving the reaction performance. Numerous Rh(I) catalyst
precursors,24 solvents,25 additives,26 and hydrogen pressures
were evaluated (Figure 1). For the hydrogenation using Rh-L6,
ethereal and ester solvents were found to be optimal, with
MeOAc providing the best conversion and selectivity (entries
10−12). Optimization of the solvent enabled a 9-fold
reduction in catalyst loading and minimized the levels of
dechlorinated byproduct 15. Since negligible difference in
performance was observed among different Rh sources (Table
S6), [Rh(cod)Cl]2 was chosen to conduct subsequent studies
due to its lower cost and greater air stability (entry 13). A
survey of Brønsted and Lewis acid additives (Table S7)
indicated that reactions conducted in the presence of KOTf
under high H2 pressures (≥1000 psi) provided better
conversion and selectivities, giving full conversion to (S)-8 in
90% ee after 24 h (entries 14−15).27 The optimized conditions
using Rh-L6 were performed on a multigram scale to prepare
(S)-8 in 69% yield and >99% ee.
Scheme 3. Attempt to Prepare Enantiopure 2,4-
Dichloropyrimidine 12 from 10
conditions to prepare 2,4-dichloropyrimidine 12, racemization
of the C7 stereocenter occurred giving racemic product. DFT
calculations predicted a pKa of 14 for the benzylic proton of
12, explaining the configurational lability, and halting further
investigation of the enantioselective hydroacylation ap-
proach.14−16
Despite this setback, we envisioned that racemic ketone 7
could be a valuable intermediate for other asymmetric routes.
To obtain 7 via acid-catalyzed cycloisomerization of 6, we
evaluated 120 Brønsted and Lewis acids, and various other
catalysts (Table 2).17 Although Brønsted acids performed
Table 2. Acid-Catalyzed Cyclization of Pyrimidine-5-
carbaldehyde 6
a
b
b
entry
acid
HCl
MsOH
ZnCl2
BF3−AcOH
AlCl3
HfCl4
GaCl3
TiCl4
product (%)
conv. (%)
c
1
3
6
2
60
55
30
23
21
81
80
30
9
3
78
79
36
27
98
98
88
2
3
4
5
6
7
d
8
9
10
SnCl4
ZrCl4
a
Reaction conditions: 0.5 equiv catalyst, 25 vol DCE, 40 °C, 16 h.
b
Determined by HPLC analysis of the crude reaction mixture. On the
basis of mol % of product or remaining SM vs an internal standard.
c
d
Anhydrous HCl in dioxane. Decomposed to several unidentified
Nucleophilic aromatic substitution of final intermediate (S)-
8 with triazol-1-yl-aniline 2 under acidic conditions (H2SO4,
NMP, 100 °C) led to substitution at the 2-chloro position to
give (S)-1 in 86% yield and >99% ee.28,29 Importantly, there
was no loss of enantiopurity, demonstrating a viable
enantioselective route that provided (S)-1 in >99% ee and
16.3% overall yield over 7 steps.30
Synthesis of (S)-1 by Dynamic Kinetic Resolution.
Despite the development of a successful synthesis through
alkene hydrogenation, our team investigated if a shorter
synthesis, which would not require Rh, an expensive ligand, or
high-pressure (1000 psi) capabilities could be developed. A
potential route would involve kinetic resolution of racemic 7
through ketone reduction (Scheme 5). A variety of bio- or
chemocatalysts could affect this transformation, but as with any
kinetic resolution, the yield would be capped at 50%.31
byproducts.
poorly, giving only trace cyclized product (entries 1 and 2),
metal chloride Lewis acids were effective for the cyclization of
6. Among the Lewis acids, weakly electrophilic species such as
ZnCl2 showed almost no reactivity (entry 3) while strongly
electrophilic species such as SnCl4 were the most effective
(entry 9). Due to the toxicity and environmental impact of Sn,
we chose to proceed with ZrCl4 (entry 10) and conditions
were optimized to provide cyclopenta[d]pyrimidin-5-one 7 in
90% yield using 10 mol % ZrCl4 in DCM at 40 °C. Reduction
of 7 with DIBAL-H gave a mixture of diastereomeric alcohols
13, which upon dehydration with Eaton’s reagent (MSA/
P2O5) selectively afforded cyclopentene 14 (Scheme 4). This
C
DOI: 10.1021/acs.joc.8b01734
J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
Scheme 4. Route to (S)-1 via Enantioselective Hydrogenation
Table 3. Optimization of Enantioselective Hydrogenation of Alkene 14
a
b
b
b
entry
metal precursor
metal (mol %)
ligand
solvent
H2 (psi)
additive (mol %)
conv. (%)
ee (%)
15 (%)
1
2
3
4
5
6
7
8
9
10
11
12
Ir-1
Rh-1
5.0
5.0
4.5
4.5
4.5
4.5
4.5
4.5
0.8
0.8
0.8
0.5
0.5
0.75
0.50
DCE
750
750
750
750
750
750
750
750
750
750
750
750
750
1050
1000
17
55
35
40
32
36
53
82
6
93
89
>99
97
>99
>99
+18
−26
+40
−49
−23
+60
+66
+76
MeOH
MeOH
MeOH
toluene
toluene
toluene
toluene
DMAc
DME
EtOAc
MeOAc
MeOAc
MeOAc
MeOAc
2.8
4.8
6.3
8.4
4.3
5.6
5.7
[Rh(nbd)2]BF4
[Rh(nbd)2]BF4
[Rh(nbd)2]BF4
[Rh(nbd)2]BF4
[Rh(nbd)2]BF4
[Rh(nbd)2]BF4
[Rh(nbd)2]BF4
[Rh(nbd)2]BF4
[Rh(nbd)2]BF4
[Rh(nbd)2]BF4
[Rh(cod)Cl]2
[Rh(cod)Cl]2
[Rh(cod)Cl]2
L1
L2
L3
L4
L5
L6
L6
L6
L6
L6
L6
L6
L6
c
c
+89
+87
+86
+85
+90
+90
0.9
0.9
0.7
<0.1
2.1
0.8
c
d
d
13
14
15
KOTf (2.3)
KOTf (0.7)
a
Reaction conditions: 14 (10−100 μmol), metal precursor, ligand (1.1 ligand:metal ratio), solvent (125−500 μL), 40 °C, 750−1050 psi H2, 15−24
b
c
d
h. Determined by HPLC analysis of the crude reaction mixture. 30 °C. 50 °C.
Alternatively, we hypothesized that a dynamic kinetic
resolution (DKR) via ketone reduction could be effective,
allowing for both enantiomers of ketone 7 to be converted to
chiral alcohol 13, and offering the potential for higher yield and
selectivity.32 Rather than trying to avoid epimerization of the
benzylic stereocenter of 7 as in the asymmetric hydroacylation
approach, we could embrace this reactivity and possibly use it
to our advantage. Typically, DKR of ketones via reduction
involves epimerization of an alpha stereocenter, which is more
acidic in the ketone substrate than in the alcohol product (e.g.,
Scheme 6).33 The selective epimerization of the substrate is
driven by a large ΔpKa, allowing for the product to be
configurationally stable. This type of process is well established
for contiguous centers, but no such 1,2 relationship existed
between the ketone and the C7 benzylic center of 7. We
considered that the 1,4 relationship of the two centers through
the pyrimidine ring would allow for epimerization of C7 via a
dienolate intermediate (Scheme 7). This type of reactivity
would be consistent with the principle of vinylogy introduced
by Ludwig Claisen in 1926, involving the transmission of
electronic effects through a conjugated organic bonding
system.34 However, a vinylogous DKR process (VDKR)
would only be possible if there were a significant difference
in pKa between ketone 7 and alcohol 13, as well as a steric bias
imparted by the 4-fluorophenyl substituent.
At the outset, we employed computational predictions to
assess the viability of the VDKR. DFT calculations predicted a
pKa of 14 for the benzylic proton of ketone 7, a pKa of 21 for
the benzylic proton of the trans-alcohol (5S,7R)-13, and a pKa
of 22 for the cis-alcohol (5S,7S)-13.14 Importantly, the
carbonyl bond length increased considerably going from the
ketone (1.229 Å) to the ketone enolate (1.251 Å), consistent
with dienolate character leading to anion stability. With the
hope that a pKa difference of approximately 7−8 units could be
D
DOI: 10.1021/acs.joc.8b01734
J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
Figure 1. Selected Complexes and Ligands Evaluated for the Enantioselective Hydrogenation of 14.
Scheme 5. Kinetic Resolution of Racemic Ketone 7
sufficient for a successful VDKR process, we began
experimental investigations into this approach. The correct
choice of catalyst system would be necessary to allow for
epimerization of the starting material ketone 7, but not the
product alcohols 13.
The Noyori transfer hydrogenation catalyst RuCl(p-
cymene)(S,S)-TsDPEN (Ru-1)35 was selected in the initial
attempt at this novel transformation. This catalyst can function
with mixtures of formic acid and triethylamine under either
acidic or basic conditions, allowing the pH of the medium to
be tuned.36 A promising result under mildly acidic conditions
Scheme 6. DKR of a Representative α-Keto ester
1
showed high conversion and a 29:1 isomer ratio by H NMR
analysis (Scheme 8). The cis configuration was established
Scheme 8. VDKR of Ketone 7 with a Transfer
Hydrogenation Catalyst
Scheme 7. Potential Vinylogous DKR with a 1,4-
Relationship between Ketone and the Benzylic Center
through detection of NOE interaction between the protons at
C5 and C7. The high cis-selectivity is indicative of inherent
substrate control where the catalyst approaches the face
opposite to the 4-fluorophenyl substituent at C7. Additionally,
chiral HPLC analysis showed 99% ee for this transformation,
consistent with high facial selectivity associated with the
transfer hydrogenation catalyst for acetophenone-type sub-
strates. With a successful proof of concept for the VDKR, we
directed our efforts to further optimize this transformation.
A screen of solvents, amines, and acid−base ratios was
undertaken to determine if the selectivity and reactivity could
be further improved (Table 4).37 Except for i-PrOH (entry 2),
which offered poor solubility, all of the other solvents gave
high conversion and high levels of enantio- and diastereose-
lectivity. However, dechlorinated species 15 was observed as a
prominent byproduct and varied greatly with solvent choice.
Alcohol solvents such as MeOH provided the highest levels of
dechlorination (entry 5) while polar aprotic solvents such as
DMAc provided the lowest amounts (i.e., 0.2%, entry 6).
While dehalogenation is common for Ru hydrogenation
E
DOI: 10.1021/acs.joc.8b01734
J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
Table 4. Screen of Ru-Catalyzed Transfer Hydrogenation Conditions
a
c
c
c
c
entry
solvent
amine
HCO2H:amine
conv. (%)
dr
ee (%)
>99
15 (%)
1
2
3
4
5
6
7
8
MeCN
i-PrOH
DCM
TEA
TEA
TEA
TEA
TEA
TEA
TEA
TEA
TEA
TEA
4:3
4:3
4:3
4:3
4:3
4:3
4:3
4:3
5:3
3:5
3:5
5:3
3:5
97
33
97
98
80
98
85
91
97
98
21
27
94
42
5
15
14
26
60
0.2
25
31
0.3
0.2
5
112
134
31
>99
>99
93
>99
>99
>99
>99
>99
THF
MeOH
DMAc
EtOAc
PhOMe
DMAc
DMAc
DMAc
DMAc
DMAc
67
85
98
56
9
10
11
12
57
N,N-dimethyl-p-toluidine
N,N-dimethyl-p-toluidine
pyridine
5
11
b
13
a
b
[7] = 0.19 M, 18 vol solvent, 40 °C, 16 h, 4% RuCl(p-cymene)(S,S)-TsDPEN (Ru-1). Although all other reactions showed >90% mass balance,
c
the reaction with pyridine gave <1% product and several unidentified impurities. Determined by HPLC analysis of the crude reaction mixture.
Table 5. Screen of Ru-Catalyzed Transfer Hydrogenation Conditions for Reaction Optimization
a
b
[7] = 0.26, 12 vol DMAc, 2:1 HCO2H-Et3N, 40 °C, 16 h. Determined by HPLC analysis of the crude reaction mixture.
catalysts, it has not been previously reported with Noyori-type
transfer hydrogenation catalysts.38 The highly electrophilic
nature of the chloropyrimidine moiety may be responsible for
this unusual behavior of the nucleophilic Ru-hydride. The
effect of the relative pH on the system was evaluated by
varying the formic acid-triethylamine proportion. Reversal of
the acid−base ratio showed no impact on selectivity (cf. entries
9 and 7) demonstrating that, across this range, racemization of
substrate 7 is sufficiently fast and epimerization of product 13
sufficiently slow to obtain a high dr. Weaker amine bases like
N,N-dimethyl-p-toluidine or pyridine gave either very low
conversion or undesired side reactions (entries 11−13).
Although the ee was quite high with most catalysts, the dr
was critical to consider since, upon removal of the hydroxyl
group at C5, both diastereomers with the R configuration at
C7 would converge to the undesired (R)-8 enantiomer.
Using the optimized conditions, we evaluated other Ru
transfer hydrogenation catalysts (Table 5). In general, reducing
F
DOI: 10.1021/acs.joc.8b01734
J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
Scheme 9. Optimized Conditions for the VDKR of Ketone 7 and Deoxygenation of 13
Scheme 10. Two Asymmetric Routes to BMS-932481 ((S)-1)
the catalyst loading offered improved enantio- and diaster-
eoselectivity, but resulted in a decrease in reaction conversion
(entries 1 and 4). Among the catalysts tested, RuCl(mesityl)-
(S,S)-TsDPEN was selected for further development based on
its high reactivity and stereoselectivity, and for its wide
commercial availability (entry 5). Further optimization
reduced the catalyst loading to 0.5% Ru and allowed for the
reaction to be run at high concentrations.39 On a 400 g scale, a
high selectivity of 99.4% ee and 146:1 dr was observed for the
reaction. After workup and isolation, (5S,7S)-13 was obtained
in 84% yield, with 99.7% ee and >200:1 dr (Scheme 9).
With the conditions for the VDKR optimized, only a
deoxygenation of (5S,7S)-13 remained to provide (S)-8. Due
to the importance of maintaining the stereochemical integrity
at the C7 benzylic position, we tested a series of acids and
silanes.40 Although most acids were capable of mediating the
deoxygenation, the highest yield was obtained using a
combination of BF3−OEt2 and Et3SiH, accompanied by no
epimerization at the C7 stereocenter (>99% ee). Additional
optimization led to a reaction in 40:1 DCM-sulfolane at 40 °C
that gave 89% yield without erosion in enantioselectivity. The
final coupling of (S)-8 and 2 under acidic conditions provided
(S)-1 in 88% yield, > 99% purity, and >99% ee (Scheme 10).
The synthetic route using the VDKR gave (S)-1 in 38.2%
overall yield over six steps.41
CONCLUSIONS
■
We have developed two efficient asymmetric routes to γ-
secretase modulator BMS-932481 [(S)-1]. Some important
elements of both routes were the regioselective amination of
trichloropyrimidine 3, the regioselective Suzuki coupling of
dichloropyrimidine 4, and a Lewis acid mediated intra-
molecular hydroacylation to generate the cyclopentanone
fragment of 7. The key asymmetric step for each of the two
routes was capable of setting the C7 benzylic center with high
enantioselectivity. First, through a combination of high-
throughput experimentation and process optimization, a
challenging asymmetric hydrogenation of unfunctionalized
G
DOI: 10.1021/acs.joc.8b01734
J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
2,4,6-trichloropyrimidine-5-carbaldehyde (3) (1.20 kg, 5.45 mol), and
a solution of KHCO3 (0.64 kg, 6.41 mol) in water (3 L). After cooling
the solution to 15 °C, a solution of methylamine (40 wt % in water,
0.44 kg, 5.71 mol, 1.05 equiv) was charged slowly while maintaining
the temperature below 20 °C. After stirring for an additional 12 h, the
DCM layer was set aside and the aqueous layer was extracted with
DCM (2 L). The combined DCM layer was swapped to IPA through
a vacuum distillation. The product crystallized, was collected by
filtration, and was dried under vacuum to give 4 (CAS: 1007197−30−
9)5 as a pale yellow crystalline solid (905 g, 77% yield). MP: 127 °C
(DSC). 1H NMR (500 MHz, CDCl3) δ ppm 10.31 (s, 1H), 9.28 (br
s, 1H), 3.15 (d, J = 5.0 Hz, 3H). 13C NMR (125.75 MHz, CDCl3) δ:
190.5, 165.9, 162.8, 162.3, 106.9, 28.2. IR (cm−1) 3293, 1664, 1606,
1553, 1415, 1370, 1290, 1112, 956, 836.
Preparation of 2-Chloro-4-(1-(4-fluorophenyl)vinyl)-6-
(methylamino)pyrimidine-5-carbaldehyde (6). To a 5 L reactor
were charged 2-MeTHF (1.13 L, 17 vol), 2-(1-(4-fluorophenyl)-
vinyl)-4,4,5,5-tetramethyl-1,3,2-dioxaborolane (5, CAS:850567−55−
4, 455.5 mmol, 113 g), 2,4-dichloro-6-(methylamino)pyrimidine-5-
carbaldehyde (4, 114 g, 546.5 mmol, 1.2 equiv) and (Xantphos)PdCl2
(7.03 g, 9.11 mmol, 2 mol %). The suspension was degassed through
three evacuation/nitrogen backfill cycles and warmed to 50 °C. An
aqueous solution of K3PO4 (4.0 M, 256 mL, 1025 mmol, 2.25 equiv)
was added in one portion via addition funnel, and the mixture was
allowed to stir until the consumption of the vinylboronate was
confirmed by HPLC (approximately 1.5 h). The reaction mixture was
cooled to rt, water (1.1 L) was added and the biphasic system was
stirred for 30 min. The aqueous phase was separated and discarded,
and the organic layer was stirred with an aqueous solution of 2-
(dimethylamino)ethanethiol hydrochloride (10% w/w, 550 mL) for
12 h to remove residual Pd. The aqueous layer was separated and
discarded, and the organic layer was washed with 10% w/w brine (2 ×
550 mL). The 2-MeTHF solvent was swapped to MeCN (1.1 L), and
water (1.1 L) was added to induce crystallization. Isolation of the
trisubstituted alkene 14 was accomplished. While this reaction
is known for substrates with directing groups, the trans-
formation represents a very rare case of achieving high
selectivity on a complex substrate with only aryl and alkyl
substituents. Second, a unique vinylogous dynamic kinetic
resolution of ketone 7 was developed, where the carbonyl and
the epimerizable chiral center were not contiguous but
conjugated through a pyrimidine fragment. This process,
which allowed for both the alcohol and benzylic stereocenters
to be set in a single operation, required catalyst control to add
to the desired face of the carbonyl, as well as substrate control
to add to the face opposite to the aryl substituent. It also
required that the nonadjacent benzylic center of the ketone
substrate be more acidic than that of the alcohol product. DFT
computations aided in the design of the VDKR by predicting
the fulfillment of the relative acidity requirement, despite
attenuation of the electron-withdrawing character of the
ketone in a vinylogous system. This novel application of a
DKR to establish a remote chiral center led to a highly efficient
route to BMS-932481, which was suitable for long-term
implementation.
EXPERIMENTAL SECTION
■
General Materials and Methods. All reactions were carried out
using dry glassware under a nitrogen atmosphere unless otherwise
noted. All reagents and solvents were purchased from commercial
vendors and employed without further purification. 1H NMR
chemical shifts are given in ppm with respect to the residual CDCl3
peak (δ 7.26 ppm), residual DMSO-d6 (δ 2.50 ppm), or an internal
TMS standard (δ 0.00 ppm), 13C{1H} NMR shifts are given in ppm
with respect to CDCl3 (δ 77.16 ppm), or DMSO-d6 (δ 39.52 ppm).
Coupling constants are reported as J-values in Hz. Mass spectral data
were obtained using an Orbitrap mass spectrometer. High-resolution
mass spectra were collected on an Agilent 6200 series TOF/6500
series Q-TOF B.06.01 (B6172 SP1). Melting points were measured
with a capillary melting Stuart SMP10 instrument and are
uncorrected. Moisture determinations were performed with a Karl
Fischer titrator. Infrared spectra were recorded on an FT instrument
at 4.00 cm−1 resolution. Optical rotation values were measured were
recorded on a PerkinElmer Model 341 polarimeter (1 mL cell, 1 dm
path length); concentration (c) is in g/100 mL and [α]D values are in
degrees. HPLC analyses were performed using reversed-phase
techniques. Chiral and achiral analysis was performed on a Shimadzu
LC-20 AT liquid chromatograph with conditions as noted for
individual compounds. Area percent (AP) refers to HPLC area
percent purity.
solids by Buchner filtration followed by drying overnightin a vacuum
̈
oven at 50 °C provided pyrimidin-5-carbaldehyde derivative 6 (131 g,
1
82% yield). MP: 152−153 °C. H NMR (500 MHz, CDCl3) δ: ppm
9.97 (s, 1H), 9.13 (br s, 1H), 7.37−7.28 (m, 2H), 7.08−6.97 (m,
2H), 6.03 (s, 1H), 5.46 (s, 1H), 3.15 (d, J = 5.0 Hz, 3H). 13C NMR
(125.75 MHz, CDCl3) δ: 191.7, 174.6, 164.0, 162.9 (d, J = 232),
162.3, 142.5, 134.0 (d, J = 3.4 Hz), 128.4 (d, J = 8.3 Hz), 120.2, 115.9
(d, J = 21.8 Hz), 109.06, 27.8. IR (cm−1) 3310, 1655, 1610, 1437,
1303, 1228, 1170, 965, 809, 711. HRMS (ESI-TOF)m/z: [M + H]+
calcd for C14H11ClFN3O 292.0647; found 292.0662.
Preparation of 2-Chloro-7-(4-fluorophenyl)-4-(methylami-
no)-6,7-dihydro-5H-cyclopenta[d]pyrimidin-5-one (7). To a
250 mL reactor were charged DCM (120 mL, 6 vol) and
pyrimidin-5-carbaldehyde derivative 6 (18 g, 62 mmol). Once fully
dissolved, ZrCl4 (1.6 g, 6.9 mmol, 0.10 equiv) was added in a single
portion, and the reaction mixture was heated at reflux for 16 h. While
still near reflux, IPA (120 mL, 6 vol) was added dropwise over 1 h,
and then aged for 1 h. The resulting suspension was cooled to 5 °C
and filtered, washed with IPA (2 × 60 mL), and dried under vacuum
to give pyrimidin-5-one 7 as a pale yellow crystalline solid (16.6 g,
92% yield). MP: 230 °C (dec). 1H NMR (500 MHz, CDCl3) δ: ppm
7.35 (d, J = 7.3 Hz, 1H), 7.15−7.05 (m, 2H), 7.05−6.96 (m, 2H),
4.41 (dd, J = 8.0, 3.1 Hz, 1H), 3.21 (dd, J = 19.2, 8.0 Hz, 1H), 3.17
(d, J = 5.1 Hz, 3H), 2.66 (dd, J = 19.2, 3.1 Hz, 1H). 13C NMR
(125.75 MHz, CDCl3) δ: 202.6, 186.3, 166.5, 162.1 (d, J = 246 Hz),
159.8, 135.5 (d, J = 3.4 Hz), 129.2 (d, J = 8.2 Hz), 116.0 (d, J = 21.3
Hz), 110.8, 45.6, 45.3, 27.5. IR (cm−1) 3350, 1686, 1628, 1570, 1424,
1317, 1228, 1116, 912, 809. HRMS (ESI-TOF) m/z: [M + H]+ calcd
for C14H11ClFN3O 292.0647; found 292.0659.
Catalyst and ligand screenings were carried out in a nitrogen-filled
glovebox. A 96-well block was loaded with 1 mL glass vials containing
the appropriate preformed catalyst (0.5 μmol) or ligand (0.5 μmol for
bidentate ligands, 1.0 μmol for monodentate ligands). To each vial
containing ligand only was added 45 μL (0.45 μmol) of a solution of
the appropriate metal precursor, and the resulting mixture was aged at
room temperature for 20 min. A solution of the substrate was then
added (10 μmol per vial), and the resulting mixtures were
concentrated to dryness using a Genevac vacuum centrifuge. A
micro stir bar was charged to each vial, followed by solvent (125−150
μL). For the enantioselective hydrogenation screening, the 96-well
block was placed inside an aluminum hydrogenation block, which was
then sealed under nitrogen and removed from the glovebox. The
hydrogenation block was set on a shaker plate and placed under a
hydrogen atmosphere. The pressurized block was then heated to
target temperature with orbital shaking for 15−20 h. Upon cooling to
room temperature and venting the hydrogen atmosphere, the reaction
mixtures were concentrated to dryness using a Genevac vacuum
centrifuge, redissolved in MeCN and analyzed by chiral HPLC.
Preparation of 2,4-Dichloro-6-(methylamino)pyrimidine-5-
carbaldehyde (4). To a 20 L reactor were charged DCM (12 L),
Preparation of 4-(1-(4-Fluorophenyl)vinyl)-2,6-dimethoxy-
pyrimidine-5-carbaldehyde (9). To a 50 mL three-necked flask
were charged 2-MeTHF (22 mL), 2-(1-(4-fluorophenyl)vinyl)-
4,4,5,5-tetramethyl-1,3,2-dioxaborolane (5, 8.60 mmol, 2.13 g), 4-
chloro-2,6-dimethoxypyrimidine-5-carbaldehyde (CAS:134221−52−
6, 2.09 g, 10.3 mmol, 1.2 equiv) and (Xantphos)PdCl2 (133 mg, 0.172
mmol, 2 mol %). The suspension was degassed and warmed to 50 °C.
H
DOI: 10.1021/acs.joc.8b01734
J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
An aqueous solution of K3PO4 (4.0 M, 4.8 mL) was added and the
reaction mixture was stirred for 4 h. After allowing to cool to 20 °C,
the organic phase was washed with water (20 mL) and then stirred
with an aqueous solution of 2-(dimethylamino)ethanethiol hydro-
chloride (10% w/w, 10 mL) for 18 h. After removing the aqueous
layer, the organic layer was washed with 10% w/w brine (2 × 10 mL)
and was concentrated by rotary evaporation. The residue was
dissolved in MeCN (20 mL), and water (20 mL) was added to
induce crystallization. The crystals were isolated as an off-white
powder of pyrimidin-5-carbaldehyde derivative 9 (2.20 g, 74% yield).
1H NMR (400 MHz, CDCl3) δ: ppm 7.28−7.32 (m, 2H), 7.01 (t, J =
10.0 Hz, 2H), 5.92 (bs, 1H), 5.42 (bs, 1H), 4.15 (s, 3H), 4.08 (s,
3H). 13C NMR (100.6 MHz, CDCl3) δ: 186.9, 174.2, 171.3, 165.0 (d,
point 2,4-dichloropyrimidine 12 crystallized as a red solid (502 mg,
43% yield). H NMR (400 MHz, DMSO-d6) δ: ppm 7.29 (m, 1H),
7.16 (m, 1H), 4.57 (t, J = 8.0 Hz, 1H), 2.90−3.10 (m, 2H), 2.66 (m,
1H), 2.11 (m, 1H); 13C NMR (100.5 MHz, DMSO-d6) δ: 180.4,
162.4, 160.0, 157.5, 156.9, 137.6, 133.8, 130.3, 130.2, 115.4, 115.2,
50.8, 32.2, 26.86. HRMS (ESI-TOF) m/z: [M + H]+ calcd for
C13H9Cl2FN2 283.0200; found 283.0192.
1
Preparation of 2-Chloro-7-(4-fluorophenyl)-N-methyl-5H-
cyclopenta[d]pyrimidin-4-amine (14). Step 1. Preparation of
2-Chloro-7-(4-fluorophenyl)-4-(methylamino)-6,7-dihydro-5H-
cyclopenta[d]pyrimidin-5-ol (rac-13). . A 100 mL three-necked flask
equipped with a thermocouple was charged with pyrimidin-5-one 7 (4
g, 13.75 mmol) and DCM (50 mL). The flask was cooled to 0 °C,
evacuated-nitrogen filled, and 16.5 mL of 1 M DIBAL-H
incyclohexane (16.5 mmol, 1.2 equiv) was added via syringe while
maintaining the internal temperature below 10 °C. Upon complete
addition, a reddish orange slurry formed that was aged at the same
temperature for 2 h. Then, a solution of citric acid (65 mL, 20 wt
%/vol in water, 68.75 mmol, 5 equiv) was added via addition funnel
(CAUTION!: Vigorous reaction! Gas evolution!), and the mixture
was transferred to a separatory funnel. Extraction of the phases with
DCM (2 × 40 mL) followed by concentration under vacuum afforded
a solid residue that was slurried in 5 vol MTBE, filtered and dried (50
°C, 20 mmHg) to give crude pyrimidin-5-ol 13 as a 9:1 mixture of
J
C−F = 194 Hz), 161.5, 144.4, 134.3, 128.3 (d, JC−F = 8.0 Hz), 118.9,
115.5 (d, JC−F = 21.1 Hz), 110.8, 55.6, 55.0. LC-MS (ESI-MS) m/z:
[M + H]+ calcd for C15H14FN2O3 289.10; found 289.10.
Preparation of 7-(4-Fluorophenyl)-2,4-dimethoxy-6,7-dihy-
dro-5H-cyclopenta[d]pyrimidin-5-one (10). Step 1: Catalyst
Preparation. To a 40 mL scintillation vial were charged [Rh(nbd)2]-
BF4 (380 mg, 1.0 mmol), R-TolBINAP (700 mg, 1.3 mmol), and 1,2-
dichloroethane (10 mL) in a nitrogen-filled glovebox to form a deep
red solution that was reduced at rt under 2 atm H2 for 30 min.
Step 2: Hydroacylation. The catalyst mixture was added under
nitrogen to a vial containing aldehyde 9 (1.5 g, 5.2 mmol) in 1,2-
dichloroethane (10 mL). The resulting solution was heated to 75 °C
and allowed to stir at the same temperature for 16 h. The crude
reaction was filtered through a silica plug and purified by flash
chromatography (hexanes-AcOEt gradient) to afford ketone 10 as a
yellow solid (1.26 g, 84% yield, 95% ee). 1H NMR (400 MHz,
CDCl3) δ: ppm 7.01−7.08 (m, 2H), 6.93 (t, J = 10.0 Hz, 2H), 4.34
(dd, J = 4.5, 4.0 Hz, 1H), 4.09 (s, 3H), 3.94 (s, 3H), 3.14 (dd, J =
18.0, 8.0 Hz, 1H), 2.64 (dd, J = 18.0, 4.0 Hz, 1H). 13C NMR (100.6
MHz, CDCl3) δ: 198.5, 189.6, 169.0, 167.9, 163.2, 160.7, 136.1,
129.1, 115.9, 115.7, 110.9, 55.8, 54.9, 45.4, 45.4. LC-MS (ESI-TOF)
m/z: [M + H]+ calcd for C15H14FN2O3 289.0988; found 289.0966.
Preparation of 2,4-Dichloro-7-(4-fluorophenyl)-6,7-dihy-
dro-5H-cyclopenta[d]pyrimidine (12). Step 1: Ketone to Alcohol
Reduction. NaBH4 (570 mg, 15 mmol) was added to a solution of
ketone 10 (1.2 g, 4.2 mmol) in MeOH (40 mL), and the resulting
slurry was stirred for 30 min at rt. The reaction was analyzed by
HPLC to confirm consumption of the ketone and formation of the
alcohol intermediate as a 6:1 mixture of diastereomers. The reaction
mixture was concentrated by rotary evaporation to give a crude oil. 1H
NMR (400 MHz, CDCl3) δ: 7.24 (m, 2H, major), 7.08 (m, 2H,
minor), 6.96 (m, 2H, major), 5.35 (m, 1H, major), 4.48 (m, 1H,
minor), 4.11 (m, 1H, major), 4.04 (s, 3H, major), 3.89 (s, 3H, major),
2.98 (m, 1H, major), 2.56 (m, 1H, minor), 2.38 (m, 1H, minor), 2.06
(m, 1H, major).
1
diastereomers (2.7 g, 67% yield). H NMR (400 MHz, CDCl3) δ:
ppm 7.15 (m, 2H, major), 7.09 (m, minor) 7.00 (m, 2H, major), 6.96
(m, minor), 5.77 (br s, 1H, major), 5.67 (br s, minor), 5.44 (q, J = 6.0
Hz, minor), 5.25 (q, J = 6.0 Hz, 1H, major), 5.40 (m, minor), 4.08 (t,
J = 8.0 Hz, 1H, major), 3.13 (m, 3H, minor), 3.26 (m, 3H, major),
1.88 (m, 2H).
Step 2. Preparation of Pyrimidin-4-amine (14). The crude 13
from the previous step was dissolved in 1,2-DCE (200 mL) and
Eaton’s reagent (67.5 g, 18.4 mmol, 2 equiv) was added under
nitrogen at rt. The resulting mixture was stirred overnight at rt and
quenched with aqueous K2CO3 (20% w/v) to adjust the pH to 7−8.
Extraction with DCM (2 × 100 mL) and evaporation under vacuum
afforded a solid that was slurried in MTBE (10 mL) and stirred at rt
for 3 h. The product was then isolated by filtration and dried in a
1
vacuum oven at 40 °C overnight to (2.28 g, 90% yield). H NMR
(500 MHz, DMSO-d6) δ: ppm 8.07 (m, 2H), 7.29 (s, 1H), 7.27 (t, J =
10.0 Hz, 2H), 5.22 (br s, 1H), 3.38 (br s, 2H), 2.90 (s, 3H). 13C
NMR (125.75 MHz, CDCl3) δ: 167.2, 162.8, 160.9, 159.1, 158.2,
139.6, 137.0, 129.6, 129.2, 116.0, 115.3, 115.1, 33.4, 27.2 HRMS
(ESI-TOF) m/z: [M + H]+ calcd for C14H11ClFN3 276.0698; found
276.0710.
Preparation of (S)-2-Chloro-7-(4-fluorophenyl)-N-methyl-
6,7-dihydro-5H-cyclopenta[d]pyrimidin-4-amine [(S)-8]
through enantioselective alkene hydrogenation. Catalyst
Preparation. In a nitrogen-filled glovebox, a 40 mL vial was charged
with [Rh(cod)Cl]2 (178 mg, 361 mmol, 0.50 mol % Rh), DTBM-
SEGPHOS-R (942 mg, 799 mmol, 0.55 mol %) and 30 mL MeOAc.
The vial was gently agitated to dissolve the solids, and the mixture was
aged for 10 min to give a dark red solution that was added to a
separate 40 mL vial containing KOTf (163 mg, 866 mmol, 0.6 mol
%). The vial containing the original Rh-ligand solution was rinsed
with 4 mL MeOAc and the rinse was transferred to the vial containing
the Rh-L-KOTf mixture, which was aged for 5 min to give a dark
orange solution.
Step 2: Deoxygenation of the Alcohol Intermediate. To a 40 mL
scintillation vial containing a solution of alcohol intermediate in DCM
(20 mL) were added Et3SiH (2.0 mL, 12 mmol) and BF3·Et2O (1.6
mL, 13 mmol) at rt. The vial was capped and stirred over 24 h, at the
end of which time the reaction was quenched with a saturated
aqueous solution of NaHCO3, extracted with AcOEt and purified by
flash chromatography (hexanes-AcOEt 80:20) to afford 11 as a clear
1
oil. H NMR (400 MHz, CDCl3) δ ppm 7.02−7.09 (m, 2H), 6.85−
6.94 (m, 2H), 4.15 (t, J = 8.0 Hz, 1H), 3.94 (s, 3H), 3.82 (s, 3H),
2.82 (m, 1H), 2.70 (m, 1H), 2.55 (m, 1H), 2.00 (m, 1H).
Hydrogenation. A 17 mL aliquot of the catalyst solution was
added to each of 2 × 350 mL stainless-steel autoclaves each charged
with 20.0 g (72.5 mmol) alkene 14 and 163 mL MeOAc. The
autoclaves were sealed with a lid containing a suspended magnetic stir
bar, then removed from the glovebox and placed in an HEL polyblock
inside a fume hood. Using the HEL WinISO software program,
stirring was begun at 600 rpm and the autoclaves were pressurized
with 1000 psi of H2. The pressurized autoclaves were heated to 40 °C
for 20 h, then allowed to cool to RT and carefully vented. Isolation at
rt: The mother liquor from the first autoclave was removed via pipet
to leave a fine white solid. The solid was transferred to a 60 mL fritted
funnel and washed with heptane (200 mL). The cake was dried under
Step 3: Methyl Cleavage and Dichlorination. To a 40 mL
scintillation vial containing the crude product of step 2 was added 1 M
HCl (10 mL), and the reaction mixture was stirred at 60 °C for 16 h.
The mixture was extracted with EtOAc, and the organic stream was
washed with aq NaHCO3. The organic layer was concentrated by
rotary evaporation to give a crude oil. This material was dissolved in
POCl3 (7 mL) and heated to 100 °C for 7 h. The POCl3 was removed
by rotary evaporation, and MTBE (10 mL) was charged to the
residue. This mixture was poured over ice and aged overnight. The
organic layer was separated, and the aqueous was extracted with
MTBE (10 mL). The combined organic layers were dried over
sodium sulfate, and concentrated by rotary evaporation, at which
I
DOI: 10.1021/acs.joc.8b01734
J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
vacuum for 5 min to give 12.8 g (63% yield) of (S)-8 as a white solid
(99.4 AP, 99.6% ee, 0.15 AP dechlorinated byproduct 15). Prior to
isolation, the second autoclave was cooled to 0 °C for 1.5 h with 450
rpm stirring. The mother liquor was removed via pipet to leave a fine
white solid, which was transferred to a 60 mL fritted funnel, reslurried
in heptane (3 × 20 mL) and then washed with heptane (140 mL).
The cake was dried under vacuum for 5 min to give 14.0 g (69%
yield) of (S)-8 as a white solid (98.9 AP, 99.1% ee, 0.26 AP
allowed to stir for 1 h, and then the aqueous layer was separated and
discarded. The organic phase was washed with brine (2 × 1.5 L). The
THF solvent was swapped for MeCN (1.5 L) by distillation, and
water (1.5 L) was added to induce crystallization. Isolation of the
product was achieved through Buchner filtration, washing with 1:1
̈
MeCN-water (750 mL) and tert-butylmethyl ether (200 mL) at 0 °C.
The product was dried in a vacuum oven overnight to afford (S)-8
(123.0 g, 89% yield) with 99.74% HPLC purity and >99.5% potency.
MP: 250−252 °C. 1H NMR (500 MHz, DMSO-d6) δ: ppm 7.51 (q, J
= 4.7 Hz, 1H), 7.18 (dd, J = 8.6, 5.6 Hz, 2H), 7.11 (t, J = 8.8 Hz, 2H),
4.21 (t, J = 8.0 Hz, 1H), 2.87 (d, J = 4.7 Hz, 3H), 2.78 (td, J = 9.8,
8.8, 4.7 Hz, 1H), 2.58 (dddd, J = 27.3, 18.6, 14.8, 8.8 Hz, 2H), 1.95
(dq, J = 16.1, 6.6 Hz, 1H). 13C NMR (125.75 MHz, DMSO-d6) δ:
171.9, 160.9 (d, J = 242 Hz), 160.6, 158.6, 139.4 (d, J = 3.0 Hz),
129.7 (d, J = 8.0 Hz), 115.6, 115.1 (d, J = 21 Hz), 49.9, 32.1, 27.3,
25.2. IR: 3279, 2941, 2896, 1584, 1508, 1401, 1317, 1268, 1112, 916.
HRMS (ESI-TOF) m/z: [M + H]+ calcd for C14H13ClFN3 278.0855;
found 278.0847.
1
dechlorinated byproduct 15). H NMR (400 MHz, CDCl3) δ: ppm
7.08 (m, 2H), 6.69 (m, 2H), 4.76 (br s, 1H), 4.24 (t, J = 8.0 Hz, 1H),
3.09 (d, J = 4.0 Hz, 3H), 2.60−2.76 (m, 3H), 2.08 (m, 1H). 13C
NMR (125.75 MHz, CDCl3) δ: 172.6, 162.6, 160.9, 159.8, 138.4,
129.2, 115.4, 115.2, 50.7, 32.7, 27.8, 25.0. HRMS (ESI-TOF) m/z:
[M + H]+ calcd for C14H13ClFN3 278.0855; found 278.0847. [α]20
D
+1.12 (c 1.16, DMSO). Chiral HPLC method: ChiralPak OJ-Rh, 5
um, 4.6 × 150 mm, 0.8 mL/min, 254 nm, solvent A: 0.01 M NH4OAc
in MeOH-water (20:80), solvent B: 0.01 M NH4OAc in MeOH-
water-MeCN (20:5:75), 52% to 100% B over 4 min, hold 6 min, RT =
5.366 min (undesired (R)-8), 7.003 min (desired (S)-8).
Preparation of (S)-1. To a 5 L reactor were charged NMP (83
mL, 3 L/kg), chloropyrimidine (S)-8 (27.7 g, 100 mmol), aniline 2
(22.1 g, 108 mmol, 1.08 equiv), and sulfuric acid (7.36 g, 75.0 mmol,
0.75 equiv). The reaction mixture was heated to 100 °C. After 30 min,
a slurry was formed and aging at 90−100 °C was continued for 4 h.
The slurry was cooled to 60 °C and water (125 mL, 4.5 L/kg), brine
(33 mL, 1.5 L/kg), and 2 M NaOH (100 mL, 200 mmol, 2 equiv)
were added. Seed of BMS-932481 (0.1 g, 0.5%) was charged, followed
by THF (13.9 mL, 0.5 L/kg). After stirring for 30 h, the reaction
mixture was cooled to 20 °C. BMT-932481 [(S)-1] was isolated as
off-white crystals (39.2 g, 88% yield) in >98.9% purity and >99% ee.3a
1H NMR (400 MHz, DMSO-d6) δ: ppm 9.20 (br s, 1H), 8.57 (br s,
1H), 8.17 (s, 1H), 7.30 (d, J = 8.0 Hz, 1H), 7.22 (m, 2H), 7.15 (m,
2H), 6.93 (dd, J = 8.0, 4.0 Hz, 1H), 4.16 (br s, 1H), 3.63 (s, 3H), 2.98
(br s, 3H), 2.76 (m, 1H), 2.61 (m, 2H), 2.31 (s, 3H), 1.90 (m, 1H).
13C NMR (125.75 MHz, DMSO-d6) δ: 171.2, 160.3, 160.1, 159.7,
151.9, 145.6, 143.5, 141.0, 130.4, 130.3, 125.1, 118.6, 115.5, 115.3,
110.2, 108.7, 102.2, 55.8, 51.0, 33.2, 27.9, 26.0, 14.1. HRMS (ESI-
TOF) m/z: [M + H]+ calcd for C24H25FN7O 446.2099; found
Preparation of (5S,7S)-2-Chloro-7-(4-fluorophenyl)-4-
(methylamino)-6,7-dihydro-5H-cyclopenta[d]pyrimidin-5-ol
((5S,7S)-13). A 20 L reactor was charged with DMAc (12.3 L, 3.00
L/kg), ketone 4 (410 g, 99.0 wt %, 1.41 mol), RuCl(mesityl)(S,S)-
TsDPEN (4.37 g, 7.03 mmol, 0.005 equiv), and DIPEA (245 mL,
182g, 1.41 mmol, 1 equiv). The reaction mixture was heated to 40 °C
and degassed by sparging with nitrogen for 20 min. Over a period of 2
h, formic acid (129 g, 2.81 mmol, 2 equiv) was added dropwise. After
completion of the addition, the reaction mixture was aged for an
additional 3 h until it showed <1% of starting ketone. At this point,
the end of reaction mixture contained 98.8 AP of 5S,7S)-13 in 99.4%
ee and 146:1 dr. 2-MeTHF (2.05 L, 5 L/kg) was added to the
reaction mixture, which was subsequently cooled to 20 °C. A solution
of citric acid (148 g citric acid, 1.03 L brine, 1.03 L water) was added
over 20 min and the stirred mixture was warmed to 20 °C. The
aqueous layer was removed, and the organic layer was extracted with a
saturated NaHCO3 solution (2.05 L), a half brine solution (2.05 L),
and water (2.05 L) to fully remove the DMAc. 2-MeTHF was
swapped for toluene (5.0 L) under reduced pressure, maintaining the
temperature below 40 °C. After cooling to 20 °C and aging for 3 h,
the product slurry was filtered and the cake was washed with toluene
(2.05 L). Overnight drying in a vacuum oven gave (5S,7S)-13 (346 g,
84% yield) in 99.37% HPLC purity, > 200 dr, 99.7% ee, and >99.5%
446.2084. [α]20 −69.13 (c 1.08, CHCl3). Chiral HPLC method:
D
Phenomenex Lux Cellulose-3, 3 um, 4.6 × 150 mm2, 0.8 mL/min, 220
nm, solvent A: 0.01 M NH4OAc in MeOH-water (20:80), solvent B:
0.01 M NH4OAc in MeOH-water-MeCN (20:5:75), 75% B isocratic,
15 min, RT = 3.17 min (desired S), 4.08 min (undesired R).
1
potency. MP: 181−183 °C. H NMR (500 MHz, CD3CN) δ: 7.31−
7.19 (m, 2H), 7.14−7.02 (m, 2H), 6.19 (br s, 1H), 5.15 (q, J = 4.9
Hz, 1H), 4.06 (t, J = 8.2 Hz, 1H), 3.66 (d, J = 6.1 Hz, 1H), 3.05−3.00
(m, 1H), 2.99 (d, J = 4.9 Hz, 3H), 1.84 (ddd, J = 13.5, 7.8, 5.8 Hz,
1H). 13C NMR (125.75 MHz, CD3CN) δ: 172.4, 162.6 (d, J = 242
Hz), 162.5, 161.7, 139.8 (d, J = 3.1 Hz), 131.1 (d, J = 8.1 Hz), 117.67,
116.1 (d, J = 21.4 Hz), 71.3, 49.6, 44.9, 27.8. IR: 3435, 3257, 3061,
2945, 1619, 1521. 1513, 1388, 1263, 1165, 1108, 1041, 987. 849, 814.
HRMS (ESI-TOF) m/z: [M + H]+ calcd for C14H14ClFN3O
ASSOCIATED CONTENT
■
S
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.joc.8b01734.
NMR spectra for all new compounds, computational
details, chiral HPLC chromatograms, reaction screening
results, and PMI and process greenness comparison of
routes (PDF)
294.0809; found 294.0815. [α]20 +2.00 (c 1.08, DMSO). Chiral
D
HPLC method: Phenomenex Lux Cellulose-1, 3 um, 4.6 × 150 mm, 1
mL/min, 220 nm, solvent A: 0.01 M NH4OAc in MeCN-water
(5:95), solvent B: 0.01 M NH4OAc in MeCN-water (95:5), 25 to
100% B over 30 min, RT = 8.20 min (major trans), 9.22 min (minor
trans), 11.69 min (major cis (5S, 7S)-13), 13.72 min (minor cis),
19.73 min (7), 24.24 min (7).
Preparation of (S)-8 through Deoxygenation. To a 5 L
reactor was charged alcohol (5S,7S)-13 (150 g, 97.0 wt %, 495
mmol), DCM (1.5 L), sulfolane (37.5 mL), and Et3SiH (200 mL,
1.24 mol, 2.5 equiv). The resulting suspension was warmed to 42 °C
and BF3·OEt2 was added (190 mL, 1.49 mol, 3.0 equiv). The reaction
mixture became homogeneous and was allowed to stir for 21 h, at
which point the consumption of starting material was confirmed by
HPLC. The reaction mixture was cooled to 0 °C and
isobutyraldehyde (110 mL, 1.24 mol, 2.5 equiv) was added (exotherm
= 20 °C). After stirring for 30 min, the solvent was swapped from
DCM to THF (3.0 L). To the resulting solution was added 20% w/w
brine (750 mL) and 20% w/w aqueous KHCO3. The phases were
X-ray data for (S)-1 (CIF)
AUTHOR INFORMATION
■
Corresponding Author
*E-mail: Neil.Strotman@bms.com (N.A.S.).
ORCID
Neil A. Strotman: 0000-0002-5350-8735
Eric M. Simmons: 0000-0002-3854-1561
Andrew T. Parsons: 0000-0002-0320-0919
Thomas E. La Cruz: 0000-0002-9745-4580
Notes
The authors declare no competing financial interest.
J
DOI: 10.1021/acs.joc.8b01734
J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
(13) The poor selectivity with 6 prompted us to explore protecting
groups on nitrogen (e.g., Bn, Cbz), but no effective conditions were
identified.
ACKNOWLEDGMENTS
■
We thank Carolyn Wei for helpful discussions during the
preparation of the manuscript. We also acknowledge the
analytical support provided by Michael Peddicord during the
development of this work.
(14) (a) The estimated acidities correspond to pKa1 values
computed using PCM single point calculations (DMSO, ε = 46.826)
on B3LYP/6-31+G(d) optimized geometries: Ding, F.; Smith, J. M.;
Wang, H. First-Principles Calculation of pKa Values for Organic Acids
in Nonaqueous Solution. J. Org. Chem. 2009, 74, 2679−2691. (b)
For a discussion on the relationship between the microscopic pKa
values of polyprotic acids and reactivity, see: Bodner, G. M. Assigning
the pKa’s of Polyprotic Acids. J. Chem. Educ. 1986, 63, 246−247.
(15) All calculations were performed with Gaussian 09. See SI for
details and coordinates.
REFERENCES
■
(1) Alzheimer’s Association. 2017 Alzheimer’s Disease Facts and
Figures. Alzheimer's Dementia 2017, 13, 325−373.
(2) For clinical trial results of BMS-932481, see: Soares, H. D.;
Gasior, M.; Toyn, J. H.; Wang, J.-S.; Hong, Q.; Berisha, F.; Furlong,
M. T.; Raybon, J.; Lentz, K. A.; Sweeney, F.; Zheng, N.; Akinsanya,
B.; Berman, R. M.; Thompson, L. A.; Olson, R. E.; Morrison, J.;
Drexler, D. M.; Macor, J.E.; Albright, C. F.; Ahlijanian, M. K.;
AbuTarif, M. The γ-Secretase Modulator, BMS-932481, Modulates
Aβ Peptides in the Plasma and Cerebrospinal Fluid of Healthy
Volunteers. J. Pharmacol. Exp. Ther. 2016, 358, 138−150.
(3) (a) Boy, K. M.; Guernon, J. M.; Macor, J. E.; Olson, R. E.; Shi, J.;
Thompson, III, L. A.; Wu, Y.-J.; Xu, L.; Zhang, Y.; Zuev, D. S.
Compounds for the Reduction of β-Amyloid Production. Int. Patent
WO 2011/014535 A1. (b) Boy, K. M.; Guernon, J. M.; Macor, J. E.;
Olson, R. E.; Shi, J.; Thompson, III, L. A.; Wu, Y.-J.; Xu, L.; Zhang,
Y.; Zuev, D. S. Compounds for the Reduction of β-Amyloid
Production. Patent US 2012/0028994 A1.
(16) A pKa value of 24 was estimated for more configurationally
stable 11.
(17) See SI for complete details.
(18) (a) Friedfeld, M. R.; Shevlin, M.; Margulieux, G. W.; Campeau,
L.-C.; Chirik, P. J. Cobalt-Catalyzed Enantioselective Hydrogenation
of Minimally Functionalized Alkenes: Isotopic Labeling Provides
Insight into the Origin of Stereoselectivity and Alkene Insertion
Preferences. J. Am. Chem. Soc. 2016, 138, 3314−3324. (b) Cui, X.;
Burgess, K. Chem. Rev. 2005, 105, 3272−3296. (c) Margarita, C.;
Andersson, P. G. Catalytic Homogeneous Asymmetric Hydro-
genations of Largely Unfunctionalized Alkenes. J. Am. Chem. Soc.
2017, 139, 1346−1356.
(19) (a) Zhang, A.; Butt, N. A.; Zhang, W. Asymmetric
Hydrogenation of Nonaromatic Cyclic Substrates. Chem. Rev. 2016,
116, 14769−14827. (b) Woodmansee, D. H.; Pfaltz, A. Asymmetric
Hydrogenation of Alkenes Lacking Coordinating Groups. Chem.
Commun. 2011, 47, 7912−7916. (c) Roseblade, S. J.; Pfaltz, A.
Iridium- Catalyzed Asymmetric Hydrogenation of Olefins. Acc. Chem.
Res. 2007, 40, 1402−1411.
(4) 2,4-Dichloro-6-(methylamino)-5-pyrimidinecarboxalde hyde,
CAS: 1007197−30−9. Hossain, M. S.; Le, C. Q.; Joseph, E.;
Nguyen, T. Q.; Johnson-Winters, K.; Foss, F. W., Jr. Convenient
Synthesis of Deazaflavin Cofactor FO and Its Activity in F420-
dependent NADP Reductase. Org. Biomol. Chem. 2015, 13, 5082−
5085.
(5) Beingessner, R. L.; Deng, B.-L.; Fanwick, P. E.; Fenniri, H. A
Regioselective Approach to Trisubstituted 2 (or 6)-Arylaminopyr-
imidine-5-carbaldehydes and Their Application in the Synthesis of
Structurally and Electronically Unique G∧C Base Precursors. J. Org.
Chem. 2008, 73, 931−939.
(6) Gao, F.; Hoveyda, A. H. α-Selective Ni-Catalyzed Hydro-
alumination of Aryl- and Alkyl-Substituted Terminal Alkynes:
Practical Syntheses of Internal Vinyl Aluminums, Halides, or
Boronates. J. Am. Chem. Soc. 2010, 132, 10961−10963.
(7) For reviews, see: (a) Willis, M. C. Transition Metal Catalyzed
Alkene and Alkyne Hydroacylation. Chem. Rev. 2010, 110, 725−748.
(b) Murphy, S. K.; Dong, V. M. Enantioselective Hydroacylation of
Olefins with Rhodium Catalysts. Chem. Commun. 2014, 50, 13645−
13649.
(20) Menges, F.; Pfaltz, A. Threonine-Derived Phosphinite-Oxazo-
line Ligands for the Ir-Catalyzed Enantioselective Hydrogenation.
Adv. Synth. Catal. 2002, 344, 40−44.
(21) Almena, J.; Monsees, A.; Kadyrov, R.; Riermeier, T. H.; Gotov,
̈
B.; Holz, J.; Borner, A. Highly Enantioselective Hydrogenation of
Itaconic Acid Derivatives with a Chiral Bisphospholane-Rh (I)
Catalyst. Adv. Synth. Catal. 2004, 346, 1263−1266.
(22) nbd = norbornadiene; cod = 1,5-cyclooctadiene.
(23) Dechlorinated byproduct 15 was not isolated, and the structure
was proposed based on LC-MS analysis (ESI-MS) calcd for
C14H15FN3 244.1 [(M + H)+], found 244.1.
(24) (a) Hong, L.; Sun, W.; Yang, D.; Li, G.; Wang, R. Additive
Effects on Asymmetric Catalysis. Chem. Rev. 2016, 116, 4006−4123.
(b) Brown, J. M. Rhodium Asymmetric Hydrogenation Observed
during its Exponential Growth Phase. Organometallics 2014, 33,
5912−5923. (c) Lin, S.-T.; Siegel, S. Stereochemistry and the Use of
H2/D2Mixtures as Probes into the Mechanism of Hydrogenations
Catalyzed by Cationic Rhodium(DIPHOS) Complexes. Kinet. Catal.
2006, 47, 83−92.
(8) Nucleophilic aromatic substitution of related 2-chloropyrimi-
dines was previously demonstrated. See ref 3.
(9) Kundu, K.; McCullagh, J. V.; Morehead, A. T., Jr. Hydro-
acylation of 2-Vinyl Benzaldehyde Systems: An Efficient Method for
the Synthesis of Chiral 3-Substituted Indanones. J. Am. Chem. Soc.
2005, 127, 16042−16043.
́
́
(25) For recent references, see: (a) Leon, F.; Gonzalez-Liste, P. J.;
García-Garrido, S. E.; Arribas, I.; Rubio, M.; Cadierno, V.; Pizzano, A.
Broad Scope Synthesis of Ester Precursors of Nonfunctionalized
Chiral Alcohols Based on the Asymmetric Hydrogenation of α,β-
Dialkyl-, α,β-Diaryl-, and α-Alkyl-β-aryl-vinyl Esters. J. Org. Chem.
2017, 82, 5852−5867. (b) Molinaro, C.; Scott, J. P.; Shevlin, M.;
(10) Tanaka, M.; Imai, M.; Fujio, M.; Sakamoto, E.; Takahashi, M.;
Eto-Kato, Y.; Wu, X. M.; Funakoshi, K.; Sakai, K.; Suemune, H.
Concurrent Induction of Two Chiral Centers from Symmetrical 3,4-
Disubstituted and 3,3,4-Trisubstituted 4-Pentenals Using Rh-Cata-
lyzed Asymmetric Cyclizations. J. Org. Chem. 2000, 65, 5806−5816.
(11) Yang, X.-H.; Dong, V. M. Rhodium-Catalyzed Hydro-
functionalization: Enantioselective Coupling of Indolines and 1,3-
Dienes. J. Am. Chem. Soc. 2017, 139, 1774−1777.
(12) There is literature precedent for the sluggish reactivity of 2-
aminobenzaldehydes in related Rh-catalyzed hydroacylations:
(a) Stemmler, R. T.; Bolm, C. An Unprecedented Rhodium-Catalyzed
Asymmetric Intermolecular Hydroacylation Reaction with Salicylal-
dehydes. Adv. Synth. Catal. 2007, 349, 1185−1198. (b) Arnold, J. S.;
Mwenda, E. T.; Nguyen, H. M. Rhodium-Catalyzed Sequential Allylic
Amination and Olefin Hydroacylation Reactions: Enantioselective
Synthesis of Seven-Membered Nitrogen Heterocycles. Angew. Chem.,
Int. Ed. 2014, 53, 3688−3692.
́
Wise, C.; Menard, A.; Gibb, A.; Junker, E. M.; Lieberman, D.
Catalytic, Asymmetric, and Stereodivergent Synthesis of Non-
Symmetric β,β-Diaryl-α-Amino Acids. J. Am. Chem. Soc. 2015, 137,
999−1006.
(26) See Table S7 in the SI.
(27) In contrast, incomplete conversion and slightly decreased
enantioselectivities were observed at lower pressures (e.g., 450 psi
H2), see Table S5.
(28) When a 0.6 M solution of (S)-1 in NMP was treated with 1
equiv H2SO4 at 100 °C, < 1% (R)-1 formed in 24 h.
(29) (S)-1 racemized under strongly basic conditions and elevated
temperatures (2.5 equiv t-BuONa in n-PrOH at 96 °C, 12 h). DFT
K
DOI: 10.1021/acs.joc.8b01734
J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
calculations that predicted an approximate pKa value of 26 for the
benzylic proton at C7.
the BMS Process Greenness Scorecard. Green Chem. 2017, 19, 5163−
5171.
(30) The absolute configuration of (S)-1 was determined by X-ray
crystallography, see SI.
(31) (a) Pellissier, H. Catalytic Non-Enzymatic Kinetic Resolution.
Adv. Synth. Catal. 2011, 353, 1613−1666. (b) Dehli, J. R.; Gotor, V.
Parallel Kinetic Resolution of Racemic Mixtures: A New Strategy for
the Preparation of Enantiopure Compounds? Chem. Soc. Rev. 2002,
31, 365−370.
(32) (a) Noyori, R.; Tokunaga, M.; Kitamura, M.; Ohkuma, T.
Dynamic Kinetic Resolution in BINAP−Ruthenium(II) Catalyzed
Hydrogenation of 2-Substituted 3-Oxo Carboxylic Esters. Tetrahe-
dron: Asymmetry 1990, 1, 1−4. (b) Steward, K. M.; Gentry, E. C.;
Johnson, J. S. Dynamic Kinetic Resolution of α-Keto Esters via
Asymmetric Transfer Hydrogenation. J. Am. Chem. Soc. 2012, 134,
7329−7332. (c) Echeverria, P. G.; Ayad, T.; Phansavath, P.;
Ratovelomanana-Vidal, V. Recent Developments in Asymmetric
Hydrogenation and Transfer Hydrogenation of Ketones and Imines
through Dynamic Kinetic Resolution. Synthesis 2016, 48, 2523−2539.
̈
(d) Verho, O.; Backvall, J. E. Chemoenzymatic Dynamic Kinetic
Resolution: A Powerful Tool for the Preparation of Enantiomerically
Pure Alcohols and Amines. J. Am. Chem. Soc. 2015, 137, 3996−4009.
(e) Applegate, G. A.; Berkowitz, D. B. Exploiting Enzymatic Dynamic
Reductive Kinetic Resolution (DYRKR) in Stereocontrolled Syn-
thesis. Adv. Synth. Catal. 2015, 357, 1619−1632.
(33) Experimentally determined pKa values for ethyl-3-oxobutanoate
and ethyl acetate: (a) Bordwell, F. G. Equilibrium Acidities in
Dimethyl Sulfoxide Solution. Acc. Chem. Res. 1988, 21, 456−463.
(b) Zhang, X. M.; Bordwell, F. G.; Van Der Puy, M.; Fried, H. E.
Equilibrium Acidities and Homolytic Bond Dissociation Energies of
the Acidic C−H Bonds in N-Substituted Trimethylammonium and
Pyridinium Cations. J. Org. Chem. 1993, 58, 3060−3066.
(34) Claisen, L. Zu den O-Alkylderivaten des Benzoyl-acetons und
den aus ihnen entstehenden Isoxazolen. (Entgegnung an Hrn. O.
Weygand.). Ber. Dtsch. Chem. Ges. B 1926, 59, 144−153.
(35) Noyori, R.; Hashiguchi, S. Asymmetric Transfer Hydrogenation
Catalyzed by Chiral Ruthenium Complexes. Acc. Chem. Res. 1997, 30,
97−102.
(36) (a) Fujii, A.; Hashiguchi, S.; Uematsu, N.; Ikariya, T.; Noyori,
R. Ruthenium(II)-Catalyzed Asymmetric Transfer Hydrogenation of
Ketones Using a Formic Acid−Triethylamine Mixture. J. Am. Chem.
Soc. 1996, 118, 2521−2522. (b) Haack, K.-J.; Hashiguchi, S.; Fujii, A.;
Ikariya, T.; Noyori, R. The Catalyst Precursor, Catalyst, and
Intermediate in the RuII-Promoted Asymmetric Hydrogen Transfer
between Alcohols and Ketones. Angew. Chem., Int. Ed. Engl. 1997, 36,
285−288.
(37) Treatment of alcohol 13 with t-BuOK consistently resulted in
an equilibrium ratio of ∼5:1 cis:trans, regardless of the initial ratio.
DFT calculations indicate that cis-13 is 1.0 kcal/mol more stable than
its trans isomer.
(38) (a) You, T.; Wang, Z.; Chen, J.; Xia, Y. Transfer Hydro-
dehalogenation of Organic Halides Catalyzed by Ruthenium(II)
Complex. J. Org. Chem. 2017, 82, 1340−1346. (b) Paul, B.;
Chakrabarti, K.; Shee, S.; Maji, M.; Mishra, A.; Kundu, S. A Simple
and Efficient In Situ Generated Ruthenium Catalyst for Chemo-
selective Transfer Hydrogenation of Nitroarenes: Kinetic and
Mechanistic Studies and Comparison with Iridium Systems. RSC
Adv. 2016, 6, 100532−100545.
(39) DIPEA was substituted for TEA to prevent displacement of the
chlorine atom by TEA, which was observed under some conditions.
(40) See Table S10 in the SI.
(41) The VDKR route showed significant improvements in step
count, yield, and process mass intensity (PMI) compared to the
enabling discovery route. Moreover, the “Process Greenness Score”
benchmarking metric, which incorporates factors not captured by
PMI alone, shows substantial improvements in efficiency and
sustainability (Table S11). Leahy, D. K.; Simmons, E. M.; Hung,
V.; Sweeney, J. T.; Fleming, W. F.; Miller, M. Design and Evolution of
L
DOI: 10.1021/acs.joc.8b01734
J. Org. Chem. XXXX, XXX, XXX−XXX