Development of a Commercial Process to Prepare AMG 232 Using a Green Ozonolysis-Pinnick Tandem Transformation

Article abstract of DOI:10.1021/acs.joc.8b02390

A robust process to manufacture AMG 232 was developed to deliver drug substance of high purity. Highlights of the commercial process development efforts include the following: (i) use of a novel bench-stable Vilsmeier reagent, methoxymethylene-N,N-dimethyliminium methyl sulfate, for selective in situ activation of a primary alcohol intermediate; (ii) use of a new crystalline and stable isopropyl calcium sulfinate reagent ensuring robust preparation of a sulfone intermediate; (iii) development of a safe ozonolysis process conducted in an aqueous solvent mixture in either batch or continuous manufacturing mode; and (iv) control of the drug substance purity by crystallization of a salt rejecting impurities effectively. The new process was demonstrated to afford the drug substance (99.9 LC area %) in 49.8% overall yield from starting material DLAC (1).

Full text of DOI:10.1021/acs.joc.8b02390

Subscriber access provided by University of Winnipeg Library
Article
Development of a Commercial Process to Prepare AMG 232
Using A Green Ozonolysis-Pinnick Tandem Transformation
Brian M. Cochran, Michael T. Corbett, Tiffany L. Correll, Yuan-Qing Fang, Tawnya G.
Flick, Sian C. Jones, Maria V. Silva Elipe, Austin Gerald Smith, John L. Tucker, Filisaty
Vounatsos, Greg Wells, David Yeung, Shawn D. Walker, Matthew Michael Bio, and Seb Caille
J. Org. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.joc.8b02390 • Publication Date (Web): 17 Dec 2018
Downloaded from http://pubs.acs.org on December 19, 2018
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 37
The Journal of Organic Chemistry
1
2
3
4
5
6
Development of a Commercial Process to Prepare AMG 232 Using A
Green Ozonolysis-Pinnick Tandem Transformation
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Brian M. Cochran, Michael T. Corbett, Tiffany L. Correll, Yuan-Qing Fang, Tawnya G. Flick, Siân C. Jones,
Maria V. Silva Elipe, Austin G. Smith, John L. Tucker, Filisaty Vounatsos, Greg Wells,Δ David Yeung, Shawn
D. Walker, Matthew M. Bio, and Seb Caille*
Pivotal Drug Substance Technologies, Amgen Inc., One Amgen Center Drive, Thousand Oaks, California,
91320
Abstract
A robust process to manufacture AMG 232 was developed to deliver drug substance of high
purity. Highlights of the commercial process development efforts include: i. use of a novel bench-stable
Vilsmeier reagent, methoxymethylene-N,N-dimethyliminium methyl sulfate, for selective in situ
activation of a primary alcohol intermediate; ii. use of a new crystalline and stable isopropyl calcium
sulfinate reagent ensuring robust preparation of a sulfone intermediate; iii. development of a safe
ozonolysis process conducted in an aqueous solvent mixture in either batch or continuous
manufacturing mode; iv. control of the drug substance purity by crystallization of a salt rejecting
impurities effectively. The new process was demonstrated to afford the drug substance (99.9 LC area%)
in 49.8% overall yield from starting material DLAC (1).
Me
O2S
Me
i. L-Valinol
Me
O2S
Me
ii.
O
Me
Me
N
OMe
O
O
Me
O
Me
Me
Me
Ozonolysis-Pinnick
Tandem Reaction
Me
OSO3Me
N
N
CO2H
Me
Me
iii. Ms2O
Cl
Continuous or Semi-Batch
Manufacturing
iv.
O
S
O
S
Cl
Cl
Me
Me
Me
Cl
Ca
O
Cl
Air/Aqueous CH3CN
20 kg scale
O
Cl
DLAC (1)
Me
SUL (4)
AMG 232
80%
4 steps
64%
99.9 LC area%
Introduction
1
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 2 of 37
1
2
3
4
5
6
7
8
The goal of commercial process development is to elaborate a robust process capable of
delivering steady supply of drug substance meeting defined quality attributes at the lowest possible cost
per kilo. Achieving these requirements not only enables stable commercial supplies but fulfills
regulatory expectations for new drug applications. This article details the development of a process to
manufacture AMG 232,1 an inhibitor of the p53-MDM2 protein-protein interaction2 and a clinical
candidate in multiple oncology indications. It starts with the presentation of the first-in-human (FIH)
synthetic process and outlines the improvements brought forward to meet the requirements for a
dependable commercial manufacturing process. These improvements include the development of new
chemical transformations and a reliable isolation strategy.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The gram-scale synthesis of this class of inhibitors has previously been reported by our discovery
group in 2012.3 Based on this work, a few kilograms of drug substance were prepared by the synthetic
process illustrated in Scheme 1. Intermediate OXOS (3) was used as regulatory starting material for the
investigational new drug application. The development of a process to synthesize DLAC (1), setting the
three stereogenic centers of this molecule using an enantioselective dynamic kinetic resolution and
highly diastereoselective alkylation, will be the topic of a separate report.4 Ring-opening of DLAC (1)
using excess L-valinol (3 equivalents) at elevated temperature afforded amide ABA (2), which was
extracted in dichloromethane. Reduction of the equivalents of L-valinol used in this transformation was
identified as a development objective going forward considering the high cost of this raw material. Of
note is the fact that the preparation of analogues ABA1 (6) from the corresponding analogues DLAC1
(5), bearing side-chains with a functional group of the same oxidation state as that of AMG 232, was not
successful due to the formation of undesired succinimides SUC1 (7) (Scheme 2). Considering the similar
rates observed for the formation of desired products ABA1 (6) and their transformation to side-products
SUC1 (7), it was not possible to isolate amides ABA1 (6) in productive yields.
Scheme 1. First-in-human process to manufacture AMG 232
2
ACS Paragon Plus Environment

Page 3 of 37
The Journal of Organic Chemistry
1
2
3
4
O3S
5
6
7
8
HO
O
O
Me
Me
O
Me
Me
Me
Me
i. p-Toluenesulfonic Anhydride
Thionyl chloride/2,6-Lutidine
1,2-Dichloroethane/120 C
Me
N
O
N
i. L-Valinol/110 C
H
Me
HO
H
ii. Aqueous HCl/CH2Cl2
ii. Na Naphthyl Sulfonate
9
Cl
Cl
Cl
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1/2
Cl
Cl
Cl
OXOS (3)
DLAC (1)
ABA (2)
70% yield
90% yield
O
S
i.
ii. Acetonitrile/H2O
Me
Me
OH
Isopropylsulfinic acid
tBuONa/THF/Toluene/120 C
Me
O2S
Me
Me
O2S
Me
Me
O2S
Me
O
O
i. RuCl3/NaIO4
EtOAc/H2O/Acetonitrile
Me
Me
O
Me
Me
Me
Me
i. MeOH/H2O
N
CO2H
N
CO2H
EtOH
N
Me
Me
Me
iii. Ethanol/Heptane
Cl
Cl
Cl
Cl
Cl
Cl
AMG 232
90% yield
AMG 232
SUL (4)
Ethanol Solvate
80% yield
70% yield
Scheme 2. Attempted preparation of analogues of ABA (2) (k1 ≈ k2)
HO
OH
O
Me
N
Me
OH
Me
O
H
Y
Me
Me
Me
Me
O
X
NH2
X
N
H
HO
Me
OH
O
Me
Cl
Heat
k2
Heat
k1
Cl
Cl
DLAC1 (5)
= COOtBu, CONiPr2,
CO2H, CN
Cl
Cl
Cl
ABA1 (6)
= COOtBu, CONiPr2,
CO2H, CN
X
X
SUC1 (7)
= O or NH
Y
Oxoiminium salt OXOS (3) was prepared from ABA (2) by double activation with two equivalents
of toluenesulfonic anhydride (Ts2O) and 2,6-lutidine at elevated temperature. The cation thus prepared
was isolated as a 1-naphthylsulfonic acid salt, which offered satisfactory impurity rejection properties.
One attractive option considered for market application was to isolate crystalline intermediate DHO (8)
(Scheme 3) to increase control over the rejection of impurities for the manufacturing sequence.
However, this would necessitate the use of a reagent enabling chemoselective activation of the primary
alcohol of ABA (2) while leaving the secondary benzylic alcohol untouched. Sulfonic anhydrides
reagents do not offer this advantage. Identifying an alternative reagent to Ts2O would also avoid
3
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 4 of 37
1
2
3
4
5
6
formation of the potentially mutagenic impurity DHO-OTs (9), which undergoes slow conversion to
OXOS (3) at elevated temperature (120 °C).
7
8
Scheme 3. Intermediates in the Preparation of OXOS (3) from ABA (2) using Ts2O
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
HO
O
Me
O
Me
O
H
Me
Me
Me
Me
Me
Me
N
TsO
N
Me
H
N
2 equiv. Ts2O
120 C
H
HO
2 equiv. 2,6-lutidine
120 C/1 hrs
10 hrs
Cl
Cl
Cl
Cl
DHO-OTs (9)/PMI
Cl
via:
O
Cl
ABA (2)
OXOS (3)
Me
Me
N
HO
Me
H
Cl
Cl
DHO (8)
The preparation of SUL (4) from OXOS (3) was carried out by treatment with isopropylsulfinic
acid in the presence of sodium t-butoxide. This transformation proceeds via reversible formation of a
diastereomeric pair of sulfinate intermediates (SULFI, 10) and rearrangement to the thermodynamic
product SUL (4) (Scheme 4). Side-product ALC (11) is formed under these conditions by reaction with
residual water. Isopropylsulfinic acid, an oil at 20 °C, is prepared from isopropylmagnesium chloride and
isolated after an aqueous work-up. Unfortunately, azeotropic drying of this reagent is an unreliable
process as isopropylsulfinic acid decomposes via disproportination5 upon drying. Consequently, the
discovery of a crystalline salt of isopropylsulfinic acid stable under drying conditions and which may be
designated as a commercial regulatory starting material was sought. Alternatively, a process for in situ
preparation of a sulfinic acid salt from isopropyl magnesium chloride without an aqueous work-up and
further reaction with OXOS (3) was considered.
Scheme 4. Intermediates and Side-Product Generated in the Preparation of SUL (4) from OXOS (3)
4
ACS Paragon Plus Environment

Page 5 of 37
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
Me
Me
O2S
Me
N
O
Me
Me
Me
S
O
Me
Me
Me
Me
Me
Me
Me
O
SO2Na
O
SO2Na
Me
N
Me
O
Me
Me
N
Cl
Cl
Cl
Cl
Cl
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
OXOS (3)
Cl
SUL (4)
SULFI (10)
(2 diastereomers)
H2O
HO
O
Me
Me
Me
N
Cl
Cl
ALC (11)
Oxidation of the alkene group of SUL (4) was carried out via treatment with catalytic ruthenium
chloride (2 mol%) and excess sodium periodate (5 equiv.). The crude product was isolated as a
crystalline ethanol solvate. Several features of this step were not desirable. First, the heavy metal used
in this chemical step had to be scavenged, which was carried out using a DARCO-G resin for the first-in-
human delivery. Additionally, multiple equivalents of sodium periodate were necessary to conduct this
process and had to be charged to the reaction vessel in portions to minimize impurity formation. A
complex downstream treatment was necessary to remove the large amounts of salts used. Several
dimeric impurities were generated in this last chemical transformation and made control of drug
substance purity challenging. An AMG 232 Ethanol Solvate was used as a crystalline control point,
however it was problematic. It was only moderately effective at rejecting the impurities present in the
mixture and its crystallization had to be operated as an evaporative process due to the low ethanol
composition (5% v/v) necessary to ensure limited mother liquor losses upon filtration. The use of
ethanol in the crystallization reduced process robustness due to the formation of the corresponding
ethyl ester at temperatures above 30 C. The ester impurity was poorly rejected in the crystallization of
the ethanol solvate. The development of a more dependable and greener oxidation method to prepare
AMG 232 from SUL (4) as well as the elaboration of a robust isolation strategy for the drug substance
allowing effective control of critical quality attributes were sought as part of commercial process
development.
Table 1. Process Changes from the FIH Process to the Commercial Process to Prepare AMG 232
5
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 6 of 37
1
2
3
4
5
CP1 Process Step
FIH Process Issue
Commercial Process Solution
ABA (2) from DLAC (1)
Three equivalents of L-Valinol were used
Reduction of L-Valinol equivalents to two
6
7
8
9
OXOS (3) from ABA (2)
DHO (8) was not isolated
Replacement of Ts2O with chemoselective
reagent enabling isolation of DHO (8)
Replacement of Ts2O with a reagent
OXOS (3) from ABA (2)
PMI and long-lasting intermediate DHO-OTs
(9) was formed
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
enabling the formation of an intermediate
undergoing rapid conversion to OXOS (3)
Discovery of a crystalline salt of
Isopropylsulfinic acid stable under drying
conditions
SUL (4) from OXOS (3)
Isopropylsulfinic acid is a liquid at 20 °C and
was unstable to azeotropic drying
conditions
AMG 232 from SUL (4)
AMG 232 from SUL (4)
Catalyst ruthenium was used in last step
Excess (5 equiv.) sodium periodate was
used in last step
Change of reagents for the last step
Change of reagents for the last step
Isolation of AMG 232
Ethanol Solvate
The isolation was moderately effective at
impurity rejection and not well suited for
crystallization design
Discovery of a salt of AMG 232 well suited
to crystallization design and having
superior impurity rejection properties
Develop a new crystallization of the drug
substance free of impurity formation
Isolation of AMG 232
The isolation was poorly effective at
rejecting the corresponding methyl ester
which can be formed in the crystallization
solvent system (MeOH/H2O)
Results and Discussion
6
ACS Paragon Plus Environment

Page 7 of 37
The Journal of Organic Chemistry
1
2
3
4
a. Development of a Commercial Process to Prepare Intermediate OXOS (3)
5
6
7
8
The thermal amidation of DLAC (1) to ABA (2) with L-valinol proceeds through a multistep
mechanism via intermediate ester EST (12) (Scheme 5). The initial transesterification of DLAC (1) to EST
(12) was determined to be a reversible process (k1 > k-1) leading to a build-up of EST (12) prior to
rearrangement to the amide product ABA (2). Upon performing the reaction at 60 °C, rapid conversion
of DLAC (1) to EST (12) is observed at the start of the reaction followed by slow conversion of EST (12) to
ABA (2) over the course of several days (k1 > k2) (Figure 1). At elevated temperatures (115 °C), the
rearrangement of EST (12) to the more stable ABA (2) is faster, resulting in an increase in the overall
rate of the reaction.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 5. Kinetics of Thermal Amidation of DLAC (1) to ABA (2)
i-Pr
HO
i-Pr
H2N
O
O
L-Valinol
k1
O
Me
Me
Me
k2
O
N
O
slow
H
H
fast
HO
HO
H
k-1
Cl
Cl
Cl
Cl
Cl
Cl
DLAC (1)
EST (12)
ABA (2)
Figure 1. Conversion Rates of DLAC (1) to ABA (2) at 60 °C and 115 °C with 2 equiv. of L-Valinol
120
100
80
60
40
20
0
0
20
40
60
80
Time (hr)
DLAC/115 °C
DLAC/60 °C
EST/115 °C
EST/60 °C
ABA/115 °C
ABA/60 °C
The first-in-human process utilized a thermal melt with 3 equivalents of L-valinol to ensure rapid
conversion of DLAC (1) to ABA (2) at 110 °C. Consistent with our mechanistic understanding of this
transformation, decreasing L-valinol loading leads to a decrease in the overall rate of the reaction by
7
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 8 of 37
1
2
3
4
5
6
7
8
9
lowering the relative concentration of EST (12). The loading of L-valinol can be reduced to 2
equivalents,6 although the reaction takes 36-48 hours to reach conversion at 115 °C. This longer
processing time is nonetheless justifiable based on cost reductions.7 After an aqueous work-up, ABA (2)
is isolated as a solution in toluene (>90% assay yield). The solution also contains DHO (8) (<3 LC area%),
starting material DLAC (1) (~1 LC area%), and EST (12) (~1 LC area%) (Scheme 6). Direct thermolysis of
ABA (2) to prepare DHO (8) led to complex mixtures of products.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 6. Preparation of ABA (2) from DLAC (1)
i-Pr
HO
i-Pr
H2N
O
O
H
O
O
Me
Me
Me
Me
Me
Me
i. L-valinol (2 equiv.)
Toluene (1V)
N
N
O
O
H
HO
H
HO
H
115 °C/72 hrs
HO
~1 LC area%
DLAC (1)
ii. Aqueous HCl
Toluene (4V)
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
ABA (2)
>90% assay yield
EST (12)
~1 LC area%
DHO (8)
<3 LC area%
DLAC (1)
The isolation of intermediate DHO (8) would provide additional opportunity to remove
impurities from the process stream and strengthen the overall control strategy to deliver the drug
substance for market application. Paramount to this strategy was the identification of conditions that
would untether the dehydrative double-cyclization of ABA (2) to OXOS (3) into two distinct mono-
cyclization reactions through the development of a chemoselective activation of the primary alcohol of
ABA (2) (conditions A) that would enable the isolation of DHO (8) (Scheme 7).
Scheme 7. Sequential Dehydration of ABA (2) to OXOS (3)
first-in-human process
HO
O
Me
Me
O
Me
O
H
Me
Me
Me
Me
Me
Me
N
N
N
HO
H
conditions A
conditions B
H
HO
Cl
Cl
Cl
Cl
Cl
Cl
DHO (8)
Crystalline Isolated
OXOS (3)
Crystalline Isolated
ABA (2)
Sulfonyl chloride and sulfonic anhydride reagents were found to be unselective in discriminating
between the primary and secondary alcohols of ABA (2). Furthermore, the use of acid catalysts afforded
complex mixtures of products. However, a novel Vilsmeier salt reagent,8 methoxymethylene-N,N-
8
ACS Paragon Plus Environment

Page 9 of 37
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
dimethyliminium methyl sulfate, successfully provided the desired outcome (Scheme 8). This salt is
easily prepared with no special precautions to exclude moisture and can be stored at 20 °C for several
months with no erosion in titer.9 Additionally, it possesses milder reactivity and improved
chemoselectivity upon comparison with the common halide-derived Vilsmeier salt chloromethylene-
N,N-dimethyliminium chloride and does not cause the formation of alkyl halide side-products.10
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 8. Mono-Cyclization of ABA (2) to DHO (8)
OSO3Me
Me
HO
Me
Me
Me
Me
O
O
Me
Me
N
N
OMe
N
Me
H
H
HO
HO
H
(1.5 equiv.)
NaOAc (1.5 equiv.)
Cl
Cl
Toluene (5V)
30 °C
Cl
Cl
ABA (2)
DHO (8)
>91% assay yield
The desired chemoselectivity of this transformation is possible through to the unique ability of
methoxymethylene-N,N-dimethyliminium methyl sulfate to enable reversible activation of alcohols via
the generation of labile imidate intermediates. This reversibility was investigated in the activation of 4-
chlorobenzyl alcohol (CHA, 13) with the deuterated reagent DEU (14) to generate imidate IMI (15),
which was found to reach equilibration at a 2.5/1 ratio of CHA (13)/IMI (15) (Scheme 9).8 Any imidate
formed by derivatization of the secondary alcohol group of ABA (2) group does not undergo further
cyclization to OXOS (3) at the operating temperature (30 °C). Residual quantities of DHO imidate
intermediate are hydrolyzed as part of the aqueous work-up.
Scheme 9. Reversible Activation of Alcohols with Vilsmeier Reagents
D
D3C
N
OMe
CD3
MeO3SO
CHA IMI
Time
/
D
DEU (14)
HO
D3C
30 min
60 min
18 h
5.4/1.0
4.0/1.0
2.5/1.0
N
O
MeOH
Cl
CD3
DMF-d7
23°C
MeO3SO
Cl
CHA (13)
IMI (15)
Equilibrium solubility measurement were gathered for DHO (8) in various solvents and all values
obtained were above 20 mg/mL at 20 °C (including heptane) except for water (<0.1 mg/mL), which was
thus selected as anti-solvent. Acetonitrile was selected as solvent for the crystallization as it offered
9
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 10 of 37
1
2
3
4
5
6
good impurity rejection in combination with water. A curve showing solubility values at different time
points in the crystallization process is presented in Figure 2. Using this protocol, crystalline DHO (8) is
isolated in >87% yield and >98 LC area% from DLAC (1) (Scheme 10).
7
8
9
Figure 2. Solubility of DHO (8) during the Crystallization Process at 25 °C
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 10. Commercial Process to Prepare DHO (8) from DLAC (1)
HO
Me
Me
Me
Me
O
i. NaOAc (1.2 equiv.)
Vilsmeier Salt (1.5 equiv.)
Toluene (5V)
O
O
Me
i. L-Valinol (2.0 equiv.)
Toluene (1V)
115 °C/72 h
Me
Me
N
O
N
H
HO
30 °C, 2 h
H
HO
H
ii. Toluene (4V)
iii. 1N HCl wash
iv. Azeotropic drying
ii. Aqueous NH4Cl
iii. ACN solvent exchange
iv. Cyrstallization (ACN/H2O)
Cl
Cl
Cl
Cl
Cl
Cl
DLAC (1)
ABA (2)
>90% assay yield
DHO (8)
DLAC (1)
>87% yield from
>98 LC area%
Having untethered the double dehydrative cyclization of ABA (2) to OXOS (3), development of a
method to convert DHO (8) to OXOS (3) was necessary. Methanesulfonic anhydride (Ms2O) was found
to promote a faster conversion to OXOS (3) compared to Ts2O, as potentially mutagenic mesylate
intermediate DHO-OMs (16) is completely consumed at 75 °C in 10 hours (Scheme 11). This is likely due
to the reduced steric hindrance experienced in the transition state leading from DHO-OMs (16) to OXOS
(3) compared to that involved in the cyclization of DHO-OTs (9) to OXOS (3). The transformation was
found to proceed well with 2,6-lutidine as a base in toluene.
Scheme 11. Comparison of Required Temperatures for Conversion of Alkyl Sulfonates to OXOS (3)
10
ACS Paragon Plus Environment

Page 11 of 37
The Journal of Organic Chemistry
1
2
3
4
O
Me
Me
5
6
N
MsO
Me
H
75 C
10 hrs
7
8
Cl
O
Me
Me
Me
9
Cl
N
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
DHO-OMs (16)/PMI
Cl
O
Me
Me
Cl
N
OXOS (3)
TsO
Me
H
120 C
10 hrs
Cl
Cl
DHO-OTs (9)/PMI
The mesylate salt of OXOS (3) generated during the transformation is poorly soluble in toluene
and forms a separate liquid layer as the reaction progresses. To enable further processing, it is
necessary to dilute the reaction mixture with dichloromethane prior to removal of mesylate salts using
aqueous sulfuric acid washes. Salt metathesis with aqueous sodium 1-naphthalenesulfonate is followed
by distillation of dichloromethane, leading to the crystallization of the 1-naphthalenesulfonate toluene
hemi-solvate OXOS (3) salt in >90% yield, >99.5 LC area%, and >99.5 wt% from DHO (8) (Scheme 12).
Scheme 12. Commercial Process to Prepare OXOS (3) from DHO (8)
O3S
Me
Me
O
i. Ms2O (1.2 equiv.)
Me
2,6-Lutidine (2.0 equiv.)
Toluene (5V)/75°C/10 h
O
Me
Me
N
Me
H
HO
N
Me
ii. Dichloromethane (8V)
iii. Aqueous H2SO4
iv. Aqueous Sodium
1-Naphthylsulfonate (1.5 equiv.)
v. Distillation
Cl
Cl
Cl
1/2
Cl
DHO (8)
OXOS (3)
vi. Filtration
>90% yield
>99.5 LC area%
>99.5 wt%
b. Development of a Commercial Process to Prepare Penultimate Intermediate SUL (4)
11
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 12 of 37
1
2
3
4
5
6
7
8
9
The treatment of OXOS (3) with an isopropylsulfinate salt at elevated temperature leads to the
formation of the diastereomeric pair of sulfinate esters SULFI (10) (Scheme 4) which rearrange to the
more thermodynamically stable product SUL (4). This type of sulfinate-sulfone rearrangement has been
described to occur via an ion pair formation and recombination mechanism for benzhydryl sulfinate
esters.11 However, in the case at hand, the results of a cross-over experiment, shown in Scheme 13,
revealed that the rearrangement involves dissociated ions. Considering that OXOS (3) reacts
quantitatively with water at temperatures above 70 °C, there is ample opportunity for alcohol ALC (11)
to be generated. Unless OXOS (3) can be regenerated from ALC (11), water must be rigorously excluded
from the reaction mixture. One avenue to achieve this objective is to prepare a salt of isopropylsulfinic
acid which is stable to azeotropic drying conditions, can be isolated in high purity, and efficiently reacts
with OXOS (3) to generate SUL (4).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 13. Intermediates and Side-Product Generated in the Preparation of SUL (4) from OXOS (3)
Me
Me
O2S
Me
N
Me
O2S
O
Me
Me
Me
S
O
S
O
O
O
O
Me
Me
Et
OMgCl
Me
Me
Me
(4 equiv.)
N
N
Me
Me
DMAc (3V)
120 C
12 hrs
Cl
Cl
Cl
Cl
Cl
Cl
Ethyl-SUL (17)
>80% yield
SUL (4)
none observed
SULFI (10)
(2 diastereomers)
After the evaluation of several isopropylsufinic acid salt candidates to achieve that goal,
including lithium, sodium, potassium, magnesium and ammonium salts, a calcium sulfinate dihydrate
salt stood out as a stable and crystalline species. This salt is prepared from reaction of isopropyl
magnesium chloride with sulfur dioxide (Scheme 14) affording isopropylsulfinic acid after an aqueous
hydrochloric acid quench.12 This material is treated with calcium acetate, thus prompting reactive
crystallization of calcium isopropylsulfinate dihydrate (CALID, 18).13 Of course, a hydrate is not the ideal
species to use to prepare SUL (4) considering the water sensitivity of the process described above.
However, no alternative solvate could be identified and the calcium salt is chemically stable (no
disproportionation product observed in 48 hours) upon azeotropic drying in toluene at elevated
temperature (up to 110 °C). Thus, drying of the reagent suspension can be used as part of the process
prior to addition of OXOS (3) and preparation of SUL (4). X-Ray powder diffraction analyses of oven-
dehydrated samples show that CALID (18) undergoes a polymorphic change upon complete drying (<100
12
ACS Paragon Plus Environment

Page 13 of 37
The Journal of Organic Chemistry
1
2
3
4
5
6
ppm water) at <15% RH. However, it is stable as a dehydrated material and converts back to the original
hydrate form upon water re-adsorption at >20% RH.
Scheme 14. Preparation of Calcium Isopropylsulfinate Dihydrate (CALID, 18)
7
8
9
O
S
O
S
Ca(OAc)2
EtOH/H2O
Me
Me
Me
Me
SO2
THF
Ca
HCl
Me
Me
Me
Me
Me
Me
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
O
O
SO2H
MgCl
SO2MgCl
(2x H2O)
Step 2
Step 1
>80% yield
CALID (18)
>75% yield
The process to manufacture SUL (4) includes drying of CALID (18) and OXOS (3) toluene
suspensions by azeotropic distillation under reduced pressure to prepare mixtures containing <100 ppm
of water. The dry suspensions are combined and a solvent exchange to dimethylacetamide is performed
(Scheme 15). The resultant solution is heated to 120 °C and agitated for up to 20 hours. During this
time, the sulfinate esters, generated within the first hour at 120 °C, rearrange to SUL (4). The typical
level of impurity ALC (11) formed under these conditions is 3 LC area%. SUL (4) is isolated in >80% yield
and >99.5 LC area% after an aqueous work-up and crystallization from acetonitrile and water.14
Scheme 15. Preparation of SUL (4) using CALID (18)
O3S
O
Me
Me
Toluene
N
Me
Azeotropic Drying
Me
O2S
Me
N
60-80 C
HO
Cl
<100 ppm water
O
O
i. Exchange to DMAc
ii. 120 C
Me
Me
Me
Me
Me
Cl
N
Me
OXOS (3)
iii. Aqueous Work-Up
iv. ACN/water crys.
Toluene
O
O
Me
Me
Me
Me
Cl
Cl
Azeotropic Drying
S
Ca
S
Cl
Cl
O
O
70-95 C
SUL (4)
(2x H2O)
CALID (18)
<100 ppm water
ALC (11)
<0.5 LC area%
in isolated solids
>99.5% LC area%
>80% yield
In situ prepared isopropyl magnesium chloride is a reagent that can be employed directly to
prepare SUL (4) and thus offers an alternative to the use of CALID (18). A comparison was made
between these two methods to manufacture SUL (4) with the aim of selecting the most robust process
for commercial application. As shown in Scheme 14, a tetrahydrofuran solution of isopropyl magnesium
chloride can be treated with sulfur dioxide to prepare isopropylsulfinate magnesium chloride. Solvent
exchange to N-methylpyrrolidinone (NMP) was followed by addition of OXOS (3) to afford SUL (4) at 120
13
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 14 of 37
1
2
3
4
5
6
7
8
9
°C or 180 °C (Scheme 16). Elimination of moisture or magnesium hydroxide from the reaction mixture
thus prepared is challenging. For instance, using three equivalents of isopropylsulfinate magnesium
chloride relative to OXOS (3), approximately 2.5 mol% of magnesium hydroxide are present in the
starting Grignard solution.15 NMP and OXOS (3) incorporate 5-10 mol% of water. It is thus difficult to
avoid the formation of a minimum of 15 mol% of ALC (11) using this reagent. Interestingly, the level of
ALC (11) observed during the formation of SUL (4) exceeds the measured amount of water and
hydroxides contained in the reaction mixture, particularly while operating at 120 °C (Figure 3). A second
mechanism, in addition to the direct opening of OXOS (3) with water or a hydroxide salt, must be
invoked to account for the formation of ALC (11). The postulated mechanism involves the reaction of
SULFI (10) with isopropylsulfinate magnesium chloride, involving a S-O bond breakage to afford ALC (11)
(Scheme 17).16
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
There is an operative mechanism converting ALC (11) to SUL (4) at 180 °C that does not occur at
a productive rate at 120 °C (Figure 3). Treatment of OXOS (3) with isopropylsulfinate magnesium
chloride (3 equivalents) in NMP at 180 °C allows the formation of SUL (4) and ALC (11) in 77% and 18%
assay yields, respectively, after 6 minutes. After 80 minutes, 90% assay yield of SUL (4) and 5% assay
yield of ALC (11) are formed (Figure 3). Unfortunately, isopropylsulfinate magnesium chloride has
displayed poor stability at elevated temperatures (85% degradation over 1 hour at 200 °C). Additionally,
the elevated temperatures required to convert ALC (11) to SUL (4) under these conditions represent a
significant hurdle for industrial application.
Scheme 16. Preparation of SUL (4) from OXOS (3) using Isopropylsulfinate Magnesium Chloride
14
ACS Paragon Plus Environment

Page 15 of 37
The Journal of Organic Chemistry
1
2
3
4
Me
MgCl
Me
5
6
SO2
THF
7
8
9
Me
Verify Absence of SO2 (FTIR)
and Alk Grignard (Phenanthroline)
SO2MgCl
Me
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Exchange
NMP
Me
O2S
Me
N
Me
Me
HO
O3S
SO2MgCl
3 equiv.
5V NMP
O
O
Me
Me
Me
Me
Me
O
Me
Me
N
Me
N
Me
120 C or 180 C
Cl
Cl
Cl
Cl
Cl
Cl
SUL (4)
45-90% yield
ALC (11)
OXOS (3)
5-45% yield
Figure 3. Preparation of SUL (4) using Mg Sulfinate at 120 °Ca or 180 °Cb
120
100
80
60
40
20
0
0
50
100
150
200
250
300
Time (minutes)
OXOS/180 °C
OXOS/120 °C
SUL/180 °C
SUL/120 °C
ALC/180 °C
ALC/120 °C
a14 mol% of water or hydroxide (vs OXOS) in rxn mixture
b11 mol% of water or hydroxide (vs OXOS) in the rxn mixture
Scheme 17. Proposed Mechanisms for Generation of ALC and its Conversion to SUL (4)
15
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 16 of 37
1
2
3
4
5
6
7
8
Me
O
Me
Me
Me
S
O
ClMgO
Me
Me
Me
O
O
S
OMgCl
O
Me
O O
Me
Me
Me
Me
Me
N
N
S
S
Me
120 C
O
9
SSO (20)
Cl
Cl
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Cl
Cl
ALC-MgCl (19)
SULFI (10)
(2 diastereomers)
MgCl2
180 C
Me
O2S
Me
N
O
O
Me
Me
Me
Me
120-180 C
Me
S
OMgCl
O
N
Me
Me
MgCl2
MgO
MgClO
Me
Cl
Cl
Cl
OXOS (3)
Cl
SUL (4)
Use of alternative isopropylsulfinate salts to modify the reactivity and increase the stability of
isopropylsulfinate magnesium chloride was evaluated. Treatment of one equivalent of the in situ
prepared reagent with one equivalent of zinc chloride afforded promising results. A commercially
available zinc chloride tetrahydrofuran solution (0.5 M) was added to the solution of isopropylsulfinate
magnesium chloride prepared as previously described (Scheme 14).17 The new species formed was
1
observed to be distinct from isopropylsulfinate magnesium chloride and zinc diisopropylsulfinate by H
NMR.18 Based on this data, the active species is postulated to be isopropylsulfinate zinc chloride. A
solvent exchange to NMP was performed and OXOS (3) was added to the reaction mixture (Scheme 18).
Using this mixed reagent, the mechanism enabling the conversion of ALC (11) to SUL (4) is operative at
120 °C, thus the transformation can be performed at that temperature while avoiding rigorous
moisture-free conditions (Figure 4). Moreover, there is no evidence of an alternative mechanism
affording ALC (11) without the involvement of water with this mixed magnesium-zinc reagent, as all the
ALC (11) generated during the process can be accounted for from the water and hydroxides incoming in
reagents and solvents. At a temperature of 140 °C a good reaction rate was achieved while using
limited equivalents (1.5 equivalents) of the magnesium-zinc species and affording a 90% yield of SUL (4)
with 7% ALC (11) in 7 hours. The magnesium-zinc mixed species was shown to be stable at 150 °C for 16
1
hrs by H NMR experiments. Despite the advantages detailed for use of the latter reagent, a narrow
operating range was uncovered with respect to sulfur dioxide equivalents and CALDI (18) was thus
selected to prepare SUL (4) for commercial manufacture of AMG 232.19
16
ACS Paragon Plus Environment

Page 17 of 37
The Journal of Organic Chemistry
1
2
3
4
5
6
7
Scheme 18. Preparation of SUL (4) from OXOS (3) using the Isopropylsulfinate Magnesium Chloride-Zinc
Chloride Reagent
8
9
Me
MgCl
Me
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
SO2
THF
Me
Verify Absence of SO2
and Alkyl Grignard
SO2MgCl
Me
i. ZnCl2/THF
ii. Exchange
NMP
Me
SO2ZnCl
Me
Me
O2S
Me
N
3.0 equiv.
O3S
HO
MgCl2
O
O
Me
Me
Me
Me
3.0 equiv.
5V NMP
O
Me
Me
Me
Me
N
N
Me
120 C or 180 C
Cl
Cl
Cl
Cl
Cl
Cl
SUL (4)
90-93% yield
ALC (11)
OXOS (3)
5-10% yield
Figure 4. Preparation of SUL (4) using Mg Sulfinate-ZnCl2 at 120 °C or 180 °Ca
120
100
80
60
40
20
0
0
200
400
600
800
1000
Time (minutes)
OXOS/180 °C
OXOS/120 °C
SUL/180 °C
SUL/120 °C
ALC/180 °C
ALC/120 °C
a Both experiments at 17 mol% of water (vs OXOS) in reaction mixture
c. Development of a Commercial Process to Prepare AMG 232 Drug Substance
17
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 18 of 37
1
2
3
4
5
6
7
8
9
An ozonolysis of the alkene group of SUL (4) followed by oxidation of the resultant aldehyde to
the corresponding carboxylic acid group of AMG 232 would present a greener alternative to the
ruthenium oxide/sodium periodate method used for initial manufacture of the drug substance (Scheme
1). In addition, it would likely eliminate the formation of several dimeric impurities difficult to reject via
crystallization.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
To develop a safe process to run the ozonolysis reaction, an aqueous mixture was selected
(Scheme 19).20 Under these conditions, the high energy ozonide intermediate (OZO, 22) is hydrolyzed
and its accumulation is avoided. The LC area% of OZO (22) accumulated was measured relative to the
volumetric percentage of water used in the acetonitrile/water reaction mixture and the results are
reported in Figure 5. Using 10% of water, the total energy release is limited for the ozonolysis mixture
(92 J/g at the decomposition temperature 240 °C). Another parameter requiring control is the gas phase
concentration of oxygen in the vessel during processing. The limiting oxygen concentration (LOC) for
combustion of the reaction mixture has been measured at 10.75 vol%.21 The ozonolysis process is
conducted at half the LOC (~5 vol% oxygen), thus ensuring a comfortable margin to avoid possible
combustion.
Scheme 19. Preparation of AMG 232 from SUL (4) using an Ozonolysis-Pinnick Tandem Process
Me
Me
Me
Me
Me
Me
O2S
O2S
O2S
O
O
Me
Me
O
Me
Me
Me
Me
Me
Me
O3
Me
O
N
O
O
N
N
O
O
10% H2O/MeCN
O
20-30 C
Cl
Cl
Cl
Cl
Cl
Cl
SUL (4)
MOL (21)
OZO (22)
HCHO
H2O
Me
O2S
Me
O
Me
O2S
Me
Me
O2S
Me
N
i. NaClO2 (4 equiv.)
H2O (4V)
Me
O
O
Me
Me
Me
Me
Me
+ H2O2
-H2O2
OOH
OH
N
CO2H
N
CHO
Me
Me
Me
ii. NaHSO3 (4 equiv.)
H2O (4V)
Cl
Cl
Cl
Cl
Cl
Cl
AMG 232
PAC (23)
ALD (24)
Figure 5. OZO (22) LC area% Relative to Water wt% in Reaction Mixture at 20 °C
18
ACS Paragon Plus Environment

Page 19 of 37
The Journal of Organic Chemistry
1
2
3
4
5
6
15
10
5
7
8
9
0
0
5
10
15
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Water Content (wt%)
Two different modes of processing have been practiced in GMP setting for this transformation; i.
a semi-batch ozonolysis using ozone sparging in a batch vessel and; ii. a continuous stirred-tank reactor
process. At the outset, a continuous process appears attractive to alleviate the general safety concerns
associated with using an ozonolysis on manufacturing scale.22 A schematic and a picture of the
continuous ozonolysis apparatus used are presented in Figure 6 and 7, respectively. A CFS-3 ozone
generator model marketed by Ozonia was utilized to process 4.8 kg of SUL (4) while producing
approximately 0.9 mol of ozone per hour. The ozone is generated from air supply and introduced via a
valve located at the bottom of the continuous stirred-tank reactor (CSTR), as shown in Figure 6. The
starting material solution is introduced at a flow rate of 60 mL/minute using a dip tube with an outlet
above the glass frit located at the bottom of the vessel (0.9 L capacity). Vigorous agitation of the
mixture is important for proper gas dispersion and an example can be seen in Figure 7.23 Low agitation
rates (<50 rpm) leads to poor conversion to SUL (4) as gas distribution is visually poor. A nitrogen
headspace purge having three times the flow rate of the air flow introduced at the bottom is
maintained, thus ensuring that less than 5 vol% of oxygen/ozone is present in the gas phase.24
Accumulated ozonolysis reaction solution portions25 were charged to a 2M aqueous solution of sodium
chlorite (4 equivalents) and the resultant mixture agitated for 16 hours at 20 °C. Due to the low
solubility of the process intermediates (ALD, PAC) in aqueous solutions, addition of the aqueous solution
of sodium chlorite to the ozonolysis reaction stream is preferable to avoid initial precipitation of those
intermediates. This is the mode of addition used for the semi-batch ozonolysis process described below.
Figure 6. Schematic of Apparatus for Continuous Mode Ozonolysis and Pinnick Oxidation
19
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 20 of 37
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 7. Picture of the Continuous Ozonolysis Processing Apparatus
1
2
3
4
5
6
7
SUL Inlet
O3 Inlet
Glass Frit
Receiving flask
Raman Probe Propeller Cooling Jacket
20
ACS Paragon Plus Environment

Page 21 of 37
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
One disadvantage of the continuous mode of manufacturing compared to the semi-batch mode
for this ozonolysis is that the entire process is operated in the presence of excess ozone (approximately
two equivalents of ozone are used). The formation of several impurities results from using an excess of
ozone in this transformation, the most notable being the de-homologated carboxylic acid (DHCA, 26)
shown in Scheme 20. This side-product is formed in approximately 1.6 LC area% in the continuous
process performed under the conditions described above, however it has been generated in up to 3 LC
area% using larger ozone excess. A proposal for the mechanism of formation of this impurity is shown in
Scheme 20.26 α-Oxidation of the aldehyde product27 under ozonolysis conditions generates HAL (25)
(hydroxy or keto-aldehyde), which may be oxidized to the corresponding acid upon treatment with
sodium chlorite. The decarboxylation of keto-acid derivatives with sodium chlorite has been
documented previously.28
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 20. Proposed Mechanism for the Formation of DHCA (26) from ALD (24)
Me
Me
Me
Me
Me
Me
O2S
Me
O2S
Me
O2S
Me
X
Me
O
Me
O
O
O
Me
O3
NaClO2
N
CHO
N
CO2H
N
CHO
Me
Me
Me
Cl
Cl
Cl
Cl
Cl
Cl
HAL (25)
ALD (24)
X = OH and H
X = O
- HClO2 + HClO2
Me
Me
Me
Me
O2S
Me
O2S
Me
OClO
O
O
Me
CO2H
OH
Me
- CO2
N
N
CO2H
Me
Me
- HClO
Cl
Cl
Cl
Cl
DHCA (26)
A semi-batch approach to perform the ozonolysis of SUL (4) carries the advantages of processing
with an excess of alkene starting material for most of the transformation and employing a simpler
manufacturing footprint. A maximum of 0.4 LC area% of impurity DHCA (26) was observed to be formed
using this processing mode. As described above, safe processing conditions for this reaction manifold
are provided by employing an aqueous medium and by maintaining the vessel oxygen concentration
below 5 vol%. The air/ozone gas stream is diluted with nitrogen downstream of the ozone generator
but upstream of the introduction of the reacting gas through a dip tube in the reaction vessel, as seen in
Figure 8. Four times the gas flow of nitrogen is used compared to the air flow, ensuring that gas
21
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 22 of 37
1
2
3
4
5
6
7
8
9
entering the vessel contains no more than 5 vol% of oxygen/ozone. This heavy gas flow (15.6 SCFM)
necessitated custom design of a 1/8 inch diameter holes sparger to avoid surface cooling and concurrent
SUL (4) and ALD (24) crystallization during the operation. This processing mode requires a dependable
approach to monitor reaction completion. Starting material consumption has been found to be linear
for this transformation (Figure 9), thus enabling HPLC analysis of a few samples during the
transformation coupled with the knowledge of the ozone generator output to accurately predict the
time for reaction completion. A CFS-14 ozone generator model marketed by Ozonia was utilized at a
setting of 285 g of ozone/hour to process a 23 kg batch of SUL (4). The predicted time to reaction
completion using the ozone gas output (7.4 hours) was exceeded by 10% (8.2 hours actual). Ozone
reacts with water to generate hydroxy radical29 and some of the ozone is consumed in this manner,
explaining the need to use excess ozone. The process was thus not found to be mass transfer limited
using these flow and agitation rates. Upon completion of the ozonolysis process, addition of a 2M
aqueous sodium chlorite solution affords AMG 232. The mixture is treated with 2M aqueous sodium
bisulfite to eliminate all oxidants for further processing. The drug substance is obtained in >97% assay
yield and >98 LC area% purity as a solution in isopropyl acetate after completion of an aqueous work-up.
Figure 8. Schematic of Apparatus for Semi-Batch Mode Ozonolysis and Pinnick Oxidation
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 9. SUL (4) Consumption Rate for Semi-Batch Mode Ozonolysis
SUL Consumption Plot
150
100
50
0
0
500
1000
Time (minutes)
22
ACS Paragon Plus Environment

Page 23 of 37
The Journal of Organic Chemistry
1
2
3
4
d. Isolation of AMG 232
5
6
7
8
Considering that a drug substance control point enabling robust impurity rejection is unavailable
for the free-acid, several salts of AMG 232 were prepared and tested to uncover an effective candidate
to reject impurities. Over thirty organic salts and five inorganic salts of the drug substance were
evaluated for that purpose, enabling the identification of a hemi-DABCO salt (Scheme 21) and a hemi-
calcium salt as promising contenders. However, the hemi-calcium salt displayed only moderate
rejection of starting material SUL (4) or impurity DHCA (26) whereas the hemi-DABCO isopropyl acetate
salt (232-DAB, 27) showed very efficient rejection of both species. Five LC area% of SUL (4) and 1 LC
area% of DHCA (26) could be completely rejected as part of the isolation of the latter salt, levels which
are significantly higher than those observed in a typical ozonolysis-Pinnick oxidation reaction mixture
prepared in semi-batch mode. 232-DAB (27) has been observed to be a stable crystalline mono-solvate
(isopropyl acetate) hemi-DABCO salt by TGA and 1H NMR. A single polymorph of the material has been
identified. The polymorphic form and isopropyl acetate level of the material was unchanged by a
dynamic vapor sorption experiment carried out from 0 to 90% RH. In addition, a robust temperature
and anti-solvent based crystallization protocol could be designed for the hemi-DABCO isopropyl acetate
salt using isopropyl acetate and heptane. A solubility curve displaying values at different time intervals
in the crystallization process is shown in Figure 10. Using this crystallization procedure, 232-DAB (27) is
isolated in >82% isolated yield and >99.9 LC area% purity from SUL (4).30
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 21. Isolation of AMG 232
Me
O2S
Me
N
Me
O2S
Me
Me
O2S
Me
N
i. Aqueous HCl/IPAc
ii. Aqueous Sodium
Phosphate (pH 6)
i. DABCO (0.5 equiv.)
55 C
O
O
O
Me
Me
Me
Me
iii. Exchange to AcOH
Me
Me
H
N
N
CO2H
N
CO2H
CO2
Me
Me
Me
iv. AcOH/Water (6.6V)
v. Water (4.4V)/55 C
vi. 20 C
ii. 20 C
iii. Heptane (4V)
H
IPAC
Cl
Cl
Cl
vii. Water (1V)
Cl
Cl
AMG 232 rxn mixture
Cl
232-DAB (27)
SUL (4)
>82% yield from
>99.9 LC area%
2
AMG 232
>90% yield
SUL (4)
>95% assay yield from
>98 LC area%
>99.9 LC area%
100 wt%
<200 ppm AcOH
<200 water
with 4 volumes of IPAc
23
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 24 of 37
1
2
3
4
Figure 10. Solubility of 232-DAB (27) during the Crystallization Process
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
An aqueous crystallization protocol was desired to isolate the drug substance as an orthogonal
mean of purification since the crystallization of 232-DAB (27) was performed in organic solvents.
Alcohol-water solvent mixtures could not be utilized for this isolation considering the general instability
of AMG 232 in alcohol solvents above 30 °C caused by Fischer esterification processes. Aqueous
mixtures with polar aprotic solvents showed very steep solubility curves that were not conducive to
crystallization design. On the other hand, acetic acid and water could be utilized to crystallize the
material without the drawbacks detailed above and with good crystal growth characteristics (Scheme
21). A robust temperature and anti-solvent based crystallization in this solvent mixture was designed
and operated after salt break using an aqueous hydrochloric acid (2 equivalents)/isopropyl acetate
mixture, additional washes and solvent exchange from isopropyl acetate to acetic acid. A curve showing
solubility values at different time points in the crystallization process is represented in Figure 11. AMG
232 is isolated31 using this protocol in >90% isolated yield (100 wt%) and >99.9 LC area% purity with
<200 ppm of residual water and <200 ppm of residual acetic acid.
Figure 11. Solubility of AMG 232 during the Crystallization Process
24
ACS Paragon Plus Environment

Page 25 of 37
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
The material is milled using a Pallman Universal Mill (wing beater) and the results are
summarized in Table 2. The target d50 for the drug substance was set at <35 µm based on oral
absorption modeling (GastroPlus v 9.0)32 for the range of clinical doses evaluated (60 to 480 mg) to
provide complete absorption at fasted gastric pH (1.3).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 2. AMG 232 Particle Size Distribution Pre and Post Dry-Milling
Material
d10 (µm)
d50 (µm)
d90 (µm)
AMG 232 unmilled
AMG 232 milled
14.6
3.2
42.5
19.1
102
43.7
25
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 26 of 37
1
2
3
4
Conclusion
5
6
7
8
9
In summary, a robust process to manufacture AMG 232 was developed to deliver drug
substance of high purity. Significant outputs of the commercial process development efforts include: i.
use of a novel bench-stable Vilsmeier reagent, methoxymethylene-N,N-dimethyliminium methyl sulfate,
for selective in situ activation of a primary alcohol intermediate; ii. isolation of crystalline intermediate
DHO (8), enhancing the process capacity to reject impurities; iii. use of a new crystalline and stable
isopropyl calcium sulfinate reagent ensuring robust preparation of a sulfone intermediate; iv.
development of a safe ozonolysis process conducted in an aqueous solvent mixture in either batch or
continuous manufacturing mode; v. control of the drug substance purity by crystallization of a salt
rejecting impurities effectively. The new process was demonstrated to afford 18 kg of drug substance
(99.9 LC area%) in 49.8% overall yield from DLAC (1), representing a significant improvement relative to
the first-in-human process (32% overall yield).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Experimental Section
(S)-2-((2R,3R)-2-(3-chlorophenyl)-3-(4-chlorophenyl)-3-hydroxypropyl)-N-((S)-1-hydroxy-3-
methylbutan-2-yl)-2-methylpent-4-enamide (ABA, 2). L-valinol (110.8 g, 1.06 mol, 2.0 equiv.) and
toluene (205 mL) were added to (3S,5R,6R)-3-allyl-5-(3-chlorophenyl)-6-(4-chlorophenyl)-3-
methyltetrahydro-2H-pyran-2-one (DLAC, 1) (201.8 g, 0.53 mol, 98.6 wt%, 1.0 equiv.) and the resultant
mixture was agitated at reflux (115 °C controlled via jacketed vessel) for 72 hours. Upon reaction
completion, the mixture was cooled to 20 °C and diluted with toluene (800 mL). The mixture was
acidified by addition of aqueous 1N HCl (1000 mL). The phases were separated and the organic layer
was washed twice with aqueous sodium chloride (2 x 500 mL, 23 wt%). The organic phase was
concentrated to a volume of approximately 500 mL. Toluene was added (1 L) and the organic phase was
concentrated to a volume of approximately 800 mL. The ABA (2) solution (229.4 g, 0.48 mol, 90.5% AY)
1
was verified to contain less than 500 ppm of water and polished filtered. Data for isolated ABA (2): H
NMR (600 MHz, DMSO-d6) δ 7.21 (d, J = 8.5 Hz, 2H), 7.12 (dt, J = 7.6, 1.5 Hz, 1H), 7.09 (t, J = 7.6 Hz, 1H),
7.06 (t, J = 1.5 Hz, 1H), 7.00 (d, J = 8.5 Hz, 2H), 6.85 (d, J = 8.8 Hz, 1H), 6.83 (dt, J = 7.6, 1.5 Hz, 1H), 5.53
(dddd, J = 17.3, 10.6, 8.0, 6.5 Hz, 1H), 5.45 (d, J = 3.8 Hz, 1H), 4.91 (dd, J = 10.6, 2.2 Hz, 1H), 4.88 (dd, J =
17.3, 2.2 Hz, 1H), 4.78 (t, J = 3.8 Hz, 1H), 4.55 (bs, 1H), 3.56 (dq, J = 8.8, 6.4 Hz, 1H), 3.36 (q, J = 6.4 Hz,
2H), 2.87 (dt, J = 7.9, 3.8 Hz, 1H), 2.29 (dd, J = 13.7, 6.5 Hz, 1H), 2.14 (dd, J = 14.5, 3.8 Hz, 1H), 1.97 (dd, J
26
ACS Paragon Plus Environment

Page 27 of 37
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
= 14.5, 7.9 Hz, 1H), 1.89 (dd, J = 13.7, 8.0 Hz, 1H), 1.76 (oct, J = 6.4 Hz, 1H), 0.98 (s, 3H), 0.81 (d, J = 6.4
13
Hz, 3H), 0.76 (d, J = 6.4 Hz, 3H); C {1H} NMR (151 MHz, DMSO-d6) δ 175.2, 144.5, 143.4, 135.2, 131.7,
130.7, 129.2, 128.7, 128.4, 127.9, 127.3, 125.6, 117.3, 75.1, 61.2, 55.8, 48.7, 45.1, 42.2, 41.3, 28.4, 22.8,
19.5, 18.6; HRMS (ESI-TOF) m/z: [M + H]+ Calcd for C26H34Cl2NO3 478.1910; Found 478.1914.
Characterization data for isolated EST (12) [isolated via silica chromatography with
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
dichloromethane/methanol (9/1) as eluents]: H NMR (600 MHz, CDCl3) δ 7.22 (d, J = 8.5 Hz, 2H), 7.15
(dd, J = 7.4, 1.2 Hz, 1H), 7.14 (d, J = 1.2 Hz, 1H), 7.11 (t, J = 7.4 Hz, 1H), 7.04 (d, J = 8.5 Hz, 2H), 6.89 (dt, J
= 7.4, 1.2 Hz, 1H), 5.53 (ddt, J = 17.1, 10.7, 7.4 Hz, 1H), 4.98 (dd, J = 10.7, 1.9 Hz, 1H), 4.96 (dd, J = 17.1,
1.9 Hz, 1H), 4.62 (d, J = 6.2 Hz, 1H), 3.56 (dd, J = 11.0, 9.0 Hz, 1H), 3.44 (dd, J = 11.0, 3.5 Hz, 1H), 2.90
(ddd, J = 8.7, 6.2, 4.3 Hz, 1H), 2.56 (ddd, J = 9.0, 5.3, 3.5 Hz, 1H), 2.35 (dd, J = 13.8, 7.4 Hz, 1H), 2.04 (dd, J
= 13.8, 7.4 Hz, 1H), 2.02 (dd, J = 14.3, 8.7 Hz, 1H), 1.94 (dd, J = 14.3, 4.3 Hz, 1H), 1.52 (dspt, J = 6.8, 5.3
Hz, 1H), 1.03 (s, 3H), 0.85 (d, J = 6.8 Hz, 3H), 0.84 (d, J = 6.8 Hz, 3H); C {1H} NMR (151 MHz, CDCl3) δ
13
176.1, 142.8, 141.1, 133.6, 133.2, 133.1, 129.2, 129.0, 128.1, 128.0, 126.8, 118.5, 76.2, 67.8, 60.3, 55.0,
49.6, 44.9, 43.5, 40.5, 31.0, 21.7, 19.0, 17.6.
(1R,2R,4S)-2-(3-chlorophenyl)-1-(4-chlorophenyl)-4-((S)-4-isopropyl-4,5-dihydrooxazol-2-yl)-4-
methylhept-6-en-1-ol (DHO, 8). ABA (2) (0.48 mol, 1.0 equiv.) in toluene (1145 mL) was warmed to 30
°C. NaOAc (48.3 g, 0.59 mol, 1.2 equiv.) and N,N-dimethylformamide dimethyl sulfate adduct (174.1 g,
0.72 mol, 82.8 wt%, 1.5 equiv.) were added. The mixture was agitated for 2 hours and cooled to 20 °C.
The reaction was treated with aqueous NH4Cl (1250 mL, 23 wt%). The phases were separated and the
organic layer was washed twice with aqueous sodium chloride (2 x 750 mL, 11 wt%). The organic phase
was concentrated to a volume of approximately 400 mL. Acetonitrile was added (2 L) and the organic
phase was concentrated to a volume of approximately 400 mL. Acetonitrile was added (2 L) and the
organic phase was concentrated to a volume of approximately 400 mL. The solution was verified to
contain less than 5000 ppm of toluene and polished filtered. Acetonitrile (1 L) and water (330 mL) were
added to the mixture. The solution was seeded with DHO (8) (11 g, 5 mol%) at 25 °C and agitated for 1
hour. Water (1.2 L) was added over the course of 6 hours and the mixture was agitated for an additional
1 hour. The mixture was filtered and the cake was washed with an acetonitrile/water (50% v/v, 700 mL)
solution. The cake was washed with an acetonitrile/water (65% water v/v, 900 mL) solution and dried
for 24 hours under nitrogen. The residual water content of the solid was verified to be less than 1 wt%
and DHO (8) was isolated as a white crystalline solid (206.5 g, 0.45 mol, 87.7% yield over 2 steps, 98.8 LC
1
area%): H NMR (600 MHz, CDCl3) δ 7.21 (d, J = 8.4 Hz, 2H), 7.17 (ddd, J = 7.7, 1.8, 1.1 Hz, 1H), 7.13 (t, J =
7.7 Hz, 1H), 7.10 (bt, J = 1.8 Hz, 1H), 7.00 (d, J = 8.4 Hz, 2H), 6.90 (dt, J = 7.7, 1.1 Hz, 1H), 5.52 (ddt, J =
27
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 28 of 37
1
2
3
4
5
6
7
8
9
17.2, 10.2, 7.3 Hz, 1H), 4.96 (dd, J = 10.2, 1.9 Hz, 1H), 4.89 (dd, J = 17.2, 1.9 Hz, 1H), 4.74 (d, J = 5.6 Hz,
1H), 4.05 (dd, J = 9.2, 8.2 Hz, 1H), 3.73 (btd, J = 9.2, 6.8 Hz, 1H), 3.67 (t, J = 8.2 Hz, 1H), 3.19 (ddd, J = 8.2,
5.6, 4.0 Hz, 1H), 2.37 (dd, J = 13.9, 7.3 Hz, 1H), 2.16 (dd, J = 14.7, 4.0 Hz, 1H), 2.11 (dd, J = 13.9, 7.3 Hz,
1H), 1.87 (dd, J = 14.6, 8.2 Hz, 1H), 1.55 (oct, J = 6.7 Hz, 1H), 1.10 (s, 3H), 0.96 (d, J = 6.8 Hz, 3H), 0.84 (d, J
13
= 6.8 Hz, 3H); C {1H} NMR (151 MHz, CDCl3) δ 172.0, 143.4, 140.4, 133.7, 133.3, 132.9, 129.5, 129.1,
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
128.2, 127.8, 127.7, 126.7, 118.3, 76.7, 71.8, 69.8, 49.4, 42.2, 39.7, 38.6, 32.5, 24.1, 19.1, 18.3; HRMS
(ESI-TOF) m/z: [M + H]+ Calcd for C26H32Cl2NO2 460.1815; Found 460.1805; melting point = 71-75 °C.
(3S,5S,6R,8S)-8-allyl-6-(3-chlorophenyl)-5-(4-chlorophenyl)-3-isopropyl-8-methyl-2,3,5,6,7,8-
hexahydrooxazolo[3,2-a]pyridin-4-ium naphthalene-1-sulfonate toluene hemi-solvate (OXOS, 3). DHO
(8) (186 g, 0.40 mol, 1.0 equiv.) was dissolved in toluene (1000 mL). Methanesulfonic anhydride (88.2 g,
0.49 mol, 1.2 equiv.) and 2,6-lutidine (95.0 mL, 0.82 mol, 2.0 equiv.) were added to the reaction mixture.
The mixture was heated to 75 °C (controlled via jacketed vessel) and agitated for 12 hours. Upon
reaction completion, the mixture was cooled to 20 °C and dichloromethane (1600 mL) was added. A
solution of concentrated sulfuric acid (45.0 mL, 0.82 mol, 2.0 equiv.) in water (955 mL) was added and
the resultant mixture was agitated for 30 minutes. The phases were separated and the organic layer
was washed twice with an aqueous solution of sodium 1-naphthalenesulfonate (2 x 72.5 g, 0.31 mol,
0.75 equiv.) in H2O (2 x 800 mL). The organic phase was polished filtered and concentrated to a volume
of approximately 400 mL. Toluene was added (600 mL) and the organic phase was concentrated to a
volume of approximately 400 mL. The solution was verified to contain less than 2000 ppm of
dichloromethane and agitated at 20 °C for 24 hours. The mixture was filtered, the cake was washed
twice with toluene (2 x 400 mL), and dried for 24 hours under nitrogen. The residual water content of
the solid was verified to be less than 500 ppm and OXOS (3) was isolated as an off-white crystalline solid
1
(260.1 g, 0.37 mol, 90.0% yield, 99.5 LC area%): H NMR (400 MHz, CDCl3) δ ppm: 9.14 (d, J=8.50 Hz,
1H), 8.35 (dd, J=7.26, 1.24 Hz, 1H), 7.86 (t, J=8.71 Hz, 2H), 7.57 (t, J=7.70 Hz, 1H), 7.43-7.50 (m, 2H), 7.13-
7.39 (m, 7.5H), 7.03-7.10 (m, 3H), 6.07 (d, J=11.20 Hz, 1H), 5.80 (ddt, J=17.00, 9.90, 7.39, 7.39 Hz, 1H),
5.51 (t, J=9.74 Hz, 1H), 5.26-5.34 (m, 2H), 4.76 (ddd, J=10.37, 4.66, 2.18 Hz, 1H), 4.62 (dd, J=9.12, 4.77
Hz, 1H), 3.51-3.60 (m, 1H), 2.86 (t, J=13.68 Hz, 1H), 2.65-2.71 (m, 1H), 2.55-2.60 (m, 1H), 2.35 (s, 1.5H),
13
1.95 (dd, J=13.89, 3.52 Hz, 1H), 1.52 (s, 3H), 0.54-0.67 (m, 7H); C {1H} NMR (100 MHz, CDCl3) δ ppm:
183.3, 142.2, 140.0, 137.7, 135.9, 134.2, 134.1, 133.3, 132.2, 130.4, 130.3, 130.0, 129.6, 129.4, 129.1,
128.9, 128.3, 128.1, 127.9, 127.7, 127.4, 127.2, 126.4, 125.9, 125.6, 125.5, 125.2, 124.3, 121.2, 73.1,
67.3, 67.1, 43.6, 43.0, 38.7, 38.6, 26.6, 22.1, 21.3, 18.1, 13.7; HRMS (ESI-TOF) m/z: M+ Calcd for
C26H30Cl2NO 442.1699; Found 442.1710; melting point = 202-206 °C. Characterization data for isolated
28
ACS Paragon Plus Environment

Page 29 of 37
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
1
DHO-OMs (16): H NMR (400 MHz, CDCl3) δ ppm: 7.31 (d, J=8.4 Hz, 2H), 7.24-7.18 (m, 2H), 7.16 (s, 1H),
7.08 (d, J=8.2 Hz, 2H), 7.07-7.01 (m, 1H), 5.59 -5.43 (m, 2H), 5.01-4.83 (m, 2H), 3.84 (dd, J=8.1, 9.5 Hz,
1H), 3.55-3.45 (m, 1H), 3.42-3.34 (m, 1H), 3.24-3.13 (m, 1H), 2.46 (s, 3H), 2.39-2.28 (m,1H), 2.28-2.14 (m,
1H), 1.98 (br dd, J=7.8, 13.6 Hz, 1H), 1.72 (dd, J=2.4, 14.3 Hz, 1H), 1.26 (br s, 1H), 1.06 (s, 3H), 0.88 (d,
J=6.7 Hz, 3H), 0.71 (d, J=6.7 Hz, 3H); C {1H} NMR (100 MHz, CDCl3) δ ppm: 169.7, 141.5, 135.6, 135.0,
13
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
133.9, 133.4, 129.8, 129.3, 128.8, 128.5, 127.8, 127.3, 118.2, 87.0, 77.2, 69.8, 48.0, 44.6, 40.0, 39.3,
38.5, 32.8, 21.9, 19.4, 18.3; HRMS (ESI-TOF) m/z: [M + H]+ Calcd for C27H34Cl2NO4S 538.1580; Found
1
538.1571. Characterization data for isolated DHO-OTs (9): H NMR (600 MHz, CDCl3) δ 7.38 (d, J = 8.1
Hz, 2H), 7.16 (d, J = 8.9 Hz, 2H), 7.12 (ddd, J = 8.0, 2.0, 1.3 Hz, 1H), 7.12 (d, J = 8.1 Hz, 2H), 7.09 (t, J = 8.0
Hz, 1H), 6.94 (t, J = 2.0 Hz, 1H), 6.91 (d, J = 8.9 Hz, 1H), 6.91 (d, J = 8.0 Hz, 2H), 5.50 (dddd, J = 17.0, 10.2,
7.9, 6.8 Hz, 1H), 5.32 (d, J = 6.9 Hz, 1H), 4.94 (ddt, J = 10.2, 2.3, 1.0 Hz, 1H), 4.88 (ddt, J = 17.0, 2.3, 1.4
Hz, 1H), 3.81 (dd, J = 9.7, 8.3 Hz, 1H), 3.46 (ddd, J = 9.7, 8.3, 7.5 Hz, 1H), 3.34 (t, J = 8.3 Hz, 1H), 3.09 (ddd,
J = 9.8, 6.9, 2.3 Hz, 1H), 2.38 (s, 3H), 2.28 (dd, J = 13.7, 6.8 Hz, 1H), 2.10 (dd, J = 14.0, 9.8 Hz, 1H), 1.92
(dd, J = 13.7, 7.9 Hz, 1H), 1.64 (dd, J = 14.4, 2.3 Hz, 1H), 1.23 (dqui, J = 7.5, 6.7 Hz, 1H), 1.00 (s, 3H), 0.86
(d, J = 6.7 Hz, 3H), 0.69 (d, J = 6.7 Hz, 3H); 13C {1H} NMR (151 MHz, CDCl3) δ 169.8, 144.5, 141.2, 135.6,
134.4, 133.7, 133.5, 129.5, 129.1, 128.5, 128.3, 127.9, 127.5, 127.0, 118.1, 86.6, 72.1, 69.7, 48.0, 44.6,
39.8, 39.3, 32.8, 21.7, 21.5, 19.4, 18.2; HRMS (ESI-TOF) m/z: [M + H]+ Calcd for C33H38Cl2NO4S 614.1893;
Found 614.1890.
Calcium Propane-2-Sulfinate Dihydrate (CALDI, 18). Sulfur dioxide (5.23 kg, 81.6 mol, 1.3 equiv.) was
added via subsurface addition to tetrahydrofuran (14.4 L) over 2.5 hours while maintaining the reaction
mixture temperature below –15 °C. Isopropyl magnesium chloride (2 M solution in tetrahydrofuran)
(28.4 kg, 58.2 mol, 1.0 equiv.) was added to the mixture over 5 hours while maintaining the temperature
at –10 ± 5 °C and the resultant solution was agitated for 30 minutes. Upon reaction completion,
aqueous 2.5N HCl (25.7 L, 64.2 mol, 1.1 equiv.) was added over 1 hour while maintaining the reaction
mixture temperature at –10 ± 10 °C and the mixture was warmed to 20 °C. MTBE (30 L) was added and
the mixture was agitated for 30 minutes. The phases were separated and the aqueous phase was
extracted with MTBE (20 L). The combined organic phases were washed with aqueous sodium chloride
(20 L, 5 wt%), polished filtered, and concentrated under reduced pressure to afford 8.8 kg of crude
isopropylsulfinic acid (72 wt% product and 1.4 wt% tetrahydrofuran). Ethanol (58 L) was added to the
crude product and the reaction mixture was agitated at 20 °C for 15 minutes. A solution of calcium
acetate monohydrate (4.632 Kg, 26.3 mol, 0.5 equiv.) in water (11.4 L) was added to the mixture over 15
minutes and the resultant suspension was agitated for 1 hour. The mixture was filtered and the cake
29
ACS Paragon Plus Environment