Stereoselective Bulk Synthesis of CCR2 Antagonist BMS-741672: Assembly of an All-cis (S,R,R)-1,2,4-Triaminocyclohexane (TACH) Core via Sequential Heterogeneous Asymmetric Hydrogenations

Article abstract of DOI:10.1021/acs.oprd.6b00282

A concise bulk synthesis of stereochemically complex CCR2 antagonist BMS-741672 is reported. A distinct structural feature is the chiral all-cis 1,2,4-triaminocyclohexane (TACH) core, which was assembled through consecutive stereocontrolled heterogeneous

Full text of DOI:10.1021/acs.oprd.6b00282

Subscriber access provided by CORNELL UNIVERSITY LIBRARY
Full Paper
Stereoselective bulk synthesis of CCR2 antagonist BMS-741672:
Assembly of an all-cis (S,R,R)-1,2,4-triamino-cyclohexane (TACH)
core via sequential heterogeneous asymmetric hydrogenations.
Joerg Deerberg, Siva J. Prasad, Chris Sfouggatakis, Martin D. Eastgate, Yu Fan,
Ramakrishnan Chidambaram, Praveen Kumar Sharma, Li Li, Richard L. Schild,
Jale Muslehiddinoglu, Hyei-Jha Chung, Simon W. Leung, and Victor W. Rosso
Org. Process Res. Dev., Just Accepted Manuscript • DOI: 10.1021/acs.oprd.6b00282 • Publication Date (Web): 13 Oct 2016
Downloaded from http://pubs.acs.org on October 14, 2016
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 free 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 accessible to all
readers and 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.
Organic Process Research & Development 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 41
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
Stereoselective bulk synthesis of CCR2 antagonist BMSꢀ741672: Assembly of
an all-cis (S,R,R)ꢀ1,2,4ꢀtriaminoꢀcyclohexane (TACH) core via sequential
heterogeneous asymmetric hydrogenations
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
Joerg Deerberg,* Siva J. Prasad, Chris Sfouggatakis, Martin D. Eastgate, Yu Fan, Ramakrishnan
Chidambaram,† Praveen Sharma,†† Li Li, Richard Schild, Jale Müslehiddinoğlu,# HyeiꢀJha Chung,##
Simon Leung, and Victor Rosso
.
Chemical and Synthetic Development, BristolꢀMyers Squibb Company, One Squibb Drive, New Brunswick, New
Jersey 08901, U.S.A.
ACS Paragon Plus Environment
1

Organic Process Research & Development
Page 2 of 41
1
2
3
4
5
TOC FIGURE:
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
ACS Paragon Plus Environment
2

Page 3 of 41
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
ABSTRACT: A concise bulk synthesis of stereochemically complex CCR2 antagonist BMSꢀ741672 is reported.
A distinct structural feature is the chiral allꢀcis 1,2,4ꢀtriaminocyclohexane (TACH) core, which was assembled
through consecutive stereocontrolled heterogeneous hydrogenations: efficient Ptꢀcatalyzed reduction of a
βꢀenaminoester, directed by (S)ꢀαꢀmethylbenzylamine as a lowꢀcost chiral template, and reductive amination of a
3,4ꢀcisꢀdisubstituted cyclohexanone over sulfided Pt/C introduced a tertꢀamine, setting the third stereocenter in
the allꢀcis cyclohexane core. The heterogeneous catalysts were recycled. Ester hydrolysis produced a γꢀ
aminoacid, isolated as its Na salt. A challenging Curtius reaction to introduce the remaining C−N bond at Cꢀ2 was
strongly influenced by the presence of the basic tertꢀamine, providing a stereoelectronically highly activated
isocyanate. Detailed mechanistic and process knowledge was required to enable clean trapping with an alcohol
(tꢀBuOH) while avoiding formation of side products, particularly an unusual carbamoyl phosphate. Deprotection,
Nꢀacetylation, and uncatalyzed SNAr coupling with known 4ꢀchloroquinazoline provided the final product. The
resulting 12ꢀstep synthesis was used to prepare 50 kg of the target compound in an average yield of 82% per step.
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
KEYWORDS: CCR2 Antagonist, allꢀcis Triaminocyclohexane (TACH), Freidinger lactam, Curtius
rearrangement, carbamoyl phosphate.
ACS Paragon Plus Environment
3

Organic Process Research & Development
Page 4 of 41
1
2
3
4
5
6
7
8
9
■
INTRODUCTION
The chemokine family of proteins and their receptors play a pivotal role in directing leukocyte migration,
activation, and angiogenesis.1 Chemotactic chemokine receptor 2 (CCR2),2 a Gꢀproteinꢀcoupled receptor,3 has
elicited sustained interest in the pharmaceutical industry as a prospective therapeutic target for the treatment of
inflammatory, cardiovascular and metabolic diseases, such as rheumatoid arthritis, multiple sclerosis,
atherosclerosis, and diabetes mellitus.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
Figure 1. Structure of Target Compound 1
AcHN
N
N
(S)
(R)
N
(S)
N
H
O
(R)
N
1
Me
CF3
BMSꢀ741672
BMSꢀ741672, 1 (Figure 1), is a highly selective CCR2 antagonist (IC50=1.4 nM) featuring a complex array of
four stereocenters. The key synthetic challenge was efficient assembly of the densely functionalized 1,2,4ꢀ
triaminocyclohexane (TACH) core in a minimum number of linear steps. While the basic substitution type was
known,5 the only precedent to access the allꢀcis arrangement with full stereocontrol and differentiation of
nitrogens was provided by our medicinal chemistry team.6 This diversityꢀoriented concept (Scheme 1) required 19
linear steps to 1 and involved enzymatic desymmetrization of mesoꢀanhydride 2,6a Curtius rearrangement of the
secoacid, γꢀiodolactamization to set the stereochemistry at C4 in 3, dehalogenation with tributyltin hydride, and
Hofmann degradation to establish the C2ꢀN bond in 4, followed by sequential reductive aminations at N4, and
SNAr quinazoline coupling to give 1.6b
To support expeditious development of BMSꢀ741672, we required a more concise enantioꢀ and
diastereocontrolled synthesis that was fully amenable to plant scale operations. Ideally, despite the target’s
complexity, the linear sequence would not exceed 12 steps. Through such a synthesis, we sought to support a fastꢀ
moving campaign to deliver 50 kg of highꢀquality drug substance 1.
From a strategic perspective, in order to advance our key objectives − stereocontrol, product quality, and
synthetic brevity − we opted for a new, more streamlined disconnection strategy (Scheme 2): In our design,
convergent quinazoline SNAr coupling in the final step and contraction of the ester at C2 to a C−N bond via
Curtius reaction would lead to allꢀcis 1,4ꢀdiaminoꢀ2ꢀcarboxylate 5, which we expected to access from ketone 6
through direct reductive amination with isopropylmethylamine, thus setting the stereochemistry at C4.
ACS Paragon Plus Environment
4

Page 5 of 41
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
Scheme 1. Medicinal Chemistry Synthetic Strategy towards 1
O
O
AcHN
9 steps
7 steps
NHCbz
2
4
2
1
N
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
O
BocꢀN
2
4
NHCbz
4
1
O
Ref. 6a
Ref. 6b
HN
Boc
O
2
3
4
Ι
3 steps
1
Our confidence in a Siꢀfacial reduction of 6 was based on observations with 3,4ꢀcisꢀdisubstitued
cyclohexanones employed to access an earlier chemotype of CCR2 antagonists.7 To introduce the (S)ꢀ3ꢀaminoꢀ
pyrrolidone subunit, we expected to utilize established methodologies from common aminoꢀacid building blocks
8,8 which would allow for further simplification to cyclic βꢀaminoester 7.
For assembly of this structural motif, we herein demonstrate an effective stereocontrolled methodology we
developed9 using heterogeneous catalytic hydrogenation of chiral cyclic βꢀenaminoesters, such as 9, readily
accessible through condensation of inexpensive, optically pure αꢀmethylbenzylamine with known monoketalꢀ
protected βꢀketoester 10.
■ RESULTS AND DISCUSSION
Heterogeneous catalytic hydrogenation. Introduction of the first two stereocenters. Following the
above strategy, we first assembled βꢀenaminoester 15 (Scheme 3), which required efficient access to its precursor,
βꢀketoester 13. Based on bulk availability of 1,4ꢀcyclohexadione monoꢀethylene ketal 12, we elected to prepare
13 by direct acylation10 with diethylcarbonate, in preference to Dieckmann cyclization from diester 11.11 Key
advantages of 12 are its commercial availability, crystallinity, and relatively high melting point (74 °C).
ACS Paragon Plus Environment
5

Organic Process Research & Development
Page 6 of 41
1
2
3
4
5
6
Scheme 2. Disconnection strategy for the all-cis 1,2,4ꢀtriaminocyclohexane towards 1
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
Scheme 3. Preparation of chiral βꢀenaminoester 15
Step 1
Step 2
EtO2C
EtO2C
(S)
H
14
O
OH
N
H2N
Ph
(EtO)2CO
2
O
(S)
1
O
O
Ph
NaH, ꢀ
cat. HOAc
MeOH, 40−45°C
O
THF, cat. EtOH
O
O
12
13
15
(87%, 2 steps)
(bulk commercial)
CO2Et
Ref. 11
O
O
CO2Et
11
While the acylation was precedented,10 we felt a need for closer examination of this reaction prior to application
on large scale (6 m3 equipment). A variety of homogeneous conditions led to significant amounts of selfꢀ
condensation and unacceptably low yields of 13. Among heterogeneous bases, sodium hydride10aꢀc stood out,
providing 13 in nearꢀquantitative yield. This presented challenges from a perspective of both safety (reagent
handling, hydrogen evolution), and quality (removal of mineral oil from 60 wt% NaH dispersions commercially
available in bulk).
To address these issues, we profiled heat flow and gas evolution rates in an RCꢀ1 calorimeter equipped with a
gas flow meter (Figure 2). Using slow addition of substrate 12 over 2 h to a slurry of NaH (2.2 equiv) in THF and
diethylꢀcarbonate (2.0 equiv) at 60 °C, we found that mild heat evolution (ꢀH= −160 kJ/mol) occurred with full
addition control, so long as catalytic amounts of alcohol (EtOH, 5 mol%) were present. Omitting the alcohol
ACS Paragon Plus Environment
6

Page 7 of 41
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
resulted in reagent accumulation, delayed exotherms and inconsistent gas evolution. In the presence of alcohol,
90% of heat and 96% of gas evolved evenly throughout the addition period, with the hydrogen mass balance
effectively closed (>98%) within 10−20 min of additional hold time. Upon completion of the acylation, the
reaction mass was quenched in a doseꢀcontrolled manner into an aqueous solution of dilute HOAc, rendering 13
in solution in nearꢀquantitative yield after solvent exchange to MeOH.
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
We then developed conditions to remove mineral oil through a phase split as part of the final workup after the
subsequent amination (Step 2, vide infra). This allowed standard dry handling of the NaH dispersion at the initial
setup charge for the twoꢀstage process, obviating the need to wash the dispersion with hydrocarbons prior to use,
effectively eliminating associated hazards. To ensure process safety on scale (Step 1), offꢀgases evolved from
both the reaction vessel and the quench tank were vented through a flame arrestor. Upon completion of the
quench, subsurface sparging with nitrogen gas was used to drive out any residual hydrogen, then vents were
switched back to a thermal oxidizer.
The next step involved condensation of 13 with (S)ꢀαꢀmethylbenzylamine 14 (αꢀMBA). We found catalytic
amounts of HOAc (~3 mol%) significantly accelerated the reaction when performed in alcoholic media (e.g.,
MeOH) at only slightly elevated temperatures (40−45°C, 8 h).12 By contrast, standard DeanꢀStark conditions13 in
refluxing PhMe led to very slow reactions (>18 h), accompanied by significant amounts of amide byꢀproduct.
Conversely, acids stronger than HOAc (e.g., TFA) led to incomplete reaction and partial decomposition of 15.
Thus, under optimized conditions, the crude product was obtained in MeOH solution, the supernatant mineral
oil layer was removed (phase split), and pure 15 was isolated as a crystalline solid upon addition of water,
filtration, and drying in 87% yield across the two steps (6 m3 scale). Using this protocol, we manufactured 2,116
kg of 15 in five batches during our campaign, starting from 1,150 kg of commercial ketone 12.
ACS Paragon Plus Environment
7

Organic Process Research & Development
Page 8 of 41
1
2
3
4
5
6
Figure 2. Additionꢀcontrolled H2 evolution during acylation of 12
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
This efficient route to 15 now set the stage for introduction of the first two stereocenters (C1, C2) in the
cyclohexane core. With a largeꢀscale application in mind, we were intrigued by the prospects of using
inexpensive (S)ꢀαꢀmethylbenzylamine 14 to direct faciality in the cisꢀselective reduction of enaminoesters I
(Scheme 4). Palmieri and coworkers14 had introduced this concept using borohydrides as reductants in protic
media (Step A), giving rise to a mixture of four diastereomers in which isomer II predominates. In the case of a
cyclic enaminoester II (with R=iꢀBu; R1,R2= ), Xu et al.14c subsequently demonstrated selective removal
−(CH2)4−
of minor diastereomers IIIꢀV via salt formation and crystallization (Step B) of desired II·HBr salt.
Hydrogenolytic removal of the auxiliary yielded optically pure βꢀaminoesters VI·HBr (Step C).
Scheme 4. Palmieri’s approach to chiral βꢀaminoesters14
ACS Paragon Plus Environment
8

Page 9 of 41
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
Despite its formal simplicity and readily available reagents, the approach often provided relatively moderate
facial selectivity (~39ꢀ85% d.e., depending on substitution type R1,R2) and a highly volumeꢀintensive workup
(>50 L/kg) for quench and removal of boronate waste after the first reduction (Step A) using sodium
triacyloxyborohydrides.
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
Indeed, when 15 was reduced with NaBH(OAc)3 (STAB), a mixture of all four possible diastereomers of the
βꢀaminoester 16 was obtained. The desired cisꢀproduct 16a predominated, though selectivity was only moderate
(77% diastereomeric purity, Table 1, Entry 1). Salt formation with TsOH15 led to enriched, crystalline 16a•TsOH
(97.1% diastereomeric purity) in low yield (40%). Monitoring of the mother liquor showed that, even under
optimized crystallization conditions, the amount of desired diastereomer 16a entrained was larger than the
proportion of the three undesired isomers16bꢀd combined.
Table 1. Pt catalyzed reduction of cyclic βꢀenaminoester 15
Reductant/
% Aryl
% Enamine
Crude d.r. %
% Yield
Entry
(H2 pressure) Reductiona Reductiona 16a vs. other isom.b (purity)c
NaBH(OAc)3
1.5 equiv
40%
(97.1%)
1
2
n.d.
12.0
0.3
100
88.0
99.7
77: 16: 7 : <0.2
94 : 3 : 3 : <0.2
88 : 6 : 6 : <0.2
PtO2 , 13 mol%
(50 psig)
58%
(99.6%)
Pt/C, 0.4 mol%
(50 psig)
70%
(99.5%)
3
a
b
c
Determined by GC; Determined by HPLC; % Diastereomeric purity of
isolated product 16a (TsOH salt) by HPLC.
To improve selectivity, we were intrigued by early reports from Melillo16 and Lhommet17 describing a high
degree of stereoselectivity (~90% diastereomeric purity) in the hydrogenation of structurally diverse βꢀ
enaminoesters derived from 14 on unsupported platinum oxide. While overꢀreduction of the auxiliary’s phenyl
ring was observed, byꢀproducts from debenzylation were not reported. Intriguingly, prior to our work,9 this
hydrogenation concept had not been described for cyclic βꢀenaminoester substrates of type I (Scheme 4).
When applied to 15, hydrogenation on PtO2 in EtOH in the presence of HOAc (2 equiv) immediately led to
improvements in stereoselectivity (94% diastereomeric purity, Table 1, Entry 2). As we had hoped, this boost in
ACS Paragon Plus Environment
9

Organic Process Research & Development
Page 10 of 41
1
2
3
4
5
selectivity directly translated to strong gains in isolated yields, leading to desired isomer 16a•TsOH in high purity
6
7
(99.6% diastereomeric purity) and 58% yield after TsOH salt formation.
8
9
However, the considerable cost of PtO2, combined with a high catalyst loading (>10 mol%) required to
complete the reaction, and a significant amount of overꢀreduction of the phenyl substituent (~12%, GC),
prompted us to search for a more practical alternative. We found that when using a supported platinum catalyst (5
wt% Pt/C, 50% wet), loadings could be dramatically reduced to 0.4 mol%, while Nꢀdebenzylation (<0.1%) and
overꢀreduction of the phenyl ring (<0.3%) became practically insignificant.18 A slight drop in selectivity to 88%
diastereomeric purity was more than offset by an increase in isolated yields to 70% (Table 1, Entry 3). On scale,
the optimized process performed very reliably in a 6 m3 batch hydrogenator, allowing us to produce metricꢀton
quantities (1,747 kg) of highꢀquality 16a•TsOH in >99.5% diastereomeric purity. The heterogeneous catalyst was
recovered by simple filtration and the precious metal was recycled.
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 3. Rationale: Preferential Reꢀface reduction at Cβ of 15
We rationalize the observed facial bias through preferential Reꢀface attack of a (metal)ꢀhydride at C
(Figure 3), preꢀorganized via hydrogen bonding between the amine and the ester carbonyl, as well as through
conformational lock of the benzylic CꢀN bond through A1,3ꢀstrain against ꢀCH2.17b,19 Presumably, the hydride
donor in the catalyzed variant exerts increased steric demand compared to small borohydrides, leading to
enhanced selectivity.20 Kinetic protonation of the resulting transient ketene acetal at C
by the protic reaction
β of 15
γ
α
solvent from the same (top) face as the hydride donor then provides preferentially cisꢀ16a.
ACS Paragon Plus Environment
10

Page 11 of 41
Organic Process Research & Development
1
2
3
4
5
6
Scheme 5. Preparation of chiral cisꢀβꢀaminoester 21a
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
Deprotection of the benzyl group via hydrogenolysis using Pd/C (10 wt% Pd, 50% wet; 0.9 mol%) proceeded
smoothly to yield isomerically pure 17a, which was isolated after solvent exchange to iꢀPrOAc (IPAc), filtration
and drying in 89% yield (Scheme 5). Thus, in our campaign, homochiral cisꢀβꢀaminoester (1S,2R)ꢀ17a (1,158 kg)
was obtained from readily accessible βꢀenaminoester 15 in two synthetic steps and 62% overall yield.
Introduction of the (S)-3-amino-pyrrolidone moiety. Chiral, conformationally constrained lactams of
the Freidinger pyrrolidone type8,21 are commonly assembled starting from primary amines through one of two
approaches: (a) The Freidinger synthesis,8a,b consisting of Nꢀacylation with a protected methionine (Scheme 2, 8:
X=H,SMe), followed by Sꢀmethylation and baseꢀpromoted cyclization with extrusion of dimethylsulfide (DMS),
or, less frequently: (b) via reductive amination with a protected aspartic aldehyde (Scheme 2, 8: X=O), followed
by thermal lactamization.8cꢀe Given our time constraints and the ready bulk availability of LꢀNꢀCbzꢀMetꢀOH (18)
we selected this option for our campaign.
Following the above concept (Scheme 6, Step 5), volumeꢀefficient amide coupling (MeCN 2.2 L/kg) between
βꢀaminoester 17a and 18 proceeded smoothly using EDAc·HCl (1.1 equiv) and HOBt·monohydrate (1.0 equiv).
To avoid potential epimerization of the aminoacid, the exothermic reaction was controlled to ≤30 °C through slow
addition of DIPEA (2.2 equiv), providing amide 19 as a single isomer in high yield (96%) after workꢀup. The
resulting EtOAc solution was directly carried forward into the subsequent Sꢀalkylation (Step 6).
Among several alkylating reagents tested (methyl bromoꢀacetate, dimethylsulfate, MeI), we found that only
MeI provided meaningful amounts of lactam in the subsequent cyclization step. Using excess MeI (15 equiv.) at
20−25 °C, pure sulfonium salt 20 precipitated from the reaction mass, driving the (partially reversible) alkylation
to completion within ~20h. The product was isolated by filtration and the cake was washed with sufficient MTBE
ACS Paragon Plus Environment
11

Organic Process Research & Development
Page 12 of 41
1
2
3
4
5
6
7
8
9
(~11 L/kg) until residual MeI was no longer detectable in the rinses (criteria: ≤ 0.005% v/v, GC).22 The wetcake
was dried in a Nutsche filter using a stream of nitrogen gas while maintaining drying temperatures of 20−25 °C in
order to avoid regeneration of MeI from thermal dealkylation of 20. Pure dry cake (>99.6% HPLC purity) was
discharged under full containment into sealed containers, which were directly amenable to the equipment used in
the next step. Overall, a total of 1,294 kg of 20 was produced in 93% yield from 17a.
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
For the Freidinger cyclization (Step 7a), we tested bases under homogeneous and heterogeneous conditions
(Table 2), as well as inert dipolar solvents, in an effort to identify viable substitutes for Cs2CO3 and DMSO,
which are frequently used for this transformation.8a,b Significant amounts of byꢀproducts 23−25 (Scheme 6) were
observed under most conditions tried. While the structurally differentiated βꢀelimination product 25 was readily
purged during downstream isolation post deketalization (Step 7b, vide infra), tolerance for the C2ꢀepimeric ester
24 was below 1%.
Table 2. Screening of cyclization conditions
− Impurity profile
Base (equiv)/
Time
h
20a
21a
23a
24a
25a
Entry
1
Solvent, Temp.
%
%
%
%
%
KOtꢀBu (1.2)
DMF, 20−25 °C
2.2
6.8
6.3
7.2
1.0
ꢀ
74.8
71.0
87.2
91.7
n.d.d 10.0
n.d.d 18.2
11.0
4.7
NaHMDS (1.0)
DMF, 20−25 °C
2
3
4
LiHMDS (1.0)
DMF, 20−25 °C
ꢀ
n.d.d
n.d.d
4.8
4.7
5.0
Cs2CO3 (1.2)
DMF, 20−25 °C
0.5
2.8
b
n.d.d
1.8
6.3
Cs2CO3 (1.7)
5
37
0.9
90.9
DMSO, 20 °C
c
Cs2CO3 (1.1)
6
7.0
1.1
92.4
0.5
0.1
3.1
DMSO, 20 °C
a
b
HPLC area% of crude reaction mass; Partial suspension of solids in glassꢀlined
c
reactor with 2ꢀblade retreatꢀcurve impeller; Full suspension in glassꢀlined reactor
d
equipped with highꢀcapacity pumpꢀaround loop; Not determined.
ACS Paragon Plus Environment
12

Page 13 of 41
Organic Process Research & Development
1
2
3
4
5
6
Scheme 6. Freidinger synthesis of (S)ꢀ3ꢀaminoꢀpyrrolidone 2226
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
We therefore reverted to Cs2CO3 in DMSO (Table 2, Entries 5 & 6), which produced the lowest amounts of
epimer 24. At the observed levels, elimination byꢀproduct 25 was of no concern, though it detracted from overall
yields. However, downstream tolerance for Nꢀmethylated byꢀproduct 23 was also below 1%. We found that
formation of this impurity correlated with thermal Sꢀdemethylation of 20 in DMSO solution above 35 °C. To
prevent this sideꢀreaction, the temperature for the cyclization was controlled in the 20−25 °C range.
A systematic investigation of mixing regimes for the heterogeneous system showed that a stoichiometric
amount of Cs2CO3 base was sufficient to complete the reaction, so long as the heavy solids remained evenly
suspended throughout the reaction.23 These conditions (Table 2, Entry 6) produced the least amount of byꢀ
products and were thus selected for scaleꢀup. We found that particle size of Cs2CO3 played only a minor role for
reaction performance; Wetꢀmilling was ineffective for particle attrition and was marginally beneficial only
inasmuch as it aided with solids suspension. Working in a fully suspended regime, we found the product 21 to be
slightly unstable towards excess base, giving rise to gradually increasing amounts of 24 and 25 during prolonged
hold times, with the rate of sideꢀreactions increasing with the amount of excess Cs2CO3 used.23
At plant scale, we limited the Cs2CO3 charge to 1.1 equiv. To accommodate the mixing requirements, we
implemented a highꢀcapacity pumpꢀaround loop, equipped with a centrifugal pump (min. pump rate: 300 L/min)
as an assist to the poor mixing action provided by the retreatꢀblade impeller in the glassꢀlined reactor alone.
Pumping the suspension out through the bottom valve while rapidly feeding it back through the top of the vessel
provided continuous turnover of the reactor contents in <10 min. Under optimized conditions, typical reaction
times were 6−8 h (completion criteria: <4% 20, HPLC), providing highꢀquality crude ketal 21 in DMSO solution
(HPLC purity: >92%).
To control the emission of Me2S, a constant stream of nitrogen was sparged through the suspension, with vent
lines directed into a twoꢀstage scrubber system containing aqueous bleach and caustic solutions (range: pH
ACS Paragon Plus Environment
13

Organic Process Research & Development
Page 14 of 41
1
2
3
4
5
6
7
10−14). Throughout the operation, the pH was monitored to ensure effectiveness of the scrubbers, adding an
aqueous solution of NaOH as required.
8
9
Post reaction completion, subꢀsurface sparging of nitrogen was maintained for a period of time (~30 min), then
solids were removed by filtration to further protect the product from unreacted excess base.24 The spent cake was
washed with EtOAc and the crude product in the filtrate was extracted into EtOAc after brine workꢀup. The rich
organic phase was directly introduced into the ketal deprotection using 1N aq. HCl in acetone. Isolation was
achieved by concentration in vacuo, addition of water, and filtration. The highly crystalline ketone 22 was
obtained as a single isomer (>99.8% diastereomeric purity, >98.3% purity) in 69% yield over four steps (from
17a), producing a total of 611 kg of 22 in the campaign.
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
Reductive amination of 22. For introduction of the remaining stereocenter at C4 of the core, we opted for a
reductive amination approach utilizing substrate 22. Based on our experience with 3,4ꢀcisꢀdisubstituted
cyclohexanones,7 we expected a selective Siꢀfacial reduction to occur. Indeed, when this ketone was condensed
26
with iꢀPrN(Me)H under Mattson dehydration conditions25 using Ti(OiꢀPr)4 (1.5 equiv) in DCM, followed by
reduction with NaBH4 (1.1 equiv) in EtOH, provided a ~5.6:1 ratio of diastereomers (4R)ꢀ26 vs. (4S)ꢀ27, i.e.,
favoring the desired allꢀcis cyclohexane product (Table 3, Entry 1). The modest selectivity, combined with the
need to quench the leftover hydride reagent after reaction completion, resulting in a volumeꢀintensive workup
(~35 L/kg vs. 22), rendered isolation of the product inefficient. Upon crystallization (EtOAc/nꢀheptane), 26 was
obtained in modest yield (56%) and low isomeric purity (96.9% diastereomeric purity) − well below our target of
99.5% diastereomeric purity.
To boost throughput and purity, we examined the reduction step in more detail. Gratifyingly, we found that
hydrogenation on Ptꢀbased catalysts led to marked improvements in facial selectivity. Sulfided platinum on
carbon (Table 3, Entry 5), selected to suppress adventitious deprotection of the NꢀCbz group, yielded the largest
proportion of desired isomer 26, enhancing the diastereomeric ratio to ~15:1 while sharply reducing solvent use
(~11 L/kg vs. 22). These conditions were explored for potential scaleꢀup. However, isomeric purity of 26 after
crystallization remained unacceptably low (97.0% diastereomeric purity, Entry 5). The option of recrystallizing
the crude material was unattractive due to significant losses to the mother liquor (~10%). We therefore explored
telescoping the reduction step into the subsequent ester hydrolysis (Step 9, Scheme 7).
ACS Paragon Plus Environment
14

Page 15 of 41
Organic Process Research & Development
1
2
3
4
5
Table 3. Reductive amination of 3,4ꢀcis disubstituted ketone 22
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
Reductant (equiv) p[H2] Time 22a
26b
27b
26 Yield %
Entry
Solvent / temp.
(psig)
h
SM %
%
%
(purity)c
NaBH4 (1.1 equiv)
EtOH/RT
56%
(96.9%)
1
2
3
4
n/a
2
<0.5 85.0 15.0
PtO2 (15 mol%)
EtOH/45 ºC
n/d d
n/d d
n/d d
45
30
30
30
20 <0.5 87.5 12.5
5%Pt/Al2O3 (1 mol%)
DCM/RT
12
6
0.8
0.7
0.7
89.8 10.8
90.3 9.7
93.8 6.2
5%Pt(S)/C (1 mol%)
DCMꢀEtOH 70:30, RT
5%Pt(S)/C (0.5 mol%)
DCM-EtOAc/RT
71%
(97.0%)
5
6
a
b
HPLC area% (in crude reaction mass); Relative area% between isomers;
% Diastereomeric purity of intermediate (4R)ꢀ26 isolated by crystallization;
Not isolated.
c
d
Prior to telescoping the sequence on scale, to suppress formation of nonꢀaminated byꢀproduct 28 (Scheme 7),
we needed to understand the kinetics for the initial dehydration step. In addition, a workable protocol for removal
of TiO2 was required. FTꢀIR monitoring of the dehydration (Figure 4) showed buildꢀup of a distinct new species
at 1,636 cm−1 within 5−6 h.27 Under these optimized conditions, hydrogenation was performed using 0.45 mol%
Pt/(S)/C. Upon completion, the catalyst was removed by inert filtration and was recycled.
Figure 4. FTꢀIR: Intermediate from Dehydration (1,636 cm−1
)
t = 5 h
t = 3 h
Wavenumbers [cm−1]
ACS Paragon Plus Environment
15

Organic Process Research & Development
Page 16 of 41
1
2
3
4
5
6
Scheme 7. Reductive amination of 3,4ꢀcis disubstituted cyclohexanone 22 and isolation of γꢀaminoacid 31
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
For removal of TiO2, we found that fully inerted,28 waterꢀsaturated EtOAc was most effective at furnishing a
filterable solid with minimal entrainment of product (<0.5%). Post addition of wet EtOAc, a sufficiently long
incubation time (6−8 h) was critical to ensure high recovery. Inert filtration upon discharge from the vessel
yielded productꢀrich liquors (85−90% solution assay for desired 26, 15:1 d.r.), essentially free of residual Ti
(<100 ppm) and Pt (<20 ppm).29 The filtrate was carried forward directly into the subsequent ester hydrolysis
(Step 9, Scheme 7).
Scheme 8. Rationale: Preferential Siꢀface reduction of 33
ACS Paragon Plus Environment
16

Page 17 of 41
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
We rationalize the increased faciality (Scheme 8) for the heterogeneous reduction catalyst through preferential
equatorial attack30 of the bulkier Ptꢀhydride (relative to borohydrides) from the less crowded Siꢀface of
iminium/enamine 33. This intermediate would be locked into a substrateꢀlike conformation with the bulky
pyrrolidone substituent occupying the C1eq position. (c.f.: 1ꢀHeq in 22: J=12,4,4 Hz). Interestingly, the pyrrolidone
moiety in the allꢀcis reduction product 26 was oriented axially (1ꢀHeq: J=4,4,4 Hz),31 likely due to unfavorable
transannular interactions between the ester group and the branched secꢀamine in their alternate 1,3ꢀaxial
orientations.
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
Ester hydrolysis and isolation of isomerically pure γ-aminoacid. Given the 1,2ꢀcis substitution in 26,
saponification of the purified ethyl ester under a variety of basic conditions was predictably complicated by
significant epimerization at C2 (Table 4). A control experiment equilibrating 26 with NaOtꢀBu (THF, RT, 4h)
showed the thermodynamic diastereomeric ratio (d.r.) for the input esters to be completely favoring the transꢀ
epimer 30 (>95:5 d.r.; Scheme 7).
Even under mild basic conditions at 0 °C (Table 4, Entry 3) epimerization could not be controlled. Approx.
16% of undesired acid transꢀ(2S)ꢀ32 were incurred, while the unreacted ester starting material (~11%) was found
to have epimerized almost entirely to transꢀ(2S)ꢀ30.
Table 4. Ester Hydrolysis of 26
Reagent(s)/
Solvent
Time 26a 29a (2S)ꢀ30a (2S)ꢀ32a 29 Yield %
SM Prod. (epimer) (epimer) (purity)c
Entry
h
1
2
12 N.R.b N.R.b
10 7.5 56.3
ꢀ
ꢀ
n/d d
n/d d
LiI, 80 ºC, nꢀBuOAc
LiOH, THF/water,
7.5
28.2
0 ºC
→ RT
KOH (5 equiv)/1,2ꢀ
propanediol (4 equiv), 24 <0.5 72.2
THF/water, 0 ºC
68%
(98.6%)
3
11.2
16.1
66% e
(99.6%)
2N aq. HCl (10 L/kg),
20 4.0 95.0
4
<0.5
<0.5
60 ºC, then NaOMe
a
b
c
HPLC rel. area% (in crude reaction mass); No reaction, % Diastereomeric
d
e
purity of isomer (2R)ꢀ29 enriched by extraction; Not isolated; Over two steps
from 22, isolated as 31.
As a key observation during workup, we found that the desired allꢀcis acid (2R)ꢀ29 was readily separated from
its C2 transꢀepimer 30 through acidic aqueous extraction (aq. HCl), followed by selective partitioning32 into
DCM at pH 5−6, yielding neutral 29 as an amorphous solid upon evaporation of the solvent in 68% yield and
98.6% diastereomeric purity. Attempts to crystallize the neutral γꢀaminoacid 29 and upgrade its purity to the
ACS Paragon Plus Environment
17

Organic Process Research & Development
Page 18 of 41
1
2
3
4
5
6
7
target range (≥99.5% diastereomeric purity) remained unsuccessful. A broad screen of hydrolytic enzymes tried to
saponify 26 did not provide meaningful conversion to the aminoacid.
8
9
Gratifyingly, acidic hydrolysis of 26 in water (2N aq. HCl) at 55−60 ºC yielded desired 29 without detectable
epimerization and little adventitious cleavage of the NꢀCbz group.33 Interestingly, when applied to the crude C4ꢀ
epimeric mixture of γꢀaminoesters (26 + 27) from the above reductive amination step, undesired 27 proved
largely resistant to acid hydrolysis and remained 70−80% unchanged. These findings now confirmed the ability to
telescope the reductive amination and hydrolysis (Steps 8 & 9) while removing all impurities (leftover 26/27, and
traces of 30/32) in the final workup through sequential extractions.
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
Indeed, using the crude EtOAc/DCM stream from Step 8 (post filtration of Ptꢀcatalyst and Ti residue),
extraction with 2N aq. HCl entrained esters 26 and 27 (~15:1 ratio) into the aqueous layer. After phase split, the
acidic solution was heated (60 °C, 20 h), providing hydrolysis product 29 essentially free of undesired C2ꢀepimer
32 (<0.5%). Upon cooling of the reaction mass to RT, the aqueous layer was basified (pH 9−10) and unreacted
esters 26 and 27 were removed by extraction into toluene. Adjustment of the aqueous phase to pH 6.5−7.5,
followed by selective backꢀextraction into DCM provided isomerically pure allꢀcis 29 (~99.6% diastereomeric
purity) in 70ꢀ72% solution yield (over two steps, from 22).
Having achieved isomeric purity, our next strategic goal for the sequence (Steps 8 & 9) was isolation of the
crystalline γꢀaminoacid to ensure stability during storage and transport, as well as flowability of the solid. Highꢀ
throughput screening identified Naꢀcarboxylate salt 31 as a nonꢀhygroscopic34 crystalline solid. Among a number
of sodium bases investigated (NaHMDS, NaOtꢀBu, NaOPh, NaOMe), use of a 25−30 wt% solution of NaOMe in
MeOH was found to yield the highest quality solid 31 for both purity and potency. Combination of the salt
formation with the above extractive purification procedure for 29 was achieved through solvent exchange from
DCM to THF (criteria: ≤5% v/v DCM), concentration to a ~20 wt% solution, followed by addition of methanolic
NaOMe (30 wt%; 1.1 equiv vs. 29 assay) at RT.
ACS Paragon Plus Environment
18

Page 19 of 41
Organic Process Research & Development
1
2
3
4
5
6
Figure 5. Filtration rate vs. solvent composition (vol% MeOH in THF) during isolation of Na Salt 31
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
While this procedure routinely provided desired 31, we experienced issues with very slow desaturation times
(>12 h), small particle size, and intractably poor filtration rates for the resulting slurry (flux rate ~7 L/min/m2,
Figure 5), which was not helped by prior seeding. Investigating the root cause for the inhibition of crystal growth,
we noticed that 31 was virtually insoluble in dry THF (~0.6 mg/mL, or ~1,600 L/kg). We reasoned that crystal
growth under these conditions would be kinetically hindered and may be helped by addition of strong coꢀsolvents.
Indeed, a dramatic increase in solubility was observed upon addition of even small proportions of methanol
and/or water (Figure 6). Using this data, a charge of 15−25 vol% MeOH to the reaction mass prior to addition of
NaOMe solution led to significantly improved filtration rates due to increase in crystal growth and particle size
(Figure 5). To ensure robustness on scale, we opted for a 25 vol% MeOH charge. Product losses to the mother
liquor were minimized by controlling the moisture content of the medium (KF<0.1 wt%). Using these parameters
in our campaign, a total of 124 kg of highꢀquality aminoacid Naꢀsalt 31 (≥99.6% diastereomeric purity) was
produced in 66% yield over two steps, starting from of 167 kg of ketone 22.
ACS Paragon Plus Environment
19

Organic Process Research & Development
Page 20 of 41
1
2
3
4
5
6
Figure 6. Solubility of Naꢀsalt 31 in THF/MeOH/water at 25 °C
100
90
80
70
60
50
40
30
20
10
0
7
8
9
0% Water
1% Water
1.5% Water
2% Water
2.5% Water
3% Water
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
0
5
10
15
20
25
30
MeOH (vol%)
Curtius rearrangement of the γ-aminoacid. To complete the assembly of the TACH core towards 1, we
intended to introduce the remaining C−N bond through direct Curtius rearrangement of the carboxylate function
at C2. From a safety perspective, we opted for use of the thermally stable DPPA as reagent of choice: when
added slowly to a preꢀheated mass of the substrate held above the thermal trigger point for the Curtius
rearrangement, DPPA provides an opportunity to combine the steps of carbonyl activation, azidation, and thermal
rearrangement to the corresponding isocyanate into a single operation.35 With this design, the potentially
hazardous acylazide36 intermediate would be consumed rapidly inꢀsitu as it forms, thus avoiding accumulation.
In practice, successful plantꢀscale application of the Curtius protocol to substrates 29/31 required us to develop
a detailed mechanistic understanding of the overall sequence. Several competing reaction pathways needed to be
controlled.
Our initial goal for the Curtius cascade was the direct introduction of the requisite acetamide at C2. Using free
acid 29, trapping of the intermediate isocyanate 3437 with HOAc was attempted (Scheme 9, Eq. 1).38 However,
using established methods (HOAc/Ac2O/DPPA),39 dimeric urea 36 was obtained as the major component with
little or none of the desired acetamide 35 present.
A switch from HOAc to the less acidic tꢀBuOH as nucleophile initially provided no apparent advantage. Even
under rigorously anhydrous conditions (NEt3, PhMe, KF<150 ppm), virtually none of the desired NꢀBoc
carbamate 37 (Scheme 9, Eq. 2) was observed. Instead, a new unexpected byꢀproduct predominated: carbamoyl
phosphate 39,40 admixed with dimeric urea 36.
ACS Paragon Plus Environment
20

Page 21 of 41
Organic Process Research & Development
1
2
3
4
5
6
Scheme 9. Formation of urea dimer 36 and carbamoyl phosphate 39 in anhydrous media
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
The formation of urea dimer 36 in an anhydrous medium required explanation since it implied intermediacy of
free amine 42. While not observed directly, any 42 formed inꢀsitu would react with transient isocyanate 34 (or,
alternatively, 39) to produce urea dimer 36. To rationalize the intermediacy of free amine 42 in the absence of
water, a more detailed analysis was required. We hypothesized that 42 could emerge from nucleophilic attack of
diphenylphosphate 38 (liberated from DPPA azidation of carboxylate 29) at phosphorus of carbamoyl phosphate
39, to produce phosphonic anhydride 41, a sequence of events previously described by Cramer et al.41 In this
scenario, decarboxylation of 39 would act as a key driving force for this cascade (Scheme 9).
While carbamoyl phosphates are known,42,43 few accounts exist of their formation through direct action of
phosphates on isocyanates.44 Indeed, a control experiment with cyclohexane carboxylic acid as substrate showed
no evidence of carbamoyl phosphates. Therefore, to explain the formation of 39 from 29, a more detailed
rationale was required. Based on kinetic work by Schwetlick et al.,45 weak nucleophiles would be expected to
react with isocyanates under general base catalysis via an associative proton transfer mechanism. Specifically, in
the case of the conformationally flexible isocyanate 34, we rationalize its unusually high reactivity towards
diphenylphosphate 38 (weak nucleophile) through internal base catalysis by the impending tertꢀamine at C4, i.e.,
via a cyclic diaxial transition state such as 40.
If operative, termination of the cascade from 39 to 42 would require proton transfer from the strongest donor
+
present, (i.e., HNEt3 ). In principle, an attack of 38 on 39 should be prevented if 39 either could not form or
remained deprotonated at nitrogen. We posited that the simplest strategy to curtail accumulation of 39 (and
ACS Paragon Plus Environment
21

Organic Process Research & Development
Page 22 of 41
1
2
3
4
5
subsequent undesired sideꢀreactions) would be to remove the carboxylate proton from the initial aminoacid
6
7
substrate 29, e.g., by switching to an alkali salt.
8
9
Scheme 10. Curtius Rearrangement of γꢀAminoacid NaꢀSalt 31
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
Indeed, reacting sodium salt 31 with DPPA in the absence of the amine base led to a clean reaction profile,
yielding predominantly isocyanate 34 with little or no dimeric urea 36.
We therefore opted to prepare acetamide 35 via a twoꢀstage process (Scheme 10), i.e., through initial trapping
of isocyanate 34 with tꢀBuOH as NꢀBocꢀcarbamate 37, followed by Nꢀdeprotection (MSA) to 42 and subsequent
Nꢀacetylation (TEA/Ac2O). In the laboratory, this procedure proved successful at first, producing crystalline 35 in
acceptable yield (50ꢀ60%, from 31).
However, upon further examination, the protocol revealed several inconsistencies. First, we found that some
laboratory batches of 31 suffered complete (nonꢀhazardous) decomposition upon addition of DPPA. The root
cause was traced to the presence of residual free alkali base, presumably from a base overcharge during
preparation of Naꢀsalt, whereby residual active NaOMe (and/or NaOH) would remain trapped in the isolated 31.
Strong alkali bases are known to catalyze the oligomerization of isocyanates.46 To circumvent this issue, we
added a small amount of diphenylphosphoric acid 38 to the batch (~7 mol%).47 Indeed, using this modification,
previously failing lots of 31 now performed reliably, comparable to reference batches that did not require addition
of the acid.
Secondly, successful preparation of 37 hinged on a rigorously dry reaction medium to further control levels of
dimeric urea 36. To address this issue, we implemented azeotropic distillative drying of each tꢀBuOH batch prior
to use in the Curtius reaction. Using PhMe as coꢀdistillation solvent, the moistureꢀrich distillate return was
percolated through a bed of molecular sieves (MS 4Å) contained in a Nutsche filter, polishꢀfiltered (1 ꢁm pore
size), and reꢀcirculated back into the main reactor until drying criteria were met (KF < 150 ppm).48
In the finalized Curtius procedure, slow addition of DPPA to a preꢀheated mixture (75−82 °C) containing 31
and (PhO)2P(=O)OH 38 (7 mol%) in preꢀdried 85% tꢀBuOH/PhMe provided desired NꢀBoc 37 in excellent
ACS Paragon Plus Environment
22

Page 23 of 41
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
solution yield (85%ꢀ90%). Distillative solvent switch to PhMe, aqueous workꢀup (KH2PO4), and crystallization
(nꢀheptane) yielded solid 37, which was filtered and reꢀslurried in an agitated Nutsche (PhMe/nꢀheptane), then
deliquored with a stream of nitrogen. The wetcake49 was redissolved in warm PhMe (55−65 °C) and the NꢀBoc
group was removed by addition of MeSO3H (MSA) and iꢀPrOH,50 yielding 42 in solution. Upon completion of
the deprotection, crude 42 was acetylated inꢀsitu by addition of Ac2O (1.2 equiv) and NEt3 (5.5 equiv vs. 31) to
give 35. The mixture was quenched into water and AcOH (2.5 L/kg) was added to extract product 35 into the
aqueous phase. The organic layer was discarded and the aqueous phase was basified (NaOH). Addition of
MeTHF extracted 35 back into the organic phase. After solvent switch to IPA, the solution was used directly in
the subsequent NꢀCbz deprotection (Step 11). In our campaign, a total of 78.6 kg of 35 (71.5 % yield, from 31)
were obtained in IPA solution as a single isomer (>99.9% diastereomeric purity) and in high chemical purity
(99.7%).
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 11. Completion of Synthesis: N-CBz Deprotection of 35 and SNAr Coupling to Final API BMSꢀ
741672
Deprotection of triaminocyclohexane 35 and final SNAr coupling to BMS-741672 (1).
Hydrogenolytic cleavage of Nꢀbenzylcarbamate 35 with 10% Pd on carbon (1.1 mol%) in iꢀPrOH51 cleanly
provided free base 43 (Scheme 11). Inert filtration to remove the spent catalyst and addition of freshly titrated 5N
solution of HCl in iꢀPrOH52 provided crystalline bisꢀhydrochloride salt 43•2HCl. The product was isolated by
filtration and dried in vacuo to give 43•2HCl (61.8 kg; partial iꢀPrOH solvate) in 95% yield (i.e., 68% from 31).
In the final step of the synthesis, we found that uncatalyzed SNAr coupling of free base 43 with chloroꢀ
quinazoline 44 proceeded smoothly at ambient temperature in MeCN solution. The requisite 44 was prepared inꢀ
situ through reaction of commercially available quinazolone 4553 with oxalyl chloride in the presence of a small
amount of DMF (0.34 equiv.) as catalyst (Scheme 12). To preclude the occurrence of redox reactions with the
chlorinating agent, the use of tertꢀamine bases was avoided in this part of the process.54 Instead, we first prepared
Naꢀsalt 45a from 45 by addition of stoichiometric 25% NaOMe solution in MeOH.
ACS Paragon Plus Environment
23

Organic Process Research & Development
Page 24 of 41
1
2
3
4
5
6
Scheme 12. Chlorination of Quinazolone 45
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
After distillative removal of free MeOH through solvent switch to MeCN, doseꢀcontrolled addition of (COCl)2
to the mixture of 45a at 40 °C was performed at such a rate as to ensure adequate equipment venting for the
evolving process gases (CO, CO2). To avoid accumulation of dimeric 46,55 an excess of chlorinating agent (1.6
equiv. vs. 45) was required, which effectively sequestered adventitious 46 by inꢀsitu conversion to 44. Upon
completion of the chlorination, excess (COCl)2 was removed by distillation and the mass was neutralized through
quench into aqueous KH2PO4 solution.
SNAr coupling was performed by direct introduction of the rich organic phase containing the crude 44 into a
solution of 43•2HCl and excess DIPEA in MeCN at ambient temperature. Reactions typically completed within
7–10 h. A key property of the coupling was its high tolerance for water (up to 20 wt%), which made drying of the
solution containing 43 and 44 unnecessary.
Utilizing the differences in acidity between product 1 and quinazolines 44 and 45, an extractive workup was
designed to remove traces of the heterocycles: Upon acidification with 10% aqueous AcOH, excess 44 was
removed by extraction with DCM. MTBE was added and the acidic aqueous layer was basified with 2N LiOH
solution56 to remove adventitious 45.57 The rich MTBE phase was dried azeotropically (criteria: KF ≤ 0.1%).
Seeded crystallization with addition of nꢀheptane as antisolvent then yielded isomerically pure BMSꢀ741672 (1)
of the desired physical form in 76.8% yield (from 43•2HCl).
■ CONCLUSION
A concise enantioselective 12ꢀstep synthesis of stereochemically complex CCR2 antagonist BMSꢀ741672 is
presented. Key feature of the target is an allꢀcis 1,2,4ꢀtriaminocyclohexane core.
The route was based on a set of design criteria, all of which were met, allowing full control over the genesis of
each new stereocenter in excellent enantiomeric and diastereomeric purity (>99.5% each).
ACS Paragon Plus Environment
24

Page 25 of 41
Organic Process Research & Development
1
2
3
4
5
6
7
8
A Ptꢀcatalyzed reduction of a βꢀenaminoester, directed by (S)ꢀαꢀmethylbenzylamine as lowꢀcost chiral
template, and reductive amination of a 3,4ꢀcisꢀdisubstituted cyclohexanone with a secondary amine on a sulfided
Ptꢀcatalyst established the stereochemistry in the cyclohexane core. Controlled Curtius reaction of γꢀaminoacid
salt 31 to NꢀBoc carbamate 37 introduced the remaining nitrogen into the core. Elaboration to the Nꢀacetylamide
35, N’ꢀCbz deprotection, and uncatalyzed SNAr coupling with chloroꢀquinazoline 44 then completed the
synthesis. The route was used to manufacture 50 kg of target 1 in 9 % overall yield.
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
■ EXPERIMENTAL SECTION
All reagents and solvents were commercially available and used without further purification. Unless indicated
otherwise, all reactions were performed using plantꢀscale reactors in an atmosphere of nitrogen. All charges were
calculated in mol/mol or L/kg relative to the limiting input reagent for each step. 1H NMR spectra were recorded
with TMS as an internal standard. HPLC analyses were performed using Agilent 1100 systems. All reaction
yields were calculated as corrected for purity, unless otherwise noted. All isomeric ratios and reaction yields were
analyzed by HPLC. Analytical conditions are described below.
HPLC Conditions: Method A (16a): Waters Xterra MS C18, (4.6 i.d. × 150 mm; 3.5 ꢁm), Eluent: (a) 0.1%
NH4OH in MeCN/ water 95:5, pH 9, (b) 0.1% NH4OH in MeCN/water 5:95, pH 9, Flow rate: 1.0 mL/min,
Temperature: 22 ºC, Gradient: (b/a) 40/60 (0 min) − 40/60 (20 min) − 85/15 (40 min) − 85/15 (50 min), UV
detection at 210 nm, (epiꢀ16): 17.6, (16b): 20.2, (16a): 26.4, (epiꢀ16): 27.8, (15): 30.4; Method B (22): Waters
Xterra MS C18, (4.6 i.d. × 150 mm; 3.5 ꢁm), Eluent: (a) 10 mM NH4OAc in MeCN/ water 95:5, (b) 10 mM
NH4OAc in MeCN/water 5:95, (b) 10 mM NH4OAc in MeCN/water 95:5, Flow rate: 1.0 mL/min, Temperature:
ambient, Gradient: (b/a) 5/95 (0 min) − 30/70 (10 min) − 40/60 (18 min) − 95/5 (35 min), UV detection at 210
nm, (24): 13.6, (20): 14.4, (22): 16.6, (2ꢀepiꢀ22): 17.6, (21): 20.2, (2ꢀepiꢀ21): 20.9, (2ꢀepiꢀ19): 22.8, (19): 23.5,
(3’ꢀepiꢀ19): 23.7; Method C (26/29): YMC Proꢀpack C18 ODS (4.6 i.d. × 150 mm; 3 ꢁm), Eluent: (a) 0.05% TFA
in MeCN/water 95:5, (b) 0.05% TFA in MeCN/water 95:5, Flow rate: 1.0 mL/min, Temperature: ambient,
Gradient: (b/a) 10/90 (0 min) − 15/85 (15 min) − 95/5 (35 min), UV detection at 210 nm, (29) major (1S,2R,4R):
16.6, (27) minor (1S,2R,4S): 18.8 min, (26) (1S,2R,4R): 21.9 min; Method D (35): Waters Xterra MS C18, (4.6
i.d. × 150 mm; 3.5 ꢁm), Eluent: (a) 0.05% NH4OH and 0.025% TFA in MeCN/ water 95:5, (b) 0.1% NH4OH and
0.025% TFA in MeCN/water 5:95, Flow rate: 1.0 mL/min, Temperature: ambient, Gradient: (b/a) 15/85 (0 min) −
25/75 (10 min) − 15/75 (28 min) − 100/0 (38 min), UV detection at 210 nm, (2ꢀepiꢀ35): 24.2, (35): 27.9, (3’ꢀepiꢀ
35): 31.8.
Ethyl 8ꢀOxoꢀ1,4ꢀdioxaspiro[4.5]decaneꢀ7ꢀcarboxylate (13). To a 6 m3 reactor at 20 °C was charged dry THF
(2,090 kg, KF<150 ppm), 60 wt% NaH dispersion in mineral oil (131.4 kg; 3.28 kmol, 2.2 eq.), and CO(OEt)2
(351.9 kg, 1.79 kmol, 2.0 eq.), followed by a catalytic amount of absolute EtOH (3.4 kg, 0.074 kmol, 0.05 eq.).
ACS Paragon Plus Environment
25

Organic Process Research & Development
Page 26 of 41
1
2
3
4
5
6
7
8
9
The resulting suspension was warmed to 60 °C and a solution of 1,4ꢀcyclohexanedione monoethylene ketal (12,
232.6 kg, 1.49 kmol, limiting reagent) in THF (700 kg) was added over 2h. The resulting mixture was held at 60
°C for another 3 h, then cooled to 0 °C and quenched into a separate tank at 0 °C containing HOAc (273.1 kg,
4.55 kmol, 3.0 eq.) and water (1,721 kg) while maintaining the internal temp. in the receiver below +10 °C. The
phases were allowed to settle and the aq, layer was removed. To the organic phase was added nꢀheptane (510 kg),
followed by a 7.5 wt% aq. solution of NaHCO3 (1,162 kg). The aq. layer was removed and the organic phase was
concentrated in vacuo (~200 mm Hg) below 40 °C to a total volume of ~950 L. MeOH (922 kg) was added and
the mixture was concentrated in vacuo (~150 mm Hg) below 67 °C. Upon collecting the distillate (1,180 kg), the
resulting methanolic solution (529 kg) was cooled to RT. The bulk solution containing 13 (316 kg, 1.39 kmol,
~93% yield, GC) was used directly in the next step. For characterization, an aliquot of an equivalent solution
obtained in the laboratory was evaporated and purified by flash chromatography on silica gel (nꢀheptane/EtOAc
4:1) to give 13 as a white crystalline solid after evaporation of solvents. Mp 50.5 ºC. 1H NMR (400 MHz, CDCl3)
δ 12.24 (s, 1H), 4.20 (q, J=7.2 Hz, 2H), 4.03ꢀ3.98 (m, 4H), 4.05ꢀ3.96 (m, 1H), 2.54ꢀ2.45 (m, 4H), 1.84 (t, J=6.8
Hz, 2H), 1.29 (t, J=7.3 Hz, 3H). 13C NMR (101 MHz, CDCl3) δ 172.0, 170.9, 107.2, 95.2, 64.3, 60.3, 32.6, 30.2,
27.8, 14.2. HRꢀESI(pos)ꢀMS: calcd for C11H17O5 229.1071 [M+H]+, found 229.1074. IR (KBr): ν=2983 (m),
2903 (m), 1742 (m), 1655 (s), 16.13 (m), 1290 (s), 1230 (s), 1197 (s), 1060(s), 852 (m), 831 (m). Anal. Calcd for
C11H16O5: C, 57.88; H, 7.07. Found: C, 57.89; H, 7.00.
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
Ethyl (S)ꢀ8ꢀ(1ꢀphenylethylamino)ꢀ1,4ꢀdioxaspiro[4.5]decꢀ7ꢀeneꢀ7ꢀcarboxylate (15). To the methanolic
solution (~670 L) of ketoester 13 from the previous acylation step was charged a catalytic amount of HOAc (3.6
kg, 0.060 kmol, 0.04 eq. vs. 12) at RT, followed by (S)ꢀαꢀmethylbenzylamine (14, 216.3 kg, 1.78 kmol, 1.20 eq.
vs. 12, 99.5% e.r.). The resulting homogeneous mass was warmed to 40 °C for 8h. The reaction was analyzed
for completion (criteria: <2% 13 (GC), then cooled to RT. MeOH (440 kg) and MeCN (440 kg) were charged;
agitation was halted to allow the mineral oil to separate. The oil layer (~40 kg) was removed and the lower phase
was cooled to 0 °C. Seed crystals of 15 (4.0 kg) were charged. After 30 min, water (744 kg) was added slowly
over 2h. The resulting thick slurry was filtered and the cake was washed with MeOH/water 1:1 (v/v; 1,350 kg) at
0 °C, then dried at 45 °C/50 mm Hg for 24h, yielding 15 (436.4 kg, 1.30 kmol, 88% yield, 2 steps) as a crystalline
white solid containing small amounts of residual oil (note slightly elev. carbon content in Elemental Analysis),
which did not affect downstream performance. Mp 86.7 ºC. 1H NMR (400 MHz, CDCl3) δ 9.42 (d, J=7.07 Hz,
1H), 7.28ꢀ7.36 (m, 2H), 7.19ꢀ7.28 (m, 3H), 4.62 (quin, J=6.88 Hz, 1H), 4.14 (q, J=7.07 Hz, 2H), 3.85ꢀ4.01 (m,
4H), 2.45ꢀ2.60 (m, 3H), 1.57ꢀ1.73 (m, 2H), 1.48 (d, J=6.82 Hz, 3H), 1.28 (t, J=7.07 Hz, 3H). 13C NMR (101
MHz, CDCl3) δ 170.2, 157.6, 145.4, 128.7 (2C), 126.8, 125.3 (2C), 107.4, 87.5, 64.4, 64.3, 58.8, 52.3, 33.8, 30.1,
25.4, 25.3, 14.5. IR (KBr): ν=3434 (w, br.); 2981 (w), 2948 (w), 2927 (w), 1655 (s), 1600 (s); 1452 (m), 1234 (s),
1207 (s), 1121 (m), 1054 (m), 845 (w), 698 (s). HRꢀESI(pos)ꢀMS: calcd for C19H26NO4 332.1856 [M+H]+, found
ACS Paragon Plus Environment
26

Page 27 of 41
Organic Process Research & Development
1
2
3
4
5
6
332.1857. [α]20 + 346.8 (c 1.0, CHCl3). Anal. Calcd for C19H25NO4: C,68.86; H,7.60; N, 4.22. Found: C, 69.64;
D
H, 7.95; N, 4.12.
7
8
9
Ethyl (7R,8S)ꢀ8ꢀ((S)ꢀ1ꢀphenylethylamino)ꢀ1,4ꢀdioxaspiro[4.5]decꢀaneꢀ7ꢀcarboxylate 4ꢀtoluene sulfonate
(16a). To a 6 m3 hydrogenator was charged an inerted slurry of 15 (282.5 kg, 0.85 kmol) in EtOH (990 kg). From
a separate, fully inerted tank was added an agitated slurry of Pt/C catalyst (32 kg, 0.0048 eq., 5 wt% Pt on dry
basis, 50% wet, JM Type 5R128M) in iꢀPrOAc (424 kg). The tank was flushed with EtOH (565 kg) and rinses
were transferred to the hydrogenator. Under full inertion, HOAc (118 kg, 1.96 kmol, 2.3 eq.) was pumped into the
hydrogenator, followed by EtOH (141 kg) via line flush. The hydrogenator was purged at RT with pressurized
nitrogen (30 psig, 3x), followed by hydrogen gas (75 psig, 3x). The reaction mass was warmed to 35 °C and
allowed to react for ~10h. The reaction was monitored for completion (criteria: ≤3% 15), then cooled to RT.
Hydrogen was removed and the reactor was inerted with nitrogen (30 psig, 3x). The spent catalyst was removed
by inert filtration through a bag filter, followed by an inꢀline polish filter (0.5 ꢁm pore size). The filters were
rinsed with iꢀPrOAc (860 kg) and combined filtrates were collected in a clean tank. A solution assay showed
~0.83 eq. of desired isomer 16a. EtOH was removed by concentration in vacuo at ≤40 °C to ~450 L, and solvent
exchange to iꢀPrOAc (Total iꢀPrOAc use: 7,490 kg). Criteria: ≤0.5% EtOH (GC). Anhydrous 2ꢀMe THF (1,420
kg) was added, followed by slow addition a solution of TsOH•H2O (143 kg, 1.0 eq. vs. soln. assay of 16a) in 2ꢀ
MeTHF (622 kg) over 2h. The resulting slurry was stirred at RT for 10h, filtered, and the cake was rinsed with 2ꢀ
MeTHF (995 kg), then dried in vacuo at ≤40 °C to give 16a (305.2 kg, 0.60 kmol, 71% yield) as a crystalline,
white solid (99.0 % diastereomeric purity). Mp 184.0 ºC. 1H NMR (400 MHz, CDCl3) δ 9.16 (br. s., 1H), 8.90
(br. s., 1H), 7.81 (d, J=8.08 Hz, 2H), 7.50ꢀ7.58 (m, 2H), 7.34ꢀ7.43 (m, 3H), 7.18 (d, J=8.08 Hz, 2H), 4.29ꢀ4.39
(m, 1H), 4.09ꢀ4.28 (m, 2H), 3.91ꢀ3.98 (m, 1H), 3.82ꢀ3.91 (m, 2H), 3.72ꢀ3.82 (m, 1H), 3.39 (d, J=3.54 Hz, 1H),
3.32 (dd, J=4.80, 10.61 Hz, 1H), 2.40 (td, J=3.03, 14.15 Hz, 1H), 2.35 (s, 3H), 1.98ꢀ2.16 (m, 2H), 1.78 (d, J=6.82
Hz, 3H), 1.69ꢀ1.75 (m, 1H), 1.49ꢀ1.67 (m, 2H), 1.27 (t, J=7.20 Hz, 3H). 13C NMR (101 MHz, CDCl3) δ 173.9,
142.6, 139.6, 136.2, 129.5 (2C), 129.4, 128.6 (2C), 127.2 (2C), 125.8 (2C), 106.2, 64.7, 64.2, 61.6, 56.7, 54.5,
38.7, 34.2, 33.0, 23.1, 21.2, 19.8, 13.8. IR (KBr): ν=3435 (w, br.), 2983 (m), 1737 (m), 1613 (w), 1233 (s), 1158
(s), 1032, (m), 1009 (m), 682 m). HRꢀESI(pos)ꢀMS: calcd for C19H28NO4 334.2013 [M+H]+, found 334.2013.
[α]20D +9.80 (c 1.0, CHCl3). Anal. Calcd. for C26H35NO7S: C, 61.76; H, 6.98; N, 2.77; S, 6.34. Found: C, 61.86;
H, 7.15; N, 2.80; S, 6.26.
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
Ethyl (7R,8S) 8ꢀaminoꢀ1,4ꢀdioxaspiro[4.5]decaneꢀ7ꢀcarboxylate 4ꢀtoluene sulfonate (17a). To a 6 m3
hydrogenator was charged an inerted solution of 16a (436.7 kg; 0.86 kmol) in EtOH (1,430 kg). From a separate,
fully inerted tank was added an agitated slurry of Pd/C catalyst (17.4 kg, 0.0095 eq., 10 wt% Pd on dry basis, 50%
wet, JohnsonꢀMatthey Type A501023ꢀ10) in iꢀPrOAc (1,000 kg). The tank was flushed with iꢀPrOAc (520 kg)
and rinses were transferred to the hydrogenator, which was subsequently purged at RT with nitrogen (30 psig,
ACS Paragon Plus Environment
27

Organic Process Research & Development
Page 28 of 41
1
2
3
4
5
3x), followed by hydrogen gas (45 psig, 3x). The reaction mass was warmed to 40 °C and allowed to react for
6
7
8
9
~5h at high agitation. The reaction was monitored for completion (criteria: ≤0.1% XIV vs. PhEt), then cooled to
RT. Hydrogen was removed and the reactor was inerted with nitrogen (30 psig, 3x). The spent catalyst was
removed by inert filtration through a bag filter, followed by an inꢀline polish filter (0.5ꢁm pore size). The filters
were rinsed with iꢀPrOAc (1,520 kg) and combined filtrates were collected in a clean tank. EtOH was removed by
concentration in vacuo (~200 mm Hg) to ~620 L at 40°ꢀ60 °C, solvent exchange to iꢀPrOAc, concentration to
~600L, and reꢀsupply of iꢀPrOAc (Total iꢀPrOAc use: 5,310 kg). Criteria: ≤0.5% EtOH (GC). The resulting
slurry was heated to 80 °C to dissolve solids and cooled to 65 °C. Seed crystals of 17a (1.7 kg) were charged.
The resulting slurry was cooled to 50 °C over 2h, followed by slow ramping to 20 °C over 4h. The slurry was
stirred at RT for 10h, filtered, and the cake was rinsed with iꢀPrOAc (390 kg), followed by drying in vacuo at
35 °C to give 17a (292.2 kg, 0.73 kmol, 84.3% yield) as a crystalline, white solid. Mp 120.2 ºC. 1H NMR (400
MHz, CDCl3) δ 7.96 (br. s., 3H), 7.76 (d, J=8.08 Hz, 2H), 7.15 (d, J=8.34 Hz, 2H), 4.11 (m, 2H), 3.75ꢀ3.98 (m,
4H), 3.41ꢀ3.51 (m, 1H), 3.15 (q, J=4.29 Hz, 1H), 2.35 (s, 3H), 2.26ꢀ2.33 (m, 1H), 2.05ꢀ2.18 (m, 1H), 1.97 (td,
J=4.39, 8.65 Hz, 1H), 1.63ꢀ1.78 (m, 2H), 1.50ꢀ1.62 (m, 1H), 1.21 (t, J=7.20 Hz, 3H). 13C NMR (101 MHz,
CDCl3) δ 173.0, 141.7, 140.1, 128.8 (2C), 125.9 (2C), 106.6, 64.5, 64.2, 61.0, 49.1, 40.1, 33.8, 32.3, 24.6, 21.2,
13.8. IR (KBr): ν=3436 (m, br.), 2981 (m), 1731 (m), 1606 (w), 1233 (s), 1189 (s), 1036, (s), 1012 (m), 916 (m),
683 (m). HRꢀESI(pos)ꢀMS: calcd for C11H20NO4 230.1387 [M+H]+, found 230.1386. [α]20D +16.0 (c 1.0, CHCl3).
Anal. Calcd. for C18H27NO7S: C, 53.85; H, 6.78; N, 3.49; S, 7.99. Found: C, 54.00; H, 6.90; N, 3.49; S, 7.92.
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
Ethyl (7R,8S)ꢀ8ꢀ((S)ꢀ2ꢀ(benzyloxycarbonylamino)ꢀ4ꢀ(methylthio)ꢀbutanamido)ꢀ1,4ꢀdioxaspiro[4.5] decaneꢀ
7ꢀcarboxylate (19). To a 4 m3 reactor were charged 17a (131 kg; 0.33 kmol), MeCN (227 kg, 290 L, 2.2 L/kg vs.
17a), EDAC·hydrochloride (68.8 kg, 0.36 kmol, 1.10 eq.), HOBt·monohydrate (49.8 kg, 0.33 kmol, 1.0 eq.), and
LꢀNꢀCbzꢀMetꢀOH (18, 97.3 kg, 0.34 kmol, 1.05 eq., 99.5% e.r.). The homogeneous, colorless solution was stirred
at RT and DIPEA (93 kg, 0.72 kmol, 2.20 eq.) was added slowly over 1h. During the addition, the resulting
exotherm was controlled to ≤30 °C. After 2 h, the reaction was monitored for completion (Criteria: ≤4 Area% of
18 vs. 19). The mixture was diluted with EtOAc (870 kg, 970 L, 7.4 L/kg vs. 17a), followed by washing with aq.
H2SO4 (4.8 wt%, 2x144 kg). The combined aq. layers (pH ~4) were removed and the organic phase was washed
with aq. KHCO3 (20 wt%, 2x234 kg) and water (2x640 kg). The aq. layers were removed (pH 8.5ꢀ9) and the
organic phase (1,120 kg) was assayed for 19 (13.8 wt%, 155.4 kg, 0.31 kmol, 96.3% soln. yield). The mass was
further concentrated to give 444 kg of solution. The bulk solution was used asꢀis in the next step. For
characterization, an aliquot of an equivalent solution obtained in the laboratory was evaporated without further
purification to give the product as a light yellow crystalline solid. Mp 110.2 ºC. 1H NMR (400 MHz, CDCl3) δ
7.26ꢀ7.41 (m, 5H), 7.05 (d, J=8.59 Hz, 1H), 5.71 (d, J=7.83 Hz, 1H), 5.10 (s, 2H), 4.25ꢀ4.39 (m, 2H), 4.07ꢀ4.20
(m, 2H), 3.81ꢀ4.00 (m, 4H), 2.84 (br. s., 1H), 2.43ꢀ2.60 (m, 2H), 2.18 (dd, J=6.82, 13.89 Hz, 1H), 2.08 (s, 3H),
1.98ꢀ2.06 (m, 2H), 1.92 (td, J=7.17, 14.21 Hz, 1H), 1.61ꢀ1.87 (m, 4H), 1.26 (t, J=7.07 Hz, 3H). 13C NMR (101
ACS Paragon Plus Environment
28

Page 29 of 41
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
MHz, CDCl3) δ 173.0, 170.6, 155.9, 136.1, 128.4 (2C), 127.9 (2C), 128.0, 107.2, 66.9, 64.4, 64.3, 60.5, 53.9,
46.4, 42.8, 33.9, 32.1, 31.8, 29.8, 26.9, 15.0, 14.0. IR (KBr): ν=3380 (m, br.), 3266 (w), 2977 (w), 1720 (s), 1641
(s), 1546 (m), 1243 (s), 1044 (m), 752 (w), 697 (w). HRꢀESI(pos)ꢀMS: calcd for C24H35N2O7S 495.2159 [M+H]+,
found 495.2158. [α]20D +7.8 (c 1.0, CHCl3). Anal. Calcd. for C24H34N2O7S: C, 58.28; H, 6.93; N, 5.66; S, 6.48.
Found: C, 58.29; H, 6.69; N, 5.67; S, 6.41.
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
((S)ꢀ3ꢀ(Benzyloxycarbonylamino)ꢀ4ꢀ((7R,8S)ꢀ7ꢀ(ethoxycarbonyl)ꢀ1,4ꢀdioxaspiro[4.5]decanꢀ8ꢀylꢀamino)ꢀ4ꢀ
oxobutyl)dimethylsulfonium iodide (20). To a 4 m3 reactor was charged the EtOAc solution of 19 (444 kg) from
the previous step, followed by a line rinse with EtOAc (30 kg). To capture and neutralize volatile MeI fumes, the
equipment train vent manifold was connected to two consecutive inline scrubbers, each containing water (600
kg), DMAP (7.2 kg) and morpholine (300 kg). Neat MeI (694 kg, 4.89 kmol, 15.5 eq. vs. 19 assay) was charged
below 25 °C over 30 min and the reactor was closed off. After 20 h, the reaction was monitored for completion
(Criteria: ≤1 Area% XVI). MTBE (606 kg) was charged and the resulting slurry was filtered. The cake was
washed with MTBE (4x300 kg), sampled and analyzed for residual MeI (Criteria: ≤ 0.005% v/v), then dried
inside the filter using a constant stream of nitrogen to yield 194 kg of 20 (0.31 kmol, 97%). For characterization,
an aliquot of an equivalent solid obtained in the laboratory was dried at RT using a stream of nitrogen to give the
sulfonium iodide salt as a white, crystalline solid. Mp 123.2 ºC. 1H NMR (400 MHz, CDCl3) δ 7.75 (d, J=8.59
Hz, 1H), 7.24ꢀ7.41 (m, 5H), 6.43 (d, J=7.07 Hz, 1H), 5.04ꢀ5.15 (m, 2H), 4.71 (s, 1H), 4.43 (br. s., 1H), 3.82ꢀ4.15
(m, 6H), 3.69ꢀ3.82 (m, 1H), 3.54ꢀ3.68 (m, 1H), 3.16ꢀ3.28 (m, 3H), 3.13 (s, 3H), 2.81ꢀ2.91 (m, 1H), 2.41 (s, 1H),
2.12ꢀ2.30 (m, 2H), 2.00 (d, J=9.85 Hz, 1H), 1.67ꢀ1.93 (m, 3H), 1.49ꢀ1.66 (m, 1H), 1.24 (t, J=7.07 Hz, 3H). 13C
NMR (101 MHz, CDCl3) δ 173.0, 169.8, 156.4, 136.2, 128.4 (2C), 128.1, 127.8 (2C), 107.8, 67.0, 64.3, 64.2,
60.6, 52.6, 45.8, 43.6, 40.0, 32.8, 30.7, 28.4, 27.7, 26.6, 25.9, 14.1. IR (KBr): ν=3431 (m, br.), 3264 (m), 2940
(w), 1721 (s), 1670 (s), 1529 (m), 1247 (m), 1090 (m), 1048 (m), 747 (w), 699 (w). HRꢀESI(pos)ꢀMS: calcd for
C25H37N2O7S 509.2316 [M+H]+, found 509.2314. [α]20D+16.2 (c 1.0, CHCl3). Anal. Calcd. for C25H37IN2O7S: C,
47.17; H, 5.85; I, 19.93; N, 4.40; S, 5.03. Found: C, 47.04; H, 5.68; I, 20.30; N, 4.43; S, 5.31.
Ethyl (7R,8S)ꢀ8ꢀ((S)ꢀ3ꢀ(benzyloxycarbonylamino)ꢀ2ꢀoxopyrrolidinꢀ1ꢀyl)ꢀ1,4ꢀdioxaspiro[4.5]decane ꢀ7ꢀ
carboxylate (21). A 4 m3 reactor was equipped with a highꢀcapacity pumpꢀaround loop with a centrifugal pump
capable of operating at a min. pump rate of 300 L/min. A constant stream of nitrogen was introduced into the
reactor subꢀsurface and entrained volatiles were vented and neutralized through a twoꢀstage scrubber, each
containing aq. bleach and caustic solutions, maintained at pH 10ꢀ14 through addition of more caustic as needed.
Nitrogenꢀdried 20 (~104.7 kg, 0.17 kmol) was charged to the reactor and dissolved in DMSO (1,100 kg; 10 L/kg),
thermostated at 21ꢀ23 °C (to avoid thermal demethylation of 20). The pumpꢀaround loop was activated and
Cs2CO3 (59.0 kg, 0.18 kmol, 1.1 eq. vs. 20) was charged to the reactor. Throughout the reaction, the suspension
was maintained at 21−23 °C. After 7h, the reaction was monitored for conversion (Criteria: ≤5 Area% 20). Upon
completion, the contents of the loop were emptied into the reactor and the circulation pump was switched off.
ACS Paragon Plus Environment
29