Process Development and Scale-up of Fully Synthetic Tetracycline TP-2758: A Potent Antibacterial Agent with Excellent Oral Bioavailability

Article abstract of DOI:10.1021/acs.oprd.5b00404

Process research and development of the fully synthetic broad spectrum tetracycline TP-2758, with a chiral pyrrolidine side chain at the C-8 position, is described. The process utilizes two key intermediates, 7 and 10, in a convergent approach that allows for manufacturing of sufficient quantities of API to supply preclinical and early clinical development. The pyrrolidine moiety was introduced into the left-hand piece (LHP) 10 with high enantioselectivity using Ellman's sulfinamide chemistry, and the absolute configuration was confirmed by X-ray crystal structure analysis.

Full text of DOI:10.1021/acs.oprd.5b00404

Subscriber access provided by UNIV OF NEBRASKA - LINCOLN
Full Paper
Process Development and Scale-up of Fully Synthetic Tetracycline
TP-2758: A Potent Antibacterial Agent with Excellent Oral Bioavailability
Wu-Yan Zhang, Cuixiang Sun, Diana Hunt, Minsheng He, Yonghong Deng, Zhijian Zhu, Chi-Li Chen,
Christopher Katz, John Niu, Philip Hogan, Xiao-Yi Xiao, Nicholas Dunwoody, and Magnus Ronn
Org. Process Res. Dev., Just Accepted Manuscript • DOI: 10.1021/acs.oprd.5b00404 • Publication Date (Web): 28 Dec 2015
Downloaded from http://pubs.acs.org on December 29, 2015
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 43
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
Process Development and Scaleꢀup of Fully
Synthetic Tetracycline TPꢀ2758: A Potent
Antibacterial Agent with Excellent Oral
Bioavailability
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
Wu-Yan Zhang*, Cuixiang Sun, Diana Hunt, Minsheng He, Yonghong Deng, Zhijian Zhu, Chi-
Li Chen, Christopher E. Katz, John Niu, Philip C. Hogan, Xiao-Yi Xiao, Nicholas Dunwoody,
and Magnus Ronn
Tetraphase Pharmaceuticals Inc, 480 Arsenal Street, Suite 110, Watertown, Massachusetts,
02472, U.S.A.
ACS Paragon Plus Environment
1

Organic Process Research & Development
Page 2 of 43
1
2
3
4
Table of Contents Grpahic
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
ACS Paragon Plus Environment
2

Page 3 of 43
Organic Process Research & Development
1
2
3
4
Abstract
5
6
7
8
9
Process research and development of the fully synthetic broad spectrum tetracycline TPꢀ2758,
with a chiral pyrrolidine side chain at the Cꢀ8 position, is described. The process utilizes two key
intermediates, 7 and 10, in a convergent approach that allows for manufacturing of sufficient
quantities of API to supply preꢀclinical and early clinical development. The pyrrolidine moiety
was introduced into the leftꢀhand piece (LHP) 10 with high enantioselectivity using Ellman’s
sulfinamide chemistry, and the absolute configuration was confirmed by xꢀray crystal structure
analysis.
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
Key words: tetracycline, chiral pyrrolidine, antibacterial, MichaelꢀDieckmann, regioselective
methylation
ACS Paragon Plus Environment
3

Organic Process Research & Development
Page 4 of 43
1
2
3
4
Introduction
5
6
7
8
9
Using our fully synthetic tetracycline technology,1 Tetraphase has synthesized more than 3000
novel, highly active tetracycline analogs, including our lead clinical candidate eravacycline (1),
which is currently undergoing phase 3 clinical studies.2 TPꢀ2758 (2) is another new tetracycline
analog with novel substitutions at Cꢀ7 and Cꢀ8, which are difficult to install by semisynthetic
methodologies. TPꢀ2758 demonstrated potent activity against tetracyclineꢀresistant isolates of
Gramꢀnegative bacteria including Enterobacteriaceae strains containing extendedꢀspectrum βꢀ
lactamases, Proteus and Acinetobacter, but not Pseudomonas spp. TPꢀ2758 also retains
significant Gramꢀpositive coverage. TPꢀ2758 exhibited excellent oral and IV efficacy in
multiple murine infection models, including pyelonephritis models with either uropathogenic E.
coli or ESBLꢀproducing K. pneumoniae and lung infection models with either tetracyclineꢀ
resistant S. aureus or S. pneumoniae. The favorable IV/oral pharmacokinetic profile of TPꢀ2758
further supports its candidacy for development as an IV/oral antibiotic for the empiric treatment
of serious infections with a high incidence of Gramꢀnegative pathogens.3 The CMC group was
tasked to supply hundreds of grams of TPꢀ2758 for toxicity studies and phase I clinical trials.
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. Tetracyclines developed at Tetraphase
ACS Paragon Plus Environment
4

Page 5 of 43
Organic Process Research & Development
1
2
3
4
Results and Discussion
5
6
7
8
9
The discovery route used to prepare TPꢀ2758 as well as a number of other analogs is shown in
Scheme 1.4 The racemic left hand piece (LHP) 6 was synthesized in 13 steps from commercially
available but expensive starting material 3, including one reaction to install the methoxy group
(from 4 to 5) with modest yield (41%), even at gram scale. The product of the Michaelꢀ
Dieckmann reaction (8), which is obtained as a ~1:1 diastereomeric mixture, required extensive
flash chromatography to isolate the desired diastereomer, a process considered costly at large
scale. Large scale manufacturing required a more robust and economical approach.
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 1. Discovery Route to TP-2758.
For an efficient synthesis, we planned to prepare the LHP as a fully built single enantiomer 10
(Scheme 2), which should undergo Michael Dieckmann reaction with enone 7,5 followed by
deprotection steps to produce desired tetracycline 2. Stereoselective introduction of the chiral
pyrrolidine moiety within LHP 10 is envisioned from bromide 9.
ACS Paragon Plus Environment
5

Organic Process Research & Development
Page 6 of 43
1
2
3
4
Scheme 2. Retrosynthetic Analysis of TP-2758.
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
Research and Manufacture of Intermediate 9.
To manufacture material for GLP toxicity studies and a phase 1 clinical trial, several routes for
intermediate 9 were investigated simultaneously using commercially available starting materials.
The main goal for the compound synthesis was to make the route easily accessible and robust for
early manufacturing and to minimize the cost of goods. Three major preliminary process
research routes explored are described below.
Synthesis of Intermediate 9 Starting with 2,5-Dimethoxybenzoic Acid.
Outlined in Scheme 3 is one of the initial routes we investigated using 2,5ꢀdimethoxybenzoic
acid 11 as the starting material. Regioselective methylation of acid 11 was achieved by
modifying a procedure from the eravacycline program at Tetraphase.6 The key transformation in
this route is the regioselective bromination at the Cꢀ4 position. After extensive investigations, we
concluded that bromination of the Cꢀ4 position could only be achieved under the directing effect
of a 3ꢀhydroxy group. Owing to the electron withdrawing phenyl ester group, the Cꢀ6 methoxy
within compound 13 could be selectively demethylated but not the Cꢀ3 one. Thereby both
methoxy groups were cleaved with BBr3 to provide compound 14. Attempts to selectively
protect the Cꢀ6 hydroxyl group with methyl, acetyl, benzyl, and TBS groups respectively failed.
Finally, it was found that the Cꢀ3 hydroxyl group can be selectively protected using Boc2O to
give compound 15 along with ~7% bisꢀBoc byproduct and ~7% starting material. Protection of
ACS Paragon Plus Environment
6

Page 7 of 43
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
the Cꢀ6 hydroxyl group as a benzyl ether followed by Boc removal gave compound 16.
Regioselective bromination of compound 16 with bromine in the presence of NaOAc gave the
desired bromide 17, which was then methylated to provide intermediate 9. The protecting group
manipulations made this 8ꢀstep route laborious, and a few steps required chromatographic
purifications. In addition, the starting material, 2,5ꢀdimethoxybenzoic acid, is much more
expensive than the 1,4ꢀdimethoxybenzene in Scheme 5 (vide infra). The cost and length of the
route led to discontinuation of its optimization.
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. Synthesis of Intermediate 9 Starting with 2,5-Dimethoxybenzoic Acid.
Synthesis of Intermediate 9 Starting with 2,5-Dihydroxybenzoic Acid.
An approach outlined in Scheme 4 using commercially available 2,5ꢀdihydroxybenzoic acid 18
as starting material was explored simultaneously with the above route. The key step in this
approach is the regioselective methylation of compound 20 to produce compound 21, which was
achieved using 1 equivalent of MeLi followed by quenching of the resulting aryllithium with
iodomethane. Other reagents such as iꢀPrMgCl·LiCl or iꢀPrMgCl were not as regioselective as
ACS Paragon Plus Environment
7

Organic Process Research & Development
Page 8 of 43
1
2
3
4
5
6
7
8
9
MeLi. However, the formation of acetonide 19 suffered from low yield and low throughput while
the phenyl ester 23 formation required a large excess of TFA and TFAA, rendering this route
less attractive. Further optimization of these two steps are required if this route were to be used
for future production.
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 4. Synthesis of Intermediate 9 from 2,5-Dihydroxybenzoic Acid.
Research and Manufacturing of Intermediate 9 Starting with 1,4-Dimethoxybenzene.
Intermediate 9 was manufactured in kilogram scale according to the route shown in Scheme 5,
starting from cheap and readily available 1,4ꢀdimethoxybenzene. In this 6ꢀstep sequence, the
yields in all the steps are above 90% except for one bottleneck reaction, the methylation of
phenyl ester 27 to produce compound 28. The methylation reactions of the corresponding acid
(26), methyl and tertꢀbutyl ester substrates were also investigated. None of these proved to be a
better substrate than phenyl ester 27; therefore phenyl ester 27 was investigated thoroughly
(Table 1). For all experiments the anion was generated by addition of phenyl ester 27 to a freshly
prepared base solution at or below ‒70 °C. In general, the major issues causing low yield are low
ACS Paragon Plus Environment
8

Page 9 of 43
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
conversion (starting material was recovered) and the formation of byproduct amides. The amount
of phenol as a byproduct represents the total amount of various amide byproducts in the reaction;
thereby the quantity of phenol formed was measured as well as the starting material 27 to
evaluate the reaction. We first looked into reverse addition of the cold anion to MeI solution at
different temperatures (entry 1ꢀ4), and found that ‒40 °C is the optimal reaction temperature. On
the other hand, addition of MeI to the cold anion solution gave lower yield (entry 5). Additives
also play a key role in this reaction. It was known that the presence of a small amount of LiCl
could enhance the kinetic activity of LDA.7 In this case, a combination of TMEDA (1.1 eq) and
NEt3·HCl (0.03 eq) gave better results than either one or neither of these additives present
(entries 2 and 6ꢀ9). Use of THF as solvent proved critical too, as reaction in MeTHF gave much
lower yield (entry 10). To control the amide formation several other secondary amines with
varied steric hindrance were screened to make the lithium amides for the deprotonation (entries
12ꢀ16). When smaller diethylamine was used, amide formation dominated with no desired
methylation product (entry 13). Bulkier amines, diꢀsecꢀbutylamine and dicyclohexylamine,
emerged as the two most promising, with high conversion and lower levels of amide byproducts.
Both amines provided the desired product 28 in yields above 70% (entries 14, 15). However,
when the amines are too bulky (entry 12, 16), more starting material and phenol were observed.
In the case of tertꢀbutylꢀisoꢀpropylamine, undesired regioisomer with methylation at Cꢀ5 was
observed as well, indicating a weakened directing effect from the ester group (entry 16). While
dicyclohexylamine is desirable due to its low cost, its poor solubility in THF at low temperature
resulted in the selection of diꢀsecꢀbutylamine for this production cycle. Finally, a large excess of
MeI (6 eq) was necessary to achieve the best yield; a lower number of equivalents resulted in
lower yields (entries 8, 11 and entries 15, 17).
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
9

Organic Process Research & Development
Page 10 of 43
1
2
3
4
5
6
7
8
9
In this methylation step, it is critical to keep the anion cold and maintain good agitation during
the addition of anion into the MeI solution. This reaction typically gave above 70% yield of the
desired methylation product at 200 g scale. In an inꢀhouse kiloꢀlab campaign (~4 kg scale), a
lower yield (49%) was obtained presumably caused by poor mixing and temperature control
during the transfer of the 27ꢀanion into the MeI solution. The capacity and cylindrical shape of
the reactor (a 50 L reactor that was nearly full) and the solid formation, which had appeared to be
less significant at a 200 g scale, made agitation very inefficient. This mixing issue could likely be
addressed in future manufacturing using a lager vessel. Nevertheless, compound 28 could be
isolated by treatment with sodium hydroxide to selectively hydrolyze the starting material 27 to
acid 26 and a silica gel plug to remove the highly polar amide byproducts. The chemistry in all
other steps is high yielding and features convenient isolations if the optimized conditions are
used. For the dibromination step, it is critical to keep the temperature cold to prevent the
formation of tribromide byproduct. For the selective deꢀmethylation step, the stoichiometry of
BBr3 (1 equiv) is critical. Excess amount of BBr3 reagent leads to bisꢀdemethylation resulting in
a hydroquinone byproduct which could not be easily removed. For the benzylation step, a
solution of phenol 23 was added to a heated mixture of K2CO3, NaI and BnBr in NMP to
minimize the concentration of 23 in the reaction so as to avoid the formation of a homodimeric
byproduct. It is critical to use fine powder K2CO3 to facilitate the benzylation. Intermediate 9 has
been produced at 2 kg scale in house.
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
10

Page 11 of 43
Organic Process Research & Development
1
2
3
4
Scheme 5. Manufacture of Intermediate 9 from 1,4-Dimethoxybenzene
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
Table 1. Summary of Methylation Reactions
entry
Base
Additivesa Solvent Conditionsb
MeI
(equiv)
28/27/PhOH
Yield of
28
(isolated)
(1.05 eq)
(by 1H NMR)
1
2
3
4
5
6
7
8
9
LDA
LDA
LDA
LDA
LDA
LDA
LDA
LDA
LDA
TMEDA
TMEDA
TMEDA
TMEDA
TMEDA
THF
THF
THF
THF
THF
THF
THF
THF
THF
A, –60 °C
A, –40 °C
A, –20 °C
A, 0 °C
6
6
6
6
6
6
6
6
6
1:0.51:0.81
1:0.22:0.38
1:0.35:0.50
<45%
54%
49%
49%
B, –78 °C
A, –40 °C
B, –78 °C
A, –40 °C
A, –40 °C
1:0.37:0.48
1:0.24:0.59
Mostly SM 27
1:0.18:0.44
1:0.19:0.16
44%
<50%
MgCl2
Et3N·HCl
TMEDA/
Et3N·HCl
<50%
64%
ACS Paragon Plus Environment
11

Organic Process Research & Development
Page 12 of 43
1
2
3
4
5
6
7
8
9
10
11
LDA
LDA
TMEDA/
Et3N·HCl
TMEDA/
Et3N·HCl
TMEDA
2ꢀ
MeTHF
A, –40 °C
A, –40 °C
6
1:0.43:1.4
<35%
<40%
THF
1.5
1:0.63:0.87
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
12
13
14
LTMP
LiNEt2
THF
THF
THF
A, –40 °C
A, –40 °C
A, –40 °C
6
6
6
1:0.77:1.45
mostly amide
1:0.15:0.16
<30%
0%
LiNCy2 TMEDA/
Et3N·HCl
72%
15
LiNsBu2 TMEDA/
Et3N·HCl
THF
THF
A, –40 °C
B, –78 °C
6
6
1:0:0.25
73%
16 LiNtBuiP TMEDA/
1:1.63:1.55
<20%
r
Et3N·HCl
(0.57
undesired
methylated
regioisomer)
17
LiNsBu2 TMEDA/
Et3N·HCl
THF
A, –40 °C
3
1:0.05:0.37
<70%
a. For TMEDA or MgCl2, 1.1 equiv was added; for Et3N·HCl, 0.03 equiv was added. b. A:
The 27ꢀanion solution was added to a solution of MeI; B: MeI was added to the 27ꢀanion
solution.
Stereoselective Synthesis of LHP 10 from Intermediate 9.
Based on literature research, we envisioned that the chiral center could be introduced by
asymmetric reduction of cyclic imine,8 enamine or encarbamate,9 asymmetric Negishi
coupling,10 or induction by a chiral auxiliary (e.g., (S)ꢀtꢀbutylsulfinamide).11 After careful
evaluation of the literature information and a brief exploration, the Negishi coupling and
Ellman’s sulfinamide chemistry were selected as potential enantioselective manufacturing routes
ACS Paragon Plus Environment
12

Page 13 of 43
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
and were investigated simultaneously. At this stage of development, as LHP 10 was the cGMP
starting material, the ability to control purity (chiral and chemical) was our top priority.
Research of LHP 10 using Negishi Chemistry
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
Introduction of the chirality through an asymmetric Negishi coupling (Scheme 6)10 could
potentially provide a short and efficient sequence. Enantioselective deprotonation of NꢀBoc
s
pyrrolidine with BuLi in the presence of the natural (‒)ꢀsparteine followed by Negishi coupling
with arylbromide 9 via Li/Zn/Pd transmetalation provided the undesired enantiomer of
compound (R)ꢀ30 (consistent with the literature prediction) with >90% ee in 63% isolated yield,
along with some protoꢀdebromination and dimeric byꢀproducts. Thus, a known (S,S)ꢀdiamine
2912 was explored as the (+)ꢀsparteine surrogate to achieve the desired chirality. A number of
different reaction conditions were screened including solvents (ether, 2ꢀmethylꢀTHF, toluene,
NMP, THF, DME), ratio of reagents, Pd precursors and phosphorus ligands. It was found that: 1)
Excess Zn reagent is needed to consume all starting materials; 2) Reactions using ether and THF
are faster than those in DME and THF/PhMe media; 3) Pd(OAc)2 and PdCl2(COD) are superior
to other catalysts explored including PdCl2(MeCN)2, PdCl2(BnCN)2, Pd2dba3, and Pd(TFA)2; 4)
The typical Negishi coupling ligand P(tBu)3 is the best choice followed by Buchwald phosphine
ligands.13 Using the optimized conditions, the Negishi coupling using (S,S)ꢀdiamine 29 gave the
desired product 30 in 50% isolated yield with 93% ee, which is borderline acceptable at this
stage of development. Many attempts to improve the yield failed. This chemistry was not
pursued further in development not only because the crossꢀcoupling was not robust but also due
to the use of hazardous sBuLi in the route.
ACS Paragon Plus Environment
13

Organic Process Research & Development
Page 14 of 43
1
2
3
4
Scheme 6. Negishi Coupling Approach
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
Research and Manufacturing of LHP 10 Using Ellman’s Sulfinamide Chemistry
LHP 10 is currently manufactured from intermediate 9 using the 5ꢀstep sequence depicted in
Scheme 7. The chiral stereocenter was installed via Ellman’s asymmetric synthesis of 2ꢀ
substituted pyrrolidines using tertꢀbutylsulfinamide.11 Formylation of intermediate 9 by
i
treatment with PrMgCl·LiCl in DME followed by addition of DMF cleanly gave desired
aldehyde 31 with only ~3ꢀ4 mol% protoꢀdebromination byproduct, whereas the magnesium
bromide exchange was sluggish in THF. The crude formylation product was taken directly to the
Nꢀsulfinyl aldimine formation step, where it was heated with CuSO4 and (S)ꢀtertꢀ
butylsulfinamide in toluene at 45 °C to afford the desired aldimine 32 in good yield and purity
after recrystallization from MTBE and heptane. In our case Ellman’s standard Ti(OEt)4
condensation conditions14 suffered from a tedious workꢀup due to the emulsion caused by
titanium. Addition of freshly prepared Grignard reagent 33 to aldimine 32 at cryogenic
temperature generated the desired addition product 34 with high diastereoselectivity in high
yield. The absolute configuration of this compound was unambiguously determined by single
crystal Xꢀray structure (Figure 2). It is consistent with what was predicted when (S)ꢀtertꢀ
butylsulfinamide is used.14
ACS Paragon Plus Environment
14

Page 15 of 43
Organic Process Research & Development
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 2. Single crystal Xꢀray structure of compound 34
With the desired sulfinamide 34 in hand, conditions were optimized to effect acidꢀinduced
deprotections and subsequent intramolecular reductive amination to afford the chiral pyrrolidine
35. Using Ellman’s procedure (95:5 TFA:H2O; then Et3SiH), the pyrrolidine ring was formed in
a clean manner, but the benzyl protecting group was cleaved. In order to avoid an additional reꢀ
protection step, the reaction conditions were modified. Treatment of sulfonamide 34 with 1:1
TFA:H2O at ꢀ10 to 0°C overnight gave the cyclic imine, which was then reduced in one pot with
NaBH(OAc)3, providing the desired pyrrolidine 35 without any deꢀbenzylation byproduct
detected. Installation of the ethyl group to compound 35 to produce LHP 10 generally proceeded
in good yield but with a low percentage of several impurities. Polar amine LHP 10, a thick oil,
proved very difficult to purify via column chromatography. Salt formation and recrystallization
was briefly explored but proved unsuccessful. This route required significant development of the
isolation to enable a viable control strategy.
ACS Paragon Plus Environment
15

Organic Process Research & Development
Page 16 of 43
1
2
3
4
5
6
7
8
9
After substantial investigation, it was found that using methanol as reaction solvent is the key
to a clean reaction. Compared to reactions in commonly used solvents (e.g. DCM,
dichloroethane), the reaction in methanol required more reducing reagent (3 equiv.), but was
very clean when a cold solution of acetaldehyde in MeOH was added to the reaction mixture
containing 35 and NaBH(OAc)3 while maintaining the temperature at ‒20 °C. Later we found
this reaction could be performed at ambient temperature if employing picoline borane or pyridine
borane as reducing agent.15 High purity leftꢀhand piece 10 (99% ee) was obtained under the
optimal reaction conditions with only a trace of impurity at the baseline, easily removed via
purification on a silica gel plug. LHP 10 has been made in >700 g batches. The scale was limited
only by the reactor capacity (50 L).
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 7. Manufacture of LHP 10
iPrMgCl—LiCl
DME; DMF
OCH3
OCH3
(S)ꢀtBuSONH2
CuSO4
Toluene, 45 C
O
S
H
OCH3
Br
CH3
OHC
CH3
CH3
tBu
N
ꢀ10 C to 0 C
MTBE/heptane
recrystallization
COOPh
COOPh
CO2Ph
OBn
OBn
OBn
9
31
32
77% over
2 steps
O
O
MgBr
90%
33
THF, <ꢀ70 C
CH3CHO
NaBH(OAc)3
AcOH, CH3OH
ꢀ20 C
O
O
OCH3
OBn
OCH3
OBn
H
1. TFA, water
6 C
H
CH3
CH3
O
OCH3
CH3
N
N
H
S
tBu
N
2. NaBH(OAc)3
ꢀ20 C
CH3
CO2Ph
CO2Ph
H
90% over
2 steps
CO2Ph
OBn
10
35
34
In summary, complex LHP 10, containing one chiral center, was manufactured in a total of 11
robust steps starting from cheap and commercially available dimethoxybenzene.
ACS Paragon Plus Environment
16

Page 17 of 43
Organic Process Research & Development
1
2
3
4
5
6
7
8
TP-2758 Research, Manufacturing and Impurity Control
TPꢀ2758 was manufactured by a threeꢀstep sequence that begins with the coupling of two key
intermediates, LHP 10 and enone 7, with deprotection and recrystallization steps to follow (
Scheme 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 8. Manufacture of TP-2758
H3C
H
CH3
H3C
H
CH3
LDA/NEt3—HCl
TMEDA, <ꢀ70 oC
LHMDS
N
OCH3
N
OCH3
H
H
H
CH3
O
O
N
N
N
N
+
CH3
CO2Ph
CH3
85% to 90%
OBn
OBn
O
O
OTBS
OBn O
OH
O
OBn
OTBS
LHP (10)
enone (7)
36
48% HF
90% to 95%
CH3CN
H3C
H2, Conc. HCl, Pd/C
CH3OH/H2O
SilaBond DMT
CH3OH/MTBE
H3C
H
CH3
OH
CH3
OCH3
N
O
OCH3
H
N
H
H
H
H
O
N
N
—2HCl
NH2
N
CH3
CH3
90% to 95%
OH
OBn
OH
O
OH
O
OBn O
OH
O
OH
TPꢀ2758 (2)
37
MeOH/EtOH/MTBE/water
recrystallization
90%
TPꢀ2758 API
MichaelꢀDieckmann (MꢀD) reaction of LHP 10 and enone 7 afforded the desired coupling
product 36. Similar to the MꢀD procedure used for eravacycline synthesis,2 triethylamine
hydrochloride was added to ensure reproducible and complete deprotonation of the LHP 10.
TMEDA was also added to potentially stabilize the anion, although this stabilization has not
been conclusively verified. Cryogenic conditions were maintained throughout the anion
ACS Paragon Plus Environment
17

Organic Process Research & Development
Page 18 of 43
1
2
3
4
5
6
7
8
9
formation and the subsequent transfer of the 7ꢀanion via a preꢀchilled cannula into the preꢀcooled
solution of enone 7. The crude product was purified by a methanol/water slurry. Because of the
relatively good solubility of MꢀD product 36 in methanol, water was added to maximize the
recovery. It is worth noting that despite the low solubility of 10 and 7 in methanol in one
incomplete reaction we were amazed to find that reslurry of 36 improved the purity from ~80%
to 95% HPLC purity. Furthermore this reslurry had no loss of 36 to the mother liquors. During
our reaction condition optimization at small scale (<5 g) this step only gave 70 to75% yield,
which is significantly lower than the >90% yield in the eravacycline program2 using the same
enone 7, presumably due to the instability of 10-anion. We worried that this issue might be
exacerbated at larger scale, so the reaction was scaled gradually and carefully. To our relief, after
scaling up this reaction to > 20 g, we have consistently obtained 36 in 85% to 90% yield,
inferring that the 10ꢀanion was most susceptible at the commencement of the reverse transfer
when the transfer line was not sufficiently cold, which should become less of a concern at larger
scale.
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
MichaelꢀDieckmann reaction using LHP 10 with 99% ee consistently gave 36 containing ~
1.5% of side chain epimer 38 (see Figure 3) irrespective of scale (ranging from 10 g to 340 g of
enone 7). This result indicated that there was minor chiral center inversion during the MꢀD
reaction. The side chain epimer 38 would further undergo desilylation and hydrogenolysis to
afford 39 as one of the major impurities in the API.
The second step, treatment of compound 36 with aqueous hydrofluoric acid in acetonitrile at
room temperature, gave the desilylation product 37 in near quantitative yield and with ~95%
HPLC purity. In the final step, global deprotection and isoxazole opening were achieved by
hydrogenation using Pd/C as catalyst in the presence of conc. HCl. In this step one major
ACS Paragon Plus Environment
18

Page 19 of 43
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
impurity with a molecular weight (MW) of Mꢀ18 (M represents the MW of TPꢀ2758) was
identified and the structure was tentatively assigned as 40 (Figure 3) based on the knowledge
acquired from other similar hydrogenations at Tetraphase.2 On one occasion, this impurity was
observed in amounts >10%.16 Impurity 40 could be partially reduced to 41 with a MW of Mꢀ16
under the same hydrogenation condition. We have found that in general for the tetracycline
analogs synthesized at Tetraphase, impurities Mꢀ16 are particularly difficult to purge, therefore
impurity 41 should be controlled during the hydrogenation.
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. Major impurities formed en route to TPꢀ2758
We examined the earlier small batches which generated less impurity 40 and found that for all
those batches the MꢀD product 36 was purified either by column chromatography or by silica gel
plug in combination with methanol slurry. This indicated there might be traces of contaminants
poisoning the catalyst in the instance where more Mꢀ18 impurity 40 was generated. It was also
confirmed that once the hydrogenation was complete the sum of impurities 40 and 41 would not
increase with extended reaction time. It appeared that after the isoxazole ring was opened an
ACS Paragon Plus Environment
19

Organic Process Research & Development
Page 20 of 43
1
2
3
4
5
6
7
8
9
elimination of benzyl alcohol occurred instead of benzyl deprotection. We hypothesized that an
accelerated hydrogenolysis of benzyl ether could suppress this elimination. To avoid the catalyst
poisoning, the crude desilylation product 37 was purified and then tested for hydrogenation. First
we attempted recrystallization of 37ꢀdiꢀHCl salt. It was not successful as the solid was very fine
and was not filterable. We then focused on purification of the free base and the results are
summarized in Table 2. The reaction solvent system was simultaneously investigated and it was
discovered that impurity 40 could be reduced by half by simply adding 1 to 1.5 vol. of water into
the hydrogenation reaction (entries 4 to 7). Thus all the experiments after entry 7 used 12 vol of
methanol and 1.5 vol of water. Charcoal treatment was not effective (entry 2). Acetone/water
recrystallization improved the reaction only marginally (entry 11). MTBE slurry gave low
recovery (entry 3). Both methanol and ethanol slurry significantly lowered 40+41 to ~1% but
suffered >20% material loss (entries 8 and 12). The optimal purification condition was a simple
acid (HCl)/base (K2HPO4) extraction that gave ~95% recovery but also reduced 40+41 to ~1%
(entry 5). After acid/base extraction, further purification of compound 37 by ethanol or methanol
slurry only slightly reduced 40+41 in the hydrogenation step (entries 9, 10, 15, 16). This extra
purification, with significant loss of material, was considered neither necessary nor worthwhile
as impurities could also be purged in the final recrystallization. In the production process, the
final EtOAc solution of compound 37 is passed through a pad of silica gel (0.5× weight) before
evaporation. This is because the first batch of TPꢀ2758 contained lots of inorganic salt (ROI:
~1%). The silica gel filtration can effectively remove the inorganic salt and possibly some
baseline impurities.
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 also investigated other catalysts and catalyst loadings. Nonꢀpalladium catalysts (i.e., Pt/C,
Rh/C and Rh/Al2O3) couldn’t remove the benzyl groups. LC/MS showed only one product with a
ACS Paragon Plus Environment
20

Page 21 of 43
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
MW of plus 2, indicating the isoxazole ring opened. The readily available catalyst of 10% Pd/C
(wet) from Aldrich or Alfa Aesar was the best for this hydrogenation. Other types of Pd/C and
Pd(OH)2/C were also tested, and in all the reactions impurity 40 nearly doubled. The optimal
catalyst loading was 10 wt% of 10% Pd/C (wet). With more catalyst (20 wt%, entry 13), the total
amount of 40 and 41 was not reduced while the conversion of 40 to 41 was accelerated. With
less catalyst (5%, entry 14) more 40 and 41 were generated.
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. Purification of 37 and Impurity Level in the Hydrogenation Step
Experiment PreꢀTreatment of 37
Hydrogenation Sum of 40 and
Solventsa
41
1
2
3
4
5
6
7
8
9
No treatment
Charcoal slurry
MTBE slurry (60% recovery)
Acid/base extraction
Acid/base extraction
MeOH Reꢀslurry 75% recovery
MeOH Reꢀslurry
MeOH/water Reꢀslurry
Acid/base extraction and MeOH reꢀslurry
Acid/base extraction and MeOH/water reꢀ
slurry
MeOH
MeOH
MeOH
5.6%
5%
1.8%
3.2%
1.1%
2.1%
1.2%
1.4%
0.8%
MeOH
MeOH/water
MeOH
MeOH/water
MeOH/water
MeOH/water
10
MeOH/water
0.9%
11
12
Acetone/water recrystallization
EtOH slurry (80% recovery)
EtOH slurry
MeOH/water
MeOH/water
MeOH/water
(20% catalyst)
MeOH/water
(5% catalyst)
MeOH/water
2.2%
0.9%
13
0.9%
EtOH slurry
14
15
16
1.3%
0.6%
0.8%
Acid/base extraction and EtOH reꢀslurry
Acid/base extraction and EtOH/water reꢀ
slurry
Acid/base extraction and EtOH/hexane
reꢀslurry
MeOH/water
MeOH/water
17
0.9%
a. All reaction use 10% Pd/C (10%, wet) unless stated
Finally alternative pharmaceutically acceptable acids were screened for the hydrogenation.
HBr was comparable to HCl. With MsOH, TsOH, sulfamic or sulfuric acid, 40 was rapidly
converted to 41. In addition, with MsOH, TsOH, sulfamic acid the Cꢀ4 position was more prone
ACS Paragon Plus Environment
21

Organic Process Research & Development
Page 22 of 43
1
2
3
4
5
6
7
8
9
to epimerization forming Cꢀ4 epimer 42. Cꢀ4 epimerization is a wellꢀknown issue for
tetracyclines and has to be controlled.17 Reactions involving benzenesulfonic acid, TFA, 1ꢀ
naphthalenesulfonic acid, 2ꢀnaphthalenesulfonic acid, oxalic acid, Nꢀcyclohexyl sulfamic acid or
hippuric acid were sluggish, generating significant amount of Cꢀ4 epimer 42 along with other
unknown impurities. Thus we focused our optimization on HCl.
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
Using the optimized conditions, the impurities 40 and 41 have been consistently controlled at <
2% and ~0.1% levels respectively, along with < 2% of epimer 42 and ~2% of side chain epimer
39 after the hydrogenation. The Pd/C was removed by filtration through a pad of celite. The
filtrate was further treated with scavenger SiliaBond DMT to reduce the Pd to an acceptable
level (10 ppm). After distillation of the methanol, crude TPꢀ2758 was reꢀdissolved in methanol
and precipitated out by addition of MTBE. The precipitation, with only ~1% loss of product,
could purge significant amount of epimers 39 (~50%) and 42 (>50%), but not impurities 40 and
41.
A brief salt screen using 7 acids (i.e., HCl, sulfate, fumaric acid, tartaric acid, ethanesulfonic
acid and 1,2ꢀethanedisulfonic acid) was performed, and it was found that TPꢀ2758 could form
crystal diꢀsalt with H2SO4, TsOH and HCl. A three month accelerated stability study showed that
for the diꢀHCl salt the Cꢀ4 epimer didn’t significantly increase under a 40 °C/75% RH storage
condition. Considering this stability, which is rare among the tetracycline family, as well as the
preference of HCl for the hydrogenation, the diꢀHCl salt was ultimately selected as the final API
form. TPꢀ2758 diꢀHCl salt is readily soluble in water and methanol (both >100 mg/mL), and
slightly soluble in ethanol (~2.5 mg/mL). Brief investigation revealed that TPꢀ2758 diꢀHCl salt
could be recrystallized using MeOH/acetone, MeOH/THF and MeOH/EtOH; however only the
MeOH/EtOH solvent system could effectively purge impurity 40 and epimer 42. To improve the
ACS Paragon Plus Environment
22

Page 23 of 43
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
recovery MTBE was also added as a more powerful antiꢀsolvent. After several batches we found
the final product after drying always contained ~3 wt% of water corresponding to 1 equivalent.
We suspected that the crystalline form was a monohydrate and water might facilitate the
crystallization. We were pleased to find that after addition of ~3.5 equiv of water to the
recrystallization system the yield was boosted from ~80% to ~90% without compromise of the
purity. This recrystallization can effectively purge ~90% of epimer 42, ~50% of side chain
epimer 39 and ~40% of impurity 40. Impurity 41 is only purged slightly.
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
A single crystal of TPꢀ2758 was grown by slow diffusion of acetone to TPꢀ2758 solution in
methanol over a 6 month period. The absolute structure of TPꢀ2758 was unambiguously
determined by the single Xꢀray structure in Figure 4. The unit also contained two HCl molecules
and one H2O molecule.
Figure 4. Single Xꢀray structure of TPꢀ2758
Conclusion
We have developed an efficient and scalable process for the synthesis of a fully synthetic
tetracycline TPꢀ2758 for preclinical and clinical studies. The complex LHP (10) was synthesized
in 11 steps with high yield for all the steps except the moderate methylation step. The chiral
pyrrolidine moiety was introduced into the LHP using Ellman’s sulfinamide chemistry with
excellent enantiomeric selectivity (99% ee). Crude TPꢀ2758 is manufactured convergently from
ACS Paragon Plus Environment
23

Organic Process Research & Development
Page 24 of 43
1
2
3
4
5
6
7
8
9
LHP (10) and enone (7) in 3 steps with 83% overall yield and good control of impurities,
especially the Mꢀ18 impurity 40 and Mꢀ16 impurity 41. Finally a recrystallization process using
methanol, ethanol, MTBE and water was developed to purify the TPꢀ2758 diꢀHCl salt with good
impurity purging and ~90% recovery. The API typically has an HPLC area purity of greater than
98.0%.
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
General Information
All commercially available reagents and solvents, including anhydrous solvents, were used
without further purification. If anhydrous conditions were required, the reactions were performed
under nitrogen; jacketed reactor was dried by sweeping with nitrogen while keeping the jacket
1
13
temperature at 50 °C; flasks were flame dried. H and C NMR spectra were recorded at room
temperature (RT) on a 400 MHz JEOL ECXꢀ400 spectrometer. Chemical shifts (δ) are expressed
in parts per million (ppm) relative to the corresponding deuterated solvent peak. HPLC analyses
were conducted on an Agilent 1200 series (SunFire C18 columns 5 ꢁm, 4.6 mm × 100 mm; and
mobileꢀphase gradients consisting of 0.1% TFA in water and 0.1% TFA in acetonitrile) or
Waters Alliance system (SunFire C18 column, 5 ꢁm, 4.6 mm × 50 mm; mobileꢀphase gradients
consisting of 0.1% formic acid in water and 0.1% formic acid in acetonitrile; MS detector,
Waters 3100). Analytical characterizations were obtained on materials that were representative
or purified for that purpose.
Dibromination of 1,4-dimethoxybenzene. A 50 L jacketed reactor was charged with 1,4ꢀ
dimethoxybenzene (1.386 kg, 10 mol, 1.0 equiv) and DCM (13.8 L) while keeping the vessel
under a blanket of nitrogen. After agitation the resulting solution was cooled to 2 °C. The vessel
ACS Paragon Plus Environment
24

Page 25 of 43
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
was vented to a NaOH solution (25%, 3 L) for HBr scrubbing. Neat Br2 (1.13 L, 22 mol, 2.2
equiv) was added to the reaction under nitrogen over 120 min while keeping the temperature
below 5 °C. After addition, reaction was stirred at 4 °C for 50 min at which time the reaction was
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
deemed complete by H NMR analysis. Towards the end of reaction some white solid
precipitated out. A chilled solution of K2HPO4 (3.48 kg was dissolved in water to make a 10 L
solution) was slowly added to the reaction over 15 min at < 8 °C to adjust the pH to ~7. Sodium
sulfite solution (10% in water, 5 L) was then added; the dark red color of the solution
disappeared after agitation for 15 min. The reaction mixture was allowed to warm to 20 °C, after
which it was stirred for an additional hour. The organic layer was separated and washed
sequentially with water (10 L) and brine (10 L). The solution was dried over Na2SO4 and
concentrated on a rotary evaporator to give a white solid, which was further dried under high
1
vacuum to give product 25 (2.918 kg, 98.6 % yield; 97% HPLC purity). H NMR (400 MHz,
CDCl3) δ 7.10 (s, 2H), 3.84 (s, 6H); 13C NMR (100 MHz, CDCl3) δ 150.5, 117.1, 110.5, 57.0.
Carboxylation of 25. A 50 L jacketed reactor was charged with solid 25 (3.73 kg, 12.61 mol,
i
1.0 equiv) and anhydrous THF (7.5 L). A solution of PrMgCl·LiCl in THF (11.6 L, 1.3 M,
15.15 mol, 1.2 equiv) was added to the suspension maintaining internal temperature between 24
and 27 °C. The resulting homogeneous dark solution was stirred at 25 °C for 2 h 45 min, at
which time LCꢀMS analysis showed the MgꢀBr exchange was complete. After the reaction
mixture was cooled to –40 °C, a stream of CO2 gas was bubbled through the reaction mixture via
Teflon tubing at –39 to –40 °C for 3 h 30 min until the reaction was deemed complete by HPLC
analysis. The resulting reaction mixture was warmed to 23 °C. Because of limited reactor
capacity, the subsequent work up was divided into two portions. Half of the reaction mixture was
concentrated on a rotary evaporator to give a white foamy solid. To the residue was added 1.2 N
ACS Paragon Plus Environment
25

Organic Process Research & Development
Page 26 of 43
1
2
3
4
5
6
7
8
9
aqueous HCl (9.4 L) to give a suspension, which was stirred for approximate 1h. The white solid
was collected by filtration and washed with water (4 L). LCꢀMS indicated that the product
contained ~3% desꢀbromo byproduct. The other half of the reaction mixture was processed the
same way and combined with the first half. To remove the desꢀbromo byproduct, the resulting
solid was stirred with 1 N NaOH solution (27 L) and MTBE (13 L) for 2 h at RT. The aqueous
layer was separated and then acidified with 2 N HCl (15 L) at < 30 °C to give a white slurry (pH
of the aqueous solution < 2), which was stirred for additional 1 h and filtered. The solid was
washed with water (6 L) and was suction dried in the air for 2 days, and then dried in a vacuum
oven (50 °C) for 2 days to give product 26 (3.03 kg, 92% yield, 99% HPLC purity), containing ~
3% of water by KF. MS (ESI) m/z calcd for C9H10BrO4 260.98, 262.97 (M+H+); found 283.14,
285.13. 1H NMR (400 MHz, CDCl3) δ: 10.67 (br s, 1H), 7.67 (s, 1H), 7.28 (s, 1H), 4.04 (s, 3H),
3.91 (s, 3H). 13C NMR (100 MHz, CDCl3) δ: 165.0, 152.1, 151.5, 118.9, 117.8, 117.5, 115.4,
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
+
57.8, 57.1. MS (ESI) m/z calcd for C9H10BrO4 260.98, 262.97 [M+H]; found 261.10, 263.10.
Phenyl ester formation. A 50 L jacketed reactor was equipped with a condenser through
which off gas was vented to a scrubber containing 4 N aqueous NaOH (3.5 L). The reactor was
charged with solid 26 (3.17 kg, 12.14 mol, 1.0 equiv) followed by DCM (32 L). To the resulting
suspension DMF (9.25 mL, 0.12 mol, 0.01 equiv) was added. Oxalyl chloride (1.27 L, 14.57
mol, 1.2 equiv) was added to the reaction mixture over 80 min at 27 to 24 °C. Vigorous gas
evolution was observed. At the end of addition, the reaction mixture became a thinner yellow
suspension. After another 2 h of stirring, gas evolution ceased. An aliquot of the reaction
mixture quenched with anhydrous MeOH and checked with HPLC showed ~9% of remaining
starting material 26. Additional oxalyl chloride (212 mL, 0.2 equiv) was added to drive the
reaction to completion.
ACS Paragon Plus Environment
26

Page 27 of 43
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
PhOH (1.60 kg, 17 mol, 1.4 equiv) was added in one portion. The reaction temperature
dropped from 25 °C to 21 °C in 2 min. Then DMAP (29.7 g, 0.24 mol, 0.02 equiv) was added in
one portion at 22.5 °C after which the reaction temperature rose to 26 ºC in 2 min and vigorous
gas evolution was observed. The resulting reaction mixture was stirred for 30 min at < 26 °C.
An aliquot of the reaction mixture was quenched with anhydrous MeOH and checked with
HPLC, showing methyl ester : SM 26 : product 27 = ~8 : 10 : 82. TEA (1 L, 7.17 mol, 0.59
equiv) was added via an addition funnel to the reaction mixture over 10 min.18 A lot of white
smoke was observed, and the reaction temperature rose from 25 °C to 33 °C. The resulting
reaction mixture was stirred at that temperature for 5 min. HPLC analysis indicated the reaction
was complete (product 27 : PhOH : diphenyl oxalate=~90.6 : 7.4 : 1.9).
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
Aqueous HCl solution (1 N, 12 L) was added to the reaction mixture. After addition it
was stirred at 23 to 25 °C for 10 min. The organic layer was separated and then treated with 1 N
aqueous NaOH solution (12.7 L) at RT for 20 min to hydrolyze the diphenyl oxalate and to
remove the phenol. The organic layer was then washed with brine (12 L) and concentrated on a
rotary evaporator to give a white solid, which was dried in a vacuum oven (50 °C) for 2 days to
1
give product 27 (3.96 kg, 96% yield, ~98% HPLC purity). H NMR (CDCl3) showed product
1
containing PhOH (~1.3 %) and DMAP (0.8%). H NMR (400 MHz, CDCl3) δ: 7.54 (s, 1H),
7.44ꢀ7.40 (m, 2H), 7.28ꢀ7.20 (m, 4H), 3.91 (s, 3H), 3.88 (s, 3H). 13C NMR (100 MHz, CDCl3) δ:
163.7, 154.2, 150.7, 149.7, 129.3, 125.8, 121.7, 118.3, 118.2, 118.0, 114.6, 56.9, 56.8. MS (ESI)
+
m/z calcd for C15H13BrNaO4 358.99, 360.99 (M+Na); found 359.04, 361.03.
Methylation reaction. A 3ꢀL flask was charged with Et3N·HCl (2.57 g, 19.05 mmol,
0.03 equiv), HNsBu2 (141.7 mL, 825.2 mmol, 1.3 equiv) and anhydrous THF (500 mL). The
resulting reaction mixture was cooled to –60 °C with a dry ice/acetone bath. A solution of nBuLi
ACS Paragon Plus Environment
27

Organic Process Research & Development
Page 28 of 43
1
2
3
4
5
6
7
8
9
in hexanes (311 mL, 2.45 M, 716.6 mmol, 1.2 equiv) was added slowly, maintaining solution
temperature below –50 °C. The resulting light yellow solution was warmed to –10 °C and then
cooled to –70 °C. TMEDA (123.7 mL, 825 mmol, 1.3 equiv) was added, maintaining solution
temperature below –66 °C. The resulting reaction mixture was cooled to –75 ºC.
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
A solution of ester 27 (214.39 g, 634.7 mmol, 1.0 equiv) in anhydrous THF (700 mL)
was precooled to –74 °C and was then transferred to the above LDA solution via a cannula
(wrapped with insulation tape packed with dry ice) slowly maintaining solution temperature
below –75 °C over 21 min. The flask and transfer line was rinsed forward with 20 mL THF.
The resulting dark brownish solution was then stirred at –74 to –78 °C for 30 min.
A 5 L flask was charged with anhydrous THF (500 mL) and MeI (237 mL, 3.8 mol, 6.0
equiv). The MeI solution was cooled to –40 °C with a dry ice/acetone bath. The above anion
was transferred to the MeI solution via Teflon tubing (wrapped with dry iceꢀcooled sleeve)
slowly while maintaining the temperature at –38 to –41 °C over 30 min. The cold bath was then
removed. The resulting brown reaction mixture was allowed to warm to RT over 1 h to give an
orange suspension. The reaction mixture was then cooled to 8 °C with an ice/water bath.
Aqueous Me3N solution (25 wt%, 900 mL, 3.8 mol, 6 equiv) was added over 5 min, and the
temperature rose to 29 °C. A lot of white solid formed and the resulting suspension was stirred
at RT for 30 min. The mixture was again cooled to 11 °C with an ice/water bath. Then 2 N HCl
(1.6 L) was added and the temperature rose to 24 °C and more solid formed. The resulting
mixture was filtered, and the cake was washed with MTBE (1.6 L). The combined filtrate was
transferred to a separatory flask. The two layers were separated.
1
An aliquot of the organic phase was concentrated. H NMR (CDCl3) showed the ratio of
product 28:SM 27:PhOH=~1 : 0.034 : 0.28.
ACS Paragon Plus Environment
28

Page 29 of 43
Organic Process Research & Development
1
2
3
4
5
6
The organic phase was washed with 1 N aqueous NaOH solution twice (800 mL and 400
mL), and concentrated on a rotary evaporator to give an orange oil.
7
8
9
The crude product was subjected to selective hydrolysis to remove starting material 27.
The crude product was dissolved in 2ꢀMeTHF (1 L) and stirred with 2 N aqueous NaOH (170
mL) at 72 °C for 2 h. The organic phase of the reaction mixture was checked with LCꢀMS and
showed no compound 27. After cooling to RT the layers were separated, and the aqueous layer
was extracted with MTBE (200 mL). The combined organics were sequentially washed with 1 N
aqueous NaOH (150 mL), water (300 mL) and brine (200 mL), and then concentrated on a rotary
evaporator. The residue was diluted with toluene, and the resulting solution was filtered through
a silica gel plug (430 g); the plug was washed with EtOAc/heptane (1:12 v/v). The filtrate was
concentrated and dried under vacuum to give the desired product 28 as a pale yellow oil (156.8g,
70% yield, > 98% HPLC purity). (White crystalline product was achieved in a ~4 kg scale
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
reaction when the product was triturated from heptane/MTBE). H NMR (400 MHz, CDCl3) δ:
13
7.44ꢀ7.40 (m, 2H), 7.28ꢀ7.22 (m, 3H), 7.02 (s, 1H), 3.84 (s, 3H), 3.78 (s, 3H), 2.41 (s, 3H). C
NMR (100 MHz, CDCl3) δ: 165.5, 152.8, 150.6, 149.3, 131.2, 129.4, 126.0, 123.6, 121.5, 119.2,
+
113.7, 60.5, 56.4, 13.6. MS (ESI) m/z calcd for C15H13BrNaO4 373.01, 375.00 (M+Na); found
373.09, 375.05.
De-methylation reaction of compound 28. In a 50 L reactor, a solution of compound 28
(1734 g, 4.937 mol, 1.0 equiv) in DCM (12.34 L) was cooled to ‒65 °C. BBr3 solution in DCM
(4.765 L, 1.0 M, 4.765 mol, 0.965 equiv) was slowly added over 45 min while maintaining the
internal temperature at ‒65 to ‒53 °C. After addition, the mixture was stirred at ‒50 °C for 35
min after which it was allowed to warm up to ‒5 °C over 55 min, and then kept at that
temperature for 1 h. Cold water (0 °C, 8.6 L) was added to the reaction mixture over 2 min, the
ACS Paragon Plus Environment
29