Process Research and Development of TP-808: A Key Intermediate for the Manufacture of Synthetic Tetracyclines

Article abstract of DOI:10.1021/acs.oprd.7b00003

Process research, development, and manufacture of TP-808 (1), a key intermediate for the discovery and manufacture of tetracycline analogues, is described. The process used for the preparation of 1 avoids chromatographic purifications and has been substan

Full text of DOI:10.1021/acs.oprd.7b00003

Subscriber access provided by UB + Fachbibliothek Chemie | (FU-Bibliothekssystem)
Full Paper
Process Research and Development of TP-808: A Key
Intermediate for the Manufacture of Synthetic Tetracyclines
Wu-Yan Zhang, Chi-Li Chen, Minsheng He, Zhijian Zhu, Philip
Hogan, Olga Gilicky, Nicholas Dunwoody, and Magnus Ronn
Org. Process Res. Dev., Just Accepted Manuscript • DOI: 10.1021/acs.oprd.7b00003 • Publication Date (Web): 10 Feb 2017
Downloaded from http://pubs.acs.org on February 11, 2017
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 30
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
Process Research and Development of TPꢀ808: A
Key Intermediate for the Manufacture of Synthetic
Tetracyclines
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*, Chi-Li Chen, Minsheng He, Zhijian Zhu, Philip Hogan, Olga Gilicky, Nicholas
Dunwoody, and Magnus Ronn
Tetraphase Pharmaceuticals, Inc, 480 Arsenal Way, Suite 110, Watertown, Massachusetts 02472,
United States
* To whom correspondence should be addressed: wzhang@tphase.com
ORCID: WuꢀYan Zhang, 0000ꢀ0002ꢀ4858ꢀ4784
ACS Paragon Plus Environment
1

Organic Process Research & Development
Page 2 of 30
1
2
3
4
5
6
7
8
9
H3C
H
CH3
H3C
CH3
no chromatographic
N
O
N
purification
O
O
O
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
CHO
N
N
+
56% yield
OBn
OCH3
OBn
O
OTBS
TP-808
ACS Paragon Plus Environment
2

Page 3 of 30
Organic Process Research & Development
1
2
3
4
Abstract
5
6
7
8
9
Process research, development and manufacture of TPꢀ808 (1), a key intermediate for the
discovery and manufacture of tetracycline analogs is described. The process used for the
preparation of 1 avoids chromatographic purifications and has been substantially improved over
the previously reported preparation. The robustness of the process was demonstrated in a 76.1 kg
manufacturing campaign with 56% overall yield.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Key words: tetracycline, magnesiation, isoxazole, OmuraꢀSharmaꢀSwern oxidation, Parikhꢀ
Doehring, enone, eravacycline
Introduction
As part of an ongoing effort to develop an efficient process for the manufacture of novel fully
synthetic tetracyclines unavailable from traditional semi synthetic routes, our group has
extensively investigated the process for the manufacture of enone 1, a key intermediate for the
synthesis of tetracyclines of wide structural diversity (Scheme 1).1 Several synthetic approaches
to the tetracycline antibiotics have been published previously, however most suffer from low
overall yields as well as lengthy linear sequences.2 The convergent, discoveryꢀenabling synthetic
approach to the tetracycline class of antibiotics reported by Myers group combines 1 and 2, a
substituted phenyl benzoate, in a baseꢀpromoted MichaelꢀDieckmann reaction to give
pentacyclic intermediates 3 followed by deprotections to give the tetracycline analogues 4. Our
lead compound, eravacycline (5) is manufactured using this approach.3
Scheme 1. Myers’ Route to Tetracycline Analogues and the Structure of Eravacycline.
ACS Paragon Plus Environment
3

Organic Process Research & Development
Page 4 of 30
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
In 2007 Myers and Brubaker published a rapid and convergent route to 1 from two heterocyclic
precursors, isoxazole 6 and 3ꢀmethoxyꢀ2ꢀfuraldehyde (7) (Scheme 2).4 This route was used to
generate the first batches of material for research at Tetraphase, but all the steps needed
improvement for large scale manufacturing. The lithiation of the isoxazole 6 and addition of the
resulting anion to aldehyde 7 to give a mixture of alcohols 8a and 8b required a temperature of
−100 °C; above this temperature the deprotonation is not regioselective leading to deleterious
side reactions. The intramolecular DielsꢀAlder reaction of the mixture of 8a and 8b required a
minimum of 7 days to reach completion. The overall selectivity for the favored endo isomers 9a
and 9b relative to the unfavored exo isomers 9c and 9d was modest (~3:1) and the subsequent
oxidation of the alcohols (9a-d) utilized cryogenic Swern conditions
.
The boron trichlorideꢀ
promoted demethylation/ring opening step was not robust and at times provided low yields of
11a. After each operation the desired product was isolated by silica gel chromatography. 10a and
10b were particularly difficult to separate in this manner. An overall yield of 29% was reported.
As Tetraphase’s lead candidate, eravacycline (5) progressed through development, the
demand for 1 required significant improvements in its manufacturing process.
ACS Paragon Plus Environment
4

Page 5 of 30
Organic Process Research & Development
1
2
3
4
Scheme 2. The MyersꢀBrubaker route to 1
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
Results and Discussion
Step 1 Coupling:
At temperatures higher than −100 °C, allylic metallation (i.e., removal of proton α to the
dimethylamino group) of 6 is competitive with metallation of the isoxazole ring. We attempted
to improve the regioꢀselectivity of the deprotonation of 6 by using different reagents including nꢀ
butyllithium, nꢀhexyllithium, phenyllithium, methyllithium, KHMDS, and LiTMP in different
solvents including THF, ethyl ether, DME and toluene. We also explored the use of additives
such as TMEDA. To determine the regioselectivity of deprotonation, the reaction was quenched
ACS Paragon Plus Environment
5

Organic Process Research & Development
Page 6 of 30
1
2
with D2O (Scheme 3) and the selectivity was determined by 1H NMR. It was found that nꢀhexyl
lithium using THF as solvent gave the best selectivity at −78 ºC (82% 12, 18% 13 and 14) which
was slightly better than when nꢀBuLi was used. An analogous distribution of products was seen
when the resulted anions reacted with furaldehyde 7. All other conditions gave lower
regioselectivity. In many cases the deprotonation of the allylic position was predominant.
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
Scheme 3. NMR Determination of RegioꢀSelectivity
We proceeded to investigate the direct magnesiation of 6 using a magnesium base. Hauser,5
and later Eaton6 and Knochel7 have investigated the use of magnesium dialkylamides for the
direct magnesiation of acidic CꢀH bonds. These bases have been used in the direct and selective
magnesiation of acidic CꢀH bonds within arenes, heteroarenes and even certain alkenes and
alkanes under mild conditions. To our knowledge, no metallations of isoxazoles had been
examined with these types of bases at the initiation of this work.
Three different bases derived from tetramethylpiperidine (TMP), TMPMgClꢁLiCl,
TMPMgBrꢁLiBr and TMPMgBrꢁLiCl8 were examined with regards to the regioselective
metallation of 6 and the subsequent reaction of the organomagnesium species derived from 6
(will be referred to as 6ꢀanion) towards aldehyde 7 with respect to both conversion and
stereoselectivity. The stereoselectivity of the addition of 6 to 7 plays a key role in the
stereoselctivity of the DielsꢀAlder reaction (vide infra). All three bases metallated 6 exclusively
ACS Paragon Plus Environment
6

Page 7 of 30
Organic Process Research & Development
1
2
on the isoxazole ring at moderate temperature.9 The bases containing bromide counter ion,
TMPMgBrꢁLiBr and TMPMgBrꢁLiCl, provided slightly more favorable ratios of 8a to 8b,
however the reactions with these bases were often incomplete and less clean, thus offered no
advantage over TMPMgClꢁLiCl, whose reactivity was superior in terms of providing the most
consistent high yielding coupling reactions. For these reasons TMPMgClꢁLiCl was chosen for
further investigation and scale up.
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
Regioselective deprotonation of 6 proceeded rapidly at 0 °C using 1.2 to 1.4 equiv of
TMPMgClꢁLiCl. The furaldehyde 7 (1.1 equiv) was then added at −30 to −20 ºC to furnish the
coupled product with ~2.3 : 1 ratio of 8a to 8b. These reaction conditions were successfully
scaled to 1.6 kg scale with ~80% assay yields. An ensuing larger scale manufacturing run
provided a 10 ~ 20% lower yield. We suspected that 6 and/or 6-anion might be unstable under
these conditions. To investigate the stability, three experiments were performed. After
deprotonation of 6 with 1.4 equiv of TMPMgClꢁLiCl at 0 ºC, the reactions were quenched at 0
°C at 0.5h, 1h and 2h, respectively. Although no significant impurities were detected by HPLC
1
and H NMR, the recovered assay yields of all three samples were approximately 15% lower
than expected, indicating some nonꢀspecific decomposition under these conditions. The fact that
deprotonation would not achieve completion without the extra 0.2 to 0.4 equiv of base suggested
that at 0 °C the decomposition occurred concurrently to the deprotonation and consumed base.
We proceeded to examine the metallation at a lower temperature. The magnesiation of 6 using
1.4 equiv of TMPMgClꢁLiCl was carried out at approximately −25 ºC and was complete in ~ 40
min. The aldehyde 7 (1.1 equiv) was added and an IPC indicated ~30% of 6 was not consumed
after overnight at room temperature. Additional 7 (0.3 equiv) was added leading to nearly
quantitative consumption of 6 without the previously observed nonꢀspecific decomposition of 6.
ACS Paragon Plus Environment
7

Organic Process Research & Development
Page 8 of 30
1
2
3
4
5
6
7
8
9
This indicated substantial stability of 6ꢀanion even at room temperature. We speculate that 7
reacts stoichiometrically with the excess TMPMgClꢁLiCl first and consequently both the base
and aldehyde 7 are not available for reaction with 6ꢀanion. Indeed, adding a slight excess of 7
relative to TMPMgClꢁLiCl in general leads to a smooth and complete coupling. We noted that 7
was partially recovered after quench. In house experimentation revealed that the amount of
TMPMgClꢁLiCl and 7 used in this coupling could be lowered to 1.1 equiv each without
sacrificing conversion and yield with the benefit of not wasting substantial amounts of 7.
Without the extra base the metallation took longer but still reached completion in 1.5 h. During
the manufacturing normally 1.2 to 1.3 equiv of base is needed possibly due to generally less
anhydrous conditions.
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. Coupling Reaction selectivity and conversion
Temperature
(°C)
Selectivity
8a/8b
Conversion
Entry*
Solvent
Additive
of 6
1
2
THF
THF
5.0:1
14%
46%
53%
99%
93%
85%
16%
16%
19%
98%
93%
−70
4.1 to 1
3.3 to 1
3.5 to 1
3.2 to 1
3.3 to 1
5.1 to 1
6.4 to 1
5.9 to 1
3.7 to 1
4.1 to 1
−60
3
THF
−50
4
THF
−70 to −10
−70 to −10
−70 to −10
−70
5
2ꢀMeTHF
toluene
THF
6
7
LiCl (2equiv)
LiBr (3equiv)
MgBr2 (2equiv)
LiBr (3equiv)
MgBr2 (2equiv)
8
2ꢀMeTHF
2ꢀMeTHF
2ꢀMeTHF
2ꢀMeTHF
−70
9
−70
10
11
−70 to −10
−70 to −10
ACS Paragon Plus Environment
8

Page 9 of 30
Organic Process Research & Development
1
2
3
4
*All reactions used 1.2 equiv of base and 1.2 equiv of furaldehyde 7.
5
6
7
8
9
Another aspect of the reaction concerns the stereoselectivity since the Sꢀalcohol (8a) is favored
in the subsequent DielsꢀAlder reaction (vide infra). Our results for the coupling between 6 and 7
are summarized in Table 1. The coupling at −70 °C (entry 1) using TMPMgClꢁLiCl provided the
most favorable ratio of 8a/b observed (5:1), however the reaction stalled with 86% of 6
remaining. Raising the temperature of the reaction to −60 °C and −50 °C (entries 2 and 3)
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
provided somewhat better conversion but with decreased diastereoselectivity.
The best
compromise between stereoselectivity and conversion was achieved by addition of 7 to a cold
solution (−70 °C) of 6ꢀanion and allowing the reaction temperature to gradually rise to −10 °C
(entry 4). The use of other solvents, toluene and 2ꢀmethyltetrathydrofuran (2ꢀMeTHF) (entries 5
and 6), suffered lower selectivity and lower conversion. Reactions with additives were also
investigated.10 The reactions in 2ꢀMeTHF with LiBr and MgBr2 as additives gave better
selectivity at low temperature, whereas the selectivity did not change in THF with LiCl as
additive (entries 7ꢀ9). The use of LiBr in THF improved the ratio slightly when the reaction was
gradually warmed up from −70 °C to −10 °C (entry 10). Although the selectivity of 8a/b was
improved to 4.1:1 with MgBr2 as additive, it was negated by lower conversion (about 7% of 6
recovered, entry 11). Other additives (e.g. Me3Al, Et2AlCl, BF3OEt2, TiCl4, and Ti(OꢀiPr)4) were
examined, however none improved the selectivity significantly. Overall none of the additives
benefited the reaction. Our optimal conditions employ THF as solvent with the metallation of 6
conducted at −25 °C and the subsequent coupling of 7 with 6ꢀanion conducted from −70 °C to
−10 °C to give a mixture of 8a/b with 3.5:1 selectivity favoring 8a. The products 8a and 8b are
not stable below pH 6 thereby the use of mineral acids for work up must be avoided. We also
ACS Paragon Plus Environment
9

Organic Process Research & Development
Page 10 of 30
1
2
3
4
5
6
7
8
9
observed that the presence of residual tetramethylpiperidine is detrimental to the subsequent
DielsꢀAlder reaction. Using saturated aqueous ammonium chloride solution for quenching the
reaction avoids strong acidic conditions and also effectively removes the tetramethylpiperidine
from the organics during phase cuts. In a typical manufacturing, 8a/b are obtained in greater than
90% purity by HPLC and ~90% yield, and the mixture can be directly used in the next step
without further purification.
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
Step 2 Diels-Alder Reaction
The originally reported DielsꢀAlder conditions for the conversion of 8a/b to 9a-d required 7
days and provided modest endo selectivity (8a gave 9a and 9c in a ratio of 6.2 to 1 and 8b gave
9b and 9d in a ratio of 1.6 to 1). Additionally, the volumetric efficiency was low as 26 volumes
of solvent were required and increasing the initial concentration of 8a/b provided lower
selectivity. We proceeded to investigate improving the endo/exo selectivity, the volumetric
efficiency and the reaction rate in order to reduce the manufacturing cycle.
We briefly investigated the effect of Lewis and mild Brönsted acids with regards to reaction
time and endo/exo selectivity.11 Lewis acids (Yb(OTf)3, Al(OꢀiPr)3 and silica gel) as additives as
well as halogenated alcohols12 ((CF3)2CHOH, CF3CH2OH and ClCH2CH2OH) as solvents were
examined. In line with our previous observations of the acid sensitivity of 8a/b, these
additives/solvents generally led to substantial decomposition of the starting materials. We
proceeded with an extensive solvent screen in the presence of 1.1 equiv of diisopropylethylamine
and 1% of BHT (Table 2). Reactions in chlorobenzene and α,α,αꢀtrifluorotoluene gave a similar
product distribution as toluene (entries 1 to 3). Excellent endoꢀselectivity was obtained at small
scale in acetonitrile (entry 4). However, at larger scale the selectivity dropped unexpectedly
ACS Paragon Plus Environment
10

Page 11 of 30
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
(entry 5). The reaction could be accelerated at higher temperature by using a sealed tube or
microwave reactor (entries 6 and 7). Notwithstanding the lower selectivity at higher temperature,
this provided quick access to DielsꢀAlder products. Excellent endo selectivity was also observed
when isopropyl acetate, methyl isobutyl ketone, butyl acetate, or tꢀbutyl acetate was employed at
small scale (entries 8 to 11), however the selectivity was not realized at a larger scale (entry 12).
The inconsistent results discouraged us from investigating these solvents for further scale up.
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. DielsꢀAlder Reaction Conditions and Results
Entrya
Solvent
Temperature 9a/9c
9b/9d
endo/exo
Reaction
time
(from 8a)b (from 8b)b selectivityc
1
2
3
Toluene
95 ꢀ 110 °C
95 °C
6:1
7:1
7:1
1.6:1
1.4:1
1.9:1
78.5:21.5
78.7:21.3
80.8: 19.2
7 days
6 days
6 days
chlorobenzene
α,α,αꢀ
95 °C
trifluorotoluene
4
acetonitrile
acetonitrile
acetonitrile
acetonitrile
reflux
reflux
105 °C
37:1
13:1
8:1
2:1
89.9:10.1
84.9:15.1
80.2:19.8
78.1:21.9
4 days
5 days
40 h
5d
6
2:1
1.5:1
1.3:1
7
150 ºC
7:1
1.5h
(microwave)
8
9
isopropyl acetate 95 ꢀ 100 °C
34:1
35:1
1.7:1
1.6:1
86.8:13.2
86.5:13.5
5days
methyl isobutyl 95 ꢀ 100 °C
ketone
5 days
10
11
12e
nꢀbutyl acetate
tꢀbutyl acetate
95 ꢀ 100 °C
95 ꢀ 100 °C
32:1
26:1
17:1
1.7:1
1.7:1
1.3:1
86.6:13.4
86.2:13.8
83.1:16.9
5 days
5 days
6 days
isopropyl acetate 95 °C
ACS Paragon Plus Environment
11

Organic Process Research & Development
Page 12 of 30
1
2
3
4
5
6
7
8
9
13
14
15
16
17
18
19
20
2ꢀpropanol (IPA) reflux
30:1
29:1
23:1
11:1
30:1
25:1
45:1
10:1
1:1
82.7:17.3
82.5:17.5
81.2:18.8
78.3:21.7
81.5:18.5
81.6:18.4
83.4:16.6
79.3:10.7
6 days
6 days
1 day
tꢀbutanol
95 °C
95 °C
95 °C
95 °C
95 °C
95 °C
95 °C
1:1
1,3ꢀpropanediol
Ethylene glycol
1,4ꢀdioxane
DMF
1:1.1
1:1.1
1:1
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
20 h
3 days
2 days
2 days
2 days
1.1:1
1: 1
DMSO
EtN(iPr)2
1.1:1
a: All the reactions used 30 vol of solvents, 1.1 equiv of i-Pr2NEt and 1% of BHT. Unless
noted, 100mg of starting material was used.
b: The selectivity is determined by the integration of the methoxy group by 1H NMR.
c: The endo/exo selectivity is calculated based on the endo/exo selectivity, and the 2.3 : 1
ratio of 8a to 8b (Sꢀ8 to Rꢀ8) in the starting material.
d: Starting material: 17 g.
e: Starting material: 10 g.
In alcoholic solvents (IPA and tꢀbutanol, entries 13 and 14), reaction of 8a gave excellent
selectivity (~30:1), while 8b was not selective (~1:1). The reaction proceeded without the
formation of substantial impurities by HPLC, however some tarry materials were observed at
kilogram scale. These materials could be removed by a filtration through a small pad of silica
gel. We were gratified that the reaction rate and selectivity were not affected by lowering the
solvent volume from 30 vol to 10 vol. Thus IPA was chosen as the solvent for a laboratory scale
campaign (1.5 kg) and larger scale pilot plant runs. This DielsꢀAlder reaction was complete in 6
days at gram scale but needed 7 to 8 days in the pilot plant. After distillation the residue was
purified by passage through a short silica gel (3x) column. We investigated the use of other,
higher boiling point alcohols for decreasing the reaction time. The use of primary alcohols (nꢀ
ACS Paragon Plus Environment
12

Page 13 of 30
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
butanol, 3ꢀmethylꢀ1ꢀbutanol) resulted in decomposition of starting materials and/or products. If
the reaction temperature was kept at ~95 ºC, other secondary and tertiary alcohols (e.g., tꢀamyl
alcohol, 3ꢀpentanol) gave similar results as the use of IPA. When the reaction was conducted at
110 ºC all reactions were complete in 2 days with increased impurities and lower selectivity (8a,
15:1 endo/exo; 8b: 1:1 endo/exo). The reactions in ethylene glycol and in 1,3ꢀpropanediol were
attractive because of the shorter reaction time (<1 day, entries 15 and 16), however the workups
were complicated by the formation of emulsions and consequently the yields were low.
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
Examination of the DielsꢀAlder reaction in polar aprotic solvents dioxane, DMF and DMSO
gave good selectivity and the process was substantially accelerated (entries 17ꢀ19). The use of
DMSO was particularly attractive, as the oxidation in the next step uses a substantial amount of
DMSO as a reagent. The reaction in DMSO was completed after 2 days at 95 °C. After the
reaction reached completion, it was stressed by keeping the reaction at 100 ºC for additional 2
days. Even after this extended reaction time, yield and purity were not compromised. We tested
these conditions with crude 8a/b containing residual 6 and 7 and monitored the process closely
1
by H NMR and LC/MS. Only products and impurities carried over from the starting material
were observed with both residual 6 and 7 unchanged. The reaction was run using 3, 5 and 10
volumes of DMSO respectively and no significant difference in selectivity or reaction rate was
noted. Because MTBE, ethyl acetate or isopropyl acetate could be used as an extraction solvent
for the coupling step, one volume of isopropyl acetate, ethyl acetate or MTBE was added to the
reaction respectively and again no adverse effect was observed.
In the presence of water, we observed in the course of these studies the formation of a
substantial byproduct 15 (Scheme 4) which, unfortunately, is not a productive intermediate. In
fact, 15 could be easily prepared by treatment of 9a/b with acids (HCl, HCOOH or TFA) and
ACS Paragon Plus Environment
13

Organic Process Research & Development
Page 14 of 30
1
2
3
4
5
6
water. In our hands 15 did not undergo any productive ring opening reactions, nor did ketone 16
which was independently prepared by oxidation of 15.
7
8
9
As water is also detrimental to the subsequent oxidation step, the moisture content before
DielsꢀAlder is controlled at <0.1% by KF, a level that ensures both steps proceed smoothly. If
the coupling product contains too much water it can be dried with a desiccant (e.g., 4 Ǻ
molecular sieves) or by azeotropic distillation. For large scale manufacturing, isopropyl acetate is
selected as the extraction solvent for the coupling step because it is efficient for azeotropic
removal of water. After the DielsꢀAlder reaction is complete the reaction mixture is diluted with
10 volumes of ethyl acetate and the resultant solution is used directly in the next 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
Scheme 4. Formation and Attempted Use of 15
Step 3. Oxidation of Diels-Alder Products
Although the Swern oxidation of alcohols 9aꢀd to ketones 10a/b using oxalyl chloride and
DMSO as reagents under cryogenic conditions was reported to proceed with good yield, on
multiple occasions in our hands this reaction condition gave significant byproducts. We
postulated that since in the Swern process the base (i.e., triethylamine) was added at the end, the
ACS Paragon Plus Environment
14

Page 15 of 30
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
reaction environment could be acidic which was known to be detrimental to 9a-d and 10a/b. We
investigated two other DSMO based oxidation methods. The ParikhꢀDoehring13 oxidation
(DMSO/SO3.Py/EtN(iꢀPr)2) could be performed at room temperature and provided an essentially
quantitative yield with 2 equivalents of SO3ꢁPy. After the DielsꢀAlder reaction was completed, to
the reaction mixture extra amine base and SO3ꢁPy were simply added at mild temperatures (20 to
40 °C) to furnish the oxidation product. When this material was used to complete the synthesis
of enone 1 an impurity was present that could not be removed by our normal purification (short
column chromatography followed by heptane slurry). This impurity was identified as
BCl3.pyridine complex, confirmed by simply mixing pyridine and BCl3 together. This complex
was difficult to quench, surviving treatment with acids or bases. To prevent the formation of this
byproduct, pyridine must be completely removed after SO3ꢁPy/DMSO oxidation. Azeotropic
distillation with toluene effectively removed most of the pyridine but not completely. The
pyridine could be extracted by washing the organics with aqueous CuSO4, but emulsions formed
at larger scales. The need to deliver 1 in excellent quality prompted us to continue screening
oxidation methods. As an alternative we investigated the possibility of using the Omuraꢀ
SharmaꢀSwern oxidation [Trifluoroacetic anhydride (TFAA)/DMSO/base].14 To assure a basic
environment throughout the reaction, the reagent charging sequence normally used in the
literature (i.e., alcohol is added to premixed DMSO and TFAA followed by an amine base at the
end) was modified. We investigated a reagent charging sequences in which TFAA was added to
a mixture of 9aꢀd, DMSO and base (NEt3 or i-Pr2NEt) in a solvent. This charging sequence
proved effective on larger laboratory scales. It also allows for the DielsꢀAlder reaction mixture,
which contains starting materials 9aꢀd, i-Pr2NEt and DMSO, to be directly used in the oxidation
without work up. Charging additional amine base followed by TFAA to the reaction mixture
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
15

Organic Process Research & Development
Page 16 of 30
1
2
3
4
5
6
7
8
9
after dilution with an appropriate solvent provides 10a/b. The reaction is complete as soon as the
charging of TFAA is complete. Typically 1.6 equiv of TFAA is sufficient to provide a complete
reaction on laboratory scales. Since the product is not stable under acidic conditions, sufficient
base should be added first to ensure that the mixture never becomes acidic. However, during
large scale manufacturing, the reactions using TFAA often stalled with ~3% of 9a remaining
even after 3 equiv of TFAA were charged, presumably because the reaction mixture was less
anhydrous at large scale. As a result the assay yield of the TFAA oxidation is generally 5%
lower compared to that of SO3ꢁPy oxidation. Additionally for the TFAA reaction, a temperature
range of 0 ꢀ 10 °C needs to be maintained to ensure a clean reaction and to avoid DMSO
freezing. For these reasons SO3ꢁPy is preferred although charging liquid is operationally
preferred in a plant setting. Using either TFAA or SO3ꢁPy, the same solvents and work up
procedure can be used. Several solvents performed satisfactorily for this oxidation including
THF, DCM, toluene and ethyl acetate. Ethyl acetate was selected as the solvent due to the ease
of phase cuts during workup. In addition, we observed that when using ethyl acetate as the
organic phase during the workup, DMSO and residual 7 from the coupling step above are easily
extracted into the aqueous washes. After extraction the crude ethyl acetate solution is filtered
through a pad of silica gel to remove any trace of polar impurities and then concentrated.
Only 10a is a productive intermediate for the manufacture of enone 1. The two isomers are not
easily separated chromatographically. We discovered that the compound resulting from
continued operations on 10b could easily be removed after the last step of the manufacturing
route of 1 (vide infra), allowing the mixture to be carried forward without purification in our
earlier manufacturing campaigns. Since ethyl acetate is not compatible with BCl3 in the next
step, the residual ethyl acetate is completely removed by azeotropic distillation with toluene.
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
16

Page 17 of 30
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
If the oxidation product is not purified, the final product 1 needs to be purified by column
chromatography followed by a heptane slurry. In order to remove chromatography from the
entire process, recrystallization of the crude oxidation product was investigated. A
recrystallization of 10a/b was developed using an ethanol/MTBE mixture yielding pure 10a.
This process is capable of purging impurities resulting from early manufacturing using unꢀ
optimized reaction conditions, however ~20% of 10a is lost to the mother liquor. After
optimizing the three steps leading to 10a/b, especially when SO3ꢁPy was used, the purity of the
crude material improved and an alternative recrystallization/reꢀslurry process using toluene and
heptane was developed. After the work up, the ethyl acetate solution is concentrated followed by
switching the solvent to toluene. The ratio of toluene to 10a (mL/g) is determined by IPC using
NMR and should be between 3 and 6. To the toluene solution, heptane (10 vol) is added at 70 to
80 °C followed by cooling to room temperature. The isolated product is typically > 90% pure as
a light brown solid, containing 1 to 7% of 10b depending on the quantity of toluene used. The
yield loss to mother liquor is normally 2 to 5%. This recrystallization reduced impurity carry
over, resulted in better control in subsequent reactions and allowed us to purify enone 1 without
chromatography. In addition, it completely removed the residual pyridine resulting from the use
of SO3ꢁPy.
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 twoꢀstep (DielsꢀAlder and oxidation) in one pot process has been successfully scaled up
with a typical 10a yield of 70%.
Step 4. Demethylative Oxygen Bridge Opening
Treating 10a/b with BCl3 at −40 °C gave crude tertiary alcohol 11a/b in moderate yield along
with varying amounts (up to 40%) of dieneꢀcontaining side products 18a and presumably minor
ACS Paragon Plus Environment
17

Organic Process Research & Development
Page 18 of 30
1
2
3
4
5
6
7
8
9
amount of 18b (Scheme 5). In general, more 18a/b is formed at lower reaction temperatures. The
quality and source of BCl3 also affected the outcome of the products. However, we could not
explain or predict the variation of the amount of 18a/b formed.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 5. Conversion of 10a/b to 11a/b
Since the amount of 18a/b formed is significant and variable, conversion of 18a/b to 11a/b
was seen as critical. Experimentation revealed that the vinyl ether 18a/b could be easily
converted to 11a/b with acid. Variable amounts (1 to 10%) of β,ϒꢀenone 19a/b were observed
when 18a/b was treated with TFA or TsOH in DCM, or HCl in methanol. Since DCM is used as
solvent in the ring opening reaction and workꢀup, it is easy to execute the conversion of 18a/b to
11a/b by addition of TFA or TsOH to the DCM solution after workup. Typically, treatment of
18a/b with TFA in DCM gives the least amount of 19a/b (1 to 5%) whereas TsOH gives up to
10%. However, the hydrolysis usually required 5 equiv of TFA and the reaction time varied
significantly (6 to 21 h) dependent on the water content of the reaction mixture. On the other
ACS Paragon Plus Environment
18

Page 19 of 30
Organic Process Research & Development
1
2
3
4
5
6
hand, the hydrolysis only needs 2 equiv of TsOH.H2O, and the reaction consistently reached
completion in 2h irrespective of the water content.
7
8
9
A time course study confirmed enone 11a, as well as the newly formed 19a/b, were stable
under the TFA or TsOH condition for >24h. The downstream product arising from 19a/b is
difficult to purge from 1 thus we examined the possibility of converting 19a/b to 11a/b. In the
presence of catalytic amount (10 mol%) of Et3N, isomerization of 19a/b to 11a/b was complete
at room temperature in 1 h. Weaker organic bases such as 2,6ꢀlutidine were not effective for this
conversion.
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 next examined the BCl3 reaction parameters such as equivalents, reaction temperature and
workꢀup. The ring opening reaction typically required 1.1 equiv of BCl3 if purified 10a was used
and 1.3 equiv of BCl3 when crude 10a/b was used. Excess BCl3 led to benzyl ether cleavage as
evidenced by the appearance of a benzyl chloride peak on HPLC, however other sideꢀproducts
were not detected indicating further nonspecific decomposition of the molecule other than the
benzyl removal. The effect of the purity of the starting material on the yield of the reaction was
also evaluated. When the oxidation product was purified by chromatography or recrystallization,
the yield of BCl3 reaction was improved by 5 to 10%. Later, it was found that a simple filtration
of the crude oxidation product through a silica gel pad (2x) removed the baseline polar impurities
and this resulted in a similar high yield. It is hypothesized that the baseline impurities may
partially quench BCl3 rendering it ineffective for the demethylative ring opening reaction while
still leading to nonspecific decompositions of the reaction intermediates.
The reaction rate at different temperatures was studied using 1.3 equiv of BCl3 relative to the
sum of 10a and 10b (Table 3). The reaction was rapid at 0 to −25 °C (entries 1ꢀ2) but slower
below −35 ºC (entries 3ꢀ5). There was no benefit when the reaction was started at low
ACS Paragon Plus Environment
19

Organic Process Research & Development
Page 20 of 30
1
2
3
4
5
6
7
8
9
temperature and then warmed (entry 6). To examine the product stability under reaction
conditions, a time course to monitor the yield of reactions was performed. At −10 to 0 °C the
yield at 2 h was significantly lower than it was at 0.5 h but at below −20 °C the yield did not
change with extended reaction time. Therefore, a temperature range of −20 to −30 °C is selected
for this reaction.
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 3: Temperature Effect on BCl3 Reaction
Entry
Conditionsa
−15 oC, 45 min
−25 oC, 30 min
−35 oC, 45 min
−45 oC, 45 min
−55 oC, 45 min
−70 to −25 oC, 80 min
11ab yield
81%
10ac
0%
1
2
3
4
5
6
80%
1.6%
11%
28%
34%
10%
73%
63%
55%
77%
a. Crude starting material was used. After the ring opening reaction, 18a was converted to 11a
by TFA treatment.
b. Yield was determined by HPLC assay analysis.
c. Amount of SM after BCl3 treatment.
Although 11a is stable toward acid, lower recoveries of 11a were obtained if the reaction
mixture was exposed to acidic conditions, e.g., when the reaction was quenched with water the
yield of 11a was low. This indicated that the intermediate before quench, presumably a borate
complex, is susceptible to the acid generated from reaction of BCl3 and water. A stress study in
which aqueous potassium phosphate tribasic was added over an extended period also resulted in
low yield. To ensure a basic environment during the work up a reverse quench was instituted for
manufacturing. To our satisfaction, addition of the reaction mixture into an aqueous potassium
phosphate tribasic solution at 30 to 40 ºC consistently gave high yields.
ACS Paragon Plus Environment
20

Page 21 of 30
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
The quality of BCl3 is critical for this reaction. We have encountered very low yield (< 30%) of
11a with older BCl3 solution presumably due to the HCl generated from the hydrolytic
decomposition of BCl3. To confirm this hypothesis HCl in dioxane was added to 10a/b in
anhydrous DCM solution and this indicated that 10a/b decomposed quickly to many low level
impurities. Other acids (i.e, TFA, MsOH, PPh3·HBr) also decomposed 10a/b under anhydrous
conditions. High quality BCl3 solution in DCM is not commercially available on large scale and
therefore must be freshly prepared from neat BCl3 and used immediately. We tested other BCl3
solutions (toluene and heptane) without success. As a cost effective and practical alternative, we
developed the direct use of vaporized BCl3 from a cylinder. In contrast to BCl3 in DCM solution
neat BCl3 in a cylinder is stable, readily available in large quantities and easy to handle during
manufacturing. It is imperative that the BCl3 should be condensed in the head space of the
reactor. If it is bubbled into the reaction mixture the locally excess BCl3 may cause product
decomposition.
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
With control of the temperature, equivalents of BCl3, and using neat BCl3, yields of ~ 90%
have been consistently achieved at laboratory and manufacturing scales.
Step 5. TBS Protection
Scheme 6. TBS Protection
ACS Paragon Plus Environment
21

Organic Process Research & Development
Page 22 of 30
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
The reaction solution from the previous step could be used directly in the TBS protection of the
tertiary alcohol within 11a (Scheme 6). If β,Ƴꢀenone 19a content was >2%, triethylamine was
added before silylation to isomerize 19a to desired 11a. The reaction of 11a with TBSOTf in the
presence of NEt3 gives some silyl enol ether 20 as a sideꢀproduct but by controlling the reaction
temperature at 0 to 10 ºC, impurity 20 could be controlled at ~1%. Residual 18a and 19a are
converted to 21 and 22 respectively. Additionally, we discovered that 11b does not react with
TBSOTf under these conditions presumably due to the steric hindrance of the molecule and can
be removed by a silica gel filtration. Indeed no reaction occurred when pure 11b was treated with
excess TBSOTf and 2,6ꢀlutidine at room temperature. In our earlier manufacturing, the crude
enone was purified by a two stage purification process. Unreacted 11b and other polar impurities
are first removed by a short column using 5x silica gel followed by removal of residual 2,6ꢀ
lutidine and less polar impurities including 20, 21 and 22 by a reꢀslurry in heptane. The desired
product is isolated as a white to offꢀwhite solid with greater than 99% HPLC purity. In our
earlier development manufacturing up to 26 kg scale of 1, chromatography was needed but a
method to avoid the cumbersome and costly procedure was desired.
ACS Paragon Plus Environment
22

Page 23 of 30
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
After optimization of all steps, crude 1 after work up was generally > 90% pure on HPLC and
this crude material could be purified by a simple silica gel filtration followed by methanol/water
slurry to give 1 with > 98% HPLC purity. However, in many instances the product was darker
and had baseline impurities by TLC. After screening we found that the baseline TLC spot and
color could be removed by treatment of the crude material with charcoal and silica gel in toluene.
As a switch of solvent from toluene to methanol is difficult and azeotropic distillation of toluene
with IPA is effective we investigated using IPA instead of MeOH for the crystallization of 1. We
found that enone 1 has better solubility in IPA at high temperature (70 to 80 °C) and a true
recrystallization in IPA/water yielded better quality material than a methanol/water slurry. After
a combination of silica gel/charcoal preꢀtreatment and IPA/water recrystallization, the material is
generally >99% pure. The major impurities 20, 21 and 22 are normally completely purged to the
mother liquor. To maximize the benefit of toluene, which requires less TBSOTf in the protection
step compared to DCM presumably because toluene distillation removes some volatile
impurities, and streamline the process, we designed the process as follows: the DCM solution
from the previous step is distilled to a minimum volume, toluene is charged and the residual
DCM is removed by azeotropic distillation. To the toluene solution, 0.1 equiv of triethylamine is
added to facilitate the conversion of 19a to 11a. After isomerization, silylation is performed by
adding 2,6ꢀlutidine and TBSOTf at 0 to 10 °C. After aqueous work up, the toluene solution is
slurried with 1.5x silica gel and 1x charcoal to remove color and baseline impurities. After
removal of charcoal and silica, the filtrate is distilled and the residual toluene is removed by
azeotropic distillation with IPA resulting in a clear hot IPA solution at 70 to 80 °C. To this hot
solution 20% (volume relative to that of IPA) of water is added followed by cooling to 0 °C. The
desired product is collected by filtration. The loss to the mother liquor is generally <5%.
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
23

Organic Process Research & Development
Page 24 of 30
1
2
3
4
5
6
It is worth noting that when 10a/b and enone 1 are purified by recrystallization, the chiral
purity is enhanced and the chiral purity of enone 1 is consistently close to 100%.
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
Conclusion:
We have developed an improved, robust process for manufacture of enone 1, a key
intermediate for synthetic tetracyclines, and demonstrated the process by manufacturing of 76.1
kg of enone 1 with 56% overall yield (Step 1, 95.2%; Step 2/3, 73.9%; Step 4, 90.1%; Step 5,
88.4%). Key improvements include: 1) mild coupling conditions using TMP.LiCl.MgCl as base
with 3.5 :1 stereoselectivity; 2) shorter DielsꢀAlder reaction using DMSO as solvent with higher
endo to exo selectivity, allowing the resulting reaction mixture to be used in subsequent step as a
one pot procedure; 3) a mild doseꢀcontrol oxidation condition and a recrystallization process
with high recovery; 4) tighter control of the demethylation/ring opening reaction and subsequent
conversion of diene to desired product to ensure high yield; 5) using neat BCl3 from a cylinder
resulting in much lower material cost as well as a simple operation; 6) an IPA/water
crystallization process to provide high quality enone 1 with minimal product loss. These
improvements enabled a synthetic sequence with no column chromatography.
Experimental Section
General. The experiments outlined below provide representative procedure of enone 1
1
synthesis for what was run on scale in the pilotꢀplant facilities. H NMR spectra were recorded
on a JEOL ECXꢀ400 400 MHz spectrometer. Achiral HPLC analyses were performed on an
Agilent SunFire C18 (150 mm × 4.6 mm, 5 ꢂm) column with 10–100% CH3CN/H2O (+0.1%
ACS Paragon Plus Environment
24

Page 25 of 30
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
TFA) as mobile phase over 21 min at flow rate of 1.5 mL/min. Chiral HPLC analysis were
performed on Shimadzu IA (250 mm × 4.6 mm, 5 ꢂm) column using Hex/IPA/Ethansulfonic
acid (30:70:0.14) as mobile phase over 15 min at flow rate of 0.5 mL/min.
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
Free base 6. The LꢀTartrate of compound 615 (115 kg, 281.6 mol, 1.0 equiv) and water (610
kg) were added to a 3000 L reactor and the mixture was cooled to 0ꢀ10 °C. To the mixture 6 M
NaOH aqueous solution (96 kg, 563.2 mol, 2.1 equiv) was added over a period of 1 h while
maintaining the temperature at 0ꢀ10 °C. After addition the pH of the mixture was approximate 10
indicated by pH paper. After stirring for another 10 min, the mixture was extracted with toluene
twice (472 kg and 196 kg). The combined organics were washed with water (196 kg). The
toluene solution was distilled under reduced pressure (< −0.085MPa) at below 65 °C to give 86.6
kg of product as a toluene solution. The water content was measured at 0.05% by KF.
Coupling Product 8a/b. The above toluene solution of 6 (281.6 mol, 1.0 equiv) was dissolved
in THF (226 kg) and then the mixture was cooled to −20 to −30 °C at which temperature
TMPMgClꢁLiCl (320 kg, 309.6 mol, 1.10 equiv) was charged over a period of 1 h. After
addition, the mixture was stirred for another 90 min at which time IPC indicated that ~7.4% of 6
were not deprotonated. Additional TMPMgClꢁLiCl (24.0 kg, 23.2 mol, 0.08 equiv) was added
over 15 min and then the reaction mixture was stirred for another 30 min at −20 to −30℃. IPC
deemed the deprotonation was complete. The mixture was cooled to between −70 and −80 °C at
which temperature a solution of compound 7 (42.5 kg, 337.0 mol, 1.2 equiv) in THF (240 kg)
was added. After addition the reaction was stirred for 2 h and then was gradually warmed up to
−15°C over a period of 8 h and stirred at −15 °C for another 2h. The reaction was quenched with
ACS Paragon Plus Environment
25

Organic Process Research & Development
Page 26 of 30
1
2
3
4
5
6
7
8
9
670 kg of saturated NH4Cl. The mixture was filtered through a layer of celite. The celite was
washed with 200 kg of water followed by 396 kg of isopropyl acetate. The organic layer was
separated and the aqueous layer was extracted with 302 kg of isopropyl acetate. The combined
organics were washed sequentially with saturated NH4Cl twice (400 kg each), water (268 Kg)
and brine (266 Kg). The resulting isopropyl acetate solution was concentrated under reduced
pressure at below 50 °C until the water content was 0.06% by KF. During the distillation
additional 178 kg of isopropyl acetate was added. An isopropyl acetate solution (164 kg with a
HPLC assay of 62.86% corresponding to 103.1 kg of 8a/b and 95.2% yield) was obtained. The
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
ratio of 8a to 8b was 3.57:1 as determined by HPLC (74.3% of 8a and 20.8% of 8b). The H
NMR spectral data were consistent with those reported.4
Ketone 10a. The above 8a/b solution in isopropyl acetate (103.1 kg of 8a/b, 268.1 mol, 1.0
equiv), DMSO (526 Kg, KF=0.05%), diisopropylethylamine (39.1 kg, 294.9 mol, 1.1 equiv) and
BHT (930 g, 2.7 mol, 0.01 equiv) were mixed in a 2000 L reactor. After the mixture was purged
with argon it was stirred at 90 °C for 24 h and then at 95 ℃ for 24 h at which time HPLC analysis
deemed the DielsꢀAlter reaction was complete. The reaction mixture was cooled to room
temperature and diluted with ethyl acetate (412 kg, KF=0.02%) followed by
diisopropylethylamine (138 kg, 1067 mol, 4.0 equiv). Sulfur trioxide pyridine complex (120.6
kg, 758.7 mol, 2.8 equiv) was added in portions at below 30 °C over a period of 1.5 h. After the
addition, the mixture was stirred at 20 to 30 ℃ for 16 h at which time the reaction was deemed
complete by HPLC analysis. Water (310 kg) was added over a period of 1 h at below 30 °C to
quench the reaction. The mixture was transferred to a bigger reactor and additional water (814
kg) and ethyl acetate (690 kg) were added. The mixture was filtered through a layer of celite and
ACS Paragon Plus Environment
26

Page 27 of 30
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
the celite was washed with ethyl acetate (484 kg). The organic layer was separated and the
aqueous layer was extracted with ethyl acetate twice (536 kg and 236 kg). The combined
organics were washed with water (432 kg) and brine (432 kg), and then filtered through a 60.0
kg silica plug. The silica gel was washed with 588 kg of ethyl acetate. The crude solution was
concentrated to about 300 L under reduced pressure (< −0.085 MPa) at below 50 ℃. To the
residue toluene (640 kg) was added and the resulting solution was again distilled to ~ 300 L. The
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
residue was diluted with 160 kg of toluene. H NMR determined that the ratio of toluene to 10a
was approximately 4.5 L per 1 kg based on the integration of CH3 groups in toluene and 10a.
Additional toluene (50 Kg) was added and then the toluene solution was heated to 70ꢀ80 ℃ at
that temperature heptane (742 kg, approximately 10x volume of theoretical 10a/b) was added
over a period of 3 h. After addition, the mixture was cooled slowly to 10ꢀ20 ℃ at which
1
temperature it was stirred for 12 h. Sample was analyzed by H NMR which indicated a 5.4:10
(vol/vol) ratio of toluene and heptane. The slurry was filtered, washed with heptane twice (82 kg
each) and dried under vacuum at 40ꢀ50 ℃ to afford 79.14 kg of 10a as a light brown solid with a
HPLC assay of 95.77 % corresponding to 75.8 kg and 73.9% yield). HPLC showed 99.1% of
10a and 0.9% of 10b.The 1H NMR spectral data were consistent with those reported.4
Unprotected enone 11a. In a 500 L cryogenic reactor, the above 10a (39.5 kg solution with an
assay of 95.77% corresponding to 37.8 kg of 10a, 98.9 mol, 1.0 equiv) was dissolved in 380 kg
of anhydrous DCM. The solution was cooled to −50 to −40 ℃. At this temperature vaporized
neat BCl3 (13.2 Kg, 1.14 equiv) from a cylinder was directly conducted into the reactor head
space and condensed/dissolved into the reaction mixture. After addition, the mixture was
warmed to −30 to −20 ℃ and then stirred at this temperature for 15 minutes at which time the
ACS Paragon Plus Environment
27

Organic Process Research & Development
Page 28 of 30
1
2
3
4
5
6
7
8
9
reaction was deemed complete by HPLC. The reaction mixture was transferred into a reactor
containing 502 kg of 20% K3PO4 aqueous solution while maintaining the temperature at 35 to
40 ℃. The mixture was cooled to 15ꢀ25 ℃ and then combined with another reaction mixture
with the same batch size for workup.
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 organic layer was separated and the aqueous layer was extracted with 180 kg of DCM. The
combined organics were washed with 460 kg of brine. To the DCM solution was added
TsOH.H2O (75.6 kg, 397.4 mol, 2.0 equiv) and then stirred at 15ꢀ25 ℃ for 5 h at which time
HPLC showed all diene 18a was consumed. To the mixture was added 910 kg of 20% K3PO4
solution to adjust the pH to > 8. The organic layer was separated and the aqueous layer was
extracted with 182 kg of DCM. The combined DCM solutions were washed with 552 kg of brine
and distilled to give 644 kg solution of 11a with a HPLC assay of 10.2% corresponding to 65.7
kg of 11a and 90.1% yield. HPLC purity: 96.5%. The 1H NMR spectral data was consistent with
those reported.4
TP-808 (Enone 1). The above solution of 11a (65.7 kg of 11a, 178.3 mol, 1.0 equiv) was
distilled to a minimum volume. To the residue toluene (210 kg) was added and the mixture was
distilled under reduced pressure (< −0.085MPa) at below 60 °C to a minimum volume. The
distillation cycle was repeated until the water content of the residue was 0.02%. To the residue
390 kg of toluene was added followed by triethylamine (1.96 kg, 0.1 equiv). The mixture was
stirred at 15ꢀ25 ℃ for 2 h and then was cooled to 0 to10 ℃. At this temperature 2,6ꢀlutidine (38
Kg, 354.8 mol, 2.0 eq) was added followed by TBSOTf (76.5 Kg, 288.8 mol, 1.6 equiv). After
stirring for 30 min at 0 to10 ℃ the reaction was deemed complete by HPLC and was quenched
with water (312 kg). The two layers were separated. The aqueous layer was extracted with
ACS Paragon Plus Environment
28

Page 29 of 30
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
toluene (143 kg). The combined toluene layers were washed with brine (390 kg). To the toluene
solution was added silica gel (117 kg) and charcoal (78 kg). The slurry was stirred at 20ꢀ30℃ for
3 h and filtered. After filtration, the filter cake was washed with toluene (977 kg) until TLC
showed no product remained on the filter cake. The combined organic solutions were distilled
under reduced pressure (< −0.085MPa) at below 60 ℃ to approximately 100 L. To the residue
326 kg of isopropyl alcohol was added and again the solution was distilled to ~100 L. This
process was repeated for another cycle. To the residue 327 kg of IPA was added and the solution
was analyzed by GC which indicated 0.14% of residual toluene. The mixture was heated to
75 °C and the solution was analyzed by 1H NMR which determined the ratio of IPA and enone 1
was approximately 5.3 L per kg. Additional IPA (272 kg, 4.5 V) was added to make up a
solution with 10x volume of IPA assuming 90% yield. The mixture was heated to 75ꢀ85 ℃, and
to the resulting clear solution water (156 kg, 2.0 V) was added. The solution was cooled to 0 to
5 ℃ over a period of 6 h at which temperature the slurry was stirred for another 2 h. The slurry
was filtered, washed with IPA/water twice (54 kg each, V/V=5:1), and dried in a vacuum oven at
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
45ꢀ55 ℃ for 36 h to give 76.1 kg (88.4 % yield) of enone 1. The H NMR spectral data were
consistent with those reported.4 HPLC purity: 100%; Chiral HPLC purity: 100%; LC/MS:
[M+H]+ 483.2; Anal. Calcd for C26H34N2O5Si: C, 64.70; H, 7.10; N, 5.80. Found: C, 64.77; H,
7.29; N, 5.82.
A total of 936 kg mother liquor was obtained with an assay of 0.31%. The loss to the mother
liquor is 2.9 kg.
Acknowledgement:
ACS Paragon Plus Environment
29