Process Development of Tryptophan Hydroxylase Inhibitor LX1031, a Drug Candidate for the Treatment of Irritable Bowel Syndrome

Article abstract of DOI:10.1021/acs.oprd.9b00520

Two process routes for LX1031, a tryptophan hydroxylase inhibitor for the treatment of irritable bowel syndrome, were developed. They shared the same left-hand and right-hand starting materials as well as the penultimate intermediate. The chiral center in

Full text of DOI:10.1021/acs.oprd.9b00520

Subscriber access provided by UNIV OF NOTTINGHAM
Full Paper
Process Development of Tryptophan Hydroxylase Inhibitor LX1031,
a Drug Candidate for the Treatment of Irritable Bowel Syndrome
Matthew Zhao, Haiming Zhang, Shinya Iimura, Mark Bednarz,
Ramanaiah Kanamarlapudi, Jie Yan, Ngiap-Kie Lim, and Wenxue Wu
Org. Process Res. Dev., Just Accepted Manuscript • DOI: 10.1021/acs.oprd.9b00520 • Publication Date (Web): 20 Jan 2020
Downloaded from pubs.acs.org on January 20, 2020
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 44
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
Process Development of Tryptophan Hydroxylase
Inhibitor LX1031, a Drug Candidate for the Treatment
of Irritable Bowel Syndrome
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
Matthew M. Zhao,* Haiming Zhang,† Shinya Iimura,‡ Mark S. Bednarz,§ Ramanaiah C.
††
†
Ѱ
Kanamarlapudi, Jie Yan, Ngiap-Kie Lim, Wenxue Wu*
Chemical Development, Lexicon Pharmaceutical, Inc. 110 Allen Road, Basking Ridge, New Jersey
07920, USA
ACS Paragon Plus Environment

Organic Process Research & Development
Page 2 of 44
1
2
3
4
5
6
TOC Graphic
7
8
O
O
9
OH
NHBoc
MeO
OMe
NHBoc
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
O
O
OH
CF3
B
HO
O
OH
Cl
NHBoc
Right-to-Left
(First-Generation Process)
N
N
O
NH2
MeO
OH
Cl
Cl
O
NH2
N
N
CF3
N
N
NH2
LX1031
NH2
Left-to-Right
(Second-Generation Process)
MeO
O
Cl
O
CF3
N
N
Br
OH
NHBoc
MeO
NH2
O
O
B
OH
CF3
CF3
O
ACS Paragon Plus Environment

Page 3 of 44
Organic Process Research & Development
1
2
3
4
5
6
7
ABSTRACT
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
Two process routes for LX1031, a tryptophan hydroxylase (TPH) inhibitor for the treatment of irritable
bowel syndrome (IBS), were developed. They shared the same left-hand and right-hand starting
materials as well as the penultimate intermediate. The chiral center in the left-hand moiety was
established via a Noyori asymmetric hydrogenation of a trifluoromethyl aryl ketone. The right-hand
boronate was prepared via a palladium-catalyzed borylation of L-tyrosine derived aryl triflate. Union
of these two fragments to the pyrimidine core, from the right- or left-hand side, constituted the first-
and second-generation routes, respectively. Removal of the Boc-protecting group from the penultimate
intermediate gave LX1031. The challenges overcome in purification and isolation of the LX1031
zwitterion are also discussed. Both process routes were successfully performed on multi-kilogram
scales to supply LX1031 API for the preclinical and clinical studies.
KEYWORDS: Asymmetric hydrogenation, trifluoromethyl carbinol, Suzuki, SNAr, zwitterion
ACS Paragon Plus Environment

Organic Process Research & Development
Page 4 of 44
1
2
3
4
INTRODUCTION
5
6
7
8
Irritable Bowel Syndrome (IBS) is characterized by abdominal pain, cramping and changes in bowel
functions such as constipation, diarrhea, and alternating symptoms. IBS affects approximately 25–45
million Americans who have limited therapeutic options.1 Overexpression of serotonin
(5-hydroxytryptamine or 5-HT) in the GI tract is associated with IBS diarrhea.2 Serotonin is also
involved in important biological functions in the central nervous system (CNS). Existing IBS
treatments targeting serotonin receptors often impart undesirable CNS side effects, affecting mood,
appetite, and other behavioral functions. Tryptophan hydroxylase (TPH), an enzyme that catalyzes the
rate-limiting step in the biosynthesis of serotonin, exists in two distinct isoforms, TPH1 and TPH2.
Selective inhibition of TPH1, which is expressed primarily in the GI tract, can potentially provide an
effective treatment of IBS without affecting serotonin levels in the brain3 thereby minimizing
potential CNS side effects. In animal models, a locally acting TPH1 inhibitor LX1031, has been shown
to reduce 5-HT levels in the GI but not in the brain.4
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 medicinal chemistry route4,5 for LX1031 started with reduction of trifluoromethyl aryl ketone
1 with NaBH4 to give racemic alcohol (±)2 (Scheme 1). Kumada coupling of the latter with
m-methoxyphenyl magnesium chloride gave racemic biaryl alcohol (±)3. Deprotonation of (±)3 with
NaH followed by an SNAr substitution on 4,6-dichloro-2-aminopyrimidine (DCAP, 4) gave
pyrimidine chloride (±)5. Palladium catalyzed Suzuki coupling of (±)5 with 4-borono-L-phenylalanine
(L-BPA, 6) afforded a mixture of LX1031 and a diastereomer.
While this synthesis was relatively concise, several significant challenges had to be overcome before
scale-up. First, chiral resolution at the final stage by preparative HPLC or other means was inherently
inefficient, so an enantioselective synthesis of chiral alcohol 2 or 3 was desired. Second, NaH must be
replaced with a process-friendly base for the SNAr substitution. Third, L-BPA (6) was costly
ACS Paragon Plus Environment

Page 5 of 44
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
($10,000/100 g) and not readily available in kilogram quantities. A practical synthesis of L-BPA or its
surrogate had to be developed. Additionally, the microwave heating (150 °C) used for the Suzuki-
coupling of pyrimidine chloride 5 and L-BPA (6) was not amenable to scale-up. Finally, it was
desirable to identify a crystalline form of LX1031 for isolation and purity upgrade.
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 Medicinal Chemistry Route to LX1031
Br
Br
MgBr
Pd(PPh3)4
MeO
NaH
N
NaBH4
MeO
O
OH
*
OH
*
Cl
Cl
CF3
CF3
CF3
N
1
2
3
(±)
(±)
4
NH2
O
OH
O
NH2
HO
B
OH
MeO
MeO
6
OH
NH2
O
Cl
O
*
*
Pd(PPh3)2Cl2
THF/H2O, Na2CO3
Microwave, 150 °C
CF3
N
N
NH2
CF3
N
N
NH2
5
(±)
LX1031 + diastereomer
For the enantioselective reduction of the trifluoromethyl aryl ketone (1), Noyori asymmetric
hydrogenation6 is one of the most frequently used methods in the literature. Replacing sodium hydride
with a more process-friendly base for the SNAr substitution should also be feasible.7 However,
preliminary screening experiments indicated that it could be challenging for the Suzuki coupling of 5
and 6 to go to completion without microwave heating. Therefore, an alternative synthetic approach
(right-to-left) involving Suzuki-coupling with a more active substrate, DCAP (4), was explored
(Scheme 2). In this case, an L-BPA surrogate, compound 9 was synthesized via borylation of the aryl
triflate prepared from commercially available N-Boc-tyrosine methyl ester.8 An added benefit of this
approach was that intermediate 10 could be used for synthesizing other potential drug candidates being
pursued at the time.
ACS Paragon Plus Environment

Organic Process Research & Development
Page 6 of 44
1
2
3
4
5
6
7
Scheme 2 First Generation Process Approach (Right-to-left)
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
Cl
Cl
O
O
N
N
O
OH
OH
NHBoc
4
NH2
Cl
NHBoc
OMe
O
B
NHBoc
N
N
HO
O
O
NH2
8
9
10
OH
MeO
NH2
O
CF3
N
N
Br
Br
B(OH)2
NH2
MeO
11
MeO
O
OH
CF3
LX1031
OH
CF3
CF3
1
2
3
ACS Paragon Plus Environment

Page 7 of 44
Organic Process Research & Development
1
2
3
4
RESULTS AND DISCUSSION
5
6
First-Generation Process Route (Right-to-Left)
7
8
9
Preparation of boronate 9: For the synthesis of the right-hand fragment of LX1031, borylation9 of
L-tyrosine triflate 12 appeared the most straightforward (Scheme 3). Triflate 12 was prepared by
reaction of Boc-L-tyrosine with triflic anhydride. Attempts to replace pyridine used in the literature10
with K2CO3 was unsuccessful (no reaction). On the other hand, the reaction proceeded cleanly and
rapidly with either triethylamine or N-methylmorpholine (NMM) as the base, reaching completion
immediately after the addition of triflic anhydride at ‒10 °C in DCM. NMM was selected for easier
removal of colored materials by aqueous citric acid washes. The aryl triflate (12) was surprisingly
stable in the organic layer, showing no appreciable degradation when stirred with aqueous citric acid
for three days at rt. Being a low-melting solid, it was not isolated but directly used in the next step
after solvent swap to acetonitrile. The acetonitrile solution of 12 also proved to be stable for many
months at ambient temperature.
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
Borylation of 12 by treatment with bis(pinacolato)diboron (pin2B2) worked well in acetonitrile with
Pd(dppf)Cl2 as the catalyst and two equivalents of KOAc as the base. A small amount of coupling
product of dppf with 12 was observed (Figure 1). This issue was resolved by switching to
Pd(OAc)2/PCy3 catalyst. The reaction mixture was cooled and the precipitated byproducts were
filtered off. After solvent swap to MTBE, the methyl ester was hydrolyzed in-situ to the carboxylic
acid 9 by treatment with aqueous LiOH at 0 °C for 1 hour. This was to eliminate the potential risk of
racemization under basic conditions downstream.
In the initial scale-up of LX1031 (1.4 kg), the MTBE solution of compound 9 was used directly in
the next step after solvent swap to ethanol. In subsequent campaigns, boronate 9 was crystallized from
heptane/MTBE mixture in ~90% yield and with > 99% purity.
ACS Paragon Plus Environment

Organic Process Research & Development
Page 8 of 44
1
2
3
4
Scheme 3 Synthesis of pyrimidine chloride 10
5
6
7
8
O
O
Tf2O, NMM
OMe
OMe
NHBoc
NHBoc
DCM, -10 to 0 °C
TfO
HO
8
12
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
1. Pd(OAc)2, PCy3
Pin2B2, KOAc, ACN
2. aq. LiOH
O
O
Pd(PPh3)2Cl2
EtOH/THF/H2O
KHCO3, 70 °C
OH
NHBoc
OH
Cl
NHBoc
O
B
Cl
Cl
N
N
O
N
N
NH2
10
9
NH2
O
OMe
NHBoc
PPh2
PPh2
Fe
13
Figure 1 DPPF adduct
Synthesis of pyrimidine chloride (10) As expected, the Suzuki-coupling of aryl boronate 9 with
DCAP was indeed facile. With 1 mol% of Pd(PPh3)2Cl2 as the catalyst and KHCO3 as the base, the
reaction proceeded smoothly at 70 °C in 5/1 EtOH/water (Scheme 3). However, due to the high
reactivity of DCAP, several impurities such as bis-coupling product 14, amination impurity 15, and
hydrolysis/solvolysis of DCAP, 16a/16b, were generated at significant levels (Figure 2). For example,
with 1.2 equivalents of DCAP, bis-coupling product 14 was about 20%. The formation of this impurity
was suppressed to ∼ 6% by increasing DCAP to 1.8 equivalents, albeit requiring a high solvent volume
(20 vol) to fully dissolve DCAP. Unfortunately, the increased charge of DCAP led to increased
formation of amination impurity 15 (~ 4%) particularly at the end of the reaction. In fact, it increased
to 10% in 15 h in a stress experiment. Therefore, the reaction mixture should be cooled promptly at
the end of the reaction. DCAP solvolysis product 16b, typically 2% at the end of the reaction, also
ACS Paragon Plus Environment

Page 9 of 44
Organic Process Research & Development
1
2
3
4
5
6
increased to 10% during the stress test. This impurity was less of a concern since it was readily purged
in the workup and isolation.
7
8
9
Table 1 Impurities vs DCAP stoichiometry and age time
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
DCAP
(equiv)
Time
(h)
14
(A%)
15
(A%)
Entry
1
2
3
4
1.2
1.5
1.8
1.8
5
5
5
20
< 1
5.5
6
2.5
4
20
6
10
*1 mol% Pd(PPh3)2Cl2, 3 equiv KHCO3, 5/1 EtOH/H2O (∼20 vol), 75–80 °C.
O
O
HO
OH
BocHN
NHBoc
N
N
NH2
14
O
OH
NHBoc
OR
Cl
N
Cl
N
N
N
NH2
NH
Cl
N
N
R:
H
16a
16b
:
NH2
15
:
Et
Figure 2 Impurities in Suzuki coupling
Early in development, isolation of 10 was a challenge due to lack of any known crystalline forms.
After completion of the reaction, the mixture was concentrated to precipitate most of the excess DCAP
for removal by filtration. The filtrate was extracted with EtOAc to remove residual DCAP and then
treated with Darco-G60 to remove residual palladium. The product was then precipitated by
acidification with aqueous citric acid. However, this operation proved problematic in the initial 600 g
scale-up due to formation of sticky mass at pH 5.5–6.0 which blocked the stirring. Although the sticky
mass eventually hardened at pH 4 allowing it to be crushed, filtered, washed, and dried, the purity of
ACS Paragon Plus Environment

Organic Process Research & Development
Page 10 of 44
1
2
3
4
5
6
7
8
9
the isolated 10 was only 76% and residual palladium level was >1000 ppm. This problem was
remedied by slowly adding the crude aqueous solution into a citrate buffer maintained at pH 4 by
concomitant addition of aqueous citric acid to give a free-flowing suspension. Additionally, improved
residual palladium control was achieved by adding (n-Bu)3P in the EtOAc wash. These improvements
were implemented in the next scale-up batch affording 4.65 kg of 10 (∼60% yield corrected). The
main impurities were diacid 14 (5.6%) and amination impurity 15 (8.9%). The higher than typical
level of 15 was attributed to the longer reaction time (12 h vs < 5 h).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
While the diacid (14) was inert and mostly purged in the next step in the workup, 15 was expected
to propagate to new impurities, increasing impurity load in the purification and isolation of LX1031.
Thus, additional catalyst screening focused on minimizing the formation of 15. Among the catalysts
screened (PEPPSI and CombiPhos’s POPd, PXPd catalyst families), POPd6 was the most promising,
controlling 15 to 0.5–1%. However, diacid 14 was significantly higher than with Pd(PPh3)2Cl2 (10%
vs 5%). By increasing DCAP charge to 3 equivalents, 14 was controlled to ∼8%. Meanwhile, a
crystalline toluene solvate of 10 was discovered and utilized for its isolation in two scale-up batches
(25 kg and 40 kg scale) affording 10 with ∼88% purity in 50–60% yield (corrected for ∼75% w/w
assay, 10–15% w/w toluene). Amination impurity 15 was ∼0.5%, but the diacid 14 remained relatively
high at ∼8%.
While POPd6 was successful in minimizing amination impurity 15, the level of diacid 14 was rather
high, wasting significant amounts of valuable boronate 9. Therefore, other catalyst systems including
Pd(OAc)2/P(Cy)3, Pd(OAc)2/PPh3, and Pd(dppf)Cl2 were evaluated (Table 2). Even higher level of
bis-coupling product 14 (18% vs 8.8%) was observed with Pd(OAc)2/P(Cy)3 as the catalyst (entry 2).
Pd(dppf)Cl2 was less active, reaching only 41% conversion after aging for 15 h at 75–80 °C (entry 3).
Pd(OAc)2/PPh3 gave the most promising results and was selected for further optimization (entry 4).
ACS Paragon Plus Environment

Page 11 of 44
Organic Process Research & Development
1
2
3
4
Table 2 Alternative catalysts to control diacid impurity 14
5
6
7
8
Entry
Catalyst
POPd6
Reaction Time Conversion 14 (%) 15 (%)
1
15 h
10 h
15 h
10 h
100
100
41
8.8
17.8
--
0.55
1.1
--
Pd(OAc)2/P(Cy)3
(1 mol%/2 mol%)
9
2
3
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
Pd(dppf)Cl2
Pd(OAc)2/PPh3
(1 mol%/2 mol%)
100
4.3
2.5
*1 mol% Pd, 5/1 EtOH/H2O (20–35 vol), 75–80 °C
It was observed that stronger base such as K3PO4 significantly increased the formation of amination
impurity 15 (Table 3, entry 1). The Suzuki-coupling reaction was completely shut down in acetonitrile
(entry 3). The reaction was also very slow in THF or acetone perhaps due to the lower refluxing
temperatures (entry 4, 5). The reaction proceeded well in DME, dioxane, or a dioxane/EtOH mixture
(entry 6–8). But DME and dioxane are not desirable solvents due to safety and environmental concerns.
The reaction performed better in THF/EtOH mixtures than EtOH alone, controlling the diacid 14 and
amination impurity 15 to < 4% and < 2.5%, respectively (entry 9, 10). The reaction mixture contained
∼90% area 10 (excluding DCAP and DCAP-derived impurities 16a/16b) under the optimized reaction
conditions (Pd(OAc)2/PPh3, 3.5 equiv KHCO3, THF/EtOH/water, 75 °C).
Table 3 Optimization of Suzuki-coupling with Pd(OAc)2/PPh3 catalyst
Entry
Solvent
EtOH
Base
K3PO4
Temp, time
Conv (%) 14 (%) 15 (%)
1
75–80 °C, 15 h 100
3.8
3.8
--
8.2
3.5
--
2
3
4
5
6
7
8
EtOH
KHCO3
KHCO3
KHCO3
KHCO3
KHCO3
KHCO3
75–80 °C, 5 h
75–80 °C 20 h
Reflux, 8 h
99
0
Acetonitrile
THF
26
35
--
--
Acetone
DME
Reflux, 15 h
--
--
75–80 °C, 20 h 93
75–80 °C, 15 h 98
3.1
3.7
3.8
2.1
1.9
3.2
Dioxane
dioxane : EtOH KHCO3
75–80 °C 6 h
100
ACS Paragon Plus Environment

Organic Process Research & Development
Page 12 of 44
1
2
3
4
(1 : 1)
5
6
7
8
THF : EtOH
(7 : 3)
9
KHCO3
KHCO3
67–70 °C, 9 h
99
3.2
3.5
2.3
2.1
THF : EtOH
(3 : 7)
10
70–73 °C, 14 h 95
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
* ∼20 vol total solvent, 0.75 mol% Pd(OAc)2, 1.5 mol% PPh3, 75–80 °C or reflux
The reaction mixture was concentrated to a low volume, diluted with water to precipitate most of
the excess DCAP for removal by filtration. The aqueous filtrate was washed with EtOAc twice to
remove residual DCAP. After adjusting the pH to 2.5–3.5 with 6 N HCl, the product was extracted
into THF/toluene. The organic layer was treated with activated carbon and solvent swapped to toluene
to crystallize 10 as a toluene solvate in ∼65% yield with ∼95 A% purity (∼2% 14, ∼0.5% 15). On
multiple 50-kg batches, residual palladium levels in most batches were higher (> 100 ppm) than lab
batches perhaps due to more efficient removal of THF during solvent swap. Recrystallization from
THF/toluene (3.2 : 23 vol) reduced residual palladium to < 100 ppm and improved its purity to ∼97%
with ∼90% recovery.
Synthesis of biaryl chiral alcohol 3. For the left-hand fragment, the biaryl chiral alcohol (3), the
most straightforward synthesis was via asymmetric reduction of trifluoromethyl ketone 1 followed by
Suzuki-coupling with 3-methyoxyphenylboronic acid (11) (Scheme 4). Trifluoromethyl ketone 1 was
prepared from 1,4-dibromobenzene according to literature procedures.11 Several methods for
asymmetric reduction of ketone 1 were evaluated and the results summarized in Table 5.
(S)-Me-CBS12 catalyzed reduction with catecholborane gave chiral alcohol 2 with only 56% ee
(entry 1). Reduction with (+)-DIP-Cl13 afforded significantly higher enantioselectivity (86% ee) (entry
2). However, the reaction was sluggish (3 days) even with three equivalents of DIP-Cl. Increasing the
reaction temperature to 60 °C decreased the stereoselectivity to 65% and 68% ee in toluene and
heptane, respectively. Noyori asymmetric hydrogenation in IPA using (R)-Xyl-P-Phos RuCl2/(R)-
ACS Paragon Plus Environment

Page 13 of 44
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
DAIPEN catalyst in the presence of a small amount of t-BuOK afforded chiral alcohol 2 in 86% ee
(entry 6). Similarly, Noyori transfer-hydrogenation with Me5Cp*IrCl(R,R-TsDPEN) catalyst gave 2
in 82% ee (entry 7). Though not ideal, the enantioselectivity of the Noyori asymmetric hydrogenation
was satisfactory as the chiral purity of 2 could be readily upgraded to > 99% ee by crystallization from
heptane. The reaction conditions were further optimized and implemented on hundred-kilogram scale,
delivering chiral alcohol 2 in ~56% yield with > 99% ee.
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 biaryl chiral alcohol 3
Br
RuCl2[(R)Xyl-P-Phos] Br
(R)-DAIPEN
>99% ee
crystallized
O
OH
CF3
30 psi H2, IPA, 45 °C
86% ee
CF3
1
2
Pd(PPh3)2Cl2
K2CO3
EtOH/water
OH
B
MeO
OH
11
MeO
OH
CF3
3
Table 4 Asymmetric reduction of ketone 1
Entry
Conditions
2 (e.e.%)
(S)-Me-CBS, catecholborane, THF,
-78 to -30 °C
1
56%
2
3
4
5
3 equiv (+)-DIP-Cl, heptane, 3 days, RT
2.5 equiv (+)-DIP-Cl, heptane, 60 °C, 24 h
2.5 equiv (+)-DIP-Cl, toluene, 60 °C, 24 h
(R,R)-TsDPEN Ru(cymene)Cl
86%
68%
65%
41%
(R)-Xyl-P-Phos RuCl2 (R)-DAIPEN (0.01
6
7
mol%); KOtBu (0.4 mol%), 30 psi H2, IPA, 45 86%
°C
Et3N-HCO2H/Me5CpIrCl
82%
(R,R-TsDPEN)
ACS Paragon Plus Environment

Organic Process Research & Development
Page 14 of 44
1
2
3
4
5
6
7
8
For the aryl-aryl bond formation, Kumada-coupling used in the medicinal chemistry route required
excess Grignard reagent (> 2 equiv) due to the presence of an acidic proton on the chiral alcohol (2).
Therefore, Suzuki-coupling with commercially available 3-methyoxyphenylboronic acid (11) was
explored. With 1.2 equivalents of 11, 3 equivalents of K2CO3 and 0.1 mol% Pd(PPh3)2Cl2 catalyst, the
reaction was facile in EtOH/water (9/1 vol), achieving complete conversion in 2–5 h at 75–80 °C. It
should be noted that the boronic acid (11) must be charged within 1–2 h after the catalyst charge to
minimize catalyst deactivation and de-bromination of 2. The workup and isolation of biaryl chiral
alcohol 3 entailed dilution with water, concentration, extraction with MTBE, washing with 1 N NaOH
(to remove excess boronic acid 11), treatment with Darco-G60, and crystallization by solvent swap to
n-heptane. On a 120-kg batch, the biaryl chiral alcohol (3) was isolated in 93% yield with > 99% purity
and > 99% ee.
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
SNAr reaction of biaryl chiral alcohol 3 with pyrimidine chloride 10. With both pyrimidine chloride
10 and biaryl chiral alcohol 3 in hand, their coupling via an SNAr reaction14 without using NaH was
investigated (Scheme 5).
Scheme 5 SNAr reaction of pyrimidine chloride 10 with chiral biaryl alcohol 3
O
O
OH
NHBoc
MeO
OH
NHBoc
Base
Cl
O
N
N
CF3
N
N
MeO
NH2
NH2
10
OH
CF3
17
3
Organic bases such as DBU and TMG were ineffective (Table 6, entry 1, 2). Potassium carbonate
was also ineffective even in the presence of 20 mol% 18-crown-6 (entry 3). Strong bases such as
sodium t-butoxide (3 equiv) or sodium t-pentoxide (3 equiv) gave complex reaction mixtures partially
ACS Paragon Plus Environment

Page 15 of 44
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
due to cleavage of the Boc-protecting group (entry 4, 5). The most promising result was obtained with
Cs2CO3 in dioxane at 90–100 °C (entries 6–10). However, the reaction was rather slow (15 h) even
with 5 equivalents of Cs2CO3 (entry 6). Adding 20 mol% 18-crown-6 seemed to accelerate the reaction
but with substantially higher cost (entry 7). Decreasing the Cs2CO3 charge to 3 equivalents to maintain
a similar cost led to a much slower reaction (entry 8). Inexpensive phase-transfer agent Bu4NHSO4
(10 mol%) also seemed to accelerate the reaction, achieving complete conversion in 16 h with 3.5
equivalents of Cs2CO3 (entry 9). Milled Cs2CO3 performed substantially better than granular material
(entry 10). Even under the optimized conditions (1.2 equiv 3, 3.5 equiv. milled Cs2CO3, 5 vol dioxane)
the reaction generally required 15–24 h at 90–100 °C. Dimeric urea impurity 18 (Figure 3)15 and
LX1031 (premature cleavage of the Boc-protecting group), were typically observed at 2–5% each at
the end of the reaction. These impurities continued to increase with age time mandating prompt
quenching of the 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 5 SNAr reaction of chiral alcohol 3 with pyrimidine chloride 10
Additive
(mol%)
Time
(h)
Entry Base (equiv)
3
Dioxane
10
Con (%)
1
DBU (5)
TMG (5)
--
--
1.2
20
0
2
3
1.2
1.2
10
10
20
20
0
0
18-crown-6
(20 mol%)
K2CO3 (5)
4
5
6
NaOtBu (3)
--
1.2
1.2
1.2
10
10
10
2
NA*
NaOtPent (3) --
2
NA*
Cs2CO3 (5)
Cs2CO3 (5)
--
15
> 98%
18-crown-6
(20 mol%)
7
1.2
1.2
1.2
1.2
10
10
5
8
> 98%
> 98%
> 98%
> 98%
18-crown-6
(20 mol%)
8
Cs2CO3 (3)
30
16
16
Bu4NHSO4
(10%)
9
Cs2CO3 (3.5)
Cs2CO3 (3.5)
(milled)
10
--
5
ACS Paragon Plus Environment

Organic Process Research & Development
Page 16 of 44
1
2
3
4
*complex reaction mixture due to cleavage of Boc-protecting group.
5
6
7
8
NH2
N
NH2
N
CF3
N
N
CF3
9
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
O
O
O
HO
OH
N
N
H
H
O
O
18
Figure 3 Dimeric urea impurity 18.
A notable operational challenge was the severe gelling of the cesium salts of starting material 10
and product 17 below 70 °C which could completely immobilize the reaction mixture and damage the
agitator. To mitigate this problem, starting material 10 was charged into a suspension of Cs2CO3 and
the biaryl chiral alcohol (3) in dioxane at 80–90 °C. The reaction mixture was then aged at 100 °C
until reaction completion. The reaction mixture was then quenched with water at 80–90 °C before
further cooling and continuing the workup.
Serendipitously, it was found that by adding an appropriate amount of water, a clean phase split was
achieved where the cesium salt of 17 was forced into the organic layer while most of the inorganic
salts remained in the aqueous layer. This phase split obviated the need for neutralizing the large
amounts of Cs2CO3 in the reaction mixture. Diacid impurity 14 was also mostly removed in the
aqueous layer.
Since no crystalline form of 17 was identified at the time, the organic layer was directly treated with
aqueous HCl to remove the Boc-protecting group to give LX1031. Excess biaryl chiral alcohol 3 and
other neutral impurities were removed by extraction with IPAc.
Isolation and purification of LX1031 via dihydrochloride salt. The isolation of LX1031 was
particularly challenging due to its unusual physicochemical properties. Being a zwitterion with a large
hydrophobic aromatic moiety, it behaved like a surfactant. Attempted isolation by neutralization to its
ACS Paragon Plus Environment

Page 17 of 44
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
isoelectric point (pH 6.5) gave a gel that was difficult to filter and dry. Subsequently, a manageable
suspension of LX1031 zwitterion was obtained by slowly adding a solution of LX1031 in aqueous
HCl into a sodium phosphate buffer maintained at pH 6.5 by concomitant addition of aqueous NaOH.
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 the first scale-up batch, starting from 4.57 kg of pyrimidine chloride 10 (78.7% purity), 3.23 kg
of LX1031 zwitterion was obtained as a dark yellow solid with 87% purity. Initial salt screening
suggested that crystallization of LX1031 dihydrochloride could provide significant purity upgrade.
Thus, this batch of LX1031 zwitterion was converted into the dihydrochloride salt in a mixture of
THF/IPA/IPAc, improving its purity to 95.4% (2.99 kg, 82% yield as-is). Recrystallization (86%
recovery) in a similar solvent system (THF/IPA/IPAc) only slightly improved its purity to 96.3%. To
further reduce the color and residual palladium level, it was dissolved in 1 N NaOH and treated with
activated carbon again (0.3X). Unfortunately, the filtration through a pad of cellulose powder was
excruciatingly slow. Switching to a pad of diatomaceous earth led to significant breakthrough of
activated carbon. Meanwhile, nickel contamination was detected for this batch (∼600 ppm), which
was traced to the use of HCl in a new Hastelloy reactor. Fortunately, adding chelating agent EDTA to
the filtrate was effective in keeping the nickel in the mother liquor during free basing in phosphate
buffer (3 ppm in the isolated LX1031, 1.57 kg). To remove the trace amounts of activated carbon, it
was dissolved in 1 N HCl, filtered through a pad of diatomaceous earth followed by a cartridge-filter
and then precipitated from pH 6.5 phosphate buffer to give 1.45 kg of LX1031 zwitterion (97.7%
purity) as a pale-yellow solid (45% overall yield from 10).
For the 2nd scale-up batch (28.3 kg of 10), the SNAr reaction and deprotection went smoothly.
However, the isolation of LX1031 API was still problematic. The filtration/wash of the precipitated
LX1031 zwitterion from pH 6.5 sodium phosphate buffer was slow (23 h). The drying was even more
time consuming (4.5 days at 50–60 °C) affording 23.8 kg (61% yield) crude LX1031 (94.0% purity).
ACS Paragon Plus Environment

Organic Process Research & Development
Page 18 of 44
1
2
3
4
5
6
7
8
9
The purity was upgraded to 98.9% via crystallization of its dihydrochloride salt (18.1 kg, 67% yield).
Again, the filtration (6 days) and drying (3 days) were extremely slow. The dihydrochloride salt was
converted to LX1031 zwitterion (10.2 kg, 68% yield) as described previously. The low overall yield
(27.3%) from 10 was partially attributed to physical loss due to a missing dust collector in the
filter-dryer.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
To alleviate the challenges of purity upgrade at the API stage, identifying a crystalline form of
penultimate intermediate 17 was focused on. An extensive solvent screen led to the discovery a
crystalline form of 17 in acetonitrile. The crystallization was incorporated into the process stream,
allowing the isolation of 17 with 95.3% purity in a scale-up batch (85.9 kg, 81% yield as-is). The urea
impurity 18 was reduced from ∼5% to ~2%.
Meanwhile, a semi-crystalline THF solvate of LX1031 zwitterion was discovered, which could
provide significantly improved filtration rate. The crystallization was sensitive to temperature
(< 40 °C), THF/water ratio (> 1/3) and levels of residual inorganic salt (the lower the better). It also
required heavy seeding for a more consistent performance. The solvated THF was readily removed
during drying to give amorphous LX1031. In a scale-up batch (85 kg of 17), it was found that the
filtration and wash were still slow (19 h/32 h), suggesting that the purportedly THF solvate might not
have formed. The wet-cake also contained large amounts of water requiring prolonged drying (39 h)
at 45 °C under reduced pressure to give 75.3 kg of LX1031 zwitterion (91.4 % yield corrected) with
95.3% purity.
ACS Paragon Plus Environment

Page 19 of 44
Organic Process Research & Development
1
2
3
4
Scheme 6 Initial processes for deprotection and isolation of LX1031
5
6
7
8
1. HCl, dioxane, H2O
17
LX1031 zwitterion
2. phophate buffer
(pH 6.5)
1. HCl (aq)
2. pH 6.5
phosphate buffer
(pH 6.5)
HCl, IPA
THF, IPAc
9
3. THF/H2O
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
LX1031 2HCl

LX1031 THF
(partially crystaline

)
Purification of LX1031 via tosylate salt. To achieve a more robust purity upgrade, impurity purging
via isolation of a crystalline salt of LX1031 was investigated. Salt screening showed that the tosylate
salt was promising. Extensive experimentation revealed that LX1031 tosylate exhibited rather
complex polymorphism. It could be isolated as an anhydrate, monohydrate or dihydrate, depending on
the water contents of crystallization medium (THF/ACN/water) (Scheme 7). Below 5 vol% water, the
anhydrate was the preferred form. However, it readily converted to the monohydrate upon exposure
to humid air even in solid state. The monohydrate was stable in solid-state regardless of humidity
levels. But suspension equilibration experiments showed that it was stable only within a narrow range
of water contents (5 ± 1%). On the other hand, the dihydrate was proved stable in a much wider range
of water contents (6–13%), and selected for development. The optimal composition for the
crystallization medium was 5–8% water, 10–25% THF, with the remainder being acetonitrile. This
crystallization provided robust purity upgrade as demonstrated on a scale-up batch. Starting from 95.3%
purity LX1031 zwitterion (64 kg), 99.5% purity LX1031•TsOH•2H2O was obtained in 82.3% yield.
A portion of the batch (27.2 kg) was converted to LX1031 zwitterion in 87.4% yield (17.2 kg) in
THF/water. Unfortunately, the filtration/wash and drying operations were still slow.
ACS Paragon Plus Environment

Organic Process Research & Development
Page 20 of 44
1
2
3
4
Scheme 7 Polymorphism of LX1031 tosylate
5
6
7
humid air
solid state
8
9
> 6-13% H2O
suspension
0-5% H2O
Form 3
anhydrate
Form 1
monohydrate
Form 2
dihydrate
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
suspension
Crystallization of LX1031 acetonitrile solvate. As amorphous LX1031 zwitterion was used in early
clinical studies, it was desirable to continue using this form in mid-stage trials. Therefore, a better
process for the isolation of the zwitterion, particularly with improved filtration and drying performance
was sought. Attempted crystallization from 2-MeTHF gave a gel even after repeated heating/cooling
cycles with periodic sonication. Crystallization from IPA afforded a slow-filtering suspension.
Crystallization in ethanol gave a LX1031 EtOH solvate that filtered reasonably well. But the residual
EtOH was difficult to remove. After drying at 60 °C for 16 h in vacuo, 26 mol% EtOH remained.
Fortunately, it was discovered that LX1031 acetonitrile solvate monohydrate was highly crystalline
and filtered easily. Residual acetonitrile and water were also readily removed during drying at 50 °C
under reduced pressure. This discovery was utilized for the free-basing of the next portion of LX1031
tosylate dihydrate (25.5 kg). Thus, after neutralization (pH 6.5) with 8% aq. NaOH, the THF layer was
washed with 15% aqueous NaCl, concentrated and solvent swapped to acetonitrile to crystallized
LX1031 acetonitrile solvate monohydrate. The filtration, washes (water, acetonitrile) and drying were
expedient delivering amorphous LX1031 zwitterion in 91% yield (16.7 kg). Residual acetonitrile was
at undetectable level (< 100 ppm).
SECOND-GENERATION PROCESS ROUTE TO LX1031
While the first-generation process route was amenable for production after implementing the many
process improvements, the projected high dose of the drug (>2 g/day) prompted us to further reduce
the cost of the drug substance. The main cost driver of the process was the production of pyrimidine
ACS Paragon Plus Environment

Page 21 of 44
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
chloride 10. This was attributed to the tedious and volume-intensive process as well as a modest yield
(60–65%). Therefore, an alternative “left-to-right” approach like the medicinal chemistry route was
reevaluated (Scheme 8). In this approach, the main challenge was to overcome the sluggishness of the
Suzuki-coupling of pyrimidine chloride 5 with L-BPA.
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 Left-to-right Approach
MeO
DCAP
MeO
O
Cl
OH
CF3
CF3
N
N
NH2
3
5
oil
9
1. Suzuki, boronate
2. Deprotection
LX1031
SNAr reaction of chiral alcohol (3) with DCAP. The SNAr substitution reaction of biaryl chiral
alcohol 3 with DCAP was facile in 1,4-dioxane, achieving complete conversion in 5 h at 100 °C with
1.5 equivalents of Cs2CO3. Although less expensive K2CO3 could be used, a highly polar solvent such
as DMF, DMSO or DMAc was required, complicating the workup. Interestingly, in t-amyl alcohol,
the reaction proceeded even with K2CO3 as the base, albeit requiring 20 h at refluxing temperature.
As expected, less hindered alcohol EtOH gave substantial amounts (∼10%) of ethoxy adduct of DCAP
16b. In comparison, IPA and i-BuOH gave < 1% and ∼3% of corresponding adducts. The cost saving
of using K2CO3 vs. Cs2CO3 did not justify the longer cycle-time, however.
DCAP charge was reduced from 2.0 to 1.5 equivalents while still controlling bis-substitution
impurity 19 to ∼3% (Scheme 9). Further cut (1.3 equivalents) resulted in a much slower reaction and
noticeably increased 19 (4%). Reducing solvent from 10 to 3 volumes not only improved volume
productivity but also the reaction rate. Under the optimized reaction conditions (1.5 equiv Cs2CO3, 1.5
ACS Paragon Plus Environment

Organic Process Research & Development
Page 22 of 44
1
2
3
4
5
6
equiv DCAP, 3 vol dioxane, 100 °C), the reaction mixture was typically composed of 95% of 5, 3%
of 19, and two unidentified impurities at 1% and 0.6% (normalized area% excluding DCAP peak).
7
8
9
To minimize consumption of valuable boronate 9 in the subsequent Suzuki-coupling step, removal
of residual DCAP was crucial. Since acid or base extractions was deemed unfeasible, the reaction
mixture was solvent swapped to toluene and dilution with n-heptane to precipitate most of the DCAP
for removal by filtration. Unfortunately, ∼6 mol% DCAP still remained in the filtrate. Attempted
crystallization of 5 in various organic solvents or aqueous systems was unsuccessful. Therefore,
crystallization of 5 as a salt was explored. Only strong acids such as HCl, MsOH and p-TsOH were
evaluated considering the weak basicity of 5 (predicted pKa = 0.98). Neither HCl nor p-TsOH was
effective in forming a crystalline salt with 5. In contrast, adding MsOH to a solution of 5 in
toluene/dioxane produced a white precipitate nearly instantly. XRPD data of the filtered and dried
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
solid confirmed its crystallinity. The HNMR spectrum showed 1 : 1 : 1 molar ratio of 5 : MsOH :
toluene, suggesting that toluene was incorporated into the crystal lattice. Indeed, no appreciable
toluene loss was observed when the solid was dried in vacuo at 45 °C. Under more forceful drying
conditions (85 °C in vacuo), the powdery solid shrunk, melted and turned into a glassy solid upon
cooling. Similarly, suspending the solid in warm heptane (60 °C) to remove toluene also gave a
gummy solid. Thermal Gravimetric Analysis (TGA) showed 16% weight-loss between 70–180 °C,
consistent with the calculated value of 15.4% for 1.0 equivalent toluene.
Scheme 9 SNAr substitution of 3 and DCAP
Ar
O
Cl
Ar
O
O
Ar
CF3
Ar
OH
CF3
DCAP, 1,4-dioxane
Cs2CO3, 100 °C
CF3
N
N
+
CF3
N
N
NH2
oil
NH2
3
5
19
MsOH
toluene
MeO
Ar:
5·MsOH·tol
ACS Paragon Plus Environment

Page 23 of 44
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
The optimal toluene/dioxane ratio for the crystallization was determined to be 5/1 based on solubility
data and impurity purging capacity. Thus, after completion of the reaction, the reaction mixture was
concentrated, flushed with toluene and filtered to remove the precipitated DCAP as well as inorganic
salts. After adjusting the toluene/dioxane ratio to 5/1, a portion of MsOH (0.38 equiv) was charged at
50 °C to give a thin suspension. The remaining MsOH (0.76 equiv) was then slowly added and the
mixture was cooled and aged to complete the crystallization. This isolation efficiently purged both
DCAP and bis-substitution impurity 19, furnishing 5 in 87% yield (35.8 kg) with 99.7% purity.
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
Suzuki-coupling to penultimate intermediate (17). As mentioned earlier, developing a scalable
Suzuki-coupling of pyrimidine chloride 5 with boronate 9 was crucial for the success of the second-
generation process route (Scheme 10). Initial catalyst screening reactions were carried out in refluxing
EtOH/H2O using crude pyrimidine chloride 5 with KHCO3 as the base (Table 6). While the reactions
were still sluggish, they were more promising than when L-BPA (5) was used in medicinal chemistry
route. Perhaps this was due to decreased substrate inhibition when the amino group was protected by
a Boc-group. With 2 mol% POPd6 or PXPd4, after refluxing for 16 h, the conversions were 21% and
16%, respectively (entries 1, 2). Catalyst Pd(dppf)Cl2 was slightly more active, giving 28% conversion
in 16 h (entry 3). However, a dramatically higher level of amination impurity 20 was observed (20%
vs ∼0.2%). Relatively inexpensive Pd(PPh3)2Cl2 performed the best, giving 34% conversion in 16 h
and 62% in 40 h (entry 4). After additional 1.0 equivalent of boronate 9, the reaction reached
completion in 65 h. With 1.5 equivalents of 9, the conversion reached 70% after 40 h (entry 5).
ACS Paragon Plus Environment

Organic Process Research & Development
Page 24 of 44
1
2
3
4
Scheme 10 Suzuki coupling of pyrimidine chloride 5 and boronate 9
5
O
6
7
8
9
OH
NHBoc
Ar
O
Cl
9
Boronate
Ar
O
CF3
N
N
Pd(PPh3)2Cl2, (0.5mol%)
KHCO3, dioxane/H2O,
CF3
N
N
NH2
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
NH2
17
O
OH
NHBoc
Ar
O
CF3
N
N
Ar:
MeO
Ar
O
NH
CF3
N
N
20
NH2
Table 6 Catalyst Screening in EtOH/H2O/KHCO3
Entry
Catalyst
Boronate 9 Time (h) Conv (%)
(equiv)
1
POPd6
1.0
1.0
16
16
21
16
2
3
PXPd4
16
40
28
49
Pd(dppf)Cl2
1.0
1.0
40
62
100
4
5
Pd(PPh3)2Cl2
Pd(PPh3)2Cl2
+1.0
+65
1.5
40
70
*2 mol% catalyst, 3.5 equiv KHCO3, EtOH/H2O, reflux (75 °C)
To achieve acceptable reaction rate for production, higher boiling solvent systems were evaluated
(Table 7). Reaction in refluxing i-BuOH/H2O for 24 h gave 31% conversion (entry 1). In contrast, the
reaction in refluxing 5 : 1 dioxane : water reached completion in 16 h (entry 2). Decreasing charge of
boronate 9 to 1.3 equivalents led to much longer reaction time (48 h) (entry 3).
The reaction condition was further optimized to enhance the reaction rate. Decreasing the dioxane :
water ratio from 5 : 1 to 2 : 1 (total solvent volume = 9 vol) markedly increased the reaction rate,
reducing the reaction time from 12 h to 3–4 h (Figure 4). Interestingly, further decreasing dioxane :
water ratio to 1 : 1 led to a slower reaction. This was likely due to oiling of substrate 5 at the outset.
ACS Paragon Plus Environment

Page 25 of 44
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
Another surprising observation was that the reaction was faster under more dilute conditions. For
example, increasing total solvent volume from 7.5 to 12 volumes (2 : 1 dioxane : water) shorten the
reaction time from 10 h to 6 h (Figure 5). Not surprisingly, with slow agitation, the reaction required
∼15 h to reach completion compared to 8 h with vigorous stirring (Figure 6).
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 7 Screening of higher boiling solvent systems
Entry
Solvent
(15/3 vol)
Boronate 9 Time Conv
(equiv) (h) (%)
1
i-BuOH/H2O
dioxane/H2O
dioxane/H2O
1.5
1.5
1.3
24 31
2
3
16
48
100
98
Conditions: 2 mol% Pd(PPh3)2Cl2, 3.5 equiv KHCO3, reflux.
100
80
60
40
20
0
100
90
80
70
60
5/1 dioxane/water
3/1 dioxane/water
1/1 dioxane/water
2/1 dioxane/water
50
40
30
20
10
0
8X dioxane
5X dioxane
0
2
4
6
8
10
12
0
2
4
6
Time (h, 90 °C)
Time at 90 °C (h)
Figure 5. Effect of solvent volume
Figure 4. Effects of dioxane : water ratio
ACS Paragon Plus Environment

Organic Process Research & Development
Page 26 of 44
1
2
3
4
5
6
7
8
9
100
80
60
40
20
0
100
90
80
70
60
50
1.5eq ArB-pin
1.2 eq ArB-pin
1.15 eq ArB-pin
1.10 eq ArB-pin
1.05 eq ArB-pin
Slow stirring
40
30
20
10
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
Rapid stirring, Morton flask
0
5
10
15
0
1
2
3
4
5
6
7
Time (h)
Time (h, 90 °C)
Figure 6. Effect of agitation
Figure 7. Effect of boronate 9 equivalents
These observations could be rationalized by the efficiency of bring the two coupling partners into
contact in the biphasic reaction mixture. It was observed that boronate 9 partitioned mostly in the
aqueous phase while pyrimidine chloride 5 in the organic phase. Lower concentration or lower dioxane
: water ratio decreased the ionic strength of the aqueous layer which should lead to better miscibility
of the two phases. On the other hand, at a higher concentration, pyrimidine chloride 5 became a
significant portion of the organic phase and acted like a hydrophobic solvent, thereby decreasing
miscibility of the two phases.
Another operational issue was the aerosolization of boronate 9 while being charged into KHCO3 by
the CO2 generated. This risk was mitigated by replacing KHCO3 (3.5 equiv) with K2CO3 (1.75
equiv).16 Under the optimized reaction conditions (0.5 mol% Pd(PPh3)2Cl2, 1.75 equiv. K2CO3, 9 vol
2 : 1 dioxane : water, reflux), boronate charge could be reduced from 1.5 to 1.1 equivalents while still
achieving complete conversion in 6–8 h (Figure 7). This constituted a substantial cost saving for the
second-generation process.
With higher purity of the crystalline 5·MsOH, the Suzuki reaction was expected to proceed more
rapidly. To the contrary, the reaction was noticeably slower, requiring ∼10 h (vs ∼4 h) even after
ACS Paragon Plus Environment

Page 27 of 44
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
adjusting the K2CO3 charge to compensate for the neutralization of MsOH. This was attributed to less
effective mixing of the biphasic mixture due to higher ionic strength of the aqueous layer (potassium
mesylate salt). Indeed, by performing an in-situ free-basing of the mesylate salt with 0.6 equivalents
of K2CO3 in dioxane : water and removing the aqueous layer, the typical reaction rate was restored.
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
As for the workup and isolation, the reaction mixture was cooled, acidified with 5 N HCl to pH 2–
3, and saturated with NaCl to aid the phase separation. The organic layer was treated with activated
carbon and solvent swapped to acetonitrile to crystallize product 17. The main drawbacks of this
process included a slow phase separation and tedious solvent swap from higher boiling 1,4-dioxane to
lower boiling acetonitrile. Therefore, a more streamlined process was developed where the reaction
mixture was treated with activated carbon, and then acidified with aq. HCl to pH 2–3 to directly
crystallize 17 in 83% yield with ~98% purity. Residual Pd levels were somewhat higher and more
variable (100–1000 ppm) than the previous workup procedure (< 100 ppm). Retrospective data
analysis revealed an inverse correlation between residual palladium and reaction rates, i.e. lower
residual palladium levels were correlated to faster reactions. In these cases, the reaction mixtures also
turned darker more rapidly after reaction completion. This phenomenon suggested that once
pyrimidine chloride 5 was completely consumed, the catalyst became unstable and began to precipitate
as palladium black. Indeed, by extending the aging time by additional 3–5 h after reaction completion,
residual palladium levels were consistently below 100 ppm in the isolated product 17.
Deprotection and isolation of LX1031. Previously, the deprotection was carried out in THF and
aqueous HCl. As THF might not be completely stable to the strongly acidic conditions, it was
advantageous to conduct the deprotection without it. The reaction was sluggish in 2 N aqueous HCl at
50 °C but much more expeditious at 60 °C (2–3 h). For safety and practicality in the plant, penultimate
17 was charged in portions over a period of 2 h to avoid excessive foaming of the reaction mixture
ACS Paragon Plus Environment

Organic Process Research & Development
Page 28 of 44
1
2
3
4
5
6
7
8
9
and aerosolization of powdery 17 due to vigorous CO2 evolution. With higher purity penultimate 17
(∼98%) from the second-generation process route, high quality LX1031 zwitterion could be isolated
directly, obviating the need for purity upgrade via tosylate salt formation. Thus, upon reaction
completion, the reaction mixture was cooled, partially (∼60%) neutralized with 50% aqueous NaOH,
and then further neutralized to pH 6.5 after adding THF (5 vol). The organic layer was separated and
flushed with acetonitrile to crystallize LX1031 acetonitrile solvate monohydrate which upon drying
furnished amorphous LX1031 zwitterion in 86% yield (20 kg scale) and 98.7% purity.
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

Page 29 of 44
Organic Process Research & Development
1
2
3
4
CONCLUSIONS
5
6
7
8
In summary, two process routes were developed to produce LX1031, a drug candidate for the treatment
of irritable bowel syndrome. They shared the key intermediates biaryl chiral alcohol (3) and L-tyrosine
derived boronate (9), as well as the penultimate intermediate (17). Compound 3 was synthesized via
asymmetric hydrogenation of trifluoromethyl ketone 1 followed by Suzuki coupling with
3-methoxyphenylboronic acid (11). The boronate intermediate (9) was synthesized via borylation of
triflate of Boc-L-tyrosine methyl ester with pin2B2. The first-generation process route assembled the
target molecule from right-hand to left-hand side while the second-generation process from left-hand
to right-hand side to give the same penultimate intermediate (17). For the first-generation route, the
challenges in the isolation and purification of LX1031 zwitterion were overcome by isolating LX1031
tosylate dihydrate followed by free basing and crystallization of LX1031 acetonitrile solvate. With
higher quality penultimate intermediate 17 in the second-generation synthesis, high purity LX1031
zwitterion was isolated directly. The second-generation route not only employed cleaner reactions but
was also more productive (62% vs 36% overall yield). Both process routes had been carried out on
multi-kg scales to provide LX1031 drug substance for pre-clinical and early clinical programs.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
EXPERIMENTAL SECTION
General. All reagents and materials were used as received, unless otherwise noted. Reactions were
monitored by reverse-phase HPLC using a C18 or phenyl-hexyl column with water/MeCN or
water/MeOH mobile phase and TFA as a modifier. Melting points were determined by using a melting
point apparatus or based on the peak temperatures of the endotherm on a differential scanning
calorimeter (DSC). NMR spectra were acquired in deuterated solvents with a 300 or 400 MHz (1H)
NMR spectrometer. Mass spectrometry data were obtained during LC-MS analysis. Compound purity
data were determined by reverse phase HPLC analysis.
ACS Paragon Plus Environment