First, second, and third generation scalable syntheses of two potent H 3 antagonists

Article abstract of DOI:10.1021/op200005e

Our teams have recently completed scale-up campaigns for two structurally similar H₃ receptor antagonists. The first and second generation processes were developed for the synthesis of 107 and 125 g batches of (4-cyclobutyl-1,4-diazepan-1-yl)(6

Full text of DOI:10.1021/op200005e

ARTICLE
pubs.acs.org/OPRD
First, Second, and Third Generation Scalable Syntheses of Two Potent
H3 Antagonists
Daniel J. Pippel,* John E. Mills, Chennagiri R. Pandit, Lana K. Young, Hua M. Zhong, Frank J. Villani, and
Neelakandha S. Mani
Johnson & Johnson Pharmaceutical Research and Development, L.L.C., 3210 Merryfield Row, San Diego, California 92121,
United States
S Supporting Information
b
ABSTRACT: Our teams have recently completed scale-up campaigns for two structurally similar H3 receptor antagonists. The first
and second generation processes were developed for the synthesis of 107 and 125 g batches of (4-cyclobutyl-1,4-diazepan-
1-yl)(6-(4-fluorophenoxy)pyridin-3-yl)methanone HCl (1 HCl). A third generation process was utilized for production of 104 g
3
3
of 3-((5-(4-cyclobutyl-1,4-diazepane-1-carbonyl)pyridin-2-yl)oxy)benzonitrile HCl (2 HCl). The evolution from first to second
3
3
generation process was driven by a desire to minimize cost of goods through employment of symmetrical homopiperazine rather
than a more expensive monoprotected variant. Project demands for a late stage intermediate that could provide 1 or 2 led to
additional route scouting and the ultimate determination of a third scalable synthesis for these types of molecules. The use of a
lithium alkoxide for Lewis base catalysis of an ester to amide transformation represents a key improvement for the third generation
synthesis.
’ INTRODUCTION
The H3 receptor has emerged as a promising target within the
neuroscience community; antagonism of this G-protein-coupled
receptor could have therapeutic applications for various disease-
states including narcolepsy, ADHD, Alzheimer’s disease, and
obesity.1ꢀ3 With the phenomenal success of other histamine
receptor antagonists including Claritin and Benadryl (H1 recep-
tor antagonists), as well as Tagamet and Zantac (H2 receptor
antagonists), excitement over a potential third class of antihista-
mine drugs has resulted in a large body of work spanning the
laboratories of academic groups, biotechnology companies, and
several pharmaceutical companies.4 Internally, our medicinal
Figure 1. Structures of H3 receptor antagonists 1 and 2.
chemistry team has developed several series of H3 receptor
antagonists, including recently disclosed hydroxyproline-based
several factors drove us to pursue an alternative. First, even
though 5-bromo-2-fluoropyridine (3) is readily available and
inexpensive, we did not believe it could ultimately compete with
6-chloronicotinic acid (or its simple derivatives) from a cost
and aryloxypyridine-based ligands.5ꢀ7 Of the aryloxypyridine-
based H3 receptor antagonists, two compounds, 1 and 2, were
chosen for in-depth profiling based on binding affinity, brain
perspective.11 Several key intermediates utilized in the synthesis
penetration, and microsomal stability (Figure 1). Scale-up devel-
opment of these two compounds forms the focus of this paper.
of common herbicides and the neonicotinoic class of insecticides
are closely related to 6-chloronicotinic acid,12 and it is available in
Compound 1 was initially synthesized in just 2 steps (not
including the preparation of N-cyclobutylhomopiperazine 5) by
our medicinal chemistry counterparts. Nucleophilic aromatic
substitution of 5-bromo-2-fluoropyridine (3) with 4-fluorophe-
nol was followed by Larhed aminocarbonylation of 4 under
two simple steps from 3-picoline (chlorination to the corre-
sponding chloro(trichloromethyl)pyridine and hydrolysis).13
Second, the opportunity to replace a palladium-mediated ami-
nocarbonylation in the final step with a simple amide bond
forming reaction seemed desirable in light of the potential for
metal residues. Third, the chromatography employed after the
microwave conditions with 5 and carbon monoxide generated
in situ from molybdenum hexacarbonyl (Figure 2).7ꢀ10 This
route provided the starting point for our investigations.
medicinal chemistry aminocarbonylation would not be suitable
for large-scale synthesis.
’ RESULTS AND DISCUSSION
Although the sequence shown in Figure 2 was quite elegant
and concise, in considering a potential manufacturing route,
Received: January 6, 2011
Published: March 12, 2011
r
2011 American Chemical Society
638
dx.doi.org/10.1021/op200005e Org. Process Res. Dev. 2011, 15, 638–648
|

Organic Process Research & Development
ARTICLE
Figure 2. Synthesis of 1 through SNAr and aminocarbonylation
reactions.
With this alternative route in mind, a first generation synthesis
to 1 HCl H2O (Figure 3) was developed. Preparation of the
3
3
N-cyclobutylhomopiperazine dihydrochloride salt (5 2HCl) was
3
straightforward, proceeding in 86% yield over two steps. Re-
ductive amination of cyclobutanone with commercially available
N-Boc-homopiperazine (6) and sodium triacetoxyborohydride
afforded 7 as a semisolid. Treatment of 7 with 4 N HCl in dioxane
afforded 5 2HCl as a hygroscopic off-white solid. This diamine
3
and the commercially available 6-chloronicotinyl chloride (8)
were then subjected to SchottenꢀBaumann conditions to form
the amide 9 as an oil in 91% yield. Isopropyl acetate was chosen as
the organic solvent for this reaction based on comparison with
toluene and THF. In the case of the former, the nicotinyl chloride
8 was not completely soluble; in the case of the latter, hydrolysis
of 8 was observed. Nucleophilic aromatic substitution of 9 with
4-fluorophenol at 100 °C in DMA with 2 equiv of cesium
carbonate provided the desired H3 receptor antagonist 1 in
76% yield as an oil. Reactions with potassium carbonate or less
than 2 equiv of cesium carbonate did not reach completion. The
need for a crystalline salt form of the final product was addressed
with a screening campaign that ultimately led to a mono HCl,
monohydrate of 1. Salt formation was conducted with 1 in
ethanol and 2 M HCl in ether as the acid source to provide
1 HCl H2O in 82% yield. The largest campaign with this route
Figure 3. First generation scale-up synthesis of 1 HCl H2O.
3
3
heteroarylꢀaryl ether linkage in ethyl 6-(4-fluorophenoxy)
nicotinate (11). Subsequently, in the key step of the sequence,
homopiperazine was allowed to react directly with the ethyl ester
11 to provide the desired diazapanylamide 12. In this reaction, a
solution of the ethyl ester 11 and 2.4 equiv of homopiperazine
were cooled to 0 °C and hexyllithium (1.5 equiv, 2.3 M in
hexane) was slowly added. The order of operations is critical in
this reaction as addition of the hexyllithium to the homopiper-
azine followed by addition of the ester to the lithium amide
results in formation of a small amount of a des-fluoro amide
byproduct. Not surprisingly, even under optimized conditions,
the reaction does produce some of the expected diamide
byproduct.16 However, this diamide can be removed by repeated
extractions of a pH 2 aqueous layer with MTBE. Thus, with the
appropriate extraction and crystallization protocol, 12 is pro-
vided as a free-flowing solid in 56% yield. Final-stage reductive
amination of cyclobutanone with 12 and salt formation gave rise
to 1 HCl H2O in 69% yield, on 125 g scale.14
3
3
provided 107 g of the final product.14
This first generation scale-up synthesis was more than ade-
quate for progression of the project through initial stage pre-
clinical tolerance studies. However, when planning for the
synthesis of the first GMP kilogram-scale batches, our team
reassessed the process. From a cost-of-goods perspective, it was
clear that Boc-homopiperazine (6) was not an optimal starting
material as it is up to 25 times more expensive than the parent
homopiperazine.15 Additionally, employment of homopipera-
zine directly would obviate a Boc removal step. Finally, a GMP
version of the first generation synthesis would most likely have
led to outsourcing cyclobutylhomopiperazine 2HCl (5 2HCl),
3
3
Because the desymmetrization of homopiperazine is not a
trivial problem, we were especially pleased with this second
generation synthesis.6b,17 Also, direct ester to amide transforma-
tion is arguably underutilized.18ꢀ20 Another advantage of this
route is the introduction of the most expensive piece on a molar
basis, cyclobutanone, in the final chemical transformation.
As the 4-fluorophenoxy analog 1 continued to progress
through biological profiling, it became apparent that our H3
receptor antagonist team was also interested in the closely related
3-cyanophenoxy analog 2.7 In fact, the team became motivated to
design a synthesis that could proceed to either 1, 2, or a similar
3
3
which is hygroscopic and difficult to handle, making it a poor
choice as a regulatory starting material.
With the desire to utilize the least expensive possible diamine
starting material as a primary driver, our team developed the
synthesis shown in Figure 4. An initial SNAr reaction between
ethyl-6-chloronicotinate (10) and 4-fluorophenol provided the
639
dx.doi.org/10.1021/op200005e |Org. Process Res. Dev. 2011, 15, 638–648

Organic Process Research & Development
ARTICLE
analog via a common late-stage intermediate. Such a route would
provide maximum flexibility as the project continued to move
forward. Unfortunately, the second generation synthesis of 1 did
not meet this requirement as the differentiation of the aryl ether
group occurred in the first step.21 Clearly, an alternate approach,
which featured a final step heteroarylꢀaryl ether bond formation,
was required. The retrosynthetic analysis shown in Scheme 1 was
proposed as an alternative to the first generation route.
(Scheme 2). These unwanted side products were anticipated
based on two modes of possible reactivity. First, homopiperazine has
the capacity to react at both termini to form pseudodimer 14. Second,
since the 6-chloro group on the nicotinic acid derivative is potentially
labile, byproduct 15 is easily conceived. In the ethyl 6-chloronicotinate
case, side products 16, 17, and 18 are also possible. Further,
combinations of the two modes of byproduct generation, as well as
potential incorporation of the base were also feared.
Here, the key step is base-mediated reaction between 6-chlor-
onicotinyl chloride (8) or ethyl 6-chloronicotinate (10) and
homopiperazine to generate 13. However, in addition to the
desired 13, one can imagine many potential side products
The reaction between 8 and homopiperazine was explored
under a variety of conditions as shown in Table 1. With
triethylamine as a base, the product distribution ranged from
exclusively pseudodimer 14 in THF to a ∼1:1 ratio of product 13
to pseudodimer 14 in tert-amyl alcohol (Table 1, entries 1ꢀ5).
Several inorganic bases were also screened with tert-amyl alcohol
(Table 1, entries 6ꢀ8). The case with 2 equiv of potassium
carbonate (Table 1, entry 8) provided the optimal product 13 to
pseudodimer 14 ratio of 1.7:1. Interestingly a reaction with no
base at all (Table 1, entry 9) afforded identical results. All
reactions displayed clean HPLC profiles and proceeded to
completion.
With these results in hand, the tert-amyl alcohol/K2CO3
conditions shown in Table 1, entry 8, were scaled to 12 g, and
the desired product 13 was obtained in 37% yield (eq 1).
Removal of pseudodimer 14 was possible without chromatogra-
phy as the product could be selectively extracted into a pH 3.74
buffer (NaOAc/HOAc).
Despite this success, we still hoped to improve the yield
through minimization of pseudodimer 14 formation. Our initial
Figure 4. Second generation scale-up synthesis of 1 HCl H2O.
3
3
Scheme 1. Retrosynthetic Analysis for Third Generation Synthesis of 1 and 2
640
dx.doi.org/10.1021/op200005e |Org. Process Res. Dev. 2011, 15, 638–648

Organic Process Research & Development
ARTICLE
Scheme 2. Potential Products in Reaction between 6-Chloronicotinic Acid Derivatives and Homopiperazine
Table 1. Formation of Amide 13 from 8 and Homopiperazine
pseudodimer formation, we chose to explore reactions between
10 and homopiperazine.
Thus, 10 and homopiperazine were premixed and then treated
with hexyllithium, exactly according to the previous instance with
11 and homopiperazine (Table 2, entry 1). In fact, 13 was
generated as the primary product in a 8:1 ratio with pseudodimer
14; however, this reaction was not nearly as clean as the case with
tert-amyl alcohol/potassium carbonate. Preformation of the
lithium amide followed by addition of the ester 10 minimized
pseudodimer 14 formation further (14:1, 13:14), and the crude
product was purified by flash chromatography to provide a 60%
yield of material that was 88 wt % 13 and 12 wt % 16 (Table 2,
entry 2).22,23
With i-PrMgCl as an alternative base, the reaction was quite
clean providing exclusively product 13 and pseudodimer 14;
however, the 13:14 ratio was just 2.5:1 (Table 2, entry 3).
Recalling the reasonably successful reaction between 6-chloro-
nicotinyl chloride (8) and homopiperazine in the absence of base
(Table 1, entry 9), a base-free reaction with ethyl 6-chloronico-
tinate (10) was also attempted. In this case, 16 is the exclusive
product, isolated in 83% yield, with no formation of 13 (Table 2,
entry 4).
A scale-up of the conditions shown in Table 2, entry 2
(preformation of the lithium amide and then addition of the
ester 10) to a 20 g scale resulted in an unexpectedly low yield
(46%) and significant formation of side product 17, which
previously had not been seen to any significant extent
(Table 2, entry 5). In an attempt to understand the unexpected
behavior, varying equivalents of n-hexyllithium were screened on
small-scale.24 HPLC yields for the reactions indicated that 0.5
equiv of base is sufficient to push the reaction to completion, and
excess base contributes to complex reaction profiles (Table 2,
entries 6 and 7).25
entrya
base (equiv)
solvent
THF
product ratio (13:14)b
1
2
3
4
5
6
7
8
9
triethylamine (1)
triethylamine (3)
triethylamine (1)
triethylamine (1)
triethylamine (1)
2 M NaOH(aq) (2)
14 only
1:11.2
1:2.3
THF
DCM
ACN
1:1.8
tert-amyl alcohol
tert-amyl alcohol
1:1.1
1:1.37
1:1.35
1.7:1
2 M Na2CO3(aq) (2) tert-amyl alcohol
K2CO3 (2)
tert-amyl alcohol
none
tert-amyl alcohol
1.7:1
a In entries 1ꢀ6 and 9 homopiperazine, (base), and solvent were stirred
under N2 (g) at 0 °C and then a solution of 6-chloronicotinyl chloride
was slowly added. In entries 7 and 8, solid 6-chloronicotinyl chloride was
slowly added at room temperature. b Ratios are based on calibrated
HPLC area response.
experiments had been conducted with nicotinyl chloride 8
because we reasoned that the acid chloride functional group
would be more reactive than the ethyl ester group in 10, which
could help to minimize reactions at the 6-chloro position.
However, because that enhanced reactivity seemed to promote
Although we were delighted to have improved the yield of the
small-scale reaction in an unexpected way (substoichiometric
base), we were still stymied in our attempts to scale up the
reaction. In fact, a 20 g reaction with 0.5 equiv of base gave results
only moderately superior to those shown in entry 5, again with
641
dx.doi.org/10.1021/op200005e |Org. Process Res. Dev. 2011, 15, 638–648

Organic Process Research & Development
ARTICLE
Table 2. Formation of Amide 13 from 10 and Homopiperazine
entry
base (equiv)
conditionsa
yield of 13, %
commentsb,c
13:14 = 8:1
1
2
3
4
5
6
7
8
9
n-hexyllithium (1.05)
n-hexyllithium (1.05)
i-PrMgCl (1.05)
A, 1 g scale
B, 1 g scale
C, 1 g scale
D, 1 g scale
B, 20 g scale
B, 1 g scale
B, 1 g scale
B, 20 g scale
A, 20 g scale
58d
60e
59e
0
46e
75d
45d
57e
73e
13:14 = 14:1
13:14 = 2.5:1
83%e of 16
none
n-hexyllithium (1.05)
n-hexyllithium (0.5)
n-hexyllithium (2.0)
n-hexyllithium (0.5)
n-hexyllithium (0.5)
17 is major side pdt
clean reaction
messy reaction
17 is major side pdt
no 17 observed
a In condition A, homopiperazine and 10 were premixed and then base was added at 0 °C. The reaction was then warmed to room temperature and
held for 2ꢀ16 h. In condition B, homopiperazine and base were premixed and then the electrophile 10 was added at 0 °C. The reaction was then
warmed to room temperature and held for 2ꢀ16 h. In condition C, the conditions were identical to B, but with no warming after addition of the
ester and quenching after 1 h. In condition D, the reaction was warmed to 40 °C overnight. b In every case (except entries 3 and 4) impurity 16 is
formed at about 5ꢀ10%. c Ratios of 13 to 14 are based on calibrated HPLC area response. d Yield by quantitative HPLC with internal standard.
e Isolated yield.
significant formation of 17 (Table 2, entry 8).26 In an attempt to
duplicate the large-scale reaction failure on small scale, the
reaction was stressed in the following ways: addition of 2 mol %
of Pd(OAc)2 (to test the effect of possible trace Pd contamina-
tion of our glassware), addition of 10 mol % of water, poor
stirring, and aging of the lithium amide for several hours prior to
the addition of the ester 10. The only case that duplicated the
large-scale failure with formation of 17 was the reaction that
incorporated aging of the reaction mixture prior to ethyl ester 10
addition. It seems likely that the cause of the failures upon scale-
up or on small scale with extended aging relates to aggregation
states and possible precipitation of the homopiperazine lithium
amide. These reactions are not homogeneous.27
This result prompted us to return to our original conditions
wherein homopiperazine and the ester 10 were premixed prior to
addition of the n-hexyllithium (Table 2, entry 1). This addition
order does result in a slight increase in the level of pseudodimer
14; however, this problem is more than offset by the fact that
impurity 17 is formed only at trace levels upon scale-up to 20 g.
An isolated yield of 73% was obtained (Table 2, entry 9).
With this scalable approach to 13, a third generation synthesis
for 1 or 2 was realized (Figure 5). The homopiperazine reaction
with ethyl 6-chloronicotinate (10) was executed on 250 g scale,
and the desired product 13 was isolated in a yield of 69% (after
subtracting the mass of a small amount of 16 that remained after
the aqueous workup).28 The reductive amination of cyclobuta-
none with 13 was conducted in DCE with sodium triacetox-
yborohydride and proceeded uneventfully to 9.29 Because 9 was
not crystalline as its free base, the minor impurities that existed in
the 13 product were still present, either unchanged in the case of
14, or as a reductive amination product in the case of 16. Thus, a
crystalline salt form of 9 was sought for the control strategy. After
a brief screen of typically employed acids, purification was, in fact,
possible through formation of a solid L-tartrate salt in ethanol.
The salt formation proceeded in 87% yield. Recovery of the free
base through extraction (1 N NaOH/i-PrOAc) proceeded in
89% yield. The overall yield, then, for reductive amination/
purification was 77%.
With the key common intermediate in hand, it was possible to
generate both 1 (vide supra) and 2 in high yield. In the case of the
latter, treatment of the chloroamido pyridine 9 with cesium
carbonate and m-cyanophenol in dimethylacetamide gave rise to
the desired product 2 in 90% yield after ∼24 h at 125 °C. Salt
formation was accomplished with anhydrous HCl in IPA to
provide 2 HCl as a crystalline solid in 83% yield.30 It is important
3
to drive the diaryl ether forming reaction to completion, as it is
difficult to completely remove the residual chloroamido pyridine
9 starting material through final salt formation.
After completion of the campaign to provide 100 g of 2 HCl,
3
we briefly returned to the homopiperazine reaction with ethyl
6-chloronicotinate (10). On the basis of the success with 0.5
equiv of n-hexyllithium, the reaction was also explored with use of
1.0 equiv of lithium ethoxide as the base. Gratifyingly, the
reaction provided product in the highest quantitative HPLC
yield yet obtained, 84%. On 10 g scale, the product was obtained
in a calculated yield of 90% after subtracting contributions of
642
dx.doi.org/10.1021/op200005e |Org. Process Res. Dev. 2011, 15, 638–648

Organic Process Research & Development
ARTICLE
Figure 5. Third generation scale-up synthesis of 2 HCl.
3
impurities (mostly 16). An isolated yield of 67% was obtained
The obvious utility of these reactions is in sequences that
preferentially proceed through an ester and would otherwise
require hydrolysis to the corresponding acid prior to amide bond
formation under the control of a coupling system.
after reverse phase HPLC (eq 2).31
’ SUMMARY
We have demonstrated three independent syntheses for the
H3 antagonist 1 and two syntheses for its analog 2 (Scheme 3).
All of the approaches have been demonstrated to at least 100 g
scale. We have also identified HCl salt forms for both 1 and 2 that
are crystalline and likely suitable for further development. Our
first generation approach builds cyclobutylhomopiperazine from
N-Boc homopiperazine in two steps. Subsequent amidation of a
nicotinyl chloride and SNAr heteroarylꢀaryl ether formation
complete the synthesis. The route proceeds in four steps with an
overall yield of 59% and represents the first chromatography-free
synthesis of this class of molecules. Our second generation
approach proceeds from four readily available and reasonably
priced building blocks to provide 1 in three steps. The three steps
were achieved in 48% overall yield. The steps included SNAr
reaction, direct transformation of a nicotinic ester to a nicotinic
amide, and reductive amination. The key advance from the first
generation synthesis was the employment of homopiperazine
directly in the synthesis rather than the much more expensive
It is well-known that ester aminolysis may occur through
direct reaction between amines and esters under a variety of
conditions.18 Since the reaction is thermodynamically favor-
able,32 simple heating of the two components is often enough
to affect the reaction. The application of sodium methoxide as a
promoter of these reactions was demonstrated in 1937 by
Hammett during the course of his work on linear free-energy
relationships.33,34 In light of some of the more exotic variants of
this reaction that have been developed lately, this simple
Group 1A alkoxide-catalyzed variant may be underutilized.18,19
643
dx.doi.org/10.1021/op200005e |Org. Process Res. Dev. 2011, 15, 638–648

Organic Process Research & Development
ARTICLE
Scheme 3. Three Routes to the H3 Receptor Antagonists 1 and 2
Boc-homopiperazine. Our third generation approach proceeds
from the same four building blocks in just three steps to provide 2
in 48% overall yield. These steps include conversion of ethyl
6-chloronicotinate to the corresponding homopiperazine amide,
reductive amination of cyclobutanone, and SNAr reaction to form
the appropriate ether linkage. The linchpin step is formation of a
novel intermediate, (6-chloropyridin-3-yl)[1,4]diazepan-1-yl-
methanone, and is accomplished through an unusual reaction
in which ethyl 6-chloronicotinate and homopiperazine (2.5
equiv) are exposed to 0.5 equiv of n-hexyllithium, or alternatively,
to 1 equiv of LiOEt. The key advantages of the sequence are as
follows: (1) concise route with reasonable yields in each step; (2)
opportunity for purification of the step 2 product 9 through
tartrate salt formation and opportunity for purification of the
final step 3 products 1 or 2 through HCl salt formation (overall
chromatography-free approach); and (3) progression through a
final stage intermediate that is common to both 1 and 2. Further
development activities related to these compounds have pro-
vided additional opportunities for our downstream colleagues.
Although the second generation route was suitable for direct
adaptation to kilogram-scale synthesis, the third generation route
was not. In that case, the elimination of undesirable solvents
(dichloroethane and dichloromethane), the avoidance of strip-
ping to dryness operations, and a further improved synthesis/
isolation of 9 were critical for plant-scale operations. Those
achievements will form the focus of a future paper.
acetonitrile/95% water to 99% acetonitrile/1% water ramp over
3.8 min, then hold at 99% acetonitrile/1% water for 0.4 min.
Hydrochloric acid in 2-propanol was purchased as a 5ꢀ6 M
solution and used as 5 M. All solvents utilized were purchased in
anhydrous form.
First Generation Synthesis: 4-Cyclobutyl[1,4]diazepane-1-
carboxylic Acid tert-Butyl Ester (7).6b To a solution of
Boc-homopiperazine (6) (73.0 g, 365.0 mmol) and anhydrous
dichloroethane (800 mL) was added cyclobutanone (25.5 g,
363.8 mmol). The pale yellow reaction was stirred at ambient
temperature for 1 h, following which, sodium triacetoxyborohy-
dride (92.5 g, 436.3 mmol) was added portion-wise over 1 h. The
reaction mixture was stirred for 48 h. NaOH(aq) (1 N, 225 mL)
was added to the reaction mixture, which was then stirred for 1 h.
The phases were separated and the aqueous layer was extracted
with dichloroethane (2 ꢁ 100 mL). The organic layers were
pooled, washed with brine (1 ꢁ 250 mL), dried over anhydrous
sodium sulfate, and filtered. The solvent was removed by rotary
evaporation under reduced pressure to afford the crude product
as a pale yellow semisolid (92.4 g, 98%). 1H NMR (400 MHz,
CDCl3) δ 3.59ꢀ3.40 (m, 4H), 2.98ꢀ2.84 (m, 1H), 2.59ꢀ2.43
(m, 4H), 2.12ꢀ2.03 (m, 2H), 2.00ꢀ1.83 (m, 4H), 1.76ꢀ1.55
(m, 2H), 1.51ꢀ1.41 (s, 9H).
1-Cyclobutyl[1,4]diazepane Dihydrochloride (5 2HCl).6b
A
3
slurry of crude 4-cyclobutyl[1,4]diazepane-1-carboxylic acid tert-
butyl ester (7) (92.4 g, 363.8 mmol) in a mixture of dioxane/
MeOH (100 mL/50 mL) was treated with HCl in dioxane
(250 mL, Aldrich: 4M) followed by heating to 55 °C in an oil
bath. A pale orange-yellow solution resulted, and the tempera-
ture was maintained at 55 °C for 16 h. After cooling to ambient
temperature, the reaction mixture was concentrated to a pasty
solid. MTBE (200 mL) was added, and the slurry was agitated at
’ EXPERIMENTAL SECTION
HPLC conditions: column—Agilent Eclipse XDB-C8, 5 μm,
4.6 ꢁ 150 mm; flow rate—3 mL/min; mobile phases—acetoni-
trile with 0.05% TFA and water with 0.05% TFA; gradient—5%
644
dx.doi.org/10.1021/op200005e |Org. Process Res. Dev. 2011, 15, 638–648

Organic Process Research & Development
ARTICLE
ca. 55 °C in an oil bath for 1 h. The solvent was removed by rotary
evaporation under reduced pressure to afford the product as
an off-white solid (81.2 g, 98%). 1H NMR (400 MHz, DMSO-
d6) δ 11.92 (s, 1H), 9.87 (s, 1H), 9.46 (s, 1H), 3.76ꢀ3.03
(m, 9H), 2.43ꢀ2.33 (m, 2H), 2.17ꢀ2.15 (m, 4H), 1.75ꢀ1.60
(m, 2H).
iceꢀwater bath. Ethereal HCl (152 mL, 304 mmol, 0.95 equiv;
Aldrich: 2 M solution) was added dropwise via the addition
funnel. The addition was complete in 0.5 h. The reaction mixture
(a suspension) was stirred for 2 h, and then diluted with diethyl
ether (200 mL). The resulting suspension was stirred at ambient
temperature for 16 h. The suspension was recooled to 0 °C,
maintained at that temperature for 2 h with agitation, then
filtered. The filter-cake was broken and washed with Et2O/
EtOH (60:40, 3 ꢁ 100 mL) and the product was dried under
house vacuum for 1 h, then in a vacuum oven at 50 °C for 48 h.
The product (113.5 g) was suspended in Et2O (2 L) and agitated
(mechanical stirrer) for 6 h. The suspension was filtered, then the
filter-cake was broken and washed with Et2O (3 ꢁ 100 mL). The
product was dried in a vacuum oven at 45 °C for 16 h (106.8 g,
(6-Chloropyridin-3-yl)(4-cyclobutyl[1,4]diazepan-1-yl)meth-
anone (9). To a flask charged with 1-cyclobutyl[1,4]diazepane
dihydrochloride (5 2HCl) (110.0 g, 484.6 mmol), 1 N NaOH
3
(1400 mL), and isopropyl acetate (600 mL) was added a
precooled (0 °C) solution of 6-chloronicotinyl chloride (8)
(82.7 g, 470.0 mmol) in isopropyl acetate (800 mL) via an
addition funnel at a rate such that the reaction temperature was
maintained between 5 and 10 °C. After the addition was
complete, the reaction mixture was warmed to ambient tem-
perature and stirred for 2 h (pH of reaction mixture ca. 5.6). The
reaction mixture was basified with 2 N NaOH(aq) (to pH 13).
The phases were separated and the aqueous layer was extracted
with isopropyl acetate (2 ꢁ 300 mL). The organic layers were
pooled and filtered through a pad of Celite to remove some
reddish-brown flocculent material. The filtrate was washed with
brine (1 ꢁ 300 mL) and dried over anhydrous sodium sulfate.
Filtration and removal of the solvent by rotary evaporation under
reduced pressure afforded the crude product as a reddish brown
oil (126.0 g, 91%). 1H NMR (400 MHz, CDCl3) δ 8.45 (m, 1H),
7.74 (m, 1H), 7.39 (m, 1H), 3.86ꢀ3.76 (m, 2H), 3.53ꢀ3.45 (m,
2H), 3.05ꢀ2.80 (m, 1H), 2.70 (br s, 1H), 2.62ꢀ2.45 (m 3H),
2.10ꢀ1.56 (m, 8H). MS (electrospray): calcd 293.1; m/z found
294.1 [M þ H]þ.
1
82%). H NMR (400 MHz, D2O) δ 8.09ꢀ8.05 (m, 1H),
7.88ꢀ7.81 (m, 1H), 7.13ꢀ7.09 (m, 4H), 7.02 (d, J = 8.64 Hz,
1H), 4.15ꢀ3.39 (m, 7H), 3.07ꢀ2.90 (m, 2H), 2.28ꢀ1.99 (m,
6H), 1.75ꢀ1.63 (m, 2H). 13C NMR (150 MHz, D2O) δ 2
rotamers (171.1, 170.8), (164.62, 164.55), 160.0 (d, JCꢀF = 242
Hz), 148.8, (145.8, 145.4), (140.0, 139.7), (126.0, 125.7), 123.0
(d, JCꢀF = 9 Hz), 116.7 (d, JCꢀF = 23 Hz), (111.8, 111.6), 59.55,
6 carbons (51.53, 51.51, 50.6, 49.5, 48.2, 44.8, 44.7, 41.2, 26.0,
25.9, 24.2, 22.8), (12.3, 12.2). Anal. Calcd for C21H24FN3O2
3
HCl H2O: C, 59.50; H, 6.42; N, 9.91. Found: C, 59.36; H, 6.66;
3
N, 9.98.
Second Generation Synthesis: Ethyl 6-(4-Fluorophenoxy)-
nicotinate (11) To a solution of 6-chloronicotinate (10)
(100.00 g, 0.522 mol), and 4-fluorophenol (65.09 g, 0.575 mol)
in DMF (194 mL) was added Cs2CO3 (189.19 g, 0.575 mol) in
one portion. The reaction temperature increased from 20 to
30 °C over 10 min without external heating and then started
cooling down. The resulting suspension was stirred at rt for 2ꢀ3 h
and the internal reaction temperature cooled back to 23ꢀ25 °C.
The reaction mixture was then heated to 60 °C and stirred for
18ꢀ20 h. HPLC analysis indicated that the reaction was com-
plete. The heating mantle was removed and the reaction mixture
was allowed to cool to 25ꢀ30 °C. To the mixture was added
deionized water (146 mL) in a steady stream over 5 min and a
slight exotherm was observed. The resulting suspension was
stirred at rt for 15ꢀ20 min. Two additional portions of deionized
water (146 mL each) were added and the suspension was stirred
at rt for 15ꢀ30 min. The pH of the suspension was around 9ꢀ10.
The solid product was collected by vacuum filtration, rinsed
thoroughly with deionized water in portions. The filter cake was
dried in a filter funnel by pulling through air for 24 h. The product
was isolated as a white solid (133.6 g, 98%). Mp 68 °C (by DSC).
1H NMR (CDCl3) δ 8.81 (d, J = 2.6 Hz, 1H), 8.22 (dd, J = 8.5,
2.6 Hz, 1H), 7.12 (br s, 2H), 7.10 (d, J = 1.0 Hz, H), 6.94
(d, J = 8.5 Hz, 1H), 4.38 (q, J = 7.1 Hz, 2H), 4.38 (t, J = 7.1 Hz,
3H). 13C NMR (150 MHz, CDCl3) δ 166.3, 165.0, 159.9
(d, JCꢀF = 244 Hz), 150.2, 149.1 (d, JCꢀF = 3 Hz), 140.6,
123.0 (d, JCꢀF = 8 Hz), 121.6, 116.4 (d, JCꢀF = 24 Hz), 110.7,
61.2, 14.3.
(4-Cyclobutyl[1,4]diazepan-1-yl)[6-(4-fluorophenoxy)pyridin-
3-yl]methanone (1). A suspension of anhydrous DMA
(625 mL), 4-fluorophenol (57.3 g, 511.6 mmol), and Cs2CO3
(278.0 g, 853.3 mmol) was stirred for 15 min, and then a solution
of crude (6-chloropyridin-3-yl)(4-cyclobutyl[1,4]diazepan-1-
yl)methanone (9) (125.0 g, 426.6 mmol) in anhydrous DMA
(625 mL) was added via an addition funnel over 0.5 h. The
reaction mixture was heated to 100 °C and maintained at that
temperature for 12 h. The reaction mixture was cooled to room
temperature and filtered through a pad of Celite (3 in. Celite pad
in a 600 mL coarse glass frit), then the Celite pad was washed
with DMA (2 ꢁ 125 mL). The reaction mixture was diluted with
a slurry of ice/water (ca. 1 L) and 2 N NaOH(aq) (500 mL). The
pH of the reaction mixture was 13. Out of convenience, the
reaction mixture was divided into two approximately equal
portions. Each portion was extracted with MTBE (3 ꢁ
300 mL). The organic layers were pooled, dried over anhydrous
MgSO4, filtered, and concentrated by rotary evaporation under
reduced pressure to afford the crude product as a thick, orange-
colored oil (129 g). The oil was taken up in anhydrous diethyl
ether (1 L) and stirred overnight. After removal of the resulting
solids by filtration, the filtrate was concentrated to provide 1 as an
orange oil (120 g, 76%). 1H NMR (400 MHz, CDCl3) δ 8.23 (br
s, 1H), 7.79 (dd, J = 8.48, 2.39 Hz, 1H), 7.11ꢀ7.10 (m, 4H), 6.95
(d, J = 8.48 Hz, 1H), 3.76 (t, J = 5.84 Hz, 2H), 3.55ꢀ3.50 (m,
2H), 2.90ꢀ2.84 (m, 1H), 2.61 (t, J = 4.71 Hz, 1H), 2.55ꢀ2.37
(m, 3H), 2.06ꢀ1.92 (m, 3H), 1.88ꢀ1.56 (m, 5H). MS
(electrospray): calcd 369.2; m/z found 370.1 [M þ H]þ.
(4-Cyclobutyl[1,4]diazepan-1-yl)[6-(4-fluorophenoxy)pyridin-
(1,4-Diazepan-1-yl)(6-(4-fluorophenoxy)pyridine-3-yl)metha-
none (12). Homopiperazine (234.4 g, 2.34 mol) and ethyl 6-(4-
fluorophenoxy)nicotinate (11) (251.34 g, 0.962 mol) in THF
(2 L) were cooled to 0 °C and a solution of hexyllithium (2.3 M
in hexane, 440.41 g, 1.43 mol) was added via addition funnel over
1 h. The reaction mixture was warmed to 25 °C and stirred for
2 h. The reaction was cooled to 15 °C and quenched with water
(1 L). The organic layer was extracted with 600 mL of a 2 N
HCl(aq) solution. The layers were mixed then separated, and the
3-yl]methanone Hydrochloride Monohydrate (1 HCl H2O). A
3
3
solution of crude (4-cyclobutyl[1,4]diazepan-1-yl)[6-(4-fluoro-
phenoxy)pyridin-3-yl]methanone (1) (118.0 g, 319.8 mmol) in
ethanol/diethyl ether (400 mL each) was cooled to 5 °C in an
645
dx.doi.org/10.1021/op200005e |Org. Process Res. Dev. 2011, 15, 638–648

Organic Process Research & Development
ARTICLE
organic layer was extracted with additional 2 N HCl(aq)
(200 mL). The combined aqueous layers were allowed to sit at
room temperature for 12 h. The resulting solids were removed by
filtration through a glass fiber filter. The filtrate was then
extracted with tert-butyl methyl ether (7 ꢁ 180 mL) to remove
residual diamide byproduct. A solution of NaOH (50% w/w,
175.4 g, 2.19 mol) was added to the aqueous layer until a pH of
greater than 10 was reached. The layers were separated and the
aqueous layer was extracted with additional ethyl acetate (2 ꢁ
150 mL). The combined organic layers were then washed with
brine (100 mL). Ethyl acetate (100 mL) was added to remove
water through azeotropic distillation. Additional ethyl acetate
(600 mL) was added in several portions and distillation was
continued. The residue was filtered hot to remove sodium
chloride and the filtrate was stirred for several hours until a large
amount of solid had formed. The mixture was treated slowly with
heptane (500 mL) over 1 h and then stirred for an additional
hour. The resulting solids were isolated by filtration and dried to
give the title compound as an off-white to yellow solid (168.9 g,
mmol) were suspended in tert-amyl alcohol (285 mL). The
reaction mixture was cooled to 0 °C and 6-chloronicotinoyl
chloride (8) (23.4 g, 133 mmol) was added as a solid over 20 min.
The resulting mixture was stirred overnight at room temperature.
The reaction was quenched with water (285 mL) and the layers
were separated. The aqueous was extracted with additional ethyl
acetate (285 mL). The combined organics were then extracted
with a NaOAc/HOAc buffer (pH ∼3.74). The resulting aqueous
layer was basified with a 50% NaOH(aq) solution until a pH of
13.4 was reached. The aqueous was then extracted with addi-
tional isopropyl acetate (3 ꢁ 500 mL) and chloroform (1 ꢁ
500 mL). The combined organics were dried over sodium sulfate,
filtered, and concentrated under reduced pressure to provide the
desired product 13 (11.9 g, 37% yield).
(6-Chloropyridin-3-yl)[1,4]diazepan-1-yl-methanone (13)
(Method B). A solution of homopiperazine (385.62 g, 3.85 mol)
in THF (3.9 L) was cooled to an internal temperature of 0 °C and
ethyl 6-chloronicotinate (10) (285.82 g, 1.54 mol) was added in
THF (0.57 L) over 5 min. After the mixture was stirred for 10
min, n-hexyllithium (2.3 M in hexane, 335 mL, 0.77 mol) was
added over 40 min. Although the addition was exothermic, the
maximum temperature was maintained at less than 5 °C. The
reaction mixture was stirred for 2 h at 0 °C and then warmed to
20 °C over 1 h. The mixture was then stirred at 20 °C for 15 h, at
which time HPLC analysis revealed no residual starting material.
The reaction mixture was treated with 5 L of a 1 M NaOAc/
HOAc buffer prepared by diluting 47.35 g of sodium acetate and
253.2 mL of acetic acid with water to a total volume of 5 L. The
aqueous layer was found to contain the desired product and a
reduced amount of impurity 14, relative to the crude reaction
product. The aqueous layer pH was then increased from 8.0 to
11.35 with 153 mL of 50% NaOH(aq) solution. The basic layer
was extracted with dichloromethane (2 ꢁ 4 L) and the resulting
organics dried with sodium sulfate, filtered, and concentrated to a
thick oil (292.73 g crude, 87.1 wt % desired, 255.1 g actual
product, 69% yield). By 1H NMR, the crude reaction product was
also found to contain 0.3 wt % of dichloromethane, 3.7 wt % of
THF, an estimated 4.4 wt % of various impurities including 14
1
56%). Mp 95.4 °C (by DSC). H NMR (500 MHz, CDCl3)
δ 8.24 (dd, J = 2.4, 0.6 Hz, 1H), 7.79 (dd, J = 8.5, 2.4 Hz, 1H),
7.16ꢀ7.06 (m, 4H), 6.95 (d, J = 8.4 Hz, 1H), 3.76 (br s, 2H),
3.59ꢀ3.45 (br m, 2H), 3.12ꢀ2.84 (br m, 4H), 1.97ꢀ1.69 (br m,
2H). 13C NMR (150 MHz, CDCl3) δ 2 rotamers 168.8, 164.0,
159.8 (d, JCꢀF = 244 Hz), 149.3, (146.1, 145.9), (139.0, 138.8),
127.7, 122.9 (d, JCꢀF = 8 Hz), 116.4 (d, JCꢀF = 23 Hz), 111.1,
5 carbons (52.8, 50.6, 49.5, 49.3, 48.9, 48.4, 47.8, 45.5, 31.7,
29.2). HRMS (electrospray): calcd 316.1456; m/z found
316.1453 [M þ H]þ.
(4-Cyclobutyl-1,4-diazepan-1-yl)(6-(4-fluorophenoxy)pyri-
dince-3-yl)methanone Hydrochloride Monohydrate (1). A
suspension of (1,4-diazepan-1-yl)(6-(4-fluorophenoxy)pyridin-
3-yl)methanone (12) (136.2 g, 428 mmol) in ethyl acetate
(1.08 L) was warmed to 35 °C to dissolve all solids. The solution
was cooled to 5ꢀ10 °C, cyclobutanone was added (36.33 g,
513 mmol), and the resulting mixture was stirred for an addi-
tional 10 min. Sodium triacetoxyborohydride (143.09 g, 641
mmol) was then added in several portions over 20 min. The
resulting suspension was stirred at 15 °C for 15 min, and then at
ambient temperature for 3 h. The reaction was quenched with a
solution of potassium carbonate (164.15 g, 1.175 mol) in
deionized water (540 mL). After the mixture was stirred for
30 min, the layers were separated and the organic layer was further
washed with deionized water (3 ꢁ 540 mL). The organic layer
was filtered through a pad of Celite (10 g) and rinsed with ethyl
acetate (270 mL). The combined organic layer was assayed by
HPLC to contain 143.88 g of freebase, which corresponds to a
90% crude yield. The solution was diluted with ethanol (270 mL)
and a solution of concd HCl(aq) (36.73 g, 368 mmol, 0.95 equiv of
the freebase based on HPLC assay) in ethanol (67 mL) was added
slowly. The resulting solution was stirred at ambient temperature
for 24 h, during which time the desired product precipitated out.
The resulting solids were filtered, washed with ethyl acetate, and
dried in a vacuum oven at 50 °C for 24 h. The product was then
rehydrated in a sealed oven at ambient temperature in the
1
and 18, and 4.5 wt % of 16. H NMR (400 MHz, CDCl3) δ
8.46ꢀ8.45 (m, 1H), 7.75ꢀ7.72 (m, 1H), 7.40ꢀ7.38 (m, 1H),
3.80ꢀ3.75 (m, 2H), 3.49ꢀ3.44 (m, 2H), 3.09ꢀ3.06 (m, 1H),
2.96ꢀ2.89 (m, 3H), 1.95ꢀ1.88 (m, 1H), 1.75ꢀ1.70 (m, 1H).
13C NMR (125 MHz, CDCl3) δ 2 rotamers 167.9, (152.2,
152.1), (147.8, 147.5), (137.6, 137.3), 131.6, (124.2), 5 carbons
(52.6, 50.5, 49.4, 49.3, 48.8, 48.3, 47.8, 45.4, 31.6, 29.0). HRMS
(electrospray): calcd 240.0898; m/z found 240.0885 [M þ H]þ.
For impurity 16: 1H NMR (400 MHz, CDCl3) δ 8.79 (dd, J =
2.3, 0.6 Hz, 1H), 7.98 (dd, J = 9.0, 2.4 Hz, 1H), 6.46 (dd, J = 9.0,
0.7 Hz, 1H), 4.32 (q, J = 6.9 Hz, 2H), 3.82ꢀ3.75 (m, 4H),
3.04ꢀ3.01 (m, 2H), 2.85ꢀ2.82 (m, 2H), 1.91ꢀ1.86 (m, 2H),
1.36 (t, J = 7.1, 3H). The structure was assigned on the basis of
HMBC spectrum showing ethyl CH2 interacting with two
quaternary carbons. MS (electrospray): calcd 249.2; m/z found
250.2 [M þ H]þ.
(6-Chloropyridin-3-yl)[1,4]diazepan-1-yl-methanone (13)
(Method C). Homopiperazine (12.52 g, 125 mmol) and ethyl
6-chloronicotinate (10) (9.28 g, 50 mmol) were taken up in
tetrahydrofuran (150 mL). After the mixture was cooled to 0 °C,
LiOEt (1 M in THF, 50 mL, 50 mmol) was added over 20 min.
The mixture was stirred for 2 h at 0 °C and then warmed to room
temperature and held overnight. To the reaction mixture was
then added 162 mL of an aqueous solution containing sodium
presence of a saturated solution of ZnSO4 7H2O (100 g, 348
3
mmol) in deionized water (200 mL) for 24ꢀ48 h to form the
monohydrate of the desired product as a white solid (125 g, 69%).
Karl Fischer analysis: ∼4.25 wt % water. Mp 143.4 °C (by DSC).
Third Generation Synthesis: (6-Chloropyridin-3-yl)[1,4]-
diazepan-1-yl-methanone (13) (Method A) Homopiperazine
(33.35 g, 333 mmol) and potassium carbonate (36.8 g, 266
646
dx.doi.org/10.1021/op200005e |Org. Process Res. Dev. 2011, 15, 638–648

Organic Process Research & Development
ARTICLE
acetate (1.53 g) and acetic acid (8.2 mL). The layers were mixed
and then separated. To the organic layer was added hexanes
(50 mL), and it was extracted again with 162 mL of the sodium
acetate/acetic acid solution described above. The combined
aqueous layers (pH 6) were made basic (pH 10) through
addition of 50% NaOH(aq) solution (15 mL). The aqueous layer
was then extracted with dichloromethane (3 ꢁ 250 mL), and the
combined organics were dried over sodium sulfate and concen-
trated to dryness. The resulting crude oil (15.05 g) was
calculated to contain approximately 72 wt % of the desired
product (10.8 g, 90% crude yield). Further purification, if
desired, was possible through reverse phase HPLC under basic
conditions (Waters X-Bridge prep C18 5 μm 30 ꢁ 100 mm
OBD column, 80 mL/min, Solvent A = 10 mM ammonium
hydroxide in water, Solvent B = acetonitrile, 5% B for 2 min
ramping to 99% B over 15 min). The resulting eluent was
concentrated, saturated with sodium chloride, and extracted into
dichloromethane (3 ꢁ 100 mL). After concentration, 13 was
afforded in 67% overall yield.
yl)[1,4]diazepan-1-ylmethanone). 1H NMR (400 MHz,
CDCl3) δ 8.46ꢀ8.45 (m, 1H), 7.75ꢀ7.72 (m, 1H),
7.40ꢀ7.38 (m, 1H), 3.80ꢀ3.75 (m, 2H), 3.49ꢀ3.44 (m,
2H), 3.09ꢀ3.06 (m, 1H), 2.96ꢀ2.89 (m, 3H), 1.95ꢀ1.88
(m, 1H), 1.75ꢀ1.70 (M, 1H). 13C NMR (500 MHz, CDCl3)
δ 2 rotamers 167.8, (152.2, 152.1), (147.8, 147.7), (137.6,
137.5), (131.6, 131.5), 124.2, (59.60, 59.58), 4 carbons (52.8,
51.9, 51.0, 50.3, 50.1, 48.8, 46.5, 45.6), (28.9, 26.8), 28.0,
(13.5, 13.4). MS (electrospray): calcd 239.1; m/z found
240.1 [M þ H]þ.
3-[5-(4-Cyclobutyl[1,4]diazepane-1-carbonyl)pyridin-2-ylo-
xy]benzonitrile (2). To a solution of (6-chloropyridin-3-yl)(4-
cyclobutyl-[1,4]diazepan-1-yl)methanone (9) (101.0 g, 343.8
mmol) in dimethylacetamide (1.1 L) was added Cs2CO3 (224
g, 687.6 mmol) and m-cyanophenol (81.9, 687.6 mmol). The
mixture was warmed to 125 °C and stirred for 20 h. After cooling
to room temperature, the reaction mixture was filtered and acetic
acid (1.5 L) was added to the filtrate (Note: acetic acid is known
to form an azeotrope with DMA and was added to assist in its
distillation). The resulting mixture was concentrated under
reduced pressure to a brown crude, which was then partitioned
between MTBE (1.5 L) and 1 N NaOH(aq) (1.5 L). The layers
were thoroughly mixed and then separated. The organic extract
was dried over magnesium sulfate, filtered, and concentrated to
provide the title compound as a brown oil (128.4 g, 89 wt %
(6-Chloropyridin-3-yl)(4-cyclobutyl[1,4]diazepan-1-yl)met-
hanone (9). To a solution of (6-chloropyridin-3-yl)[1,4]-
diazepan-1-ylmethanone (13) (from Method B above, 255.1 g,
1.06 mol; note that the starting material also contained ∼53
mmol of impurity 16 from the previous step; thus total secondary
amine was estimated at 1.11 mol) in dichloroethane (3.0 L) was
added cyclobutanone (108.1 mL, 1.45 mol), and then the
solution was allowed to stir for 1 h. Sodium triacetoxyborohy-
dride (308.2 g, 1.45 mol) was then added in four equal portions
over 1.5 h. Each addition was accompanied by an exotherm of
about 4 °C. The reaction mixture was allowed to stir for 20 h and
then quenched with 2.5 L of an aqueous solution containing
NaOH (141.3 g, 3.53 mol). After the mixture was stirred for 30
min, the layers were separated and the organic dried with
magnesium sulfate, filtered, and concentrated to provide the
desired product as an oil (360.07 g crude, 86.5 wt % desired,
1
desired, 114.3 g actual product, 90% yield). By H NMR, the
material was also found to contain 2.5 wt % of dimethylacetamide
1
and 8.5 wt % of tert-butyl methyl ether. H NMR (400 MHz,
CDCl3) δ 8.22 (s, 1H), 7.84 (dd, J = 8.4, 2.4 Hz, 1H), 7.55ꢀ7.37
(m, 4H), 7.03 (d, J = 8.4 Hz, 1H), 3.77 (m, 2H), 3.53 (m, 2H),
2.98ꢀ2.80 (m, 1H), 2.70ꢀ2.58 (m, 1H), 2.55ꢀ2.35 (m, 3H),
2.15ꢀ1.53 (m, 8H). 13C NMR (150 MHz, CDCl3) δ 2 rotamers
(168.5, 168.4), (162.94, 162.88), 153.8, (145.93, 145.89), (139.4,
139.2), 130.6, 128.6, 128.5, 126.3, 125.0, 118.1, 113.6, 111.7,
59.6, 3 carbons (53.1, 51.9, 51.2, 50.3, 50.2, 48.9), (46.5, 45.8),
(29.0, 26.8), 28.1, 13.5.
1
311.46 g actual product, quantitative yield). By H NMR, the
material was also found to contain 4.4 wt % of the reductive
amination product of 16 as well as 5.6 wt % of dichloroethane,
and an estimated 3.5 wt % of other impurities. The material was
purified through formation of the corresponding tartrate salt as
follows: Crude (6-chloropyridin-3-yl)(4-cyclobutyl[1,4]diazepam-
1-yl)methanone (311.5 g actual desired, 1.06 mol) was taken up
in ethanol (3.0 L). To the mixture was added L-tartaric acid
(167.05 g, 1.11 mol). The heterogeneous suspension was
warmed to 80 °C over 45 min and held for 1 h. The mixture
was then cooled to 20 °C over 3 h and stirred at 20 °C for 1 h. The
resulting solids were filtered and washed with 1 L of ethanol
(Cake cracking observed—precautions must be taken to ensure
efficient cake washing). The material was dried under vacuum at
43 °C to provide an off-white solid (432.52 g crude, 95 wt %
3-[5-(4-Cyclobutyl[1,4]diazepane-1-carbonyl)pyridin-2-yl-
oxy]benzonitrile HCl (2 HCl). A solution of 3-[5-(4-cyclobutyl-
3
3
[1,4]diazepane-1-carbonyl)pyridin-2-yloxy]benzonitrile (2) (114g,
302.8 mmol) in IPA (900 mL) was warmed to 40 °C. To the
resulting solution was added anhydrous HCl (5 M solution in
IPA, 60.6 mL, 302.8 mmol). After the addition of seed crystals,
the mixture was cooled to 35 °C and held for 2 h. After cooling to
room temperature, filtering, washing with IPA (220 mL), and
drying at 50 °C, the title compound was provided as an off-white
crystalline solid (104.2 g, up to 99.6 wt % desired, 4000ꢀ6000
ppm residual IPA, 83% yield). Note that slight cake shrinkage was
observed after removal of the wash layer. Note also, if residual
chloroamido pyridine 9 persists, it may be removed by additional
trituration in IPA. 1H NMR (400 MHz, DMSO) δ 11.45 (br s,
1H), 8.29 (br s, 1H), 8.01 (br d, J = 7.8 Hz, 1H), 7.82ꢀ7.5 (m,
4H), 7.2 (d, J = 8.5 Hz, 1H), 4.1ꢀ3.3 (m, 7H), 3.1ꢀ2.8 (m, 2H),
2.49ꢀ2.25 (m, 3H), 2.25ꢀ1.9 (m, 3H), 1.8ꢀ1.55 (m, 2H). 13C
NMR (150 MHz, DMSO) δ167.8, 162.7, 153.5, 146.2, 139.7, 131.2,
128.9, 127.6, 126.9, 125.1, 118.1, 112.4, 111.2, 58.4, 50.2, 48.7, 47.7,
40.2, 25.2, 23.5, 12.6. Anal. Calcd for C22H25ClN4O2: C, 63.99; H,
6.1; N, 13.57; Cl, 8.59. Found: C, 63.78; H, 6.15; N, 13.33; Cl, 8.39.
1
desired, 411 g actual product, 87% yield). By H NMR, the
material was also found to contain 5 wt % of residual ethanol. A
portion of the total tartrate salt was then free based as follows: To
a 4 L Erlenmeyer was added (6-chloropyridin-3-yl)(4-cyclo-
butyl[1,4]diazepan-1-yl)methanone L-tartaric acid (172 g,
3
386.9 mmol), iPrOAc (1.5 L), and 1 N NaOH(aq) (1.5 L). The
layers were thoroughly mixed and then separated. The aqueous
layer was extracted with additional iPrOAc (1.5 L) and the
combined organic layers were dried over magnesium sulfate.
After filtration and concentration, the title compound was
obtained as a yellow oil (101.1 g, 89% yield for free base
procedure, 77% overall yield from (6-chloropyridin-3-
’ ASSOCIATED CONTENT
Supporting Information. 1H and 13C spectra for key
S
b
intermediates 9, 13, 11, and 12, as well as final products 1 HCl,
3
647
dx.doi.org/10.1021/op200005e |Org. Process Res. Dev. 2011, 15, 638–648

Organic Process Research & Development
ARTICLE
2, and 2 HCl. This material is available free of charge via the
Internet at http://pubs.acs.org.
65, 4740–4742. (b) Chou, W.-C.; Tan, C.-W.; Chen, S.-F.; Ku, H.
J. Org. Chem. 1998, 63, 10015–10017.
(18) For microwave-mediated ester to amide transformation (and
leading references on ester aminolysis in general), see: Varma, R. S.;
Naicker, K. P. Tetrahedron Lett. 1999, 40, 6177–6180.
3
’ AUTHOR INFORMATION
(19) For aluminum-mediated ester to amide transformations, see:
(a) Novak, A.; Humphreys, L. D.; Walker, M. D.; Woodward, S.
Tetrahedron Lett. 2006, 47, 5767–5769. (b) Huang, P.-Q.; Zheng, X.;
Deng, X.-M. Tetrahedron Lett. 2001, 42, 9039–9041. (c) Basha, A.;
Lipton, M.; Weinreb, S. M. Tetrahedron Lett. 1977, 48, 4171–4174.
(20) For sodium amide reactions with esters to form N-aryl
amides, see: (a) Walker, M. D.; Andrews, B. I.; Burton, A. J.; Humphreys,
L. D.; Kelly, G.; Schilling, M. B.; Scott, P. W. Org. Process Res. Dev. 2010,
14, 108–113.(b) Wang, J.; Rosingana, M.; Discordia, R. P.; Soundarar-
ajan, N.; Polniaszek, R. Synlett 2001, 1485ꢀ1487.
Corresponding Author
*E-mail: dpippel@its.jnj.com.
’ REFERENCES
(1) Gemkow, M. J.; Davenport, A. J.; Harich, S.; Ellenbroek, B. A.;
Cesura, A.; Hallett, D. Drug Discovery Today 2009, 14, 509–515.
(2) Wijtmans, M.; Leurs, R.; de Esch, I. Expert Opin. Invest. Drugs
2007, 16, 967–985.
(3) Leurs, R.; Bakker, R. A.; Timmerman, H.; de Esch, I. J. P. Nat.
Rev. Drug Discovery 2005, 4, 107–120.
(21) Even without the requirement for a late stage common inter-
mediate, the second generation synthesis was not directly applicable to
2. The hexyllithium-mediated amide bond forming step in that case was
not high yielding.
(4) Celanire, S.; Wijtmans, M.; Talaga, P.; Leurs, R.; de Esch, I. J. P.
Drug Discovery Today 2005, 10, 1613–1627.
(5) Bonaventure, P.; Letavic, M.; Dugovic, C.; Wilson, S.; Aluisio, L.;
Pudiak, C.; Lord, B.; Mazur, C.; Kamme, F.; Nishino, S.; Carruthers, N.;
Lovenberg, T. Biochem. Pharmacol. 2007, 73, 1084–1096.
(6) (a) Stocking, E. M.; Aluisio, L.; Atack, J. R.; Bonaventure, P.;
Carruthers, N. I.; Dugovic, C.; Everson, A.; Fraser, I.; Jiang, X.;
Leung, P.; Lord, B.; Ly, K. S.; Morton, K. L.; Nepomeceno, D.; Shah,
C. R.; Shelton, J.; Soyode-Johnson, A.; Letavic, M. A. Bioorg. Med.
Chem. Lett. 2010, 20, 2755–2760. (b) Pippel, D. J.; Young, L. K.;
Letavic, M. A.; Ly, K. S.; Naderi, B.; Soyode-Johnson, A.; Stocking, E. M.;
Carruthers, N. I.; Mani, N. S. J. Org. Chem. 2010, 75, 4463–4471.
(7) Letavic, M. A.; Aluisio, L.; Atack, J. R.; Bonaventure, P.;
Carruthers, N. I.; Dugovic, C.; Everson, A.; Feinstein, M. A.; Fraser,
I. C.; Hoey, K.; Jiang, X.; Keith, J. M.; Koudriakova, T.; Leung, P.; Lord,
B.; Lovenberg, T. W.; Ly, K. S.; Morton, K. L.; Motley, T.; Nepomuceno,
D.; Rizzolio, M.; Rynberg, R.; Sepassi, K.; Shelton, J. Bioorg. Med. Chem.
Lett. 2010, 20, 4210–4214.
(22) The side product 16 is a structural isomer of the potential side
product 18. The assignment of 16 was made after isolation of the pure
material and 2-dimensional HMBC NMR analysis (2 to 4 bond
couplings), which showed that the protons on the ethyl CH2 group
are correlated to two quaternary carbons with chemical shifts of 166.18
and 113.89 ppm (the carbonyl carbon and 3-position of the pyridine
ring). In the case of 18, correlation between the ethyl CH2 group and just
one quaternary carbon would be expected (the 6-position of the pyridine
ring).
(23) The n-hexyllithium mediated coupling of 10 and homopiper-
azine was also attempted in toluene and MTBE. The reaction in toluene
produced more of the impurity 16, relative to the THF case. The
reaction in MTBE resulted in even more of the impurity 16 and
proceeded at a retarded rate.
(24) At this point in our study, we thought that the LiOEt generated
from the desired reaction would react with 10 in the absence of sufficient
homopiperazine lithium amide to form the undesired 17. We subse-
quently found that not to be the case.
(25) A reaction with 0.1 equiv of hexyllithium provided some 13, but
16 was the major product.
(8) For a recent leading reference on this type of carbonylation, see:
Lagerlund, O.; Mantel, M. L. H.; Larhed, M. Tetrahedron 2009,
65, 7646–7652.
(9) For aminocarbonylation of heteroaromatic halides see: Letavic,
M. A.; Ly, K. S. Tetrahedron Lett. 2007, 48, 2339–2343.
(26) It should be noted that our failed large-scale reactions were all
carefully controlled with slow addition of reagents and internal tem-
perature monitoring.
(27) Even the most commonly employed lithium amide in organic
synthesis (LDA) has demonstrated exceptionally complex properties
based on solvation and aggregation, which in turn have marked effects on
its reactivity: Collum, D. B.; McNeil, A. J.; Ramirez, A. Angew. Chem., Int.
Ed. 2007, 46, 3002–3017.
(28) A drawback to this approach is the fact that we currently employ
methylene chloride for the extraction of the secondary amine during
aqueous workup. We have not yet discovered a suitable replacement.
(29) It was also possible to utilize 2-methyl tetrahydrofuran as a
solvent for this step; however, in that case, the starting material was
somewhat insoluble.
(10) HerrmannꢀBeller catalyst: trans-di(μ-acetato)bis[o-(di-o-
tolylphosphino)benzyl]dipalladium(II) [Registry No. 172418-32-5].
(11) Catalog prices for 5-bromo-2-fluoropyridine are typically high-
er than for 6-chloronicotinic acid from any given single supplier. For
example, AK Scientific lists the former for $272/100 g and the latter for
$50/100 g.
(12) Kovganko, N. V.; Kashkan, Zh. N. Russ. J. Org. Chem. 2004,
40, 1709–1726.
(13) (a) Zhang, J.; Dong, F.; Tong, L.; Yu, Y. Faming Zhuanli
Shenqing Gongkai Shuomingshu, CN Patent 1803772, 2006, CAN
145:188729. (b) Marinak, M. J.; Simonson, J. L. U.S. Patent 4497955,
1985, CAN 102:166623.
(14) This experimental procedure has been previously documented
in the form of a patent: Keith, J. M.; Letavic, M. A.; Ly, K. S.; Mani, N. S.;
Mills, J. E.; Pandit, C. R.; Villani, F. J.; Zhong, H. U.S. Patent
20070281923, 2007, CAN 148:54895.
(15) Standard Sigma-Aldrich pricing: N-Boc-homopiperazine,
$3180/mol; homopiperazine $124/mol.
(16) The reaction profile is quite clean with other possible bypro-
ducts (i.e. addition of hexyllithium to the ester) not observed to any
significant extent. Simple concentration of a typical crude reaction
mixture (starting with 3.8 mmol of 11) followed by flash chromatogra-
phy provides a fraction of the desired product 12 (2.26 mmol) and a
fraction of the pseudodimer (0.54 mmol). As 2 mmol of 11 are
incorporated into each mmol of pseudodimer, these two products
account for nearly 90% of the starting 11.
(30) An alternative, less desirable polymorph was formed when the
salt formation was conducted in ethanol.
(31) LiOEt was used as a 1 M solution in THF from a commercial
supplier. We were unsuccessful in our attempts to use commercial solid
LiOEt for these reactions. Due to competitive reaction to form 16,
reactions with substoichiometric LiOEt were not high yielding. In fact,
with older bottles of LiOEt that were likely not up to titer, an “excess” of
the reagent was required to avoid substantial 16 formation.
(32) Gresser, M. J.; Jencks, W. P. J. Am. Chem. Soc. 1977, 99, 6963–6970.
(33) Betts, R. L.; Hammett, L. P. J. Am. Chem. Soc. 1937, 59, 1568–
1572.
(34) For other early examples, see: (a) Meade, E. M. U.S. Patent
2464094, 1949, CAN 43:22693. (b) Schurman, J. V. U.S. Patent
2863888, 1958, CAN 53:42539.
(17) For desymmetrization of the more common piperazine, see: (a)
Wang, T.; Zhongxing, Z.; Meanwell, N. A. J. Org. Chem. 2000,
648
dx.doi.org/10.1021/op200005e |Org. Process Res. Dev. 2011, 15, 638–648