Synthesis of the 6-azaindole containing HIV-1 attachment inhibitor pro-drug, BMS-663068

Article abstract of DOI:10.1021/jo5016008

The development of a short and efficient synthesis of a complex 6-azaindole, BMS-663068, is described. Construction of the 6-azaindole core is quickly accomplished starting from a simple pyrrole, via a regioselective Friedel.Crafts acylation, Pictet.Spengler cyclization, and a radical-mediated aromatization. The synthesis leverages an unusual heterocyclic Noxide α-bromination to functionalize a critical C.H bond, enabling a highly regioselective copper-mediated Ullmann. Goldberg.Buchwald coupling to install a challenging triazole substituent. This strategy resulted in an efficient 11 step linear synthesis of this complex clinical candidate.

Full text of DOI:10.1021/jo5016008

Article
pubs.acs.org/joc
Synthesis of the 6‑Azaindole Containing HIV‑1 Attachment Inhibitor
Pro-Drug, BMS-663068
Ke Chen, Christina Risatti, Michael Bultman, Maxime Soumeillant, James Simpson, Bin Zheng,
Dayne Fanfair, Michelle Mahoney, Boguslaw Mudryk, Richard J. Fox, Yi Hsaio, Saravanababu Murugesan,
David A. Conlon, Frederic G. Buono,† and Martin D. Eastgate*
Chemical Development, Bristol-Myers Squibb, 1 Squibb Drive, New Brunswick, New Jersey 08903, United States
S
* Supporting Information
ABSTRACT: The development of a short and efficient synthesis of a complex 6-azaindole, BMS-663068, is described.
Construction of the 6-azaindole core is quickly accomplished starting from a simple pyrrole, via a regioselective Friedel−Crafts
acylation, Pictet−Spengler cyclization, and a radical-mediated aromatization. The synthesis leverages an unusual heterocyclic N-
oxide α-bromination to functionalize a critical C−H bond, enabling a highly regioselective copper-mediated Ullmann−
Goldberg−Buchwald coupling to install a challenging triazole substituent. This strategy resulted in an efficient 11 step linear
synthesis of this complex clinical candidate.
complexly functionalized 6-azaindole shows good activity and
bioavailability when administered as the TRIS salt of the
phosphate pro-drug.4 While this molecule may appear simple,
its reactivity is dominated by electronic perturbations resulting
from the interplay between the functional groups present on
the aromatic nucleus. The disposition of functionality around
the ring system has an impact on reactivity at each stage of the
existing synthesis (Figure 1); for example, the C4-methoxy
substituent significantly influences the reactivity at C7,
complicating the installation of the triazole through standard
SNAr chemistry. The oxalate side-chain at C3 modifies the
reactivity of the indole nitrogen (N1), which then affects N1/
N6 regioselectivity during alkylation. Additional challenges
result from the triazole itself, which is mildly acidic as a free
compound (i.e., is a good leaving group), but once alkylated at
nitrogen can be deprotonated at the ring carbon under basic
conditions (deprotonation occurring at the methine). Fur-
thermore, the three ring-nitrogen atoms of the triazole have
competitive nucleophilic reactivity, leading to regioselectivity
challenges during installation of this important fragment of the
molecule. The existing synthesis (Figure 1), used to support
clinical development, had several challenging steps, contained
several safety liabilities, was long, and considered unsuitable for
commercialization,5 despite several improvements made during
development.6 This original synthesis started from 2-amino
picoline 4 and required a significant number of challenging
INTRODUCTION
■
For many patients infected with the HIV virus, treatment with
the current antiretroviral therapies results in a high degree of
viral suppression. However, there remains a need to develop
new approaches to reduce viral replication with improved
activity, mutant coverage, reduced side-effects, and alternate
modes of action. The HIV-1 attachment inhibitor pro-drug,1
BMS-663068 1,2 is a compound currently in clinical develop-
ment for the treatment of HIV infection (Figure 1).3 This
Figure 1. Key challenges and previous disconnection strategy.5,6
Received: July 16, 2014
© XXXX American Chemical Society
A
dx.doi.org/10.1021/jo5016008 | J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
transformations to reach the C4/7-dimethoxy-6-azaindole
intermediate 3; namely N-acetylation, C5-bromination, C3-
nitration, Sandmeyer methoxylation, condensation of the C4-
methyl group with dimethylformamide dimethyl acetal (DMF·
DMA), Ullmann methoxylation (to install the methoxy group
at C4 of the forming azaindole), reduction of the C3-nitro
group (N1 of the azaindole), and finally, cyclization to form the
6-azaindole core (Figure 1). The installation of the triazole via
this strategy was complicated by the early installation of the C4-
methoxy moiety, which greatly deactivated C7 to SNAr; thus
both the chlorination (neat POCl3/100 °C) and the
subsequent triazole displacement (140 °C in 4-methyl-2-
pentanol) required harsh reaction conditions, resulting in
poor triazole N-selectivity (ca. 3:1:1). While selectivity was in
favor of the desired isomer, the low selectivity resulted in a
correspondingly modest isolated yield of the desired product 2
(ca. 56%). Finally, installation of the phosphate pro-drug was
complicated by several handling and selectivity challenges.
During the initial development of this synthetic sequence
several issues were observed with the pro-drug installation. To
obviate these challenges the indole 2 was first N1-alkylated with
ArSCH2Cl (the lack of acylation at C3 enabling high N1-
selectivity), followed by Friedel−Crafts C3-acylation with
methyl oxalyl chloride and amidation, to install the side-
chain. Treatment with chlorine converted the N1 thio-ether to
the N1 chloromethyl derivative (NCH2Cl), which could then
be treated with di-tert-butyl potassium phosphate to prepare the
pro-drug in good yield. While this approach provided BMS-
663068 1, the lengthy sequence, high number of complex steps
and numerous safety liabilities led us to interrogate a
completely different approach.
functionalizing pyridine analogues (Figure 2). During the
course of this work, we discovered an unusual aromatization
In reviewing alternate strategies for constructing the core of
the molecule it was interesting to note that while 4, 5 and 7-
azaindole derivatives have been known for some time, the 6-
azaindole analogue is a relatively uncommon structure, with
only a limited number of demonstrated synthetic approaches
reported. The Bartoli cyclization7 of nitro pyridines can be used
to prepare the 6-azaindole core, but it is a complex
transformation, often resulting in poor yields.8 3-Amino
pyridine derivatives have also been used as starting materials
through protection, lithiation, and condensation.9 This
approach often intersects with 3-amino 4-picoline, which can
be converted to a 6-azaindole through condensation of the
methyl group with an appropriate formyl derivative (such as
DMF·DMA) and subsequent cyclization.10 More complex
strategies involving Pd-mediated cross-coupling and cyclization
have also been developed.11 While some of the palladium-
mediated approaches appear, at first glance, to be appropriate in
the context of azaindole 1, we had to consider several
complicating factors in designing our approach. BMS-663068
1 was projected to require relatively high doses to reach efficacy
targets in the clinic, thus synthetic efficiency and simplicity
were prerequisites of any viable long-term synthesis. From the
outset we realized that these additional constraints may rule out
many Pd-mediated transformations (largely due to the use of
either difficult to prepare ligand architectures or high catalyst
loadings). While such limitations can be constrictive, this
restriction can often be the catalyst for synthetic innovation.
Thus, our approach to BMS-663068 1 focused on a retro-
synthetic hydrogenation of the 6-azaindole, approaching the
synthesis from a lower oxidation state.12 We believed that this
strategy would enable a condensation-based approach, starting
from a simple pyrrole, obviating the many challenges of
Figure 2. Disconnection proposal for BMS-663068 1.
process, and invented a bromination protocol to functionalize
heterocyclic pyridine N-oxides,13 which enabled the use of a
highly regioselective Cu-mediated Ullmann−Goldberg−Buch-
wald coupling.14
RESULTS AND DISCUSSION
■
Assembly of the 6-Azaindole Core. To assess the
viability of our proposed disconnection we prepared a range
of amino-pyrroles, following a known C3-selective Friedel−
Crafts acylation of N-benzenesulfonyl pyrrole with chloro-
acetyl chloride.15 Simple chloride displacement with a range of
nitrogen sources gave a series of amino ketones 5 (Scheme 1)
ready for further functionalization. Ketone reduction and
optional methanolysis thus provided three viable substrate
types (5, 6 ,and 7) for cyclization. We hoped that this strategy,
through a combination of starting material oxidation state (i.e.,
ketone vs alcohol/ether) and condensing agent oxidation state
(i.e., aldehyde, ester or carbonate), would allow us to target any
desired oxidation state of the resulting azaindole (for example;
4,5-dihydro-; 6,7-dihydro-, or tetrahydro-azaindole) and
potentially reveal additional options for installing both the
methoxy at C4 and the triazole at C7.
We initiated our study using ketone 5, hoping to obtain
cyclization while maintaining the oxidation level at C4, despite
the obvious deactivation of pyrrole nucleophilicity the C3-
carbonyl would induce. Unfortunately, Pictet−Spengler,16
B
dx.doi.org/10.1021/jo5016008 | J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
process (Scheme 3); after the initial Friedel−Crafts, displace-
ment with sodium N-formyl tosylamide 13 gave ketone 12,
Scheme 1. Preparation of Pyrrole-Based Amino-Alcohols
Scheme 3. Optimized Synthesis of 6,7-Dihydroindole 11
Bischler−Napieralski,17 or other common methods failed in the
condensative cyclization. However, it was found that the lower
oxidation state amino-alcohol derivatives (such as 6 and 7)
easily cyclized under standard Pictet−Spengler conditions (i.e.,
condensation with formaldehyde in dilute acid), forming the
tetra-hydro azaindoles 8 and 9 in excellent yield (Scheme 2).
We rationalized that sp3 hybridization at C4 was critical for
successful cyclization (potentially offering both electronic and
conformational advantages), indeed, after initial ketalization of
the C4 ketone 5 with ethylene glycol (to form dioxolane 10),
smooth cyclization was observed under standard Pictet−
Spengler conditions to give 6,7-dihydroazaindole 11; in this
case cyclization was followed by in situ deprotection of the ketal
post cyclization. Interestingly, only a Pictet−Spengler approach
was effective in the cyclization, limiting us to producing the
C6−C7 dihydro systems (by incorporation of a unit of
formaldehyde). All attempts to affect cyclization at a higher
oxidation state via Bischler−Napieralski or other similar
processes (cyclizing amides, amidines, formates, etc.) resulted
in nonproductive reactivity. With an understanding of the scope
of the cyclization strategy, we investigated methods to further
functionalize the tetrahydro derivatives (8 and 9). After
exploring several unsuccessful methods to appropriately
functionalize the tetrahydro-azaindoles, we focused on ketone
11 as the most promising substrate for downstream
manipulation, hoping that the higher level of oxidation would
ease the required aromatization process downstream. The
optimized formation of this compound followed a simple 4 step
with the formyl group ensuring monoselectivity in the chloride
displacement. Ketalization with ethylene glycol occurred with
concurrent N-formyl deprotection, setting the stage for the
Pictet−Spengler cyclization/deketalization, which occurred in
excellent yield under mild conditions.
The tosylate protection of tetrahydroazaindole 11 was
selected as it offered an additional functional handle for
rearomatization, which we proposed could occur via a redox
elimination (Scheme 4). It was envisaged that formation of the
Scheme 4. Initial Aromatization Conditions
corresponding methyl-enol 14 would facilitate thermal
elimination of sulfinic acid, aromatizing the indole 15 (i.e.,
Scheme 2. Initial Interrogation of the Key Condensation To Form 6-Azaindoles
C
dx.doi.org/10.1021/jo5016008 | J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
affecting reduction at sulfur with aromatic oxidation).18 After
some experimentation, enabling conditions were identified,
using 2,2-dimethoxypropane and ZrCl4 as a Lewis-acid (where
the zirconium promotes ketalization, enolization, and aroma-
tization). While this reaction afforded excellent solution yields,
the zirconium salts proved problematic to remove and resulted
in significant loss of our desired product. After additional
exploration, alternate conditions were found, leveraging protic
acids or their anhydrides (such as MSA or Tf2O with trimethyl
orthoformate); however, both conditions were highly variable
with reactions exhibiting great differences in reactivity. Some
experiments would proceed rapidly to completion, while others
would not progress past the intermediate enol ether 14. After a
detailed review of these results, we noted that the reaction
performance correlated closely to the equipment used for the
reaction, with poor performance in equipment which offered
more tightly controlled environments (i.e., better inertion and
removal of trace oxygen). The reaction was therefore
performed under a strictly controlled atmosphere, with
systematic variation of oxygen content ranging from <0.5
oxygen in nitrogen to atmospheric conditions (ca. 21% O2)
(Chart 1). These data showed that the rate of aromatization
Scheme 5. Optimized Conditions for the Preparation of 15
through C−O or O−O bond homolysis), resulting in low
concentrations of cumyl radical and/or peroxy radicals (the
microscopic reverse of cumene hydroperoxide production),19
which then mediate the aromatization process. Interestingly, in
contrast to the acidic thermal conditions, where methyl p-
toluenesulfinate was produced as the stoichiometric byproduct,
these conditions provided methyl p-toluenesulfonate as the
elimination product (Scheme 5). This change appeared
quantitative even though substoichiometric amounts of the
initiators (AIBN or CHP) are used. With high yielding
conditions for the aromatization in place, we turned our
attention to functionalization of C7.
Introduction of the 3-Methyl-1,2,4-Triazole. The
Reissert−Henze reaction is a common process used to
functionalize the C2 position of pyridine, often proceeding
through the corresponding pyridine N-oxide.20 This process has
been extended from the original addition of cyanide to include
a wide range of N-oxide activators and nucleophilic partners.21
In our context, N-oxide 16 was easily prepared using standard
conditions (methylrhenium trioxide (MTO)/H2O2).22 While
activation of the N-oxide 16, under standard Reissert conditions
(with tosic anhydride, in the presence of triazole 17) showed
high conversion and excellent solution yield, regioselectivity
with respect to the triazole proved difficult to control (a 1:1
mixture of N1:N2 (18:19) was generally observed, with
coupling at N4 only a minor product, Table 1). We rationalized
that the poor regioselectivity was induced by the competition of
electronic and steric influences on the triazole. We therefore
designed a survey of indole nitrogen protecting groups and
activating agents, to determine if variations within either the
electrophilic partner or activating agent could improve the
regioselectivity observed in the addition (Table 1). Electron-
poor N-protecting groups (i.e., Boc, benzenesulfonyl, Piv)
showed poor regiochemical control, with little variance from a
1:1 N1:N2 ratio. In this series the activating agent had little
effect, though it was evident that chloride was a competitive
nucleophile to the triazole. More electron-neutral N-substitu-
ents (i.e., Bn, MOM, phosphate) showed only a slight
improvement in regioselectivity, with N-Bn providing a 2.8:1
ratio in favor of the desired isomer 18. While this is only
modest selectivity, isolation of the desired isomer in a pure
form was easily achieved (∼56% overall yield). Despite these
low yields, this approach was competitive with the conditions
used in the previous approach (which used high temperatures
and long reaction times to produce similar results in terms of
yield and selectivity, vide supra).
Chart 1. Aromatization of 11 in the presence of varying
levels of oxygen
was highly sensitive to oxygen content, with faster rates being
observed in the presence of increasing oxygen concentrations.
For example, the aromatization reaction reached 90+%
conversion within 24 h under atmospheric conditions, while
the reaction lingered around 60% conversion after 72 h with an
oxygen content of <0.5%. Based on this information we
hypothesized that the aromatization was a radical-based
process. In order to avoid the complexity of using a controlled
atmosphere in our chemistry (i.e., a set ratio of O2/N2), we
chose to screen a variety of radical initiators to remove the
sensitivity of the aromatization to oxygen. While AIBN, tert-
butyl hydroperoxide and cumene hydroperoxide (CHP) were
found to be highly competent mediators of the aromatization,
we selected CHP as our optimal initiator due to its availability
and ease of use (Scheme 5). However, as CHP rapidly
undergoes the Hock Rearrangement under acidic conditions,19
we found it necessary to add the CHP in two portions to
reliably get full conversion; indeed NMR analysis of the crude
reaction mixture indicated that both phenol and acetone (the
products of the Hock rearrangement of CHP) were formed in
near quantitative amounts (based on the CHP input). Based on
the limited data available, we believe this reaction is initiated
through a minor decomposition pathway of CHP (either
While we had accomplished the desired C7-functionalization,
the outcome was less than satisfactory. During our work on this
transformation, Londregan et al. published their findings using
the coupling agent PyBroP23 as an activator in Reissert−Henze
chemistry.24 This system demonstrated reactivity across a wide
range of oxygen and nitrogen-based nucleophiles and appeared
to offer a new approach to activation, as such, we explored
D
dx.doi.org/10.1021/jo5016008 | J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
Table 1. Reissert Addition of Triazole 17 to N-Oxide 16
a
Provided 56% isolated yield as a single isomer.
Table 2. Bromination screen of N-Oxide 16
a
a
a
conditions
21 (%)
22 (%)
15(%)
note
POBr3
PPh3/Br2
PPh3/NBS
AcBr
trace
11
trace
5
trace
79
decomposition
rt or 50 °C
rt or 50 °C
25
3
69
N/A
40 AP
N/A
N/A
N/A
95
clean reaction
PhSO2Br
PyBroP
46
DCM, rt, 12 h
K3PO4, PhCF3, rt, 6 h
b
82 AP (80%)
10
a
b
Relative area percent determined by HPLC. Solution yield of a gram-scale experiment.
these conditions in our triazole coupling. While triazole regio-
selectivity was again poor, analysis of the reaction mixture
showed some evidence of C7-bromination under the PyBroP-
mediated conditions. Intrigued by the possibilities of a C7-
bromide in the downstream chemistry, we focused on
optimizing this side reaction. Simply removing the triazole 17
and optimizing base, solvent, and temperature resulted in a
good yields (ca. 70−80%) of the C7-bromide 21 (Table 2). To
the best of our knowledge, this reactivity of pyridine-like N-
oxides is unprecedented, indeed C2-bromopyridines are
significantly more challenging to prepare than their chloride
analogues. Typically, pyridine-2-ones react smoothly with
POCl3 to prepare 2-Cl pyridines, whereas significantly more
forcing conditions are required to effect bromination (i.e., using
POBr3 or a combination of P2O5/TBABr at high temper-
ature).25 Indeed, when POBr3 was used in place of POCl3 in
our original synthesis (Figure 1) it proved challenging to
produce even modest amounts of the corresponding bromide
21.26 To compare PyBroP to other standard bromination
systems, a wide range of alternate conditions were surveyed
using N-oxide 16, such as NBS/PPh3,27 POBr3,28 PPh3/Br2
among others (Table 2), all these approaches led mainly to des-
oxygenation (i.e., recovery of 15), while PyBroP was the
exception and appeared unique in its ability to smoothly
brominate this azaindole N-oxide 16. Further interrogation of
these enabling conditions resulted in the development of a
Ts2O/TBABr(Cl) based process for the bromination (or
chlorination) of heterocyclic N-oxides, chemistry which was
recently reported.29
The ability to prepare the C7-bromide 21 suggested exciting
possibilities to improve the triazole installation, not only
opening up the potential to install the triazole later in the
sequence (avoiding impurities formed through the acidic C−H
on the triazole), but also enabling the use of a ligated Cu-
mediated process (such as an Ullmann−Goldberg−Buchwald
reaction), which we hoped would improve the regiochemical
E
dx.doi.org/10.1021/jo5016008 | J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
outcome of the addition.30 With this goal in mind, we prepared
the amide 23, through N-benzenesulfonyl deprotection,
Friedel−Crafts acylation/hydrolysis and amidation with piper-
azine 2431all of which proceeded in high yield (Scheme 6).
up to a 10:1 N1/N2 ratio was obtained in the initial test
reactions with trans-DMCHDA. We presumed that this was
due to a much tighter transition state induced by the catalytic
process (Figure 3). The coupling generally performed well in
most organic solvent classes (such as hindered alcohols,
acetates, ethers and nitriles), though the nature of the cationic
component of the base proved critical. Potassium was uniquely
effective (Li or Na cationic counterions provided little to no
observable product), interestingly the anionic component of
the base was much less significant. Utilization of either tBuOK
or K2CO3 provided similar results, which were comparable to
the use of preformed potassium-triazolate. Intriguingly, it was
discovered that the presence of water significantly enhanced
both reactivity and regioselectivity. The reason for this effect is
unknown, though the impact of water on the rate of Ullmann
couplings has been documented elsewhere.36 Consequently,
replacing tBuOK with aqueous KOH, provided the required
counterion, base and water, which, along with further
optimization, delivered a reproducible and highly regioselective
process (>20:1 N1/N2 isomeric ratio and >80% in-process
yield). To isolate the desired product, LiBr was added (after a
simple aqueous workup) effecting in situ salt metathesis and
crystallization of a highly crystalline Li-salt of 25, isolated as a
KBr cocrystal (Scheme 7).
Scheme 6. Preparation of the Amide 23
Pro-Drug Formation. The conversion of Li-salt 25 to
BMS-663068 1 involves only pro-drug installation, however,
the installation of the phosphate was not without complexity.
Previously, thiomethylation, chlorination, and phosphate
displacement had been used to prepare 26, this lengthy
sequence was conducted to obviate issues of reactivity and
reagent availability, involved in the direct alkylation.5,6 In
reconsidering the installation of the last section of the molecule,
we wished to return to a direct alkylation-based approach,
utilizing a chloromethyl-phosphate derivative such as phosphate
ester 27. In order to accomplish this strategic revision,
concurrent development of new methods to prepare both the
chloromethyl-phosphate electrophile 2737 and its starting
material, chloromethyl chlorosulfate,38 were developed. The
With the key amide 23 in hand, we studied the late-stage
installation of the triazole 17.32 Extensive screening was carried
out for the Ullmann−Goldberg−Buchwald coupling, inves-
tigating the choice of ligand, base, solvent, and metal precursor.
Our initial screens mainly focused on the use of diamine,33
diketone,34 and phenanthroline35 ligands, among others; some
of our initial results are shown (Table 3). It was found that in
the presence of copper iodide (a preferred catalyst precursor),
the 1,2-diamine ligands were unique in their reactivity and were
significantly better than other ligand classes; trans-DMCHDA
(trans-dimethyl cyclohexyl diamine) proved especially effective.
We were pleased to find in our initial leads, especially with
diamino-ligands, that greatly improved regioselectivities were
generally observed vs the thermal SNAr approach (vide supra);
Table 3. Optimization of the Ullmann−Goldberg−Buchwald reaction
a
ligand
ligand equiv
CuI equiv
base
solvent
conversion (yield)
N1/N2
b
phenanthrene
2,2-bipyridine
8-quinolinol
5
5
0.5
0.5
0.5
0.5
0.2
0.2
0.5
0.2
0.2
0.2
KOtBu
KOtBu
4-methyl-2-pentanol.
4-methyl-2-pentanol
4-methyl-2-pentanol
4-methyl-2-pentanol
4-methyl-2-pentanol
4-methyl-2-pentanol
4-methyl-2-pentanol
THF or EtOAc or MeCN
MeCN
0
0
N/A
N/A
N/A
N/A
4:1
5
KOtBu
0
EDTA or MIDA
DMEDA
5
KOtBu
0
1.5
1.5
5
KOtBu
37%
90%
<5%
85−93%
>98%
c
DMCHDA
KOtBu
10:1
N/A
10:1
>20:1
22:1
d
DMCHDA
DMCHDA
DMCHDA
DMCHDA
LiOH or Na2CO3
1.5
1.5
1.5
KOtBu or K2CO3
KOtBu; 10 equiv H2O
aq KOH
e
MeCN
>98% (85%)
a
b
Conversion was based on reactions run at 80 °C over 20 h. Various phen-type ligands were tested, all of which failed to promote the coupling.
c
d
e
DMCHDA; N,N-dimethy-1,2-cyclohexanediamine. Organic bases such as pyridine or Hunig’s base also failed to promote the coupling. Solution
yield obtained from a 2 g-scale experiment.
F
dx.doi.org/10.1021/jo5016008 | J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
Figure 3. Proposed rationale for regioselectivity in the Ullmann−Goldberg−Buchwald coupling of bromide 23 to triazole 17.
solvent was critical, and alternative potassium additives (i.e.,
dibasic phosphate, bicarbonate, acetate, benzoate, formate, etc.)
led to poor reactivity. Our analysis of this trend suggested that
potassium additives with significantly less soluble lithium
analogues were critical for driving the salt metathesis process
in the desired direction. With tert-butyl phosphate 26 in-hand,
deprotection was accomplished under previously reported
conditions, to give the final compound in excellent purity.2b
Overall, this new synthesis provided BMS-663068 1 in 11
linear steps and 5.9% overall yield, setting the stage for further
optimization of this strategy for the potential manufacture of
BMS-663068 on commercial scales. The transformations
described herein have been subsequently optimized and
improved, to provide the manufacturing route to BMS-
663068 1, with significant additional yield improvements.
This work will be described in due course.
Scheme 7. End-Game Steps To Prepare BMS-663068, 1
CONCLUSION
■
An efficient synthesis of a complex 6-azaindole, BMS-663068,
has been accomplished in an 11-step sequence from simple
precursors such as pyrrole and 3-methyl-1H-1,2,4-triazole. The
route features the rapid assembly of a 6-azaindole, leveraging a
low oxidation state approach via a condensative cyclization,
followed by a radical-mediated redox-aromatization to reveal
the 6-azaindole. Other highlights include an unusual α-
bromination of a heterocyclic N-oxide, which provides a critical
functional handle for a highly regioselective Ullmann−Gold-
berg−Buchwald coupling to install the key triazole moiety.
protio-form of 25 had been utilized in an alkylation with
phosphate 27 previously, as part of an earlier synthesis of this
BMS-663068 1.2b In that case, alkylation occurred with modest
yield and N1 to N6-selectivity. During optimization of this
reaction we screened a range of alternate salt forms of 25; for
example, tetra-butyl ammonium and potassium salts. While,
none of these alternate forms provided crystallinity in the
azaindole 25, all showed reactivity with the chloromethyl-
phosphate 27. However, surprisingly the lithium salt, while
providing an isolatable crystalline form of 25, was much less
reactive in the alkylation; presumably due to the lithium
counterion preferring to coordinate (and therefore localize
charge) in the conjugated α-keto amide, as opposed to direct
nitrogen coordination and charge localization. Thus, in situ salt
metathesis from lithium to potassium was required for
productive alkylation. After extensive screening, we discovered
that the use of wet acetonitrile (KF = 1%) as solvent, in the
presence of 0.5 equiv of tetraethylammonium iodide (to
facilitate an in situ halide exchange) and either tribasic
potassium phosphate or potassium carbonate as additives, led
to efficient conversion of 25 to 26. Interestingly, the use of wet
EXPERIMENTAL SECTION
■
2-Chloro-1-(1-(phenylsulfonyl)-1H-pyrrol-3-yl)ethanone.15
Aluminum chloride (21.2 g, 1.1 equiv) and dichloromethane (300 mL)
were added to a 500 mL round-bottom flask under an atmosphere of
nitrogen. Chloro acetyl chloride (12.6 mL, 1.1 equiv) was then added
in one portion, at which point the reaction turned from white slurry
into a light-yellow homogeneous solution. After stirring at ambient
temperature for 10 min, N-benzenesulfonylpyrrole (30 g, limiting
reagent) was added in 2−3 portions over 30 min. The resulting brown
solution was stirred overnight. After cooling to 0 °C, aq sodium
bicarbonate (10 wt %, 250 mL) was added slowly while maintaining
the internal temperature below 20 °C. The biphasic mixture was
stirred for 30 min, and the layers were separated. The organic layer was
dried over sodium sulfate, filtered, and concentrated under vacuum to
afford a dark-brown oil. Purification by silica-gel chromatography
(eluting with 10−30% EtOAc in hexanes) afforded the 3-acyl-pyrrole
G
dx.doi.org/10.1021/jo5016008 | J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
as a white crystalline solid (29 g, 70% yield). Rf = 0.6 in 25% EtOAc in
hexanes; mp = 105−106 °C; 1H NMR (500 MHz, CDCl3) δ: 7.91 (d,
J = 7.6 Hz, 2 H), 7.86 (s, br, 1 H), 7.66 (t, J = 7.32 Hz, 1 H), 7.55 (dd,
J = 7.3, 7.6 Hz, 2 H), 7.17 (s, br, 1 H), 6.71 (s, br, 1 H), 4.45 (s, 2 H);
13C NMR (125 MHz, CDCl3) δ: 186.1, 137.6, 134.7, 129.7 (2), 127.1
(2), 125.6, 124.9, 121.8, 112.4, 46.0; HRMS [M − H; ESI-
ORBITRAP] calcd for C12H9ClNO3S: 281.9986; found: 281.9979.
Sodium N-tosylformamide, 13.39 A mixture of sodium
methoxide in methanol (25% w/w, 300 mL, 1.5 equiv) and toluene
sulfonamide (150 g, limiting reagent) was heated to 30 °C for 30 min,
at which point a thick slurry (sodium tosylamide) formed. Methanol
(200 mL) was added, followed by portion-wise addition of ethyl
formate (282 mL, 4 equiv). The resulting mixture was stirred
vigorously at 30 °C while the reaction conversion was monitored by
HPLC. Upon completion (typically 14 h), the reaction mixture was
cooled to 20 °C, filtered, washed with methanol (450 mL) and dried in
vacuum at 50 °C to afford sodium N-tosylformamide 13 as a white
solid (183.3 g, 83%). HPLC monitored at 256 nm using Water
Symmetry Shield RP-8, 3.5 mm 4.6 × 150 mm column, gradient
method: 10% B (0 min) to 100% B (12 min), then 100% B (15 min).
Mobile phase A: water (80%), methanol (20%) with 0.05%
trifluoroacetic acid, Mobile phase B: acetonitrile (80%), methanol
(20%) with 0.05% trifluoroacetic acid. HPLC retention time: 8.35 min.
reaction mixture was cooled to 20 °C and aged for 1 h to afforded a
thick slurry. The product was filtered, washed with methanol (80 mL),
and dried under vacuum at 60 °C to afford dioxolane 10 (9.56 g, 97.2
wt %, 90% yield). HPLC at 232 nm using Ascentis Express C18 2.7
μm 4.6 × 50 mm column, gradient method 40% B (0−4.5 min) to
100% B (6 min). Mobile phase A: 0.01% NH4OAc in H2O: MeOH
(80:20), Mobile phase B: 0.01% NH4OAc in H2O: MeOH: CH3CN
(5:20:75). Retention time: 3.0 min. Melting point (mp) = 188−199
°C; 1H NMR (500 MHz, CDCl3) δ: 7.9 (d, J = 7.5 Hz, 2 H), 7.7 (d, J
= 7.8 Hz, 2 H), 7.6 (t, J = 7.5 Hz, 1 H), 7.5 (t, J = 7.5 Hz, 2 H), 7.2 (s,
br, 2 H), 7.1 (s, 1 H), 7.0 (s, 1 H), 6.2 (s, 1 H), 4.7 (s, br, 1 H), 3.9 (s,
br, 2 H), 3.8 (s, br, 2 H), 3.2 (d, J = 6.1 Hz, 2 H), 2.4 (s, 3 H). 13C
NMR (125 MHz, CDCl3) δ: 143.30, 138.71, 137.31, 134.11 (2),
129.58 (2), 129.53 (2), 128.48, 127.00 (2), 126.91 (2), 118.67, 111.70,
105.47, 65.35 (2), 49.35, 21.53. CHNC21H22N2O6S2: calcd: 54.53% C,
4.79% H, 6.06% N; found: 54.53% C, 4.88% H, 5.93% N. HRMS [M +
H; ESI-ORBITRAP] calcd for C21H23N2O6S2: 463.0992; found:
463.0989.
1-(phenylsulfonyl)-6-tosyl-6,7-dihydro-1H-pyrrolo[2,3-c]-
pyridin-4(5H)-one, 11. To a slurry of dioxolane 10 (10.00 g, limiting
reagent) and paraformaldehyde (1.6 g) in dichloromethane (100 mL)
at 20 °C was added trifluoroacetic acid (4.90 mL, 3 equiv). Once the
reaction was complete (typically 2 h), the reaction was quenched with
aqueous solution of dibasic potassium phosphate (10 wt %, 100 mL)
and further stirred for 20 min. The aqueous layer was discarded, and
the organic layer was polished filtered to remove remaining
paraformaldehyde. The organic layer was then concentrated to ca.
50 mL, at which point isopropyl acetate (24 mL) and methyl tert-
butylether (30 mL) were added. The solution was distilled with
constant methyl tert-butylether addition to remove dichloromethane.
The resulting slurry was then cooled to 0 °C over a course of 2 h. The
product was filtered, washed with methyl tert-butylether (2 × 50 mL),
and dried at 50 °C in a vacuum oven to give 1-(phenylsulfonyl)-6-
tosyl-6,7-dihydro-1H-pyrrolo[2,3-c]pyridin-4(5H)-one 11 as a white
solid (7.56 g, 81%). HPLC at 230 nm using Phenomenex Kinetex, C18
100A, 4.6 × 150 mm column with gradient method 15% B (0 min) to
60% B (20 min) to 100% B (25 min) then 100% B (30 min). Mobile
phase A: water (80%), methanol (20%) with 0.01 M NH4OAc. Mobile
phase B: acetonitrile (75%), methanol (15%), water (5%) with 0.01 M
NH4OAc, retention time: 20.33 min; Melting point (mp) = 168 °C;
1H NMR (500 MHz, CDCl3) δ: 8.01 (dd, J = 8.4, 3.0 Hz, 2H), 7.78
1
Melting point (mp) = 261 °C; H NMR (400 MHz, DMSO-d6) δ:
8.75 (br s, 1H), 7.52 (d, br, J = 8.3 Hz, 2H) 7.22 (d, br, J = 7.8 Hz,
2H) 2.32 (s, 3 H); 13C NMR (100 MHz, DMSO-d6) δ: 170.0, 144.2,
139.6, 128.7 (2), 125.4 (2), 20.8; HRMS [M + H; ESI-ORBITRAP]
calcd for C8H10NO3S: 200.0381; found: 200.0379.
N-(2-oxo-2-(1-(phenylsulfonyl)-1H-pyrrol-3-yl)ethyl)-N-to-
sylformamide, 12. To a slurry of sodium N-tosylformamide 13 (9.36
g, 1.2 equiv) in THF (100 mL) was added tetra-N-butylammonium
iodide (1.3 g, 0.1 equiv) and the chloro-ketone (10.0 g, limiting
reagent). The reaction was heated to 60 °C for 3 h, at which point
reaction completion was determined via HPLC. The solution was then
cooled to 20 °C and washed with hydrochloric acid (0.5 M, 50 mL)
and then sodium chloride solution (25% w/w, 50 mL). The collected
organic layer was concentrated to ca. 60 mL, at which point isopropyl
alcohol was added (57 mL). After stirring at ambient temperature for
20 min, additional isopropyl alcohol was added (57 mL), and the
resulting slurry was cooled to 10 °C to enable crystallization. After
stirred at 10 °C for another 20 h, the product was filtered, washed with
isopropyl alcohol/THF (50/50, 3 × 30 mL), and dried under vacuum
at 50 °C to afford N-(2-oxo-2-(1-(phenylsulfonyl)-1H-pyrrol-3-
yl)ethyl)-N-tosylformamide 12 as a white solid (12.95 g, 82%).
HPLC at 256 nm using Water Symmetry Shield RP-8, 3.5 m 4.6 × 150
mm column, gradient method 10% B (0 min) to 100% B (12 min),
then 100% B (15 min), Mobile phase A: water (80%), methanol
(20%) with 0.05% trifluoroacetic acid, Mobile phase B: acetonitrile
(80%), methanol (20%) with 0.05% trifluoroacetic acid, HPLC
(dt, J = 5.3, 1.2 Hz, 1H), 7.68 (t, J = 7.8 Hz, 2H), 7.30 (d, J = 8.4 Hz,
2H), 7.14 (d, J = 3.5 Hz, 1H), 7.07 (d, J = 8.1 Hz, 2H), 6.39 (d, J = 3.3
Hz, 1H), 4.85 (s, 2H), 3.95 (s, 2H) 2.33 (s, 3H); 13C NMR (125
MHz, CDCl3) δ: 187.0, 144.1, 137.9, 137.5, 135.3, 134.0, 130.2 (2),
129.8 (2), 127.4 (2), 127.1 (2), 122.7, 122.5, 108.3, 53.0, 43.2, 21.5;
HRMS [M + H; ESI-ORBITRAP] calcd for C20H19N2O5S2: 431.0730;
found: 431.0722.
4-Methoxy-1-(phenylsulfonyl)-1H-pyrrolo[2,3-c]pyridine,
15. Ketone 11 (10 g, limiting reagent), methanol (120 mL), and
methanesulfonic acid (2.9 g, 1.5 equiv) were sequentially charged to a
250 mL round-bottom flask. The mixture was then heated to 30 °C.
Trimethylorthoformate (10.5 g, 5 equiv) was added, followed by
cumene hydroperoxide (1.5 g, 0.5 equiv in two equal portions, 1 h
apart). The reaction progress was monitored by HPLC. Upon
completion (typically 2 h), the crude mixture was cooled to 20 °C.
Triethylamine (3.4 g, 1.7 equiv) was slowly added over a course of 30
min. Aqueous sodium thiosulfate (1 wt %, 100 mL) was slowly added
over the course of 2 h, resulting in the formation of a thick slurry. The
product was filtered, washed with water/methanol (45/55; 50 mL),
and then dried under vacuum at 50 °C. Azaindole 15 was isolated as a
white crystalline solid (4.9 g, 85% yield). HPLC at 230 nm using
Phenomenex Kinetex, C18 100A, 4.6 × 150 mm column gradient
method 15% B (0 min) to 60% B (20 min) to 75%B (25 min). Mobile
phase A: water (80%), methanol (20%) with 0.01 M NH4OAc. Mobile
phase B: acetonitrile (75%), methanol (15%), water (5%) with 0.01 M
NH4OAc, retention time: 15.19 min; Melting point (mp) = 152 °C;
1H NMR (500 MHz, CDCl3) δ: 9.01 (s, 1H), 8.03 (s, 1H), 7.92 (m,
1
retention time: 11.90 min. Melting point (mp) = 142 °C; H NMR
(500 MHz, CDCl3) δ: 9.14 (br s, 1H), 7.92 (d, br, J = 8.7 Hz, 2H)
7.76−7.75 (m, 3H) 7.69 (t, J = 6.6 Hz, 1H) 7.58 (t, J = 7.6 Hz, 2H)
7.35 (d, br, J = 8.5 Hz, 2H) 7.16 (dd, J = 3.2,2.2 Hz, 1H) 6.62 (dd, J =
3.5 1.6 Hz, 1H) 4.75 (s, 2H) 2.46 (s, 3H); 13C NMR (126 MHz,
CDCl3) δ: 184.4, 160.7, 145.7, 137.8, 134.8, 134.7, 130.1 (2), 129.8
(2), 127.7 (2), 127.3 (2), 125.5, 124.3, 121.8, 112.1, 48.2, 21.7; HRMS
[M + H; ESI-ORBITRAP] calcd for C20H 19N2O6S2: 447.0679; found:
447.0669.
4-Methyl-N-((2-(1-(phenylsulfonyl)-1H-pyrrol-3-yl)-1,3-diox-
olan-2-yl)methyl)benzenesulfonamide, 10. To a mixture of of N-
(2-oxo-2-(1-(phenylsulfonyl)-1H-pyrrol-3-yl)ethyl)-N-tosylformamide
12 (10.1 g, 22.4 mmol) and methanol (100 mL) was added a
methanolic sulfuric acid solution (4.5 mL, 2.23 mmol). [Note: The
H2SO4-MeOH was prepared by dissolving H2SO4 (96 wt %, 0.228 g,
2.23 mmol) in MeOH (3.44 g).] The reaction was warmed to 60 °C
and stirred for 3h. Ethylene glycol (13.9 g, 10 equiv) was then added,
followed by trimethyl orthoformate (7.26 g, 3 equiv). As the reaction
progressed, 4-methyl-N-((2-(1-(phenylsulfonyl)-1H-pyrrol-3-yl)-1,3-
dioxolan-2-yl)methyl)benzenesulfonamide 10 crystallized from sol-
ution. Once complete reaction was observed (typically 1 h), the
2H), 7.68 (d, J = 3.9 Hz, 1H), 7.60 (dt, J = 7.8, 1.0 Hz, 1H), 7.49 (dt, J
= 8.3, 1.9 Hz, 2H), 6.84 (d, J = 3.0 Hz, 1H), 4.00 (s, 3H); 13C NMR
H
dx.doi.org/10.1021/jo5016008 | J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
(125 MHz, CDCl3) δ: 149.6, 137.5, 134.5, 132.3, 129.6 (2), 129.1,
128.1, 127.6, 126.9 (2), 122.8, 105.6, 56.3; HRMS [M + H; ESI-
ORBITRAP] calcd for C14H13N2O3S: 289.0641; found: 289.0639.
7-Bromo-4-methoxy-1H-pyrrolo[2,3-c]pyridine hydrochlor-
ide salt monohydrate, 21. N-Oxidation. To a solution of azaindole
15 (10 g, limiting reagent) in dichloromethane (200 mL) at 25 °C was
added hydrogen peroxide (6.1 mL, 35 wt %, 2 equiv) followed by
MTO (17 mg, 0.2 mol %). The mixture was stirred at 25 °C until
reaction reached completion by HPLC monitoring (typically 24 h). If
needed, additional 0.2 mol % MTO was added to drive reaction to
completion. The mixture was cooled to 0 °C and quenched with aq
sodium sulfite solution (20 wt %, 50 mL). The layers were separated,
and the product-rich lower DCM layer was washed with water (50
mL). The collected DCM solution was diluted with trifluorotoluene
(100 mL). The resulting thin slurry was distilled with constant
trifluorotoluene addition until residual DCM was <1 vol % by GC
analysis. Adjust the overall volume of the reaction crude to ca. 150 mL,
and submit the thick slurry to the subsequent bromination/hydrolysis
step without further purification.
137.6, 133.0, 123.6, 121.6, 116.8, 114.0, 56.2. IR (KBr, cm−1) 3297,
3159, 1688, 1557, 1419, 1166, 1125, 978, 889, 756, 702, 636; HRMS
[M + H; ESI-ORBITRAP] calcd for C10H8BrN2O4: 298.9667; found:
298.9678.
To the HBr salt of the oxalic acid derivative (5.03 g, limiting
reagent) and dimethylformamide (68 mL) was stirred vigorously at
ambient temperate until a homogeneous solution formed. The
solution was cooled to 10 °C, at which point 1,1′-carbonyl diimidazole
(3.46 g, 1.7 equiv) was added. After stirring at 10 °C for 3 h,
benzoylpiperzine hydrochloric salt (3.70 g, 1.3 equiv) was added, and
the mixture was warmed to 25 °C and stirred at 25 °C until reaction
reached completion by HPLC monitoring (typically 2 h). The reaction
was quenched with water (10 mL) and warmed up to 40 °C.
Additional water (70 mL) was slowly added to enable the
crystallization of amide 23. After cooling to 25 °C, the slurry was
held at 25 °C for 10 h, then filtered, and washed with
dimethylformamide/water (1:1 v/v, 15 mL). The solid was collected,
reslurried with water/tetrahydrofuran (10:1 v/v, 25 mL), rinsed with
tetrahydrofuran (15 mL, 3 mL/g), and dried at 60 °C in a vacuum to
afford 1-(4-benzoylpiperazin-1-yl)-2-(7-bromo-4-methoxy-1H-pyrrolo-
[2,3-c]pyridin-3-yl)ethane-1,2-dione 23 as a tan solid (5.49 g, 92 wt %,
85.2% yield). HPLC at 230 nm using Ascentis Express C18, 2.7 μm,
4.6 × 150 mm column following gradient method 0% B (0 min) to
30% B (5 min) to 45%B (20 min) to 100%B (25 min) then 100% B
(30 min). Mobile phase A: water (80%), methanol (20%) with 0.01 M
NH4OAc. Mobile phase B: acetonitrile (75%), methanol (15%), water
(5%) with 0.01 M NH4OAc, retention time: 10.55 min. Melting point
(mp) = 252.5 °C: 1H NMR (500 MHz, CDCl3) δ: 8.31 (s, 1H), 7.84
(s, 1H), 7.45 (s, br, 5H), 3.94 (s, 3H), 3.66 (s, br, 4H), 3.40 (s, br,
4H); 13C NMR (1250 MHz, CDCl3) δ: 185.4, 169.3, 166.2, 149.9,
138.3, 135.5, 133.2, 129.7, 128.4 (2), 127.1 (2), 123.8, 121.1, 116.9,
114.9, 56.8, 45.1 (2, br), 40.5 (2, br); HRMS [M + H; ESI-
ORBITRAP] calcd for C21H20BrN4O4: 471.0662; found: 471.0654.
Lithium 3-(2-(4-benzoylpiperazin-1-yl)-2-oxoacetyl)-4-me-
thoxy-7-(3-methyl-1H-1,2,4-triazol-1-yl)pyrrolo[2,3-c]pyridine-
1-ide, potassium bromide, 25. To a mixture of amide 23 (12 g,
22.656 mmol) and acetonitrile (96 mL) at 20 °C was added 3-methyl-
1H-1,2,4-triazole 17 (2.83 g, 33.98 mmol), followed by aqueous
potassium hydroxide (45 wt %, 5.93 g, 47.58 mmol), water (4.08 g,
226.56 mmol), and N,N′-dimethyl-1,2-cyclohexanediamine (4.92 g,
33.98 mmol). The overall mixture was stirred at 20 °C for 1 h before
adding copper(I) iodide (0.863 g, 4.53 mmol). The reaction mixture
was heated to 75 °C and stirred until reaction reached completion by
HPLC monitoring (typically 18 h). The mixture was cooled to 45 °C
and diluted with acetonitrile (264 mL). At 45 °C, the crude slurry was
washed with a solution of EDTA (6.70 g, 22.66 mmol) in 45 wt %
aqueous potassium hydroxide (2×, 60 mL), followed by 45 wt %
aqueous potassium hydroxide (45 mL). The organic crude was
distilled at constant volume until KF reached <1.3 wt %. The resulting
solution was cooled to 30 °C and filtered to remove inorganic
particulates. The liquors were then further concentrated to ca. 240 mL.
Water was added to the concentrated mixture until a KF of 4.5 wt %
was achieved. Lithium bromide (2.102 g, 24.22 mmol) was added, at
which point the lithium salt of 25 readily crystallized out the solution.
After aging overnight, the solids were collected by filtration, washed
sequentially with acetonitrile/water (98/2 v/v, 65 mL) and then
acetonitrile (44 mL), and dried at 45 °C under vacuum. Azaindole 25
was obtained as a pale yellow solid (8.94 g, 14.94 mmol, 66% yield).
1H NMR (400 MHz, DMSO-d6) δ: 9.51 (s, 1 H), 8.07 (s, 1 H), 7.56
Bromination. To a slurry of N-oxide 16 in trifluorotoluene (ca. 150
mL) at 20 °C was added K3PO4 (36.5 g, 5 equiv) and PyBroP (19.3 g,
1.1 equiv). The resulting mixture was stirred at 20 °C for 3 h, then
warmed to 50 °C, and stirred at that temperature until bromination
reached completion by HPLC analysis (typically 2 h).
Hydrolysis. IPA (80 mL) was then added, followed by aqueous
sodium hydroxide (2 N, 80 mL). The mixture was warmed to 80 °C
and stirred at that temperature. Upon completion of hydrolysis by
HPLC analysis (typically 12 h), the reaction was cooled to 20 °C. The
upper organic phase was separated and distilled under reduced
pressure (150−200 Torr, 50 °C) to ∼100 mL. The resulting slurry was
filtered to remove residual inorganics. The temperature was adjusted
to 50 °C, at which point HCl/IPA (5 N, 10 mL, 1.5 equiv) was added.
The resulting slurry was stirred at 50 °C for 30 min, then cooled to 5
°C over 2 h, and maintained at 5 °C overnight. The product was
collected by filtration and dried at 50 °C for 12 h to afford bromide 21
as its HCl salt monohydrate (7.0 g, 70% yield). HPLC at 256 nm using
YMC Pro C18 3 μm 4.6 × 150 mm column with a gradient method:
20% B (0−1 min) to 55% B (10 min), then 100% (16.5 min). Mobile
phase A: water with 0.05% trifluoroacetic acid. Mobile phase B:
acetonitrile with 0.05% trifluoroacetic acid, retention time: 5.31 min.
Decomposed at 160 °C; 1H NMR (500 MHz, DMSO-d6) δ: 12.80 (s,
1 H), 7.84 (s, br, 1 H), 7.68 (s, 1 H), 6.99 (s, br, 4 H), 6.73 (s, br, 1
H), 3.97 (s, 3 H); 13C NMR (125 MHz, DMSO-d6) δ: 149.8, 133.7,
131.8, 126.8, 115.8, 114.0, 101.0, 56.8; HRMS [M + H; ESI-
ORBITRAP] calcd for C8H8BrN2O (as free base): 226.9820; found:
226.9813.
1-(4-Benzoylpiperazin-1-yl)-2-(7-bromo-4-methoxy-1H-
pyrrolo[2,3-c]pyridin-3-yl)ethane-1,2-dione, 23. To a slurry of
AlCl3 (3.75 g, 28.1 mmol, 4.3 equiv) in DCM (14 mL) at 0 °C was
added bromide 21 (2.00 g, 92 wt %, 6.50 mmol, 1.0 equiv), followed
by a mixture of DCM (2.6 mL) and nitromethane (3.2 mL). After 20
min, methyl oxalyl chloride (1.60 g, 13.0 mmol, 2.0 equiv) was added,
and the resulting reaction mixture was further stirred at 0 °C until
reaction was complete by HPLC analysis (typically 5 h). The reaction
crude was then slowly added into a biphasic mixture of THF (30 mL)
and aqueous solution of Na2SO4 (1.0 M, 30 mL) which was precooled
with an ice/water bath. The organic phase was collected. After washing
with water (30 mL), the organic crude was distilled with constant
water addition until residual nitromethane was <7% (mol/mol to the
ester intermediate) by NMR analysis. To the resulting slurry of the
methyl ester was added aqueous NaOH (10.0 N, 2.3 mL, 3.5 equiv).
After the complete saponification, the resulting sodium salt was
acidified by a slow addition of aqueous HBr solution (48%, 3.7 mL) at
45 °C, leading to the crystallization of the product as its HBr salt. The
resulting slurry was cooled to 21 °C over 3 h. The product was filtered,
washed with aqueous HBr solution (0.5 M, 10 mL), and dried under
vacuum at 50 °C to provide the oxalate as its HBr salt (1.97 g, 79%
yield). Melting point (mp) = 107 °C; 1H NMR (400 MHz, DMSO-d6)
δ: 13.04 (s, 1 H), 11.34 (s, br, 2 H), 8.34 (s, 1 H), 7.82 (s, 1 H), 3.92
(s, 3 H); 13C NMR (100 MHz, DMSO-d6) δ: 182.8, 166.2, 149.9,
(s, 1 H), 7.44 (br s, 5 H), 3.87 (s, 3 H), 3.62 (m, 4 H), 3.38 (m, 4 H),
2.38 (s, 3 H); 13C NMR (100 MHz, DMSO-d6) δ: 169.29, 168.68,
159.46, 149.46, 144.76, 138.02, 135.61, 133.62, 129.64, 128.42, 127.06,
125.66, 119.15, 113.42, 56.75, 13.79; HRMS [M + H; ESI-
ORBITRAP] calcd for C24H23N7O4: 474.1884; found: 474.1876. Li:
1.49 wt %; calcd for C24H22BrKLiN7O4: 1.16 wt %; K: 6.82 wt %; calcd
for C24H22BrKLiN7O4: 6.53 wt %.
(3-(2-(4-benzoylpiperazin-1-yl)-2-oxoacetyl)-4-methoxy-7-
(3-methyl-1H-1,2,4-triazol-1-yl)-1H-pyrrolo[2,3-c]pyridin-1-yl)-
methyl di-tert-butyl phosphate, 26. To a 125 mL reactor
equipped with an overhead stirbar was charged azaindole 25 (12.5 g
I
dx.doi.org/10.1021/jo5016008 | J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
[Limiting Reagent], 20.34 mmol), potassium phosphate tribasic
(comilled, 0.045 in. mesh) (4.32 g, 20.35 mmol, 1.0 equiv), and
tetraethylammonium iodide (2.61 g, 10.15 mmol, 0.50 equiv).
Sequentially acetonitrile (KF = 0.85%, 63.5 mL) and a solution of
phosphate 27 in DCM (1.48 M, 17.9 mL, 26.52 mmol, 1.3 equiv) were
added, and the resulting mixture was warmed to 40−45 °C. After 21 h,
the reaction mixture was concentrated to ∼30 mL. Dichloromethane
(50 mL) was added, followed by phosphoric acid (0.2 M in water, 60
mL). The lower DCM layer was collected and washed with phosphoric
acid (0.2 M in water, 70 mL) and then water (70 mL). IPAc (40 mL)
was added to the DCM solution, and the resulting solution distilled at
35 °C to remove DCM, while additional IPAc (100 mL) was added to
effect a constant volume distillation. Once the distillation was
complete, the mixture was cooled to 20 °C over 1.5 h and aged
overnight. The product was filtered, washed sequentially with IPAc (50
mL) and MTBE (50 mL), and dried in vacuum oven at 50 °C to give
phosphate 26 as a white solid (8.59 g, 61% yield). Melting point (mp)
= 198−202 °C; 1H NMR (400 MHz, CDCl3) δ: 8.51 (s, 1 H), 8.17 (s,
1 H), 7.88 (s, 1 H), 7.39 (br s, 5 H), 5.92 (d, J = 16 Hz, 2 H), 4.03 (s,
3 H), 3.90−3.35 (m, 8 H), 2.47 (s, 3 H), 1.25 (s, 18 H); 13C NMR
(100 MHz, CDCl3) δ: 184.56, 170.50, 165.81, 161.56, 150.66, 145.34,
141.79, 134.86, 130.05, 129.52, 128.53, 127.46, 126.96, 124.64, 122.55,
115.07, 83.70 (d, J = 8 Hz), 73.54 (J = 6 Hz), 56.82, 45.88, 41.55,
29.47, 29.43, 13.80; HRMS [M + H; ESI-ORBITRAP] calcd for
C33H43N7O8P: 696.2911; found: 696.2885.
Vakkalagadda, B. J.; Lai, C.; Kuang, S.-M.; US 20060293304,
December 28, 2006.
(6) Tripp, J. C.; Fanfair, D. D.; Schultz, M. J.; Murugesan, S.; Fox, R.
j.; Chen, C-P. H.; Ivy, S. E.; Payack, J. F.; Doubleday, W. W.; WO
2012106189, August 9, 2012.
(7) (a) Bartoli, G.; Palmieri, G.; Bosco, M.; Dalpozzo, R. Tetrahedron
Lett. 1989, 30, 2129. (b) Bartoli, G.; Bosco, M.; Dalpozzo, R.; Palmieri,
G.; Marcantoni, E. J. Chem. Soc., Perkin Trans 1. 1991, 2757.
(8) (a) Blaazer, A. R.; Lange, J. H. M.; van der Neut, M. A. W.;
Mulder, A.; den Boon, F. S.; Werkman, T. R.; Kruse, C. G.; Wadman,
W. J. Eur. J. Med. Chem. 2011, 46, 5086. (b) Ganser, C.; Lauermann,
E.; Maderer, A.; Stauder, T.; Kramb, J.-P.; Plutizki, S.; Kindler, T.;
Moehler, M.; Dannhardt, G. J. Med. Chem. 2012, 55, 9531.
(9) (a) Huestis, M. P.; Fagnous, K. Org. Lett. 2009, 11, 1357.
(b) Hands, D.; Bishop, B.; Cameron, M.; Edwards, J. S.; Cottrell, I. F.;
Wright, S. H. B. Synthesis 1996, 877.
(10) (a) Sakamoto, T.; Satoh, C.; Kondo, Y.; Yamanaka, H.
Heterocycles 1992, 34, 2379. (b) Mahadevan, I.; Rasmussen, M. J.
Heterocycl. Chem. 1992, 29, 359.
(11) Whelligan, D. K.; Thomson, D. W.; Taylor, D.; Hoelder, S. J.
Org. Chem. 2010, 75, 11.
(12) Eastgate, M. D.; Bultman, M. S.; Chen, K. C.; Fanfair, D. D.;
Fox, R. J.; La Cruz, T. E.; Mudryk, B. M.; Risatti, C. A.; Simpson, J. H.;
Soumeillant, M. C.; Tripp, J. C.; Xiao, Y.; US 20130203992, August 8,
2013.
(13) Wengryniuk, S. E.; Weickgenannt, A.; Reiher, C.; Strotman, N.
A.; Chen, K.; Eastgate, M. D.; Baran, P. S. Org. Lett. 2013, 15, 792.
(14) (a) Ullmann, F.; Bielecki, J. Chem. Ber. 1901, 34, 2174.
(b) Ullmann, F. Chem. Ber. 1903, 36, 2382. (c) Ullmann, F.; Sponahel,
P. Chem. Ber. 1905, 38, 2211. (d) Goldberg, I. Chem. Ber. 1906, 39,
1691.
ASSOCIATED CONTENT
* Supporting Information
■
S
Copies of 1H/13C NMR spectra and HPLC chromatograms
for new compounds. This material is available free of charge via
the Internet at http://pubs.acs.org.
(15) DeAmici, M.; DeMicheli, C.; Platini, F.; Della Bella, D.;
Caramazza, I. Eur. J. Med. Chem. 1988, 23, 511.
AUTHOR INFORMATION
Corresponding Author
*E-mail: martin.eastgate@bms.com.
■
(16) Pictet, A.; Spengler, T. Chem. Berichte 1911, 44, 2030.
(17) Fodor, G.; Nagubandi, S. Tetrahedron 1980, 10, 1279.
(18) (a) Chang, M.-Y.; Tai, H.-Y. Heterocycles 2011, 83, 2373.
(b) Silveira, C.; Bernardi, A.; Kaufman, T. Synlett 2002, 6, 907.
(19) Hock, H.; Lang, S. Ber. Dtsch. Chem. Ges. 1944, 77, 257.
(20) Fife, W. J. Org. Chem. 1983, 48, 1375.
Present Address
†Boehringer Ingelheim Pharmaceuticals, Inc., 900 Ridgebury
Rd, Ridgefield, CT 06877, United States.
Notes
(21) (a) Cobb, R. L.; McEwen, W. E. Chem. Rev. 1955, 55, 511.
(b) Keith, J. M. J. Org. Chem. 2010, 75, 2722.
The authors declare no competing financial interest.
(22) Banerjee, S.; Zeller, M.; Bruckner, C. J. Org. Chem. 2009, 74,
4283.
(23) El-Faham, A.; Albericio, F. Chem. Rev. 2011, 111, 6557.
(24) Londregan, A. T.; Jennings, S.; Wei, L. Org. Lett. 2011, 13, 1840.
(25) Kato, Y.; Okada, S.; Tomimoto, K.; Mase, T. Tetrahedron Lett.
2001, 42, 4849.
(26) Bromination of bis-methoxy azaindole 3 with POBr3 could be
accomplished in 50% yield using freshly distilled POBr3 (6 equiv) in
anisole at 130 °C.
ACKNOWLEDGMENTS
■
The authors would like to thank Drs. David Kronenthal,
Rajendra Deshpande, Rodney Parsons, and Robert Waltermire
for support, Prof. Phil Baran, Prof. Marty Burke, Dr. Nicolas
Cuniere, and Dr. Wendel Doubleday for useful discussions, Dr.
Qi Gao for X-ray crystallography, and Dr. Charles Pathirana
and Mr. Michael Peddicord for assistance with characterization.
(27) Sugimoto, O.; Mori, M.; Tanji, K.-I. Tetrahedron Lett. 1999, 40,
7477.
REFERENCES
■
(1) Wyatt, R.; Kwong, P. D.; Desjardins, E.; Sweet, R. W.; Robinson,
J.; Hendrickson, W. A.; Sodroski, J. G. Nature 1998, 393, 705.
(2) (a) Wang, T.; Zhang, Z.; Meanwell, N. A.; Kadow, J. F.; Yin, Z.;
Xue, Q. M.; Reguerio-Ren, A.; Matiskella, J. D.; Ueda, Y.; US
20040110785, June 10, 2004. (b) Ueda, Y.; Connolly, T. P.; Kadow, J.
F.; Meanwell, N. A.; Wang, T.; Chen, C-P. H.; Yeung, K.-S.; Zhang, Z.;
Leahy, D. K.; Pack, S. K.; Soundararajan, N.; Sirard, P.; Levesque, K.;
Thoraval, D.; US 20050209246, September 22, 2005.
(3) Bristol-Myers Squibb R&D Home Page; http://www.bms.com/
research/investigational/HIV/Pages/BMS-663068.aspx, accessed July
2014.
(4) Li, Z.; Zhou, N.; Sun, Y.; Ray, N.; Lataillade, M.; Hanna, G. J.;
Krystal, M. Antimicrob. Agents Chemother. 2013, 57, 4172.
(5) Soundararajan, N.; Qiu, Y.; Hu, W.; Kronenthal, D. K.; Sirard, P.;
Lajeunesse, J.; Droghini, R.; Chidambaram, R.; Qian, X.; Natalie, K. J.;
Pack, S. K.; Reising, N.; Tang, E.; Fakes, M. G.; Gao, Q.; Qian, F.;
(28) Lazzaroni, R.; Pini, D.; Bertozzi, S.; Fatti, G. J. Org. Chem. 1986,
51, 505.
(29) Wengryniuk, S. E.; Weickgenannt, A.; Reiher, C.; Strotman, N.
A.; Chen, K.; Eastgate, M. D.; Baran, P. S. Org. Lett. 2013, 15, 792.
(30) Jones, G. O.; Liu, P.; Houk, K. N.; Buchwald, S. L. J. Am. Chem.
Soc. 2010, 132, 6205.
(31) For a synthesis of the N-benzoyl piperidine see: Hendrick, C. E.;
McDonald, S. L.; Wang, Q. Org. Lett. 2013, 15, 3444.
(32) For a synthesis of the triazole see: Faridoon Hussein, W. M.;
Vella, P.; Ul Islam, N.; Ollis, D. L.; Schenk, G.; McGeary, R. P. Bioorg.
Med. Chem. Lett. 2012, 22, 380.
(33) (a) Klapars, A.; Huang, X.; Buchwald, S. L. J. Am. Chem. Soc.
2002, 124, 7421. (b) Antilla, J. C.; Klapars, A.; Buchwald, S. L. J. Am.
Chem. Soc. 2002, 124, 11684.
(34) Shafir, A.; Buchwald, S. L. J. Am. Chem. Soc. 2006, 126, 8742.
J
dx.doi.org/10.1021/jo5016008 | J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
(35) (a) Goodbrand, H. B.; Hu, N. X. J. Org. Chem. 1999, 64, 670.
(b) Kiyomori, A.; Marcoux, J.-F.; Buchwald, S. L. Tetrahedron Lett.
1999, 40, 2657.
(36) Xie, X.; Chen, Y.; Ma, D. J. Am. Chem. Soc. 2006, 128, 16050.
(37) Zheng, B.; Fox, R. J.; Sugiyama, M.; Fritz, A.; Eastgate, M. D.
Org. Process Res. Dev. 2014, 18, 636.
(38) Zheng, B.; Sugiyama, M.; Eastgate, M. D.; Fritz, A.; Murugesan,
S.; Conlon, D. A. Org. Process Res. Dev. 2012, 16, 1827.
(39) Hoffman, R. W.; Bruckner, D. New J. Chem. 2001, 25, 369.
K
dx.doi.org/10.1021/jo5016008 | J. Org. Chem. XXXX, XXX, XXX−XXX