An efficient, regioselective amination of 3,5-disubstituted pyridine N-oxides using saccharin as an ammonium surrogate

Article abstract of DOI:10.1021/ol303218p

A process for the regioselective amination of unsymmetrical 3,5-substituted pyridine N-oxides has been developed utilizing cheap, readily available saccharin as an ammonium surrogate. High conversions of the corresponding saccharin adducts have been achieved under mild reaction conditions. In situ deprotection under acidic conditions allows for a one-pot process to substituted aminopyridines. High regioselectivities were obtained from a variety of 3,5-disubstituted pyridine N-oxides.

Full text of DOI:10.1021/ol303218p

ORGANIC
LETTERS
XXXX
Vol. XX, No. XX
000–000
An Efficient, Regioselective Amination
of 3,5-Disubstituted Pyridine N‑Oxides
Using Saccharin as an Ammonium
Surrogate
Robert P. Farrell,* Maria Victoria Silva Elipe, Michael D. Bartberger, Jason S. Tedrow,
and Filisaty Vounatsos
Chemical Process Research and Development, Amgen Inc., One Amgen Center Drive,
Thousand Oaks, California 91320, United States
rfarrell@amgen.com
Received November 21, 2012
ABSTRACT
A process for the regioselective amination of unsymmetrical 3,5-substituted pyridine N-oxides has been developed utilizing cheap, readily available
saccharin as an ammonium surrogate. High conversions of the corresponding saccharin adducts have been achieved under mild reaction
conditions. In situ deprotection under acidic conditions allows for a one-pot process to substituted aminopyridines. High regioselectivities were
obtained from a variety of 3,5-disubstituted pyridine N-oxides.
Aminopyridines are common structural motifs in mole-
cules of high interest to the pharmaceutical industry.1
An attractive, direct approach to 2-aminopyridines is
via their corresponding pyridine N-oxides. The develop-
ment of conditions for the reaction of pyridine N-oxides
with various activating agents2 and nitrogen nucleophiles
to obtain 2-amino,3 2-amido,4 and 2-N-hetero5 pyridines
has improved access to 2-aminopyridines and derivatives.
While seeking an efficient, lower cost manufacturing route
for a clinical candidate, we explored a regioselective amina-
tion of a 3-arylthio, 5- aryloxy pyridine N-oxide (2) readily
prepared from dibromopyridine N-oxide (1) to obtain
compound 3 (Scheme 1).
(1) (a) Cui, J.; Tran-Dube, M.; Shen, H.; Nambu, M.; Kung, P.;
Pairish, M.; Jia, L.; Meng, J.; Funk, L.; Botrouis, I.; McTique, M.;
Grodsky, N.; Pyan, K.; Padrique, E.; Alton, G.; Timofeevski, S.;
Yamazaki, S.; Li, Q.; Zou, H.; Christensen, J.; Mroxzkowski, B.;
Bender, S.; Kania, R.; Edwards, M. J. Med. Chem. 2011, 54, 6342–
6363. (b) May, B.; Zorn, J.; Witkop, J.; Sherrill, J.; Wallace, G.; Prusiner,
S.; Cohen, F. J. Med. Chem. 2007, 50, 65–73. (c) Malamas, M.; Barnes,
K.; Hui, Y.; Johnson, M.; Lovering, F.; Condon, J.; Fobare, W.;
Solvibile, W.; Turner, J.; Hu, Y.; Manas, E.; Fan, K.; Olland, A.;
Chopra, R.; Bard, J.; Pangalos, M.; Reinhart, P.; Robichaud, A. Bioorg.
Med. Chem. Lett. 2010, 20, 2068–2073. (d) Zhang, D.; Ai, J.; Liang, Z.;
Li, C.; Peng, X.; Ji, Y.; Jiang, H.; Geng, M.; Luo, C.; Liu, H. Bioorg.
Med. Chem. 2012, 20, 5169–5180.
Scheme 1. Process Development Route to a 3-Arylthio,
5-Aryloxy Aminopyridine
(2) Activating agent examples include Ts2O, TsCl, POCl3, oxalyl
chloride, Ac2O, trifluoroacetic anhydride, and PyBroP.
(3) (a) Yin, J.; Xiang, B.; Huffman, M.; Raab, C.; Davies, I. J. Org.
Chem. 2007, 72, 4554–4557. (b) Londregan, A.; Jennings, S.; Wei, L.
Org. Lett. 2010, 12 (22), 5254–5257. (c) Wachi, K.; Terada, A. Chem.
Pharm. Bull. 1980, 28 (2), 465–472.
A few examples of regioselective additions to pyridine
N-oxides using a variety of C, N, and O based nucleophiles
(4) (a) Abramovitch, R.; Singer, G. J. Am. Chem. Soc. 1969, 91 (20),
5672–5673. (b) Abramovitch, R.; Rogers, R. J. Org. Chem. 1974, 39 (13),
1802–1807. (c) Manley, P.; Bilodeau, M. Org. Lett. 2002, 4 (18), 3127–
3129.
(5) (a) Keith, J. J. Org. Chem. 2008, 73, 327–330. (b) Keith, J. J. Org.
Chem. 2010, 75, 2722–2725. (c) Londregan, A.; Jennings, S.; Wei, L. Org.
Lett. 2011, 13 (7), 1840–1843.
r
10.1021/ol303218p
XXXX American Chemical Society

have been reported;3c,4b,5a,b,6,7 however the substrate
scope for such transformations has not been well-defined
nor has a clear understanding been made of what key
elements drive regioselectivity (e.g., substrate, nucleophile,
activator). For our purposes, regioselective aminations of
3,5-disubstituted pyridine N-oxideshave notbeen reported
in the literature. Coupled with the challenge to develop a
highly regioselective amination, it was vital to our process
to minimize reaction side products commonly observed
with direct aminations of pyridine N-oxides which often
lead to lower yields and complicated isolations. These side
products arise from the inherent reactivity of the amine
nucleophile with the activator as well as from counterion
attack of the activated species 5 leading to side product 9
(Scheme 2). It was recently reported that utilization of
PyBroP3b mitigates the side reaction of the activator reac-
ting with the nucleophile; however, the high cost and in-
stability of the phosphonium reagent8 make this approach
undesirable for large scale development.
to overcome problems with formation of the 6-tosylated side
product and low reaction conversions.
Having observed improved results with sterically bulky,
less nucleophilic HMDS we began to screen alternative
ammonia surrogates with similar characteristics but which
surprisingly have not been published as nucleophiles for
thisapplication. A major breakthrough wasachieved using
N-(tert-butoxycarbonyl)-p-toluenesulfonamide as the nu-
cleophile under Ts2O activation10 affording >99% con-
version to the corresponding aminated product in a 11.6:1
ratio of the 6 and 2 regioisomer respectively. Removal
of the sulfonamide group to generate the aminopyridine
was problematic leading us to evaluate other Gabriel-type
reagents11 including saccharin, phthalimide, and diethyl-
N-(tert-butoxycarbonyl)phosphoramidate. Gratifyingly,
saccharin also displayed excellent reactivity toward our
activated starting material (>99% conversion) with good
regioselectivity under unoptimized conditions (5.8:1). Fur-
thermore, direct treatment of the unisolated saccharin
adduct with aqueous HCl at 80 °C resulted in clean
deprotection to the desired aminopyridine. After initial reac-
tion optimization we also found both phthalimde and diethyl-
N-(tert-butoxycarbonyl)phosphoramidate to give high con-
version to their corresponding adducts although, importantly,
regioselectivity was significantly impacted by the choice of
nucleophile. This variability is highlighted on a model sub-
strate (Table 1). Although diethyl-N-(tert-butoxycarbonyl)-
Scheme 2. Side Products Observed During Development of a
Regioselective Direct Amination
Table 1. Effect of Nucleophile on Amination Regioselectivity
We thus examined alternative literature conditions3a
in which we treated our substrate with tert-butyl amine
(4.0 equiv) and Ts2O (2.0 equiv) in CH2Cl2 at 0 °C. Un-
fortunately, despite the excess of reagents, we obtained low
conversions to product and observed many reaction impu-
rities leading to a complex reaction mixture upon TFA
deprotection. However, switching to HMDS (4.0 equiv) as
the nucelophile with the otherwise same reaction conditions
gave our desired 6-substituted product in 45% assay yield.9
While the 2-substituted regioisomer was also observed,
substitution at the 4-position was only detected in very trace
quantities (<0.5%) presumably due to the steric bulk of the
3,5-substituents. Despite this initial success, we were unable
entry
nucleophile
regioselectivitya conversion
1
2
3
saccharin
11.8:1
17.0:1b
2.6:1
97%
99%
100%
diethyl-N-(Boc)phosphoramidate
phthalimide
a Reaction conditions: Combined N-oxide (1.0 equiv) with nucleo-
phile (1.1.equiv), and the mixure was slurried in CH2Cl2 (10 mL/g
starting material). Added iPr2EtN (2.0 equiv) with subsequent cooling
to 0 °C. Added tosyl chloride (1.2 equiv) as a solid in one portion
followed by stirring at 0 °C for 3À16 h. Regioselectivity and conversions
were calculated by analysis of HPLC/MS peak area%. b Regioselectivity
after deprotection to aminopyridine.
phosphoramidate12delivers excellent regioselectivity, its
relatively high cost made it unattractive for our purposes,
and thus saccharin was chosen for further development.
In an effort to maximize regioselectivity, optimization of
our saccharin process included evaluation of base, activa-
tor, solvent, and reaction temperature.13
(6) Ujjainwalla, F.; Walsh, T. Tett. Lett. 2001, 42, 6441–6445.
(7) For a regioselective cyanation of a 3-alkoxy, 5-thiazole pyridine
N-oxide using TMSCN/Et3N, see: Umemura, K.; Noda, H.; Yoshimura,
J.; Konn, A.; Yonezawa, Y.; Shin, C. Tetrahedron Lett. 1997, 38 (20), 3539–
3542.
(8) Bromotripyrrolidinophosphonium hexafluorophosphate PyBroP:
CAS# 132705-51-2: Available from Aldrich: 25 g, $671. Storage condi-
tions: À20 °C. Decomposition of PyBroP over prolonged storage gen-
erates pyrrolidine which can act as a nucleophile under the reaction
conditions.
(9) Under the reaction conditions the TMS groups cleaved to give the
unprotected aminopyridine in 45% solution assay yield with 6 HPLC
area % tosylated side product and 31 HPLC area % unreacted starting
material. 2-Aminopyridine regioisomer content was not quantified due
to overlap with the starting material.
(10) Conditions: 2 equiv of Ts2O, 1.3 equiv of DBU, 1.1 equiv of
N-(tert-butoxycarbonyl)-P-toluenesulfonamide, CH2Cl2 (10 mL/g starting
material), 0 °C.
(11) For a review of Gabriel reagents see: Ragnarsson, U.; Grehn, L.
Acc. Chem. Res. 1991, 24, 285–289.
(12) Diethyl-N-(tert-butoxycarbonyl)phosphoramidate (CAS# 85232-
02-6) avialable from TCI America: 5 g, $260.
B
Org. Lett., Vol. XX, No. XX, XXXX

for our purposes.15 The choice of solvent16 was another
important parameter, with CH2Cl2 delivering significantly
higher regioselectivity for the desired isomer than THF,
acetone, DMF, and DME. Due to the high polarity of
the starting material, other less polar solvents were not
evaluated. iPr2NEt slightly outperformed pyridine, DBU,
NMM, and Et3N. Theoretically, 2 equiv of base are
necessary for deprotonation of the nucleophile and scaveng-
ing the TsOH byproduct from rearomatization. During
initial screening, we evaluated the use oflessthan2 equiv of
base allowing our substate, which contains a basic nitro-
gen, to also serve as a base (see closely related structure
entry 1, Table 2). However, the unintended consequence of
protonating our substrate was lower regioselectivity (6.7:1
vs 9.4:1) using 1.3 and 2.3 equiv of iPr2NEt respectively.
This result implied substrate electronics is an important
Table 2. 3-Arylthio, 5-Aryloxy Pyridine N-Oxide Aminations
Using Saccharin
Table 3. Regioselective Amination of 3,5 Disubstituted Pyridine
N-Oxidesa
a Reaction conditions: Combined N-oxide (1.0 equiv) with saccharin
(1.1equiv), and the mixture was slurried in CH2Cl2 (6 mL/g N-oxide).
Added iPr2EtN (2.0 equiv) with subsequent cooling to 0 °C. Added tosyl
chloride (1.2 equiv) as a solution in CH2Cl2 (4 mL/g N-oxide) followed
by stirring at 0 °C for 3 h. Regioisomer ratio refers to 2 and 6 substituted
isomers. Regioisomer ratio and conversions were calculated by analysis
of HPLC/MS peak area%. b Reported yields are for a single regioisomer
using an unoptimized crystallization. c Isolated as the 6-aminopyridine
derivative (Table 4, entry 1).
Of these variables, the choice of electrophilic activator
was the most influential in the regioselective outcome. The
highest regioselectivities were obtained using alkyl substi-
tuted benzenesulfonyl chlorides such as 4-tert-butylsulfo-
nyl chloride (6.7:1), while remarkably little regioselectivity
was observed with triflic anhydride (1.3:1).14 Tosyl chlor-
ide (5.9:1) possessed the best mix of performance and price
(13) Regioselectivity was improved from 9.4:1 to 11.1:1 with our
development substrate by lowering the reaction temperature from 0 to
À15 °C. Reaction completion time in this case required 6 h.
(14) The following relative regioselectivities were observed using our
development substrate: TfO, 1.3:1; Ms2O, 3.9:1; 4-NO2-benzenesulfonyl
chloride, 4.3:1; Ts2O, 5.4:1; 4-MeO-benzenesulfonyl chloride, 5.5:1; tosyl
chloride, 5.9:1; mesitylene sulfonyl chloride, 6.4:1; 4-tert-butylbenzenesulfo-
nyl chloride, 6.7:1 (All reactions were carried out with 1.1 equiv of saccharin,
2.0 equiv of iPr2NEt, in CH2Cl2 (10 mL/g starting material) at 0 °C).
(15) Aldrich pricing: p-Toluenesulfonyl chloride, 1 kg = $66; 4-tert-
butylbenzenesulfonyl chloride, 50 g = $127.
(16) It is noteworthy that these amination conditions are tolerant of
water with full conversions realized using a 9:1 acetone/water solvent
system.
a Refer to Table 2 for conditions and footnotes.
Org. Lett., Vol. XX, No. XX, XXXX
C

factor for regioselectivity. To test this hypothesis, a series
of analogous 3-aryloxy, 5-arylthio pyridine N-oxides were
prepared. Their performance under the optimized amina-
tion conditions is shown (Table 2).
Table 4. Through Process Amination/Deprotection to
3,5-Disubstituted Aminopyridines
While for all these examples substitution is favored
ortho to the aryl oxy substituent (see entries 1À6, Table 2),
the regioselectivity is indeed highly influenced by altering
the electronics of the substituents as highlighted in entries
4À6, Table 2. From these findings we reasoned that the
scope of our regioselective amination protocol may well
apply to other electronically differentiated 3,5-disubstituted
pyridine N-oxides (Table3). Inthe case of 3,5-diarylethers,
marginal regioselectivity was observed ortho to the more
electron-rich aryloxy substituent (entry 1, Table 3).
3-Bromo, 5-aryloxy, and alkyloxy substrates (readily
prepared from dibromopyridine N-oxide) aminate predomi-
nately ortho to the electron-rich 5-substituent (entries 2À4,
Table 3) making for a useful class of intermediates. Cur-
iously, in the case of 3-bromo, 5-arylthio pyridine N-oxide
(entry 6, Table 3) substitution was modestly favorable
ortho to the bromine substituent. These amination con-
ditions gave both lower selectivity and conversion when
applied to monosubstituted substrates such as 3-methoxy
pyridine N-oxide (entry 7, Table 3).
In conclusion, an efficient and regioselective amination
of electronically differentiated 3,5-pyridine N-oxides has
been developed using novel amination reaction conditions.
Saccharinisideally suitedasanammonia surrogate forthis
application, delivering both high conversions and high
regioselectivity. This protocol overcomes challenges asso-
ciated with direct amination of pyridine N-oxides caused
by the competetive reaction of the nucleophile with the
activator. Cleavageofthe saccharinintermediates hasbeen
achieved in a one-pot fashion upon treatment with aq. HCl
or H2SO4. The substrate examples presented offer insight
into the applicability of this protocol toward related 3,5-
disubstituted pyridine N-oxides.
With acid soluble substrates, deprotection of the saccharin
adducts can be achieved by direct addition of aqueous
HCl or H2SO4 to the reaction mixture. In this way,
primary aminopyridines were obtained in a single isola-
tion (Table 4).
In a recent publication, various electrophilicity indices
computed via quantum mechanics for unactivated sub-
strates have been found to correspond well to experimen-
tally observed regiochemical ratios,5c and we observed the
same phenomenon. Curiously, however, our own calcula-
tions of both electrophilic Fukui indices and the LUMO
distribution for activated (tosyl, triflyl) substrates suggest
a significant decrease in inherent 2- versus 6-position
reactivity far below that which is experimentally observed.
This result may be suggestive of other, nonobvious
mechanistic factors beyond the relative electrophilicity of
the competing reactive centers upon substrate activation
(refer to the Supporting Information for experimental
details and results).
Acknowledgment. The authors thank our Amgen re-
search colleagues Eric Bercot, Matthew Bio, Andrew
Cosbie, Kelly Hu, and Kevin Turney for support and
helpful discussions.
Supporting Information Available. General experimen-
tal procedures and spectral data for all products. This
material is available free of charge via the Internet at
http://pubs.acs.org.
The authors declare no competing financial interest.
D
Org. Lett., Vol. XX, No. XX, XXXX