Assembly of 4-aminoquinolines via palladium catalysis: A mild and convenient alternative to SNAr methodology

Article abstract of DOI:10.1021/jo062168u

4-Aminoquinolines, classically prepared via S_NAr chemistry from an amine and 4-haloquinoline, are important scaffolds in medicinal chemistry. Interest in these compounds prompted us to explore palladium catalysis as an alternative to the existing methods for their preparation. Initial results followed by an iterative screening paradigm confirmed Pd(OAc)₂/ DPEphos/K₃PO₄ as a mild and convenient alternative for the formation of the C-N bond in 4-aminoquinolines. A description of the screen and the scope of this methodology are discussed herein.

Full text of DOI:10.1021/jo062168u

Assembly of 4-Aminoquinolines via Palladium
Catalysis: A Mild and Convenient Alternative to
SNAr Methodology
Brandon J. Margolis,* Kimberly A. Long,† Dana L. T. Laird,
J. Craig Ruble, and Shon R. Pulley
Lilly Research Laboratories, Eli Lilly and Company,
Indianapolis, Indiana 46285
FIGURE 1. Marketed drugs/drug candidates containing a 4-AQ core.
chloro). The SNAr reaction is often carried out neat, or in the
presence of phenol, at elevated temperatures (∼100-160 °C).5
Isolating the desired product can be difficult since the reaction
mixtures tend to solidify upon cooling and require an acid/base
extraction to remove the excess phenol. Interest in these
structures prompted us to explore palladium-catalyzed amination
chemistry6 as a mild and convenient alternative to classical
methodologies for the synthesis of 4-AQ’s. In these cases, one
can simply adsorb the crude reaction mixture onto silica gel
and purify directly, thus avoiding a formal aqueous workup and
making this attractive for small scale or parallel reactions.
Encouraged by an initial result with Pd(OAc)2/DPEphos/K3-
PO4 in toluene to couple 4-bromoquinoline (4a) with a primary
alkylamine, a series of iterative screens to optimize this catalyst
system were undertaken in a controlled fashion selecting the
R-methyl amine (6) present in chloroquine as a representative
amine in these studies (eq 1).
margolis_brandon_j@lilly.com
ReceiVed October 18, 2006
4-Aminoquinolines, classically prepared via SNAr chemistry
from an amine and 4-haloquinoline, are important scaffolds
in medicinal chemistry. Interest in these compounds prompted
us to explore palladium catalysis as an alternative to the
existing methods for their preparation. Initial results followed
by an iterative screening paradigm confirmed Pd(OAc)2/
DPEphos/K3PO4 as a mild and convenient alternative for the
formation of the C-N bond in 4-aminoquinolines. A
description of the screen and the scope of this methodology
are discussed herein.
There are numerous therapeutic agents which highlight the
significance of 4-aminoquinolines (4-AQ’s) as important scaf-
folds in medicinal chemistry, including: those able to treat
malaria,1 cancer,2 gastric disorders,3 and those able to activate
targets in the central nervous system (Figure 1).4
Recognizing that an exhaustive matrix evaluation was im-
practical given the number of possible permutations, we took a
three-step approach in the following order: (1) solvents, (2)
bases, and (3) phosphine ligands. The decision was made to
use only one palladium source [precatalyst Pd(II)(OAc)2] for
this work. In addition, for convenience, an artificial end point
of 18 h was imposed for the reactions. For the base and
phosphine screens, reactions were analyzed by analytical HPLC,
and yields were calculated relative to an internal standard (IS).7
Using the Pd(OAc)2/DPEphos/K3PO4 system, several solvents
including dioxane, PhMe, and DME were evaluated. Either
dioxane or PhMe was found to be optimal (these results were
not quantitated). Of these, dioxane was chosen for further study
mainly for its relatively low boiling point and low UV cutoff.
A small survey of bases commonly used in palladium-catalyzed
aminations was preformed as the second iteration; K3PO4 was
found to be optimal for this system (Table 1).
Historically, the assembly of 4-AQ’s involves a SNAr reaction
between the requisite amine and a 4-haloquinoline (usually
† Participant in the 2006 Lilly Corporate Summer Intern Program.
(1) Chloroquine exemplifies a class of 4-AQ antimalarial agents. Also
see: (a) O’Neill, P. M.; Ward, S. A.; Berry, N. G.; Jeyadevan, J. P.; Biagini,
G. A.; Asadollaly, E.; Park, K. B.; Bray, P. G. Curr. Top. Med. Chem.
2006, 6 (5), 479-507. (b) Madrid, P. B.; Sherrill, J.; Liou, A. P.; Weisman,
J. L.; DeRisi, J. L.; Guy, R. K. Bioorg. Med. Chem. Lett. 2005, 15, 1015-
1018. (c) Pradines, B.; Tall, F.; Ramiandrasoa, A.; Spiegel, C.; Sokhna,
C.; Fusai, J.; Mosnier, J.; Daries, W.; Trape, J. F.; Kunesch, G.; Parzy, D.;
Rogier, C. J. Antimicrob. Chemother. 2006, 57 (6), 1093-1099. (d) Biot,
C.; Daher, W.; Chavain, N.; Fandeur, T.; Khalife, J.; Dive, D.; De Clercq,
E. J. Med. Chem. 2006, 49 (9), 2845-2849.
(2) Ruchelman, A. L.; Singh, S. K.; Abhijit, R.; Wu, X. H.; Yang, J.;
Li, T.; Liu, A.; Liu, L. F.; LaVoie, E. J. Bioorg. Med. Chem. 2003, 11,
2061-2073.
(3) SKF-97574 was in Phase III clinical trials for the treatment of upper
gastrointestinal ulcers. (a) Leach, C. A.; Brown, T. H.; Ife, R. J.; Keeling,
D. J.; Parsons, M. E.; Theobald, C. J.; Wiggall, K. J. J. Med. Chem. 1995,
38 (14), 2748-2762. (b) Atkins, R. J.; Breen, G. F.; Crawford, L. P.; Grinter,
T. J.; Harris, M. A.; Hayes, J. F; Moores, C. J.; Saunders, R. N.; Share, A.
C.; Walsgrove, T. C.; Wicks, C. Org. Process Res. DeV. 1997, 1, 185-
197.
(5) (a) Surrey, A. R.; Hammer, H. F. J. Am. Chem. Soc. 1946, 68, 113-
116. (b) Surrey, A. R.; Cutler, R. A. J. Am. Chem. Soc. 1951, 73, 2623-
2626.
(4) Tacrine is approved to treat the symptoms of Alzheimer’s disease.
For a series of 4-aminoquinoline-based H3 antagonists, see: Turner, S. C.;
Esbenshade, T. A.; Bennani, Y. L.; Hancock, A. A. Bioorg. Med. Chem.
Lett. 2003, 13, 2131-2135.
(6) (a) Hartwig, J. F. In Handbook of Organopalladium Chemistry for
Organic Synthesis; Negiahi, E. I., Ed.; Wiley-Interscience: New York, 2002;
Vol. 1, p 1051. (b) Muci, A. R.; Buchwald, S. L. Top. Curr. Chem. 2002,
219, 131-209.
10.1021/jo062168u CCC: $37.00 © 2007 American Chemical Society
Published on Web 02/13/2007
2232
J. Org. Chem. 2007, 72, 2232-2235

TABLE 1. Survey of Common Bases
TABLE 4. Quinoline Substrate Scope (1)
entrya
baseb
% yield 7ac
1
2
3
K3PO4
NaOt-Bu
K2CO3
89
78
39
a Reactions were run in dioxane, using 4 mol % Pd(OAc)2/8 mol %
DPEphos, then analyzed after 18 h at 85 °C. b Molar ratio of base to bromide
was 2.5. c Yield as detected by quantitative HPLC using 4,4′-dimethoxy-
biphenyl as IS.
entrya
X
HPLC yield, %b
isolated yield, % 7ac
1
2
3
Cl
Br
I
93
95
52
92
88
N/A
TABLE 2. Screen of Phosphine Ligands
a Reactions were run in dioxane, using 4 mol % Pd(OAc)2/8 mol %
DPEphos, then analyzed after 18 h at 85 °C. b Yield relative to IS 4,4′-
dimethoxybiphenyl. c Isolated yield averaged from 2 runs which were
adsorbed onto silica and purified after 18 h at 85 °C
%
%
yield
yield
entrya
ligandb
7ac
entrya
ligandb
7ac
1
DPEphos
(R)-BINAP
P(t-Bu)3
S-Phos
95
94
92
85
10g
11h
12
dtbf
50
50
49
44
2
PCy2 biphenyl
CTC-Q-Phos
PPh2 NMe2
biphenyl
PEPPSI
(IPr)Pd(acac)Cl
X-Phos
TABLE 5. Quinoline Substrate Scope (2)
3d
4
13i
5
DPPF
84
73
71
148,j
159,j
16
40
37
18
6
DavePhos
PCy2 Me
biphenyl
Xantphos
7e
8
64
17k
P(t-Bu)2 Me
biphenyl
PPh3
11
9f
dippf
58
18
5
a Each reaction ran for 18 h at 85 °C with 5 mol % Pd(OAc)2. b The
per-atom ratio of P:Pd for this screen was held constant at 4:1. Therefore,
20 mol % of L was used for monodentate ligands, and 10 mol % for
bidentate ligands. Except where indicated by reference or CAS number,
names are as listed on www.strem.com. c Yield as detected by quantitative
HPLC using 4,4′-dimethoxybiphenyl as IS. d Also used was the [(t-
Bu3PH)BF4] salt which gave an identical yield.10 e CAS [251320-86-2].
f CAS [97239-80-0]. g CAS [84680-95-5]. h CAS [247940-06-3]. i CAS
[240417-00-9]. j KOt-Bu was used as the base in DME. k CAS [255837-
19-5].
A
B
C
isolated yield, % 7a
isolated yield, % 10a
H
88 (7a)
63 (7b)
78 (7c)
74 (7d)
58b,c (7e)
32c (7f)
51d (7h)
75 (7i)
SMe
OMe
Cl
Br
65b (10e)
74 (10f)
85 (10h)
NO2
CN
CO2Me
F
TABLE 3. Varying the Ratio of P:Pd
F
74 (7j)
79 (7k)
entrya
% yield 7ab
P:Pd ratioc
OMe
OMe
1
2
3
95
89
35
4:1
2:1
1:1
a Average of 2 runs; reactions were cooled and adsorbed onto silica gel,
then purified after 18 h at 85 °C; when appropriate, yields have been adjusted
down to account for any trace solvents present as detected by 1H NMR.
b Isolated as a mixture of 4-amino-6-bromo/4-bromo-6-amino with a ratio
of 15:1 by 1H NMR. c Reaction did not go to completion after 18 h.
d Reduced quinoline was detected in crude reaction by ESMS but never
isolated.
a Reactions were run in dioxane, using 4 mol % of Pd(OAc)2, then
analyzed after 18 h at 85 °C. b Yield as detected by quantitative HPLC
using 4,4-dimethoxybiphenyl as IS. c Expressed on a per-atom basis.
A diverse sampling of phosphine ligands was selected to build
upon the DPEphos result. Several of these gave high HPLC
yields (identical, within experimental error) of coupled product
7a (Table 2, entries 1-3).
Cost was selected as a filter criterion, and DPEphos was
chosen for further study.11 Since our initial experiments used a
per-atom ratio of P:Pd of 4:1, we wondered if a lower phosphine
loading would work equally as well in the chosen 18 h time
window. We found that a ratio of 4:1 was superior to a lower
phosphine loading; however, good conversion was also obtained
with a 2:1 ratio (Table 3, entries 1 and 2).
This process most likely proceeds by a Pd-catalyzed mech-
anism as demonstrated by a control experiment lacking the Pd-
(OAc)2. Under these conditions, no 7a was detected even after
one week at 85 °C. To summarize the results of the screen
iterations, an optimized set of conditions for this reaction
follow: 1.0 equiv of halide, 8 mol % DPEphos/4 mol % Pd-
(OAc)2, 1.5 equiv of amine, 2.5 equiv of K3PO4 in dioxane at
85 °C.12
In an attempt to demonstrate the generality of these reaction
conditions, additional quinolines were prepared and/or pur-
chased. Scheme 1 depicts the synthetic route used to generate
(7) Yields are reported with a 95% measurement error prediction interval.
The prediction interval reflects both the uncertainty in the measure for each
reaction condition as well as the two-point calibration used within each
batch. The interval was derived assuming normally distributed errors for
the AUC measurements, scaling of the reaction product AUC values to the
spiked internal standard AUC value within each injection, a correlation
coefficient of 0.7 between the AUC of the product of interest and the internal
standard across repeated injections, and a constant AUC total coefficient
of variation estimated from the between and within batch variance from
the spiked internal standard AUC values.
(9) Marion, N.; Ecarnot, E. C.; Navarro, O.; Amoroso, D.; Bell, A.;
Nolan, S. P. J. Org. Chem. 2006, 71, 3816-3821.
(10) Netherton, M. R.; Fu, G. C. Org. Lett. 2001, 3, 4295-4298.
(11) Strem Catalog No. 20 (2004-2006).
(12) During the course of our investigation, Beletskaya and co-workers
reported the amination of 4-chloroquinoline derivatives under the influence
of palladium catalysis employing either BINAP or DPPF derivatives as
ligands, see: Beletskaya, I. P.; Tsvetkob, A. V.; Latyshev, G. V.; Lukashev,
N. V. Russ. Chem. Bull. Int. Ed. 2005, 54 (1), 215-219.
(8) Aldrich Chem. Files 2006, 6 (3).
J. Org. Chem, Vol. 72, No. 6, 2007 2233

SCHEME 1. Preparation of 4-Haloquinolines
TABLE 6. Direct Comparison to Literature Examples
SCHEME 2. Alternative Preparation of Substituted
4-Haloquinolines
the 4-haloquinolines from their respective anilines.13 Noteworthy
is the possibility of regioisomers in the instances when the
desired quinoline is derived from a meta-substituted aniline (for
example, 2f and 2g). In our hands, these isomers were separated
by silica gel chromatography after the bromination step to give
pure 4f and 4g, respectively.
Scheme 2 shows a complementary methodology particularly
useful for accessing quinolines in a regiospecific fashion, starting
from the requisite substituted o-aminoacetophenone.14
As alluded to in Scheme 1, one can easily convert a quinolin-
4-ol to its corresponding 4-haloquinoline (chloro or bromo) via
treatment with either POCl3 at reflux5 or PBr3/DMF, respec-
tively.15 Operationally, we feel that preparation of the bromide
over the chloride is advantageous for two reasons: (1) the
reaction conditions are mild and rapid (ambient temperature
versus refluxing POCl3) and (2) products are easily isolated from
the reaction mixture (via a precipitation from water or sodium
bicarbonate (aq) then filtration versus in Vacuo removal of POCl3
followed by careful neutralization and a formal aqueous
workup).
a Average of 2 runs; reactions were cooled and adsorbed onto silica gel,
then purified after 18 h at 85 °C.
(8) was detected along with coupled product (7a) and a small
amount of dehalogenated quinoline (amount not quantitated).
In general, the coupling reaction tolerates a range of electronic
effects on the quinoline as shown in Table 5. Several examples
did give a low yield when the coupling partner was R-methy-
lamine 6 (compounds 7e,f,h). These bromoquinolines were
evaluated with another amine lacking the R-substitution (9). As
shown by the notable increase in yield (Table 5, 10e,f,h), the
R-methylamine is a more challenging substrate for this reaction.
Direct comparison to a literature example with 4-chloro-6,7-
methylenedioxyquinoline and two different amines demonstrates
the advantages of the Pd-catalyzed amination route to 4-AQ’s
(Table 6).
In general, the coupling reaction with 8 mol % DPEphos/4
mol % Pd(OAc)2, 1.5 equiv of amine, and 2.5 equiv of K3PO4
in dioxane at 85 °C tolerates both 4-chloroquinoline (3) and
4-bromoquinoline (4a) (Table 4). Interestingly, this catalyst
system did not work well with an iodo group (Table 4, entry
3). After the 18 h time period, unreacted iodo starting material
(13) (a) Davies, D. T.; Elder, J. S.; Forrest, A. K.; Jarvest, R. L.; Pearson,
N. D.; Sheppard, R. J. WO 2004/014361A1 (PCT Int. Appl.), Feb 19, 2004.
(b) Walz, A. J.; Sundberg, R. J. J. Org. Chem. 2000, 65 (23), 8001-2010.
(c) Gordon, H. J.; Martin, J. C.; McNab, H. J. Chem. Soc., Chem. Commun.
1983, 957-958.
(14) (a) Kubo, K.; Shimizu, T.; Ohyama, S.; Murooka, H.; Iwai, A.;
Nakamura, K.; Hasegawa, K.; Kobayashi, Y.; Takahashi, N.; Takahashi,
K.; Kato, S.; Izawa, T.; Isoe, T. J. Med. Chem. 2005, 48 (5), 1359-1366.
(b) Ruchelman, A. L.; Houghton, P. J.; Zhou, N.; Liu, A.; Liu, L. F.; LaVoie,
E. J. J. Med. Chem. 2005, 48 (3), 792-804. (c) Osawa, T.; Kubo, K.;
Murooka, H.; Nakajima, T. WO 98/47873 (PCT Int. Appl.), Oct 29, 1998.
(d) Nakajima, T.; Kamimasahara, M.; Matsunaga, N. JP 2002/030083A (Jpn.
Kokai Tokkyo Koho), Jan 29, 2002.
An additional advantage is realized with polyhaloquinolines
(eq 2). In our hands, the SNAr conditions reported by Surrey
and Cutler5b with 4-bromo-6,8-difluoroquinoline (4j) generated
a mixture of 4-substituted 7j and 8-substituted 15 in a ratio of
approximately 1:2 after chromatography with the desired
4-substituted 7j being the minor component. The Pd-catalyzed
conditions are specific for the formation of 7j (74 % isolated
yield with no detectable 15 present).
Amine substrate scope was expanded to include a secondary
amine and an aniline, both of which ran in moderate to good
yield (Table 7).
(15) Axten, J. M.; Daines, R. A.; Davies, D. T.; Gallagher, T. F.; Jones,
G. E.; Miller, W. H.; Pearson, N. D.; Pendrak, I. (PCT Int. Appl) WO2004/
002992 A1, Jan 8, 2004.
2234 J. Org. Chem., Vol. 72, No. 6, 2007

TABLE 7. Amine Substrate Scope
convenient than the existing SNAr methodology. The described
conditions were rapidly optimized by using an iterative screening
technique, and the catalyst system was shown to be effective
for a variety of quinoline/amine combinations.
Experimental Section
General Procedure for Palladium Catalyzed Amination of
4-Haloquinolines. An oven-dried 40-mL vial was charged with
the 4-haloquinoline (1.0 mmol), Pd(OAc)2 (4 mol %), DPEphos (8
mol %), K3PO4 (2.5 mmol), and the requisite amine (1.5 mmol).
An upside-down 24/40 septum was placed over the vial and an 18
gauge needle inserted (as a vent) while the resulting mixture was
purged with argon or nitrogen for several minutes through a second
needle. Dioxane (4 mL) was introduced through the septum. The
resulting suspension was sparged with argon or nitrogen for 3-5
min. The vial was quickly capped, then heated to 85 °C (block
temp) for 18 h and cooled. The mixture was adsorbed onto silica
gel and purified. Solvent systems for chromatography were based
on either mixtures of EtOH (contains 10% NH4OH)/CH2Cl2 or THF/
hexanes (THF/hexanes mix contains 2% dimethylethylamine, and
THF is inhibitor free).
a Average of 2 runs; reactions were cooled and adsorbed onto silica gel,
then purified after 18 h at 85 °C; when appropriate, yields have been adjusted
down to account for any trace solvents present as detected by 1H NMR.
b This example gives a challenging mixture to purify; several mixed fractions
were discarded in both cases.
Finally, the racemization of R-chiral amines has been
observed in Pd-catalyzed reactions.16 It was proposed that the
coupling reaction of (R)-(+)-R-methylbenzylamine and 4a
would be a good measure of this effect with this catalyst system
(eq 3). As expected with a bidentate phosphine ligand, the
reaction proceeded with only a small erosion of enantiomeric
excess (from 98 % ee to 92 % ee).
Acknowledgment. We thank Michael Kalbfleisch for helpful
discussions and guidance in the development of our quantitative
HPLC assay and Robert D. Boyer for helpful discussions
1
surrounding H, 13C, and 19F NMR leading up to the charac-
terization of several of the compounds reported herein. We also
acknowledge the efforts of Richard E. Higgs for performing
the statistical analysis on the data set generated from our screen.
Finally, we thank James R. McCarthy for helpful discussions
surrounding the preparation of this manuscript.
Herein we have demonstrated an alternative methodology for
the synthesis of 4-aminoquinolines. The key step in this
methodology exploits palladium catalysis to form the desired
C-N bond under conditions that are milder and debatably more
Supporting Information Available: Additional experimental
procedures and spectral characterization data. This material is
available free of charge via the Internet at http://pubs.acs.org.
(16) Wagaw, S.; Rennels, R. A.; Buchwald, S. L. J. Am. Chem. Soc.
1997, 119, 8451-8458.
JO062168U
J. Org. Chem, Vol. 72, No. 6, 2007 2235