Rh(III)-catalyzed C-H activation and double directing group strategy for the regioselective synthesis of naphthyridinones

Article abstract of DOI:10.1021/ja405140f

A general Rh(III)-catalyzed synthesis of naphthyridinone derivatives is described. It relies on a double-activation and directing approach leveraging nicotinamide N-oxides as substrates. In general, high yields and selectivities can be achieved using low

Full text of DOI:10.1021/ja405140f

Communication
pubs.acs.org/JACS
Rh(III)-Catalyzed C−H Activation and Double Directing Group
Strategy for the Regioselective Synthesis of Naphthyridinones
,†
John R. Huckins,* Eric A. Bercot,* Oliver R. Thiel, Tsang-Lin Hwang, and Matthew M. Bio
Chemical Process Research and Development, Amgen Inc., One Amgen Center Drive, Thousand Oaks, California 91320, United
States
*
S Supporting Information
Scheme 1. Heterocycle Synthesis via Rh(III)-Catalyzed C−H
Activation with Internal Directing Group/Oxidant
ABSTRACT: A general Rh(III)-catalyzed synthesis of
naphthyridinone derivatives is described. It relies on a
double-activation and directing approach leveraging
nicotinamide N-oxides as substrates. In general, high
yields and selectivities can be achieved using low catalyst
loadings and mild conditions (room temperature) in the
couplings with alkynes, while alkenes require slightly more
elevated temperatures.
itrogen-containing heterocycles are ubiquitous in both
N
natural products and pharmaceuticals; therefore, new,
selective methods for their preparation are an important focus of
research. Traditional methods often require harsh reaction
conditions, rendering them incompatible with functionalized
1
substrates. Transition metal catalysis offers an attractive
alternative. In the past decade, directed C−H activation
approaches have enabled more-direct access to heterocyclic
2
,3
scaffolds.
Recently, Rh(III)-catalyzed [4+2] annulation
reactions of aromatic carboxamides with alkynes have emerged
as a powerful and efficient method to construct fused 2-pyridone
4
heterocycles. A newer strategy employs an internal, multifunc-
tional group that both directs the metal to the desired site of C−
H activation and reoxidizes it after the formation of a C−N bond,
5
typically by cleavage of an N−O bond.
The utility of this method is highlighted in the ability to rapidly
access isoquinolones from rather simple starting materials.
Extension of this reaction to other heterocycles was met with
partial success; examples in which the parent phenyl ring has
been replaced by electron-rich heterocycles (furan, thiophene,
We postulated that we could gain access to the desired
naphthyridinones by C−H activation of functionalized nicotin-
amide derivatives 2 (eq 1) in the presence of an unsaturated
9
coupling partner [norbornadiene or (trialkylsilyl)acetylene].
4a,5a,b
indoles) have recently been reported.
In contrast, extension
The use of nicotinamide derivatives in the Rh(III)-catalyzed
annulation with norbornadiene furnished a mixture of products
resulting from C−H activation at C-2 and C-4 (Chart 1).
Interestingly, the regioselectivity with regard to the C−H
activation could only be influenced subtly by introducing varying
substituents at the 5-position. This is in contrast to reports of
excellent regioselectivity for meta-substituted substrates in the
of this method to pyridine derivatives has been only partially
successful. Rovis observed that the reaction failed to afford any of
4a
the naphthyridinone product under the conditions studied,
while Glorius reported the isolation of a 1:1 mixture of products
resulting from the unselective activation of the C-2 and C-4
5b
positions of the nicotinamide ring (Scheme 1). Li also reported
3c,d
6
parent phenyl system. The use of silylalkynes provided only
trace amounts of the desired product; however, 4-octyne
furnished adduct 1e in 78% yield with 1.7:1 regioselectivity.
In addition to poor selectivity, poor reaction rates were also
observed, in contrast to the reasonable rates reported for the
corresponding all-carbon congeners. We reasoned that employ-
a mixture of regioisomers at the 2- and 4-positions. We describe
herein a general solution to this problem that addresses both
reactivity and selectivity challenges by employing pyridine N-
oxides as substrates. This general and highly regioselective
Rh(III)-catalyzed annulation leads to naphthyridone derivatives
in high yields under mild conditions (Scheme 1).
We recently required efficient access to naphthyridinones such
7
as 1 (eq 1). Naphthyridinones are popular scaffolds in drug
Received: May 22, 2013
Published: September 10, 2013
8
discovery; thus, new methods for their synthesis are of interest.
©
2013 American Chemical Society
14492
dx.doi.org/10.1021/ja405140f | J. Am. Chem. Soc. 2013, 135, 14492−14495

Journal of the American Chemical Society
Communication
Chart 1. Rh(III)-Catalyzed Annulation of Nicotinamide
Derivatives
naphthyridinone N-oxide adduct 6a in 63% yield with ≥20:1
regioselectivity favoring the regioisomer depicted. Similar results
were obtained using an alkynyl coupling partner 4-octyne (Chart
1
, 5a).
Having established proof of concept, reaction optimization
was undertaken. Unlike norbornadiene or 4-octyne, unsym-
metrical coupling partners could lead to the formation of four
regioisomers arising from C−H activation at the 2- vs 4-positions
of the ring and from alkene/alkyne insertion into the Rh−C
bond (Figure 1).
To probe the global reaction selectivity while accessing value-
10
added products, nicotinamide N-oxide 7a and TES-acetylene
11
were chosen as substrates for optimization experiments. The
annulation took place efficiently at room temperature in the
presence of CsOAc and 2 mol% rhodium catalyst and tolerated a
number of different solvents with little to no impact on product
yield (Table 1, entries 1−7). The exceptions were polar, aprotic
solvents, which were not well tolerated and afforded the desired
product in depressed yields (entries 4 and 7). Similarly, the
nature of the counterion present in the exogenous base had little
effect on product yields, with Cs-, K-, Na-, and LiOAc all
affording similar results (entries 1, 8−10). Additional opti-
mization demonstrated that catalyst loading could be reduced to
General conditions: 1.1 equiv of norbornadiene, 2.0 equiv of CsOAc.
1
Regioselectivity determined by H NMR of crude reaction mixture.
a
b
Isolated, combined yields. Used 2.5 mol% [Cp*RhCl ] . Used 5 mol
2
2
c
%
[Cp*RhCl ] . Hydroxyl group was protected as TBS ether. Silyl
2 2
d
group was deprotected during reaction. Used 0.5 equiv of NaOAc
e
1
and 1 mol% [Cp*RhCl ] . Regioselectivity determined by H NMR
2
2
0
.5 mol% and base stoichiometry could be reduced to 0.5 equiv
entries 11 and 12).
of isolated product.
12
(
Reaction selectivity was uniformly high in all optimization
experiments, favoring 5b as the major product. The minor isomer
observed is the result of inverted alkyne insertion (e.g., 8b) and
not C−H activation at the 4-pyridyl carbon (e.g., 8d). The
methanol/sodium acetate system was selected as optimal based
on product yield, selectivity, and reagent handling. With
optimized reaction conditions in hand, the substrate scope was
examined.
A variety of substituents on the nicotinamide core were well
tolerated in the reaction (Chart 2). The electronic nature of a
substituent at the 5-position of the ring had little effect on the
outcome of the reaction, providing products in good yield and
Figure 1. Potential regioisomers arising from unsymmetrical alkenes/
alkynes.
Table 1. Reaction Optimization
≥
20:1 regioselectivity in all cases. Electron-withdrawing (5c−5f)
Chart 2. Effects of N-Oxide Ring Substituents
a
b
entry
base
solvent
yield (%)
regioisomeric ratio
1
2
3
4
5
6
7
8
9
CsOAc
CsOAc
CsOAc
CsOAc
CsOAc
CsOAc
CsOAc
KOAc
MeOH
89
85
88
32
84
85
24
85
89
84
85
87
23:1
31:1
19:1
NA
2-PrOH
CH CN
3
DMF
THF
NA
PhMe
NMP
47:1
23:1
23:1
23:1
23:1
24:1
23:1
MeOH
MeOH
MeOH
MeOH
MeOH
NaOAc
LiOAc
CsOAc
NaOAc
10
11
12
c
d
General conditions: 1.1 equiv of Si(Et) -acetylene, 2.0 equiv of base,
3
a
and 2 mol% [Cp*RhCl ] . Assay yields of major regioisomer
2
2
b
c
determined by HPLC. Based on HPLC area percent values. Used 0.5
mol% [Cp*RhCl ] . Used 0.5 equiv of NaOAc.
d
2
2
General conditions: 1.1 equiv of Si(Et) -acetylene, 0.5 equiv of
3
ing a nicotinamide N-oxide not only would result in increased
reaction rates due to its more electron rich nature but also should
favor the desired regioselectivity. To our delight, the N-oxide 7a
smoothly added across norbornadiene to furnish the dihydro-
1
NaOAc, and 1 mol% [Cp*RhCl ] . Regioselectivity determined by H
2
2
a
NMR of crude reaction mixtures. Yields are isolated. Reduced yields
reflect difficulty in isolation. Assay yields for compounds 5f and 5h
1
were 80 and 89%, respectively, as determined by H NMR.
1
4493
dx.doi.org/10.1021/ja405140f | J. Am. Chem. Soc. 2013, 135, 14492−14495

Journal of the American Chemical Society
Communication
Chart 3. Reaction with Alkynes
Chart 4. Reaction with Alkenes
General conditions: 1.5 equiv of alkene, 0.5 equiv of NaOAc, and 2.5
a
mol% [Cp*RhCl ] . Yields are isolated. Selectivity of C−H activation
2
2
1
at the 2- vs 4-position on the ring as determined by H NMR of
isolated product. Selectivity of alkene insertion into the Rh−C bond
as determined by H NMR of crude reaction mixture. Low yields
b
1
c
General conditions: 1.1 equiv of alkyne, 0.5 equiv of NaOAc, and 1
reflect difficulty in isolation. Assay yields for 6b and 6c were 84 and
mol% [Cp*RhCl ] . Selectivity of alkyne insertion into the Rh−C
2
2
1
1
82%, respectively, for the major regioisomers as determined by
NMR of crude reaction mixtures.
H
bond as determined by H NMR of crude reaction mixture. Yields are
a
b
isolated. Reaction at 50 °C. Selectivity of C−H activation at the 2-
c
d
vs 4-position on the ring. 1.5 equiv of alkyne. 2.5 mol%
Cp*RhCl ] . Reaction at 65 °C. Selectivity determined by
e
f
1
[
H
lactam nitrogen selectively. Like their alkyne counterparts,
regioisomeric impurities observed with unsymmetrical alkenes
were the result of inverted olefin insertion at C-2. For
symmetrical norbornadiene, the minor regioisomer arose from
C−H activation at C-4.
The resulting naphthyridinone N-oxides are smoothly reduced
to the parent naphthyridinone under mild conditions, as
exemplified by treatment of 5b with Zn powder and aqueous
2
2
NMR of isolated product.
and electron-donating substituents (5g and 5i) were tolerated
without impact to yields. A base-labile ester moiety (5e) was
tolerated as was a hydroxyl (5h) and acetamide (5f)
functionality. Substitution at the 6-position was also tolerated,
as exemplified in 5j, which gave 90% yield and ≥20:1 selectivity.
The reaction proved to be general for a variety of terminal and
internal alkynes to afford naphthyridinone N-oxides (Chart 3).
The regioselectivity of alkyne insertion was consistent with
previous studies; sterically demanding alkyne substituents were
positioned adjacent to the lactam nitrogen in the product. A
variety of functional groups on the alkynyl coupling partner were
tolerated, including strained rings, esters, and free hydroxyl
groups. Reactions with terminal alkyne coupling partners gave
yields of 45−99% and provided products in uniformly high
selectivity. Internal alkynes were also acceptable substrates. Both
dialkyl (5a) and alkyl-aryl (5o and 5p) alkynes cyclized in 80−
NH Cl in THF to afford naphthyridinone 9 in 75% yield (eq
4
14
2
). Importantly, the vinylsilane functional group was preserved
during the reduction, demonstrating the mild nature of the
reaction conditions.
To probe the mechanism of the C−H insertion reaction, an
experiment was performed in deuterated solvent. It was
anticipated that the reaction proceeds via a mechanism similar
to the related Rh(III)-mediated oxidative cycloaddition manifold
9
0% yield, with selectivities ranging from 13:1 to ≥20:1. In cases
2
where the alkyne possessed a vicinal sp -hybridized center, that
center was selectively installed adjacent to the lactam nitrogen in
the product (5n−5p). Elevated temperatures were required in
cases where either electron-deficient or sterically demanding
coupling partners were employed (5a and 5l−5n). At elevated
temperatures, we observed a small amount of C−H activation at
C-4 in the case of symmetrical alkyne 4-octyne (Chart 3, 5a),
while for unsymmetrical alkynes regioselectivity was dictated by
alkyne insertion into the Rh−C bond at C-2, suggesting an
interesting selectivity dichotomy based on reaction temperature.
Extending the reaction scope to include alkenes, other than
norbornadiene, would offer access to the useful dihydronaph-
thyridinone core structure. Rewardingly, alkenes performed well
in the reaction, contrary to related systems where alkenes have
been challenging substrates due to competing β-hydride
5b,d
developed by Fagnou and co-workers. Nicotinamide N-oxide
7a was subjected to the C−H insertion reaction conditions in
CD OD at 20 °C in the absence of alkyne [Scheme 2, (1)]. At the
3
2-position, 12% deuterium incorporation was observed while at
the 4-position <5% was observed. Interestingly 8% deuterium
incorporation was also observed at the 6-position. Performing
the same reaction in the presence of an alkyne coupling partner at
ambient temperature led to no deuterium incorporation in the
isolated product [Scheme 2, (2)]. These results suggest the
cyclometalation of 7a under the reaction conditions is slow
15
compared to the subsequent alkyne insertion.
The postulated mechanism that gives rise to the major
observed regioisomer commences with a preferred C−H
activation at the 2-position of the N-oxide ring of I to afford
metallocycle II (Scheme 2). Regioselective alkyne/alkene
insertion into the Rh−C bond would then furnish intermediate
III, followed by reductive elimination to give the naphthyr-
idinone N-oxide product IV with concomitant N−O bond
cleavage to re-oxidize the metal center. Based on the labeling
experiment in the presence of the alkyne, it is proposed that
13
elimination giving rise to Heck-type adducts. The conditions
were further optimized for alkenes; the reactions required a
slightly higher catalyst loading (2.5 mol%) in addition to heating
(
Chart 4). Both acyclic and strained cyclic alkenes participated,
with selectivity ranging from 4:1 using styrene (6b) to ≥20:1
with methyl acrylate and 2,3-dihydrofuran (6c and 6d). As with
alkynes, vicinal sp carbon centers were installed next to the
2
1
4494
dx.doi.org/10.1021/ja405140f | J. Am. Chem. Soc. 2013, 135, 14492−14495

Journal of the American Chemical Society
Communication
Scheme 2. Deuterium Labeling Studies and Postulated
Mechanism
REFERENCES
■
(
1) General overviews: (a) Bull, J. A.; Mousseau, J. J.; Guillaume, P.;
Charrette, A. B. Chem. Rev. 2012, 112, 2642. (b) Eicher, T.; Hauptmann,
S.; Speicher, A. The Chemistry of Heterocycles: Structures, Reactions,
Synthesis, and Applications, 3rd ed.; Wiley-VCH: Weinheim, Germany,
2
(
012.
2) Recent reviews: (a) Patureau, F. W.; Glorius, F. Angew. Chem., Int.
Ed. 2011, 50, 1977. (b) Yamaguchi, J.; Yamaguchi, A. D.; Itami, K.
Angew. Chem., Int. Ed. 2012, 51, 8960. (c) Colby, D. A.; Tsai, A. S.;
Bergman, R. G.; Ellman, J. A. Acc. Chem. Res. 2012, 43, 814.
(3) (a) Umeda, N.; Tsurgi, H.; Satoh, T.; Miura, M. Angew. Chem., Int.
Ed. 2008, 47, 4019. (b) Fukutani, T.; Umeda, M.; Hirano, K.; Satoh, T.;
Miura, M. Chem. Commun. 2009, 5141. (c) Stuart, D. R.; Bertrand-
Laperle, M.; Burgess, K. M. N.; Fagnou, K. J. Am. Chem. Soc. 2008, 130,
1
6474. (d) Colby, D. A.; Bergman, R. G.; Ellman, J. A. J. Am. Chem. Soc.
2
008, 130, 3645.
(
4) (a) Hyster, T. K.; Rovis, T. J. Am. Chem. Soc. 2010, 132, 10565.
b) Hyster, T. K.; Rovis, T. Chem. Sci. 2011, 2, 1606. (c) Guimond, N.;
Fagnou, K. J. Am. Chem. Soc. 2009, 131, 12050.
5) (a) Rakshit, S.; Grohmann, C.; Besset, T.; Glorius, F. J. Am. Chem.
Soc. 2011, 133, 2350. (b) Wang, H.; Glorius, F. Angew. Chem., Int. Ed.
012, 51, 7318. (c) Wang, H.; Grohmann, C.; Nimphius, C.; Glorius, F.
(
(
2
J. Am. Chem. Soc. 2012, 134, 19592. (d) Guimond, N.; Gouliaras, C.;
Fagnou, K. J. Am. Chem. Soc. 2010, 132, 6908. (e) Guimond, N.;
Gorelsky, S. I.; Fagnou, K. J. Am. Chem. Soc. 2011, 133, 6449. (f) Ye, B.;
Cramer, N. J. Am. Chem. Soc. 2013, 135, 636. (g) Xu, X.; Liu, Y.; Park, C.
M. Angew. Chem., Int. Ed. 2012, 51, 9372.
(6) Song, G.; Gong, X.; Li, X. J. Org. Chem. 2011, 76, 7583.
(7) Fang, Y.-Q.; Bio, M. M.; Hansen, K. B.; Potter, M. S.; Clausen, A. J.
Am. Chem. Soc. 2010, 132, 15525.
8) (a) Johnson, D. S.; Chung, C.; Fay, L. K.; Favor, D. A.; Repine, J. T.;
(
insertions at the 4- and 6-position are slower. To the extent they
occur under the reaction conditions, they appear to lead to
intermediates that are either slower in the insertion step (4-
position) or unproductive (6-position).
White, A. D.; Akunne, H. C.; Fitzgerald, L.; Nicholls, K.; Snyder, B. J.;
Whetzel, S. Z.; Zhang, L.; Serpa, K. A. Bioorg. Med. Chem. Lett. 2011, 21,
2
621. (b) Siu, T.; Kozina, E. S.; Jung, J.; Altman, M. D.; Rush, T.;
Dinsmore, C. J.; Young, J. R.; Rosenstein, C.; Chan, G.; Mathur, A.;
Bachman, E.; Mo, J.-R.; Xu, L.; Bouthillette, M.; Marshall, C. G. Bioorg.
Med. Chem. Lett. 2010, 20, 7421. (c) Li, Y.; Rossi, M. A.; Liang, J.; Siu, T.;
Hu, E.; Cosford, N.; Prasit, P.; Barnett, S. F.; Defeo-Jones, D.; Jones, R.
E.; Robinson, R. G.; Leander, K.; Huber, H. E.; Mittal, S. Bioorg. Med.
Chem. Lett. 2009, 19, 834.
(9) Pd-catalyzed C−H activations of pyridine N-oxides: (a) Cho, S. H.;
Hwang, S. J.; Chang, S. J. Am. Chem. Soc. 2008, 130, 9254. (b) Campeau,
L.-C.; Rousseaux, S.; Fagnou, K. J. Am. Chem. Soc. 2005, 127, 1820.
In conclusion, a Rh(III)-catalyzed C−H activation of
nicotinamide N-oxides in the presence of alkenes or alkynes to
afford naphthyridinone N-oxide or dihydronaphthyridinone N-
oxide products has been realized. The reaction proceeds with
nearly perfect regioselectivity on the pyridine and high
16
regioselectivity for the olefin insertion. Furthermore, the
reaction proceeds under mild conditions with relatively low
catalyst loadings (1 mol% for alkyne, 2.5 mol% for alkenes).
Additional mechanistic and computational studies to fully
rationalize the observed products are planned and will be
reported in due course.
(
8
c) Kanyiva, K. S.; Nako, Y.; Hiyama, T. Angew. Chem., Int. Ed. 2007, 46,
872.
10) The use of terminal alkynes is known to be problematic in similar
(
systems that employ external oxidants (typically Cu(II)) due to alkyne
dimerization (Glaser coupling).
(
11) This would allow access to the desired naphthyridinone motif
ASSOCIATED CONTENT
Supporting Information
Experimental details and characterization data. This material is
available free of charge via the Internet at http://pubs.acs.org.
■
upon protonolysis of the silyl group and reduction of the pyridine N-
oxide.
(12) The (CH CN) Cp*Rh(III)SbF catalyst was equally effective at
promoting the transformation as [Cp*Rh(III)Cl ] at the same catalyst
loading.
*
S
3
3
6
2 2
AUTHOR INFORMATION
(13) (a) Umeda, N.; Hirano, K.; Satoh, T.; Miura, M. J. Org. Chem.
■
2
009, 74, 7094. (b) Mochida, S.; Hirano, K.; Satoh, T.; Miura, M. Org.
Corresponding Authors
jhuckins@amgen.com
ebercot@suterra.com
Lett. 2010, 12, 5776. (c) Ueura, K.; Satoh, T.; Miura, M. Org. Lett. 2007,
9
, 1407. (d) Patureau, F. W.; Glorius, F. J. Am. Chem. Soc. 2010, 132,
9
982. (e) Patureau, F. W.; Besset, T. Angew. Chem., Int. Ed. 2010, 50,
Present Address
E.A.B.: Suterra, LLC, 20950 Talus Place, Bend, OR 97701
Notes
1064. (f) Tsai, A. S.; Brasse, M.; Bergman, R. G.; Ellman, J. A. Org. Lett.
2011, 13, 540.
†
(14) Aoyagi, Y.; Abe, T.; Ohta, A. Synthesis 1997, 891.
(15) The presence of different reaction pathways in the absence and
The authors declare no competing financial interest.
presence of coupling partner cannot be ruled out and is the subject of
further studies.
ACKNOWLEDGMENTS
■
(16) After completion of this research, a single example of
regioselective functionalization of a nicotinamide was reported in
closely related chemistry: Presset, M.; Oehlreich, D.; Rombouts, F.;
Molander, G. A. Org. Lett. 2013, 15, 1528.
We gratefully acknowledge Dr. Jacqueline E. Milne for useful
discussions and Dr. Chul Yoo for assistance with obtaining
HRMS data.
1
4495
dx.doi.org/10.1021/ja405140f | J. Am. Chem. Soc. 2013, 135, 14492−14495