Manganese-catalyzed late-stage aliphatic C-H azidation

Article abstract of DOI:10.1021/jacs.5b01983

We report a manganese-catalyzed aliphatic C-H azidation reaction that can efficiently convert secondary, tertiary, and benzylic C-H bonds to the corresponding azides. The method utilizes aqueous sodium azide solution as the azide source and can be performed under air. Besides its operational simplicity, the potential of this method for late-stage functionalization has been demonstrated by successful azidation of various bioactive molecules with yields up to 74%, including the important drugs pregabalin, memantine, and the antimalarial artemisinin. Azidation of celestolide with a chiral manganese salen catalyst afforded the azide product in 70% ee, representing a Mn-catalyzed enantioselective aliphatic C-H azidation reaction. Considering the versatile roles of organic azides in modern chemistry and the ubiquity of aliphatic C-H bonds in organic molecules, we envision that this Mn-azidation method will find wide application in organic synthesis, drug discovery, and chemical biology.

Full text of DOI:10.1021/jacs.5b01983

Communication
pubs.acs.org/JACS
Manganese-Catalyzed Late-Stage Aliphatic C−H Azidation
Xiongyi Huang, Tova M. Bergsten, and John T. Groves*
Department of Chemistry, Princeton University, Princeton, New Jersey 08544 United States
S
* Supporting Information
hydrocarbons, primarily due to the harsh reaction conditions
and/or instability.
ABSTRACT: We report a manganese-catalyzed aliphatic
C−H azidation reaction that can efficiently convert
secondary, tertiary, and benzylic C−H bonds to the
corresponding azides. The method utilizes aqueous
sodium azide solution as the azide source and can be
performed under air. Besides its operational simplicity, the
potential of this method for late-stage functionalization has
been demonstrated by successful azidation of various
bioactive molecules with yields up to 74%, including the
important drugs pregabalin, memantine, and the anti-
malarial artemisinin. Azidation of celestolide with a chiral
manganese salen catalyst afforded the azide product in
70% ee, representing a Mn-catalyzed enantioselective
aliphatic C−H azidation reaction. Considering the versatile
roles of organic azides in modern chemistry and the
ubiquity of aliphatic C−H bonds in organic molecules, we
envision that this Mn-azidation method will find wide
application in organic synthesis, drug discovery, and
chemical biology.
Jiang et al. have reported an allylic, Pd-catalyzed C−H
azidation method with NaN3 as the azide source.8 Gade et al.
developed an enantioselective C−H azidation reaction of β-keto
esters with an iron boxim catalyst and an azidoiodinane.9
Bollinger et al. devised a variant of the SyrB2 halogenase enzyme
that is capable of performing C−H azidation reaction on its
preferred substrate.10 The former two methods are restricted to
allylic C−H bonds and β-keto ester α-positions, respectively.
The enzymatic approach requires substrates to be bound to the
carrier protein. Very recently, Hartwig et al. reported an elegant
protocol for late-stage azidation of tertiary and benzylic C−H
bonds using an iron catalyst and an azidoiodinane reagent.11
We report here a practical and complementary manganese-
catalyzed C−H azidation reaction that is applicable to secondary,
tertiary, and benzylic C−H bonds. The method uses easily
handled aqueous sodium azide as the azide source. By harnessing
the ubiquitous C−H bonds in organic molecules and the
synthetic versatility of alkyl azides, this reaction provides a
powerful chemical tool for late-stage diversification of drug-like
molecules as well as for developing new biochemical reporters.
Conceptually, this azidation reaction derives from mechanistic
analyses of the manganese-catalyzed C−H fluorination reported
recently by our group (Scheme 1).12 For C−H fluorination, the
rganic azides are of significant importance in organic
O
synthesis, chemical biology, pharmaceutical discovery, and
materials science. These versatile synthetic intermediates provide
convenient access to a variety of functionalities such as amines,
imines, amides, and triazoles.1 Additionally, organic azides
perform irreplaceable roles in chemical biology and drug
discovery, due to the broad applications of azide−alkyne Huisgen
cycloaddition and Staudinger ligation in “click” chemistry.2 In
materials science, azide-based transformations are widely used
for surface modification, macromolecular engineering, and
synthesis of novel polymeric materials.3
Scheme 1. Concept of Mn-Catalyzed C−H Azidation
Since the preparation of the first organic azide, phenyl azide, in
1864, numerous azidation reactions have been developed,
enabling the synthesis of organic azides from a variety of
functionalities, including alkyl halides, amines, hydrazines,
triazenes, alcohols, epoxides, aziridines, alkenes, diazonium
ions, aldehydes, etc.1 These advances have significantly expanded
the synthetic availability of this functional group. More recently,
direct aryl C−H azidation has been realized using transition-
metal catalysis (Cu, Pd or Rh)4 or with hypervalent iodine
reagents through a Friedel−Crafts reaction process.5 Moreover,
recent developments in the field of catalytic amination and
amidation have provided a plethora of methods for constructing
aliphatic C−N bond through C−H activation.6 In contrast, direct
aliphatic C−H azidation reactions are noticeably lacking.
Although hypervalent iodine reagents like IN3 have been
known to perform direct aliphatic C−H azidation,7 the
application of this reagent has been limited to simple
presence of fluoride as the axial ligand to manganese facilitates
the fluorine transfer to incipient substrate radicals formed after a
manganese(V)oxo-induced hydrogen abstraction step. Diversion
of the reaction to afford the fluorination product rather than the
usual hydroxylation product is made possible by the unusually
Received: February 23, 2015
© XXXX American Chemical Society
A
DOI: 10.1021/jacs.5b01983
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
low barrier to fluorine atom transfer from manganese(IV)-F
species.12a,13 Early results by Hill et al. and by us showed that
manganese porphyrins were capable of mediating low-con-
version C−H azidation of simple hydrocarbons along with
oxygenation products.13b,14 We considered that replacing
fluoride with a suitable azide source might promote the more
efficient formation of an analogous MnIV−N3 intermediate,
which would trap the substrate radical and form the desired C−
N3 bond.
Manganese-catalyzed azidation screening using exploratory
conditions are shown in Table 1. A modest 3% yield of azidation
Table 1. Reaction Optimization with Cyclooctane
catalyst
NaN3 (M)
solvent
yield (%) Az/Ox
1
2
Mn(TPP)Cl
Mn(TMP)Cl
Mn(TDCPP)Cl
Mn(TMP)Cl
Mn(TMP)Cl
Mn(TMP)Cl
Mn(TMP)Cl
Mn(TMP)Cl
Mn(TMP)Cl
Mn(salen)Cl
Mn(TMP)Cl
6.3
6.3
6.3
6.3
6.3
6.3
6.3
1.5
1.5
1.5
1.5
CH2Cl2
CH2Cl2
CH2Cl2
PhCF3
acetone
benzene
EtOAc
EtOAc
EtOAc
EtOAc
EtOAc
3
26
18
20
3
1:1
5:1
3
2:1
4
1.5:1
1:2
5
6
20
25
29
55
51
53
2:1
Figure 1. Substrate scope of Mn-catalyzed C−H azidation. (a)
Conditions: catalyst (method A: 1.5 mol % Mn(TMP)Cl; method B:
5 mol % Mn(salen)Cl), substrate (0.6 mmol), 1.5 mL of NaN3 (aq., 1.5
M, 4 equiv), 1 mL of EtOAc, 23 °C. PhIO (3−6 equiv) was added in
portions (0.8−1.0 equiv PhIO each portion). The reaction was
monitored by TLC, GC-MS or LC-MS. Isolated yields of major
products are reported unless notified otherwise. A and B superscripts
refer to methods A and B, respectively. Azide to oxygenated product
ratios were 2−4:1. (b) Yields were determined relative to starting
material by GC-MS. (c) Methyl acetate was used as solvent.
7
4:1
8
5.3:1
3.9:1
3.7:1
3.6:1
b
b
b
9
,
,
c
10
11
d
a
Yields, based on starting substrate, and product distributions were
b
c
determined by GC-MS. 4 equiv of PhIO were added sequentially. 5
mol % catalyst loading. Reaction was carried out under air.
d
product with 1:1 azidation/oxygenation ratio (abbreviated as
Az/Ox) was obtained with Mn(TPP)Cl as catalyst (Table 1,
entry 1). The yield surged to 26% (Az/Ox = 5) by changing
catalyst from Mn(TPP)Cl to Mn(TMP)Cl. Since CH2Cl2
solvent is not desirable for azidation due to the potential
formation of explosive diazidomethane,15 various solvents were
screened and ethyl acetate was found to be optimal, affording a
25% yield (Az/Ox = 4). The effect of aqueous phase azide
concentration on reaction yield and selectivity was investigated.
Interestingly, decreasing the azide concentration from 6.3 to 1.5
M led to a slightly higher yield (29%, Az/Ox = 5.3). Serial
addition of more oxidant further increased the yield. With 4 equiv
of PhIO, the yield was increased to 55% (Az/Ox = 3.9) (Table 1,
entry 9). Besides manganese porphyrins, manganese salen-type
Jacobsen catalysts16 were also found to be reactive for this C−H
azidation reaction, albeit with higher catalyst loading (Table 1,
entry 10). While most of the screening was performed under
nitrogen atmosphere, carrying out the reaction under air led to
only slightly decreased conversion and azidation selectivity
(Table 1, entry 11), greatly simplifying the reaction setup. No
azidation products were detected in control experiments in
which the manganese catalysts were omitted.
transfer catalyst. Functionalities including amide, ester, ketone,
carbamate, tertiary alcohol, terminal alkyne, and heteroaryl
groups including pyridine, thiophene, and pyrimidine were well
tolerated. Secondary alcohols were oxidized to ketones and
cyclohexene afforded a 1:1 ratio of allylic azide and epoxide. The
regio-selectivity of C−H activation could be regulated by the
electronic and steric properties of the ligand, as we have
previously shown for C−H halogenations.12a,13c For example,
azidation of adamantane with Mn(TMP)Cl led to a 1.5:1 ratio of
tertiary to secondary products compared to 4:1 tertiary/
secondary ratio with the manganese salen catalyst (Figure 1,
11). The reaction of trans-decalin with the bulky Mn(TMP)Cl
catalyst afforded mostly secondary azidation products with a 3.2-
to-1 ratio of C3 over C2 (Figure 1, 2). The efficiency of C−H
activation was significantly affected by the electronic properties
of the substrate. While a 45% yield (based on the substrate) was
observed for azidation of cyclohexane, the more electron-
deficient cyclohexanone only yielded 2% azidation product
under the same conditions. The potential synthetic application of
this Mn-azide method was demonstrated by the successful
azidation of common structural motifs in bioactive molecules
such as tetrahydronaphthalene, dibenzocycloheptane, and
adamantane.
The optimized reaction conditions were established based on
these reaction screening results (Figure 1). A wide range of
substrates was efficiently converted to the corresponding azides.
Reaction times were generally 3−12 h with no need of a phase-
With these promising results in hand, we set out to apply the
method to more complex bioactive molecules (or their protected
B
DOI: 10.1021/jacs.5b01983
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
analogs), encompassing the important drugs memantine
(Namenda) and pregabalin (Lyrica), the Parkinson’s disease
drug rasagiline (Azilect), the opium alkaloid antispasmodic drug
papaverine (Pavabid), ibuprofen, the terpenoid sclareolide, the
antimalarial artemisinin, and the estrogenic hormone estrone.
Gratifyingly, subjecting these molecules to the Mn-azide method
afforded corresponding azides with yield up to 74% yield.
Pregabalin was selectively functionalized at the C−H bond
adjacent to the NHBoc protecting group, as a result of the
radical-stabilizing ability and weak electron-withdrawing prop-
erty of the carbamate nitrogen (Figure 2, 24). Sclareolide
Figure 2. Late-stage azidation of bioactive molecules. A and B
superscripts refer to methods A and B, respectively. Azide to oxygenated
product ratios were 2−4:1. (a) Manganese salen catalyst was used. (b)
Methyl acetate was used as solvent.
Figure 3. (a) KIE studies of Mn-catalyzed C−H azidation. (b) Azidation
with norcarane substrate. (c) Enantioselective C−H azidation with a
chiral manganese salen catalyst. (d) DFT calculations of the azide-
transfer step (kcal/mol at 298 K). 3Nquintet refers to azide transfer
through nitrogen N3 on quintet surface and other pathways are labeled
accordingly.
afforded C2 azidation as the major product in 57% yield and a
high α/β ratio of 7.5 (Figure 2, 25). Although estrone acetate has
several possible reactive sites, the C9-α azide was obtained as the
major product in 38% yield. Interestingly, the major side product
was determined to be a diazidation product with both C9 and C6
being activated (Figure 2, 27). The method is also applicable to
fragile molecules like artemisinin. With the manganese salen
catalyst, the C10 tertiary azide was found to be the exclusive
azidation product (Figure 2, 29). These results highlight the
enabling power of this method to late-stage functionalization.
We conducted a number of experiments to elucidate the
reaction mechanism. Stirring the Mn(TMP)Cl ethyl acetate
solution with an aqueous solution of NaN3 led to facile ligand
exchange to form Mn(TMP)N3, as observed by UV−vis
spectroscopy (479−485 nm, Figure S1). The dependence of
the regioselectivity on the catalyst ligand architecture suggests a
catalyst-based intermediate for the hydrogen abstraction step,
presumably an oxomanganese(V) species, rather than the azide
radical observed for C−H activation with azidoiodinane reagents.
This conclusion was further supported by measurement of the
deuterium kinetic isotope effect (KIE). Reaction of a 1:1 mixture
of toluene and toluene-d8 produced competitive intermolecular
KIEs of 7.8 and 9.3, respectively, for Mn(TMP)Cl and
Mn(salen)Cl (Figure 3a). The values are significantly larger
than the KIE value observed for hydrogen abstraction by the
azide radical (∼5).11,17 The radical nature of the reaction was
demonstrated by using norcarane as a probe substrate (Figure
3b).18 The ratios of rearranged product, 3-azidomethylcyclohex-
ane (34) to unrearranged product, 2-azidonorcaranes (35) are
7.2 and 5.2 for Mn(TMP)Cl and Mn(salen)Cl, respectively.
Given that ring-opening rate constant of 2-norcaranyl radical is 2
× 108 M−1 s−1, the obtained ratios indicate rather long radical
lifetimes of 36 and 26 ns.
To study the azido-transfer step, the chiral manganese salen
Jacobsen catalyst was used.16 Surprisingly, azidation of
celestolide with this catalyst afforded a 63% ee at 25 °C and
70% ee at 4 °C (Figure 3c), representing a Mn-catalyzed
enantioselective aliphatic C−H azidation. The enantioselectivity
was reversed when changing from (R,R)-Mn(salen1)Cl to (S,S)-
Mn(salen1)Cl. This high ee clearly demonstrates the involve-
ment of a manganese-bound azide intermediate in the azido-
transfer step.
DFT calculations with an unsubstituted azidomanganese(IV)
salen model and a tolyl radical were employed to probe the facile
azide ligand transfer from MnIV−N3 to the substrate radicals
(B3LYP/6-31G(d) (SDD for Mn), see computational details in
the SI). The calculated bond lengths for MnIV−N3 and the Mn−
N1−N2 bond angle were very similar to the previous reported X-
ray crystal structures of MnIV−N3 salen complexes, 1.99 Å and
120.7°, respectively (Figure 3d).16 We considered four
reasonable pathways by which a substrate radical could interact
C
DOI: 10.1021/jacs.5b01983
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
with MnIV−N3: approach to either the manganese-bound (N1)
or terminal (N3) azido-nitrogen on either the triplet or quintet
energy surfaces. All of these trajectories had low barriers, with the
lowest energy transition state obtained for azido-transfer through
N3 on the quintet surface, which had an energy barrier of only 7.0
kcal/mol. The azido-transfer transition state at the N1 nitrogen
was found just above at 8.3 kcal/mol on the triplet energy surface.
Closer inspection of the structures of these transition states
revealed that the benzylic carbon of the tolyl radical is in close
proximity to the aryl group of the salen ligand in both structures
(3.27 and 3.10 Å), consistent with the observed enantioselec-
tivity. Based on these mechanistic considerations, we propose a
reaction mechanism similar to that of Mn−F C-H fluorination,12a
wherein PhIO first oxidizes the resting Mn(III) catalyst to the
hydrogen-abstracting oxoMn(V) intermediate (Figure 4). The
ACKNOWLEDGMENTS
■
This research was supported by the US National Science
Foundation award CHE-1148597. X.H. thanks the Howard
Hughes Medical Institute for fellowship support. The authors
thank Prof. A. G. Doyle for access to chiral HPLC.
REFERENCES
■
(1) (a) Brase, S.; Gil, C.; Knepper, K.; Zimmermann, V. Angew. Chem.,
Int. Ed. 2005, 44, 5188. (b) Organic Azides: Syntheses and Applications;
John Wiley & Sons, Ltd: Chichester, U.K., 2010.
(2) (a) Thirumurugan, P.; Matosiuk, D.; Jozwiak, K. Chem. Rev. 2013,
113, 4905. (b) Grammel, M.; Hang, H. C. Nat. Chem. Biol. 2014, 10, 239.
(3) (a) Binder, W. H.; Kluger, C. Curr. Org. Chem. 2006, 10, 1791.
(b) Xi, W.; Scott, T. F.; Kloxin, C. J.; Bowman, C. N. Adv. Funct. Mater.
2014, 24, 2572.
(4) (a) Tang, C.; Jiao, N. J. Am. Chem. Soc. 2012, 134, 18924. (b) Xie,
F.; Qi, Z.; Li, X. Angew. Chem., Int. Ed. 2013, 52, 11862. (c) Zheng, Q.-Z.;
Feng, P.; Liang, Y.-F.; Jiao, N. Org. Lett. 2013, 15, 4262. (d) Fan, Y.;
Wan, W.; Ma, G.; Gao, W.; Jiang, H.; Zhu, S.; Hao, J. Chem. Commun.
2014, 50, 5733. (e) Yao, B.; Liu, Y.; Zhao, L.; Wang, D.-X.; Wang, M.-X.
J. Org. Chem. 2014, 79, 11139.
(5) (a) Lubriks, D.; Sokolovs, I.; Suna, E. J. Am. Chem. Soc. 2012, 134,
15436. (b) Telvekar, V. N.; Sasane, K. A. Synth. Commun. 2012, 42,
1085.
(6) (a) Lescot, C.; Darses, B.; Collet, F.; Retailleau, P.; Dauban, P. J.
Org. Chem. 2012, 77, 7232. (b) Li, J.; Cisar, J. S.; Zhou, C.-Y.; Vera, B.;
Williams, H.; Rodriguez, A. D.; Cravatt, B. F.; Romo, D. Nat. Chem.
2013, 5, 510. (c) Yu, X. Q.; Huang, J. S.; Zhou, X. G.; Che, C. M. Org.
Lett. 2000, 2, 2233. (d) Espino, C. G.; Fiori, K. W.; Kim, M.; Du Bois, J. J.
Am. Chem. Soc. 2004, 126, 15378. (e) Davies, H. M. L.; Long, M. S.
Angew. Chem., Int. Ed. 2005, 44, 3518. (f) Wasa, M.; Yu, J.-Q. J. Am.
Chem. Soc. 2008, 130, 14058. (g) Ochiai, M.; Miyamoto, K.; Kaneaki, T.;
Hayashi, S.; Nakanishi, W. Science 2011, 332, 448. (h) Che, C.-M.; Lo, V.
K.-Y.; Zhou, C.-Y.; Huang, J.-S. Chem. Soc. Rev. 2011, 40, 1950.
(i) Roizen, J. L.; Harvey, M. E.; Du Bois, J. Acc. Chem. Res. 2012, 45, 911.
(j) Paradine, S. M.; White, M. C. J. Am. Chem. Soc. 2012, 134, 2036.
(k) Michaudel, Q.; Thevenet, D.; Baran, P. S. J. Am. Chem. Soc. 2012,
134, 2547. (l) Dequirez, G.; Pons, V.; Dauban, P. Angew. Chem., Int. Ed.
2012, 51, 7384. (m) Jeffrey, J. L.; Sarpong, R. Chem. Sci. 2013, 4, 4092.
(n) Hennessy, E. T.; Betley, T. A. Science 2013, 340, 591. (o) Louillat,
M.-L.; Patureau, F. W. Chem. Soc. Rev. 2014, 43, 901.
(7) (a) Zhdankin, V. V.; Krasutsky, A. P.; Kuehl, C. J.; Simonsen, A. J.;
Woodward, J. K.; Mismash, B.; Bolz, J. T. J. Am. Chem. Soc. 1996, 118,
5192. (b) Viuf, C.; Bols, M. Angew. Chem., Int. Ed. 2001, 40, 623.
(8) Chen, H.; Yang, W.; Wu, W.; Jiang, H. Org. Biomol. Chem. 2014, 12,
3340.
(9) Deng, Q.-H.; Bleith, T.; Wadepohl, H.; Gade, L. H. J. Am. Chem.
Soc. 2013, 135, 5356.
(10) Matthews, M. L.; Chang, W.-c.; Layne, A. P.; Miles, L. A.; Krebs,
C.; Bollinger, J. M., Jr. Nat. Chem. Biol. 2014, 10, 209.
Figure 4. Proposed Mn-azide reaction mechanism.
substrate radical formed after hydrogen abstraction is then
captured by MnIV−N3 intermediate to form C−N3 bond and
regenerate the catalyst in accordance with the energy landscape
in Figure 3d.
In conclusion, we have developed a facile and convenient
manganese-catalyzed aliphatic C−H azidation reaction. The
reported reaction uses easily handled aqueous NaN3 solutions as
the azide source and is operationally simple. The potential of this
method for applications in organic synthesis, chemical biology
and drug discovery has been demonstrated by the successful late
stage azidation of several bioactive molecules. With an initial high
enatioselectivity with chiral manganese salen catalyst, we are now
working to further improve the enantioselectivity of this reaction
and expanding its substrate scope. Furthermore, efforts are being
undertaken to extend the concept illustrated in this paper to the
installation of other pseudohalogen functional groups to
molecules via manganese-catalyzed C−H activation.
(11) Sharma, A.; Hartwig, J. F. Nature 2015, 517, 600.
(12) (a) Liu, W.; Huang, X.; Cheng, M.-J.; Nielsen, R. J.; Goddard, W.
A., III; Groves, J. T. Science 2012, 337, 1322. (b) Huang, X.; Liu, W.;
Ren, H.; Neelamegam, R.; Hooker, J. M.; Groves, J. T. J. Am. Chem. Soc.
2014, 136, 6842.
(13) (a) Groves, J. T.; Kruper, W. J.; Haushalter, R. C. J. Am. Chem. Soc.
1980, 102, 6375. (b) Kruper, W. J., Jr. Ph.D. Dissertation, University of
Michigan, 1982. (c) Liu, W.; Groves, J. T. J. Am. Chem. Soc. 2010, 132,
12847.
(14) Hill, C. L.; Smegal, J. A.; Henly, T. J. J. Org. Chem. 1983, 48, 3277.
(15) Conrow, R. E.; Dean, W. D. Org. Process Res. Dev. 2008, 12, 1285.
(16) (a) Kurahashi, T.; Fujii, H. Inorg. Chem. 2008, 47, 7556.
(b) Kurahashi, T.; Hada, M.; Fujii, H. J. Am. Chem. Soc. 2009, 131,
12394.
ASSOCIATED CONTENT
■
S
* Supporting Information
Detailed experimental procedures, spectroscopic data for all new
compounds, and details for the DFT calculations. This material is
available free of charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Author
(17) Pedersen, C. M.; Marinescu, L. G.; Bols, M. Org. Biomol. Chem.
2005, 3, 816.
*jtgroves@princeton.edu
(18) Groves, J. T. J. Inorg. Biochem. 2006, 100, 434.
Notes
The authors declare no competing financial interest.
D
DOI: 10.1021/jacs.5b01983
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX