Radical C-H Fluorination Using Unprotected Amino Acids as Radical Precursors

Article abstract of DOI:10.1021/acs.orglett.7b01188

We report a unique example of utilizing unprotected amino acids for benzylic C-H fluorination via a radical process. α-Aminoalkyl radicals are readily generated via oxidative decarboxylation of unprotected amino acids using a simple silver(I) catalyst and Selectfluor, which serves as both a mild oxidant and source of electrophilic fluorine. Mechanistic investigation shows that coordination of the unprotected amino acid plays a crucial role in lowering the oxidation potential of Ag(I), enabling oxidation under mild conditions. Mono- or difluorination is possible by controlling the stoichiometry of amino acid and fluorine source.

Full text of DOI:10.1021/acs.orglett.7b01188

Letter
pubs.acs.org/OrgLett
Radical C−H Fluorination Using Unprotected Amino Acids as Radical
Precursors
Alyssa M. Hua, Duy N. Mai, Ramon Martinez, and Ryan D. Baxter*
Department of Chemistry and Chemical Biology, University of California, 5200 North Lake Road, Merced, California 95343, United
States
*
S Supporting Information
ABSTRACT: We report a unique example of utilizing unprotected amino acids for benzylic C−H fluorination via a radical
process. α-Aminoalkyl radicals are readily generated via oxidative decarboxylation of unprotected amino acids using a simple
silver(I) catalyst and Selectfluor, which serves as both a mild oxidant and source of electrophilic fluorine. Mechanistic
investigation shows that coordination of the unprotected amino acid plays a crucial role in lowering the oxidation potential of
Ag(I), enabling oxidation under mild conditions. Mono- or difluorination is possible by controlling the stoichiometry of amino
acid and fluorine source.
α-Aminoalkyl radicals are important and highly reactive
intermediates for the synthesis of substituted amines. Common
strategies for generating α-aminoalkyl radicals involve direct
oxidation of an amine to an N-centered radical cation via UV light
Scheme 1. α-Aminoalkyl Radicals from Amino Acids
1
or photocatalyst, followed by α-deprotonation. However, site-
specific formation of the α-aminoalkyl radical can be problematic
withsubstratescontainingmultiplesitesofdeprotonation. Recent
work has shown that radical decarboxylation of amino acids is an
attractive, regioselective alternative (Scheme 1). Because α-
aminoalkyl radicals are highly reducing and prone to oxidation,
their synthetic utility depends on substitution at nitrogen to
2
attenuate stability and reactivity. Strong chemical oxidants and
light-mediated photocatalysis have both been shown to be
effective for the radical decarboxylation of N-substituted amino
acids. The resulting α-aminoalkyl radicals have been shown to
participate in a range of C−C bond-forming reactions such as
conjugate additions, cross-coupling, and Minisci reactions with
3
−5
electron-deficient heteroarenes (Scheme 1a).
To the best of
our knowledge, unprotected α-aminoalkyl radicals have not been
successfully used as synthetic intermediates due to their
propensity for rapid oxidation. Our interest in developing new
and mild methods for C−H radical functionalizations led us to
question whether it was possible to use unprotected α-aminoalkyl
radicals as useful synthetic intermediates. Herein we report a new
strategy for radical C−H fluorination using unprotected α-amino
acids as stable radical precursors. Our mild protocol enables
radical fluorination in the presence of sensitive functional groups
and is effective under mild reaction conditions.
decarboxylation, oxidation, and iminium hydrolysis (Scheme
6
1
b). Our laboratory has recently utilized this transformation as a
meanstogenerateandusealdehydesasalkylradicalprecursorsfor
7
C−H heterocycle functionalization. During our studies, we
contemplated the fate of the α-aminoalkyl radical species formed
immediately after amino acid decarboxylation. We hypothesized
The oxidative decomposition of α-amino acids by strong
oxidants has been known for more than eight decades. The
Streckerdegradationproducesaldehydesviaasequenceinvolving
Received: April 19, 2017
©
XXXX American Chemical Society
A
DOI: 10.1021/acs.orglett.7b01188
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
a
that if a mild oxidant could be employed to facilitate radical
decarboxylation, this would diminish the prevalence for α-
aminoalkyl radical oxidation, allowing alternative radical
termination pathways to proceed.
Table 1. Optimization of Benzylic Fluorination
Based on analysis of bond dissociation energies, we believed
that α-aminoalkyl radicals should be capable of hydrogen atom
3
8
entry
deviation from standard conditions
yield (%)
abstraction from benzylic sp C−H bonds. Hydrogen atom
transfer (HAT) strategies have been used to generate α-oxy, α-
amino, or α-aryl radicals from metal complexes, but the use of
1
2
3
4
5
6
7
8
9
none
76
46
48
62
72
59
0
50 °C
open to air
9
,10
carbon-centered radicals as HAT agents is less common.
proline instead of glycine
valine instead of glycine
phenylalanine instead of glycine
N-Boc-glycine instead of glycine
mandelic acid instead of glycine
NFSI as fluorine source
CH CI /H O
Because electrophilic sources of fluorine can serve as both oxidant
and a source of atomic fluorine, we sought to establish the
feasibility of using α-aminoalkyl radicals as HAT reagents by
developing a new protocol for C−H fluorination (Scheme
1
1,12
0
1c).
0
At the outset of our studies, we considered two possible
mechanistic scenarios for reaction initiation (Scheme 2). Direct
10
11
12
13
14
15
0
2
2
2
acetone/H O
48
0
2
20 mol % Ag(Bpy) OTf
2
Scheme 2. Mechanistic Scenarios for Ag(I) Oxidation
20 mol % Ag(Phen) OTf
0
2
no AgNO3
no glycine
0
0
a
Yields refer to chromatographically pure material. Reaction
conditions: 4-methylbiphenyl (2) (0.2 mmol), glycine (0.4 mmol),
Selectfluor (0.4 mmol), AgNO (0.04 mmol), 2 mL of CH CN/H O
3
3
2
(1:1).
o
16
V) compared to Selectfluor (E = −0.04 V) (entry 9).
addition of Ag(I) into the N−F bond of Selectfluor, as originally
proposed by Li, would promote fluorination via a Ag(I)/(III)/
Alternative solvents led to diminished reaction conversions
(entries 10−11). Other silver(I) catalysts, such as Ag(Bpy) OTf
2
1
3
(
II) catalytic cycle (eq 1). Alternatively, single-electron
or Ag(Phen) OTf, did not promote radical fluorination (entries
2
oxidation of Ag(I) by Selectfluor, as suggested by Flowers,
12−13). Control reactions showed that both a silver catalyst and
an amino acid were necessary for reaction (entries 14−15).
With optimized experimental conditions established, the scope
of radical fluorination was explored. As shown in Scheme 3, good
yields and functional group tolerance were observed. In most
cases, unreacted starting material accounted for the mass balance
of incomplete reactions. Mono- or difluorinated products may be
accessed for select electron-rich substrates by controlling reaction
stoichiometry (2a, 2b, both from 4-methylbiphenyl; 3a, 3b, both
from 4-ethylbiphenyl). Although secondary sites can be suitably
fluorinated (5a, 14a), primary benzylic positions were shown to
be the most reactive under our conditions. Tertiary benzylic sites
were not reactive, yielding only trace product (6a). Analogous
selectivity (1° > 2° > 3°) has been shown in previous radical
fluorinations and may be the result of a proton-coupled electron
would facilitate decarboxylation via aAg(I)/(II) catalyst cycle (eq
1
4
2
). This second scenario could be complicated by the
stoichiometric formation of radical dication 1 from Selectfluor,
potentially obscuring the identity of the dominant HAT agent.
Lectka has previously implicated dication 1 in HAT processes
involving radical chain reactions for C−H fluorinations involving
1
5
Selectfluor. Studies involving ReactIR, analytical electro-
chemistry, and deuterium labeling have led us to support a
Ag(I)/(II) cycle involving HAT predominately from an α-
aminoalkyl radical (vide infra).
As shown in Table 1, 4-methylbiphenyl (2) was chosen as a
model system to develop radical C−H fluorinations using amino
acids. Through our optimization, we discovered that catalytic
AgNO , Selectfluor, glycine, anda1:1mixtureofCH CN/H Oat
3
3
2
12c
slightlyelevatedtemperaturesledtoformationof2aingoodyield.
The reaction was shown to be ineffective at room temperature,
and higher temperatures led to diminished conversion (entry 2).
The reaction may be performed open to air (entry 3), although
higher conversion was observed under an inert atmosphere.
Several unprotected amino acids were capable of benzylic
fluorination (entries 4−6) although competing oxidation was
observed in some cases, and glycine generally led to the highest
conversion across all substrates examined. Interestingly, N-
protected amino acids, such as N-Boc-glycine, did not promote
radical fluorination (entry 7), although these results were later
rationalized through mechanistic investigation (vide infra).
Carboxylic acids were not effective at promoting benzylic
fluorination, suggesting the presence of nitrogen plays a key
role in the reaction (entry 8). N-Fluorobenzenesulfonamide
transfer mechanism that is sensitive to stereoelectronic effects.
Both electron-rich (2a−8a) and electron-poor (9a, 10a) arenes
are tolerated, although higher conversions are noted for electron-
rich substrates, consistent with fluorination from a nucleophilic
benzylic radical. Boronic acid pinacol esters are typically sensitive
to oxidative cleavage, but are tolerated under our reaction
conditions (12a−13a). Although direct heterobenzylic fluorina-
tion is a challenge, nitrogen-containing substrates were
successfully fluorinated (15a−17a), supporting the feasibility of
late-stage fluorination on pharmaceutically relevant scaffolds.
Basedsimplyontheanalysisofreductionpotentials, Selectfluor
should not be capable of generating Ag(II) from Ag(I) via single-
electron transfer. To provide spectroscopic evidence for the
formation of Ag(III)−F, we studied the radical fluorination of
17
substrate10byinsituReactIR. Weweresurprisedtofindthatno
significant change in Selectfluor concentration was observed
(
NFSI) was ineffective as an electrophilic fluorine source, which
o
may be rationalized by its lower reduction potential (E = −0.78
when exposed to stoichiometric AgNO (Figure 1a). Upon the
3
B
DOI: 10.1021/acs.orglett.7b01188
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
Scheme 3. Scope of Benzylic Fluorination
Figure 2. Cyclic voltammetry of Ag(I)/Ag(II) systems. Electrochemical
conditions: AgNO3 (0.4 mmol) in 5 mL of CH CN, tetrabutyl-
3
ammoniumhexafluorophosphate supportingelectrolyte(0.1M), glycine
(0.4 mmol). See Supporting Information for additional details.
These experiments suggest that amino acid ligation of the Ag(I)
catalyst is necessary for decarboxylation. Although N-Boc-glycine
is incapable of promoting radical fluorination on its own, the
addition of 1 equiv of pyridine enabled the fluorination of 10 in
19
6
7% isolated yield (Scheme 4).
Scheme 4. Catalyst Activation via Pyridine as an Additive
a
Yields refer to chromatographically pure material. Reaction
conditions: 10 (0.2 mmol), N-Boc-glycine (1.0 mmol), Selectfluor
a
Yields refer to chromatographically pure material unless otherwise
(1.0 mmol), pyridine (0.2 mmol), AgNO
3
(0.04 mmol), 2 mL of
noted. Reaction conditions: arene (0.2 mmol), glycine (0.4 mmol),
CH CN/H O (1:1).
3
2
Selectfluor (0.4 mmol), AgNO (0.04 mmol), 2 mL of CH CN/H O
3
3
2
b
(
1:1). 5.0 equiv of Selectfluor (1.0 mmol) and glycine (1.0 mmol).
To validate the role of the α-aminoalkyl radical as a HAT
reagent, we conducted deuterium-labeling studies. In situ carbon
NMR provides evidence for the formation of methylamine from
glycine during a standard reaction using toluene to produce 4a.
The analogous experiment using toluene-d8 (4-D) produces
deuterated methylamine upon fluorination under standard
c
NMR yield compared to benzotrifluoride as an internal standard.
19
conditions. Unfortunately, several attempts at in situ protection
of the resulting free amine were unsuccessful, precluding
quantitative analysis of deuterium incorporation. As shown in
Scheme 2, a consequence of a Ag(I)/(II) initiation would be the
stoichiometric formation of dication 1. It is possible that
unprotected amino acids act merely to initiate a chain radical
process that involves 1 as the active HAT agent. However,
running fluorinations with catalytic amounts of glycine (20−40
mol %) yielded only trace amounts of product, supporting its
action as a stoichiometric participant beyond radical chain
initiation.
Figure 1. ReactIR monitoring of selectfluor concentration. Reaction
conditions: 10 (0.2 mmol), glycine (0.4 mmol), Selectfluor (0.4 mmol),
AgNO (0.4 mmol), 2 mL of CH CN/H O (1:1). Reaction showed 25%
Based on our studies, we propose the following mechanism for
radical C−H fluorination using amino acids as radical precursors
(Figure 3). Precoordination of Ag(I) to one or more unprotected
glycines generates an electron-rich silver catalyst capable of
3
3
2
conversion to 10a by GCMS.
20
undergoing single-electron oxidation with Selectfluor. As
shown in Table 1 (entries 12−13), fully coordinated Ag(I)
complexes do not participate in radical fluorination, suggesting
that glycine reversibly associates with Ag(I). Fluorine NMR
shows that an aqueous fluoride anion is produced continuously
throughout the fluorination reactions, supporting a single-
electron transfer event with Selectfluor, although at this time a
Ag(I)/Ag(III) catalytic cycle cannot be fully ruled out. Rapid
decarboxylation from Ag(II) produces the α-aminoalkyl radical,
and hydrogen-atom transfer with a benzylic C−H bond generates
a benzylic radical and methylamine. Finally, the nucleophilic
benzylic radical reacts with Selectfluor to provide the desired
addition of glycine to this reaction mixture, a rapid decrease in
Selectfluor concentration was observed, suggesting that glycine
plays an important role in the reaction between Ag(I) and
Selectfluor (Figure 1b).
This resultled ustoreconsiderthe possibility ofaAg(I)/Ag(II)
catalystcyclefordecarboxylation ofunprotected aminoacids with
Selectfluor. Measuring the onset oxidation potential of AgNO in
3
the presence of glycine, we noticed a significant decrease in the
onset oxidation potential for the Ag(I)/Ag(II) couple (Figure
1
8
2
b). Conversely, whenN-Boc-glycine wasaddedtoasolutionof
AgNO , a negligible shift in oxidation potential was observed.
3
C
DOI: 10.1021/acs.orglett.7b01188
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
REFERENCES
■
(
1) For a recent review on the generation and reaction of α-aminoalkyl
radicals, see: Nakajima, K.; Miyake, Y.; Nishibayashi, Y. Acc. Chem. Res.
016, 49, 1946−1956.
2)Armstrong, D. A.;Rauk, A.;Yu,D. J. Am. Chem. Soc.1993, 115,666−
73.
3) For conjugate additions from α-aminoalkyl radicals derived from
amino acids, see: (a) Noble, A.; MacMillan, D. W. C. J. Am. Chem. Soc.
2
(
6
(
2
014, 136, 11602−11605. (b) McCarver, S. J.; Qiao, J. X.; Carpenter, J.;
Borzilleri, R. M.; Poss, M. A.; Eastgate, M. D.; Miller, M.; MacMillan, D.
W. C. Angew. Chem., Int. Ed. 2017, 56, 728−732.
(4) For cross-coupling from α-aminoalkyl radicals derived from amino
acids, see: (a) Zuo, Z.; MacMillan, D. W. C. J. Am. Chem. Soc. 2014, 136,
5
257−5260. (b) Zuo, Z.; Ahneman, D. T.; Chu, L.; Terrett, J. A.; Doyle,
A. G.; MacMillan, D. W. C. Science 2014, 345, 437−440. (c) Zuo, Z.;
Cong, H.; Li, W.; Choi, J.; Fu, G. C.; MacMillan, D. W. C. J. Am. Chem.
Soc. 2016, 138, 1832−1835.
Figure 3. Proposed mechanism for benzylic fluorination.
(
5) For Minisci reactions from α-aminoalkyl radicals derived from
amino acids, see: (a) Cowden, C. J. Org. Lett. 2003, 5, 4497−4499.
b) Shore, D. G. M.; Wasik, K. A.; Lyssikatos, J. P.; Estrada, A. A.
Tetrahedron Lett. 2015, 56, 4063−4066.
(
product. As illustrated in Figure 2, amino acid coordination to
Ag(I) is required to yield a catalyst of the appropriate oxidation
potential for reaction with Selectfluor. Because the oxidation of
glycine by Ag(II) is believed to occur rapidly from a metal-bound
species, the concentration of Ag(II) in solution is expected to be
lowthroughout the reaction. This has implications for the lifetime
of the α-aminoalkyl radical which is expected to be readily
oxidized by Ag(II). Based on the mechanism shown in Figure 3,
Selectfluor and 1 are the strongest oxidants the α-aminoalkyl
radical is likely to encounter, allowing HAT to occur prior to
unwanted oxidation.
In conclusion, we have developed a new method for C−H
benzylic fluorination using unprotected amino acids as radical
precursors. Fluorination proceeds under mild conditions and is
effective for several electron-rich benzylic substrates. Amino acid
binding to the silver precatalyst is critical for lowering the
oxidation potential of Ag(I), allowing oxidative decarboxylation
under mildly oxidizing conditions. Additional work exploring the
precise role of the amino acid throughout the reaction is ongoing.
(6) (a) Schonberg, A.; Moubacher, R. Chem. Rev. 1952, 50, 261−277.
(
(
b) Zelechonok, Y.; Silverman, R. B. J. Org. Chem. 1992, 57, 5787−5790.
c) Rizzi, G. P. Food Rev. Int. 2008, 24, 416−435. (d) Nashalian, O.;
Yaylayan, V. A. J. Agric. Food Chem. 2014, 62, 8518−8523.
(7) Mai, D. N.; Baxter, R. D. Org. Lett. 2016, 18, 3738−3741.
(8)Luo,Y.-R. ComprehensiveHandbookofChemicalBondEnergies;CRC
Press: Boca Raton, FL, 2007.
(9) For examples of metal complexes as HAT agents, see: (a) Mayer, J.
M. Acc. Chem. Res. 2011, 44, 36−46. (b) Jin, J.; MacMillan, D. W. C.
Angew. Chem., Int. Ed. 2015, 54, 1565−1569. (c) Qvortrup, K.;Rankic, D.
A.; MacMillan, D. W. C. J. Am. Chem. Soc. 2014, 136, 626−629. (d) Ding,
W.; Lu, L. − Q.; Liu, J.; Liu, D.; Song, H. − T.; Xiao, W. − J. J. Org. Chem.
2
2
016, 81, 7237−7243. (e) Zhang, Y.; Teuscher, K. B.; Ji, H. Chem. Sci.
016, 7, 2111−2118.
(10) For a discussion of α-oxy radicals generated from carbon-centered
radicals via the Bergman cyclization, see: Togo, H. Advanced Free Radical
Reactions for Organic Synthesis; Elsevier: Amsterdam, 2004; p 235.
(
11) For arecent reviewon radical fluorination, see: Chatalova-Sazepin,
C.; Hemelaere, R.; Paquin, J.-F.; Sammis, G. Synthesis 2015, 47, 2554−
569.
12) For alternative methods of radical benzylic fluorinations using
Selectfluor, see: (a) Xia, J.-B.; Zhu, C.; Chen, C. J. Am. Chem. Soc. 2013,
2
(
ASSOCIATED CONTENT
■
1
35, 17494−17500. (b) Bloom, S.; Pitts, C. R.; Woltornist, R.; Griswold,
*
S
Supporting Information
A.; Holl, M. G.; Lectka, T. Org. Lett. 2013, 15, 1722−1724. (c) Bloom, S.;
McCann, M.; Lectka, T. Org. Lett. 2014, 16, 6338−6341. (d) Xu, P.; Guo,
S.; Wang, L.; Tang, P. Angew. Chem., Int. Ed. 2014, 53, 5955−5958.
(
1
TheSupportingInformationisavailablefreeofchargeontheACS
Publications website at DOI: 10.1021/acs.orglett.7b01188.
13) Yin, F.; Wang, Z.; Li, Z.; Li, C. J. Am. Chem. Soc. 2012, 134, 10401−
0404.
Detailed experimental procedures, full characterization,
and copies of all spectra data (PDF)
(14) Patel, N. R.; Flowers, R. A. J. Org. Chem. 2015, 80, 5834−5841.
(15) (a) Pitts, C. R.; Bloom, S.; Woltornist, R.; Auvenshine, D. J.;
Ryzhkov, L. R.; Siegler, M. A.; Lectka, T. J. Am. Chem. Soc. 2014, 136,
9780−9791. (b) Pitts, C. R.; Ling, B.; Woltornist, R.; Liu, R.; Lectka, T. J.
Org. Chem. 2014, 79, 8895−8899.
AUTHOR INFORMATION
■
*
Corresponding Author
(
16) Liang, T.; Neumann, C. N.; Ritter, T. Angew. Chem., Int. Ed. 2013,
2, 8214−8264.
17) Substrate 10 maintained homogeneous conditions throughout all
E-mail: rbaxter@umcerced.edu.
5
(
ORCID
reaction progressions studied and was therefore chosen as a model
substrate for study via ReactIR.
Ryan D. Baxter: 0000-0002-1341-5315
Notes
(18) The reverse waveform of Ag(II) to Ag(I) was not observed under
these conditions due to the instability of aqueous Ag(II). For a detailed
explanation, see: Panizza, M.; Duo, I.; Michaud, P. A.; Cerisola, G.;
Comnellis, Ch. Electrochem. Solid-State Lett. 1999, 3, 550−551.
The authors declare no competing financial interest.
(
(
19) See Supporting Information for additional details.
20) (a) Kumar, A.; Neta, P. J. Am. Chem. Soc. 1980, 102, 7284−7289.
ACKNOWLEDGMENTS
■
Financial support for this work was provided by the University of
California, Merced, The Hellman Foundation, and ACS-PRF No.
6225-DNI5. We are grateful to Dr. Luke Reed (UC Merced) for
(b) Rendosova, M.; Vargova, Z.; Kuchar, J.; Sabolova, D.; Levoca, S.;
Kudlacova, J.; Paulikova, H.; Hudecova, D.; Helebrandtova, V.; Almasi,
M.; Vilkova, M.; Dusek, M.; Bobalova, D. J. Inorg. Biochem. 2017, 168, 1−
5
1
2.
assistance with cyclic voltammetry.
D
DOI: 10.1021/acs.orglett.7b01188
Org. Lett. XXXX, XXX, XXX−XXX