The merger of decatungstate and copper catalysis to enable aliphatic C(sp 3)–H trifluoromethylation

Article abstract of DOI:10.1038/s41557-020-0436-1

The introduction of a trifluoromethyl (CF₃) group can dramatically improve a compound’s biological properties. Despite the well-established importance of trifluoromethylated compounds, general methods for the trifluoromethylation of alkyl C–H bonds remain elusive. Here we report the development of a dual-catalytic C(sp³)–H trifluoromethylation through the merger of light-driven, decatungstate-catalysed hydrogen atom transfer and copper catalysis. This metallaphotoredox methodology enables the direct conversion of both strong aliphatic and benzylic C–H bonds into the corresponding C(sp³)–CF₃ products in a single step using a bench-stable, commercially available trifluoromethylation reagent. The reaction requires only a single equivalent of substrate and proceeds with excellent selectivity for positions distal to unprotected amines. To demonstrate the utility of this new methodology for late-stage functionalization, we have directly derivatized a broad range of approved drugs and natural products to generate valuable trifluoromethylated analogues. Preliminary mechanistic experiments reveal that a ‘Cu–CF₃’ species is formed during this process and the critical C(sp³)–CF₃ bond-forming step involves the copper catalyst. [Figure not available: see fulltext.].

Full text of DOI:10.1038/s41557-020-0436-1

Articles
https://doi.org/10.1038/s41557-020-0436-1
The merger of decatungstate and copper catalysis
to enable aliphatic C(sp3)–H trifluoromethylation
Patrick J. Sarverꢀ ꢀ1,4, Vlad Bacauanuꢀ ꢀ1,4, Danielle M. Schultzꢀ ꢀ2, Daniel A. DiRocco2, Yu-hong Lam3,
3
1
ꢀ✉
Edward C. Shererꢀ ꢀ and David W. C. MacMillanꢀ ꢀ
The introduction of a trifluoromethyl (CF3) group can dramatically improve a compound’s biological properties. Despite the
well-established importance of trifluoromethylated compounds, general methods for the trifluoromethylation of alkyl C–H
bonds remain elusive. Here we report the development of a dual-catalytic C(sp3)–H trifluoromethylation through the merger
of light-driven, decatungstate-catalysed hydrogen atom transfer and copper catalysis. This metallaphotoredox methodology
enables the direct conversion of both strong aliphatic and benzylic C–H bonds into the corresponding C(sp3)–CF3 products in
a single step using a bench-stable, commercially available trifluoromethylation reagent. The reaction requires only a single
equivalent of substrate and proceeds with excellent selectivity for positions distal to unprotected amines. To demonstrate the
utility of this new methodology for late-stage functionalization, we have directly derivatized a broad range of approved drugs
and natural products to generate valuable trifluoromethylated analogues. Preliminary mechanistic experiments reveal that a
‘Cu–CF3’ species is formed during this process and the critical C(sp3)–CF3 bond-forming step involves the copper catalyst.
t has long been established that the incorporation of the trifluo- steps22, has allowed the invention of robust methods for the synthesis
romethyl (CF3) group into drug molecules can often improve of formerly elusive C(sp3)–N (ref. 23) and C(sp3)–CF3 bonds24,25
.
I
their pharmacokinetic properties, including permeability, meta-
Inspired by these collective studies, we recently sought to combine
bolic stability and protein binding affinities (Fig. 1a)1,2. As such, the copper catalysis with a strong C(sp3)–H bond-cleaving HAT catalyst
development of general technologies that enable aryl and alkyl triflu- in the presence of a suitable CF3 radical source. Given the remarkable
oromethylation has become a major area of both industrial and aca- efficiency with which photoexcited decatungstate (*[W10O32]4−) can
demic research in recent years3,4. Significant efforts have focused on cleave strong C–H bonds26,27, it is no surprise that this polyoxometal-
the direct conversion of C–H bonds to C–CF3 bonds, which removes late has found broad application in a number of synthetically useful
the need for prefunctionalization of drug intermediates and allows C(sp3)–H functionalizations28, including oxidations29,30, dehydro-
derivatization at carbon sites that lack traditional functional handles. genations31, fluorinations32, azidations33 and conjugate additions34.
To this end, aromatic C(sp2)–H trifluoromethylations via Minisci- Moreover, the highly electrophilic ligand-to-metal charge transfer
type pathways have become widely utilized in a variety of contexts5,6; excited state of decatungstate is known to engage aliphatic substrates
however, such methods are not amenable to aliphatic C(sp3)–H bond with predictable selectivity for abstraction of the most sterically
functionalization (Fig. 1b). Although C(sp3)–H trifluoromethylation accessible, electron-rich C(sp3)–H bond35. Recently, our laboratory
protocols have been developed for allylic7,8 and benzylic systems9–11
,
utilized this oxometallate HAT agent in the development of a nickel/
as well as substrates that incorporate Lewis basic directing groups12, decatungstate dual-catalytic C(sp3)–H arylation methodology, which
there have been no reports so far of general methods for the non- demonstrates the potential to combine decatungstate photocatalysis
directed C(sp3)–H trifluoromethylation of strong aliphatic C–H with transition metal-catalysed cross-coupling36. Herein, we describe
bonds. With this in mind, we recently sought to design a new, direct the first merger of decatungstate anion photochemistry with copper
trifluoromethylation technology that allows the rapid synthesis of catalysis and introduce a new protocol for the direct trifluorometh-
CF3-bearing fragments from simple feedstocks, as well as provides ylation of C(sp3)–H aliphatic substrates (Fig. 1c). To ensure utility in
single-step access to trifluoromethyl analogues of pharmaceutically a medicinal chemistry setting, we sought to identify a dual catalysis
important molecules via late-stage functionalization.
platform that would tolerate a range of polar functional groups, such
Over the last five years, metallaphotoredox has emerged as a valu- as unprotected amines and carboxylic acids37, while implementing
able catalysis platform that enables native functional groups, such a commercially available, bench-stable trifluoromethylation reagent.
as carboxylic acids and alcohols, to be employed as C(sp3)-coupling Perhaps as important, this methodology employs only a single equiv-
partners via the intermediacy of open-shell species13. Within this par- alent of the C(sp3)–H-bearing substrate, an unusual and especially
adigm, a range of highly modular C–H functionalizations have also attractive feature for late-stage C–H functionalization of complex or
been developed by combining photoredox-mediated hydrogen atom medicinally relevant molecules.
transfer (HAT) with nickel-catalysed radical cross-coupling to deliver
a series of aliphatic-arylation protocols14–17. Beyond nickel, a number Results and discussion
of metallaphotoredox mechanisms have successfully employed copper A detailed description of our reaction design is provided in Fig. 2.
as the coupling catalyst18, which, given its remarkable capacity for both Photoexcitation of [W10O32]4− (1) was expected to afford the oxo-
radical capture19–21 and traditionally difficult reductive elimination metallate excited state 2. In the case of amine substrates such as
1Merck Center for Catalysis at Princeton University, Princeton, NJ, USA. 2Process Research and Development, Merck & Co., Inc., Rahway, NJ, USA.
3Computational and Structural Chemistry, Merck & Co., Inc., Rahway, NJ, USA. 4These authors contributed equally: Patrick J. Sarver, Vlad Bacauanu.
✉
e-mail: dmacmill@princeton.edu
NATuRE CHEMiSTRY | www.nature.com/naturechemistry

NaTure CHemisTry
Articles
a
c
F3C
I
O
O
Decatungstate
photocatalysis
Copper
CF3
Me
Me
N
metal catalysis
O
N
Me Me
HN
S
Low potency
Alpelisib
Anticancer
NH2
W
Cu
Commercial
CF3 source
N
Metabolically labile
O
b
Aliphatic C–H substrates
HN
Trifluoromethyl products
HN
H
CF3
H
H
CF3
F3C
N
N
N
CO2H
CO2H
CO2H
CO2H
Me
F3C
N
Me
Me
O
O
H
N
N
H
F3C
Me
Me
Aromatic C–H sites
Well established
Biologically relevant
substrate
Aliphatic C–H sites
No general methods
H
N
N
H
O
H
N
N
H
O
Me
Me
Fig. 1 | Dual decatungstate/copper metallaphotoredox catalysis enables direct C(sp3)–H trifluoromethylation. Despite the broadly appreciated
importance of trifluoromethylated compounds in the pharmaceutical industry, general catalytic methods for C(sp3)–H trifluoromethylation remain
elusive. This transformation has now been achieved through the merger of decatungstate and copper metallaphotoredox catalysis. a, The synthesis of
trifluoromethylated compounds has become an invaluable tool for the development of potent, metabolically stable pharmaceuticals, as demonstrated
by the approved pharmaceutical alpelisib2. b, To enable more efficient syntheses of these valuable compounds, recent efforts have focused on the
development of broadly applicable C–H trifluoromethylations. Despite significant progress in aryl C–H trifluoromethylation, no general catalytic methods
for alkyl C–H trifluoromethylation have been developed. c, By combining decatungstate-catalysed HAT with copper catalysis, we have developed a broadly
applicable C(sp3)–H trifluoromethylation of strong aliphatic and benzylic C–H bonds.
4–
CF3
10
O
O
e–
CF3
4 Na+
W
O
W
O
C(sp3)–CF3
O
O
O
O
source
O
O
Electron
O
product
N
X
O
O
O
W
W
H
H
O
W
O
O
W O
O
O
W
O
O
W
O
O
Redox-mediated
CuII–CF3 formation
O
O
O
O
O
O
W
O
W
O
4–
Ln CuI
[W10O32
]
– H
1
7
W
NaDT, 11
SET
390 nm
light source
Cu
Photo-HAT
catalytic
cycle
Copper
catalytic
cycle
Cl
Cl
No exogenous
ligand
H
Cu
CF3
X
H
N
H+ [W10O32
]
5– (6)
Ln CuII
*[W10O32
HAT reagent
]
4– (2)
CuCl2
Ln CuIII
CF3
Reductant
X
8
9
H
H
3
4
HAT
5
H
H
H
N
H
N
N
Polarity-driven
selectivity
H
H
H
H
Alkyl
Neutral C–H bond
Hydridic, favoured
Neutral, favoured
Protic, disfavoured
radical
Fig. 2 | Mechanistic design for direct C(sp3)–H trifluoromethylation via decatungstate and copper catalysis. Excitation of decatungstate (1) with
near-ultraviolet light affords highly reactive excited state 2. Pyrrolidine (3), which is protonated under acidic reaction conditions to afford pyrrolidinium
4, undergoes a subsequent HAT by 2 at the more electron-rich β-position to afford a reactive alkyl radical (5) and reduced decatungstate (6). Formal
reduction of an electrophilic trifluoromethylation reagent by 6 in the presence of a copper(i) catalyst (7) generates a Cu(ii)–CF3 species (8), which can
rapidly trap alkyl radical 5 to produce key alkyl–Cu(iii)–CF3 species 9, which undergoes facile reductive elimination to regenerate the Cu(i) catalyst 7 and
C(sp3)–CF3 product 10.
pyrrolidine (3), protonation of the basic nitrogen atom renders the This polarity modulation strategy has previously been implemented
adjacent hydrogen atoms both stronger and less hydridic38, enabling in the direct, distal oxidation of aliphatic amines to generate ketones
reactivity at the traditionally less reactive distal position39–41
.
in lieu of amide moieties29. Thus, subsequent HAT from the more
NATuRE CHEMiSTRY | www.nature.com/naturechemistry

NaTure CHemisTry
Articles
polarity-matched β-C(sp3)–H bond of pyrrolidinium 4 by electro-
philic excited state decatungstate 2 would produce the alkyl radical 5
andthereduceddecatungstate6.Formalreductionofanelectrophilic
CF3 source by the ground-state polyoxometallate 6 in the presence
of a Cu(i) catalyst (7) should then afford a Cu(ii)–CF3 intermedi-
ate (8) and regenerate the decatungstate anion (1) (Supplementary
Fig. 25). At this stage, we assumed the copper(ii)–CF3 species 8
would capture the carbon-centred radical 542,43 at near-diffusion
rates to produce the critical alkyl–Cu(iii)–CF3 complex (9)20, which,
upon reductive elimination, would afford the desired C(sp3)–CF3
product 10 while regenerating the Cu(i) catalyst11,44. Computational
studies support the feasibility of the radical capture and reductive
elimination steps (Supplementary Section 10), although alternative
mechanisms have not been ruled out at this time.
1.2 equiv. H2SO4
a
b
F3C
H
I
O
CF3
1 mol% NaDT, 5 mol% CuCl2
O
+
9:1 H2O/MeCN (0.025 M)
N
N
H
H
390 nm Kessil lamp, 20–30 °C
12
1 equiv.
66% yield
1.25 equiv.
H
H
3
13
14
H
NH2
Initial
N
evaluation
of amines
N
H
Cl
H
No oxidation
side products
Oxidation side products observed
Preliminary studies into this new C–H trifluoromethyl-
ation reaction were performed with pyrrolidine, Togni reagent II
(1.25equiv., 12)4, sodium decatungstate (1mol%, 11), a copper(ii)
chloride precatalyst (5mol%), an acidic medium (1.2equiv. H2SO4)
of water/acetonitrile (90:10) and a Kessil 40W 390nm lamp to pro-
vide the desired β-CF3 product in 66% yield as a single regioiso-
mer (Fig. 3). Further investigations revealed a surprising finding for
decatungstate chemistry: specifically that the presence of chloride
anion (1–10equiv.) can substantially improve the yield of this trans-
formation for a number of substrate classes (vide infra). Control
experiments highlighted that this trend is general for a broad range
chloride sources, but specific to the chloride anion in lieu of other
metal counterions (Supplementary Tables 13 and 14). Notably, this
effect is not observed uniformly, as the presence of chloride was
not found to be beneficial for pyrrolidine functionalization (pos-
sibly due to unproductive quenching of excited decatungstate45),
and the reaction is competent with numerous other copper precata-
lysts (Supplementary Table 6). Although the precise origin of this
effect remains under investigation, preliminary studies suggest that
chloride anion alters the ligand sphere of copper under these reac-
tion conditions. This proposal is supported by changes observed
in the ultraviolet–vis spectrum of aqueous copper(ii) chloride at
reaction concentrations upon addition of sodium chloride (from
predominantly Cu2+ to CuCl+ based on literature spectra in water46;
Supplementary Figs. 14–17) and effects on diastereoselectivity that
trend with chloride concentration (Supplementary Table 15). We
presume that the presence of chloride should make the resulting
copper ensemble less cationic, and therefore less oxidizing, a char-
acteristic that would disfavour the formation of alkene or solvoly-
sis products associated with the well-established oxidation of alkyl
radicals by Cu(ii) salts47,48. Our hypothesis finds further support
in that the addition of sodium chloride (1–10equiv.) substantially
improves reaction efficiencies for substrates that can readily ren-
der oxidative byproducts such as ketones, alcohols and alkenes.
An alternative hypothesis for the beneficial role of chloride ion is
that the HAT step is mediated by chlorine radicals generated via
oxidation by excited decatungstate. Although certainly plausible
based on observed quenching rate constants45, efforts to mediate
the reaction with chlorine radicals by replacing decatungstate with
other oxidizing photocatalysts have thus far proven unsuccessful
(Supplementary Table 16).
NaCl
Yield
Yield
Yield
additive
from 3 (%)
from 13 (%)
from 14 (%)
0 equiv.
2 equiv.
10 equiv.
66
61
49
35
63
57
39
48
54
c
+
2+
OH2
OH2
Cu
OH2
Cu
OH2
Cl
Cl
H2O
H2O
H2O
OH2
Cl
Cl
Cu
OH2
Cl
OH2
OH2
H2O
H2O
OH2
OH2
More oxidizing copper species
Aqueous Cu2+
Aqueous Cu2+ with NaCl
O
CF3
NH2
NH2
NH2
Oxidation side products
Radical intermediate
Desired CF3 product
Fig. 3 | Effect of chloride anion on reaction efficiency. During the course
of optimization, the presence of chloride was found to dramatically affect
reaction efficiency. See Supplementary Sections 4, 6 and 7 for experimental
details and further discussion. a, Irradiating a solution of pyrrolidine and
Togni reagent II (12) in acidic water/acetonitrile with near-ultraviolet
light in the presence of catalytic sodium decatungstate and copper(ii)
chloride affords the desired 3-trifluoromethylpyrrolidine product in 66%
analytical yield. b, Subsequent investigation of the reaction scope revealed
a substrate-dependent effect of chloride anion concentration on the yield
of this transformation. c, A potential mechanistic basis for this trend is the
effect of chloride on the ligand sphere of copper. In the presence of chloride,
a less oxidizing copper chloride species appears to produce fewer oxidative
side products, resulting in improved reaction efficiency.
Given the importance of the amine moiety in biomedically rel- anion’s ability to discriminate between hydrogen atoms three and four
evant molecules and the unique HAT selectivity enabled by proton- bonds away from an electron-withdrawing group. Moreover, trans-
ation at nitrogen, we first examined our reaction scope on a diverse 1,2-diaminocyclohexane, a common ligand scaffold49, was directly
set of alkyl amines (Table 1). Despite the wide range of C–H bond converted to the 4-trifluoromethylated product (23) in good yield
dissociation energies and bond polarities, our conditions proved (78%) and excellent selectivity (94%). This C–H trifluoromethylation
general for a broad array of cyclic and linear amines, with good to methodology was further found to be applicable to a range of medici-
excellent selectivities for trifluoromethylation at positions distal to nally relevant bicyclic amines (24–26, 32–44% yield, single regioi-
the protonated nitrogen (15–23, 52–83% yield, 61 to >95% selectiv- somer in all cases). For example, nortropinone, a common scaffold
ity, defined as the percentage of the major regioisomer over all regioi- in bioactive molecules50, was successfully functionalized to generate
somers). In the case of n-pentylamine (20, 62% yield), a 2.5:1 ratio of a single CF3-addition product (25, 32% yield). Notably, in cases for
δ to γ functionalization was observed, illustrating the decatungstate which the synthesis of the trifluoromethyl analogue was previously
NATuRE CHEMiSTRY | www.nature.com/naturechemistry

NaTure CHemisTry
Articles
Table 1 | Scope of the direct trifluoromethylation of strong aliphatic C–H bonds
O
O
Cu
W
O
W
H+
CF3
H
W
O
F3C
I
O
Cu
W
O
O
O
O
Cl
Cl
O
O
O
Cu
+
O
O O
O
W
W
O
H2O/MeCN, 20–30 °C
N
H
N
O
W
O
O
W O
H
O
O
W
W
O
O
O
O
O
O
C–CF3 product
No exogenous ligand
CF3
1 equiv.
1.25 equiv.
O
O
O
W
O
W
O
Minor regioisomer
Aliphatic amines
H
CF3
H
H
CF3
HO2C
HO2C
N
N
H
N
H
N
H
H2N
H2N
H
27 62% (60%)d,e
(
)-15 68% (66%)
21 66% (58%)a
Proline
Single regioisomer
61% selectivity
58% selectivity
CO2H
CO2H
CF3
H
H
CF3
H
H
CF3
HN
HN
H2N
H2N
HN
HN
16 83% (80%)
75% selectivity
(
)-22 52%b
89% selectivity
(
)-28 65% (58%)f
88% selectivity
H
CF3
O
O
CF3
H2N
H
H2N
CF3
N
N
H2N
H2N
H
H
N
H
N
H
(
)-17 56% (51%)
(
)-23 78%c
(
)-29 45%e,g
83% selectivity
94% selectivity
Single regioisomer
H
CF3
HN
HN
CF3
H
H
CF3
H2N
H
H2N
Me
N
N
O
O
(
)-18 71%
(
)-24 44% (39%)
(
)-30 51% (46%)e
76% selectivity
Single regioisomer
Single regioisomer
NH
NH
H
H
CF3
CF3
H
H
H2N
CO2Me
O
H2N
CO2Me
O
H2N
Me
H2N
CF3
)-25 32% (31%)d
(
)-31 55%h
46% selectivity
(
)-19 66%
Nortropinone
GABA methyl ester
H
(
81% selectivity
Single regioisomer
CF3
H
N
H
N
CF3
H
H
CF3
H2N
H
H2N
Me
NH2
N
NH2
N
)-26 44% (33%)d
32 50%i,j
(
)-20 62%
(
63% selectivity
Single regioisomer
82% selectivity
Neutral unactivated substrates
H
O
O
CF3
H
CF3
Me
N
N
CF3
HO2C
HO2C
HO2C
HO2C
S
S
O
O
O
O
O
O
34 59%k
97% selectivity
35 53% (50%)k
(
)-33 64% (64%)
Single regioisomer
All yields and selectivities were determined by 19F NMR versus 2,2,2-trifluoroethanol (average of two trials). Isolated yields appear in parentheses. Selectivity is reported as the percentage of the major
regioisomer over all regioisomers. ‘Single regioisomer’ refers to substrates for which >95% selectivity was observed. Volatile amines were isolated as the corresponding trifluoroacetate or chloride salt.
Standard conditions: substrate (0.5ꢀmmol, 1ꢀequiv.), Togni reagent II (12, 1.25ꢀequiv.), NaDT (1–3ꢀmol%), CuCl2 (5–10ꢀmol%), H2SO4 (0–2.5ꢀequiv.), NaCl (0–10ꢀequiv.), H2O/MeCN (9:1 to 1:1, 0.025ꢀM),
8–12ꢀh of irradiation with a Kessil 40ꢀW 390ꢀnm lamp at 20–30ꢀ°C. See Supplementary Section 15 for full experimental details. a1.5:1 d.r. (major), >20:1 d.r. (minor). b1.7:1 d.r. (major), >20:1 d.r. (minor).
c1.2:1 d.r. (major), >20:1 d.r. (minor). d>20:1 d.r. (all regioisomers). etwo additions of Togni reagent and NaDT. f1.3:1 d.r. (major), >20:1 d.r. (minor). gisolated and characterized following reduction with
NaBH4. h3:1 substrate:Togni reagent (0.5ꢀmmol). i2.5:1 d.r. (major), 1.7:1 d.r. (minor). j0.1ꢀmmol scale. k20–25% CD3CN in H2O as solvent.
reported, the capacity to selectively engage the C–H bonds of simple,
Our subsequent studies focused on the direct trifluoromethyl-
unprotected amines greatly improves the step efficiency relative to ation of aliphatic amines that incorporate a wide array of functional
traditional functional group conversion protocols51,52
.
groups that are often challenging in metal-mediated couplings.
NATuRE CHEMiSTRY | www.nature.com/naturechemistry

NaTure CHemisTry
Articles
For example, proline was directly engaged at positions distal to the functionalization with ortho-substituted aryl systems was also
protonated amine, generating CF3-bearing analogues in a single step possible to efficiently generate the CF3-adduct 47 from the corre-
and with excellent diastereoselectivity (27, 62% yield, 58% selectiv- sponding toluene derivative (52% yield).
ity, >20:1 d.r.). The reaction’s tolerance for carboxylic acids proved
Next, investigations into medicinally relevant heterocyclic scaf-
general, with 28 formed in good efficiency (65% yield) and excellent folds demonstrated that a range of pyridines (48–50, 51–64% yield)
C–H abstraction selectivity (88%). The presence of other electron- can be employed to generate the corresponding C(sp3)–CF3 adducts
withdrawing groups, including ketones and esters, further improved in good yields. Remarkably, electron-poor heteroarenes such as
the regiocontrol of this transformation, enabling the synthesis of pyrimidines (51, 71% yield) and pyridazines (52, 43% yield) were
the corresponding trifluoromethyl adducts (29–31) in good yields also efficiently functionalized, highlighting the potency of deca-
(45–55%) and good to excellent selectivities (46 to >95%). For tungstate as a HAT catalyst, even in cases of electronically mis-
example, the methyl ester of γ-aminobutyric acid (GABA), an matched C–H abstraction. Five-membered heteroarenes were also
important neurotransmitter, was successfully engaged to afford 31 competent under our protocol, as demonstrated by the trifluoro-
(55% yield, 46% selectivity), a notable result given the prevalence of methylation of a bicyclic pyrazole (53, 68% yield).
lipophilic GABA analogues among approved drugs53. Additionally,
Furthermore, we set out to demonstrate the applicability of this
a pyridine-bearing cyclobutylamine was successfully subjected to transformation to the late-stage functionalization of pharmaceuti-
trifluoromethylation with excellent regioselectivity for the most cals and natural products (Fig. 4c). For example, nicotine readily
electron-rich position (32, 50% yield, 82% selectivity), underscor- underwent trifluoromethylation at positions distal to the proton-
ing both the selectivity of our transformation and the broad tol- ated amine and pyridine moieties to afford previously unreported
erance for coordinating functional groups provided by the mildly analogues in a single step (54, 40% yield, 67% selectivity). In this
acidic reaction medium.
case, selectivity was observed for the strong, electron-neutral C–H
A number of non-amine-bearing compounds were also com- bonds over the α-ammonium heterobenzylic position due to the
petent substrates for this new CF3-installation technology (33–35, high degree of polarity mismatch56. Sclareolide, a well-studied ses-
53–64% yield, >95% selectivity), demonstrating that this protocol is quiterpenoid, reacted to afford a single regio- and diastereoisomer
also applicable to organic substrates, broadly defined. For example, (55, 25% yield). The gabapentin analogue gababutin53 was directly
both sulfolane and cyclobutane-1,1-dicarboxylic acid afforded the trifluoromethylated with comprehensive regiocontrol to afford two
corresponding C(sp3)–CF3 products with excellent regioselectiv- diastereomeric derivatives, albeit in modest yield (56, 30% yield).
ity, favouring the positions furthest from the electron-withdrawing For pharmaceuticals bearing benzylic C–H bonds less polarity-mis-
groups (33 and 34, 64 and 59% yield, respectively). Implementation matched than those in nicotine, excellent selectivity was observed
of N-methylphthalimide yielded the corresponding C–H trifluoro- for benzylic functionalization (57–60, 33–72% yield). Moreover,
methylation product (35, 53% yield), indicating that our transforma- lidocaine was successfully engaged at both benzylic positions to
tion can effectively functionalize electron-poor α-heteroatom C–H afford the bis-CF3 analogue 57 (72% yield). Perhaps most notably,
bonds. It should be noted that more electron-rich α-heteroatom torsemide afforded the corresponding benzylic trifluoromethyl-
positions are not successful using this protocol (Supplementary ation adduct (60, 33% yield), further highlighting the remarkable
Fig. 45), presumably due to facile copper-mediated oxidation of the functional group tolerance and predictable selectivity of this new
intermediate α-amino or α-oxy radical.
C(sp3)–CF3 installation protocol.
We next sought to evaluate the scope of this transformation with
Given the broad range of copper-catalysed reactions employing
respect to substrates bearing benzylic C–H bonds, a common motif Togni reagent II but surprisingly limited mechanistic understand-
found among pharmaceutical agents. Given the metabolic lability ing of such transformations4, we concluded our studies by investi-
of benzylic C–H bonds to cytochrome P450-mediated oxidation54, gating the mechanism of this dual-catalytic system. Consistent with
we envisioned that an operationally simple trifluoromethylation of our mechanistic design (Fig. 2), these studies provide substantial
such compounds would provide a useful tool for medicinal chem- evidence for a mechanism in which ‘Cu–CF3 complexes’ are gener-
ists. Furthermore, the lower bond dissociation energy of benzylic ated via the interaction of copper(i) with Togni reagent II, and that
C–H bonds compared to alkyl C–H bonds should enable robust, the copper catalyst is involved in the critical C(sp3)–CF3 bond-form-
predictable selectivity in late-stage functionalization. Fortunately, ing step. Key components of this evidence include (1) the privileged
our methodology proved effective for a broad range of benzylic nature of copper as a catalyst for this transformation; (2) the gen-
and heterobenzylic substrates, affording primary and secondary eration of Cu–CF3 species via the reaction of copper(i) with Togni
trifluoromethylation adducts with complete selectivity observed reagent II, the competency of these species in our transformation,
for the benzylic position (>20:1 regioisomeric ratio in all cases; and the observation of such species under reaction conditions; (3)
Fig. 4a,b). Phenethylamines, an important class of neuroactive mol- the observation of enantioinduction and improvements to diastere-
ecules55, provided the corresponding benzylic C–CF3 products in oselectivity in the presence of a chiral ligand.
good to high efficiency (36–38, 47–72% yield) and with complete
To elaborate, control experiments indicated that copper is unique
preservation of enantiopurity in the case of (S)-1-aminoindane (38, in its effectiveness for this C(sp3)–H trifluoromethylation when com-
>99% e.e.). Similarly, aliphatic systems bearing ester, chloride or pared to water-stable Lewis acids57 or other metals known to reduce
acetoxy substituents were selectively converted to the benzylic tri- Togni reagent II (Fig. 5a and Supplementary Tables 9 and 10)4. These
fluoromethyl analogues (39–41, 52–59% yield). A broad range of data suggest that copper probably plays a more significant role than
functional groups were also tolerated on the aryl ring, including merely single-electron reduction of Togni reagent II58, consistent
ketones, esters, sulfonamides and tetrazoles (42–45, 65–77% yield). with our mechanistic design invoking the generation of Cu(ii)–CF3.
Notably, selectivity for the benzylic position was still observed In support of this hypothesis, the mixing of copper(i) triflate with
when strongly electron-withdrawing groups were incorporated Togni reagent II under reaction-relevant conditions was shown to
at the aryl para positions, suggesting that the polarity preference generate a 19F NMR-active Cu(iii)–CF3 species and a new electron
for electron-rich sites can be overridden by the bond dissociation paramagnetic resonance (EPR)-active copper(ii) species (Fig. 5b,c).
energy differential between benzylic and primary C–H bonds. Based on the previously reported equilibrium between relevant
Moreover, an unsubstituted aniline, which would typically quench CuIII– and CuII–CF3 complexes in the presence of copper(i)59, we
the excited state of a photocatalyst under basic or neutral condi- attribute this new EPR signal to a species of the form CuII(CF3)x;
tions45, was efficiently functionalized at the benzylic position further studies into the exact structure of this copper complex are
when an acidic medium was employed (46, 71% yield). Benzylic ongoing. Subsequent experiments conclusively demonstrate that
NATuRE CHEMiSTRY | www.nature.com/naturechemistry

NaTure CHemisTry
Articles
a
b
CF3
Me
CF3
Me
CF3
Me
CF3
NH2
CF3
O
CF3
Me
Cl
N
OMe
Me
N
N
N
Cl
N
NH
( )-45 70% yield
O
N
)-36 60% yielda,b
CF3 Me
(
)-48 51% yieldb
(
)-51 71% yieldb
(
)-39 52% yield
(
)-42 70% yield
(
CF3
Me
Cl
CF3
CF3
CF3
Me
N
N
CF3
H2N
Me
NH2
Cl
MeO
Me
Cl
N
O
)-49 64% yieldb,f
(
)-37 72% yielda,c
(
)-40 54% yield
(
)-43 65% yield
(
)-46 71% yield
52 43% yieldg
(
CF3
Me
CF3
CF3
CF3
CF3
H
N
CF3
O
N
OAc
H2N
MeO
N
S
N
H
Me
NH2
O
O
(
O
38 47% yielda,d,e
50 61% yieldf
(
)-41 59% yield
)-44 77% yield
47 52% yield
(
)-53 68% yield
c
O
O
O
H
CF3
H
CF3
Me
Me
O
H
CF3
N
N
Me
H
H
Me
H
H
NH2
NH2
Me
N
Me
N
CO2H
CO2H
Me Me
Me Me
55 25%
54 40%h
(
)-56 30%f,i
Nicotine
Sclareolide
Gababutin
Single regioisomer
67% selectivity
Single regioisomer
CF3
F3C
Me
Me
Me
H
Me
H
N
N
H
N
H
N
Me
Me
Me
N
Me
N
Me
N
H
N
O
H
O
O
O
Me
Me
CF3
Lidocaine
57 72%
Prilocaine
(
)-58 63%
Single regioisomer
Single regioisomer
H
N
H
Me
O
H
H
CF3
O
N
Me
N
N
Me
S
S
N
N
O
O
Me
F3C
O
O
Me
F3C
N
N
NH
N
NH
N
O
O
S
S
CF3
Me
O
NH2
O
NH2
60 (45%)j 33%
Celecoxib
59 57%
Torsemide
Single regioisomer
Fig. 4 | Extension to benzylic C–H trifluoromethylation and application to natural products and pharmaceuticals. This methodology was found to be
applicable to a broad range of benzylic substrates, pharmaceuticals and natural products. Standard conditions: substrate (0.5ꢀmmol, 1ꢀequiv.), Togni
reagent II (12, 1.25ꢀequiv.), NaDT (1ꢀmol%), CuCl2 (5–10ꢀmol%), H2SO4 (0–1.5ꢀequiv.), H2O/MeCN (9:1 to 1:1, 0.025ꢀM), 12ꢀh of irradiation with a Kessil 40ꢀW
390ꢀnm lamp at 20–30ꢀ°C. All yields are isolated unless otherwise noted. aIsolated as the chloride salt. bPerformed with NaCl (1–10ꢀequiv.). c1:1 d.r. d16:1 d.r.
ePerformed with Cu(OTf)2 (Supplementary Table 15). fThe 19F NMR yield versus 2,2,2-trifluoroethanol. g20% CD3CN in H2O as solvent. h1.5:1 d.r. (major).
i2.3:1 d.r. j2:1 substrate:Togni reagent. See Supplementary Section 15 for full experimental details. a, Compounds bearing benzylic C–H bonds are directly
trifluoromethylated with excellent selectivity for the benzylic position in all cases. A diverse range of functionality on the aryl ring and alkyl chain is tolerated.
b, Heterobenzylic substrates, including electron-poor pyridines and pyrimidines, were directly functionalized at the benzylic position. c, A diverse array of
approved pharmaceuticals, bioactive compounds and natural products are directly trifluoromethylated to afford valuable C(sp3)–CF3 analogues. Selectivity is
reported as the percentage of the major regioisomer over all regioisomers. Blue circles denote sites where minor regioisomers are observed.
this ensemble of ‘Cu–CF3’ species is a competent replacement for complexes are potentially catalytic intermediates. Subsequent
the copper(ii) chloride and trifluoromethyl source employed under spectroscopic investigations revealed that the aforementioned
the optimized conditions (Fig. 5b), indicating that such Cu–CF3 CuIII(CF3)x complex is also observed at partial conversion under the
NATuRE CHEMiSTRY | www.nature.com/naturechemistry

NaTure CHemisTry
Articles
a
c
1.2 equiv. H2SO4
e
H
F3C
I
O
CF3
Equlibrium
reported for related
Cu–CF3
1 mol% NaDT, 5–10 mol% metal
CuIII(CF3)x
CuII(CF3)x
O
+
9:1 H2O/MeCN (0.025 M)
(ref. 59
)
N
N
SET from
CuI or [W10O32
H
H
12
390 nm Kessil lamp, 20–30 °C
5–
]
Detected under
Consistent with
EPR data
reaction conditions
Lewis acids
Fe3+ Zn2+ Sc3+ Gd3+
Copper sources
Low-valent metals
Addition spectrum:
(CuI + Togni)
+
CuI + Togni
Ti3+
Cr2+
Fe2+
Ni2+
All 14 compounds
evaluated
(t = 0 min)
Y3+
La3+ Eu3+ Hf3+
<3% yield
Mn2+
Co2+
>60% yield
<3% yield
Standard reaction
t = 0 min
Standard reaction
t = 120 min
b
1.15 equiv. CuIOTf
3 equiv. TFA
F3C
I
O
CF3
N
H
CuIII(CF3)x
Standard reaction
O
t = 120 min
6:4 D2O/CD3CN
N
H
12
W
Observed
by 19F NMR
1 equiv.
20% yield
1,500 2,000 2,500 3,000 3,500 4,000 4,500
1,500 2,000 2,500 3,000 3,500 4,000 4,500
B (G)
B (G)
d
H
CF3
H
CF3
Ligated-copper experiments
Cu
Cu
Me
Me
Modulation of selectivity supports
copper-mediated bond formation
*
Cu
W
NH2
NH2
Ac
Ac
O
N
Alk
CuI
CF3
61
38
62
63
CuIII(CF3)x
Alk
[Cu•LS]2+
> 20:1 d.r.
8:1 d.r.
[Cu•LS]2+
+5% e.e.
–6% e.e.
N
iPr
LS
Cu2+
[Cu•LR]
2+
Fig. 5 | Mechanistic investigations into the role of copper. Studies were conducted into the role of copper in this dual-catalytic transformation. a, Control
experiments indicated that copper is unique in its effectiveness when compared with water-stable Lewis acids or other metals known to reduce Togni
reagent II. b, Competent Cu–CF3 species are formed in stoichiometric experiments. The combination of Togni reagent II with copper(i) triflate under
reaction-relevant conditions affords a mixture containing a Cu(iii)–CF3 complex observable by 19F NMR. This mixture was shown to be a competent
replacement for the copper(ii) chloride and trifluoromethyl source employed under our optimized reaction conditions. c, The same CuIII(CF3)x species is
also observed under standard reaction conditions and exists in equilibrium with an EPR-active copper(ii) compound, provisionally assigned as CuII(CF3)x
based on a reported equilibrium for a related Cu(iii)–CF3 species59. A simple linear combination of the EPR signal for solvated Cu2+ (standard reaction
before irradiation) and the ‘Cu(ii)–CF3’ complex derived from the reaction of copper(i) with Togni reagent II shows excellent agreement with the
spectrum for the standard reaction at partial conversion. d, Significant improvement in diastereoselectivity, as well as moderate but reproducible
enantioinduction, is observed when employing an iPr-QuinOx ligand (LS or LR, absolute stereochemistry indicated in subscript). The modulation of
selectivity as a function of copper ligand environment supports a mechanism in which the bond-forming event is mediated by copper. Asterisk indicates
the stereocentre at which enantioenrichment is observed.
standard C(sp3)–H trifluoromethylation conditions (Supplementary of the existence of a reactive Cu–CF3 species under our standard
Figs. 22 and 29), clearly indicating that Cu–CF3 complexes are pres- conditions, and salient evidence that copper is involved in the key
ent throughout the course of the reaction. Furthermore, the EPR bond-forming step.
signal previously assigned to CuII(CF3)x was shown to develop rap-
In summary, the merger of decatungstate photocatalysis with
idly (<2min; Supplementary Fig. 32) and remain constant during copper catalysis has enabled the development of a general C(sp3)–H
irradiation (Supplementary Fig. 33), as evidenced by the excellent trifluoromethylation using a widely available CF3 source and requir-
fit of the partial conversion EPR spectrum provided by a simple ing only a single equivalent of the C(sp3)–H-bearing substrate. This
linear combination of the spectrum prior to irradiation and the method provides a valuable approach to biorelevant, synthetically
spectrum of the reaction of CuOTf with Togni reagent II (Fig. 5c). challenging aliphatic trifluoromethyl compounds, as evidenced by
The loss of this signal upon reaction completion is again consistent its effective application to the direct C–H trifluoromethylation of
with a mechanism involving a Cu–CF3 intermediate. Finally, and natural products and medicinal agents. Mechanistic studies have
perhaps most compelling, a significant improvement in diastereose- provided evidence that C(sp3)–CF3 bond formation is mediated by
lectivity (from 8:1 to >20:1 d.r.) and reproducible enantioinduction copper, a finding with significant implications for the broad field
(5–6% e.e.) were observed when employing a chiral QuinOx ligand60 of copper-catalysed trifluoromethylations employing hypervalent
(Fig. 5d). These results reveal that copper is involved in the criti- iodine-based reagents. Furthermore, the successful development of
cal C(sp3)–CF3 bond-forming process, consistent with the radical a decatungstate/copper dual-catalytic transformation illustrates the
capture/reductive elimination sequence proposed in Fig. 2. In sum- generality of decatungstate metallaphotoredox catalysis, suggesting
mary, these studies provide significant insight into the nature of the that this platform could enable a number of other elusive C(sp3)–H
interaction between copper and Togni reagent II, definitive proof engagement protocols.
NATuRE CHEMiSTRY | www.nature.com/naturechemistry

NaTure CHemisTry
Articles
27. De Waele, V., Poizat, O., Fagnoni, M., Bagno, A. & Ravelli, D. Unraveling the
key features of the reactive state of decatungstate anion in hydrogen atom
transfer (HAT) photocatalysis. ACS Catal. 6, 7174–7182 (2016).
28. Tzirakis, M. D., Lykakis, I. N. & Orfanopoulos, M. Decatungstate as an
efcient photocatalyst in organic chemistry. Chem. Soc. Rev. 38,
2609–2621 (2009).
Online content
Any Nature Research reporting summaries, source data, extended data,
supplementary information, acknowledgements, peer review information; details
of author contributions and competing interests; and statements of data and code
availability are available at https://doi.org/10.1038/s41557-020-0436-1.
29. Schultz, D. M. et al. Oxyfunctionalization of the remote C–H bonds of
aliphatic amines by decatungstate photocatalysis. Angew. Chem. Int. Ed. 56,
15274–15278 (2017).
Received: 30 May 2019; Accepted: 3 February 2020;
Published: xx xx xxxx
30. Laudadio, G. et al. Selective C(sp3)−H aerobic oxidation enabled by
decatungstate photocatalysis in fow. Angew. Chem. Int. Ed. 57,
4078–4082 (2018).
31. West, J. G., Huang, D. & Sorensen, E. J. Acceptorless dehydrogenation of
small molecules through cooperative base metal catalysis. Nat. Commun. 6,
10093 (2015).
32. Halperin, S. D., Fan, H., Chang, S., Martin, R. E. & Britton, R. A convenient
photocatalytic fuorination of unactivated C–H bonds. Angew. Chem. Int. Ed.
53, 4690–4693 (2014).
33. Zwick, C. R. III & Renata, H. Evolution of biocatalytic and chemocatalytic
C–H functionalization strategy in the synthesis of manzacidin C. J. Org.
Chem. 83, 7407–7415 (2018).
References
1. Meanwell, N. A. Fluorine and fuorinated motifs in the design and
application of bioisosteres for drug design. J. Med. Chem. 61,
5822–5880 (2018).
2. Furet, P. et al. Discovery of NVP-BYL719 a potent and selective
phosphatidylinositol-3 kinase α inhibitor selected for clinical evaluation.
Bioorg. Med. Chem. Lett. 23, 3741–3748 (2013).
3. Alonso, C., Martínez de Marigorta, E., Rubiales, G. & Palacios, F. Carbon
trifuoromethylation reactions of hydrocarbon derivatives and heteroarenes.
Chem. Rev. 115, 1847–1935 (2015).
4. Charpentier, J., Früh, N. & Togni, A. Electrophilic trifuoromethylation by use
of hypervalent iodine reagents. Chem. Rev. 115, 650–682 (2015).
5. Ji, Y. et al. Innate C–H trifuoromethylation of heterocycles. Proc. Natl Acad.
Sci. USA 108, 14411–14415 (2011).
34. Ravelli, D., Protti, S. & Fagnoni, M. Decatungstate anion for photocatalyzed
‘window ledge’ reactions. Acc. Chem. Res. 49, 2232–2242 (2016).
35. Ravelli, D., Fagnoni, M., Fukuyama, T., Nishikawa, T. & Ryu, I. Site-selective
C–H functionalization by decatungstate anion photocatalysis: synergistic
control by polar and steric efects expands the reaction scope. ACS Catal. 8,
701–713 (2018).
6. Nagib, D. A. & MacMillan, D. W. C. Trifuoromethylation of arenes
and heteroarenes by means of photoredox catalysis. Nature 480,
224–228 (2011).
7. Wang, X. et al. Copper-catalyzed C(sp3)–C(sp3) bond formation using a
hypervalent iodine reagent: an efcient allylic trifuoromethylation.
J. Am. Chem. Soc. 133, 16410–16413 (2011).
36. Perry, I. B. et al. Direct arylation of strong aliphatic C–H bonds. Nature 560,
70–75 (2018).
37. Blakemore, D. C. et al. Organic synthesis provides opportunities to transform
drug discovery. Nat. Chem. 10, 383–394 (2018).
8. Parsons, A. T. & Buchwald, S. L. Copper-catalyzed trifuoromethylation of
unactivated olefns. Angew. Chem. Int. Ed. 50, 9120–9123 (2011).
9. Xiao, H. et al. Copper-catalyzed late-stage benzylic C(sp3)–H
trifuoromethylation. Chem 5, 940–949 (2019).
38. Song, K.-S., Liu, L. & Guo, Q.-X. Efects of α-ammonium, α-phosphonium
and α-sulfonium groups on C–H bond dissociation energies. J. Org. Chem.
68, 4604–4607 (2003).
39. Lee, M. & Sanford, M. S. Palladium-catalyzed, terminal-selective C(sp3)–H
oxidation of aliphatic amines. J. Am. Chem. Soc. 137, 12796–12799 (2015).
40. Howell, J. M., Feng, K., Clark, J. R., Trzepkowski, L. J. & White, M. C.
Remote oxidation of aliphatic C–H bonds in nitrogen-containing molecules.
J. Am. Chem. Soc. 137, 14590–14593 (2015).
10. Paeth, M. et al. Copper-mediated trifuoromethylation of benzylic Csp3–H
bonds. Chem. Eur. J. 24, 11559–11563 (2018).
11. Guo, S., AbuSalim, D. I. & Cook, S. P. Aqueous benzylic C–H
trifuoromethylation for late-stage functionalization. J. Am. Chem. Soc. 140,
12378–12382 (2018).
41. Lee, M. & Sanford, M. S. Remote C(sp3)–H oxygenation of protonated
aliphatic amines with potassium persulfate. Org. Lett. 19, 572–575 (2017).
42. Tan, X. et al. Silver-catalyzed decarboxylative trifuoromethylation of aliphatic
carboxylic acids. J. Am. Chem. Soc. 139, 12430–12433 (2017).
43. Shen, H. et al. Trifuoromethylation of alkyl radicals in aqueous solution.
J. Am. Chem. Soc. 139, 9843–9846 (2017).
12. Liu, Z. et al. Copper‐catalyzed remote C(sp3)−H trifuoromethylation of
carboxamides and sulfonamides. Angew. Chem. Int. Ed. 58, 2510–2513 (2019).
13. Twilton, J. et al. Te merger of transition metal and photocatalysis. Nat. Rev.
Chem. 1, 0052 (2017).
14. Shaw, M. H., Shurtlef, V. W., Terrett, J. A., Cuthbertson, J. D. & MacMillan,
D. W. C. Native functionality in triple catalytic cross-coupling: sp3 C-H bonds
as latent nucleophiles. Science 352, 1304–1308 (2016).
44. Paeth, M. et al. Csp3–Csp3 bond-forming reductive elimination from
well-defned copper(iii) complexes. J. Am. Chem. Soc. 141, 3153–3159 (2019).
45. Texier, I., Delaire, J. A. & Giannotti, C. Reactivity of the charge transfer
excited state of sodium decatungstate at the nanosecond time scale.
Phys. Chem. Chem. Phys. 2, 1205–1212 (2000).
15. Heitz, D. R., Tellis, J. C. & Molander, G. A. Photochemical nickel-catalyzed
C–H arylation: synthetic scope and mechanistic investigations. J. Am. Chem.
Soc. 138, 12715–12718 (2016).
16. Shields, B. J. & Doyle, A. G. Direct C(sp3)–H cross coupling enabled by
catalytic generation of chlorine radicals. J. Am. Chem. Soc. 138,
12719–12722 (2016).
46. Khan, M. A. & Schwing-Weill, M. J. Stability and electronic spectra of the
copper(ii) chloro complexes in aqueous solutions. Inorg. Chem. 15,
2202–2205 (1976).
47. Kochi, J. K., Bemis, A. & Jenkins, C. L. Mechanism of electron transfer
oxidation of alkyl radicals by copper(ii) complexes. J. Am. Chem. Soc. 90,
4616–4625 (1968).
17. Shen, Y., Gu, Y. & Martin, R. sp3 C–H arylation and alkylation enabled by the
synergy of triplet excited ketones and nickel catalysts. J. Am. Chem. Soc. 140,
12200–12209 (2018).
18. McLean, E. B. & Lee, A.-L. Dual copper- and photoredox-catalysed reactions.
Tetrahedron 74, 4881–4902 (2018).
19. Ferraudi, G. Photochemical generation of metastable methylcopper
complexes. Oxidation–reduction of methyl radicals by copper complexes.
Inorg. Chem. 17, 2506–2508 (1978).
48. Jenkins, C. L. & Kochi, J. K. Solvolytic routes via alkylcopper intermediates in
the electron-transfer oxidation of alkyl radicals. J. Am. Chem. Soc. 94,
843–855 (1972).
49. Surry, D. S. & Buchwald, S. L. Diamine ligands in copper-catalyzed reactions.
Chem. Sci. 1, 13–31 (2010).
20. Freiberg, M. & Meyerstein, D. Reactions of aliphatic free radicals with copper
cations in aqueous solution. Part 2. Reactions with cupric ions: a pulse
radiolysis study. J. Chem. Soc. Faraday Trans. 1, 1825–1837 (1980).
21. Navon, N., Golub, G., Cohen, H. & Meyerstein, D. Kinetics and reaction
mechanisms of copper(i) complexes with aliphatic free radicals in
aqueous solutions. A pulse-radiolysis study. Organometallics 14,
5670–5676 (1995).
22. Casitas, A. & Ribas, X. Te role of organometallic copper(iii) complexes in
homogeneous catalysis. Chem. Sci. 4, 2301–2318 (2013).
23. Liang, Y., Zhang, X. & MacMillan, D. W. C. Decarboxylative sp3 C-N coupling
via dual copper and photoredox catalysis. Nature 559, 83–88 (2018).
24. Kautzky, J. A., Wang, T., Evans, R. W. & MacMillan, D. W. C. Decarboxylative
trifuoromethylation of aliphatic carboxylic acids. J. Am. Chem. Soc. 140,
6522–6526 (2018).
50. Grynkiewicz, G. & Gadzikowska, M. Tropane alkaloids as medicinally useful
natural products and their synthetic derivatives as new drugs. Pharmacol.
Rep. 60, 439–463 (2008).
51. Yarmolchuk, V. S. et al. Synthesis and characterization of β‐trifuoromethyl‐
substituted pyrrolidines. Eur. J. Org. Chem. 2013, 3086–3093 (2013).
52. Jacobs, R. T. et al. Synthesis, structure–activity relationships, and
pharmacological evaluation of a series of fuorinated 3-benzyl-5-
indolecarboxamides: identifcation of 4-[[5-[((2R)-2-methyl-4,4,4-
trifuorobutyl)carbamoyl]-1-methylindol-3-yl]methyl]-3-methoxy-N-[(2-
methylphenyl)sulfonyl]benzamide, a potent orally active antagonist of
leukotrienes D4 and E4. J. Med. Chem. 37, 1282–1297 (1994).
53. Blakemore, D. C. et al. Synthesis and in vivo evaluation of 3-substituted
gababutins. Bioorg. Med. Chem. Lett. 20, 362–365 (2010).
54. Zhang, Z. & Tang, W. Drug metabolism in drug discovery and development.
Acta Pharm. Sin. B 8, 721–732 (2018).
25. Kornflt, D. J. P. & MacMillan, D. W. C. Copper-catalyzed trifuoromethylation
of alkyl bromides. J. Am. Chem. Soc. 141, 6853–6858 (2019).
26. Duncan, D. C. & Fox, M. A. Early events in decatungstate photocatalyzed
oxidations: a nanosecond laser transient absorbance reinvestigation. J. Phys.
Chem. A 102, 4559–4567 (1998).
55. McLean, T. H. et al. 1-Aminomethylbenzocycloalkanes: conformationally
restricted hallucinogenic phenethylamine analogues as functionally selective
5-HT2A receptor agonists. J. Med. Chem. 49, 5794–5803 (2006).
NATuRE CHEMiSTRY | www.nature.com/naturechemistry

NaTure CHemisTry
Articles
56. Fukuyama, T. et al. Photocatalyzed site-selective C(sp3)–H functionalization
of alkylpyridines at non-benzylic positions. Org. Lett. 19, 6436–6439 (2017).
57. Kobayashi, S., Nagayama, S. & Busujima, T. Lewis acid catalysts stable in
water. Correlation between catalytic activity in water and hydrolysis constants
and exchange rate constants for substitution of inner-sphere water ligands.
J. Am. Chem. Soc. 120, 8287–8288 (1998).
58. Ling, L., Liu, K., Li, X. & Li, Y. General reaction mode of hypervalent iodine
trifuoromethylation reagent: a density functional theory study. ACS Catal. 5,
2458–2468 (2015).
59. Nebra, N. & Grushin, V. V. Distinct mechanism of oxidative
trifuoromethylation with a well-defned Cu(ii) fuoride promoter: hidden
catalysis. J. Am. Chem. Soc. 136, 16998–17001 (2014).
60. Yang, G. & Zhang, W. Renaissance of pyridine-oxazolines as chiral ligands for
asymmetric synthesis. Chem. Soc. Rev. 47, 1783–1810 (2018).
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in
published maps and institutional affiliations.
© The Author(s), under exclusive licence to Springer Nature Limited 2020
NATuRE CHEMiSTRY | www.nature.com/naturechemistry

NaTure CHemisTry
Articles
Data availability
Author contributions
The data supporting the findings of this study are available within the article and its
Supplementary Information.
P.J.S. and V.B. performed and analysed the experiments. P.J.S., V.B., D.M.S., D.A.D. and
D.W.C.M. designed the experiments. Y.-h.L. and E.C.S. performed the computational
analysis. P.J.S., V.B. and D.W.C.M. prepared the manuscript.
Acknowledgements
We acknowledge financial support from the National Institute of General Medical
Sciences (NIGMS), the NIH (award no. R01 GM078201-05) to D.W.C.M., V.B. and
P.J.S., and gifts from MSD, Abbvie, Pfizer and Janssen. P.J.S. and V.B. acknowledge
Princeton University, E. Taylor and the Taylor family for an Edward C. Taylor Fellowship.
We thank L. Wilson (Lotus Separations) for assistance with compound purification,
D. Abrams for helpful discussions, I. Pelczer for assistance with NMR spectroscopy,
O. Garry for synthesizing a batch of sodium decatungstate, I. Perry for providing a
graphical rendering of the decatungstate anion and R. Martinie, H. Zhong and J. Eng for
assistance with EPR spectroscopy. The content is solely the responsibility of the authors
and does not necessarily represent the official views of NIGMS.
Competing interests
The authors declare no competing interests.
Additional information
Supplementary information is available for this paper at https://doi.org/10.1038/
s41557-020-0436-1.
Correspondence and requests for materials should be addressed to D.W.C.M.
Reprints and permissions information is available at www.nature.com/reprints.
NATuRE CHEMiSTRY | www.nature.com/naturechemistry