Enantioselective radical C–H amination for the synthesis of β-amino alcohols

Article abstract of DOI:10.1038/s41557-020-0482-8

Asymmetric, radical C–H functionalizations are rare but powerful tools for solving modern synthetic challenges. Specifically, the enantio- and regioselective C–H amination of alcohols to access medicinally valuable chiral β-amino alcohols remains elusive. To solve this challenge, a radical relay chaperone strategy was designed, wherein an alcohol was transiently converted to an imidate radical that underwent intramolecular H-atom transfer (HAT). This regioselective HAT was also rendered enantioselective by harnessing energy transfer catalysis to mediate selective radical generation and interception by a chiral copper catalyst. The successful development of this multi-catalytic, asymmetric, radical C–H amination enabled broad access to chiral β-amino alcohols from a variety of alcohols containing alkyl, allyl, benzyl and propargyl C–H bonds. Mechanistic experiments revealed that triplet energy sensitization of a Cu-bound radical precursor facilitates catalyst-mediated HAT stereoselectivity, enabling the synthesis of several important classes of chiral β-amines by enantioselective, radical C–H amination. [Figure not available: see fulltext.]

Full text of DOI:10.1038/s41557-020-0482-8

Articles
https://doi.org/10.1038/s41557-020-0482-8
Enantioselective radical C–H amination for
the synthesis of β-amino alcohols
Kohki M. Nakafukuꢀ ꢀ1,2, Zuxiao Zhangꢀ ꢀ1,2, Ethan A. Wappes1,2, Leah M. Stateman1, Andrew D. Chen1
1
ꢀ✉
and David A. Nagibꢀ ꢀ
Asymmetric, radical C–H functionalizations are rare but powerful tools for solving modern synthetic challenges. Specifically,
the enantio- and regioselective C–H amination of alcohols to access medicinally valuable chiral β-amino alcohols remains elu-
sive. To solve this challenge, a radical relay chaperone strategy was designed, wherein an alcohol was transiently converted to
an imidate radical that underwent intramolecular H-atom transfer (HAT). This regioselective HAT was also rendered enanti-
oselective by harnessing energy transfer catalysis to mediate selective radical generation and interception by a chiral copper
catalyst. The successful development of this multi-catalytic, asymmetric, radical C–H amination enabled broad access to chiral
β-amino alcohols from a variety of alcohols containing alkyl, allyl, benzyl and propargyl C–H bonds. Mechanistic experiments
revealed that triplet energy sensitization of a Cu-bound radical precursor facilitates catalyst-mediated HAT stereoselectivity,
enabling the synthesis of several important classes of chiral β-amines by enantioselective, radical C–H amination.
ive of the six most commonly employed chemical reactions
in the discovery of new medicines entail the construction of
a C–N bond1. To supplement these classical methods for ami-
Results and discussion
Design plan. In designing a HAT-mediated synthesis of chiral
β-amino alcohols via radical C–H amination, we noted several
challenges, as outlined in Fig. 1b. First, the α-C–H of an alcohol
is weaker than the β- and γ-methylene C–H bonds (bond disso-
ciation energy: 90 versus 98kcalmol−1)34. We have shown that a
radical relay chaperone strategy, wherein an alcohol is temporarily
converted to an imidate radical, facilitates selective β-radical forma-
tion via 1,5-HAT to overcome this thermodynamic bias16. Yet, this
approach has only been rendered racemic because of its reliance on
non-selective, iodine atom transfer to generate and trap the requi-
site radicals. Alternatively, we envisioned that a chiral catalyst could
mediate asymmetric C–H functionalization via an enantioselective
HAT only if radical generation is dependent on the presence of a
chiral catalyst, such as through a multi-catalytic, triplet sensitiza-
tion mechanism. Moreover, the low barrier for inversion of alkyl
radicals (0.5kcalmol−1; >108 s−1)35 necessitates that this β-radical
also be rapidly and stereoselectively trapped by the chiral catalyst.
In this manner, and in contrast with concerted, nitrene insertion
mechanisms, we envisioned that a radical-mediated approach could
enable stereoselectivity by exercizing catalyst control over two ele-
mentary steps: (1) H-atom abstraction; and (2) alkyl radical trap-
ping. We anticipated that catalytic stereocontrol over both of these
radical-generating and -terminating events would provide highly
enantioenriched β-amino alcohols.
F
nation of preformed C–X bonds, modern synthetic efforts have
focused on designing strategies for direct C–H amination2,3. To
this end, aminations of aryl4–6 and alkyl7–9 C–H bonds have been
developed. In the latter cases, stereochemical control of sp3 C–H
amination affords an added challenge, with rare solutions entailing
nitrenes and heavy metals, porphyrins or enzymes10–15. Nonetheless,
the synthesis of β-amino alcohols by enantio- and regioselective
β-C–H amination remains an unsolved problem. Here, we report
a new radical relay chaperone strategy, wherein an alcohol is tran-
siently converted to an imidate radical16 that enables intramolecular
H-atom transfer (HAT)17. This regioselective 1,5-HAT18,19 has now
also been rendered enantioselective—a rare example for C–H ami-
nation—by harnessing energy transfer catalysis to mediate selective
radical generation and interception by a chiral copper catalyst20–22
.
To selectively prepare chiral β-amino alcohols—a privileged
motif in nature, medicine and catalysis—we sought to develop an
enantio- and regioselective, radical C–H amination of alcohols,
as shown in Fig. 1a. This strategy complements current methods
to access vicinal amino alcohols that rely on the chiral pool (for
example, amino acids), chiral auxiliaries (for example, imines) or
multi-step conversion of chiral diols23–25. However, we were cog-
nizant of the dearth of methods for stereoselective, radical C–H
amination. In fact, asymmetric radical C–H functionalizations in
general are rare, with pioneering solutions existing only for alkyla- Proposed mechanism. The chemo-, regio- and enantioselective
tion, arylation and deracemization26–32. Our analysis of this problem radical C–H amination described herein is mediated by multiple
led us to propose that the challenge lies in the typical generation catalysts working in concert to enable this valuable, asymmetric
of radicals by non-selective H abstraction of weak C–H bonds (for transformation via an energy transfer mechanism, as illustrated in
example, benzylic and α-hetero), leading to radical chain processes Fig. 1c. In this strategy, an alcohol is converted to an oxime imidate
that are difficult to render stereoselective—especially in the case by combination with a bench-stable imidoyl chloride chaperone
of C–H amination33. We proposed that if these carbon radicals are prepared in two steps from H2N–OPh36. This radical precursor is
instead mildly generated by intramolecular HAT from N-centred then coordinated by copper catalyst A bearing a chiral bisoxazoline
radicals, catalytic systems could be engineered to exercize complete ligand, as well as a chiral carboxylate, to generate Cu imidate com-
regio- and stereoselective control over radical C–H amination.
plex B. Upon excitation with visible light, an Ir photocatalyst may
1Department of Chemistry and Biochemistry, The Ohio State University, Columbus, OH, USA. 2These authors contributed equally: Kohki M. Nakafuku,
✉
Zuxiao Zhang, Ethan A. Wappes. e-mail: nagib.1@osu.edu
NAtuRE CHEMiStRy | www.nature.com/naturechemistry

NATure CHeMiSTry
Articles
a
Classic
NH2
NH2
*
NSOR′
OH
*
Asymmetric
C–H amination
approaches
HO
R
HO
O
HO
R
R
R
R
*
*
OH
Amino alcohols
(medicines, catalysts)
Alcohols
(Feedstock)
Amino acids
(Chiral pool)
Imines
(Chiral auxiliary)
Diols
(Limitations)
b
Goal
Challenges
Design
CuL*
N
H
Regioselectivity
Enantioselectivity
R2N
HO
H
H
H
HO
R
H
HO
O
R
R
R
= weakest C–H bond
= similar C–H bonds
Radical
Regio- and
enantioselective HAT
Racemizable
>108 s–1
C–H amination
c
d
1% Ir1·BArF
4
hv
Ln*CuI
2% CuBArF
Ir*
4
Ph
Cu
Ph
N
4% chiral ligand L1
25% chiral acid
N
N
Energy
transfer
OPh
Ph
coordination
R′Cl
OPh
OPh
N
O
HO
O
R
O
O
Ir
2:1 pentane: Et2O
Blue LED, 23 °C
R
R
Ph
B
1
Imidate
Controls
Ln*CuIIOPh
N
No photocatalyst
No conversion
Asymmetric
C–H amination
via radical relay
L*
HA*
No ligand L1
20%, 0% e.e.
CuIBArF
Ln*CuI
Counterions
4
chaperone catalysis
O
Cu(OTf)2 versus CuBArF
Ir1·PF6 versus Ir1·BArF
87%, 83% e.e.
94%, 92% e.e.
4
R
A
4
C
Acids
PhOH
Trifluoroacetic acid
Phosphoric acid
None
41%, 18% e.e.
15%, 36% e.e.
16%, 85% e.e.
75%, 83% e.e.
84%, 92% e.e.
93%, 93% e.e.
95%, 94% e.e.
Ln*CuIIOPh
N
H+
Regio- and
enantioselective HAT
NH2
N
*
Acetic acid
O
Stereoselective
amination
Adamantyl acid
(–)-Camphoric acid
(+)-Camphoric acid
HO
O
R
R
R
*
D
Oxazoline
Fig. 1 | Design of a selective, radical C–H amination for the synthesis of β-amino alcohols. a, Radical C–H amination of cheap and abundant alcohols
enables access to the privileged β-amino alcohol architecture, complementing classic approaches that are limited by the availability of chiral pool
precursors (for example, amino acids) or stoichiometric chiral auxiliaries (for example, imines) or multi-step manipulation of chiral reagents (for example,
diols). b, Dual challenges for a radical approach include selective formation and trapping of a transient radical intermediate. Specifically, β-regioselectivity
is disfavoured compared with abstraction of the weaker α-C–H bond by an N-centred radical. Moreover, enantioselectivity is challenging to both control
and retain in this process, as it may rapidly erode by radical epimerization. c, Proposed mechanism. Alcohol addition to an imidoyl chloride chaperone
furnishes an oxime imidate, which binds to a chiral Cu catalyst. This complex then forms an N-centred radical via selective, triplet sensitization by
an excited Ir photocatalyst. Regio- and enantioselective HAT, followed by stereoselective amination, affords a chiral oxazoline. Hydrolysis yields the
enantioenriched β-amino alcohol. d, The development of this asymmetric β-C–H amination required a multi-catalytic strategy, wherein a photocatalyst
selectively excites a chiral Cu catalyst complex when it is bound to an imidate-activated alcohol. Mechanistic probes illustrate the necessity of each
component in ensuring the efficiency and selectivity of this radical C–H amination.
then sensitize this organocopper complex via an energy transfer multi-catalytic approach that includes chiral copper and acid
mechanism37, which generates N-centred radical C by triplet sensi- co-catalysts working in concert with an energy transfer photocat-
tization of the oxime N–O bond to enable a net photoinduced oxida- alyst to selectively initiate and trap an intramolecular HAT mech-
tive addition. The resulting chiral Cu-bound imidate radical C can anism. Specifically, the benzoic acid-derived oxime imidate of
undergo enantio- and regioselective HAT to afford chiral β-radical 2-phenyl ethanol was converted to chiral oxazoline 1 (95% yield;
D. Amination may then occur with further stereocontrol by either: 94% e.e.) by combination of 1% Ir1 photocatalyst (Ir{[dF(CF3)
(1) diastereoselective metalation of the alkyl radical38 and reduc- ppy]2dtbbpy}BArF ), 2% CuBArF , 4% bisoxazoline ligand L1
4
tive elimination of the ensuing alkyl Cu(iii) intermediate39; or (2) and 25% camphoric acid in a 2:14mixture of pentane:Et2O, with
diastereoselective coupling of the radical within a Cu(ii) complex40. blue LED irradiation for 1 h. Although asymmetric, α-amido
Both pathways would afford enantioenriched oxazoline, along with amination has been shown by direct photoexcitation of Cu car-
PhOH and regenerated Cu catalyst A. Finally, acidic hydrolysis fur- bazolides40, neither Cu acetate A nor Cu imidate B absorbs vis-
nishes the chiral β-amino alcohol from this multi-catalytic radical ible light (λabs > 400 nm). Therefore, an external triplet sensitizer
relay mechanism.
(Ir1) was used to mediate this reaction with blue LED irradia-
tion. Control experiments confirmed that this photosensitizer
Reaction development. The execution of our design for asym- was required (starting material is unconsumed without Ir1) and
metric β-C–H amination, as shown in Fig. 1d, requires a that other non-redox active, triplet sensitizers increased the reac-
NAtuRE CHEMiStRy | www.nature.com/naturechemistry

NATure CHeMiSTry
Articles
Table 1 | Synthetic scope of alcohols in enantioselective, radical C–H amination
1% Ir1·BArF
4
OPh
nPr nPr
2% CuBArF
4
N
4% ligand L1
25% camphoric acid
Ph
O
O
O
Ph
N
NH2
HCl
Ph
Cl
NaH
OPh
HO
N
N
N
R
HO
O
R
2:1 pentane: Et2O
Blue LED, 23 °C
R
R
Ph
Ph
L1
Ph
Ph
Ph
Ph
Ph
R = H, 1, 80%, 94% e.e.
F, 2, 98%, 93% e.e.
Cl, 3, 87%, 94% e.e.
Br, 4, 89%, 95% e.e.
I, 5, 79%, 93% e.e.
O
N
O
N
O
N
O
N
O
N
Me
CF3
Me, 6, 89%, 91% e.e.
OMe, 7, 80%, 84% e.e.
CF3, 8, 70%, 95% e.e.
OCF3, 9, 94%, 92% e.e.
OMe
CF3
R
10, 95%, 91% e.e.
11, 81%, 97% e.e.
12, 79%, 93% e.e.
13, 85%, 91% e.e.
Ph
Ph
Ph
Ph
Ph
Ph
O
N
O
N
O
N
O
N
O
N
O
N
Cl
CF3
Cl
Cl
Cl
F
OMe
CF3
14, 88%, 98% e.e.
15, 87%, 93% e.e.
16, 90%, 89% e.e.
17, 79%, 80% e.e.
18, 92%, 93% e.e.
19, 99%, 94% e.e.
Ph
Ph
Ph
Ph
Ph
Ph
O
N
O
N
O
N
O
N
O
N
O
N
Me
S
S
N
20, 65%, 82% e.e.
21, 83%, 90% e.e.
22, 92%, 95% e.e.
23, 82%, 94% e.e.
24, 80%, 83% e.e.
25, 99%, 89% e.e.
Ph
Ph
Ph
Ph
Ph
Ph
O
N
O
N
O
N
O
N
O
N
O
N
OMe
Me Me
Me
Me Me
Me
Me
Me
26, 96%, 83% e.e.
Ph
27, 89%, 83% e.e.
28, 36%, 89% e.e.
29, 47%, 94% e.e.
30, 86%, 88% e.e.
31, 50%, 79% e.e.
Conditions: 1% Ir{[dF(CF3)ppy]2dtbbpy}BArF4, 2% CuBArF4, 4% ligand L1, 25% camphoric acid and pentane:Et2O (2:1) under blue LED irradiation at room temperature for 1ꢀh. See general procedure 3 in
Supplementary Information section II for full experimental details. Isolated yield and e.e. values are indicated below each entry.
tion efficiency under ultraviolet irradiation (for example, naph- data indicate that a bulky—albeit not necessarily chiral—acid is
thalene and xanthone; see Supplementary Information section required, probably to stabilize the higher oxidation state of Cu
VIII ‘Mechanistic experiments’). Additionally, chiral ligand L1 intermediates.
impacted both the reaction efficiency and selectivity (20% yield
and 0% e.e. in its absence), and the BArF counterion improved Synthetic scope. To probe the synthetic utility of this enantioselec-
stereocontrol (83–94% e.e.)41. Finally, the a4cid co-catalyst played a tive β-C–H amination, we subjected a range of sterically and elec-
pivotal role in controlling efficiency and selectivity. For example, tronically diverse alcohols to this catalytic, radical relay sequence.
acetic acid notably improved the yield relative to the absence of As illustrated in Table 1, the HAT mechanism was robust and fur-
acid or the use of strong acids, whose anions were poorly coor- nished β-amino variants of a wide range of para-substituted 2-aryl
dinating (75% versus 16% yield). Notably, bulkier acids (for ethanols (1–9)—spanning Hammett σp constants of –0.3 (OMe) to
example, adamantyl or camphoric) afforded optimal efficiency +0.5 (CF3)—with little effect on efficiency or selectivity (70–98%;
(84–95% yield) and stereocontrol (92–94% e.e.). Together, these up to 95% e.e.). Intrigued by the potential medicinal value of these
NAtuRE CHEMiStRy | www.nature.com/naturechemistry

NATure CHeMiSTry
Articles
Table 2 | Synthetic scope and effect of chaperone in enantioselective, radical C–H amination
1% Ir1·BArF
4
OPh
2% CuBArF
nPr nPr
4
N
4% ligand L1
25% camphoric acid
Ar
O
O
O
Ar
N
NH2
HCl
Ar
Cl
NaH
OPh
HO
N
N
N
R
HO
O
R
2:1 pentane: Et2O
Blue LED, 23 °C
R
R
Ph
Ph
L1
OMe
Cl
O
O
Cl
N
O
N
O
N
O
N
O
N
O
N
Ph
Ph
Ph
Ph
Ph
Ph
37, 80%, 97% e.e.
32, 56%, 97% e.e.
33, 75%, 94% e.e.
34, 99%, 93% e.e.
35, 49%, 98% e.e.
36, 64%, 94% e.e.
versus 1, 80%, 94% e.e.
Ar
Ar
Ar
Ar
Ar
Ar
O
N
O
N
O
N
O
N
O
N
O
N
Cl
CF3
S
S
Cl
OCF3
OMe
CF3
Np: 43, 78%, 93% e.e.
Ph: 21, 83%, 90% e.e.
Np: 38, 64%, 99% e.e.
Ph: 14, 88%, 98% e.e.
Np: 39, 85%, 99% e.e.
Ph: 22, 92%, 95% e.e.
Np: 40, 77%, 97% e.e.
Ph: 9, 94%, 92% e.e.
Np: 41, 82%, 88% e.e.
Ph: 7, 80%, 84% e.e.
Np: 42, 65%, 97% e.e.
Ph: 17, 79%, 80% e.e.
Ar
Ar
Ar
Ar
Ar
Ar
O
N
O
N
O
N
O
N
O
N
O
N
Me
Me
Me Me
Me
Ph
Np: 49, 30%, 91% e.e.
Ph: 30, 86%, 88% e.e.
Np: 44, 82%, 94% e.e.
Ph: 23, 82%, 94% e.e.
Np: 45, 91%, 87% e.e.
Ph: 24, 80%, 83% e.e.
Np: 46, 87%, 91% e.e.
Ph: 25, 99%, 89% e.e.
Np: 47, 92%, 84% e.e.
Ph: 26, 96%, 83% e.e.
Np: 48, 96%, 90% e.e.
Ph: 27, 89%, 83% e.e.
Conditions: 1% Ir{[dF(CF3)ppy]2dtbbpy}BArF4, 2% CuBArF4, 4% ligand L1, 25% camphoric acid and pentane:Et2O (2:1) under blue LED irradiation at room temperature for 1ꢀh. See general procedure 3 in
Supplementary Information section II for full experimental details. Isolated yield and e.e. values are indicated below each entry. 1-naphthyl afforded improved enantioselectivity in many cases.
Np, 1-naphthyl; Ph, phenyl; Ar, aryl.
β-amino alcohols (for example, 7 is an isomeric analogue of the 2-alkyl ethanols with large β-substituents afforded enantioselectivi-
neurotransmitter octopamine, wherein the N and O atoms of the ties as high as the 2-aryl ethanols (28–30; up to 94% e.e.), suggest-
β-amino alcohol are inverted), we then explored various ortho and ing that the role of the aryl group is merely to act as a sterically
meta substitutions (10–13) as well as disubstituted arenes (14–17), repulsive, catalyst-recognition element. When the β-substituent
which were also smoothly aminated. Similarly, alcohols contain- size was decreased (for example, to an iPr group), the stereoselec-
ing polyaromatics (18 and 19) and heteroarenes (20–22) afforded tivity dropped (31; 79% e.e.), yet complete β-C–H regioselectivity
β-C–H amination with high enantioselectivity.
remained intact as the kinetically favoured 1,5-HAT mechanism out-
Having established robust reactivity for benzylic C–H amina- competed 1,6-HAT of the weaker γ-tertiary C–H (Δ: 2kcalmol−1)34.
tion, we next probed this strategy for alcohols with stronger allylic, Having investigated the alcohol scope exclusively using simple
propargylic and aliphatic C–H bonds. In the allylic case, a more benzimidate chaperones (derived from benzoyl chloride), we were
sterically encumbered 1,1-disubstituted alkene afforded greater interested in exploring the synthetic robustness of this β-C–H ami-
enantioselectivity than a simple vinyl substituent (23 (94% e.e.) nation in accessing other chiral oxazolines. To this end, a variety
versus 24 (83% e.e.)), although both were similarly efficient (80– of aryl imidates were prepared and subjected to the multi-catalytic
82% yield). To our surprise, the smaller sp-hybridized substituent sequence, as shown in Table 2. Interestingly, a smaller 3-furanyl
of propargylic alcohols also afforded β-C–H amination with high substituent afforded unexpectedly high selectivity (32; 97% e.e.).
selectivity (25–27; 83–89% e.e.). The observation that the smallest Other arenes with stereoelectronically diverse substitution also
ethynyl substituent afforded the greatest efficiency and selectivity afforded highly enantioenriched oxazolines (33–35; 93–98% e.e.),
among the trio suggests that these alkynes may interact with the Cu albeit with varying efficiency (49–99%). To our delight, we observed
catalyst in an additional manner. With hopes of removing all such that a 1-naphthyl substituent (37) afforded similar efficiency to the
possible coordination from the substrate, we then probed a series Ph-imidate (80% in both cases), as well as boosting stereoselectivity
of aliphatic alcohols bearing β-methylene units. Notably, simple (97% e.e. (versus 94% e.e. for 1). To further investigate this effect, we
NAtuRE CHEMiStRy | www.nature.com/naturechemistry

NATure CHeMiSTry
Articles
a
b
SET
OPh
EnT
favoured
OPh
OR
OPh
disfavoured
N
N
N
H
L*Cu
L*Cu
OPh
OR
OPh
OR
OPh
OR
N
N
N
N
Ered: –1.6 V
ET: 47 kcal
Ph
OR
Ph
Ph
OR
Triplet
N–O: –18 kcal
Ph
Ph
Ph
Ph
OR
Radical anion
N–O: 6 kcal
Imidate
N–O: 35 kcal
PCET
ET
Cu-bound ET
Ksv: 901
Cu-bound
N radical
Ksv: 1
Ksv: 430
Photocatalyst
Ered
ET
τ
Yield
Ir1
Ir2
Xanthone
–1.4 V
–1.5 V
62 kcal
49 kcal
74 kcal
80%, 94% e.e.
14%, 94% e.e.
45%, 94% e.e.
2,300 ns
Not likely
Possible
Most likely
557 ns
c
d
H2N Me
NH2
CuL*
OH
H2N H/D
Ph
OH
O
N
O
NCuL*
HO
HO
Ph
HO
NH2
5:1
>20:1
CO2Me
OMe
Kinetic isotope
effect, 51
Stereoselective HAT
of (S)-Aleve, 52
Diastereoselectivity
probe, 53
KIE: 6.1
Racemic cat: 2:1 e.r.
S,S cat: 1:3 e.r.
S,S cat: 2:1 d.r.
R,R cat: 8:1 d.r.
Achiral
alcohol
Desymmetrization by HAT
Amino alcohol, 50
5:1 e.r., >20:1 d.r.
R,R cat: 5:1 e.r.
e
f
Et
Ph
HO
HO
HO
R
HO
HO
HO
Me
Me
C12H25
Et
cis
Z
E
γ-functionality
R = CF3, 55
Ph, 56
γ-secondary
γ-tertiary
NH2
NH2
NH2
NH2
NH2
Et
HO
HO
HO
R
HO
HO
HO
Me
Me
C12H25
Et
Ph
54, 66%, 55% e.e.
>20:1β
31, 50%, 79% e.e.
>20:1β
55, 74% e.e., >20:1β
56, 59% e.e., 4:1β:γ
57, 87%, 80% e.e.
10:1 Z:E
58, 85%, 86% e.e.
20:1 E:Z
59, <10% amination
versus ring opening
Fig. 2 | Mechanistic experiments. a, An energy transfer (EnT) pathway (versus single-electron transfer (SET)) is supported by cyclic voltammetry
and density functional theory data. Specifically, a less reducing photocatalyst (ir1 versus ir2) provides greater efficiency due to a higher triplet energy
(ET) and longer-lived excited state (τ). The triplet sensitizer xanthone also affords reactivity, indicating that the imidate triplet, which is computed
to have a greatly weakened N–O bond, is the source of the N-centred radical. Ered, reduction potential. b, Photo-quenching experiments illustrate the
favourability of Cu catalyst coordination during the triplet sensitization and subsequent N radical generation. c, Desymmetrization of a secondary alcohol
demonstrates the ability of this Cu-bound N radical to discriminate among H atoms by enantioselective HAT (5:1 e.r.). High diastereoselectivity (>20:1
d.r.) illuminates the role of Cu radical trapping in further upgrading the high enantioselectivity observed for primary alcohols. d, A large KIE indicates
HAT is the product-determining step. Stereoselectivity of the HAT was further probed with two chiral alcohols. Selectivity and efficiency are highly
catalyst-dependent for both (S)-Aleve, which contains a β-stereocentre, as well as an enantioenriched secondary alcohol. cat, catalyst. e, A highly
regioselective 1,5-HAT (>20:1 β) outcompetes 1,6-HAT even when much weaker γ-C–H bonds are present. r.r., regioisomeric ratio. f, Radical clocks
indicate that radical trapping by Cu is more rapid than allyl radical isomerization, but slower than cyclopropyl ring opening.
probed the naphthyl imidates of a variety of alcohols from Table 1. electrode)36. Yet, the more efficient complex (Ir1) has a notably
In several cases, we observed near-complete enantiocontrol (38–39; higher triplet energy (62 versus 49kcalmol−1) and longer-lived
99% e.e.), with the improvement in enantioselectivity ranging from excited state (2,300 versus 557ns)42, suggesting that energy trans-
modest (38; Δ: 1% e.e.) to large (42; Δ: 17% e.e.). This effect was fer is more likely to be operative than an electron transfer pathway.
recapitulated for the vinyl-, alkynyl- and alkyl-substituted alcohols Importantly, catalyst counterion effects (BArF >PF6 >OTf)—on
4
(44–49), which were boosted to >90% e.e. in many cases.
both yield and enantioselectivity—were also observed, probably due
to increased triplet energies or lifetimes, which may facilitate more
Mechanistic investigations. The rationale for the proposed mecha- efficient energy transfer41. Concerning the proposed mechanism of
nism illustrated in Fig. 1c is provided by a combination of data from radical formation, photoinduced oxidative addition into the N–O
cyclic voltammetry, as well as computational, photo-quenching, bond of the oxime imidate was probably a result of its bond weaken-
isotopic labelling, competition, desymmetrization and radical clock ing. For example, upon catalytic triplet sensitization, the N–O bond
experiments, as shown in Fig. 2. Probing the role of the photocat- of the triplet state was notably weaker than the ground state (–18
alyst in radical initiation (Fig. 2a), we noted that a more strongly versus 35kcalmol−1), facilitating N–O scission (see Supplementary
reducing Ir complex Ir2 (–1.5V; see Supplementary Information Information section IX ‘Computational studies’ for details).
section VIII ‘Mechanistic experiments’ for details) was less effi-
The interaction of the excited Ir catalyst with the imidate was
cient than a less reducing catalyst Ir1 (–1.4V) (14% versus 80% further investigated through Stern–Volmer quenching experiments
yield), which was surprising, given that the oxime imidate reduc- (Fig. 2b). Since an acid co-catalyst was pivotal in obtaining high
tion potential is –1.6V (all potentials versus the saturated calomel efficiency and selectivity (see Fig. 1d), a proton-coupled electron
NAtuRE CHEMiStRy | www.nature.com/naturechemistry

NATure CHeMiSTry
Articles
O
a
Ph
NH2
Ph
HN
Ph
NHBz
Ph
HN
Ph
NHBz
Ph
N
O
HO
HO
O
HO
Ph
Ph
R
OH
α-amino acid
X
β-amine
via HCl, 100 °C
β-amide
via HCl, 23 °C
β-benzyl amine
iBu2AlH
β-amino halide
Si–X
Oxazoline 1
99% e.e.
via HCl; RuCl3
, NaIO4
via
via Me3
60, 77%, 98% e.s.
61, 91%, 99% e.s.
62, 47%, 99% e.s.
63, 79%, 98% e.s.
64, Cl, 88%, 98% e.s.
65, Br, 89%, 99% e.s.
66, I, 66%, 99% e.s.
Ref. 49
b
PPh2
Double C–H
amination
Inverted
bisoxazoline
68, 66%, 98% e.e.
15:1 d.r.
OPh
O
PhO
N
N
N
O
O
O
N
N
Ph
O
Ph
R
Ph
Ph
67
PHOX ligand
via nBuLi; Ph2PCl
Fig. 3 | Synthetic applications. a, The enantioenriched oxazoline intermediate is rapidly converted to a family of valuable chiral products, including amino
acids and β-amines, by acidic hydrolysis, oxidation, reduction or halide substitution. Reagents, yield and e.s. values are provided below each example (see
Supplementary Information section VII ‘Post-functionalization of oxazolines’ for full details). Alternatively, directed lithiation and phosphinylation may
afford access to phosphinooxazoline (PHOX) ligands49. b, Double C–H amination of a diol affords a new class of chiral catalysts.
transfer mechanism was first considered18. However, addition of acid and electronically diverse γ substituents—including those with
to the imidate decreased quenching (KSV: 1 versus 430), discount- weaker γ-C–H bonds—were investigated. However, β-selectivity
ing the involvement of the Brønsted acid in radical generation. In (via 1,5-HAT) was exclusively observed (>20:1) for both γ-tertiary
contrast, addition of the copper catalyst increased quenching (KSV: (31) and γ-CF3 substituents (55). Even in the case of a 10kcalmol−1
901 versus 430), suggesting that the Cu-coordinated imidate is the weaker γ-benzyl C–H bond, 4:1 β:γ regioselective amination was
most likely acceptor of triplet energy from the excited Ir photocata- observed (56).
lyst. This Cu-bound energy transfer mechanism also explains the
Since stereoselective C–H amination by HAT is much rarer
proximity of chiral information at the outset of radical generation to than nitrene insertion, we wished to investigate whether these
overcome racemic mechanisms, involving either energy transfer43 or metal-bound nitrogen radicals behave more like radicals or
Cu catalysis44, as well as to mediate a stereoselective HAT pathway.
nitrenes. If the former case were operative, we envisioned that a
To test the hypothesis that a Cu-bound N-centred radical medi- radical pathway could allow for catalyst stereocontrol over two
ates enantioselective HAT, we subjected an achiral secondary alco- steps (both HAT and alkyl radical trapping), rather than a single,
hol to the C–H amination (Fig. 2c). Given the likely irreversibility of C–H insertion step as in nitrene reactivity. Thus, a homoallylic
the HAT, the observed desymmetrization (50; 5:1 e.r.) indicated that alcohol was subjected to the amination (Fig. 2f). Interestingly, this
the coordinated chiral Cu is involved in promoting enantioselective probe exclusively afforded C–H amination (57–58; >85% yield;
HAT. The subsequent diastereoselectivity (>20:1 d.r.) illustrated >80% e.e.) without any aziridination, as is sometimes observed in
the Cu catalyst’s additional role in stereoselectively trapping this nitrene mechanisms10,11,45. Interestingly, a Z-alkene was tolerated,
intermediate. The higher enantioselectivity observed for primary and only slightly isomerized (57; 10:1 Z from 20:1 Z starting mate-
alcohols versus this secondary alcohol (see Tables 1 and 2) can be rial) despite a low inversion barrier for the allyl radical intermediate
considered a result of these additive effects.
(15.7kcalmol−1)46, indicating that radical trapping by the Cu cata-
Further insight into the key HAT mechanism (Fig. 2d) was pro- lyst was rapid. As a control, an E-alkene completely retained its con-
vided by a strong, primary kinetic isotope effect (KIE) observed figuration (58; 20:1 E). As further evidence of a radical-mediated
in the intramolecular H/D competition (51; KIE: 6.1), which indi- HAT mechanism, a β-cyclopropyl ring predominantly underwent
cated that HAT is the product-determining step. To further test ring opening or reduction (59), rather than C–H amination, unlike
the likelihood of a stereoselective HAT, an enantiopure β-chiral with Rh-nitrenes11. Combining these observations, we conclude
alcohol derived from the non-steroidal anti-inflammatory drug that alkyl radical trapping by the Cu catalyst probably occurs at
(S)-Aleve was converted to β-amine 52. A pronounced catalyst a rate between 108 s–1 (tBu radical isomerization)35 and 1011 s–1
match/mismatch scenario was observed wherein the S,S catalyst (cyclopropyl ring opening)47, which is rapid enough to retain (and
afforded 1:3 e.r. while the R,R catalyst afforded 5:1 e.r., and the enhance) chiral memory via diastereoselective catalyst coordina-
racemic catalyst afforded 2:1 e.r.—all with varying levels of effi- tion38,48. The large KIE (6.1) also contrasts with those observed for
ciency (R,R>rac>S,S), indicating stereocontrol of both HAT and concerted C–H insertion by an Ir-nitrene (KIE: 1.5)9. Lastly, unlike
subsequent alkyl radical trapping. Likewise, subjecting an enan- nitrene-based methods that require azides or heterocyclic nitrene
tioenriched secondary alcohol as a diastereoselectivity probe, 53, precursors10–15, this Cu-catalysed radical relay strategy has the syn-
afforded between 2:1 d.r. (S,S catalyst) and 8:1 d.r. (R,R catalyst).
thetic benefit of enabling C–H amination of ubiquitous alcohols via
To probe the regioselectivity of the intramolecular HAT, sev- their readily available imidates.
eral classes of alcohols were subjected to this amination (Fig. 2e).
Notably, n-hexadecanol afforded a single β-aminated regioisomer Diversification of chiral products. Finally, the synthetic utility of
(54) resulting from exclusive 1,5-HAT. Additionally, α-amination this enantioselective β-C–H amination for accessing medicinally
was not observed in any case, indicating that 1,5-HAT is highly relevant motifs is illustrated in Fig. 3 via derivatization of oxazo-
favoured over 1,4-HAT despite an 8kcalmol−1 weaker α-oxy C–H line 1. For example, subjecting 1 to HCl at 100°C afforded free
bond. To elucidate 1,5- versus 1,6-HAT regioselectivity, sterically β-amino alcohol 60, while hydrolysis with HCl at 23°C afforded
NAtuRE CHEMiStRy | www.nature.com/naturechemistry

NATure CHeMiSTry
Articles
16. Wappes, E. A., Nakafuku, K. M. & Nagib, D. A. Directed β C–H amination
of alcohols via radical relay chaperones. J. Am. Chem. Soc. 139,
10204–10207 (2017).
β-benzamide 61. This amide may then be oxidized by RuCl3/NaIO4
to α-amino acid 62. Conversely, instead of an acidic work-up, dou-
ble reduction by iBu2AlH at 0°C yielded Bn-protected β-amine 63,
and C–O displacement by silyl halides (Me3Si–X) afforded vicinal
benzamide halides (64–66). Importantly, in each of these oxazoline
derivatizations, the stereocentre was retained with greater than 98%
e.s. Lastly, applying this method to streamlined access of privileged
catalyst architectures, oxazoline may be selectively deprotonated
17. Stateman, L. M., Nakafuku, K. M. & Nagib, D. A. Remote C–H
functionalization via selective hydrogen atom transfer. Synthesis 50,
1569–1586 (2018).
18. Choi, G. J., Zhu, Q., Miller, D. C., Gu, C. J. & Knowles, R. R. Catalytic
alkylation of remote C–H bonds enabled by proton-coupled electron transfer.
Nature 539, 268–271 (2016).
19. Chu, J. C. K. & Rovis, T. Amide-directed photoredox-catalysed
C–C bond formation at unactivated sp3 C–H bonds. Nature 539,
272–275 (2016).
n
with BuLi, then quenched with Ph2PCl, to afford phosphinooxa-
zoline ligand49. Alternatively, the bis-imidate 67 of a meta-diol was
converted to inverted bisoxazoline 68 (66% yield; 98% e.e.; 15:1
d.r.), providing rapid access to a new C2 symmetric ligand class.
20. Twilton, J. et al. Te merger of transition metal and photocatalysis. Nat. Rev.
Chem. 1, 0052 (2017).
21. Wang, F., Chen, P. & Liu, G. Copper-catalyzed radical relay for asymmetric
radical transformations. Acc. Chem. Res. 51, 2036–2046 (2018).
22. Hossain, A., Bhattacharyya, A. & Reiser, O. Copper’s rapid ascent in
visible-light photoredox catalysis. Science 364, eaav9713 (2019).
23. Ager, D. J., Prakash, I. & Schaad, D. R. 1,2-Amino alcohols and their
heterocyclic derivatives as chiral auxiliaries in asymmetric synthesis. Chem.
Rev. 96, 835–875 (1996).
Conclusions
In summary, we expect that this enantio- and regioselective, radical
β-C–H amination will streamline the synthesis of important mol-
ecules with the chiral, vicinal amino alcohol motif. Notably, this
approach bypasses the use of chiral auxiliaries (imines) and chiral
pool precursors (amino acids) and enables the enantioselective syn-
thesis of several new β-amino alcohols. Moreover, this radical relay
chaperone strategy will also enable the development of other classes
of stereoselective C–H functionalization.
24. Robak, M. T., Herbage, M. A. & Ellman, J. A. Synthesis and applications of
tert-butanesulfnamide. Chem. Rev. 110, 3600–3740 (2010).
25. Chang, H. T. & Sharpless, K. B. A practical route to enantiopure
1,2-aminoalcohols. Tetrahedron Lett. 37, 3219–3222 (1996).
26. Zhang, W. et al. Enantioselective cyanation of benzylic C–H bonds via
copper-catalyzed radical relay. Science 353, 1014–1018 (2016).
27. Murphy, J. J., Bastida, D., Paria, S., Fagnoni, M. & Melchiorre, P. Asymmetric
catalytic formation of quaternary carbons by iminium ion trapping of
radicals. Nature 532, 218–222 (2016).
Online content
Any methods, additional references, Nature Research reporting sum-
maries, source data, extended data, supplementary information,
acknowledgements, peer review information; details of author con-
tributions and competing interests; and statements of data and code
28. Wang, C., Harms, K. & Meggers, E. Catalytic asymmetric Csp3−H
functionalization under photoredox conditions by radical translocation
and stereocontrolled alkene addition. Angew. Chem. Int. Ed. 55,
13495–13498 (2016).
availability are available at https://doi.org/10.1038/s41557-020-0482-8. 29. Proctor, R. S. J., Davis, H. J. & Phipps, R. J. Catalytic enantioselective
Minisci-type addition to heteroarenes. Science 360, 419–422 (2018).
30. Zhang, W., Wu, L., Chen, P. & Liu, G. Enantioselective arylation of benzylic
Received: 7 November 2019; Accepted: 4 May 2020;
Published: xx xx xxxx
C−H bonds by copper-catalyzed radical relay. Angew. Chem. Int. Ed. 58,
6425–6429 (2019).
31. Shin, N. Y., Ryss, J. M., Zhang, X., Miller, S. J. & Knowles, R. R. Light-driven
deracemization enabled by excited-state electron transfer. Science 366,
References
364–369 (2019).
1. Roughley, S. D. & Jordan, A. M. Te medicinal chemist’s toolbox: an analysis
32. Li, J. et al. Site-specifc allylic C–H bond functionalization with a
of reactions used in the pursuit of drug candidates. J. Med. Chem. 54,
copper-bound N-centred radical. Nature 574, 516–521 (2019).
3451–3479 (2011).
33. Ni, Z. et al. Highly regioselective copper-catalyzed benzylic C–H
2. Park, Y., Kim, Y. & Chang, S. Transition metal-catalyzed C–H amination:
amination by N-fuorobenzenesulfonimide. Angew. Chem. Int. Ed. 51,
scope, mechanism, and applications. Chem. Rev. 117, 9247–9301 (2017).
1244–1247 (2012).
3. Davies, H. M. L. & Manning, J. R. Catalytic C–H functionalization by metal
34. Luo, Y. R. Comprehensive Handbook of Chemical Bond Energies
carbenoid and nitrenoid insertion. Nature 451, 417–424 (2008).
(CRC Press, 2007).
4. Romero, N. A., Margrey, K. A., Tay, N. E. & Nicewicz, D. A. Site-selective arene
35. Griller, D., Ingold, K. U., Krusic, P. J. & Fischer, H. Confguration of the
C–H amination via photoredox catalysis. Science 349, 1326–1330 (2015).
tert-butyl radical. J. Am. Chem. Soc. 100, 6750–6752 (1978).
5. Paudyal, M. P. et al. Dirhodium-catalyzed C–H arene amination using
36. Nakafuku, K. M., Fosu, S. C. & Nagib, D. A. Catalytic alkene
hydroxylamines. Science 353, 1144–1147 (2016).
difunctionalization via imidate radicals. J. Am. Chem. Soc. 140,
11202–11205 (2018).
6. Rufoni, A. et al. Practical and regioselective amination of arenes using alkyl
amines. Nat. Chem. 11, 426–433 (2019).
37. Welin, E. R., Le, C., Arias-Rotondo, D. M., McCusker, J. K. &
MacMillan, D. W. C. Photosensitized, energy transfer-mediated
organometallic catalysis through electronically excited nickel(ii). Science 355,
380–385 (2017).
7. McNally, A., Hafemayer, B., Collins, B. S. L. & Gaunt, M. J.
Palladium-catalysed C–H activation of aliphatic amines to give strained
nitrogen heterocycles. Nature 510, 129–133 (2014).
8. Sharma, A. & Hartwig, J. F. Metal-catalysed azidation of tertiary C–H bonds
suitable for late-stage functionalization. Nature 517, 600–604 (2015).
9. Hong, S. Y. et al. Selective formation of γ-lactams via C–H amidation enabled
by tailored iridium catalysts. Science 359, 1016–1021 (2018).
10. Reddy, R. P. & Davies, H. M. L. Dirhodium tetracarboxylates derived from
adamantylglycine as chiral catalysts for enantioselective C–H aminations.
Org. Lett. 8, 5013–5016 (2006).
38. Gutierrez, O., Tellis, J. C., Primer, D. N., Molander, G. A. & Kozlowski, M. C.
Nickel-catalyzed cross-coupling of photoredox-generated radicals: uncovering
a general manifold for stereoconvergence in nickel-catalyzed cross-couplings.
J. Am. Chem. Soc. 137, 4896–4899 (2015).
39. Kochi, J. K. Electron-transfer mechanisms for organometallic intermediates in
catalytic reactions. Acc. Chem. Res. 7, 351–360 (1974).
40. Kainz, Q. M. et al. Asymmetric copper-catalyzed C–N cross-couplings
induced by visible light. Science 351, 681–684 (2016).
11. Zalatan, D. N. & Du Bois, J. A chiral rhodium carboxamidate
catalyst for enantioselective C–H amination. J. Am. Chem. Soc. 130,
9220–9221 (2008).
12. Milczek, E., Boudet, N. & Blakey, S. Enantioselective C–H amination
using cationic ruthenium(ii)–pybox catalysts. Angew. Chem. Int. Ed. 47,
6825–6828 (2008).
41. Farney, E. P. et al. Discovery and elucidation of counteranion dependence in
photoredox catalysis. J. Am. Chem. Soc. 141, 6385–6391 (2019).
42. Strieth-Kalthof, F., James, M. J., Teders, M., Pitzer, L. & Glorius, F. Energy
transfer catalysis mediated by visible light: principles, applications, directions.
Chem. Soc. Rev. 47, 7190–7202 (2018).
13. Park, Y. & Chang, S. Asymmetric formation of γ-lactams via C–H
amidation enabled by chiral hydrogen-bond-donor catalysts. Nat. Catal. 2,
219–227 (2019).
43. Blum, T. R., Miller, Z. D., Bates, D. M., Guzei, I. A. & Yoon, T. P.
Enantioselective photochemistry through Lewis acid-catalyzed triplet energy
transfer. Science 354, 1391–1395 (2016).
14. Hu, Y. et al. Enantioselective radical construction of 5-membered cyclic
sulfonamides by metalloradical C–H amination. J. Am. Chem. Soc. 141,
18160–18169 (2019).
44. Wang, Y.-F., Chen, H., Zhu, X. & Chiba, S. Copper-catalyzed aerobic aliphatic
C–H oxygenation directed by an amidine moiety. J. Am. Chem. Soc. 134,
11980–11983 (2012).
15. Yang, Y., Cho, I., Qi, X., Liu, P. & Arnold, F. H. An enzymatic platform for
the asymmetric amination of primary, secondary and tertiary C(sp3)–H
bonds. Nat. Chem. 11, 987–993 (2019).
45. Alderson, J. M., Corbin, J. R. & Schomaker, J. M. Tunable, chemo- and
site-selective nitrene transfer reactions through the rational design of silver(i)
catalysts. Acc. Chem. Res. 50, 2147–2158 (2017).
NAtuRE CHEMiStRy | www.nature.com/naturechemistry

NATure CHeMiSTry
Articles
46. Korth, H. G., Trill, H. & Sustmann, R. [1-2H]-Allyl radical: barrier
to rotation and allyl delocalization energy. J. Am. Chem. Soc. 103,
4483–4489 (1981).
47. Newcomb, M., Johnson, C. C., Manek, M. B. & Varick, T. R. Picosecond
radical kinetics. Ring openings of phenyl-substituted cyclopropylcarbinyl
radicals. J. Am. Chem. Soc. 114, 10915–10921 (1992).
48. Gloor, C. S., Dénès, F. & Renaud, P. Memory of chirality in reactions
involving monoradicals. Free Radic. Res. 50, S102–S111 (2016).
49. Wüstenberg, B. & Pfaltz, A. Homogeneous hydrogenation of tri- and
tetrasubstituted olefns: comparison of iridium–phospinooxazoline [Ir–PHOX]
complexes and Crabtree catalysts with hexafuorophosphate (PF6) and
tetrakis[3,5-bis(trifuoromethyl) phenyl]borate (BArF) as counterions. Adv. Synth.
Catal. 350, 174–178 (2008).
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
Methods
Data availability
Enantioselective C–H amination. An oven-dried vial was charged with a stir
bar, imidate (0.1mmol) and (+)-camphoric acid (25mol%). Catalyst solutions of
Ir1 (1mol%), as well as CuBArF4 (2mol%) and L1 (4mol%) in Et2O (see below
for details) were added in a glove box. Afer dilution with pentane (4ml), the
vial was irradiated with a 455-nm blue LED for 1h while cooling by fan. Upon
completion, the reaction mixture was purifed by column chromatography to
aford the enantioenriched oxazoline. Further hydrolysis with HCl aforded the
enantioenriched amino alcohol.
The data that support the findings of this study are available within the article and
its Supplementary Information files.
Acknowledgements
Financial support was provided by the National Institutes of Health (NIH R35
GM119812) and National Science Foundation (NSF CAREER 1654656). L.M.S. is
supported by an NSF graduate fellowship. Calculations were performed using resources
at the Ohio Supercomputer Center.
Preparation of catalyst solutions.
Author contributions
L1·CuBArF4. A flame-dried flask was charged with a stir bar, CuBArF4 (0.04mmol)
and L1 (0.08mmol) in a glove box. Et2O (20ml) was added and stirred at 23°C for
1h to afford a 0.002M solution (1ml was used in each experiment).
K.M.N. designed and discovered the enantioselective C–H amination. K.M.N., E.A.W.
and Z.Z. developed the optimized method. Z.Z. evaluated the synthetic scope. Z.Z. and
L.M.S. performed the mechanistic experiments and derivatizations. A.D.C. performed
the calculations. All authors contributed to writing the manuscript.
Ir1. A flame-dried flask was charged with Ir{[dF(CF3)ppy]2dtbbpy}BArF
photocatalyst (0.02mmol) in a glove box. Et2O (20ml) was added to afford a
0.001M solution (1ml was used in each experiment).
4
Competing interests
The authors declare no competing interests.
Synthesis of imidate starting materials. A flame-dried flask was sequentially
charged with a stir bar, alcohol (2mmol), tetrahydrofuran (20ml) and NaH
(1.5equiv.). After stirring for 1h, imidoyl chloride (1.1equiv.) was added and the
resulting solution was further stirred. Upon complete consumption of alcohol,
the reaction mixture was purified by column chromatography to afford the
corresponding oxime imidate in 80–99% yields.
Additional information
Supplementary information is available for this paper at https://doi.org/10.1038/
s41557-020-0482-8.
Correspondence and requests for materials should be addressed to D.A.N.
Reprints and permissions information is available at www.nature.com/reprints.
NAtuRE CHEMiStRy | www.nature.com/naturechemistry