ADVERTISEMENT

 

 BACK TO VOL. 386, NO. 6717

 RESEARCH ARTICLE ORGANIC CHEMISTRY

 & ' 

A chiral hydrogen atom
abstraction catalyst for
the enantioselective
epimerization of meso-
diols
ANTTI S. K. LAHDENPERÄ , JYOTI DHANKHAR , [...], AND
ROBERT J. PHIPPS  +4 authors Authors Info & ARliations

SCIENCE 3 Oct 2024 Vol 386, Issue 6717 pp. 42-49

DOI: 10.1126/science.adq8029

10,516

   

       

Editor’s summary

Chiral diols (double alcohols) are impor-

tant building blocks for pharmaceutical
synthesis. Lahdenperä et al. report a
metal-free method to transform sym-
metrical diols selectively into one of two
possible mirror-image isomers. Specifi-
cally, they used a photoredox dye to acti-
vate a chiral amine catalyst, which then
abstracts a hydrogen atom preferentially
from one of the two diol carbons. Unse-
lective quenching from a thiol then fun-
nels the product stream over the course
of the reaction toward the desired prod-
uct. —Jake S. Yeston

Abstract

Hydrogen atom abstraction is an impor-

tant elementary chemical process but is
very difficult to carry out enantioselec-
tively. We have developed catalysts,
readily derived from the Cinchona alka-
loid family of natural products, which
can achieve this by virtue of their chiral
amine structure. The catalyst, following
single-electron oxidation, desym-
metrizes meso-diols by selectively ab-
stracting a hydrogen atom from one car-
bon center, which then regains a hydro-
gen atom by abstraction from a thiol.
This results in an enantioselective
epimerization process, forming the chi-
ral diastereomer with high enantiomeric
excess. Cyclic and acyclic 1,2-diols are
compatible, as are acyclic 1,3-diols. Ad-
ditionally, we demonstrate the viability
of combining our approach with carbon-
carbon bond formation in Giese addition.
Given the increasing number of synthet-
ic methods involving hydrogen atom
transfer steps, we anticipate that this
work will have a broad impact in the field
of enantioselective radical chemistry.

SIGN UP FOR THE AWARD-WINNING

SCIENCEADVISER NEWSLETTER
The latest news, commentary, and research, free to
your inbox daily

SIGN UP 

Hydrogen atom transfer (HAT) is a funda-

mental chemical reaction that forms the ba-
sis of numerous key processes in chemistry
and biology. Synthetic chemists have long
been aware of the potential of harnessing
HAT to develop new disconnections, and the
pursuit of such strategies has accelerated
sharply in the past decade, as photoredox
catalysis has facilitated exploration of reac-
tions involving open shell intermediates (1–
3). Issues of selectivity in radical reactions,
including HAT processes, have been exten-
sively considered (4). HAT involving the ab-
straction of a hydrogen atom from a closed-
shell substrate raises the obvious question of
site selectivity, as most molecules contain
numerous hydrogen atoms. Although by no
means a solved problem, a framework for
understanding site selectivity now exists to
guide practitioners based on the electronic
or steric environment of a given hydrogen
atom (5), spatial position relative to other
functional groups (6, 7), and the nature of
the incoming radical that executes the ab-
straction (3, 8, 9). In contrast to this, ap-
proaches to enantioselectivity in HAT pro-
cesses remain extremely limited. Enantiose-
lective HAT can be broadly classified into
two main types (Fig. 1A). The first type is
selective delivery of a hydrogen atom from a
chiral molecule to a previously generated,
prochiral radical intermediate, termed here
hydrogen atom delivery (HAD) (Fig. 1A, left).
Major breakthroughs in this area have re-
cently been achieved; biocatalysis using
flavin-dependant “ene” reductases, as pio-
neered by Hyster, allows precise and enanti-
ocontrolled delivery of a hydrogen atom
within the chiral environment of the enzyme
active site (10–16). In the area of chemo-
catalysis, highly effective peptide-based thi-
ols developed by Miller and Knowles have
demonstrated the potential of small-mol-
ecule chiral catalysts for this purpose (17,
18). In the second type of enantioselective
HAT process, termed here hydrogen atom
abstraction (HAA), a chiral radical abstracts
a hydrogen atom from a closed-shell sub-
strate, inducing asymmetry by means of ei-
ther desymmetrization or selective HAA
from one enantiomer of a racemic mixture
(kinetic resolution) (Fig. 1A, right). In con-
trast to the recent progress made in enan-
tioselective HAD, there are very few exam-
ples of enantioselective HAA to form dis-
crete radical intermediates. Yet the desym-
metrization of meso compounds is an impor-
tant strategy in asymmetric synthesis, par-
ticularly because multiple stereocenters can
be defined in a single operation (19–22).
Roberts and co-workers carried out early pi-
oneering studies using radicals generated
from chiral borane-complexed amines in the
context of kinetic resolution (23–26). En-
couraging results were obtained for selective
HAA from one enantiomer of a racemic mix-
ture, although selectivity (s) factors re-
mained low in most cases, and these early
discoveries have not since been followed up.
A mechanistic paradigm involving enantios-
elective HAA is the oxygen rebound mecha-
nism for C–H hydroxylation using high-va-
lent metal oxo complexes, occurring in nu-
merous enzymes as well as engineered small
molecule catalysts (27). Notably, Costas, Bi-
etti, and co-workers have recently reported
several elegant applications of the latter to
achieve enantioselective desymmetrization
reactions in which HAA is thought to be the
enantiodetermining step (28–30). Despite
their effectiveness for oxygenation, a limita-
tion is that the HAA step is closely coupled
with the subsequent radical rebound step,
meaning that it can be challenging to divert
the desymmetrized radical intermediate to
participate in asymmetric transformations
beyond oxygenation (31–35). Therefore, the
development of chiral HAA catalysts that are
able to form a discrete radical intermediate
amenable to broader functionalization re-
mains an outstanding challenge.

Fig. 1 . Background to enantioselective HAT and over-

view of this work.
(A) Two key modes in which enantioselective HAT can
occur: HAD and HAA. (B) Quinuclidine has emerged as
a versatile HAA catalyst. (C) A chiral quinuclidine-con-
taining HAA catalyst could be derived from Cinchona
alkaloids. (D) Application of chiral HAA catalyst for
enantioselective epimerization of meso-diols.

Quinuclidine has emerged as a highly effec-

tive catalyst for HAA since it was reported
for this purpose by MacMillan and co-work-
ers in 2015 (36). Forming a strong N–H bond
in the resulting quinuclidinium salt, the in-
termediary electrophilic radical cation excels
at abstracting hydridic hydrogen atoms, such
as those sharing a carbon atom with hydrox-
yl groups (Fig. 1B) (9, 37). We imagined that
a chiral version of quinuclidine, possessing
stereochemical information at the periphery
of the abstracting nitrogen atom, could form
the basis for an asymmetric HAA catalyst.
Fortuitously, the Cinchona alkaloid family of
natural products (38) present just such a
scaffold, and this parallel has been exploited
to great effect in the context of transition
metal catalysis (39). More broadly, these al-
kaloids have formed the basis of numerous
catalytic asymmetric processes spanning a
diverse range of mechanisms, repeatedly
proving themselves to be a privileged class
of chiral catalyst (Fig. 1C) (40). In selecting a
reaction in which to evaluate putative chiral
HAA catalysts, we were drawn to the inspir-
ing studies of Wendlandt and co-workers in
which cyclic cis-diols are isomerized to
trans-diols through non-enantioselective
HAA followed by HAD (41). In that work,
which built on their studies on carbohydrate
epimerization (42, 43), the optimized condi-
tions were deduced to be operating under
thermodynamic control, whereby an achiral
thiol acted as a HAD catalyst, and the corre-
sponding thiyl radical acted as a HAA cata-
lyst. We were intrigued by alternative condi-
tions identified in that work in which quinu-
clidine was used as a HAA catalyst alongside
adamantyl thiol as a HAD catalyst. That
combination was deduced to lead to a kineti-
cally controlled product outcome in which
the cis:trans ratio was approximately 1:1 for
the substrate under study. We surmised that
chiral HAA catalysts could be evaluated on
similar cis-diols under these conditions of
kinetic control, and the enantioselectivity of
any trans-diol produced could be evaluated.
Furthermore, it seemed plausible that a more
elaborate HAA catalyst than simple quinucli-
dine may feasibly influence the kinetically
controlled ratio of cis- to trans-diol ultimate-
ly obtained by influencing the rate of HAA
from each, although the direction in which
this might travel would be difficult to predict
de novo. In this work, we disclose a chiral
HAA catalyst that desymmetrizes a diverse
range of meso-diols by generating a radical
intermediate that can be trapped with a hy-
drogen atom donor to give the correspond-
ing chiral diols (Fig. 1D). We additionally
trapped the intermediate radical with elec-
tron-deficient olefins, demonstrating the
viability of carbon-carbon bond formation.
The desymmetrization of meso-diols has
been subject to extensive study by using all
manner of asymmetric catalysis methods
but, in the vast majority of cases, it is
achieved through functionalization on the
oxygen atom; methods that hinge on reac-
tivity at carbon are rare (44). Furthermore,
our method is complementary to current
leading approaches for accessing chiral diols
in C2-symmetric molecules: for cyclic sub-
strates, trans-1,2-diol products can be ob-
tained through epoxidation of a cis-alkene
followed by desymmetrizing, hydrolytic ring
opening (45). Jacobsen’s oligomeric
Co(salen) catalysts can achieve this effec-
tively but do not translate well to acyclic sys-
tems (46). In these cases, Sharpless asym-
metric dihydroxylation of a symmetrical
trans-alkene would be required (47). Neither
protocol is suitable for accessing chiral 1,3-
diols, yet our method through epimerization
of the corresponding meso-diols provides a
unified approach for these varied substitu-
tion patterns. Most saliently, given the wide-
spread use of HAT steps in many radical
methods, we believe that this catalytic ap-
proach to asymmetric hydrogen atom ab-
straction with easily accessed catalysts will
have broad impact in the burgeoning area of
enantioselective radical chemistry (48, 49).

Reaction development

We began our studies using the meso com-

pound cis-1,2-cyclohexanediol (1a) as sub-
strate, applying a triple catalyst system com-
prising photocatalyst 1,2,3,5-Tetrakis(car-
bazol-9-yl)-4,6-dicyanobenzene (4CzIPN; 5
mol%), a Cinchona alkaloid derivative (10
mol%) as chiral HAA catalyst, and 1-dode-
canethiol (25 mol%) as an achiral HAD cata-
lyst under blue light-emitting diode (LED)
irradiation at room temperature (Fig. 2A).
We also included as an additive 10 mol% of
tetrabutylammonium dihydrogen phosphate
(Bu4NH2PO4), as previous studies have sug-
gested that this can accelerate HAA adjacent
to alcohols (36, 50). For the chiral HAA cata-
lyst, we began with dihydrocinchonidine
(DHCD) and found that this promoted some
formation of the desired trans-diol 2a at
room temperature after 16 hours, but with
no enantioselectivity (entry 1). Lowering the
temperature to –35°C (see the supplemen-
tary materials for experimental setup) pro-
duced a higher yield of 2a, and an encourag-
ing enantiomeric excess (ee) of 24% was ob-
served (entry 2). At this point, we proceeded
to evaluate structural changes to the HAA
catalyst to gauge the impact on both the ra-
tio of 2a:1a and the ee of 2a (entry 3) (Fig.
2B). Inverting the hydroxyl stereocenter (epi-
DHCD) gave a product ratio now in favor of
2a, but ee was only slightly increased (33%).
The next modification replaced the hydroxyl
group with a butoxycarbonyl (Boc)–protect-
ed amine while retaining the original stereo-
chemistry (NHBoc-DHCD). This gave lower
conversion to 2a over the same time frame
with broadly similar ee, in favor of the oppo-
site enantiomer (–30%). Notably, inverting
the amine stereocenter (epi-NHBoc-DHCD)
gave a much-improved performance, yield-
ing >20:1 2a:1a and –74% ee of 2a. With this
highly encouraging result in hand, we evalu-
ated the pseudoenantiomer of this catalyst.
Epi-NHBoc-DHCN, derived from cinchonine
(DHCN, dihydrocinchonine), gave the an-
tipode of 2a, as anticipated, and did so with
substantially increased ee (92%). The poten-
tial for discrepancy between pseudoenan-
tiomeric Cinchona alkaloid–derived cata-
lysts is well acknowledged (51). Moving from
cinchonine to the analogous quinidine-de-
rived catalyst (epi-NHBoc-DHQD; DHQD,
dihydroquinidine), which differs by inclusion
of a methoxy group on the quinoline portion
of the catalyst, gave an extremely poor out-
come with a 1:1 ratio of 1a:2a and only 11%
ee of 2a. Notably, its pseudoenantiomer (epi-
NHBoc-DHQN; DHQN, dihydroquinine)
proved to be an excellent catalyst, giving –
86% ee; the origin of this sharp divergence in
effectiveness is unclear at this time. With
optimal catalyst and conditions in hand (en-
try 4), we next carried out a series of control
reactions to establish the necessity of the
various components, omitting each reaction
component sequentially: blue light irradia-
tion, thiol catalyst, photocatalyst, and HAA
catalyst (entries 5 to 8). In the absence of
any of these, conversion to 2a was negligible.
Removal of Bu4NH2PO4 or exploration of
several other tetrabutylammonium phos-
phate salts had relatively little effect on re-
activity or enantioselectivity (table S2).
However, we observed that the inclusion of
Bu4NH2PO4 for some other substrates ap-
pears to increase reactivity toward the
epimerization (fig. S1) and therefore chose to
retain it during scope exploration (entry 9).
We replaced epi-NHBoc-DHCN with simple
quinuclidine which gave a 1:1 ratio of 1a:2a,
in line with Wendlandt’s precedent under
similar conditions (41), demonstrating that
the chiral HAA catalyst epi-NHBoc-DHCN is
not only inducing enantioselectivity but is
also having a profound and fortuitous effect
on the final diastereomeric ratio (dr) of prod-
ucts for this substrate (entry 10).

Fig. 2 . Reaction optimization and chiral HAA catalyst

exploration.
Yields were determined by 1H–nuclear magnetic reso-
nance (NMR) with reference to an internal standard. Ee
values were determined by chiral supercritical fluid
chromatography (SFC) analysis after derivatization of
the diol to the corresponding dibenzoyl ester (see the
supplementary materials for details). (A) Evaluating the
effects of temperature and Cinchona alkaloid–derived
catalyst, as well as control experiments. Rt, room tem-
perature. (B) Schematic of Cinchona alkaloid–derived
HAA catalysts evaluated and associated reaction out-
comes.

Scope exploration for the enan-

tioselective epimerization of
cyclic meso-diols

With optimal conditions in hand for the

enantioselective epimerization of 1a to 2a
using the chiral HAA catalyst epi-NHBoc-
DHCN, we evaluated the scope and limita-
tions of the process on cyclic meso-diols (Fig.
3A). The reaction translated well to a larger
seven-membered ring (2b), with an even
higher enantioselectivity being obtained. For
eight- and five-membered rings, lower yields
of the trans-diol were obtained, but the ee
values obtained were still outstanding (2c
and 2d). We found that as we moved away
from the optimization substrate, increasing
the amount of thiol was beneficial for yield,
as it reduced deleterious side reactions, pos-
sibly by faster trapping of the intermediate
radical by the hydrogen atom donor. Gemi-
nal substituents could be included on the
cycloheptane ring, as manifested with di-
esters (2e), an acetonide-protected diol (2f),
and a benzoyl-protected diol (2g). High ee
was maintained when a benzene ring was
fused to the cyclohexane (2h). We proceeded
to investigate meso-cyclohexane-1,2-diols
possessing more complex substitution (Fig.
3B). For some substrates, yields could be in-
creased by increasing the HAA catalyst load-
ing (see the supplementary materials for
specific details). In most cases, stereospecific
dihydroxylation of the parent alkenes di-
astereoselectively gave the meso-diol sub-
strate in which the introduced cis-diol had a
trans relationship with the existing sub-
stituents on the cyclohexane. Compatible
substrates included 1i, bearing two esters;
1j, bearing a fused cyclopentane; and 1k,
featuring an acetonide-protected cis-diol in
addition to the diol that underwent epimer-
ization. Notably, this protocol proved com-
patible with a range of labile C–H bonds that
may be susceptible to HAT processes. For
example, substrates containing benzylic
(1h), other hydridic (1f and 1g), and protic
(1i) C–H bonds all smoothly gave the
epimerized products. Fused pyrrolidine sub-
strate 1l is notable in that it possesses po-
tentially abstractable hydrogen atoms adja-
cent to nitrogen but still gives the epimer-
ized product 2l in high yield with good enan-
tioselectivity. In 1m, a cyclobutane ring can
be incorporated into the substrate without
issue for the epimerization. For substrates 1i
to 1m, four stereocenters are defined in an
absolute sense during the epimerization
process. In the synthesis of 1n, an insepara-
ble mixture of diastereomers was obtained
on dihydroxylation, and this mixture was
used in the epimerization reaction to give a
low yield but high ee of cyclopropane-con-
taining 2n. In several cases, both cis and
trans meso-diols were formed during dihy-
droxylation and could be separated, allowing
separate examination. For 1o, featuring a
pair of acetonide protected primary alcohols,
both cis and trans isomers converged to give
the same enantiomer of 2o upon epimeriza-
tion (Fig. 3C). Both isomers of starting mate-
rial were similarly evaluated for the eight-
membered acetonide-protected diol 1p, each
delivering excellent ee of the common prod-
uct 2p. To test the limits of the stereochemi-
cal complexity tolerated by the HAA catalyst,
we evaluated meso-diol 1q, readily synthe-
sized through Diels-Alder reaction followed
by dihydroxylation in which all carbons of
the core cyclohexane bear substituents (Fig.
3D). This reaction gave 99% ee of the trans-
diol product 2q in which six stereocenters
are defined in an absolute sense in this sin-
gle catalytic asymmetric step, underlining
the potential of enantioselective HAA catal-
ysis for the rapid asymmetric synthesis of
stereochemically complex compounds.

Fig. 3 . Reaction scope for enantioselective epimeriza-

tion of cyclic meso-diols.
(A) Simpler cyclic diol substrates. (B) More substituted
cyclic diol substrates. (C) Common products obtained
from different meso stereoisomers. (D) Defining six ab-
solute stereocenters in a single step. Yields are isolat-
ed yields of the diol product. Ee values were determined
Expand for more 

Translation to acyclic meso-di-

ols

Given the excellent outcomes obtained by

using our chiral HAA catalyst in the desym-
metrization of cyclic meso-diols, we were
intrigued as to whether this process may
translate to analogous acyclic meso-diols.
The much greater conformational freedom
enjoyed by such substrates would suggest
that this may be a rather more challenging
undertaking. We were pleased to discover
that when meso-hexane-3,4-diol (1r) was
evaluated under the reaction conditions, the
chiral (S,S) isomer 2r could be obtained in
99% ee and high diastereoselectivity, with
only traces of starting meso-diol remaining
in the crude reaction mixture, although
volatility reduced the isolated yield (Fig. 4A).
The alkyl chain on each side of the diol could
be elongated by one and two methylene
units with similarly successful outcomes (2s
and 2t). We proceeded to evaluate several
substrates that bear functionality on either
side of the diol; 2u and 2v constitute partic-
ularly interesting products, as both possess
benzylic sites that could potentially undergo
HAA in competition with the desired site,
and yet, these reactions still functioned to
epimerize the alcohol with high ee. In the
case of 2u, even bulky phenyl rings close to
the site of HAA gave only a small decrease in
ee, although conversion was moderate,
which we speculate could be due to the rela-
tively large size of our HAA catalyst reducing
the rate of abstraction. Esters could be toler-
ated (2w), as could silyl-protected tertiary
alcohols (2x). For the ester-substituted 2w,
conversion was again moderate, which we
attribute to electronic effects disfavoring
HAA. For those substrates in which some
inseparable meso-diol remained, quoted
yields are adjusted to take into account the
fraction of chiral isomer present in the iso-
lated mixture (see the supplementary mate-
rials for full details). Owing to the success of
the acyclic 1,2-diols, we speculated as to
whether 1,3-diols may also be amenable to
the enantioselective epimerization using our
Cinchona alkaloid–derived catalyst. We sub-
jected the meso diastereomer of pentane-
2,4-diol to the same conditions and were de-
lighted to find that, although meso isomer
remained in the reaction mixture, the chiral
isomer 4a could be isolated in 51% yield and
99% ee (Fig. 4B). The extremely high level of
stereoinduction was retained moving from
methyl to isopropyl substituents on the 1,3-
diol unit (4b). This could be expanded to in-
corporate a substituent on the central car-
bon, giving 4c in 99% ee.

Fig. 4 . Reaction scope for enantioselective epimeriza-

tion of acyclic meso-diols.
(A) Scope of acyclic 1,2-diols. (B) Scope of acyclic 1,3-
diols. Yields are isolated yields of the diol product ex-
cept where indicated with an asterisk, which denotes
that the chiral diol product was inseparable from re-
maining meso-diol, and yield was calculated based on
Expand for more 

Further experiments and mech-

anistic considerations

Having been successful in trapping the puta-

tive desymmetrized radical intermediate
with a hydrogen atom, we sought to demon-
strate the viability of C–C bond formation by
trapping with an electrophilic alkene in a
Giese addition, the archetypical C–C bond-
forming radical reaction (52). Initially, by
using 1a as substrate, we found that two dif-
ferent acceptors, acrylamide 5a and vinyl
sulfone 5b, reacted smoothly under the opti-
mized conditions (in the absence of thiol) to
generate 6a and 6b (Fig. 5A, top). In both
cases, no diastereocontrol was obtained in
the trapping, but each of the diastereomers
was obtained in excellent ee, in line with our
mechanistic hypothesis of enantiodetermin-
ing HAA followed by an unselective trapping
of the resultant radical, in analogy with Fig.
1D. Oxidation of a diastereomeric mixture of
6a prior to separation gave a ketone with a
minimal ee that reflected the low dr of 6a
(see the supplementary materials for de-
tails), verifying that the catalyst is not con-
trolling the facial selectivity of the Giese ad-
dition but rather setting the absolute config-
uration of the adjacent hydroxyl-containing
stereocenter by enantioselective HAA. We
next evaluated Giese addition with the
acyclic 1,2- and 1,3-diols 1r and 3a to com-
pare their behavior with the cyclic diol 1a.
Similar outcomes were achieved in products
6c to 6f; we generally obtained low dr in the
trapping of the radical but excellent enan-
tioselectivity for each diastereomer, again
supporting the hypothesis of highly enan-
tioselective HAA by the catalyst from the
meso-diol starting material (Fig. 5A, middle
and bottom). Practically, it is important that
the opposite product enantiomer can be ob-
tained if desired, and we demonstrate that
epi-NHBoc-DHCD, the cinchonidine-derived
version of our optimal cinchonine-derived
catalyst, is an effective pseudoenantiomer
for a number of substrates from the reaction
scope, to give cyclic products 2j, 2k, and 2q
and also acyclic products 2r and 4a (Fig. 5B).
Stern-Volmer quenching experiments (Fig.
5C) showed that photoexcited 4CzIPN is ef-
fectively quenched by epi-NHBoc-DHCN but
not by dodecanethiol or the diol substrate 1a
(the phosphate additive was excluded from
the experiments because it had no signifi-
cant impact on enantioselectivity or reactivi-
ty for 1a). This is in line with the proposed
scenario whereby the chiral HAA catalyst is
first oxidized to the radical cation by the
photocatalyst (36). A combination of HAA
and thiol displays similar quenching to the
HAA alone, providing support for the alkyl
thiol acting as a hydrogen atom donor source
to quench the desymmetrized radical inter-
mediate rather than undergoing deprotona-
tion to form an alternative achiral HAA cata-
lyst, which may be expected with a more
acidic thiol, per Wendlandt’s prior studies
(41). The HAD from the dodecanethiol may
not be diastereoselective, but the alternative
to formation of desired trans-diol 2a is ref-
ormation of meso-1a, the starting material,
which can re-enter the reaction (Fig. 1D).
Further experimental evidence was obtained
for our proposed mechanism by subjecting
deuterium-labeled 1a (1a-d2, 96% deutera-
tion) to the standard reaction conditions
(Fig. 5D, left). With one equivalent of dode-
canethiol as the HAD source, the reaction
was stopped after 4 hours, and product 2a
was isolated in 41% yield and 92% ee, ex-
hibiting 51% deuteration. Remaining 1a was
recovered in 54% yield with 84% deuteration.
These observations provide support for the
above-described mechanistic scenario in-
volving nonselective hydrogen atom delivery
to the desymmetrized radical intermediate,
reforming starting material (indicated by
partial erosion of isotopic purity) as well as
product. Although the progressive erosion of
the deuteration level of 1a-d2 as the reaction
progresses makes determining a precise ki-
netic isotope effect (KIE) value difficult, we
monitored the rate of product formation
from 1a-d2 and compared this with the rate
from standard 1a. This suggested a primary
KIE of approximately 2.4, implicating HAA
by chiral catalyst as being rate limiting (53).
The high trans diastereoselectivity obtained
in the epimerization of 1a using our chiral
catalyst contrasts sharply with quinuclidine,
and we demonstrate this using a series of
simple experiments (Fig. 5D, lower). Al-
though quinuclidine reacts readily with cis-
diol 1a (34% yield of 2a after 1 hour), it re-
acts similarly readily with trans-diol 2a to
form 1a (33% yield after 1 hour), in full ac-
cordance with Wendlandt’s prior observa-
tions (41). The chiral catalyst forms substan-
tial amounts of 2a on the same time scale
(20%), but, in contrast with quinuclidine,
there is negligible back-reaction when 2a is
submitted to the reaction conditions. We
also evaluated some of the suboptimal cata-
lysts from Fig. 2B in the reaction with 2a and
saw varying levels of conversion back to 1a
(see the supplementary materials for de-
tails). Additionally, we submitted deuterated
2a (rac-2a-d2, 95% deuterium incorporation)
to the reaction conditions, and after 4 hours,
it was recovered with negligible hydrogen
incorporation, demonstrating that on the
epimerization reaction timescale, HAA from
2a is not occurring to any significant degree
(Fig. 5D, right). Although dynamic processes
do not appear be playing a significant role in
the selectivity outcome for this substrate,
this mechanistic paradigm involving sepa-
rate hydrogen atom abstractors and donors,
either or both of which could be chiral, offers
valuable opportunities for further develop-
ment of catalytic enantioselective methods.

 OPEN IN VIEWER

Fig. 5 . Additional experiments.

(A) Trapping of desymmetrized radical intermediates
with electron-deficient olefins in a Giese addition to
form a new C–C bond. (B) Access to product antipodes
using a pseudoenantiomeric catalyst. (C) Stern-Volmer
quenching experiments indicating excited photocata-
lyst quenching with HAA catalyst but not thiol or 1a. (D)
Experiments using deuterated substrates and the com-
parison of epi-NHBoc-DHCN with quinuclidine.

We anticipate that this disclosure of a highly

effective catalyst for enantioselective diol
epimerization through HAA, giving broad
scope on cyclic and acyclic 1,2- and 1,3-diols,
will feed into exciting developments in
enantioselective radical chemistry. It has the
potential to open valuable opportunities for
combination with further bond forming pro-
cesses that allow the rapid construction of
stereochemically rich molecules through
single electron pathways.

Acknowledgments

We are grateful to R. Cabot and T. Chisholm

for assistance with the Stern-Volmer
quenching experiments, P. Gierth and D.
Howe for assistance with NMR experiments,
and the Dydio group for use of their GC.

Funding: Horizon 2020 ERC Starting Grant

NonCovRegioSiteCat (757381); UKRI EPSRC
grant EP/Y02348X/1; Swiss National Science
Foundation Postdoctoral Fellowship (J.D.);
EPSRC Centre for Doctoral Training - Syn-
Tech (EP/S024220/1)and Vertex Pharmaceu-
ticals Ph.D. Studentship (D.J.D.); Trinity Hall,
Cambridge Research Fellowship (N.Y.S.L.);
Ph.D. Studentship through the AstraZeneca-
Cambridge Ph.D. Program (P.D.B.); European
Union (Horizon 2020 Marie Skłodowska-
Curie COFUND grant agreement no. 801474)
and Ministerio de Ciencia e Inno-
vación/Agencia Estatal de Investigación
(MCIN/AEI/10.13039/501100011033,
CEX2019-000925-S) (K.V.H.).

Author contributions: A.S.K.L. and R.J.P.

conceived the work. All authors designed and
analyzed the experiments. A.S.K.L., J.D.,
D.J.D., N.Y.S.L., P.D.B., and K.V.H. performed
the experiments. R.J.P. directed the research
and prepared the manuscript with input
from all authors.

Competing interests: The authors declare

that they have no competing interests.

Data and materials availability: All data

are available in the main text or the supple-
mentary materials.

License information: Copyright © 2024 the

authors, some rights reserved; exclusive li-
censee American Association for the Ad-
vancement of Science. No claim to original
US government works. https://www.sci-
ence.org/about/science-licenses-journal-ar-
ticle-reuse. The author has made the Author
Accepted Manuscript version of this article
available under a CC BY license in the Uni-
versity of Cambridge institutional repository
Apollo.

Supplementary Materials

This PDF _le includes:

Materials and Methods

Fig. S1 to S3
Tables S1 and S2
NMR Spectra and SFC/HPLC Traces
References (54–111)

DOWNLOAD 19.87 MB

References and Notes

1 L. Capaldo, D. Ravelli, Hydrogen Atom Transfer
(HAT): A Versatile Strategy for Substrate Activation
in Photocatalyzed Organic Synthesis. Eur. J. Org.
Chem. 2017, 2056–2071 (2017).

 GO TO REFERENCE • CROSSREF • PUBMED

• WEB OF SCIENCE • GOOGLE SCHOLAR

2 H. Cao, X. Tang, H. Tang, Y. Yuan, J. Wu, Photoin-

duced intermolecular hydrogen atom transfer reac-
tions in organic synthesis. Chem Catal. 1, 523–598
(2021).

CROSSREF • GOOGLE SCHOLAR

3 L. Capaldo, D. Ravelli, M. Fagnoni, Direct Photocat-

alyzed Hydrogen Atom Transfer (HAT) for Aliphatic
C-H Bonds Elaboration. Chem. Rev. 122, 1875–1924
(2022).

 SEE ALL REFERENCES • CROSSREF • PUBMED

• WEB OF SCIENCE • GOOGLE SCHOLAR

4 A. L. J. Beckwith, Centenary Lecture. The pursuit of

selectivity in radical reactions. Chem. Soc. Rev. 22,
143–151 (1993).

 GO TO REFERENCE • CROSSREF • WEB OF SCIENCE

• GOOGLE SCHOLAR

SHOW ALL REFERENCES

eLetters (0)

eLetters is a forum for ongoing peer review. eLet-

ters are not edited, proofread, or indexed, but they
are screened. eLetters should provide substantive
and scholarly commentary on the article. Neither
embedded dgures nor equations with special char-
acters can be submitted, and we discourage the
use of dgures and equations within eLetters in
general. If a dgure or equation is essential, please
include within the text of the eLetter a link to the
dgure, equation, or full text with special characters
at a public repository with versioning, such as Zen-
odo. Please read our Terms of Service before sub-
mitting an eLetter.

LOG IN TO SUBMIT A RESPONSE

No eLetters have been published for this article

yet.

We recommend

O-coordinated W-Mo dual-atom catalyst for pH-universal elec-

trocatalytic hydrogen evolution
Yang Yang, Sci Adv, 2020

Magnetism of the Hydrogen Atom

Science, 1926

Magnetism of the Hydrogen Atom

Science, 1926

Electron diffraction and the hydrogen atom

Lynne B. McCusker, Science, 2017

Fine Structure of the Hydrogen Atom

Willis E. Lamb, Science, 1956

Chiral titanium complexes for enantioselective catalysis

Koichi Mikami, Oxford Academic Books, 1999

The hydrogen atom

Keith Hannabuss, Oxford Academic Books, 1997

The Hydrogen Atom

Richard W Robinett ed., Oxford Academic Books, 2006

The hydrogen atom

Geoffrey Brooker, Oxford Academic Books, 2021

The hydrogen atom

C J Foot, Oxford Academic Books, 2004

Powered by

CURRENT ISSUE

A glutamine metabolic switch supports

erythropoiesis
BY JUNHUA LYU, ZHIMIN GU, ET AL.

SPL13 controls a root apical meristem phase

change by triggering oriented cell divisions
BY BAOJUN YANG, YANBIAO SUN, ET AL.

Chemical genetic approaches to dissect microbiota

mechanisms in health and disease
BY XINGLIN YANG, HOWARD C. HANG

TABLE OF CONTENTS 
ADVERTISEMENT

Sign up for ScienceAdviser

Get Science’s award-winning newsletter with the
latest news, commentary, and research, free to your
inbox daily.

SUBSCRIBE


LATEST NEWS
SCIENCESHOTS 15 NOV 2024

Saber-toothed kitten frozen in Siberian ice

reveals new details about species

SCIENCEINSIDER 15 NOV 2024

Prospect of RFK Jr. as head of HHS panics

many in medical science community

NEWS 15 NOV 2024

The _rst people on Tasmania brought _re and

forever changed the land

SCIENCEINSIDER 14 NOV 2024

When it comes to science, U.S. Senate’s new

leader has a buried past

NEWS 14 NOV 2024

With _rst mechanical qubit, quantum

computing goes steampunk

NEWS 14 NOV 2024

News at a glance: Replication troubles,

_nancial confict disclosures, and a public
health shake-up

ADVERTISEMENT

RECOMMENDED
REPORT MARCH 2020

Enantioselective remote C–H activation

directed by a chiral cation

PERSPECTIVE JANUARY 2020

Stitching two chiral centers with one catalyst

RESEARCH ARTICLE APRIL 2019

Delayed catalyst function enables direct

enantioselective conversion of nitriles to
NH2-amines

REPORT DECEMBER 2011

Asymmetric Electrophilic Fluorination Using

an Anionic Chiral Phase-Transfer Catalyst

SPONSORED WEBINAR TECHNOLOGY 29 OCT 2024

Autophagy's hidden drivers: Exploring the role

of post-translational modi_cations

ADVERTISEMENT

View full text | Download PDF

FOLLOW US

 
 

  &

GET OUR NEWSLETTER

NEWS

All News

ScienceInsider

News Features

Subscribe to News from Science

News from Science FAQ

About News from Science

Donate to News

CAREERS

Careers Articles

Find Jobs

Employer Hubs

COMMENTARY

Opinion

Analysis

Blogs

JOURNALS
 Science

Science Advances

Science Immunology

Science Robotics

Science Signaling

Science Translational Medicine

Science Partner Journals

AUTHORS & REVIEWERS

Information for Authors

Information for Reviewers

LIBRARIANS

Manage Your Institutional Subscription

Library Admin Portal

Request a Quote

Librarian FAQs

ADVERTISERS

Advertising Kits

Custom Publishing Info

Post a Job

RELATED SITES

AAAS.org

AAAS Communities

EurekAlert!

Science in the Classroom

ABOUT US

Leadership

Work at AAAS

Prizes and Awards

HELP

FAQs

Access and Subscriptions

Order a Single Issue

Reprints and Permissions

TOC Alerts and RSS Feeds

Contact Us

© 2024 American Association for the Advancement of Science. All

rights reserved. AAAS is a partner of HINARI, AGORA, OARE,
CHORUS, CLOCKSS, CrossRef and COUNTER. Science ISSN 0036-
8075.

Terms of Service | Privacy Policy | Accessibility