Ligand coordination- and dissociation-induced divergent allylic alkylations using alkynes

Article abstract of DOI:10.1016/j.chempr.2021.02.018

Efficient allylic alkylation of vinyl epoxides/aziridines with terminal alkynes can be achieved in the presence of nickel and copper catalysts. Intriguing ligand-induced divergent reactions are observed to produce either 1,4-enynes from bimolecular allylic alkynylation or dienynes from trimolecular alkyne dimerization/allylic alkylation. The allylic alkynylation reaction works well for a wide range of alkynes bearing a myriad of functionalities. An enantioconvergent transformation of racemic vinyl carbonates to enantioenriched 1,4-enynes is also realized.In transition metal catalysis, the nature of ligand coordination plays a vital role in defining the reactivity of the catalytic system. Achieving multi-substrate coordination and coupling by leveraging dynamic ligand dissociation is rare. Here, we report the discovery of such a phenomenon in nickel/copper-co-catalyzed divergent allylic alkylation of vinyl epoxides and aziridines using alkynes. Under otherwise identical conditions, the use of either strong bisphosphine ligands or hemilabile P,N-ligands leads to bimolecular allylic alkynylation or trimolecular dienyne formation, respectively. DFT calculations provide key insights for these ligand-induced divergent reactivities, particularly ligand-dissociation-enabled coordination of three substrates on nickel for trimolecular coupling. This catalytic system couples a remarkably broad range of readily available terminal alkynes with vinyl epoxides, carbonates, and aziridines in high regio- and stereo-selectivity. The utility of the versatile enyne and dienyne products, coupled with the use of abundant nickel/copper catalysts, makes this a practical method in synthetic and medicinal chemistry.Catalysis is a central concept in chemistry as it enables reactivity not attainable under reasonable conditions by other means and allows for more sustainable and economical processes. In transition metal catalysis, the nature of ligand coordination plays a vital role in the reactivity of the catalytic system. Ligand dissociation from the metal center is typically deemed a destructive scenario, and its alteration to achieve well-controlled reactivity is difficult to accomplish. We report here an intriguing example of leveraging dynamic ligand dissociation on nickel for divergent bimolecular or trimolecular allylic alkylation using terminal alkynes. This catalytic system shows a remarkable scope with superb functional group tolerance and may find wide use in the functionalization of alkyne-containing molecules. This discovery and the detailed mechanistic understanding obtained from DFT calculations will also direct future endeavors in the development of novel, divergent metal-catalyzed reactions.

Full text of DOI:10.1016/j.chempr.2021.02.018

ll
Article
Ligand coordination- and dissociation-
induced divergent allylic alkylations using
alkynes
Yuan Huang, Chao Ma, Song
Liu, Li-Cheng Yang, Yu Lan, Yu
Zhao
lanyu@cqu.edu.cn (Y.L.)
zhaoyu@nus.edu.sg (Y.Z.)
HIGHLIGHTS
Ligand coordination/dissociation-
induced divergent bimolecular or
trimolecular reactions
A general Ni/Cu-catalyzed allylic
alkylation of vinyl epoxides/
aziridines using alkynes
A remarkable scope of alkynes
substituted with a myriad of
functionalities
Enantioconvergent allylic
alkynylation of vinyl carbonates
with alkynes
Efficient allylic alkylation of vinyl epoxides/aziridines with terminal alkynes can be
achieved in the presence of nickel and copper catalysts. Intriguing ligand-induced
divergent reactions are observed to produce either 1,4-enynes from bimolecular
allylic alkynylation or dienynes from trimolecular alkyne dimerization/allylic
alkylation. The allylic alkynylation reaction works well for a wide range of alkynes
bearing a myriad of functionalities. An enantioconvergent transformation of
racemic vinyl carbonates to enantioenriched 1,4-enynes is also realized.
Huang et al., Chem 7, 812–826
March 11, 2021 ª 2021 Elsevier Inc.
https://doi.org/10.1016/j.chempr.2021.02.018

ll
Article
Ligand coordination- and dissociation-induced
divergent allylic alkylations using alkynes
1,7,9,
Yuan Huang,1,2,8 Chao Ma,1,3,8 Song Liu,4,5,8 Li-Cheng Yang,1 Yu Lan,4,6, and Yu Zhao
*
*
SUMMARY
The bigger picture
Catalysis is a central concept in
chemistry as it enables reactivity
not attainable under reasonable
conditions by other means and
allows for more sustainable and
economical processes. In
In transition metal catalysis, the nature of ligand coordination plays
a vital role in defining the reactivity of the catalytic system.
Achieving multi-substrate coordination and coupling by leveraging
dynamic ligand dissociation is rare. Here, we report the discovery
of such a phenomenon in nickel/copper-co-catalyzed divergent
allylic alkylation of vinyl epoxides and aziridines using alkynes. Un-
der otherwise identical conditions, the use of either strong bisphos-
phine ligands or hemilabile P,N-ligands leads to bimolecular allylic
alkynylation or trimolecular dienyne formation, respectively. DFT
calculations provide key insights for these ligand-induced divergent
reactivities, particularly ligand-dissociation-enabled coordination of
three substrates on nickel for trimolecular coupling. This catalytic
system couples a remarkably broad range of readily available termi-
nal alkynes with vinyl epoxides, carbonates, and aziridines in high re-
gio- and stereo-selectivity. The utility of the versatile enyne and di-
enyne products, coupled with the use of abundant nickel/copper
catalysts, makes this a practical method in synthetic and medicinal
chemistry.
transition metal catalysis, the
nature of ligand coordination
plays a vital role in the reactivity of
the catalytic system. Ligand
dissociation from the metal center
is typically deemed a destructive
scenario, and its alteration to
achieve well-controlled reactivity
is difficult to accomplish. We
report here an intriguing example
of leveraging dynamic ligand
dissociation on nickel for
divergent bimolecular or
trimolecular allylic alkylation using
terminal alkynes. This catalytic
system shows a remarkable scope
with superb functional group
tolerance and may find wide use in
the functionalization of alkyne-
containing molecules. This
INTRODUCTION
The efficient generation of compound libraries bearing a broad structural diversity
remains an important objective of chemical synthesis in support of medicinal chem-
istry and material science.1 As an effective strategy to serve such endeavors,
achieving divergent reactivities of the same set of substrates to access different
products has attracted much attention in synthetic method development.2 In partic-
ular, transition metal catalysis has given great impetus to delivering elegant exam-
ples of chemo-, regio-, or diastereo-divergent transformations to access products
that are isomeric in structure.3–5 In these catalytic methods, it was often the identity
of the ligands that played an essential role in altering the steric and electronic prop-
erties of the metal complexes leading to different selectivities. It is important to
note, however, ligand association on the metal at all times is considered to be the
key for efficient catalysis, while the dissociation of ligand typically results in loss of
catalytic activity or uncontrolled side reaction pathways. Herein, we present an
intriguing yet rarely explored scenario for divergent reactions as shown in Figure 1A:
the coordination/dissociation of a ligand may result in different numbers of open
binding sites on the metal to accommodate two or more substrates to yield products
of completely different molecular scaffolds. By the use of strongly coordinating bi-
dentate ligand, the metal catalyst tends to incorporate only two molecules of sub-
strates and, thus, promotes bimolecular reaction to yield product A. Alternatively,
the metal complex bearing (hemi)labile ligand may undergo ligand dissociation
at an appropriate stage in the catalytic cycle to allow the coordination of
more than two substrate molecules, delivering, e.g., product B derived from a
discovery and the detailed
mechanistic understanding
obtained from DFT calculations
will also direct future endeavors in
the development of novel,
divergent metal-catalyzed
reactions.
812 Chem 7, 812–826, March 11, 2021 ª 2021 Elsevier Inc.

ll
Article
three-component coupling. To the best of our knowledge, the successful adoption
of such a scenario in divergent synthesis remains elusive in the literature. We report
herein our discovery and development of Ni and Cu co-catalyzed allylic alkylation of
vinyl epoxides and aziridines using terminal alkynes, which represents an interesting
example of these ligand coordination- and dissociation-induced divergent transfor-
mations, as supported by both experimental observations and detailed DFT
calculations.
Transition-metal-catalyzed allylic substitution reactions are of significant utility in
organic synthesis.6–11 Various nucleophiles have been reported to undergo regio-
and enantio-selective allylic alkylation, providing access to a wide range of valuable
chiral building blocks. However, the use of alkynyl nucleophiles in allylic substitution
has remained underdeveloped, even though the 1,4-enyne structure that is acces-
sible from this transformation is a versatile synthon in chemical synthesis. To date,
only a few examples of allylic alkynylation were reported under the catalysis of cop-
per or iridium to produce the branched 1,4-enynes in an enantioselective fashion
(Figure 1B).12–16 Alternatively, Pd-catalyzed dehydrative coupling of alkynes with
allylic alcohols was also documented to deliver linear 1,4-enynes.17 However, the
products from these systems are limited to simple 1,4-enynes, and the access to
related compounds bearing additional functionalities is still highly desired. We
were intrigued by this challenge and set out to explore allylic alkynylation of func-
tionalized substrates — such as vinyl epoxides,18 vinyl ethylene carbonates,19 and
vinyl aziridines20 — which present attractive, atom-economical access to syntheti-
cally versatile alcohol- or amine-containing skipped enynes. Notably, allylic alkyny-
lation of these well-explored substrates with high regio- and stereo-selectivity still
remain elusive in the literature. With an aim to promote a more sustainable process,
we were particularly interested in the use of terminal alkynes as the nucleophiles and
base metals, such as nickel, as the catalysts for our method development.21
1Department of Chemistry, National University of
Singapore, 3 Science Drive 3, Singapore 117543,
Singapore
We report herein not only our development of the first allylic alkynylation of vinyl ep-
oxides, carbonates, and aziridines to yield functionalized branched 1,4-enynes
(product A; Figure 1C) but also our discovery of ligand-induced switch of reactivity
to deliver enyne-containing allyl alcohols and amines (product B; Figure 1C). In
our studies, dual catalysis of nickel and copper proved to be essential for the high
efficiency and selectivity for the reactions, providing an important example of coop-
erative bimetallic catalysis22,23 using nonprecious metals. The remarkable scope of
this system, the use of commercially available alkynes, and the abundance of nickel
and copper catalysts make this a practical method in chemical synthesis. Highly
enantioselective allylic alkynylation has also been achieved. Importantly, DFT calcu-
lations provided key insight on the cause of this divergent bimolecular or trimolec-
ular reactivities, making it a remarkable demonstration of the effect of ligand coor-
dination and dissociation on substrate participation in transition metal catalysis.
2School of Pharmacy, Xi’an Jiaotong University,
No. 76, Yanta West Road, Xi’an 710061, China
3Ministry of Education Key Laboratory for the
Synthesis and Application of Organic Functional
Molecules, College of Chemistry and Chemical
Engineering, Hubei University, Wuhan 430062,
China
4School of Chemistry and Chemical Engineering
and Chongqing Key Laboratory of Theoretical
and Computational Chemistry, Chongqing
University, Chongqing 400030, China
5Chongqing Key Laboratory of Environmental
Materials and Remediation Technologies,
Chongqing University of Arts and Sciences,
Chongqing 402160, China
6College of Chemistry and Institute of Green
Catalysis, Zhengzhou University, Zhengzhou
450001, China
RESULTS AND DISCUSSION
Establishing divergent allylic alkylation of vinyl epoxide using alkynes
7Joint School of National University of Singapore
and Tianjin University, International Campus of
Tianjin University, Binhai New City, Fuzhou
350207, China
To identify an efficient and regioselective catalytic system for allylic alkynylation, we
decided to first explore reactions using the strong nucleophile potassium alkynyltri-
fluoroborate 2a with vinyl epoxide 1a. Despite the fact that much work has been
done on vinyl epoxides,24 highly regio- and stereo-selective transition-metal-cata-
lyzed allylic alkylation using carbon-based nucleophiles is still underdeveloped.25,26
The representative optimization studies are summarized in Figure 2A. Out of the
various metal complexes examined, the use of cobalt/scandium salts or Pd2(dba)3
8These authors contributed equally
9Lead contact
*Correspondence: lanyu@cqu.edu.cn (Y.L.),
zhaoyu@nus.edu.sg (Y.Z.)
https://doi.org/10.1016/j.chempr.2021.02.018
Chem 7, 812–826, March 11, 2021 813

ll
Article
A
B
C
Figure 1. Divergent reactivities induced by coordination or dissociation of ligands
(A) Bimolecular versus trimolecular divergent reactivities determined by the nature of ligand coordination.
(B) Previous allylic alkynylation delivered linear or branched isomeric skipped enynes.
(C) Our discovery of ligand-induced Ni-catalyzed allylic alkylation of vinyl epoxides, carbonates, and aziridines using alkynes: bimolecular allylic
alkynylation with strongly coordinating bisphosphine ligand versus trimolecular alkyne dimerization/allylic vinylation with hemilabile P,N-ligand.
led to no reaction at all, whereas using CuI, Ni[cod]2, or NiBr2 resulted in moderate
reactivity and poor regioselectivity. When PdCl2(PPh3)2 was used, a perfectly
branched regioselectivity could be realized; however, the efficiency remained low.
To our delight, high reactivity and regioselectivity could be observed in the pres-
ence of NiCl2(PPh3)2 to yield 3a in a high yield of 73%.
With the nickel catalysis for allylic alkynylation of 1a established, we then switched to
exploring our desired transformation between 1a and commercially available al-
kynes, which possesses much practical advantage (Figure 2B). We first examined
the reaction of 1a with phenylacetylene 5a by using NiCl2(PPh3)2 or Ni[cod]2
together with CuI as a co-catalyst for alkyne activation. Unfortunately, no conversion
to the desired product was observed at all. Through a systematic screening of
various conditions, we were excited to find out that the ligand had a significant effect
on the reactivity. By the use of DPPP with Ni[cod]2 and CuI, the alkynylation pro-
ceeded smoothly to deliver 3a and 4a in good yields but in a moderate branch to
linear ratio (66% versus 17%, first column in Figure 2B). Control reactions were car-
ried out to examine the effect of the dual catalysts at this point (second and third col-
umns in Figure 2B). The reaction without CuI only proceeded to a lower conversion,
although the regioselectivity of this reaction remained the same (39% and 10% for 3a
814 Chem 7, 812–826, March 11, 2021

ll
Article
A
B
C
D
Figure 2. Discovery, optimization, and control experiments for divergent synthesis
(A) Validation of different metals to promote allylic alkynylation of vinyl epoxide using alkynyltrifluoborate 2 identified nickel as an optimal choice.
(B) Examination of various ligands for Ni/Cu bimetallic catalysis identified divergent reactions to deliver 3a or 6a. While stable bidentate ligands favored
the formation of 3a/4a, the use of mono-dentate or hemilabile bidentate ligands led to the formation of trimolecular reaction product 6a.
(C) The presence of copper co-catalyst in the reactions using DPPP ligand led to only improved reactivity with the same level of regioselectivity for allylic
alkynylation. In sharp contrast, the presence of copper in the reactions using P,N-ligand completely changed a non-selective formation of three
products to, exclusively, formation of 6a.
(D) Control reactions for the formation of 6a. The attempted reaction of 4a and 5a under the standard conditions led to no formation of 6a, which ruled
out the possibility of a hydroalkynylation of 4 for the formation of 6. The use of deuterated 5a led to a clean conversion to 6a with deuterium on the enyne
alkene, which is consistent with alkyne dimerization followed by allylic alkylation mechanism leading to 6a formation.
and 4a, respectively). In contrast, no reaction was observed at all in the absence of Ni
[cod]2, which should be involved in the key Ni-p-allyl intermediate formation.
Clearly, the Ni/Cu dual catalysis was needed to achieve high efficiency in this allylic
alkynylation.
In order to further improve the regioselectivity to favor the branched product 3a, a more
thorough screening of ligands was carried out. While the commonly used bisphosphines
such as DPPF, BINAP, and DPEPhos completely shut down the alkynylation reaction, the
use of NiXantphos with a relatively large bite angle resulted in much higher
Chem 7, 812–826, March 11, 2021 815

ll
Article
regioselectivity (69% and 12% for 3a and 4a). The use of Xantphos further enhanced the
regioselectivity (ꢀ20:1 ratio) and produced 3a in a high yield of 83%. The catalyst
loading was successfully reduced to 5 mol % Ni/Cu and 10 mol % Xantphos to yield
3a with the same yield and regioselectivity (last blue column in Figure 2B).
Based on these results, we were curious whether the use of more flexible ligands
could lead to selective formation of the linear product 4a. Surprisingly, an intriguing
switch of reactivity to produce an enyne-containing allyl alcohol 6a was observed by
the use of monodentate P(Et)3 or N-heterocyclic carbene ligand IPr (first two pink col-
umns in Figure 2B). Further screening of other types of ligands identified the P,N-
ligand Me-PHOX as the optimal choice, which produced 6a exclusively. Under opti-
mized conditions using a lower catalyst loading of 5 mol % with extended reaction
time of 48 h and 3:1 ratio of 5a versus 1a, desired product 6a was obtained in a
good yield of 79% (last pink column in Figure 2B).
Control experiments were carried out to shed light on the operating mechanism of
this Ni-catalyzed divergent allylic alkylation using terminal alkynes. As shown in Fig-
ure 2C, by the use of bisphosphine ligands such as DPPP, the CuI co-catalyst only
improved the efficiency while maintaining the regioselectivity. In sharp contrast,
while a Ni/Cu cooperative system using Me-PHOX as the ligand produced 6a exclu-
sively, the omission of CuI led to a complete loss of selectivity, resulting in a ꢀ1:1:1
ratio of 3a, 4a, and 6a. The effect of copper is clearly crucial for this transformation.
More control experiments provided some evidence for the mechanism of 6a forma-
tion (Figure 2D). Hypothetically, 6a should be formed through an alkyne dimeriza-
tion followed by terminal allylic alkylation. To rule out an alternative sequence of ter-
minal allylic alkynylation followed by hydroalkynylation of the internal alkyne
intermediate, we attempted the reaction of pre-synthesized terminal allylic alkynyla-
tion product 4a with 5a under the standard Ni/Cu-catalyzed conditions. As shown in
Equation 1, no formation of 6a was observed at all, which disfavored the possibility
of a late hydroalkynylation step. In addition, deuterium labeled 5a was subjected to
the standard reaction, leading to the formation of 6a with a clean deuterium labeling
at the vinylic position. This was also consistent with the mechanism of alkynylnicke-
lation of a terminal alkyne before allylic alkylation.
Understanding divergent allylic alkylations using DFT calculations
To further probe the underlying causes of these divergent reactivities, we carried out
DFT27 calculations to investigate the mechanism for the formation of 3a, 4a, and 6a
for this Ni-catalyzed allylic alkylation reaction. The calculated results are shown in
Figures 3 and 4.
The mechanism for the formation of 3a and 4a catalyzed by Xantphos-Ni was calcu-
lated first. The calculated Gibbs energy profiles for 3a versus 4a formation are given
in Figure 3, where Xantphos-Ni(0) species A was set as the relative zero point. Ligand
exchange of vinyl epoxide 1a with cyclooctadiene in A gave p-complex B with
8.8 kcal/mol exergonic. The subsequent oxidative addition of vinyl epoxide onto
Ni(0) center occurred via transition state with an energy barrier of 13.0 kcal/mol,
from which the penta-coordinate allyl-Ni(II) intermediate C was generated. At this
stage, the Ni-complex could engage alkyne 5a directly to form the nickel acetylide
intermediate D through TS2 with a free energy barrier of 18.1 kcal/mol. In this step,
the alkoxide moiety served an important role as base for the deprotonation of 5a.
Alternatively, the inclusion of copper co-catalyst as a p-acid to 5a (in the form of
E) greatly facilitated the deprotonation by Ni-alkoxide C via TS3 with a minimal
816 Chem 7, 812–826, March 11, 2021

ll
Article
Figure 3. DFT calculations on 3a versus 4a formation catalyzed by Ni/Cu-Xanthpos
Following the formation of Ni-p-allyl alkoxide complex C, incorporation of Cu-alkyne complex E proceeds via alkoxide base-assisted ligand exchange
to give Ni-acetylide D. At this stage, direct C(alkynyl)-C(allyl internal) reductive elimination takes place via TS4 to produce 3a. For the formation of
alternative regioisomer 4a, a conformational change is required via H and TS5 to generate Ni-p-allyl acetylide complex I/J before reductive elimination
can take place. This provided a convincing rationale for the regio-selectivity (3 versus 4) for Ni-catalyzed allylic alkynylation.
activation barrier of 0.3 kcal/mol; the resultant intermediate G then underwent trans-
metalation with alkynyl copper F to yield alkynyl nickel D exergonic. The(alkynyl)-
C(allyl internal) reductive elimination then took place via TS4 with a barrier of
18.4 kcal/mol to give the internal alkylation product 3a. In TS4, the length of forming
˚
C(alkynyl)-C(allyl internal) bond is 1.89 A. Ligand exchange of vinyl epoxide 1a with
product 3a then gave intermediate B to complete the catalytic cycle.
For the formation of linear product 4a, direct reductive elimination of intermediate D
using the terminal carbon on the allyl unit was not viable due to the anti-relationship
of C(alkynyl) and C(allyl terminal). A conformational change had to take place first. As
shown by the gray potential energy surface in Figure 3, a ligand exchange between
hydroxyl and terminal alkenyl in intermediate D gave complex H with 13.4 kcal/mol
endergonic. The subsequent rotation of allyl unit in H then occurred via TS5, with an
free energy barrier of 7.3 kcal/mol, to generate the allyl-Ni(II) complex I, which
readily converted to J. The relatively high activation barrier in this step (D to I) is
most likely due to the involvement of a Ni-complex in which only one phosphine
in Xantphos kept coordination. The following C(alkynyl)-C(allyl) reductive elimina-
tion then occurred via TS6 with a barrier of 10.0 kcal/mol to give the terminal alkyl-
˚
ation product 4a. In TS6, the length of forming C(alkynyl)-C(allyl) bond is 1.88 A.
These calculated results show that the relative energy of TS5 is 2.3 kcal/mol higher
than that of TS4, which is fully consistent with the experimental observations that 3a
was formed as the major regioisomer (3a:4a ꢀ20:1).
DFT calculations were also carried out to better understand the mechanism for 6a
formation by the use of Ni-MePHOX in combination with CuI (Figure 4). On the basis
Chem 7, 812–826, March 11, 2021 817

ll
Article
A
B
C
D
E
818 Chem 7, 812–826, March 11, 2021

ll
Article
Figure 4. Proposed pathway for 6a formation and DFT calculations on 5a dimerization, 6a, or 3a/4a formation catalyzed by Ni/Cu-Me-PHOX
(A) The proposed pathways for 6a formation catalyzed by Ni/Me-PHOX. The typical stepwise alkyne dimerization followed by allylic alkylation (pathway
A) was determined to be unfavorable by DFT calculations. The alternative trimolecular reaction (pathway B) was identified to be the favorable reaction
pathway.
(B) The dimerization of 5a catalyzed by Ni/Me-PHOX was energetically unfavorable as determined by DFT calculations.
(C) The Gibbs energy profiles for 6a formation through a trimolecular Ni-catalyzed pathway assisted by Cu co-catalyst. Ligand dissociation from D0 takes
place to allow the formation of key intermediate P with the coordination of three molecules of substrates.
(D) The Gibbs energy profiles for 3a and 4a formation catalyzed by Ni/Me-PHOX as a comparison.
(E) Structure information of the key intermediates for the reaction profiles.
of the mechanistic studies (Figure 2D), alkyne dimerization should precede allylic
alkylation. A stepwise pathway A (Figure 4A) was considered first, which includes
oxidative addition of alkyne 5a onto Ni(0) A0, migratory insertion to another 5a fol-
lowed by allylic alkylation of 1a. To probe the feasibility of this mechanism, the Gibbs
energy profiles for 5a dimerization using nickel/copper cooperative catalysis was
calculated first (Figure 4B). Ligand exchange of phenylacetylene 5a with cycloocta-
diene in A0 gave p-complex K with exergonic of 16.9 kcal/mol. Subsequent oxidative
addition of a C–H bond in 5a onto Ni(0) center occurred via transition state TS7 with a
free energy barrier of 17.1 kcal/mol to generate the alkynyl-Ni(II) intermediate L. The
following migratory insertion of phenylacetylene into C(alkynyl)-Ni(II) bond, acti-
vated by Cu(I) as a p-acid, took place via transition state TS8 to generate the dimer
alkenyl-Ni(II) intermediate N. The overall free energy barrier for 5a dimerization is
40.7 kcal/mol. Such an inhibitively high barrier indicates that this pathway is unlikely
the operating mechanism for the formation of 6a. As the high activation barrier for 5a
dimerization partly comes from an unfavorable oxidative addition of terminal alkyne
onto Ni(0) (TS7 in Figure 4B), we hypothesized an alternative mechanism as shown in
pathway B in Figure 4A. The key factor was the involvement of vinyl epoxide at the
beginning to convert Ni(0) to Ni-p-allyl alkoxide complex, which then engages
alkyne 5a via alkoxide-assisted ligand exchange, analogous to that in Figure 3.
Another alkyne coordination to the Ni-acetylide then set the stage for migratory
insertion followed by allylic alkylation to form 6a. DFT calculations were then carried
out to probe the feasibility of this mechanism.
As shown in Figure 4C, coordination of vinyl epoxide 1a to A0 to give B0, oxidative
addition to give C0, deprotonation of E0 with C0 followed by transmetalation pro-
duced alkynyl-Ni(II) complex D0 with a similar potential energy surface to that in Fig-
ure 3 (via TS30). It is noteworthy that the nitrogen atom in Me-PHOX does not coor-
dinate with the Ni(II) center in D0. At this stage, coordination of another alkyne
directly could form intermediate O with 24.6 kcal/mol endergonic. In contrast, an
alternative alkyne coordination with the release of Me-PHOX ligand was identified
to be energetically favorable, leading to P via TS9 with an energy barrier of
19.0 kcal/mol. The hemilabile property of this P,N-ligand was a key factor for this
intriguing new pathway to take place.
From intermediate P, migratory insertion of the acetylide ligand onto the alkyne,
activated by Cu(I) as a p-acid, occurred via transition state TS10 to give the al-
kenyl-Ni(II) intermediate Q with a free energy barrier of 4.1 kcal/mol. In TS10, the
˚
lengths of the forming C-C bond and C-Ni(II) bond are 1.96 and 1.94A, respectively.
Re-association of the Me-PHOX ligand with nickel center in Q to replace the alkyne
coordination generated intermediate R with 9.2 kcal/mol exergonic. The following
C(alkenyl)-C(allyl terminal) reductive elimination occurred via transition state TS11
with a barrier of 11.9 kcal/mol to give terminal alkenylation 6a. In transition state
˚
TS11, the length of the forming C(alkenyl)-C(allyl) bond is 1.89 A. For the formation
of internal alkenylation product 6a0, reductive elimination had to take place via
Chem 7, 812–826, March 11, 2021 819

ll
Article
A
B
Figure 5. Proposed divergent catalytic pathways for the formation of 3/4 or 6 using bisphosphine or P,N-ligand
transition state TS12 bearing a h-1 allyl unit with a free energy barrier of 20.3 kcal/
mol. The much higher free energy of TS12 versus TS11 (8.4 kcal/mol) was consistent
with experimental observation of the exclusive formation of linear 6a (without the
corresponding branched regioisomer). Overall, this pathway proceeded through
an intriguing mechanism involving the dynamic association and dissociation of the
hemilabile P,N-ligand, enabling an efficient trimolecular reaction.
To provide further support for the exclusive formation of 6a (but not 3a/4a) under our
catalytic conditions, the formation of 3a and 4a catalyzed by Ni-MePHOX was also
calculated. As illustrated in Figure 4D, a similar pathway to that in Figure 3 was deter-
mined to be the case. Notably, the weak coordination capacity of the nitrogen atom
in Me-PHOX resulted in a similar free energy of TS40 and TS50 that would lead to the
formation of 3a and 4a as a mixture. However, the activation barrier of TS10 for 6a
formation (Figure 4C) was much lower than that of TS40 and TS50, which is experimen-
tally consistent with the exclusive formation of 6a.
As shown by the control experiment in Figure 2C, right, the presence of copper co-
catalyst made a significant difference for the product distribution of this reaction us-
ing Ni-MePHOX catalyst. Instead of exclusive 6a formation in the presence of cop-
per, 3a, 4a, and 6a were formed in equal amounts in the absence of CuI. To better
understand the effect of CuI, the activation barrier for the migratory insertion step
in 6a formation was further calculated without activation by copper. As shown by
TS13 in Figure 4C, the activation barrier (24.2 kcal/mol from D0) was determined
to be comparable to that of TS40 and TS50 (22.2/23.0 kcal/mol from D0). Thus, this
agrees well with the mechanistic studies with or without copper co-catalyst.
To provide a full picture of this ligand-induced divergent reaction system, the ligand-
dissociation pathway for trimolecular reaction with XantPhos, experimentally not
observed, was also probed by DFT calculation (see Figure S132). The activation barrier
for the ligand-dissociation pathway analogous to that in Figure 4C with bisphosphine
was determined to be much less favorable. Overall, these data from DFT calculations
provided a conclusive overall picture for this ligand-induced divergent allylic alkylation.
Based on the above DFT calculations, the catalytic pathways for the ligand-induced
divergent synthesis of 3/4 or 6 are summarized in Figures 5A and 5B for direct
820 Chem 7, 812–826, March 11, 2021

ll
Article
comparison. Most notably, the dynamic dissociation of hemilabile P,N-ligand from
VI0 enabled the formation of the key intermediate VII (Figure 5B), from which a trimo-
lecular reaction takes place, leading to the formation of product 6.
Scope of divergent allylic alkylation of vinyl epoxides and aziridines with
alkynes
With the optimal conditions in hand, we turned our attention to exploring the scope
of these ligand-induced divergent allylic alkylations. As illustrated in Figure 6, it is
important to note that the only variation in the reaction parameters leading to
different product formation was the identity of the ligand.
Allylic alkynylationofvinylepoxidestoyieldskippedenyneswasexploredfirst. The scope
of terminal alkynes for this transformation turned out to be remarkably broad (Figure 6A).
Under the same set of optimal conditions, a wide range of phenylacetylene derivatives 5
bearing substituents of different electronic characters on different positions of the aryl
ring underwent efficient alkynylation to produce 3a–3g in good to high yields. Terminal
alkynes bearing heterocyclic substituents, such as pyridine, imidazole, and thiophene, all
participated in the alkynylation reaction smoothly to yield 3h–3j in high yields. Further-
more, alkenyl-, alkyl-, and silyl-substituted alkynes were good substrates as well (70%–
83% for 3k–3m). More importantly, a wide spectra of common functionalities such as
chloro-, cyano-, acetyl-, and hydroxyl-amines were all well tolerated to deliver 3n–3u in
65%–85% yields. The generality of this method prompted us to examine alkyne-contain-
ing complex drug compounds for the alkynylation as well. As a representative example,
the tyrosine kinase inhibitor erlotinib28 underwent a smooth reaction to yield 3v in 66%
yield. With a simple procedure utilizing commercially available catalysts and reagents,
we are confident that this method may find wide use as a reliable modification of terminal
alkynes bearing diverse substituents.
To our delight, this catalytic system was also successfully extended to the alkynyla-
tion of vinyl aziridine 7 (Figure 6B). Under the same set of optimal conditions, various
terminal alkynes participated in the reaction smoothly. As shown by the representa-
tive examples, aryl-, heteroaryl-, and alkyl-substituted terminal alkynes were all suit-
able nucleophiles to produce the amine-containing 1,4-enynes 8 in good efficiency
and excellent regioselectivity.
The scope of the synthesis of dienynes 6 was explored next (Figure 6C). It is noteworthy
that a single geometric isomer at the tri-substituted alkene was observed for all cases. A
wide range of aryl- and heteroaryl-alkynes participated in this reaction to produce 6a–6k
in good to excellent yields. Alkenyl-substituted alkynes also could be used to yield 6l and
6m in good yields. As one limitation for this series, alkyl-substituted alkynes failed to un-
dergo reactions to produce the desired products 6. Moreover, we were delighted to find
that this transformation could also tolerate substituted vinyl epoxides. Compounds 6n
and 6o could be produced from 1b and 1c in good to high yields. The conversion of vinyl
aziridine 7 to allyl amine 9 also proceeded smoothly under the standard conditions. The
general scope, coupled with a simple procedure, makes this reaction a useful method for
the preparation of these synthetically valuable functionalized enyne-containing allyl alco-
hols and amines.
Allylic alkynylation of vinyl ethylene carbonates to yield products bearing a
quaternary center
To further expand the scope of this methodology, we aimed toward the synthesis of
skipped enynes bearing a quaternary center (10 in Figure 7). However, these efforts
were met with much trouble by the use of substituted vinyl epoxide, such as 1d,
Chem 7, 812–826, March 11, 2021 821

ll
Article
A
C
B
Figure 6. The scope of divergent allylic alkylation of vinyl epoxides/aziridines with alkynes
The reactions were carried out on a 0.2 mmol scale. Isolated yields of the pure products were reported. See Supplemental information for the detailed
reaction conditions.
(A) In the presence of Xantphos ligand, highly efficient and regioselective allylic alkynylation of vinyl epoxides is achieved with a wide range of alkynes
bearing various functionalities. The reactions were carried out using 2 equiv of alkynes versus vinyl epoxides. aFor this reaction, NiXantphos was used as
the ligand due to the separation issue of Xantphos from the desired product. The use of TBME as the solvent was beneficial over THF to yield 3a in 75%
isolated yield.
(B) Allylic alkynylation of vinyl aziridines worked with the same efficiency and selectivity to deliver skipped enynes bearing an amine functionality.
(C) In the presence of Me-PHOX ligand, a wide range of alkynes and vinyl epoxides and aziridines underwent smooth reaction to deliver the trimolecular
products 6 and 9 in high efficiency. The reactions were carried out using 3 equiv of alkynes versus vinyl epoxides.
which underwent epoxide to aldehyde rearrangement under the standard condi-
tions instead of alkynylation to deliver the desired 1,4-enyne 10a (Figure 7A). In an
effort to overcome this problem, we turned our attention to the related vinyl
ethylene carbonate (VEC) 11, which is known to be less prone to rearrangement
822 Chem 7, 812–826, March 11, 2021

ll
Article
A
C
B
D
Figure 7. Reactions of vinyl ethylene carbonates (VECs) to access products bearing a quaternary center and enantioselective allylic alkynylation
(A) The allylic alkynylation of substituted vinyl epoxide 1d failed due to alternative side reaction.
(B) By the use of aryl-substituted VECs as the substrate, the allylic alkynation proceeded smoothly to yield 10 bearing a quaternary center in good to high
efficiency. Further optimization identified DPPF and DABCO as the best choices of ligand and base. See Supplemental information for detailed reaction
conditions.
(C) The allylic alkynylation of enantiopure 1a led to the formation of racemic product 3e, suggesting a stereoablative pathway via Ni-p-allyl
intermediate.
(D) An enantioconvergent allylic alkynylation of VECs 11 produced enantioenriched skipped enynes 3 in good to excellent yields and
enantioselectivities. (R)-SIPHOS was identified to be the optimal chiral ligand for this transformation. A lower temperature of ꢁ20ꢂC was needed to
achieve high enantioselectivity. Under these conditions, the branched to linear ratio ranged from 2.5:1 to 5:1, and the pure branched product 3 was
isolated in pure form. See Supplemental information for detailed reaction conditions. aThe reaction used 10 mol % nickel and copper and 20 mol %
ligand and the reaction time was 120 h. bAmbient temperature was used for this reaction due to the low solubility of this terminal alkyne substrate.
under Lewis acidic conditions and may provide a different chemo-selectivity to favor
the desired allylic alkynylation. It is noteworthy that allylic substitution of these com-
pounds using heteroatom-based nucleophiles including enantioselective variants
has been well-established in recent years.29–34 In contrast, the use of carbon-nucle-
ophiles in this reaction remains rare in the literature.35
The allylic alkynylation of 11a (R0 = Ph) using 5a under the standard conditions from
Figure 6A turned out to be very sluggish, leading to minimal formation of the desired
10a. After much optimization (see Supplemental information for details), the use of
DPPF as the ligand and a stronger base—DABCO—proved to be important for an
Chem 7, 812–826, March 11, 2021 823

ll
Article
efficient Ni/Cu-catalyzed alkynylation of 11 to deliver 10 in high efficiency. As shown
in Figure 7B, 10a was produced in a good yield of 82%. Similar to the previous sub-
strate scope, alkynes bearing different aryl, heteroaryl, alkenyl, and silyl substituents
were all well tolerated in this reaction to yield 10b–10m in good to high yields. In
addition, various aryl-substituted VECs underwent reactions smoothly to deliver
10n–10r in moderate to good yields. VEC with a cyclohexyl group could also partic-
ipate in this reaction, albeit with a low yield for 10s.
Enantioconvergent allylic alkynylation of vinyl ethylene carbonates
In nickel-catalyzed allylic alkylation or cross coupling of aryl ethers, both pathways of ste-
reospecifictransferofthe substrate chiralitytothe product36 and stereoablative synthesis
of enantioenriched products have been reported.37 When we tested allylic alkynylation
using the enantiopure vinyl epoxide 1a under the standard conditions (Figure 7C), prod-
uct 3e was obtained in a racemic form in 70% yield. This suggests that the nickel-p-allyl
intermediate for this catalytic alkynylation is prone to racemization. This observation
prompted us to explore an enantioconvergent synthesis of the valuable skipped enynes.
Again,muchoptimizationwithvinylepoxide1aledtolowconversiontothedesiredprod-
uct 3a by the use of various chiral ligands. Further extensive optimization with VEC 11h
(see Supplemental information for details) identified that the use of (R)-SIPHOS and
DABCO at ꢁ20ꢂC produced 3a in good yield and high enantioselectivity (70% yield,
94.5:5.5 er; Figure 7D). The scope of this enantioselective allylic alkynylation was then
explored. A broad scope of terminal alkynes was achieved. Arylalkynes bearing different
substituents underwent efficient alkynylation to yield 3a–3g with up to 98:2 er. Pyridyl-
containing alkyne worked similarly well to produce 3h in 93.5:6.5 er, albeit with a slightly
reduced yield. Alkynes bearing alkyl substituents turned out to be much less reactive. A
higher catalyst loading plus longer reaction time or higher temperature was necessary to
produce 3l and 3t in reasonable yield and good 91:9 or 92:8 er. The enantioselective al-
kynylation of substituted VEC 11a was also attempted. However, no reactivity was ob-
tained for these bulkier substrates. It is noteworthy that in these enantioselective reac-
tions, the branched to linear regioselectivity was lower than the racemic series in
Figure 6A. However, the pure branched products could be easily obtained by chroma-
tography separation. The absolute configuration of 3t was assigned unambiguously by
single crystal X-ray analysis.
Conclusions
We present herein an unprecedented Ni/Cu-catalyzed allylic alkynylation of vinyl ep-
oxides, vinyl carbonates, and vinyl aziridines using readily available terminal alkynes.
The scope of this catalytic method, especially toward the alkynes, is remarkably
broad. The enantioselective variant of this process has also been realized. By switch-
ing to a hemilabile P,N-ligand, the formation of enyne-containing allyl alcohols and
amines from a trimolecular reaction was also realized with a broad scope. For this
intriguing ligand effect leading to divergent bimolecular or trimolecular reactivities,
evidence from DFT calculations pointed to the dynamic ligand coordination/disso-
ciation, which influenced the number of open sites on the metal center at an appro-
priate stage in the catalysis to bind varied number of substrates and induce different
reactivities. With this important discovery, further application of this valuable princi-
ple for divergent transition metal catalysis will be explored in our laboratories.
EXPERIMENTAL PROCEDURES
Resource availability
Lead contact
Further information and requests for resources should be directed to and will be ful-
filled by the lead contact, Yu Zhao (zhaoyu@nus.edu.sg).
824 Chem 7, 812–826, March 11, 2021

ll
Article
Materials availability
All unique reagents generated in this study are available from the Lead Contact
without restriction.
Data and code availability
All data needed to support the conclusions of this manuscript are included in the
main text or Supplemental information. Crystallographic data generated during
this study have been deposited in the Cambridge Crystallographic Data Centre
(CCDC) under accession numbers CCDC:1920222 (3g) and 1920233 (R-3t). These
data can be obtained free of charge from the CCDC at http://www.ccdc.cam.ac.
uk/data_request/cif.
Representative procedure Ni/Cu-catalyzed alkynylation of vinyl epoxides
To a vial equipped with a dried stir bar was added Ni[cod]2 (0.01 mmol), Xantphos
(0.02 mmol), and anhydrous THF (1.0 mL) in the glovebox. The reaction mixture
was then allowed to stir for 15 min, followed by the addition of CuI (0.01 mmol), vinyl
epoxide 1a (0.20mmol), acetylene 5a (0.40 mmol), and DIPEA (0.40 mmol). The re-
action tube was sealed and taken outside the glovebox and allowed to stir at
room temperature for 24 h. The crude reaction mixture was concentrated under
reduced pressure and directly purified by silica gel chromatography (ethyl ace-
tate/hexanes = 1:8).
SUPPLEMENTAL INFORMATION
Supplemental information can be found online at https://doi.org/10.1016/j.chempr.
2021.02.018.
ACKNOWLEDGMENTS
We are grateful for the financial support from Ministry of Education of Singapore (R-
143-000-A94-112), National University of Singapore (R-143-000-A57-114), Xi’an
Jiaotong University Youth Talent Support Program (0800-712110510716), National
Natural Science Foundation of China (21901066, 21822303, and 22001203), and the
111 Project (D20003). We acknowledge the allocation of computing resources at the
Henan Province Supercomputing Center.
AUTHOR CONTRIBUTIONS
Y.H. and C.M. designed and performed the experimental work, with help from L.-
C.Y.; S.L. and Y.L. performed the DFT calculation. Y.Z. directed the project and
wrote the manuscript with Y.H., C.M., S.L., and Y.L.
DECLARATION OF INTERESTS
The authors declare no competing interests.
Received: September 15, 2020
Revised: December 13, 2020
Accepted: February 12, 2021
Published: March 11, 2021
REFERENCES
1. O’Connor, C.J., Beckmann, H.S.G., and Spring,
D.R. (2012). Diversity-oriented synthesis:
producing chemical tools for dissecting
biology. Chem. Soc. Rev. 41, 4444–4456.
2. Mahatthananchai, J., Dumas, A.M., and
Bode, J.W. (2012). Catalytic selective
synthesis. Angew. Chem. Int. Ed. Engl. 51,
10954–10990.
3. Krautwald, S., and Carreira, E.M. (2017).
Stereodivergence in asymmetric catalysis.
J. Am. Chem. Soc. 139, 5627–5639.
Chem 7, 812–826, March 11, 2021 825

ll
Article
4. Beletskaya, I.P., Na´jera, C., and Yus, M. (2018).
Stereodivergent catalysis. Chem. Rev. 118,
5080–5200.
alkylation of terminal alkynes by synergistic Rh/Cu
catalysis. J. Am. Chem. Soc. 142, 15276–15281.
29. Guo, W., Martınez-Rodrıguez, L., Kuniyil, R.,
´ ´
Martin, E., Escudero-Ada´n, E.C., Maseras, F.,
and Kleij, A.W. (2016). Stereoselective and
versatile preparation of tri- and tetrasubstituted
allylic amine scaffolds under mild conditions.
J. Am. Chem. Soc. 138, 11970–11978.
17. Li, Y.X., Xuan, Q.Q., Liu, L., Wang, D., Chen,
Y.J., and Li, C.J. (2013). A Pd(0)-catalyzed direct
dehydrative coupling of terminal alkynes with
allylic alcohols to access 1,4-enynes. J. Am.
Chem. Soc. 135, 12536–12539.
5. Na´jera, C., Beletskaya, I.P., and Yus, M. (2019).
Metal-catalyzed regiodivergent organic
reactions. Chem. Soc. Rev. 48, 4515–4618.
30. Cai, A., Guo, W., Martı´nez-Rodrı´guez, L., and
Kleij, A.W. (2016). Palladium-catalyzed Regio-
and enantioselective synthesis of allylic
amines featuring tetrasubstituted tertiary
carbons. J. Am. Chem. Soc. 138, 14194–
14197.
6. Trost, B.M., and Crawley, M.L. (2003).
Asymmetric transition-metal-catalyzed allylic
alkylations: applications in total synthesis.
Chem. Rev. 103, 2921–2944.
18. He, J., Ling, J., and Chiu, P. (2014). Vinyl
epoxides in organic synthesis. Chem. Rev. 114,
8037–8128.
7. Lu, Z., and Ma, S. (2008). Metal-catalyzed
enantioselective allylation in asymmetricsynthesis.
Angew. Chem. Int. Ed. Engl. 47, 258–297.
`
19. Guo, W., Go´ mez, J.E., Cristo` fol, A., Xie, J., and
Kleij, A.W. (2018). Catalytic transformations of
functionalized cyclic organic carbonates.
Angew. Chem. Int. Ed. Engl. 57, 13735–13747.
31. Khan, A., Khan, S., Khan, I., Zhao, C., Mao, Y.,
Chen, Y., and Zhang, Y.J. (2017).
Enantioselective construction of tertiary C-O
bond via allylic substitution of Vinylethylene
carbonates with water and alcohols. J. Am.
Chem. Soc. 139, 10733–10741.
8. Butt, N.A., and Zhang, W. (2015). Transition
metal-catalyzed allylic substitution reactions
with unactivated allylic substrates. Chem. Soc.
Rev. 44, 7929–7967.
20. Ohno, H. (2014). Synthesis and applications of
vinylaziridines and ethynylaziridines. Chem.
Rev. 114, 7784–7814.
9. Hethcox, J.C., Shockley, S.E., and Stoltz, B.M.
(2016). Iridium-catalyzed Diastereo-, enantio-,
and regioselective allylic alkylation with
32. Khan, A., Zhao, H., Zhang, M., Khan, S., and
Zhao, D. (2020). Regio- and enantioselective
synthesis of sulfone-bearing quaternary carbon
stereocenters by Pd-catalyzed allylic
substitution. Angew. Chem. Int. Ed. Engl. 59,
1340–1345.
21. Tasker, S.Z., Standley, E.A., and Jamison, T.F.
(2014). Recent advances in homogeneous
nickel catalysis. Nature 509, 299–309.
prochiral enolates. ACS Catal 6, 6207–6213.
22. Pye, D.R., and Mankad, N.P. (2017). Bimetallic
catalysis for C-C and C-X coupling reactions.
Chem. Sci. 8, 1705–1718.
10. Cheng, Q., Tu, H.F., Zheng, C., Qu, J.P.,
Helmchen, G., and You, S.L. (2019). Iridium-
catalyzed asymmetric allylic substitution
reactions. Chem. Rev. 119, 1855–1969.
33. Rong, Z.Q., Yang, L.C., Liu, S., Yu, Z., Wang,
Y.N., Tan, Z.Y., Huang, R.Z., Lan, Y., and Zhao,
Y. (2017). Nine-membered benzofuran-fused
heterocycles: enantioselective synthesis by Pd-
catalysis and rearrangement via transannular
bond formation. J. Am. Chem. Soc. 139, 15304–
15307.
23. Wu, Y., Huo, X., and Zhang, W. (2020).
Synergistic Pd/Cu catalysis in organic synthesis.
Chemistry 26, 4895–4916.
11. Dutta, S., Bhattacharya, T., Werz, D.B., and
Maiti, D. (2020). Transition metal catalyzed C-H
allylation reactions. Chem. https://doi.org/10.
1016/j.chempr.2020.10.020.
24. Trost, B.M., Bunt, R.C., Lemoine, R.C., and
Calkins, T.L. (2000). Dynamic kinetic
asymmetric transformation of diene
monoepoxides: a practical asymmetric
synthesis of Vinylglycinol, vigabatrin, and
ethambutol. J. Am. Chem. Soc. 122, 5968–
5976.
12. Dabrowski, J.A., Gao, F., and Hoveyda, A.H.
(2011). Enantioselective synthesis of alkyne-
substituted quaternary carbon stereogenic
centers through NHC-Cu-catalyzed allylic
substitution reactions with (i-Bu)2(alkynyl)
aluminum reagents. J. Am. Chem. Soc. 133,
4778–4781.
34. Wang, Y.N., Yang, L.C., Rong, Z.Q., Liu, T.L.,
Liu, R., and Zhao, Y. (2018). Pd-catalyzed
enantioselective [6+4] cycloaddition of vinyl
oxetanes with azadienes to access ten-
membered heterocycles. Angew. Chem. Int.
Ed. Engl. 57, 1596–1600.
25. Du, C., Li, L., Li, Y., and Xie, Z. (2009).
Construction of two vicinal quaternary carbons
by asymmetric allylic alkylation: total synthesis
of hyperolactone C and (ꢁ)-biyouyanagin A.
Angew. Chem. Int. Ed. Engl. 48, 7853–7856.
13. Hamilton, J.Y., Sarlah, D., and Carreira, E.M.
(2013). Iridium-catalyzed enantioselective
allylic alkynylation. Angew. Chem. Int. Ed. Engl.
52, 7532–7535.
35. Khan, A., Yang, L., Xu, J., Jin, L.Y., and Zhang,
Y.J. (2014). Palladium-catalyzed asymmetric
decarboxylative cycloaddition of vinylethylene
carbonates with Michael acceptors:
construction of vicinal quaternary
26. Ma, C., Huang, Y., and Zhao, Y. (2016).
Stereoselective 1,6-conjugate addition/
annulation of para-quinone methides with vinyl
epoxides/cyclopropanes. ACS Catal 6, 6408–
6412.
14. Harada, A., Makida, Y., Sato, T., Ohmiya, H.,
and Sawamura, M. (2014). Copper-catalyzed
enantioselective allylic alkylation of terminal
alkyne pronucleophiles. J. Am. Chem. Soc. 136,
13932–13939.
stereocenters. Angew. Chem. Int. Ed. Engl. 53,
11257–11260.
36. Zhou, Q., Srinivas, H.D., Zhang, S., and
Watson, M.P. (2016). Accessing both
retention and inversion pathways in
stereospecific, nickel-catalyzed Miyaura
borylations of allylic pivalates. J. Am. Chem.
Soc. 138, 11989–11995.
27. Frisch, M.J., Trucks, G.W., Schlegel, H.B.,
Scuseria, G.E., Robb, M.A., Cheeseman, J.R.,
Scalmani, G., Barone, V., Petersson, G.A.,
Nakatsuji, H., et al. (2013). Gaussian 09, revision
D.01, Gaussian. The full author list is shown in
Supporting information.
15. Cui, X.Y., Ge, Y., Tan, S.M., Jiang, H., Tan, D.,
Lu, Y., Lee, R., and Tan, C.H. (2018). (Guanidine)
copper complex-catalyzed enantioselective
dynamic kinetic allylic alkynylation under
biphasic condition. J. Am. Chem. Soc. 140,
8448–8455.
37. Son, S., and Fu, G.C. (2008). Nickel-catalyzed
asymmetric Negishi cross-couplings of
secondary allylic chlorides with alkylzincs.
J. Am. Chem. Soc. 130, 2756–2757.
28. Dowell, J., Minna, J.D., and Kirkpatrick, P.
(2005). Erlotinib hydrochloride. Nat. Rev. Drug
Discov. 4, 13–14.
16. Huang, W.Y., Lu, C.H., Ghorai, S., Li, B., and Li, C.
(2020). Regio- and enantioselective allylic
826 Chem 7, 812–826, March 11, 2021