Highly Enantioselective Cobalt-Catalyzed (3+2) Cycloadditions of Alkynylidenecyclopropanes

Article abstract of DOI:10.1002/anie.202015202

Low-valent cobalt complexes equipped with chiral ligands can efficiently promote highly enantioselective (3+2) cycloadditions of alkyne-tethered alkylidenecyclopropanes. The annulation allows to assemble bicyclic systems containing five-membered rings in good yields and with excellent enantiomeric ratios. We also present a mechanistic discussion based on experimental and computational data, which support the involvement of Co^I/Co^III catalytic cycles.

Full text of DOI:10.1002/anie.202015202

Angewandte
Research Articles
Chemie
How to cite: Angew. Chem. Int. Ed. 2021, 60, 8182–8188
International Edition:
German Edition:
doi.org/10.1002/anie.202015202
doi.org/10.1002/ange.202015202
Homogeneous Catalysis
Highly Enantioselective Cobalt-Catalyzed (3+2) Cycloadditions of
Alkynylidenecyclopropanes
Eduardo Da Concepción, Israel Fernµndez, JosØ L. MascareÇas,* and Fernando López*
Abstract: Low-valent cobalt complexes equipped with chiral
ligands can efficiently promote highly enantioselective (3+2)
cycloadditions of alkyne-tethered alkylidenecyclopropanes.
The annulation allows to assemble bicyclic systems containing
five-membered rings in good yields and with excellent
enantiomeric ratios. We also present a mechanistic discussion
based on experimental and computational data, which support
the involvement of CoI/CoIII catalytic cycles.
cycloadditions between ACPs and tethered alkenes or 1,3-
dienes (Scheme 1a).[8] Unfortunately, the success of these
reactions is limited to substrates featuring electron-with-
drawing groups at the terminal position of the alkene (or
diene), and the enantiomeric ratios (er) are strongly depen-
dent on the nature of the connecting tether. Therefore, the
development of new, efficient and versatile enantioselective
annulations of ACPs remains a major goal in current synthetic
chemistry.
Introduction
In this context, we describe herein a highly enantioselec-
tive formal (3+2) intramolecular cycloaddition of alkynyli-
dene-cyclopropanes (Scheme 1b). The reaction, which is
promoted by non-expensive cobalt catalysts featuring chiral
bisphosphine ligands, is the first enantioselective cycloaddi-
tion of ACPs catalyzed by an earth-abundant first-row metal
complex, and the first (3+2) cycloaddition of ACPs promoted
by cobalt.[9] Moreover, it also constitutes one of the very few
asymmetric (3+2) annulations promoted by cobalt re-
agents.[10]
Transition-metal-catalyzed (TMC) cycloadditions are
among the most practical and versatile methods to construct
polycyclic products from simple acyclic materials, in an atom
economical manner.[1] A number of enantioselective variants
have been disclosed throughout the last decades, most of
which relying on the use of second- and third-row precious
metal catalysts.[2,3]
Among the different cycloaddition strategies developed
À
so far, those involving the activation of C C bonds of
The method provides a straightforward approach to five-
membered fused bicyclic scaffolds (2) from simple, readily
accessible ACP precursors (1, Scheme 1b).[11] Remarkably,
we have found that the performance of the Co catalysts is
much superior to that of Pd or Rh alternatives. We also
demonstrate that a simple hydrogenation protocol allows the
stereospecific transformation of the adducts into enantioen-
riched fused-bicarbocyclic systems bearing up to four stereo-
centers. Importantly, we include experimental and computa-
tional results that allow us to propose a CoI/CoIII catalytic
cycle encompassing cobaltacyclopentene and p-allyl cobalt
designed carbocyclic reactants are especially attractive.[4] In
this context, we and others have shown that readily accessible
alkylidenecyclopropanes (ACPs) can behave as versatile
three-carbon components in a wide variety of synthetically
powerful (3+n) and (3+n+m) annulations, promoted by Pd,
Ni, Ru, or Rh catalysts.[5,6]
Curiously, despite the clear synthetic potential of these
annulations, progress in the development of enantioselective
variants has been extremely scarce and essentially limited to
an isolated case of a Rh-catalyzed (3+2+1) cycloaddition,[7]
and to our recent report on Pd-catalyzed (3+2) and (3+4)
[*] E. Da Concepción, Prof. J. L. MascareÇas, Dr. F. López
Centro Singular de Investigación en Química Biolóxica e Materiais
Moleculares (CiQUS) and Departamento de Química Orgµnica,
Universidade de Santiago de Compostela
15782 Santiago de Compostela (Spain)
E-mail: joseluis.mascareÇas@usc.es
fernando.lopez@csic.es
Prof. I. Fernµndez
Departamento de Química Orgµnica I and Centro de Innovación en
Química Avanzada (ORFEO-CINQA), Facultad de Ciencias Quími-
cas, Universidad Complutense de Madrid
28040 Madrid (Spain)
Dr. F. López
Misión Biológica de Galicia
Consejo Superior de Investigaciones Científicas (CSIC)
36080 Pontevedra (Spain)
Supporting information and the ORCID identification number(s) for
the author(s) of this article can be found under:
https://doi.org/10.1002/anie.202015202.
Scheme 1. Enantioselective cycloadditions of ACPs.
Angew. Chem. Int. Ed. 2021, 60, 8182 –8188
8182
ꢀ 2021 Wiley-VCH GmbH

Angewandte
Research Articles
Chemie
intermediates and shed light into the origin of the enantio-
explore their performance with our alkynyl-tethered ACPs.
selectivity.
Cobalt catalysts are highly attractive from a sustainability
standpoint, and also because of their rich coordination and
redox properties, which differ from those of their group IX
congeners.[16]
Results and Discussion
Gratifyingly, after a thorough screening, we found that the
cobalt catalyst generated in situ from CoBr2, dppp, and
a mixture of Zn and ZnI2 promotes the cycloaddition of 1a (in
acetonitrile at 1008C), delivering 2a in a good 73% yield
(entry 7). Curiously, related ligands such as dppe or Xantphos
failed to promote the reaction (Supporting Information,
Table S2), whereas alternative cobalt(II) precursors such as
CoCl2, CoI2 and Co(acac)2, under otherwise identical con-
ditions, provided incomplete conversions and lower yields of
2a (34–70%; Supporting Information, Table S2).[13]
Likewise, in the absence of ZnI2, the efficiency of the
reaction also decreased (entry 8). On the other hand, from all
other reducing agents tested (Supporting Information, Ta-
ble S3), only indium worked satisfactorily and, interestingly, it
did not require the use of a Lewis acid additive to provide
a good yield of 2a (entry 9 vs. 10).
Encouraged by these promising results, we analyzed the
viability of a Co-catalyzed enantioselective variant. Asym-
metric cobalt catalysis has received increasing attention in
recent years;[17] however, most examples consist of addition
processes such as hydrofunctionalizations of unsaturated
bonds. Indeed, enantioselective Co-promoted annulations
are rather scarce, and none of the reported cases involve the
We started the study using alkynes as ACP cycloaddition
partners considering that their well-known metal coordina-
tion properties could enable a more efficient enantiodiscri-
minating step.[12] Thus, we chose as substrates alkynyl-
tethered ACPs of type 1, which had never been explored in
asymmetric reactions. We initially tested the reactivity of 1a
using our recently developed Pd-phosphoramidite chiral
catalysts.[8] Heating 1a in refluxing toluene for 3 h, in the
presence of the catalyst generated in situ from Pd2(dba)3 and
L1, afforded the desired adduct 2a in moderate yield, and
with a low 67:33 er (Table 1, entry 1). Related monodentate
ligands like L2, or different Pd0 sources, did not improve these
values, whereas bidentate chiral ligands led to poor conver-
sions (Supporting Information, Table S1).[13] On the other
hand, alternative Rh and Ni catalysts, which had previously
been shown effective in a variety of ACP cycloadditions,[6] led
to the recovery of 1a (entries 3–6), thus precluding the
exploration of chiral variants.
Considering the seminal work by Liebeskind and Jones,[14]
as well as more recent studies by Dong and Yoshikai,[15]
showing the potential of low-valent cobalt complexes for
activating strained carbocyclic scaffolds, we were curious to
À
activation of C C bonds of ACPs or related
strained ring systems.[18]
Table 1: Initial evaluation of suitable metal precursors for the development of an
asymmetric cycloaddition of 1a.[a]
After screening several bidentate chiral ligands
in combination with CoBr2 and Zn/ ZnI2, we were
glad to identify a number of ligands that allow the
cycloaddition of 1a to be performed with moderate
to high enantioselectivity. As can be deduced from
Figure 1 and the Supporting Information, Table S4,
most of them are C2-symmetric bisphosphines, with
the more common BINAP providing the best
results. Using this ligand, the product 2a was
obtained in 73% yield, and an exciting 95:5 er
(acetonitrile at 1108C, 3 h). The reaction could also
be carried out at lower temperatures but, curiously,
it provides slightly lower er (89:11 er at 908C;
Supporting Information„ Table S5).[13] Finally, the
use of 1,2-DCE, instead of acetonitrile, increased
the reaction yield up to 83% (96:4 er), while with
Indium reductants (In or In/InI3, instead of Zn/
ZnI2) the reaction was similarly efficient (Fig-
ure 1).[19]
With these optimized conditions in hand, we
evaluated the scope of the method (Table 2). The
presence of different alkyl groups at the terminal
position of the alkyne was well tolerated (such as
adducts 2a–2e). The reaction proceeds well either
with a bulky OTBS or with an unprotected hydroxyl
moiety at the propargylic position, so that 2c and 2d
were obtained in almost identical er and yields.
Likewise, the cycloaddition of precursor 1e, bearing
an O-allyl group, was also efficient and gave 2e in
Entry M (x mol%)
L (y mol%)
Additive
Solvent
toluene
t
T
Yield
[h] [8C] [%]
1
Pd2dba3 (6)
L1 (13)
L2 (13)
P(OPh)3 (12)
–
–
–
dppp (12)
dppp (12)
dppp (12)
dppp (12)
–
–
–
–
–
–
3
110 66[b]
2
Pd2dba3 (6)
toluene 16 110 12
3
4
5
6
Rh(COD)Cl]2 (10)
RhClCO(PPh3)2 (10)
RhCl(PPh3)3 (10)
Ni(COD)2 (10)
CoBr2 (10)
toluene 20 110
toluene 20 110
toluene 20 110
toluene 12 110
CH3CN
CH3CN
CH3CN
CH3CN
–
–
–
–
[c]
7
Zn/ZnI2
3
3
3
3
100 73
100 24
100 82
100 86
8
CoBr2 (10)
Zn[d]
9
CoBr2 (10)
In[d]
[e]
10
CoBr2 (10)
In/InI3
[a] Conditions: A solution of 1a, [M] (x mol%), L (y mol%) and additive, in the
corresponding solvent, was heated for 3–20 h at the indicated temperature.
[b] Enantiomeric ratio=67:33. [c] Carried out with Zn (50 mol%) and ZnI2
(20 mol%). [d] Carried out with 50 mol% of the additive. [e] Carried out with In
(50 mol%) and InI3 (20 mol%).
Angew. Chem. Int. Ed. 2021, 60, 8182 –8188
ꢀ 2021 Wiley-VCH GmbH
www.angewandte.org
8183

Angewandte
Research Articles
Chemie
73% yield and 96:4 er. Notably, despite low-valent cobalt
[20]
À
catalysts could insert into Csp H bonds of alkynes, the use
of a terminal alkyne as a cycloaddition partner is also
tolerated. Thus, precursor 1 f was efficiently transformed into
its corresponding cycloadduct, 2 f, in 72% yield, and with an
excellent 95:5 er. The reaction of precursor 1g, bearing a TMS
substituent, required the use of In/InI3, instead of Zn/ZnI2, to
provide a good yield of the corresponding adduct (2g, 73%
yield, 5 h at 1308C). Despite the bulkiness of the TMS moiety,
the er was not significantly eroded (91:9 er). Aromatic
substituents with electron-neutral, donating or withdrawing
groups (1h–1j), as well as heteroaromatic rings, such as an
ortho-pyridine or a 2-thiofuryl group, are also well tolerated.
Their corresponding products (2h–2l) were isolated with
good yields and er above 91:9, up to 97:3. Importantly, the
presence of geminal diesters at the connecting tether is not
essential; indeed, the cycloaddition also works with substrates
that bear oxygen- or nitrogen-based groups, such as 1m–1q.
The expected 5,5-fused bicyclic systems 2m–2q were ob-
tained in yields above 70%, and with very good enantiomeric
ratios (erꢀ 93:7). Interestingly, enynes like in 1m, which
might engage in other side processes,[18b] are also excellent
partners, and exclusively provided the desired cycloadduct
2m in good yield and 94:6 er. The absolute configuration of
the N-tosyl derivate 2q could be determined by X-ray
crystallographic analysis (Table 2).[13]
Figure 1. Ligands tested for the enantioselective cycloaddition of 1a.
Conditions: A solution of 1a (1 equiv), CoBr2 (10 mol%), ligand
(12 mol%), Zn (50 mol%) and ZnI2 (20 mol%) was heated in CH3CN
at 1108C for 3 h, unless otherwise noted. Yields determined by 1H-
NMR of the crude mixtures with an internal standard. [b] Carried out
in 1,2-DCE as solvent. [c] Identical result obtained when using In
(50 mol%) instead of Zn/ZnI2, in 1,2-DCE as solvent. [d] Carried out
with In (50 mol%), InI3 (20 mol%), in 1,2-DCE.
Table 2: Scope of the Co-catalyzed enantioselective (3+2) cycloadditio-
n.[a]
The robustness of the transformation was further demon-
strated with a precursor that holds a methyl at the internal
position of the ACP alkene (1r), which allowed to build
a bicycle bearing a fully substituted quaternary carbon center
at the ring fusion (2r, 94:6 er). Finally, the method could also
be extended to the preparation of 6,5-fused bicyclic systems,
just by elongating the connecting tether by one methylene
unit. Thus, the reaction of 1s afforded the corresponding
adduct (2s) in 76% yield and 95:5 er, whereas the reaction of
the N-tosyl derivative 1t, which was better when carried out
using In/InI3, instead of Zn/ZnI2, provided 2t in 85% yield,
and a 95:5 er.
To further prove the synthetic potential of our method, we
did a brief exploration on the elaboration possibilities of the
cycloadducts. Therefore, azabicycle adduct 2q was submitted
to classical hydrogenation conditions using Pd/C. The reac-
tion proceeded smoothly to provide the saturated bicycle 3q
as a single isomer in 62% yield (Scheme 2). The stereochem-
istry of this compound could be determined by X-ray
diffraction analysis, which confirmed that the hydrogenation
of both alkenes occurred through the same face. These
reducing conditions can also be applied to other cycloadducts
providing their respective saturated bicycles in good yields
(Scheme 2). Overall, several enantioenriched 5,5- and 6,5-
fused bicyclic skeletons, bearing up to four stereocenters,
could be easily obtained in just two steps from simple acyclic
precursors.
According to previous work on low-valent cobalt com-
plexes generated in situ from CoII precursors,[16] the active
catalyst might consist of a cationic bisphosphine-CoI species
of type [(P-P)Co(L)2]+ (L = solvent or substrate molecule, P-
P = bisphosphine ligand). This species would be the result of
a Zn- (or In-) promoted reduction of the initially formed CoII
[a] Conditions: A solution of 1, CoBr2 (10%), (S)-BINAP (12%), Zn
(50%) and ZnI2 (20%) in 1,2-DCE were heated at 1108C for 3 h, unless
otherwise noted. Yields of isolated products after chromatography; er
values determined by HPLC. E’=CO2Et; E=CO2Me.[13] [b] Carried out
with In/ InI3, instead of Zn/ZnI2. [c] Reaction heated at 1308C for 5 h.
8184
www.angewandte.org
ꢀ 2021 Wiley-VCH GmbH
Angew. Chem. Int. Ed. 2021, 60, 8182 –8188

Angewandte
Research Articles
Chemie
sluggishly, requiring 18 h to achieve full conversion, and
providing a mixture of the (3+2) adduct 2a (24% yield), and
the isomeric product 2a’ (40% yield). The enantiomeric ratio
of 2a turned out to be 73:27, significantly lower than that
obtained with the cobalt catalyst. Therefore, the cobalt
reagents clearly outperform the homologous Rh counterparts,
both in terms of reaction efficiency and enantioselectivity.
To gain more insight into the mechanism of this trans-
formation, we carried out Density Functional Theory (DFT)
calculations at the dispersion-corrected PCM(DCE)-B3LYP-
D3/def2-SVP level.[13] We used the alkynylidenecyclopropane
precursor 1a’ and the cationic CoI complex [(dppp)Co]+ as
model reactant and catalyst, respectively (Figure 2). Initially,
we explored the viability of the insertion of the cobalt(I)
catalyst into the distal bond of the ACP, in analogy to the
routes previously identified for related Pd- or Rh-catalyzed
processes.[24] We first found out a pathway involving a distal
coordination and subsequent insertion of cationic cobalt
complex, followed by alkyne-coordination leading to the five-
coordinate 18-electron CoIII trimethylenemethane (TMM)
complex INT1 (via TS1). This intermediate is easily converted
Scheme 2. Hydrogenation of the cycloadducts, [X=NTs, O or C-
(CO2Me)2, n=1 or 2].
precursor [(P-P)CoBr2], followed by halide abstraction by
ZnI2 (or InI3). Nonetheless, the participation of a Co0
precursor such as [(P-P)Co0L2][21] cannot be ruled out, since
the cycloaddition of 1a also takes place to some extent just by
adding Zn or In, without ZnI2 or InI3 salts (see Table 1,
entries 8 and 9).[22] To shed some light onto the nature of the
active catalyst, we screened different low-valent cobalt
precursors, and other reducing agents, in the cycloaddition
of 1a (Scheme 3). Notably, the use of Et2Zn as reductant,
which has been previously shown to deliver [(P-P Co0] species
from their respective CoII precursors [(P-P)CoX2],[22] led to
the recovery of the starting material. Likewise, treatment of
1a with Co0 complexes such as Co(PMe3)4 or Co(PMe3)4/
BINAP, neither provided 2a nor any other identifiable
product. However, with the CoI precursor Co(PPh3)3Cl, in
the presence of BINAP and ZnI2, we isolated a 52% yield of
2a (92:8 er). Using identical conditions, but without ZnI2, the
reaction did not proceed at all. Overall, these results strongly
suggest the participation of a cationic CoI catalyst of the type
[(P-P)Co(L)2]+.[23]
À
into the p-allyl complex INT2, from which a C C reductive
elimination occurs with a very accessible barrier (12.7 kcal
molÀ1, via TS3) to deliver the observed bicyclic product
coordinated to the cobalt(I) catalyst. Finally, metal decoordi-
nation releases the observed cycloadduct and regenerates the
active cobalt(I) catalyst (Figure 2).[13]
Notably, the oxidative addition step of this pathway, which
directly yields the TMM metal complex, differs from that
proposed for related processes involving RhI or Pd0 catalysts,
which were found to generate metallacyclobutane intermedi-
ates.[24]
More remarkably, further computational scrutiny led us to
discover two energetically more accessible pathways towards
the intermediate INT2. In particular, we computationally
located the intermediate INT0’, a highly stable CoI-complex
in which both the alkene and the alkyne units of 1a’ are
coordinated to [Co(dppp)]+. Our calculations indicate that
this species can be transformed into the exo-methylene
cobaltacyclobutane intermediate INT1’, via TS1’, with an
energy barrier of ca. 31 kcalmolÀ1 in a highly endergonic
process (DGR = 17.2 kcalmolÀ1). Should this intermediate
INT1’ be formed, it would readily evolve to the significantly
more stable CoIII-TMM complex INT1 through a barrierless
ligand slippage process.[13]
Since this type of square-planar d8 reagents is well
established with RhI, we analyzed the performance of
[(BINAP)Rh(cod)]+, homologous to the above cobalt com-
plex, in the cycloaddition of 1a. The reaction took place
More importantly, besides this pathway, we additionally
found that INT0’ can alternatively undergo an oxidative
cyclometallation to a cobaltacycle intermediate, INT1’’,
through TS1’’ with a significantly more accessible energy
barrier of 14.1 kcalmolÀ1. Interestingly, this spiro-metallacy-
clic intermediate (INT1’’) can be transformed into the above-
mentioned intermediate INT2, via a cyclopropyl to p-allyl
rearrangement. This step is not a canonical b-carboelimina-
tion, but instead consists of a distal opening of the cyclopropyl
ring (C···C bond distance of 2.10 Š in TS2’’), a process
reminiscent to that previously proposed for cyclopropyl RhIII
complexes and other transition metal derivatives.[25]
Although the pathways proceeding through TS1 and TS1’
cannot be discarded, the high stability of INT0’, and
Scheme 3. Results using different cobalt sources, and reaction with
a cationic RhI complex.
Angew. Chem. Int. Ed. 2021, 60, 8182 –8188
ꢀ 2021 Wiley-VCH GmbH
www.angewandte.org
8185

Angewandte
Research Articles
Chemie
Figure 2. Computed reaction profile for the reaction model ACP 1a’ and [Co(dppp)]+. Relative free energies (DG, at 298 K) are given in kcalmolÀ1
.
All data have been computed at the PCM(DCE)-B3LYP-D3/def2-SVP level.
especially, the low activation barriers associated with the
subsequent steps, clearly favor the mechanism involving the
cobaltacyclopentene INT1’’.
intermediate of the catalytic cycle, upon heating at 608C, the
complex smoothly delivers the bicyclic product 2p
(Scheme 4).[13]
Interestingly, the computational study showed that the
cyclopropyl to p-allyl rearrangement is also viable using an
homologous RhI complex (difference in activation barriers of
approx. 1 kcalmolÀ1). However, the final reductive elimina-
tion step was much higher in the case of the RhI counterpart
(almost 9 kcalmolÀ1 more costly). This is consistent with the
observed lower reactivity of RhI catalysts (see above,
Scheme 3), and might be partially ascribed to the stronger
In consonance with the isolation of this diamagnetic
intermediate, computational analysis of profiles involving
cobalt triplet states (S = 1) revealed kinetically less favored
pathways. In addition, the stationary points along the singlet
hypersurface are more stable than their corresponding triplet
analogs, and very much in particular in the case of INT2.[13]
However, some eventual single-triplet crossovers along the
pathway cannot be fully discarded.
À
À
and less polarized nature of C Rh compared to C Co
bonds.[26] Indeed, analysis of the Wiberg bond indexes (WBIs)
on the corresponding intermediates (INT2 and its Rh analog
Finally, to shed some light into the origin of the
enantioinduction, we computationally analyzed the migratory
insertion process leading to species INT1’’, via TS1’’, as this is
the enantiodiscriminating step, using 1a’ and the chiral cobalt
complex [(S)-BINAP-Co]+. Our preliminary calculations
indicate that the transition state TS1’’(S,S), eventually leading
À
INT2-Rh), confirms that the two C M bonds involved in the
reductive elimination are significantly stronger for Rh (WBI
of 0.80 and 0.39) than for cobalt (WBI of 0.65 and 0.32).[13]
With this information at hand, we pursued the experi-
mental identification of some of the computationally located
cobalt complex intermediates. Therefore, we mixed equimo-
lar amounts of CoBr2, dppp, Zn, ZnI2 and 1p, at rt. After 2 h,
we observed the formation of a novel cobalt-containing
species, as deduced from the major molecular ion peak
observed by ESI-MS at m/z 746.2. Gratifyingly, this species
turned out to be rather stable, and could be analyzed by 1D-
and 2D-NMR spectroscopy, confirming its identity as the p-
allyl cobaltacyclic species 1p-INT2, analogous to the theo-
retically calculated intermediate INT2.[27] As expected for an
Scheme 4. Stoichiometric experiment: Isolation of p-allyl intermediate
1p-INT2.
8186
www.angewandte.org
ꢀ 2021 Wiley-VCH GmbH
Angew. Chem. Int. Ed. 2021, 60, 8182 –8188

Angewandte
Research Articles
Chemie
to the observed (S)-product, lies 1.6 kcalmolÀ1 below its
corresponding diastereomeric counterpart TS1’’(S,R), which
is qualitatively consistent with the high enantioselectivity of
the process. This energy difference may be in part ascribed to
Acknowledgements
This work received financial support from the Spanish
MINECO (SAF2016-76689-R, CTQ2016-78205-P,
À
the ocurrence of stabilizing noncovalent C H···p interactions
CTQ2017-84767-P, PID2019-106184GB-I00 and PID2019-
108624RB-I00), the Xunta de Galicia (ED431C 2017/19,
2015-CP082, Centro Singular de Investigación de Galicia
accreditation 2019–2022, ED431G 2019/03, and a predoctoral
Fellowship to E. da C.) and the ERDF, ERC (Adv. Grant No.
340055). The Orfeo-Cinqa network (CTQ2016-81797-REDC)
and Dr. Juan Durµn (preliminary work) are also acknowl-
edged. We sincerely thank the Reviewers of this manuscript
for their very constructive and fruitful comments.
in TS1’’(S,S), established between one of the phenyl groups of
À
BINAP and a C H bond adjacent to the reactive carbon atom
(see green surfaces in Figure 3), interactions that are absent in
TS1’’(S,R). Moreover, analysis of TS2, which is the enantio-
determining step in the alternative TMM mechanistic path-
way, also indicates a strong preference for the S,S-diastereo-
isomeric transition state [TS2(S,S)], which also features
À
a similar stabilizing C H···p interaction.
Conflict of interest
The authors declare no conflict of interest.
Keywords: alkylidenecyclopropane ·cobalt ·cycloaddition ·
earth-abundant metals ·enantioselectivity
[1] a) M. Lautens, W. Klute, W. Tam, Chem. Rev. 1996, 96, 49 –92;
b) M. Gulías, F. López, J. L. MascareÇas, Pure Appl. Chem. 2011,
83, 495 –506; c) M. Gulías, J. L. MascareÇas, Angew. Chem. Int.
Ed. 2016, 55, 11000 –11019; Angew. Chem. 2016, 128, 11164 –
11184.
[2] For selected reviews covering this topic, see: a) P. Le Marquand,
W. Tam, Angew. Chem. Int. Ed. 2008, 47, 2926 –2928; Angew.
Chem. 2008, 120, 2968 –2970; b) P. A. Inglesby, P. A. Evans,
Chem. Soc. Rev. 2010, 39, 2791 –2805; c) A. Marinetti, H. Jullien,
A. Voituriez, Chem. Soc. Rev. 2012, 41, 4884 –4908; d) F. López,
J. L. MascareÇas, Beilstein J. Org. Chem. 2013, 9, 2250 –2264.
[3] Most enantioselective TMC annulations have been focused on
intermolecular processes, while synthetically powerful intra-
molecular cases remained far less explored. For selected cases,
see: a) L. Souillart, E. Parker, N. Cramer, Angew. Chem. Int. Ed.
2014, 53, 3001 –3005; Angew. Chem. 2014, 126, 3045 –3049; b) L.
Liu, N. Ishida, M. Murakami, Angew. Chem. Int. Ed. 2012, 51,
2485 –2488; Angew. Chem. 2012, 124, 2535 –2538; c) P. A.
Wender, L. O. Haustedt, J. Lim, J. A. Love, T. J. Williams, J. Y.
Yoon, J. Am. Chem. Soc. 2006, 128, 6302 –6303; d) T. Xu, H. M.
Ko, N. A. Savage, G. Dong, J. Am. Chem. Soc. 2012, 134, 20005 –
20008.
[4] a) P. H. Chen, B. A. Billett, T. Tsukamoto, G. Dong, ACS Catal.
2017, 7, 1340 –1360; b) C. Aïssa, Comprehensive Organic Syn-
thesis II, Elsevier, Amsterdam, 2014, pp. 1738 –1771; c) G.
Fumagalli, S. Stanton, J. F. Bower, Chem. Rev. 2017, 117,
9404 –9432; d) L. Souillart, N. Cramer, Chem. Rev. 2015, 115,
9410 –9464; e) J. Liu, R. Liu, Y. Wei, M. Shi, Trends Chem. 2019,
1, 779 –793.
[5] For reviews, see: a) A. Brandi, S. Cicchi, F. M. Cordero, A. Goti,
Chem. Rev. 2003, 103, 1213 –1269; b) H. Pellissier, Tetrahedron
2014, 70, 4991 –5031; c) L.-Z. Yu, K. Chen, Z.-Z. Zhu, M. Shi,
Chem. Commun. 2017, 53, 5935 –5945; d) P. Binger, H. M. Buch,
Top. Curr. Chem. 1987, 135, 77 –151.
[6] For selected examples, see: a) A. Delgado, J. R. Rodríguez, L.
Castedo, J. L. MascareÇas, J. Am. Chem. Soc. 2003, 125, 9282 –
9283; b) F. López, A. Delgado, J. Rodríguez, L. Castedo, J.
MascareÇas, J. Am. Chem. Soc. 2004, 126, 10262 –10263; c) M.
Gulías, R. García, A. Delgado, L. Castedo, J. L. MascareÇas, J.
Am. Chem. Soc. 2006, 128, 384 –385; d) M. Gulías, J. Durµn, F.
López, L. Castedo, J. L. MascareÇas, J. Am. Chem. Soc. 2007,
Figure 3. Contour plots of the reduced density gradient isosurfaces
(density cutoff of 0.03 a.u.) for TS1’’(S,S). The green surfaces indicate
attractive non-covalent interactions.
Conclusion
We have developed the first enantioselective cycloaddi-
tion of alkylidenecyclopropanes (ACPs) promoted by any
type of earth-abundant metal complex, and the first (3+2)
annulation of ACPs promoted by Co catalysts. Moreover, our
results also represent the first asymmetric annulations of
ACPs with alkynes. The reactions, promoted by a chiral low-
valent CoI catalyst generated in situ from an air-stable CoII
precursor, provide 5,5- and 6,5- fused bicyclic products in
good yields and with excellent enantioselectivities. Mecha-
nistic studies, supported by DFT calculations, allowed us to
propose a pathway involving a bisphosphine cobalt(I) cationic
complex as the active catalyst, and a key cobaltacycle
intermediate (INT1’’) that evolves via cyclopropane opening
to a highly stable p-allyl cobaltacycle, which could even be
isolated and characterized by spectroscopic methods. In
addition to the novelty and synthetic significance of this
method, our results pave the way for the development of
novel enantioselective multicomponent cycloadditions based
on earth-abundant, non-expensive cobalt catalysts.
Angew. Chem. Int. Ed. 2021, 60, 8182 –8188
ꢀ 2021 Wiley-VCH GmbH
www.angewandte.org
8187

Angewandte
Research Articles
Chemie
129, 11026 –11027; e) L. Saya, G. Bhargava, M. A. Navarro, M.
Gulías, F. López, I. Fernµndez, L. Castedo, J. L. MascareÇas,
Angew. Chem. Int. Ed. 2010, 49, 9886 –9890; Angew. Chem.
2010, 122, 10082 –10086; f) G. Bhargava, B. Trillo, M. Araya, F.
Lopez, L. Castedo, J. L. Mascarenas, Chem. Commun. 2010, 46,
270 –272; g) M. Araya, M. Gulías, I. Fernµndez, G. Bhargava, L.
Castedo, J. L. MascareÇas, F. López, Chem. Eur. J. 2014, 20,
10255 –10259; h) P. A. Evans, P. A. Inglesby, J. Am. Chem. Soc.
2008, 130, 12838 –12839; i) P. A. Evans, D. E. Negru, D. Shang,
Angew. Chem. Int. Ed. 2015, 54, 4768 –4772; Angew. Chem.
2015, 127, 4850 –4854; j) S. Saito, K. Maeda, R. Yamasaki, T.
Kitamura, M. Nakagawa, K. Kato, I. Azumaya, H. Masu, Angew.
Chem. Int. Ed. 2010, 49, 1830 –1833; Angew. Chem. 2010, 122,
1874 –1877; k) S. Saito, M. Masuda, S. Komagawa, J. Am. Chem.
Soc. 2004, 126, 10540 –10541; l) D. H. Zhang, M. Shi, Tetrahe-
dron Lett. 2012, 53, 487 –490; m) S. Kim, Y. K. Chung, Org. Lett.
2014, 16, 4352 –4355; n) T. Yoshida, Y. Tajima, M. Kobayashi, K.
Masutomi, K. Noguchi, K. Tanaka, Angew. Chem. Int. Ed. 2015,
54, 8241 –8244; Angew. Chem. 2015, 127, 8359 –8362; o) L. G.
Zhao, A. de Meijere, Adv. Synth. Catal. 2006, 348, 2484 –2492;
p) B. Trillo, M. Gulias, F. López, L. Castedo, J. L. MascareÇas,
Adv. Synth. Catal. 2006, 348, 2381 –2384; q) F. Verdugo, E.
da Concepción, R. RodiÇo, M. Calvelo, J. L. MascareÇas, F.
López, ACS Catal. 2020, 10, 7710 –7718.
2018, 130, 6668 –6672; c) J. Yang, Q. Sun, N. Yoshikai, ACS
Catal. 2019, 9, 1973 –1978; d) J. Yang, Y. Mori, M. Yamanaka, N.
Yoshikai, Chem. Eur. J. 2020, 26, 8302 –8307.
[16] “Cobalt Catalysis in Organic Synthesis”: Methods and Reactions
(Eds.: M. Hapke, G. Hilt), Wiley-VCH, Weinheim, 2020.
[17] For reviews, see: a) H. Pellissier, H. Clavier, Chem. Rev. 2014,
114, 2775 –2823; b) N. Yoshikai, Synthesis 2019, 51, 135 –145.
[18] > Co-catalyzed enantioselective formal cycloadditions are es-
sentially limited to cyclopropanations, alkyne (2+2+2) cyclo-
additions and intermolecular (4+2) and (2+2) cycloadditions.
See Ref. [17] and a) M. M. Parsutkar, V. V. Pagar, T. V. Rajan-
Babu, J. Am. Chem. Soc. 2019, 141, 15367 –15377; b) V. V. Pagar,
T. V. RajanBabu, Science 2018, 361, 68 –72; For an enantiose-
lective example with a CoIII catalyst, see: c) K. Ozols, Y. S. Jang,
N. Cramer, J. Am. Chem. Soc. 2019, 141, 5675 –5680.
[19] Other cobalt sources such as CoI2, CoCl2 or Co(acac)2 proved to
be less efficient than CoBr2.
[20] See, for instance: T. Sakurada, Y.-K. Sugiyama, S. Okamoto, J.
Org. Chem. 2013, 78, 3583 –3591.
[21] a) H. Zhong, M. R. Friedfeld, P. J. Chirik, Angew. Chem. Int. Ed.
2019, 58, 9194 –9198; Angew. Chem. 2019, 131, 9292 –9296;
b) M. R. Friedfeld, H. Zhong, R. T. Ruck, M. Shevlin, P. J.
Chirik, Science 2018, 360, 888 –893.
[22] See, for instance: D. K. Kim, J. Riedel, R. S. Kim, V. M. Dong, J.
Am. Chem. Soc. 2017, 139, 10208 –10211.
[23] a) The formation of 2a in the absence of ZnI2 (Table 1, entry 8)
or InI3 (entry 9) can be rationalized considering the in situ
generation of ZnBr2 (or InBr3) from the reduction of [(P-
P)CoBr2] by Zn (or In); b) Zinc (II) halides might be key for the
stability of the CoI species, avoiding undesired disproportiona-
tion processes, see: S. Seka, O. Buriez, J. Perichon, Chem. Eur. J.
2003, 9, 3597 –3603.
[24] For mechanistic studies on TMC-ACP cycloadditions, see: a) H.
Yu, Q. Lu, Z. Dang, Y. Fu, Chem. Asian J. 2013, 8, 2262 –2273;
b) R. García-FandiÇo, M. Gulías, J. L. MascareÇas, D. J. Cµrde-
nas, Dalton Trans. 2012, 41, 9468 –9481; c) R. Yamasaki, M.
Ohashi, K. Maeda, T. Kitamura, M. Nakagawa, K. Kato, T.
Fujita, R. Kamura, K. Kinoshita, H. Masu, I. Azumaya, S.
Ogoshi, S. Saito, Chem. Eur. J. 2013, 19, 3415 –3425; d) T.
Suzuki, H. Fujimoto, Inorg. Chem. 2000, 39, 1113 –1119. See also
references [7], [6e] and [6g].
[25] a) R. A. Periana, R. G. Bergman, J. Am. Chem. Soc. 1986, 108,
7346 –7355; b) T. Omrcen, N. J. Conti, W. M. Jones, Organo-
metallics 1991, 10, 913 –917; c) N. J. Conti, D. J. Crowther, S.
Tivakornpannarai, W. M. Jones, Organometallics 1990, 9, 175 –
184; d) K. D. J. Parker, S. Labat, J.-M. Sotiropoulos, K. Miqueu,
V. Pimienta, L. Vendier, M. Etienne, Eur. J. Org. Chem. 2019,
4555 –4563.
[7] S. Mazumder, D. Shang, D. E. Negru, M.-H. Baik, P. A. Evans, J.
Am. Chem. Soc. 2012, 134, 20569 –20572.
[8] a) F. Verdugo, L. Villarino, J. Durµn, M. Gulías, J. L. MascareÇas,
F. López, ACS Catal. 2018, 8, 6100 –6105; For a (3+2) enantio-
selective annulation between structurally related vinylidenecy-
clopropanes and alkenes, see: b) K.-H. Rui, S. Yang, Y. Wei, M.
Shi, Org. Chem. Front. 2019, 6, 2506 –2513.
[9] There is only one previous report on formal cycloadditions of
ACPs with cobalt, dealing with the insertion of CO into their
proximal bond; however stoichiometric amounts of the metal
are usually required to achieve good yields, see: a) T. Kurahashi,
A. de Meijere, Angew. Chem. Int. Ed. 2005, 44, 7881 –7884;
Angew. Chem. 2005, 117, 8093 –8096; For a cycloisomerization
À
of ACPs based on a C H activation promoted by a high-valent
Co catalyst, see: b) M. Li, F. Y. Kwong, Angew. Chem. Int. Ed.
2018, 57, 6512 –6516; Angew. Chem. 2018, 130, 6622 –6626; For
an annulation that does not involve cyclopropane cleavage, see:
c) A. Dey, N. Thrimurtulu, C. M. R. Volla, Org. Lett. 2019, 21,
3871 –3875.
[10] For one report on an unrelated, enantioselective Co-catalyzed
(3+2) annulation, see: C.-H. Wei, S. Mannathan, C.-H. Cheng,
Angew. Chem. Int. Ed. 2012, 51, 10592 –10595; Angew. Chem.
2012, 124, 10744 –10747.
[11] Very few TMC enantioselective cycloadditions have been
reported to provide 5-membered fused bicarbocyclic systems,
see: a) M. Lin, G.-Y. Kang, Y.-A. Guo, Z.-X. Yu, J. Am. Chem.
Soc. 2012, 134, 398 –405; b) N. Iwasawa, Comprehensive Organic
Synthesis II, Elsevier, Amsterdam, 2014, pp. 273 –350.
[12] Overall, alkynes coordinate more tightly than alkenes to
transition metals, see: a) R. H. Crabtree, in Th e Organometallic
Chemistry of the Transition Metals, Wiley, Hoboken, 2014,
pp. 134 –162; b) J. P. Collman, L. S. Hegedus, J. R. Norton, R. G.
Finke, Principles and Applications of Organotransition Metal
Chemistry, University Science Books, Sausalito, 1987.
[13] See the Supporting Information for further details.
[14] a) L. S. Liebeskind, J. P. Leeds, S. L. Baysdon, S. Iyer, J. Am.
Chem. Soc. 1984, 106, 6451 –6453; b) M. A. Huffman, L. S.
Liebeskind, J. Am. Chem. Soc. 1990, 112, 8617 –8618; c) C.
Perthuisot, B. L. Edelbach, D. L. Zubris, W. D. Jones, Organo-
metallics 1997, 16, 2016 –2023.
À
[26] The bond dissociation energy of M C bonds increases down the
group. See: a) C. T. To, K. S. Chan, Eur. J. Org. Chem. 2019,
6581 –6591; b) J. A. M. Simoes, J. L. Beauchamp, Chem. Rev.
1990, 90, 629 –688; c) P. E. M. Siegbahn, J. Phys. Chem. 1995, 99,
12723 –12729.
[27] a) The species 1p-INT2 was isolated when carrying out the
experiment in acetonitrile, suggesting that this solvent might
play a role in further stabilizing this type of intermediate;
b) Alkyl, instead of alkenyl, p-allyl rhodacycles had been
previously isolated in the context of (3+2+2) and (3+2+1)
cycloadditions, see: P. A. Inglesby, J. Bacsa, D. E. Negru, P. A.
Evans, Angew. Chem. Int. Ed. 2014, 53, 3952 –3956; Angew.
Chem. 2014, 126, 4033 –4037.
Manuscript received: November 13, 2020
Revised manuscript received: December 20, 2020
Accepted manuscript online: January 19, 2021
Version of record online: March 5, 2021
[15] a) Z. Zhu, X. Li, S. Chen, P. H. Chen, B. A. Billett, Z. Huang, G.
Dong, ACS Catal. 2018, 8, 845 –849; b) C. Wu, N. Yoshikai,
Angew. Chem. Int. Ed. 2018, 57, 6558 –6562; Angew. Chem.
8188
www.angewandte.org
ꢀ 2021 Wiley-VCH GmbH
Angew. Chem. Int. Ed. 2021, 60, 8182 –8188