Stereoselective gold(I) domino catalysis of allylic isomerization and olefin cyclopropanation: Mechanistic studies

Article abstract of DOI:10.1021/acs.joc.5b00666

Gold(I) catalysis of olefin activation is still a rare feature in the repertoire of that metal. Mechanistic studies on the gold(I)-catalyzed cyclopropanation of allyl-substituted sulfonium ylides, including kinetic analysis as well as detailed computational studies, reveal that the reaction proceeds through activation of the alkene moiety. Furthermore, a novel competitive allylic isomerization pathway that interconverts "linear" and "branched" allylic isomers is uncovered. The subtle interplay of cyclopropanation and olefin isomerization results in an intriguing domino process where two independent catalytic transformations combine with near-perfect regio- and stereoselectivities.

Full text of DOI:10.1021/acs.joc.5b00666

Article
pubs.acs.org/joc
Stereoselective Gold(I) Domino Catalysis of Allylic Isomerization and
Olefin Cyclopropanation: Mechanistic Studies
Sebastian Klimczyk,† Xueliang Huang,‡ Hanspeter Kahlig,† Luís F. Veiros,*,§ and Nuno Maulide*,†
̈
†Faculty of Chemistry, Institute of Organic Chemistry, University of Vienna, Wahringer Strasse 38, 1090 Vienna, Austria
̈
‡State Key Laboratory of Structural Chemistry, Fujian Institute of Research on the Structure of Matter, Chinese Academy of Sciences,
155 Yangqiao Road West, Fujian 350002, China
§Centro de Química Estrutural, Instituto Superior Tec
́
nico, Universidade de Lisboa, Av. Rovisco Pais 1, 1049-001 Lisbon, Portugal
S
* Supporting Information
ABSTRACT: Gold(I) catalysis of olefin activation is still a rare
feature in the repertoire of that metal. Mechanistic studies on
the gold(I)-catalyzed cyclopropanation of allyl-substituted
sulfonium ylides, including kinetic analysis as well as detailed
computational studies, reveal that the reaction proceeds
through activation of the alkene moiety. Furthermore, a
novel competitive allylic isomerization pathway that inter-
converts “linear” and “branched” allylic isomers is uncovered. The subtle interplay of cyclopropanation and olefin isomerization
results in an intriguing domino process where two independent catalytic transformations combine with near-perfect regio- and
stereoselectivities.
independent, facile gold(I)-catalyzed allylic isomerization that
proceeds under similar conditions and ultimately results in a
domino process of high efficiency and selectivity.
INTRODUCTION
■
The development and application of gold(I) catalysis in organic
synthesis has witnessed an explosive growth over the last
decades. The unique, almost unrivalled ability of various
gold(I) species, ranging from simple salts to ornate organo-
metallic complexes, to selective activate C−C multiple bonds
under mild conditions is largely responsible for this.1 A very
large body of literature exists on the gold(I)-catalyzed
activation of alkynes and allenes; comparatively, however, the
activation of simple olefins for direct nucleophilic addition is by
far less common.2 Even in these rare instances, C−O and C−N
bond-forming events appear to dominate the landscape;3
gold(I) olefin activation leading to C−C nucleophilic bond
formation is a rare class of transformations.4
In the course of our studies on gold(I)-catalyzed trans-
formations of sulfur ylides, we discovered a catalytic cyclo-
propanation reaction (Scheme 1).5,4f Unusually, direct and
indirect evidence suggested that this reaction proceeds by olefin
activation rather than by metallocarbene formation. Herein, we
present a thorough mechanistic study of the gold(I) stereo-
convergent cyclopropanation of sulfonium ylides yielding
reactivity insights that should prove useful for gold(I) catalysis
of alkene activation in general. Additionally, we uncover an
RESULTS AND DISCUSSION
■
We began our studies by carrying out kinetic measurements on
the model substrate system 5 in the presence of Echavarren’s
catalyst 6.6 As shown in Figure 1, heating these two species at
80 °C in toluene delivered the cyclopropane 7 in a very clean
reaction. This process reached 95% conversion after 6.5 h
(Figure 1).
1
Reaction monitoring was also possible employing H NMR
spectroscopy (Figure 2), with a clear disappearance of the
characteristic olefin resonances at δ = 4.6−6.0 ppm.
Furthermore, the total concentration of 5 and 7 during the
reaction was consistent with the concentration of 5 at t = 0,
indicating that virtually no byproducts were formed. A
logarithmic plot of [5] against time was characteristic of first-
order behavior (Figure 3).
Encouraged by these observations, we performed further
measurements at this temperature, obtaining consistent values
for the rate constant (kav(80 °C) = 8.27 × 10−3). Carrying out
kinetic measurements at different temperatures allowed the
determination of the corresponding rate constants (Figure 4).
This in turn enabled an estimation of the activation energy for
this reaction of Eact = 21.11 kcal/mol.
Scheme 1. Serendipitously Discovered Gold(I)-Catalyzed
Cyclopropanation Reaction
Interested in the role of the catalyst in this reaction, we
measured the rate constants of the reaction at various catalyst
loadings. A plotting of the observed rate constants against the
Received: March 24, 2015
© XXXX American Chemical Society
A
DOI: 10.1021/acs.joc.5b00666
J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
Figure 3. Representative plot of ln([5]t/[5]0) against the reaction
time (min) for the conversion of 5 showing first-order dependence in
5.
Figure 1. Gold(I)-catalyzed cyclopropanation of sulfonium ylide 5
1
■
▲
●
where = [5]; = [7], and = [5] + [7], monitored by H NMR
spectroscopy.
catalyst loading (see the Supporting Information) afforded a
straight line, documenting first-order behavior for the catalyst.
Finally, we followed the reaction by 31P NMR spectroscopy in
order to assess the resting state of the catalyst during the
process (Figure 5).
Figure 4. Plot of ln(k) against 1/T (K).
As shown, it would appear that the catalyst readily exchanges
coordination to either the in situ liberated diphenyl sulfide
(spectrum 2) or to the olefin (spectrum 3).7
The mechanism of cyclopropanation was further studied by
DFT calculations.8 The results obtained for the unsubstituted
substrate 5 were the subject of a preliminary publication,4f,9 and
the corresponding profile is presented in Figure 6. The reaction
pathway starts with a complex (A) where the substrate
coordinates the metal fragment (M = [Au(PPh3)3]+) by
means of the CC double bond in a η2 mode. A rotation of
the carbon backbone (A → A′) brings the ylide carbon and the
inner carbon of the olefin into ideal proximity for the formation
of the corresponding C−C bond. This bond is formed in the
following step (A′ → B), going through an early transition
state, TSA′B, as shown by a very long distance (3.04 Å) and a
negligible Wiberg index,10 WI = 0.01.
Figure 5. 31P kinetics of the gold(I)-catalyzed cyclopropanation,
spectra 4−18. Free catalyst 6 (spectrum 1), 6 + diphenyl sulfide
(spectrum 2), and 6 + allyl acetate (spectrum 3).
Figure 2. Enlargement of the allylic and aliphatic regions of the 400 MHz 1H NMR spectra of the conversion of 5 to 7. The aromatic region and the
area between 2.2 and 3.3 ppm were omitted for clarity.
B
DOI: 10.1021/acs.joc.5b00666
J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
Figure 6. Free energy profile (PBE0) for cyclopropanation of the unsubstituted substrate. The minima and the transition states were optimized, and
the structures obtained are presented with the more relevant bond distances in each step. Free energy values (kcal/mol) are referred to the initial
intermediate (A).
Intermediate B corresponds to a 5-membered lactone
carrying a gold−alkyl complex appended to the lactone β-
carbon. The mechanism proceeds with a further conformational
change (B → B′), by which the original ylidic carbon atom and
the carbon bound to the metal center are brought to the right
conformation for the establishment of the final C−C bond and,
thus, cyclopropanation and loss of SPh2 in the final step.
Formation of the cyclopropane ring, from B′ to C′ occurs via
transition state TSB′C′ where cleavage of the C−S bond is well
advanced (d = 2.51 Å, WI = 0.34), whereas the new C−C bond
is only incipient (d = 2.23 Å, WI = 0.23). The final product
coordinates the metal moiety weakly in a σ-complex using the
newly formed C−C bond of the cyclopropane in C′ and C.11a
The mechanism depicted in Figure 6 entails two consecutive
C−C bond-forming events: the first leads to the lactone ring in
intermediates B and B′ while the second corresponds to the
formation of the cyclopropane ring in C′ and C, and both
present similar energy barriers (within 0.4 kcal/mol). The first
C−C bond results from a formal nucleophilic attack of the
ylidic carbon to the internal carbon of the activated CC
double bond. This is indicated by the respective atomic charges
(NPA, see the Computational Details in the Supporting
Information) calculated for intermediate A′, which clearly
show a more electron-rich ylide carbon with a charge of C =
−0.60, compared with the olefin carbon atom with C = −0.17.
This is in agreement with the known reactivity of sulfur ylides.
The nature of the second C−C bond formation event,
effectively resulting in cyclopropanation, is perhaps more
surprising. In fact, the atomic charges calculated for the
relevant C atoms in intermediate B′ present a more electron-
rich coordinated carbon, CAu (C = −0.93), when compared
with the carbon carrying the Ph2S moiety (C = −0.37). This
suggests a nucleophilic attack from the bound metal−alkyl in a
reactivity pattern similar to that of a typical organometallic C-
nucleophile.12 Frontier orbital analysis of intermediate B′
(Figure 7) further confirms this assumption.
The HOMO of B′ is a Au−C σ orbital with a significant
(23%) contribution on the coordinated carbon, while the
LUMO represents a C−S σ* interaction with 12% of its
electron density on the ylide C atom. This corroborates a
cyclopropanation step occurring with electron transfer from the
Figure 7. Frontier orbitals of intermediate B′.
C
DOI: 10.1021/acs.joc.5b00666
J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
coordinated C atom to the originally ylidic carbon, with
consequent breaking of the C−S bond and loss of the leaving
group. It is important to note that the reaction intermediate B/
B′ bears two stereogenic centers (see Figure 7). Therefore, four
different stereoisomers corresponding to two pairs of
enantiomers are possible. In one pair (B′), the carbon bound
to gold (CAu) and the SPh2 group are trans to each other; in the
opposite diastereomer, intermediate D, CAu and the SPh2 group
occupy the same face of the lactone ring (Figure 8).
Figure 9. Transition states for the cyclopropanation step of the
mechanism involving the two diastereomeric lactone intermediates, B′
(left) and D (right). Free energy values (kcal/mol) relative to A are
presented.
good to excellent yields and in excellent stereoselectivities
(typically >12:1).
Scheme 2. Gold-Catalyzed Cyclopropanation of Substituted
(Allyloxy)Sulfonium Ylides
Figure 8. Diastereoisomeric intermediates B′ and D.
Naturally, the geometrical constraints of the cyclopropane
ring imply that the final product will be the same regardless of
the stereochemistry of the intermediate. Nevertheless, different
energy profiles should result for each case, and in the case of
substrates with substituents (cf. Figure S14) the selectivity of
the reaction will depend on the stereochemical features of the
intermediate. Consequently, we also addressed the mechanism
of cyclopropanation via intermediate D, and the profile
obtained is presented as Supporting Information (cf. Figure
S12). The main characteristics of the mechanism remain the
same as in the path depicted in Figure 6 and discussed above.
Thus, there are two consecutive C−C bond-forming events.
Both energy barriers are higher than the ones calculated for the
previous mechanism, via B/B′, being 19.9 and 25.4 kcal/mol
for the two steps, respectively. Consequently, two main
differences between these pathways emerge: (1) the mechanism
proceeding via B/B′ is more favorable than the one that
includes intermediate D and (2) instead of two equally difficult
steps, as observed in the first path, the second step
(cyclopropanation) becomes clearly rate determining in the
second mechanism. A closer observation of the transition states
for the cyclopropanation step in both mechanisms explains
those differences (Figure 9). While for TSB′C′ formation of the
C−C bond and loss of the SPh2 leaving group occur in nearly
antiperiplanar fashion, in the case of TSDE′ those two processes
take place on the same face of the lactone ring, the latter
resulting in a significantly higher barrier.
Unexpectedly, we found that the use of terminally substituted
olefins such as 1h in this cyclopropanation reaction leads to
formation of 4d, the same product (with similarly high
diastereoselectivity) we had previously observed when employ-
ing branched allyl moieties. This was the case regardless of the
E/Z geometric purity of the starting materials (Scheme 3b).
Scheme 3. Apparent Allylic Rearrangement upon Attempted
Cyclopropanation of Terminally Substituted Olefins
Use of Substituted Allyl Moieties. If substituted allyl
moieties are employed, interesting observations can be made.
In case of substitution at the allylic position (Scheme 2), the
corresponding cyclopropanation products could be isolated in
D
DOI: 10.1021/acs.joc.5b00666
J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
With this finding in mind, we were interested in cyclizing the
enantiopure ylide (R)-1j (ee >99%), hoping to thus obtain the
corresponding enantiopure cyclopropane 4j. Surprisingly,
however, the resulting product, while still essentially a single
diastereomeric substance, was nearly devoid of optical activity
(Scheme 4a). As such a result must imply transient loss of
between branched and linear substrates, for which the profile
presented in Figure 10 was obtained.
The initial species in the reaction profile, F, is an η2 complex
where the branched substrate binds to the metal atom through
the CC double bond, similarly to what happens in complex A
for the unsubstituted substrate (Figure 6). From F, a rotation
around the COO−C bond brings the O atom of the ester
carbonyl group close to the terminal carbon of the olefin,
ultimately resulting in intermediate G. This corresponds to a
cyclic 6-membered acetal carrying a metal−alkyl fragment. In
the corresponding transition state, TSFG, the new C−O bond is
still far from established with a long distance (2.13 Å)
indicating a weak interaction (WI = 0.22), while the metal
center is clearly sliding toward the internal olefinic carbon, with
differentiated Au−C distances of 2.58 and 2.17 Å (compared
with 2.27 and 2.32 Å in F). The second step (G → H) is
equivalent to first one. Here, the other C−O bond of the acetal
breaks, regenerating the carbonyl group but now bearing a
linear allylic moiety. The overall process depicted in Figure 10
corresponds to a facile reaction with a barrier of 14.6 kcal/mol
associated with the first step.
Scheme 4. (a) Cyclopropanation of an Enantiopure
Sulfonium Ylide. (b) Isotope-Labeling Experiment
To complete the isomerization study, one has to consider the
interconversion between the two isomers involved in the
previous mechanism (branched and linear E) with the third
stereoisomer, i.e., the linear Z species. Since a direct
interconversion between the E and Z isomers of the linear
substrate would involve a prohibitive rotation around a CC
double bond, the actual change occurs via the branched isomer.
However, the shift of the ester group from one olefin C atom to
the other, through a cyclic acetal intermediate similar to G
(Figure 10), must take place with the appropriate conformation
of the carbon chain in order to yield the Z isomer of the linear
substrate.
tetrahedral character at C1′, we prepared the deuterated
substrate 1k and submitted it to the cyclopropanation reaction
conditions (Scheme 4b). In the resulting cyclopropane product,
12% of deuterium scrambling could be unambiguously detected
by NMR analysis of the crude mixture.13
This interesting result suggested that not only an allylic
rearrangement occurs during the reaction with gold(I) catalyst
6, but also that the interconversion between allylic isomers of
the substrate competes with the cyclopropanation reaction. In
order to gain further understanding of this phenomenon, the
mechanisms of interconversion between substrates were
studied by DFT and compared with those calculated for the
formation of each one of the four possible isomers of the
product. The first process addressed is the isomerization
In the acetal intermediate that leads to the linear E isomer
(G), the metal center and the methyl group have a trans
orientation along the six-membered ring (see Scheme 5a),
Figure 10. Free energy profile (PBE0) for interchange between the branched substrate and its E linear isomer. The minima and the transition states
were optimized, and the structures obtained are presented with the more relevant bond distances in each step. Free energy values (kcal/mol) are
referred to the initial intermediate (F).
E
DOI: 10.1021/acs.joc.5b00666
J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
practically cleaved (2.70 Å, WI = 0.04), while the Au atom is
much closer to the inner olefin C atom (2.17 Å) than to the
outer one (2.58 Å). In the final step of the mechanism, the
acetal ring is opened through the breaking of one O−C bond,
yielding the η2 complex of the Z linear isomer of the substrate
in P. This step is equivalent to the one that leads to the
coordinated E linear isomer (via TSGH), discussed above. The
highest barrier is associated with the ring-expansion step, going
through TSNO, and its value (16.3 kcal/mol, relative to F) is
only 1.7 kcal/mol higher than the one obtained for the
branched-linear E isomerization path (Figure 10). In summary,
the entire isomerization process interconverting the branched
and linear E and Z allylic isomers is represented by the
combination of the energy profiles in Figures 10 and 11 and
Figure S15 (Supporting Information) and, thus, requires the
overcoming of a maximum energy barrier of 16.3 kcal/mol
(corresponding to TSNO). This indicates a process that should
occur easily under the experimental conditions of the reaction.
Further experimental evidence for this facile isomerization
(Figure 12) was readily obtained by exposure of the simple but-
2-enyl methylmalonate 8 (not carrying a sulfur ylide moiety) to
the action of gold catalyst 6 in deuterated chloroform.14
Comparison of the NMR spectra of pure “branched” isomer 9
with that of the reaction mixture unambiguously reveals the
presence of the “linear” isomer in both (E) and (Z) forms, thus
corroborating the computational data. Further, it confirms this
isomerization process as being independent of the sulfur ylide
appendage.
Scheme 5. Relevant Conformations of the Intermediates in
the E−Z Isomerization Process
Having established and understood the grounds for isomer-
ization, we now sought to understand the remarkable stereo-
and regioconvergence of this transformation that enables the
formation of a single product from an interconverting mixture
of isomers (and regardless of the substrate isomer employed).
The DFT profile illustrated in Figure 13 represents the
mechanism of formation of product R, i.e., the major regio- and
stereoisomer of the cyclopropane product that is experimentally
obtained. The path calculated for the formation of R from the
branched substrate is entirely equivalent to the mechanism
calculated for the unsubstituted substrate, discussed above
(Figure 6).
Thus, it comprises two consecutive steps with C−C bond
formations in each. After a rearrangement of the carbon chain
of the substrate, from F to F′, there is formation of the lactone
ring in intermediate Q. In all mechanisms calculated for the
cyclopropanation of substituted substrates (R = Me) the
lactone intermediate forms as the stereoisomer with the
coordinated C atom (CAu) and the SPh2 is trans relative
positions, since this was shown to be the most favorable
intermediate from the stereochemical point of view (see Figure
9 and the corresponding discussion above). In transition state
TSF′Q the new C−C bond is only incipient as shown by a very
large distance (2.63 Å) and a negligible Wiberg index (WI =
0.09). The final step of the mechanism corresponds to the
formation of the C3 ring and loss of SPh2, from Q to R′. In the
transition state, TSQR′, the relevant bond lengths (dC−C = 2.19
Å, dC−S = 2.50 Å) indicate that both C−S bond breaking and
C−C bond formation are halfway through once the transition
state is reached. The highest barrier in the profile depicted in
Figure 12 corresponds to the formation of the lactone ring
(TSF′Q), being 16.4 kcal/mol.
while in the equivalent intermediate that yields the Z isomer
(O) those two fragments are cis to each other. This relates to
two different conformations of the Me−C−CHCH2 chain in
the initial η2 complex of the branched substrate that
interchange through a rotation around the allylic C−C bond.
One conformer (Scheme 5b, left) will lead to the E linear
substrate (cf. F in Figure 10), while the other (Scheme 5b,
right) will produce the Z linear substrate, corresponding to
intermediate M in the profile depicted in Figure 11. The energy
profile calculated for the interconversion between the two
relevant conformers of the η2 complex of the branched
substrate is provided in the Supporting Information as Figure
S15 and attests to a facile process with a maximum barrier of
7.5 kcal/mol.
The energy profile calculated for the branched-to-linear Z
isomerization is depicted in Figure 11. The first step in the
mechanism depicted in Figure 11 corresponds to rotation
around the COO−C bond bringing the ester carbonyl oxygen
into proximity of the internal olefin carbon, converting M into
the intermediate N. The transition state, TSMN, is reached when
the carbonyl group passes by the methyl substituent and, thus,
at a rather long O−C distance (3.98 Å). Intermediate N is a
five-membered cyclic acetal that coordinates the metal as an
alkyl ligand. In the following step, from N to O, there is a
simultaneous 1,2-shift of the O−C and of the Au−C bonds
such that the metal and the O atom exchange carbon atoms.
The result is an expansion of the acetal from a 5-membered ring
in N to a 6-membered ring in O. In the corresponding
transition state, TSNO, formation of the new O−C bond (2.14
Å, WI = 0.21) is well advanced and the previous one is
The mechanisms that lead to the formation of each one of
the other three isomers of the cyclopropane product, T, V, and
Y, were also calculated, and the corresponding energy profiles
F
DOI: 10.1021/acs.joc.5b00666
J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
Figure 11. Free energy profile (PBE0) for interchange between the branched substrate and its Z linear isomer. The minima and the transition states
were optimized and the structures obtained are presented with the more relevant bond distances, in each step. Free energy values (kcal/mol) are
referred to intermediate F.
Figure 12. 1H NMR spectra of (E)-8 (spectrum 1), pure 9 (spectrum 2), mixtures of pure (E)-8 + 5 mol % of gold complex 6 after 2 h (spectrum 3)
and after 7 h (spectrum 4; here, a mixture of 9 and (E)- and (Z)-8 is visible).
are presented in the Supporting Information (Figures S16, S17,
and S18). The main characteristics of those paths are similar to
the one obtained for product R, discussed above and, thus,
require no further discussion. However, the energy barriers
obtained for those mechanisms (19.0−22.2 kcal/mol) are all
higher than the ones calculated for the formation of R,
indicating that this is the most favorable of all four
cyclopropanation mechanisms.
Thus, on one hand, it emerges that cyclopropanation is
always preferred from the “branched” allyl isomer rather than
from the “linear” analogue (i.e., T/R preferred over Y/V, Figure
S14, Supporting Information). This is reflected in higher energy
barriers for the second mechanistic step, namely the formation
of the three-membered ring. The energy barriers practically
double from one type of substrate to the other, going from 13.5
kcal/mol (TSQR′) and 12.0 kcal/mol (TSST′) to 18.8 and 22.2
kcal/mol in TSUV′ and TSWY′, respectively. In fact, for the
“linear” substrates the second step of each mechanism has an
energy barrier either similar to (Figure S17) or even higher
than the first step, becoming rate determining in the formation
of product Y from the Z isomer of the “linear” substrate (Figure
S18). Again, it is the stereochemical disposition of the various
groups around the C−C bond that is being formed that dictates
the differences observed (Figure S20). Here, the steric
hindrance caused by the presence of a methyl substituent in
the terminal CAu atom hampers the formation of the
corresponding C−C bond, destabilizing the transition state
and raising the barrier.
The two cyclopropanation mechanisms for the branched
substrate, leading to the two possible different diastereoisomers,
are represented in the profiles of Figure 13 (cis product R) and
Figure S16 (trans product T). The major difference between
the two profiles is the first step (i.e., formation of the lactone
intermediate). Here, the stereochemical repulsion between the
methyl and the acyl groups in the corresponding transition
G
DOI: 10.1021/acs.joc.5b00666
J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
Figure 13. Free energy profile (PBE0) for cyclopropanation of the branched substrate yielding product R. The minima and the transition states were
optimized, and the structures obtained are presented with the more relevant bond distances in each step. Free energy values (kcal/mol) are referred
to the initial intermediate (F).
states (see the Supporting Information, Figure S19) results in a
difference of 2.6 kcal/mol in the corresponding barriers.
Validation of the Computational Predictions. The
mechanistic studies described above afford two main
conclusions. First, the isomerization process has an overall
energy barrier similar (TSNO, Figure 12) to the one presented
Scheme 6. Cyclopropanations of “Branched” Substrates
by the cyclopropanation reaction that forms product R (TSF′Q
,
Figure 13). In fact, the values are within 0.1 kcal/mol and can
be considered equal at the level of theory employed. The
second equally important result is that all cyclopropanation
paths that lead to the formation of the other three isomers of
the product (T, V, and Y) have higher energy barriers and, thus,
are less favorable than both substrate isomerization and
formation of R. This means that all substrate isomers
interconvert faster than any cyclopropanation reaction occurs,
except the one that produces R thus explaining the high regio-
and stereoselectivity of the process. The ability of gold to not
only catalyze both processes in what is effectively a domino
allylic isomerization/cyclopropanation, but also to do so with
such high selectivity is remarkable.
To validate these conclusions, we prepared several “linear”-
allyl-containing sulfur ylides and submitted them to the
cyclopropanation conditions. As shown in Schemes 3 and 6,
all linear substrates yield the corresponding “branched”
cyclopropanes, with substituents in the 4-position, in very
good to excellent diastereoselectivities. As predicted by DFT,
using either the E- or Z-olefin isomer of the substrate led to the
same product isomer in virtually identical yields and
diastereoselectivities (Schemes 3b and 6a). Isomerization
under these conditions appears to be limited to disubstituted
olefins, as use of a prenylated substrate blocks the reaction and
only starting material could be recovered (Scheme 6b).
Two additional experiments shed more light on the
peculiarities of these coupled processes (Scheme 7a). The
failure of double-substituted substrate 1o to undergo any
reaction (substrate recovered unchanged) implies that cyclo-
propanation of a “linear” allyl substrate has a prohibitively high
barrier under these conditions. It is worthy of note that the
allylic rearrangement of 1o is a degenerate process.
H
DOI: 10.1021/acs.joc.5b00666
J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
(m, 4H), 7.55−7.39 (m, 6H), 5.76−5.59 (m, 1H), 5.32−5.21 (m, 1H),
5.17 (dt, J = 17.2, 1.5 Hz, 1H), 5.03 (dt, J = 10.5, 1.5 Hz, 1H), 3.67 (s,
3H), 1.61−1.41 (m, 2H), 1.29−1.14 (m, 6H), 0.83 (t, J = 6.7 Hz, 3H).
13C NMR (75 MHz, CDCl3): δ = 167.2, 165.3, 137.8, 131.4, 131.3,
130.9, 130.8, 129.6, 129.5, 129.4, 115.7, 73.8, 51.2, 34.5, 31.8, 24.9,
22.7, 14.1. MS (EI): m/z = 285 (18), 186 (100). HRMS-(ESIpos): [M
+ Na]+ calcd for C24H28O4SNa 435.1600, found 435.1604. IR (neat):
3064, 2931, 2858, 2349, 1984, 1719, 1684, 1629, 1579, 1475, 1442,
1432, 1296, 1278, 1225, 1180, 1120, 1075, 1053, 998, 954, 921, 813,
769, 742, 684, 664. The enantiomeric excess was determined by chiral
HPLC using a Chiralpack AS-3 150 × 4.6 mm coloumn. Solvent
system: n-heptane + 0.1% i-PrOH/i-PrOH 9/1; flow rate 0.7 mL/min;
T = 25 °C.
Scheme 7. (a) Blocking the Cyclopropanation with 1,3
Disubstitution (b) and Using the Amide Bond in 10 To
Force Cyclopropanation of a “Branched” Substrate
Allyl-1-d2 Methyl 2-(Diphenylsulfuranylidene)malonate (1k). The
compound was prepared following the general procedure. Purification
by column chromatography over silica (isohexane/ethyl acetate 1/1)
afforded the title compound as a colorless oil (93.2 mg, 0.27 mmol,
54%). Rf = 0.2. 1H NMR (400 MHz, CDCl3): δ = 7.65−7.59 (m, 4H),
7.54−7.43 (m, 6H), 5.89 (dd, J = 17.2, 10.5 Hz, 1H), 5.31 (dd, J =
17.2, 1.7 Hz, 1H), 5.14 (dd, J = 10.5, 1.7 Hz, 1H), 3.68 (s, 3H) ppm.
13C NMR (101 MHz, CDCl3): δ = 166.9, 165.8, 133.4, 131.5, 130.6,
129.6, 129.5, 117.1, 77.4, 51.2 ppm. The deuterated C cannot be
observed. MS-(EI): m/z = 344 (1), 186 (100), 119 (15). HRMS-
(ESIpos): [M + Na]+ calcd for C19H16D2O4SNa 367.0943, found
367.0946.
(E)-Hept-2-en-1-yl Methyl 2-(Diphenylsulfuranylidene)malonate
(1l). The compound was prepared following the general procedure.
Purification by column chromatography over silica (n-pentane/ethyl
acetate 1/1) afforded the title compound as a colorless oil (97.1 mg,
0.24 mmol, 49%). Rf = 0.20. 1H NMR (300 MHz, CDCl3): δ = 7.69−
7.56 (m, 4H), 7.56−7.40 (m, 6H), 5.78−5.61 (m, 1H), 5.57−5.42 (m,
1H), 4.50 (dd, J = 6.0, 0.9 Hz, 2H), 3.67 (s, 3H), 2.06−1.92 (m, 2H),
1.37−1.18 (m, 4H), 0.87 (t, J = 7.1 Hz, 3H) ppm. 13C NMR (75 MHz,
CDCl3): δ = 167.0, 165.9, 134.9, 131.4, 130.8, 129.5, 129.5, 125.1,
64.5, 60.2, 51.2, 32.1, 31.3, 22.3, 14.0 ppm. MS (EI): m/z = 398 (0.2),
186 (100). HRMS (ESIpos): calcd for C23H26O4NaS 421.1443, found
421.1446. IR (neat): ν = 3061, 2952, 2927, 2871, 2297, 1720, 1685,
1630, 1579, 1475, 1442, 1432, 1377, 1299, 1278, 1225, 1181, 1072,
1024, 999, 971, 915, 824, 769, 744, 685 cm−1.
In contrast to this, the crotyl amide 10 underwent
cyclopropanation and afforded the 6-methyl cyclopropanation
product 11 (as determined by 2D-NMR analysis) in
quantitative yield and a dr of 4:1 (Scheme 7b). This is the
only instance where we have observed cyclopropanation of an
internal olefin. It is reasonable to assume that the hypothetical
isomerization process (converting the amide in 10 to an O-
allylated imidate) becomes energetically much less favorable (or
even prohibitive).15 In addition, the conformational rigidity of
amides may also contribute to a lowering of the energy barrier
for cyclopropanation. The conjunction of these two factors
presumably brings the hurdle for cyclopropanation of an
internal olefin to a surmountable value in 10.
CONCLUSIONS
■
In conclusion, we have presented a detailed study on a gold-
catalyzed olefin cyclopropanation that unusually proceeds by a
two-step mechanism critically hinging on olefin activation by
the metal catalyst. The mechanistic consequences of that
include not only a very high specificity of the individual steps
but also the emergence of a parallel, rapid (under the reaction
conditions), gold-catalyzed allylic isomerization process. The
latter is independent of the cyclopropanation transform; when
coupled to cyclopropanation, it results in very high regio- and
stereoselectivity. That the same metal promotes two
independent transformations which can function in tandem
with such high selectivity is testimony to the versatility of gold
in catalysis.
Methyl 3-Methylbut-2-en-1-yl 2-(Diphenylsulfuranylidene)-
malonate (1m). The compound was prepared following the general
procedure. Purification by column chromatography over silica
(isohexane/ethyl acetate 1/1) afforded the title compound as a
1
colorless oil (109 mg, 0.29 mmol, 59%). Rf = 0.16. H NMR (300
MHz, CDCl3): δ = 7.65−7.57 (m, 4H), 7.52−7.42 (m, 6H), 5.31−
5.23 (m, 1H), 4.55 (d, J = 7.0 Hz, 2H), 3.66 (s, 3H), 1.69 (d, J = 0.7
Hz, 3H), 1.63 (s, 3H) ppm. 13C NMR (75 MHz, CDCl3): δ = 167.0,
166.1, 137.2, 131.3, 130.8, 129.5, 129.5, 120.1, 60.7, 60.0, 51.2, 25.8,
18.1 ppm. MS (EI): m/z = 370 (2), 285 (10), 186 (100), 161 (12),
121 (13). HRMS-(ESIpos): [M + Na]+ calcd for C21H22O4SNa
393.1131, found 393.1134. IR (neat): ν = 3062, 2945, 2243, 1975,
1715, 1682, 1646, 1628, 1579, 1475, 1442, 1432, 1374, 1331, 1301,
1276, 1225, 1121, 1057, 1024, 999, 977, 952, 913, 837, 769, 745, 685
cm−1.
EXPERIMENTAL SECTION
■
Unless otherwise indicated, all glassware was flame-dried before use,
and all reactions were performed under an atmosphere of Argon. All
solvents were distilled from appropriate drying agents prior to use. All
reagents were used as received from commercial suppliers unless
otherwise stated. HRMS measurements were carried out with an ESI
QqTOF mass analyzer. All the ylides were prepared according to the
procedures reported by our group.16 Compounds 4b−g were prepared
following a procedure reported by our group.4f
Methyl Pent-3-en-2-yl 2-(Diphenylsulfuranylidene)malonate
(1n). The compound was prepared following the general procedure.
Purification by column chromatography over silica (n-pentane/ethyl
acetate 1/1) afforded the title compound as a colorless oil (156 mg,
1
0.42 mmol, 84%). Rf = 0.32. H NMR (300 MHz, CDCl3): δ = 7.61
General Procedure for the Preparation of Sulfonium Ylides.
To a solution of Martin’s sulfurane (336 mg, 0.5 mmol) in CH2Cl2 was
added a solution of 1,3-dicarbonyl compound (0.5 mmol). After the
mixture was stirred at room temperature for 2 h, all volatiles were
removed in vacuo, and the crude mixture was purified by column
chromatography over silica.
(dt, J = 3.8, 2.5 Hz, 4H), 7.54−7.41 (m, 6H), 5.60 (dt, J = 9.3, 6.5 Hz,
1H), 5.39−5.27 (m, 2H), 3.68 (s, 3H), 1.61 (dd, J = 6.5, 0.6 Hz, 3H),
1.15 (d, J = 6.3 Hz, 3H) ppm. 13C NMR (75 MHz, CDCl3): δ = 167.3,
165.2, 131.9, 131.3, 131.0, 129.5, 129.4, 126.8, 70.1, 60.2, 51.2, 20.7,
17.8 ppm. MS (EI): m/z = 285 (13), 186 (100), 161 (16), 121 (14).
HRMS-(ESIpos): [M + Na]+ calcd for C21H22O4SNa 393.1131, found
393.1134. IR (neat): ν = 3061, 2974, 2945, 1717, 1682, 1630, 1579,
1476, 1442, 1432, 1331, 1297, 1227, 1180, 1136, 1062, 1039, 999, 965,
895, 857, 804, 770, 744, 685, 663 cm−1.
(R)-Methyl Octen-3-yl 2-(Diphenylsulfuranylidene)malonate (1j).
The compound was prepared following the general procedure.
Purification by column chromatography over silica (n-pentane/ethyl
acetate 1/1) afforded the title compound as colorless oil (206 mg, 0.5
Methyl 3-(Benzyl(but-2-en-1-yl)amino)-2-(diphenylsulfuranyli-
dene)-3-oxopropanoate (10). The compound was prepared following
1
mmol, 99%). Rf = 0.46. H NMR (300 MHz, CDCl3): δ = 7.67−7.57
I
DOI: 10.1021/acs.joc.5b00666
J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
the general procedure. Purification by column chromatography over
silica (isohexane/ethyl acetate 2/1) afforded the title compound as a
yellow oil (140 mg, 0.34 mmol, 67%). Rf = 0.19. 1H NMR (300 MHz,
CDCl3): δ = 7.74−7.65 (m, 4H), 7.53−7.39 (m, 6H), 7.21−7.11 (m,
3H), 6.99 (dd, J = 7.3, 2.0 Hz, 2H), 5.43−5.30 (m, 2H), 4.58 (s, 2H),
3.83 (d, J = 4.4 Hz, 2H), 3.56 (s, 3H), 1.61 (d, J = 4.6 Hz, 3H) ppm.
13C NMR (75 MHz, CDCl3): δ = 168.4, 166.0, 138.6, 131.9, 131.1,
129.9, 129.4, 128.3, 128.2, 127.7, 127.5, 126.6, 59.4, 53.5, 50.5, 17.8
ppm. MS (EI): m/z = 200 (16), 186 (100), 91 (26). HRMS-(ESIpos):
[M + Na]+ calcd for C27H27NO3SNa 468.1603, found 468.1608. IR
(neat): ν = 3061, 3026, 2944, 2916, 2854, 2239, 1961, 1739, 1645,
1588, 1574, 1494, 1475, 1441, 1431, 1407, 1359, 1279, 1233, 1179,
1131, 1098, 1075, 1023, 999, 970, 950, 910,728, 686 cm−1.
General Procedure for Cyclopropanation. A solution of
sulfonium ylide (0.2 mmol) and Echavaren’s catalyst (0.01 mmol)
was dissolved in dry toluene. The resulting mixture was heated to 100
°C until TLC showed complete consumption of the ylide. The
resulting solution was cooled to room temperature and directly
purified by column chromatography affording the desired cyclo-
propanation product.
ASSOCIATED CONTENT
* Supporting Information
Crystallographic data, details of mechanistic experiments,
documentation of theoretical calculations, NMR spectra, and
chromatograms. The Supporting Information is available free of
charge on the ACS Publications website at DOI: 10.1021/
acs.joc.5b00666.
■
S
AUTHOR INFORMATION
Corresponding Authors
*E-mail: veiros@ist.utl.pt.
*E-mail: nuno.maulide@univie.ac.at.
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
Generous support of this research by the Deutsche
Forschungsgemeinschaft (DFG, Grants MA 4861/4-1 and/4-
2) and the University of Vienna is acknowledged. Mrs. U.
Methyl 2-Oxo-4-pentyl-3-oxabicyclo[3.1.0]hexane-1-carboxylate
(4j). The compound was prepared according to the general procedure
using ylide 1j. Purification by column chromatography over silica (n-
pentane/ethyl acetate 9/1 to 1/1) afforded the title compound as a
colorless oil (45 mg, 199 μmol, 99%, dr = 95:5). Rf = 0.68 (n-pentane/
Rosentreter (MPI Mulheim) and Mrs. E. Macoratti (U.
̈
Vienna) are acknowledged for analytical determination. Parts
of this research were carried out at the Max-Planck-Institut fur
Kohlenforschung, which we gratefully acknowledge. L.F.V.
̈
1
ethyl acetate 1/1). H NMR (300 MHz, CDCl3): δ = 4.28 (t, J = 6.2
Hz, 1H), 3.80 (s, 3H), 2.50 (dd, J = 8.1, 5.5 Hz, 1H), 2.02 (dd, J = 8.1,
4.7 Hz, 1H), 1.74−1.62 (m, 1H), 1.48−1.23 (m, 8H), 0.87 (t, J = 6.9
Hz, 3H) ppm. 13C NMR (75 MHz, CDCl3): δ = 170.2, 167.5, 79.0,
53.0, 36.0, 32.7, 31.5, 29.8, 23.8, 22.5, 21.1, 14.0. MS (EI): m/z = 155
(100), 127 (23), 95 (12) ppm. HRMS (ESIpos): calcd for
C12H18O4Na: 249.1097, found 249.1096. IR (neat): ν (cm−1) =
2955, 2932, 2860, 2318, 1776, 1728, 1439, 1377, 1349, 1308, 1264,
1231, 1202, 1174, 1125, 1095, 1083, 1045, 991, 945, 912, 790, 771,
729 cm−1. The enantiomeric excess was determined by chiral GC using
a IVADEX 1/PS 086 G 548 (25 m) column: temp 80 1N 180 10/MIN
220; 0.4 bar H2. Minor enantiomer tR = 72.28 min.Major enantiomer:
tR = 73.32 min.
Methyl 4-Butyl-2-oxo-3-oxabicyclo[3.1.0]hexane-1-carboxylate
(4l). The compound was prepared according to the general procedure
using ylide 1l. Purification by column chromatography over silica (n-
pentane/ethyl acetate 5/1 to 1/1) afforded the title compound as a
colorless oil (27 mg, 0.13 mmol, 82%, dr = 95:5). Rf = 0.68 (n-
pentane/ethyl acetate 1/1). 1H NMR (300 MHz, CDCl3): δ 4.29 (t, J
= 6.2 Hz, 1H), 3.82 (s, 3H), 2.54−2.45 (m, 1H), 2.04 (dd, J = 8.1, 4.7
Hz, 1H), 1.71 (ddd, J = 13.2, 7.9, 4.9 Hz, 2H), 1.49−1.22 (m, 5H),
0.92 (t, J = 7.0 Hz, 3H) ppm. 13C NMR (75 MHz, CDCl3): δ 170.2,
167.6, 79.1, 53.1, 35.8, 32.7, 29.9, 26.3, 22.5, 21.2, 14.0 ppm. MS (EI):
m/z = 155 (100), 127 (32), 123 (10), 95 (17), 59 (10). HRMS
(ESIpos): calcd for C11H16O4Na 212.24, found 212.24. IR (neat): ν
(cm−1) = 2956, 2933, 2862, 1777, 1728, 1439, 1376, 1349, 1309, 1265,
1202, 1174, 1094, 1047, 979, 937, 911, 788, 767, 734 cm−1.
Methyl 3-Benzyl-6-methyl-2-oxo-3-azabicyclo[3.1.0]hexane-1-
carboxylate (11). The compound was prepared according to the
general procedure using ylide 10. Purification by column chromatog-
raphy over silica (isohexane/ethyl acetate 2/1 to 1/1) afforded the title
compound as a colorless oil (45 mg, 199 μmol, 99%, dr = 4.5:1). Rf =
0.34 (isohexane/ethyl acetate 1/1). 1H NMR (300 MHz, CDCl3): δ =
7.31−7.07 (m, 5H), 4.43 (d, J = 14.6 Hz) and 4.39 (d, J = 14.6 Hz)
(1H), 4.18 (d, J = 14.6 Hz) and 4.16 (d, J = 14.6 Hz) (1H), 3.75 (s)
and 3.72 (s) (1H), 3.38 (dd, J = 11.0, 6.4 Hz) and 3.32 (dd, J = 10.2,
5.8 Hz) (1H), 3.02 (d, J = 10.2 Hz) and 2.90 (d, J = 11.0 Hz) (1H),
2.27−2.15 (m) and 2.05 (t, J = 5.4 Hz, 1H), 1.39−1.27 (m, 1H), 1.19
(d, J = 6.1 Hz) and 0.88 (d, J = 6.4 Hz) (1H) ppm. 13C NMR (75
MHz, CDCl3): δ = 170.2, 169.7, 168.0, 167.3, 136.5, 136.0, 128.8,
128.8, 128.3, 127.9, 127.8, 52.6, 52.5, 46.7, 46.6, 43.1, 37.2, 36.9, 29.6,
27.7, 26.6, 25.6, 12.2, 7.1 ppm. MS (EI): m/z = 259 (64), 227 (67),
199 (60), 118 (30), 109 (12), 95 (19), 91 (100), 65 (20). HRMS
(ESIpos): calcd for C15H17O3NNa 282.1106, found 282.1099.
acknowledges Fundaca
̧
o para a Cien
̂
cia e Tecnologia, Projecto
̃
Estrateg
́
ico - PEst-OE/QUI/UI0100/2013.
REFERENCES
■
(1) For selected reviews on gold(I) catalysis, see: (a) Dyker, G.
Angew. Chem., Int. Ed. 2000, 39, 4237. (b) Hoffmann-Roder, A.;
Krause, N. Org. Biomol. Chem. 2005, 3, 387. (c) Hashmi, A. S. K.;
Hutchings, G. Angew. Chem., Int. Ed. 2006, 45, 7896. (d) Furstner, A.;
Davies, P. W. Angew. Chem., Int. Ed. 2007, 46, 3410. (e) Jimenez-
Nunez, E.; Echavarren, A. M. Chem. Commun. 2007, 333. (f) Hashmi,
A. S. K. Chem. Rev. 2007, 107, 3180. (g) Li, Z.; Brouwer, C.; He, C.
Chem. Rev. 2008, 108, 3239. (h) Arcadi, A. Chem. Rev. 2008, 108,
3266. (i) Shen, H. C. Tetrahedron 2008, 64, 3885. (j) Shen, H. C.
̈
̈
Tetrahedron 2008, 64, 7847. (k) Furstner, A. Chem. Soc. Rev. 2009, 38,
̈
3208. (l) Shapiro, N.; Toste, F. D. Synlett 2010, 675. (m) Bandini, M.
Chem. Soc. Rev. 2011, 40, 1358.
(2) (a) Chiarucci, M.; Bandini, M. Beilstein J. Org. Chem. 2013, 9,
2586. (b) Marion, N.; Nolan, S. P. Chem. Soc. Rev. 2008, 37, 1776.
(3) For selected examples, see: (a) Yang, C.; He, C.-G. J. Am. Chem.
Soc. 2005, 127, 6966. (b) LaLonde, R. L.; Brenzovich, W. E., Jr.;
Benitez, D.; Tkatchouk, E.; Kelley, K.; Goddard, W. A., III; Toste, F.
D. Chem. Sci. 2010, 1, 226. (c) Brouwer, C.; He, C. Angew. Chem., Int.
Ed. 2006, 45, 1744. (d) Bender, C.-F.; Widenhoefer, R. A. Chem.
Commun. 2006, 45, 4143. (e) Han, X.; Widenhoefer, R. A. Angew.
Chem., Int. Ed. 2006, 45, 1747. (f) Liu, X.-Y.; Li, C. H.; Che, C. M. Org.
Lett. 2006, 8, 2707. (g) Li, H.; Song, F.; Widenhoefer, R. A. Adv. Synth.
Catal. 2011, 353, 955. (h) Kanno, O.; Kuriyama, W.; Wang, Z. J.;
Toste, F. D. Angew. Chem., Int. Ed. 2011, 50, 9919.
(4) For examples, see: (a) Zhou, C.-Y.; Che, C.-M. J. Am. Chem. Soc.
2007, 129, 5828. (b) Wang, M.-Z.; Wong, M.-K.; Che, C.-M. Chem.
Eur. J. 2008, 14, 8353. (c) Xiao, Y.-P.; Liu, X.-Y.; Che, C.-M. Angew.
Chem., Int. Ed. 2011, 50, 4937. (d) Xiao, Y.-P.; Liu, X.-Y.; Che, C.-M.
Beilstein J. Org. Chem. 2011, 7, 1100. (e) Rao, W.; Susanti, D.; Chan, P.
W. H. J. Am. Chem. Soc. 2011, 133, 15248. (f) Huang, X.; Klimczyk, S.;
Veiros, L. F.; Maulide, N. Chem. Sci. 2013, 4, 1105.
(5) Huang, X.; Peng, B.; Luparia, M.; Gomes, L.; Veiros, L. F.;
Maulide, N. Angew. Chem., Int. Ed. 2012, 51, 8886.
(6) (a) Nieto-Oberhuber, C.; Lop
́ ́
ez, S.; Munoz, M. P.; Cardenas, D.
̃
J.; Bunuel, E.; Nevado, C.; Echavarren, A. M. Angew. Chem., Int. Ed.
̃
́ ́ ́
2005, 44, 6146. (b) Perez-Galan, P.; Delpont, N.; Herrero-Gomez, E.;
Maseras, F.; Echavarren, A. M. Chem.Eur. J. 2010, 16, 5234.
(c) Lop
́
ez-Carrillo, V.; Echavarren, A. M. J. Am. Chem. Soc. 2010, 132,
J
DOI: 10.1021/acs.joc.5b00666
J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
9292. (d) Homs, A.; Obradors, C.; Leboeuf, D.; Echavarren, A. M.
Adv. Synth. Catal. 2014, 356, 1.
(7) The 31P NMR shift for the free ligand 6 is δ = 21.4 ppm; for
comparison, see: (a) Huang, X.; Anderson, K. W.; Zim, D.; Jiang, L.;
Klapars, A.; Buchwald, S. L. J. Am. Chem. Soc. 2003, 125, 6653. The
31P NMR shift for the phosphine oxide of ligand 6 is δ = 53.7 ppm; for
comparison, see: (b) Zhang, D.; Gua, A.; Doan, C.; Stewart, T.; Bau,
R.; Selke, M. Org. Lett. 2010, 10, 3100.
(8) Parr, R. G.; Yang, W. In Density Functional Theory of Atoms and
Molecules; Oxford University Press: New York, 1989.
(9) The single-point free energy values presented here and in ref 4f
result from different solvation models. In ref 4f, it was used the PCM
model as implemented in Gaussian 03 with a molecular cavity based
on the united atom topological model applied on UAHF radii,
optimized for the HF/6-31G(d). Here, we used the PCM model as
implemented in Gaussian 09 with with radii and nonelectrostatic terms
of the SMD solvation model, developed by Truhler et al. (see the
Computational Details in the Supporting Information).
(10) (a) Wiberg, K. B. Tetrahedron 1968, 24, 1083. (b) Wiberg
indices are electronic parameters related to the electron density
between atoms. They can be obtained from a natural population
analysis and provide an indication of the bond strength..
(11) For examples of metal−cyclopropane σ-complexes, see:
(a) Youngs, W. J.; Simms, B. L.; Ibers, J. A. J. Organomet. Chem.
1994, 272, 295. (b) Youngs, W. J.; Ibers, J. A. Organometallics 1983, 2,
979. (c) Cutler, A.; Fish, R. W.; Giering, W.; Rosenblum, M. J. Am.
Chem. Soc. 1972, 94, 4354.
(12) For literature precedents on SN2-type displacement at tertiary
sp3-carbon electrophiles, see: (a) Pronin, S. V.; Reiher, C. A.; Shenvi,
R. A. Nature 2013, 501, 195. (b) Mascal, M.; Hafezi, N.; Toney, M. D.
J. Am. Chem. Soc. 2010, 132, 10662.
(13) See the Supporting Information for further details.
(14) For a similar rearrangement, see: (a) Marion, N.; Gealageas, R.;
Nolan, S. P. Org. Lett. 2007, 9, 2653. For a computational study of
related processes, see: (b) Paton, R. S.; Maseras, F. Org. Lett. 2009, 11,
2237.
(15) For the reverse of this hypothetical reaction, the Overman
rearrangement, see: (a) Overman, L. E. J. Am. Chem. Soc. 1976, 98,
2901. (b) Anderson, C. E.; Overman, L. E. J. Am. Chem. Soc. 2003,
125, 92697. For a gold-catalyzed version, see: (c) Xing, D.; Yang, D.
Beilstein J. Org. Chem. 2011, 7, 781.
(16) For our previously reported ylide transfer, see: Huang, X.;
Goddard, R.; Maulide, N. Angew. Chem., Int. Ed. 2010, 49, 8979.
K
DOI: 10.1021/acs.joc.5b00666
J. Org. Chem. XXXX, XXX, XXX−XXX