Mechanistic insights into enantioselective gold-catalyzed allylation of indoles with alcohols: The counterion effect

Article abstract of DOI:10.1021/ja3086774

Enantioselective gold-catalysis is emerging as a powerful tool in organic synthesis for the stereoselective manipulation of unfunctionalized unsaturated hydrocarbons. Despite the exponential growth, the molecular complexity of common chiral gold complexes

Full text of DOI:10.1021/ja3086774

Subscriber access provided by FORDHAM UNIVERSITY
Article
Mechanistic Insights into Enantioselective Gold-Catalyzed
Allylation of Indoles with Alcohols: the Counterion Effect
Marco Bandini, Andrea Bottoni, Michel Chiarucci, Gianpiero Cera, and Gian Pietro Miscione
J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/ja3086774 • Publication Date (Web): 30 Nov 2012
Downloaded from http://pubs.acs.org on December 10, 2012
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of the American Chemical Society is published by the American Chemical
Society. 1155 Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 13
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
Mechanistic Insights into Enantioselective Gold-Catalyzed Al-
lylation of Indoles with Alcohols: the Counterion Effect
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Marco Bandini,* Andrea Bottoni, Michel Chiarucci, Gianpiero Cera, Gian Pietro Miscione*
Dipartimento di Chimica “G. Ciamician”, Alma Mater Studiorum – Università di Bologna, via Selmi 2, 40126 Bologna,
Italy.
KEYWORDS Alcohols, Asymmetric Catalysis, Catalytic Mechanism, Counterions, Computational Investigation, Gold, In-
dole.
Supporting Information Placeholder
ABSTRACT: Enantioselective gold-catalysis is emerging as a powerful tool in organic synthesis for the stereoselective manipula-
tion of unfunctionalized unsaturated hydrocarbons. Despite the exponential growth, the molecular complexity of common chiral
gold complexes generally prevents a complete description of the mechanism steps and activation modes being documented. In this
study, we present the results of a combined experimental-computational (DFT) investigation of the mechanism of the enantioselec-
tive gold-catalyzed allylic alkylation of indoles with alcohols. A step-wise SN2’-process (i.e. anti-auroindolination of the olefin,
proton-transfer and subsequent anti-elimination [Au]-OH) is disclosed, leading to a library of tricyclic-fused indole derivatives. The
pivotal role played by the gold counterion, in terms of molecular arrangement (i.e. “folding effect”) and proton-shuttling in restor-
ing the catalytic species, is finally documented.
OTf, BF4, OBz, NTf2, ClO4, etc.) turns out to be remarkably
INTRODUCTION
Over the recent years, enantioselective [gold(I)] catalysis has
affected by the counterion utilized and, with the exception of a
few leading examples,8 the real role of the anions over both
become a concrete reality in the realm of asymmetric synthe-
sis.1 The “re-discovering” of this coinage metal in homogene-
ous catalysis allowed to expand significantly the current syn-
thetic portfolio for the selective manipulation of unactivated
unsaturated hydrocarbons. Moreover, a concrete contribution
toward the replacement of “old” chemistry with more efficient
and sustainable chemical methodologies was documented.2
Worth noting is the concomitant and rapid diffusion of enanti-
oselective gold catalysis in organic synthesis, that can be also
ascribable to the current large volume of synthetically flexible
and sterically congested P- and NCH-based chiral ligands ca-
pable of generating stable and fine-tunable chiral gold com-
plexes. In particular, chiral C1-symmetric mononuclear, C2-
symmetric binuclear and C3-symmetric trinuclear3 gold(I) sys-
tems have already shown particular pertinence in the stereose-
lective manipulation of unactivated π-systems.1b
kinetic and stereochemical reaction profiles still remains ob-
scure.9
In line with our ongoing interests on the heterocyclic-oriented
methodology development,10 some of us have recently report-
ed on the use of cationic chiral binuclear [Au(I)] complexes as
efficient electrophilic activators of unfunctionalized carbon-
carbon double bonds towards the intramolecular allylic func-
tionalization of indoles.11 In particular, environmentally be-
nign allylic alcohols were employed in additive-free enanti-
oselective nucleophilic substitution reactions (Figure 1).12
Despite efficiency, the unambiguous elucidation of reaction
intermediates in gold catalysis still remains a challenging
task,4 and even more pronounced are the mechanistic ambigui-
ties within enantioselective gold-catalyzed processes. The state
of the art is partly ascribable to the complex architectures of
the chiral [Au(I)] π-acids used in asymmetric transformations.
In particular, binuclear gold complexes of general formula [P-
P(AuX)2]5 lead to the possibility of multiple reaction path-
ways/activation modes, but restrict at the same time the ap-
plicability of experimental/computational treatments. Fur-
thermore, additional issues in mechanism elucidation lie in the
mandatory activation of the chiral gold chloride pre-catalysts
with silver-salts (i.e. halide scavengers)6 in order to ensure sat-
isfying levels of carbophilicity.7 In this context the catalytic
performance of the resulting [P-P(AuX)2] adducts (X: SbF6,
Figure 1. Gold-catalyzed enantioselective allylic alkylation of
indoles with primary alcohols.
The intramolecular methodology worked smoothly for both
C(2)- and C(3)-alkylation of indolyl cores, several tethering
units as well as functional groups were also well tolerated de-
livering a rapid route to key building blocks for biologically
active alkaloids.13 In particular, a range of 1-vinyl- (2),
4-vinyl-tetrahydrocarbazoles (4) and 4-vinyl-tetrahydro-β-
carbolines (3, vinyl-THBCs) were isolated in good yields and
excellent enantiomeric ratios (Chart 1). Moreover, the meth-
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 2 of 13
Gold-catalyzed organic manipulations of π-systems commonly
deal with a pronounced reagent-controlled stereoselectivi-
ty.15,16 Some exceptions concern the dehydrative nucleophilic
substitutions17 with enantioenriched π-alcohols as alkylating
agents in which the starting material undergoes partial or total
racemization.18
Recent reports by Widenhoefer15a,g,i and Aponick15c,f on the
[Au(I)]-assisted nucleophilic allylic substitution with alcohols,
have further pointed out the influence of stereochemistry of
the reaction partners on the entire process. However, despite
these inspiring records, reagent-controlled stereochemistry in
gold-promoted enantioselective allylic alkylation reactions
still remains barely investigated.15j,19
In our original communication,12a we discussed on the highly
stereoselective synthesis of 1-vinyl-tetrahydrocarbazole (R)-2a
(yield = 95%, ee = 90%), by subjecting the corresponding acy-
clic precursor (Z)-1a to the dehydrative ring-closing allylic
odology found application also in the enantioselective synthe-
sis of 1-vinyl-morpholines via intramolecular dehydrating oxa-
allylic alkylation with diols 12c
1
2
3
4
5
6
7
8
CHART 1. A representative collection of tricyclic indolyl
scaffolds available through the intramolecular enantioselective
allylic alkylation procedure (in red the newly formed C-C
bonds, with (R)-DTBM-MeObiphep, toluene, 0 °C/rt, 16 h).
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
alkylation,
in
the
presence
of
(R)-DTBM-
MeObiphep(AuCl)2/AgOTf (10 mol%, in toluene at 0 °C,
Scheme 1). On the contrary, the diastereomeric alcohol (E)-1a
proved to be completely inert under best reaction conditions.
SCHEME 1. Proofs of reagent-controlled stereochemistry in
the synthesis of 1-vinyl-tetrhydrocabazole 2a.
Preliminary experimental evidences suggested that some spe-
cific structural features of the acyclic precursors and gold-pre-
catalysts, namely: i) configuration of the carbon-carbon double
bond, ii) nature of the gold-counterion and iii) leaving hydrox-
yl group played a pivotal role over the chemical and optical
outcome of the process.12a These aspects open up intriguing
mechanistic perspectives concerning the coordination mode of
the binuclear gold species with the allylic alcohols (i.e. single
site- vs bidentate interaction) and lead us to envisage the pres-
ence of a complex interplay of secondary interactions involv-
ing the counterion during the enantiodiscriminating step of the
reaction. Moreover, important issues concerning the mechanis-
tic course are still unsolved. Firstly, both SN1 and SN2’ (step-
wise or concerted) reaction channels constitute reasonable
mechanistic alternatives to rationalize the overall chemical
output (Figure 2). Secondly, the regiochemistry of the initial
nucleophilic attack of the indole core on the alkylating agent
(i.e C(2) vs C(3) position) is still matter of debate.
At first instance, the recorded dichotomy in reactivity could be
rationalized in terms of an unfavorable spatial arrangement of
the sterically demanding chiral bimetallic catalyst with the
trans allylic alcohol. However, no further information was
available.
Moreover, the pivotal role of the counterion in affecting the
reaction profile, was demonstrated by screening a range of sil-
ver salts in the model reaction ((Z)-1a ꢀ 2a). The results are
collected in Table 1.
TABLE 1. Counterion effect in the gold-catalyzed enantiose-
lective synthesis of tetrahydrocarbazole 2a.a
Figure 2. Still open tasks in the mechanistic description of the
gold-catalyzed allylic alkylation of indoles with alcohols.
In this investigation, we wish to elucidate these intriguing and
still pending problems, by documenting a combined experi-
mental and theoretical investigation in which, the enantiose-
lective synthesis of 1-vinyl-tetrahydrocarbazoles 2, was used
as the model-reaction.14
Entry
AgX
Yield (%)b
Ee (%)c
87
L
L1
“
1
2
3
AgBF4
AgSbF6
AgNTf2
55
75
95
62
RESULTS AND DISCUSSION
“
65
2
ACS Paragon Plus Environment

Page 3 of 13
Journal of the American Chemical Society
4
“
AgOTf
AgOTs
AgOPNB
AgX1
78
88
92
nd
nd
nd
1
2
3
4
5
6
7
8
“
10
“
“
traces
traces
traces
5
dppm
AgX1
6
7
8
9
a All reactions were carried out under nitrogen atmosphere with
b
anhydrous toluene, unless otherwise specified. Isolated yields
c
after flash chromatography. Determined by HPLC analysis with
chiral column. nd = not determined.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Generally, moderate to high enantiomeric excesses were rec-
orded with fluoro-based counterions (entries 1-4, 62-88%).
Excellent stereoinduction was also obtained in the presence of
AgOTs (ee = 92%, entry 5) albeit along with modest yield.
Here, the lower turnovers observed for AgOTs and AgOPNB
(PNB: p-nitrobenzoate, entry 6)20 can be rationalized in terms
of the higher coordinating character of these anions towards
[Au(I)] atoms with respect to triflate and hexafluoroantimo-
nate. Finally, the pertinence of the chiral counterion strategy,
The small model-system. We discuss first the results obtained
for MODEL 1. A schematic representation of two possible
low energy structures of the starting reactants is given in Fig-
ure 3. A point concerning Figure 3A and 3C should be
stressed. Since in these figures we have adopted a two-
dimensional representation, several atomic distances are not
realistic and appear much longer (or shorter) than in the real
molecules. Also, the labels used to identify the various atoms
do not follow the order usually employed for aromatic rings.
Furthermore, within the carbon framework we have explicitly
indicated only the hydrogen atoms needed to identify the ste-
reochemical features of the process. A more realistic three-
dimensional picture of the model-system is given in Figure 3B
and 3D.
8b,21
in our ring-closing protocol was assessed in the presence
of both chiral [L1(Au2Cl2)] and achiral [dppm(Au2Cl2)] gold
complexes and chiral silver salt AgX1 (entries 7-8). However,
the desired tetracyclic scaffold was formed only in traces.
Therefore, beside the expected role of the counterion on the
kinetic reaction profile, we can first conclude that the anions
must be involved in some of the key-intermediates which are
responsible for the stereochemical translation.
To elucidate the above described points (in particular reagent-
controlled stereochemistry and counterion effect) we have car-
ried out a computational investigation of the reaction mecha-
nism using a DFT approach and two different model-systems
differing in the size of the ligand coordinating the two gold
atoms. The basic framework of the system is formed by com-
pound 1a (where we have replaced the two CO2Et groups on
the side-chain with two hydrogen atoms) and the gold com-
plex L1(AuOTf)2. In the smaller system (MODEL 1) Ar is a
simple benzene ring while in the larger system (MODEL 2)
Ar = 3,5-(t-Bu)2-4-OMe-Ph.
In order to verify the reliability of our computational models
we firstly investigated experimentally the role of the malonyl
tethering unit. To this aim, alcohol (Z)-1f proved to be syn-
thetically accessible in five steps starting from indolyl-alcohol
5 (Scheme 2, unoptimized yields are reported). Interestingly,
by subjecting (Z)-1f to optimal conditions, a prolonged reac-
tion time (48 h, yield = 58%) was required, with respect to
model substrates (1a-e), supporting the presence of a Thorpe-
Ingold effect22 played by the malonyl group. However, satisfy-
ing enantiocontrol was still produced (ee = 70%), underlying
the reliability of our model-system.
Figure 3. Two-dimensional and three-dimensional representations
of m0-SiSi-cis and m0-ReRe-cis (bond lengths are in Ångstroms).
SCHEME 2. Proving the role of the malonyl groups in the
enantioselective ring-closing process.
In these molecules each gold atom has a formal positive
charge +1, while each triflate group bears a negative charge;
therefore the total charge is zero. The initial gold complex (re-
actants) can exist as different isomers. Some of them are con-
formational isomers that mainly differ in the relative position
of the two reacting moieties (indole and allylic alcohol group)
and in the nature of triflate-substrate interactions. Further
3
ACS Paragon Plus Environment

Journal of the American Chemical Society
structural arrangements originate from the trans-cis configura-
Page 4 of 13
1
tion of the C2-C3 double bond.
2
3
4
5
6
7
8
9
The energies of the various isomeric forms of the starting reac-
tant species are reported in the upper side (A) of the diagram
of Figure 4. To denote the various complexes we have adopted
appropriate names that contain information concerning the in-
dole and alkene faces (Re or Si) involved in the reaction and
the configuration of the C2-C3 double bond. Thus, for in-
stance, m0-SiSi-cis indicates that the C2-C3 double bond has a
cis configuration and the reaction involves the two Si faces of
the indole moiety and olefin C2-C3 fragment. In Figure 4A we
have collected the energy levels of the cis isomers on the left
and those of the trans isomers on the right. m0-SiSi-cis is the
most stable isomer and corresponds to the structure depicted in
Figure 3A and 3B where it is easy to recognize the strong η2
interaction between the double bond C2-C3 and one (Au1) of
the two gold atoms. This interaction is not symmetrical, the
Au1-C3 bond length (2.32 Å) being shorter than Au1-C2 (2.41
Å). Interestingly, in this structure one triflate ion is bridging
the indole hydrogen H1 and the hydroxyl oxygen O3 by means
of two rather strong hydrogen-bonds: NH1…O1 (H1…O1 =
1.96 Å) and O3H2…O2 (H2…O2 = 1.84 Å). In the following
discussion we shall refer to this bridging interaction as the
“folding effect” or “chelating effect” of the anion since it
forces the two reactive sites of the molecule to move closer
adopting the right orientation to react. As a consequence the
m0-SiSi-cis complex assumes a U-turn-type geometry, as evi-
denced in Figure 3A and B, where the C1…C2 distance is 3.34
Å. The other triflate ion is bonded to the second gold atom
Au2. This atom does not interact with the C2-C3 double bond
and we shall see that it behaves like a “spectator” throughout
the entire process. The computed Mulliken charges reveal that,
even if both C2 and C3 are negative, the former (which is sub-
jected to the nucleophilic attack) is less negative than the latter
(net charges are -0.16 and -0.35, respectively).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 4. (A) Energies of the cis (left) and trans (right) starting
adducts m0 in the various conformations, the corresponding tran-
sition states ts1 and the resulting intermediates m1 (MODEL 1).
Values in parenthesis are relative to m0-SiSi-cis. (B) Energy pro-
file for the reaction path originating from m0-ReRe-cis. Energies
in kcal mol-1.
Starting adducts featuring a trans C2-C3 double bond have
also been considered. Even if the overall bonding pattern of
these structures is similar to that of the corresponding cis spe-
cies, the energy spectrum significantly varies (see right side of
Figure 4A): m0-ReSi-trans is now the most stable isomer (on-
ly 1.7 kcal mol-1 above m0-SiSi-cis), while the two isomers
where the reaction occurs on the Re face of the C2-C3 olefin
fragment, m0-ReRe-trans and m0-SiRe-trans, are 5.5 and 5.9
kcal mol-1 higher than m0-SiSi-cis, respectively. The highest
energy structure becomes m0-SiSi-trans, 8.4 kcal mol-1 above
The diagram of Figure 4A shows that the isomers where the
Re face of the indole moiety (carbon C1) is involved in the
reaction i.e. m0-ReRe-cis and m0-ReSi-cis are higher in ener-
gy: 2.4 and 3.1 kcal mol-1 above m0-SiSi-cis, respectively. The
structural features of m0-ReRe-cis (see Figure 3C and 3D) are
similar to those of m0-SiSi-cis, except for a larger Au1-Au2
distance: 5.24 Å instead of 3.59 Å. Although well below the
sum of the van del Waals radii (3.80 Å), Au-Au distances
greater than 3.50 Å are only marginally ascribed to aurophilic
interactions.23 However, a contribution of this kind cannot be
completely ruled out in our system to account for the slight
energy difference between the m0-ReRe-cis and m0-SiSi-cis
isomers. The triflate ion exerts here the same “folding effect”
previously described. This effect seems to be more important
in m0-ReRe-cis since the C1…C2 distance is shorter (3.18 Å)
with respect to m0-SiSi-cis.
m0-SiSi-cis
(a
two-dimensional
representation
of
m0-ReRe-trans is given in Figure S1 of the Supporting Infor-
mation). Since the energies of the cis and trans structures are
not so far one from the other and the approximations adopted
in the simpler model-system could seriously affect the energy
spectrum, all species have been initially considered as possible
starting complexes of alternative reaction channels.
Our computations have shown the possibility of a different
coordination mode involving the gold atoms and the allyl OH
group. This alternative molecular arrangement is character-
ized by a 7-membered ring that includes Au1, Au2, O4, H2,
O3, C4, C3 and can be observed when the allylic hydroxyl
points towards the Au2 atom and the second triflate group.
However, these molecular aggregates (deriving from the two-
4
ACS Paragon Plus Environment

Page 5 of 13
Journal of the American Chemical Society
ts1-ReRe-cis is the most stable transition state located for the
site binding coordination mode) resulted quite unstable, lying
about 10 kcal mol-1 above m0-SiSi-cis regardless of their con-
figuration. Also, they originate very high reaction pathways.
For these reasons they will not be considered in the following
discussion.
first reaction step in which the new C1-C2 bond is forming
(the C1-C2 distance is 2.05 Å). A schematic representation of
ts1-ReRe-cis is given in Figure 5. The structural changes oc-
curring in this step are clearly evidenced by the values of the
geometrical parameters: since C2-C3 is turning into a single
bond, its lengths is greater (1.46 Å) with respect to reactants.
According to the formation of a Au1-C3 σ bond, the asym-
metry featuring the interaction between Au1 and the C2 and
C3 atoms becomes more pronounced (the Au1-C3 and
Au1-C2 distances are 2.17 and 2.90 Å, respectively). Further-
more, a strengthening of the hydrogen-bond between the in-
dolic hydrogen H1 and the triflate oxygen O1 is observed
(H1…O1 = 1.78 Å; it was 1.92 Å in m0-ReRe-cis) as a conse-
quence of the negative charge displacement from the indole
ring toward C2, which increases the H1 acidity. The computed
energy barrier for ts1-ReRe-cis is rather low (only 9.5 kcal
mol-1). This can be explained by the π-acidity of Au1, which
enhances the electrophilicity of C2 and the pronounced “fold-
ing effect” of the triflate ion forcing C1 and C2 to the best
relative position for the attack.
1
2
3
4
5
6
7
8
The preliminary computational studies pointed to the above
mentioned “folding effect” (played by the triflate counterion as
shown in m0-SiSi-cis and m0-ReRe-cis), favoring the best
relative orientation of the C1 and C2 atoms to form the new
C-C bond. This effect should directly involve only one gold
atom of the binuclear catalyst in the electrophilic activation of
the olefin bond (single-site coordination).
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Seeking for experimental evidence to support this hypothesis,
we investigated modified systems characterized by the impos-
sibility of establishing the above mentioned network of
H-bonds responsible for the “folding effect”. Following our
mechanistic model, this impossibility should cause a decrease
of chemical yield and stereoinduction. Thus, indole-alcohols
carrying either the N(1)-substituted indole rings ((Z)-1g,h) or
the OMe group in place of the allyl OH (1i) were prepared and
reacted under optimal conditions (Scheme 3). Interestingly,
while no conversion was observed for N(1)-Me (1g) and
methylether 1i, N(1)-allyl compound 1h yielded 2h in 22%
yield but in racemic form. All these experimental evidences
support the idea that hydrogen interactions (as shown in Fig-
ure 3) play an essential role both in the rate limiting and ste-
reo-differentiating steps of the reaction (vide infra for mecha-
nistic details).
Unlike what found for the starting reactants for which the
three most stable isomers are rather close in energy (within a
range of 3.1 kcal mol-1), ts1-ReRe-cis turns out to be highly
favored over the other possible transition structures originating
from the various cis starting adducts. To have an exhaustive
picture of the energetics of the various reaction pathways, we
have reported in Figure 4A the energy levels of all possible
transitions states (originating from the different isomeric m0
species) and the resulting m1 intermediates. In particular, the
transition state ts1-SiSi-cis stemming from the most stable
isomer m0-SiSi-cis, lies 4.6 kcal mol-1 above ts1-ReRe-cis and
is characterized by a much larger intrinsic energy barrier (16.5
kcal mol-1). Thus, it cannot compete with ts1-ReRe-cis. It is
interesting to note that the two hydrogen-bonds involving the
triflate fragment i.e. H1…O1 and H2…O2, are weaker in ts1-
SiSi-cis with respect to ts1-ReRe-cis (the corresponding dis-
tances are 1.82 Å and 1.89 Å, respectively). Higher energy
values feature the two additional cis transition states ts1-ReSi-
cis and ts1-SiRe-cis.
SCHEME 3. Denying the establishment of hydrogen interac-
tions via the acid N1(H) and OH protons. Both reaction rate
and stereoselection are markedly eroded.
The only reaction channel which, in principle, could compete
with m0-ReRe-cis ꢀ ts1-ReRe-cis is that associated with
ts1-ReRe-trans, whose structural features are similar to those
of ts1-ReRe-cis except for the reversed configuration of the
C2-C3 double bond (a two-dimensional representation of
ts1-ReRe-trans is given in Figure S1). It is interesting to ex-
amine in this case the mechanism of the “folding effect” ex-
erted by the triflate ion. Because of the trans configuration of
the C2-C3 double bond, the triflate oxygen O2 forms a strong
H-bond with the allylic hydroxyl (O3H2…O2 = 1.77 Å), while
the triflate O1 is too far away to interact simultaneously with
the indole NH1 bond, as observed in the cis structure. Howev-
er, since the H2…O2 interaction makes O3 basic enough to es-
tablish a hydrogen bond with the indole ring (NH1…O2 = 1.81
Å), the triflate “folding effect” is still effective.
The knowledge of the optimized reagent structures, together
with the above described experimental results, has made pos-
sible to envisage a reasonable mechanistic scheme. Because of
the single-site coordination of the cationic gold complex and
relative low oxophilicity, the formation of allylic carbocation
intermediates (required by a SN1-type mechanism) seems un-
likely. A SN2’ mechanism represents a more feasible hypothe-
sis, albeit ambiguity between concerted and step-wise profiles
remains unsolved.
First reaction step: outer-sphere carboauration of the C-C
double bond. In the first reaction step the indole C1 carbon
attacks the C2 olefin carbon, affording a new C1-C2 bond and
a six-membered cycle. This event is triggered by the electro-
philic activation of the olefin played by the gold catalyst, with
some important consequences: (i) the loss of aromaticity of the
pyrrole ring; (ii) the formation of a formally covalent σ Au1-
C3 bond; (iii) the transfer of the formal positive charge from
Au1 to C5; (iv) the formation of three new stereocenters (C1,
C2 and C3).
5
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 6 of 13
Figure 4A). In m1 the pyrrole ring has lost its aromatic charac-
1
2
3
4
5
6
7
8
ter and bears now a positive charge formally located on C5.
Although a new σ C-C bond has been formed, the loss of aro-
matic character determines an overall destabilization of m1
with respect to the starting complex. Three new stereocenters
are generated in m1: C1, C2 and C3. The configuration at C1
is crucial because it determines the position of the H3 proton
that will be abstracted in the subsequent re-aromatization step.
Depending on its position (corresponding to the existence of
two possible stereoisomers), H3 can be facing the allylic hy-
droxyl (O3H2) or the triflate ion. Importantly, C2 is the only
stereocenter that will be maintained during the whole reaction
course: thus, the first m0 ꢀ m1 step determines the stereo-
chemical outcome of the process.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The energetic order of all possible m1 isomers (Figure 4A)
replicates that of the corresponding transition states. The most
stable m1 structure (depicted in Figure 5) is m1-ReRe-cis
which originates from the most stable transition state ts1-
ReRe-cis. As one can expect on the basis of the Hammond
postulate, being the first reaction step endothermic, ts1 and
m1 have very similar structural features. Inspection of the m1-
ReRe-cis structure shows that the six-membered ring is com-
plete, even if the C1-C2 bond is slightly longer (1.60 Å) than a
normal C-C σ bond. The C2…Au1 interaction has become neg-
ligible (C2-Au1 = 3.05 Å) and Au1 is definitely σ-bonded to
C3 (Au-C3 = 2.13 Å). Interestingly, in addition to the persist-
ing interaction between the triflate group, the indole ring and
the allylic hydroxyl O3H2, a strong hydrogen bond of this
group with H3 can be observed (O3…H3(C1) = 2.11 Å). This
interaction anticipates the hydrogen abstraction occurring in
the next step.
Figure 5. Two-dimensional representations of the critical points
ts1, m1, ts2 and m2 located along the reaction path originating
from m0-ReRe-cis (bond lengths are in Ångstroms).
The next two most stable m1 isomers are m1-ReRe-trans and
m1-SiSi-cis. These minima originate from ts1-ReRe-trans and
ts1-SiSi-cis, respectively i.e. the second and third most stable
transition structures. Every other m1 structure is much higher
in energy, confirming that the corresponding reaction channels
cannot be competitive. Inspection of the structural features of
these intermediates suggests that their instability could be due
to unfavorable steric interactions between the phenyl rings of
the phosphine ligands and the groups directly involved in the
reaction (triflate and allylic hydroxyl).
A comparison between the cis and trans structures shows that
in both cases the most stable starting adducts (m0-SiSi-cis and
m0-ReSi-trans) do not originate the most likely reaction paths
(in particular, for the trans structure the intrinsic activation
barrier for m0-ReSi-trans is very high i.e. 23.1 kcal mol-1).
Thus, for both cis and trans configuration the ”active” species
does not corresponds to the most populated one and an equi-
librium must exist between this conformation and the actually
“active” conformation. Since the energy difference between
these two conformational isomers is lower in the cis case (2.4
kcal mol-1) than in trans case (3.8 kcal mol-1, see Figure S1)
and the intrinsic activation barrier is almost identical in the
two cases (9.5 and 9.4 kcal mol-1), the simpler model-system
seems to suggest that both cis and trans adducts should be re-
active, but that the former should react faster than the latter.
Since all remaining trans transition structures are much higher
in energy, the corresponding reaction channels can be imme-
diately discarded.
Second step: proton transfer and rearomatization. In the
second step (transition state ts2) a proton (H3) is transferred
from C1 to the hydroxylic O3 oxygen, affording the interme-
diate m2 where the indole aromatic character is re-established.
The intrinsic energy barrier for the transformation
m1-ReRe-cis ꢀ ts2 ꢀ m2 is 7.7 kcal mol-1. The proton trans-
fer, favored by the relative position of C1 and O3 which are
facing each other in the m1-ReRe-cis complex, is assisted by
the negatively charged triflate ion which, acting as a Lewis
base, enhances the basicity of the O3 oxygen through a strong
H2…O2 hydrogen-bond (H2…O2 = 1.65 Å). The migrating
proton is approximately half-way between C1 and O3 (C1-H3
and O3-H3 ≈ 1.32 Å): this distance is rather large but avoids
destabilizing geometrical distortions in the six-member ring-
like structure C1-C2-C3-C4-O3-H3.
Before considering a more complex and realistic model-
system (MODEL 2), to obtain an exhaustive mechanistic sce-
nario we have completed the investigation of the reaction
mechanism for MODEL 1 following the favored reaction
channel originating from m0-ReRe-cis. A schematic represen-
tation of intermediates and transition states that feature our
proposed mechanism is reported in Figures 5 and 6, while the
corresponding energy profile is depicted in Figure 4B.
Since ts2 is 2.0 kcal mol-1 higher than ts1, it is responsible for
the overall reaction-rate (rate-determining step of the process).
The formation of the C1-C2 bond is completed in the m1 in-
termediates that are characterized by the newly formed tetra-
hydrocarbazolyl skeleton (the m1 energy levels are reported in
6
ACS Paragon Plus Environment

Page 7 of 13
Journal of the American Chemical Society
Thus, before discarding the reaction path originating from the 2.00 Å) while that involving the indole ring becomes weaker
1
2
3
4
5
6
7
most stable isomer m0-SiSi-cis, it is important to check even
in that case the energy of ts2. We found that this transition
state along the SiSi-cis pathway lies 5.2 kcal mol-1 higher.
Thus, the SiSi-cis path is very unlikely and can be definitely
ruled out. This result clearly indicates that the favored path is
ReRe-cis, which leads to configuration R at C2 in agreement
with the experimental evidence.
(H1-O1 = 2.04 Å).
The resulting m2 intermediate is characterized by the restored
aromatic indole system and the presence of two stereocenters
C2 and C3, while C1, after deprotonation, is now an sp2 hy-
bridized carbon. The positively charged water molecule is
bonded to C4 (O3-C4 distance = 1.55 Å) and forms two quite
strong hydrogen bonds: one with the indole nitrogen (N…H3 =
1.73 Å) and the other (even stronger) with one of the triflate
oxygen atoms (H2…O2 = 1.47 Å, H2-O3 = 1.05 Å), which out-
lines once again the crucial role of the triflate ion in stabilizing
the positive charge now formally located on the O3 atom. The
aromatic character of the indole ring mainly determines the
high stability of m2 (1.3 kcal mol-1 below m0-ReRe-cis). Cer-
tainly the strong interactions involving the triflate ion provide
non-negligible contributions to this stability.
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Third step: deauration process. In the following third step
(m2 ꢀ ts3 ꢀ m3, see Figure 6) we observe the cleavage of
the O3-C4 bond and the consequent release of the water mole-
cule (C4…O3 distance = 1.83 Å in ts3). The triflate counterion
was found to assist the displacement of the water molecule by
the persisting strong H2…O2 hydrogen bond (H2…O2 distance
= 1.61 Å in ts3). This transformation is very fast, (the intrinsic
activation barrier being only 1.6 kcal mol-1) and leads to a very
stable structure m3 (14.3 kcal mol-1 below the initial reac-
tants). In m3 the C3-C4 double bond is almost re-established
(C3-C4 = 1.38 Å) and the gold atom Au1, where the positive
charge is now formally located, is again characterized by an η2
interaction with the newly formed C3-C4 double bond (C3-
Au1 = 2.31 Å and C4-Au1 = 2.40 Å). After displacement, the
water molecule is now far away from the C3-C4 moiety
(C4…O3 distance = 3.30 Å), but maintains a strong interaction
with the negatively charged triflate ion (H2…O2 distance =
1.82 Å).
Figure 6. Two-dimensional representations of the critical points
ts3, m3, ts4 and m4 located along the reaction path originating
from m0-ReRe-cis (bond lengths are in Ångstroms).
Alternative mechanism: carboauration at the C3 position
of the indole. In principle, the initial nucleophilic attack of the
adjacent C3 indole carbon atom (C5 in our numbering
scheme) on the [Au(I)]-activated olefin should also be consid-
ered. As a matter of fact, even C3-substitued indoles have
been found to react regioselectively at the more nucleophilic
C3 position.14b,24,25
Fourth step: restoring of the catalytic species. In this last
step (m3 ꢀ ts4 ꢀ m4) the bimetallic gold catalyst is restored.
Since both ts4 and m4 (see Figure 6) are below the starting
reactants (-3.5 and -8.8 kcal mol-1, respectively) and the intrin-
sic activation barrier is 10.8 kcal mol-1, this step must be rather
fast indicating an effective catalytic cycle. The mechanism
bringing back the catalytic species outlines once again the cru-
cial role of the triflate counterion. In ts4, the gold atom Au1
moves from the C3-C4 double bond toward the triflate group:
the Au1…O4 distance significantly decreases (2.33 Å) while
the distances between the gold atom and C3 and C4 increase
(Au1…C3 = 2.78 Å and Au1…C4 = 2.96 Å). Thus, the triflate
ion gradually replaces the C3-C4 double bond as a ligand. In
the final resulting complex m4 the bimetallic catalytic com-
plex is definitely re-established (Au1…C3 = 6.32 Å and
Au1…C4 = 6.00 Å) and the product is released. The hydrogen-
bond of the water molecule to triflate is maintained (O2-H2 =
Scheme 4. Exploring the initial C(3)-nucleophilic attack of the
indole core on the allylic alcohol group.
Following this hypothesis (pictorially represented in Scheme
4), the initial carbonauration of the C-C double bond via the
nucleophilic attack of the carbon C5 was also investigated.
This reaction path would lead to the formation of a spiro-
compound featuring the new C2-C5 and C3-Au1 bonds. Even
in this case, either the Re or Si face of carbon C5 can be in-
7
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 8 of 13
volved in the nucleophilic attack. However, since for a given ure 7) show that all the starting reactants, with the exception of
1
2
3
4
5
6
7
8
orientation of the indole ring the C5 Re face corresponds to the
C1 Re face, we can directly compare the transition structures
leading to the 5- and 6-member cycles. We have located four
possible transition states (denoted as ts1’) for this alternative
nucleophilic attack. The corresponding energy values relative
to the most stable transition state yielding the 6-member cycle
(ts1-ReRe-cis), are reported in Table 2. Interestingly, even if
the intrinsic activation barriers are of the same order of magni-
tude as those previously computed, all these new structures are
significantly higher in energy than the transition states leading
to the 6-member ring. An additional calculation shows that the
5-member cycle intermediate m1’ (originating from ts1’-
ReRe-cis), lies 5.6 kcal mol-1 above the most stable 6-member
cycle intermediate m1-ReRe-cis. Hence, these alternative re-
action channels cannot effectively compete with those leading
to the formation of the 6-member cycle.
m0-ReRe-trans, are very close in energy. This is probably due
to the quite large distance between the indole group and the
sterically hindered phosphines so that different steric patterns
cannot lead to different stabilities. The situation changes in the
transition state structures, where the indole approaches the C2-
C3 double bond linked to Au1 and consequently the cumber-
some ligand side-arms.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
TABLE 2. Energies (kcal mol-1) of the transition states ts1’
affording a 5-member cycle. The values are relative to the
most stable transition state (ts1-ReRe-cis) affording the
6-member cycle. Intrinsic activation barriers are reported in
parenthesis.
ts1’ transition states
C2 Re
C2-C3 configuration
cis
trans
ts1’-SiRe
ts1’-ReRe
C2 Si
13.1 (17.1)
2.9 (12.4)
cis
11.8 (18.1)
/
trans
/
ts1’-SiSi
7.5 (17.0)
Figure 7. (A) Energies of the cis (left side) and trans (right side)
starting adducts m0 in the various conformations, the correspond-
ing transition states ts1 and the resulting intermediates m1 ob-
tained for MODEL 2. Values in parenthesis are relative to
m0-ReRe-cis. (B) Energy profile for the reaction path originating
from m0-ReRe- cis. Energies in kcal mol-1.
The large model system: MODEL 2. The results obtained for
MODEL 1 indicate that the two most likely reaction channels
are those originating from m0-ReRe-cis and m0-ReRe-trans,
that involve the same indole and alkene faces but are associat-
ed with opposite configurations of the C-C double bond. This
result correctly predicts the experimental enantioselectivity for
the cis isomer. However, even if the reaction on the trans spe-
cies is predicted to be slower (as outlined in the previous sec-
tion), this finding does not explain the total inability (experi-
mentally observed) of the trans-alcohol (E)-1a to afford the
ring closure. This discrepancy between theory and experiment
suggested that MODEL 1 could be an oversimplification of
the real system and, thus, was unable to describe properly the
stereo-electronic environment generated by the chiral ligand
DTBM-MeO-biphep.
In the case of the cis configuration the favored path is again
that associated with the most stable m0-ReRe-cis species (ac-
tivation energy of 13.0 kcal mol-1). Even if the other cis spe-
cies are rather close in energy, the corresponding activation
barriers are significantly higher suggesting that the relating
pathways can be ruled out.
To elucidate this point i.e. how much likely are the reaction
pathways originating from the other m0 cis structures (in par-
ticular that associated with the next most stable transition state
ts1-SiSi), we have computed the energy of m1-ReRe, ts2 and
m2 originating from m0-ReRe-cis and that of the transition
state ts1-SiSi-cis and intermediate m1-SiSi-cis (originating
from m0-SiSi-cis). The reaction profile m0-ReRe-cis ꢀ m2 is
represented in the lower part of Figure 7. Two important
points are evidenced by these computations: (i) since ts2 is
16.4 kcal mol-1 higher than the starting reactants, even for
MODEL 2, it represents the rate-determining step of the pro-
cess; (ii) m1-SiSi (the intermediate along the SiSi-cis path)
lies 8.1 kcal mol-1 above m1-ReRe and 1.3 above ts2 located
To evaluate the effect of the phenyl ring substituents on the
reaction profile, we have recomputed the most important criti-
cal points using the larger model-system MODEL 2 that coin-
cides with the real system. In this case, the four phenyl rings
(two for each phosphine group), that are not bridging the
phosphorous atoms bear a tert-butyl group in the two meta-
positions and one OMe group in the para-position. Obviously
these four cumbersome rings increase the phosphine steric
hindrance: in particular, since at least one phenyl ring is close
to the reaction center, in principle they could significantly af-
fect the relative stability of the various critical points. The re-
sults of the computations on MODEL 2 (summarized in Fig-
8
ACS Paragon Plus Environment

Page 9 of 13
Journal of the American Chemical Society
on the ReRe-cis path. Thus, the SiSi-cis path (and the two re- The mechanism of the enantioselective gold(I)-catalyzed al-
1
2
3
4
5
6
7
8
maining cis paths) can be definitely discarded.
lylic alkylation of indoles with alcohols has been investigated
using a combined experimental-computational DFT approach.
To this purpose two model-systems MODEL 1 and MODEL
2, differing in the structure of the gold complex L1(AuOTf)2,
have been used. In the former case the Ar groups not bridging
the two P atoms are simple benzene rings while in latter case
(corresponding to the real adduct used in the experiment) Ar =
3,5-(t-Bu)2-4-OMe-Ph.
To understand why the ReRe-cis path (leading to configura-
tion R at C2) is at a lower energy with respect to the SiSi-cis
path, it is useful to compare the structures of ts1-ReRe-cis and
ts1-SiSi-cis. When the C2 Si face (in place of the Re face) is
involved in the attack, rotations around the C3-C4 and C4-O3
bonds are necessary to activate the H-bond between O3-H2
and the adjacent triflate oxygen. However, this forces one t-Bu
group bonded to the P1 non-bridging phenyl ring to move too
much closer to the reaction site and the C4 methylene group,
thus increasing the steric hindrance and the energy of ts1-SiSi-
cis. The structural differences between ts1-ReRe-cis and ts1-
SiSi-cis are evidenced in Figures S2 and S3 of the Supplemen-
tary Information file.
9
The results obtained with the two model-systems can be sum-
marized as follows:
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(i) The fully explored reaction surface for MODEL 1 reveals
that the favoured mechanism is a step-wise SN2’ type mecha-
nism based on indole-auration of the C-C double bond (outer-
sphere like), re-aromatization of the indole moiety and subse-
quent β-elimination of [Au(I)]-H2O.
It is interesting to note that the reaction profile obtained with
the more realistic chiral ligand does not change notably with
respect to that computed for the smaller model. As found for
MODEL 1, m2 is strongly stabilized by the restored aromatic
character of the indole moiety and is only 1.5 kcal mol-1 above
m0-ReRe-cis. A larger activation barrier (13.0 kcal mol-1 in-
stead of 9.5 kcal mol-1) has been found for the first step m0 ꢀ
m1. This is probably due to the more cumbersome phosphine
groups, which come closer to the substrate as the reaction pro-
ceeds. For the same reason, the first intermediate m1-ReRe
lies 9.6 kcal mol-1 above the initial reactants (this value was
3.8 kcal mol-1 when the smaller model system was employed).
(ii) The triflate gold counterion has been demonstrated to play
a pivotal role several times along the reaction coordinate. First
of all this ion exerts a sort of “folding effect” which forces the
two reactive sites of the starting adduct to move closer adopt-
ing the right orientation to react (U-turn-type geometry). This
is achieved by means of two rather strong H-bonds involving
the indole NH bond, the allylic hydroxyl group and two triflate
oxygen atoms. This hypothesis has been confirmed by addi-
tional experimental results on modified reacting systems
where the above described H-bonds cannot be established. The
negatively charged triflate ion plays an additional important
role in “assisting“ the proton transfer from C1 (C(2) indole
position) to the allylic hydroxyl oxygen O3 and, thus, restor-
ing the aromatic character of the pyrrolyl ring (second step of
the process, transition state ts2). The triflate ion strongly inter-
acts with the hydroxyl hydrogen H2 (H2…O2 = 1.65 Å) and
enhances the basicity of the hydroxyl oxygen. The role of the
triflate counterion is once again crucial in the third reaction
step (ts3) where the release of the water molecule (needed to
restore a C-C double bond) is helped by the persisting strong
hydrogen bond between this ion (oxygen O2) and the water
hydrogen H2 (H2…O2 distance = 1.61 Å in ts3). All these as-
pects contribute to make this transformation very fast.
In the case of the trans configuration we have computed the
two transition states ts1-SiRe-trans (associated with the most
stable trans m0 isomer) and ts1-ReRe-trans, since for
MODEL 1 we found that the energetics of this path was simi-
lar to that of the path originating from m0-ReRe-cis. In both
cases the two transition states are very high in energy (i.e. 66.9
and 60.1 kcal mol-1 above m0-ReRe-cis, respectively) and
large activation barriers feature the corresponding reaction
channels (66.1 and 51.9 kcal mol-1, respectively). This compu-
tational finding accounts for the negligible reactivity experi-
mentally recorded for the trans species. Inspection of the
structural features of ts1-ReRe-cis and ts1-ReRe-trans shows
that the key-factor responsible for the high energy of the latter
transition state is again the steric hindrance caused by two t-
Bu groups bonded to the P1 non-bridging phenyl rings and
pointing towards the reaction site: in ts1-ReRe-trans these
groups are forced to move rather close to the double bond and
the methylene unit (carbon C4) bearing the hydroxyl group.
(iii) The low energy barrier that features the transition state
corresponding to the formation of the new carbon-carbon bond
(ts1) can be explained by the electrophilic character of the
C2-C3 double bond, enhanced by the interaction with the gold
atom and by the pronounced “folding effect” found in the tran-
sition structure.
In summary, the calculations on MODEL 2 show that, even if
the smaller system provided precious information on the gen-
eral mechanism of this process (in particular the role played by
the gold atoms and triflate ions), a complete model-system in-
cluding the real cumbersome ligand L1, is essential to explain
the reagent-controlled stereochemistry of these gold-promoted
enantioselective transformations.
(iv) Being ts2 the highest in energy transition state, the corre-
sponding step (where the aromatic character of the indole ring
is restored) determines the overall kinetics and is the rate-
determining step of the process.
Finally, a validation of the accuracy of our computational
treatment in predicting the overall stereochemical profile of
the process was ascertained by determining the absolute con-
figuration of 1-vinyl-tetrahydrocarbazoles 2 (which was found
to be R), via comparison of the optical rotation values of (R)-
2a and (R)-2f with the known compounds in literature.14b
(v) The new stereocenter at C2 is formed in the first reaction
step and is maintained along the whole reaction path. The ab-
solute configuration of this carbon in the final products is R in
agreement with the experimental finding.
(vi) Even if the simpler model provides a complete mechanis-
tic scenario for these reactions, it does not correctly predict the
CONCLUSIONS
9
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 10 of 13
sviluppo di sistemi catalitici innovativi” (MIUR, Rome) and Alma
Mater Studiorum - Università di Bologna.
influence of the stereochemistry of the reaction partners and
the dichotomy in reactivity of the two diastereomeric Z and E
acyclic precursors 1a. To explain this aspect it is essential to
use a larger model corresponding to the real system used in the
experiment and including the cumbersome t-Bu groups bond-
ed to the non-bridging phenyl groups of the ligand. When the
C2-C3 double bond has a trans configuration (ts1-ReRe-
trans) two of these groups (pointing toward the reaction site)
are forced too much close to the double bond and the meth-
ylene unit (carbon C4) bearing the hydroxyl group, thus caus-
ing a destabilizing steric hindrance. The steric hindrance
caused by the cumbersome tBu groups is also responsible for
lower energy of the ReRe-cis path (leading to configuration R
at C2) compared to the SiSi-cis path and, thus, for the e.e. ex-
perimentally observed.
1
2
3
4
5
6
7
8
ABBREVIATIONS
dppm: 1,1-bis(diphenylphosphino)methane; DTBM: 3,5-tBu2-4-
OMe; ee: enantiomeric excess; LA: Lewis acid; MeObiphep: 2,2'-
bis(diphenylphosphino)-6,6'-dimethoxy-1,1'-biphenyl; nr: no
reaction; NTf2: bis(trifluoromethylsulfonyl)amide; OPBN: p-
nitrobenzoate; OTf: trifluoromethanesulfonate; OTs: p-
toluenesulfonate; THBCs: tetrahydro-β-carbolines.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
REFERENCES
(1) (a) R. A. Widenhoefer, Chem. Eur. J. 2008, 14, 5382-5391. (b) D.
J. Gorin, B. D. Sherry, F. D. Toste, Chem. Rev. 2008, 108, 3351-
3378. (c) N. Bongers, N. Krause, Angew. Chem. Int. Ed. 2008, 47
2178-2181. (d) S. Sengupta, X. Shi, Chem. Cat. Chem. 2010,
,
,
2
609-619. (e) A. S. K. Hashmi, C. Hubbert, Angew. Chem. Int. Ed.
2010, 49 1010-1012. (f) A. Pradal, P. Y. Toullec, V. Michelet,
,
(vii) The nucleophilic attack of the C3 indole carbon atom (C5
in our numbering scheme) on the [Au(I)]-activated olefin
(formation of a 5-member cyclic intermediate) seems disfa-
vored with respect to the attack of the C2 indole carbon (C1 in
our model-system).
Synthesis 2011, 1501-1514.
(2) For general reviews on gold-catalyzed organic transformations
see: (a) A. S. K. Hashmi, Angew. Chem. In. Ed. 2005, 44
6993. (b) A. S. K. Hashmi, G. J. Hutchings, Angew. Chem. Int. Ed
2006, 45 7896-7936. (c) D. J. Gorin, F. D. Toste, Nature 2007,
446 395-403. (d) A. S. K. Hashmi, Chem. Rev. 2007, 107 3180-
3211. (e) A. Fürstner, O. D. Davies, Angew. Chem. Int. Ed. 2007,
, 6990-
.
,
,
,
46
,
4310-4317. (f) Z. Li, C. Brouwer, C. He, Chem. Rev. 2008,
3239-3265. (g) E. Jiménez-Núñez, A. M. Echavarren, Chem.
108
,
Rev. 2008, 108, 3326-3350. (h) H. C. Shen, Tetrahedron 2008, 64
,
COMPUTATIONAL DETAILS
3885-3903. (i) R. Skouta, C.-J. Li, Tetrahedron 2008, 64 4917-
,
4938. (j) A. Arcadi, Chem. Rev. 2008, 108, 3266-3325. (k) P. Bel-
mont, E. Parker, Eur. J. Org. Chem. 2009, 6075-6089. (l) N. D.
All the reported DFT computations have been carried out with
the Gaussian 0926 series of programs using the B3LYP27 func-
tional that has been demonstrated to provide a reliable descrip-
tion of systems involving metal atoms and hydrogen bond in-
teractions.28 According to a locally dense basis set (LDBS)29
approach, the model-system has been partitioned into different
regions, which were assigned basis sets of different accuracy.
For MODEL 1 all atoms, except the gold atoms, have been
described by the DZVP basis,30 which is a Local Spin Density
(LSD)-optimized basis set of double-zeta quality including
polarization functions. For the two metal atoms we used the
LANL2DZ basis set.31 For MODEL 2 we have maintained the
same accuracy level of MODEL 1 except for the methyl
groups added in the larger model. Thus, the carbon and hydro-
gen atoms of the methyl belonging to t-Bu and OMe have
been described by the simple STO-3G basis, but we have used
the DZVP basis for oxygen in OMe and the central carbon in
t-Bu. We have performed frequency computations to deter-
mine the nature of the various critical points.
Shapiro, F. D. Toste, Synlett
Leyva-Pérez, M. J. Sabater, Chem. Rev. 2011, 111, 1657-1712. (n)
T. de Haro, C. Nevado, Synthesis 2011, 2530-2539. (p) M. Ban-
, 2010, 675-691. (m) A. Corma, A.
,
dini, Chem. Soc. Rev. 2011, 40, 1358-1367. (q) T. C. Boorman, I.
Larrosa, Chem. Soc. Rev. 2011, 40, 1910-1925.
(3) L.-I. Rodríguez, T. Roth, J. Lloret Fillol, H. Wadepohl, L. H.
Gade, Chem. Eur. J. 2012, 18, 3721-3728.
(4) For
Tkatchouk, Y. Wang, W. A. Goddard, F. D. Toste, Nat. Chem. 2009,
482. (b) A. S. K. Hashmi, Angew. Chem. Int. Ed. 2010, 49,
a general review see: (a) D. Benitez, N. D. Shapiro, E.
1
,
5232-5241. For a selection of articles see: (c) G. Seidel, R. My-
nott, A. Fürstner, Angew. Chem. Int. Ed. 2009, 48, 2510-2513. (d)
D. Weber, M. A. Tarselli, M. R. Gagné, Angew. Chem. Int. Ed.
2009, 48
2009, 11, 2237-2240. (f) Z. J. Wang, D. Benitez, E. Tkatchouk, W.
A. Goddard, F. D. Toste, J. Am. Chem. Soc. 2010, 132 13064-
, 5733-5736. (e) R. S. Paton, F. Maseras, Org. Lett.
,
13071. (g) R. L. LaLonde, W. E. Brenzovich, D. Benitez, E.
Tkatchouk, K. Kelly, W. A. Goddard, F. D. Toste, Chem. Sci. 2010,
1
,
226-233. (h) P. Pérez-Galán, E. Herrero-Gómez, D. T. Hog, N.
J. A. Martin, F. Maseras, A. M. Echavarren, Chem. Sci. 2011,
141-149. (i) P. Nun, S. Gaillard, A. Poater, L. Cavallo, S. P.
Nolan, Org. Biomol. Chem. 2011, , 101-104. (j) A. S. K. Hashmi,
M. Pernpointner, M. M. Hansmann, Faraday Discussion 2011, 152
179-184. (k) R. Döpp, C. Lothschütz, T. Wurm, M. Pernpointner, S.
Keller, F. Rominger, A. S. K. Hashmi, Organometallics, 2011, 30
2
,
9
,
,
,
5894-8903. (l) E. L. Noey, X. Wang, K. N. Houk, J. Org. Chem.
2011, 76, 3477-3483. (m) O. A. Egorova, H. Seo, Y. Kim, D. Moon, Y. M.
Rhee, K. H. Ahn, Angew. Chem. Int. Ed. 2011, 50, 11446-11450. (n) E. L. No-
ASSOCIATED CONTENT
Supporting Information. Experimental methods, compound
characterizations and detailed computational results. This material
is available free of charge via the Internet at http://pubs.acs.org.
ey, Y. Luo, L. Zhang, K. N. Houk, J.Am. Chem. Soc. 2012, 134
1078-1084 (p) J. E. Heckler, M. Zeller, A. D. Hunter, T. G.
Gray, Angew. Chem. Int. Ed. 2012, 51, 5924-5928.
,
.
(5) Chiral monometallic gold complexes are also emerging as valuable
alternative see: 4a, (a) H. Teller, S. Flügge, R. Goddard, A.
AUTHOR INFORMATION
Fürstner, Angew. Chem. Int. Ed. 2010, 49
, 1949-1953. (b) H.
Corresponding Authors
Teller, A. Fürstner, Chem. Eur. J. 2011, 17, 7764-7767.
(6) (a) H. Schmidbaur, A. Schier, Z. Naturforsch. 2011, 66b, 329-350.
(b) D. Wang, R. Cai, S. Sharma, J. Jirak, S. K. Thummanapelli, N.
G. Akhmedov, H. Zhang, X. Liu, J. L. Petersen, X. Shi, J. Am.
Chem. Soc. 2011, 134, 9012-9019.
*E-mail: gianpiero.miscione@unibo.it (G.M.),
marco.bandini@unibo.it (M.B.).
(7) A. L. Pérez, A. Corma, Angew. Chem. Int. Ed. 2012, 51, 614-635.
(8) (a) M. J. Johansson, D. J. Gorin, S. T. Staben, F. D. Toste, J.
Am. Chem. Soc. 2005, 127, 18002-18003. (b) G. L. Hamilton, E. J.
Kang, M. Mba, F. D. Toste, Scienze, 2007, 317, 496-499.
ACKNOWLEDMENTS
Acknowledgment is made to Progetto FIRB-2008 “Futuro in
Ricerca” Innovative sustainable synthetic methodologies for C-H
activation processes, Progetto PRIN-2009 “Progettazione e
(9) D. Zuccaccia, L. Belpassi, F. Tarantelli, A. Macchioni, J. Am.
Chem. Soc. 2009, 131, 3170-3171, and references therein.
10
ACS Paragon Plus Environment

Page 11 of 13
Journal of the American Chemical Society
(10) (a) M. Bandini, A. Eichholzer, P. Kotrusz, M. Tragni, S. Troisi,
A. Umani-Ronchi, Adv. Synth. Catal 2009, 351 319-324. (b) M.
Bandini, M. Tragni, A. Umani-Ronchi, Adv. Synth. Catal. 2009,
351 2521-2524. (c) M. Bandini, A. Eichholzer, A. Gualandi, T.
Quinto, D. Savoia, Chem. Cat. Com. 2010, 661-665. (d) M.
Chiarucci, M. Locritani, G. Cera, M. Bandini, Beilstein J. Org.
Chem. 2011, 1198-1204. (e) G. Cera, S. Piscitelli, M.
G. Zheng, J. L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R.
.
,
Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao,
H. Nakai, T. Vreven, J. A. Montgomery, Jr., J. E. Peralta, F.
Ogliaro, M. Bearpark, J. J. Heyd, E. Brothers, K. N. Kudin, V. N.
Staroverov, R. Kobayashi, J. Normand, K. Raghavachari, A.
Rendell, J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi, N.
Rega, J. M. Millam, M. Klene, J. E. Knox, J. B. Cross, V. Bakken,
C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev,
A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, R. L.
Martin, K. Morokuma, V. G. Zakrzewski, G. A. Voth, P. Salvador,
J. J. Dannenberg, S. Dapprich, A. D. Daniels, O. Farkas, J. B.
Foresman, J. V. Ortiz, J. Cioslowski, D. J. Fox, Gaussian, Inc.,
Wallingford CT, 2009.
1
2
3
4
5
6
7
8
9
,
2
,
7
,
Chiarucci, G. Fabrizi, A. Goggiamani, R. S. Ramón, S. P. Nolan,
M. Bandini, Angew. Chem. Int. Ed. 2012, 51, 9891-9895.
(11) For a review see: (a) M. Bandini, Angew. Chem. Int. Ed. 2011, 50
,
994-995. (b) M. Bandini, G. Cera, M. Chiarucci, Synthesis, 2012,
504-512.
(12) (a) M. Bandini, A. Eichholzer, Angew. Chem. Int. Ed. 2009, 48
,
9533-9537. (b) M. Bandini, A. Gualandi, M. Monari, A. Romaniello,
D. Savoia, M. Tragni, J. Organomet. Chem. 2011, 696, 338-347. (c)
M. Bandini, M. monari, A. Romaniello, M. Tragni, Chem. Eur. J.
2010, 16, 14272-14277.
(27) (a) C. Lee, W.T. Yang, R.G. Parr, Phys Rev B: Condens Matter
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1988;37
5652
, 785-789. (b) A.D. Becke. J Chem Phys 1993, 98, 5648-
(28) (a) L. Fan, T. Ziegler, J. Am. Chem. Soc. 1992, 114, 10890-10897.
(b) A. Bottoni, A. Perez Higueruelo, G. P. Miscione, J. Am. Chem.
(13) (a) M. Somei, F. Yamada, Nat. Prod. Rep. 2005, 22, 73–103. (b) G.
R. Humphrey, J. T. Kuethe, Chem. Rev
.
2006, 106
,
2875-2911. (c)
Soc. 2002, 124
,
5506-5513. (c) A. Bottoni, C. Z. Lanza, G. P.
1542-1550.
2002,
TB. M. rost, M. K. Brennan, Synthesis, 2009, 3003-3025. (d) A. J.
Miscione, D. Spinelli, J. Am. Chem. Soc. 2004, 126
,
Kochanowska-Karamyan, M. T. Hamann, Chem. Rev. 2010, 110
4497.
,
4489-
(d) R. D. Bach, C. Thorpe, O. Dmitrenko, J. Phys. Chem.
106, 4325-4335.
B
(14) Recent catalytic enantioselective synthesis of tetrahydrocarba-
zoles: (a) X.-Y. Zhu, X.-L. An, C.-F. Li, F.-G. Zhang, Q.-L. Hua,
(29) G. A. DiLabio, D. A. Pratt, J. S. Wright, Chem. Phys. Lett. 1998,
297, 181-186.
J.-R. Chen, W.-J. Xiao, Chem. Cat. Chem. 2011,
3
,
679-683. (b)
(30) N. Godbout, D. R. Salahub, J. Andzelm, E. Wimmer, Can. J. Chem.
1992, 70, 560-571.
Q.-F. Wu, C. Zheng, S.-L. You, Angew. Chem. Int. Ed. 2012, 51,
1680-1683.
(31) (a) P. J. Hay, W. R. Wadt, J. Chem. Phys. 1985, 82, 270-283. (b)
P. J. Hay, W. R. Wadt, J. Chem. Phys. 1985, 82, 284-299. (c) P.
J. Hay, W. R. Wadt, J. Chem. Phys. 1985, 82, 299C.
(15) For some leading examples see: (a) C. Liu, R. A. Widenhoefer,
Org. Lett
. 2007, 9, 1935-1938. (b) Ö. Aksin, N. Krause, Adv.
Synth. Catal. 2008, 350, 1106-1112. (c) A. Aponick, C.-Y. Li, B.
Biannic, Org. Lett. 2008, 10, 669-671. (d) F. Volz, S. H. Wadman,
A. Hoffmann-Röder, N. Krause, Tetrahedron
(e) P. Mukherjee, R. A. Widenhoefer, Org. Lett. 2010, 12
1187. (f) A. Aponick, B. Biannic, Org. Lett. 2011, 13, 1330-1334.
(g) P. Mukherjee, R. A. Widenhoefer, Org. Lett 2011, 13 1334-
,
2009, 65
,
1902-1910.
,
1184-
.
,
1337. (h) S. R. K. Minkler, B. H. Lipshutz, N. Krause, Angew. Chem. Int. Ed.
2011, 50, 7820-7823. (i) P. Mukherjee, R. A. Widenhoefer, Angew. Chem.
Int. Ed. 2012, 51, 1434-1436.
(16) Modern Gold Catalyzed Synthesis, (A. S. K. Hashmi, F. D. Toste, Eds.) Wiley-
VCH, Weinheim, 2012.
(17) B. Biannic, A. Aponick, Eur. J. Org. Chem. 2011, 6821-6824.
(18) (a) M. Georgy, V. Boucard, J.-M. Campagne, J. Am. Chem. Soc.
2005, 127, 14180–14181. (b) M. Georgy, V. Boucard, O. Debleds, C.
Dal Zotto b, J.-M. Campagne, Tetrahedron
(c) O. Debleds, E. Gayon, E. Vrancken, J.-M. Campagne, Beilstein
J. Org. Chem. 2011, , 866-877.
, 2009, 65, 1758-1766.
7
(19) For a recent review: N. T. Patil, Chem. Asian J. 2012,
2194.
7, 2186-
(20) R. L. LaLonde, B. D. Sherry, E. J. Kang, F. D. Toste, J. Am.
Chem. Soc. 2007, 129, 2452-2453.
(21) For representative examples see: (a) K. Aikawa, M. Kojima, K. Mi-
kami, Angew. Chem. Int. Ed. 2009, 48
Lalonde, J. Z. Wang, M. Mba, A. D. Lackner, F. D. Toste, Angew.
Chem. Int. Ed. 2010, 49 598-601. (c) K. Aikawa, M. Kojima, K.
, 6073-6077. (b) R. L.
,
Mikami, Adv. Synth. Catal. 2010, 352, 3131-3135. (d) C. H. Cheon,
O. Kanno, F. D. Toste, J. Am. Chem. Soc. 2010, 133, 13248-13251.
(22) (a) R. M. Beesley, C. K. Ingold, J. F. Thorpe, J. Chem. Soc.
1915, 107, 1080-1106. (b) C. K. Ingold, J. Chem. Soc., 1921, 119
305-329. (c) C. K. Ingold, S. Sako, J. F. Thorpe, J. Chem. Soc.
1922, 1117-1198.
,
,
,
(23) H. Schmidbaur, A. Schier, Chem. Soc. Rev. 2012, 41, 370-412, and
references therein.
(24) For general reviews: (a) E. D. Cox, J. M. Cook, Chem. Rev. 1995,
95
,
1797-1842. (b) M. Bandini, A. Eichholzer, Angew. Chem. Int.
9608-9644. (c) B. M. Trost, M. K. Brennan,
Ed. 2009, 48
,
Synthesis, 2009, 3003-3025.
(25) For a selection of recent examples of stereo- and regioselective
functionalization of C3-substituted indoles see: (a) B. M. Trost,
J. Quancard J. Am. Chem. Soc. 2006, 128
Trost, M. K. Brennan Synthesis, 2009, 3003-3025. (c) Q.-F. Wu, H.
He, W.-B. Liu, S.-L. You, J. Am. Chem. Soc. 2010, 132 11418-
11419. (d) Q. Cai, C. Zheng, J.-W. Zhang, S.-L. You, Angew. Chem.
Int. Ed. 2011, 50 8665-8669. (e) B. Jones, B. Simmons, A.
, 6314-6315. (b) B. M.
,
,
Mastracchio, D. W. C. MacMillan, Nature, 2011, 475, 183-188, and
references therein. (f) G. Cera, M. Chiarucci, A. Mazzanti, M.
Mancinelli, M. Bandini, Org. Lett. 2012, 14, 1350-1353.
(26) Gaussian 09, Revision A.02, M. J. Frisch, G. W. Trucks, H. B.
Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, G.
Scalmani, V. Barone, B. Mennucci, G. A. Petersson, H. Nakatsuji,
M. Caricato, X. Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino,
11
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 12 of 13
1
2
TOC
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
12
ACS Paragon Plus Environment

Page 13 of 13
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
43x14mm (300 x 300 DPI)
ACS Paragon Plus Environment