Annulated N-Heterocycles by Tandem Gold(I)-Catalyzed [3,3]-Rearrangement/Nazarov Reaction of Propargylic Ester Derivatives: An Experimental and Computational Study

Article abstract of DOI:10.1002/ejoc.201500462

The gold(I)-catalyzed tandem rearrangement/Nazarov reaction of propargylic ester derivatives is a useful strategy for the synthesis of cyclopenta-fused N-heterocyclic structures present in many natural compounds. Readily available lactams are converted in

Full text of DOI:10.1002/ejoc.201500462

FULL PAPER
DOI: 10.1002/ejoc.201500462
Annulated N-Heterocycles by Tandem Gold(I)-Catalyzed [3,3]-Rearrangement/
Nazarov Reaction of Propargylic Ester Derivatives: an Experimental and
Computational Study
[a]
[a]
[b]
[b]
´
Martina Petrovic, Dina Scarpi, Béla Fiser, Enrique Gómez-Bengoa, and
Ernesto G. Occhiato*[a]
Keywords: Homogeneous catalysis / Gold / Annulation / Enynes / Nitrogen heterocycles
The gold(I)-catalyzed tandem rearrangement/Nazarov reac-
tion of propargylic ester derivatives is a useful strategy for
the synthesis of cyclopenta-fused N-heterocyclic structures
present in many natural compounds. Readily available
lactams are converted into enol phosphates and triflates and
ion which undergoes a 4π electrocyclization (Nazarov reac-
tion) leading to the target compound in good to excellent
yield. This process has been studied in details both experi-
mentally and computationally to understand the influence of
both the reaction conditions and substrate structural features
coupled to propargyl alcohols under Sonogashira conditions. on the reaction rate and regioselectivity, as well as the tor-
After acetylation, the gold-catalyzed rearrangement of the
enynyl acetates readily occurs when using 3–5 mol-% of a
gold(I) catalyst. The rearrangement generates a divinyl cat-
quoselectivity in the ring closure step. A series of examples
illustrates at the end the scope of the reaction.
Introduction
The widespread presence of the 2-cyclopentenone moiety
in natural products has always been a stimulus for synthetic
organic chemists to find new methods for efficiently build-
ing this structure, with a varying degree of substitution and
control of the stereochemistry.[1] Among the many ap-
proaches to 2-cyclopentenones, the Nazarov reaction ranks
as one of the most important and versatile since the requi-
site 4π electron pentadienyl cation can be generated not
only from classical dienones but also from a steadily in-
creasing variety of unconventional substrates or processes,[2]
including gold-catalyzed transformations.[3,4] Suitably as-
sembled propargylic esters are particularly useful as sub-
strates for the Nazarov reaction, since the transition metal-
catalyzed migration of the carboxylic group to any of the
two unsaturated positions leads to competent pentadienyl
cations. 5-Acyloxy-1,3-enynes 1 in particular (Scheme 1),
undergo – under remarkably mild conditions – a gold(I)-
catalyzed [3,3]-rearrangement to pentadienyl cations 2
which, after Nazarov reaction and eventual protode-
auration, provide acetyloxy-substituted cyclopentadiene
products 4. Hydrolysis (in situ or after work-up) of the lat-
ter leads to the target cyclopentenones 5.[5] This strategy
[a] Dipartimento di Chimica “U. Schiff”, Università degli Studi di
Firenze,
Via della Lastruccia 13, 50019 Sesto Fiorentino (FI), Italy
E-mail: ernesto.occhiato@unifi.it
www.egocch.it/
[b] Departamento de Química Orgánica I, Universidad del País
Vasco/UPV-EHU,
Manuel de Lardizabal 3, 20018 Donostia-San Sebastián, Spain
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejoc.201500462.
Scheme 1. Sequential gold(I)-catalyzed [3,3]-rearrangement/Naza-
rov reaction of 5-acyloxy-1,3-enynes 1 (above) and 7 (below).
Eur. J. Org. Chem. 2015, 3943–3956
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3943

E. G. Occhiato et al.
FULL PAPER
successfully provides cyclopenta-fused carbacycles 6 when thus posing a concern about a possible competition in the
the enyne double bond is embedded in a five-, six- and nucleophilic attack to the activated triple bond. Because the
seven-membered ring (Scheme 1).[5a,5b,6] Our interest in the cyclization leading to intermediate 8 involves a vinylgold
synthesis of cyclopenta-fused heterocycles by the Nazarov species, a further question arises about the conservation of
reaction,[7] as well as in gold-catalysis,[8a] prompted us to the high torquoselectivity we have previously observed in
investigate if the same approach could still furnish annu- Brønsted acid-catalyzed Nazarov reactions, which led to cis
lated systems when embodying the same double bond into disubstituted cyclopenta-fused N-heterocycles.[7a,7b,7d,7g]
N-heterocycles as 7 (Scheme 1).
and which we have exploited for the formal enantioselective
This in fact would represent a new synthetic approach to synthesis of roseophilin.[10f] Finally, to establish this ap-
molecular structures (e.g. [1]pyrindines, annulated pyrrol- proach as a reliable synthetic strategy for cyclopenta-annu-
ines and azepines) present in many natural compounds lated N-heterocycles, the scope of the reaction had to be
(Figure 1),[9] some of which (e.g. cephalotaxine and roseo- assessed on the alkyne chain R1 and the heterocycle ring
philin) already synthesized via some of the Nazarov reac- size and substituents. In this paper we demonstrate that the
tion variants.[10]
tandem gold(I)-catalyzed propargylic rearrangement/Naza-
rov reaction of propargylic esters 7 is a useful strategy for
the synthesis of annulated N-heterocycles but, also, that the
picture for this process is much more complex than that
previously reported for non-heterocyclic systems, with a
whole series of elements [N-protecting group, heterocycle
ring size, gold(I) counterion, etc.] all influencing reaction
rate and selectivity (regio- and stereoselectivity) in one way
or another. To assist us in the comprehension of the experi-
mental results, the energies and structures of the intermedi-
ates and transition states involved in this process were de-
termined by a computational study, the results of which are
well in accordance with the experimental data and which
we present in this paper.
Results and Discussion
Evaluating the Reactivity of Model Compounds
We have recently shown that N-Boc-protected enynes
closely related to 7 (but lacking the acetyloxy group) are
excellent substrates for a gold(I)-catalyzed oxyauration in-
volving the N-Boc carbonyl oxygen which eventually af-
fords exocyclic vinylogous amides.[8a] Thus, in order to
avoid any possible interference by a N-alkoxycarbonyl
group, we decided to find the best reaction conditions for
the sequential [3,3]-rearrangement/Nazarov cyclization with
compound 16a, bearing a N-Tosyl protecting group. The
synthesis of this model compound (Scheme 2) was carried
by converting the N-Ts-substituted δ-valerolactam 13a into
Figure 1. Natural compounds containing a cyclopenta-fused N-
heterocyclic ring.
To accomplish this, however, a few issues have to be dealt the corresponding enol phosphate 14a which was immedi-
with. First, because of the likely contribution of the N atom ately subjected to Sonogashira coupling[11] with 1-heptyn-
in stabilizing positively charged intermediates, whether and 3-ol to give alcohol 15a in 62% after chromatography. This
how the presence of the heteroatom influences the rate and was eventually treated with acetic anhydride to provide
the regiochemical outcome of the reaction had to be as- acetate 16a in 79% yield.[12]
sessed. In our previous studies on the Brønsted acid-cata-
To find the best catalyst, we first screened a series of
lyzed Nazarov reaction of masked and classical dienones,[7] silver and copper salts as sources of the non-coordinating
deprotonation of the oxyallyl cation exclusively led to com- anion with 3 mol-% of Ph3PAuCl in CH2Cl2 at room tem-
pounds of type 12 having the more substituted double bond perature (Table 1).[13] The reactions were monitored by TLC
and not to even traces of its regioisomer 10, in strong con- and, after consumption of 16a, they were left whilst stirring
trast with the outcome of the gold-catalyzed reaction of the (usually 16 h at room temperature) to allow hydrolysis (if
corresponding carbacyclic systems (Scheme 1). Moreover, occurred) of the acetate intermediate(s) to the final Nazarov
the choice of the N protecting group is not of secondary product(s). With all screened catalysts, compound 16a was
importance, as its involvement in gold-catalyzed rearrange- a competent substrate. However, the expected acetate 17a[14]
ment of closely related N-Boc enynes has been shown,[8] was always obtained in mixture with cyclopentenone 20a,
3944
www.eurjoc.org
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2015, 3943–3956

Gold(I)-Catalyzed Synthesis of Annulated N-Heterocycles
acetate 17a in 61% yield and 20a in 18% yield after
chromatography. Decreasing the reaction temperature to
0 °C the ratio between 17a and 20a was further improved
(4:1), but the yield of 17a was not changed much (60%) as
degradation took place during the reaction.
In sharp contrast with the results obtained by using Ag-
OTf and even more with those reported for the correspond-
ing carbacyclic systems 1 (Scheme 1),[5a,5b] with all of the
other silver salts the major product was always cyclopen-
tenone 20a (entries 4–6) accompanied by lower amounts of
acetate 17a and only traces (less than 5%) of cyclopen-
tenone 18a. The best silver salt was AgSbF6 (entry 6) as the
reaction provided cyclopentenone 20a in 70% yield after
chromatography together with some residual acetate 17a
(14%). Again contrarily to expectations, the reaction was
Scheme 2. [a] KHMDS, –78 °C, 1.5 h; then (PhO)2POCl, –78 °C,
1 h. [b] heptyn-3-ol, 15 mol-% CuI, 6 mol-% (Ph3P)2PdCl2, CHCl3/
Et3N, 1:2, 55 °C, 5–7 h. [c] Ac2O, Et3N, DMAP, DCM, 0 °C to much slower when using “wet” CH2Cl2.[5a] In this case, only
after 5 h the conversion was complete.[16] Instead, the con-
room temp., 1–7 h.
current hydrolysis of acetate 19a was faster to provide 20a
possessing the more substituted double bond in 4a–7a posi-
tion and possibly deriving from the hydrolysis in situ of its
acetate precursor 19a. Intriguingly, we never isolated this
acetate.[14] With AgOTf[15] (entry 1) complete conversion
was observed in less than 1.5 hours, whereas hydrolysis oc-
curred only partially in 16 hours, providing acetate 17a in
again as the major product but in a lower ratio with resid-
ual acetate 17a (55% and 30% respectively after chromatog-
raphy) than in commercial CH2Cl2. Similarly, a lower ratio
between 20a and 17a was obtained when the reaction was
carried out with AgSbF6 in chloroform (entry 10).
To demonstrate that the formation of ketone 20a occurs
51% yield and cyclopentenone 20a in 35% yield after via hydrolysis in situ of acetate intermediate 19a, we set up
chromatography. Trying to increase the relative amount of one experiment in CDCl3 and monitored the progress of
acetate 17a, the reaction was carried out in different sol- the reaction by 1H NMR spectroscopy (Figure 2). Aliquots
vents (THF, 1,4-dioxane, acetonitrile, dichloroethane). The of the reaction mixture were filtered through a Celite layer
best result was achieved in toluene which provided pure and diluted after 0.5, 1 and 2.5 h from the addition of the
Table 1. Sequential gold(I)-catalyzed rearrangement/Nazarov reaction of acetate 16a.[a]
Entry
Catalyst[b]
Conditions[c]
Time [h][d,e]
Yield[f]
17a
18a
19a
20a
[g]
1
2
3
4
5
6
7
8
Ph3PAuCl/AgOTf
CH2Cl2
toluene
toluene, 0 °C
CH2Cl2
CH2Cl2
1.5
1
5
3
6
2.5
5 (0)
5[i] (0)
3
51
61
60
18
18
14
30
18[h]
18
30
–
–
–
–
–
–
–
–
–
–
–
35
18
15
62
50
70
Ph3PAuCl/AgBF4
Ph3PAuCl/AgNTf2·ACN
Ph3PAuCl/AgSbF6
5
5
CH2Cl2
3
–
–
wet CH2Cl2
dry CH2Cl2
CH2Cl2
CHCl3
CH2Cl2
CH2Cl2, reflux
CH2Cl2
CH2Cl2
55
22[h]
–
–
–
–
–
19[h]
64
[j]
9
Ph3PAuSbF6
–
10
11
12
13
14
15
Ph3PAuCl/AgSbF6
Ph3PAuCl/Cu(OTf)2
2 (0.5)
–
60
29
6[k]
19
58[h]
–
0.7 (3)
3
3
–
–
–
–
42[h]
Ͻ 5[h]
Ͻ 5[h]
–
AgOTf
AgSbF6
Cu(OTf)2
–
–
–
–
CH2Cl2
16 (0)
[a] Reactions were carried out on 0.1–0.15 mmol scale, at 25 °C and left standing whilst stirring 16 h after consumption of the starting
material. An aqueous work-up was carried out to recover the products from the reaction mixture. [b] Catalysts were prepared by adding
the silver salt to a 0.004 m solution of the gold(I) chloride in the reaction solvent. [c] Solvents were not dried before use unless otherwise
indicated. [d] Time to reach complete conversion of the starting material. [e] In brackets, time left for hydrolysis. [f] Yield after chromatog-
1
1
raphy unless otherwise indicated. [g] Not detected by H NMR analysis of the crude reaction mixture. [h] By H NMR analysis of the
crude reaction mixture. [i] The reaction stopped at a 70% conversion. [j] The reaction was carried out with 6 mol-% of catalyst and after
filtration of AgCl. [k] The conversion was not complete and degradation of the starting material occurred.
Eur. J. Org. Chem. 2015, 3943–3956
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
3945

E. G. Occhiato et al.
FULL PAPER
1
catalyst. After 30 min, H NMR analysis actually showed triflate as a counterion from both AgOTf and Cu(OTf)2,
three new sets of signals, of which one for cyclopentenone there is an appreciable increase of the relative amount of
20a and two for the two acetates (17a and 19a). As we could either acetate 17a or its product of hydrolysis 18a. A [1,5]-
anticipate from our previous results, the ratio between inter- H shift in 19a to form 17a triggered by triflic acid (as
mediates 19a and 17a decreased during the reaction concur- H3O+TfO–) generated in situ can be excluded because by
rently to a relative increase of 20a. At the end of the reac- acidic treatment we instead observe some conversion of 17a
tion, only the signals of 20a and acetate 17a were present into 19a. In fact, when acetate 17a was dissolved in dichlo-
in the spectrum.
romethane/MeOH and treated with a catalytic amount of
monohydrate pTsOH (Scheme 3), it slowly (16 h) provided
the corresponding cis fused (see later) cyclopentenone 18a
as the main product and in a 13:1 ratio with 20a, the forma-
tion of which could be accounted for by an acid catalyzed
[1,5]-H shift in 17a.[24] On the contrary, attempts at basic
hydrolysis (K2CO3 in MeOH) led to decomposition of the
starting material.[25] Similarly, we observed decomposition
of cyclopentenone 18a when we tried to convert it into its
isomer 20a upon treatment with bases.[26]
Figure 2. Monitoring by 1H NMR of the reaction of 16a in CDCl3.
Having in mind to prevent hydrolysis of the acetate inter-
mediates and to isolate 19a we also carried out one experi-
ment in anhydrous CH2Cl2 (entry 8) but the reaction was
slower and, after 5 h, only reached an approximately 50%
conversion into the acetates (roughly in a 1:1 ratio between
1
17a and 19a) by H NMR analysis of the crude reaction
mixture.[17] Degradation of enynyl acetate 16a occurred to
a great extent under these conditions, though, and we ob-
Scheme 3. Hydrolysis of acetates 17a and 17d.
served the formation also of a small amount of final prod-
The best reaction conditions (3 mol-% Ph3PAuCl/
1
uct 20a (19% by H NMR of the crude reaction mixture) AgSbF6, in DCM at room temperature) found for enyne
probably due to the presence of adventitious water.[18]
16a were then extended to the corresponding substrates
Given the role that silver can have in gold-catalyzed reac- protected as N-Boc (16b), N-Cbz (16c) and N-CO2Me
tion,[19] we carried out an experiment with a catalyst ob- (16d), all prepared in good overall yield as described for 16a
tained after filtration through a Celite layer (entry 9), in starting from the corresponding protected lactams
order to exclude AgCl from the reaction medium. Since de- (Scheme 2). Given our previous results on the gold-cata-
composition of the catalyst during such an operation has lyzed oxyauration of simple N-Boc protected enynes,[8a] we
been reported[20] we opted for a larger amount of initial anticipated a possible interaction of the carbamate carbonyl
catalyst (6 mol-%). We were glad to see that the reaction group with the activated triple bond before the acetate re-
occurred as usual, thus demonstrating that the silver halide arrangement and so we were not surprised by the much
does not affect neither the reaction rate nor the selectiv- lower reactivity of substrates 16b–d. The reaction of N-Boc
ity.[21] On the other hand, AgOTf alone (entry 13) as well derivative 16b (Table 2, entries 1–3) only provided degrada-
as AgSbF6 (entry 14) did catalyze the reaction, albeit to a tion products plus a minor amount of the oxyauration
very low extent (less than 5% conversion after 3 h).[22]
product (about 20%) when the reaction was carried out in
Finally, when the reaction was carried out using Cu- toluene.[27] The formation of this byproduct, in particular,
(OTf)2 to generate the active gold(I) catalyst (entry 11),[23] is consistent with our hypothesis of a carbamate carbonyl
both isomers 18a and 20a were obtained in a 1.5:1 ratio group competing for the activated triple bond (see later for
and 48% yield (degradation of the starting material was DFT calculation on this issue). Better results, but only with
observed under these conditions). Repeating the reaction 8 mol-% of catalyst, were obtained with N-Cbz derivative
in refluxing dichloromethane, consumption of the starting 16c (entry 5) which provided cyclopentenone 20c (61%
material was complete in 40 min and hydrolysis in 3 h, pro- yield after chromatography) as the sole product. In this
viding again a mixture of the two isomers in a 1:1.4 ratio. case, both consumption of the starting material and hydrol-
It is interesting to note here that only with Cu(OTf)2 we ysis were complete after 3 h in refluxing DCM. Also in this
observe hydrolysis of acetate 17a. Unlike AgOTf, copper case, some degradation of the starting material occurred,
triflate remains in solution and due to its Lewis acid charac- albeit in a much lower extent than with N-Boc compound
ter it can accelerate the hydrolysis of the acetates. Also, with 16b. Similarly, with the N-CO2Me protected compound the
3946
www.eurjoc.org
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2015, 3943–3956

Gold(I)-Catalyzed Synthesis of Annulated N-Heterocycles
Table 2. Sequential gold(I)-catalyzed rearrangement/Nazarov reaction of acetates 16b–d.[a,b]
Entry
Substrate
Conditions[c]
Time [h][d]
Yield[e]
17
18
19
20
[g]
1
2
3
16b
CH2Cl2, 25 °C
16[f]
3[f]
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
48
61
–
64
50
45
toluene, 25 °C to reflux
acetone, 25 °C to reflux
CH2Cl2, 25 °C
CH2Cl2, reflux
CH2Cl2, 25 °C
CH2Cl2, reflux
toluene, reflux
THF, reflux
3[f]
4[h]
5[h,i]
6
16c
16d
20[j]
3
–
2
–
7
8
9
2[k]
1.5[k]
1.5[k]
14
11
15
[a] Reactions carried out on 0.15–0.2 mmol, with 5 mol-% of catalyst unless otherwise indicated. [b] Catalysts were prepared by adding
the silver salt to a 0.004 m solution of the gold(I) chloride in the reaction solvent. [c] Solvents were not dried before use unless otherwise
indicated. [d] Time to reach complete conversion of the starting material. [e] Yield after chromatography unless otherwise indicated. [f]
Complete degradation of the starting material occurred. [g] Not detected by 1H NMR analysis of the crude reaction mixture. [h] Reaction
carried out with 8 mol-% of catalyst. [i] Reaction carried out on a 0.075 mmol scale. [j] Degradation occurred to a certain extent. [k]
Then left standing for 16 h at room temperature.
reaction did not take place if not by heating in refluxing this result with those reported by Zhang et al. on the corre-
dichloromethane (entry 7) with 5 mol-% of catalyst. Again, sponding carbacyclic systems 1 (Scheme 1), in which the
Nazarov compound 20d with the double bond at 4a-7a po- formation of a unique cyclopentenone (6) with the less sub-
sition was the major product after chromatography (64%) stituted double bond was observed.[5a]
and in this case we isolated also a smaller amount of acetate
17d (14%).[28] Similarly to acetate 17a, when we subjected
acetate 17d to hydrolysis in DCM/MeOH (1:1) and in the
presence of catalytic amount of monohydrate pTsOH
(Scheme 3), its quantitative conversion into the correspond-
ing Nazarov product 18d was slow (16 h) and this was ob-
tained in a 4:1 mixture with its isomer 20d.
Mechanistic Considerations and DFT Calculations
The fact that we could in most cases isolate acetates 17a
and 17d and never their isomers (19a and 19d) from the
crude reaction mixtures of the gold-catalyzed reactions, to-
gether with the observation that acetates 17 are very slowly
converted into the corresponding cyclopentenones when
subjected to hydrolytic conditions (Scheme 3), suggests that
Scheme 4. The two competing pathways to 18a and 20a.
In order to understand this process in more detail, and
acetates 17, after their formation in the protodeauration help identifying the structures and energies of the critical
step involving oxyallyl cation 8 (Scheme 4), only slowly hy- steps of the mechanism, we studied the potential reaction
drolyze under the conditions of the gold-catalyzed reaction. coordinates computationally. The structures were located
Instead, isomers 19 are quickly converted into the corre- using the B3LYP[29] density functional theory method as
sponding cyclopentenones 20 and so we never isolated them implemented in the Gaussian suite of programs.[30]
(see the NMR study above, Figure 2). Since cyclopentenone
For the sake of simplicity, two model complexes were
20 is the major or only product in the reactions carried out considered for the computational studies, that is
I
in DCM when using AgSbF6 as the anion source we can (Scheme 5), which contains the smallest substituent
hypothesize that, under these conditions, formation of acet- (methyl) in the propargylic position and a N-CO2Me moi-
ate 19 from oxyallyl cation 8 is faster than formation of its ety; and the corresponding one with the N-tosyl group. In
isomer 17 and that then a fast hydrolysis of 19 under the the former case, the methyl carbamate is conformationally
reaction conditions occurs. It is interesting to compare here simpler than the tosylate group, and would allow us to
Eur. J. Org. Chem. 2015, 3943–3956
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
3947

E. G. Occhiato et al.
FULL PAPER
Scheme 5. Top: acetate rearrangement in the initial steps of the mechanism. Bottom: cyclization step from III and protodeauration step.
mimic also compound 16d and related carbamates 16b–c. Although not shown in the Scheme the results with the N-
The alkynyl–gold(I) cationic complex I (Scheme 5) was thus tosyl derivative are comparable, showing activation energies
considered as the starting point of the mechanism (G = of 11.4 and 10.9 kcal/mol for the two step acetate re-
0 kcal/mol), and all reported energies in the following dis- arrangement processes, a reaction energy of +0.7 kcal/mol
cussion are relative to it. The energy values correspond to in the formation of III, and a low activation barrier for the
ΔG Gibbs Free energies, computed at B3LYP/6-31G** level Nazarov cyclization (4.0 kcal/mol).
(LANL2DZ[31] for gold atom). The overall discussion is
At this point, we attempted to explain the diene forma-
based on model I. The results with the corresponding N- tion through a single-step intramolecular hydride shift with
tosyl derivative are very similar and reinforce the conclu- concomitant C–Au bond breaking. In fact, the correspond-
sions.
ing transition structure TS4 was located (Scheme 5), but its
Initially, the coordination of the gold atom to the triple accompanying activation energy (30 kcal/mol from IV to
bond induces a rapid two step acetate rearrangement to TS4) is too high to be feasible under the experimental reac-
form the pre-Nazarov cyclization complex III (Scheme 5). tion conditions.[33] Therefore, an external base is needed for
The attack of the acetate carbonyl oxygen of I to the gold- the deprotonation, and it is important to highlight that sev-
activated alkyne leads to the formation of a cyclic interme- eral possible candidates exist in the reaction media, like the
diate II, and the subsequent C–O bond breaking event ren- gold counterion in the gold(I) salt or the anion forming the
ders an allylic cation (III), stabilized by the presence of the silver or copper salt. Even some water molecules present in
gold atom.[32] Noteworthy, the computed energies indicate the reaction media could play a significant role and, in fact,
that I and III are isoenergetic (0.1 kcal/mol difference), and the intervention of water molecules in the deprotonation/
that the activation energies of both steps are fairly low, ca. protodeauration steps has been suggested for the tandem
10 kcal/mol, meaning that in the absence of further evol- process involving enynyl acetates 1 (Scheme 1) in “wet”
ution, I and III would be in an almost 1:1 equilibrium. CH2Cl2.[5b,34] We were particularly interested in the proton
However, complex III evolves through an easy cyclization abstraction step since, based on our experimental findings,
to IV, a process that presents a very low barrier of 5.1 kcal/ the regioselectivity of the reaction, which was not an issue
mol (TS3, Scheme 5). In accordance to the conrotatory na- with enynyl acetates 1, seems to be determined with our
ture of the Nazarov reaction under thermal conditions, the substrates just at this stage. We have in fact shown (vide
reaction is predicted to be diastereoselective, with formation supra) that once the final diene-acetates VI and VIII (mod-
of the C–C bond that presents the two H atoms in a trans els for acetates 19 and 17, respectively, Figure 3) are formed,
relationship. The alternative diastereoisomeric transition there are not evident signs of significant equilibration be-
state, in which the two H are cis to each other is 3.9 kcal/ tween them during the reaction. So, as a plausible approxi-
mol higher in energy, and thus, not operative. The low acti- mation, and without knowing the exact nature of the mo-
vation barriers can explain the fast rate of a reaction that lecule responsible for the abstraction, to demonstrate that
is completed at room temperature in less than two hours. this is actually the regiodetermining stage, we decided to
3948
www.eurjoc.org
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2015, 3943–3956

Gold(I)-Catalyzed Synthesis of Annulated N-Heterocycles
Figure 3. Transition states TS5 and TS6 in the presence of triflate anion and compounds formed thereafter.
compute the deprotonation step by using the triflate anion
as the base (Figure 3 and Figure S2 in Supporting Infor-
mation).[35]
Our computational results indicate that the triflate medi-
ated hydrogen abstraction is, in fact, not highly regioselec-
tive, the difference between TS5 and TS6 being of only
0.8 kcal/mol in favor of the formation of VII isomer (en
route to 17). In the case of the N-tosyl group, the difference
is 0.4 kcal/mol (favoring TS6). The experimental data
(Table 1) show a regioselectivity that ranges from 4:1 (17a
over 20a, entry 3), when using AgOTf as the source of the
non-coordinating anion, to 1:5 (20a over 17a, entry 6),
when instead using AgSbF6, which computationally ac-
counts for a difference of less than 1 kcal/mol in each sense.
As we will show later, with different types of substrates the
regioselectivity we observed when using AgSbF6 was in-
stead complete in favour of ketones derived from hydrolysis
of acetate VI (i.e. 19).
Scheme 6. Hydrolysis of diene-acetates VI and VIII in the presence
of a model acid (TfOH or H3O+).
After the formation of VI- and VIII-type dienes (Fig-
ure 3), the most logical process would follow via hydrolysis
to the final products (Scheme 6). There is a final interesting
question at this point, regarding the very different experi-
mental hydrolysis rates of the two diene types, 17 and 19.
Concerning the lower reaction rates of N-alkoxycarbonyl
Once again, as the exact nature of the protonating species protected substrates 16b–d, as we have hypothesized, DFT
is unknown, we chose triflic acid and H3O+ as simple mod- calculations predict the formation of a non-productive cy-
els to study computationally this issue, and the transition clized intermediate (from carbonyl oxygen attack), which is
states for the hydrogen transfer were located. Noteworthy, more stable than the starting material, thus sequestering the
the protonation of VI (model of diene 19) is predicted to gold catalyst for a while and reducing the reaction rate
be three or four orders of magnitude faster than the corre- (Scheme 7). The starting compound I can proceed through
sponding protonation of VIII (model of 17), as derived acetate rearrangement to form III or through cyclization
from a 3.5 kcal/mol (H3O+) or 5.7 kcal/mol (TfOH) energy with the methyl ester to form XI. The three complexes XI,
difference in favor of the former. These data easily explain I, and III are in equilibrium, which is shifted towards the
why non-hydrolyzed 17 and hydrolyzed 20 are the final non-productive, but low in energy (–6.1 kcal/mol) side com-
products of the reaction. The result can be understood in plex XI. Only the irreversible Nazarov cyclization from III
light of the dienamine structure of compound 17 (VIII), to IV is finally able to displace this equilibrium towards
and the donor character of the nitrogen atom, which can the formation of the bicyclic adducts. Thus, XI is partially
induce a stabilization of that structure making it less prone sequestering the gold catalyst and decreasing the reaction
to hydrolysis.
rate.
Eur. J. Org. Chem. 2015, 3943–3956
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
3949

E. G. Occhiato et al.
FULL PAPER
Scheme 7. DFT predicted formation of the non-productive XI in-
termediate.
The Torquoselectivity
To evaluate the torquoselectivity[36] in the ring closure we
synthesized model compounds 16i and 16j (Scheme 8) from
the corresponding lactams,[37] which should provide Naza-
rov compounds with defined stereochemistry we had al-
ready prepared in the past.[7a] The reaction of enyne 16i was
first carried out in DCM at room temperature in the pres-
ence of 5 mol-% of Ph3PAuSbF6 as the catalyst and pro-
vided, quite interestingly, only cyclopentenone 23i in 78%
yield after chromatography (Scheme 9) with no traces of the
acetate 21i.
Scheme 9. [a] Reaction carried out under reflux. [b] 83% yield,
pure cis after chromatography. Reaction carried out with
(c-Hex)3PAuSbF6.
1H NMR analysis of 23i revealed that a 6:1 mixture of
diastereomers was present, with the cis compound being
predominant. As the pre-Nazarov complex is a vinylgold
species, a few ligands with different properties were scree-
ned to evaluate their possible role in the ring closure selec-
tivity. Best results were obtained with the electron-rich li-
gand (c-hex)3P (10:1 cis/trans ratio, separable mixture, 91%
yield) whereas with electron-poor [(pCF3C6H4)3P] and
NHC-carbene ligands [1,3-bis(2,6-diisopropylphenyl)imid-
azol-2-ylidene], not only the selectivity did not ameliorate
but the reaction provided a mixture of cyclopentenones,
acetates and degradation products. Moreover, the reaction
with 5 mol-% (c-hex)3PAuSbF6 was much cleaner than with
Scheme 8. Reagents and conditions: [a] KHMDS, –78 °C, 1.5 h;
then (PhO)2POCl or PhNTf2, –78 °C, 1 h. [b] alkynol, 15 mol-%
CuI, 6 mol-% (Ph3P)2PdCl2, CHCl3/Et3N, 1:2, 55 °C, 5–7 h. [c]
Ac2O, Et3N, DMAP, DCM, 0 °C to room temp., 1–7 h. [d] Yield
over two steps. [e] Prepared from 16k. [f] Yield over three steps.
3950
www.eurjoc.org
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2015, 3943–3956

Gold(I)-Catalyzed Synthesis of Annulated N-Heterocycles
Ph3P. [38] These results are consistent with our previous ob- atives 16o and 16p, and enyne 16q, which were obtained
servations on the torquoselectivity in the Nazarov reac- from the corresponding lactam-derived enol triflate-
tion,[7a,7b,7d,7g] although the cis selectivity was in those cases s.[7a,7b,40,41] The results are shown in Scheme 9.
complete.[7a] For stereoelectronic reasons, the ring closure
The number of alkyl groups at 3Ј position seems to affect
occurs in a way to form the new bond on the opposite side the reaction outcome and reactivity. Unsubstituted acetate
of the axially oriented 6-Me group.[7d,7g] The bases for the 16e reacted slower at room temperature and it required
lower torquoselectivity in the gold-catalyzed process are not higher temperature (refluxing DCM) to be completely con-
easy to understand, and could be either kinetic (i.e. both verted into 23e (83%) only. This could be due to a slower
clockwise and counterclockwise ring closures could take rearrangement of the propargyl acetate moiety as a positive
place at different rate in the putative intermediate XII, charge develops on a primary C atom (C3Ј) in the transition
Scheme 10) or related to the geometry of the oxyallyl cation state. Interestingly, no traces of acetate 21e (or the corre-
before ring closure. In the latter case a counterclockwise sponding ketone) were found in the crude reaction mixture,
ring closure of the isomer XIII would lead to the trans com- meaning that proton abstraction is much more favored if
pound. However, DFT calculations revealed that the prefer- occurs at the more substituted position to generate the most
ence for XII is clear, with a difference over XIII of 2.3 kcal/ substituted double bond. The calculated energy difference
mol because of steric reasons.[3g] Thus the major cis di- between the two competing transition state is 1.2 kcal/mol
methyl diastereoisomer (XIV) is formed by a counterclock- (as usually using triflate as the base) which corresponds to
wise conrotatory ring closure of XII, with a low activation a 7:1 selectivity (Figure 4).
energy of 6.4 kcal/mol (TS9), whereas its trans diastereomer
XV is formed by conrotatory clockwise ring closure from
XII (TS10, 9.4 kcal/mol). Because of the predominance of
intermediate XII it is likely that the alternative pathway
from XIII is never operative.[39]
Figure 4. DFT study on the competing proton abstraction pathway
in the reaction of 16e.
The results obtained with 3Ј-methyl-substituted 16f were
in line with those for our model compound 16a whereas,
curiously, dimethyl-substituted acetate 16g provided acetate
22g (86%) only, which did not hydrolyze spontaneously.
This was later done by treatment with pTsOH (Scheme 11)
which furnished Nazarov compound 23g in 86% yield.
With the 3Ј-phenyl-substituted enyne 16h the reaction pro-
vided expected cyclopentenone 23h in 55% after
chromatography. Acetate 21h, formed during the reaction
in mixture with 23h in ca. 1:1 ratio, during the purification
underwent [1,5]-H shift giving a complex mixture of few
acetate isomeric products. A small amount of pure 21h was
treated with pTsOH (Scheme 11), and it was converted into
a mixture of 24h (95%) and 23h. With compound 24h it
Scheme 10. DFT study on the torquoselectivity of the ring closure.
The torquoselectivity with 4-methyl substitute enyne 16j
was in line (6:1) with that observed for 16i, as this substrate
provided a separable 6:1 mixture of cis (major) and trans
(minor) compounds 23j (66% after chromatography) when
using Ph3PAuSbF6 as the catalyst, together with some acet-
ate 21j (14% after chromatography). In this case, the use
of other ligands did not improve the torquoselectivity and,
moreover, caused the formation of a complex mixture of all
acetates and cyclopentenone isomers.
1
was possible to demonstrate by H NMR studies the cis
fusion of the rings (in analogy to the corresponding carba-
cyclic systems)[5a] and the structure of most populated con-
former. 7a-H is a doublet with a low coupling constant
(6.8 Hz) with 4a-H indicating its equatorial orientation, fur-
ther confirmed by the lack of any NOE of 7a-H with 2-Hax
and 4-Hax. Consequently, 4a-H is axially oriented. Simi-
larly, in compound 18a (Scheme 3) 7a-H is a doublet with
a low coupling constant (6.9 Hz) with 4a-H confirming the
Assessing the Scope of the Reaction
To assess the scope of the reaction, a series of enynyl cis fusion also for this compound. Hydroxy-substituted pip-
acetates (Scheme 8) was prepared by varying the substitu- eridine derivatives, both protected (16k) and unprotected
ents on the alkyne moiety and on the piperidine ring as well (16l) on the OH group, were compatible with the reaction
as the heterocycle. All acetates 16 were prepared in good conditions. However, while 16k provided cyclopentenone
yield over two or three steps from lactam-derived enol 23k as the sole products in 75% yield after chromatography,
phosphates 14, with the exception of pyrrolidin-2-one deriv- substrate 16l furnished the acetylated derivatives 25 as the
Eur. J. Org. Chem. 2015, 3943–3956
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
3951

E. G. Occhiato et al.
FULL PAPER
major compound (75%), accompanied by a minor amount under Sonogashira conditions. After acetylation, the gold-
of the corresponding alcohol 23l (16%). Clearly, the free catalyzed rearrangement of the enynyl acetates readily (and
OH group acts a nucleophile by trapping the acyl cation best) occurs when using hexafluoroantimonate as the non-
which is released in the final step of the process (Scheme 6). coordinating anion. This generates a divinyl cation which
undergoes a 4π electrocyclization forming the target annu-
lated N-heterocyclic compound in good to excellent yield.
This process has been studied in details both experimentally
and computationally, and the influence of the reaction con-
ditions and the structure of the substrates on the reaction
rate, regio- and stereoselectivity have been evaluated. The
main features can be summarized as follows: (a) compared
to the tandem process of carbacyclic enynyl acetates, the
presence of the N atom clearly favors the nearly exclusive
formation of the Nazarov product having the most substi-
tuted double bond, i.e. at the 4a-7a position. This is true
with most catalysts we used, although its isomer (i.e. with
the double bond at the 5–6 position) was the major product
when triflate was the gold(I) counterion. (b) In contrast to
similar cases in literature, suitably prepared “wet” dichloro-
methane seems to be not necessary to have a complete hy-
drolysis of the acetate intermediates. However the presence
of some water in the solvent seems in any case essential for
the process to occur properly, as in dry CH2Cl2 the conver-
sion into the Nazarov product mixture was much lower.
Under the best conditions (with Ph3PAuSbF6) the hydroly-
sis of only one of the two regioisomeric acetates occurs,
leading to the 4a-7a unsaturated compound. (c) All N-pro-
tecting groups are compatible with the propargyl acetate
rearrangement, with exception of the N-Boc group, and
with the N-Ts being the best. The other N-alkoxycarbonyl
groups somehow interfere, though, and heating is necessary
to achieve complete conversion. (d) With piperidine rings
substituted at 4 or 6 position, moderate to high torquoselec-
tivity in the ring closure in favor of the cis diastereomers is
observed when the propargyl moiety bears a substituent at
C3Ј. This is in accordance to the results previously reported
by us for the classical Nazarov reaction, meaning that the
gold atom on the divinyl cation has a little effect on the
stereoelectronically preferred conrotation mode. (e) The
scope of the reaction ranges from six to seven-membered
N-heterocyclic rings and various substituents at C3Ј on the
propargyl moiety, but is not extended to pyrroline deriva-
tives which proved unsuitable for this approach. Instead, an
indole containing enynyl acetate reacted smoothly to pro-
vide the cyclopenta[b]indole nucleus. Because of the combi-
nation of easily accessible propargyl alcohols and readily
available lactams in the assemblage of the substrates re-
quired for the tandem process, this methodology is surely
suited for the preparation of natural and biologically active
compounds possessing a cyclopenta-annulated N-heterocy-
clic nucleus.
Scheme 11. [a] pTsOH (cat.), DCM-MeOH, 1:1, room temp., 16 h.
Also seven-membered rings as in 16m and 16n were com-
patible as these substrates provided, after complete hydroly-
sis of the corresponding acetate intermediates, Nazarov
compounds 23m (81%) and 23n (83%) only and in excellent
yield after chromatography. As in the case of 16e, the reac-
tion of 16n required refluxing conditions, too. In both cases,
disappearance of the starting material was faster than with
the corresponding six-membered heterocycles and no traces
1
of acetate isomers were detected by H NMR analysis the
crude reaction mixture. Not unexpectedly,[7a] the ring clo-
sure of five-membered derivatives 16o and 16p did not oc-
cur both at room temperature and in refluxing CH2Cl2, but
we only recovered residual starting material with a certain
amount of unidentified degradation compounds. The reac-
tion of 16p was also carried out in refluxing toluene, but to
no avail. We have already reported the greater difficulty in
the ring closure to give a 5–5 fused system, presumably due
to ring strain in the intermediate azabicyclo[3.3.0]octenyl
cation.[7a] On the other hands, we were very glad to observe
that propargyl acetate 16q containing a N-CO2Me-pro-
tected indole nucleus reacted smoothly in refluxing CH2Cl2,
to provide the corresponding cyclopenta-fused aromatic
system 23q in 84% yield after 1.5 h, thus paving a new way
for the synthesis of natural compounds containing the cy-
clopenta[b]indole nucleus.
Conclusions
In this paper we have demonstrated that the tandem
gold(I)-catalyzed rearrangement/Nazarov reaction of prop-
argylic ester derivatives is a useful strategy for the synthesis,
in just four steps, of cyclopenta-fused N-heterocyclic struc-
tures present in many natural compounds. First, readily
available lactams are converted into the corresponding enol
Experimental Section
General: Anhydrous solvents were prepared according to the stan-
dard techniques. Commercially available reagents were used with-
phosphates and triflates and coupled to propargyl alcohols out further purification. Chromatographic separations were per-
3952 Eur. J. Org. Chem. 2015, 3943–3956
www.eurjoc.org
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Gold(I)-Catalyzed Synthesis of Annulated N-Heterocycles
formed under pressure on silica gel 60 (Merck, 70–230 mesh) by
using flash column techniques; Rf values refer to TLC carried out
on 0.25 mm silica gel plates with the same eluent as indicated for
column chromatography. 1H NMR spectra were recorded at 200 or
(6 mL) was cooled (ice bath) and Ac2O (0.28 mL, 3 mmol) was
added. The reaction mixture was stirred at room temperature and
monitored by TLC. When the conversion was complete, a satd
solution of NaHCO3 (15 mL) was added and the product extracted
400 MHz and 13C NMR spectra at 100.4 MHz, both in CDCl3 with DCM (3ϫ 15 mL). The combined organic extracts were dried
solution. Mass spectra were carried out by direct inlet of a 10 ppm
solution in CH3OH on an Ion Trap LC/MS system with electro-
spray ionization (ESI) interface in the positive ion mode.
with anhydrous K2CO3 for 30 min. After filtration and evaporation
of the solvent, the crude was purified by flash chromatography (n-
hexane/EtOAc, 8:1 + 1% Et3N; Rf = 0.21) affording pure 16a as a
pale yellow oil (455 mg, 79%). This was stored at 4 °C as 0.1 m
solution in the eluent containing 1% Et3N until use. 16a: 1H NMR
(400 MHz, CDCl3): δ = 7.77 (d, J = 8.2 Hz, 2 H, Ts), 7.29 (d, J =
8.2 Hz, 2 H, Ts), 5.65 (t, J = 4.2 Hz, 1 H, 3-H), 5.45 (t, J = 6.6 Hz,
1 H, 1Ј-H), 3.64–3.62 (m, 2 H, 6-H), 2.42 (s, 3 H, CH3 Ts), 2.08 (s,
3 H, CH3 Ac), 2.06–2.03 (m, 2 H, 4-H), 1.72–1.74 (m, 2 H, 5-H),
1.65–1.59 (m, 2 H, 2Ј-H), 1.44–1.31 (m, 4 H, 3Ј-H and 4Ј-H), 0.90
(t, J = 7.2 Hz, 3 H, 5Ј-H) ppm. 13C NMR (100.4 MHz, CDCl3): δ
= 170.0 (s, CO), 143.4 (s, Ts), 137.2 (s, Ts), 129.5 (d, 2 C, Ts), 127.5
(d, 2 C, Ts), 124.4 (d, C3), 120.2 (s, C2), 97.2 (s, C2b), 81.1 (s, C2a),
64.4 (d, C1Ј), 46.0 (t, C6), 34.2 (t, C2Ј), 27.0 (t, C3Ј), 23.2 (t, C4),
22.2 (t, C4Ј), 21.5 (q, CH3 Ts), 21.0 (q, CH3 Ac), 21.0 (t, C5), 13.9
(q, C5Ј) ppm. MS (ESI): m/z (%) = 801 (100) [2M + Na]+, 412 (13)
[M + Na]+.
Diphenyl 1-(4-Tolylsulfonyl)-1,4,5,6-tetrahydropyridin-2-yl Phos-
phate (14a): A 0.5 m solution of KHMDS (8.9 mL, 4.44 mmol) in
toluene was diluted in anhydrous THF (28 mL) and cooled to
–78 °C.
A solution of N-tosyl δ-valerolactam 13a (900 mg,
3.55 mmol) in anhydrous THF (14 mL) was added dropwise. The
resulting mixture was stirred for 1.5 h at –78 °C and then diphenyl-
chlorophosphate (0.92 mL, 4.44 mmol) was slowly added and the
stirring continued below –70 °C for 1 h. The mixture was first
warmed to 0 °C and then quenched with aqueous 10% NaOH
(88 mL). The product was extracted with Et2O (4ϫ 40 mL); the
combined organic extracts were washed with 10% NaOH (40 mL)
and dried with anhydrous K2CO3 for 30 min. After filtration and
evaporation of the solvent, the crude was purified by flash
chromatography (n-hexane/EtOAc, 2:1 + 1% Et3N; Rf = 0.29) and
product 14a was obtained (1.38 g, 80%) as a colorless oil. Phos-
phate 14a was stored at 4 °C as 0.1 m solution in the eluent contain-
5-Butyl-1-(4-tolylsulfonyl)-1,2,3,4,5,6-hexahydro-[1]pyrindin-7-one
(20a): Precatalyst Ph3PAuCl (2.2 mg, 4.5 μmol, 3 mol-%) was dis-
solved in DCM (1.2 mL) and AgSbF6 (1.5 mg, 4.5 μmol, 3 mol-%)
was added. The formed suspension was left to stir at room tempera-
ture under nitrogen atmosphere. After 20 min a solution of enynyl
acetate 16a (59 mg, 0.15 mmol) in DCM (1.9 mL) was added and
reaction mixture was stirred at room temperature. The progress of
the reaction was monitored by TLC. After complete consumption
of the enynyl acetate (2.5 h) the reaction mixture was left to stir at
room temperature overnight (16 h). Water (4 mL) was added and
the product extracted with DCM (3ϫ 4 mL). The combined or-
ganic extracts were dried with anhydrous Na2SO4, filtered and con-
centrated. Chromatography (n-hexane/EtOAc, 6:1) afforded pure
ketone 20a (Rf = 0.17; 37 mg, 70%) as a yellow oil and acetate 17a
(Rf = 0.24; 8 mg, 14%) as an orange oil.
1
ing 1% Et3N until use. H NMR (400 MHz, CDCl3): δ = 7.75 (d,
J = 8.2 Hz, 2 H, Ts), 7.35–7.17 (m, 12 H, Ts and Ph), 5.25 (dd, J
= 6.6, 3.8 Hz, 1 H, 3-H), 3.66–3.64 (m, 2 H, 6-H), 2.37 (s, 3 H,
CH3 Ts), 2.08–2.03 (m, 2 H, 4-H), 1.56–1.51 (m, 2 H, 5-H) ppm.
13C NMR (100.4 MHz, CDCl3): δ = 150.4 (s, C2), 143.8 (s, Ts),
139.6 (s, 2 C, Ph), 137.0 (s, Ts), 129.7 (d, 4 C, Ph), 129.6 (d, 2 C,
Ts), 127.6 (d, 2 C, Ts), 125.5 (d, 2 C, Ph), 120.4 (d, 4 C, Ph), 100.5
(d, C3), 47.5 (t, C6), 21.5 (q, CH3 Ts), 21.3 (t, C4), 20.9 (t, C5)
ppm. MS (ESI): m/z (%) = 993 (100) [2M + Na]+, 508 (13.1) [M +
Na]+, 486 (8) [M + 1]+.
1-[1-(4-Tolylsulfonyl)-1,4,5,6-tetrahydropyridin-2-ylethynyl]pentyl
Acetate (16a): Phosphate 14a (1.16 g, 2.4 mmol) was dissolved in
an anhydrous 2:1 Et3N/CHCl3 mixture (14 mL), and (Ϯ)-heptyn-3-
ol (0.32 mL, 2.4 mmol), CuI (46 mg, 0.24 mmol) and (Ph3P)2PdCl2
(84 mg, 0.12 mmol) were added under nitrogen atmosphere. The
reaction mixture was heated at 55 °C (external) for 3 h and then a
second portion of (Ϯ)-heptyn-3-ol (0.16 mL, 1.2 mmol), CuI
(23 mg, 0.12 mmol) and (Ph3P)2PdCl2 (17 mg, 0.024 mmol) was
added. Heating was continued at 55 °C for 3 h. The mixture was
cooled to room temperature and water (36 mL) was added. The
product was extracted with Et2O (3ϫ 36 mL) and the combined
organic extracts were dried with anhydrous K2CO3 for 30 min. Af-
ter filtration and evaporation of the solvent, the crude was purified
by flash chromatography (n-hexane/Et2O, 2:1 + 1% Et3N; Rf =
0.30) affording enynyl alcohol 15a as a pale yellow oil (515 mg,
20a: 1H NMR (400 MHz, CDCl3): δ = 7.99 (d, J = 8.2 Hz, 2 H,
Ts), 7.29 (d, J = 8.2 Hz, 2 H, Ts), 3.50–3.36 (m, 2 H, 2-H), 2.72–
2.68 (bm, 1 H, 5-H), 2.60 (dd, J = 18.2, 6.4 Hz, 1 H, 6-H), 2.47
(dt, J = 19.9, 6.8 Hz, 1 H, 4-H), 2.41 (s, 3 H, CH3 Ts), 2.28 (dt, J
= 19.9, 6.4 Hz, 1 H, 4-HЈ), 2.13 (dd, J = 18.2, 2.2 Hz, 1 H, 6-HЈ),
2.01–1.99 (m, 2 H, 3-H), 1.76–1.71 (m, 1 H, 1Ј-H), 1.43–1.21 (m,
5 H, 1Ј-HЈ, 2Ј-H and 3Ј-H), 0.90 (t, J = 7.0 Hz, 3 H, 4Ј-H) ppm.
13C NMR (100.4 MHz, CDCl3): δ = 200.1 (s, C7), 162.6 (s, C7a),
143.6 (s, Ts), 138.2 (s, C4a), 137.5 (s, Ts), 129.5 (d, 2 C, Ts), 127.9
(d, 2 C, Ts), 46.2 (t, C2), 40.1 (t, C6), 39.3 (d, C5), 32.9 (t, C1Ј),
29.0 (t, C2Ј), 24.1 (t, C4), 22.9 (t, C3Ј), 21.9 (t, C3), 21.7 (q, CH3
Ts), 14.1 (q, C4Ј) ppm. MS (ESI): m/z (%) = 717 (100) [2M +
Na]+, 370 (12) [M + Na]+. C19H25NO3S (347.47): calcd. C 65.68,
H 7.25, N 4.03; found C 65.36, H 7.43, N 4.07.
1
62%). 15a: H NMR (400 MHz, CDCl3): δ = 7.72 (d, J = 8.2 Hz,
2 H, Ts), 7.30 (d, J = 8.2 Hz, 2 H, Ts), 5.63 (t, J = 4.3 Hz, 1 H, 3-
H), 4.43 (t, J = 6.6 Hz, 1 H, 1Ј-H), 3.68–3.65 (m, 2 H, 6-H), 2.42
(s, 3 H, CH3 Ts), 2.06 (td, J = 6.6, 4.3 Hz, 3 H, 4-H and OH),
17a: 1H NMR (400 MHz, CDCl3): δ = 7.67 (d, J = 8.0 Hz, 2 H,
Ts), 7.25 (d, J = 8.0 Hz, 2 H, Ts), 6.03 (br. s, 1 H, 6-H), 4.16–4.09
1.76–1.59 (m, 4 H, 5-H and 2Ј-H), 1.47–1.23 (m, 4 H, 3Ј-H and 4Ј- (m, 1 H, 2-H), 3.05 (td, J = 13.6, 2.8 Hz, 1 H, 2-HЈ), 2.41 (s, 3 H,
H), 0.91 (t, J = 7.2 Hz, 3 H, 5Ј-H) ppm. 13C NMR (100.4 MHz,
CH3 Ts), 2.21–2.03 (m, 7 H, CH3 Ac, 4-H, 4a-H and 1Ј-H), 1.64–
CDCl3): δ = 143.5 (s, Ts), 137.3 (s, Ts), 129.5 (d, 2 C, Ts), 127.4 (d, 1.56 (m, 1 H, 3-H), 1.48–1.23 (m, 5 H, 3-HЈ, 2Ј-H and 3Ј-H), 1.04
2 C, Ts), 123.6 (d, C3), 120.4 (s, C2), 90.9 (s, C2a), 80.9 (s, C2b), (qd, J = 12.8, 3.6 Hz, 1 H, 4-HЈ), 0.88 (t, J = 7.2 Hz, 4Ј-H) ppm.
62.8 (d, C1Ј), 46.0 (t, C6), 37.2 (t, C2Ј), 27.2 (t, C5), 23.2 (t, C4), 13C NMR (100.4 MHz, CDCl3): δ = 168.5 (s, CO), 150.3 (s, C7),
22.4 (t, C3Ј), 22.3 (q, CH3 Ts), 21.5 (t, C4Ј), 13.9 (q, C5Ј) ppm. 143.6 (s, Ts), 137.4 (s, Ts), 129.5 (d, 2 C, Ts), 127.4 (d, 2 C, Ts),
MS (ESI): m/z (%) = 717 (100) [2M + Na]+, 348 (14) [M + 1]+. A 123.7 (s, C7a), 122.9 (d, C6), 49.5 (t, C2), 45.5 (d, C4a), 30.7 (t,
solution of enynyl alcohol 15a (515 mg, 1.48 mmol), DMAP C2Ј), 29.7 (s, C5), 28.4 (t, C4), 28.1 (t, C1Ј), 25.0 (t, C3), 22.4 (t,
(37 mg, 0.30 mmol) and Et3N (0.57 mL, 4.44 mmol) in DCM C3Ј), 21.5 (q, CH3 Ts), 20.7 (q, CH3 Ac), 13.9 (q, C4Ј) ppm. MS
Eur. J. Org. Chem. 2015, 3943–3956
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
3953

E. G. Occhiato et al.
FULL PAPER
(ESI): m/z (%) = 801 (100) [2M + Na]+, 390 (12) [M + Na]+.
C21H27NO4S (389.51): calcd. C 64.75, H 6.99, N 3.06; found C
64.82, H 7.08, N 3.04.
[5] a) L. Zhang, S. Wang, J. Am. Chem. Soc. 2006, 128, 1442–
1443; b) F. Q. Shi, X. Li, Y. Xia, L. Zhang, Z. X. Yu, J. Am.
Chem. Soc. 2007, 129, 15503–15512, see also ref.[3g,3m]
[6] The metal-catalyzed rearrangement of 3-acyloxy-1,4-enynes,
called the Rautenstrauch rearrangement, occurs instead via ini-
tial 1,2-shift of the carboxylate and is another excellent way to
prepare cyclopenta-fused systems. For a very recent example,
see: a) W. Zi, H. Wu, F. D. Toste, J. Am. Chem. Soc. DOI:
10.1021/jacs.5b00613; see also: b) O. N. Faza, C. S. López, R.
Álvarez, Á. R. de Lera, J. Am. Chem. Soc. 2006, 128, 2434–
2437; c) V. Rautenstrauch, J. Org. Chem. 1984, 49, 950–952; d)
X. Shi, D. J. Gorin, F. D. Toste, J. Am. Chem. Soc. 2005, 127,
5802–5803; e) K. Kato, T. Kobayashi, T. Fujimami, S. Motod-
ate, T. Kusakabe, T. Mochida, H. Akita, Synlett 2008, 1081–
1085; f) P. A. Caruana, A. J. Frontier, Tetrahedron 2007, 63,
10646–10656; g) Y. Nakanishi, K. Miki, K. Ohe, Tetrahedron
2007, 63, 12138–12148; h) D. Nakagawa, M. Miyashita, K.
Tanino, Tetrahedron Lett. 2010, 51, 2771–2773, see also ref.[4c]
Acknowledgments
Financial support from the University of Florence is acknowl-
edged. M. P. and B. F. thank the European Commission for a Ma-
rie Curie fellowship (FP7-PEOPLE-2012-ITN, project ECHONET
“Expanding capability in Heterocyclic Organic Synthesis”, project
number 316379). Ente Cassa di Risparmio di Firenze is acknowl-
edged for granting a 400 MHz NMR instrument. SGI/IZO-SGIker
(UPV-EHU) is acknowledged for allocation of computational re-
sources. Mr. Emanuele Chiti is acknowledged for carrying out some
experiments. Dr. Maurizio Passaponti is acknowledged for techni-
cal assistance.
[7]
a) E. G. Occhiato, C. Prandi, A. Ferrali, A. Guarna, P. Ven-
turello, J. Org. Chem. 2003, 68, 9728–9741; b) C. Prandi, A.
Ferrali, A. Guarna, P. Venturello, E. G. Occhiato, J. Org.
Chem. 2004, 69, 7705–7709; c) C. Prandi, A. Deagostino, P.
Venturello, E. G. Occhiato, Org. Lett. 2005, 7, 4345–4348; d)
A. Cavalli, M. Masetti, M. Recanatini, C. Prandi, A. Guarna,
E. G. Occhiato, Chem. Eur. J. 2006, 12, 2836–2845; e) P. Larini,
A. Guarna, E. G. Occhiato, Org. Lett. 2006, 8, 781–784; f) L.
Bartali, P. Larini, A. Guarna, E. G. Occhiato, Synthesis 2007,
1733–1737; g) A. Cavalli, A. Pacetti, M. Recanatini, C. Prandi,
D. Scarpi, E. G. Occhiato, Chem. Eur. J. 2008, 14, 9292–9304.
a) A. Oppedisano, C. Prandi, P. Venturello, A. Deagostino, G.
Goti, D. Scarpi, E. G. Occhiato, J. Org. Chem. 2013, 78, 11007–
11016; b) E.-S. Lee, H.-S. Yeom, J.-H. Hwang, S. Shin, Eur. J.
Org. Chem. 2007, 3503–3507; c) A. Buzas, F. Gagosz, Synlett
2006, 2727–2730.
[1] For a recent and excellent review see: D. J. Aitken, H. Eijsberg,
A. Frongia, J. Ollivier, P. P. Piras, Synthesis 2014, 46, 1–24.
[2] For the most recent reviews, see: a) D. R. Wenz, J. R. de Alaniz,
Eur. J. Org. Chem. 2015, 23–37; b) M. A. Tius, Chem. Soc. Rev.
2014, 43, 2979–3002; c) M. J. Di Grandi, Org. Biomol. Chem.
2014, 12, 5331–5345; d) W. T. Spencer, T. Vaidya, A. J. Frontier,
Eur. J. Org. Chem. 2013, 3621–3633; e) T. Vaidya, R. Eisenberg,
A. J. Frontier, ChemCatChem 2011, 3, 1531–1548. For an older
but excellent review on the Nazarov reaction covering the past
decade, see: f) A. J. Frontier, C. Collison, Tetrahedron 2005, 61,
7577–7606.
[8]
[9]
[3] For homogeneus AuI-catalyzed processes involving a Nazarov
reaction see: a) M. Hoffmann, J.-M. Weibel, P. de Frémont, P.
Pale, A. Blanc, Org. Lett. 2014, 16, 908–911; b) W. Rao, M. J.
Koh, D. Li, H. Hirao, P. W. H. Chan, J. Am. Chem. Soc. 2013,
135, 7926–7932; c) E. Álvarez, D. Miguel, P. García-García,
M. A. Fernández-Rodríguez, F. Rodríguez, R. Sanz, Synthesis
2012, 44, 1874–1884; d) R. Sanz, D. Miguel, M. Gohain, P.
García-García, M. A. Fernández-Rodríguez, A. González-
Pérez, O. Nieto-Faza, Á. R. de Lera, F. Rodríguez, Chem. Eur.
J. 2010, 16, 9818–9828; e) C. H. Oh, S. Karmakar, J. Org.
Chem. 2009, 74, 370–374; f) S. E. An, J. Jeong, B. Baskar, J.
Lee, J. Seo, Y. H. Rhee, Chem. Eur. J. 2009, 15, 11837–11841;
g) G. Lemière, V. Gandon, K. Cariou, A. Hours, T. Fukuyama,
A.-L. Dhimane, L. Fensterbank, M. Malacria, J. Am. Chem.
Soc. 2009, 131, 2993–3006; h) S. Suárez-Pantiga, E. Rubio, C.
Alvarez-Rúa, J. M. González, Org. Lett. 2009, 11, 13–16; i)
G. Y. Lin, C. W. Li, S. H. Hung, R. S. Liu, Org. Lett. 2008, 10,
5059–5062; j) J. H. Lee, F. D. Toste, Angew. Chem. 2007, 119,
930–932; k) J. H. Lee, F. D. Toste, Angew. Chem. Int. Ed. 2007,
46, 912–914; Angew. Chem. 2007, 119, 930; l) G.-Y. Lin, C.-Y.
Yang, R.-S. Liu, J. Org. Chem. 2007, 72, 6753–6757; m) G.
Lemière, V. Gandon, K. Cariou, T. Fukuyama, A.-L. Dhim-
ane, L. Fensterbank, M. Malacria, Org. Lett. 2007, 9, 2207–
2209; n) G. Y. Lin, C. Y. Yang, R. S. Liu, J. Org. Chem. 2007,
72, 6753–6757; o) Z.-X. Ma, S. He, W. Song, R. P. Hsung, Org.
Lett. 2002, 14, 5736–5739. For homogeneus AuI-catalyzed iso-
Nazarov reactions, see: p) C. C. Lin, T. M. Teng, C. C. Tsai,
H. Y. Liao, R. S. Liu, J. Am. Chem. Soc. 2008, 130, 16417–
16423; q) C. C. Lin, T. M. Teng, A. Odedra, R. S. Liu, J. Am.
Chem. Soc. 2007, 129, 3798–3798.
The [1]pyrindine system that is formed is contained in a few
natural compounds. For example, for streptazolin, streptazon
C, and streptazon A, isolated from Streptomyces viridochrom-
ogenes, S. luteogriseus, and FORM5, Strain A1, see: a) H.
Drautz, H. Zähner, E. Kupfer, W. Keller-Schierlein, Helv.
Chim. Acta 1981, 64, 1752–1768; b) S. Grabley, H. Kluge, H.-
U. Hoppe, Angew. Chem. Int. Ed. Engl. 1987, 26, 664–665; An-
gew. Chem. 1987, 99, 692; c) S. Grabley, P. Hammann, H.
Kluge, J. Wink, P. Kricke, A. Zeeck, J. Antibiot. 1991, 44, 797–
800; d) C. Puder, P. Krastel, A. Zeeck, J. Nat. Prod. 2000, 63,
1258–1260; e) C. Puder, S. Loya, A. Hizi, A. Zeeck, J. Nat.
Prod. 2001, 64, 42–45. For the synthesis of streptazolin and
related natural products, see: f) B. M. Trost, C. K. Chung, A. B.
Pinkerton, Angew. Chem. Int. Ed. 2004, 43, 4327–4329; Angew.
Chem. 2004, 116, 4427, and references cited therein. For pyrin-
dicin, isolated from Streptomyces sp. SCC 2313, see: g) S.
Omura, H. Tanaka, J. Awaya, Y. Narimatsu, Y. Konda, T.
Hata, Agric. Biol. Chem. 1974, 38, 899–906. For abikoviromy-
cin, isolated from S. abikoensis, see: h) A. I. Gurevich, M. N.
Kolosov, V. G. Korobko, V. V. Onoprienko, Tetrahedron Lett.
1968, 9, 2209–2212. For 1-pyrindine alkaloids from Rhizocto-
nia leguminicola, see: i) F. P. Guengerich, S. J. Di Mari, H. P.
Broquist, J. Am. Chem. Soc. 1973, 95, 2055–2056. For pandam-
arilactone-32 and analogues from Pandanus amaryllifolius, see:
j) M. G. Nonatoa, M. J. Garson, R. J. W. Truscott, J. A. Carver,
Phytochemistry 1993, 34, 1159–1163. For proline-specific Mail-
lard compounds, see: k) C.-W. Chen, G. Lu, C.-T. Ho, J. Agric.
Food Chem. 1997, 45, 2996–2999. For decahydrocyclopenta[b]-
azepines, for example Drupacine, isolated from Cephalotaxus
harringtonia, see: l) R. G. Powell, R. V. Madrigal, C. R.
Smith Jr., K. L. Mikolajczak, J. Org. Chem. 1974, 39, 676–679.
Sessilistemonamine F, from Stichoneuron caudatum, see: m)
R. A. Ramli, W. Lie, S. G. Pyne, J. Nat. Prod. 2014, 77, 894–
901. For cyclopenta-fused pyrroline rings, for rexample roseo-
philin, isolated from S. griseoviridis, see: n) A. Fürstner, Angew.
Chem. Int. Ed. 2003, 42, 3582–3603; Angew. Chem. 2003, 115,
[4] For homogeneous AuIII-catalyzed Nazarov reactions, see: a) T.
Vaidya, R. Cheng, P. N. Carlsen, A. J. Frontier, R. Eisenberg,
Org. Lett. 2014, 16, 800–803; b) M. E. Krafft, D. V. Vidhani,
J. W. Cran, M. Manoharan, Chem. Commun. 2011, 47, 6707–
6709; c) S. Karmakar, A. Kim, C. H. Oh, Synthesis 2009, 194–
198; d) T. Jin, Y. Yamamoto, Org. Lett. 2008, 10, 3137–3139,
see also ref.[3a]
3954
www.eurjoc.org
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2015, 3943–3956

Gold(I)-Catalyzed Synthesis of Annulated N-Heterocycles
3706. A plethora of natural, cyclopentannulated indole alka-
loids exist; for example Bruceolline J, isolated from Brucea
mollis, see: o) H. Chen, J. Bai, Z.-F. Fang, S.-S. Yu, S.-G. Ma,
S. Xu, Y. Li, J. Qu, J.-H. Ren, L. Li, Y.-K. Si, X.-G. Chen, J.
Nat. Prod. 2011, 74, 2438–2445.
five-membered N-heterocycles. See: M. M. Domostoj, E.
Irwing, F. Scheimann, K. J. Hale, Org. Lett. 2004, 6, 2615–
2618.
[27] The cyclic urethane intermediate was not isolated. Diagnostic
signals are the triplet at δ = 4.71 ppm and the singlet at δ =
5.56 ppm assigned to protons Ha and Hb, respectively. Proton
Hc resonates at δ = 5.15 ppm as a triplet:
[10]
Cephalotaxine synthesis: a) S. H. Kim, J. K. Cha, Synthesis
2000, 2113–2116. For formal and total syntheses of roseophilin
by the Nazarov reaction, see: b) D. J. Kerr, L. B. Flynn, Org.
Lett. 2012, 14, 1740–1743; c) C. Song, H. Liu, M. Hong, Y.
Liu, F. Jia, L. Sun, Z. Pan, J. Chang, J. Org. Chem. 2012, 77,
704–706; d) M. Hong, H. Liu, L. Sun, F. Jia, Y. Liu, Q. Jiang,
C. Song, Synlett 2011, 2995–2996; e) A. Y. Bitar, A. J. Frontier,
Org. Lett. 2009, 11, 49–52; f) E. G. Occhiato, C. Prandi, A.
Ferrali, A. Guarna, J. Org. Chem. 2005, 70, 4542–4545; g) P. E.
Harrington, M. A. Tius, J. Am. Chem. Soc. 2001, 123, 8509–
8514.
[8a]
[28] In this particular case, monitoring the reaction by TLC re-
vealed the formation of two spots corresponding to the two
acetate isomers and only one of this disappeared to form a
more polar spot corresponding to Nazarov product 20d.
[29] a) C. Lee, W. Yang, R. G. Parr, Phys. Rev. B 1988, 37, 785–
789; b) A. D. Becke, J. Chem. Phys. 1993, 98, 5648–5652; c) W.
Kohn, A. D. Becke, R. G. Parr, J. Phys. Chem. 1996, 100,
12974–12980.
[30] 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, G. Zheng, J. L. Son-
nenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hase-
gawa, 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. Starov-
erov, T. Keith, 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. Strat-
mann, 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 09, revision D.01, Gaussian, Inc., Wall-
ingford CT, 2013.
[31] a) P. J. Hay, W. R. Wadt, J. Chem. Phys. 1985, 82, 270–283; b)
W. R. Wadt, P. J. Hay, J. Chem. Phys. 1985, 82, 284–298; c) P. J.
Hay, W. R. Wadt, J. Chem. Phys. 1985, 82, 299–310.
[32] This species could also be considered as a gold-carbene, al-
though it should be more adequately described as a gold-stabi-
lized carbocation, see: a) A. Fürstner, L. Morency, Angew.
Chem. Int. Ed. 2008, 47, 5030–5033; Angew. Chem. 2008, 120,
5108. See also ref.[3g] and references cited therein.
[33] Interestingly, this [1,2]-H shift process has been postulated to
be the rate-limiting step in the tandem [3,3]-rearrangement/Na-
zarov reaction of simpler enynyl acetates in the gas phase.[5b]
[34] Interestingly, when the reaction of 16a was carried out in anhy-
drous DCM, we observed a significant decrease in the conver-
sion rate of the substrate into any of the acetate intermediates
accompanied by the formation of degradation products.
[35] The results for the water assisted process are shown in the Sup-
porting Information, and the conclusions are in total agree-
ment with the ones withdrawn from Figure 3. With water as a
base, the process leading to VII is favoured by 0.6 kcal/mol.
[36] For the most recent studies on the torquoselectivity in Nazarov
reactions, see: a) T. D. R. Morgan, L. M. LeBlanc, G. H.
Ardagh, R. J. Boyd, D. J. Burnell, J. Org. Chem. 2015, 80,
1042–1051; b) B. L. Flynn, N. Manchala, E. H. Krenske, J. Am.
Chem. Soc. 2013, 135, 9156–9163, and references cited therein,
see also ref.[2b]
[11]
[12]
We used the procedure reported by us in ref. for simple alk-
ynes.
[8a]
As with simpler enynes,
compounds 16 decompose when
neat, but they can be stored for months in solution of the chro-
matographic eluant containing 1% Et3N. Because of their in-
stability when neat, we cannot provide elemental analysis for
these compounds.
[13]
[14]
This was the best solvent for the corresponding carbacyclic
substrates, see ref.[5a]
Monitoring the reaction by TLC we observed the formation of
only one spot which corresponded to both acetates. Acetate
17a and 19a are easily distinguished by 1H NMR as in 17a the
bridgehead proton 4a-H can be identified at 2.21–2.30 ppm
and the olefinic 6-H proton resonates at δ = 6.03 ppm as a
singlet. In 19a this proton is more shielded and resonates at δ
= 5.88 ppm as a doublet (with a very low coupling constant)
because of the coupling with 5-H (which is a multiplet at 2.98–
2.92 ppm).
The Ph3PAuOTf catalyst was the best in the oxyauration of
the triple bond in N-Boc-protected 6-alkynyl-3,4-dihydro-2H-
pyridines, see ref.[8a]
Degradation of the active gold(I) catalyst has been report to
increase with the relative amount of water in organic solvents.
M. Kumar, J. Jasinski, G. B. Hammond, B. Xu, Chem. Eur. J.
2014, 20, 3113–3119.
[15]
[16]
[17]
[18]
No aqueous work up, only filtration through Celite.
Adventitious water was also responsible for the hydrolysis of
acetate 4 in carbacyclic systems (Scheme 1), see ref.[5a]
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. 2012, 134, 9012–9019.
[19]
[20]
[21]
M. Kumar, G. B. Hammond, B. Xu, Org. Lett. 2014, 16, 3452–
3455.
An opposite result has been recently reported for the [3,3]-re-
arrangement of propargyl ester derivatives, see: L.-J. Wang, H.-
T. Zhu, A.-Q. Wang, Y.-F. Qiu, X.-Y. Liu, Y.-M. Liang, J. Org.
Chem. 2014, 79, 204–212.
For the silver-catalyzed [3,3]-rearrangement of propargyl esters
see: Silver in Organic Chemistry (Ed.: M. Harmata), John
Wiley & Sons, 2010, p. 402.
A. Guérinot, W. Fang, M. Sircoglou, C. Bour, S. Bezzenine-
Lafollée, V. Gandon, Angew. Chem. Int. Ed. 2013, 52, 5848–
5852; Angew. Chem. 2013, 125, 5960.
In the hydrolysis of 17a with 1% TfOH the shift was more
pronounced as we obtained a mixture of 18a and 20a in 2:1
ratio.
[22]
[23]
[24]
[25]
This degradation could be due to the acidity of the bridgehead
proton in acetates 17. Deprotonation at that position leads to
an aromatic cyclopentadienyl anion and from this, probably, to
various unidentified byproducts during work-up.
[26]
Two experiments were performed in MeOH with 0.8 equiv. of [37] See Supporting Information for the preparation of all lactams.
MeONa and with 2 equiv. of DBU in THF. Isomerization has
been instead successfully performed with the corresponding
carbacyclic systems (see ref.[5a]) and it has been reported for
[38] We noticed that the solution maintains its initial orange colour
(whereas it turns blackish when Ph3P is used as the ligand),
which is probably accounted for by a greater stability of the
Eur. J. Org. Chem. 2015, 3943–3956
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
3955

E. G. Occhiato et al.
FULL PAPER
cationic gold complex. The same ligand was tested with model
compound 16a but the regioisomeric ratio did not change com-
pared to Ph3P ligand.
[40] T. Luker, H. Hiemstra, W. N. Speckamp, J. Org. Chem. 1997,
62, 8131–8140.
[41] X.-H. Xu, X. Wang, G.-K. Liu, E. Tokunaga, N. Shibata, Org.
Lett. 2012, 14, 2544–2547.
[39] Clearly our calculations overestimate the kinetic preference for
XIV (by 3 kcal/mol), which is not experimentally so high.
Received: April 10, 2015
Published Online: May 15, 2015
3956
www.eurjoc.org
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2015, 3943–3956