Combining Zn ion catalysis with homogeneous gold catalysis: An efficient annulation approach to N-protected indoles

Article abstract of DOI:10.1039/c2sc21333h

The Fischer indole synthesis is perhaps the most powerful method for indole preparation, but it often suffers from low regioselectivities with unsymmetrical aliphatic ketone substrates and strongly acidic conditions and is not suitable for α,β-unsaturated ketones. In this edge article, we disclose an efficient synthesis of N-protected indoles from N-arylhydroxamic acids/N-aryl-N-hydroxycarbamates and a variety of alkynes via cooperative gold and zinc catalysis. The zinc catalysis is similar to the related zinc ion catalysis in metalloenzymes such as human carbonic anhydrase II and substantially enhances the O-nucleophilicity of N-acylated hydroxylamine by forming the corresponding Zn chelates. The Zn chelates can attack gold-activated alkynes to form O-alkenyl-N-arylhydroxamates, which can undergo facile 3,3-sigmatropic rearrangements and subsequent cyclodehydrations to yield N-protected indole products. This new chemistry offers several important improvements over the Fischer indole synthesis: (a) the reaction conditions are mildly acidic and can tolerate sensitive groups such as Boc; (b) broader substrate scopes including substrates with pendant carbonyl groups (reactive in the Fischer chemistry) and alkyl chlorides; (c) better regioselectivities for the formation of 2-substituted indoles under much milder conditions; (d) 2-alkenylindoles can be prepared readily in good to excellent yields, for which Fischer chemistry could not be used; (e) with internal alkynes both steric and electronic controls are available for achieving good regioselectivities, while Fischer chemistry is in general problematic. The Royal Society of Chemistry 2013.

Full text of DOI:10.1039/c2sc21333h

Chemical Science
View Article Online
View Journal
EDGE ARTICLE
Combining Zn ion catalysis with homogeneous gold
catalysis: an efficient annulation approach to
N-protected indoles†
Cite this: DOI: 10.1039/c2sc21333h
*
Yanzhao Wang, Lianzhu Liu and Liming Zhang
The Fischer indole synthesis is perhaps the most powerful method for indole preparation, but it often
suffers from low regioselectivities with unsymmetrical aliphatic ketone substrates and strongly acidic
conditions and is not suitable for a,b-unsaturated ketones. In this edge article, we disclose an efficient
synthesis of N-protected indoles from N-arylhydroxamic acids/N-aryl-N-hydroxycarbamates and a variety
of alkynes via cooperative gold and zinc catalysis. The zinc catalysis is similar to the related zinc ion
catalysis in metalloenzymes such as human carbonic anhydrase II and substantially enhances the O-
nucleophilicity of N-acylated hydroxylamine by forming the corresponding Zn chelates. The Zn chelates
can attack gold-activated alkynes to form O-alkenyl-N-arylhydroxamates, which can undergo facile 3,3-
sigmatropic rearrangements and subsequent cyclodehydrations to yield N-protected indole products.
This new chemistry offers several important improvements over the Fischer indole synthesis: (a) the
reaction conditions are mildly acidic and can tolerate sensitive groups such as Boc; (b) broader substrate
scopes including substrates with pendant carbonyl groups (reactive in the Fischer chemistry) and alkyl
chlorides; (c) better regioselectivities for the formation of 2-substituted indoles under much milder
conditions; (d) 2-alkenylindoles can be prepared readily in good to excellent yields, for which Fischer
chemistry could not be used; (e) with internal alkynes both steric and electronic controls are available
for achieving good regioselectivities, while Fischer chemistry is in general problematic.
Received 23rd August 2012
Accepted 7th November 2012
DOI: 10.1039/c2sc21333h
www.rsc.org/chemicalscience
On the other hand, analogous indole syntheses based on 3,3-
rearrangements of O-alkenyl-N-arylhydroxylamines or its
Introduction
The Fischer indole synthesis1 is arguably the most versatile
method for indole synthesis and has been subjected to exten-
sive modications/improvements2 since the original report by
Fischer’s group3 more than one hundred years ago. This one-
step annulation between an arylhydrazine and a ketone under
strongly acidic conditions relies on a key 3,3-sigmatropic rear-
rangement of the N-alkenyl-N0-arylhydrazine intermediate, and
many syntheses of indole alkaloids have used it as the key
transformation.4 Notwithstanding, its synthetic potential is
limited by several drawbacks associated with the formation of
N-alkenylhydrazine precursors via hydrazone tautomerization;
the most notable is the oen poor regioselectivities with
unsymmetrical aliphatic ketones. In addition, 2-alkenylindoles
cannot be prepared via this method except for a few special
cases.5 While catalytic hydroamination of alkynes has been
developed as an alternative approach to accessing N-arylhy-
drazones, the same drawbacks remain, and the reported reac-
tion scopes are mostly limited to arylalkynes.6
derivatives7 such as the Bartoli indole synthesis8 have been
developed; while these rearrangements involving fragmentations
of weaker N–O bonds typically proceed at much lower tempera-
tures than in the Fischer indole synthesis, there is a lack of a
general and straightforward method for the generation of O-
alkenylhydroxyamine intermediates.7 For the Bartoli reaction, it is
severely limited by the requirement of an ortho substituent in the
nitroarene substrate, and its functional group tolerance is rather
poor due to the use of excess Grignard reagents (at least 3
equivalents).
Department of Chemistry and Biochemistry, University of California, Santa Barbara,
CA 93106, USA. E-mail: zhang@chem.ucsb.edu
† Electronic supplementary information (ESI) available: Experimental procedures
and NMR spectra for all the new compounds. See DOI: 10.1039/c2sc21333h
Scheme 1 Regiospecific formation of 2-alkylindoles from aliphatic terminal
alkynes via gold-catalyzed O-additions of hydroxylamines to C–C triple bonds.
This journal is ª The Royal Society of Chemistry 2012
Chem. Sci.

View Article Online
Chemical Science
Edge Article
We9 reported previously that this type of intermediate (i.e., 1, 6 h. Instead, the hydration product, methyl ketone 4, became
Scheme 1) could be formed via the addition of the HO group of signicant over time (Table 1, entry 1). This result conrmed
an N-arylhydroxylamine onto a terminal aliphatic alkyne in the our initial fear that heating the reaction could not compensate
presence of a gold catalyst.10,11 Followed by one-pot sequential the diminished nucleophilicity of 2a. Further increasing the
3,3-rearrangement and dehydrative cyclization, this gold catal- reaction temperature led to gold decomposition, and variation
ysis affords 2-alkylindole products at ambient temperature. Two of other reaction conditions did not offer much improvement.
notable features of this method are the exceedingly mild reac- We concluded that a new strategy had to be developed to
tion conditions (tolerance of a THP ether) and the excellent dramatically improve this chemistry.
regioselectivities. In comparison, the Fischer synthesis requires
Inspired by the role of metal ions in enhancing the nucleo-
heating and harsh acidic conditions (e.g., 5% P2O5 per neat philicity of H2O by forming metal hydroxides in metalloenzyme
MsOH12 or neat PPA13) to convert methyl ketones into 2-alky- catalysis (Scheme 2A),18 we wondered whether the nucleophi-
lindoles, albeit only with low to moderate regioselectivities.
licity of 2a would be enhanced in a similar manner. As shown in
Although this chemistry represents a signicant improve- Scheme 2B, 2a could react with a metal ion to form metal
ment in the synthesis of 2-alkylindoles over the Fischer indole chelate 2a-M and a proton reversibly. Due to the bidentate
synthesis, our attempts to extend it to the synthesis of other nature of the deprotonated 2a, 2a-M should be formed in a
types of indoles were thwarted by inherent slow reactions and signicant amount.19 As in metalloenzyme catalysis, 2a-M
limited thermostabilities of N-arylhydroxylamines.9 The should be more nucleophilic than 2a due to the increased
aliphatic terminal alkynes used in our previous studies appear negative charge on the deprotonated oxygen. The most common
to be the best substrates for the chemistry, but the reactions still metal employed in hydrolytic or hydration metalloenzymes (e.g.,
required 18 h or more to reach completion at room tempera- human carbonic anhydrase II) is Zn2+. To our delight, when
ture. Increasing reaction temperature was detrimental as N- Zn(OTf)2 (5 mol%) was added, the reaction was drastically
arylhydroxylamines easily underwent disproportionation and/ accelerated, and 3a was formed in 62% yield in 6 h (Table 1,
or decomposition. When less reactive internal alkynes, ary- entry 2). Other metal salts (Table 1, entries 3–5), however, did
lalkynes and enynes were employed, the reaction yields were not work nearly as well. Subsequent variation of the gold
<15%. As a result, the strategy outlined in Scheme 1 is very
limited and could only be applied to aliphatic terminal alkynes.
A new synthetic approach to O-alkenyl-N-arylhydroxylamines or
their derivatives is imperative in order to establish a broadly
Table 1 Screening gold catalysts and reaction conditionsa
applicable and improved/complementary alternative to the
Fischer indole synthesis. Additional drawbacks with the pub-
lished work are the need to use excess alkynes (typically 1.8
equiv.) in order to compensate the competing C^C triple bond
hydration and oen difficulties in separating the formed methyl
ketones.
Yieldb
We reasoned that the thermostability of N-arylhydroxyl-
amines can be much improved by converting them into
hydroxamic acids or N-hydroxycarbamates; however, this
modication is at the expense of the nucleophilicity of the HO
group and hence the reaction rates. Although increasing the
reaction temperature might compensate the rate loss, little
improvement over the already long reaction duration could be
expected unless a new design can be introduced. Herein, we
disclose a solution by employing a rare cooperative gold and
zinc catalysis,14 which enables an efficient synthesis of N-pro-
tected indoles by using hydroxamic acids or N-hydrox-
ycarbamates as substrates; notably, the Zn catalysis is similar to
metal ion catalysis in metalloenzymes.
Entry Catalyst
Additive (5 mol%) 3a
4
1
Ph3PAuNTf2
—
2%c 26%
62% 45%
20% 60%
2
3
4
5
6
7
Ph3PAuNTf2
Ph3PAuNTf2
Ph3PAuNTf2
Ph3PAuNTf2
(4-CF3Ph)3PAuNTf2
IPrAuNTf2
Zn(OTf)2
Sc(OTf)3
Cu(OTf)2
Dy(OTf)3
Zn(OTf)2
Zn(OTf)2
Zn(OTf)2
Zn(OTf)2
Zn(OTf)2
Zn(OTf)2
4%
54%
23% 62%
77% 37%
40%c 60%
91%f 36%
90% 38%
41%c 11%
83% 39%
0%c 1%
d
e
8
9
(ArO)3PAuNTf2
(ArO)3PAuOTfd
Et3PAuNTf2
e
10
11
12
13
14
15
16
17
18
BrettphosAuNTf2
Dichloro(2-picolinato)gold(III) Zn(OTf)2
Ph3PAuCl/AgClO4
PtCl2
Zn(OTf)2
Zn(OTf)2
Zn(OTf)2
20%c 21%
5%c 24%
Results and discussion
—
—
—
4%c 76%
6%c 64%
48%c 27%
d
g
(ArO)3PAuNTf2
(ArO)3PAuNTf2
(ArO)3PAuNTf2
—
d
HOTf (10%)h
We began the reaction condition studies by using tert-butyl N-
hydroxy-N-phenylcarbamate 2a as the substrate. While 2a and
its congeners in this work were prepared via sequential nitro-
arene reduction and protection,15 they could also be prepared
via Cu-16 or Pd-catalyzed17 cross coupling reactions. In the
presence of a slight excess of 1-dodecyne and catalytic
Ph3PAuNTf2, little indole 3a was formed in toluene at 60 ꢀC aer
d
e,i
Zn(OTf)2
a
Reaction run in oven-dried vials with anhydrous toluene; [2a] ¼ 0.1 M.
b
Estimated by 1H NMR using diethyl phthalate as the internal
c
reference; to be consistent, the yield of 4 is also based on 2a. 2a le.
d
e
f
Ar ¼ 2,4-di-tert-butylphenyl. Reaction time: 2 h. 90% isolated
g
h
i
yield. Reaction time: 12 h. Reaction time: 3 h. NaHCO3 (1 equiv.)
added; reaction time: 12 h; 51% of the alkyne remained.
Chem. Sci.
This journal is ª The Royal Society of Chemistry 2012

View Article Online
Edge Article
Chemical Science
led to 25% yield of the desired product, which is attributed to
sterics but can be complemented by the Bartoli reaction. A meta-
Me group led to a good combined yield, albeit with a low
regioselectivity (entry 14). While this chemistry provides indoles
with the nitrogen protected by readily removable Boc groups,
other groups such as methoxycarbonyl (entries 12, 14 and 16)
and acetyl (entries 15 and 17) can also be incorporated. Since
protection of the indole NH group is oen required in multi-
step synthesis and acyl or alkoxycarbonyl groups are frequently
used for this purpose, this chemistry, directly offering such N-
protected indoles, could prove to be advantageous in the total
synthesis of indole alkaloids.
Although the above studies have rmly established that this
novel dual Zn/Au catalysis is better than our previous study in
the synthesis of 2-alkylindoles in terms of reaction scope,
synthetic efficiency, reaction time and ease of purication, a
categorical superiority by the new approach would be ascer-
tained by expanding the reaction scope beyond what the
previous method would achieve. With this in mind, we exam-
ined terminal enynes and internal alkynes, both of which were
not suitable substrates in our previous work.
Scheme 2 Enhancing the nucleophilicity of 2a by a metal ion.
catalyst (entries 6–13) showed that (ArO)3PAuNTf2 (Ar ¼ 2,4-di-
tert-butylphenyl)20 gave the best yield of the desired product
(Table 1, entry 8). While the related gold triate was nearly as
effective (entry 9), PtCl2 was ineffective (Table 1, entry 14), and
no reaction was detected without gold (entry 15). The impor-
tance of Zn(OTf)2 was again conrmed as only a little product
was formed in its absence (Table 1, entry 16) or by replacing it
with HOTf (10 mol%) (Table 1, entry 17). The latter result
indicates that the Brønsted acid generated in Scheme 2B is not
responsible for the dramatic rate acceleration, but it is crucial
for catalyst turnover as the reaction was substantially slowed in
the presence of NaHCO3 (1 equiv., entry 18).
In the case of terminal enyne substrates, 2-alkenylindoles
were the expected products. Notably, they could not be prepared
via the Fischer indole synthesis,5 either, as the ketone
substrates, a,b-unsaturated ketones, typically react with aryl-
hydrazines to form pyrazolines.23
It is noteworthy that only 5 mol% of Zn(OTf)2 was needed,
therefore making this reaction an interesting cooperation
between a gold catalysis and a Zn catalysis. Notably, this type of
cooperative dual metal catalysis involving gold is rare.21
The scope of the reaction was then studied. As shown in
Table 2, various functionalized aliphatic terminal alkynes were
readily allowed, and the desired 2-substituted indoles 3 were
formed regiospecically in mostly good to excellent yield
(entries 1–10). Comparing to our previous chemistry based on
N-arylhydroxylamines, this method requires shorter reaction
time (2–8 h vs. ꢁ20 h) and less alkyne (1.4 equiv. vs. 1.8 equiv.),
generally results in higher yields and allows much easier
product purication. Albeit the raised reaction temperature
To our delight, with this cooperative Au/Zn catalysis, these
synthetically useful diene-type indoles24 were readily prepared
by using terminal enynes as substrates. As shown in Table 3,
different alkenyl groups (entries 1–5) including sterically
demanding cyclohexenyl (entry 1) and conjugated b-styryl (entry
3) were smoothly installed at the indole 2-position. Similar to
aliphatic terminal alkynes, the reaction was regiospecic, and
no 3-alkenylindole was observed. In entry 5, the increased
product E/Z ratio might be due to acid-promoted isomerization.
Both electron-withdrawing and electron-donating substituents
on the benzene ring were readily allowed (entries 6–11), and
other nitrogen protecting groups such as methoxycarbonyl
(entries 8, 10 and 12) and acetyl (entries 2, 11 and 13) could
easily replace the Boc group. In the cases of N-methoxycarbonyl
groups, 1,2-dichloroethane was a better solvent. Of note, for all
the cases involving more sterically demanding cyclohexenyl
groups, the cyclodehydrative indole formation did not complete
ꢀ
(60 C) and the use of Lewis acid Zn(OTf)2, the reaction condi-
tions were still very mild as the Boc group was not touched. In
comparison, the Fischer method, in addition to the aforemen-
tioned issues (i.e., harsh acidic conditions and moderate
regioselectivities), could not afford the N-unprotected forms of
chloroindole 3f (entry 5), due to the interference of the chloro
group,22 and may not be applicable to the synthesis of indole 3h
(entry 7), due to likely elimination of benzyl alcohol from the
corresponding methyl ketone substrate under strongly acidic
conditions.
To our delight, substituents on the benzene ring para to the
hydroxylamine moiety can vary from electron-withdrawing ones
(entries 11 and 12) to electron-donating ones (entries 13–15)
with good to excellent reaction efficiencies. With an electron-
donating p-MeO group, the N-Boc hydroxylamine was labile at
the reaction temperature, and the yield was only 40%; however,
swapping the Boc group with an acetyl provided sufficient
substrate stability, and the yield was improved to 80% in the
presence of 10% Zn(OTf)2 (entry 15). Notably, the corresponding
free hydroxylamine, N-(4-methoxyphenyl)hydroxylamine, is
labile even at ambient temperature and, therefore, could not be
employed in our previous indole chemistry.9 An ortho-Me group
ꢀ
until further heating at 80 C for 4–8 h.
An inherent issue with the Fischer indole synthesis is the low
regioselectivity25 with unsymmetric ketones with the following
generic formulae, R1CH2C(O)CH2R2 (R1, R2 s H, Ar). To probe
whether our chemistry offers a solution to this long standing
problem, we examined various internal alkynes by using N-
hydroxy-N-phenylacetamide (2b) as the substrate. The results
are shown in Table 4. First of all, internal alkynes with moderate
steric demand were allowed, but due to their decreased reac-
tivities compared to the terminal ones, the reactions were
slower (18–30 h) even with increased amounts of Zn(OTf)2
(20 mol%); moreover, much more thermostable IPrAuOTf,
though less reactive, was a better catalyst than (ArO)3PAuNTf2 as
the reaction could not proceed to completion with the latter as
This journal is ª The Royal Society of Chemistry 2012
Chem. Sci.

View Article Online
Chemical Science
Edge Article
Table 2 Formation of N-protected 2-alkylindolesa
1
2
5
8
3
6
9
3b, 83%
3c, 89%
3f, 87%
3i, 78%
3d, 84%
4
3e, 96%
3g, 95%
3j, 93%
7
3h, 83%
10
11
12
3k, 82%
3l, 94%
3m, 80%
13
14b
3n, 91%
3o, 78% (1.25 : 1)
15b,c
16b,d
17
3p, 80%
3q, 81%c
3r, 90%
a
Reactions run in oven-dried vials; isolated yields are reported; Ar ¼ 2,4-di-tert-butylphenyl. b 2 equivalents of dodecyne. c 10% Zn(OTf)2. d DCE as
solvent.
the catalyst. With the symmetric 6-dodecyne, the indole product the alkyne (i.e., but-1-ynylcyclohexane) was apparently sterically
6a was isolated in 74% yield (entry 1). With methylalkynes too demanding, and the reaction was very sluggish (21% yield).
(entries 2–7), the regioisomers with the methyl group at the 2 With a smaller isobutyl group, the reaction proceeded well
position were always favored, apparently due to its smaller size. (entry 8), and the small steric difference between the two alkyne
With primary groups at the other end, the selectivities were ends was reected by a 2.5/1 regioselectivity, favoring the one
around 6–8 : 1 (entries 2, 4, and 5). Interestingly, a similar with ethyl at the indole 2 position. With differently protected
selectivity was observed with a Ph group (entry 7), suggesting hydroxyethyl groups (entries 9–11), the trend of regioselectiv-
that steric control is dominating over electronics. By increasing ities clearly correlated with the inductive effect of the protecting
the steric size of the group (i.e., a cyclohexyl group, entry 3), the groups. Hence, a synthetically useful selectivity (10.3 : 1, entry
selectivity was enhanced to >19 : 1. There is a notable electronic 9) was achieved by using a triuoroacetate substrate. A some-
effect when a strongly inductive PhthN group is present (entry 6, what synergy between sterics and induction resulted in an even
vide infra), and the selectivity and the yield were synthetically better regioselectivity for the indole product 6l (entry 12). It is
desirable. While the above products could be obtained selec- important to note that the triuoroacetyl group can be easily
tively by the Fischer indole synthesis using the corresponding removed, therefore making this chemistry a convenient and
methyl ketones in most cases, the successful formation of the regioselective approach to various 2-substituted indole-3-etha-
ketone product 6d shows another example of functional group nols. In comparison, a recent gold catalysis using phenyl-
tolerance offered by this chemistry but unattainable by the hydrazine and hex-3-yn-1-ol under strong conditions (TsOH,
Fischer approach.
toluene, 100 ꢀC) led to the fully deprotected 6i as a minor
For internal alkynes with one end substituted by a primary isomer,26 and the Fischer method would most likely offer a
alkyl group, the one with a cyclohexyl group on the other end of similar result. These results demonstrated that unattainable
Chem. Sci.
This journal is ª The Royal Society of Chemistry 2012

View Article Online
Edge Article
Chemical Science
Table 3 Formation of N-protected 2-alkenylindolesa
3c
1
2b,c
5a, 83%
5b, 68%
5e, 86%
5c, 74%
5f, 75%
4c
5c,d
6
9
5d, 90%
7
8e
5g, 65%
5h, 77%
5i, 80%
5k, 75%
10e
11b,f
5j, 71% (1.5 : 1)
12e
13
5l, 86%
5m, 92%
a
Reactions run in oven-dried vials; isolated yields are reported; Ar ¼ 2,4-di-tert-butylphenyl. b 2 equiv. of the enyne. c The reaction was run at 60 ꢀC
for 12 h. The enyne E/Z ¼ 9 : 1, and the product E/Z > 19 : 1. e Solvent: DCE. f 10 mol% of Zn(OTf)2.
d
regioselectivities in the Fischer chemistry could be realized by the difference of the natural charges (NC) of the alkyne ends
judicial choices of protecting groups in this dual metal computed by using DFT calculations (B3LYP/6-31G*) is 0.22
catalysis.
with R ¼ Et, Fig. 1B], which is also reected by the 13C chemical
Compared to the methyl group in entry 7, the use of an Et shis.28 The electronic difference totally overrides the steric
group in entry 13 expectedly led to a decreased regioselectivity, disparity, resulting in an exclusive nucleophilic attack at the
although the Ph group at the 3 position was still favored. A more more electrophilic internal end. For alkynes with conjugated
electronically biased triple bond in entry 14 led selectively to the electron-withdrawing groups, the polarization of the triple bond
regioisomer following an initial Michael-type addition.27
is again signicant; for example, in Fig. 1C, DNC and Dd(13C)
A reaction mechanism is proposed in Scheme 3. First, the between the two sp-hybridized carbons are 0.249 and 15.33
hydroxamic acid or related acylated substrate 2 is activated via ppm, respectively; the electronic difference, quantitatively
the formation of the zinc chelate 2-Zn. This more O-nucleo- similar to that in the terminal alkyne 1-butyne (Fig. 1B), is
philic intermediate can then attack the gold-activated alkyne consistent with the observed excellent selectivity. For those with
effectively, forming the precursor (i.e., 7) for the 3,3-sigmatropic non-conjugated electron-withdrawing groups, the oen over-
rearrangement upon subsequent protodeauration. Aer the looked inductive effects29 can be signicant. Such an induction
rearrangement, a subsequent cyclodehydration, most likely results in less electron density at the distal end of the C^C
facilitated by the Lewis acidic Zn(OTf)2, leads to the desired triple bond than at the proximal end, which is conrmed by the
indole product.
calculated NCs and the observed 13C chemical shis of the
The observed regioselectivities are rationalized in Fig. 1: rst alkyne substrates for entries 6 and 9–11 in Table 4 (Fig. 1D).
of all, the zinc hydroxamate 2-Zn is sterically bulky, likely due to This electronic bias predicts correctly the selective attack by the
additional ligands on the metal, and preferably attacks the less nucleophilic hydroxamate at the distal end in the absence of
hindered end of internal alkynes (Fig. 1A). With terminal signicant steric disparity and, moreover, correlates in a
alkynes including enynes, the internal end of the C^C triple roughly quantitative manner with the regioselectivity observed
bond is much less negatively charged than the terminus [e.g., in entries 9–11,Table 4.
This journal is ª The Royal Society of Chemistry 2012
Chem. Sci.

View Article Online
Chemical Science
Edge Article
Table 4 Regioselective formation of N-acetyl-2,3-disubstituted indolesa
1
2
3
6a, 74%
6b, 61% (7.4 : 1)b
6e, 52% (6 : 1)b
6h, 75% (2.5 : 1)c,d
6k, 86% (2 : 1)c
6c, 67% (>19 : 1)b
4
5
6
6d, 68% (6.8 : 1)b
6g, 61% (8 : 1)b
6j, 77% (5.6 : 1)c
6f, 80% (16.6 : 1)b
6i, 65% (10.3 : 1)c
6l, 69% (12 : 1)c
7
8
9
10
11
12
13
14
15
6m, 70% (2.9 : 1)c
6n, 66% (>20 : 1)
a
Reactions run in oven-dried vials; isolated yields of the major isomers are reported if not specied. b Regioselectivity; the major isomer is shown.
Overall yields of both isomers are reported. 2 equiv. of 6-methylhept-3-yne was used.
c
d
Scheme 3 Proposed reaction mechanism.
To substantiate the notion that catalytic Zn(OTf)2 enhances
the nucleophilicity of N-arylhydroxamic acids/N-aryl-N-hydrox-
ycarbamates, we probed the competition between the addition
of 2a to 1-dodecyne and its hydration by varying the amount of
the metal salt. As shown in Table 5, the reactions were stopped
aer 30 min in order to glean early stage rate differences. When
toluene saturated with H2O was used as the solvent, there was a
clear correlation between the amount of Zn(OTf)2 and the ratio
of products due to 2a addition (i.e., 3a and 8) and hydration (i.e.,
4), suggesting that Zn(OTf)2 increases the rate of 2a addition
Fig. 1 Rationales for the reaction regioselectivity on alkynes.
(entries 1–5). Notably, the amount of the hydration product,
methyl ketone 4, did not change much as the amount of the zinc
salt increased, indicating that hydration was not much affected
by the salt. Under anhydrous conditions, water was generated
upon the cyclodehydration of 8. Again, a positive correlation
was observed when the amount of Zn(OTf)2 was varied from 0%
Chem. Sci.
This journal is ª The Royal Society of Chemistry 2012

View Article Online
Edge Article
Chemical Science
a
Table 5 Examination of the role of Zn(OTf)2
protecting groups, good to excellent regioselectivities with
internal alkynes can be achieved. Importantly, this chemistry is
a rare example of cooperative dual catalysis involving gold and
the rst example of gold-catalyzed intermolecular addition of
hydroxamic acids/N-aryl-N-hydroxycarbamates to alkynes.
Catalytic Zn(OTf)2 enhances the nucleophilicity of these
hydroxylamine derivatives via the formation of deprotonated
chelates, which is similar to the metal ion catalysis in
metalloenzymes.
Yieldb
Entry
Toluene
Zn(OTf)2
3a
4
8
(3a + 8)/4
1
2
3
4
5
6
7
8
9
10
Wetc
0
<1
9%
29%
50%
67%
0
13%
41%
57%
78%
18%
19%
21%
21%
21%
1.4%
7%
9%
11%
17%
<1
—
Acknowledgements
Wetc
1 mol%
5 mol%
10 mol%
20 mol%
0
1 mol%
5 mol%
10 mol%
20 mol%
16%
27%
25%
19%
0
16%
25%
20%
13%
1.3
2.7
3.6
4.1
0
4.1
7.3
7.0
5.4
Wetc
The generous nancial support of NIGMS (R01 GM084254) is
greatly appreciated.
Wetc
Wetc
Anhydrousd
Anhydrousd
Anhydrousd
Anhydrousd
Anhydrousd
Notes and references
1 (a) B. Robinson, The Fischer Indole Synthesis, Wiley,
Chichester, New York, 1982; (b) R. S. Downing and
P. J. Kunkeler, in Fine Chemicals Through Heterogenous
Catalysis, ed. R. A. Sheldon and H. Bekkum, Wiley-VCH,
Weinheim; New York, 2001, pp. 178–183; (c) D. L. Hughes,
Org. Prep. Proced. Int., 1993, 25, 607–632.
a
b
[2a] ¼ 0.1 M; Ar ¼ 2,4-di-tert-butylphenyl. NMR yield estimated by
c
using diethyl phthalate as the internal reference. Saturated with
d
deionized water, and reactions run in vials. Reaction run in Schlenk
tubes.
2 (a) T. Lipin'ska, Chem. Heterocycl. Compd., 2001, 37, 231–236;
(b) C.-Y. Chen, C. H. Senanayake, T. J. Bill, R. D. Larsen,
T. R. Verhoeven and P. J. Reider, J. Org. Chem., 1994, 59,
3738–3741; (c) A. R. Katritzky, S. Rachwal and S. Bayyuk,
Org. Prep. Proced. Int., 1991, 23, 357–363; (d) M. Inman and
C. J. Moody, Chem. Commun., 2011, 47, 788–790; (e)
B. Narayana, B. V. Ashalatha, K. K. Vijaya Raj, J. Fernandes
and B. K. Sarojini, Bioorg. Med. Chem., 2005, 13, 4638–4644.
3 (a) E. Fischer and F. Jourdan, Ber. Dtsch. Chem. Ges., 1883, 16,
2241–2245; (b) E. Fischer and O. Hess, Ber. Dtsch. Chem. Ges.,
1884, 17, 559–568.
(entry 6) to 5 mol% (entry 8); moreover, the preference toward
2a addition was expectedly better than in wet toluene. The
decreased preferences in entries 9 and 10 reect the signicant
increase of the amount of by-product H2O, which leads to
increasing hydration. These results are consistent with the
proposed role of Zn(OTf)2. Expectedly, it, being Lewis acidic,
also facilitated the cyclodehydration (comparing the ratios of 8/
3a); notably, the gold catalyst can also promote the process,
albeit much slower (93% 8 converted inꢀto 3a in the presence of 5
mol% (ArO)3PAuNTf2 in toluene at 60 C for 6 h).
4 (a) J. Bonjoch, J. Catena and N. Valls, J. Org. Chem., 1996, 61,
7106–7115; (b) R. Iyengar, K. Schildknegt and J. Aube, Org.
Lett., 2000, 2, 1625–1627; (c) C. W. Roberson and
K. A. Woerpel, J. Am. Chem. Soc., 2002, 124, 11342–11348;
(d) T. Gan, R. Liu, P. Yu, S. Zhao and J. M. Cook, J. Org.
Chem., 1997, 62, 9298–9304; (e) H. Ueda, H. Satoh,
K. Matsumoto, K. Sugimoto, T. Fukuyama and
H. Tokuyama, Angew. Chem., Int. Ed., 2009, 48, 7600–7603.
5 J. Bergman and B. Pelcman, Tetrahedron, 1988, 44, 5215–
5228.
Conclusions
We have developed an efficient and general synthesis of N-
protected indoles via the annulation of N-arylhydroxamic acids/
N-aryl-N-hydroxycarbamates and alkynes. It is synthetically
superior to our previous chemistry based on N-hydroxylamines.
The reaction scope is broad, and the reaction conditions are
mild. The indole nitrogens are substituted with acyl or alkox-
ycarbonyl groups, which could be advantageous as these groups
oen offer desirable N-protection in complex molecule
synthesis. With terminal alkynes, this reaction is regiospecic
and offers 2-substituted indoles and tolerates a range of func-
tional groups. Notably, various 2-alkenylindoles, useful
substrates for the Diels–Alder reaction as dienes but not
accessible via the Fischer indole synthesis, can be readily
prepared from 3-en-1-ynes. This chemistry also works with
internal alkynes of moderate steric demand, and the regiose-
lectivity can be controlled both by sterics and/or by electronics:
the small substituent on the C^C triple bond ends up selec-
tively at the indole 2-position, and the electron-withdrawing
groups direct the substituents to the indole 3 positions via
either induction or conjugation. With judicial choices of
6 (a) C. Cao, Y. Shi and A. L. Odom, Org. Lett., 2002, 4, 2853–
2856; (b) L. Ackermann and R. Born, Tetrahedron Lett.,
2004, 45, 9541–9544.
7 J. A. Joule, Sci. Synth., 2000, 10, 380–383.
8 (a) G. Bartoli, R. Leardini, A. Medici and G. Rosini, J. Chem.
Soc., Perkin Trans. 1, 1978, 692–696; (b) G. Bartoli,
G. Palmieri, M. Bosco and R. Dalpozzo, Tetrahedron Lett.,
1989, 30, 2129–2132; (c) G. Bartoli, M. Bosco, R. Dalpozzo,
G. Palmieri and E. Marcantoni, J. Chem. Soc., Perkin Trans.
1, 1991, 2757–2761; (d) R. Dalpozzo and G. Bartoli, Curr.
Org. Chem., 2005, 9, 163–178.
9 Y. Wang, L. Ye and L. Zhang, Chem. Commun., 2011, 47,
7815–7817.
This journal is ª The Royal Society of Chemistry 2012
Chem. Sci.

View Article Online
Chemical Science
Edge Article
¨
10 (a) M. M. A. Pereira, S. Prabhakar and A. M. Lobo, J. Nat.
Prod., 1996, 59, 744–747; (b) J. R. Hwu, H. V. Patel, R. J. Lin
S. Schafer, L. Schultes, F. Rominger and W. Frey, Chem.–
Eur. J., 2012, 18, 10480–10486.
and M. O. Gray, J. Org. Chem., 1994, 59, 1577–1582; (c) 21 (a) M. Egi, Y. Yamaguchi, N. Fujiwara and S. Akai, Org. Lett.,
M. Toyota and K. Fukumoto, J. Chem. Soc., Perkin Trans. 1,
1992, 547–552; (d) M. Toyota and K. Fukumoto,
Heterocycles, 1990, 31, 1431–1433.
2008, 10, 1867–1870; (b) L. Ye and L. Zhang, Org. Lett., 2009,
11, 3646–3649; (c) A. S. Demir, M. Emrullahoglu and
K. Buran, Chem. Commun., 2010, 46, 8032–8034.
11 H.-S. Yeom, E. So and S. Shin, Chem.–Eur. J., 2011, 17, 1764– 22 (a) M. D. Bowman, J. R. Schmink, C. M. McGowan,
1767.
C. M. Kormos and N. E. Leadbeater, Org. Process Res. Dev.,
2008, 12, 1078–1088; (b) T. Benincori, E. Brenna and
F. Sannicolo, J. Chem. Soc., Perkin Trans. 1, 1991, 2139–
2145.
12 D. L. Hughes and D. Zhao, J. Org. Chem., 1993, 58, 228–233.
13 J. Bosch, J. Bonjoch, A. Diez, A. Linares, M. Moral and
M. Rubiralta, Tetrahedron, 1985, 41, 1753–1762.
`
14 G. M. Sammis, H. Danjo and E. N. Jacobsen, J. Am. Chem. 23 F. Sannicolo, Tetrahedron Lett., 1984, 25, 3101–3102.
Soc., 2004, 126, 9928–9929.
15 A. Porzelle, M. D. Woodrow and N. C. O. Tomkinson, Synlett,
2009, 798–802.
16 K. L. Jones, A. Porzelle, A. Hall, M. D. Woodrow and
N. C. O. Tomkinson, Org. Lett., 2008, 10, 797–800.
17 A. Porzelle, M. D. Woodrow and N. C. O. Tomkinson, Org.
Lett., 2009, 11, 233–236.
24 C.-B. Chen, X.-F. Wang, Y.-J. Cao, H.-G. Cheng and
W.-J. Xiao, J. Org. Chem., 2009, 74, 3532–3535.
25 (a) M. P. Prochazka, L. Eklund and R. Carlson, Acta Chem.
Scand., 1990, 44, 610–613; (b) E. S. Balenkova,
E. P. Zakurdaev and V. G. Nenajdenko, Russ. Chem. Bull.,
2008, 57, 2220–2222.
26 N. T. Patil and A. Konala, Eur. J. Org. Chem., 2010, 6831–6839.
18 D. W. Christianson and J. D. Cox, Annu. Rev. Biochem., 1999, 27 (a) P. Martin, Helv. Chim. Acta, 1989, 72, 1554–1582; (b)
68, 33–57.
M. Toyota and K. Fukumoto, Heterocycles, 1990, 31, 1431–
1433.
´
¨
19 D. A. Brown, N. J. Fitzpatrick, H. Muller-Bunz and A. T. Ryan,
Inorg. Chem., 2006, 45, 4497–4507.
28 (a) H. Matsumoto, Y. Hoshino and Y. Nagai, Chem. Lett.,
1982, 1663–1666; (b) C. A. Tsipis, J. Organomet. Chem.,
1980, 187, 427–446; (c) M. Rubin, A. Tromov and
V. Gevorgyan, J. Am. Chem. Soc., 2005, 127, 10243–10249.
´
´
´
´
20 (a) S. Lopez, E. Herrero-Gomez, P. Perez-Galan, C. Nieto-
Oberhuber and A. M. Echavarren, Angew. Chem., Int. Ed.,
2006, 45, 6029–6032; (b) J. H. Teles, S. Brode and
M. Chabanas, Angew. Chem., Int. Ed., 1998, 37, 1415–1418; 29 D. A. Rooke and E. M. Ferreira, Angew. Chem., Int. Ed., 2012,
¨
(c) A. S. K. Hashmi, T. Haffner, W. Yang, S. Pankajakshan,
51, 3225–3230.
Chem. Sci.
This journal is ª The Royal Society of Chemistry 2012