Allenyl azide cycloaddition chemistry. 2,3-Cyclopentennelated indole synthesis through indolidene intermediates

Article abstract of DOI:10.1021/jo900659w

(Chemical Equation Presented) The thermal, photochemical, and photochemical/CuI-mediated cascade cyclizations of a range of substituted 1-(2-azidophenyl)-3-alkenylallenes are described. These reactions provide both 1,2- and 2,3-cyclopentennelated indole products in varying ratios. In most cases, high regioselectivity for the 2,3-annelated isomer can be achieved under the h?/CuI conditions. Computational studies of this multistep reaction support the intermediacy of indolidene intermediates whose electrocyclizations (with or without copper present) define the regioselectivity branch point in the sequence.

Full text of DOI:10.1021/jo900659w

Allenyl Azide Cycloaddition Chemistry. 2,3-Cyclopentennelated
Indole Synthesis through Indolidene Intermediates
Ken S. Feldman,*,† D. Keith Hester II,† Malliga R. Iyer,† Paul J. Munson,†
Carlos Silva Lo´pez,*,‡ and Olalla Nieto Faza‡
Department of Chemistry, The PennsylVania State UniVersity, UniVersity Park, PennsylVania 16802, and
Departmento de Quimica Organica, UniVersidade de Vigo, Lagoas Marcosende,
36200 Vigo, Galicia, Spain
ksf@chem.psu.edu; csilVal@uVigo.es
ReceiVed March 30, 2009
The thermal, photochemical, and photochemical/CuI-mediated cascade cyclizations of a range of substituted
1-(2-azidophenyl)-3-alkenylallenes are described. These reactions provide both 1,2- and 2,3-cyclopen-
tennelated indole products in varying ratios. In most cases, high regioselectivity for the 2,3-annelated
isomer can be achieved under the hν/CuI conditions. Computational studies of this multistep reaction
support the intermediacy of indolidene intermediates whose electrocyclizations (with or without copper
present) define the regioselectivity branch point in the sequence.
SCHEME 1. Cyclization Cascade of
1-(2-Azidophenyl)-3-vinylallenesa
Introduction
Cascade cyclization sequences offer great promise for the
expeditious assembly of polycyclic products. To the extent that
these processes convert readily accessible cascade precursors
to polycyclic products with predictable control of relevant
selectivity issues (i.e., chemoselectivity, regioselectivity, ste-
reoselectivity, enantioselectivity), these transformations may
impact favorably on total synthesis efforts directed toward
complex polycyclic natural products. It is within this context
that the allenyl azide cascade cyclization sequence was devel-
oped, as the potential for accessing cyclopentennelated indole
products from simple ortho-positioned (3-alkenyl)allene phe-
nylazide substrates was apparent, Scheme 1. Preliminary results
documented that species 1 under diverse reaction conditions
rapidly and cleanly proceeded to the regioisomeric indole
products 4a and 4b.1a Furthermore, reaction conditions could
be identified that led to high levels of regioselectivity for the
C-C-bonded indole product 4a.1b Computational tools were
brought to bear on this complex cascade sequence, and the initial
mechanistic picture that emerged from these calculated models
implicated indolidenes 3a and 3b as pivotal intermediates linking
allenyl azide 1 to products 4a and 4b.2 In addition, these in
a The new bonds are indicated in red.
silico models provided a rationale for the regioselectivity (or
lack thereof in certain circumstances) observed.
(1) (a) Feldman, K. S.; Iyer, M. R.; Hester, D. K., II. Org. Lett. 2006, 8,
3113–3116. (b) Feldman, K. S.; Hester, D. K., II; Lo´pez, C. S.; Faza, O. N.
Org. Lett. 2008, 10, 1665–1668.
† Pennsylvania State University.
‡ Universidade de Vigo.
4958 J. Org. Chem. 2009, 74, 4958–4974
10.1021/jo900659w CCC: $40.75  2009 American Chemical Society
Published on Web 05/27/2009

Allenyl Azide Cycloaddition Chemistry
TABLE 1. Simple Vinyl- and Phenyl-Substituted
(o-Azidophenyl)allene Synthesis
SCHEME 2. Representative 2,3-Cyclopentennelated Indole
Natural Products and the Historical Precedent for Invoking
Indolidenes in Indole-Containing Natural Product Synthesis
yield of yield of yield of
12a (%) 13a (%) 14a (%)
entry
R
R1
a
b
c
d
e
f
g
h
i
CH3
H
H
H
H
48
91
73
78
48
48
79
96
82
79
48
48
48
58
55
57
63
67
59
69
62
49
CH2OTBS
(CH2)2OTBS
t-Bu
CH3
CH3
Ph
TMS
H
Ph
CH3
Ph
CH3
CH3
CH3
CH3
CH3
p-(CH3O)C6H4
m-(CH3O)C6H4
p-FC6H4
b
j
k
l
81
c
c
CH3
CH3
m
c
a Isolated, chromatographically purified material. b CH3MgBr/CuI/LiBr
on 12j (R ) Ph). c The allenes 14k-m were not isolated but rather
carried forward to the thermolysis reaction directly. See Scheme 5 for
details.
Potential natural product targets for this developing meth-
odology are plentiful, provided that both stereochemical and
regiochemical requirements can be met by the allenyl azide
cascade cyclization. For example, the Fisher indoles (cf. 5),3
the indole diterpenes (cf. 6),4 and the unique triindole yeast
isolate malasseziacitrin (7)5 all fall under this general structural
class, Scheme 2. Reaction chemistry that evolves from reactive
indolidene intermediates of the type 3 has not been well
described.6 One notable exception involves the proposed
intermediacy of indolidene (or protonated indolidene) species
in the syntheses of ibogaine derivatives and dimeric alkaloids
of the vinblastine series, Scheme 2.7 The fact that this putative
intermediate is reactive (electrophilic) enough to efficiently trap
the sterically hindered and modestly nucleophilic aryl ring of
vindoline suggests that further development of the reaction
chemistry of indolidenes has the potential for advances in several
areas of indole synthesis/functionalization.
several new allenyl azide substrates that probe the limits of the
methodology. In addition, new calculational results that aid in
elucidating mechanistic subtleties are presented; these calcula-
tions examine the “real system” and so extend beyond the model
system calculations reported earlier.
Results and Discussion
The exploratory studies commenced with the preparation of
some simple 1-(2-azidophenyl)-3-alkenylallenes 13 and the
analogous 1-(2-azidophenyl)-3-arylallenes 14, Table 1. The
propargylic acetate displacement chemistry of Krause proved
completely serviceable in this task, using either commercially
available zinc nucleophiles or zinc species generated in situ from
the corresponding Grignard reagent and ZnCl2.8 All of the
allynyl azide substrates prepared in this study are racemic, and
these species are a mixture of diastereomers when a second
stereogenic element is present. The steric character of the
substituent R directly attached to the allene was varied
significantly (CH3 f CH2OTBS f Ph f TMS f t-Bu) in order
to probe the consequences of steric bulk on the partitioning
between 3a/3b and/or 4a/4b. As it transpired, an unexpected
diversion of reactive intermediates with the R ) TMS case (vide
infra) limited the value of this entry in the steric comparison
series. In this series of substrates, the 1-alkenyl position (i.e.,
R1) was subjected to some variation also (H f CH3 f Ph)
with the expectation that R//R1 clashes (cf. 3a/3b) might also
influence the regioselectivity profile of the reaction. The aryl
series 14j-m was designed to probe for electronic effects on
product formation (yield, regioselectivity), but the route il-
lustrated by the conversion of 12 (R ) CH3) into 14 did not
In this full accounting of our work in this area, the scope of
cyclopentennelated indole synthesis is investigated by examining
(2) Lo´pez, C. S.; Faza, O. N.; Feldman, K. S.; Iyer, M. R.; Hester, D. K., II.
J. Am. Chem. Soc. 2007, 129, 7638–7646.
(3) (a) Park, A.; Moore, R. E.; Patterson, G. M. L. Tetrahedron Lett. 1992,
33, 3257–3260. (b) Stratmann, K.; Moore, R. E.; Bonjouklian, R.; Deeter, J. B.;
Patterson, G. M. L.; Shaffer, S.; Smith, C. D.; Smitka, T. A. J. Am. Chem. Soc.
1994, 116, 9935–9942.
(4) Ondeyka, J. G.; Helms, G. L.; Hensens, O. D.; Goetz, M. A.; Zink, D. L.;
Tsipouras, A.; Shoop, W. L.; Slayton, L.; Dombrowski, A. W.; Polishook, J. D.;
Ostlind, D. A.; Tsou, N. N.; Ball, R. G.; Singh, S. B. J. Am. Chem. Soc. 1997,
119, 8809–8816.
(5) Irlinger, B.; Bartsch, A.; Kra¨mer, H.-J.; Mayser, P.; Steglich, W. HelV.
Chim. Acta 2005, 88, 1472–1485.
(6) Gross, G.; Wentrup, C. J. Chem. Soc., Chem. Commun. 1982, 360.
(7) (a) Bu¨chi, G.; Manning, R. E. J. Am. Chem. Soc. 1966, 88, 2532. (b)
Kutney, J. P.; Beck, J.; Bylsma, F.; Cretney, W. J. J. Am. Chem. Soc. 1968, 90,
4504. (c) Potier, P.; Langlois, Y.; Gueritte, F. J. Chem. Soc., Chem. Commun.
1975, 670–671. (d) Kutney, J. P.; Ratcliffe, A. H.; Treasurywala, A. M.;
Wunderly, S. Heterocycles 1975, 3, 639–649. (e) Schill, G.; Priester, C. U.;
Windhovel, U. F.; Fritz, H. Tetrahedron 1987, 43, 3765–3786. (f) Magnus, P.;
Stamford, A.; Ladlow, M. J. Am. Chem. Soc. 1990, 112, 8210–8212. (g)
Yokoshima, S.; Ueda, T.; Kobayashi, S.; Sato, A.; Kuboyama, T.; Tokuyama,
H.; Fukuyama, T. J. Am. Chem. Soc. 2002, 124, 2137–2139. (h) Ishikawa, H.;
Colby, D. A.; Boger, D. L. J. Am. Chem. Soc. 2008, 130, 420–421.
(8) Jansen, A.; Krause, N. Synthesis 2002, 1987–1992.
J. Org. Chem. Vol. 74, No. 14, 2009 4959

Feldman et al.
TABLE 2. Synthesis of More Complex
1-(2-Azidophenyl)-3-alkenylallenes
TABLE 3. Preparation of Functionalized
Cyclohexenyl-Substituted 1-(2-Azidophenyl)-3-(1-cyclohexenyl)allenes
yield of
yield of
entry
R
R1
Ph
Ph
-(CH2)3-
-(CH2)4-
R2
15a (%)
16a (%)
a
b
c
d
H
Ph
H
Ph
H
47
78
49
31
34
38
30
37
H
a Isolated, chromatographically pure material.
yield of
yield of
entry
R
R1
R2
27a (%)
28a (%)
SCHEME 3. Synthesis of Additional (o-Azidophenyl)allenes
a
b
c
d
e
TBS
TIPS
SEM
TBS
TBS
H
H
H
CN
H
H
H
H
H
79
92
63
76
79
92
89
86
81
85
TMS
a Isolated, chromatographically pure material.
SCHEME 4. Preparation of Functionalized Cyclohexenyl
Iodides
afford products that survived attempts at chromatographic (SiO2)
purification. Therefore, the crude allenyl azides 14j-m were
subjected to thermal rearrangement without further purification.
Subsequently, we determined that slightly modifying the
sequence so that a methyl nucleophile was added to a phenyl-
substituted propargyl acetate (12j, R ) Ph) did furnish an allenyl
azide 14j that tolerated passage over SiO2 and could be isolated
as a pure, discrete compound. However, disappointing results
upon thermolysis (vide infra, Scheme 5) did not encourage
revisiting the syntheses of the other aryl substrates 14k-m.
A related series of 1-(2-azidophenyl)-3-alkenylallenes 16 and
23-25 (Table 2 and Scheme 3, respectively) were prepared by
slightly different routes than those described above. For the
synthesis of 15, a preformed alkenyl acetylide anion was
employed to introduce the propargyl acetate functionality, and
then methyl cuprate was utilized as the allene-forming nucleo-
phile. The preparation of 23-25 featured a Sonogashira coupling
of the preformed propargyl acetate 12i with the appropriate
alkenyl iodide fragment 17-19 to deliver the enyne acetates
20-22. The cuprate chemistry was relied upon again to deliver
the allene product. These substrates introduced one (or two)
substituents at the alkene terminus, a position destined to
participate directly in C-C or C-N bond formation (cf. the
red bonds in 4a/4b). In this way, the influence of steric effects
(primarily) and possibly electronic effects can be assessed upon
attempted cyclopentennelated indole formation. The substrate
16b in particular will test the premise that even very challenging
all-carbon quaternary centers might be forged through this
methodology.
for Scheme 3; Sonogashira coupling of 12i with the requisite
cyclohexenyl iodide 26 followed by alkyl cuprate-mediated
propargylic displacement of the acetate function in 27 to deliver
the allene product 28. Since the cuprate required to prepare 37
(see eq 1) was not available, we resorted to the Krause chemistry
with the highly functionalized cyclohexenyl zincate derived from
35a and the propargyl acetate 36, itself prepared from 11 and
the TBS ether of propargyl alcohol. The various cyclohexenyl
iodides 34 and 35a-c were synthesized from the commercially
available dione 31 using routine and well-documented chemistry,
Scheme 4.9 The free hydroxyl substrate 29 was conveniently
available by desilylation of 28a. The ketone 30, prepared via
oxidation of 29, was synthesized as well. The rationale behind
the choice of these cyclohexenyl substrates was 2-fold; (1) the
relationship between substrate structure and product regiochem-
istry was still underdeveloped at this early juncture of the
research, and so probing the influence of allylic oxygen
substitution (sterics, electronics?) on C-C vs C-N bond
formation seemed appropriate, and (2) for the first time, the
issue of relative stereochemistry, as is relevant to the C(10)/
C(11) stereogenic centers in the Fisher indoles (cf. 5, Scheme
2) can be addressedshow will the preexisting C(11) stereogenic
center (-OR) influence the stereochemical outcome at C(10)
A series of cyclohexenyl alcohol-based substrates were
prepared in order to test the feasibility of accessing the
tetracyclic skeleton of the Fisher indoles via this methodology
(cf. 5, Scheme 2, Table 3, and eq 1). The majority of these
species were prepared by application of the chemistry developed
(9) Piers, E.; Grierson, J. R.; Lau, C. K.; Nagakura, I. Can. J. Chem. 1982,
60, 210–223.
4960 J. Org. Chem. Vol. 74, No. 14, 2009

Allenyl Azide Cycloaddition Chemistry
TABLE 4. Cyclization Cascade of Substituted Vinyl (o-Azidophenyl)allenes To Furnish Indole and Pyrrole Products
110 °Ca
hνb
hν/CuIc
entry allenyl azide
R
R1
R2
R3 38d (%) 39d (%) 40d (%) 38d (%) 39d (%) 40d (%) 38d (%) 39d (%)
a
b
c
d
e
f
13a
13c
13d
13b
13e
13f
13g
23
CH3
(CH2)2OTBS
t-Bu
CH2OTBS
CH3
CH3
H
H
H
H
Ph
CH3
CH3
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
40
52
57
22
40
36
47
24
56
43
20
29
30
36
39
24
24
53
24
44
38
23
20
42
36
59
55
67
54
69f
16f
38
33
31
3
35
g
h
Ph
CH3
11e, f
CH2OTBS
34
decomp
25
decomp
decomp
decomp
i
j
25
16a
CH3
CH3
H
H
TBS
Ph
H
H
33
7
9
44f
(33)e
35
40
k
l
16c
16d
CH3
CH3
-(CH2)3-
-(CH2)4-
H
H
(36)e
not examined
not examined
36
51
40
32
(20)e, g
51
27
(40)e, g
31
54
42g
53
m
n
24
16b
CH3
CH3
CH3 (CH2)2OTBS
Ph Ph
H
Ph
7
45
(42)e, g
26
21
a 100 mM in toluene. b 254 nm in CH3CN (5 mM). c 254 nm in CH3CN (5 mM), 150 mol % of CuI. d Isolated, chromatographically purified material.
e
Not isolated; estimated by integration of the H NMR spectrum of the crude reaction mixture. f Irradiated at 300 nm. g Irradiated at 350 nm.
1
upon C(3)/C(10) bond formation during electrocyclization of
an indolidene species related to 3a?
present, with a methylene singlet at 4.53 ppm. The structure of
this major isolated isomer tentatively was assigned as the pyrrole
40a after comparison with the spectroscopic data from the
desmethyl analogue of 40a, a compound reported earlier by
Kashulin and Nifant’ev11 and later secured by single-crystal
X-ray analysis (see the Supporting Information for details). It
appears that the indole isomer 39a was formed first as expected,
but it suffered tautomerization to the pyrrole form 40a upon
exposure to SiO2. Thus, this index example offered the hope
that the cyclization cascade might be developed into a high-
yielding route to cyclopentennelated indoles (96% yield overall),
but it also suggested that modifications were necessary in order
to meet the goal of regioselective formation of the C-C-bonded
isomer 38.
One such modification might involve changing the size and/
or electronic character of the substituents R, R1, and R2 in the
hope that steric discrimination in the transition states for bond
formation might translate into regioselectivity. Toward that end,
the size of the allene substituent “R” was increased from methyl
(entry a) to (CH2)2OTBS to t-Bu (entries b and c, respectively).
Similar to the parent 13a, thermolysis of the substrates 13b and
13c furnished products in high chemical yield. The regioselec-
tivity did respond to these steric changes, with the preference
for the C-C bonded isomer 38 increasing as the size of the R
unit increased (1:1.4 for CH3 to 1:1.2 for (CH2)2OTBS to 2.9:1
for t-Bu). The regioselectivity for the related R ) CH2OTBS
case (entry d) was similar to the R ) CH3 case, but these results
may be compromised by the overall lower yield.
The initial attempts at cascade cyclization of the 1-(2-
azidophenyl)-3-alkenylallenes followed the experimental pro-
tocol identified previously for the successful reaction of the
simple saturated-linker 1-azido-3,4,6-hepatrienes.10 Those pro-
cesses furnished cyclopentennelated dihydropyrrole products via
an azatrimethylenemethane intermediate, and at the outset of
this work, there was no reason, a priori, to expect that the
phenyl-linked analogues would proceed through an alternative
pathway. In fact, this argument-by-analogy proved to be
mechanistically misleading (vide infra), although it led to
satisfactory results in a practical sense. Specifically, heating a
100 mM toluene solution of the parent allenyl azide 13a led
efficiently to two new products in a 1.4:1 ratio following SiO2
chromatography, Table 4. Spectroscopic analysis of the minor
component left little doubt that an indole product 38a was
formed, presumably via formation of a C-C bond between C(3)
and C(10) (Fisher indole numbering). However, the structure
of the major product was a bit of a puzzle, as the methylene
unit of this species appeared at a somewhat smaller value in
the 1H NMR spectrum (3.77 ppm) than expected for the
anticipated C-N cyclized product 39a. Adding to the mystery
was the observation that examination of the crude thermosylate
by 1H NMR spectroscopy did not reveal any of the signals for
this compound; rather, a similar but still different species was
Examination of steric effects at the R1 site of 13 extended to
just two compounds, R1 ) CH3 and R1 ) Ph, entries f and e,
(11) Kashulin, I. A.; Nifant’ev, I. E. J. Org. Chem. 2004, 69, 5476–5479.
(12) Cambridge Crystallographic Date Centre deposition numbers: 39n,
649706; 40a, 619622; 42b, 650218, 42d, 652945; 43e, 650217; 54, 651276.
The data can be obtained free of charge from the Cambridge Data Centre via
http://www.ccdc.cam.ac.uk/data_request/cif.
(13) (a) Hirsch, J. A. Top. Stereochem. 1967, 1, 199–222. (b) Eliel, E. L.;
Satici, H. J. Org. Chem. 1994, 59, 688–689. (c) Crich, D.; Jayalath, P.; Hutton,
T. K. J. Org. Chem. 2006, 71, 3064–3070.
(10) Feldman, K. S.; Iyer, M. R. J. Am. Chem. Soc. 2005, 127, 4590–4591.
J. Org. Chem. Vol. 74, No. 14, 2009 4961

Feldman et al.
respectively. Thermolysis of the R1 ) CH3 compound 13f
furnished the two products 38f and 40f in equal proportions
(Table 4, entry f). Enlarging R1 to a phenyl substituent (Table
4, entry e) marginally increased the proportion of the C-C-
bonded product 38e upon thermolysis of 13e. However, in
neither case did the ratio of C-C to C-N bonded products
vary significantly from the index case 13a (R1 ) H). Thus,
introducing steric bulk at the R1 site had little effect. As a
variation on this theme, the R and R1 substituents were switched
in entry g (R ) Ph, R1 ) CH3) in order to see if the same
interaction (CH3//Ph) when approached from the “other” direc-
tion had any different consequence. In fact, the C-C/C-N ratio
(38g/40g ) 1.2) upon thermolysis of 13g fell in line with the
earlier values of entry e (38e/40e ) 1.3). Thus, there is no
evidence to support the notion that a steric interaction between
the R and R1 substituents has any material influence on the
overall regiochemistry of product formation.
disubstituted alkene substrate 24 was subjected to thermolysis,
leading to a preponderance of the sensitive N-cyclized product
40m and only a trace of the C-C-bonded isomer 38m.
A further extension of this alkene substitution theme led to
the trisubstituted alkene analogue 16b. This species was
designed to test the premise that an all-carbon quaternary center
might be accessible through this methodology. Thermally
initiated cascade cyclization of 16b proceeded in good yield to
afford an approximately 2:1 ratio of the C-C/C-N-bonded
products 38n and 39n, respectively. The structural assignment
of the C-N-bonded isomer 39n was based on single-crystal
X-ray analysis (see the Supporting Information). Overall, the
successful thermal cascade cyclization of the disubstituted alkene
substrates 16c and 16d and the trisubstituted higher homologue
16b do extend the scope of the methodology in potentially useful
ways, but they do not lend any insight into the factors that might
influence the key issue of regioselectivity upon terminal bond
formation. In the final analysis, all of the efforts to probe the
relationship between steric bulk and reaction regioselectivity
for the three conceivable substrate variables R, R1, and R2 did
not provide either promising results or even enough encouraging
information to craft a useful model for further developing a
substrate-based strategy to favor the formation of the C-C-
bonded product 38 over the C-N-bonded alternative 39 (or 40).
This discouraging state of affairs improved markedly when
we turned to photochemistry to initiate the cascade process. Even
then, however, success had to await one further chance
discovery. The photochemistry experiments commenced without
a clear notion of how light initiation might favorably influence
the regiochemistry of the transformation, since a clear grasp of
the mechanistic intricacies of the process still eluded us.
Nevertheless, since ignorance should not tarnish hope, we
proceeded to examine a few of the allenyl azide substrates under
irradiation and found that, indeed, light at 254 nm did promote
the cyclization cascade and furnish the expected suite of
cyclopentennelated indole products (Table 4, entries a-g).
However, no material improvement in the reaction regiochem-
istry followed from these examplessthe usual unremarkable
ratio of C-C 38 to C-N 39/40 bonded products were detected
in yields comparable to the thermal cases. Thus, we learned
that the sequence could be initiated with photons of appropriate
energy, but it seemed likely that the same intermediate(s) were
being accessed under either thermolysis or photolysis, led to
the same nearly unbiased mixture of regioisomeric products.
Variation in the steric bulk at the R2 position was expected
to have the most profound effect on product regioselectivity
since R2 was right at the bond-forming site. This expectation
was partially supported experimentally, as thermolysis of allenyl
azide substrates with R2 ) CH2OTBS and R2 ) Ph led to
product regioisomer ratios that scarcely varied from 1:1, but
when R2 ) the larger TBS, the C-C bonded product was
strongly favored (∼ 5:1) over the C-N alternative. Note that
the C-C-cyclized species 38j did not survive SiO2 chromatog-
raphy, and so the reported yield derives from integration of
1
characteristic signals in the H NMR spectrum of the crude
thermolysate. The N-cyclized species featured an indole unit
39 and not the previously observed pyrrole moiety of isomer
40. Speculation about the absence of isomerization in this series
might include the argument that planarization of the R2-bearing
carbon upon pyrrole-forming isomerization would bring R2 and
the peri-positioned aryl hydrogen in conflict, a steric clash that
is absent in the R2 ) H substrates. The structural assignment
of 39j was confirmed by single crystal X-ray analysis of the
derived hydrogenation product.1a The lack of correlation
between steric bulk at R2 and regioselectivity of terminal bond
formation was confounding at first, but a more thorough
mechanistic analysis, guided by density functional calculations,
provided a consistent rationale by placing the regioselectivity-
determining event earlier on the reaction coordinate than the
final C-C (f 38) vs C-N (f39/ 40) bond closures (vide infra).
Three disubstituted alkene substrates, 16c, 16d, and 24, were
examined in the thermolysis reaction in order to extend the scope
of the transformation. The cyclic substrates 16c (cyclopentenyl)
and 16d (cyclohexenyl) both participated in the cyclization
cascade, but provided divergent types of products. The cyclo-
hexenyl case (Table 4, entry l) delivered the expected C-C
cyclized product 38l and the isomerized C-N-cyclized regio-
isomer 40l in a ratio slightly favoring the latter species. The
cyclopentenyl substrate 16c, on the other hand (Table 4, entry
k), provided only the unisomerized C-N bonded indole product
39k following SiO2 chromatography. Signals characteristic of
the tentatively assigned C-C bonded product 38k could be
detected in the 1H NMR spectrum of the crude thermolysate in
nearly equal amounts with 39k, but this species did not survive
the purification procedure. The differing stabilities of the similar
tetracycles 38k and 38l to chromatography are surprising.
Identifiable decomposition products from 38k (or 38j, another
C-cyclized species intolerant to SiO2) were not isolated, and so
the basis for this sensitivity remains unknown. The final
During the course of these photolysis experiments, however,
an unexpected observation was made. Every once in a while, a
batch of allenyl azide substrate proceeded to indole product with
remarkably high regioselectivity (>10:1) favoring the C-C-
cyclized product 38. What distinguished these “exceptional”
batches of substrate from the normal, disappointing ones? The
key parameter was purity; the early nonregioselective runs
utilized scrupulously purified and colorless allenyl azides that
were obtained through careful SiO2 chromatography even at the
inevitable expense of material loss due to the sensitivity of
these species. In some cases, however, batches of allenyl azide
were obtained that had a light pink color despite appearing to
be impurity-free by 1H and 13C NMR spectroscopy. It was these
colored batches that delivered indole product with high regi-
oselectivity for the C-C-bonded isomer 38. Once this connec-
tion was made, the search for the apparently critical “impurity”
in the pink samples led to examination of test case photolysis
reactions with substrate 28a in the presence of various ingre-
4962 J. Org. Chem. Vol. 74, No. 14, 2009

Allenyl Azide Cycloaddition Chemistry
TABLE 5. Discovery and Optimization of the Cui-Mediated Photochemical Cascade Cyclization of 29a into 42b
entry
irrada (nm)
heatb (°C)
conditions
no additives
no additives
excess CuI, LiBr, Mg powder
LiBr (1 equiv)
yieldc (%)
42b/43b ratiod
a
b
b
d
e
f
g
h
i
j
k
l
m
n
o
p
q
r
110
80
67
33
68
(33)e
f
1:1.6
1:1
254
254
254
254
254
254
254
254
254
254
254
254
254
>10:1
1:1.1
1:1.2
1:1.4
5.4:1
2.8:1
1.6:1
5:1
4:1
6: 1
9:1
10:1
1:1.6
3.5:1
5:1
Mg powder
LiBr (1 equiv), CuI (0.05 equiv)
LiBr (1 equiv), CuI (1 equiv)
CuBr ·SMe2 (1 equiv)
CuBr ·SMe2 (0.25 equiv)
Cu(OAc)2 (1 equiv)
CuI (0.5 equiv)
CuI (0.8 equiv)
CuI (1.2 equiv)
CuI (1.5 equiv)
CuI (1.5 equiv), 5 mM
CuI (1.5 equiv), 35 mM
CuI (1.5 equiv), 51 mM
[Rh(OAc)2]2 (1 equiv)
21
f
(17)e
f
67
72
67
70
75
66
65
(48)e
110
110
110
254
1:1
a Irradiated through a quartz vessel as a 5 mM CH3CN solution of 28a. b 100 mM solution in toluene, unless otherwise noted. c Isolated,
chromatographically purified material. The N-cyclized material had undergone elimination to furnish the diene 44b (see Table 6). d Ratio by integration
of the 1H NMR spectrum of the crude reaction mixture prior to chromatography. e Only the N-cyclized material was isolated; 42b decomposed upon
attempted chromatography. f Products were not isolated.
dients of the allenyl azide-forming reaction, which included CuI,
LiBr, and Mg (see Table 5). Eventually, the key role of Cu(I)
became apparent, and the salt CuI was chosen for its ready
availability and solubility in the reaction solvent CH3CN.
Furthermore, it is mechanistically noteworthy that the improve-
ment in regioselectivity scaled with the amount of CuI present.
Reproducibly high regioselectivity favoring the C-C-bonded
product 42b was obtained when 150 mol % of CuI was included
in the photolysis reaction mixture. None of the other additives
tested, whether present in the allenyl azide synthesis or not,
had any significant effect on the reaction regiochemistry. The
role that copper plays in this complex process was still to be
elucidated (vide infra), but this empirically derived solution to
the regiochemical issue begged further study. In this light,
reexamination of several of the allenyl azide substrates (Table
4) under the hν/CuI conditions led to favorable results overall,
with a few notable exceptions. Simple unsubstituted-alkene
substrates (Table 4, entries a-d) responded well to the new
protocol, and the C-C-bonded regioisomer 38 was formed in
good yield in ratios of at least 18:1 compared to the C-N-
bonded isomer 40. The monosubstituted-alkene substrates tested,
13e-g, provided tricyclic indole product that appeared to consist
exclusively of the C-C-bonded isomers 38e-g, respectively.
The C-N bonded products 40e-g could not be detected in the
1H NMR spectrum of the crude photolysate. The cyclohexenyl-
substituted allenyl azide 16d also provided only the C-C-
bonded tetracycle 38l. Two exceptions to this dramatic im-
provement in reaction regioselectivity can be seen in Table 4,
entries m and n. Both the disubstituted alkene substrate 24 and
the trisubstituted analogue 16b did furnish good yields of
tricyclic indole-containing products, but there was no apparent
improvement in regioselectivity for 16b over the thermally
initiated case. Disubstituted alkene substrate 24 did proceed to
indole product with an enhanced selectivity for the C-C-bonded
isomer 38m (C-C/C-N ) 1:1) compared to the thermal case
(C-C/C-N ) 1:5), but the result was far short of the goal of
strongly favoring the C-C-bonded isomer. Thus, it appears that
these particular di- and trisubstituted alkene substrates can be
induced to form indole products under the CuI/hν reaction
conditions, but the high levels of regioselectivity for the C-C-
bonded product experienced by simpler species do not translate
for these entries.
The positive effect of copper inclusion on reaction regiose-
lectivity in the photochemical series could be extended to the
thermal chemistry as well, Table 5, entries o-q. Substrate
concentration appeared to be a key variable, as a thermolysis
experiment run at the photochemical experiment’s concentration
of allenyl azide did not show any influence of the copper
additive. Upon increasing the substrate concentration 10-fold,
however, the copper effect became manifest; the C-C-bonded
product 42b was formed in a 5:1 ratio compared to the C-N-
bonded isomer 43b (Table 5, entry q). These concentration
effects, and our suspicion that both photochemistry and thermal
chemistry proceeded through common intermediates (vide
supra), formed the basis of speculation on the role of copper in
this transformation. This argument will be elaborated with the
aid of density function calculations as discussed with Scheme
9 below.
The final series of allenyl azide substrates examined in this
study featured functionalized cyclohexenyl rings as model
systems for the Fisher indoles, Table 6. The structural variables
in this series included the allylic alcohol/ether substituent R,
the carbinol group R1, and the alkyl residue appended to C(3)
of the allene, R2. The reaction features that were tested by these
structural probes comprised the overriding regioselectivity issue
(C-C vs C-N terminal bond formation) and the relative
stereochemistry emerging at C(10) under the influence of the
adjacent C(11) stereogenic center. In addition, the sensitivity
J. Org. Chem. Vol. 74, No. 14, 2009 4963

Feldman et al.
TABLE 6. Cyclization Cascade of Substituted 1-(2-Azidophenyl)-3-cyclohexenylallenes To Furnish Indole Products
110 °Ca
43d (%)
hνb
43d (%)
hν/CuIc
43d (%)
entry
allenyl azide
R
R1
R2
42d (%)
44d (%)
42d (%)
44d (%)
42d (%)
44d (%)
a
b
c
d
e
29
28a
28b
28d
28e
H
H
H
H
CN
H
H
H
H
H
34
34
25
25
51
21
33
31
24
34
45
60
62
64
62
TBS
TIPS
TBS
TBS
8
5
27
35
48
35
27
43
50
7
f
TMS
41 (45)
(25)e
11 (45)
f
g
28c
37
SEM
TBS
H
H
H
OTBS
51
38
51
60
59
not examined
a 100 mM in toluene. b 254 nm in CH3CN (5 mM). c 254 nm in CH3CN (5 mM), 150 mol % CuI. d Isolated, chromatographically purified material.
e
Not isolated (unstable); estimated by integration of the H NMR spectrum of the crude reaction mixture. f A 1.6:1 ratio of 42e/45 was observed in the
1
1H NMR spectrum of the crude reaction mixture; chromatographic purification of this CuI-mediated reaction gave irreproducible isolation results.
TABLE 7. Cyclization Cascade with the H-Substituted- and TMS-Substituted 1-(2-Azidophenyl)-3-(2-propenyl)allene Substrates 13h and 13i,
Respectively
entry
allenyl azide
conditions (CH3CN)
51a/ba(%)
52a/ba(%)
53a(%)
54a(%)
a
b
c
d
e
f
13h
13h
13h
13h
13i
13i
13i
110 °C
53
40
6
8
9
23
6
110 °C/150 mol % CuI
3
28
254 nm
35
254 nm/150 mol % CuI
110 °C
254 nm
48
12b
decomp
decomp
g
254 nm/150 mol % CuI
b
a Isolated, chromatographically purified material. Observed in the H NMR spectrum of the crude reaction mixture, but decomposed upon attempted
1
chromatographic isolation.
of the C(11) oxygen function to fulvene-forming elimination
promoted by the potentially acidic C(10) cyclopentadiene proton
was a concern. The preliminary CuI optimization studies with
the OTBS variant 28a detailed in Table 5 suggested that this
latter issue might not arise as a general problem, as 42b could
be isolated without event following SiO2 chromatography. The
data displayed in Table 6 indicate that all of the cyclohexenyl
alcohol derivatives 28a-e, 29, and 37 performed satisfactorily
under the new hν/CuI conditions to deliver good yields of the
C-C-cyclized products 42a-g, respectively, accompanied by
only trace amounts of the C-N-bonded regioisomer. The
mildness of this protocol was evidenced by the functional group
tolerance, which included a free alcohol, silyl ethers, and acid-
sensitive alkyl ethers. The stereochemical outcome of all of the
cyclizations examined was completely biased toward the C(10)/
C(11) syn isomer shown as 42; no evidence for minor species
1
bearing an alternative stereochemistry was detected in the H
NMR spectrum of the crude photolysates. Surprisingly, even
4964 J. Org. Chem. Vol. 74, No. 14, 2009

Allenyl Azide Cycloaddition Chemistry
SCHEME 5. Cascade Cyclization Attempts with Aryl-Substituted 1-(2-Azidophenyl)allene Substrates 14j-m
the cyanohydrin 42d, whose structure was determined by single-
crystal X-ray analysis (see the Supporting Information), did not
deviate from this trend, perhaps because of the steric distinction
between the OTBS and CN groups (A values: CN 0.15 kcal/
mol (298 K, t-BuOH);13a OSi(CH3)2-t-Bu 1.06 kcal/mol (179-188
K, CD2Cl2))13b-1.5 kcal/mol (193 K, CS2 solvent).13c
hydride as a nucleophile leaves C(11) in 47 devoid of
functionality. Application of this methylation procedure in Fisher
indole synthesis therefore will have to employ a nucleophilic
species with fulvene 46 that preserves the opportunity to further
functionalize C(11) as is required for 5.
In counterpoint, cyclohexenol-based allenyl azide cascade
cyclizations performed under strictly thermal conditions afforded
the usual mixture of the C-C- and C-N-cyclized regioisomers
42 and 43/44, respectively, with little preference for one mode
of cyclization over the other. Similarly, photochemistry without
copper furnished indole products with nearly identical regioi-
somer ratios as in the thermal cases. These contrasting results
emphasize the importance of copper(I) inclusion in the pho-
tolysis solution and are consistent with our preliminary inter-
pretation that this metal salt plays a critical role in steering the
partitioning of a reactive intermediate(s) between the two
regioisomeric products.
The TMS-substituted simple allenyl azide 13h was examined
with the expectation that an uneventful reaction would deliver
products of the type 51a and 52a, Table 7. Less certain was
the fate of the single allenyl-H-substituted system 13i, given
the propensity for H-substituted triazolines (cf. 2, R ) H) to
tautomerize into stable aromatic triazoles.14 Whereas the
standard indole products 51a/b and 52a/b were produced under
certain circumstances, the big surprise with 13h involved
formation of the mechanistically informative triazoles 53 and
54. The triazole 53 also was formed as a minor component from
13i. The structural assignments of 51a/b, 52a/b, and 53 followed
The two substrates that included functionality on the allenyl
methyl unit proceeded to indole products as expected, but not
to the same type of indole products, which was unexpected
(Table 6). The R2 ) TMS substrate 28e was the outlier within
this series. Reaction under thermal conditions did not afford
any of the C-C-bonded indole product 42e. Rather, the C-N-
bonded regioisomer 43e (X-ray analysis; see the Supporting
Information) and a new type of product, the alkene 45,
1
from their characteristic H NMR data, but the structure of 54
remained a puzzle until single-crystal X-ray analysis revealed
it to be the formal acetonitrile adduct shown (see the Supporting
Information for details). Under thermolytic conditions with or
without copper, the triazoles were the major species formed from
13h. The formation of these compounds irrefutably implicates
the intermediacy of a triazoline species 48a/b, previously posited
but otherwise unconfirmed. Apparently, the elevated temperature
of these reaction conditions with 13h provided enough energy
for the nucleophilic capability of the allyl silane moiety within
48a to be expressed; trapping by either adventitious protons
or, remarkably, by acetonitrile, led to 53 and 54, respectively.
With 13i, simple tautomerization within 48b, presumably
catalyzed by the same adventitious acid, delivers 53. Irradiation
of 13h without copper(I) generates a reaction solution temper-
ature of approximately 35 °C, and under this temperature regime
the allyl silane moiety of 48a participates only slightly in triazole
formation. In this instance, the major products formed are the
C-C- and C-N-bonded indoles 51a and 52a, respectively, in
a ratio similar to that seen earlier with related substrates (Table
4). Copper-mediated photolysis of 13h again produces a minor
amount of the triazole 53, but the major product is the C-C-
bonded regioisomer 51a, formed to the exclusion of the C-N-
bonded alternative 52a. The H-substituted substrate 13i decom-
posed under irradiation and no further information was
forthcoming from this species. The ( CuI photochemistry of
1
predominated. H NMR analysis of the crude thermolysate
suggested that the first-formed alkene product actually had a
TMS moiety on the indole nitrogen but that labile group suffers
protodesilylation upon SiO2 chromatography. The formation of
this alkene can be rationalized by passage through an indolidene
3a (cf. Scheme 1), which engages in silyl transfer in preference
to C-C bond-forming electrocyclization. Irradiation of 28e sans
copper did furnish evidence for the C-C-bonded indole product
42e in the 1H NMR spectrum of the crude product mixture, but
the C-N-bonded regioisomer 43e predominated. In the presence
of copper(I), however, the photochemically derived product
suite is entirely different, with the C-C-bonded regioisomeric
indole 42e largely favored and the C-N regioisomer 43e absent.
These observations provide further clues about the role of Cu(I)
in influencing the regiochemical outcome of the cascade
cyclization, a topic taken up in more detail below.
Using the Fisher indoles as a context for these studies
introduces the need to convert sp2-hybridized C(16) of 42a into
a quaternary carbon bearing two methyl groups, eq 2. Toward
this end, we discovered that hydride addition to an intermediate
fulvene 46, a species inadvertently derived from the alcohol
42a upon N-BOC protection conditions, provided a cyclopen-
tadienylide anion intermediate that methylates exclusively at
C(16). This intermediate anion could, in principle, methylate
at C(10) as well, but an incipient 1,3 diaxial interaction between
the incoming methyl electrophile and the axial CH3 of C(12)
apparently thwarts this reaction trajectory. Of course, use of
(14) Mukai, C.; Kobayashi, M.; Kubota, S.; Takahashi, Y.; Kitagaki, S. J.
Org. Chem. 2004, 69, 2128–2136.
J. Org. Chem. Vol. 74, No. 14, 2009 4965

Feldman et al.
SCHEME 6. Attempt To Synthesize a
(Z)-Phenyl-Substituted 1-(2-Azidophenyl)-3-alkenylallene
Substrate 63
tion. Thus, the conformational consequences of the R//R1 steric
interaction () CH3//Ph in 16b, but only ) H//CH3 in 62) may
be the decisive factor in both the stability and the successful
reaction of 16b and the failure of same with 62. A second
irretrievable failure attended the attempted cascade cyclization
of the enone 30 (cf. Table 3). Under any conditions examined
(thermal, photochemistry, photochemistry with CuI), this sub-
strate was consumed quickly without formation of any charac-
terizable products. Presumably, the electronic activation con-
ferred by the carbonyl functionality introduced lower energy
competitive pathways that drained away allenyl azide faster than
the desired [3 + 2] cycloaddition process.
Mechanistic Analysis. The accumulated experimental evi-
dence to date can be framed within a unified mechanistic
scenario featuring a sequence of discrete triazoline and then
indolidene intermediates as hypothesized in Scheme 1. However,
many subtleties remain unresolved in this crude picture; for
example, (a) which step is the slow step? (b) is loss of N2 from
the triazoline 2 concerted or stepwise? (c) what is the relation-
ship between substituent R in 2 and the alkene geometry in
3a/3b?, (d) what is the relationship between the alkene geometry
of 3a/3b and the regiochemistry of terminal bond formation,
C-C (f4a) or C-N (f4b)?, and finally, (e) what is the role
of Cu(I) in redirecting the regiochemistry of bond formation to
strongly favor 4a? In order to gain more insight into the details
of this complex process and address these questions, we turned
to density functional theory in its Kohn-Sham formulation,
using the B3LYP/6-311+G(d,p)//B3LYP/6-31G(d,p) computa-
tional approach.15 Second derivatives of the energy with respect
to the Cartesian nuclear coordinates were computed for all
stationary points and subsequent harmonic analysis confirmed
the nature (minimum or transition state) of each structure.
Unscaled frequency values obtained from the second derivatives
were employed for thermochemical analysis. The calculated
reaction profile for the noncopper mediated thermal conversion
of allenyl azide 13a into the precursor 66 of the C-C bonded
indole product 38a and the precursor 39a of the observed C-N
bonded regioisomer 40a is detailed in Scheme 7.
Interpretation of the energy values displayed on this reaction
coordinate diagram provides an entry point for answering the
mechanistic questions listed above. In order: (a) The rate-
determining step appears to be the initial [3 + 2] cycloaddition
to form triazoline 64. (b) The conversion of triazoline 64 into
the indolidenes 65a and 65b can, in principle, occur through
either concerted or stepwise loss of N2. The calculated barriers
for concerted N2 loss, 16.5 kcal/mol for 65a and 16.7 kcal/mol
for 65b, are likely to be smaller than the barrier for stepwise
N2 loss, although a transition state for the conversion of
triazoline 64 into the singlet diyl 67 could not be located by
this computational approach (see the Supporting Information
for details). This claim may seem surprising at first glance, since
the concerted loss of N2 from 64 is equivalent to a formal
suprafacial 12π-electron cycloreversion, a process that falls into
the “forbidden” regime of the Woodward-Hoffmann (W-H)
rules.16 How can these calculated results be reconciled with the
otherwise infallibility of the W-H dogma? The key to
understanding this apparent exception is appreciating the
13h proceeded in a manner completely analogous to the other
simple alkene-containing substrates. The anomalous behavior
of the silyl-substituted substrate 13h occurs at higher temper-
atures, where an alternative pathway involving electrophilic
desilylation intervenes, and in so doing, preserves evidence for
the intermediacy of the triazoline 48a. Likewise, the thermal
chemistry of 13i argues for reaction through triazoline 48b, but
curiously, N2 loss is competitive with tautomerization for this
intermediate.
The final group of substrates examined feature an aryl
substituent on the allenyl azide framework, 14j - 14m, Scheme
5. The challenge with this type of substrate lies in the
competition within the indolidene intermediates 55a/55b be-
1,7
tween
H-shift and electrocyclization through the aryl ring.
Whereas the H-shift has always been a potentially problematic
side-reaction, the electrocyclization through an alkene partner
apparently won out energetically in all of the previous cases.
With 14j - 14m, however, breeching the aromaticity of the
aryl group evidently is not competitive, and only the [1,7]
H-shift products 57j - 57m are formed in modest yields.
Photochemistry with and without CuI present was explored with
14j, although no evidence for redirection of the reaction course
was detected. Thus, irradiation (254 nm) of 14j in CH3CN
afforded only 57j in 56% yield, whereas similar reaction in the
presence of 1.5 equivalents of CuI provided this same product
in 41% yield.
Finally, a discussion of the scope of the cascade cyclization
would not be complete without acknowledging the 1-(2-
azidophenyl)-3-alkenyl allene substrates that, for various reasons,
did not participate productively in the transformation. For
example, attempts to prepare a terminally disubstituted alkene
substrate 62 were derailed by the facility of a 6π-electron
electrocyclization that affords the naphthalene product 61
following tautomerization within 63 (Scheme 6). This example
can be contrasted with the successful formation and unexcep-
tional cascade cyclization of the triphenylalkenyl allenyl azide
substrate 16b. A rationale for the lack of similar electrocycliza-
tion with this (Z)-phenyl-containing alkene might be traced to
the additional phenyl group at position R1 (cf. 16b, Table 4),
which might prevent access to the approximately planar
conformation of the triene that should facilitate electrocycliza-
(15) (a) Hohenberg, P.; Kohn, W. Phys. ReV. 1964, 136, B864-B871. (b)
Kohn, W.; Sham, L. Phys. ReV. A 1965, 140, A1133-A1138. (c) Stephens, P. J.;
Devlin, F. J.; Chabalowski, C. F.; Frisch, M. J. J. Phys. Chem. 1994, 98, 11623–
11637. (d) Becke, A. D. J. Chem. Phys. 1993, 98, 5648–5652. (e) Lee, C.; Yang,
W.; Parr, R. G. Phys. ReV. B 1988, 37, 785–789.
(16) Woodward, R. B.; Hoffmann, R. The ConserVation of Orbital Symmetry;
VCH: Weinheim, 1970.
4966 J. Org. Chem. Vol. 74, No. 14, 2009

Allenyl Azide Cycloaddition Chemistry
SCHEME 7. Qualitative Representation of the Free Energy Profile for the Conversion of Azidoallene 13a into the C-C and
C-N Cyclized Products 66 and 39a, Respectivelya
a Energies of ground-state structures and transition states were calculated using the B3LYP/6-311+G(d,p)//B3LYP/6-31G(d,p) functional. All energy
values are in kcal/mol.
SCHEME 8. Steric Interactions and Rotational Direction
during Indolidene Formation
orthogonal relationship between the two π-systems (the indole
and the N2). From this perspective, as the C-N and N-N bonds
stretch during the approach to the transition state, the two π
systems scarcely “communicate” and there is no appreciable
passage of electron density between them. This supposition is
supported by the results of ancillary ACID17 and NICS18
calculations of N2 loss from the desmethyl version of 64.2 These
computational techniques provide “measurements” of electron
density in off-atom regions of space, such as the region between
the cleaving atoms in the transition state for N2 extrusion. These
calculations clearly indicate that almost no electron density
occupies this key locale in the transition state for N2 extrusion.2
It appears that as the C-N bond stretches in 64, the remaining
N-N bond begins to stretch before the transition state to 67 is
reached, and, in what might be characterized as a non-least-
motion process, C(16) (cf. 64) begins to rotate until the planarity
of 65a/65b is achieved. Overall, this N2 loss process appears to
be accidentally concerted, and the lack of electronic com-
munication between the two components removes it from the
W-H regime of cycloadditions.
The relationship between indolidene alkene geometry and
substituent R (question (c)) is illustrated in Scheme 8. The
direction of rotation about the indicated bond in 69 determines
the eventual alkene geometry; rotation in direction A as shown
leads to the (E)-isomer 70a, whereas the alternative, rotation in
direction B, delivers the (Z)-isomer 70b. These rotational
directions engender different steric interactions between the
proximate substituents, and it is reasonable to suppose that the
energetic trade-offs between these interactions ultimately de-
termine the alkene geometry ratio. Rotational direction A
increases the vinyl//H clash but relieves the R//H interaction,
whereas rotational direction B has the opposite effect. To the
extent that the substituent R is larger than a vinyl group, rotation
away from the R//H interaction as in 69-(rotation A) would be
expected to be favored, leading to preferred formation of 70a
over 70b. The limited series of substrates 13a-c (R ) CH3,
(CH2)2OTBS, t-Bu, respectively) provide results consistent with
this analysis. As the R substituent increases in size, the C-C
bonded product emerging from 70a is increasingly preferred;
R ) CH3 1:1.4 38a/40a, R ) (CH2)2OTBS 1:1.2 38b/40b, and
R ) t-Bu 2.9:1 38c/40c (A values: CH3 ) 1.7 kcal/mol, ethyl
) 1.75 kcal/mol, vinyl ) 1.35 kcal/mol, t-Bu ≈ 4 kcal/mol).13a
DFT calculations largely support these conclusions. Thus, the
calculated ratio of 65a/65b derived from 13a (R ) CH3, R1 )
H) is 55:45, a value that is not too dissimilar to the experimental
ratio (42:58) of eventual isolable products 38a:40a (Table 4,
entry a). The calculated/experimental values for the analogous
tert-butyl case 13c (Table 4, entry c), remarkably, are completely
coincident; 74:26 for 38c/40c. Finally, calculations of the
barriers for 65a/65b formation from 13g (R ) Ph, R1 ) CH3)
predict a 90:10 ratio of C- to N-cyclized material; the experi-
mental ratio of 38g/40g is 55:45. The basis for this discrepancy
has not been identified.
(17) (a) Herges, R.; Geuenich, D. J. Phys. Chem. A 2001, 105, 3214–3220.
(b) Geuenich, D.; Hess, K.; Koehler, F.; Herges, R. Chem. ReV. 2005, 105, 3758–
3772.
(18) Schleyer, P. v. R.; Maerker, C.; Dransfeld, A.; Jiao, H.; Hommes, N. J. R.
v. E. J. Am. Chem. Soc. 1996, 118, 6317–6318.
The importance of the alkene geometry of 65a and 65b on
eventual product regiochemistry (question (d)) is highlighted
as a result of these calculations. Each alkene isomer can proceed
J. Org. Chem. Vol. 74, No. 14, 2009 4967

Feldman et al.
SCHEME 9. Qualitative Representation of the Free Energy Profile for the Conversion of Allenyl Azide 13a into the C-C and
C-N Cyclized Products 72 and 73, Respectively, in the Presence of Cu(I) ·HCNa
a Energies of ground-state structures and transition states were calculated using the B3LYP/6-311+G(d,p)//B3LYP/6-31G(d,p) methodology.
via an electrocyclization to give indole products, 65a f 66 and
65b f 39a, though concerted processes that require surmounting
activation barriers of only 14.5 kcal/mol for each transformation.
In contrast, the barrier for interconversion of these alkene
isomers is much higher; 33.2 kcal/mol, and so a non-
Curtin-Hammet situation arises in which 65a and 65b lead
directly to product without equilibration. Thus, these calculations
provide numerical support for a reactivity model in which the
ultimate product regioisomer ratio (4a/4b, Scheme 1) replicates
the indolidene alkene isomer population, which in turn reflects
the differential steric interactions illustrated in Scheme 8.
The introduction of Cu(I) into this complex reaction sequence
(question (e)) could, in principle, impact the chemistry at several
different stages, Scheme 9. For example, the copper might play
a defining role in the initial azide chemistry by forming a copper
nitrene-type intermediate 74 by analogy with much prior copper-
mediated amination chemistry.19 Alternatively, the copper might
become involved in a much more passive capacity by providing
a Lewis acidic binding site for various nitrogen donor ligands,
as suggested by 71a/b. Initially, the nature of the Cu(I) ligation
was explored through calculation. An assumption underlying
all of the Cu-ligand models examined for Cu solvation was
made; the I- was dissociated in the polar CH3CN solvent, and
so the active species was a version of Cu(1+)(CH3CN)n. From
n ) 0 up to three coordinated HCN molecules (as a model for
the CH3CN solvent used in the reactions) were explored
computationally with several of the putative intermediates
illustrated in Scheme 9. The structures and energies of species
with no coordinating ligands (“naked copper(I)”) reflected the
extreme electrophilicity of the metal, which was an unrealistic
attribute for catalytic activity. Tricoordinated Cu(I) appeared
to be too sterically hindered to participate in any productive
interactions with the organic compounds present. Little energy
difference was found between di- and monocoordinated copper
with select intermediates of Scheme 9, but the singly coordinated
Cu(I)·HCN complex consistently provided species with slightly
lower energies, and so that motif was adopted for the calcula-
tional approach as a whole.
The initial question of importance focuses on the competition
between copper-mediated nitrene 74 formation and orthodox
[3 + 2] cycloaddition (f 64) within 13a to start the cascade
cyclization off. The calculated activation barriers are similar,
with a slight edge going to the nitrene mechanism. The
subsequent electrocyclization of 74 and reductive elimination
within the Cu(III) species 75 are essentially barrierless processes
by these calculations. Since the numerical values for the two
distinct pathways from 13a are within computational error, there
is no basis to exclude either from further consideration. In fact,
both starting points lead to the same intermediates, the Cu(I)-
bound indolidenes 71a and 71b. The nitrene-based mechanism
delivers these key species directly, whereas the [3 + 2]
cycloaddition route requires a bimolecular step, Cu(I) ligation
(not explicitly shown in Scheme 9; barrierless by these
calculations), after the indolidenes 65a/65b are formed by the
previously discussed N2 extrusion reaction. In neither chemistry
is there yet any reason to link copper participation to a change
in the regiochemistry of indole formation, the pivotal issue of
this discussion.
The link between copper mediation and reaction regiochem-
istry comes at the next stage of the sequence, during which the
indolidenes 71a and 71b electrocyclize to product tricycles 72
and 73, respectively. The inclusion of Cu(I) in this part of the
transformation perturbs the system in two profound ways: (1)
The barrier to electrocyclization to form the C-N bonded
species (71b f 73) is now significantly raised compared to the
noncopper case of Scheme 7 (24.9 kcal/mol for 71b vs 14.5
(19) (a) Li, Z.; Quan, R. W.; Jacobsen, E. N. J. Am. Chem. Soc. 1995, 117,
5889–5890. (b) Brandt, P.; So¨dergren, M. J.; Andersson, P. G.; Norrby, P.-O.
J. Am. Chem. Soc. 2000, 122, 8013–8020. (c) Bertz, S. H.; Cope, S.; Murphy,
M.; Ogle, C. A.; Taylor, B. J. J. Am. Chem. Soc. 2007, 129, 7208–7209.
4968 J. Org. Chem. Vol. 74, No. 14, 2009

Allenyl Azide Cycloaddition Chemistry
Gaussian0321 in conjunction with the 6-31G(d,p) basis set. To
further improve the electronic energy values, an energy refinement
was computed with the larger triple-ꢀ quality 6-311+G(d,p) basis
set. This scheme usually is noted as B3LYP/6-311+G(d,p)//B3LYP/
6-32G(d,p).
Due to the potential diradical character of some of the structures
considered in this work, the internal and external stability of the
wave functions was computed via the Hermitian stability matrices
A and B in all cases.22 For all of the structures exhibiting unstable
restricted wave functions, the spin symmetry constraint of the wave
function was released (i.e., expanding the SCF calculation to an
unrestricted space, UB3LYP), leading to stable unrestricted wave
functions.
kcal/mol for 65b). (2) The barrier to interconversion between
the two indolidenes is now lower than the same barrier in the
copper-free case (23.5 kcal/mol for 71a f 71b, compared to
33.4 kcal/mol for 65a f 65b). The barrier to C-C bond forming
electrocyclization (71a f 72) remains largely unchanged by
copper ligation (12.5 kcal/mol vs 14.5 kcal/mol for 65a f 66).
The confluence of these energy changes has significant conse-
quences for the regioselectivity of the reaction. With Cu(I)
present, the fate of the pre-C-N forming indolidene 71b is now
different than its fate with Cu absent. In the former instance,
isomerization of 71b to the pre-C-C bond forming indolidene
71a is calculated to be faster than electrocyclization to afford
73. Since the barrier to electrocyclization of 71a is largely
unperturbed by copper, its closure is now the fastest option
available to this (Z)-indolidene. Thus, in contradistinction to
the noncopper-mediated chemistry of Scheme 7, a Curtin-
Hammett situation appears to be at hand here; 71b interconverts
to its geometrical isomer 71a faster than it proceeds to indole
product 73, while 71a forms tricyclic product 72 faster than it
returns to 71b. The sum of these factors then favors formation
of the C-C bonded product 72. Thus, these calculations suggest
that copper(I) inclusion in the cascade cyclization reaction may
have two effects: (1) redirecting indolidene formation through
a copper-nitrene intermediate and (2) ligating to, and retarding
the rate of, electrocyclization of the pre-C-N bond forming
indolidene. It is only this latter effect that influences reaction
regioselectivity to favor the C-C bonded product 4a. The two
substrates 16b and 24, whose regiochemistry upon photocy-
clization did not benefit from CuI incorporation, may have
suffered from unfavorable steric interactions between the alkene
substituents and Cu(L)n that mitigated against formation of the
pivotal complex 71a.
The cascade cyclizations of 1-(2-azidophenyl)-3-alkenylal-
lenes provide 2,3-annelated indole-containing products under a
range of experimental conditions. These transformations likely
commence through initial [3 + 2] cycloaddition of the azide to
the allene and then proceed through loss of nitrogen gas and
electrocyclization of the derived indolidene. Computational
studies of the energy surface for this complex transformation
suggest that the central issue of C-C vs C-N regiochemistry
of ultimate bond formation appears to be related to the
differences in the activation energies for indolidene alkene
isomer interconversion compared with the barriers for final C-C
or C-N bond-forming electrocyclization. Incorporation of a
nitrogen binding metal (CuI) can influence these energetic trade-
offs and in favorable cases steer the reaction toward high levels
of C-C-bonded cyclopentennelated indole product. One family
of substrates bearing functionalized cyclohexene rings as the
“alkene” portion were particularly suitable for C-C bond
forming cyclization, and the derived tetracyclic structures may
find use in the synthesis of cognate natural products such as
the Fisher indoles, inter alia.
This methodology has been tested extensively and compared
against very robust and specialized methodology (CASSCF,
CASPR2) for the study of oxa and aza derivatives of the
trimethylenemethane motif.23
General Procedure 1. Allenyl Azide Formation. To a solution
of ZnCl2 (2.5 equiv) in THF (0.3 M solution) was added the
appropriate Grignard reagent (2.5 equiv), and the mixture was stirred
at room temperature for 1 h. Pd(PPh3)4 (5 mol %) and the
propargylic acetate (1 equiv) in THF (0.1 M solution) were added
sequentially. The reaction mixture was allowed to stir at room
temperature for 20 min (monitored by TLC for starting material
consumption). After addition of an equal volume of a saturated
NH4Cl solution, the organic layer was extracted into Et2O and
washed with water and brine, dried over Na2SO4, and evaporated
under reduced pressure maintaining a water bath temperature below
40 °C. The resulting oil was purified by flash chromatography
(hexanes f 2% Et2O/hexanes).
General Procedure 2. Sonagashira Coupling. To a 0.5 M
solution of alkenyl iodide in THF were added PdCl2(PPh3)2 (10
mol %), 1-(2-azidophenyl)prop-2-ynyl acetate (12i) (1.6 equiv),
Et3N (7 equiv), and CuI (3 mol %). The mixture was stirred at
room temperature under a nitrogen atmosphere for 16 h unless
otherwise noted. The reaction solution was then diluted with an
equivalent volume of Et2O, and an equivalent volume of saturated
NH4Cl solution was added. The organic layer was washed (3 ×
H2O, 3 × brine), dried over MgSO4, and evaporated under reduced
pressure. The resulting oil was purified by flash chromatography
(1% Et2O/hexanes f 15% Et2O/hexanes).
General Procedure 3. Allenyl Azide Formation. CH3MgBr (3.0
M in Et2O, 10 equiv) was added dropwise to an ice-cold 0.1 M
solution of CuI (10 equiv) and LiBr (10 equiv) in THF, and the
resulting solution was stirred at that temperature for 30 min. A
solution of the alkynyl azide in THF (0.1 M) was added slowly
via cannula. The ice-cold solution was stirred for 30 min (monitored
by TLC for starting material consumption). Ice-cold saturated
NH4Cl solution was then added dropwise until gas evolution ceased,
and then the reaction mixture was diluted with an equivalent volume
of Et2O. The organic layer was washed (3 × ice-cold H2O, 3 ×
ice-cold brine), dried over MgSO4 and evaporated under reduced
pressure maintaining a water bath temperature below 40 °C. The
resulting oil was purified by flash chromatography (hexanes f 2%
Et2O/hexanes).
General Procedure 4. Copper-Mediated Photolysis of
Allenyl Azides. A flame-dried quartz vessel containing a 5 mM
solution of allenyl azide and 1.5 equiv of CuI in CH3CN was
vigorously purged with N2 or Ar for 5 min. The vessel was
irradiated at the specified wavelength for a minimum of 1 h with
monitoring by TLC for starting material consumption. The reaction
Experimental Section
Computational Methodology. All of the stationary points were
located with the B3LYP20 density functional as implemented in
(21) Frisch, M. J. et al. Gaussian03, reVision C.02; Gaussian, Inc.: Wall-
ingford, CT, 2004.
(22) Bauernschmitt, R.; Ahlrichs, R. J. Chem. Phys. 1996, 22, 9047–9052.
(23) (a) Hess, B. A., Jr.; Eckart, U.; Fabian, J. J. Am. Chem. Soc. 1998, 120,
12310–12315. (b) Hess, B. A., Jr.; Smentek, L.; Brash, A. R.; Cha, J. K. J. Am.
Chem. Soc. 1999, 121, 5603–5604. (c) Lo´pez, C. S.; Faza, O. N.; York, D. M.;
de Lera, A. J. Org. Chem. 2004, 69, 3635–3644. (d) Li, J.; Worthington, S. E.;
Cramer, C. J. J. Chem. Soc., Perkin Trans. 2 1998, 1045–1052.
(20) (a) Hohenberg, P.; Kohn, W. Phys. ReV. 1964, 136, B864-B871. (b)
Kohn, W.; Sham, L. Phys. ReV. A 1965, 140, A1133-A1138. (c) Stephens, P. J.;
Devlin, F. J.; Chabalowski, C. F.; Frisch, M. J. J. Phys. Chem. 1994, 98, 11623–
11637. (d) Becke, A. D. J. Chem. Phys. 1993, 98, 5648–5652. (e) Lee, C.; Yang,
W.; Parr, R. G. Phys. ReV. B 1988, 37, 785–789.
J. Org. Chem. Vol. 74, No. 14, 2009 4969

Feldman et al.
solution was diluted with an equivalent volume of Et2O and washed
2× with ice-cold 10% NH4OH in saturated NH4Cl (1:4), 3× with
ice-cold H2O (until the aqueous phase exhibited pH ) 7), and 3×
with brine, dried over Na2SO4, and evaporated under reduced
pressure. An alternative aqueous workup also was found to be
effective: ice-cold 1 M Na2S2O3, 2 × ice-cold 1 N H3PO4, 3 ×
ice-cold H2O, and 3 × ice-cold brine. The residue was purified by
flash chromatography (2% Et2O/hexanes f 10% Et2O/hexanes).
General Procedure 5. Thermolysis of Allenyl Azides. A flame-
dried round-bottomed flask or sealed tube containing a 0.1 M
solution of allenyl azide in toluene was refluxed for a minimum of
30 min under a N2 atmosphere with monitoring by TLC for starting
material consumption. The solvent was evaporated under reduced
pressure and the residue was purified by flash chromatography (2%
Et2O/hexanes f 10% Et2O/hexanes).
General Procedure 6. Photolysis of Allenyl Azides. A flame-
dried quartz vessel containing a 5 mM solution of allenyl azide in
CH3CN was vigorously purged with N2 or Ar for 5 min. The vessel
was irradiated at the specified wavelength for a minimum of 1 h
with monitoring by TLC for starting material consumption. The
reaction solution was diluted with an equivalent volume of Et2O,
washed 3× with ice-cold H2O and 3× with brine, dried over
Na2SO4, and evaporated under reduced pressure. The residue was
purified by flash chromatography (2% Et2O/hexanes f 10% Et2O/
hexanes).
(relative intensity) 141.1 (M - I, 20), 210.9 (M - t-Bu, 100), 268.0
(M+, 20); HRMS (EI+) calcd for C8H17SiI 268.0144, found
268.0146.
Acetic Acid 1-(2-Azidophenyl)-6-(tert-butyldimethylsilany-
loxy)hex-4-en-2-ynyl Ester (20). Following general procedure 2,
with a reaction time of 4 h, 2.40 g (11.2 mmol) of 12i and 2.08 g
(6.97 mmol) of tert-butyldimethylsilyl-(E)-3-iodo-2-propenyl ether
(17)24 produced 1.99 g (74%) of 20 as a yellow oil: IR (neat) 2127,
1
1747 cm-1; H NMR (300 MHz, CDCl3) δ 7.66 (dd, J ) 7.6, 1.6
Hz, 1H), 7.37 (td, J ) 7.7, 1.6 Hz, 1H), 7.17 (m, 2H), 6.73 (d, J
) 1.8 Hz, 1H), 6.25 (dt, J ) 15.8, 4.1 Hz, 1H), 5.79 (dq, J ) 15.8,
2.2 Hz, 1H), 4.20 (ddd, J ) 4.0, 2.2, 0.4 Hz, 1H), 2.07 (s, 3H),
0.89 (s, 9H), 0.03 (s, 6H); 13C NMR (75 MHz, CDCl3) δ 169.3,
143.9, 137.7, 130.1, 129.2, 128.1, 124.9, 118.2, 107.5, 85.3, 85.0,
62.6, 61.2, 25.7, 20.9, 18.2, -5.5; ES+ m/z (relative intensity) 408.1
(M + Na, 100); HRMS (ES+) calcd for C20H27N3O3NaSi 408.1719,
found 408.1729.
Acetic Acid 1-(2-Azidophenyl)-7-(tert-butyldimethylsilany-
loxy)-4-methylhept-4-en-2-ynyl Ester (21). Following general
procedure 2, 605 mg (2.81 mmol) of alkyne 12i and 573 mg (1.76
mmol) of iodide 1824 produced 514 mg (71%) of 21 as a yellow
1
oil: IR (neat) 2126, 1747 cm-1; H NMR (360 MHz, CDCl3) δ
7.66 (d, J ) 7.6 Hz, 1H), 7.34 (td, J ) 7.7, 1.4 Hz, 1H), 7.17-7.10
(m, 2H), 6.72 (s, 1H), 5.90 (t, J ) 7.2 Hz, 1H), 3.59 (t, J ) 6.8
Hz, 2H), 2.28 (app q, J ) 6.9 Hz, 2H), 2.05 (s, 3H), 1.77 (s, 3H),
0.86 (s, 9H), 0.02 (s, 6H); 13C NMR (90 MHz, CDCl3) δ 169.2,
137.7, 135.8, 130.0, 129.3, 128.3, 124.8, 118.3, 118.1, 89.8, 81.6,
61.9, 61.2, 32.2, 25.8, 20.8, 18.2, 17.0, -5.5; ES+ m/z (relative
intensity) 436.3 (M + Na, 100); HRMS (ES+) calcd for
C22H31N3O3SiNa 436.2032, found 436.2029.
1-Azido-2-(4-methylpenta-1,2,4-trienyl)benzene (13i). Follow-
ing general procedure 1, isopropenylmagnesium bromide (0.5 M
in THF, 4.6 mL, 2.3 mmol) and acetate 12i (200 mg, 0.929 mmol)
produced 90 mg (49%) of 13i as a yellow oil: IR (neat) 2123, 1925
1
cm-1; H NMR (400 MHz, CDCl3) δ 7.40 (dd, J ) 7.8, 1.5 Hz,
Acetic Acid 1-(2-Azidophenyl)-5-(tert-butyldimethylsilanyl)-
pent-4-en-2-ynyl Ester (22). Following general procedure 2, using
20 mol % of PdCl2(PPh3)2, 1.14 g (4.25 mmol) of iodide 19, and
1.46 g (6.80 mmol) of alkyne 12i produced 332 mg (22%) of 22
1H), 7.24 (td, J ) 7.7, 1.5 Hz, 1H), 7.13 (dd, J ) 8.0, 1.1 Hz, 1H),
7.08 (td, J ) 7.5, 1.1 Hz, 1H), 6.70 (d, J ) 6.4 Hz, 1H), 6.34 (d,
J ) 6.5 Hz, 1H), 5.02 (d, J ) 0.6 Hz, 1H), 4.93 (q, J ) 1.5 Hz,
1H), 1.82 (t, J ) 0.6 Hz, 3H); 13C NMR (100 MHz, CDCl3) δ
208.5, 138.5, 136.3, 128.2, 128.1, 125.5, 124.9, 118.4, 114.6, 101.5,
91.8, 19.7; EI+ m/z (relative intensity) 169.1 (M - N2, 70%);
HRMS (EI+) calcd for C12H11N 169.0891, found 169.0973.
1-Azido-2-(3-phenylbuta-1,2-dienyl)benzene (14j). Following
general procedure 3, 400 mg (1.37 mmol) of acetate 12j produced
1
as a yellow oil: IR (neat) 2127, 1746 cm-1; H NMR (360 MHz,
CDCl3) δ 7.67 (dd, J ) 7.6, 1.4 Hz, 1H), 7.37 (td, J ) 7.8, 1.5 Hz,
1H), 7.19-7.13 (m, 2H), 6.75 (d, J ) 1.7 Hz, 1H), 6.50 (d, J )
19.4 Hz, 1H), 5.98 (dd, J ) 19.3, 1.8 Hz, 1H), 2.08 (s, 3H), 0.86
(s, 9H), 0.01 (s, 6H); 13C NMR (90 MHz, CDCl3) δ 169.3, 144.9,
137.7, 130.2, 129.2, 128.0, 124.9, 123.5, 118.2, 87.0, 85.1, 61.2,
26.3, 20.9, 16.5, -6.5; ES+ m/z (relative intensity) 295.1 (M -
TBS + MeOH + Na, 100), 378.2 (M + Na, 10); HRMS (ES+)
calcd for C14H14N3O3Na 295.0933, found 295.0941.
[6-(2-Azidophenyl)-4-methylhexa-2,4,5-trienyloxy]-tert-bu-
tyldimethylsilane (23). Following general procedure 3, 263 mg
(0.682 mmol) of acetate 20 produced 167 mg (72%) of 23 as a
yellow oil: IR (neat) 2123, 1929 cm-1; 1H NMR (400 MHz, CDCl3)
δ 7.35 (dd, J ) 7.7, 1.4 Hz, 1H), 7.24 (td, J ) 7.2, 1.3 Hz, 1H),
7.13 (dd, J ) 8.0, 1.0 Hz, 1H), 7.08 (td, J ) 7.6, 1.0 Hz, 1H), 6.53
(s, 1H), 6.28 (dd, J ) 15.6, 0.7 Hz, 1H), 5.78 (dtd, J ) 15.6, 5.3,1.2
Hz, 1H), 4.31 (t, J ) 1.4 Hz, 1H), 4.29 (t, J ) 1.4 Hz, 1H), 1.96
(s, 3H), 0.95 (s, 9H), 0.12 (s, 6H); 13C NMR (100 MHz, CDCl3) δ
209.7, 136.3, 128.9, 128.4, 128.0, 127.3, 126.0, 124.8, 118.4, 103.3,
88.6, 63.9, 25.9, 18.4, 15.3, -5.2; ES+ m/z (relative intensity) 314.2
(M - N2 + H, 80%); HRMS (AP+) calcd for C19H28NOSi:
314.1940, found 314.1936.
275 mg (81%) of 14j as a yellow oil: IR (neat) 2122, 1931 cm-1
;
1H NMR (300 MHz, CDCl3) δ 7.36-7.33 (m, 2H), 7.31 (dd, J )
8.5, 0.7 Hz, 1H), 7.25-7.20 (m, 2H), 7.15-7.10 (m, 2H), 7.03
(ddd, J ) 8.1, 1.3, 0.4 Hz, 1H), 6.96-6.91 (m, 1H), 6.65 (q, J )
2.9 Hz, 1H), 2.12 (d, J ) 3.0 Hz, 3H); 13C NMR (75 MHz, CDCl3)
δ 207.5, 136.4, 136.1, 128.4, 128.2, 128.1, 127.0, 125.80, 125.77,
124.8, 118.5, 104.5, 91.0, 16.7; ES+ m/z (relative intensity) 220.2
(M - N2 + H, 20), 248.1 (M + H, 10); HRMS calcd for C16H14N
220.1126, found 220.1122.
tert-Butyl-(2-iodovinyl)dimethylsilane (19). To an ice-cold
stirring solution of Zr(Cp)2Cl2 (1.82 g, 6.22 mmol) in 20 mL of
THF was added DIBAL-H (1.0 M in hexanes, 6.22 mL, 6.22 mmol).
The reaction was stirred at that temperature for 30 min, and then
a solution of (tert-butyldimethylsilyl)acetylene (1.06 mL, 5.66
mmol) in 3 mL of THF was added slowly and the reaction allowed
to warm to room temperature. Upon warming, the reaction mixture
turned green and then red. The reaction solution was cooled to -78
°C, a solution of I2 (1.87 g, 7.36 mmol) in 5 mL of THF was added
slowly, and stirring was continued at that temperature for 1 h. The
reaction solution was then allowed to warm to room temperature,
stirring was continued for 2.5 h, ice-cold 1 M HCl (35 mL) was
slowly added, and the mixture was diluted with Et2O (50 mL). The
organic layer was washed with 1 M HCl (50 mL), 1 M Na2S2O3
solution (2 × 50 mL), saturated NaHCO3 solution (50 mL), H2O
(3 × 50 mL), and brine (3 × 50 mL), dried over MgSO4, and
concentrated under reduced pressure. The residue was purified by
flash chromatography (hexanes) to provide 1.14 g (75%) of 19 as
a pale yellow oil: 1H NMR (360 MHz, CDCl3) δ 7.04 (d, J ) 16.1
[7-(2-Azido-phenyl)-4,5-dimethyl-hepta-3,5,6-trienyloxy]-tert-
butyl-dimethyl-silane (24). Following general procedure 3, 512
mg (1.24 mmol) of acetate 21 produced 368 mg (80%) of 24 as a
1
red oil: IR (neat): 2122, 1929 cm-1; H NMR (360 MHz, CDCl3)
δ 7.31 (dd, J ) 7.7, 1.4 Hz, 1H), 7.17 (td, J ) 7.6, 1.2 Hz, 1H),
7.07 (dd, J ) 7.6, 1.2 Hz, 1H), 7.02 (t, J ) 7.5 Hz, 1H), 6.54 (s,
1H), 5.50 (t, J ) 7.2 Hz, 1H), 3.66 (t, J ) 6.9 Hz, 2H), 2.37 (app
q, J ) 6.9 Hz, 2H), 1.94 (d, J ) 2.6 Hz, 3H), 1.75 (s, 3H), 0.89 (s,
9H), 0.06 (s, 6H); 13C NMR (90 MHz, CDCl3) δ 207.9, 136.1,
132.3, 128.0, 127.7, 126.6, 124.7, 122.6, 118.3, 107.6, 89.8, 62.7,
32.4, 25.9, 18.3, 16.2, 15.2, -5.3; ES+ m/z (relative intensity) 342.2
Hz, 1H), 6.65 (d, J ) 16.2 Hz, 1H), 0.86 (s, 9H), 0.02 (s, 6H); 13
C
NMR (90 MHz, CDCl3) δ 147.6, 90.1, 26.2, 16.6, -6.3; EI+ m/z
(24) Huang, Z.; Negishi, E.-i. Org. Lett. 2006, 8, 3675–3678.
4970 J. Org. Chem. Vol. 74, No. 14, 2009

Allenyl Azide Cycloaddition Chemistry
(M - N2 + H, 100%), 387.3 (M + NH4, 60%), 392.2 (M + Na,
30%); HRMS (ES+) calcd for C21H35N4OSi: 387.2580, found
387.2578.
[2-(3-Iodo-6,6-dimethyl-cyclohex-2-enyloxymethoxy)-ethyl]-
trimethylsilane (35c). To a stirring solution of alcohol 33 (355
mg, 1.41 mmol) in 5 mL of CH2Cl2 was added Bu4NI (624 mg,
1.69 mmol), N,N-diisopropylethylamine (dried over Na2SO4 and
decolorized with charcoal) (279 µL, 1.69 mmol), and SEMCl (tech.
grade, 499 µL, 2.82 mmol). The reaction mixture was stirred for
16 h and then the solvent was removed in vacuo. The resulting oil
was dissolved in Et2O (10 mL) and washed (3 × 10 mL H2O, 3 ×
10 mL brine), dried over MgSO4, and then the solvent was removed
in vacuo. The resulting oil was purified by flash chromatography
(hexanes f 10% Et2O/hexanes) to yield 515 mg (96%) of 35c as
a colorless oil: IR (neat): 1635 cm-1;1H NMR (360 MHz, C6D6) δ
6.45 (m, 1H), 4.57 (d, J ) 7.0 Hz, 1H), 4.47 (d, J ) 7.0 Hz, 1H),
3.66 - 3.56 (m, 2H), 3.46 (m, 1H), 2.28 (m, 2H), 1.28 (m, 1H),
1.06 (m, 1H), 0.91 (m, 5H), 0.77 (s, 3H), -0.02 (m, 9H); 13C NMR
(90 MHz, C6D6) δ 138.8, 99.3, 94.6, 81.4, 65.2, 37.7, 35.9, 32.8,
26.3, 21.8, 18.2, -1.2; EI+ m/z (relative intensity) 309.0 (M -
SiMe3, 40%); HRMS (EI+) calcd for C11H18O2I: 309.0352, found
309.0360.
[5-(2-Azido-phenyl)-3-methyl-penta-1,3,4-trienyl]-tert-butyl-
dimethyl-silane (25). Following general procedure 3, 66 mg (0.19
mmol) of acetate 22 produced 34 mg (59%) of 25 as a yellow oil:
1
IR (neat): 2121, 1929 cm-1; H NMR (400 MHz, CDCl3) δ 7.31
(d, J ) 7.6 Hz, 1H), 7.22 (t, J ) 7.7 Hz, 1H), 7.11 (d, J ) 7.9 Hz,
1H), 7.06 (t, J ) 7.3 Hz, 1H), 6.54 (d, J ) 18.9 Hz, 1H), 6.51 (s,
1H), 5.84 (d, J ) 19.0 Hz, 1H), 1.91 (d, J ) 2.5 Hz, 1H), 0.88 (s,
9H), 0.04 (s, 6H); 13C NMR (75 MHz, CDCl3) δ 210.8, 142.3,
136.3, 128.5, 128.0, 126.8, 125.9, 124.8, 118.4, 105.4, 88.6, 26.4,
16.7, 14.6, -6.0, -6.1; ES+ m/z (relative intensity) 283.2 (M -
N2, 50%).
1-Azido-2-{3-[3-(tert-butyl-dimethyl-silanyloxy)-4,4-dimethyl-
cyclohex-1-enyl]-4-trimethylsilanyl-buta-1,2-dienyl}-benzene (28e).
Following general procedure 3 using TMSCH2MgCl (1.0 M in
Et2O), 115 mg (0.251 mmol) of 27e produced 115 mg (95%) of
1
28e as a yellow oil: IR (neat): 2123, 1919 cm-1; H NMR (300
1-Azido-2-{4-(tert-butyl-dimethyl-silanyloxy)-3-[3-(tert-butyl-
dimethyl-silanyloxy)-4,4-dimethyl-cyclohex-1-enyl]-buta-1,2-di-
enyl}-benzene (37). To a -78 °C solution of iodide 35a (2.078 g,
8.243 mmol, 2.5 equiv) in 82 mL of THF was added dropwise
t-BuLi (1.7 M in pentane, 10.7 mL, 18.1 mmol, 5.5 equiv) and
stirring continued at that temperature for 30 min, during which time
the solution turned deep yellow. A solution of ZnCl2 (1.124 g, 8.243
mmol, 2.5 equiv) in 40 mL of THF was then added slowly and the
resulting solution warmed to room temperature during which time
the solution turned colorless. Pd(PPh3)4 (190 mg, 0.165 mmol, 5
mol %) and a solution of acetate 36 (1.185 g, 3.297 mmol, 1 equiv)
in 33 mL of THF were added sequentially and the reaction mixture
was stirred for 45 min (as monitored by TLC). Ice-cold saturated
NH4Cl solution (30 mL) was added slowly and the organic layer
was diluted with Et2O (30 mL). The organic layer was washed with
H2O (3 × 30 mL) and brine (3 × 30 mL), dried over MgSO4 and
concentrated in vacuo. The residue was purified by flash chroma-
tography (2% Et2O/hexanes) to produce 1.132 g (64%) of 37 as a
yellow oil: IR (neat): 2123, 1929 cm-1; 1H NMR (360 MHz, C6D6)
δ 7.41 - 7.49 (m, 1H), 6.85 (d, J ) 5.8 Hz, 1H), 6.78 - 6.81 (m,
2H), 6.65 - 6.68 (m, 1H), 5.96 (s, 1H), 4.49 - 4.60 (m, 2H), 3.96
(s, 1H), 2.02 - 2.15 (m, 2H), 1.44 - 1.49 (m, 1H), 1.28 - 1.34
(m, 2H), 1.01 (m, 9H), 0.98 (s, 3H), 0.93 - 0.97 (m, 11H), 0.18
(s, 3H), 0.10 (s, 3H), 0.07 (m, 6H); 13C NMR (90 MHz, C6D6) δ
208.1, 208.0, 136.7, 136.6, 131.4, 131.3, 128.5, 128.40, 128.35,
128.1, 126.2, 126.1, 125.03, 124.95, 118.71, 118.67, 111.6, 92.54,
92.49, 76.0, 75.8, 62.7, 62.6, 34.4, 34.3, 34.0, 33.9, 27.7, 27.5, 26.2,
26.0, 25.4, 25.3, 20.8, 20.5, 18.4, -3.7, -4.6, -5.0, -5.1; ES+
m/z (relative intensity) 512.3 (M + H, 35%), 562.3 (M + Na,
100%); HRMS (ES+) calcd for C30H49N3O2Si2Na: 562.3261, found
562.3242.
MHz, C6D6) δ 7.45 (m, 1H), 6.81 (m, 3H), 6.71 (m, 1H), 5.68 (d,
J ) 1.2 Hz, 1H), 3.96 (s, 1H), 2.25 - 2.10 (m, 2H), 1.75 (m, 2H),
1.43 (m, 1H), 1.31 (m, 1H), 1.00 (s, 9H), 0.96 (m, 3H), 0.94 (s,
3H), 0.15 (m, 3H), 0.10 (s, 12H);13C NMR (75 MHz, CDCl3) δ
208.1, 207.9, 136.2, 133.2, 133.0, 128.0, 127.7, 127.6, 127.5, 126.7,
124.8, 124.7, 118.4, 108.21, 108.18, 90.71, 90.65, 75.8, 75.7, 34.3,
34.23, 34.18, 34.1, 27.72, 27.66, 25.9, 24.9, 19.9, 19.7, 18.2, 18.1,
17.6, -0.8, -4.2, -4.9; APCI+ m/z (relative intensity) 454.3 (M
+ H - N2, 100%), 482.3 (M + H, 15%); HRMS (AP+) calcd for
C27H44NOSi2: 454.2961, found 454.2979.
3-[3-(2-Azido-phenyl)-1-methyl-propa-1,2-dienyl]-6,6-dimeth-
yl-cyclohex-2-enol (29). To a 0 °C stirring solution of 28a (429
mg, 1.05 mmol) in 10 mL of THF was added n-Bu4NF (1.0 M in
THF, 10 mL, 10 mmol). The reaction mixture was stirred at 0 °C
for 54 h (monitored by TLC for starting material consumption).
Ice-cold saturated NaHCO3 solution (10 mL) was added and then
the reaction mixture was diluted with Et2O (10 mL). The organic
layer was washed (ice-cold brine × 2), dried over Na2SO4 and
evaporated under reduced pressure maintaining a water bath
temperature below 40 °C. The resulting oil was purified by flash
chromatography (neutral alumina, 25% EtOAc/hexanes) to give 289
mg (93%) of 29 as an orange oil: IR (neat): 3393, 2124, 1930 cm-1
;
1H NMR (400 MHz, CDCl3) δ 7.31 (m, 1H), 7.21 (m, 1H), 7.11
(dd, J ) 8.0, 0.9 Hz, 1H), 7.05 (t, J ) 7.8 Hz, 1H), 6.57 (s, 1H),
5.68 (m, 1H), 3.89 (d, J ) 7.9 Hz, 1H), 2.11 - 2.03 (m, 2H), 1.96
(d, J ) 2.7 Hz, 3H), 1.73 (br s, 1H), 1.50 (m, 1H), 1.39 (m, 1H),
0.96 (m, 3H), 0.90 (m, 3H); 13C NMR (100 MHz, CDCl3) δ 207.9,
207.8, 136.11, 136.09, 135.1, 135.0, 127.93, 127.91, 126.1, 125.0,
124.8, 124.7, 118.4, 105.6, 90.2, 74.9, 74.6, 33.6, 33.5, 33.3, 32.9,
26.5, 26.2, 24.6, 24.5, 21.2, 20.7, 15.9; ES+ m/z (relative intensity)
318.1 (M + Na, 100%), 296.2 (M + H, 10%); HRMS (ES+) calcd
for C18H21N3ONa: 318.1582, found 318.1585.
2-Methyl-1-phenyl-8H-3a-aza-cyclopenta[a]indene (40g).
Method A: Following General Procedure 4, 48 mg (0.18 mmol) of
13g provided 30 mg (69%) of 38g1b as a yellow oil. Method B:
Following General Procedure 5, 193 mg (0.71 mmol) of 13g
provided 81 mg (47%) of 38g1b as a yellow oil and 67 mg (39%)
of 40g as a white solid. Method C: Following General Procedure
6, 100 mg (0.37 mmol) of 13g provided 21 mg (23%) of 38g1b as
3-[3-(2-Azido-phenyl)-1-methyl-propa-1,2-dienyl]-6,6-dimeth-
yl-cyclohex-2-enone (30). To a stirring solution of alcohol 29 (57
mg, 0.20 mmol) in 1 mL of CH2Cl2 was added a solution of Dess-
Martin periodinane (170 mg, 0.40 mmol) in 2 mL of CH2Cl2 and
the reaction was stirred for 1 min. Saturated NaHCO3 (3 mL) was
added and the organic layer washed with H2O (3 × 5 mL), and
brine (3 × 5 mL), dried over MgSO4 and concentrated under
reduced pressure. The residue was purified by flash chromatography
(5% EtOAc/hexanes) to provide 24 mg (41%) of 30 as a yellow
oil: IR (neat): 2124, 1927, 1664 cm-1; 1H NMR (360 MHz, C6D6)
δ 7.21 (dd, J ) 7.3, 1.7 Hz, 1H), 6.84 - 6.76 (m, 2H), 6.70 - 6.66
(m, 2H), 6.07 (d, J ) 1.1 Hz, 1H), 2.20 (app pent, J ) 6.1 Hz,
2H), 1.68 (d, J ) 2.7 Hz, 3H), 1.46 (t, J ) 6.2 Hz, 2H), 1.06 (s,
3H), 1.04 (s, 3H); 13C NMR (90 MHz, C6D6) δ 210.4, 202.7, 153.5,
136.9, 129.0, 128.4, 125.2, 124.9, 123.3, 118.9, 106.4, 91.5, 40.6,
36.4, 24.9, 24.3, 15.5; ES+ m/z (relative intensity) 316.1 (M +
Na, 100%); HRMS (ES+) calcd for C18H19N3ONa: 316.1426, found
316.1429.
1
a yellow oil. 40g: mp. 89 - 90 °C; H NMR (400 MHz, CDCl3)
δ 7.40 (m, 2 H), 7.31 (m, 2H), 7.26 (d, J ) 7.4 Hz, 1H), 7.19 -
7.11 (m, 2H), 7.08 (d, J ) 7.7 Hz, 1H), 6.95 (t, J ) 7.3 Hz, 1H),
6.87 (s, 1H), 3.84 (s, 2H), 2.23 (s, 3H); 13C NMR (100 MHz,
CDCl3) δ 140.9, 136.2, 133.8, 132.9, 128.4, 127.5, 127.4, 125.7,
125.2, 122.7, 122.1, 117.5, 109.2, 109.1, 29.7, 12.6; ES+ m/z
(relative intensity) 246.1 (M + H, 100%); HRMS (ES+) calcd for
C18H16N: 246.1283, found 246.1286.
1-(tert-Butyl-dimethyl-silanyloxymethyl)-3-methyl-1,4-dihy-
dro-cyclopenta[b]indole (38 h) and 3-(tert-Butyl-dimethyl-
silanyloxymethyl)-1-methyl-3H-3a-aza-cyclopenta[a]indene (39
h). Following General procedure 5, 95 mg (0.28 mmol) of 23
J. Org. Chem. Vol. 74, No. 14, 2009 4971

Feldman et al.
produced 21 mg of 38h (24%) and 30 mg of 39h (34%). 38h:
2H, J ) 7.4 Hz), 2.07 (s, 3H), 2.02 (s, 3H), 0.93 (s, 9H), -0.02 (s,
6H); 13 C NMR (C6D6) δ ppm: 145.0, 136.1, 130.9, 126.0, 121.8, 120.8,
119.6, 110.5, 110.4, 63.3, 29.4, 27.5, 26.1, 18.1, 9.9, 9.7, -5.4; ES+
m/z (relative intensity) 342.2 (M + H, 100%); HRMS (ES+) calcd
for C21H32NOSi: 342.2253, found 342.2253.
1
(yellow oil): IR (neat): 3401 cm-1; H NMR (400 MHz, C6D6) δ
7.81 (d, J ) 7.4 Hz, 1H), 7.26 (t, J ) 7.0 Hz, 1H), 7.21 (m, 2H),
6.73 (br s, 1H), 6.28 (s, 1H), 4.06 (dd, J ) 9.1, 7.2 Hz, 1H), 3.75
(t, J ) 7.1 Hz, 1H), 3.62 (t, J ) 8.7 Hz, 1H), 1.84 (s, 3H), 1.00 (s,
9H), 0.05 (s, 3H), 0.02 (s, 3H); 13C NMR (90 MHz, C6D6) δ 147.5,
140.4, 134.5, 132.9, 125.6, 122.4, 120.6, 120.3, 119.4, 112.2, 65.3,
48.5, 26.2, 18.6, 13.2, -5.1, -5.3; ES+ m/z (relative intensity)
314.2 (M + H), 100%); HRMS (ES+) calcd for C19H28NOSi:
314.1940, found 314.1944.
6,9,9-Trimethyl-5,7,8,9,10,10a-hexahydro-indeno[2,1-b]indol-
10-ol (42a) and 3,3,1,1-Trimethyl-2,3-dihydro-1H-indolo[1,2-
a]indole (44). Method A: Following general procedure 4, 109 mg
(0.369 mmol) of allene 29 produced 59 mg (60%) of 42a as a brown
solid. Method B: Following general procedure 5, 350 mg (1.18
mmol) of allene 29 produced 107 mg (34%) of 42a as a brown
solid and 72 mg (25%) of 441b as a brown solid. Method C:
Following general procedure 6, 78 mg (0.26 mmol) of allene 29
produced 15 mg (21%) of 42a as a brown solid and 16 mg (24%)
of 44 as a brown solid. 42a: mp. 76 °C (dec.); IR (neat): 3400,
1
(39 h). (yellow oil): H NMR (400 MHz, C6D6) δ 7.75 (d, J )
7.7 Hz, 1H), 7.47 (d, J ) 8.0 Hz, 1H), 7.25 (t, J ) 7.2 Hz, 1H),
7.19 (t, J ) 7.6 Hz, 1H), 6.28 (s, 1H), 5.79 (s. 1H), 4.48 (t, J )
6.0 Hz, 1H), 3.78 (dd, J ) 10.0, 6.0 Hz, 1H), 3.48 (dd, J ) 10.0,
6.0 Hz, 1H), 1.87 (s, 3H), 0.89 (s, 9H), -0.05 (s, 3H), -0.12 (s,
3H); 13C NMR (100 MHz, C6D6) δ 148.5, 135.4, 133.7, 129.9,
122.1, 121.5, 119.6, 110.5, 90.9, 65.2, 64.3, 26.0, 18.4, 12.4, -5.3,
-5.6; ES+ m/z (relative intensity) 314.2 (M + H, 100%), 627.4
(2 M + H, 25%); HRMS (ES+) calcd for C19H28NOSi: 314.1940,
found 314.1951.
1
3301 cm-1; H NMR (400 MHz, C6D6) δ 7.94 (d, J ) 7.7 Hz,
1H), 7.29 - 7.20 (m, 3H), 6.99 (br s, 1H), 3.26 (d, J ) 10.7 Hz,
1H), 2.56 (dd, J ) 10.7, 4.8 Hz, 1H), 2.34 (dd, J ) 14.2, 3.0 Hz,
1H), 2.16 (td, J ) 14.2, 3.2 Hz, 1H), 1.80 - 1.75 (m, 1H), 1.77 (s,
3H), 1.44 (dd, J ) 13.1, 3.2 Hz, 1H), 1.08 (s, 3H), 1.08 - 0.99 (m,
1H), 0.91 (s, 3H); 13C NMR (90 MHz, CDCl3) δ 148.9, 145.5,
139.3, 125.0, 122.5, 120.7, 120.1, 119.5, 118.3, 111.8, 82.6, 49.1,
40.7, 36.2, 28.0, 22.7, 17.5, 10.2; APCI+ m/z (relative intensity)
268.2 (M + H, 100%); HRMS (ES+) calcd for C18H22NO:
268.1701, found 268.1714.
1-(tert-Butyl-dimethyl-silanyl)-3-methyl-1,4-dihydro-cyclo-
penta[b]indole (38i) and 3-(tert-Butyl-dimethyl-silanyl)-1-meth-
yl-8H-3a-aza-cyclopent[a]indene (40i). Method A: Following
general procedure 4, with column chromatography performed at
-78 °C, 84 mg (0.27 mmol) of allene 25 produced 34 mg (44%)
of 38i as a yellow oil. Method B: Following general procedure 5,
75 mg (0.23 mmol) of allene 25 produced 22 mg (34%) of 38i as
a yellow oil. Method B: Following general procedure 6, 1005 mg
(0.30 mmol) of allene 25 produced 21 mg (25%) of 38i as a yellow
oil and 7.4 mg (9%) of 40i as a light yellow oil. 38i: IR (neat):
4-(tert-Butyl-dimethyl-silanyloxy)-3,3-dimethyl-11-trimethyl-
silanylmethyl-2,3,4,4a-tetrahydro-1H-indolo[1,2-a]indole (43e)
and 2-{1-[3-(tert-Butyl-dimethyl-silanyloxy)-4,4-dimethyl-cyclo-
hex-1-enyl]-vinyl}-1H-indole (45e). Method A: Following General
Procedure 5, 135 mg (0.28 mmol) of 28e produced 45 mg (35%)
of 43e as a white solid and 44 mg (41%) of 45e as a brown oil.
Method B: Following General Procedure 6, 60 mg (0.12 mmol) of
28e produced 28 mg (50%) of 43e as a white solid and 5 mg (11%)
of 45e as a brown oil. Crystals of 43e suitable for X-ray
crystallographic analysis were obtained by slow evaporation of an
Et2O/hexanes solution of 43e over a period of 24 h at rt. mp. 105
- 106 °C. 43e: IR (neat): 1456 cm-1; 1H NMR (300 MHz, CDCl3)
δ 7.49 (m, 2H), 7.00 (m, 2H), 6.08 (s, 1H), 4.55 (d, J ) 9.3 Hz,
1H), 3.10 (d, J ) 9.4 Hz, 1H), 2.48 (dd, J ) 14.2, 5.5 Hz, 1H),
2.31 (td, J ) 12.9, 5.1 Hz, 1H), 1.87 (s, 2H), 1.63 (dd, J ) 13.4,
4.0 Hz, 1H), 1.33 (m, 1H), 1.15 (s, 3H), 1.00 (s, 12H), 0.06 (s,
9H), -0.06 (s, 3H), -0.84 (s, 3H); 13C NMR (75 MHz, CDCl3) δ
151.3, 139.6, 137.0, 132.3, 125.0, 120.7, 119.7, 118.8, 112.3, 90.8,
84.1, 65.6, 39.1, 37.8, 29.9, 26.6, 22.7, 18.5, 18.4, 15.5, -0.9, -2.3,
-3.7; ES+ m/z (relative intensity) 454.3 (M + H, 100%); HRMS
(ES+) calcd for C27H44NOSi2: 454.2961, found 454.2949.
1
3410 cm-1; H NMR (360 MHz, CDCl3) δ 7.95 (br s, 1H), 7.57
(m, 1H), 7.37 (m, 1H), 7.07 (m, 2H), 6.35 (app pent, J ) 1.8 Hz,
1H), 3.43 (app pent, J ) 1.8 Hz, 1H), 2.21 (t, J ) 1.8 Hz, 3H),
0.87 (s, 9H), 0.24 (s, 3H), -0.36 (s, 3H);13C NMR (75 MHz,
CDCl3) δ 147.6, 140.3, 133.2, 127.9, 124.6, 123.2, 119.9, 119.3,
119.2, 111.7, 38.0, 27.3, 18.0, 12.8, -4.5, -8.1; MS (ES+) m/z
(relative intensity) 284.2 (M + H, 90%), 567.4 (2 M + H, 100%),
850.6 (3 M + H, 25%); HRMS (ES+) calcd for C18H26NSi:
284.1835, found 284.1837.
40iIR (neat): 2123, 1606 cm-1; 1H NMR (360 MHz, C6D6) δ ppm:
7.68 (d, 1H, J ) 8.3 Hz), 7.21 (s overlap w solvent, 1H), 7.11 (d, 1H,
J) 7.6 Hz), 6.74 (s, 1H), 3.34 (s, 2H), 2.18 (s, 3H), 1.11 (s, 9H), 0.50
(s, 6H); 13C NMR (90 MHz, C6D6) δ ppm: 143.5, 138.7, 135.6, 127.1,
126.1, 122.4, 122.4, 121.0, 111.9, 111.6, 27.6, 27.4, 18.2, 11.1, -3.1;
MS (ES+) m/z (relative intensity) 284.2 (M + H, 100%); HRMS
(ES+) calcd for C18H26NSi: 284.1835, found 284.1816.
1-[2-(tert-Butyl-dimethyl-silanyloxy)-ethyl]-2,3-dimethyl-1,4-
dihydro-cyclopenta[b]indole (38m) and 3-[2-(tert-Butyl-dimeth-
yl-silanyloxy)-ethyl]-1,2-dimethyl-8H-3a-aza-cyclopenta[a]in-
dene (40m). Method A: Following general procedure 4, with
column chromatography performed at -78 °C and irradiation
performed at 350 nm, 75 mg (0.20 mmol) of allene 24 produced
29 mg (42%) of 38m as a yellow oil. Method B: Following general
procedure 5, with column chromatography on deactivated silica gel
(2% H2O/SiO2), 100 mg (0.27 mmol) of allene 24 produced 6 mg
(7%) of 38m as a yellow oil and 37 mg (40%) of 40m as a colorless
45e: IR (neat): 3412 cm-1; 1H NMR (300 MHz, CDCl3) δ 8.05
(br s, 1H), 7.56 (dd, J ) 8.2, 1.1 Hz, 1H), 7.31 (dd, J ) 8.0, 0.8
Hz, 1H), 7.16 (td, J ) 7.0, 1.2 Hz, 1H), 7.08 (td, J ) 7.1, 1.2 Hz,
1H), 6.52 (dd, J ) 2.1, 0.9 Hz, 1H), 5.79 (pent, J ) 1.6 Hz, 1H),
5.34 (s, 1H), 5.22 (s, 1H), 3.88 (dd, J ) 4.9, 2.0 Hz, 1H), 2.25 (m,
2H), 1.65 (m, 1H), 1.45 (m, 1H), 0.96 (s, 3H), 0.92 (s, 3H), 0.90
(s, 9H), 0.05 (s, 3H), 0.00 (s, 3H); 13C NMR (75 MHz, CDCl3) δ
142.3, 137.4, 136.4, 136.0, 131.2, 128.6, 122.2, 120.6, 119.9, 111.9,
110.6, 102.4, 74.7, 33.8, 33.3, 26.8, 25.9, 25.4, 21.4, 18.2, -4.1,
-4.9; ES+ m/z (relative intensity) 382.2 (M + H, 100%); HRMS
(ES+) calcd for C24H36NOSi: 382.2566, found 382.2555.
1
oil. 38m: IR (neat): 3401 cm-1; H NMR (300 MHz, CDCl3) δ
7.94 (br s, 1H), 7.53 (d, J ) 7.7 Hz, 1H), 7.34 (dd, J ) 7.0, 0.9
Hz, 1H), 7.08 (td, J ) 7.1, 0.9 Hz, 1H), 7.01 (t, J ) 7.2 Hz, 1H),
3.83 - 3.71 (m, 2H), 3.34 - 3.31 (m, 1H), 2.35 - 2.24 (m, 1H), 2.03
(d, J ) 1.0 Hz, 3H), 2.01 (d, J ) 0.7 Hz, 3H), 1.60 - 1.55 (m, 1H),
0.89 (s, 9H), 0.04 (s, 3H), 0.02 (s, 3H);13C NMR (90 MHz, CDCl3)
δ 148.4, 145.5, 139.3, 125.0, 124.1, 120.7, 119.9, 119.2, 117.9,
111.7, 61.4, 44.7, 33.9, 26.0, 18.3, 12.9, 10.4, -5.15, -5.22; ES+
m/z (relative intensity) 342.2 (M + H, 100%); HRMS (AP+) calcd
for C21H32NOSi: 342.2253, found 342.2239.
6,9,9-Trimethyl-10-(2-trimethylsilanyl-ethoxymethoxy)-
5,7,8,9,10,10a-hexahydro-indeno[2,1-b]indole (42f) and 3,3,1,1-
Trimethyl-2,3-dihydro-1H-indolo[1,2-a]indole (44). Method A:
Following general procedure 4, 84 mg (0.20 mmol) of allene 28c
produced 47 mg (60%) of 42f as a yellow solid. Method B:
Following general procedure 5, 67 mg (0.16 mmol) of allene 28c
produced 22 mg (35%) of 42f as a yellow solid and 20 mg (51%)
of 441b as a yellow solid. Method C: Following general procedure
6, 57 mg (0.13 mmol) of allene 28c produced 20 mg (38%) of 42f
as a yellow solid and 17 mg (51%) of 44 as a yellow solid.42f:
40m: IR (neat): 2857, 1607, 1098 cm-1; 1H NMR (300 MHz, C6D6)
δ ppm: 7.32 (d, 1H, J ) 3.9 Hz), 7.05 (t, 2H, J ) 7.0 Hz), 6.86 (dt,
1H, J ) 3.7, 0.9 Hz), 3.81 (t, 2H, J ) 7.5 Hz), 3.32 (bs, 2H), 3.13 (t,
1
mp. 139 - 140 °C; IR (neat): 3405 cm-1; H NMR (400 MHz,
CDCl3) δ 7.99 (br s, 1H), 7.55 (d, J ) 7.8 Hz, 1H), 7.32 (d, J )
4972 J. Org. Chem. Vol. 74, No. 14, 2009

Allenyl Azide Cycloaddition Chemistry
7.8 Hz, 1H), 7.09 (t, J ) 7.1 Hz, 1H), 7.03 (t, J ) 7.6 Hz, 1H),
4.91 (d, J ) 6.8 Hz, 1H), 4.61 (d, J ) 6.8 Hz, 1H), 3.87 (m, 1H),
3.52 (m, 1H), 3.38 (d, J ) 10.5 Hz, 1H), 2.73 (d, J ) 10.7 Hz,
1H), 2.54 (m, 1H), 2.36 (m, 1H), 2.01 (s, 3H), 1.66 (m, 1H), 1.30
mL of CH3CN was irradiated at 254 nm for 2 h and then
concentrated under reduced pressure. 1H NMR analysis of the
reaction mixture indicated that it contained a 5: 1 mixture of the E
and Z isomers, respectively. The residue was purified by flash
chromatography (hexanes f 4% EtOAc/hexanes f 8% EtOAc/
hexanes) to provide 157 mg (14%) of the starting material 58a and
400 mg (36%) of 58b as a white solid: mp. 58 -59 °C; IR (neat):
(m, 1H), 1.21 (s, 3H), 1.08 (s, 3H), 0.91 (m, 1H), 0.00 (s, 9H); 13
C
NMR (100 MHz, CDCl3) δ 149.2, 145.9, 139.3, 125.3, 122.3, 120.1,
120.0, 119.4, 118.6, 111.7, 96.8, 89.7, 65.8, 48.7, 41.3, 36.5, 28.7,
22.7, 18.7, 18.0, 10.2, -1.5; ES+ m/z (relative intensity) 398.7
(M + H, 100%), 420.7 (M + Na, 40%); HRMS (ES+) calcd for
C24H36NO2Si: 398.2515, found 398.2521.
1
3321 cm-1; H NMR (360 MHz, CDCl3) δ 7.28 - 7.21 (m, 3H),
7.11 - 7.09 (m, 2H), 6.49 (s, 1H), 4.19 (s, 2H), 2.27 (br s, 1H); 13
C
NMR (90 MHz, CDCl3) δ 151.7, 139.5, 128.5, 128.1, 128.0, 78.3,
67.3; EI+ m/z (relative intensity) 260.3 (M+, 100%); HRMS (EI+)
calcd for C9H9OI 259.9698, found 259.9704.
10-(tert-Butyl-dimethyl-silanyloxy)-6-(tert-butyl-dimethyl-si-
lanyloxymethyl)-9,9-dimethyl-5,7,8,9,10,10a-hexahydro-indeno[2,1-
b]indole (42 g). Following general procedure 4, allene 37 (99 mg,
0.18 mmol) produced 54 mg (59%) of 42g as a tacky yellow solid:
(E)-tert-Butyldimethyl(2-phenylpent-2-en-4-ynyloxy)silane (59).
To a stirring solution of alcohol 58b (395 mg, 1.52 mmol) in 10
mL of Et3N were added trimethylsilylacetylene (433 µL, 3.04
mmol), PdCl2(PPh3)2 (21 mg, 0.30 mmol), and CuI (3 mg, 0.02
mmol), and the reaction mixture was stirred for 24 h at rt and then
diluted with Et2O. The mixture was washed with saturated NH4Cl
solution (3 × 50 mL), H2O (3 × 50 mL), and brine (3 × 50 mL),
dried over MgSO4, and concentrated under reduced pressure. The
crude residue was purified by flash chromatography (hexanes f
5% Et2O/hexanes f 25% Et2O/hexanes) to provide 215 mg (54%)
of the corresponding TMS-alkyne. TBAF (1.0 M in THF, 1 mL, 1
mmol) was added slowly to a solution of the above TMS-alkyne
(215 mg, 0.933 mmol) in 10 mL of THF and stirred for 20 min.
Saturated NaHCO3 (10 mL) was added, and the organic layer was
washed with H2O (3 × 30 mL) and brine (3 × 30 mL), dried over
MgSO4, and concentrated under reduced pressure. The residue was
dissolved in CH2Cl2 (6 mL). Imidazole (95 mg, 1.4 mmol) and
TBSCl (211 mg, 1.40 mmol) were added sequentially, and the
reaction solution was stirred for 20 min, then H2O (10 mL) was
added slowly, and the reaction mixture was diluted with Et2O (50
mL). The organic layer was washed with H2O (3 × 50 mL) and
brine (3 × 50 mL), dried over MgSO4, and concentrated under
reduced pressure. The residue was purified by flash chromatography
(hexanes f 5% Et2O/hexanes) to provide 228 mg (90%) of 59 as
a yellow oil: IR (neat) 3294 cm-1; 1H NMR (400 MHz, CDCl3) δ
7.49 - 7.46 (m, 2H), 7.39 - 7.29 (m, 3H), 5.94 (app q, J ) 2.3 Hz,
1H), 4.45 (dd, J ) 2.0, 0.7 Hz, 2H), 2.90 (d, J ) 2.5 Hz, 1H), 0.94
(s, 9H), 0.10 (s, 6H); 13C NMR (90 MHz, CDCl3) δ 153.1, 137.0,
128.2, 128.1, 127.7, 103.5, 81.6, 80.6, 65.4, 25.8, 18.3, -5.4; ES+
m/z (relative intensity) 273.2 (M + H, 100%); HRMS (ES+) calcd
for C17H25OSi: 273.1675, found 273.1664.
1
IR (neat): 3483 cm-1; H NMR (300 MHz, CDCl3) δ 8.20 (br s,
1H), 7.65 (m, 1H), 7.34 (m, 1H), 7.01 - 7.10 (m, 2H), 4.73 (d, J
) 13.2 Hz, 1H), 4.64 (d, J ) 13.2 Hz, 1H), 3.41 (d, J ) 10.3 Hz,
1H), 3.07 (d, J ) 10.3 Hz, 1H), 2.52 (ddd, J ) 14.2, 5.6, 1.8 Hz,
1H), 2.36 - 2.47 (m, 1H), 1.67 (ddd, J ) 13.1, 5.3, 1.8 Hz, 1H),
1.36 (dd, J ) 13.1, 5.6 Hz, 1H), 1.19 (s, 3H), 1.07 (s, 9H), 1.02 (s,
3H), 0.95 (s, 9H), 0.13 (s, 3H), 0.11 (s, 3H), -0.02 (s, 3H), -0.45
(s, 3H); 13C NMR (75 MHz, CDCl3) δ 148.5, 146.2, 139.2, 127.2,
126.0, 121.8, 120.4, 119.5, 118.7, 111.4, 82.9, 58.6, 50.0, 40.3,
37.7, 29.9, 26.6, 26.0, 23.4, 18.6, 18.4, 18.3, -1.9, -3.0, -5.1,
-5.2; ES+ m/z (relative intensity) 512.2 (M + H, 100%); HRMS
(ES+) calcd for C30H50NO2Si2: 512.3380, found 512.3379.
6,9,9-Trimethyl-8,9-dihydro-7H-indeno[2,1-b]indole-5-car-
boxylic Acid tert-Butyl Ester (46). To a 0 °C stirring solution of
alcohol 42a (198 mg, 0.741 mmol) in 7.5 mL of CH2Cl2 was added
DMAP (904 mg, 7.41 mmol) and Boc2O (403 mg, 1.85 mmol)
sequentially and then the reaction was warmed to rt and stirred for
1 h. The solvent was then removed under reduced pressure and the
residue was dissolved in Et2O (15 mL) and washed with H2O (3 ×
15 mL), and brine (3 × 15 mL), dried over MgSO4, and
concentrated under reduced pressure. The residue was purified by
flash chromatography (5% Et2O/hexanes) to provide 150 mg (58%)
of 46 as an orange oil: IR (neat): 1736 cm-1; 1H NMR (400 MHz,
CDCl3) δ 8.01 (dd, J ) 8.3, 0.8 Hz, 1H), 7.51 (d, J ) 7.6 Hz, 1H),
7.17 (td, J ) 7.5, 1.0 Hz, 1H), 7.10 (ddd, J ) 8.5, 7.3, 1.4 Hz,
1H), 6.58 (s, 1H), 2.64 (td, J ) 6.0, 1.3 Hz, 2H), 2.23 (s, 3H), 1.75
(t, J ) 6.7 Hz, 2H), 1.69 (s, 9H), 1.20 (s, 6H);13C NMR (100 MHz,
CDCl3) δ 150.2, 148.7, 141.9, 139.7, 133.6, 132.4, 125.5, 124.8,
123.0, 121.7, 117.6, 116.8, 116.3, 83.9, 38.6, 33.7, 28.4, 28.3, 20.5,
14.0; ES+ m/z (relative intensity) 350.2 (M + H, 100%); HRMS
(ES+) calcd for C23H28NO2: 350.2120, found 350.2124.
Acetic Acid 1-(2-Azidophenyl)-6-(tert-butyldimethylsilany-
loxy)-5-phenylhex-4-en-2-ynyl Ester (60). To a -30 °C stirring
solution of alkyne 59 (278 mg, 0.837 mmol) in 8 mL of THF was
added slowly n-BuLi (2.5 M in hexanes, 340 µL, 0.84 mmol).
Stirring was continued for 1 h, and then a solution of 2-azidoben-
zaldehyde (123 mg, 0.837 mmol) in 8 mL of THF was added.
Stirring was continued for 20 min (as monitored by TLC for starting
material consumption), and then Ac2O (119 µL, 1.26 mmol) was
added, and the reaction solution was allowed to warm to rt. Ice-
cold saturated NH4Cl (10 mL) was added dropwise, and the mixture
was diluted with Et2O (25 mL). The organic layer was washed with
H2O (3 × 25 mL) and brine (3 × 25 mL), dried over MgSO4, and
concentrated under reduced pressure. The residue was purified by
flash chromatography (5% Et2O/hexanes f 15% Et2O/hexanes) to
provide 320 mg (83%) of 60 as a yellow oil: IR (neat) 2126, 1746
6,6,9,9-Tetramethyl-7,8,9,10-tetrahydro-6H-indeno[2,1-b]in-
dole-5-carboxylic Acid tert-Butyl Ester (47). To a -40 °C stirring
solution of fulvene 46 (27 mg, 0.077 mmol) in 4 mL of THF was
added LiAlH4 (1.0 M in THF, 120 µL, 0.12 mmol) and the reaction
solution was allowed to warm to rt. CH3I (15 µL, 0.24 mmol) was
then added and stirring was continued for 30 min, then ice-cold
saturated NH4Cl (4 mL) was added slowly and the mixture was
diluted with Et2O (15 mL). The organic layer was washed with
H2) (3 × 15 mL) and brine (3 × 15 mL), dried over MgSO4 and
concentrated under reduced pressure. The residue was purified by
flash chromatography (5% Et2O/hexanes) to provide 24 mg (85%)
1
of 47 as a yellow oil: IR (neat): 1729 cm-1; H NMR (360 MHz,
CDCl3) δ 8.14 (m, 1H), 7.56 (m, 1H), 7.20 - 7.17 (m, 2H), 2.41 (t,
J ) 2.3 Hz, 2H), 2.15 - 2.12 (m, 2H), 1.71 (s, 9H), 1.54 (t, J ) 6.2
Hz, 2H), 1.39 (s, 6H), 1.00 (s, 6H); 13C NMR (90 MHz, CDCl3) δ
151.8, 150.1, 145.0, 139.5, 127.4, 126.4, 124.4, 122.7, 122.5, 118.6,
116.4, 83.7, 48.5, 37.5, 35.9, 29.7, 28.4, 28.2, 20.6, 19.0; ES+ m/z
(relative intensity) 366.3 (M + H, 100%); HRMS (ES+) calcd for
C24H32NO2: 366.2433, found 366.2446.
1
cm-1; H NMR (360 MHz, CDCl3) δ 7.35-7.32 (m, 2H), 7.30
(dd, J ) 7.7, 1.1 Hz, 1H), 7.24 (td, J ) 7.9, 1.4 Hz, 1H), 7.21-7.18
(m, 3H), 7.00-6.92 (m, 2H), 6.59 (d, J ) 1.5 Hz, 1H), 5.91 (app
q, J ) 2.0 Hz, 1H), 4.37 (d, J ) 1.7 Hz, 2H), 1.94 (s, 3H), 0.84 (s,
9H), 0.00 (s, 6H); 13C NMR (90 MHz, CDCl3) δ 169.3, 153.1,
137.6, 137.0, 130.0, 129.4, 128.02, 127.95, 127.8, 127.7, 124.7,
118.0, 103.6, 88.0, 85.3, 65.3, 61.4, 25.8, 20.8, 18.2, -5.5; ES+
m/z (relative intensity) 484.1 (M + Na, 80), 507.2 (M + 2Na, 70);
HRMS (ES+) calcd for C26H31N3O3SiNa 484.2032, found 484.2033.
[4-(2-Azidobenzyl)-3-methylnaphthalen-1-ylmethoxy]-tert-bu-
tyldimethylsilane (61). Following general procedure 3, 320 mg
(E)-3-Iodo-2-phenyl-prop-2-en-1-ol (58b). A solution of (Z)-
3-Iodo-2-phenyl-prop-2-en-1-ol (58a)25 (1.12 g, 4.31 mmol) in 160
(25) Duboudin, J. G.; Jousseaume, B.; Bonakdar, A. J. Organomet. Chem.
1979, 168, 227–232.
J. Org. Chem. Vol. 74, No. 14, 2009 4973

Feldman et al.
(0.693 mmol) of 60 produced 120 mg (41%) of 61 as a tacky yellow
Facility (CHE 0131112), is gratefully acknowledged. C.S.L. is
grateful to CESGA for supercomputer time.
1
solid: IR (neat) 2120 cm-1; H NMR (360 MHz, CDCl3) δ 7.73
(dd, J ) 7.5, 1.7 Hz, 1H), 7.54 (dd, J ) 7.6, 1.8 Hz, 1H), 7.17-7.13
(m, 2H), 6.97-6.90 (m, 3H), 6.56 (td, J ) 7.1, 2.0 Hz, 1H), 6.23
(d, J ) 7.8 Hz, 1H), 4.96 (s, 2H), 4.08 (s, 2H), 2.17 (s, 3H), 0.74
(s, 9H), -0.08 (s, 6H); 13C NMR (90 MHz, CDCl3) δ 137.7, 135.2,
134.2, 132.9, 131.53, 131.48, 129.9, 129.0, 127.28, 127.25, 127.1,
125.9, 124.8, 124.6, 123.7, 117.6, 63.4, 28.9, 26.0, 20.4, 18.5, -5.2;
ES+ m/z (relative intensity) 418.3 (M + H, 100), 435.3 (M + NH4,
70), 440.3 (M + Na, 70); HRMS (ES+) calcd for C25H32N3OSi
418.2315, found 418.2322.
Supporting Information Available: General experimental
1
procedures, copies of H and 13C NMR spectra for 13i, 14j,
19-25, 27c, 28c, 29, 30, 35c, 37, 38h,i,m, 39n, 40g, 42a,f,g,
43e, 45e, 46, 47, 58b, and 59-61, details of the X-ray
crystallographic structural determinations of 39n, 40a, 42b,d,
43e, and 54, 2-D energy profiles for the 64 f 67 conversion
and its desmethyl lower homologue, and Cartesian coordinates
and SCF energies of all of the computed structures. This material
is available free of charge via the Internet at http://pubs.acs.org.
Acknowledgment. Support from the National Institutes of
Health, General Medical Sciences division (GM72572), and the
National Science Foundation for the X-ray Crystallography
JO900659W
4974 J. Org. Chem. Vol. 74, No. 14, 2009