Allenl azide cycloaddtion chemistry. Photochemical initiation and cul mediation leads to improved regioselectivity

Article abstract of DOI:10.1021/ol800429j

Irradiation of 2-(3-alkenyl)allenylphenyl azides in the presence of excess Cul furnished functionalized 2,3-cyclopentenylindoles in good yield with only trace amounts of competitive C-N-bonded regioisomeric products. These results represent a significant departure from the modest to-nonexistent regioselectivity that attended thermal cyclization of these allenyl azide substrates.

Full text of DOI:10.1021/ol800429j

ORGANIC
LETTERS
2008
Vol. 10, No. 8
1665-1668
Allenyl Azide Cycloaddition Chemistry.
Photochemical Initiation and CuI
Mediation Leads to Improved
Regioselectivity
Ken S. Feldman,*,† D. Keith Hester II,† Carlos Silva Lo´pez,‡ and
Olalla Nieto Faza‡
Department of Chemistry, The PennsylVania State UniVersity, UniVersity Park,
PennsylVania 16802, Departmento de Quimica Organica, UniVersidade de Vigo,
Lagoas Marcosende, 36200 Vigo, Galicia, Spain, and Department of Chemistry,
UniVersity of Minnesota, 207 Pleasant Street SE, Minneapolis, Minnesota 55455-0431
ksf@chem.psu.edu
Received February 25, 2008
ABSTRACT
Irradiation of 2-(3-alkenyl)allenylphenyl azides in the presence of excess CuI furnished functionalized 2,3-cyclopentenylindoles in good yield
with only trace amounts of competitive C N-bonded regioisomeric products. These results represent a significant departure from the modest-
−
to-nonexistent regioselectivity that attended thermal cyclization of these allenyl azide substrates.
The synthesis of annelated indoles from thermolysis of
2-allenylphenyl azides was described recently, Scheme 1,
Scheme 1. Initial Photochemical and Thermal Results
entry a.1 This cascade reaction sequence displayed little
regioselectivity during the terminal bond formation, and both
C-C- and C-N-bonded products were observed in roughly
equal proportions. An ongoing interest in the total synthesis
of indole-terpene-derived alkaloids exemplified by, inter alia,
the fisherindoles,2 yuehchukene,3 and the nodulisporanes4
† Pennsylvania State University.
‡ Universidade de Vigo and University of Minnesota.
(1) Feldman, K. S.; Iyer, M. R.; Hester, S. K., II. Org. Lett. 2006, 8,
3113-3116.
(2) (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.
would be advanced if selectivity for the C-C-bonded isomer
could be increased. As detailed below, new studies have
revealed that the combination of photochemical initiation and
CuI together provide this enhanced regioselectivity, with
exclusive formation of the C(2)-C(3) annelated indole
product analogous to 2 in the most favorable cases.
(3) Kong, Y.-C.; Cheng, K.-F.; Cambie, R. C.; Waterman, P. G. J. Chem.
Soc., Chem. Commun. 1885, 47-48.
(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.
10.1021/ol800429j CCC: $40.75
© 2008 American Chemical Society
Published on Web 03/18/2008

The initial foray into allenyl azide photochemistry did not
provide promising results, Scheme 1, entry b. A nearly equal
mixture of the C-C- and C-N-bonded products resulted
from irradiation of the unadorned cyclohexenyl substrate 1
in CH3CN (Rayonet photochemical reactor, 0.005 M, N2
purge) for 2 h. Although photochemistry offered no advan-
tage over thermolysis for this substrate, these results did
demonstrate, for the first time, that the allenyl azide cycli-
zation cascade could be triggered at lower temperatures
through simple irradiation.
Extension of this photochemistry to the more highly
functionalized substrate 8 led to surprising and irreproducible
results: 9/11 regioisomer ratios from 1:1 to 6:1, depending
upon batch (Scheme 2). This variability was traced to the
presence of contaminants that survived chromatographic
Table 1. Optimization of the Formation of C-C-Bonded
Tetracycle 9 from Allenyl Azide 8
9/10
ratiob
9
(%)c
11
entry
conditionsa
(%)c
a
b
c
d
e
f
g
h
i
110 °C
254 nm
254 nm, 50% CuI
254 nm, 80% CuI
254 nm, 120% CuI
254 nm, 150% CuI
110 °C, 150% CuI
110 °C, 150% CuI, 35 mM
110 °C, 150% CuI, 51 mM
1:1.6
1:1
4:1
6:1
9:1
10:1
1:1.6
3.5:1
5:1
30
33
51
62
61
62
31
48
52
50
34
16
10
6
8
44
18
13
a CH3CN solvent, 5 mM in 8 unless otherwise noted, quartz vessel for
the irradiations; mol % CuI indicated. The temperature of the photochemical
reaction solution was 30-35 °C. b From integration of diagnostic signals
in the 1H NMR spectrum of the crude reaction mixture. c Chromatographi-
cally pure material.
Scheme 2. Synthesis and Cyclization of Allenyl Azide
Substrate 8
examined (Table 1, entries g-i). Indeed, the inclusion of
CuI in the thermal process had the same favorable effect as
in the photochemical series, but higher concentrations of
substrate were necessary to achieve satisfactory results.
A survey of various allenyl azide substrates revealed that,
with one exception, the high selectivity for C-C bond
formation observed with 8 carries over to a range of
substituent patterns, Table 2 and Schemes 2 and 3. The
syntheses of the allenyl azide substrates examined in this
Table 2. Copper-Mediated Photochemical Cyclization of
Representitive Allenyl Azides
purification of the starting allene, which was formed in the
presence of a large excess of cuprate reagent. Photochemical
trials in the presence of MgBr2, LiBr, and various copper
salts, alone and in combination, eventually led to the
observation that irradiation of 8 in the presence of stoichio-
metric (and greater) quantities of CuI reproducibly furnished
tetracyclic product that was strongly biased toward the C-C-
bonded regioisomer 9, Table 1, entries c-f.
13/14
13
14
entry
R
R1
R2
R3
ratioa
(%)b (%)b
a
b
c
dd
e
CH3
CH3
-(CH2)4-
Ph
H
CH3
Ph
H
H
H
H
>10:1c
>10:1c
10:1
> 10:1c
2.2:1
54
67
55
69
53
H
H
H
Ph
CH2OTBS
Ph
3
This tetracyclic material was formed as a single stereo-
isomer whose relative stereochemistry was first suggested
CH3
Ph
26
1
by analysis of H coupling constants and later verified by
a From integration of diagnostic signals in the 1H NMR spectrum of the
crude reaction mixture. b Chromatographically pure material. c Estimated
single-crystal X-ray analysis5 (see Supporting Information).
The ether 10 did not survive SiO2 chromatography, and only
the dehydration product 11 could be recovered. The isolated
yields of 9 and 11 did not always reflect the crude 9/10 ratios,
as variable losses upon chromatography occurred, and so both
sets of values are reported.
from H NMR detection limits. d Irradiated at 300 nm.
1
study either were described earlier1 or followed chemistry
similar to that developed for 8 (see Supporting Information
for experimental details). That the allylic oxygen function
of 8 did not play a role in enhancing the regioselectivity of
bond formation is implied by the cyclization results of the
simple cyclohexenyl substrate 1 (Table 2, entry a; compare
to Scheme 1), a species that proceeded to product with similar
yield and indistinguishable regioselectivity as the allylic ether
Is it the photochemistry or is it the copper? To probe this
question, thermolysis of 8 in the presence of CuI was
(5) Cambridge Crystallographic Data Centre deposition numbers: 9,
650218; 14e, 649706; 16b, 652945; 22, 651276. The data can be obtained
free from the Cambridge Cyrstallographic Data Centre via www.ccdc.
cam.ac.uk/data_request/cif.
1666
Org. Lett., Vol. 10, No. 8, 2008

porting Information),5 whereas the relative stereochemistry
of 17 remains unassigned.
Scheme 3. Copper Iodide-Mediated Photochemical
Cyclization of Some Oxygenated Cyclohexenyl Allenyl Azides
The silylated allene substrate 18 provides some insight
into the possible role that copper plays in steering the product
regiochemistry strongly toward the C-C-bonded product 20
(Scheme 4). Upon thermolysis with or without copper, 18
provides only a trace of 20 and none of the C-N-bonded
Scheme 4. Silylated Allenyl Azide Substrate 18 Highlights the
Distinct Consequences of Copper Mediation
a Ratio of isolated, chromatographically pure material (1H NMR
ratio from the crude reaction mixture.
index case. Substrates 12b, 12c, and 12d (Table 2, entries
b-d) demonstrate that the cyclohexene ring plays no decisive
role in the enhanced selectivity, as these simple alkene
substrates also provide almost exclusively the C-C-bonded
product 13. Furthermore, entry c demonstrates that a func-
tionalized variant of the allene substituent, R ) CH2OTBS
(Table 2, entry c), is tolerated and furnishes only the C-C
bonded product in measurable amounts. The last entry of
Table 2 provides the first example of this cyclization cascade
that fashions a C-C bond to a quarternary carbon. The
structural assignment of 14e derives from a single-crystal
X-ray analysis5 (see Supporting Information). This latter case
is distinctive for two reasons: (1) There appears to be
significant erosion in the regioselectivity for C-C bond
formation even under the copper-mediated conditions (com-
pare 12e with hν/no copper, 1.9:1 13e/14e). (2) Substrate
12e did not engage in undesired electrocyclizations of the
allenyl-vinyl-(Z)-phenyl assembly, despite some indirect
precedent for this type of process.6 In contrast, substrates
bearing a methyl group in the (Z)-alkenyl position did not
survive long enough to test in the allenyl azide cyclization
cascade, as they appeared to rearrange through [1,5] H-shifts.7
Two further examples of cyclohexenol-derived allenyl
azide substrates were examined, Scheme 3. TIPS ether 15a
was designed to test the relationship between steric bulk at
the allylic ether position and selectivity. The formation of a
mixture of the C-C-bonded product 16a and the elimination
product 11 in a ratio even more favorable than the OTBS
case 8 suggests that the regioselectivy may be susceptible
to fine-tuning by peripheral steric influences. The stereo-
chemistry of 16a was assigned by comparison of its 1H NMR
spectral data to those of 9, whose structural assignment was
secured by X-ray analysis. The silylated cyanohydrin sub-
strate 15b delivered the expected C-C-bonded product 16b
as the strongly favored isomer as expected, but in this
instance, elimination of OTBS from the C-N-bonded isomer
17 was not observed. Tetracycles 16b and 17 were isolated
as single stereoisomers; the relative stereochemistry of 16b
was ascertained by single-crystal X-ray analysis (see Sup-
regioisomer 21. Rather, the aromatic triazole 19 predominates
along with lesser amounts of the unexpected formal aceto-
nitrile adduct 22 (structure by single-crystal X-ray analysis;5
see Supporting Information). Apparently, the very nucleo-
philic allylic silane moiety of 23 is readily trapped by either
adventitious protons (to form 19) or, quite remarkably, by
acetonitrile to provide 22.8 Upon irradiation without copper,
all products observed can be rationalized by citing reaction
through the expected [3 + 2] cycloaddition intermediate 23,
although the lower temperature of the photochemical reaction
may suffice to minimize allylsilane reactivity. On the other
hand, reaction of 18 under the copper-mediated photochemi-
cal conditions leads to a different result; formation of the
C-C-bonded indole 20 as the major product accompanied
by only traces of the triazole. The triazole in this instance
may be formed by competitive non-copper-mediated chem-
istry, but the formation of 20 upon copper intercession clearly
requires consideration of an alternative reaction course
compared to the non-copper case.
The role that copper might play in this complex transfor-
mation is illustrated with substrate 1 (Scheme 5). The key
point of departure as a result of copper mediation may be
the formation of the copper-bound indolidenes 25a/26a.
These species might originate through standard dipolar
cycloaddition/N2 extrusion to form 25/26, followed by
capture of the reactive indolidenes by ligated copper, or they
might owe their genesis to an entirely different mechanistic
course featuring a copper nitrene 289 and a formal copper-
(8) Mukai, C.; Kobayashi, M.; Kubota, S.; Takahashi, Y.; Kitagaki, S.
J. Org. Chem. 2004, 69, 2128-2136.
(9) (a) Li, Z.; Quan, R. W.; Jaconsen, 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.
(6) Elnager, H. Y.; Okamura, W. H. J. Org. Chem. 1988, 53, 3060-
3066.
(7) Barrack, S. A.; Okamura, W. H. J. Org. Chem. 1986, 51, 3201-
3206.
Org. Lett., Vol. 10, No. 8, 2008
1667

energetics are perturbed. The activation energy for the
conversion of the Cu-N-bound intermediate 26a into 3‚CuLn
(not shown) is elevated substantially compared to the copper-
free case (27 vs 18 kcal/mol, respectively), whereas the
barrier for the 25a f 27 cyclization is essentially insensitive
to the presence of copper (ca. 15 kcal/mol). Furthermore,
the calculated barrier to 25a/26a equilibration (24 kcal/mol)
is now lower than this 26a cyclization value (27 kcal/mol).
Thus, the notable advance of these studies, that photochem-
istry in the presence of CuI leads to high levels of selectivity
for the C-C bond cyclopentannelated indole product, can
be attributed to formation of a readily equilibrating pair of
indolidene isomers 25a/26a from which the E isomer 25a
proceeds to product much faster than the Z isomer 26a.
Scheme 5. Mechanistic Analysis of the Copper-Mediated
Photochemical Cyclization of Allenyl Azides
In summary, the inclusion of CuI in the photochemical
decomposition of 2-allenylphenyl azides diverts the complex
cascade sequence down a mechanistic pathway that appears
to be distinct from that of the non-copper-mediated process.
High levels of regiochemical control for the C-C-bonded
indole product can be achieved, possibly signifying the
intermediacy of a copper nitrene species. Whereas the mech-
anistic intricacies still remain to be unraveled, the significant
improvement in reaction regiochemistry may render this
transformation useful in the synthesis of C(2)-C(3) cyclo-
pentannelated indole natural products.
(III) intermediate 29.10 The experimental results do not allow
differentiation between these two pathways in the photo-
chemical series, but the predominant formation of 19/22 from
18 under thermolysis/CuI is only consistent with operation
of the former process.
Density functional theory calculations on a model system
support this view (HCN instead of CH3CN as a copper
ligand, vinyl and proton as allene terminal substituents; see
Supporting Information for calculational details and further
discussion). In the absence of copper, prior calculations
indicated that the indolidene pair 25/26 does not equilibrate
faster than each electrocyclizes to product (2 from 25 and 3
from 26), and so the 2/3 product mixture reflects the
relatively unbiased partitioning of 24 into 25 and 26.11
However, with copper present, calculations suggest that these
Acknowledgment. Support from the National Institutes
of Health, General Medical Sciences division (GM72572),
and from the National Science Foundation for the X-ray
Crystallography Facility (CHE 0131112) is gratefully
acknowledged. C.S.L. is grateful to CESGA for supercom-
puter time.
Supporting Information Available: Full experimental
details and spectroscopic data for 5-9, 11, 12d,e, 13d,e, 14e,
15a,b, 16a,b, and 17-22; X-ray crystallographic data and
thermal ellipsoid polts for 9, 14e, 16b, and 22; computational
details and energy level diagrams for Scheme 6. This material
is available free of charge via the Internet at http://pubs.acs.org.
OL800429J
(10) Bertz, S. H.; Cope, S.; Murphy, M.; Ogle, C. A.; Taylor, B. J. J.
Am. Chem. Soc. 2007, 129, 7208-7209.
(11) 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.
1668
Org. Lett., Vol. 10, No. 8, 2008