Indolyne experimental and computational studies: Synthetic applications and origins of selectivities of nucleophilic additions

Article abstract of DOI:10.1021/ja1086485

Efficient syntheses of 4,5-, 5,6-, and 6,7-indolyne precursors beginning from commercially available hydroxyindole derivatives are reported. The synthetic routes are versatile and allow access to indolyne precursors that remain unsubstituted on the pyrrole ring. Indolynes can be generated under mild fluoride-mediated conditions, trapped by a variety of nucleophilic reagents, and used to access a number of novel substituted indoles. Nucleophilic addition reactions to indolynes proceed with varying degrees of regioselectivity; distortion energies control regioselectivity and provide a simple model to predict the regioselectivity in the nucleophilic additions to indolynes and other unsymmetrical arynes. This model has led to the design of a substituted 4,5-indolyne that exhibits enhanced nucleophilic regioselectivity.

Full text of DOI:10.1021/ja1086485

Published on Web 11/29/2010
Indolyne Experimental and Computational Studies: Synthetic
Applications and Origins of Selectivities of Nucleophilic
Additions
G-Yoon J. Im, Sarah M. Bronner, Adam E. Goetz, Robert S. Paton,
Paul H.-Y. Cheong, K. N. Houk,* and Neil K. Garg*
Department of Chemistry and Biochemistry, UniVersity of California, Los Angeles,
California 90095, United States
Received September 24, 2010; E-mail: houk@chem.ucla.edu; neilgarg@chem.ucla.edu
Abstract: Efficient syntheses of 4,5-, 5,6-, and 6,7-indolyne precursors beginning from commercially
available hydroxyindole derivatives are reported. The synthetic routes are versatile and allow access to
indolyne precursors that remain unsubstituted on the pyrrole ring. Indolynes can be generated under mild
fluoride-mediated conditions, trapped by a variety of nucleophilic reagents, and used to access a number
of novel substituted indoles. Nucleophilic addition reactions to indolynes proceed with varying degrees of
regioselectivity; distortion energies control regioselectivity and provide a simple model to predict the
regioselectivity in the nucleophilic additions to indolynes and other unsymmetrical arynes. This model has
led to the design of a substituted 4,5-indolyne that exhibits enhanced nucleophilic regioselectivity.
Introduction
The indole heterocycle (1, Figure 1) is observed in an
astonishing number of natural products and medicinal agents.1,2
More than 10,000 biologically active indole derivatives have
been discovered to date, of which over 200 are currently
marketed as pharmaceuticals or are undergoing clinical trials.3
These include the common chemotherapeutics vincristine,
vinorelbine, and vinblastine.4 Despite the development of an
impressive array of strategies to synthesize C2- and C3-
functionalized indoles,5 methods to access benzenoid-substituted
indoles remain limited.1
Figure 1. Indole reactivity and indolyne targets 4-6.
Our approach to benzenoid-substituted indoles contrasts with
the usual paradigm of indole reactivity. The indole heterocycle
is typically exploited for its nucleophilic character (1 f 2, Figure
1);6 however, methods for rendering indoles susceptible to attack
by nucleophiles are rare (1 f 3).7 We envisioned that aryne
derivatives of indoles, or “indolynes” (e.g., 4-6), could serve
as electrophilic indole surrogates. This umpolung of the indole
reactivity would not only be conceptually interesting but also
provide unconventional strategies to access polysubstituted
indoles. Such motifs are often considered medicinally ‘privi-
leged’ and are also observed in countless natural products, such
as the complex bioactive alkaloids 7-18 (Figure 2).
Whereas arynes have been studied extensively over the past
century,8,9 indole derivatives of arynes have received less
attention. Indolynes were first reported in the 1960s by Igolen
and co-workers.10 In a key experiment, it was found that C3-
unsubstituted 4,5-indolyne 20 could be generated from 5-bro-
(1) (a) Sundberg, R. J. The Chemistry of Indoles; Academic Press: New
York, 1970. (b) Sundberg, R. J. Pyrroles and Their Benzoderivatives:
Synthesis and Applications. In ComprehensiVe Heterocyclic Chemistry;
Katritzky, A. R., Rees, C. W., Eds.; Pergamon Press: Oxford, 1984;
Vol. 4, pp 313-376. (c) Sundberg, R. J. Indoles (Best Synthetic
Methods); Academic Press: New York, 1996; pp 7-11. (d) Joule, J. A.
Indole and its Derivatives. In Science of Synthesis: Houben-Weyl
Methods of Molecular Transformations; Thomas, E. J., Ed.; George
Thieme Verlag: Stuttgart, 2000; Category 2, Vol. 10, Chapter 10.13.
(2) (a) Horton, D. A.; Bourne, G. T.; Smyth, M. L. Chem. ReV. 2003,
103, 893–930. (b) de Sá Alves, F. R.; Barreiro, E. J.; Fraga, C. A. M.
Mini-ReV. Med. Chem. 2009, 9, 782–793.
(3) MDL Drug Data Report; MDL Information Systems Inc.: San Leandro,
CA.
(4) (a) Kidwai, M.; Venkataramanan, R.; Mohan, R.; Sapra, P. Curr. Med.
Chem. 2002, 9, 1209–1228. (b) Guerite, F.; Fahy, J. The Vinca
Alkaloids. In Anticancer Agents from Natural Products; Cragg, G. M.,
Kingston, D. G. I., Newman, D. J., Eds.; CRCC Press LLC: Boca
Raton, 2005; pp 123-135. (c) American Cancer Society Chemotherapy
Principles. http://www.cancer.org/docroot/eto/eto_1_3_chemotherapy_
principles.asp (accessed Oct 10, 2009).
(6) For the nucleophilicity of N-methylindole, see: Mayr, H.; Kempf, B.;
Ofial, A. R. Acc. Chem. Res. 2003, 36, 66–77.
(7) For the activation of indoles with stoichiometric chromium, see: (a)
Semmelhack, M. F.; Rhee, H. Tetrahedron Lett. 1993, 34, 1399–1402.
(b) Semmelhack, M. F.; Knochel, P.; Singleton, T. Tetrahedron Lett.
1993, 34, 5051–5054. (c) Semmelhack, M. F.; Wulff, W.; Garcia, J. L.
J. Organomet. Chem. 1982, 240, C5–C10. For an example of Pd-
catalyzed arylation in complex molecule synthesis, see: (d) MacKay,
J. A.; Bishop, R. L.; Rawal, V. H. Org. Lett. 2005, 7, 3421–3424.
(5) For a recent review, see: Bandini, M.; Eichholzer, A. Angew. Chem.,
Int. Ed. 2009, 48, 9608–9644.
9
10.1021/ja1086485  2010 American Chemical Society
J. AM. CHEM. SOC. 2010, 132, 17933–17944 17933

A R T I C L E S
Im et al.
Figure 2. Benzenoid-substituted indole natural products 7-18.
Buszek’s laboratory demonstrated that C3-substituted indolynes
could be generated from dihaloindoles in the presence of
butyllithium reagents.12a The presumed indolyne intermediates
were trapped with furan to afford Diels-Alder products (e.g.,
23 + 24 f 26). Further studies have been reported by Buszek,12
which include indolyne Diels-Alder reactions of substituted
furans and cyclopentadiene, the total synthesis of trikentrin A
and herbindole A using an indolyne Diels-Alder reaction, and
access to N- and C3-substituted indolynes from silyltriflate
precursors for use in the furan Diels-Alder reaction.
In this article,13 we report efficient syntheses of 4,5-, 5,6-,
and 6,7-indolyne precursors beginning from commercially
available hydroxyindole derivatives. The synthetic routes are
versatile and for the first time allow access to indolynes that
remain unsubstituted at N1, C2, and C3 of the pyrrole ring.
Indolynes can be generated under mild fluoride-mediated
conditions, trapped by a variety of reagents, and used to access
a number of novel substituted indoles. Nucleophilic addition
reactions to indolynes proceed with varying degrees of regi-
oselectivity. A computational and experimental study reveals
that distortion energies control regioselectivity and provides a
simple model for predicting the regioselectivity in the nucleo-
philic addition to indolynes and unsymmetrical arynes. This
Figure 3. Previous indolyne studies.
moindole (19) and KNH2 in ammonia to afford a complex
mixture of products, which upon purification furnished 4- and
5-aminoindole products 21 and 22 (Figure 3). The field of
indolyne chemistry lay dormant for over four decades,11 until
our groups and Buszek’s laboratory entered the area. In 2007,
(8) For reviews, see: (a) Pellissier, H.; Santelli, M. Tetrahedron 2003,
59, 701–730. (b) Wenk, H. H.; Winkler, M.; Sander, W. Angew. Chem.,
Int. Ed. 2003, 42, 502–528. (c) Sanz, R. Org. Prep. Proc. Int. 2008,
40, 217–291.
(9) For reviews regarding heteroaromatic arynes, see: (a) Reinecke, M. G.
Tetrahedron 1982, 38, 427–498. (b) Kauffmann, T.; Wirthwein, R.
Angew. Chem., Int. Ed. 1971, 10, 20–33.
(12) (a) Buszek, K. R.; Luo, D.; Kondrashov, M.; Brown, N.; VanderVelde,
D. Org. Lett. 2007, 9, 4135–4137. (b) Brown, N.; Luo, D.; Vander-
Velde, D.; Yang, S.; Brassfield, A.; Buszek, K. R. Tetrahedron Lett.
2009, 50, 63–65. (c) Buszek, K. R.; Brown, N.; Luo, D. Org. Lett.
2009, 11, 201–204. (d) Brown, N.; Luo, D.; Decapo, J. A.; Buszek,
K. R. Tetrahedron Lett. 2009, 50, 7113–7115. (e) Garr, A. N.; Luo,
D.; Brown, N.; Cramer, C. J.; Buszek, K. R.; VanderVelde, D. Org.
Lett. 2010, 12, 96–99.
(10) (a) Julia, M.; Huang, Y.; Igolen, J. C. R. Acad. Sci., Ser. C 1967, 265,
110–112. (b) Igolen, J.; Kolb, A. C. R. Acad. Sci., Ser. C 1969, 269,
54–56, For related studies, see: (c) Julia, M.; Goffic, F. L.; Igolen, J.;
Baillarge, M. C. R. Acad. Sci., Ser. C 1967, 264, 118–120. (d) Julia,
M.; Igolen, J.; Kolb, M. C. R. Acad. Sci., Ser. C 1971, 273, 1776–
1777.
(11) Attempts to synthesize 2,3-indolynes were put forth, but were not
successful; see: (a) Muller, H. Dissertation, University of Heidelberg,
1964. (b) Hoffman, R. W. Dehydrobenzene and Cycloalkynes;
Academic Press: New York, 1967. (c) Conway, S. C.; Gribble, G. W.
Heterocycles 1992, 34, 2095–2108.
(13) For preliminary communications involving this work, see: (a) Bronner,
S. M.; Bahnck, K. B.; Garg, N. K. Org. Lett. 2009, 11, 1007–1010.
(b) Cheong, P. H.-Y.; Paton, R. S.; Bronner, S. M.; Im, G.-Y.; Garg,
N. K.; Houk, K. N. J. Am. Chem. Soc. 2010, 132, 1267–1269.
9
17934 J. AM. CHEM. SOC. VOL. 132, NO. 50, 2010

Indolyne Experimental and Computational Studies
A R T I C L E S
Scheme 1
droxyindole 28 with isopropyl isocyanate, in the presence of
catalytic Et3N, afforded carbamate 34. Of note, conversion of
33 to carbamate 34 proceeds in 85% yield and requires only
one final chromatographic purification. Following the protocol
disclosed by Snieckus and Hoppe,19 carbamate 34 underwent
o-lithiation/silylation to provide silyl carbamate 35.20 Testament
to the outstanding ortho-directing ability of carbamates,21 the
relatively acidic C2 proton of the N-methylindole is not disturbed
in this process. Although initial attempts to elaborate silyl
carbamate 35 to silyltriflate 27 in a stepwise fashion were met
with difficulty,22 an efficient one-pot deprotection/triflation
sequence proved successful. Our optimized route to silyltriflate
27 can be carried out on multigram scale and proceeds in 63%
overall yield. To confirm that silyltriflate 27 would function as
a suitable precursor to the targeted 4,5-indolyne 29, silyltriflate
27 was reacted with TBAF in the presence of furan (24) to afford
Diels-Alder product 36 in 85% yield.
Scheme 2
A number of heteroatom- and carbon-based nucleophiles
undergo smooth reaction with indolyne 29, which indeed
functions as an electrophilic indole surrogate (Table 1). Treat-
ment of silyltriflate 27 with p-cresol in the presence of CsF led
to the formation of products 37a and 37b (entry 1). Similarly,
the use of aniline as the nucleophilic trapping agent provided
adducts 38a and 38b (entry 2),23 whereas trapping with
p-methylthiophenol afforded sulfur-containing products 39a and
39b (entry 3). With respect to carbon-based nucleophiles, the
use of a cyclic ꢀ-enaminoketone generated indole products 40a
and 40b (entry 4).24 Furthermore, trapping with potassium
cyanide afforded cyanoindoles 41a and 41b (entry 5). The latter
result is notable because cyanide has rarely been used in
nucleophilic addition reactions to arynes.25
A variety of formal cycloaddition processes involving in-
dolyne 29 were used to access several unique 4,5-disubstituted
indole derivatives (Table 2). For instance, reaction of benzy-
lazide and silyltriflate 27, in the presence of TBAF, provided
access to indolyltriazoles 42a and 42b in 86% yield via a formal
aryne cycloaddition (entry 1).26 Moreover, a formal [2 + 2]
cycloaddition provided indolylcyclobutanones 43a and 43b
(entry 2), whereas a variant involving cycloaddition followed
by fragmentation provided ketoesters 44a and 44b (entry 3).27
Of note, each of the examples shown in Tables 1 and 2 reflect
an interesting general preference for initial nucleophilic attack
at C5 of the presumed 4,5-indolyne intermediate 29 with
selectivity as high as 12.5:1 (Table 1, entry 2). An explanation
for this observation is presented in the latter part of this
manuscript.
model has led to the design of a substituted 4,5-indolyne that
exhibits enhanced nucleophilic regioselectivity.
Results and Discussion
Synthesis of
a 4,5-Indolyne Precursor and Synthetic
Applications. To initiate studies, 4,5-indolyne 29 was selected
as our primary target (Scheme 1). If accessible under mild
reaction conditions, indolyne 29 could undergo nucleophilic
additions and pericyclic reactions to provide a range of synthetic
applications. We elected to utilize Kobayashi’s approach to
aryne generation14 and hypothesized that indolyne 29 could be
obtained from indolyl silyltriflate precursor 27. At the time of
our studies, routes to indolyl silyltriflates were not known15 but
could plausibly begin from commercially available hydroxyin-
dole derivatives (e.g., 28).
The synthesis of an appropriate indolyl silyltriflate proved
challenging, primarily because of the acidity of the C2 hydrogen
and the electron-rich nature of the indole ring.16 Nonetheless,
extensive experimentation led to the development of an efficient
route to silyltriflate 27 (Scheme 2). Commercially available
5-benzyloxyindole (33) was converted to hydroxyindole 28
following a known two-step sequence.17,18 Reaction of hy-
(18) 5-Hydroxyindoles can also be readily prepared by the classic Nen-
itzescu indole synthesis; for a review, see: Allen, G. R. Org. React.
1973, 20, 337–454.
(19) (a) Kauch, M.; Snieckus, V.; Hoppe, D. J. Org. Chem. 2005, 70, 7149–
7158. (b) Kauch, M.; Hoppe, D. Synthesis 2006, 1578–1589.
(20) For the selective C4 lithiation of a related N-silylated substrate, see:
Griffen, E. J.; Roe, D. G.; Snieckus, V. J. Org. Chem. 1995, 60, 1484–
1485.
(14) Himeshima, Y.; Sonoda, Y.; Kobayashi, H. Chem. Lett. 1983, 1211–
1214.
(21) Snieckus, V. Chem. ReV. 1990, 90, 879–933.
(22) Cleavage of the carbamate of 35 was routinely accompanied by loss
of the C4 silyl substituent under a variety of reaction conditions.
(23) Liu, Z.; Larock, R. C. J. Org. Chem. 2006, 71, 3198–3209.
(24) Ramtohul, Y. K.; Chartrand, A. Org. Lett. 2007, 9, 1029–1032.
(25) Scardiglia, F.; Roberts, J. D. Tetrahedron 1958, 3, 197–208.
(26) (a) Shi, F.; Waldo, J. P.; Chen, Y.; Larock, R. C. Org. Lett. 2008, 10,
2409–2412. (b) Campbell-Verduyn, L.; Elsinga, P. H.; Mirfeizi, L.;
Dierckx, R. A.; Feringa, B. L. Org. Biomol. Chem. 2008, 6, 3461–
3463.
(15) Subsequently, Buszek reported routes to N-Me-C3-phenyl-substi-
tuted indolyl silyltriflates using a Fischer indolization strategy; see
ref 12b.
(16) The electron-rich indole was prone to undesired reactions at C3 and
protodesilylation at C4.
(17) (a) Stadlwieser, J. F.; Dambaur, M. E. HelV. Chim. Acta 2006, 89,
936–946. (b) Gwaltney, S. L.; Imade, H. M.; Barr, K. J.; Li, Q.;
Gehrke, L.; Credo, R. B.; Warner, R. B.; Lee, J. Y.; Kovar, P.; Wang,
J.; Nukkala, M. A.; Zielinski, N. A.; Frost, D.; Ng, S. C.; Sham, H. L.
Bioorg. Med. Chem. Lett. 2001, 11, 871–874.
(27) Tambar, U. K.; Stoltz, B. M. J. Am. Chem. Soc. 2005, 127, 5340–
5341.
9
J. AM. CHEM. SOC. VOL. 132, NO. 50, 2010 17935

A R T I C L E S
Im et al.
Table 1. Nucleophilic Additions to 4,5-Indolyne 29
Table 3. Diels-Alder Reactions of 4,5-Indolyne 29
a Conditions: 27, TBAF, CH3CN, 23 °C. b Conditions: 27, CsF,
CH3CN, 100 °C. c Conditions: 27, CsF, CH3CN, 80 °C.
Scheme 3
a Conditions: 27, CsF, CH3CN, 50 °C. b Conditions: 27, CsF,
CH3CN, 40 °C.
Table 2. Formal Cycloaddition Reactions of 4,5-Indolyne 29
of CO2. Finally, trapping of indolyne 29 with anthracene
furnished indolyltriptycene 48 in 72% yield (entry 4).
Stability and Functionalization of Indolyne Precursors. Our
studies (vide supra) show that indolyne methodology can be
used to access a range of unique indole derivatives. It should
be noted that C3 in all indolyne adducts remains unfunction-
alized and could be easily substituted if so desired. Alternatively,
we hypothesized that silyltriflate 27 could be further substituted
before indolyne generation and trapping. As shown in Scheme
3, silyltriflate 27 is stable to a range of reaction conditions and
can be readily functionalized at C3 through bromination (27 f
49),28 formylation (27 f 50),29 or Lewis-acid-mediated acety-
lation (27 f 51).7d Furthermore, in each case, the silyltriflate
a Conditions: 27, TBAF, CH3CN, 23 °C. b Conditions: 27, CsF,
CH3CN, 23 °C. c Conditions: 27, CsF, CH3CN, 80 °C.
Diels-Alder reactions of indolyne 29 were also investigated
as a means to access additional 4,5-disubstituted indole products
from silyltriflate 27 (Table 3). In addition to furan (see Scheme
2), N-Boc-pyrrole and cyclopentadiene could also be used to
trap indolyne 29, thus affording adducts 45 and 46, respectively
(entries 1 and 2). Furthermore, reaction in the presence of
R-pyrone produced benzoindole 47 (entry 3), presumably via a
Diels-Alder/retro-Diels-Alder process with concomitant loss
(28) Garg, N. K.; Sarpong, R.; Stoltz, B. M. J. Am. Chem. Soc. 2002, 124,
13179–13184.
(29) Vilsmeier, A.; Haack, A. Ber. 1927, 60, 119–122.
9
17936 J. AM. CHEM. SOC. VOL. 132, NO. 50, 2010

Indolyne Experimental and Computational Studies
A R T I C L E S
Scheme 4
Scheme 6
Scheme 7
Scheme 5
remains intact, allowing for subsequent indolyne generation and
trapping. By employing furan (24) as a trapping agent, indolyne
cycloadducts 52, 53, and 54 could be obtained in good yield.
The reaction sequence highlighted in Scheme 4 demonstrates
the viability of carrying indolyl silyltriflates through multistep
synthesis. 3-Bromo silyltriflate 49 was treated with t-BuLi at
-78 °C to facilitate halogen-metal exchange. Subsequent
quenching with borolane 55 delivered boronic ester 56 in 72%
yield.28 Next, boronic ester 56 underwent Pd-catalyzed
Suzuki-Miyaura coupling with iodobenzene at 100 °C to
furnish 3-phenyl silyltriflate 57. It is noteworthy that the
silyltriflate functionality withstands both harsh base (i.e., t-BuLi)
and Pd(0). Treatment of silyltriflate 57 with CsF in the presence
of furan (24) provided adduct 26 in 70% yield, thus demonstrat-
ing that silyltriflate 57 serves as an efficient precursor to indolyne
25.
Generation and Trapping of 5,6- and 6,7-Indolynes. Our
strategy for preparing indolyl silyltriflates from hydroxyindole
precursors proved amenable to the synthesis of 5,6- and 6,7-
indolyne precursors. The synthesis of 5,6-indolyne precursor
59 is depicted in Scheme 5. As described earlier, o-lithiation/
silylation of carbamate 34 afforded C4-silylated indole 35 in
84% yield, which could be elaborated to 4,5-indolyne precursor
27 (see Scheme 2). However, the C6-silylated indole 58 also
formed in this process in 10% yield. Carbamate cleavage,
followed by stirring with PhNTf2, provided silyltriflate 59 in
70% yield. The overall route to 59, although low yielding,
enabled our early studies of 5,6-indolynes. Higher yielding
routes to 5,6-indolyne precursors were later developed (vide
infra).
A suitable silyltriflate precursor to a 6,7-indolyne proved most
difficult to access. Exploratory efforts beginning from 7-hy-
droxyindole derivatives failed. For example, 7-benzyloxyindole
60 was elaborated to carbamate 61 using our standard two-step
sequence (Scheme 6). Attempts to effect the o-lithiation/
silylation of carbamate 61 led to competitive C2 lithiation and
decomposition products rather than to the desired silylcarbamate
62. Thus, silyltriflate 63 was inaccessible.
Unsuccessful efforts to synthesize silyltriflate 63 (see Scheme
6) led us to pursue the synthesis of regioisomeric silyltriflate
68 (Scheme 7). Readily available 6-benzyloxyindole 64 was
smoothly converted to carbamate 65 in 77% yield. Subsequent
treatment of carbamate 65 with TBSOTf and TMEDA, followed
by n-BuLi, led primarily to C7 lithiation.30 Quenching with
(30) Quenching of the presumed carbanion with MeOD led to complete
deuterium incorporation at C7.
(31) The intermediate silyl alcohol readily undergoes protodesilylation by
DBU-H+.
(32) Removal of the N-TIPS of 76 occurs more readily than indolyne
formation.
Figure 4. Generation and trapping of 5,6- and 6,7-indolynes 69 and 71.
9
J. AM. CHEM. SOC. VOL. 132, NO. 50, 2010 17937

A R T I C L E S
Im et al.
Scheme 8
Scheme 9
TMSCl delivered silylcarbamate 66. Although our previously
described protocol for elaborating aryl carbamates to aryl triflates
led primarily to hydroxyindole 67,31 a modified procedure using
n-BuLi in place of DBU furnished silyltriflate 68 in 91% yield.
To confirm that silyltriflates 59 and 68 serve as efficient
indolyne precursors, each compound was independently sub-
jected to furan (24) in the presence of fluoride sources (Figure
4). Trapping of 5,6-indolyne 69 delivered adduct 70 in 92%
yield. Similarly, 6,7-indolyne 71 underwent [4 + 2] cycload-
dition to provide adduct 72 in 95% yield. Nucleophilic additions
and formal cycloadditions of indolynes 69 and 71 were also
carried out and, as will be described below, found to proceed
efficiently and with interesting regioselectivities.
Indolyne Precursors with Alternative N-Substituents. Having
discovered that indolynes can be generated under mild fluoride-
mediated conditions and serve as building blocks for the
synthesis of polysubstituted indoles, we synthesized indolyne
precursors with alternative N-substituents. Indolyl silyltriflates
with removable groups, such as Boc and TIPS, were pursued,
in addition to N-unsubstituted silyltriflates. Syntheses of the
targeted indolyne precursors in the 4,5-indolyne series are shown
in Scheme 8 and parallel our synthesis of 4,5-indolyne precursor
27 (see Scheme 2). Beginning with TIPS protection of 5-ben-
zyloxyindole (33), N-TIPS-silyltriflate 76 was prepared in 85%
yield over five steps. N-TIPS-silyltriflate 76 served as the
precursor to the corresponding N-H and N-Boc compounds 77
and 78, respectively. Of note, when treated with CsF in the
presence of furan (24), both substrates 76 and 77 are converted
to the identical unprotected bicycle 80, likely via N-H indolyne
intermediate 79.32 Free N-H indolynes of the type 79 have never
been synthesized previously but provide a valuable synthetic
tool for the synthesis of unprotected, benzenoid-substituted
indoles.
the 5-substituted carbamate was observed to favor formation
of the 4,5-substituted species 84 over the desired 5,6-substituted
product 85, we explored the use of 6-carbamylated substrates.
Although C7 lithiation of 86 was favored when R ) Me, it
was hypothesized that the use of a bulky TIPS protective group
would disfavor C7 lithiation, ultimately allowing access to
5-silylated product 88.
The execution of this approach to 5,6-indolynes is shown in
Scheme 10. Commercially available 6-benzyloxyindole (89) was
TIPS protected under standard conditions. Conversion of 90 to
carbamate 91 proceeded in 82% yield using our robust two-
step sequence. In the key o-lithiation/silylation experiment,
functionalization of 91 occurred exclusively at C5 to deliver
silylcarbamate 92. Analogous to previous studies, carbamate
92 was elaborated to N-TIPS silyltriflate 93, which in turn was
readily converted to N-H and N-Boc silyltriflates 94 and 95,
respectively.
As described earlier, our synthesis of an N-methyl-5,6-
indolyne precursor proceeded in low yield (see Scheme 5). To
address this limitation and access 5,6-indolyne precursors with
alternative N-substitutents, we considered the strategies depicted
in Scheme 9. Indolyne 5 could be derived from either regioi-
someric silyltriflates 81 or 82, which in turn would be derived
from 5- or 6-substituted precursors 83 or 86, respectively. Since
The synthesis of 6,7-indolyne precursors with variable
N-substituents proved challenging. Nevertheless, two routes
were developed beginning from 6-hydroxyindole derivatives.
The first-generation approach is shown in Scheme 11. Boc
protection of benzyloxyindole 89 followed by treatment with
H2 and Pd/C furnished a hydroxyindoline intermediate. Subse-
9
17938 J. AM. CHEM. SOC. VOL. 132, NO. 50, 2010

Indolyne Experimental and Computational Studies
A R T I C L E S
Scheme 10
Scheme 12
calculations were performed using Gaussian.35 The arynes and
their transition structures for nucleophilic addition were opti-
mized with B3LYP36 and M06-2X37 hybrid density functionals
employing 6-31G(d) or 6-311+G(d,p) basis sets. In addition,
optimizations were performed with the B2PLYP double hybrid
functional38 and with MP2 perturbation theory39 using a
6-31G(d) basis set. All density functional calculations used a
fine grid for numerical integration. Solvation effects were
evaluated through single-point calculations using an implicit
CPCM model of acetonitrile,40 defining the solute surface by
UAKS radii.41 In the manuscript we discuss the results of the
B3LYP/6-31G(d) calculations, although optimizations at B3LYP/
6-311+G(d,p) and M06-2X/6-311+G(d,p) levels of theory (see
Supporting Information) give very similar aryne geometries and
also predict the same sense of regioselectivity in each case
studied.
Scheme 11
The results of several nucleophilic additions to 4,5-indolyne
29 are shown in Table 4, in addition to the results of density
functional computations performed at the B3LYP/6-31G(d) level
of theory. The experimental preference for attack at C5 was
matched well by the activation free energies of the competing
transition structures for each of the nucleophiles we studied,
while the magnitudes are usually exaggerated. Activation
barriers are very low, in some cases zero, so that variational
effects and dynamics may cause selectivities to be less than
predicted by ∆∆G‡. Cyanide addition has no energetic barrier,
but approach of the nucleophile at C5 is favored over C4 for
all distances.42 The major indolotriazole adduct formed from
azide cycloaddition involves attack at C5 by the nucleophilic
internal nitrogen. Ion pairing and entropic effects should cause
these reactions to have positive barriers.
quent reaction with i-PrNCO in the presence of Et3N afforded
carbamate 96. o-Lithiation/silylation of carbamate 96 delivered
6,7-disubstituted indoline 97, albeit with competitive function-
alization at C5.33 Nonetheless, DDQ oxidation furnished indole
98, which was readily converted to silyltriflate 99. The N-Boc
group of 99 was removed by thermolysis to provide N-H
indolyne precursor 100.
A higher yielding route to N-H indolyne 100 is depicted in
Scheme 12. O-Silylation and N-allylation of 6-hydroxyindole
(101) afforded protected indole 102. Subsequent elaboration of
TBS ether 102 to carbamate 103 proceeded without event.
Lithiation of carbamate 103, followed by quenching with
TMSOTf, furnished silylcarbamate 104 in 78% yield. After
installing the silyltriflate (104 f 105), the N-allyl group was
removed34 to deliver N-H silyltriflate 100.
Figures 5 and 6, respectively, show the calculated structures
of 4,5-indolyne 29 and transition state (TS) geometries for
Regioselectivity in Nucleophilic Additions to Indolynes. Our
studies demonstrate that nucleophilic additions to 4,5-indolynes
can occur with significant regioselectivity. A combination of
experimental and computational studies was undertaken to
determine the origin of these observations. Quantum chemical
(35) Frisch, M. J. et al. Gaussian 03, revision C.02; Gaussian, Inc.:
Wallingford, CT, 2004.
(36) (a) Becke, A. D. J. Chem. Phys. 1993, 98, 5648–5652. (b) Lee, C.;
Yang, W.; Parr, R. G. Phys. ReV. B 1988, 37, 785–789. (c) Vosko,
S. H.; Wilk, L.; Nusair, M. Can. J. Phys. 1980, 58, 1200–1211. (d)
Stephens, P. J.; Devlin, F. J.; Chabalowski, C. F.; Frisch, M. J. J.
Phys. Chem. 1994, 98, 11623–11627.
(33) Attempts to lithiate the corresponding indole led to competitive C2
functionalization.
(37) Zhao, Y.; Truhlar, D. G. Theor. Chem. Acc. 2008, 120, 215–241.
(38) Grimme, S. J. Chem. Phys. 2006, 124, 034108-1-034108-16.
(39) Head-Gordon, M.; Pople, J. A.; Frisch, M. J. Chem. Phys. Lett. 1988,
153, 503–506.
(34) For ruthenium-mediated olefin isomerization, see: (a) Kamijo, S.; Huo,
Z.; Jin, T.; Kanazawa, C.; Yamamoto, Y. J. Org. Chem. 2005, 70,
6389–6397. (b) For oxidative cleavage, see: Alcaide, B.; Almendros,
P.; Alonso, J. M. Tetrahedron Lett. 2003, 44, 8693–8695. (c) Alcaide,
B.; Almendros, P.; Alonso, J. M. Chem.sEur. J. 2006, 12, 2874–
2879. (d) For Et3N-promoted N-deformylation, see: Aizpurua, J. M.;
Coss´ıo, F. P.; Lecea, B.; Palomo, C. Tetrahedron Lett. 1986, 27, 4359–
4362.
(40) (a) Barone, V.; Cossi, M. J. Phys. Chem. A 1998, 102, 1995–2001.
(b) Cossi, M.; Rega, N.; Scalmani, G.; Barone, V. J. Comput. Chem.
2003, 24, 669–681.
(41) Takano, Y.; Houk, K. N. J. Chem. Theory Comput. 2005, 1, 70–77.
(42) See Supporting Information for details.
9
J. AM. CHEM. SOC. VOL. 132, NO. 50, 2010 17939

A R T I C L E S
Im et al.
Table 4. Nucleophilic Additions to 4,5-indolyne 29a
a Conditions: see Supporting Information. Computed ratios obtained
from Boltzmann factors using B3LYP/6-31G(d) free energies including
CPCM solvation by MeCN. b Attack at C5 is favored (see Supporting
Information).
Figure 5. B3LYP/6-31G(d)-optimized structure of 4,5-indolyne 29.
Figure 6. B3LYP/6-31G(d)-optimized TSs for addition of aniline, p-cresol,
and methyl azide to indolyne 29. Energies in kcal/mol.
nucleophilic addition of aniline, p-cresol, and methylazide (as
a model for benzylazide used in experiment). For each of the
nucleophiles, the bond angle of the terminus of the aryne that
undergoes attack is flattened in the transition structure, whereas
the angle at the other terminus is compressed. This picture can
be understood in terms of the polarization of the Ct C bond,
with developing positive charge at the site of attack and negative
charge at the adjacent carbon. We also found that the computed
tivities of additions to arynes,13b one of us previously explored
the LUMO-lowering effect of bending distortions on alkynes
and benzynes that leads to their greater reactivity toward
nucleophiles over electrophiles.48
Internal angles of 4,5-indolyne 29 show that the aryne is
distorted, particularly at C3a (θCCC ) 110° vs θHCC ) 126° in
pyrrole). Nucleophilic attack at C5 relieves some of the strain
at C3a (e.g., opening from 110° to 118° in TS-1), but attack at
C4 adds to the unfavorable distortion at C3a (e.g., closing from
110° to 108° in TS-2). Attack at C4 has 1.4 kcal/mol greater
distortion energy than at C5 due to compression at C3a, leading
to a lower activation barrier.
‡
distortion energies, ∆Ed , are useful in quantifying the transition
state stabilities. These distortion energies (also termed deforma-
tion energy43 or activation strain44) are defined as the energies
required to distort the indolyne and nucleophile into their TS
geometries without interaction between them. These are shown
in Figure 6. Computed distortion energies have been shown to
correlate well with activation barriers in a number of bimolecular
reactions: 1,3-dipolar cycloadditions,45 1,4-dihydrogenations,
Diels-Alder cycloadditions,46 and cycloadditions involving
fullerenes and nanotubes.47 While this is the first time that
distortion energies have been used to understand regioselec-
The types of distortions present in the indolyne TSs shown
in Figure 6 can be understood from the geometry for nucleo-
philic attack on benzyne (106). As shown in Figure 7, benzyne
is flattened at the point of attack of aniline in TS-7 (θCCC
)
136°), giving the orbital at the site of attack significant p
character and a slight positive charge, while the adjacent angle
is compressed (θCCC ) 113°), increasing s character to stabilize
the developing carbanion (see TS-8). We reasoned that nucleo-
(43) Morokuma, K. J. Chem. Phys. 1971, 55, 1236–1244.
(44) Bickelhaupt, F. M. J. Comput. Chem. 1999, 20, 114–128.
(45) (a) Ess, D. H.; Houk, K. N. J. Am. Chem. Soc. 2007, 129, 10646–
10647. (b) Ess, D. H.; Houk, K. N. J. Am. Chem. Soc. 2008, 130,
10187–10198.
(48) (a) Strozier, R. W.; Caramella, P.; Houk, K. N. J. Am. Chem. Soc.
1979, 101, 1340–1342. (b) Rondan, N. G.; Domelsmith, L. N.; Houk,
K. N.; Bowne, A. T.; Levin, R. H. Tetrahedron Lett. 1979, 35, 3237–
3240.
(46) Hayden, A. E.; Houk, K. N. J. Am. Chem. Soc. 2009, 131, 4084–
4089.
(47) Osuna, S.; Houk, K. N. Chem.sEur. J. 2009, 15, 13219–13231.
9
17940 J. AM. CHEM. SOC. VOL. 132, NO. 50, 2010

Indolyne Experimental and Computational Studies
A R T I C L E S
Figure 7. B3LYP/6-31G(d)-optimized benzyne (106) and TS for aniline
addition. Energies in kcal/mol.
Table 5. Nucleophilic Additions to N-Me-5,6-indolynea
a Conditions: see Supporting Information. b Attack at C5 is favored.
philic addition to unsymmetrical arynes (e.g., indolynes) is
favored for attack at the carbon that requires minimal geo-
metrical and energetic change from aryne to TS structure. This
is the carbon at which the internal angle is larger as it is in the
nucleophilic addition transition state. In the structure of 4,5-
indolyne (see Figure 5), C5 is flatter than C4 (129° vs 124°),
and accordingly, nucleophilic addition is preferred at C5.
Regioselectivities of nucleophilic attack on 5,6- and 6,7-
indolynes were studied computationally prior to experiment.
There is uniformly lower unsymmetrical distortion and lower
addition selectivity in nucleophilic attack on 5,6-indolyne seen
in the results of our calculations and also in experiment (Table
5). Figure 8 shows the computed structures of 5,6-indolyne 69
and the TSs for nucleophilic addition. The slightly greater
flattening of 69 at C5 than at C6 (internal angles are 130° and
127°, respectively) can be attributed to electron withdrawal by
the indole nitrogen (the charges at C5 and C6 are -0.01 and
-0.03e, respectively). This facilitates attack at C5 in TS-10,
TS-12, and TS-14, albeit to only a small extent, paralleling the
reactivity depicted in TS-9.
In contrast to the 5,6-indolyne, the computed internal angles
at the Ct C termini of 6,7-indolyne 71 are very different from
each other with θCCC ) 135° and 117° (Figure 9). Since the
flatter of the two termini is more susceptible to nucleophilic
attack, we expected that the 6,7-indolyne would display a high
regiochemical preference for addition to C6. Nucleophilic attack
on 6,7-indolyne at C6 is highly favored because of the smaller
distortion energy, as the indolyne is already distorted into the
TS geometry.49 In fact, the structure of 6,7-indolyne is practi-
cally identical to that of the TS for aniline addition to benzyne
(see Figure 7, TS-7). There is no calculated energetic barrier
for the attack at C6 by aniline, while C7 attack (TS-12) has a
Figure 8. B3LYP/6-31G(d)-optimized 5,6-indolyne 69 and TSs for aniline
addition. Energies in kcal/mol.
∆Ed of 8.8 kcal/mol, which is higher than that for any other of
the TSs computed for the indolyne isomers. These results are
in accord with the greater selectivity for C6 attack shown here.
Extremely high levels of regioselectivity are observed experi-
mentally (Table 6), and even azide cycloaddition gives very
high selectivity. The 6,7-indolynes also display remarkably high
regioselectivity in cycloadditions with 2-substituted furans,
unlike 4,5- and 5,6-indolynes.12e Complementary to our ratio-
nalization based on the distortion energies required to achieve
the transition state geometry, Buszek and Cramer have also
shown that regioselectivity of additions of electron-rich furans
to indolynes are influenced by the computed polarization of the
Ct C bond.12e This polarization is related to the unsymmetrical
bending distortions of the indolyne.
Model for Predicting Regioselectivity in the Nucleophilic
Addition to Indolynes and Unsymmetrical Arynes. These studies
show that computed atomic charges at the triple bond termini
are related to the asymmetric distortion of indolynes. On this
basis, a distortion model for regioselective nucleophilic addition
to fused arynes is depicted in Figure 10. Regioselectivity is
controlled by the relative ease of distorting the aryne into the
two possible transition structures shown. Unsymmetrical distor-
tion is already present in ring-fused arynes, and electron-
withdrawing groups (X) can further enhance this distortion by
(49) The nucleophilic attack of p-cresol on indolynes is accompanied by
proton transfer, and the transition states appear to be 4-centered, since
the proton transfer will occur without a barrier upon nucleophilic attack
of the oxygen.
9
J. AM. CHEM. SOC. VOL. 132, NO. 50, 2010 17941

A R T I C L E S
Im et al.
Figure 10. Distortion model for regioselective nucleophilic addition to
fused arynes (X ) electron-withdrawing group). The effect of X is larger
than the more remote X′.
Figure 9. B3LYP/6-31G(d)-optimized 6,7-indolyne 71 and TSs for aniline,
p-cresol, and methylazide addition. Energies in kcal/mol.
Figure 11. Internal angles for unsymmetrical arynes obtained from B3LYP/
6-31G(d)-optimized geometries.
Table 6. Nucleophilic Additions to 6,7-Indolyne 71a
nucleophilic attack. The difference in internal angles also relates
to the degree of regioselectivity, with higher selectivity observed
in the addition to arynes that possess large differences between
internal angles. If angle differences are small (4° or less),
regioselectivity may be modest or somewhat variable. This
predictive model functions at various different levels of theory,
where the angles are essentially the same as in these cases.
Influence of N-Substituents on the Regioselectivity of
Indolyne Reactions. With a predictive model for determining
the regioselectivity in nucleophilic additions to indolynes, we
examined the role of substituents on the presumed indolyne
intermediates. Figure 12 shows the results of B3LYP/6-31G(d)
geometry optimization on 4,5-, 5,6-, and 6,7-indolynes as a
function of the nitrogen substituent. Computed geometries for
the N-H indolynes and N-Me indolynes are very similar,
indicating that their regioselectivities would also be very similar.
Effects of Boc protection on the indolyne geometry were more
noticeable: N-Boc-4,5-indolyne is slightly more distorted than
the N-H and N-Me structures, while 5,6- and 6,7-indolynes were
a Conditions: see Supporting Information. b C6 attack has no
energetic barrier; C7 attack has ∆∆H‡ ) 5.5. c Attack at C6 is favored
(see Supporting Information).
polarizing the aryne. The unsymmetrical distortion biases
nucleophilic attack to the flatter, more electropositive end of
the aryne, leading to an energetically favorable decrease in ring
distortion.
Furthermore, we advocate a comparison of computed internal
angles in indolynes as a rapid and reliable method to predict
nucleophilic regioselectivity. As exemplified in Figure 11, this
predictive tool holds for other unsymmetrical arynes with fused
rings,50,51 in addition to arynes with adjacent polar substituents.52
In all cases, the terminus of the aryne that is flatter, or more
distorted, correlates to the experimentally favored site of
(50) For experimental studies of naphthalyne, see: (a) Kauffmann, K.;
Fischer, H.; Nu¨rnberg, R.; Withwein, R. Justus Liebigs Ann. Chem.
1970, 731, 23–26. (b) Kolomeitsev, A. A.; Vorobyev, M.; Gillandt,
H. Tetrahedron Lett. 2008, 49, 449–454.
(51) For studies pertaining to benzynocycloalkanes, see: Hamura, T.;
Ibusuki, Y.; Sato, K.; Matsumoto, T.; Osamura, Y.; Suzuki, K. Org.
Lett. 2003, 5, 3551–3554.
(52) For an experimental study involving 3-methoxybenzyne, see: Yoshida,
H.; Fukushima, H.; Ohshita, J.; Kunai, A. J. Am. Chem. Soc. 2006,
128, 11040–11041, and ref 8.
9
17942 J. AM. CHEM. SOC. VOL. 132, NO. 50, 2010

Indolyne Experimental and Computational Studies
A R T I C L E S
Table 8. Nucleophilic Additions to 5,6-Indolynes with Varying
N-Substituentsa
Figure 12. Internal angles for 4,5-, 5,6-, and 6,7-indolynes with varying
N-substitutents from B3LYP/6-31G(d)-optimized geometries.
Table 7. Nucleophilic Additions to 4,5-Indolynes with Varying
N-Substituentsa
a Conditions: see Supporting Information. b Yield reported after
Boc-protection.
Table 9. Nucleophilic Additions to 6,7-Indolynes with Varying
N-Substituentsa
a Conditions: see Supporting Information. b Yield reported after
Boc-protection.
slightly less distorted. Although some minor perturbations are
observed, these computational results suggest that regioselec-
tivity in nucleophilic additions to each indolyne would not vary
significantly by changing the N-substituent from methyl to H
or Boc.
As described earlier, our robust route to indolyl silyltriflates
provided access to a variety of indolyne precursors with differing
N-substituents. By employing these indolyne precursors in
nucleophilic additions and formal cycloaddition reactions, we
verified our computational predictions. Experimental results for
the 4,5-, 5,6-, and 6,7-indolynes are shown in Tables 7, 8, and
9, respectively. As expected, only minor variations in regiose-
lectivity were observed in all cases. It should be noted that in
all reactions involving the N-H-6,7-indolyne (Table 9), products
indicative of initial attack at C7 were also observed. Although
it is not possible to significantly improve regioselectivities based
on the N-substituent, these studies validate our predictive
regioselectivity model and demonstrate that an array of indolyl
silyltriflates can be used as building blocks for the synthesis of
substituted indoles.
a Conditions: see Supporting Information.
tial53 was used) for 3-bromo-4,5-indolyne 107 are shown in
Figure 13. The internal angles at C5 and C4 were found to be
131° and 124°, respectively. This finding suggests that bro-
moindolyne 107 should display a greater preference for nu-
cleophilic attack at C5 compared to its C3-unsubstituted
derivative (for comparison, the desbromo derivative of 107
possesses internal angles of 129° and 125°; see Figure 5).
Our computational prediction was tested experimentally by
the reaction of 3-bromosilyltriflate 49 (see Scheme 3 for
preparation) with benzylazide in the presence of CsF (Scheme
13). Adduct 108a was the major product obtained, showing that
initial attack of the presumed indolyne intermediate 107 is
Influence of
a
C3 Halide Substituent on the
Regioselectivities of 4,5-Indolyne Reactions. As a means to
further probe our predictive regioselectivity model and ulti-
mately improve regioselectivities, the influence of C3 halide
substituents on 4,5-indolynes was evaluated. The results of
geometry optimization (B3LYP/6-31G(d) for all atoms except
Br, for which the Hay-Wadt LANL2DZ effective core poten-
(53) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 299–310.
(54) Trapping of indolyne 107 with cyanide led to a 6.5:1 mixture of
products, favoring attack at C5 over C4. This result reflects a notable
improvement in regioselectivity compared to that seen in the reaction
of the parent desbromo compound with cyanide (3.3:1 selectivity, see
Table 1).
9
J. AM. CHEM. SOC. VOL. 132, NO. 50, 2010 17943

A R T I C L E S
Im et al.
the benzenoid ring. Our strategy relies on the generation of
indolynes, which act as electrophilic indole surrogates, thus
opposing the typical mode of indole reactivity. A series of 4,5-,
5,6-, and 6,7-indolynes are accessible from indolyl silyltriflate
precursors and can be readily employed in both nucleophilic
additions and formal cycloaddition processes. Varying degrees
of regioselectivity are observed, and we have shown that these
regioselectivities are controlled by distortion energies. These
studies have led to the discovery of a simple model to predict
the regioselectivity in the nucleophilic additions to indolynes
and other unsymmetrical arynes. Applications of indolyne
methodology in the synthesis of complex molecules are currently
being explored in our laboratories.
Figure 13. 3-Bromo-4,5-indolyne 107 and B3LYP/6-31G(d)-optimized
geometry.
Scheme 13
Acknowledgment. We are grateful to the National Science
Foundation (CHE-0548209 to K.N.H. and CHE-0955864 to N.K.G.)
and the National Institute of General Medical Sciences, National
Institutes of Health Grant (GM-36700 to K.N.H.), Dupont (N.K.G),
Boehringer Ingelheim (N.K.G.), the Foote Fellowship (S.M.B.), the
Fulbright Commission, AstraZeneca, and the Royal Commission
for 1851 (R.S.P) for financial support.
favored at C5. Of note, the degree of regioselectivity was higher
compared to that seen for the corresponding desbromo species
(6.8:1 versus 2.4:1), consistent with our calculations. Analogous
results were obtained in the nucleophilic addition of cyanide.54
The fact that a bromo substituent could lead to improvements
in regioselectivity is significant since the C3 bromide is easy
to introduce. Moreover, after the necessary indolyne reaction,
the bromide substituent could be either removed reductively or
used as a handle for the installation of other functional groups.
Supporting Information Available: Experimental details and
compound characterization data; comparison of optimizations
at different levels of theory, absolute energies, Cartesian
coordinates, and full authorship of ref 35. This material is
available free of charge via the Internet at http://pubs.acs.
org.
Conclusion
In summary, we developed an efficient method for accessing
a variety of indole derivatives that bear functional groups on
JA1086485
9
17944 J. AM. CHEM. SOC. VOL. 132, NO. 50, 2010