The role of a Lewis acid in the Nenitzescu indole synthesis

Article abstract of DOI:10.1016/j.tetlet.2008.09.087

A highly efficient Lewis acid-catalyzed method for the Nenitzescu synthesis of 5-hydroxyindoles with a range of substituents at N-1 and C-3 and symmetric 5,5′-dihydroxydiindoles has been developed. The amount of the catalyst (10-100 mol %) required depend

Full text of DOI:10.1016/j.tetlet.2008.09.087

Tetrahedron Letters 49 (2008) 7106–7109
Contents lists available at ScienceDirect
Tetrahedron Letters
journal homepage: www.elsevier.com/locate/tetlet
The role of a Lewis acid in the Nenitzescu indole synthesis
b
a
a
Valeriya S. Velezheva a, , Andrey I. Sokolov , Albert G. Kornienko , Konstantin A. Lyssenko ,
*
Yulia V. Nelyubina a, Ivan A. Godovikov a, Alexander S. Peregudov a, Andrey F. Mironov b
a A. N. Nesmeyanov Institute of Organoelement Compounds, Russian Academy of Sciences, 28 Vavilov Street, 119991 GSP-1 Moscow, Russia
b M. V. Lomonosov Moscow State Academy of Fine Chemical Technology, 86 Vernadskogo Street, 119571 Moscow, Russia
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 16 July 2008
Revised 4 September 2008
Accepted 15 September 2008
Available online 19 September 2008
A highly efficient Lewis acid-catalyzed method for the Nenitzescu synthesis of 5-hydroxyindoles with a
range of substituents at N-1 and C-3 and symmetric 5,50-dihydroxydiindoles has been developed. The
amount of the catalyst (10–100 mol %) required depended on the nature of the enaminone component.
It has been shown that Lewis acid plays a role in enaminone component activation through an enam-
ine-ZnC12 complex followed by its deprotonation.
Ó 2008 Elsevier Ltd. All rights reserved.
The Nenitzescu reaction has proven to be important in drug
discovery for 5-hydroxyindole-based derivatives bearing new
additional functions at various positions of the ring system.1-4
5-Hydroxyindoles with a high density of functional and pharmaco-
phoric groups are of interest as novel inhibitors for protein tar-
gets.5 However, 3-acyl-5-hydroxyindoles, especially those with
N-functionalized alkyl tails, and symmetric 5,50-dihydroxydiin-
doles bearing a linker between the indole nitrogens, are not easily
attainable due to formation of the corresponding 3-acylbenzo-
furans instead of the target indoles.6 The use of diketodienamines
derived from acetyl acetone and ethylene- or butylenediamines in
the Nenitzescu reaction in acetic acid has been reported to afford
3-acetyl-5-hydroxy-2-methyl-benzofuran/naphthofuran, instead
of symmetric 5,50-dihydroxydiindoles.7 Earlier, Grinev managed
to obtain 3-acetyl-5-hydroxy-1-(2-hydroxyethyl)-2-methylindole
in very low yield along with 3-acetyl-5-hydroxy-2-methylbenzofu-
ran.8 We reasoned that employing Lewis acids in the Nenitzescu
reaction with enaminones would direct the process to the indoliza-
tion pathway.9 The method is simple, rapid, efficient, and allows
the preparation of hydroxyindoles from p-benzoquinone (PBQ)
and simple enaminones in good to excellent yields with the use
of low polarity solvents in the presence of weak Lewis acid
catalysts (hereafter referred to as ‘standard conditions’). The time
required for the indolization decreases by a factor of three or more,
compared to the non-catalyzed process, while the formation of
3-acylbenzofurans is supressed. The formation of 5-hydroxy-
indoles under such mild conditions is explained in terms of a
non-redox mechanism. It was noted that the corresponding
5-hydroxybenzofurans were obtained by the condensation of PBQ
with acetyl acetone or ethyl acetoacetate in the presence of Lewis
acid catalysts.10
We chose a series of enaminones 1-5 with the structural unit
Z–N–C@C–COR (Z = (CH2)3CH3 1, (CH2)2NHTs, 2, (CH2)2OH 3,
(CH2)nNHC(CH3)@CHCOR 4 (n = 2), 5 (n = 6); R = Me, OEt, Ph)
(Schemes 1 and 2). Under standard conditions (10 mol % ZnCl2/
CH2Cl2 system and PBQ (1 equiv)), the method9 worked well only
with enaminone 2a of enaminones 1–3a to give pure indole 7a11
in 88% isolated yield (Table 1, entries 1, 5 and 7). Moreover, 2a
reacted with PBQ faster than did the enaminone 3a and other
simple enaminones.9 Similarly, enaminones 2b,c reacted smoothly
under the same conditions to give indoles 7b,c in high yields
(Table 2).
Poor yields of crude 7a and 8a were also obtained, when AlCl3
was employed as the catalyst (Table 1, entries 4 and 6). As
expected, enaminones 1–5 reacted poorly with PBQ in the absence
* Corresponding author. Tel.: +7 499 135 9333; fax: +7 499 135 5085.
E-mail address: vel@ineos.ac.ru (V. S. Velezheva).
Scheme 1.
0040-4039/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2008.09.087

V. S. Velezheva et al. / Tetrahedron Letters 49 (2008) 7106–7109
7107
Figure 1. The general view of 9a, atoms represented by thermal ellipsoids at 50%
probability level.
reaction between diketodien-amines 4–5 and PBQ proceeded via
the indolization pathway, however, the crude monoindoles 9a,c
and diindoles 10b, 11a–c were isolated in low yields under the
standard conditions. A 50–100% amount of the catalyst appeared
to be critical for the efficient synthesis of compounds 8–11, and
under these conditions, 8–11 were obtained in good to high iso-
lated yields (Tables 1 and 2). For both diketo- and diethoxycarbon-
yl diendiamines 5, the number of intervening methylene groups (6)
did not influence the result, since diindoles 11 were obtained as
single products (Table 2). Further studies on concentration effects
revealed that a three-fold excess of 4a appears to be optimum to
obtain indole 9a as a pure product in 78% isolated yield (Table
2). Finally, we showed that a stoichiometric amount of ZnCl2 was
required to obtain pure (>95%) N-butylindole 6a in good yield (Ta-
ble 1, entry 2). All reactions were performed in refluxing CH2Cl2
without exclusion of air or moisture. The results were evaluated
through TLC and 1H NMR data. The molecular structure of com-
pound 9a was confirmed by an X-ray analysis (Fig. 1).12 The struc-
tures of compounds 6–11 were deduced from elemental analyses
and 1H and 13C NMR data. The mass spectra of these compounds
displayed molecular ion peaks at appropriate m/z values.
Scheme 2.
Table 1
Yields of indoles 6–8a depending on the type and mol % of catalysta
Entry Enaminone Catalyst (mol %) Time (min) Product, isolated yield (%);
purity (%) by 1H NMRb
1
2
3
4
5
6
7
8
9
1a
1a
2a
2a
2a
3a
3a
3a
3a
3a
3a
ZnCl2 (10)
ZnCl2 (100)
No catalyst
AlCl3 (10)
ZnCl2 (10)
AlCl3 (10)
ZnCl2 (10)
ZnCl2 (20)
ZnCl2 (50)
ZnCl2 (100)
ZnI2 (100)
20
20
210
90
30
90
90
60
40
20
20
6a, 79; <95
6a, 60; >95
7a, 23; <95
7a, 29; <95
7a, 88; >95
8a, 22; <95
8a, 37; <95
8a, 71; <95
8a, 81; <95
8a, 56; >95
8a, 65; >95
10
11
a
1:1.1 PBQ:1–3a.
b
Based on 1H NMR analysis (600 MHz, DMSO-d6) of the crude product. Purity of
isolated 6–8a determined by integration of the signals corresponding to the main
compound and by-product.
As compared with the results under the standard conditions, all
the reactions with enaminones 3–5 in the presence of 10–
100 mol % of ZnCl2 proceeded smoothly and rapidly to give the
corresponding indoles in good to high yields as practically pure
products. Several examples illustrating this novel method are
of a catalyst (Table 1, entry 3). We observed a reduction of the reac-
tion time and a rise in the yield of 8a with an increase in the
amount of ZnCl2 relative to the standard conditions (Table 1, en-
tries 8–10), 20 mol % of ZnCl2 giving 71% of 8a after 60 min. If more
catalyst was employed (50 mol %), the yield rose to 81% after
40 min. The use of a stoichiometric amount of ZnCl2 gave the pur-
est product after 20 min but in 56% yield. A similar result was
achieved when ZnI2 was used as a catalyst (Table 1, entry 11).
The application of 50 mol % of ZnCl2 produced a positive effect
relative to catalytic amounts in the reaction of enaminones 3b,c
to afford indoles 8b,c in high yields (Table 2). As expected, the
Table 2
Yields of indoles 7–11 depending on reaction conditions
Entry
Enaminone
Equiv PBQ:
2–5b,c (mmol)
ZnCl2,
(mol %)
Time
(min)
Product
Yield
(%)
1
2
3
4
5
6
7
8
9
2b
2c
3b
3c
4a
4b
4c
5a
5b
5c
1:1.1
1:1.1
1:1.1
1:1.1
1:3
2.2:1
1:1.1
2.2:1
2.2:1
2.2:1
10
10
50
50
100
50
100
50
50
50
25
30
60
60
20
30
30
30
25
30
7b
7c
8b
8c
9a
10b
9c
11a
11b
11c
85
83
83
82
78
80
77
78
75
82
Figure 2. The general view of 12, atoms represented by thermal ellipsoids at 50%
probability level.
10

7108
V. S. Velezheva et al. / Tetrahedron Letters 49 (2008) 7106–7109
Scheme 3.
summarized in Tables 1 and 2. High yields of indoles 7 could be
achieved when a catalytic amount of ZnC12 was employed (Tables
1 and 2).
to the enaminone nitrogens; Zn(II) is known to have a high affinity
for amines.15 Furthermore, Lewis acid catalysts can promote the
reactions of both N-alkyl functionalized enaminones and simple
ones with PBQ and other quinones through type 12–14 catalyst—
enaminone complexes.
In summary, we have developed an efficient and versatile meth-
od to obtain 5-hydroxyindoles containing a range of substituents at
the N-1 and C-3 positions as well as 5,50-dihydroxydiindoles bear-
ing a linker between the ring nitrogens. The method requires the
use of 10–100 mol % of a Lewis acid. The amount of the catalyst
employed depends on the nature of the enamine component. We
showed that the catalytic effect occurs due to enamine component
activation through a diketodienamine-ZnC12 complex, in particu-
lar, followed by its deprotonation. The structure of a similar depro-
tonated ethylenediamine-based complex with Zn–N bonds was
reported recently.19 Further studies on the utility of this reaction
are in progress and will be reported in due course.
The yields of indoles 7a and 8a obtained in the presence of the
more azaphilic Lewis acids (ZnC12 or ZnI2) exceeded those
achieved in the presence of oxophilic AlCl3. We tried to test
whether the origin of the catalytic effect was due to enamine acti-
vation, in accordance with a hypothesis involving an enamine-
Michael reaction, that is, the nucleophilic addition of a coordinated
enamine complex (prepared from a b-diketone and an amino acid
derivative) to methyl vinyl ketone.13,14 This hypothesis prompted
us to examine whether indole 9a could result from a diketodien-
amine-ZnC12 complex and, if so, whether it depended on the pres-
ence or absence of a base. Recently, Lectka et al. successfully used
bifunctional catalyst systems, in which cinchona alkaloid deriva-
tives worked best when paired with Lewis acids based on, in par-
ticular, Zn(II) salts.15 An enaminone-ZnCl2 complex was reported
earlier.16 We observed that treatment of the diketodienamine 4a
with one equivalent of anhydrous ZnCl2 in THF or CH2Cl2 generated
a complex 12 (Scheme 3) in which, according to X-ray diffraction
data, ligand 4a coordinated to Zn(II) via two oxygen atoms
(Fig. 2).17 The complex alone failed to react with PBQ, but a high
level of acceleration for formation of indole 9a was achieved in
the presence of triethylamine as a base (>four-fold reduction in
reaction time). Thus, a deprotonated enaminone-ZnC12. complex
turned out to be an intermediate in the Nenitzescu reaction with
diketodienamine 4a and probably with the other enaminones as
well. Our findings confirmed that 4a was activated via precomplex-
ation with ZnC12 followed by deprotonation with a base.
Acknowledgment
The work was supported in part by the Russian Foundation for
Basic Research (Project 06-03-32557).
References and notes
1. Böhme, T. M.; Augelli-Szafran, C. E.; Hallak, H.; Pugsley, T.; Serpa, K.; Schwarz,
R. D. J. Med. Chem. 2002, 45, 3094–3102.
2. Shefali, S.; Srivastava, S. K.; Husbands, S. M.; Lewis, J. W. J. Med. Chem. 2005, 48,
635–638.
3. Zhao, C.; Zhao, Y.; Chai, H.; Gong, P. Bioorg. Med. Chem. 2006, 14, 2552–
2558.
4. Rönn, M.; McCubbin, Q.; Winter, S.; Veige, M. K.; Grimster, N.; Alorati, T.;
Plamondon, L. Org. Process Res. Dev. 2007, 11, 241–245.
5. Erlanson, D.; Wells, J.; Braisted, A. Tethering: Fragment-Based Drug Discovery.
Annu. Rev. Biophys. Biomol. Struct. 2004, 33, 199–223.
6. Allen, G. R. Org. React. 1973, 20, 337–454.
7. Gadaginamath, G. S.; Kavali, R. R.; Pujar, S. R. Synth. Commun. 2003, 33, 2285–
2292.
On the basis of the results obtained and the literature prece-
dents for the Lewis acid-catalyzed enamine-Michael reac-
tion,13,14,18 a plausible mechanism for the novel modification of
the Lewis acid-catalyzed Nenitzescu reaction is proposed in
Scheme 3.
The route involves in situ formation of a coordination complex
12 followed by its rearrangement into a complex with even more
acidic NH proton(s). When the latter is generated, however, it is
rapidly deprotonated by a base such as the enaminone component
itself or an additional base, to a highly nucleophilic complex 13.
Then, the first C–C bond is formed via an enamine-Michael reaction
between nucleophilic complex 13 and PBQ. The complex formation
is likely to prevent the reversion of this addition due to increasing
the rate of direct reaction through a shift of the equilibrium. It is
noted that a large or even stoichiometric amount of a Lewis acid
is necessary to promote the enamine-Michael reaction with methyl
vinyl ketone.18 The high nucleophilicity of complex 13 can also
facilitate the formation of cyclic intermediate 15 through a ring
closure step. A non-redox mechanism seems to occur on account
of a high rate for the cyclization step. An excess of PBQ proved
not to give rise to oxidative side products as the reaction took place
under conditions related to a redox mechanism.8,10,11 Our explana-
tion also underlines why Zn(II) works so efficiently as compared to
other Lewis acids, Al(III), in particular. This result is almost cer-
tainly due to the preferential binding of the more azaphilic Zn(II)
8. Grinev, A. N.; Shvedov, V. I.; Sugrobova, I. P. Zh. Obsch. Khim. 1961, 31, 2298–
2303.
9. Velezheva, V. S.; Kornienko, A. G.; Turashev, A. D.; Topilin, S. V.; Peregudov, A.
S.; Brennan, P. J. J. Heterocycl. Chem. 2006, 43, 873–879.
10. Bernatek, E.; Ledaal, T. Acta Chem. Scand. 1958, 12, 2053.
11. 3-Acetyl-5-hydroxy-2-methyl-1-(N-tosyl-2-aminoethyl)-indole (7a). To a solution
of 1,4-benzoquinone (0.33 g, 3.00 mmol) in CH2Cl2 (6 ml) was added ZnCl2
(0.04 g, 0.30 mmol). The resulting mixture was heated to reflux and then a
solution of enamine 2a (0.98 g, 3.30 mmol) in CH2Cl2 (10 ml) was added
drop by drop with stirring over 5–10 min. The mixture was stirred at reflux
for an additional 25 min and cooled to 0–5 °C for 2–3 h. The precipitated
crystals were filtered off and washed with CH2Cl2 (2 ꢀ 1 ml) and acetone
(2 ꢀ 1 ml) to afford 1.02 g (88%) of 7a. IR (KBr): mmax 3262 (NH), 1603 (C@O)
cmꢁ1 1H NMR (600 MHz, DMSO-d6): d 2.36 (3H, s, CH3 (Ts)), 2.48 (3H, s,
;
C(O)CH3), 2.65 (3H, s, 2-CH3), 2.98 (2H, q, J = 6.2 Hz, NH–CH2), 4.16 (2H, t,
J = 6.2 Hz, N–CH2), 6.62 (1H, dd, J = 8.6 Hz, 2.3 Hz, H6), 7.18 (1H, d, J = 8.6 Hz,
H7), 7.29, 7.55 (4H, 2d, J = 8.1 Hz, H2,3 (SO2-p-Tol)), 7.34 (1H, d, J = 2.3 Hz, H4),
7.77 (1H, t, J = 6.2 Hz, NH), 8.96 (1H, s, OH); 13C NMR (150 MHz, DMSO-d6): d
13.11, 21.41, 31.70, 42.09, 43.17, 106.00, 110.88, 111.73, 113.79, 126.86,
127.67, 130.10, 130.36, 137.64, 143.27, 145.10, 153.43, 193.27; MS m/z 386.
Anal. Calcd for C20H22N2O4S: C, 62.16; N, 7.25; H, 5.74. Found: C, 62.07; N, 7.19;
H, 5.81.
12. Crystallographic data for 9a (C18H22N2O3 ½(C3H6O), M = 343.42): crystals are
monoclinic, space group C2/c, at 120 K: a = 20.817(3), b = 10.6378(13),
c = 16.799(2) Å, b = 98.608(5)°, V = 3678.1(8) Å3, Z = 8, dcalc = 1.240 g cmꢁ3
,

V. S. Velezheva et al. / Tetrahedron Letters 49 (2008) 7106–7109
7109
l
(Mo
K
a
) = 0.85 cmꢁ1
,
F(000) = 1472. The refinement of 9a using 4864
17. Crystallographic data for 12 (C12H20Cl2N2O2Zn, M = 360.57): crystals are
monoclinic, space group P21/n, at 100 K: a = 8.1250(2), b = 18.9964(5),
independent reflections (2h < 58°) is converged to wR2 = 0.1650, GOF = 1.005
and R1 = 0.0622 (for 3001 observed reflections with I > 2
are available from CCDC 694501.
r(I)). Further details
c = 10.3853(3) Å,
b = 101.3350(10)°,
V = 1571.66(7) Å3,
Z = 4,
dcalc
=
1.524 g cmꢁ3
,
l
(Mo
K
a
) = 19.01 cmꢁ1
,
F(000) = 744. The refinement of
A
13. Christoffers, J.; Mann, A. Angew. Chem., Int. Ed. 2000, 39, 2752–2754.
14. Christoffers, J.; Mann, A. Angew. Chem., Int. Ed. 2001, 40, 4591–4597.
15. France, S.; Shah, M. H.; Weatherwax, A.; Wack, H.; Roth, J. P.; Lectka, T. J. Am.
Chem. Soc. 2005, 127, 1206–1215.
16. Haider, S. Z.; Malik, K. M. A.; Islam, M. A.; Hursthouse, M. B. J. Bangladesh Acad.
Sci. 1984, 8, 69–75.
using 4999 independent reflections (2h < 62°, Rint = 0.0359) is converged to
wR2 = 0.0505, GOF = 1.042 and R1 = 0.0191 (for 4720 observed reflections with
I > 2r(I)). Further details are available from CCDC 694502.
18. Borowka, J.; van Wullen, C. J. Organomet. Chem. 2006, 691, 4474–4479.
19. Liu, P.; Shi, J.-c.; Tong, Q.; Feng, Y.; Huang, H.; Jia, L. Inorg. Chim. Acta. 2008, in
press, doi:10.1016/j.ica.2008.03.113.