An extremely facile aza-Bergman rearrangement of sterically unencumbered acyclic 3-aza-3-ene-1,5-diynes

Article abstract of DOI:10.1021/jo0267192

The factors that affect the kinetics of the aza-Bergman cyclization of aza-enediynes (C,N-dialkynyl imines) have not previously been elucidated. Here we report our kinetic studies of the aza-Bergman reactions of a series of 6-triisopropylsilyl and 6-unsub

Full text of DOI:10.1021/jo0267192

An Extr em ely F a cile Aza -Ber gm a n Rea r r a n gem en t of Ster ica lly
Un en cu m ber ed Acyclic 3-Aza -3-en e-1,5-d iyn es
Liping Feng, Dalip Kumar, and Sean M. Kerwin*
Division of Medicinal Chemistry, College of Pharmacy, The University of Texas at Austin,
Austin, Texas 78712
skerwin@mail.utexas.edu
Received November 14, 2002
The factors that affect the kinetics of the aza-Bergman cyclization of aza-enediynes (C,N-dialkynyl
imines) have not previously been elucidated. Here we report our kinetic studies of the aza-Bergman
reactions of a series of 6-triisopropylsilyl and 6-unsubstituted 1-phenyl-4-aryl-3-aza-3-ene-1,4-diynes
in which the aryl group is phenyl, o-(methoxy)phenyl, or p-(methoxy)phenyl. These aza-enediynes
are prepared as single isomers in modest yield from the corresponding 1-aryl-3-(triisopropylsilyl)-
propynone oximes. These aza-enediynes undergo aza-Bergman reaction followed by a rapid retro-
aza-Bergman cyclization to afford â-alkynyl acrylonitrile products. In no case are products
corresponding to trapping the intermediate 2,5-didehydropyridine diradical isolated. While the rate
of aza-Bergman cyclization is not greatly affected by the nature of the 4-aryl substituent, the rate
is very dependent on the nature of the 6-substituent. 1-Phenyl-4-aryl-3-aza-3-ene-1,5-diynes that
lack a 6-substituent undergo aza-Bergman cyclization spontaneously at 20 °C with first-order half-
lives of 36-78 min. The effect of solvent on the kinetics of aza-Bergman cyclization of 1,4-diphenyl-
3-aza-3-ene-1,5-diyne was investigated. The rate of this cyclization is solvent dependent, proceeding
more rapidly in less polar solvents.
SCHEME 1
In 1972 Robert Bergman and co-workers reported the
gas-phase thermal rearrangement of substituted 3-hex-
ene-1,5-diynes (Scheme 1, A, X ) CH) and proposed the
existence of a 1,4-didehydrobenzene diradical (Scheme
1, B, X ) CH) in this process (Bergman cyclization).1
Interest in the Bergman cyclization has increased since
the mid-1980s due to the discovery of the naturally
occurring enediyne anticancer antibiotics, such as cali-
cheamicin,2 esperamicin,3 dynemicin,4 and kedarcidin.5
These enediynes are among the most potent antibiotic
antitumor agents ever investigated.6 The phenomenal
biological activity of these substances against bacterial
and tumor cells results from the fact that enediynes can
undergo a Bergman cyclization to benzenoid diradicals,
which abstract hydrogen atoms from the sugar phosphate
backbone of DNA, leading to oxidative DNA strand
scission.7 The extraordinary reactivity and cytotoxicity
of the enediynes has sparked considerable interest, yet
the lack of tumor specificity of these compounds has
complicated their use in cancer chemotherapy. One
approach that has been successfully employed to increase
the specificity of these agents is their conjugation to
cancer cell targeting antibodies.8 Alternative approaches
involve designing novel triggering mechanisms that may
lead to enediynes that only undergo diradical cyclization
in cancer cells.9 A more recent approach involves the
redesign of the enediyne core to provide both selective
triggering and more discriminating diradical intermedi-
ates.10 However, many factors which affect the Bergman
cyclization reaction still remain unresolved.
* To whom correspondence should be addressed. Phone: (512)-471-
5074. Fax: (512)-232-2606.
(1) (a) J ones, R. R.; Bergman, R. G. J . Am. Chem. Soc. 1972, 94,
660. (b) Bergman, R. G. Acc. Chem. Res. 1973, 6, 25.
(2) (a) Lee, M. D.; Dunne, T. S.; Siegel, M. M.; Chang, C. C.; Morton,
G. O.; Borders, D. B. J . Am. Chem. Soc. 1987, 109, 3464. (b) Lee, M.
D.; Ellestad, G. A.; Borders, D. B. Acc. Chem. Res. 1992, 24, 235.
(3) Konishi, M.; Ohkuma, H.; Saitoh, K.; Kawaguchi, H.; Golik, J .;
Dubay, G.; Groenewold, G.; Krishnan, B.; Doyle, T. W. J . Antibiot.
1985, 38, 1605.
Although the most studied factors which affect the
Bergman cyclization of many natural and synthetic
(4) Konishi, M.; Ohkuma, H.; Matsumoto, K.; Tsuno, T.; Kamei, H.;
Miyaki, T.; Oki, T.; Kawaguchi, H.; Van Duyne, G. D.; Clardy, J . J .
Antibiot. 1989, 42, 1449.
(5) Leet, J . E.; Schroeder, D. R.; Hofstead, S. L.; Golik, J .; Colson,
K. L.; Huang, S.; Klohy, S. E.; Doyle, T. W.; Matson, J . A. J . Am. Chem.
Soc. 1992, 114, 7946.
(7) De Voss, J . J .; Townsend, C. A.; Ding, W. D.; Morton, G. O.;
Ellestad, G. A.; Zein, N.; Tabor, A. B.; Schreiber, S. L. J . Am. Chem.
Soc. 1990, 112, 9669.
(8) Sievers, E. L. Exp. Opin. Biol. Therap. 2001, 1, 893.
(9) Dai, W. M.; Lai, K. W.; Wu, A. X.; Hamaguchi, W.; Lee, M. Y.
H.; Zhou, L.; Ishii, A.; Nishimoto, S. J . Med. Chem. 2002, 45, 758.
(10) Tuntiwechapikul, W.; David, W. D.; Kumar, D.; Salazar, M.;
Kerwin, S. M. Biochemistry 2002, 41, 5283.
(6) (a) Nicolaou, K. C.; Dai, W.-M. Angew. Chem. 1991, 103, 1453.
(b) Nicolaou, K. C.; Smith, A. L. Acc. Chem. Res. 1992, 25, 497.
10.1021/jo0267192 CCC: $25.00 © 2003 American Chemical Society
Published on Web 02/15/2003
2234
J . Org. Chem. 2003, 68, 2234-2242

Facile Aza-Bergman Rearrangement
enediynes are molecular strain energy11 and the distance
separating the acetylenic termini,6 electronic effects also
influence this process.12 Electron-withdrawing substitu-
ents either attached to triple-bond termini12a or associ-
ated with the double bond12,13 of enediynes can lower the
activation enthalpy of the Bergman reaction by decreas-
ing the degree of repulsion of the in-plane π-orbitals in
the transition state. We have investigated a new class of
compounds, the aza-enediynes, or C,N-dialkynyl imines
(Scheme 1, A, X ) N). The replacement of an sp2 carbon
with a nitrogen to afford such aza-enediynes was pre-
dicted to facilitate by a similar electronic effect their
cyclization by an aza-Bergman process. We reported the
first synthesis of the C,N-dialkynyl imines (2a ,b, eq 1)
The factors that affect the kinetics of the aza-Bergman
cyclization of aza-enediynes have not been elucidated.
There are only two reports of kinetic studies of this
process, and both of these were carried out on similar
aza-enediynes.14,20 Here we report our kinetic studies of
the aza-Bergman reactions of a series of 6-triisopropyl-
silyl and 6-unsubstituted 4-aryl-3-aza-3-ene-1,4-diynes.
These aza-enediynes are prepared as single isomers by
an adaptation of our previously reported route to the aza-
enediyne system. These aza-enediynes undergo aza-
Bergman reaction followed by a rapid retro-aza-Bergman
cyclization to afford â-alkynyl acrylonitrile products. In
no case are products corresponding to trapping the
intermediate 2,5-didehydropyridine diradical isolated.
While the rate of aza-Bergman cyclization is not affected
by the nature of the aryl substituent, the rate is very
dependent on the nature of the 6-substituent. 3-Aza-3-
ene-1,5-diynes that lack a 6-substituent undergo an
extremely facile aza-Bergman cyclization. The rate of this
cyclization is solvent dependent, proceeding more rapidly
in less polar solvents. The implications of these findings
to drug design efforts are discussed.
and their aza-Bergman cyclization in 1997.14 Although
these aza-enediynes underwent an aza-Bergman cycliza-
tion to generate 2,5-didehydropyridine intermediates
(Scheme 1, B, X ) N), all attempts to trap these
intermediates only resulted in the formation of the retro-
aza-Bergman products (Scheme 1, C, X ) N).14 Numerous
computational studies of the aza-Bergman cyclization and
the resulting 2,5-didehydropyridine intermediate have
been carried out.15-18 Significantly, the only experimental
support for the existence of the elusive 2,5-didehydropy-
ridine intermediate19 has come from studies of the aza-
Bergman cyclization reaction. Chen and co-workers20
reported the detection of miniscule amounts of pyridine
product by GC/MS in the thermolysis of the aza-enediyne
2a under acidic conditions. These workers propose that
the ability to trap the intermediate 2,5-didehydropyridine
is critically dependent upon this diradical’s singlet-
triplet gap, which is predicted to be larger than that of
the corresponding p-benzyne diradical. Protonation of the
2,5-didehydropyridine intermediate is predicted to lower
the singlet-triplet gap, which makes hydrogen atom
abstraction more competitive with the retro-aza-Bergman
process.20
Resu lts a n d Discu ssion
1. Syn th esis of 3-Aza -3-en e-1,5-d iyn es. Our previous
synthesis of C,N-dialkynyl imine aza-enediynes pro-
ceeded from the (E)-oxime 1 (eq 1), which was obtained
from the addition of lithium phenylacetylide to the tert-
butyldimethylsilyl nitronate ester of nitroethane.14 Ac-
tivation of the oxime 1 as the sulfonate ester followed by
cuprate coupling led to the aza-enediynes 2a ,b in modest
yields (eq 1). Attempts to employ this same route to
access other aza-enediynes were met with difficulty.
Additions of lithium phenylacetylide to the silyl nitronate
derived from nitromethyl benzene led to intractable
mixtures from which the desired oxime could not be
isolated. Condensation of 1,3-diphenylpropynone with
hydroxylamine only afforded 3,5-diphenylisoxazole. To
avoid the problem of isoxazole formation, we chose to
prepare propynone oximes bearing a large substituent
on the terminal acetylenic carbon. Thus, cuprous iodide-
catalyzed addition21 of triisopropylsilyl acetylene to ben-
zoyl chloride, p-anisoyl chloride, or o-anisoyl chloride
afforded the corresponding propynones 3a -c in good
yield (Scheme 2). The conversion of these acetylenic
ketones to the corresponding oximes 4a -c proceeded in
good yield, although the reactions required 10 days for
completion. The oximes 4a -c were isolated as single
isomers, as judged from 1H and 13C NMR spectra. The
corresponding mesylates 5a -c were also isolated as
single isomers of undetermined stereochemistry. Addition
of the cuprate reagent22 derived from phenylacetylene to
the mesylates 5a -c affords the C,N-dialkynyl imines
6a -c in modest yields along with variable amounts of
1-aryl-3-(triisopropylsilyl)-prop-2-ynylidenamine. The aza-
enediynes 6a -c can be purified by chromatography and
isolated as relatively stable, yellow oils that can be stored
at 0 °C for several months without any signs of decom-
position. The aza-enediynes 6a -c are isolated as pre-
(11) (a) Magnus, P.; Fortt, S.; Pitterna, T.; Snyder, J . P. J . Am.
Chem. Soc. 1990, 112, 4986. (b) Magnus, P.; Eisenbeis, S. A. J . Am.
Chem. Soc. 1993, 115, 953. (c) Snyder, J . P. J . Am. Chem. Soc. 1990,
112, 5367. (d) Chen, W.-C.; Chang, N.; Yu, C. J . Phys. Chem. 1998,
102, 2584. (e) Schreiner, P. R. J . Am. Chem. Soc. 1998, 120, 4148.
(12) (a) Schmittel, M.; Kiau, S. Chem. Lett. 1995, 953. (b) Semmel-
hack, M. F.; Neu, T.; Foubelo, F. Tetrahedron Lett. 1992, 3, 3277. (c)
Kim, C.-S.; Russell, K. C. J . Org. Chem. 1998, 63, 8229. (d) Choy, N.;
Russell, K. C. Heterocycles 1999, 51, 13.
(13) (a) J ones, G. B.; Plourde, G. W. Org. Lett. 2000, 2, 1757. (b)
J ones, G. B.; Warner, P. M. J . Am. Chem. Soc 2001, 123, 2134.
(14) David, W. M.; Kerwin, S. M. J . Am. Chem. Soc. 1997, 119, 1464.
(15) Nam, H. H.; Leroi, G. E.; Harrison, J . F. J . Phys. Chem. 1991,
95, 6514.
(16) (a) Cramer, C. J . J . Am. Chem. Soc. 1998, 120, 6261. (b)
Debbert, S. L.; Cramer, C. J . Int. J . Mass Spectrom. 2000, 201, 1.
(17) (a) Kraka, E.; Cremer, D. J . Am. Chem. Soc. 2000, 122, 8245.
(b) Kraka, E.; Cremer, D J . Comput. Chem. 2001, 22, 216.
(18) Schutz, M. J . Chem. Phys. 2002, 4, 3941.
dominantly one isomer, as judged by H and 13C NMR.
1
(19) Reinecke, M. Tetrahedron 1982, 38, 427 and references therein.
(20) Hoffer, J .; Schottelius, M. J .; Feichhtinger, D.; Chen, P. J . Am.
Chem. Soc. 1998, 120, 376.
(21) Wang, J .-X.; Wei, B.; Hu, Y.; Liu, Z.; Fu, Y. Synth. Commun.
2001, 31, 3527.
(22) Wu¨rthwein, E.-U.; Weigmann, R. Angew. Chem. 1987, 99, 918.
J . Org. Chem, Vol. 68, No. 6, 2003 2235

Feng et al.
SCHEME 2
SCHEME 3
2. Aza -Ber gm a n Kin etics. Thermolysis of 6a was
performed by heating compound 6a in chlorobezene in
the presence of 20 equiv of 1,4-cyclohexadiene (1,4-CHD)
as proton donor in a sealed tube at 150 °C. The disap-
pearance of starting material was monitored by TLC,
which demonstrated that the reaction required 3 days
to reach completion. We were unable to identify the
product derived from trapping of the intermediate 2,5-
didehydropyridine diradical from the reaction mixture;
however, the retro-Bergman cyclization product, 2,5-
diphenyl-3-triisopropylsilyl-pent-2-en-4-ynenitrile 9 (eq
2), was isolated in 35% yield.
F IGURE 1. The conversion of aza-enediyne 7a (alkyne
resonance at 3.9 pm) to nitrile 8a (alkene resonance at 6.8
ppm) followed by 1H NMR (500 MHz) in CDCl3 at 20 °C.
Spectra were acquired every 1010 s, starting with the bottom
spectrum.
The conversion of the aza-enediyne 7a to nitrile 8a
could also be followed by UV spectroscopy. Nitrile 8a has
a UV absorbance peak at 345 nm and the aza-enediyne
7a has an absorbance peak at 390 nm. Dilute (µM)
solutions of the product mixture from the deprotection
of 6a in CHCl3 were quickly prepared and placed in
temperature-controlled cuvettes to monitor the UV ab-
sorbance change with time. As shown in Figure 2, the
absorbance spectra demonstrate isobestic behavior and
the expected first-order kinetics. At 20 °C, the half-life
for the conversion of 7a to 8a determined spectrophoto-
metrically (71 ( 1 min, average of two runs) is in
Initial attempts to deprotect aza-enediyne 6a by treat-
ment with TBAF in THF afforded only the rearranged
nitrile 8a as the sole product in 89% yield (Scheme 3).
Despite the inability to isolate and purify the intermedi-
ate deprotected aza-enediyne 7a , rapid workup of the
deprotection reactions at low temperature followed by 1H
NMR of the unpurified reaction product in CDCl3 dem-
onstrated the presence of the deprotected aza-enediyne
7a (Figure 1). The terminal alkyne resonance for 7a at
3.9 ppm diminished with time and was replaced by a
resonance for the alkene proton at 6.8 ppm as 7a
underwent conversion to the nitrile 8a over a period of
minutes (Figure 1). The conversion of 7a to 8a follows
first-order kinetics for both the disappearance of 7a and
the appearance of 8a , and the kinetic yield of 8a (ratio
of the rate of disappearance of 7a over the rate of
appearance of 8a ) is >99%. At 20 °C, the 1H NMR-
determined half-life is 79 ( 4 min (average of two runs).
1
agreement with that determined by H NMR.
3. Su bstitu en t Effects. A combination of 1H NMR and
spectrophotometric kinetic data were used to follow the
1
conversion of 7b to 8b at 20 °C. The time-dependent H
NMR spectra of crude 7b in CDCl3 demonstrates the
same features noted above for 7a , namely, the first-order
disappearance of a peak at 3.9 ppm corresponding to the
terminal alkyne proton in 7b and the appearance of a
new peak at 6.9 ppm, corresponding to the alkene proton
in 8b. In addition, in the case of 7b, the methoxy protons’
resonance for the aza-enediyne (3.86 ppm) is well re-
solved from the methoxy protons’ resonance for the nitrile
8b (3.82 ppm), and these signals also undergo a first-
order decay and increase, respectively, with a kinetic
yield of >99% (data not shown). The time-dependent UV
absorbance spectra of dilute solutions of 7b in CHCl3 are
2236 J . Org. Chem., Vol. 68, No. 6, 2003

Facile Aza-Bergman Rearrangement
calculations of a variety of aza-enediynes in which the
imine double bond is replaced with an amide, amidine,
or amidinium group predict that the barrier to aza-
Bergman cyclization is relatively insensitive to these
changes.17 The lack of an observable difference in the rate
of aza-Bergman reaction of 7a and 7b is not surprising
in light of these theoretical studies.
The conversion of the o-(methoxy)phenyl-substituted
1
aza-enediyne 7c to nitrile 8c can also be followed by H
1
NMR spectroscopy in CDCl3. The overlap of the H NMR
signals due to the terminal alkyne proton and methoxy
group of 7c and the methoxy group of 8c preclude the
use of these signals to accurately follow the course of the
1
reaction. Instead, a distinctive aromatic H NMR reso-
nance for 6c at 7.9 ppm can be used to determine a first-
order half-life of 35 min at 20 °C (one run). The UV
absorbance changes accompanying the conversion of 7c
to 8c in CHCl3 are similar to those described above for
the para-isomer 7b, and give a first-order half-life of 36
( 1 min at 20 °C (for two runs). Thus, the aza-Bergman
reaction of 7c at 20 °C occurs approximately two-times
faster than that of either 7a or 7b (Table 1). The origin
of this modest substituent effect is not known; however,
it is likely that the steric interaction between the o-
methoxy group and the adjacent alkyne moiety in 7c
plays a role.
The relative insensitivity of the rate of aza-Bergman
cyclization to the nature of the imine double bond
substituent indicates that this region of the aza-ene-
diynes provides a great deal of flexibility in the design
of potential warheads for targeting DNA and other
receptors. This is significant because the imine double
bond of aza-enediynes is a site of hydrolytic instability.
Chen and co-workers note that thermolysis reactions of
2a under acidic conditions afford primarily products of
imine hydrolysis.20 This hydrolytic instability could be
remedied by replacing the imine double bond with the
amidine group as proposed by Kraka and Cremer,17 or
by incorporating the imine double bond into a heterocyclic
ring.26 The latter approach has the advantage of enforcing
the proper orientation of the alkynyl substituents re-
quired for aza-Bergman cyclization while providing a
diverse set of chemotypes with which to design specific
targeting compounds.
F IGURE 2. The conversion of aza-enediyne 7a to nitrile 8a
followed by UV-absorption spectroscopy in CHCl3 at 45 °C. The
spectra were taken every 1.8 min for 20 min. The arrows
indicate the decrease in absorbance due to the aza-enediyne
7a at 390 nm and the increase in absorbance due to the nitrile
8a at 345 nm with time.
TABLE 1. Ha lf-Lives (t1/2) for th e Aza -Ber gm a n
Cycliza tion of En ed iyn es 2a , 7a , 7b, a n d 7c
aza-enediyne
temp (°C)
t1/2 (min)
7a
7a
7a
7a
7a
7a
7a
7b
7c
2a
0
10
20
25
30
37
45
20
20
667a
276 ( 13b
75 ( 6c
40.6 ( 0.1d
21.5 ( 0.1d
10.4 ( 0.1d
4.3 ( 0.1d
77 ( 3e
35.6 ( 0.5e
47f
110
a
b
Based on one 1H NMR run in CDCl3. Average value for two
1H NMR runs in CDCl3. c Based on two 1H NMR runs in CDCl3
d
and two spectrophotometric runs in CHCl3. Average of two
spectrophotometric runs in CHCl3. e Based on one 1H NMR run
in CDCl3 and two spectrophotometric runs in CHCl3. f Data from
ref 14 in CH3CN.
characterized by a decrease in a peak at 415 nm,
corresponding to the aza-enediyne, and an increase in a
peak at 360 nm, corresponding to the nitrile 8b . The
first-order half-life for the conversion of 7b to 8b is
77 ( 3 min. Thus, there is no apparent electronic
substituent effect on the rate of the aza-Bergman cy-
clization when comparing 7a and 7b (Table 1). This result
is in contrast to a number of reports of substituent effects
in the Bergman cyclization. Maier and Greiner23 have
reported a pronounced retardation of the Bergman cy-
clization of the bicyclo[7.3.1]enediyne 10 when the double
bond is substituted with an electron-donating group (R
) p-C6H4OMe). J ones and Plourde have reported the
retardation of the Bergman cyclization of a series of cyclic
enediynes 11 when the double bond is substituted with
chlorine.24 A systematic study of substituent effects in
the Bergman cyclization of 1,2-dialkynylbenzene deriva-
tives 12 by Russell and co-workers demonstrated a linear
free energy relationship between the cyclization rate and
the Hammett σm substituent coefficient.25 In contrast,
In contrast to the modest substituent effects observed
when comparing the aza-Bergman cyclization rate of
7a -c, the nature of the alkyne substituents appears to
play an important role in the rate of aza-Bergman cycliza-
tion. The aza-enediyne 6a , which bears a bulky triiso-
(23) Maier, M. E.; Greiner, B. Liebigs Ann. Chem. 1992, 855.
(24) J ones, G. B.; Plourde, G. W. Org. Lett. 2000, 2, 1757.
(25) Choy, N.; Kim, C. S.; Ballestero, C.; Artigas, L.; Diez, C.;
Lichtenberger, F.; Shapiro, J .; Russell, K. C. Tetrahedron Lett. 2000,
41, 6955.
(26) Nadipuram, A. K.; David, W. M.; Kumar, D.; Kerwin, S. M. Org.
Lett. 2002, 4, 4543.
J . Org. Chem, Vol. 68, No. 6, 2003 2237

Feng et al.
propylsilyl substituent on the C-alkynyl group, undergoes
aza-Bergman cyclization only under rather forcing condi-
tions (eq 2). Removal of the triisopropylsilyl group from
6a to give 7a results in facile aza-Bergman cyclization.
The phenyl-substituted aza-enediyne 2a undergoes aza-
Bergman cyclization at a rate that is intermediate
between that for 6a and 7a (Table 1). Schreiner27 and
co-workers have performed calculations on substituent
effects on the Bergman cyclization of (Z)-1,5-hexadiyne-
3-enes. According to these calculations, a diphenyl-
substituted enediyne (Figure 1, A, X ) CH, R1 ) R3 )
Ph, R2 ) H) has a Bergman cyclization transition state
that is 10 kcal/mol higher than that for the phenyl-
substituted enediyne (Figure 1, A, X ) CH, R1 ) Ph, R2
) R3 ) H), an effect that is due to a combination of
additional stabilization of the starting enediyne and
destabilization of the transition state due to steric
repulsion between the phenyl groups in the diphenyl-
substituted case. In the parent enediyne case, the effect
of changing the size of the terminal substituents on the
facility of Bergman cyclization cannot be employed as a
triggering mechanism at physiological temperatures due
to the large barrier to Bergman cyclization of the unsub-
stitued enediyne;28 however, in the case of the aza-
enediynes studied here, the terminal alkyne substituent
effect can be viewed as a potential triggering device due
to the facility with which sterically unencumbered aza-
enediynes such as 7a -c undergo aza-Bergman cycliza-
tion. Of course, such a triggering device would only lead
to hydrogen atom abstraction or cyclization chemistry if
the intermediate 2,5-didehydropyridine diradical could
enter into free radical reactions prior to retro-aza-
Bergman cyclization. This is not the case under the
conditions employed in this work. Chen and co-workers
have demonstrated that the large singlet-triplet gap for
2,5-didehydropyridine is related to its inability to par-
ticipate in hydrogen atom abstraction reactions. Proto-
nation of the nitrogen atom of the 2,5-didehydropyridine
lowers the singlet-triplet gap and increases the barrier
to retro-aza-Bergman rearrangement. Chen and co-
workers have reported the trapping of these diradicals
when the aza-Bergman cyclization is performed under
acidic conditions.20
F IGURE 3. Arrhenius plot of the aza-Bergman cyclization of
aza-enediyne 7a . Fifteen points at seven temperatures (0, 10,
20, 25, 30, 37, 45 °C) yield Ea ) 19.9 ( 0.3 kcal/mol and log A
) 11.1 ( 0.2 (R2 ) 0.993).
equilibration during their preparation and isolation, the
predominant isomers observed for 7a -c are the more
thermodynamically stable (Z)-isomer.
As expected for an aza-Bergman process, the spontane-
ous rearrangement of 7a -c to the nitriles 8a -c occurs
stereoselectively to afford the nitrile products as single
isomers. Assignment of the (Z)-stereochemistry of these
nitriles is based upon their isomerization to a mixture
1
of (E)- and (Z)-isomers. The H NMR spectum of 8a in
CD3CN contains a singlet resonance at 7.1 ppm, corre-
sponding to the alkene hydrogen. When this solution is
subjected to irradiation at 371 nm, using the excitation
beam of a spectroflourimeter,30 the 1H NMR spectrum
shows a new single peak at 6.8 ppm, which is due to the
alkene proton of the isomer of 8a . Because the resonance
of the alkene proton of (E)-8a is predicted to lie 0.34 ppm
upfield of that for the (Z)-isomer,31 we assign the 7.1-
ppm peak to the (Z)-isomer and the 6.8-ppm peak to the
(E)-isomer. After 4.5 h of irradiation, 34% (E)-8a was
formed based on the ratio of the peaks at 6.8 and 7.1
ppm. For conversion of imine 7b to nitrile 8b, the 1H
NMR (CDCl3) of the product mixture prior to purifica-
tion showed one singlet peak at 6.8 ppm, due to the
(Z)-isomer of 8b . Photochemical isomerization of 8b
results in the formation of an isomer with a new 1H
NMR resonance at 6.6 ppm, which is associated with the
(E)-nitrile (20%).
4. Ster eoch em ica l Asp ects. The aza-Bergman cy-
clization of C,N-dialkynyl imine aza-enediynes can only
occur through the (Z)-imine isomer. Previous studies of
the aza-Bergman cyclization have employed aza-ene-
diynes (e.g., 2a ,b, eq 1) that were mixtures of stereo-
chemical isomers about the imine double bond.14,20 As a
result, the observed kinetics for the conversion of these
aza-enediynes to products of aza-Bergman cyclization
may have been affected by the rate of (E)- to (Z)-imine
isomerization. The deprotected aza-enediynes 7a -c ap-
5. Activa tion En er gy for th e Aza -Ber gm a n Cy-
cliza tion . The kinetics of the conversion of 7a to 8a were
determined as a function of temperature. At 10 °C, the
1H NMR-determined half-life is 4.6 h, and at 0 °C the
half-life is 11 h (Table 1). The half-life for the aza-
Bergman reaction of 7a at 37 °C is only 10.4 min, as
determined spectrophotometrically (Table 1). An Arrhe-
nius plot for the aza-Bergman reaction of 7a to nitrile
8a , shown in Figure 3, gives Ea ) 19.9 ( 0.3 kcal/mol
and log A ) 11.1 ( 0.2. The activation energy for the
aza-Bergman reaction of 7a is lower than that for 1,6-
bis(4-tert-butylphenyl)-3-aza-4-methylhex-3-ene-1,5-
diyne (Figure 1, A, X ) N, R1 ) R3 ) 4-(t-Bu)Ph, R2 )
Me), which has been studied by Chen and co-workers,
who determined Ea ) 23.1 ( 1.5 kcal/mol.20 The lower
activation energy for the aza-Bergman reaction of 7a
1
pear to be single isomers (>9:1) by H NMR (Figure 1).
Density functional theory calculations employing the
B3LYP functional29 predict that the (Z)-isomer of 7a is
more stable than the (E)-isomer by 4.4 kcal/mol. Assum-
ing that these aza-enediynes undergo thermodynamic
(27) Prall, M.; Wittkopp, A.; Fokin, A. A.; Schreiner P. J . Comput.
Chem. 2001, 22, 1605.
(28) For an example of steric effects in the Bergman cyclization of
simple enediyne see: Rawat, D. S.; Zaleski, J . M. Chem Commun 2000,
2493.
(30) Yu, H.-T.; Hurley, L. H.; Kerwin, S. M. J . Am. Chem. Soc. 1996,
118, 7040.
(31) Predictions of 1H NMR resonances based on the ChemDraw
Ultra program.
(29) Becke, A. D. J . Chem. Phys. 1993, 98, 5648.
2238 J . Org. Chem., Vol. 68, No. 6, 2003

Facile Aza-Bergman Rearrangement
TABLE 2. F ir st-Or d er Ha lf-Lives (t1/2) for th e
Aza -Ber gm a n Cycliza tion of Aza -En ed iyn e 7a in Va r iou s
Solven ts a t 45 °C
enediyne 2a was isolated as a mixture of (E)- and (Z)-
isomers. While the aza-Bergman reaction of 2a would be
expected to show a similar solvent dependence as 7a , the
(E)- to (Z)-isomerization rate may have a more compli-
cated solvent dependence,34 resulting in a fortuitous
canceling of solvent polarity effects for the two solvents
investigated.
Although the solvent dependence of the rate of aza-
Bergman cyclization of 7a is small, the increased facility
of aza-Bergman cyclization in nonpolar solvents implies
that aza-enediynes such as 7a would be somewhat more
stable to aza-Bergman cyclization in aqueous solution as
compared to when bound to a target receptor, where the
microenvironment is less polar.
half-lifeb
(t1/2, min)
dielectric
constant (D)
solventa
hexanes
chlorobenzene
chloroform
THF
3.61 ( 0.03
4.2 ( 0.1
1.89
2.71
4.81
7.60
18.30
37.50
4.3 ( 0.1
6.4 ( 0.2
isopropyl alcohol
acetonitrile
7.91 ( 0.04
8.14 ( 0.03
a
All solvents were HPLC or ACS grade and used directly
without further purification. b Average value ( standard deviation
for two spectrophotometric runs.
relative to the 1,6-diaryl aza-enediyne studied by Chen
and co-workers is most likely due to the substituent
effects discussed above in the comparison of the facility
of cyclization of 7a versus 2a . The activation energy for
the aza-Bergman cyclization of 7a is lower than the
barrier for aza-Bergman cyclization of the parent aza-
enediyne system (Figure 1, A, X ) N, R1 ) R2 ) R3 ) H)
calculated at the CASMP2 level of theory (27.7 kcal/
mol),20 but close to the value determined by B3LYP
calculations (21.8 kcal/mol),17 and that calculated by
Cramer using a composite energy term (19.1 kcal/mol).16
It is probable that the true energy barrier for cyclization
in the parent aza-enediyne system is even lower than
that determined for 7a , in which case the later calcula-
tions may also overestimate this barrier.
6. Solven t Effects. We have performed kinetic studies
of 7a in various solvents to determine solvent effects on
the aza-Bergman cyclization. The conversion of 7a to 8a
at 45 °C in various solvents was followed by UV-vis
spectroscopy. Kinetic data in all solvents were first-
order and analysis of the reaction mixtures by TLC and
mass spectroscopy confirmed the presence of 8a as the
sole reaction product. The aza-Bergman cyclization rates
were found to be solvent dependent, with slightly faster
rates in nonpolar solvents (Table 2). The effect is not
large; the half-lives varied from 3.6 min in hexanes to
8.2 min in acetonitrile. There are examples of enediynes
that undergo solvent-dependent Bergman cycloaromati-
zation rates,32,33 but these cases are complicated by the
solvent effects on the rate of hydrogen atom abstraction
by the diradical intermediate, which impacts the ob-
served rate of disappearance of enediyne. In the case of
the conversion of 7a to 8a , it is clear that the rate
determining step is the initial aza-Bergman cyclization.
We believe that the solvent-dependent rate of aza-
Bergman reaction of 7a shown in Table 3 is due to
preferential solvation of 7a by more polar solvents,
resulting in ground-state stabilization and concomitant
decreased rate.
Con clu sion
Since the initial report14 of the aza-Bergman reaction
of aza-enediynes, these compounds have been the subject
of a number of theoretical studies.16-18,20 These studies
have focused on two issues associated with these aza-
enediynes and their aza-Bergman cyclization: the facility
of the cyclization and the potential reactivity of the
resulting 2,5-didehydropyridine intermediate. For these
aza-enediynes to serve as potential warheads for target-
ing cancer cells, the aza-Bergman cyclization rate must
be sufficiently fast at physiological temperature, and the
resulting diradical must be capable of entering into free
radical chemistry in competition with its collapse via a
retro-aza-Bergman cyclization to the thermodynamically
stable â-alkynyl acrylonitrile product. There have only
been two experimental reports on aza-enediynes that
addressed either of these issues, and both of these were
carried out with similar aza-enediynes.14,20 Thus, the
effects of substituents on the rate of aza-Bergman cy-
clization had not previously been explored. Here we
report the synthesis and kinetic studies of the aza-
Bergman reactions of a series of 6-triisopropylsilyl and
6-unsubstituted 4-aryl-3-aza-3-ene-1,4-diynes. There are
two types of substituent effects in this system: The aza-
Bergman reaction rate is not greatly affected by changing
the 4-aryl substituent on the aza-enediyne imine double
bond, but is very sensitive to the nature of the 6-substi-
tutent at the alkyne terminus. The aza-Bergman cycliza-
tion of sterically unencumbered, 6-unsubstituted aza-
enediynes occurs readily at temperatures as low as 0 °C.
These observations provide the basis for the design of a
novel aza-Bergman triggering device involving the con-
version of a sterically blocked aza-enediyne to a sterically
unencumbered one mediated by a cancer cell-specific
enzymatic activity.
(34) Asano, T.; Okada, T.; Herkstroeter, W. G. J . Org. Chem. 1989,
52, 379.
(35) Frisch, M. J .; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.;
Robb, M. A.; Cheeseman, J . R.; Zakrzewski, V. G.; Montgomery, J . A.,
J r.; Stratmann, R. E.; Burant, J . C.; Dapprich, S.; Millam, J . M.;
Daniels, A. D.; Kudin, K. N.; Strain, M. C.; Farkas, O.; Tomasi, J .;
Barone, V.; Cossi, M.; Cammi, R.; Mennucci, B.; Pomelli, C.; Adamo,
C.; Clifford, S.; Ochterski, J .; Petersson, G. A.; Ayala, P. Y.; Cui, Q.;
Morokuma, K.; Rega, N.; Salvador, P.; Dannenberg, J . J .; Malick, D.
K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J . B.; Cioslowski, J .;
Ortiz, J . V.; Baboul, A. G.; Stefanov, B. B.; Liu, G.; Liashenko, A.;
Piskorz, P.; Komaromi, I.; Gomperts, R.; Martin, R. L.; Fox, D. J .; Keith,
T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Challacombe, M.;
Gill, P. M. W.; J ohnson, B.; Chen, W.; Wong, M. W.; Andres, J . L.;
Gonzalez, C.; Head-Gordon, M.; Replogle, E. S.; Pople, J . A. Gaussian
98, Revision A.11.4; Gaussian, Inc.: Pittsburgh, PA, 2002.
In our previous studies of the aza-Bergman reaction
of aza-enediyne 2a , we noted that the reaction rates were
nearly identical in acetonitrile and heptane, the only two
solvents investigated.14 The lack of an apparent solvent
dependence on the aza-Bergman reaction of 2a may be
due to its decreased interaction with polar solvents when
compared to 7a . In addition, as noted above, the aza-
(32) Yoshida, K.; Minami, Y.; Otani, T.; Tada, Y.; Hirama, M.
Tetrahedron Lett. 1994, 35, 5253.
(33) Kim, C.; Russell, K. C. Tetrahedron Lett. 1999, 40, 3835.
J . Org. Chem, Vol. 68, No. 6, 2003 2239

Feng et al.
1-(2-Meth oxyp h en yl)-3-(tr iisop r op ylsilyl)-p r op yn on e
(3c). Propynone 3c was prepared from o-anisoyl chloride by a
procedure analogous to that described for the preparation of
3a . The product was purified by flash chromatography (10%
EtOAc in hexanes) to afford 4.62 g (74%) of 3c as a transparent
oil: Rf 0.85 (CH2Cl2); 1H NMR (CDCl3) δ 0.95-1.20 (m, 21H),
3.85 (s, 3H), 6.90-7.01 (m, 2H), 7.45 (ddd, J ) 8.2, 6.0, 2.0
Hz, 1H), 7.99 (dd, J ) 8.2, 2.0 Hz, 1H); 13C NMR (CDCl3) δ
10.94, 18.34, 55.51, 95.50, 105.11, 111.84, 119.96, 126.25,
132.56, 134.72, 159.96, 176.11; HRMS m/z 317.1944 (calcd
317.1936, C19H28O2Si).
One issue that remains to be addressed is the reactivity
of the 2,5-didehydropyridine intermediates involved in
the aza-Bergman cyclization of these aza-enediynes. In
this study, there is no evidence of any products corre-
sponding to trapping of these 2,5-didehydropyridine
intermediates. These results are in accord with many
theoretical predictions of an extremely low barrier to
retro-aza-Bergman cyclization of these intermediates to
afford the observed â-alkynyl acrylonitrile products.15,16,20,27
The lack of diradical reactivity associated with the 2,5-
didehydropyridine intermediate has been rationalized
based upon its large singlet-triplet gap.20 Substituting
the aza-enediyne imine double bond17,20 or incorporating
it within a heterocyclic ring26 has the potential to both
increase the barrier to retro-aza-Bergman reaction and
increase the diradical character of the aza-Bergman-
derived 2,5-didehydropyridine intermediates. The results
presented here indicate that such substitutions may be
compatible with physiologically relevant aza-Bergman
cyclization rates.
1-P h en yl-3-(tr iisop r op ylsilyl)-p r op yn on e Oxim e (4a ).
To a solution of 3a (0.64 g, 2.23 mmol) in 8 mL of EtOH was
added hydroxylamine hydrochloride (186 mg, 2.68 mmol) and
sodium acetate (220 mg, 2.68 mmol). The mixture was stirred
at room temperature under argon for 10 days. The reaction
mixture was poured into water (20 mL), and the water layer
was extracted by CH2Cl2 (3 × 15 mL). The residue upon drying
and concentration of the organic layer was purified by flash
chromatography (0-20% CH2Cl2 in hexanes) to afford 0.51 g
1
(77%) of oxime 4a as a transparent oil: Rf 0.54 (CH2Cl2); H
NMR (CDCl3) δ 1.04-1.30 (m, 21H), 7.38-7.42 (m, 3H), 7.80-
7.89 (m, 2H), 8.55 (s, 1H); 13C NMR (CDCl3) δ 11.13, 18.61,
94.61, 107.42, 126.46, 128.46, 129.82, 133.12, 141.97; HRMS
m/z 302.1946 (calcd 302.1940, C18H27NOSi).
Exp er im en ta l Section
1-(4-Met h oxyp h en yl)-3-(t r iisop r op ylsilyl)p r op yn on e
Oxim e (4b). Oxime 4b was prepared from propynone 3b by a
procedure analogous to that described for the preparation of
4a . The product was purified by flash chromatography on silica
gel (0-20% CH2Cl2 in hexanes) to afford 2.01 g (79%) of oxime
4b as a light yellow oil: Rf 0.50 (CH2Cl2); 1H NMR (CD2Cl2) δ
0.96-1.35 (m, 21H), 3.85 (s, 3H), 6.93 (d, J ) 8.1 Hz, 2H),
7.82 (d, J ) 8.1 Hz, 2H), 8.80 (s, 1H); 13C NMR (CD2Cl2) δ
11.58, 18.80, 55.74, 97.90, 107.23, 114.26, 126.14, 128.26,
All commercial chemicals were used without further puri-
fication unless otherwise noted. Tetrahydrofuran (THF) was
re-distilled from Na/benzophenone, methylene chloride and
pyridine were re-distilled from calcium hydride, and diethyl
ether was re-distilled from lithium aluminum hydride (LAH)
prior to use. All reactions were performed under argon. Unless
otherwise noted, organic extracts were dried with Na2SO4,
filtered through a fritted glass funnel, and concentrated with
a rotary evaporator (20-30 mmHg). Rf values are reported for
thin-layer chromatography (TLC) performed on silica gel TLC
plates, using the indicated mobile phase. Flash chromatogra-
142.00, 161.52; HRMS m/z 332.2036 (calcd 332.2045, C19H29
-
NO2Si).
1-(2-Met h oxyp h en yl)-3-(t r iisop r op ylsilyl)p r op yn on e
Oxim e (4c). Oxime 4c was prepared from propynone 3c by a
procedure analogous to that described for the preparation of
4a . The product was purified by flash chromatography (0-
20% CH2Cl2 in hexanes) to afford 1.24 g (79%) of oxime 4c as
a white solid: Rf 0.40 (CH2Cl2); mp 120.2-121.0 °C; 1H NMR
(CD2Cl2) δ 1.10-1.22 (m, 21H), 3.89 (s, 3H), 6.97-7.08 (m, 2H),
7.39 (ddd, J ) 8.0, 7.0, 2.0 Hz, 1H), 7.80 (dd, J ) 8.0, 2.0 Hz,
1H); 13C NMR (CD2Cl2) δ 11.57, 18.78, 55.71, 96.46, 105.56,
111.84, 120.77, 122.20, 131.01, 131.39, 140.42, 157.68; HRMS
m/z 332.2058 (calcd 332.2045, C19H29NO2Si).
1
phy was preformed with EM silica gel. H NMR spectra were
recorded at 500, 300, and 250 MHz and 13C NMR spectra were
recorded at 75 and 62 MHz. Spectroscopic studies were
performed on a spectrophotometer with a Peltier thermostated
cuvette holder. All mass spectra were obtained by chemical
ionization with methane as the ionizing gas. Melting points
are uncorrected. Density functional theory computations were
carried out with Guassian 98.35
1-P h en yl-3-(tr iisopr opylsilyl)-pr opyn on e (3a). To a mix-
ture of triisopropylsilyl acetylene (0.5 g, 2.74 mmol) and
cuprous iodide (26.6 mg, 0.14 mmol) in triethylamine (9 mL)
was added benzoyl chloride (0.5 g, 3.56 mmol). The reaction
mixture was stirred at room temperature under argon for 30
h. After the removal of the solvent, methanol (3 mL) was added
and the mixture was stirred for an additional 5 min, during
which time the mixture became clear. The methanol was
distilled off under vacuum, and the residue was poured into
water and extracted with dichloromethane (2 × 20 mL) and
dried over Na2SO4. The solvent was distilled off and the
product was purified by Kugelrohr distillation (110 °C, 1
mmHg) to afford 0.693 g (87%) of propynone 3a as a clear oil:
1H NMR (CDCl3) δ 1.01-1.22 (m, 21H), 7.46-7.55 (m, 3H),
8.16-8.23 (m, 2H); 13C NMR (CDCl3) δ 11.07, 18.53, 97.91,
103.01, 128.53, 129.49, 134.00, 136.70, 177.43; HRMS m/z
287.1826 (calcd 287.1831, C18H26OSi).
1-P h en yl-3-(tr iisopr opylsilyl)-pr opyn on e Mesylate (5a).
To a solution of oxime 4a (2.7 g, 8.96 mmol) in 60 mL of dry
CH2Cl2 was added dry pyridine (3.6 mL, 44.8 mmol). The
reaction mixture was cooled to 0 °C, followed by dropwise
addition of a solution of methanesulfonyl chloride (1.28 g, 11.2
mmol) in 5 mL of dry CH2Cl2. The reaction mixture was stirred
at room temperature for 3 days. The solvent volume was
reduced by rotary evaporation, and the residue was poured
into ice-water. The aqueous layer was extracted by CH2Cl2
three times. The residue upon drying and concentration of the
organic layer was purified by flash chromatography (0-20%
CH2Cl2 in hexanes) to afford 2.6 g (78%) of mesylate 5a as a
1
yellow oil: Rf 0.76 in CH2Cl2; H NMR (CDCl3) δ 1.04-1.28
(m, 21H), 3.25 (s, 3H), 7.39-7.52 (m, 3H), 7.94 (dd, J ) 8.0,
1.7 Hz, 2H); 13C NMR (CDCl3) δ 11.01, 18.50, 36.71, 93.37,
111.53, 127.52, 128.68, 130.89, 131.79, 148.58; HRMS m/z
380.1718 (calcd 380.1715, C19H29NOSSi).
1-(4-Meth oxyp h en yl)-3-(tr iisop r op ylsilyl)-p r op yn on e
(3b). Propynone 3b was prepared from p-anisoyl chloride by
a procedure analogous to that described for the preparation
of 3a . The product was purified by flash chromatography (0-
30% CH2Cl2 in hexanes) to afford 4.8 g (92%) of 3b as a
1-(4-Meth oxyp h en yl)-3-(tr iisop r op ylsilyl)-p r op yn on e
Mesyla te (5b). Mesylate 5b was prepared from oxime 4b by
a procedure analogous to that described for the preparation
of 5a . The product was purified by flash chromatography on
silica gel (0-30% CH2Cl2 in hexanes) to afford 450 mg (50%)
of mesylate 5b as a yellow solid: mp 85.8-86.9 °C; Rf 0.73
(CH2Cl2); 1H NMR (CDCl3) δ 1.05-1.20 (m, 21H), 3.20 (s, 3H),
3.81 (s, 3H), 6.91 (d, J ) 8.7 Hz, 2H), 7.87 (d, J ) 8.7 Hz, 2H);
1
transparent oil: Rf 0.85 (CH2Cl2); H NMR (CD2Cl2) δ 1.11-
1.30 (m, 21H), 3.86 (s, 3H), 6.95 (d, J ) 8.0 Hz, 2H), 8.13 (d,
J ) 8.0 Hz, 2H); 13C NMR (CD2Cl2) δ 11.54, 18.74, 55.99, 97.21,
103.87, 114.24, 130.80, 132.12, 165.04, 176.23; HRMS m/z
317.1930 (calcd 317.1936, C19H28O2Si).
2240 J . Org. Chem., Vol. 68, No. 6, 2003

Facile Aza-Bergman Rearrangement
13C NMR (CDCl3) δ 10.98, 18.46, 36.59, 55.35, 93.53, 110.61,
114.05, 123.27, 129.19, 148.20, 162.53; HRMS m/z 410.1828
(calcd 410.1821, C20H31NO4SSi).
125.99, 127.77, 128.11, 130.81, 131.27, 132.64, 157.66, 158.60;
HRMS m/z 416.2404 (calcd 416.2409, C27H33NOSi).
(Z)-2,5-Dip h en yl-p en t-2-en -4-yn en itr ile (8a ). A solution
of 6a (68 mg, 0.176 mmol) in dry THF (6 mL) was cooled to
-78 °C and then a 1 M solution of TBAF (0.194 mmol) was
added slowly. After being stirred at -78 °C for 5-10 min the
mixture was poured into ice-water (10 mL) and extracted by
CH2Cl2 (2 × 15 mL). The dry organic solution was plugged
through a small column before evaporation. The conversion
1-(2-Meth oxyp h en yl)-3-(tr iisop r op ylsilyl)-p r op yn on e
Mesyla te (5c). Mesylate 5c was prepared from oxime 4c by
a procedure analogous to that described for the preparation
of 5a . The product was purified by flash chromatography (0-
15% CH2Cl2 in hexanes) to afford 90 mg of a yellow oil
consisting of an inseparable mixture of mesylate 5c along with
Beckman rearrangement product 3-(triisopropylsilyl)-propy-
noic acid (2-methoxyphenyl)amide in 6 to 1 ratio, as judged
by 1H NMR. A pure sample of 3-(triisopropylsilyl)-propynoic
acid (2-methoxyphenyl)amide was obtained as a yellow solid
in 72% yield after chromatography from a reaction allowed to
1
of aza-enediyne 7a to (Z)-nitrile 8a was followed by H NMR
and/or UV-vis spectroscopy. After the conversion was com-
plete, the (Z)-nitrile 8a was purified by flash chromatography
on silica gel (0-10% CH2Cl2 in hexanes) to afford 36 mg (89%)
of nitrile 8a as a yellow solid: Rf 0.41 (50% CH2Cl2 in hexanes);
mp 79.3-80.7 °C; 1H NMR (CD2Cl2) δ 6.92 (s, 1H), 7.38-7.50
1
proceed for 6 days. Mesylate 5c: Rf 0.65 (CH2Cl2); H NMR
1
(CDCl3) δ 1.01-1.19 (m, 21H), 3.18 (s, 3H), 3.83 (s, 3H), 6.93-
7.05 (m, 2H), 7.43 (ddd, J ) 9.3, 7.5, 1.8 Hz, 1H), 7.55 (dd, J
) 7.5, 1.5 Hz, 1H); 13C NMR (CDCl3) δ 10.93, 18.38, 36.57,
55.55, 94.39, 110.12, 111.55, 120.45, 120.59, 130.64, 132.43,
(m, 6H), 7.55-7.67 (m, 4H); H NMR (CD3CN) δ 7.1 (s, 1H),
7.41-7.50 (m, 6H), 7.54-7.61 (m, 2H), 7.65-7.71 (m, 2H); 13
C
NMR (CD2Cl2) δ 86.45, 102.23, 116.90, 121.87, 122.43, 124.09,
126.03, 128.99, 129.54, 130.15, 130.52, 132.48, 133.04; HRMS
m/z 230.0972 (calcd 230.0969, C17H11N).
147.87, 157.91; HRMS m/z 410.1825 (calcd 410.1821, C20H31
-
NO4SSi). 3-(Triisopropylsilyl)-propynoic acid (2-methoxyphen-
yl)amide: Rf 0.67 (CH2Cl2); mp 32.0-32.9°C; 1H NMR δ
(CDCl3) 1.00-1.19 (m, 21H), 3.88 (s, 3H), 6.82-6.98 (m, 2H),
7.00-7.08 (m, 1H), 8.05 (s, 1H), 8.29 (dd, J ) 8.7, 2.0, 1H);
13C NMR (CDCl3) δ 10.98, 18.45, 55.76, 89.10, 100.31, 109.86,
120.21, 121.08, 124.32, 127.15, 147.43, 149.81; HRMS m/z
332.2042 (calcd 332.2045, C19H29NO2Si).
(Z)-5-(4-Met h oxyp h en yl)-2-p h en yl-p en t -2-en -4-yn en i-
tr ile (8b). (Z)-Nitrile 8b was prepared from aza-enediyne 6b
by a procedure analogous to that described for the preparation
of 8a . The conversion of imine 6b to nitrile 8b was followed
by 1H NMR and/or UV-vis. The nitrile 8b was purified by
flash chromatography on silica gel (0-5% CH2Cl2 in hexanes)
to afford 4 mg (72%) of (Z)-nitrile 8b as a light yellow solid:
1,4-D ip h e n y l-6-(t r iis o p r o p y ls ily l)-3-a za -3-e n e -1,5-
d iyn e (6a ). To a solution of phenylacetylene (410 mg, 4.02
mmol) in 5 mL of dry ether at -78 °C was added dropwise a
solution of n-butyllithium (2.89 mL, 1.6 M in hexanes, 4.02
mmol) such that the temperature was maintained at -78 °C.
The reaction mixture was stirred for 20 min at the same
temperature. In a separate flask, a suspension of cuperous
iodide (251 mg, 1.32 mmol) in 5 mL of dry ether was cooled to
-40 °C. The phenylacetylide solution was added slowly via
cannula to this suspension. After the addition was complete,
the mixture was stirred at room temperature for 40 min. The
resulting light yellow suspension was added to a solution of
mesylate 5a (500 mg, 1.32 mmol) in 5 mL of dry ether at 0 °C.
The reaction mixture was stirred at 0 °C for 1 h, and then at
room temperature for 2 h. The solvent was evaporated, and
the product was purified by flash chromatography on silica
gel (hexanes) to afford 110 mg (26%) of aza-enediyne 6a as a
yellow oil: Rf 0.56 (40% CH2Cl2 in hexanes); 1H NMR (CDCl3)
δ 1.10-1.25 (m, 21H), 7.28-7.39 (m, 3H), 7.39-7.55 (m, 5H),
8.14 (dd, J ) 7.6, 1.7 Hz, 2H); 13C NMR (CDCl3) δ 11.55, 18.79,
93.75, 100.27, 100.60, 106.79, 124.90, 128.26, 128.52, 128.66,
128.92, 131.77, 132.49, 136.22, 158.96; HRMS m/z 386.2310
(calcd 386.2304, C26H31NSi).
1-P h en yl-4-(4-m eth oxyp h en yl)-6-(tr iisop r op ylsilyl)-3-
a za -3-en e-1,5-d iyn e (6b). Aza-enediyne 6b was prepared
from mesylate 5b by a procedure analogous to that described
for the preparation of 6a . The product was purified by
preparative TLC on silica gel (hexanes) to afford 12 mg (15%)
of aza-enediyne 6b as a yellow oil: Rf 0.35 (30% CH2Cl2 in
hexanes); 1H NMR (CDCl3) δ 1.08-1.25 (m, 21H), 3.85 (s, 3H),
6.92 (d, J ) 9.0 Hz, 2H), 7.25-7.34 (m, 3H), 7.44 (dd, J ) 8.4,
1.5 Hz, 2H), 8.08 (d, J ) 9.0 Hz, 2H); 13C NMR (CDCl3) δ 11.16,
18.65, 55.47, 93.60, 97.86, 100.00, 105.35, 113.94, 124.90,
127.82, 128.14, 128.99, 129.86, 131.38, 131.92, 163.00; HRMS
m/z 416.2407 (calcd 416.2409, C27H33NOSi).
1
Rf 0.35 (40% CH2Cl2 in hexanes); mp 85.5-86.6 °C; H NMR
(CDCl3) δ 3.85 (s, 3H), 6.84 (s, 1H), 6.88 (d, J ) 8.9 Hz, 2H),
7.36-7.44 (m, 3H), 7.52 (d, J ) 8.9 Hz, 2H), 7.57-7.62 (dd, J
) 7.7, 1.5 Hz, 2H); HRMS m/z 260.1070 (calcd 260.1075,
C
18H13NO).
(Z)-5-(2-Met h oxyp h en yl)-2-p h en yl-p en t -2-en -4-yn en i-
tr ile (8c). (Z)-Nitirle 8c was prepared from aza-enediyne 6c
by a procedure analogous to that described for the preparation
of 8a . The conversion of imine 7c to nitrile 8c was followed by
1H NMR and/or UV-vis. The nitrile 8c was purified by flash
chromatography (0-10% CH2Cl2 in hexanes) to afford 10 mg
(91%) of (Z)-nitrile 8c as a light yellow solid: Rf 0.30 (40%
CH2Cl2 in hexanes); mp 57.7-58.6 °C; 1H NMR (CDCl3) δ 3.92
(s, 3H), 6.88-6.98 (m, 3H), 7.33-7.48 (m, 4H), 7.50-7.57 (dd,
J ) 7.8, 1.8 Hz, 1H), 7.58-7.62 (m, 2H); 13C NMR (CDCl3) δ
55.87, 90.44, 99.41, 110.74, 111.21, 120.65, 120.82, 121.56,
122.95, 125.61, 129.14, 129.94, 131.41, 132.50, 134.27, 160.54;
HRMS m/z 260.1077 (calcd 260.1075, C18H13NO).
(E)-2,5-Dip h en yl-3-(tr iisop r op ylsilyl)-p en t-2-en e-4-yn e-
n itr ile (9). To a solution of the aza-enediyne 6a (20 mg, 0.0518
mmol) in chlorobenzene (1.5 mL) was added 1,4-cyclohexadiene
(83.19 mg, 1.036 mmol, 20 equiv). The mixture was heated at
150 °C for 3 days. The solvent was evaporated and the residue
purified by chromatography to afford nitrile 9 (7 mg, 35%) as
a yellow oil: Rf 0.51 (40% CH2Cl2 in hexanes); 1H NMR (CDCl3)
δ 1.10-1.22 (d, J ) 7.5 Hz, 18H), 1.68-1.85 (p, J ) 7.5 Hz,
3H), 7.24-7.34 (m, 5H), 7.37-7.48 (m, 3H), 7.84 (dd, J ) 7.7,
1.9, 2H); 13C NMR (CDCl3) δ 12.27, 18.78, 91.31, 108.25,
120.38, 122.98, 128.12, 128.51, 128.70, 129.23, 129.42, 131.44,
134.10, 136.07, 139.69; HRMS m/z 386.2317 (calcd 386.2304,
C26H31NSi).
Sp ectr oscop ic Kin etics. To a solution of aza-enediyne 6a
or 6b (0.0052 mmol) in dry THF (2 mL) cooled to -78 °C was
added slowly a 1 M solution of tetrabutylammonium floride
in dry THF (0.0059 mmol). The reaction mixture was allowed
to stir at -78 °C for 5-10 min, and the mixture was poured
into ice-water (10 mL) and extracted by CH2Cl2 (2 × 15 mL).
The dry organic solution was chromatographed through a
small plug of silica gel to remove tetrabutylammonium species
before evaporation. The residue was storred at -80 °C and
quickly redissolved in CHCl3 or other solvents (THF, iPrOH,
chlorobenzene, hexanes, CH3CN) and the resulting solution
transferred to a thermostated cuvette at 20, 25, 30, 37, or 45
°C. The conversion of aza-enediyne 7a /7b to (Z)-nitrile 8a /8b
1-P h en yl-4-(2-m eth oxyp h en yl)-6-(tr iisop r op ylsilyl)-3-
a za -3-en e-1,5-d iyn e (6c). Aza-enediyne 6c was prepared from
mesylate 5c by a procedure analogous to that described for
the preparation of 6a . The product was purified by flash
chromatography (0-20% CH2Cl2 in hexanes) to afford 22 mg
(31%) of aza-enediyne 6c as a yellow oil: Rf 0.45 (40% CH2Cl2
1
in hexanes); H NMR (CDCl3) δ 1.05-1.19 (m, 21H), 3.88 (s,
3H), 6.95-7.02 (m, 2H), 7.26-7.34 (m, 3H), 7.36-7.48 (m, 3H),
7.86 (d, J ) 7.6, 1.5 Hz, 1H); 13C NMR (CDCl3) δ 11.19, 18.60,
55.70, 93.21, 100.66, 101.82, 105.05, 111.77, 120.44, 124.86,
J . Org. Chem, Vol. 68, No. 6, 2003 2241

Feng et al.
was followed spectrophotometrically by periodically acquiring
UV absorbance spectra (250-500 nm) over a period of at least
2 half-lives. The first-order rate constants were obtained by
fitting plots of the normalized absorbance at 390 nm (7a ) or
415 nm (7b) versus time to an exponential decay.
Foundation (F-1298) and the Petroleum Research Fund,
administered by the American Chemical Society (AC-
34115).
Su p p or tin g In for m a tion Ava ila ble: Spectroscopic data
1
for compounds 3a ,b,c-9; H NMR and spectroscopic kinetic
data for the aza-Bergman reaction of 7a , 7b, and 7c; and
Cartesian coordinates and energies for (Z)-7a and (E)-7a
(B3LYP/6-31G(d,p) level). This material is available free of
charge via the Internet at http://pubs.acs.org.
Ack n ow led gm en t. We acknowledge Steve Sorey for
his assistance with NMR experiments and Dr. Miguel
Salazar for use of the UV-vis spectrophotometer. This
work was supported by grants from the Robert Welch
J O0267192
2242 J . Org. Chem., Vol. 68, No. 6, 2003