Synthesis and Evaluation of CC-1065 and Duocarmycin Analogs Incorporating the 1,2,3,4,11,11a-Hexahydrocyclopropa[c]naphtho[2,1-b]azepin-6-one (CNA) Alkylation Subunit: Structural Features that Govern Reactivity and Reaction Regioselectivity

Article abstract of DOI:10.1021/jo9707085

The synthesis of 1,2,3,4,11,11a-hexahydrocyclopropa[c]naphtho[2,1-b]azepin-6-one (CNA), a seven-membered C-ring analog of the alkylation subunits of CC-1065 and the duocarmycins, is detailed. The core structure of CNA was prepared through the implementation of an intramolecular Heck reaction for assemblage of the key tricyclic tetrahydronaphtho[2,1-6]azepine skeleton and a final Winstein Ar-3′ spirocyclization for introduction of the reactive cyclopropane. A study of the solvolysis reactivity of N-BOC-CNA revealed that incorporation of the seven-membered fused C-ring system increased the reactivity 4750 × compared to the corresponding five-membered C-ring analog. Solvolysis occurs with S_N2 nucleophilic attack at the more substituted carbon of the activated cyclopropane to afford exclusively the abnormal ring expansion product in a reaction that was shown to proceed with complete inversion of configuration at the reaction center. Single crystal X-ray structure analyses of N-CO₂Me-CNA (29) and CNA (11) and their comparisons with X-ray structures of the corresponding five- and six-membered C-ring analogs revealed the structural origins of the solvolysis regioselectivity and reactivity. The regioselectivity may be attributed to the stereoelectronic alignment of the two available cyclopropane bonds with the cyclohexadienone π-system which for 29 resides with the bond that extends to the more substituted cyclopropyl carbon. The increased reactivity may be due in part to the geometric alignment of the cyclopropane but more significantly is linked to a twist in the N² amide. X-ray analysis provides documentation of the disruption in the vinylogous amide stabilization as measured by a lengthening of the diagnostic C-N bond that accompanies the twist in the χ₁ dihedral angle of the N² amide. As the cross-conjugated vinylogous amide stabilization is diminished, the cyclopropane conjugation, bond lengths, and resulting reactivity increase. The unusual stability of the five-membered C-ring bearing alkylation subunits characteristic of the natural products is intimately linked to the extent of this vinylogous amide conjugation, and the studies support the proposal that catalysis for the DNA alkylation reaction may be due to a DNA binding-induced conformational change in the agents which serves to twist the linking N² amide, disrupting the vinylogous amide stabilization, and activating the agents for S_N2 nucleophilic attack.

Full text of DOI:10.1021/jo9707085

J . Org. Chem. 1997, 62, 5849-5863
5849
Syn th esis a n d Eva lu a tion of CC-1065 a n d Du oca r m ycin An a logs
In cor p or a tin g th e
1,2,3,4,11,11a -Hexa h yd r ocyclop r op a [c]n a p h th o[2,1-b]a zep in -6-on e
(CNA) Alk yla tion Su bu n it: Str u ctu r a l F ea tu r es th a t Gover n
Rea ctivity a n d Rea ction Regioselectivity
Dale L. Boger* and Philip Turnbull
Department of Chemistry, The Scripps Research Institute, 10550 North Torrey Pines Road,
La J olla, California 92037
Received April 21, 1997
The synthesis of 1,2,3,4,11,11a-hexahydrocyclopropa[c]naphtho[2,1-b]azepin-6-one (CNA), a seven-
membered C-ring analog of the alkylation subunits of CC-1065 and the duocarmycins, is detailed.
The core structure of CNA was prepared through the implementation of an intramolecular Heck
reaction for assemblage of the key tricyclic tetrahydronaphtho[2,1-b]azepine skeleton and a final
Winstein Ar-3’ spirocyclization for introduction of the reactive cyclopropane. A study of the solvolysis
reactivity of N-BOC-CNA revealed that incorporation of the seven-membered fused C-ring system
increased the reactivity 4750× compared to the corresponding five-membered C-ring analog.
Solvolysis occurs with SN2 nucleophilic attack at the more substituted carbon of the activated
cyclopropane to afford exclusively the abnormal ring expansion product in a reaction that was shown
to proceed with complete inversion of configuration at the reaction center. Single crystal X-ray
structure analyses of N-CO2Me-CNA (29) and CNA (11) and their comparisons with X-ray structures
of the corresponding five- and six-membered C-ring analogs revealed the structural origins of the
solvolysis regioselectivity and reactivity. The regioselectivity may be attributed to the stereoelec-
tronic alignment of the two available cyclopropane bonds with the cyclohexadienone π-system which
for 29 resides with the bond that extends to the more substituted cyclopropyl carbon. The increased
reactivity may be due in part to the geometric alignment of the cyclopropane but more significantly
is linked to a twist in the N2 amide. X-ray analysis provides documentation of the disruption in
the vinylogous amide stabilization as measured by a lengthening of the diagnostic C-N bond that
accompanies the twist in the ø1 dihedral angle of the N2 amide. As the cross-conjugated vinylogous
amide stabilization is diminished, the cyclopropane conjugation, bond lengths, and resulting
reactivity increase. The unusual stability of the five-membered C-ring bearing alkylation subunits
characteristic of the natural products is intimately linked to the extent of this vinylogous amide
conjugation, and the studies support the proposal that catalysis for the DNA alkylation reaction
may be due to a DNA binding-induced conformational change in the agents which serves to twist
the linking N2 amide, disrupting the vinylogous amide stabilization, and activating the agents for
SN2 nucleophilic attack.
CC-1065 (1)1 and the duocarmycins 2 and 32-4 are
novel natural products that exhibit potent biological
properties (Figure 1). The agents derive their biological
activity through a characteristic DNA alkylation reaction
that has been shown to proceed by a reversible adenine
N3 addition to the least-substituted carbon of the acti-
vated cyclopropane at selected AT-rich sites within the
minor groove.5-14
We have examined a series of functional analogs of the
X Abstract published in Advance ACS Abstracts, August 1, 1997.
(1) Hanka, L. J .; Dietz, A.; Gerpheide, S. A.; Kuentzel, S. L.; Martin,
D. G. J . Antibiot. 1978, 31, 1211. Chidester, C. G.; Krueger, W. C.;
Mizsak, S. A.; Duchamp, D. J .; Martin, D. G. J . Am. Chem. Soc. 1981,
103, 7629.
(2) Yasuzawa, T.; Muroi, K.; Ichimura, M.; Takahashi, I.; Ogawa,
T.; Takahashi, K.; Sano, H.; Saitoh, Y. Chem. Pharm. Bull. 1995, 43,
378. Takahashi, I.; Takahashi, K.; Ichimura, M.; Morimoto, M.; Asano,
K.; Kawamoto, I.; Tomita, F.; Nakano, H. J . Antibiot. 1988, 41, 1915.
Yasuzawa, T.; Iida, T.; Muroi, K.; Ichimura, M.; Takahashi, K.; Sano,
H. Chem. Pharm. Bull. 1988, 36, 3728. Ichimura, M.; Muroi, K.; Asano,
K.; Kawamoto, I.; Tomita, F.; Morimoto, M.; Nakano, H. J . Antibiot.
1988, 41, 1285.
(3) Ichimura, M.; Ogawa, T.; Takahashi, K.; Kobayashi, E.; Kawa-
moto, I.; Yasuzawa, T.; Takahashi, I.; Nakano. H. J . Antibiot. 1990,
43, 1037. Ichimura, M.; Ogawa, T.; Katsumata, S.; Takahashi, K.;
Takahashi, I.; Nakano. H. J . Antibiot. 1991, 44, 1045.
F igu r e 1.
(4) Ohba, K.; Watabe, H.; Sasaki, T.; Takeuchi, Y.; Kodama, Y.;
CC-1065 and duocarmycin alkylation subunits in efforts
to define the fundamental relationships between struc-
ture, chemical reactivity, and the corresponding biological
Nakazawa, T.; Yamamoto, H.; Shomura, T.; Sezaki, M.; Kondo, S. J .
Antibiot. 1988, 41, 1515. Ishii, S.; Nagasawa, M.; Kariya, Y.; Yama-
moto, H.; Inouye, S.; Kondo, S. J . Antibiot. 1989, 42, 1713.
S0022-3263(97)00708-1 CCC: $14.00 © 1997 American Chemical Society

5850 J . Org. Chem., Vol. 62, No. 17, 1997
Boger and Turnbull
properties.5 Replacing the A-ring pyrrole found in 4-6
with a phenyl group allowed for the simplified synthesis15
of a variety of modified alkylation subunits (Figure 2).
Our first example of such a simplified derivative in this
series, CBI (8), demonstrated remarkable biological and
chemical properties.16 In addition to being 4× more
stable than CC-1065 and related CPI-derived agents, it
proved considerably more regioselective in its reaction
with nucleophiles. Derivatization of CBI with the ap-
propriate DNA binding subunits gave agents that are 4×
more potent than (+)-CC-1065, and selected agents
within this series have exhibited efficacious antitumor
(5) Boger, D. L.; J ohnson, D. S. Angew. Chem., Int. Ed. Engl. 1996,
35, 1439. Boger, D. L.; J ohnson, D. S. Proc. Natl. Acad. Sci. U.S.A.
1995, 92, 3642. Boger, D. L. Acc. Chem. Res. 1995, 28, 20. Boger, D. L.
Chemtracts: Org. Chem. 1991, 4, 329. Boger, D. L. Proc. Robert A.
Welch Found. Conf. Chem. Res., XXXV. Chem. Front. Med. 1991, 35,
137. Boger, D. L. In Advances in Heterocyclic Natural Products
Synthesis; Pearson, W. H., Ed.; J AI: Greenwich, CT, 1992; Vol. 2, p 1.
(6) Warpehoski, M. A.; Hurley, L. H. Chem. Res. Toxicol. 1988, 1,
315. Hurley, L. H.; Needham-VanDevanter, D. R. Acc. Chem. Res. 1986,
19, 230. Warpehoski, M. A. In Advances in DNA Sequence Specific
Agents; Hurley, L. H., Ed.; J AI: Greenwich, CT, 1992; Vol. 1, p 217.
Hurley, L. H.; Draves, P. In Molecular Aspects of Anticancer Drug-
DNA Interactions; Neidle, S., Waring, M., Eds.; CRC: Ann Arbor, 1993;
Vol. 1, p 89. Aristoff, P. A. In Advances in Medicinal Chemistry; J AI:
Greenwich, CT, 1993; Vol. 2, p 67. Warpehoski, M. A.; McGovren, P.;
Mitchell, M. A. In Molecular Basis of Specificity in Nucleic Acid-Drug
Interactions; Pullman, B., J ortner, J ., Eds.; Kluwer: Netherlands, 1990;
p 531.
F igu r e 2.
(7) Boger, D. L.; Ishizaki, T.; Zarrinmayeh, H.; Kitos, P. A.; Sun-
tornwat, O. J . Org. Chem. 1990, 55, 4499. Boger, D. L.; Ishizaki, T.;
Zarrinmayeh, H.; Munk, S. A.; Kitos, P. A.; Suntornwat, O. J . Am.
Chem. Soc. 1990, 112, 8961. Boger, D. L.; Munk, S. A.; Zarrinmayeh,
H.; Ishizaki, T.; Haught, J .; Bina, M. Tetrahedron 1991, 47, 2661.
Boger, D. L.; Ishizaki, T.; Zarrinmayeh, H. J . Am. Chem. Soc. 1991,
113, 6645. Boger, D. L.; Yun, W.; Terashima, S.; Fukuda, Y.; Nakatani,
K.; Kitos, P. A.; J in, Q. Bioorg. Med. Chem. Lett. 1992, 2, 759. Boger,
D. L.; Yun, W. J . Am. Chem. Soc. 1993, 115, 9872. Boger, D. L.;
J ohnson, D. S.; Yun, W. J . Am. Chem. Soc. 1994, 116, 1635.
(8) Boger, D. L.; J ohnson, D. S. J . Am. Chem. Soc. 1995, 117, 1443.
Boger, D. L.; J ohnson, D. S.; Yun, W.; Tarby, C. M. Bioorg. Med. Chem.
1994, 2, 115. Boger, D. L.; Zarrinmayeh, H.; Munk, S. A.; Kitos, P. A.;
Suntornwat, O. Proc. Natl. Acad. Sci. U.S.A. 1991, 88, 1431. Boger,
D. L.; Munk, S. A.; Zarrinmayeh, H. J . Am. Chem. Soc. 1991, 113,
3980. Boger, D. L.; Coleman, R. S.; Invergo, B. J .; Sakya, S. M.;
Ishizaki, T.; Munk, S. A.; Zarrinmayeh, H.; Kitos, P. A.; Thompson, S.
C. J . Am. Chem. Soc. 1990, 112, 4623.
(9) Sugiyama, H.; Hosoda, M.; Saito, I.; Asai, A.; Saito, H. Tetrahe-
dron Lett. 1990, 31, 7197. Sugiyama, H.; Ohmori, K.; Chan, K. L.;
Hosoda, M.; Asai, A.; Saito, H.; Saito, I. Tetrahedron Lett. 1993, 34,
2179. Yamamoto, K.; Sugiyama, H.; Kawanishi, S. Biochemistry 1993,
32, 1059. Asai, A.; Nagamura, S.; Saito, H. J . Am. Chem. Soc. 1994,
116, 4171.
(10) Reynolds, V. L.; Molineux, I. J .; Kaplan, D. J .; Swenson, D. H.;
Hurley, L. H. Biochemistry 1985, 24, 6228. Hurley, L. H.; Lee, C.-S.;
McGovren, J . P.; Warpehoski, M. A.; Mitchell, M. A.; Kelly, R. C.;
Aristoff, P. A. Biochemistry 1988, 27, 3886. Hurley, L. H.; Warpehoski,
M. A.; Lee, C.-S.; McGovren, J . P.; Scahill, T. A.; Kelly, R. C.; Mitchell,
M. A.; Wicnienski, N. A.; Gebhard, I.; J ohnson, P. D.; Bradford, V. S.
J . Am. Chem. Soc. 1990, 112, 4633.
activity.17 The ring expansion of the fused five-membered
C-ring found in CBI to a six-membered ring with CBQ
(9) resulted in a substantial increase in solvolytic reactiv-
ity, a loss of the characteristic solvolysis regioselectivity,
and a corresponding loss in cytotoxic potency.18 Herein,
we detail the extension of these studies to the synthesis
and examination of 1,2,3,4,11,11a-hexahydrocyclopro-
pa[c]naphtho[2,1-b]azepin-6-one (CNA, 11), a fused
seven-membered C-ring analog, and its comparison with
prior agents.
Syn th esis of N-BOC-CNA (10) a n d CNA (11). At-
tempts to extend the radical cyclization methodology
employed for the five- and six-membered C-ring analogs
in this class were unsuccessful.16-19 Reaction of the
appropriately functionalized naphthalene20 with Bu3SnH-
AIBN (R ) H, + TEMPO) resulted in simple reduction
(eq 1). The metal hydride reduction of the resulting aryl
radical proved faster than the required 7-exo-trig radical
cyclization even with substrates bearing an activated
acceptor alkene.
(11) Boger, D. L.; Hertzog, D. L.; Bollinger, B.; J ohnson, D. S.; Cai,
H.; Goldberg, J .; Turnbull, P. J . Am. Chem. Soc. 1997, 119, 4977.
(12) Boger, D. L.; Bollinger, B.; Hertzog, D. L.; J ohnson, D. S.; Cai,
H.; Mesini, P.; Garbaccio, R. M.; J in, Q.; Kitos, P. A. J . Am. Chem.
Soc. 1997, 119, 4987.
(13) Boger, D. L.; Boyce, C. W.; J ohnson, D. S. Bioorg. Med. Chem.
Lett. 1997, 7, 233.
(14) Boger, D. L.; J ohnson, D. S.; Wrasidlo, W. Bioorg. Med. Chem.
Lett. 1994, 4, 631.
(15) Boger, D. L.; Boyce, C. W.; Garbaccio, R. M.; Goldberg, J . Chem.
Rev. 1997, 97, 263.
(16) Boger, D. L.; Ishizaki, T.; Wysocki, R. J .; Munk, S. A.; Kitos, P.
A.; Suntornwat, O. J . Am. Chem. Soc. 1989, 111, 6461. Boger, D. L.;
Ishizaki, T.; Kitos, P. A.; Suntornwat, O. J . Org. Chem. 1990, 55, 5823.
Boger, D. L.; Ishizaki, T. Tetrahedron Lett. 1990, 31, 793. Boger, D.
L.; Ishizaki, T.; Zarrinmayeh, H.; Kitos, P. A.; Suntornwat, O. Bioorg.
Med. Chem. Lett. 1991, 1, 55. Boger, D. L.; Ishizaki, T.; Sakya, S.;
Munk, S. A.; Kitos, P. A.; J in, Q.; Besterman, J . M. Bioorg. Med. Chem.
Lett. 1991, 1, 115. Boger, D. L.; Munk, S. A. J . Am. Chem. Soc. 1992,
114, 5487. Boger, D. L.; Yun, W. J . Am. Chem. Soc. 1994, 116, 7996.
Boger, D. L.; Yun, W.; Han, N.; J ohnson, D. S. Bioorg. Med. Chem.
1995, 3, 611. Boger, D. L.; Yun, W.; Cai, H.; Han, N. Bioorg. Med.
Chem. 1995, 3, 761.
(17) Boger, D. L.; Yun, W.; Han, N. Bioorg. Med. Chem. 1995, 3,
1429.
(18) Boger, D. L.; Mesini, P.; Tarby, C. M. J . Am. Chem. Soc. 1994,
116, 6461. Boger, D. L.; Mesini, P. J . Am. Chem. Soc. 1994, 116, 11335.
Boger, D. L.; Mesini, P. J . Am. Chem. Soc. 1995, 117, 11647.
(19) Boger, D. L.; Yun, W.; Teegarden, B. R. J . Org. Chem. 1992,
57, 2873. Boger, D. L.; McKie, J . A. J . Org. Chem. 1995, 60, 1271.
(20) Full experimental details are provided in Supporting Informa-
tion.

Synthesis and Evaluation of CC-1065 and Duocarmycin Analogs
J . Org. Chem., Vol. 62, No. 17, 1997 5851
Sch em e 1
chloride 18 upon treatment with Ph3P-CCl4 proceeded
in modest yields (ca. 50%, eq 3). Much higher conver-
sions were realized by treatment of 15 with mesyl
chloride in the presence of Et3N to provide 16 in 89%.
Removal of the benzyl group through transfer hydro-
genolysis23 (4 wt equiv of 10% Pd-C, aqueous HCO2NH4-
THF, 25 °C, 2 h, 89%) afforded 17 and set the stage for
the final spirocyclization.
The successful approach to the preparation of the CNA
nucleus rested on the implementation of an alternative
intramolecular Heck reaction.21 The starting material
for both the Heck reaction and the radical cyclization
attempts was prepared by N-alkylation of N-(tert-butyl-
oxycarbonyl)-4-(benzyloxy)-1-iodo-2-naphthylamine (12),
readily available in three steps (60% overall) from 1,3-
dihydroxynaphthalene,19 with the mesylate of 4-penten-
1-ol to provide 13 (Scheme 1). A full 2 mol equiv of
benzyltributylammonium chloride was required to pro-
vide high yields of the alkylated product using phase
transfer conditions. An intramolecular Heck reaction
conducted with 3 mol % of Pd[P(Ph3)]4 as catalyst
furnished the desired cyclization product 14 in 89%.
Careful removal of O2 from the reaction mixture provided
reproducible high yields, and no palladium hydride
induced isomerization of 14 to give the endocyclic olefin
was detectable by 1H NMR of the crude reaction mixture.
Winstein Ar-3′ spirocyclization was effected by treat-
ment of 17 with DBU (3 equiv, CH3CN) and smoothly
provided 10 (87%). Due to its unusual reactivity, careful
chromatographic conditions were required and pretreat-
ment of the chromatographic support (SiO2) with Et3N
resulted in much higher yields (i.e., 87% versus 40-50%).
Treatment of 17 with 3.9 M HCl-EtOAc removed the
BOC group to afford the unstable amine hydrochloride
which was taken directly into the subsequent cyclization
step. Addition of 10 equiv of DBU as a dilute solution in
CH3CN resulted in good conversion to 11 (77%). Con-
ventional chromatography provided the relatively stable
CNA which gave crystals suitable for X-ray analysis24
(CH3CN-CH3OH).
Syn t h esis of N-CO2Me-CNA (29) a n d CNA-TMI
(30). Efforts to acylate the amine hydrochloride resulting
from BOC deprotection of 17 at this late stage in the
synthesis were unsuccessful. The only azepine amenable
to N-acylation was the free amine derived from 14 which
bore the exocyclic olefin. Incorporation of the sp2-
hybridized carbon presumably alters the conformation of
the fused seven-membered ring in a manner that allows
N-acylation. Synthesis of the amine hydrochloride 20
through treatment with 3.9 N HCl-EtOAc (25 °C, 40
min) provided a relatively well-behaved solid amine salt
that served as a short term storage intermediate for
further acylations (Scheme 2). Liberation of the free
amine with NaHCO3 followed by treatment with methyl
chloroformate provided 21 in 94% yield. High yields
(80%) were also realized for EDCI coupling of 20 and
Similar observations were made with the substrate (R
) CO2CH3) bearing an activated electron-deficient ac-
ceptor alkene, but the reaction proved unsuccessful with
the substrate (R ) OTHP) bearing an electron-rich
acceptor alkene (eq 2).20 In principle, both successful
cyclization products can be converted into appropriately
C-1 substituted and functionalized tetrahydronaph-
tho[2,1-b]azepines for use in the preparation of 10 and
11. In our efforts, 14 was employed since its conversion
to 10 and 11 would follow protocols we introduced with
the synthesis of CC-1065 and consequently were familiar
with.16,22
Hydroboration followed by oxidative workup converted
the exocyclic olefin into the desired alcohol 15 (82%).16,22
Efforts to convert this alcohol to the corresponding
(23) Ram, S.; Ehrenkaufer, R. E. Synthesis 1988, 91. Bieg, T.; Szeja,
W. Synthesis 1985, 76.
(21) Sakamoto, T.; Kondo, Y.; Uchiyama, M.; Yamanaka, H. J .
Chem. Soc., Perkin Trans. 1 1993, 1941. Abelman, M. M.; Oh, T.;
Overman, L. E. J . Org. Chem. 1987, 52, 4130.
(24) The atomic coordinates for this structure have been deposited
with the Cambridge Crystallographic Data Centre and may be obtained
upon request from the Director, Cambridge Crystallographic Data
Centre, 12 Union Road, Cambridge, CB2 1EZ, UK.
(22) Boger, D. L.; Coleman, R. S. J . Am. Chem. Soc. 1988, 110, 4796.

5852 J . Org. Chem., Vol. 62, No. 17, 1997
Boger and Turnbull
Sch em e 2
key to understanding of the structural features that
dictate the properties of the duocarmycin and CC-1065
alkylation subunits.5 Comparison of the rate of solvolysis
allows for the identification of the functional features that
contribute to the stabilization or activation of the cyclo-
propane while studies regarding the regioselectivity of
the acid-catalyzed ring opening have proven key to
understanding the mechanism of the DNA alkylation
reaction. To date, the preferred nucleophilic addition
sites appear to be dictated by the relative stereoelectronic
alignment of the scissile cyclopropyl bonds with the
cyclohexadienone π-system but may be influenced by the
relative reactivity. Moreover, both SN2 and SN1 mech-
anisms have been advanced to account for the generation
of the abnormal ring expansion solvolysis products.11,18,25-27
In instances of mixed reaction regioselectivity, the clean
generation of single DNA alkylation products derived
from adenine N3 addition exclusively to the least-
substituted cyclopropane carbon indicates that additional
subtle features inherent in the reaction contribute to its
regioselectivity.11,18,26
The solvolysis reactivities of N-BOC-CNA (10) and
CNA (11) were followed spectrophotometrically by UV
at both pH 3 (50% CH3OH-buffer, buffer ) 4:1:20 v/v/v
0.1 M citric acid, 0.2 M Na2HPO4, H2O) and pH 7 (50%
CH3OH-H2O), Figure 3. The reactivity of both 10 (t1/2
)
1.7 min at pH 3) and 11 (t1/2 ) 37 min at pH 3) proved
remarkable. Consistent with past observations, CNA
(11) was considerably more stable than N-BOC-CNA (10).
Unlike prior studies, both were sufficiently reactive to
undergo rapid solvolysis even at pH 7. Comparison of
the solvolytic reactivity of N-BOC-CNA with representa-
tive prior analogs reveals a reactivity that is between that
of N-BOC-CBQ and the exceptionally reactive N-BOC-
CI (Table 2). Particularly informative are comparisons
of the benzo series of agents that differ only in the C-ring
size (i.e., 8-11). Solvolysis reactivity smoothly increases
through the series CBI < CBQ < CNA with both the
N-BOC derivatives and the free NH derivatives. The
reactivity of both N-BOC-CNA and CNA proved to be
4750 and 1500× more reactive than the corresponding
CBI derivatives. Moreover, N-BOC-CNA exhibited a
first-order rate constant for the solvolysis at pH 7 which
is of the same order of magnitude as the characteristically
rapid DNA alkylation reactions of 1-327 without deliber-
ate added catalysis. The agents containing the free NH,
and thus the greatest degree of vinylogous amide conju-
gation, were 10-40× more stable to solvolysis than the
corresponding N-BOC derivatives.28
Ta ble 1. Resolu tion of 16, 19, a n d 24a
agent
% i-PrOH-hexane, flow rate
tR (min)
R
16
19
24
10%, 5 mL/min
6%, 6 mL/min
50%, 8 mL/min
60%, 8 mL/min
70%, 8 mL/min
38.0, 45.0
22.6, 26.8
14.4, 16.3
16.3, 18.5
20.9, 24.0
1.18
1.19
1.13
1.13
1.15
a
10 µm, 2 × 25 cm semipreparative Chiracel OD column.
5,6,7-trimethoxyindole-2-carboxylic acid to form the ad-
vanced analog 22. Exposure of the acylated products to
the same sequence of synthetic transformations detailed
for 10 and 11 gave good to excellent yields of the desired
products 23-28. Unlike N-BOC-CNA (10), N-CO2Me-
CNA (29) proved to be a crystalline solid and recrystal-
lization from 10% EtOAc-hexanes gave X-ray quality
crystals.24 Spirocyclization of TMI derivative 28 gave a
lower yield of ring closed product 30 (48%) due to its high
reactivity and more challenging chromatography require-
ments for purification.
Resolu tion . In order to obtain the separate enanti-
omers of both 10 and 28, various synthetic intermediates
were subjected to direct chromatographic resolution on
a semipreparative chiracel OD column (Table 1). The
intermediates 16, 19, and 24 proved to have the greatest
separation and were resolved (>99.9% ee) in sufficient
quantities to be carried on to the respective final prod-
ucts. The assignment of the absolute configuration for
(+)-10 was based initially on the relative cytotoxic
potencies of (+)- and ent-(-)-N-BOC-CNA, with the
former exhibiting more potent activity which is consistent
with observations made for 1-9. Ultimately, the absolute
configuration for (+)-10 was confirmed and unambigu-
ously established by X-ray structure analysis24 of the
resolved seco-derivative (-)-(1R)-19 bearing a heavy atom
and subsequent conversion to (+)-10.
Ch em ica l Solvolysis: Regioselectivity a n d Mech -
a n ism . Treatment of N-BOC-CNA (10) with 0.12 equiv
of CF3SO3H in a mixture of H2O and THF resulted in
clean solvolysis (>96%) to provide a single product
(Scheme 3). No N-BOC deprotection or olefin formation
was observed. Comparison of this single product to 33
that would result from H2O attack at the less-substituted
cyclopropyl carbon indicated that ring expansion had
(25) Boger, D. L.; Goldberg, J .; McKie, J . A. Bioorg. Med. Chem. Lett.
1996, 6, 1955.
(26) Boger, D. L.; McKie, J . A.; Nishi, T.; Ogiku, T. J . Am. Chem.
Soc. 1997, 119, 311.
(27) Warpehoski, M. A.; Harper, D. E. J . Am. Chem. Soc. 1994, 116,
7573. Warpehoski, M. A.; Harper, D. E. J . Am. Chem. Soc. 1995, 117,
2951.
(28) In previous studies this rate difference was attributed to N-
versus O-protonation, see: Warpehoski, M. A.; Gebhard, I.; Kelly, R.
C.; Krueger, W. C.; Li, L. H.; McGovern, J . P.; Prairie, M. D.;
Wicnienski, N.; Wierenga, W. J . Med. Chem. 1988, 31, 590.
Ch em ica l Solvolysis: Rea ctivity. Both the study
of the rate of acid-catalyzed solvolysis and the regio-
selectivity of cyclopropyl ring opening have proven to be

Synthesis and Evaluation of CC-1065 and Duocarmycin Analogs
J . Org. Chem., Vol. 62, No. 17, 1997 5853
Sch em e 3
expansion derived from cleavage of the internal cyclo-
propane bond. Similarly, treatment of 10 with 1.5 equiv
of HCl in THF (-78 °C, 15 min) cleanly provided 34 (>20:
1, 99%) and not 19. This observation is especially
significant. In some,11,27 but not all,18,26 of the prior
instances of mixed solvolysis regioselectivity, the addition
of HCl still cleanly provided the typical addition product
derived from attack at the least-substituted carbon rather
than the ring expansion addition observed with the
generation of 34. The clean addition of HCl to 10 to
provide the ring expansion product 34 illustrates that
the stereoelectronic control of the addition directs even
the larger nucleophiles to the more-hindered cyclopropyl
carbon.
F igu r e 3. Solvolysis study (UV) spectra of N-BOC-CNA (10,
top) and CNA (11, bottom) in 50% CH3OH-aqueous buffer (pH
3.0, 4:1:20 (v/v/v) 0.1 M citric acid, 0.2 M NaH2PO4, and H2O,
respectively). The spectra were recorded at regular intervals,
and only a few are shown for clarity. Top: 1, 0 min; 2, 0.5
min; 3, 1.0 min; 4, 1.5 min; 5, 2.5 min; 6, 3.5 min; 7, 8.5 min.
Bottom: 1, 0 min; 2, 10 min; 3, 18 min; 4, 30 min; 5, 46 min;
6, 64 min; 7, 90 min; 8, 152 min; 9, 302 min.
Ta ble 2
t1/2 (h,
pH 3)
k (s-1
pH 3}
,
t1/2 (h,
pH 7)
k (s-1
pH 7)
,
agent
In an effort that established the mechanistic course of
the reaction, both racemic and optically active (+)-10
were subjected to acid-catalyzed solvolysis (0.12 equiv of
CF3SO3H, H2O-THF, 25 °C, 5 min) to provide 31.
Resolution on a Diacel Chiracel OG analytical HPLC
column separated both enantiomers of the reaction
product. The product derived from optically active (+)-
10 provided one enantiomer of the ring expanded product
(Figure 4). The generation of a single enantiomer of 31
establishes that the cleavage of the internal cyclopropane
bond does not proceed with generation of a free carboca-
tion (SN1) but instead occurs with clean inversion of the
reaction center stereochemistry in an SN2 ring-opening
reaction. These results are in agreement with the
similarly unambiguous results of solvolysis studies of
N-BOC-CBQ,18 N-BOC-DSA,11 and N-BOC-DA26 which
also afford a single enantiomer of the minor ring expan-
sion products but contrasts the studies with a CPI-
derivative where racemization is reported to occur.27 This
growing set of observations and especially that of the
exceptionally reactive N-BOC-CNA detailed herein sug-
gest the latter CPI studies should be reexamined, em-
N-BOC Derivatives
177
37
N-BOC-DSA (5)
N-BOC-CPI (4)
N-BOC-DA (6)
N-BOC-CBI (8)
N-BOC-CBQ (9)
N-BOC-CNA (10)
N-BOC-Cl (7)
1.08 × 10-6 stable stable
5.26 × 10-6 stable stable
1.75 × 10-5 stable stable
1.45 × 10-6 stable stable
11
133
2.1
9.07 × 10-5 544
3.54 × 10-7
0.03 6.90 × 10-3 2.1
9.16 × 10-5
0.01 1.98 × 10-2 5.2
3.67 × 10-5
N-H Derivatives
CBI
CBQ
CNA (11)
930
91.2
2.07 × 10-7 stable stable
2.11 × 10-6 stable stable
0.62 3.10 × 10-4 563
3.42 × 10-7
occurred exclusively to afford 31. None of the typical
seven-membered ring solvolysis product 33 could be
detected in the crude reaction mixture by H NMR or
HPLC. Treatment of 10 with CH3OH under similar
conditions gave an analogous result, providing excellent
conversion exclusively to the abnormal ring expansion
product 32. Although minor amounts of the ring expan-
sion solvolysis products have been observed with related
systems, the solvolysis of N-BOC-CNA represents the
first observation of either predominant or exclusive ring
1

5854 J . Org. Chem., Vol. 62, No. 17, 1997
Boger and Turnbull
F igu r e 5.
carbon is nearly perpendicular to the plane of the
cyclohexadienone and consequently overlaps well with
developing π-system of the phenol solvolysis product
(Figure 6). In contrast, the cyclopropane bond extending
to the more-substituted carbon is less effectively aligned
and nearly orthogonal to the π-system. Reflecting this
relative alignment, the cleaved C8b-C9 bond is the
longest (1.544 versus 1.521 Å) and weakest of these two
bonds. Thus, the cyclopropane cleavage occurs under
stereoelectronic control with preferential addition of the
nucleophile at the least-substituted carbon (>20:1). This
stereoelectronic control overrides any intrinsic electronic
preference for ring expansion ring opening with partial
positive charge development on the more substituted
cyclopropane carbon but may also benefit from the
characteristic preference for SN2 nucleophilic attack to
occur at the least-substituted carbon.
F igu r e 4. Chiral phase HPLC separation of the product 31
of acid-catalyzed addition of H2O to racmic 10 (top) and (+)-
10 (bottom). Chiracel OG HPLC column (10 µm, 0.46 × 25 cm),
5% i-PrOH-hexane, 1 mL/min.
ploying a more definitive basis for establishing the optical
purity of the ring expansion solvolysis product.
In contrast, the previously obtained X-ray structure18
for N-BOC-CBQ (9) containing the six-membered C-ring
revealed that the cyclopropane is ideally conjugated and
aligned with the π-system and the two potential cyclo-
propane cleavage bonds are perfectly bisected by the
cyclohexadienone π-system (Figure 6). Such a bond
orientation is consistent with the loss of solvolytic regio-
selectivity for N-BOC-CBQ which shows only a slight
preference for nucleophilic addition at the less-substi-
tuted cyclopropane carbon (3:2). Given an inherent
preference for partial positive charge delocalization onto
a secondary versus primary center, one might have
anticipated preferential cleavage of the C9b-C10a bond.
However, the bond extending to the less-substituted
cyclopropyl carbon is weaker as judged by its longer bond
length (1.543 Å versus 1.528 Å) suggesting that any
inherent preference for cleavage of the C9b-C10a bond
is offset by this lower bond strength of the C9b-C10
bond. In addition, SN2 nucleophilic attack at the more-
substituted tertiary center is sterically disfavored. Given
that the ring expansion solvolysis occurs with clean
inversion of stereochemistry, the developing torsional
strain that accompanies nucleophilic addition with ring
expansion may be especially significant and also contrib-
uting to the small regioselectivity preference.
X-r a y St r u ct u r a l Cor r ela t ion s w it h Solvolysis
Regioselectivity. The single-crystal X-ray structure
determination of N-CO2Me-CNA (29) and CNA (11) was
conducted in expectations of providing structural insights
into its reversed solvolysis regioselectivity and extraor-
dinary reactivity, Figure 5.24 To insure the comparisons
were accurately made between the appropriately acylated
and nonacylated derivatives, we have also extended the
efforts to secure a total of eight X-ray crystal structures24
including both a simple N-acyl and the free NH derivative
of CNA, CBQ, and CBI, as well as two seco precursors.
Comparison of the X-ray structures allows for a con-
firmation of the conclusions that have been drawn
between the relative stereoelectronic alignment of the
reacting cyclopropane bonds and solvolysis regioselectiv-
ity.11,18,26,29,30 The newly obtained X-ray structure for
N-CO2Me-CBI containing the five-membered C-ring, like
that of CBI itself,16 reveals that the bent orbital of the
cyclopropane bond extending to the least substituted
(29) Boger, D. L.; McKie, J . A.; Cai, H.; Cacciari, B.; Baraldi, P. G.
J . Org. Chem. 1996, 61, 1710. Boger, D. L.; Han, N.; Tarby, C. M.;
Boyce, C. W.; Cai, H.; J in, Q.; Kitos, P. A. J . Org. Chem. 1996, 61,
4894.
(30) Boger, D. L.; J enkins, T. J . J . Am. Chem. Soc. 1996, 118, 8860.

Synthesis and Evaluation of CC-1065 and Duocarmycin Analogs
J . Org. Chem., Vol. 62, No. 17, 1997 5855
tuted center. Diagnostic of the regioselectivity of nu-
cleophilic addition, the cleaved C10b-C11a bond is longer
(1.565 Å versus 1.525 Å) and weaker than the C10b-
C11 bond extending to the less-substituted cyclopropane
carbon. Like the geometric alignment of N-BOC-CBQ
but unlike N-BOC-CBI, N-CO2Me-CNA possesses a near
perfect geometric orientation with respect to the plane
bisecting the cyclohexadienone, deviating little from a
perfect backside alignment (Figure 6).
Thus, consistent with comparisons that can be made
from all such structural studies to date,11,18,26,29,30 the
solvolysis reaction regioselectivity accurately reflects the
relative stereoelectronic alignment of the two available
cyclopropane bonds. This is beautifully illustrated with
the clean, smooth, and complete reversal of the reaction
regioselectivity as one progresses through the series
N-BOC-CBI (>20:1), N-BOC-CBQ (3:2), and N-BOC-CNA
(<1:20).
X-r a y St r u ct u r a l Cor r ela t ion s w it h Solvolysis
Rea ctivity. The solvolysis reactivity increases that
occur in the series CNA > CBQ > CBI (Table 1) are the
consequence of a previously unappreciated structural
feature of the CC-1065 and duocarmycin alkylation
subunits. Although the alkylation subunit vinylogous
amide has been recognized as a structural feature
contributing to its unusual stability, the extent of this
stabilization has not been established nor has the struc-
tural, chemical, and biological consequences of its pres-
ence or disruption been defined.5 We recently highlighted
that one of the most prominent structural features of the
alkylation subunits that is directly observable in their
X-ray structures is the alternating shortened and length-
ened bonds within the vinylogous amide with the most
diagnostic feature being the shortened C-N bond length.31
This provides the opportunity to establish by X-ray the
relative extent of the vinylogous amide conjugation that
accompanies structural modifications within the alkyla-
tion subunit and, ultimately, the ability to correlate this
with the properties and chemical reactivity of the agents.
Such comparisons within the CBI, CBQ, and CNA series
proved especially revealing.
The direct comparisons of the full set of X-ray struc-
tures in the series provide a superb assessment of the
extent of vinylogous amide conjugation and ultimately,
the relative reactivity of the agents. First, N-acylation
(e.g., N-CO2Me-CBI versus CBI) reduces the vinylogous
amide conjugation, lengthens bond c, and results in a
substantial increase in reactivity. This is observed with
each of the three sets of agents (Figure 7). Typical C-N
bond lengths for a fully conjugated vinylogous amide are
1.312-1.337 Å.32 Thus, that of CBI (bond c, 1.337 Å) is
diagnostic of a fully engaged vinylogous amide while that
of N-CO2Me-CBI (1.390 Å) is substantially diminished.
Accompanying this reduction in the vinylogous amide
conjugation that results from N-acylation is an increase
in length of the reacting cyclopropane bonds and a
readjustment of the cyclopropane alignment to a more
idealized conjugation with the cyclohexadienone π-sys-
tem. This illustrates that both the cyclopropane conjuga-
tion and its inherent reactivity increase as the cross-
conjugated vinylogous amide π-overlap is diminished.
One of the more important conclusions that can be drawn
from these correlations is that the pronounced solvolysis
F igu r e 6. Stick models of the side view and 90° rotation view
of the activated cyclopropane of N-CO2Me-CBI, N-BOC-CBQ,
and N-CO2Me-CNA illustrating data taken from the X-ray
crystal structures and highlighting the stereoelectronic and
geometric alignment of the cyclopropane with the cyclohexa-
dienone π-system.
Inspection of the X-ray crystal structure of N-CO2Me-
CNA reveals that the cyclopropane possesses alignment
features of both N-CO2Me-CBI and N-BOC-CBQ (Figure
6). Like N-CO2Me-CBI, the cyclopropane is nonideally
conjugated with the π-system but rises above rather than
dips below the plane of the cyclohexadienone such that
now the bond to the more-substituted cyclopropane
carbon enjoys the better stereoelectronic alignment.
Thus, the smooth reversal of solvolysis regioselectivity
with N-BOC-CNA may be attributed to the relative
stereoelectronic alignment of the breaking C10b-C11a
bond combined with any intrinsic electronic stabilization
inherent in its greater substitution. Notably, this com-
bination of stereoelectronic and electronic features is
sufficient to overcome the destabilizing steric interactions
that must accompany SN2 addition at the more substi-
(31) Boger, D. L.; Garbaccio, R. M. Bioorg. Med. Chem. 1997, 5, 263.
(32) The Chemistry of Enamines; Rappaport, Z., Ed.; Wiley: New
York, 1994; Part 1.

5856 J . Org. Chem., Vol. 62, No. 17, 1997
Boger and Turnbull
F igu r e 8. Stick models of the vinylogous amide profile
illustrating data taken from X-ray crystal structures and
highlighting the ø1 twist angle of the vinylogous amide.
(Note: corresponding dihedral angle incorporating N-H is the
same assuming sp2 hybridization.)
in the ø1 dihedral angle (Figures 7 and 8). Notably, the
ø1 dihedral angle of N-CO2Me-CNA is so large that the
acyl group is nearly perpendicular to the plane of the
cyclohexadienone π-system, and the agent benefits from
little, if any, vinylogous amide conjugation. This is
reflected in its bond c length of 1.428 Å. Throughout this
series, the ø2 dihedral angle is maintained at ca. 0°,
illustrating the preferential maintenance of the N car-
bamate conjugation versus that of the vinylogous amide.
However, N-BOC-CBQ exhibits a slightly larger ø2 dihe-
dral angle, suggesting some carbamate distortion in
efforts to maintain the vinylogous amide conjugation.
These observations have important implications on the
source of catalysis for the DNA alkylation reaction which
is discussed in the following section.
F igu r e 7.
stability of the free NH derivatives is due to this cross-
conjugated and fully engaged vinylogous amide. Despite
their increased basicity which would facilitate C-4 car-
bonyl protonation, they are much less reactive than the
corresponding N-acyl derivatives toward acid-catalyzed
nucleophilic addition reactions.28
More importantly, these same trends are observed
within the N-acyl series. As one moves across the series
of N-CO2Me-CBI, N-BOC-CBQ, and N-CO2Me-CNA, the
length of bond c increases (1.390, 1.415, and 1.428 Å)
diagnostic of the loss of the cross-conjugated vinylogous
amide (Figure 7). Correspondingly, the length, conjuga-
tion, and reactivity of the scissile cyclopropane bonds
increase tracking with the relative reactivity of the
agents. Accompanying these changes and responsible for
this loss of vinylogous amide conjugation is an increase
An additional and more subtle structural feature that
contributes to reactivity is seen in the structural com-
parisons of the NH derivatives. This series does not

Synthesis and Evaluation of CC-1065 and Duocarmycin Analogs
J . Org. Chem., Vol. 62, No. 17, 1997 5857
to the perfect geometrical (backside) alignment of the
CBQ cyclopropane not accessible to CBI.18 Presumably,
this increases the CBQ cyclopropane conjugation, length-
ens the scissile cyclopropane bonds, and results in an
increased reactivity. In contrast to CBI and CBQ which
have similar c bond lengths but different cyclopropyl
alignments, CBQ and CNA have similar cyclopropyl
alignments but substantially different ø1 dihedral angles
and different c bond lengths of 1.336 versus 1.376 Å,
respectively. This difference, diagnostic of the extent of
vinylogous amide conjugation, is accompanied by a large
increase in the CNA scissile cyclopropane bond lengths
indicative of a much greater degree of conjugation and
accounts for over a hundred-fold difference in reactivity.
Thus, the degree of cyclopropane conjugation and its
resulting reactivity is not only related to its geometrical
alignment but also to the extent of the cross-conjugated
vinylogous amide stabilization.
Thus, the geometrical constraints imposed by the fused
five-membered C-ring found in the natural products
dictate the regioselectivity of cleavage of the cyclopropane
ring. More importantly, the agents display a beautiful
interplay between the cross-conjugated stability provided
by the vinylogous amide and the extent of cyclopropane
conjugation that is central to their functional reactivity.
Structural perturbations that diminish the vinylogous
amide conjugation increase the cyclopropyl conjugation
and its inherent reactivity. One particularly important
structural perturbation is the ø1 dihedral angle of the
linking amide of the N-acyl derivatives of the alkylation
subunits. As the ø1 dihedral angle is increased, the
nitrogen lone pair remains conjugated with the acyl
group carbonyl disrupting the vinylogous amide conjuga-
tion resulting in very substantial increases in the cyclo-
propane reactivity. One fundamental insight gained
from these comparisons is the extent of the vinylogous
amide conjugation and the contribution it makes to the
unusual stability of the CBI and DSA alkylation sub-
units.
Ca ta lysis of th e DNA Alk yla tion Rea ction . The
remarkable chemical stability of 1-3 and the acid-
catalysis requirement for addition of typical nucleophiles
have led to the assumption that the DNA alkylation must
also be an acid-catalyzed reaction. Although efforts have
gone into supporting the extent and role of this acid
catalysis,27 it remains largely undocumented for the DNA
alkylation reaction. At pH 7.4, the DNA phosphate
backbone is fully ionized (0.0001-0.00004% protonated).
Consequently, it is unlikely that the catalysis is derived
from a phosphate backbone delivery of a proton to the
C4 carbonyl as advanced in related efforts.6 Although
increases in the local hydronium ion concentrations
surrounding “acidic domains” of DNA have been invoked
to explain DNA-mediated acid-catalysis,33 nucleotide
reactivity,33 and extrapolated in studies with 1 to alky-
lation site catalysis,27 the remarkable stability of 1-3
even at pH 5 suggests that it is unlikely to be the source
of catalysis. Consistent with this, the rate of the DNA
alkylation reaction exhibits only a very modest pH
dependence below pH 7 and essentially no dependence
in the more relevant pH 7-8 range.11,13
F igu r e 9. Stick models of the side view and 90° rotation view
of the activated cyclopropane of CBI, CBQ, and CNA illustrat-
ing data taken from the X-ray crystal structures and high-
lighting the stereoeloectronic and geometric alignment of the
cyclopropane with the cyclohexadienone π-system.
exhibit a completely smooth correlation between the ø1
torsion angle, the length of bond c, and reactivity (Figure
7). CBI boasts a torsion angle of 15.7° but is significantly
less reactive than CBQ which has a smaller torsion angle
of 6.9°. In the case of the NH derivatives, the X-ray ø1
torsional angles reveal nothing about the pyramidaliza-
tion of the nitrogen and the resulting orientation of its
lone pair. Therefore, it is difficult to assess the relative
extent of π-overlap simply by measuring the ø1 angle.
However, the identical C-N bond lengths for CBI and
CBQ suggest that both benefit from comparable vinylo-
gous amide conjugation. The structural difference that
accounts for their 10-fold different reactivities lies in the
lengths of the scissile cyclopropyl bonds (Figure 9). CBI
has bond lengths of 1.508 Å and 1.532 Å for the C8b-
C9a and C8b-C9 bonds, respectively, while those of CBQ
are 1.525 Å and 1.539 Å. In turn, this may be attributed
In conjunction with the results of these studies which
document the lack of pH dependence on the rate of DNA
(33) J ayaram, B.; Sharp, K. A.; Honig, B. Biopolymers 1989, 28, 975.
Lamm, G.; Pack, G. R. Proc. Natl. Acad. Sci. U.S.A. 1990, 87, 9033.
Lamm, G.; Wong, L.; Pack, G. R. J . Am. Chem. Soc. 1996, 118, 3325.

5858 J . Org. Chem., Vol. 62, No. 17, 1997
Boger and Turnbull
Ta ble 3. In Vitr o Cytotoxic Acitivity
alkylation and related studies that demonstrated that a
rigid N2 amide substituent is required for catalysis,11,12
an alternative source of catalysis became apparent. The
studies detailed herein along with a number of additional
unrelated observations have led us to propose that
catalysis for the DNA alkylation reaction is derived from
a DNA binding-induced conformational change in the
agent that disrupts the vinylogous amide stabilization
of the alkylation subunit and activates the agent for
nucleophilic addition.31 This conformational change re-
sults from adoption of a helical bound conformation that
follows the curvature and pitch of the DNA minor groove.
The helical rise in the bound conformation of the rigid
agents is adjusted by twisting the linking N2 amide which
is the only available flexible site. The twisting of the ø1
dihedral angle of the linking amide (ø2 ∼ 0°) diminishes
the N2 lone pair conjugation with the cyclohexadienone,
disrupts the vinylogous amide stabilization of the alky-
lation subunit, and increases its inherent reactivity. An
alternative possibility involves a twisting of the ø2
dihedral angle, diminishing the amide conjugation and
increasing the N2 vinylogous amide conjugation. This
would increase the basicity of the C4 carbonyl leading to
more effective protonation. Notably, both are consistent
with the studies that demonstrate even subtle perturba-
tions in the vinylogous amide have a remarkably large
impact on reactivity (F ) -3.0).34 Although our present
studies do not directly distinguish between these two
possibilities, the latter reflects changes that occur upon
N-deacylation (e.g., N-BOC-CBI to CBI, decreased reac-
tivity) while the former reflects the changes observed in
going to the product of the reaction (fully engaged amide,
ø2 ) 0°; no vinylogous amide, ø1 ∼20-35° and lengthened
bond c). It is consistent with a DNA bound conformation
agent
configuration
IC50 (L1210)
(+)-N-BOC-CNA
(-)-N-BOC-CNA
(+)-CNA-TMI
natural
5 µM (4 µM)a
unnatural
unnatural
natural
16 µM (29 µM)a
10 µM (11 µM)a
1 µM (0.4 µM)a
(-)-CNA-TMI
a
The value in parentheses is the IC50 value for the correspond-
ing seco-precusor.
This has important ramifications on the source of the
DNA alkylation selectivity. The inherent twist and
helical rise of the bound conformation of the agent is
greatest within the narrower, deeper AT-rich minor
groove. This leads to preferential activation of the agent
for DNA alkylation within extended AT-rich minor groove
sites and complements their preferential AT-rich nonco-
valent binding selectivity.36 Thus, both shape-selective
recognition (preferential AT-rich noncovalent binding)
and shape-dependent catalysis (extended AT-rich > GC-
rich activation by twist in N2 amide) combine to restrict
SN2 alkylation to accessible adenine N3 nucleophilic sites
within the preferred binding sites. Importantly, this
ground state destabilization of the substrate only acti-
vates the agent for a rate determining SN2 nucleophilic
addition and requires the subsequent proper positioning
and accessibility to an adenine N3 site.
This source of catalysis requires an extended and rigid
N2 amide substituent,12 and the absence of such a
substituent with 4-6 accounts nicely for their slow and
ineffective DNA alkylation. The noncovalent binding
derived from the attached right-hand subunits accounts
for a much smaller part of the difference in the rates of
DNA alkylation between 4-6 and 1-3.12 More impor-
tantly, this source of catalysis would lead to distinctions,
not similarities, in the DNA alkylation selectivities of 4-6
versus 1 contrary to the consequences of alternative
proposals that have been advanced.6,10,37
In Vitr o Cytotoxic Activity. The in vitro cytotoxic
activity of the agents proved to be consistent with past
observations that have illustrated a direct correlation
between solvolysis stability and cytotoxic potency. The
CNA-based agents exhibited cytotoxic activity that was
less potent than the corresponding duocarmycin-based
agents but more potent than the corresponding CI-based
agents (Table 3 and Figure 10). The natural enantiomers
of the CNA-based agents were found to be more potent
than the unnatural enantiomers.38 Interestingly, the
natural enantiomer of the full analog of duocarmycin SA,
CNA-TMI (30), was only 5-10× more potent than
N-BOC-CNA (10), and the difference in potency of the
unnatural enantiomers was even smaller (1-2×). While
this may simply be due to the extraordinary reactivity
of the agents which preclude effective DNA alkylation,
1
of duocarmycin SA established by H NMR which exhib-
ited at 44 ( 2° twist between the planes of the two
subunits with the bulk of the twist being accommodated
in ø1.35 The remarkably large and appropriate reactivity
changes observed herein that accompany such a decou-
pling of the vinylogous amide including that resulting
from a twist in the ø1 dihedral angle is consistent with
this as a source of catalysis. The reactivity of N-CO2Me-
CNA is extraordinary exhibiting a t1/2 of only 2.1 h at
pH 7 in the absence of deliberate added acid catalysis. It
is 103-104× more reactive than N-BOC-DSA and repre-
sents an agent that benefits from little, if any, vinylogous
amide stabilization. This level of reactivity is greater
than that required. In fact, it is the reactivity and ø1
dihedral angle of N-BOC-CBQ that may more closely
approximate that required for the DNA alkylation ca-
talysis provided by the DNA binding-induced conforma-
tional change in 1-3. Its inherent reactivity at pH 7
coupled with the rate enhancements afforded a bound
species that might provide a further 102× rate accelera-
tion approximating the rates observed with the DNA
alkylation reaction.
(36) Boger, D. L.; Coleman, R. S.; Invergo, B. J .; Zarrinmayeh, H.;
Kitos, P. A.; Thompson, S. C.; Leong, T.; McLaughlin, L. W. Chem.
Biol. Interact. 1990, 73, 29. Boger, D. L.; Zhou, J .; Cai, H. Bioorg. Med.
Chem. 1996, 4, 859.
(37) Lin, C. H.; Sun, D.; Hurley, L. H. Chem. Res. Toxicol. 1991, 4,
21. Lee, C.-S.; Sun, D.; Kizu, R.; Hurley, L. H. Chem. Res. Toxicol.
1991, 4, 203. Lin, C. H.; Hill, G. C.; Hurley, L. H. Chem. Res. Toxicol.
1992, 5, 167. Ding, Z.-M.; Harshey, R. M.; Hurley, L. H. Nucl. Acids
Res. 1993, 21, 4281. Sun, D.; Lin, C. H.; Hurley, L. H. Biochemistry
1993, 32, 4487. Thompson, A. S.; Sun, D.; Hurley, L. H. J . Am. Chem.
Soc. 1995, 117, 2371.
(38) The tentative assignment of the absolute configuration of CNA-
TMI (30) is based on both the more potent cytotoxic activity and the
DNA alkylation selectivity of (-)-CNA-TMI which corresponds to that
of the natural enantiomers. The potential that this may be reversed
by virtue of the reversed regioselectivity of adenine N3 nucleophilic
addition is presently under investigation.
(34) Boger, D. L.; Yun, W. J . Am. Chem. Soc. 1994, 116, 5523. The
increased stability and decreased acid-catalyzed solvolysis reactivity
that smoothly correlates with the electron-withdrawing strength of the
N2 substituent detailed in this work suggests the interesting possibility
that solvolysis catalysis may be derived from protonation of the N2
acyl substituent carbonyl rather than C4 carbonyl protonation. The
net effect of this protonation would be further deconjugation of the
vinylogous amide. Moreover, the rate of solvolysis would be dependent
upon the relative ease of protonation of the N2 acyl substituent carbonyl
(CONH2 > CO2CH3 > COCH3).
(35) Eis, P.; Smith, J . A.; Rydzewski, J . M.; Case, D. A.; Boger, D.
L.; Chazin, W. J . J . Mol. Biol., in press.

Synthesis and Evaluation of CC-1065 and Duocarmycin Analogs
J . Org. Chem., Vol. 62, No. 17, 1997 5859
precipitated salts were removed by filtration and thoroughly
rinsed with the solvent mixture. Concentration of the filtrate
and radial chromatography (SiO2, 2 mm plate, 5% EtOAc-
hexane) afforded 14 as a colorless oil (0.343 g, 0.384 g
theoretical, 89%) which slowly crystallized upon storage: mp
99-100 °C; 1H NMR (400 MHz, CDCl3) major rotamer δ 8.37-
8.35 (m, 1H), 8.13-8.11 (m, 1H), 7.54-7.34 (m, 7H), 6.72 (s,
1H), 5.52 (s, 1H), 5.24 (s, 2H), 4.94 (s, 1H), 4.50-4.10 (m, 1H),
3.70-3.30 (m, 1H), 2.95-1.70 (m, 4H), 1.26 (s, 9H); 13C NMR
(100 MHz, CDCl3) major rotamer δ 153.1, 145.3, 137.2, 137.0,
132.2, 129.8, 128.7 (3C), 128.0 (2C), 127.7, 127.2, 126.6, 126.1,
124.9, 122.1, 117.3, 106.5, 79.7, 70.2, 47.4, 35.1, 29.8, 28.3 (3C);
IR (film) νmax 3071, 3031, 2971, 2931, 2852, 1694, 1590 cm-1
;
FABHRMS (NBA-NaI) m/z 438.2065 (M + Na+, C27H29NO3
requires 438.2045).
7-(Ben zyloxy)-5-(ter t-b u t yloxyca r b on yl)-1-(h yd r oxy-
m eth yl)-1,2,3,4-tetr ah ydr o-5H-n aph th o[1,2-b]azepin e (15).
A solution of 14 (0.209 g, 0.504 mmol, 1 equiv) in THF (5 mL)
was cooled to 0 °C prior to dropwise addition of BH3•SMe2
(0.151 mL, 10 M, 3(9) equiv). The cooling bath was removed
after 5 min, and the mixture was stirred at 25 °C for 12 h.
The excess borane was quenched with slow addition of H2O
(0.6 mL). Oxidative workup was accomplished by the addition
of 2.5 M aqueous NaOH (0.605 mL, 3 equiv) followed by 30%
H2O2 (0.514 mL, 9 equiv), and the resulting heterogeneous
solution was stirred rapidly at 25 °C (1 h) and 50 °C (1 h).
The cooled reaction mixture was treated with saturated
aqueous NaCl (0.5 mL), and the layers were separated. The
aqueous portion was extracted with EtOAc (2 × 5 mL), and
the combined organic portions were dried (MgSO4), filtered,
and concentrated. Radial chromatography (SiO2, 2 mm plate,
30% EtOAc-hexane) provided 15 as a colorless oil (0.179 g,
0.218 g theoretical, 82%) which slowly crystallized upon
F igu r e 10.
this also suggests that the twisted conformation of the
full analogs may be sufficient to disfavor minor groove
binding providing agents that are comparable to analogs
lacking the attached DNA binding subunits altogether.
1
storage: mp 108-110 °C; H NMR (400 MHz, CDCl3) major
rotamer δ 8.40 (d, J ) 8.3 Hz, 1H), 8.18 (d, J ) 8.6 Hz, 1H),
7.57-7.34 (m, 7H), 6.62 (s, 1H), 5.28 (d, J ) 12.1 Hz, 1H),
5.17 (d, J ) 12.0 Hz, 1H), 4.41-4.37 (m, 1H), 4.09-4.06 (m,
1H), 3.88-3.86 (m, 2H), 2.81-2.75 (m, 1H), 2.23-2.03 (m, 2H),
1.71-1.48 (m, 3H), 1.29 (s, 9H); 13C NMR (100 MHz, CDCl3)
major rotamer δ 154.0, 153.2, 139.9, 136.9, 133.6, 128.7 (2C),
128.0 (2C), 127.7, 127.0 (2C), 125.4, 125.0, 123.7, 122.7, 107.3,
80.2, 70.1, 63.3, 47.5, 39.1, 28.3 (3C), 26.3, 24.1; IR (film) νmax
3429, 3067, 3037, 2967, 2926, 2866, 1694, 1674, 1619, 1594
cm-1; FABHRMS (NBA-CsI) m/z 566.1333 (M + Cs+,
C27H31NO4 requires 566.1307). Anal. Calcd for C27H31NO4:
C, 74.79; H, 7.21; N, 3.23. Found: C, 74.49; H, 7.37; N, 3.44.
7-(Ben zyloxy)-5-(ter t-bu tyloxyca r bon yl)-1-[((m eth a n e-
su lfon yl)oxy)m eth yl]-1,2,3,4-tetr ah ydr o-5H-n aph th o[1,2-
b]a zep in e (16). A solution of 15 (0.138 g, 0.319 mmol, 1
equiv) in CH2Cl2 (3 mL) was cooled to 0 °C prior to sequential
addition of Et3N (0.222 mL, 1.59 mmol, 5 equiv) and CH3SO2Cl
(0.049 mL, 0.637 mmol, 2 equiv). The cooling bath was
removed after 5 min, and the mixture was stirred at 25 °C for
1 h. The reaction was quenched with the addition of saturated
aqueous NaHCO3 (0.5 mL), and the layers were separated. The
aqueous portion was extracted with EtOAc (2 × 5 mL), and
the combined organic portions were dried (MgSO4), filtered,
and concentrated. Radial chromatography (SiO2, 2 mm plate,
30% EtOAc-hexane) furnished 16 (0.143 g, 0.163 g theoretical,
Exp er im en ta l Section
4-(Ben zyloxy)-N-(ter t-b u t yloxyca r b on yl)-1-iod o-N-(4-
p en ten -1-yl)-2-n a p h th yla m in e (13). A solution of 1219 (1.00
g, 2.11 mmol, 1 equiv) in benzene (100 mL) was treated with
50% w/v aqueous NaOH (20 mL). Benzyltributylammonium
chloride (1.32 g, 4.22 mmol, 2 equiv) was added followed by
5-[(methanesulfonyl)oxy]-1-pentene (1.73 g, 10.53 mmol, 5
equiv). The resulting biphasic mixture was stirred vigorously
at 25 °C for 12 h. The layers were allowed to separate, and
the aqueous portion was extracted with EtOAc (2 × 30 mL).
The combined organic portions were washed with H2O (2 ×
100 mL) and saturated aqueous NaCl (50 mL). Drying
(MgSO4), filtration, and concentration was followed by flash
chromatography (SiO2, 3.5 × 14 cm, 5% EtOAc-hexane) to
furnish 13 (0.975 g, 1.15 g theoretical, 85%) as a pale yellow
solid: mp 69-71 °C; 1H NMR (500 MHz, CDCl3) major rotamer
δ 8.34-8.32 (m, 1H), 8.23-8.21 (m, 1H), 7.62-7.49 (m, 4H),
7.42-733 (m, 3H), 6.70 (s, 1H), 5.81-5.71 (m, 1H), 5.34-5.18
(m, 2H), 5.01-4.93 (m, 2H), 3.84 (ddd, J ) 17.5, 12.6, 7.6 Hz,
1H), 3.37-3.31 (m, 1H), 2.07-2.02 (m, 2H), 1.74-1.65 (m, 2H),
1.31 (s, 9H); 13C NMR (125 MHz, CDCl3) major rotamer δ
155.1, 154.0, 143.2, 137.9, 136.4, 135.4, 132.8, 128.7 (2C),
128.5, 128.2 (2C), 127.2, 126.2, 122.5, 115.0, 107.8, 95.1, 80.1,
70.3, 49.1, 31.3, 28.6, 28.3 (3C), 27.5; IR (film) νmax 3067, 3036,
2974, 2923, 1697, 1615, 1590 cm-1; FABHRMS (NBA-NaI)
m/z 544.1337 (M + Na+, C27H30INO3 requires 544.1349). Anal.
Calcd for C27H30INO3: C, 59.65; H, 5.57; N, 2.58. Found C,
60.02; H, 5.27; N, 2.54.
1
89%) as a white solid: mp 134-135 °C; H NMR (400 MHz,
CDCl3) major rotamer δ 8.40 (d, J ) 8.2 Hz, 1H), 8.13 (d, J )
8.6 Hz, 1H), 7.60-7.35 (m, 7H), 6.63 (s, 1H), 5.26 (d, J ) 11.7
Hz, 1H), 5.17 (d, J ) 11.6 Hz, 1H), 4.49-4.44 (m, 2H), 4.38-
4.30 (m, 2H), 2.69-2.64 (m, 4H), 2.26-2.05 (m, 2H), 1.73-
1.70 (m, 2H), 1.32 (s, 9H); 13C NMR (100 MHz, CDCl3) major
rotamer δ 153.8, 153.6, 140.5, 136.7, 133.0, 128.8 (2C), 128.6,
128.1 (2C), 127.7, 127.6, 127.1, 125.2, 123.3, 122.8, 107.1, 80.8,
71.4, 70.2, 47.7, 37.1, 36.1, 28.3 (3C), 26.5, 24.4; IR (film) νmax
3056, 2974, 2933, 2862, 1692, 1615, 1590 cm-1; FABHRMS
(NBA-NaI) m/z 534.1914 (M + Na+, C28H33NO6S requires
534.1926). Anal. Calcd for C28H33NO6S: C, 65.73; H, 6.51;
N, 2.74; S, 6.25. Found: C, 65.68; H, 6.59; N, 2.55; S, 6.31.
A solution of 16 (0.055 g) in 50% i-PrOH-hexane was
resolved on a semipreparative Diacel Chiracel OD column (10
µm, 2 × 25 cm) using 10% i-PrOH-hexane as eluent (7 mL/
min). The effluent was monitored at 254 nm and the enan-
7-(Ben zyloxy)-5-(ter t-bu tyloxycar bon yl)-1-m eth yliden e-
1,2,3,4-tetr a h yd r o-5H-n a p h th o[1,2-b]a zep in e (14). A so-
lution of 13 (0.503 g, 0.926 mmol, 1 equiv) in CH3CN (18 mL,
0.051 M, degassed and purged with Ar) in a thick-walled
reaction tube was treated with Et3N (0.257 mL, 1.85 mmol, 2
equiv) followed by (Ph3P)4Pd (0.032 g, 0.0277 mmol, 0.03
equiv). The reaction vessel was sealed, and the mixture was
warmed at 130 °C for 14 h. Concentration gave a yellow
semisolid that was suspended in 5% EtOAc-hexanes. The

5860 J . Org. Chem., Vol. 62, No. 17, 1997
Boger and Turnbull
tiomers eluted with retention times of 38.0 and 45.0 min,
respectively (R ) 1.18). The fractions were collected and
concentrated to afford ent-(+)-(1S)-16 (tR ) 38.0 min, 0.0251
g) and (-)-(1R)-16 (tR ) 45.0 min, 0.0248 g) with a 91%
recovery (>99.9% ee).
1693, 1663 cm-1; FABHRMS (NBA-NaI) m/z 384.1352 (M +
Na+, C20H24ClNO3 requires 384.1342).
A solution of 19 (0.0096 g) in 50% i-PrOH-hexane was
resolved on a semipreparative Diacel Chiracel OD column (10
µm, 2 × 25 cm) using 6% i-PrOH-hexane as eluent (6 mL/
min). The effluent was monitored at 254 nm, and the
enantiomers were eluted with retention times of 22.6 and 26.8
min, respectively (R ) 1.19). The fractions were collected and
concentrated to afford (-)-(1R)-19 (tR ) 22.6 min, 0.0033 g)
and ent-(+)-(1S)-19 (tR ) 26.8 min, 0.0031 g) with a 67%
recovery (>99.9% ee). The absolute configuration of (-)-(1R)-
19 was established by X-ray analysis.24
(-)-(1R)-16: [R]25 -41 (c 0.12, CHCl3).
D
ent-(+)-(1S)-16: [R]25 +46 (c 0.13, CHCl3).
D
5-(t er t -Bu t yloxyca r b on yl)-7-h yd r oxy-1-[((m e t h a n e -
su lfon yl)oxy)m eth yl]-1,2,3,4-tetr ah ydr o-5H-n aph th o[1,2-
b]a zep in e (17). A solution of 16 (0.096 g, 0.188 mmol, 1
equiv) in THF (3.0 mL) was treated with 10% Pd-C (0.042 g,
0.04 mmol, 0.2 equiv) followed by aqueous HCO2NH4 (1.0 mL,
25% w/v, 21 equiv) and the resulting mixture was stirred at
25 °C for 2.5 h. The mixture was filtered through Celite to
afford a biphasic filtrate which was partitioned. The aqueous
phase was extracted with EtOAc (2 × 2 mL), and the combined
organic portions were dried (MgSO4), filtered, and concentrated
to give 17 as an analytically pure white solid (0.072 g, 0.079
g theoretical, 91%): mp 141-142 °C; 1H NMR (400 MHz, DMF-
d7) major rotamer δ 8.29 (d, J ) 8.2 Hz, 1H), 8.26 (d, J ) 8.7
Hz, 1H), 7.62-7.59 (m, 1H), 7.54-7.50 (m, 1H), 6.82 (s, 1H),
4.59-4.47 (m, 1H), 4.43-4.25 (m, 3H), 3.02 (s, 3H), 2.98-2.76
(m, 1H), 2.35-2.20 (m, 1H), 2.13-1.95 (m, 1H), 1.71-1.67 (m,
2H), 1.39 (s, 9H); 13C NMR (100 MHz, DMF-d7) major rotamer
δ 154.0, 153.5, 141.8, 134.0, 127.7, 125.2, 125.1, 124.3, 123.3,
123.0, 110.3, 80.4, 71.4, 48.2, 36.8, 36.4, 28.2 (3C), 27.2, 25.0;
(-)-(1R)-19: [R]25 -67 (c 0.017, THF).
D
ent-(+)-(1S)-19: [R]25 +67 (c 0.016, THF).
D
N-(ter t-Bu tyloxyca r bon yl)-1,2,3,4,11,11a -h exa h yd r ocy-
clopr opa[c]n aph th o[2,1-b]azepin -6-on e (N-BOC-CNA, 10).
A sample of 17 (0.025 g, 0.059 mmol, 1 equiv) was suspended
in CH3CN (1.0 mL) and treated with DBU (0.027 mL, 0.178
mmol, 3 equiv) at 25 °C to instantly afford a homogeneous pale
yellow solution. The solvent was removed with a gentle stream
of N2 after 5 min. Flash chromatography (SiO2, 1 × 7 cm, 30%
EtOAc-hexane with 5% Et3N) furnished 10 as a colorless oil
(0.017 g, 0.019 g theoretical, 87%): 1H NMR (400 MHz,
CD3CN) δ 8.12 (dd, J ) 7.9, 1.5 Hz, 1H), 7.61 (ddd, J ) 8.3,
7.1, 1.5 Hz, 1H), 7.41 (ddd, J ) 8.0, 7.2, 1.0 Hz, 1H), 7.14 (d,
J ) 8.1 Hz, 1H), 6.48 (s, 1H), 4.09-4.02 (m, 1H), 2.69-2.54
(m, 1H), 2.28-2.22 (m, 1H), 2.10-2.07 (m, 2H), 1.95-1.91 (m,
2H overlapping with solvent), 1.75-1.72 (m, 1H), 1.62-1.57
(m, 1H), 1.26 (broad s, 9H); 13C NMR (100 MHz, CD3CN) δ
186.4, 162.1, 146.8, 133.7, 132.8, 129.3, 126.8, 126.6, 123.3,
81.0, 47.8, 40.2, 33.8, 28.4, 28.3, 26.8 (3C), 26.0, 25.9; IR (film)
νmax 2923, 2852, 1698, 1647, 1596 cm-1; UV (THF) λmax 304 (ꢀ
8600), 258 (ꢀ 11100), 220 (ꢀ 15500) nm; FABHRMS (NBA-
NaI) m/z 326.1766 (M + H+, C20H24NO3 requires 326.1756).
IR (film) νmax 3262, 2974, 2933, 2872, 1692, 1662, 1594 cm-1
;
FABHRMS (NBA-NaI) m/z 444.1468 (M + Na+, C21H27NO6S
requires 444.1457).
(-)-(1R)-17: [R]25 -63 (c 0.084, THF).
D
ent-(+)-(1S)-17: [R]25 +58 (c 0.084, THF).
D
7-(Be n zyloxy)-5-(t er t -b u t yloxyca r b on yl)-1-(ch lor o-
m eth yl)-1,2,3,4-tetr ah ydr o-5H-n aph th o[1,2-b]azepin e (18).
A solution of 15 (0.105 g, 0.24 mmol, 1 equiv) in CH2Cl2 (2
mL) was treated with CCl4 (0.225 mL, 2.33 mmol, 9.7 equiv)
followed by Ph3P (0.212 g, 0.81 mmol, 3.3 equiv), and the
mixture was stirred at 25 °C for 2.5 h. The crude reaction
solution was passed through a plug of silica gel and concen-
trated. Radial chromatography (SiO2, 1 mm plate, 5% EtOAc-
hexane) provided 18 as a pale yellow oil (0.048 g, 0.108 g
theoretical, 44%): 1H NMR (250 MHz, CDCl3) major rotamer
δ 8.44-8.36 (m, 1H), 8.15-7.12 (m, 1H), 7.62-7.33 (m, 7H),
6.64 (s, 1H), 5.31-5.15 (m, 2H), 4.46-4.41 (m, 1H), 4.17-4.08
(m, 1H), 3.92-3.81 (m, 1H), 3.74-3.63 (m, 1H), 2.78-2.69 (m,
1H), 2.41-2.35 (m, 1H), 2.17-2.05 (m, 1H), 1.72-1.67 (m, 2H),
1.30 (s, 9H); IR (film) νmax 2971, 2930, 2859, 1693, 1617, 1592
cm-1; FABHRMS (NBA-NaI) m/z 451.1933 (M+, C27H30ClNO3
requires 451.1914).
(+)-10: [R]25 +36 (c 0.009, EtOAc).
D
ent-(-)-10: [R]25 -36 (c 0.016, EtOAc).
D
1,2,3,4,11,11a -H exa h yd r ocyclop r op a [c]n a p h t h o[2,1-
b]a zep in -6-on e (CNA, 11). A sample of 17 (0.030 g, 0.07
mmol, 1 equiv) was suspended in 3.9 M HCl-EtOAc (1.5 mL).
After 40 min, the volatiles were removed with a gentle stream
of N2 followed by high vacuum. The resulting white foam was
dissolved in CH3CN containing DBU (0.105 mL, 0.70 mmol,
10 equiv in 1 mL of CH3CN). The solvent was removed with
a gentle stream of N2 after 5 min. Radial chromatography
(SiO2, 1 mm plate, 95% CH2Cl2-CH3OH) furnished 11 as a
pale yellow solid (0.012 g, 0.016 g theoretical, 77%). Recrys-
tallization from 10% CH3OH-CH3CN gave pale yellow plates
suitable for X-ray analysis:24 mp 180-183 °C with decomposi-
tion; 1H NMR (400 MHz, CD3OD) δ 8.07 (dd, J ) 7.8, 1.5 Hz,
1H), 7.50 (ddd, J ) 8.7, 7.2, 1.6 Hz, 1H), 7.33 (ddd, J ) 8.0,
7.2, 1.0 Hz, 1H), 7.13 (d, J ) 8.1 Hz, 1H), 6.00 (s, 1H), 3.54-
3.48 (m, 1H), 3.00-2.90 (m, 1H), 2.78-2.74 (m, 1H), 2.50 (dd,
J ) 8.7, 5.0 Hz, 1H), 2.41-2.36 (m, 1H), 2.24-2.15 (m, 1H),
1.92-1.79 (m, 3H); 13C NMR (150 MHz, CD3OD) δ 184.6, 175.0,
146.3, 133.0, 132.7, 126.6, 126.1, 122.4, 107.9, 45.4, 38.2, 31.0
(2C), 26.4, 26.1; IR (film) νmax 3285, 2924, 2855, 1608, 1584
cm-1; UV (CH3OH) λmax 342 (ꢀ 11900), 308 (ꢀ 8300), 245 (ꢀ
17100) nm; FABHRMS (NBA-NaI) m/z 226.1238 (M + H+,
C15H16NO requires 226.1232).
5-(ter t-Bu tyloxyca r bon yl)-1-(ch lor om eth yl)-7-h yd r oxy-
1,2,3,4-tetr a h yd r o-5H-n a p h th o[1,2-b]a zep in e (19). A so-
lution of 18 (0.043 g, 0.095 mmol, 1 equiv) in THF (1.5 mL)
was treated with 10% Pd-C (0.020 g, 0.019 mmol, 0.2 equiv)
followed by aqueous HCO2NH4 (0.5 mL, 25% w/v, 1.98 mmol,
21 equiv), and the resulting black suspension was stirred
rapidly at 25 °C for 2.5 h. The crude mixture was filtered
through a plug of Celite to afford a biphasic filtrate which was
partitioned. The aqueous portion was extracted with EtOAc
(2 × 1 mL), and the combined organic portions were dried
(MgSO4), filtered, and concentrated to provide 19 as an
analytically pure white solid (0.032 g, 0.034 g theoretical, 94%).
Recrystallization from 10% EtOAc-hexane gave colorless
plates suitable for X-ray analysis:24 mp 157-159 °C; 1H NMR
(400 MHz, CDCl3) major rotamer δ 7.98-7.94 (m, 1H), 7.37-
7.32 (m, 1H), 7.22-7.19 (m, 1H), 6.96-6.90 (m, 1H), 6.47 (s,
1H), 4.41-4.26 (m, 1H), 4.19-4.07 (m, 1H), 3.89-3.81 (m, 2H),
2.88-2.78 (m, 1H), 2.51-2.46 (m, 1H), 2.18-2.05 (m, 1H),
7-(Ben zyloxy)-1-m et h ylid en e-1,2,3,4-t et r a h yd r o-5H -
n a p h th ol[1,2-b]a zep in e Hyd r och lor id e (20). A solution
of 14 (0.226 g, 0.55 mmol, 1 equiv) was dissolved in 3.9 M HCl-
EtOAc (7.0 mL). After 40 min, the volatiles were removed with
a gentle stream of N2 followed by high vacuum. The resulting
crude pale green solid was used directly for the synthesis of
21 and 22 without further characterization.
7-(Ben zyloxy)-5-(m et h oxyca r b on yl)-1-m et h ylid en e-
1,2,3,4-tetr a h yd r o-5H-n a p h th o[1,2-b]a zep in e (21). A so-
lution of 20 (0.049 g, 0.14 mmol, 1 equiv) in THF (1.5 mL)
was treated with NaHCO3 (0.026 g, 0.31 mmol, 2.2 equiv)
followed by methyl chloroformate (0.022 mL, 0.28 mmol, 2.0
equiv), and the mixture was stirred at 25 °C for 5 h. The
reaction was quenched with the addition of saturated aqueous
NaHCO3 (0.5 mL), and the resulting layers were separated.
The aqueous portion was extracted with EtOAc (2 × 2 mL),
and the combined organic portions were dried (MgSO4),
1
1.71-1.67 (m, 2H), 1.40 (s, 9H); H NMR (400 MHz, acetone-
d6) major rotamer δ 9.17 (s, 1H), 8.27 (d, J ) 8.1 Hz, 1H), 8.19-
8.17 (m, 1H), 7.59-7.54 (m, 1H), 7.50-7.46 (m, 1H), 6.75 (s,
1H), 4.39-4.36 (m, 1H), 4.19-4.15 (m, 1H), 3.92-3.87 (m, 1H),
3.75-3.63 (m, 1H), 2.75-2.68 (m, 1H), 2.37-2.33 (m, 1H),
2.06-2.04 (m, 1H), 1.69-1.67 (m, 2H), 1.40 (s, 9H); 13C NMR
(100 MHz, CDCl3) major rotamer δ 155.8, 152.9, 151.6, 139.3,
132.3, 126.5, 124.9, 124.7, 123.5, 122.4, 109.4, 81.5, 48.9, 44.6,
38.6, 28.4 (3C), 26.3, 24.2; IR (film) νmax 3287, 2965, 2915, 2844,

Synthesis and Evaluation of CC-1065 and Duocarmycin Analogs
J . Org. Chem., Vol. 62, No. 17, 1997 5861
filtered, and concentrated. Trituration with 10% EtOAc-
hexane furnished 21 as a white solid (0.049 g, 0.052 g
theoretical, 94%): mp 99-100 °C; 1H NMR (500 MHz, CDCl3)
major rotamer δ 8.38-8.36 (m, 1H), 8.14-8.13 (m, 1H), 7.54-
7.36 (m, 7H), 6.73 (s, 1H), 5.56 (s, 1H), 5.25 (s, 2H), 4.99 (s,
1H), 4.50-4.20 (m, 1H), 3.59 (s, 3H), 3.10-2.35 (m, 1H), 2.22-
1.70 (m, 4H); 13C NMR (125 MHz, CDCl3) major rotamer δ
155.5, 153.4, 144.7, 136.8, 136.2, 132.2, 129.9, 128.6 (2C),
127.9, 127.3 (2C), 126.8, 126.1, 125.2 (2C), 122.2, 118.0, 105.8,
70.2, 52.8, 48.1, 35.0, 29.7; IR (film) νmax 3067, 2923, 2862,
1703, 1590, 1508 cm-1; FABHRMS (NBA-NaI) m/z 373.1688
(M+, C24H23NO3 requires 373.1678).
7-(Ben zyloxy)-1-(h ydr oxym eth yl)-5-(m eth oxycar bon yl)-
1,2,3,4-tetr a h yd r o-5H-n a p h th o[1,2-b]a zep in e (23). Fol-
lowing the general procedure detailed for 15, 21 (0.049 g, 0.013
mmol, 1 equiv) was treated with BH3•SMe2 (0.039 mL, 10 M,
3(9) equiv), quenched with H2O (0.16 mL), and oxidized with
aqueous 2.5 M NaOH (0.16 mL, 0.39 mmol, 3 equiv) and 30%
aqueous H2O2 (0.134 mL, 9 equiv). Radial chromatography
(SiO2, 1 mm plate, 50% EtOAc-hexane) provided 23 as a white
solid (0.033 g, 0.051 g theoretical, 64%): mp 90-92 °C; 1H
NMR (500 MHz, CDCl3) major rotamer δ 8.41 (d, J ) 8.5 Hz,
1H), 8.18 (d, J ) 8.5 Hz, 1H), 7.57-7.35 (m, 7H), 6.63 (s, 1H),
5.25-5.17 (m, 2H), 4.46-4.43 (m, 1H), 4.09-4.07 (m, 1H),
3.87-3.76 (m, 2H), 3.59 (s, 3H), 2.93-2.83 (m, 1H), 2.21-2.08
(m, 2H), 1.70-1.64 (m, 3H); 13C NMR (125 MHz, CDCl3) major
rotamer δ 155.6, 153.4, 139.0, 136.7, 133.5, 128.6 (2C), 128.0
(2C), 127.6, 127.3 (3C), 125.2, 123.7, 122.8, 106.9, 70.1, 63.4,
52.9, 48.3, 39.1, 26.4, 24.1; IR (film) νmax 3453, 3063, 3032,
2940, 2868, 1696, 1593 cm-1; FABHRMS (NBA-NaI) m/z
392.1871 (M + H+, C24H25NO4 requires 392.1862).
Hz, 1H), 7.42 (ddd, J ) 8.0, 7.2, 1.5 Hz, 1H), 7.20 (d, J ) 8.5
Hz, 1H), 6.52 (s, 1H), 4.26-4.15 (m, 1H), 3.61 (s, 3H), 2.79-
2.73 (m, 1H), 2.32-2.28 (m, 1H), 2.18-2.09 (m, 3H), 1.96-
1.94 (m, 1H), 1.80-1.78 (m, 1H), 1.69-1.65 (m, 1H); 13C NMR
(125 MHz, acetone-d6) δ 185.6, 155.2, 146.5 (2C), 133.6, 132.9,
129.7, 126.7, 126.5, 123.2, 53.1, 48.6, 39.8, 33.5, 26.8, 25.8, 25.5;
IR (film) νmax 3067, 2923, 2851, 1708, 1646, 1600, 1446 cm-1
;
FABHRMS (NBA-NaI) m/z 284.1281 (M + H+, C17H17NO3
requires 284.1287).
7-(Ben zyloxy)-5-[(5,6,7-tr im eth oxyin d ol-2-yl)ca r bon yl]-
1-m et h ylid en e-1,2,3,4-t et r a h yd r o-5H -n a p h t h o[1,2-b]-
a zep in e (22). A solution of 20 (0.142 g, 0.405 mmol, 1 equiv)
in DMF (3 mL) was treated with [[3-(dimethylamino)propyl]-
ethyl]carbodiimide hydrochloride (EDCI, 0.233 g, 1.22 mmol,
3 equiv) followed by 5,6,7-trimethoxyindole-2-carboxylic acid
(0.122 g, 0.486 mmol, 1.2 equiv), and the reaction mixture was
stirred at 25 °C for 4.5 h. The crude reaction mixture was
diluted with EtOAc (4 mL) and extracted with H2
O (2 mL).
The aqueous portion was extracted with EtOAc (2 × 1 mL).
The combined organic portions were dried (MgSO4), filtered,
and concentrated. Radial chromatography (SiO2, 2 mm plate,
50% EtOAc-hexane) furnished 22 (0.177 g, 0.222 g theoretical,
80%) as a pale yellow solid: mp 88-90 °C; 1H NMR (500 MHz,
CDCl3) major rotamer δ 9.30 (s, 1H), 8.47-8.44 (m, 1H), 7.98-
7.97 (m, 1H), 7.63-7.57 (m, 2H), 7.39-7.36 (m, 2H), 7.26-
7.18 (m, 3H), 6.72 (s, 1H), 6.46 (s, 1H), 6.23-6.18 (m, 1H),
5.31-5.27 (m, 1H), 5.19-5.14 (m, 1H), 5.09-5.04 (m, 1H),
4.94-4.89 (m, 1H), 4.02 (s, 3H), 3.88 (s, 3H), 3.75 (s, 3H), 3.67-
3.61 (m, 1H), 2.18-2.10 (m, 5H); 13C NMR (125 MHz, CDCl3)
major rotamer δ 162.1, 154.5, 149.4, 144.3, 139.8, 138.6, 137.3,
136.5, 136.4, 136.3, 132.0, 129.5, 128.4 (2C), 127.9, 127.1 (2C),
126.9, 126.3, 125.7, 123.4, 122.9, 118.2, 107.3, 106.9, 97.8, 70.2,
61.4, 56.0, 54.7, 47.3, 34.7, 25.2, 22.9; IR (film) νmax 3448, 3235,
3062, 2930, 1607, 1581, 1495 cm-1; FABHRMS (NBA-CsI) m/z
681.1378 (M + Cs+, C34H32N2O5 requires 681.1366).
7-(Be n zyloxy)-1-[((m e t h a n e su lfon yl)oxy)m e t h yl]-5-
(m eth oxyca r bon yl)-1,2,3,4-tetr a h yd r o-5H-n a p h th o[1,2-
b]a zep in e (25). Following the general procedure detailed for
16, 23 (0.029 g, 0.074 mmol, 1 equiv) was treated with Et3N
(0.052 mL, 0.371 mmol, 5 equiv) and CH3SO2Cl (0.017 mL,
0.148 mmol, 2 equiv). Radial chromatography (SiO2, 1 mm
plate, 50% EtOAc-hexane) furnished 25 (0.032 g, 0.035 g
7-(Ben zyloxy)-1-(h yd r oxym eth yl)-5-[(5,6,7-tr im eth oxy-
in dol-2-yl)car bon yl]-1,2,3,4-tetr ah ydr o-5H-n aph th o[1,2-
b]a zep in e (24). Following the general procedure detailed for
15, 22 (0.200 g, 0.365 mmol, 1 equiv) was treated with
BH3•SMe2 (0.110 mL, 10 M, 9 equiv), quenched with H2O (0.44
mL), and oxidized with aqueous 2.5 M NaOH (0.44 mL, 1.10
mmol, 3 equiv) and 30% aqueous H2O2 (0.373 mL, 9 equiv).
Radial chromatography (SiO2, 2 mm plate, 70% EtOAc-
hexane) provided 24 (0.107 g, 0.207 g theoretical, 52%) as an
amber oil. Recrystallization from EtOAc gave a white micro-
crystalline solid: mp 178-180 °C with decomposition; 1H NMR
(500 MHz, CDCl3) major rotamer δ 9.26 (s, 1H), 8.48 (d, J )
10.4 Hz, 1H), 8.28 (d, J ) 10.8 Hz, 1H), 7.68-7.65 (m, 1H),
7.60-7.57 (m, 1H), 7.35-7.33 (m, 2H), 7.22-7.13 (m, 3H), 6.66
(s, 1H), 6.43 (s, 1H), 5.23-5.21 (m, 1H), 5.13-5.10 (m, 1H),
5.03-4.99 (m, 1H), 4.95-4.90 (m, 1H), 4.24-4.21 (m, 2H), 4.03
(s, 3H), 3.87 (s, 3H), 3.73 (s, 3H), 3.26-3.20 (m, 1H), 2.46-
2.40 (m, 1H), 1.95-1.92 (m, 1H), 1.30-1.23 (m, 3H); 13C NMR
(125 MHz, CDCl3) major rotamer δ 160.7, 153.9, 149.5, 140.0,
139.8, 138.6, 136.3, 134.3, 129.0, 128.4 (2C), 127.8, 127.7, 127.1
(2C), 126.2, 125.9, 125.6, 124.8, 123.7, 123.3, 123.2, 108.4,
107.6, 97.9, 70.6, 70.2, 61.4, 61.1, 56.1, 41.3, 38.8, 31.8, 17.3;
1
theoretical, 91%) as a white solid: mp 130-132 °C; H NMR
(500 MHz, CDCl3) major rotamer δ 8.40 (d, J ) 8.5 Hz, 1H),
8.13 (d, J ) 8.5 Hz, 1H), 7.61-7.35 (m, 7H), 6.63 (s, 1H), 5.27-
5.17 (m, 2H), 4.51-4.33 (m, 4H), 3.60 (s, 3H), 2.86-2.81 (m,
1H), 2.54 (s, 3H), 2.32-2.23 (m, 1H), 2.07-2.04 (m, 1H), 1.76-
1.70 (m, 2H); 13C NMR (125 MHz, CDCl3) major rotamer δ
155.3, 153.7, 139.5, 136.6, 133.1, 128.6 (2C), 128.0 (2C), 127.7,
127.6, 127.3, 125.4, 124.3, 123.4, 122.8, 106.9, 71.1, 70.1, 53.0,
48.3, 36.8, 36.2, 26.7, 24.5 ; IR (film) νmax 3063, 3022, 2938,
2858, 1701, 1593 cm-1; FABHRMS (NBA) m/z 470.1648 (M +
H+, C25H27NO6S requires 470.1637).
7-Hyd r oxy-1-[((m eth a n esu lfon yl)oxy)m eth yl]-5-(m eth -
oxyca r b on yl)-1,2,3,4-t e t r a h yd r o-5H -n a p h t h o[1,2-b]-
a zep in e (27). Following the general procedure detailed for
17, 25 (0.028 g, 0.060 mmol, 1 equiv) was treated with 10%
Pd-C (0.010 g, 0.009 mmol, 0.16 equiv) followed by aqueous
HCO2NH4 (0.265 mL, 25% w/v, 18 equiv). Filtration and
concentration gave 27 as an analytically pure white solid
(0.0204 g, 0.0225 g theoretical, 91%): mp 113-115 °C; 1H NMR
(500 MHz, CD3OD) major rotamer δ 8.25 (d, J ) 7.5 Hz, 1H),
8.14 (d, J ) 8.5 Hz, 1H), 7.55-7.45 (m, 2H), 6.60 (s, 1H), 4.41-
4.33 (m, 4H), 3.66 (s, 3H), 2.85-2.79 (m, 1H), 2.64 (s, 3H),
IR (film) 3446, 3290, 3062, 2927, 1732, 1608, 1587 cm-1
;
FABHRMS (NBA-CsI) m/z 699.1482 (M + Cs+, C34H34N2O6
requires 699.1471).
2.35-2.20 (m, 1H), 2.13-1.95 (m, 1H), 1.71-1.67 (m, 2H); 13
C
A solution of 24 (0.025 g) in 50% i-PrOH-hexane was
resolved on a semipreparative Diacel Chiracel OD column (10
µm, 2 × 25 cm) using 30% i-PrOH-hexane as eluent (8 mL/
min). The effluent was monitored at 254 nm, and the
enantiomers were eluted with retention times of 20.9 and 24.0
min, respectively (R ) 1.15). The fractions were collected and
concentrated to afford (+)-(1R)-24 (tR ) 20.9 min, 0.0088 g)
and ent-(-)-(1S)-24 (tR ) 24 min, 0.0093 g) with a 72% recovery
(>99.9% ee).
NMR (125 MHz, CD3OD) major rotamer δ 157.2, 154.4, 140.9,
134.8, 128.3, 126.2, 125.8 (2C), 124.5, 124.3, 123.9, 72.5, 53.6,
49.9, 37.3, 36.9, 27.9, 25.5; IR (film) νmax 3283, 2938, 2857,
1672, 1621, 1591 cm-1; FABHRMS (NBA-NaI) m/z 379.1081
(M+, C18H21NO6S requires 379.1090).
N-(Met h oxyca r b on yl)-1,2,3,4,11,11a -h exa h yd r ocyclo-
pr opa[c]n aph th o[2,1-b]azepin -6-on e (N-CO2Me-CNA, 29).
Following the general procedure detailed for 10, 27 (0.0071 g,
0.019 mmol, 1 equiv) was treated with DBU (0.008 mL, 0.057
mmol, 3 equiv). Flash chromatography (SiO2, 1 × 7 cm, 50%
EtOAc-hexane with 5% Et3N) furnished 29 as a colorless oil
(0.0048 g, 0.0052 g theoretical, 90%). Recrystallization from
10% EtOAc-hexane gave colorless plates suitable for X-ray
(+)-(1R)-24: [R]25 +58 (c 0.044, EtOAc).38
D
ent-(-)-(1S)-24: [R]25 -56 (c 0.047, EtOAc).
D
7-(Be n zyloxy)-1-[((m e t h a n e su lfon yl)oxy)m e t h yl]-5-
[(5,6,7-tr im eth oxyin dol-2-yl)car bon yl]-1,2,3,4-tetr ah ydr o-
5H-n a p h th o[1,2-b]a zep in e (26). Following the general
procedure detailed for 16, 24 (0.040 g, 0.071 mmol, 1 equiv)
was treated with Et3N (0.049 mL, 0.354 mmol, 5 equiv) and
1
analysis:24 mp 259-261 °C ; H NMR (500 MHz, acetone-d6)
δ 8.13 (dd, J ) 8.0, 1.5 Hz, 1H), 7.63 (ddd, J ) 8.5, 7.5, 1.5

5862 J . Org. Chem., Vol. 62, No. 17, 1997
Boger and Turnbull
CH3SO2Cl (0.011 mL, 0.141 mmol, 2 equiv). Radial chroma-
tography (SiO2, 1 mm plate, 70% EtOAc-hexane) furnished
26 (0.034 g, 0.046 g theoretical, 75%) as a white solid: mp
1
175-177 °C with decomposition; H NMR (500 MHz, CDCl3)
major rotamer δ 9.30 (s, 1H), 8.49 (dd, J ) 8.0, 1.0 Hz, 1H),
8.25 (d, J ) 9.0 Hz, 1H), 7.68 (ddd, J ) 8.0, 6.5, 1.0 Hz 1H),
7.61-7.58 (m, 1H), 7.36-7.34 (m, 2H), 7.22-7.16 (m, 3H), 6.69
(s, 1H), 6.43 (s, 1H), 5.28 (d, J ) 2.0 Hz, 1H), 5.27-5.26 (m,
1H), 5.14 (d, J ) 11.5 Hz, 1H), 5.03-4.99 (m, 2H), 4.52-4.49
(m, 1H), 4.04 (s, 3H), 3.88 (s, 3H), 3.74 (s, 3H), 3.21-3.17 (m,
1H), 2.78 (s, 3H), 2.60-2.59 (m, 1H), 2.20-2.18 (m, 1H), 1.34-
1.32 (m, 3H); 13C NMR (125 MHz, CDCl3) major rotamer δ
160.7, 154.0, 149.6, 140.1, 139.9, 138.6, 136.2, 133.8, 128.7,
128.4 (2C), 127.9, 127.8, 127.1 (2C), 125.9, 125.7, 125.1, 124.9,
123.5, 123.3, 123.2, 108.6, 107.4, 97.8, 79.9, 70.2, 61.4, 61.1,
56.0, 41.4, 38.4, 36.7, 30.5, 16.9; IR (film) νmax 3453, 3299, 3001,
2929, 2827, 1608, 1588 cm-1; FABHRMS (NBA-CsI) m/z
777.1267 (M + Cs+, C35H36N2O6 requires 777.1247).
(+)-(1R)-26: [R]25 +26 (c 0.044, EtOAc).38
D
ent-(-)-(1S)-26: [R]25 -28 (c 0.042, EtOAc).
D
7-Hyd r oxy-1-[((m eth a n esu lfon yl)oxy)m eth yl]-5-[(5,6,7-
t r im et h oxyin d ol-2-yl)ca r b on yl]-1,2,3,4-t et r a h yd r o-5H -
n a p h th o[1,2-b]a zep in e (28). Following the general proce-
dure detailed for 17, 26 (0.024 g, 0.037 mmol, 1 equiv) was
treated 10% Pd-C (0.006 g, 0.15 equiv) followed by aqueous
HCO2NH4 (0.165 mL, 25% w/v, 18 equiv). Filtration and
concentration gave 28 as an analytically pure colorless oil in
quantitative yield (0.020 g, 0.020 g theoretical): 1H NMR (500
MHz, acetone-d6) major rotamer δ 10.17 (s, 1H), 9.20 (s, 1H),
8.37-8.34 (m, 2H), 7.68-7.56 (m, 2H), 6.66 (s, 1H), 6.51 (s,
1H), 5.42 (s, 1H), 5.23-5.22 (m, 1H), 4.95-4.92 (m, 1H), 4.60-
4.59 (m, 1H), 3.98 (s, 3H), 3.79 (s, 3H), 3.68 (s, 3H), 3.21-3.13
(m, 1H), 3.05 (s, 3H), 2.55-2.43 (m, 1H), 2.19-2.16 (m, 1H),
1.40-1.37 (m, 3H); 13C NMR (125 MHz, acetone-d6) major
rotamer δ 161.1, 153.4, 150.8, 141.3, 141.1, 139.8, 135.1, 130.6,
128.3, 126.0, 125.8 (2C), 125.0, 124.6, 124.3, 123.8, 110.7,
108.8, 98.7, 81.6, 61.4, 61.3, 56.3, 41.8, 38.1, 37.5, 31.2, 17.4;
IR (film) νmax 3313, 2936, 2836, 1731, 1592, 1584 cm-1
;
FABHRMS (NBA-CsI) m/z 687.0751 (M + Cs+, C28H30N2O8S
requires 687.0774).
(+)-(1R)-28: [R]25 +143 (c 0.036, THF).38
D
ent-(-)-(1S)-28: [R]25 -147 (c 0.036, THF).
D
N-[(5,6,7-Tr im eth oxyin dol-2-yl)car bon yl]-1,2,3,4,11,11a-
h exah ydr ocyclopr opa[c]n aph th o[2,1-b]azepin -6-on e (30,
CNA-TMI). Following the general procedure detailed for 10,
28 (0.0054 g, 0.001 mmol, 1 equiv) was treated with DBU
(0.004 mL, 0.030 mmol, 3 equiv). Flash chromatography (SiO2,
1 × 7 cm, 50% EtOAc-hexane with 5% Et3N) furnished 30 as
a yellow solid (0.0022 g, 0.0046 g theoretical, 48%). Recrys-
tallization from EtOAc gave yellow needles: mp 222-227 °C;
1H NMR (CD3CN, 400 MHz) major rotamer δ 9.67 (br s, 1H),
8.15 (dd, J ) 7.8, 1.3 Hz, 1H), 7.65-7.61 (m, 1H), 7.49-7.31
(m, 1H), 7.32 (d, J ) 8.1 Hz, 1H), 6.77 (s, 1H), 6.63 (s, 1H),
6.12 (s, 1H), 4.29 (ddd, J ) 12.0, 6.3, 5.3 Hz, 1H), 3.99 (s, 3H),
3.83 (s, 3H), 3.76 (s, 3H), 3.37 (ddd, J ) 13.0, 8.5, 7.0 Hz, 1H),
2.86-2.83 (m, 1H), 2.54-2.48 (m, 1H), 2.32-2.27 (m, 2H),
1.31-1.26 (m, 3H); IR (film) νmax 3292, 2923, 2841, 1626, 1600,
1523, 1456 cm-1; FABHRMS (NBA-CsI) m/z 459.1934 (M +
H+, C27H26N2O5 requires 459.1920).
F igu r e 11.
lized upon storage: mp 235-237 °C; 1H NMR (400 MHz,
acetone-d6) major rotamer δ 9.08 (s, 1H), 8.26 (d, J ) 8.3 Hz,
1H), 8.11 (d, J ) 8.5 Hz, 1H), 7.57-7.54 (m, 1H), 7.49-7.44
(m, 1H), 6.68 (s, 1H), 4.38-4.35 (m, 1H), 4.07-3.95 (m, 1H),
3.40-3.35 (m, 1H), 3.02-2.80 (m, 2H), 2.01-1.93 (m, 1H),
1.71-1.63 (m, 1H), 1.52-1.44 (m, 1H), 1.28 (s, 9H), 1.20-1.17
(m, 2H); IR (film) νmax 3309, 2977, 2927, 2867, 1689, 1664,
1619, 1589 cm-1; FABHRMS (NBA-CsI) m/z 476.0848 (M +
Cs+, C20H25NO4 requires 476.0838).
Acid -Ca ta lyzed Ad d ition of H2O to (+)-10: (-)-6-(ter t-
Bu t yloxyca r b on yl)-2,8-d ih yd r oxy-1,2,3,4,5-p en t a h yd r o-
6H-n aph th o[1,2-b]azocin e (31). A solution of (+)-10 (0.0030
g, 0.009 mmol, 1 equiv) in THF (0.38 mL) was treated with
H2O (0.13 mL) followed by CF3SO3H (0.011 mL, 0.1 M in THF,
0.12 equiv) at 25 °C, and the mixture was stirred for 5 min.
Workup as described for racemic 31 and radial chromatogra-
phy (SiO2, 1 mm plate, 50% EtOAc-hexane) provided (-)-31
as a colorless oil (0.0032 g, 0.0032 g theoretical, 99%) which
slowly crystallized upon storage. This material was identical
(-)-30: [R]25 -146 (c 0.004, THF).38
D
ent-(+)-30: [R]25 +146 (c 0.004, THF).
D
Acid -Ca t a lyzed Ad d it ion of H2O t o 10: 6-(ter t-
Bu t yloxyca r b on yl)-2,8-d ih yd r oxy-1,2,3,4,5-p en t a h yd r o-
6H-n a p h th o[1,2-b]a zocin e (31). A solution of 10 (0.0074
g, 0.022 mmol, 1 equiv) in THF (0.75 mL) was treated with
H2O (0.25 mL) followed by CF3SO3H (0.027 mL, 0.1 M in THF,
0.12 equiv) at 25 °C, and the mixture was stirred for 5 min.
The reaction mixture was treated with NaHCO3 (0.01 g)
followed by H2O (1.0 mL). The aqueous portion was extracted
with EtOAc (2 × 1 mL), and the combined organic portions
were dried (MgSO4), filtered, and concentrated to a colorless
oil. 1H NMR analysis of the crude mixture and comparison
with 33 indicated the presence of the ring expansion solvolysis
product exclusively (g40-20:1). Radial chromatography (SiO2,
1 mm plate, 50% EtOAc-hexane) provided 31 as a colorless
oil (0.0077 g, 0.0078 g theoretical, 99%) which slowly crystal-

Synthesis and Evaluation of CC-1065 and Duocarmycin Analogs
J . Org. Chem., Vol. 62, No. 17, 1997 5863
with racemic 31 in all respects. The solvolysis of (+)-10
1.68-1.67 (m, 1H), 1.43-1.41 (m, 1H), 1.29 (s, 9H); IR (film)
provided a single enantiomer (Figure 4) established by chiral
phase HPLC separation on a ChiralCel OG column (10 µm,
0.46 × 25 cm, 5% i-PrOH-hexane, 1 mL/min).
ν
max 3251, 2974, 2933, 1692, 1662, 1621, 1585 cm-1; FABHRMS
(NBA-NaI) m/z 384.1333 (M + Na+, C20H24CINO3 requires
384.1342).
(-)-(2R)-31: [R]25 -9 (c 0.009, THF).
Aqu eou s Solvolysis of N-BOC-CNA (10) a n d CNA (11).
Samples of 10 (150 µg) and 11 (150 µg) were dissolved in
CH3OH (1.5 mL), and the resulting solutions were mixed with
aqueous buffer (pH 3, 1.5 mL, 4:1:20 (v:v:v) 0.1 M citric acid,
0.2 M Na2HPO4, and deionized H2O, respectively). Similarly,
samples of 10 (150 µg) and 11 (150 µg) in CH3OH (1.5 mL)
were mixed with deionized H2O (pH 7, 1.5 mL). The UV
spectra of the solutions were measured immediately after
mixing with the appropriate aqueous solution. The blank and
the solvolysis reaction solutions were stoppered, protected from
light, and allowed to stand at 25 °C. The total reaction times
reflect those required to observe no further change in absor-
bance. For the solvolysis of 10 at pH 3, the UV spectrum was
taken every 30 s for 10 min and then every 3 min for the next
40 min. The decrease of the absorbance at 319 nm was
monitored. The solvolysis rate was calculated from the least
squares treatment (r2 ) 0.998) of the slope of a plot of time
D
5-(ter t-Bu tyloxyca r bon yl)-7-h yd r oxy-1-(h yd r oxym eth -
yl)-1,2,3,4-tetr ah ydr o-5H-n aph th o[1,2-b]azepin e (33). Fol-
lowing the general procedure detailed for 17, 15 (0.015 g, 0.035
mmol, 1 equiv) was treated with 10% Pd-C (0.001 g, 0.3 equiv)
followed by aqueous HCO2NH4 (0.2 mL, 25% w/v, 23 equiv).
Filtration and concentration gave 33 as an analytically pure
colorless oil in quantitative yield (0.012 g, 0.012 g theoretical)
which crystallized upon storage: mp 202-203 °C; 1H NMR
(400 MHz, CDCl3) major rotamer δ 9.19 (s, 1H), 8.07-8.05 (m,
1H), 7.26-7.22 (m, 1H), 6.81-6.75 (m, 2H), 6.50 (s, 1H), 4.28-
4.24 (m, 1H), 4.18-4.08 (m, 2H), 3.89-3.86 (m, 1H), 2.85-
2.79 (m, 1H), 2.11-2.07 (m, 1H), 1.94-1.91 (m, 1H), 1.70-
1.67 (m, 2H), 1.62 (s, 9H); 1H NMR (400 MHz, acetone-d6)
major rotamer 9.03 (s, 1H), 8.26 (d, J ) 8.2 Hz, 1H), 8.07 (d,
J ) 8.6 Hz, 1H), 7.55-7.51 (m, 1H), 7.46-7.43 (m, 1H), 6.74
(s, 1H), 4.37-4.33 (m, 1H), 4.03-3.99 (m, 1H), 3.91-3.80 (m,
1H), 3.61-3.52 (m, 2H), 2.73-2.66 (m, 1H), 2.36-2.32 (m, 1H),
2.20-2.10 (m, 1H), 1.60-1.55 (m, 1H), 1.33 (s, 9H); IR (film)
νmax 3272, 2976, 2925, 1662, 1621, 1596 cm-1; FABHRMS
(NBA-NaI) m/z 344.1855 (M + H+, C20H25NO4 requires
344.1862).
versus ln[(Af - Ai)/(Af - A)] (Figure 11); k ) 6.90 × 10-3 s-1
,
t1/2 ) 1.7 min, 0.028 h. For the solvolysis of 10 at pH 7, the
UV spectrum was monitored every 5 min for 19 h. The rate
was calculated from the least squares treatment of the same
type of plot as above (r2 ) 0.999); k ) 9.16 × 10-5 s-1, t1/2
)
Acid -Ca ta lyzed Ad d ition of CH3OH to 10: 6-(ter t-
Bu tyloxyca r bon yl)-8-h yd r oxy-2-m eth oxy-1,2,3,4,5-p en -
ta h yd r o-6H-n a p h th o[1,2-b]a zocin e (32). A solution of 10
(0.0069 g, 0.021 mmol, 1 equiv) in freshly distilled CH3OH (1.0
mL) was treated with CF3SO3H (0.027 mL, 0.1 M in CH3OH,
0.12 equiv) at 25 °C for 5 min. Workup as described for 31
and radial chromatography (SiO2, 1 mm plate, 50% EtOAc-
hexane) provided 32 as a colorless oil (0.0070 g, 0.0076 g
theoretical, 92%) which slowly crystallized upon storage: mp
199-200 °C with decomposition; 1H NMR (500 MHz, acetone-
d6) major rotamer δ 9.09 (s, 1H), 8.27 (d, J ) 8.5 Hz, 1H), 8.11
(d, J ) 8.5 Hz, 1H), 7.59-7.56 (m, 1H), 7.49-7.47 (m, 1H),
6.69 (s, 1H), 4.37-4.29 (m, 1H), 3.57-3.56 (m, 1H), 3.49-3.46
(m, 1H), 3.43 (s, 3H), 2.85-2.79 (m, 2H), 1.99-1.90 (m, 1H),
1.81-1.75 (m, 1H), 1.28 (s, 9H overlapping with m, 2H); IR
(film) νmax 3268, 2971, 2868, 1695, 1664, 1587 cm-1; FABHRMS
(NBA-CsI) m/z 357.1951 (M + Cs+, C21H27NO4 requires
357.1940).
Ad d ition of HCl to 10: 6-(ter t-Bu tyloxyca r bon yl)-2-
ch lor o-8-h ydr oxy-1,2,3,4,5-pen tah ydr o-6H-n aph th ol[1,2-
b]a zocin e (34). A solution of 10 (0.0031 g, 0.0095 mmol, 1
equiv) in THF (0.31 mL) was treated with 3.3 M HCl-THF
(0.0043 mL, 1.5 equiv) at -78 °C for 5 min. Evaporation of
the volatiles and chromatography (SiO2 1 × 7 cm, 30% EtOAc-
hexane) furnished 34 as a pure white foam (0.0034 g, 0.0034
g theoretical, 100%): 1H NMR (400 MHz, acetone-d6) major
rotamer δ 9.27 (br s, 1H), 8.27 (d, J ) 8.5 Hz, 1H), 8.06 (d, J
) 8.5 Hz, 1H), 7.63-7.59 (m, 1H), 7.57-7.49 (m, 1H), 6.72 (s,
1H), 4.49-4.43 (m, 1H), 3.67-3.63 (m, 1H), 3.24-3.21 (m, 1H),
2.88-2.82 (m, 2H), 2.02-1.90 (m, 1H), 1.81-1.75 (m, 1H),
126 min, 2.1 h. For the solvolysis of 11 at pH 3, the UV
spectrum was monitored every 2 min for the first 110 min and
then every 10 min for the next 4 h. The decrease in the
absorbance at 345 nm was recorded. The solvolysis rate
constant was calculated from the least squares treatment of
the same type of plot as above (r2 ) 0.999); k ) 3.10 × 10-4
s-1, t1/2 ) 37 min, 0.62 h. For the solvolysis of 11 at pH 7, the
UV spectrum was taken every 24 h for 21 d. The rate was
calculated from the least squares treatment of the same type
of plot as above (r2 ) 0.999); k ) 3.42 × 10-7 s-1, t1/2 ) 563 h.
Ack n ow led gm en t. We gratefully acknowledge the
financial support of the National Institutes of Health
(CA55276) and the Skaggs Institute for Chemical Biol-
ogy. We thank Dr. P. Me´sini and N.-E. Haynes for
preliminary studies (eq 2), Dr. Raj K. Chadha for the
X-ray structure determinations, and Dr. Qing J in for
the in vitro cytotoxicity evaluations summarized in
Table 4.
Su p p or tin g In for m a tion Ava ila ble: Full experimental
details for the preparation of substrates found in eqs 1 and 2,
summary tables of the data employed in Figure 10, and 1H
NMR spectra of 10, 11, 13-33 (31 pages). This material is
contained in libraries on microfiche, immediately follows this
article in the microfilm version of the journal, and can be
ordered from the ACS; see any current masthead page for
ordering information.
J O9707085