afford the C2–C3 unsaturated C2-aryl PBD 3. In the new
synthetic strategy the triflation/Suzuki coupling reactions are
carried out prior to B-ring closure (Scheme 2) thus avoiding the
dilactam intermediate and the risk of over-reduction to the
biologically inactive N10–C11 secondary amine.
C11 secondary amine formation and allowing reduction-
sensitive groups to be included on the C2-aryl ring. Fur-
thermore, a convenient, new, and high yielding oxidation
method has been discovered for B-ring cyclization, which
should be applicable to other PBD systems.
Preliminary modelling studies suggest that there is an
appreciable difference in C-ring conformation between the 2,3-
and 1,2-unsaturated structures. In the former case, the molecule
has a crescent shape with each of the four rings closely
following the curve. However, with the new 1,2-unsaturated
arrangement reported here, the molecule is more linear in shape
with the component rings much less aligned, suggesting that it
does not fit so well into the minor groove of DNA. Comparison
of the biophysical and biological properties of the two structures
is presently underway and the results will be reported
elsewhere.
The PBD backbone was established by coupling 4,5-dime-
thoxy-2-nitrobenzoic acid to the amine 4 derived from 4- -
L
trans-hydroxyproline9 in the presence of DCC/HOBt in 70%
yield. The critical pro-C2 ketone (6) was obtained by oxidising
the 2-hydroxy group with BAIB/TEMPO in 90% yield.10 At
this stage a number of triflation strategies were explored.
Trifluoromethanesulfonic anhydride with pyridine failed to
give appreciable amounts of enol triflate with little consumption
of the starting ketone. The use of LDA and 5-chloro-N,N-
(trifluoromethanesulfonyl)aminopyridine11 afforded a mixture
of 1,2 (8) and 2,3 (7) unsaturated enol triflates in a 4 : 1 ratio.
However, use of the same triflating agent with NaHMDS as
base afforded almost exclusively the 1,2-unsaturated product 8
in 50% yield. The NMR spectra of the isomeric triflates were
quite distinct with the two H3 signals clearly visible in the
spectrum of 8.12 The 1,2-product 8 was subjected to Suzuki
coupling conditions in the presence of 4-methoxybenzene-
boronic acid. The reaction proceeded in excellent yield ( > 80%)
but required heating unlike the Suzuki reactions involving the
dilactam-based enol triflate reported previously.7 The nitro
group was then reduced with sodium dithionite in 80% yield
without jeopardising the C1–C2 double bond. The newly
formed amine 10 was protected as an allyl (Alloc) carbamate
(11) and the TBDMS group removed (12) in order to permit a
Fukuyama B-ring cyclization. Normally, this reaction is
performed under Swern conditions, but we found that oxidation
with BAIB/TEMPO10 achieved B-ring closure to 13 in 84%
yield. This is a significant improvement as, unlike the Swern
reaction, the TEMPO/BAIB method does not require anhydrous
conditions or an inert atmosphere, thus making the cyclization
easier to perform and monitor. The 1H spectrum of 13 revealed
the diagnostic H-11 signal at 5.7 ppm. Removal of the allyl
carbamate group under Deziel conditions9 afforded the novel
C2-aryl substituted 1,2-unsaturated PBD 1413 in 80% yield.14
The NMR spectrum of 14 was compared to that of the
previously reported 3. In addition to the expected aromatic and
methoxy signals, it revealed a doublet at d 7.82 for the 11-H
imine proton, a singlet for 1-H at 6.19 with some evidence of
fine coupling, two overlapping doublets at 4.81–4.78 for 3-H/HA
and a broad singlet for H-11a at 4.63 ppm.15 The orientation of
C-ring unsaturation was confirmed by a 2D proton–proton
NMR experiment. The diagnostic H-11 imine signal correlated
with the multiplet at 4.63 ppm, revealing it as H-11a. This signal
correlated in turn with the alkenic multiplet at 6.19 ppm
confirming it as H-1 and supporting assignment of C1–C2
unsaturation. These experiments showed that the double bond
orientation set-up in the enol triflate forming reaction was not
isomerised by subsequent steps involving palladium catalysis.
In conclusion, altering the sequence of synthetic events by
triflating prior to B-ring closure affords a novel type of C2-aryl
substituted PBD with 1,2-endo unsaturation, thus complement-
ing the C2–C3 unsaturated compounds produced by the reverse
sequence of events.7 A further advantage is that this synthetic
route avoids a reduction step16 thus removing the risk of N10–
Notes and references
1 D. E. Thurston, Advances in the Study of Pyrrolo[2,1-c][1,4]benzodia-
zepine (PBD) Antitumour Antibiotics, in Molecular Aspects of Anti-
cancer Drug–DNA Interactions, ed. S. Neidle and M. J. Waring, The
MacMillan Press Ltd., London, 1993, pp. 54–88.
2 D. E. Thurston, Br. J. Cancer, 1999, 80, 65.
3 D. E. Thurston and D. S. Bose, Chem. Rev., 1994, 94, 433.
4 B. S. P. Reddy, Y. Damayanthi, B. S. N. Reddy and J. W. Lown, Anti-
Cancer Drug Design, 2000, 15, 225.
5 A. Kamal, N. Laxman, G. Ramesh, K. Neelima and A. K. Kondapi,
Chem. Commun., 2001, 437.
6 S. J. Gregson, P. W. Howard, J. A. Hartley, N. A. Brooks, L. J. Adams,
T. C. Jenkins, L. R. Kelland and D. E. Thurston, J. Med. Chem., 2001,
44, 737.
7 N. Cooper, D. R. Hagan, A. Tiberghien, T. Ademefun, C. S. Matthews,
P. W. Howard and D. E. Thurston, Chem. Commun., 2002, 1764.
8 The simple C1–C2 unsaturated C2-unsubstituted parent compound is
known in the literature (see ref. 16) but the synthetic route used is
restricted to this parent.
9 S. J. Gregson, P. W. Howard, K. E. Corcoran, S. Barcella, M. M. Yasin,
A. A. Hurst, T. C. Jenkins, L. R. Kelland and D. E. Thurston, Bioorg.
Med. Chem. Lett., 2000, 10, 1845.
10 A. DeMico, R. Margarita, L. Parlanti, A. Vescovi and G. Piancatelli, J.
Org. Chem., 1997, 62, 6974.
11 K. Ritter, Synthesis (Stuttgart), 1993, 735.
12 The 2,3-unsaturated product 7 exhibited spin systems for H-1 and H-1A
at 2.9 and 3.1 ppm, and an H-3 signal was observed at 6.0 ppm. In
contrast, the 1,2-unsaturated compound 8 revealed two 3-H protons at
4.1 and a H-1 alkenic signal at 5.85 ppm.
13 Selected data for 14: (250 MHz, CDCl3): d 7.82 (d, 1H, J = 4.1, H11),
7.56 (s, 1H, H6), 7.43 (d, 2H, J = 8.7, H12), 6.93 (d, 2H, J = 8.7, H13),
6.88 (s, 1H, H9), 6.19 (m, 1H, H1), 4.81–4.78 (m, 2H, H3), 4.63 (m, 1H,
H11a), 3.98 (s, 3H, 8-OCH3), 3.94 (s, 3H, 7-OCH3), 3.84 (s, 3H,
14-OCH3); m/z (ES+): 383 ([M·+ carbinolamine form, 65%), 365 ([M +
H]·+, imine form, 100%); IR (CHCl3): 3307, 2931, 2842, 1643, 1604,
1513, 1454, 1433, 1380, 1254, 1218, 1179, 1098, 1033, 911, 827, 768,
728 cm21
.
14 Exposure of 14 to heat during work-up should be avoided as this can
provoke C-ring aromatization.
15 In contrast the spectrum of 3 clearly shows two distinct spin systems for
H1 and H1A at 3.4 and 3.6 ppm, and the diagnostic enamide 3-H at 7.4
ppm.
16 L. H. Hurley, T. Reck, D. E. Thurston, D. R. Langley, K. G. Holden, R.
P. Hertzberg, J. R. E. Hoover, G. Gallagher, L. F. Faucette, S. M. Mong
and R. K. Johnson, Chem. Res. Toxicol., 1988, 1, 258.
CHEM. COMMUN., 2003, 1688–1689
1689