Evaluation of alkene isomerization as a trigger for enediyne activation

Article abstract of DOI:10.1016/S0040-4039(01)02139-6

The concept of non-conjugated-to-conjugated double bond isomerization as a triggering mechanism for a calicheamicin model was tested. A bicyclo[7.3.1] enediyne framework was prepared with a bridgehead double bond and an acetonyl side chain. Base-promoted exchange failed to produce the conjugated isomer. Computational analysis suggested that additional conjugating groups would favor the isomerization, but experiment proved that not to be correct.

Full text of DOI:10.1016/S0040-4039(01)02139-6

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 43 (2002) 541–544
Evaluation of alkene isomerization as a trigger for
enediyne activation
M. F. Semmelhack,* Richmond Sarpong, Jeffrey Bergman and Douglas M. Ho†
Department of Chemistry, Princeton University, Princeton, NJ 08544, USA
Received 10 October 2001; accepted 22 October 2001
Abstract—The concept of non-conjugated-to-conjugated double bond isomerization as a triggering mechanism for a calicheamicin
model was tested. A bicyclo[7.3.1] enediyne framework was prepared with a bridgehead double bond and an acetonyl side chain.
Base-promoted exchange failed to produce the conjugated isomer. Computational analysis suggested that additional conjugating
groups would favor the isomerization, but experiment proved that not to be correct. © 2002 Published by Elsevier Science Ltd.
1. Introduction
2. Results and discussion
The removal of a bridgehead double bond in the
calicheamicin/esperamicin family of enediyne toxins is
a primary step in the activation toward arene-1,4-diyl
formation, H-atom abstraction, and DNA cleavage.1
With the natural toxins very high toxicity is observed
but low selectivity.2 It has been challenging to design
functional analogs of the enediynes which preserve
the toxicity, but take advantage of alternative trigger-
ing mechanisms which might be tailored for useful
selectivity. In a simple analog of the calicheamicin
structural type, one might consider the endo-to-exo
isomerization of the alkene unit in a bicyclo[7.3.1]
framework (i.e. 1“2) as the basis of a new triggering
mechanism. It is proposed that the thermodynamic
(relative stability of 1 versus 2) and kinetic features
(i.e. rate of rearrangement under physiological condi-
tions) can be tuned with the substituent Y to provide
chemical control for activation at the target site
(Scheme 1).
Our initial target was compound 4, which bears the
propargylic hydroxyl group appropriate for the stan-
dard methods of ring closure to form the cyclic
enediyne, and important as a point of attachment for
side chains to optimize DNA binding. It also bears a
ketal group or equivalent, which could be unmasked
to initiate the endo–exo alkene isomerization via eno-
lization–ketonization. The ketone 5 would then equili-
brate with enone 6 (Scheme 2).
Even with an unfavorable equilibrium, the further
steps through 7 are expected to be essentially irre-
versible. The other structural features of 4 result from
our synthesis strategy in a related system.3 Based on
the earlier work, enone 6 is predicted to have a half-
lifetime for cycloaromatization of 30–60 min at 37°C.
The synthesis is shown in Scheme 3. A Diels–Alder
reaction of Danishefsky’s diene with methacrolein fol-
H
Y
H
Y
Y
H
Y
H
fast?
H•
2
3
1
Scheme 1. endo–exo Isomerization as a trigger.
Keywords: enediynes; alkene isomerization; enol–keto isomerization.
* Corresponding author. Tel.: 609 258-5501; fax: 609 258-3409; e-mail: mfshack@princeton.edu
† Director, Princeton Small Molecule X-ray Facility.
0040-4039/02/$ - see front matter © 2002 Published by Elsevier Science Ltd.
PII: S0040-4039(01)02139-6

542
M. F. Semmelhack et al. / Tetrahedron Letters 43 (2002) 541–544
O
X
O
O
X
OR
OH
OR
OR
2
Keq
?
12
10
•
5
9
6
•
4
5
7
6
Scheme 2. The ketal/ketone target.
OTf
O
OMe
+
CO2Me
c
CO2Me
d
CHO
CO2Me
OH
a
b
TMSO
O
O
O
O
O
8
11
9
10
O
O
e
O
O
O
O
O
O
O
O
O
O
OH
CO2Me
OH
CO2Me
CO2Me
i
f
h
g
H
15
14
13
12
16
j
O
O
O
OTBS
OTBS
k
18
17
Scheme 3. Synthesis of the ketal/ketone pair. (a) i. D, benzene, ii. HOCH2CH2OH, 0.10 mol equiv. CuCl2, 60% (two steps); (b)
i. LiHMDS, HMPA, MeOCOCN, 84%, ii. NaH, PhN(Tf)2, 91%; (c) Allyl2Cu(CN)Li2, 71%; (d) i. 30% HOAc, ii. 0.37 mol equiv.
NaBH4 in EtOH, 76%; (e) i. 0.3 mol equiv. PdCl2, 0.04 mol equiv. CuCl, DMF, H2O, O2, ii. HOCH2CH2OH, 0.10 mol equiv.
t
PPTS, 70% (two steps); (f) i. Dess–Martin, ii. (MeO)2POCHN2, BuOK, 57% (two steps); (g) i. 0.20 mol equiv. CuI, 0.05 mol
equiv. Pd(PPh3)4, (Z)-ClCHCHCCTMS, ii. K2CO3 in MeOH, 84% (two steps); (h) i. DIBALH, ii. RuCl2(PPh3)3, N-methylmor-
pholine N-oxide, 74% (two steps); (i) 2.5 mol equiv. CeCl3, 25 mol equiv. LiHMDS, 79%; (j) 2,6-lutidine, TBStriflate, 77%; (k)
2 mol equiv. CuCl2·2H2O in CH3CN, 85%.
lowed by acid treatment and selective protection of the
aldehyde group furnishes the enone–ketal 8 in 60%
overall yield.4 Generation of the enolate of 8 followed
by trapping with methyl cyanoformate5 gave an enol
ester which was treated with N-phenyl trifluoro-
methanesulfonimide6 to give the triflate 9 in 76% yield
overall. Selective allyl coupling with a higher order
cuprate7 converted 9 to 10 in 71% yield. The acetal unit
was hydrolyzed and the resulting aldehyde group was
reduced to the alcohol (in 11) as a protecting group
strategy (76% yield overall). Then selective Wacker
oxidation of the terminal alkene followed by protection
of the resulting ketone as the ethylene ketal gave 12 in
70% yield. The alcohol was re-oxidized with Dess–Mar-
tin periodinane8 and the aldehyde unit was converted to
the alkyne with dimethyl diazomethylphosphonate9
(57%, two steps). Sonogashira coupling10 of the alkyne
with 1-chloro-4-trimethylsilyl-(Z)-but-1-en-3-yne fol-
lowed by desilylation gave the acyclic enediyne 14 in
84% yield. A reduction/re-oxidation protocol was opti-
mal for preparation of the precursor (15, 74%) for ring
closure. Ring closure to give 16 was accomplished in
79% yield using high dilution conditions in THF with
Ce(III) and a large excess of lithium hexamethyldisil-
azide.11 The overall yield to this point in Scheme 3 is
about 5%, and the procedures can be carried out on
multi-gram scale. The structure of 16 was confirmed by
X-ray diffraction analysis.12 Protection of the alcohol as
the tert-butyldimethylsilyl derivative (17) was useful for
further manipulations. Deprotection of the ketal unit to
expose the ketone functionality proved delicate, but
treatment with CuCl2·2H2O (20 min, 23°C) produced
18 in 85% yield.
The isolation of the non-conjugated isomer 18 was
surprising in light of the acidic conditions employed in
the deprotection, and suggested a thermodynamic pref-
erence for the non-conjugated ketone. This preference
was established by treatment of 18 with NaOCD3 in
CD3OD until deuterium introduction was complete
(giving 20 without a trace of a conjugated analog, e.g.
21). It is also supported by computational analysis,
using a comparison of the DHf values obtained by AM1
calculations (Table 1, compare structures 22a and 23a,
R=H).13 The non-conjugated isomer 22a is favored by
1.9 kcal/mol. A more elaborate calculation using den-
sity functional theory reversed the preference, and sug-
gested 23a is the more stable by 1.7 kcal/mol.14 The

M. F. Semmelhack et al. / Tetrahedron Letters 43 (2002) 541–544
543
Table 1. AM1 estimates of relative stability of conjugated and non-conjugated isomers (P=H)
O
O
R
R
OP
OP
OP
OP
13
10
∆∆Hf
11
∆∆Hf
R
R
(kcal/mol)
(kcal/mol)
22
23
24
25
a. H
-0.2
+0.2
+2.0
+0.5
+4.6
a. H
+1.9
-2.9
0
-2.0
-3.4
b. SPh
c. CO2Me
d. Ph
b. SPh
c. CO2Me
d. Ph
O
O
R
R
e. NO2
e. NO2
computations suggest at best a small difference (Scheme
4).
no change in structure other than introduction of four
D atoms.
In an effort to develop derivatives more likely to favor
the isomer with the exocyclic double bond, similar
calculations were carried out with groups at C13 (R in
22, Table 1) expected to have a stabilizing interaction
with the double bond. In addition, the effect of the
double bond at C10ꢀC11 was considered. Table 1 dis-
plays the derivatives evaluated and gives the DDHf
between the non-conjugated isomer (e.g. 22, 24) and the
most stable conformer of the conjugated isomer (23,
25). For the cases 22b–e, the calculations suggest that
the substituent is predicted to favor (or not disfavor, as
for 22c) the exocyclic isomer.
While the AM1 calculations (Table 1) did not favor
rearrangement for the dihydro versions, 24, the test
cases were readily synthesized and were evaluated. A
modified synthesis starting by reduction16 of 8 to give
26 ran parallel with that in Scheme 3. The ketone 24a
was obtained by ketal hydrolysis, and the addition of a
carbomethoxy group proceeded as before, to give
ketoester 24c. Again, the nonconjugated isomers 24a
and 24c showed no tendency toward alkene isomeriza-
tion while undergoing proton exchange in basic
CD3OD (Scheme 5).
In a preliminary set of experiments, we evaluated an
alternative mechanism for dislodging the double bond
from the bridgehead position. The epoxide was
obtained in essentially quantitative yield by reaction of
16 with m-chloroperbenzoic acid,17 and protection gave
28. We anticipated that ketal hydrolysis (to 29) would
lead to ring opening of the epoxide to give a structure
activated toward cycloaromatization. However, under
all conditions tested, the solvolysis of the epoxide was
faster than ketal hydrolysis and the proposed product
(30) underwent diradical formation (to 31) rapidly
(25°C). For example, CuCl2/H2O for 20 min at 25°C (in
the presence of excess 1,4-cyclohexadiene) gave the
cycloaromatized diol 31 in 62% yield, based on 33%
To test this analysis, 22b and 22c were prepared as the
TBS ethers by direct functionalization of 18. Treatment
of 18 with lithium hexamethyldisilylazide (LiHMDS)
generated the enolate at C13, and reaction with S-
phenylbenzenethiosulfonate15 gave 22b (TBS ether) in
59% yield, while reaction with methyl cyanoformate5
gave 22d (TBS ether) in 93% yield. Efforts to induce
rearrangement to the conjugated isomers (23) using
NaOCD3 in CD3OD failed, as the acidic protons adja-
cent to the ketone group simply exchanged without
structural rearrangement; this is clearly inconsistent
with the results of the calculations, The products were
1
isolated and shown by H NMR, IR and TLC to have
O
O
O
O
D
D
OTBS
OTBS
OTBS
OTBS
D
D3C
NaOCD3
CD3OD
not:
18
20
19
21
Scheme 4. Attempted ketone–enone rearrangement.
O
O
CO2Me
OTBS
CO2Me
CO2Me
NaOCD3
O
O
O
O
O
OTBS
OTBS
24c
as in
Scheme 3
CO2Me
NaOCD3
O
O
24a
26
27
Scheme 5. Preparation and evaluation of the dihydro series.

544
M. F. Semmelhack et al. / Tetrahedron Letters 43 (2002) 541–544
O
O
O
O
O
OH
OTBS
OH
OTBS
OH
OTBS
OTBS
O
c
a,b
•
16
HO
HO
HO
•
30
29
31
28
Scheme 6. Activation via epoxidation. (a) mCPBA, 0°C; (b) TBSOTf, lutidine; (c) CuCl2, H2O, 25°C.
recovery of the starting epoxide 28. This result confirms
the high reactivity of this framework toward cyclo-
aromatization, and suggests the basis for a chemically
triggerable derivative (Scheme 6).
5. Mander, L.; Sethi, S. P. Tetrahedron Lett. 1983, 24, 5425.
6. McMurry, J.; Scott, W. Tetrahedron Lett. 1983, 24, 979.
7. Lipshutz, B. H.; Elworthy, T. J. Org. Chem. 1990, 55,
1695.
8. Dess, D. B.; Martin, J. C. J. Am. Chem. Soc. 1991, 113,
7277.
9. Gilbert, J.; Weerasooriya, U. J. Org. Chem. 1979, 44,
4997.
Acknowledgements
10. Kende, A.; Smith, C. Tetrahedron Lett. 1988, 29, 4217.
11. Myers, A.; Harrington, P.; Kuo, E. J. Am. Chem. Soc.
1991, 113, 694.
We thank the National Institutes of Health for financial
support (CA 54819).
12. Results of the X-ray structure determination were
deposited at the Cambridge Crystallographic Database
(CCDC 175664).
References
13. The calculations were done with MacSpartan Pro. The
results are complicated by the presence of E/Z isomers
about the exocyclic double bond in 23. Where relevant,
the isomer with lowest DHf was identified and used in the
comparison.
14. B3LYP/6-31G(d). We thank Professor Robert Pascal,
Princeton University, for these calculations.
15. For the general procedure, see: Hiroi, K.; Matsuda, Y.;
Sato, S. Chem. Pharm. Bull. 27, 1979, 2338. Our example
was more efficient with LiHMDS in place of NaH for
enolate generation.
1. For reviews, see: (a) Grissom, J. W.; Gunawardena, G.
U.; Klingberg, D.; Huang, D. Tetrahedron 1996, 52, 6453;
(b) Lhermitte, H.; Grierson, D. S. Contemp. Org. Synth.
1996, 3, 41 and 1996, 3, 93.
2. (a) Myers, A. G.; Cohen, S. B.; Kwon, B. M. J. Am.
Chem. Soc. 1994, 116, 1255; (b) Walker, S.; Landovitz,
R.; Ding, W.-d.; Ellestad, G. A.; Kahne, D. Proc. Natl.
Acad. Sci. USA 1992, 89, 4608; (c) Sievers, E. L.; Bern-
stein, I. D.; Spoielberger, R. T.; Forman, S. J.; Shannon-
Dorcy, K.; Appelbaum, F. R. Eur. J. Cancer. 1997, 33,
398.
16. Tsuda, T.; Hayashi, T.; Satomi, H.; Kawamoto, T.; Sae-
gusa, T. J. Org. Chem. 1986, 51, 537.
17. Mastalerz, H.; Doyle, T. W.; Kadow, J. F.; Vyas, D. M.
Tetrahedron Lett. 1996, 37, 8683 and 8687.
3. Semmelhack, M. F.; Gallagher, J. J.; Minami, T.; Date,
T. J. Am. Chem. Soc. 1993, 115, 11618.
4. Danishefsky, S.; Kitahara, T.; Yan, C.; Morris, J. J. Am.
Chem. Soc. 1979, 101, 6996.