A direct and stereocontrolled route to conjugated enediynes

Article abstract of DOI:10.1021/ja993766b

A unified synthetic route to 3-hex-en-1,5-diynes, a key building block found in many of the enediyne antitumor agents and designed materials, was developed. The method, which relies on a carbenoid coupling-elimination strategy is tolerant of a wide range of functionalities, and was applied to the synthesis of a variety of linear and cyclic enediynes. Reaction parameters can be adjusted to control stereoselectivity of the process, producing linear enediynes from 1:12 to >100:1 E:Z ratio, and in the case of cyclic enediynes, giving the exclusively Z C-9, C-10, or C-11 products. Key features of the process are the ready availability of precursors and the mildness and efficiency of the reaction. Application of the process in the design of materials precursors and preparation of enediyne antitumor agents are presented.

Full text of DOI:10.1021/ja993766b

J. Am. Chem. Soc. 2000, 122, 1937-1944
1937
A Direct and Stereocontrolled Route to Conjugated Enediynes
Graham B. Jones,* Justin M. Wright, Gary W. Plourde, II, George Hynd,
Robert S. Huber, and Jude E. Mathews
Contribution from the Bioorganic and Medicinal Chemistry Laboratories, Department of Chemistry,
Northeastern UniVersity, Boston, Massachusetts 02115
ReceiVed October 20, 1999
Abstract: A unified synthetic route to 3-hex-en-1,5-diynes, a key building block found in many of the enediyne
antitumor agents and designed materials, was developed. The method, which relies on a carbenoid coupling-
elimination strategy is tolerant of a wide range of functionalities, and was applied to the synthesis of a variety
of linear and cyclic enediynes. Reaction parameters can be adjusted to control stereoselectivity of the process,
producing linear enediynes from 1:12 to >100:1 E:Z ratio, and in the case of cyclic enediynes, giving the
exclusively Z C-9, C-10, or C-11 products. Key features of the process are the ready availability of precursors
and the mildness and efficiency of the reaction. Application of the process in the design of materials precursors
and preparation of enediyne antitumor agents are presented.
Introduction
Scheme 1. Proposed Strategy for In Situ Enediyne Synthesis
Conjugated 3-ene-1,5-diynes represent important chemical
building blocks, the subunit found in systems as diverse as
antitumor antibiotics, synthetic polymers, and designed nano-
structures. Examples include the recently discovered class of
antitumor agents which contain cyclic C-10 or C-9 enediyne
subunits,1 exemplified by calicheamicinone 1, cross-conjugated
macrocycles including radialene 2,2 and polyacetylenes including
the carbon-rod polydiacetylenes 3.3
a direct route to the core structures 7 and 8 from readily available
precursors, and envisaged monometallocarbenoid species 5,
derived in turn from 4, that under appropriate circumstances
might be encouraged to participate in an inter/intramolecular
addition to give halodiynes 6 (Scheme 1). E2 elimination would
then give access to enediynes 7/8 directly. Such a method could
provide unified access to both linear and cyclic systems, and
vinyl group stereochemistry could be influenced by geometric
constraints imposed in 6. Although generation of such metal-
lohalocarbenoids has precedent, thermal decomposition might
be expected to limit their application.5 Conditions that would
permit coupling followed by in situ elimination to proceed at
low temperature were thus sought.
Linear Enediynes. To establish proof of principal for the
coupling process, commercially available propargylic bromide
9 was employed for model studies (Scheme 2). An initial survey
of metallating agents, temperature of reaction, and stoichiometry
of reactants quickly revealed that the coupling-elimination
process was indeed feasible, albeit giving only modest yields
Although a number of methods are available for the construc-
tion of E and Z enediyne subunits in both linear and cyclic form,
they often present practical limitations and can be specific to a
narrow range of substrates.4 We became interested in developing
* Corresponding author (phone: 617/373-8619; fax: 617/373-8795;
E-mail: grjones@lynx.neu.edu).
(1) (a) Grissom, J. W.; Gunawardena, G. U.; Klingberg, D.; Huang, D.
Tetrahedron 1996, 52, 6453. (b) Nicolaou, K. C.; Smith, A. L. Acc. Chem.
Res. 1992, 25, 497. (c) Zein, N.; Schroeder, D. R. In AdVances in DNA
Sequence Specific Agents; Jones, G. B., Ed.; Jai Press Inc., Greenwich, CT,
1998; Vol. 3.
(2) Boldi, A. M.; Diederich, F. Angew. Chem., Int. Ed. Engl. 1994, 33,
468.
(3) Anthony, J. A.; Boudon, C.; Diederich, F.; Gisselbrecht, J. P.;
Gramlich, V.; Gross, M.; Hobi, M.; Seiler, P. Angew. Chem., Int. Ed. Engl.
1994, 33, 763.
(5) For discussion of carbenes see: (a) Seyferth, D.; Lambert, R. L.;
Massol, M. J. Organomet. Chem. 1975, 88, 255. (b) Brandsma, L. In
PreparatiVe Polar Organometallic Chemistry 2; Springer-Verlag: New
York, 1990. (c) Olmstead, K. K.; Nickon, A. Tetrahedron 1999, 55, 3585.
Applications see: (d) Hopf, H.; Kreutzer, M.; Jones, P. G. Chem. Ber. 1991,
124, 1471. (e) Hauptmann, H. Tetrahedron Lett. 1974, 3587.
(4) Maier, M. E. Synlett 1995, 13.
10.1021/ja993766b CCC: $19.00 © 2000 American Chemical Society
Published on Web 02/08/2000

1938 J. Am. Chem. Soc., Vol. 122, No. 9, 2000
Jones et al.
Scheme 2. Preparation of
HMPA and LiBr, 11 is formed cleanly, and all starting material
Bis-trimethylsilyl-3-hexen-1,5-diyne
is consumed (entry 9).
The ratio of Z:E isomers proved to be a function of
temperature, but more suprisingly, high yields of product were
still attainable when the reaction was conducted at ambient
temperature (entries 9-13)! Addition of LiBr to reactions
containing HMPA still had a negative influence, resulting in a
drop in both yield and stereoselectivity (entry 14). Changing
counterion of the base gave no improvement (entry 15). Because
under appropriate conditions, bromodiyne 10 could be recovered,
the stereoselectivity of the elimination step was independently
probed (entries 16-23). To our suprise, elimination from 10
using LiHMDS resulted in inferior Z:E ratio, although a
temperature dependence on this ratio was evident (entries 16-
18). The nature of the base used and requirement for cosolvent
are clearly important, LDA or KCN/dimethyl sulfoxide (DMSO)
giving little or no selectivity (entries 20-22), and omission of
HMPA resulting in lowering of yield (entry 19). Conducting
the reaction at lower temperature gave an improvement in
stereocontrol, but the response to addition of LiBr was favorable,
suggesting that this may play a role in the elimination process
(entry 23).
of product (Table 1). In addition to halodiyne 10 and the
corresponding â elimination products 11, triyne 12 was also
recovered in low yield when lithium diisopropylamide (LDA)
was employed as base. This product can be envisioned by R
deprotonation of 10, and subsequent addition to 9. More
encumbered bases were examined, and as expected, the hin-
dered, nonnucleophilic, Li hexamethyldisilazane (HMDS) proved
more effective, resulting in appreciable conversion to 11
directly.6 Based on the assumption that a carbenoid intermediate
is involved, and because LiBr has been reported to have a
stabilizing effect on such species,7 it seemed reasonable that
LiBr could serve to inhibit the process and account for the
modest yields of product. Under these circumstances, should
the rate of coupling decrease substantially, all available 9 would
become metallated, leaving a deficit of electrophilic component.
Deliberate introduction of LiBr to the reaction medium sup-
ported this hypothesis, resulting in a sharp loss of yield (entry
7). Initial attempts to remove liberated lithium halide by
precipitation using nonpolar solvent mixtures proved ineffective,
as was the case using alternate proparglic halides as substrates
(vide infra). Remedy was finally found by introducing the
electron donor additive hexamethylphosphoramide (HMPA),
which was reported to have a destabilizing effect on carbenoids.7
Thus when a tetrahydrofuran (THF) solution of 1.1 equiv
LiHMDS and 1.1 equiv HMPA is slowly added to a solution
of 9 in THF at -95 °C, maintaining equal concentrations of
The origin of stereochemical bias toward the thermodynami-
cally less stable Z isomer deserves comment. Indeed Sondheimer
had reported that the base-induced elimination of the propargyl
tosylate of 1,5-hexadiyne-3-ol preferentially produces the E
isomeric enediyne.8 Although direct elimination from 10 gave
isomer ratios comparable with the in situ process, rationalization
using conventional models reveals that the Z stereoisomer is
clearly not predicted based on transition state considerations.
The antiperiplanar model A for E2 elimination suffers from
gauche interactions, suggesting that a synperiplanar elimination
(model C) predominates. However, because an unencumbered
E2 antiperiplanar model can be invoked to give rise to the E
isomer (model B), it is entirely possible that an alternative
process to the â-elimination may be responsible for the
formation of the predominant Z isomer. A viable possibility
would be an R-elimination by means of the formation and
subsequent degradation of a second carbenoid, resulting in a
highly transient carbene (model D). Such a carbene, if it formed,
would be expected to undergo a rapid insertion into an adjacent
Table 1. Effects of Base on Carbenoid Coupling-Elimination
% yields
entry
substrate
base
equiv
T (°C)
solvent/additive
10
11
Z/E
12
1
2
9
9
n-BuLi
1.1
1.0
0.3
0.5
1.1
2.0
1.1
1.1
1.1
1.1
1.1
1.1
1.1
1.1
1.1
1.1
1.1
1.1
1.1
1.8
1.2
1.1
1.1
-100
-100
-85
-100
-85
-85
-95
-85
-95
-85
-45
0
THF
THF
THF
THF
THF
THF
0
0
10
0
43
64
94
25
3
96
95
94
92
90
27
92
93
95
92
64
91
95
74
93
n.a.
0
5
16
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
LDA
5
1.1
3
9
LDA
16
0
n.a.
4
9
LiHMDS
LiHMDS
LiHMDS
LiHMDS
LiHMDS
LiHMDS
LiHMDS
LiHMDS
LiHMDS
LiHMDS
LiHMDS
KHMDS
LiHMDS
LiHMDS
LiHMDS
LiHMDS
LDA
2.3:1
1.9:1
2.1:1
1.5:1
1:2
5
9
0
6
9
0
7
9
THF/LiBr (5)
0
8
9
THF/LiBr (5)
4
9
9
THF/HMPA (1.1)
THF/HMPA (1.1)
THF/HMPA (1.1)
THF/HMPA (1.1)
THF/HMPA (1.1)
THF/HMPA (1.1) LiBr (5)
THF/HMPA (1.1)
THF/HMPA (1.1)
THF/HMPA (1.1)
THF/HMPA (1.1)
THF
0
2.2:1
2.1:1
1.6:1
1.3:1
1.2:1
1.5:1
1.8:1
1.6:1
1.5:1
1.2:1
1.9:1
1:1
10
11
12
13
14
15
16
17
18
19
20
21
22
23
9
0
9
0
9
0
9
25
0
9
-95
-95
-95
-78
0
0
9
0
10
10
10
10
10
10
10
10
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
-85
-95
-95
25
THF/HMPA (1.1)
THF
DMSO
THF/HMPA (1.1) LiBr (5)
LDA
KCN
LiHMDS
1:1
1.2:1
2.2:1
-95

Conjugated Enediynes
J. Am. Chem. Soc., Vol. 122, No. 9, 2000 1939
Scheme 3. Substituted Enediynes via Propargylic Halides
using the TES group could be duplicated using the HMPA
substitute 1,3-dimethyl-3,4,5,6-tetrahydro-2(1H)-pyrimidinone
(DMPU), and even improved upon by employing the powerful
electron-donating solvent trispyrrolidinophosphoramide (TPP)
(entries 10-13). Finally, the influence of the base was examined,
using a series of lithium disilazides 17-20 with the TES
substrate (entries 14-19). Striking changes in selectivity were
revealed, with the n-alkyl base 17d proving optimal, yet aryl
bases 17b and 19 resulting in reVersal of selectivity. Conducting
the reactions at even lower temperature (where use of the TPP
additive became essential to prevent precipitation) resulted in
even higher selectivity (entry 20). One possibility is that specific
interactions between the alkylsilyl groups on the base and
substrate play a key role in selectivity. This hypothesis is
supported by results obtained with alternate substrates (entries
21-23) where up to 12:1 Z selectivity is attainable (entry 21).
These results are significant in that they tentatively suggest that
appropriately chosen base/substrate combinations might be
tailored to allow access to otherwise thermodynamically inac-
cessible products.
C-H bond, yielding the observed product 11, and would also
account for the observed formation of byproduct 12. Regrettably,
attempts to intercept this carbene by introduction of trapping
agents were unsuccessful.
For R-substituted enediynes, as predicted (models E and F)
the subtle stereochemical bias that the vinyl substituents impart
had a profound influence on product geometric isomer ratio
(entries 24-34). Surprisingly, the ethyl substituent retained
stereoselectivity in favor of the Z isomer (entries 24-26).
However, moving to the isopropyl derivatives and beyond,
selectivity was reversed and greatly enhanced in favor of the E
isomer. Because of its robust nature and popularity in alkyne-
based materials chemistry, the triisopropylsilyl group was
adopted for series of analogues. The route provided ready access
to disubstituted primary, secondary, tertiary, cycloalkyl, aryl,
and heteroaryl enediynes. Despite efforts to intercept the
intermediate bromodiyne, in all cases examined, in situ elimina-
tion ensued, resulting in direct formation of the enediynes 16,
in good to excellent yield. The stereoselectivity in the process
correlates well with steric bulk of the vinyl substituent,
consistent with antiperiplanar alignment for in situ E2 elimina-
tion (F). Exceptionally high levels of stereoselectivity are
obtained in many cases (entries 27-28, 30) and recovered yields
are high even with sensitive substrates (entries 29, 32). To
demonstrate the utility of the process, in the case of the diphenyl
enediyne (entry 31) Pd-catalyzed coupling of 13 [R ) TIPS]
with E 1,2-dichlorostilbene was conducted. EVen under a range
of conditions, using 1:1 stoichiometry, the maximum yield of E
enediyne obtained was <10%, a competing reaction being
alkyne dimerization, confirming the benefit of the current
method. (Catalysts examined included [Pd(PPh3)4], [PdCl2-
(PPh3)2], [PdCl2(PhCN)2].20a Improved yields are obtained when
>5 equiv of vinyl chloride is employed or the corresponding
bromides are synthesized and used.20b) The synthesis of tet-
(Interception with a range of arene traps failed to give the
corresponding cycloheptatriene adducts.)
Regardless, the models and results suggested that subtle
effects may be responsible for the selectivity observed. We
therefore elected to undertake a comprehensive probe of the
effects of (i) different terminal alkyne functionality (R1), (ii)
nature of halide leaving group, (iii) substitution at the propar-
gylic position (R2), (iv) alternate disilazide bases, and (v) various
cosolvents. The substituents alone may be expected to either
consolidate syn elimination/R insertion to form the Z isomer
(when R2 ) H), or in cases where R2 is bulky to encourage E2
elimination to give the E isomer (model F), which would be
preferred to that indicated in model E. A variety of propargylic
bromides were thus prepared (Scheme 3). Using the optimized
conditions, a variety of strategically important enediyne building
blocks could be accessed, and in every case examined, chemical
yields of products were good, and separation of isomers were
routinely achieved either using distillative or chromatographic
methods. Little effect of the halide bulk on the process was
identified, both in terms of chemical yield and product Z:E ratio
(Table 2, entries 1-2). However, stereochemical outcome
proved sensitive to bulk of the alkyne termini such that the
triethylsilyl (TES) analogue (Table 2, entry 3) was more
selective for the Z isomer, but additional bulk was detrimental
(entries 4-8). Analysis of both the TES and triisopropylsilyl
(TIPS) analogues revealed a strong temperature dependence on
selectivity (Figure 1). The effect of carbenoid destabilizing
additives was then scrutinized, and the excellent results obtained
(6) Jones, G. B.; Huber, R. S. Tetrahedron Lett. 1994, 35, 2655.
(7) Tarhouni, R.; Kirschleger, B.; Rambaud, M.; Villieras, J. Tetrahedron
Lett. 1984, 25, 835.
(8) Okamura, W. H.; Sondheimer, F. J. Am. Chem. Soc. 1967, 89, 5991.

1940 J. Am. Chem. Soc., Vol. 122, No. 9, 2000
Jones et al.
Scheme 4. Unified Procedure for Synthesis of Linear Enediyne Building Blocksa
a Conditions: (a) LiHMDS, HMPA; (b) TBAF, THF; (c) BuLi, THF; (d) Na(OCH3)3BH; (e) BuLi, then (CH2O)n.
Table 2. In Situ Elimination of Propargylic Halides 15 to Give Enediynes 16
entry
R1
R2
base
hal
temp
additive
% yield
Z:E
1
2
TMS
TMS
TES
TBDMS
TIPS
TPS
DMTS
tBu
Ph
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
Et
Et
Et
iPr
tBu
17a
17a
17a
17a
17a
17a
17a
17a
17a
17a
17a
17a
17a
17b
17c
17d
18
Cl
I
-85
-85
-95
-95
-95
-95
-95
0
HMPA(1.1)
HMPA(1.1)
HMPA(1.1)
HMPA(1.1)
HMPA(1.1)
HMPA(1.1)
HMPA(1.1)
HMPA(1.1)
HMPA(1.1)
DMPU(1.1)
TPP(1.1)
95
93
94
90
92
89
91
25
90
89
93
91
95
87
91
90
92
89
94
88
92
89
90
65
75
85
87
70
79
80
95
82
95
80
1.9:1
1.9:2
4.2:1
4.0:1
2.9:1
1.2:1
1.1:1
1.2:1
1.0:1
4.0:1
4.0:1
4.0:1
6.1:1
1:1.6
3.3:1
5.1:1
1.8:1
1:1.2
2.0:1
9.0:1
12.0:1
1.8:1
1.5:1
2.0:1
2.0:1
3.0:1
1.0:100
1.0:100
1.0:15
1.0:100
1.0:11
1.0:6
N/A
3
Br
Br
Br
Br
Br
Br
Br
Br
Br
Br
Br
Br
Br
Br
Br
Br
Br
Br
Br
Br
Br
Br
Br
Br
Br
Br
Br
Br
Br
Br
Br
Br
4
5
6
7
8
9
-95
-95
-95
-95
-95
-95
-95
-95
-95
-95
-95
-115
-115
-115
-95
-95
-80
-80
-80
-80
-80
-80
-80
-80
-80
-80
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
TES
TES
TES
TES
TES
TES
TES
TES
TES
TES
TES
TBS
TMS
TPS
TMS
TMS
TIPS
TIPS
TIPS
TIPS
TIPS
TIPS
TIPS
TIPS
TIPS
HMPA(5.0)
TPP(5.0)
HMPA(1.1)
HMPA(1.1)
HMPA(1.1)
HMPA(1.1)
HMPA(1.1)
HMPA(1.1)
TPP(5.0)
19
20
17d
17d
17d
19
TPP(5.0)
TPP(5.0)
TPP(5.0)
17a
17a
17a
17a
17a
17a
17a
17a
17a
17a
17a
HMPA(1.1)
HMPA(1.1)
HMP(1.1)
HMPA(1.1)
HMPA(1.1)
HMPA(1.1)
HMPA(1.1)
HMPA(1.1)
HMPA(1.1)
HMPA(1.1)
HMPA(1.1)
c(C3H5)
c(C6H11)
Ph
3-pyridyl
TIPS ethynyl
TBS ethynyl
1:1
raalkynyl systems is also noteworthy (entries 33-34), in that it
method offers a general alternative to classical Pd-mediated
cross-coupling reactions; its merits include speed, economy, and
versatility.
complements existing stepwise methods.9 Overall, the present
(9) (a) Vollhardt, K. P. C.; Winn, L. S. Tetrahedron Lett. 1985, 26, 709.
(b) Wang, K. K.; Wang, Z.; Gu, Y. G. Tetrahedron Lett. 1993, 34, 8391.
(c) Stracker, E. C.; Zweifel, G. Tetrahedron Lett. 1991, 32, 3329. (d)
Magriotis, P. A.; Scott, M. E.; Kim, K. D. Tetrahedron Lett. 1991, 32,
6085. (e) Alami, M.; Crousse, B.; Linstrumelle, G. Tetrahedron Lett. 1994,
35, 3543. (f) Shibuya, M.; Sakai, Y.; Naoe, Y. Tetrahedron Lett. 1995, 36,
897. (g) Nuss, J. M.; Murphy, M. M. Tetrahedron Lett. 1994, 35, 37. (h)
Dai, W.-M.; Fong, K. C.; Danjo, H.; Nishimoto, S. Angew. Chem., Int. Ed.
Engl. 1996, 35, 779. (i) Anthony, J.; Knobler, C. B.; Diederich, F. Angew.
Chem., Int. Ed. Engl. 1993, 32, 406. (j) Rubin, Y.; Knobler, C. B.; Diederich,
F. Angew. Chem., Int. Ed. Engl. 1991, 30, 698.
Having a unified route to either Z or E enediynes opens up
a number of synthetic possibilities. For E enediynes, desilylation
under standard conditions gave the corresponding terminal
alkynes including 21-24 (Scheme 4). These substrates and
others can be expected to find immediate application in metal-
catalyzed oligomerization sequences,10 and possibly in the
(10) See: Modern Acetylene Chemistry; Stang, P. J., Diederich, F., Eds.;
VCH: Weinheim, 1995.

Conjugated Enediynes
J. Am. Chem. Soc., Vol. 122, No. 9, 2000 1941
design of various enediyne based nanomaterials.11 The stereo-
selective route to Z enediynes demonstrated in turn can be
expected to find application in the preparation of synthetic
enediyne antitumor agents. Z-1,6-Dilithio-3-hexene-1,5-diyne
25, available via deprotection and metallation, has been used
by others for the construction of natural enediynes,12 and
differential deprotection allows access to the key Myers type
precursor 26 13 and synthon 27, which have been used widely
in the synthesis of enediyne hybrids.14
Cyclic Enediynes. The naturally ocurring enediyne antitumor
antibiotics typically possess either a 9- or 10-membered cyclic
enediyne moiety. For the cyclohex-3-ene-1,5-diyne subunit
present in 10-membered enediynes, a variety of methods for
construction have been investigated. In contrast to the natural
systems where the enediynes are strained and amenable to a
variety of synthetic approaches, the parent subunit 28, which
undergoes Bergman type cyclization to give diyl radical 29, has
an observed half-life of 18 h at 37 °C, warranting special
circumstances during its synthesis.15 Methods examined to date
are illustrated. Although the Ramberg-Backlund reaction (route
a) has been used successfully by Nicolaou,16 for parent enediyne
28 the chemical yield is low, and involves a multistep procedure
to assemble the substrate. To address the former issue, an elegant
alternate route was developed involving a base-induced retro
Diels-Alder reaction (route b).17 The Corey-Winter procedure
(route c) was employed by Semmelhack,18 which likewise gives
the desired product, but requires a multistep sequence to prepare
the substrate. Palladium-mediated vinyl coupling methods (route
d) have also been attempted, but for 28 the prolonged conditions
required to effect cyclization are incompatible with the (ther-
mally labile) product.19 Likewise, although linear enediynes 34
(Y ) Li) have been used as nucleophiles for addition to carbonyl
substrates, cyclization to give 28 has not been effected (route
e).
Figure 1. Z/E temperature dependence.
Scheme 5. Attemped Acetylide-Mediated Closures
Scheme 6. Distribution of Acetylide Alkylation Products
without success (vide infra). Therefore, we elected to concentrate
on use of the linear enediyne 25, which was available in
multigram quantities using the intermolecular carbenoid method.
Reaction with 1,4-diiodobutane under a variety of conditions
led primarily to monoalkylated species 35, with only traces of
the cyclic species 28 detected following prolonged reaction at
ambient temperature (Scheme 5). Indeed, under such conditions,
product cycloaromatization would be competitive with ring
closure. Attempts to reduce the degrees of freedom for the
cyclization were made, using cyclic substrates 36, however, even
under a range of conditions, bicyclic product 37 could not be
detected. The sluggish nucleophilicity of the acetylide was
presumably responsible for the poor yields. This was supported
by a model study where the lithio acetylide ion of phenylethyne
was reacted with 1,4-diiodobutane (Scheme 6). Using 2:1
stoichiometry, which eventually required extended exposure at
ambient temperature to achieve coupling, the balance of material
was chiefly monoalkylated product 38, with only moderate yield
of dialkylated species 39. Addition of acetone as a quench after
20 h confirmed that acetylide was still present, leading to
formation of 1,2 addition product 40.
We initially attempted formation of 28 using an intramo-
lecular version of the carbenoid coupling-elimination, but
(11) (a) Gobbi, L.; Seiler, P.; Diederich, F. Angew. Chem., Int. Ed. Engl.
1999, 38, 674. (b) Diederich, F.; Gobbi, L. Top. Curr. Chem. 1999, 201,
43.
(12) Haseltine, J. N.; Cabal, M. P.; Mantlo, N. B.; Iwasawa, N.;
Yamashita, D. S.; Coleman, R. S.; Danishefsky, S. J.; Schulte, G. K. J.
Am. Chem. Soc. 1991, 113, 3850.
(13) (a) Myers, A. G.; Harrington, P. M.; Kuo, E. Y. J. Am. Chem. Soc.
1991, 113, 694. (b) Myers, A. G.; Parrish, C. A. Bioconjugate Chem. 1996,
7, 322.
Accordingly, we sought milder conditions to effect cycliza-
tion, and elected to pursue an enolate-based strategy, as
illustrated in Scheme 7. The ester enolate was reasoned to be
appropriate, because intramolecular 10-exo-tet or 10-exo-trig
processes present viable vector approaches for formation of the
desired ring. The requisite linear enediyne substrate was
prepared from the tetrahydropyranyl (THP) ether of propargylic

1942 J. Am. Chem. Soc., Vol. 122, No. 9, 2000
Scheme 7. Exocyclic Enolate Ring Closure Strategy
Jones et al.
Scheme 9. Carbenoid Route to Cyclic-3-ene-1,5-diynes
Scheme 8. Preparation of Intramolecular Enolate Substratesa
accessible from the commercially available alkadiynes 45
(Scheme 9). Should conditions to allow low-temperature cy-
clization be developed, they might also be amenable to
production of C-9 enediyne [X ) CH2] in addition to the more
heavily studied C-10 [X ) CH2CH2] and C-11 enediynes [X )
CH2CH2CH2].
a X ) OMs: MsCl, Et3N, CH2Cl2, 72%. X ) OTos: TosCl, Py,
85%. X ) Cl: CCl4, Ph3PCH2Cl2, 78%. X ) Br: CBr4, Ph3P, CH2Cl2,
59%. X ) [CdO]: Dess-Martin periodinane, 97%.
In initial experiments, employing typical conditions for
intermolecular coupling, only regenerated starting material could
be isolated from the workup mixtures (Table 3, entries 1 and
2). By increasing the reaction temperatures from -85 to -45
°C however, moderate yields of the desired cyclic species 47
could be obtained, along with varying amounts of starting
material, together with intermolecular coupling products (entry
3). As in the case of the intermolecular variant of this reaction,
the hindered LiHMDS proved to be the base of choices
substitution for LDA, lithium-2,2,6,6-tetramethyl piperidine
(LTMP), NaHMDS, or KHMDS resulting in lowering of the
overall yields of the desired products (entries 3-9). A key
feature for increasing efficiency of the cyclization proved to be
the concentration of the carbenoid destabilization agent,7 and
when the ratio of additive/base was increased from 1:1 to 10:1
moderate yields of product were produced with a variety of
different agents. However, optimal conditions were found to
involve slow addition of base to a premixed solution of substrate/
additive (method C), resulting in near quantitative formation
of the enediyne product (entry 15).
alcohol (Scheme 8). Palladium coupling with Z dichloroethene
under Sonogashira conditions gave chloroeneyne 41,20 to which
was attached the ester function using commercially available
methyl-5-hexynoate. Hydroxy ester 42 was converted into a
variety of functional group analogues 43 using traditional
methods, then the intramolecular cyclization was probed. In no
instance was the desired ester 44 obtained (or the R-hydroxy
ester where X ) [CdO]). As previously, a range of conditions
were investigated, leading chiefly to recovery of starting
material, or intermolecular reactions, which could not be
suppressed on manipulation of substrate concentration.
Surprised by this outcome, we serendipitously elected to
reinvestigate an intramolecular variant of the carbenoid coupling-
elimination process, and concentrated efforts on dibromides 46,
(14) (a) Lu, Y.-F.; Harwig, C. W.; Fallis, A. G. J. Org. Chem. 1993, 58,
4202. (b) Schoenen, F. J.; Porco, J. A., Jr.; Schreiber, S. L. Tetrahedron
Lett. 1989, 30, 3765. (c) Porco, J. A., Jr.; Schoenen, F. J.; Stout, T. J.;
Clardy, J.; Schreiber, S. L. J. Am. Chem. Soc. 1990, 112, 7410. For other
syntheses using terminally protected linear enediynes, see: (d) Magnus,
P.; Fortt, S. M. J. Chem. Soc. Chem. Commun. 1991, 544. (e) Cabal, M.
P.; Coleman, R. S.; Danishefsky, S. J. J. Am. Chem. Soc. 1990, 112, 3253.
(15) For the contribution of strain to Bergman cycloaromatization see:
Magnus, P.; Fortt, S.; Pitterna, T.; Snyder, J. P. J. Am. Chem. Soc. 1990,
112, 4986.
(16) Nicolaou, K. C.; Zuccarello, G.; Ogawa, Y.; Schweiger, E. J.;
Kumazawa, T. J. Am. Chem. Soc. 1988, 110, 4866.
(17) Bunnage, M. E.; Nicolaou, K. C. Angew. Chem., Int. Ed. Engl. 1996,
35, 1110.
In some instances intermolecular reactions served to lower
the yields of the desired cyclic species. Where this was found
to be the case, the yields of the desired products could generally
be improved by increasing the dilution of the reaction medium.
Because of the thermal instability of 47 (X ) CH2), in situ
protection as its cobalt carbonyl complex was effected, by
adding dicobalt octacarbonyl to the (predried) workup solvent
mixtures (Et2O containing THF from the cyclization). This
(18) Semmelhack, M. F.; Gallagher, J. Tetrahedron Lett. 1993, 34, 4121.
(19) For discussion of these limitations see: Nicolaou, K. C.; Zuccarello,
G.; Riemer, C.; Estevez, V. A.; Dai, W.-M. J. Am. Chem. Soc. 1992, 114,
7360.
(20) (a) Sonogashira, K.; Tohda, Y.; Hagihara, N. Tetrahedron Lett. 1975,
16, 4467. (b) Chemin, D.; Linstrumelle, G. Tetrahedron Lett. 1994, 50,
5335.

Conjugated Enediynes
J. Am. Chem. Soc., Vol. 122, No. 9, 2000 1943
Table 3. Preparation of Cyclic Enediynes and their Complexes 48 via Coupling Dibromides 46
entry
X
hal
base
equiv
additive
equiv
temp
methoda
%47
%48
%50
1
2
CH2 CH2
CH2 CH2
CH2 CH2
CH2 CH2
CH2 CH2
CH2 CH2
CH2 CH2
CH2 CH2
CH2 CH2
CH2 CH2
CH2 CH2
CH2 CH2
CH2 CH2
CH2 CH2
CH2 CH2
CH2 CH2
CH2CH2CH2
CH2CH2CH2
CH2
Br
Br
Br
Br
Br
Br
Br
Br
Br
Br
Br
Br
Br
Br
Br
Br
Br
Br
Br
Br
LiHMDS
LiHMDS
LiHMDS
LiHMDS
LiHMDS
NaHMDS
KHMDS
LDA
2.2
2.2
2.2
4
NA
1.1
1.1
1.1
1.1
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
-78
-85
-45
-45
-45
-45
-45
-45
-45
-45
-45
-45
-45
-45
-45
-45
-45
-45
-45
-45
A
B
B
B
B
B
B
B
B
B
B
B
B
B
C
D
B
C
D
C
<1
0
14
15
20
15
5
10
15
22
10
8
10
8
∼
∼
∼
∼
∼
∼
∼
∼
∼
∼
∼
∼
∼
∼
∼
92
∼
∼
11
∼
∼
∼
∼
∼
∼
∼
∼
∼
∼
∼
∼
∼
∼
∼
∼
∼
∼
∼
55
89
HMPA
HMPA
HMPA
HMPA
HMPA
HMPA
HMPA
HMPA
TPP
3
4
5
2.2
2.2
2.2
2.2
2.2
2.2
2.2
2.2
2.2
2.2
2.2
2.2
2.2
2.2
2.2
2.2
6
7
8
9
LTMP
10
11
12
13
14
15
16
17
18
19
20
LiHMDS
LiHMDS
LiHMDS
LiHMDS
LiHMDS
LiHMDS
LiHMDS
LiHMDS
LiHMDS
LiHMDS
LiHMDS
TMEDA
DMPU
DMI
HMPA
TEP
HMPA
HMPA
HMPA
HMPA
HMPA
95
∼
12
90
∼
CH2
∼
a Method A: base added to 46; method B: base + additive added to 46; method C: base added to 46 + additive; method D: as method C, then
immediate complexation with Co2CO8.
We opted to synthesize a preformed C-10 cyclic enediyne
with a functionalized spacer group attached, then couple this
unit to an aryl moiety via its cobalt carbonyl complex. A route
was thus developed for the synthesis of appropriate complexes
(Scheme 10). Commercially available methyl-5-hexynoate (51)
was first converted to diyne 52. Hydroxymethylation gave key
diol 53, which when followed by bromination allowed carbenoid
coupling to give the desired enediyne 54 in high yield. For
practical purposes it was immediately protected as the corre-
sponding cobalt carbonyl complex. Alternatively, it was found
that in situ deprotective bromination of 53 gave rise to the
corresponding tribromide, which itself underwent cyclization
to give the isolatable bromoenediyne in moderate yield, and
which could likewise be converted to the complex (56). The
key deprotection of the tert-butyldimethylsilyl ether of the cobalt
carbonyl complex derived from 54 was then investigated. Initial
concerns regarding the acid sensitivity of the cobalt moiety
proved to be unfounded; conventional conditions (TsOH/MeOH/
25 °C) successfully afforded the corresponding alcohol 55 in
high yield.
Routine functional group interconversions such as Swern
oxidation to the aldehyde 57 (72%), displacement to give the
bromide 56 (84%), or esterification to give 58 were not impeded
by the cobalt carbonyl group. In principle, these species offer
a variety of regimens for attachment to biomolecule binding
agents to be pursued, allowing the development of libraries of
hybrid enediynes using shelf-stable precursors. Prior to initiating
these studies, we wished to confirm the ability of the enediyne
core itself to undergo cycloaromatization. Accordingly, substrate
55 was subjected to decomplexation. We had previously found
that tetrabutylammonium fluoride (TBAF) is an effective agent
for this transformation using linear enediynes,23 and were
delighted to observe that quantitative deprotection at low-
Figure 2.
allowed chromatographic purification of the masked enediynes
48 where necessary, and more importantly, permitted shelf-
storage for extended periods. The process was successful for
the C-11 and C-9 enediynes, but in the latter case, concomitant
cycloaromatization to produce 50 mandates isolation as the
cobalt complex (entries 19 and 20). The isolation of this C-9
enediyne complex is noteworthy, constituting the first synthesis
of this interesting species, the Bergman cycloaromatization
profile of which has yet to be studied.21
With an efficient synthesis of cyclic enediynes secure, we
turned our attention to the preparation of cyclic enediynes which
can be potentially tethered to molecules with specific biomo-
lecular targets. For DNA binding agents, several investigations
of this nature have been reported, that used an eclectic mix of
protocols to synthesize the desired cyclic enediyne component.22
Encouraged by the near quantitative yields of 48 obtained
following optimization of the cyclization-complexation strat-
egy, we turned our attention toward incorporation of this
methodology in the preparation of shelf-stable synthons (Figure
2).
(21) For computational analysis of cyclic enediynes see: Snyder, J. P.
J. Am. Chem. Soc. 1990, 112, 5367.
(22) (a) Nicolaou, K. C.; Smith, A. L.; Yue, E. W. Proc. Natl. Acad.
Sci. U.S.A. 1993, 90, 5881. (b) Toshima, K.; Ohta, K.; Ohashi, A.;
Nakamura, T.; Nakata, M.; Matsumura, S. J. Chem. Soc. Chem. Commun.
1993, 1525. (c) Tokuda, M.; Fujiwara, K.; Gomibuchi, T.; Hirama, M.;
Uesugi, M.; Sugiura, Y. Tetrahedron Lett. 1993, 34, 669. (d) Semmelhack,
M. F.; Gallagher, J. J.; Ding, W.; Krishnamurthy, G.; Babine, R.; Ellestad,
G. A. J. Org. Chem. 1994, 59, 4357. (e) Boger, D. L.; Zhou, J. J. Org.
Chem. 1993, 58, 3018. (f) Wender, P. A.; Zercher, C. K.; Beckham, S.;
Haubold, E. M. J. Org. Chem. 1993, 58, 5867. (g) Funk, R. L.; Young, E.
R. R.; Williams, R. M.; Flanagan, M. F.; Cecil, T. L. J. Am. Chem. Soc.
1996, 118, 3291. (h) Takahashi, T.; Tanaka, H.; Matsuda, A.; Yamada, H.;
Matsumoto, T.; Sugiura, Y. Tetrahedron Lett. 1996, 37, 2433. (i) Myers,
A. G.; Cohen, S. B.; Tom, N. J.; Madar, D. J.; Fraley, M. E. J. Am. Chem.
Soc. 1995, 117, 7574. (j) Hay, M. P.; Wilson, W. R.; Denny, W. A. Bioorg.
Med. Chem. Lett. 1995, 5, 2829.
(23) Jones, G. B.; Wright, J. M.; Rush, T. M.; Plourde, G. W.; Kelton,
T. F.; Mathews, J. E.; Huber, R. S.; Davidson, J. P. J. Org. Chem. 1997,
62, 9379.
(24) (a) Zein, N.; Casazza, A. M.; Doyle, T. W.; Leet, J. E.; Schroeder,
D. R.; Solomon, W.; Nadler, S. G. Proc. Natl. Acad. Sci. U.S.A. 1993, 90,
8009. (b) Zein, N.; Solomon, W.; Casazza, A. M.; Kadow, J. F.; Krishnan,
B. S.; Tun, M. M.; Vyas, D. M.; Doyle, T. W. Bioorg. Med. Chem. Lett.
1993, 3, 1351. (c) Zein, N.; Reiss, P.; Bernatowicz, M.; Bolgar, M. Chem.
Biol. 1995, 2, 45. (d) Jones, G. B.; Kilgore, M. W.; Pollenz, R. S.; Li, A.;
Mathews, J. E.; Wright, J. M.; Huber, R. S.; Tate, P. L.; Price, T. L.; Sticca,
R. P. Bioorg. Med. Chem. Lett. 1996, 6, 1971-1976. For mechanistic
insights see: (e) Braslau, R.; Anderson, M. O. Tetrahedron Lett. 1998, 39,
4227.

1944 J. Am. Chem. Soc., Vol. 122, No. 9, 2000
Jones et al.
Scheme 10. Carbenoid Route to Cyclodec-3-ene-1,5-diynesa
Scheme 12. Proof of Principal for Arene-enediyne Library
Design
tetrahydronaphthoic acid. A similar protocol was then applied
to the ester 58, resulting in cycloaromatization of the aryl-
enediyne conjugate to give 63, isolated in high yield following
incubation with 1,4-cyclohexadiene (Scheme 12), and identical
with authentic material prepared from 61 (PhCOCl, Et3N).
These short sequences establish the concept of taking a
thermally stable enediyne complex and attaching the diyl
progenitor to a molecule (arene) with potential affinity for a
biomolecular target. The biological effects of enediyne-derived
diyl radicals have been the focus of intense research over the
last decade. In addition to DNA cleavage, diyl-induced pro-
teolysis has also been reported,24 suggesting that numerous
targets may be envisioned. Extensions of the current work to
develop affinity cleavage systems, by preparing sequence-
specific hybrids is underway. Because the Bergman cycloaro-
matization proceeds via a late-state transition state, a reasonable
strategy will be to screen cycloaromatized products (more
elaborate analogues of 63, available by coupling alcohol 61 with
appropriate arenes) for affinity to specific biological targets.
Once identified, the corresponding enendiyne precursor to it
can then be investigated for its ability to induce cleavage of
that target. The efficiency with which shelf-stable enediynes
55-58 can be produced will afford considerable flexibility to
such a screening process, and a number of applications are
anticipated.
a Conditions: (a) LDA, propargyl bromide; (b) LAH, Et2O; (c)
TBDMSCl, imidazole, DMF; (d) nBuLi, EtOCOCl; (e) Ph3P:Br2/2,6-
lutidine; (f) LiHMDS/HMPA/THF; (g) Co2(CO)8; (h) TsOH/MeOH;
(i) Ph3P:Br2; (j) (COCl)2/Et3N/DMSO; (k) PhCOCl/Et3N.
Scheme 11. Cycloaromatization of C-10 Enediyne
Conclusion
The carbenoid coupling of lithiated propargylic halides was
studied and optimized to give ready access both to linear and
cyclic enediynes. Direct application of the chemistry can be
expected both in the design of antitumor agents and in the
preparation of enediyne nanomaterials.25 The chemical efficiency
and mildness of the process, coupled with predictable stereo-
chemical control, render it an attractive alternative to conven-
tional methods for the preparation of enediynes. Indeed in some
instances no comparable alternate route exists, justifying its place
in the synthetic arsenal.
temperature ensued, to give enediyne alcohol 59 (Scheme 11).
As expected, analogous reaction with freshly prepared enediyne
54 also produced this product. Using an NMR-based assay, the
half-life of this enediyne was determined to be 18 h at
physiological temperature, giving adduct 61 on trapping with
1,4-cyclohexadiene. The identity of this adduct was confirmed
by comparison with a sample obtained from reduction (lithium
aluminum hydride (LAH), THF) of commercially available
Acknowledgment. We acknowledge the efforts of Aiwen
Li, Dennis Parrish, Brant Chapman, Mark Whitten, and Farid
Fouad in the preparation of the substrates used for the carbenoid
couplings. This work was generously supported by the National
Institutes of Health [R01GM57123].
Supporting Information Available: Full experimental
details for all compounds described herein, together with
associated 13C NMR spectra for all enediyne products (PDB).
This material is available free of charge via the Internet at
http://pubs.acs.org.
(25) For recent applications of the carbenoid coupling see: (a) Dubois,
F.; Gingras, M. Tetrahedron Lett. 1998, 39, 5039. (b) McPhee, M. M.;
Kerwin, S. M. J. Org. Chem. 1996, 61, 9385. (c) Mikami, K.; Feng, F.;
Matsueda, H.; Yoshida, A.; Grierson, D. S. Synlett 1996, 833. (d) Konig,
B. Angew. Chem., Int. Ed. Engl. 1996, 35, 165.
JA993766B