Allenes and transition metals: A diverging approach to heterocycles

Article abstract of DOI:10.1021/ol0492391

An alkynyl allene has been converted to heterocycles possessing an α-alkylidene cyclopentenone, a 4-alkylidene cyclopentenone, or a cross-conjugated triene. Thus, a common intermediate has been converted to three structurally unique compounds by changing

Full text of DOI:10.1021/ol0492391

ORGANIC
LETTERS
2004
Vol. 6, No. 13
2245-2248
Allenes and Transition Metals: A
Diverging Approach to Heterocycles
Kay M. Brummond* and Branko Mitasev
Department of Chemistry, UniVersity of Pittsburgh, Pittsburgh, PennsylVania 15260
kbrummon@pitt.edu
Received April 26, 2004
ABSTRACT
An alkynyl allene has been converted to heterocycles possessing an r-alkylidene cyclopentenone, a 4-alkylidene cyclopentenone, or a cross-
conjugated triene. Thus, a common intermediate has been converted to three structurally unique compounds by changing only the reaction
conditions and, therefore, controlling various reaction pathways.
Diversity-oriented synthesis (DOS) is a recognized tactic for
synthesizing pools of new small molecules that will impact
Scheme 1. Nature’s Construction of Secondary Metabolites
research, either as biological tools or pharmacological
agents.1 The DOS approach used most frequently involves
sequential reactions where groups, structural and/or func-
tional, are incorporated into the products, thereby generating
diversity. An alternative and less common approach involves
the construction of a pivotal compound that, when subjected
to the appropriate reagents or catalysts, can give structurally
distinct products. Nature has made extensive use of this latter
diversification strategy, for example, enzyme-catalyzed
construction of the various sesquiterpene carbon skeletons
of secondary metabolites from one common precursor,
farnesyl pyrophosphate (Scheme 1).
alkynyl allene, when subjected to rhodium biscarbonyl
chloride dimer [(Rh(CO)2Cl)2] results in a selective reaction
with the distal double bond of the allene (pathway b).3 This
unusual regiocontrol element has subsequently been applied
Recently, investigations aimed at increasing the scope of
the allenic Pauson-Khand reaction have led to an interesting
regiocontrol element that results in compounds not easily
accessible using existing protocols. For example, treatment
of an alkynyl allene with molybdenum hexacarbonyl results
in a selective reaction with the internal double bond of the
allene in most cases (pathway a, Scheme 2).2 The same
(2) In some cases, a steric bias can direct the reaction toward the other
double bond. (a) Kent, J. L.; Wan, H.; Brummond, K. M. Tetrahedron Lett.
1995, 36, 2407. (b) Brummond, K. M.; Wan, H. Tetrahedron Lett. 1998,
39, 931. (c) Brummond, K. M.; Wan, H.; Kent, J. L. J. Org. Chem. 1998,
63, 6535. (d) Brummond, K. M. In AdVances in Cycloaddition; Harmata,
M., Ed.; JAI Press, Inc.; Stamford, CT, 1999; Vol. 6, p 211. (e) Brummond,
K. M.; Lu, J. J. Am. Chem. Soc. 1999, 121, 5087. (f) Brummond, K. M.;
Lu, J.; Petersen, J. L. J. Am. Chem. Soc. 2000, 122, 4915. Brummond, K.
M.; Sill, P.; Rickards, B.; Geib, S. J. Tetrahedron Lett. 2002, 43, 3735.
(1) Burke, M. D.; Schreiber, S. L. Angew. Chem., Int. Ed. 2004, 43,
46-58.
10.1021/ol0492391 CCC: $27.50 © 2004 American Chemical Society
Published on Web 06/04/2004

to other carbocyclization reactions. For example, if an argon
or nitrogen atmosphere is used instead of carbon monoxide
during the Rh(I)-catalyzed reaction of an alkynyl allene, high
yields of cross-conjugated trienes are afforded via a postu-
lated oxidative addition/â-hydride elimination/reductive elimi-
nation sequence (pathway c).4 The facility of this latter
reaction prompted us to consider an analogous reaction with
an alkene group instead of an alkyne and a tether containing
a nitrogen or an oxygen atom, which afforded azepines and
oxepines, respectively, in good yields.5
exploring the allenic amino acid protocols reported by
Castehano and Krantz (method I)7 and Kazmaier (method
II)8 for the preparation of 4 (Scheme 3). Both methods start
Scheme 3. Synthesis of Allenic Amino Acids
Scheme 2. Diverging Reaction Pathways
from the propargylic ester 3, which is prepared by the
condensation of a protected amino acid 1 and substituted
propargyl alcohol 2 in yields ranging from 78 to 95%.9 In
addition, each method produces an allenic amino acid 4 via
a Claisen rearrangement of compound 3. The two methods
differ only in the protecting groups on the amine and the
conditions used to effect the Claisen rearrangement. Method
I uses a dehydrating protocol (PPh3, CCl4, NEt3) that requires
that the amine be protected as a benzamide in order to
generate the oxazolone intermediate required for the Claisen
rearrangement. A shortcoming of this method is that the
rearrangement proceeds with low diastereoselectivity (dr ∼3:
2). Method II utilizes a chelate-controlled ester-enolate
Claisen rearrangement (LDA, ZnCl2) to provide allene 4.
The advantages to this protocol are the range of amine
protecting groups that are compatible with the Claisen
rearrangement conditions (Cbz, Boc and Ts), and the reaction
proceeds with excellent stereocontrol, showing a single
We believe that these discoveries are ideally suited for
the preparation of small-molecule libraries, because in
principle, one compound can be converted to at least three
structurally unique scaffolds, an R-alkylidene cyclopenten-
one, a 4-alkylidene cyclopentenone, and a cross-conjugated
triene. Moreover, transition metals are used extensively in
the preparation of libraries of compounds; however, very few
protocols utilize cycloisomerization processes. The high
substrate specificity exhibited by these metals can be a
serious limitation given that, in some cases, changing a single
substituent can turn a high-yielding process into one that does
not proceed at all. Thus, we report our results regarding the
scope and limitations of this transition metal-catalyzed
diverging strategy.
Since a long-term goal of this approach is to prepare
compounds that are biologically relevant, the incorporation
of a larger number of heteroatoms in the carbocyclization
substrate was essential. Thus, inclusion of an amino acid
subunit between the alkyne and the allene appeared to be
ideal for two reasons: (1) Bolton has used a similar strategy
for the construction of Pauson-Khand and Heck reaction
precursors6 and (2) there are known protocols for the
preparation of allenic amino acids. First we began by
1
diastereomer by H NMR. However, this method required
substituting the terminus of alkyne 2 with a trimethylsilyl
group (R ) TMS) since the Claisen rearrangement step was
not tolerant of a terminal alkyne.10 The synthesis of the target
alkynyl allene 6 was then completed by N-alkylation of the
N-protected allenic amino acid. Standard N-alkylation condi-
tions were employed using NaH as a base and propargylic
halides. The yields for the latter reaction and the substrates
prepared are included in Table 1.
Target substrates (15) have been synthesized in an effort
to establish the effect of substrate specificity on the course
of these transition metal-catalyzed reactions. Diversity was
incorporated into the amino acid portion of the molecules
by using amino acids comprising three different classes:
alanine (nonpolar aliphatic R2 group, entries a-e, l-n),
phenylalanine (aromatic R2 group, entries f-k), or TBS-
protected serine (polar uncharged R2 group, entry o). The
(3) Brummond, K. M.; Chen, H.; Fisher, K. D.; Kerekes, A. D.; Rickards,
B.; Sill, P. C.; Geib, S. J. Org. Lett. 2002, 11, 1931. Brummond, K. M.;
Gao, D. Org. Lett. 2003, 5, 3491. Anomalies have been observed, especially
if a coordinating group resides near the proximal double bond of the allene,
which then directs the reaction to that double bond (unpublished results
from our group).
(6) Bolton, G. L. Tetrahedron Lett. 1996, 37, 3433. Bolton, G. L.;
Hodges, J. C.; Rubin, J. R. Tetrahedron 1997, 53, 6611. Bolton, G. L.;
Hodges, J. C. J. Comb. Chem. 1999, 1, 130.
(4) Brummond, K. M.; Chen, H.; Sill, P.; You, L. J. Am. Chem. Soc.
2002, 124, 15186. Shibata subsequently reported a similar finding: Shibata,
T.; Takesue, Y.; Kadowaki, S.; Takagi, K. Synlett 2003, 268.
(5) For preliminary investigations, see: Brummond, K. M.; Chen, H.;
Mitasev, B.; Casarez, A. D. Org. Lett. 2004, 6, 0000. Initial disclosure of
this type of cyclization: Abstracts of Papers, 226th National Meeting of
the American Chemical Society, New York, NY, Sept 7, 2003; American
Chemical Society: Washington, DC, 2003. For related observations, see:
Makino, T.; Itoh, K. Tetrahedron Lett. 2003, 44, 6335. Makino, T.; Itoh,
K. J. Org. Chem. 2004, 69, 395.
(7) Castelhano, A. L.; Pliura, D. H.; Taylor, G. J.; Hsieh, K. C.; Krantz,
A. J. Am. Chem. Soc. 1984, 106, 2734. Castelhano, A. L.; Horne, S.; Taylor,
G. J.; Tetrahedron 1988, 44, 5451.
(8) Kazmaier, U.; Gorbitz, C. H. Synthesis 1996, 1489.
(9) See Supporting Information section for experimental/spectroscopic
data for compounds that were not prepared by the authors in refs 7 and 8.
(10) Method II could be used to generate the allenic amino acids by
increasing the equivalents of LDA (2.5 equiv instead of 1.2 equiv), but the
product was formed as a 1:1 mixture of diastereomers. After the Claisen
rearrangement, the silicon group was removed.
2246
Org. Lett., Vol. 6, No. 13, 2004

terminus of the alkyne (R3) has been systematically varied
to include H, alkyl, aryl, or silyl groups. For this initial study,
the R1 group on the allene terminus has only been varied
slightly and is either an H or an alkyl group. The amine
protecting group, while not considered to be a diversification
element, is instead designed so that it is not limiting after
carbocyclization (Cbz, Boc, and Bz).
terminus of the alkyne (R3), the reaction conditions were
tolerant to a hydrogen (entries a, f), alkyl (entries b, e, g, n,
o), TMS (entries c, h), or aryl groups (entries d, i). The only
variations made so far on the terminus of the allene (R1) are
a methyl and an isopropyl group. A major reason for this is
that the R4 groups on the appending olefin are the same,
and thus E/Z isomers are not an issue.11
The pivotal substrates 6a-o were then systematically
subjected to the transition metal-catalyzed reaction conditions
enumerated in Scheme 2. Initially, the Alder-ene reaction
was investigated. Treatment of a representative group of
alkynyl allenes 6 with 5 mol % [Rh(CO)2Cl]2 in toluene at
room temperature afforded the cross-conjugated trienes 7 in
high yields (74-95%) in less than 10 min for each entry in
Table 2 (entries a-j, o). This is a substantial rate increase
Next, we examined these substrates in the Rh(I)-catalyzed
Pauson-Khand reaction to form 4-alkylidene cyclopenten-
ones. Previously, we had shown that by using rhodium
biscarbonyl chloride dimer and simply changing the atmo-
sphere to carbon monoxide, only CO insertion products were
obtained without any of the cross-conjugated triene. How-
ever, due to the exceedingly fast rate of triene 7 formation
for substrates 6, trienes were the only products isolated even
under a CO atmosphere, with CO only serving to slow the
triene formation (1 h vs 10 min). Since it is postulated that
the triene 7 and alkylidene 8 arise from the same rhodium
metallocycle, it was hoped that by systematically changing
the metal, ligands, metal counterion, and solvent, we could
identify conditions for the desired CO insertion process. The
best conditions were found to be 10 mol % rhodium
biscarbonyl chloride dimer, 30 mol % triphenylphosphine,
and 22 mol % silver tetrafluoroborate in dichloroethane at
40 °C. By using these newly developed conditions, alkynyl
allene 6 was converted to 4-alkylidene cyclopentenone 8
selectively in good yields for all entries in Table 3 except
Table 1. Assembly of Alkynyl Allene 6
Table 3. Formation of 4-Alkylidene Cyclopentenones
compared to those published previously (minutes vs hours).
We attribute this in part to a Thorpe-Ingold effect arising
from the carbomethoxy and alkyl/aryl functionality. Interest-
ingly, these facile cyclizations did not require an inert
atmosphere. All functional groups and substituents present
in compound 6 were tolerant of the Alder-ene reaction
conditions. Both N-benzoyl- (entries f-i, o) and carbamate-
protected (entries a-e, n) substrates were tested with no
apparent difference in the rate of the reaction. For the
a Triene formation was not possible, so conditions used were 5 mol %
[Rh(CO)2Cl]2, DCE, CO, rt, 2.5 h. b Diastereomeric ratio was 1.7:1. c Ratio
of [6,5]:[5,5] CO insertion products was 5:1. d For entries 6a, f, k, and m,
the reaction was not attempted since conditions are not compatable with a
terminal alkyne; For entry 6j, the reaction was not attempted but should
work based upon entry l.
Table 2. Formation of Cross-Conjugated Trienea
entry f, showing that these reaction conditions were not
tolerant of a hydrogen atom on the terminus of the alkyne.
All other changes to R3 groups occurred without problems;
i.e., R3 ) TMS (entries c, h), alkyl (entries b, e, g, l-o),
and aryl (entries d, i). Both benzoyl and carbamate protecting
groups and all three classes of amino acids were tolerated
during this reaction.
Next, conversion of a subset of substrates 6 to R-alkylidene
cyclopentenones 9 was accomplished by heating these
(11) E/Z selectivity of this double bond can be controlled using an Ir(I)
catalyst (ref 4); however, a TMS group on the terminus of the alkyne is
required, thus limiting the diversity of R3.
a Entries j, k, l, and m cannot react to give triene 7.
Org. Lett., Vol. 6, No. 13, 2004
2247

allenynes with molybdenum hexacarbonyl and DMSO (Table
4). Unfortunately, a mixture of E/Z isomers of 9 were
product mixtures obtained were so complex that they could
not be separated (entries f, g, o). While these substrates give
some of the highest yields observed to date for the
molybdenum-mediated Pauson-Khand reaction, the low
diastereoselectivities, the nonselective reaction with the
double bonds of the allene, the mixture of E/Z isomers, and
the use of stoichiometric molybdenum make this reaction
pathway less than ideal for application to library synthesis
in its present state of development.
Table 4. Formation of R-Alkylidene Cyclopentenones
In conclusion, by taking advantage of competing reaction
pathways available via transition metal-catalyzed carbocy-
clization processes, we have shown the potential generality
of converting a common intermediate to at least three
structurally unique and heretofore unknown hetereocycles.
Furthermore, as a result of this study, we have developed
optimized conditions to control reaction pathway selectivity
by changes to the transition metals. Finally, while the
compounds generated thus far are new and will be tested
for biological activity, it is the unique reactivity profile of
each scaffold (R-alkylidene cyclopentenone, 4-alkylidene
cyclopentenone, and cross-conjugated triene) and its potential
to be differentially functionalized that are deemed to be the
most useful feature of this approach.
a Mixture (20:1) of 9:8 gave a 9Z:9E ratio of 7:1. b 9E:9Z ratio was
6.4:1. c 9E:9Z ratio was 7.6:1. d Mixture (5.7:1) of 9:8; dr of 9 was 6.4:1.
e Mixture (3:1) of 9:8. f These entries afforded an inseparable mixture of
diastereomers of 9Z, 9E, and 8. g Mixture (14:1) of 9:8; dr for 9 was 11:1.
h Mixture (14:1) of 9:8; dr for 9 was 10:1. i 9Z:9E ) 5:1. j For entries d
and i, the reaction afforded a very complex mixture of products. For entries
c and h, the reaction afforded a complex mixture of silylated and desilylated
products.
Acknowledgment. We gratefully acknowledge the fi-
nancial support provided by the National Institutes of Health
(P50 GM067982), the University of Pittsburgh Center for
Chemical Methodologies and Library Development, and
Stefan Werner, from the Curran group, for providing
compounds for entry n, Table 1, and entry j, Table 2.
obtained for entries a, b, e, f-i, n, o. Changing the R1 group
on the terminus of the allene to a hydrogen solved this
problem but led to the formation of both 8 and 9, with 9
predominating (entries j-m). This increase in reaction with
the distal double bond of the allene was attributed to the
quaternary center next to the internal double bond by
sterically directing the reaction to the more accessible olefin.
This result is unique from our previously published examples,
where none of the cases possessed a quaternary center vicinal
to the allene. Finally, the low diastereoselectivities observed
in the formation of the benzamide-protected allenic amino
acids proved to be a serious problem. In some cases, the
Supporting Information Available: Characterization
data and full experimental procedures for all compounds in
Tables 1-4 except entry n, Table 1, and entry j, Table 2,
which will be the topic of a future publication. This material
is available free of charge via the Internet at http://pubs.acs.org.
OL0492391
2248
Org. Lett., Vol. 6, No. 13, 2004