Asymmetric synthesis of dihydroindanes by convergent alkoxide-directed metallacycle-mediated bond formation

Article abstract of DOI:10.1021/ja2105043

A convergent synthesis of highly substituted and stereodefined dihydroindanes is described from alkoxide-directed Ti-mediated cross-coupling of internal alkynes with substituted 4-hydroxy-1,6-enynes (substrates that derive from 2-directional functionaliza

Full text of DOI:10.1021/ja2105043

Article
pubs.acs.org/JACS
Asymmetric Synthesis of Dihydroindanes by Convergent Alkoxide-
Directed Metallacycle-Mediated Bond Formation
Stephen N. Greszler, Holly A. Reichard, and Glenn C. Micalizio*
Department of Chemistry, The Scripps Research Institute, Scripps Florida, Jupiter, Florida 33458, United States
S
* Supporting Information
ABSTRACT: A convergent synthesis of highly substituted
and stereodefined dihydroindanes is described from alkoxide-
directed Ti-mediated cross-coupling of internal alkynes with
substituted 4-hydroxy-1,6-enynes (substrates that derive from
2-directional functionalization of readily available epoxy alcohol
derivatives). In addition to describing a new and highly stereo-
selective approach to bimolecular [2 + 2 + 2] annulation that
delivers products not available with other methods in this area of chemical reactivity, evidence is provided to support annulation by way
of regioselective alkyne−alkyne coupling, followed by metal-centered [4 + 2] rather than stepwise alkene insertion and reductive
elimination. Overall, the reaction proceeds with exquisite stereochemical control and defines a convenient, convergent, and
enantiospecific entry to fused carbocycles of great potential value in target-oriented synthesis and medicinal chemistry.
Pioneered by Vollhardt, metal-mediated [2 + 2 + 2] annulation
has surfaced as a useful strategy for the synthesis of dihydro-
indanes.10 First demonstrated in a wholly intramolecular fashion,
recent studies have begun to describe the utility of such processes
in intermolecular settings,11 where the strategic benefit of con-
vergency can dramatically enhance both the efficiency and
flexibility of synthetic pathways to complex carbocyclic targets.
This area of chemical reactivity, however, remains in its infancy,
with intermolecular reactions being plagued by challenges
associated with overcoming barriers associated with chemical
reactivity (i.e., sluggish reactivity of substrates bearing substituted
alkenes and/or sterically hindered internal alkynes), controlling
regioselectivity (i.e., site of C−C bond formation on unsymmetrical
systems) and stereoselection.12 As such, the utility of [2 + 2 + 2]
annulation in target-oriented synthesis has been broadly constrained
to the world of intramolecular cycloisomerization chemistry.10
In the context of a program aimed at the control of intermolecular
metallacycle-mediated C−C bond formation, we have been
actively engaged in the study of alkoxide-directed approaches to
achieve fine control of metallacycle reactivity. These studies
have led to the description of new intermolecular or convergent
coupling reactions that have overcome the substantial
limitations associated with prior art [including the difficulty
in achieving sufficient reactivity and selectivity to accomplish
bimolecular union of differentially substituted internal alkynes
(1 + 2 → 3; Figure 2A)13 and the cross-coupling of internal
alkynes with substituted alkenes (4 + 2 → 5; Figure 2B)14].
Here, we describe a general solution to the asymmetric
synthesis of densely functionalized hydroindanes that builds on
these earlier successes. In short, we report a highly stereoselective
alkoxide-directed Ti-mediated intermolecular annulation between
INTRODUCTION
■
Highly substituted and stereodefined hydroindanes define a
structural motif that is encountered in natural and synthetic small
molecules of broad pharmacological and biological relevance.
Examples include complex natural products like cortistatin,1 batra-
chotoxin,2 ouabain,3 and cephalostatin4 (Figure 1A), as well as a
Figure 1. Dihydroindanes in natural products and pharmaceuticals.
large collection of less functionalized steroids that are prescribed
daily for pharmaceutical intervention across a diverse landscape of
therapeutic areas (Figure 1B).5 Robust chemical methods based on
cycloaddition,6 cation-olefin cyclization,7 and Robinson annulation8
are typically embraced as strategies of choice for the assembly of
such systems, yet these venerable methods can be challenged when
confronted with the goal of accessing enantiodefined, highly sub-
stituted and/or oxygenated systems.9
Received: November 8, 2011
Published: January 10, 2012
© 2012 American Chemical Society
2766
dx.doi.org/10.1021/ja2105043 | J. Am. Chem.Soc. 2012, 134, 2766−2774

Journal of the American Chemical Society
Article
Figure 2. Metallacycle-mediated cross-coupling of alkenes and alkynes.
readily available 4-hydroxy-1,6-enynes and internal alkynes as a
strategy to access densely functionalized hydroindanes (Figure 2C).
This reaction, while related to previous metallacycle-mediated
transformations (i.e., [2 + 2 + 2] annulation chemistry), is to
our knowledge the first that is capable of delivering stereo-
defined angularly substituted hydroindanes in a convergent and
highly selective manner using unsymmetric and nontethered
reaction partners.
Figure 3. Established metallacycle-mediated annulation chemistry of
relevance to the present advance.
and substituted alkenes. While a number of metallacycle-mediated
reactions of this ilk are known, all available transformations are sub-
stantially limited in substrate scope. While reductive cross-coupling
of alkynes is typically accomplished with one terminal alkyne,18
related union of internal alkynes with substituted alkenes has
remained useful in only the simplest of cases, that is, with un-
hindered terminal mono-substituted alkenes.19 In fact, from a
broader perspective, few reactions in organic chemistry are suitable
for the regioselective carbometalation of an electronically unactivated
alkene or alkyne, especially if such systems are 1,2-disubstituted.
In an attempt to overcome the well established barrier to
bimolecular reactivity of preformed metal−alkyne complexes
with other electronically unactivated π-systems, we speculated
that an alkoxide-directed approach as depicted in Figure 4A,B may
offer a convenient solution. In this way, rapid and reversible ligand
exchange would serve as a mechanism to render the carbo-
metalation reaction intramolecular, and provide a means for step-
wise encapsulation of the metal center by an orchestrated sequence
of metal−ligand, metal−carbon, and carbon−carbon bond-forming
processes. Importantly, in addition to providing a pathway to
overcome the barriers to reactivity associated with scores of
potential metallacycle-mediated cross-coupling processes, this
strategy would define a mechanism to control regioselection that
is distinct from established processes based on steric or electronic
differentiation of the reacting π-systems.
As depicted in Figure 3, Ti-alkoxide-based methods have
been described for intermolecular [2 + 2 + 2], yet these
methods yield substituted aromatics15 (Figure 3A) or nearly
symmetric cyclohexadienes16 (Figure 3B). In this latter case,
intramolecular diyne cyclization delivers a symmetrical metal-
lacyclopentadiene that is converted to a carbocyclic product by
way of a process that avoids any challenges with regard to
regioselection. Other metallacycle-mediated chemistry has been
reported to access angularly substituted hydroindane products
as depicted in Figure 3C,D.10c,17 These methods represent
impressive demonstrations of intramolecular chemistry but, like
many powerful modes of reactivity suitable for intramolecular
C−C bond formation, these metallacycle-mediated bond
constructions have not been demonstrated to be similarly
useful in intermolecular settings. Given the significance of
convergency in realizing step-economical pathways to complex
molecules, this restriction to intramolecularity marks a
significant limitation in the synthetic utility of these latter methods.
In short, the chemistry described herein marks a substantial
advance in metallacycle-mediated annulation chemistry, offer-
ing a reaction process that enables bimolecular [2 + 2 + 2]
annulation as an entry to complex angularly substituted
hydroindanes, structural motifs that are encountered in a
diverse array of natural products and synthetic molecules of
pharmacological relevance.
To accomplish such a transformation, we accepted that
catalytic versions of metallacycle-mediated coupling would not
be reasonable as a general solution to the problem. This
conclusion derived from the expected difficulty of favoring
cross-coupling over homodimerization in a system where both
RESULTS
■
Initial investigation of intermolecular metallacycle-mediated coupling
chemistry was focused on the union of internal alkynes with
notoriously sluggish reaction partners, that is, other internal alkynes
2767
dx.doi.org/10.1021/ja2105043 | J. Am. Chem.Soc. 2012, 134, 2766−2774

Journal of the American Chemical Society
Article
Figure 4. Alkoxide-directed alkyne−alkyne and alkene−alkyne reductive cross-coupling.
reactive π-systems are available (in large excess) to the catalytic
metal center. Also, we aimed to achieve ligand-directed C−C
bond formation with a ubiquitous free hydroxy group rather
than with phosphines, thioethers, or terminal alkenes.20 The
combination of these factors led to the selection of established
and inexpensive Ti(IV) alkoxides as the metal of choice to
accomplish alkoxide-directed metallacycle-mediated cross-cou-
pling as generalized in Figure 4B,C. In addition to Ti(Oi-Pr)4
being easy to handle and purify in the absence of a glovebox,
this reagent is inexpensive, nontoxic, and on aqueous workup
delivers byproducts that are easily removed from product
mixtures (TiO2, i-PrOH, and MgX2 salts−from the RMgX-
mediated reduction of the Ti(IV) alkoxide). For these reasons,
we moved forward with the development of stoichiometric
Ti(IV)-mediated alkoxide-directed metallacycle-mediated cross-
coupling chemistry.
group (i.e., → 16). And, most interestingly, coupling of two
unsymmetrical alkyne coupling partners proved to be highly
selective, delivering tetrasubstituted 1,3-dienes 17 and 18 as
single isomers. Here, exquisite regioselectivity is observed in the
carbometalation of both π-systems. This observation was a
welcomed surprise, as related coupling reactions of internal
alkynes with carbonyl electrophiles do not typically proceed
with such high levels of regioselection.21 While the mechanistic
origins of this phenomenon remain to be determined, we
suspect that the transition states associated with the present
coupling reactions are significantly more product-like, as one
would suspect that these processes are less exothermic than the
analogous reactions of metal−alkyne complexes with aldehydes.
As such, the closer distance between the reactive loci in the
transition state should result in an enhancement of regio-
selection due to steric differences between the alkyne termini.
We note that these reductive cross-coupling reactions
uniformly proceed with exceptional levels of selectivity, albeit
in modest yield (51−68%). This compromise is associated with
the frequent inability to push these reactions to completion
and/or the observation of homodimer derived from the alkynyl
alcohol coupling partner. This second observation likely arises
from the use of a slight excess of Ti(Oi-Pr)4 to efficiently
convert the first alkyne coupling partner (2.5 equiv) to the
metal−alkyne complex.
Moving on to the more demanding coupling reactions
between alkynes and alkenes, we were delighted to find that the
general approach taken with alkyne−alkyne coupling was also
effective here. As illustrated in Figure 5B, a range of products
can be generated from the union of internal alkynes with:
(1) terminal-(→ 19), (2) (E)-disubstituted-(→ 20), (3) (Z)-
disubstituted-(→ 20), and (4) 1,1-disubstituted alkenes (→ 21).
In all cases, the site of C−C bond formation is completely
controlled by the position of the free hydroxy group in the
homoallylic alcohol coupling partners. Further, as seen in the
alkyne−alkyne coupling process, the union of unsymmetrical
alkynes with unsymmetrical homoallylic alcohols also proceeds
As summarized in Figure 5, this strategy for the control of
metallacycle-mediated C−C bond formation proved to be quite
effective and general.13,14 In the case of alkyne−alkyne coupling
(Figure 5A), successful reactions were achieved with homo-
propargylic and bis-homopropargylic alcohols, delivering 1,3-
diene products as single regio- and stereoisomers. In the case of
products 9 and 10, the regioselectivity reported (≥42:1 and
1
≥65:1) is based on the signal-to-noise ratio of each H NMR
spectrum of the crude product mixture, as no evidence could be
found for presence of the other regioisomeric product.
Symmetrical nonconjugated alkynes are also viable coupling
partners in this diene synthesis (→ 11; → 12). Further, while
demonstrating a powerful means to overcome the barriers
associated with reactivity for the cross-coupling process, this
directing group strategy proved quite effective for overriding
steric effects in the site-selective C−C bond forming process
(i.e., for 13−15, C−C bond formation occurs at the site distal
to the free hydroxyl, independent of the nature of the proximal
substituent: Et, i-Pr, t-Bu).
This general strategy also proved to be effective in coupling
substrates that contain a hindered secondary alcohol directing
2768
dx.doi.org/10.1021/ja2105043 | J. Am. Chem.Soc. 2012, 134, 2766−2774

Journal of the American Chemical Society
Article
Figure 5. Examples of alkoxide-directed reductive cross-coupling.
with exquisite levels of regioselection with respect to both
reacting π-systems (→ 25 and 26). In these final cases, we
observed a general lack of influence of double asymmetric
relationships on the stereochemical course of bond formation.
With a firm foundation of preliminary data that confirms the
ability to employ a hydroxy-group directing strategy to enable
intermolecular metallacycle-mediated union of preformed
metal−alkyne complexes with internal alkynes or substituted
alkenes, we moved on to investigate the inherent reactivity of
substrates that could react by either mode of reactivity in
isolation, or by a process that engaged all three π-systems in an
annulation process.
by addition of the lithium alkoxide of an enyne coupling partner
(27) may result in either: (1) alkene−alkyne coupling (to
deliver A),14 (2) alkyne−alkyne coupling (to furnish B),13 or
(3) [2 + 2 + 2] annulation (to produce the functionalized
carbocycle C).15,16 As illustrated in Figure 6B, alkene
substitution plays a dominant role in affecting product
distribution.22 If the alkene is unsubstituted, reaction proceeds
by alkene−alkyne coupling to deliver 30 (eq 1). Alternatively,
substrates that contain terminal substitution on the alkene react
by alkyne−alkyne coupling (eqs 2 and 3). Finally, a substrate
bearing a simple 1,1-disubstituted alkene (35) undergoes reaction
by yet a different course; here, convergent union engages all three
π-systems and proceeds by way of a highly stereoselective [2 + 2 + 2]
As illustrated in Figure 6A, preformation of a Ti-alkyne complex
of 28 (Ti(Oi-Pr)4, c-C5H9MgCl, Et2O, −78 to −30 °C), followed
Figure 6. Path selectivity in alkoxide-directed metallacycle-mediated cross-coupling. Reaction conditions: (a) (PMBOCH2CH2C)2, Ti(Oi-Pr)4,
c-C5H9MgCl, (−78 to −30 °C), then add enyne as the corresponding Li-alkoxide (−30 °C); (b) identical procedure as described for “a”, except that
this reaction was warmed to 0 °C. If this reaction (35 → 36) was quenched at −30 °C, little change in overall yield was observed (77%). For entries
2 and 3, if these reactions were warmed to 0 °C prior to quenching, no hydroindane could be identified from the product mixture.
2769
dx.doi.org/10.1021/ja2105043 | J. Am. Chem.Soc. 2012, 134, 2766−2774

Journal of the American Chemical Society
Article
annulation, delivering the functionalized dihydroindane product
36 in 83% yield as a single stereoisomer.
case (rs = 2:1), the stereochemically defined product 39 possesses
substitution of significance in the context of de novo steroid
synthesis [see C8, C9 and C11 (steroid numbering)]. In efforts to
explore the scope of this annulation process, we observed
enhancements in regioselection as a function of the coupling
partners employed. First, regioselection was seen to increase slightly
when using the TMS-substituted alkyne 41 that contains an
aliphatic substituent (CH2CH2OPMB) rather than Ph (as in 38).
Coupling of 40 with alkyne 41 proceeds in 75% yield to deliver a
3:1 mixture of regioisomeric carbocycles, the major isomer of which
is the hydroindane product 42 (dr ≥ 20:1; entry 2). Regioselection
also improved in the coupling reaction of TMS-substituted enyne
44 with 4-hydroxy-1,6-enyne 43 (entry 3). Here, the annulation
event proceeds with 4:1 regioselection, ≥20:1 diastereoselection,
and the major isomer 45 can be isolated in 53% yield as a single
isomer. Finally, when the alkene of the 4-hydroxy-1,6-enyne
possesses a more sterically demanding substituent (as compared
to Me), regioselection also increases (entry 4). Here, union of the
benzyl-substituted substrate 46 with TMS-alkyne 47 proceeds with
5:1 regioselection, ≥20:1 diastereoselection, and the major product
48 can be isolated in 55% yield as a single isomer.23b While in
entries 3 and 4 of Table 1 the major isomer could be separated from
the minor regioisomer, it is important to note that in all cases,
purification of the major isomer is facile after chemoselective
protodesilylation of the major products’ vinylsilane (see Supporting
Information for details).
At this point, we speculated that the unique sequence of
bond-forming events inherent to these alkoxide-directed
coupling reactions that was previously described as a “stepwise
encapsulation of the metal center by metal−ligand, metal−
carbon, and carbon−carbon bond-forming reactions” may
define a powerful new direction for the control of metal-
centered [2 + 2 + 2] chemistry. On the basis of this perspective,
we pursued the study of regio- and stereoselection in this
unique annulation process in hopes of achieving a robust
strategy for the synthesis of densely functionalized and
stereodefined hydroindane systems that are not accessible
with current [2 + 2 + 2] annulation methods.
As illustrated in Table 1, a range of 4-hydroxy-1,6-enynes
and substituted alkynes are effective coupling partners in this
Table 1. Initial Studies Regarding Regioselection and
Enantiospecificity
Notably, this annulation process can be employed to generate
hydroindane products possessing C17 β-substitution (Table 2). As
depicted in entry 1, coupling of enyne 49 with alkyne 50 results in
the formation of hydroindane products possessing a C17 β-Me
substituent in 63% yield and ≥20:1 diastereoselection, albeit as a 3:1
mixture of regioisomers. Interestingly, when exploring the utility of
this annulation process for the synthesis of hydroindanes possessing
C17 β-aryl substitution (a prominent architectural feature of the
cortistatins), we discovered that this process can proceed with
exquisite regiochemical control. As illustrated in entry 2, coupling of
hydroxy-enyne 52 with alkyne 41 delivers hydroindane product 53
in 64% yield, with ≥20:1 ds and 13:1 rs. While the mechanistic
origin of this long-range effect on regioselection remains unclear, all
enyne substrates bearing an allylic Ph-substituent underwent Ti-
mediated annulation reactions with high levels of regioselectivity
(typically ≥20:1), and consistently exquisite levels of diastereose-
lection (uniformly ≥20:1) (entries 2−7).
While this annulation process is quite effective at generating
17-β-aryl substituted hydroindanes as essentially single regio
and stereoisomers (i.e., 53, 56, 58, 59, 61 and 63), the process
is sensitive to the relative stereochemistry of the enyne starting
material. As illustrated in entry 8 of Table 2, annulation of
enyne 64 with TMS-phenylacetylene (50) produced a complex
mixture of products from which only 22% yield of the C17
α-substituted hydroindane product (65) could be obtained.
While this stereochemical pattern certainly represents a pre-
parative limitation to the current chemistry, product 65 could
easily be isolated as a single regio- and stereoisomer.
a
Reaction conditions: alkyne, Ti(Oi-Pr)4, c-C5H9MgCl, toluene (−78
b
to −30 °C). Reaction conditions: cool to −78 °C and add lithium
alkoxide of enyne, warm to a final temperature of between −15 and
c
d
23 °C. Yield reported is for the mixture of regioisomers. Yield
reported is for the major regioisomer. See Supporting Information for
additional information.
Mechanistic Hypothesis. Metal-centered [2 + 2 + 2]
annulation chemistry has long been proposed to occur by initial
formation of a metallacyclopentadiene, followed by either:
stereoselective annulation process. Further, this reaction proceeds in
an enantiospecific fashion.23a As depicted in entry 1, union of the
optically active enyne 37 (95% ee) with Me2PhSi-alkyne 38 occurs
in 74% yield to deliver a mixture of regioiomeric hydroindane
products, the major isomer of which is the stereodefined
hydroindane 39 (95% ee; dr ≥ 20:1). While regiochemical control
in the functionalization of alkyne 38 was not particularly high in this
(1) [4 + 2]/cycloreversion (Figure 7A), or
(2) alkene insertion/reductive elimination (Figure 7B).
In our efforts to explore the scope of this Ti-mediated annul-
ation, we have come upon an experiment that distinguishes
between these two pathways, and further provides evidence in
2770
dx.doi.org/10.1021/ja2105043 | J. Am. Chem.Soc. 2012, 134, 2766−2774

Journal of the American Chemical Society
Article
Table 2. C17 β-Aryl Substitution and Regioselection
Figure 7. Potential pathways for metal-centered annulation.
Figure 8. Development of an empirical model for the Ti-mediated
intermolecular hydroindane synthesis.
selective annulation is then proposed to occur through a sequence
of steps that includes alkoxide-promoted cleavage of the oxametall-
acyclopentane (68 → 69), followed by stereoselective intra-
molecular [4 + 2] (69 → 70), a highly stereoselective process that
is thought to be controlled by intramolecular coordination of the
pendant alkoxide to a ligand on Ti. Unlike previous examples
depicted in Table 1, intermediate 70 is suitably functionalized to
undergo vinylogous elimination of the PMB ether to generate an un-
stable tertiary allyltitanium species 71. Finally, 1,3-metallotropic
shift generates allyltitanium species 72 which, on protonolysis with
allylic transposition, is poised to deliver the observed product 67.
Alternatively, if annulation proceeded by a sequence defined by
alkene insertion and reductive elimination (Figures 7B and 8C),
a suitable intermediate is not readily apparent to support the
facile elimination of the PMB ether [i.e., 73 (Figure 8B) vs 70
(Figure 8C)].24
a
Reaction conditions: alkyne, Ti(Oi-Pr)4, c-C5H9MgCl, toluene (−78
b
to −30 °C). Reaction conditions: cool to −78 °C and add lithium
alkoxide of enyne, warm to a final temperature of between −15 and
c
d
23 °C. Yield reported is for the mixture of regioisomers. Precise
1
selectivities could not be determined from the H NMR spectrum of
the crude product; hydroindane product 65 was isolated as a single
regio- and stereoisomer. See Supporting Information for additional
information
support of a sequential [4 + 2]/cycloreversion in preference to
alkene insertion/reductive elimination.
As illustrated in Figure 8A, coupling of enyne 66 with alkyne
55 delivers a carbocyclic product (67) that lacks the PMB ether
of the starting material and contains an exocyclic alkene.
This observation is consistent with the empirical model depicted in
Figure 8B, where initial alkyne−alkyne coupling delivers metal-
lacyclopentadiene 68 in a regioselective fashion.13 sequent stereo-
Overall, alkoxide-directed intermolecular [2 + 2 + 2]
annulation between an internal alkyne and a 1,6-enyne defines
a unique and powerful approach to controlling regio- and
stereoselection in this class of chemical reactivity. The virtues of
stepwise encapsulation of the metal center (through a process
of sequential ligand exchange and intramolecular carbometala-
2771
dx.doi.org/10.1021/ja2105043 | J. Am. Chem.Soc. 2012, 134, 2766−2774

Journal of the American Chemical Society
Article
tion) set the stage for a highly controlled series of bond
forming events en route to a functionalized carbocyclic system
of great potential utility in natural product synthesis. Further,
the mechanistic course of this process (which is more
consistent with [4 + 2]/cycloreversion than alkene insertion/
reductive elimination) may afford the opportunity to divert
chemical reactivity away from the expected [2 + 2 + 2]
products (i.e., Table 1) to stereodefined products that take
advantage of the unique reactivity of bridged polycyclic allylic
metal intermediates (as demonstrated in Figure 8A).
Utility of the Carbocyclic Products. The present annulation
process delivers complex carbocycles that can be differentially
functionalized to address a range of common substitution
patterns about the hydroindane core of natural products and
small molecules of pharmacological relevance. As depicted in
Figure 9, deoxygenation of the D-ring hydroxy group proceeds
from two-directional functionalization of chiral epoxy alcohol
derivatives (including epichlorohydrin), the synthetic pathway
secured is also amenable to the production of optically active
hydroindanes that bear a variety of functionality of likely utility
in target-oriented synthesis. While mechanistic ambiguity has
traditionally surrounded related [2 + 2 + 2] annulation
chemistry, our studies have led to the conclusion that sequential
[4 + 2] cycloaddition/cycloreversion is likely operative in
preference to alkene insertion/reductive elimination. From a
preparative perspective, the convergent annulation described is
efficient (isolated yields ranging from 48 to 83%), regioselective
(up to ≥20:1), and highly stereoselective (dr is uniformly
≥20:1). We are unaware of another intermolecular metallacycle-
mediated annulation reaction capable of delivering stereo-
defined products of related structure to that observed here
structural motifs that are core skeletal components of a vast
array of pharmacologically active natural and synthetic
substances. On the basis of the considerations discussed, the
current science defines a significant advance in metallacycle-
mediated annulation chemistry, and marks the establishment of
an exceptionally concise means for the convergent assembly of
densely functionalized hydroindanes. As such, we anticipate
that the current findings will be of great potential utility in
chemical synthesis and medicinal chemistry.
ASSOCIATED CONTENT
■
S
* Supporting Information
Experimental procedures and tabulated spectroscopic data for
new compounds. This material is available free of charge via the
Internet at http://pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Author
micalizio@scripps.edu
ACKNOWLEDGMENTS
■
We gratefully acknowledge financial support of this work by the
National Institutes of Health−NIGMS (GM80266 and
GM80266-04S1). The authors also acknowledge Hayley
Browdy for assistance in synthesizing the panel of enynes
depicted in Figure 6B.
Figure 9. Some functionalization processes suitable for elaboration of
the dihydroindane products.
REFERENCES
■
(1) (a) Narayan, A. R. H.; Simmons, E. M.; Sarpong, R. Eur. J. Org.
Chem. 2010, 3553−3567. (b) Flyer, A. N.; Si, C.; Myers, A. G. Nat.
Chem. 2010, 2, 886−892. (c) Aoki, S.; Watanabe, Y.; Sanagawa, M.;
Setiawan, A.; Kotoku, N.; Kobayashi, M. J. Am. Chem. Soc. 2006, 128,
3148−3149. (d) Shenvi, R. A.; Guerrero, C. A.; Shi, J.; Li, C.-C.;
Baran, P. S. J. Am. Chem. Soc. 2008, 130, 7241−7243. (e) Lee, H. M.;
Nieto-Oberhuber, C.; Shair, M. D. J. Am. Chem. Soc. 2008, 130,
16864−16866. (f) Nicolaou, K. C.; Peng, X.-S.; Sun, Y.-P.; Polet, D.;
Zou, B.; Lim, C. S.; Chen, D. Y.-K. J. Am. Chem. Soc. 2009, 131,
10587−10597. For an approach to the core skeleton of the cortistatins
by coupling of a functionalized hydroindane with an aromatic A-ring
precursor, see: (g) Dai, M.; Danishefsky, S. J. Tetrahedron Lett. 2008,
49, 6610−6612.
readily by the Barton procedure (61 → 75).25 Alternatively,
C11 functionality (steroid numbering) can easily be introduced
by protodesilylation26 and hydroboration (61→ 76). Finally, a
sequence of oxidation, isomerization, and protodesilylation is
suitable to access a CD-ring system bearing a dienone (59 → 77).
While certainly not comprehensive, the transformations
depicted in Figure 9 serve to introduce the great potential
utility of the functionalized dihydroindane products derived
from the Ti-mediated annulation reaction in the stereoselective
synthesis of complex carbocycles.
(2) (a) Daly, J. W.; Witkop, B.; Bommer, P.; Biemann, K. J. Am.
Chem. Soc. 1965, 87, 124−126. (b) Dumbachaer, J. P.; Beehler, B. M.;
Spande, T. F.; Garraffo, H. M.; Daly, J. W. Science 1992, 258, 799−801.
(c) Tokuyama, T.; Daly, J. J. Am. Chem. Soc. 1969, 91, 3931−3938.
(d) Albuquerque, E. X.; Daly, J. W.; Witkop, B. Science 1971, 172,
995−1002. (e) Kurosu, M.; Marcin, L. R.; Grinsteiner, T. J.;
Kishi, Y. J. Am. Chem. Soc. 1998, 120, 6627−6628.
CONCLUSION
■
We have described a convenient and highly stereoselective
pathway to a range of functionalized hydroindanes by the union
of simple 4-hydroxy-1,6-enynes with silyl-substituted alkynes.
Since the enyne-containing starting materials are easily derived
2772
dx.doi.org/10.1021/ja2105043 | J. Am. Chem.Soc. 2012, 134, 2766−2774

Journal of the American Chemical Society
Article
(3) (a) Prassas, I.; Diamandis, E. P. Nat. Rev. Drug Discovery 2008, 7,
926−935. (b) Newman, R. A.; Yang, P.; Pawlus, A. D.; Block, K. I.
Mol. Interventions 2008, 8, 36−49. (c) Riganti, C.; Campia, I.; Kopecka,
J.; Gazzano, E.; Doublier, S.; Aldieri, E.; Bosia, A.; Ghigo, D. Curr.
Med. Chem. 2011, 18, 872−885. (d) Overman, L. E.; Rucker, P. V.
Heterocycles 2000, 52, 1297−1314. (e) Reddy, M. S.; Zhang, H.;
Phoenix, S.; Deslongchamps, P. Chem. Asian J. 2009, 4, 725−741.
(4) (a) Lee, S.; LaCour, T. G.; Fuchs, P. L. Chem. Rev. 2009, 109,
2275−2314. (b) LaCour, T. G.; Guo, C.; Bhandaru, S.; Boyd, M. R.;
Fuchs, P. L. J. Am. Chem. Soc. 1998, 120, 692−707. (c) Moser, B. R.
J. Nat. Prod. 2008, 71, 487−491. (d) Rudy, A.; Lopez-Anton, N.;
Dirsch, V. M.; Vollmar, A. M. J. Nat. Prod. 2008, 71, 482−486.
(e) Fortner, K. C.; Kato, D.; Tanaka, Y.; Shair, M. D. J. Am. Chem. Soc.
2010, 132, 275−280.
(5) (a) Zeleen, F. J. Medicinal Chemistry of Steroids. In Principles of
Medical Biology, Volume 8B, Molecular and Cellular Pharmacology; JAI
Press, Inc.: Stamford, 1997; pp427−463. For reviews on the synthesis
of steroids, see: (b) Chapelon, A.-S.; Moraleda, D.; Rodriguez, R.;
Ollivier, C.; Santelli, M. Tetrahedron 2007, 63, 11511−11616.
(c) Jankowski, P.; Marczak, S.; Wicha, J. Tetrahedron 1998, 54,
12071−12150.
(6) (a) Parker, K. A.; Iqbal, T. J. Org. Chem. 1982, 47, 337−342.
(b) Wilson, S. R.; Jacob, L. J. Org. Chem. 1992, 57, 4380−4385.
(c) Wender, P. A.; Smith, T. E. Tetrahedron 1998, 54, 1255−1275.
(d) Taber, T. F.; Song, Y. J. Org. Chem. 1996, 61, 7508−7512.
(7) (a) Snider, B. B.; Kirk, T. C. J. Am. Chem. Soc. 1983, 105, 2364−
2368. (b) Johnson, W. S.; Elliot, J. D.; Hanson, G. J. J. Am. Chem. Soc.
1984, 106, 1138−1139. (c) Hatakeyama, S.; Numata, H.; Osanai, K.;
Takano, S. J. Chem. Soc., Chem. Commun. 1989, 1893−1895.
(d) Browder, C. C.; West, F. G. Synlett 1999, 9, 1363−1366.
(8) (a) Hajos, Z. G.; Parrish, D. R. J. Org. Chem. 1974, 39, 1612−
1615. (b) Stork, G.; Shiner, C. S.; Winkler, J. D. J. Am. Chem. Soc.
1982, 104, 310−312. (c) Stork, G.; Winkler, J. D.; Shiner, C. S. J. Am.
Chem. Soc. 1982, 104, 3767−3768. (d) Yamamoto, K.; Iijima, M.;
Ogimura, Y.; Tsuji, J. Tetrahedron Lett. 1984, 25, 2813−2816.
(e) Kotoku, N.; Sumii, Y.; Hayashi, T.; Kobayashi, M. Tetrahedron
Lett. 2008, 49, 7078−7081.
(11) (a) Evans, P. A.; Lai, K. W.; Sawyer, J. R. J. Am. Chem. Soc. 2005,
127, 12466−12467. (b) Shibata, T.; Arai, Y.; Tahara, Y. Org. Lett.
2005, 7, 4955−4957.
(12) For recent reviews that describe these challenges, see:
(a) Reichard, H. A.; McLaughlin, M.; Chen, M. Z.; Micalizio, G. C.
Eur. J. Org. Chem. 2010, 391−409. (b) Reichard, H. A.; Micalizio, G.
C. Chem. Sci. 2011, 2, 573−589.
(13) Ryan, J.; Micalizio, G. C. J. Am. Chem. Soc. 2006, 128, 2764−
2765.
(14) (a) Reichard, H. A.; Micalizio, G. C. Angew. Chem., Int. Ed.
2007, 46, 1440−1443. (b) Canterbury, D. P.; Micalizio, G. C. J. Am.
Chem. Soc. 2010, 132, 7602−7604.
(15) (a) Suzuki, D.; Urabe, H.; Sato, F. J. Am. Chem. Soc. 2001, 123,
7925−7926. (b) Suzuki, D.; Tanaka, R.; Urabe, H.; Sato, F. J. Am.
Chem. Soc. 2002, 124, 3518−3519. (c) Tanaka, R.; Yuza, A.; Watai, Y.;
Suzuki, D.; Takayama, Y.; Sato, F.; Urabe, H. J. Am. Chem. Soc. 2005,
127, 7774−7780. For earlier examples demonstrating the ability of Ti-
aryloxides to promote [2 + 2 + 2] annulation en route to substituted
aromatics, see: (d) Hill, J. E.; Balaich, G.; Fanwick, P. E.; Rothwell, I. P.
Organometallics 1993, 12, 2911−2924.
(16) (a) Sung, M. J.; Pang, J.-H.; Park, S.-B.; Cha, J. K. Org. Lett.
2003, 5, 2137−2140. For early examples demonstrating the ability of
Ti−aryloxides to generate cyclohexadienes from the coupling of two
alkynes and one alkene, see: (b) Balaich, G. J.; Rothwell, I. P. J. Am.
Chem. Soc. 1993, 115, 1581−1583. (c) Johnson, E. S.; Balaich, G. J.;
Rothwell, I. P. J. Am. Chem. Soc. 1997, 119, 7685−7693.
(17) Wender, P. A.; Smith, T. E. Tetrahedron 1998, 54, 1255−1275.
(18) (a) Hamada, T.; Suzuki, D.; Urabe, H.; Sato, F. J. Am. Chem. Soc.
1999, 121, 7342−7344. (b) See ref 15c. (c) Buchwald, S. L.; Watson,
B. T.; Huffman, J. C. J. Am. Chem. Soc. 1987, 109, 2544−2546.
(d) Buchwald, S. L.; Nielsen, R. B. J. Am. Chem. Soc. 1989, 111, 2870−
2874. (e) Shimp, H. L.; Micalizio, G. C. Org. Lett. 2005, 7, 5111−5114.
(f) Perez, L. J.; Shimp, H. L.; Micalizio, G. C. J. Org. Chem. 2009, 74,
7211−7219. For Zr-mediated cross-coupling of two different internal,
yet symmetrical, alkynes, see: (g) Xi, Z.; Hara, R.; Takahashi, T. J. Org.
Chem. 1995, 60, 4444−4448.
(19) For addition reactions of preformed benzyne−ZrCp2 complexes
to simple alkenes, see: (a) Aoki, K.; Peat, A. J.; Buchwald, S. L. J. Am.
Chem. Soc. 1998, 120, 3068−3073. For addition reactions of
preformed Ti−vinylsilane complexes to alkynes, see: (b) Mizojiri,
R.; Urabe, H.; Sato, F. J. Org. Chem. 2000, 65, 6217−6222. For a
review of Co-catalyzed coupling chemistry of alkynes with alkenes, see:
Jeganmohan, M.; Cheng, C.-H. Chem.Eur. J. 2008, 14, 10876−
10886.
(20) For a review of directed reactions in organic chemistry, see:
(a) Hoveyda, A. H.; Evans, D. A.; Fu, G. C. Chem. Rev. 1993, 93,
1307−1370. For a recent example of an alkene-directed metallacycle-
mediated coupling, see: (b) Moslin, R. M.; Miller, K. M.; Jamison, T. F.
Tetrahedron 2006, 62, 7598−7610. For examples of heteroatom-
directed intermolecular Pauson−Khand reactions, see: (c) Gibson, S. E.;
Mainolfi, N. Angew. Chem., Int. Ed. 2005, 44, 3022−3037.
(21) (a) Harada, K.; Urabe, H.; Sato, F. Tetrahedron Lett. 1995, 36,
3203−3206. (b) Bahadoor, A. B.; Flyer, A.; Micalizio, G. C. J. Am.
Chem. Soc. 2005, 127, 3694−3695. (c) Bahadoor, A. B.; Micalizio, G. C.
Org. Lett. 2006, 8, 1181−1184.
(22) For an example of how alkene substitution plays a significant
role in dictating siteselectivity in the reductive cross-coupling of 1,5-
dienes with alkynes,see: Diez, P. S.; Micalizio, G. C. J. Am. Chem. Soc.
2010, 132, 9576−9578.
(23) (a) See the Supporting Information for the opposite
antipodea demonstration that supports the enantiospecific nature
of the process. (b) Overall, isolated yields of the annulation products
typically range from 52−83%. These less than quantitative yields
reflect: (i) incomplete consumption of the enyne, (ii) alkyne−alkyne
coupling without annulation (the product of this is a functionalized
triene, i.e., Figure 6B, eqs 2 and 3), and occasionally (iii) partial
deprotection of the PMB/silyl ether of the annulation product.
(24) Metallacycle-mediated [2 + 2 + 2] annulation is generally thought
to proceed through initial formation of a metallacyclopentadiene
(9) For example, see: refs 1d, 5b, 5c, 6b, 6d, 7b, 7c, 8d, 8e, and
(a) Corey, E. J.; Huang, A. X. J. Am. Chem. Soc. 1999, 121, 710−714.
(b) Rychnovsky, S. D.; Mickus, D. E. J. Org. Chem. 1992, 57, 2732−
2736. (c) Clasby, M. C.; Craig, D.; Jaxa-Chamiec, A. A.; Lai, J. Y. Q.;
Marsh, A.; Slawin, A. M. Z.; White, A. J. P.; Williams, D. J. Tetrahedron
1996, 52, 4769−4802. (d) Nemoto, H.; Matsuhashi, N.; Imaizumi, M.;
Nagai, M.; Fukumoto, K. J. Org. Chem. 1990, 55, 5625−5631.
(10) (a) Sternberg, E. D.; Vollhardt, K. P. C J. Am. Chem. Soc. 1980,
102, 4839−4841. (b) Sternberg, E. D.; Vollhardt, K. P. C J. Org. Chem.
1982, 47, 3447−3450. (c) Clinet, J.-C.; Dunach, E.; Vollhardt, K. P. C
̃
J. Am. Chem. Soc. 1983, 105, 6710−6712. (d) Butenschon, H.;
̈
Winkler, M.; Vollhardt, K. P. C J. Chem. Soc., Chem. Commun. 1986,
388−390. (e) Grotjahn, D. B.; Vollhardt, K. P. C J. Am. Chem. Soc.
1986, 108, 2091−2093. (f) Germanas, J.; Aubert, C.; Vollhardt, K. P. C
J. Am. Chem. Soc. 1991, 113, 4006−4008. (g) Johnson, E. P.;
Vollhardt, K. P. C J. Am. Chem. Soc. 1991, 113, 381−382. For reviews,
see: (h) Vollhardt, K. P. C Acc. Chem. Res. 1977, 10, 1−8. (i) Aubert,
C.; Buisine, O.; Petit, M.; Slowinski, F.; Malacria, M. Pure Appl. Chem.
1999, 71, 1463−1470. (j) Rubin, M.; Sromek, A. W.; Gevorgyan, V.
Synlett 2003, 2265−2291. (k) Gandon, V.; Aubert, C.; Malacria, M.
Chem. Commun. 2006, 2209−2217. For examples of catalytic
asymmetric intramolecular [2 + 2 + 2] annulation, see: (l) Shibata,
T.; Kurokawa, H.; Kanda, K. J. Org. Chem. 2007, 72, 6521−6525.
(m) Shibata, T.; Tahara, Y. J. Am. Chem. Soc. 2006, 128, 11766−
11767. For recent reviews of [2 + 2 + 2] annulations, see: (n) Shibata,
T.; Tsuchikama, K. Org. Biomol. Chem. 2008, 6, 1317−1323.
(o) Tanaka, K. Chem. Asian J. 2009, 4, 508−518. (p) Leboeuf, D.;
Gandon, V.; Malacria, M. Transition Metal-Mediated [2 + 2 + 2]
Cycloaddtions. In Handbook of Cyclization Reactions; Ma, S., Ed.;
Wiley−VCH: Weinheim, 2009; Vol. 1, p 367.
2773
dx.doi.org/10.1021/ja2105043 | J. Am. Chem.Soc. 2012, 134, 2766−2774

Journal of the American Chemical Society
Article
followed by either alkene insertion/reductive elimination or [4 + 2]/
cycloreversion. We have opted to invoke a model based on [4 + 2]/
cycloreversion, where stereochemistry is controlled by a coordination
event between the homoallylic alkoxide and an alkoxide bound to
Ti. This proposition does not preclude the involvement of an alkene
insertion mechanism, and only intends to offer an empirical model to
predict the stereochemical course of annulation and that is consistent
with the observations summarized in Figure 8A. Future studies are
aimed at deconvoluting these mechanistically distinct pathways.
(25) Crich, D.; Quintero, L. Chem. Rev. 1989, 89, 1413−1432.
(26) Fleming, I.; Dunogues, J.; Smithers, R. Org. React. 1989, 37, 57−
193.
2774
dx.doi.org/10.1021/ja2105043 | J. Am. Chem.Soc. 2012, 134, 2766−2774