Enantioselective syntheses of tremulenediol A and tremulenolide A

Article abstract of DOI:10.1021/ol051945u

(Chemical Equation Presented) An enantioselective entry to the skeleton of the tremulane sesquiterpenes is described. The approach features a series of efficient transition metal-catalyzed reactions commencing with an enantioselective rhodium(II)-catalyze

Full text of DOI:10.1021/ol051945u

ORGANIC
LETTERS
2005
Vol. 7, No. 20
4535-4537
Enantioselective Syntheses of
Tremulenediol A and Tremulenolide A
Brandon L. Ashfeld and Stephen F. Martin*
Department of Chemistry and Biochemistry, The UniVersity of Texas at Austin,
Austin, Texas 78712
sfmartin@mail.utexas.edu
Received August 11, 2005
ABSTRACT
An enantioselective entry to the skeleton of the tremulane sesquiterpenes is described. The approach features a series of efficient transition
metal-catalyzed reactions commencing with an enantioselective rhodium(II)-catalyzed intramolecular cyclopropanation followed by a regioselective
allylic alkylation and a diastereoselective rhodium(I)-catalyzed [5
syntheses of tremulenediol A and tremulenolide A.
+
2] cycloaddition. This strategy was applied to the first enantioselective
Tremulenolide A (1) and tremulenediol A (2), two repre-
sentative members of the novel tremulane class of sesqui-
terpenes, were isolated in 1993 from the fungal pathogen
Phellinus tremulae as part of a project to develop methods
for controlling fungal decay and staining in trembling or
quaking aspen (Populus tremuloides).1 Although the com-
mercial advantages associated with the potential biological
activity that these two natural products may possess is
intriguing in itself, the orientation and stereochemistry of
the substituents on the hydroazulene skeleton was the primary
motivation for our interest in 1 and 2 as synthetic targets.
Although Davies reported a nice approach to racemic 1 and
2 in 1998,2 there has been no report of an enantioselective
synthesis of these interesting targets. We now report the
enantioselective syntheses of 1 and 2 exploiting several
transition metal-catalyzed reactions that have been of interest
to us and others for a number of years.
2 was inspired by the pioneering work of Wender.3 We
envisioned that 3 would be accessible via a transition metal-
catalyzed allylic alkylation of the vinyl cyclopropyl lactone
Scheme 1. Retrosynthesis of 1 and 2
The key features of our approach to tremulenolide A (1),
which would be produced upon selective allylic oxidation
of tremulenediol A (2), are outlined in retrosynthetic format
in Scheme 1. The diastereoselective, Rh(I)-catalyzed [5 +
2] cycloaddition of the cyclopropyl 1,6-enyne 3 leading to
10.1021/ol051945u CCC: $30.25
© 2005 American Chemical Society
Published on Web 08/30/2005

4 with the substituted malonate 5 prepared from 2-butyne-
1,4-diol (7). The lactone 4 would then be prepared by the
enantioselective, intramolecular rhodium(II)-catalyzed cyclo-
propanation of diazoester 6, a reaction developed by our
group over the past decade.4 We were optimistic that this
synthetic strategy could be amended to incorporate the
domino [Rh(CO)2Cl]2-catalyzed allylic alkylation/[5 + 2]
cycloaddition reaction sequence that we recently reported,5
thus enabling a very rapid entry to the tremulane carbon
skeleton.
in 99% combined yield and 94% ee for each diastereomer
as determined by chiral HPLC analysis.
The synthesis of malonate 5 began with the monobenzyl-
ation of commercially available 2-butyne-1,4-diol (7) to give
11 (Scheme 3). In accord with literature reports,9 we found
Scheme 3. Synthesis of Malonate 5
The first phase of the synthesis entailed the enantioselec-
tive construction of cyclopropyl lactone 4 in a straightforward
four-step sequence from commercially available 2-methyl-
2-vinyl oxirane 8 (Scheme 2). Thus, treating oxirane 8 with
Scheme 2. Synthesis of Cyclopropyl Lactone 4
that 7 could be converted into 11 in 51% yield upon treatment
with benzyl bromide and Ag2O. However, given the modest
yield and cost of a silver-mediated reaction, we developed
the less expensive alternative of reacting 7 with benzyl
bromide, NaH, and tetrabutylammonium iodide (TBAI) in
DMF to give 11 in 53% yield. Subsequent mesylation of 11
gave 12 (92%), which was then allowed to react with
sodiodimethyl malonate to furnish 5 in 97% yield. Using
this optimized three-step sequence, multigram quantities of
5 were readily obtained.
At this juncture, the stage was set for the preparation of
13 via a transition metal-catalyzed allylic alkylation of the
vinyl cyclopropyl lactone 4 with a salt of malonate 5.
Inasmuch as [Rh(CO)2Cl]2 was known to catalyze the
projected [5 + 2] cycloaddition that would form the
hydroazulene ring, we queried whether it might also catalyze
the desired allylic alkylation, thus potentially enabling a
tandem allylic alkylation and [5 + 2] cycloaddition. We were
of course cognizant of the work of Evans,10 who had shown
that a modified Wilkinson’s catalyst promoted allylic alkyl-
ations to give preferentially products in which substitution
occurred at the more encumbered terminus of the allylic
moiety. Not dissuaded by this unfavorable literature prece-
dent, we discovered that when 4 was treated with the sodium
enolate of 5 in the presence of [Rh(CO)2Cl]2, 13 was obtained
with complete regiocontrol as a mixture (1:1) of E/Z isomers
(Scheme 4), albeit in only a 20% yield that defied our
attempts at optimization. It may be noted that this result led
us to developing [Rh(CO)2Cl]2 as a novel catalyst having
unique properties for allylic alkylations.11
the sulfur ylide generated from trimethylsulfonium iodide
provided divinyl carbinol 9 in 84% yield.6 Because the
tertiary alcohol group in 9 was resistant to acylation, it was
not possible to prepare the diazoester 6 directly from 9 using
the Corey-Myers diazoesterification procedure,7 and there-
fore, it became necessary to resort to an alternate protocol
we had previously developed. In the event, acylation of 9
with diketene in the presence of a catalytic amount of
4-(dimethylamino)pyridine (DMAP) gave the â-ketoester 10
in 93% yield.8 Subsequent conversion of 10 to the corre-
sponding 1,3-dicarbonyl diazo compound followed by hy-
droxide-induced cleavage of the methyl ketone moiety
provided diazoester 6 in 97% overall yield. Cyclization of 6
via intramolecular cyclopropanation in the presence of
Rh2[5(R)-MEPY]4 (0.1 mol %) proceeded smoothly to yield
the cyclopropyl lactone 4 as a mixture (1:1) of C4 epimers
(1) Ayer, W. A.; Cruz, E. R. J. Org. Chem. 1993, 58, 7529.
(2) Davies, H. M. L.; Doan, B. D. J. Org. Chem. 1998, 63, 657.
(3) Wender, P. A.; Dyckman, A. J.; Husfield, C. O.; Kadereit, D.; Love,
J. A.; Rieck, H. J. Am. Chem. Soc. 1999, 121, 10442.
(4) Doyle, M. P.; Austin, R. E.; Bailey, A. S.; Dwyer, M. P.; Dyatkin,
A. B.; Kalinin, A. V.; Kwan, M. M. Y.; Liras, S.; Oalmann, C. J.; Pieters,
R. J.; Protopopova, M. N.; Raab, C. E.; Roos, G. H. P.; Zhou, Q.-L.; Martin,
S. F. J. Am. Chem. Soc. 1995, 117, 5763.
Scheme 4. Rh(I)-Catalyzed Allylic Alkylation of Cyclopropyl
Lactone 4 with Malonate 5
(5) Ashfeld, B. L.; Miller, K. A.; Smith, A. J.; Tran, K.; Martin, S. F.
Org. Lett. 2005, 7, 1661.
(6) Alcaraz, L.; Harnett, J. J.; Mioskowski, C.; Martel, J. P.; Le Gall,
T.; Shin, D.; Falck, J. R. Tetrahedron Lett. 1994, 35, 5449.
(7) Corey, E. J.; Myers, A. G. Tetrahedron Lett. 1984, 25, 3559.
(8) All new compounds were purified (>95%) by distillation, recrys-
tallization, or preparative HPLC and were characterized by 1H and 13C NMR,
IR, and HRMS. Complete experimental details and characterization of new
compounds will be published in a full account of this work.
4536
Org. Lett., Vol. 7, No. 20, 2005

Given the poor efficiency with which [Rh(CO)2Cl]2
catalyzed the allylic alkylation of 4 with 5, we turned to the
use of more conventional Pd(0) catalysts and found that the
reaction of 4 with the sodium enolate of 5 in the presence
of 10 mol % of Pd(PPh3)4, and additional PPh3 gave enyne
13 in 71% yield (Scheme 5). Although we were unable to
Scheme 6. Completion of the Total Syntheses of 2 and 1
Scheme 5. Entry into the Tremulane Carbon Skeleton
MeOH)) was comparable to that of natural 2 ([R]24D ) +41.3
(c 0.24 MeOH)). Subsequent treatment of 2 with MnO2
provided lactone 1, which exhibited spectral characteristics
(e.g., 500 MHz 1H and 125 MHz 13C NMR) consistent with
those reported in the literature. The optical rotation of
synthetic 1 ([R]24D ) +99.6 (c 0.14 MeOH)) was comparable
effect the [5 + 2] cycloaddition of the carboxylic acid 13 to
give the requisite hydroazulene, the [Rh(CO)2Cl]2-catalyzed
[5 + 2] cycloaddition of 3, which was prepared in two steps
and 84% overall yield from 13, underwent facile intra-
molecular [5 + 2] cycloaddition upon heating in the presence
of [Rh(CO)2Cl]2 to give 14 as the sole isolated product. The
regiochemistry in this cycloaddition was consistent with the
precedent of Wender.12
to that of the isolated natural product ([R]24 ) +110.7 (c
D
0.14 MeOH)).
In summary, the first enantioselective syntheses of the two
representative tremulane sesquiterpenes tremulenolide A (1)
and tremulenediol A (2) have been achieved. The syn-
thetic route is highlighted by a chiral rhodium(II)-catalyzed
enantioselective cyclopropanation to establish the requisite
absolute stereochemistry of the natural products. A transition
metal-catalyzed allylic alkylation is then utilized to assemble
the complete carbon ensemble and to set up a diastereo-
selective rhodium(I)-catalyzed [5 + 2] intramolecular cyclo-
addition. Use of this trio of transition metal-catalyzed
operations leads to the convergent and efficient total
syntheses of 2 and 1 in 6% (16 steps) and 5.2% (17 steps)
overall yields, respectively. Other applications of these and
other sequential and domino transition metal-catalyzed
reactions are in progress and will be reported in due course.
At this stage, our efforts focused on the task of reducing
the diester moiety in 14 to install the requisite gem-dimethyl
substitution present in tremulanes 1 and 2. Toward this end,
hydride reduction of 14 followed by protection of the
resultant alcohol gave 15 in 81% overall yield (Scheme 6).
Treatment of 15 with LiAlH4 followed by mesylation of the
resultant diol gave the corresponding bismesylate 16 in 72%
overall yield. Initial attempts to reduce the bismesylate
functionality to provide the gem-dimethyl moiety utilizing
LiAlH4 did not proceed to completion, even after extended
reaction times (>24 h). However, reaction of 16 with
LiBHEt3 followed by removal of the TBS protecting group
delivered 17 in 70% yield over the two steps. When 17 was
reduced via catalytic hydrogenation using base-washed Pd/C
(10 mol %) under an atmosphere of H2, the trisubsituted
olefin was reduced stereoselectively with concomitant re-
moval of the benzyl protecting group to provide tremulene-
diol A (2) in 82% yield and as the only isolable product.
The spectral data (e.g., 500 MHz 1H and 125 MHz 13C NMR
and IR) for 2 were consistent with those reported in the
Acknowledgment is made to the National Institute of
General Medical Sciences (GM 31077), the Robert A. Welch
Foundation, Pfizer, Inc., and Merck Research Laboratories
for their generous support of this research.
Supporting Information Available: Experimental pro-
1
cedures for preparing 13 and 14 and copies of H NMR
literature,1 and the optical rotation ([R]24 ) +40.0 (c 0.24
D
spectra for all new compounds and synthetic 1 and 2. This
material is available free of charge via the Internet at
http://pubs.acs.org.
(9) Bouzide, A.; Sauve, G. Tetrahedron Lett. 1997, 38, 5945.
(10) Evans, P. A.; Nelson J. D. J. Am. Chem. Soc. 1998, 120, 5581.
(11) Ashfeld, B. L.; Miller, K. A.; Martin, S. F. Org. Lett. 2004, 6, 1321.
(12) Wender, P. A.; Dyckman, A. J. Org. Lett. 1999, 1, 2089.
OL051945U
Org. Lett., Vol. 7, No. 20, 2005
4537