Total synthesis of gracilioether F: Development and application of lewis acid promoted ketene-alkene [2+2] cycloadditions and late-stage C-H oxidation

Article abstract of DOI:10.1002/anie.201408055

The first synthesis of gracilioether F, a polyketide natural product with an unusual tricyclic core and five contiguous stereocenters, is described. Key steps of the synthesis include a Lewis acid promoted ketene-alkene [2+ 2] cycloaddition and a late-stage carboxylic acid directed C(sp³)-H oxidation. The synthesis requires only eight steps from norbornadiene.

Full text of DOI:10.1002/anie.201408055

Angewandte
Chemie
DOI: 10.1002/anie.201408055
Total Synthesis
Total Synthesis of Gracilioether F: Development and Application of
Lewis Acid Promoted Ketene–Alkene [2+2] Cycloadditions and Late-
À
Stage C H Oxidation**
Christopher M. Rasik and M. Kevin Brown*
Abstract: The first synthesis of gracilioether F, a polyketide
natural product with an unusual tricyclic core and five
contiguous stereocenters, is described. Key steps of the syn-
two key features of our synthesis have emerged. The first is
the development of a new method for Lewis acid promoted
[2+2] cycloadditions between alkenes and monosubstituted
thesis include
a
Lewis acid promoted ketene–alkene
ketenes, and the second is the strategic use of a late-stage
3
À
[2+2] cycloaddition and a late-stage carboxylic acid directed
carboxylic acid directed C(sp ) H oxidation. Finally, the
3
À
C(sp ) H oxidation. The synthesis requires only eight steps
synthesis of gracilioether F (1) was pursued because we
envisioned that it could be a potential intermediate to access
other, more functionalized members of the natural product
family (e.g., 2–4).
from norbornadiene.
T
he gracilioethers (e.g., 1–4, Figure 1) are a class of recently
isolated polyketides from several families of marine sponge.[1]
Common to this class of molecules is the presence of an
Our retrosynthetic analysis for gracilioether F (1) is
illustrated in Scheme 1. Two points were central to our
design: 1) Cyclobutanone 6 as a key intermediate. Our
laboratory has taken an interest in the development of
Figure 1. Several members of the gracilioether family.
unusual tricyclic core and at least five contiguous stereogenic
centers (two of which are quaternary). Several members of
the family also exhibit considerable biological activity.[1c,d]
Our interest in these molecules stems primarily from the
perspective that the structural uniqueness of the tricyclic core
and the dense array of stereogenic centers would provide
a unique challenge to state-of-the-art chemical synthesis. No
studies towards the synthesis of any member of the tricyclic
gracilioether family have been reported to date.[2]
Scheme 1. Retrosynthetic analysis of gracilioether F.
Lewis acid promoted ketene–alkene [2+2] cycloadditions.[3]
These promoters permit efficient reactions between compo-
nents that do not undergo traditional thermally induced
cycloadditions.[3,4] We envisioned that preparation of cyclo-
butanone 6 from alkenyl ethyl ketene 7 and alkene 8 would be
an ideal testing ground for our newly developed method.
Herein, we disclose an eight-step synthesis of gracilioe-
ther F (1) from norbornadiene. In the context of these studies,
À
À
2) Strategic disconnection of the C O bond by C H oxida-
tion. Implementation of this non-traditional strategy has
allowed for great simplification of the target structure by
ultimately exploiting latent symmetry.[5] It should be noted
[*] C. M. Rasik, Prof. M. K. Brown
Department of Chemistry, Indiana University
800 E. Kirkwood Ave, Bloomington, IN 47401 (USA)
E-mail: brownmkb@indiana.edu
À
that strategic late-stage C H oxidations have rarely been
[**] We thank Indiana University for financial support. We also thank
Prof. M. Christina White (UIUC) for helpful discussions. Data
acquisition for the X-ray structure of 5 was carried out at the
Advanced Photon Source (supported by the U.S. Department of
Energy, Office of Science, Office of Basic Energy Sciences, DE-AC02-
06CH11357) in collaboration with ChemMatCARS Sector 15, which
is principally supported by the National Science Foundation/
Department of Energy (NSF/CHE-1346572). We acknowledge Dr.
Yu-Sheng Chen for data collection.
employed in total synthesis settings.[6,7]
As mentioned above, our laboratory has recently devel-
oped the first method for Lewis acid promoted ketene–alkene
[2+2] cycloadditions.[3] The method offers several advantages
over more traditional thermally induced variants, such as
increased reactivity and selectivity, which has consequently
allowed for the synthesis of highly congested cyclobutanones.
The [2+2] cycloaddition needed for the synthesis of graci-
lioether F requires the use of an alkenyl ketene. This
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201408055.
Angew. Chem. Int. Ed. 2014, 53, 1 – 6
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1
These are not the final page numbers!

.
Angewandte
Communications
substrate, however, was not investigated in our original
study,[3] and therefore, studies were initiated on model
systems to probe for reactivity.
One of the key challenges with reactions of alkenyl
ketenes is that they tend to be more reactive than their alkyl
or aryl counterparts.[8] As such, alkenyl ketenes are always
generated in situ in the presence of the reacting partner (e.g.,
alkene or imine).[8] Furthermore, owing to the propensity of
alkenyl ketenes to undergo dimerization through
methyl group, which presumably reduces the rates of ketene
dimerization/polymerization. Upon confirmation of the ste-
reochemical identities of cyclobutanones 11b and 11c, how-
ever, it was discovered that a diastereomer of the one
necessary for the synthesis of gracilioether F (1) had been
generated. This diastereomer is likely formed by approach of
the alkene syn to the alkenyl group, which is consistent with
related cycloadditions involving aryl substituents.[3]
A revised route to key cyclobutanone intermediate 6 was
then devised (Scheme 3). It was reasoned that if cyclobuta-
none 13 could be prepared from a [2+2] cycloaddition
a
[4+2] cycloaddition, [2+2] cycloadditions with alkenes
under thermally induced conditions are rare and typically
require forcing conditions and a large excess of the alkene.[8]
For Lewis acid promoted ketene–alkene [2+2] cycloaddi-
tions, the original method required that the ketene be
generated prior to use (isolation of the ketene not required).[3]
Therefore, we initiated our studies by searching for a suitable
set of conditions and a ketene precursor (i.e., unsaturated acid
chloride) that would allow for clean generation of an alkenyl
ketene.
As illustrated in Scheme 2, attempts to generate an
alkenyl ketene from a,b-unsaturated acid chloride 9a and
Et3N did not lead to formation of the desired alkenyl ketene
Scheme 3. Revised strategy for the Lewis acid promoted [2+2] cyclo-
addition. KHMDS=potassium hexamethyldisilazide.
Scheme 2. Original strategy for the [2+2] cycloaddition.
employing a monosubstituted alkenyl ketene (e.g., 12),
subsequent diastereoselective a-alkylation with EtI would
lead to formation of the desired diastereomer. Pre-generation
of the requisite monosubstituted alkenyl ketene (like most
monosubstituted ketenes), however, was not possible.[10] Thus,
it was necessary to establish a new method for Lewis acid
promoted ketene–alkene [2+2] cycloadditions with mono-
substituted ketenes.
After considerable experimentation we found that
through slow addition of acid chloride 14 to a solution of
AlMe3, iPr2NEt, and cyclopentene, cyclobutanone 16 could be
generated in 69% yield (8:1 d.r., inconsequential to the total
synthesis; Scheme 3). These conditions function well because
the ketene, when generated in the presence of the Lewis
acid,[11] is consumed through cycloaddition at a rate that is
similar to its formation; therefore, side reactions resulting
from polymerization of the ketene are minimized. Similar to
previous studies outlined in Scheme 2, the presence of
g-Ph,Me-substituted acid chloride 14 was necessary for
rapid generation of the corresponding ketene at low temper-
atures. To complete the model studies, cyclobutanone 16 was
deprotonated with KHMDS and the resulting enolate trapped
owing to a slow dehydrohalogenation process. The slow rate
of ketene generation from acid chloride 9a is likely due to the
poorly acidic g-protons. To increase the acidity of the
g-protons and hence the rate of ketene generation a phenyl
group was installed at the g-position. It should be noted that
additional substitution at the g-position is inconsequential to
the total synthesis, as the alkene will ultimately be cleaved
under oxidative conditions (see below). For ease of synthesis,
the b,g-unsaturated acid chlorides 9b–c were prepared (as
opposed to the analogous a,b-unsaturated acid chlorides).[9]
Treatment of acid chloride 9b with Et3N led to rapid
formation of the ketene at 228C. Lewis acid promoted
ketene–alkene cycloaddition with AlMe3 and a model alkene,
namely cyclopentene (2.0 equiv), was then carried out and led
to the formation of cyclobutanone 11b in 42% yield and 7:1
d.r. Further studies revealed that acid chloride 9c, which
incorporates an additional methyl substituent, led to very
clean generation of the alkenyl ketene and consequently to
a higher yield of cycloadduct 11c (61%). The increase in yield
is likely due to the increase in steric bulk imposed by the
2
www.angewandte.org
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 1 – 6
These are not the final page numbers!

Angewandte
Chemie
with EtI to provide cyclobutanone 17 in 91% yield and > 20:1
d.r. (Scheme 3).
With a method for the synthesis of the desired diastereo-
mer of the key cyclobutanone in a model system, our
attention then focused on the total synthesis of gracilioether F
(1; Scheme 4). The synthesis of the requisite diethylcyclopen-
> 20:1 d.r. (Scheme 4). The high selectivity of the alkylation
reaction is likely the result of the EtI approaching on the
convex face of the enolate. Regioselective Baeyer–Villiger
oxidation provided lactone 21 in 61% yield over two steps.
One-pot ozonolysis and quenching under Pinnick oxidation
conditions led to the formation of carboxylic acid 5 in 83%
yield. The entire sequence to carboxylic acid 5 required only
seven steps from norbornadiene and, with the exception of
the conversion of 21 into 5, was carried out on gram scale.
To complete the synthesis of gracilioether F, a selective
À
À
C H oxidation was necessary. Analysis of the desired C H
oxidation reveals two key points: 1) Based on electronic
À
considerations, oxidation is likely to occur at the desired C H
3
À
bond because it is the tertiary C(sp ) H bond that is most
distant from the electron-withdrawing groups.[6] 2) The
À
desired C H bond is located on the concave face of the
bicyclic scaffold. Thus its accessibility to external oxidants is
likely to be hindered.[6]
Considering primarily the former point, we investigated
common methods that typically oxidize the most electron-rich
3
[6]
À
C(sp ) H bond, barring consideration of any steric effects.
However, treatment of carboxylic acid 5 with RuO4 (gen-
erated form RuCl3 and KBrO3)[13] or TFDO[14] led to either
mostly consumption or complete recovery of the carboxylic
acid 5, respectively, with < 2% product formation (Table 1,
3
À
Table 1: Late-stage C(sp ) H oxidation.
Entry Conditions[a]
Yield[b]
[%]
RSM[b]
[%]
[c]
1
RuCl3 (10 mol%), KBrO3, pyridine, MeCN/ <2
–
H2O
2
3
4
5
TFDO, CH2Cl2, À458C to 228C
<2
9
<2
<2
15 (15)
10
>90
48
ca. 80
ca. 80
51
[Fe(PDP)] (20 mol%), H2O2, MeCN, 228C
Fe(OAc)2 (1 equiv), H2O2, MeCN, 228C
Mn(OAc)2 (1 equiv), H2O2, MeCN, 228C
Cu(OAc)2 (1 equiv), H2O2, MeCN, 228C
Cu(OAc)2 (1 equiv), H2O2, MeCN, 08C
Cu(OAc)2 (1 equiv), H2O2, MeCN, 228C
6
7
88
>90
8[d]
<2
Scheme 4. Synthesis of gracilioether F. Ts=para-toluenesulfonyl.
[a] See the Supporting Information for experimental details. [b] Deter-
mined by NMR analysis using 1,3,5-trimethoxybenzene as the internal
standard. Yield of isolated product given in parentheses. [c] Conversion
of starting material >80%. [d] The methyl ester of 5 was used as the
starting material. RSM=recovered starting material, TFDO=trifluoro-
methyldioxarane.
tene 8 was accomplished in two steps from known diol 18[12] by
tosylation and subsequent displacement with Me2CuLi·LiI.
Lewis acid promoted cycloaddition between cyclopentene 8
(2 equiv) and acid chloride 14 led to cyclobutanone 19 in 65%
yield and > 20:1 d.r. (5:1 d.r. at the a-carbon atom,
inconsequential to the synthesis) over two steps from the
carboxylic acid precursor to 14. The diastereoselectivity is
easily rationalized by inspection of the pre-transition state
assembly I.
entries 1 and 2). The lack of selectivity observed in the
reaction with RuO4 is likely due to the vastly different steric
À
environments of the C H bond that should be oxidized and
À
the C H bonds whose oxidation is undesired. Thus efforts
were directed towards utilizing the neighboring carboxylic
À
acid substituent to perhaps direct C H oxidation to occur at
Cycloadduct 19 was converted into carboxylic acid 5
through a straightforward sequence, starting with a-alkylation
of cyclobutanone 19 with KHMDS and EtI to provide 20 in
the desired site.[15,16] To the best of our knowledge, only one
À
report details the carboxylic acid directed C H oxidation of
3
À
tertiary C(sp ) H bonds. The White group has demonstrated
Angew. Chem. Int. Ed. 2014, 53, 1 – 6
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3
These are not the final page numbers!

.
Angewandte
Communications
that [Fe(PDP)] (22) is effective for carboxylic acid directed
[17]
[1] a) R. Ueoka, Y. Nakao, S. Kawatsu, J. Yaegashi, Y. Matsumoto,
S. Matsunaga, K. Furihata, R. W. M. van Soest, N. Fusetani, J.
Org. Chem. 2009, 74, 4203; b) C. Festa, S. De Marino, M. V.
DꢀAuria, E. Deharo, G. Gonzalez, C. Deyssard, S. Petek, G.
Bifulco, A. Zampella, Tetrahedron 2012, 68, 10157; c) C. Festa,
G. Lauro, S. De Marino, M. V. DꢀAuria, M. C. Monti, A.
Casapullo, C. DꢀAmore, B. Renga, A. Mencarelli, S. Petek, G.
Bifulco, S. Fiorucci, A. Zampella, J. Med. Chem. 2012, 55, 8303;
d) C. Festa, C. DꢀAmore, B. Renga, G. Lauro, S. Marino, M.
DꢀAuria, G. Bifulco, A. Zampella, S, Fiorucci, S. Marine Drugs
2013, 11, 2314.
À
C H oxidation. When carboxylic acid 5 was treated with
[Fe(PDP)] ((S,S) or racemate) under the reported conditions,
9% of gracilioether F (1) was generated with 48% recovered
starting material (entry 3). The loss in mass balance may be
due to undesired decarboxylation. The formation of unde-
sired oxidation products was limited to < 5%.
In the context of our efforts to identify alternative
À
conditions that might improve the yield of the C H oxidation,
we reasoned that perhaps if metals without ancillary ligands
were used, the desired C-H-oxidation might be favored owing
to fewer adverse steric interactions. Therefore, we evaluated
[2] For studies towards the synthesis of a different structural
subclass of the gracilioethers, see: M. D. Norris, M. V. Perkins,
Tetrahedron 2013, 69, 9813.
several first-row transition metals (e.g., FeII, MnII, CuII) for
[18,19]
À
their ability to promote C H oxidation (entries 4–6).
[3] a) C. M. Rasik, M. K. Brown, J. Am. Chem. Soc. 2013, 135, 1673;
b) C. M. Rasik, M. K. Brown, Synlett 2014, 760; for studies
regarding the mechanism of Lewis acid promoted ketene –
alkene [2+2] cycloadditions, see: c) Y. Wang, D. Wei, Z. Li, Y.
Zhu, M. Tang, J. Phys. Chem. A 2014, 118, 4288; d) C. M. Rasik,
Y. J. Hong, D. J. Tantillo, M. K. Brown, Org. Lett. 2014, 16, 5168.
[4] For excellent reviews, see: a) B. B. Snider, Chem. Rev. 1988, 88,
793; b) J. A. Hyatt, P. W. Raynolds, Org. React., Wiley, New
York, NY, 1994; c) T. T. Tidwell, Ketenes, Wiley, New York, NY,
1995; d) T. T. Tidwell, Ketenes II, Wiley Interscience, Hoboken,
NJ, 2006; e) T. T. Tidwell, Eur. J. Org. Chem. 2006, 563; f) R. L.
Danheiser, Science of Synthesis: Compounds with Four and
Three Carbon Heteroatom Bonds, Vol. 23, Georg Thieme,
Stuttgart, 2006.
Among these metal salts, the use of Cu(OAc)2 in combination
with H2O2 led to the formation of the desired product in 15%
yield with 51% recovered starting material (entry 6, 84 mg
scale). The iron- and copper-promoted hydroxylations are
significant for several reasons: 1) Their strategic implemen-
tation has enabled the synthesis of gracilioether F through
unconventional means in only eight steps. 2) A large amount
of starting material can be recovered thus counterbalancing
the low yield of isolated product (this is a common scenario
for many C H oxidation reactions).[6] Furthermore, when the
À
copper-promoted oxidation was carried out at 08C, 10% of
gracilioether F was formed with 88% recovered starting
[5] For selected recent syntheses that exploit latent symmetry, see:
a) M. Ball, M. J. Gaunt, D. F. Hook, A. S. Jessiman, S. Kawahara,
P. Orsini, A. Scolaro, A. C. Talbot, H. R. Tanner, S. Yamanoi,
S. V. Ley, Angew. Chem. Int. Ed. 2005, 44, 5433; Angew. Chem.
2005, 117, 5569; b) J. T. Malinowski, R. Sharpe, J. S. Johnson,
Science 2013, 340, 180.
[6] For reviews, see: a) K. Godula, Science 2006, 312, 67; b) W. R.
Gutekunst, P. S. Baran, Chem. Soc. Rev. 2011, 40, 1976; c) T.
Newhouse, P. S. Baran, Angew. Chem. Int. Ed. 2011, 50, 3362;
Angew. Chem. 2011, 123, 3422.
material (entry 7). 3) To the best of our knowledge, the use of
3
À
copper salts and H2O2 to promote selective C(sp ) H
oxidation reactions is unknown.[18–20]
À
The mechanism of the copper-promoted C H oxidation is
not clear at this time, but is likely related to Fenton-type
reactions.[18,19] Several additional key points are noteworthy:
À
1) The carboxylic acid moiety is necessary for C H oxidation.
When the corresponding methyl ester of 5 was used, < 2% of
1 was observed with > 90% recovery of starting material
(entry 8). 2) The use of catalytic quantities of copper led to
inferior yields. 3) The portion-wise addition of Cu(OAc)2 and
H2O2 was necessary to obtain optimal yields. 4) The formation
of undesired oxidation products was observed only in minimal
quantities (< 5%). Similar to the iron-catalyzed reactions, the
loss in mass balance may be due to decarboxylation and
subsequent decomposition.[21]
À
[7] For recent examples of strategic C H oxidations in total
synthesis, see: a) D. A. Siler, J. D. Mighion, E. J. Sorensen,
Angew. Chem. Int. Ed. 2014, 53, 5332; Angew. Chem. 2014, 126,
5436; b) B. R. Rosen, L. R. Simke, P. S. Thuy-Boun, D. D. Dixon,
J.-Q. Yu, P. S. Baran, Angew. Chem. Int. Ed. 2013, 52, 7317;
Angew. Chem. 2013, 125, 7458.
[8] a) R. L. Danheiser, C. Martinez-Davila, H. Sard, Tetrahedron
1981, 37, 3943; b) R. Huston, M. Rey, A. S. Dreiding, Helv.
Chim. Acta 1982, 65, 1563.
[9] a) A. H. Mermerian, G. C. Fu, J. Am. Chem. Soc. 2005, 127,
5604; b) T. Kapferer, R. Brꢁckner, Eur. J. Org. Chem. 2006, 2119.
[10] M. Rey, S. Roberts, A. Dieffenbacher, A. S. Dreiding, Helv.
Chim. Acta 1970, 53, 417.
[11] For examples of ketene generation in the presence of a Lewis
acid, see: a) T. P. Yoon, V. M. Dong, D. W. C. MacMillan, J. Am.
Chem. Soc. 1999, 121, 9726; b) S. G. Nelson, Z. Wan, Org. Lett.
2000, 2, 1883.
[12] C. A. Grob, H. R. Pfaendler, Helv. Chim. Acta 1970, 53, 2156.
[13] E. McNeill, J. Du Bois, J. Am. Chem. Soc. 2010, 132, 10202.
[14] R. Curci, L. DꢀAccolti, C. Fusco, Acc. Chem. Res. 2006, 39, 1.
In summary, we have prepared gracilioether F (1) with
a longest linear sequence of eight steps from norbornadiene
without the use of protecting groups. The synthesis was
enabled through the development of a new variant of a Lewis
acid promoted ketene–alkene [2+2] cycloaddition and the
3
À
development of a carboxylic acid directed C(sp ) H oxida-
tion. This synthesis also highlights the state of the art in
C(sp ) H oxidation as a powerful transformation and under-
scores the need for the development of new methods.
3
À
Received: August 7, 2014
Published online: && &&, &&&&
À
[15] For a review on carboxylic acid directed C H functionalizations,
see: G. Shi, Y. Zhang, Adv. Synth. Catal. 2014, 356, 1419.
[16] For a review on remote functionalization reactions involving
ˇ
alkoxy radicals (e.g., Barton and hypohalite reactions), see: Z.
ˇ
´
Cekovic, Tetrahedron 2003, 59, 8073.
À
Keywords: C H oxidation · cycloaddition · gracilioether F ·
ketenes · total synthesis
[17] a) M. S. Chen, M. C. White, Science 2007, 318, 783; b) M. A.
Bigi, S. A. Reed, M. C. White, J. Am. Chem. Soc. 2012, 134, 9721.
.
4
www.angewandte.org
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 1 – 6
These are not the final page numbers!

Angewandte
Chemie
[18] a) A. E. Shilov, G. B. Shulꢀpin, Chem. Rev. 1997, 97, 2879; b) T.
Punniyamurthy, L. Rout, Coord. Chem. Rev. 2008, 252, 134.
[19] a) T. Kurata, Y. Watanabe, M. Katoh, Y. Sawaki, J. Am. Chem.
Soc. 1988, 110, 7472; b) A. Sobkowiak, A. Qui, X. Liu, A. Llobet,
D. T. Sawyer, J. Am. Chem. Soc. 1993, 115, 609.
[20] Copper-catalyzed allylic and benzylic oxidations are well known;
for a review, see: M. B. Andrus, J. C. Lashley, Tetrahedron 2002,
58, 845.
[21] The corresponding reaction with 4-methylpentanoic acid pro-
vided < 5% of the oxidation products. Further investigations are
ongoing.
Angew. Chem. Int. Ed. 2014, 53, 1 – 6
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5
These are not the final page numbers!

.
Angewandte
Communications
Communications
Total Synthesis
C. M. Rasik,
M. K. Brown*
&&&& — &&&&
Total Synthesis of Gracilioether F:
Development and Application of Lewis
Acid Promoted Ketene–Alkene
À
[2+2] Cycloadditions and Late-Stage C H
Oxidation
Controlled oxidation: Gracilioether F can
be synthesized from norbornadiene in
only eight steps while avoiding the use of
protecting groups. Key steps of the syn-
thesis include a Lewis acid promoted
[2+2] cycloaddition of a ketene with an
olefin and a late-stage carboxylic acid
3
À
directed C(sp ) H oxidation.
6
www.angewandte.org
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 1 – 6
These are not the final page numbers!