A concise total synthesis of (R)-puraquinonic acid

Article abstract of DOI:10.1002/anie.201208792

Distant relative: The configuration of the quaternary stereocenter in puraquinonic acid (see scheme, right) is defined by groups far removed from the point of asymmetry. The use of a bicyclic thioglycolate lactam to set the quaternary stereocenter at an e

Full text of DOI:10.1002/anie.201208792

Angewandte
Chemie
DOI: 10.1002/anie.201208792
Natural Product Synthesis
A Concise Total Synthesis of (R)-Puraquinonic Acid**
Erica A. Tiong, Daniel Rivalti, Benjamin M. Williams, and James L. Gleason*
In memory of Robert E. Ireland (1929–2012)
The construction of quaternary carbon stereocenters is among
the most difficult challenges in synthesis, one which is
compounded when two or more groups at the stereocenter
are similar in size and electronics.[1] Puraquinonic acid (1,
Scheme 1), a 15-norilludalane fungal metabolite, which was
Scheme 1. Structure of (R)-puraquinonic acid and the illudalane skel-
eton.
Scheme 2. A bicyclic lactam auxiliary for selective quaternary stereo-
center formation. LDA=lithium diisopropylamide.
isolated from mycelial cultures of Mycena Pura and which
possesses mild differentiation-inducing activity towards HL-
60 cells,[2] is an intriguing example of a molecule containing
a challenging quaternary stereocenter. The majority of
illudalane sesquiterpenoids contain geminal dimethyl groups
at C11, which are either prochiral or diastereotopic.[3] In
contrast, in the structure of 1, one of the methyl groups of the
geminal dimethyl groups has undergone oxidation resulting in
a new quaternary stereocenter. Importantly, the stereodefin-
ing groups in 1, the methyl and hydroxyethyl groups, are far
removed from the stereocenter and offer only a minimum of
electronic differentiation. Thus, while 1 may appear at first
glance to be a simple molecule, its synthesis in enantiopure
form is a significant challenge. Indeed, while an efficient 10-
step synthesis of racemic 1 has been reported,[4] the only
enantioselective synthesis of 1 exceeds 30 steps in length.[5]
We have previously developed a method for the formation
of quaternary stereocenters based on the stereoselective
generation and alkylation of a,a-disubstituted amide enolates
a C2-pseudosymmetric auxiliary and undergoes highly diaste-
reoselective alkylations (6) and Mannich additions (7).[6c,8]
Importantly, the stereoselectivity of enolate generation and
subsequent alkylation is based on the order of alkylation and
not the size of the substituents, such that quaternary
stereocenters bearing three groups of nearly identical size
(e.g. ethyl, propyl, and allyl) can be formed with excellent
stereoselectivity. We felt that this method would be ideal for
the synthesis of 1, because it would allow for an early-stage
introduction of the quaternary stereocenter.
Our retrosynthetic analysis (Scheme 3) suggested that the
quinoid ring in 1 might be generated by oxidation of a Diels–
Alder adduct of dimethyl acetylenedicarboxylate 10 and
diene 9. Diene 9 could be envisaged as coming from a tandem
ring-closing ene–yne/diene–ene cross metathesis of 1,6-enyne
11 and 3-buten-1-ol.[9] Finally, we fully expected that 11 could
be generated stereoselectively through a dialkylation/reduc-
tion/alkylation sequence using thioglycolate lactam 3. Impor-
from easily prepared bicyclic thioglycolate lactam
3
(Scheme 2).[6,7] Sequential double alkylation of bicyclic
lactam 3 followed by a dissolving metal reduction results in
stereoselective formation of an a,a-disubstituted enolate (5),
where the enolate geometry is regulated by the combination
of the conformation of the starting bicycle and the config-
uration of the a-stereocenter. The resulting enolate possesses
[*] E. A. Tiong, D. Rivalti, B. M. Williams, Prof. J. L. Gleason
Department of Chemistry, McGill University
801 Sherbrooke W., Montreal, QC, H3A 0B8 (Canada)
E-mail: jim.gleason@mcgill.ca
[**] This work was supported by the NSERC. E.A.T. thanks NSERC and
FQRNT, and B.M.W. thanks NSERC for postgraduate fellowships.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201208792.
Scheme 3. Retrosynthetic approach to (R)-puraquinonic acid.
Angew. Chem. Int. Ed. 2013, 52, 1 – 5
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1
These are not the final page numbers!

.
Angewandte
Communications
tantly, in this design, a quaternary stereocenter bearing two
groups of similar size and similar electronics (allyl and
propargyl) would be set at an early stage, and the stereo-
chemistry would be transmuted through the subsequent
metathesis and Diels–Alder sequences to eventually produce
the distally differentiated quaternary stereocenter.
Diels–Alder precursor 9 was prepared in six steps from
bicyclic lactam 3 in 63% overall yield. Sequential alkylation
of 3 with allyl bromide followed by methyl iodide provided 4a
in 86% yield as a single stereoisomer (Scheme 4). Reductive
buten-1-ol with no cross-metathesis product observed.[11]
However, subjection of the benzoate ester of 3-buten-1-ol to
the metathesis conditions allowed the formation of cross-
metathesis product 9 in 89% yield as a single E isomer.
Diels–Alder cycloaddition of diene 9 with dimethyl
acetylenedicarboxylate proceeded readily in refluxing tolu-
ene without need for Lewis acid catalysts. The resulting
diastereomeric mixture of 1,4-cyclohexadienes was directly
oxidized with DDQ in benzene to afford arene 15 in 87%
yield. Given the need for an ester-protecting group in the
metathesis step, we attempted to develop a cascade sequence
using the 3-buten-1-yl ester of acetylene dicarboxylic acid,
which would undergo the metathesis sequence followed by
Diels–Alder cycloaddition. Unfortunately, all attempts at this
cascade resulted only in isolation of simple ring-closing
product 14. However, we were able to develop an efficient
one-pot process. Conducting the metathesis reaction in
dichloromethane followed by a solvent switch to toluene
and addition of DMAD followed by treatment with DDQ
afforded 15 in 83% yield from 11.
To differentiate the two methyl esters, 15 was subjected to
global saponification to reveal a diacid alcohol, which under-
went lactonization using catalytic camphor sulfonic acid in
dichloromethane to afford 8 in 83% yield (Scheme 5). To set
Scheme 4. Formation of the quaternary carbon stereocenter and
a metathesis/Diels–Alder sequence. MOMCl=methoxymethyl chloride,
DIPEA=N,N-diisopropylethylamine, Grubbs-1=Grubbs catalyst of the
first generation, DMAD=dimethyl acetylenedicarboxylate, DDQ=2,3-
dichloro-5,6-dicyano-1,4-benzoquinone.
Scheme 5. Core functionalization. CSA=camphor sulfonic acid,
DPPA=diphenylphosphorylazide, Red-Al=sodium bis(2-methoxy-
ethoxy)aluminum hydride.
enolization of 4a using lithium in ammonia followed by
addition of propargyl bromide formed the desired quaternary
stereocenter in > 95:5 d.r.[10] The aminal in 13 was hydrolyzed
to simplify subsequent NMR spectra (by eliminating amide
rotamers) and also to excise the additional propargyl group to
avoid possible complications in subsequent metathesis and
Diels–Alder chemistry. Treatment with aq. 1m HCl in dioxane
at 238C afforded selective cleavage of the aminal without
affecting the amide. Protection of the resulting primary
alcohol with methoxymethyl chloride and Hꢀnigꢁs base
afforded metathesis precursor 11 in 82% yield over three
steps. Attempted ring-closing ene–yne/diene–ene cross meta-
thesis[9] of 11 with 3-buten-1-ol using either Grubbs first- or
second-generation catalysts afforded only intramolecular
ene–yne metathesis product 14 and the homodimer of 3-
the stage for eventual oxidation to the quinone, the remaining
carboxylic acid was transformed into an amine through
a Curtius rearrangement. Treatment of 8 with diphenylphos-
phorylazide followed by aqueous hydrolysis afforded 16 in
88% yield. In a model system, we noted that oxidation of
a related aniline to the corresponding quinone could only be
achieved when the ring did not bear any electron-withdrawing
groups, and thus we sought to reduce the lactone in 16. We
were pleased to find that treatment of 16 with Red-Al in
toluene at reflux reduced the lactone to the desired methyl
group, thereby affording 18 in 69% yield. Importantly, as long
as a large excess of hydride was not employed, the reaction
2
www.angewandte.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 1 – 5
These are not the final page numbers!

Angewandte
Chemie
proceeded without any noticeable reduction of the secondary
amide. The reduction was accompanied by the formation of
a small amount of dihydropyran 19 (15%). That this was a by-
product and not an intermediate on the reduction pathway
was supported by the fact that resubmission of 19 to the
reduction conditions only resulted in slow reduction of the
amide. We presume that the reduction of the lactone to the
methyl occurs via an ortho-iminoquinone methide intermedi-
ate (17), which is not easily formed from 19 owing to poor
orbital alignment.
To complete the synthesis of 1, all that was seemingly
required was to oxidize to the quinone and remove the chiral
auxiliary. Oxidation of 18 could be easily achieved by using
Fremyꢁs salt in water/acetone to give the valinol amide of
puraquinonic acid (Scheme 6). However, hydrolysis of 20 with
An improved route could be achieved by hydrolyzing the
auxiliary first. Heating of 18 at reflux in aq. 4m H2SO4/
dioxane afforded the desired carboxylic acid 23 in 80% yield.
This reaction was accompanied by the formation of dihydro-
pyran 22 in 15% yield, presumably arising from electrophilic
aromatic substitution with the formaldehyde released during
MOM group cleavage. Finally, the aniline could be cleanly
oxidized to the quinone to afford 1 in 83% yield.
1
Quinone 1 had identical H, 13C NMR, and IR spectro-
scopic characteristics to those reported.[13] However, the
measured specific rotation of + 1.5 (c = 0.3, CHCl3) was
opposite to that expected. Based on our methodology, we
expected that the synthesis beginning with reduction and
alkylation of 3 would ultimately produce (R)-1, as depicted. In
contrast, Clive et al. reported that their synthesis of (S)-1,
wherein the quaternary stereocenter was established via an
Evans aldol followed by a radical cyclization, resulted in
a positive rotation.[5,14]
Given the discrepancy between our observed rotation and
the prior assignment, we reconfirmed the stereochemical
outcome of our alkylation sequence. We initially assigned the
stereochemistry of the alkylation sequence by comparing the
optical rotation of an alkylation product to literature data.[6]
As noted above, we have recently extended our enolate
chemistry to include Mannich additions to benzenesulfonyl-
protected imines.[8] Fortuitously, the stereochemistry of Man-
nich addition products was assigned unambiguously by X-ray
crystallography, and we reasoned that a deamination process
would allow direct comparison to products 24a and 24b
formed from a standard alkylation sequence (Scheme 7).
Thus, reduction of Me/Et-substituted lactam 4b followed by
addition to the benzylsulfonylimine of benzaldehyde and
subsequent acetal hydrolysis afforded Mannich addition
product 25 as reported. The stereochemistry of 25 was
reconfirmed by X-ray crystallography and found to be
consistent with our prior assignment. Direct hydrogenolytic
deamination of 25 proved to be difficult.[15] However a two-
step deamination could be achieved by desulfonylation with
LiDBB followed by deamination via in situ formation of
Scheme 6. Completion of the synthesis of puraquinonic acid.
aq. 4m H2SO4 in dioxane at reflux failed to provide the natural
product. Although the residual auxiliary was hydrolyzed, the
quinone had undergone reductive etherification with the
pendant hydroxyethyl group affording 21 in low yield
(Scheme 6). Switching the co-solvent from dioxane to iso-
propanol improved the yield of 21, but did not solve the
reduction problem. Reduction of quinone dimethyl acetals
under acidic conditions has been reported, with the most
likely reduction source being hydride donation by released
methanol,[12] and in the present case, we presume that dioxane
and isopropanol can also fill this role. Unfortunately, con-
ducting the hydrolysis without organic co-solvents was
inefficient owing to insolubility. Ultimately, hydroquinone
ether 21 could be transformed to puraquinonic acid (1) by
oxidation with Fremyꢁs salt. However, overall this final
sequence was inefficient owing to the need to re-oxidize.
Scheme 7. Stereochemical proof for alkylation chemistry. LiDBB=
lithium di-tert-butylbiphenylide.
Angew. Chem. Int. Ed. 2013, 52, 1 – 5
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3
These are not the final page numbers!

.
Angewandte
Communications
5969 – 5994; h) M. Bella, T. Gasperi, Synthesis 2009, 1583 – 1614;
i) C. Hawner, A. Alexakis, Chem. Commun. 2010, 46, 7295 –
7306; j) J. P. Das, I. Marek, Chem. Commun. 2011, 47, 4593 –
4623.
a monoalkyl diazene by using hydroxylamine O-sulfonic
acid.[16] The deamination product proved to be identical to
24a, which is the major product derived from reduction and
alkylation of 4b with benzyl bromide followed by partial
aminal hydrolysis, and was clearly different than 24b, which is
the product derived from an identical sequence on lactam 4c.
This result confirmed that our original assignment of stereo-
chemistry of the alkylation sequence was correct and strongly
supports the conclusion that we have prepared (R)-puraqui-
nonic acid. We presume that the source of discrepancy
between our observed rotation and the literature lies in the
small absolute value for rotation, which may render the
determination of sign within the error limits of standard
polarimetry on small samples.
In conclusion, the stereoselective synthesis of (R)-pur-
aquinonic acid has been accomplished in an efficient 12 steps
and 20% overall yield. This is the first application of our
bicyclic thioglycolate lactam method towards the synthesis of
a natural product and highlights the ability to fashion
quaternary stereocenters that possess groups with very little
steric or electronic difference. In the context of puraqinonic
acid, the bicyclic lactam allowed the preparation of a stereo-
center bearing allyl and propargyl units, which, through
a metathesis and Diels–Alder sequence, could be transmuted
to set the deceptively difficult stereochemistry of the natural
product.
[2] U. Becker, G. Erkel, T. Anke, O. Sterner, Nat. Prod. Lett. 1997, 9,
229 – 236.
[3] W. R. Abraham, Curr. Med. Chem. 2001, 8, 583 – 606.
[4] G. A. Kraus, P. K. Choudhury, J. Org. Chem. 2002, 67, 5857 –
5859.
[5] a) D. L. J. Clive, M. Yu, Chem. Commun. 2002, 2380 – 2381;
b) D. L. J. Clive, M. Yu, M. Sannigrahi, J. Org. Chem. 2004, 69,
4116 – 4125; For an additional racemic synthesis, see: c) D. L. J.
Clive, M. Sannigrahi, S. Hisaindee, J. Org. Chem. 2001, 66, 954 –
961; d) S. Hisaindee, D. L. J. Clive, Tetrahedron Lett. 2001, 42,
2253 – 2255.
[6] a) J. M. Manthorpe, J. L. Gleason, J. Am. Chem. Soc. 2001, 123,
2091 – 2092; b) J. M. Manthorpe, J. L. Gleason, Angew. Chem.
2002, 114, 2444 – 2447; Angew. Chem. Int. Ed. 2002, 41, 2338 –
2341; c) A. Arpin, J. M. Manthorpe, J. L. Gleason, Org. Lett.
2006, 8, 1359 – 1362.
[7] Thioglycolate lactam 3 is also commercially available from
Aldrich (Cat #667587).
[8] E. A. Tiong, J. L. Gleason, Org. Lett. 2009, 11, 1725 – 1728.
[9] a) H.-Y. Lee, H. Y. Kim, H. Tae, B. G. Kim, J. Lee, Org. Lett.
2003, 5, 3439 – 3442; b) F. Royer, C. Vilain, L. Elkaꢂm, L.
Grimaud, Org. Lett. 2003, 5, 2007 – 2009; c) S. E. Kurhade, A. I.
Sanchawala, V. Ravikumar, D. Bhuniya, D. S. Reddy, Org. Lett.
2011, 13, 3690 – 3693.
[10] Stereoselectivity was determined by HPLC versus the authentic
diastereomer formed by a sequence of methylation, allylation,
and reductive propargylation.
Received: November 1, 2012
Published online: && &&, &&&&
[11] In a model system, an identical observation was made with tert-
butyldimethylsilyl-protected 3-buten-1-ol.
[12] M. P. Capparelli, J. S. Swenton, J. Org. Chem. 1987, 52, 5360 –
5364.
[13] No authentic sample of the natural product was available for
comparison.
[14] Upon the suggestion of a reviewer that the rotation might be due
to contamination by a sodium salt of 1, the synthetic sample was
partitioned between 0.1n HCl and ethyl acetate, concentrated,
and repurified by preparative TLC using 9:1 ethyl acetate/
dichloromethane (conditions identical to Ref. [5]). A positive
rotation for the sample was again observed.
Keywords: alkylation · puraquinonic acid ·
quaternary stereocenter · quinones · total synthesis
.
[1] For reviews, see: a) K. Fuji, Chem. Rev. 1993, 93, 2037 – 2066;
b) E. J. Corey, A. Guzman-Perez, Angew. Chem. 1998, 110,
2092 – 2118; Angew. Chem. Int. Ed. 1998, 37, 388 – 401; c) J.
Christoffers, A. Mann, Angew. Chem. 2001, 113, 4725 – 4732;
Angew. Chem. Int. Ed. 2001, 40, 4591 – 4597; d) I. Denissova, L.
Barriault, Tetrahedron 2003, 59, 10105 – 10146; e) C. J. Douglas,
L. E. Overman, Proc. Natl. Acad. Sci. USA 2004, 101, 5363 –
5367; f) B. M. Trost, C. Jiang, Synthesis 2006, 369 – 396; g) P. G.
Cozzi, R. Hilgraf, N. Zimmermann, Eur. J. Org. Chem. 2007,
[15] D. C. Johnson, T. S. Widlanski, Tetrahedron Lett. 2004, 45, 8483 –
8487.
[16] G. A. Doldouras, J. Kollonitsch, J. Am. Chem. Soc. 1978, 100,
341 – 342.
4
www.angewandte.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 1 – 5
These are not the final page numbers!

Angewandte
Chemie
Communications
Natural Product Synthesis
E. A. Tiong, D. Rivalti, B. M. Williams,
J. L. Gleason*
&&&& — &&&&
A Concise Total Synthesis of (R)-
Puraquinonic Acid
Distant relative: The configuration of the
quaternary stereocenter in puraquinonic
acid (see scheme, right) is defined by
groups far removed from the point of
asymmetry. The use of a bicyclic thiogly-
colate lactam to set the quaternary ste-
reocenter at an early stage cuts the length
of the synthesis of the natural product by
nearly two thirds.
Angew. Chem. Int. Ed. 2013, 52, 1 – 5
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5
These are not the final page numbers!