A general route to 5-substituted-2-furylacetic acids: a brief synthesis of plakorsin B

Article abstract of DOI:10.1016/j.tetlet.2009.11.120

3,4-Dihydroxy-5-alkynylcarboxylic acids, readily obtained by the addition of lithium acetylides to α-acetoxysuccinic anhydride followed by reduction and hydrolysis, undergo smooth silver(I)-catalysed 5-endo-dig cyclisations and in situ dehydration to give excellent overall yields of 5-substituted-2-furylacetic acids, including the natural metabolite plakorsin B.

Full text of DOI:10.1016/j.tetlet.2009.11.120

Tetrahedron Letters 51 (2010) 717–719
Contents lists available at ScienceDirect
Tetrahedron Letters
journal homepage: www.elsevier.com/locate/tetlet
A general route to 5-substituted-2-furylacetic acids: a brief synthesis of
plakorsin B
a
b
Simon J. Hayes a, David W. Knight a, , Andrew W. T. Smith , Mark J. O’Halloran
*
a School of Chemistry, Cardiff University, Main College, Park Place, Cardiff CF10 3AT, UK
b Technical & Business Development, GlaxoSmithKline, Shewalton Road, Irvine KA11 5AP, UK
a r t i c l e i n f o
a b s t r a c t
Article history:
3,4-Dihydroxy-5-alkynylcarboxylic acids, readily obtained by the addition of lithium acetylides to a-acet-
Received 9 October 2009
Revised 17 November 2009
Accepted 27 November 2009
Available online 1 December 2009
oxysuccinic anhydride followed by reduction and hydrolysis, undergo smooth silver(I)-catalysed 5-endo-
dig cyclisations and in situ dehydration to give excellent overall yields of 5-substituted-2-furylacetic
acids, including the natural metabolite plakorsin B.
Ó 2009 Elsevier Ltd. All rights reserved.
Keywords:
2-Furylacetic acids
5-endo-dig
Silver-catalysed
Furan
Plakorsin
In the preceding paper,1 we reported that 2-furylethanols 1
have a strong tendency to rearrange into the keto-tetrahydrofurans
2, probably by sequential oxidative ring opening to the corre-
sponding (Z)-enediones followed by 5-exo-trig Michael-type ring
closure, when treated with a variety of oxidising agents, especially
the Jones reagent (Scheme 1).
This may account for the fact that the oxidation of 2-furyletha-
nols to the corresponding 2-furylacetic acids is not represented in
the current chemical literature. We were therefore left in something
of a quandary as to how to apply our very efficient furan synthesis,
wherein 3-alkyne-1,2-diols 3 are converted directly into furans 4
and water using heterogeneous silver(I) catalysts (Scheme 2),2 to
the synthesis of naturally occurring 2-furylacetic acids.
Specific examples of interest were the fatty acid derivatives
plakorsins A and B 5 (Scheme 3), isolated from the marine sponge,
Plakortis simplex.3 While these are simple enough metabolites, they
do display some useful cytotoxic activities against various cancer
cells lines and, of more chemical interest, are the precursors of
the much more densely functionalised metabolite manzamenone
A 6, into which they are converted following sequential oxidative
ring opening (the first step in Scheme 1), aldol condensation,
dehydrative dimerisation and finally a retro-Dieckmann ring clo-
sure.4 Additional members of this complex group of metabolites
are members of the pyrrolidine alkaloids, the plakoridines; by a
neat piece of logical deduction, it has been proposed that these
compounds are all interrelated biosynthetically.5
O
R
Jones
R
OH
O
O
O
1
2
Scheme 1.
R2
R1
OH
R2
R1
R3
Ag(I)
> 95%
H2O
+
R3
O
OH
3
4
Scheme 2.
MeO2C
C16H33
CO2Me
H
H
C16H33
CO2R
C16H33
O
5a plakorsin A; R = Me
5b plakorsin B; R = H
O
HO2C
6 manzamenone A
* Corresponding author.
E-mail address: knightdw@cf.ac.uk (D.W. Knight).
Scheme 3.
0040-4039/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2009.11.120

718
S. J. Hayes et al. / Tetrahedron Letters 51 (2010) 717–719
Li
5
C15H31
O
O
O
O
R
R
13
OH
CN
O
CO2Me
THF, - 78 °C
0.5 h
7
8
9
AcO
CO2Li
Scheme 4.
14
i) NaBH4, EtOH, -78 °C
ii) 1 M NaOH, warm to 0 °C
The Whitehead group synthesis of the plakorsins 5 occupied
their attention for some time. The successful five-step synthe-
sis (Scheme 4) was based on classical steps and consisted of
chlorination of very cheap 2-furylmethanol 7 and homologation
to the nitrile 8, hydrolysis to the corresponding methyl ester, intro-
duction of the C16 fatty side chain by Friedel–Crafts acylation to
give the keto-ester 9 and finally a Wolff–Kishner reduction. Ini-
tially, and perhaps unsurprisingly given the sensitive nature of fur-
ans in general, this was a rather inefficient synthesis but the group
have subsequently achieved remarkable improvements to this by
very careful optimisation such that it is now a very effective ap-
proach to the plakorsins 5 and, presumably, to many other 5-
substituted-2-furylacetic acids in general.6
In spite of these improvements, this is still quite a demanding
synthesis which utilises a large range of reagents and solvents, so
we were still keen to apply our silver(I)-catalysed method and
reasoned that, in view of the foregoing problems associated with
oxidation chemistry (Scheme 1), a direct approach might be
better. We therefore wondered if the generalised 5-substituted-
2-furylacetic acids 10 could be obtained from the corresponding
dihydroxyalkynoic acids 11, without interference from the free
carboxylic acid group and further of course, if such precursors
could be easily synthesised in a general manner (Scheme 5).
Fortunately, the required precursors 11 are available using a
OH
R
HO
11
CO2H
Scheme 7.
temperature, acidified and the products were extracted into ethyl
acetate. The resulting polar dihydroxy acids 11 were obtained in
around 60–70% yields and were essentially a clean mixture of dia-
stereoisomers according to 1H NMR analysis and were not further
purified beyond drying under high vacuum.
We were delighted to find that subsequent exposure of the
dihydroxy acids 11 to 10 mol % of 10% w/w silver(I) nitrate ad-
sorbed on silica gel in dichloromethane generated excellent yields
of the desired 2-furylacetic acids 10.2 As far as we could judge,
these were all well in excess of 90% and hence the overall yields
quoted almost entirely reflect the losses at the acetylide addition
stage. The results are collected in the Table 1.8
It should be noted that these are all unoptimised yields. The
method was equally effective for aryl- and alkyl-substituted al-
kynes and the products (entries 1 and 2) and tert-butyldimethylsi-
lyl groups were completely stable to all conditions used (entries 3–
5), even when there was a possibility of competing cyclisation of
this function onto the alkyne (entry 5). A sensitive citronellyl ter-
pene residue also survived unmolested (entry 6). Finally, the meth-
od delivered a decent yield of plakorsin B (entry 7), without the
need to modify the reaction conditions in view of the presence of
a lengthy alkyl chain.
In general, the dihydroxy acids 11 were isolated as 1:1–2:1 mix-
tures of diastereoisomers; we noticed little difference in the rates
at which each underwent cyclisation and dehydration to the furyl-
acetic acids 10. In an effort to probe this aspect further, the mixture
of dihydroxy acids 11 [R = Bu] derived from 1-hexyne was treated
with catalytic p-toluenesulfonic acid in toluene at 50 °C for two
hours,7 after which we isolated a 51:49 mixture of the two possible
lactones 15a,b. When this mixture was treated with 10% w/w sil-
one-pot synthesis from
a-acetoxysuccinic anhydride 13, which is
readily derived from cheap (DL)-malic acid (hydroxysuccinic acid)
12 by recrystallisation from acetyl chloride (Scheme 6).7
Subsequent addition of a lithio acetylide to the anhydride 13
follows a Bürgi–Dunitz trajectory to give the keto-acetylides 14
(Scheme 7). In our hands, if such a reaction was worked up at this
stage, these were the only products (as the free acids), according to
1H NMR analysis, indicating that the attack of the acetylide is
extremely regioselective in terms of both the functional group
selectivity and the position and angle of attack. Minor constituents
of the crude product were the starting acetylide and succinic acid
residues, indicating that a side reaction might be deprotonation
of the anhydride by the acetylide, however, this did not represent
a serious loss. In normal circumstances however, we followed the
original Scheme 7 and added ethanolic sodium borohydride to
the reaction mixture to reduce the ketone group. This was followed
by the addition of aqueous sodium hydroxide to hydrolyse all the
esters present after which the mixture was warmed to ambient
Table 1
The conversion of 3,4-dihydroxy-5-alkynoic acids 11 into 5-substituted-2-furylacetic
acids 10 using silver(I) nitrate-SiO2
OH
OH
?
AgNO3-SiO2
R
Scheme 7
R
CO2H
13
R
R
O
O
CH2Cl2, 20 °C
~ 4 h
HO
11
CO2H
HO
CO2H
CO2H
10
10
11
Overall yielda (%)
Scheme 5.
Entry
R
1
2
3
4
5
6
7
Ph
Bu
TBSOCH2
TBSO(CH2)2
TBSO(CH2)3
Me2C:CH(CH2)2CH(Me)(CH2)2
Me(CH2)16
85
82
59
67
68
71
55
OH
OAc
AcCl
HO2C
CO2H
recryst.
O
O
O
13
12
a
From the anhydride 13.
Scheme 6.

S. J. Hayes et al. / Tetrahedron Letters 51 (2010) 717–719
719
silver-catalysed cyclisations of 3,4-dihydroxy-5-alkynoic acids 11
were completely regioselective: no products from competing 6-
exo-dig cyclisations of the carboxylic acid group onto the alkyne
were observed. In the case of lower homologues such as (Z)-alk-
2-en-4-ynoic acids, silver-catalysed 5-exo-dig cyclisations of these
have long been established as a useful route to ylidenebuteno-
lides.11 In the current examples, as mentioned above, it seems very
likely that the presence of a hydroxy group adjacent to the alkyne
assists silver complexation to the latter and hence in accelerating
the 5-endo-dig process.12
11
[R = Bu]
H+, Δ
Bu
Bu
O
O
O
O
HO
HO
15b
15a
45%
Acknowledgements
Ag+
Ag+
We are grateful to Dr. Roger C. Whitehead (University of Man-
chester) for providing helpful comments and to GSK and the EPSRC
for financial support.
H
H
H
H
O
O
Bu
Bu
O
Bu
O
O
O
O
References and notes
16a
16b
1. Hayes, S. J.; Knight, D. W.; Smith. A.W.T.; O’Halloran, M. J. Tetrahedron Lett.
2009, 50.
2. Hayes, S. J.; Knight, D. W.; Menzies, M. D.; O’Halloran, M.; Tan, W.-F.
Tetrahedron Lett. 2007, 48, 7709–7712.
3. Shen, Y. C.; Prakash, C. V.; Kuo, Y. H. J. Nat. Prod. 2001, 64, 324–327.
4. Al-Busafi, S.; Doncaster, J. R.; Drew, M. G. B.; Regan, A. C.; Whitehead, R. C. J.
Chem. Soc., Perkin Trans. 1 2002, 476–484.
5. Etchells, L. L.; Sardarian, A.; Whitehead, R. C. Tetrahedron Lett. 2005, 46, 2803–
2807.
CO2H
45%
10 [R = Bu]
6. Etchells, L. L.; Helliwell, M.; Kershaw, N. M.; Sardarian, A.; Whitehead, R. C.
Tetrahedron 2006, 62, 10914–10927.
Scheme 8.
7. Rej, R.; Nguyen, D.; Go, B.; Fortin, S.; Lavallée, J.-F. J. Org. Chem. 1996, 61, 6289–
6295.
8. 2-(5-Butylfuran-2-yl)acetic acid [10;
R = Bu] (Table 1; entry 2): 3,4-
ver(I) nitrate-silica gel, only one of the diastereoisomeric lactones
underwent cyclisation, while the other remained unaltered
(Scheme 8).
Dihydroxydec-5-ynoic acid [11; R = Bu] (0.72 g, 3.61 mmol) was dissolved in
CH2Cl2 (5 ml) and the flask was wrapped in aluminium foil. Silver nitrate on
silica gel (0.613 g, 0.36 mmol of 10% w/w) was added and the solution was
stirred in the dark for 4 h, then filtered through Celite and the combined
filtrates and washings were evaporated under reduced pressure to leave the 2-
This must be due to the high degree of strain associated with
trans-fused 5/5 cyclic systems and hence the isomer 16b is not
formed under these relatively mild conditions. Each compound,
10 [R = Bu] and lactone 15b was isolated in 45% yield. Clearly, this
is mainly of mechanistic interest and does reveal a limitation of
this silver-catalysed chemistry, which is not able to trigger cyclisa-
tions leading to highly strained systems, although it could prove
useful in separating diastereoisomeric mixtures of lactones of type
15, but obviously somewhat wastefully. It has been established
previously that propargylic hydroxy groups assist the binding of
silver(I) ions to alkynes.9 Clearly this is not occurring in the case
of cyclisation of the lactone 15a although its cyclic nature may well
be an alternative facilitating feature.
furylacetic acid [10; R = Bu] as an orange oil (0.599 g, 91%): m
max/cmꢀ1 (film)
3422, 2960, 2873, 1716, 1566, 1466, 1420, 1232, 1015, 906,733, 650; dH
(400 MHz; CDCl3) 6.06 (1H, d, J 3.0 Hz, 3-H), 5.85 (1H, d, J 3.0 Hz, 4-H), 3.62 (2H,
s, 10-CH2), 2.52 (2H, t, J 7.6 Hz, 100-CH2), 1.58–1.48 (2H, m, 200-CH2), 1.36–1.24
(2H, m, 300-CH2), 0.85 (3H, t, J 7.3 Hz, 400-Me); dC (125 MHz; CDCl3) 175.6 (C),
156.5 (C), 144.8 (C), 108.9 (3-CH), 105.5 (4-CH), 33.9 (10-CH2), 30.1 (100-CH2),
27.7 (200-CH2), 22.3 (300-CH2) and 13.8 (400-Me); m/z (EI) 182 (M+, 50%), 137
(100) [Found M+, 182.0946. C10H14O3 requires M, 182.0943].
9. Dalla, V.; Pale, P. New J. Chem. 1999, 23, 803.
10. For the elaboration of synthetic analogues of the structures described in this
paper, see: Doncaster, J. R.; Etchells, L. L.; Kershaw, N. M.; Nakamura, R.; Ryan,
H.; Takeuchi, R.; Sakaguchi, K.; Sardarian, A.; Whitehead, R. C. Bioorg. Med.
Chem. Lett. 2006, 16, 2877–2881.
11. For an application to the synthesis of Patulin oxime, see: Serratosa, F.
Tetrahedron 1961, 16, 185–191; For
a modified method, see: Kotora, M.;
Negishi, E. Synthesis 1997, 121–128. and for a very recent application; Doroh,
B.; Sulikowski, G. A. Org. Lett. 2006, 8, 903–906; For reviews of
ylidenebutenolides synthesis, see: Negishi, E.; Kotora, M. Tetrahedron 1997,
53, 6707–6738. and references therein.
In summary, we contend that the present chemistry represents
a rather brief, direct and viable route to 5-substituted-2-furylacetic
acids, which should find many applications in the synthesis of
these compounds, which are not easy to prepare by the oxidation
of the corresponding 2-furylethanols.10 It is notable that these
12. For an excellent review of this aspect of silver-catalysed cyclisations and of
such reactions in general, see: Weibel, J.-M.; Blanc, A.; Pale, P. Chem. Rev. 2008,
108, 3149–3173.