Precursors to oak lactone: Synthesis of gallate ester derivatives of 3-methyl-4-hydroxyoctanoic acid

Article abstract of DOI:10.1016/S0040-4039(01)01890-1

(3R,4R- and 3S,4S)-3-Methyl-4-(3′-O-methylgalloyloxy)octanoic acid (2b), the corresponding methyl ester (2c) and the straight galloyl derivative, (3R,4R- and 3S,4S)-3-methyl-4-galloyloxyoctanoic acid (2a) were synthesised from cis-oak lactone (1). Analysis by TLC and mass spectrometry established that the original assignment of structure (2c) to a methylated natural oak component was in error.

Full text of DOI:10.1016/S0040-4039(01)01890-1

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 42 (2001) 8717–8719
Precursors to oak lactone: synthesis of gallate ester derivatives
of 3-methyl-4-hydroxyoctanoic acid
Michael Raunkjær, D. Sejer Pedersen, Gordon M. Elsey, Mark A. Sefton* and
George K. Skouroumounis
The Australian Wine Research Institute, PO Box 197, Glen Osmond, South Australia 5064, Australia
Received 22 June 2001; revised 24 September 2001; accepted 5 October 2001
Abstract—(3R,4R- and 3S,4S)-3-Methyl-4-(3%-O-methylgalloyloxy)octanoic acid (2b), the corresponding methyl ester (2c) and the
straight galloyl derivative, (3R,4R- and 3S,4S)-3-methyl-4-galloyloxyoctanoic acid (2a) were synthesised from cis-oak lactone (1).
Analysis by TLC and mass spectrometry established that the original assignment of structure (2c) to a methylated natural oak
component was in error. © 2001 Elsevier Science Ltd. All rights reserved.
Results and discussion
These observations indicate the presence of one or
more precursor forms of oak lactone in oak. Several
potential precursors have been observed: Otsuka et al.4
isolated, from oak wood powder, a compound which
they termed a precursor to oak lactone. On the basis of
rather limited degradation studies, they assigned a
structure to this compound. Although the structure of
the natural species was claimed as (2b), the compound
which was actually isolated was thought to be the
corresponding methyl ester (2c), formed as an artefact
of the isolation process. Despite the limited evidence
presented, this compound as a precursor to oak lactone
has gained credence through literature citation.1
The 4S,5S-(cis) and 4S,5R-(trans) isomers of 5-n-butyl-
4-methyl-4,5-dihydro-2(3H)-furanone (1), also known
as ‘oak lactone’ or ‘whisky lactone’, are natural oak
components, extracted into wine and spirits during oak
barrel maturation. Of the two, the cis-isomer is consid-
ered to be the more important in sensory terms, and is
thought to impart coconut and vanilla-like nuances to
wine.1,2 Despite their importance, the origin of these
compounds remains unclear. cis- and trans-Oak lactone
are already present in green oakwood, but additional
quantities of these compounds can be generated in the
wood during the drying (seasoning) and coopering pro-
cesses,1 in model wine oak extracts heated to 50°C,3
and even in the injector block of a gas chromatograph
during analysis of oak extracts.2
More recently Tanaka and co-workers5 isolated several
interesting compounds from the wood of Platycarya
strobilacea, a member of the walnut (Juglandaceae)
O
O
O
O
OR1
OH
O
O
O
O
O
O
OH
OH
3, R=H
4, R=
OH
O
cis-1
trans-1
OR2
OH
RO
HO
OH
OH
OH
2a, R1=H R2=H
2b, R1=H R2=Me
2c, R1=Me R2=Me
Keywords: oak lactone; flavour precursors; wine; gallate esters.
* Corresponding author. E-mail: msefton@awri.adelaide.edu.au
0040-4039/01/$ - see front matter © 2001 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(01)01890-1

8718
M. Raunkjær et al. / Tetrahedron Letters 42 (2001) 8717–8719
family of trees. Among this collection of compounds
were two which could conceivably serve as precursors
to oak lactone, viz. (3) and (4). Hydrolysis of the
glycosides (3) and (4), albeit under strongly acidic con-
ditions, did indeed produce the (4S,5S) cis-oak lac-
tone.5,6 The identification of the galloyl glucoside (4) in
this wood, and its subsequent isolation from oak6 is of
particular interest, as the structural similarity to (2b)
allows the possibility of the misidentification of this
compound by Otsuka et al. Our interest was aroused by
this potential ambiguity and, therefore, we set out to
synthesise (2b) and (2c) in an unambiguous fashion.
Because of the possibility that O-methylation on the
aromatic ring might also have occurred as an artefact
of the isolation by Otsuka et al., the straight gallate
ester (2a) was also synthesised; 3-O-methylgallic acid
derivatives are relatively uncommon in nature, whereas
gallate esters are major constituents of oak and other
woods.
tively. Deprotection with fluoride regenerated the
primary alcohols which were oxidised to the acids by
the use of the free radical TEMPO and bis-ace-
toxyiodobenzene (BAIB) in aqueous acetonitrile.10 The
acid (7b) was esterified to give the corresponding
methyl ester in 97% yield. Removal of the benzyl
protecting groups to produce the desired gallic acid
derivatives (2a–c)11 was effected in excellent yields by
hydrogenolysis over palladium in acetone.12,13
Because of the absence of any definitive structural
evidence in the original report concerning the role of
(2b) as precursor to oak lactone, it was necessary to
repeat the TLC work of Otsuka et al. on the corre-
sponding methyl ester (2c) as this was the crux of his
argument. The results are collected in Table 1.
Notwithstanding the difficulty in reproducing exactly
the conditions reported by Otsuka,4 it is clear that the
Rf values for the compound reported by him as (2c), are
not in agreement with those of the authentic sample
Initial attempts involved ring-opening of the oak lac-
tone and protection of the acid functionality as its
isopropyl ester.7 Our attempts at coupling this ester
with various substituted gallic acid chlorides were, on
the whole unsuccessful; the major problem appeared to
be the overwhelming preference for the ester to undergo
relactonisation to the oak lactone, even when strongly
buffered the oak lactone was the major product. To
overcome this lactonisation problem, an alternative
strategy was employed (Scheme 1) which involved ini-
tial reduction of the ester function with lithium alu-
minium hydride.8 Coupling of the protected species (5)
with the gallic acid derivatives (8) and (9) in the pres-
ence of DCC and 4-DMAP9 gave (6a) and (6b), respec-
Table 1. Comparison of Rf values between this work and
Ref. 4
This work
Gallic acid
Otsuka Ref. 4
2c
Gallic acid
2c
Sol6ent system
A
Sol6ent system
0.39
0.13
0.64
0.69
0.51
0.25
0.40
0.19
B
A: 40:10:2:10 (benzene:dioxane:acetic acid:methanol).
B: 50:30:20:0.2 (chloroform:ethyl acetate:methanol:1N NH4OH).
O
O
OTBDPS
OTBDPS
a)-b)
c) or d)
OH
O
O
R
(±) cis-1
5
6
e)-f)
O
OMe
OBn
OBn
OBn
OBn
OBn
OH
a, R=
b, R=
O
O
R
7
O
OH
O
OH
BnO
OBn
BnO
OMe
OBn
OBn
8
9
Scheme 1. Reagents and conditions: (a) LiAlH4, THF, D, 97%; (b) TBDPSCl, pyr., 92%; (c) DCC, DMAP, 8, CH2Cl2, 88%; (d)
DCC, DMAP, 9, CH2Cl2, 96%; (e) TBAF, THF (a 88%, b 94%); (f) TEMPO, BAIB, CH3CN/H2O (a 84%, b 92%).

M. Raunkjær et al. / Tetrahedron Letters 42 (2001) 8717–8719
matogr. A 1999, 857, 239.
8719
prepared in this study. In both solvent systems the
authentic (2c) elutes much more rapidly than gallic
acid, whereas in the original report the converse was
found to be the case. In addition to the non-congruence
of the TLC data, the mass spectral data for the authen-
tic sample in no way matches the reported spectrum.
The spectrum of authentic (2c) is dominated by ions
due to aromatic fragments; the base peak (m/z 167) for
this compound appears as a very minor peak in the
published spectrum, whereas the published base peak
(m/z 99) does not appear at all in the spectrum of
authentic (2c).
3. Pollnitz, A. P. Ph.D. Thesis; University of Adelaide,
2000.
4. Otsuka, K.; Sato, K.; Yamashita, T. J. Ferment. Technol.
1980, 58, 395.
5. Tanaka, T.; Kuono, I. J. Nat. Prod. 1996, 59, 997.
6. Masson, E.; Baumes, R.; Le Guerneve´, C.; Puech, J.-L. J.
Agric. Food. Chem. 2000, 48, 4306.
7. Gu¨nther, C.; Mosandl, A. Liebigs Ann. Chem. 1986, 2112.
8. Skouroumounis, G.; Winter, B. Helv. Chim. Acta 1996,
79, 1095.
9. Nieses, B.; Steglich, W. Angew. Chem., Int. Ed. Engl.
1978, 17, 522.
These results indicate that the compound isolated from
oakwood by Otsuka et al.4 did not have the structure
(2c) as proposed by these authors. It is possible that
their compound was an artefactually methylated ana-
logue of the known oak compound (4), or was some
other galloylated derivative of 3-methyl-4-hydroxy-
octanoic acid. The status of the simple gallate ester (2a)
as a possible oak component is under investigation.
10. (a) De Mico, A.; Margarita, R.; Parlanti, L.; Vescovi, A.;
Piancatelli, G. J. Org. Chem. 1997, 62, 6974; (b) Epp, J.
B.; Widlanski, T. S. J. Org. Chem. 1999, 64, 293.
1
11. Selected spectral data for 2a: H NMR (acetone-d6, 300
MHz): l 7.18 (2H, s, ArH), 5.07 (1H, ddd, J=8.5, 4.7
and 3.8 Hz, H4), 2.46 (1H, dd, J=15.2 and 5.4 Hz, H2a),
2.39–2.28 (1H, m, H3), 2.20 (1H, dd, J=15.2 and 8.1 Hz,
H2b), 1.78–1.64 (2H, m, H5), 1.44–1.28 (4H, m, H6, H7),
1.08 (3H, d, J=6.8 Hz, Me), 0.91 (3H, t, J=6.7 Hz, H8);
13C NMR (acetone-d6, 75.5 MHz): l 173.2 (C1), 165.9
(CꢀO), 145.4, 138.1, 121.5, 109.2 (Ar), 76.4 (C4), 37.8
(C2), 33.9 (C3), 31.2 (C5), 28.3 (C6), 22.7 (C7), 14.3 (Me),
13.8 (C8).
Acknowledgements
We wish to thank the Grape and Wine Research and
Development Corporation for financial assistance. We
wish also to thank Professors R. Baumes (Montpellier)
and R. Prager (Flinders) for useful discussions. Dr. A.
Pollnitz, Ms. Dimi Capone, Mr. K. Pardon and Ms. K.
Wilkinson are thanked for help in various areas around
the lab, and Professor Peter Høj is thanked for his
ceaseless encouragement.
Selected spectral data for 2b: 1H NMR (acetone-d6): l
7.29 (1H, d, J=1.8 Hz, ArH), 7.23 (1H, d, J=1.8 Hz,
ArH), 5.10 (1H, ddd, J=8.6, 4.6 and 3.8 Hz, H4), 3.91
(3H, s, OMe), 3.35 (3H, br s, OH), 2.48 (1H, dd, J=15.2
and 5.5 Hz, H2a), 2.40–2.30 (1H, m, H3), 2.20 (1H, dd,
J=15.2 and 8.0 Hz, H2b), 1.82–1.64 (2H, m, H5), 1.44–
1.28 (4H, m, H6, H7), 1.10 (3H, d, J=6.8 Hz, Me), 0.92
(3H, t, J=6.7 Hz, H8); 13C NMR (acetone-d6): l 173.3
(C1), 165.9 (CꢀO), 148.1, 145.3, 139.2, 121.4, 111.1, 105.4
(Ar), 76.7 (C4), 56.1 (OMe), 37.9 (C2), 34.0 (C3), 31.2
(C5), 28.4 (C6), 22.8 (C7), 14.5 (Me), 13.9 (C8).
12. Bieg, T.; Szeja, W. Synthesis 1985, 76.
References
1. Maga, J. A. Food Rev. Int. 1996, 12, 105.
13. All intermediates gave satisfactory spectral and elemental
2. Pollnitz, A. P.; Jones, G. P.; Sefton, M. A. J. Chro-
data.