Studies on the biosynthesis of secobotryane skeleton

Article abstract of DOI:10.1016/S0040-4020(03)01025-1

The labelling and coupling patterns of secobotrytrienediol, biosynthesised from [1-¹³C] and [1,2-¹³C₂]-acetate by the fungus Botrytis cinerea, have been used to define the mode of formation and the biogenetic origin of secobotrytrienediol. [10-²H]-Botrydiol was not incorporated into the secobotryane skeleton. In addition, this feeding experiment led to the isolation of three new unlabelled derivatives possessing a secobotryane skeleton, secobotrydiene-3,10,15-triol, secobotrydiene-3,4,10,15-tetraol, and secobotrytriene-10,12,15-triol.

Full text of DOI:10.1016/S0040-4020(03)01025-1

Tetrahedron 59 (2003) 6267–6271
Studies on the biosynthesis of secobotryane skeleton
a
b
a
c
´
´
Rosa Duran-Patron, Ana J. Colmenares, Antonio Montes, James R. Hanson,
a
Rosario Hernandez-Galan and Isidro G. Collado
a
,
*
´
´
a
´ ´ ´ ´
Departamento de Quımica Organica, Facultad de Ciencias, Universidad de Cadiz, Republica Saharaui s/n, apdo. 40,
´
11510 Puerto Real, Cadiz, Spain
Departamento de Quımica, Universidad del Valle, A.A. 25360 Cali, Colombia
b
´
cThe School of Chemistry, Physic and Environmental Science, University of Sussex, Brighton, Sussex BN1 9QJ, UK
Received 31 March 2003; revised 4 June 2003; accepted 24 June 2003
Abstract—The labelling and coupling patterns of secobotrytrienediol, biosynthesised from [1-13C] and [1,2-13C2]-acetate by the fungus
Botrytis cinerea, have been used to define the mode of formation and the biogenetic origin of secobotrytrienediol. [10-2H]-Botrydiol was not
incorporated into the secobotryane skeleton. In addition, this feeding experiment led to the isolation of three new unlabelled derivatives
possessing a secobotryane skeleton, secobotrydiene-3,10,15-triol, secobotrydiene-3,4,10,15-tetraol, and secobotrytriene-10,12,15-triol.
q 2003 Elsevier Ltd. All rights reserved.
1. Introduction
2. Results and discussion
The major characteristic metabolites of the plant pathogen
Botrytis cinerea possess the botryane skeleton. In addition
there are some other minor phytotoxic metabolites.1
Recently we have reported2 the isolation of secobotry-
trienediol (1), a new slightly phytotoxic compound which
possesses a novel carbon skeleton. From the biosynthetic
studies reported,3 the structure of compound 1 suggests that
it might be a biodegradation product arising from botryanes.
However in our studies4 on the stages at which the
metabolites are formed during the growth of B. cinerea,
secobotrytrienediol (1) had disappeared when botrydial (2)
reached the maximum concentration. Furthermore, when
(4E,8R)-caryophyll-4(5)-en-8-ol5 was fed to B. cinerea,
compound 1 was isolated in good yield while botrydial (2)
could not be detected.
The structure of compound 1 and its derivatives do not obey
the simple isoprene rule. However, if we assumed a
relationship to the botryane skeleton, then the biosynthetic
evidence for the origin of the latter from farnesyl diphos-
phate3 may be extended to these compounds. In order to
confirm this relationship we have carried out feeding
experiments with sodium [1-13C]- and [1,2-13C2]-acetates.
The optimum time for feeding acetate to resting cell cultures
of B. cinerea and for isolating the metabolites was
previously established. As a result of the time:course
study for the production of secobotrytrienediol (1), 48 h
old resting cell cultures of B. cinerea were fed with 13C
labelled sodium acetates. Purification of the fermentation
broths led the isolation of compound 1, which was analysed
by 13C NMR spectroscopy. The 13C–13C couplings which
were observed are given in Table 1. Specific incorporations
were calculated by HREIMS (Table 2).
These observations are intriguing and place some con-
straints on proposals for the formation of this new skeleton.
Consequently, we are investigating the biosynthesis of
secobotrytrienediol (1) in B. cinerea. In this paper, we
describe the incorporation of sodium [1-13C]- and
[1,2-13C2]-acetate into secobotrytrienediol (1), the results
of feeding [10-2H]-botrydiol (3) and botrydial (2) to
B. cinerea and the conclusions that may be drawn from
this work.
When sodium [1-13C]-acetate was fed to B. cinerea as a
pulse at the carefully determined time of maximum
metabolite production, significant enrichments were
observed at C-2, C-4 and C-10. In addition to the
enrichments, couplings were observed in the 13C NMR
spectrum between C-6 and C-7 and between C-7 and C-8
(Scheme 1). A second feeding experiment was carried out
with a 1:2 mixture of sodium [1,2-13C2]-acetate and
unlabelled sodium acetate. The incorporation of intact
13C–13C units was established by analysis of the satellite
signals arising from the 13C–13C coupling. Thus, the
signals for the carbons pairs C-1:C-10, C-2:C-11, C-4:C-5,
Keywords: Botrytis cinerea; secobotryane; biosynthesis; labelling;
isolation.
*
Corresponding author. Tel.: þ34-956-016368; fax: þ34-956-016193;
e-mail: isidro.gonzalez@uca.es
0040–4020/$ - see front matter q 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0040-4020(03)01025-1

´ ´
R. Duran-Patron et al. / Tetrahedron 59 (2003) 6267–6271
6268
Table 1. 100 MHz 13C NMR data, and enrichment and coupling patterns of
compound 1 after feeding of 13C-labelled sodium acetate
quently, it is possible that compound 1 might be
biosynthesised from botrydial (2) by means of a bio-
degradation or detoxification mechanism, as is shown in
Schemes 1 and 2. In order to explore this possibility
we examined the incorporation of [10-2H]-botrydiol (3)
into 1 and carried out a feeding experiment with botrydial
(2).
Carbon
dC (ppm)
Coupling constant (Hz)a
[1-13C]-Acetate [1,2-13C2]-Acetate
1
2
3
4
5
6
7
8
136.1 (s)
126.6 (d)
115.3 (t)
131.1 (d)
148.6 (s)
45.2 (s)
52.3 (t)
52.5 (s)
140.0 (s)
68.3 (t)
14.7 (q)
30.7 (q)c
29.2 (q)c
24.0 (q)
70.3 (t)
45.4
43.5
en
55.7
55.7
35.1
en
35.9
en
45.4
43.5
en
en
Labelled 3 was synthesised in 53% yield by the reduction of
dihydrobotrydial (4) with sodium borodeuteride.6 The triol
was fed to resting cell cultures of B. cinerea at a final
concentration of 0.5 mM. The level of incorporation was
determined by the incorporation of deuterium into
10-dehydroxydihydrobotrydialone7 (5) after 5 days of
fermentation. The isotopic composition of 5, determined
by integration of the resultant 1H NMR spectrum, was
88.1%. Nevertheless, no incorporation of the isotope was
apparent in secobotrytrienediol (1).
en
34.3
34.3;b
b
9
10
11
12
13
14
15
en
35.1
35.9
en
a
en¼enriched.
However, a detailed study of this fermentation led to the
isolation of three new metabolites with the secobotryane
b
c
J7–8 not calculated (C-7 is superimposed on C-8).
Interchangeable signals.
Table 2. Incorporation of label determined by HREIMS from labelled sodium acetate into 1
Labelled precursor
Incorporation into 1 (%)
M
Mþ1
Mþ2
Mþ3
Mþ4
Mþ5
Mþ6
Mþ7
Sodium [1-13C]-acetate
25.6
46.9
31.9
16.4
18.2
19.3
13.1
7.6
7.0
5.6
3.5
2.7
,1
1.5
–
,1
Sodium [1,2-13C2]-acetate
C-6:C-13 and C-8:C-14 showed the expected incorporation
of an intact 13C2 unit (Scheme 2). These results together
with the couplings that were generated between C-6, C-7
and C-8 in the experiment with the [1-13C]-acetate suggests
that secobotrytrienediol (1) is biosynthesised from farnesyl
diphosphate by the same folding as that utilised in the
biosynthesis of botryanes (see Schemes 1 and 2).3 Conse-
skeleton. These were secobotrydiene-3,10,15-triol (6),
secobotrydiene-3,4,10,15-tetraol (7) and secobotrytriene-
10,12,15-triol (8). Some of these metabolites have now been
detected in other fermentation experiments. Their structures
were established by mass spectroscopy and NMR studies.
Their spectra showed no incorporation of deuterium from
[10-2H]-botridiol (3).
Scheme 1. Pattern of folding of farnesyl diphosphate obtained from sodium [1-13C]-acetate.

´ ´
R. Duran-Patron et al. / Tetrahedron 59 (2003) 6267–6271
6269
Scheme 2. Pattern of folding of farnesyl diphosphate obtained from sodium [1,2-13C2]-acetate.
Compounds 6, 7 and 8 were colourless oils. Their IR, 1H and
13C NMR spectra showed that they possessed structures
which were similar to that of secobotrytrienediol2 (1). The
principal difference between the 1H NMR spectra of 6 and 7
and that of 1 was the absence of signals characteristic of the
C-3:C-4 double bond. The presence in the 13C NMR
spectrum of two methylene signals at dC 61.9 and 28.7 in the
spectrum of 6 and a methylene (dC 66.0) and a methyne (dC
69.8) in 7 indicated that these compounds were mono-
hydroxylated at C-3 and dihydroxylated at C-3 and C-4,
respectively. Because of the small amount of compound 7
which was isolated, we were unable to determinate its
configuration at C-4.
hydroxymethyl resonance indicated that compound 8 was
hydroxymethyl derivative of 1. The new hydroxymethyl
group was assigned to C-12 by comparing the 13C NMR
spectrum with that of compound 1. The stereochemistry at
C-6 was shown to be R based on the deshielding of the H-7b
resonance from dH 2.08 in 1 to dH 2.28 in 8. This signal had
been assigned to H-7b because a doublet, which was
correlated with it in the COSY spectrum, was enhanced in
an NOE experiment on irradiation of H-14.
The formation of a hydroxymethyl group at this position has
been found previously in the aromatic botryanes,2 and is
characteristic of the biotransformation of fungitoxic com-
pounds by B. cinerea. These oxidised derivatives possess a
much reduced biological activity when compared to the
parent compounds.8
In order to determine the phytotoxic activity of compounds
with the secobotryane skeleton, in vivo phytotoxicity assays
were carried out with Phaseolus vulgaris using previously
described methodology.9 Although the phytotoxicity of
secobotrytrienediol (1) has already been reported in vitro on
tobacco leaves,2 we reassayed this compound in this
investigation to obtained comparable results. Interestingly,
compounds 1 and 6–8 were inactive at the maximum
concentration tested (1000 ppm).
When botrydial (2) (22 mg) was fed to a culture of B.
cinerea, dihydrobotrydial (4) and a higher amount
(33.2 mg) of botrydial (2) were obtained. However,
secobotrytrienediol (1) was not detected in a careful study
of the fermentation broth.
These data together with the fact that the secobotryanes did
not display phytotoxicity, suggests that these compounds
are biodegradation products arising from a branch in the
biosynthetic pathway that leads to botrydial (2). The triene
1
The absence of a methyl signal in the H and 13C NMR
spectra of compound 8 together with the presence of a new

´ ´
R. Duran-Patron et al. / Tetrahedron 59 (2003) 6267–6271
6270
could arise by an electrocyclic ring opening of a
cyclohexadiene reminiscent of the relatioship between
vitamin D and dehydrocholesterol. The dienes could also
be the precursors of the aromatic botryanes.2 Work is in
progress to establish the stage in the botryane pathway at
which these secobotryanes emerge.
containing 200 ml of Czapek–Dox medium.2 The pH of the
medium was adjusted to 7.0 and each flask was inoculated
with 3£106 fresh conidia. After 48 or 72 h of incubation (see
detailed experiments below) the mycelia was transferred
into the same number of 500 ml Erlenmeyer flasks, each
containing 200 ml of Czapek–Dox medium (without
glucose) and an aseptic solution of the labelled precursor
in H2O or EtOH. After 48 h from the administration of the
precursor, the culture medium and mycelia were separated
by filtration. The broth was saturated with NaCl, and
extracted with EtOAc (£4). The EtOAc extract was washed
with H2O (£3) and then dried over anhydrous Na2SO4.
Evaporation of the solvent at reduced pressure gave a
yellow oil that was separated by means of column
chromatography, with an increasing gradient of ethyl
acetate in petroleum ether. Final purification was carried
out with HPLC.
3. Experimental
3.1. General experimental procedures
Melting points were measured with a Reichert–Jung Kofler
block and are uncorrected. Optical rotations were deter-
mined with a Perkin–Elmer 241 polarimeter. IR spectra
were recorded on a Mattson Genesis spectrophotometer,
1
series FTIRe. H and 13C NMR spectra were recorded on
Varian Gemini 200 MHz (1H at 199.975 MHz, 13C at
50.289 MHz) and Varian Unity 400 MHz (1H at
399.952 MHz, 13C at 100.570 MHz) spectrometers. 2H
NMR Spectra were recorded on a Varian Unity 400 MHz
(2H at 61.395 MHz) spectrometer. Chemical shifts are
quoted relative to TMS (Me4Si) in CDCl3. Mass spectra
were recorded on GC-MS Thermoquest, model Voyager,
and VG-Autospec spectrometers. HPLC was performed
with a Hitachi/Merck L-6270 apparatus equipped with an
UV-VIS detector (L 4250) and a differential refractometer
detector (RI-71). TLC was performed on Merck Kiesegel 60
F254, 0.25 mm thick. Silica gel (Merck) was used for column
chromatography. Purification by means of HPLC was carried
out with a Si gel column (LiChrospher Si-60, 10 mm, 1 cm
wide, 25 cm long or 5 mm, 0.4 cm wide, 25 cm long).
3.5. Feeding of sodium [1-13C]-acetate to B. cinerea
Eleven cultured flasks with B. cinerea were fed with 360 ml
of an aseptic aqueous solution of sodium [1-13C]-acetate to
give a final concentration of 450 ppm per flask on day 2.
Extraction of the broth yielded a crude extract (61 mg),
which was purified as described in the general method to
afford secobotrytrienediol2 (1, 2.3 mg). The isotopic
composition of 1 was determined by HREIMS and is
shown in the Table 2.
3.6. Feeding of sodium [1,2-13C2]-acetate to B. cinerea
Six cultured flasks with B. cinerea were fed with 400 ml of
an aseptic aqueous solution of sodium [1,2-13C2]-acetate:
sodium acetate (1:2) to give a final concentration of
450 ppm per flask on day 2. Extraction of the broth yielded
a crude extract (51 mg), which was purified as described in
the general method to afford 12 (2.1 mg). The isotopic
composition of 1 was determined by HREIMS and is shown
in the Table 2.
3.2. Labelled precursors
Sodium [1-13C]-acetate (99 at.% 13C) and sodium
[1,2-13C2]-acetate (99 at.% 13C) were obtained from the
Aldrich Chemical Company. [10-2H]-Botrydiol (3)6 was
synthesised by treatment of dihydrobotrydial (4, 116 mg,
0.37 mmol) in MeOH (6 ml) with NaBD4 (3 mg,
0.07 mmol) ($99 at.% D, purchased from the Fluka
Chemical Company). After stirring for 2 h at room
temperature, the reaction mixture was poured onto ice,
acidified with 2N HCl (10 ml) and stirred for 10 min. The
solution was then diluted with H2O (45 ml) and extracted
with CHCl3 (30 ml, £3). The solvent was evaporated and
the crude extract subjected to column chromatography to
afford [10-2H]-botrydiol (3, 62 mg, 0.20 mmol, 53%, mp
170–1718C). The isotopic composition, calculated by
3.7. Feeding of [10-2H]-botrydiol to B. cinerea
Eleven cultured flasks with B. cinerea were fed with 375 ml
of an aseptic solution of [10-2H]-botrydiol (3) in EtOH to
give a final concentration of 150 ppm per flask on day 3.
Extraction of the broth yielded a crude extract (590 mg),
which was purified as described in the general method to
afford secobotrytrienediol2 (1, 2.8 mg), secobotrydiene-
3,10,15-triol (6, 1.2 mg), secobotrydiene-3,4,10,15-tetraol
(7, 0.5 mg), secobotrytriene-10,12,15-triol (8, 0.7 mg) and
[10-2H]-dehydroxydihydrobotrydialone7 (5, 14.5 mg). The
isotopic composition of 5, calculated by integration of the
1
integration of the resultant H NMR spectrum, was 98.9%.
3.3. Organism
1
resultant H NMR spectrum, was 88.1%.
´
B. cinerea 2100 was obtained from the Coleccion Espan˜ola
3.7.1. Secobotrydiene-3,10,15-triol (6). Colourless oil;
[a]2D5¼þ57 (c 0.1, CHCl3); IR (KBr) nmax 3354, 2925,
2862, 1739, 1451, 1373, 1261, 1033, 801 cm21; EIMS m/z
(%) 236 (1) [M2H2O]þ, 223 (10) [M2CH2OH]þ, 206 (90)
[M2CH2OH–OH]þ, 175 (82) [M22£CH2OH–OH]þ, 161
(58) [M23£CH2OH]þ, 145 (71), 133 (88), 119 (100);
HREIMS calcd. for C15H26O3 [M]þ 254.1882, found
254.1904. See Tables 3 and 4 for 1H and 13C NMR
spectroscopic data.
´
de Cultivos Tipo (CECT), Facultad de Biologıa, Universi-
dad de Valencia, Spain. Conidial stock suspensions of this
strain were maintained viable in 80% glycerol at 2208C.
3.4. Feeding experiments of labelled precursors:
extraction and isolation of metabolites
General method. B. cinerea 2100 was grown at 24–268C on
shake culture at 250 rpm in 500 ml Erlenmeyer flasks, each

´ ´
R. Duran-Patron et al. / Tetrahedron 59 (2003) 6267–6271
6271
Table 3. 400 MHz 1H NMR data of compounds 6–8 in CDCl3
3.7.3. Secobotrytriene-10,12,15-triol (8). Colourless oil;
[a]2D5¼þ23 (c 0.07, CHCl3); IR (KBr) nmax 3409, 2925,
2855, 1733, 1457, 1375, 1260, 1037 cm21; EIMS m/z (%)
234 (2) [M2H2O]þ, 222 (43) [M22£CH3]þ, 221 (20)
[M2CH2OH]þ, 204 (24) [M22£CH3–H2O]þ, 191 (61)
[M2CH2OH–2£CH3]þ, 173 (100) [M2CH2OH–H2O–
2£CH3]þ. See Tables 3 and 4 for 1H and 13C NMR
spectroscopic data.
Proton
dH (ppm)
6
7
8
2
3
30
4
40
7a
7b
10
100
11
12
120
13
14
15
150
5.60 (qdd)
3.69 (ddd)
3.80 (ddd)
2.07 (ddd)
2.32 (ddd)
1.55(d)
2.13 (d)
4.13 (d)
4.33 (dt)
1.53 (dd)
1.16 (s)a
5.54 (q)
5.71 (q)
3.55 (dd)
3.72 (dd)
4.21 (dd)
5.07 (dd)
5.40 (dd)
6.30 (dd)
1.60 (d)
2.11 (d)
4.17 (d)
4.26 (dt)
1.57 (dd)
1.25 (s)a
1.71(d)
3.8. Feeding of botrydial (2)
2.28 (d)
4.08 (d)
4.34 (dt)
1.53 (dd)
3.45 (d)
4.02 (d)
1.27 (s)
0.99 (s)
3.44 (d)
3.60 (d)
B. cinerea 2100 was grown in two flasks containing
Czapek–Dox medium under the conditions described
above. Each flask was inoculated with 1.1£107 fresh
conidia. An aseptic solution of botrydial (22 mg) (2),
obtained from previous fermentation experiments,9 in
EtOH (450 ml) was fed to the fermentation in one pulse to
give a final concentration of 55 ppm per flask 3 days after
inoculation. The broth was harvested after a further 2 days
as above to afford botrydial6 (2, 33.2 mg) and dihydro-
botrydial6 (4, 0.5 mg).
1.15 (s)a
0.92(s)
3.17 (d)
3.58 (d)
1.12 (s)a
0.94 (s)
3.17 (d)
3.51 (d)
0
0
J (Hz): 6: J2–10¼0.9; J2–10 ¼2.0; J2–11¼6.7; J3–3 ¼10.4; J3–4¼2.3; J3–
0
0
0
0
0
0
¼10.7; J3 –4¼4.5; J3 –4 ¼4.2; J4–4 ¼14.6; J7a–7b¼12.8; J10–10 ¼13.7;
4
0
0
0
0
J10 –11¼2.1; J15–15 ¼11.7. 7: J2–11¼6.7; J3–3 ¼10.8; J3–4¼3.0, J3 –4¼9.8,
J7a–7b¼13.3; J10–10 ¼13.8; J10 –11¼2.3; J15–15 ¼11.5. 8: J2–11¼6.7; J3–
0
0
0
0
0
0
0
¼1.4; J3–4¼11.9; J3 –4¼18.4; J7a–7b¼14.3; J10–10 ¼13.3; J10 –11¼1.6;
3
0
0
J12–12 ¼10.6; J15–15 ¼11.5.
a
Interchangeable signals.
Acknowledgements
Table 4. 100 MHz 13C NMR data of compounds 6–8 in CDCl3
This research was supported by a grant from the DGICYT
AGL2000-0635-C02-01.
Carbon
dC (ppm)
6
7
8
1
2
3
4
5
6
7
8
137.1 (s)
123.1 (d)
61.9 (t)
137.0 (s)
122.9 (d)
66.0 (t)
69.8 (d)
148.8 (s)
46.0 (s)
50.1 (t)
54.2 (s)
137.7 (s)
66.5 (t)
16.3 (q)
31.1 (q)a
28.3 (q)a
24.4 (q)
70.4 (t)
References
126.2 (d)
114.9 (t)
131.2 (d)
28.7 (t)
´ ´ ´ ´
1. Collado, I. G.; Aleu, J.; Hernandez-Galan, R.; Duran-Patron, R.
Curr. Org. Chem. 2000, 4, 1261–1286.
´
46.4 (s)
49.2 (t)
52.8 (s)
´ ´ ´
2. Duran-Patron, R.; Hernandez-Galan, R.; Collado, I. G. J. Nat.
49.3 (t)
Prod. 2000, 63, 182–184.
3. (a) Bradshaw, A. P. W.; Hanson, J. R.; Nyfeler, R. J. Chem. Soc.
Perkin Trans. 1 1981, 1469–1472. (b) Hanson, J. R. Pure Appl.
9
10
11
12
13
14
15
67.4 (t)
14.3 (q)
30.3 (q)a
29.2 (q)a
23.2 (q)
70.0 (t)
68.0 (t)
14.4 (q)
69.3 (t)
26.5 (q)
25.1 (q)
71.9 (t)
´
Chem. 1981, 53, 1155–1162. (c) Duran-Patron, R.;
´
´ ´
Colmenares, A. J.; Hernandez-Galan, R.; Collado, I. G.
Tetrahedron 2001, 57, 1929–1933.
4. Unpublished results.
´ ´ ´
5. Duran-Patron, R.; Aleu, J.; Hernandez-Galan, R.; Collado, I. G.
a
´
Interchangeable signals.
J. Nat. Prod. 2000, 63, 44–47.
6. Felhaber, H. W.; Geipel, R.; Mercker, H. J.; Tschesche, R.;
Welmar, K. Chem. Ber. 1974, 107, 1720–1730.
3.7.2. Secobotrydiene-3,4,10,15-tetraol (7). Colourless oil;
[a]2D5¼þ14 (c 0.05, CHCl3); IR (KBr) nmax 3350, 2832,
1460, 1378, 1723, 824 cm21; EIMS m/z (%) 271 (6)
[Mþ1]þ, 221 (99) [M2CH2OH–H2O]þ, 204 (61)
[Mþ12CH2OH–2£H2O]þ, 191 (65) [Mþ122£CH2OH–
H2O]þ, 173 (45) [Mþ122£CH2OH–2£H2O], 119 (100);
HREIMS calcd. for C15H27O4 [Mþ1]þ 271.1909, found
271.1898. See Tables 3 and 4 for 1H and 13C NMR
spectroscopic data.
´ ´
7. Collado, I. G.; Hernandez-Galan, R.; Prieto, V.; Hanson, J. R.;
Rebordinos, L. G. Phytochemistry 1996, 41, 513–517.
´ ´
8. Aleu, J.; Hanson, J. R.; Hernandez-Galan, R.; Collado, I. G.
J. Nat. Prod. 1999, 62, 437–440.
´ ´
9. Colmenares, A. J.; Aleu, J.; Duran-Patron, R.; Collado, I. G.;
´
Hernandez-Galan, R. J. Chem. Ecol. 2002, 28, 997–1005.
´