2-Hydroxyflavanones from Leptospermum polygalifolium subsp. polygalifolium Equilibrating sets of hemiacetal isomers

Article abstract of DOI:10.1016/j.phytochem.2003.09.006

2,5-Dihydroxy-6-methyl-7-methoxyflavanone and its isomer 2,5-dihydroxy-8-methyl-7-methoxyflavanone have been identified as the principal components in the ethanol extract of L. polygalifolium subsp. polygalifolium. The former compound has been separated from the latter for the first time by crystallisation. Other flavonoids identified include an equilibrating mixture of novel 2,3,5-trihydroxyflavanones; the first discovery of this type of compound from a natural source. This study has also confirmed the presence of the triketones, flavesone and leptospermone, along with the closely related isoleptospermone as anti-microbial components.

Full text of DOI:10.1016/j.phytochem.2003.09.006

Phytochemistry 64 (2003) 1285–1293
www.elsevier.com/locate/phytochem
2-Hydroxyflavanones from Leptospermum polygalifolium
subsp. polygalifolium
Equilibrating sets of hemiacetal isomers
Kamarul’Ain Mustafaa, Nigel B. Perryb, Rex T. Weaversa,*
aDepartment of Chemistry, University of Otago, PO Box 56, Dunedin, New Zealand
bPlant Extracts Research Unit, New Zealand Institute for Crop and Food Research Limited, Department of Chemistry, University of Otago, PO Box 56,
Dunedin, New Zealand
Received 25 June 2003; received in revised form 2 September 2003
Abstract
2,5-Dihydroxy-6-methyl-7-methoxyflavanone and its isomer 2,5-dihydroxy-8-methyl-7-methoxyflavanone have been identified
as the principal components in the ethanol extract of L. polygalifolium subsp. polygalifolium. The former compound has been
separated from the latter for the first time by crystallisation. Other flavonoids identified include an equilibrating mixture of
novel 2,3,5-trihydroxyflavanones; the first discovery of this type of compound from a natural source. This study has also
confirmed the presence of the triketones, flavesone and leptospermone, along with the closely related isoleptospermone as
anti-microbial components.
# 2003 Elsevier Ltd. All rights reserved.
Keywords: L. polygalifolium subsp. polygalifolium; Myrtaceae; Flavonoids; Hemiacetals; 2,5-Dihydroxyflavanones; 2,3,5-Trihydroxyflavanones;
Antimicrobial activity
1. Introduction
the absence of oxygenation in the B-ring. Our interest in
the chemistry of this genus originated in the anti-
microbial activity of the volatile oil of L. scoparium, due
to triketones (1)–(4) (Perry et al., 1997; Porter and
Wilkins, 1998).
We previously chose to study L. recurvum Hook. f.
because its ethanol extract also showed anti-microbial
activity in addition to some anti-viral action (Mustafa et
al., 2003). This study established the presence of the
dihydrochalcones (5) and (6), flavone (7) (but not the
isomeric (8)) accompanied by the corresponding flava-
none (9), and an inseparable mixture of 2,5-dihydroxy-
6-methyl-7-methoxyflavanone (10) and its isomer 2,5-
dihydroxy-8-methyl-7-methoxyflavanone (11).
We extended our study to include the Australian
plant, L. polygalifolium subsp. polygalifolium Salisb.,
because the HPLC trace of its ethanol extract suggested
a similar compound profile. This ethanol extract also
showed anti-microbial, anti-viral and anti-fungal activ-
ity, and we had no more L. recurvum foliage to study
the 2,5-dihydroxyflavanones (10) and (11). Previous
reports described the volatile components of L. poly-
galifolium subsp. polygalifolium as a-pinene and the a, b
and g-eudesmols (Brophy and Goldsack, 1993; Brophy
Leptospermum J. R. et G. Forst. is a genus in the
Myrtaceae family consisting of 83 species that is widely
distributed throughout Australia (Wrigley and Fagg,
1993). One tropical species (L. amboinense) extends to
South-east Asia, and L. scoparium, a species common in
Tasmania, is widespread in New Zealand. Three species,
L. javanicum, L. parviflorum and L. recurvum, are en-
demic to South-east Asia, the latter being found only on
Mt. Kinabalu in the state of Sabah, Malaysia (Lee and
Lowry, 1980).
Chemical studies on this genus have concentrated on
volatile oils, with little work having been published on
non-volatiles. To the best of our knowledge, the flavo-
noids of only two species, L. scoparium (Haberlein and
Tschiersch, 1994a, b, 1998; Haberlein et al., 1994;
Mayer, 1990, 1993, 1996) and L. laevigatum (Wollen-
weber et al., 1996), have been examined. These species
yielded flavonoids characterised by C-methylation. In
the case of L. scoparium this feature was coupled with
* Corresponding author. Tel.: +64-34797925; fax: +64-34797906.
E-mail address: rweavers@alkali.otago.ac.nz (R.T. Weavers).
0031-9422/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.phytochem.2003.09.006

1286
K. Mustafa et al. / Phytochemistry 64 (2003) 1285–1293
mers (10) and (11) and triketones (1)–(3). Base
extraction of one chromatography fraction yielded fur-
ther samples of (1)–(3) and further chromatography on
the neutral fraction gave a mixture of the isomeric fla-
vones (7) and (8). Chloroform extraction of the residue
from the ethanol extract gave an extract that was parti-
tioned into neutral and acid fractions. The latter yielded
further samples of the triketone mixture (1)–(3). Com-
pounds isolated to this point were identified by com-
parison with authentic samples; by HPLC using co-
injection, and by comparison of 1H NMR spectra.
Chromatography of the neutrals gave fractions con-
taining further samples of isomeric mixtures (7)/(8) and
(10)/(11), but also a fraction from which oleanolic acid
(16) could be crystallised. Data for this compound mat-
ched those reported previously (Mahato and Kundu,
1994). The filtrate from this crystallisation, and also
another chromatography fraction, yielded a mixture of
four new isomeric compounds, the cis and trans isomers
of 2,3,5-trihydroxy-6-methyl-7-methoxyflavanone (12)
and (13) and those of its position isomer 2,3,5-trihy-
droxy-8-methyl-7-methoxyflavanone (14) and (15).
2,5-Dihydroxyflavanones (10) and (11) were first
reported from Friesodielsia enghiana (Annonaceae), as
an approximate 1:1 mixture (Fleischer et al., 1997).
Given the propensity for such cyclic hemiacetals to open
in solution, it seems likely that these two compounds
would inter-convert as in Fig. 2. This pathway also
provides a mechanism for racemisation. Therefore it
surprised us to find that the previous workers found
non-zero optical rotation ([ꢀ]2D1 +20.7ꢀ). Our samples of
(10) and (11), isolated from L. recurvum (Mustafa et al.,
2003), and from L. polygalifolium subsp. polygalifolium,
were also 1:1 mixtures as shown by paired peaks in
NMR spectra. These also had significant non-zero
optical rotations, but of opposite signs ([ꢀ]2D1 +24.2ꢀ
and [ꢀ]2D0 ꢁ15.9ꢀ respectively). Vapour diffusion crystal-
lisation (ethanol/water) on a sample of the 1:1 mixture
of (10) and (11) from L. polygalifolium subsp. poly-
galifolium gave a crystalline material, in a yield of
greater than 50% of the original mass. 1H NM R
experiments on a freshly prepared benzene-d6 solution
of this solid showed the presence of only one isomer
that proved to be the 6-methyl derivative (10). In the
NMR tube, equilibrium was established after about 390
min, giving approximately equal amounts of (10) and
(11). 2D NMR methods on this equilibrated mixture
et al., 2000). The only reports of non-terpenoid meta-
bolites stated that the leaf oils contain the triketones,
flavesone (1) and leptospermone (3) (Briggs et al., 1938,
1945; Penfold, 1921), although Brophy found no evi-
dence of these two compounds (Brophy et al., 2000).
In this paper, we report isolation of several flavonoids
from L. polygalifolium subsp. polygalifolium, including
the desired (10) and (11). We describe the first isolation
of hemiacetal (10) from its position isomer (11), along
with a study of its solution behaviour. In addition, we
report a more complex system of hemiacetal isomers,
some novel 2,3,5-trihydroxyflavanones.
1
enabled assignment of the H and 13C resonances for
both (10) and (11). Compound (10) showed a corre-
lation in the CIGAR-HMBC spectrum between the C-5
hydroxyl proton (ꢁ 12.65) and the oxygenated carbon
resonating at ꢁ 160.2 (C-5). The latter in turn showed
correlation to the methyl proton signal (ꢁ 2.24), thereby
establishing that the methyl group was at C-6. Further
verification was obtained by correlation between the
methyl proton signal and a further oxygenated carbon
2. Results and discussion
Extraction of the dried foliage of L. polygalifolium
subsp. polygalifolium with ethanol followed by a
sequence of chromatographic separations (Fig. 1) affor-
ded separate mixtures of 2,5-dihydroxyflavanone iso-

K. Mustafa et al. / Phytochemistry 64 (2003) 1285–1293
1287
Fig. 1. Separation Tree for L. polygalifolium subsp. polygalifolium showing masses (mg) and compounds identified from each fraction. ySamples
contaminated with uncharacterised materials.
Fig. 3. CIGAR correlations establishing the methyl group positions in
(10) and (11).
in the formation of more (10) yielding a higher than
50% return.
The average optical rotation of five equilibrated solu-
tions prepared from the crystallised sample in chloro-
form ([ꢀ]2D1=+8ꢀ {ꢂ3}) was significantly lower in
Fig. 2. Solution equilibria of 2-hydroxyflavanones (10) and (11).
magnitude than that of the parent mixture, but rotation
values at longer wavelengths were very substantial
peak at ꢁ 165.7 (C-7) that was further linked to the
methoxyl proton signal (ꢁ 3.24). Similar correlations
established that (11) was methyl substituted at C-8
20.9
405 nm
([ꢀ]
ꢁ324ꢀ {ꢂ4}). Because of the proposed solu-
tion behaviour (Fig. 2), the optical activity seems
incompatible with the structures (10) and (11), and may
be due to the presence of trace impurities. These would
necessarily have very large rotations as the sample of
(10) was analytically pure, and solutions showed only
one peak in an HPLC trace (detection at 206, 254, 280
and 325 nm) and only signals for (10) and (11) in the 1H
NMR spectrum. Optical activity in samples of similar 2-
hydroxyflavanones has been reported on at least three
other occasions (Fleischer et al., 1997; Lee et al., 1995;
Wang et al., 1999). Wang also isolated a single position
1
(Fig. 3). The signals in the H NMR spectrum of the
freshly prepared sample of the selectively crystallised
compound matched those assigned to (10). In more
acidic solvents such as CDCl3 equilibration between
(10) and (11) was attained much more rapidly (<5 min),
although equilibrium composition did not vary appre-
ciably over the solvents examined (benzene-d6, CDCl3
and acetone-d6). The results of the vapour diffusion
crystallisation indicated that, as the less soluble (10)
crystallised from the 1:1 mixture, equilibration resulted

1288
K. Mustafa et al. / Phytochemistry 64 (2003) 1285–1293
isomer from one of these equilibrating systems by
selective crystallisation, a 6-methoxy analogue of (10)
(Wang et al., 1999). The X-ray structure proved the
compound to be racemic.
unsubstituted B-phenyl ring, which was supported by
the fragment at m/z 77.0385 in the HR-EIMS. A fea-
ture of the spectrum of these new compounds, the
presence of singlets at ꢁ 4.65/4.59, had no analogue in
the spectrum of (10) and (11). Also lacking were the
3-CH2 signals of (10) and (11). Both features were
consistent with hydroxylation at C-3. Assignment of
individual peaks to a 6-methyl structure (12) and its 8-
methyl isomer (14) was achieved by 2D NMR (HSQC,
CIGAR-HMBC and NOESY). Two and three bond
1H–13C correlations (CIGAR-HMBC) confirmed that
the signals at ꢁ 4.65/4.59 belonged to H-3 for (12) and
(14) respectively, consistent with the presence of a C-3
hydroxyl group (Fig. 4). Only in isomer (12) were
correlations from the 5-OH proton signal (ꢁ 11.06)
observed. These linked to one oxygen bearing carbon
(ꢁ 159.4, C-5) and two others, a carbon with a peak at
ꢁ 106.9 (C-6), which also showed correlation to the
methyl proton system resonating at ꢁ 2.02 (8-Me), and
to a carbon signal at ꢁ 100.7, a chemical shift typical
of C-4a of flavanone systems. This suggested a 6-
methyl derivative. Correlations from the aromatic
proton singlet at ꢁ 6.13 (Fig. 4) were consistent with
this deduction and showed that the contributing pro-
ton was located at C-8. Correlation between the
methoxyl proton signal at ꢁ 3.86 and the carbon signal
at ꢁ 166.5 (C-7), which in turn correlated to the H-8
signal, established C-7 methoxylation. A similar set of
correlations (Fig. 4) established that (14) was methy-
lated at C-8 and had a methoxyl group at C-7.
Mixture (12)–(15) showed similarity to the hemiacetal
mixture (10) and (11) in that pairing of the dominant
1
peaks was observed in both the H and the 13C spectra
(Table 1). It showed one spot on silica TLC (cyclohex-
ane:EtOAc; 3:1), at lower RF (0.13) than that obtained
for the mixture of (10) and (11) (RF 0.28), consistent
with further hydroxylation. Reversed-phase HPLC on
C-18 showed a single peak of shorter retention time
(21.2 min) than that for the mixture of (10) and (11)
(26.4 min). This peak was present at low level in the
HPLC trace of the crude ethanol foliage extract. The
HR-EIMS gave a single molecular ion peak at m/z
316.0947, corresponding to C17H16O6 (one more oxygen
than (10) and (11)). UV analysis showed a strong
absorption at 294 nm and a shoulder at 316 nm, which
is typical of flavanones, (Markham and Mabry, 1975)
while the IR spectrum indicated hydroxyl (3385 cmꢁ1
)
and conjugated carbonyl (1636 cmꢁ1) functionality.
1
The H NMR spectrum in CDCl3 showed peaks con-
sistent with two major isomers (12) and (14) in
approximately equal proportions (Table 1), in particular
the hydrogen-bonded proton singlets (5-OH) at ꢁ 11.06/
11.02, the methoxyl singlets (7-OMe) at ꢁ 3.86/3.87, the
methyl singlets (6-Me/8-Me) at ꢁ 2.02 and the aromatic
proton singlets (H-8/H-6) at ꢁ 6.13/6.16 ppm. Further-
more, the aromatic multiplets at ꢁ 7.45–7.47 and ꢁ 7.76–
7.80, which integrated in a 3:2 ratio, suggested an
The next task was to assign the relative orientation of
1
the 2-OH and 3-OH groups. The H NMR spectrum of
our sample was obtained in acetone-d6 to compare with
similar compounds reported in the literature where
couplings had been reported between the 2-OH proton
and H-3 (Hauteville et al., 1979) (2-OH signals were not
observed in our spectra of the (12)–(15) mixture recor-
ded in CDCl3). The isomeric composition of the mixture
proved to be solvent dependent. In CDCl3 the major
isomers (12) and (14) dominated over two minor iso-
mers (13) and (15) to the extent of 97:3. In acetone-d6
the ratio was about 7:3.
Conformational searching of the cis-isomer of the 8-
methyl compound (14) generated several low energy
conformations. These all had the O–H bond at C-2 in
an anti relationship with the C–H bond at C-3 (Fig. 5).
This W planar arrangement would favour long-range
Table 1
1H NMR and 13C NMR data for (12) and (14) at 25 ꢀCa
Position
1H
(12)
–
13C
(14)
(12)
(14)
2
–
4.59, s
103.4
75.4
194.0
100.7
159.4
106.9
166.5
92.3
157.4
139.8
126.1
128.5
129.5
–
102.6
75.4
194.3
100.7
161.3
92.8
166.9
106.7
154.9
139.0
126.3
128.5
129.6
–
3
4.65, s
–
4
4a
–
–
–
5
–
–
6
–
6.16, s
–
7
–
8
8a
10
20,60
40
30,50
2-OH
3-OH
5-OH
6-Me
8-Me
7-OMe
6.13, s
–
–
–
–
–
7.76, m
7.47, m
7.45, m
n.d.
7.80, m
7.47, m
7.45, m
n.d.
n.d.
n.d.
11.02, s
–
–
–
11.06, s
2.02, s
–
–
6.9
–
–
7.5
2.02, s
3.87, s
–
56.0
3.86, s
56.1
a
Solvent CDCl3; 500 MHz (1H); 125 MHz
(
13C); n.d.=not
Fig. 4. Significant long-range (1H–13C) correlations (CIGAR) for
isomers (12) and (14).
detected.

K. Mustafa et al. / Phytochemistry 64 (2003) 1285–1293
1289
Fig. 5. Predicted most stable conformations of the cis- and trans-8-methyl-7-methoxy-2,3,5-trihydroxyflavanones, from molecular modelling,
showing the spatial relationship between the 2-OH proton and H-3.
coupling. Conformations calculated for the trans-
arrangement (15) had a gauche relationship between the
O–H bond at C-2 and the C–H bond at C-3, incompa-
tible with a substantial long-range coupling. Published
data for some synthetic 2,3-hydroxyflavanones (17) and
(18) (Hauteville et al., 1979) (Table 2) are consistent
with this conclusion. The fact that the major isomers
showed significant long-range couplings between H-3
and the C-2 OH proton (1.5 Hz, Table 2) suggests that
these both have the cis geometry as in (12) and (14).
The 2,3,5-trihydroxyflavanone mixture (12)–(15) gave
a small optical rotation of [ꢀ]D +4ꢀ. These compounds
have two chiral centres and like (10)/(11) they will
interchange C-2 stereochemistries readily (Fig. 6).
Although, C-3 is not implicated in the tautomeric equi-
libria responsible for this interchange it too may lose its
stereochemical integrity through enolisation in the ring-
opened b-diketone state (the absolute stereochemistry
has not been determined and the configuration shown at
C-3 is arbitrarily chosen). These compounds have so far
resisted attempts at separation, but it may be possible to
isolate one or more components by selective crystal-
lisation as was achieved for the mixture of (10) and
(11). The limited amount of material available from
this plant has precluded further investigation of this
possibility.
Table 2
Key 1H NMR signals of cis and trans-2,3-dihydroxyflavanonesa
(17)b
cis
(18)b
(12)/(14)
cis
(13)/(15)
trans
trans
(ꢃ70%)
(ꢃ30%)
(ꢃ70%)
(ꢃ30%)
2-OH 7.02 (d,2)
7.10 (s)
7.04/6.99 (d,1.5)
7.12/7.08 (s)
3-OH 5.23 (d,6.5)
H-3
5.90 (d,6.5) 5.13/5.11 (d,6.5)
4.80 (dd,2,6.5) 4.03 (d,6.5) 4.77/4.70 (dd,2.5,6.5) 4.00/3.97 (d,6.5)
5.88/5.85 (d,6.5)
Solvent acetone-d6; 500 MHz; ꢁ28 ꢀC; expressed as chemical shift in ppm
a
relative to TMS (multiplicity, coupling constant(s) in Hz).
b
Hauteville et al. (1979).
Fig. 6. Inter-conversion of hemiacetal isomers (12)–(15).
acid (16) is a common triterpene in plants and has been
reported once previously from a Leptospermum species,
from the bark of L. scoparium (Corbett and McDowall,
1958).
3. Conclusions
As in L. recurvum (Mustafa et al., 2003), the mixture
of hemiacetal 2-hydroxyflavanones (10) and (11) dom-
inates the flavonoid composition of the L. poly-
galifolium subsp. polygalifolium foliage. These
compounds are accompanied by the the isomeric flava-
nones (7) and (8). However, the dihydrochalcones found
in L. recurvum were not observed in this taxon. The fact
that solution equilibria are rapidly established in 2-
hydroxyflavanone systems such as (10)/(11) indicates
The presence of triketones (1) and (3) in this specimen
of L. polygalifolium subsp. polygalifolium foliage oil,
along with the closely related (2), agrees with early
results on this sub-species (Briggs et al., 1938, 1945;
Penfold, 1921), but contrasts with Brophy0s findings
(Brophy et al., 2000). It seems likely that this is another
example of infraspecific variation of triketone levels, as
found in L. scoparium (Perry et al., 1997). Oleanolic

1290
K. Mustafa et al. / Phytochemistry 64 (2003) 1285–1293
that reported optical rotations probably result from
residual levels of co-occurring metabolites with high
rotation values. Flavonoids with hydroxyl groups at
both C-2 and C-3 are rare in nature. Of the four exam-
ples known prior to this work, two are biflavonoids
(Arens et al., 1986). Compounds of this type have been
generated in the laboratory by enzymic transformations
of flavanols (Hosel and Barz, 1972).
Biological activities noted for the crude plant extract
were not concentrated in the flavonoid fractions isolated
and the anti-microbial activities were associated with
the triketones (1)–(3) (Perry et al., 1997; Porter and
Wilkins., 1998). No significant activities were noted for
the 2-hydroxyflavanones reported in this study.
was used for reversed-phase (RP) chromatography.
Melting points were determined on a hot bench Leica
AG melting point apparatus first calibrated with stan-
dard samples of known melting points. Optical rota-
tions were performed on
a
Perkin-Elmer 241
polarimeter using deuterium (Na, 589 nm) and mercury
(Hg 577, 546, 435, 405 nm) lamps with specified filters.
The instrument was first calibrated using cholesterol
(20 mg/ml). Biological assays were performed as
described previously (Lorimer et al., 1996). Molecular
mechanics calculations were performed with PCModel
version 7 incorporating p-calculations. Conformational
searches used the MMX force field, with the mixed
Monte Carlo coordinate movements/bond rotations
strategy (Saunders et al., 1990) for the generation of
initial structures. The default cut-off criteria were
employed.
4. Experimental
4.1. General experimental procedures
4.2. Plant material
UV and IR spectra were recorded on a Jasco 7800
UV–vis spectrometer and a Perkin-Elmer 1600 FTIR
instrument respectively. NMR spectra were recorded on
a Varian INOVA-500 spectrometer operating at 500
L. polygalifolium subsp. polygalifolium was collected
from Lincoln Landcare Research Garden, New Zealand
in November 1997. A voucher specimen (code 971108-
23) has been deposited at the Plant Extract Research
Unit, University of Otago, Dunedin, New Zealand.
1
MHz for H and 125 MHz for 13C. NMR spectra were
recorded on ca 0.075 Msolutions in (CD ) CO (refer-
3 2
enced to solvent, ꢁ 2.05 for 1H and ꢁ 29.9 for 13C), C6D6
(referenced to solvent, ꢁ 7.16 for 1H and ꢁ 128.6 for 13C)
4.3. Extraction and isolation of flavonoids of
L. polygalifolium subsp. polygalifolium
1
or CDCl3 (referenced to solvent, ꢁ 7.25 for H and ꢁ
ꢀ
77.0 for 13C) at 25 C. Spectra were assigned with the
Details of the separation process are summarised in
Fig. 1.
aid of double-quantum filtered COSY (1H–1H corre-
lations), HSQC (one bond 1H–13C correlations), and
HMBC-CIGAR experiments (two and three bond
1H–13C correlations) (Hadden et al., 2000). Analytical
HPLC was performed on HPLC equipment consisting
of a Waters 717 auto-sampler, 600 pump and controller,
and a 490E UV programmable multi-wavelength detec-
tor controlled by Millennium software. Samples were
analysed using an RP-18 analytical column (Merck 100
RP-18, LichroCART 250ꢄ4 mm, 5 mm) fitted with a
guard column (Merck 100 RP-18, LichroCART 4ꢄ4
mm, 5 mm). The mobile phase flow rate was 1.0 ml/min,
with UV detection at 206, 254, 280 and 325 nm. The
initial solvent mix was MeCN–H2O (20:80) (both sol-
vents containing 0.1% H3PO4), with a linear ramp up to
100% MeCN over 40 min, a return to the initial mix
over 10 min, and 10 min equilibration before the next
analysis. Semi-preparative HPLC used a Waters 717
auto-sampler, 600 pump and controller, a 2487 l
absorbance detector and fraction collector II. Samples
were separated using an RP-18 semi-preparative column
(Merck 100 RP-18, LichroCART 250ꢄ10 mm, 10 mm)
fitted with a guard column (Merck 100 RP-18, Lichro-
CART 4ꢄ4 mm, 5 mm). Silica gel 60, 200–400 mesh, 40–
63 mm (Merck) was used in the column chromatography
and octadecyl-functionalised silica gel (C-18, Aldrich)
4.3.1. EtOH extract
Foliage (109 g) was dried, ground and extracted with
EtOH (650 ml, then 350 ml). Removal of the solvent
from the filtered extracts gave a green gum (14.9 g). In
disk diffusion assays this crude extract showed activity
against Bacillus subtilis, Candida albicans and Tricho-
phyton mentagrophytes at 600 mg/disk. Fractionation by
reversed-phase column chromatography (50 g, C-18)
was carried out with gradient elution from H2O through
CH3CN to CHCl3. The anti-microbial active fractions
were pooled into five groups: E1 (H2O–CH3CN; 1:3),
E2 (H2O–CH3CN; 1:3 ), E3 (H2O–CH3CN; 1:3), E4
(H2O–CH3CN; 1:9) and E5 (H2O–CH3CN; 1:9 to 0:1)
based on their UV-active spots on TLC silica (cyclo-
hexane–EtOAc; 3:1). Compositions of fractions were
1
determined by H NMR and by HPLC. Fraction E1
(256 mg) contained only the isomeric 2,5-dihydroxy-
flavanones (10) and (11). Fraction E2 (202 mg) also
contained (10) and (11) along with triketone (1). Frac-
tion E3 (202 mg) consisted of compounds (10) and (11)
and triketones (1)–(3). Fraction E4 (183 mg) was further
fractionated on a silica gel column (5 g) using a cyclo-
hexane–EtOAc gradient to give triketones (1)–(3) (59
mg, cyclohexane–EtOAc; 19:1), a mixture of com-

K. Mustafa et al. / Phytochemistry 64 (2003) 1285–1293
1291
pounds (7), (8), (10) and (11) (37 mg, cyclohexane–
4.3.4. Flavonoids (7), (8), (10) and (11)
(CAS No. (7)-55969-57-8; (8)-14004-48-9;
(10)-186906-54-7; (11)-186906-53-6)
Identity was established by comparison by HPLC and
NMR spectroscopy with samples isolated from L.
recurvum (Mustafa et al., 2003).
EtOAc; 18:2) and a fraction of the isomeric (10) and
(11) (8 mg, cyclohexane–EtOAc; 18:2, 15:1). Fraction
E5 (446 mg) was dissolved in Et2O and extracted with
satd. NaHCO3 (3ꢄ10 ml). The Et2O phase was parti-
tioned with 5% Na2CO3 (3ꢄ10 ml) dried (MgSO4) and
evaporated to give an oil (257 mg) that was fractionated
on a silica gel column (6 g, cyclohexane–EtOAc; 19:1)
yielding a mixture of isomers (7) and (8) (20 mg) and
another of isomers (10) and (11) (15 mg). The Na2CO3
extracts from this work-up were acidified and extracted
with Et2O. The Et2O extracts were dried (MgSO4) and
evaporated to give a mixture of triketones (1)–(3) (96
mg). The NaHCO3 extracts were treated similarly to
yield triketones (1)–(3) (15 mg).
4.3.5. cis and trans-2,3,5-Trihydroxy-6-methyl-7-
methoxyflavanone (12) and (13) and cis and trans-2,3,5-
trihydroxy-8-methyl-7-methoxyflavanone (14) and (15)
The mixture of (12) (cis-2,3,5-trihydroxy-7-methoxy-6-
methyl-2-phenylchroman-4-one), (13) (trans-2,3,5-trihy-
droxy-7-methoxy-6-methyl-2-phenylchroman-4-one), (14)
(cis-2,3,5-trihydroxy-7-methoxy-8-methyl-2-phenylchro-
man-4-one) and (15) (trans-2,3,5-trihydroxy-7-methoxy-
8-methyl-2-phenylchroman-4-one) was obtained as a
yellow oil; [ꢀ]1D9.7 +4.1ꢀ, [ꢀ]
ꢁ7.5ꢀ, [ꢀ]
20.0
20.3
546 nm
4.3.2. CHCl3 extract
577 nm
20.6
435 nm
20.9
405 nm
The residue from the EtOH extraction was extracted
with CHCl3 (265 ml). The extract was dried (MgSO4),
dissolved in Et2O and extracted with satd. NaHCO3 and
Na2CO3 as above. The NaHCO3 and Na2CO3 extracts
gave mixtures of triketones (1)–(3) (170 and 191 mg
respectively). The neutral fraction was fractionated on a
silica gel column (20 g) to give six flavonoid containing
fractions, C1 (cyclohexane–EtOAc; 19:1), C2 (cyclohex-
ane–EtOAc; 19:1), C3 (cyclohexane–EtOAc; 19:1), C4
(cyclohexane–EtOAc; 18:2), C5 (cyclohexane–EtOAc;
18:2) and C6 (cyclohexane–EtOAc; 17:3 and 16:4).
Fraction C1 (23 mg) contained the isomeric flavones (7)
and (8) along with unidentified terpene derivatives.
Fraction C2 (87 mg) contained isomers (10) and (11)
contaminated with fatty acid derivatives. Fraction C3
(251 mg) consisted of (10) and (11), while fraction C4
(16 mg) comprised a mixture containing the isomers
(12)–(15). Fraction C5 (116 mg) consisted of a mixture
of hemiacetals (12)–(15) and oleanolic acid (16). The
composition of fraction C6 (154 mg) was similar to that
of fraction C5. Further chromatography on C5 failed to
separate the triterpene from the flavonoids, but careful
addition of H2O to a soln. of the mixture (60 mg) in a
1:1 mixture of MeOH and CH3CN (1.0 ml) crystallised
out the triterpene, (16) as white fluffy material (6.5 mg).
A sub-sample of the filtrate (22 mg) was further purified
by preparative RP HPLC (C-18) (CH3CN:MeOH; 1:1)
to give the isomeric flavonoids (12)–(15) (3 mg). A
sample of C6 (154 mg) was passed through a short C-
18 column which removed the major component (16),
and the effluent was also subjected to preparative
HPLC to give a further portion of the mixture of (12)–
(15) (2 mg).
ꢁ24.0ꢀ, [ꢀ]
ꢁ136.4ꢀ, [ꢀ]
ꢁ156.5ꢀ (CHCl3; c
0.20); UV lmax MeOH nm (log "): 294 (3.83), 316 (3.18)
(shoulder); IR ꢂmax (film) cmꢁ1: 3385 br, 3015, 2923,
1636, 1585, 1451, 1143; 1H NMR (acetone-d6), (12)/(14)
(cis): 11.54/11.49 (2ꢄ1H, s, 5-OH), 7.81 (4H, m, 20-H
and 60-H), 7.41–7.43 (6H, m, 30-H, 40-H and 50-H), 6.56/
6.59 (2ꢄ1H, d, J=1.5 Hz, 2-OH), 6.20 (2H, s, H-6 and
H-8), 4.74/4.68 (2ꢄ1H, br. d, J=4.5 Hz, H-3), 4.591/
4.61 (2ꢄ1H, d, J=6.0 Hz, 3-OH), 3.90 (6H, s, 7-OMe),
1.96 (6H, s, 6-Me and 8-Me); (13)/(15) (trans): 11.80/
11.78 (2ꢄ1H, s, 5-OH), 7.76 (4H, m, 20-H and 60-H),
7.41–7.43 (6H, m, 30-H, 40-H and 50-H), 6.68/6.65
(2ꢄ1H, s, 2-OH), 6.24/6.18 (2ꢄ1H, s, H-6 and H-8),
4.07/4.05 (2ꢄ1H, d, J=6.5 Hz, H-3), 5.38/5.35 (2ꢄ1H,
d, J=6.5 Hz, 3-OH), 3.92 (6H, s, 7-OMe), 2.03/1.96
(2ꢄ3H, s, 6-Me and 8-Me); 13C NMR (acetone-d6), (12)/
(14) (cis): 197.8/197.6 (C-4), 167.1/166.8, (C-7), 162.3/
160.2 (C-5), 159.1/156.7 (C-8a), 141.8/141.5 (C-10),
129.6/129.5 (C-30/C-50), 128.7/128.6 (C-40), 127.8 (C-20/
C-60), 106.3/106.1 (C-6/C-8), 105.4/104.8 (C-2), 102.3/
102.2 (C-4a), 93.1/93.0 (C-6 /C-8), 77.4 (C-3), 56.6 (7-
OMe), 7.9/7.1 (8-Me/6-Me); (13)/(15) (trans): 197.5/
197.3 (C-4), 167.0/166.7 (C-4), 163.7/161.4 (C-5), 159.2/
156.6 (C-8a), 141.4/141.0 (C-10), 129.6/129.5 (C-30/C-50),
128.8/128.7 (C-40), 127.8/127.7 (C-20/C-60), 105.9/105.8
(C-6/C-8), 103.7/103.4 (C-2), 101.6/101.5 (C-4a), 93.0
(C-6 /C-8), 75.4 /75.2 (C-3), 56.6 (7-OMe), 8.0 (8-Me/6-
1
Me); H and 13C NM R (CDCl) for (12) and (14), see
3
Table 1. HR-EIM S:m/z (rel. int.): 316.0947 [M+] (14)
(calc. for C17H16O6 316.0947), 298.0849 (20), 181.0509
(100), 105.0347 (81), 77.0385 (39).
4.3.6. Oleanolic acid (16) (CAS No. 508-02-1)
Compound (16) (Mahato and Kundu, 1994) was
ꢀ
4.3.3. Triketones (1)–(3) (CAS No. (1)-22595-45-5;
(2)-5009-05-2; (3)-567-75-9)
Identity was established by comparison with authentic
samples by HPLC and NMR spectroscopy (van Klink
et al., 1999).
obtaꢀined as a white fluffy solid; mp 288 C (Lit. 295–
297 C (Corbett and McDowall, 1958)) IR (KBr) ꢂmax
3500–3000, 3412, 2924, 2851, 1683, 1636, 1614, 1457
1
cmꢁ1; ESI (-ve): m/z 455 [M+–H]; H NMR and 13C
NMR as reported by Mahato and Kundu (1994).

1292
K. Mustafa et al. / Phytochemistry 64 (2003) 1285–1293
4.4. 2,5-Dihydroxy-6-methyl-7-methoxyflavanone (10)
Thomas and W. Redmond for assistance with NMR
experiments, B. Clark for MS analysis; G. Ellis for bio-
logical assays, and M. Dick for microanalysis. This
research was supported in part by the New Zealand
Foundation for Research, Science and Technology and
by the New Zealand Official Development Assistance
(NZODA)/New Zealand Agency for International
Development (NZAID), Ministry of Foreign Affairs
and Trade for a PhD scholarship.
A soln. of a mixture containing compounds (10) and
(11) (263 mg) in EtOH (2 ml) in a 4 ml vial was placed
in a 16 ml capped-vial containing H2O (2.5 ml). After
cooling in a refrigerator for a week, the crystals were
collected by removal of supernatant liquid and air-dry-
ing to give 2,5-dihydroxy-6-methyl-7-methoxyflavanone
(2,5-dihydroxy-7-methoxy-6-methyl-2-phenylchroman-4-
one) (10) as a milky-yellow crystalline powder (92 mg).
The supernatant liquor was stored in the refrigerator for
another week to give a further crop of (10) (50 mg). Mp
143ꢀ; IR ꢂmax (KBr) cmꢁ1: 3412, 3274, 3083, 2992, 2923,
1645, 1603, 1573, 1508, 1497, 1447, 1325, 1295, 1203,
1154, 1120, 769, 697; HR-EIMS: m/z (rel. int.): 300.0994
[M+] (45) (calc. for C17H16O5 300.0998), 282.0834 (31),
181.0396 (25), 180.0333 (28), 154.0545 (33), 152.0412
(21), 105.0296 (100), 77.0365 (42); found: C, 67.9; H,
5.1, C17H16O5 requires: C, 68.0; H, 5.4%. The remain-
ing data were obtained on equilibrated mixtures of (10)
and (11) derived from pure (10): optical rotations
References
Arens, H., Ulbrich, B., Fischer, H., Parnham, M.J., Romer, A., 1986.
Novel antiflammatory flavonoids from Podophyllum versipelle cell
culture. Planta Medica 6, 468–473.
Briggs, L.H., Hassall, C.H., Short, W.F., 1945. Leptospermum. Part II.
Journal of the Chemical Society 706–709.
Briggs, L.H., Penfold, A.R., Short, W.F., 1938. Leptospermum. Part I.
Journal of the Chemical Society 1193–1195.
Brophy, J.J., Goldsack, R.J., 1993. The essential oil of the indigenous
Myrtaceae of Lord Howe Island. Flavour and Fragrance Journal 8,
153–158.
(average of 5 measurements) [ꢀ]2D4.2 8 (ꢂ3)ꢀ, [ꢀ]
25.1
Brophy, J.J., Goldsack, R.J., Bean, A.R., Forster, P.I., Lepschi, B.J.,
2000. Leaf essential oils of the genus Leptospermum (Myrtaceae) in
Eastern Australia. Part 6. Leptospermum polygalifolium and Allies.
Flavour and Fragrance Journal 15, 271–277.
577 nm
ꢀ
26.2
ꢀ
25.5
546 nm
25.9
435 nm
ꢁ 14 (ꢂ4) , [ꢀ]
ꢁ45 (ꢂ4) , [ꢀ]
ꢁ269
ꢀ
(ꢂ4) , [ꢀ]
ꢁ324 (ꢂ4)ꢀ (CHCl3; c 0.10); UV lmax
405 nm
MeOH nm (log "): 290 (4.14), 335 (3.55) (shoulder);
1H and 13C NMR (CDCl3) as reported by Mustafa et
Corbett, R.E., McDowall, M.A., 1958. Extractives from the New
Zealand Myrtaceae. Part III. Triterpene acids from the bark of Lep-
tospermum scoparium. Journal of the Chemical Society 3715–3716.
Fleischer, T.C., Waigh, R.D., Waterman, P.G., 1997. Bisabolene ses-
quiterpenes and flavonoids from Friesodielsia enghiana. Phy-
tochemistry 44, 315–318.
al. (2003): 1H NM R (CD6) ꢁ 12.65 (1H, s, 5-OH),
6
7.35 (2H, m, H-20/H-60), 7.09 (3H, m, H-30/H-40/H-50),
5.98 (1H, s, H-8), 3.24 (3H, s, 7-OMe), 2.24 (3H, s, 6-
Me); 13C NM R (CD6): ꢁ 194.2 (C-4), 165.7 (C-7),
6
Haberlein, H., Tschiersch, K.-P., 1994a. 2,5-Dihydroxy-7-methoxy-
6,8-dimethylflavan-3-one a novel flavonoid from Leptospermum
scoparium: in vitro affinity to the benzodiazepine binding site of the
GABAA receptor-chloride channel complex. Pharmazie 49, 860.
Haberlein, H., Tschiersch, K.-P., 1998. On the occurrence of methy-
lated and methoxylated flavonoids in Leptospermum scoparium.
Biochemical Systematics and Ecology 26, 97–103.
160.2 (C-5), 157.7 (C-8a), 141.9 (C-10), 129.2 (C-40),
128.8 (C-30/C-50), 125.0 (C-20/C-60), 106.8 (C-6), 102.3
(C-4a), 101.1 (C-2), 91.6 (C-8), 55.8 (7-OMe), 48.4 (C-
3), 6.9 (6-Me).
4.5. NMR study of equilibration of (10) and (11)
Haberlein, H., Tschiersch, K.-P., 1994b. Triterpenoids and flavonoids
from Leptospermum scoparium. Phytochemistry 35, 765–768.
Haberlein, H., Tschiersch, K.-P., Schafer, H.L., 1994. Flavonoids
from Leptospermum scoparium with affinity to the benzodiazepine
receptor characterized by structure activity relationships and in vivo
studies of a plant extract. Pharmazie 49, 912–922.
The pre-acquisition time delay was arrayed (0, 60
sꢄ5, 120 sꢄ5, 180 sꢄ5, 240 sꢄ5, 300 sꢄ5, 360 sꢄ5, 420
sꢄ5, 480 sꢄ5, 540 sꢄ5, 600 sꢄ5, 660 sꢄ5) to create a set
of time intervals for data collection. The spectrometer
was shimmed on a blank sample of C6D6 (0.75 ml)
immediately prior to conducting the experiment. C6D6
(0.75 ml) was added to a sample of 2,5-dihydroxy-6-
methyl-7-methoxyflavanone (10) (1.0 mg, 3.4 mmol) in a
5 mm NMR tube that was immediately inserted into the
NMR probe and the data collection commenced. In
total, 56 experiments were run over a period of 20 h.
Equilibrium was reached after ca 390 min when an
approximately 1:1 mixture of (10) and (11) was obtained.
Hadden, C.E., Martin, G.E., Krishnamurthy, V.V., 2000. Constant
time Inverse-detection Gradient Accordion Rescaled Heteronuclear
Multiple Bond Correlation Spectroscopy: CIGAR-HMBC. Mag-
netic Resonance in Chemistry 38, 143–147.
Hauteville, M., Chadenson, M., Chopin, J., 1979. Un nouveau type de
flavonoides naturels: les dihydroxy-2,5 flavanones. VI. Synthese et
etude structurale de trihydroxy-2,3,5 flavanones. Bulletin de la
Societe Chimique de France II, 125–131.
Hosel, W., Barz, W., 1972. Enzymatic transformation of flavonols
with a cell-free preparation from Cicer arietinum L. Biochimica et
Biophysica Acta 261, 294–303.
Lee, D.W., Lowry, J.B., 1980. Plant speciation on tropical mountains:
Leptospermum (Myrtaceae) on Mount Kinabalu, Borneo. Biological
Journal of the Linnean Society 80, 223–224.
Lee, S.-S., Tsai, F.-Y., Chen, I.-S., 1995. Chemical constituents from
Berchemia formosana. Journal of the Chinese Chemical Society 42,
101–105.
Acknowledgements
Thanks to W. Harris and J. van Klink for plant
material, D. Orlovich for taxonomic identification, M.
Lorimer, S.D., Barns, G., Evans, A.C., Foster, L.M., May, B.C.H.,
Perry, N.B., Tangney, R.S., 1996. Cytotoxic and antimicrobial

K. Mustafa et al. / Phytochemistry 64 (2003) 1285–1293
1293
activity of plants from New Zealand’s Subantarctic Islands. Phyto-
medicine 2, 327–333.
Mahato, S.B., Kundu, A.P., 1994. 13C NMR spectra of pentacyclic
triterpenoids—a compilation and some salient features. Phy-
tochemistry 37, 1517–1575.
M.H., McGimpsey, J.A., Smallfield, B.M., Anderson, R.E., 1997.
Essential oils from New Zealand manuka and kanuka: chemotax-
onomy of Leptospermum. Phytochemistry 44, 1485–1494.
Porter, N.G., Wilkins, A.L., 1998. Chemical, physical and anti-
microbial properties of essential oils of Leptospermum scoparium
and Kunzea ericoides. Phytochemistry 50, 407–415.
Markham, K.R., Mabry, T.J., 1975. Ultraviolet-visible and proton
magnetic resonance spectroscopy of flavonoids. In: Harborne, J.B.,
Mabry, T.J., Mabry, H. (Eds.), The Flavonoids. Chapman and
Hall, London, pp. 50–77.
Saunders, M., Houk, K.N., Wu, Y.D., Still, W.C., Lipton, M., Chang,
G., Guida, W.C., 1990. Conformations of cycloheptadecane. A
comparison of methods for conformational searching. Journal of
the American Chemical Society 112, 1419–1427.
Mayer, R., 1993. A b-hydroxychalcone from Leptospermum scopar-
ium. Planta Medica 59, 269–271.
van Klink, J.W., Brophy, J.J., Perry, N.B., Weavers, R.T., 1999. b-
Triketones from Myrtaceae: isoleptospermone from Leptospermum
scoparium and papuanone from Corymbia dallachiana. Journal of
Natural Products 62, 487–489.
Mayer, R., 1990. Flavonoids from Leptospermum scoparium. Phy-
tochemistry 29, 1340–1342.
Mayer, R., 1996. Three lupane derivatives from Leptospermum sco-
parium. Archiv der Pharmazie 329, 447–450.
Wang, Q., Terreaux, C., Marston, A., Tan, R.X., Stoeckli-Evans, H.,
Hostettmann, K., 1999. A new 2-hydroxyflavanone from Mosla
soochouensis. Planta Medica 65, 729–731.
Mustafa, K., Kjaergaard, H.G., Perry, N.B., Weavers, R.T., 2003.
Hydrogen-bonded rotamers of 20,40,60-trihydroxy-30-formyldihydro-
chalcone, an intermediate in the synthesis of a dihydrochalcone from
Leptospermone recurvum. Tetrahedron 59, 6113–6120.
Penfold, A.R., 1921. On the occurrence of a new phenol in the essen-
tial oils of the ‘‘Leptospermum’’. Journal of the Proceedings of the
Royal Society of New South Wales 55, 49–51.
Wollenweber, E., Mann, K., Roitman, J.N., 1996. C-Methyl flavones
from the leaf wax of Leptospermum laevigatum (Myrtaceae). Zeits-
chrift fur Naturforschung C: Journal of Biosciences 51, 8–10.
Wrigley, J.W., Fagg, M., 1993. Bottlebrushes, Paperbacks and Tea
Trees and All Other Plants in the Leptospermum Alliance. Angus
and Robertson, Christchurch.
Perry, N.B., Brennan, N.J., van Klink, J.W., Harris, W., Douglas,