Isoprenyl phenyl ethers from liverworts of the genus Trichocolea: Cytotoxic activity, structural corrections, and synthesis

Article abstract of DOI:10.1021/np9603378

The main cytotoxic component in New Zealand collections of the liverwort Trichocolea mollissima was identified as methyl 4-[(5-oxogeranyl)oxy]-3- methoxybenzoate, a structure that has not been reported previously. Two double-bond isomers of this geranyl e

Full text of DOI:10.1021/np9603378

J . Nat. Prod. 1996, 59, 729-733
729
Isop r en yl P h en yl Eth er s fr om Liver w or ts of th e Gen u s Tr ich ocolea : Cytotoxic
Activity, Str u ctu r a l Cor r ection s, a n d Syn th esis
Nigel B. Perry,* Lysa M. Foster, Stephen D. Lorimer, Barnaby C. H. May, and Rex T. Weavers
Plant Extracts Research Unit, New Zealand Institute for Crop & Food Research Limited, Department of Chemistry,
University of Otago, Box 56, Dunedin, New Zealand
Masao Toyota, Eri Nakaishi, and Yoshinori Asakawa
Faculty of Pharmaceutical Sciences, Tokushima Bunri University, Yamashiro-cho, Tokushima 770, J apan
The main cytotoxic component in New Zealand collections of the liverwort Trichocolea
mollissima was identified as methyl 4-[(5-oxogeranyl)oxy]-3-methoxybenzoate, a structure that
has not been reported previously. Two double-bond isomers of this geranyl ether were present
at lower levels. Reinvestigation of the benzoates from J apanese collections of Trichocolea
tomentella led to the identification of four geranyl ethers (including two of the three compounds
identified from T. mollissima), which had previously been assigned incorrect geranyl ester
structures. One compound, previously reported as a 3,3-dimethylallyl ester, could not be re-
isolated from T. tomentella, but was found in a New Zealand collection of Trichocolea lanata.
It was shown to be a 3,3-dimethylallyl ether by synthesis from methyl vanillate. Several of
these compounds were active in cytotoxic and antifungal assays.
Liverworts have yielded many new compounds, mainly
terpenoids or lipophilic aromatic compounds, a good
proportion of which have been found to be bioactive.1
Recently we began a screening program searching for
new bioactive natural products from the rich liverwort
flora of New Zealand, as lead compounds for new
pharmaceuticals or agrochemicals.2 Early in our liver-
wort research in Tokushima, we discovered from a
J apanese collection of Trichocolea tomentella (Ehrh.)
Dum. (family Trichocoleaceae) some compounds with
the rare combination of both terpenoid and aromatic
portions.3,4 These were identified as isoprenyl esters
1
1-4 based on IR, UV, H NMR, and MS data (but not
13C NMR spectra), and on degradative reactions. Pathak
and Khanna synthesized compound 1 (tomentellin) by
reacting 5-oxogeranyl bromide with veratric acid (5), but
they were not able to make a direct comparison of
synthetic 1 with the natural product.5 We later reported
more isoprenyl benzoates from J apanese and European
collections of T. tomentella, including geranyl ester (6).6
Some of these compounds were also found in a collection
of Trichocolea pluma Mont. from Malaysia, and 3,4-
dimethoxybenzoates were proposed as significant chemi-
cal markers of the Trichocoleaceae.7 Becker reported
that compounds 1 and 2 were major products of cultured
T. tomentella.8
Renewed interest in Trichocolea was sparked by the
re-investigation of the aromatic compounds from J apa-
nese collections of T. tomentella, for which the corrected
structures 7, 9, 10, and 11 are now reported. The
compound previously reported as 4 could not be re-
isolated from T. tomentella, but it was isolated from a
New Zealand collection of T. lanata. Its correct struc-
ture has now been confirmed as 12 by synthesis from
methyl vanillate (13).
cytotoxic activity of crude extracts from New Zealand
collections of T. mollissima, a species whose chemistry
has not been reported. This foliose liverwort grows
throughout New Zealand in rainforest and beech forest,
hanging from old tree trunks or growing erect on the
forest floor.9 Four other Trichocolea species have been
recorded from New Zealand, of which T. lanata (Hook.)
Nees and T. hatcheri Hodgs. are also common. We now
report that the main cytotoxic component in New
Zealand collections of T. mollissima is the monoterpene
phenyl ether (7), and that closely related compounds 8
and 9 occur at lower levels. This result prompted the
Resu lts a n d Discu ssion
Crude extracts of various New Zealand collections of
T. mollissima consistently showed cytotoxic effects
against monkey kidney (BSC) cells and antifungal
activity against the dermatophyte Trichophyton men-
tagrophytes. We used cytotoxicity against BSC cells to
* To whom correspondence should be addressed. Phone: +64-3-
4798354. FAX: +64-3-4798543. E-mail: perryn@alkali.otago.ac.nz.
X Abstract published in Advance ACS Abstracts, J uly 1, 1996.
S0163-3864(96)00337-0 CCC: $12.00
© 1996 American Chemical Society and American Society of Pharmacognosy

730 J ournal of Natural Products, 1996, Vol. 59, No. 8
Perry et al.
Ta ble 1. 1H NMR Data for Compounds from Trichocolea Speciesa
signal
7b
8b
9c
10c
11b
12c
1
2
4
4.71(d, 6)
5.60(tm, 6, 1)
3.12(s)
4.20(t, 7)
2.67(t, 7)
6.13(m, 1)
4.30(t, 7)
3.11(t, 7)
6.16(br s)
4.68(d, 7)
5.50(br t, 7)
2.10(m)
4.69(d, 7)
5.56(t, 7)
3.15(s)
4.65(d, 7)
5.52(br t, 7)
1.79(s)
5
2.10(m)
1.75(s)
6
8
9
10
2′
5′
6′
6.07(m, 1)
1.85(d, 1)
2.13(d, 1)
1.75(d, 1)
7.53(d, 2)
6.86(d, 9)
7.63(dd, 9, 2)
3.88(s)
6.06(m, 1)
1.88(d, 1)
2.16(d, 1)
2.23(d, 1)
7.54(d, 2)
6.80(d, 8)
7.65(dd, 8, 2)
3.885(s)
6.07(br s)
1.89(d, 1)
2.17(d, 1)
2.04(d, 1)
7.53(d, 2)
7.06(d, 8)
7.67(dd, 8, 2)
3.89(s)
5.07(m)
1.66(s)
1.59(s)
1.74(s)
7.54(d, 2)
6.88(d, 9)
7.65(dd, 9, 2)
3.89(s)
6.08(br s)
1.86(br s)
2.14(br s)
1.76(br s)
7.56(m)
6.85(d, 9)
7.56(m)
3.85(s)
7.55(d, 2)
6.89(d, 9)
7.66(dd, 9, 2)
3.90(s)
7′-OMe
3′-OMe
3.91(s)
3.895(s)
3.90(s)
3.92(s)
3.92(s)
a
b
In CDCl3, δ in ppm (J in Hz). 300 MHz spectrometer. c 400 MHz spectrometer.
Ta ble 2. 13C NMR Data for Compounds from Trichocolea
Speciesa
signal
7b,c
8b
9c,d
10c,d
11b,c
12c,d
1
2
3
4
5
6
7
8
65.6
123.8
135.5
55.0
197.9
122.7
156.6
27.7
66.9
40.2
67.9
33.6
65.9
119.1
141.3
39.5
65.6
123.0
136.6
54.7
197.7
122.7
157.0
27.7
65.8
119.3
138.4
25.8
N.O.e
127.6f
191.4
126.1f
N.O.
27.8
154.5
127.6
190.7
125.9
155.3
27.8
26.2
18.3
F igu r e 1. Important NMR correlations establishing the
structure of 7. T, selected NOE interactions; s, selected
HMBC correlations.
123.7
131.8
25.6
9
20.8
20.6
20.6
17.7
20.8
10
1′
2′
3′
4′
5′
6′
7′
17.1
19.5
27.0
16.7
17.2
(Figure 1). The critical features that distinguish these
structures are the linking of the 5-oxogeranyl group to
the aromatic ring via an ether linkage instead of an
ester and the replacement of a methyl ether with a
methyl ester. Unequivocal evidence of these features
was obtained from the HMBC experiment, in which
correlations were observed between the ester carbonyl
(δ 166.8) and one methoxyl proton signal (δ 3.88, 3H)
and between a quaternary oxygenated aromatic signal
(δ 152.0) and the H2-1 protons (δ 4.71, d, J ) 6 Hz) of
the geranyl group (Figure 1). A further NOE interaction
between the geranyl H-10 and H-1 signals showed that
the 2,3 double bond has E stereochemistry (Figure 1).
We could find no references to this monoterpene phenyl
ether (7) in the literature.
122.6
112.2
148.9
152.0
111.7
123.4
166.8
52.0
123.0
112.6
152.1g
152.3g
111.8
123.4
166.8
52.0
122.4
112.2
148.8
152.4
111.6
123.6
167.0
51.9
122.4
112.1
148.9
152.3
111.7
123.4
167.0
51.9
123.3
115.6f
145.4
149.5
111.0
122.6f
166.7
51.9
122.5
112.1
148.9
152.3
111.6
123.4
167.0
52.0
7′-OMe
3′-OMe
56.0
56.1
56.0
56.0
56.0
a
b
In CDCl3, δ in ppm. 75 MHz spectrometer. c Assignments
d
confirmed by HMQC and HMBC experiments. 100 MHz spec-
trometer. e Not observed. Assignments interchangeable within
f,g
columns.
direct the isolation of the main cytotoxic component.
Reversed-phase flash chromatography on a crude ex-
tract of T. mollissima gave most of the cytotoxic activity
in a fraction eluted with CH3OH-H2O 9:1. This fraction
contained one main UV-active compound by TLC. This
compound, which was responsible for the cytotoxic
activity of this fraction, was purified by Si gel flash
chromatography.
Another collection of T. mollissima was extracted to
give more of compound 7 for biological assays. During
this isolation, an additional TLC spot was noticed at
higher Rf than 7 on Si gel, but with a similar color
reaction to that of 7. The compound responsible for this
spot and another with a very similar Rf on Si gel to 7
were purified by preparative reversed-phase HPLC and
shown to be isomeric with 7 by MS. Both of the new
HRMS showed a molecular ion at 332.1624 daltons,
corresponding to the formula C19H24O5. The H NMR
1
1
spectrum (Table 1) and UV and IR spectra were similar
to those reported for 1.4 The 13C NMR spectrum (Table
2) of our compound showed signals appropriate for a
trisubstituted aromatic ring, a 5-oxogeranyl group, an
ester carbonyl, and two methoxyl groups, as expected
for 1. Because the 13C NMR spectrum of 1 had not been
compounds had H NMR (Table 1), UV, and IR spectra
similar to those reported for a compound from J apanese
T. tomentella, which was assigned monoterpene ester
structure (2), with a cross-conjugated dienone in the
monoterpene portion.3,4 However, the 13C NMR spectra
(Table 2) showed aromatic and methoxyl signals very
similar to those of 7. Our suspicion that these com-
pounds were also monoterpene phenyl ethers was
confirmed by NOE interactions between H-5′ and H-1
for both new compounds. The compound with closest
Si gel and C18 retention behavior to 7 showed NOE
interactions that demonstrated E stereochemistry about
the 3,4 double bond, and was assigned the previously
unreported structure 8. Comparison of the 13C NMR
spectra of 8 and 9 (Table 2) showed changes in chemical
shift for C-2 and C-10 consistent with a change from E
to Z stereochemistry about the 3,4 double bond.10 As
reported, we rigorously assigned the H- and 13C NMR
spectra with the aid of HMQC, HMBC, and NOE
difference experiments.
The NOE difference experiments produced a surpris-
ing result. Irradiation of the two methoxyl signals,
which were barely resolved, only gave enhancement of
the H-2′ aromatic proton signal. Structure 1 would be
expected to give enhancements of both the H-2′ and H-5′
signals. An NOE interaction between H-5′ and H-1
protons of the geranyl group was also inconsistent with
structure 1, but could be explained by structure 7
1

Cytotoxic Isoprenyl Phenyl Ethers
J ournal of Natural Products, 1996, Vol. 59, No. 8 731
Ta ble 3. Biological Activity Data for Phenyl Ethersa
(Table 3). Crude extracts of T. lanata were much less
cytotoxic than extracts of T. mollissima, but showed
some antifungal activity. This fact was explained by
the presence of the simpler isoprenyl ether 12 in T.
lanata, because 12 was less cytotoxic than 7 but had
similar mild antifungal activity (Table 3).
assay
7b
7c
8b 9b 10c 11c 12d 13d
BSC cytotoxicitye
100 100 10 10
0
100
75f
0
50
25
75f
0
75f
0
C. albicansg
0
0
0
0
0
0
1
2
0
0
T. mentagrophytesg
2
1
6
a
b
Tested at 60 µg/disk, unless otherwise stated. From T.
mollissima. c From T. tomentella. Synthetic. e Percentage of well
d
showing cytotoxic effects. f At 15 µg/disk. g Width of growth inhibi-
tion zone (mm).
this stereochemistry was confirmed by NOE difference
experiments, this compound has the previously unre-
ported structure 9. Purified samples of both 8 and 9
were found to undergo E-Z isomerization, to mixtures
of 8 and 9, on storage.
As the spectral data for compound 7 were very similar
to those reported for tomentellin,4 we re-isolated the
compounds previously assigned structures 1 (toment-
ellin), 2 (isotomentellin), 3 (demethoxytomentellin), and
6 (deoxytomentellin) from J apanese collections of T.
tomentella. Tomentellin proved identical to the cyto-
toxic compound 7 from T. mollissima. HMBC and NOE
difference experiments showed that the structures of
isotomentellin, demethoxytomentellin, and deoxytomen-
tellin had also been incorrectly assigned as monoterpene
esters. The correct structures are 9 for isotomentellin,
11 for demethoxytomentellin, and 10 for deoxytomen-
tellin.
We did not find the compound assigned structure 4
(trichocolein) in our reinvestigation of J apanese T.
tomentella. Therefore, we undertook a synthesis of
compound 12, previously unreported, which we expected
to be the correct structure of trichocolein. Coupling of
4-bromo-2-methylbut-2-ene to methyl vanillate (13) gave
12 in good yield. Structure 12 was confirmed by
spectral data (Tables 1 and 2), including HMBC and
NOE difference experiments. Comparison of the origi-
1
nal 1H NMR spectrum of trichocolein with the H NMR
spectrum of the synthetic compound supported our
conclusion that the correct structure of trichocolein is
12. Compound 12 was confirmed as a natural product
of Trichocolea when we isolated it as the main isoprenyl
phenyl ether in a New Zealand collection of T. lanata,
a species that had not been investigated previously. We
are currently investigating the correlation of isoprenyl
phenyl ether content with taxonomic divisions of the
New Zealand Trichocolea species.
Compound 7 was the major cytotoxic component of
T. mollissima, active against BSC cells at 15 µg/disk
with no apparent antiviral effects against Herpes sim-
plex or Polio viruses. The only other isoprenyl phenyl
ether with similar cytotoxic potency was compound 11
(Table 3). Therefore, both the allylic ether and the
conjugated enone substructures seem to be required for
cytotoxic activity. It may be that the cytotoxic activity
is due to hydrolysis of the ether linkage to release
â-ocimenones (15), which we have found to be cytotoxic
during our synthetic studies on 7 (unpublished results).
Compounds 7 and 11 were less toxic to fast-growing
P-388 leukemia cells, with IC50’s > 25 µg/mL. The
cytotoxic activity of 7 was tested in the U.S. National
Cancer Institute’s screen against 60 tumor cell lines,11
but the results did not warrant further investigation.
Natural products 7 and 11 were also mildly antifungal
against the dermatophyte Trichophyton mentagrophytes
Exp er im en ta l Section
Gen er a l Exp er im en ta l P r oced u r es. General ex-
perimental and instrumental methods have been de-
scribed previously,2 as have details of antimicrobial,
BSC (antiviral), and P-388 assays.12 Si gel TLC was
carried out using Merck DC-Plastikfolien Kieselgel 60
F254, with hexane-EtOAc (4:1) as the mobile phase,
followed by visualization first with a UV lamp, and then
by dipping in 1% vanillin + 1% H2SO4-EtOH solution
and heating. NMR spectra, of CDCl3 solutions at 25
1
°C, were recorded either at 300 MHz for H and 75 MHz
for 13C on a Varian VXR-300 spectrometer, referenced
to the solvent peaks CHCl3 at 7.25 ppm and CHCl3 at
77.00 ppm; or at 400 MHz for H and 100 MHz for 13C
1
on a J EOL J NM GX400 spectrometer, referenced to
TMS at 0.00 ppm.
P la n t Ma t er ia l. T. mollissima was first collected
from Lake Ellery, Haast, on the west coast of the South
Island, New Zealand, in October 1992 [University of
Otago Herbarium (OTA) specimen no. 046586]. Initial
screening was carried out using an extract produced by
shaking air-dried (30 °C), ground material (5.0 g)
overnight in EtOH (50 mL). T. mollissima was recol-
lected from the same location in J une 1993 (OTA

732 J ournal of Natural Products, 1996, Vol. 59, No. 8
Perry et al.
046636). T. lanata was collected from the Catlins, on
the coast of South Otago, New Zealand, in December
1994 (OTA 046799). T. tomentella was collected in
December 1994, at Kamiyama-cho, Tokushima, J apan
(dry wt 36.6 g, Tokushima specimen no. 94245); and in
J uly 1994, at Ozaka-cho, Gifu, J apan (4 g, no. 94123).
6.0 min peak yielded 7 (18 mg), the 6.5 min peak yielded
8 (5 mg), and the 8.3 min peak yielded 9 (5 mg).
Meth yl 4-[(3E)-3,7-d im eth yl-5-oxo-3,6-octa d ien -
yl)oxy]-3-m eth oxyben zoa te (8): brown oil; TLC Rf
0.3; UV (MeOH) λ max (log ꢀ) 262 (3.78) nm; IR (dry
film) ν max 2925, 2850, 1715, 1625, 1600, 1515, 1435,
1270, 1220, 1135, 1115, 1030, 875, 765 cm-1 ; 1H NMR
in Table 1; 13C NMR in Table 2; EIMS (70 eV) [M]+
332.1619 (20, C19H24O5 req 332.1624), 300.1339 (17,
[M]+ - HOCH3), 182.0582 (48, [M]+ - C10H14O), 151.1145
(97, C10H15O), 151.0424 (40, C8H7O3), 83 (100) 55 (45);
assay results in Table 3.
Met h yl 4-[(3Z)-3,7-d im et h yl-5-oxo-3,6-oct a d ien -
yl)oxy]-3-m eth oxyben zoa te (9): yellow oil; TLC Rf
0.4; UV (MeOH) λ max (log ꢀ) 262 (3.89) nm; IR (dry
film) ν max 3020, 2925, 2855, 1710, 1625, 1600, 1515,
1435, 1295, 1270, 1215, 1115, 1035, 880, 760 cm-1 ; 1H
NMR in Table 1; 13C NMR in Table 2; EIMS (70 eV)
[M]+ 332.1629 (12, C19H24O5 req 332.1624), 300.1329
(10, [M]+ - HOCH3), 182.0582 (53, [M]+ - C10H14O),
151.1109 (100, C10H15O), 151.0419 (20, C8H7O3), 83 (75)
55 (60); assay results in Table 3.
Bioa ct ivit y-Dir ect ed Isola t ion of 7. Dried T.
mollissima from the second collection (31.5 g) was
extracted with EtOH (600 mL, then 3 × 150 mL) and
CHCl3 (1 × 100 mL) by homogenizing and filtering to
give a dark green gum (1.4 g, 25% BSC cytotoxicity at
150 µg/disk, abbreviated as 25% cyt at 150 µg). Reversed-
phase flash chromatography over C18 (1.4 g precoated
on 5.6 g C18, loaded on a 14 g C18 column) was
developed in 20 mL steps from H2O through CH3OH to
CHCl3. The most cytotoxic fraction was eluted with
H2O-CH3OH 1:9 (75 mg, brown oil, 100% cyt at 150
µg). A subsample of this fraction was chromatographed
over Si gel (58 mg precoated on 180 mg Si gel, loaded
on a 2 g column) developed with hexane-EtOAc 9:1 (10
× 2 mL fractions) and then hexane-EtOAc 4:1 (15 × 2
mL fractions). Fractions 14 to 16, which showed a
single UV active spot on TLC (lilac with vanillin/H2SO4),
were combined and solvent removed to give 7 (12 mg).
Re-isola tion of 7, 9, a n d 10 fr om T. tom en tella
(n o. 94245). The diethyl ether extract (1.2 g) was
chromatographed on Si gel using an n-hexane-EtOAc
gradient, giving eight fractions. Fraction 5 (446 mg)
was chromatographed on Sephadex LH-20 using CH2-
Cl2-MeOH and further purified by Si gel HPLC using
n-hexane-EtOAc 4:1 to give 7 (152 mg). Fraction 4 (70
mg) was chromatographed on Sephadex LH-20 using
CH2Cl2-MeOH and further purified by Si gel HPLC
using n-hexane-EtOAc 9:1 to give 9 (13 mg). Fraction
3 (76 mg) was chromatographed on Sephadex LH-20
using CH2Cl2-MeOH and further purified by Si gel
HPLC using n-hexane-EtOAc 9:1 to give 10 (29 mg).
Meth yl 4-[(2E)-3,7-d im eth yl-2,6-octa d ien yl)oxy]-
3-m eth oxyben zoa te (10): yellow crystals, mp 41-42
°C; TLC Rf 0.5 (pale blue with vanillin/H2SO4); UV
(EtOH) λ max (log ꢀ) 223 (4.4), 262 (4.3), 293 (4.1) nm;
IR (dry film) ν max 1720, 1600, 1510, 1440, 1300, 1140,
Met h yl 4-[(2E)-3,7-d im et h yl-5-oxo-2,6-oct a d ien -
yl)oxy]-3-m eth oxyben zoa te (7): colorless oil; TLC Rf
0.25; UV (MeOH) λ max (log ꢀ) 250 (4.39), 288 (3.99)
nm; (MeOH/NaOH) 300 (4.59); IR (dry film) ν max 2950,
2833, 1715, 1687, 1620, 1600, 1511, 1435, 1293, 1271,
1219, 1174, 1134, 1107, 990, 764 cm-1
;
1H NMR in
Table 1; 13C NMR in Table 2; EIMS (70 eV) [M]+
332.1624 (3, C19H24O5 req 332.1624), 301.1429 (3, [M]+
- OCH3), 264.0996 (29, [M]+ - C5H8), 182.0558 (82,
[M]+ - C10H14O), 151.1123 (38, C10H15O), 151.0393 (49,
C8H7O3), 83 (100, C5H7O); CIMS (C4H10) [MH]+ 333 (4),
301 (1), 265 (6), 264 (7), 183 (21), 182 (13), 151 (49), 83
(100); assay results in Table 3.
Isola tion of 8 a n d 9. Dried T. mollissima from the
first collection (40.0 g) was extracted with EtOH (2 ×
300 mL) and CHCl3 (2 × 200 mL) by homogenizing and
filtering to give a combined dark green-brown solution.
Most of the organic solvents were removed by evapora-
tion, then the remaining suspension was partitioned
between H2O (150 mL) and CHCl3 (150 mL). The
CHCl3-soluble material was subjected to reversed-phase
flash chromatography over C18 (1.1 g precoated on 4.0
g of C18, loaded on a 20 g C18 column). This column
was developed in steps of H2O-CH3CN 1:1 (4 × 20 mL),
H2O-CH3CN 1:3 (2 × 20 mL), H2O-CH3CN 1:9 (2 ×
20 mL), CH3CN (2 × 20 mL), CH3CN-CHCl3 3:1 (2 ×
20 mL), and CHCl3 (4 × 20 mL). Compound 7 was
detected by TLC in fractions eluted with H2O-CH3CN,
1:3 so these were combined and the solvent removed (69
mg, yellow oil). TLC of fractions eluted with H2O-CH3-
CN 1:9 showed some 7 plus a higher Rf UV-active spot,
lilac with vanillin/H2SO4. These fractions were com-
bined and solvent removed (47 mg). The CHCl3-soluble
materials from these two samples were subjected to
preparative reversed-phase HPLC (Merck Lichrospher
100 C18, 250 × 10 mm, with 25 × 4 mm guard column).
The mobile phase was H2O:CH3CN 1:3 (5 mL/min) with
UV detection at 280 nm. Samples, as 100 mg/mL
solutions in CH3CN, were injected in amounts of up to
10 mg per injection. Peaks at 6.0, 6.5, and 8.3 min were
collected separately. Combined fractions from the
1
990, 760 cm-1 ; H NMR in Table 1; 13C NMR in Table
2; EIMS (70 eV) [M]+ 318.1826 (1, C19H26O4 req
318.1831), 287.1632 (1, [M]+ - OCH3), 182 (100), 167
(5), 151 (13), 136 (32), 121 (10), 93 (50), 81 (41), 69 (99);
assay results in Table 3.
Re-isola tion of 11 fr om T. tom en tella (n o. 94123).
The diethyl ether extract (0.11 g) was directly chro-
matographed on Sephadex LH-20 to give a mixture
containing 11 (63 mg). This mixture was re-chromato-
graphed on Si gel using an n-hexane-EtOAc gradient
and further purified by Si gel HPLC using n-hexane-
EtOAc 7:3 to give 11 (25 mg).
Meth yl 4-[(2E)-3,7-d im eth yl-5-oxo-2,6-octa d ien -
yl)oxy]-3-h yd r oxyben zoa te (11): colorless oil; TLC Rf
0.15 (lilac with vanillin/H2SO4); UV (EtOH) λ max (log
ꢀ) 212 (4.4), 238 (4.4), 250 (4.0), 293 (3.5) nm; IR (dry
film) ν max 3550, 1710, 1690, 1620, 1590, 1510, 1460,
1
1440, 1290, 1130, 1000 cm-1 ; H NMR in Table 1; 13C
NMR in Table 2; EIMS (70 eV) [M]+ 318.1460 (0.3,
C18H22O5 req 318.1467), 287.1284 (1, [M]+ - OCH3),
250.0853 (2, [M]+ - C5H8), 168.0424 (3, [M]+ - C10H14O),
151.1158 (9, C10H15O), 137.0265 (8, C7H5O3), 83 (100);
assay results in Table 3.
Syn th esis of 12. 4-Bromo-2-methylbut-2-ene (reg-
istry no. 870-63-3, 0.059 g, 0.4 mmol) in dry DMF (0.40

Cytotoxic Isoprenyl Phenyl Ethers
J ournal of Natural Products, 1996, Vol. 59, No. 8 733
mL) was added dropwise, with stirring, to a flame-dried
flask containing methyl 3-methoxy-4-hydroxybenzoate
(13) (registry no. 3943-74-6, 0.060 g, 0.33 mmol) and
NaH (0.013 g, 0.36 mmol) under N2. Stirring was
continued at 25 °C for 24 h, then the reaction mixture
was partitioned between H2O (10 mL) and CH2Cl2 (10
mL). The H2O fraction was re-extracted with CH2Cl2
(10 mL), the organic fractions combined, dried with
MgSO4, and the solvent removed. The crude product
(0.183 g) was precoated on Si gel (Merck 9385, 0.50 g),
loaded onto a dry-packed Si gel column (5 g, 55 × 15
mm), which was developed with hexane-CH2Cl2 (1:4,
2 mL fractions). Fractions 1-5 were combined and the
solvent removed to give 12 (0.060 g, 73%).
using the conditions described above. The major peak
(1 mg, retention time 6.2 min) was collected and proved
to be 12 from H NMR, TLC, and reversed-phase HPLC
1
(coinjection) comparisons with synthetic 12.
Ack n ow led gm en t. We thank the Department of
Conservation’s Westland Conservancy for permission to
collect, R. Tangney for taxonomic expertise, E. Burgess
for some spectroscopic measurements, G. Barns for
biological assays, R. Cunninghame and associates for
microanalysis, B. Clark for MS, and D. Newman and
associates for antitumor testing. This research was
supported in part by the New Zealand Foundation for
Research, Science and Technology; and in part by the
Grant-in-Aid for Cancer Research (04-01) from the
Ministry of Health and Welfare of J apan.
Meth yl 4-(3-m eth yl-2-bu ten oxy)-3-m eth oxyben -
zoa te (12): colorless oil; bp 120 °C/0.15 mm Hg; anal.,
calcd for C14H18O4, C 67.2%, H 7.2%, found C 67.4%, H
7.4%; TLC Rf 0.4 (pale blue with vanillin/H2SO4); UV
(MeOH) λ max (log ꢀ) 262 (4.03), 290 (3.73) nm; IR (dry
film) ν max 2977, 2934, 1711, 1598, 1512, 1464, 1418,
1384, 1366, 1343, 1269, 1218, 1180, 1134, 1105, 992,
Refer en ces a n d Notes
(1) Asakawa, Y. In Progress in the Chemistry of Organic Natural
Products; Herz, W., Kirby, G. W., Moore, R. E., Steglich, W.,
Tamm, C., Eds.; Springer: Vienna, 1995; Vol. 65, and references
therein.
(2) Perry, N. B., Foster, L. M. J . Nat. Prod. 1995, 58, 1131-1135,
1
932, 877, 764 cm-1; H NMR in Table 1; 13C NMR in
and references therein.
Table 2; EIMS (70 eV) [M]+ 250 (2), 182 (100), 151 (72),
69 (53); assay results in Table 3.
(3) Asakawa, Y., Toyota, M., Takemoto, T. Experientia 1978, 34,
155-156.
(4) Asakawa, Y., Takemoto, T. Bryophytorum Bibl. 1978, 13, 335-
Isola tion of 12 fr om T. la n a ta . Dried T. lanata (3.3
g) was extracted with EtOH (66 mL) by homogenizing
and filtering, to give a dark green gum (76 mg). This
extract was partitioned between H2O (40 mL) and
CHCl3 (40 mL), the H2O fraction re-extracted with
CHCl3 (40 mL), and the organic fractions combined and
solvent removed to give a green solid (60 mg). This was
subjected to reversed-phase flash chromatography over
C18 (sample pre-coated on 0.17 g of C18, loaded on 2.00
g C18 column, 65 × 8 mm) developed in six steps from
1:1 H2O-CH3CN through CH3CN to CHCl3. 1H NMR
and Si gel TLC suggested that a fraction eluted with
H2O-CH3CN 1:1 (2 mg) contained 12, which was
further purified by preparative reversed-phase HPLC
353.
(5) Pathak, V. P., Khanna, R. N. Indian J . Chem. 1980, 19B, 1077-
1078.
(6) Asakawa, Y., Toyota, M., Takemoto, T., Mues, R. Phytochemistry
1981, 20, 2695-2699.
(7) Asakawa, Y., Lin, X., Kondo, K., Fukuyama,Y. Phytochemistry
1991, 30, 4019-4024.
(8) Becker, H. In Bryophytes. Their Chemistry and Chemical
Taxonomy; Zinsmeister, H. D., Mues, R., Eds.; Clarendon
Press: Oxford, 1990; pp 339-347.
(9) Allison, K. W., Child, J . The Liverworts of New Zealand;
University of Otago Press: Dunedin, 1975.
(10) Barrow, C. J ., Blunt, J . W., Munro, M. H. G., Perry, N. B. J .
Nat. Prod. 1988, 51, 275-281.
(11) Boyd, M. R., Paull, K. D. Drug Dev. Res. 1995, 34, 91-109.
(12) Lorimer, S. D., Perry, N. B., Tangney, R. S. J . Nat. Prod. 1993,
56, 1444-1450.
NP9603378