Antiviral phenylpropanoid glycosides from the medicinal plant Markhamia lutea

Article abstract of DOI:10.1021/np9703914

Three new phenylpropanoid glycosides, named luteoside A (3), luteoside B (4), and luteoside C (5), were isolated together with the known compounds verbascoside (1) and isoverbascoside (2) from the roots of the medicinal: plant Markhamia lutea. The structures of the new compounds were determined to be 1-O-(3,4-dihydroxyphenyl)ethyl β-D-apiofuranosyl(1→2)-αL- rhamnopyranosyl(1→3)-4-O-caffeoyl-6-acetyl-β-D-glucopyranoside, 1-O-(3,4- dihydroxyphenyl)ethyl β-D-apiofuranosyl(1→2)-α-L-rhamnopyranosyl(1→3)-6- O-cafeoyl-β-D-glucopyranoside, and 1-O-(3,4-dihydroxyphenyl)ethyl β-D- apiofuranosyl(1→2)-α-L-rhamnopyranosyl(1→3)-6-O-feruloyl-β-D- glucopyranoside, respectively, on the basis of chemical and spectroscopic data. All five phenylpropanoid glycosides exhibited potent in vitro activity against respiratory syncytial virus.

Full text of DOI:10.1021/np9703914

564
J . Nat. Prod. 1998, 61, 564-570
An tivir a l P h en ylp r op a n oid Glycosid es fr om th e Med icin a l P la n t
Ma r kh a m ia lu tea
Michael R. Kernan,*,† Ambrose Amarquaye,† J ian Lu Chen,† J ody Chan,† David F. Sesin,† Nigel Parkinson,†
Zhi-jun Ye,† Marilyn Barrett,† Cheryl Bales,† Cheryl A. Stoddart,†,‡ Barbara Sloan,† Phillipe Blanc,†
Charles Limbach,† Salehe Mrisho,§ and Edward J . Rozhon†
Shaman Pharmaceuticals, 213 East Grand Avenue, South San Francisco, California, 94080-4812, and Tanga AIDS Working
Group (c/ o David Scheinman), Rural Development Association, P.O. Box 859 Tanga, Tanzania, East Africa
Received August 19, 1997
Three new phenylpropanoid glycosides, named luteoside A (3), luteoside B (4), and luteoside C
(5), were isolated together with the known compounds verbascoside (1) and isoverbascoside
(2) from the roots of the medicinal plant Markhamia lutea. The structures of the new
compounds were determined to be 1-O-(3,4-dihydroxyphenyl)ethyl â-D-apiofuranosyl(1f2)-R-
L-rhamnopyranosyl(1f3)-4-O-caffeoyl-6-acetyl-â-D-glucopyranoside, 1-O-(3,4-dihydroxyphenyl)-
ethyl â-D-apiofuranosyl(1f2)-R-L-rhamnopyranosyl(1f3)-6-O-caffeoyl-â-D-glucopyranoside, and
1-O-(3,4-dihydroxyphenyl)ethyl â-D-apiofuranosyl(1f2)-R-L-rhamnopyranosyl(1f3)-6-O-feru-
loyl-â-D-glucopyranoside, respectively, on the basis of chemical and spectroscopic data. All
five phenylpropanoid glycosides exhibited potent in vitro activity against respiratory syncytial
virus.
The drug discovery process at Shaman Pharmaceu-
ticals relies on gathering ethnobotanical and ethno-
medical information through interactions between west-
ern trained botanist-physician scientific teams and
traditional healers.1,2 Our interest has focused on
treatments for respiratory syncytial virus (RSV), an
important cause of severe lower respiratory tract infec-
tions in infants and young children.3 In a continuation
of our antiviral drug discovery program,4 the plant
Markhamia lutea Seemann ex Baillor (Bignoniaceae)
was identified as a potential treatment for viral respira-
tory infections, including RSV. In traditional prepara-
tions, the roots are soaked in cold water for about 30
min, and the resulting tea is taken three times daily to
alleviate symptoms of watery, bloodless diarrhea. When
tested in our viral cytopathic effect (CPE) assays,5,6 we
found that several extracts of M. lutea exhibited potent
and selective inhibition of RSV in cell culture. Bioassay-
guided fractionation led to the isolation of five phenyl-
propanoid glycosides responsible for the antiviral ac-
tivity of the extract. They were identified by analysis
of their spectral data, including 2D-NMR, as the known
compounds verbascoside (1) and isoverbascoside (2), and
three new compounds luteoside A (3), luteoside B (4),
and luteoside C (5). All of these compounds had potent
in vitro activity against RSV, and they appear to inhibit
the virus by an intracellular mechanism of action. The
new compounds are related to other phenylpropanoid
glycosides, including forthysioside B (6),7 myricoside
(7),8 and pedicularioside A (8),9 that contain glucose,
rhamnose, and apiose units.
* To whom correspondence should be addressed. Tel.: (650) 952-
7070. Fax: (650) 873-8377. E-mail: mkernan@shaman.com.
† Shaman Pharmaceuticals.
Resu lts a n d Discu ssion
‡ Current address: Gladstone Institute of Virology and Immunology,
University of California, San Francisco, CA 94141.
§ Tanga AIDS Working Group.
The roots of M. lutea were ground to a powder and
extracted with CH2Cl2-i-PrOH (1:1) followed by aque-
S0163-3864(97)00391-1 CCC: $15.00
© 1998 American Chemical Society and American Society of Pharmacognosy
Published on Web 04/10/1998

Antiviral Phenylpropanoid Glycosides from Markhamia
J ournal of Natural Products, 1998, Vol. 61, No. 5 565
Ta ble 1. 1H NMR Assignments (400 MHz) for 3-5 (DMSO-d6 + TFA 0.1%) and 11-13 (CDCl3) (Splitting Patterns and J Values
(Hz) Are Given in Paretheses)
H
3
4
5
11
12
13
caffeoyl
2
5
6
7
7.03 (d, 2)
6.75 (d, 8)
6.98 (dd, 8, 2)
7.46 (d, 16)
6.19 (d, 16)
7.05 (d, 2)
6.74 (d, 8)
6.96 (dd, 8, 2)
7.46 (d, 16)
6.29 (d, 16)
7.32 (d, 2)
6.77 (d, 8)
7.05 (dd, 8, 2)
7.52 (d, 16)
6.50 (d, 16)
3.80 (s)
7.37 (d, 2)
7.23 (d, 8)
7.40 (dd, 8, 2)
7.65 (d, 16)
6.34 (d, 16)
7.37 (d, 2)
7.22 (d, 8)
7.39 (dd, 8, 2)
7.63 (d, 16)
6.40 (d, 16)
7.06 (d, 2)
7.07 (d, 8)
7.05 (dd, 8, 2)
7.65 (d, 16)
6.40 (d, 16)
3.87 (s)
8
OMe
aglycone
2A
6.59 (d, 2)
6.62 (d, 8)
6.47 (dd, 8, 2)
2.68 (t, 8)
6.57 (d, 2)
6.60 (d, 8)
6.54 (dd, 8, 2)
2.66 (t, 7)
6.58 (d, 2)
6.56 (d, 8)
6.44 (dd, 8, 2)
2.66 (t, 7)
7.13 (d, 2)
7.07 (d, 8)
7.14 (dd, 8, 2)
2.98 (t, 7)
7.07 (d, 2)
7.06 (d, 8)
7.10 (dd, 8, 2)
2.97 (t, 7)
7.07 (d, 2)
7.07 (d, 8)
7.07 (dd, 8, 2)
2.97 (t, 7)
5
A
6
A
7
A
8
3.80 (m); 3.58 (m) 3.79 (m); 3.55 (m)
3.78 (m); 3.55 (m) 4.03 (dt, 12, 7);
3.76 (dt, 12, 7)
4.03 (dt, 12, 7);
3.76 (dt, 12, 7)
4.03 (dt, 12, 7);
3.77 (dt, 12, 7)
A
glucose
1′
2′
3′
4′
5′
6′
4.54 (d, 8)
4.39 (d, 8)
3.27 (dd, 9, 8)
3.52 (t, 9)
3.31 (t, 9)
3.49 (m)
4.39 (d, 8)
3.32 (dd, 9, 8)
3.51 (t, 9)
3.29 (t, 9)
3.48 (m)
4.38 (d, 8)
3.82 (dd, 8, 10)
3.97 (t, 10)
5.13 (t, 10)
3.58 (m)
4.37 (d, 8)
4.37 (d, 8)
3.42 (dd, 9, 8)
3.88 (dd, 10, 9)
4.85 (t, 10)
3.78 (m)
4.01 (dd, 12, 5),
3.92 (dd, 12, 3)
3.80 (dd, 8, 9)
3.89 (dd, 10, 9)
5.03 (t, 10)
3.55 (ddd, 10, 6, 2) 3.56 (ddd, 10, 6, 2)
4.27 (dd, 12, 2),
4.21 (dd, 12, 6)
3.81 (dd, 8, 10)
3.89 (t, 10)
5.03 (t, 10)
4.37 (br d, 12),
4.18 (dd, 12, 5)
4.35 (br d, 12),
4.21 (dd, 12, 6)
4.13 (m)
4.28 (dd, 12, 2),
4.21 (dd, 12, 6)
rhamnose
1′′
4.90 (br s)
4.86 (br s)
4.86 (br s)
5.04 (d, 2)
4.99 (d, 2)
4.99 (d, 2)
2′′
3′′
4′′
3.74 (br d, 3)
3.29 (dd, 10, 3)
3.12 (t, 10)
3.75 (br d, 3)
3.49 (dd, 9, 3)
3.21 (t, 9)
3.77 (dr d, 3)
3.47 (dd, 9, 3)
3.21 (t, 9)
5.19 (dd, 3, 2)
5.07 (dd, 10, 3)
4.98 (t, 10)
5.18 (dd, 3, 2)
5.12 (dd, 10, 3)
4.99 (t, 10)
5.18 (dd, 3, 2)
5.12 (dd, 10, 3)
5.00 (t, 10)
5′′
6′′
3.34 (dq, 10, 6)
0.97 (d, 6)
3.89 (dq, 9, 7)
1.10 (d, 7)
3.88 (dq, 9, 6)
1.10 (d, 6)
3.87 (dq, 10, 6)
1.12 (d, 6)
3.90 (dq, 10, 6)
1.12 (d, 6)
3.90 (dq, 10, 6)
1.17 (d, 6)
apiose
1′′′
2′′′
4′′′
5′′′
5.06 (d, 2)
3.76 (d, 2)
3.95, 3.60 (d, 10)
3.39, 3.37 (d, 10)
1.97 (s)
5.05 (d, 2)
3.75 (d, 2)
3.92, 3.58 (d, 10)
3.40, 3.37 (d, 10)
5.05 (d, 2)
3.74 (d, 2)
3.92, 3.58 (d, 10)
3.40, 3.37 (d, 10)
5.27 (br s)
5.31 (br s)
4.35, 4.13 (d, 10)
4.69 (s)
5.32 (br s)
5.31 (br s)
4.34, 4.15 (d, 10)
4.73, 4.67 (d, 12)
b
5.32 (br s)
5.31 (br s)
4.35, 4.15 (d, 10)
4.74, 4.67 (d, 12)
c
OAc
a
a
b
11: 2.32, 2.31, 2.30, 2.86, 2.12, 2.10, 2.08, 2.06, 2.03, 1.94, 1.77 (all s). 12: 2.32, 2.31, 2.30, 2.86, 2.12, 2.10, 2.08, 2.06, 1.94, 1.77 (all
s). c 13: 2.33, 2.27, 2.25, 2.15, 2.10, 2.09, 2.08, 2.04, 2.03, 1.96 (all s).
ous i-PrOH or with aqueous EtOH. These extracts
exhibited significant inhibition against RSV in the
standard antiviral cytopathic effect (CPE) assay.5,6 The
extracts were also active when administered 3 h after
infection of the cells with RSV, indicating that they may
inhibit the virus by an intracellular mechanism of
action. To determine the identity of the compound(s)
responsible for the anti-RSV activity of M. lutea ex-
tracts, bioassay-guided fractionation of these extracts
using an RSV CPE assay was conducted. Initially, the
extracts were purified using a combination of liquid-
liquid extraction, reversed-phase chromatography, cen-
trifugal partition chromatography (CPC), and prepara-
(10), respectively. The spectral data for 9 was also
identical to literature values;12 although the NMR
assignments for 10 had not been previously reported,
the spectral data for this compound were consistent with
the structure.
The spectral data, including UV, IR, MS, and NMR,
of luteosides A-C (3-5) indicated that they were related
to 1 and 2. For example, the UV spectrum of 3 was
similar to that of 1, with λmax at 325, 292, 245, 238, 218,
and 203 nm, while the IR spectrum of 3 had absorbances
due to hydroxyl groups (ν 3420 cm-1, br), ester (1720,
1690 cm-1), and phenol groups (1636 cm-1). Their
molecular formulas, C36H46O20 for 3, C34H44O19 for 4,
and C35H46O19 for 5, were determined from their high-
resolution negative-ion FAB mass spectra. Comparison
of the NMR data, including data from 2D NMR experi-
ments, to similar data for 1 and 2, identified the
presence of a â-D-glucose, an R-L-rhamnose, and a â-(3,4-
dihydroxyphenyl)ethoxy group in all three compounds
(Tables 1 and 2). In addition, signals due to an acetyl
group in 3, a caffeoyl ester in 3 and 4, and a feruloyl
ester in 5 were observed. All three compounds also
contained signals attributed to an â-D-apiose group [e.g.,
for 3, δ 5.06 (d, J ) 2 Hz, H-1′′′), 109.4 (C-1′′′); 3.76 (d,
J ) 2 Hz, H-2′′′), 76.7 (C-2′′′); 78.9 (C-3′′′); 3.95 (d, J )
10 Hz, H-4a′′′), 3.60 (d, J ) 10 Hz, H-4b′′′), 73.7 (C-4′′′);
3.39 (d, J ) 10 Hz, H-5a′′′), 3.37 (d, J ) 10 Hz, H-5b′′′),
63.4 (C-5′′′)]. Acid hydrolysis of each of the compounds
3-5 with 5 N HCl gave three sugars identified as â-D-
glucose, R-L-rhamnose, and â-D-apiose. Exact mass
tive TLC to isolate the active consitituents 1-4.
A
method more amenable to larger scale production using
reversed-phase chromatography and preparative HPLC
was subsequently conducted to give compounds 1-4
along with a related minor compound 5. Compounds
1-5 all had potent in vitro activity against RSV.
Similar extracts prepared from bark also had anti-RSV
activity and were shown to contain compounds 1-5.
Verbascoside (1) and isoverbascoside (2) were identi-
fied from their spectral data, which were identical in
all respects to literature values.10,11 The 1H and 13C
NMR assignments for these compounds were deter-
mined using 2D NMR experiments (COSY, HMQC,
HMBC) and consistent with reported values. As ex-
pected, acid hydrolysis of both 1 and 2 gave â-D-glucose
and R-L-rhamnose. Acetylation of 1 and 2 gave verbas-
coside nonaacetate (9) and isoverbascoside nonaacetate

566 J ournal of Natural Products, 1998, Vol. 61, No. 5
Kernan et al.
Ta ble 2. 13C NMR Assignments for Phenylpropanoids 3-5
(100 MHz, DMSO-d6 + TFA 0.1%) and Acetates 10-13 (CDCl3)
carbon
3
4
5
10
11
12
13
caffeoyl
1
2
3
4
5
6
7
8
125.4 125.5 125.5 133.2 132.8 133.1 133.2
114.7 115.0 110.9 122.9 122.8 122.8 121.5
145.8 145.5 147.9 142.5 142.5 142.4 140.6
148.6 145.0 149.3 143.7 143.8 143.6 151.4
115.5 115.8 115.4 123.9 124.1 123.9 111.2
121.5 121.4 123.4 126.6 126.4 126.6 133.2
145.6 148.4 145.0 143.6 144.2 143.5 144.7
113.3 113.8 114.3 118.6 118.2 118.6 117.7
165.7 166.5 166.6 166.2 165.0 166.1 166.4
CO
OMe
55.7
56.0
aglycone
1A
2A
3A
4A
5A
6A
129.0 129.1 129.0 137.6 136.8 136.8 136.8
115.8 115.5 115.4 123.1 123.8 123.3 123.3
145.0 145.3 145.2 140.5 140.6 140.6 141.5
143.6 143.5 143.5 141.8 141.9 141.9 141.9
116.2 116.2 116.2 123.8 123.3 123.8 123.8
119.5 119.5 119.5 127.2 127.0 127.0 127.0
7A
8A
35.0
70.7
35.1
70.3
35.1
70.5
35.4
69.8
35.5
70.0
35.5
69.7
35.5
70.0
glucose
1′
101.2 101.3 101.3 100.7 101.3 101.3 101.4
2′
3′
4′
5′
78.6
78.5
69.3
70.2
62.4
77.6
82.1
70.5
73.5
63.3
77.6
82.1
68.6
73.4
63.2
77.2
81.6
69.8
72.2
62.5
76.4
80.4
69.8
72.0
62.6
75.5
82.1
70.1
72.1
62.8
75.7
82.0
70.7
72.1
62.7
6′
rhamnose
1′′
2′′
3′′
4′′
101.7 101.4 101.4
99.6
69.7
68.8
70.5
67.5
17.3
98.4
69.5
68.8
70.5
67.3
17.2
98.8
70.0
68.8
70.7
67.3
17.0
98.8
69.7
68.8
70.1
67.3
17.1
70.9
70.5
71.5
68.8
18.1
70.6
70.5
71.9
68.7
17.8
70.5
70.5
71.9
68.6
17.8
5′′
6′′
F igu r e 1. Structures of luteoside A (3) and B (4) showing key
HMBC correlations (1H f 13C) and HRFABMS fragments.
apiose
1′′′
109.4 109.3 109.2
105.3 105.2 105.2
2′′′
3′′′
4′′′
5′′′
COCH3
COCH3
76.7
78.9
73.7
63.4
170.1
20.6
76.6
79.0
73.7
63.7
76.6
79.0
73.7
63.7
75.8
83.9
73.1
63.6
b
75.9
84.0
73.2
63.7
c
75.9
84.0
73.2
63.7
d
4.54 (H-1′): 70.7 (CA-8)], the apiose was attached to C-2′
[δ 3.42 (H-2′): 109.4 (C-1′′′)], the rhamnose was attached
to C-3′ [δ 3.88 (H-3′): 101.7 (C-1′′)], and the caffeoyl
group was attached to C-4′ on the glucose [δ 4.85 (H-
4′): 165.7 (C-9)]. The chemical shifts at C-6′ (63.9 ppm)
and H-6′ [δ 4.01 (dd, J ) 12, 5 Hz), 3.92 (dd, J ) 12, 3
Hz)] are consistent with acetylation at C-6′ on the
a
a
b
c
d
a
10: CO, 170.1, 170.0, 169.5, 169.4, 169.4, 168.4, 168.3, 168.0,
b
167.9; COCH3, 21.1, 20.9, 20.8, 20.63 (3C), 20.59 (3C). 11: CO,
170.7, 170.6, 170.1, 170.0, 169.8, 169.7, 169.4, 168.3, 168.2, 168.0,
167.9; COCH3, 21.1, 20.9, 20.79, 20.80, 20.6 (3C), 20.58, 20.52, 20.4.
c 12: CO, 170.6, 170.0, 169.9, 169.83, 169.77, 169.7, 169.4, 168.32,
168.28, 168.0, 167.9; COCH3, 21.2, 21.1, 20.9, 20.8 (2C), 20.65 (2C),
20.62 (2C), 20.59 (2C). 13: CO, 170.6, 170.0, 169.9, 169.8, 169.74,
169.70, 169.4, 168.7, 168.3, 168.0; COCH3, 21.2, 21.1, 20.9, 20.8
1
glucose. The H NMR chemical shift of the rhamnose
secondary methyl group (δ 0.97 ppm) are consistent with
attachment of the caffeoyl moeity at C-4′ as in verbas-
coside (1) rather than at C-6′ as in isoverbascoside (2).13
On the basis of these data, the structure of luteoside A
was determined to be 1-O-(3,4-dihydroxyphenyl)ethyl
â-D-apiofuranosyl(1f2)-R-L-rhamnopyranosyl(1f3)-4-O-
caffeoyl-6-acetyl-â-D-glucopyranoside (3).
d
(2C), 20.63 (3C), 20.58 (2C).
measurements of fragmentation ions observed in the
negative FABMS of 3 and 4 (Figure 1) established that
both apiose and rhamnose were terminal units in both
3 and 4 and that the acetate was attached to the glucose
in 3 [e.g., for 3, m/z 797 ([M - H]-, dev 6.9 ppm), 665
([M - api - H]-, dev 4.4 ppm), 651 ([M - rha - H]-,
dev - 1.7 ppm), and 459 ([M - api - rha - HOAc -
H]-, dev - 1.7 ppm)].
The structures of luteosides A-C (3-5) were deter-
mined by careful analysis of 2D NMR data of the
natural products and their peracetate derivatives 11-
13. For example, in luteoside A (3), most of the 1H NMR
signals, including those due to the sugars, were assigned
from a COSY spectrum (see Table 1). Heteronuclear
1H-13C correlation experiments (HMQC, HMBC) then
allowed assignment of the structure and the remaining
NMR signals (Figure 1, Tables 1 and 2). Key HMBC
correlations observed for 3 required that the â-(3,4-
dihydroxyphenyl)ethoxy group was attached to C-1′ [δ
Analysis of the NMR spectra for luteoside A unde-
caacetate (11) provided further support for the proposed
structure. Although some of the sugar protons gave
1
overlapping signals in the H NMR spectrum of 3, the
corresponding signals in 11 were well dispersed and
were assigned using 2D NMR, including HMBC and
HMQC-TOCSY experiments. Correlations in the
HMBC spectrum of 11 from H-1′ to CA-8, H-2′ to C-1′′′,
H-3′ to C-1′′, H-1′′ to C-3′, and H-1′′′ to C-2′ confirmed
the positions of the sugars. The expected upfield shifts
in the 13C NMR spectrum due to the â effect of adjacent
1
acetyl groups and downfield shifts in the H NMR due
to the acetyl groups were observed in the NMR spectra
of 11.
The structures of luteoside B and luteoside C were
determined to be 1-O-(3,4-dihydroxyphenyl)ethyl â-D-
apiofuranosyl(1f2)-R-L-rhamnopyranosyl(1f3)-6-O-caf-

Antiviral Phenylpropanoid Glycosides from Markhamia
J ournal of Natural Products, 1998, Vol. 61, No. 5 567
Ta ble 3. In Vitro Antiviral Activity (EC50) and Cytotoxicity
(IC50) of Phenylpropanoids and Various Control Compounds
feoyl-â-D-glucopyranoside (4) and 1-O-(3,4-dihydroxy-
phenyl)ethyl â-D-apiofuranosyl(1f2)-R-L-rhamnopyran-
osyl(1f3)-6-O-feruloyl-â-D-glucopyranoside (5), respec-
tively, by interpretation of the spectral data for 4, 5,
and the acetates 12 and 13, as described in the structure
elucidation of 3. Compounds 4 and 5 are differentiated
from 3 in that they are analogues of isoverbascoside (2)
rather than verbascoside (1). The HMBC spectrum of
4 included correlations from H-1′ to CA-8, H-2′ to C1′′′,
H-3′ to C-1′′, H-1′′ to C-3′, and H1′′′ to C-2′. All of these
HMBC correlations were also observed for 5, with the
exception of the correlation from H-3′ to C-1′′, and with
an additional correlation from the methoxy signal to
C-3. The HMBC correlations observed for 4 (Figure 1)
and 5 are consistent with the proposed structures. Key
differences in the NMR spectra are signals due to C-4′
and C-6′ on the glucose and C-6′′ on rhamnose, with
additional signals in 5 due to the methoxy group. The
HMQC-TOCSY and HMBC experiments allowed un-
ambiguous assignment of all of the signals in com-
pounds 11-13 and provided further support for the
proposed structures. The 1H and 13C NMR assignments
of 3-5 and the acetates 11-13 are summarized in
Tables 1 and 2, with key HMBC correlations for 3 and
4 shown in Figure 1.
Comparison of the 13C NMR assignments for luteo-
sides A-C (3-5) with the related compounds 6-8,7-9
14, and 1514 revealed that the major differences, due to
the attached subunits, are in the signals assigned to
C-2′, C-3′, and C-6′ of the glucose. The signals assigned
to the apiose group are similar to the apiose signals in
6-8, 14, and 15. Compound 4 is isomeric with pedicu-
larioside A (6), myricoside (7), and forsythioside A (8),
with the differences being the positions of the rhamnose,
apiose, and caffeoyl groups on the central glucose.
a
b
compd
virus
assay
EC50
IC50
SIc
n
verbascoside (1)
RSVd
RSV
CPEe 0.80
plaquef 9.7
76.9
n.t.
44
85
2
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
mCMVg CPE
none
none
0.62
HSV-2h CPE
71
isoverbascoside (2) RSV
2
2
CPE
51.4
n.t.
75
84
RSV
plaque 1.1
plaque none
VZVi
2
2
2
3
4
4
4
5
9
10
16
17
mCMV CPE
HSV-1j CPE
HSV-2 CPE
none
none
none
0.87
3.4
none
none
15.5
none
3.40
5.87
1.93
89
30
27
77
>67
93
90
189
33.7
11.2
26.2
15.9
35 ( 4.6
25 ( 1.7
>100
RSV
RSV
CPE
CPE
89
>19
HSV-1 CPE
HSV-2 CPE
RSV
RSV
RSV
RSV
RSV
RSV
RSV
CPE
CPE
CPE
CPE
CPE
CPE
12
3.3
4.4
8.2
19 23
11 21
>20 20
>250
ribavirin
ribavirin
gancyclovir
acyclovir
acyclovir
gancyclovir
1.8 ( 0.2
plaque 2.2 ( 0.2
mCMV CPE
HSV-1 CPE
HSV-2 CPE
5.0 ( 0.4
0.26 ( 0.01 >67
2
2.3 ( 0.3
>10
840
>4.3 14
74 12
VZV
plaque 26.7
a
b
Antiviral activity, 50% effective concentration (µg/mL). Cy-
totoxicity, 50% inhibitory concentration (µg/mL). c Selective index
d
) IC50/EC50
.
Respiratory syncytial virus, Long (CPE) or A-2
(plaque) strain in Hep-2 cells. e Cytopathic effect. f Plaque-neu-
tralization assay. g Murine cytomegalovirus, RM-461 strain. Her-
h
i
pes simplex virus type 2. Varicella zoster virus, 3CV-1 strain.
Herpes simplex virus type 1.
j
Ta ble 4. Effect of Time of Administration on Antiviral
Activity of Compounds and Extracts
time of administration
before infection (t ) 0)
after infection (t ) 3 h)
extract or
compd
a
b
a
b
EC50
IC50
SIc
EC50
IC50
SIc
A ro E2d
0.3
8.2
2.1
0.80
0.62
0.87
3.4
15.5
2.9
42.8 143
1.7
140
>830
130
108
125
180
180
>330
26
82
>15
43
415
735
16
A ro E3e
385
130
76.9
46.9 55
62
85
The in vitro antiviral activities of the five isolated
compounds are summarized in Table 3. Compounds
1-4 all have similar or better antiviral activity (EC50)
against RSV in the in vitro assay than ribavirin, an
approved drug for the treatment of RSV infections in
humans; these compounds also have better selectivity
against RSV than ribavirin, as demonstrated by the
large selective index, or the ratio of cytotoxicity (IC50)
to antiviral activity (SI ) IC50/EC50). The compounds
were active when adminstered 3 h after infection of the
cells in a neutral red anti-RSV assay, whereas SP-303,
a mixture of proanthocyanidin oligomers known to
inhibit RSV via an extracellular mechanism,5 had no
activity against RSV when administered 3 h postinfec-
tion (see Table 4). Hence, it is likely that the phenyl-
propanoid glycosides, like ribavirin, inhibit RSV through
an intracellular antiviral mechanism of action. Com-
pound 5 is considerably less active than 1-4; this drop
in activity could be due to methylation of the caffeoyl
ester. It is noteworthy that there was a considerable
drop in the anti-RSV activity of two of the extracts when
they were administered postinfection (Table 4). This
is probably due to the presence of compounds in the
extracts that are active via an extracellular mechanism
of action and have no activity in the post infection anti-
RSV assay. Other compounds containing caffeoyl ester
groups, including caffeic acid, have been reported to
have antiviral activity against herpesviruses.15 How-
B ro E4f
1
2
3
4
5
3.0
0.26
0.17
11
0.23
12
51.4
77
84
89
>67
189
28
>19
780
>28
12
SP-303
ribavirin
9
none
1.8 ( 0.2 35 ( 4.6
19
2.3 ( 0.2 25 ( 1.8
10.8
a
b
Antiviral activity, 50% effective concentration (µg/mL). Cy-
totoxicity, 50% inhibitory concentration (µg/mL). c Selective index
d
) IC50/EC50
.
First collection, roots, i-PrOH-CH2Cl2 extract.
e First collection, roots, i-PrOH-H2O extract. f Second collection,
roots, EtOH extract.
ever, none of the phenylpropanoid glycosides tested had
any activity against the herpesviruses HSV, CMV, or
VZV. We prepared some derivatives of verbascoside and
isoverbascoside for preliminary SAR studies. The non-
aacetates 9 and 10 were hydrolyzed with triethylamine
to give the pentaacetates 16 and 17, respectively. Both
16 and 17 had moderate activity against RSV but were
considerably less active than the natural products.
These results suggest that the inhibition of RSV by
phenylpropanoid glycosides requires two catechol groups
and a central sugar unit and that the mechanism of
anti-RSV inhibition may be different than the antiher-
petic activity reported for other caffeoyl esters. Further
details on the anti-RSV activity of related compounds
containing a caffeoyl ester group will be published
elsewhere.

568 J ournal of Natural Products, 1998, Vol. 61, No. 5
Kernan et al.
We have demonstrated that M. lutea, a medicinal
plant used for symptoms attributed to viral respiratory
infections, contains five phenylpropanoid glycosides with
antiviral activity against respiratory syncytial virus.
The active constituents of M. lutea include the com-
monly occurring compounds16 verbascoside and isover-
bascoside and the new compounds 3-5. It is likely that
most of the phenylpropanoid glycosides, a commonly
occurring group of plant-derived natural products,16
have activity against RSV. The phenylpropanoid gly-
coside echinaside, from Echinacea pallida, was reported
to have in vitro activity against vesicular stomatitis
virus (VSV);17 it is possible that other phenylpropanoid
glycosides are also active against VSV. We report for
the first time the isolation of phenylpropanoid glycosides
from the genus Markhamia.
37 °C for 3 h to allow virus absorption and penetration.
After incubation, diluted sample (50 µL) was added to
each well, and plates were incubated at 37 °C for 3 days
until complete destruction of the cell monolayer occurred
in virus control wells. The plates were then stained
with neutral red and the antiviral and cytotoxic activity
measured.18 The antiviral activity of each sample is
expressed in µg/mL as 50% effective concentration
(EC50), and cytotoxicity is expressed in µg/mL as 50%
inhibitory concentration (IC50).
Extr a ction a n d Isola tion . The roots of the first
collection (J an 1993, 200 g) were air-dried, ground to a
powder, and then extracted with i-PrOH-CH2Cl2 (1:1)
for ∼24 h. The resulting extract was evaporated (4.67
g), suspended in water (40 mL), and extracted succes-
sively with CH2Cl2 (3 × 35 mL) and EtOAc (4 × 35 mL).
The two organic-soluble fractions were evaporated to
dryness. The aqueous fraction was evaporated to
remove residual solvent and purified on an HP-20
column (125 g, 3.5 × 45 cm), eluting with increasing
amounts of MeOH in H2O. The resulting fractions were
evaporated to dryness. Samples of all fractions were
then submitted for testing in the RSV CPE assay and
analyzed by TLC (40% aqueous MeOH, C18).
Exp er im en ta l Section
Gen er a l Exp er im en ta l P r oced u r es. The general
experimental procedures have been described previ-
ously.4 NMR data for compounds 16 and 17 were
collected using a Varian nanoprobe. 13C multiplicities
for all compounds were determined from DEPT spectra.
TLC was conducted using EM Separations Si60 and RP-
18WFs (0.2 mm, 10 × 10 cm; preparative TLC on 1.0
mm (RP-18WFs) or 2.0 mm (Si60) plates, 20 × 20 mm)
HPTLC plates, visualized using UV and/or 5% vanillin/
EtOH and 5% H2SO4/EtOH with heating (100 °C, 5-10
min). Column chromatography was carried out using
HP-20 polystyrene divinylbenzene gel obtained from
Mitsubishi Kasei Corp. or Bakerbond C18 obtained from
Fisher Scientific. CPC was carried out using a Sanki
Series 1000 instrument equipped with a Linear UVIS
detector. HPLC was conducted on a Rainin Dynamax
system.
P la n t Ma ter ia l. The roots of M. lutea were collected
on J an 1, 1993, in Misima District, Tanzania, by Dr.
Charles Limbach, and on Sep 4, 1993, in Muheza
District, Tanzania, by Salehe Waziri. Plant specimens
were identified by Charles Mabula of Tanzania and B.
Verdcourt, A. Gereau, R. Liesner, and Mary Merello
of the Missouri Botanical Garden. Voucher specimens
(voucher nos. CL 257, ROA 14, and ROA 33) are
deposited in the reference collection, Ethnobotany and
Conservation Department, Shaman Pharmaceuticals.
An tivir a l a n d Cytotoxicity Assa ys. The antiviral
activities of compounds 1-5 as well as control antiviral
compounds were determined using viral cytopathic
effect (CPE) assay and the plaque-neutralization (plaque)
assay. The procedures used for the antiviral and
cytotoxicity assays have been previously described.5,6
The RSV “post infection” CPE assay was conducted with
a neutral red endpoint as follows. The day before the
assay was performed, cells were seeded into 96-well
plates (10 000 cell in 50 µL/well) in high-glucose con-
taining Dulbecco’s modified Eagle’s medium (DMEM)
supplemented with 2% or 10% heat-inactivated fetal
bovine serum (FBS), L-glutamine, and antibiotics. On
the day of the assay, samples were dissolved in an
appropriate solvent (10-50 mg/mL), and serial 3-fold
dilutions (in 200 µL volumes) were made in DMEM-FBS
for a final range of eight concentrations (830-0.38 or
130-0.13 µg/mL). Virus (3000 plaque-forming units in
50 µL of DMEM-FBS with 20 mM HEPES buffer) or
medium alone was added to each well and incubated at
Fractions eluting from the HP-20 column with 40%
aqueous MeOH (318 mg) and 60% aqueous MeOH
(208.3 mg) had anti-RSV activity. The 40% aqueous
MeOH fraction was purified by column chromatography
on C18 (Bakerbond, 60 g, 2 × 45 cm) eluting with 40%
aqueous MeOH. Fractions were monitored by TLC
(C18, 40% MeOH; SiO2, CH2Cl2-MeOH-H2O 43:37:20,
lower phase) and with the RSV CPE assay. Fractions
containing 1 had anti-RSV activity and were combined,
evaporated (48.3 mg), and purified by preparative TLC
on silica (CH2Cl2-MeOH-H2O 40:40:20, lower phase)
to give 1 (12.2 mg, 0.006%). The 60% aqueous MeOH
fraction from the HP-20 column was purified by CPC
(1200 rpm, 2 mL/min), eluting with the lower phase of
CH2Cl2-MeOH-H2O (40:40:20); the eluent was moni-
tored by TLC (C18, 40% aqueous MeOH). Fractions
containing 2 and 4 had anti-RSV activity and were
combined, evaporated (24.2 mg), and purified by pre-
parative TLC on C18 (40% aqueous MeOH) to give 2
(8.0 mg, 0.004%) and 4 (3.5 mg, 0.002%). Fractions
containing 3 had anti-RSV activity and were combined,
evaporated (14.9 mg), and purified by preparative TLC
on C18 (40% aqueous MeOH) to give 3 (2.0 mg, 0.001%).
The marc was extracted with i-PrOH-H2O (1:1) for
∼24 h to give an extract (5.08 g) that had anti-RSV
activity. Purification and analysis of the fractions by
TLC and in the RSV CPE assay established that the
active constituents in the i-PrOH-H2O extract were
compounds 1-4.
A larger collection (Sep 1993, 15 kg) was ground to a
powder, extracted with 80% aqueous EtOH (60 L) for
24 h, and evaporated to give an EtOH extract (742.4
g). This was partitioned between CH2Cl2 and H2O to
give a dark brown CH2Cl2-soluble portion (35.1 g), an
insoluble residue (26.3 g), and an aqueous layer. The
aqueous layer was purified on an HP-20 column (18 ×
70 cm), eluting with increasing amounts of MeOH in
H2O (0-100%); the resulting fractions were evaporated
to dryness and samples of all fractions analyzed by TLC
(C18, 40% aqueous MeOH). The fraction eluting with

Antiviral Phenylpropanoid Glycosides from Markhamia
J ournal of Natural Products, 1998, Vol. 61, No. 5 569
40% aqueous MeOH (52.5 g) was purified on a C18
column (9 × 50 cm) eluting with 40% aqueous MeOH
followed by 75% aqueous MeOH (15 mL/min). Two
enriched fractions containing 1 and 2 (18 g) and 4 (14
g) eluted with 40% aqueous MeOH and were purified
on a second C18 column (9 × 50 cm, 20-100% aqueous
MeOH) followed by preparative HPLC on C18 (Prime-
sphere, 50 × 250 mm, 20% aqueous MeCN, 45 mL/min)
to give 1 (1.7 g, 0.011%), 2 (1.3 g, 0.0087%), and 4 (1.1
g, 0.0073%). A fraction containing 3 and 5 (1.64 g)
eluted from the first C18 column with 75% aqueous
MeOH; a portion (0.70 g) of this was purified by
preparative HPLC on C18 (Kromasil, 20 × 250 mm, 25%
aqueous MeCN, 15 mL/min) to give 3 (351 mg, 0.0056%)
and 5 (76.5 mg, 0.0012%).
Ver ba scosid e (1): yellowish powder; [R]D -74.0°
(MeOH, c 0.424). All spectral data were identical to
those in the literature.10,11
Isover ba scosid e (2): off-white powder; [R]D -47.2°
(MeOH, c 0.212). All spectral data were identical to
those in the literature.11
R-L-rhamnose, and â-D-apiose in the sample. The pres-
ence of ferulic acid was detected by TLC analysis of the
EtOAc phase.
Ver b a scosid e Non a a cet a t e (9). Acetylation of 1
(25.0 mg) [Ac2O/pyridine (1:1), room temperature, 24 h]
and purification on silica gel (CH2Cl2-EtOAc) gave
verbascoside nonaacetate 9 as a yellow solid (28.0 mg,
70%). All spectral data were identical to those in the
literature.12
Isover ba scosid e Non a a ceta te (10). Isoverbasco-
side (25.5 mg) was acetylated following the procedure
used to prepare 9 to give isoverbascoside nonaacetate
10 as an amorphous white powder (27.3 mg, 67%): 1H
NMR (DMSO-d6 + 0.1% TFA) δ 7.65 (d, H-7, J ) 16
Hz), 7.42 (dd, H-6, J ) 8, 2 Hz), 7.39 (d, H-2, J ) 2 Hz),
7.23 (d, H-5, J ) 8 Hz), 7.06 (dd, H-6A, J ) 8, 2 Hz),
7.05 (d, H-5A, J ) 8 Hz), 7.03 (d, H-2A, J ) 2 Hz), 6.34
(d, H-8, J ) 16 Hz), 5.11 (dd, H-3′′, J ) 10, 3 Hz), 5.09
(dd, H-2′′, J ) 3, 2 Hz), 5.08 (dd, H-2′, J ) 10, 8 Hz),
5.06 (t, H-4′, J ) 10 Hz), 4.80 (d, H-1′′, J ) 2 Hz), 4.32
(dd, H-6′, J ) 12, 3 Hz), 4.25 (dd, H-6′, J ) 12, 6 Hz),
4.12 (m, HA-8a), 3.88 (dq, H-5′, J ) 10, 6 Hz), 3.79 (t,
H-3′, J ) 10 Hz), 3.65 (m, HA-8b), 3.62 (m, H-5′), 2.88
(m, HA-7), 2.32 (s, 3 H), 2.31 (s, 3 H), 2.28 (s, 3 H), 2.27
(s, 3 H), 2.14 (s, 3 H), 2.11 (s, 3 H), 2.10 (s, 3 H), 1.96 (s,
3 H), 1.95 (s, 3 H), 1.15 (d, H-6′′, J ) 6 Hz); 13C NMR
see Table 2; negative FABMS m/z 1001.9 [M - H]-.
Lu teosid e A Un d eca a ceta te (11). Acetylation of
3 (35.5 mg) [Ac2O/pyridine (1:1), room temperature, 24
h] and purification by HPLC on diol (CH2Cl2-MeOH)
gave the undecaacetate 11 (41.2 mg, 76%) as an
amorphous white powder: 1H NMR see Table 1; 13C
NMR see Table 2; positive FABMS m/z 1241 [M + Na]+,
1219 [M + H]+, 1178, 940.
Lu teosid e B Un d eca a ceta te (12). Luteoside B
(32.6 mg) was acetylated following the procedure used
to prepare 11 to obtain the undecaacetate 12 (46.4 mg,
90%) as an amorphous powder: 1H NMR see Table 1;
13C NMR see Table 2; positive FABMS m/z 1219 [M +
H]+, 1178, 940.
Lu teosid e C Deca a ceta te (13). Luteoside C (10.1
mg) was acetylated following the procedure used to
prepare 11 to obtain the decaacetate 13 (11.0 mg, 69%)
as an amorphous powder: 1H NMR see Table 1; 13C
NMR see Table 2; positive FABMS m/z 1191 [M + H]+,
1149, 982, 954, 910, 766.
Lu teosid e A (3): off-white powder; [R]D -81.0°
(MeOH, c 0.648); λmax (log ꢀ, MeOH) 203 (4.19), 218
(3.50), 238 (sh, 3.47), 245 (3.53), 292 (3.89), 325 (4.08)
nm; IR (KBr) νmax 3420 (br, OH), 2937, 1720, 1690, 1636,
1605, 1522, 1448, 1370, 1280, 1157, 1040, 812 cm-1; 1H
NMR see Table 1; 13C NMR see Table 2; negative
HRFABMS m/z 797.2559 [M - H]-, 665.2111 [M -
api]-, 651.1924 [M - rha]-, 459.1283 [M - rha - api -
HOAc]-, calcd for C36H45O20 797.2504, for C31H37O16
665.2082, for C30H35O16 651.1925, for C23H23O10 459.1291.
Lu teosid e B (4): off-white powder; [R]D -51.5°
(MeOH, c 0.824); λmax (MeOH) 204 (4.01), 212 (sh, 3.80),
237 (3.58), 291 (3.83), 331 (3.98) nm; IR (KBr) νmax 3420
(br, OH), 2930, 2870, 1700, 1686, 1609, 1522, 1458,
1
1364, 1284, 1040, 812 cm-1; H NMR see Table 1; 13C
NMR see Table 2; negative HRFABMS m/z 755.2477 [M
- H]-, 623.1931 [M - api]-, 609.1836 [M - rha]-,
459.1320 [M - rha - api - H2O]-, calcd for C34H43O19
755.2399, for C29H35O15 623.1976, for C28H33O15 609.1819,
for C23H23O10 459.1291.
Lu teosid e C (5): off-white powder; [R]D -42.8°
(MeOH, c 0.236); λmax (log ꢀ, MeOH) 204 (3.90), 213 (sh,
3.62), 236 (3.55), 289 (3.80), 327 (3.94) nm; 1H NMR see
Table 1; 13C NMR see Table 2; negative HRFABMS m/z
770.2633 [M - H]-; calcd for C35H45O19 769.2555.
Acid Hyd r olysis of 3. Luteoside A (1.3 mg) was
dissolved in 5 M HCl and heated for 2.5 h at 90 °C. The
reaction mixture was cooled to room temperature and
then extracted with EtOAc (2 × 200 mL). The aqueous
phase was neutralized with 1 M Na2CO3, freeze-dried,
and extracted with pyridine (100 mL). The pyridine
extract was then analyzed by TLC on silica (EtOAc-
MeOH-H2O-HOAc 13:3:3:4; vis. with 1% p-anisalde-
hyde, 2% H2SO4 in HOAc, 100 °C). Analysis indicated
the presence of â-D-glucose (Rf ) 0.65), R-L-rhamnose
(Rf ) 0.77), and â-D-apiose (Rf ) 0.73).
Acid Hyd r olysis of 4. Hydrolysis and TLC analysis
of luteoside B (4) (0.8 mg) using the method described
for 3 resulted in the detection of â-D-glucose, R-L-
rhamnose, and â-D-apiose in the sample.
Acid Hyd r olysis of 5. Hydrolysis and TLC analysis
of the luteoside C (5) (0.5 mg) using the method
described for 3 resulted in the detection of â-D-glucose,
Hyd r olysis of 9. Compound 9 (10.0 mg) was dis-
solved in anhydrous MeOH (1 mL); Et3N (20 µL) was
added and the mixture stirred at room temperature
under argon for 90 min. The sample was neutralized
by adding formic acid (20 µL) and evaporated in vacuo
and the residue purified on a cyano SPE column (0-
10% MeOH in CH2Cl2) and evaporated under vacuum
to give the pentaacetate 16 (6.0 mg, 72%) as an
amorphous white powder: 1H NMR (CDCl3) δ 7.60 (d,
H-7, J ) 16 Hz), 7.12 (br s), 6.95 (br d, J ) 8 Hz), 6.90
(d, J ) 8 Hz), 6.81 (br s), 6.79 (d, J ) 8 Hz), 6.59 (dd, J
) 8, 2 Hz), 6.21 (d, H-8, J ) 16 Hz), 5.27 (dd, H-2′, J )
10, 8 Hz), 5.13 (dd, H-3′′, J ) 10, 3 Hz), 5.09 (dd, H-4′′,
J ) 10, 9 Hz), 5.04 (dd, H-2′′, J ) 3, 2 Hz), 4.96 (t, H-4′,
J ) 10 Hz), 4.88 (d, H-1′′, J ) 2 Hz), 4.39 (d, H-1′, J )
8 Hz), 4.19 (m, H-6′), 4.10 (m, HA-8a), 3.91 (t, H-3′, J )
10 Hz), 3.82 (dq, H-5′′, J ) 9, 7 Hz), 3.68 (m, H-5′), 3.57
(m, HA-8b), 2.75 (m, HA-7), 2.10 (s, 3 H), 2.08 (s, 3 H),

570 J ournal of Natural Products, 1998, Vol. 61, No. 5
Kernan et al.
2.04 (s, 3 H), 1.99 (s, 3 H), 1.92 (s, 3 H), 1.04 (d, H-6′′,
J ) 7 Hz); 13C NMR (CDCl3) δ 170.1 (s), 169.9 (s), 165.9
(s), 147.5 (s), 146.9 (s), 144.6 (s), 130.8 (s), 126.5 (s),
122.3 (d), 120.7 (d), 116.5 (d), 115.6 (d), 115.1 (d), 114.4
(d), 113.4 (d), 100.6 (d), 98.5 (d), 79.5 (d), 72.8 (d), 71.8
(d), 70.7 (d), 70.5 (t), 70.0 (d), 68.9 (d), 68.6 (d), 67.1 (d),
62.4 (d), 35.1 (t), 20.8 (q), 20.6 (q), 20.5 (q), 17.4 (q);
negative FABMS m/z 833.227 [M - H]-, 727, 582, 461,
389, 334.
scientists of Tanzania and other African countries for
their intellectual and scientific contributions to Sha-
man’s research program. We thank Dr. Raymond Coo-
per for help in preparing this manuscript.
Refer en ces a n d Notes
(1) King, S.; Tempesta, M. Ciba Found. Symp. 1994, 195, 197-213.
(2) King, S.; Carlson, T. Interscience 1995, 20, 134-139.
(3) Hall, C. B.; McCarthy, C. A. In Infectious Diseases; Mandell, G.
L., Bennett, J . E., Dolin, R., Eds.; Churchill Livingstone, Inc.:
New York, 1995; pp 1501-1519.
(4) Sendl, A.; Chen, J . L.; J olad, S. D.; Stoddart, C.; Rozhon, E.;
Kernan, M. R.; Nanakorn, W.; Balick, M. J . Nat. Prod. 1996,
59, 808-811.
(5) Ubillas, R.; J olad, S. D.; Bruening, R. C.; Kernan, M. R.; King,
S. R.; Sesin, D. F.; Barrett, M.; Stoddart, C. A.; Flaster, T.; Kuo,
J .; Ayala, F.; Meza, E.; Castan˜el, M.; McMeekin, D.; Rozhon,
E.; Tempesta, M. S.; Barnard, D.; Huffman, J .; Smee, D.; Sidwell,
R.; Soike, K.; Brazier, A.; Safrin, S.; Orlando, R.; Kenny, P. T.
M.; Berova, N.; Nakanishi, K. Phytomedicine 1994, 1, 77-106.
(6) Wyde, P.; Meyerson, L.; Gilbert, B. Drug Devel. Res. 1993, 28,
467-472.
(7) Endo, K.; Takahashi, K.; Abe, T.; Hikino, H. Heterocycles 1982,
19, 261-265.
(8) Cooper, R.; Solomon, P. H.; Kubo, I.; Nakanishi, K.; Shoolery,
J . N.; Occolowitz, J . L. J . Am. Chem. Soc. 1980, 102, 7953-7955.
(9) Zimin, L.; Zhongjian, J . Phytochemistry 1991, 30, 1341-1344.
(10) Andary, C.; Wylde, R.; Laffite, C.; Privat, G.; Winternitz, F.
Phytochemistry 1982, 21, 1123-1127.
(11) Kobayashi, H.; Karasawa, H.; Miyase, T.; Fukushima, S. Chem.
Pharm. Bull. 1984, 32, 3009-3014.
(12) Kobayashi, H.; Oguchi, H.; Takizawa, N.; Miyase, T.; Ueno, A.;
Usmanghani, K.; Ahmad, M. Chem. Pharm. Bull. 1987, 35,
3309-3314.
(13) Pettit, G. R.; Numata, A.; Takemura, T.; Ode, R. H.; Narula, A.
S.; Schmidt, J . M.; Cragg, G. M.; Pase, C. P. J . Nat. Prod. 1990,
53, 456-458.
Hyd r olysis of 10. Compound 10 (10.0 mg) was
hydrolyzed following the procedure used to prepare 16
to give the pentaacetate 17 (8.0 mg, 96%) as a colorless
oil: 1H NMR (CDCl3) δ 7.53 (d, H-7, J ) 16 Hz), 7.10
(br s), 6.84 (br s), 6.74 (br s), 6.51 (br d, J ) 8 Hz), 6.19
(d, H-8, J ) 16 Hz), 4.96-5.10 (m, 5 H), 4.79 (d, H-1′′,
J ) 2 Hz), 4.37 (d, H-1′, J ) 8 Hz), 4.28 (dd, H-6′a, J )
12, 2 Hz), 4.18 (dd, H-6′b, J ) 12, 5 Hz), 3.92 (m, H-5′′),
3.79 (m, HA-8), 3.77 (t, H-3′, J ) 10 Hz), 3.61 (m, H-5′),
2.71 (m, HA-7), 2.10 (s, 3 H), 2.08 (s, 3 H), 2.04 (s, 3 H),
1.99 (s, 3 H), 1.92 (s, 3 H), 1.11 (d, H-6′′, J ) 7 Hz); 13
C
NMR (CDCl3) δ 170.1 (CO), 170.0 (CO), 169.9 (CO),
169.7 (CO), 169.6 (CO), 167.0 (s), 147.7 (s), 146.0 (d),
144.9 (s), 142.8 (s), 130.2 (s), 126.4 (s), 122.5 (s), 120.6
(s), 116.3 (d), 115.5 (d, 2C), 114.1 (d), 113.8 (d), 100.7
(d), 99.2 (d), 81.0 (d), 71.9 (d), 71.0 (t), 70.4 (d), 69.6 (d),
68.7 (d), 67.3 (d), 62.2 (d), 58.6 (d), 35.2 (d), 21.0 (q),
20.8 (q), 20.7 (q), 20.6 (q), 17.1 (q); negative FABMS m/z
833 [M - H]-, 727, 670, 582, 492, 461, 429, 390, 337,
334.
(14) Kasai, R.; Ogawa, K.; Ohtani, K.; Ding, J .-K.; Chen, P.-Q.; Fei,
C.-J .; Tanaka, O. Chem. Pharm. Bull. 1991, 39, 927-929.
(15) Ko¨nig, B.; Dustmann, J . H. Naturwissenschaften 1985, 72, 659-
661.
(16) Cometa, F.; Tamassini, L.; Nicoletti, M.; Pieretti, S. Fitoterapia
1993, 44, 195-217.
(17) Cheminat, A.; Zawatsky, Y.; Becker, H.; Broullard, R. Phy-
tochemistry 1988, 27, 2787-2794.
Ack n ow led gm en t. The authors acknowledge Dr.
J ohn Kuo and Dr. Connie J ohn from the Spectroscopy
Department at Shaman Pharmaceuticals for conducting
NMR and mass spectral experiments, Dr. Shivinand
J olad for preparing some of the extracts, and Dr. J ohn
Huffman and Dr. Robert Sidwell at the Institute for
Antiviral Research, Utah State University, for conduct-
ing the in vitro anti-VZV assays. We also wish to
acknowledge and thank the traditional healers and
(18) Cavenaugh, P. R., J r.; Moskwa, P. S.; Donish, W. H.; Pera, P.
J .; Richardson, D.; Andreses, A. P. Invest. New Drugs 1990, 8,
347-354.
NP9703914