Antimycobacterial activities of dehydrocostus lactone and its oxidation products

Article abstract of DOI:10.1021/np970333i

In an attempt to study the structural dependence of antimycobacterial activity of the guaianolide dehydrocostus lactone and its derivatives, m- chloroperoxybenzoic acid oxidations of dehydrocostus lactone (1a) were performed. Three new monoepoxides, one previously synthesized diepoxide, and two new diepoxides were obtained. Two of the monoepoxides are C-10 epimers (3a, 3b), while the 4(15)-monoepoxide (2) has the 4α-O-configuration. The known diepoxide (4a) contains a C-10 α-epoxide and a β-epoxide at C-4. The diepoxides 4b and 4c, each with a C-4 α-epoxy group, differ in the configuration of the epoxide ring at C-10. Allylic oxidation of dehydrocostus lactone (1a) with selenium dioxide/tert-butyl hydroperoxide afforded the known 3-epizaluzanin C (1b). The relative configurations of compounds 1b-4c were established by 1D and 2D NMR techniques (¹H, ¹³C, COSY, NOESY, HMQC, and HMBC) as well as comparison with literature data. The molecular structures of lactones 1b, 4a, and 4c were determined by single-crystal X- ray diffraction. In radiorespirometric bioassays against Mycobacterium tuberculosis and Mycobacterium avium, dehydrocostus lactone (1a) exhibited minimum inhibitory concentrations of 2 and 16 μg/mL, respectively. In contrast, its monoepoxides (2, 3a, and 3b) and diepoxides (4a-c), as well as its hydrogenated derivatives and other analogues (1b, 1c, 5, and 6), showed significantly lower activities against M. tuberculosis.

Full text of DOI:10.1021/np970333i

© Copyright 1998 by the American Chemical Society and the American Society of Pharmacognosy
Volume 61, Number 10
October 1998
Full Papers
An tim ycoba cter ia l Activities of Deh yd r ocostu s La cton e a n d Its Oxid a tion
P r od u cts
Charles L. Cantrell,† Isabel S. Nun˜ez,† J ose´ Castan˜eda-Acosta,† Maryam Foroozesh,† Frank R. Fronczek,†
Nikolaus H. Fischer,*,† and Scott G. Franzblau‡
Department of Chemistry, Louisiana State University, Baton Rouge, Louisiana 70803-1804, and
G. W. L. Hansen’s Disease Center, P.O. Box 25072, Baton Rouge, Louisiana 70894
Received J uly 14, 1997
In an attempt to study the structural dependence of antimycobacterial activity of the guaianolide
dehydrocostus lactone and its derivatives, m-chloroperoxybenzoic acid oxidations of dehydrocostus lactone
(1a ) were performed. Three new monoepoxides, one previously synthesized diepoxide, and two new
diepoxides were obtained. Two of the monoepoxides are C-10 epimers (3a , 3b), while the 4(15)-
monoepoxide (2) has the 4R-O-configuration. The known diepoxide (4a ) contains a C-10 R-epoxide and
a â-epoxide at C-4. The diepoxides 4b and 4c, each with a C-4 R-epoxy group, differ in the configuration
of the epoxide ring at C-10. Allylic oxidation of dehydrocostus lactone (1a ) with selenium dioxide/tert-
butyl hydroperoxide afforded the known 3-epizaluzanin C (1b). The relative configurations of compounds
1b-4c were established by 1D and 2D NMR techniques (1H, 13C, COSY, NOESY, HMQC, and HMBC)
as well as comparison with literature data. The molecular structures of lactones 1b, 4a , and 4c were
determined by single-crystal X-ray diffraction. In radiorespirometric bioassays against Mycobacterium
tuberculosis and Mycobacterium avium, dehydrocostus lactone (1a ) exhibited minimum inhibitory
concentrations of 2 and 16 µg/mL, respectively. In contrast, its monoepoxides (2, 3a , and 3b) and
diepoxides (4a -c), as well as its hydrogenated derivatives and other analogues (1b, 1c, 5, and 6), showed
significantly lower activities against M. tuberculosis.
Worldwide, the number of tuberculosis cases is currently
on the rise, and it is estimated that new infections of
humans by Mycobacterium tuberculosis are greater than
8 million annually, and more than 3 million people will
die of this disease each year.1 This devastating pandemic
is worsened by the fact that strains of M. tuberculosis have
developed resistance to currently administered therapeutic
agents,2 creating a need for the discovery and development
of new and more effective antituberculosis drugs.
activity. Chemical investigation of active fractions resulted
in the isolation of epoxycycloartane-type triterpenes with
minimum inhibitory concentrations (MICs) of 8 µg/mL.3 In
a search for new chemotypes of antimycobacterial natural
products, in vitro tests against M. tuberculosis of a series
of natural and semisynthetic guaianolide-type sesquiter-
pene lactones were performed. Dehydrocostus lactone (1a ),
a major constituent of Saussurea lappa Clark (Aster-
aceae),4,5 gave MICs of 2 and 16 µg/mL against M.
tuberculosis and M. avium, respectively. It was therefore
of interest to learn whether epoxidation of the nonlactonic
double bonds of 1a would result in derivatives with
increased activity, because antimycobacterial triterpenes
from B. frutescens correlated well with the presence of an
epoxyprenyl group.3 Therefore, 1a was used as a starting
compound for various oxidative modifications in an attempt
In a previous paper we reported on bioassay-guided
fractionations of crude plant extracts of Borrichia frutescens
(L.) DC. (Asteraceae) with significant antimycobacterial
* To whom correspondence should be addressed. Tel: (504) 388-2695.
Fax: (504) 388-2695. E-mail: fischer@chemgate.chem.lsu.edu.
† Louisiana State University.
‡ G. W. L. Hansen’s Disease Center.
10.1021/np970333i CCC: $15.00
© 1998 American Chemical Society and American Society of Pharmacognosy
Published on Web 08/19/1998

1182 J ournal of Natural Products, 1998, Vol. 61, No. 10
Cantrell et al.
a
Ta ble 1. 300 MHz 1H NMR Spectral Data for Compounds 2-4c Performed in CDCl3
proton
1
2
3a
3b
4a
4b
4c
3.21 ddd
(8.1, 8.1, 8.1)
2.93 ddd
(4.0, 9.0, 9.0)
2.05 m
2.43 ddd
(8.1, 8.1, 8.1)
2.85 m
2.45 ddd
(8.7, 8.7, 8.7)
1.70 m
2.74 ddd
(8.1, 8.1, 8.1)
1.68 m
2a
2b
3a
1.91 m
2.11 m
1.69 m
1.6-1.85b
1.66 m
1.78 m
1.49 m
2.05 m
1.86 ddd
2.37 m
2.02 m
1.92 m
1.6-1.85b
1.60-1.78b
1.55-1.70b
(6.5, 9.0, 14.5)
3b
5
2.25 m
2.13 dd
2.26 m
2.64 dd
2.37 m
2.88 dd
2.14 m
2.69 dd
2.11-2.25b
2.12 m
2.02-2.12b
2.22 dd
(8.1, 11.0)
4.04 dd
(8.8, 11.0)
2.74 m
1.45 m
(9.6, 9.6)
4.08 dd
(9.6, 9.6)
2.86 m
1.42 dddd
(9.3, 9.3)
4.12 dd
(9.0, 10.2)
2.98 m
1.48 dddd
(6.1, 8.6, 11.3, 13.6)
2.27 dddd
(4.3, 6.0, 6.0, 13.6)
1.78 m
(9.9, 9.9)
4.10 dd
(9.7, 9.8)
2.85 m
2.23 dddd
(3.0, 3.0, 4.7, 13.7)
(8.8, 10.7)
4.23 dd
(9.4, 9.4)
2.74 m
6
4.13 dd
(8.8, 11.2)
2.99 m
7
8a
1.52 m
1.61 m
(5.0, 11.9, 11.9, 13.0)
2.22 m
8b
2.25 m
1.45 m
2.29 m
2.11 m
9a
9b
2.19 m
2.46 dd
(6.5, 12.7)
5.46 d
(3.4)
2.09 m
2.56 ddd
(3.5, 5.0, 14.6)
5.48 d (3.1)
1.62 m
2.10 m
1.70 m
2.02 m
1.55-1.70b
2.02-2.12b
1.91 m
13a
5.53 d (3.1)
5.53 d (3.1)
5.51 d (3.1)
5.50 d (2.9)
13b
14a
14b
15a
15b
6.17 d (3.6)
4.94 s
4.97 s
2.84 d (4.4)
3.36 d (4.5)
6.19 d (3.6)
2.75 d (4.9)
3.10 d (4.9)
5.00 s
6.26 d (3.1)
2.54 d (4.7)
2.62 dd (1.2, 4.5)
5.02 d (1.7)
5.26 d (1.8)
6.24 d (3.5)
2.55 d (4.3)
2.85 d (4.3)
2.80 d (4.4)
3.27 d (4.5)
6.22 d (3.6)
2.63 d (4.8)
2.68 d (4.8)
2.87 d (4.6)
3.33 d (4.4)
6.22 d (3.67)
2.60 d (4.8)
2.78 d (4.7)
2.86 d (4.4)
3.32 d (4.7)
5.00 s
a
b
Coupling constants (Hz) are given in parentheses. Overlap with other signals.
Ta ble 2. 75.4 MHz 13C NMR Spectral Data for Compounds
2-4c Performed in CDCl3
to find sesquiterpene lactones with increased antimyco-
bacterial activity.
a
carbon
2
3a
3b
4a
4b
4c
Resu lts a n d Discu ssion
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
46.8 d
46.6 d
46.5 d
43.8 db 45.1 db 44.4 d
28.4 tb 29.5 tb 25.5 tb 22.4 t
25.4 t
29.6 t
66.3 s
24.5 tb
31.1 t
66.8 s
52.8 d
81.6 d
Dehydrocostus lactone (1a ) was previously isolated from
costus root oil in our laboratory as described earlier.5 The
reaction of a solution of 1a in CH2Cl2 with 1.5 molar
equivalents of m-chloroperbenzoic acid (m-CPBA) gave a
mixture of reaction products that were separated by
vacuum-liquid chromatography (VLC), providing the mo-
noepoxides 3a , 2, and 3b in the less polar fractions,
followed by the more polar diepoxides 4a , 4b, and 4c.
Inspection of the 1H NMR spectrum of 3a indicated a
monoepoxide that differed from 1a in the absence of the
two olefinic methylene protons at C-14 with singlets at δ
4.82 and 4.90. Instead, two mutually coupled doublets
appeared at δ 2.75 and 3.10 (J ) 4.9 Hz), suggesting
epoxidation at the C-10(14) position. This was supported
by the 13C NMR spectrum of 3a , in which the two olefinic
carbon signals (C-10 and C-14) in 1a were replaced by
signals corresponding to the oxygen-bearing C-10 (δ 65.1,
s) and C-14 (δ 48.4, t) in 3a . The stereochemistry of the
epoxide ring was determined by NOESY spectroscopy with
correlations being observed between H-14b and H-5R,
suggesting a C-10 â-epoxide. Further 1H and 13C NMR
assignments were made by inspection of COSY, HMQC,
and HMBC spectra, and the chemical shifts obtained are
summarized in Tables 1 and 2, respectively.
31.4 tc 33.5 tc 32.2 tc 32.6 t
66.4 s 149.1 s 150.9 s
65.1 s
53.2 d
81.8 d
45.8 d
49.2 d
82.0 d
44.9 d
51.0 d
84.1 d
44.7 d
47.4 db 52.8 d
81.1 d
81.1 d
44.5 db 46.4 db 47.4 d
29.8 tb 31.5 tb 28.8 tb 27.5 t
25.6 t
30.7 t
58.3 s
24.7 tb
32.6 t
58.7 s
33.1 tc 38.5 tc 33.4 tc 38.1 t
148.1 s
65.1 s
58.1 s
56.8 s
139.3 s 139.3 s 139.7 s 138.5 s 139.6 s 139.4 s
169.8 s 170.4 s 170.2 s 169.7 s 169.8 s 169.8 s
120.2 t 120.6 t 120.8 t 120.9 t
114.1 t
50.1 t 113.2 t 109.8 t
121.0 t
50.3 tc
54.6 tc
120.5 t
50.7 t
53.8 t
48.4 t
51.6 t
48.1 tc
48.9 tc
a
Peak multiplicities were determined by DEPT experiments.
b,c Data in same column are interchangeable.
oxirane methylene doublets at δ 2.84 and 3.36 (J ) 4.5 Hz),
which were absent in 1a . The 13C NMR data also indicated
that the epoxide group in 2 could be placed between the
C-4(15) positions. The stereochemistry at C-4 was deter-
mined by NOESY spectroscopy, which indicated a strong
NOE correlation between H-15b and H-6â. Further sup-
port for a C-4R-epoxide was provided by reaction of 2 with
excess m-CPBA to produce diepoxides 4b and 4c. Because
the stereochemistry of 4c was determined by single-crystal
X-ray diffraction, this conversion unambiguously estab-
lished the R-epoxide stereochemistry at C-4(15) in 2. 1H
and 13C NMR assignments of 2 were based on inspection
of COSY, HMQC, and HMBC spectra and are summarized
in Tables 1 and 2, respectively.
In the 1H NMR spectrum of compound 4a , the olefinic
H-14 and H-15 signals present in 1a were replaced by two
sets of mutually coupled oxirane proton doublets at δ 2.55
and 2.85 (J ) 4.3 Hz) corresponding to H-14a and H-14b
and doublets (H-15a and H-15b) at δ 2.80 and 3.27 (J )
4.4 Hz), suggesting a C-10(14),C-4(15) diepoxide. The
relative stereochemistry of the epoxide groups was unam-
biguously determined by single-crystal X-ray diffraction.
1H and 13C NMR analysis of epoxide 3b also indicated
epoxidation at C-10(14) based on reasoning similar to that
described for 3a . The stereochemistry of 3b was confirmed
by NOESY spectroscopy, which showed interactions be-
tween H-6â and both epoxide protons at C-14, suggesting
a C-10 R-epoxide. The application of 2D COSY, HMQC,
and HMBC permitted all assignments in the 1H and 13C
NMR spectra, which are summarized in Tables 1 and 2,
respectively.
1
The H NMR spectrum of compound 2 indicated that it
differed from 1a in the absence of the olefinic methylene
protons at C-15. Instead, 2 exhibited mutually coupled

Antimycobacterials of Dehydrocostus Lactone
J ournal of Natural Products, 1998, Vol. 61, No. 10 1183
(14),C-4(15) isomers based on arguments made above for
4a . The epoxide stereochemistry of compound 4b was
determined by NOESY spectroscopy with correlations
observed between H-6â and H-14b as well as between H-6â
and H-15b, both indicating an R-orientation for the two
epoxide oxygens. The relative stereochemistry of com-
pound 4c was determined by single-crystal X-ray diffrac-
tion (Figure 1), which showed a C-4(15) R-epoxide and a
C-10(14) â-epoxide. Therefore, the epimers 4b and 4c
differed only in the configuration of the C-10(14) epoxide
group. X-ray coordinates for 4c are listed in Table 6 and
1H and 13C NMR spectral assignments by various 2D NMR
techniques for 4b and 4c are summarized in Tables 1 and
2, respectively.
3-Epizaluzanin C (1b) was prepared from 1a by a
previously described method7 using SeO2 and t-BuOOH in
1
CH2Cl2. The H and 13C NMR spectral data of 1b were in
full agreement with reported values.8 Single-crystal X-ray
crystallographic analysis of 1b (Figure 1) was performed,
and its atomic coordinates are listed in Table 4.
The crystal structures of 1b, 4a , and 4c are illustrated
in Figure 1. Lactone 1b is essentially isomorphous with
dehydrocostus lactone,9 which lacks the OH group at C-3,
despite the intermolecular hydrogen bonding by the OH
group. That hydrogen bond is linear and involves the
lactone carbonyl oxygen O-2 as acceptor, having an O‚‚‚O
distance of 2.858(2) Å and an angle about H of 179(3)°.
Although the conformations of the seven-membered and
lactone rings in 1b and dehydrocostus lactone (1a ) are
nearly identical, the conformations of their other five-
membered rings differ as a result of the C-3-OH substitu-
tion. While this ring in dehydrocostus lactone has a
distorted half-chair conformation, it has the envelope
conformation with C-5 in the flap position in 1b.
The conformation of 4c is similar to that of 1b, with a
twist-chair seven-membered ring having C-8 on the twist
axis and a half-chair lactone. Diepoxide 4a has two
independent molecules in the asymmetric unit. Their
conformations are nearly identical, but differ markedly
from that of 4c. Lactone 4a has its seven-membered ring
in a distorted twist-chair with C-5 on the twist axis, and
an envelope cyclopentane ring with C-4 at the flap position.
The MICs of the guaianolides were determined in ra-
diorespirometric bioassays against M. tuberculosis (H37Rv),
and 1a and 1b were also tested against M. avium.3,10 The
MICs against M. tuberculosis of dehydrocostus lactone (1a ),
its oxidative derivatives (2, 3a , 3b, and 4a -4c), and
hydroxylated analogues (1b, 1c, 5, and 6) are listed in
Table 7. The most active lactone 1a , with an MIC of 2 µg/
mL, is also the most lipophilic among this group of
guaianolides. It is evident that the presence of a hydroxyl
group at different positions of the guaianolide skeleton, as
shown for 1b, 1c, 5, and 6, significantly reduces activity
against M. tuberculosis. This might be due to the increase
in polarity, which possibly reduces transport through the
outer lipid layer of the mycobacterium.10
F igu r e 1. Molecular structure of compounds 1b, 4a , and 4c.
Its molecular structure is shown in Figure 1 and the X-ray
crystallographic atomic coordinates are given in Table 5.
The molecular structure indicated a C-4(15) â-epoxide,
while the C-10(14) epoxide oxygen is R-oriented. Com-
pound 4a has been previously prepared from 1a as a
mixture of C-10(14) epoxide epimers with an undefined
stereochemistry of the products.6 1H and 13C NMR spectra
of pure 4a were assigned by various 2D NMR techniques,
and the spectral data are reported in Tables 1 and 2,
respectively.
Because, in our previous study on cycloartane triterpenes
from Borrichia frutescens, the presence of an epoxide group
in the molecule significantly augmented the antitubercu-
losis activity,3 monoepoxide- and diepoxide derivatives of
dehydrocostus lactone (1a ) were tested for their activity.
Compared to 1a , the monoepoxides 2, 3a , and 3b showed
significantly decreased activity against M. tuberculosis with
MICs of 32, 32, and 64 µg/mL, respectively. The more polar
diepoxides 4a , 4b, and 4c exhibited MICs of 64, 128, and
128 µg/mL, respectively, supporting the correlation between
Examination of the 1H and 13C NMR spectra of diep-
oxides 4b and 4c indicated that they represented C-10-

1184 J ournal of Natural Products, 1998, Vol. 61, No. 10
Cantrell et al.
Ta ble 3. Crystal Data and X-ray Collection Parameters of 1b, 4a , and 4c
compound
formula
1b
15H18O3
246.3
4a
C15H18O4
4c
C15H18O4
262.3
C
FW
262.3
crystal system
space group
cell constants
a, Å
orthorhombic
P212121
orthorhombic
P212121
orthorhombic
P212121
7.650(1)
11.8057(7)
14.2153(6)
1283.8(4)
4
1.274
6.7
24
6.7140(5)
15.980(2)
24.931(3)
2674.9(9)
8
1.303
7.3
24
8.3193(6)
10.1704(8)
15.919(3)
1346.9(5)
4
1.293
7.3
22
b, Å
c, Å
V, Å3
Z
Dc, g cm-3
µ(Cu KR), cm-1
T
crystal size, mm
θ limits, deg.
octants collected
unique data
obsd data
criterion
0.11 × 0.13 × 0.50
0.18 × 0.38 × 0.60
0.15 × 0.27 × 0.55
2-75
2-75
2-75
(h, k, (l
2603
2427
I > 3σ(I)
236
h, k, l
3224
2509
I > 3σ(I)
344
0.091
(h, (k, (l
2763
2427
I > 1σ(I)
245
variables
R
0.031
0.045
Rw
0.040
0.107
0.049
resid density, e- Å-3
extinction
hydrogen atoms
0.23
0.66
0.27
3.8(3) × 10-6
1.7(2) × 10-6
2.7(1) × 10-6
refined iso
calcd
refined iso
Ta ble 4. Coordinates and Equivalent Isotropic Thermal
Parameters for 1b
atom
x
y
z
Beq (Å2)a
O-1
O-2
O-3
C-1
C-2
C-3
C-4
C-5
C-6
C-7
0.0798(1)
0.2231(1)
0.94141(7)
1.09870(8)
0.51947(9)
0.7666(1)
0.6747(1)
0.6345(1)
0.70896(9)
0.81357(9)
0.90910(9)
1.01729(9)
1.0285(1)
0.9187(1)
0.8494(1)
1.1091(1)
1.0548(1)
1.2180(1)
0.8583(2)
0.6811(1)
0.92791(7)
0.96659(8)
0.8894(1)
0.78196(9)
0.8548(2)
0.9026(1)
0.86393(9)
0.82126(8)
0.89195(8)
0.84877(9)
0.8588(1)
0.8368(1)
0.7588(1)
0.89462(8)
0.93343(9)
0.9046(1)
0.6717(1)
0.8619(1)
3.69(2)
4.89(2)
6.45(3)
3.57(2)
6.06(4)
3.97(3)
3.24(2)
2.94(2)
2.91(2)
3.04(2)
4.59(3)
4.77(3)
3.62(2)
3.29(2)
3.54(2)
4.38(3)
5.79(4)
4.12(3)
-0.0907(2)
-0.2437(2)
-0.2973(2)
-0.1310(2)
0.0138(2)
-0.0674(2)
-0.0930(2)
-0.1695(2)
-0.3672(2)
-0.4654(2)
-0.3871(2)
-0.0640(2)
0.0949(2)
C-8
C-9
C-10
C-11
C-12
C-13
C-14
C-15
-0.0961(2)
-0.4460(3)
0.1812(2)
a
Beq ) (8π2/3)∑i∑jUijai*aj*a i‚a j.
lipophilicity and antituberculosis activity within this gua-
ianolide series.
In conclusion, the above structure-activity data suggest
that the antimycobacterial activity of the tested guaiano-
lides is not as strongly determined by the exocyclic meth-
ylene lactone moiety as by polarity factors. Activity is
reduced when at least one hydroxyl group is present in the
molecule (compounds 1b, 1c, 5, and 6). The successive
replacement of double bonds with epoxide groups also
caused a significant loss of activity, with the more polar
diepoxides being less active than the monoepoxide ana-
logues. Because the most lipophilic guaianolide 1a gives
the highest antimycobacterial activity, it can be concluded
that activity is most strongly influenced by the polarity of
a given molecule. Therefore, within this structurally
related group of sesquiterpene lactones antimycobacterial
activity is most likely controlled by the transport through
the outer lipid layer of mycobacteria.
MHz spectrometer or a Bruker AM 400 MHz spectrometer.
MS were obtained on a Hewlett-Packard 5971A GC-MS. IR
spectra were run on a Perkin-Elmer 1760X spectrometer as
a film on KBr plates. VLC separations were carried out on
TLC grade Si gel (MN Kieselgel).11
m -CP BA Oxid a tion of Deh yd r ocostu s La cton e (1a ). A
solution of 980 mg of dehydrocostus lactone (1a ) in 20 mL of
CH2Cl2 was added to a solution of 1.5 molar equivalents
m-CPBA in 20 mL CH2Cl2 and stirred in an ice bath for 2 h.
The reaction mixture was washed twice with 40 mL of 0.01 M
NaOH solution and once with 40 mL of distilled H2O. TLC of
the reaction mixture revealed at least five products. Accord-
ingly, the crude mixture was adsorbed onto Si gel and
chromatographed on a VLC column (3.5 cm i.d.) using solvent
mixtures of hexane and EtOAc of increasing polarity. Fraction
7 (100 mL hexane-EtOAc, 88:12) contained 70 mg of 3a , and
Exp er im en ta l Section
Gen er a l Exp er im en ta l P r oced u r es. 1H and 13C NMR
spectra were recorded in CDCl3 on either a Bruker ARX 300

Antimycobacterials of Dehydrocostus Lactone
J ournal of Natural Products, 1998, Vol. 61, No. 10 1185
Ta ble 5. Coordinates and Equivalent Isotropic Thermal
Parameters for 4a
Ta ble 6. Coordinates and Equivalent Isotropic Thermal
Parameters for 4c
atom
x
y
z
Beq (Å2)a
atom
x
y
z
Beq (Å2)a
O-1A
O-2A
O-3A
O-4A
C-1A
C-2A
C-3A
C-4A
C-5A
C-6A
C-7A
C-8A
C-9A
C-10A
C-11A
C-12A
C-13A
C-14A
C-15A
O-1B
O-2B
O-3B
O-4B
C-1B
C-2B
C-3B
C-4B
C-5B
C-6B
C-7B
C-8B
C-9B
C-10B
C-11B
C-12B
C-13B
C-14B
C-15B
0.2190(6)
0.2238(8)
0.3605(8)
0.3271(9)
0.094(1)
0.052(1)
0.030(1)
0.155(1)
0.1081(9)
0.2498(9)
0.2090(9)
0.3402(9)
0.278(1)
0.271(1)
0.2355(9)
0.2267(9)
0.265(1)
0.459(1)
0.229(1)
0.7720(7)
0.7719(8)
0.9224(9)
0.838(1)
0.621(1)
0.595(2)
0.591(1)
0.713(1)
0.6541(9)
0.7904(9)
0.7388(9)
0.855(1)
0.785(1)
0.787(1)
0.7772(9)
0.7745(9)
0.813(1)
0.983(1)
0.794(2)
0.4585(3)
0.3525(3)
0.6486(3)
0.5685(3)
0.5504(3)
0.6452(4)
0.6832(4)
0.6263(4)
0.5395(3)
0.4712(3)
0.3858(3)
0.3713(4)
0.4257(4)
0.5183(4)
0.3269(4)
0.3743(4)
0.2452(4)
0.5651(5)
0.6496(4)
0.5216(3)
0.6165(4)
0.3467(3)
0.4558(4)
0.4560(4)
0.3635(5)
0.3129(5)
0.3648(4)
0.4543(4)
0.5206(4)
0.6105(4)
0.6368(4)
0.5910(5)
0.4988(5)
0.6588(4)
0.6023(4)
0.7399(5)
0.4563(7)
0.3340(5)
0.6081(1)
0.6657(2)
0.5576(2)
0.3922(2)
0.4695(2)
0.4611(3)
0.5155(3)
0.5512(2)
0.5322(2)
0.5502(2)
0.5241(2)
0.4751(2)
0.4279(2)
0.4392(2)
0.5693(3)
0.6191(2)
0.5686(3)
0.4375(3)
0.6047(3)
0.1331(1)
0.0675(2)
0.1984(2)
0.3558(2)
0.2752(2)
0.2893(3)
0.2402(3)
0.2018(2)
0.2123(2)
0.1919(2)
0.2097(2)
0.2599(3)
0.3107(3)
0.3064(2)
0.1601(3)
0.1156(2)
0.1558(4)
0.3123(3)
0.1524(3)
4.90(9)
7.6(1)
6.7(1)
7.1(1)
4.6(1)
6.9(2)
6.8(2)
5.6(2)
4.0(1)
3.8(1)
4.1(1)
4.5(1)
5.5(2)
4.5(1)
4.9(1)
5.3(1)
7.2(2)
7.0(2)
6.6(2)
5.26(9)
7.8(1)
7.8(1)
8.9(2)
5.6(2)
9.5(3)
8.5(2)
6.2(2)
4.4(1)
4.2(1)
4.3(1)
5.8(2)
6.9(2)
5.6(2)
4.9(1)
5.5(1)
8.7(2)
8.6(2)
8.3(2)
O-1
O-2
O-3
O-4
C-1
C-2
C-3
C-4
C-5
C-6
C-7
C-8
C-9
C-10
C-11
C-12
C-13
C-14
C-15
0.5714(2)
0.7116(2)
0.5560(2)
-0.0431(2)
0.2248(3)
0.2003(3)
0.3615(3)
0.4800(3)
0.3959(2)
0.4015(2)
0.3146(2)
0.1412(3)
0.0429(3)
0.0917(3)
0.4261(3)
0.5842(2)
0.3982(3)
0.0568(3)
0.6526(4)
0.7098(1)
0.8958(2)
0.3325(2)
0.5139(2)
0.4807(2)
0.3974(3)
0.3791(2)
0.4308(2)
0.5394(2)
0.6694(2)
0.7833(2)
0.8038(2)
0.6768(3)
0.5773(2)
0.8959(2)
0.8408(2)
1.0239(2)
0.6065(4)
0.4132(3)
0.39907(9)
0.3942(1)
0.3468(1)
0.2787(1)
0.3402(1)
0.4198(2)
0.4603(2)
0.3980(2)
0.3495(1)
0.3953(1)
0.3526(1)
0.3835(2)
0.3853(2)
0.3197(2)
0.3666(1)
0.3868(1)
0.3637(2)
0.2316(2)
0.4015(2)
4.80(3)
7.04(4)
8.42(4)
9.48(5)
5.54(5)
7.53(6)
6.49(6)
5.58(5)
4.35(4)
3.75(3)
4.07(3)
5.25(4)
6.40(6)
5.99(5)
4.46(4)
4.86(4)
6.55(6)
8.23(8)
8.58(7)
a
Beq ) (8π2/3)∑i∑jUijai*aj*a i‚a j.
Ta ble 7. Minimum Inhibitory Concentrations of Compounds
1a -6 against Mycobacterium tuberculosis
compound
MIC (µg/mL)
compound
MIC (µg/mL)
1a a
1b
1c
2
3a
3b
2
>128
>128
32
4a
4b
4c
5
64
128
128
50
32
64
6a
128
a
Lactones 1a and 6 were also tested against M. avium and gave
MICs of 16 and 128 µg/mL, respectively.
Deh yd r ocostu s la cton e, 4â(15),10r(14)-d iep oxid e (4a ):
colorless crystals (hexane-EtOAc); mp 97 °C; IR νmax 1765 (Cd
O), 1254, 996, 867 cm-1; H NMR, see Table 1; 13C NMR, see
1
Table 2; EIMS 70 eV m/z 262 [M]+ (1), 244 [M - H2O]+ (2),
231 (35), 214 (27), 203 (13), 185 (26), 173 (17), 159 (24), 105
(47), 91 (100), 79 (79), 67 (59), 53 (80).
a
Beq ) (8π2/3)∑i∑jUijai*aj*a i‚a j.
fraction 8 (hexane-EtOAc, 86:14) provided 138 mg of 2.
Fraction 9 (hexane-EtOAc, 84:16) gave 18 mg 3b, and fraction
11 (hexane-EtOAc, 4:1) contained 160 mg of crystalline 4a .
Fraction 14 (hexane-EtOAc, 3:2) contained a mixture of two
diepoxides that required further purification. Crude fraction
14 was adsorbed onto Si gel and rechromatographed on a VLC
column (2.3 cm i.d.) using hexane or hexane-EtOAc mixtures
of increasing polarity. Fraction 14-10 (50 mL hexane-EtOAc,
77:23) contained 33 mg of 4b, and fraction 14-11 (hexane-
EtOAc, 76:24) gave 20 mg of crystalline 4c.
Deh yd r ocostu s la cton e, 4r(15),10r(14)-d iep oxid e (4b):
gum; IR νmax 1763 (CdO), 1258, 732 cm-1; 1H NMR, see Table
1; 13C NMR, see Table 2; EIMS 70 eV m/z 262 [M]+ (1), 261
(3), 247 [M - CH3]+ (1), 244 [M - H2O]+ (2), 231 (14), 215 (7),
203 (19), 187 (13), 175 (20), 151 (29), 131 (33), 117 (44), 105
(47), 91 (100), 79 (67), 67 (54), 53 (58).
Deh yd r ocostu s la cton e, 4r(15),10â(14)-d iep oxid e (4c):
colorless crystals (hexane-EtOAc); mp 133-135 °C; IR νmax
1765 (CdO), 1258, 999 cm-1; 1H NMR, see Table 1; 13C NMR,
see Table 2; EIMS 70 eV m/z 262 [M]+ (1), 244 [M - H2O]+
(2), 231 (4), 215 (5), 203 (9), 151 (19), 131 (32), 117 (34), 105
(41), 91 (100), 79 (80), 67 (76), 53 (99).
Selen iu m Dioxid e Oxid a tion of Deh yd r ocostu s La c-
ton e (1a ). Oxidation was performed as previously described,
and the crude product was purified by preperative TLC to yield
pure 1b.7 The NMR data were essentially identical with
values reported in the literature,7 and the molecular structure
was unambiguously established by single-crystal X-ray dif-
fraction.
Com p ou n d s 1c, 5, a n d 6. The natural sesquiterpene
lactones 7R-hydroxydehydrocostus lactone (1c),12 micheliolide
(5),13 and pumilin (6)14 have been obtained previously in our
laboratory.
Deh yd r ocostu s la cton e, 4r(15)-ep oxid e (2): gum; IR νmax
1764 (CdO), 1637, 1258, 1138 cm-1; 1H NMR, see Table 1; 13
C
NMR, see Table 2; EIMS 70 eV m/z 246 [M]+ (12), 228 [M -
H2O]+ (5), 217 (10), 188 (20), 161 (24), 150 (37), 124 (64), 91
(100), 79 (67), 67 (43), 53 (81).
m -CP BA Oxid a tion of 2. Compound 2 (40 mg) was
dissolved in 5 mL of CH2Cl2 and added to 5 mL of CH2Cl2
solution containing 4 equivalents of m-CPBA. The reaction
mixture was placed in an ice bath and stirred until TLC
indicated that all 2 had reacted. The reaction was stopped
after 2 h and worked up as described previously. Crude
reaction products were chromatographed as indicated above
to yield 16 mg of 4b and 11 mg of 4c.
Deh yd r ocostu s la cton e, 10â(14)-ep oxid e (3a ): gum; IR
νmax 1763 (CdO), 1255, 996 cm-1; H NMR, see Table 1; 13C
X-r a y Cr ysta llogr a p h ic An a lyses.15 Intensity data were
collected for 1b, 4a , and 4c on an Enraf-Nonius CAD4
diffractometer with graphite-monochromated Cu KR radiation
(λ ) 1.54184 Å) by ω-2θ scans. Crystals of 4a were of
considerably lower quality than those of the other two com-
pounds, yielding broad diffraction peaks. The resulting struc-
ture determination is of lower precision, but is sufficient to
establish the molecular structure. Two octants of data were
collected for 4a and five octants for 4c. Data reduction
1
NMR, see Table 2; EIMS 70 eV m/z 246 [M]+ (2), 228 [M -
H2O]+ (7), 215 (44), 199 (9), 187 (12), 173 (14), 159 (21), 150
(99), 123 (41), 105 (45), 91 (100), 79 (76), 53 (84).
Deh yd r ocostu s la cton e, 10r(14)-ep oxid e (3b): gum; IR
νmax 1764 (CdO), 1254, 995 cm-1; H NMR, see Table 1; 13C
1
NMR, see Table 2; EIMS 70 eV m/z 246 [M]+ (2), 228 [M -
H2O]+ (3), 216 (13), 200 (8), 171 (13), 159 (12), 120 (36), 105
(40), 91 (69), 80 (100), 53 (47).

1186 J ournal of Natural Products, 1998, Vol. 61, No. 10
Cantrell et al.
(4) Romanuk, M.; Herout, V.; Sˇorm, F. Collect. Czech. Chem. Commun.
1956, 21, 894-901.
(5) Lu, T.; Fischer, N. H. Spectroscopy Lett. 1996, 29, 437-448.
(6) Kalsi, P. S.; Kaur, P.; Chhabra, B. R. Phytochemistry 1979, 18, 1877-
1878.
(7) Kalsi, P. S.; Kaur, G.; Sharma, S.; Talwar, K. K. Phytochemistry 1984,
23, 2855-2861.
(8) Kalsi, P. S.; Sharma, S.; Kaur, G. Phytochemistry 1983, 22, 1993-
1995.
(9) Fronczek, F. R.; Vargas, D.; Parodi, F. J .; Fischer, N. H. Acta Cryst.
1989, C45, 1829-1831.
(10) Connell, N. D.; Nikaido, H. In Tuberculosis: Pathogenesis, Protection,
and Control; Bloom, B. R., Ed.; American Society of Microbiology:
Washington, DC, 1994; pp 333-352.
(11) Coll, J . C.; Bowden, B. F. J . Nat. Prod. 1986, 49, 934-936.
(12) Fronczek, F. R.; Vargas, D.; Fischer, N. H.; Hostettmann, K. J . Nat.
Prod. 1984, 47, 1036-1039.
(13) Castan˜eda-Acosta, J .; Fischer, N. H.; Vargas, D. J . Nat. Prod. 1993,
56, 90-98.
included corrections for background, Lorentz, polarization,
decay (for 4a , 19%), and absorption (for 4c) effects. Absorption
corrections were based on ψ scans, which, for 4c, indicated no
change in intensity with rotation about the diffraction vector.
Crystals of 1b are isomorphous with dehydrocostus lactone;
all atoms but C-3 of that structure were used as a beginning
refinement model. Structures 4a and 4c were solved by direct
methods. Refinement was by full-matrix least squares, with
neutral-atom scattering factors and anomalous dispersion
corrections. All nonhydrogen atoms were refined anisotropi-
cally, while H atoms were treated as specified in Table 3.
Refinement of the absolute structures of 1b and 4c is consis-
tent with the expected absolute configurations. Crystal data,
details of data collection and refinement, and final agreement
indices for the three structures are given in Table 3.
Ra d ior esp ir om etr ic Bioa ssa ys. Bioassays were per-
formed essentially as described previously.3,16-18 Experiments
for M. avium were usually completed within 5 days and those
for M. tuberculosis in 10 days.
(14) Korp, J . D.; Bernal, I.; Fischer, N. H.; Leonard, C.; Lee, I. Y.; LeVan,
N. J . Heterocyclic Chem. 1982, 19, 181-187.
(15) Atomic coordinates, thermal parameters, bond distances and angles,
and observed and calculated structure parameters have been depos-
ited with the Cambridge Crystallographic Data Centre and can be
obtained upon request from Dr. Olga Kennard, University Chemical
Laboratory, 12 Union Road, Cambridge CB2 IEZ, UK.
(16) Inderleid, C. B.; Nash, K. A. In Antibiotics in Laboratory Medicine;
Lorian, V., Ed.; Williams and Wilkins: Baltimore, 1996; pp 127-
175.
(17) Inderleid, C. B.; Salfinger, M. In Manual of Clinical Microbiology;
Murray, P. R., Baron, E. J ., Pfaller, M. A., Tenover, F. C., Yolken, R.
H., Eds.; American Society of Microbiology: Washington, DC, 1995;
pp 1385-1404.
Ack n ow led gm en t. Financial support from the National
Institute of Allergy and Infectious Diseases, National Insti-
tutes of Health, under intra-agency agreement Y1-AI-5016, is
gratefully acknowledged. We thank Anita N. Biswas and Ken
E. White for technical assistance.
Refer en ces a n d Notes
(1) Sepkowitz, K. A.; Raffalli, J .; Riley, L.; Kiehn, T. E.; Armstrong, D.
Clin. Microbiol. Rev. 1995, 8, 180-199.
(18) Siddiqi, S. H. In Clinical Microbiology Procedures Handbook; Isen-
berg, H. D., Ed.; American Society for Microbiology: Washington, DC,
1992; pp 5.14.2-5.14.25.
(2) Cohn, D. L.; Bustreo, F.; Raviglione, M. C. Clin. Infect. Dis. 1997, 24
(Suppl. 1), S121-130.
(3) Cantrell, C. L.; Lu, T.; Fronczek, F. R.; Fischer, N. H.; Adams, L. B.;
Franzblau, S. G. J . Nat. Prod. 1996, 59, 1131-1136.
NP970333I