Modes of methyleneoxy bridging and their effect on tetrahydronaphthalene lignan cytotoxicity

Article abstract of DOI:10.1021/jm020158p

Dioxatricyclodecane, oxabicyclooctane, and benzodihydropyran derivatives of α-conidendrin (ACON), podophyllotoxin (PT), and sikkimotoxin (SK) were prepared to learn which methyleneoxy bridging modes and arene and aryl substituents coincided with high cytotoxicity. PT-derived dioxatricyclodecane 14 showed in vitro activity at 10^-8 M. SK analogue 12 was less active, and ACON analogue 11 was inactive at 10^-4 M. In vivo intraperitoneal and subcutaneous activities of 14 were observed. In vitro cytotoxicities were higher for oxabicyclooctanes when hydroxymethyl group and methyleneoxy bridge were cis, as in deoxypicropodophyllin analog 20, rather than trans, as in PT analogue 5. Acetylation of the hydroxymethyl group of 20 lowered activities, whereas acetylation of 5 increased or lowered activities. Reduction of the hydroxymethyl group of 5 to a methyl group increased cytotoxicities. Molecular dynamics indicated the THN scaffold of benzodihydropyrans was conformationally mobile, but scaffolds of oxabicyclooctanes and dioxatricyclodecanes were immobile. Each of three PT-benzodihydropyrans was less active than its oxabicyclooctane counterpart.

Full text of DOI:10.1021/jm020158p

1180
J . Med. Chem. 2003, 46, 1180-1190
Mod es of Meth ylen eoxy Br id gin g a n d Th eir Effect on Tetr a h yd r on a p h th a len e
Lign a n Cytotoxicity
Robert T. LaLonde,* Frank Ramdayal, Anatole Sarko, Koichi Yanai,§ and Mianji Zhang
Department of Chemistry, State University of New York, College of Environmental Science and Forestry,
Syracuse, New York 13210-2726
Received April 15, 2002
Dioxatricyclodecane, oxabicyclooctane, and benzodihydropyran derivatives of R-conidendrin
(ACON), podophyllotoxin (PT), and sikkimotoxin (SK) were prepared to learn which methyl-
eneoxy bridging modes and arene and aryl substituents coincided with high cytotoxicity. PT-
derived dioxatricyclodecane 14 showed in vitro activity at 10-8 M. SK analogue 12 was less
active, and ACON analogue 11 was inactive at 10-4 M. In vivo intraperitoneal and subcutaneous
activities of 14 were observed. In vitro cytotoxicities were higher for oxabicyclooctanes when
hydroxymethyl group and methyleneoxy bridge were cis, as in deoxypicropodophyllin analog
20, rather than trans, as in PT analogue 5. Acetylation of the hydroxymethyl group of 20 lowered
activities, whereas acetylation of 5 increased or lowered activities. Reduction of the hydroxy-
methyl group of 5 to a methyl group increased cytotoxicities. Molecular dynamics indicated
the THN scaffold of benzodihydropyrans was conformationally mobile, but scaffolds of oxabi-
cyclooctanes and dioxatricyclodecanes were immobile. Each of three PT-benzodihydropyrans
was less active than its oxabicyclooctane counterpart.
In tr od u ction
for engaging the preparation and cytotoxicity screening
of these bridged compounds were future prospects for
including additional THN lignans which, initially pos-
sessing â-pendant-ring stereochemistry, could be con-
verted to R-stereochemistry directly with resulting
cytotoxic activation without further hydrogenolyses and
oxolane formation. The foregoing prospects have been
examined. Additionally, the cytotoxicity of a group of
lignans having a conformationally mobile mode of
bridging represented by benzodihydropyrans has been
compared to the analogous group of conformationally
immobilized oxabicyclooctanes.
Among structural components most important for
high cytotoxicity of a tetrahydronaphthalene (THN)
lignan are configuration and substitution of the pendant
aryl group located at C-4, the benzhydrylic carbon of
podophyllotoxin (PT, 1, Scheme 1). This has been
demonstrated1 through SAR, which was facilitated by
oxidative2 methyleneoxy bridging of PT and dimethyl-
R-conidendrin (2) through their respective diols 3 and
4 to oxabicyclooctane derivatives 5 and 6. Succeeding
hydrogenolyses were directed through the use of the
appropriate reagents to proceed with retention or inver-
sion at C-4 of 5 or 6 and resulted in R- or â-stereochem-
istry for the pendant aryl group as required. Simulta-
neously, the THN was regenerated in 7 and 8, and
subsequently conformational rigidity of the THN scaf-
fold was restored by dehydration, giving trans-fused
oxolane derivatives 9 and 10, which replaced the
original lactones.
The C-4 configuration in PT, its oxabicyclooctane 5,
and the oxabicyclooctane 6 derived from R-conidendrin
(ACON) is the same. Conceivably, conformational im-
mobilization of the THN scaffold could be achieved
through an oxabicyclooctane just as it is by a trans-fused
lactone or oxolane, although molecular models reveal
the spacial relation of pendant to fused aromatic rings
attached to an oxabicyclooctane scaffold is somewhat
different from what it is in the corresponding lactone
or oxolane derivatives.3 These considerations prompted
our examination of methyleneoxy-bridging within oxa-
bicyclooctanes and dioxatricyclodecanes and its effect
on cytotoxicities. Both bridging modes immobilize the
conformation of the THN scaffold. An added incentive
Resu lts
Ch em istr y. ACON-derived diol 44 (Scheme 1) was
converted directly to dioxatricyclodecane 11 (Scheme 2)
in 38% yield by CuSO4/K2S2O8. Using the same oxidant
couple, a 36% yield of the sikkimotoxin (SK) analogue
12 was obtained directly from the corresponding diol
131. However, when this one-step procedure was applied
to 9-deoxypodophyllol, 3 (Scheme 1), the dioxatricyclo-
decane 14 dropped to 0.7% yield. The overall yield of
14 was 29% from a two-step process involving dehydra-
tion of podophyllol 15 (Scheme 2) to the known oxabi-
cyclooctane 165 and then treating the latter with the
oxidant couple in the usual manner to obtain 14.
However, 14 and oxabicyclooctanes 5 (Scheme 1) and
16 (Scheme 2), all required in smaller amounts for
initial cytotoxicity evaluations, could be obtained in the
respective yields of 3.7, 44, and 27% from a single
oxidation of diol 3 with DDQ. Oxabicyclooctane 5 was
converted to its acetate 17 (Scheme 1), which was
required for cytotoxicity comparisons. DDQ also con-
verted 9-deoxypicropodophyllol (picroPT) 18 (Scheme 3)
to a mixture of the isomeric pair of oxabicyclooctanes
19 and 20 in 10 and 20% yields, respectively. Likewise,
DDQ converted SK-based diol 13 to oxabicyclooctanes
* To whom correspondence should be sent. Tel: (315) 470-6823.
Fax: (315) 470-6856, e-mail: rtlalond@mailbox.syr.edu.
§ Present address: Nippon Paper Industries Co. Ltd.
10.1021/jm020158p CCC: $25.00 © 2003 American Chemical Society
Published on Web 02/22/2003

Tetrahydronaphthalene Lignan Cytotoxicity
J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 7 1181
Sch em e 1
alcohols. Also, the hydroxymethyl groups of PT-derived
5 and ACON-derived 6 were replaced by methyl groups
to learn the influence of both the C-3 hydroxy and
acetoxy groups on the cytotoxicity levels of oxabicyclo-
octanes lacking these more polar groups. Thus, the five-
step route summarized in Scheme 4, starting from
dimethyl ACON (2) and involving the four intermediates
24-27, produced target 28 in 29% overall yield.6
A
similar seven-step sequence started from 9-deoxy PT
(29, Scheme 4) proceeded through intermediates 30-
35 and yielded target 36 in 4% overall yield. Relative
to the preparation of 28, the two additional steps leading
to 36 included protection and deprotection of the C-1
hydroxyl group, which was required for final methyl-
eneoxy bridging to C-4.
Access to PT and SK benzodihydropyrans was through
precursor diols 39 and 42 (Scheme 5). Steps required
to obtain diol 39 included Pd/C-catalyzed hydrogenolysis
of the C-9 hydroxyl group of 9-epi-4′-demethyl PT (37)
followed by LAH reduction of the resulting lactone 38.
Preparation of the required SK-based diol 42 involved
demethylation at the C-4′ oxygen of 9-deoxysikkimo-
toxin, 40, followed by LAH reduction of the resulting
lactone 41. Phenolic diols 39 and 42 were oxidized by
sodium metaperiodate7 (Scheme 6) to their respective
o-quinones 43 and 44, which resulted through linking
of a C-3 hydroxyl group to a C-2′ position of a pendant,
o-quinone ring. The intermediate, PT-derived o-quinone
43 was isolated and characterized. Subsequent reduc-
tion of both quinone intermediates by sodium hydro-
sulfite in the presence of sodium hydrogen phosphate,
followed by treatment with dimethyl sulfate, provided
the corresponding PT and SK benzodihydropyrans 45
and 46. The PT benzodihydropyran alcohol 45 was
derivatized as its acetate, 47, and converted to its
methanesulfonate, 48. The last of these three was
treated with sodium iodide and zinc powder to obtain
49, the product in which the hydroxymethyl group of
the precursor 45 had been replaced as a methyl group.
Molecu la r Mod elin g. Molecular dynamics (MD)
simulations were carried out with dioxatricyclodecane
14, oxabicyclooctanes 36 and 16, and benzodihydropy-
ran 45, which were chosen as representative structures
resulting from various modes of methyleneoxy bridging.
The trajectories of 400 ps simulations for all structures
showed that a single skeletal molecular conformation
characterized the dioxatricyclodecanes and each of the
21 and isomeric 22, in yields of 32% and 43%, respec-
tively. Compound 20 was converted to its acetate 23.
The cytotoxicities of the PT and picroPT acetates would
be compared with those of their immediate precursor
Sch em e 2

1182 J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 7
LaLonde et al.
Sch em e 3
Sch em e 4
flexibility. The dioxatricyclodecane and oxabicyclooctane
conformations were stable and showed little torsional
variability. In contrast, the benzodihydropyran skeleton
underwent multiple transitions between four conform-
ers, all relatively close in potential energy but exhibiting
widely varying torsion angles in the nonaromatic por-
tions of the structure. The minimum-energy conforma-
tions of the dioxatricyclodecane and oxabicyclooctane
structures and the four benzodihydropyran conformers
are illustrated in Figure 1. Comparing these conforma-
tional features with the respective cytotoxicity levels
could indicate that conformational rigidity in the THN
scaffold in these compounds may be among the neces-
sary structural features that determine potential for
high levels of cytotoxicity.
Cytotoxicities, SAR, a n d Discu ssion . The anti-
proliferative activities of the eighteen dioxatricyclodec-
anes, oxabicyclooctanes, and benzodihydropyrans were
determined through the National Cancer Institute
(NCI), Cancer Drug Discovery and Development pro-
gram. Each compound was evaluated against approxi-
mately 55 cell lines of different tumor origins. The
resulting mean graph midpoint (MGM), delta, and range
values given in Table 1 allowed a preliminary, general
comparison of the 18 compound activities across all cell
lines. The MGM is calculated as the average GI50 for
all tested cell lines. It ranged in value from 10-4 to 10-8
M and is expressed in Table 1 as log GI50, which had
values of >-4 to <-8 M. The PT-derived dioxatricy-
clodecane, 14, was generally the most cytotoxic (MGM
-7.74 M) of the 18 compounds. Also, 14 had the lowest
delta value (0.26), which indicated relatively little
difference between this compound’s MGM and the log
GI50 value (<-8.00 M) observed for 49 of the 57 cell
lines. However, the relatively high range value (3.37)
for 14 pointed to a considerable activity difference in
the most and least sensitive of the 57 cell lines. Since
cytotoxicity is cell line sensitive, relative activities of
the 18 compounds were compared more specifically on
the basis of log GI50 values given in Table 2. All
compounds except 6, 11, and 22 were cytotoxic, with log
GI50 levels between >-4 and <-8 M. Dioxatricyclodec-
ane 14 was the most active of the 18 compounds in all
eight cell panels.8,9 In vivo activities of 5 and 14 were
appraised by the NCI fiber assay,10 which involves the
intraperitoneal (IP) and subcutaneous (SC) implanta-
two types of oxabicyclooctanes, whereas the benzodihy-
dropyrans exhibited a high degree of conformational

Tetrahydronaphthalene Lignan Cytotoxicity
J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 7 1183
Sch em e 5
Ta ble 1. Mean Cytotoxicities of Dioxatricyclodecane,
Oxabicyclooctane, and Benzodihydropyran Derivatives of THNa
Lignans toward Human Cancer Cells
seven cell lines for which activity data had been
obtained. A further replacement of the trimethoxyphen-
yl group of 12 by the dimethoxyphenyl group of ACON-
derived 11 resulted in inactivity for all cell lines for the
highest concentration level (10-4 M) used in routine
testing. The same two replacements of fused ring sub-
stituents and pendant rings in the oxabicyclooctane
series of 5, 21, and 6 resulted in a similar stepwise loss
of activity.
cytotoxicity (log GI50, M)b
compd
MGMc
deltad
rangee
cell linesf,g
5
6
-5.39
>-4.00h
>-4.00h
-4.38
-7.74
-4.63
-5.41
-4.49
-4.98
-4.32
>-4.00h
-4.91
-5.42
-6.07
-4.51
-4.70
-4.39
-4.42
1.14
2.53
58
47
51
56
57
53
58
57
56
56
11
12
14
16
17
19
20
21
22
23
28
36
45
46
47
49
0.41
0.26
1.15
0.92
1.09
0.73
0.44
0.79
3.37i
1.51
2.33
1.58
1.71
0.76
Table 2 data also reveals that the PT-derived dioxa-
tricyclodecane 14 is more active than its oxabicyclooc-
tane counterpart 5 by factors in excess of 100. However,
in MDA-MB-435, 14 is only 30 times more active than
5 for the reason that MDA-MB-435 is much more
sensitive than any of the other cell lines to 5, and to
several of the other compounds as well. Oxabicyclooc-
tanes 5 and 20 differ in the stereochemical configuration
of their hydroxymethyl groups. Oxabicyclooctane 20 was
approximately 8- to 70-times more active than 5 in cell
lines HOP-62, HCT-116, UACC-62, and OVCAR-3.
However, 20 appeared marginally less active than 5 in
SF-539 and DU-145, but was 25-times less active than
5 in MDA-MB-435. Also, MDA-MD-435 was the most
sensitive to 5 and the least sensitive to 20 than were
any of the other seven cell lines. Of the three additional
oxabicyclooctanes, including PT-derived 16, picropodo-
phyllin-derived 19, and SK-derived 22, only diastereo-
mers 16 and 19 were sufficiently active in preliminary
screening for subsequent 60 cell line evaluations. These
comparisons revealed that activities for both 16 and 19
were at the same low level (log GI50 -4.23 to -4.84 M)
in seven cell lines. However, the cytotoxicity levels
increased for MDA-MB-435, as indicated by log GI50
values of -5.41 for 16 and -5.17 for 19. The hydroxy-
methyl group and the methyleneoxy bridge have a cis
stereochemical relationship in both oxabicyclooctane 19
and its isomer 20. Activities were lower for 19 than 20
by factors ranging broadly from 6 to 700, except in MDA-
MB-435 where activities were virtually the same for
both oxabicyclooctanes.
3.09
0.49
1.01
0.78
1.01
0.65
0.52
4.00
1.21
3.08
1.08
1.52
1.03
0.82
48
56
51
58
58
59
58
a
b
Tetrahydronaphthalene. Cytotoxicity values are molar con-
centrations corresponding to 50% growth inhibition. c MGM is the
mean graph midpoint for growth inhibition of all human cancer
d
lines successfully tested. Delta is the logarithm difference be-
tween the MGM and the most sensitive cell line. e Range is the
logarithm difference between the log GI50 of the most resistant
and the most sensitive cell lines. f The several cell lines are
distributed among nine human cancer cell panels, which include
leukemia, melanoma, and nonsmall cell lung, colon, CNS, ovarian,
g
h
renal, prostrate, and breast cancers. See refs 13 and 14. Cell
lines giving log MGM values representing concentrations greater
than log MGM -7.74 M are HOP-92 (-4.91), SF-268 (-7.07),
MALME-3M (-7.08), SK-MEL-28 (-4.63), OVCAR-4 (-5.94),
i
OVCAR-5 (-4.63), and BT-549 (-7.03). Compound 22 was insuf-
ficiently active in the three cell line, one-dose preliminary assay
to warrant further testing in the 60 cell line assay.
tion of polyvinylidene fluoride hollow fibers in mice.
Each fiber contains the various cancer cell cultures, and
generally a minimum of 12 human cancer cell lines are
used. Compounds are introduced in solution by the IP
route. After termination of the four-day dosing routine,
the fibers are evaluated for cell mass. The respective
IP and SC scores for 5 were 14 and 2, while the
corresponding scores for 14 were 10 and 6. Net cell kill
was observed for both compounds.
Molecular modeling11 indicated insignificant differ-
ences between the oxabicyclooctane scaffold of 5 and the
same scaffold incorporated within the dioxatricyclodec-
ane 14. This indication coupled to the lower activity of
5 compared with that of 14, and the variable cell line
responses to 5 in relation to those of its diastereomer
Replacement of the methylenedioxy group of 14 by
two methoxy groups resulted in the sikkimotoxin (SK)-
derived 12 and at least a 1000-fold activity loss in the

1184 J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 7
LaLonde et al.
F igu r e 1. (A) Stereoviews of minimum-energy conformations for three types of methyleneoxy-bridged lignans. Top to bottom:
oxabicyclooctane 16, dioxatricyclodecane 14, and oxabicyclooctane 5. (B) Stereoviews of the four minimum energy conformations
of benzodihydropyran 45. Aromatic ring substituents are omitted.
Sch em e 6
20, suggested that cytotoxicities could be responding to
differing cell line interactions with the hydroxymethyl
group, which possibly could function as a hydrogen bond
donor or acceptor. The stereochemical relation of the
hydroxymethyl group to methyleneoxy bridge is trans
in 5, but cis in 20. Acetylation of the hydroxymethyl
group would allow the resulting ester to function only
as a hydrogen bond acceptor. Activities of acetate 17
were only 1.8 times higher or lower than those of its
precursor 5 in seven cell lines. In contrast conversion
of alcohol 20 to acetate 23 decreased activities by factors
ranging from 115 to 316 for HOP-62, HCT-116, and
UACC-62, but by only a factor of approximately 5 for
SF-539. However, this acetylation also increased activity
by a factor of near 4 for MDA-MB-435. Reduction of the
hydroxymethyl group of 5 to the methyl group in 36
removed a polar group from the oxabicyclooctane scaf-
fold and increased activity by factors ranging from
approximately 4 to 10 in five of the seven cell lines
allowing comparison. However, the previously noted
inversion of the hydroxymethyl group in 5, giving 20,
had enhanced activity by greater amounts and in more
cell lines than had the reduction of the hydroxymethyl
group in 5. Activity enhancements ranging from 8 (SN12
C) to 66 (MDA-MB-435) resulted from reducing the
hydroxymethyl group in ACON-derived 6 to the methyl
group of 28.
The group of benzodihydropyrans 45, 46, 47, and 49
differ structurally from the oxabicyclooctanes in having
a more conformationally mobile THN scaffold and a
more highly oxygen-substituted aromatic ring. However,
these four compounds share with the oxabicyclooctanes
the same THN substituents, which include hydroxy-
methyl, acetoxymethyl, and methyl groups. Overall, the

Tetrahydronaphthalene Lignan Cytotoxicity
J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 7 1185
Ta ble 2. Cytotoxicity of Methyleneoxy-Bridged Tetrahydronaphthalene Lignan Analogues
cytotoxicity (log GI50
melanoma ovarian
)
lung
colon
CNS
renal
SN12C
prostate
DU-145
breast
MDA-MB-435
compd
HOP-62
HCT-116
SF-539
UACC-62
OVCAR-3
5
6
-5.36
>-4.00
>-4.00
-4.15
-5.54
>-4.00
>-4.00
-4.55
-5.60
>-4.00
>-4.00
-4.50
-5.37
>-4.00
>-4.00
-4.16
-5.84
>-4.00
>-4.00
-4.52
a
-5.58
>-4.00
>-4.00
-4.53
-6.53
>-4.00
>-4.00
-4.71
>-4.00
11
12
14b
16
17
19
20
21
22c
23
28
36
45
46
47
49
a
-4.46
a
<-8.00
-4.72
<-8.00
-4.53
<-8.00
-4.81
<-8.00
-4.57
<-8.00
-4.84
<-8.00
-4.40
<-8.00
-5.41
-4.50
-5.30
-4.23
-5.32
-4.23
-5.37
-5.63
-5.76
-5.53
-5.67
-5.34
-6.33
-4.32
-4.52
-4.75
-4.74
-4.62
-4.52
-5.17
-6.66
-7.36
-5.41
-7.11
-6.74
-5.35
-5.13
-4.32
-4.52
-4.75
-4.20
-4.62
-4.52
-5.17
-4.60
-5.62
-6.10
-4.28
-4.22
-4.17
>-4.12
-4.86
-5.56
-6.36
-4.52
-4.65
-4.21
-4.50
-4.72
-5.71
d
-4.68
-4.61
-4.24
-4.68
-5.00
-5.16
-6.39
-4.59
-4.70
-4.50
-4.52
a
-4.59
-4.92
-5.27
-4.35
-4.60
-4.39
>-4.12
a
-5.73
-5.82
-7.08
-4.72
-4.74
-4.40
-4.94
-5.62
-6.73
-4.60
-4.82
-4.25
-4.48
-5.41
-6.32
-4.43
-4.64
-4.15
>-4.12
a
b
This particular cell line was lacking from the cell panel at the time of screening. Etoposide is reported (see ref 8) to give a log GI50
value of -5.2 in CAKI-1 (renal). Compounds 14, 5, and 46 give the respective log GI50 values of <-8.00, -5.24, and -4.60 in this cell line.
c Compound 22 was insufficiently active in the three cell line, one-dose preliminary assay to warrant further testing in the 60 cell line
d
assay. Two 60-cell line evaluations of compound 36 cytotoxicity resulted in widely divergent results for SF-539. Therefore, the log GI50
value was omitted from the table for this particular cell line.
Ta ble 3. Cytotoxicity Comparisons for Benzodihydropyrans 45,
46, 47, and 49 and Their Oxabicyclooctane Analogues 5, 21, 17,
and 36
arene substitution patterns of sikkimotoxin. In vitro
cytotoxicities of the PT-derived dioxatricyclodecane 14
were the highest at a log GI50 value of <-8 M in most
cytotoxicity (log GI50
)
cell lines. In vitro cytotoxicities for the dioxatricyclo-
decanes and oxabicyclooctanes were lowered through
replacement of the methylenedioxy group of the fused
aromatic ring by two methoxy groups and were further
lowered by replacement of one of two m-methoxy groups
by hydrogen in the pendant aromatic ring. These results
clearly indicated the requirement for particular aro-
matic substitution patterns in both fused and pendant
aromatic rings for highest cytotoxicity of oxabicyclooc-
tanes and dioxatricyclodecanes.
leukemia
SR
lung
NCI-322M
CNS
SNB-75
ovarian
OVCAR-4
compound
45
5
-5.29
-6.23
-4.74
-4.35
-4.66
-6.12
-4.72
-6.83
-4.32
-5.21
-5.06
-4.36
-4.77
-5.36
>-4.12
-5.67
-4.37
-5.57
-5.44
a
-4.61
-5.70
-4.91
-6.42
-4.40
-4.63
-5.71
-4.11
-4.32
-5.37
-4.43
-6.42
46
21
47
17
49
36
a
SNB-75 was lacking from the CNS cell panel when 21 was
Inversion of the hydroxymethyl group of PT-derived
oxabicyclooctane 5 resulted in mixed activity outcomes
for its diastereomer, 20, with some cells becoming more
sensitive to 20, and others less, when compared to 5.
Relevant to the configuration of the hydroxymethyl
group in 5 and 20 were the differing activity responses
to this group’s acetylation. While acetylation of 5
resulted in some very small activity changes, acetylation
of 20 lowered some activities by factors of 100 or more.
Furthermore, reduction of the hydroxymethyl group of
5 to the methyl group of 36 enhanced activities, al-
though the number of cell lines affected were fewer and
the activity increases were smaller than those resulting
from inversion of the hydroxymethyl group of 5. These
manifestations of activities responding to hydroxymeth-
yl group manipulation point to this group’s interactions
with the various cell lines. However, incorporation of
the hydroxymethyl group of oxabicyclooctanes 5 or 16
as a second methyleneoxy bridge in dioxatricyclodecane
14 enhanced cytotoxicities by factors of 10 to 1000 and
was responsible for the greatest activity enhancement
across all cell lines. This result might be attributed to
some gain in scaffold rigidity that is provided by double
bridging, the absence of the hydroxymethyl group, or a
combination of both. Although in vitro activities of
dioxatricyclodecane 14 and oxabicyclooctane 5 differed
considerably, their in vivo activities were similar.
screened.
four benzodihydropyrans produced were less active than
the other types of methyleneoxy-bridged compounds
except in cell line MDA-MB-435. The somewhat higher
MGM and delta values (Table 1) for 46 pointed to higher
sensitivities of other cell lines not included in Table 2.
Four such cell lines indicating sensitivities to test
compound at concentrations lower than 10-5 M were
found for 45 and 46. These cell lines were used in
comparing activities (Table 3) of all four benzodihydro-
pyrans to activities of their oxabicyclooctane counter-
parts. The comparison was deliberately biased in choos-
ing cell lines having the highest sensitivities to the
benzodihydropyrans. The pair wise comparison demon-
strated consistently higher activities for PT-derived
oxabicyclooctanes than PT-derived benzodihydropyrans.
However, the SK-derived benzodihydropyran 46 was
consistently more active than its oxabicylooctane coun-
terpart, 21, in three comparisons involving cell lines SR,
NCI-322M, and OVCAR-4.
Su m m a r y a n d Con clu sion s. Podophyllotoxin (PT)
and R-conidendrin can be converted through semisyn-
thesis that includes intramolecular methyleneoxy bridg-
ing to dioxatricyclodecanes, oxabicyclooctanes, and ben-
zodihydropyrans, including those having the aryl and

1186 J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 7
LaLonde et al.
(1), 110.56 (1), 109.48 (1), 87.30 (C-4), 80.90 (C-9), 64.25 (C-
3), 63.58 (C-1), 56.06 (C-3a), 55.94 (OCH3), 55.93 (OCH3), 55.88
(OCH3), 55.85 (OCH3), 49.69 (C-9a). HRMS [M+] Calcd for
C22H24O6: 384.1573; Found: 384.1572. Dioxa tr icyclod eca n e
12. Diol 13 (100 mg, 0.239 mmol, 36 mL CH3CN), CuSO4‚5H2O
(60 mg, 0.239 mmol, 4 mL H2O), and K2S2O8 (129 mg, 0.478
mmol, 11 mL H2O) refluxed 0.5 h gave, after processing and
Activities of the group of oxabicyclooctanes (5, 17, and
36), which included hydroxymethyl-, acetoxymethyl-,
and methyl-substituted members, were higher than
those of their benzodihydropyran counterparts, which
unlike the oxabicyclooctanes, lacked the measure of
immobilization of the THN scaffold, as indicated by
molecular dynamics. Also, unlike the dioxatricyclodec-
anes and oxabicyclooctanes, it was a SK-derivative (46)
that was the most active of the four-compound series of
benzodihydropyrans.
MPLC, 36 mg (36%) of 12 (glasslike solid): [R]25 +20.0° (c
D
0.5, CHCl3); HPLC tRI1 6.3, tRG1 7.0; 1H NMR (CDCl3) 7.13 (br,
1, H-2′ or H-6′), 6.81 (s, 1, H-8), 6.51 (s, 1, H-5), 6.43 (br, 1,
H-6′ or 2′), 4.89 (d, J ) 5.49, 1, H-9), 4.25 (dd, J ) 6.36, 9.35,
1, H-1), 4.10 (dd, J ) 6.24, 9.60, 1, H-3), 3.91 (s, 3, OCH3),
3.88 (s, 6, 2 × OCH3), 3.83 (d, J ) 9.66, 1, H-3), 3.77 (s, 3,
OCH3), 3.67 (s, 3, OCH3), 3.51 (d, J ) 9.25, 1, H-1), 3.20 (ddd,
J ) 1.19, 5.96, 6.70, 1, H-9a), 2.88 (t, J ) 6.14, 1, H-3a); 13C
NMR (CDCl3) 153.2 (C-3′, 5′) 149.4 (C-7), 149.3 (C-6), 137.3
(C-4′), 136.4 (C-1′), 128.5 (C-4a), 127.0 (C-8a), 110.8 (C-8), 110.6
(C-5), 103.1 (C-2′, 6′), 87.5 (C-4), 80.8 (C-9), 64.3 (C-3), 63.6
(C-1), 60.8 (OCH3), 56.1 (C-3a), 56.0 (OCH3), 55.9 (OCH3), 55.8
(OCH3), 49.6 (C-9a); HRMS [M+ + Na] Calcd for C23H26O7Na:
437.1576; Found: 437.1583. Also formed was oxabicyclooctane
Exp er im en ta l Section
Gen er a l. NMR data were obtained from Bruker 300 and
600 spectrometers and recorded in CDCl3 solution, unless
indicated otherwise. Chemical shift values (δ) are reported in
1
ppm and in relation to TMS (δ 0.00) and CDCl3 (δ 77.0) for H
and 13C NMR, respectively; J values are in hertz (Hz).
Quaternary, methine, methylene, and methyl carbons were
differentiated by DEPT and when unassigned to a specific
carbon are designated within parentheses as 0, 1, 2, and 3 in
association with 13C NMR δ values. 1H-1H correlation and one
bond 1H-13C connectivity were determined by COSY and
HMQC or HETCOR experiments, respectively, while multiple-
bond 1H-13C connectivity was determined by HMBC. HRMS
were determined by the Nebraska Center for Mass Spectrom-
etry, University of Nebraska, Lincoln, NE. Preparative TLC
was performed using 0.5-mm thickness silica gel plates
containing fluorescent indicator and were viewed under 254-
nm irradiation. Isocratic and gradient HPLC was determined
for each sample submitted for cytotoxicity evaluation. HPLC
employed a 5 µm, C18 (2), 250 × 4.60 mm column. Isocratic
HPLC used solvents MeOH/H2O (65/35) for compounds 5, 6,
11, 12, 14, 16, 19-23, 45, 46, and 49, but MeOH/H2O (80/20)
for 17, 28, 36, and 47. Retention times in minutes are
designated tRI1 and tRI2, respectively, for the two isocratic
solvent systems. A 15-min linear gradient HPLC used CH3-
CN/H2O (50/50-95/5) for compounds 5, 6, 11, 12, 14, 16, 19-
23, 45, 46, and 49, but CH3CN/H2O (70/30-95/5) for 17, 28,
36, and 47. Compound retention times are designated tRG1 and
21: 26 mg (26%); [R]25 +66.3° (c 0.3, CHCl3). Dioxa tr icy-
D
clod eca n e 14: Diol 31 (2.67 g, 6.63 mmol, 300 mL CH3CN),
CuSO4 ‚5H2O (1.66 g, 6.63 mmol, 100 mL H2O), and K2S2O8
(3.58 g, 13.26 mmol, 100 mL H2O) heated 0.5 h gave, after
processing and MPLC, 20 mg (0.7%) of 14 (crystalline solid
from ether): mp 158-160 °C; [R]25 -5.2° (c 0.59, acetone);
D
1
HPLC tRI1 9.5, tRG1 8.9; H NMR (CDCl3) 7.08 (br, 1, H-2′ or
H-6′), 6.76 (s, 1, H-8), 6.44 (s, 1, H-5), 6.39 (br, 1, H-6′ or H-2′),
5.94 (d, J ) 1.39, 1, OCH2O), 5.92 (d, J ) 1.39, 1, OCH2O),
4.85 (d, J ) 5.54, 1, H-9), 4.23 (dd, J ) 6.32, 9.36, 1, H-1),
4.12 (dd, J ) 9.61, 6.17, 1, H-3), 3.90 (br, 3, OCH3), 3.88 (br,
4, OCH3, H-3), 3.77 (br, 3, OCH3), 3.52 (d, J ) 9.35, 1, H-1),
3.20 (m, 1, H-9a), 2.89 (m, 1, H-3a); 13C NMR (CDCl3) 153.23
(C-5′ or C3′), 152.79 (C-3′ or C-5′), 148.07 (C-6 or C-7), 147.62
(C-7 or C-6), 137.28 (C-4′), 136.45 (C-1′), 130.43 (C-8a), 128.24
(C-4a), 108.26 (C-8), 107.96 (C-5), 103.25 (C-2′ or 6′), 102.78
(C-6′ or 2′), 101.20 (OCH2O), 87.42 (C-4), 80.87 (C-9), 64.40
(C-3), 63.49 (C-1), 60.87 (OCH3), 56.15 (OCH3 ), 55.69 (C-3a),
49.58 (C-9a). HRMS [M+] Calcd for C22H22O7: 398.1365;
Found: 398.1365. The major product from this reaction was
oxabicyclooctane 5: 1.06 g (40%).
t
RG2, respectively, for the two gradient solvent systems.
F or m a tion of Dioxa tr icyclod eca n e 14 fr om Oxa bicy-
cloocta n es 16 a n d 5. Podophyllol 15 (185 mg, 0.442 mmol)
was dehydrated by the method of Castro et al.5 giving 16 [153
Ma ter ia ls. Podophyllotoxin and R-conidendrin were ob-
tained respectively from Toronto Research Chemicals, Inc.
(Ontario, Canada) and Raisio Chemicals, Raisio, Finland.
mg (86.4%), mp 252-257 °C, [R]25 +18.3° (c 2.9, CHCl3)]. A
D
Gen er a l Rea ction , Extr a ction , a n d Sep a r a tion P r oce-
d u r es. Unless indicated otherwise, reactions were conducted
under dry N2, and organic reaction solvents were dried except
when a reagent required water as a solvent or the reaction
was quenched with water. Product extracts were dried over
anhyd MgSO4, and the solvents were removed under vacuum.
Products from the residue were separated and purified by silica
gel MPLC and/or preparative TLC.
warmed solution of 16 (150 mg, 0.374 mmol) in CH3CN (55
mL) was cooled to 45 °C, and to it was added a solution of
CuSO4‚5H2O (94 mg, 0.374 mmol) in H2O (6 mL) and K2S2O8
(203 mg, 0.749 mmol) in H2O (18 mL). The resulting mixture
was heated to reflux under N2 for 0.5 h. Processing and MPLC
(EtOAc/CH2Cl2, 1/10) gave 50 mg (0.125 mmol) of 14 (34%
yield) ([R]25 -28.9° (c 3.4, CHCl3)). To a solution of 312 mg
D
(0.779 mmol) of 5 in CH3CN (115 mL) was added a solution of
CuSO4‚5H2O (195 mg, 0.779 mmol) in H2O (13 mL) and K2S2O8
(422 mg, 1.55 mmol) in H2O (37 mL). The resulting mixture
was heated to reflux under N2 for 0.5 h, cooled to 25 °C, diluted
with H2O (60 mL), processed, and MPLC in the usual manner
to obtain 34 mg (11% yield) of 14.
Cu SO4/K2S2O8 Oxid a tion s of 1,4-Bu ta n ed iols to Dioxa -
tr icyclod eca n es: Dir ect Con ver sion s of Diols 4, 13, a n d
3, to th e Resp ective Dioxa tr icyclod eca n es 11, 12, a n d 14.
The CuSO4‚5H2O and K2S2O8 in H2O were added together in
one portion to the stirred diol in CH3CN solution. The mixture
was heated to reflux for 0.5 h, cooled to 25 °C, diluted with
H2O, and extracted with EtOAc. Dioxa tr icyclod eca n e 11:
Diol 4 (116 mg, 0.3 mmol, 45 mL CH3CN), CuSO4‚5H2O (75
mg, 0.3 mmol, 5 mL H2O), and K2S2O8 (116 mg, 0.6 mmol, 14
mL H2O) refluxed 0.5 h gave, after processing and MPLC (CH2-
Cl2/EtOAc, 3:1), 44 mg (38%) 11: mp 165-166 °C; [R]25D -6.07°
(c 2.5, acetone); HPLC tRI1 6.3, tRG1 6.7; 1H NMR (CDCl3) 7.24-
6.90 (br, 2, H-2′, 6′), 6.89 (d, J ) 7.72, 1, H-5′), 6.81 (s, 1, H-8),
6.48 (s, 1, H-5), 4.89 (d, J ) 5.51, 1, H-9), 4.24 (dd, J ) 9.33,
6.33, 1, H-1), 4.11 (dd, J ) 9.57, 6.24, 1, H-3), 3.913 (s, 3,
OCH3), 3.906 (s, 6, 2 × OCH3), 3.84 (d, J ) 9.74, 1, H-3), 3.66
(s, 3, OCH3), 3.52 (d, J ) 9.42, 1, H-1), 3.20 (m, 1, H-9a), 2.90
(t, J ) 6.10, 1, H-3a); 1H NMR (CDCl3, 333 K) 7.18 (br, 1) and
7.01 (br, 1) with collapse of 7.24-6.90 observed at 298 K; 13C
NMR (CDCl3) 149.36 (0), 149.20 (0), 148.80 (0), 148.38 (0),
133.47 (C-1′), 128.91 (C-4a), 126.87 (C-8a), 118.15 (1), 110.87
F or m a tion of Oxa bicycloocta n e 5 fr om 9-Deoxyp od o-
p h yllol 3. A solution of DDQ (597 mg, 2.63 mmol) in 80 mL
of CH2Cl2 was added to a solution of diol 3 (706 mg, 1.75 mmol)
in 200 mL of CH2Cl2. The dark blue solution was stirred under
N2 at 25 °C and became progressively lighter in color. TLC
indicated the absence of diol 3 (after 5.5 h). The solvent was
removed under vacuum, and the residue was dissolved in
EtOAc (100 mL). The solution was washed with aq 5%
NaHCO3 (120 mL), H2O (2 × 50 mL), and brine (30 mL). The
EtOAc extract was dried. The combined aq layers were
extracted with Et2O. Vacuum removal of solvents from com-
bined extracts left a residue, which by MPLC (EtOAc/CH2Cl2,
1/5-1/1) followed by preparative TLC (EtOAc/CH2Cl2, 1/1) gave
oxabicyclooctane 5 (306 mg, 44%): mp 144-146 °C, [R]25
D
+15.2° (c 1.2, CHCl3), HPLC tRI1 7.6, tRG1 6.4, Anal. (C22H24O7)

Tetrahydronaphthalene Lignan Cytotoxicity
J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 7 1187
C: calcd, 65.99, found, 65.63; H: calcd, 6.04, found, 6.03; 16
Cl2/EtOAc, 1/1) followed by preparative TLC (CH2Cl2/EtOAc,
(190 mg, 27%): mp 256-257 °C, [R]25 +21° (c 0.2, dioxane),
1/1). 19: mp 71-72 °C (glasslike solid); [R]25 +88.5° (c 1.1,
D
D
HPLC tRI1 5.7, tRG1 4.3; and dioxatricyclodecane 14 (26 mg,
3.7%). Oxa bicycloocta n e Aceta te 17. Ac2O (26 mg, 0.255
mmol) in dry THF (0.5 mL) was added to a solution of 5 (51
mg, 0.127 mmol) and DMAP (31 mg, 0.255 mmol) in dry THF
(0.5 mL). The resulting solution was stirred under N2 at 25
°C for 2 h. The liquid was evaporated at reduced pressure.
CHCl3); HPLC tRI1 6.6, tRG1 5.2; 1H NMR (CDCl3) 6.71 (s, 1,
H-8), 6.49 (s, 1, H-5), 6.20 (s, 2, H-2′, 6′), 5.96 (d, J ) 1.19, 1,
OCH2O), 5.90 (d, J ) 1.20, 1, OCH2O), 4.75 (s, 1, H-9), 4.10
(br, 1, H-4), 4.06 (dd, J ) 8.63, 5.94, 1, H-3), 3.83 (s, 3, OCH3),
3.76 (s, 6, 2 × OCH3), 3.71 (d, J ) 8.69, 1, H-3), 3.59 (dd, J )
8.91, 10.50, 1, H-1), 3.47 (dd, J ) 5.78, 10.68, 1, H-1), 2.48 (d,
J ) 5.56, 1, H-3a), 2.47 (dd, J ) 5.88, 8.69, 1, H-9a); 13C NMR
(CDCl3) 153.0 (C-3′, 5′), 147.6 (C-7), 146.4 (C-6), 140.4 (C-1′),
136.6 (C-4′), 133.8 (C-4a), 129.1 (C-8a), 110.9 (C-8), 107.7 (C-
5), 106.0 (C-2′, 6′), 101.0 (OCH2O), 78.1 (C-9), 71.0 (C-3), 62.5
(C-1), 60.8 (OCH3), 56.1 (OCH3), 53.6 (C-3a), 45.0 (C-4), 43.9
(C-9a); HRMS [M+] Calcd for C22H24O7: 400.1522; Found:
[MPLC] (CH2Cl2/EtOAc 15/1) gave 48 mg (85%) of 17: [R]25
D
+67.7° (c 0.9, CHCl3); HPLC tRI2 5.7, tRG2 6.4; 1H NMR (CDCl3)
7.14 (br, 1, H-2′ or 6′), 6.64 (s, 1, H-8 or 5), 6.41 (s, 1, H-5 or
8), 6.20 (br, 1, H-6′ or 2′), 5.89 (d, J ) 1.28, 1, OCH2O), 5.84
(d, J ) 1.28, 1, OCH2O), 4.38 (dd, J ) 5.36, 11.19, 1, H-3),
4.25 (ddd, J ) 2.41, 5.53, 8.06, 1, H-1), 4.17 (t, J ) 10.64, 1,
H-3), 3.88-3.81 (m, 10, 3 × OCH3, H-1), 3.20 (dt, J ) 2.73,
17.09, 1, H-9), 2.82 (m, 1, H-9a), 2.76 (dd, J ) 1.74, 17.37, 1,
H-9), 2.46 (m, 1, H-3a), 1.99 (s, 3, CH3); 13C NMR (CDCl3)
170.80 (CO), 147.14 (C-7), 145.69 (C-6), 137.03 (C-1′), 136.07
(C-4′), 132.05 (C-8a), 128.70 (C-4a), 109.00 (C-5 or 8), 107.68
(C-8 or 5), 104.06 (C-2′, 6′), 100.81 (OCH2O), 72.48 (C-1), 61.81
(C-3), 60.74 (OCH3), 56.12 (OCH3), 49.03 (C-3a), 36.82 (C-9a),
33.03 (C-9), 20.70 (CH3). HRMS [M+] Calcd for C24H26O8:
442.1628; Found: 442.1623.
400.1521. 20: mp 84-86 °C (glasslike solid); [R]25 -23.7° (c
D
1
0.9, CHCl3); HPLC tRI1 8.4, tRG1 6.5; H NMR (CDCl3) 7.03 (s,
1, H-2′), 6.64 (s, 1, H-8), 6.25 (s, 1, H-6′), 6.11 (s, 1, H-5), 5.87
(s, 1, OCH2O), 5.83 (s, 1, OCH2O), 4.23 (t, J ) 6.95, 1, H-1),
3.90 (s, 3, OCH3), 3.89 (s, 3, OCH3), 3.77 (t, J ) 4.06, 1, H-1),
3.75 (s, 3, OCH3), 3.48 (dd, J ) 5.21, 11.27, 1, H-3), 3.34 (dd,
J ) 7.30, 11.20, 1, H-3), 3.25 (d, J ) 15.65, 1, H-9), 2.90 (d,
J ) 15.65, 1, H-9), 2.88 (m, 1, H-9a), 2.58 (t, J ) 6.20, 1, H-3a),
1.49 (br, 1, OH); 13C NMR (CDCl3) 153.1 (C-3′ or 5′), 153.0 (C-
5′ or 3′), 147.1 (C-6), 145.7 (C-7), 137.0 (C-4′), 136.3 (C-4a),
135.1 (C-1′), 128.3 (C-8a), 109.0 (C-8), 107.3 (C-5), 104.6 (C-
2′) 104.0 (C-6′), 100.9 (OCH2O), 86.4 (C-4), 71.5 (C-1), 62.3 (C-
3), 60.9 (OCH3), 56.2 (OCH3), 56.1 (OCH3), 52.7 (C-3a), 38.5
(C-9), 37.5 (C-9a); HRMS [M+] Calcd for C22H24O7: 400.1522;
Found: 400.1534.
Oxa bicycloocta n e Aceta te 23. Ac2O (24 mg, 0.231 mmol)
in dry THF (0.5 mL) was added dropwise to 20 (42 mg, 0.105
mmol) and DMAP (26 mg, 0.210 mmol) in dry THF (1.0 mL),
and the resulting solution was stirred under N2 for 2 h. The
solvent was evaporated at reduced pressure to dryness, and
the residue was dissolved in ether (20 mL). The solution was
washed with 5% aq NaHCO3 (2 × 3 mL), 1 N aq HCl (5 × 3
mL), H2O (3 mL) and brine (3 mL) then dried (anhyd MgSO4).
Evaporation of solution at reduced pressure and MPLC of the
Similarly, treatment of 9-deoxysikkimol 13 (164 mg, 0.392
mmol) with DDQ (98 mg, 0.431 mmol) in 80 mL of CH2Cl2
gave 21 and 22. 21: (53 mg, 33%), [R]25D + 66.5° (c 3.4, CHCl3);
1
HPLC tRI1 4.8, tRG1 4.8; H NMR (CDCl3) 7.20 (br, 1, H-2′ or
H-6′), 6.68 (s, 1, H-8), 6.47 (s, 1, H-5), 6.21 (br, 1, H-6′ or H-2′),
4.27 (ddd, J ) 2.42, 5.69, 8.20, 1, H-1), 3.92 (dd, J ) 5.39, 10.90,
1, H-3), 3.88 (s, 6, 2 × OCH3), 3.86 (s, 3, OCH3), 3.82 (d, J )
8.27, 1, H-1), 3.74 (s, 3, OCH3), 3.72 (d, J ) 10.57, 1, H-3),
3.62 (s, 3, OCH3), 3.30 (dt, J ) 2.71, 17.04, 1, H-9), 2.90 (m, 1,
H-9a), 2.77 (dd, J ) 2.18, 17.10, 1, H-9), 2.32 (ddd, J ) 4.95,
5.00, 9.52, 1, H-3a), 1.64 (br, 1, OH); 13C NMR (CDCl3) 153.4
(C-3′ or C-5′), 152.0 (C-5′ or C-3′), 148.6 (C-6), 146.9 (C-7), 137.0
(C-1′), 136.8 (C-4′), 131.0 (C-4a), 127.9 (C-8a), 111.9 (C-5), 110.7
(C-8), 104.3 (C-2′ or C-6′), 104.0 (C-6′ or C-2′), 84.1 (C-4), 72.6
(C-1), 60.9 (OCH3), 60.0 (C-3), 56.2 (OCH3), 55.9 (OCH3), 55.8
(OCH3), 53.2 (C-3a), 36.7 (C-9a), 32.7 (C-9); HRMS [M+ + Na]
Calcd for C23H28O7Na: 439.17327; Found: 439.1737. 22: (72
residue (CHCl2/EtOAc, 10/1) gave 23 (35 mg, 75%). 23: [R]25
D
-27.0° (c 2.3, CHCl3); HPLC tRI1 18.9, tRG1 11.7; 1H NMR
(CDCl3) 6.99 (d, J ) 1.82, 1, H-2′), 6.63 (s, 1, H-8), 6.27 (d,
J ) 1.82, 1, H-6′), 6.09 (s, 1, H-5), 5.87 (d, J ) 1.37, 1, OCH2O),
5.84 (d, J ) 1.37, 1, OCH2O), 4.26 (dt, J ) 2.24, 8.34, 1, H-1),
3.94-3.88 (m, 1, H-1), 3.90 (s, 3, OCH3), 3.89 (s, 3, OCH3),
3.82-3.75 (m, 1, H-3), 3.76 (s, 3, OCH3), 3.71 (dd, J ) 10.18,
11.36, 1, H-3), 3.23 (d, J ) 16.56, 1, H-9), 2.90 (dd, J ) 2.43,
mg, 44%), [R]25 - 8.6° (c 1.8, CHCl3); HPLC tRI1 4.1, tRG1 3.7;
D
1H NMR (CDCl3) 6.68 (s, 1, H-8), 6.61 (s, 1, H-5), 6.37 (s, 2,
H-2′, 6′), 4.72 (d, J ) 4.78, 1, H-9), 4.45 (d, J ) 4.05, 1, H-4),
3.91-3.88 (m, 1, H-3), 3.89 (s, 3, OCH3), 3.85 (s, 3, OCH3), 3.77
(s, 6, 2 × OCH3), 3.79-3.73 (m, 1, H-1), 3.72 (s, 3, OCH3), 3.73-
3.68 (m, 2, H-1, 3), 2.76 (sep, J ) 4.56, 1, H-3a), 2.69 (dd, J )
4.49, 8.98, 1, H-9a); 13C NMR (CDCl3) 153.1 (C-3′, 5′), 148.8
(C-7), 147.8 (C-6), 138.9 (C-1′), 136.7 (C-4′), 131.1 (C-4a), 128.0
(C-8a), 112.8 (C-5), 110.7 (C-8), 106.6 (C-2′, 6′), 77.7 (C-9), 68.7
(C-3), 60.8 (OCH3), 60.2 (C-1), 56.2 (OCH3), 56.1 (OCH3), 55.8
(OCH3), 48.1 (C-3a), 47.0 (C-4), 43.3 (C-9a); HRMS [M+ + Na]
Calcd for C23H28O7Na: 439.17323; Found: 439.1731.
16.71, 1, H-9), 2.84-2.76 (m, 2, H-3a, 9a), 2.00 (s, 3, CH3); 13
C
NMR (CDCl3) 170.94 (CO), 152.89 (C-3′ or 5′), 152.77 (C-5′ or
3′), 147.04 (C-7 or 6), 145.60 (C-6 or 7), 136.70 (C-4′), 136.05
(C-4a), 134.58 (C-1′), 127.80 (C-8a), 108.75 (C-8), 107.27 (C-
5), 104.56 (C-2′), 103.72 (C-6′), 100.80 (OCH2O), 85.89 (C-4),
70.97 (C-1), 63.40 (C-3), 60.81 (OCH3), 56.07 (OCH3), 55.98
(OCH3), 44.29 (C-3a), 37.97 (C-9), 36.91 (C-9a), 20.80 (CH3);
HRMS [M+] Calcd for C24H26O8: 442.1628; Found: 442.1624.
P r ep a r a tion of 4′-Dem eth yld eoxyp od op h yllotoxin , 38.
4′-Demethyl-9-epipodophyllotoxin (37) (200 mg, 0.50 mmol) in
15 mL of glacial HOAc at 95 °C was stirred 5.5 h under H2 (1
atm) with 200 mg of Pd/C (10%). After the catalyst was filtered
off and the solvent was removed in a vacuum, the residue was
recrystallized from CH3OH/THF to give 38,13 120 mg (62%).
4′-Dem eth yl-9-d eoxyp od op h yllol 39 by Red u ction of
4′-Dem eth yld eoxyp od op h yllotoxin 38. 4′-Demethyldeoxy-
podophylotoxin 38 (400 mg, 1.04 mmol) in 35 mL of dry THF
was treated with LiAlH4 (320 mg, 8.32 mmol) gave 39 (333
mg, 82%).
F or m a tion of Ben zod ih yd r op yr a n 45. A solution of diol
39 (100 mg, 0.258 mmol) in (CH3)2CO (20 mL) was added to a
stirred solution of NaIO4 (124 mg, 0.568 mmol) in H2O (40 mL),
and stirring was continued at 25 °C for 6 h. Brine (10 mL)
was added, and the mixture was extracted with CHCl3 (5 ×
30 mL). The combined organic phase was washed with H2O
(2 × 20 mL), dried over MgSO4, and evaporated to dryness.
Preparative TLC of one-third of the residue gave red solid
quinone (15 mg) 43: 1H NMR (CDCl3) 6.62 (s, 1, H-8), 6.38 (s,
Oxa bicycloocta n e 6. This compound was prepared as
previously described and exhibited physical properties, includ-
ing 1H and 13C NMR, consistent with those reported.1 Anal.
(C22H26O6) C: calcd, 68.38, found, 67.54; H: calcd, 6.78, found,
6.90.
F or m a tion of Dioxa tr icyclod eca n e 12 fr om Oxa bicy-
cloocta n e 21. Oxabicyclooctane 21 (53 mg, 0.127 mmol)
dissolved in acetonitrile (20 mL) was refluxed with CuSO4‚
5H2O (34 mg, 0.134 mmol) in H2O (2 mL) and K2S2O8 (73 mg,
0.269 mmol) in water (6 mL) for 0.5 h to give 28 mg of
oxatricyclooctane 12 (53%); [R]25 +21.6° (c 1.3, CHCl3).
D
F or m a tion of Dioxa tr icyclod eca n e 12 fr om Oxa bicy-
cloocta n e 22. Oxabicyclooctane 22 (70 mg, 0.168 mmol, 28
mL CH3CN) was oxidized by CuSO4‚5H2O (43 mg, 0.168 mmol,
3 mL of H2O) and K2S2O8 (91 mg, 0.336 mmol, 9 mL of H2O),
43 mg of dioxatricyclodecane 12 was obtained (60%); [R]25
+19.6° (c 3.4, CHCl3).
D
F or m a t ion of Oxa b icyclooct a n es 19 a n d 20 fr om
9-Deoxyp icr op od op h yllol 18. DDQ (85 mg, 0.372 mmol) and
1812 (100 mg, 0.248 mmol) in CH2Cl2 (120 mL) solution were
heated to reflux under N2 for 22 h to give 195 (10 mg, 10%)
and 20 (20 mg, 20%) which were separated by MPLC (CH2-

1188 J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 7
LaLonde et al.
1, H-5), 5.88 (d, J ) 1.41, 1, OCH2O), 5.84 (d, J ) 1.42, 1,
OCH2O), 5.77 (s, 1, H-6′), 4.16 (ddd, J ) 1.91, 4.07, 10.98, 1,
H-3), 3.80 (s, 3, OCH3), 3.78-3.72 (m, 3, H-1, 4), 3.52 (t, J )
11.50, 1, H-3), 2.67 (dd, J ) 5.13, 14.46, 1, H-9), 2.46 (dd, J )
10.94, 14.32, 1, H-9), 2.26 (dt, J ) 4.86, 16.80, 1, H-3a), 1.86-
1.68 (br, 1, OH), 1.40-1.34 (m, 1, H-9a); 13C NMR (CDCl3)
176.2 (C-5′), 175.7 (C-4′), 167.7 (C-2′), 153.5 (C-3′), 146.3 (C-
6), 146.0 (C-7), 130.6 (C-4a), 129.9 (C-8a), 108.9 (C-1′), 108.6
(C-5), 108.0 (C-8), 106.8 (C-6′), 100.7 (OCH2O), 69.8 (C-3), 66.1
(C-1), 56.1 (OCH3), 38.6 (C-9a), 34.1 (C-3a), 31.0 (C-4), 30.4
(C-9); HRMS [M+] Calcd for C20H20O7 (reduced form): 372.1209;
Found: 372.1210. The remaining 2/3 of the residue was
redissolved in CHCl3 (30 mL) and stirred with aq (15 mL)
Na2S2O4 (164 mg, 0.80 mmol) and Na2HPO4 (60 mg, 0.42
mmol) for 0.5 h following a known method.14 The CHCl3 phase
was separated, and the aq phase was extracted with CHCl3
(3 × 10 mL). The combined CHCl3 phase was dried over
MgSO4. Evaporation of the CHCl3 gave a residue, which was
methylated with (CH3)2SO4 to afford 45 (21 mg, three-step
(C-2′), 145.80 (C-7), 145.40 (C-6), 135.36 (C-4′), 132.76 (C-4a),
131.35 (C-8a), 108.82 (C-5), 107.58 (C-1′), 107.52 (C-8), 100.46
(OCH2O), 95.65 (C-6′), 66.83 (C-3), 60.93 (OCH3), 60.89 (OCH3),
55.81 (OCH3), 39.22 (C-3a), 36.51 (C-9), 32.88 (C-4), 31.57 (C-
9a), 21.55 (C-1); HRMS [M+] Calcd for C22H24O5: 384.1573;
Found: 384.1580.
P r ep a r a tion of 4′-Dem eth yl-9-d eoxysik k im otoxin , 41.
According to a known procedure,10 treating 9-deoxysikkimo-
toxin 40 (550 mg, 1.33 mmol) with 30% HBr-HOAc (3 mL) in
CH2ClCH2Cl (25 mL) for 13.7 h gave 41 (338 mg, 64%).
P r epa r a tion of 4′-Dem eth yl-9-deoxysikk im ol, 42. Treat-
ing 4′-demethyl-9-deoxysikkimotoxin (41, 338 mg, 0.844 mmol)
in dry THF (25 mL) with LiAlH4 (257 mg, 6.752 mmol) gave
42 (210 mg, 62%).
F or m a tion of Ben zod ih yd r op yr a n 46. A solution of diol
42 (208 mg, 0.514 mmol) and NaIO4 (242 mg, 1.131 mmol) in
(CH3)2CO/H2O (40/80 mL) was stirred at 25 °C for 9 h.
Processing the reaction mixture in the manner previously
described for conversion of 39 to 43 followed by MPLC (CH2-
Cl2/EtOAc, 1/1) gave red, solid quinone (100 mg, 50%) 44: 1H
NMR (CDCl3) 6.59 (s, 1, H-8), 6.40 (s, 1, H-5), 5.75 (s, 1, H-6′),
4.16 (d, J ) 9.07, 1, H-3), 3.96 (br, 1, OH), 3.78 (s, 3, OCH3),
3.77 (s, 3, OCH3), 3.69 (s, 3, OCH3), 3.82-3.52 (m, 4, H-1, 3,
4), 2.65 (dd, J ) 5.51, 15.22, 1, H-9), 2.44 (dd, J ) 8.84, 15.08,
1, H-9), 2.32-2.24 (m, 1, H-3a), 1.53-1.48 (m, 1, H-9a); 13C
NMR (CDCl3) 176.2 (C-5′), 175.9 (C-4′), 167.9 (C-2′), 153.3 (C-
3′), 147.6 (C-6), 147.5 (C-7), 129.0 (C-4a), 128.0 (C-8a), 112.0
(C-8), 111.4 (C-5), 109.5 (C-1′), 106.9 (C-6′), 69.8 (C-3), 65.5
(C-1), 56.2 (OCH3), 56.0 (OCH3), 55.9 (OCH3), 37.7 (C-9a), 33.3
(C-3a), 30.0 (C-4), 29.2 (C-9). The quinone 44 (100 mg, 0.259
mmol) in CHCl3 (30 mL) and an aq solution of Na2S2O4 (477
mg, 2.329 mmol) and Na2HPO4 (184 mg, 1.295 mmol) in H2O
(8 mL) were stirred rapidly for 0.5 h. The resulting catechol
was methylated in (CH3)2CO (30 mL) with dimethyl sulfate
yield from 4′-demethyl-9-deoxypodophyllol, 39, 30%). 45: [R]25
D
-126.6° (c 4.6, CHCl3); HPLC tRI1 5.5, tRG1 4.7; 1H NMR (CDCl3)
6.67 (s, 1, H-8), 6.29 (s, 1, H-5), 6.21 (s, 1, H-6′), 5.86 (s, 2,
OCH2O), 3.87 (s, 3, OCH3), 3.87-3.83 (m, 2, H-3, 4), 3.83 (s,
6, 2 × OCH3), 3.83-3.76 (m, 2, H-1), 3.26 (d, J ) 10.44, 11.76,
1, H-3), 2.75 (dd, J ) 4.99, 13.97, 1, H-9), 2.54 (dd, J ) 12.10,
13.53, 1, H-9), 2.20-2.16 (m, 1, H-3a), 1.65-1.25 (m, 1, H-9a);
13C NMR (CDCl3) 153.4 (C-5′), 151.9 (C-3′), 150.6 (C-2′), 146.0
(C-7), 145.6 (C-6), 135.4 (C-4′), 133.3 (C-4a), 130.3 (C-8a), 108.8
(C-5), 107.8 (C-8), 107.4 (C-1′), 100.5 (OCH2O), 95.7 (C-6′), 67.1
(C-3), 66.6 (C-1), 60.94 (OCH3), 60.91 (OCH3), 55.8 (OCH3),
39.6 (C-9a), 34.6 (C-3a), 33.2 (C-4), 31.2 (C-9); HRMS [M+]
Calcd for C22H24O7: 400.1522; Found: 400.1531.
Ben zod ih yd r op yr a n Aceta te 47. Ac2O (33 mg, 0.32 mmol)
in dry THF (0.5 mL) was added to 45 (64 mg, 0.16 mmol) and
DMAP (39 mg, 0.32 mmol) in THF (1 mL). The resulting
solution was stirred under N2 at 25 °C for 2 h. The residue
from evaporation of THF at reduced pressure was dissolved
in ether (30 mL). The mixture was washed with aq NaHCO3
(5%) (2 × 5 mL), 1 N HCl (2 × 5 mL), water (5 mL), and brine
(5 mL). The organic phase was dried (anhyd MgSO4). The ether
was evaporated at reduced pressure. MPLC (CH2Cl2/EtOAc,
(300 mg, 2.38 mmol) gave 46 (60 mg, 56%): [R]25 -242.3° (c
D
1.1, CHCl3); 1H NMR (CDCl3) 6.70 (s, 1, H-8), 6.36 (s, 1, H-5),
6.22 (s, 1, H-6′), 3.90 (d, J ) 5.44, 1, H-4), 3.87 (s, 3, OCH3),
3.85 (s, 3, OCH3), 3.83 (s, 3, OCH3), 3.82 (m, 1, H-3), 3.81 (s,
3, OCH3), 3.81-3.68 (m, 2, H-1), 3.68 (s, 3, OCH3), 3.29 (t, J )
11.18, 1, H-3), 2.76 (dd, J ) 5.15, 14.18, 1, H-9), 2.56 (dd, J )
11.25, 13.96, 1, H-9), 2.22 (m, 1, H-3a), 1.72 (br, 1, OH), 1.35
(m, 1, H-9a); 13C NMR (CDCl3) 152.1 (C-5′), 150.9 (C-3′), 149.5
(C-2′), 146.1 (C-7 or 6), 146.0 (C-6 or 7), 134.1 (C-4′), 130.5
(C-4a), 127.8 (C-8a), 110.9 (C-5), 109.7 (C-8), 106.5 (C-1′), 94.6
(C-6′), 66.1 (C-3), 65.4 (C-1), 59.8 (OCH3), 59.7 (OCH3), 54.9
(OCH3), 54.7 (OCH3), 54.6 (OCH3), 38.3 (C-9a), 33.3 (C-3a),
31.5 (C-4), 29.4 (C-9); HRMS [M+ + Li] Calcd for C23H28O7Li:
423.1995; Found: 423.2001.
20/1) gave pure 47 (60 mg, 85%): [R]25 -122.2° (c 1.04,
D
CHCl3); HPLC tRI2 9.4, tRG2 8.5; 1H NMR (CDCl3) 6.67 (s, 1,
H-8), 6.29 (s, 1, H-5), 6.22 (s, 1, H-6′), 5.87 (s, 2, OCH2O), 4.24-
4.17 (m, 2, H-1), 3.88 (s, 3, OCH3), 3.93-3.86 (m, 1, H-4), 3.86-
3.82 (m, 1, H-3), 3.84 (s, 6, 2 × OCH3), 3.25 (dd, J ) 10.78,
11.63, 1, H-3), 2.70 (dd, J ) 4.99, 14.02, 1, H-9), 2.54 (dd, J )
11.97, 13.81, 1, H-9), 2.23-2.18 (m, 1, H-3a), 2.10 (s, 3, CH3),
1.43-1.38 (m, 1, H-9a); 13C NMR (CDCl3) 171.1 (CO), 153.4
(C-5′), 151.9 (C-3′), 150.4 (C-2′), 146.1 (C-7), 145.6 (C-6), 135.4
(C-4′), 133.0 (C-4a), 129.7 (C-8a), 108.8 (C-5), 107.7 (C-8), 107.1
(C-1′), 100.6 (OCH2O), 95.6 (C-6′), 67.6 (C-1), 66.8 (C-3), 60.9
(OCH3), 55.8 (OCH3), 36.3 (C-9a), 34.8 (C-3a), 33.0 (C-4), 31.2
(C-9), 20.9 (CH3); HRMS [M+] Calcd for C24H26O8: 442.1628;
Found: 442.1631.
F or m a tion of Oxa bicycloocta n e 36 fr om 9-Deoxyp od o-
p h yllotoxin . Ethanolysis of 9-Deoxypodophyllotoxin 29. 29
(530 mg, 1.33 mmol) in 30 mL of absolute EtOH and 10 drops
of concentrated H2SO4 was heated to reflux for 2 h after which
the reaction mixture was cooled to 25 °C and the EtOH was
evaporated. The residue was dissolved in EtOAc (40 mL) and
successively washed with 5% aq NaHCO3 (5 mL), water (10
mL), and brine (5 mL). The organic phase was dried (MgSO4).
Evaporation of the solvent and MPLC of the residue (eluting
first with CH2Cl2/EtOAc, 10/1, then EtOAc) gave 30 (140 mg,
24%) and starting 29 (389 mg, 73%). Hyd r oxyl Gr ou p
P r otection of 30. DIPEA (643 mg, 4.96 mmol) dissolved in
CH2Cl2 (1 mL) was added to an ice-cooled solution of 30 (552
mg, 1.24 mmol) in CH2Cl2 (10 mL). After stirring the resulting
solution under N2 for 10 min, a solution of BrCH2OCH3 (689
mg, 90%, 4.96 mmol) in CH2Cl2 (4 mL) was added dropwise
over 5 min, after which time stirring under N2 continued for
7 h. Solvent was removed at reduced pressure below 30 °C,
and the residue was dissolved in EtOAc (40 mL). The EtOAc
solution was washed with H2O (2 × 5 mL) and brine (5 mL),
and the organic phase was dried (anhyd MgSO4). Reduced
pressure solvent evaporation gave a residue, which was
purified by MPLC (CH2Cl2/EtOAc, 5/1) to give 31 (589 mg,
97%). Ester Red u ction in 31. Ester 31 (617 mg, 1.26 mmol)
in dry THF was stirred with LiAlH4 (252 mg, 6.31 mmol) for
4.5 h. The residue obtained after standard workup was purified
Ben zod ih yd r op yr a n 49 th r ou gh Mesyla tion of Alcoh ol
45 Givin g 48. Alcohol 45 (162 mg, 0.40 mmol) was treated
with MsCl (66 mg, 0.57 mmol) in CH2Cl2 (5 mL) under N2 and
at 0 °C for 2 h in the presence of Et3N (58 mg, 0.57 mmol).
Mesylate 48 (193 mg, 99.7%) was obtained after the standard
workup procedure. Reductive cleavage15 of mesylate 48 giving
49. A mixture of mesylate 48 (95 mg, 0.20 mmol), NaI (150
mg, 1.0 mmol), zinc powder (131 mg, 2.0 mmol), and glyme (5
mL) was heated to reflux under N2 for 3 h. Standard processing
of the reaction mixture followed by MPLC (CH2Cl2/EtOAc 20/
1) gave 49 (52 mg, 68%) and unconverted 48 (12 mg, 13%).
49: mp 152-154 °C; [R]25 -175.0° (c 0.7, CHCl3); HPLC tRI1
D
18.3, tRG1 12.9; 1H NMR (CDCl3) 6.64 (s, 1, H-8), 6.31 (s, 1,
H-5), 6.21 (s, 1, H-6′), 5.85 (s, 2, OCH2O), 3.90 (d, J ) 5.57, 1,
H-4), 3.88 (s, 3, OCH3), 3.84 (s, 3, OCH3), 3.83 (s, 3, OCH3),
3.82 (dd, J ) 1.62, 4.01, 1, H-3), 3.22 (t, J ) 11.23, 1, H-3),
2.61 (dd, J ) 4.87, 14.09, 1, H-9), 2.47 (dd, J ) 11.33, 13.93,
1, H-9), 2.02 (m, 1, H-3a), 1.23 (d, J ) 6.67, 3, H-1), 1.13 (m,
1, H-9a); 13C NMR (CDCl3) 153.28 (C-5′), 152.03 (C-3′), 150.63

Tetrahydronaphthalene Lignan Cytotoxicity
J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 7 1189
by MPLC (CH2Cl2/EtOAc, 1/1) to obtain 32 (494 mg, 88%).
Mesyla tion of Alcoh ol 32 Givin g 33. MsCl (234 mg, 2.04
mmol) dissolved in CH2Cl2 (1 mL) was added dropwise to an
ice-cooled solution of 32 (455 mg, 1.02 mmol) and triethylamine
(310 mg, 3.06 mmol) in CH2Cl2 (7 mL). The resulting mixture
was stirred under N2 at 0 °C for 1.5 h. Water was added and
the organic phase separated. The aq phase was extracted with
CHCl3 (3 × 5 mL), and the combined organic phase was dried
(anhyd MgSO4). Evaporation of the solvent at reduced pressure
and below 30 °C gave 33 (520 mg, 97%), which was used in
the next step without further purification. Red u ctive Clea v-
a ge15 of Mesyla te 33 Givin g 34. A mixture of 33 (510 mg,
0.97 mmol), NaI (730 mg, 4.86 mmol), and zinc powder (636
mg, 9.72 mmol) in glyme (10 mL) was refluxed under N2 for 3
h. The mixture was filtered to remove excess NaI and zinc.
The filtrate was added to water and then extracted with EtOAc
(3 × 30 mL). The combined EtOAc extract was washed with
brine and dried (anhyd Mg SO4), and EtOAc was evaporated
at reduced pressure. The residue was separated by MPLC
(CH2Cl2/EtOAc 10/1) to obtain 34 (238 mg, 57%) and uncon-
verted 33 (50 mg, 10%). Dep r otection in 34 Givin g 35. A
mixture of 34 (238 mg, 0.55 mmol), aq HCl (37%, 10 drops),
and MeOH (10 mL) was refluxed under N2 for 1 h and then
cooled to 25 °C. The MeOH was evaporated at reduced
pressure, and the residue was dissolved in EtOAc (25 mL).
The solution was washed with 5%, aq NaHCO3 (6 mL), H2O
(5 mL), and brine. The organic phase was dried (anhyd
MgSO4). The EtOAc was evaporated at reduced pressure, and
alcohol 35 (170 mg, 80%) was separated from the residue by
MPLC (CH2Cl2/EtOAc 5/1). Meth ylen eoxy Br id gin g in 35
Givin g 36. Employing the standard DDQ oxidation procedure
described above, alcohol 35 (156 mg, 0.40 mmol) was treated
with DDQ (110 mg, 0.48 mm) in CH2Cl2 (30 mL) for 4.5 h.
MPLC (CH2Cl2/EtOAc 15/1) gave 36 (60 mg, 39%). 36:
mmol) in CH2Cl2 in the manner indicated above gave 28 (67
mg, 56%) and an aromatized aldehyde as minor product (13
mg, 11%). 28: [R]25 +50.6° (c 1.8, CHCl3); HPLC tRI2 6.0, tRG2
D
7.2; 1H NMR (CDCl3) 7.48 (br, 1, H-5′), 6.87 (br, 1, H-6′), 6.66
(s, 1, H-8), 6.64 (br, 1, H-2′), 6.43 (s, 1, H-5), 4.22 (ddd, J )
2.28, 5.65, 8.01, 1, H-1), 3.90 (s, 3, OCH3), 3.85 (s, 3, OCH3),
3.76 (d, J ) 8.10, 1, H-1), 3.97-3.74 (br, 3, OCH3), 3.59 (s, 3,
OCH3), 3.21 (dt, J ) 2.70, 16.89, 1, H-9), 2.71 (dd, J ) 2.03,
16.91, 1, H-9), 2.57 (m, 1, H-9a), 2.12 (m, 1, H-3a), 1.08 (d,
J ) 6.88, 3, H-3); 13C NMR (CDCl3) 148.10 (C-3′), 147.57 (0),
146.70 (0), 133.90 (C-1′), 131.53 (C-4a), 127.36 (C-8a), 119.14
(C-6′), 111.71 (C-5), 110.84 (C-8), 109.90 (C-2′, 5′), 85.10 (C-
4), 72.13 (C-1), 55.76 (OCH3), 55.74 (OCH3), 55.60 (OCH3),
45.28 (C-9a), 39.72 (C-3a), 32.74 (C-9), 10.74 (C-3). HRMS [M+]
Calcd for C22H26O5: 370.1780; Found: 370.1779.
Molecu la r Mod elin g. Initial structural models of all
compounds were constructed using Alchemy (versions II and
2000)11 or Chem3D (version 3.5.1)16 molecular modeling pro-
grams. Atomic charges were computed from a Mulliken pop-
ulation analysis using the Gaussian94 program.17 Final mini-
mum-energy structures were computed with the Discover
(version 2.9.5) program18 using its built-in Amber force field.
Molecular dynamics (MD) simulations were also carried out
with the Discover-Amber combination. All MD simulations
were run at 300 K (NVT conditions) for a duration of 400 ps
(after an initial equilibration period of 5 ps) using a time-step
of 1 fs. Preliminary MD runs comparing fully substituted
molecules with their skeletal equivalents, which were devoid
of substituents on the two aromatic rings but contained all
other substituents, showed no measurable differences. Con-
sequently, all MD results are reported for the unsubstituted
(skeletal) molecules.
Ack n ow led gm en t. This work was supported by the
McIntire-Stennis USDA-CREES program and SUNY-
ESF. Determination of high-resolution mass spectra by
the Nebraska Center for Mass Spectrometry, University
of Nebraska, Lincoln, NE is also acknowledged. Com-
pounds were evaluated for cytotoxic activities through
the Developmental Therapeutics Program, Division of
Cancer Treatment and Diagnosis, National Cancer
Institute, National Institute of Health, Bethesda, MD.
[R]25 + 10.5° (c 1.0, CHCl3); HPLC tRI2 7.6, tRG2 8.4; 1H NMR
D
(CDCl3) 7.15 (br, 1, H-2′ or 6′), 6.64 (s, 1, H-8), 6.38 (s, 1, H-5),
6.27 (br, 1, H-6′ or 2′), 5.88 (d, J ) 1.42, 1, OCH2O), 5.83 (d,
J ) 1.42, 1, OCH2O), 4.21 (ddd, J ) 2.42, 5.54, 8.07, 1, H-1),
3.89 (s, 3, OCH3), 3.95-3.74 (m, 7, 2 × OCH3, H-1), 3.20 (dt,
J ) 2.79, 17.03, 1, H-9), 2.73 (dd, J ) 2.21, 17.04, 1, H-9), 2.55
(m, 1, H-9a), 2.17 (m, 1, H-3a), 1.16 (d, J ) 6.93, 3, H-3); 13C
NMR (CDCl3) 153.18 (C-3′ or 5′), 152.51 (C-5′ or 3′), 146.84
(C-7), 145.67 (C-6), 137.12 (C-1′), 136.89 (C-4′), 132.93 (C-8a),
128.80 (C-4a), 108.96 (C-5), 108.12 (C-8), 104.53 (C-2′ or 6′),
103.75 (C-6′ or 2′), 100.78 (OCH2O), 85.43 (C-4), 72.25 (C-1),
60.88 (OCH3), 56.18 (OCH3), 44.81 (C-9a), 39.97 (C-3a), 32.21
(C-9), 11.21 (C-3); HRMS [M+] Calcd for C22H24O6: 384.1573;
Found: 384.1577.
Su p p or tin g In for m a tion Ava ila ble: HPLC data for
compounds 5, 6, 11, 12, 14, 16, 17, 19-23, 28, 36, 45, 46, 47,
and 49. This material is available free of charge via the
Internet at http://pubs.acs.org.
F or m a tion of Oxa bicycloocta n e 28 fr om Dim eth yl-R-
con id en d r in , 2. Meth a n olysis of 2. 2 (204 mg, 0.53 mmol)
in 24 mL of absolute MeOH and 5 drops of concentrated H2-
SO4 was heated to reflux for 24 h and then processed in the
manner used for the conversion of 9-deoxypodophyllotoxin to
30. Obtained 24 (98 mg, 44%) and recovered 2 (104 mg, 51%).
Mesyla tion of Alcoh ol 24 Givin g 25. MsCl (129 mg, 1.12
mmol) dissolved in CH2Cl2 (2 mL) was added dropwise to an
ice-cooled solution of alcohol 24 (234 mg, 0.56 mmol) and
triethylamine (114 mg, 1.12 mmol) in CH2Cl2 (8 mL). The
mixture was stirred under N2 at 0 °C for 2.5 h. Subsequent
processing of the mixture in the manner used in the conversion
of 32 to 33 gave 25 (268 mg, 96%), which was used without
further purification. Red u ctive Clea va ge of Mesyla te 25
Givin g 26. A mixture of 25 (258 mg, 0.52 mmol), NaI (391
mg, 2.61 mmol), Zn powder (171 mg, 2.61 mmol), and glyme
(15 mL) was heated to reflux under N2 for 4 h. Subsequent
processing of the mixture in the manner used in the conversion
of 33 to 34 gave 26 (155 mg, 74%) and recovered 25 (24 mg,
9%) from MPLC (CH2Cl2/EtOAc, 10/1). Red u ction of Ester
26 to 27. Ester 26 (153 mg, 0.382 mmol) in 7 mL of dry THF
and LiAlH4 (73 mg, 1.910 mmol) was stirred for 14 h followed
by processing in the normal manner as noted above gave 27
(130 mg, 91%). Meth ylen eoxy Br id gin g in 27 Givin g 28.
27 (120 mg, 0.322 mmol) treated with DDQ (88 mg, 0.387
Refer en ces
(1) Dantzig, A.; LaLonde, R. T.; Ramdayal, F. D.; Shepard, R. L.;
Yanai, K.; Zhang, M. Cytotoxic responses to aromatic ring and
configurational variations in R-conidendrin, podophyllotoxin, and
sikkimotoxin derivatives. J . Med. Chem. 2001, 44, 180-185.
(2) Ramdayal, F. D.; Kiemle, D. J .; LaLonde, R. T. Directed DDQ-
promoted benzylic oxygenations of tetrahydronaphthalenes. J .
Org. Chem. 1999, 64, 4607-4609.
(3) One indicator of the spacial relation between fused and pendant
aromatic ring is the torsion angle C-5, C-4a, C-4, all within the
tetrahydronaphthalene scaffold, and C-1′ of the pendant ring.
The measure of this torsion angle indicates the position of C-1′
relative to the plane of the fused aromatic ring. The energy-
minimized, Alchemy models of oxolane 10 and oxabicyclooctane
5 (Scheme 1) and dioxatricyclodecane 14 and oxabicyclooctane
16 (Scheme 2) have the torsion angles as follows: 10, 73.6°; 5,
24.6°; 14, 26.1°; and 16, 48.9°. The torsion angle C-4a, C-4, C-1′,
and C-2′(or 6′) defines the twist orientation of the pendant ring
plane relative to the plane of the fused aromatic ring.
(4) Hearon, W. M.; MacGregor, W. S. The naturally occurring
lignans. Chem. Rev. 1955, 55, 957-1068.
(5) Castro, M. A.; Gordaliza, M.; del Corral, J . M. M.; San Feliciano,
A. Preparation of triols and ethers related to podophyllotoxin.
Org. Prep. Proced. Int. 1994, 26, 539-547.
(6) A shorter route to 28, involving initial mesylation of oxabicy-
clooctane 6 followed by reduction, was infeasible since the
mesylate of 6 was found unstable at temperatures considerably
lower than that required for a NaI/Zn-reduction.

1190 J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 7
LaLonde et al.
(7) Nemec, J . U.S. Patent 4609644, Sep. 2, 1986.
(13) Terada, T.; Fujimoto, K.; Nomura, M.; Yamashita, J .; Kobunai,
T.; Takeda, S.; Wierzba, K.; Yamada, Y.; Yamaguchi, H. Anti-
tumor agents. I. DNA topoisomerase II inhibitory activity and
the structural relationship of podophyllotoxin derivatives as
antitumor agents. Chem. Pharm. Bull. 1992, 40, 2720-2727.
(14) Cooksey, C. J .; Land, E. J .; Riley, P. A. A simple one-pot
preparation of 4-alkoxy- and 4-alkylthio-catechols and benzo-
quinones. Org. Prep. Proced. Int. 1996, 28, 463-498.
(15) Fujimoto, Y.; Tatsuno, T. A novel method for reductive removal
of tosyloxy and mesyloxy groups. Tetrahedron Lett. 1976, 37,
3325-3326.
(8) Dioxatricyclodecane 14 gives a log GI50 value of <-8 in CAKI-
1, whereas the antineoplastic agent Etoposide gives a value of
-5.2 in the same cell line, as reported by Van Vliet, D. S.;
Tachichibana, Y.; Bastow, K. F.; Huang, E. S.; Lee, K. H. J . Med.
Chem. 2001, 44, 1422-1428.
(9) Preliminary investigations by Dr. Susan Bane (Chemistry
Department, SUNY Binghamton, Binghamton, NY) indicate that
unlike PT its cytotoxic dioxatricyclodecane derivative 14 has no
effect on microtubule polymerization at concentrations up to 55
µM.
(10) Hollingshead, M. G.; Alley, M. C.; Camalier, R. F.; Abbott, B.
J .; Mayo, J . G.; Malspeis, L.; Grever, M. R. In vivo cultivation
of tumor cells in hollow fibers. Life Sci. 1995, 57, 131-141.
(11) Tripos Inc, 1699 South Hanley Road, St. Louis, MO 63144.
(12) Murphy, S. T.; Ritchie, E.; Taylor, W. C. Some constituents of
Australobaileya scandens (Austrobailyaceae) structures of seven
new lignans. Aust. J . Chem. 1975, 28, 81-90.
(16) CambridgeSoft. Com, 100 Cambridge Park Drive, Cambridge,
MA 02140.
(17) Gaussian, Inc., Carnegie Office Park, Building 6, Suite 230,
Carnegie, PA 15106.
(18) Accelrys Inc., 9685 Scranton Road, San Diego, CA 92121.
J M020158P