New homocamptothecins: Synthesis, antitumor activity, and molecular modeling

Article abstract of DOI:10.1016/j.bmc.2007.10.046

Homocamptothecins (hCPTs) represent a class of new emerging antitumor agents, which contains a seven-membered β-hydroxylactone in place of the conventional six-membered α-hydroxylactone ring (E ring) of camptothecins. Some novel 7-substituted hCPTs were designed and synthesized based on a newly developed synthetic route which couples ring A with ring C, E and D. Most of the synthesized compounds exhibit very high cytotoxic activity on tumor cell line A549. Some compounds, such as 9b, 9l, and 9y, show broad in vitro antitumor spectrum and are more potent than topotecan. Three-dimensional quantitative structure-activity relationship (3D-QSAR) methods, CoMFA and CoMSIA, were applied to explain the structure-activity relationship (SAR) of the synthesized compounds. Furthermore, molecular docking was used to clarify the binding mode of the synthesized compounds to human DNA topoisomerase I. The important hydrophobic, base-pair stacking, and hydrogen-bonding interactions were observed between the hCPT derivatives and their receptor. The results from molecular modeling will guide the design of novel hCPTs with higher antitumor activity.

Full text of DOI:10.1016/j.bmc.2007.10.046

Available online at www.sciencedirect.com
Bioorganic & Medicinal Chemistry 16 (2008) 1493–1510
New homocamptothecins: Synthesis, antitumor activity,
and molecular modeling
Zhenyuan Miao,a,ꢀ Chunquan Sheng,a,ꢀ Wannian Zhang,a,* Haitao Ji,b Jing Zhang,a
Lucheng Shao,a Liang You,a Min Zhang,a Jianzhong Yaoa and Xiaoyin Chea
¨
aSchool of Pharmacy, Military Key Laboratory of Medicinal Chemistry, Second Military Medical University,
325 Guohe Road, Shanghai 200433, People’s Republic of China
bDepartment of Chemistry, Department of Biochemistry, Molecular Biology, and Cell Biology, and Center for Drug Discovery
and Chemical Biology, Northwestern University, Evanston, IL 60208-3113, USA
Received 24 July 2007; revised 15 October 2007; accepted 16 October 2007
Available online 22 October 2007
Abstract—Homocamptothecins (hCPTs) represent a class of new emerging antitumor agents, which contains a seven-membered
b-hydroxylactone in place of the conventional six-membered a-hydroxylactone ring (E ring) of camptothecins. Some novel 7-substi-
tuted hCPTs were designed and synthesized based on a newly developed synthetic route which couples ring A with ring C, E and D.
Most of the synthesized compounds exhibit very high cytotoxic activity on tumor cell line A549. Some compounds, such as 9b, 9l,
and 9y, show broad in vitro antitumor spectrum and are more potent than topotecan. Three-dimensional quantitative structure–
activity relationship (3D-QSAR) methods, CoMFA and CoMSIA, were applied to explain the structure–activity relationship
(SAR) of the synthesized compounds. Furthermore, molecular docking was used to clarify the binding mode of the synthesized com-
pounds to human DNA topoisomerase I. The important hydrophobic, base-pair stacking, and hydrogen-bonding interactions were
observed between the hCPT derivatives and their receptor. The results from molecular modeling will guide the design of novel
hCPTs with higher antitumor activity.
ꢀ 2007 Elsevier Ltd. All rights reserved.
1. Introduction
Human DNA topoisomerase I (Topo I) is the molecular
target of CPT derivatives.11 These compounds act by
binding to a transient Topo I–DNA covalent complex,
leading to an accumulation of DNA strand breaks upon
replication, ultimately causing cell death during the S
phase of the cell cycle.12–14 Although CPT analogues re-
main a promising class of antitumor agents, the intrinsic
instability of the highly electrophilic a-hydroxylactone
of the E ring undergoes rapid hydrolysis to the biologi-
cally inactive carboxylate form under physiological con-
ditions.15 This chemical feature diminishes efficacy of
various CPT derivatives in vivo compared to the specta-
cular results often obtained from in vitro studies and
xenograft models.16 Thus, extensive efforts have been
put to synthesize new CPT analogues with a prolonged
biological life maintaining its active lactonic form. In
1997, Lavergne et al. reported the insertion of a methy-
lene spacer between alcohol and carboxyl functions of
CPT to obtain homocamptothecin (hCPT).17 The
replacement of the six-membered a-hydroxylactone with
a seven-membered b-hydroxylactone reduces consider-
ably the electrophilicity of the lactone and hence de-
creases the rate of lactone hydrolysis to the
Camptothecin (CPT) was first isolated from the Chi-
nese tree Camptotheca acuminata by Wall et al. in
1966.1 The compound exhibits potent antitumor activ-
ity and has been evaluated in clinic.2 However, the
therapeutic application of unmodified CPT is hindered
due to its poor aqueous solubility and high in vivo
hepatotoxicity. Subsequently, the structural modifica-
tion of natural CPT has generated many new CPT
derivatives with improved pharmacological or phar-
macokinetic profile.3–5 Among those, topotecan6 and
irinotecan7 (Chart 1) have entered clinical use and sev-
eral drug candidates, such as rubitecan,8 lurtotecan,9
and exatecan,10 are currently in different stages of
the clinical trials.
Keywords: Homocamptothecins; Antitumor activity; 3D-QSAR;
Molecular docking.
*
Corresponding author. Tel./fax: +86 21 25074460; e-mail:
zhangwnk@hotmail.com
These two authors contributed equally to this work.
ꢀ
0968-0896/$ - see front matter ꢀ 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2007.10.046

1494
Z. Miao et al. / Bioorg. Med. Chem. 16 (2008) 1493–1510
N
N
7
9
10
HO
N
O
O
O
N
N
O
11
N
N
O
O
O
O
OH
OH
irinotecan
topotecan
CH
3
N
F
F
H C
3
O
O
N
N
N
Cl
N
O
O
O
O
OH
diflomotecan
OH
BN80927
Chart 1.
corresponding inactive carboxylate form. Moreover,
hCPT is found to be much more stable than CPT in
plasma because the affinity of hCPTs to human serum
albumin is lower than that of the corresponding CPTs.
As compared to CPT, hCPT exhibits better biological
and pharmacological properties.18 For example, hCPT
conserves Topo I-mediated activity and was found to
promote additional DNA breaks which are not seen
with CPT.19 It proves to be much more cytotoxic (up
to 10 times) than CPT against a variety of tumor cells.
Furthermore, the in vivo evaluation showed that hCPT
exhibited superior antitumor activities compared to
CPT in the HT-29 xenograft model.20 On the basis of
these encouraging results, several hCPT analogues
with the modified A and/or B rings have been synthe-
sized and evaluated in vitro and in vivo as the antitumor
agents.21–25 Two of them, BN80915 (Diflomotecan)26
and BN8092727 (Chart 1) are in clinical trial and encour-
aging results have been obtained. The exploration of
new hCPT derivatives has recently emerged as a promis-
ing field to find better antitumor agents.
ing reduction, oxidative cleavage, reformatsky reaction,
and ring closure, was used to synthesize the racemic
hCPT. However, the semi-synthetic approach was not
appropriate for the preparation of diversely substituted
hCPT analogues. Therefore, the total synthesis of hCPT
analogues was more practical for the structural optimi-
zation. The first total synthesis of hCPT was reported
by Lavergne et al. in 1998.21 This route was based on
the coupling of a DE heterobicycle with an AB quino-
line. Another synthetic route of hCPT was developed
by Curran and co-workers in 2002,28 which was based
on the coupling of an A ring with a DE heterobicycle
by using the cascade radical annulation approach. How-
ever, the overall yield of these two synthetic routes is rel-
atively low and some expensive or dangerous reagents
have to be used. Therefore, the development of a new to-
tal synthesis route with mild reaction condition and high
yield is particularly urgent.
Scheme 1 shows a new synthetic strategy for hCPTs
based on the coupling of ring A with ring CDE. The
key intermediate CDE ring 1 was synthesized according
to the reported route29 and the synthetic process was
optimized in our previous studies.30,31 Compound 1
was reduced by KBH4 in anhydrous methanol to afford
1,2-diol 2, which was oxidatively cleaved by NaIO4 to
give formyloxy ketone 3. Reformatsky reaction of 3 with
tert-butyl bromoacetate and Zn gave b-hydroxy ester 4
in a 97% yield and then the treatment with trifluoroace-
tic acid provided a key tricyclic intermediate 5.
Friedlander condensation of the tricyclic 5 with N-(o-
aminobenzylidene)-p-toluidine in the presence of an acid
catalyst gave the known homocamptothecin 6. The
overall yield of the present route is 12.8%, which is high-
er than those of the other two methods mentioned
above.21,28 Moreover, this route is more appropriate
for the synthesis of diversely A, B ring-modified new
hCPT derivatives.
Some 7-substituted hCPT derivatives were designed in
the present study. A new synthetic route was developed
to synthesize the designed hCPT derivatives. Three-
dimensional quantitative structure–activity relationship
(3D-QSAR) methods were applied to explain the struc-
ture–activity relationship (SAR) of the synthesized com-
pounds. Furthermore, molecular docking was used to
clarify the binding mode of the synthesized compounds
to human DNA topoisomerase I. The obtained informa-
tion is useful in the future structural optimization.
2. Chemistry
hCPT was first synthesized by a semi-synthetic ap-
proach.17 Starting from CPT, a four-step route, includ-

Z. Miao et al. / Bioorg. Med. Chem. 16 (2008) 1493–1510
1495
O
O
O
OH
OH
OH
O
O
O
O
O
a
b
N
OCHO
N
O
N
O
O
O
O
1
2
3
O
N
O
OH
COOtBu
O
O
c
d
e
O
N
OH
NH2
O
OH
O
O
O
4
5
O
N
N
O
O
OH
6
Scheme 1. Reagents and conditions: (a) KBH4, CH3OH, rt, 20 min, 80%; (b) CH3COOH, NaIO4, rt, 30 min, 93%; (c) Zn, BrCH2COOtBu, THF,
75 ꢁC, 3 h, 97%; (d) CF3COOH, rt, 10 h, 94%; (e) toluene, p-TSA, 130 ꢁC, 1.5 h, 80%.
7-Hydroxymethylhomocamptothecin 7 could be ob-
tained by the reaction of 6 with hydrogen peroxide
and methanol. The hydroxymethyl group was oxidized
to form 7-formylhomocamptothecin 8, which was al-
lowed to react with different substituted anilines to gen-
erate the corresponding derivatives 9a–y (Scheme 2).
lid tumor cell lines, A-549 (for non-small cell lung can-
cer), LOVO (for colon cancer), and MCF-7 (for breast
cancer), were chosen for testing the in vitro antitumor
activities of the 7-aryliminomethyl hCPT derivatives.
Topotecan and irinotecan were used as the reference
compounds. The inhibitory activities suggested that
the arylimino-methyl group in the 7-position can pro-
mote antitumor activity. These results revealed that
the compounds were more sensitive against A-549 and
LOVO cell lines than against MCF-7 cell line (Table
1), which is consistent with the clinical behavior of
CPT derivatives.35 For A-549 cell line, most compounds
showed high inhibitory activities with the IC50 values
lower than 0.2 nM, which were more potent than topo-
tecan and irinotecan. Many compounds also showed
better inhibitory activity against LOVO cell line than
topotecan. For example, the IC50 values of 9l and 9y
were lower than 0.2 nM. In particular, the activity of
the most promising compounds, 9b, 9l, and 9y, showed
broad in vitro antitumor spectrum and are more potent
than topotecan and irinotecan. Further pharmacological
and toxicological evaluation of these promising com-
pounds is in progress. The detailed explanation of
SAR would be discussed in the following 3D-QSAR
studies.
Various 1-(2-amino substituted phenyl)-2-chloro-etha-
nones 11a–n were obtained by reacting substituted ani-
lines 10a–n with chloroacetonitrile under the catalysis
of BCl3, which were treated with the intermediate 5 to
give the 7-chloromethylhomocamptothecin
A ring
substituted derivatives 13a–n (Scheme 3). Compounds
13a–n reacted with the substituted pyridines to afford
quaternary-salt derivatives 13a–n and 14a–d.
3. Results and discussion
3.1. 7-Aryliminomethyl hCPT derivatives
Numerous reports indicated that the CPT derivatives
with substitutions at quinoline (A/B) ring,3 especially
at C-7, C-9, and C-10, are of great interest and several
good drug candidates (such as exatecan, lurtotecan,
and silatecan) are under clinical trials. The importance
of the lipophilic groups linked to position 7 of campto-
thecin has been demonstrated to increase inhibitor intra-
cellular uptake and cytotoxic activities.32 The crystal
structures of Topo I–DNA in complex with topotecan
and CPT33,34 also revealed that the substituents on C-
7 were located into the major groove of DNA and they
could reinforce the stability of the inhibitor-Topo I–
DNA covalent complex. Moreover, some 7-aryliminom-
ethyl CPT analogues have shown good in vitro antitu-
mor activity.35 Therefore, we designed and synthesized
a series of 7-aryliminomethyl hCPT derivatives. The so-
3.2. Water soluble hCPT derivatives
The clinical success of the water soluble CPT deriva-
tives topotecan and irinotecan highlights the impor-
tance of new water soluble CPT or hCPT analogues,
which has become a major research focus at pres-
ent.36,37 We designed and synthesized a series of ring
A and/or ring B-modified 7-substituted pyridine qua-
ternary-salt derivatives in this study. The in vitro cyto-
toxicity evaluations indicated that the quaternary salt
hCPTs were less active than the above synthesized

1496
Z. Miao et al. / Bioorg. Med. Chem. 16 (2008) 1493–1510
HO
CHO
O
O
N
N
a
b
N
N
6
O
O
O
O
OH
OH
7
8
R
N
9q R = 2-OH
9r R = 2,4-2CH3
9s R = 2-Br
9t R = 3-Cl-4-F
9u R = 2-OCH3
9v R = 2,4-2Cl
9w R = 2-Cl
9h R = 3,4-2F
9i R = 3-Cl
9j R = 3-Cl-4-CH3
9k R = 2,5-2Cl
9l R = 3,5-2CH3
9m R = 3,4-2CH3
9n R = 4-Cl
9a R = H
9b R = 2-CH3
9c R = 4-CH3
9d R = 3-CH3
9e R = 4-Br
9f R = 3,5-2Cl
9g R = 3-F
N
c
O
N
9x R = 4-C(CH3)3
9y R = 4-COCH3
O
9o R = 4-CN
9p R = 3,4-2Cl
O
OH
9a-y
Scheme 2. Reagents and conditions: (a) H2O2, CH3OH, FeSO4Æ7H2O, rt, 14 h, 64.7%; (b) CH3COOH, 130 ꢁC, 5 h, 52.7%4; (c) substituted aniline, rt,
10–24 h, 14.5–30.7%.
Cl
R1
Cl
R1
R1
R2
R3
O
R2
R3
R2
R3
N
O
a
b
N
O
NH2
NH2
O
10a-n
11a-n
OH
12a-n
13a R1 = H, R2,R3 = -OCH2CH2O-
13b R1,R2 = -OCH2CH2O-, R3 = H
13c R1=R2 = H, R3 = CH3
13d R1 = H, R2=R3 = CH3
13e R1=R3 = CH3, R2 = H
13f R1=R3 = H, R2 = F
N
R1
R2
O
c
N
13g R1 = H, R2 = CH3, R3 = Cl
13h R1=R2=H, R3=Cl
13i R1=R2=H, R3=F
13j R1=R3 = H, R2 = Cl
13k R1 = H, R2=R3 = Cl
R3
N
O
O
OH
13l
R1 = H, R2 = F, R3 = Cl
13a-n
13m R1= R3 = H, R2 = Br
13n R1= R3 = H, R2 = CH3
R
N
Cl
H3C
H3C
O
O
N
14a R = 2-CH3
14b R = 3-CH3
14c R = 4-CH3
14d R = 4-COOH
N
d
N
N
O
O
O
O
OH
OH
12n
14a-d
Scheme 3. Reagents: (a) BCl3; (b) p-TSA; (c) pyridine, DMSO; (d) substituted pyridine, DMSO.
7-iminomethyl hCPTs (Table 2). Most compounds are
more sensitive against A549 cell line than LOVO and
MCF-7 with their IC50 values in the range of 19.6 lM
to 7 nM. Only 13n is more potent than topotecan with
the IC50 value against A549 lower than 0.2 nM. For
A549 cell line, the antitumor activity was increased
when A-ring was substituted with a methyl group.
However, the pyridyl group of C-7 is suitable to be
unsubstituted because the incorporation of the
methyl or carboxyl group could lead to a decrease

Z. Miao et al. / Bioorg. Med. Chem. 16 (2008) 1493–1510
1497
Table 1. In vitro antitumor activity of 7-aryliminomethyl hCPT
derivatives
Table 2. In vitro antitumor activity of 7-substituted pyridine quater-
nary-salt hCPT derivatives
Compound
IC50 (lM)
Compound
IC50 (lM)
A-549
LOVO
MCF-7
A-549
1.122
LOVO
MCF-7
7
0.017
0.026
0.002
0.050
0.024
0.014
0.0008
0.035
0.125
0.011
0.533
0.014
0.023
5.192
0.020
0.381
<0.0002
0.007
0.0004
0.368
0.032
0.002
10.729
0.018
0.040
0.015
0.060
0.003
0.021
<0.0002
0.036
9.015
4.33
3.55
13a
13b
>19.51
>19.51
>21.34
>20.72
>20.72
>21.16
7.08
>19.51
>19.51
>21.34
>20.72
>20.72
>21.16
>19.88
>20.45
>21.16
>20.45
>19.11
>19.73
>18.75
6.937
8
9a
2.017
2.391
0.313
0.290
0.381
0.007
18.282
2.850
3.49
0.488
0.517
5.882
20.855
2.324
9.291
9.245
0.441
>20
13c
13d
9b
9c
<0.0002
0.038
0.093
13e
13f
9d
9e
0.002
0.828
13g
13h
9f
9g
>20.45
>21.16
9.039
0.0008
0.003
5.822
13i
13j
9h
9i
13k
13l
19.608
>19.73
0.729
<0.0002
5.015
2.072
2.114
7.316
0.005
6.528
>19.11
>19.73
7.912
9j
9k
0.013
0.226
1.344
6.621
0.003
0.559
2.000
25.994
6.847
20.768
20.259
0.035
0.492
2.763
3.943
1.154
0.853
0.852
0.488
17.403
13m
13n
9l
9m
<0.0002
0.0006
<0.0002
0.538
3.543
9.318
14a
14b
>20.72
>20.72
>20.72
>19.51
0.488
9n
9o
>20.72
>20.72
>19.51
0.036
14c
14d
9p
9q
0.0009
<0.0002
13.926
<0.0002
<0.0002
<0.0002
0.0009
<0.0002
0.002
Topotecan
Irinotecan
9r
9.015
17.403
9s
9t
9u
3.4. 3D-QSAR models
9v
9w
A data set of the above synthesized hCPTs was used to
perform the comparative molecular field analysis (CoM-
FA)38 and the comparative molecular similarity indices
analysis (CoMSIA)39 studies. In an attempt to build ro-
bust 3D-QSAR models, variation of the parameters,
such as grid spacing and attenuation factor, was consid-
ered during the 3D-QSAR studies. However, the best re-
sults were obtained from the parameters with the default
values. It has been discussed that five different descriptor
fields are not totally independent of each other and such
dependencies of individual fields usually decrease the sta-
tistical significance of the models.40,41 A fast algorithm
SAMPLS was used to evaluate which of these five CoM-
SIA fields was actually important for a predictive model
by the combination of all possible descriptor fields42
(Fig. 2). The first five models, using a single CoMSIA
field, indicated that the steric field, the electrostatic field,
and the hydrophobic field were more important than the
other two fields, the hydrogen-bond donor and the
hydrogen acceptor. However, the addition of hydro-
gen-bond acceptor field generated the best CoMSIA
model ðq2SAMPLS ¼ 0:717Þ. The correlation coefficients of
r2 of 0.905 and 0.938, and the cross-validated coefficients
of q2 of 0.703 and 0.717 were obtained for the best CoM-
FA and CoMSIA models, respectively. The final CoM-
FA and CoMSIA models were satisfactory from the
viewpoint of the statistical significance (Table 4). The
predictive ability of the training set and test set wad eval-
uated which indicated that the predictive IC50 values of
the compounds correlated well with the experimental
values (Table 5 and Fig. 3). A real external compound
set of six previously synthesized 7-ester hCPT derivatives
(compounds 15a–f, Fig. 1) was tested to evaluate the reli-
ability of their 3D-QSAR models. The results revealed
that the biological activities of these compounds were
also predicted with good accuracy.
9x
9y
<0.0002
0.005
6.528
Topotecan
Irinotecan
of antitumor activity. Although the in vitro cytotoxic-
ities of these water soluble hCPTs are not very satis-
fying, we expect that the in vivo studies of these
compounds can provide some hints for the future
inhibitor design.
3.3. In vivo antitumor activity
For most antitumor agents, their in vivo activity usually
does not correlate well with their in vitro data. Thus, a
preliminary in vivo screening of the synthesized com-
pounds against C26 colon cancer model was performed
(data not shown). Compounds 9j and 13g were found to
be good candidates, which were selected for further de-
tail in vivo studies. Dose–effect relationship studies re-
vealed that the tumor growth inhibition rate of
compounds 9j and 13g was increased with the addition
of their dose. At the dose of 4 mg/kg, compounds 9j
and 13g achieved tumor growth inhibition rate of
30.99% and 36.21%, respectively, which was almost
equivalent to that of TPT at a dose of 1 mg/kg
(Table 3). It is worth noting that when the dose of
TPT was added to 1.5 mg/kg, all the mice in the group
died after the treatment, while all the mice in the group
of compounds 9j and 13g at the dose of 4 mg/kg re-
mained alive. According to this, compounds 9j and
13g may be safer than TPT, although they were less
effective than TPT at the same dosage. Further pharma-
cological and toxicological evaluation of these two com-
pounds is in progress.

1498
Z. Miao et al. / Bioorg. Med. Chem. 16 (2008) 1493–1510
To visualize the information content of the derived 3D-
QSAR models, the CoMFA and CoMSIA contour
maps were generated. The field energies at each lattice
point were calculated as the scalar results of the coeffi-
cient and the standard deviation associated with a par-
vored. The hydrophobic favored regions are around
the phenyl ring of the C7-side chain. Except the meta po-
sition, the ortho and para positions of the phenyl ring are
suitable to be substituted with hydrophobic groups. For
example, compounds 9n (IC50 < 0.0002 lM) and 9w
(IC50 < 0.0002 lM) are more potent than 9a
(IC50 = 0.002 lM). On the contrary, the addition of a
hydrophobic chlorine atom to the meta position of the
phenyl ring of 9a led to a decrease of antitumor activity
(compounds 9f and 9i). As shown in Figure 4d, the pur-
ple regions of hydrogen-bond acceptor field indicate
where the hydrogen-bond acceptors are favored and it
is at the ortho position of the phenyl ring, while hydro-
gen-bond acceptor disfavored area in the red color is at
the para position. It was supported by the fact that the
addition of a hydroxyl group to the ortho position of
the phenyl group of 9a led to a increase of antitumor
activity (compound 9q, IC50 < 0.0002 lM) and when
the hydroxyl group was methylated (compound 9u,
IC50 < 0.0002 lM) the activity was unchanged. On the
contrary, if a hydrogen-bond acceptor was added to
the para position of the phenyl group, the antitumor
activity was decreased (e.g., compound 9o,
IC50 = 0.538 lM).
ticular column of the data table (‘stdev coeff’), which
*
was always plotted as the percentage of the contribution
to the CoMFA or CoMSIA equation. Because the
CoMFA and CoMSIA contour maps of steric field rep-
resented similar results, only the CoMFA contour map
is displayed in Figure 4a. The green contours represent
the regions of high steric tolerance, while the yellow con-
tours represent regions of unfavorable steric effects. The
sterically favored regions are mainly around the aro-
matic ring attached to the iminomethyl group of the
C7-side chain, which indicates that larger substituents
on the iminomethyl group are favorable for high cyto-
toxicity, for example, most of the compounds with
substituted aromatic ring on the iminomethyl group
are more potent than compounds with small substitu-
ents on C-7 (compounds 7 and 8). Moreover, the addi-
tion of two methyl groups on the meta and para
position of the phenyl group of 9a (IC50 = 0.002 lM)
led to a increase of the antitumor activity (compound
9m, IC50 = 0.0006 lM). However, the length of C7-side
chain cannot be too long because there is a yellow region
above the para position of the pyridyl group of 13n
(IC50 < 0.0002 lM). When a methyl group was added
to the para position of the pyridyl group of 13n, the anti-
tumor activity was decreased to a large extent (com-
pound 14c, IC50 = 2.114 lM). There are a green region
and a yellow region around the C-10 and C-11 of the
A ring, respectively, which is consistent with the SAR
analysis of 13a–n. For example, the addition of a methyl
group to the C-10 of 13c (IC50 = 2.391 lM) generates
13d (IC50 = 0.313 lM) and its antitumor activity is im-
proved accordingly. When the 9,10-ethylenedioxy group
of 13a (IC50 = 1.122 lM) was replaced by the 10,11-eth-
ylenedioxy group (compound 13b, IC50 = 2.107 lM),
the antitumor activity was decreased much because a
larger substituent on C-11 was prohibited.
3.5. The binding mode of the hCPTs
Compound 9l, one of the most active compounds, was
used to investigate the binding mode of the synthesized
hCPT derivatives. InsightII/Affinity was used in the
present docking study because its effectiveness in clarify-
ing the interactions of hCPTs with Topo I has been
proved by previous docking studies.43,44 Moreover, the
crystal structures of apo- and ternary complex of Topo
I revealed the flexibility of the active site. Because the
flexibility of both the active site of the target enzyme
and the inhibitor is taken into account in the Affinity
method, it should be a powerful tool to study the mode
of action of the newly synthesized hCPT derivatives. A
comparison between the docking model of 9l and the
ternary complex structures of CPT reveals only minor
differences in the overall orientation of the heterocyclic
ring. Compound 9l intercalates at the site of DNA cleav-
age, forming base-stacking interactions with both
the ꢀ1 (upstream) and +1 (downstream) base pairs
(Fig. 5). The orientation of 9l in the active site cavity
is perpendicular to the main axis of the DNA and paral-
lel to the bases. Previous studies indicated that hydro-
gen-bonding interaction played an important role
during the formation of drug–DNA–enzyme ternary
complex.33,34 In the docking model, compound 9l bound
into Topo I–DNA covalent complex via three direct
hydrogen bonds. The N1 of compound 9l formed a
hydrogen bond with Arg 364 and the b-hydroxyl group
of E-ring formed another hydrogen bond with Asp 533,
which is consistent with crystal structure of CPT–DNA–
enzyme ternary complex.34 The third hydrogen bond
was observed between the lactone carbonyl group of
E-ring and Thr 718. The C-7 side chain was faced into
the major groove and formed hydrophobic interaction
with the surrounding hydrophobic residues lined with
Ala 351, Met 428, and Pro 431. The addition of the
hydrophobic groups (such as a methyl group) on the
From the CoMFA model, the contribution of electro-
static field (0.272) was lower than that of steric field
(0.728). Negative charge favored regions (the red con-
tour in Fig. 4b) were found around the C-2 and C-5 of
the phenyl ring of the 7-iminomethyl hCPTs and the po-
sitive charge favored regions (the blue contour in
Fig. 4b) were around the N atom of the imino group
and the C-4 of the phenyl ring. These regions support
the observation that the imino group is important for
the antitumor activity. Moreover, when a hydroxy
group was added to C-2 of the phenyl ring of 9a
(IC50 = 0.002 lM), the cytotoxicity was improved to a
large extent (compound 9q, IC50 < 0.0002 lM). The sim-
ilar results were obtained when a chlorine atom was
introduced to the C-5 of the phenyl ring of 9i.
Furthermore, CoMSIA reveals three other contour
maps that are not shown in CoMFA. The yellow regions
in Figure 4c indicate the areas where the hydrophobic
substitutions are preferred and the purple regions repre-
sent areas where hydrophobic substitutions are disfa-

Z. Miao et al. / Bioorg. Med. Chem. 16 (2008) 1493–1510
1499
Table 3. The in vivo antitumor activity of compounds 9j and 13g against C26 colon cancer model
Sample
Dose (mg/kg)
Mice number
Body weight (g)
End
Tumor weight (g)
Tumor growth inhibition %
Begin
End
Begin
ꢀx ꢁ SD
Control
TPT
9j
25
1
18
10
10
10
10
6
18
10
10
10
10
6
18.85 1.84
16.88 2.09
17.85 1.22
18.36 1.31
17.98 1.28
21.26 1.73
22.10 1.75
21.66 2.48
17.31 2.32
16.06 1.97
16.64 1.11
17.27 1.73
16.92 1.30
19.22 2.15
20.09 1.82
20.14 2.58
1.54 0.19
0.82 0.19
1.21 0.34
1.09 0.59
1.06 0.43
2.04 0.32
2.01 0.35
1.52 0.25
37.61
21.29
28.98
30.99
14.60
15.79
36.21
1
9j
2
9j
4
13g
13g
13g
1
2
6
6
4
6
6
would clash with Ala 351. The ortho position of the phe-
nyl ring is near the side chain of Asn 352, a hydrogen-
bond acceptor at this position is favorable to form the
hydrogen-bonding interaction with this residue (for
example, compound 9q).
O
N
R
O
15a R = -CH3
15b R = Ph
15c R = -CH2Ph
15d R = -CH2CH2Ph
15e R = -CH=CHPh
15f R = -CH2OPh
O
N
O
4. Conclusions
O
OH
In response to the emergent need for novel antitumor
agents with improved activity, a series of 7-substituted
hCPTs were designed and synthesized. A new syn-
thetic route was developed based on the coupling of
ring A with ring CDE to effectively synthesize di-
versely substituted hCPT analogues. The in vitro anti-
tumor assay revealed that most of the compounds
exhibited good cytotoxic activities against the A549
cell line and several compounds are more potent than
topotecan. In general, the 7-iminomethyl hCPT deriv-
atives were more potent than 7-substituted pyridine
quaternary-salt hCPTs. The SAR analysis of the syn-
thesized compounds was explained in detail by the
3D-QSAR studies. The best derived CoMFA and
CoMSIA models showed a predictive q2 value of
0.703 and 0.717, and the activities of compounds in
both training set and test set were predicted with high
accuracy. Molecular docking was further used to clar-
ify the binding mode of the new hCPTs. They interca-
late at the DNA cleavage site and formed base-
stacking interactions with the base pairs. Three hydro-
gen bonds are formed between the heterocyclic ring of
hCPT and the active site of Topo I. The C7-side chain
is extended into the major groove of DNA and mainly
forms hydrophobic interaction with the active site res-
idues. The addition of small hydrophobic groups or
hydrogen-bond acceptors on the proper position of
the phenyl ring of C-7 side chain can improve the
hydrophobic or hydrogen-bonding interaction with
the enzyme. The results obtained from 3D-QSAR
studies show good correlation with the docking mod-
els, which allows for the rational design of more po-
tent hCPT derivatives.
Figure 1. Structures of the 10-ester of derivatives of hCPT as external
test set in 3D-QSAR studies.
Figure 2. Results of 31 possible CoMSIA field combinations using
q2SAMPLS. e: electrosteric field; s: steric field; a: hydrogen-bond field; h:
hydrophobic field; d: hydrogen-bond donor field.
phenyl ring can reinforce the hydrophobic and van der
Waals interactions with the residues in the binding site.
However, the space at the para position of the phenyl
ring is limited and the bulky group at this position
Table 4. The statistical parameters for the final CoMFA and CoMSIA models
q2
r2
Standard error
F
n
Fraction
Steric
Electrostatic
Hydrophobic
HB acceptor
0.276
CoMFA
CoMSIA
0.703
0.717
0.905
0.938
0.395
0.299
56.599
63.818
5
5
0.728
0.196
0.272
0.150
0.379

1500
Z. Miao et al. / Bioorg. Med. Chem. 16 (2008) 1493–1510
Table 5. Experimental and calculated antitumor activity (ꢀlgIC50
,
respectively. ESI mass spectra were performed on an
API-3000 LC–MS spectrometer. Elemental analyses
were performed with a MOD-1106 instrument and were
consistent with theoretical values within 0.4%. TLC
analysis was carried out on silica gel plates GF254 (Qin-
dao Haiyang Chemical, China). Flash column chroma-
tography was carried out on silica gel 300–400 mesh.
Anhydrous solvent and reagents were all analytically
pure and dried through routine protocols.
*
lM) of compounds in training set and test set (marked with
)
Compound
Obsd
CoMFA calcd
CoMSIA calcd
7
1.7696
1.4003
1.5786
2.8781
3.8753
2.3827
2.9200
2.2313
0.2434
2.7289
2.8640
ꢀ0.7070
2.4187
1.9663
2.8683
2.4931
2.4180
0.9453
2.4134
3.2362
ꢀ1.0707
3.3716
2.5638
3.9991
2.8658
3.2248
2.1994
3.7429
0.2911
ꢀ0.3447
ꢀ0.6573
0.4774
0.5418
ꢀ0.4265
2.1168
ꢀ1.2113
ꢀ0.4752
ꢀ0.1803
ꢀ1.1207
ꢀ1.4313
0.1682
3.7292
ꢀ0.8010
ꢀ0.1812
ꢀ0.0801
ꢀ0.5512
4.1004
3.5579
3.4966
3.7823
2.6084
1.3937
2.0563
1.6033
2.9643
3.7073
2.2391
2.6737
2.8366
0.1687
2.6188
2.5817
ꢀ0.8518
2.0102
1.1745
2.6510
2.2087
2.8484
0.4856
2.6151
3.6117
ꢀ1.0228
3.5376
2.8334
3.7288
3.3798
3.2356
2.6543
3.9463
ꢀ0.0198
ꢀ0.3606
ꢀ0.0400
0.2047
0.4443
0.0424
2.1247
ꢀ1.3005
ꢀ0.2600
ꢀ0.1469
ꢀ1.1745
ꢀ1.2879
0.1112
3.6672
ꢀ0.6421
ꢀ0.2731
ꢀ0.4118
ꢀ0.7739
3.7337
6.6609
3.4747
4.1249
2.7511
0.7911
8
9a
1.5850
2.6990
3.6990
1.4202
1.0315
2.6990
0.0820
3.0969
2.5229
ꢀ0.7651
1.8861
0.6459
3.6990
3.2218
3.6990
0.2692
2.0458
3.6990
ꢀ1.1438
3.6990
3.6990
3.6990
3.0458
3.6990
2.6990
3.6990
ꢀ0.0500
ꢀ0.3237
ꢀ0.3786
0.5045
0.5376
0.4191
2.1549
ꢀ1.2620
ꢀ0.4548
ꢀ0.5428
ꢀ1.2803
ꢀ1.3010
0.1373
3.6990
ꢀ0.7003
ꢀ0.3164
ꢀ0.3251
ꢀ0.8643
3.6990
3.6990
3.6990
3.6990
2.6990
1.1675
9b
9c
9d
9e
9f*
9g
5.1.1. 1,1-(Ethylenedioxy)-8-ethyl-7,8-dihydroxy-2,3,7,8-
tetrahydro-5H-6-oxa-3a-aza-cyclopenta[b]naphthalene-1,
4-dione (2). Potassium borohydride (1.0 g, 18.6 mmol)
was added in portions, during 10 min, to a solution of
1,1-(ethylenedioxy)-8-Ethyl-8-hydroxy-2,3,5,8-tetrahy-
dro-6-oxa-3a-aza-cyclopenta[b]naphthal-ene-1,4,7-tri-
one (1) (8.0 g, 26.1 mmol) in anhydrous methanol
(170 mL). The reaction mixture was stirred for 20 min
at 25 ꢁC.The solution was concentrated in vacuo, diluted
with water, and brought to pH 4 by addition of 2 M
HCl. The solid was filtered off to give 2 as a white solid:
9h
9i*
9j
9k
9l
9m
9n
9o
9p
9q
9r*
1
yield 6.4 g (79.5%). H NMR (DMSO-d6): 0.92 (t, 3H,
9s
J = 7.6 Hz), 1.75 (q, 2H, J = 7.5 Hz), 2.41 (t, 2H,
J = 6.8 Hz), 2.72 (s, 1H), 3.31 (s, 1H), 4.11–4.25 (m,
4H), 4.19 (d, 2H, J = 7.0 Hz), 4.47–4.69 (q, 2H,
J = 16.8 Hz), 5.06 (s, 1H) and 6.71 (s, 1H). MS (ESI):
m/z 310 (M+H)+. Anal. Calcd for C15H19NO6: C,
58.25; H, 6.19; N, 4.53. Found: C, 58.27; H, 6.20; N,
4.51.
9t
9u
9v
9w
9x
9y
13a
13b
13c
13d
13e
13f
13g*
13h
13i
13j
13k
13l
13m
13n*
14a
14b
14c
14d
15a*
15b*
15c*
15d*
15e*
15f*
5.1.2. Formic acid-1,1-(ethylenedioxy)-1,5-dioxo-7-propi-
onyl-1,2,3,5-tetrahydro-indolizin-6-yl-methyl ester (3). To
a suspension of 2 (4.0 g, 12.9 mmol) in glacial acetic acid
(200 mL) was added dropwise a solution of NaIO4
(4.0 g, 18.7 mmol) in H2O (40 mL). The resulting mix-
ture was stirred at room temperature for 30 min, at
which time ethylene glycol (4 mL) was added. Water
(400 mL) was slowly added and extracted with dichloro-
methane (3· 100 mL). The combined dichloromethane
extracts were dried over sodium sulfate and concen-
trated under reduced pressure, and the residue was crys-
tallized from THF yielding 3 as a white solid: 3.7 g
1
(93.1%). H NMR (CDCl3): 1.16 (t, 3 H, J = 7.2 Hz),
2.45 (t, 2 H, J = 6.8 Hz), 2.89 (q, 2H, J = 7.1 Hz),
4.10–4.23 (m, 6H), 5.13 (s, 2H), 6.57 (s, 1H), 8.07 (s,
1H). MS (ESI): 308 (M+H)+. Anal. Calcd for
C15H17NO6: C, 58.63; H, 5.58; N, 4.56. Found: C,
58.47; H, 5.61; N, 4.58.
5.1.3. 1,1-(Ethylenedioxy)-3-hydroxy-3-(6-hydroxymethyl-
1,5-dioxo-1,2,3,5-tetrahydro-indolizin-7-yl)-pentanoic acid
tert-butyl ester (4). A suspension of zinc (9.0 g,
137.6 mmol) was stirred in anhydrous THF (120 mL)
under N2 and then heated to reflux. The heating bath
was removed and tert-butyl bromoacetate (20 mL,
133 mmol) was added dropwise at a reflux-maintaining
rate. Then a solution of 3 (5.5 g, 17.9 mmol) in anhy-
drous THF (50 mL) was added and continued for
0.5 h further. The reaction mixture was allowed to cool
to room temperature, quenched with saturated
ammonium chloride (100 mL), and extracted with
dichloromethane (3· 100 mL). The combined dichloro-
methane extracts were dried over sodium sulfate and
5. Experimental
5.1. General procedures
Melting points were measured on an electrically heated
XT4A instrument and are uncorrected. The H spectra
were recorded at 500 MHz with a Bruker instrument,
and reported with TMS as internal standard and CDCl3
or DMSO-d6 as solvents. Chemical shifts (d values) and
coupling constants (J values) are given in ppm and Hz,
1

Z. Miao et al. / Bioorg. Med. Chem. 16 (2008) 1493–1510
1501
MS (ESI): 278 (M+H)+. Anal. Calcd for C14H15NO5:
C, 60.64; H, 5.45; N, 5.05. Found: C, 60.66; H, 5.43;
N, 5.02.
5.1.5.
5-Ethyl-1,4,5,13-tetrahydro-5-hydroxy-3H,15H-
oxepino-[30,40:6,7]indolizino[1,2-b]quinoline-3,15-dione
(6). A solution of 2-[1,3]dioxolan-2-yl-phenyl-amine
(0.8 g, 4.8 mmol) and tricyclic ketone 5 (1.0 g, 3.6 mmol)
in toluene (800 mL) was refluxed using a Dean–Stark
trap for 30 min. p-Toluenesulfonic acid (0.1 g) was then
added and refluxing was continued for an additional 1 h.
The solution was allowed to cool to room temperature
and the solid was filtered off and washed by acetone
(20 mL) and methanol (20 mL) to give homocamptothe-
cin 6 as yellow solid: yield 0.98 g (75.5%). 1H NMR
(DMSO-d6): 0.87 (t, 3 H, J = 7.2 Hz), 1.86 (q, 2H,
J = 7.5 Hz), 3.07–3.48 (q, 2H, J = 14.0 Hz), 5.28 (s,
2H), 5.42–5.54 (q, 2H, J = 15.0 Hz), 6.03 (s, 1H), 7.42
(s, 1H), 7.72 (t, 1H, J = 7.3 Hz), 7.87 (t, 1H,
J = 7.2 Hz), 8.15 (q, 2H, J = 7.9 Hz), 8.69 (s, 1H). MS
(ESI): 363 (M+H)+. Anal. Calcd for C21H18N2O4: C,
69.60; H, 5.01; N 7.73. Found C, 69.37; H, 5.05; N, 7.77.
5.1.6. 7-Hydroxymethylhomocamptothecin (7). To a sus-
pension of 6 (0.98 g, 2.7 mmol) in a mixture of MeOH
(30 mL) and H2O (25 mL), 75% H2SO4 (25 mL) was
added dropwise, and then FeSO4Æ7H2O (0.80 g,
2.9 mmol) was added. To the ice-cold mixture, 30%
H2O2 (5 mL, 2.2 mmol) was added dropwise and stirred
at room temperature for 14 h and then diluted with
H2O. The solid was filtered and purified by chromatog-
raphy over silica gel (4% MeOH/CH2Cl2) to give 7 as a
yellow solid: 0.61 g (64.7%). 1H NMR (DMSO-d6): 0.88
(t, 3H, J = 7.4 Hz), 1.87 (q, 2 H, J = 7.4 Hz), 3.06–3.48
(q, 2H, J = 13.8 Hz), 5.27 (d, 2H, J = 5.3 Hz), 5.41 (s,
2H), 5.41–5.55 (q, 2H, J = 14.7 Hz), 5.80 (t, 1H,
J = 5.6 Hz), 6.03 (s, 1H), 7.40 (s, 1H), 7.71 (t, 1H,
J = 7.4 Hz), 7.86 (t, 1H, J = 7.4 Hz), 8.16 (d, 1H,
J = 8.3 Hz), 8.20 (d, 1H, J = 8.3 Hz). MS (ESI):393
(M+H)+. Anal. Calcd for C22H20N2O5: C 67.34, H
5.14, N 7.14. Found C, 67.12; H, 5.17; N, 7.19.
Figure 3. Plots of the experimental and calculated pIC50 (ꢀlgIC50, lM)
values for the CoMFA (a) and CoMSIA (b) analysis of the training set
and test set.
concentrated under reduced pressure, and the residue
was purified by chromatography over silica gel
(2%MeOH/CH2Cl2) to give ester 4 as a pale-yellow so-
lid: 6.9 g (97.5%). 1H NMR (CDCl3): 0.85 (t, 3H,
J = 7.4 Hz), 1.30 (s, 9H), 1.85–1.98 (m, 2H), 2.40 (t,
2H, J = 7.0 Hz), 2.84 (d, 2H, J = 16.0 Hz), 4.06–4.20
(m, 6H), 4.20 (s, 1H), 4.80 (s, 1H), 4.88 (q, 2H,
J = 13.0 Hz), 6.52 (s, 1H). MS (ESI): 396 (M+H)+.
Anal. Calcd for C20H29NO7: C, 60.74; H, 7.39; N,
3.54; Found: C 60.59, H 7.42, N 3.55.
5.1.7. 7-Formylhomocamptothecin (8). A suspension of 7
(0.61 g, 1.6 mmol) in AcOH (200 mL) was heated under
reflux for 6 h. The mixture was evaporated to dryness
and the residue was purified by chromatography over
silica gel (4% MeOH/CH2Cl2) to give 8 as a yellow solid:
0.32 g (52.7%). 1H NMR (DMSO-d6): 0.87 (t,
J = 7.3 Hz, 3H), 1.88 (q, 2H, J = 7.4 Hz), 3.07–3.49 (q,
1H, J = 11.5 Hz), 5.42–5.55 (q, 1H, J = 15.3 Hz), 5.52
(s, 2H), 6.06 (s, 1H), 7.46 (s, 1H), 7.92 (t, 1H,
J = 7.3 Hz), 7.98 (t, 1H, J = 7.4 Hz), 8.31 (d, 1H,
J = 8.2 Hz), 9.04 (d, 1H, J = 8.2 Hz), 11.10 (s, 1H).
MS (ESI): 391 (M+H)+. Anal. Calcd for C22H18N2O5:
C, 67.69; H, 4.65; N, 7.18. Found C, 67.49; H, 4.67;
N, 7.21.
5.1.4. 9-Ethyl-9-hydroxy-2,3,8,9-tetrahydro-5H-6-oxa-3a-aza-
cyclohepta[f]indene-1,4,7-trione (5). Ester 4 (4.6 g, 11.6
mmol) was dissolved in trifluoroacetic acid (28 mL)
and stirred at room temperature for 10 h. The solution
was concentrated under reduced pressure and the resi-
due was crystallized from acetone/water yielding 5 as a
5.1.8. 7-Phenyliminomethylhomocamptothecin (9a). To a
suspension of Yb(OTf)3 (16 mg, 0.03 mmol) in 15 mL
CH2Cl2, 8 (50 mg, 0.13 mmol) and aniline (70 mg,
0.75 mmol) were added. The resulting mixture was stir-
red at room temperature for 10 h until the reaction
was complete. The solvent was evaporated to dryness
1
white solid: 3.0 g (93.0%). H NMR (CDCl3): 0.90 (t,
3 H, J = 7.5 Hz), 1.78–1.88 (m, 2H), 2.97 (t, 2H,
J = 6.7 Hz), 3.31 (q, 2H, J = 13.9 Hz), 4.26 (t, 2H,
J = 6.7 Hz), 5.45 (q, 2H, J = 15.7 Hz), 7.04 (s, 1H).

1502
Z. Miao et al. / Bioorg. Med. Chem. 16 (2008) 1493–1510
Figure 4. Contour plots of CoMFA steric fields (a, compound 13f is shown in ball-and-stick mode), electrostatic fields (b, compound 9q is shown in
ball-and-stick mode), CoMSIA hydrophobic fields (c, compound 9n is shown in ball-and-stick mode), and CoMSIA hydrogen-bonding acceptor
fields (d, compound 9o is shown in ball-and-stick mode).
and the residue was purified by chromatography over
silica gel(2% MeOH/CH2Cl2) to give 9a as a yellow so-
5.1.9. 7-(2-Methylphenyl)iminomethylhomocamptothecin
(9b). The titled compound was prepared from 8 and 2-
methylaniline according to the method of compound
9a, yield: 24.4%. 1H NMR (DMSO-d6): 0.89 (t, 3H,
J = 7.3 Hz), 1.88 (q, 2H, J = 8.1 Hz), 2.50 (s, 3H),
3.08–3.48 (q, 2H, J = 14.0 Hz), 5.42–5.54 (q, 2H,
J = 15.0 Hz), 5.59 (s, 2H), 6.05 (s, 1H), 7.29 (t, 1H,
J = 6.7 Hz), 7.35 (t, 1H, J = 6.7 Hz), 7.38 (d, 1H,
J = 7.4 Hz), 7.46 (s, 1H), 7.50 (d, 1H, J = 7.4 Hz), 7.83
(t, 1H, J = 7.5 Hz), 7.95 (t, 1H, J = 7.5 Hz), 8.26 (d,
1H, J = 8.5 Hz), 8.99 (d, 1H J = 8.5 Hz), 9.66 (s, 1H).
MS (ESI): 480 (M+H)+. Anal. Calcd for C29H25N3O4:
1
lid: 14 mg (23.5%). H NMR (DMSO-d6): 0.89 (t, 3H,
J = 7.4 Hz), 1.88 (q, 2H, J = 7.4 Hz), 3.08–3.48 (q, 2H,
J = 14.0 Hz), 5.42–5.55 (q, 2H, J = 15.1 Hz), 5.60 (s,
2H), 6.05 (s, 1H), 7.41 (t, 2H, J = 7.2 Hz), 7.46 (s,
1H), 7.54 (d, 2H, J = 7.5 Hz), 7.59 (m, 1H), 7.84 (t,
1H, J = 7.3 Hz), 7.96 (t, 1H, J = 7.3 Hz), 8.27 (d, 1H,
J = 8.1 Hz), 9.01 (d, 1H, J = 8.1 Hz), 9.72 (s, 1H). MS
(ESI): 466 (M+H)+. Anal. Calcd for C28H23N3O4: C,
72.24; H, 4.98; N, 9.03. Found C, 72.36; H, 4.97; N,
9.01.

Z. Miao et al. / Bioorg. Med. Chem. 16 (2008) 1493–1510
1503
Figure 5. Stereoview of the binding mode of 9l in the active site of Topo I. The base pairs forming stack interaction with 9l are colored purple and
residues forming hydrogen-bonding interaction with 9l are colored red. The hydrogen bonds are displayed in dotted lines.
C, 72.64; H, 5.25; N, 8.76. Found C, 72.39; H, 5.27; N,
8.77.
bromoaniline according to the method of compound
9a, yield: 20.1%. 1H NMR (DMSO-d6): 0.89 (t, 3H,
J = 7.3 Hz), 1.88 (q, 2H, J = 7.3 Hz), 3.08–3.48 (q, 2H,
J = 14.0 Hz), 5.42–5.55 (q, 2H, J = 15.1 Hz), 5.58 (s,
2H), 6.05 (s, 1H), 7.46 (s, 1H), 7.56 (d, 2H,
J = 8.6 Hz), 7.74 (d, 2H, J = 8.6 Hz), 7.84 (t, 1H,
J = 7.5 Hz), 7.95 (t, 1H, J = 7.5 Hz), 8.27 (d, 1H,
J = 8.5 Hz), 8.99 (d, 1H, J = 8.5 Hz), 9.72 (s, 1H). MS
(ESI): 545 (M+H)+. Anal. Calcd for C28H22BrN3O4:
C, 61.77; H, 4.07; N, 7.72. Found C, 61.88; H, 4.05;
N, 7.71.
5.1.10. 7-(4-Methylphenyl)iminomethylhomocamptothe-
cin (9c). The titled compound was prepared from 8
and 4-methylaniline according to the method of com-
1
pound 9a, yield: 26.0%. H NMR (DMSO-d6): 0.87 (t,
3H, J = 7.2 Hz), 1.87 (q, 2H, J = 8.5 Hz), 2.40 (s, 3H),
3.08–3.48 (q, 2H, J = 13.7 Hz), 5.42–5.55 (q, 2H,
J = 15.1 Hz), 5.59 (s, 2H), 6.05 (s, 1H), 7.36 (d, 2H,
J = 8.0 Hz), 7.46 (s, 1H), 7.53 (d, 2H, J = 8.0 Hz), 7.84
(t, 1H, J = 7.5 Hz), 7.94 (t, 1H, J = 7.5 Hz), 8.26 (d,
1H, J = 8.5 Hz), 9.00 (d, 1H, J = 8.5 Hz), 9.72 (s, 1H).
MS (ESI): 480 (M+H)+. Anal. Calcd for C29H25N3O4:
C, 72.64; H, 5.25; N, 8.76. Found C, 72.41; H, 5.27;
N, 8.78.
5.1.13. 7-(3,5-Dichlorophenyl)iminomethylhomocampto-
thecin (9f). The titled compound was prepared from 8
and 3,5-dichloroaniline according to the method of com-
1
pound 9a, yield: 30.7%. H NMR (DMSO-d6): 0.89 (t,
3H, J = 7.0 Hz), 1.89 (q, 2H, J = 7.1 Hz), 3.08–3.47 (q,
2H, J = 13.9 Hz), 5.42–5.56 (q, 2H, J = 15.1 Hz), 5.59
(s, 2H), 6.06 (s, 1H), 7.46 (s, 1H), 7.63 (s, 1H), 7.68 (s,
2H), 7.84 (t, 1H, J = 7.9 Hz), 7.96 (t, 1H, J = 7.9 Hz),
8.27 (d, 1H, J = 8.5 Hz), 9.01 (d, 1H, J = 8.5 Hz), 9.74
(s, 1H). MS (ESI): 535 (M+H)+. Anal. Calcd for
C28H21Cl2N3O4: C, 62.93; H, 3.96; N, 7.86. Found C,
62.80; H, 3.98; N, 7.87.
5.1.11. 7-(3-Methylphenyl)iminomethylhomocamptothe-
cin (9d). The titled compound was prepared from 8
and 3-methylaniline according to the method of com-
1
pound 9a, yield: 24.4%. H NMR (DMSO-d6): 0.89 (t,
3H, J = 7.3 Hz), 1.88 (q, 2H, J = 8.2 Hz), 2.44 (s, 3H),
3.09–3.48 (q, 2H, J = 13.9 Hz), 5.42–5.55 (q, 2H,
J = 15.0 Hz), 5.60 (s,2H), 6.04 (s, 1H), 7.22 (t, 1H,
J = 7.1 Hz), 7.39 (d, 1H, J = 8.0 Hz), 7.41 (s, 1H), 7.43
(d, 1H, J = 8.0 Hz), 7.46 (s, 1H), 7.84 (t, 1H,
J = 7.4 Hz), 7.96 (t, 1H, J = 7.4 Hz), 8.27 (d, 1H,
J = 8.6 Hz), 9.00 (d, 1H, J = 8.6 Hz), 9.71 (s, 1H). MS
(ESI): 480 (M+H)+. Anal. Calcd for C29H25N3O4: C,
72.64; H, 5.25; N, 8.76. Found C, 72.83; H, 5.23; N, 8.75.
5.1.14. 7-(3-Fluorophenyl)iminomethylhomocamptothecin
(9g). The titled compound was prepared from 8 and 3-
fluoroaniline according to the method of compound
9a, yield: 14.5%. 1H NMR (DMSO-d6): 0.87 (t, 3H,
J = 7.4 Hz), 1.87 (q, 2H, J = 8.4 Hz), 3.07–3.47 (q, 2H,
J = 13.8 Hz), 5.42–5.55 (q, 2H, J = 15.1 Hz), 5.52 (s,
2H), 6.05 (s, 1H), 7.25 (d, 1H, J = 8.1 Hz), 7.46 (s,
1H), 7.49 (d, 1H, J = 8.1 Hz), 7.53 (t, 1H, J = 7.2 Hz),
5.1.12. 7-(4-Bromophenyl)iminomethylhomocamptothecin
(9e). The titled compound was prepared from 8 and 4-

1504
Z. Miao et al. / Bioorg. Med. Chem. 16 (2008) 1493–1510
7.89 (s, 1H), 7.91 (t, 1H, J = 7.5 Hz), 7.98 (t, 1H,
J = 7.5 Hz), 8.30 (d, 1H, J = 8.4 Hz), 9.03 (d, 1H,
J = 8.4 Hz), 9.73 (s, 1H). MS (ESI): 484 (M+H)+. Anal.
Calcd for C28H21FN3O4: C, 69.56; H, 4.59; N, 8.69.
Found C, 69.77; H, 4.58; N, 8.69.
J = 7.5 Hz), 7.95 (t, 1H, J = 7.5 Hz), 7.97 (s, 1H), 8.29
(d, 1H, J = 8.5 Hz), 9.06 (d, 1H, J = 8.5 Hz), 9.78 (s,
1H). MS (ESI): 535 (M+H)+. Anal. Calcd for
C28H21Cl2N3O4: C, 62.93; H, 3.96; N, 7.86. Found C,
62.88; H, 3.97; N, 7.88.
5.1.15. 7-(3,4-Difluorophenyl)iminomethylhomocampto-
thecin (9h). The titled compound was prepared from 8
and 3,4-difluoroaniline according to the method of com-
5.1.19. 7-(3,5-Dimethylphenyl)iminomethylhomocampto-
thecin (9l). The titled compound was prepared from 8
and 3,5-dimethylaniline according to the method of
compound 9a, yield: 23.7%. 1H NMR (DMSO-d6):
0.89 (t, 3H, J = 7.2 Hz), 1.89 (q, 2H, J = 8.2 Hz), 2.39
(s, 6H), 3.08–3.47 (q, 2H, J = 14.0 Hz), 5.42–5.56 (q,
2H, J = 15.0 Hz), 5.58 (s, 2H), 6.07 (s, 1H), 7.04 (s,
1H), 7.32 (s, 2H), 7.46 (s, 1H), 7.84 (t, 1H,
J = 7.7 Hz), 7.96 (t, 1H, J = 7.7 Hz), 8.27 (d, 1H,
J = 8.1 Hz), 8.99 (d, 1H, J = 8.1 Hz), 9.69 (s, 1H). 13C
NMR (DMSO-d6): 8.63 (1C), 21.31 (2C), 29.42 (1C),
36.68 (1C), 42.79 (1C), 53.00 (1C), 61.73 (1C), 73.54
(1C), 99.93 (1C), 119.84 (2C), 123.10 (1C), 124.98
(1C), 126.23 (1C), 128.76 (1 C), 129.48 (1C), 130.18
(1C), 130.79 (1C), 134.09 (1C), 138.98 (1 C), 144.43
(1C), 149.51 (1C), 150.90 (1C), 153.52 (1C), 156.17 (1
C), 159.41 (1C), 161.31 (1C), 170.87 (1C), 172.24 (1C).
MS (ESI): 494 (M+H)+. Anal. Calcd for C30H27N3O4:
C, 73.01; H, 5.51; N, 8.51. Found C, 73.21; H, 5.51;
N, 8.50.
1
pound 9a, yield: 17.1%. H NMR (DMSO-d6): 0.88 (t,
3H, J = 7.1 Hz), 1.88 (q, 2H, J = 7.4 Hz), 3.09–3.48 (q,
2H, J = 13.8 Hz), 5.42–5.55 (q, 2H, J = 15.0 Hz), 5.58
(s, 2H), 6.06 (s, 1H), 7.46 (s, 1H), 7.47 (s, 1H), 7.62 (d,
1H, J = 8.1 Hz), 7.83 (d, 1H, J = 8.1 Hz), 7.85 (t, 1H,
J = 7.5 Hz), 7.96 (t, 1H, J = 7.5 Hz), 8.28 (d, 1H,
J = 8.6 Hz), 9.01 (d, 1H, J = 8.6 Hz), 9.73 (s, 1H). MS
(ESI): 502 (M+H)+. Anal. Calcd for C28H21F2N3O4: C,
67.06; H, 4.22; N, 8.38. Found C, 67.23; H, 4.21; N, 8.35.
5.1.16. 7-(3-Chlorophenyl)iminomethylhomocamptothecin
(9i). The titled compound was prepared from 8 and 3-
chloroaniline according to the method of compound
9a, yield: 23.4%. 1H NMR (DMSO-d6): 0.89 (t, 3H,
J = 7.3 Hz), 1.88 (q, 2H, J = 8.3 Hz), 3.08–3.48 (q, 2H,
J = 13.9 Hz), 5.42–5.55 (q, 2H, J = 15.1 Hz), 5.59 (s,
2H), 6.05 (s, 1H), 7.43 (d, 1H, J = 8.2 Hz), 7.46 (s,
1H), 7.52 (d, 1H, J = 8.2 Hz), 7.57 (t, 1H, J = 7.4 Hz),
7.71 (s, 1H), 7.83 (t, 1H, J = 7.6 Hz), 7.95 (t, 1H,
J = 7.6 Hz), 8.27 (d, 1H, J = 8.5 Hz), 9.02 (d, 1H,
J = 8.5 Hz), 9.73 (s, 1H). MS (ESI): 501 (M+H)+. Anal.
Calcd for C28H22ClN3O4: C, 67.27; H, 4.44; N, 8.40.
Found C, 67.14; H, 4.45; N, 8.43.
5.1.20. 7-(3,4-Dimethylphenyl)iminomethylhomocampto-
thecin (9m). The titled compound was prepared from 8
and 3,4-dimethylaniline according to the method of
compound 9a, yield: 25.3%. 1H NMR (DMSO-d6):
0.88 (t, 3H, J = 7.3 Hz), 1.88 (q, 2H, J = 8.2 Hz), 2.31
(s, 3H), 2.34 (s, 3H), 3.09–3.48 (q, 2H, J = 13.8 Hz),
5.42–5.55 (q, 2H, J = 15.0 Hz), 5.58 (s, 2H), 6.05 (s,
1H), 7.30 (d, 1H, J = 8.3 Hz), 7.36 (d, 1H, J = 8.3 Hz),
7.43 (s, 1H), 7.46 (s, 1H), 7.83 (t, 1H, J = 7.4 Hz), 7.95
(t, 1H, J = 7.4 Hz), 8.26 (d, 1H, J = 8.6 Hz), 9.00 (d,
1H, J = 8.6 Hz), 9.72 (s, 1H). MS (ESI): 494 (M+H)+.
Anal. Calcd for C30H27N3O4: C, 73.01; H, 5.51; N,
8.51. Found C, 73.19; H, 5.49; N, 8.51.
5.1.17.
7-(3-Chloro-4-methylphenyl)iminomethylhomo-
camptothecin (9j). The titled compound was prepared
from 8 and 3-chloro-4-methylaniline according to the
method of compound 9a, yield: 19.7%. 1H NMR
(DMSO-d6): 0.88 (t, 3H, J = 7.3 Hz), 1.88 (q, 2H,
J = 7.4 Hz), 2.36 (s, 3H), 3.08–3.48 (q, 2H,
J = 13.8 Hz), 5.42–5.55 (q, 2H, J = 15.2 Hz), 5.58 (s,
2H), 6.06 (s, 1H), 7.46 (s, 1H), 7.50 (d, 1H,
J = 8.1 Hz), 7.52 (d, 1H, J = 8.1 Hz), 7.75 (s, 1H), 7.84
(t, 1H, J = 7.4 Hz), 7.95 (t, 1H, J = 7.4 Hz), 8.26 (d,
1H, J = 8.6 Hz), 9.02 (d, 1H, J = 8.6 Hz), 9.74 (s, 1H).
13C NMR (DMSO-d6): 8.56 (1C), 19.66 (1C), 29.35 (1
C), 36.61 (1C), 42.73 (1C), 52.92 (1C), 61.66 (1C),
73.47 (1C), 99.89 (1C), 121.84 (1C), 123.07 (1C),
125.05 (1C), 126.15 (1C), 128.75 (1C), 129.03 (1C),
130.09 (1C), 130.77 (1C), 132.13 (1C), 133.86 (1C),
134.30 (1C), 135.10 (1C), 144.31 (1C), 149.44 (1C),
149.79 (1C), 153.45 (1C), 156.10 (1C), 157.45 (1C),
159.34 (1C), 172.18 (1C). MS (ESI): 515 (M+H)+. Anal.
Calcd for C29H24ClN3O4: C, 67.77; H, 4.71; N, 8.18.
Found C, 67.58; H, 4.72; N, 8.20.
5.1.21. 7-(4-Chlorophenyl)iminomethylhomocamptothecin
(9n). The titled compound was prepared from 8 and 4-
chloroaniline according to the method of compound
9a, yield: 23.4%. 1H NMR (DMSO-d6): 0.89 (t, 3H,
J = 7.2 Hz), 1.88 (q, 2H, J = 8.3 Hz), 3.09–3.48 (q, 2H,
J = 13.8 Hz), 5.42–5.52 (q, 2H, J = 15.0 Hz), 5.47 (s,
2H), 6.05 (s, 1H), 7.44 (s, 1H), 7.59 (d, 2H, J = 8.2 Hz),
7.61 (d, 2H, J = 8.2 Hz), 7.81 (t, 1H, J = 7.5 Hz), 7.93 (t,
1H, J = 7.5 Hz), 8.23 (d, 1H, J = 8.6 Hz), 8.95 (d, 1H,
J = 8.6 Hz), 9.68 (s, 1H). MS (ESI): 501 (M+H)+. Anal.
Calcd for C28H22ClN3O4: C, 67.27; H, 4.44; N, 8.40.
Found C, 67.17; H, 4.44; N, 8.42.
5.1.22. 7-(4-Cyanophenyl)iminomethylhomocamptothecin
(9o). The titled compound was prepared from 8 and 4-
aminobenzonitrile according to the method of com-
5.1.18. 7-(2,5-Dichlorophenyl)iminomethylhomocampto-
thecin (9k). The titled compound was prepared from 8
and 2,5-dichloroaniline according to the method of com-
1
pound 9a, yield: 15.9%. H NMR (DMSO-d6): 0.88 (t,
1
pound 9a, yield: 24.8%. H NMR (DMSO-d6): 0.89 (t,
3H, J = 7.3 Hz), 1.88 (q, 2H, J = 8.0 Hz), 3.08–3.48 (q,
2H, J = 13.8 Hz), 5.42–5.55 (q, 2H, J = 15.2 Hz), 5.59 (s,
2H), 6.06 (s, 1H), 7.47 (s, 1H), 7.68 (d, 2H, J = 8.1 Hz),
7.85 (t, 1H, J = 7.5 Hz), 7.97 (t, 1H, J = 7.5 Hz), 8.02 (d,
2H, J = 8.1 Hz), 8.28 (d, 1H, J = 8.5 Hz), 8.98 (d, 1H,
3H, J = 7.2 Hz), 1.88 (q, 2H, J = 8.1 Hz), 3.08–3.48 (q,
2H, J = 13.9 Hz), 5.42–5.54 (q, 2H, J = 14.8 Hz), 5.63
(s, 2H), 6.06 (s, 1H), 7.46 (s, 1H), 7.48 (d, 1H,
J = 8.1 Hz), 7.69 (d, 1H, J = 8.1 Hz), 7.86 (t, 1H,

Z. Miao et al. / Bioorg. Med. Chem. 16 (2008) 1493–1510
1505
J = 8.5 Hz), 9.71 (s, 1H). MS (ESI): 491 (M+H)+. Anal.
Calcd for C29H22N4O4: C, 71.01; H, 4.52; N, 11.42. Found
C, 71.22; H, 4.51; N, 11.38.
method of compound 9a, yield: 22.6%. 1H NMR
(DMSO-d6): 0.89 (t, 3H, J = 7.4 Hz), 1.88 (q, 2H,
J = 8.0 Hz), 3.09–3.48 (q, 2H, J = 13.9 Hz), 5.42–5.55
(q, 2H, J = 15.0 Hz), 5.57 (s, 2H), 6.05 (s, 1H), 7.45 (s,
1H), 7.58 (d, 1H, J = 8.3 Hz), 7.61 (d, 1H, J = 8.3 Hz),
7.84 (t, 1H, J = 7.4 Hz), 7.94 (t, 1H, J = 7.4 Hz), 7.95
(s, 1H), 8.26 (d, 1H, J = 8.5 Hz), 9.01 (d, 1H,
J = 8.5 Hz), 9.74 (s, 1H). MS (ESI): 519 (M+H)+. Anal.
Calcd for C28H21FClN3O4: C, 64.93; H, 4.09; N, 8.11.
Found C, 65.11; H, 4.08; N, 8.10.
5.1.23. 7-(3,4-Dichlorophenyl)iminomethylhomocampto-
thecin (9p). The titled compound was prepared from 8
and 3,4-dichloroaniline according to the method of com-
1
pound 9a, yield: 24.8%. H NMR (DMSO-d6): 0.89 (t,
3H, J = 7.3 Hz), 1.88 (q, 2H, J = 7.4 Hz), 3.08–3.48 (q,
2H, J = 13.9 Hz), 5.42–5.54 (q, 2H, J = 14.6 Hz), 5.60
(s, 2H), 6.07 (s, 1H), 7.47 (s, 1H), 7.58 (d, 1H,
J = 8.2 Hz), 7.80 (d, 1H, J = 8.2 Hz), 7.85 (t, 1H,
J = 7.5 Hz), 7.95 (s, 1H), 7.96 (t, 1H, J = 7.5 Hz), 8.28
(d, 1H, J = 8.6 Hz), 9.01 (d, 1H, J = 8.6 Hz), 9.75 (s,
1H). MS (ESI): 535 (M+H)+. Anal. Calcd for
C28H21Cl2N3O4: C, 62.93; H, 3.96; N, 7.86. Found C,
63.11; H, 3.95; N, 7.86.
5.1.28. 7-(2-Methoxyphenyl)iminomethylhomocamptothe-
cin (9u). The titled compound was prepared from 8 and
2-methoxyaniline according to the method of compound
9a, yield: 25.2%. 1H NMR (DMSO-d6): 0.89 (t, 3H,
J = 7.4 Hz), 1.88 (q, 2H, J = 8.3 Hz), 3.09–3.48 (q, 2H,
J = 13.8 Hz), 3.92 (s, 3H), 5.42–5.54 (q, 2H,
J = 15.1 Hz), 5.57 (s, 2H), 6.04 (s, 1H), 7.09 (t, 1H,
J = 7.1 Hz), 7.19 (d, 1H, J = 8.0 Hz), 7.35 (t, 1H,
J = 7.1 Hz), 7.46 (s, 1H), 7.47 (d, 1H, J = 8.0 Hz), 7.83
(t, 1H, J = 7.6 Hz), 7.95 (t, 1H, J = 7.6 Hz), 8.26 (d,
1H, J = 8.5 Hz), 8.94 (d, 1H, J = 8.5 Hz), 9.67 (s, 1H).
MS (ESI): 496 (M+H)+. Anal. Calcd for C29H25N3O5:
C, 70.29; H, 5.09; N, 8.48. Found C, 70.42; H, 5.08;
N, 8.46.
5.1.24. 7-(2-Hydroxyphenyl)iminomethylhomocamptothe-
cin (9q). The titled compound was prepared from 8 and
2-aminophenol according to the method of compound
9a, yield: 16.2%. 1H NMR (DMSO-d6): 0.89 (t, 3H,
J = 7.5 Hz), 1.88 (q, 2H, J = 8.3 Hz), 3.08–3.48 (q, 2H,
J = 14.0 Hz), 5.42–5.55 (q, 2H, J = 15.4 Hz), 5.67 (s,
2H), 6.05 (s, 1H), 6.95 (t, 1H, J = 7.1 Hz), 6.96 (d, 1H,
J = 8.6 Hz), 7.01 (d, 1H, J = 8.6 Hz), 7.20 (t, 1H,
J = 7.1 Hz), 7.46 (s, 1H), 7.83 (t, 1H, J = 7.5 Hz), 7.96
(t, 1H, J = 7.5 Hz), 8.27 (d, 1H, J = 8.5 Hz), 8.93 (d,
1H, J = 8.5 Hz), 9.49 (s, 1H), 9.66 (s, 1H). MS (ESI):
482 (M+H)+. Anal. Calcd for C28H23N3O5: C, 69.84;
H, 4.81; N, 8.73. Found C, 70.00; H, 4.80; N, 8.71.
5.1.29. 7-(2,4-Dichlorophenyl)iminomethylhomocampto-
thecin (9v). The titled compound was prepared from 8
and 2,4-dichloroaniline according to the method of com-
1
pound 9a, yield: 24.8%. H NMR (DMSO-d6): 0.89 (t,
3H, J = 7.1 Hz), 1.88 (q, 2H, J = 8.3 Hz), 3.08–3.48 (q,
2H, J = 13.9 Hz), 5.42–5.54 (q, 2H, J = 14.6 Hz), 5.60
(s, 2H), 6.07 (s, 1H), 7.47 (s, 1H), 7.58 (d, 1H,
J = 8.2 Hz), 7.80 (d, 1H, J = 8.2 Hz), 7.85 (t, 1H,
J = 7.4 Hz), 7.95 (s, 1H), 7.96 (t, 1H, J = 7.4 Hz), 8.28
(d, 1H, J = 8.6 Hz), 9.01 (d, 1H, J = 8.6 Hz), 9.75 (s,
1H). MS (ESI): 535 (M+H)+. Anal. Calcd for
C28H21Cl2N3O4: C, 62.93; H, 3.96; N, 7.86. Found C,
63.10; H, 3.96; N, 7.84.
5.1.25. 7-(2,4-Dimethylphenyl)iminomethylhomocampto-
thecin (9r). The titled compound was prepared from 8
and 2,4-dimethylaniline according to the method of
compound 9a, yield: 25.3%. 1H NMR (DMSO-d6):
0.89 (t, 3H, J = 7.3 Hz), 1.88 (q, 2H, J = 8.1 Hz), 2.30
(s, 3H), 2.35 (s, 3H), 3.08–3.48 (q, 2H, J = 13.9 Hz),
5.42–5.54 (q, 2H, J = 15.0 Hz), 5.57 (s, 2H), 6.06 (s,
1H), 7.16 (d, 1H, J = 8.2 Hz), 7.20 (s, 1H), 7.45 (s,
1H), 7.49 (d, 1H, J = 8.2 Hz), 7.82 (t, 1H, J = 7.3 Hz),
7.94 (t, 1H, J = 7.3 Hz), 8.25 (d, 1H, J = 8.6 Hz), 8.99
(d, 1H, J = 8.6 Hz), 9.69 (s, 1H). MS (ESI): 494
(M+H)+. Anal. Calcd for C30H27N3O4: C, 73.01; H,
5.51; N, 8.51. Found C, 73.17; H, 5.50; N, 8.49.
5.1.30. 7-(2-Chlorophenyl)iminomethylhomocamptothecin
(9w). The titled compound was prepared from 8 and 2-
chloroaniline according to the method of compound 9a,
yield: 23.4%. 1H NMR (DMSO-d6): 0.89 (t, 3H,
J = 7.3 Hz), 1.88 (q, 2H, J = 8.2 Hz), 3.08–3.48 (q, 2H,
J = 13.9 Hz), 5.42–5.55 (q, 2H, J = 15.2 Hz), 5.64 (s,
2H), 6.05 (s, 1H), 7.41 (t, 1H, J = 7.5 Hz),7.46 (s, 1H),
7.52 (t, 1H, J = 7.5 Hz), 7.67 (d, 1H, J = 8.1 Hz), 7.75
(d, 1H, J = 8.1 Hz), 7.85 (t, 1H, J = 7.4 Hz), 7.95 (t,
1H, J = 7.4 Hz), 8.28 (d, 1H, J = 8.5 Hz), 9.03 (d, 1H,
J = 8.5 Hz), 9.75 (s, 1H). MS(ESI): 501 (M+H)+. Anal.
Calcd for C28H22ClN3O4: C, 67.27; H, 4.44; N, 8.40.
Found C, 67.10; H, 4.45; N, 8.42.
5.1.26. 7-(2-Bromophenyl)iminomethylhomocamptothecin
(9s). The titled compound was prepared from 8 and 2-
bromoaniline according to the method of compound
9a, yield: 18.6%. 1H NMR (DMSO-d6): 0.89 (t, 3H,
J = 7.6 Hz), 1.88 (q, 2H, J = 8.1 Hz), 3.08–3.48 (q, 2H,
J = 13.9 Hz), 5.42–5.54 (q, 2H, J = 15.2 Hz), 5.70 (s,
2H), 6.05 (s, 1H), 7.33 (t, 1H, J = 7.4 Hz), 7.46 (s,
1H), 7.57 (t, 1H, J = 7.4 Hz), 7.72 (d, 1H, J = 8.6 Hz),
7.84 (d, 1H, J = 8.6 Hz), 7.86 (t, 1H, J = 7.5 Hz), 7.96
(t, 1H, J = 7.5 Hz), 8.28 (d, 1H, J = 8.5 Hz), 9.02 (d,
1H, J = 8.5 Hz), 9.74 (s, 1H). MS (ESI): 545 (M+H)+.
Anal. Calcd for C28H22BrN3O4: C, 61.77; H, 4.07; N,
7.72. Found C, 61.59; H, 4.08; N, 7.74.
5.1.31. 7-(4-tert-Butylphenyl)iminomethylhomocampto-
thecin (9x). The titled compound was prepared from 8
and 4-tert-butylaniline according to the method of com-
1
pound 9a, yield: 18.0%. H NMR (DMSO-d6): 0.89 (t,
3H, J = 7.4 Hz), 1.35 (s, 9H), 1.88 (q, 2H, J = 8.5 Hz),
3.08–3.48 (q, 2H, J = 14.0 Hz), 5.42–5.55 (q, 2H,
J = 14.9 Hz), 5.59 (s, 2H), 6.05 (s, 1H), 7.46 (s, 1H),
7.54 (d, 2H, J = 8.3 Hz), 7.56 (d, 2H, J = 8.3 Hz), 7.83
(t, 1H, J = 7.4 Hz), 7.95 (t, 1H, J = 7.4 Hz), 8.27 (d,
5.1.27. 7-(3-Chloro-4-fluorophenyl)iminomethylhomo-
camptothecin (9t). The titled compound was prepared
from 8 and 3-chloro-4-fluoroaniline according to the

1506
Z. Miao et al. / Bioorg. Med. Chem. 16 (2008) 1493–1510
1H, J = 8.5 Hz), 9.00 (d, 1H, J = 8.5 Hz), 9.71 (s, 1H).
MS (ESI): 522 (M+H)+. Anal. Calcd for C32H31N3O4:
C, 73.68; H, 5.99; N, 8.06. Found C, 73.81; H, 5.98;
N, 8.03.
ether (1 mL) which precipitated the desired product.
The yellow solid was filtered off and washed once with
ethanol (2 mL) and twice with diethyl ether (5 mL) to af-
ford 13a as a yellow solid: yield 19 mg (34.7%). 1H
NMR (DMSO-d6): 0.86 (t, 3 H, J = 7.1 Hz), 1.86 (q,
2H, J = 8.1 Hz), 3.05–3.51 (q, 2H, J = 13.8 Hz), 4.25–
4.41 (m, 4H), 5.36 (s, 2H), 5.38–5.52 (q, 2H,
J = 14.8 Hz), 6.04 (s, 1H), 6.45 (q, 2H, J = 14.9 Hz),
7.39 (s, 1H), 7.57 (s, 1H), 7.65 (s, 1H), 8.13 (t, 2H,
J = 7.1 Hz), 8.62 (t, 1H, J = 7.5 Hz), 9.06 (d, 2H,
J = 5.9 Hz). MS (ESI): 512 (M+). Anal. Calcd for
C29H26ClN3O6: C, 63.56; H, 4.78; N, 7.67. Found C,
63.79; H, 4.77; N, 7.65.
5.1.32. 7-(4-Acetylphenyl)iminomethylhomocamptothecin
(9y). The titled compound was prepared from 8 and 1-
(4-amino-phenyl)-ethanone according to the method of
compound 9a, yield: 15.4%. 1H NMR (DMSO-d6):
0.89 (t, 3H, J = 7.3 Hz), 1.88 (q, 2H, J = 8.2 Hz), 2.65
(s, 9H), 3.08–3.48 (q, 2H, J = 12.4 Hz), 5.42–5.55 (q,
2H, J = 15.2 Hz), 5.61 (s, 2H), 6.06 (s, 1H), 7.46 (s,
1H), 7.65 (d, 2H, J = 8.0 Hz), 7.86 (t, 1H, J = 7.4 Hz),
7.91 (t, 1H, J = 7.4 Hz), 8.13 (d, 2H, J = 8.0 Hz), 8.30
(d, 1H, J = 8.5 Hz), 9.02 (d, 1H, J = 8.5 Hz), 9.74 (s,
1H). MS (ESI): 508 (M+H)+. Anal. Calcd for
C30H25N3O6: C, 68.82; H, 4.81; N, 8.03. Found C,
68.99; H, 4.81; N, 8.00.
5.1.34. 9,10-(Methylenedioxy)-7-(pyridiniumylmethyl)-
homocamptothecin chloride (13b). The titled compound
was prepared according to the method of compound
1
13a, yield: 39.0%. H NMR (DMSO-d6): 0.88 (t, 3H,
J = 7.3 Hz), 1.87 (q, 2H, J = 8.1 Hz), 3.06–3.52 (q, 2H,
J = 13.8 Hz), 4.14–4.26 (m, 4H), 5.41–5.51 (q, 2H,
J = 15.1 Hz), 5.51 (s, 2H), 6.07 (s, 1H), 6.41 (q, 2H,
J = 14.7 Hz), 7.41 (s, 1H), 7.50 (d, 1H, J = 7.5 Hz),
7.78 (d, 1H, J = 7.5 Hz), 8.07 (t, 2H, J = 7.2 Hz), 8.59
(t, 1H, J = 7.5 Hz), 9.03 (d, 2H, J = 6.0 Hz). MS (ESI):
512 (M+). Anal. Calcd for C29H26ClN3O6: C, 63.56;
H, 4.78; N, 7.67. Found C, 63.75; H, 4.78; N, 7.66.
5.1.33. 10,11-(Methylenedioxy)-7-(pyridiniumylmethyl)
homocamptothecin chloride (13a). A 1 L flask was
charged with dry methylene chloride (50 mL) and 1,4-
benzodioxan-6-amine (1.5 g, 10 mmol) and was cooled
to 0 ꢁC followed by slow addition of a 1 M solution of
boron trichloride in methylene chloride solution
(40 mL) while maintaining an internal temperature at
or below 10 ꢁC under N2. Aluminum chloride (1.3 g,
10 mmol) was added quickly in three portions followed
by addition of chloroacetonitrile (0. 7 mL, 11 mmol).
The reaction mixture was stirred for 30 min at 0 ꢁC
and then heated to reflux for 16 h. The reaction mixture
was allowed to cool to room temperature and then was
quenched into a mixture of ice (100 g)/2 N HC1
(50 mL). The aqueous layer was extracted by methylene
chloride (100 mL). The combined organic layers were
dried over sodium sulfate and concentrated under re-
duced pressure, and the residue was purified by chroma-
tography over silica gel (2% MeOH/CH2Cl2) to give
20-amino-2-chloro-40,50-(ethylenedioxy)acetophenone
(11a) as a yellow solid: 0.45 g (35.2%).1H NMR
(DMSO-d6): 4.22–4.24 (m, 2H), 4.34–4.36 (m, 2H),
4.75 (s, 2H), 6.22 (s, 1H), 6.87 (s, 1H). MS (ESI): 228
(M+H)+.
5.1.35. 11-Methyl-7-(pyridiniumylmethyl)-homocampto-
thecin chloride (13c). The titled compound was prepared
according to the method of compound 13a, yield: 40.4%.
1H NMR (DMSO-d6): 0.88 (t, 3H, J = 7.2 Hz), 1.87 (q,
2H, J = 8.1 Hz), 2.56 (s, 3H), 3.07-3.51 (q, 2H,
J = 13.9 Hz), 5.38–5.53 (q, 2H, J = 15.3 Hz), 5.41 (s,
2H), 6.08 (s, 1H), 6.55 (q, 2H, J = 14.6 Hz), 7.45 (s,
1H), 7.59 (d, 1H, J = 7.3 Hz), 8.03 (d, 1H, J = 7.5 Hz),
8.05 (s, 2H), 8.12 (t, 2H, J = 7.4 Hz), 8.62 (t, 1H,
J = 7.4 Hz), 9.09 (d, 2H, J = 6.0 Hz). MS (ESI): 468
(M+). Anal. Calcd for C28H26ClN3O4: C, 66.73; H,
5.20; N, 8.34. Found C, 66.55; H, 5.21; N, 8.36.
5.1.36. 10,11-Dimethyl-7-(pyridiniumylmethyl)-homo-
camptothecin chloride (13d). The titled compound was
prepared according to the method of compound 13a,
yield: 45.6%. 1H NMR (DMSO-d6): 0.87 (t, 3H,
J = 7.4 Hz), 1.87 (q, 2H, J = 8.1 Hz), 2.43 (s, 3H), 2.47
(s, 3 H), 3.06–3.51 (q, 2H, J = 13.7 Hz), 5.38 (s, 2H),
5.40–5.52 (q, 2H, J = 15.1 Hz), 6.07 (s, 1H), 6.52 (q,
2H, J = 14.7 Hz), 7.43 (s, 1H), 7.94 (s, 1H), 8.04 (s,
1H), 8.13 (t, 2H, J = 7.3 Hz), 8.62 (t, 1H, J = 7.4 Hz),
9.10 (d, 2H, J = 6.1 Hz). MS (ESI): 482 (M+). Anal.
Calcd for C29H28ClN3O4: C, 67.24; H, 5.45; N, 8.11.
Found C, 67.43; H, 5.44; N, 8.13.
A solution of 11a (0.25 g, 1.1 mmol) and tricyclic ketone
5 (0.28 g, 1.0 mmol) in toluene (85 mL) was refluxed
using a Dean–Stark trap for 30 mins. p-Toluenesulfonic
acid (0.03 g) was then added and refluxing was contin-
ued for an additional 1 h. The solution was allowed to
cool to room temperature and the solid was filtered off
and washed by acetone (20 mL) and methanol (20 mL)
to give 10,11-(methylenedioxy)-7-(chloromethyl)-homo-
camptothecin (12a) as yellow solid: yield 0.38 g
(80.6%).1H NMR (DMSO-d6): 0.88 (t, 3H, J = 7.0 Hz),
1.90 (q, 2H, J = 8.1 Hz), 3.06–3.46 (q, 2H,
J = 14.0 Hz), 4.45–4.47 (m, 4H), 5.28–5.37 (m, 4H),
5.40–5.54 (q, 2H, J = 15.4 Hz), 6.04 (s, 1H), 7.33 (s,
1H), 7.57 (s, 1H), 7.75 (s, 1H). MS (ESI): 470 (M+H)+.
5.1.37. 9,11-Dimethyl-7-(pyridiniumylmethyl)-homocam-
ptothecin chloride (13e). The titled compound was pre-
pared according to the method of compound 13a,
yield: 43.5%. 1H NMR (DMSO-d6): 0.87 (t, 3H,
J = 7.4 Hz), 1.87 (q, 2H, J = 8.3 Hz), 2.51 (s, 3H), 2.58
(s, 3H), 3.05–3.51 (q, 2H, J = 13.7 Hz), 5.27 (s, 2H),
5.40–5.52 (q, 2H, J = 15.1 Hz), 6.07 (s, 1H), 6.59 (q,
2H, J = 14.8 Hz), 7.43 (s, 1H), 7.96 (s, 1H), 8.14 (t,
2H, J = 7.5 Hz), 8.64 (t, 1H, J = 7.4 Hz), 8.93 (d, 2H,
J = 6.2 Hz). MS (ESI): 482 (M+). Anal. Calcd for
10,11-(Methylenedioxy)-7-(chloromethyl)-homocam-
ptothecin 12a (47 mg, 0.1 mmol) was added to anhy-
drous pyridine (1 mL) and DMSO (2 mL) at 50 ꢁC.
This was stirred for 12 h before the addition of diethyl

Z. Miao et al. / Bioorg. Med. Chem. 16 (2008) 1493–1510
1507
C29H28ClN3O4: C, 67.24; H, 5.45; N, 8.11. Found C,
67.41; H, 5.44; N, 8.12.
J = 7.4 Hz), 8.35 (s, 1H), 8.62 (t, 1H, J = 7.4 Hz), 9.02
(d, 2H, J = 6.1 Hz). MS (ESI): 489 (M+). Anal. Calcd
for C27H23Cl2N3O4: C, 61.84; H, 4.42; N, 8.01. Found
C, 61.66; H, 4.43; N, 8.04.
5.1.38. 10-Fluoro-7-(pyridiniumylmethyl)-homocampto-
thecin chloride (13f). The titled compound was prepared
according to the method of compound 13a, yield: 33.8%.
1H NMR (DMSO-d6): 0.87 (t, 3H, J = 7.3 Hz), 1.87 (q,
2H, J = 8.3 Hz), 3.07–3.51 (q, 2H, J = 13.8 Hz), 5.40–
5.53 (q, 2H, J = 15.2 Hz), 5.42 (s, 2H), 6.09 (s, 1H),
6.55 (q, 2H, J = 14.7 Hz), 7.47 (s, 1H), 7.87 (d, 1H,
J = 7.3 Hz), 8.06 (d, 1H, J = 7.2 Hz), 8.13 (t, 2H,
J = 7.4 Hz), 8.36 (s, 1H), 8.60 (t, 1H, J = 7.4 Hz), 9.08
(d, 2H, J = 6.2 Hz). MS (ESI): 472 (M+). Anal. Calcd
for C27H23FClN3O4: C, 63.84; H, 4.56; N, 8.27. Found
C, 63.99; H, 4.55; N, 8.27.
5.1.43. 10,11-Dichloro-7-(pyridiniumylmethyl)-homocam-
ptothecin chloride (13k). The titled compound was pre-
pared according to the method of compound 13a,
yield: 21.0%. 1H NMR (DMSO-d6): 0.86 (t, 3H,
J = 7.3 Hz), 1.87 (q, 2H, J = 8.3 Hz), 3.08–3.51 (q, 2H,
J = 13.9 Hz), 5.39 (s, 2H), 5.41–5.53 (q, 2H,
J = 15.2 Hz), 6.07 (s, 1H), 6.54 (q, 2H, J = 14.7 Hz),
7.41 (s, 1H), 7.95 (s, 1H), 8.03 (s, 1H), 8.12 (t, 2H,
J = 7.3 Hz), 8.61 (t, 1H, J = 7.4 Hz), 9.08 (d, 2H,
J = 6.0 Hz). MS (ESI): 523 (M+). Anal. Calcd for
C27H22Cl2N3O4: C, 58.03; H, 3.97; N, 7.52. Found C,
57.91; H, 3.98; N, 7.55.
5.1.39. 10-Methyl-11-chloro-7-(pyridiniumylmethyl)-
homocamptothecin chloride (13g). The titled compound
was prepared according to the method of compound
5.1.44. 11-Fluoro-10-chloro-7-(pyridiniumylmethyl)-
homocamptothecin chloride (13l). The titled compound
was prepared according to the method of compound
13a, yield: 23.7%.1H NMR (DMSO-d6): 0.87 (t, 3H,
J = 7.4 Hz), 1.87 (q, 2H, J = 8.2 Hz), 3.07–3.52 (q, 2H,
J = 13. 8 Hz), 5.41 (s, 2H), 5.42–5.51 (q, 2H,
J = 15.1 Hz), 6.09 (s, 1H), 6.56 (q, 2H, J = 14.7 Hz),
7.42 (s, 1H), 7.94 (s, 1H), 8.05 (s, 1H), 8.13 (t, 2H,
J = 7.1 Hz), 8.62 (t, 1H, J = 7.4 Hz), 9.07 (d, 2H,
J = 6.1 Hz). MS (ESI): 507 (M+). Anal Calcd for
C27H22FClN3O4: C, 59.79; H, 4.09; N, 7.75. Found C,
59.98; H, 4.08; N, 7.73.
1
13a, yield: 31.8%. H NMR (DMSO-d6): 0.87 (t, 3H,
J = 7.3 Hz), 1.87 (q, 2H, J = 8.0 Hz), 2.53 (s, 3H),
3.06–3.51 (q, 2H, J = 13.8 Hz), 5.38 (s, 2H), 5.40–5.53
(q, 2H, J = 15.0 Hz), 6.09 (s, 1H), 6.54 (q, 2H,
J = 14.6 Hz), 7.45 (s, 1H), 8.12 (t, 2H, J = 7.2 Hz),
8.14 (s, 1H), 8.35 (s, 1H), 8.63 (t, 1H, J = 7.4 Hz), 9.07
(d, 2H, J = 6.3 Hz). MS (ESI): 503 (M+). Anal. Calcd
for C28H25Cl2N3O4: C, 62.46; H, 4.68; N, 7.80. Found
C, 62.23; H, 4.69; N, 7.82.
5.1.40. 11-Chloro-7-(pyridiniumylmethyl)-homocampto-
thecin chloride (13h). The titled compound was prepared
according to the method of compound 13a, yield: 22.5%.
1H NMR (DMSO-d6): 0.87 (t, 3H, J = 7.3 Hz), 1.86 (q,
2H, J = 8.1 Hz), 3.06–3.52 (q, 2H, J = 14.2 Hz), 5.40 (s,
2H), 5.41–5.54 (q, 2H, J = 16.1 Hz), 6.10 (s, 1H), 6.56
(q, 2H, J = 14.3 Hz), 7.47 (s, 1H), 7.81 (d, 1H,
J = 7.5 Hz), 8.13 (t, 2H, J = 7.3 Hz), 8.20 (d, 1H,
J = 7.5 Hz), 8.35 (s, 1H), 8.63 (t, 1H, J = 7.3 Hz), 9.06
(d, 2H, J = 6.0 Hz). MS (ESI): 489 (M+). Anal. Calcd
for C27H23Cl2N3O4: C, 61.84; H, 4.42; N, 8.01. Found
C, 62.01; H, 4.41; N, 8.00.
5.1.45. 10-Bromo-7-(pyridiniumylmethyl)-homocampto-
thecin chloride (13m). The titled compound was pre-
pared according to the method of compound 13a,
yield: 33.7%. 1H NMR (DMSO-d6): 0.87 (t, 3H,
J = 7.2 Hz), 1.87 (q, 2H, J = 8.1 Hz), 3.06–3.51 (q, 2H,
J = 13.9 Hz), 5.37–5.53 (q, 2H, J = 15.4 Hz), 5.41 (s,
2H), 6.07 (s, 1H), 6.56 (q, 2H, J = 14.9 Hz), 7.48 (s,
1H), 8.06 (d, 1H, J = 7.5 Hz), 8.15 (t, 2H, J = 7.4 Hz),
8.20 (d, 1H, J = 7.5 Hz), 8.48 (s, 1H), 8.64 (t, 1H,
J = 7.3 Hz), 9.06 (d, 2H, J = 6.1 Hz). MS (ESI): 533
(M+). Anal. Calcd for C27H22BrClN3O4: C, 57.01; H,
4.08; N, 7.39. Found C, 57.22; H, 4.08; N, 7.37.
5.1.41. 11-Fluoro-7-(pyridiniumylmethyl)-homocampto-
thecin chloride (13i). The titled compound was prepared
according to the method of compound 13a, yield: 25.3%.
1H NMR (DMSO-d6): 0.86 (t, 3H, J = 7.3 Hz), 1.86 (q,
2H, J = 8.3 Hz), 3.06–3.50 (q, 2H, J = 14.1 Hz), 5.41 (s,
2H), 5.42–5.52 (q, 2H, J = 16.0 Hz), 6.08 (s, 1H), 6.51
(q, 2H, J = 14.2 Hz), 7.43 (s, 1H), 7.80 (d, 1H,
J = 7.3 Hz), 8.12 (t, 2H, J = 7.5 Hz), 8.21 (d, 1H,
J = 7.3 Hz), 8.32 (s, 1H), 8.60 (t, 1H, J = 7.3 Hz), 9.01
(d, 2H, J = 6.0 Hz). MS (ESI): 472 (M+). Anal. Calcd
for C27H23FClN3O4: C, 63.84; H, 4.56; N, 8.27. Found
C, 63.97; H, 4.56; N, 8.25.
5.1.46. 10-Methyl-7-(pyridiniumylmethyl)-homocampto-
thecin chloride (13n). The titled compound was prepared
according to the method of compound 13a, yield: 25.3%.
1H NMR (DMSO-d6): 0.87 (t, 3H, J = 7.4 Hz), 1.87 (q,
2H, J = 8.2 Hz), 2.50 (s, 3H), 3.06–3.51 (q, 2H,
J = 13.9 Hz), 5.40 (s, 2H), 5.43–5.53 (q, 2H,
J = 15.1 Hz), 6.06 (s, 1H), 6.54 (q, 2H, J = 14.5 Hz),
7.45 (s, 1H), 7.76 (d, 1H, J = 7.3 Hz), 7.97 (s, 1H),
8.13 (t, 2H, J = 7.4 Hz), 8.17 (d, 1H, J = 7.3 Hz), 8.63
(t, 1H, J = 7.5 Hz), 9.11 (d, 2H, J = 6.1 Hz). 13C NMR
(DMSO-d6): 8.67 (1C), 21.96 (1C), 36.78 (1C), 42.39
(1C), 50.85 (1C), 57.42 (1C), 61.61 (1C), 73.61 (1C),
100.06 (1C), 122.68 (1C), 123.12 (1C), 125.08 (1C),
128.82 (2C), 130.37 (1C), 131.60 (1 C), 133.22 (1C),
139.56 (1C), 144.87 (1C), 145.42 (2C), 146.81 (1 C),
147.76 (1C), 149.91 (1C), 150.85 (1C), 156.33 (1C),
159.67 (1 C), 172.20 (1C). MS (ESI): 468 (M+). Anal.
Calcd for C28H26ClN3O4: C, 66.73; H, 5.20; N, 8.34.
Found C, 66.95; H, 5.20; N, 8.31.
5.1.42. 10-Chloro-7-(pyridiniumylmethyl)-homocampto-
thecin chloride (13j). The titled compound was prepared
according to the method of compound 13a, yield: 23.7%.
1H NMR (DMSO-d6): 0.87 (t, 3H, J = 7.4 Hz), 1.86 (q,
2H, J = 8.3 Hz), 3.08–3.51 (q, 2H, J = 13.9 Hz), 5.41 (s,
2H), 5.41–5.52 (q, 2H, J = 15.0 Hz), 6.07 (s, 1H), 6.54
(q, 2H, J = 14.8 Hz), 7.47 (s, 1H), 7.88 (d, 1H,
J = 7.5 Hz), 8.07 (d, 1H, J = 7.5 Hz), 8.11 (t, 2H,

1508
Z. Miao et al. / Bioorg. Med. Chem. 16 (2008) 1493–1510
5.1.47. 10-Methyl-7-[(20-methyl)pyridiniumylmethyl]-
homocamptothecin chloride (14a). The titled compound
was prepared from 2-methylpyridine according to the
method of compound 13a, yield: 33.1%. 1H NMR
(DMSO-d6): 0.86 (t, 3H, J = 7.5 Hz), 1.86 (q, 2H,
J = 8.2 Hz), 2.53 (s, 3H), 3.04–3.49 (q, 2H,
J = 14.0 Hz), 3.05 (s, 3H), 5.35 (s, 2H), 5.37–5.48 (q,
2H, J = 15.0 Hz), 6.05 (s, 1H), 6.37 (q, 2H,
J = 14.7 Hz), 7.45 (s, 1H), 7.78 (d, 1H, J = 7.1 Hz),
7.79 (t, 1H, J = 7.4 Hz), 7.95 (s, 1H), 8.19 (d, 1H,
J = 7.1 Hz), 8.22 (d, 1H, J = 7.5 Hz), 8.38 (d, 1H,
J = 7.3 Hz), 8.55 (s, 1H). MS (ESI): 482 (M+). Anal.
Calcd for C29H28ClN3O4: C, 67.24; H, 5.45; N, 8.11.
Found C, 67.05; H, 5.46; N, 8.12.
tration, and 0.1% DMSO for control. After three days
of incubation, 20 lL MTT (3-[4,5-dimethylthiazol-2-
yl]-2,5-diphenyl-tetrazolium bromide) solution (5 mg
mLꢀ1) was added to each well, and after shaking for
1 min the plate was incubated further for 4 h. Formazan
crystals were dissolved with 100 lL DMSO. The absor-
bance (OD) was quantitated with microplate spectro-
photometer at 570 lM. Wells containing no drugs
were used as blanks for the spectrophotometer. The
survival of the cells was expressed as percentage of
untreated control wells.
5.3. In vivo antitumor activity
The in vivo antitumor activity of compounds 9j and 13g
was evaluated and TPT was used as reference drug.
C57BL/6 strain male mice (Certificate SCXK-2002-
0010, weighing 18–20 g) were obtained from Shanghai
Experimental Animal Center, Chinese Academy of Sci-
ences. C26 colon cancer cell suspensions were implanted
subcutaneously into the right axilla region of mice. Dai-
ly treatment with drugs or normal saline commenced 3
days after implantation of cells. Mice were administered
by sc injection once daily for 5 consecutive days. All the
mice were euthanized and the tumors were weighed on
day 14 after implantation of cells. The rate of inhibition
of tumor growth in vivo was calculated using the follow-
ing formula (ATWCG = Average Tumor Weight of
Control Group, ATWTG = Average Tumor Weight of
Test Group):
5.1.48. 10-Methyl-7-[(30-methyl)pyridiniumylmethyl]-
homocamptothecin chloride (14b). The titled compound
was prepared from 3-methylpyridine according to the
method of compound 13a, yield: 31.1%. 1H NMR
(DMSO-d6): 0.88 (t, 3H, J = 7.5 Hz), 1.88 (q, 2H,
J = 8.1 Hz), 2.44 (s, 3H), 3.06–3.51 (q, 2H,
J = 13.7 Hz), 3.32 (s, 3H), 5.36 (s, 2H), 5.39–5.53 (q,
2H, J = 15.1 Hz), 6.06 (s, 1H), 6.47 (q, 2H,
J = 14.6 Hz), 7.45 (s, 1H), 7.76 (d, 1H, J = 7.3 Hz),
7.94 (s, 1H), 8.06 (t, 1H, J = 7.1 Hz), 8.16 (d, 1H,
J = 7.3 Hz), 8.47 (d, 1H, J = 7.4 Hz), 8.89 (s, 1H), 9.02
(d, 1H, J = 7.3 Hz). MS (ESI): 482 (M+). Anal. Calcd
for C29H28ClN3O4: C, 67.24; H, 5.45; N, 8.11. Found
C, 67.01; H, 5.45; N, 8.14.
5.1.49. 10-Methyl-7-[(40-methyl)pyridiniumylmethyl]-
homocamptothecin chloride (14c). The titled compound
was prepared from 4-methylpyridine according to the
method of compound 13a, yield: 26.9%. 1H NMR
(DMSO-d6): 0.87 (t, 3H, J = 7.6 Hz), 1.88 (q, 2H,
J = 8.0 Hz), 2.60 (s, 3H), 3.06–3.51 (q, 2H,
J = 14.2 Hz), 3.32 (s, 3H), 5.40–5.54 (q, 2H,
J = 15.1 Hz), 5.41 (s, 2H), 6.06 (s, 1H), 6.44 (q, 2H,
J = 14.7 Hz), 7.45 (s, 1H), 7.76 (d, 1H, J = 7.2 Hz),
7.91 (s, 1H), 7.94 (d, 2H, J = 7.1 Hz), 8.16 (d, 1H,
J = 7.2 Hz), 8.95 (d, 2H, J = 7.4 Hz). MS (ESI): 482
(M+). Anal. Calcd for C29H28ClN3O4: C, 67.24; H,
5.45; N, 8.11. Found C, 67.45; H, 5.44; N, 8.09.
ATWCG ꢀ ATWTG
Growth inhibition ¼
ꢂ 100%
ATWTG
5.4. Computational details
5.4.1. General. The crystallographic coordinates of the
˚
ternary complex of CPT–DNA–Topo I (3.0 A resolution,
Rcryst = 0.244) were obtained from the Brookheaven Pro-
tein Databank as entries 1T8I. All calculations were per-
formed with commercially available SYBYL 6.945
software package and InsightII 200046 software package.
All calculations were performed on an Origin 300 Server.
5.1.50. 10-Methyl-7-[(30-carboxy)pyridiniumylmethyl]-
homocamptothecin chloride (14d). The titled compound
was prepared from nicotinic acid according to the meth-
od of compound 13a, yield: 30.6%.1H NMR (DMSO-
d6): 0.87 (t, 3H, J = 7.5 Hz), 1.87 (q, 2H, J = 8.3 Hz),
2.53 (s, 3H), 3.05–3.50 (q, 2H, J = 14.0 Hz), 5.37 (s,
2H), 5.39–5.52 (q, 2H, J = 15.1 Hz), 6.05 (s, 1H), 6.59
(q, 2H, J = 14.6 Hz), 7.44 (s, 1H), 7.78 (d, 1H,
J = 7.1 Hz), 8.04 (s, 1H), 8.05 (t, 1H, J = 7.4 Hz), 8.17
(d, 1H, J = 7.1 Hz), 8.86 (d, 1H, J = 7.4 Hz), 8.92 (d,
1H, J = 7.3 Hz), 9.75 (s, 1H). MS (ESI): 512 (M+). Anal.
Calcd for C29H26ClN3O6: C, 63.56; H, 4.78; N, 7.67.
Found C, 63.77; H, 4.77; N, 7.65.
5.4.2. Data set. A total of 45 hCPT derivatives were used
as a data set in the following 3D-QSAR analysis. The
selection of training set and test set is based on struc-
tural diversity and frequency of distribution of biologi-
cal data by doing principal component analysis. As a
result, five compounds were selected as a test set, which
represented a range of inhibitory activity similar to that
of a training set and was used to evaluate the predictive
power of the resulting models. In order to reinforce the
reliability of their 3D-QSAR models, six previously syn-
thesized 7-ester hCPT derivatives (compound 15a–15f)47
were selected as real external compounds (Fig. 2) and
were added into the test set. The biological activity of
each compound was expressed as inhibitory concentra-
tions (IC50) against A549 cell line and ꢀlg(IC50) was
used for the 3D-QSAR analysis.
5.2. MTT assay
One thousand two hundred cells per well were plated in
96-well plates. After culturing for 24 h, test compounds
were added onto triplicate wells with different concen-
In the 3D-QSAR studies, pharmacophoric conforma-
tion and aliguMent rule are two major critical steps to

Z. Miao et al. / Bioorg. Med. Chem. 16 (2008) 1493–1510
1509
get meaningful results. Because the crystal structure of
hCPT–DNA–Topo I ternary complex has not been re-
ported, molecular docking was used to simulate the
pharmacophoric conformation. As a result, the docked
conformation of compound 9l was used as the template
to construct the structures of the remaining compounds
in the data set. Energy minimization was performed
using the Tripos force field, Powell optimization meth-
od, and MAXIMIN2 minimizer with a convergence cri-
to be a measure of how good the model represents the data
in the training set. The cross-validated coefficient, q2, was
calculated using the following equation:
P
2
ðY predicted ꢀ Y observed
Þ
q2 ¼ 1 ꢀ
ð2Þ
P
2
ðY observed ꢀ Y mean
Þ
where Ypred, Yactual, and Ymean are predicted, actual, and
mean values of the target property (pMIC80), respec-
tively. To maintain the optimum number of PLS com-
ponents and minimize the tendency to overfit the data,
the number of components corresponding to the lowest
PRESS value was used to derive the final PLS regression
models. Conventional correlation coefficient r2 and its
standard error(s) were also computed for the final PLS
models. To graphically interpret the 3D-QSAR results
in terms of field contributions, isocontour maps were
˚
terion of 0.001 kcal/mol A. Charges were calculated by
the Gasteiger–Huckel method. Simulated annealing
¨
was then performed. The system was heated to 1000 K
for 1.0 ps and then annealed to 250 K for 1.5 ps. The
annealing function was exponential; 50 such cycles of
annealing were run and the resulting 50 conformers were
optimized using methods described above. The lowest
energy conformation was selected. Compounds were
aligned by RMS fitting of the atoms to the most active
compound, compound 9l, based on the heavy atoms in
the five-membered ring (Fig. 3). The superimposing of
all the compounds is shown in Figure 4.
generated using the field type ‘stdev coeff’ and the con-
*
tour levels were set to default values.
5.4.4. Docking analysis. Compound 9l was manually
docked into the active site of Topo I on the basis
of the binding mode of CPT. Then, the flexible ligand
docking procedure in the Affinity module within
InsightII was used to define the lowest energy position
for the substrate using a Monte Carlo docking proto-
5.4.3. CoMFA and CoMSIA Setup. In deriving the
CoMFA and CoMSIA descriptor fields, a 3D cubic lat-
˚
˚
tice with grid spacing of 2 A and extending 4 A units be-
yond the aligned molecules in all directions was created
to encompass the aligned molecules. The charges were
calculated using the MOPAC AM1 Hamiltonian semi-
empirical method.48 The CoMFA descriptors, steric
(Lennard–Jones 6–12 potential) and electrostatic (Cou-
lombic potential) field energies, were calculated by the
following parameters: an sp3 carbon probe atom
˚
col. All the atoms within a defined radius (6 A) of the
inhibitor were allowed to move. The solvation grid
supplied with the affinity Program was used. If the
resulting inhibitor/enzyme system was within a prede-
fined energy tolerance of the previous structure, the
system was subjected to minimization. The resulting
structure was accepted on the basis of energy check,
which used the Metropolis criterion, and also a check
of rms distance of the new structure versus the struc-
ture found so far. The final conformation was ob-
˚
with +1 charge and a van der Waals radius of 1.52 A,
and energy cutoff of 30 kcal/mol. CoMSIA similarity
indices descriptors (steric, electrostatic, hydrophobic,
H-bond donor, and H-bond acceptor fields) were calcu-
tained through
a simulation annealing procedure
˚
lated using a C1+ probe atom with a radius of 1.0 A
from 500 to 300 K, and then 5000 rounds of energy
minimization were performed to reach a convergence,
where the resulting interaction energy values were
used to define a rank order.
˚
placed at various grid spacing from 1 to 3 A. CoMSIA
similarity indices (AF) for a molecule j with atoms i at
a grid point q are calculated by the following equation:
X
2
iq
AqF ;K ðjÞ ¼ ꢀ
xprobe kxikeꢀar
ð1Þ
0
Acknowledgment
where k represents the following physicochemical prop-
erties: steric, electrostatic, hydrophobic, H-bond donor,
and H-bond acceptor. A Gaussian type distance depen-
dence was used between the grid point q and each atom i
of the molecule. Here, steric indices are related to the
third power of the atomic radii, electrostatic descriptors
are derived from atomic partial charges, hydrophobic
fields are derived from atom-based parameters, and H-
bond donor and acceptor indices are obtained by a
rule-based method based on experimental results. The
attenuation factor (a) was initially set to the default va-
lue of 0.3.
This research was supported by the National Natural
Science Foundation of China (No. 30371689).
References and notes
1. Wall, M. E.; Wani, M. C.; Cook, C. E.; Palmer, K. H.;
McPhail, A. T.; Sim, G. A. J. Am. Chem. Soc. 1966, 88,
3888–3890.
2. Moertel, C. G.; Schutt, A. J.; Reitemeier, R. J.; Hahn, R.
G. Cancer Chemother. Rep. 1972, 56, 649–652.
3. Li, Q. Y.; Zu, Y. G.; Shi, R. Z.; Yao, L. P. Curr. Med.
Chem. 2006, 13, 2021–2039.
4. Driver, R. W.; Yang, L. X. Mini-Rev. Med. Chem. 2005, 5,
425–439.
5. Thomas, C. J.; Rahier, N. J.; Hecht, S. M. Bioorg. Med.
Chem. 2004, 12, 1585–1604.
The CoMFA and CoMSIA descriptors were used as inde-
pendent variables, and pIC50 values were used as depen-
dent variables in partial least squares (PLS) regression
analyses to derive 3D-QSAR models using the standard
implementation in the SYBYL package. To obtain the
optimal number of principle components, the leave-one-
out (LOO) cross-validation was utilized, which is used
6. Kingsbury, W. D.; Boehm, J. C.; Jakas, D. R. J. Med.
Chem. 1991, 34, 98–107.

1510
Z. Miao et al. / Bioorg. Med. Chem. 16 (2008) 1493–1510
7. Graul, A. I.; Prous, J. R. Drug News Perspect. 2005, 1, 21–
27.
8. Giovanella, B. C.; Stehlin, J. S.; Hinz, H. R.; Kozielski, A.
J.; Harris, N. J.; Vardeman, D. M. Int. J. Oncol. 2002, 20,
81–88.
28. Du, W.; Curran, D. P.; Bevins, R. L.; Zimmer, S. G.;
Zhang, J.; Burke, T. G. Bioorg. Med. Chem. 2002, 10,
103–110.
29. Wani, M. C.; Ronman, P. E.; Lindley, J. T.; Wall, M.
E. J. Med. Chem. 1980, 23, 554.
9. Eckhardt, S. G.; Baker, S. D.; Eckardt, J. R. Clin. Cancer
Res. 1998, 4, 595–604.
10. Minami, H.; Fujii, H.; Igarashi, T.; Itoh, K.; Tamanoi, K.;
Oguma, T.; Sasaki, Y. Clin. Cancer Res. 2001, 7, 3056–3064.
11. Hsiang, Y. H.; Hertzberg, R.; Hecht, S.; Liu, L. F. J. Biol.
Chem. 1985, 260, 14873–14878.
12. Darpa, P.; Beardmore, C.; Liu, L. F. Cancer Res. 1990, 50,
6919–6924.
13. Taso, Y. P.; Darpa, P.; Liu, L. F. Cancer Res. 1992, 52,
1823–1829.
30. Yang, S.; Zhang, W. N.; Zhou, Y. J.; Yao, J. Z.; Yu, J. X.;
Sheng, C. Q. Chin. J. Med. Chem. 2004, 14, 242–244.
31. Miao, Z. Y.; Zhang, W. N.; Yao, J. Z.; Sheng, C. Q.;
Song, Y. L.; Zhang, M. Chin. J. Med. Chem. 2005, 15,
244–246.
32. De Cesare, M.; Pratesi, G.; Perego, P.; Carenini, N.;
Tinelli, S.; Merlini, L.; Penco, S.; Pisano, C.; Bucci, F.;
Vesci, L.; Pace, S.; Capocasa, F.; Carminati, P.; Zunino,
F. Cancer Res. 2001, 61, 7189–7195.
33. Staker, B. L.; Hjerrrild, K.; Feese, M. D.; Behnke, C. A.;
Burgin, A. B.; Stewart, L. Proc. Natl. Acad. Sci. U.S.A.
2002, 99, 15387–15392.
14. Liu, L. F.; Desai, S. D.; Li, T. K.; Mao, Y.; Sun, M.; Sim,
S. P. Ann. N.Y. Acad. Sci. 2000, 922, 1–10.
15. Burke, T. G. Ann. N.Y. Acad. Sci. 1996, 803, 29–31.
16. Cao, Z.; Harris, N.; Kozielski, A.; Vardeman, D.; Stehlin,
J. S.; Giovanella, B. J. Med. Chem. 1998, 41, 31–37.
17. Lavergne, O.; Lesueur-Ginot, L.; Rodas, F. P. Bioorg.
Med. Chem. Lett. 1997, 7, 2235–2238.
18. Bailly, C. Crit. Rev. Oncol. Hematol. 2003, 45, 91–108.
19. Bailly, C.; Lansiaux, A.; Dassonneville, L.; Demarquay,
D.; Coulomb, H.; Huchet, M.; Lavergne, O.; Bigg, D. C.
H. Biochemistry 1999, 38, 15556–15563.
20. Lesueur-Ginot, L.; Demarquay, D.; Kiss, R.; Kasprzyk, P.
G.; Dassonneville, L.; Bailly, C.; Camara, J.; Lavergne,
O.; Bigg, D. C. Cancer Res. 1999, 59, 2939–2943.
21. Lavergne, O.; Lesueur-Ginot, L.; Pla, R. F.; Kasprzyk, P.
G.; Pommier, J.; Demarquay, D.; Prevost, G.; Ulibarri,
G.; Rolland, A.; Schiano-Liberatore, A. M.; Harnett, J.;
Pons, D.; Camara, J.; Biqq, D. C. J. Med. Chem. 1998, 41,
5410–5419.
22. Lavergne, O.; Harnett, J.; Rolland, A.; Lanco, C.;
Lesueur-Ginot, L.; Demarquay, D.; Huchet, M.; Cou-
lomb, H.; Bigg, D. C. Bioorg. Med. Chem. Lett. 1999, 9,
2599–2602.
34. Staker, B. L.; Feese, M. D.; Cushman, M.; Pommier, Y.;
Zembower, D.; Stewart, L.; Burgin, A. B. J. Med. Chem.
2005, 48, 2336–2345.
35. Dallavalle, S.; Merlini, L.; Morini, G.; Musso, L.; Penco,
S.; Beretta, G. L.; Tinelli, S.; Zunino, F. Eur. J. Med.
Chem. 2004, 39, 507–513.
36. Lackey, K.; Sternbach, D. D.; Croom, D. K.; Emerson, D.
L.; Evans, M. G.; Leitner, P. L.; Luzzio, M. J.; Mcintyre,
G.; Vuong, A.; Yates, J.; Besterman, J. M. J. Med. Chem.
1996, 39, 713–719.
37. Zu, Y. G.; Li, Q. Y.; Fu, Y. J.; Wang, W. Bioorg. Med.
Chem. Lett. 2004, 14, 4023–4026.
38. Bringmann, G.; Rummey, C. J. Chem. Inf. Comput. Sci.
2003, 43, 304–316.
39. Cramer, R. D.; Patterson, D. E.; Bunce, J. D. J. Am.
Chem. Soc. 1988, 110, 5959–5967.
40. Klebe, G.; Abraham, U.; Mietzner, T. J. Med. Chem.
1994, 37, 4130–4146.
41. Bohm, M.; Strzebecher, J.; Klebe, G. J. Med. Chem. 1999,
42, 458–477.
42. Bush, B. L.; Nachbar, R. B. Sample-distance partial least
squares: PLS optimized for many variables, with applica-
tion to CoMFA. J. Comput. Aided Mol. Des. 1993, 7, 587–
619.
43. Laco, G. S.; Du, W.; Kohlhagen, G.; Sayer, J. M.; Jerina,
D. M.; Burke, T. G.; Curran, D. P.; Pommier, Y. Bioorg.
Med. Chem. 2004, 12, 5225–5235.
44. Song, Y. L.; Zhang, W. N.; Ji, H. T.; Sheng, C. Q.; Zhang,
M.; Yao, J. Z.; Yu, J. X.; Zhou, Y. J.; Zhu, J.; Lu, J. G.
Acta Chim. Sinica 2003, 61, 1860–1866.
45. SYBYL 6.9: Tripos Associates, Inc. 1699 S. Hanley Road,
Suite 303, St. Louis, MO 63144. 2003.
23. Lavergne, O.; Demarquay, D.; Bailly, C.; Lanco, C.;
Rolland, A.; Huchet, M.; Coulomb, H.; Muller, N.;
Baroqqi, N.; Camara, J.; Le Breton, C.; Manqinot, E.;
Cazaux, J. B.; Biqq, D. C. J. Med. Chem. 2000, 43, 2285–
2289.
24. Bom, D.; Curran, D. P.; Chavan, A. J.; Kruszewski, S.;
Zimmer, S. G.; Fraley, K. A.; Burke, T. G. J. Med. Chem.
1999, 42, 3018–3022.
25. Tangirala, R. S.; Antony, S.; Agama, K.; Pommier, Y.;
Anderson, B. D.; Bevins, R.; Curran, D. P. Bioorg. Med.
Chem. 2006, 14, 6202–6212.
26. Gelderblom, H.; Salazar, R.; Verweij, J.; Pentheroudakis,
G.; De Jonge, M. J.; Devlin, M.; van Hooije, C.; Seguy,
F.; Obach, R.; Prunonosa, J.; Principe, P.; Twelves, C.
Clin. Cancer Res. 2003, 9, 4101–4107.
27. Demarquay, D.; Huchet, M.; Coulomb, H.; Lesueur-
Ginot, L.; Lavergne, O.; Camara, J.; Kasprzyk, P. G.;
Prevost, G.; Bigg, D. C. Cancer Res. 2004, 64, 4942–4949.
46. InsightII 2000: Molecular Simulation Inc. 9685 Scranton
Road, San Diego, CA 92121-3752. 1999.
47. Miao, Z. Y.; Zhang, W. N.; Yao, J. Z.; Sheng, C. Q.; Xu,
H.; Zhang, M.; Zhang, J.; You, L.; Che, X. Y. Chin. J.
Org. Chem. 2006, 26, 1221–1224.
48. Dewar, M. J. S.; Zoebisch, E. G.; Healy, E. F.; Stewart, J.
J. P. J. Am. Chem. Soc. 1985, 107, 3902–3909.