Podophyllotoxin directly binds a hinge domain in E2 of HPV and inhibits an E2/E7 interaction in vitro

Article abstract of DOI:10.1016/j.bmc.2008.03.053

Podophyllotoxin (PT), a strong cytotoxic agent from berberidaceae, has been known to inhibit tubulin polymerization. Although PT has been used for developing anticancer drugs as one of seed compounds, clinical treatment by itself has been unsuccessful because of the side effects, except one example in the treatments of warts. In this study, we screened peptides binding to PT with T7 phage display clonings in order to obtain more information about molecular mechanism of the action. A selected phage clone has a specific amino acid sequence to be SVPSRRRPDGRTHRSSRG. A homology search by protein database BLAST showed that this sequence had a similarity to a hinge domain (HD) of E2 protein in human papillomavirus (HPV) type 1a which is known to cause plantar warts. Surface plasmon resonance (SPR) analysis showed that PT bound to a recombinant HPV 1a E2 protein giving a K_D = 24.1 μM which has compared with those of other domains of E2 protein. Also we demonstrated whether PT inhibited HD interaction or not. E7 protein of HPV has been known to be an oncoprotein and was reported to interact with HD of E2 protein. We demonstrated that an E2/E7 interaction was inhibited by the addition of PT in this report. And we showed the bindings of PT to other types of HPV. Our results suggest that PT is potential as a tool for clarifying the molecular mechanism of HPV.

Full text of DOI:10.1016/j.bmc.2008.03.053

Available online at www.sciencedirect.com
Bioorganic & Medicinal Chemistry 16 (2008) 5815–5825
Podophyllotoxin directly binds a hinge domain in E2 of HPV
and inhibits an E2/E7 interaction in vitro
Takeki Saitoh,a,b Kouji Kuramochi,b Takahiko Imai,a,b Kei-ichi Takata,b,
Masahide Takehara,a,b,à Susumu Kobayashi,a,c Kengo Sakaguchia,b
and Fumio Sugawaraa,b,
*
aGenome and Drug Research Center, Tokyo University of Science, 2641 Yamazaki, Noda, Chiba 278-8501, Japan
bDepartment of Applied Biological Science, Tokyo University of Science, 2641 Yamazaki, Noda, Chiba 278-8501, Japan
cFaculty of Pharmaceutical Sciences, Tokyo University of Science, 2641 Yamazaki, Noda, Chiba 278-8501, Japan
Received 1 February 2008; revised 17 March 2008; accepted 21 March 2008
Available online 25 March 2008
Abstract—Podophyllotoxin (PT), a strong cytotoxic agent from berberidaceae, has been known to inhibit tubulin polymerization.
Although PT has been used for developing anticancer drugs as one of seed compounds, clinical treatment by itself has been unsuc-
cessful because of the side effects, except one example in the treatments of warts. In this study, we screened peptides binding to PT
with T7 phage display clonings in order to obtain more information about molecular mechanism of the action. A selected phage
clone has a specific amino acid sequence to be SVPSRRRPDGRTHRSSRG. A homology search by protein database BLAST
showed that this sequence had a similarity to a hinge domain (HD) of E2 protein in human papillomavirus (HPV) type 1a which
is known to cause plantar warts. Surface plasmon resonance (SPR) analysis showed that PT bound to a recombinant HPV 1a E2
protein giving a KD = 24.1 lM which has compared with those of other domains of E2 protein. Also we demonstrated whether PT
inhibited HD interaction or not. E7 protein of HPV has been known to be an oncoprotein and was reported to interact with HD of
E2 protein. We demonstrated that an E2/E7 interaction was inhibited by the addition of PT in this report. And we showed the bind-
ings of PT to other types of HPV. Our results suggest that PT is potential as a tool for clarifying the molecular mechanism of HPV.
ꢀ 2008 Elsevier Ltd. All rights reserved.
1. Introduction
tubulin polymerization.4,5 Podophyllin, which contains
PT as a constituent, is a resin from the roots and rhi-
zomes of P. peltatum, and has been widely used as a
treatment for genital warts (known as condyloma acu-
minate) caused by human papillomavirus (HPV).6 The
mechanism responsible for the clinical effects could be
Podophyllotoxin (PT), which is a lignan extracted from
roots of Podophyllum peltatum, has a strong cytotoxicity
in several cells (Fig. 1a).1,2 Etoposide (EP), one of semi-
synthetic derivatives of PT, has been widely used as an
anti-cancer drug and has given good clinical results
against several types of neoplasm (Fig. 1b).3 Although
a cytotoxic effect of EP can be based on the inhibition
of DNA topoisomerase II, PT acts as an inhibitor of
a
b
OH
9
O
HO
O
O
O
O
O
OH
O
O
O
Keywords: HPV; Podophyllotoxin; E2; Phage display; SPR.
*
O
O
Corresponding author. Tel.: +81 4 7124 1501x3400; fax: +81 4 7123
9757; e-mail: sugawara@rs.noda.tus.ac.jp
O
Present address: Department of Pharmacology, Hillman Cancer
Center, University of Pittsburgh Medical School, Pittsburgh, PA
15213-1863, USA.
H3CO
OCH3
OCH3
H3CO
OCH3
OH
à
Present address: Department of Biological Information, Graduate
School of Bioscience and Biotechnology, Tokyo Institute of Tech-
nology, 4259 Nagatsuta-cho, Midori-ku, Yokohama-shi, Kanagawa
226-8501, Japan.
Figure 1. Structures of podophyllotoxin and its derivative. (a) Podo-
phyllotoxin. (b) Podophyllotoxin derivative to inhibit topoisomerase
II, etoposide.
0968-0896/$ - see front matter ꢀ 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2008.03.053

5816
T. Saitoh et al. / Bioorg. Med. Chem. 16 (2008) 5815–5825
owed that PT induced necrosis in host cells and resulted
to disrupt viral replications.7 However, direct influence
of PT on HPV itself is still ambiguous.
ciation.15 Lai et al. showed that E2 protein of HPV 5
interacted to SR protein with RS rich region in HD and
assisted pre-mRNA splicing.16 The function of HD was
thought to work as a linker between two functional do-
mains, but the significance of HD function has been re-
cently recognized.
Human papillomavirus is a family over 100 types of
viruses, some of which cause diseases such as hand warts,
plantar warts, genital warts, and several cancers.8,31
HPVs are maintained as episomal circular double-strand
DNA in infected cell and are replicated by the virus pro-
teins E1 and E2.9,29 The E2 protein, which is one of the
functional proteins of HPV, works as a transcriptional
activator and repressor for virus. After infection to host
cells, the complex of E2 and E1 proteins down-regulates
other viral proteins on its genome and starts viral prolifer-
ation. HPV genome also contains three oncogenes, E5,
E6, and E7 and two structural proteins, L1 and L2. In
some types of HPV to be known as a responsible virus
for developing cervical cancer, E6 and E7 proteins regu-
lated by E2 protein interact with p53 and Rb proteins
and turn host cells to cancer.10,11 E2 protein composes
three domains, activator domain (AD), HD, and DNA-
binding domain (DBD). AD is an N-terminal domain of
E2 protein, and then activates transcription and replica-
tion of the site of E2 protein interacting with E1 protein
is conserved on AD.30,32 A C-terminal domain, DBD,
binds to specific sequence ACCN6GGT and formed
DNA–protein complex. The structure of DBD/DNA
complex has been reported, and DBD is known to induce
apoptotic cell death and bind to p53.12,13,38 Both do-
mains, AD and DBD, are found in many types of HPV,
and each function is well studied.12,33 On the contrary to
them, HD is constituted with flexible amino acids se-
quence and the homology among HPV types is lower than
those of other two domains, and its function is still un-
known. In a recent study, however, direct interactions
of HDs in each HPV 11 and 16 E2 protein with E7 protein
was observed in vivo and in vitro by Gammoh et al.14 And
Zou et al. demonstrated that HD contained critical deter-
minates for nuclear localization and nuclear matrix asso-
In order to obtain broad information on the binding
molecules of PT, we focused on T7 phage display.17,34
T7 phage particles can exhibit a huge number of pro-
teins and peptides, which are derived from foreign
DNA insert, on coat proteins as fusion proteins. After
affinity selections of a phage library to a target molecule,
peptides or proteins that bind to the molecule can be
easily elucidated by sequencing insert DNA in the affin-
ity phages. This method allows a rapid selection and an
identification of binding peptides or protein to the target
molecules of interest. Recently, several binding proteins
or binding sites of small molecules have been determined
by using the phage display technology.18–20
We report here in a T7 phage display screening of PT-
binding molecules and possibilities of direct interaction
between PT and papillomavirus proteins.
2. Result
2.1. Synthesis of biotynilated podophyllotoxin and assay
for cytotoxicity
Studies on structure–activity relationships of PT by
´
´
Lopez-Perez et al. established that the acylation of the
hydroxyl group at C-9 did not influence on the cytotox-
icity.22 Therefore we decided to connect a biotin moiety
to the hydroxyl group C-9 in PT. PT was coupled with
biotinylaminocaproic acid to furnish biotinylated PT
(Bio-PT) in 23% yield (70% yield based on the recovered
PT) (Scheme 1).
OH
O
O
O
NH
HN
H
O
H
O
H
N
O
S
+
HO
O
O
H3CO
OCH3
OCH3
NH
HN
H
O
H
H
N
S
O
O
O
O
EDCI
O
DMAP
O
CH2Cl2–DMF
23%
H3CO
OCH3
(Unreacted podophyllotoxin was
recovered in 70% yield.)
OCH3
Scheme 1. Synthesis of a biotinylated podophyllotoxin (Bio-PT).

T. Saitoh et al. / Bioorg. Med. Chem. 16 (2008) 5815–5825
5817
We determined the cytotoxicity against murine leukemia
a
cell line L1210. Bio-PT had cytotoxicity against L1210
cells with IC50 values of 400 nM, which was about 60-
fold less potent than that of PT (7 nM). Nevertheless,
Bio-PT still kept the cytotoxicity and would work as a
tool for identifying the binding molecules of PT.
Number of
amino acid
amino acid sequence
frequency
SVPSRRRPDGRTHRSSRG
18
15/96
2.2. Screening of peptides selectively binding to PT by
phage display and homology search
9
8
7
6
5
4
3
2
1
0
b
Peptides specifically binding to PT were screened from
T7 phage library constructed with Drosophilae random
cDNA by using Bio-PT which immobilized streptavi-
din-coated wells.20 For screening, a suspension of
1.0 · 108pfu/well of phage library 200 lL was applied
onto each well. The wells were washed with 200 lL of
100 mM Tris–HCl (pH 8.0), followed by the wash buffer
(3% Tween 20, 2 M urea, 3 M NaCl, 100 mM Tris–HCl,
pH 8.0) to eliminate unbound and non-specifically
bound phage particles. The remainings on the wells were
recovered by infection of the log phase E. coli (200 lL)
applied on the wells. After four rounds of biopanning
according to this method, the DNA sequences of ran-
domly selected 96 clones were analyzed. Thus, we found
that single clone was highly concentrated (Fig. 2a). The
PT-binding clone particles conserved a specific cDNA se-
quence to encode 18 amino acids (SVPSRRRPDG
RTHRSSRG) on the surface.
18 amino acids NoIns phage
phage clone
clone
c
1
188 189
303 304
387
AD
HD
DBD
HPV 1a E2
233
240
250
260
QRQQKTPVRRRPYGRRRSRSPRGGGRRE
* **** **. ** **
SVPSRRRPDGRTH-RSSRG
In order to check the binding specificity of the peptide
against PT, the clone was amplified and the resulting
single clone was applied for the binding assay.23 The
phage particle with no inserted cDNA fragment (NoIns
phage) was used for the same binding assay to be a neg-
ative control. We compared the recovery titers of each
clone from PT immobilized wells (PT(+) wells) and
non-immobilized wells (PT(ꢀ) wells) (Fig. 2b). The titers
of the PT-binding clone derived from PT(+) wells were
5-fold higher than those from PT(ꢀ) wells. On the other
hand, there is no difference between PT(+) wells and
PT(ꢀ) wells in the recovered titer of the NoIns phage
clone. These results indicated that the PT-binding clone
was effectively concentrated in the biopanning to be due
to the specificity.
Figure 2. Result of screening with T7 phage particles constructed from
a random cDNA library of D. melanogaster. (a) Sequence result
obtained from concentrated phage particles after four rounds of
biopanning. DNA sequence analysis of 96 phage plaques showed 18
amino acids clone concentrated up to 15 clones, 15.6%. (b) Single clone
biopanning demonstrated 18 amino acids clone had specificity of
binding to Bio-PT immobilized to streptavidin-coated wells compared
with non-insert (NoIns) phage clone. Recovery rate was calculated by
(titer of PT-immobilized well/titer of non-immobilized well). (c)
Homology of 18 amino acids clone to HPV 1a E2 HD estimated by
BLAST.
combinant proteins on CM5 sensor chip (Fig. 3a). The
binding between PT and full length 1a E2 was analyzed
with the BIAEVALUATION 4.1 software and the dis-
sociation constant KD(M) was determined by using gen-
eral fitting analysis. The KD value of PT binding to full
length 1a E2 was determined to be 24.1 lM. And then,
we also observed the interaction and determined the
KD value of EP binding to full length 1a E2 (Fig. 3b).
The binding activity of EP to HPV 1a E2 protein was
10-fold weaker than that of PT (EP: KD = 210 lM).
A homology search on BLAST demonstrated that the
amino acid sequence displayed on the PT-binding clone
was similar to that of the hinge domain (HD) in the E2
protein of human papillomavirus (HPV) type 1a
(Fig. 2c).13–16 We focused on the fact that PT were used
as a treatment for genital warts caused by HPV, and we
then decided to examine the effect of PT on HPV E2.
In order to confirm the binding domain of PT in HPV
1aE2 protein, we next expressed and purified recombi-
nant each domain, AD, HD, and DBD. Then the bind-
ings between PT and each domain were analyzed by the
Biacore apparatus (Fig. 3c and d). The KD value ob-
tained from HD, which has a similar sequence to 18
amino acid sequence displayed on PT-binding phage
clone, was estimated to be 1.05 lM. However, the KD
values obtained from AD and DBD, which have no
similarity to 18 amino acid sequence, showed that both
2.3. Kinetic analysis of the binding between PT and HPV
1a E2 protein by SPR
At first, we investigated the interaction between a re-
combinant HPV E2 protein and PT by surface plasmon
resonance (SPR) analysis. The interaction between PT
and recombinant full length HPV 1a E2 protein was
analyzed by SPR biosensor instrument, Biacore3000
(GE Healthcare). The several concentrations of PT were
employed to determine the binding to the conjugated re-

5818
T. Saitoh et al. / Bioorg. Med. Chem. 16 (2008) 5815–5825
a
50
40
30
20
10
0
-10
-50
0
50
100
150
200
250
Time (sec)
50
40
30
20
10
0
b
c
-10
-50
0
50
100
150
200
250
Time (sec)
1
1
387
387
Full length
HPV 1a E2
His6
His6
FLAG
FLAG
188
AD
189
303
304
HD
His6
FLAG
His6
DBD
FLAG
d
Domain of HPV 1a E2
KD (μM)
24.1 3.1
>400
full length
AD
1.05 0.44
>400
HD
DBD
Figure 3. SPR instrument analysis of interaction between PT and E2 protein. (a) and (b) SPR (Biacore 3000, GE healthcare) sensorgram of
compounds (a, PT; b, EP) with immobilized full length 1a E2 protein on CM5 chip. The concentrations of compounds were 100, 50, 25, 12.5, 6.25 lM
from up to bottom. Kinetic values were analyzed by BIAEVALUATION 4.1 (GE healthcare) and were determined (PT, KD = 24.1 lM; EP,
KD = 210 lM). (c) Construct for recombinant proteins of HPV 1a E2 each domain. To be solved easily, FLAG tag was inserted instead of T7 tag.
After infection by E. coli, the expressed proteins were extracted and purified on a FPLC system. (d) Kinetics values of PT and HPV 1a E2 each
domain. SPR analysis was performed at flow late 20 lL/min and at 25 ꢁC with running buffer. Each recombinant protein was immobilized on CM5
chip with amine coupling reaction, and sensorgrams were analyzed by BIAEVALUATION 4.1.
domains did not bind to PT (KD > 400 lM). When we
attempted electro mobility shift assay (EMSA) with re-
combinant E2 protein, PT had no inhibition between
the E2–DNA interaction (data not shown).24 Thus the
results demonstrated here suggested that PT specifically
binds to the HD in HPV 1a E2 protein, implying that PT
binds to the amino acid sequence, which is similar to
that selected by the phage screening.
recombinant E2 and E7 proteins of HPV 1a to be His
tag fusion proteins respectively, and analyzed the inter-
action between them. The E7 protein was immobilized
on a CM5 sensor chip, and the E2 protein was applied
as an analyte. The binding of E2 protein on E7 protein
was observed at various concentrations (Fig. 4a). The
KD value of 0.73 lM was obtained by separate fitting
of the sensorgrams to a 1:1 Langmuir model. The result
implied that the interaction between HPV 1a E2 and E7
proteins was performed in vitro by the SPR experiments.
2.4. Direct protein–protein interaction between E2 and E7
of HPV 1a observed by SPR
2.4.1. Inhibition of the E2/E7 interaction by PT. We dem-
onstrated that PT bound to HD of HPV the E2 and that
1a E2 protein interacted with E7 protein obtained from
the SPR experiments as described above. We next sur-
veyed whether PT inhibit this interaction or not. We
Gammoh et al. reported that E2 proteins of HPV 11 and
16 interact with each E7 protein directly by HD in vivo
and in vitro.14 We obtained the interacting from E2 pro-
tein and E7 protein of HPV 1a by SPR. We prepared the

T. Saitoh et al. / Bioorg. Med. Chem. 16 (2008) 5815–5825
5819
1500
1000
500
0
-500
0
30
60
90
120
150
Time (sec)
400
300
200
100
0
PT (μM)
0
50
125
250
500
500 μM without
E2 protein
buffer
-100
0
30
60
90
120
150
Time (sec)
c
120
100
80
60
40
20
0
0
50
125
250
500
PT (μM)
Figure 4. HPV 1a E2/E7 protein–protein interaction was determined with SPR and inhibited by PT. (a) Sensorgram of E2 protein with immobilized
E7 protein. The concentrations of E2 protein were 1.0, 0.5, 0.25, 0.125 lM from up to bottom. And then dissociation constant (KD) was determined
to be 0.73 lM from this sensorgram with BIAEVALUATION 4.1. (b) Inhibition analysis of E2/E7 interaction by PT with SPR. Tested solutions of
E2 protein were adjusted to 0.1 lM concentration contained with several concentration of PT in 5% DMSO. Analysis was performed at flow rate
30 lL/min and at 25 ꢁC. Buffer and 500 lM PT without E2 protein were showed no effect on sensorgram. (c) Comparison with bound E2 protein
amounts at 115 s after injection. Resonance unit of control (PT 0 lM) was regarded as 100% and the amounts of bound E2 protein with immobilized
E7 protein were reduced by the presence of PT.
monitored the amounts of E2 protein bound to immobi-
lized E7 protein in the presence or absence of PT by
SPR.35 After a solution of 0.1 lM E2 protein was pre-
mixed with several concentrations of PT in 5% DMSO,
the resulting solution of the mixture was employed for
analysis of the binding to E7 protein. The bound
amounts of E2 protein was determined from the re-
sponse units at 115 s (Fig. 4b and c). When the bound
amounts of E2 protein on E7 protein in the absence of
PT was regarded as 100%, the binding amounts of E2
protein decreased in the presence of PT. Finally, the rate
of binding by E2 protein to E7 protein at concentration
of 500 lM PT reached up to 49.1%. We also tested
whether PT bound to E7 protein, however, PT did not
bind or weakly bound to E7 protein (data not shown).
These results suggested that PT bound to directly E2
protein and inhibited the protein–protein interaction be-
tween E2 and E7 in vitro.
2.4.2. Comparison the abilities to inhibit E2/E7 interac-
tion between PT and EP. The KD value in the EP/E2
interaction was 10-fold weaker than that in the PT/E2
interaction (Fig. 3a and b). And the inhibition of HPV
1a E2/E7 interaction by PT was as above mentioned

5820
T. Saitoh et al. / Bioorg. Med. Chem. 16 (2008) 5815–5825
(Fig. 4b). Then, we also checked the inhibition by EP.
The amounts of E2 protein bound to E7 protein was de-
creased to 82.0% and 60.6% by the addition of 250 lM
of EP and PT, respectively (Fig. 5a and b). These results
suggested that the binding abilities of the proteins can be
responsible to the difference of inhibition activities.
3. Discussion
Podophyllotoxin (PT) is a natural lignan that has strong
cytotoxicity against several cancer lines. However, the
clinical use of PT in the treatment of cancer has been
limited by severe toxic side effects.4 The cytotoxicity of
PT is considered to be due to its inhibitory effect of
tubulin polymerization. On the other hand, etoposide
(EP), a semisynthetic derivative of PT, is clinically used
as an anticancer drug. EP induces topoisomerase II-
mediated DNA cleavage, however, it has no effect on
the tubulin assembly. We have been interested in the dif-
ference of biological properties between PT and EP.
2.5. Kinetic analysis of the binding between PT and other
HPV E2 proteins, HPV 11 and 16, by SPR
Having found that PT bound to E2 proteins of HPV 1a,
we assessed whether PT bound other E2 proteins de-
rived from other HPV types. Amino acid sequences of
AD and DBD in HPV E2 have been widely known to
be highly conserved between many types.13,25 On the
other hand, HD has been known to be scarcely con-
served between many HPV types. And the homologies
between 18 amino acids and amino acid sequences of
HD in other types such as high risk type HPV 16 and
low risk type HPV 11 has been checked (Fig. 6a). The
similarities in both HD to 18 amino acid sequences were
lower than that in HPV 1a. However, the KD values for
the binding between PT and recombinant HPV 11 and
16 E2 proteins which expressed in Escherichia coli were
not so much difference to that of HPV 1a (Fig. 6b and
c). These results would indicate that PT bound to other
types of E2 protein in spite of the lower homologies be-
tween E2 proteins and amino acid sequence obtained
from phage display screening in this text.
In order to screen binding molecules of PT, we prepared
the biotinylated PT derivative (Bio-PT) (Scheme 1). Bio-
tin moiety was substituted on a hydroxyl group at C-9
position in PT by a condensation reaction. It has been
reported that the hydroxyl group at C-9 position in
PT was essential for the inhibitory activity of tubulin
polymerization.27 However, some 9-O-acyl derivatives
still have cytotoxicity, while they lost the inhibitory
activity of tubulin polymerization.4,26–28 Actually, our
Bio-PT showed cytotoxicity against L1210 cell line with
IC50 value of 400 nM. Although the cytotoxicity of Bio-
PT was less potent than that of PT (IC50 value of 7 nM),
it was still strong. These results strongly suggested the
presence of other target molecules of PT. We applied a
T7 phage display method to screen for PT-binding mol-
a
2500
2000
1500
1000
500
0
DMSO
250 μM EP
250 μM PT
0
20
40
60
80
100
120
140
Time (sec)
120
b
100
80
60
40
20
0
DMSO
PT
EP
Figure 5. Comparison with PT and EP at inhibition of E2/E7 interaction. The concentration of E2 was adjusted to 0.5 lM. (a) EP less inhibited E2/
E7 interaction than PT. (b) At 115 s after injection, a solution of 250 lM PT inhibited E2 protein binding activity about 60% compared with a
solution of 5% DMSO, but a solution of 250 lM EP inhibited about 80%.

T. Saitoh et al. / Bioorg. Med. Chem. 16 (2008) 5815–5825
5821
a
230
HPV 11 SACTTEDGVSAPPRKRARGPSTNNTLCVANIRSVDSTINNIVTDNY
*.*.*:*. * *: *:
---------SVPSRRRPDG------------RTHRSSRG-------
240
250
260
270
202 210 220 230 240
250
270
HPV 16 SNEVSSPEIIRQHLANHPAATHTKAVALGTEETQTTIQRPRSEPDTGNPCHTTKLLHRDSVDS
* * .* .** **.*
----SVP--------------------------------SRRRPDG--------RTHRSSRG-
1
1
1
387
367
Full length
HPV 1a E2
Full length
HPV 11 E2
b
c
His6
His6
His6
FLAG
FLAG
FLAG
365
Full length
HPV 16 E2
HPV type
KD (μM)
1a
11
24.1 3.1
20.1
70.5
0.7
8.5
16
Figure 6. Analysis for interaction between PT and HPV 11 and 16 E2 proteins. (a) Alignment between 18 amino acids sequence and hinge domain of
HPV 11 and 16 E2 proteins obtained by CLUSTALW. The results of alignments showed that 18 amino acids sequence was similar to the part of
hinge domain in HPV 11 E2 (S235-G243). But this peptide was not matched enough to HPV 16 E2 on the primary structure. (b) Construction of
recombinant HPV E2s. These proteins were expressed and purified same as HPV 1a E2 protein. (c) Comparison among KD values by Biacore 3000 at
three types HPV E2 proteins with PT. These results were obtained by same methods of Figure 3a and b.
ecules. After screening, we obtained a phage clone,
which conserved 18 amino acids (SVPSRRRPD
GRTHRSSRG), and we proved that this clone bound
PT by binding assay. Although the NoIns clone had
no affinity in the PT-immobilized wells, the selected
clone showed affinity in the PT-immobilized wells 5-fold
greater than that in non-immobilized wells.
vator domain (AD) in the amino-terminal domain, a
hinge domain (HD) of the intermediate region between
two functional domains, and a DNA-binding domain
(DBD) in the carboxyl-terminal domain. E2 interacted
at its HD region with several cellular proteins, such as
E7 protein, nuclear matrix, and pre-mRNA splicing
factor.
One advantage to display the drug-binding amino acid se-
quences on phage particles is that the technique can also
identify a binding site within the binding protein.18,23 In-
versely, the candidates for the binding proteins can be
estimated by analyzing the peptide sequences selected
by phage display screening. Thus, we analyzed the se-
lected sequence (SVPSRRRPDGR THRSSRG) to ob-
tain clues for PT-binding proteins. A homology search
of the selected sequence by using BLAST programs gave
several candidates for PT-binding proteins. Among them,
our tension has focused on a HPV 1a E2 protein. The se-
lected peptide was similar to a part of the hinge domain of
HPV 1a E2 protein. HPV 1a has been known to cause
plantar warts, and PT has been clinically used for the
treatment of plantar and genital warts caused by
HPVs.2,36 However, the interactions between PT and pro-
teins of HPV have never been studied before.
The binding between PT and HPV 1a E2 protein was con-
firmed by a SPR experiment (Fig. 3). PT binds HPV 1a E2
protein with a dissociation constant (KD) of 24.1 lM. PT
is known as a tubulin polymerization inhibitor, and the
KD value for PT binding to tubulin was reported to be
0.6–0.7 lM.37 The KD value for tubulin is 40-fold stron-
ger than the value for full length HPV 1a E2 protein,
but these values were determined with purified proteins
in vitro, and clarifying the relationship between KD values
and efficacy of PT depends on the further studies. We also
found that PT selectively interacted with the hinge do-
main of HPV 1a E2 protein, while it did not show any
affinity with other domains (AD and DBD) (Fig. 2 and
4c). These results indicated that PT selectivity bound to
the hinge domain of HPV 1a E2 protein, and suggested
that PT binds to the amino acid sequence, which was sim-
ilar to that selected by phage display screening. Interest-
ingly, EP weakly interacted with HPV 1aE2 protein
(Fig. 3b). Thus, the trimethoxyphenyl moiety at E ring
or the stereochemistry at C-9 in PT will be important
for the interaction between PT and HPV 1a E2 protein.
The E2 protein plays an important role as a major reg-
ulator of both viral transcription and DNA replication.
The E2 protein has three domains, a transcriptional acti-

5822
T. Saitoh et al. / Bioorg. Med. Chem. 16 (2008) 5815–5825
Gammoh et al. reported that E2 protein of HPV 11 or
16 directly bound with the corresponding E7 protein
at HD in vitro and in vivo.14 Thus, we next investigated
the effect of PT on the interaction between HPV 1a E2
protein and E7 protein. We confirmed that HPV 1a E2
protein bound with HPV 1a E7 protein in vitro by
SPR (Fig. 4a). The obtained KD value for E2 protein
with E7 protein was 0.73 lM. The interaction between
E2 protein and E7 protein was inhibited by addition
of PT (Fig. 4b and c). The relative binding ratio de-
creased about half when 500 lM of PT was added, indi-
cating that PT apparently inhibited the interaction
between E2 protein and E7 protein. We also found that
the inhibitory effect of EP was much weaker than that of
PT (Fig. 5).
protein in HPV 1a. We also demonstrated that PT
bound with HD region in E2. Thus, PT inhibits the
E2/E7 interaction by binding to HD of E2 protein.
These results suggest that PT has a potency to develop
a seed for novel derivatives to investigate the molecular
mechanism of HPV life cycle with chemical genetics ap-
proach. Structural optimization of PT as a specific
inhibitor of the E2/E7 interaction and further biological
effects of PT are currently underway.
4. Experimental
4.1. Compounds
Podophyllotoxin and etoposide were obtained from Sig-
ma–Aldrich (St. Louis, MO).
We identified that PT inhibited the interaction between
HPV 1a E2 protein and E7 protein. Although the inhib-
itory effect is weak, this is the first example of E2 protein
inhibitor which inhibits E2/E7 protein–protein interac-
tion. Known E2 protein inhibitors target the interactions
between E2 protein and DNA, or between E2 protein
and E1 protein.39,40 Since our results showed that PT
bound to the HD of HPV 1a E2, and the HD was re-
ported to interact with several transcription factors, PT
might inhibit the interaction of E2 protein and other pro-
teins. The interaction between E2 protein and other cel-
lular factor will be studies in due course. The molecular
mechanism of HPV infection, proliferation, and cancer
formation has not been clarified yet. The modification
and structure optimization of PT as an E2 protein inhib-
itor will develop powerful tools to reveal the function of
E2 protein as well as the virus life cycle.
4.2. T7 phage library
The preparations of T7 phage library by cDNA from
Drosophila melanogaster and synthetic random DNAs
were followed by a previous report.20
4.3. Synthesis of biotinylated podophyllotoxin
To a solution of podophyllotoxin (10.8 mg, 26.0 lmol)
and 6-biotinylaminocaproic acid (9.6 mg, 26.9 lmol) in
CH2Cl2-DMF (4:1, 2.5 mL) was added EDCI
(14.5 mg, 75.6 lmol) and DMAP (10.9 mg, 89.2 lmol)
and the mixture was stirred at room temperature for 2
days. Solvents were removed under a reduced pressure.
The residue was purified on a silica gel chromatography
(CHCl3/MeOH = 9: 1) to yield podophyllotoxin–biotin
PT has been used as a treatment for genital warts includ-
ing condyloma acuminate. HPV 1a is known to cause
plantar warts, while genital warts are caused by HPV
11 mainly.7 Our results suggest that PT influences to
E2 function HPV 1a by binding. Furthermore, since
we aimed to assess the relationship between the potency
of PT and HPV 11, at first, we searched a homology be-
tween HPV 11 E2 protein and 18 amino acids, which we
obtained by screening (Fig. 6a). When we observed the
interaction between HPV 11 E2 protein and PT, we
tested about E2 protein of HPV 16 in order to compare
to HPV 1a and 11. HPV 16 is the most famous and re-
searched type of HPVs and is known as the virus to
cause the cervical cancer. Even though the HD domains
of both proteins have low similarity if compared to the
selected peptide, PT interacted with both E2 proteins
interestingly. And KD values obtained for PT-binding
E2 proteins of HPV 11 and 16 were 20.1 lM and
70.5 lM, respectively (Fig. 6c). We assume that PT
interacts with the HD of HPV 11 and 16 at the motif,
which are spatially close together in the three-dimen-
sional structure. The motif is composed of amino acids,
which is similar to a part of the peptide selected by
phage display. The structure of HD has not been re-
vealed, however, further structure studies are required
to clarify the prediction.
conjugate (4.5 mg, 23%) and recovered podophyllotoxin
25
D
(7.6 mg, 70%). ½aꢁ ꢀ66 (c 0.20, MeOH); 1H NMR
(600 MHz, CD3OD) d 6.82 (1H, s), 6.52 (1H, s), 6.43
(2H, s), 5.97 (1H, s), 5.95 (1H, s), 5.91 (1H, d,
J = 13.6 Hz), 4.60 (1H, d, J = 4.6 Hz), 4.46 (1H, dd,
J = 7.7 Hz, 5.0 Hz), 4.36 (1H, m), 4.28 (1H, dd,
J = 7.7 Hz, 4.5 Hz), 4.22 (1H, t, J = 9.8 Hz), 3.72 (9H,
s), 3.34 (1H, s), 3.19 (3H, m), 2.89 (1H, dd,
J = 12.7 Hz, 5.0 Hz), 2.80 (1H, m), 2.67 (1H, d,
J = 12.7 Hz), 2.48 (2H, m), 2.17 (2H, t, J = 7.3 Hz),
1.70 (2H, m), 1.61 (2H, m), 1.53 (2H, m), 1.40 (4H,
m). 13C NMR (100 MHz, CDCl3–CD3OD) d 174.1,
173.8, 173.2, 163.8, 152.5, 152.5, 148.0, 147.5, 137.0,
134.8, 132.2, 128.2, 109.6, 108.0, 108.0, 106.9, 101.5,
73.4, 71.3, 61.7, 60.6, 59.9, 56.1, 56.1, 55.2, 45.5, 43.6,
40.3, 39.0, 38.7, 35.7, 34.0, 29.1, 28.1, 27.9, 26.3, 25.3,
24.4. IR (film) cmꢀ1 3288, 2934, 1775, 1699, 1650,
1589, 1542, 1505, 1484, 1460, 1421, 1376, 1332, 1239,
1158, 1126, 1037, 998, 931, 866. HRMS (ESI) calcd
for C38H47N3O11SNa ([M+Na]+) 776.2823; found:
776.2841.
4.4. Cell line and cytotoxicity assay
The murine lymphocytic leukemia L1210 cell line was
supported by Cell Resource Center for Biomedical Re-
search Institute of Development, Aging and Cancer To-
hoku University. Cells were maintained in DMEM
supplemented with 10% horse serum and grown in this
In this report, we screened the peptide binding PT by T7
phage display and it was similar to a part of HD of E2

T. Saitoh et al. / Bioorg. Med. Chem. 16 (2008) 5815–5825
5823
medium at 37 ꢁC in 5% CO2/95% air. Cells were sus-
pended at 1.0 · 104 cells in 100 lL of the media and cul-
tured 24 h as preincubation. After cells were exposed to
graded concentration of the compounds for 48 h, Cell
Count Reagents SF, WST-8 (Nacalai tesque, Kyoto, Ja-
pan), was added 10 lL and cells were incubated for
4 h.21 Then, absorbance at 450 nm was measured. Media
and supplement were from Nacalai tesque (Kyoto,
Japan).
on a streptavidin-coated 96-well plate. After washing
by a same manner as above described, binding phage
particles were eluted with 200 lL of Tris buffer including
1% SDS. The titers of the eluted phage particles were
determined by a method described in an previous
paper.23
4.7. Plasmid constructs
An expression vector pDF-28a (+), which has a FLAG
tag, a 6 · His tag, and a multicloning site of pET-28a
(+), was constructed by a method based on pET
DUET-1 (Novagen). A FLAG tag fragment was gener-
ated with cloning ligated complementary synthetic oligo-
nucleotides: 50-TCCATGGGCAGCCATCATCATCA
TCATCACAGCAGCGACTACAAG-30 and 50-AGCC
ATATGTTTGTCATCGTCGTCCTTGTAGTCGCT
GCTGTGATG-30 into the NcoI/NdeI sites of pET
DUET-1 instead of thrombin recognized site, and then
a multicloning site of pET-28a (+) cloned with NdeI/
XhoI was ligated.
4.5. Biopanning and sequence analysis
A biotinylated derivative of PT was immobilized on a
streptavidin-coated 96-well plate (Nalge Nunc Interna-
tional, Wiesbaden, Germany) overnight at 4 ꢁC, and
unbound PT was removed by washing three times
with 200 lL of 100 mM Tris–HCl (pH 8.0). Then,
immobilized wells were blocked with 200 lL of Tris
buffer containing 3% skimmed milk for 1 h, and
washing three times with Tris buffer. Phage library
solution contain 1% skimmed milk was added in well
and incubated at rt for 3 h. After washing 10 times
with Tris buffer in order to unbound phage particles,
wells washed another ten times to eliminate non-spe-
cific bound phage particles with washing buffer (3%
Tween 20, 2 M urea, 3 M NaCl, 100 mM Tris–HCl,
pH 8.0). The remaining phage particles were recov-
ered by infection to 200 lL of E. coli BLT5615
(Novagen, Darmstadt, Germany), were grown a log
phase, and were amplified at 37 ꢁC until E. coli be-
came bacteriolysis. Lysis solutions were used in next
biopanning round and picking a single plaqueup.
After four rounds of this method, 96 single plaques
were picked up at random from LB with 50 lg/mL
carbenicillin plates and each was dissolved in a phage
extraction solution (20 mM Tris–HCl, pH 8.0,
100 mM NaCl, 6 mM MgSO4). Phage DNA inserted
random cDNAs was amplified by PCR using forward
primer (50-GCTAACTTCCAAGCGGACC-30) and re-
verse primer (50-TTGCCCAGAACTCCCCAA-30, T7
select Down, Novagen). The PCR products were puri-
fied with ExoSAP-It (GE healthcare Bio-Science
Corp, Piscataway, NJ) and used cycle sequence PCR
with BigDyeꢂ Terminator v3.1 Cycle Sequencing Kit
(Applied BioSystems, Foster City, CA), then frag-
ments were precipitated with 100% ethanol. The puri-
fied PCR products were sequenced on an ABI
PRISMꢂ 3100 Genetic Analyzer (Applied BioSystems,
Foster City, CA). From results, the amino acid se-
HPV genomes, human papillomavirus type 1a (VG008),
human papillomavirus type 11 (VG012), and human
papillomavirus type 16 (VG013) were provided from
the Japan Health Sciences Foundation (HSRRB, Osaka,
Japan). E2 and E7 proteins of HPV in full length were
derived by PCR with primers conjugated with NdeI/
XhoI or NheI/XhoI restriction sites referred to National
Center for Biotechnology Information (NCBI, http://
www.ncbi.nlm.nih.gov/) (1a E2 forward primer: 50-CA
TATGAACCTATATGAACAGGA-30, 1a E2 reverse
primer: 50-CTCGAGTTACAATAAATGTAATGA-30,
11 E2 forward primer: 50-CATATGAACCTATAT
GAACAGGAC-30, 11 E2 reverse primer: 50-CTC
GAGTTAAGACCCATTAAACTG-30, 16 E2 forward
primer: 50-GCTAGCATGGAGACTCTTTGCCAA-30,
16 E2 reverse primer: 50-CTCGAGTCATATAGAC
ATAAATCCAG-30, 1a E7 forward primer: 50-CAT
ATGGTGGGCGAAATGCCA-30, 1a E7 reverse pri-
mer:
50-CTCGAGTTACTGTCGCTGTAGGGTGC
-30), and were cloned into pDF-28a (+).
Three domains of HPV 1a E2 were constructed by PCR
referred to Pfam of European Molecular Biology Labo-
ratory (EMBL) with primers conjugating with NdeI/
XhoI (1a E2 AD 1–188 forward primer to be same as
full length forward primer, reverse primer: 50-CTC
GAGTTAGCTAGTGGAAGACATAAC-30, HD 189-
303 forward primer: 50-CATATGTCCCCAAGGGCT
GCTGGG-30, reverse primer: 50-CTCGAGTTATGG
ATCCCAAGCTTCCTGCA-30, DBD 304–387 forward
primer: 50-CATATGCCCGTGGTCTGTGTAAAA-30,
reverse primer to be same as full length reverse primer),
and were cloned into pDF-28a (+).
quence on T7 phage capsid was determined.
A
homology search (BLAST) demonstrated that the
amino acid sequence obtained from the selected clone,
SVPSRRRPDGRTHRSSRG, is similar to that of
HPV 1a hinge domain.
4.6. Affinity for PT by a selected phage single clone
4.8. Expression and purification of recombinant HPV
proteins
The affinity of a selected clone was compared to that of
a non-displayed clone which was inserted with only stop
codon. The non-displayed clone was merely selected by
biopanning. Both titers each clones were adjusted to
5.0 · 1010 pfu/mL. Then, 200 lL of each phage suspicion
(i.e., 109 pfu) was incubated with Bio-PT immobilized
Protein expression was performed by transforming the
constructs into Rosetta (DE3) pLysS (Novagen).
These bacteria were grown up in 1 L of LB medium
containing 50 lg/mL carbenicillin and 100 lg/mL

5824
T. Saitoh et al. / Bioorg. Med. Chem. 16 (2008) 5815–5825
5. Burden, D. A.; Kingma, P. S.; Froelich-Ammon, S. J.;
Bjornsti, M.-A.; Patchan, M. W.; Thompson, R. B.;
Osheroff, N. J. Biol. Chem. 1996, 271, 29238–29244.
6. Bernard, H. U. J. Antimicrob. Chemother. 2004, 53, 137–139.
7. Scheinfeld, N.; Lehman, S. L. Dermatol. Online J. 2006,
12, 5.
8. Oliveira, J. G.; Colf, L. A.; McBride, A. A. Proc. Natl.
Acad. Sci. U.S.A. 2006, 103, 1047.
9. Del Vecchio, A. M.; Romanczuk, H.; Howley, P. M.;
Baker, C. C. J. Virol. 1992, 66, 5949.
chloramphenicol. Cells were grown at 37 ꢁC until
OD600 came up to 0.5, and then treated 1 mM of iso-
propyl thio-b-D-galactoside (IPTG). After 3 h, cells
were harvested by centrifugation at 3000g for
30 min, and were suspended up to 50 mL with the
buffer (20 mM sodium phosphate, pH 7.4, 0.5 M
NaCl, 20 mM imidazole, 2 M urea, 5% glycerol,
0.05% Triton X-100) and sonicated. The cell lysates
were centrifuged at 20,000g for 10 min and the pro-
teins were purified with 5 mL His-Trap HP column
10. Band, V.; De Caprio, J. D.; Delmolino, L.; Kulesa, V.;
Sager, R. J. Virol. 1991, 65, 6671.
¨
(GE healthcare) of the FPLC system (AKTA ex-
11. Huang, P. S.; Patric, D. R.; Edwards, G.; Goodhart, P. J.;
Huber, H. E.; Miles, L.; Garsky, V. M.; Oliff, A.;
Heimbrook, D. C. Mol. Cell. Biol. 1993, 13, 953.
12. Giri, I.; Yaniv, M. EMBO J. 1988, 7, 2823.
13. Hedge, R. H. Annu. Rev. Biophys. Biomol. Struct. 2002, 31, 343.
14. Gammoh, N.; Grm, H. S.; Massimi, P.; Banks, L. J. Virol.
2006, 80, 1787.
15. Zou, N.; Lin, B. Y.; Duan, F.; Lee, K.-Y.; Jin, G.; Guan,
R.; Yao, G.; Lefkowitz, E. J.; Broker, T. R.; Chow, L. T.
J. Virol. 2000, 74, 3761.
plorer, GE healthcare) using the elution buffer
(20 mM sodium phosphate, pH 7.4, 0.5 M NaCl,
500 mM imidazole, 5% glycerol, 0.05% Triton X-
100). The eluted fraction was used to confirm the
purity by SDS–polyacrylamide gel electrophoresis.
The buffers of eluted proteins were exchanged to the
storage buffer (PBS, pH7.4, 5% glycerol) with PD-10
columns (GE healthcare) and stored at ꢀ80 ꢁC.
16. Lai, M.-C.; Teh, B. H.; Tarn, W.-Y. J. Biol. Chem. 1999,
274, 11832.
4.9. Surface plasmon resonance (SPR) analysis
17. Smith, G. P.; Petrenko, V. A. Chem. Rev. 1997, 97, 391.
18. Rodi, D. J.; Janes, R. W.; Sanganee, H. J.; Holton, R. A.;
Wallace, B. A.; Makowski, L. J. Mol. Biol. 1999, 285, 197.
19. Jin, Y.; Yu, J.; Yu, Y. G. Chem. Biol. 2002, 9, 157.
20. Morimura, T.; Noda, N.; Kato, Y.; Watanabe, T.; Saitoh,
T.; Yamazaki, T.; Takada, K.; Aoki, S.; Ohta, K.;
Ohshige, M.; Sakaguchi, K.; Sugawara, F. J. Antibiot.
2006, 59, 625.
21. Orita, T.; Tsunoda, H.; Yabuta, N.; Nakano, K.; Yosh-
ino, T.; Hirata, H.; Ohtomo, T.; Nezu, J.; Sakumoto, H.;
Ono, K.; Saito, M.; Kumagai, E.; Nanami, M.; Yoshiku-
bo, M.; Tsuchiya, M. Blood 2005, 105, 562.
Proteins were immobilized on CM5 chips using an
amine coupling kit (GE Healthcare) and a Biacore
3000 apparatus (GE Healthcare). The surface of chips
was activated by injecting a solution containing 0.2 M
N-ethyl-N0-(3-dimethylaminopropyl)
carbodiimide
hydrochloride (EDC) and 50 mM N-hydroxy-
succinimide (NHS) for 10 min. Proteins diluted with
10 mM sodium acetate buffer at pH 4.0–6.0 were in-
jected and the surface was then blocked by injecting
1 M ethanolamine at pH 8.5 for 10 min. PT at vari-
ous concentration in running buffer (PBS, pH 7.4,
0.005% Tween20, 5% glycerol, 5% DMSO) were flo-
wed at 20–30 lL/min. Regeneration of the sensor chip
was used 25 mM NaOH and 0.5 mM NaCl and rep-
licates of each experiments were performed over two
times. Response curves of dissociation constants
(KD) analysis were generated by subtraction of the
back ground signal generated simultaneously on the
control flow cell. BIAEVALUATION 4.1 software
(GE Healthcare) was used to determine the kinetic
parameters.
´
´
22. Lopez-Perez, J. L.; del Olno, E.; de Pascual-Teresa, B.;
Abad, A.; San Feliciano, A. Bioorg. Med. Chem. Lett.
2004, 14, 1283.
23. Aoki, S.; Ohta, K.; Yamazaki, T.; Sugawara, F.; Sakag-
uchi, K. FEBS J. 2005, 272, 2132.
24. Hou, S. Y.; Wu, S.-Y.; Chiang, C.-M. J. Biol. Chem. 2002,
277, 45619.
25. McBride, A. A.; Romanczuk, H.; Howley, P. M. J. Biol.
Chem. 1991, 266, 18411.
26. Thurston, L. S.; Imakura, Y.; Haruna, M.; Li, D.-H.; Liu,
Z.-C.; Liu, S.-Y.; Cheng, Y.-C.; Lee, K.-H. J. Med. Chem.
1989, 32, 604.
27. Terada, T.; Fujimoto, K.; Nomura, M.; Yamashita, J.;
Kobunai, T.; Takeda, S.; Wierzba, K.; Yamada, Y.;
Yamaguchi, H. Chem. Pharm. Bull. 1992, 40, 2720.
28. Kiselev, V. V. Chem. Nat. Compd. 1987, 23, 638.
29. Gopalakrishnan, V.; Khan, S. A. Proc. Nati. Acad. Sci.
U.S.A. 1994, 91, 9597.
At the observation of inhibition of E2/E7 protein–pro-
tein interaction, the flow cell of E7 immobilized was
used in order to determine the amount of binding
E2 protein. To avoid mass transport limitation, the
flow rate was 30 lL/min. The E2 protein solution con-
taining with DMSO, PT or EP was set just before
observation.
30. Storey, A.; Piccini, A.; Massimi, P.; Bouvard, V.; Banks,
L. J. Gen. Virol. 1995, 76, 819.
31. De Villies, E.-M.; Fauquet, C.; Broker, T. R.; Bernard,
H.-U.; Zur Hausen, H. Virology 2004, 324, 17.
32. Harris, S. F.; Botchan, M. R. Science 1999, 284, 1673.
33. Liang, H.; Petros, A. M.; Meadows, R. P.; Yoon, H. S.;
Egan, D. A.; Walter, K.; Holzman, T. F.; Robins, T.;
Fesik, S. W. Biochemistry 1996, 25, 2095.
34. Danner, S.; Belascco, J. G. Proc. Nati. Acad. Sci. U.S.A.
2001, 98, 12954.
35. Karin, M. M.; Svitkin, Y. V.; Kahvejian, A.; De
Crescenzo, G.; Costa-Mattiolo, M.; Sonenberg, N. Proc.
Nati. Acad. Sci. U.S.A. 2006, 103, 9494.
References and notes
1. Hartwell, J. L.; Scherecker, A. W. J. Am. Chem. Soc. 1951,
73, 2909–2916.
2. Bohrin, L.; Rosen, B. Drug Discov. Today 1996, 1, 343–
351.
3. Meresse, P.; Dechaux, E.; Monneret, C.; Bertounesque, E.
Curr. Med. Chem. 2004, 11, 2443–2466.
´
36. Sterling, J. C.; Handfield-Jones, S.; Hudson, P. M. Br. J.
Dermatol. 2001, 144, 4.
4. Desbene, S.; Giorgi-Renault, S. Curr. Med. Chem. 2002, 2,
71–90.

T. Saitoh et al. / Bioorg. Med. Chem. 16 (2008) 5815–5825
5825
37. Lobert, S.; Ingram, J. W.; Correia, J. J. Can. Res. 1999, 59,
4816.
38. Parish, J. L.; Kowalczyk, A.; Chen, H.-T.; Roeder, G. E.;
Sessions, R.; Buckle, M.; Gaston, K. J. Virol. 2006, 80,
4580.
39. Schaal, T. D.; Mallet, W. D.; McMinn, D. L.; Nguyen, N.
V.; Sopko, M. M.; John, S.; Parekh, B. S. Nucleic Acids
Res. 2003, 31, 1282.
40. Kasukawa, H.; Howley, P. M.; Benson, J. D. J. Virol.
1998, 72, 8166.