Maintaining potent HTLV-I protease inhibition without the P3-cap moiety in small tetrapeptidic inhibitors

Article abstract of DOI:10.1016/j.bmcl.2011.01.048

The human T cell lymphotropic/leukemia virus type 1 (HTLV-I) causes adult T cell lymphoma/leukemia. The virus is also responsible for chronic progressive myelopathy and several inflammatory diseases. To stop the manufacturing of new viral components, in our previous reports, we derived small tetrapeptidic HTLV-I protease inhibitors with an important amide-capping moiety at the P₃ residue. In the current study, we removed the P₃-cap moiety and, with great difficulty, optimized the P₃ residue for HTLV-I protease inhibition potency. We discovered a very potent and small tetrapeptidic HTLV-I protease inhibitor (KNI-10774a, IC₅₀ = 13 nM).

Full text of DOI:10.1016/j.bmcl.2011.01.048

Bioorganic & Medicinal Chemistry Letters 21 (2011) 1832–1837
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Maintaining potent HTLV-I protease inhibition without the P3-cap moiety
in small tetrapeptidic inhibitors
⇑
Jeffrey-Tri Nguyen, Keiko Kato, Henri-Obadja Kumada, Koushi Hidaka, Tooru Kimura, Yoshiaki Kiso
Department of Medicinal Chemistry, Center for Frontier Research in Medicinal Science, Kyoto Pharmaceutical University, Yamashina-ku, Kyoto 607-8412, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
The human T cell lymphotropic/leukemia virus type 1 (HTLV-I) causes adult T cell lymphoma/leukemia.
The virus is also responsible for chronic progressive myelopathy and several inflammatory diseases. To
stop the manufacturing of new viral components, in our previous reports, we derived small tetrapeptidic
HTLV-I protease inhibitors with an important amide-capping moiety at the P3 residue. In the current
study, we removed the P3-cap moiety and, with great difficulty, optimized the P3 residue for HTLV-I pro-
tease inhibition potency. We discovered a very potent and small tetrapeptidic HTLV-I protease inhibitor
(KNI-10774a, IC50 = 13 nM).
Received 15 December 2010
Revised 11 January 2011
Accepted 12 January 2011
Available online 15 January 2011
Keywords:
Human T cell lymphotropic virus
Adult T cell leukemia
Aspartic protease
Inhibitor
Ó 2011 Elsevier Ltd. All rights reserved.
Retrovirus
The human T cell lymphotropic/leukemia virus type 1 (HTLV-I)
is the only human retrovirus known to cause cancer. It is estimated
that 20–30 million people are infected with HTLV-I worldwide,1
with endemic foci in southern Japan, western Africa, Central and
South America,2 transmitted by close contact with infected T lym-
phocytes via breast-feeding,3 sexual intercourse, and parenteral
exposure to blood.4 Infection with HTLV-I is lifelong with an
asymptomatic carrier state lasting 20–50 years5 that progresses
to aggressive conditions: either adult T cell lymphoma/leukemia
(ATL)6 or HTLV-I associated myelopathy/tropical spastic parapare-
sis (HAM/TSP).7 The retrovirus in ATL patients is resistant to con-
ventional chemotherapy,8 and there is presently no cure for
HTLV-I infection. Current treatment regimens for HIV-1 infection
in highly active anti-retroviral therapy (HAART) include HIV-1 pro-
tease inhibitors. HTLV-I protease grossly resembles HIV-1 protease
with both being homodimers that form in a pincer shape with a
flap region and a central active site region where a trapped precur-
sor protein is cleaved into peptides.9 Inhibiting the protease with
an antagonist prevents the processing of precursor proteins and
consequently stops viral replication. Several commercial HIV-1
protease inhibitors such as ritonavir,10 amprenavir, saquinavir,
indinavir, and nelfinavir,11 were shown ineffective against HTLV-I
protease, most likely because HTLV-I protease is larger than
HIV-1 protease (125 vs 99 amino acids per chain), and shares only
an overall 28% sequence identity and 45% sequence identity at
the active site.9 Based on the matrix–capsid cleaving region of
precursor proteins, one of our initial HTLV-I protease inhibitor de-
sign (1) had a hydroxymethylcarbonyl isostere as a central inhibi-
tory unit (‘warhead’) against the protease’s two catalytic aspartic
acid residues (Fig. 1).10 After several structure-based drug design
optimization to improve potency while decreasing molecular size
from octapeptidic inhibitors, we derived an isoleucine laden inhib-
itor (2).12 Further optimization and removal of natural amino acids
led to a small yet very potent inhibitor against HTLV-I protease,
KNI-10635 (3), comprising of four amino acid residues with two
end-capping moieties (Table 1, Fig. 2).13–15 Herein, we describe
our arduous attempt to remove the N-terminal cap moiety which
significantly contributes to inhibitory potency.
HTLV-I protease was shown to display a high degree of specific-
ity for substrates over that of HIV-1 protease.9 Similar to the
substrate study, in a retrospective study of 128 inhibitors, we
observed a high degree of specificity for inhibitors over that of
⇑
Corresponding author. Tel.: +81 75 595 4635; fax: +81 75 591 9900.
E-mail address: kiso@mb.kyoto-phu.ac.jp (Y. Kiso).
Figure 1. The precursor protein sequence was used to design HTLV-I protease
inhibitor 1 that was potency optimized to inhibitor 2.16
0960-894X/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2011.01.048

J.-T. Nguyen et al. / Bioorg. Med. Chem. Lett. 21 (2011) 1832–1837
1833
Table 1
Screening for potent HTLV-I protease inhibition potency in synthesized compounds.
Compounds17
R
HTLV-I protease
HIV-1 protease
inhibition at 50 nM16 (%)
inhibition at 50 nM18 (%)
3
4
5
KNI-10635
93a
69a
43
99
KNI-10680
KNI-10748
>99
92
6
7
8
KNI-10747
KNI-10777
KNI-10751
10
33
8
82
76
22
9
KNI-10750
KNI-10765
4
80
a. 15b
b. 19
a. 70b
b. 32
10
11
12
KNI-10776
KNI-10755
26
28
34
>99
13
KNI-10775
65a
99
(continued on next page)

1834
J.-T. Nguyen et al. / Bioorg. Med. Chem. Lett. 21 (2011) 1832–1837
Table 1 (continued)
Compounds17
R
HTLV-I protease
HIV-1 protease
inhibition at 50 nM16 (%)
inhibition at 50 nM18 (%)
a. 24
b. 11
a. 93
b. 54
14
15
KNI-10817
KNI-10754
KNI-10842
31
97
95
16
86a
17
18
KNI-10814
KNI-10813
19
26
95
96
a. 84a
b. 18
a. >99
b. 68
19
20
KNI-10774
KNI-10843
a. 81a
b. 20
a. >99
b. 80
a
The IC50 values for HTLV-I protease inhibition were determined for compounds 3, 4, 13, 16, 19a, and 20a (Fig. 2).
1:2 mixture with compound 10b.
b
in potency among compounds has made HTLV-I protease inhibitor
design much difficult. As a way to overcome this hurdle, in our re-
cent study in which most inhibitors exhibited similarly potent
HTLV-I protease inhibition, we used statistical analysis to deter-
mine which functional groups were more favored for inhibition,
and settled with inhibitor 3.15 Along with drug optimization,
throughout the years, we also optimized our enzyme assay system
to reduce experimental errors.10,13 About half of the HTLV-I prote-
ase expression protein readily autodigests between Leu40 and Pro41
during isolation and purification.20 In fact, three out of the seven
known cleavage sites in the precursor protein are between a
Leu–Pro amide bond, namely matrix–capsid, (transcription factor
1)–protease and protease–p3. Fortunately, autolysis is preventable
by a Leu40 to Ile40 mutation without affecting enzyme kinetics
parameters. Our current HTLV-I protease inhibition assay method
uses an L40I mutation HTLV-I protease to decrease autoproteolysis,
and a fluorescent substrate to increase detection.21 With our cur-
rent assay method, we observed that within a narrow range of test
inhibitor concentrations between 1 and 100 nM, the percent
HTLV-I protease inhibition appreciably changes from none to near
complete inhibition. On a side note, another research group re-
Figure 2. IC50 values of potent HTLV-I protease inhibitors.16
HIV-1 protease.19 In other words, HTLV-I protease inhibitors gener-
ally also exhibit potent HIV-1 protease inhibition potency, whereas
compounds with potent HIV-1 protease inhibition do not necessar-
ily potently inhibit HTLV-I protease. The HTLV-I and HIV-1 prote-
ase inhibition profiles in the current study with 18 compounds
also support this observation (Table 1). Hence, from a structure-
based drug design perspective for HTLV-I protease inhibitors, in
our experience, any slight modification in the design of the refer-
ence compound tends to drastically affect HTLV-I protease inhibi-
tion profile. HIV-1 protease inhibitors, on the other hand, are more
forgiving to unfavorable changes. This lack of observed gradation
ported a chemically synthesized, autocleavage-resistant [Ile40
,
Ala90, Ala109]HTLV-I protease mutant22 in which the C90A and
C109A mutations improved enzyme stability by preventing intra-
molecular disulfide bond formation without significantly altering
enzyme kinetic properties.20 According to a recent X-ray crystal-
lography study on several of our tetrapeptidic HTLV-I protease
inhibitors in complex with a des-(117–125)-[Ile40]HTLV-I prote-
ase21 and computer-assisted modeling experiments,15 the oxygen

J.-T. Nguyen et al. / Bioorg. Med. Chem. Lett. 21 (2011) 1832–1837
1835
of the P3-cap carbonyl function is involved in hydrogen bond inter-
valine can be accommodated by the S3 subsite. However, com-
pound 18 with a P3 valine residue did not exhibit potent HTLV-I
protease inhibition. The fluorine atom is often used as a bioisostere
of the hydrogen atom that modifies the drug’s hydrophilicity
profile. In obtaining the X-ray crystallography data of inhibitor–
protease complexes in our related compounds, we observed that
the more hydrophobic inhibitors were disfavored for co-crystalli-
zation with the HTLV-I protease mutant.21 A perfluorinated ana-
action with Leu57 while the alkyl group of the P3-cap, as a minor
0
determinant of potency, interacts favorably at the ‘entrance’ (S4
subsite) to the active site. Removing the N-terminal P3-cap moiety
from reference inhibitor 3 was shown to decrease inhibition po-
tency against HTLV-I protease, as exemplified by reference inhibi-
tor KNI-10680 (4, Table 1).
The goal of the current study is to remove the P3-cap moiety in
our HTLV-I protease inhibitor design. The loss of favorable interac-
tions from the P3-cap moiety requires compensation by other
favorable interactions. General observations from substrate speci-
ficity studies on retroviral aspartic proteases23 along with our
study on HTLV-I protease inhibitors12 suggest that distant subsites
logue with
a free amino group may provide an improved
hydrophilicity profile. Although the modification was minimal,
we were pleasantly surprised that perfluorinated inhibitor 19a
(IC50 = 13 nM) was significantly more potent than its reference
compound 4. Extending the N-terminal with an n-butyramide cap-
ping moiety afforded inhibitor 20a (IC50 = 11 nM) that was slightly
more potent or equipotent to free amine inhibitor 19a, and was
somewhat less potent than n-butyramide reference inhibitor 3.
The reactant 3,4,5-trifluorophenylglycine used as the P3 residue
in compounds 19 and 20 is only available commercially as a race-
mic mixture, and thus the exact configuration for the P3 residue in
the inhibitors, respectively, was not ascertained. However, the
HTLV-I protease inhibition profiles of compounds 3–6 and 13,
along with the X-ray diffraction crystallography data of closely re-
lated compounds21 strongly suggest that the more potent isolated
‘a’ fraction in compounds 19 and 20 are likely in an S-configura-
tion. Assuming that the P3 residue is in an S-configuration, compar-
ative computer-assisted modeling experiments24 on inhibitors 3, 4,
19a, and 20a in complex with a truncated L40I mutation HTLV-I
protease were performed using the X-ray crystallography data of
similar compounds (PDB IDs: 3LIN and 3LIV).21 In all inhibitors,
hydrogen bond interactions are present throughout the backbone
anchoring the docked compound with the catalytic Asp32 and
0
from the catalytic S1–S1 subsites are more tolerant to changes in
the substrate or inhibitor. In reference compound 4, the P3 residue
is farthest from the inhibitor’s ‘warhead’ and is thus relatively
more accommodating to modifications. Keeping a phenyl group at-
tached to the alpha carbon atom from the P3 carbonyl as a pharma-
cophore, we investigated several modifications (Table 1, 5–11.) The
higher potency found in compound 5 over compound 6 supported
our X-ray diffraction crystallography data21 that the S-configura-
tion, also found in reference inhibitors 3 and 4, was better oriented
over the R-configuration for interactions with the S3 subsite. Com-
puter-assisted modeling experiments13–15 and X-ray diffraction
crystallography data21 of related compounds indicate that the
amino group in L-phenyglycine in reference inhibitor 4 acts as a
hydrogen bond donor to the protease’s Asp360 side-chain. In consid-
eration that an amino group is a stronger hydrogen bond donor
than a hydroxyl group, the relatively weaker inhibition in hydroxyl
compound 5 than reference amino compound 4 highlights the
importance of this Asp36 interaction. A carbonyl group acts as a
0
Asp32 residues from each respective chain of the homodimeric
0
hydrogen bond acceptor rather than a donor, and thus phen-
ylglyoxylamide 7 which has lower hydrogen bond donating poten-
protease, along with Gly340, Asp360, and Leu57 (Fig. 3). Hydrogen
0
bond interactions with Asp36 and the flap residues Ala59 and Ala59
0
tial than hydroxyl compound
5
predictably exhibited lower
inhibition against HTLV-I protease, although bond geometry differ-
ences also played a role. In compounds 8 and 9 that cannot form
hydrogen bond with Asp360, inhibition potency against HTLV-I pro-
tease was low. Compounds 10 and 11 were designed to interact
are mediated through water. Hydration of the terminal free amino
group in inhibitors 4 and 19a mediates a link with the amide nitro-
57
gen of Leu570. This same Leu amide nitrogen forms a hydrogen
0
bond interaction with the oxygen atom of the P3-cap in inhibitors
with Leu57 in a similar way as the P3-cap carbonyl in reference
3 and 20a. The P3 phenyl group is ideally located and oriented for a
0
inhibitor 3. However, their inhibition against HTLV-I protease
was inadequate. The P3 phenylglycine residue in reference com-
pound 4 was substituted by phenylalanine in compound 12 to ob-
non-covalent binding cation–
observed hydrogen bond and cation–
p
interaction with Arg10. Overall, the
p
interaction patterns are
identical in all four compounds (3, 4, 19a, and 20a) as well as those
found in 3LIN and 3LIV, with differences in Leu57 amide nitrogen
0
serve any change in the reported cation–
p
interaction between the
phenyl group and Arg10’s cationic guanidinyl group,21 and any
appreciative improvement in HTLV-I protease inhibition was not
seen. We reconsidered the significance of the cation–
and opted for a cyclohexyl group. Inhibitor 13 (IC50 = 36 nM, Fig. 2)
with P3 -cyclohexylglycine was unexpectedly equipotent to refer-
p interaction
L
ence inhibitor 4 (IC50 = 37 nM). The IC50 value was determined for
compound 13 and any compound that passed our initial screening
assay, that is, compounds with greater than 50% HTLV-I protease
inhibition at 50 nM. Believing this to be a breakthrough after so
many misses, we examined six-membered ring DL-pipecolic acid
(14) only to be once again disappointed by a low potency profile.
The P3 L-phenylglycine in reference inhibitors 3 and 4 was initially
designed from inhibitor 2 as an isostere of isoleucine (Fig. 1).13
Although the P3 isoleucine derivative (15) of inhibitor 4 exhibited
low potency, the P3 isoleucine analogue (16, IC50 = 6 nM) of inhib-
itor 3 (IC50 = 6 nM) was equipotent to its reference. These observa-
tions indicate that the P3-cap n-butyramide moiety considerably
improves HTLV-I protease inhibition. In the design of reference
inhibitors 3 and 4 from compound 2,
other isostere of isoleucine at the P2 position. In the current study,
the P3 -tert-leucine derivative (17) of reference inhibitor 4 did not
present a desirable HTLV-I protease inhibition profile. According to
the precursor protein sequence near the protease–p3 cleave site,
L-tert-leucine was used as an-
L
Figure 3. Hydrogen bond (black dotted lines) and cation–p (cyan) interactions in a
computer-assisted model of inhibitor 19a (gray), water (red spheres) and the two
chains of HTLV-I protease (purple and green).24

1836
J.-T. Nguyen et al. / Bioorg. Med. Chem. Lett. 21 (2011) 1832–1837
interaction, accordingly. The active site of the protease forms a nar-
row tunnel with hydrogen bond interactions along the inhibitor’s
en the interaction,25 via primarily through-space effects of the sub-
stituents and secondarily resonance, instead of being mainly
polarization dependent.26 Applied to our case, perfluorination of
the P3 phenyl group weakens the cation– interaction. In the P3
p
p
backbone perpendicular to subsite interactions (Fig. 4). The inhib-
itor’s P3 phenyl group is slightly obstructed by Arg10 and Asp36
.
p
0
Our previous work demonstrates that the hydrophobic alkyl
portion of the P3-cap moiety is a minor determinant of HTLV-I pro-
tease inhibition and thus does not contribute greatly to the inhibi-
tion profile.14,15 The current computer-assisted modeling study
indicates that the presence of a P3-cap moiety in the inhibitor gen-
erally greatly increases inhibition against HTLV-I protease due to a
direct hydrogen bond interaction between the P3-cap amide
interaction network, the interaction at the P3 phenyl group with
Arg10 indirectly affects the interaction at the P3 amino group with
10
Asp360, because Arg forms hydrogen bonds with Asp36 (Fig. 3).
0
In free amine compounds, a weak P3 cation–p interaction may
permit sufficient flexibility for the P3 amino group to finely adjust
for a more ideal hydrogen bond interaction with Asp360. Conse-
quently, trifluorophenylglycine inhibitor 19a is more potent than
phenylglycine inhibitor 4, because the hydrogen bond interaction
from the P3 amino group is stronger. Although cyclohexylglycine
oxygen and Leu57 amide nitrogen atoms (Fig. 2). In compounds
0
lacking a P3-cap moiety, the interaction between the compounds’
N-terminal and the Leu57 amide nitrogen atom is mediated by a
inhibitor 13 lacks the cation–p interaction, the compound is less
0
water molecule (Fig. 3). Because water-mediated interactions are
weaker than direct hydrogen bond interactions, P3-capped inhibi-
tors 3, 16 and 20a exhibited more potent inhibition potency
against HTLV-I protease than free amines 4, 15, and 19a,
respectively.
potent than trifluorophenyglycine inhibitor 19a, most likely be-
cause compound 13 is bulkier than compound 19a in regard to
hydrophobic interactions at the S3 subsite, and its bulkiness may
also restrict the flexibility of its P3 amino group. Nevertheless,
cyclohexylglycine inhibitor 13 is as potent as reference phenylgly-
cine inhibitor 4.
Perfluorination of the P3 phenyl group affects the cation–
p
interaction between the aromatic ring and cationic Arg10. Compu-
tational studies indicate that electron withdrawing groups weaken
In the presence of a P3-cap moiety, the P3 amide nitrogen is con-
strained by the adjacent strong (P3-cap oxygen)–(Leu57 nitrogen)
0
the cation–
p
interaction while electron donating groups strength-
hydrogen bond interaction. Consequently, the P3 amide nitrogen
is not as flexible as a free amine nitrogen. As a result, the cation–
p
interaction contributes additively to the inhibitor–protease
interactions, and phenylglycine inhibitor 3 with a P3-cap moiety
is slightly more potent than its perfluorinated derivative (20a).
The relatively higher potency in isoleucine inhibitor 16 over triflu-
orophenylglycine inhibitor 20a may be a result of favored hydro-
phobic interactions with the S3 subsite when a P3-cap moiety is
present. In either case, the IC50 value range between the compara-
tive P3-capped inhibitors 3, 16, and 20a is small (6–11 nM) and we
may be over-interpreting the inhibition profiles, especially in light
that our goal is to discard the P3-cap moiety.
In summary, the purpose of the current study is to remove the
P3-cap moiety from a very potent tetrapeptidic HTLV-I protease
reference inhibitor (3, KNI-10635) and to optimize the P3 phenyl-
glycine moiety for potency by structure-based drug design. Lacking
a P3-cap moiety, the cyclohexylglycine derivative (13, KNI-10775)
exhibits equipotency to the reference phenylglycine inhibitor (4,
KNI-10680). To the best of our knowledge, our most potent inhib-
itor with a 3,4,5-trifluorophenylglycine moiety at the P3 position
(19a, KNI-10774a, IC50 = 13 nM) represents the smallest and most
potent peptidic HTLV-I protease inhibitor in the global scientific
community.
Figure 4. Cross-section top view of computer-assisted model of HTLV-I protease in
complex with inhibitor 19a (black slice plane).24 Subsite interactions with the
inhibitor are perpendicular to hydrogen bond interactions found above (Leu57
,
0
Ala59, and Ala590) and below (Asp360, Gly340, Asp320, Asp32, and Asp36) the plane of the
page. See Figure 3 for a side view.
Figure 5. A typical synthetic route for exemplified compounds 19 and 20.17 Reaction conditions: (a) 1.2 equiv carboxylic acid, 1.3 equiv BOP, DMF, 1.5 equiv Et3N, rt
overnight, and microwave 50 °C 1 h if low reactivity; (b) 4 N HCl in dioxane, rt 30 min; (c) 1.2 equiv EDC, 1.3 equiv HOBt, DMF, rt 30 min, 1.1 equiv carboxylic acid in DMF gtt,
overnight; (d) 1.1 equiv Boc2O, THF, water, 1.5 equiv Et3N, rt overnight.

J.-T. Nguyen et al. / Bioorg. Med. Chem. Lett. 21 (2011) 1832–1837
1837
14. Zhang, M.; Nguyen, J.-T.; Kumada, H. O.; Kimura, T.; Cheng, M.; Hayashi, Y.;
Kiso, Y. Bioorg. Med. Chem. 2008, 16, 5795.
Acknowledgments
15. Zhang, M.; Nguyen, J.-T.; Kumada, H. O.; Kimura, T.; Cheng, M.; Hayashi, Y.;
Kiso, Y. Bioorg. Med. Chem. 2008, 16, 6880.
This study was supported in part by the ‘Academic Frontier’
Project for Private Universities, a matching fund subsidy from the
Ministry of Education, Culture, Sports, Science and Technology
(MEXT), Japan. We thank Mr. T. Hamada for performing HIV-1 pro-
tease inhibition assay.
16. Recombinant L40I mutation HTLV-I protease percent inhibition potency at
50 nM of the test compound was evaluated as single determinations using
previously reported procedures.21 IC50 values were calculated from the
sigmoid plot derived by percent inhibition data at 1, 5, 10, 20, 50 and
100 nM of the test compound, as single determinations, using Synergy
Software’s KaleidaGraph (Supplementary data). The error range for IC50
values was calculated from the root mean square deviation (RMSD) of the
plot, i.e., 50% RMSD inhibition.
Supplementary data
17. The synthesis of reference compounds 3 and 4 was previously described.15
Compounds 5–20 were synthesized by standard solution phase peptide
synthesis by which sequential elongation and coupling of an amine to a
carboxylic acid was performed in DMF with benzotriazol-1-yl-oxy-tris-
(dimethylamino)phosphonium hexafluorophosphate (BOP) as coupling
reagent and Et3N as base (Fig. 5). Peptide coupling to the P1 residue was
Supplementary data (IC50 plot for compound 19a; HPLC profiles
and MS-data for all target compounds) associated with this article
can be found, in the online version, at doi:10.1016/
j.bmcl.2011.01.048.
accomplished
using
1-ethyl-3-(3-dimethylaminopropyl)
carbodiimide
hydrochloride (EDC) and 1-hydroxybenzotriazole (HOBt) as additive, without
a base, to avoid reported side-reactions.27 Attachment of a Boc protection
group to an amino group was achieved with Boc2O and Et3N in THF/water,
while the removal of the Boc protection group was achieved with 4 N HCl in
dioxane. After preparative HPLC purification, all target compounds (3–20) were
>95% pure by analytical HPLC (Supplementary data). The identities of the
compounds were confirmed by TOF MS and ESI-Q MS.
References and notes
1. Nicot, C. Am. J. Hematol. 2005, 78, 232.
2. Proietti, F. A.; Carneiro-Proietti, A. B.; Catalan-Soares, B. C.; Murphy, E. L.
Oncogene 2005, 24, 6058.
3. Ureta-Vidal, A.; Angelin-Duclos, C.; Tortevoye, P.; Murphy, E.; Lepère, J. F.;
Buigues, R. P.; Jolly, N.; Joubert, M.; Carles, G.; Pouliquen, J. F.; de Thé, G.;
Moreau, J. P.; Gessain, A. Int. J. Cancer 1999, 82, 832.
4. Manns, A.; Wilks, R. J.; Murphy, E. L.; Haynes, G.; Figueroa, J. P.; Barnett, M.;
Hanchard, B.; Blattner, W. A. Int. J. Cancer 1992, 51, 886.
5. Nitta, T.; Kanai, M.; Sugihara, E.; Tanaka, M.; Sun, B.; Nagasawa, T.; Sonoda, S.;
Saya, H.; Miwa, M. Cancer Soc. 2006, 97, 836.
6. Takatsuki, K. Retrovirology 2005, 2, 16.
7. Jacobson, S. J. Infect. Dis. 2002, 186, 187.
8. Ravandi, F.; Kantarjian, H.; Jones, D.; Dearden, C.; Keating, M.; O’Brien, S. Cancer
2005, 104, 1808.
9. Kádas, J.; Weber, I. T.; Bagossi, P.; Miklóssy, G.; Boross, P.; Orozslan, S.; Tözsér, J.
J. Biol. Chem. 2004, 279, 27148.
10. Maegawa, H.; Kimura, T.; Arii, Y.; Matsui, Y.; Kasai, S.; Hayashi, Y.; Kiso, Y.
Bioorg. Med. Chem. Lett. 2004, 14, 5925.
11. Bagossi, P.; Kádas, J.; Miklóssy, G.; Boross, P.; Weber, I. T.; Tözsér, J. J. Virol.
Methods 2004, 119, 87.
12. Kimura, T.; Nguyen, J.-T.; Maegawa, H.; Nishiyama, K.; Arii, Y.; Matsui, Y.;
Hayshi, Y.; Kiso, Y. Bioorg. Med. Chem. Lett. 2007, 17, 3276.
13. Nguyen, J.-T.; Zhang, M.; Kumada, H. O.; Itami, A.; Nishiyama, K.; Kimura, T.;
Cheng, M.; Hayashi, Y.; Kiso, Y. Bioorg. Med. Chem. Lett. 2008, 18, 366.
18. Recombinant HIV-1 protease percent inhibition potency at 50 nM of the test
compound was evaluated as single determinations using previously reported
procedures.10
19. Nguyen, J.-T.; Kiso, Y. In Lendeckel, U., Hooper, N. M., Eds.; Viral Proteases and
Antiviral Protease Inhibitor Therapy, Proteases in Biology and Disease; Springer
Science: Dordretch, 2009; Vol. 8, pp 83–100.
20. Louis, J. M.; Oroszlan, S.; Tözsér, J. J. Biol. Chem. 1999, 274, 6660.
21. Satoh, T.; Li, M.; Nguyen, J.-T.; Kiso, Y.; Gustchina, A.; Wlodawer, A. J. Mol. Biol.
2010, 401, 626.
22. Teruya, K.; Kawakami, T.; Akaji, K.; Aimoto, S. Tetrahedron Lett. 2002, 43, 1487.
23. Tözsér, J.; Zahuczky, G.; Bagossi, P.; Copeland, J. M.; Oroszlan, S.; Harrison, R.
W.; Weber, I. T. Eur. J. Biochem. 2000, 267, 6287.
24. Computer-assisted modeling experiments were performed using Molecular
Operating Environment 2009.1001. Visualization of the subsites (Fig. 4) was
assisted by UCSF Chimera 1.4.1.
25. Mecozzi, S.; West, A. P., Jr.; Dougherty, D. A. J. Am. Chem. Soc. 1996, 118, 2307.
26. Wheeler, S. E.; Houk, K. N. J. Am. Chem. Soc. 2009, 131, 3126.
27. Hayashi, Y.; Kinoshita, Y.; Hidaka, K.; Kiso, A.; Uchibori, H.; Kimura, T.; Kiso, Y. J.
Org. Chem. 2001, 66, 5537.