A novel bis-tetrahydrofuranylurethane-containing nonpeptidic protease inhibitor (PI), GRL-98065, is potent against multiple-PI-resistant human immunodeficiency virus in vitro

Article abstract of DOI:10.1128/AAC.01413-06

We designed, synthesized, and identified GRL-98065, a novel nonpeptidic human immunodeficiency virus type 1 (HIV-1) protease inhibitor (P1) containing the structure-based designed privileged cyclic ether-derived nonpeptide P2 ligand, 3(R),3a(S),6a(R)-bis-

Full text of DOI:10.1128/AAC.01413-06

A Novel
Bis-Tetrahydrofuranylurethane-Containing
Nonpeptidic Protease Inhibitor (PI),
GRL-98065, Is Potent against
Multiple-PI-Resistant Human
Immunodeficiency Virus In Vitro
Masayuki Amano, Yasuhiro Koh, Debananda Das, Jianfeng
Li, Sofiya Leschenko, Yuan-Fang Wang, Peter I. Boross,
Irene T. Weber, Arun K. Ghosh and Hiroaki Mitsuya
Antimicrob. Agents Chemother. 2007, 51(6):2143. DOI:
10.1128/AAC.01413-06.
Published Ahead of Print 19 March 2007.
Updated information and services can be found at:
http://aac.asm.org/content/51/6/2143
These include:
REFERENCES
This article cites 33 articles, 10 of which can be accessed free
at: http://aac.asm.org/content/51/6/2143#ref-list-1
CONTENT ALERTS
Receive: RSS Feeds, eTOCs, free email alerts (when new
articles cite this article), more»
Information about commercial reprint orders: http://journals.asm.org/site/misc/reprints.xhtml
To subscribe to to another ASM Journal go to: http://journals.asm.org/site/subscriptions/

ANTIMICROBIAL AGENTS AND CHEMOTHERAPY, June 2007, p. 2143–2155
0066-4804/07/$08.00ϩ0 doi:10.1128/AAC.01413-06
Copyright © 2007, American Society for Microbiology. All Rights Reserved.
Vol. 51, No. 6
A Novel Bis-Tetrahydrofuranylurethane-Containing Nonpeptidic
Protease Inhibitor (PI), GRL-98065, Is Potent against
Multiple-PI-Resistant Human Immunodeficiency
Virus In Vitroᰔ
Masayuki Amano,1,2 Yasuhiro Koh,1,2 Debananda Das,3 Jianfeng Li,4 Sofiya Leschenko,4
Yuan-Fang Wang,5 Peter I. Boross,5,6 Irene T. Weber,5 Arun K. Ghosh,4
and Hiroaki Mitsuya1,2,3
*
Department of Infectious Diseases1 and Department of Hematology,2 Kumamoto University School of Medicine, Kumamoto 860-8556,
Japan; Experimental Retrovirology Section, HIV and AIDS Malignancy Branch, National Cancer Institute, National Institutes of Health,
Bethesda, Maryland 208923; Departments of Chemistry and Medicinal Chemistry, Purdue University, West Lafayette, Indiana 479074;
Department of Biology, Georgia State University, Atlanta, Georgia 303035; and Department of Biochemistry and
Molecular Biology, Faculty of Medicine, University of Debrecen, Debrecen, Hungary6
Received 13 November 2006/Returned for modification 5 January 2007/Accepted 12 March 2007
We designed, synthesized, and identified GRL-98065, a novel nonpeptidic human immunodeficiency virus
type 1 (HIV-1) protease inhibitor (PI) containing the structure-based designed privileged cyclic ether-derived
nonpeptide P2 ligand, 3(R),3a(S),6a(R)-bis-tetrahydrofuranylurethane (bis-THF), and a sulfonamide isostere,
which is highly potent against laboratory HIV-1 strains and primary clinical isolates (50% effective concen-
tration [EC50], 0.0002 to 0.0005 ␮M) with minimal cytotoxicity (50% cytotoxicity, 35.7 ␮M in CD4؉ MT-2
cells). GRL-98065 blocked the infectivity and replication of each of the HIV-1NL4-3 variants exposed to and
selected by up to a 5 ␮M concentration of saquinavir, indinavir, nelfinavir, or ritonavir and a 1 ␮M
concentration of lopinavir or atazanavir (EC50, 0.0015 to 0.0075 ␮M), although it was less active against
HIV-1NL4-3 selected by amprenavir (EC50, 0.032 ␮M). GRL-98065 was also potent against multiple-PI-resistant
clinical HIV-1 variants isolated from patients who had no response to existing antiviral regimens after having
received a variety of antiviral agents, HIV-1 isolates of various subtypes, and HIV-2 isolates examined.
Structural analyses revealed that the close contact of GRL-98065 with the main chain of the protease active-site
amino acids (Asp29 and Asp30) is important for its potency and wide-spectrum activity against multiple-PI-
resistant HIV-1 variants. The present data demonstrate that the privileged nonpeptide P2 ligand, bis-THF, is
critical for the binding of GRL-98065 to the HIV protease substrate binding site and that this scaffold can
confer highly potent antiviral activity against a wide spectrum of HIV isolates.
Highly active antiretroviral therapy (HAART) has had a
major impact on the AIDS epidemic in industrially advanced
nations; however, no eradication of human immunodeficiency
virus type 1 (HIV-1) appears to be currently possible, in part
due to the viral reservoirs remaining in blood and infected
tissues. Moreover, we have encountered a number of chal-
lenges in bringing the optimal benefits of the currently avail-
able therapeutics of AIDS and HIV-1 infection to individuals
receiving HAART (2, 29, 30). They include (i) drug-related
toxicities; (ii) partial restoration of immunologic functions
once individuals have developed AIDS; (iii) development of
various cancers as a consequence of survival prolongation; (iv)
flameup of inflammation in individuals receiving HAART or
immune reconstruction syndrome; and (v) increased cost of
antiviral therapy. Such limitations and flaws of HAART are
exacerbated by the development of drug-resistant HIV-1 vari-
ants (1, 5, 12, 13, 20).
effects by interacting with viral receptors, virally encoded en-
zymes, viral structural components, viral genes, or their tran-
scripts without disturbing the cellular metabolism or function.
However, at present, no antiretroviral drugs or agents are
likely to be completely specific for HIV-1 or to be devoid of
toxicity or side effects in the therapy of AIDS, which has been
a critical issue because patients with AIDS and its related
diseases will have to receive antiretroviral therapy for a long
period of time, perhaps for the rest of their lives. Thus, the
identification of new class of antiretroviral drugs which have a
unique mechanism(s) of action and produce no or minimal
side effects remains an important therapeutic objective.
We have been focusing on the design and synthesis of non-
peptidyl protease inhibitors (PIs) that are potent against
HIV-1 variants resistant to the currently approved PIs. One
such anti-HIV-1 agent, darunavir (DRV)/TMC114, contains a
structure-based designed privileged nonpeptidic P2 ligand,
3(R),3a(S),6a(R)-bis-tetrahydrofuranylurethane (bis-THF) (6,
7, 18). DRV has recently been approved as a therapeutic agent
for the treatment of individuals who harbor multidrug-resistant
HIV-1 variants and do not respond to previously existing
HAART regimens. Incorporation of bis-THF also conferred
on other PIs, including brecanavir/GW640385, potent antiviral
Successful antiviral drugs, in theory, exert their virus-specific
* Corresponding author. Mailing address: Departments of Infec-
tious Diseases and Hematology, Kumamoto University School of Med-
icine, 1-1-1 Honjo, Kumamoto 860-8556, Japan. Phone: (81) 96-373-
5156. Fax: (81) 96-363-5265. E-mail: hm21q@nih.gov.
ᰔ Published ahead of print on 19 March 2007.
2143

2144
AMANO ET AL.
ANTIMICROB. AGENTS CHEMOTHER.
FIG. 1. Structures of GRL-98065, darunavir, and amprenavir.
(Research Triangle Park, NC). Nelfinavir (NFV), indinavir (IDV), and lopinavir
(LPV) were kindly provided by Japan Energy, Inc., Tokyo, Japan. Atazanavir
(ATV) was a kind gift from Bristol Myers Squibb (New York, NY).
activity against a wide spectrum of PI-resistant HIV-1 variants
(9, 10, 23), although clinical development of brecanavir has
been discontinued due to its inherent formulation difficulty.
In the present work, we report the synthesis and biological
properties of a potent nonpeptidic HIV-1 protease inhibitor,
GRL-98065, which also contains bis-THF and a sulfonamide
isostere. GRL-98065 exerts highly potent activity against a
wide spectrum of laboratory HIV-1 strains and primary clinical
isolates, including multiple-PI-resistant variants, with minimal
cytotoxicity. GRL-98065 was also active against HIV-1 isolates
of various subtypes, as well as the HIV-1 isolates examined.
Structural analyses revealed that the close contact (backbone
hydrogen bonding) of GRL-98065 with the main chain of the
protease active-site amino acids (Asp29 and Asp30) is critical
for its potency and wide-spectrum activity against multiple-PI-
resistant HIV-1 variants.
Synthesis of GRL-98065. The synthesis of GRL-98065 is summarized in Fig. 2.
In brief, to a stirred solution of secondary amine 1 (8) (83 mg, 0.25 mmol),
N,N-diisopropylethylamine (65 ␮l, 0.37 mmol) and 4-(N,N-dimethylamino)pyri-
dine (3 mg, 0.03 mmol) in THF (2 ml) at room temperature, sulfonyl chloride 2
(21) (60 mg, 0.27 mmol) in THF (1 ml) was added. The mixture was stirred at
room temperature for 4 h and then concentrated in vacuum, and the residue was
chromatographed with 25% ethyl acetate in hexanes to give sulfonamide 3 (105
mg, 81%) as a white amorphous solid. 1H nuclear magnetic resonance (NMR)
(CDCl3, 500 MHz): ␦, 7.33 to 7.17 (m, 7H), 6.86 (d, 1H, J ϭ 8 Hz), 6.06 (s, 2H),
4.67 (d, 1H, 8 Hz), 3.78 (d, 2H, J ϭ 24 Hz), 3.08 to 3.06 (m, 2H), 3.02 to 2.98 (m,
1H), 2.96 to 2.86 (m, 2H), 2.84 to 2.80 (m, 1H), 1.88 to 1.82 (m, 1H), 1.34 (s, 9H),
0.88 (dd, 6H, J ϭ 6.5 Hz, 15.5 Hz). 13C NMR (CDCl3, 500 MHz): ␦, 189.4, 151.3,
148.1, 137.8, 131.6, 129.4, 128.4, 126.3, 108.2, 107.5, 102.2, 79.6, 72.7, 58.6, 54.6,
53.6, 35.4, 28.1, 27.1, 20.0, 19.8. To a stirred solution of sulfonamide 3 (57 mg,
0.11 mmol) in CH2Cl2 (3 ml), trifluoroacetic acid (1 ml) was added. The resulting
solution was stirred at room temperature for 1 h, and then the solvent was
removed in a vacuum. The residue was dissolved in acetonitrile (2 ml), and
MATERIALS AND METHODS
Cells and viruses. MT-2 and MT-4 cells were grown in RPMI 1640-based
culture medium supplemented with 10% fetal calf serum (PAA Laboratories
GmbH, Linz, Austria) plus 50 U of penicillin and 100 ␮g of kanamycin per ml.
The following HIV strains were used for the drug susceptibility assay: HIV-1LAI
,
HIV-1NL4-3, HIV-2EHO, HIV-2ROD, clinical HIV-1 strains from drug-naive pa-
tients with AIDS (HIV-1ERS104pre) (28), and six HIV-1 clinical isolates that were
originally isolated from patients with AIDS who had received anti-HIV-1 therapy
heavily (for 32 to 83 months) and that were genotypically and phenotypically
characterized as multiple-PI-resistant HIV-1 variants. We also used five HIV-1
isolates of different subtypes: HIV-192UG029 (subtype A; X4), HIV-192UG037
(subtype A; R5), HIV-1Ba-L (subtype B; R5), HIV-197ZA003 (subtype C; R5), and
HIV-192TH019 (subtype E; R5). These five HIV-1 isolates were obtained from the
National Institutes of Health (NIH) AIDS Research and Reference Reagent
Program, Division of AIDS, National Institute of Allergy and Infectious Dis-
eases, NIH. To determine whether each clinical HIV-1 isolate used in the
present study was a syncytium-inducing (X4 virus) or non-syncytium-inducing
(R5 virus) strain, MT-2 cells (105) were exposed to an aliquot of viral stock
supernatant containing 100 50% tissue culture infectious doses (TCID50s) of the
virus and cultured in 12-well culture plates. Cultures were maintained for 4
weeks and were examined under an inverted microscope to determine the syn-
cytium-inducing or non-syncytium-inducing nature of the virus, as described
previously (33).
Antiviral agents. GRL-98065 (Fig. 1), a novel nonpeptidic PI containing bis-
THF, was designed and synthesized by Ghosh and coworkers as described below.
Saquinavir (SQV) and ritonavir (RTV) were kindly provided by Roche Products,
Ltd. (Welwyn Garden City, United Kingdom) and Abbott Laboratories (Abbott
Park, IL), respectively. Amprenavir (APV) was a kind gift from GlaxoSmithKline
FIG. 2. Synthesis of GRL-98065.

VOL. 51, 2007
Drug
BIS-THF-CONTAINING PROTEASE INHIBITOR POTENT AGAINST HIV
2145
TABLE 1. Antiviral activities of GRL-98065 against HIV-1LAI, HIV-2EHO, and HIV-2ROD and cytotoxicitiesa
EC50 (␮M) for:
CC50 (␮M)
Selectivity indexb
HIV-1LAI
HIV-2EHO
HIV-2ROD
GRL-98065
SQV
0.0005 Ϯ 0.0001
0.008 Ϯ 0.001
0.054 Ϯ 0.001
0.048 Ϯ 0.007
0.032 Ϯ 0.004
0.036 Ϯ 0.002
0.007 Ϯ 0.001
0.0048 Ϯ 0.0001
0.0039 Ϯ 0.0009
0.0032 Ϯ 0.0007
0.0030 Ϯ 0.0004
0.21 Ϯ 0.05
0.024 Ϯ 0.005
0.030 Ϯ 0.006
0.25 Ϯ 0.08
0.0026 Ϯ 0.0008
0.005 Ϯ 0.002
0.0080 Ϯ 0.0009
0.0045 Ϯ 0.0004
0.0043 Ϯ 0.0002
0.26 Ϯ 0.01
0.054 Ϯ 0.003
0.240 Ϯ 0.009
0.57 Ϯ 0.01
0.0049 Ϯ 0.0008
0.013 Ϯ 0.006
0.0068 Ϯ 0.0004
35.7
16.4
71,400
2,050
RTV
31.1
580
IDV
69.8
8.1
1,450
250
NFV
APV
Ͼ100
Ͼ100
27.6
Ͼ2,780
Ͼ14,300
5,750
LPV
ATV
DRV
83.1
21,310
a MT-2 cells (2 ϫ 103) were exposed to 100 TCID50s of HIV-1LAI or each HIV-2 isolate and cultured in the presence of various concentrations of each PI, and EC50
s
were determined by using the MTT assay. All assays were conducted in duplicate, and data shown represent mean values (Ϯ1 standard deviation) derived from results
of three independent experiments.
b Each selectivity index denotes a ratio of CC50 to EC50 against HIV-1LAI
.
triethylamine (45 ␮l, 0.32 mmol) and mixed carbonate 4 (8) (30 mg) were added.
The mixture was stirred at room temperature for 4 h and then concentrated in
vacuum. Column chromatography over silica gel with 30% and then 50% of ethyl
acetate in hexanes gave the inhibitor GRL-98065 (5, 51 mg, 82%) as a white
amorphous solid. 1H NMR (CDCl3, 500 MHz): ␦, 7.35 to 7.17 (m, 7H), 6.89 (d,
1H, J ϭ 8.5 Hz), 6.09 (s, 2H), 5.64 (d, 1H, J ϭ 5.5 Hz), 5.05 to 4.97 (m, 2H), 3.97
to 3.94 (m, 1H), 3.89 to 3.83 (m, 3H), 3.72 to 3.67 (m, 2H), 3.16 to 2.95 (m, 4H),
2.93 to 2.87 (m, 1H), 2.83 to 2.79 (m, 2H), 1.86 to 1.81 (m, 1H), 1.68 to 1.55 (m,
1H), 1.50 to 1.42 (m, 1H), 0.94 to 0.88 (dd, 6H, J ϭ 6.5 Hz, 21 Hz). 13C NMR
(CDCl3, 500 MHz): ␦, 155.4, 151.5, 148.3, 137.5, 131.2, 129.3, 129.2, 128.6, 126.5,
123.0, 109.2, 108.3, 107.3, 102.3, 73.0, 72.8, 70.6, 69.6, 58.8, 55.0, 53.7, 45.2, 35.5,
27.1, 25.7, 19.2.
was generally increased two- to threefold. Proviral DNA samples obtained from
the lysates of infected cells were subjected to nucleotide sequencing. This drug
selection procedure was carried out until the drug concentration reached 1 or 5
␮M. In the experiments for selecting drug-resistant variants, MT-4 cells were
exploited as target cells, since HIV-1 in general replicates at greater levels in
MT-4 cells than in MT-2 cells.
Determination of nucleotide sequences. Molecular cloning and determination
of the nucleotide sequences of HIV-1 strains passaged in the presence of anti-
HIV-1 agents were performed as described previously (18). In brief, high-mo-
lecular-weight DNA was extracted from HIV-1-infected MT-4 cells by using the
InstaGene Matrix (Bio-Rad Laboratories, Hercules, CA) and was subjected to
molecular cloning, followed by sequence determination. The primers used for
the first round of PCR with the entire Gag- and protease-encoding regions of the
HIV-1 genome were LTR F1 (5Ј-GAT GCT ACA TAT AAG CAG CTG C-3Ј)
and PR12 (5Ј-CTC GTG ACA AAT TTC TAC TAA TGC-3Ј). The first-round
PCR mixture consisted of 1 ␮l of proviral DNA solution, 10 ␮l of Premix Taq (Ex
Taq version; Takara Bio, Inc., Otsu, Japan), and 10 pmol of each of the first PCR
primers in a total volume of 20 ␮l. The PCR conditions used were an initial 3 min
at 95°C, followed by 30 cycles of 40 s at 95°C, 20 s at 55°C, and 2 min at 72°C, with
a final 10 min of extension at 72°C. The first-round PCR products (1 ␮l) were
used directly in the second round of PCR with primers LTR F2 (5Ј-GAG ACT
CTG GTA ACT AGA GAT C-3Ј) and Ksma2.1 (5Ј-CCA TCC CGG GCT TTA
ATT TTA CTG GTA C-3Ј) under the PCR conditions of an initial 3 min at 95°C,
followed by 30 cycles of 30 s at 95°C, 20 s at 55°C, and 2 min at 72°C, with a final
10 min of extension at 72°C. The second-round PCR products were purified with
spin columns (MicroSpin S-400 HR columns; Amersham Biosciences Corp.,
Piscataway, NJ), cloned directly, and subjected to sequencing with a model 3130
automated DNA sequencer (Applied Biosystems, Foster City, CA).
Determination of replication kinetics of GRL-98065-resistant HIV-1NL4-3 vari-
ant and wild-type HIV-1NL4-3. MT-4 cells (2.4 ϫ 105) were exposed to the
GRL-98065-selected HIV-1 variant at passage 40 (HIV-1GRL98065p40) or wild-
type HIV-1NL4-3 preparation containing 30 ng p24 in six-well culture plates for
3 h, and the MT-4 cells were divided into three fractions, each cultured with or
without GRL-98065 (final concentration of MT-4 cells, 104/ml; drug concentra-
tions, 0, 0.01, and 0.1 ␮M). Amounts of p24 were measured every 2 days for up
to 9 days.
Crystallographic analysis. Recombinant HIV-1 protease was expressed and
purified as described previously (31). GRL-98065 was dissolved in dimethyl
sulfoxide. Crystals were grown by the hanging-drop vapor diffusion method from
4.9 mg/ml protease solution buffered at pH 4.8 with 25 mM sodium acetate in the
presence of 10% (wt/vol) sodium chloride, 6% dioxane, and 10% (vol/vol) di-
methyl sulfoxide. The crystal was mounted in a fiber loop with 20 to 30% (vol/vol)
glycerol as a cryoprotectant. Diffraction data were collected at the National
Synchrotron Light Source, beamline X-26C. The data were processed in the
space group P21212 with unit cell parameters of a ϭ 58.25 Å, b ϭ 85.83 Å, and
c ϭ 45.97 Å by using the HKL2000 program (25). The structure was solved by
molecular replacement with AMoRe (24) using 1FG6, from the Protein Data
Bank, as the starting model. Refinement was carried out using SHELX-97 (27)
and manual adjustment with O (16). Alternate conformations for protease res-
idues, inhibitor, water, and other solvent molecules were modeled when ob-
served, as described previously (31). Anisotropic B factors were applied, and
Drug susceptibility assay. The susceptibilities of HIV-1LAI, HIV-2EHO, and
HIV-2ROD to various drugs and the cytotoxicities of those drugs were deter-
mined by using the MTT assay. Briefly, MT-2 cells (2 ϫ 104/ml) were exposed to
100 TCID50s of HIV-1LAI, HIV-2EHO, or HIV-2ROD in the presence or absence
of various concentrations of drugs in 96-well microculture plates and were incu-
bated at 37°C for 7 days. After 100 ␮l of the medium was removed from each
well, 3-(4,5-dimetylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) solu-
tion (10 ␮l, 7.5 mg/ml in phosphate-buffered saline) was added to each well in the
plate, followed by incubation at 37°C for 4 h. After incubation to dissolve the
formazan crystals, 100 ␮l of acidified isopropanol containing 4% (vol/vol) Triton
X-100 was added to each well, and the optical density was measured in a kinetic
microplate reader (Vmax; Molecular Devices, Sunnyvale, CA). All assays were
performed in duplicate or triplicate. In some experiments, MT-2 cells were
chosen as target cells in the MTT assay, since these cells undergo greater
HIV-1-elicited cytopathic effects than MT-4 cells. To determine the sensitivities
of HIV-1Ba-L, HIV-1ERS104pre, clinical multidrug-resistant HIV-1 isolates, and
different subtypes of HIV-1 isolates to drugs, phytohemagglutinin-activated pe-
ripheral blood mononuclear cells (PHA-PBMs) (106/ml) were exposed to 50
TCID50s of each HIV-1 isolate and cultured in the presence or absence of
various concentrations of drugs in 10-fold serial dilutions in 96-well microtiter
culture plates. To determine the drug susceptibilities of certain laboratory HIV-1
strains (HIV-1NL4-3), MT-4 cells were used as target cells. MT-4 cells (105/ml)
were exposed to 100 TCID50s of wild-type HIV-1NL4-3 and PI-resistant HIV-
1NL4-3 in the presence or absence of various concentrations of drugs and were
incubated at 37°C. On day 7 of culture, the supernatant was harvested and the
amount of p24 Gag protein was determined by using a fully automated chemi-
luminescent enzyme immunoassay system (Lumipulse F; Fujirebio, Inc., Tokyo,
Japan) (22). The drug concentrations that suppressed the production of the p24
Gag protein by 50% (50% effective concentrations [EC50s]) were determined by
comparison with the level of p24 production in drug-free control cell cultures. All
assays were performed in duplicate or triplicate. PHA-PBMs were derived from
a single donor in each independent experiment. Thus, to obtain the data, three
different donors were recruited.
Generation of PI-resistant HIV-1 in vitro. MT-4 cells (105/ml) were exposed to
HIV-1NL4-3 (500 TCID50s) and cultured in the presence of various PIs, each at
an initial concentration of its EC50. Viral replication was monitored by deter-
mination of the amount of p24 Gag produced by MT-4 cells. The culture super-
natants were harvested on day 7 and used to infect fresh MT-4 cells for the next
round of culture in the presence of increasing concentrations of each drug. When
the virus began to propagate in the presence of the drug, the drug concentration

2146
AMANO ET AL.
ANTIMICROB. AGENTS CHEMOTHER.
hydrogen atoms were calculated in the last round of crystallographic refinement
by using SHELXL.
Analysis of GRL-98065 interactions with mutant proteases with molecular
docking. A model was generated from the crystal structure. Hydrogens were
added and optimized, with constraints on heavy atom positions, using the
OPLS2005 force field as implemented in MacroModel, version 9.1. Structural
figures were generated using Maestro, version 7.5. The interactions of GRL-
98065 with six mutant HIV-1 proteases were elucidated with molecular docking
using Glide version 4.0 (Schro¨dinger, LLC, New York, NY). The crystal struc-
tures of these mutant proteases were accessed from the Protein Data Bank
(PDB), and the native ligand was removed. Close interaction in the protease was
annealed, and the docking grid was set up. The conformation of GRL-98065 in
its complex with wild-type protease was taken as the starting ligand conforma-
tion. The conformational flexibility of GRL-98065 when it binds to protease was
taken into account during the docking calculations. The extra-precision mode of
Glide, which has a higher penalty for unphysical interactions, was used (4).
RESULTS
Antiviral activity of GRL-98065 against HIV-1LAI and
HIV-2. We designed and synthesized GRL-98065 and exam-
ined its antiviral activity against a variety of HIV-1 isolates. We
found that GRL-98065 was highly potent in vitro against a
laboratory wild-type HIV-1 strain, HIV-1LAI, compared to
clinically available Food and Drug Administration (FDA)-ap-
proved PIs, with EC50s of ϳ0.0005 ␮M, as examined with the
MTT assay using MT-2 target cells, while its cytotoxicity was
seen only at high concentrations, with 50% cytotoxicities
(CC50s) of 35.7 ␮M and a selectivity index of 71,400 (Table 1).
In contrast, FDA-approved PIs had EC50s ranging from 0.0039
to 0.054 ␮M. The selectivity index of GRL-98065 hence proved
to be very high at 71,400. GRL-98065 was also examined in
comparison with two different strains of HIV-2, HIV-2EHO and
HIV-2ROD. The potency of GRL-98065 against the HIV-2
strains examined was less than that against HIV-1LAI by factors
of 6 to 9; however, its absolute EC50s were comparable to those
of four FDA-approved PIs (SQV, LPV, ATV, and DRV)
which showed similar antiviral potencies against HIV-1LAI and
HIV-2 strains (Table 1).
GRL-98065 is potent against PI-selected laboratory HIV-1
variants. We also examined GRL-98065 against a variety of
HIV-1 variants in vitro selected with each of seven FDA-
approved PIs (SQV, RTV, IDV, NFV, APV, LPV, and ATV).
These variants were selected by propagating HIV-1NL4-3 in the
presence of increasing concentrations of each of these PIs (up
to 1 or 5 ␮M) in MT-4 cells (18), and such variants proved to
have acquired various PI resistance-associated amino acid sub-
stitutions in the protease-encoding region of the viral genome
(Table 2). Each of the variants (HIV-1SQV5␮M, HIV-1RTV5␮M
HIV-1IDV5␮M, HIV-1NFV5␮M, and HIVAPV5␮M), except HIV-
LPV1␮M and HIV-1ATV1␮M, was highly resistant to the PI with
,
1
which the variant was selected and showed significant resis-
tance, with an EC50 of Ͼ1 ␮M (n-fold differences of 29 to 143).
Interestingly, HIV-1LPV1␮M, which was only moderately resis-
tant to LPV, with an EC50 of 0.31 ␮M, was highly resistant to
both RTV and IDV, with an EC50 of Ͼ1 ␮M. HIV-1ATV1␮M
was resistant to ATV, with an EC50 of 0.33 ␮M (79-fold dif-
ference). The activities of GRL-98065 against all of the vari-
ants except HIV-1APV5␮M were found to be relatively well
maintained, with n-fold differences of 5 to 25. It was of note
that even with the 5- to 25-fold differences in the EC50s from
those against wild-type HIV-1NL4-3, EC50s were all Ͻ0.0075
␮M except against HIV-1APV5␮M. GRL-98065 was relatively

VOL. 51, 2007
BIS-THF-CONTAINING PROTEASE INHIBITOR POTENT AGAINST HIV
2147
less potent against HIV-1APV5␮M, with an EC50 of 0.032 ␮M
(107-fold difference), presumably due to the structural resem-
blance between GRL-98065 and APV, both of which contain a
sulfonamide isostere (Fig. 1).
GRL-98065 exerts potent activities against highly PI-resis-
tant clinical HIV-1 strains. In our previous work, we isolated
highly multiple-PI-resistant primary HIV-1 strains, HIV-1MDR/TM
,
HIV-1MDR/MM, HIV-1MDR/JSL, HIV-1MDR/B, HIV-1MDR/C
,
and HIV-1MDR/G, from patients with AIDS who had failed
then-existing anti-HIV regimens after receiving 9 to 11 anti-
HIV-1 drugs over 32 to 83 months (32). These primary strains
contained 9 to 14 amino acid substitutions in the protease-
encoding region which have reportedly been associated with
HIV-1 resistance to various PIs (see footnote a of Table 3).
The EC50s of RTV, IDV, and NFV against clinical multidrug-
resistant HIV-1 strains were mostly Ͼ1 ␮M, and the activities
of the other four PIs (SQV, APV, LPV, and ATV) were also
significantly compromised, as examined in PHA-PBMs with
target cells using p24 production inhibition as an end point.
However, GRL-98065 exerted quite potent antiviral activity,
and its EC50s against those clinical variants were as low as
0.006 ␮M or less (Table 3). The potency of GRL-98065 proved
that the compound was most potent against the six represen-
tative multidrug-resistant clinical HIV-1 variants compared to
the currently available approved PIs, including the two recently
approved PIs LPV and ATV. Generally, GRL-98065 exerted
more-potent antiviral activities against various wild-type
HIV-1 strains, drug-resistance variants, and HIV-2 strains than
DRV by 2- to 10-fold.
GRL-98065 is potent against HIV-1 strains of diverse sub-
types. GRL-98065 was further examined as to whether the
compound exerted antiviral activity against HIV-1 strains of
diverse subtypes in vitro. It was found that GRL-98065 exerted
highly potent activity against HIV-1 isolates of all subtypes
(subtypes A, B, C, and E) examined (Table 3), with EC50s from
0.0002 to 0.0005 ␮M. It is noteworthy that GRL-98065 was
significantly more potent than SQV, LPV, and ATV, whose
EC50s were fairly low compared with other FDA-approved PIs
(RTV, IDV, NFV, and APV) by factors of 8 to 41.5, 11 to 26.5,
and 6.5 to 12, respectively.
In vitro selection of HIV-1 variants resistant to GRL-98065.
We attempted to select HIV-1 variants with GRL-98065 by
propagating a laboratory HIV-1 strain, HIV-1NL4-3, in MT-4
cells in the presence of increasing concentrations of GRL-
98065 as described previously (32). HIV-1NL4-3 was initially
exposed to 0.0005 ␮M GRL-98065 and underwent 40 passages
to be capable of propagating at a 1,000-fold greater concen-
tration (0.5 ␮M). Judging by the amounts of p24 Gag protein
secreted in the culture medium, the replicative capacity of
HIV-1NL4-3 at passage 40 was generally well maintained (ϳ900
ng/ml). We compared whether the emergence of GRL-98065-
resistant HIV-1 is delayed in comparison with the emergence
of resistant HIV-1 upon APV, LPV, or ATV selection (Fig. 3).
HIV-1 variants resistant to APV, LPV, or ATV, which repli-
cated at Ͼ1 ␮M, emerged by passages 25, 30, and 39, respec-
tively, while resistance to GRL-98065 emerged by passage 42,
strongly suggesting that the emergence of GRL-98065-resis-
tant HIV-1 variants was substantially delayed compared to that
for the three PIs tested. Genetic characterization of the pro-
tease-encoding region disclosed that those variants resistant to

2148
AMANO ET AL.
ANTIMICROB. AGENTS CHEMOTHER.
FIG. 3. In vitro selection of PI-resistant HIV-1 variants. HIV-1NL4-3 was propagated in MT-4 cells in the presence of increasing concentrations
of amprenavir (E), lopinavir (F), atazanavir (‚), or GRL-98065 (Œ). Each passage of virus was done in a cell-free fashion.
each of the three PIs had acquired previously reported muta-
tions (Table 3). The protease-encoding region of the proviral
DNA isolated from infected MT-4 cells was cloned and se-
quenced at passages 5, 10, 15, 20, 25, 30, 33, and 40 upon
GRL-98065 selection. Individual protease sequences and their
frequency at each passage are depicted in Fig. 4. By passage 10
(HIV-1GRL98065p10), the wild-type protease gene sequence was
seen in 8 of 13 clones, although one or two sporadic amino acid
substitutions were noted in 5 of the 13 clones. However, by
passage 15 and beyond, the virus acquired the K43I substitu-
tion. As the passage proceeded, more amino acid substitutions
emerged. In HIV-1GRL98065p25, K43I, M46I, V82I, I85V, and
L89M were seen, along with A28S (9 of 20 clones). Val82 is an
active-site amino acid residue whose side chain has direct con-
tacts with inhibitor atoms (33), and the V82I substitution has
been shown to be effective in conferring resistance when
combined with a second active-site mutation, such as V32I
(17). By passage 30, more amino acid substitutions, such as
E21K and E34K, were seen, while the latter was not seen in
HIV-1GRL98065p33. The A28S substitution, which was first seen
in HIV-1GRL98065p20, never became predominant in the later
passages, and the percentage of HIV-1 carrying A28S
remained around 50% (45% in HIV-1GRL98065p25, 60% in
HIV-1GRL98065p30, 36% in HIV-1GRL98065p33, and 64% in
HIV-1 GRL98065p40). It should be noted that as we described
previously (32), the A28S substitution, located at the active site
of the enzyme, was seen early (at passage 15) in HIV-1 selected
in the presence of TMC126, the prototype of GRL-98065, and
this particular mutation never disappeared but was consistently
seen at frequencies of ϳ50%, suggesting that the A28S substi-
tution was critical in conferring resistance to TMC126 (32).
E21K coexisted with A28S by passage 30 and beyond, being
seen in four of six clones at passage 30. The substitution I50V,
seen in HIV-1 resistant to APV, did not coexist with A28S
throughout the passage. This profile was previously seen in the
case of TMC126-selected HIV-1 variants, as described previ-
ously (32). The M46I substitution first emerged at passage 25
and was present in 4 of 10 clones at passage 30 (Fig. 4). Met46
is located on the flap region of the enzyme. The I47V substi-
tution reportedly emerges with viral resistance to APV but was
not seen in GRL-98065-resistant variants. We examined
whether the virus acquired mutations in the Gag region at
passages 5, 10, 15, 20, 25, 30, 33, and 40 of GRL-98065 selec-
tion. It was found that by passage 25, the virus had acquired the
R275K substitution. By passage 33 and beyond, the G412D
substitution emerged and persisted. By passage 40, the p7/p1
cleavage site substitution, I437T, was seen in four of nine
clones (Fig. 5).
GRL-98065-resistant HIV-1NL4-3 variant maintained ro-
bust replicative activity. We determined replication kinetics
of HIV-1GRL98065p40 and wild-type HIV-1NL4-3
.
HIV-
1GRL98065p40 generally propagated well regardless of the pres-
ence of GRL-98065 in culture medium. As shown in Fig. 6,
when HIV-1GRL98065p40 was propagated in MT-4 cells in the
presence or absence of 0.01 or 0.1 ␮M GRL-98065, there was
no marked difference observed in the replication kinetics of
HIV-1GRL98065p40 compared to that of HIV-1NL4-3 without
GRL-98065.
Reduced sensitivities of GRL-98065-selected HIV-1 variants
to various PIs. We also examined the susceptibilities of HIV-
1GRL98065p40 to eight FDA-approved PIs with MT-4 cells (Ta-
ble 2). HIV-1GRL98065p40 was highly resistant to GRL-98065,
with a 600-fold-greater EC50 (0.18 ␮M) relative to the EC50 of
GRL-98065 against HIV-1NL4-3. However, HIV-1GRL98065p40
was still susceptible to SQV and ATV, with relatively low
EC50s of 0.032 ␮M (fivefold difference relative to that for
HIV-1NL4-3) and 0.011 ␮M (threefold difference), respectively.
However, APV was no longer active against HIV-1GRL98065p40
,
with an EC50 of Ͼ1 ␮M. The loss of APV’s activity against

VOL. 51, 2007
BIS-THF-CONTAINING PROTEASE INHIBITOR POTENT AGAINST HIV
2149
FIG. 4. Amino acid sequences of protease-encoding regions of HIV-1NL4-3 variants selected in the presence of GRL-98065. The amino acid
sequence of protease, deduced from the nucleotide sequence of the protease-encoding region of each proviral DNA isolated at each indicated time,
is shown. The amino acid sequence of wild-type HIV-1NL4-3 protease is illustrated at the top as a reference.

2150
AMANO ET AL.
ANTIMICROB. AGENTS CHEMOTHER.
FIG. 5. Amino acid sequences of Gag-encoding regions of HIV-1 variants selected in the presence of GRL-98065. The amino acid sequence
of Gag, deduced from the nucleotide sequence of the Gag-encoding region of each proviral DNA isolated at each indicated time, is shown. The
amino acid sequence of wild-type HIV-1NL4-3 Gag is illustrated at the top as a reference.

VOL. 51, 2007
BIS-THF-CONTAINING PROTEASE INHIBITOR POTENT AGAINST HIV
2151
FIG. 6. Replication kinetics of GRL-98065-resistant HIV-1 variant and HIV-1NL4-3. MT-4 cells (2.4 ϫ 105) were exposed to an HIV-
1GRL98065p40 or wild-type HIV-1NL4-3 preparation containing 30 ng p24 in six-well culture plates for 3 h, and these MT-4 cells were divided into
three fractions, each cultured with or without GRL-98065 (final concentration of MT-4 cells, 104/ml; drug concentrations, 0, 0.01, and 0.1 ␮⌴).
Amounts of p24 were measured every 2 days for up to 9 days.
HIV-1GRL98065p40 was thought to be due to APV’s structural
relatedness to GRL-98065.
(Fig. 7A). The protease formed very similar hydrogen bond
interactions with GRL-98065 and DRV, with a few exceptions.
The equivalent atoms of GRL-98065 and DRV superimpose
with a root mean square deviation of 0.04 Å, excluding the
aniline group of DRV and the 1,3-benzodioxole group of
GRL-98065. The 1,3-benzodioxole group of GRL-98065 and
the aniline group of DRV each formed a hydrogen bond with
Asp30Ј—however, they interact with different atoms of
Asp30Ј. GRL-98065 interacts with the Asp30Ј amide, while the
aniline of DRV interacts with the carbonyl oxygen of Asp30Ј
(Fig. 7B). More significantly, the other oxygen of the 1,3-
benzodioxole group of GRL-98065 formed a water-mediated
interaction with the amide of the flap residue, Gly48Ј, while
DRV had no equivalent interaction with Gly48Ј. These addi-
tional interactions of GRL-98065 with Gly48Ј in the flexible
flap region should stabilize its binding to protease and mimic
the interactions of the peptide substrates more closely than
does DRV.
Crystal structure analysis of HIV-1 protease with GRL-
98065. The crystal structure of HIV-1 protease complexed with
GRL-98065 was refined to a residual factor of 0.147 at a 1.6-Å
resolution in order to determine the molecular basis for the
inhibitor potency. The crystallographic statistics are listed in
Table 4. The inhibitor was found to bind in two overlapping
conformations with equivalent interactions with protease, as
observed for DRV (18, 31). GRL-98065 has hydrogen bond
interactions with the backbone atoms of Asp29, Asp30, Gly27,
and Asp30Ј and with the side chain atoms of Asp25 and Asp25Ј
TABLE 4. Crystallographic data collection and refinement statistics
Value for wild-type
protease
Parametera
Space group ......................................................................
P21212
˚
Unit cell dimensions (A)
Structural analysis of interaction with mutant protease. We
finally attempted to predict the molecular interactions of
GRL-98065 with a variety of mutant proteases by molecular
docking. We first analyzed the hydrogen bond interactions of
various PIs (including DRV) with wild-type protease, employ-
ing previously published coordinates for each PI complexed
with wild-type protease (PDB identifiers, 1S6G, 1HXB,
1HXW, 1SDT, 1OHR, 1HPV, 1MUI, and 2AQU) (Table 5). It
was noted that GRL-98065 and DRV have four hydrogen bond
interactions with backbone atoms of Asp29, Asp30, and
Asp30Ј. None of the other PIs examined have more than two
hydrogen bond interactions with these residues. These results
should corroborate that GRL-98065 and DRV exert highly
potent antiviral activities against wild-type HIV-1 strains.
We next examined six mutant proteases with 2 to 11 amino
acid substitutions (PDB identifiers, 2FDD, 1SGU, 1HSH,
2AZC, 1B6K, and 2AVV). Even though the mutations in each
of these proteases do not exactly match the mutations shown in
Table 2 and Table 3, they cover the range of amino acid
substitutions observed in multidrug-resistant protease. The
a......................................................................................
b .....................................................................................
c......................................................................................
58.25
85.83
45.97
˚
Resolution range (A) ......................................................
50–1.60
Unique reflections............................................................31,128
Rmerge (%) overall (final shell).......................................
ϽIϾ/Ͻ␴(I)Ͼ overall (final shell)...................................
Completeness (%) overall (final shell) .........................
7.1 (38.8)
11.0 (4.7)
99.7 (99.4)
10–1.60
14.7
˚
Data range for refinement (A) ......................................
Rfactor (%) .....................................................................
Rfree (%)........................................................................
20.8
No. of solvent atoms (total occupancies) ..................... 207.7
Root-mean-square deviation from ideal
˚
Bonds (A) .....................................................................
0.009
0.030
˚
Angle distance (A).......................................................
Average B factors (A )
2
˚
Main chain....................................................................
Side chain......................................................................
Inhibitor ........................................................................
15.5
19.0
10.7
28.7
Solvent...........................................................................
Residual density (max/min) (eAϪ3)............................... 0.38/Ϫ0.40
˚
a Rmerge, ⌺͉I Ϫ ϽIϾ͉/⌺(I); Rfactor, ⌺ʈFobs ͉Ϫ ͉Fcalcʈ/⌺͉Fobs͉; Rfree, Rfactor calcu-
lated for 5% reference set; max/min, maximum/minimum.

2152
AMANO ET AL.
ANTIMICROB. AGENTS CHEMOTHER.
FIG. 7. Selected hydrogen bond interactions of GRL-98065 with wild-type HIV-1 protease. (A) The bis-THF group forms hydrogen bond
interactions with backbone atoms of Asp29 and Asp30. There is a hydrogen bond with the backbone atom of Gly27. The hydroxyl group forms
hydrogen bonds with the side chains of the catalytic aspartates. One oxygen of the benzodioxole group forms a hydrogen bond interaction with
Asp30Ј, and the other oxygen of the benzodioxole group forms a water-mediated hydrogen bond interaction with Gly48Ј. (B) Hydrogen bond
interactions between DRV and protease (PDB identifier, 1S6G) are shown. Most interactions between GRL-98065 and DRV are similar, except
for interactions with Asp30Ј and Gly48Ј. GRL-98065 interacts with the Asp30Ј amide, while DRV interacts with the Asp30Ј carbonyl oxygen. The
benzodioxole oxygen of GRL-98065 has a water-mediated interaction with Gly48Ј in the flap. This interaction appears to stabilize the binding site
more for GRL-98065 and may be partly responsible for its greater antiviral potency than that of DRV.
docking calculations show that for these mutant proteases,
GRL-98065 is able to maintain most of the four hydrogen bond
interactions observed for the wild-type protease. In particular,
we observed that the hydrogen bond interaction with Asp29 is
maintained in four out of six mutant proteases. Taken to-
gether, these results are likely to explain why GRL-98065 is
able to show greater potency than other clinically approved
PIs against a wide spectrum of multiple-PI-resistant HIV-1
variants.
DISCUSSION
GRL-98065, which contains a unique component, bis-THF,
and a sulfonamide isostere, suppressed a wide spectrum of
HIV-1, HIV-2, and primary HIV-1 strains of different subtypes
over a very narrow spread of EC50s ranging from 0.0002 to
0.0045 ␮M (Tables 1 and 3). GRL-98065 was highly potent
against a variety of multidrug-resistant clinical HIV-1 isolates,
with EC50s of 0.003 to 0.006 ␮M, while the existing FDA-
approved PIs examined either failed to suppress the replication
of those isolates or required much higher concentrations for
viral inhibition (Table 3). When examined against laboratory
PI-selected HIV-1 variants (except against HIV-1APV5␮M),
GRL-98065 also exerted potent activity, with EC50s ranging
0.0015 to 0.0075 ␮M (Table 2). It is of note that GRL-98065
was less potent against APV-resistant HIV-1APV5␮M; however,
it is thought that this relative cross-resistance is due to the
structural similarities of GRL-98065 with APV. It is intriguing
that the activity of SQV against laboratory PI-selected variants,
except for the SQV-selected variant, was fairly well main-
tained, a profile generally consistent with our results and those
of other groups (11, 32, 33). However, when SQV was exam-
ined against multidrug-resistant primary HIV-1 strains, high
concentrations of SQV were required to suppress the replica-
tion of four of six strains tested (EC50s ranging from 0.14 to
0.29 ␮M) (Table 3). In contrast, GRL-98065 exerted highly
potent activity against all the six primary strains examined. As
shown in Table 3, even the replication of the most PI-resistant
primary strain, HIV-1JSL, against which EC50s of the eight PIs,
including ATV and DRV, were 0.027 to Ͼ1 ␮M, was effec-
tively suppressed by GRL-98065 at a fairly low concentration,
with an EC50 of 0.006 ␮M.
TABLE 5. Hydrogen bond distances of protease inhibitors with
selected active-site residuesa
˚
Hydrogen bond distance(s) (A)
Inhibitor
PDB ID
Asp29
Asp30
Asp30Ј
GRL-98065
DRV
SQV
1.9/2.4
2.3/2.4
1.9
2.4
2.4
2.2
NP
NP
NP
2.6
NP
NP
2.5
2.5
NP
2.2
NP
NP
NP
NP
NP
1S6G
1HXB
1HXW
1SDT
1OHR
1HPV
1MUI
2AQU
RTV
2.1
IDV
2.1
NFV
NP
APV
2.8
LPV
1.7
ATV
1.9
a Hydrogen atoms were added and optimized with constraints on heavy atoms
using the OPLS2005 force field (MacroModel, version 9.1; Schro¨dinger, LLC).
˚
Hydrogen bond tolerances used were as follows: 3.0 A for HOA distance;
DOHOA angle greater than 90°; and HOAOB angle greater than 60°, where H
is the hydrogen atom, A is the acceptor, D is the donor, and B is a neighbor atom
bonded to the acceptor. Values for separate interactions are separated by a shill.
NP (not present) denotes that a hydrogen bond is not present between the
inhibitor and the particular residue. ID, identifier.

VOL. 51, 2007
BIS-THF-CONTAINING PROTEASE INHIBITOR POTENT AGAINST HIV
2153
tion with GRL-98065 (Fig. 4). In GRL-98065-selected HIV-1,
neither of the active-site amino acid substitutions, I84V or V32I,
emerged. These two substitutions are known to confer high levels
of PI resistance on HIV-1, in particular when combined with V82I
(17). The absence of these mutations may contribute to the ob-
served delayed acquisition and relatively low level of resistance to
GRL-98065.
It is also of note that during selection with GRL-98065, the
unique A28S mutation in the active site of the enzyme emerged.
The A28S mutation was seen during the selection of HIV-1NL4-3
with TMC126, where the mutation never became predominant
but persisted within TMC126-selected HIV-1 variants at frequen-
cies of ϳ50%. In a previous biochemical study conducted by
Hong et al. (14), the A28S mutation in HIV protease caused a
more than 1,500-fold decrease in kcat/Km values for peptide sub-
strates. These results suggest that A28S represents a critical mu-
tation for GRL-98065 resistance but also confers a severe repli-
cation disadvantage on the virus. It should be noted, however,
that the population size of HIV-1 in a culture is relatively small
and the appearance of mutations can be affected by stochastic
phenomena, i.e., rates and orders of appearance of mutations. In
order to address the issue of mutation appearance, clinical studies
on GRL-98065 are ultimately needed.
The crystal structure reveals that GRL-98065 has a series
of hydrogen bond interactions with backbone atoms of
Asp29, Asp30, Asp30Ј, and Gly27 of the protease (Fig. 7).
Like other protease inhibitors, GRL98065 also has hydro-
gen bond interactions with the side chain atoms of Asp25
and Asp25Ј. Besides the water-mediated hydrogen bond in-
teractions with Ile50 and Ile50Ј, there is a water-mediated
hydrogen bond interaction with the flap residue Gly48Ј.
Thus, GRL-98065 makes favorable polar interactions with
Asp29 and Asp30 as well as with the flap residues. These
hydrogen bond interactions, besides various favorable van
der Waals contacts, are likely to be responsible for the
strong binding of the inhibitor and its potent antiviral ac-
tivity observed in the present work. Comparison of the crys-
tal structure of HIV-1 protease with that of GRL-98065 and
the crystal structure of the complex with the recently ap-
proved inhibitor DRV shows that the interactions with the
S2 site of the protease are shared by the two PIs, but the
nature of the hydrogen bonds with residues in the S2Ј site
differs (Fig. 7). The water-mediated interaction of GRL-
98065 with flap residue 48Ј is not observed for DRV. These
differences in interactions might be partly responsible for
the low EC50 of GRL-98065 compared to that of DRV
(Tables 1 to 3).
FIG. 8. Interactions between GRL-98065 and wild-type pro-
tease. van der Waals surfaces of GRL-98065 (green), Val82 (red),
and Ile85 (magenta) are shown. There are strong van der Waals
interactions of GRL-98065 with Val82 and Val82Ј. Note that Val82
was replaced with isoleucine as a primary resistance mutation dur-
ing in vitro passage of HIV-1 in the presence of GRL-98065. How-
ever, Ile85 does not have van der Waals contact with the inhibitor,
suggesting that I85V emerged as a secondary mutation during in
vitro selection with the inhibitor.
The observed greater potency of GRL-98065 than those of
the existing FDA-approved PIs examined in the present study
appears to stem, at least in part, from the ability of the two
conformationally constrained ring oxygen atoms in its bis-THF
group to form hydrogen bonds with the main chain amide
hydrogen atoms of Asp29 and Asp30 in the S2 subsite (Fig. 7).
Since the main chain atoms cannot be changed by viral amino
acid substitutions, the interactions of GRL-98065 and the two
catalytic site amino acids are unlikely to be substantially af-
fected, perhaps resulting in GRL-98065’s broad spectrum of
activity against multidrug-resistant variants.
It is noteworthy that Asp30 can be mutated to asparagine
when HIV-1 is exposed to NFV (26). This mutation, D30N, is
a primary resistance mutation for NFV that results in forma-
tion of a hydrogen bond with the side chain of Asp30 (26).
GRL-98065 does not have direct interaction with the side
chain of Asp30 (Fig. 7). Consistent with this observation, ex-
posure of HIV-1 to GRL-98065 did not select mutations at
codon 30, and GRL-98065 was active against D30N-carrying
HIV-1NFV5␮M, which was highly resistant to NFV, with an
EC50 value of Ͼ1 ␮M (Table 2).
In the present HIV-1 selection experiment with GRL-98065, by
passage 30 and beyond, 10 major amino acid substitutions (E21K,
A28S, K43I, M46I, I50V, D60N, A71V, V82I, I85V, and L89M)
were identified. It is noteworthy that mutation of Val82, whose
side chain makes direct contacts with a number of PIs (3), was not
seen in HIV-1 selected with TMC126 that has bis-THF and exerts
potent activity against a wide spectrum of HIV-1 strains (32).
Presumably, V82I arises due to the fact that GRL-98065 has a
tight and direct contact with Val82 (Fig. 8), while Ile85 does not
have van der Waals contact with the inhibitor, suggesting that
I85V emerged as a secondary mutation during the in vitro selec-
We also attempted to gain a structural understanding of why
GRL-98065 is able to maintain a highly favorable potency against
a variety of laboratory PI-resistant HIV-1 variants and multidrug-
resistant clinical isolates. The resistance of PIs due to mutations
arises because of possible loss of direct hydrogen bond interac-
tions with specific residues (e.g., D30N for NFV and G48V for
SQV) or loss of van der Waals contact (e.g., with V82A and I84V
for first-generation PIs). Analysis of mutant protease crystal struc-
tures in comparison with that of wild-type protease showed that
the backbone atoms of mutant protease undergo minimal con-
formational changes on mutation (9, 15, 19). The loss of binding
in many cases seems to be due to loss of weaker van der Waals
contacts between the inhibitor and the protease. We hypothesize

2154
AMANO ET AL.
ANTIMICROB. AGENTS CHEMOTHER.
6. Ghosh, A. K., J. F. Kincaid, W. Cho, D. E. Walters, K. Krishnan, K. A.
Hussain, Y. Koo, H. Cho, C. Rudall, L. Holland, and J. Buthod. 1998. Potent
HIV protease inhibitors incorporating high-affinity P2-ligands and (R)-(hy-
droxyethylamino)sulfonamide isostere. Bioorg. Med. Chem. Lett. 8:687–690.
7. Ghosh, A. K., K. Krishnan, D. E. Walters, W. Cho, H. Cho, Y. Koo, J.
Trevino, L. Holland, and J. Buthod. 1998. Structure based design: novel
spirocyclic ethers as nonpeptidal P2-ligands for HIV protease inhibitors.
Bioorg. Med. Chem. Lett. 8:979–982.
8. Ghosh, A. K., S. Leshchenko, and M. Noetzel. 2004. Stereoselective photo-
chemical 1,3-dioxolane addition to 5-alkoxymethyl-2(5H)-furanone: synthe-
sis of bis-tetrahydrofuranyl ligand for HIV protease inhibitor UIC-94017
(TMC-114). J. Org. Chem. 69:7822–7829.
9. Ghosh, A. K., G. Schiltz, R. S. Perali, S. Leshchenko, S. Kay, D. E. Walters,
Y. Koh, K. Maeda, and H. Mitsuya. 2006. Design and synthesis of novel
HIV-1 protease inhibitors incorporating oxyindoles as the P2Ј-ligands.
Bioorg. Med. Chem. Lett. 16:1869–1873.
10. Ghosh, A. K., P. R. Sridhar, S. Leshchenko, A. K. Hussain, J. Li, A. Y.
Kovalevsky, D. E. Walters, J. E. Wedekind, V. Grum-Tokars, D. Das, Y. Koh,
K. Maeda, H. Gatanaga, I. T. Weber, and H. Mitsuya. 2006. Structure-based
design of novel HIV-1 protease inhibitors to combat drug resistance. J. Med.
Chem. 49:5252–5261.
that if an inhibitor maintains strong hydrogen bond interactions
with the wild-type protease, particularly with backbone atoms of
multiple residues that are conserved (e.g., Asp29 and Gly27), then
the loss of van der Waals contacts due to mutations may not result
in a drastic loss of binding affinity. Thus, inhibitors without mul-
tiple strong hydrogen bond interactions with wild-type protease
would be more susceptible to loss of binding due to loss of weaker
van der Waals contacts than inhibitors with multiple hydrogen
bond interactions. In this respect, we analyzed the hydrogen bond
interactions of several PIs with wild-type protease (Table 5). It is
noteworthy that only GRL-98065 and DRV have four hydrogen
bond interactions with backbone atoms of Asp29 and Asp30 and
of Asp30Ј. None of the other clinically approved PIs studied here
have more than two hydrogen bond interactions with these resi-
dues. Thus, GRL-98065 is likely to preserve the hydrogen bond
interactions and bind tightly with mutant protease.
The present data suggest that GRL-98065 has several ad-
vantages: (i) it exerts potent activity against a wide spectrum of
drug-resistant HIV-1 variants, presumably due to its interac-
tions with the main chains of the active-site amino acids Asp29
and Asp30; (ii) its unique contact with HIV-1 protease differs
from that of other PIs; (iii) the viral acquisition of resistance is
substantially delayed; and (iv) at least several PIs, including
SQV and ATV, remain active in vitro against the virus selected
in vitro with GRL-98065. It is of note that GRL-98065 pos-
sesses substantially favorable features as a potential therapeu-
tic for AIDS, as described above; however, its oral bioavail-
ability, pharmacokinetics/pharmacodynamics, biodistribution,
etc., are yet to be determined in further rigorous preclinical
and clinical testing.
11. Gong, Y. F., B. S. Robinson, R. E. Rose, C. Deminie, T. P. Spicer, D. Stock,
R. J. Colonno, and P. F. Lin. 2000. In vitro resistance profile of the human
immunodeficiency virus type 1 protease inhibitor BMS-232632. Antimicrob.
Agents Chemother. 44:2319–2326.
12. Grabar, S., L. Weiss, and D. Costagliola. 2006. HIV infection in older
patients in the HAART era. J. Antimicrob. Chemother. 57:4–7.
13. Hirsch, H. H., G. Kaufmann, P. Sendi, and M. Battegay. 2004. Immune
reconstitution in HIV-infected patients. Clin. Infect. Dis. 38:1159–1166.
14. Hong, L., J. A. Hartsuck, S. Foundling, J. Ermolieff, and J. Tang. 1998.
Active-site mobility in human immunodeficiency virus, type 1, protease as
demonstrated by crystal structure of A28S mutant. Protein Sci. 7:300–305.
15. Hong, L., X. C. Zhang, J. A. Hartsuck, and J. Tang. 2000. Crystal structure
of an in vivo HIV-1 protease mutant in complex with saquinavir: insights into
the mechanisms of drug resistance. Protein Sci. 9:1898–1904.
16. Jones, T. A., J. Y. Zou, S. W. Cowan, and Kjeldgaard. 1991. Improved
methods for building protein models in electron density maps and the loca-
tion of errors in these models. Acta Crystallogr. A 47:110–119.
17. Kaplan, A. H., S. F. Michael, R. S. Wehbie, M. F. Knigge, D. A. Paul, L.
Everitt, D. J. Kempf, D. W. Norbeck, J. W. Erickson, and R. Swanstrom.
1994. Selection of multiple human immunodeficiency virus type 1 variants
that encode viral proteases with decreased sensitivity to an inhibitor of the
viral protease. Proc. Natl. Acad. Sci. USA 91:5597–5601.
ACKNOWLEDGMENTS
18. Koh, Y., H. Nakata, K. Maeda, H. Ogata, G. Bilcer, T. Devasamudram, J. F.
Kincaid, P. Boross, Y. F. Wang, Y. Tie, P. Volarath, L. Gaddis, R. W.
Harrison, I. T. Weber, A. K. Ghosh, and H. Mitsuya. 2003. Novel bis-
tetrahydrofuranylurethane-containing nonpeptidic protease inhibitor (PI)
UIC-94017 (TMC114) with potent activity against multi-PI-resistant human
immunodeficiency virus in vitro. Antimicrob. Agents Chemother. 47:3123–
3129.
19. Kovalevsky, A. Y., Y. Tie, F. Liu, P. I. Boross, Y. F. Wang, S. Leshchenko,
A. K. Ghosh, R. W. Harrison, and I. T. Weber. 2006. Effectiveness of
nonpeptide clinical inhibitor TMC-114 on HIV-1 protease with highly
drug resistant mutations D30N, I50V, and L90M. J. Med. Chem. 49:1379–
1387.
20. Little, S. J., S. Holte, J. P. Routy, E. S. Daar, M. Markowitz, A. C. Collier,
R. A. Koup, J. W. Mellors, E. Connick, B. Conway, M. Kilby, L. Wang, J. M.
Whitcomb, N. S. Hellmann, and D. D. Richman. 2002. Antiretroviral-drug
resistance among patients recently infected with HIV. N. Engl. J. Med.
347:385–394.
21. Lube, A., W. P. Neumann, and M. A. Niestroj. 1995. New and regioselective
method for the synthesis of aromatic, heteroaromatic, and olefininc sulfon-
amides by electrophilic destannylation. Chem. Ber. 128:1195–1198.
22. Maeda, K., K. Yoshimura, S. Shibayama, H. Habashita, H. Tada, K. Sagawa,
T. Miyakawa, M. Aoki, D. Fukushima, and H. Mitsuya. 2001. Novel low
molecular weight spirodiketopiperazine derivatives potently inhibit R5
HIV-1 infection through their antagonistic effects on CCR5. J. Biol. Chem.
276:35194–35200.
We thank the Center for Information Technology, National Insti-
tutes of Health, for providing computational resources.
This work was supported in part by the Intramural Research Program
of the Center for Cancer Research, National Cancer Institute, National
Institutes of Health grants GM62920 and GM53386, and in part by a
Grant-in-aid for Scientific Research (Priority Areas) from the Ministry of
Education, Culture, Sports, Science, and Technology of Japan (Monbu-
Kagakusho), a Grant for Promotion of AIDS Research from the Ministry
of Health, Welfare, and Labor of Japan (Kosei-Rohdosho; H15-AIDS-
001), and a grant to the Cooperative Research Project on Clinical and
Epidemiological Studies of Emerging and Re-emerging Infectious
Diseases (Renkei Jigyo; no. 78, Kumamoto University) of Monbu-
Kagakusho, the Georgia State University Molecular Basis of Disease
Program, the Georgia Research Alliance, the Georgia Cancer Coalition,
and National Institute of Health grants GM62920 and GM53386. The
X-ray diffraction data were collected at beamline X-26C, National Syn-
chrotron Light Source. Use of the National Synchrotron Light Source,
Brookhaven National Laboratory, was supported by the U.S. Department
of Energy, Office of Science, Office of Basic Energy Sciences, under
contract no. DE-AC02-98CH10886.
REFERENCES
23. Miller, J. F., C. W. Andrews, M. Brieger, E. S. Furfine, M. R. Hale, M. H.
Hanlon, R. J. Hazen, I. Kaldor, E. W. McLean, D. Reynolds, D. M.
Sammond, A. Spaltenstein, R. Tung, E. M. Turner, R. X. Xu, and R. G.
Sherrill. 2006. Ultra-potent P1 modified arylsulfonamide HIV protease in-
hibitors: the discovery of GW0385. Bioorg. Med. Chem. Lett. 16:1788–1794.
24. Navaza, J. 1994. AMoRe: an automated package for molecular replacement.
Acta Crystallogr. A 50:157–163.
1. Carr, A. 2003. Toxicity of antiretroviral therapy and implications for drug
development. Nat. Rev. Drug Discov. 2:624–634.
2. De Clercq, E. 2002. Strategies in the design of antiviral drugs. Nat. Rev.
Drug Discov. 1:13–25.
3. Erickson, J. W., and S. K. Burt. 1996. Structural mechanisms of HIV drug
resistance. Annu. Rev. Pharmacol. Toxicol. 36:545–571.
4. Friesner, R. A., J. L. Banks, R. B. Murphy, T. A. Halgren, J. J. Klicic, D. T.
Mainz, M. P. Repasky, E. H. Knoll, M. Shelley, J. K. Perry, D. E. Shaw, P.
Francis, and P. S. Shenkin. 2004. Glide: a new approach for rapid, accurate
docking and scoring. 1. Method and assessment of docking accuracy. J. Med.
Chem. 47:1739–1749.
25. Otwinowski, Z., and W. Minor. 1997. Processing of X-ray diffraction data in
oscillation mode. Methods Enzymol. 276:307–326.
26. Patick, A. K., H. Mo, M. Markowitz, K. Appelt, B. Wu, L. Musick, V. Kalish,
S. Kaldor, S. Reich, D. Ho, and S. Webber. 1996. Antiviral and resistance
studies of AG1343, an orally bioavailable inhibitor of human immunodefi-
ciency virus protease. Antimicrob. Agents Chemother. 40:292–297.
5. Fumero, E., and D. Podzamczer. 2003. New patterns of HIV-1 resistance
during HAART. Clin. Microbiol. Infect. 9:1077–1084.

VOL. 51, 2007
BIS-THF-CONTAINING PROTEASE INHIBITOR POTENT AGAINST HIV
2155
27. Sheldrick, G. M., and T. R. Schneider. 1997. SHELXL: high resolution
refinement. Methods Enzymol. 277:319–343.
28. Shirasaka, T., M. F. Kavlick, T. Ueno, W. Y. Gao, E. Kojima, M. L. Alcaide,
S. Chokekijchai, B. M. Roy, E. Arnold, R. Yarchoan, et al. 1995. Emergence
of human immunodeficiency virus type 1 variants with resistance to multiple
dideoxynucleosides in patients receiving therapy with dideoxynucleosides.
Proc. Natl. Acad. Sci. USA 92:2398–2402.
29. Siliciano, J. D., R. F. Siliciano, J. N. Blankson, D. Persaud, R. F. Siliciano,
T. Pierson, J. McArthur, R. F. Siliciano, and R. Siliciano. 2004. A long-term
latent reservoir for HIV-1: discovery and clinical implications. J. Antimicrob.
Chemother. 54:6–9.
A. K. Ghosh, J. M. Louis, R. W. Harrison, and I. T. Weber. 2004. High
resolution crystal structures of HIV-1 protease with a potent non-peptide
inhibitor (UIC-94017) active against multi-drug-resistant clinical strains. J.
Mol. Biol. 338:341–352.
32. Yoshimura, K., R. Kato, M. F. Kavlick, A. Nguyen, V. Maroun, K. Maeda,
K. A. Hussain, A. K. Ghosh, S. V. Gulnik, J. W. Erickson, and H. Mitsuya.
2002. A potent human immunodeficiency virus type 1 protease inhibitor,
UIC-94003 (TMC-126), and selection of a novel (A28S) mutation in the
protease active site. J. Virol. 76:1349–1358.
33. Yoshimura, K., R. Kato, K. Yusa, M. F. Kavlick, V. Maroun, A. Nguyen,
T. Mimoto, T. Ueno, M. Shintani, J. Falloon, H. Masur, H. Hayashi, J.
Erickson, and H. Mitsuya. 1999. JE-2147: a dipeptide protease inhibitor
(PI) that potently inhibits multi-PI-resistant HIV-1. Proc. Natl. Acad. Sci.
USA 96:8675–8680.
30. Simon, V., and D. D. Ho. 2003. HIV-1 dynamics in vivo: implications for
therapy. Nat. Rev. Microbiol. 1:181–190.
31. Tie, Y., P. I. Boross, Y. F. Wang, L. Gaddis, A. K. Hussain, S. Leshchenko,