A possible improvement for structure-based drug design illustrated by the discovery of a Tat HIV-1 inhibitor

Article abstract of DOI:10.1016/j.bmcl.2003.12.095

The HIV-1 Tat protein is a promising target for AIDS therapy, due to its extra-cellular roles against the immune system. From the 2D-NMR structure of Tat, we have designed molecules, called TDS, able to bind to Tat and inhibit HIV-1 replication in vitro. This new family of antivirals is composed of a triphenylene aromatic ring substituted with at least one carbon chain bearing a succinimide group. These ligands are prepared from triphenylene or 2,6,10-trimethylphenylene in 3-6 steps depending on the target molecule.

Full text of DOI:10.1016/j.bmcl.2003.12.095

Bioorganic & Medicinal Chemistry Letters 14 (2004) 1543–1546
A possible improvement for structure-based drug design illustrated
by the discovery of a Tat HIV-1 inhibitor
Mickael Montembault,a Giang Vo-Thanh,a Abdallah Deyine,a Valerie Fargeas,a
Monique Villieras,a Ane Adjou,b Didier Dubreuil,a Didier Esquieu,c Catherine Gregoire,c
Sandrine Opi,c Jean-Marie Peloponese,c Grant Campbell,c Jennifer Watkins,c
Jean de Mareuil,c Anne-Marie Aubertin,d Christian Bailly,e Erwann Loretc,*
and Jacques Lebretona,*
a
`
Laboratoire de Synthese Organique, CNRS UMR 6513, Faculte des Sciences, 2 rue de la Houssiniere, BP 92208,
´
`
44322 Nantes Cedex 3, France
´
b
Laboratoire de Chimie Organique Structurale, Faculte des Sciences et des Techniques, 22 BP 582 Abidjan 22,
Ivory Coast
c
´
FRE CNRS Univ. Med./CNRS Faculte de Pharmacie, 27, Bd Jean Moulin, 13385 Marseille Cedex 5, France
´ ´
Pathogenie des Infections Virales Persistantes, INSERM U 74, Institut de Virologie, 3 rue Koberle, 67000 Strasbourg, France
d
eINSERM U 524, Place de Verdun, 59045 Lille, France
Received 1 October 2003; revised 5 December 2003; accepted 24 December 2003
Abstract—The HIV-1 Tat protein is a promising target for AIDS therapy, due to its extra-cellular roles against the immune system.
From the 2D-NMR structure of Tat, we have designed molecules, called TDS, able to bind to Tat and inhibit HIV-1 replication in
vitro. This new family of antivirals is composed of a triphenylene aromatic ring substituted with at least one carbon chain bearing a
succinimide group. These ligands are prepared from triphenylene or 2,6,10-trimethylphenylene in 3–6 steps depending on the target
molecule.
# 2004 Elsevier Ltd. All rights reserved.
A decade ago, structure-based drug design or computer-
aided drug discovery (CADD) raised the hope that the
discovery of new drugs could be accelerated. The reality,
however, is far from the expectation since the rate of
introduction of new chemical entities to the pharma-
ceutical market is currently lower than at any time since
the Second World War.1 Focusing on antiviral drugs
alone, the current repertoire now includes more than 30
specialized products but only one has been obtained
from CADD.2 The problem is that CADD, based on
the use of a single rigid protein structure, faces the
difficulty of the dynamic behavior of a receptor, thus
accommodating protein flexibility in CADD is a major
issue.3 We have developed a new approach that could
contribute to improve CADD by taking into account
the dynamic behavior of a receptor and local structural
changes induced by mutations. The best way to illus-
trate the potential of our approach was to choose a
highly flexible protein as a target, with possible mutants
still having a biological activity. This is the case for the
HIV-1 Tat protein.4
AIDS is due to the loss of CD4 cells resulting from HIV
cytotoxicity but it is also related to the extra-cellular
effect of Tat, which is a toxic protein secreted early by
HIV infected cells.4 Tat was first identified as a regulatory
protein, essential for the HIV viral cycle by its action of
dramatically increasing the HIV gene expression.5
However, the interest to target Tat for AIDS therapy is
due to the discovery of the extra-cellular roles of this
protein.6 Tat acts on different cell lines, such as macro-
phages and CTL, which are essential for the cellular
response of the immune system to eliminate virus-infected
cells.7 The extra-cellular roles of Tat are considered to
Keywords: HIV; Inhibitor; Tat protein.
* Corresponding authors. E-mail: erwann.loret@pharmacie.univ-mrs.
fr; lebreton@chimie.univ-nantes.fr
0960-894X/$ - see front matter # 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2003.12.095

1544
M. Montembault et al. / Bioorg. Med. Chem. Lett. 14 (2004) 1543–1546
be the major reason for the maintenance of HIV infected
cells (or reservoir cells), in spite of the use of antiviral
therapy that efficiently blocks the HIV cycle.7 Conse-
quently, tremendous efforts are currently being made to
develop HIV vaccines targeting Tat.7
tests showed that 30 mM of TDS1 is necessary to inhibit
the HIV-1 cytotoxicity with MT4 cells and cytotoxicity
in human cells appears at 100 mM.12 However, TDS1
can inhibit the Tat TAR interaction (Fig. 2A). The
small TDS molecule prevents binding of the protein to
its 59-bp TAR RNA target, obtained by in vitro tran-
scription.15 It is interesting to note that the Tat TAR
interaction is not inhibited by TDS1 if the complex has
already formed (Fig. 2B). To observe the inhibition of
the Tat–TAR complex, Tat must be pre-incubated with
TDS1 for 30 min before adding TAR. Structural changes
are required for Tat to bind to TAR.16 This experiment
suggests that TDS1 inhibits Tat function by blocking
Tat structural changes. Moreover, it shows that the
TDS1 binding site is preserved in two different Tat
variants, one from Europe (Tat Bru) and one from
Africa (Tat Mal). These two variants have a similar folding
but mutations induce local structural variations.17 The
TDS binding site is not affected by these structural
variations.12
Only two attempts to obtain inhibitors binding on Tat have
been reported during the last decade. A benzodiazepine
derivative (Ro5-3335 and Ro24-7429) was described as
a potential Tat inhibitor,8 but it turned out that this
compound was not binding to Tat but to cycline T, a
cellular cofactor essential for Tat.9 The second study
described a tetrahydropyridine derivative able to bind
to a polyarginine peptide.10 This polyarginine peptide
binds to TAR but there was no evidence that tetra-
hydropyridine could bind to the basic region of Tat.
However, a number of studies describe molecules that
bind to TAR and act as Tat competitors. The most
interesting is a compound recently discovered called
TR87 that inhibits HIV replication at 5 mM.11 These
molecules can inhibit only the Tat–TAR interaction but
have no effect on the other Tat functions such as the
immunosupressive effects.
Synthesis of the TDS0 ligand was accomplished in a
four-step sequence from the commercially available tri-
phenylene 1 as described in Scheme 1. Triphenylene 1
was treated with glutaric acid monomethyl ester chloride
This study describes the synthesis of Tat inhibitors that
we call TDS. TDS compounds were designed from
docking with a Tat NMR structure.12 A full Tat protein
(86 residues) was chemically synthesized with the six
glycine Ca labeled with 13C. This synthetic protein had
an activity similar to a regular Tat variant. Homo-
nuclear and heteronuclear 1H–13C 2D Nuclear Mag-
netic Resonance (NMR) spectra enabled the assignment
of the 86 spin systems and the determination of
915 NMR distance constraints.13 The NMR structure
shows that the basic region and the cysteine-rich region
are well exposed to the solvent while a part of the N-
terminal region and the C-terminal region constitute the
core of Tat. The two regions adjacent to the basic region
are highly folded and play a hinge-like role to allow the
basic region to fit into the Tat nucleotide target TAR.
Target regions were selected on Tat to design TDS.11
The rationale was to design allosteric inhibitors to block
the structural changes required for Tat functions.14
2
18 in the presence of an excess of anhydrous aluminum
TDS1 turned out to be our first lead compound (Fig. 1).
This new family of antivirals is composed of a triphenylene
ring with functional groups, such as a succinimide,
connected to the triphenylene by an aliphatic chain (Fig. 1).
Fluorescence experiments showed that TDS compounds
bind in a hydrophobic pocket in Tat and that the
succinimide is essential for this binding.12 Virological
Figure 2. Inhibition of Tat–TAR binding with TDS1. In the panels A,
the two Tat variants were incubated with TDS1 for 30 min prior to
adding the TAR RNA. In the panels B, the two Tat variants were
incubated with TAR RNA prior to adding TDS1. The letters b and f
indicate respectively a free Tat RNA and a Tat–TAR complex. The 59
nucleotide TAR RNA containing the essential UUU pyrimidine bulge
was prepared by in vitro T3 RNA polymerase transcription. Binding
reaction mixtures (20 mL) contained 0.2 nmol of radiolabeled TAR
RNA, 0–100 ng of Tat in TK buffer (50 mM Tris pH 7.4, 20 mM KCl,
0.1% Triton X-100). Complexes were separated from unbound RNA
by electrophoresis in non-denaturing 8% polyacrylamide gels con-
taining 0.1% Triton X-100. The gels were pre-run for 30 min before
loading the sample (25 mL). The electrophoresis was continued for 90
min at about 200 V. The relative amounts of free (f) and/or bound (b)
RNA were determined by phosphor imaging.
Figure 1. TDS compounds.

M. Montembault et al. / Bioorg. Med. Chem. Lett. 14 (2004) 1543–1546
1545
Scheme 2. Reagents and conditions: (a) BPO, AIBN (cat.), CCl4,
reflux, 2–3 h, 40%; (b) 9 MgCl-(CH2)4-OMgBr, CuLiBr2, THF,
À78 ꢀC to rt, overnight; (c) N-succinimide, PPh3, DEAD, THF, rt,
overnight, 50%.
Scheme 1. Reagents and conditions: (a) 2, AlCl3, CH2Cl2, rt,
overnight, 65%; (b) Zn dust, HgCl2 (cat.), HCl concd, Toluene/H2O,
reflux, 24 h, 70%; (c) LAH, Et2O, 0 ꢀC to rt, 3 h, 95%; (d) N-succini-
mide, PPh3, DEAD, THF, À30 ꢀC to rt, overnight, 95%.
trichloride in dichloromethane to provide the ketoester
3 with 65% yield. In our previous approach, we used
glutaric anhydride instead of 2, but the corresponding
ketoacid 4 was isolated with only 25% yield. The
ketoester 3 was submitted to the Clemmensen reduction,
under typical acidic conditions with zinc dust and cata-
lytic amounts of mercuric chloride, leading to reduction
of the ketone with simultaneous hydrolysis of the
methyl ester to afford the acid 5 with 70% yield. The
acid 5 was reduced with an excess of lithium aluminum
hydride (LAH) to give the primary alcohol 6, which was
converted to the first target molecule TDS0 with 41%
yield under Mitsunobu conditions19 using succinimide
in the presence of diethyl azodicarboxylate (DEAD)
and triphenylphosphine (Scheme 1).
Scheme 3. Reagents and conditions: (a) Na2Cr2O7, H2O, autoclave,
250 ꢀC, 18 h, 88%; (b) LAH, THF, reflux, 12 h, 85%; (c) 2 equiv
TBDMS–Cl, imidazole, DMF, rt, overnight, 46%; (d) NBS, Me2S,
CH2Cl2, À20 ꢀC to 0 ꢀC, 3 h, 89%; (e) 9, CuLiBr2, THF, À10 ꢀC, 50%;
(f) N-succinimide, PPh3, DEAD, THF, À10 ꢀC to rt, 48 h, 80%.
TDS1 and TDS2 were obtained from 2,6,10-trimethyl-
triphenylene20 7 as described in Schemes 2 and 3.
Treatment of 2,6,10-trimethyltriphenylene 7 with a
slight excess of N-bromosuccinimide (NBS) and 10%
benzoylperoxide (BPO) as radical initiators in diluted
CCl4 (0.02 M) under reflux led to the corresponding
monobromomethyl derivative 8, which was isolated
with 40% yield after purification, with 20% recovery of
the starting material 7. Elaboration of the chain was
accomplished using the coupling reaction of the mono-
bromomethyl 8 with an excess (6 equiv) of Grignard
reagent 9,21 in the presence of a catalytic amount of
LiCuBr2, to furnish the desired alcohol 10 with 50%
yield. Finally, TDS1 was synthesized from alcohol 10
using the Mitsunobu procedure previously described for
TDS0 (Scheme 2).
of TDS2, triol 12 was treated with 2 equiv of tert-
butyldimethylsilyl chloride (TBDMSCl) in the presence
of imidazole in DMF. Careful purification of the crude
mixture by chromatography on a silica gel column
resulted in the isolation of the desired disilyl ether 13
with 25% yield, as well as the corresponding monosilyl
ether (25%) and trisilyl ether (25%). These were recycled
after silyl ether cleavage with tetra-n-butylammonium
fluoride (TBAF). The hydroxyl group of 13 was cleanly
converted to the corresponding bromide derivative 14
with a good yield using NBS and dimethylsulfide. Using
the coupling reaction previously described for TDS1,
alcohol 15 was isolated in 55% yield from the bromide
14. TDS2 was finally obtained from 15, after treatment
under Mitsunobu conditions with succinimide and
removal of both TBDMS protecting groups under
standard TBAF cleavage conditions (Scheme 3).
It was anticipated that the last target ligand TDS2 could
be prepared from the triol 12, which would arise from
the reduction of the corresponding triacid 11. Following
a known procedure, 2,6,10-trimethyltriphenylene 7 was
oxidized with aqueous Na2Cr2O7 at 250 ꢀC in an auto-
clave,22 to afford the triacid 11 with 88% yield. After
careful purification,23 the triacid 11 was reduced with an
excess of LAH in refluxing THF to the corresponding
triol 12 in moderate yield. To complete the preparation
The application of this strategy to prepare analogues
with various chain lengths or containing aromatic systems
is currently being explored to improve the biological
activity of these ligands.

1546
M. Montembault et al. / Bioorg. Med. Chem. Lett. 14 (2004) 1543–1546
In conclusion, the great hope that drug design or
CADD could replace random tests in drug discovery
has not yet been fulfilled due to the flexibility of pro-
teins. This study suggests that the discovery of ligands
could be improved by the design of chemical functional
groups (or probes) not rigidified to a central structure
(or scaffold). The rationale is that different possible
orientations of a probe could compensate for the struc-
tural variations of proteins that are due to flexibility or
mutations. We hope that this approach will contribute
to improving CADD by taking into account the
dynamic behavior of a receptor and local structural
changes induced by mutations. Tat was certainly the
best way to illustrate the potential of our approach, due
to its high flexibility.
6. Ensoli, B.; Barillari, G.; Salahuddin, S. Z.; Gallo, R. C.;
Wong-Staal, F. Nature 1990, 345, 84.
7. Gallo, R. C. Proc. Natl. Acad. Sci. U.S.A. 1999, 96, 8224.
8. Hsu, M. C.; Schutt, A. D.; Holly, M.; Slice, L. W.; Sher-
man, M. I.; Richman, D. D.; Potash, M. J.; Volsky, D. J.
Science 1991, 254, 1799.
9. Hsu, M. C.; Dhingra, U.; Earley, J. V.; Holly, M.; Keith,
D.; Nalin, C. M.; Richou, A. R.; Schutt, A. D.; Tam,
S. Y.; Potash, M. J.; Volsky, D. J.; Richman, D. D. Proc.
Natl. Acad. Sci. U.S.A. 1993, 90, 6395.
10. Lapidot, A.; Ben-Asher, E.; Eisenstein, M. FEBS Lett.
1995, 367, 33.
11. Hwang, S.; Tamilarasu, N.; Kibler, K.; Cao, H.; Ali, A.;
Ping, Y. H.; Jeang, K. T.; Rana, T. M. J. Biol. Chem.
2003, 278, 39092.
12. Esquieu, D.; Peloponese, J. M.; Opi, S.; Gregoire, C.; de;
Mareuil, J.; Dunot, J. P.; Sturgis, J.; Witvrouw, M.; Pan-
necouque, C.; Chols, D.; De; Clercq, E.; Montembault,
M.; Vo-Thanh, G.; Villieras, M.; Fargeas, V.; Lebreton,
J.; Loret, E. P. Spectroscopy 2003, 17, 639.
Acknowledgements
13. Peloponese, J. M.; Gregoire, C.; Opi, S.; Esquieu, D.;
Sturgis, J.; Lebrun, E.; Meurs, E.; Collette, Y.; Olive, D.;
Aubertin, A. M.; Witvrow, M.; Pannecouque, C.; De
Clercq, E.; Bailly, C.; Lebreton, J.; Loret, E. P. C. R.
Acad. Sci. III 2000, 323, 883.
14. Gregoire, C.; Loret, E. P. J. Biol. Chem. 1996, 271, 22641.
15. Peloponese, J.-M.; Collette, Y.; Gregoire, C.; Bailly, C.;
Campese, D.; Meurs, E.; Olive, D.; Loret, E. P. J. Biol.
Chem. 1999, 274, 11472.
16. Loret, E. P.; Georgel, P.; Johnson, W. C.; Ho, P. S. Proc.
Natl. Acad. Sci. U.S.A. 1992, 89, 9724.
17. Gregoire, C.; Peloponese, J. M.; Esquieu, D.; Solomiac,
M.; Opi, S.; Lebrun, E.; Lebreton, J.; Loret, E. P. Bio-
polymers 2001, 62, 324.
The authors are deeply thankful to the American Scien-
tific Committee, represented by Drs. David Ho, Michael
Saag and John Bartlett, for having selected the TDS
project for the GlaxoSmithKline (GSK) Drug Dis-
covery and Development Award. E.L.P. and J.L. thank
ANVAR (Agence Nationale de la Valorisation de la
Recherche), ANRS (Agence Nationale de la Recherche
sur le SIDA) and SIDACTION for their financial support.
In addition, the authors are deeply indebted to the CNRS
(Centre National de la Recherche Scientifique) for its con-
stant support of the TDS project. We also would like to
thank Dr. Harald Bock (UPR-CNRS 864, Pessac, France)
for experimental details of the compound 7 procedure.
18. The compound 2 was easily prepared in a two-step
sequence from glutaric anhydride : MeOH/CHCl3 (1/10),
reflux, 5 h. ii) SOCl2, reflux, 2 h.
19. Mitsunobu, O. Synthesis 1981, 1.
20. Shirai, H.; Amano, N.; Hashimoto, Y.; Fukui, E.; Ishii,
T.; Ogawa, M. J. Org. Chem. 1991, 56, 2253.
References and notes
1. Horrobin, D. F. Brit. Med. J. 2001, 222.
2. De Clerck, E. Nature Reviews Drug Discovery 2002, 1, 13.
3. Anderson, A. C.; O’Neil, R. H.; Surti, T. S.; Stroud, R. M.
Chem. Biol. 2001, 5, 445.
21. This reagent was prepared from commercial 4-chloro-1-
butanol, which was first treated with 1 equiv of n-BuMgCl
in THF, previous to adding magnesium metal to form the
Grignard reagent.
4. Jeang, K. T.; Xiao, H.; Rich, E. A. Tat. J. Biol. Chem.
1999, 274, 28827.
5. Fisher, A.; Feinberg, M.; Josephs, S.; Harper, M.;
Marselle, L.; Reyes, G.; Gonda, M.; Aldovini, A.;
Debouk, C.; Gallo, R. C.; Wong-staal, F. Nature 1986,
220, 267.
22. Hassheider, T.; Benning, S. A.; Kitzerow, H.-S.; Achard,
M.-F.; Bock, H. Angew. Chem., Int. Ed. 2001, 40, 2060.
23. Surprisingly, all attempts to reduce crude triacid with a
large excess of LAH in refluxing THF afforded the start-
ing material. The crude triacid should be refluxing in
toluene and washed with hot toluene previous to reduction.