Synthesis and assay of isoquinoline derivatives as HIV-1 Tat-TAR interaction inhibitors

Article abstract of DOI:10.1016/j.bmcl.2005.01.068

Four new isoquinoline derivatives bearing guanidinium group or amino group-terminated side chain were synthesized to target the HIV-1 TAR element. Their abilities to bind TAR RNA and inhibit Tat-TAR RNA interaction were determined by CE analysis, a Tat-dependent HIV-1 LTR-driven CAT assay and SIV-induced syncytium evaluation.

Full text of DOI:10.1016/j.bmcl.2005.01.068

Bioorganic & Medicinal Chemistry Letters 15 (2005) 3978–3981
Synthesis and assay of isoquinoline derivatives as HIV-1
Tat–TAR interaction inhibitors
Meizi He, Dekai Yuan, Wei Lin, Ruifang Pang, Xiaolin Yu and Ming Yang*
National Research Laboratory of Natural and Biomimertic Drugs, Peking University, Beijing 100083, China
Received 26 November 2004; revised 25 January 2005; accepted 27 January 2005
Available online 20 July 2005
Abstract—Four new isoquinoline derivatives bearing guanidinium group or amino group-terminated side chain were synthesized to
target the HIV-1 TAR element. Their abilities to bind TAR RNA and inhibit Tat–TAR RNA interaction were determined by CE
analysis, a Tat-dependent HIV-1 LTR-driven CAT assay and SIV-induced syncytium evaluation.
ꢀ 2005 Elsevier Ltd. All rights reserved.
The early phase of the HIV-1 replication is controlled by
the interaction of trans-activator of transcription (Tat)
protein with the trans-activation responsive region
(TAR) RNA.1,2 A three-nucleotide bulge (U23, C24,
and U25) of HIV-1 TAR is essential for high affinity
and specific binding of Tat protein.3 The Arginine resi-
due at position 52 of Tat is the only sequence-specific
contact mediating the complex formation between Tat
and TAR, and its guanidine group interacts with the
residue U23 of the trinucleotide bulge.4–6 On the basis
of the structural information from HIV-1 Tat–TAR
interaction as well as the design principle put forward
previously,7 we defined requirements for potential low-
molecular weight inhibitor of TAR recognition by the
Tat protein. A series of isoquinoline derivatives with
fused aromatic moiety bearing guanidinium group or
amino group-terminated side chain were designed and
synthesized under the molecular modeling direction.
Their bioactivities were evaluated by CE, a Tat-depen-
dent HIV-1 LTR-driven CAT assay and SIV-induced
syncytium evaluation.
Isoquinoline derivatives 3 and 4 were synthesized
according to the procedure described previously.8,9
Compounds I1 and I2 were obtained by isoquinoline
methyl ester reacted with ethylenediamine and 1,3-prop-
anediamine, respectively. Finally, coupling of isoquino-
line carboxamide derivatives with S-methylisothiourea
yielded compounds IG1 and IG2. Compounds I1, I2,
IG1 and IG2 were characterized by MS and 1H NMR.10
Capillary electrophoresis (CE) was used in the analysis
of RNA–protein interactions, which provided a quick,
sensitive, and precise method to study the binding of
Tat–TAR RNA and drug-TAR RNA.11–14 Herein, we
used CE to study the four compoundsÕ binding specific-
ity of TAR RNA and to determine their binding
constants.15 The result of IG2–TAR interaction is pro-
vided here as an example.
TAR, Tat, and IG2 migrated at about 10 min, 6 min,
and 3 min, respectively. As shown in Table 1, the bind-
ing of compound IG2 with TAR RNA was observed
with the decrease of the area of free compoundÕs peak
in the presence of TAR. Then the inhibition of com-
pound IG2 on the binding of Tat–TAR complex was
probed. Figure 1h showed the height and area of TatÕs
peak was greater than those of Figure 1c, which implied
that compound IG2 could not only bind to TAR but
also hinder the Tat–TAR interaction. The CE analysis
of other compounds provided similar results as well.
The route used for the preparation of N-(2-amino-
ethyl)-isoquinoline-3-carboxamide, N-(3-aminopropyl)-
isoquinoline-3-carboxamide, N-(2-guanidinoethyl)-iso-
quinoline-3-carboxamide bisulfite, and N-(3-guanidino-
propyl)-isoquinoline-3-carboxamide bisulfite (I1, I2,
IG1, and IG2) was carried out as outlined in Scheme 1.
We assume for simplicity that TAR RNA has a single
site for Tat, then define a simplified reaction model for
the TAR–Tat system:
Keywords: Tat–TAR interaction; Isoquinoline; CE.
*
Corresponding author. Tel.: +86 10 82801569; fax: +86 108
2802062; e-mail: yangm@bjmu.edu.cn
0960-894X/$ - see front matter ꢀ 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2005.01.068

M. He et al. / Bioorg. Med. Chem. Lett. 15 (2005) 3978–3981
3979
CO2Me
CO2H
CO2H
c
a
b
NH
NH
d
NH2
3
2
1
CONH n NH2
CO2Me
g
e
f
I1
I2
n=2
n=3
N
N
4
NH
CONH nN
H
H2SO3
NH2
IG1
IG2
n=2
n=3
N
Scheme 1. Reagents and conditions: (a) HCHO, HCl, reflux, 3 h; (b) 10% NH3, pH 7; (c) SOCl2, MeOH, 3 h; (d) saturated NaHCO3 solution, pH 8;
(e) S, xylene, reflux, 48 h; (f) NH2(CH2)nNH2, CHCl3; (g) EtOH, 35 ꢁC, 4 h.
Table 1. CE analysis of IG2–TAR interaction
1
a
Peaks
Migration time (min)
Peak area
2
1
2
9.750
6.017
6.022
8.583
2.498
2.352
7.898
2.223
9.300
2.982
12.777
2.453
6.422
10.026
1496436.501
174852.453
30264.600
117549.797
53183.648
30466.404
112402.95
23299.400
164893.842
22429.500
36775.083
30264.600
575355.408
100301.579
b
c
3
3
4
4
5
5
6
7
8
9
d
10
11
12
13
14
6
7
e
f
9
8
R þ P ¼ RP Kp ¼ ½PRꢀ=½Rꢀ½Pꢀ ¼ cP=½Pꢀ;
ð1Þ
11
10
where [PR], [R], and [P] is the concentration of TAR–
Tat complex, nonbonding TAR, and nonbonding Tat
in the mixture.
g
13
14
12
By the law of mass action, the molar concentration of all
TAR–Tat complexes ([PR]) is equal to the concentration
of bound Tat peptide and the net concentration of
bound Tat peptide is equal to the total Tat peptide
concentration (CP) less the concentration of free Tat
([P]) remaining at equilibrium. The TAR binding ratio
cP is calculated as Eq. 2.
h
9
8
11
12 13 14 15
0
4
5
6
7
10
1
2
3
Migration time (min)
Figure 1. CE elctrophorograms of IG2–TAR interaction. (a) TAR
alone (300 lM); (b) Tat alone (300 lM); (c) Tat + TAR (300 lM, each)
incubated for 30 min; (d) IG2 alone (300 lM); (e), (f) and (g)
IG2 + TAR (300 lM, each), incubated for 15 min, 30 min and
45 min, respectively.
cP ¼ ½PRꢀ=½Rꢀ ¼ ðCP ꢁ ½PꢀÞ=½Rꢀ:
ð2Þ
The concentration of free TAR RNA ([R]) in the equi-
librium mixture was set equal to the total concentration
of TAR (CR) added to the mixture minus the amount of
bound RNA in the complex (equal to [PR]). To estimate
the binding constant of TAR–Tat complex using our CE
method, we developed Eq. 3.
binding constant K were calculated and summarized in
Table 2.
As seen in Table 2, compounds bearing guanidinium
group-terminated side chain had stronger affinities than
those with terminal amino group, thus the guanidinium
group played an important role in the interaction with
TAR. Furthermore, the length of the side chain also
affected their affinities, but the effect did not appear so
obviously as that of guanidinium group.
cP ¼ ðCP ꢁ ½PꢀÞ=ðCR ꢁ CP þ ½PꢀÞ:
ð3Þ
In our method, linear equations by which concentration
of free Tat or the compounds in the mixture could be
calculated were obtained.16 Thus, the ratio cP and the

3980
M. He et al. / Bioorg. Med. Chem. Lett. 15 (2005) 3978–3981
Table 2. Binding constants of the compounds and Tat with TAR
Compounds
Linear equation
R2
Peak area
Concentration (·10ꢁ4 M)
K
Tat
I1
A = 22184c + 3759.9
A = 37074c ꢁ 1645.1
A = 18417c + 3577.5
A = 48670c ꢁ 2215.4
A = 78175c ꢁ 7385.9
0.9881
0.9824
0.9977
0.9822
0.9783
30264.600
28515.699
23958.539
15154.543
23299.400
1.195
0.814
1.107
0.357
0.393
1.26 · 104
3.30 · 104
1.54 · 104
2.07 · 105
1.69 · 105
I2
IG1
IG2
Table 3. Computer simulation of the interaction between the title
compounds and TAR RNA
a
Compounds E1 (kcal/mol) E2 (kcal/mol) H-bond number
b
I1
93.8721
79.2974
15.2278
ꢁ4.20675
ꢁ1.70263
ꢁ4.37024
ꢁ6.40564
ꢁ47.2764
2
I2
N
N
2
IG1
IG2
a Optical conformation energy.
b Minimization binding energy. All the compounds, TAR RNA and
their complexes were minimized by using DISCOVERwith a steepest
descents algorithm in an AMBER FRC potential forcefield on SGI
(Silicon Graphics Inc.) IRIX6.5 workstation, respectively.
In order to reveal the mechanism of their TAR RNA-
binding properties, molecular modeling experiments
were also performed. The docking calculation of the
four small molecules, I1, I2, IG1, and IG2, with target
TAR RNA sequences was carried out using the Biosym
modeling package (Biosym Technologies) in Insight II.
The theoretical calculating energies are presented in
Table 3.
Figure 2. Effect of compounds I1, IG1, I2, and IG2 on Tat-mediated
transactivation in 293T cell.
Table 4. Inhibition effect of compounds on SIV induced syncytium
Compounds
EC50 (lM)
Inhibition %
I1
12.9
6.86
8.74
3.92
44.2
56.2
58.7
60.7
IG1
I2
The modeling results implied that IG2 was a good match
to the trinucleotide bulge (U23-C24-U25) of TAR RNA
and two hydrogen bonds formed between the terminal
guanidinium group of the side chain and the phosphate
of C24 at the intercalation site. IG2 exhibited the lowest
optical conformation energy and minimization binding
energy among the four compounds, which indicated that
the complex IG2–TAR was more stable than the other
analogues. Those data agree well with our CE results.
Usingthe program, DISCOVER, water molecular andmetal
cation have been excluded in the calculation, by which our
experimental data of the binding between the compounds
and TAR RNA are consistent with the estimation.
IG2
formation as well. Also, we found that a side chain with
a suitable length linking isoquinoline and guanidine
group played an important part in the inhibition.
All the experiments reported here show that the newly
designed compound IG2 bearing guanidinium
group-terminated side chain could block the Tat–TAR
interaction and have the anti-HIV potency. These stud-
ies provide a new idea for the design of HIV-1 Tat–TAR
inhibitors.
In order to probe whether the title compounds could
block the Tat-mediated transactivation in human cells,
the activities of compounds I1, I2, IG1, and IG2 were
examined by using Tat dependent HIV-1 LTR-driven
CAT gene expression colorimetric enzyme assays.14,17
As shown in Figure 2, the decreased CAT activities in
the presence of the title compounds reflect their compe-
tition with Tat for TAR RNA binding and the inhibi-
tion of Tat function in vivo.
Acknowledgements
We thank Professor Xinxiang Zhang, Dr. Xiaojian Yao,
and Dr. Li Ding for their excellent technical assistance.
This project was supported by the National Natural
Science Foundation of China (No. 30370323,
No.20332010) and Special Fund for Promotion Educa-
tion. Ministry of Education of China (No.
20030001041).
Moreover, the inhibitions of SIV-induced syncytium in
CEM174 cells with the compounds were tested to con-
firm their antiviral effects,18 and the data were summa-
rized in Table 4. From the two tests above, it is
indicated that compounds containing guanidine group
were more potent in inhibiting Tat–TAR RNA interac-
tion and showed stronger inhibition of the syncytium
References and notes
1. Rana, T. M.; Jeang, K. Arch. Biochem. Biophys. 1999, 365,
175.

M. He et al. / Bioorg. Med. Chem. Lett. 15 (2005) 3978–3981
3981
2. Turpin, J. A. Expert Rev. Anti-infect. 2003, 1, 97.
3. Aboul-ela, F.; Karn, J.; Varani, G. Nucleic Acid Res. 1996,
24, 3974.
4. Calnan, B. J.; Tidor, B.; Biancalana, S.; Hudson, D.;
Frankel, A. D. Science 1991, 252, 1167.
5. Karn, J.; Graeble, M. A. Trends Genet. 1992, 8, 365.
6. Tassew, N.; Thompson, M. Anal. Chem. 2002, 74, 5313.
7. Chow, C. S.; Bogdan, F. Chem. Rev. 1997, 97, 1489.
8. Kato, H.; Miyazawa, K.; Etchu, E. Japan. 7307, 115 (C1.
C07d, A61k)
ning buffer for 4 min each. Solutions were filtered through
a 0.22 lm PTFE membrane prior to use. To ensure proper
folding of the TAR RNA structure, the RNA solutions
were annealed by heating for 3 min at 95 ꢁC and cooled
slowly. TAR-compound or TAR–Tat peptide complex
was incubated for 30 min at 4 ꢁC (binding buffer: 10 mM
Tris–HCl, 70 mM NaCl, 0.2 mM EDTA, and 5% glycerol,
pH 7.4) before CE analysis.
16. The samples were prepared with a wide range of Tat
(or the compounds) diluted by binding buffer (10 mM
Tris–HCl, 70 mM NaCl, 0.2 mM EDTA, and 5% glycerol,
pH 7.4) and CE procedure was as described in Ref. 15. All
experiments were in triplicate. Relative standard deviation
(RSD) was calculated from a series of three experiments
carried out with the same sample in 1 day. Linear
equations of the concentration and the peak area were
calculated. With the peak areas of free Tat or the free
compounds were obtained as described in Ref. 15, their
concentration in the mixture could be calculated.
9. Cain, M.; Weber, R. W.; Guzman, F. J. Med. Chem. 1982,
25, 1081.
10. Melting points were determined on a XA-4 instrument and
1
are uncorrected. All H NMR spectra were on a Varian
1
Unity 300 NMR spectrometer. Chemical shifts (d) for H
spectra were expressed in parts per million relative to
tetramethylsilane (TMS) as an internal standard. Mass
spectra were measured on a VG-ZAB-HS spectrometer or
an ABI QSTAR spectrometer. 1H NMR, MS, mp for
representative compounds: I1: 1H NMR (DMSO-d6) d
9.39 (s, 1H, 1-H), 8.56 (s, 1H, 4-H), 8.26 (m, J = 7.8 Hz,
1H, 8-H), 8.19 (m, J = 7.8 Hz, 1H, 5-H), 7.88 (m,
J = 7.2 Hz, 1H, 6-H), 7.81 (m, J = 7.2 Hz, 1H, 7-H), 3.43
(m, 2H, CONHCH2), 2.74 (m, 2H, NHCH2), 1.23 (m, 2H,
17. Transient transfection and CAT assays: 293T cells were
grown as monolayer in DulbeccoÕs modified EagleÕs
medium (DMEM) (Gibco BRL) supplemented with 10%
(v/v) fetal calf serum, penicillin (100 U · mLꢁ1), and
streptomycin (100 U · mLꢁ1
)
at 37 ꢁC in 5% CO2
1
CH2); MS (EI) (m/z): 216 (M+H); mp: jelly. I2: H NMR
containing humidified air. The cells were seeded at a
six-well plate 24 h prior to transfection which was
performed by standard calcium phosphate co-precipita-
tion techniques with optimum amounts of the plasmids
pLTRCAT and pSVCMVTAT. Twenty four hours later,
the culture medium was removed and the cells were
washed twice with phosphate-buffered saline (PBS). Then
the transfected cells were added to fresh medium together
with diluted compounds of final concentration 30 lM,
respectively, and incubated for another 24 h. After forty
eight hours post-transfection, the cells were harvested and
analyzed for CAT activity using a commercial CAT
ELISA kit (Roche Molecular Biochemicals) in accordance
with the manufacturerÕs protocol. All data were reported
as a percentage of CAT activity ( SD). Results shown are
representative of three independent experiments.
(DMSO-d6) d9.38 (s, 1H, 1-H), 8.56 (s, 1H, 4-H), 8.25 (m,
J = 7.8 Hz, 1H, 8-H), 8.20 (m, J = 7.8 Hz, 1H, 5-H), 7.88
(m, J = 7.2 Hz, 1H, 6-H), 7.81 (m, J = 7.2 Hz, 1H, 7-H),
3.46–1.71 (m, 6H, CH2 · 3); MS (EI) (m/z): 229 ; mp: jelly.
IG1: (FABMS) (m/z): 258.0 (M+H); mp: 136–137 ꢁC. IG2:
(FABMS) (m/z): 271.9 (M+H); mp: 136–137 ꢁC.
11. Zhang, B.; Xue, H.; Lin, B. C. Chromatographia 1998, 48,
268.
12. Mucha, P.; Szyk, A.; Rekowski, P.; Barciszewski, J. J.
Chromatogr. A 2002, 968, 211.
13. Zhao, H.; Dai, D.; Li, J.; Chen, Y.; Jiang, L. BBRC 2003,
312, 351.
14. Yu, X.; Lin, W.; Lin, J.; Yang, M. Bioorg. Med. Chem.
Lett. 2004, 14, 3127.
15. Capillary electrophoresis assay: CE experiments were
carried out on a Beckman P/ACE 2100 capillary electro-
phoresis system using a 50 cm · 50 lm ID bare fused-silica
capillary (Beckman). Phosphate buffer (50 mM, pH 8.0)
was used as running buffer. Electrophoresis was started at
15 kV and 20 0.1 ꢁC. Samples were injected at 10 kV for
20 s and detected at 214 nm. Prior to use, the capillary was
pre-treated successively with 0.1 M NaOH for 60 min,
water for 30 min, and finally with running buffer until the
baseline becomes smooth. Between runs the capillary was
washed sequentially with 0.1 M NaOH, water, and run-
18. Inhibition of SIV-induced syncytium in CEM174 cell
cultures was measured in a 96-well microplate containing
2 · 105 CEM cells/mL infected with 100 TCID50 of SIV
per well and containing appropriate dilutions of the tested
compounds. After 5 days of incubation at 37 ꢁC in 5%
CO2 containing humidified air, CEM giant(syncytium) cell
formation was examined microscopically(COIC). The
EC50 was defined as the compound concentration required
to protect cells against the cytopathogenicity of SIV by
50%.