Design and synthesis of dimeric HIV-1 integrase inhibitory peptides

Article abstract of DOI:10.1016/S0960-894X(03)00679-6

Dimers of known HIV-1 integrase inhibitory hexapeptide H-His-Cys-Lys-Phe-Trp-Trp-NH₂ containing different lengths of cross linkers in the place of cysteine residue, were designed, and synthesized. The inhibitory potency of these dimeric peptides is consistently higher than the lead hexapeptide. The dimeric peptide with djenkolic acid linker exhibited IC₅₀ values of 5.3 and 6.5 μM, for 3′-end processing and strand transfer, respectively.

Full text of DOI:10.1016/S0960-894X(03)00679-6

Bioorganic & Medicinal Chemistry Letters 13 (2003) 3203–3205
Design and Synthesis of Dimeric HIV-1 Integrase Inhibitory
Peptides
Krzysztof Krajewski,a,y Ya-Qiu Long,b Christophe Marchand,c Yves Pommierc
and Peter P. Rollera,*
aLaboratory of Medicinal Chemistry, CCR, NCI-Frederick, NIH, Frederick, MD 21702, USA
bShanghai Institute of Materia Medica, Chinese Academy of Sciences, Shanghai 200031, China
cLaboratory of Molecular Pharmacology, CCR, NCI, NIH, Bethesda, MD 20892, USA
Received 16 April 2003; accepted 26 June 2003
Abstract—Dimers of known HIV-1 integrase inhibitory hexapeptide H-His-Cys-Lys-Phe-Trp-Trp-NH2 containing different lengths
of cross linkers in the place of cysteine residue, were designed, and synthesized. The inhibitory potency of these dimeric peptides is
consistently higher than the lead hexapeptide. The dimeric peptide with djenkolic acid linker exhibited IC50 values of 5.3 and
6.5 mM, for 30-end processing and strand transfer, respectively.
# 2003 Elsevier Ltd. All rights reserved.
HIV-1 integrase is essential for the HIV replication
cycle,1 furthermore it is a key enzyme for the ability of
the HIV virus to infect non-dividing cells.2 The
mechanism of integrase enzymatic activity involves two
steps. First, the enzyme removes two nucleosides from
each of the viral DNA ends (30-processing) then, after
migration of the DNA-protein complex to the nucleus,
integrase catalyzes the insertion of the viral DNA
through its 30-end into the cellular chromosome (strand
transfer, referred to as integration or 30-end joining). In
addition to reverse transcriptase and protease, integrase
has also been the focus of attention for HIV antiviral
chemotherapy. Integrase is an attractive target because
it has no counterpart in mammalian cells; therefore,
selective integrase inhibitors should not produce any
side effects. Many integrase inhibitors have been repor-
ted to date.3 However, progress with the design of
effective inhibitors of HIV-integrase is slower than in the
case of reverse transcriptase or protease inhibitors and
no clinically useful inhibitors have yet been reported.
that a dimeric peptide (2c) was a more potent inhibitor
of integrase than the monomeric hexapeptide 1.5 This
suggests that a dimeric peptide may occupy the two
neighboring catalytic sites in the integrase oligomer. In
order to study the relationship between the linker length
or linker functionality, and the inhibitory activity of
dimeric peptide, we designed several linkers and devel-
oped an efficient strategy for dimeric peptide synthesis.
Here, we present the syntheses and results of HIV-1
integrase inhibitory activity assays of these newly gen-
erated dimeric peptides 2a–e (Fig. 1) with five different
dimerization linkers: l-lanthionine 4a, l-cystine 4b,
l-cystathionine 4c, l-homolanthionine 4d, and l-djenko-
lic acid 4e. Linkers 4a and 4c have been found pre-
viously in natural product peptides, and are used
frequently as non reducible analogues of 4b6À8 mostly as
a part of cyclic structures. Linkers 4d9 and 4e10 have
found only infrequent usage, mostly as components in
cyclic peptides.
The lead peptide 1 was synthesized by an automated
SPPS using Rink amide resin and Fmoc chemistry; the
dimeric cystine containing peptide 2b was isolated as the
oxidative byproduct after preparative HPLC of 1.
During our structure–activity studies of analogues of
the hexapeptide integrase inhibitor, His-Cys-Lys-Phe-
Trp-Trp-NH2 (1), identified earlier by peptide library
methodology by Plasterk and coworkers,4 we found
The method of synthesis of dimeric peptides 2a, 2c, 2d,
and 2e is presented on Scheme 1 (synthesis of the
l-homolanthionine containing peptide 2d is shown as an
example). The di-Fmoc derivatives 5a, 5c, 5d, and 5e
were prepared by reaction of the corresponding amino
*Corresponding author: Tel.: +1-301-846-5904; fax: +1-301-846-
6033; e-mail: proll@helix.nih.gov
yPermanent address: Faculty of Chemistry, Wroclaw University,
Joliot-Curie 14, 50-383 Wroclaw, Poland.
0960-894X/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0960-894X(03)00679-6

3204
K. Krajewski et al. / Bioorg. Med. Chem. Lett. 13 (2003) 3203–3205
Scheme 1. The synthesis of dimeric peptide 2d: (i) 1% TFA, 1% EDT
in DCM (15Â2 min); (ii) FEP, DIEA in DCM (3 h, À10 ꢀC!rt); (iii)
50% piperidine in DMF (1 h, rt), preparative HPLC; (vi) (1) Boc-
His(Boc)-OH, HATU, HOAt, DIPEA (3 h rt); (2) 2.5% TIS, 2.5%
H2O in TFA (2 h, rt).
2.94 (1H, dd, 13.2 Hz, 4.2 Hz), 2.77 (1H, dd, 13.2 Hz,
9.7 Hz), 2.65–2.53 (2H, m), 1.98–1.84 (2H, m); 5d: 7.88
(4H, d, 7.4 Hz), 7.71 (4H, d, 7.4 Hz), 7.41 (4H, dd,
2Â7.4 Hz), 7.32 (4H, dd, 2Â7.4 Hz), 4.29–4.17 (6H, m),
4.12–4.07 (2H, m) 2.62–2.47 (4H, m), 1.98–1.83 (4H, m);
5e: 7.88 (4H, d, 7.5 Hz), 7.72 (4H, d, 7.5 Hz), 7.41 (4H,
dd, 2Â7.5 Hz), 7.32 (4H, dd, 2Â7.5 Hz), 4.34–4.18 (8H,
m), 3.83 (2H, s), 3.11 (2H, dd, 13.7 Hz, 4.8 Hz), 2.83
(2H, dd, 13.7 Hz, 9.9 Hz)). The observed monoisotopic
masses for 5a, 5c, 5d, and 5e (FAB-MS) were in good
agreement with their theoretical masses (in parentheses):
5a 653.5, 675.5 (653.2 M+H+, 675.2 M+Na+) Da; 5c
667.2 (667.2 M+H+) Da; 2d 681.2 (681.5 M+H+) Da;
2e 699.4 (699.2 M+H+) Da.
Figure 1. Schematic representation of prepared peptides.
acids with FmocOSu (2.3 equiv) and Et3N (5 equiv) in
acetonitrile/water (1:1 vol) mixture (rt, 3 h). The products
were purified by MPLC11 and the purity of each pro-
duct was >95% (determined by RP HPLC, C8 col-
umn). The dimeric amino acids 4c and 4e are
commercially available.12 The amino acids 4a and 4d
were prepared fromthe methyl esters of Boc 2-l-cystine
and Boc2-l-homocystine by the sulfur extrusion reac-
tion with tris(diethylamino)phosphine in dry benzene13
(conditions for cystine derivative: 1 h, rt; for homocys-
tine derivative: reflux 5 min, then 24 h at 30 ꢀC), puri-
fication by MPLC and deprotection (LiOH in THF/
water, rt, 1 h, then 50% TFA in DCM À10 ꢀC to rt, 2
h). 1H- and 1H,1H-COSY NMR spectra of 5a (mp
Typical Procedure for Dimeric Peptides Synthesis
The tetrapeptide Lys(Boc)-Phe-Trp(Boc)-Trp(Boc)-
NH2 (3) was synthesized by manual SPPS on Sieber
amide resin (Aldrich, 0.35 mmol/g) using Fmoc chem-
istry, cleaved fromthe resin by treatment with, 1%
TFA, 1% EDT in DCM (15Â2 min), and purified14 by
MPLC.15 The di-Fmoc protected linker 5 was coupled
with 3 (2.5 equiv) in DCM solution, with 2-fluoro-1-
ethylpyridiniumtetrafluoroborate (FEP) 16 as a coupling
reagent (FEP gave better results than HATU/HOAt)
and DIPEA. The coupling product 6 was purified by RP
HPLC, after deprotection of N-terminal NH2 groups
(50% piperidine in DCM). After purification and lyo-
philization, the peptide 7 was coupled with Boc-His
(Boc)-OH in DMF (HATU/HOAt/DIPEA). After
completion of the reaction water was added to pre-
cipitate the fully protected peptide. This peptide was
washed three times with distilled water, dried under
111–114 ꢀC, ½aŠ2D0=+10ꢀ,
c 0.15 in MeOH), 5c
(mp 129–130 ꢀC, ½aŠD20=À6ꢀ, c 0.15 in MeOH), 5d (mp
198–204 ꢀC dec., ½aŠ2D0=À26ꢀ, c 0.45 in DMF), and 5e
(mp 123–125 ꢀC, ½aŠ2D0=À57ꢀ, c 0.45 in DMF) con-
firmed their structures (1H NMR, 400 MHz, in
(CD3)2SO; 5a: 7.88 (4H, d, 7.4 Hz), 7.72 (4H, d, 7.4 Hz),
7.41 (4H, dd, 2Â7.4 Hz), 7.32 (4H, dd, 2Â7.4 Hz), 4.31–
4.11 (8H, m), 3.01–2.95 (2H, m), 2.85–2.77 (2H, m); 5c:
7.88 (4H, d, 7.5 Hz), 7.71 (4H, d, 7.5 Hz), 7.40 (4H, dd,
2Â7.5 Hz), 7.32 (4H, dd, 2Â7.5 Hz), 4.34–4.18 (8H, m),

K. Krajewski et al. / Bioorg. Med. Chem. Lett. 13 (2003) 3203–3205
3205
Table 1. Results of in vitro HIV-1 integrase inhibitory assay for
peptides 1, 2a–e
also for the strand transfer process. The linker moieties
of the dimeric peptides in this study include the thio-
ethers, the dithio linkage and the dithiomethylene lin-
kages. The peptide backbone to backbone lengths
consist of four, five, or six bonds. Significantly, the
dithiomethylene linker containing dimer (peptide 2e)
showed an approximately 20-fold higher potency com-
pared to the monomer peptide 1. This inhibitory
potency was sustained both in the 30-processing and the
strand transfer process of the integrase enzyme. The
mechanistic implications of inhibitory action suggest
that the dimeric inhibitory peptide may act as a bivalent
inhibitor, simultaneously occupying two neighboring
catalytic sites in tandem, with an entropic advantage,
within the integrase oligomeric complex. Indeed, recent
X-ray structure of a small molecule inhibitor bound to
the three integrase core domains showed that the active
site is very close to the active site of the neighboring
integrase subunit.18
Peptides [IC50] (mM)a
1
2a
2b
2c
2d
2e
30-End processing 102ꢁ20 67ꢁ7 81ꢁ10 77ꢁ21 59ꢁ7 5.3ꢁ0.7
Strand transfer
90ꢁ15 19ꢁ6 58ꢁ15 87ꢁ23 63ꢁ7 6.5ꢁ1.9
aValues are means of at least three experiments, and are given with
standard deviations.
reduced pressure, and deprotected by mixing with a
solution of 2.5% TIS and 2.5% water in TFA. After 2 h,
the solvents were removed under reduced pressure, the
residue was washed six times with ether and dissolved in
water/acetonitrile mixture (0.05% TFA), and lyophi-
lized, The product was dissolved again in 0.05% TFA in
water, kept at rt 6 h (to complete Trp residue deprotec-
tion), and then purified by preparative RP HPLC
(Vydac C4 or C18 column). Purity of all peptides was
between 90 and 95%, (RP HPLC on C8 and C4 col-
umns), MALDI-TOF-MS spectra verified molecular
masses of all peptides. The observed monoisotopic
masses were in good agreement with theoretical masses
(in parentheses): 1 905.3 (905.4 M+H+) Da; 2a 1775.3
(1775.8 M+H+) Da; 2b 1808.0 (1807.8 M+H+) Da; 2c
1789.7 (1789.9 M+H+) Da; 2d 1803.9 (1803.9 M+H+)
Da; 2e 1821.8 (1821.8 M+H+) Da.
References and Notes
1. Turner, B. G.; Summers, M. F. J. Mol. Biol. 1999, 285, 1.
2. Vodicka, M. A. Somat. Cell. Mol. Genet. 2001, 26, 35.
3. Pommier, Y.; Marchand, C.; Neamati, N. Antiviral Res.
2000, 47, 139.
4. Lutzke, R. A. P.; Eppens, N. A.; Weber, P. A.; Houghten,
R. A.; Plasterk, R. H. A. Proc. Natl. Acad. Sci. U.S.A. 1995,
92, 11456.
5. Roller, P. P.; Long, Y.-C.; Lung, F.-D. T.; Pommier, Y.;
Neamati, N. In Peptides 2000; Martinez, J., Fehrentz, J. A.
Eds., EDK, Paris, 2001; p 725.
6. Tromelin, A.; Moutiez, M.; Meziane-Cherif, D.; Aumer-
cier, M.; Tartar, A.; Sergheraert, C. Bioorg. Med. Chem. Lett.
1993, 3, 1971.
HIV-1 Integrase Inhibition Assay
The HIV-1 integrase assays were performed as descri-
bed previously,17 with the following modifications. The
peptides were pre-incubated with 500 nM HIV-1 inte-
grase for 20 min at room temperature in a buffer con-
taining 50 mM MOPS, pH 7.2, 7.5 mM NaCl, 7.5 mM
MnCl2, and 5% glycerol. Reactions were started by
adding 20 nM of the 50-end 32P-labeled 21-mer double-
stranded DNA template in a final volume of 10 mL, and
reactions were carried out for 1 h at 37 ꢀC. Reactions
were quenched by adding 10 mL of denaturing loading
dye (formamide 99%, SDS 1%, bromophenol blue 0.2
mg/mL, xylene cyanol FF 0.2 mg/mL). Samples were
loaded onto 20% (19:1) denaturing polyacrylamide gels.
Gels were dried, exposed overnight and analyzed using
Molecular Dynamics PhosphorImager (Sunnyvale, CA,
USA). The densitometric analysis was performed using
ImageQuant v1.2 from Molecular Dynamics software
package. Each lane was quantified to determine the
amount of 30-processing and strand transfer products
(Table 1).
7. Li, H.; Jiang, X.; Howell, S. B.; Goodman, M. J. Peptide
Sci. 2000, 40, 26.
8. Johnes, D. S.; Gamino, C. A.; Randow, M. E.; Victoria,
E. J.; Yu, L.; Coutts, S. M. Tetrahedron Lett. 1998, 39, 6107.
9. Fotouhi, N.; Joshi, P.; Tilley, J. W.; Rowan, K.; Schwinge,
V.; Wolitzky, B. Bioorg. Med. Chem. Lett. 2000, 10, 1167.
10. Ueki, M.; Ikeo, T.; Iwadate, M.; Asakura, T.; Williamson,
M. P.; Slaninova, J. Bioorg. Med. Chem. Lett. 1999, 9, 1767.
11. Isco, CombiFlash Separation System Sg100c (silica gel,
hexane–EtOAc, 0–100% EtOAc/30 min).
12. l-Cystathionine fromAldrich ChemCo, Milwaukee, WI,
USA, l-djenkolic acid fromTCI America, Portland, OR, USA.
13. Harpp, D. N.; Gleason, J. G. J. Org. Chem. 1971, 36, 73.
14. FAB-MS: 987.2 [987.5 (M+Na+)] Da, HPLC (after
deprotection): 1 peak.
15. Isco, CombiFlash Separation System Sg100c (silica gel,
EtOAc–MeOH, 0–100% MeOH/30 min).
16. Li, P.; Xu, J.-C. Tetrahedron 2000, 56, 8119.
17. Marchand, C.; Neamati, N.; Pommier, Y. Methods Enzy-
mol. 1999, 340, 624.
In conclusion, HIV-1 integrase inhibitory assays on
peptides 1 and 2a–2e demonstrate that the dimeric pep-
tides possess higher inhibitory potency than monomeric
peptide 1 for the integrase enzyme 30-processing and
18. Goldgur, Y.; Craigie, R.; Cohen, G. H.; Fujiwara, T.;
Yoshinaga, T.; Fujishita, T.; Sugimoto, H.; Endo, T.; Murai,
H.; Davies, D. R. Proc. Natl. Acad. Sci. U.S.A. 1999, 96,
13040.