Design, synthesis and evaluation of potential inhibitors of HIV gp120-CD4 interactions

Article abstract of DOI:10.1016/j.bmcl.2004.02.091

This paper describes an approach to prevent HIV-cell fusion by disrupting the interaction between HIV protein gp120 and CD4 receptor. The CD4 residues Phe43 and Arg59 make important interactions with gp120. Small molecule analogues were made to mimic the crucial features of these residues. The analogues were assayed using a cellular 'FIGS' assay to measure inhibition of cell fusion and caused some inhibition.

Full text of DOI:10.1016/j.bmcl.2004.02.091

Bioorganic & Medicinal Chemistry Letters 14 (2004) 2673–2676
Design, synthesis and evaluation of potential inhibitors of HIV
gp120–CD4 interactions
Cyrille Boussard,a Thomas Klimkait,b Naheed Mahmood,c Martin Pritchardd and
Ian H. Gilberta,*
aWelsh School of Pharmacy, Cardiff University, Redwood Building, King Edward VII Avenue, Cardiff CF10 3XF, UK
bInstitute for Medical Microbiology, University of Basel, CH-4003 Basel, Switzerland
cTurner Building, St Batholomews and The Royal London School of Medicine and Dentistry, Turner Street, London E1 2AD, UK
dParke–Davis Neuroscience Research Centre, Addenbrookes Hospital Site, Hills Road, Cambridge, UK
Received 24 November 2003; revised 2 February 2004; accepted 4 February 2004
Abstract—This paper describes an approach to prevent HIV-cell fusion by disrupting the interaction between HIV protein gp120
and CD4 receptor. The CD4 residues Phe43 and Arg59 make important interactions with gp120. Small molecule analogues were
made to mimic the crucial features of these residues. The analogues were assayed using a cellular ‘FIGS’ assay to measure inhibition
of cell fusion and caused some inhibition.
Ó 2004 Elsevier Ltd. All rights reserved.
1. Introduction
alternative approach, we decided to see if we could
design small compounds, which bind to HIV gp120 and
prevent interaction with CD4.
HIV-AIDS is a major cause of mortality and due to the
potential for development of resistance to existing
therapies,1 discovery of new therapeutic agents is cru-
cial. While current therapies are based on inhibition of
reverse transcriptase and protease,2 we were interested
in investigating the fusion step as a potential drug target.
Entry of the virus into the cell is a complex process.
Initially, the viral surface protein gp120 interacts with
the cellular receptor CD4.3 This causes a change in the
conformation of gp120,4 which binds to a second cel-
lular receptor of the family of chemokine receptors,5
allowing the HIV protein gp41 to penetrate the cell
membrane, leading to membrane fusion.6
The gp120–CD4 interaction is an example of a protein–
protein interaction. Binding occurs often over a rela-
tively large surface area and identifying specific binding
sites is crucial in the development of a new protein
antagonist.13 The crystal structure of gp120 bound to
CD4 has been solved14 and indicates that two residues
from CD4 seem to be of crucial importance in interac-
tion with gp120; these are Phe43 and Arg59. For these
residues, which are spatially in close proximity, the
important interactions are through the phenyl ring and
the guanidine group, respectively.15
A number of groups have been working on developing
inhibitors of fusion.5;7 Much of this work has concen-
trated on development of antagonists of chemokine
receptors CCR58 and CXCR4.2;9;10 More recently, new
compounds targeting the fusion process, have shown
promising results in clinical trials, leading to the
approval of the first entry inhibitor (Enfuvirtide, a linear
36 amino acid peptide mimetic of the HR2 region of
gp41) by the US FDA in March 2003.11;12 As an
In a previous paper, the identification of lead peptides as
potential inhibitors of gp120–CD4 interactions was
described.16 A series of tri- and tetra-peptides based
around the amino acids arginine and phenylalanine and
designed to mimic the original spatial arrangement of
the gp120 residues Arg59 and Phe 43, were shown to
have moderate inhibition of this interaction in a cell-free
assay. Biological testing of these libraries led to IC50
values from 0.17 mM for HPhe-Asp-ArgNH2 to
9.75 mM for HPhe-Gly-Phe-ArgNH2. It has been
noticed that shorter peptides showed higher activity and
that constrained and hydrophobic residues enhanced the
inhibitory properties of these peptides.
* Corresponding author. Fax: +44-29-2087-4149; e-mail: gilbertih@
cardiff.ac.uk
0960-894X/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2004.02.091

2674
C. Boussard et al. / Bioorg. Med. Chem. Lett. 14 (2004) 2673–2676
Figure 1. (a) Molecules designed to mimic the relative orientation of Phe43 and Arg59 in their binding to gp120. (b) The structure of the designed
molecule A minimised using Monte Carlo conformational searching, followed by minimisation (Macromodel 6.0). (c) Superimpositions of original
CD4 residues Phe43 and Arg59 (original spatial arrangement, in blue) and designed peptoid structure A (in grey) after minimisation (using
Macromodelâ).
As peptides are limited for use as drugs because of their
susceptibility to peptidase degradation, poor oral bio-
availability and their difficulty in crossing the blood–
brain barrier, development of nonpeptidic compounds
by a peptoid approach was envisaged. This paper
describes the design, synthesis and biological evaluation
of new potential inhibitors of the HIV gp120–CD4
interaction by using the previous reported work.
iline. The use of three guanylating reagents on the 3-
phenoxyaniline 1 gave, respectively, moderate yields with
the diprotected thiourea 419 and the bis-Boc protected
guanyl pyrazole 5;20 and quantitative yields with the bis-
Boc protected S-methyl isothiourea 621 (Scheme 1).
The final step of the guanidine synthesis was the
deprotection of the guanidine moiety. Boc deprotection
in acidic conditions using trifluoroacetic acid (TFA) in
DCM22 was carried out and led to partial cleavage of the
Boc groups, giving a mixture of the mono-protected
form and other impurities. However, the method of Miel
and Rault23 using tin(IV) chloride followed by tritur-
ation with ether, yielded the desired product 3 without
any of the mono-Boc guanidine intermediate.
2. Compound design
To inhibit the gp120–CD4 interaction, a crucial feature
of the designed compounds was to hold the phenyl ring
and the guanidine groups in their correct stereochemical
alignment. A good mimic of the stereochemical orien-
tation of these groups was obtained by attaching the
phenyl and the guanidine groups to a benzene ring. (Fig.
1a). This was confirmed in silico by performing Monte–
Carlo conformational searching followed by minimisa-
tion (Fig. 1b). The lowest energy structure was then
superimposed onto the coordinates Phe43 and Arg59 as
found in CD4 in the gp120–CD4 complex (Fig. 1c) and
gave a good fit.
3.2. SAR studies
Having established a suitable pathway for the prepara-
tion of the phenoxy guanidine derivatives, this meth-
odology was applied to the corresponding benzyl
derivatives. In order to probe interactions of the benzyl
moiety with the protein, it was decided to prepare a
broad series of compounds with different substituents on
the benzyl ring.
Owing to the preliminary activity of the lead compounds
(compound 3 and 12a), a number of variations of the
structure of the lead compound were proposed; the
guanidine moiety could be replaced by the amidine
group, which is a potential bio-isostere17 and various
substituents could be added to the benzyloxy moiety to
change electronic and steric properties.
3. Chemistry18
3.1. Synthesis of the guanidine series
Scheme 1. Reagents and conditions: (i) method 1: 4/Mukaiyama’s
reagent/TEA/DCM, 46%; method 2: 5/anhydrous THF, 55%; method
3: 6/THF, 95%; (ii) SnCl4/AcOEt, MeOH/Et2O, 50%; (iii) TFA/DCM,
partial cleavage.
Investigation of potential methods for guanylation were
performed on the commercially available 3-phenoxyan-

C. Boussard et al. / Bioorg. Med. Chem. Lett. 14 (2004) 2673–2676
2675
Scheme 2. Reagents and conditions: (i) TBDMSCl/anhyd pyridine, 90%; (ii) method 3: Guanylating reagent 6/THF, 95%; (iii) TBAF/DCM, 95%;
(iv) NaH/DMF; (v) a: benzyl chloride; b: 4-methoxybenzyl bromide, 80%; c: 4-bromobenzyl bromide, 40%; d: 4-chlorobenzyl chloride, 88%; e: 3-
nitrobenzyl bromide, 35%; f: 4-methylbenzyl bromide, 46%; g: 4-chloromethyl phenyl acetate, no yield available; h: 4-benzyloxybenzyl chloride, 86%;
(vi) SnCl4/AcOEt, MeOH/Et2O, 95% over two steps for 12a; (vii) for 13c: 2 equiv TFA/DCM, quantitative; (viii) for 13d: 10 equiv TFA/DCM, 89%.
The most efficient synthesis would proceed via a com-
mon intermediate, which would allow attachment of a
variety of substituted benzyl groups. Starting from the
commercially available 3-aminophenol 7, the protection
of the hydroxyl in position 3 was achieved by action of
TBDMSCl in pyridine. Then the bis-Boc protected
guanidine derivative 9 was prepared via the use of the
bis-Boc protected S-methyl isothiourea 6 in high yield
(95%). The last step of the preparation of compound 10
was the cleavage of the tert-butyldimethylsilyl group,
achieved by using TBAF in DCM (Scheme 2).
hydroxyl using either the appropriately substituted
benzyl bromide or chloride22 in 35–88% (Scheme 2).
The Boc-protected guanidines were deprotected. Ini-
tially tin(IV) chloride was used and yielded 12a but
11b,c,e decomposed under these conditions. Use of
excess of TFA (10 equiv) in DCM yielded the chloro-
derivative 13d but led to decomposition of 11b,f and h.
Finally 2 equiv of TFA yielded the bromoderivative 13c
but led to decomposition of 11b,e,f and h (Scheme 2).
The next step was to introduce the substituted benzyl
groups. A variety of substituents were introduced onto
the phenyl ring, mainly in the para-position. These were
selected to give a variety of types of substituent––elec-
tron withdrawing, electron donating, bulky and allow-
ing further additional chemical modifications such as
reduction or acylation. Starting from compound 10,
derivatives 11a–h were prepared by alkylation of the
3.3. Amidine synthesis (Scheme 3)
Amidines are an isostere of the guanidine group. They
were prepared by treatment of aldehyde 15 with
hydroxylamine hydrochloride in ethanol in presence of
HCl to give a good yield (75%) of the nitrile 16.24 The
nitriles 16 were reacted with the etherate complex of
a lithium salt of hexamethyldisilazane (LiN(SiCH3)2
N
C
H
O
(i)
R
O
R
O
15
16: R=C6H5
17: R=H
18a:
18b:
18c:
18d:
18e:
R=CH2C6H5
(iv)
R=p(Br)C6H5CH2
H2N
NH
R=p(OCH2C6H5)C6H5CH2
R=p(OCOCH3)C6H5CH2
R=p(CH3)C6H5CH2
R=p(Cl)C6H5CH2
R
18f:
O
18g: R=p(NO2)C6H5CH2
18h:
R=m(NO2)C6H5CH2
R=C6H5
19:
R=CH C H
20a:
20b:
20c:
20e:
20f:
20h:
R=p(B2r)C H CH
6
6
2
5
5
2
R=p(OCH C H )C H CH
6
5
5
2
R=p(CH )C H CH6
R=p(Cl)C6H5CH2
3
6
5
2
R=m(NO )C H CH
2
2
6
5
Scheme 3. Reagents and conditions: (i) NH2OHÆHCl/EtOH/HCl/heat, 75%; (ii) NaH/DMF; (iii) a: benzyl bromide, 93%; b: 4-bromobenzyl bromide,
75%; c: 4-benzyloxybenzyl chloride, 84%; d: 4-chloromethyl phenyl acetate, 91%; e: 4-methylbenzyl bromide, 59%; f: 4-chlorobenzyl chloride, 68%; g:
4-nitrobenzyl bromide, 15%; h: 3-nitrobenzyl bromide, 92%; (iv) 4 equiv of corresponding compounds 16, 18a–c,e,f and h and 2 equiv of
LiN(SiCH3)2Et2O prepared in situ by addition of n-BuLi to a solution of hexamethyldisilazane in Et2O at 0 °C, HCl, NaOH, 90%–95%.

2676
C. Boussard et al. / Bioorg. Med. Chem. Lett. 14 (2004) 2673–2676
Et2O)25 to give the free amidine 19 pure with excellent
yield (90%, Scheme 3).25
Acknowledgements
We would like to acknowledge Ingrid Ziekau and Doris
€
and the Welsh School of Pharmacy for funding and the
EPSRC National Mass Spectrometry Centre at Swansea
for accurate mass spectrometry.
Again the SAR of the phenyl ring was probed using
substituted benzyl derivatives. These were prepared by
alkylation of 3-cyanophenol 17 with various substituted
benzyl chlorides followed by treatment with lithium
hexamethyldisilazane. The acetyl 18d and para-nitro 18g
derivatives did not yield the required amidines.
Low for excellent technical assistance, Parke–Davies
References and notes
4. Biological assays
1. http://www.WHO.org
2. Cooley, L. A.; Lewin, S. R. J. Clin. Virol. 2002, 1–12.
3. Dalgeish, A. G.; Beverley, P. C.; Clapham, P. R.;
Crawford, D. H. Nature 1984, 312, 763–767.
4. Wyatt, R.; Kwong, P. D.; Desjardins, E.; Sweet, R. W.;
Robinson, J.; Hendrickson, W. A.; Sodroski, J. G. Nature
1998, 393, 705–711.
Compounds were principally evaluated using a cellular
assay to study cell fusion, the so-called FIGS assay.26
Essentially in this assay HIV infected cells are cultured
in the presence of a reporter cell. The reporter cell
expresses a b-galactosidase when fusion occurs with an
HIV-infected cell. However, when fusion is prevented,
the expression is prevented. The levels of expression of
b-galactosidase are measured using a colourimetric
assay. Encouragingly, compounds 3 and 12a showed
inhibition of fusion with IC50 of 167 and 131 lM,
respectively. However the remaining compounds
showed cytotoxicity at levels where activity was found,
minimising a ‘therapeutic window’.
5. LaBranche, C. C.; Galasso, G.; Moore, J. P.; Bolognesi,
D. P.; Hirsch, M. S. Antiviral Res. 2001, 50, 95–115.
ꢀ
6. Yachou, A.; Sekaly, R. P. Biochem. Biophys. Res.
Commun. 1999, 265, 428–433.
7. Jiang, S. B.; Zhao, Q.; Debnath, A. K. Curr. Pharm. Des.
2002, 8, 563–580.
8. Tremblay, C. L.; Giguel, F.; Kollmann, C.; Guan, Y. B.;
Chou, T. C., et al. Antimicrob. Agents Chemother. 2002,
46, 1336–1339.
9. Tamamura, H.; Omagari, A.; Oishi, S.; Kanamoto, T.;
Yamamoto, N.; Peiper, S. C.; Nakashima, H.; Otaka, A.;
Fujii, N. Bioorg. Med. Chem. Lett. 2000, 10, 2633–2637.
10. Ruff, M. R.; Melendez-Guerrero, L. M.; Yang, Q.;
Ho, W.; Mikovits, J. W., et al. Antiviral Res. 2001, 52,
63–75.
11. Fletcher, C. V. The Lancet 2003, 361, 1577–1578.
12. Labonte, J.; Lebbos, J.; Kirkpatrick, P. Nature Rev. Drug
Discov. 2003, 2, 345–346.
Compounds 3 and 12a were then investigated further
and assayed using an ELISA assay16 to detect the inhi-
bition of the HIV-specific CD4–gp120 interaction using
recombinant protein. The compounds showed weak
inhibition at concentrations of 5 mM. They were also
tested for their activity against HIV-1 IIIB infections of
C8166 cells.27 In this last assay the compounds gave
some anti-HIV activity (EC50 ꢀ 4–8 lM), although with
a low therapeutic index (TC50 ¼ 20 and 40 lM, respec-
tively).
13. Toogood, P. L. J. Med. Chem. 2002, 45, 1543–1558.
14. Kwong, P. D.; Wyatt, R.; Robinson, J.; Sweet, R. W.;
Sodroski, J.; Hendrickson, W. A. Nature 1998, 393, 648–
659.
15. Moore, J. P.; Binley, J. Nature 1998, 393, 630–631.
16. Boussard, C.; Doyle, V. E.; Mahmood, N.; Klimkait, T.;
Pritchard, M.; Gilbert, I. H. Eur. J. Med. Chem. 2002, 37,
883–890.
Compounds 13c,d, 19, 20a–c,e,f and h were all found to
be inactive at 10 lM and toxic at 100 lM in the FIGS
assay and did not show any antiviral activity.
17. Magalhes, A.; Monteiro, M. R.; Magalhes, H. P.; Mares-
Guia, L.; Rogana, E. Toxicon. 1997, 35, 1549–1559.
18. Compounds were characterised by 1H, 13C, mass spec-
trometry and accurate mass analysis.
19. Yong, Y. F.; Kowalski, J. A.; Lipton, M. A. J. Org. Chem.
1997, 62, 1540–1542.
5. Conclusion
In this paper we have focused on preparation of small,
nonpeptide analogues that contain functionality corre-
sponding to the side chains of Phe43 and Arg59. Two of
the compounds prepared showed some activity in a
reporter cell assay for prevention of HIV––cellular
fusion and weak activity in an ELISA assay to measure
CD4–gp120 interaction using recombinant protein.
These compounds could be a starting point for further
synthesis.
20. Bernatowicz, M. S.; Wu, Y.; Matsueda, G. R. Tetrahedron
Lett. 1993, 34, 3389–3392.
21. Miel, H.; Rault, S. Tetrahedron Lett. 1998, 39, 1565–
1568.
22. Greene, T. W.; Peter, G. M. Protective Groups in Organic
Synthesis, 2nd ed.; Wiley-Interscience: New York, USA,
1991.
23. Miel, H.; Rault, S. Tetrahedron Lett. 1997, 38, 7865–7866.
24. Findlay, J. A.; Tang, C. S. Can. J. Chem. 1967, 45, 1014–
1015.
25. Lappert, M. F.; Slade, M. J.; Singh, A. J. Am. Chem. Soc.
1983, 105, 302–304.
26. Klimkait, T.; Stauffer, F.; Lupo, E.; Sonderegger-Rubli, C.
Arch. Virol. 1998, 143, 2109.
27. HIV, Biochemistry, Molecular Biology and Drug Discov-
ery; Karn, J., Ed.; Oxford University Press: New York,
USA, 1995; Vol. 2.
The compounds showed higher potency in an infection
assay, than in the fusion assay characterised by less
stringent conditions. However, these data taken together
suggest that the main mode of antiviral action of the
compound is not by inhibition of CD4–gp120 interac-
tion.