Lysine derivatives as potent HIV protease inhibitors. Discovery, synthesis and structure-activity relationship studies

Article abstract of DOI:10.1016/j.bmcl.2004.12.068

A screening assay program on HIV-protease was carried out on more than fifty commercially available N-protected amino acids and has revealed that those with a long side chain such as lysine, ornithine and arginine exhibited significant inhibition of HIV protease enzyme. The presence of an Fmoc group was found to be essential to obtain micromolar inhibitors and the addition of an alkyl group at the Nα-position resulted in the discovery of the lead compound 11 displaying a 5 nM inhibition constant. Although this new inhibitor series is not categorized among those mimicking the substrate with a non-hydrolyzable transition-state isoster, it was found very specific to inhibit HIV protease enzyme in comparison to the mammalian aspartyl proteases pepsin, renin and cathepsin. Furthermore, these inhibitors did not show any cytotoxicity at a concentration below 75 μM.

Full text of DOI:10.1016/j.bmcl.2004.12.068

Bioorganic & Medicinal Chemistry Letters 15 (2005) 1509–1513
Lysine derivatives as potent HIV protease inhibitors.
Discovery, synthesis and structure–activity relationship studies
a
b
Abderrahim Bouzide,a, Gilles Sauve and Jocelyn Yelle
*
´
aPharmacor Inc., 535, Cartier West Blvd., Laval, Que, Canada H7V 3S8
bProcyon Biopharma Inc., 1650, Trans Highway, Suite 200, Dorval, Que, Canada H9P 1H7
Received 3 December 2004; revised 20 December 2004; accepted 21 December 2004
Available online 25 January 2005
Abstract—A screening assay program on HIV-protease was carried out on more than fifty commercially available N-protected
amino acids and has revealed that those with a long side chain such as lysine, ornithine and arginine exhibited significant inhibition
of HIV protease enzyme. The presence of an Fmoc group was found to be essential to obtain micromolar inhibitors and the addition
of an alkyl group at the Na-position resulted in the discovery of the lead compound 11 displaying a 5 nM inhibition constant.
Although this new inhibitor series is not categorized among those mimicking the substrate with a non-hydrolyzable transition-state
isoster, it was found very specific to inhibit HIV protease enzyme in comparison to the mammalian aspartyl proteases pepsin, renin
and cathepsin. Furthermore, these inhibitors did not show any cytotoxicity at a concentration below 75 lM.
Ó 2005 Elsevier Ltd. All rights reserved.
1. Introduction
new and improved drugs to target HIV protease due
to increasing viral resistance, a matter that is now of
great concern.5 The high cost of synthesis is today a bar-
rier to the widespread use of the currently approved pro-
tease inhibitors, notably in the less developed countries.
It is thus important that new improved protease inhibi-
tors accessible at low cost are developed.6
Since its discovery two decades ago, acquired immuno-
deficiency syndrome (AIDS) has increased to epidemic
proportions throughout the world, affecting more than
40 million people today and killing so far more than
22 million (UNAIDS, 2004).1 Among various strategies
to combat this devastating disease, therapeutic inhibi-
tion of the virally encoded human immunodeficiency
virus (HIV) protease became an attractive target. Its
inactivation leads to the formation of immature and
non-infectious virions.2 A number of reports on the de-
sign and synthesis of HIV protease inhibitors have been
published.3 In general, in these inhibitors, the scissile
bond has been replaced by a non-hydrolyzable transi-
tion-state isoster. Unfortunately most of these pepti-
domimetic HIV protease inhibitors retain a substantial
amount of peptide character, and as a result, their oral
bioavailability is low and a short plasma half-time is
observed.
2. Lead compound discovery
During a continuous effort in our laboratory to develop
small molecules with low cost as potent HIV protease
inhibitors, we initiated a broad screening program by
testing more than 50 commercially available N-pro-
tected amino acids.7 It was noticed that those with a
long side chain, such as lysine, ornithine and arginine,
showed decent micromolar inhibition constants (Ki),
especially when they are protected with bulky protecting
groups such us an Fmoc carbamate.
Despite the success of the FDA-approved HIV protease
inhibitors,4 saquinavir, ritonavir, indinavir, nelfinavir,
amprenavir and lopinavir, there is an urgent need for
Some commercially available amino acids that showed
inhibitions are presented in Table 1. Amino acids pro-
tected with small protecting groups such as Boc- or
Cbz-carbamates did not exhibit any activity (1 and 2).
However, the presence of an Fmoc group whether at
the Na-position or at the end of the side chain was
always accompanied with some activity. The first
*
Corresponding author. Present address: Neurochem, 275, Armand
Frappier Blvd., Laval, Que, Canada H7V 1B7; e-mail: abouzide@
neurochem.com
0960-894X/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2004.12.068

1510
A. Bouzide et al. / Bioorg. Med. Chem. Lett. 15 (2005) 1509–1513
Table 1. HIV protease inhibition data for N-protected amino acids8
the ornithine analogue 7 (vide infra). On the other hand,
the 10 lM potency displayed by inhibitor 8 containing a
naphtylsulfonamide moiety on the side chain was a wor-
thy indication that compounds attached with bulkier
groups at the Ne-position might exhibit some optimiza-
tion. In fact, compound 9 where the Fmoc group was
switched from the Na- to the Ne-position exhibited an
affinity 2-fold higher than its analogue 5. Based on these
valuable indications revealed by the commercially avail-
able amino acids, we started the preparation of our
Ôhome-madeÕ inhibitors. First we prepared compound
10, which is a derivative of 5 bearing an additional
isobutyl group. Interestingly, this inhibitor exhibited
an 8-fold increase in potency compared to 5. After sev-
eral modifications, we gratifyingly found that the addi-
tion of an isobutyl group at the Na-position while
keeping the Fmoc group at Ne-position led to a dra-
matic enhancement in potency resulting in the discovery
of the lead candidate 11.10 To the best of our knowledge,
this is the first example of a single amino acid that inhi-
bit HIV protease in nanomolar range. As we anticipated
and hoped,11 the D-isomer 12 displayed a potency 67
fold lower than 11 indicating that the natural L-stereo-
chemistry is the preferred one.
No
N-Protected amino acid
Ki (lM)a
CO2H
1
>100
CbzHN
CO2H
NHBoc
2
3
4
5
>100
28.0
21.5
17.4
TsHN
NHBoc
CO2H
CO2H
CO2H
FmocHN
FmocHN
FmocHN
NHBoc
NHCbz
NHTs
CH2O2H
H
N
NHTs
NH
6
7
14.2
26.2
FmocHN
CO2H
CO2H
NHTs
FmocHN
FmocHN
8
10.5b
8.0
NHNaph
The synthesis of 11 was completed in four steps starting
from the commercially available a-amino-caprolactam
13 (Scheme 1).12 Reductive N-alkylation of 13 in the
presence of isobutyraldehyde followed by N-tosylation
afforded the corresponding sulfonamide 14 in excellent
yield. Ring opening hydrolysis proceeded quantitatively
and without any detectable racemization to afford
amino acid 15. Treatment of 15 with FmocOSuc under
aqueous conditions gave the desired product 11 in high
yield. Initial attempts to use FmocCl reagent were
unfruitful and poor yields were obtained.
CO2H
9
TsHN
NHFmoc
CO2H
Ts
N
10
11
12
2.0
FmocHN
CO2H
Ts
Ts
0.005
0.33
N
N
NHFmoc
NHFmoc
CO2H
3. Structure–activity relationship
a All values are the average of at least two runs.
b Naph = 1-naphtyl–SO2.
To investigate the structure–activity relationships (SAR)
of this new HIV protease inhibitor series, we synthesized
and examined various analogues of 11 by modifying the
Na- and Ne-substituents as well as the side chain length
and the carboxylic acid function.
encouraging result was exhibited by the commercially
available Na-Fmoc-Ne-Boc-L-lysine 3, which displayed
a 28 lM Ki. Compounds with Ne-Cbz carbamate 4 or
tosylamide 5 instead of Boc group displayed the same
potency. The Na-Fmoc-Nx-tosylarginine 6, which was
reported during the progress of this project by Ortiz
de Montellano and co-workers9 showed a similar po-
tency than 5 and slight loss in activity was noticed in
3.1. Modifications of Na-substituents
Compounds 16–28 in Table 2 were prepared in the same
manner as for 11. Replacement of the isobutyl group in
11 with benzyl (16), 3-ethyl-butyl (17) or 3-methyl-butyl
CO2H
CO2H
O
O
O
Ts
N
H2N
Ts
N
Ts
NH
N
N
H
O
NH
c
d
a, b
HCl
NH2
11
13
14
15
Scheme 1. Reagents and conditions: (a) i-PrCHO, MeOH, NaCNBH3, pH 4; (b) TsCl, NEt3, CH2Cl2, 88% (two steps); (c) 6 M HCl, reflux; (d)
FmocOSuc, K2CO3 (1 M), THF 82% (two steps).

A. Bouzide et al. / Bioorg. Med. Chem. Lett. 15 (2005) 1509–1513
1511
Table 2. Structure and HIV Protease inhibitory activities of N-functionalized natural and unnatural amino acids8
O
OH
H
N
R2
R3
N
R1
(CH2)n
No
R1
R2
R3
n
Ki (nM)a
11
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
i-PrCH2–
C6H5CH2–
Et2CHCH2–
MeEtCHCH2–
c-C3H5CH2–
c-C5H9CH2–
i-PrCH2–
i-PrCH2–
i-PrCH2–
i-PrCH2
4-MeC6H4SO2
4-MeC6H4SO2
4-MeC6H4SO2
4-MeC6H4SO2
4-MeC6H4SO2
4-MeC6H4SO2
C6H5SO2
Fmoc
Fmoc
4
4
4
4
4
4
4
4
4
4
4
4
4
4
3
5
2
1
4
4
4
4
5.0
19.6
15.2
13.5
3.9
Fmoc
Fmoc
Fmoc
Fmoc
5.2
18.7
Fmoc
Fmoc
COC6H5
>300
257
54.9
4-t-BuC6H4SO2
1-Naphtyl-SO2
4-FC6H5SO2
Fmoc
Fmoc
i-PrCH2–
i-PrCH2–
i-PrCH2–
i-PrCH2–
i-PrCH2–
i-PrCH2–
i-PrCH2–
i-PrCH2–
i-PrCH2–
i-PrCH2–
i-PrCH2–
i-PrCH2–
Fmoc
Fmoc
17.4
42
2-ThiopheneSO2
4-NO2C6H4SO2
4-NH2C6H4SO2
4-MeC6H4SO2
4-MeC6H4SO2
4-MeC6H4SO2
4-MeC6H4SO2
4-MeC6H4SO2
4-MeC6H4SO2
4-MeC6H4SO2
4-MeC6H4SO2
Fmoc
Fmoc
7.2
2.1
Fmoc
Fmoc
180
>250
>250
>250
Fmoc
Fmoc
CO–CH2–9-Fluorene
CO–9-Fluorene
t-Boc
89.7
105
>250
>250
Cbz
a All values are the average of at least two runs.
O
(18) moieties resulted in an alteration in activity, while
slight improvement was noticed in the case of cycloprop-
ylmethyl group (19). Of particular interest, the compar-
ison between the benzenesulfonamide 21 and
benzenecarboxamide 22 showed that while the first
maintain some potency, the second lost completely its
activity, suggesting that in addition to its hydrogen bond
acceptor character, –SO2– group may provide an appro-
priate conformation essential for the potency of this
series. The utilization of more sterically demanding sul-
fonamide such as tert-butyl benzenesulfonamide 23 or
naphtylsulfonamide 24 resulted in substantial loss in
activity. An 8-fold decrease in affinity was noticed when
the tosyl was replaced with a thiophene group as in 26
while 2-fold improvement in potency was attained by
replacing the 4-Me group of 11 with 4-NH2 as in 28.
CO2Me
a, b, c, d
H2N
O
Ts
N
OH
38
37
e, f
CO2H
Ts
CO2Me
c, g, h Ts
N
NHFmoc
N
N3
31
39
Scheme 2. Reagents and conditions: (a) i-PrCHO, MeOH, NaCNBH3;
(b) TsCl, NEt3, CH2Cl2, 82% (two steps); (c) NaOH (1 M), THF; (d)
CH2N2, Et2O (77% two steps); (e) TsCl, NEt3, CH2Cl2; (f) NaN3,
DMF 66% (two steps); (g) H2, Pd/C, EtOAc; (h) FmocOSuc, K2CO3
(1 M), THF 73% (three steps).
3.2. Modification of the side chain length
Compound 32 was prepared in six steps from L-serine 40
as outlined in Scheme 3. Its esterification and reductive
N-alkylation furnished 41. Treatment of the amino alco-
hol in the presence of an excess of TsCl provided in one
pot the acrylate 42 in moderate yield. Michael addition
in the presence of dissolved ammonia in EtOH followed
by saponification and reaction with FmocOsuc afforded
racemate 32.
Compounds 29 and 30 were prepared, respectively, from
ornithine and racemic homolysine13 as described in Ref.
12. Compound 31 was prepared as described in Scheme
2. Reductive N-alkylation of the commercially available
c-lactone 37 followed by sulfonamidation, ring opening
hydrolysis and esterification in the presence of diazome-
thane afforded 38. Tosylation of the alcohol and dis-
placement of the resulting tosylate with sodium azide
gave the corresponding azide 39. Alkaline hydrolysis
of the methyl ester followed by a catalytic hydrogena-
tion gave the corresponding amino acid, which in the
presence of FmocOSuc afforded 31 in good yield.
The side chain length was found to be hypercritical for
the activity of this series of N-Fmoc protected amino
acid. In fact, when it was shortened by one carbon as
in the ornithine derivative 29, 36-fold loss in potency

1512
A. Bouzide et al. / Bioorg. Med. Chem. Lett. 15 (2005) 1509–1513
Table 3. Inhibition constant of lysine derivative modified on the
carboxyl group8
CO2Me
OH
CO2H
OH
a,b
HN
Ts
H2N
R
O O
S
41
40
N
NHFmoc
c
CO2Me
CO2H
d, e, f
Ki (nM)a
Ts
NHFmoc
No
R
N
N
32
42
11
43
44
45
46
47
48
49
50
51
52
COOH
5.0
>300
>300
105
COOMe
CHO
CH2OH
Scheme 3. Reagents and conditions: (a) TMsCl, MeOH, D; (b)
i-PrCHO, MeOH, NaCNBH3, pH 4, 89% (two steps); (c) TsCl,
NEt3, CH2Cl2, 45%; (d) NH3, MeOH; (e) NaOH (1 M), THF; (f)
FmocOSuc, K2CO3 (1 M), THF, 33% (three steps).
CONH2
CSNH2
107
127
CNHNH2
CONHOH
CONHNH2
Tetrazole
CH2NH2.HCl
280
245
254
>300
>300
compared to 11 was noticed. No activity, however, was
noticed at a concentration below 250 nM when the side
chain was shortened by two carbons as in 31 or by three
carbons as in 32. On the other hand, no inhibition was
detected when the side chain was extended by one more
carbon as in 30, suggesting that the side chain length is
optimal at four carbons as in the natural amino acid ly-
sine. The role of the side chain remains an enigma in the
absence of an X-ray of the enzyme–inhibitor complex.
One may speculate that in the case where the inhibitor
adopt an extended conformation in the active site, the
side chain may act only as a spacer assisting the Fmoc
to reach S02 and S03 subsites, while in the case of con-
strained conformation, the side chain may itself fit in
S01 subsite.
a All values are the average of at least two runs.
Table 4. Specificity of 11 and 28 as inhibitors of different aspartyl
proteases
Enzyme
IC50 of 11 (nM)
IC50 of 28 (nM)
HIV PR
Pepsin
Renin
5.0
2.1
5400
6500
15,600
4100
5200
10,300
Cathepsin
potency of this series as it was corroborated by the car-
boxamide 46 and thioamide 47, where the NH2 might be
engaged in a similar kind of hydrogen bond interaction
as an OH. In the case of compounds containing two
HBDs as in the amidine 48, N-hydroxyamide 49 and
hydrazide 50, moderate binding affinity was observed,
which could be explained by the contrasting interaction
induced by the second HBD. Although tetrazols are
known to mimic the carboxylic acid, no activity was ob-
tained in the case of 51. Finally, the positively charged
amine 52 did not show any activity.
3.3. Modification of Ne-Fmoc
The carbamate oxygen atom, as well as the –OCH2–
group of the Fmoc were essential for the high potency
as it is substantiated by the loss in activity of 33 and
34 (18-fold and 21-fold less active than 11). As we al-
ready mentioned, no inhibition was noticed when the
Fmoc group was replaced with small Boc- and Cbz-car-
bamate (35, 36). For more details on the modifications
achieved on Ne-Fmoc, see Ref. 14.
3.4. Modification of the carboxylic acid
Although this new series of inhibitors is not categorized
among those mimicking the substrate with a non-hydro-
lyzable transition-state isoster, we were very pleased to
find it very specific as HIV protease inhibitors in com-
parison to the mammalian aspartyl proteases pepsin, re-
nin and cathepsin (see Table 4). Furthermore, inhibitors
11 and 28 for example, did not show any cytotoxicity at
a concentration below 75 and 94 lM, respectively.
Finally the carboxylic acid group was modified by
various functional groups, such as methyl ester 43,
carboxaldehyde 44, alcohol 45, carboxamide 46, thio-
carboxamide 47, amidine 48, N-hydroxyamide 49,
hydrazide 50, tetrazol 51 and amine 52 (see Table 3).
The carboxylic acid was found so far the best function
to afford potent inhibition. Presumably the carbonyl
and the OH groups of the carboxylic acid may interact
with asp25 and asp250 in the active site of the enzyme.
In fact, its protection as a methyl ester 43 resulted in
complete loss in activity indicating the primordial role
of the OH group in the binding affinity. This was sub-
stantiated by the loss in activity of the aldehyde 44 lack-
ing the OH moiety, whereas the lysinol counterpart 45
maintained some potency albeit its 20-fold decrease in
activity compared to 11. The presence of a hydrogen
bond donor (HBD) was essential for the inhibitory
4. Conclusion
In summary, a screening assay program carried out on
commercially available N-protected amino acids showed
that Na-sulfonamide-Ne-Fmoc-L-lysine 9 displayed a
decent 8 lM inhibition constant. Addition of an isobu-
tyl group at Na-position allowed the discovery of the
lead candidate 11 exhibiting a 5.0 nM Ki. Furthermore
this new inhibitor series is very specific to HIV protease

A. Bouzide et al. / Bioorg. Med. Chem. Lett. 15 (2005) 1509–1513
1513
11. In terms of cost, 100 g of L
lysine cost $700 at Novabiochem.
12. The same compound was obtained readily from commer-
cially available L-Lys(e-Cbz)-OBn as described bellow.
-lysine cost $30 and 100 g of D-
enzyme and did not show any cytotoxicity below a
75 lM concentration.15
Acknowledgements
CO2Bn
CO2H
CO2Bn
i, ii
Ts
Ts
iii
The authors would like to thank Dr. Guy Sevigny and
Mr. Patrick Soucy for assistance with the assay, Mr.
Nicolas LeBerre and Dr. Fadil Dahhani for analysis of
compound purity.
N
N
H2N
AcOH
NH2
NHCbz
Ts
NHCbz
iv
CO2H
References and notes
N
1. http://www.unaids.org.
2. McQuade, T. J.; Tomaselli, A. G.; Liu, L.; Karacostas, V.;
Moss, B.; Sawyer, T. K.; Heinrikson, R. L.; Tarpley, W.
G. A. Science 1990, 247, 454.
3. (a) Leung, D.; Abbenate, G.; Fairlie, D. P. J. Med. Chem.
2000, 43, 305; (b) Babine, R. E.; Bender, S. L. Chem. Rev.
1997, 97, 1359.
NHFmoc
(i) i-PrCHO, MeOH, NaCNBH3. (ii) TsCl, NEt3, CH2Cl2,
3 days 88% (two steps). (iii) H2, Pd/C, AcOH. (iv)
FmocOSuc, K2CO3 (1 M), THF 82% (two steps).
13. Payn, L. S.; Boger, J. Synth. Commun. 1985, 15, 1277.
4. De Clercq, E. J. Clin. Virol. 2001, 22, 73.
5. (a) De Clercq, E. Nat. Drug Discov. 2002, 1, 13; (b) Molla,
A.; Granneman, G. R.; Sun, E.; Kempf, D. J. Antiviral
Res. 1998, 1; (c) Flexner, C. J. N. Eng. J. Med. 1998, 338,
1281.
´
14. (a) Stranix, B. R.; Sauve, G.; Bouzide, A.; Cote, A.;
Sevigny, G.; Yelle, J. Bioorg. Med. Chem. Lett. 2003, 13,
´
´
3601; (b) Stranix, B. R.; Sauve, G.; Bouzide, A.; Cote, A.;
Sevigny, G.; Yelle, J.; Perron, V. Bioorg. Med. Chem. Lett.
´
2004, 14, 3971. The first author in these two papers
published hastily the results in which he was involved
without waiting the preliminary work to be published.
6. Sham, H. L.; Chen, X. Exp. Opin. Invest. Drugs 1996, 5,
977.
7. The Ki values were determined with a fluorometric assay
according to Matayoshi test. Matayoshi, E. D.; Wang, G.
T.; Kraft, G. A.; Erickson, J. Science 1990, 247, 957.
8. Inhibitory value for clinically used saquinavir16 and
amprenavir17 in this assay are 0.18 and 0.25 nM,
respectively.
9. Rutenber, E. E.; De Voss, J. J.; Hoffman, L.; Stroud, M.
R.; Lee, K. H.; Alvarez, J.; McPhee, F.; Craik, C.; Ortiz de
Montellano, P. R. Bioorg. Med. Chem. 1997, 5, 1311.
1
15. All new compounds were characterized by H NMR, 13C
NMR and HRMS.
16. Roberts, N. A.; Martin, J. A.; Kinchington, D.; Broad-
hurst, A. V.; Craig, J. C.; Duncun, I. B.; Galpin, S. A.;
Handa, B. K.; Kay, J.; Krohn, A.; Redchaw, S.; Ritchie,
A. J.; Taylor, D. L.; Thomas, G. J.; Machin, P. J. Science
1990, 248, 358.
17. Kim, E. E.; Baker, C. T.; Dwyer, M. D.; Murcko, M. A.;
Rao, B. J.; Tung, R. D.; Navia, M. A. J. Am. Chem. Soc.
1995, 117, 1181.
´
10. Bouzide, A.; Sauve, G.; Stranix, B.; Sevigny, G.; Yelle, J.
U.S. Patent 6,455,587 B1; Pharmacor, Inc., USA, 2002.
´