Modular construction of quaternary hemiaminal-based inhibitor candidates and their in cellulo assessment with HIV-1 protease

Article abstract of DOI:10.1016/j.bmc.2013.06.018

Non-peptidomimetic drug-like protease inhibitors have potential for circumventing drug resistance. We developed a much-improved synthetic route to our previously reported inhibitor candidate displaying an unusual quaternized hemi-aminal. This functional group forms from a linear precursor upon passage into physiological media. Seven variants were prepared and tested in cellulo with our HIV-1 fusion-protein technology that result in an eGFP-based fluorescent readout. Three candidates showed inhibition potency above 20 μM and toxicity at higher concentrations, making them attractive targets for further refinement. Importantly, our class of original inhibitor candidates is not recognized by two major multidrug resistance pumps, quite in contrast to most clinically applied HIV-1 protease inhibitors.

Full text of DOI:10.1016/j.bmc.2013.06.018

Bioorganic & Medicinal Chemistry 21 (2013) 5407–5413
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Modular construction of quaternary hemiaminal-based inhibitor
candidates and their in cellulo assessment with HIV-1 protease
Guillaume Gros a, Lorena Martinez b, Anna Servat Gimenez a, Paula Adler a, , Philippe Maurin a,
Roland Wolkowicz c, Pierre Falson b, Jens Hasserodt a,
⇑
a Laboratoire de Chimie, Université de Lyon – ENS, 46 allée d’Italie, 69364 Lyon, France
b Drug Resistance Mechanism and Modulation Laboratory, BMSSI, UMR 5086 CNRS, Université Lyon 1 – IBCP, 7 passage du Vercors, 69367 Lyon, France
c Department of Biology, San Diego State University, 5500 Campanile Drive NLS304, 92182 San Diego, CA, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Non-peptidomimetic drug-like protease inhibitors have potential for circumventing drug resistance. We
developed a much-improved synthetic route to our previously reported inhibitor candidate displaying an
unusual quaternized hemi-aminal. This functional group forms from a linear precursor upon passage into
physiological media. Seven variants were prepared and tested in cellulo with our HIV-1 fusion-protein
technology that result in an eGFP-based fluorescent readout. Three candidates showed inhibition potency
Received 28 February 2013
Revised 30 May 2013
Accepted 6 June 2013
Available online 14 June 2013
above 20 lM and toxicity at higher concentrations, making them attractive targets for further refinement.
Keywords:
Cyclic urea
HIV-1 protease inhibitors
Quaternary hemiaminal
Transition-state isostere
Antiviral agents
Importantly, our class of original inhibitor candidates is not recognized by two major multidrug resis-
tance pumps, quite in contrast to most clinically applied HIV-1 protease inhibitors.
Ó 2013 Elsevier Ltd. All rights reserved.
Chemical synthesis
Cellular assay
1. Introduction
We have originated the exploration of an unusual functional group,
or rather, the unusual interaction of two functional groups, for its
Over the past 40 years, a limited number of functional groups
have been considered and explored for the simulation of the tran-
sition states of peptide bond hydrolysis.1 Only very few have been
considered in earnest for drug design targeting the class of aspartic
capacity to simulate electrostatic properties found in the transition
states stabilized by aspartic proteases.5–7 The weak interaction be-
tween a tertiary amine and a ketone gives rise to a highly polar and
stable molecular motif, provided the two functional groups belong
to the same molecule, that their cyclic interaction is pre-organized,
and that the molecule is dissolved in polar protic media (MeOH).8,9
Simple synthetic molecules showing this interaction have sporadi-
cally surfaced in the literature.10 Only a few examples of natural
products (alkaloids) showing this interaction are known.11–17 In
water, the pKa range of such species is estimated to range from 7
to 9,18–21 so that one can reasonably assume this functional group
to become protonated at physiological pH and adopt the status of
a quaternized hemiaminal, a highly unusual molecular species.
Any preliminary conclusion on the hypothesis that this group mim-
ics the electrostatics of the transition states of peptide hydrolysis
may be drawn from our ability to observe elevated affinities with
a model peptidase whose choice is motivated by a great supply of
SAR studies with inhibitor candidates, namely HIV-1 protease
(HIV-1 PR).2 In 2008, we reported on the diversity-orientated syn-
thesis of a hydrazino urea that displays this N–C@O interaction.22
This molecular constitution was chosen for two reasons: (a) because
the carbonyl contained in a urea motif found in the HIV-1 PR
inhibitor DMP45023,24 has been proven to interact directly via
proteases, including hydroxyethylene,
a,a-difluoroketones, phos-
phinates, and
a
-keto-carboxamides.2 More recently, the focus has
veered away from linear-shaped peptidomimetics incorporating
these groups, aiming more at drug-like test candidates.2,3 Expected
benefits of this measure include smaller size, reduced bio-degrada-
bility and preventing development of pathogen resistance.4
δ−O
H
R
R
δ+
R
N
N
N
R
R
O
⇑
Corresponding author. Tel./fax: +33 472728860.
E-mail address: jens.hasserodt@ens-lyon.fr (J. Hasserodt).
On leave from the University of Ottawa, Canada.
0968-0896/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.bmc.2013.06.018

5408
G. Gros et al. / Bioorg. Med. Chem. 21 (2013) 5407–5413
O
O
hydrogen-bonding with two protein flaps of HIV-1 protease cover-
ing the active site; this results in the extrusion of a key structural
water molecule usually found in X-ray crystal structures of com-
plexes of HIV-1 PR with linear, peptidomimetic inhibitors25 and
thus contributes to high binding affinity,1 and (b) in order to max-
imize chances of observing total passage to the interacted form in
methanol, and indeed this was proven by NMR studies. The mea-
sures that achieved this include the choice of (1) a urea that ensure
coplanarity of all substituents on the two nitrogens, (2) only two
possible orientations (linear or concave) for both displayed func-
tional groups, an aldehyde and a tertiary hydrazine nitrogen, and
(3) a distance between both functional groups that allows for clo-
sure of a (favored) six-membered ring. These entropy-related pre-
cautions allowed us to see almost quantitative ring formation
when transferring the compound to methanol, thus depriving the
equilibrium of the presence of any free aldehyde, a prerequisite
for consideration as an inhibitor/drug candidate. In 2009, we re-
ported the first in vitro inhibition data on this system with HIV-1
R = H : a
c
R = OH : a, b
OH
NHCbz
O
NHCbz
R
R
R
1 R = H
6 R = OTBS
R = H : d
R = OTBS : d, e
O
OMe
OMe
OMe
OMe
NHCbz
f
NHCbz
NH2
R
R
4 R = H
10 R = OTBS
2 R = H
7 R = OTBS
3 R = H
9 R = OTBS
Scheme 2. Preparation of
a-amino acid precursors before urea coupling; Cbz
protection; reagents and conditions: (a) MeOH, H2SO4 cat, reflux, 12 h (1: 92%); (b)
TBSCl, imidazole, DMF, rt, 12 h (6: 73% two steps); (c) DIBAL-H, CH2Cl2, À78 °C, 2 h;
(d) MeOH, PTSA, rt, 12 h (3: 84% two steps); (e) TBSCl, imidazole, DMF, rt, 12 h (9:
52% three steps); and (f) H2, 5% Pd/C, MeOH, 3 h (quant).
O
O
protease and obtained values down to 25 l
M.6 Here, we describe
c
a, b
OH
O
NH2
O
NHFmoc
the preparation of seven new variants on this system, including
the first example of an exhaustively substituted one, a feat we
had previously been unable to accomplish.
O2N
O2N
O2N
11
OMe
OMe
NHFmoc
OMe
OMe
d
e
NHFmoc
NH2
O2N
2. Results and discussion
2.1. Chemistry
O2N
12
13
14
Scheme 3. Preparation of
a
-amino acid precursor before urea coupling; Fmoc
protection; reagents and conditions: (a) MeOH, SOCl2, reflux, 12 h (quant); (b)
FmocCl, K2CO3, H2O–dioxane, rt, 12 h (79%); (c) DIBAL-H, CH2Cl2, À78 °C, 2 h; (d)
MeOH, PTSA, rt, 12 h (53% two steps); (e) piperidine, DMF, 40 min (80%).
Our hydrazino urea can be built up in a highly modular fashion
from three different building blocks (Scheme 1). To this end, re-
duced derivatives of any a-amino acid may serve if protected prop-
erly, as well as any N,N-disubstituted hydrazine, and further
alkylating reagents that will introduce substituents into the two
urea nitrogens found in the target structures.
We were very much interested in extending the substituent on
the amino aldehyde building block in order to possibly satisfy more
remote binding pockets of the protease for higher affinities. How-
ever, as the introduction of a carboxamide junction in the para
position of a phenylalanine side chain cannot be done prior to
DIBAL-H reduction, this junction has to be established post-urea
coupling by working initially with a nitro group. However, an aro-
matic nitro group is highly susceptible to catalytic hydrogenation,
and we therefore chose to work in this particular case with a Fmoc
protective group that is as resistant to DIBAL-H as is Cbz and that
can be removed with piperidine instead of hydrogenation. The cor-
responding five-step synthesis of the aminoacetal precursor (14,
Scheme 3) gives an overall yield of 33% from commercial unpro-
tected p-nitrophenylalanine.
The second building block (Scheme 1) was varied so as to allow
for the presentation of respectively, a cyclic N-piperidinyl substitu-
ent, a cyclic N-morpholinyl substituent, and an open-chain dimeth-
ylamine unit in the final target compounds. Hydrazines 15,22 16,
and 17 were prepared by a two-step procedure consisting of imine
formation from benzaldehyde followed by reduction with sodium
cyanoborohydride (80% average yield over two steps for the cyclic
substituents and 40% for the open-chain). These two sets of build-
ing blocks then served in the key coupling step towards a urea unit.
The three-step protocol (Scheme 4) using carbonyl-di-imidazol
(CDI) followed by activation/quaternization of the mono-reacted
intermediate with iodomethane and final reaction with the second
building block (tri-substituted hydrazine) was previously reported
by us.22 We thus prepared five out of the eight variants contained
in this work (ureic acetals 21, 22, 26, 27 and 29) in yields ranging
from 43% to 71%.
While we were previously unable to produce significant
amounts of the first generation of a hydrazino urea,22 we could de-
velop a highly efficient route to the protected
a-amino aldehyde
building blocks allowing us to run the synthesis at the multi-gram
scale.26 This became possible due to the choice of simple Cbz-pro-
tected
a-amino acids (Schemes 2 and 3) that are by far the most
affordable among all commercialized protected
a
-amino acids. Ini-
tial esterification of Cbz-protected phenylalanine and tyrosine, fol-
lowed by protection of the tyrosine hydroxyl with a TBS group gave
intermediates 1 and 6. Their partial reduction to aldehydes 2 and 7
with DIBAL-H is high-yielding while retaining the enantiopurity.27
These are easily transformed into their dimethyl acetals but the
conditions cause the loss of the TBS protective group in the tyro-
sine series. After its reprotection, both systems could be prepared
for introduction into the key urea coupling step by liberation of
the amino groups under mild catalytic hydrogenation conditions,
and produce aminoacetals 4 and 10 in respectively 77% and 38%
yield over four and six steps. Aminoacetal 4 in particular could thus
be produced at the level of 10 grams.
O
R
N
R
N
N
X
R'
Y
O
carbo-di-imidazol (CDI)
Br
O
OMe
OMe
NH2
R
R
Initially, we designed our cyclic ureas in accordance with in-
sights gained from the inhibitor DMP450 (see Section 1). However,
up to now, our three-step coupling protocol did not allow for the
preparation of a derivative of our N–CO interacted hydrazino ureas
that displays benzyl groups on both nitrogens of the urea motif as
N
or
+
+
NH2
X
Y
Scheme 1. Modular built-up of target compounds capable of showing the N–CO
interaction.

G. Gros et al. / Bioorg. Med. Chem. 21 (2013) 5407–5413
5409
MeO
N
H
OMe
OMe
O
OMe
OMe
R3
R4
OMe
OMe
NH2
N
NH2
N
a
HN
N
HN
O
R1
X
R2
b, c
X
HN
N
18
: X = H
19 : X = OTBS
20
N
4 : X = H
O
21
10
: X = OTBS
N
: X = NO2
14 : X = NO2
O
N
R5
R3
O
R4
b, c
N
16
R2
N
N
N
R3
O
R4
b, c
NHBn
NHBn
Me3SiCl, NaI
MeCN
b, c
R2
N
N
N
17
N
NHBn
15
O
R1
OMe
OMe
OMe
OMe
R1
MeO
OMe
R5
CDCl3
OMe
O
OMe
O
HN
O
HN
HN
R5
X
N
N
N
N
N
N
δ−O
26
O
H
23 : X = H
22
HO
N
H
24
: X = OTBS
R3
R4
δ+
R3
R4
25 : X = NO2
N
N
N
N
N
R1
R1
for 24 : d
R2
R2
for 25 : e
OMe
OMe
O
O
OMe
OMe
f
OMe
O
R5
OMe
O
MeOH-d4
H2O < pH 9
HN
O
R5
AcNH
HN
HN
HO
H2N
N
N
N
N
N
Scheme 6. Deprotection of all acetal precursors to hydrazino-aldehydes, constitu-
tional equilibrium in different solvents, and hypothetical form in neutral aqueous
media.
N
27
28
29
A simple TBS protected version was chosen and synthesized from
commercial 3-hydroxybenzoate.28
Scheme 4. Protected precursors of target compounds capable of showing the N–CO
interaction; reagents and conditions: (a) CDI, THF, rt, 15 min; (b) MeI, MeCN, rt,
12 h; (c) hydrazine, NEt3, CH2Cl2, rt, 12 h; (d) TBAF, THF, rt, 10 min; (e) H2, 5% Pd/C,
MeOH, 2 h; Ac2O, Et3N, DCM, rt, 30 min.
Acetals obtained by this process can be stored indefinitely,
while their corresponding free aldehydes run the risk of racemiza-
tion over longer storage times. In order to prepare our target alde-
hydes for study we ran the deprotection (35b was not deprotected
in view of the small amounts available) in anhydrous media using
Me3SiI generated in situ from Me3SiCl and sodium iodide
(Scheme 6). While six out of eight targets were obtained as oils,
two turned out to be solids (39 and 41 that have as a common fea-
ture a phenol substituent). LCMS monitoring of the deprotection
reactions revealed the formation of an intermediate that could in
two cases be isolated: a quaternized N/O acetal (Scheme 6). Its
salt-like character may be responsible for the total retention of
the two products at the base of a silicagel column if chromato-
graphic purification is attempted. The existence of this unusual
species may be explained with the high pre-organization that we
deliberately chose for this system. Aqueous workup however
hydrolyzes it to the desired aldehyde. This tendency to halt cleav-
age at the quaternized N/O acetal is most pronounced for the
exhaustively substituted targets 33b further proof for our strategy
of maximum pre-organization in order to achieve 100% cyclization
and the absence of free aldehyde in physiological media.
seen in DMP450 (Scheme 7); we believe steric hindrance to be
responsible for the lack of any reactivity of a secondary amine con-
tained in the aminoacetal (rather than a primary one as seen in
Scheme 4) towards the quaternized imidazolyl urea intermediate.
Even our measure of switching our coupling reagent from CDI to
carbonyl-di-triazol did not trigger the reaction. In the present re-
port, we therefore opted for the simultaneous introduction of
two identical pendent arms starting from a urea with only two
substituents (21, Scheme 5). Use of sodium hydride and an excess
of substituted or unsubstituted benzyl bromide did indeed yield
small amounts of the tetra-substituted ureas (33b, 34b, 35b)
alongside tri-substituted material. 33b and 35b were isolated
and characterized while the formation of 34b was only confirmed
by LCMS monitoring. The exclusive presence of tri-substituted ur-
eas having received a benzyl substituent in the hydrazine portion
proves that alkylation of this nitrogen is faster than the one on
the opposite side of the urea unit. For the synthesis of 34a and
34b, a protected version of meta-hydroxy benzyl bromide proved
necessary in order to resist the harsh condition of the substitution.
2.2. Folding of deprotected 39
As previously reported, the NCO interaction is observable only
in polar protic solvents. In the presence of a urea moiety, it has
been shown that the adoption of the required configuration for
NCO interaction requires time. In deuterated chloroform the pro-
ton NMR spectrum for compound 39 corresponds solely to the
non-interacted form, which is confirmed by the presence of a char-
acteristic aldehyde signal at 9.6 ppm (Fig. 1). On the other hand,
this aldehydic signal disappears over time when 39 is dissolved
in deuterated methanol; a new signal around 4.1 ppm takes the
place of that at 9.6 ppm. The kinetics are slow enough to be mon-
itored by NMR. This demonstrates that about 40% is cyclized after
25 min at room temperature, 60% conversion is reached after
35 min, and cyclization is complete in 4 h. In view of the similarity
of the structure to that of the benchmark structure published by
Waibel et al. (37) it was not surprising that 39 also folds to a
Br
OMe
OMe
OMe
OMe
Y
OMe
OMe
Y
+
HN
O
HN
O
N
O
HN
30 : Y = H,
N
N
N
N
N
31
: Y = OTBS,
21
32 : Y = OMe
Y
Y
33a = 23 Y = H
34a Y = OH
33b Y = H
34b Y = OH
35b
35a
Y = OMe
Y = OMe
Scheme 5. Outcome of a double alkylation reaction to reach an exhaustively
derivatized urea; reagents and conditions: 6 equiv bromide compound, 20 equiv
NaH, DMF, rt, 48 h.

5410
G. Gros et al. / Bioorg. Med. Chem. 21 (2013) 5407–5413
δ−O
δ−O
δ−
O
H
H
H
Nδ+ O
δ+
Nδ+
N
HN
N
HN
N
HN
N
O
O
O
36
37
38
δ−O
δ−O
δ−O
H
H
H
Nδ+
Nδ+
Nδ+
HN
N
HN
N
HN
N
HN
HO
O
O
39
41
40
O
O
OH
δ−O
δ−O
H
HO
N
OH
N
H
Nδ+
Nδ+
HN
N
N
N
O
O
O
H2N
NH2
O
42
43
DMP 450
Scheme 7. Formula of all eight target compounds as present in methanol, including 37 (reported previously) and DMP 450.
[IDV], µM
0,0001 0,001
100
0,01
0,1
1
10
100
#36
#37
#38
#39
#41
#42
#43
IDV
80
60
40
20
0
-20
-40
0
20
40
60
80
100
[compound], µM
Fig. 2. Assay response to inhibitors. SupT1 HIV-1 protease-Gal4 cells were grown
for 24 h in the presence of 1–100 M of compounds 36–43 or 0.0001–10 lM
l
Indinavir (IDV) as reference inhibitor, and then activated with 1 mg/mL doxocy-
cline. eGFP fluorescence was quantified 48 h later by flow cytometry.
Fig. 1. 1H spectra (500 MHz) of 39 in CDCl3, and at certain time intervals beginning
with initial dissolution in MeOH-d4. (j) aldehydic proton, (d) proton of the NCO
carbonyl moiety.
Fig. 2 (black circles), this inhibitor triggers eGFP expression via
HIV-1 protease inhibition at concentrations as low as 0.1 lM.
The same experiments carried out with compounds 36–43 re-
veal an HIV-1 protease inhibition pattern with 37, 41 and 42 above
20 lM. The other candidates did not display significant inhibition
activity. For comparison, in vitro inhibition data obtained by a clas-
sic, FRET-based spectrometer assay was included in the Supporting
information. In cellulo assays contribute an extra level of reliability
to the inhibition data in view of the variable quality of purified pro-
tease samples.
100%. This cyclization process is as reversible as has been reported
for 37: MeOH evaporation followed by a further dissolution in
CDCl3 gives rise to a proton NMR spectrum in full congruence with
open chain form.
2.3. Inhibition assays of HIV-1 Protease activity in T-Cells
To investigate the inhibitory activity of our compounds, we uti-
lized a new system for monitoring the HIV-1 protease activity in T-
cells, natural targets of HIV.29 In this system, the HIV-1 protease is
fused to Gal4 and placed under the control of a tetracycline-induc-
ible promoter. When expressed in T cells, the fusion protein binds
to a Gal4 Responsive Element which controls the expression of a
gene coding for Enhanced Green Fluorescent Protein (eGFP). When
the HIV-1 protease is active, it prevents the binding of GAL4. The
more the HIV-1 protease activity is inhibited by our molecule can-
didates, the more eGFP is expressed. Cells are grown for 24 h in the
presence of inhibitor, and then eGFP fluorescence is quantified by
flow cytometry. Indinavir is used as positive control; as shown in
2.4. Cytotoxicity and ABC pumps efflux
One obstacle to efficient chemotherapy and antiretroviral ther-
apeutic is the multidrug resistance (MDR) phenotype developed by
cells to prevent cytotoxicity. This mechanism involves ATP-binding
cassette (ABC) transporters30 such as P-gp, MRP131 and BCRP.32–34
Most of antiretroviral drugs used to treat HIV-infected patients are
unfortunately substrates of such efflux pumps.35,36 We thus evalu-
ated the efflux of the present compounds by these pumps by look-
ing for the compounds cytotoxicity and a cross-resistance induced
by P-gp or BCRP expression (Table 1).

G. Gros et al. / Bioorg. Med. Chem. 21 (2013) 5407–5413
5411
Table 1
Compounds cytotoxicity. Cell survival was estimated by MTT assays as described in the experimental section. Half-maximal cytotoxic concentrations, IC50
mean sd of at least three independent experiments
l
M, were the
#
P-gp
BCRP
Expressing cells (%)
Control cells (%)
Expressing cells (%)
Control cells (%)
Saquinavir
4937
3737
—
—
36
37
38
39
41
42
43
100 1.8
48.6 0.8
—
71.2 1.7
43.5 0.7
46.8 0.9
12.8 0.4
68.8 2.1
42.6 0.9
79.8 1.2
58.6 1.6
42.1 1.5
39.4 0.7
11.2 0.2
57.2 0.8
34.8 2.1
—
47.9 1.6
23.1 1.1
37.2 0.9
12.8 0.2
56.3 0.9
37.6 1.1
—
52.4 0.4
35.3 1.5
42.6 0.7
17.6 0.4
As shown, these compounds displayed either no (38), limited
(>30–50 M, 43)
cytotoxicity. Cytotoxicity of the more active compounds 37, 41
and 42 (Fig. 2) is in the same range of concentrations showing that
the inhibition power needs to be improved and the cytotoxicity
reduced.
The specific expression of P-gp or BCRP did not change the
cytotoxicity pattern of all products, indicating that multidrug
resistance pumps do not transport this type of compound, con-
trary to what is being observed for established, clinically applied
HIV-1 protease inhibitors such as Saquinavir (Table 1) or Indinavir
for which Pgp confers a twofold resistance ratio.38 Finally, we
checked if such compounds could inhibit the transport activity
of the same ABC pumps by looking at their effect on the efflux
of mitoxantrone mediated by P-gp or BCRP (Supporting informa-
tion Table 1). We could not observe any effect of them on the
mitoxantrone efflux suggesting that they probably are not inhibi-
tors of such pumps.
CH2Cl2 (50 mL, 50 mmol) over 40 min. The reaction mixture, mon-
itored by TLC, was quenched at À78 °C with MeOH (30 mL) after
2 h and then warmed to room temperature. The mixture was
poured into 300 mL of an ice cold 1.2 M HCl aqueous solution, ex-
tracted twice with 300 mL of CH2Cl2 and the combined organic
phases were then washed with 300 mL brine, dried over MgSO4
and concentrated under vacuum. Because of the presence of a chi-
ral center in alpha of the aldehyde, the next step was performed
immediately with the crude product.
lM, 36, 37, 39, 41 and 42) or rather high (ꢀ10
l
4.1.2. (S)-(1-(4-Hydroxy)benzyl-2,2-dimethoxy-ethyl)-carbamic
acid benzyl ester (8)
The crude 7 (23.2 mmol) was dissolved in 180 mL of anhydrous
MeOH, and p-toluenesulfonic acid monohydrate (3.0 g, 15.8 mmol)
was added. The reaction was stirred overnight, and then most of
the solvent was removed under vacuum. 100 mL of a solution of
aqueous saturated NaHCO3 was added, and the mixture was ex-
tracted twice with 100 mL of CH2Cl2. The combined organic layers
were dried over MgSO4, and the solvent was evaporated under vac-
uum. Purification by flash chromatography (gradient 1:9 to 3:7,
EtOAc:Cyclohexane) gave 8 as a white solide (5.67 g, 61%). 1H
NMR (500 MHz, CDCl3) d 7.30 (m, 5H), 7.04 (d, J = 8.0 Hz, 2H),
6.73 (d, J = 8.0 Hz, 2H), 5.07 (m, 2H), 4.20 (d, J = 2.8 Hz, 1H), 4.09
(m, 1H), 3.45 (s, 3H), 3.40 (s, 3H), 2.87 (dd, J = 5.8 Hz, JAB = 14.1 Hz,
1H), 2.71 (dd, J = 5.8 Hz, JAB = 14.1 Hz, 1H). 13C NMR (127 MHz,
CDCl3) d 156.5, 154.8, 136.4, 130.3, 128.5, 128.1, 127.9, 115.4,
104.9, 68.9, 55.8, 55.7, 52.4, 35.3. HRMS (ESI) calcd for C19H23-
NNaO5 [M+Na]+ 368.1468, found 368.1463.
3. Conclusion
In this report, we significantly improved synthetic access to the
previously reported 37 that allows the investigator to have access
to this compound at the gram scale. We demonstrated the quick
access to a number of substitution variants and thus the modular-
ity of our synthetic route. Our in cellulo assay of HIV-1 PR activity
allowed us to position the inhibitory potency of our molecule can-
didates with respect to two established and clinically applied
inhibitors; the values for our candidates have to be interpreted in
light of the considerably reduced size compared to the benchmark
inhibitors (60%) and the associated decrease in interacting surface.
Among our new molecule variants, 41 and 42 are promising hits, as
they display a significant inhibition potential and toxicity profile
that encourage further refinement. Significantly, they are not
transported out of cells by multidrug efflux pumps, quite in con-
trast to clinically applied inhibitors. Future attempts to obtain an
X-ray diffraction structure of an enzyme-inhibitor complex may
furnish the molecular insights to shift the inhibitory potency into
the nanomolar range.
4.1.3. General coupling procedure
A solution of amine in anhydrous THF (1 mL per mmol) was
added dropwise to a stirred suspension of CDI (1.1 equiv) in
anhydrous THF (1 mL per mmol) at room temperature. After
30 min the solvent was evaporated, and the residue redissolved
in CH2Cl2 (5 mL per mmol) and washed twice with water (2 mL
per mmol). The organic phase was dried over MgSO4, and the sol-
vent evaporated under vacuum to give pure imidazole
carboxamide.
The imidazole carboxamide compound was dissolved in anhy-
drous acetonitrile (1.5 mL per mmol) and iodomethane (4 equiv)
was added. The reaction was stirred overnight at room tempera-
ture, then the solvent was evaporated, and the yellow orange oil
was dried under vacuum. The residue was redissolved in dry CH2-
Cl2 (5 mL per mmol), and the hydrazine (1 equiv) and triethyl-
amine (1 equiv) were added. After stirring for 24 h at room
temperature, the reaction was quenched by adding saturated aque-
ous NaHCO3 (5 mL per mmol), and the mixture was extracted CH2-
Cl2 (twice with 5 mL per mmol). The combined organic extracts
were dried over MgSO4, and the solvent was removed under vac-
uum. After purification by flash chromatography (1:4 to 1:1,
EtOAc:Cyclohexane), the urea compound was obtained.
4. Experimental
4.1. Synthesis
Further synthetic protocols and analytical data of the target
compounds are contained in the Supplementary information.
4.1.1. (S)-(1-(4-tert-Butyldimethylsilyloxy)benzyl-2-oxo-ethyl)-
carbamic acid benzyl ester (7)
To a cold (À78 °C) solution of 6 (10.3 g, 23.2 mmol) in dry CH2-
Cl2 (100 mL) was added dropwise a 1.0 M solution of DIBAL-H in

5412
G. Gros et al. / Bioorg. Med. Chem. 21 (2013) 5407–5413
4.1.4. (S)-Imidazole-1-carboxylic Acid (1-(p-tert-
butyldimethylsily-loxy)benzyl-2,2-dimethoxy-ethyl)-amide
(19)
(m, 3H; NCH2Ph + NHCH), 3.10 (dd, J = 6.2 Hz, JAB = 14.0 Hz, 1H;
CHAHBPhenol), 3.02 (dd, J = 7.0 Hz, JAB = 14.0 Hz, 1H; CHAHBPhenol),
2.75 (m, 1H; NNCH2piperidine), 2.60 (m, 1H; NNCH2pip), 2.50 (m, 2H;
NNCH2pip), 1.61 (m, 3H; CH2pip), 1.45 (m, 2H; CH2pip), 1.00 (m, 1H;
CH2pip). 13C NMR (127 MHz, CDCl3) d 200.9 (CHO), 158.3 (NCON),
155.7 (ArPhenolCOH), 139.8 (ArPhenylCCH2N), 130.3 (ArPhenol), 128.4
(2ArPhenyl), 127.4 (2ArPhenyl), 126.8 (ArPhenyl), 115.7 (ArPhenol), 60.5
(NHCH), 53.6 (NNCH2pip), 53.4 (NNCH2pip), 42.5 (NCH2Ph), 34.7
(CH2Phenol), 26.4 (CH2pip), 26.3 (CH2pip), 23.1 (CH2pip). HRMS
(ESI) calcd for C22H27N3NaO3 [M+Na]+ 404.1945, found 404.1941.
Yellow oil, yield: 86%. 1H NMR (500 MHz, CDCl3) d 8.06 (s, 1H),
7.29 (m, 1H), 7.09 (m, 3H), 6.79 (d, J = 8.2 Hz, 2H), 4.35 (m, 1H),
4.26 (d, J = 2.7 Hz, 1H), 3.49 (s, 3H), 3.40 (s, 3H), 2.93 (dd,
J = 7.0 Hz, JAB = 14.4 Hz, 1H), 2.90 (dd, J = 7.6 Hz, JAB = 14.4 Hz,
1H), 0.98 (s, 9H), 0.19 (s, 6H). 13C NMR (127 MHz, CDCl3) d 154.5,
148.6, 135.9, 130.5, 130.1, 129.6, 120.3, 115.8, 104.1, 56.0, 55.7,
53.7, 35.3, 25.7, 18.2, À4.4.
4.1.5. (S)-1-Benzyl-3-(1-(p-tert-butyldimethylsilyloxy)benzyl-
2,2-dimethoxy-ethyl)-1-piperidin-1-yl-urea (24)
4.2. Cell culture
From 19 + 15. Colorless oil, yield: 82%. 1H NMR (500 MHz,
CDCl3) d 7.27 (m, 2H), 7.19 (m, 3H), 7.12 (d, J = 8.5 Hz, 2H), 6.78
(d, J = 8.5 Hz, 2H), 6.64 (d, J = 9.0 Hz, 1H; NH), 4.62 (d, JAB = 16.4 Hz,
1H), 4.50 (d, JAB = 16.4 Hz, 1H), 4.31 (d, J = 3.7 Hz, 1H), 4.28 (m, 1H),
3.48 (s, 3H), 3.47 (s, 3H), 3.00 (dd, J = 5.0 Hz, JAB = 13.8 Hz, 1H), 2.74
(m, 2H), 2.50 (m, 2H), 2.39 (m, 1H), 1.63 (m, 3H), 1.51 (m, 2H), 1.01
(m, 10H), 0.19 (s, 6H). 13C NMR (127 MHz, CDCl3) d 157.9, 154.0,
140.6, 131.2, 130.4, 128.3, 127.3, 126.5, 119.8, 105.8, 56.2, 54.9,
53.5, 53.2, 52.2, 42.1, 34.7, 26.5, 25.7, 18.2, À4.4. HRMS (ESI) calcd
for C30H47N3NaO4Si [M+Na]+ 564.3228, found 564.3221.
Human T-cell line SupT1, PR/Gal4 and Gal4 was from Dr. R.
Wolkowicz, San Diego, California, USA. Cells were maintained in
complete RPMI 1640 media with Stable Glutamine (PAA laborato-
ries) supplemented with 10% fetal bovine serum (PAA laborato-
ries), 1% penicillin/streptomycin (PAA laboratories). Cells were
incubated at 37 °C in humidified 5% CO2.
The NIH3T3, drug-sensitive, parental cell line and that transfec-
ted with human MDR1-G185, NIH3T3 Pgp,39 derived from NIH
Swiss mouse embryo cultures, were from ATCC and used as previ-
ously described.40 Human embryonic kidney (HEK293) cell lines
transfected with either BCRP (HEK293-BCRP) or the empty vector
(HEK293-pcDNA3) were obtained as previously described.38 Cells
were maintained in Dulbecco’s modified Eagle’s medium (DMEM
high glucose, PAA laboratories), supplemented with 10% fetal bo-
vine serum (FBS, PAA laboratories), 1% penicillin/streptomycin
(PAA laboratories). The NIH3T3 Pgp growth medium was supple-
mented with 60 ng/mL colchicine. The HEK293-pcDNA3 and
HEK293-BCRP was supplemented with 0.75 mg/mL G418 (PAA lab-
oratories). Cells were incubated at 37 °C in humidified 5% CO2.
4.1.6. General procedure for the deprotection of all acetal
precursors to hydrazino-aldehydes
To a solution of the acetal precursor (63 mg, 0.14 mmol) in
anhydrous acetonitrile (20 mL per mmol) were added NaI (3 equiv)
and TMSCl (2 equiv). The reaction was stirred at room temperature
for 1.5 h and was then quenched by adding NaHCO3 (saturated
aqueous solution, 20 mL per mmol). The mixture was extracted
with CH2Cl2 (three times 60 mL per mmol), and the combined or-
ganic extracts were washed with Na2S2O3 (saturated solution,
100 mL per mmol) and then with water (60 mL per mmol). The or-
ganic phase was dried over Na2SO4 and the solvent was evaporated
under vacuum to give the crude aldehyde.
4.3. Flow cytometry
4.3.1. HIV-1 protease activity in T-cells
Doxycycline (Sigma Aldrich) was dissolved in water at 10 mg/
mL stock concentration and stored at À20 °C. Indinavir was a gift
from Merck and Saquinavir from Roche. They were dissolved in
100% DMSO at 10 mM stock concentration and stored at À20 °C.
In a 96 well plate, 50,000 cells were incubated for 10 min in
growth medium and then inhibitors were added at varying concen-
trations and incubated for at least 10 min. Cells were expressing the
4.1.7. (S)-1-Benzyl-3-(1-(4-hydroxy)benzyl-2,2-dimethoxy-
ethyl)-1-piperidin-1-yl-urea (27)
Compound 26 (415 mg, 0.77 mmol) was dissolved in 25 mL THF
and 1.5 mL of a 1 M solution of tetra-n-butylammonium fluoride in
THF was added dropwise. The reaction was stirred for 10 min and
the solvent was evaporated. The residue was redissolved in CH2Cl2
(20 mL) and washed with water (20 mL). The aqueous phase was
extracted with CH2Cl2 (10 mL), and the combined organic phase
was dried over MgSO4 and concentrated under vacuum. After puri-
fication by flash chromatography (gradient 1:4 to 1:1,
EtOAc:Cyclohexane), 27 (332 mg, quant) was obtained as a color-
less gel-oil. 1H NMR (500 MHz, CDCl3) d 7.25 (m, 2H), 7.17 (m,
3H), 7.02 (d, J = 8.2 Hz, 2H), 6.77 (d, J = 9.1 Hz, 1H; NH), 6.69 (d,
J = 8.5 Hz, 2H), 4.61 (d, JAB = 16.0 Hz, 1H), 4.54 (d, JAB = 16.0 Hz,
1H), 4.29 (d, J = 3.3 Hz, 1H), 4.26 (m, 1H), 3.48 (s, 3H), 3.46 (s,
3H), 2.95 (dd, J = 5.0 Hz, JAB = 14.0 Hz, 1H), 2.70 (m, 2H), 2.51 (m,
2H), 2.39 (m, 1H), 1.63 (m, 3H), 1.53 (m, 2H), 1.01 (m, 1H). 13C
NMR (127 MHz, CDCl3) d 158.3, 155.3, 140.2, 130.2, 129.3, 128.3,
127.2, 126.6, 115.4, 105.9, 56.2, 55.3, 53.4, 53.3, 52.8, 42.3, 35.0,
26.4. HRMS (ESI) calcd for C24H33N3NaO4 [M+Na]+ 450.2363, found
450.2355.
HIV-1–eGFP fusion protein with 1 lg/ml doxycycline. After 48 h,
cells were washed with PBS and the expression of eGFP was quan-
tified by flow cytometry carried out with a FACS Calibur cytometer
(Becton Dickinson). Excitation and emission were set up at 488 and
530 nm, respectively. Data were collected on CellQuest Pro (version
4.0) software and then exported to FlowJo for analysis.
4.4. Cytotoxicity assays
Parental cell line (NIH3T3 and HEK293-pcDNA3) and transfec-
ted with each pump (NIH2T3 Pgp and HEK293-BCRP) were incu-
bated for 72 h in the presence of increasing concentrations of the
novel HIV-1 protease inhibitors, and then cells survival was esti-
mated by a 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium
bromide (MTT) (Sigma–Aldrich) colorimetric assay as detailed
previously.38
4.1.8. (S)-1-Benzyl-3-(1-(4-hydroxy)benzyl-2-oxo-ethyl)-1-
piperidin-1-yl-urea (39)
Acknowledgments
From acetal 27 (32 mg, 0.07 mmol). Purification by flash chro-
matography (gradient 1:4 to 1:1, EtOAc:Cyclohexane), gave 39 as
a colorless oil (17 mg, 59%). 1H NMR (500 MHz, CDCl3) d 9.71 (s,
1H; CHO), 7.28 (m, 5H; ArPhenyl), 7.04 (d, J = 7.8 Hz, 1H; NH), 6.98
(d, J = 8.6 Hz, 2H; ArPhenol), 6.71 (d, J = 8.6 Hz, 2H; ArPhenol), 4.61
We thank T. ANDRIEU and S. DUSSURGEY (Tour Inserm CERVI,
Lyon France) for their help with the flow cytometer analysis. We
would like to acknowledge the kind supply with indinavir by
Merck Sharp & Dohme and GF120918 by GlaxoSmithKline.

G. Gros et al. / Bioorg. Med. Chem. 21 (2013) 5407–5413
5413
[16]. Cimino, G.; Destefano, S.; Scognamiglio, G.; Sodano, G.; Trivellone, E. Bull. Soc.
Chim. Belg. 1986, 95, 783.
[17]. Becker, M. H.; Chua, P.; Downham, R.; Douglas, C. J.; Garg, N. K.; Hiebert, S.;
Jaroch, S.; Matsuoka, R. T.; Middleton, J. A.; Ng, F. W.; Overman, L. E. J. Am.
Chem. Soc. 2007, 129, 11987.
[18]. Leonard, N. J.; Fox, R. C.; Oki, M. J. Am. Chem. Soc. 1954, 76, 5708.
[19]. Hine, J.; Kokesh, F. C. J. Am. Chem. Soc. 1970, 92, 4383.
[20]. McCrindle, R.; McAlees, A. J. J. Chem. Soc., Chem. Commun. 1983, 61.
[21]. Kirby, A. J.; Komarov, I. V.; Bilenko, V. A.; Davies, J. E.; Rawson, J. M. Chem.
Commun. 2002, 2106.
[22]. Waibel, M.; Hasserodt, J. J. Org. Chem. 2008, 73, 6119.
[23]. DeLucca, G. V.; Ericksonviitanen, S.; Lam, P. Y. S. Drug Discovery Today 1997, 2,
6.
This work was supported by the Centre National de la Recher-
che Scientifique (CNRS) and University of Lyon 1 (UMR5086), and
the French Ministry of Research (EA3741, Lyon 1). It was funded
by the National Research Agency (ANR-06-BLAN_0420, ANR-06-
PCVI-0019-01, ANR piribio09_444706), the Association pour la
Recherche sur le Cancer (ARC), and the Ligue Nationale Contre le
Cancer (Labellisation 2009–14).
G.G. and L.M. were recipients of doctoral fellowships from the
Ministère de la Recherche (France) and the Région Rhône-Alpes
(France), respectively.
[24]. Nair, A. C.; Jayatilleke, P.; Wang, X.; Miertus, S.; Welsh, W. J. J. Med. Chem.
2002, 45, 973.
[25]. Dunitz, J. D. Science 1994, 264, 670.
Supplementary data
[26]. Jurczak, J.; Golebiowski, A. Chem. Rev. 1989, 89, 149.
[27]. Dinh, T. Q.; Armstrong, R. W. J. Org. Chem. 1995, 60, 8118.
[28]. Kwon, S.; Myers, A. G. J. Am. Chem. Soc. 2005, 127, 16796.
[29]. Hilton, B. J.; Wolkowicz, R. PLoS ONE 2010, 5, e10940.
[30]. Dean, M.; Hamon, Y.; Chimini, G. J. Lipid Res. 2001, 42, 1007.
[31]. Cole, S. P.; Bhardwaj, G.; Gerlach, J. H.; Mackie, J. E.; Grant, C. E.; Almquist, K.
C.; Stewart, A. J.; Kurz, E. U.; Duncan, A. M.; Deeley, R. G. Science 1992, 258,
1650.
Supplementary data associated (further synthetic protocols,
NMR data, protocol for ABC pump inhibition) with this article
can be found, in the online version, at http://dx.doi.org/10.1016/
j.bmc.2013.06.018.
References and notes
[32]. Allikmets, R.; Schriml, L. M.; Hutchinson, A.; Romano-Spica, V.; Dean, M.
Cancer Res. 1998, 58, 5337.
[33]. Doyle, L. A.; Yang, W.; Abruzzo, L. V.; Krogmann, T.; Gao, Y.; Rishi, A. K.; Ross,
D. D. Proc. Natl. Acad. Sci. U.S.A. 1998, 95, 15665.
[1]. Tyndall, J. D. A.; Nall, T.; Fairlie, D. P. Chem. Rev. 2005, 105, 973.
[2]. Brik, A.; Wong, C. H. Org. Biomol. Chem. 2003, 1, 5.
[3]. Bursavich, M. G.; Rich, D. H. J. Med. Chem. 2002, 45, 541.
[4]. De Clercq, E. Nat. Rev. Drug Disc. 2007, 6, 1001.
[5]. Gautier, A.; Pitrat, D.; Hasserodt, J. Bioorg. Med. Chem. 2006, 14, 3835.
[6]. Waibel, M.; Pitrat, D.; Hasserodt, J. Bioorg. Med. Chem. 2009, 17, 3671.
[7]. Hasserodt, J.; Gautier, A.; Barbe, R.; Waibel, M. Attanazi, O. A.; Spinelli, D.,
Eds.; Targets in Heterocyclic Systems, Societa Chimica Italiana: Rome, 2009;
Vol. 13, pp 1–26.
[34]. Miyake, K.; Mickley, L.; Litman, T.; Zhan, Z.; Robey, R.; Cristensen, B.; Brangi,
M.; Greenberger, L.; Dean, M.; Fojo, T.; Bates, S. E. Cancer Res. 1999, 59, 8.
[35]. Weiss, J.; Haefeli, W. E. In International Review of Cell and Molecular Biology;
Jeon, K. W., Ed.; Academic Press, 2010; Vol. 280, pp 219–279. T2.
[36]. Kis, O.; Robillard, K.; Chan, G. N. Y.; Bendayan, R. Trends Pharmacol. Sci. 2010,
31, 22.
[37]. Kim, A. E.; Dintaman, J. M.; Waddell, D. S.; Silverman, J. A. J. Pharmacol. Exp.
Ther. 1998, 286, 1439.
[38]. Kim, R. B.; Fromm, M. F.; Wandel, C.; Leake, B.; Wood, A. J.; Roden, D. M.;
Wilkinson, G. R. J. Clin. Invest. 1998, 101, 289.
[8]. Leonard, N. Rec. Chem. Prog. 1956, 17, 243.
[9]. Leonard, N. J. Acc. Chem. Res. 1979, 12, 423.
[10]. Zhang, X. A.; Song, D. T.; Lippard, S. J. J. Org. Chem. 2008, 73, 734.
[11]. Gadamer, J. Arch. Pharm. 1920, 258, 148.
[12]. Kermack, W.; Robinson, R. J. Chem. Soc. 1922, 121, 427.
[13]. Huisgen, R.; Wieland, H.; Eder, H. Liebigs Ann. Chem. 1949, 561, 193.
[14]. Boit, H. G.; Paul, L. Chem. Ber. Recl. 1954, 87, 1859.
[15]. Birnbaum, G. I. J. Am. Chem. Soc. 1974, 96, 6165.
[39]. Cardarelli, C. O.; Aksentijevich, I.; Pastan, I.; Gottesman, M. M. Cancer Res.
1995, 55, 1086.
[40]. Arnaud, O.; Koubeissi, A.; Ettouati, L.; Terreux, R.; Alamé, G.; Grenot, C.;
Dumontet, C.; Di Pietro, A.; Paris, J.; Falson, P. J. Med. Chem. 2010, 53, 6720.