Discovery of novel peptidomimetics as irreversible CHIKV NsP2 protease inhibitors using quantum mechanical-based ligand descriptors

Article abstract of DOI:10.1111/cbdd.12621

Chikungunya virus (CHIKV) is a mosquito-borne alphavirus. Recent outbreaks of CHIKV infections have been reported in Asia, Africa, and Europe. The symptoms of CHIKV infection include fever, headache, nausea, vomiting, myalgia, rash, and chronic persistent arthralgia. To date, no vaccines or selective antiviral drugs against this important emerging virus have been reported. In this study, the design, synthesis, and antiviral activity screening of new topographical peptidomimetics revealed three potential prototype agents 3a, 4b, and 5d showing 93-100% maximum inhibition of CHIKV replication in cell-based assay having EC₉₀ of 8.76-9.57 μg/mL. Intensive molecular modeling studies including covalent docking, lowest unoccupied molecular orbital energies, and the atomic condensed Fukui functions calculations strongly suggested the covalent binding of peptidomimetics 3a, 4b, and 5d to CHIKV nsP2 protease leading to permanent enzyme inactivation via Michael adduct formation between α/β-unsaturated ketone functionality in our designed peptidomimetics and active site catalytic cysteine1013. Furthermore, small molecular weight peptidomimetics 3a and 4b satisfied the Lipinski rule of five for drug-likeness and showed promising intestinal absorption and aqueous solubility via computational admet studies making them promising hits for further optimization. Covalent docking of the inhibitor 3a with the nsP2 enzyme showing the proposed Michael adduct formed between inhibitors 3a and the side chain of catalytic Cysteine1013.

Full text of DOI:10.1111/cbdd.12621

Chem Biol Drug Des 2015; 86: 1518–1527
Research Article
Discovery of Novel Peptidomimetics as Irreversible
CHIKV NsP2 Protease Inhibitors Using Quantum
Mechanical-Based Ligand Descriptors
Eman M. El-labbad1, Mohammed A. H. Ismail1,
Key words: ADMET, atomic condensed Fukui functions,
Chikungunya virus, covalent docking, lowest unoccupied
molecular orbital energies, peptidomimetics
Dalal A. Abou El Ella1, Marawan Ahmed2,
Feng Wang3, Khaled H. Barakat2,4,5 and Khaled
A. M. Abouzid1,*
Received 22 April 2015, revised 25 June 2015 and accepted
for publication 10 July 2015
1Department of Pharmaceutical Chemistry, Faculty of
Pharmacy, Ain Shams University, Abbassia, Cairo 11566,
Egypt
2Faculty of Pharmacy and Pharmaceutical Sciences,
Chikungunya virus (CHIKV) is a mosquito-borne alphavirus
University of Alberta, Edmonton, AB, Canada
3Molecular Model Discovery Laboratory, Department of
Chemistry and Biotechnology, Faculty of Science,
Engineering and Technology, Swinburne University of
Technology, Melbourne, Vic. 3122, Australia
4LiKaShing Institute of Virology, University of Alberta,
Edmonton, AB, Canada
transmitted to humans primarily by Aedes aegypti and
Aedes albopictus (1). After the first isolation of CHIKV in
Tanzania in 1953 (2), a few localized outbreaks involving
relatively few individuals in Africa and Asia were reported
(3). Starting from 2005, CHIKV infection unexpectedly
arose as a major epidemic human pathogenic infection
throughout tropical regions such as central and southern
Africa, India, Indonesia, and Malaysia. CHIKV is now fre-
quently being transmitted into other temperate regions of
Asia, Europe, Australia, and America (4,5). Clinical mani-
festations of CHIKV infection include fever, headache, nau-
sea, vomiting, myalgia, rash, and chronic persistent
arthralgia similar to rheumatoid arthritis arthralgia (6,7).
Researches were carried out to test the efficacy of some
antiviral drug candidates against CHIKV (8–12). But up-to-
date, no vaccines or a selective antiviral drug against this
potential emerging epidemic infection has been reported.
5LiKaShing Applied Virology Institute, University of Alberta,
Edmonton, AB, Canada
*Corresponding author: Khaled A. M. Abouzid,
khaled.abouzid@pharma.asu.edu.eg
Chikungunya virus (CHIKV) is a mosquito-borne alpha-
virus. Recent outbreaks of CHIKV infections have been
reported in Asia, Africa, and Europe. The symptoms of
CHIKV infection include fever, headache, nausea,
vomiting, myalgia, rash, and chronic persistent arthral-
gia. To date, no vaccines or selective antiviral drugs
against this important emerging virus have been
reported. In this study, the design, synthesis, and
antiviral activity screening of new topographical pep-
tidomimetics revealed three potential prototype agents
3a, 4b, and 5d showing 93–100% maximum inhibition
of CHIKV replication in cell-based assay having EC90
of 8.76–9.57 lg/mL. Intensive molecular modeling
studies including covalent docking, lowest unoccupied
molecular orbital energies, and the atomic condensed
Fukui functions calculations strongly suggested the
covalent binding of peptidomimetics 3a, 4b, and 5d to
CHIKV nsP2 protease leading to permanent enzyme
inactivation via Michael adduct formation between a/
b-unsaturated ketone functionality in our designed
peptidomimetics and active site catalytic cys-
teine1013. Furthermore, small molecular weight pep-
tidomimetics 3a and 4b satisfied the Lipinski rule of
five for drug-likeness and showed promising intestinal
absorption and aqueous solubility via computational
ADMET studies making them promising hits for further
optimization.
CHIKV genome consists of a linear, positive-sense, single-
stranded RNA molecule of about 11.8 kb (13). The non-
structural proteins, required for viral replication, are
encoded in the 50 two-thirds of the genome, while the
structural genes are within the 30 one-third (1). There are
four non-structural genes, designated nsP1–4, which are
necessary for viral replication; each has specific function
during the replication process. The nsP1 protein is
required for initiation of synthesis of minus-strand RNA
and also functions as a methyltransferase to cap the geno-
mic and subgenomic RNAs during transcription (14). The
nsP2 protein is made of three domains, the first containing
helicase, RNA triphosphatase, and nucleoside triphos-
phatase activities (15,16) whereas the second and third
domains are a papaine-like protease and a non-functional
methyltransferase, respectively (17). The functions, roles,
and activities of the nsP3 protein are less well understood.
It may be involved in the transcription process at an early
1518
ª 2015 John Wiley & Sons A/S. doi: 10.1111/cbdd.12621

Irreversible CHIKV NsP2 Protease Inhibitors
stage of the infection (18). The nsP4 protein contains the
RNA-dependent RNA polymerase, involved in genome
replication and transcription (19). Alphavirus nsP2 protease
domain is considered an interesting antiviral target
because it is essential for alphavirus replication. NsP2 is
cysteine protease plays crucial role in the cleavage and
polyprocessing of the non-structural polyprotein. The inhi-
bition of nsP2 protease will lead to prevention of the cleav-
age of viral polyprotein chain, halting the viral replication.
Recently, interactions between CHIKV nsP3 or nsP4 pro-
teins with the host stress-pathway chaperone HSP-90
protein has been reported (20). Furthermore, nsP2 pro-
tease exhibits some degree of sequence specificity among
alphaviruses making it an attractive broad spectrum
antiviral target.
electrostatic and hydrophobic properties of binding site. The
designed inhibitors were superimposed or aligned with
AGGY tetrapeptide using FIELDALIGN software (Cresset
ꢀ
group, Cambridgeshire, UK). The alignment pattern was
matched with the rational design strategy explained earlier
(Figure S1B,C). All designed compounds showed reason-
able field points, similarity index and shape similarity with the
AGGY tetrapeptide natural substrate. The idea behind this
type of alignment is that two molecules which both perfectly
aligned to each other based on electric field properties tend
to make similar interactions with the binding sites.
The designed compounds were synthesized according to
Schemes 1–3. Evaluation for antiviral activity in a CHIKV
virus-cell-based assay was carried out according to previ-
ously reported method (10). Our study suggested that
these inhibitors may act as covalent inhibitors for the
CHIKV nsP2 protease based on the fact that the a/b-un-
saturated ketone functionality adopted to replace the scis-
sile amide bond in the natural substrate (AGGY
tetrapeptide) had been previously reported in many cova-
lent cysteine protease inhibitors. The a/b-unsaturated
ketone functionality tend to form Michael adduct with
active cysteine in the binding site leading to irreversible
alkylation of this residue and permanent enzyme inactiva-
tion (26). Furthermore, extensive molecular modeling stud-
ies including covalent docking study, lowest unoccupied
molecular orbital (LUMO) energies and the atomic con-
densed Fukui functions calculations were performed.
These calculations had been previously reported in deter-
mining the reactivity in biological systems and studying the
binding of wide range of covalent inhibitors (27), including
cysteine protease inhibitors (28).
This study focused on nsP2 protease as target for design-
ing new small peptidomimetics as prototype covalent inhi-
bitors of Chikungunya virus nsP2 protease enzyme. CHIKV
polynucleotide sequence of non-structural polyprotein
(nsP; uniprot database; accession number: Q8JUX6) was
investigated and showed that nsP polyprotein has 2474
amino acid and contains four proteins, nsP1 (1–535), nsP2
(536–1333), nsP3 (1334–1863), and nsP4 (1864–2474).
The catalytic active site for nsP2 is composed of cys-
teine1013 and histidine1083. NsP2 protease hydrolyzes
peptide bonds at three different sites, the first is the junc-
tion between nsP1 and nsP2 (535–536), the second site is
nsP2–nsP3 junction (1333–1334), and the third site is
nsP3–nsP4 junction (1863–1864) (21). Biochemical studies
of the alphavirus nsP2 protease showed considerable dif-
ferences in cleavage efficiency at the nsP12, nsP23, and
nsP34 polyprotein cleavage sites (22–24) with the sub-
strate preference of both full-length nsP2 and the isolated
protease domain being nsP34 > nsP12 ≫ nsP23 (24).
Even though 15 residues flanking the scissile bond were
required for optimal activity, controlled biochemical studies
showed that only substrate residues P10–P4 specifically
impacted substrate recognition by the nsP2 protease (22).
Therefore, nsP34 tetrapeptide natural substrate at P10–P3
residues Ala1861-Gly1862-Gly1863-Tyr1864 (AGGY) (25)
was selected as prototype lead compound for designing
topographical peptidomimetic inhibitors. This was achieved
by performing the several rational design strategies (Fig-
ure S1A). Tyrosine1864 at P10 residue was replaced with
scaffold (A). Scaffold (A) aryl/heteroaryl groups were
designed to mimic interaction pattern of phenolic group of
tyrosine1864 at P10 residue while a/b-unsaturated ketone
functionality was adopted to replace scissile amide bond in
substrate for possible reaction with the active site cys-
teine1013 forming a Michael adduct causing an irreversible
alkylation of this residue leading to permanent enzyme
inactivation (26). Glycine1863 at P1 residue was replaced
with pseudopeptide composed of ketone group of a/b-un-
saturated ketone attached to two carbon atom linker.
Scaffold (B) was designed as dipeptide analogue of
glycine1862 and alaine1861 at P2–P3 residues having
Methods and Materials
Chemistry
The designed compounds were synthesized according to
Schemes 1–3. All the starting materials were purchased
from Sigma-Aldrich (Stockholm, Sweden). Melting points
were determined on Stuart Scientific apparatus and are
uncorrected. Reactions were monitored using thin-layer
chromatography (TLC), performed on 0.255 mm silica gel
plates, with visualization under UV light (254 nm). FT-IR
spectra were recorded on a Perkin–Elmer spectropho-
tometer. 1H-NMR spectra were recorded on a Perkin–
Elmer 300 MHz spectrometer using CDCl3 or DMSO-d6
as a solvent; chemical shifts (d) were reported in parts per
million (ppm) downfield from TMS; multiplicities are abbre-
viated as follows: s, singlet; d, doublet; q, quartet; m,
multiplet; dd, doublet of doublet; br, broad. MS spectra
were recorded on Finnigan Mat SSQ 7000 (70 eV) mass
spectrometer and Triple Quadrupole LC/Ms/Ms mass
spectrometer API 200 (AB Sciex Instrument). Elemental
analyses were performed at the Micro-analytical Centre,
Cairo University. Detailed experimental procedures are
explained in supplementary information.
additional aryl/heteroaryl
2
groups to investigate
Chem Biol Drug Des 2015; 86: 1518–1527
1519

El-labbad et al.
Scheme 1: Synthesis of target compounds. (i) Levulinic acid/morpholine/Gl.acetic acid/dry benzene, reflux for 8 h; (ii) Benzylamine or
phenylethylamine/Et3N/ethyl chloroformate/CH2Cl2, stirring 10–12 h; (iii) L-tyrosine ethyl ester/Et3N/ethyl chloroformate/CH2Cl2, stirring
10–12 h; (iv) L-tryptophan ethyl ester/Et3N/ethyl chloroformate/CH2Cl2, stirring 10–12 h; and (v) LiOH/MeOH/H2O, 50 °C for 4 h.
Virus-cell-based antiviral assay
of L-glutamine, and 5 mL of sodium bicarbonate) at a ratio
of 1:4 and grown for 7 days in 150 cm2 tissue culture
flasks Techno Plastic Products). The assay plates were
returned to the incubator for 6–7 days (37 °C, 5% CO2,
and 95–99% relative humidity), a time at which maximal
virus-induced cell death or cytopathic effect (CPE) is
observed in untreated, infected controls. Subsequently,
the assay medium was aspirated, replaced with 75 lL of
a 5% MTS (Promega) solution in phenol red-free medium,
and incubated for 1.5 h. Absorbance was measured at a
wavelength of 498 nm (Safire2; Tecan); optical densities
(OD values) reached 0.6–0.8 for the untreated, uninfected
controls). Raw data were converted to percentage of
Throughout the experiment, Vero (African green monkey
kidney) was used as cell line for Chikungunya virus strain
899. In a 96-well microtiter plate (Falcon, BD), serial dilu-
tions of compounds were prepared in assay medium
[MEM Rega3 (cat. no. 19993013; Invitrogen) supple-
mented with 2% FCS (Integro), 5 mL of 200 mM L-glu-
tamine, and 5 mL of 7.5% sodium bicarbonate]. Fifty
microliters of a 49 virus dilution in assay medium was
added, followed by 50 lL of a cell suspension. This sus-
pension, with a cell density of 25 000 cells/50 lL, was
prepared from a Vero cell line subcultured in cell growth
medium (MEM Rega3 supplemented with 10% FCS, 5 mL
1520
Chem Biol Drug Des 2015; 86: 1518–1527

Irreversible CHIKV NsP2 Protease Inhibitors
Scheme 2: Synthesis of target compounds. (i) POCl3/DMF; (ii) Ethyl bromide or benzyl chloride/K2CO3/CH3COCH3; (iii) Levulinic acid/
Morpholine/Gl.acetic acid/dry benzene/reflux 8 h; (iv) L-tyrosine ethyl ester/Et3N/ethyl chloroformate/CH2Cl2/stirred 10–12 h; and (v)
L-tryptophan methyl or ethyl ester/Et3N/ethyl chloroformate/CH2Cl2/stirred 10–12 h.
Scheme 3: Synthesis of target compounds. (i) Indole/I2 & ethanol stirring at RT for 6 h and (ii) L-tyrosine ethyl ester or L-tryptophan
methyl ester/Et3N/ethyl chloroformate/CH2Cl2, stirring 10–12 h.
controls, and the EC50 (50% effective concentration or
concentration that is calculated to inhibit virus-induced cell
death by 50%) and CC50 (50% antimetabolic concentra-
tion or concentration that is calculated to inhibit the overall
cell metabolism by 50%) were derived from the dose–
response curves. All assay conditions producing an
Chem Biol Drug Des 2015; 86: 1518–1527
1521

El-labbad et al.
antiviral effect exceeding 50% were checked microscopi-
cally for minor signs of a CPE or adverse effects on the
host cell (i.e. altered cell morphology). A compound is only
considered to elicit a selective antiviral effect on virus repli-
cation when, following microscopic quality control, at least
at one concentration of compound no CPE nor any
adverse effect is observed (image resembling untreated,
uninfected cells) (10).
binding of wide range of covalent inhibitors (27), including
cysteine protease inhibitors (28).
The ‘response of molecular fragment’ approach (44) for
calculating the atomic condensed Fukui function evaluates
the atomic condensed Fukui function of a given atom (k) in
a molecule as follows:
fkþ ¼ PkðN þ 1Þ ꢀ PkðNÞ
and
Covalent docking
The crystal structure of the Chikungunya virus nsP2 pro-
tease was downloaded from the protein data bank (PDB
ID: 3TRK). The protein preparation wizard in Maestro
(29,30) was used to repair the protein, adding missing
residues and saturate with hydrogens, similar to what have
been described in our previous studies (31,32). The
geometries of the ligands were optimized at the RM1 (33)
level of theory as implemented in the SCHRODINGER package
(34). RM1 optimized ligands were used for the primary
non-covalent docking with Glide (data not shown) (35,36).
Docking has been carried out using the Glide standard
precision mode (Glide-SP) with the default settings as
implemented in SCHRODINGER. The ultimate goal of this non-
covalent docking step was to explore the accessibility of
the binding pocket.
fkꢀ ¼ PkðNÞ ꢀ PkðN ꢀ 1Þ
where f+ is the susceptibility of atom k to nucleophilic
attack; fꢀ is the susceptibility of atom k to electrophilic
attack; P is the atomic electronic population; and N is the
total number of electrons of the system.
The ligand were subjected to a more robust geometry
optimization at the B3LYP/6-31G* level and a single-point
calculations at the B3LYP/6-31 + G* level of theory to cal-
culate the LUMO energies and extract the natural bond
orbital (NBO) populations using the GAUSSIAN package (27).
For cysteine1013, the active cysteine, we assumed a neg-
atively charged species (cysteine anion) and the calcula-
tions were carried out at the same level of theory and
basis set as the studied ligands. The presence of both
polarized and diffused function on the basis was selected
to give consistent results for all studied molecular species
(neutral and anions).
Based on previous studies (37,38) docking was restricted
to binding site residues asparagine1011, cysteine1013,
tyrosine1079, asparagine1082, histidine1083, trypto-
phan1084, and leucine1205. For the covalent docking, the
CovalentDock server (39) was used. The CovalentDock
method is well parameterized against wide range of cova-
lently bound inhibitors and achieves excellent results in
reproducing the native poses of several covalent drug–pro-
tein complexes (39). The method also gives superior
results, estimated by calculating the root mean square
deviation (RMSD) of the reproduced complexes from the
native complexes (average RMSD = 1.68), over other
commonly used covalent docking methods such as Gold
Results and Discussion
Chemistry
The designed compounds were synthesized according to
Schemes 1–3. The synthesis of (E)-6-Aryl-4-oxohex-5-
enoic acid 2a–d, Scheme 1, was adopted according
general procedure previously reported (45,46). This
involved d-condensation of the appropriate aldehyde 1a–d
with levulinic acid using catalytic amounts of morpholine
and glacial acetic acid. The same procedure was adopted
in Scheme 2 to prepare unreported intermediates (E)-6-
(1H-indol-3-yl)-4-oxo-hex-5-enoic acid 11 and (E)-6-(sub-
stituted indol-3-yl)-4-oxo-hex-5-enoic acid 12, 13. These
intermediates were prepared after condensation of 1-in-
(average
RMSD = 2.49 A) (39).
RMSD = 3.69)
or
Autodock
(average
ꢀ
Quantum mechanical-based ligand descriptors
From the basis of conceptual density functional theory
(40,41) and the frontier molecular orbital (FMO) theory (42),
Michael addition reaction can be treated as a nucleophilic
attack of the highest occupied molecular orbital (HOMO)
electrons pair from the Michael donor (Sꢀ1 of cys-
teine1013) to the LUMO to form the covalent enzyme–in-
hibitor complex. In this regard, the LUMO energies and
the atomic condensed Fukui functions (43) of the Michael
acceptor species (the inhibitors) are two good metrics to
explain the reactivity and hence the inhibitory activity of the
irreversible covalent inhibitors. Particularly, the atomic con-
densed Fukui function has several applications in deter-
mining the reactivity in biological systems and studying the
dole-3-carbaldehyde
8 or 1-alkyl-1H-indole-3-carbalde-
hyde 9, 10 with levulinic acid. The general synthesis of final
compounds (E)-N-phenylalkyl-6-aryl-4-oxohex-5-enoic acid
amides 3a, b, 4a, b Scheme 1, (S, E)-3-substituted-2-[6-
aryl-4-oxo-hex-5-enoylamino]-propionic acid ethyl ester
5a–d, 6a–c, Scheme 1 and (S, 5E)-3-(substituted)-2-[6-(in-
dol-3-yl)-4-oxo-hex-5-enoylamino] propionic acid alkyl
esters 14–16, Scheme 2 was carried out through formation
of amide bond between corresponding acids 2a–d, 11–13
with different amines (viz; benzylamine and phenylethy-
lamine) or with amino acid esters (viz, L-tryptophan methyl/
ethyl ester and L-tyrosine ethyl ester), using ethyl chlorofor-
1522
Chem Biol Drug Des 2015; 86: 1518–1527

Irreversible CHIKV NsP2 Protease Inhibitors
mate and triethylamine, adopting the mixed anhydrides
method (47). Elucidation of the structures of the target
amides 5a–d, 6a–c, and 14–16 was proved by different
analytical and spectral data. The IR spectra showed the
absence of the broad hydroxꢀyl1 group of the carboxylic
(38). The second challenge is that the inhibitor should be
able to react with the activated cysteine residue to form
the Michael adduct causing an irreversible alkylation of this
residue. In that case, the activated cysteine residue (anion)
will be the Michael donor, and the irreversible inhibitor will
be the Michael acceptor that covalently binds the protein
causing a permanent enzyme inactivation (26). Figure S3
illustrates the proposed mechanism of action of the
designed inhibitors. The b-carbon of the active a/b-unsatu-
rated ketone functional group acts as an electrophile
(Michael acceptor, E+) that can be attacked by a nucle-
ophile (Michael donor, Nuꢀ), which is the active cys-
teine1013 anion. Table 1 gives the in vitro measured
inhibitory activity of the synthesized inhibitors. As can be
seen in the table, the most active inhibitor identified in vitro
is compound 3a, with a maximum inhibition of 100% at
68.2 lg/mL and an EC90 of 16.4 lg/mL. Compound 4b
achieves a maximum inhibition of 100% at a slightly lower
concentration than compound 3a, and is given by
59.3 lg/mL and an EC90 of 16.6 lg/mL. From the 5a–d
series, compound 5d achieves a maximum inhibition of
93.8% at a concentration of 42.4 lg/mL and an EC90 of
18.7 lg/mL. Our results show that small molecule inhibi-
tors achieve a better success in inhibiting the nsP2 activity,
perhaps due to their superior accessibility to the binding
pocket.
~
function at v ꢁ 3200–2800 cm and the presence of the
carboxylic C=O fuꢀn1ction at shorter wave number at
~
v ꢁ 1630–1600 cm corresponding to the formed amidic
~
C=O and appearance of amidic NH at v ꢁ 3320–
3200 cmꢀ1. The MS of compounds showed molecular ions
1
peaks matched with their molecular weights. H-NMR data
showed the vinylic hydrogens at d ꢁ 6.6 and 7.6 ppm,
with J constant = 16 Hz, proving that the stereochemistry
of such alkene group was the (E) isomer (48). Final com-
pounds 18 and 19 were prepared as illustrated in
Scheme 3. The unreported Michael adduct (R S)-6-(1H-In-
dol-3-yl)-6-(4-methoxyphenyl)-4-oxo-hexanoic acid 17 was
prepared via Michael addition of indole to (5E)-6-(4-meth-
oxyphenyl)-4-oxo-hex-5-enoic acid 2b in non-stereo-speci-
fic manner. A modified Wang et al. (49) procedure was
performed using iodine as catalyst. This procedure suc-
cessfully produced the titled compound 17 with 73% yield.
The structure of the Michael adduct 17 was proved by the
different analytical and spectral data. The IR spectra of 17
showed the absence of the C=C at ύ ꢁ 1490 cmꢀ1. The
MS showed its molecular ion peaks at 351.4 amu. 1H
NMR data were consistent with the structure as indicated
1
by the absence of the vinyl hydrogens in its H NMR spec-
To determine how the new synthesized inhibitors could
covalently bind the nsP2 protease, the CovalentDock web
server was used (39). The online tool is able to identify the
active b-carbon from the a/b-unsaturated group that acts
as the Michael acceptor; however, the sequence number
of the active cysteine residue must be supplied by the
user; thus, a primary knowledge of the binding site is
important. The corresponding two-dimensional (2D) and
the three-dimensional (3D) figures of the inhibitor–enzyme
covalent complexes formed after in silico covalent docking
study using CovalentDock web server are given in (Fig-
ure S4) for the aforementioned compounds (3a, 4b and
5d). As we can see in the figure, all ligands are able to
bind the cysteine1013 residue covalently and partially pro-
truding inside the binding pocket. The extra phenyl group
in compound 5d seems to facilitate optimal orientation of
the active b-carbon so that the Michael addition reaction
can take place. This might partially explain the higher
activity of compound 5d, which achieves a maximum inhi-
bition of 93.8%, compared to similar inhibitors in the same
series that achieves a much lower maximum inhibition of
16% (compound 5a), 33.7% (compound 5b) and 39.3%
(compound 5c). The DG values of the inhibitors estimated
with CovalentDock do not give strong correlation with
in vitro measured inhibitory activities and will not be dis-
cussed further.
trum and the presence of new duplets and triplet signals at
d 3.2 and 4.8 ppm, respectively, corresponding to the –
CH–CH2–CO moiety. Amide bond between carboxylic
group of Michael adduct (R S)-6-(1H-indol-3-yl)-6-(4-meth-
oxyphenyl)-4-oxo-hexanoic acid 17 and amino group of L-
tyrosine ethyl ester or L-tryptophan methyl ester was
formed to give final compounds (6RS,2S)-3-(4-hydrox-
yphenyl)-2-[6-(1H-indol-3-yl)-6-(4-methoxyphenyl)-4-oxo-hex-
anoylamino]-propionic acid ethyl ester 18 and (6RS,2S)-3-
(1H-indol-3-yl)-2-[6-(1H-indol-3-yl)-6-(4-methoxyphenyl)-4-oxo-
hexanoylamino]-propionic acid methyl ester 19. Based on
different analytical and spectra data, the structures of the
amides 18 and 19 were proved. The FT-IR spectra of 18
and 19 showed the disappearance of the broad stretching
vibration band of hydrꢀo1xyl group of the carboxylic function
~
at v = 3300–2900 cm with the shifting of their carbonyl
groups vibrꢀa1tion to shorter wave number at
~
v ꢁ 1650 cm
. The MS of compounds 18 and 19
showed matched molecular ion peaks with their molecular
weights, at 542.40 and 549.00, respectively.
The covalent binding mode of the new inhibitors
Designing covalent inhibitors for CHIKV cysteine nsP2 pro-
tease is a challenging task due to several reasons; first,
the designed inhibitor should be capable of accessing the
binding site, which is a tunnel-shaped pocket near the
protein surface (Figure S2). Substrate access to this
Based on the fact that classical molecular modeling tech-
niques such as covalent docking do not include bond
breaking or bond formation, a more accurate method to
estimate the expected potency of the inhibitors had been
pocket is restricted by
a b-hairpin at residues ala-
nine1080–histidine1083, as has been shown previously
Chem Biol Drug Des 2015; 86: 1518–1527
1523

El-labbad et al.
Table 1: Antiviral evaluation in a
LUMO
energy
Chikungunya virus-cell-based assay
together with the estimated free energy of
binding (DG) from CovalentDock, the f+ of
the Michael acceptor active carbon
(denoted as f+CMichael) and the LUMO
energies, the EC90, and DG values are
given only for the three most active inhibi-
tors identifies from the cell-based assay
a
b
Compound Max, %
no.
Conc.
EC50
EC90
DGc
inhibition (lg/mL) (lg/mL) (lg/mL) (kcal/mol) f+CMichaeld (kcal/mol)d
3a
3b
4a
4b
5a
5b
5c
5d
6a
6b
6c
7
100
1.36
81
100
16
68.2
12.4
65.1
59.3
10.1
8.76 16.4
>100
10.9
8.94 16.6
>100
ꢀ8.92
0.152
0.147
0.137
0.143
0.131
0.134
0.145
0.103
0.131
0.158
0.129
0.161
0.141
0.131
0.101
–
ꢀ59.457
ꢀ47.251
ꢀ52.592
ꢀ46.869
ꢀ52.133
ꢀ47.986
ꢀ39.207
ꢀ53.596
ꢀ52.918
ꢀ46.398
ꢀ39.627
ꢀ49.128
ꢀ44.177
ꢀ44.051
ꢀ43.348
–
–
–
–
–
ꢀ7.85
–
–
–
–
33.70
235
206
42.4
1.91 >100
>100
–
–
39.30
93.80
ꢀ1.91
75.10
ND
2.40
86.50
35.30
0
>100
9.57 18.7
ꢀ9.48
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
44.6
ND
1.9
216
1.8
183
11.8
ND
>100
52.3
<0.8
14
15
16
18
19
>100
1.47 >100
181 >100
0.52
8.32
–
–
LUMO, lowest unoccupied molecular orbital; ND, not determined.
aConcentration at which 50% inhibition of virus replication is observed.
bConcentration at which 90% inhibition of virus replication is observed.
cDG values are obtained from the CovalentDock webserver.
dAt the B3LYP/6-31 + G* level of theory.
carried out using quantum mechanical-based ligand
descriptors. From the basis of conceptual density func-
tional theory (40,41) and the FMO theory (42), Michael
addition reaction can be treated as a nucleophilic attack of
the HOMO electrons pair from the Michael donor (Sꢀ1 of
cysteine1013) to the LUMO to form the covalent enzyme–
inhibitor complex. In this regard, the LUMO energies and
the atomic condensed Fukui functions (43) of the Michael
acceptor species (the inhibitors) are two good metrics to
explain the reactivity and hence the inhibitory activity of the
irreversible covalent inhibitors. The atomic condensed
Fukui function f+ of the Michael acceptor (active b-carbon,
reflected by their f+CM values, such that the f+CM value in
5c is given by 0.145, which is slightly higher than in 5b
where it is given by 0.134. The additional phenyl ring in
compound 5d results in higher in vitro activity (Max. Inhibi-
tion 93.8% and EC50 = 9.57 lg/mL). Similar results are
obtained in the 6a–c series where compound 6b achieves
the highest in vitro activity (Max. Inhibition 75.1% and
EC50 = 11.8 lg/mL) and the highest f+CM as well
(f+CM = 0.158), and also in compounds 14, 15, and 16.
Please note that we did not consider calculating the f+CM
or LUMO energies for compounds 18 and 19 as they do
not have an active Michael acceptor site.
C
Michael) on the corresponding inhibitor is given in Table 1,
the table also gives the LUMO orbital energies of the cor-
responding inhibitors. The f+ value of the active Michael
acceptor carbon will be denoted as f+CM. As we can see
in the table, atomic condensed Fukui function f+CM
achieves an excellent correlation with the in vitro measured
inhibitory activities. With the exception of 5d in the 5a–d
series, the value of f+CM in any close series of ligands is
always higher in the more active inhibitor, another strong
The values of the LUMO energies of the inhibitors calcu-
lated at the B3LYP/6-31 + G* level of theory are also
listed in Table 1. The LUMO energy values show that two
inhibitors from the most active inhibitors (3a and 5d)
indeed have the lowest LUMO energies (LUMO-
3a = ꢀ59.457 kcal/mol and LUMO-5d = ꢀ53.596 kcal/
mol). In the present case, LUMO energies can provide
good enrichment; however, the LUMO energies represent
only a static picture of the actual wave function and thus
cannot be considered a true good metric in situations in
which a chemical reaction takes place, that is wave func-
tion reorganization. Nevertheless, the LUMO energies can
be considered as a primary metric especially at the early
next phase of development in these new series of inhibi-
tors to screen large number of compounds where
calculating the f+CM values will be very time consuming.
We also observed that the LUMO energies calculated
using the faster RM1 semi-empirical method achieves sim-
ilar performance. It is noted, however, that the two param-
eters used to assess the potency of the compounds. That
proof for the proposed mode of action of the synthesized
+CM
series of compounds. For example, the f
value of in
compound 3a is 0.152, which is higher than that in 3b
(0.147). Also, in the case of the 4a/4b pair of ligands, the
f+CM is in favor of 4b (f+CM 0.137, Max. Inhibition 100%
and EC50 = 8.94 lg/mL) over 4a (f+CM 0.143, Max. Inhibi-
tion 81% and EC50 = 10.9 lg/mL). In the 5a–c series, the
in vitro results are perfectly explained by the f+CM values,
that is the least active inhibitor from this series, compound
5a, has the lowest f+CM (0.131). Also, compound 5c (Max
% inhibition = 39.30%) is slightly more active than com-
pound 5b (Max % inhibition = 33.70%), which is also
1524
Chem Biol Drug Des 2015; 86: 1518–1527

Irreversible CHIKV NsP2 Protease Inhibitors
is, LUMO energies and f+CM cannot accurately reproduce
the trend for very different molecular scaffolds. Neverthe-
less, they can be valuable metrics to predict the potency
for focused compound libraries or for library enrichment.
that this compound is too soluble. The analysis of ADMET
model results (Table S1) revealed that all biologically active
peptidomimetics 3a, 4a–b, 5d, 6b, and 14 showed marked
improvement of the predictive intestinal absorption level
(good moderate absorption) compared to very poor
absorption level of tetrapeptide AGGY. Although this was
on the expense of aqueous solubility which was decreased
compared to tetrapeptide AGGY, still compound 3a and
4a–b showed good aqueous solubility while compounds
5d, 6b, and 14 were still soluble but with low predictive
aqueous solubility.
To investigate the orbital distribution of the FMO, the orbi-
tal characters of the frontier interacting orbitals, that is
HOMO of cysteine1013 and LUMO of the inhibitors were
identified. Figure S5 displays the orbital characters of the
HOMO orbital of cysteine1013 together with the LUMO
orbitals of the three most active inhibitors identified in vitro,
that is 3a, 4b, and 5d. As can be seen in the figure, the
HOMO orbital of cysteine1013 is dominated by exclusive
contribution from the Sꢀ1 anion. On the other hand, the
LUMO orbitals of the given inhibitors has a much more dif-
fused character with a major contribution from the p-elec-
trons on the active double bond and with other
contribution from the surrounding group but with different
phases and amplitudes.
Conclusions
In summary, new series of inhibitors that covalently bind
and inactivate the Chikungunya viral nsP2 protease have
been designed, synthesized, and in vitro tested. Several
inhibitors from the new series show micromolar inhibitory
activities, which are strongly, correlated with the reactivity
of the Michael acceptor active carbons. The reactivity of
the Michael acceptor active carbon has been quantified by
calculating some reactivity indices, such as the atomic
condensed Fukui functions and LUMO orbital energies
and characters. Based on the recent success of the sub-
strate envelope hypothesis in designing several protease
inhibitors (53,54), small active inhibitors from the new ser-
ies (3a and 4b) represent excellent hits for further opti-
mization, as they will be less prone to viral resistance due
to binding site residues’ mutations. This is in contrast to
large size active inhibitors, such as 5d that partially
protrudes outside the binding pocket.
In silico human intestinal absorption & aqueous
solubility study
To assess the druggability of the new series of inhibitors, in
silico druggability assessment has been carried out. Small
molecular weight peptidomimetics 3a, 4b, and 5d, satisfied
the Lipinski rule of five (RO5) for drug-likeness compared to
tetrapeptide natural substrate AGGY Table S1. The RO5
deals with orally active compounds and defines four simple
physicochemical parameter ranges (MWT ≤ 500, log p ≤ 5,
H-bond donors ≤ 5, H-bond acceptors ≤ 10) associated
with 90% of orally active drugs that have achieved phase II
clinical status (50). Further investigation was carried out
using Accelrys Discovery Studio software ADMET module to
predict human intestinal absorption after oral administration
& aqueous solubility. Accelrys Discovery Studio software AD-
MET human intestinal absorption model was developed
using 182 compounds in the training set, with descriptors
that include AlogP98 and 2D polar surface area (PSA_2D)
(51). There are four prediction levels for ADMET absorption
ranging from 0 to 3 described as the following: (0) for good
absorption, (1) for moderate absorption, (2) for poor
absorption, and (3) for very poor absorption. The second AD-
MET model used was ADMET aqueous solubility model. This
model uses linear regression to predict the solubility of
compounds in water at 25 °C. The model was generated
using a dataset containing 775 compounds (molecular
weight between 50 and 800) with training set compounds
cover numerous classes, including alkanes, alkenes, alky-
nes, halogens, amines, alcohols, N-containing compounds,
ketones, aldehydes, and organic acids (52). ADMET solubility
model can be described by six levels ranging from 0 to 5
described as following: (0) indicated compounds may have
extremely low solubility, (1) indicates insoluble compounds
or compounds with very low but possible solubility, (2) indi-
cates compound will have low solubility, (3) indicates com-
pound may have good solubility, (4) indicates that this
compound may have optimal solubility, and (5) indicates
Acknowledgments
The authors are grateful to Dr. Pieter Leyssen at Labora-
tory for Virology and Experimental Chemotherapy, Rega
Institute for Medical Research (KU Leuven), Leuven,
Belgium for performing CHIKV virus-cell-based assay.
References
1. Powers A.M. (2008) Togaviruses: Alphaviruses. In: Edi-
tors-in-Chief Mahy B.W.J., van Regenmortel M.H.V.,
editors. Encyclopedia of Virology, 3rd edn. Oxford:
Academic Press; p. 96–100.
2. Ross R.W. (1956) The Newala epidemic. III. The virus: iso-
lation, pathogenic properties. J Hyg (Lond);54:177–191.
ꢁ
ꢂ
€
3. Pialoux G., Gauzere B.-A., Jaureguiberry S., Strobel M.
(2007) Chikungunya, an epidemic arbovirosis. Lancet
Infect Dis;7:319–327.
4. Parola P., De Lamballerie X., Jourdan J., Rovery C.,
Vaillant V., Minodier P., Brouqui P., Flahault A., Raoult
D., Charrel R. (2006) Novel chikungunya virus variant in
travelers returning from Indian Ocean islands. Emerg
Infect Dis;12:1493–1499.
Chem Biol Drug Des 2015; 86: 1518–1527
1525

El-labbad et al.
5. Rezza G., Nicoletti L., Angelini R., Romi R., Finarelli
A.C., Panning M., Cordioli P. et al. (2007) Infection with
chikungunya virus in Italy: an outbreak in a temperate
region. Lancet;370:1840–1846.
6. Chopra A., Anuradha V., Lagoo-Joshi V., Kunjir V.,
Salvi S., Saluja M. (2008) Chikungunya virus aches and
20. Rathore A.P.S., Haystead T., Das P.K., Merits A., Ng
M.-L., Vasudevan S.G. (2014) Chikungunya virus nsP3
& nsP4 interacts with HSP-90 to promote virus replica-
tion: HSP-90 inhibitors reduce CHIKV infection and
inflammation in vivo. Antiviral Res;103:7–16.
21. Strauss J.H., Strauss E.G. (1994) The alphaviruses:
gene expression, replication, and evolution. Microbiol
Rev;58:491–562.
pains:
an
emerging
challenge.
Arthritis
Rheum;58:2921–2922.
7. Waymouth H.E., Zoutman D.E., Towheed T.E. (2013)
Chikungunya-related arthritis: case report and review of
the literature. Semin Arthritis Rheum;43:273–278.
22. Lulla A., Lulla V., Tints K., Ahola T., Merits A. (2006)
Molecular determinants of substrate specificity for
Semliki Forest virus nonstructural protease.
J
8. Leyssen P., Litaudon M., Guillemot J.-C., Rasoanaivo P.,
Smadja J., Gurib-Fakim A., Canard B., Gueritte F. (2011)
PHYTOCHIK: biodiversity as a source of selective inhibi-
tors of CHIKV replication. Antiviral Res;90:A36–A37.
9. Bourjot M., Leyssen P., Eydoux C., Guillemot J.-C.,
Virol;80:5413–5422.
23. Vasiljeva L., Merits A., Golubtsov A., Sizemskaja V.,
€€ €
Kaariainen L., Ahola T. (2003) Regulation of the
sequential processing of Semliki Forest virus replicase
polyprotein. J Biol Chem;278:41636–41645.
ꢂ
€€ €
Canard B., Rasoanaivo P., Gueritte F., Litaudon M.
24. Vasiljeva L., Valmu L., Kaariainen L., Merits A. (2001)
(2012) Chemical constituents of Anacolosa pervilleana
and their antiviral activities. Fitoterapia;83:1076–1080.
Site-specific protease activity of the carboxyl-terminal
domain of Semliki Forest virus replicase protein nsP2.
J Biol Chem;276:30786–30793.
ꢂ
10. Bourjot M., Delang L., Nguyen V.H., Neyts J., Gueritte
F., Leyssen P., Litaudon M. (2012) Prostratin and 12-
O-tetradecanoylphorbol 13-acetate are potent and
selective inhibitors of Chikungunya virus replication. J
Nat Prod;75:2183–2187.
25. Pastorino B.A.M., Peyrefitte C.N., Almeras L., Gran-
dadam M., Rolland D., Tolou H.J., Bessaud M. (2008)
Expression and biochemical characterization of nsP2
cysteine protease of Chikungunya virus. Virus
Res;131:293–298.
11. Bourjot M., Leyssen P., Eydoux C., Guillemot J.-C.,
ꢂ
Canard B., Rasoanaivo P., Gueritte F., Litaudon M.
26. Santos M.M.M., Moreira R. (2007) Michael acceptors
as cysteine protease inhibitors. Mini Rev Med
Chem;7:1040–1050.
27. Faver J., Merz K.M. Jr (2010) Utility of the hard/soft
acid–base principle via the Fukui function in biological
systems. J Chem Theory Comput;6:548–559.
(2012) Flacourtosides A-F, phenolic glycosides isolated
from flacourtia ramontchi. J Nat Prod;75:752–758.
12. Masrinoul P., Puiprom O., Tanaka A., Kuwahara M.,
Chaichana P., Ikuta K., Ramasoota P., Okabayashi T.
(2014) Monoclonal antibody targeting chikungunya
virus envelope 1 protein inhibits virus release. Virol-
ogy;464–465:111–117.
13. Khan A.H., Morita K., del Carmen Parquet M., Hasebe
F., Mathenge E.G.M., Igarashi A. (2002) Complete
nucleotide sequence of chikungunya virus and evi-
dence for an internal polyadenylation site. J Gen
Virol;83:3075–3084.
28. Shah F., Mukherjee P., Gut J., Legac J., Rosenthal
P.J., Tekwani B.L., Avery M.A. (2011) Identification of
novel malarial cysteine protease inhibitors using struc-
ture-based virtual screening of a focused cysteine pro-
tease inhibitor library. J Chem Inf Model;51:852–864.
29. Wodicka L.M., Ciceri P., Davis M.I., Hunt J.P., Floyd M.,
Salerno S., Hua X.H., Ford J.M., Armstrong R.C., Zar-
rinkar P.P., Treiber D.K. (2010) Activation state-dependent
binding of small molecule kinase inhibitors: structural
insights from biochemistry. Chem Biol;17:1241–1249.
30. Madhavi Sastry G., Adzhigirey M., Day T., Annabhimoju
R., Sherman W. (2013) Protein and ligand preparation:
parameters, protocols, and influence on virtual screen-
ing enrichments. J Comput Aided Mol Des;27:221–234.
31. Ahmed M., Sadek M.M., Abouzid K.A., Wang F. (2013)
In silico design: extended molecular dynamic simula-
tions of a new series of dually acting inhibitors against
EGFR and HER2. J Mol Graph Model;44:220–231.
32. Ahmed M., Sadek M.M., Serrya R.A., Kafafy A.H.N.,
Abouzid K.A., Wang F. (2013) Assessment of new anti-
HER2 ligands using combined docking, QM/MM scor-
ing and MD simulation. J Mol Graph Model;40:91–98.
33. Rocha G.B., Freire R.O., Simas A.M., Stewart J.J.P.
(2006) RM1: a reparameterization of AM1 for H, C, N, O,
P, S, F, Cl, Br, and I. J Comput Chem;27:1101–1111.
34. Bochevarov A.D., Harder E., Hughes T.F., Greenwood
J.R., Braden D.A., Philipp D.M., Rinaldo D., Halls
€€ €
14. Ahola T., Kaariainen L. (1995) Reaction in alphavirus
mRNA capping: formation of a covalent complex of
nonstructural protein nsP1 with 7-methyl-GMP. Proc
Natl Acad Sci;92:507–511.
15. Wengler G., Wengler G. (1993) The NS 3 nonstructural
protein of flaviviruses contains an RNA triphosphatase
activity. Virology;197:265–273.
€
€€ €
16. Rikkonen M., Peranen J., Kaariainen L. (1994) ATPase
and GTPase activities associated with Semliki Forest
virus nonstructural protein nsP2. J Virol;68:5804–5810.
17. Strauss E.G., De Groot R.J., Levinson R., Strauss J.H.
(1992) Identification of the active site residues in the
nsP2 proteinase of sindbis virus. Virology;191:932–940.
18. Wang Y.F., Sawicki S.G., Sawicki D.L. (1994) Alphavirus
nsP3 functions to form replication complexes transcrib-
ing negative-strand RNA. J Virol;68:6466–6475.
19. Rubach J.K., Wasik B.R., Rupp J.C., Kuhn R.J., Hardy
R.W., Smith J.L. (2009) Characterization of purified
Sindbis virus nsP4 RNA-dependent RNA polymerase
activity in vitro. Virology;384:201–208.
1526
Chem Biol Drug Des 2015; 86: 1518–1527

Irreversible CHIKV NsP2 Protease Inhibitors
M.D., Zhang J., Friesner R.A. (2013) Jaguar: a high-
performance quantum chemistry software program
with strengths in life and materials sciences. Int J
Quantum Chem;113:2110–2142.
49. Wang S.-Y., Ji S.-J., Loh T.-P. (2003) The Michael
addition of indole to a, b-unsaturated ketones cat-
alyzed by iodine at room temperature. Syn-
lett;15:2377–2379.
35. Keledjian K., Garrison J.B., Kyprianou N. (2005) Doxa-
zosin inhibits human vascular endothelial cell adhesion,
migration, and invasion. J Cell Biochem;94:374–388.
36. Friesner R.A., Murphy R.B., Repasky M.P., Frye L.L.,
Greenwood J.R., Halgren T.A., Sanschagrin P.C.,
Mainz D.T. (2006) Extra precision glide: docking and
scoring incorporating a model of hydrophobic enclo-
50. Lipinski C.A. (2004) Lead- and drug-like compounds:
the rule-of-five revolution. Drug Discov Today Tech-
nol;1:337–341.
51. Egan W.J., Merz K.M., Baldwin J.J. (2000) Prediction
of drug absorption using multivariate statistics. J Med
Chem;43:3867–3877.
52. Cheng A., Merz K.M. (2003) Prediction of aqueous sol-
ubility of a diverse set of compounds using quantitative
sure
for
protein–ligand
complexes.
J
Med
Chem;49:6177–6196.
structure–property
relationships.
J
Med
37. Bassetto M., De Burghgraeve T., Delang L., Massarotti
A., Coluccia A., Zonta N., Gatti V., Colombano G.,
Sorba G., Silvestri R., Tron G.C., Neyts J., Leyssen P.,
Brancale A. (2013) Computer-aided identification,
design and synthesis of a novel series of compounds
with selective antiviral activity against chikungunya
virus. Antiviral Res;98:12–18.
Chem;46:3572–3580.
53. Nalam M.N.L., Ali A., Altman M.D., Reddy G.S.K.K.,
Chellappan S., Kairys V., Ozen A., Gilson M.K., Tridor
B., Rana T.M., Schiffer C.A. (2010) Evaluating the sub-
strate-envelope hypothesis: structural analysis of novel
HIV-1 protease inhibitors designed to be robust
against drug resistance. J Virol;84:5368–5378.
38. Russo A.T., Malmstrom R.D., White M.A., Watowich S.J.
(2010) Structural basis for substrate specificity of alpha-
virus nsP2 proteases. J Mol Graph Model;29:46–53.
39. Ouyang X., Zhou S., Su C.T.T., Ge Z., Li R., Kwoh
C.K. (2013) CovalentDock: automated covalent dock-
ing with parameterized covalent linkage energy estima-
tion and molecular geometry constraints. J Comput
Chem;34:326–336.
54. Shen Y., Altman M.D., Ali A., Nalam M.N.L., Cao H.,
Rana T.M., Schiffer C.A., Tidor B. (2013) Testing the
substrate-envelope hypothesis with designed pairs of
compounds. ACS Chem Biol;8:2433–2441.
Supporting Information
Additional Supporting Information may be found in the
online version of this article:
40. Liu S.B. (2009) Conceptual density functional theory
and some recent developments. Acta Phys Chim
Sin;25:590–600.
Appendix S1. Materials and methods.
41. Roos G., Geerlings P., Messens J. (2009) Enzymatic
catalysis: the emerging role of conceptual density func-
tional theory. J Phys Chem B;113:13465–13475.
42. Li Y., Evans J.N.S. (1995) The Fukui function: a key
concept linking frontier molecular orbital theory and the
Figure S1. Design strategy of new CHIKV nsP2 protease
peptidomimetic covalent inhibitors.
Figure S2. Binding site of nsP2 CHIKV protease, (A) the
tunnel shaped binding site of nsP2 CHIKV protease, (B)
binding site residues in ball and stick representation. The
b-hairpin structure designated is thought to restrict the
access of substrates.
hard–soft-acid–base
Soc;117:7756–7759.
43. Fukui K. (1982) Role of Frontier orbitals in chemical
reactions. Science;218:747–754.
principle.
J
Am
Chem
44. Bultinck P., Fias S., Van Alsenoy C., Ayers P.W.,
ꢂ
Carbo-Dorca R. (2007) Critical thoughts on computing
Figure S3. The proposed mechanism of action of the
designed inhibitor as illustrated by the Michael addition
reaction between cys1013 and compound 3a.
atom condensed Fukui functions.
Phys;127:034102–034111.
J
Chem
45. Zaheer S.H., Kacker I.K., Rao N.S. (1956) The con-
densation of levulinic acid with aromatic aldehydes.
Chem Ber;89:351–354.
46. Abouzid K., Frohberg P., Lehmann J., Decker M.
(2007) 6-Aryl-4-oxohexanoic acids: synthesis, effects
on eicosanoid biosynthesis, and anti-inflammatory
in vivo-activities. Med Chem;3:433–438.
47. Ismail M., Lehmann J., Abou El Ella D., Albohy A.,
Abouzid K. (2009) Lonazolac analogues: molecular
modeling, synthesis, and in vivo anti-inflammatory
activity. Med Chem Res;18:725–744.
Figure S4. Covalent docking of the three most active inhi-
bitors defined in vitro (3a, 4b and 5d) with the nsP2
enzyme.
Figure S5. The orbital distribution of the frontier molecular
orbitals involved in the Michael addition reaction, that is
HOMO of Cys1013 and the LUMO orbitals of 3a, 4b and
5d. Orbital are calculated at the B3LYP/6-31 + G*.
Table S1. Principal parameters calculated by ADMET module.
48. Silverstein R.M., Webster F.X., Kiemie D. (2005) Spec-
trometric Identification of Organic Compounds, 7th edn.
Hoboken: John Wiley & Sons, Inc.
Chem Biol Drug Des 2015; 86: 1518–1527
1527