Structure-based drug design, synthesis and biological assays of P. falciparum Atg3–Atg8 protein–protein interaction inhibitors

Article abstract of DOI:10.1007/s10822-018-0102-5

The proteins involved in the autophagy (Atg) pathway have recently been considered promising targets for the development of new antimalarial drugs. In particular, inhibitors of the protein–protein interaction (PPI) between Atg3 and Atg8 of Plasmodium falc

Full text of DOI:10.1007/s10822-018-0102-5

Journal of Computer-Aided Molecular Design
https://doi.org/10.1007/s10822-018-0102-5
Structure-based drug design, synthesis and biological assays of P.
falciparum Atg3–Atg8 protein–protein interaction inhibitors
Stefania Villa1 · Laura Legnani2 · Diego Colombo · Arianna Gelain · Carmen Lammi
3
1
1
·
Daniele Bongiorno1 · Denise P. Ilboudo
4,5
· Kellen E. McGee · Jürgen Bosch
6
6,7
· Giovanni Grazioso
1
Received: 27 July 2017 / Accepted: 24 January 2018
©
Springer International Publishing AG, part of Springer Nature 2018
Abstract
The proteins involved in the autophagy (Atg) pathway have recently been considered promising targets for the develop-
ment of new antimalarial drugs. In particular, inhibitors of the protein–protein interaction (PPI) between Atg3 and Atg8
of Plasmodium falciparum retarded the blood- and liver-stages of parasite growth. In this paper, we used computational
techniques to design a new class of peptidomimetics mimicking the Atg3 interaction motif, which were then synthesized
by click-chemistry. Surface plasmon resonance has been employed to measure the ability of these compounds to inhibit the
Atg3–Atg8 reciprocal protein–protein interaction. Moreover, P. falciparum growth inhibition in red blood cell cultures was
evaluated as well as the cyto-toxicity of the compounds.
Keywords Malaria · Autophagy · Atg8 inhibitors · Docking · PPI inhibitors · Peptidomimetics · 1,2,3-Triazole
Abbreviations
MD Molecular dynamics
PE
Phosphatidylethanolamine
SPR Surface plasmon resonance
PPI Protein–protein interactions
Introduction
Electronic supplementary material The online version of this
article (https://doi.org/10.1007/s10822-018-0102-5) contains
supplementary material, which is available to authorized users.
Malaria is one of the deadliest parasitic diseases with an esti-
mated 212 million clinical cases in 2015, causing 429,000
deaths [1]. More than two-thirds (70%) of these deaths are
children under 5 years old, who are particularly susceptible
to infection and illness. Malaria is thus a major killer of
children in this age group, taking the life of a child every
*
Giovanni Grazioso
giovanni.grazioso@unimi.it
1
2
3
Dipartimento di Scienze Farmaceutiche, Università degli
Studi di Milano, Via L. Mangiagalli 25, 20133 Milan, Italy
2
min [1].
Malaria burden is caused by the Plasmodium parasites,
Dipartimento di Chimica, Università degli Studi di Pavia, Via
Taramelli 12, 27100 Pavia, Italy
transmitted through the bite of the infected female anophe-
line mosquitoes. The species able to infect humans are Plas-
modium falciparum, P. vivax, P. malariae, and P. ovale, the
first being the most lethal one. Together with P. vivax, P.
falciparum accounts for >95% of malaria cases in the world
[2]. Moreover, in South East Asia, monkey-to-human trans-
mission of P. knowlesi infections have also been reported [3].
Several treatments of malaria are available, but it has
become evident that all classes of drugs provoke parasite
resistance, such as that recently reported for artemisinins
Dipartimento di Biotecnologie Mediche e Medicina
Traslazionale, Università degli Studi di Milano, Via Saldini
5
0, 20133 Milano, Italy
4
5
6
7
Dipartimento di Scienze Farmacologiche e Biomolecolari,
Università degli Studi di Milano, Via Balzaretti 9,
2
0133 Milano, Italy
Present Address: Centre Universitaire Polytechnique,
Université de Ouagagadougou, Fada N’Gourma,
Burkina Faso
Pediatric Pulmonology Division, Department of Pediatrics,
Case Western Reserve University School of Medicine,
Cleveland, OH, USA
[
4, 5], limiting the efficacy of the applied therapy. After
intense research efforts by big pharmaceutical compa-
nies and nonprofit organizations, the new malaria vaccine
InterRayBio, LLC, Baltimore, MD, USA
Vol.:(0
1
123456789)
3

Journal of Computer-Aided Molecular Design
RTS,S/AS01, developed by GlaxoSmithKline Biologicals
and the “PATH Malaria Vaccine Initiative”, will be rolled
out in pilot projects in three countries in sub-Saharan
Africa [6]. However, current available data from these
studies suggest limited protection through the vaccine
of about 30–40% [7]. It follows that new targets for the
design and development of antimalarial drugs are urgently
needed [8].
Atg8–PE then prompts the assembly of source membranes
into the membranes of autophagosomes [17].
In parasites such as P. falciparum [13] and T. gondii [15]
autophagy is carried out by a simplified cellular processes,
yet it is known that Atg8 was found to be essential to the
parasite survival and it is expressed during all stages of their
growth [18, 19]. Consequently, the Atg8–Atg3 protein–pro-
tein interactions (PPIs) have been deemed a promising target
for the development of new drugs active in blood and liver
stages of the parasite infection [20, 21].
A promising strategy to develop new drugs is to target the
proteases of malaria parasites, which are involved in the pro-
cesses of host erythrocyte rupture, erythrocyte invasion and
hemoglobin degradation. In particular, attention has been
focused on falcipain-2, for which some potent inhibitors
have been reported [9, 10]. More recently, new experimental
evidence suggested the importance of autophagy in differ-
entiation, development and survival of pathogenic parasites
The X-ray crystal structure of the P. falciparum (Pf)
Atg3–Atg8 complex (PDB ID: 4EOY [13]) paved the way to
the computer-aided design of new potential drug candidates,
revealing that PfAtg8 provides multiple sites (W and L) that
anchor PfAtg3 (Fig. 1). An additional region of the pro-
tein, the A-loop, has high sequence diversity among differ-
ent parasites and humans. This has been the object of Hain
et al. [20] studies leading to the identification of ALC25
(Fig. 1), which significantly inhibits the PfAtg3–PfAtg8 PPI,
in contrast to the 4-pyridin-2-yl-1,3-thiazol-2-amine (PTA)
class of compounds [21], which binds to the W- and L-sites.
In this paper, taking into account the structural insights
from the PfAtg3–PfAtg8 X-ray complex, we report the
computational design and the synthesis of new potential
PfAtg3–PfAtg8 PPI inhibitors, and their subsequent test-
ing by surface plasmon resonance (SPR) and P. falciparum
growth inhibition assays.
[
[
11, 12] such as P. falciparum [13] or Toxoplasma gondii
14, 15]. Autophagy is an intracellular degradation process
in which the cellular contents are enveloped by phospholipid
bilayers (autophagosome) and then fused with lysosomes
or vacuoles. In mammals, over thirty different subtypes of
autophagy-related proteins (Atg) are involved in this process.
Among them, Atg8 seems to be essential for the formation of
autophagosomal membranes. The association of Atg8 to the
membrane results from its coupling with phosphatidyletha-
nolamine (PE). This event, triggered by a post-translational
modification process known as “lipidation,” is accomplished
by the Atg8 conjugation system. This process involves Atg
proteins such as Atg4 (member of the caspase family), Atg7,
Atg3 and the Atg5–Atg12 complex. The first of these (Atg4),
undergoes cleavage of the C-terminal residues, which acti-
vates Atg8 and promotes the binding of a cysteine residue
of Atg7 to the Atg8–G116. This forms a thioester bond in an
ATP-dependent manner. Then, the activated form of Atg8 is
available for association with Atg3 by a thioester bond. Ulti-
mately, the Atg5–Atg12 complex assists [16] the Atg8–Atg3
dissociation and the coupling of the Atg8′ N-term to a PE
molecule composing the membrane via an amide bond.
Results and discussion
Design of PPI inhibitors
PPI inhibitors are compounds capable of impairing the
molecular interactions between protein partners [22]. Pep-
tidomimetics belong to a class of PPI inhibitors that retain
the molecular features of the peptide to be mimicked, in
which peptide bonds are replaced by more enzymatically
Fig. 1 X-ray structure of PfAtg8
(
gray surface) in complex with
PfAtg3 (green stick model,
PDB ID: 4EOY). Chemical
structures of ALC25 and of a
representative PTA containing
derivative, known as Atg8–Atg3
PPI inhibitors. PTA moiety is
colored in blue
1
3

Journal of Computer-Aided Molecular Design
stable bonds. Therefore, the ideal PfAtg3′ peptidomimetic
is a compound able to reproduce the contacts formed by the
template structure (sequence NDWLLPSY), while one or
more peptide bonds are replaced by new moieties. Due to the
variety of interactions shaped by the amino acids of PfAtg3
in the X-ray complex, the hot spots responsible for the PPI
should be identified in order to design ligands that mimic
the essential contacts only. To this aim, we identified the
minimal active sequence of PfAtg3 by performing molecular
dynamics (MD) simulations on the solvated PfAtg3/PfAtg8
X-ray complex. This analysis revealed that PfAtg3 interacts
with PfAtg8 by: (i) the H-bond between the side chains of
PfAtg3–N103 and PfAtg8–E17, (ii) the projection of the
PfAtg3–W105 indole ring into the W-site of PfAtg8 shaped
by residues E17, T18, I21, P30, K58 and Y113, (iii) the
H-bonds of PfAtg3–L106 with the backbone atoms of
PfAtg8–L50 and PfAtg8–K58, (iv) the H-bond between
PfAtg3–Y110 and PfAtg8–Met55, (v) the van der Waals
(vdW) contacts between the side chain of PfAtg3–L106
and those of PfAtg8–F49 and PfAtg8–H67, and finally (vi)
the vdW contacts of the PfAtg3–P108′ pyrrolidine ring
with the PfAtg8′ pocket sized by residues F49, V51, F60
and M55 (L-site). MD simulations on this complex also
showed that the main contributors to the PfAtg8–PfAtg3
PPI were located in the middle of the PfAtg3 sequence
(core region, Fig. 2a), i.e., the amino acids WLLP. In fact,
after over 100 ns of MD simulations, the root mean square
fluctuation (RMSF) of the Cα atoms belonging to PfAtg3′
core region was lower than 1.1 Å, at variance with the resi-
dues at the PfAtg3′ C- and N-terms showing higher values
(Fig. 2b). This was the consequence of the stability showed
by the interactions (H-bonds and vdW contacts) engaged
Fig. 2 a WLLP segment of
PfAtg3 (yellow stick model) in
complex with PfAtg8 (colored
as green stick model). Hydrogen
bonds are represented as yellow
dashed lines. b Cα-RMSF of
PfAtg3 within the PfAtg8 bind-
ing site over MD simulations
1
3

Journal of Computer-Aided Molecular Design
by PfAtg3–WLLP, further reinforced by the creation of
two additional H-bonds between the backbone atoms of
PfAtg3–D104 and those of PfAtg8–K47 and PfAtg8–K48.
Conversely, the high conformational mobil-
ity showed by residues belonging to the PfAtg3′ ter-
minal residues was caused by the loss of the H-bond
between the side chains of PfAtg3–N103 and
PfAtg8–E17. Moreover, an H-bond between PfAtg3–Y110
and PfAtg8–E59 was only occasionally shaped, at the
expense of the PfAtg3–Y110/PfAtg8–Met55 contact found
in the crystallized complex. All distance fluctuations over
MD simulations can be found in Fig. S1, Electronic sup-
plementary material 1.
suitable H-bonds with the guanidine group of PfAtg8–R28;
the optimal length of the spacer linking these heterocycles
and the 1,2,3-triazole ring remained to be established. We
therefore undertook the prediction of the size of the spacer
by performing docking calculations on a small series of suit-
ably designed peptidomimetics (Fig. 3).
The results (Fig. S2, Electronic supplementary mate-
rial 1) showed that the pyridine and pyridazine derivatives,
linked by an ethylene spacer to the 1,2,3-triazole ring, dis-
played the highest theoretical score in binding the target
protein. Noteworthy, they displayed docking score identi-
cal to the one acquired by WLLP, showing experimental
K value of 326 nM [13]. We supposed that: (i) the pyri-
D
In concordance with the calculated predictions, the
WLLP sequence of PfAtg3 was identified as the most suit-
able structure to be mimicked by new PfAtg3–PfAtg8 PPI
inhibitors. The chemistry of peptidomimetics provides sev-
eral molecular fragments capable of reproducing the pep-
tide’s primary and secondary structures [23, 24]. Among
them, 1,2,3-triazoles are heterocycles with interesting
properties for application in peptide sciences since they are
amide bioisosteres and are resistant to enzymatic degrada-
tion. These properties make 1,2,3-triazoles promising candi-
dates for the development of novel PfAtg3 peptidomimetics
with potentially improved biological profiles [23]. Further-
more, 1,4-disubstituted 1,2,3-triazoles have gained relevance
in medicinal chemistry due to their synthetic accessibility
by click 1,3-dipolar cycloaddition reaction between azides
and alkynes under copper salts catalysis. Consequently, we
designed new PfAtg3 analogues in which: (i) a 1,4-disubsti-
tuted 1,2,3-triazole moiety replaced the PfAtg3′ L106–P107
amide bond, (ii) a phenyl ring replaced the isopropyl chain
of PfAtg3–L106, aiming to engender τ or π–π interactions
with the PfAtg8–F49 phenyl ring (Fig. 2a), and (iii) a tryp-
tophan was preserved in the N-term (Fig. 3), to optimally
fill the W-site.
dazine/pyridine ring could interact with the PfAtg8–R28
side chain by hydrogen bonds, (ii) the triazole ring could
create a hydrogen bond with the NH of PfAtg8–L50, (iii) the
N-term group (protonated) of the ligands interacted with the
carbonyl group of PfAtg8–K48 by a hydrogen bond, and (iv)
the tryptophan moiety occupied the PfAtg8′ W-site (Fig. 4).
The superimposition of the new peptidomimetics and the
PfAtg3 X-ray pose showed a good overlapping between the
template structure PfAtg3 and the newly designed ligand
structures (Fig. 4a).
MD simulations were also performed on the (S,S)-
pyridazine analogue within PfAtg8 and, at the end of the
simulations (100 ns), the obtained trajectories showed that
the ligand was almost stable in the binding site. In fact, the
phenyl ring of the ligand created vdW contacts with the side
chain of the PfAtg8–F49 (Fig. S3A, Electronic supplemen-
tary material 1). The triazole ring retained the H-bond with
the NH group of PfAtg8–L50 (Fig. S3B, Electronic supple-
mentary material 1), and the indole ring was firmly inserted
in the W-site (Fig. S3C, Electronic supplementary material
1). Similarly, the H-bond between the N-term of the ligand
and the backbone of PfAtg8–K48 was maintained (Fig. S3D,
Electronic supplementary material 1). On the other hand, the
pyridazine ring of the ligand was not firmly inserted in the
L-site, in fact the predicted H-bond of the pyridazine ring
with PfAtg8–R28 during docking calculations was lost dur-
ing the early stages of MD simulations (Fig. S3E, Electronic
supplementary material 1).
The substituent to be introduced in position 1 of the
1
,2,3-triazole ring should occupy the PfAtg8′ L-site, and
is expected to promote additional interactions with the
PfAtg8–R28 side chain surrounding that site (see Fig. 2a). In
our hypothesis, a pyridine or a pyridazine ring should create
Fig. 3 General structure of the small set of designed compounds, in comparison with the template structure (PfAtg3–WLLP)
1
3

Journal of Computer-Aided Molecular Design
Fig. 4 a Superimposition of
pyridazine derivative (orange
stick models) and PfAtg3 (yel-
low stick models) within the
binding site of PfAtg8 (green
stick models). Hydrogen bonds
are represented as yellow
dashed lines. b Hypothetical
binding mode of the pyridazine
analogue in the PfAtg8 binding
site. The protein surface is
colored accordingly with the
partial charge of the residues,
red and blue for negative and
positive charges, respectively.
Surface and atom partial
charges were acquired by Pymol
software
Taking these theoretical outcomes into considera-
tion, we synthesized the designed peptidomimetics. The
planned synthetic strategy led to two pairs of diastere-
oisomers (Fig. 5, compounds 1, 2 and 3, 4). Additional
molecular modeling studies, supported by high-field
NMR spectroscopy, were necessary to assign the abso-
lute configurations to the stereocenters, after which the
biological activity of the peptidomimetics was evaluated
by SPR and P. falciparum growth inhibition tests.
Fig. 5 Chemical structure of target compounds
1
3

Journal of Computer-Aided Molecular Design
Synthesis of compounds 1–4
Configuration assignment of diastereoisomers 1, 2
and 3, 4
Synthesis of compounds 1, 2 and 3, 4 (Fig. 5) were per-
formed by click reactions from alkyne 7 (Scheme 1) and
the azides 12 and 13 (Scheme 2) respectively. Alkyne 7
was obtained from 1-phenyl-2-propyn-1-ol (Scheme 1)
that, by treatment with acetonitrile, in presence of sul-
furic acid and Na SO , was converted into the amide
To assign the configuration of diastereoisomers 1,2 and
3,4 we decided to combine NMR spectroscopy and com-
putational methods, as described in previous publications
1
[27–29]. Here, the H NMR spectra of 1 versus 2 and 3
versus 4 showed significantly different resonances of the
triazole and the ortho phenyl protons (see Table 1), sug-
gesting that the chemical environments of these protons is
sufficiently different for the two diastereoisomers to deeply
influence their chemical shifts. For this reason we decided
to accomplish a complete assignment for compounds 1–4
2
4
5
[25]. The latter was hydrolyzed in acidic conditions
and the resulting amine 6 was coupled with N-α-Boc-l-
tryptophan, in the presence of EDAC and HOBt, to give
the key intermediate 7.
Azide 12 was synthesized from the commercially avail-
able 2-(pyridin-2-yl)ethanol 8, meanwhile azide 13 was
obtained from 3-methyl pyridazine that was converted in
the corresponding 2-(pyridazin-3-yl)ethanol 9 in the pres-
ence of formaldehyde (Scheme 2). The alcohols 8 and 9
were treated with methanesulfonyl chloride to afford the
intermediates 10 and 11, which were then converted into
the corresponding azides 12 and 13 by treatment with
1
H resonances (Electronic Supplementary material 2) and
compare the obtained data with the corresponding calculated
chemical shifts based on previously optimized geometries.
1
The H NMR chemical shifts of compounds 1–4 are
shown in Table S1. Some NOE correlations (NOE1-4, Elec-
tronic supplementary material 2) were also related to the
conformations of 1–4 and were used to drive the following
modeling experiments. The conformational study was per-
formed on the simplified structures (R,S)-16 and (S,S)-17
(Fig. 6) that correspond to the two epimers for compounds
1–4. All the degrees of conformational freedom were con-
sidered; in particular, the orientation around the single
bonds exemplified by the curved arrows in Fig. 6. The vari-
ous geometries were all optimized at the B3LYP/6-31G(d)
level [30, 31].
NaN [26].
3
Finally, the cycloaddition of alkyne 7 to azides 12 and
1
3, in the presence of CuSO and sodium l-ascorbate,
4
provided the intermediates 14 and 15; removal of the tert-
butoxycarbonyl group by treatment with trifluoroacetic
acid, afforded the pairs of products 1, 2 and 3, 4 respec-
tively. The diasteroisomers were quantitatively separated
by flash chromatography (Scheme 3).
For construction of the starting geometries, atten-
tion was initially focused on conformations showing the
Scheme 1 Reagents and conditions: (a) H SO , Na SO , CH CN, N , r.t., 48 h; (b) 3.5 M HCl, 90 °C, 4 h; (c) N-α-Boc-l-tryptophan,
2
4
2
4
3
2
HOBt·H O, EDAC, CH Cl , N , r.t., 16 h
2
2
2
2
Scheme 2 Reagents and condi-
tions: (a) 37% CH O, piperi-
2
dine, H O, N , MW, 165 °C,
2
2
3
0 min; (b) MsCl, DIPEA,
CH Cl , r.t., 3 h; (c) NaN ,
2
2
3
DMF, 70 °C, 3 h
1
3

Journal of Computer-Aided Molecular Design
Scheme 3 Reagents and conditions: (a) CuSO , sodium l-ascorbate, tert-butanol, H O, rt, 18 h; (b) TFA, CH Cl , r.t., 3 h
4
2
2
2
1
Table 1 B3LYP/6-31G(d) GIAO calculated H NMR chemical shift
(
δ, in ppm relative to TMS) of some key protons of compounds 16
and 17 based on the geometries optimized at the same level in com-
parison with the experimental values from the recorded spectra in
compounds 1–4
Compound
δ
δ
δ
HαTRZ
Ha,e-Bn
HaTHR
1
1
1
2
3
4
6 (R,S)
6.56
7.65
6.85
7.35
6.97
7.52
7.13
6.92
7.09
6.83
7.13
6.85
6.96
6.86
7.07
6.97
7.09
7.00
7 (S,S)
Fig. 6 Simplified structures corresponding to the epimers of com-
pounds 1–4
experimentally determined NOESY correlations (see
above). Epimer 16 had a greater distribution of the popu-
lation between the different conformations, while in the
case of 17 only three geometries were significantly popu-
lated (Electronic supplementary material 2, Fig. S1 and
assignment: 1 and 3 were the diastereoisomer (R,S) while
2 and 4 were the diastereoisomer S,S.
Biophysical and biological investigations
on compounds 1–4
1
Table S2). We finally computed the H NMR chemical
shifts for each populated conformation of compounds 16
and 17 (Table 1) using GIAO NMR calculations [32, 33]
at the same DFT B3LYP/6-31G(d) [30, 31] level already
used for the optimizations. The close agreement between
experimental and theoretical data of the triazole and the
ortho phenyl protons (Table 1) allowed the configurational
We tested compounds 1–4 for in vitro inhibition of the
plasmodial Atg8–Atg3 PPI using our established SPR com-
petition assay [13, 20, 21]. Recombinant proteins for the
SPR assay were purified as previously described [13]. Each
inhibitor was dissolved in the running buffer and tested in
triplicates in a 12-point, two-fold dilution series ranging
1
3

Journal of Computer-Aided Molecular Design
from 0.49 to 1000 µM (Fig. 7a). To assess the SPR chip’s
integrity over the course of the experiment, the uninhibited
in an in culture context, an approximately ten-fold increase
in potency was observed both times the (R,S) diastereoiso-
mer was used to prevent P. falciparum growth, implying the
in culture processes are more efficient at transporting the
(R,S) diastereoisomer into the cell and thereby providing
compound 1 with access to the PfAtg3 and PfAtg8 inter-
face. The counterintuitive result of 2 emphasizes the need
for the cross-validation of assays using multiple readout
methods, enabling the early differentiation of putative lead
compounds.
(
control) interaction of Atg8 with Atg3 in the presence of
DMSO was run at regular intervals between the inhibition
tests. The response of the control interaction was maintained
throughout the experiment. Therefore, the chip surface was
not compromised by any of the compounds, or the time it
took to complete the dose–response experiment. At high
concentrations (500–1000 µM), the compounds started pre-
cipitating in the SPR running buffer and those data were
thus omitted for the IC calculation of each compound. A
5
0
dose-dependent inhibition of the Atg8–Atg3 interaction was
observed in all four compounds with increasing amounts of
the inhibitory molecules, indicating that the described mol-
ecules are indeed able to specifically block the interaction of
PfAtg3 with PfAtg8. The IC calculated from the normal-
Compounds 1 and 2 do not show any cytotoxic
effect on HepG2 viability
MTT experiments were performed in order to exclude
the potential toxic effects of compounds 1 and 2 on the
HepG2 cell line (Fig. 8). Compounds 1 and 2 were tested at
5
0
ized SPR response were 1 20, 2 3.8, 3 12.4, 4 17.4 µM. No
stereoselectivity of the PPI inhibition was observed in the
SPR assay between the (R,S) and the (S,S) diastereoisomers,
however, we found stereoselectivity when we evaluated the
growth inhibition of P. falciparum 3D7-infected blood-stage
cultures using the SYBR green assay, which measures bind-
ing of a fluorescent dye to parasite-derived DNA (Fig. 7b).
That the growth inhibition assays demonstrate a clear prefer-
ence for the (R,S) diastereoisomer while the SPR assay does
not is consistent with a situation in which some up-stream
transporter or process, present only in in culture context, is
stereoselective even though the PPI inhibition itself is not.
This may be observed in the result that, as measured by
SPR, compound 2 had the most prominent inhibition with
an IC of 3.8 µM, yet was found to be essentially inactive
5
0
in the in culture SYBR green assay with an IC of about
5
0
Fig. 8 Bar graphs indicating the results of MTT cell viability assay of
HepG2 cells after compounds 1 and 2 treatment for 24 and 48 h. The
data points represent the averages± SD of three independent experi-
ments in triplicate
5
60 µM. In culture, compound 1 had the most prominent,
albeit still modest, inhibitory effect with an IC of about
5
0
4
0 µM. In support of stereoselectivity of the inhibitory effect
Fig. 7 a SPR PPI inhibition study. The normalized SPR response
for each compound is plotted versus the logarithmic inhibitory con-
centration in M. All four compounds were tested in the range of
0.4–250 µM. b SYBR green growth inhibition assay of P. falciparum
3D7 strain with calculated IC values. Compound 1 and 3 are the R,S
5
0
diastereoisomers while 2 and 4 are the S,S diastereoisomers
1
3

Journal of Computer-Aided Molecular Design
concentrations ranging from 1 to 100 µM and no significant
cell mortality was observed at all tested doses after 24 and
molecular fragment more stably bound to the PfAtg8′ L-site.
Computational and synthetic studies aiming to identify more
efficient PfAtg3–PfAtg8 PPI inhibitors are still ongoing.
4
8 h treatment versus vehicle (C, DMSO). These data sug-
gest that both compounds do not induce any cell mortality
in this dose range.
Experimental protocol
Conclusion
Molecular dynamics simulations
The crystal structure of the PfAtg3–PfAtg8 complex ush-
ered in the possibility of computer-aided drug design of
new PfAtg3′ peptidomimetics with potential antimalarial
activity. In this paper, the X-ray crystal structure of the
PfAtg3–PfAtg8 complex was geometrically optimized by
energy minimization and MD simulations. The analysis of
the resulting trajectories supported the evidence that the
WLLP sequence of PfAtg3 displayed the lowest conforma-
tional freedom in the PfAtg3–PfAtg8 complex; therefore
WLLP was chosen as template for the design of new pepti-
domimetics incorporating a substituted 1,2,3-triazole ring,
which is a known amide bioisostere. A small set of com-
pounds was designed and screened by docking calculations,
and two pairs of diastereoisomers (compounds 1–4) were
then selected and synthesized through a click-chemistry
approach. In each diastereomeric pair, NMR and conforma-
tional analysis studies allowed unequivocal assignment of
the absolute configuration to the stereocentres.
MD simulations of the PfAtg3–PfAtg8 and the PfAtg8/4
complexes were performed by assigning the force field
parameters of AMBER12 [34] package. Accordingly, the
proteins were treated by the ff12SB force fields, and GAFF
was employed for the treatment of ligand [35], and the
TIP3P model [36] was used to represent the solvent mol-
ecules. A cubic box of water, with a minimum distance of
20 Å from the protein surface, was built. Prior to starting
MD simulations, a minimization of the bulk water mol-
ecules was carried out, using a gradient criterion con-
−
1
−1
vergence of 0.2 kcal mol Å . Then, the whole system
was optimized by applying a convergence criterion of
−
1
−1
0.0001 kcal mol
Å
and next equilibrated gradually
increasing the temperature to 300 K, over 60 ps. Finally, a
production run of 100 ns was performed. van der Waals and
short-range electrostatic interactions were taken into account
within a 10 Å cutoff, while long-range electrostatic inter-
actions were assessed using particle mesh Ewald method
[37]. Berendsen’s coupling algorithms [38] were active to
maintain the temperature at 300 K and the pressure of the
system at 1 atm, by a coupling constant set to 1.5 ps. Pmemd.
cuda module of AMBER12 performed the production runs
of MD simulations.
Biological assays showed that the PfAtg3–PfAtg8 asso-
ciation was inhibited in the SPR assay by both diastereoiso-
mers. Noteworthy, the (R,S) diastereoisomers 1 and 3 were
about tenfold more potent in inhibiting parasite growth
than their corresponding (S,S) counterparts 2 and 4. This
suggests that inhibitors with the (R,S) configuration of the
peptidomimetic drug could be either more efficiently trans-
ported across the red blood cell membrane and/or the para-
site plasma membrane, or that inhibitors having the (S,S)
configuration are degraded faster in a cellular context by
peptidases.
Docking studies
Docking calculations of the (S,S) pyridazine analogue
(Fig. 5) were performed by GOLD5.2.2 algorithm [39] using
the crystal structure of PfAtg8 deposited in the Protein Data
Bank with the code 4EOY, chain B [13]. The binding site
was defined by selecting the residues included in a sphere
with a radius of 13.0 Å from the Cα atom of Phe49. Amino
and carboxyl groups were considered in the ionized form
for consistency with the expected protonation state at physi-
ological pH. ChemPLP scoring function [40] was employed
while the genetic algorithm parameters were kept at their
default values, as already done in previous studies [41]. At
the end of docking calculations, poses were clustered by
the complete-linkage method [39]. This is an agglomerative
hierarchical clustering algorithm in which, initially, each ele-
ment is located in a different singleton cluster, then clusters
are combined into larger clusters until all elements are in
the same cluster. At each step, the two clusters separated by
the shortest distance are combined [39]. Finally, the clusters
The most active isomers 1 and 3 showed a comparable
IC value of about 40 µM in a P. falciparum growth inhibi-
5
0
tion assay. MTT experiments on compounds 1 and 2 showed
that they do not induce any cell mortality. However, fur-
ther efforts are needed to design more active compounds,
capable to fully and stably occupy the L-site of the target
protein PfAtg8. In fact, as suggested by MD simulations,
the pyridazine moiety of the pyridazine derivatives (3–4)
could not be firmly bound in the L-site of PfAtg8, and this
could explain the low biological activity showed by these
compounds.
The results attained in this initial study point at the
research of peptidomimetics in this area to novel inhibitors
endowed with the (R,S) absolute configuration, with the
pyridazine moiety that should conceivably be replaced by a
1
3

Journal of Computer-Aided Molecular Design
1
were ranked depending on the obtained score. The solution
representative of each cluster was the one which obtained
the highest docking score.
Yield: quantitative; M.p. 88.8–89.8 °C; H NMR
(300 MHz, CDCl ) δ 1.62 (s, 1H), 2.02 (s, 3H), 2.49 (d,
3
J=3 Hz, 1H), 6.02 (d, J=3 Hz), 7.23–7.55 (m, 5H).
Synthesis of 1-phenylprop-2-yn-1-amine (6) [42]
Chemistry
N-(1-phenylprop-2-yn-1-yl)acetamide (5) (465 mg,
Reagents and solvents were purchased from Sigma-Aldrich
and used without further purification. Reactions were per-
formed through conventional conditions or by a microwave
synthesizer (Initiator 2.0 Biotage). The reactions involving
air-sensitive reagents were carried out under nitrogen atmos-
phere and anhydrous solvents were used when necessary.
Reactions were monitored by thin layer chromatography
analysis on aluminum-backed Silica Gel 60 plates (70–230
mesh, Merck), using an ultraviolet fluorescent lamp at 254
and 365 nm. Visualization was aided by opportune staining
reagents. Purification of intermediates and final compounds
was performed by flash chromatography using Geduran® Si
2
.684 mmol) was suspended in 3.5M HCl (15.5 mL) and
heated to 90 °C under stirring for 5 h. After cooling, the
solution was extracted with diethyl ether (1 × 5 mL) and
the aqueous layer was basified by adding solid NaHCO
3
until pH 8.5 and subsequently extracted with diethyl ether
(
3×10 mL). The organic layer was dried and concentrated
in vacuo. The crude was purified by flash chromatography
(
eluent mixture:diethyl ether/petroleum ether, 5:5) to obtain
amine 6 as purple oil (242 mg, 1.845 mmol).
1
Yield: 69%; H NMR (300 MHz, CDCl ) δ 2.057 (s, 2H),
3
2
.49 (d, J = 3 Hz, 1H), 4.79 (d, J = 3 Hz, 1H), 7.23–7.58
(
m, 5H).
6
0 (40–63 µm, Merck).
1
13
H and C NMR spectra were recorded in CDCl and
3
CD OD on a Variant 300 MHz Oxford equipped with a non-
Synthesis of tert-butyl
((2S)-3-(1H-indol-3-yl)-1-oxo-1-((1-phenylprop-2-yn-1-yl)
amino)propan-2-yl)carbamate (7)
3
reverse probe at 25 °C. Chemical shifts are expressed as δ
(
ppm) and are referenced to residual solvent proton/carbon
peak. Multiplicity is reported as s (singlet), br s (broad sin-
glet), d (double), t (triplet), q (quartet), m (multiplet), dd
A solution of N-α-Boc-l-tryptophan (235 mg, 0.772 mmol)
dissolved anhydrous dichloromethane (30 mL) under nitro-
gen atmosphere was cooled by an ice bath, and 1-phenyl-
prop-2-yn-1-amine (6) (160 mg, 1.220 mmol) was added.
When the mixture reached the room temperature, monohy-
drated HOBt (178 mg, 1.165 mmol) and EDAC (231 mg,
1.206 mmol) were added. The mixture was stirred at room
temperature for 16 h and then the solvent was evaporated
under reduced pressure. The crude was purified by flash
chromatography (eluent mixture:ethyl acetate/petroleum
ether 3:7) to afford a yellow oil 7 (305 mg, 0.731 mmol) as
(
double of doublets), dt (doublet of triplets). The coupling
constants (J-values) are given in Hertz (Hz). All spectro-
scopic data match the assigned structures. ESI-MS analyses
were performed by using a LQT-Orbitrap XL mass spec-
trometer (Thermo Fisher Scientific) with an electrospray
ionization source (Finningan IonMax). Melting points were
determined on a Buchi Melting Point B540 instrument. Opti-
cal rotations were determined by the polarimeter JASCO
P-1010, at 25 °C.
diastereoisomeric mixture.
1
Procedure for the synthesis of tert-butyl
Yield: 95%; H NMR (300 MHz, CDCl ) δ 1.41 (s,
3
((2S)-3-(1H-indol-3-yl)-1-oxo-1-((1-phenylprop-2-yn-1-yl)
9H), 1.72 (br s, 1H), 2.40 (s, 1H), 3.14–3.23 (m, 1H),
3.27–3.37 (m, 1H), 4.45 (br s, 1H), 6.29 (dd, J=9, 24 Hz,
1H), 7.11–7.37 (m, 10H), 7.67 (d, J = 9 Hz, 1H), 7.99 (d,
amino)propan-2-yl)carbamate (7). Synthesis
of N-(1-phenylprop-2-yn-1-yl)acetamide (5) [25]
1
3
J=24 Hz, 1H); C NMR (75 MHz, CDCl ) δ 28.01 (CH ,
3
2
To a suspension of 1-phenyl-2-propyn-1-ol (500 mg,
THR β), 28.20 (3CH ), 44.40 and 44.44 (diast. CH, Bn),
3
3
.783 mmol) and Na SO (537 mg, 3.783 mmol) in CH CN
54.22 (CH, THR α), 73.04 (≡CH), 81.26 (≡C– or C–O),
81.29 (C–O or ≡C–), 110.10 (C, THR h), 111.21 (CH, THR
b), 118.70 and 118.76 (diast. CH, THR e), 119.75 (CH, THR
d), 122.18 (CH, THR c); 123.34 and 123.41 (diast. CH,
THR a), 126.86 and 126.89 (diast. 2CH, Bn a, e), 127.29
(C, THR g), 128.08 (CH, Bn c), 128.56 and 128.59 (diast.
2CH, Bn b, d), 136.15 and 136.21 (diast. C, Bn f), 137.71
(C, THR f or C=O Boc), 137.79 (THR f or C=O Boc),
170.77 (C=O). Chemical formula: C H N O . Molar mass:
2
4
3
(
(
6 mL) at −25 °C under stirring, a solution of 97% H SO
2 4
1 mL) in CH CN (3.9 mL) was added dropwise. The mix-
3
ture was stirred at room temperature for 48 h. Afterwards
the solvent was concentrated and ice was added to the sus-
pension that was extracted by diethyl ether (3×5 mL) and
dichloromethane (3×5 mL). The organic layer was dried and
evaporated under vacuum. The crude was purified by flash
chromatography (eluent mixture:diethyl ether/petroleum
ether 5:5) to give the white solid 5 (655 mg, 3.782 mmol).
2
5
27
3
3
−
1
+
417.51 g mol ; MS (ESI) m/z 440.2 [M+Na] .
1
3

Journal of Computer-Aided Molecular Design
Procedure for the synthesis of 2-(pyridazin-3-yl) ethanol (9)
General procedure for the syntheses of tert-butyl ((2S)-
-(1H-indol-3-yl)-1-oxo-1-((phenyl(1-(2-(pyridin-2-yl)
3
To 3-methylpyridazine (500 mg, 5.313 mmol) in a micro-
wave vessel under nitrogen atmosphere, 37% formalde-
hyde (730 mg, 7.975 mmol), water (1 mL) and piperidine
ethyl)-1H-1,2,3-triazol-4-yl)methyl)amino)propan-2-yl)
carbamate (14) and tert-butyl ((2S)-3-(1H-indol-3-yl)-1-oxo
-1-((phenyl(1-(2-(pyridazin-3-yl)ethyl)-1H-1,2,3-triazol-4-yl)
methyl)amino)propan-2-yl)carbamate (15)
(
0.20 mL) were added. The reaction mixture was heated
at 165 °C for 30 min in a microwave synthesizer (300 W).
The solvent was then evaporated under reduced pressure
and the crude was purified by flash chromatography (eluent
mixture:dichloromethane/methanol, 9:1) to give the alcohol
To the solution of the alkyne 7 (0.5 mmol) and the oppor-
tune azide (0.5 mmol) dissolved in tert-butanol (120 mL),
CuSO (0.8 mmol), sodium l-ascorbate (2 mmol) and water
4
9
(259 mg, 2.086 mmol) as a yellow oil.
(60 mL) were respectively added. After stirring at room tem-
perature for 18 h, the mixture was extracted with dichlo-
romethane (3×40 mL) and washed with brine (1×20 mL).
The organic layer was dried and concentrated to reduced
pressure. The crudes were purified by flash chromatography
(eluent mixture:dichloromethane/methanol 95:5) obtaining
the required triazoles as yellow oils.
1
Yield: 40%; H NMR (300 MHz, CDCl ) δ 2.99 (br s,
3
1
H), 3.19 (t, J=6 Hz, 2H), 4.11 (t, J=6 Hz, 2H), 7.37–7.46
(
m, 2H), 9.03–9.07 (m, 1H).
General procedure for the syntheses of 2-(pyridin-2-yl)
ethyl methanesulfonate (10) and 2-(pyridazin-3-yl)ethyl
methanesulfonate (11)
1
Intermediate 14. Yield: 61%; H NMR (300 MHz,
CD OD) δ 1.09–1.25 (m, 1H), 1.38 (s, 9H), 3.01–3.12
3
To the opportune alcohol (2 mmol) in dichloromethane
(m, 1H), 3.14–3.23 (m, 1H), 3.28–3.32 (m, 2H), 3.51 (t,
(
5 mL), DIPEA (4 mmol) was added and the solution was
J=6.6 Hz, 2H), 4.40 (t, J=7.2 Hz, 1H), 6.13 (d, J=7.2 Hz,
cooled at 0 °C. Subsequently methanesulfonyl chloride
1H), 6.82–7.63 (m, 15H), 9.00 (s, 1H).
1
(
2.2 mmol) was added and the mixture was stirred at room
Intermediate 15. Yield: 92%; H NMR (300 MHz, CDCl )
3
temperature for 3 h. Afterwards water (5 mL) was added
and the organic layer was separated, washed with a solution
δ 1.39–1.45 (m, 10H), 3.11–3.22 (m, 1H), 3.27–3.40 (m,
3H), 4.53 (br s, 1H), 4.64–4.74 (m, 2H), 6.09–6.21 (m, 1H),
6.56–7.36 (m, 14H), 7.53–7.70 (m, 2H), 8.51–8.59 (m, 1H).
of NaHCO (5 mL), dried and concentrated under vacuum.
3
The obtained yellow oils were used in the next step with-
out further purification.
General procedure for the syntheses of (2S)-2-ami
no-3-(1H-indol-3-yl)-N-(phenyl(1-(2-(pyridin-2-yl)
ethyl)-1H-1,2,3-triazol-4-yl)methyl)propanamides (1,2)
and (2S)-2-amino-3-(1H-indol-3-yl)-N-(phenyl(1-(2-(pyrida
zin-3-yl)ethyl)-1H-1,2,3-triazol-4-yl)methyl)propanamides
(3,4)
1
Intermediate 10. Yield: quantitative; H NMR (300 MHz,
CDCl ) δ 2.95 (s, 3H), 3.42 (t, J = 6 Hz, 2H), 4.75 (t,
3
J=6 Hz, 2H), 7.39–7.48 (m, 2H), 9.08–9.14 (m, 1H);
1
Intermediate 11. Yield: 79%; H NMR (300 MHz, CDCl )
3
δ 2.87 (s, 3H), 3.20 (t, J=6.3 Hz, 2H), 4.64 (t, J=6.3 Hz,
2
1
H), 7.12–7.23 (m, 2H), 7.58–7.66 (m, 1H) 9.08–9.14 (m,
H).
To the solution of the opportune Boc-derivative (0.3 mmol)
in dichloromethane (3 mL) trifluoroacetic acid (3 mL) was
slowly added dropwise under stirring, at room temperature.
General procedure for the syntheses of 2-(2-azidoethyl)
pyridine (12) and 3-(2-azidoethyl)pyridazine (13)
After 3 h, the solution was treated with aqueous NaHCO
3
until pH 8.5 and the organic layer was dried and concen-
trated under vacuum. The crudes were purified by flash chro-
matography (eluent mixture:dichloromethane/methanol 9:1)
affording the expected products.
To a solution of the suitable mesylate (1.3 mmol) in DMF
(
3 mL) under stirring, NaN (4 mmol) was added and the
3
mixture was heated at 70 °C for 3 h, then the solvent was
evaporated under reduced pressure. The crudes were purified
by flash chromatography (eluent mixture:dichloromethane/
methanol 95:5) to give the azides as yellow oils.
Product 1. Chemical formula: C H N O. Molar mass:
2
6
26
8
−
1
466.55 g mol . Yield: 44%; light yellow solid; M.p. 230 °C
+
25
dec.; MS (ESI) m/z 489.4 [M + Na] . [α] = + 16.2
D
1
1
13
Intermediate 12. Yield: 63% (from intermediate 10); H
(c = 0.4 M, CH OH). H and C NMR spectra data are
3
NMR (300 MHz, CDCl ) δ 3.23 (t, J = 6.6 Hz, 2H), 3.83
reported in Table S1 in Electronic supplementary materi-
3
(
t, J=6.6 Hz, 2H), 7.39–7.48 (m, 2H), 9.08–9.14 (m, 1H);
als 2.
1
Intermediate 13. Yield: 57% (from intermediate 11); H
Product 2. Chemical formula: C H N O. Molar
2
6
26
8
−
1
NMR (300 MHz, CDCl ) δ 3.06 (t, J = 6.9 Hz, 2H), 3.71
mass: 466.55 g mol . Yield: 39%; light yellow
3
+
(
t, J=6.9 Hz, 2H), 7.13–7.22 (m, 2H), 7.58–7.66 (m, 1H),
solid; M.p. 182.4 °C; MS (ESI) m/z 489.4 [M + Na] .
2
5
1
13
9
.52–9.58 (m, 1H);
[α]
= + 9.3 (c = 0.4 M, CH OH). H and C NMR
3
D
1
3

Journal of Computer-Aided Molecular Design
spectra data are reported in Table S1 in Electronic sup-
plementary materials 2.
Conformational analysis
Product 3. Chemical formula: C H N O. Molar mass:
All the calculations were carried out using the GAUSS-
IAN09 program package [43]. Optimizations were per-
formed at the B3LYP/6-31G(d) level [30, 31], considering
all degrees of conformational freedom and on the basis of
experimental NOESY correlations. Vibrational frequencies
were computed at the same level of theory to verify that the
optimized structures were minima. The population percent-
ages were calculated through the Boltzmann equation. GIAO
NMR calculations were carried out at the B3LYP/6-31G(d)
level [30, 31].
2
7
27
7
−
1
4
65.56 g mol . Yield: 32%; light pink foam; MS (ESI)
+
25
1
m/z 466.2 [M + 1] . [α] = + 7.4 (c = 0.8 M, CH OH). H
D
3
1
3
and C NMR spectra data are reported in Table S1 in
Electronic supplementary materials 2.
Product 4. Chemical formula: C H N O. Molar mass:
2
7
27
7
−
1
4
65.56 g mol . Yield: 50%; light yellow foam; MS (ESI)
+
25
m/z 466.2 [M + 1] . [α] = + 13.02 (c = 0.8 M, CH OH).
D
3
1
13
H and C NMR spectra data are reported in Table S1in
Electronic supplementary materials 2.
Surface plasmon resonance assays
NMR spectroscopy
SPR runs were conducted on a Biacore T100 instrument (GE
−
1
NMR spectra of compounds 1–4 and 7 were recorded
Healthcare) at 25 °C with a flow rate of 50 µL min , unless
otherwise specified. Running buffer (RB) consisted of 1x
PBS (1 mM KH PO , 5.6 mM Na HPO , 154.5 mM NaCl,
with a Bruker AVANCE-500 spectrometer operating at
1
13
5
00.13 MHz for H or at 125.76 MHz for C NMR spec-
2
4
2
4
tra using a 5 mm z-PFG (pulsed field gradient) broadband
reverse probe. Chemical shifts are reported on the δ (ppm)
pH 7.4), 0.01% v/v P20, and varying amounts of DMSO
(Quality Biologicals). Binding and equilibrium constants
were determined with Scrubber (BioLogic™). A double-
referencing method was applied to correct for nonspecific
binding to the chip with interspersed blank injections cor-
recting for baseline drifts. Changes in refractive index due to
DMSO were accounted for with a DMSO calibration curve.
Furthermore, a positive control experiment using a defined
amount of the binding partner of Atg8 in the presence of
DMSO was interspersed every 4th injection. Using this
positive control allows monitoring the decay of the ligand
attached to the SPR chip. No decay was observed throughout
the experiments.
scale and are relative to residual CH OD at 3.30 ppm (cen-
3
1
tral line) for H NMR spectra, and relative to CD OD at
3
1
3
4
7.0 ppm (central line) for C NMR spectra. The data
were collected and processed by XWIN-NMR software
Bruker) running on a PC with Microsoft Windows xp.
Compounds (about 15 mg) were dissolved in CD OD
(
3
(
0.75 mL). The solutions were put in a 5 mm NMR tube
and the spectra recorded at 296 K. The signal assignment
was given by a combination of 1D and 2D (COSY, NOESY
and HSQC) experiments, using standard Bruker pulse
1
1
13
1
programs. The H– H and C– H bond correlations were
confirmed by COSY and HSQC experiment using Z-PFGs.
1
The pulse widths were 7.50 µs (90°) for H and 14.75 µs
P. falciparum blood stage culturing
1
3
(
90°) for C. Typically 32 K data points were collected
for one-dimensional spectra. Spectral width was 14.0 ppm
Cultures of P. falciparum 3D7 strain were maintained using
modified, previously published methods at 37 °C, 2% hema-
tocrit of human red blood cells [44]. Complete culture media
consisted of sterile RPMI 1640 media (Life Technologies)
supplemented with 10% human serum and 0.005% hypox-
anthine and buffered with final concentrations of 0.6%
1
(
7002 Hz) for H NMR (digital resolution: 0.21 Hz per
point).
D experiments parameters were as follows. For H– H
correlations: relaxation delay 2.0 s, data matrix 1 K ×1 K
512 experiments to 1 K zero filling in F , 1 K in F ), 2
1
1
2
(
1
2
or 32 transients in each experiment for COSY and NOESY
respectively, spectral width 9.00 ppm (4496.4 Hz). The
NOESY spectra were generated with a mixing time of 1.0 s
and acquired in the TPPI mode. There were not signifi-
cant differences in the results obtained at different mixing
HEPES and 0.26% NaHCO . The P. falciparum 3D7 strain
3
was maintained at 5% CO , 5% O , and 90% N atmosphere.
2
2
2
SYBR green I growth inhibition assay
1
3
1
times (2.0–3.0 s). For C– H correlations (HSQC): relaxa-
tion delay 2.5 s, data matrix 1 K × 1 K (512 experiments
to 1 K zero filling in F , 1 K in F ), six transients in each
Ten microliters of 10× compound diluted in RPMI 1640
media (Gibco) with a constant concentration of 1% DMSO
were added to a 96 well plate (Costar), 90 µL of 1.5% ring
stage, synchronized with 5% w/v sorbitol P. falciparum
3D7 parasites, 1% hematocrit, in culture media with 10%
1
2
experiment, spectral width 7.0 ppm (3501 Hz) in the proton
domain and 180.0 ppm (22,638 Hz) in the carbon domain.
A sinebell weighting was applied to each dimension. All 2D
spectra were processed with the Bruker software package.
−
1
v/v human serum with 10 µg mL gentamycin. Each com-
pound concentration and 1% v/v DMSO controls were run
1
3

Journal of Computer-Aided Molecular Design
in triplicate. Plates were incubated at 37 °C in 5% O , 5%
References
2
CO , and 90% N for 72 h. Plates were frozen, thawed,
2
2
and incubated with 100 µL 2× SYBR green in lysis buffer
1. World Health Organization (2017) http://www.who.int/mediacentr
e/factsheets/fs094/en/. Accessed 16 may 2017
(
20 mM Tris pH 7.5, 5 mM EDTA, 0.008% Saponin,
.08% TritonX-100) in the dark for at least 1 h. Fluo-
2
.
Ashley E, McGready R, Proux S, Nosten F (2006) Malaria.
Travel Med Infect Dis 4:159–173. https://doi.org/10.1016/j.tmaid
.2005.06.009
0
rescence was measured with a plate reader (SpectraMax
I3X, Molecular Dimensions) at excitation/emission wave-
lengths of 485/535 nm. Analysis was carried out using
GraphPad Prism and treating the DMSO vehicle control
as 100% growth and the 100 µM Chloroquine control as
3. Maeno Y, Culleton R, Quang NT, Kawai S, Marchand RP, Naka-
zawa S (2016) Plasmodium knowlesi and human malaria parasites
in Khan Phu, Vietnam: gametocyte production in humans and
frequent co-infection of mosquitoes. Parasitology 144:527–535.
https://doi.org/10.1017/S0031182016002110
0
% growth. Automatic outlier control was employed using
4. Dondorp AM, Nosten F, Yi P, Das D, Phyo AP, Tarning J, Lwin
KM, Ariey F, Hanpithakpong W, Lee SJ, Ringwald P, Silamut K,
Imwong M, Chotivanich K, Lim P, Herdman T, An SS, Yeung S,
Singhasivanon P, Day NPJ, Lindegardh N, Socheat D, White NJ
log (inhibitor) versus the normalized response.
(
2009) Artemisinin resistance in Plasmodium falciparum malaria.
MTT assay
New Engl J Med 361:455–467. https://doi.org/10.1056/NEJMo
a0808859
5
.
Dondorp AM, Yeung S, White L, Nguon C, Day NPJ, Socheat D,
von Seidlein L (2010) Artemisinin resistance: current status and
scenarios for containment. Nat Rev Microbiol 8:530–530. https
://doi.org/10.1038/nrmicro2385
The HepG2 cell line was bought from ATCC (HB-8065,
ATCC from LGC Standards, Milan, Italy) and was cul-
tured in DMEM high glucose with stable L-glutamine,
−
1
6. Beesley T, Gascoyne N, Knott-Hunziker V, Petursson S, Waley
SG, Jaurin B, Grundström T (1983) The inhibition of class C
β-lactamases by boronic acids. Biochem J 209:229–233. https://
doi.org/10.1042/bj2090229
supplemented with 10% FBS, 100 U mL penicillin,
−
1
1
00 µg mL streptomycin (complete growth medium)
with incubation at 37 °C under 5% CO atmosphere.
2
HepG2 cells were used for no more than 20 passages after
thawing, because the increase in number of passages may
change the cell characteristics and impair assay results. A
7. World Health Organization (2017) Malaria vaccine: WHO posi-
tion paper, January 2016—Recommendations. Vaccine. https://
doi.org/10.1016/j.vaccine.2016.10.047
8
.
Jana S, Paliwal J (2007) Novel molecular targets for antimalarial
chemotherapy. Int J Antimicrob Agents 30:4–10. https://doi.
org/10.1016/j.ijantimicag.2007.01.002
4
total of 3 × 10 HepG2 cells/well were seeded in 96-well
plates and treated with 100.0 40.0 and 1.0 µM of com-
pounds 1 and 2, respectively, or vehicle (DMSO) in com-
plete growth media for 24 and 48 h. Subsequently, the
treatment solvent was aspirated and 100 µL/well of MTT
filtered solution added for 2 h. After the incubation time,
9
.
Ettari R, Micale N, Grazioso G, Bova F, Schirmeister T, Grasso
S, Zappalà M (2012) Synthesis and molecular modeling studies
of derivatives of a highly potent peptidomimetic vinyl ester as
falcipain-2 inhibitors. ChemMedChem 7:1594–1600. https://doi.
org/10.1002/cmdc.201200274
−
1
10. Marco M, Coteron JM (2012) Falcipain inhibition as a promising
antimalarial target. Curr Top Med Chem 12:408–444
0
.5 mg mL MTT solution was aspirated and 100 µL/well
of MTT lysis buffer (8 mM HCl + 0.5% NP-40 in DMSO)
added. After 5 min of slow shaking, the absorbance at
1
1. Alvarez VE, Kosec G, Sant’Anna C, Turk V, Cazzulo JJ, Turk B
(2008) Autophagy is involved in nutritional stress response and
differentiation in Trypanosoma cruzi. J Biol Chem 283:3454–
5
70 nm and 630 nm (reference wavelength) were read by a
3
464. https://doi.org/10.1074/jbc.M708474200
microplate reader Synergy H1 from BioTek Germany (Bad
Friedrichshall, Germany).
1
2. Sinai AP, Roepe PD (2012) Autophagy in Apicomplexa: A life
sustaining death mechanism?. Trends Parasitol 28:358–364. https
:
//doi.org/10.1016/j.pt.2012.06.006
Acknowledgements We acknowledge the CINECA and the Regione
Lombardia award under the LISA initiative, for the availability of high
performance computing resources and support. We thank the Johns
Hopkins Malaria Research Institute parasite and insectary facility
for assistance with our experiments. This work was partially funded
through The Bloomberg Family Foundation (J.B.). We thank Professors
D. Taramelli and M. De Amici for helpful discussion. L. L. thanks the
University of Pavia for partial financial support.
13. Hain AUP, Weltzer RR, Hammond H, Jayabalasingham B,
Dinglasan RR, Graham DRM, Colquhoun DR, Coppens I, Bosch
J (2012) Structural characterization and inhibition of the Plasmo-
dium Atg8–Atg3 interaction. J Struct Biol 180:551–562. https://
doi.org/10.1016/j.jsb.2012.09.001
14. Besteiro S (2012) Which roles for autophagy in Toxoplasma gon-
dii and related apicomplexan parasites? Mol Biochem Parasitol
184:1–8. https://doi.org/10.1016/j.molbiopara.2012.04.001
1
5. Chen D, Lin J, Liu Y, Li X, Chen G, Hua Q, Nie Q, Hu X, Tan
F (2016) Identification of TgAtg8–TgAtg3 interaction in Toxo-
plasma gondii. Acta Trop 153:79–85. https://doi.org/10.1016/j.
actatropica.2015.09.013
Author contributions The manuscript was written through contribu-
tions of all authors. All authors have given approval to the final version
of the manuscript.
1
1
6. Geng J, Klionsky DJ (2008) The Atg8 and Atg12 ubiquitin-like
conjugation systems in macroautophagy. Embo Rep 9:859
7. Abada A, Elazar Z (2014) Getting ready for building: signaling
and autophagosome biogenesis. Embo Rep 15:839
Compliance with ethical standards
Conflict of interest The authors declare no competing financial inter-
18. Le Roch KG, Zhou Y, Blair PL, Grainger M, Moch JK, Haynes
JD, De La Vega P, Holder AA, Batalov S, Carucci DJ, Winzeler
ests.
1
3

Journal of Computer-Aided Molecular Design
EA (2003) Discovery of gene function by expression profiling of
the malaria parasite life cycle. Science 301:1503–1508. https://
doi.org/10.1126/science.1087025
32. Wolinski K, Hinton JF, Pulay P (1990) Efficient implementation of
the gauge-independent atomic orbital method for NMR chemical
shift calculations. J Am Chem Soc 112:8251–8260. https://doi.
org/10.1021/ja00179a005
1
9. Duszenko M, Ginger ML, Brennand A, Gualdrón-López M,
Colombo MI, Coombs GH, Coppens I, Jayabalasingham B, Lang-
sley G, Lisboa de Castro S, Menna-Barreto R, Mottram JC, Nav-
arro M, Rigden DJ, Romano PS, Stoka V, Turk B, Michels PAM
33. Ditchfield R (1974) Self-consistent perturbation theory of dia-
magnetism. Mol Phys 27:789–807. https://doi.org/10.1080/00268
977400100711
(
2011) Autophagy in protists. Autophagy 7:127–158. https://doi.
34. Case DA, Darden TA, Cheatham TE III, Simmerling CL, Wang
J, Duke RE, Luo R, Walker C, Zhang W, Merz KM, Roberts B,
Hayik S, Roitberg A, Seabra G, Swails J, Goetz AW, Kolossváry
I, Wong KF, Paesani F, Vanicek J, Wolf RM, Liu J, Wu X, Brozell
SR, Steinbrecher T, Gohlke H, Cai Q, Ye X, Wang J, Hsieh MJ,
Cui G, Roe DR, Mathews DH, Seetin MG, Salomon-Ferrer R,
Sagui C, Babin V, Luchko T, Gusarov S, Kovalenko A, Kollman
PA (2012) AMBER 12. University of California, San Francisco
35. Wang J, Wolf RM, Caldwell JW, Kollman PA, Case DA (2004)
Development and testing of a general amber force field. J Comp
Chem 25:1157–1174. https://doi.org/10.1002/jcc.20035
36. Jorgensen WL, Chandrasekhar J, Madura JD, Impey RW, Klein
ML (1983) Comparison of simple potential functions for sim-
ulating liquid water. J Chem Phys 79:926–935. https://doi.
org/10.1063/1.445869
org/10.4161/auto.7.2.13310
2
2
0. Hain AU, Miller AS, Levitskaya J, Bosch J (2016) Virtual screen-
ing and experimental validation identify novel inhibitors of the
Plasmodium falciparum Atg8–Atg3 protein–protein interaction.
ChemMedChem 11:900–910. https://doi.org/10.1002/cmdc.20150
0
515
1. Hain AUP, Bartee D, Sanders NG, Miller AS, Sullivan DJ, Lev-
itskaya J, Meyers CF, Bosch J (2014) Identification of an Atg8–
Atg3 Protein–Protein interaction inhibitor from the medicines
for malaria venture malaria box active in blood and liver stage
Plasmodium falciparum parasites. J Med Chem 57:4521–4531.
https://doi.org/10.1021/jm401675a
2
2. Scott DE, Bayly AR, Abell C, Skidmore J (2016) Small mol-
ecules, big targets: drug discovery faces the protein–protein inter-
action challenge. Nat Rev Drug Discov 15:533–550. https://doi.
org/10.1038/nrd.2016.29
37. Essmann U, Perera L, Berkowitz ML, Darden T, Lee H, Pedersen
LG (1995) A smooth particle mesh Ewald method. J Chem Phys
103:8577–8593
2
2
3. Valverde IE, Mindt TL (2013) 1,2,3-Triazoles as amide-bond sur-
rogates in peptidomimetics. Chimia 67:262–266
38. Berendsen HJC, Postma JPM, van Gunsteren WF, DiNola A, Haak
JR (1984) Molecular dynamics with coupling to an external bath.
J Chem Phys 81:3684–3690. https://doi.org/10.1063/1.448118
39. Jones G, Willett P, Glen RC, Leach AR, Taylor R (1997) Develop-
ment and validation of a genetic algorithm for flexible docking. J
Mol Biol 267:727–748. https://doi.org/10.1006/jmbi.1996.0897
40. Korb O, Stutzle T, Exner TE (2009) Empirical scoring functions
for advanced protein-ligand docking with PLANTS. J Chem Inf
Model 49:84–96. https://doi.org/10.1021/ci800298z
4. Stucchi M, Grazioso G, Lammi C, Manara S, Zanoni C, Arnoldi
A, Lesma G, Silvani A (2016) Disrupting the PCSK9/LDLR
protein–protein interaction by an imidazole-based minimalist
peptidomimetic. Org Biomol Chem 14:9736–9740. https://doi.
org/10.1039/C6OB01642A
2
2
5. Ni Z, Giordano L, Tenaglia A (2014) Cyclobutene formation
in PtCl2-catalyzed cycloisomerizations of heteroatom-tethered
1
,6-enynes. Chem-Eur J 20:11703–11706. https://doi.org/10.1002/
chem.201403643
41. Dallanoce C, Magrone P, Bazza P, Grazioso G, Rizzi L, Riganti
L, Gotti C, Clementi F, Frydenvang K, De Amici M (2009) New
analogues of epiboxidine incorporating the 4,5-dihydroisoxazole
nucleus: synthesis, binding affinity at neuronal nicotinic acetyl-
choline receptors, and molecular modeling investigations. Chem
Biodivers 6:244–259. https://doi.org/10.1002/cbdv.200800077
42. Nilsson BM, Vargas HM, Ringdahl B, Hacksell U (1992) Phenyl-
substituted analogues of oxotremorine as muscarinic antagonists.
J Med Chem 35:285–294
6. Masciocchi D, Gelain A, Porta F, Meneghetti F, Pedretti A, Celen-
tano G, Barlocco D, Legnani L, Toma L, Kwon B-M, Asaid A,
Villa S (2013) Synthesis, structure–activity relationships and
stereochemical investigations of new tricyclic pyridazinone
derivatives as potential STAT3 inhibitors. MedChemComm
4
:1181–1188
2
2
2
7. Toma L, Legnani L, Rencurosi A, Poletti L, Lay L, Russo G
(
2009) Modeling of synthetic phosphono and carba analogues of
N-acetyl-alpha-D-mannosamine 1-phosphate, the repeating unit of
the capsular polysaccharide from Neisseria meningitidis serovar
A. Org Biomol Chem 7:3734–3740. https://doi.org/10.1039/b9070
43. Frisch MJ, Trucks GW, Schlegel HB, Scuseria GE, Robb MA,
Cheeseman JR, Scalmani G, Barone V, Mennucci B, Petersson
GA, Nakatsuji H, Caricato M, Li X, Hratchian HP, Izmaylov AF,
Bloino J, Zheng G, Sonnenberg JL, Hada M, Ehara M, Toyota K,
Fukuda R, Hasegawa J, Ishida M, Nakajima T, Honda Y, Kitao
O, Nakai H, Vreven T, Montgomery JJA, Peralta JE, Ogliaro F,
Bearpark M, Heyd JJ, Brothers E, Kudin KN, Staroverov VN,
Kobayashi R, Normand J, Raghavachari K, Rendell A, Burant JC,
Iyengar SS, Tomasi J, Cossi M, Rega N, Millam JM, Klene M,
Knox JE, Cross BJ, Bakken V, Adamo C, Jaramillo J, Gomperts
R, Stratmann, ER, Yazyev O, Austin AJ, Cammi R, Pomelli C,
Ochterski WJ, Martin RL, Morokuma K, Zakrzewski VG, Voth
GA, Salvador P, Dannenberg JJ, Dapprich S, Daniels AD, Farkas
Ö, Foresman JB, Ortiz JV, Cioslowski J, Fox DJ (2009) Gauss-
ian09. Revision A.02
0
0a
8. Luparia M, Legnani L, Porta A, Zanoni G, Toma L, Vidari G
(
2009) Enantioselective synthesis and olfactory evaluation of
bicyclic alpha- and gamma-ionone derivatives: the 3D arrange-
ment of key molecular features relevant to the violet odor of ion-
ones. J Org Chem 74:7100–7110. https://doi.org/10.1021/jo901
4
936
9. Legnani L, Colombo D, Venuti A, Pastori C, Lopalco L, Toma L,
Mori M, Grazioso G, Villa S (2017) Diazabicyclo analogues of
maraviroc: synthesis, modeling, NMR studies and antiviral activ-
ity. MedChemComm 8:422–433. https://doi.org/10.1039/C6MD0
0
575F
3
3
0. Lee C, Yang W, Parr RG (1988) Development of the Colle-Salvetti
correlation-energy formula into a functional of the electron den-
sity. Phys Rev B 37:785–789
44. Trager W, Jensen JB (1976) Human malaria parasites in continu-
ous culture. Science 193:673–675
1. Becke AD (1993) Density-functional thermochemistry. III. The
role of exact exchange. J Chem Phys 98:5648–5652. https://doi.
org/10.1063/1.464913
1
3