WISDOM: A Grid-Enabled Drug Discovery Initiative against Malaria - Grid Computing: Infrastructure, Service, and Applications - page 364

Following the successful test, the 5000 best compounds coming out of the
i rst data challenge on plasmepsin II were rei ned and reranked. Computing
resources from the EGEE Biomed virtual organization were used exclu-
sively; 100 CPUs were used in parallel, all of them belonged to our local
cluster at Clermont Ferrand, due to licensing issues. One single simulation
was consuming ~20 CPU minutes on an Intel Xeon 3.05 machine. The
estimated CPU time if the simulations were to be performed on one
machine was therefore expected to be 124 days. By using EGEE infrastruc-
ture, the simulation time was signii cantly brought down to 7 days.

14.3.3.4 Results

After rescoring the 5000 best docking results by Molecular Dynamics
with Amber and MM-PBSA and MM-GBSA, the next step was to select the
best compounds in the perspective of in vitro tests. The starting points
were the two ranked lists of compounds, one according to MM-PBSA and
the other according to MM-GBSA free energies. One hundred complexes of
each list were analyzed manually. Each complex was visualized in 3D with
UCSF Chimera software in order to determine the molecular interactions
between protein and ligand in the complex. The major criteria for selec-
tion were the ligand-making interactions to the two catalytic residues of
plasmesin II, Asp 34 and Asp214. Second, the interaction with other key
amino acids was checked: Gly36, Val78, Gly216, Ser79. The complexes with
no interaction with at least one of the two amino acids of the catalytic
dyad were rejected. The complexes that were kept had at least one main
interaction to amino acids of the catalytic dyad. In total, 30 out of 200 com-
pounds were selected for in vitro tests.

14.3.4 In Vitro Tests

A subset of the 30 compounds selected for testing showed submicromolar
or nanomolar IC 50 values against recombinant P. falciparum plasmepsin II,
using the inhibition assay based on FRET substrate degradation, which is
well documented in the literature [16,17]. In the parasite, plasmepsin II is
translated as an inactive zymogen containing a 124 amino acid-long
N-terminal prosequence that has a membrane-spanning domain. Within
the food vacuole the prosequence is removed by a calpain-like maturase,
and active plasmepsin II is released [18]. Plasmepsin II was expressed well
from a pET3d construct that contained Glu124 after initiator Met, and the
inclusion body of the protein was refolded and purii ed to near homoge-
neity as judged by SDS-PAGE. Based on molecular mass standards, recom-
binant plasmepsin II migrates at about 37 kDa.
Pepstatin A, a general inhibitor of aspartic proteases of microbial origin
[19], was also reported to inhibit hemoglobin degradation by extracts
of digestive vacuoles of P. falciparum [20]. In the in vitro inhibition test, the