4. Discussion
Antibody mimetics is a new technique based on highly-structured
scaffolds originating from natural proteins. Mutation in their key
residues, especially in variable loops, can alter their affinity towards
new targets [18]. In the present study, we conducted a CDR grafting
technique to find new binders with a high affinity toward EGFR. In our
report, we retrieved the extracellular domain 3 of EGFR from
EGFR/panitumumab complex (5SX4) to be used as the target in our
structure-based virtual screening method for 36 CDR-acceptors designed
by loop alteration technique. Further, the docking hits were applied for
molecular dynamic (MD) simulation, and the result was analyzed by
different parameters such as RMSF, H-bond, and MMPBSA. Therefore, the
current study was undertaken to analyze the behavior of the designed
complexes using protein-receptor complex against domain III EGFR. In
addition to the scFv structure originating from panitumumab, three
grafting scaffolds among all designed ones were selected to be docked
against EGFR DIII. Then, the complex of protein-EGFR DIII was simulated
through molecular dynamic analysis to be compared in their binding site.
Based on the previous research on the crystal structure of the
panitumumab-EGFR complex[19], three CDR loops H2, H3 from the heavy
chain, and L3 from the light chain of panitumumab with the most
effective interaction against receptor selected to be grafted in new
scaffolds. The docking scores and analyzing the CDR loop positions
against receptor were used as the main criteria for the binder
selection. Therefore, we focused on the three top hits with the best
docking scores that share a similar binding site to the reference
molecules. The epidermal growth factor protein binding by EGFR occurs
within a cleft between two extracellular domains, DI and DIII [20].
The binding site of panitumumab partially overlaps the EGF binding site
and prevents the dimerization of the EGFR ectodomain and its activation
[19]. The overall binding of panitumumab to domain III is similar to
cetuximab. Based on a mutational analysis, Voigt et al. described
critical amino acids overlap in panitumumab and cetuximab epitopes with
EGF binding site: K443 and D355 in panitumumab and D355, K443, Q408,
H409, S468 in cetuximab binding site [21]. The panitumumab complex
with EGFR DIII shows that upon binding, all 6 CDRs in the heavy and
light chains interact without any conformational alteration. According
to previous studies, CDRs L3, H2, and H3 commonly make the largest
contribution to antigen binding [9]. The L3 CDR lines up adjacent to
the final B-strand of domain III with two interactions of D92 and L94.
The heavy chain showed most of the specific hydrogen bond interactions
made by H2 and H3. Most notably, three tyrosine residues interacted with
the B-strand of domain III at the same position as EGF. Y54, Y55, N58,
and T59 in H2 interacted with N384, D420, K443, respectively, while in
H3, D100 has a water-mediated hydrogen bond with S468. H3 makes an
additional hydrogen bond between T103 and K465 of domain III [19].
It can be visualized that interaction with S468 had the highest
interaction rate (Fig.5). This residue interacts with EGF, panitumumab,
cetuximab, scFv, and all three designed scaffolds. The scFv structure
interacted with critical residues of S468, Q408, and H409 through 5
H-bonds in CDR regions which overlap with the cetuximab binding site.
Scaf1 interacted with the same epitope as cetuximab through 3 H-bonds in
inserted CDR loops against S468 and Q408. Scaf2, in addition to S468,
interacted with D355, which is involved in both panitumumab and
cetuximab binding sites. It indicates that this scaffold mimics the
function of these two antibodies. In contrast to the other two
scaffolds, Scaf3 could not make any hydrogen interaction in CDR regions
with critical residues in the panitumumab binding site. However, Q1 in
Scaf3 interacted with S468 and made hydrophobic contact with F30 against
F412 in the panitumumab binding site. It shows that Scaf3 can be used as
an alternative to panitumumab but with lower binding affinity.
Therefore, all three scaffolds that exhibit the interaction with the
same amino acid residues (S468, D355, Q408, and H409) are subjected to a
50 ns simulation process. The MD results confirmed the stability of scFv
and three selected scaffolds throughout the simulation. The RMSF was
calculated for domain III of EGFR, scFv, and three designed binders. It
refers to the stability of the complex as high fluctuations related to
more flexible and unstable bounds. Although there are some fluctuations
in Scaf2 amino acid residues, they are not at the inserted CDR loops nor
involved in protein interaction. The fluctuation during all interactions
was below 0.2 nm, which is totally acceptable. To validate the docking
energy of the protein-receptor complex, an MMPBSA calculation was
performed. The designed structures presented comparatively acceptable
MM-PBSA scores compared to the scFv structure. The calculated binding
free energy of these binders were -26.4, -39.89, -9.47, -13.51 KJ
mol-1, for scFv, Scaf1, Scaf2, and Scaf3,
respectively. Therefore, they represent excellent candidates for further
investigation in vitro analysis. The only exception is the Scaf3 complex
which showed a slightly lower MMPBSA value which is in accordance with
its visual analysis results. By analyzing the binding energy and
stability through dynamic simulation, we have shown that scFv, Scaf1,
and Scaf2 may be potential binders that mimic the function of
panitumumab. We tested in vitro and in vivo function of
scFv structure against EGFR (unpublished data), which confirmed the
results of in silico report. However, we also believe that the
function of the other three designed scaffolds in vitro also should be
investigated to approve their binding potential to EGFR protein.