Screening of possible antiviral peptides to bind SARS Covid 19 spike protein

Content
Intoduction
Sructure of SARS-CoV-2
Aim of the study
Methods
VMD
CABS Dock
5. Results

Introduction
1 https://www.rki.de/DE/Content/InfAZ/N/Neuartiges_Coronavirus/Steckbrief.html
2 https://www.nature.com/articles/d41586-020-02544-6
Single-stranded RNA-enveloped virus
It bends to the angiotensin converting enzyme 2 (ACE2)
Basic reproduction number (R0) is around 3,8 [1]
accumulates two single-letter mutations per month [2]

Sructure of SARS-CoV-2 spike protein
spike protein mediates the membrane fusion process
Spike protein has 2 subunits – S1 and S2
S1 catalyzes attachment,
S2 -subunit fusion
S2 forms a six-helical bundle via the two-heptad repeat domain, HR1 and HR2 (“fusion core region”).
S2 subunit

Antiviral drugs
Classes:
Spike maturation inhibitor
Protease inhibitor
Fusion inhibitor
Polymerase inhibitor
Only Remdesivir was approved by FDA, but was later exluded from the guidline [4]
4. https://apps.who.int/iris/handle/10665/336729

Aim of the study
to screen a list of peptides which were designed to bind the HR1 domains
A peptide with the largest number of contacts with the HR1 domains would inhibit the membrane fusion, and therefore infection

Methods

VMD - Visual Molecular Dynamics
for structure visualization
displaying, animating, and analyzing large biomolecular systems using 3-D graphics

CABS Dock
for protein-peptide docking
coarse-grained model (it decreases a time of long simulations)
advantages:
Can be used without knowing the binding site and peptide conformation
Peptide conformation is allowed to be fully flexible
It is possible to simulate significant conformational changes

CABS Dock
Steps:
Generating random structures
Simulation of binding and docking
Selection of the final models

Results
Original structure:
Fraction of contacts – 51 % (cutoff – 5Å)
number of residues in chain D, that are in contact – 20
Chain D of HR2 has 41 amino acids
The docking was run with a shorten version of the peptide HR2 (chain D): VVNIQKEIDRLNEVAKNLNESLID:CCCHHHHHHHHHHHHHHHHHHCCC
24 amino acids, 18 of them form helices,
The docking was run 2 times, first time with a number of cycles 100, second time – 200,
chain C was excluded.
Parameters (100 cycles): model 5 (fraction 62, contacts 15), model 6 (fraction 70, contacts 17).
Parameters (200 cycles): model 4 (fraction 75, contacts 18), model 6 (fraction 66, contacts 16),
model 9 (fraction 54, contacts 13).
Parameters (200 cycles+residues beside helix exluded): model 1 (fraction 62, contacts 15),
model 4 (fraction 70, contacts 17)
Chain A and B
Chain C
Chain D
HR1
HR2

Original Model 4
200 cycles were also chosen for docking the derivatives.

10 derivatives were tested (additional parameter - all amino acids form helices)
Best peptide structures need to have higher fraction of contacts than in original peptide (>51)
First number-number of contacts
Second number-fraction of contacts

Ranking according to the predicted number of models, fractions and position of a modal in a rank:
1. DERIVATIVE_3630
2. DERIVATIVE_1086
3. DERIVATIVE_3200
4. DERIVATIVE_3494
5. DERIVATIVE_6089
6. DERIVATIVE_7060
7. DERIVATIVE_9119
8. DERIVATIVE_6154
9. DERIVATIVE_218
10. DERIVATIVE_7407

PPI-Affinity
The most promising candidates:
DERIVATIVE_1086
DERIVATIVE_3200
DERIVATIVE_3630
original
-8.4

Vmd images of top 3 candidates
DERIVATIVE_3630,
Model_full 0 (73%, 17 contacts)
DERIVATIVE_1086,
Model_full 0 (66%, 14 contacts)
DERIVATIVE_3200
Model_full 2 (57%, 12 contacts)