Evolutionary artificial intelligence based peptide discoveries for effective Covid-19 therapeutics

Abstract

An epidemic caused by COVID-19 in China turned into pandemic within a short duration affecting countries worldwide. Researchers and companies around the world are working on all the possible strategies to develop a curative or preventive strategy for the same, which includes vaccine development, drug repurposing, plasma therapy, and drug discovery based on Artificial intelligence. Therapeutic approaches based on Computational biology and Machine-learning algorithms are specially considered, with a view that these could provide a fast and accurate outcome in the present scenario. As an effort towards developing possible therapeutics for COVID-19, we have used machine-learning algorithms for the generation of alignment kernels from diverse viral sequences of Covid-19 reported from India, China, Italy and USA. Using these diverse sequences we have identified the conserved motifs and subsequently a peptide library was designed against them. Of these, 4 peptides have shown strong binding affinity against the main protease of SARS-CoV-2 (M^pro) and also maintained their stability and specificity under physiological conditions as observed through MD Simulations. Our data suggest that these evolutionary peptides against COVID-19 if found effective may provide cross-protection against diverse Covid-19 variants.

Keywords: Artificial intelligence; Computational biology; Covid-19; Evolutionary peptides; SARS-CoV-2.

Graphical abstract

Fig. 1

(a) Life cycle of SARS-CoV-2. (1) The replication of SAR-CoV-2 begins with the binding of its spike protein (S) with receptor angiotensin-converting enzyme 2 (ACE2) on the cell surface molecules of the host. (2) The viral S protein undergoes conformational changes leading to a fusion between the viral envelope with the plasma or endosomal membranes, entering the cell's cytoplasm, and releasing RNA. (3) The positive-strand viral RNA serves as mRNA for translation of ORF 1a and 1b, encoding units of the RNA-dependent RNA polymerase. (4) RNA polymerase transcribes full-length complementary negative-sense RNA. Viral proteins (S, M, E, and N) get translated (5) The proteins are further processed in the Endoplasmic Reticulum and Golgi Apparatus. (6) Packaging of viral proteins and positive-strand RNA. (7) Assembling complete virion surrounded by the plasma membrane. (8) Apoptosis of cells with the release of the virion. (b) Role of Main protease M^pro in SARS-CoV-2.

Fig. 2

Various therapeutic strategies underway against COVID-19 (a) and their proposed outcomes (b).

Fig. 3

Overview of the algorithm used for peptide designing and filtering.

Fig. 4

Docking results of identified sequences with proteases wherein peptide is in yellow color and receptor in brown. The upper panel shows the docked poses with the 4 shortlisted peptides (protein represented in brown ribbon form while peptide in yellow color) and the lower panel represents their respective protein-peptide interactions.

Fig. 5

M^pro residues interacting with peptides. (a) to (d) are protein-peptide complexes before MD Simulations and (e) to (h) are protein-peptide complexes post MD Simulations. Peptide is represented in orange stick form while residues of M^pro protein interacting with peptides are shown in cyan color.

Fig. 6

Short scale Molecular Dynamics Simulation studies to check the stability, strength and amount of interactions being maintained between protein and peptide. The first row represents the protein – ligand (here peptide) RMSDs with protein in blue and ligand in red color. The middle row represents the type and fraction of interactions maintained by various protein residues during simulation time. H-bond is represented in green, hydrophobic in purple, ionic in pink and water bridges in blue color. The bottom row also represents the extent of protein-peptide interactions for a given residue at a particular simulation time. In the timeline graph, darker the color more is the fraction of contacts between protein and peptide.