Exploring a bioinformatics approach to assess the binding affinities of potential therapeutic decoys for COVID-19

In a recent study published in Scientific reportsthe researchers developed a computational workflow based on molecular dynamics (MD) simulations and an artificial neural network (ANN) to evaluate the severe acute respiratory syndrome viral syndrome 2 (SARS-CoV-2) spike (S) protein receptor-binding domain (RBD)-angiotensin-converting enzyme Human 2 (hACE2) variants associated with SARS-CoV-2.

Study: Improving therapeutic decoys for SARS-CoV-2 variants using deep learning-guided molecular dynamics simulations.  Image credit: CROCOTHERY/Shutterstock
Stady: Improving therapeutic decoys for SARS-CoV-2 variants using deep learning-guided molecular dynamics simulations.. Image credit: CROCOTHERY/Shutterstock


Studies have reported that S-hACE2 binding interactions facilitate SARS-CoV-2 entry and subsequent replication in the host. Thus, coronavirus disease 2019 (COVID-19) can be prevented by inhibiting S-ACE2 binding.

Accordingly, human soluble ACE2 (hsACE2) that binds to pre-entry SARS-CoV-2 viruses may prevent COVID-19; However, the approach requires optimization and adaptation to new SARS-CoV-2 variants.

about studying

In this study, the researchers devised a workflow by combining normal methods with cloud-based technology in order to optimize the development of therapeutic baits specific for the SARS-CoV-2 variant.

MD simulations were performed to identify enzyme-2 substitutions of human angiotensin-converting amino acids that enhance S RBD-hACE2 interactions, for which ESF (experimental register function) was used in close relationship with the LIE (linear reaction energy) technique. in the laboratory SARS-CoV-2 neutralization assays were performed to evaluate the inhibition of wild-type SARS-CoV-2 strain and transduction of the beta variant by hACE2 variants that were bound to the amorphous region (Fc) of human immunoglobulin G1 (hACE2-Fc).

Some variants of hACE2-Fc were also expressed in Nicotiana bentamiana Factory to investigate the feasibility of large-scale production. Molecular dynamics running data were combined with hACE2 halos and S RBD halos to train the ANN (Artificial Neural Network). The model was used to estimate the binding affinities of SARS-CoV-2 S with hACE2 variants based on the S halos RBD and hACE2. Should a new variant emerge, hACE2 variants can be quickly screened by the artificial neural network and verified by MD simulations so that COVID-19 treatment strategies can be designed based on the human soluble ACE2 variant with the greatest affinity for novel SARS-CoV. -2 strain.

The potential of the system was evaluated to estimate the effects of mutations of the S RBD variant of the same hACE2 decoys using the BA.1 and BA.2 subvariants of the SARS-CoV-2 Omicron variant as examples. All potential hACE2 mutations were screened, and 300 promising estimates were validated by MD simulations. In addition to wild-type hACE2, promising pfhACE2 variants, together with a C-terminal human IgG Fc tag, were also expressed in Chinese hamster ovary (CHO) cells.

SARS-CoV-2 RNA was quantified by quantitative reverse transcription polymerase chain reaction (RT-qPCR) and immunohistochemistry (IHC) analysis. The SARS-CoV-2 neutralizing potential of hACE2 variants was expressed in Nicotiana bentamiana (hACE2-Fc K31W_NB) plant leaves were tested using enzyme-linked immunosorbent assays (ELISA). in silico Analyzes were performed to evaluate the binding affinities of hACE2 variants with Omicron BA.3, BA.4/5 and Omicron BA.2.75 RBD proteins.

The crystal structure of wild type SARS-CoV-2 S RBD bound to hACE2 was downloaded from the Protein Data Bank (PDB) database. The ΔG value estimated in the model was calculated based on the electrostatic and van der Waals forces. The sequences used for ANN training consisted of S RBD (n = 1165) and hACE2 (n = 95) sequences retrieved from visual inspection, literature search, or the Global Initiative for Sharing All Influenza Data (GISAID) database by January 4, 2022.


The hACE2-Fc K31W, hACE2 T27Y_L79T_N330Y_K31W and hACE2 T27Y_L79T_K31W hACE2 variants were identified as high affinity candidates. Produced candidates benthamiana showed a 5.0-fold and 6.0-fold lower IC50 (semi-maximal inhibitory concentration) compared to the same variant produced in CHO and wild-type hACE2-Fc cells, respectively. The results indicated that hACE2-Fc variants with the correct fold could be produced in f benthamiana And plant-produced soluble ACE2 variants represent a promising and cost-effective therapeutic option against SARS-CoV-2.

The ESF estimates have been validated in the laboratory by virus neutralization assays. The experimental data correlated well with the estimate of Gadvance (Gibbs free energies) in the form. Compared with wild-type hACE2, the majority of hACE2 variants showed enhanced binding affinities with the SARS-CoV-2 Beta variant, the Delta variant, Omicron’s BA.1 variant, and the BA.2 variant. hACE2-K31W was the only mutant with lower Gibbs free energy, suggesting that the K31W mutant may contribute to S RBD interactions. The K31W mutation is observed in most of the high-affinity-association mutants.

Variants with 3.0 to 5.0 mutations showed the greatest association with S RBD. hACE2 T27Y_L79T_K31W and hACE2 T27Y_L79T_N330Y_ K31W showed significantly high binding affinities for the BA.2 S RBD (ΔGadvance value −71.0 kJ/mol) compared to hACE2 wild type (−52.0 kJ/mol). With estimated binding affinities of −62.0 and 67.0 kJ/mol, the hACE2 variants T27Y_L79T_K31W and hACE2 T27Y_L79T_N330Y_K31W were the highest-affinity variants of BA.3, and the binding affinities of Omicron BA.4/5 and Omicron BA.2.75 were the lowest. The highest outliers (MD G values ​​of <−70 kJ/mol) were mapped by the model, to the highest binding affinity value observed.

The results indicated that ANN was not only able to better estimate values ​​closer to the bulk of the binding affinity distribution than extrapolated from closely related variables but also reliably map high-affinity variants to the highest affinity bracket of −68.0 kJ/mol. An artificial neural network can learn meaningful physical insights from Halos with much better performance than simply learning a regression-to-average or transcription function, and the model can combine insights gained from relatively different inputs (distant SARS-CoV-2 sequences). The model identified single mutations that were comparable to the best hACE2 mutations found in primary MD runs.

Overall, the results of the study shed light on a bioinformatics approach to combine MD simulations, in the laboratory Competitive inhibition assays, live virus infection assays, and ANN, for the rapid, cost-effective, and effective assessment of binding affinity of hACE2 decoys of novel SARS-CoV-2 strains in an upstream stage, reducing hACE2-decoy adaptation periods and sample requirements for in the laboratory selections.

Source link

Related Posts