P02-04
Interaction Analysis between pHLA and TCR using MD Simulation and Fragment Molecular Orbital Calculation
Suzu ITAMI *1, Yoshiki ARITSU2, Mizuki KITAMATSU3, Chihiro MOTOZONO2, Norihito KAWASHITA3
1Graduate School of Science and Engineering, Kindai University
2Joint Research Center for Human Retrovirus infection, Kumamoto University
3Faculty of Science and Engineering, Kindai University
( * E-mail: 2333310130r@kindai.ac.jp )
Keywords: Fragment Molecular Orbital Method, Molecular Dynamics, SARS-CoV-2
The T cell receptor (TCR) on cytotoxic T cells recognizes peptide fragments derived from viral proteins presented on HLA class I molecules on virus-infected cells as antigens. T cells that sense virus-infected cells play an important role in controlling viral infection by killing these infected cells. In mRNA vaccines used for novel coronavirus infections, it has become clear that vaccine-induced T cells are also involved in controlling viral infection. In future vaccine boosters, selective induction and reactivation of functional T cells may be important for controlling infection and preventing severe disease.
The 9-residue peptide NF-9 (NYNYLYRLF), derived from the SARS-CoV-2 spike protein, forms a complex (pHLA) with HLA, which is recognized by the TCR. The delta/omicron BA.5 variant also has an NF9-5R (NYNYRYRLF) containing a Leu5Arg mutation, and pHLAs with this mutation have been found not to bind to the TCR[1]. In this study, MD simulations and Fragment Molecular Orbital (FMO) calculations[2] were performed on pHLA before and after the mutation to elucidate the molecular mechanism by which with the Leu5Arg mutation evades TCR recognition.
Based on the crystal structure of HLA/NF-9, two structures, WT and NF9-5R, were created using MOE 2020. Force field were applied using Ambertools22 and GROMACS 2021.5 was used for MD simulations and their analysis. Simulations of 50 ns were run three times for each structure, sampling every 1 ns between 10-50 ns obtaining 41 structures per simulation. FMO calculations were performed on the sampled structures, and IFIE (Inter-Fragment Interaction Energy) and PIEDA (Pair Interaction Energy Decomposition Analysis) component values[3] were calculated using ABINIT-MP Open Ver. 1 Rev. 22 at a calculation level of MP2/6-31G*.
MD simulation results showed that the RMSF of Cα atoms of the 4th-6th residues in NF-9 was larger in NF9-5R, indicating that the mutation made it more prone to fluctuations. The average distance between hydroxyl group of Tyr6 and carbonyl group of Asn3 in NF9-5R was about 2 Å smaller than in WT. FMO calculations showed that the average ES values of Tyr4 and Tyr6 were −4.9±6.4 kcal/mol in WT and −12.6±8.8 kcal/mol, indicating that the mutation results in stronger electrostatic interactions. These results suggest that Tyr6 of NF9-5R is more likely to form hydrogen bonds with Asn3, thus making it less likely to face the TCR side and no longer bind to the TCR.
This study was conducted as part of the activities of the FMO Drug Design Consortium (FMODD), and the supercomputer "Fugaku" was used for MD simulations and FMO calculations (project numbers: hp230131, hp240162).
[1] Motozono C., et al., Cell Host & Microbe. 29(7), 1124-1136.e11, (2021).
[2] Kitaura K., et al., Chem Phys Lett. 313(3-4), 701-708, (1999).
[3] Fedorov D. G., et al., J. Comput. Chem. 28(1), 222-237, (2007).