(T1430-01-02) Coarse-Grained Molecular Dynamics Simulations of ssDNA Loaded Adeno-Associated Virus

Purpose: Adeno-associated virus (AAV) packaging a single-stranded (ssDNA) genome has emerged as a promising vector and is the most common vehicle for in vivo gene transfer. They are well established in clinical trials for in vivo gene therapy with a good safety profile, low immunogenicity, and long-term stability. However, little is known about the mechanism for the aggregation of AAV capsids during purification, their tendency to be absorbed by hydrophobic materials during the manufacturing process and storage, and the role of free ssDNA in the aggregation and adsorption processes. Therefore, we implemented coarse-grained molecular dynamics (CG-MD) simulations to probe the underlying mechanism of the interactions between capsids as well as the adsorption process of the capsid and hydrophobic surfaces, such as polyethylene (PE) and polypropylene (PP).

Methods: The CG-MD simulations and their analyses were implemented using GROMACS software package with the MARTINI force-fields¹. The AAV Serotype 8 (AAV8) was chosen as the model viral capsid (PDB: 6v12) with 2.2kb ssDNA, and the CHARMM-GUI² was used to convert it to CG model. The initial ssDNA model was constructed using the Avogadro software. The steepest descent algorithm was used in energy minimization followed by a 10 ns equilibration step in isothermal−isochoric ensemble at ambient temperature. The simulation production runs were performed in isothermal-isobaric ensemble using the Nosé−Hoover thermostat and the Parrinello−Rahman barostat at a pressure of 1 bar for each 20-fs time step. The simulations were carried out using periodic boundary conditions. Computations were performed in High Performance Center Supercomputer Cluster.

Results: The CG-MD simulations revealed the general behavior of capsid aggregation. The full capsid required more time to stabilize than the empty capsid under the same temperature and pressure conditions. This has been demonstrated by the analysis of the root-mean-square deviation (RMSD) and radius of gyration for the entire capsid, as well as root-mean-square fluctuation (RMSF) for the individual residues. Furthermore, a larger radius was captured for the full capsid than that of empty capsid, which is a resultant of the increased core size of loaded ssDNA molecule. In addition, due to the encapsulated negatively charged ssDNA, the full capsid has more anionic charges on the surface than that of the empty capsid, which was evaluated by the electrostatic potential map. Moreover, the behavior of a full capsid adsorbed onto the hydrophobic surface was captured successfully. It was noted that the capsid was rotated, driven toward, and attached to the surface with stronger interactions on the PP surface than that of the PE surface, which was assessed by the relative shape anisotropy.

Conclusion: CG-MD simulations established a better understanding of capsid aggregation associated with multiple asparagine and glutamine residues. Furthermore, the adsorption mechanisms were revealed to be related to multiple hydrophobic residues, such as leucine, isoleucine and phenylalanine. By using this developed computational model, we were able to evaluate the stability and electrostatic charges of the capsids and free ssDNA molecule. Our simulations provide general dynamical information about the structure of the ssDNA and capsids on the microsecond timescale, revealing behavior that is impossible to be obtained from experimental studies. The CG-MD simulations are demonstrated to be a powerful tool in the simulation of macromolecular complexes and capable of guiding the interpretation, explanation, and direction of experiments.

References: 1. Marrink, Siewert J., H. Jelger Risselada, Serge Yefimov, D. Peter Tieleman, and Alex H. De Vries. "The MARTINI force field: coarse grained model for biomolecular simulations." The journal of physical chemistry B 111, no. 27 (2007): 7812-7824.
2. Choi, Yeol Kyo, Tibo Duran, and Wonpil Im. "CHARMM-GUI Synthetic Polymer Modeler for Modeling and Simulation of Synthetic Polymers." Biophysical Journal 116, no. 3 (2019): 572a.

Acknowledgements: We are grateful to Sarepta Therapeutics for providing funding for this research. The authors also thank HPC facility of University of Connecticut.