Recombinant adeno-associated viruses (rAAV) are leading drugs in gene therapy with three products approved by the European Medicines Agency (EMA) and the Food and Drug Administration (FDA). Currently, rAAV vectors are evaluated in many ongoing clinical trials for treating rare genetic diseases and other therapies. Mammalian and insect cells have been utilized for rAAV production, but scaling to for the rapidly growing demand for rAAV in clinical trials remains challenging. Thus, a new AAV production technology is urgently needed. Meanwhile, AAV empty capsids as virus-like particles (VLP) produced in mammalian cells have raised increasing interest for vaccine development and drug delivery. This thesis established a new approach to produce AAV VLPs using E. coli and in vitro capsid formation. The use of these AAV VLPs for vaccine development and therapeutic delivery was also studied and strategies for DNA encapsidation were tested.

Among 13 known human and nonhuman primate AAV serotypes, AAV serotype 2 (AAV2) is the most studied and AAV serotype 5 (AAV5) is the most genetically divergent serotype. To analyze AAV capsid assembly, AAV2 and AAV5 VP3 capsid proteins were expressed in E. coli. AAV2 VP3 proteins formed inclusion bodies in E. coli, which were successfully purified and assembled into AAV2 VLPs in a chemically defined reaction. Contrary, AAV5 VP3 developed AAV5 VLPs inside E. coli cells. Recovery of the assembled AAV5 VLPs was straightforward via a one-step purification. Moreover, the supplement of AAV assembly-activating protein (AAP) into the AAV2 VLP assembly in vitro and the AAV5 VLP assembly in E. coli led to an increase of capsid formation yield. These findings provide the first evidence that AAV2 capsids can be developed in a defined reaction and AAV5 capsids can be formed inside E. coli cells.

The SARS-CoV-2 spike protein, in particular, its receptor-binding domain (RBD) containing the receptor-binding motif (RBM) mediates virus binding to the host cell receptor, and therefore RBD and RBM are lead candidates for SARS-CoV-2 subunit vaccine development. VLPs have been widely used in vaccine research due to their ability to strongly induce an immune response compared to the subunit protein alone. To assess the potential of AAV VLPs in vaccine research, RBD and RBM fused with AAV2 VP3 capsid proteins were expressed in E. coli. These fusion proteins were purified, then refolded and concomitantly assembled to VLPs. The refolded samples of AAV2 VP3_RBD and VP3_RBM containing VLPs were imaged by atomic force microscopy (AFM). AAV2 VP3_RBD VLPs were recognized by an antibody binding the folded RBD. Immunization of mice with VP3_RBM VLPs resulted in a high level of RBD-specific anti-bodies indicating SARS-CoV-2 specificity. However, VP3_RBD VLPs induced strong antibody responses against the VP3 VLP scaffold, but a relatively low level of RBD-specific antibodies. These data suggest that VP3_RBM VLP is a promising vaccine candidate for SARS-CoV-2.

Packaging of DNA in AAV is proposed as an only partially understood active procedure in which the DNA enters formed capsids with the help of proteins. In vitro, passive encapsidation would be easier and thus an arginine-rich motif (ARM), a DNA-binding domain known to passively aid DNA packaging of many viruses, was incorporated at the N-terminus of the AAV5 VP3 capsid protein. AAV ssDNA was produced using phagemids and M13 helper phage. The produced circular ssDNA with an expression cassette flanked by AAV inverted terminal repeat sequences (ITRs) mediated protein expression in CHO-K1 cells. The ARM_VP3 fusion protein was successfully obtained from E. coli and the in-vitro capsid assembly and DNA encapsidation resulted in VLPs and in protection of DNA from DNase I. However, these VLPs did not mediate protein expression in CHO-K1 cells. More studies are needed to productively encapsidate DNA in AAV capsids in vitro.

Fluorescent VLPs provide a tool for the analysis of VLP-cell interaction. In this study, FITC was chemically conjugated onto the formed particles via covalent bonds to visualize and image the AAV VLPs. The labeled VLPs internalized into human HeLa cells. This success opens opportunities to chemically couple a drug to AAV VLPs and label rAAV for different biological applications.

In summary, these studies show the first evidence that AAV capsids can be assembled in a chemically defined reaction or inside E. coli cells. The potentials of AAV VLPs for vaccine development, DNA encapsidation and imaging were also explored. These findings pave the way for using AAV VLPs in many different biological applications in the future.