12. Assembly-Activating Protein Is Not an Essential Requirement for Capsid Assembly of Adeno-Associated Virus Serotypes 4, 5, and 11

Abstract

The mechanisms of capsid assembly for adeno-associated virus (AAV) is still incompletely understood, but delineating the role of the assembly-activating protein (AAP) in this process may lead to insight on this important step in vector production. AAP is a non-structural protein encoded within the cap gene in an overlapping open reading frame (ORF) that initiates near the N-terminus of the VP2 ORF and continues into the VP3 ORF. This protein has been shown to be essential for assembly of AAV2, 8, and 9 capsids; however, interestingly, we have recently found that AAV5 is capable of forming VP3-only capsids without AAP, raising a new question of what role AAP plays in the AAV5 virus life cycle. To further understand the functional role of AAP of AAV5 and other serotypes, it is important to investigate whether AAV5 would require AAP to form infectious capsids containing all three capsid proteins (VP1, VP2, and VP3) and whether other AAV serotypes exist that could form AAP-independent capsids. Here we show that AAV5 particles assembled in the absence of AAP are infectious and AAP-independent capsid assembly takes place for AAV4 and 11 as well but not for AAV12, a serotype that is phylogenetically closely related to AAV4 and 11. In the study, we created an AAV5 helper plasmid expressing AAV2 Rep and AAV5 Cap in which the AAP5 ORF was codon-modified extensively to prevent AAP expression while leaving the VP ORF intact. Using this AAP-KO helper plasmid and the standard AAV5 helper plasmid with an intact AAP ORF, we produced two types of AAV5 vectors expressing a marker gene, AAV5(AAP-) and AAV5(AAP+) vectors, respectively. These vectors were applied on CHO cells, and the marker gene expression was quantified 2 days post-infection. This assay revealed that the AAP-independent AAV5(AAP-) vector was able to efficiently transduce CHO cells although the transduction efficiency was about a half of that with the AAV5(AAP+) vector. The yield of AAV5(AAP-) vector production was lower than AAV5(AAP+) vector by ~10 fold but restored by co-expression of AAP in trans, indicating that AAP retains the assembly promoting role even though AAP is not an absolute requirement for AAV5 capsid assembly. We next sought to determine if AAV4 and 11 could also assemble capsids without AAP. To address this question, we transfected HEK293 cells with codon-modified AAV4 or AAV11 VP3-expressing plasmids and purified the potential viral particles by cesium-chloride ultracentrifugation. Electron microscopy revealed viral particles consistent with AAV capsids. Given the close evolutionary relationship between AAV4, 11 and 12, we hypothesized that AAV12 would also be capable of AAP-independent capsid formation, but preliminary evidence has shown this hypothesis to be incorrect. We anticipate that closer examination of the dependence or independence of viral capsid assembly on AAP among various serotypes will reveal differences in VP structure or the assembly process that will illuminate the mechanistic function of AAP.