Abstract
Many lentiviral vectors used for gene therapy are derived from HIV-1. An optimal vector genome would include only the viral sequences required for transduction efficiency and gene expression to minimize the amount of foreign sequence inserted into a patient's genome. However, it remains unclear whether all of the HIV-1 sequence in vector genomes is essential. To determine which viral sequences are required, we performed a systematic deletion analysis, which showed that most of the gag region and over 50% of the env region could be deleted. Because the splicing profile for lentiviral vectors is poorly characterized, we used long-read sequencing to determine canonical and cryptic splice site usage. Deleting specific regions of env sequence reduced the number of splicing events per transcript and increased the proportion of unspliced genomes. Finally, combining a large deletion in gag with repositioning the Rev-response element downstream of the 3' R to prevent its reverse transcription showed that 1201 nucleotides of HIV-1 sequence can be removed from the integrated vector genome without substantially compromising transduction efficiency. Overall, this allows the creation of lentiviral vector genomes that contain minimal HIV-1 sequence, which could improve safety and transfer less viral sequence into a patient's DNA.