Vectorless supercoiled DNA: a new scalable abacterial in vitro system to improve gene therapy safety and production efficiency

PROJECT SUMMARY/ ABSTRACT
Bacterial plasmid manufacture is now a major bottleneck in viral gene therapy production workflows.
Although supercoiled plasmid-based vectors are the current industry standard for transient transfection
of packaging cells, minimized DNAs offer substantial safety and efficiency advantages. Currently, there
is a dearth of technologies to produce packaging and payload DNAs in a completely synthetic, abacterial
manner. In fact, no minimized DNA technology capable of scalable synthetic production of supercoiled
and completely scarless DNAs exists. Supercoiling is the preferred topology for circular DNAs given that
their performance in transient transfections is superior to other DNA forms. Minicircles or minivectors
generated by recombination in bacteria are supercoiled but require extensive and expensive purification
and yield a final product ‘scarred’ by a variable length prokaryotic sequence. Doggybone DNAs (dbDNAs)
comprised of linear double stranded DNA with circularized single-stranded ends are produced
synthetically but are not supercoiled. The long-term goal of this project is to develop an efficient bacteria-
independent workflow to produce, at a commercially viable scale, completely vectorless supercoiled
DNAs (VLSDNAs). This will be achieved using a new DNA assembly technology we have developed
termed cyclic heteroduplex thermostable ligation assembly (CHTLA). The objectives of the work
proposed here are to 1) optimize CHTLA reaction conditions to maximize supercoil production and
eliminate byproducts; and 2) to produce VLSDNAs encoding all packaging, replication, and payload
components to generate adeno-associated virus (AAV) particles loaded with a green fluorescent protein
(GFP) reporter gene. Rolling circle amplification (RCA) of a standardized green fluorescent protein (GFP)
gene expression cassette will be used to generate milligram quantities of overlapping and offset DNA
precursors for CHTLA reactions. To maximize supercoiling of CHTLA products, a two-enzyme, one-pot
system will be developed containing both thermostable DNA ligase and thermostable DNA gyrase activity.
The work proposed here is highly innovative because it represents a substantial departure from the status
quo by developing a robust new technology to produce, entirely in vitro, DNAs with a supercoiled topology
that are comprised exclusively of the sequence of interest. VLSDNA versions of pHelper-Kan, pAAV-Rep-
Cap, and pAAV ITR-GFP encoding adenovirus E1A, E1B, E2A, E4 and VA RNA open reading frames will
be generated using standard as well as optimized conditions identified in Aim 1. VLSDNAs will be
quantitatively compared to standard bacterially-sourced pHelper-Kan, pAAV-Rep-Cap, and pAAV ITR-
GFP triple plasmid transfection by commercial collaborator. Upon completion of these Aims we will have
determined optimal conditions to generate functional VLSDNAs at a scale that is commercially viable. In
Phase 2, we will seek to further scale production toward the gram+ quantities that will be required to
serve customers in the AAV and lentivirus (LV) gene therapy sectors.

Grant Number: 1R41GM154562-01
NIH Institute/Center: NIH
Principal Investigator: Haibo Bai