DNA Purification by Triple-Helix Affinity Precipitation Matteo D. Costioli, Igor Fisch, Fre ´ de ´ ric Garret-Flaudy, Frank Hilbrig, Ruth Freitag* Center of Biotechnology, Swiss Federal Institute of Technology Lausanne, 1015 Ecublens, Switzerland; telephone: +41 21 693 6108; fax: +41 21 693 6030; e-mail: ruth.freitag@epfl.ch Received 27 September 2001; accepted 16 July 2002 DOI: 10.1002/bit.10497 Abstract: Recent advances in DNA-based medicine (gene therapy, genetic vaccination) have intensified the neces- sity for pharmaceutical-grade plasmid DNA purification at comparatively large scales. In this contribution triple- helix affinity precipitation is introduced for this purpose. A short, single-stranded oligonucleotide sequence (namely (CTT) 7 ), which is capable of recognizing a complementary sequence in the double-stranded target (plasmid) DNA, is linked to a thermoresponsive N- isopropylacrylamide oligomer to form a so-called affinity macroligand (AML). At 4°C, i.e., below its critical solution temperature, the AML binds specifically to the target molecule in solution; by raising the temperature to 40°C, i.e., beyond the critical solution temperature of the AML, the complex can be precipitated quantitatively. After re- dissolution of the complex at lower temperature, the tar- get DNA can be released by a pH shift to slightly alkaline conditions (pH 9.0). Yields of highly pure (plasmid) DNA were routinely between 70% and 90%. Non-specific co- precipitation of either the target molecule by the non- activated AML precursor or of contaminants by the AML were below 7% and presumably due to physical entrap- ment of these molecules in the wet precipitate. Ligand efficiencies were at least 1 order of magnitude higher than in triple-helix affinity chromatography. © 2003 Wiley Periodicals, Inc. Biotechnol Bioeng 81: 535–545, 2003. Keywords: downstream processing; lower critical solu- tion temperature; plasmid DNA; PNIPAM; precipitation; thermoresponsive polymer; triple-helix affinity INTRODUCTION DNA-based medicine, i.e., gene therapy and genetic vacci- nation using nucleic acids, is currently one of the most innovative and exciting areas of clinical research. The treat- ment of a disease on the genetic rather than the phenotypic level is a promising option to obtain enhanced therapeutics (Schleef, 2001). In non-viral gene therapy, the gene transfer system is usually naked or complexed plasmid DNA. While such plasmid vectors are less efficient in terms of transfec- tion success than viral systems, they have the advantage of a very low immunogenicity and an excellent safety profile. Their clinical benefit has been clearly demonstrated and great progress has been made in improving these vectors (Mountain, 2000). Recently a method for orally adminis- tered gene therapy has been disclosed, and dosings between 0.1 mg and 1 g plasmid DNA have been mentioned in this context (Robertson, 2001). If such doses are indeed rou- tinely required in gene therapy, pharmaceutical-grade plas- mid DNA will have to be produced and purified at a scale that is several orders of magnitude larger than anything existing today. While plasmids can conceivably be manufactured in bac- teria at large scale, the downstream process, i.e., the isola- tion and purification of the plasmid DNA to the required specifications, may well prove to be a major bottleneck. Attempts to scale up traditional (preferably chromato- graphic) methods for the plasmid purification (mainly an- ionic and size-exclusion chromatography) have shown some inherent limitations (Ferreira et al., 2000a; Freitag and Vogt, 1999; Levy et al., 2000; Prazeres et al., 1999). When anion-exchange chromatography is used, binding is in com- petition to polyanionic impurities such as RNA, genomic DNA, proteins, and endotoxins. The separation of the plas- mid DNA from these compounds is almost inevitably beset with some difficulties, especially if the impurities are simi- lar in size, which is the case, e.g., for genomic DNA frag- ments, RNA, and the endotoxins. RNA removal by exog- enous/endogenous RNases, protein removal, e.g., by a “salt- ing out” step, as well as a concentration step and a buffer exchange are therefore often necessary before a chromato- graphic purification of the plasmid DNA can be attempted at process scale. In addition, anion-exchange chromatogra- phy produces a complex elution profile and significant losses when working at high flow rates. In the case of size-exclusion chromatography the separa- tion of genomic DNA and plasmid is only possible at very low relative genomic DNA concentrations. Process-scale size-exclusion chromatography is also hampered by the re- striction to low flow rates and loading factors. Hydroxyapa- tite has been suggested as an alternative stationary phase for *Correspondence to: Ruth Freitag Contract grant sponsors: Swiss National Science Foundation, Commis- sion for Technology and Innovation Contract grant numbers: 20-63530.00 (to R.F.), 4731.2 SUS (to R.F.) © 2003 Wiley Periodicals, Inc.