Reverse micellar extraction systems for the purification of pharmaceutical grade plasmid DNA

Abstract

Plasmid DNA as an active pharmaceutical ingredient (API) is gaining more and more importance. For the production of multigram quantities of this substance robust and scalable processes comprising several purification steps have to be designed. One main challenge is the initial separation of plasmid DNA and RNA in such a purification scheme. In this study we investigated the distribution of plasmid DNA and RNA in reverse micellar two-phase systems which is considered to be the basis for the development of an extractive purification step that can easily be integrated into common processes. For this purpose the distribution of the 4.6 kb plasmid pUT649 and Escherichia coli RNA in systems comprising isooctane, ethylhexanol, and the surfactant methyltrioctylammoniumchloride (TOMAC) under the influence of different salts was studied. Anion concentrations at which the partitioning behaviour for nucleic acids inverted (inversion point) were identified. Systems capable of separating RNA from plasmid DNA were further analysed and applied to extract RNA from plasmid DNA out of a preconditioned cleared lysate. The capability of reverse micellar systems for plasmid form separation was also shown by capillary and agarose gel electrophoresis.

Introduction

As the identification of functional genes evolves so do prospective novel medical treatments such as genetic vaccination or gene therapy. Diseases like cystic fibrosis, gaucher disease, ADA deficiency, etc. might be cured at the genetic level and some gene therapy products using viral and non-viral vector systems for gene delivery are already in clinical trial (Mountain, 2000). After tragic setbacks in gene medicine (Check, 2002, Raper et al., 2002) this prospective novel treatment got a new boost with the approval of two gene therapeutics in China in 2006 (Jia). More approved therapeutics, not only in human but also in veterinary medicine, will further increase the interest in gene therapy and thus lead to a growing demand for gene therapy products. For their production robust and scalable manufacturing processes have to be designed. Non-viral vector systems like plasmid DNA are considered to be safer in terms of oncogene activation and unintended immunological reactions. Considering the production process plasmid DNA is also superior to viral delivery systems because the manufacture, storage, and application possess fewer process and quality control problems. An even higher potential for plasmid DNA is expected in genetic vaccination. In this case the expression of a plasmid coded antigen leads to an immune response in the treated organism.

Clinical trials and future pharmaceuticals demand high quality plasmid DNA in multigram quantities. Plasmid DNA manufacturing processes comprise cultivation of Escherichia coli host cells (Voss et al., 2004), cell disruption by alkaline lysis (Birnboim and Doly, 1979), and product purification (Voß et al., 2005). For the latter, several multistep purification processes have been described. The challenge for biochemical engineering in this case is the initial separation of the desired supercoiled plasmid DNA form from structurally related impurities like RNA, host chromosomal DNA, and lipopolysaccharides (endotoxines). Since RNA is the largest contaminant, its removal is of particular importance, because it is difficult to separate from plasmid DNA and also significantly reduces the capacity of anion exchange chromatographic material that are used for subsequent purification. For non-pharmaceutical use, RNA can be removed from plasmids by digestion with RNase A. However, in case of plasmid DNA for pharmaceutical applications regulatory guidelines imply to forbear from using any animal derived substances throughout the whole production process (EMEA, 2001, FDA, 1998). Therefore, several other strategies have been applied for RNA removal, like hydrophobic interaction chromatography (Urthaler et al., 2005), gel filtration (Lemmens et al., 2003), precipitation (Eon-Duval et al., 2003a), ultrafiltration (Eon-Duval et al., 2003b), or the application of a recombinant RNase (Voss et al., 2006). Most of these methods are either tedious and time consuming, afflicted with high product loss, or difficult to scale up to the multigram purification scale. Extraction processes on the contrary are highly scalable and work with simple and inexpensive equipment and chemicals. For the purification of proteins two different kinds of extraction systems are described that allow the partitioning of the target molecule between two aqueous phases, i.e. aqueous two-phase systems and reverse micellar systems. In the first case, phases are formed by the addition of two polymers or a polymer and a salt whose solutions are not miscible. Polyethylene glycol (PEG)/potassium phosphate or PEG/dextran are well known systems for this purpose. For the extraction of plasmid DNA PEG/phosphate systems have also been investigated (Ribeiro et al., 2002, Frerix et al., 2005). While in most cases high concentrations of PEG and salts resulted in significant loss of product due to precipitation at the interface even at low DNA concentrations in the feed stock, well optimized systems are capable to separate RNA from plasmid DNA without product loss (Frerix et al., 2007). Other aqueous two-phase extraction processes use a thermoseparating polymer (Kepka et al., 2004). In this case the authors apply ultrafiltration and diafiltration for buffer exchange and extract the plasmid DNA in an aqueous two-phase system containing an ethyleneoxid propyleneoxid copolymer and dextran T-500. The capability of these systems to extract nucleic acids into the polyethylene glycol rich upper phase is known for over 20 years (Ohlsson et al., 1978). However, when using the ethyleneoxid propyleneoxid/dextran T-500 system the authors were unable to completely remove the contaminating RNA. An additional chromatographic step specially designed to remove RNA was necessary. This chromatographic step proved to be capable of completely removing the RNA, thus making extraction obsolete in this case (Gustavsson et al., 2004). Since the type of chromatographic resin called lid-beads applied in this case is also a porous particulate material it suffers from limitations these materials have when applied in the purification of plasmid DNA, i.e. low dynamic capacities and tedious and time consuming operation.

The principle of extraction with reverse micellar phases has originally been described for proteins (Hatton, 1989). Here, the electrostatic interaction between a polar head group of a surfactant and the countercharged protein results into transfer of the protein from the aqueous feed into an aqueous pool encapsulated in reversed micelles in the organic phase. The distribution can be influenced mainly by changing pH and ionic strength of the aqueous phase. This extraction method has also been applied to small fragments of genomic DNA (Goto et al., 1999). Reverse micellar two-phase systems are particularly interesting for plasmid purification because separation is based on electrostatic interaction and size differences of the target molecules. Salt concentrations in these systems are comparably low and the obtained aqueous phases can easily be subjected to further purification by e.g. anion exchange chromatography. Here we analysed the distribution of nucleic acids in reverse micellar two-phase systems under the influence of different salts with respect to separation efficiency, capacity, and the capability to discriminate between different plasmid forms. The results obtained from these model systems were successfully applied to remove RNA from a preconditioned bacterial lysate.