Generation of chromosomal DNA during alkaline lysis and removal by reverse micellar extraction
- 185 Downloads
The separation of structurally related impurities from pharmaceutical plasmid DNA by highly scalable purification techniques is a challenge for biochemical engineering. Next to RNA, proteins, and lipopolysaccharides, the chromosomal DNA of the plasmid replicating host has to be removed. Here, we describe the application of reverse micellar extraction for the separation of chromosomal from plasmid DNA. By applying different procedures for alkaline lysis, bacterial lysates with different amounts of chromosomal DNA were generated. A reverse micellar extraction step enabled us to deplete the concentration of this impurity below the required level of 50 mg g−1 of plasmid DNA with almost complete plasmid recovery.
KeywordsAlkaline lysis Chromosomal DNA Extraction Plasmid production Reverse micelles
Plasmid DNA has gained considerable attention as a vector system for gene delivery in gene therapy and genetic vaccination. Production processes at the gram to kilogram scale are required to meet the demands for clinical trials and future therapeutics. Next to RNA, proteins, and lipopolysaccharides (LPS; endotoxins), the chromosomal DNA of the Escherichia coli host cells has to be separated from the product. The concentration of chromosomal DNA (chrDNA) is determined by the lysis procedure (Levy et al. 2000). Alkaline lysis (Birnboim and Doly 1979) is applied for this purpose in most cases although other procedures like thermal lysis (O'Mahony et al. 2005, 2007) and beed milling (Carlson et al. 1995) have been utilized. Several papers have described the influence of alkaline lysis conditions on the stability of plasmid DNA (intact supercoiled form) and on the contamination with chrDNA (Chamsart and Karnjanasorn 2007; Chamsart et al. 2001; Meacle et al. 2007; Zhang et al. 2007). The high molecular weight chrDNA is considered to be very susceptible to shear during the whole purification process and the lysis step in particular resulting in soluble fragments having a broad size distribution. These fragments are difficult to remove by common downstream processing protocols. Therefore, currently applied purification strategies try to minimize shear and in this way the generation of chrDNA fragments during lysis and to deplete their concentration during subsequent process steps.
Recently, extraction procedures have been described for pharmaceutical grade plasmid DNA production. Especially reverse micellar extraction has proven to efficiently separate RNA from the product. The principle of extraction with reverse micellar phases has originally been described for proteins (Hatton 1989). The electrostatic interaction between the polar head group of a surfactant and the countercharged protein resulted in a transfer of the protein from the aqueous feed into an aqueous pool encapsulated in reversed micelles solubilized in an apolar phase. Separation can take place during forward as well as back extraction into a new aqueous phase. The distribution can be influenced by ionic strength of the aqueous phase, type of salt, and in the case of protein extraction, by pH. Variables of the reverse micellar phase influencing the extraction process are concentration and type of surfactant, organic solvent, and the type of cosurfactant (Krishna et al. 2002). Reverse micellar extraction has already been applied to small fragments of genomic DNA (Goto et al. 1999) and their capability to separate plasmid DNA from RNA has also been demonstrated (Streitner et al. 2007, 2008) thus making them an interesting alternative to common purification strategies. In addition, extraction processes are highly scalable and work with simple inexpensive equipment and chemicals.
In here, we describe how reverse micellar extraction is capable of separating chrDNA from plasmid DNA independent on the initial chrDNA concentration or alkaline lysis conditions respectively. By applying this extraction procedure as an initial separation step in downstream processing, we are able to deplete chrDNA concentration below the desired level.
Materials and methods
Chemicals and media
All chemicals were obtained from SIGMA (Deisenhofen, Germany) or MERCK (Darmstadt, Germany) and were of analytical grade if not stated otherwise. For inoculation and bioreactor cultures, a semidefined glycerol (HSG) medium (Voß et al. 2003) was used consisting of glycerol (15 g L−1; Cognis Oleochemicals, Düsseldorf, Germany), yeast extract (7 g L−1; Deutsche Hefewerke, Nürnberg, Germany), vegetable peptone (13.5 g L−1; UD Chemie, Wörrstadt, Germany), KH2PO4 (1.5 g L−1), K2HPO4 (2.3 g L−1), NaCl (2.5 g L−1), and MgSO4 × 7 H2O (0.25 g L−1).
Bacterial strains, plasmids, and growth conditions
E. coli strain DH5α (DSMZ No. 6897) was used for the cultivations without a plasmid as well as for plasmid DNA production. Transformation of the plasmid pUT649 (4.6 kb, Eurogentec, Liege, Belgium) was carried out using standard laboratory techniques (Sambrook et al. 1989). Shake flask cultures were carried out on the semidefined glycerol medium at 37°C for 6 to 8 h to obtain inoculation cultures for bioreactor cultivation.
Cultivation of the E. coli strains was carried out in a 30-L (20 L working volume) stirred tank reactor (MBR, Wetzikon, Switzerland) using the semidefined glycerol medium. The stirrer frequency was automatically controlled by means of the dissolved oxygen concentration. The pH of the culture was controlled at 7.0 using phosphoric acid (100 g L−1) and 2 M sodium hydroxide solution. Biomass was harvested after the culture entered the stationary growth phase (between 16 and 18 h cultivation time) by centrifugation for 10 min at 9,600 min−1 and 4°C in a SIGMA K70 centrifuge.
The wet cell paste was resuspended in cell resuspension buffer (50 mM Tris-HCl, 10 mM ethylenediaminetetraacetic acid, pH 8.0) at a concentration of 100 g L−1. Cells were lysed by the addition of the same volume of lysis buffer (0.2 M NaOH, 1% (w/w) sodium dodecyl sulfate (SDS)) for 3 min followed by neutralization with an equal volume of potassium acetate buffer (3.0 M potassium acetate, acetic acid, pH 5.5) for another 5 min. Precipitated material was removed by centrifugation at 9,600 min−1 for 10 min and filtration to completely remove suspended particles. The alkaline lysis has been accomplished under varying shear. E. coli cells were disrupted by continuous lysis, gentle overhead shaking in a 1 L bottle, stirring in a beaker at a stirrer speed of 200 and 600 min−1 (volume 3 L) using an Intermig stirrer (EKATO Rühr- und Mischtechnik GmbH, Schopfheim, Germany), and stirring in a bottle using a magnetic stirrer at 400 min−1 (volume 1 L). For the lysis procedure with Intermig stirrers, two stirrers were applied in a 5-L beaker with an inner diameter of 17 cm. The diameter of the stirrers was 11 cm, resulting in a ratio of 0.65. The lower impeller was located 3.3 cm above the bottom while the second impeller was located 8.5 cm above the lower impeller. For the continuous lysis, cell suspension (200 g cell paste in 2 L lysis buffer) and 2 L alkaline lysis buffer were mixed at equal flow rates in a simple T-connector operated as a T-mixer (Voß et al. 2005). Lysis took place in the volume of a subsequent tube (length 224 cm, diameter 0.5 cm, residence time 27.3 s). Neutralization and separation from the precipitate was achieved using simple froth flotation by sparging air through a sintered metal plate at the bottom of a column with a diameter of 10 cm, whereas the sintered metal plate had a diameter of 9.2 cm. The height of the column was 100 cm.
Quantification of chromosomal DNA
Quantification of chromosomal DNA was achieved using a QuantiTect® SYBR® Green PCR-Kit (Qiagen, Hilden, Germany). The standards for calibration were prepared using a Genomic DNA Purification Kit (Promega, Madison, WI, USA) to isolate chromosomal DNA from E. coli DH5α. Standards were prepared by establishing DNA concentrations of 150, 50, 10, 5, and 1 ng μL−1 by measuring the absorption at 260 nm prior to diluting these solutions further by a factor of 1:1,000. For polymerase chain reaction (PCR), 10 μL Mastermix (HotStarTaq®DNA-Polymerase, QuantiTect SYBR Green PCR buffer, dNTP-Mix, SYBR Green I, Rox-Passive Reference, MgCl2 (5 mM)), 1 μL forward primer pyka fwd (GCGTCAGCTAAACCGAGCGGTAATCAC; 20 mg L−1), 1 μL reverse primer pyka rev (GGCGTTTGCTACGTCCATGACTTCTGC; 20 mg L−1), 3 μL RNase-free water, and 5 µL sample or standard were used. For samples as well as standards, threefold measurements have been accomplished. PCR was carried out in a Rotorgene-Thermocycler (Corbett Life Science, Sydney, Australia) and analyzed with a Rotorgene Software 6.0.19 (Corbett Life Science, Sydney, Australia).
Reverse micellar extraction
The bacterial lysates were diluted by a factor of 2 with 10 mM Tris-HCl (pH 7.0) and extracted with an equal volume of a reverse micellar solution containing 40 mM methyltrioctylammonium chloride and 0.05% (v/v) 2-propanol in isooctane for 5 min at room temperature on an overhead shaker at a frequency of 100 min−1. Phase separation was achieved by centrifugation at 5,000 min−1 for 5 min. In the following back extraction, the reverse micellar phase was mixed with a new aqueous phase consisting of 1 M NaCl and 10% (v/v) 2-propanol in 10 mM Tris-HCl (pH 7.0). Extraction and phase separation was achieved as described above.
Because of its similarity in chemical composition and structure, chrDNA is difficult to separate from plasmids in purification processes for pharmaceutical grade plasmid DNA. It has already been shown that high molecular weight chromosomal DNA is very susceptible to shear leading to fragmentation into smaller parts that contaminate the subsequent product stream during plasmid purification (Levy et al. 2000). Therefore, a gentle lysis procedure was recommended and the occurring contamination with chrDNA was depleted by subsequent purification steps. Recently, we demonstrated the capability of reverse micellar extraction for plasmid purification (Streitner et al. 2007, 2008). In order to verify if these systems are capable to deplete chrDNA from plasmid preparations, we generated bacterial lysates with different alkaline lysis procedures and subjected them to reverse micellar extraction.
According to Levy et al. (2000), the alkaline lysis procedure is the process step in which chrDNA is most susceptible to degradation by shear. First of all, the liberation of high molecular weight compounds after treating the resuspended E. coli with NaOH and SDS results in the formation of a viscous liquid with non-Newtonian properties. Under these conditions, chrDNA is degraded by radial shear as well as pressure drops occurring during liquid processing. Subsequent neutralization generates a precipitate consisting of protein dodecyl sulfate complexes, cell debris, and chromosomal DNA which shows viscoelastic properties and is difficult to separate from the lysate fluid by filtration procedures. Therefore, initial centrifugation and subsequent clarification by filtration are commonly applied for generating a cleared lysate. It should be considered in this context that a high speed centrifugation step inflicts shear force on the precipitated chrDNA and poses a potential contamination risk.
In our study, bacterial cleared lysates were prepared by applying an alkaline lysis in different ways. As anticipated from the results published by Levy et al. (2000), gentle lysis procedures like overhead shaking and continuous lysis with subsequent froth flotation of the debris gave low concentrations of chrDNA in the lysate while an increased agitation with an Intermig stirrer resulted in a significantly higher concentration of chrDNA. The results obtained from overhead shaking (397 mg g−1 plasmid) were in the same order of magnitude as the results observed by Lemmens et al. (2003) with 150 mg g−1 plasmid. However, the concentration was relatively high in comparison to the concentrations of less than 20 mg g−1 reported by Chamsart et al. (2001) using a Rushton turbine for mixing. The reason for this discrepancy is probably located in the different analytical methods applied for the quantitation of chrDNA. Nevertheless, continuous lysis combined with froth flotation constituted the gentlest procedure in our study, generating the lowest chrDNA concentration. In addition, the easy removal of the solid debris by flotation resulted in a perfectly cleared lysate. As demonstrated before (Voß 2007), optical density at 600 nm showed a value of 0.002 after flotation while a lysate prepared by overhead shaking and subsequent centrifugation and filtration only gave an OD600 value of 0.03. The flotation step also constitutes a procedural advantage because centrifugation and filtration and therefore an additional risk of contamination with chrDNA are avoided. The low recovery of lysate liquid after flotation could be circumvented by rinsing the flotate with an equal volume of buffer. As expected, the concentration of chromosomal DNA in the volume obtained after washing did not exceed the concentration found in the drained lysate. By rinsing the flotated material, plasmid DNA recovery could be increased significantly without additional contamination of chromosomal DNA.
The relative amount of chrDNA to plasmid DNA was ranging from 295 to 1,125 mg g−1 plasmid in the different alkaline lysates and exceeding the limit for pharmaceutical plasmid preparations of 50 mg g−1 (Voß 2008). Since reverse micellar extraction was found to be capable to separate RNA from plasmid DNA, it was considered reasonable to analyze the capability of this purification step for the removal of chrDNA. The principle of reverse micellar extraction for the separation of proteins was originally introduced by Hatton (1989) and is based on the electrostatic interaction of a biopolymer (protein or nucleic acid) with a countercharged surfactant forming reverse micelles in an apolar solvent. The separation efficiency can be influenced through modulating the electrostatic interaction by modifying the ionic strength in the extraction system. In addition, this change also influences the repulsion of the polar surfactant headgroups and in this way modifies the size of the reverse micelles in the apolar solvent. Therefore, a size exclusion process is considered to be the second separation principle in reverse micellar extraction. When trying to separate chrDNA from plasmid DNA, it should be taken into account that chrDNA cannot only occur in different concentration but also in varying size distribution which could also depend on the mode of mixing during alkaline lysis. Since reverse micellar extraction is considered to separate by different electrostatic interaction and molecular weight, the aforementioned fact should also be kept in mind.
The extraction system optimized for RNA removal was applied for our purpose without additional modifications using the different alkaline lysates as aqueous feed. Interestingly, the most efficient depletion was observed with the lysate generated by stirring at 600 rpm and the one generated by flotation. Higher relative amounts of chrDNA were observed after extraction with the other lysates. Possible reasons for the disparity in extraction efficiency were probably differences in the structure of the chromosomal DNA as a result of the lysis procedures. High shear forces during alkaline lysis degraded the precipitated chromosomal DNA and smaller fragments were solubilized again in the lysate. It could be assumed that these fragments were not extracted due to their size and structure.
Large fragments of chromosomal DNA were presumably sterically not suited for a reverse micellar extraction due to their high molecular mass. To circumstantiate this, the structure and size distribution of the chromosomal DNA have to be further investigated. Nevertheless, the depletion of chromosomal DNA was in all cases successful. Next to continuous lysis in combination with flotation, an alkaline lysis using high shear forces seemed to be an interesting alternative to common lysis procedures for large scale plasmid DNA production. In case of reverse micellar extraction, a gentle lysis procedure and cautious removal of the precipitate were not required thus facilitating the process management on a large scale. Alternative, scalable, continuous processes for cell lysis could be applied, since the chromosomal DNA could be efficiently removed by reverse micellar extraction while plasmid DNA was completely recovered.
The research project was funded by the German Research Foundation (Deutsche Forschungsgemeinschaft) and the Max-Buchner-Forschungsstiftung. Their support is gratefully acknowledged.
- Hatton TA (1989) Reversed micellar extraction of proteins. In: Scamehorn JF, Harwell JH (eds) Surfactant based separation processes. Marcel Dekker, New YorkGoogle Scholar
- Krishna SH, Srinivas ND, Raghavarao KS et al (2002) Reverse micellar extraction for downstream processing of proteins/enzymes. Adv Biochem Eng Biotechnol 75:119–183Google Scholar
- O'Mahony K, Freitag R, Hilbrig F et al (2007) Integration of bacteria capture via filtration and in situ lysis for recovery of plasmid DNA under industry-compatible conditions. Biotechnol Prog 23:895–903Google Scholar
- Sambrook J, Fritsch EF, Maniatis T (1989) Molecular cloning. A laboratory manual. Cold Spring Harbor Laboratory, Cold Spring HarborGoogle Scholar
- Zhang H, Kong S, Booth A et al (2007) Prediction of shear damage of plasmid DNA in pump and centrifuge operations using an ultra scale-down device. Biotechnol Prog 23:858–865Google Scholar