Minicircle DNA purification using a CIM® DEAE‐1 monolithic support

Minicircle DNA is a new biotechnological product with beneficial therapeutic perspectives for gene therapy because it is constituted only by the eukaryotic transcription unit. These features improve minicircle DNA safety and increase its therapeutic effect. However, being a recently developed product, there is a need to establish efficient purification methodologies, enabling the recovery of the supercoiled minicircle DNA isoform. Thus, this work describes the minicircle DNA purification using an anion exchange monolithic support. The results show that with this column it is possible to achieve a good selectivity, which allows the isolation of the supercoiled minicircle DNA isoform from impurities. Overall, this study shows a promising approach to obtain the minicircle DNA sample with adequate quality for future therapeutic applications.

Anion exchange chromatography; Minicircle DNA; Monoliths; Parental plasmids; Recombinant DNA technology

Abbreviations

gDNA genomic DNA

mcDNA Minicircle DNA

mP mini plasmid

oc open circular

pDNA plasmid DNA

PP parental plasmid

sc supercoiled

The discovery of gene‐dependent diseases encouraged the development of DNA‐based therapies [1] . In recent years, there has been an increasing focus on the use of nonviral vectors, and minicircle DNA (mcDNA) is currently considered a new biotechnological product with promising therapeutic perspectives [2] , [3] . mcDNA is a circular DNA molecule exclusively composed of eukaryotic sequences [4] , [5] that was isolated for the first time by Cozzarelli and collaborators, from Escherichia coli (E. coli) T‐15 [6] . The absence of prokaryotic sequences in this vector, namely, the antibiotic resistance region and the origin of replication, improves its biocompatibility and efficacy in comparison with plasmid DNA (pDNA) [2] , [4] . In addition, these DNA molecules are smaller, which makes them more stable and increases the bioavailability [4] , [7] , thereby allowing improvement in the transfection efficiency and biological activity [8] , [9] . Moreover, due to the absence of CpG motifs, the gene will be expressed over several weeks [5] , reducing the death of transfected cells [5] . Another advantage lies in obtaining higher expression levels of the therapeutic protein [3] , [10] . The production of this biomolecule results from the intramolecular recombination of the parental plasmid (PP) in a bacterial culture [2] , [7] . Recombination is induced by L‐arabinose which provides phiC31 serine recombinase expression [4] , [11] , resulting in the formation of mcDNA (containing the therapeutic gene) and of mini plasmid (mP, containing the bacterial elements). The simultaneous expression of I‐SceI endonuclease promotes the digestion of PP and mP [3] , [7] . However, trace quantities of mP and PP resulting from the in vivo recombination process may still remain, suggesting that recombination is not complete [5] , [9] . Therefore, it is necessary to perform the purification to isolate the mcDNA with suitable purity to be applied in therapeutics [4] . Several efforts have been made in an attempt to purify the mcDNA, due to the complexity of the extract, an efficient purification strategy to recover pure mcDNA has not yet been described. Taking into account the diversity of species present in the E. coli lysate and the similarity between the mcDNA, PP, mP, and other host impurities, namely, RNA and genomic DNA (gDNA), only chromatographic methods will be suitable and acceptable to obtain this biomolecule, according to recommendations of regulatory agencies [1] .

In this work, a commercially available anion exchange monolithic column (CIM® DEAE‐1) was used to establish a new methodology for purification of the supercoiled (sc) isoform of mcDNA. Actually, it was considered useful to describe a novel application for this commercial column, attempting the purification of mcDNA. In general, monolithic supports enable fast and reproducible separations, providing higher binding capacities, and recovery of large biomolecules with higher yields and purities, which motivated their use for the purification of mcDNA [12] , [13] , [14] , [15] , [16] , [17] , [18] .

For mcDNA biosynthesis, the E. coli strain ZYCY10P3S2T and the vector pMC.CMV‐MCS‐EF1‐GFP‐SV40 PolyA of 7.06 Kbp were used (System Biosciences). Briefly, after PP amplification in Terrific Broth medium (12 g/L tryptone, 24 g/L yeast extract, 4 mL/L glycerol, 0.017 M KH2PO4, and 0.072 M K2HPO4) at 42°C, the synthesis of mcDNA was performed, through the addition of an inducing mixture and 0.01% of L‐arabinose at 32°C, as previously described by Gaspar and co‐workers [4] . For the preparation of the clarified E. coli lysate containing the mcDNA, it was used the modified alkaline lysis procedure described by Diogo and collaborators [19] , with the inclusion of slight modifications in the clarification step. The precipitation with ammonium sulfate was carried out to a final concentration of 2 M. After incubation on ice for 15 min, the precipitated proteins and RNA were removed by centrifugation at 16 000 × g for 20 min at 4°C, and the supernatant was used in the chromatographic experiments.

Chromatographic experiments were performed using an AKTA Purifier system with UNICORNTM version 5.11 software (GE Healthcare Biosciences, Sweden). The commercial CIM® DEAE‐1 monolithic column (BIA Separations, Ajdovscina, Slovenia) is described as an ideal chromatographic support for the purification of plasmid DNA of any size. This support is a weak anion exchanger with a bed volume of 1 mL. In addition, according to the information provided by the manufacturer, CIM® DEAE support provides high surface accessibility and very high binding capacities, fast large‐scale separations, flow unaffected resolution, pH stability (working range 2–13) and, finally, low backpressures even at very high flow rates. All buffers used in the chromatographic experiments were freshly prepared using deionized water ultra‐pure grade, filtered through a 0.2 μm pore size membrane (Schleicher Schuell, Germany) and degassed ultrasonically. Initially, several injection volumes (100, 200, and 500 μL) were tested, but for the higher volumes, it was verified that more contaminants were retained within the matrix together with mcDNA, which hindered the selectivity. Thus, in all experiments described, the sample was applied onto the column using a 100 μL loop at a flow rate of 1 mL/min. For mcDNA purification, the column was equilibrated with 0.71 M NaCl in 50 mM Tris‐HCl (pH 8.0). Nonretained species were washed out using loading buffer after which retained biomolecules were eluted by a 15 min linear gradient from 0.71 to 0.75 M of NaCl. A final elution step was performed, using 1 M NaCl, to elute strongly bound species from the column. All chromatographic experiments were accomplished at room temperature and the absorbance was continuously monitored at 260 nm. Fractions were pooled according to the chromatograms obtained and analyzed by horizontal electrophoresis using 0.8% agarose gel, to assess the purity of the mcDNA sample. Regeneration of monolithic column was executed by washing it with deionized water, followed by 20 column volumes of a buffer containing 2 M NaCl at a flow rate of 5 mL/min. This procedure enabled the efficient removal of contaminants from the CIM® DEAE‐1 column, which contributes to extremely repeatable and reproducible chromatographic separations.

The protein content in the fractions collected from the chromatographic experiments was evaluated using micro‐BCA protein assay kit from Pierce (Thermo Fisher Scientific), according to manufacturer's instructions. The calibration curve was prepared using BSA standards (20–2000 μg/mL). Genomic DNA was quantified through real time‐PCR, following changes in fluorescence of the DNA binding dye Maxima® SYBR Green/Fluorescein qPCR Master Mix (Thermo Fisher Scientific), as described by Sousa and co‐workers [14] . The calibration curve to achieve the gDNA concentration was constructed by serial dilutions of the E. coli gDNA in the range of 0.005 to 50 ng/μL. Endotoxins levels were evaluated by using the ToxinSensorTM Chromogenic Limulus Amoebocyte Lysate assay kit (GenScript, USA), in accordance with the manufacturer's instructions. The calibration curve was constructed by preparing several dilutions, in the range from 0.005 to 0.1 EU/mL, from a 10 EU/mL stock solution provided with the kit.

The zeta potential of the purified mcDNA and PP sample was determined using a Zetasizer Nano ZS particle analyzer (Malvern Instruments, UK), equipped with a He‐Ne laser, at 25°C. Zeta potential measurements were performed in disposable capillary cells and computed by using Henry's [F(Ka) 1.5] and Smoluchowsky models. All the data were examined in Zetasizer software v 7.03. The experiments were performed in triplicate and an average of 30 measurements was acquired individually for each sample.

The structural features of the mcDNA were visualized with a transmission electron microscope (TEM) (Hitachi HT7700, Japan) at 80 kV. Briefly, one drop of the solution containing the mcDNA samples was deposited for 2 min on the surface of copper grid.

CIM® DEAE‐1 column contains the polymeric amine ligand (diethylamine) positively charged, establishing electrostatic interactions with the phosphate groups present in negatively charged DNA [12] , [20] . Initially, the binding and elution behavior of mcDNA was studied, to confirm the establishment of interactions with the ligands. Then, new chromatographic experiments were performed to obtain the desired selectivity, consisting in the isolation of the sc mcDNA isoform, since it is considered the most biologically active conformation for gene transfer [1] .

The first chromatographic experiments were performed with the objective of finding the best conditions to promote the direct elution of RNA, given that functional groups of the column induce stronger interactions with DNA than with RNA. Thus, various chromatographic tests were outlined in an attempt to separate RNA from DNA, based on the premise that in general, a lower ionic strength can favor the interactions between positively charged ligands and nucleic acids [1] . The desired selectivity was achieved by applying a two steps‐based gradient of 0.68 and 1 M NaCl in 100 mM Tris‐HCl, pH 8.0, resulting in complete elution of RNA in the first peak, as shown in Fig. [NaN] A. In this figure, it is also visible that there is some loss of DNA in the first peak (Fig. [NaN] A, lane 1), but it is not significant in view of the elimination of RNA and the recovery of DNA in the second peak (Fig. [NaN] A, lane 2). The weaker interaction of RNA species with the positively charged ligands can be explained by the fact that after alkaline lysis the predominant RNAs in the extract are of low molecular weight, thus presenting lower density charge than DNA, which results in a lower ability to interact with the column [1] .

Accomplished the selectivity between RNA and DNA, new chromatographic tests were performed to optimize the gradient conditions which enabled the total separation between the DNA isoforms (open circular (oc) and sc isoforms), both from mcDNA and PP. This selectivity was achieved by applying a linear gradient from 0.68 to 0.74 M NaCl in 100 mM Tris‐HCl, pH 8.0, as shown in Fig. [NaN] B. The anion exchange monolithic column was first equilibrated with 0.68 M NaCl and after the injection of clarified lysate a first peak corresponding to the unbound species was obtained, which corresponded to RNA (Fig. [NaN] B, lane 1, peak 1). Then, a 15‐min linear gradient was performed, enabling the selectivity between the isoforms of DNA. The elution order of the species occurred in accordance to what was expected, considering the higher degree of compaction and higher density charge of the sc isoform. These characteristics promote a higher retention of the sc topology (Fig. [NaN] B lane 3, peak 3) in comparison to the oc isoform (Fig. [NaN] B lane 2, peak 2) [12] , [21] . Finally, a last elution step was performed (1 M NaCl) to retrieve strongly retained species (Fig. [NaN] B lane 4, peak 4).

Taking into account the previous results, the separation of the sc isoform of PP and the sc isoform of mcDNA was not satisfactory, co‐eluting in the conditions tested (Fig. [NaN] B lane 3, peak 3). Trying to obtain this separation some adjustments were performed in the concentration of the Tris‐HCl buffer. According to the results presented in Fig. [NaN] , the desired selectivity was obtained by performing a linear gradient from 0.71 to 0.75 M NaCl in 50 mM Tris‐HCl, pH 8.0, followed by a step of 1 M NaCl. The monolith was equilibrated with 0.71 M NaCl and after the injection of clarified lysate, a first peak corresponding to the unbound species was obtained, including some RNA and DNA (Fig. [NaN] lane 1, peak 1). Then, a 15‐min linear gradient up to 1 M NaCl was established and a peak was rapidly obtained in the flowthrough due to the elution of species with lower affinity to the column, namely most of RNA (Fig. [NaN] lane 2, peak 2). At the end of the linear gradient, the elution of sc mcDNA (Fig. [NaN] lane 3, peak 3) was achieved, obtaining the required selectivity, since the sc PP was eluted later (Fig. [NaN] lane 4, peak 4). Finally, the elution of highly bound species, mostly PP, was then achieved by increasing the ionic strength of the buffer up to 1 M NaCl (Fig. [NaN] lane 5, peak 5). In particular, although a small band of oc mcDNA was visualized in this chromatographic step (Fig. [NaN] , lane 3), it is possible to observe that sc mcDNA was isolated and suitably purified with high integrity and purity. In addition, sc mcDNA recovery is high in comparison with the slight loss of sc mcDNA isoform (Fig. [NaN] lane 4 and lane 5). To confirm the presence and identity of mcDNA, the purified mcDNA sample corresponding to the lane 3 of the Fig. [NaN] was linearized with the restriction enzyme BamHI. After digestion, only one band was visualized on agarose gel electrophoresis (results not shown). On the other hand, the agarose gels show bands above oc PP, what can be explained by the presence of gDNA fragments or by the presence of PP and/or mcDNA aggregates. Therefore, to discard the possibility of being gDNA, the sample recovered from the purification process was linearized with the restriction enzyme, BamHI. After digestion, as only one band was visualized on agarose gel electrophoresis, confirming that the initial bands visualized may correspond to PP and or mcDNA aggregates (results not shown). In anion exchange chromatography, the elution order depends on the charge, the size and conformation of biomolecules. The elution sequence found in this work follows this principle, since the mcDNA sc isoform elutes before the PP sc isoform, which is justified by the size and charge of these biomolecules. In fact, the mcDNA (3.06 kbp) is smaller than the PP (7.06 kbp), being also less negatively charged (–8.59 ± 0.07 mV) than PP (–32.63 ± 0.25 mV), resulting in a weaker interaction of mcDNA with the anion exchange monolithic column. In summary, the purification strategy used, was accomplished in about 30 min, reducing the contact time of sample with the monolithic support, avoiding the structural changes of mcDNA [12] , [14] . These findings can be explained by the presence of the diethylamino groups (–N+(C2H5)2) on support, which promote the retention of mcDNA, being isolated from an initial highly complex extract. Overall, it is suggested that although electrostatic interactions could play an important role on mcDNA retention, the contacts with nucleotide bases (due to the higher base exposure and availability for interactions) are also involved and modulate some favored interactions, responsible for the recognition found in CIM® DEAE‐1 column.

The sc mcDNA sample resulting from the purification with the CIM® DEAE‐1 column should fulfill the quality criteria established by regulatory agencies. To verify the purity of this sample, the levels of gDNA, proteins and endotoxins were determined. As shown in Table [NaN] , the anion exchange column used was effective in removing the impurities, allowing the recovery of a sc mcDNA sample that meets the requirements of regulatory agencies. The spectrophotometric quantification of the mcDNA revealed that this sample presented a concentration of about 152 μg/mL of sc mcDNA.

Quantification of genomic DNA, proteins, and endotoxins in the purified sc mcDNA sample

CIM® DEAE‐1 monolithic column is a weak anion exchanger that allows the purification of mcDNA (sc isoform) under non‐denaturing conditions using a mobile phase composed by NaCl in concentrations up to 1 M. However, to evaluate if the optimal experimental conditions, optimized for mcDNA purification, affect its structural stability and biological activity, circular dichroism (CD) experiments were performed, as described by Sousa and co‐workers [22] . CD spectrum of the purified sc mcDNA sample is similar to the typical B‐DNA CD spectrum, with characteristic bands around 220 and 275 nm (positive signals) and a 245 nm (negative signal). Thus, our results demonstrate that the purification procedure allowed the maintenance of the structural stability of the sc mcDNA isoform (data not shown) [22] . This study was also complemented with the observation of the purified sc and oc mcDNA samples through TEM. As depicted in Fig. [NaN] , the purified samples of sc and oc isoforms show no conformational change, thus confirming its structural stability. In fact, these results revealed that the elution conditions employed in the chromatographic experiments, besides having provided the desired selectivity, did not influence the structural stability of mcDNA, allowing the recovery of pure and stable sc mcDNA.

In this work, a purification strategy for mcDNA using anion exchange chromatography is described. The results obtained suggest that it is possible to establish a purification strategy to obtain the pure sc mcDNA isoform, eliminating the impurities resulting from the recombination process (PP and mP), as well as, impurities arising from the production in E. coli. Furthermore, by using this monolithic support it was promoted a rapid separation, which ensures structural stability of the sc mcDNA isoform. Thus, an important contribution is given regarding the purification of mcDNA.

This work was supported by FCT, the Portuguese Foundation for Science and Technology (UID/Multi/00709/2013). A.S. acknowledges the post‐doctoral fellowship (SFRH/BPD/102716/2014) from FCT, the Portuguese Foundation for Science and Technology. The authors also acknowledge to BIA Separations to kindly provide the monolithic support.

The authors have declared no conflict of interest.

1 Prazeres, D. M. F., Ferreira, G. N. M., Design of flowsheets for the recovery and purification of plasmids for gene therapy and DNA vaccination. Chem. Eng. Process. 2004, 43, 615 – 630.

• 2 Gaspar, V. M., Cruz, C., Queiroz, J. A., Pichon, C., Correia, I. J., Sousa, F., Sensitive detection of peptide‐minicircle DNA interactions by surface plasmon resonance. Anal. Chem. 2013, 85, 2304 – 2311.
• 3 Maniar, L. E. G., Maniar, J. M., Chen, Z. Y., Lu, J., Fire, A. Z., Kay, M. A., Minicircle DNA vectors achieve sustained expression reflected by active chromatin and transcriptional level. Mol. Ther. 2013, 21, 131 – 138.
• 4 Gaspar, V. M., Maia, C. J., Queiroz, J. A., Pichon, C., Correia, I. J., Sousa, F., Improved minicircle DNA biosynthesis for gene therapy applications. Hum. Gene Ther. Methods. 2014, 25, 93 – 105.
• 5 Kay, M. A., He, C. Y., Chen, Z. Y., A robust system for production of minicircle DNA vectors. Nat. Biotechnol. 2010, 28, 1287 – 1289.
• 6 Cozzarelli, N. R., Kelly, R. B., Kornberg, A., A minute circular DNA from Escherichia coli 15. Proc. Natl. Acad. Sci. USA. 1968, 60, 992 – 999.
• 7 Dong, Y., Aied, A., Li, J., Wang, Q., Hu, X., Wang, W., An in vitro approach for production of non‐scar minicircle DNA vectors. J. Biotechnol. 2013, 166, 84 – 87.
• 8 Chabot, S., Orio, J., Schmeer, M., Schleef, M., Golzio, M., Teissie, J., Minicircle DNA electrotransfer for efficient tissue‐targeted gene delivery. Gene Ther. 2013, 20, 62 – 68.
• 9 Kobelt, D., Schleef, M., Schmeer, M., Aumann, J., Schlag, P. M., Walther, W., Performance of high quality minicircle DNA for in vitro and in vivo gene transfer. Mol. Biotechnol. 2013, 53, 80 – 89.
• 10 Dietz, W. M., Skinner, N. E., Hamilton, S. E., Jund, M. D., Heitfeld, S. M., Litterman, A. J., Hwu, P., Chen, Z. Y., Salazar, A. M., Ohlfest, J. R., Blazar, B. R., Pennell, C. A., Osborn, M. J., Minicircle DNA is superior to plasmid DNA in eliciting antigen‐specific CD8+ T‐cell responses. Mol. Ther. 2013, 21, 1526 – 1535.
• 11 Sanei Ata‐Abadi, N., Dormiani, K., Khazaie, Y., Ghaedi, K., Forouzanfar, M., Lachinani, L., Rezaei, N., Kiani‐Esfahani, A., Nasr‐Esfahani, M. H., Construction of a new minicircle DNA carrying an enhanced green florescent protein reporter gene for efficient expression into mammalian cell lines. Mol. Biol. Rep. 2015, 42, 1175 – 1185.
• 12 Smrekar, F., Podgornik, A., Ciringer, M., Kontrec, S., Raspor, P., Strancar, A., Peterka, M., Preparation of pharmaceutical‐grade plasmid DNA using methacrylate monolithic columns. Vaccine. 2010, 28, 2039 – 2045.
• 13 Ongkudon, C. M., Danquah, M. K., Process optimization for anion exchange monolithic chromatography of 4.2 kbp plasmid vaccine (pcDNA3F). J. Chromatogr. B. 2009, 878, 2719 – 2725.
• 14 Sousa, Â., Tomaz, C. T., Sousa, F., Queiroz, J. A., Successful application of monolithic innovative technology using a carbonyldiimidazole disk to purify supercoiled plasmid DNA suitable for pharmaceutical applications. J. Chromatogr. A. 2011, 1218, 8333 – 8343.
• 15 Nunes, C., Sousa, Â., Nunes, J. C., Morão, A. M., Sousa, F., Queiroz, J. A., Supercoiled plasmid DNA purification by integrating membrane technology with a monolithic chromatography. J. Sep. Sci. 2014, 37, 1229 – 1236.
• 16 Sousa, Â., Sousa, F., Queiroz, J. A., Advances in chromatographic supports for pharmaceutical‐grade plasmid DNA purification. J. Sep. Sci. 2012, 35, 3046 – 3058.
• 17 Jungbauer, A., Hahn, R., Monoliths for fast bioseparation and bioconversation and their applications in biotechnology. J. Sep. Sci. 2004, 27, 767 – 778.
• 18 Tarmann, C., Jungbauer, A., Adsorption of plasmid DNA on anion exchange chromatography media. J. Sep. Sci. 2008, 31, 2605 – 2618.
• 19 Diogo, M. M., Queiroz, J. A., Monteiro, G. A., Martins, S. A., Ferreira, G. N., Prazeres, D. M., Purification of a cystic fibrosis plasmid vector for gene therapy using hydrophobic interaction chromatography. Biotechnol. Bioeng. 2000, 68, 576 – 583.
• 20 Stadler, J., Lemmens, R., Nyhammar, T., Plasmid DNA purification. J. Gene Med. 2004, 6, S54 – S66.
• 21 Bicho, D., Sousa, Â., Sousa, F., Queiroz, J. A., Tomaz, C. T., Effect of chromatographic conditions and plasmid DNA size on the dynamic binding capacity of a monolithic support. J. Sep. Sci. 2014, 00, 1 – 9.
• 22 Sousa, Â., Pereira, P., Sousa, F., Queiroz, J. A., Binding mechanisms for histamine and agmatine ligands in plasmid deoxyribonucleic acid purifications. J. Chromatogr. A. 2014, 1366, 110 – 119.

Graph: (A) Chromatographic profile resulting from the application of the clarified lysate onto the CIM® DEAE‐1 monolithic column. Elution was performed by using an increasing stepwise gradient of 0.65 and 1 M NaCl in 100 mM Tris‐HCl, pH 8.0. Agarose gel electrophoresis shows the species eluting in each chromatographic peak. Lane A: clarified lysate sample injected onto the column; Lanes 1 and 2: Fractions corresponding to peaks 1 and 2 of the chromatogram, respectively. (B) Chromatogram representing the application of the lysate sample in an anion exchange column. Elution was performed by increasing NaCl concentration in the eluent from 0.68 to 0.74 M NaCl in 100 mM Tris‐HCl, pH 8.0. Agarose gel electrophoresis analysis of each peak is represented in the respective figure. Lane A: sample injected onto the column; Lanes 1, 2, 3, and 4: Fractions corresponding to peaks 1, 2, 3, and 4 of the chromatogram, respectively.

Graph: Chromatographic profile resulting from the application of the clarified lysate onto the CIM® DEAE‐1 monolithic column. Elution was performed by increasing NaCl concentration in the eluent from 0.71 to 0.75 M NaCl in 50 mM de Tris‐HCl, pH 8.0. Agarose gel electrophoresis analysis of each peak is represented in the respective figure. Lane A: lysate sample injected onto the column; Lanes 1, 2, 3, 4, and 5: Fractions corresponding to peaks 1, 2, 3, 4, and 5 of the chromatogram, respectively.

Graph: Purified sample of oc and sc mcDNA, observed by TEM.