- Email: [email protected]
A multimodal histamine ligand for chromatographic purification of plasmid DNA
Journal of Chromatography A, 1281 (2013) 87–93
Contents lists available at SciVerse ScienceDirect
Journal of Chromatography A journal homepage: www.elsevier.com/locate/chroma
A multimodal histamine ligand for chromatographic purification of plasmid DNA夽 a a,b ˇ ˇ Urh Cernigoj , Urˇska Vidic a , Miloˇs Barut a,b,∗ , Aleˇs Podgornik a,b , Matjaˇz Peterka a,b , Aleˇs Strancar a b
BIA Separations d.o.o., Mirce 21, SI-5270 Ajdovˇscˇ ina, Slovenia The Centre of Excellence for Biosensors, Instrumentation and Process Control – COBIK, Velika pot 22, SI-5250 Solkan, Slovenia
a r t i c l e
i n f o
Article history: Received 3 October 2012 Received in revised form 11 January 2013 Accepted 11 January 2013 Available online 23 January 2013 Keywords: Plasmid DNA Chromatography Immobilized histamine Multimodal interactions Monoliths
a b s t r a c t To exploit different chromatographic modes for efficient plasmid DNA (pDNA) purification a novel monolithic chromatographic support bearing multimodal histamine (HISA) groups was developed and characterized. Electrostatic charge of HISA groups depends on the pH of the mobile phase, being neutral above pH 7 and becoming positively charged below. As a consequence, HISA groups exhibit predominantly ion-exchange character at low pH values, which decreases with titration of the HISA groups resulting in increased hydrophobicity. This feature enabled separation of supercoiled (sc) pDNA from other plasmid isoforms (and other process related impurities) by adjusting salt or pH gradient. The dynamic binding capacity (DBC) for a 5.1 kbp large plasmid at pH 5 was 4.0 mg/ml under low salt binding conditions, remaining relatively high (3.0 mg/ml) even in the presence of 1.0 M NaCl due to the multimodal nature of HISA ligand. Only slightly lower DBC (2.7 mg/ml) was determined under preferentially hydrophobic conditions in 3.0 M (NH4 )2 SO4 , pH 7.4. Open circular and sc pDNA isoforms were baseline separated in descending (NH4 )2 SO4 gradient. Furthermore, an efficient plasmid DNA separation was possible both on analytical as well as on preparative scale by applying the descending pH gradient at a constant concentration (above 3.0 M) of (NH4 )2 SO4 . © 2013 Elsevier B.V. All rights reserved.
1. Introduction The purification strategy of supercoiled plasmid DNA (sc pDNA), a biological macromolecule with increasing pharmaceutical potential [1–3], involves challenges and constraints that are rarely encountered with other classes of biomolecules [4]. The main reasons are: (1) a low content of pDNA in the cells, (2) the mechanical lability of sc pDNA and (3) impurities sharing common characteristics to pDNA (negative charge – RNA, open circular pDNA (oc pDNA), genomic DNA (gDNA) and endotoxins; high molecular mass – oc pDNA, gDNA and endotoxins; hydrophobicity – endotoxins) [5]. Purification process commonly starts with an alkaline lysis that removes part of the cell walls, organelles, proteins and gDNA, but leaves RNA as the main contaminant together with some proteins, endotoxins and gDNA [6]. Further purification steps can be non-chromatographic e.g., aqueous two-phase extraction [7], tangential flow filtration [8], selective precipitation of pDNA [6], but despite some advantages they are usually followed by a
夽 Presented at the MSS 2012, Ajdovˇscˇ ina, Portoroˇz, Slovenia, 1–6 June 2012. ∗ Corresponding author at: BIA Separations d.o.o., Mirce 21, SI-5270 Ajdovˇscˇ ina, Slovenia. Tel.: +386 59 699 500; fax: +386 59 699 699. E-mail address: [email protected] (M. Barut). 0021-9673/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.chroma.2013.01.058
polishing chromatographic step to obtain the product matching the FDA recommendations [9]. Due to these stringent regulatory requests and due to process economics, the majority of downstream processes are based on chromatography. There are different chromatographic separation mechanisms implemented for pDNA purification, such as size-exclusion [10], anion-exchange [11], hydrophobic interaction chromatography (HIC) [12], IMAC [13] or affinity chromatography [14]. However, to completely separate the pDNA from all other impurities on preparative scale, they must be combined in a suitable, robust and efficient purification process. Even the use of highly specific affinity ligands in some cases lacks sufficient selectivity [15] and especially suffers from slow binding kinetics and low recovery of pDNA [16]. Besides, the ligand costs are prohibitively high. A novel approach of immobilizing only 16mer peptide representing helix II of DNA binding domain of lac repressor was recently used in pDNA purification [17], which seems to overcome the above mentioned limitations but on the expense of high ligand costs. As an alternative, pseudo affinity ligands for pDNA purification are often studied as cheaper substitutes for affinity ligands. Thiophilic chromatography [18] was successfully implemented into a chromatographic process of sc pDNA purification, but it still needed two additional chromatographic steps to complete the plasmid isolation. Certain immobilized antibiotics and anticancer agents
88
ˇ U. Cernigoj et al. / J. Chromatogr. A 1281 (2013) 87–93
binding DNA with high specificity were tested as pseudo affinity ligand as well [19]. An example of such a ligand is immobilized berenil [20], where the purification of pDNA directly from the lysate is achieved in descending (NH4 )2 SO4 gradient. Its benefits were proven for different plasmid isoforms separation, but no results were shown for the process efficiency on preparative scale. Even certain amino acids, such as lysine, arginine and histidine interact with DNA through multimodal mechanism [21,22] and they were successfully applied for the separation of pDNA isoforms on an analytical scale. To achieve high binding capacity, target functionalities have to be introduced onto proper chromatographic support. Monoliths exhibit several advantages over beaded supports when implemented for sc pDNA purification due to large interconnected convective pores and high surface accessibility [23–25]. The methacrylate based monolithic chromatographic supports have already been successfully implemented for pDNA isoforms separation using a pseudo affinity ligand (imidazole) on a laboratory scale [26]. The main purpose of the present work was to further investigate the performance of amino acid based multimodal ligands. Histamine (HISA), a derivative of amino acid histidine, containing an imidazole ring on a spacer arm was used as a ligand of choice due to its unique chemical properties. Its applicability for the separation of pDNA isoforms under different chromatographic conditions on analytical as well as on preparative scale was studied in detail.
2. Materials and methods 2.1. Materials All solutions were freshly prepared using purified water which meets the requirements for European Pharmacopoeia (AQUATEHNA Biro, Zgornja Kungota, Slovenia) and analytical grade reagents. Buffer solutions were filtered through a 0.22 m PES filter (TPP, Trasadingen, Switzerland). The samples injected in the column were filtered through cellulose mixed esters syringe filters, pore size 0.45 m (Macherey-Nagel, Düren, Germany). Agarose gel was from SeaKem LE Agarose (Lonza Group, Basel, Switzerland). NaOH, TRIS, potassium acetate, CaCl2 , EDTA, (NH4 )2 SO4 , and boric acid were from Merck (Darmstadt, Germany). SDS, Na2 HPO4 ·2H2 O, NaH2 PO4 ·2H2 O, acetic acid, NaOH and NaCl were from Sigma–Aldrich (St. Louis, MO, USA). 1 ,1 Carbonyldiimidazol (CDI) was purchased from Tokyo Chemical Industry (Tokyo, Japan), while HISA hydrochloride from AppliChem (Darmstadt, Germany).
2.2. Alkaline cell lysis and sample preparation The bacterial pellet from Escherichia coli strain HMS174(DE3) containing pJ plasmid (5.17 kbp) was provided by the Microbial Fermentation Group at the Department of Biotechnology (BOKU, Vienna, Austria). The biomass was resuspended in 50 mM TRIS buffer, pH 8.0, containing 10 mM EDTA and treated with cell lysis buffer containing 0.2 M NaOH and 1% SDS followed by a neutralizing solution (3.0 M potassium acetate, adjusted to pH 5.0 with glacial acetic acid, chilled to 4 ◦ C). CaCl2 solution (4.0 M) was slowly added into suspension during constant mixing to obtain 0.5 M overall concentration of CaCl2 . Suspension was left at 4 ◦ C for 15 min. The precipitated material including cell debris, majority of chromosomal DNA, and some RNA and proteins was removed by centrifugation at 9000 × g for 10 min using Sorvall RC5C Plus (Kendro, Newtown, CT) followed by clarification through an inline Sartobran 300 (pores 0.65/0.45 m) filter (Sartorius Stedim, Goettingen, Germany).
2.3. Preparative chromatography The high-pressure gradient LC system (Knauer, Berlin, Germany) used for preparative runs consisted of two Pumps 64, a variable wavelength UV–Vis detector with a 10 mm optical path, with response time set to 0.1 s, connected by means of 0.5 mm I.D. capillary tubes and an HPLC hardware/software. A specially functionalized CIM HISA column with the column bed volume of 1 ml and an average pore size of 2200 nm in diameter was prepared from a CIM CDI column. In a similar way, a CIM HISA disk with the column bed volume of 0.34 ml and average pore sizes of 1500 nm was prepared from a CIM CDI disk. The immobilization procedure consisted of pumping the HISA solution through the monolithic columns (10 g HISA dihydrochloride dissolved in 100 ml of 1.0 M NaOH) followed by 48 h thermostating of the column at 25 ◦ C. Finally the monolithic columns were washed with water and the remaining CDI groups were hydrolyzed by soaking with 1.0 M NaOH for 30 min. The solution of plasmid was always diluted with the appropriate buffer before applying it onto a column at the flow rate of (a) 4.0 ml/min (4 CV/min) for CIM HISA 1 ml column or (b) 2.0 ml/min (5.9 CV/min) for CIM HISA disk. The breakthrough was monitored using the absorbance reading at 260 nm in the case of pure plasmid solutions, while the electrophoresis or chromatographic analysis was used for the evaluation of the pDNA breakthrough in case of loading samples after the lysis that contained other impurities. The loading conditions depended on the chromatographic mode and the exact conditions are described in detail in Section 3. Regeneration of CIM columns were performed after each preparative chromatographic run by washing them with DI water followed by 10 column volumes of 1.0 M NaOH at a flow rate of 1.0 ml/min. At the end, monolithic columns were again thoroughly washed with the deionized water until the proper pH was restored. 2.4. pDNA analysis 2.4.1. Analytical chromatography – HPLC analysis The HPLC analyses were conducted on an Agilent HP 1200 Series chromatograph, coupled with a multiwavelength detector. The chromatographic separations were run on a CIMacTM pDNA analytical column (BIA Separations, Ajdovˇscˇ ina, Slovenia) with the bed volume of 0.318 ml. The quantification of the plasmid was performed using a calibration curve with pure plasmid samples of known concentrations. A pure plasmid was obtained from the lysate using CIMTM HiP2 Plasmid Process Pack 8 ml columns (BIA Separations, Ajdovˇscˇ ina, Slovenia) according to the instruction manual (http:// www.biaseparations.com/library/includes/file.asp?FileId=266). The samples for the calibration curve were additionally treated with Amicon® Ultra centrifugal filters (10,000 MWCO) (Millipore, Cork, Ireland) to exchange buffer to 20 mM TRIS, pH 7.4, and to concentrate the plasmid concentration up to 1.2 mg/ml. Samples with pDNA concentrations between 1.0 and 120 g/ml were prepared by diluting a stock solution with 200 mM TRIS, pH 8.0. All pure pDNA concentration determinations were made spectrophotometrically with a Smart Spec 3000 spectrophotometer (Bio-Rad, Richmond, VI, USA). One unit of OD 260 nm in a 10 mm cuvette is assumed to correspond to 50 g/ml of a double stranded DNA. Concentration obtained by spectrophotometer was correlated with the corresponding peak area on the chromatogram obtained by CIMac analytical column. 2.4.2. Agarose electrophoresis (AGE) pDNA fractions were analyzed by a horizontal AGE (Bio-Rad, Richmond, VI, USA) using SYBR Safe DNA gel stain (Invitrogen, Eugene, OG, USA) and DNA Molecular Weight Marker III, DIGlabeled (Roche Applied Science, Mannheim, Germany). The running
ˇ U. Cernigoj et al. / J. Chromatogr. A 1281 (2013) 87–93
89
Fig. 1. The skeletal formula of immobilized HISA and the schematic presentation of its proteolytic equilibrium in aqueous solutions.
buffer was TBE (40 mM TRIS, 20 mM boric acid, 1 mM EDTA, pH 8.0). Gel was prepared with 7% agarose in TBE buffer. Staining was carried out after 60 min at 80 V for 1 h. The sample buffer was firstly exchanged to 20 mM TRIS, pH 7.4, using PD midiTrapTM G-25 (GE Healthcare, Buckinghamshire, UK). 3. Results and discussion 3.1. Histamine modification The main motive for the research was to find a ligand, presumably with the potential to exhibit both ion-exchange and hydrophobic character that is capable of separating pDNA isoforms on a preparative scale. Different combinations of salt and pH ascending and/or descending gradients were implemented to extensively characterize the selected ligand. Firstly, a multimodal ligand capable of adjusting different ratios between hydrophobic and electrostatic interactions in the vicinity of neutral pH had to be chosen. After that, the selected ligand had to be efficiently and readily immobilized to the stationary phase. Imidazole was found to be a good candidate, because pKa of its conjugate acid is 7.05 at 25 ◦ C [27] being positively charged below pH 7 (strengthening electrostatic interactions) or neutral above pH 7 (strengthening hydrophobic interactions). HISA is decarboxylated histidine and as such it contains an imidazole ring as well as an aminated spacer arm. HISA was immobilized on monolith via CDI activation leading to the neutral carbamate bond (Fig. 1). To deactivate residual CDI groups, a monolith was kept in 1.0 M NaOH for 30 min and the efficiency of the deactivation was confirmed by titrating the remaining CDI groups [28]. 3.2. pDNA analytics
Fig. 2. Analytical separation of a clarified pDNA lysate sample on a CIMac pDNA column and on AGE (insert). Conditions for chromatographic run: Buffer A: 200 mM TRIS, pH 8.0; Buffer B: 200 mM TRIS + 1.0 M NaCl, pH 8.0. Sample: a clarified pDNA lysate as described in Section 2.2. Gradient: buffer A for 2 min, linear gradient from 0% to 60% buffer B (0.6 M NaCl) in 1 min, 5 min hold at 60% buffer B, a linear gradient from 60% to 70% buffer B (0.7 M NaCl) in 10 min, followed by a step gradient to 100% buffer B (1.0 M NaCl) and 1 min hold at 100% buffer B. Flow rate: 1.0 ml/min. Detection: UV at 260 nm. Injection volume: 100 l.
plasmid DNA were base-line separated with the oc isoform eluting at 1.67 M (NH4 )2 SO4 , while the sc isoform eluted at 1.49 M (NH4 )2 SO4 (Fig. 3). To stimulate electrostatic interactions the plasmid was dissolved in 50 mM TRIS, 10 mM EDTA, pH 7.4, and after loading to the column, a linear gradient from 0 to 3.0 M (NH4 )2 SO4 was performed. In this case (NH4 )2 SO4 was not acting as a kosmotrope, but as an additive that increases the ionic strength of the
A CIMac pDNA analytical column was used for all the quantification analyses of sc and oc isoforms in the samples, while AGE was used mainly for qualitative analyses. With a chromatographic method both plasmid isoforms were base-line separated and could therefore be quantified with high precision. Fig. 2 compares the analytical chromatogram and agarose electrophoresis photograph of a clarified lysate containing significant amount of RNA and both plasmid isoforms. While the elution of proteins and RNA takes place below 0.6 M NaCl, both pDNA isoforms are separated within the NaCl concentration window between 0.6 and 0.7 M NaCl. pDNA was quantified as described in Section 2.4.1. 3.3. HISA chromatographic performance at analytical loadings The initial chromatographic experiments with the HISA monolith were performed with analytical amounts of pDNA to investigate the possibility of pDNA isoforms separation. The pure sample containing both sc and oc pDNA isoforms was loaded on the 0.34 ml HISA monolith under two different conditions. In the first experiment, a pDNA sample was adjusted to 3.0 M (NH4 )2 SO4 , 50 mM TRIS, 10 mM EDTA, pH 7.4, to enhance hydrophobic interactions, loaded on the column, and a linear gradient from 3.0 M (NH4 )2 SO4 to 0 M (NH4 )2 SO4 was carried out. The sc and oc isoforms of
Fig. 3. Performance of HISA disk for pure pDNA separation in ascending and descending (NH4 )2 SO4 gradient. MF A: 3.0 M (NH4 )2 SO4 , 50 mM TRIS, 10 mM EDTA, pH 7.4; MF B: 50 mM TRIS, 10 mM EDTA, pH 7.4. Vinj = 100 l, (pDNA) = 0.04 mg/ml, = 260 nm, ˚ = 2.0 ml/min; linear gradient from 0% to 100% MF B and from 100% to 0% MF B in 47 CV.
ˇ U. Cernigoj et al. / J. Chromatogr. A 1281 (2013) 87–93
90
Table 1 The pDNA elution in descending (NH4 )2 SO4 gradient with and without addition of NaCl in the buffers; the experimental conditions the same as described in Fig. 3, except the addition of 0.25 M NaCl in MF A and MF B in experiment B. Buffer composition
oc elution (NH4 )2 SO4 concentration [M]
sc elution (NH4 )2 SO4 concentration [M]
Gradient from 3.0 to 0 M (NH4 )2 SO4 , pH 7.4 Gradient from 3.0 to 0 M (NH4 )2 SO4 in 0.25 M NaCl, pH 7.4
1.67 1.85
1.49 1.68
mobile phase. The pDNA was eluted at 0.85 M (NH4 )2 SO4 and the isoforms were only partially separated (Fig. 3). It is interesting to notice that the concentration of (NH4 )2 SO4 for the pDNA elution was different for two chromatographic modes, thus demonstrating a different binding mechanism of pDNA with the stationary phase. To investigate more in detail the relative contributions of hydrophobic and electrostatic interactions an additional experiment was performed, this time introducing a combination of NaCl in tandem with (NH4 )2 SO4 . A chromatographic method used for enhancing hydrophobic interactions during pDNA loading was used, but in both buffers NaCl concentration was maintained at 0.25 M. Elution of both plasmids occurred at higher (NH4 )2 SO4 concentration (the difference is app. 0.2 M) compared to the experiment where mobile phase contained no NaCl (Table 1). This leads to the conclusion that pDNA interaction with the HISA is partially of ionic character. If this was not the case, the elution of pDNA in the presence of NaCl would be expected to occur at lower (NH4 )2 SO4 concentration due to an additional kosmotropic effect provided by NaCl. To explore even further the ion-exchange characteristics of pDNA loaded on a HISA monolith a NaCl gradient at different pH values, namely 5.0, 6.2 and 7.4, was tested. A partial separation between sc and oc isoforms of the plasmid DNA was observed regardless of the pH used, but a higher NaCl concentration was needed to elute pDNA at a lower pH (Table 2). This can be explained by the titration state of imidazole. Although an approximate value of HISA isoelectric point is 7.0, at pH 7.4 there are still around 30% of positively charged groups. In contrast, at pH 5 the ratio between neutral and positively charged ligand is 1:100, so 99% of ligand molecules are charged. The described behavior results in two important conclusions. Operating the ligand below pH 6.0 allows an effective plasmid DNA binding to occur even when the NaCl concentration is increased up to 1.0 M. The reason for this is the ion-exchange binding due to a completely charged surface. Additionally, changing a charge of the monolith surface enables the application of a pH gradient from neutral to slightly alkaline values to elute the electrostatically bound biomolecules at relatively mild conditions. 3.4. Dynamic binding capacities (DBC) for pure plasmid samples To investigate the applicability of the novel matrix for preparative pDNA purification a dynamic binding capacity (DBC) for pure plasmid at different conditions was carried out. The flow rate on HISA disks was fixed at 11 CV/min throughout the experimental work, because it was demonstrated that it had little effect on DBC [5]. Four different experiments were performed at two different pH values. The results are presented in Table 2. Two experiments were
done in the absence of NaCl, but at two different pH values of 5.0 and 6.2. The DBC was around 4 mg/ml regardless the buffer composition or the pH value. Additionally, we performed an experiment in the presence of high NaCl concentration in the loading buffer. 0.75 M NaCl was added at pH 6.2, while 1.0 M NaCl at pH 5.0. As expected, the DBC (between 2.5 and 3.0 mg/ml) was lower than in the absence of salt (4.0 mg/ml), but in comparison to the conventional weak anion exchanger, such as DEA, which loses the capacity for the plasmid at 1.0 M NaCl completely, the use of HISA ligand is advantageous [29]. Binding at such high NaCl concentration is important to capture pDNA from clarified lysate at high conductivity in presence of proteins and RNA, which are not retained by the stationary phase. The DBC was also measured in the case, where the plasmid was loaded in 3.0 M (NH4 )2 SO4 buffer, pH 7.4. The DBC at 10% breakthrough was 2.7 mg of pDNA/ml of support, which is comparable to a DBC on conventional CIM C4 monoliths, measured at the same conditions (internal results).
3.5. Separation of pDNA isoforms in descending (NH4 )2 SO4 gradient Based on the preliminary results, an ascending NaCl gradient and the use of acidic buffers were selected in a capture step for pDNA from the lysate (removing RNA, proteins). On the contrary, plasmid isoforms separation is more effective in the descending (NH4 )2 SO4 gradient and it would therefore preferably be used in a polishing step. The important consideration is the separation of the isoforms at a preparative loading in descending (NH4 )2 SO4 gradient. In the first experiment, 2.64 mg of pure pDNA, dissolved in 3.0 M (NH4 )2 SO4 , pH 7.4, was loaded (the start of breakthrough) followed by the elution of the pDNA in a linear gradient to 0 M of (NH4 )2 SO4 . Only one broad peak was obtained (data not shown), where both oc and sc isoforms were overlapping. The agarose gel electrophoresis analysis of fractions and chromatographic pDNA analytics of the elution confirmed pure oc isoform in the first fractions, followed by fractions of mixtures of both isoforms and finally by pure sc isoform. A step gradient was applied instead of a linear one and its optimization enabled the separation of pDNA isoforms to a high purity. The optimal (NH4 )2 SO4 concentration for oc wash was determined to be 2.3 M and the results are shown in Table 3. The ratio between sc and oc isoforms was increased from 5.6 in the load to 70 in the main sc elution fraction. A complete separation of both isoforms was not achievable. The homogeneity of the sc isoform was 98.5%, with 62% of the loaded sc pDNA eluting in this fraction. The overall recovery of the plasmid DNA (all elution fractions divided by the load) for both isoforms is higher than 90%.
Table 2 Concentrations of NaCl for pDNA elution in buffers with different pH values in ascending salt gradient; CIM HISA disk, pJ pDNA (0.045 mg/ml) loading in appropriate buffer, = 260 nm, ˚ = 2.0 ml/min; linear gradient (0.013 M NaCl per CV). Buffer A composition
Buffer B composition
pH of mobile phase
c (NaCl) for sc pDNA elution [M]
DBC at 10% breakthrough for pDNA (mg plasmid/ml of support)
50 mM acetate 50 mM phosphate 50 mM TRIS
50 mM acetate + 3.0 M NaCl 50 mM phosphate + 2.0 M NaCl 50 mM TRIS + 1.0 M NaCl
5.0 6.2 7.4
2.4 1.55 0.55
4.2 (in absence of NaCl) 2.9 (in presence of 1.0 M NaCl) 4.0 (in absence of NaCl)2.6 (in presence of 0.75 M NaCl) Not measured
ˇ U. Cernigoj et al. / J. Chromatogr. A 1281 (2013) 87–93
91
Table 3 Step-gradient separation of both pDNA isoforms in descending (NH4 )2 SO4 gradient after preparative loading in 3.0 M (NH4 )2 SO4 , pH 7.4: 1 ml CIM HISA column, ˚ = 4.0 ml/min, pDNA (0.06 mg/ml). Step
Buffer
V [ml]
oc amount (mg)
% of oc
sc amount (mg)
% of sc
Load (0.06 pDNA mg/ml) Flow through fraction Washing (removal of oc isoform) Elution of sc isoform Washing with 2 M NaCl Regeneration Overall recovery
3.0 M (NH4 )2 SO4 , 50 mM TRIS, 10 mM EDTA 3.0 M (NH4 )2 SO4 , 50 mM TRIS, 10 mM EDTA 2.3 M (NH4 )2 SO4 , 50 mM TRIS, 10 mM EDTA 50 mM TRIS, 10 mM EDTA 50 mM TRIS, 10 mM EDTA, 2.0 M NaCl 0.5 M NaOH
44 44 9.3 11.4 8 4.5
0.4 0 0.35 0.02 0.01 0 0.38
100 0 87.3 5.6 1.4 0 94.3
2.24 0.02 0.57 1.40 0.05 0.01 2.05
100 1.0 25.6 62.6 1.9 0.3 91.3
3.6. Separation of pDNA isoforms in a pH gradient It was found that the oc isoform of the plasmid can be selectively eluted at constant 3.0 M (NH4 )2 SO4 , while changing the pH of the mobile phase. A representative chromatogram is shown in Fig. 4. Plasmid DNA was completely retained by the monolith and the only eluted compound was an oc pDNA isoform at the end of the pH gradient. In this way, the oc form can be selectively removed from the sc form. To confirm this result and to check the efficiency of the oc removal a following experiment was performed. After
implementation of the pH gradient where the oc form was selectively removed, the column was reequilibrated to initial conditions (3.0 M (NH4 )2 SO4 , 50 mM TRIS, 10 mM EDTA, pH 7.4) and a linear gradient to 0 M (NH4 )2 SO4 , at a constant pH value of 7.4 was carried out. At a retention time, where the oc peak should occur, there was almost no detectable peak, while the sc peak was eluted normally (see Fig. 4B). This served as a proof that the oc form was efficiently removed. Additionally, as a control experiment, the same sample without a pre-step (pH gradient) was separated in the (NH4 )2 SO4 gradient from 3.0 to 0 M, at a constant pH of 7.4. In this experiment a well resolved oc peak appeared, while the sc peak had the same area and retention time as in case of a pretreated sample. 3.7. Applying pH gradient in preparative pDNA isoform separation
Fig. 4. Elution of oc isoform of the pJ in pH gradient. Monolith: HISA disk, ˚ = 2.0 ml/min, = 260 nm, Vinj = 500 l, pDNA (0.045 mg/ml) in 3.0 M (NH4 )2 SO4 , pH 7.4. (A) Elution of oc pDNA in pH gradient: MF A: 50 mM TRIS, 3.0 M (NH4 )2 SO4 , 10 mM EDTA, pH = 7.4; MF B: 3.0 M (NH4 )2 SO4 , 50 mM acetate, 10 mM EDTA, pH = 4.5, gradient from 0% to 100% MF B in 50 CV; (B) a proof for selective removal of oc pDNA isoform from the mixture: gradient from 3.0 to 0 M (NH4 )2 SO4 in 50 mM TRIS, 10 mM EDTA, pH 7.4; ( ) initial sample; ( ) sample firstly pretreated according to procedure in (A), then elution in descending (NH4 )2 SO4 gradient.
Based on the results previously obtained the plasmid DNA sample was loaded under conditions, where oc isoform was not retained by the column. In this case, all the oc was supposed to be in the flow-through fraction, while the pure sc isoform would bind to the chromatographic support. When a pDNA isoform mixture dissolved in 2.9 M (NH4 )2 SO4 , 50 mM acetate, 10 mM EDTA, pH 5, was loaded on the column, the DBC for oc isoform was around 0.1 mg/ml of support. After the oc breakthrough a plateau was reached, where only sc isoform was retained. The calculated DBC for the sc isoform was between 0.5 and 1 mg/ml of HISA chromatographic support, depending on the small variations in (NH4 )2 SO4 concentration. We decided to optimize the novel method in a way that the plasmid was loaded at high (NH4 )2 SO4 concentration, pH 7.4 (preferential hydrophobic interactions) with the following washing step of oc isoform at high (NH4 )2 SO4 concentration, but in a slightly acidic buffer. Three main parameters were optimized during this study: the amount of loaded plasmid, the concentration of (NH4 )2 SO4 in the washing buffer and the pH of the washing buffer. Changing a pH in the washing buffer between 4.0 and 5.5 did not influence the elution pattern. The titration of HISA is completed below pH 6 and the surface is almost fully charged with protonated HISAs. At the same time, the plasmid is almost fully negatively charged at pH above 3, therefore the robust conditions for pDNA elution were achieved between pH 4.0 and 5.5. Optimization of the amount of loaded plasmid followed. The DBC for sc isoform is app. 0.7 mg of plasmid/ml of the bed volume in 3.0 M (NH4 )2 SO4 , pH 4.5. The only possibility to increase the capacity of the column at high (NH4 )2 SO4 concentration and slightly acidic pH was to increase the (NH4 )2 SO4 concentration to achieve higher hydrophobicity of the mobile phase, where more plasmid could be retained on the column. After optimization of the parameters the following conditions were obtained: if app. 1.6 mg of pDNA per ml of chromatographic bed volume was loaded in 3.0 M (NH4 )2 SO4 , pH 7.4, more than 90% of oc isoform together with app. 10% of sc isoform was washed with 3.2 M (NH4 )2 SO4 , pH 5. Another challenge was how to perform an efficient elution of sc pDNA (to achieve as high as possible recovery and at the same time as high as possible concentration of the eluted
ˇ U. Cernigoj et al. / J. Chromatogr. A 1281 (2013) 87–93
92
Table 4 Step-gradient separation of pDNA isoforms in pH gradient at high (NH4 )2 SO4 concentration after preparative loading in 3.0 M (NH4 )2 SO4 , pH 7.4. The experimental parameters are described in caption of Fig. 5.
Load (1.58 mg of pDNA/ml of bed volume) Washing step (3.2 M (NH4 )2 SO4 , pH 5.0) sc elution step (0.6 M (NH4 )2 SO4 , pH 7.4) Recovery (%)
oc [mg]
% of oc compared to the oc amount in load
sc [mg]
% of sc compared to the sc amount in load
sc + oc [mg]
sc:oc
(sc) [mg/ml]
0.23 0.20 0.02 96
100 87 9
1.33 0.03 1.29 99
100 2 97
1.58 0.23 1.31 98
5.7 0.2 64
0.053 0.003 0.156
plasmid) after the washing of oc isoform. After optimization the only well-working procedure was to increase the pH back to 7.4, while keeping the (NH4 )2 SO4 concentration at 3.2 M. In this way, the sc pDNA was retained by the column. Then the elution was carried out by descending (NH4 )2 SO4 gradient at the same pH. The whole process (Fig. 5A) was analyzed by pDNA chromatographic analysis (Fig. 5B) and quantified (Table 4). Fig. 5A confirms the absence of pDNA elutions in FT and washing steps as well as in buffer exchange steps. Fig. 5B is a direct proof of purity of both fractions containing eluted pDNA isoforms while the quantification
results are gathered in Table 4. There is 97% recovery of the sc pDNA in the main elution fraction. Only 2% of the sc isoform was lost in an oc washing fraction. The sc:oc ratio was improved from 5.7 in the load to 64 in the main elution fraction. The plasmid concentration was increased approximately 3 times from load to the main elution fraction. 4. Conclusions HISA was immobilized to a monolithic chromatographic support and studied as a multimodal ligand for pDNA purification. In ascending NaCl gradient (preferentially electrostatic interactions) an elution pattern of pDNA was studied in different buffers with the pH value of the buffer as the main changing parameter. Due to the appropriate pKa value of HISA ligand a NaCl concentration for pDNA elution increases considerably from pH 7.4 to pH 5.0. Operating the ligand below pH 6 allows effective plasmid binding to occur at NaCl concentrations higher than 1.0 M, enabling the preparative loading of plasmid at high conductivities and at a slightly acidic pH of the loading buffer. The column is capable of separating efficiently the pDNA isoforms in the descending (NH4 )2 SO4 gradient (preferentially hydrophobic interactions) as well. Additionally, novel pH gradient conditions were proven to selectively elute both plasmid isoforms. Instead of applying (NH4 )2 SO4 descending gradient, a pH gradient at constant 3.0 M (NH4 )2 SO4 was used with the similar efficiency. After optimizaton of the chromatographic parameters we were able to purify and concentrate 1.3 mg of sc pDNA per ml of column bed volume with recovery higher than 95% and less than 2% oc pDNA impurities. The main goal of the subsequent research will be the development of a chromatographic pDNA purification process from the clarified lysate to the final sc pDNA fraction using HISA monolithic column, where the emphasis will be focused to the efficient removal of other process and product related impurities besides the oc isoform. Acknowledgements
Fig. 5. The separation of oc and sc pDNA isoforms on preparative scale employing the novel chromatographic conditions. 1 ml CIM HISA column, ˚ = 4.0 ml/min. Loading: pDNA (0.06 mg/ml) in 3.0 M (NH4 )2 SO4 , 50 mM TRIS, 10 mM EDTA, pH 7.4; washing step: 3.2 M (NH4 )2 SO4 , 50 mM acetate, 10 mM EDTA, pH 5.0; buffer exchange step: 3.2 M (NH4 )2 SO4 , 50 mM TRIS, 10 mM EDTA, pH 7.4; sc elution step: 0.6 M (NH4 )2 SO4 , 50 mM TRIS, 10 mM EDTA, pH 7.4. (A) UV signal at 260 nm during pDNA isoforms separation process; (B) analysis of some fractions using CIMac pDNA column.
We are very grateful to Dr. Fani Sousa and Dr. Angela Sousa from Centro de Investigac¸ão em Ciências da Saúde, Universidade da Beira Interior, for the fruitful discussions on pDNA pseudoaffinity ligands behavior and for the help in immobilization of amino acids and their derivates to monolithic chromatographic supports. We are greatly thankful to Dr. Gerald Striedner and Dr. Markus Luchner, both from the Microbial Fermentation Group at Department of Biotechnology, BOKU, Vienna, for providing us a pDNA sample. We would also like to thank Boˇstjan Gabor and Dr. Nika Lendero Krajnc from BIA Separations for giving us practical advice from their pDNA purification knowledge. Tina Jakop, Sandra Kontrec and Jernej Gaˇsperˇsiˇc from BIA Separations and Rok Oblak are acknowledged for their help in laboratory work. This work was partially financially supported by the Ministry of Education, Culture, Science and Sport of the Republic of Slovenia within the research programme P4-0469. The Centre of Excellence for Biosensors, Instrumentation and Process Control is an operation financed by the European Union, European Regional Development Fund and
ˇ U. Cernigoj et al. / J. Chromatogr. A 1281 (2013) 87–93
Republic of Slovenia, Ministry of Education, Science, Culture and Sport. References [1] S. Gurunathan, D.M. Klinman, R.A. Seder, Annu. Rev. Immunol. 18 (2000) 927. [2] I. Ruzila, N.A. Zeenathul, M.L. Mohd-Azmi, Vaccine 30 (2012) 5914. [3] C.A. Holladay, T. O’Brien, A. Pandit, Wiley Interdiscip. Rev.: Nanomed. Nanobiotechnol. 2 (2010) 721. [4] A. Ghanem, R. Healey, F.G. Adly, Anal. Chim. Acta 760 (2013) 1. ˇ A. Podgornik, J. Chromatogr. A 1218 [5] N. Lendero Krajnc, F. Smrekar, A. Strancar, (2011) 2413. [6] P.-O. Wahlund, P.-E. Gustavsson, V.A. Izumrudov, P.-O. Larsson, I.Yu. Galaev, Biotechnol. Bioeng. 87 (2004) 675. [7] G.A. Gomes, A.M. Azevedo, M.R. Aires-Barros, D.M.F. Prazeres, Sep. Purif. Technol. 65 (2009) 22. [8] D.W. Kahn, M.D. Butler, D.L. Cohen, M. Gordon, J.W. Kahn, M.E. Winkler, Biotechnol. Bioeng. 69 (2000) 101. [9] J. Stadler, R. Lemmens, T. Nyhammar, J. Gene Med. 6 (2004) S54. [10] G.N.M. Ferreira, J.M.S. Cabral, D.M.F. Prazeres, Biotechnol. Tech. 11 (1997) 417. [11] C.M. Ongkudon, M.K. Danquah, J. Chromatogr. B 878 (2010) 2719. [12] L.R. Pereira, D.M.F. Prazeres, M. Mateus, J. Sep. Sci. 33 (2010) 1175. [13] M.J. Shin, L. Tan, M.H. Jeong, J.-H. Kim, W.-S. Choe, J. Chromatogr. A 1218 (2011) 5273.
93
[14] F. Sousa, D.M.F. Prazeres, J.A. Queiroz, Trends Biotechnol. 26 (2008) 518. [15] R.A. Darby, G.M. Forde, N.K. Slater, A.V. Hine, Biotechnol. Bioeng. 98 (2007) 1103. [16] T. Schluep, C.L. Cooney, Nucleic Acids Res. 26 (1998) 4524. [17] Y. Han, G.M. Forde, J. Chromatogr. B 874 (2008) 21. [18] L.M. Sandberg, A. Bjurling, P. Busson, J. Vasi, R. Lemmens, J. Biotechnol. 109 (2004) 193. [19] C. Caramelo-Nunes, M.F. Gabriel, P. Almeida, J.C. Marcos, C.T. Tomaz, J. Chromatogr. B 904 (2012) 81. [20] C. Caramelo-Nunes, T. Tente, P. Almeida, J.C. Marcos, C.T. Tomaz, Anal. Biochem. 412 (2011) 153. [21] F. Sousa, C.T. Tomaz, D.M.F. Prazeres, J.A. Queiroz, Anal. Biochem. 343 (2005) 183. [22] F. Sousa, T. Matos, D.M.F. Prazeres, J.A. Queiroz, Anal. Biochem. 374 (2008) 432. ˇ A. Podgornik, in: B. [23] N. Lendero Krajnc, F. Smrekar, V. Frankoviˇc, A. Strancar, Mattiasson, A. Kumar, I. Galaev, I. Yu (Eds.), Monolithic Macroporous Polymers as Chromatographic Matrices, CRC Press, Florida, 2010, p. 291. [24] J. Urthaler, R. Schlegl, A. Podgornik, A. Strancar, A. Jungbauer, R. Necina, J. Chromatogr. A 1065 (2005) 93. ˇ [25] F. Smrekar, A. Podgornik, M. Ciringer, S. Kontrec, P. Raspor, A. Strancar, M. Peterka, Vaccine 28 (2010) 2039. [26] A. Sousa, D. Bicho, C.T. Tomaz, F. Sousa, J.A. Queiroz, J. Chromatogr. A 1218 (2011) 1701. [27] H. Walba, R.W. Isensee, J. Org. Chem. 26 (1961) 2789. ˇ M. Benˇcina, J. Sep. Sci. 27 (2004) 811. [28] K. Benˇcina, A. Podgornik, A. Strancar, [29] C.M. Ongkudon, M.K. Danquah, Sep. Purif. Technol. 78 (2011) 303.
Contents lists available at SciVerse ScienceDirect
Journal of Chromatography A journal homepage: www.elsevier.com/locate/chroma
A multimodal histamine ligand for chromatographic purification of plasmid DNA夽 a a,b ˇ ˇ Urh Cernigoj , Urˇska Vidic a , Miloˇs Barut a,b,∗ , Aleˇs Podgornik a,b , Matjaˇz Peterka a,b , Aleˇs Strancar a b
BIA Separations d.o.o., Mirce 21, SI-5270 Ajdovˇscˇ ina, Slovenia The Centre of Excellence for Biosensors, Instrumentation and Process Control – COBIK, Velika pot 22, SI-5250 Solkan, Slovenia
a r t i c l e
i n f o
Article history: Received 3 October 2012 Received in revised form 11 January 2013 Accepted 11 January 2013 Available online 23 January 2013 Keywords: Plasmid DNA Chromatography Immobilized histamine Multimodal interactions Monoliths
a b s t r a c t To exploit different chromatographic modes for efficient plasmid DNA (pDNA) purification a novel monolithic chromatographic support bearing multimodal histamine (HISA) groups was developed and characterized. Electrostatic charge of HISA groups depends on the pH of the mobile phase, being neutral above pH 7 and becoming positively charged below. As a consequence, HISA groups exhibit predominantly ion-exchange character at low pH values, which decreases with titration of the HISA groups resulting in increased hydrophobicity. This feature enabled separation of supercoiled (sc) pDNA from other plasmid isoforms (and other process related impurities) by adjusting salt or pH gradient. The dynamic binding capacity (DBC) for a 5.1 kbp large plasmid at pH 5 was 4.0 mg/ml under low salt binding conditions, remaining relatively high (3.0 mg/ml) even in the presence of 1.0 M NaCl due to the multimodal nature of HISA ligand. Only slightly lower DBC (2.7 mg/ml) was determined under preferentially hydrophobic conditions in 3.0 M (NH4 )2 SO4 , pH 7.4. Open circular and sc pDNA isoforms were baseline separated in descending (NH4 )2 SO4 gradient. Furthermore, an efficient plasmid DNA separation was possible both on analytical as well as on preparative scale by applying the descending pH gradient at a constant concentration (above 3.0 M) of (NH4 )2 SO4 . © 2013 Elsevier B.V. All rights reserved.
1. Introduction The purification strategy of supercoiled plasmid DNA (sc pDNA), a biological macromolecule with increasing pharmaceutical potential [1–3], involves challenges and constraints that are rarely encountered with other classes of biomolecules [4]. The main reasons are: (1) a low content of pDNA in the cells, (2) the mechanical lability of sc pDNA and (3) impurities sharing common characteristics to pDNA (negative charge – RNA, open circular pDNA (oc pDNA), genomic DNA (gDNA) and endotoxins; high molecular mass – oc pDNA, gDNA and endotoxins; hydrophobicity – endotoxins) [5]. Purification process commonly starts with an alkaline lysis that removes part of the cell walls, organelles, proteins and gDNA, but leaves RNA as the main contaminant together with some proteins, endotoxins and gDNA [6]. Further purification steps can be non-chromatographic e.g., aqueous two-phase extraction [7], tangential flow filtration [8], selective precipitation of pDNA [6], but despite some advantages they are usually followed by a
夽 Presented at the MSS 2012, Ajdovˇscˇ ina, Portoroˇz, Slovenia, 1–6 June 2012. ∗ Corresponding author at: BIA Separations d.o.o., Mirce 21, SI-5270 Ajdovˇscˇ ina, Slovenia. Tel.: +386 59 699 500; fax: +386 59 699 699. E-mail address: [email protected] (M. Barut). 0021-9673/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.chroma.2013.01.058
polishing chromatographic step to obtain the product matching the FDA recommendations [9]. Due to these stringent regulatory requests and due to process economics, the majority of downstream processes are based on chromatography. There are different chromatographic separation mechanisms implemented for pDNA purification, such as size-exclusion [10], anion-exchange [11], hydrophobic interaction chromatography (HIC) [12], IMAC [13] or affinity chromatography [14]. However, to completely separate the pDNA from all other impurities on preparative scale, they must be combined in a suitable, robust and efficient purification process. Even the use of highly specific affinity ligands in some cases lacks sufficient selectivity [15] and especially suffers from slow binding kinetics and low recovery of pDNA [16]. Besides, the ligand costs are prohibitively high. A novel approach of immobilizing only 16mer peptide representing helix II of DNA binding domain of lac repressor was recently used in pDNA purification [17], which seems to overcome the above mentioned limitations but on the expense of high ligand costs. As an alternative, pseudo affinity ligands for pDNA purification are often studied as cheaper substitutes for affinity ligands. Thiophilic chromatography [18] was successfully implemented into a chromatographic process of sc pDNA purification, but it still needed two additional chromatographic steps to complete the plasmid isolation. Certain immobilized antibiotics and anticancer agents
88
ˇ U. Cernigoj et al. / J. Chromatogr. A 1281 (2013) 87–93
binding DNA with high specificity were tested as pseudo affinity ligand as well [19]. An example of such a ligand is immobilized berenil [20], where the purification of pDNA directly from the lysate is achieved in descending (NH4 )2 SO4 gradient. Its benefits were proven for different plasmid isoforms separation, but no results were shown for the process efficiency on preparative scale. Even certain amino acids, such as lysine, arginine and histidine interact with DNA through multimodal mechanism [21,22] and they were successfully applied for the separation of pDNA isoforms on an analytical scale. To achieve high binding capacity, target functionalities have to be introduced onto proper chromatographic support. Monoliths exhibit several advantages over beaded supports when implemented for sc pDNA purification due to large interconnected convective pores and high surface accessibility [23–25]. The methacrylate based monolithic chromatographic supports have already been successfully implemented for pDNA isoforms separation using a pseudo affinity ligand (imidazole) on a laboratory scale [26]. The main purpose of the present work was to further investigate the performance of amino acid based multimodal ligands. Histamine (HISA), a derivative of amino acid histidine, containing an imidazole ring on a spacer arm was used as a ligand of choice due to its unique chemical properties. Its applicability for the separation of pDNA isoforms under different chromatographic conditions on analytical as well as on preparative scale was studied in detail.
2. Materials and methods 2.1. Materials All solutions were freshly prepared using purified water which meets the requirements for European Pharmacopoeia (AQUATEHNA Biro, Zgornja Kungota, Slovenia) and analytical grade reagents. Buffer solutions were filtered through a 0.22 m PES filter (TPP, Trasadingen, Switzerland). The samples injected in the column were filtered through cellulose mixed esters syringe filters, pore size 0.45 m (Macherey-Nagel, Düren, Germany). Agarose gel was from SeaKem LE Agarose (Lonza Group, Basel, Switzerland). NaOH, TRIS, potassium acetate, CaCl2 , EDTA, (NH4 )2 SO4 , and boric acid were from Merck (Darmstadt, Germany). SDS, Na2 HPO4 ·2H2 O, NaH2 PO4 ·2H2 O, acetic acid, NaOH and NaCl were from Sigma–Aldrich (St. Louis, MO, USA). 1 ,1 Carbonyldiimidazol (CDI) was purchased from Tokyo Chemical Industry (Tokyo, Japan), while HISA hydrochloride from AppliChem (Darmstadt, Germany).
2.2. Alkaline cell lysis and sample preparation The bacterial pellet from Escherichia coli strain HMS174(DE3) containing pJ plasmid (5.17 kbp) was provided by the Microbial Fermentation Group at the Department of Biotechnology (BOKU, Vienna, Austria). The biomass was resuspended in 50 mM TRIS buffer, pH 8.0, containing 10 mM EDTA and treated with cell lysis buffer containing 0.2 M NaOH and 1% SDS followed by a neutralizing solution (3.0 M potassium acetate, adjusted to pH 5.0 with glacial acetic acid, chilled to 4 ◦ C). CaCl2 solution (4.0 M) was slowly added into suspension during constant mixing to obtain 0.5 M overall concentration of CaCl2 . Suspension was left at 4 ◦ C for 15 min. The precipitated material including cell debris, majority of chromosomal DNA, and some RNA and proteins was removed by centrifugation at 9000 × g for 10 min using Sorvall RC5C Plus (Kendro, Newtown, CT) followed by clarification through an inline Sartobran 300 (pores 0.65/0.45 m) filter (Sartorius Stedim, Goettingen, Germany).
2.3. Preparative chromatography The high-pressure gradient LC system (Knauer, Berlin, Germany) used for preparative runs consisted of two Pumps 64, a variable wavelength UV–Vis detector with a 10 mm optical path, with response time set to 0.1 s, connected by means of 0.5 mm I.D. capillary tubes and an HPLC hardware/software. A specially functionalized CIM HISA column with the column bed volume of 1 ml and an average pore size of 2200 nm in diameter was prepared from a CIM CDI column. In a similar way, a CIM HISA disk with the column bed volume of 0.34 ml and average pore sizes of 1500 nm was prepared from a CIM CDI disk. The immobilization procedure consisted of pumping the HISA solution through the monolithic columns (10 g HISA dihydrochloride dissolved in 100 ml of 1.0 M NaOH) followed by 48 h thermostating of the column at 25 ◦ C. Finally the monolithic columns were washed with water and the remaining CDI groups were hydrolyzed by soaking with 1.0 M NaOH for 30 min. The solution of plasmid was always diluted with the appropriate buffer before applying it onto a column at the flow rate of (a) 4.0 ml/min (4 CV/min) for CIM HISA 1 ml column or (b) 2.0 ml/min (5.9 CV/min) for CIM HISA disk. The breakthrough was monitored using the absorbance reading at 260 nm in the case of pure plasmid solutions, while the electrophoresis or chromatographic analysis was used for the evaluation of the pDNA breakthrough in case of loading samples after the lysis that contained other impurities. The loading conditions depended on the chromatographic mode and the exact conditions are described in detail in Section 3. Regeneration of CIM columns were performed after each preparative chromatographic run by washing them with DI water followed by 10 column volumes of 1.0 M NaOH at a flow rate of 1.0 ml/min. At the end, monolithic columns were again thoroughly washed with the deionized water until the proper pH was restored. 2.4. pDNA analysis 2.4.1. Analytical chromatography – HPLC analysis The HPLC analyses were conducted on an Agilent HP 1200 Series chromatograph, coupled with a multiwavelength detector. The chromatographic separations were run on a CIMacTM pDNA analytical column (BIA Separations, Ajdovˇscˇ ina, Slovenia) with the bed volume of 0.318 ml. The quantification of the plasmid was performed using a calibration curve with pure plasmid samples of known concentrations. A pure plasmid was obtained from the lysate using CIMTM HiP2 Plasmid Process Pack 8 ml columns (BIA Separations, Ajdovˇscˇ ina, Slovenia) according to the instruction manual (http:// www.biaseparations.com/library/includes/file.asp?FileId=266). The samples for the calibration curve were additionally treated with Amicon® Ultra centrifugal filters (10,000 MWCO) (Millipore, Cork, Ireland) to exchange buffer to 20 mM TRIS, pH 7.4, and to concentrate the plasmid concentration up to 1.2 mg/ml. Samples with pDNA concentrations between 1.0 and 120 g/ml were prepared by diluting a stock solution with 200 mM TRIS, pH 8.0. All pure pDNA concentration determinations were made spectrophotometrically with a Smart Spec 3000 spectrophotometer (Bio-Rad, Richmond, VI, USA). One unit of OD 260 nm in a 10 mm cuvette is assumed to correspond to 50 g/ml of a double stranded DNA. Concentration obtained by spectrophotometer was correlated with the corresponding peak area on the chromatogram obtained by CIMac analytical column. 2.4.2. Agarose electrophoresis (AGE) pDNA fractions were analyzed by a horizontal AGE (Bio-Rad, Richmond, VI, USA) using SYBR Safe DNA gel stain (Invitrogen, Eugene, OG, USA) and DNA Molecular Weight Marker III, DIGlabeled (Roche Applied Science, Mannheim, Germany). The running
ˇ U. Cernigoj et al. / J. Chromatogr. A 1281 (2013) 87–93
89
Fig. 1. The skeletal formula of immobilized HISA and the schematic presentation of its proteolytic equilibrium in aqueous solutions.
buffer was TBE (40 mM TRIS, 20 mM boric acid, 1 mM EDTA, pH 8.0). Gel was prepared with 7% agarose in TBE buffer. Staining was carried out after 60 min at 80 V for 1 h. The sample buffer was firstly exchanged to 20 mM TRIS, pH 7.4, using PD midiTrapTM G-25 (GE Healthcare, Buckinghamshire, UK). 3. Results and discussion 3.1. Histamine modification The main motive for the research was to find a ligand, presumably with the potential to exhibit both ion-exchange and hydrophobic character that is capable of separating pDNA isoforms on a preparative scale. Different combinations of salt and pH ascending and/or descending gradients were implemented to extensively characterize the selected ligand. Firstly, a multimodal ligand capable of adjusting different ratios between hydrophobic and electrostatic interactions in the vicinity of neutral pH had to be chosen. After that, the selected ligand had to be efficiently and readily immobilized to the stationary phase. Imidazole was found to be a good candidate, because pKa of its conjugate acid is 7.05 at 25 ◦ C [27] being positively charged below pH 7 (strengthening electrostatic interactions) or neutral above pH 7 (strengthening hydrophobic interactions). HISA is decarboxylated histidine and as such it contains an imidazole ring as well as an aminated spacer arm. HISA was immobilized on monolith via CDI activation leading to the neutral carbamate bond (Fig. 1). To deactivate residual CDI groups, a monolith was kept in 1.0 M NaOH for 30 min and the efficiency of the deactivation was confirmed by titrating the remaining CDI groups [28]. 3.2. pDNA analytics
Fig. 2. Analytical separation of a clarified pDNA lysate sample on a CIMac pDNA column and on AGE (insert). Conditions for chromatographic run: Buffer A: 200 mM TRIS, pH 8.0; Buffer B: 200 mM TRIS + 1.0 M NaCl, pH 8.0. Sample: a clarified pDNA lysate as described in Section 2.2. Gradient: buffer A for 2 min, linear gradient from 0% to 60% buffer B (0.6 M NaCl) in 1 min, 5 min hold at 60% buffer B, a linear gradient from 60% to 70% buffer B (0.7 M NaCl) in 10 min, followed by a step gradient to 100% buffer B (1.0 M NaCl) and 1 min hold at 100% buffer B. Flow rate: 1.0 ml/min. Detection: UV at 260 nm. Injection volume: 100 l.
plasmid DNA were base-line separated with the oc isoform eluting at 1.67 M (NH4 )2 SO4 , while the sc isoform eluted at 1.49 M (NH4 )2 SO4 (Fig. 3). To stimulate electrostatic interactions the plasmid was dissolved in 50 mM TRIS, 10 mM EDTA, pH 7.4, and after loading to the column, a linear gradient from 0 to 3.0 M (NH4 )2 SO4 was performed. In this case (NH4 )2 SO4 was not acting as a kosmotrope, but as an additive that increases the ionic strength of the
A CIMac pDNA analytical column was used for all the quantification analyses of sc and oc isoforms in the samples, while AGE was used mainly for qualitative analyses. With a chromatographic method both plasmid isoforms were base-line separated and could therefore be quantified with high precision. Fig. 2 compares the analytical chromatogram and agarose electrophoresis photograph of a clarified lysate containing significant amount of RNA and both plasmid isoforms. While the elution of proteins and RNA takes place below 0.6 M NaCl, both pDNA isoforms are separated within the NaCl concentration window between 0.6 and 0.7 M NaCl. pDNA was quantified as described in Section 2.4.1. 3.3. HISA chromatographic performance at analytical loadings The initial chromatographic experiments with the HISA monolith were performed with analytical amounts of pDNA to investigate the possibility of pDNA isoforms separation. The pure sample containing both sc and oc pDNA isoforms was loaded on the 0.34 ml HISA monolith under two different conditions. In the first experiment, a pDNA sample was adjusted to 3.0 M (NH4 )2 SO4 , 50 mM TRIS, 10 mM EDTA, pH 7.4, to enhance hydrophobic interactions, loaded on the column, and a linear gradient from 3.0 M (NH4 )2 SO4 to 0 M (NH4 )2 SO4 was carried out. The sc and oc isoforms of
Fig. 3. Performance of HISA disk for pure pDNA separation in ascending and descending (NH4 )2 SO4 gradient. MF A: 3.0 M (NH4 )2 SO4 , 50 mM TRIS, 10 mM EDTA, pH 7.4; MF B: 50 mM TRIS, 10 mM EDTA, pH 7.4. Vinj = 100 l, (pDNA) = 0.04 mg/ml, = 260 nm, ˚ = 2.0 ml/min; linear gradient from 0% to 100% MF B and from 100% to 0% MF B in 47 CV.
ˇ U. Cernigoj et al. / J. Chromatogr. A 1281 (2013) 87–93
90
Table 1 The pDNA elution in descending (NH4 )2 SO4 gradient with and without addition of NaCl in the buffers; the experimental conditions the same as described in Fig. 3, except the addition of 0.25 M NaCl in MF A and MF B in experiment B. Buffer composition
oc elution (NH4 )2 SO4 concentration [M]
sc elution (NH4 )2 SO4 concentration [M]
Gradient from 3.0 to 0 M (NH4 )2 SO4 , pH 7.4 Gradient from 3.0 to 0 M (NH4 )2 SO4 in 0.25 M NaCl, pH 7.4
1.67 1.85
1.49 1.68
mobile phase. The pDNA was eluted at 0.85 M (NH4 )2 SO4 and the isoforms were only partially separated (Fig. 3). It is interesting to notice that the concentration of (NH4 )2 SO4 for the pDNA elution was different for two chromatographic modes, thus demonstrating a different binding mechanism of pDNA with the stationary phase. To investigate more in detail the relative contributions of hydrophobic and electrostatic interactions an additional experiment was performed, this time introducing a combination of NaCl in tandem with (NH4 )2 SO4 . A chromatographic method used for enhancing hydrophobic interactions during pDNA loading was used, but in both buffers NaCl concentration was maintained at 0.25 M. Elution of both plasmids occurred at higher (NH4 )2 SO4 concentration (the difference is app. 0.2 M) compared to the experiment where mobile phase contained no NaCl (Table 1). This leads to the conclusion that pDNA interaction with the HISA is partially of ionic character. If this was not the case, the elution of pDNA in the presence of NaCl would be expected to occur at lower (NH4 )2 SO4 concentration due to an additional kosmotropic effect provided by NaCl. To explore even further the ion-exchange characteristics of pDNA loaded on a HISA monolith a NaCl gradient at different pH values, namely 5.0, 6.2 and 7.4, was tested. A partial separation between sc and oc isoforms of the plasmid DNA was observed regardless of the pH used, but a higher NaCl concentration was needed to elute pDNA at a lower pH (Table 2). This can be explained by the titration state of imidazole. Although an approximate value of HISA isoelectric point is 7.0, at pH 7.4 there are still around 30% of positively charged groups. In contrast, at pH 5 the ratio between neutral and positively charged ligand is 1:100, so 99% of ligand molecules are charged. The described behavior results in two important conclusions. Operating the ligand below pH 6.0 allows an effective plasmid DNA binding to occur even when the NaCl concentration is increased up to 1.0 M. The reason for this is the ion-exchange binding due to a completely charged surface. Additionally, changing a charge of the monolith surface enables the application of a pH gradient from neutral to slightly alkaline values to elute the electrostatically bound biomolecules at relatively mild conditions. 3.4. Dynamic binding capacities (DBC) for pure plasmid samples To investigate the applicability of the novel matrix for preparative pDNA purification a dynamic binding capacity (DBC) for pure plasmid at different conditions was carried out. The flow rate on HISA disks was fixed at 11 CV/min throughout the experimental work, because it was demonstrated that it had little effect on DBC [5]. Four different experiments were performed at two different pH values. The results are presented in Table 2. Two experiments were
done in the absence of NaCl, but at two different pH values of 5.0 and 6.2. The DBC was around 4 mg/ml regardless the buffer composition or the pH value. Additionally, we performed an experiment in the presence of high NaCl concentration in the loading buffer. 0.75 M NaCl was added at pH 6.2, while 1.0 M NaCl at pH 5.0. As expected, the DBC (between 2.5 and 3.0 mg/ml) was lower than in the absence of salt (4.0 mg/ml), but in comparison to the conventional weak anion exchanger, such as DEA, which loses the capacity for the plasmid at 1.0 M NaCl completely, the use of HISA ligand is advantageous [29]. Binding at such high NaCl concentration is important to capture pDNA from clarified lysate at high conductivity in presence of proteins and RNA, which are not retained by the stationary phase. The DBC was also measured in the case, where the plasmid was loaded in 3.0 M (NH4 )2 SO4 buffer, pH 7.4. The DBC at 10% breakthrough was 2.7 mg of pDNA/ml of support, which is comparable to a DBC on conventional CIM C4 monoliths, measured at the same conditions (internal results).
3.5. Separation of pDNA isoforms in descending (NH4 )2 SO4 gradient Based on the preliminary results, an ascending NaCl gradient and the use of acidic buffers were selected in a capture step for pDNA from the lysate (removing RNA, proteins). On the contrary, plasmid isoforms separation is more effective in the descending (NH4 )2 SO4 gradient and it would therefore preferably be used in a polishing step. The important consideration is the separation of the isoforms at a preparative loading in descending (NH4 )2 SO4 gradient. In the first experiment, 2.64 mg of pure pDNA, dissolved in 3.0 M (NH4 )2 SO4 , pH 7.4, was loaded (the start of breakthrough) followed by the elution of the pDNA in a linear gradient to 0 M of (NH4 )2 SO4 . Only one broad peak was obtained (data not shown), where both oc and sc isoforms were overlapping. The agarose gel electrophoresis analysis of fractions and chromatographic pDNA analytics of the elution confirmed pure oc isoform in the first fractions, followed by fractions of mixtures of both isoforms and finally by pure sc isoform. A step gradient was applied instead of a linear one and its optimization enabled the separation of pDNA isoforms to a high purity. The optimal (NH4 )2 SO4 concentration for oc wash was determined to be 2.3 M and the results are shown in Table 3. The ratio between sc and oc isoforms was increased from 5.6 in the load to 70 in the main sc elution fraction. A complete separation of both isoforms was not achievable. The homogeneity of the sc isoform was 98.5%, with 62% of the loaded sc pDNA eluting in this fraction. The overall recovery of the plasmid DNA (all elution fractions divided by the load) for both isoforms is higher than 90%.
Table 2 Concentrations of NaCl for pDNA elution in buffers with different pH values in ascending salt gradient; CIM HISA disk, pJ pDNA (0.045 mg/ml) loading in appropriate buffer, = 260 nm, ˚ = 2.0 ml/min; linear gradient (0.013 M NaCl per CV). Buffer A composition
Buffer B composition
pH of mobile phase
c (NaCl) for sc pDNA elution [M]
DBC at 10% breakthrough for pDNA (mg plasmid/ml of support)
50 mM acetate 50 mM phosphate 50 mM TRIS
50 mM acetate + 3.0 M NaCl 50 mM phosphate + 2.0 M NaCl 50 mM TRIS + 1.0 M NaCl
5.0 6.2 7.4
2.4 1.55 0.55
4.2 (in absence of NaCl) 2.9 (in presence of 1.0 M NaCl) 4.0 (in absence of NaCl)2.6 (in presence of 0.75 M NaCl) Not measured
ˇ U. Cernigoj et al. / J. Chromatogr. A 1281 (2013) 87–93
91
Table 3 Step-gradient separation of both pDNA isoforms in descending (NH4 )2 SO4 gradient after preparative loading in 3.0 M (NH4 )2 SO4 , pH 7.4: 1 ml CIM HISA column, ˚ = 4.0 ml/min, pDNA (0.06 mg/ml). Step
Buffer
V [ml]
oc amount (mg)
% of oc
sc amount (mg)
% of sc
Load (0.06 pDNA mg/ml) Flow through fraction Washing (removal of oc isoform) Elution of sc isoform Washing with 2 M NaCl Regeneration Overall recovery
3.0 M (NH4 )2 SO4 , 50 mM TRIS, 10 mM EDTA 3.0 M (NH4 )2 SO4 , 50 mM TRIS, 10 mM EDTA 2.3 M (NH4 )2 SO4 , 50 mM TRIS, 10 mM EDTA 50 mM TRIS, 10 mM EDTA 50 mM TRIS, 10 mM EDTA, 2.0 M NaCl 0.5 M NaOH
44 44 9.3 11.4 8 4.5
0.4 0 0.35 0.02 0.01 0 0.38
100 0 87.3 5.6 1.4 0 94.3
2.24 0.02 0.57 1.40 0.05 0.01 2.05
100 1.0 25.6 62.6 1.9 0.3 91.3
3.6. Separation of pDNA isoforms in a pH gradient It was found that the oc isoform of the plasmid can be selectively eluted at constant 3.0 M (NH4 )2 SO4 , while changing the pH of the mobile phase. A representative chromatogram is shown in Fig. 4. Plasmid DNA was completely retained by the monolith and the only eluted compound was an oc pDNA isoform at the end of the pH gradient. In this way, the oc form can be selectively removed from the sc form. To confirm this result and to check the efficiency of the oc removal a following experiment was performed. After
implementation of the pH gradient where the oc form was selectively removed, the column was reequilibrated to initial conditions (3.0 M (NH4 )2 SO4 , 50 mM TRIS, 10 mM EDTA, pH 7.4) and a linear gradient to 0 M (NH4 )2 SO4 , at a constant pH value of 7.4 was carried out. At a retention time, where the oc peak should occur, there was almost no detectable peak, while the sc peak was eluted normally (see Fig. 4B). This served as a proof that the oc form was efficiently removed. Additionally, as a control experiment, the same sample without a pre-step (pH gradient) was separated in the (NH4 )2 SO4 gradient from 3.0 to 0 M, at a constant pH of 7.4. In this experiment a well resolved oc peak appeared, while the sc peak had the same area and retention time as in case of a pretreated sample. 3.7. Applying pH gradient in preparative pDNA isoform separation
Fig. 4. Elution of oc isoform of the pJ in pH gradient. Monolith: HISA disk, ˚ = 2.0 ml/min, = 260 nm, Vinj = 500 l, pDNA (0.045 mg/ml) in 3.0 M (NH4 )2 SO4 , pH 7.4. (A) Elution of oc pDNA in pH gradient: MF A: 50 mM TRIS, 3.0 M (NH4 )2 SO4 , 10 mM EDTA, pH = 7.4; MF B: 3.0 M (NH4 )2 SO4 , 50 mM acetate, 10 mM EDTA, pH = 4.5, gradient from 0% to 100% MF B in 50 CV; (B) a proof for selective removal of oc pDNA isoform from the mixture: gradient from 3.0 to 0 M (NH4 )2 SO4 in 50 mM TRIS, 10 mM EDTA, pH 7.4; ( ) initial sample; ( ) sample firstly pretreated according to procedure in (A), then elution in descending (NH4 )2 SO4 gradient.
Based on the results previously obtained the plasmid DNA sample was loaded under conditions, where oc isoform was not retained by the column. In this case, all the oc was supposed to be in the flow-through fraction, while the pure sc isoform would bind to the chromatographic support. When a pDNA isoform mixture dissolved in 2.9 M (NH4 )2 SO4 , 50 mM acetate, 10 mM EDTA, pH 5, was loaded on the column, the DBC for oc isoform was around 0.1 mg/ml of support. After the oc breakthrough a plateau was reached, where only sc isoform was retained. The calculated DBC for the sc isoform was between 0.5 and 1 mg/ml of HISA chromatographic support, depending on the small variations in (NH4 )2 SO4 concentration. We decided to optimize the novel method in a way that the plasmid was loaded at high (NH4 )2 SO4 concentration, pH 7.4 (preferential hydrophobic interactions) with the following washing step of oc isoform at high (NH4 )2 SO4 concentration, but in a slightly acidic buffer. Three main parameters were optimized during this study: the amount of loaded plasmid, the concentration of (NH4 )2 SO4 in the washing buffer and the pH of the washing buffer. Changing a pH in the washing buffer between 4.0 and 5.5 did not influence the elution pattern. The titration of HISA is completed below pH 6 and the surface is almost fully charged with protonated HISAs. At the same time, the plasmid is almost fully negatively charged at pH above 3, therefore the robust conditions for pDNA elution were achieved between pH 4.0 and 5.5. Optimization of the amount of loaded plasmid followed. The DBC for sc isoform is app. 0.7 mg of plasmid/ml of the bed volume in 3.0 M (NH4 )2 SO4 , pH 4.5. The only possibility to increase the capacity of the column at high (NH4 )2 SO4 concentration and slightly acidic pH was to increase the (NH4 )2 SO4 concentration to achieve higher hydrophobicity of the mobile phase, where more plasmid could be retained on the column. After optimization of the parameters the following conditions were obtained: if app. 1.6 mg of pDNA per ml of chromatographic bed volume was loaded in 3.0 M (NH4 )2 SO4 , pH 7.4, more than 90% of oc isoform together with app. 10% of sc isoform was washed with 3.2 M (NH4 )2 SO4 , pH 5. Another challenge was how to perform an efficient elution of sc pDNA (to achieve as high as possible recovery and at the same time as high as possible concentration of the eluted
ˇ U. Cernigoj et al. / J. Chromatogr. A 1281 (2013) 87–93
92
Table 4 Step-gradient separation of pDNA isoforms in pH gradient at high (NH4 )2 SO4 concentration after preparative loading in 3.0 M (NH4 )2 SO4 , pH 7.4. The experimental parameters are described in caption of Fig. 5.
Load (1.58 mg of pDNA/ml of bed volume) Washing step (3.2 M (NH4 )2 SO4 , pH 5.0) sc elution step (0.6 M (NH4 )2 SO4 , pH 7.4) Recovery (%)
oc [mg]
% of oc compared to the oc amount in load
sc [mg]
% of sc compared to the sc amount in load
sc + oc [mg]
sc:oc
(sc) [mg/ml]
0.23 0.20 0.02 96
100 87 9
1.33 0.03 1.29 99
100 2 97
1.58 0.23 1.31 98
5.7 0.2 64
0.053 0.003 0.156
plasmid) after the washing of oc isoform. After optimization the only well-working procedure was to increase the pH back to 7.4, while keeping the (NH4 )2 SO4 concentration at 3.2 M. In this way, the sc pDNA was retained by the column. Then the elution was carried out by descending (NH4 )2 SO4 gradient at the same pH. The whole process (Fig. 5A) was analyzed by pDNA chromatographic analysis (Fig. 5B) and quantified (Table 4). Fig. 5A confirms the absence of pDNA elutions in FT and washing steps as well as in buffer exchange steps. Fig. 5B is a direct proof of purity of both fractions containing eluted pDNA isoforms while the quantification
results are gathered in Table 4. There is 97% recovery of the sc pDNA in the main elution fraction. Only 2% of the sc isoform was lost in an oc washing fraction. The sc:oc ratio was improved from 5.7 in the load to 64 in the main elution fraction. The plasmid concentration was increased approximately 3 times from load to the main elution fraction. 4. Conclusions HISA was immobilized to a monolithic chromatographic support and studied as a multimodal ligand for pDNA purification. In ascending NaCl gradient (preferentially electrostatic interactions) an elution pattern of pDNA was studied in different buffers with the pH value of the buffer as the main changing parameter. Due to the appropriate pKa value of HISA ligand a NaCl concentration for pDNA elution increases considerably from pH 7.4 to pH 5.0. Operating the ligand below pH 6 allows effective plasmid binding to occur at NaCl concentrations higher than 1.0 M, enabling the preparative loading of plasmid at high conductivities and at a slightly acidic pH of the loading buffer. The column is capable of separating efficiently the pDNA isoforms in the descending (NH4 )2 SO4 gradient (preferentially hydrophobic interactions) as well. Additionally, novel pH gradient conditions were proven to selectively elute both plasmid isoforms. Instead of applying (NH4 )2 SO4 descending gradient, a pH gradient at constant 3.0 M (NH4 )2 SO4 was used with the similar efficiency. After optimizaton of the chromatographic parameters we were able to purify and concentrate 1.3 mg of sc pDNA per ml of column bed volume with recovery higher than 95% and less than 2% oc pDNA impurities. The main goal of the subsequent research will be the development of a chromatographic pDNA purification process from the clarified lysate to the final sc pDNA fraction using HISA monolithic column, where the emphasis will be focused to the efficient removal of other process and product related impurities besides the oc isoform. Acknowledgements
Fig. 5. The separation of oc and sc pDNA isoforms on preparative scale employing the novel chromatographic conditions. 1 ml CIM HISA column, ˚ = 4.0 ml/min. Loading: pDNA (0.06 mg/ml) in 3.0 M (NH4 )2 SO4 , 50 mM TRIS, 10 mM EDTA, pH 7.4; washing step: 3.2 M (NH4 )2 SO4 , 50 mM acetate, 10 mM EDTA, pH 5.0; buffer exchange step: 3.2 M (NH4 )2 SO4 , 50 mM TRIS, 10 mM EDTA, pH 7.4; sc elution step: 0.6 M (NH4 )2 SO4 , 50 mM TRIS, 10 mM EDTA, pH 7.4. (A) UV signal at 260 nm during pDNA isoforms separation process; (B) analysis of some fractions using CIMac pDNA column.
We are very grateful to Dr. Fani Sousa and Dr. Angela Sousa from Centro de Investigac¸ão em Ciências da Saúde, Universidade da Beira Interior, for the fruitful discussions on pDNA pseudoaffinity ligands behavior and for the help in immobilization of amino acids and their derivates to monolithic chromatographic supports. We are greatly thankful to Dr. Gerald Striedner and Dr. Markus Luchner, both from the Microbial Fermentation Group at Department of Biotechnology, BOKU, Vienna, for providing us a pDNA sample. We would also like to thank Boˇstjan Gabor and Dr. Nika Lendero Krajnc from BIA Separations for giving us practical advice from their pDNA purification knowledge. Tina Jakop, Sandra Kontrec and Jernej Gaˇsperˇsiˇc from BIA Separations and Rok Oblak are acknowledged for their help in laboratory work. This work was partially financially supported by the Ministry of Education, Culture, Science and Sport of the Republic of Slovenia within the research programme P4-0469. The Centre of Excellence for Biosensors, Instrumentation and Process Control is an operation financed by the European Union, European Regional Development Fund and
ˇ U. Cernigoj et al. / J. Chromatogr. A 1281 (2013) 87–93
Republic of Slovenia, Ministry of Education, Science, Culture and Sport. References [1] S. Gurunathan, D.M. Klinman, R.A. Seder, Annu. Rev. Immunol. 18 (2000) 927. [2] I. Ruzila, N.A. Zeenathul, M.L. Mohd-Azmi, Vaccine 30 (2012) 5914. [3] C.A. Holladay, T. O’Brien, A. Pandit, Wiley Interdiscip. Rev.: Nanomed. Nanobiotechnol. 2 (2010) 721. [4] A. Ghanem, R. Healey, F.G. Adly, Anal. Chim. Acta 760 (2013) 1. ˇ A. Podgornik, J. Chromatogr. A 1218 [5] N. Lendero Krajnc, F. Smrekar, A. Strancar, (2011) 2413. [6] P.-O. Wahlund, P.-E. Gustavsson, V.A. Izumrudov, P.-O. Larsson, I.Yu. Galaev, Biotechnol. Bioeng. 87 (2004) 675. [7] G.A. Gomes, A.M. Azevedo, M.R. Aires-Barros, D.M.F. Prazeres, Sep. Purif. Technol. 65 (2009) 22. [8] D.W. Kahn, M.D. Butler, D.L. Cohen, M. Gordon, J.W. Kahn, M.E. Winkler, Biotechnol. Bioeng. 69 (2000) 101. [9] J. Stadler, R. Lemmens, T. Nyhammar, J. Gene Med. 6 (2004) S54. [10] G.N.M. Ferreira, J.M.S. Cabral, D.M.F. Prazeres, Biotechnol. Tech. 11 (1997) 417. [11] C.M. Ongkudon, M.K. Danquah, J. Chromatogr. B 878 (2010) 2719. [12] L.R. Pereira, D.M.F. Prazeres, M. Mateus, J. Sep. Sci. 33 (2010) 1175. [13] M.J. Shin, L. Tan, M.H. Jeong, J.-H. Kim, W.-S. Choe, J. Chromatogr. A 1218 (2011) 5273.
93
[14] F. Sousa, D.M.F. Prazeres, J.A. Queiroz, Trends Biotechnol. 26 (2008) 518. [15] R.A. Darby, G.M. Forde, N.K. Slater, A.V. Hine, Biotechnol. Bioeng. 98 (2007) 1103. [16] T. Schluep, C.L. Cooney, Nucleic Acids Res. 26 (1998) 4524. [17] Y. Han, G.M. Forde, J. Chromatogr. B 874 (2008) 21. [18] L.M. Sandberg, A. Bjurling, P. Busson, J. Vasi, R. Lemmens, J. Biotechnol. 109 (2004) 193. [19] C. Caramelo-Nunes, M.F. Gabriel, P. Almeida, J.C. Marcos, C.T. Tomaz, J. Chromatogr. B 904 (2012) 81. [20] C. Caramelo-Nunes, T. Tente, P. Almeida, J.C. Marcos, C.T. Tomaz, Anal. Biochem. 412 (2011) 153. [21] F. Sousa, C.T. Tomaz, D.M.F. Prazeres, J.A. Queiroz, Anal. Biochem. 343 (2005) 183. [22] F. Sousa, T. Matos, D.M.F. Prazeres, J.A. Queiroz, Anal. Biochem. 374 (2008) 432. ˇ A. Podgornik, in: B. [23] N. Lendero Krajnc, F. Smrekar, V. Frankoviˇc, A. Strancar, Mattiasson, A. Kumar, I. Galaev, I. Yu (Eds.), Monolithic Macroporous Polymers as Chromatographic Matrices, CRC Press, Florida, 2010, p. 291. [24] J. Urthaler, R. Schlegl, A. Podgornik, A. Strancar, A. Jungbauer, R. Necina, J. Chromatogr. A 1065 (2005) 93. ˇ [25] F. Smrekar, A. Podgornik, M. Ciringer, S. Kontrec, P. Raspor, A. Strancar, M. Peterka, Vaccine 28 (2010) 2039. [26] A. Sousa, D. Bicho, C.T. Tomaz, F. Sousa, J.A. Queiroz, J. Chromatogr. A 1218 (2011) 1701. [27] H. Walba, R.W. Isensee, J. Org. Chem. 26 (1961) 2789. ˇ M. Benˇcina, J. Sep. Sci. 27 (2004) 811. [28] K. Benˇcina, A. Podgornik, A. Strancar, [29] C.M. Ongkudon, M.K. Danquah, Sep. Purif. Technol. 78 (2011) 303.