3.2. 2D-AEC×SEC experimental operation conditions
Salt concentration, flow rate, pH, and sample volume are considered as the most important parameters which should be optimized for developing a proficient 2D-AEC×SEC methodology. Regarding the self-dissociation feature and highest immunity of FMDV particles reported at pH below 5.6 and also partially on neutral pH (Liang et al., 2014), the mean pH = 7.3 was selected at all of the optimization steps for the initial and elution buffers in both column dimensions.
As it was already mentioned, co-elution behavior of virus particles and BSA indicated that the p I values are very close together, and accordingly the selection of pH =7.3 about three units higher than of the p I = 4.8 for BSA can guarantee absorption active ingredient of vaccine to ion-exchange media.
Regarding to theoretical concepts of IEX approach, sample should be loaded with the minimum ionic strength on the ion exchange media in order to prevent from competition between the ions in the medium and the target protein for adsorption on the ion exchanger. Therefore, 1.2 mS/cm, as the electrical conductivity of the initial buffer (Tri-HCl 20 mM with pH = 7.3) was considered as the reference in order to set the conductivity of primary feed stock. Based on an experimental study, a salt adjustment for the primary sample could provide a conductivity equivalent to 1.2 mS/cm as the initial buffer in AEC. In the elution step, potassium chloride (KCl) was used to adjust the ionic strength of the applied elution buffer. Accordingly, a buffer of Tris-HCl (20 mM with p pH = 7.3) containing 500 mM KCl was used in a stepwise gradient mode for elution absorbed proteins with suitable conductivity.
Considering the Q-Sepharose base media specifications, the mean flow rates of 3 mL/min and 8 mL/min were selected for loading and eluting steps the sample, respectively. In general, loading 20–30% of the total binding capacity of the column should provide the optimal resolution with gradient elution mode in IEC (Bailly & Tondeur, 1981). The total protein concentration of 0.473 mg/mL in the feed stock was determined by Bradford assay and performance of Q-Sepharose matrix was evaluated for the sample volume in the range of 50-100 mL of the primary sample in combination with Sup-200 as the final polishing step. Then, the total fractions obtained from capturing step were analyzed by HP-SEC as our last work (Bagheri et al., 2018), and virus infectivity assay was determined by Reed and Muench’s method (Frenz & Horváth, 1985).Fig. 2 represents the elution pattern of primary sample for minimum sample loading on Q-Sepharose media and Sup-200. After analyzing the obtained fractions from the first dimension, the highest amount of FMDV particles was washed out from the column between 120-180 minutes for the sample loading which was two times more than the maximum loading in the previous study (Bagheri et al., 2018). Therefore, pooled virus fraction was concentrated by the concentration factor 10, and then 4 mL of this concentrated sample was loaded on the Sup-200. The comparison of elution pattern for 2D-AEC×SEC and 2D-SEC×SEC indicates a completely same elution pattern for both methods in the second dimension. Since the minimum sample volume loaded on the first dimension (Vinj = 50 mL) is about two times more than the maximum loading capacity in SEC for 2D-SEC×SEC, it seems that AEC is considered as a better candidate for capturing step to fulfill speed and capacity as the two most important criteria in this step. Based on the obtained data, the overall purification time in the first dimension reduced by a factor of 0.7 in 2D-AEC×SEC than 2D-SEC×SEC. In order to ensure the efficiency of anion exchange media in the proposed three-step strategy, the volume of the sample loaded on the resin increased in different experiments to maximum 100 mL (equal to four times of the related amount in 2D-SEC×SEC (Bagheri et al., 2018).
Fig. 3 (A ) and (B ) represent the experimental data, as well as the simulation results of the elution pattern in the primary feed stock for maximum sample loading on a combination of Q-Sepharose and Sup-200 media. In the final polishing step for Sup-200, as the second dimension, the volumetric flow rate decreased to 2.5 mL/min for obtaining a better resolution between virus particles and NSPs marker, BSA, due to the higher concentration of the concentrated sample in an intermediate concentration step. The simulation results for retention behavior of BSA, virus, rDNA and IgG components in the first and second dimension were also depicted in Fig. 3 (A ) and (B ). In AEC the coupled set of Equations. (1), (4), (5) and (6) incorporating plate model (Eq. 2, considering convection with dispersion based on plate number), was discretized and solved using finite element and orthogonal collocation techniques. It is worth mention that the number of theoretical plates was estimated experimentally from a pulse injection of acetone and following the UV chromatogram for determination peak width at half-height and retention time. Accordingly, the height of theoretical plate (HETP) was calculated as 0.01177 cm. In order to assess the effect of controlling diffusion versus convection mechanisms,Pe number, as an important parameter affecting the performance of separation process, was considered and estimated by Eq. (3). Therefore, for the total height of the bed\(H_{B}=40\ \text{cm}\), internal diameter of the bed column\(\text{ID}\ =\ 1.6\ \text{cm}\) and particle radius of adsorbent\(\text{\ Rp}\ =\ 45-165\ \text{μm}\), the optimum amount for flow rate (or Pe number) and eluents concentration for FMDV particles elution from Q-Sepharose media were studied by mathematical modeling to explore the most appropriate condition in AEC column. Acquire data demonstrated that a flow rate equal to 8 mL/min and a combination of 73% of buffer A (buffer without salt) and 27% of buffer B (buffer with 0.5 M salt concentration) can be applied as the most suitable operational condition to elute virus particles from Q-Sepharose in the first dimension. Accordingly, the concentration of KCl at the time of departure of the virus peak from AEC column was predicted equal to 0.129 M. This is nearly close to experimental data as indicated 0.135 M KCl concentration for FMDV elution from AEC column. Moreover, as illustrated in Fig. 3 (A ), simulation results in the first dimension demonstrated that peaks for virus, BSA and rDNA have interfering areas in the output of the AEC, which is satisfactory matched with experimental data. Same results for experimental and simulated investigations can be explained by very closepI values for these components at natural pH. Similarly, mathematical modeling and simulation process was also applied for total bed height equal to\(\ H_{B}=60\ \text{cm}\), internal diameter of the bed column. \(\text{ID}\ =\ 2.6\ \text{cm}\), particle radius of adsorbent \(\text{Rp}\ =\ 24-44\ \text{μm}\) in the second dimension for prediction and obtaining optimal flow rate. Eventually a flow rate equal to 2.5 mL/min was obtained for optimized operation condition and achieved data in Fig. 3 (B ) showed that the simulation results are well-matched with the experimental counterparts. Therefore, this model can be applied to scrutinize the scaling-up issue in term of geometry (enlarging the cross-section and length of the columns), flow rate and sample loading capacity in the columns.