618
J. Sep. Sci. 2014, 37, 618–624
Muhammad Aasim1,2∗ Prasad Babu Kakarla1∗ Roy N. D’Souza1 Noor Shad Bibi1 Tanja Yvonne Klein3 Laura Treccani3 Kurosch Rezwan3 1 ´ Marcelo Fernandez-Lahore
Research Article
1 Downstream
Protein adsorption onto hydrophobic chromatographic supports has been investigated using a colloid theory surface energetics approach. The surface properties of commercially available chromatographic beads, Toyopearl Phenyl 650-C, and Toyopearl Butyl 650-C, have been experimentally determined by contact angle and zeta potential measurements. The adsorption characteristics of these beads, which bear the same backbone matrix but harbor different ligands, have been studied toward selected model proteins, in the hydrated as well as dehydrated state. There were two prominent groups of proteins observed with respect to the chromatographic supports presented in this work: loosely retained proteins, which were expected to have lower average interaction energies, and the strongly retained proteins, which were expected to have higher average interaction energies. Results were also compared and contrasted with calculations derived from adsorbent surface energies determined by inverse liquid chromatography. These results showed a good qualitative agreement, and the interaction energy minima obtained from these extended Derjaguin, Landau, Verwey and Overbeek calculations were shown to correlate well with the experimentally determined adsorption behavior of each protein.
Bioprocessing Laboratory, School of Engineering and Science, Jacobs University, Campus Ring 1, Bremen, Germany 2 Department of Biotechnology, University of Malakand, Chakdara, Dir, Khyber Pakhtunkhwa, Pakistan 3 Keramische Werkstoffe und Bauteile/Advanced Ceramics, University of Bremen, Am Biologischen Garten 2, Bremen, Germany
Received December 16, 2013 Revised January 9, 2014 Accepted January 9, 2014
The role of ligands on protein retention in adsorption chromatography: A surface energetics approach
Keywords: Adsorption chromatography / Hydrophobic interactions / Protein adsorption / Surface energetics DOI 10.1002/jssc.201301338
Additional supporting information may be found in the online version of this article at the publisher’s web-site
1 Introduction Hydrophobic interaction chromatography (HIC) has widely been employed in the separation of proteins and peptides with high resolution, high yield, and in nondenaturing environments, on laboratory and industrial scales [1, 2]. It has been presumed that almost 50% of the surface of soluble proteins contains hydrophobic residues, which can be exploited by HIC [3]. HIC was initially reported in 1949 by Shepard and Tiselius as salting-out chromatography [4], and takes advantage of intrinsic protein hydrophobicity to achieve separation, where the hydrophobic patches on the surface of the protein interact with hydrophobic ligands on the chromatographic matrix [5]. These seemingly weak interactions are important in sustaining the native structure and activity ´ Correspondence: Professor Marcelo Fernandez-Lahore, Downstream BioProcessing Laboratory, Jacobs University, Campus Ring 1, D-28759 Bremen, Germany E-mail: [email protected] Fax: +49 421 200 3600
Abbreviations: AB, Acid–base; HIC, hydrophobic interaction chromatography; IgG, total human Immunoglobulin G; ILC, inverse liquid chromatography; LW, Lifshitz–Van der Waals; Lys, lysozyme; -Gal, -galactosidase C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
of the proteins [3]. The proteins are retained at high salt concentration and eluted at decreasing salt concentration, with the least hydrophobic protein eluting first and the most hydrophobic protein released at the end [6]. Several concepts have been developed to explain the interaction process during HIC. Solvophobic theory proposes a linear relationship between protein adsorption and salt concentration in the mobile phase [6]. Therein, the interaction strength was attributed the molar surface tension increment due to the kosmotropic salt concentration [6]. In contrast, the preferential interaction model excludes salt ions surrounding the ligand and proteins and is based on the protein’s ability to either interact with another protein or with the ligand matrix [7]. Furthermore, the dependence of protein retention on the protein and adsorbent properties [8], salt concentration [9], temperature [10, 11], pH of the buffer [12, 13], ligand type [12, 14], as well as the density of the ligand immobilized on the base matrix [2, 12] has been widely investigated. Microcalorimetry has also been employed to probe protein adsorption thermodynamics [15, 16], and association equilibrium constants derived from Langmuir adsorption isotherms [8] can serve a good tool to calculate the association interaction ∗ These authors have contributed equally to this work. Colour Online: See the article online to view Figs. 2, 4 and 5 in colour.
www.jss-journal.com
Liquid Chromatography
J. Sep. Sci. 2014, 37, 618–624
energies of proteins to hydrophobic supports [17, 18]. The retention of the protein was also predicted by the determination of “dimensionless retention times (DRT)” by calculating the surface hydrophobicity of the proteins [19]. Predictive quantitative structure–retention relationship (QSRR) models have also been previously employed [20, 21], and in one case, the primary sequence of the protein was shown to predict retention behavior [22]. Noncovalent van der Waals forces have been considered to be the major contributing factor to protein adsorption during HIC [23, 24]. These interaction forces are considered as nonspecific weak interactions [25, 26], which are important to retain protein structure and function. Despite sophisticated laboratory equipment and chromatographic media, difficulties still persist in finding optimum purification conditions [27]. This study investigates the applicability of interfacial chemistry techniques as a diagnostic tool to understand protein interactions at the nanoscale. The underlying assumption of proteins being nano-colloidal particles is instrumental to apply extended Derjaguin, Landau, Verwey and Overbeek (DLVO) theory [26, 28–30], and subsequently investigate the interaction of these proteins to chromatographic beads via surface thermodynamic calculations. According to extended DLVO theory, the interaction energy between a colloidal particle (protein) and a solid support (chromatographic bead) is the cumulative sum of Lifshitz–van der Waals (LW), acid– base (AB), and electrostatic (EL) energies [31] as shown in Eq. (1).
U XDLVO = U LW + U EL + U AB
(1)
The energy components shown in Eq. (1) are derived from contact angle measurements (LW and AB) and zeta potential (EL) determinations. Details of the calculations of different energy components are given in our previous paper [38]. The surface thermodynamics approach has already been used for investigating the interaction of colloidal particles to a broad range of solid substrata in environmental, biomedical, and bioprocesses [32,33]. It has also helped study the interaction of various types of biomass with expanded bed chromatographic beads in several modes of operation [34–37]. Protein interactions to certain chromatographic supports have also been studied, where early or late eluting proteins could only be marginally discriminated [28, 29, 38]. The aim of this work was to investigate the fundamental effect of ligand chemistry on protein retention and test the validity of extended DLVO (xDLVO) calculations toward predicting protein–adsorbent interactions. The basic parameters for the calculations were assessed by contact angle measurements and zeta potential determinations by electrophoretic mobility and streaming potentials. These results were compared with those obtained using parameters assessed by inverse liquid chromatography (ILC) [39]. Details of the xDLVO calculations are provided in the Supporting Information. C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
619
2 Materials and methods 2.1 Chemicals and reagents Toyopearl Phenyl 650-C and Toyopearl Butyl 650-C were purchased from TOSOH Bioscience (Stuttgart, Germany). 1-Bromonaphthalene and formamide were purchased from Fluka (Buchs, Switzerland). Ultrapure water was used for contact angle determinations. All other chemicals were of analytical grade. Hen’s egg white lysozyme (HEWL) and BSA were obtained from Sigma-Aldrich (Munich, Germany). PolyR 3XL) and galacturonase from Aspergillus niger (Pectinex R ) were Aminopeptidase from Aspergillus oryzae (Flavorzyme purchased from Novozymes (Lund, Sweden). Chymosin from R ), Aspartic proAspergillus niger var. awamori (CHY-MAX R L 205), and tease from Rhizomucor miehei (Hannilase -galactosidase from Kluyveromyces lactis (HA-Lactase 5200) were from CHR. HANSEN (Holdorf, Germany). Total huR ) was from Ocman mmunoglobulin G (hT-IgG/Octagam tapharma (Langenfeld, Germany). Lactoferrin was obtained from Biopole (Namur, Belgium). All the protein solutions were extensively dialyzed against water to remove salts and preservatives and stored at 4⬚C as lyophilized powders until further use.
2.2 Contact angle measurements Protein layers were deposited on microscopic glass slides according to a previously reported method [26]. Proteins were also deposited onto a plastic Petri dish support, as previously described [28]. For generating uniform protein-coated surfaces, protein solutions (2 mg/mL in 20 mM phosphate buffer at pH 7) were deposited on a glass slide or Petri dish and dried overnight at room temperature in a dust-free environment. Excess buffer was carefully removed with adsorbent paper and the protein-coated surface was then dried under vacuum. Proteins were additionally deposited on an ultrafiltration (cellulose) membrane (10 kDa MWCO; Millipore) to measure the contact angles in the hydrated state [26]. Subsequently, the membrane was removed and kept at room temperature for 20 min before measurements. Contact angles were measured using a goniometer (OCA20, DataPhysics Instruments, Germany) according to the sessile drop technique [34], employing water, formamide, and 1-bromonaphthalene as probe liquids to determine surface characteristics [40, 41]. Small drops of the probe liquids (2–5 L) were dispensed on the samples. Contact angle measurements were performed six times for each probe liquid at room temperature. The measured contact angle values were used for surface energy calculations, according to the van Oss AB theory [31], using the instrument’s software (SCA 20). All contact angle measurements were also performed on the homogenous base surfaces (agar plates or glass/plastic/membrane surfaces). www.jss-journal.com
620
J. Sep. Sci. 2014, 37, 618–624
M. Aasim et al.
raphy data were stored and analyzed using a commercial software package (UNICORNTM 5).
2.3 Zeta potential determinations Zeta potential values were measured with a Zetasizer Nano ZS from Malvern Instruments (Worcestershire, UK) and all measurements were performed in triplicate. For protein samples, measurements were carried out using 2 mg/mL solutions prepared in a 20 mM phosphate buffer (pH 7). Electrophoretic mobility data were used to calculate zeta potentials within the instrument software using Henry’s equation [31].
2.4 Streaming potential measurements The charge characteristics of intact chromatographic beads were determined by streaming potential measurements (SurPASS EKA, Anton Paar, Austria). These electrokinetic measurements were converted to zeta potentials using Eq. (2) [42] at different pH conditions, where is the viscosity, is the specific electric conductivity of the electrode solution, g is the dielectric constant of the electrolyte, and dU/dP represents the variation of the streaming potential with applied pressure. =
dU · g0g d P
(2)
A 1 mM KCl solution was used as an electrolyte for the determination of streaming potentials. A 2–3 mm layer of the material was packed in the cylindrical cell. The pH value was varied within the range of pH 3–10 by the addition of 0.05 M HCl or 0.05 M NaOH solutions [43]. The values for zeta potentials at high salt concentration were extrapolated from the experimentally measured values, employing known correlations [44, 45].
2.5 Chromatographic experiments Tricon chromatography columns (5 mm internal diameter ¨ and 100 mm length) and the AKTA FPLC system were purchased from GE Healthcare (Munich, Germany). The hydrophobic beads were packed in a commercially available chromatographic column (2.0 mL bed volume; aspect ratio 4.0) and the packing quality was evaluated by residence time distribution analysis employing 1% acetone as a tracer [46]. The mobile phase was composed of two buffer solutions at pH 7: buffer A (binding buffer) with 20 mM phosphate buffer containing 1.7 M ammonium sulfate, and buffer B (elution buffer) with 20 mM phosphate buffer. Buffers were filtered and degassed before use. After equilibration with 10 column volumes (CV) of buffer A, a 200 L protein sample (4 mg/mL in buffer A) was injected into the column. Unbound material was eluted in 5 CV of buffer A and bound protein was eluted using a linear gradient from 0–100% of buffer B (10 CV). Reequilibration was performed with 5 CV of buffer A. The flow rate for all steps was 1 mL/min (300 cm/h). The eluate was monitored with a UV/Vis detector at 280 nm. Chromatog C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
3 Results and discussion 3.1 Contact angle and surface energies The deposited layers of the proteins and chromatographic bead fragments were characterized by contact angle measurements employing three diagnostic liquids: water, formamide and 1-bromonaphthalene [32]. The contact angle values and surface energy components for proteins, in the hydrated and dehydrated states have been published elsewhere [28]. The contact angle values of Toyopearl Phenyl 650-C and Toyopearl Butyl 650-C along with their nonfunctionalized analog are listed in Table 1. The surface free energy components, that is, the Lifshitz–van der Waals (␥LW ) and AB (␥+ and ␥− ) components of the proteins and chromatographic beads are directly calculated from their contact angles using Young’s equation (see Supporting Information). The energy components for the beads determined by contact angle measurements (on fragmented beads) were compared with those previously determined by ILC [39] (while these measurements were performed on intact beads of slightly different particle size, they possess identical surface chemistries). Table 1 shows that most of the chromatographic materials are hydrophilic in nature. However, the backbone of Toyopearl beads is composed of methacrylate polymers that generally have a relatively hydrophobic character. This effect is attributed to the presence of ether and hydroxyl groups on the surface of the modified bead (Toyopearl instruction manual IM02, TOSOH Bioscience). Moreover, there is a slight difference observed between the beads functionalized with different ligands. The material that has phenyl ligands shows a more hydrophilic character than the one that has butyl ligands. The phenyl ligand is generally considered to be more hydrophobic than the butyl ligand [47], but in this specific case, the backbone chemistry of the matrix might play an important role in contributing to the hydrophobicity of the butyl-functionalized matrix. This effect has been verified by interaction energy calculations as well as chromatographic experiments. The hydrophobic LW energy component was higher for the butyl ligand, while the polar AB component was higher for the phenyl ligand. There was only a minor difference between the total surface energy of both materials. It is interesting to note that the surface energies measured via ILC, while maintaining a similar qualitative trend in terms of individual energy components, are nearly 2–4 times larger in magnitude than those measured by contact angles.
3.2 Zeta potential determinations Contact angle measurements fail to account for surface electrostatic effects, which play an important role in the degree www.jss-journal.com
Liquid Chromatography
J. Sep. Sci. 2014, 37, 618–624
621
Table 1. Contact angles (measured by goniometry) and corresponding surface energies of various chromatographic beads in 20 mM phosphate buffer at pH 7.0. Surface energies determined by inverse liquid chromatography have also been listed for comparison [40]
Material
Toyopearl phenyl 650-Ca) Toyopearl phenyl 650-Mc) Toyopearl butyl 650-C Toyopearl butyl 650-Mc) Toyopearl HW 65b)
Surface tension (mJ·m−2 )
Contact angle WR
FM
AB
␥LW
␥+
␥−
␥AB
␥Total
1.5 ± 0.5
2.5 ± 0.8
48.8 ± 1.3
9.9 ± 1.1
7.5 ± 1.0
43.9 ± 2.0
5.4 ± 0.4
10.2 ± 0.4
66.5 ± 0.8
30.5 ± 0.6 124.0 32.9 ± 0.9 144.0 21.7 ± 0.4
3.9 ± 0.2 22.5 3.1 ± 0.3 20.1 7.7 ± 0.2
53.8 ± 0.1 52.3 52.7 ± 0.4 53.8 53.4 ± 0.1
28.9 ± 0.8 68.6 25.4 ± 1.0 65.8 40.4 ± 0.5
59.4 ± 0.2 192.6 58.2 ± 0.1 209.8 62.1 ± 0.1
WR stands for water; FM is formamide; AB is 1-bromonaphthalene. a) and b) Taken from reference [28]. c) Taken from reference [40] and measured by inverse liquid chromatography.
Toyopearl Butyl 650-C. Its zeta potential at pH 7 is around −20.0 mV.
3.3 Extended DLVO interaction energies as a function of distance
Figure 1. Zeta potential of Toyopearl Phenyl 650-C and Toyopearl Butyl 650-C obtained by steaming potential measurements in 1 mM KCl as a function of pH.
of interaction between surfaces. While the measurement of the surface charge of a material is not practically feasible, the electrostatic behavior of a surface can be estimated by its zeta potential, which is the electrostatic potential at the hydrodynamic plane of shear [48]. The zeta potential of a material can be conveniently determined by measuring its electrophoretic mobility [45]. This technique was successfully used to electrostatically characterize the model proteins employed in this study [28] (see Supporting Information). The zeta potential of larger particulates can be determined in varying pH conditions by using streaming potential measurements. The advantage of this method is that it exploits the beads in their intact form. Toyopearl beads have a broad pH stability range, that is, from pH 1 to 13 [43], and their surface charge can be adequately determined within this range. Figure 1 depicts the zeta potential of Toyopearl Phenyl 650-C and Toyopearl Butyl 650-C as a function of pH. It is evident that the phenyl derivative displays a transition from a negative to a positive surface charge (from −15.0 to 1.0 mV) with decreasing pH (from pH 11 to 3). Its zeta potential at pH 7 is around −10.0 mV. The same trend, but a higher surface charge (0 to −25.0 mV from pH 3 to 11) was observed for C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
By combining the experimental parameters determined in the previous sections, the total interaction energy of the proteins and chromatographic beads can be calculated as a function of the distance between them (see Supporting Information). These calculations are performed with well-defined equations, using sphere-to-plate geometry. The interaction energy components and the zeta potential values at high salts have been incorporated into the calculations. Our calculations have considered hydrated as well as dehydrated protein states. For example, Fig. 2 shows the interfacial energy of interaction (U) of human immunoglobulin G (IgG) and Lysozyme (Lys) with Toyopearl Butyl 650-C, as a function of distance (H). The average secondary energy minimum of IgG on three supporting surfaces on Toyopearl Butyl (−0.17 kT, where k is the Boltzmann constant and T is the absolute temperature) is comparatively much higher than that of Lys (−0.02 kT). Similar effects are observed for all model proteins, where interaction energy minima were higher in the case of adsorbents harboring butyl ligands. This very same effect was verified in actual chromatographic experiments (see Section 3.4). Table 2 shows a detailed comparison of calculated total interaction energies for proteins and Toyopearl beads alongside their respective experimental retention volumes.
3.4 Hydrophobic interaction chromatography The model proteins were subjected to hydrophobic interaction chromatography experiments using the Toyopearl Phenyl and Toyopearl Butyl. Standard chromatographic conditions were applied using a linear gradient from 1.7 M www.jss-journal.com
622
J. Sep. Sci. 2014, 37, 618–624
M. Aasim et al.
Figure 2. Interaction energy (U) versus distance (H) profile for lysozyme (blue area) and immunoglobulin G (red area) with Toyopearl Butyl 650-C immobilized on different surfaces. The interaction profiles are represented by shaded areas rather than single lines as they represent the complete range of energies considering the dehydrated (glass and plastic) as well as hydrated (membrane) states of the protein layers onto which contact angle measurements were performed.
ammonium sulfate in 20 mM phosphate buffer at pH 7 to a 20 mM phosphate buffer within 10 CV. The separation behavior of all the proteins on Toyopearl Phenyl has been described previously. The retention volumes on Toyopearl Butyl also follow the same sequence as for the phenyl derivative, except that the retention volumes are higher (see Table 2). While the phenyl ligand, due to its larger size and solvent-accessible surface area, can be considered more hydrophobic than the butyl ligand, the model proteins display a higher affinity to Toyopearl Butyl 650-C as compared to Toyopearl Phenyl 650-C, thereby reflecting the greater hydrophobic nature of the adsorbent having the butyl ligand [20, 21, 49]. TOSOH Bioscience, the
Figure 3. Elution profile for different proteins on a 2 mL column Tricon 5/100 packed with Toyopearl Phenyl 650-C. % Buffer B is the gradient length of 10 CV from buffer A (20 mM phosphate at pH 7 with 1.7 M ammonium sulfate) to buffer B (20 mM phosphate at pH 7).
producer of these materials, also claims the same behavior in their product data sheets, which confirms our contact angle data and extended DLVO calculations. Figure 3 represents an overlay of different protein peaks in one chromatogram. It shows the protein peaks (listed in Table 2) on Toyopearl Phenyl support. A range of not retained (PGase), loosely retained (lysozyme, BSA) and strongly retained (Lactoferrin, -Gal, and IgG) proteins are shown. Figure 4 depicts two superimposed chromatograms of Lys on both the chromatographic materials, where Toyopearl Butyl 650-C retains the protein longer than Toyopearl Phenyl 650-C.The separation behavior is recorded under the same chromatographic conditions, having the same mobile phase, column dimensions, flow rate and gradient length. Similar separation experiments on other proteins resulted in two general groups of proteins based on their retention volumes (see
Table 2. Interaction energies of HIC adsorbents with various proteins and retention volumes of model proteins in a gradient elution of 10 CV
Protein
APR PGase CHY APP Lys BSA -Gal LF IgG
Toyopearl phenyl
Toyopearl butyl
Retention volumea) , b) (mL)
|U|ILC (10−3 ·kT)
|U|CA (10−3 ·kT)
Retention volumeb) (mL)
2.0 ± 0.1 2.0 ± 0.0 2.1 ± 0.3 2.3 ± 0.1 14.8 ± 1.6 15.0 ± 0.5 16.6 ± 1.3 17.8 ± 0.8 21.6 ± 0.3
150.2 162.1 61.6 121.8 20.6 95.1 91.0 88.9 175.0
10.8 ± 0.4 11.5 ± 0.7 4.1 ± 0.3 11.2 ± 0.9 1.4 ± 0.1 6.6 ± 0.4 6.5 ± 0.4 6.4 ± 0.4 12.6 ± 0.8
2.0 2.2 3.0 2.3 17.0 28.0 20.0 21.3 24.0
± ± ± ± ± ± ± ± ±
0.2 0.0 0.0 0.0 0.1 0.1 0.1 0.0 0.2
|U|ILC (10−3 ·kT)
|U|CA (10−3 ·kT)
172.7 186.3 70.5 141.9 23.6 109.3 104.2 102.6 200.0
13.8 14.8 5.3 16.7 1.8 8.5 8.3 8.3 13.8
± ± ± ± ± ± ± ± ±
1.2 1.3 0.5 2.0 0.2 0.8 0.8 0.8 1.4
a) Taken from reference [28]. b) Averaged over three separate runs. APR, aspartic protease; PGase, polygalacturonase; CHY, chymosin; APP, aminopeptidase; Lys, lysozyme; BSA, bovine serum albumin; -Gal, -galactosidase; LF, lactoferrin; and IgG, immunoglobulin G. |U|ILC and |U|CA represent the total xDLVO interaction energies calculated using inverse liquid chromatography (ILC) and contact angle measurements, respectively.
C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.jss-journal.com
Liquid Chromatography
J. Sep. Sci. 2014, 37, 618–624
623
tently have lower interaction energies to the model proteins than Toyopearl Butyl beads, and consequently, lower retention volumes. The interaction energy values are able to discriminate between the properties of the adsorbent beads, and are associated with chemical structure of the ligand, as the both possess identical base matrices. Furthermore, the effect of the base matrix on chromatographic retention volumes has already been previously reported.
4 Conclusions Figure 4. Elution profile for Lysozyme (Lys) on a 2 mL column Tricon 5/100 packed with Toyopearl Phenyl 650-C (blue line) and Toyopearl Butyl 650-C (red line). % Buffer B is the gradient length of 10 CV (dotted line) from buffer A (20 mM phosphate at pH 7 with 1.7 M ammonium sulfate) to buffer B (20 mM phosphate at pH 7).
Table 2). The first group of proteins are relatively hydrophilic in nature and do not bind to either adsorbent even at high salt concentrations. These proteins include aminopeptidase, chymosin, aspartic protease, and polygalacturonase. The other group consisted of moderate to highly hydrophobic proteins and were retained at certain concentrations of ammonium sulfate upon elution. These proteins were BSA, lactoferrin, lysozyme, IgG, and -Gal. As mentioned earlier, the retention volumes for all these proteins were higher in case of Toyopearl Butyl 650-C. The average protein–bead interaction energies for each of the model proteins correlate well to their corresponding retention volumes on each of the chromatographic supports (see Fig. 5). Interestingly, calculations from two separate sets of data (contact angle measurements versus ILC) yielded nearly the same trend, which validates the general qualitative surface behavior of these materials. Toyopearl Phenyl beads consis-
The approach presented in this study was proposed to understand protein adsorption onto HIC supports functionalized with different ligands. Theoretical calculations were made employing extended DLVO theory by experimental surface characterization of all interacting surfaces involved using contact angle and zeta potential measurements. These theoretical interaction energies between model proteins and the chromatographic supports were shown to correlate with their respective retention volumes in real chromatographic experiments. Furthermore, the same predictions were reached using two separate surface characterization methods, thereby validating the surface behavior of the chromatographic beads. The data presented also compared two groups of proteins on the basis of retention volumes and interaction energies, namely, loose-binding proteins showing low retention and strong-binding proteins showing high retention volumes. Loose-binding proteins were unable to discriminate between the hydrophobicity of either chromatographic bead. These results indicate that the extended DLVO approach toward investigating interfacial interactions may provide a diagnostic tool toward understanding protein adsorption onto chromatographic supports and to guide future high-throughput adsorbent design, while minimizing costly and time-consuming full-scale chromatographic experiments.
Figure 5. Correlation graph between the depth of secondary energy pocket of model proteins and their corresponding retention volume with different chromatographic supports Toyopearl Phenyl 650-C (blue squares) and Toyopearl Butyl 650-C (red circles), which have been calculated using (A) ILC, and (B) contact angle measurements.
C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.jss-journal.com
624
M. Aasim et al.
Muhammad Aasim is thankful to the Higher Education Commission of Pakistan (University of Malakand) for financial support. Marcelo Fern´andez-Lahore is member of the Consejo Nacional de Investigaciones Cientificas (CONICET) (Buenos Aires, Argentina). This work has been partially supported be the European Commission under the Project FP7-SME-2007-1 ELECTROEXTRACTION 222220 and DFG Project FE-3 AFMDLVO-Theorien, Project No. 50364. The authors have declared no conflict of interest.
5 References
J. Sep. Sci. 2014, 37, 618–624
[22] Salgado, J. C., Rapaport, I., Asenjo, J. A., J. Chromatogr. A 2005, 1098, 44–54. [23] Srinivasan, R., Ruckenstein, E., Sep. Purif. Methods 1980, 9, 267–370. [24] Van Oss, C. J., Good, R. J., Chaudhury, M. K., J. Chromatogr. 1986, 376, 111–119. [25] Helm, C. A., Knoll, W., Israelachvili, J. N., Proc. Natl. Acad. Sci. U.S.A. 1991, 88, 8169–8173. [26] Van Oss, C. J., Moore, L. L., Good, R. J., Chaudhury, M. K., J. Protein Chem. 1985, 4, 245–263. [27] Janson, J. C., Protein Purification: Principles, High Resolution Methods, and Applications, John Wiley & Sons, New Jersey, USA, 2011. [28] Aasim, M., Poondi, G. R., Vennapusa, R. R., FernandezLahore, M., Int. Rev. Chem. Eng. 2010, 2, 263–271.
[1] Jungbauer, A., Machold, C., Hahn, R., J. Chromatogr. 2005, 1079, 221–228.
[29] Aasim, M., Bibi, N. S., Vennapusa, R. R., Fernandez Lahore, M., J. Chem. Tech. Biotech. 2013, 88, 1421–1428.
[2] Kato, Y., Nakamura, K., Kitamura, T., Moriyama, H., Hasegawa, M., Sasaki, H., J. Chromatogr. A 2002, 971, 143–149.
[30] Van Oss, C. J., Good, R. J., Chaudhury, M. K., J. Colloid Interface Sci. 1986, 111, 378–390.
[3] O’Farrell, P. A., Hydrophobic Interaction Chromatography, Springer, Berlin 1998. [4] Shepard, C. C., Tiselius, A., Discuss. Faraday Soc. 1949, 7, 275–285. [5] Berna, P. P., Berna, N., Porath, J., Oscarsson, S., J. Chromatogr. A 1998, 800, 151–159.
[31] Van Oss, C. J., Interfacial Forces in Aqueous Media, Taylor and Francis, New York 2006. [32] Bos, R., Mei, H. C., Busscher, H. J., FEMS Microbiol. Rev. 1999, 23, 179–230. [33] Sharma, P. K., Hanumantha, R. K., Adv. Colloid Interface Sci. 2002, 98, 341–463.
´ [6] Melander, W., Horvath, C., Arch. Biochem. Biophys. 1977, 183, 200–215.
[34] Vennapusa, R. R., Hunegnaw, S. M., Cabrera, R., Fernandez-Lahore, M., J. Chromatogr. A 2008, 1181, 9–20.
[7] Arakawa, T., Timasheff, S. N., Biochemistry 1982, 21, 6545–6552.
[35] Vennapusa, R. R., Aasim, M., Cabrera, R., FernandezLahore, M., Biotech. Biopro. Eng. 2009, 14, 419–428.
[8] To, B. C. S., Lenhoff, A. M., J. Chromatogr. A 2007, 1141, 191–205.
[36] Vennapusa, R. R., Fernandez-Lahore, M., J. Biosci. Bioeng. 2010, 110, 564–571.
[9] Rippel, G., Szepesy, L., J. Chromatogr. A 1994, 664, 27–32.
[37] Vennapusa, R. R., Tari, C., Cabrera, R., Fernandez-Lahore, M., Biochem. Eng. J. 2009, 43, 16–26.
[10] Goheen, S. C., Engelhorn, S. C., J. Chromatogr. A 1984, 317, 55–65.
[38] Aasim, M., Bibi, N. S., Vennapusa, R. R., FernandezLahore, M., J. Sep. Sci. 2012, 35, 1068–1078.
[11] Chen, W. Y., Huang, H. M., Lin, C. C., Lin, F. Y., Chan, Y. C., Langmuir 2003, 19, 9395–9403.
[39] Bednar, I., Tscheließnig, R., Berger, E., Podgornik, A., Jungbauer, A., J. Chromatogr. A 2012, 1220, 115–121.
[12] Fausnaugh, J. L., Kennedy, L. A., Regnier, F. E., J. Chromatogr. A 1984, 317, 141–155.
[40] Van Oss, C. J., Giese, R. F., J. Dispers. Sci. Technol. 2004, 25, 631–655.
[13] Wu, S., Figueroa, A., Karger, B. L., J. Chromatogr. A 1986, 371, 3–27.
[41] Van Oss, C. J., Good, R. J., J. Macromol. Sci. Part A Pure Appl. Chem. 1989, 26, 1183–1203.
[14] Schmuck, M. N., Nowlan, M. P., Gooding, K. M., J. Chromatogr. A 1986, 371, 55–62.
[42] Igarashi, C., Nishizawa, K., Colloids Surf. Physicochem. Eng. Aspects 2007, 303, 235–240.
[15] Lin, F., Chen, W., Ruaan, R., Huang, H., J. Chromatogr. A 2000, 872, 37–47.
[43] Bismarck, A., Kumru, M. E., Springer, J., J. Colloid Interface Sci. 1999, 217, 377–387.
[16] Lin, F., Chen, W., Anal. Chem. 2001, 73, 3875–3883.
[44] Huang, A. Y., Berg, J. C., J. Colloid Interface Sci. 2006, 296, 159–164.
[17] Van Oss, C. J., Curr. Opin. Colloid Interface Sci. 1997, 2, 503–512. [18] Van Oss, C. J., Mol. Immunol. 1995, 32, 199–211.
[45] Kirby, B. J., Hasselbrink Jr., E. F., Electrophoresis 2004, 25, 203–213.
´ [19] Lienqueo, M. E., Mahn, A., Vasquez, L., Asenjo, J. A., J. Chromatogr. A 2003, 1009, 189–196.
[46] Bak, H., Thomas, O. R. T., J. Chromatogr. B 2007, 848, 116–130.
[20] Chen, J., Yang, T., Cramer, S. M., J. Chromatogr. A 2008, 1177, 207–214.
[47] Queiroz, J. A., Tomaz, C. T., Cabral, J. M. S., J. Biotechnol. 2001, 87, 143–159.
[21] Ladiwala, A., Xia, F., Luo, Q., Breneman, C. M., Cramer, S. M., Biotechnol. Bioeng. 2006, 93, 836–850.
[48] Oliveira, R., Exp. Therm. Fluid Sci. 1997, 14, 316–322.
C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
[49] Chen, J., Yang, T., Luo, Q., Breneman, C. M., Cramer, S. M., React. Funct. Polym. 2007, 67, 1561–1569.
www.jss-journal.com
J. Sep. Sci. 2014, 37, 618–624
Muhammad Aasim1,2∗ Prasad Babu Kakarla1∗ Roy N. D’Souza1 Noor Shad Bibi1 Tanja Yvonne Klein3 Laura Treccani3 Kurosch Rezwan3 1 ´ Marcelo Fernandez-Lahore
Research Article
1 Downstream
Protein adsorption onto hydrophobic chromatographic supports has been investigated using a colloid theory surface energetics approach. The surface properties of commercially available chromatographic beads, Toyopearl Phenyl 650-C, and Toyopearl Butyl 650-C, have been experimentally determined by contact angle and zeta potential measurements. The adsorption characteristics of these beads, which bear the same backbone matrix but harbor different ligands, have been studied toward selected model proteins, in the hydrated as well as dehydrated state. There were two prominent groups of proteins observed with respect to the chromatographic supports presented in this work: loosely retained proteins, which were expected to have lower average interaction energies, and the strongly retained proteins, which were expected to have higher average interaction energies. Results were also compared and contrasted with calculations derived from adsorbent surface energies determined by inverse liquid chromatography. These results showed a good qualitative agreement, and the interaction energy minima obtained from these extended Derjaguin, Landau, Verwey and Overbeek calculations were shown to correlate well with the experimentally determined adsorption behavior of each protein.
Bioprocessing Laboratory, School of Engineering and Science, Jacobs University, Campus Ring 1, Bremen, Germany 2 Department of Biotechnology, University of Malakand, Chakdara, Dir, Khyber Pakhtunkhwa, Pakistan 3 Keramische Werkstoffe und Bauteile/Advanced Ceramics, University of Bremen, Am Biologischen Garten 2, Bremen, Germany
Received December 16, 2013 Revised January 9, 2014 Accepted January 9, 2014
The role of ligands on protein retention in adsorption chromatography: A surface energetics approach
Keywords: Adsorption chromatography / Hydrophobic interactions / Protein adsorption / Surface energetics DOI 10.1002/jssc.201301338
Additional supporting information may be found in the online version of this article at the publisher’s web-site
1 Introduction Hydrophobic interaction chromatography (HIC) has widely been employed in the separation of proteins and peptides with high resolution, high yield, and in nondenaturing environments, on laboratory and industrial scales [1, 2]. It has been presumed that almost 50% of the surface of soluble proteins contains hydrophobic residues, which can be exploited by HIC [3]. HIC was initially reported in 1949 by Shepard and Tiselius as salting-out chromatography [4], and takes advantage of intrinsic protein hydrophobicity to achieve separation, where the hydrophobic patches on the surface of the protein interact with hydrophobic ligands on the chromatographic matrix [5]. These seemingly weak interactions are important in sustaining the native structure and activity ´ Correspondence: Professor Marcelo Fernandez-Lahore, Downstream BioProcessing Laboratory, Jacobs University, Campus Ring 1, D-28759 Bremen, Germany E-mail: [email protected] Fax: +49 421 200 3600
Abbreviations: AB, Acid–base; HIC, hydrophobic interaction chromatography; IgG, total human Immunoglobulin G; ILC, inverse liquid chromatography; LW, Lifshitz–Van der Waals; Lys, lysozyme; -Gal, -galactosidase C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
of the proteins [3]. The proteins are retained at high salt concentration and eluted at decreasing salt concentration, with the least hydrophobic protein eluting first and the most hydrophobic protein released at the end [6]. Several concepts have been developed to explain the interaction process during HIC. Solvophobic theory proposes a linear relationship between protein adsorption and salt concentration in the mobile phase [6]. Therein, the interaction strength was attributed the molar surface tension increment due to the kosmotropic salt concentration [6]. In contrast, the preferential interaction model excludes salt ions surrounding the ligand and proteins and is based on the protein’s ability to either interact with another protein or with the ligand matrix [7]. Furthermore, the dependence of protein retention on the protein and adsorbent properties [8], salt concentration [9], temperature [10, 11], pH of the buffer [12, 13], ligand type [12, 14], as well as the density of the ligand immobilized on the base matrix [2, 12] has been widely investigated. Microcalorimetry has also been employed to probe protein adsorption thermodynamics [15, 16], and association equilibrium constants derived from Langmuir adsorption isotherms [8] can serve a good tool to calculate the association interaction ∗ These authors have contributed equally to this work. Colour Online: See the article online to view Figs. 2, 4 and 5 in colour.
www.jss-journal.com
Liquid Chromatography
J. Sep. Sci. 2014, 37, 618–624
energies of proteins to hydrophobic supports [17, 18]. The retention of the protein was also predicted by the determination of “dimensionless retention times (DRT)” by calculating the surface hydrophobicity of the proteins [19]. Predictive quantitative structure–retention relationship (QSRR) models have also been previously employed [20, 21], and in one case, the primary sequence of the protein was shown to predict retention behavior [22]. Noncovalent van der Waals forces have been considered to be the major contributing factor to protein adsorption during HIC [23, 24]. These interaction forces are considered as nonspecific weak interactions [25, 26], which are important to retain protein structure and function. Despite sophisticated laboratory equipment and chromatographic media, difficulties still persist in finding optimum purification conditions [27]. This study investigates the applicability of interfacial chemistry techniques as a diagnostic tool to understand protein interactions at the nanoscale. The underlying assumption of proteins being nano-colloidal particles is instrumental to apply extended Derjaguin, Landau, Verwey and Overbeek (DLVO) theory [26, 28–30], and subsequently investigate the interaction of these proteins to chromatographic beads via surface thermodynamic calculations. According to extended DLVO theory, the interaction energy between a colloidal particle (protein) and a solid support (chromatographic bead) is the cumulative sum of Lifshitz–van der Waals (LW), acid– base (AB), and electrostatic (EL) energies [31] as shown in Eq. (1).
U XDLVO = U LW + U EL + U AB
(1)
The energy components shown in Eq. (1) are derived from contact angle measurements (LW and AB) and zeta potential (EL) determinations. Details of the calculations of different energy components are given in our previous paper [38]. The surface thermodynamics approach has already been used for investigating the interaction of colloidal particles to a broad range of solid substrata in environmental, biomedical, and bioprocesses [32,33]. It has also helped study the interaction of various types of biomass with expanded bed chromatographic beads in several modes of operation [34–37]. Protein interactions to certain chromatographic supports have also been studied, where early or late eluting proteins could only be marginally discriminated [28, 29, 38]. The aim of this work was to investigate the fundamental effect of ligand chemistry on protein retention and test the validity of extended DLVO (xDLVO) calculations toward predicting protein–adsorbent interactions. The basic parameters for the calculations were assessed by contact angle measurements and zeta potential determinations by electrophoretic mobility and streaming potentials. These results were compared with those obtained using parameters assessed by inverse liquid chromatography (ILC) [39]. Details of the xDLVO calculations are provided in the Supporting Information. C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
619
2 Materials and methods 2.1 Chemicals and reagents Toyopearl Phenyl 650-C and Toyopearl Butyl 650-C were purchased from TOSOH Bioscience (Stuttgart, Germany). 1-Bromonaphthalene and formamide were purchased from Fluka (Buchs, Switzerland). Ultrapure water was used for contact angle determinations. All other chemicals were of analytical grade. Hen’s egg white lysozyme (HEWL) and BSA were obtained from Sigma-Aldrich (Munich, Germany). PolyR 3XL) and galacturonase from Aspergillus niger (Pectinex R ) were Aminopeptidase from Aspergillus oryzae (Flavorzyme purchased from Novozymes (Lund, Sweden). Chymosin from R ), Aspartic proAspergillus niger var. awamori (CHY-MAX R L 205), and tease from Rhizomucor miehei (Hannilase -galactosidase from Kluyveromyces lactis (HA-Lactase 5200) were from CHR. HANSEN (Holdorf, Germany). Total huR ) was from Ocman mmunoglobulin G (hT-IgG/Octagam tapharma (Langenfeld, Germany). Lactoferrin was obtained from Biopole (Namur, Belgium). All the protein solutions were extensively dialyzed against water to remove salts and preservatives and stored at 4⬚C as lyophilized powders until further use.
2.2 Contact angle measurements Protein layers were deposited on microscopic glass slides according to a previously reported method [26]. Proteins were also deposited onto a plastic Petri dish support, as previously described [28]. For generating uniform protein-coated surfaces, protein solutions (2 mg/mL in 20 mM phosphate buffer at pH 7) were deposited on a glass slide or Petri dish and dried overnight at room temperature in a dust-free environment. Excess buffer was carefully removed with adsorbent paper and the protein-coated surface was then dried under vacuum. Proteins were additionally deposited on an ultrafiltration (cellulose) membrane (10 kDa MWCO; Millipore) to measure the contact angles in the hydrated state [26]. Subsequently, the membrane was removed and kept at room temperature for 20 min before measurements. Contact angles were measured using a goniometer (OCA20, DataPhysics Instruments, Germany) according to the sessile drop technique [34], employing water, formamide, and 1-bromonaphthalene as probe liquids to determine surface characteristics [40, 41]. Small drops of the probe liquids (2–5 L) were dispensed on the samples. Contact angle measurements were performed six times for each probe liquid at room temperature. The measured contact angle values were used for surface energy calculations, according to the van Oss AB theory [31], using the instrument’s software (SCA 20). All contact angle measurements were also performed on the homogenous base surfaces (agar plates or glass/plastic/membrane surfaces). www.jss-journal.com
620
J. Sep. Sci. 2014, 37, 618–624
M. Aasim et al.
raphy data were stored and analyzed using a commercial software package (UNICORNTM 5).
2.3 Zeta potential determinations Zeta potential values were measured with a Zetasizer Nano ZS from Malvern Instruments (Worcestershire, UK) and all measurements were performed in triplicate. For protein samples, measurements were carried out using 2 mg/mL solutions prepared in a 20 mM phosphate buffer (pH 7). Electrophoretic mobility data were used to calculate zeta potentials within the instrument software using Henry’s equation [31].
2.4 Streaming potential measurements The charge characteristics of intact chromatographic beads were determined by streaming potential measurements (SurPASS EKA, Anton Paar, Austria). These electrokinetic measurements were converted to zeta potentials using Eq. (2) [42] at different pH conditions, where is the viscosity, is the specific electric conductivity of the electrode solution, g is the dielectric constant of the electrolyte, and dU/dP represents the variation of the streaming potential with applied pressure. =
dU · g0g d P
(2)
A 1 mM KCl solution was used as an electrolyte for the determination of streaming potentials. A 2–3 mm layer of the material was packed in the cylindrical cell. The pH value was varied within the range of pH 3–10 by the addition of 0.05 M HCl or 0.05 M NaOH solutions [43]. The values for zeta potentials at high salt concentration were extrapolated from the experimentally measured values, employing known correlations [44, 45].
2.5 Chromatographic experiments Tricon chromatography columns (5 mm internal diameter ¨ and 100 mm length) and the AKTA FPLC system were purchased from GE Healthcare (Munich, Germany). The hydrophobic beads were packed in a commercially available chromatographic column (2.0 mL bed volume; aspect ratio 4.0) and the packing quality was evaluated by residence time distribution analysis employing 1% acetone as a tracer [46]. The mobile phase was composed of two buffer solutions at pH 7: buffer A (binding buffer) with 20 mM phosphate buffer containing 1.7 M ammonium sulfate, and buffer B (elution buffer) with 20 mM phosphate buffer. Buffers were filtered and degassed before use. After equilibration with 10 column volumes (CV) of buffer A, a 200 L protein sample (4 mg/mL in buffer A) was injected into the column. Unbound material was eluted in 5 CV of buffer A and bound protein was eluted using a linear gradient from 0–100% of buffer B (10 CV). Reequilibration was performed with 5 CV of buffer A. The flow rate for all steps was 1 mL/min (300 cm/h). The eluate was monitored with a UV/Vis detector at 280 nm. Chromatog C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
3 Results and discussion 3.1 Contact angle and surface energies The deposited layers of the proteins and chromatographic bead fragments were characterized by contact angle measurements employing three diagnostic liquids: water, formamide and 1-bromonaphthalene [32]. The contact angle values and surface energy components for proteins, in the hydrated and dehydrated states have been published elsewhere [28]. The contact angle values of Toyopearl Phenyl 650-C and Toyopearl Butyl 650-C along with their nonfunctionalized analog are listed in Table 1. The surface free energy components, that is, the Lifshitz–van der Waals (␥LW ) and AB (␥+ and ␥− ) components of the proteins and chromatographic beads are directly calculated from their contact angles using Young’s equation (see Supporting Information). The energy components for the beads determined by contact angle measurements (on fragmented beads) were compared with those previously determined by ILC [39] (while these measurements were performed on intact beads of slightly different particle size, they possess identical surface chemistries). Table 1 shows that most of the chromatographic materials are hydrophilic in nature. However, the backbone of Toyopearl beads is composed of methacrylate polymers that generally have a relatively hydrophobic character. This effect is attributed to the presence of ether and hydroxyl groups on the surface of the modified bead (Toyopearl instruction manual IM02, TOSOH Bioscience). Moreover, there is a slight difference observed between the beads functionalized with different ligands. The material that has phenyl ligands shows a more hydrophilic character than the one that has butyl ligands. The phenyl ligand is generally considered to be more hydrophobic than the butyl ligand [47], but in this specific case, the backbone chemistry of the matrix might play an important role in contributing to the hydrophobicity of the butyl-functionalized matrix. This effect has been verified by interaction energy calculations as well as chromatographic experiments. The hydrophobic LW energy component was higher for the butyl ligand, while the polar AB component was higher for the phenyl ligand. There was only a minor difference between the total surface energy of both materials. It is interesting to note that the surface energies measured via ILC, while maintaining a similar qualitative trend in terms of individual energy components, are nearly 2–4 times larger in magnitude than those measured by contact angles.
3.2 Zeta potential determinations Contact angle measurements fail to account for surface electrostatic effects, which play an important role in the degree www.jss-journal.com
Liquid Chromatography
J. Sep. Sci. 2014, 37, 618–624
621
Table 1. Contact angles (measured by goniometry) and corresponding surface energies of various chromatographic beads in 20 mM phosphate buffer at pH 7.0. Surface energies determined by inverse liquid chromatography have also been listed for comparison [40]
Material
Toyopearl phenyl 650-Ca) Toyopearl phenyl 650-Mc) Toyopearl butyl 650-C Toyopearl butyl 650-Mc) Toyopearl HW 65b)
Surface tension (mJ·m−2 )
Contact angle WR
FM
AB
␥LW
␥+
␥−
␥AB
␥Total
1.5 ± 0.5
2.5 ± 0.8
48.8 ± 1.3
9.9 ± 1.1
7.5 ± 1.0
43.9 ± 2.0
5.4 ± 0.4
10.2 ± 0.4
66.5 ± 0.8
30.5 ± 0.6 124.0 32.9 ± 0.9 144.0 21.7 ± 0.4
3.9 ± 0.2 22.5 3.1 ± 0.3 20.1 7.7 ± 0.2
53.8 ± 0.1 52.3 52.7 ± 0.4 53.8 53.4 ± 0.1
28.9 ± 0.8 68.6 25.4 ± 1.0 65.8 40.4 ± 0.5
59.4 ± 0.2 192.6 58.2 ± 0.1 209.8 62.1 ± 0.1
WR stands for water; FM is formamide; AB is 1-bromonaphthalene. a) and b) Taken from reference [28]. c) Taken from reference [40] and measured by inverse liquid chromatography.
Toyopearl Butyl 650-C. Its zeta potential at pH 7 is around −20.0 mV.
3.3 Extended DLVO interaction energies as a function of distance
Figure 1. Zeta potential of Toyopearl Phenyl 650-C and Toyopearl Butyl 650-C obtained by steaming potential measurements in 1 mM KCl as a function of pH.
of interaction between surfaces. While the measurement of the surface charge of a material is not practically feasible, the electrostatic behavior of a surface can be estimated by its zeta potential, which is the electrostatic potential at the hydrodynamic plane of shear [48]. The zeta potential of a material can be conveniently determined by measuring its electrophoretic mobility [45]. This technique was successfully used to electrostatically characterize the model proteins employed in this study [28] (see Supporting Information). The zeta potential of larger particulates can be determined in varying pH conditions by using streaming potential measurements. The advantage of this method is that it exploits the beads in their intact form. Toyopearl beads have a broad pH stability range, that is, from pH 1 to 13 [43], and their surface charge can be adequately determined within this range. Figure 1 depicts the zeta potential of Toyopearl Phenyl 650-C and Toyopearl Butyl 650-C as a function of pH. It is evident that the phenyl derivative displays a transition from a negative to a positive surface charge (from −15.0 to 1.0 mV) with decreasing pH (from pH 11 to 3). Its zeta potential at pH 7 is around −10.0 mV. The same trend, but a higher surface charge (0 to −25.0 mV from pH 3 to 11) was observed for C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
By combining the experimental parameters determined in the previous sections, the total interaction energy of the proteins and chromatographic beads can be calculated as a function of the distance between them (see Supporting Information). These calculations are performed with well-defined equations, using sphere-to-plate geometry. The interaction energy components and the zeta potential values at high salts have been incorporated into the calculations. Our calculations have considered hydrated as well as dehydrated protein states. For example, Fig. 2 shows the interfacial energy of interaction (U) of human immunoglobulin G (IgG) and Lysozyme (Lys) with Toyopearl Butyl 650-C, as a function of distance (H). The average secondary energy minimum of IgG on three supporting surfaces on Toyopearl Butyl (−0.17 kT, where k is the Boltzmann constant and T is the absolute temperature) is comparatively much higher than that of Lys (−0.02 kT). Similar effects are observed for all model proteins, where interaction energy minima were higher in the case of adsorbents harboring butyl ligands. This very same effect was verified in actual chromatographic experiments (see Section 3.4). Table 2 shows a detailed comparison of calculated total interaction energies for proteins and Toyopearl beads alongside their respective experimental retention volumes.
3.4 Hydrophobic interaction chromatography The model proteins were subjected to hydrophobic interaction chromatography experiments using the Toyopearl Phenyl and Toyopearl Butyl. Standard chromatographic conditions were applied using a linear gradient from 1.7 M www.jss-journal.com
622
J. Sep. Sci. 2014, 37, 618–624
M. Aasim et al.
Figure 2. Interaction energy (U) versus distance (H) profile for lysozyme (blue area) and immunoglobulin G (red area) with Toyopearl Butyl 650-C immobilized on different surfaces. The interaction profiles are represented by shaded areas rather than single lines as they represent the complete range of energies considering the dehydrated (glass and plastic) as well as hydrated (membrane) states of the protein layers onto which contact angle measurements were performed.
ammonium sulfate in 20 mM phosphate buffer at pH 7 to a 20 mM phosphate buffer within 10 CV. The separation behavior of all the proteins on Toyopearl Phenyl has been described previously. The retention volumes on Toyopearl Butyl also follow the same sequence as for the phenyl derivative, except that the retention volumes are higher (see Table 2). While the phenyl ligand, due to its larger size and solvent-accessible surface area, can be considered more hydrophobic than the butyl ligand, the model proteins display a higher affinity to Toyopearl Butyl 650-C as compared to Toyopearl Phenyl 650-C, thereby reflecting the greater hydrophobic nature of the adsorbent having the butyl ligand [20, 21, 49]. TOSOH Bioscience, the
Figure 3. Elution profile for different proteins on a 2 mL column Tricon 5/100 packed with Toyopearl Phenyl 650-C. % Buffer B is the gradient length of 10 CV from buffer A (20 mM phosphate at pH 7 with 1.7 M ammonium sulfate) to buffer B (20 mM phosphate at pH 7).
producer of these materials, also claims the same behavior in their product data sheets, which confirms our contact angle data and extended DLVO calculations. Figure 3 represents an overlay of different protein peaks in one chromatogram. It shows the protein peaks (listed in Table 2) on Toyopearl Phenyl support. A range of not retained (PGase), loosely retained (lysozyme, BSA) and strongly retained (Lactoferrin, -Gal, and IgG) proteins are shown. Figure 4 depicts two superimposed chromatograms of Lys on both the chromatographic materials, where Toyopearl Butyl 650-C retains the protein longer than Toyopearl Phenyl 650-C.The separation behavior is recorded under the same chromatographic conditions, having the same mobile phase, column dimensions, flow rate and gradient length. Similar separation experiments on other proteins resulted in two general groups of proteins based on their retention volumes (see
Table 2. Interaction energies of HIC adsorbents with various proteins and retention volumes of model proteins in a gradient elution of 10 CV
Protein
APR PGase CHY APP Lys BSA -Gal LF IgG
Toyopearl phenyl
Toyopearl butyl
Retention volumea) , b) (mL)
|U|ILC (10−3 ·kT)
|U|CA (10−3 ·kT)
Retention volumeb) (mL)
2.0 ± 0.1 2.0 ± 0.0 2.1 ± 0.3 2.3 ± 0.1 14.8 ± 1.6 15.0 ± 0.5 16.6 ± 1.3 17.8 ± 0.8 21.6 ± 0.3
150.2 162.1 61.6 121.8 20.6 95.1 91.0 88.9 175.0
10.8 ± 0.4 11.5 ± 0.7 4.1 ± 0.3 11.2 ± 0.9 1.4 ± 0.1 6.6 ± 0.4 6.5 ± 0.4 6.4 ± 0.4 12.6 ± 0.8
2.0 2.2 3.0 2.3 17.0 28.0 20.0 21.3 24.0
± ± ± ± ± ± ± ± ±
0.2 0.0 0.0 0.0 0.1 0.1 0.1 0.0 0.2
|U|ILC (10−3 ·kT)
|U|CA (10−3 ·kT)
172.7 186.3 70.5 141.9 23.6 109.3 104.2 102.6 200.0
13.8 14.8 5.3 16.7 1.8 8.5 8.3 8.3 13.8
± ± ± ± ± ± ± ± ±
1.2 1.3 0.5 2.0 0.2 0.8 0.8 0.8 1.4
a) Taken from reference [28]. b) Averaged over three separate runs. APR, aspartic protease; PGase, polygalacturonase; CHY, chymosin; APP, aminopeptidase; Lys, lysozyme; BSA, bovine serum albumin; -Gal, -galactosidase; LF, lactoferrin; and IgG, immunoglobulin G. |U|ILC and |U|CA represent the total xDLVO interaction energies calculated using inverse liquid chromatography (ILC) and contact angle measurements, respectively.
C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.jss-journal.com
Liquid Chromatography
J. Sep. Sci. 2014, 37, 618–624
623
tently have lower interaction energies to the model proteins than Toyopearl Butyl beads, and consequently, lower retention volumes. The interaction energy values are able to discriminate between the properties of the adsorbent beads, and are associated with chemical structure of the ligand, as the both possess identical base matrices. Furthermore, the effect of the base matrix on chromatographic retention volumes has already been previously reported.
4 Conclusions Figure 4. Elution profile for Lysozyme (Lys) on a 2 mL column Tricon 5/100 packed with Toyopearl Phenyl 650-C (blue line) and Toyopearl Butyl 650-C (red line). % Buffer B is the gradient length of 10 CV (dotted line) from buffer A (20 mM phosphate at pH 7 with 1.7 M ammonium sulfate) to buffer B (20 mM phosphate at pH 7).
Table 2). The first group of proteins are relatively hydrophilic in nature and do not bind to either adsorbent even at high salt concentrations. These proteins include aminopeptidase, chymosin, aspartic protease, and polygalacturonase. The other group consisted of moderate to highly hydrophobic proteins and were retained at certain concentrations of ammonium sulfate upon elution. These proteins were BSA, lactoferrin, lysozyme, IgG, and -Gal. As mentioned earlier, the retention volumes for all these proteins were higher in case of Toyopearl Butyl 650-C. The average protein–bead interaction energies for each of the model proteins correlate well to their corresponding retention volumes on each of the chromatographic supports (see Fig. 5). Interestingly, calculations from two separate sets of data (contact angle measurements versus ILC) yielded nearly the same trend, which validates the general qualitative surface behavior of these materials. Toyopearl Phenyl beads consis-
The approach presented in this study was proposed to understand protein adsorption onto HIC supports functionalized with different ligands. Theoretical calculations were made employing extended DLVO theory by experimental surface characterization of all interacting surfaces involved using contact angle and zeta potential measurements. These theoretical interaction energies between model proteins and the chromatographic supports were shown to correlate with their respective retention volumes in real chromatographic experiments. Furthermore, the same predictions were reached using two separate surface characterization methods, thereby validating the surface behavior of the chromatographic beads. The data presented also compared two groups of proteins on the basis of retention volumes and interaction energies, namely, loose-binding proteins showing low retention and strong-binding proteins showing high retention volumes. Loose-binding proteins were unable to discriminate between the hydrophobicity of either chromatographic bead. These results indicate that the extended DLVO approach toward investigating interfacial interactions may provide a diagnostic tool toward understanding protein adsorption onto chromatographic supports and to guide future high-throughput adsorbent design, while minimizing costly and time-consuming full-scale chromatographic experiments.
Figure 5. Correlation graph between the depth of secondary energy pocket of model proteins and their corresponding retention volume with different chromatographic supports Toyopearl Phenyl 650-C (blue squares) and Toyopearl Butyl 650-C (red circles), which have been calculated using (A) ILC, and (B) contact angle measurements.
C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.jss-journal.com
624
M. Aasim et al.
Muhammad Aasim is thankful to the Higher Education Commission of Pakistan (University of Malakand) for financial support. Marcelo Fern´andez-Lahore is member of the Consejo Nacional de Investigaciones Cientificas (CONICET) (Buenos Aires, Argentina). This work has been partially supported be the European Commission under the Project FP7-SME-2007-1 ELECTROEXTRACTION 222220 and DFG Project FE-3 AFMDLVO-Theorien, Project No. 50364. The authors have declared no conflict of interest.
5 References
J. Sep. Sci. 2014, 37, 618–624
[22] Salgado, J. C., Rapaport, I., Asenjo, J. A., J. Chromatogr. A 2005, 1098, 44–54. [23] Srinivasan, R., Ruckenstein, E., Sep. Purif. Methods 1980, 9, 267–370. [24] Van Oss, C. J., Good, R. J., Chaudhury, M. K., J. Chromatogr. 1986, 376, 111–119. [25] Helm, C. A., Knoll, W., Israelachvili, J. N., Proc. Natl. Acad. Sci. U.S.A. 1991, 88, 8169–8173. [26] Van Oss, C. J., Moore, L. L., Good, R. J., Chaudhury, M. K., J. Protein Chem. 1985, 4, 245–263. [27] Janson, J. C., Protein Purification: Principles, High Resolution Methods, and Applications, John Wiley & Sons, New Jersey, USA, 2011. [28] Aasim, M., Poondi, G. R., Vennapusa, R. R., FernandezLahore, M., Int. Rev. Chem. Eng. 2010, 2, 263–271.
[1] Jungbauer, A., Machold, C., Hahn, R., J. Chromatogr. 2005, 1079, 221–228.
[29] Aasim, M., Bibi, N. S., Vennapusa, R. R., Fernandez Lahore, M., J. Chem. Tech. Biotech. 2013, 88, 1421–1428.
[2] Kato, Y., Nakamura, K., Kitamura, T., Moriyama, H., Hasegawa, M., Sasaki, H., J. Chromatogr. A 2002, 971, 143–149.
[30] Van Oss, C. J., Good, R. J., Chaudhury, M. K., J. Colloid Interface Sci. 1986, 111, 378–390.
[3] O’Farrell, P. A., Hydrophobic Interaction Chromatography, Springer, Berlin 1998. [4] Shepard, C. C., Tiselius, A., Discuss. Faraday Soc. 1949, 7, 275–285. [5] Berna, P. P., Berna, N., Porath, J., Oscarsson, S., J. Chromatogr. A 1998, 800, 151–159.
[31] Van Oss, C. J., Interfacial Forces in Aqueous Media, Taylor and Francis, New York 2006. [32] Bos, R., Mei, H. C., Busscher, H. J., FEMS Microbiol. Rev. 1999, 23, 179–230. [33] Sharma, P. K., Hanumantha, R. K., Adv. Colloid Interface Sci. 2002, 98, 341–463.
´ [6] Melander, W., Horvath, C., Arch. Biochem. Biophys. 1977, 183, 200–215.
[34] Vennapusa, R. R., Hunegnaw, S. M., Cabrera, R., Fernandez-Lahore, M., J. Chromatogr. A 2008, 1181, 9–20.
[7] Arakawa, T., Timasheff, S. N., Biochemistry 1982, 21, 6545–6552.
[35] Vennapusa, R. R., Aasim, M., Cabrera, R., FernandezLahore, M., Biotech. Biopro. Eng. 2009, 14, 419–428.
[8] To, B. C. S., Lenhoff, A. M., J. Chromatogr. A 2007, 1141, 191–205.
[36] Vennapusa, R. R., Fernandez-Lahore, M., J. Biosci. Bioeng. 2010, 110, 564–571.
[9] Rippel, G., Szepesy, L., J. Chromatogr. A 1994, 664, 27–32.
[37] Vennapusa, R. R., Tari, C., Cabrera, R., Fernandez-Lahore, M., Biochem. Eng. J. 2009, 43, 16–26.
[10] Goheen, S. C., Engelhorn, S. C., J. Chromatogr. A 1984, 317, 55–65.
[38] Aasim, M., Bibi, N. S., Vennapusa, R. R., FernandezLahore, M., J. Sep. Sci. 2012, 35, 1068–1078.
[11] Chen, W. Y., Huang, H. M., Lin, C. C., Lin, F. Y., Chan, Y. C., Langmuir 2003, 19, 9395–9403.
[39] Bednar, I., Tscheließnig, R., Berger, E., Podgornik, A., Jungbauer, A., J. Chromatogr. A 2012, 1220, 115–121.
[12] Fausnaugh, J. L., Kennedy, L. A., Regnier, F. E., J. Chromatogr. A 1984, 317, 141–155.
[40] Van Oss, C. J., Giese, R. F., J. Dispers. Sci. Technol. 2004, 25, 631–655.
[13] Wu, S., Figueroa, A., Karger, B. L., J. Chromatogr. A 1986, 371, 3–27.
[41] Van Oss, C. J., Good, R. J., J. Macromol. Sci. Part A Pure Appl. Chem. 1989, 26, 1183–1203.
[14] Schmuck, M. N., Nowlan, M. P., Gooding, K. M., J. Chromatogr. A 1986, 371, 55–62.
[42] Igarashi, C., Nishizawa, K., Colloids Surf. Physicochem. Eng. Aspects 2007, 303, 235–240.
[15] Lin, F., Chen, W., Ruaan, R., Huang, H., J. Chromatogr. A 2000, 872, 37–47.
[43] Bismarck, A., Kumru, M. E., Springer, J., J. Colloid Interface Sci. 1999, 217, 377–387.
[16] Lin, F., Chen, W., Anal. Chem. 2001, 73, 3875–3883.
[44] Huang, A. Y., Berg, J. C., J. Colloid Interface Sci. 2006, 296, 159–164.
[17] Van Oss, C. J., Curr. Opin. Colloid Interface Sci. 1997, 2, 503–512. [18] Van Oss, C. J., Mol. Immunol. 1995, 32, 199–211.
[45] Kirby, B. J., Hasselbrink Jr., E. F., Electrophoresis 2004, 25, 203–213.
´ [19] Lienqueo, M. E., Mahn, A., Vasquez, L., Asenjo, J. A., J. Chromatogr. A 2003, 1009, 189–196.
[46] Bak, H., Thomas, O. R. T., J. Chromatogr. B 2007, 848, 116–130.
[20] Chen, J., Yang, T., Cramer, S. M., J. Chromatogr. A 2008, 1177, 207–214.
[47] Queiroz, J. A., Tomaz, C. T., Cabral, J. M. S., J. Biotechnol. 2001, 87, 143–159.
[21] Ladiwala, A., Xia, F., Luo, Q., Breneman, C. M., Cramer, S. M., Biotechnol. Bioeng. 2006, 93, 836–850.
[48] Oliveira, R., Exp. Therm. Fluid Sci. 1997, 14, 316–322.
C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
[49] Chen, J., Yang, T., Luo, Q., Breneman, C. M., Cramer, S. M., React. Funct. Polym. 2007, 67, 1561–1569.
www.jss-journal.com
