1 Protein Binding Kinetics in Multimodal Systems - ACS Publications

Subscriber access provided by READING UNIV

Article

Protein Binding Kinetics in Multimodal Systems: Implications for Protein Separations Kartik Srinivasan, Mirco Sorci, Lars Sejergaard, Swarnim Ranjan, Georges Belfort, and Steven M. Cramer Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.7b04158 • Publication Date (Web): 07 Dec 2017 Downloaded from http://pubs.acs.org on December 12, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Analytical Chemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 21 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Protein Binding Kinetics in Multimodal Systems: Implications for Protein Separations Kartik Srinivasan,†,§ Mirco Sorci,†,§ Lars Sejergaard,†,§ Swarnim Ranjan,†,§ Georges Belfort,†,§ Steven M. Cramer*,†,§, †

Howard P. Isermann Department of Chemical and Biological Engineering, Rensselaer Polytechnic

Institute, Troy, NY 12180, USA §

Center for Biotechnology and Interdisciplinary Studies, Rensselaer Polytechnic Institute, Troy, NY

12180, USA

* Corresponding author: Fax (518) 276-4030; e-mail: [email protected]

1 ACS Paragon Plus Environment

Analytical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 21

Abstract In this work, quartz crystal microbalance with dissipation (QCM-D) was employed to study the kinetic processes involved in the interaction of proteins with self-assembled monolayers (SAMs) of multimodal (MM) ligands. SAMs were fabricated to mimic two chromatographic multimodal resins with varying accessibility of the aromatic moiety to provide a well-defined model system. Kinetic parameters were determined for two different proteins in the presence of the arginine and guanidine and a comparison was made with chromatographic retention data. The results indicated that the accessibility of the ligand’s aromatic moiety can have an important impact on the kinetics and chromatographic retention behavior. Interestingly, arginine and guanidine had very different effects on the protein adsorption and desorption kinetics in these MM systems. For cytochrome C, arginine resulted in a significant decrease and increase in the adsorption and desorption rates, respectively, while guanidine produced a dramatic increase in the desorption rate, with minimal effect on the adsorption rate. In addition, at different concentrations of arginine, two distinct kinetic scenarios were observed. For α-chymotrypsin, the presence of 0.1 M guanidine in the aromatic exposed ligand system produced an increase in the adsorption rate and only a moderate increase in the desorption rate which helped to explain the surprising increase in the chromatographic salt elution concentration. These results demonstrate that protein adsorption kinetics in the presence of different mobile phase modifiers and MM ligand chemistries can play an important role in contributing to selectivity in MM chromatography.


Page 3 of 21 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Introduction The development of efficient bioseparation processes for production of high purity biopharmaceuticals is an important challenge facing the biotechnology industry. Recent advancements in the development of multimodal (MM), or mixed mode chromatographic resins have enabled the identification of unique windows of selectivity as compared with traditional single mode chromatographic systems.1-7 This has contributed to improvements in product quality and process efficiency in industrial-scale manufacturing.8,9 While MM resins offer significant potential for bioseparations, there is a lack of understanding of the kinetic processes that govern protein-ligand interactions in these systems. The majority of commercially available MM resins employ various combinations of hydrophobic, electrostatic and/or hydrogen bonding interactions. A variety of moieties and geometries are currently being explored and a combination of these different interaction modes can create unexpected selectivity trends in protein purification.10,11 The lack of understanding of the interactions between MM ligands and biomolecules has hindered the design and optimization of improved MM ligand systems and chromatographic processes. Accordingly, there is an urgent need to develop biophysical tools to provide a deeper understanding of the interactions and selectivities obtained in these systems.12,13 The interaction of proteins with solid surfaces has been examined using a range of experimental tools such as

circular dichroism, infrared spectroscopy,

fluorescence and NMR.14 Hydrogen-deuterium

exchange mass spectroscopy has also been employed to identify solvent accessible regions on proteins during binding to chromatographic surfaces15,16 and to probe unfolding and aggregation of monoclonal antibodies in the adsorbed state.17,18 Self-assembled monolayers (SAMs) have also emerged as an important platform to probe protein-surface interactions.19-22 SAMs enable the precise control of the chemical properties of a solid surface by varying the concentrations and relative compositions of the functionalized head groups.22 SAMs have also been used in concert with nanoparticles (NPs) to provide useful systems for modeling solid surfaces.23-26 The adsorption of proteins to SAMs of MM ligands in the presence of various salt conditions and displacers has also been examined through the use of surface plasmon resonance (SPR).27,28 We have examined protein binding in ion exchange and MM chromatographic systems using free solution NMR in concert with chromatographic data for a library of ubiquitin mutants.12 Molecular dynamics (MD) simulations have also been used to study the binding of MM ligands to protein surfaces29 and a spherical harmonics method has been used to enable a quantitative analysis of ligand and moiety binding.30 We have also used MD simulations in concert with NMR to provide deeper insight into



Page 4 of 21

protein-ligand interactions in MM systems.13,31 Recently, we performed isothermal titration calorimetry and NMR spectroscopy studies on NPs with functionalized MM SAMs.32 We have also examined the behavior of mobile phase modifiers in MM systems using MD simulations and chromatographic data.33 Although these approaches provided structural and thermodynamic insights into protein interactions in MM chromatography, they have not explicitly accounted for the kinetic processes that may play an important role in these systems. Quartz crystal microbalance with dissipation (QCM-D) has been employed by several investigators to probe interactions of biomolecules with functionalized surfaces.34-37 We have previously used QCM-D to test the viscoelastic properties of (i) gels versus brushes,38 (ii) adsorbed and cross-linked polypeptide and protein layers39 and (iii) polycationic and polyanionic layers.40 QCM-D has been employed to extract kinetics of interaction between biomolecules such as DNA.41 Moreover, QCM-D has been used to examine the kinetics of pH dependent enzyme catalysis and to probe the disulfide bond formation pathway.42,43 Further, investigators have studied protein-surface interactions through the use of QCM-D in concert with complementary techniques like SPR.44,45 In the present work, a QCM-D based technique was developed to probe the kinetics of protein-MM ligand interactions by obtaining real time adsorption data using SAMs of MM ligands that mimic a chromatographic resin surface. SAMs of different ligand chemistries and densities were created on gold coated quartz crystals in order to study subtle differences in protein binding kinetics. Experiments were then carried out for two model proteins at various protein concentrations in the presence of the two mobile phase modifiers, arginine and guanidine. A global optimization fit was then used to determine the kinetic parameters which were evaluated over a range of conditions. Finally, a qualitative comparison was carried out between retention data from column experiments and the kinetic rates obtained from the QCM-D experiments in order to develop a deeper understanding about the kinetic effects in these selective MM systems. Experimental section Materials: Hexaethylene glycol thiols terminated in hydroxyl (EG6OH) or N-hydroxysuccinimide groups (EG6NHS) were purchased from Prochimia (Sopot, Poland). Sodium thiocyanate, 4-amino hippuric acid, L-arginine hydrochloride, guanidine hydrochloride, potassium phosphate, tris (hydroxymethyl) aminomethane, cytochrome C (horse heart), α-chymostrypsin (bovine pancreas), hydrochloric acid, sodium chloride, sodium azide, acetic acid, sodium hydroxide, sodium acetate, were purchased from Sigma-Aldrich (St. Louis, MO). N-benzoyl-L-lysine was purchased from Chem-Impex


Page 5 of 21 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


International (Wood Dale, IL). Ethanol 200 proof was purchased from EMD Millipore (Billerica, MA). Acetonitrile was purchased from Acros Organics (Geel, Belgium). Compressed N2 gas was supplied by Airgas (Albany, NY). AT-cut gold coated QCM-D sensors with a fundamental frequency of 5 MHz were purchased from Biolin Scientific Inc. (Paramus, NJ). Equipment: Reverse phase chromatography was performed using a Waters (Milford, MA) High Performance Liquid Chromatography system consisting of a 600 multisolvent delivery system, a 717 Waters Intelligent Sample Processor auto injector, and a 996 photodiode array detector controlled by a Millennium chromatography software manager. Mass adsorption and dissipation QCM-D experiments were performed using a Q-Sense E4-Auto system from Biolin Scientific Inc. (Paramus, NJ) at 25 ± 0.1oC. Preparation of MM ligand functionalized surfaces: Cleaning of gold coated QCM-D sensors included: (i) washing with Milli-Q, ethanol, and drying with a stream of nitrogen; (ii) UV/ozone exposure (Novascan Technologies, PSD-UVT, Ames, IA) for 15 minutes at 65oC; (iii) treatment with a 5:1:1 mixture of Milli-Q water, ammonia (25 v/v %) and hydrogen peroxide (30 v/v %) for 10 minutes at 65oC; (iv) rinsing thoroughly with Milli-Q water, ethanol and drying with nitrogen. The sensors were then immersed into ethanol solutions containing the thiol terminated MM ligand linkers, i.e. “Capto ligand” or “Nuvia ligand” (Figure 1), mixed with EG6OH at a total concentration of 1 mM (total thiol) in compositions of 40 or 100% MM ligand (molar basis) for 24 h at RT. The surfaces were then rinsed with ethanol, Milli-Q water, dried with a stream of nitrogen and used immediately for the QCM-D experiments. The “Capto ligand” and “Nuvia ligand” linkers were synthesized and purified according to previously described procedures.32 X-ray photoelectron spectroscopy: X-ray photoelectron spectroscopy (XPS) was employed to determine the atomic composition of the MM ligand functionalized quartz crystal surfaces. A PHI 5400 XPS instrument equipped with a Mg Kα probe beam was used to obtain the XPS spectra (Physical Electronics Inc., Chanhassen, MN). As shown in Table S1, the major peaks of interest were C (1s), O (1s), N (1s), S (2p) and Au (4f). Spectra were obtained over a surface area of 1 µm2. Survey scans with a pass energy of 89.45 eV were first performed to identify elements present on the surface, followed by acquisition of element-specific spectra with a pass energy of 1 eV. Quartz crystal microbalance with dissipation experiments: After the crystals were mounted in the flow cell, they were first equilibrated with 25 mM sodium acetate (pH 5.0) for an hour followed by a high strength buffer wash (2 M sodium thiocyanate (pH 11.0)) for 10 minutes to remove non-specifically adsorbed impurities from the surface. For the protein bind and elute experiments, lyophilized protein was



Page 6 of 21

dissolved in the running buffer (25 mM sodium acetate) at concentrations of 0.2, 0.12, 0.08, 0.04 and 0.02 mg/ml. Experiments were performed on the functionalized surfaces by flowing the running buffer containing dissolved protein over the surfaces for 5 minutes, followed by the wash step (running buffer without the protein), which was then perfused for 6 minutes. A constant flow rate of 500 µL/min was employed during the binding and wash steps. The wash was followed by a regeneration of the MM ligand surface by introducing a step of 2 M sodium thiocyanate (pH 11.0) for 2 minutes at a flow rate of 500 µL/min. A representative QCM-D experiment is in Figure S1. Protein binding /unbinding experiments were also performed in the presence of different mobile phase modifiers. These modifiers included arginine or guanidine at concentrations of 25 and 100 mM in the 25 mM sodium acetate pH 5.0 buffer (note: the pH of the 200 mM arginine and guanidine stock solutions were pre-adjusted to pH 5.0). Results and Discussion Multimodal Surfaces As described in the experimental section, MM SAMs were prepared on gold coated quartz crystals to produce systems that were representative of MM chromatographic surfaces (Figure 1). This was done to enable an investigation of the kinetics of protein interactions with MM surfaces which have been previously shown to create unique selectivity trends for complex protein separations.46 The MM cation exchange ligands employed in this study (Figure 1) were selected to represent two industrially important MM ligand systems, Capto MMC from GE Healthcare and Nuvia cPrime from Bio-Rad Laboratories.32 While the “Nuvia ligand” displays identical chemistry to the ligand employed in the commercial resin material, the “Capto ligand” is missing a thioether moiety at the point of attachment. As described in previous studies, the ligands used in this study were immobilized using an activated ester to attach the ligand head group to the linker, followed by attachment of the linker to the gold surfaces via the thiol moiety.19-22,32 Surfaces with controlled MM ligand densities were prepared by the use of mixed SAMs consisting of a combination of active linkers terminated with the MM ligand head groups and the inert linkers terminated with the hydroxyl group (EG6OH). As shown in the schematic in Figure 1A and 1B the major difference between both surfaces is the exposed aromatic group for the “Capto ligand” system as compared to the geometrically inaccessible aromatic group for the “Nuvia ligand” SAM. Moreover, the more linear structure of the “Nuvia ligand” coupled with its ability to efficiently π-stack could potentially result in a more aligned ligand surface with the charged moiety facing out into solution. Chromatographic systems often employ hydrophilic spacer arms to make the ligand more accessible and to minimize non-specific adsorption to the base polymer 6 ACS Paragon Plus Environment

Page 7 of 21 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


matrix. Therefore to prevent non-specific binding of proteins to the underlying hydrophobic decane surface (necessary for the SAMs assembly), a PEG linker with six monomer units was employed in our systems.25,26,47 To evaluate the chemistry of the functionalized quartz crystal surfaces, XPS was carried out to provide an elemental analysis for the MM SAM surfaces at different ligand densities (Table S1). The results confirm what was expected for both the 100 and 40% ligand densities (details are in the SI). Kinetic parameter estimation Figure 2 and S2A show representative real time protein binding curves under increasing concentrations of the protein in solution. A modified form of the Langmuir isotherm that incorporates a contribution from protein-protein self interactions was employed in this work to facilitate a more precise fitting of the protein binding curves under a wide range of conditions. A kinetic form of the isotherm is presented in the SI. Ligand density and mobile phase modifier effects In order to investigate effects of co-operativity and avidity in these systems, SAMs with two different MM ligand densities (i.e. 40 and 100%) were evaluated. Mobile phase modifier effects were also investigated by carrying out these binding experiments on the functionalized surfaces in the presence of arginine or guanidine over a range of concentrations. A quantitative analysis of binding behavior was carried out by extracting the kinetic rates of adsorption and desorption of two model proteins, αchymotrypsin and cytochrome C on these tailored surfaces. Figure 2 shows representative data for the binding of α-chymotrypsin at various concentrations to the “Capto ligand” (Figure 2A) and the “Nuvia ligand” (Figure 2B) surfaces at 40% ligand density in the presence of increasing concentrations of arginine in solution (0, 25 and 100 mM). The different colors in Figure 2 signify different protein concentrations (‫ܥ‬௜௡ ) and the fits of the model to the data are represented by the dotted lines. As can be seen from Figure 2, the total frequency change, Δ݂ (proportional to total mass adsorbed) onto the surface decreased, in terms of absolute values, with increasing arginine concentration for both surfaces. These results indicate that arginine is playing some role in decreasing the protein binding to the MM surfaces. Interestingly, for most of the sensograms, there was minimal difference for different protein concentrations, indicating that these experiments were being carried out under non-linear adsorption conditions. Importantly, on comparing the response of the two different surfaces to the presence of arginine, it was observed that at the highest arginine concentration (100 mM) only the “Nuvia ligand” surface showed strong protein concentration dependence. This concentration dependent behavior suggests that the affinity of α-chymotrypsin for the “Nuvia ligand” surface in the presence of this elevated arginine concentration decreased to an extent that the binding mechanism



Page 8 of 21

transitioned from the non-linear part of the Langmuir isotherm to the linear region. In contrast, binding to the “Capto ligand” surface was still sufficiently strong to result in minimal protein concentration dependent behavior. Figures 3 and 4 present adsorption and desorption rates of the proteins cytochrome C and αchymotrypsin, respectively, along with the corresponding error bars calculated at 95% confidence interval. Figure 3A shows kinetic adsorption rates for binding of cytochrome C to the “Capto ligand” (Figure 3A) at two different surface ligand densities (i.e. 40 and 100%). As shown in the figure, while the adsorption rate on the “Capto ligand” surface was unaffected by the presence of modifiers at the 100% ligand density, a decrease was observed at a ligand density of 40% in the presence of 0.1 M arginine. In addition, in the presence of 0.025 M guanidine, a slight increase in the adsorption rate was observed at 40% density. The kinetic desorption rates for cytochrome C to the “Capto ligand” are given in Figure 3C. As expected, as the modifier concentrations increased, the desorption rates also increased. In addition, the desorption rates were affected by the presence of modifiers to a greater extent for the 40% ligand density surface as compared to the 100% ligand density. Interestingly, the results also suggest that the desorption rates for the 40% ligand density surface were significantly lower than the 100% ligand density surface in the absence of modifiers and at arginine and guanidine concentrations of 0.025 M. On the other hand, at 0.1 M modifier concentrations, the desorption rates for the 40% ligand densities were comparable (arginine) or enhanced (guanidine) as compared to the 100% ligand density results. The adsorption rates for cytochrome C binding to the “Nuvia ligand” surface (Figure 3B) show a decrease for the 40% surface in the presence of both modifiers at 0.025 M and 0.1 M. However, an interesting behavior was observed for the 100% ligand density surface where in the presence of guanidine, the adsorption rates showed a steady increase in binding, with the rate of adsorption being the highest in the presence of 0.1 M guanidine. Further, for the 100% ligand density surface, 0.025 M arginine produced an increase in adsorption rate which was followed by a marked decrease in the rate of adsorption in the presence of 0.1 M arginine. The kinetic rates of desorption of cytochrome C from the “Nuvia ligand” surface are given in Figure 3D. Interestingly these results suggest that the desorption rates for the 40% ligand density surface were significantly lower than the 100% ligand density surface in the absence of modifiers and in the presence of guanidine. The desorption rates for the 40% ligand density surface in the presence of arginine were either lower (0.025 M arginine) or comparable (0.1 M arginine) to the 100% ligand density surface. Previous work in our lab has investigated the chromatographic behavior of a commercial protein library on a Capto MMC resin system in the absence and presence of various arginine concentrations.48 For the 8 ACS Paragon Plus Environment

Page 9 of 21 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


present study elution salt concentrations obtained with the two proteins cytochrome C and αchymotrypsin are presented in Table 1. The results are presented for two commercial chromatographic resin systems, the Capto MMC from GE Healthcare and Nuvia cPrime from Bio-Rad Laboratories. It was of interest to compare this chromatographic retention data with the kinetic rates of adsorption and desorption obtained from the QCM-D experiments. Since a ligand density of 40% on the Quartz crystal surface corresponds approximately to the chromatographic ligand density on the Capto MMC resin, we focused our comparison on this system. It is important to note that the comparison between the kinetic results with the SAM surfaces and the chromatography system that is presented below is qualitative due to differences in the operating conditions. While the chromatography experiments were carried out under the same arginine or guanidine concentrations as the QCM experiments, the results were reported as the salt concentration that resulted in elution of the protein during a gradient separation. On the other hand, the QCM experiments were carried out under low salt conditions in order to be able to observe the kinetic behavior since at higher salts there was insufficient binding to extract the kinetic parameters. The retention data in Table 1 shows that in the presence of 0.025 M arginine, a minor drop in binding occurred for cytochrome C on the Capto MMC column. A comparison of this behavior with the adsorption and desorption rates for cytochrome C on the 40% ligand density SAM (Figures 3A and 3C), indicate that this minor decrease in retention was likely due to a moderate increase in the desorption rate. At 0.1 M arginine, a more significant decrease in retention was observed for cytochrome C on the Capto MMC column (Table 1). The corresponding kinetic data given in Figure 3A and 3C, clearly indicate that this more pronounced drop in binding affinity was due to a significant decrease and increase in the adsorption and desorption rates, respectively. This represents two distinct kinetic scenarios for these different concentrations of the mobile phase modifier arginine. Data in Table 1 for cytochrome C in the presence of 0.025 M guanidine, showed a marginal drop in binding on the Capto MMC resin. A comparison with the kinetic data (Figures 3A and 3C), indicates that this was likely due to an overshadowing of the moderate increase in adsorption rate (1.4 fold) by a more marked increase in desorption rate (10 fold). In the presence of 0.1 M guanidine there was a clear decrease in binding of cytochrome C to the Capto MMC resin (Table 1). A comparison with the kinetic data in Figure 3A and 3C suggests that this was clearly due to a dramatic increase in the desorption rate, with minimal effect form the adsorption rate change. These results are quite interesting in that they demonstrate that arginine and guanidine can potentially have a very different effect on the adsorption and desorption kinetics of proteins on multimodal systems, even when they exhibit relatively similar retention behavior.



Page 10 of 21

A similar analysis was carried out on the Nuvia cPrime resin and Nuvia ligand SAM systems. Again, our analysis focuses on the results with the 40 % ligand density Nuvia ligand SAM since this corresponds to the ligand density on the Nuvia cPrime resin surface. As can be seen in Table 1, the presence of arginine (0.025 M and 0.1 M) and guanidine (0.025 M and 0.1 M) both resulted in significant drops in cyrochrome C retention as indicated by the reduced elution salt concentration. For both modifiers, the kinetic data shown in Figure 3B and 3D indicated a significant drop in the adsorption rate (3 fold) along with a dramatic increase in the desorption rates (10 fold) at a guanidine concentration of 0.1 M. This suggest that both adsorption and desorption rate processes played an important role for cytochrome C binding in these systems. Figure 4A shows kinetic adsorption rates for binding of α-chymotrypsin to the “Capto ligand” surface at two different surface ligand densities (40 and 100%). As can be seen in the figure, the adsorption rates for the 100% and 40% ligand density surfaces increased in the presence of both guanidine and arginine. The kinetic rates of desorption of α-chymotrypsin from the “Capto ligand” surface at both ligand densities are presented in Figure 4C. As can be seen, while the increase in desorption rate was moderate for the 0.025 M modifier concentrations, in the presence of 0.1 M arginine and guanidine the rates were significantly enhanced. Figure 4B shows the adsorption rates for α-chymotrypsin binding to the “Nuvia ligand” surface. The results were quite different than those obtained with Capto ligand surface. While a minor increase was observed in the presence of 0.025 M arginine and guanidine, a decrease in the adsorption rates occurred at 0.1 M for both modifiers and for both ligand densities. Figure 4D presents the kinetic rates of desorption of α-chymotrypsin from the “Nuvia ligand” surface. As can be seen, while there was minimal change in the desorption rate for the 0.025 M modifier concentrations, a dramatic increase was observed at 0.1 M modifier concentration (200 fold for the arginine concentration and 60 fold for the guanidine concentration). Interestingly, a much greater increase in desorption rates was observed for the 40% ligand density surface as compared to the 100% ligand density surface in the presence of 0.1M arginine. These results are in sharp contrast to the Capto MMC kinetic data, where there was not a marked difference between the 40 % and 100 % ligand density surfaces or between the two modifiers. The salt elution data presented in Table 1 indicates that the binding of α-chymotrypsin to the Capto MMC resin decreased moderately at 0.025 M and more substantially at 0.1 M arginine. These results taken in concert with the kinetic data in Figure 4A and 4C, indicated that although the adsorption rate increased in the presence of arginine, it was the more substantial increase in the desorption rate (12 fold) that resulted in the reduced retention in the resin system under these conditions. In contrast, for guanidine at a 10 ACS Paragon Plus Environment

Page 11 of 21 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


concentration of 0.1 M there was an increase in the salt elution concentration observed for αchymotrypsin. Thus, for this protein although there was an increase in the desorption rate, it was most likely the increase in the adsorption rate (2 fold) that ultimately resulted in the increased retention for this protein in the resin system under these conditions. The retention data in Table 1 indicates that the binding of α-chymotrypsin to the Nuvia cPrime resin decreased moderately in the presence of 0.025 M arginine and substantially in the presence of 0.1 M arginine. These results viewed in concert with the 40% ligand density kinetic data in Figure 4B and 4D, indicated that although the decrease in adsorption rate contributed to the reduced binding in the resin system, it was in fact the significantly enhanced desorption rate in the presence of 0.1 M arginine (200 fold) that strongly controlled the binding under these conditions. Further, data in Table 1 suggests that binding of α-chymotrypsin to the Nuvia cPrime resin decreased moderately in the presence of both 0.025 and 0.1 M guanidine. A comparison between these results and the kinetic rates in Figure 4B and 4D suggests that, since there was a decrease in the adsorption rate (1.5 fold) coupled with an increase in the desorption rate (60 fold), both adsorption and desorption rates played a role in controlling the binding in the Nuvia cPrime system. It is quite interesting that the Capto MMC and Nuvia results are so different with respect to αchymotrypsin in the presence of 0.1 M guanidine. As shown in Figure 5, the front face of α-chymotrypsin has a hydrophobic patch proximal to a strongly negative charged region. This likely makes this region of the protein less likely to bind to multimodal cation exchange materials. In a recent paper49, we have presented a hypothesis based on MD simulations that indicates that the increased retention of αchymotrypsin on Capto MMC in the presence of guanidine is potentially driven by a reduced electrostatic repulsion on this face of the protein, which in turn enables hydrophobic interactions to then occur with the resin surface. Since hydrophobic interactions play a more important role with Capto MMC, it would be more likely for this binding to occur in the Capto MMC system than with the Nuvia resin. As described above, this unexpected behavior in the Capto MMC system at 0.1 M guanidine is associated with a more prominent increase in the adsorption rate in concert with a less pronounced increase in desorption rate for α-chymotrypsin as compared to other conditions. While these results are intriguing, more studies with different protein multimodal systems will be required to in order to establish deeper connections between the kinetic aspects and the various molecular interactions that create selectivity in multimodal chromatography. Conclusions



Page 12 of 21

QCM-D was employed in this study as a label-free detection method for investigating the kinetic processes involved in the interaction of proteins with SAMs of MM ligands. The SAMs were fabricated to mimic the chromatographic ligand surface and to provide a well-defined and controlled system to study MM-protein interactions in a continuous flow mode. In addition, SAMs were synthesized with different ligand densities to further study these protein surface interactions. Quantitative kinetic parameters were extracted by fitting a modified form of the Langmuir isotherm to the adsorption data and kinetic parameters were determined and analyzed for two different proteins and two different MM cation exchange surfaces. A qualitative comparison between chromatographic retention data from column experiments and the rates of adsorption and desorption obtained from the QCM-D experiments was carried out provide insight into the rates that controlled the binding processes in the presence of the mobile phase modifiers arginine and guanidine. The results indicated that the significant drop in binding for cytochrome C on the Capto MMC resin in the presence of arginine was due to a significant decrease and increase in the adsorption and desorption rates, respectively. On the other hand, in the presence of guanidine, the reduced binding was due primarily to a dramatic increase in the desorption rate. These results indicated that arginine and guanidine can have a very different effect on the adsorption and desorption kinetics of proteins on multimodal systems. The results also indicated that the drop in binding for cytochrome C on the Nuvia C prime resin in the presence of both modifiers suggested that both adsorption and desorption rate processes played an important role in this system. For the binding of αchymotrypsin in the presence of arginine, the results indicated that it was primarily the increase in the desorption rate that produced the reduced retentions in the Capto MMC and Nuvia cPrime systems. In contrast, for the binding of α-chymotrypsin in the presence of guanidine, both adsorption and desorption rates played a significant role in controlling the binding, particularly in the Nuvia cPrime system. These results indicate that the kinetic behavior of these systems in the presence of different mobile phase modifiers and different MM ligand chemistries is likely playing an important role in creating selectivity in these multimodal chromatographic systems. Future work will attempt to connect these kinetic results with a more molecular level understanding of the protein interactions with these MM surfaces as well as the influences of mobile phase modifiers in these systems. If successful, this would potentially offer new opportunities for using kinetic information for the design of improved multimodal ligands and processes. For example, one could avoid slow desorption kinetic scenarios which would minimize chromatographic tailing. Further, perhaps selectivity could be improved under certain scenarios by exploiting kinetic differences in the behavior of closely eluting proteins. Finally, a deeper understanding of the effect of kinetics on protein selectivity in multimodal systems may have important implications for the development of selective multimodal membrane and monolith systems, where kinetic limitations can often play a particularly important role. 12 ACS Paragon Plus Environment

Page 13 of 21 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Supporting Information The Supporting Information is available free of charge on the ACS Publications website. Surface characterization method including XPS analysis of MM SAM surfaces; description of the QCM-D method including a representative binding/unbinding experiment and schematic of the binding process; kinetic parameter estimation procedure. Acknowledgements The authors acknowledge funding from the National Science Foundation (CBET-1150039 to SMC, CBET-1122780 to GB), Bio-Rad Laboratories Inc. and the US Department of Energy (DOE DE-FG0209ER16005; DE-SC0006520 to GB). References (1) Melander, W. R.; El Rassi, Z.; Horvath, C. J. Chromatogr. A 1989, 469, 3-27. (2) Hancock, W. S.; Sparrow, J. T. J. Chromatogr. A 1981, 206, 71-82. (3) Mclaughlin, L. W. Chem. Rev. 1989, 89, 309-319. (4) Burton, S. C.; Haggarty, N. W.; Harding, D. R. K. Biotechnol. Bioeng. 1997, 56, 45-55. (5) Burton, S. C.; Harding, D. R. K. J. Chromatogr. A 1997, 775, 39-50. (6) Johansson, B. L.; Belew, M.; Eriksson, S.; Glad, G.; Lind, O.; Maloisel, J. L.; Norrman, N. J. Chromatogr. A 2003, 1016, 21-33. (7) Ghose, S.; Allen, M.; Hubbard, B.; Brooks, C.; Cramer, S. M. Biotechnol. Bioeng. 2005, 92, 665-673. (8) Chen, J.; Tetrault, J.; Zhang, Y.; Wasserman, A.; Conley, G.; Dileo, M.; Haimes, E.; Nixon, A. E.; Ley, A. J. Chromatogr. A 2010, 1217, 216-224. (9) Kaleas, K. A.; Schmelzer, C. H.; Pizarro, S. A. J. Chromatogr. A 2010, 1217, 235-242. (10) Woo, J.; Parimal, S.; Brown, M. R.; Heden, R.; Cramer, S. M. J. Chromatogr. A 2015, 1412, 33-42. (11) Woo, J. A.; Chen, H.; Snyder, M. A.; Chai, Y.; Frost, R. G.; Cramer, S. M. J. Chromatogr. A 2015, 1407, 58-68. (12) Chung, W. K.; Freed, A. S.; Holstein, M. A.; McCallum, S. A.; Cramer, S. M. Proc. Natl. Acad. Sci. U. S. A. 2010, 107, 16811-16816. (13) Holstein, M. A.; Chung, W. K.; Parimal, S.; Freed, A. S.; Barquera, B.; McCallum, S. A.; Cramer, S. M. J. Chromatogr. A 2012, 1229, 113-120. (14) Gray, J. J. Curr. Opin. Struct. Biol. 2004, 14, 110-115. (15) Tibbs Jones, T.; Fernandez, E. J. J. Colloid Interface Sci. 2003, 259, 27-35. (16) McNay, J. L. M.; Fernandez, E. J. Biotechnol. Bioeng. 2001, 76, 224-232. (17) Guo, J.; Carta, G. J. Chromatogr. A 2014, 1356, 129-137. (18) Guo, J.; Zhang, S.; Carta, G. J. Chromatogr. A 2014, 1356, 117-128. (19) Xia, N.; Hu, Y.; Grainger, D. W.; Castner, D. G. Langmuir 2002, 18, 3255-3262. (20) Chen, S.; Liu, L.; Zhou, J.; Jiang, S. Langmuir 2003, 19, 2859-2864. (21) Li, L.; Chen, S.; Jiang, S. Langmuir 2003, 19, 2974-2982. (22) Ostuni, E.; Grzybowski, B. A.; Mrksich, M.; Roberts, C. S.; Whitesides, G. M. Langmuir 2003, 19, 1861-1872. (23) Hostetler, M. J.; Wingate, J. E.; Zhong, C.-J.; Harris, J. E.; Vachet, R. W.; Clark, M. R.; Londono, J. D.; Green, S. J.; Stokes, J. J.; Wignall, G. D.; Glish, G. L.; Porter, M. D.; Evans, N. D.; Murray, R. W. Langmuir 1998, 14, 17-30. 13 ACS Paragon Plus Environment


Page 14 of 21

(24) Templeton, A. C.; Wuelfing, W. P.; Murray, R. W. Acc. Chem. Res. 1999, 33, 27-36. (25) De, M.; You, C.-C.; Srivastava, S.; Rotello, V. M. J. Am. Chem. Soc. 2007, 129, 10747-10753. (26) You, C.-C.; De, M.; Han, G.; Rotello, V. M. J. Am. Chem. Soc. 2005, 127, 12873-12881. (27) Vutukuru, S.; Bethi, S. R.; Kane, R. S. Langmuir 2006, 22, 10152-10156. (28) Vutukuru, S.; Kane, R. S. Langmuir 2008, 24, 11784-11789. (29) Freed, A. S.; Garde, S.; Cramer, S. M. J. Phys. Chem. B 2011, 115, 13320-13327. (30) Parimal, S.; Cramer, S. M.; Garde, S. J. Phys. Chem. B 2014, 118, 13066-13076. (31) Holstein, M. A.; Parimal, S.; McCallum, S. A.; Cramer, S. M. Langmuir 2012, 29, 158-167. (32) Srinivasan, K.; Parimal, S.; Lopez, M. M.; McCallum, S. A.; Cramer, S. M. Langmuir 2014, 30, 1320513216. (33) Parimal, S. Molecular Investigations of Multimodal Ligand-Protein Interactions. Dissertation, Rensselaer Polytechnic Institute, 2014. (34) Höök, F.; Kasemo, B.; Nylander, T.; Fant, C.; Sott, K.; Elwing, H. Anal. Chem. 2001, 73, 5796-5804. (35) Höök, F.; Rodahl, M.; Brzezinski, P.; Kasemo, B. Langmuir 1998, 14, 729-734. (36) Zhang, Y.; Telyatnikov, V.; Sathe, M.; Zeng, X.; Wang, P. G. J. Am. Chem. Soc. 2003, 125, 9292-9293. (37) Turon, X.; Rojas, O. J.; Deinhammer, R. S. Langmuir 2008, 24, 3880-3887. (38) Dutta, A. K.; Belfort, G. Langmuir 2007, 23, 3088-3094. (39) Dutta, A. K.; Nayak, A.; Belfort, G. J. Colloid Interface Sci. 2008, 324, 55-60. (40) Dutta, A. K.; Belfort, G. Sens. Actuators, B 2009, 136, 60-65. (41) Okahata, Y.; Kawase, M.; Niikura, K.; Ohtake, F.; Furusawa, H.; Ebara, Y. Anal. Chem. 1998, 70, 12881296. (42) Furusawa, H.; Takano, H.; Okahata, Y. Anal. Chem. 2008, 80, 1005-1011. (43) Yazawa, K.; Furusawa, H.; Okahata, Y. J. Biol. Chem. 2013, 288, 35969-35981. (44) Tamerler, C.; Oren, E. E.; Duman, M.; Venkatasubramanian, E.; Sarikaya, M. Langmuir 2006, 22, 7712-7718. (45) Rogers, K. Mol. Biotechnol. 2000, 14, 109-129. (46) Sheth, R. D.; Morrison, C. J.; Cramer, S. M. J. Chromatogr. A 2011, 1218, 9250-9259. (47) Hong, R.; Fischer, N. O.; Verma, A.; Goodman, C. M.; Emrick, T.; Rotello, V. M. J. Am. Chem. Soc. 2004, 126, 739-743. (48) Holstein, M. A.; Parimal, S.; McCallum, S. A.; Cramer, S. M. Biotechnol. Bioeng. 2012, 109, 176-186. (49) Parimal, S.; Garde, S.; Cramer, S. M. Biotechnol. Prog. 2017, 33, 435-447.


Page 15 of 21 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 1. Schematic of experimental setup with A. “Capto ligand” and B. “Nuvia ligand” SAMs.



Page 16 of 21

Figure 2. Representative QCM-D binding curves under increasing concentrations of protein (lowest concentration displayed in red and highest in magenta) wherein the fits to the data are represented by dotted lines. Data are presented for binding of α-chymotrypsin to 40% ligand density surfaces of A. the “Capto ligand” and B. the “Nuvia ligand” in the presence of increasing concentrations of arginine in solution (0 , 25 and 100 mM).


Page 17 of 21 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 3. Kinetic adsorption rates for the binding of cytochrome C to A. the “Capto ligand” and B. the “Nuvia ligand” surfaces in the presence of different modifiers. Kinetic rates for desorption of cytochrome C from C. the “Capto ligand” and D. the “Nuvia ligand” surfaces in the presence of different modifiers. Parameters are also reported for binding to surfaces with different ligand densities. Data is reported at two different MM ligand densities of 100% and 40%.



Page 18 of 21

Figure 4. Kinetic adsorption rates for the binding of α chymotrypsin to A. the “Capto ligand” and B. the “Nuvia ligand” surfaces in the presence of different modifiers. Kinetic rates for desorption of cytochrome C from C. the “Capto ligand” and D. the “Nuvia ligand” surfaces in the presence of different modifiers. Parameters are also reported for binding to surfaces with different ligand densities. Data is reported at two different MM ligand densities of 100% and 40%.


Page 19 of 21 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 5. Electrostatic potential (left) and spatial aggregation propensity (right) maps for α-chymotrypsin with the front and back faces shown on the top and bottom rows, respectively (reprinted with permission from Reference 49 © 2017 American Institute of Chemical Engineers).



Page 20 of 21

Table 1. Salt concentration (NaCl) linear gradient elution data (M) for cytochrome C and α-chymotrypsin in the absence and presence of various concentrations of arginine or guanidine from the Capto MMC and Nuvia cPrime columns.

No modifier Capto MMC Nuvia cPrime

Cytochrome C Arginine Guanidine 0.025 M 0.1 M 0.025 M 0.1 M

No modifier

α-Chymotrypsin Arginine Guanidine 0.025 M 0.1 M 0.025 M 0.1 M

0.45

0.35

0.20

0.40

0.30

1.00

0.80

0.55

1.05

1.30

0.58

0.52

0.32

0.55

0.25

0.74

0.64

0.4

0.73

0.65


Page 21 of 21 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


for TOC only


1 Protein Binding Kinetics in Multimodal Systems - ACS Publications

Recommend Documents