Journal of Chromatography A, 1369 (2014) 64–72

Contents lists available at ScienceDirect

Journal of Chromatography A journal homepage: www.elsevier.com/locate/chroma

Dual-ligand affinity systems with octapeptide ligands for affinity chromatography of hIgG and monoclonal antibody Wei-Wei Zhao a , Qing-Hong Shi a,b , Yan Sun a,b,∗ a Department of Biochemical Engineering and Key Laboratory of Systems Bioengineering of the Ministry of Education, School of Chemical Engineering and Technology, Tianjin University, Tianjin 300072, China b Collaborative Innovation Center of Chemical Science and Engineering (Tianjin), Tianjin 300072, China

a r t i c l e

i n f o

Article history: Received 13 July 2014 Received in revised form 22 September 2014 Accepted 27 September 2014 Available online 7 October 2014 Keywords: Affinity chromatography Octapeptide ligand Dual-ligand system Human immunoglobulin G Monoclonal antibody Molecular simulation

a b s t r a c t This work reports the development of affinity systems with dual octapeptide ligands for affinity adsorption and purification of human IgG (hIgG) and monoclonal antibody (mAb). The three octapeptide ligands, FYWHCLDE (1), FYCHWALE (2), and FYCHTIDE (3), identified earlier by the biomimetic design strategy were used; any two of the three were mixed and coupled to Sepharose gel, leading to the formation of three dual-ligand affinity systems. Research emphasis was first placed on hIgG adsorption isotherms and the results were compared to the three single-ligand affinity systems. It was found that there was synergistic effect of the two peptide ligands in a dual-ligand system, so the affinity of a dual-ligand resin for hIgG was higher than those of its counterparts, single-ligand resins. Of the three dual-ligand systems, the FYWHCLDE (1)–FYCHTIDE (3) resin showed the highest affinity, so it was selected for investigating the effects of ligand density and molar ratio on hIgG adsorption equilibrium. It was found that the synergistic effect increased with increasing the total ligand density of the two peptides in the dual-ligand affinity system. Moreover, the FYWHCLDE (1)–FYCHTIDE (3) system at a molar ratio of 2:1 displayed the highest affinity for hIgG (0.69 ␮M at a total ligand density of 31.1 ␮mol/mL), indicating that the synergistic effect reached the maximum at this ratio. This dual-ligand affinity column was then used for the purification of hIgG and mAb by affinity chromatography, resulting in over 95% pure hIgG and mAb at recovery yield over 90%. Molecular docking of the two peptides to the Fc fragment simultaneously showed that FYWHCLDE (1) stood still but FYCHTIDE (3) shifted aside the CH2 CH3 inter-domain. Molecular dynamics simulation of the binding process of the two octapeptides to Fc revealed that both the peptide ligands kept stable interactions with Fc. The synergistic effect of the dual-ligand affinity system was thus elucidated by the molecular simulations. © 2014 Elsevier B.V. All rights reserved.

1. Introduction The ability to produce substantial quantities of high-purity, safe, and efficacious therapeutic proteins from cell culture of isolated genes is an ongoing challenge for the biotechnology industry [1,2]. Among the therapeutic proteins in the current medicinal market, the demands for antibodies and related proteins are increasing continuously at an astonishing rate, and the antibody-based biopharmaceuticals are used for therapeutics, diagnostics and analytical applications [3]. Nowadays the commercial success and

∗ Corresponding author at: Department of Biochemical Engineering and Key Laboratory of Systems Bioengineering of the Ministry of Education, School of Chemical Engineering and Technology, Tianjin University, Tianjin 300072,China. Tel.: +86 2227404981; fax: +86 22 27403389. E-mail address: [email protected] (Y. Sun). http://dx.doi.org/10.1016/j.chroma.2014.09.083 0021-9673/© 2014 Elsevier B.V. All rights reserved.

dramatic improvements in cell culture productivity have put massive pressure on the downstream processes for the introduction of high throughput, cost-effective, and flexible manufacturing processes. Highly selective techniques, such as affinity chromatography, play a crucial role in downstream processing [4]. The majority of the affinity adsorbents for IgG currently adopted is based on natural biological ligands such as proteins A, G, and L which present high affinity for IgG-Fc and IgG-Fab domains [5]. However, these biological ligands tend to be fragile and extremely expensive to produce and optimize [6]. Thus, a strong effort has been made by manufacturers and researchers to find alternative ligands with improved capacity and chemical stability that could offer similar selectivity at lower cost [1,7,8]. To date, several synthetic affinity ligands have been developed [5,9–11] and among them, peptides are a growing class [12–15]. Peptide ligands have several advantages over protein-based

W.-W. Zhao et al. / J. Chromatogr. A 1369 (2014) 64–72

ligands. For example, they are cost-effective in production, ligand leaching caused by enzymatic degradation can be either eliminated or dramatically reduced, and they can in general better withstand harsh cleaning in place (CIP) conditions such as 0.1 M NaOH [14]. In earlier reports, we presented that the linear octapeptide affinity ligands FYWHCLDE, FYCHWALE, and FYCHTIDE for human IgG (hIgG) identified from a pool of 14 potential candidates by a biomimetic design strategy [16–18], exhibited affinities to selectively bind human polyclonal antibodies (hpAbs) and human monoclonal antibody (hmAb) from serum feed and cell culture supernatant, respectively [16–18]. The three octapeptide ligands could selectively bind to Fc fragment (not Fab fragment) of hIgG [18]. However, it was found that electrostatic interactions contributed mainly to the binding of the three peptides to IgG [16,18], which was different from Protein A that binds IgG by hydrophobic interactions [11]. The cause was probably due to the limitation of the docking software Rosetta FlexPepDock used in the peptide biomimetic design strategy, and the flexibility of the charged Cand N-terminal residues of the peptides as discussed in detail previously [16]. In addition, ligand binding competition showed that the octapeptides could not retain a Protein A-hIgG complex, and IgG could be eluted from the peptide columns using Protein A, indicating the peptide ligands were interacting with the binding sites of IgG similar to those for Protein A [18]. The previous results [18] have also revealed the possibilities of the presence of two binding sites of the peptide ligands to Fc. In addition, though the ligands bind to the binding sites of Fc similar to those for Protein A, their binding sites to Fc may not be exactly the same as each other because of the small ligands as compared to the large binding area of Protein A. These suggest that combinations of two different ligands in a single affinity gel would result in a synergistic binding effect, thus enhancing the binding affinity for Fc. Hence, in this work we have attempted to establish dual-ligand affinity systems by immobilizing any two of the three octapeptides on a porous support, Sepharose gel, at proper molar ratios. Since the synthesized peptides would bind to the Fc fragment of IgG, the dual-ligand systems are expected to show higher affinities than their counterparts, the single-ligand affinity systems. Dissociation constants and equilibrium binding capacities obtained from adsorption isotherms were used for identifying the dual-ligand affinity gel with the highest affinity for hIgG, through comparison with those of the single-ligand gels. Then, the selected dual-ligand affinity gel was studied in detail to probe the influence of ligand density and peptide molar ratio on hIgG adsorption equilibrium. Thereafter, this dual-ligand affinity column was used for the purification of hIgG and mAb by affinity chromatography. Finally, molecular simulations were performed to elucidate the molecular basis for the binding conformation and synergistic effect displayed by the dual-ligand affinity system.

2. Materials and methods 2.1. Materials Thiopropyl Sepharose 6B and HR 5/5 column were the products of GE Healthcare (Uppsala, Sweden). Human IgG (hIgG, polyclonal antibodies) and bovine serum albumin (BSA) in lyophilized form were purchased from Sigma–Aldrich (St. Louis, MO, USA). Human serum, human IgG enzyme-linked immunosorbent assay (ELISA) kit, Micro-BCA assay kit and the reagents for solution preparations were all purchased from Beijing Dingguo Changsheng Biotechnology (Beijing, China). The CHO cell culture supernatant containing human monoclonal IgG (hmAb) was obtained from Jinan University (Guangzhou, China). Pre-stained molecular weight markers were obtained from TransGen Biotech (Beijing, China). The Durapore

65

filters with nominal pore sizes of 0.45 and 0.22 ␮m were supplied by Millipore (Billerica, MA, USA). All buffers and protein samples were filtered with 0.45-␮m and 0.22-␮m filters, respectively, and degassed prior to use. 2.2. Preparation of peptide affinity resins The octapeptides, FYWHCLDE (1), FYCHWALE (2), and FYCHTIDE (3) were synthesized by the solid-phase synthesis method by GL Biochem (Shanghai, China), and the peptides were coupled to Thiopropyl Sepharose 6B by the method described previously [16,18]. It is noted that the molar ratio of different ligands could be exactly controlled because all peptides could be coupled to the gel due to the high coupling reaction efficiency [18]. For convenience, these resins were named as D1/2, D1/3, and D2/3 for the FYWHCLDE (1)–FYCHWALE (2), FYWHCLDE (1)–FYCHTIDE (3), and FYCHWALE (2)–FYCHTIDE (3) affinity systems, respectively. Note that the figure in the parenthesis following each peptide represents the peptide number, and it is used to denote the dual-ligand systems. For instance, D1/3 is the dual-ligand affinity resin with FYWHCLDE (1) and FYCHTIDE (3). The single-ligand affinity resins were also represented by symbols, with S1 for the FYWHCLDE (1) resin, S2 for the FYCHWALE (2) resin, and S3 for the FYCHTIDE (3) resin. All the three single-ligand and three dual-ligand affinity gels were synthesized at a ligand density of 10 ␮mol/g drained gel and the dual-ligand gels were synthesized at a peptide ratio of 1:1 (i.e., 5 ␮mol/g:5 ␮mol/g). Plus, the single FYWHCLDE (1), FYCHTIDE (3) affinity gel, and the dual-ligand affinity gels with equimolar concentrations of the two octapeptides were prepared at total ligand densities of 20 and 30 ␮mol/g. Also, the FYWHCLDE (1)–FYCHTIDE (3) affinity gels (ligand density, 30 ␮mol/g) with peptide ratios of 1:2, 2:1, and 4:1 (FYWHCLDE:FYCHTIDE) were prepared. Finally, this dual-ligand affinity gel (ligand density, 23 ␮mol/g) with a peptide ratio of 2:1 was prepared for purification experiments. The ligand density data were converted into values in the unit of ␮mol/mL with the drained gel density of Sepharose 6B (1.037 g/mL). 2.3. Measurement of adsorption isotherms Measurements of static adsorptions of hIgG included three parts. The first part employed the three single affinity resins and the three dual-ligand affinity resins. This led to the identification of the bestperforming dual-ligand affinity resin through comparison of the adsorption capacity (qm ) and dissociation constant (Kd ), which was determined to be FYWHCLDE (1)–FYCHTIDE (3) affinity resin (D1/3) that presented highest affinity to hIgG. Then, in the second part, this dual-ligand affinity resin was studied in detail to explore the influence of ligand density on hIgG adsorption. The total densities tested were 10.4, 20.7, and 31.1 ␮mol/mL drained gel at a peptide molar ratio of 1:1. In the third part, focus was put on the research into the effect of peptide mixing ratio with D1/3 affinity resin at ligand molar ratios of 1:2, 2:1 and 4:1 (FYWHCLDE (1):FYCHTIDE (3)). The measurement was following the method described previously [16–18]. The previous work [16,18] studied in detail the influence of pH on protein adsorption and it was found that at pH 6.0 FYWHCLDE and FYCHTIDE had the highest binding and selectivity for IgG. Therefore, in order to achieve optimum binding and purification, 20 mmol/L phosphate buffer (PB) (pH 6.0) was used as the equilibration buffer both in the adsorption isotherm measurement as well as in the following affinity purifications of hIgG and mAb from serum and cell culture supernatant, respectively. After equilibration with 20 mmol/L phosphate buffer (PB) (pH 6.0), 10 mg of drained wet gel was mixed with 1.0 mL of protein solution (0.2–4.0 mg/mL) prepared in the same buffer and incubated at 140 rpm and 25 ◦ C for 2 h. The supernatant was collected by

66

W.-W. Zhao et al. / J. Chromatogr. A 1369 (2014) 64–72

centrifugation at 4000 rpm for 5 min. Finally, protein concentration (c, mg/mL) in the supernatant was measured at 280 nm with Lamda 35 UV/vis spectrophotometer (Shelton, CT, USA) and the adsorption density of protein (q) was calculated by mass balance. The Langmuir model (Eq. (1)) was used to describe adsorption isotherms: q=

qm c Kd + c

(1)

where qm is the adsorption capacity (mg-protein/mL-drained gel) and Kd is the dissociation constant. 2.4. IgG and mAb purifications Human IgG and hmAb were purified by affinity chromatography from human serum and cell culture supernatant, respectively. The total concentration of the human serum proteins was as high as 87.2 mg/mL and of high viscosity, so it was 10-fold diluted with 20 mmol/L PB (pH 6.0) (equilibration buffer) prior to loading. It should be noted that the dilution was also necessary to avoid protein aggregation when loading at high concentrations, which would block the binding sites and/or make the pores inaccessible [19]. Therefore, the serum feedstock was diluted in order for smooth purification operations by the affinity chromatography [20,21]. In addition, in order to achieve the optimum purification performance, the pH value of the diluted serum was adjusted by adding 10% phosphoric acid until reaching the optimum binding pH (pH 6.0). Similar dilution and pH adjustment were conducted in the pretreatment of the mAb-containing cell culture supernatant prior to a purification process [17,18] (see below). The 1-mL HR 5/5 column packed with the D1/3 (2:1) affinity resin at a ligand density of 23 ␮mol/g (23.9 ␮mol/mL) was used for IgG purification. The column was first equilibrated with at least 10 column volumes (CVs) of the buffer till a stable UV baseline was reached, and then 6 mL of the diluted human serum (8.72 mg/mL in total protein concentration) was loaded into the column at 0.3 mL/min. The column was subsequently washed with 9 CVs of the equilibration buffer to remove unabsorbed proteins, and then was eluted with a linear gradient of 0–0.5 mol/L NaCl in the equilibration buffer for 20 min at 0.5 mL/min. The effluent was monitored with a UV-900 detector at 280 nm. Finally, the column was regenerated with 0.1 mol/L Gly–HCl buffer (pH 2.4). The flow-through fraction, three elution fractions (first two in 1.0 mL and the following one in 2.0 mL), and the regeneration fraction were collected for the analysis of protein concentration and purity. The above affinity column was also used for hmAb purification. The feedstock was prepared by five times dilution of the cell culture supernatant with the equilibration buffer. Then the pH value of the diluted cell culture supernatant was adjusted by adding 10% phosphoric acid until reaching pH 6.0. The column was first equilibrated with at least 10 CVs of the equilibration buffer till a stable UV baseline was reached, and then 27 mL of the feedstock (1.08 mg/mL in total protein concentration) was loaded onto the column at 0.5 mL/min. The column was subsequently washed with 5 CVs of the equilibration buffer to remove unabsorbed proteins, and then was eluted with 0.5 mol/L NaCl in the equilibration buffer. The following procedures were the same as those described above. 2.5. Analytical methods The peptide ligand immobilization was monitored by the peptide decrease in the reaction solution by reversed-phase high performance liquid chromatography (RP-HPLC) on Agilent 1100 series (Agilent Technologies, Santa Clara, CA) with a Waters C18 reversed-phase column (Waters, Milford, MA, USA) analyzed by UV detection at 220 nm. The identification method of the peptide coupling density was described earlier [18].

Proteins collected from chromatographic peaks were analyzed by 10% SDS-PAGE gel under non-reducing conditions [22]. The electrophoresis was run at a constant current of 25 mA until the dye was approximately 1 cm from the bottom of the gel. The gel was stained with Coomassie Blue R-250 and the electrophoresis image was analyzed by the software Gel-Pro Analyzer 3.1 (Media Cybernetics, MD) to determine the purity of IgG. Total protein concentrations in the feedstock, the flow-through pool, and the elution fractions were determined using a Micro-BCA assay kit [23]. IgG concentrations in the samples were determined by ELISA using a human IgG ELISA kit [24]. Then, the recovery yield of IgG in the purification was calculated from the concentration data. 2.6. Molecular docking and molecular dynamics simulation AutoDock Vina [25] (Vina) was chosen for the docking of two octapeptides, FYWHCLDE (1) and FYCHTIDE (3), to the Fc fragment of hIgG simultaneously. A docking box was constructed covering the CH2 CH3 inter-domain region of the Fc fragment. The spacing was 0.375 A˚ (the length between two grid points), and the box size was 26.5 points in x-dimension, 32 points in y-dimension, and 31 points in z-dimension. The initial all-atom models of octapeptides and Fc fragment were derived from the docking results that were previously obtained with Rosetta Flexpepdock [16]. Ten binding conformations for the two peptides docked to Fc were acquired. MD simulations were performed using the GROMACS 4.5.3 package [26] with GROMOS96 53a6 force field. The top ranking binding conformation of FYWHCLDE-Fc-FYCHTIDE complex from above Vina docking was served as the initial binding conformation for MD simulation. A cutoff of 1.0 nm was used for short-range Coulomb as well as Lennard–Jones interactions. The Particle Mesh Ewald (PME) method [27,28] was used to calculate the long-range electrostatic interactions with a grid-spacing of 0.16 nm and an interpolation order of 4. The simple point charge (SPC) [29] model was used to present water molecule. Na+ and Cl− were added to neutralize the system. The pH value (pH 6.0) was mimicked by different protonation states of the charged residues. For all simulations at pH 6.0, the N terminal and basic residues (i.e., Lys, Arg, and His) were protonated and the C terminal was deprotonated. The acidic residues (i.e., Asp and Glu) were also deprotonated. Na+ was added according to the concentration of the binding buffer (20 mmol/L PB, pH 6.0) [18]. All simulations were performed under the isothermal-isobaric (NPT) ensemble. Temperature (300 K) and pressure (1 atm) were controlled by the v-rescale thermostat [30] and Parrinello–Rahman [31], respectively. An integration time step of 2 fs was used together with the LINCS constraint solver [32] for all covalent bonds. Structures were saved every 50 ps for analysis, resulting in 1000 conformations for each 50-ns simulation. MD simulations were run on a 64-CPU Dawning A620r-F server (Dawning, Tianjin, China). All the simulations were performed at least twice. 3. Results and discussion 3.1. Adsorption isotherms Adsorption isotherms of hIgG on the single-ligand and dualligand affinity resins are given in Fig. 1. Table 1 presents the Langmuir parameters derived by fitting Eq. (1) to the isotherm data in Fig. 1. As shown in Fig. 1 and Table 1, the qm (105.3 mg/mL) of D1/2 was close to that of S1 (104.2 mg/mL), and its Kd was smaller, 4/5 and 1/2 that of S1 and S2, respectively. As for D1/3 and D2/3, both the values of qm were between those of their single-ligand counterparts, while both the values of Kd were smaller than those of their single counterparts at the same ligand density. Smaller Kd values indicate that the binding strengths (affinities) of hIgG

W.-W. Zhao et al. / J. Chromatogr. A 1369 (2014) 64–72

67

Table 1 Capacity and dissociation constant of hIgG adsorption to the three octapeptide–Sepharose gels and three dual-ligand affinity gels at a total ligand density of 10.4 ␮mol/mL. Affinity gel

qm (mg/mL)

Kd (␮M)

qm /Kd (107 g/mol)

FYWHCLDE (1)a FYCHWALE (2)a FYCHTIDE (3)a D1/2b D1/3b D2/3b

104.2 87.6 63.7 105.3 94.3 76.5

3.71 6.13 5.67 2.97 2.34 2.55

2.81 1.43 1.12 3.54 4.03 3.00

a Single-ligand affinity gels (S1, S2 and S3). The figure in the parentheses represents peptide number. b Dual-ligand affinity gels coupling two peptides with the numbers represented in the above parentheses.

larger than those of their single-ligand counterparts, indicating that the binding affinity was improved by coupling two peptide ligands. Table 1 shows that although qm of D1/3 (94.3 mg/mL) was lower than that of D1/2 (105.3 mg/mL), its association constant (1/Kd ) was larger than that of D1/2, leading to the largest qm /Kd value of D1/3 among the three mixed-ligand resins. The qm /Kd value of D1/3 was about 1.5 and 3.6 times larger than those of S1 and S3, respectively. This implies that the synergistic binding effect of these two peptide ligands was the strongest. Hence, D1/3 was chosen for further studies. Effects of ligand density on the adsorption equilibrium of hIgG on D1/3 as well as S1 and S3 were first probed by measuring the isotherms at two more ligand densities, 20.7 and 31.1 ␮mol/mL. The isotherms are shown in Fig. 2, and the fitted Langmuir parameters

Fig. 1. Adsorption isotherms of hIgG onto the affinity resins of FYWHCLDE (S1) (), FYCHWALE (S2) (), FYCHTIDE (S3) (), D1/2 (䊉), D1/3 (), and D2/3 () with a ligand density of 10.4 ␮mol/mL gel. (a) Affinity resins of FYWHCLDE (S1), FYCHWALE (S2) and D1/2; (b) affinity resins of FYWHCLDE (S1), FYCHTIDE (S3) and D1/3; (c) affinity resins of FYCHWALE (S2), FYCHTIDE (S3) and D2/3.

on the dual-ligand affinity resins were enhanced. This might be due to a synergistic binding effect of the dual-ligand systems. The synergistic binding effect means the binding of an IgG molecule by two different peptide ligands at different binding sites, which resulted in bigger binding area and binding affinity (interactions). The presence of two binding sites of FYCHWALE and FYCHTIDE was observed previously [18]. This is also demonstrated in the following molecular simulations (see Section 3.3). The ratio qm /Kd can approximately represent the affinity of a specific affinity resin for the target protein [17]. It is seen in Table 1 that the values of qm /Kd of the dual-ligand affinity resins were all

Fig. 2. Adsorption isotherms of hIgG onto the affinity resins of two different total ligand densities. (a) FYWHCLDE (S1) (), FYCHTIDE (S3) () and D1/3 () with a total ligand density of 20.7 ␮mol/mL gel; (b) FYWHCLDE (S1) (), FYCHTIDE (S3) (䊉) and D1/3 () with a total ligand density of 31.1 ␮mol/mL gel.

68

W.-W. Zhao et al. / J. Chromatogr. A 1369 (2014) 64–72

Table 2 Capacity and dissociation constant of hIgG adsorption to single-ligand (S1, S3) and dual-ligand (D1/3) affinity gels at three total ligand densities. Affinity gel FYWHCLDE (1) FYCHTIDE (3) D1/3 (1:1) FYWHCLDE (1) FYCHTIDE (3) D1/3 (1:1) FYWHCLDE (1) FYCHTIDE (3) D1/3 (1:1)

Ligand density (␮mol/mL)

10.4

20.7

31.1

qm and Kd are summarized in Table 2. As can be seen in Table 2, significant increase in adsorption capacity and decrease in Kd values were observed with increasing ligand density, indicating that the synergistic binding effect of the dual-ligand system increased with increasing ligand density. This is reasonable because the possibility of an IgG molecule binding by two different peptide ligands should increase with increasing ligand density. The increased synergistic binding effect could also be shown by the increase of the qm /Kd values. For example, at the ligand density of 31.1 ␮mol/mL, the qm /Kd (8.45 × 107 g/mol) was 2.2 times that at 10.4 ␮mol/mL. Moreover, at each ligand density, the qm of D1/3 was always between those of its single-ligand counterparts, while the Kd was smaller than any of the single-ligand resins. The qm /Kd value reveals a distinct increase in the affinity of D1/3 as compared to those of the singleligand resins. For instance, at the ligand density of 31.1 ␮mol/mL, the qm /Kd (8.45 × 107 g/mol) of D1/3 was around 1.1 and 1.9 times that of S1 and S3, respectively. It is noted that when the total ligand density increased, the adsorption capacity as well as the affinity constant (1/Kd ) increased also (Table 2). However, it is clear that at the same ligand density, the capacity of D1/3 was much lower than S1, and was closer to the lower capacity resin, S3, at the two high ligand densities. This is considered to be additional evidence of the synergistic binding effect. That is, the synergistic binding effect increases with increasing ligand density, and due to the synergistic binding, the bound IgG amount was controlled by the lower capacity ligand, FYCHTIDE (3). Herein, it should be noted that the resins were prepared at ligand densities ranging from 10.4 to 31.1 ␮mol/mL gel due to the limited density of the 2-pyridyl disulphide group in Thiopropyl Sepharose 6B for the peptide coupling (see Instruction 71-7105-00 AE provided at www.gehealthcare.com). The immobilized peptide ligands should present a random distribution on the resin surface. Based on the specific surface area of Sepharose 6B gel (ca. 30 m2 /mL [33]), the mean distance between any two ligands on the resin surface was estimated to be 6.9–4.0 A˚ in the ligand density range of 10.4–31.1 ␮mol/mL gel. The molecular dimension ˚ [34] was far larger than the mean disof IgG (235 A˚ × 44 A˚ × 44 A) tance between any two ligands, therefore it is convincing that two neighboring ligands in the dual-ligand resin (D1/3) could work synergistically by binding to one IgG molecule at the same time. In order to explore the effect of peptide mixing ratio on IgG adsorption, adsorption isotherms were determined at three more peptide molar ratios for D1/3 at a total ligand density of 31.1 ␮mol/mL. The isotherms are shown in Fig. 3, together with those of S1 and S3 reported previously [18]. The fitted qm and Kd are given in Table 3. It can be seen that when the ratio of FYWHCLDE (1) increased from 1/3 to 4/5, the adsorption capacity increased more than 1.7 times (167.1 mg/mL), which was close to that of S1 (176.4 mg/mL) [17] at the same ligand density. More interestingly, it is found that the Kd showed a minimum value of 0.69 ␮M at the peptide ratio of 2:1. This is considered as further evidence for the synergistic effect of the dual-ligand system in IgG binding. Namely, there was a distribution of the binding sites on IgG, and at the ligand

qm (mg/mL)

Kd (␮M)

qm /Kd (107 g/mol)

104.2 63.7 90.1 130.6 89.3 99.9 176.4 104.4 112.4

3.71 5.67 2.34 2.90 3.60 2.04 2.35 2.32 1.33

2.81 1.12 3.85 4.50 2.48 4.90 7.51 4.50 8.45

Fig. 3. Adsorption isotherms of hIgG onto the D1/3 (total ligand density of 31.0 ␮mol/mL gel) at FYWHCLDE (1)/FYCHTIDE (3) molar ratios of 1:2 (), 1:1 (), 2:1 (), 4:1 (䊉), as well as S1 () and S3 ().

molar ratio of 2:1, their synergistic effect reached the maximum, leading to the highest binding affinity. By considering the smallest Kd value and the relatively high capacity at this peptide ratio, this dual-ligand affinity resin was chosen for the following purification studies. 3.2. Purifications of hIgG and mAb In order to compare with the purification results obtained with S1 column at 23.9 ␮mol/mL [17], D1/3 at FYWHCLDE (1)–FYCHTIDE (3) ratio of 2:1 and a total ligand density of 23.9 ␮mol/mL was adopted for the purification of hIgG from human serum and humanized mAb from cell culture supernatant. Fig. 4 shows a typical purification chromatogram and SDS-PAGE image in hIgG purification from diluted serum. It is clear that the flow-through fraction contained massive albumin and small amount of other impurities but only a little IgG. In the elution peak, only the first 1.0 mL collection (E-1) contained a few highmolecular-weight proteins and other trace impurities, and the left 3 mL elution fractions (E-2 and E-3) contained almost only IgG. As a result, the elution pool of the three collections presented 95% hIgG purity at a recovery of 94%. In the hIgG purification using S1 column Table 3 Capacity and dissociation constant of hIgG adsorption to D1/3 of a total ligand density of 31.1 ␮mol/mL at different ligand molar ratios. FYWHCLDE (1)/FYCHTIDE (3)

qm (mg/mL)

Kd (␮M)

qm /Kd (107 g/mol)

0:1 1:2 1:1 2:1 4:1 1:0

104.4 96.8 112.6 137.9 167.1 176.4

2.32 1.68 1.33 0.69 0.96 2.35

4.50 5.76 8.47 20.0 17.4 7.51

W.-W. Zhao et al. / J. Chromatogr. A 1369 (2014) 64–72

Fig. 4. (a) Affinity chromatography of hIgG purification from human serum using the D1/3 affinity column at 23.9 ␮mol/mL gel at an FYWHCLDE (1)/FYCHTIDE (3) molar ratio of 2:1. Arrow labels: FT, flow-through; E, elution; R, regeneration. (b) SDSPAGE analysis (non-reducing condition) of the flow-through fraction, three elution fractions, and the regeneration fraction. Lanes: M, protein marker; F, feedstock; FT, flow-through; E-1 to E-3, 1st to 3rd elution fractions collected in the elution peak (first two in 1.0 mL and the following one in 2.0 mL) beginning from 15 mL of effluent volume; R, regeneration.

69

Fig. 5. (a) Affinity chromatography of humanized mAb purification from cell culture supernatant using the D1/3 affinity column at 23.9 ␮mol/mL gel at an FYWHCLDE (1)/FYCHTIDE (3) molar ratio of 2:1. The feedstock (27 mL) was loaded at 0.5 mL/min. Arrow labels: FT, flow-through; E, elution; R, regeneration. (b) SDS-PAGE analysis (non-reducing conditions) of flow-through and elution fractions. Lanes: FT, flowthrough fraction collected at the point indicated by the arrow; E, the whole elution peak.

Fig. 6. (a) Binding conformation between peptides and the Fc obtained by AutoDock Vina. The peptides are shown by the NewRibbons model and colored in blue for FYWHCLDE (1) and purple for FYCHTIDE (3), respectively. The entire Fc is shown by the NewCartoon model and colored in metallic orange. (b) Electrostatic interactions between FYWHCLDE (1) and Fc. (c) Electrostatic interactions and hydrophobic interactions between FYCHWALE (2) and Fc. The peptide residues are shown by the CPK mode. The Fc residues are shown by the licorice model. Atoms are colored red for oxygen, green for carbon, blue for nitrogen and white for hydrogen. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

70

W.-W. Zhao et al. / J. Chromatogr. A 1369 (2014) 64–72

at the same condition, massive albumin in the early elution fraction was observed, and the purification resulted in 84% pure hIgG at a recovery of 87% [17]. The hIgG purification using S3 column at the same ligand density was not performed. This is because in the purification of hIgG from serum feedstock under the same condition, FYWHCLDE- and FYCHTIDE-Sepharose columns at a ligand density of 10.4 ␮mol/mL showed similar performance [16,18]. In addition, S3 presented much lower hIgG binding capacity and affinity constant than S1 at the ligand density of 10.4 ␮mol/mL [18] (Table 1). Hence, only the comparison of D1/3 with S1 at the same ligand density was provided. Clearly, D1/3 column showed higher selectivity for hIgG than S1. The higher selectivity of D1/3 was probably due to the high selectivity that FYCHTIDE (3) presented to hIgG in comparison with other two peptide ligands [18]. The higher recovery yield was mainly due to the much higher binding affinity of D1/3 (Kd = 0.69 ␮M) than S1 (Kd = 2.8 ␮M) [17]. Certainly, the higher selectivity of D1/3 was also responsible for the higher recovery yield because less wash was needed with a higher-selectivity adsorbent. Fig. 5 shows the purification of mAb by the affinity chromatography and the SDS-PAGE analysis of the collected fractions. From Fig. 5b, it is seen that the cell culture supernatant contained mAb and many low-molecular-weight proteins. In the flow-through fraction, little mAb was observed, indicating the complete capture of the target protein. Moreover, mAb was eluted in a highly concentrated peak. Analysis with the Gel-Pro Analyzer provided a purity of 97%, and protein concentrations measured with the Micro-BCA kit and hIgG ELISA kit indicated a recovery of 90%. Both the purity and the recovery yield were higher than those obtained by using S1 column at the same condition (90% and 85%, respectively) [17]. Thus, the dual-ligand affinity system presented high affinity and selectivity for the purifications of hIgG and mAb from practical feedstocks. This makes it more promising for antibody purification.

3.3. Molecular docking and MD simulation analysis The simultaneous binding of FYWHCLDE (1) and FYCHTIDE (3) to the Fc fragment of hIgG was first performed by molecular docking with AutoDock Vina. The binding conformation and the involved molecular interactions between the two peptides and Fc are probed and provided in Fig. 6. The binding conformation of single FYWHCLDE or FYCHTIDE with Fc was investigated in the previous study [18] and also provided herein in the Supplementary Material (Fig. S1). Fig. S1 shows that both the peptides bound at the CH2 CH3 inter-domain area of Fc and both interacted with [Lys99] of Fc. In the simultaneous binding of the two peptides, FYWHCLDE and FYCHTIDE also bound around the CH2 CH3 inter-domain area of Fc, as can be seen in Fig. 6a. However, the difference was that the region that FYWHCLDE bound still lay in the same site where single FYWHCLDE bound to Fc. It is seen in Fig. 6a and b that [Glu]-[Lys99] and [Asp]-[Lys97] were still the interaction residues which formed electrostatic interactions dominating in the FYWHCLDE–hIgG complex. By contrast, FYCHTIDE shifted to the side of its previous binding site, and the interaction residues also changed from the previous electrostatic interaction pairs [Glu]-[Lys99] and [His]-[Glu186] to an electrostatic interaction pair [Asp]-[Arg103] and a hydrophobic interaction pair [Phe]-[Ile95] (Fig. 6a and c). So it can be argued that as the two peptides competed for the binding sites on Fc, the higher affinity of the peptide FYWHCLDE for hIgG made it stay where it was, but peptide FYCHTIDE failed in the competition due to its lower affinity, and shifted aside to another binding site. This is consistent with the above equilibrium adsorption results that synergistic binding effect of these two peptides caused higher affinity for hIgG than its counterparts, and the synergistic binding effect was due to different binding sites of the two peptides on hIgG. The

Fig. 7. Time evolution of the atom-atom contact number (ncont ) and minimum distance (dmin ) of (a) FYWHCLDE (1)–Fc and (b) FYCHTIDE (3)–Fc complexes in the 50-ns MD simulation. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

observation of two binding sites for FYCHTIDE also supported this simulation [18]. Supplementary figure related to this article can be found, in the online version, at http://dx.doi.org/10.1016/j.chroma.2014.09.083. The dynamic binding process of Fc fragment with the two octapeptides was then studied by 50-ns MD simulation starting from the binding states represented in Fig. 6a. The structural parameters (the minimum distance and the contact number between the peptides and Fc) were monitored in the MD simulations (Fig. 7). The smaller the minimum distance (dmin ) and the larger the contact number (ncont ), the stronger the affinity between the peptide and the target protein is. As shown in Fig. 7, similar changing trends in dmin values between Fc and the two peptides were observed. Both the dmin had ˚ While small values and slight fluctuations, averaging about 1.8 A. as for the values of ncont , Fig. 7a shows that the ncont of FYWHCLDE fluctuated severely in the first 10 ns of the MD simulation, thereafter averaging around 600 with small fluctuations. The ncont values of FYCHTIDE were relatively stable in the 5–25 ns of the MD simulation, after then, its fluctuation became larger and reached an average value of 450 during the subsequent MD simulation process (Fig. 7b). Larger ncont values of FYWHCLDE reflected higher affinity of FYWHCLDE for hIgG than that of FYCHTIDE, agreeing with results in Table 1. Typical snapshots in the Fc–octapeptide binding process are provided in Fig. 8. As can be seen, the binding states between Fc and the peptides were always maintained (no dissociation) during the whole simulation. Therefore, they were predicated to have stable affinity for Fc. This also supported the speculation that when FYWHCLDE and FYCHTIDE simultaneously bound Fc they might bind at different sites, leading to the synergistic binding effect.

W.-W. Zhao et al. / J. Chromatogr. A 1369 (2014) 64–72

71

Fig. 8. Six snapshots of (a) FYWHCLDE (1)–Fc–FYCHTIDE (3) complex in the MD simulation. The peptides are shown by the NewRibbons model and colored in blue for FYWHCLDE (1) and purple for FYCHTIDE (3). The entire Fc is shown by the NewCartoon model and colored in metallic orange. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

4. Conclusions The three octapeptides, FYWHCLDE (1), FYCHWALE (2), and FYCHTIDE (3), identified previously have been used to establish dual-ligand affinity systems by mixed coupling of any two of them onto Sepharose gel. The binding equilibria of hIgG onto these affinity resins were extensively studied by exploring the effects of ligand density and mixed peptide molar ratio. It was found that there was synergistic effect of the two mixed peptide ligands so the dualligand affinity resins achieved higher binding affinities. The affinity resin with FYWHCLDE (1)–FYCHTIDE (3) (D1/3) showed the highest affinity of the three dual-ligand resins. The binding capacity and affinity also increased with increasing ligand density for D1/3. Most importantly, it was found that there was a proper molar ratio of FYWHCLDE (1) to FYCHTIDE (3) that resulted in the highest affinity (smallest dissociation constant). The proper molar ratio was identified as 2:1, at which Kd value of 0.69 ␮M was achieved. This suggests that the synergistic effect of the two peptide ligands reached the largest at this molar ratio. Therefore, D1/3 at FYWHCLDE (1) to FYCHTIDE (3) ratio of 2:1 was applied to the purification of hIgG and mAb by affinity chromatography. Higher purity and recovery yield than those obtained with the single-ligand counterparts indicate that the dual-ligand affinity resin was advantageous over the single-ligand resins for antibody purification. Molecular docking revealed that when the two peptides simultaneously bound to Fc, FYWHCLDE (1) of higher affinity stayed in the initial binding region while FYCHTIDE (3) shifted to the side.

Stable affinity of the peptides for Fc was observed by MD simulations. The results of molecular simulations were consistent with the experimental findings. The simulations well explained the synergistic effect displayed by the two peptides which bound to Fc at two binding sites.

Acknowledgements This work was supported by the High-Tech Research and Development Program of China from the Ministry of Science and Technology of the People’s Republic of China (no. 2012AA020206), Natural Science Foundation of China (no. 21236005), and the Specialized Research Fund for the Doctoral Program of Higher Education (SRFDP) (no. 20130032110028). The authors are grateful to Prof. Sheng Xiong, Jinan University, Guangzhou, China, for the gift of the cell culture supernatant containing humanized mAb.

References [1] T. Barroso, A. Lourenco, M. Araujo, V.D.B. Bonifacio, A.C.A. Roque, A. AguiarRicardo, A green approach toward antibody purification: a sustainable biomimetic ligand for direct immobilization on (bio)polymeric supports, J. Mol. Recognit. 26 (2013) 662–671. [2] K. Sproule, P. Morrill, J.C. Pearson, S.J. Burton, K.R. Hejnæs, H. Valore, S. Ludvigsen, C.R. Lowe, New strategy for the design of ligands for the purification of pharmaceutical proteins by affinity chromatography, J. Chromatogr. B: Biomed. Sci. Appl. 740 (2000) 17–33.

72

W.-W. Zhao et al. / J. Chromatogr. A 1369 (2014) 64–72

[3] R.Z. Wang, D.Q. Lin, H.F. Tong, H.L. Lu, S.J. Yao, Evaluation of mixed-mode chromatographic resins for separating IgG from serum albumin containing feedstock, J. Chromatogr. B: Anal. Technol. Biomed. Life Sci. 936 (2013) 33–41. [4] A.A. Shukla, B. Hubbard, T. Tressel, S. Guhan, D. Low, Downstream processing of monoclonal antibodies – application of platform approaches, J. Chromatogr. B: Anal. Technol. Biomed. Life Sci. 848 (2007) 28–39. [5] D. Low, R. O’Leary, N.S. Pujar, Future of antibody purification, J. Chromatogr. B: Anal. Technol. Biomed. Life Sci. 848 (2007) 48–63. [6] R.X. Li, V. Dowd, D.J. Stewart, S.J. Burton, C.R. Lowe, Design, synthesis, and application of a Protein A mimetic, Nat. Biotechnol. 16 (1998) 190–195. [7] A.C.A. Roque, M. Taipa, C. Lowe, An artificial protein L for the purification of immunoglobulins and Fab fragments by affinity chromatography, J. Chromatogr. A 1064 (2005) 157–167. [8] A.C.A. Roque, M.A. Taipa, C.R. Lowe, Synthesis and screening of a rationally designed combinatorial library of affinity ligands mimicking protein L from Peptostreptococcus magnus, J. Mol. Recognit. 18 (2005) 213–224. [9] Y.D. Clonis, Affinity chromatography matures as bioinformatic and combinatorial tools develop, J. Chromatogr. A 1101 (2006) 1–24. [10] S. Ghose, B. Hubbard, S.M. Cramer, Evaluation and comparison of alternatives to Protein A chromatography – mimetic and hydrophobic charge induction chromatographic stationary phases, J. Chromatogr. A 1122 (2006) 144–152. [11] A.C.A. Roque, C.S.O. Silva, M.A. Taipa, Affinity-based methodologies and ligands for antibody purification: advances and perspectives, J. Chromatogr. A 1160 (2007) 44–55. [12] H. Yang, P.V. Gurgel, R.G. Carbonell, Hexamer peptide affinity resins that bind the Fc region of human immunoglobulin G, J. Pept. Res. 66 (2005) 120–137. [13] H. Yang, P.V. Gurgel, R.G. Carbonell, Purification of human immunoglobulin G via Fc-specific small peptide ligand affinity chromatography, J. Chromatogr. A 1216 (2009) 910–918. [14] L.N. Lund, P.E. Gustavsson, R. Michael, J. Lindgren, L. Norskov-Lauritsen, M. Lund, G. Houen, A. Staby, P.M. St. Hilaire, Novel peptide ligand with high binding capacity for antibody purification, J. Chromatogr. A 1225 (2012) 158–167. [15] S. Menegatti, M. Hussain, A.D. Naik, R.G. Carbonell, B.M. Rao, mRNA display selection and solid-phase synthesis of Fc-binding cyclic peptide affinity ligands, Biotechnol. Bioeng. 110 (2013) 857–870. [16] W.W. Zhao, F.F. Liu, Q.H. Shi, X.Y. Dong, Y. Sun, Biomimetic design of affinity peptide ligands for human IgG based on protein A–IgG complex, Biochem. Eng. J. 88 (2014) 1–11. [17] W.W. Zhao, Q.H. Shi, Y. Sun, FYWHCLDE-based affinity chromatography of IgG: effect of ligand density and purifications of human IgG and monoclonal antibody, J. Chromatogr. A 1355 (2014) 107–114.

[18] W.W. Zhao, F.F. Liu, Q.H. Shi, Y. Sun, Octapeptide-based affinity chromatography of human immunoglobulin G: comparisons of three different ligands, J. Chromatogr. A 1359 (2014) 100–111. [19] B.V. Ayyar, S. Arora, C. Murphy, R. O’Kennedy, Affinity chromatography as a tool for antibody purification, Methods 56 (2012) 116–129. [20] M. Daoud-Attieh, H. Chaib, C. Armutcu, L. Uzun, A. Elkak, A. Denizli, Immunoglobulin G purification from bovine serum with pseudo-specific supermacroporous cryogels, Sep. Sci. Technol. 118 (2013) 816–822. [21] A.D. Naik, M. Raina, A.M. Lali, AbSep – an amino acid based pseudobioaffinity adsorbent for the purification of immunoglobulin G, J. Chromatogr. A 1218 (2011) 1756–1766. [22] U.K. Laemmli, Cleavage of structural proteins during the assembly of the head of bacteriophage T4, Nature 227 (1970) 680–685. [23] P.K. Smith, R.I. Krohn, G.T. Hermanson, A.K. Mallia, F.H. Gartner, M.D. Provenzano, E.K. Fujimoto, N.M. Goeke, B.J. Olson, D.C. Klenk, Measurement of protein using bicinchoninic acid, Anal. Biochem. 150 (1985) 76–85. [24] E. Engvall, P. Perlmann, Enzyme-linked immunosorbent assay (ELISA) quantitative assay of immunoglobulin G, Immunochemistry 8 (1971) 871–874. [25] O. Trott, A.J. Olson, AutoDock Vina: improving the speed and accuracy of docking with a new scoring function, efficient optimization, and multithreading, J. Comput. Chem. 31 (2010) 455–461. [26] B. Hess, C. Kutzner, D. van der Spoel, E. Lindahl, GROMACS 4: algorithms for highly efficient, load-balanced, and scalable molecular simulation, J. Chem. Theory Comput. 4 (2008) 435–447. [27] T. Darden, D. York, L. Pedersen, Particle mesh Ewald: an N·log(N) method for Ewald sums in large systems, J. Chem. Phys. 98 (1993) 10089. [28] U. Essmann, L. Perera, M.L. Berkowitz, T. Darden, H. Lee, L.G. Pedersen, A smooth particle mesh Ewald method, J. Chem. Phys. 103 (1995) 8577–8593. [29] Y.J. Wu, H.L. Tepper, G.A. Voth, Flexible simple point-charge water model with improved liquid-state properties, J. Chem. Phys. 124 (2006) 024503. [30] G. Bussi, D. Donadio, M. Parrinello, Canonical sampling through velocity rescaling, J. Chem. Phys. 126 (2007) 014101–014107. [31] I.R. Craig, D.E. Manolopoulos, Quantum statistics and classical mechanics: real time correlation functions from ring polymer molecular dynamics, J. Chem. Phys. 121 (2004) 3368–3373. [32] B. Hess, H. Bekker, H.J. Berendsen, J.G. Fraaije, LINCS: a linear constraint solver for molecular simulations, J. Comput. Chem. 18 (1997) 1463–1472. [33] Y. Yao, A.M. Lenhoff, Pore size distributions of ion exchangers and relation to protein binding capacity, J. Chromatogr. A 1126 (2006) 107–119. [34] B.D. Fair, A.M. Jamieson, Studies of protein adsorption on polystyrene latex surfaces, J. Colloid Interface Sci. 77 (1980) 525–534.

Dual-ligand affinity systems with octapeptide ligands for affinity chromatography of hIgG and monoclonal antibody.

This work reports the development of affinity systems with dual octapeptide ligands for affinity adsorption and purification of human IgG (hIgG) and m...
2MB Sizes 0 Downloads 5 Views