Application Note USD2824 Capture of an Acid-sensitive Recombinant Protein from E. coli Lysate using Mixed-Mode Chromatography Summary • This study highlights the use of mixed-mode chromatography sorbents for the capture of a model protein expressed in the periplasm of E. coli (recombinant Glutathione-S-Transferase). rGST was chosen as a relatively difficult protein to purify, due to its lability at acidic pH, limiting its elution conditions and separation from acidic endogenous contaminants from E. coli. Screening of mixed-mode and ion exchange sorbents was performed, and conditions for efficient capture step using HEA HyperCel™ sorbent were determined to protect the biological activity of rGST. • HEA HyperCel mixed-mode sorbent was selected and showed a high selectivity for rGST, with capacity > 10 mg rGST/mL. • A 300-fold scale-up of the separation performance on HEA HyperCel sorbent was achieved. • This approach using mixed-mode sorbents could be extended to other E. coli protein expression models. 1. Introduction The benefits of the unique selectivity and chemical stability offered by mixed-mode chromatography for the purification of monoclonal antibodies (MAbs) are well established. However, unique features of mixed-mode sorbents can also be exploited efficiently to purify “non-antibody” proteins produced in bacterial systems such as E. coli. A significant number of therapeutic proteins are produced in E. coli at large scale with high expression titers. In conventional E. coli systems, proteins are expressed as inclusion bodies, that need to be solubilized to recover the target protein. Periplasmic expression is a convenient alternative, which also offers a more favorable environment than cytoplasmic expression for S-S bridge formation and protein refolding. Nevertheless, the chromatography capture step of such recombinant proteins is highly challenging, since bacterial lysates contain a broad variety of contaminating protein species. An ideal protein capture step should be selective, have high capacity, and preserve the biological activity of the target protein. This study illustrates the power of mixed-mode sorbents (HEA HyperCel) to capture efficiently an acidsensitive protein expressed in the periplasm of E. coli. Mixed-mode chromatography sorbents MEP, HEA and PPA HyperCel sorbents carry synthetic ligands immobilized on a robust and scalable matrix (HyperCel) that confers mechanical stability and excellent stability to harsh cleaning-in-place conditions (1 M NaOH). The ligands include aliphatic (HEA –hexylamine) and aromatic (MEP –mercaptoethylpyridine, PPA – phenylpropylamine) groups. Data reported in this study is focused on the HEA HyperCel sorbent, screened as the best mixed-mode candidate for this application. Figure 1 Ligand Structure and Bind/Elute Mechanism of Mixed-Mode Sorbents Protein adsorption to the ligands is driven mainly by hydrophobic interaction. Desorption is carried out by decreasing the pH of the mobile phase, which creates a positive charge on both the ligands and the bound proteins, which elute by electrostatic charge repulsion. 2 2. Materials and Methods 2.1 E. coli lysate and target protein (rGST) E. coli lysate containing rGST (Glutathione-S-transferase) was prepared by high pressure homogenization of bacteria harvested from a 10 L culture of an E. coli BL21-GST strain. The lysate was prepared in 20 mM Tris-HCl, pH 8.0 in a final volume of 750 mL for a 10 L culture batch. The extract was frozen and each sample thawed was filtered through a 1 µm 47 mm disc (Glass Fiber, 1 µm nominal, Pall Life Sciences PN 66209) and then a 0.2 µm 47 mm disc (Supor® 200, 0.2 µm, Pall Life Sciences PN 60301). The conductivity of the lysate was 4 mS/cm and its pH 7.0. The target protein, rGST has a molecular weight of 27.7 kDa and is acidic, with a pI of 6.7. 2.2 Analytical methods Total Protein quantification was performed using the Coomassie Plus (Bradford) Assay Kit (Thermo Scientific). Quantification of rGST was done by ELISA GST 96-Well Detection Module 1 Kit (GE Healthcare). Analysis of proteins was performed by micro-fluidic electrophoresis in reducing conditions using the 2100 Bioanalyzer platform (Agilent Technologies). rGST biological activity was estimated with fractions adjusted at pH 7.0 using the Glutathione S-Tranferase Assay Kit (Sigma) .The rate of increase in the absorption is directly proportional to the GST activity in the sample. The GST activity in the lysate evaluated was 19.3 U/mL, corresponding to a specific activity of 1.3 U/mg of total protein. 2.3 3. Evaluation of pH stability of rGST stability To evaluate the pH stability of rGST, the pH of the lysate was adjusted by successive additions of 0.1 M HCl. Aliquots of the sample were collected at pH 6.0, 5.0, 4.0 and 3.0 and stored overnight at 4 °C. The fractions were neutralized by addition of 1 M Tris-HCl and rGST was quantified by GST-ELISA in the fractions at different pH. Chromatographic Runs and Columns 3.1 Chromatography Sorbents and Columns Chromatography sorbent screening and laboratory scale purifications were performed on 1 mL prepacked PRC columns of HEA HyperCel and Q HyperCel sorbents (Pall). Scale-up runs were carried out using LRC columns (Pall) with 25 mm and 50 mm ID and 25 cm length columns packed with 65 and 300 mL of HEA HyperCel sorbent for scale-up studies. 3.2 Runs for Dynamic Binding Capacity Evaluation Sorbents were equilibrated in 10 CV of 50 mM sodium phosphate, pH 7.0, 4 mS/cm for HEA HyperCel sorbent and 25 mM Tris-HCl, pH 8.5, 2 mS/cm for Q HyperCel sorbent. Amounts of lysate corresponding to 7 mg of rGST/mL sorbent for Q HyperCel and 16 mg of rGST/mL of HEA HyperCel sorbent were loaded, followed by a 10 CV wash in the same buffer. The flowthrough fractions were collected and their rGST content was evaluated to determine the DBC10%BT (Dynamic Binding Capacity at 10% Breakthrough). 3.3 rGST Capture by Mixed-Mode Chromatography on HEA Hypercel Sorbent The sorbent was equilibrated in 10 CV of equilibration 50 mM sodium phosphate, pH 7.0, 4 mS/cm). An amount of sample corresponding to 7 mg rGST/mL sorbent (60% of the DBC 10%BT) was loaded, followed by a 10 CV wash in the same buffer. Elution was carried out using 15 CV of 50 mM sodium acetate buffer at pH 4.5 or pH 5.0. A strip elution step was performed with 15 CV of 50 mM sodium acetate buffer at pH 3.0, followed by a cleaningin-place (CIP) step of 1 hour contact time using 1 M NaOH. The pH of the sorbent was then decreased using 3 CV of elution buffer and the column was re-equilibrated in 10 CV of 50 mM sodium phosphate, pH 7.0. www.pall.com/biopharm 3 4. RESULTS AND DISCUSSION 4.1 Evaluation of rGST stability as a function of pH A pH sensitivity study of rGST was carried to determine suitable pH conditions for the purification runs. As shown in Figure 2, the rGST concentration in rGST lysate was consistent with the concentration determined in the non modified lysate (pH 7.0) until pH 4.0. However, when the pH of the lysate was lowered down to pH 3.0, only 65% of the initial rGST content was detected. This suggests that rGST is sensitive to pH < 4.0, jeopardizing its detection by ELISA. At pH 3.0, precipitation occurred, therefore further experiments needed to be run at pH > 4.0. Figure 2 pH Stability of rGST. rGST in E. coli lysate % of initial concentration 120 100 80 60 40 20 0 7 6 5 4 3 pH rGST concentrations detected by GST-ELISA are expressed as % of the rGST concentration in the initial lysate at pH7.0. 4.2 Sorbent Screening: Comparison of Mixed-Mode and Anion Exchange Sorbents for rGST Capture Sorbent screening of anion exchangers and mixed-mode sorbents was performed to determine the best candidate for rGST capture. The objective was to obtain the highest DBC and, if possible, without modification of the feedstock. rGST has a slightly acidic pI (6.7), therefore anion exchange chromatography on Q HyperCel sorbent was tested. The interaction of proteins with mixed-mode sorbent like HEA HyperCel sorbent is driven predominantly by hydrophobic interactions in physiological conditions. Therefore, no adjustment of the sample pH nor conductivity was required. In contrast, binding of rGST to the Q HyperCel anion exchanger required pH adjustment to allow binding. The DBCs were evaluated using the conditions shown in Table 1. Table 1 Run Conditions Used for the Evaluation of rGST DBC on Different Anion Exchange and Mixed-Mode Sorbents. 4 Sorbent Load pH Load conductivity (mS/cm) HEA HyperCel Sorbent Q HyperCel Sorbent 7.0 8.5 4 2 As shown in Figure 3, the breakthrough (BT) curve obtained for the anion exchange sorbent is very steep, with a complete saturation reached after loading 3 mg of rGST/mL of sorbent. This suggests that the anion exchanger has a very low capacity for rGST (DBC10%BT=0.8 mg/mL). Two other anion exchange sorbents, Q Ceramic HyperD® and a rigid Q agarose resin, were also tested and the similar low DBCs were obtained in both cases (data not shown). In contrast, the breakthrough curve for the HEA HyperCel mixed-mode sorbent shows a shallower slope, indicating a higher binding capacity (DBC10%BT=11.1 mg/mL). One possible explanation of the very low DBC obtained with the anion exchange sorbent might be that, as most E. coli proteins are acidic, most of them would be charged at the loading pH used, leading to competition with rGST for binding. The different binding mechanism and selectivity of the HEA HyperCel mixedmode sorbent avoids this phenomenon. This indicates that HEA HyperCel sorbent has a real advantage for rGST capture compared to an anion exchanger, with a DBC >10-fold higher. Additionally, no sample adjustment was required. Therefore, HEA HyperCel sorbent was chosen for rGST Capture. Figure 3 rGST Breakthrough Curves of Ion Exchange (Q HyperCel) and Mixed-Mode Sorbents (HEA HyperCel). C/C0 0.8 Q HyperCel 0.6 0.4 HEA HyperCel 0.2 0 0 2 4 6 8 10 12 14 16 rGST loaded (mg/mL resin) DBC at 10% breakthrough 0.8 mg/mL 11.1 mg/mL The ratio C/C0 represents the ratio between the rGST concentration in the considered flow-through fraction and the rGST concentration in the lysate. 4.3 Optimization of Conditions for rGST Capture on HEA HyperCel Sorbent No modification of the lysate was required before loading the E. coli periplasmic extract on HEA HyperCel sorbent, which simplifies the purification process. But elution pH had to be optimized to obtain the highest yield and purity while preserving the protein activity (pH > 4). The conditions applied for elution optimization are shown in Table 2. Table 2 Run Conditions Used for Optimization of rGST Capture Step on HEA HyperCel Sorbent. Load pH Load conductivity (mS/cm) Load (mg/mL) Elution pH 7.0 7.0 4 4 7 7 4.5 5.0 www.pall.com/biopharm 5 As shown in Table 3, elution at pH 4.5 resulted in a 65% recovery of rGST from HEA HyperCel sorbent, but also prompted elution of various contaminants. In contrast, elution at pH 5.0 was more selective and resulted in a higher purity (90%, activity enrichment factor around 5), and was selected for further experiments. Table 3 Optimization of rGST Capture Step on HEA HyperCel Sorbent. Yield (%) rGST Purity (%) rGST Purification (fold) rGST Activity Yield (%) GST Activity Enrichment (fold) 50 90 4.9 51.1 4.7 The chromatogram of the run on HEA HyperCel sorbent in optimal conditions is shown in Figure 4. A large amount of proteins is found in the flowthrough fraction as expected, but important absorbance peaks are also observed in the elution at pH 3.0 and after CIP (1 M NaOH). This suggests that a large amount of proteins is bound to the column during the loading step, but that the pH 5.0 elution step is very selective and leads to a higher target purity (Figure 4B). A GST activity enzymatic assay demonstrated that 51% of the initial activity was recovered in the elution fraction, and enrichment in rGST specific activity of 4.7-fold after capture. This suggested that capture on HEA HyperCel sorbent preserved the protein integrity. Figure 4 Chromatogram and Gel Electrophoresis Profiles of rGST Capture Using HEA HyperCel Sorbent in Optimized Conditions. Flowthrough Elution pH 5 Elution pH 3 CIP pH UV280nm (mAU) 4000 11.0 3000 ys ate El u n tio pH 5 MW 10.0 GST 9.0 8.0 7.0 2000 6.0 5.0 1000 4.0 0 6 l ST rG 0 20 40 60 80 mL rGST Micro-fluidic electrophoresis gel analysis Evaluation of Multiple Successive Cycles of rGST Capture on HEA HyperCel Sorbent The reproducibility and robustness of the purification was assessed by 12 successive chromatographic runs on the same column. No loss of performance was observed (Figures 5 and 6). Figure 5 Overlaid Chromatograms of 12 Successive Runs of rGST Capture on HEA HyperCel Sorbent. pH 5000 UV280nm (mAU) 11.0 4000 10.0 GST 9.0 3000 8.0 7.0 2000 6.0 5.0 1000 0 4.0 3.0 10 20 30 40 50 60 mL As illustrated in Figure 5, the chromatograms obtained during the 12 successive runs were almost indistinguishable. Quantification of rGST in the flowthrough fractions indicated that after 12 runs, 95% of the rGST loaded was still binding (Figure 6A). Further analysis showed that, despite some limited fluctuations (respectively 15 and 13% CV), no clear trend regarding an evolution of rGST yield or purity appeared run after run (Figure 6). This shows that regeneration cycles using 1 M NaOH, that are proven effective for column regeneration and sanitization, do not affect the performance of HEA HyperCel sorbent. Figure 6 Consistency of rGST Purification on HEA HyperCel Sorbent. rGST Bound 100 % of total protein % of load 60 40 20 0 2 4 6 8 Run Number rGST Purity in Elution 100 80 10 12 10 12 80 60 40 20 0 2 4 6 8 Run Number 10 12 rGST Elution Yield 100 80 % of load 4.4 60 40 20 0 2 4 6 8 Run Number www.pall.com/biopharm 7 4.5 Scalability of rGST capture using HEA HyperCel sorbent To demonstrate the scalability of rGST purification using HEA HyperCel sorbent, two scale-up runs of 65 and 300-fold were performed (Figure 7). Figure 7 Chromatograms and Electrophoresis Profiles Obtained During the 65-fold (A and B) and 300-fold (C and D) Scale-up of rGST Capture on HEA HyperCel Sorbent. p A Flowthrough Elution pH 5 Elution pH 3 UV280nm (mAU) l ST rG pH 12.0 4000 GST 3000 h 5 3 ate H H ug ys np np hro t o o i i t t w Elu F lo Elu MW 10.0 8.0 2000 6.0 1000 rGST M 4.0 0 0 C UV280nm (mAU) B 500 1000 1500 Flowthrough C 2000 2500 Elution Elution pH 5 pH 3 3000 1000 2000 3000 4000 5000 ys 3 5 gh pH pH ou thr ion ion t t w u u o MW El Fl El ate 8.0 6.0 4.0 0 l ST rG 10.0 G 1000 D pH GST 2000 Micro-fluidic electrophoresis gel analysis mL 6000 rGST 7000 mL Micro-fluidic electrophoresis gel analysis As shown in Figure 7 A and C, in scaled-up conditions, the run profiles looked very similar to M those obtained at lab scale with 1 mL columns (Figure 4). Many E. coli proteins are eliminated during loading, and the elution fraction contains almost exclusively rGST. The second strip elution performed at pH 3.0 contains a large amount of proteins, The pH 5.0 elution is therefore very selective for rGST. The chromatographic separation performance was scalable 65 and 300-fold (Figure 8). Figure 8 Performance of HEA HyperCel Sorbent for rGST Capture During 65- and 300-fold Scale-up. rGST Purification Factor 5 90 4.5 80 4 70 3.5 60 3 50 2.5 40 2 30 1.5 20 1 10 0.5 0 8 Yield 1 65 Column Volume (mL) 300 0 rGST Purification Factor (fold) Yield (%) 100 5. Conclusions • Mixed-mode chromatography on HEA HyperCel sorbent is an efficient solution for the capture of rGST from E. coli. • Lysate can be loaded directly to the column, simplifying process. • Capture and elution of rGST on HEA HyperCel sorbent is selective and yields to DBC > 10 mg/mL with good purity. • A mild selective elution (pH 5.0) preserves biological activity. • Robustness and a 300-fold scale-up of the separation performance on HEA HyperCel sorbent were also demonstrated. • Efficient regeneration using 1 M NaOH CIP. • The model could be probably extended to other E. coli proteins. Ackowledgements Dr. Magali Toueille, Dr. Jérôme Champagne, Dr. René Gantier, Dr. Sylvio Bengio (Pall Life Sciences Chromatography Group), Dr. John Woodgate (Pall UK) for provision of the E. coli lysate. References • Antibody separation by hydrophobic charge induction chromatography. Boschetti, E., Trends in Biotechnology 2002, 20, 333. • Purification and scale-up of a recombinant heavy chain fragment C of botulinum neurotoxin serotype E in Pichia pastoris GS115. Dux, M.P., et al., J. of Prot. Expression and Pur. 2006, 45, 359. • New challenging chromatographic method for the purification of recombinant proteins fused with maltose-binding proteins in Escherichia coli. Cabanne, C., et al., Poster presented at ESTBB, Bordeaux, France (2008). • The distinctive separation attributes of mixed-mode resins and their application in monoclonal antibody downstream purification process. Chen, J., et al., J. Chromatogr. A 2010, 1217, 216. • Evaluation of Protein Engineering and Process Optimization Approaches to Enhance Antibody Drug Manufacturability. Conley, G.P., et al., Biot. Bioeng. 13 June 2011, Early View. www.pall.com/biopharm 9 Visit us on the Web at www.pall.com/biopharm E-mail us at biopharm@pall.com Corporate Headquarters Port Washington, NY, USA +1 800 717 7255 toll free (USA) +1 516 484 5400 phone biopharm@pall.com e-mail International Offices Pall Corporation has offices and plants throughout the world in locations such as: Argentina, Australia, Austria, Belgium, Brazil, Canada, China, France, Germany, India, Indonesia, Ireland, Italy, Japan, Korea, Malaysia, Mexico, the Netherlands, New Zealand, Norway, Poland, Puerto Rico, Russia, Singapore, South Africa, Spain, Sweden, Switzerland, Taiwan, Thailand, the United Kingdom, the United States, and Venezuela. Distributors in all major industrial areas of the world. To locate the Pall office or distributor nearest you, visit www.pall.com/contact. European Headquarters Fribourg, Switzerland +41 (0)26 350 53 00 phone LifeSciences.EU@pall.com e-mail The information provided in this literature was reviewed for accuracy at the time of publication. Product data may be subject to change without notice. For current information consult your local Pall distributor or contact Pall directly. Asia-Pacific Headquarters Singapore +65 6389 6500 phone sgcustomerservice@pall.com e-mail © 2011, Pall Corporation. Pall, , HyperD, HyperCel, and Supor are trademarks of Pall Corporation. ® indicates a trademark registered in the USA and TM indicates a common law trademark. Filtration.Separation.Solution.SM is a service mark of Pall Corporation. 1/12, PDF, GN11.7196 USD2824