Separating proteins with activated carbon

Transcription

1 Separating proteins with activated carbon Matthew T. Stone and Mikhail Kozlov EMD Millipore Corp. 80 Ashby Road, Bedford, MA 01730, USA Supporting Information Supporting information for Figure 1 Experimental procedure. Dextran polymer standards were purchased from Pharmacosmos A/S (Roervangsvej 30, DK-4300, Holbaek, Denmark). To 2.0 L of 50 mm potassium phosphate at ph 7.0 with 10 mg/l of sodium azide was added 6.10 g of 2,000,000 Da dextran (Dextran T200, 500 g, ), which was allowed to dissolve over 24 hours at room temperature with stirring. To this solution was added 1.80 g of 1,000 Da dextran (Dextran T1, 1 kg, ), 1.28 g of 3,500 Da dextran (Dextran T3.5, 500 g, ), 2.60 g of 10,000 Da dextran (Dextran T10, 500 g, ), 0.76 g of 40,000 Da dextran (Dextran T40, 500 g, ), 2.02 g of 70,000 Da dextran (Dextran T70, 500 g, ), and 0.44 g of 500,000 Da dextran (Dextran T10, 100 g, ). The dextrans were allowed to dissolve for another 24 hours at room temperature with stirring. Then the solution was filtered through a 0.22 µm membrane (Stericup -GP 0.22 µm Millipore Express PLUS membrane, 1000 ml, catalogue number: SCGPU11RE, EMD Millipore Corporation, Billerica, MA, 01821, USA). Six 15 ml centrifuge tubes were loaded with 0 mg, 5.0 mg, 10 mg, 20 mg, 40 mg, or 80 mg of Nuchar HD activated carbon (lot number: 1339-R-09, MeadWestVaco Corporation, Richmond, VA, USA). Three of the six tubes were loaded with 10 ml of the stock dextran solution while the other three were loaded with 10 ml of buffer (50 mm potassium phosphate at - S1-

2 ph 7.0 with 10 mg/l of sodium azide) lacking dextran to serve as controls. The tubes were allowed to rotate for 20 hours. The tubes were subsequently subjected to centrifugation and the supernatant solutions were filtered through a 0.22 µm membrane (Millex Syringe Filter Units, Millex -GV, 0.22 µm, PVDF, 33 mm, gamma sterilized, catalogue number: SLGV033RB, EMD Millipore Corporation, Billerica, MA, 01821, USA) in order to remove any activated carbon particles that might remain suspended in solution. The compositions of the resulting solutions were determined by analytical size exclusion chromatography as described in Experimental Section 2.4. Size exclusion chromatograms for the activated carbon treated buffer solutions are shown in Figure S1a. The normalized RI detector response at specific molecular weights for the dextran treated solutions plotted as a function of the volume of activated carbon added are shown in Figure S1b. The Δnormalized RI detector response/δactivated carbon volume values listed in Table S1 are the average of three measurements. Figure S1a. Buffer (50 mm potassium phosphate at ph 7.0 with 10 mg/l of sodium azide) without dextran that was treated with increasing amounts of Nuchar HD activated carbon under static conditions. Note that the control buffer solutions showed a peak that increased with increasing amounts of activated carbon. This peak might be attributed to the leaching of small amounts residual salt from the process of manufacturing the activated carbon. Rinsing the activated carbon with copious amount of water reduced, but did not eliminate, the peak. The peak does not influence any of the molecular weights examined except 1,000 Da. The effect on this peak is < 4% and does not change the overall trend that is observed. - S2-

3 Figure S1b. The normalized RI detector response at 1,000 Da, 3,500 Da, 10,000 Da, 40,000 Da, 70,000 Da, 500,000 Da, and 2,000,000 Da as a function of the volume of activated carbon added. Table S1. The Δnormalized RI detector response/δactivated carbon volume represents the ability of Nuchar HD activated carbon to adsorb a specific molecular weight of dextran. molecular weight (Da) Δnormalized RI detector response/ Δactivated carbon volume 1, ± , ± , ± , ± , ± , ± ,000, ± Supporting information for Figure 2 Experimental procedure. Stock solutions of the individual sulfonated polystyrene standards (Poly(styrene-sulfonate) Na Salt Kit, Mp (Da), 10 x 0.50 gram, code: pss-psskit, Polymer Standards Service-USA, Inc., Warwick, RI 02888) were prepared at a of 1.0 mg/ml by dissolving 40 mg of the polymer in 40 ml of 25 mm Tris-HCl at ph 7.0. The standards had weight average molecular weights of 1,100 Da, 3,610 Da, 6,520 Da, 14,900 Da, 32,900 Da, 63,900 Da, 148,000 Da, 282,000 Da, 666,000 Da, and 976,000 Da. - S3-

4 Three 15 ml centrifuge tubes for each molecular weight were loaded with 12 mg of Nuchar HD activated carbon (lot number: 1339-R-09, MeadWestVaco Corporation, Richmond, VA, USA). Three separate 15 ml centrifuge tubes for each molecular weight were used as controls and no activated carbon was added. To all tubes was added 2.5 ml of buffer (25 mm Tris-HCl at ph) and then the tubes were subjected to vortexing. Next 2.5 ml of the sulfonated polystyrene stock solution at 1.0 mg/ml in Tris-HCl at ph 7.0 was added to all tubes. The resulting solutions had a sulfonated polystyrene of 0.5 mg/ml. The tubes were allowed to rotate for 20 hours at room temperature. The tubes were subsequently subjected to centrifugation and the supernatant solutions were filtered through a 0.22 µm membrane (Millex Syringe Filter Units, Millex -GV, 0.22 µm, PVDF, 33 mm, gamma sterilized, catalogue number: SLGV033RB, EMD Millipore Corporation, Billerica, MA, 01821, USA) in order to remove any activated carbon particles that might remain suspended in solution. The polymer solutions were then analyzed by UV spectroscopy at a wavelength of 230 nm to determine the polymer s, which was used to calculate the activated carbon s polymer binding capacity. The binding capacities listed in Table S2 are the average of three measurements. Table S2. Nuchar HD activated carbon s static binding capacity for an individual sulfonated polystyrene standard under aqueous conditions as the weight average molecular weight was varied. polystyrene's weight average molecular weight activated carbon's polymer binding capacity 1, ± 1.3 3, ± 8.3 6, ± , ± , ± , ± , ± , ± , ± , ± S4-

5 Supporting information for Figure 3 Experimental procedure. Stock solutions were prepared at a of 2.0 mg/ml by dissolving 200 mg of from equine heart ( 95% by SDS-PAGE, product number: C2506, lot number: 84H7135, Sigma-Aldrich Corporation, St. Louis, MO, 63103, USA), from bovine milk ( 85% by PAGE, product number: L5385, lot number: 110M7003V, Sigma-Aldrich Corporation, St. Louis, MO, 63103, USA), lysozyme from chicken egg white ( 98% SDS-PAGE, product number: L4919, lot number: 088K13582, Sigma-Aldrich Corporation, St. Louis, MO, 63103, USA), or albumin from bovine serum ( 98% by agarose gel electrophoresis, product number: A7906, batch number: 038K0668, Sigma-Aldrich Corporation, St. Louis, MO, 63103, USA) in 100 ml of water. The stock solutions were filtered through a 0.22 µm membrane (Stericup -GP 0.22 µm Millipore Express PLUS membrane, 250 ml, catalogue number: SCGPU02RE, EMD Millipore Corporation, Billerica, MA, 01821, USA). MAb I was obtained as a solution from EMD Serono Biodevelopment. It was dialyzed into water to remove buffer salts with dialysis tubing (Standard RC Dialysis Trial Kits, Spectra/Por 1-3, 3.5K MWCO, 54 mm FLAT WIDTH, serial number: , Spectrum Laboratories, Inc. Rancho Dominguez, CA, USA). The dialysis tubing containing 0.5 L of the MAb I solution was submerged in 40 L of water for 24 hours. The dialysis tubing was then moved into a new container containing 40 L of fresh water where it remained submerged for an additional 24 hours. A portion of the dialyzed MAb I solution was then concentrated using Amicon Ultra-15 Centrifugal Filter Units (3 kda, catalogue number: UFC900324, EMD Millipore Corporation, Billerica, MA, 01821, USA). The concentrated portion of the solution was recombined with the rest of the dialyzed MAb I solution. The of monoclonal antibody was determined by measuring its absorbance at 280 nm. The combined solution was then diluted with water to give a stock solution with a of 2.0 mg/ml. The MAb I stock solution was then filtered through a 0.22 µm membrane (0.22 µm Millipore Express PLUS membrane, 1.0 L, catalogue number: SCGPU11RE, EMD Millipore Corporation, Billerica, MA, 01821, USA). Three 15 ml centrifuge tubes for each protein and ph were loaded with 10 mg of Nuchar HD activated carbon (lot number: 1339-R-09, MeadWestVaco Corporation, Richmond, VA, USA). Three separate 15 ml centrifuge tubes for each protein and ph were used as controls and no activated carbon was added. To all tubes was added 2.5 ml of buffer at the appropriate ph - S5-

6 (50 mm sodium acetate for ph 4.0, 5.0, 6.0 or 50 mm Tris-HCl for ph 7.0, 8.0, 9.0) and then the tubes were subjected to vortexing. Next 2.5 ml of the appropriate protein stock solutions at 2.0 mg/ml in water were added to the tubes. The resulting solutions had a protein of 1.0 mg/ml and a buffer of 25 mm. The tubes were allowed to rotate for 20 hours at room temperature. The tubes were subsequently subjected to centrifugation and the supernatant solutions were filtered through a 0.22 µm membrane (Millex Syringe Filter Units, Millex -GV, 0.22 µm, PVDF, 33 mm, gamma sterilized, catalogue number: SLGV033RB, EMD Millipore Corporation, Billerica, MA, 01821, USA) in order to remove any activated carbon particles that might remain suspended in solution. The protein solutions were then analyzed by UV spectroscopy at a wavelength of 280 nm to determine the of the protein, which was used to calculate the activated carbon s protein binding capacity. The binding capacities provided in Table S3 are an average of three measurements. Table S3. Nuchar HD activated carbon s static binding capacity for, α- lactalbumin, lysozyme, BSA, and a monoclonal antibody as the solution ph was varied. solution ph binding capacity for cytochrome C binding capacity for α- lactalbumin binding capacity for lysozyme binding capacity for BSA binding capacity for MAb I ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± 0.4 Supporting information for Figure 4 Experimental procedure. Two 2.0 mg/ml protein stock solutions were prepared by dissolving 200 mg of from bovine milk ( 85% by PAGE, product number: L5385, lot number: 110M7003V, Sigma-Aldrich Corporation, St. Louis, MO, 63103, USA) or 200 mg of lysozyme from chicken egg white ( 98% SDS-PAGE, product number: L4919, lot number: 088K13582, Sigma-Aldrich Corporation, St. Louis, MO, 63103, USA) in 100 ml of water. The protein stock solutions were then filtered through a 0.22 µm membrane (Stericup -GP 0.22 µm - S6-

7 Millipore Express PLUS membrane, 250 ml, catalogue number: SCGPU02RE, EMD Millipore Corporation, Billerica, MA, 01821, USA). Three 15 ml centrifuge tubes for both proteins at each solution condition were loaded with 10 mg of Nuchar HD activated carbon (lot number: 1339-R-09, MeadWestVaco Corporation, Richmond, VA, USA). Three additional 15 ml centrifuge tubes for both proteins at each solution condition were prepared as controls and no activated carbon was added. Then 2.5 ml of the stock buffer solution of the appropriate ph (50 mm sodium acetate for ph 4.0 or 5.0, 50 mm Tris-HCl for ph 9.0) and sodium chloride (0 M, 0.20 M, 0.50 M, 1.0 M and 2.0 M) was added to each tube and the tubes were subjected to vortexing. Then 2.5 ml of the appropriate protein stock solution at 2.0 mg/ml in water was added to the tube giving a solution with a protein of 1.0 mg/ml, a buffer of 25 mm, and a sodium chloride of 0 M, 0.10 M, 0.25 M, 0.50 M, or 1.0 M. The tubes were allowed to rotate for 20 hours at room temperature. The tubes were subsequently subjected to centrifugation and the supernatant solutions were filtered through a 0.22 µm membrane (Millex Syringe Filter Units, Millex -GV, 0.22 µm, PVDF, 33 mm, gamma sterilized, catalogue number: SLGV033RB, EMD Millipore Corporation, Billerica, MA, 01821, USA) in order to remove any activated carbon particles that might remain suspended in solution. The protein solutions were then analyzed by UV spectroscopy at a wavelength of 280 nm to determine the of the protein, which was used to calculate the activated carbon s protein binding capacity. The binding capacities provided in Table S4 are an average of three measurements. Table S4. Nuchar HD activated carbon s static binding capacity for at ph 5.0 and ph 9.0 and for lysozyme at ph 4.0 and ph 9.0 as the sodium chloride was varied. lysozyme sodium chloride (M) binding capacity at ph 5.0 binding capacity at ph 9.0 binding capacity at ph 4.0 binding capacity at ph ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± S7-

8 Supporting information for Figure 5 Experimental procedure. MAb I was obtained as a solution from EMD Serono Biodevelopment. It was dialyzed into water to remove buffer salts with dialysis tubing (Standard RC Dialysis Trial Kits, Spectra/Por 1-3, 3.5K MWCO, 54 mm FLAT WIDTH, serial number: , Spectrum Laboratories, Inc. Rancho Dominguez, CA, USA). The dialysis tubing containing 0.5 L of the MAb I solution was submerged in 40 L of water for 24 hours. The dialysis tubing was then moved into a new container containing 40 L of fresh water where it remained submerged for an additional 24 hours. A portion of the dialyzed MAb I solution was then concentrated using Amicon Ultra-15 Centrifugal Filter Units (3 kda, catalogue number: UFC900324, EMD Millipore Corporation, Billerica, MA, 01821, USA). The concentrated portion of the solution was recombined with the rest of the dialyzed MAb I solution. The of monoclonal antibody was determined by measuring its absorbance at 280 nm. The combined solution was then diluted with water to give a stock solution with a of 10.0 mg/ml. The MAb I stock solution was then filtered through a 0.22 µm membrane (0.22 µm Millipore Express PLUS membrane, 1.0 L, catalogue number: SCGPU11RE, EMD Millipore Corporation, Billerica, MA, 01821, USA). Stock protein mixtures were then prepared by dissolving 200 mg of from bovine milk ( 85% by PAGE, product number L5385, lot number 110M7003V, Sigma-Aldrich Corporation, St. Louis, MO, 63103, USA) or 200 mg of from equine heart ( 95% by SDS-PAGE, product number C2506, lot number 041M7008V, Sigma-Aldrich Corporation, St. Louis, MO, 63103, USA) in 100 ml of 10.0 mg/ml MAb I in water. The stock solutions were then filtered through a 0.22 µm membrane (Stericup -GP 0.22 µm Millipore Express PLUS membrane, 250 ml, catalogue number: SCGPU02RE, EMD Millipore Corporation, Billerica, MA, 01821, USA). Three 15 ml centrifuge tubes for both MAb I-protein solutions at ph 4.0, 5.0, 6.0, 7.0, 8.0, and 9.0 were loaded with 10 mg of Nuchar HD activated carbon (lot number: 1339-R-09, MeadWestVaco Corporation, Richmond, VA, USA). Three 15 ml centrifuge tubes for both MAb I-protein solutions at ph 4.0, 5.0, 6.0, 7.0, 8.0, and 9.0 were used as controls and no activated carbon was added. Then 2.5 ml of buffer at the appropriate ph (50 mm sodium acetate for ph 4.0, 5.0, 6.0 or 50 mm Tris-HCl for ph 7.0, 8.0, 9.0) was added to each tube and the tubes were subjected to vortexing. Next 2.5 ml of the appropriate protein stock solution - S8-

9 containing 10.0 mg/ml of MAb I and 2.0 mg/ml of or was added to each tube. The resulting solutions had s 5.0 mg/ml of MAb I, 1.0 mg/ml of either or, and 25 mm of buffer. The tubes were allowed to rotate for 20 hours at room temperature. The tubes were subsequently subjected to centrifugation and the supernatant solutions were filtered through a 0.22 µm membrane (Millex Syringe Filter Units, Millex -GV, 0.22 µm, PVDF, 33 mm, gamma sterilized, catalogue number: SLGV033RB, EMD Millipore Corporation, Billerica, MA, 01821, USA) in order to remove any activated carbon particles that might remain suspended in solution. The s of the proteins remaining in solution were determined by analytical size exclusion chromatography as described in Experimental Section 2.4. These s were then used to calculate the log reduction value of the lower molecular weight protein impurity from MAb I. The recoveries and LRV provided in Table S5 are an average of three measurements. Table S5. The recovery of MAb I and the LRV of or removed from solutions of MAb I with Nuchar HD activated carbon as the solution ph was varied. solution ph in MAb I Recovery of MAb I LRV of α- lactalbumin in MAb I Recovery of MAb I LRV of % ± 1.6% 0.51 ± % ± 0.4% 0.02 ± % ± 1.3% 0.57 ± % ± 1.1% 0.31 ± % ± 0.6% 0.49 ± % ± 0.8% 0.39 ± % ± 0.8% 0.47 ± % ± 0.2% 0.51 ± % ± 0.8% 0.27 ± % ± 0.7% 0.63 ± % ± 0.8% 0.13 ± % ± 1.0% 1.08 ± Supporting information for Figure 6 Experimental procedure. A stock solution was prepared by dissolving 200 mg of α- lactalbumin from bovine milk ( 85% by PAGE, product number: L5385, lot number: 110M7003V, Sigma-Aldrich Corporation, St. Louis, MO, 63103, USA) and 200 mg of from equine heart ( 95% by SDS-PAGE, product number: C2506, lot number: 84H7135, Sigma-Aldrich Corporation, St. Louis, MO, 63103, USA) in 100 ml of water. The protein stock solutions were then filtered through a 0.22 µm membrane (Stericup -GP 0.22 µm - S9-

10 Millipore Express PLUS membrane, 250 ml, catalogue number: SCGPU02RE, EMD Millipore Corporation, Billerica, MA, 01821, USA). Three 15 ml centrifuge tubes for ph 4.0, 5.0, 6.0, 7.0, 8.0, and 9.0 were loaded with 10 mg of Nuchar HD activated carbon (lot number: 1339-R-09, MeadWestVaco Corporation, Richmond, VA, USA). Three separate 15 ml centrifuge tubes for ph 4.0, 5.0, 6.0, 7.0, 8.0, and 9.0 were used as controls and no activated carbon was added. Then 2.5 ml of buffer at the appropriate ph (50 mm sodium acetate for ph 4.0, 5.0, 6.0 or 50 mm Tris-HCl for ph 7.0, 8.0, 9.0) was added to each tube and the tubes were subjected to vortexing. Next 2.5 ml of the stock protein solution containing 2.0 mg/ml of and 2.0 mg/ml of in water was added to each tube to give a solution with 1.0 mg/ml of, 1.0 mg/ml of, and a buffer of 25 mm. The tubes were allowed to rotate for 20 hours at room temperature. The tubes were subsequently subjected to centrifugation and the supernatant solutions were filtered through a 0.22 µm membrane (Millex Syringe Filter Units, Millex -GV, 0.22 µm, PVDF, 33 mm, gamma sterilized, catalogue number: SLGV033RB, EMD Millipore Corporation, Billerica, MA, 01821, USA) in order to remove any activated carbon particles that might remain suspended in solution. The s of the proteins remaining in solution were determined by analytical reverse-phase chromatography as described in Experimental Section 2.5. These s were then used to calculate the percentage of the two proteins remaining in solution after treatment with activated carbon. The s and percentages provided in Table S6 are an average of three measurements. - S10-

11 Table S6. The and percentage of and remaining after treatment of a 1:1 solution with Nuchar HD activated carbon under static binding conditions as the solution ph was varied. solution ph after treatment after treatment percentage of after treatment percentage of α- lactalbumin after treatment ± ± % ± 0.8% 23% ± 0.8% ± ± % ± 0.1% 24% ± 0.1% ± ± % ± 0.5% 42% ± 0.5% ± ± % ± 0.7% 56% ± 0.7% ± ± % ± 0.5% 73% ± 0.5% ± ± % ± 0.0% 100% ± 0.0% Supporting information for Figure 7 Experimental procedure. Stock solution #1 was prepared by dissolving 400 mg of α- lactalbumin from bovine milk ( 85% by PAGE, product number: L5385, lot number: 110M7003V, Sigma-Aldrich Corporation, St. Louis, MO, 63103, USA) and 40 mg of from equine heart ( 95% by SDS-PAGE, product number: C2506, lot number: 84H7135, Sigma-Aldrich Corporation, St. Louis, MO, 63103, USA) in 40 ml of water. Stock solution #2 was prepared by dissolving 400 mg of from equine heart ( 95% by SDS-PAGE, product number: C2506, lot number: 84H7135, Sigma-Aldrich Corporation, St. Louis, MO, 63103, USA) and 40 mg of from bovine milk ( 85% by PAGE, product number: L5385, lot number: 110M7003V, Sigma-Aldrich Corporation, St. Louis, MO, 63103, USA) in 40 ml of water. The stock solutions were then filtered through a 0.22 µm membrane (Stericup -GP 0.22 µm Millipore Express PLUS membrane, 250 ml, catalogue number: SCGPU02RE, EMD Millipore Corporation, Billerica, MA, 01821, USA). Three 15 ml centrifuge tubes for both stock solution #1 and stock solution #2 at both ph 4.0 and ph 9.0 were loaded with 10 mg of Nuchar HD activated carbon (lot number: 1339-R-09, MeadWestVaco Corporation, Richmond, VA, USA). Three separate 15 ml centrifuge tubes for both stock solution #1 and stock solution #2 at both ph 4.0 and ph 9.0 were used as controls and no activated carbon was added. Then 2.5 ml of buffer at the appropriate ph (50 mm sodium acetate for ph 4.0, 50 mm Tris-HCl for ph 9.0) was added to each tube and the tubes were subjected to vortexing. Next 2.5 ml of the appropriate stock protein solution at 10.0 mg/ml of - S11-

12 the higher protein and 1.0 mg/ml of the lower proteinaceous impurity in water was added to each tube. The resulting solutions had 5.0 mg/ml of the higher protein, 0.5 mg/ml of the lower proteinaceous impurity, and a buffer of 25 mm. The tubes were allowed to rotate for 20 hours at room temperature. The tubes were subsequently subjected to centrifugation and the supernatant solutions were filtered through a 0.22 µm membrane (Millex Syringe Filter Units, Millex -GV, 0.22 µm, PVDF, 33 mm, gamma sterilized, catalogue number: SLGV033RB, EMD Millipore Corporation, Billerica, MA, 01821, USA) in order to remove any activated carbon particles that might remain suspended in solution. The s of the proteins remaining in solution were determined by analytical reverse-phase chromatography as described in Experimental Section 2.5. The protein s, the recovery of the higher protein, and the LRV of the proteinaceous impurity are provided in Tables S7a and S7b are an average of three measurements. Table S7a. The s of and, recovery of, LRV of the removed by the static treatment of a solution composed of 100,00 ppm of in by treatment with Nuchar HD activated carbon. solution ph after treatment after treatment recovery of α- lactalbumin after treatment (ppm) LRV of ± ± % ± 1.3% 113,420 ± ± ± ± % ± 2.1% 13,794 ± ± Table S7b. The s of and, recovery of, LRV of the removed by the static treatment of a solution composed of 100,00 ppm of in by treatment with Nuchar HD activated carbon. solution ph after treatment after treatment recovery of after treatment (ppm) LRV of α- lactalbumin ± ± % ± 7.2% 5,873 ± ± ± ± % ± 3.9% 88,968 ± ± S12-

13 Supporting information for Figure 8 Experimental procedure. A solution was prepared by dissolving 750 mg of α- lactalbumin from bovine milk ( 85% by PAGE, product number L5385, lot number 110M7003V, Sigma-Aldrich Corporation, St. Louis, MO, 63103, USA) and 150 mg of from equine heart ( 95% by SDS-PAGE, product number C2506, lot number 84H7135 Sigma-Aldrich Corporation, St. Louis, MO, 63103, USA) in 75 ml of water. The protein stock solution consisting of 10.0 mg/ml of and 1.0 mg/ml of cytochrome C in water was then filtered through a 0.22 µm membrane (Stericup -GP 0.22 µm Millipore Express PLUS membrane, 250 ml, catalogue number: SCGPU02RE, EMD Millipore Corporation, Billerica, MA, 01821, USA). Three 15 ml centrifuge tubes at both ph 4.0 and ph 9.0 were loaded with either 10 mg of Nuchar HD activated carbon (lot number: 1339-R-09, MeadWestVaco Corporation, Richmond, VA, USA), 10 mg of Darco KB-G activated carbon (lot number: , Norit Americas Inc., Marshall, Texas, USA), or 10 mg of Norit CGP Super activated carbon (lot number: , Norit Americas Inc., Marshall, Texas, USA). Three separate 15 ml centrifuge tubes at ph 4.0 and ph 9.0 were used as controls and no activated carbon was added. Then 2.5 ml of buffer at the appropriate ph (50 mm sodium acetate for ph 4.0, 50 mm Tris-HCl for ph 9.0) was added to each tube and the tubes were subjected to vortexing. Next 2.5 ml of the stock protein solution having 10.0 mg/ml of and 2.0 mg/ml of in water was subsequently added to each tube. This gave a solution with 5.0 mg/ml of, 1.0 mg/ml of, and a buffer of 25 mm. The tubes were allowed to rotate for 20 hours at room temperature. The tubes were subsequently subjected to centrifugation and the supernatant solutions were filtered through a 0.22 µm membrane (Millex Syringe Filter Units, Millex -GV, 0.22 µm, PVDF, 33 mm, gamma sterilized, catalogue number: SLGV033RB, EMD Millipore Corporation, Billerica, MA, 01821, USA) in order to remove any activated carbon particles that might remain suspended in solution. The s of the proteins remaining in solution were determined by analytical reverse-phase chromatography as described in Experimental Section 2.5. The recovery of α- lactalbumin and LRV of the proteinaceous impurity provided in Table S8 are an average of three measurements. - S13-

14 Table S8. The recovery of and LRV of removed by Nuchar HD activated carbon, Darco KB-G activated carbon, or CGP Super activated carbon from a solution of 1.0 mg/ml of and 5.0 mg/ml of at ph 4.0 or ph 9.0. type of activated carbon recovery of α- lactalbumin at ph 4.0 recovery of α- lactalbumin at ph 9.0 LRV of at ph 4.0 LRV of at ph 9.0 Nuchar HD 83% ± 0.14% 93% ± 0.36% ± ± Darco KB-G 89% ± 0.42% 97% ± 0.23% ± ± CGP Super 90% ± 0.42% 97% ± 0.16% ± ± Supporting information for Figure 9 Experimental procedure. A solution was prepared by dissolving 1500 mg of α- lactalbumin from bovine milk ( 85% by PAGE, product number L5385, lot number 110M7003V, Sigma-Aldrich Corporation, St. Louis, MO, 63103, USA) and 150 mg of from equine heart ( 95% by SDS-PAGE, product number C2506, lot number 84H7135 Sigma-Aldrich Corporation, St. Louis, MO, 63103, USA) in 150 ml of water. The protein stock solution consisting of 10.0 mg/ml of and 1.0 mg/ml of cytochrome C in water was then filtered through a 0.22 µm membrane (Stericup -GP 0.22 µm Millipore Express PLUS membrane, 250 ml, catalogue number: SCGPU02RE, EMD Millipore Corporation, Billerica, MA, 01821, USA). The stock solution at ph 4.0 was prepared by mixing a 60 ml portion of the stock solution in water with 60 ml of 50 mm sodium acetate at ph 4.0 to give a solution with 5.0 mg/ml of, 0.5 mg/ml of, and 25 mm sodium acetate at ph 4.0. The stock solution at ph 9.0 was prepared by mixing a 60 ml portion of the stock solution in water with 60 ml of 50 mm Tris-HCl at ph 9.0 to give a solution with 5.0 mg/ml of α- lactalbumin, 0.5 mg/ml of, and 25 mm Tris-HCl at ph 9.0. The stock solutions were then filtered through a 0.22 µm membrane (Stericup -GP 0.22 µm Millipore Express PLUS membrane, 250 ml, catalogue number: SCGPU02RE, EMD Millipore Corporation, Billerica, MA, 01821, USA). Two glass chromatography columns (Omnifit Benchmark Column 10 mm/100 mm, 10 mm diameter, 100 mm length, SKU: 006BCC AF, Diba Industries, Danbury, CT 06810, US) were loaded with 200 mg of Nuchar HD activated carbon (lot number: 1339-R-09, MeadWestVaco Corporation, Richmond, VA, USA) that was slurried in water. The activated - S14-

15 carbon was packed into a column by first pulling the activated carbon slurry through the bottom of column and then flowing water through the top of the column until all of the activated carbon particles formed a tightly pack bed with a volume of 0.8 ml. The columns were equilibrated with a buffer corresponding to the desired ph (25 mm sodium acetate for ph 4.0 or 25 mm Tris- HCl for ph 9.0). Then 100 ml of the buffered protein solution was passed through the activated carbon column at a flow rate of 0.4 ml/min, resulting in a residence time of 2.0 minutes. Eight 12.5 ml fractions were collected. Then an additional 12.5 ml of the equilibration buffer was flowed through the column while a ninth 12.5 ml fraction was collected. The s of the two proteins in the individual fractions and a pooled sample of all nine were determined by analytical reverse-phase chromatography as described in Experimental Section 2.5. The s of and the proteinaceous impurity in the individual fractions and the pooled sample of all nine are provided in Table S9. Table S9. The of and in 12.5 ml fractions that were collected after a solution of 0.5 mg/ml of and 5.0 mg/ml of at ph 4.0 or ph 9.0 was passed through a column of Nuchar HD activated carbon. loading on activated carbon (kg/l) purification at ph 4.0 purification at ph rinse pool including rinse total protein recovered 93% (0.03 LRV) 88% 10% (1.00 LRV) 94% - S15-