"Applied Science, Faculty of"@en . "Chemical and Biological Engineering, Department of"@en . "DSpace"@en . "UBCV"@en . "Kavoosi, Mojgan"@en . "2011-02-16T21:04:12Z"@en . "2007"@en . "Doctor of Philosophy - PhD"@en . "University of British Columbia"@en . "Downstream processing of proteins and other biological products has long been dominated by packed-bed chromatography (Rankin 2003). Despite the generally high cost of the technique, chromatography remains widely used because it offers extraordinarily high resolution under conditions that do not denature or alter the chemistry of the product, an imperative for therapeutic proteins since purity and activity are strict requirements. However, the inability of patients and governments to meet the rising costs of healthcare, particularly the cost of recombinant protein therapeutics, and the sharp increase in competition over the past decade for market share of recombinant-protein based treatments of major illnesses have led to intense downward pressure on the cost of goods, especially for high volume products such as monoclonal antibodies and other recombinant proteins (Morrow 2002). Industry is therefore seeking to develop more cost effective downstream processes, including cheaper and more selective forms of chromatography. Generic affinity chromatography based on affinity-tag technology has the potential to simplify downstream processing by achieving higher yields and purities than conventional modes of chromatography. However, the high cost of current affinity tag technologies, due mainly to the expense of their associated affinity chromatography media, limits their application at production scales. This thesis addresses this problem by reporting on a novel affinity chromatography platform utilizing the family 9 carbohydrate-binding module (CBM9) of xylanase 10A from T. maritima , a new affinity tag that binds to both soluble sugars and insoluble cellulose to permit the highly efficient capture and purification of CBM9-tagged fusion proteins on a very inexpensive cellulose-based affinity media. Development of this technology has required (i) design of a generic CBM9 expression vector for production of chimeric fusions containing an N-terminal CBM9, a linker region containing a suitable processing site at its C-terminus for efficient removal of the affinity tag following affinity purification, and a C-terminal target protein, (ii) development of an effective strategy to design a linker sequence to stably connect the CBM9 tag to the target protein and to permit efficient tag removal through enzyme-catalyzed cleavage, (iii) derivation and validation of a mathematical model to predict binding and elution behavior of CBM9 fusion proteins on a high-capacity cellulose column, (iv) solutions to certain technology scale-up issues, including the synthesis of a mechanically stable stationary phase, and finally, (v) validation of the performance of the technology in terms of product yield, purity and concentration factor. Two bioinformatics-based strategies were developed to successfully identify a linker with improved resistance to endogeneous proteases of the host when compared against the popular poly-glycine based linker. A simple and effective assay was developed to identify the optimal conditions for efficient tag removal post-purification. The technique, based on Luminescence Resonance Energy Transfer (LRET) prescreens a library of linkers and processing enzymes to identify a CBM9-target protein fusion with enhanced processing efficiency. A novel two-zone model (TZM) of pore diffusion is presented to describe the rate of uptake of CBM9 fusion proteins within the stationary phase of the associated affinity chromatography column and thereby provide improved predictions of product breakthrough, including elution behavior from a bacterial lysate feed. Finally, a mechanically stable cellulose-based chromatography media was synthesized to allow preparative-scale purification of recombinant proteins using CBM9. A fixed-effect two-way response surface methodology was used to optimize the concentrations of the two primary reactants, epichlorohydrin and dimethyl sulfoxide (DMSO), required to cross-link the starting material, Perloza(TM) MT100, a compressible cellulose-based chromatography resin. This resulted in a cross-linked affinity chromatography media capable of operating at an order-of-magnitude higher linear velocity than permitted by unmodified MT100. In sharp contrast to MT100, the mechanical stability and purification performance of the cross-linked media are not diminished by scale-up or repeated column use. The results of this thesis thereby provide industry with a ready-made expression vector that can be used to express any target protein as a CBM9 fusion protein and to then inexpensively purify the target recombinant protein at an overall level of performance that is either superior or comparable to current commercially available fusion-tag technologies."@en . "https://circle.library.ubc.ca/rest/handle/2429/31363?expand=metadata"@en . "THE CBM9 FUSION TAG - A NEW TECHNOLOGY FOR INEXPENSIVE PRODUCTION AND AFFINITY PURIFICATION OF RECOMBINANT PROTEINS by Mojgan Kavoosi B.Sc, The University of British Columbia, 1993 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES (Chemical and Biological Engineering) THE UNIVERSITY OF BRITISH COLUMBIA October 2007 \u00C2\u00A9 Mojgan Kavoosi, 2007 Abstract Downstream processing of proteins and other biological products has long been dominated by packed-bed chromatography (Rankin 2003). Despite the generally high cost of the technique, chromatography remains widely used because it offers extraordinarily high resolution under conditions that do not denature or alter the chemistry of the product, an imperative for therapeutic proteins since purity and activity are strict requirements. However, the inability of patients and governments to meet the rising costs of healthcare, particularly the cost of recombinant protein therapeutics, and the sharp increase in competition over the past decade for market share of recombinant-protein based treatments of major illnesses have led to intense downward pressure on the cost of goods, especially for high volume products such as monoclonal antibodies and other recombinant proteins (Morrow 2002). Industry is therefore seeking to develop more cost effective downstream processes, including cheaper and more selective forms of chromatography. Generic affinity chromatography based on affinity-tag technology has the potential to simplify downstream processing by achieving higher yields and purities than conventional modes of chromatography. However, the high cost of current affinity tag technologies, due mainly to the expense of their associated affinity chromatography media, limits their application at production scales. This thesis addresses this problem by reporting on a novel affinity chromatography platform utilizing the family 9 carbohydrate-binding module (CBM9) of xylanase 10A from T. maritima, a new affinity tag that binds to both soluble sugars and insoluble cellulose to permit the highly efficient capture and purification of CBM9-tagged fusion proteins on a very inexpensive cellulose-based affinity media. Development of this technology has required (i) design of a generic CBM9 expression vector for production of chimeric fusions containing an N-terminal CBM9, a linker region containing a suitable processing site at its C-terminus for efficient removal of the affinity tag following affinity purification, and a C-terminal target protein, (ii) development of an effective strategy to design a linker sequence to stably connect the CBM9 tag to the target protein and to permit efficient tag removal through enzyme-catalyzed cleavage, (iii) derivation and validation of a mathematical model to predict ii binding and elution behavior of CBM9 fusion proteins on a high-capacity cellulose column, (iv) solutions to certain technology scale-up issues, including the synthesis of a mechanically stable stationary phase, and finally, (v) validation of the performance of the technology in terms of product yield, purity and concentration factor. Two bioinformatics-based strategies were developed to successfully identify a linker with improved resistance to endogeneous proteases of the host when compared against the popular poly-glycine based linker. A simple and effective assay was developed to identify the optimal conditions for efficient tag removal post-purification. The technique, based on Luminescence Resonance Energy Transfer (LRET) prescreens a library of linkers and processing enzymes to identify a CBM9-target protein fusion with enhanced processing efficiency. A novel two-zone model (TZM) of pore diffusion is presented to describe the rate of uptake of CBM9 fusion proteins within the stationary phase of the associated affinity chromatography column and thereby provide improved predictions of product breakthrough, including elution behavior from a bacterial lysate feed. Finally, a mechanically stable cellulose-based chromatography media was synthesized to allow preparative-scale purification of recombinant proteins using CBM9. A fixed-effect two-way response surface methodology was used to optimize the concentrations of the two primary reactants, epichlorohydrin and dimethyl sulfoxide (DMSO), required to cross-link the starting material, Perloza\u00E2\u0084\u00A2 MT100, a compressible cellulose-based chromatography resin. This resulted in a cross-linked affinity chromatography media capable of operating at an order-of-magnitude higher linear velocity than permitted by unmodified MT100. In sharp contrast to MT100, the mechanical stability and purification performance of the cross-linked media are not diminished by scale-up or repeated column use. The results of this thesis thereby provide industry with a ready-made expression vector that can be used to express any target protein as a CBM9 fusion protein and to then inexpensively purify the target recombinant protein at an overall level of performance that is either superior or comparable to current commercially available fusion-tag technologies. Rankin P: Perilous economics o f the industry. The Source 2003, Sept issue, pp. 5-10. J. Morrow: Economics o f antibody production, Genet Eng News 22 (2002), pp. 1-39 i i i Table of Contents Abstract h Table of Contents iv List of Tables ix List of Figures xi Nomenclature xviii Dimensionless numbers xxi Greek letters .' xxii Acknowledgements xxiv Co-Authorship Statement xxvi 1 Introduction, Background and Thesis Objectives 1 1.1 Overview 1 1.2 Literature Review 3 1.2.1 Downstream Processing in Biotechnology 3 1.2.2 Affinity Chromatography in Preparative-Scale Protein Purification 4 1.2.3 \"Natural\" Affinity Chromatography 5 1.2.4 Affinity Fusion-Tag Technology 6 1.2.5 Factors to Consider in Designing the Affinity Tag System 7 1.2.5.1 The Choice of Peptide Versus Protein Tags 7 1.2.5.2 Position of the Affinity Tag 8 1.2.5.3 Expression of Fusion Proteins in Host Cells 9 1.2.5.4 Tag Removal Post-Purification 10 1.2.5.5 Affinity Tag-Ligand Interaction 11 1.2.6 Affinity Tag Systems 12 1.2.7 Carbohydrate Binding Modules 15 1.2.8 The Family 9 Carbohydrate-Binding Module (CBM9) 17 1.2.9 Cellulose-Based Chromatography Resins 18 1.3 Thesis Objectives 18 1.4 Tables 20 1.5 Figures 24 1.6 References 26 2 Inexpensive One-Step Purification of Polypeptides Expressed in E. Coli as Fusions With the Family 9 Carbohydrate-Binding Module of Xylanase 10A from T. Maritima 44 2.1 Introduction 44 2.2 Materials and Methods 47 2.2.1 Reagents 47 2.2.2 Cloning of CBM9-GFP Fusion Protein 47 2.2.3 Protein Production 48 2.2.4 Affinity Chromatography 49 2.2.5 Tag Cleavage by Factor X a 49 2.2.6 Fluorescence Calibration Curves 50 2.2.7 Measurement of Binding Isotherms 50 2.3 Results and Discussion 51 2.3.1 Binding Isotherms and Thermodynamics 52 2.3.2 Fusion Protein Expression and Stability 53 2.3.3 Affinity Purification on Perloza\u00E2\u0084\u00A2 MT100 Column 54 2.3.4 Column Reusability 55 2.3.5 Removal of the CBM9-S 3Nio-IEGR Affinity Tag Using an Immobilized Factor X a Column 55 2.4 Conclusions 56 2.5 Tables 58 2.6 Figures 60 2.7 References 65 3 Strategy for Selecting and Characterizing Linker Peptides for CBM9-Tagged Fusion Proteins Expressed in E. Coli 68 3.1 Introduction 68 3.2 Materials and Methods 71 3.2.1 Reagents 71 3.2.2 Cloning of CBM9-Linker-FX a -GFP Fusion Proteins 72 3.2.3 Fusion Protein Production and Purification 72 3.2.4 Binding Isotherm Measurement and Analysis 73 3.2.5 Linker Stability Analysis 73 3.2.6 Factor X a Processing Analysis 74 3.2.7 Differential Scanning Calorimetry Studies 75 v 3.2.8 LRET Studies 75 3.3 Results and Discussion 78 3.3.1 Linker Selection and Screening Using MEROPS\u00E2\u0084\u00A2 78 3.3.2 Impact of the Linker on Fusion-Tag Performance 80 3.3.3 Determination of Characteristic Distances Using Luminescence Resonance Energy Transfer 82 3.3.4 Influence of the Linker on the Thermodynamic Stability of CBM9-Linker-FX a-GFP Fusion Proteins 83 3.4 Conclusions 84 3.5 Tables 86 3.6 Figures... 92 3.7 References 98 4 Direct Measurement of the Kinetics of CBM9 Fusion-Tag Bioprocessing Using Luminescence Resonance Energy Transfer (LRET) 107 4.1 Introduction 107 4.2 Methods and Materials 111 4.2.1 Materials I l l 4.2.2 Cloning and purification of fusion protein substrates 111 4.2.3 Preparation of apo-fusion protein 112 4.2.4 Isothermal titration calorimetry studies 112 4.2.5 LRET-based assay of fusion-tag cleavage kinetics 112 4.2.6 Gel-based analysis of EKmax\u00E2\u0084\u00A2 cleavage reaction 115 4.3 Results and Discussion 115 4.3.1 Characterization of energy transfer from bound Tb(III) to GFP 116 4.3.2 Fusion-tag bioprocessing kinetics 117 4.4 Tables 120 4.5 Figures 121 4.6 References 129 5 A Novel Two-Zone Protein Uptake Model for Affinity Chromatography and Its Application to the Description of Elution Band Profiles of Proteins Fused to a Family 9 Cellulose Binding Module Affinity Tag 133 5.1 Introduction 133 5.2 A Proposed Generalized Two-Zone Model of Affinity Chromatography 137 5.3 Two-Zone Model Solution Algorithm 140 vi 5.4 Materials and Methods 141 5.4.1 Chromatographic Media and Reagents 141 5.4.2 Scanning Electron Micrographs 141 5.4.3 Protein Production 142 5.4.4 Equilibrium binding isotherms 142 5.4.5 Confocal Laser Scanning Microscopy (CLSM) 143 5.4.6 Characterization and Application of Perloza\u00E2\u0084\u00A2 MT100/G15 Composite Media Column 144 5.5 Results and Discussion ; 144 5.5.1 Geometric and Sorption Properties of Perloza\u00E2\u0084\u00A2 MTIOO/Sephadex G15 Composite Media 144 5.5.2 Purification of CBM9-GFP on Perloza\u00E2\u0084\u00A2 MT100/G15 Composite Media Column 146 5.5.3 Characterization of Solute Mass Transfer Within Perloza\u00E2\u0084\u00A2 MT100/G15 Composite Columns 146 5.5.4 CLSM-Derived Rates of CBM9-GFP Uptake 150 5.5.5 Simulation of Breakthrough Curves 151 5.5.6 CBM9-GFP Breakthrough from a Clarified Cell Extract Feed 152 5.6 Conclusions 153 5.7 Tables 154 5.8 Figures 157 5.9 References 172 6 A Mechanically Stable Porous Cellulose Media for Affinity Purification of CBM9-Tagged Fusion Proteins 178 6.1 Introduction 178 6.2 Materials and Methods 181 6.2.1 Reagents 181 6.2.2 Cross-Linking of M T 100 181 6.2.3 Hydrodynamic Characterization 182 6.2.4 Measurement of Binding Isotherms 182 6.2.5 Affinity Purification of CBM9-GFP 183 6.3 Results and Discussion 183 6.3.1 Response Surface Methodology to Improve Mechanical Properties 183 6.3.2 Comparison of Sorption Properties 187 vii 6.3.3 Comparison of Column Parameters and Properties 187 6.3.4 Preparative Scale Affinity Purification of CBM9-GFP on CRL100-7... 189 6.3.5 Column Reusability 190 6.4 Conclusions 191 6.5 Tables 192 6.6 Figures 197 6.7 References 208 7 Conclusions and Recommendations 212 7.1 References 216 viii List of Tables Table 1.1 Affinity tag systems described in the literature. Tag size and complementary ligand immobilized on the stationary phase are given. Tags that are currently offered commercially are denoted by * 20 Table 1.2 Estimated costs of several commercially available affinity tag technologies. Prices provided by respective vendors 23 Table 2.1 Langmuir adsorption parameters (equilibrium association constant K a and binding capacity qjm a x) for binding of CBM9 and CBM9-GFP to Perloza\u00E2\u0084\u00A2 MT100 at 4\u00C2\u00B0C. Solvent contains pure protein in high-salt buffer 58 Table 2.2 Binding affinity and capacity of CBM9-GFP on various cellulosic resins ..58 Table 2.3 Summary of purification of CBM9-GFP on Perloza\u00E2\u0084\u00A2 MT100 at 4\u00C2\u00B0C 59 Table 2.4 CBM9-GFP yield and purity for consecutive purification runs through the same Perloza\u00E2\u0084\u00A2 MT100 column 59 Table 3.1 Oligonucleotides used in the construction of CBM9-Linker-IEGR-GFP fusion proteins. Restriction sites are underlined and the Factor X a recognition sequence is in bold 86 Table 3.2 Linker lengths, sequences, and susceptibility to proteolytic cleavage as predicted by MEROPS\u00E2\u0084\u00A2 and confirmed experimentally by LC/MS/MS. . . . 87 Table 3.3 Regressed Langmuir isotherm parameters for binding of CBM9-Linker-IEGR-GFP on Perloza\u00E2\u0084\u00A2 MT100 at pH 7 (4\u00C2\u00B0C). Standard deviations (a) computed from triplicate measurements 88 Table 3.4 Half-lifes for Factor X a cleavage of CBM9-Linker-IEGR-GFP fusion proteins. A l l experiments were conducted in 50 m M potassium phosphate buffer (pH 7, 21 \u00C2\u00B0C) at a fusion protein to Factor X a concentration ratio of [1000] to [1] 89 Table 3.5 Average relative distance of separation R/Rp between the bound terbiums on CBM9 and the fluorescent chromophore of GFP determined by LRET at 21\u00C2\u00B0C 90 Table 3.6 Melting temperatures (Tm) and denaturation enthalpies (A7/cai) for CBM9 on its own and as part of various CBM9-linker-IEGR-GFP fusion proteins 91 Table 4.1 Michaelis Menten kinetic constants and reaction half-lives determined by the LRET-based assay 120 IX Table 5.1 Measured Langmuir isotherm parameters for binding of CBM9-GFP and each of its fusion partners to Perloza\u00E2\u0084\u00A2 MT100, Sephadex G15, and the composite MT1007G15 media at 4 \u00C2\u00B0C. The solvent consisted of 50 mM potassium phosphate, 150 mM NaCl, pH 7. N B indicates that no binding was observed 154 Table 5.2 Yield, purity and concentration factor for the affinity purification of CBM9-GFP on a MT100/G15 composite column and on a pure MT100 column. Clarified E. coli cell lysate containing CBM9-GFP was loaded onto each column at a superficial velocity of 4.25 x 10\"3 cm s\"1. Fractions were collected and analyzed both by absorbance at 280 nm and by fluorescence intensity to obtain reported data. The total lysate volume loaded was 65 mL and contained a CBM9-GFP concentration of 5.7 u M 155 Table 5.3 Measured mass-transfer and column-geometry parameters for CBM9-GFP transport in a Perloza\u00E2\u0084\u00A2 MT100/G15 composite media column 156 Table 6.1 Cross-linked resins and reaction conditions 192 Table 6.2 Results of fitting quadratic response surface models to the data set reported in Table 6.1 193 Table 6.3 Equilibrium adsorption parameters for MT100 and CRL100-7 affinity media determined from two independent experiments 194 Table 6.4 Column properties and parameters 195 Table 6.5 Yield and purity for- consecutive purification cycles of CBM9-GFP on a CRL 100-7 column 196 List of Figures Figure 1.1 Schematic representation of the domain structure of Cellulomonas fimi cellulases and xylanases. Note: the domain structure of xylanase 10A of T. maritima is homologous to that of XynlOA of C. fimi with the family 2a CBM9 substituted by CBM9 24 Figure 1.2 Backbone secondary structure of CBM9 from T. maritima, presented in ribbon diagram form, determined from the x-ray crystallographic data of Notenboom (Notenboom et al. 2001). Trp71 and Tip 175 (solid arrow) involved in ligand binding, and Glu77, Gln96, Arg98, Glnl51, Argl61, and Asnl72 (dashed arrow), involved in hydrogen bonding, with the ligand are shown 25 Figure 2.1 Schematic representation of gene fragment coding for the CBM9-S3N10-IEGR-GFP fusion protein 60 Figure 2.2 Equilibrium adsorption isotherms for binding of CBM9 and CBM9-GFP to Perloza\u00E2\u0084\u00A2 MT100 at 4\u00C2\u00B0C. CBM9-GFP binding to Perloza\u00E2\u0084\u00A2 MT100 at 4\u00C2\u00B0C in high salt buffer (solid circle). CBM9 binding at 4\u00C2\u00B0C in high-salt buffer (open circle), where qi is the bound protein concentration and Cj is the equilibrium concentration of protein free in solution 60 Figure 2.3 Proteolytic stability of the S3N10 linker in CBM9-GFP. 12% SDS-PAGE of CBM9-GFP purified with Perloza\u00E2\u0084\u00A2 MT100 in a small batch system. PMSF treated cell extract containing CBM9-GFP was mixed end-over-end, washed with buffer and desorbed with 1-M glucose in TBS8. Despite its molecular weight of 22 kDa, CBM9 runs as a 26-31 kDa protein depending on the length of the linker fragment attached to it (Wassenberg etal. 1997) 61 Figure 2.4 Chromatogram of CBM9-GFP purification from an E. coli BL21 clarified cell lysate on Perloza\u00E2\u0084\u00A2 MT100 at 4\u00C2\u00B0C. 50 mL of clarified cell extract was loaded at 0.2 mL min\"1 on a 17 mL column packed with Perloza\u00E2\u0084\u00A2 MT100 resin, and then washed with 10 column volumes (CV) high salt buffer and 5 C V low salt buffer. Bound fusion protein was desorbed with 1 M glucose in TBS8. 10 mL fractions were collected and analyze by fluorescence (509 nm) (solid circle) and absorbance at 280 nm (line) 62 Figure 2.5 SDS-PAGE documentation of the affinity purification of CBM9-GFP. 12% SDS-PAGE of CBM9-GFP purified on a 17 mL Perloza\u00E2\u0084\u00A2 MT100 column. A l l samples dissolved in sample buffer containing 10%> SDS. Lane M : molecular mass markers in kg mol\"1. Lane 1: clarified cell extract prior to column loading. Lane 2: column flow through. Lane 3: high salt wash. Lane 4: low salt wash. Lane 5: pure CBM9-GFP eluted in TBS8 xi containing 1-M glucose. Lane 6: purified GFP after affinity-tag removal by immobilized Factor X a 63 Figure 2.6 Time course of CBM9-GFP cleavage by Factor X a at 23\u00C2\u00B0C as shown on a 12% SDS-PAGE. A fusion protein to Factor X a concentration ratio of 1000:1 was used 64 Figure 3.1 Schematic representation of the CBM9-Linker-FX a -GFP fusion protein. F X a indicates the recognition site (IEGR) for Factor X a processing 92 Figure 3.2 Intracellular Expression of CBM9-Linker-FX a -GFP fusion proteins in E. coli. Cells expressing a CBM9-Linker-FX a -GFP fusion protein were grown at 30\u00C2\u00B0C to an OD620 of -0.9 and protein expression was induced with 0.1 m M IPTG. CBM9-Linker-FX a -GFP expression was continuously monitored at 510 nm with excitation at 400 nm 92 Figure 3.3 Long-term proteolytic stability of the linkers in CBM9-Linker-FX a -GFP fusion proteins. Clarified cell lysates were incubated at 23\u00C2\u00B0C for 4 hours and then the intact and degraded CBM9-Linker-FX a -GFP fusion proteins were purified on Perloza\u00E2\u0084\u00A2 MT100 in the absence of protease inhibitors. Each lane represents proteins eluted off the Perloza\u00E2\u0084\u00A2 MT100 resin with elution buffer containing 3 M glucose. Lane M contains the molecular weight markers 93 Figure 3.4 Equilibrium adsorption isotherm for binding of CBM9 -G3 - IEGR-GFP to Perloza\u00E2\u0084\u00A2 MT100. Isotherm measured at 4\u00C2\u00B0C in high salt buffer (pH7; 50 m M potassium phosphate, 1 M NaCl) 94 Figure 3.5 F X a cleavage kinetics. CBM9-P-IEGR-GFP fusion protein was incubated at 23 \u00C2\u00B0C with Factor X a at a fusion protein/FXa concentration ratio of 1000:1. Samples were taken at each time point indicated and analyzed by SDS-PAGE 95 Figure 3.6 Sensitized fluorescence emission lifetime data for CBM9-(PT)2P-IEGR-GFP at 23\u00C2\u00B0C. Solid line indicates fit to the first-order exponential decay equation (r2 = 0.999) 96 Figure 3.7 Differential scanning calorimetry thermograms for CBM9, GFP, CBM9-P-IEGR-GFP and CBM9-(PT) 7P-IEGR-GFP in 50 m M potassium phosphate buffer (pH 7). Area under the denaturation peak(s) provides the denaturation enthalpy through the relationship A H c a l = | C p ( T ) d T , where Cp is the heat capacity 97 Figure 4.1 Schematic representation of the terbium bound fusion protein used for the LRET-based peptidase assay (not to scale). Excitation of bound Tb(III) at 235 nm results in the transfer of energy at 490 nm to GFP, leading to the xii emission of a signal at 510 nm. serine peptidase, enterokinase... D4K indicates the recognition site for the 121 Figure 4.2 Individual spectrum of donor and acceptor. Solid line is the emission spectrum of CBM9-bound terbium (donor) obtained after excitation at 235 nm and a post-excitation delay of 200 us. Dashed line is the absorbance spectrum of GFP (acceptor). The spectral overlap (480 nm to 505 nm) between CBM9-bound terbium and GFP allows for energy transfer to take place 122 Figure 4.3 Emission spectra of CBM9-bound terbium (donor domain only) (solid line) and Tb 3 +-CBM9-(PT) 2PID4K-GFP fusion protein (dashed line). Samples were excited at 235 nm followed by a 200 us delay prior to measurement of emission signal at each wavelength. Scans were performed at a rate of 1 nm/s 123 Figure 4.4 Real-time LRET-based signal from EKmax\u00E2\u0084\u00A2 mediated hydrolysis reaction. 32 u M of Tb 3 + -CBM9-PID 4 K-GFP fusion protein was incubated at 30\u00C2\u00B0C with 9.3 nM EKmax\u00E2\u0084\u00A2 (solid squares). The control reaction has buffer in place of enzyme (open squares). Sample was excited at 235 nm and emission was measured at 520 nm following a 200 us delay. Each measurement is an average of 10 excitations 124 Figure 4.5 Baseline corrected logarithmic decay in ln($) due to EKmax\u00E2\u0084\u00A2 mediated hydrolysis. Reaction conditions are the same as in Figure 4.4. The solid line represents the linear least-squares fit to the initial data 125 Figure 4.6 Michaelis-Menten analysis. Initial rates of EKmax\u00E2\u0084\u00A2-catalyzed hydrolysis measured as a function of initial substrate concentration. Tb 3 + -CBM9-PID 4 K-GFP was incubated with EKmax\u00E2\u0084\u00A2 (9.3 nM) at 30\u00C2\u00B0C.... 126 Figure 4.7 Lineweaver-Burk representation of initial rates of EKmax\u00E2\u0084\u00A2-catalyzed hydrolysis of Tb 3 + -CBM9-PID 4 K-GFP measured as a function of initial substrate concentration. Reaction conditions are the same as reported in Figure 4.4 127 Figure 4.8 SDS-PAGE documentation of EKmax\u00E2\u0084\u00A2-catalyzed hydrolysis of Tb 3 + -CBM9 -PID4K -GFP. Reaction conditions are the same as reported in Figure 4.4 128 Figure 5.1 Scanning electron micrographs of Perloza\u00E2\u0084\u00A2 MT100 beaded media. Figures A , B and C show an MT100 particle at 1.0k, 13.0k and 80.0k magnification, respectively. The pore structure shown at the center of Figure B is further magnified and shown in Figure C 157 xin Figure 5.2 Equilibrium adsorption isotherm for batch binding of CBM9-GFP to Perloza\u00E2\u0084\u00A2 MT100/G15 composite media at 4\u00C2\u00B0C. The solvent consisted of 50 m M potassium phosphate, 150 m M NaCl, pH 7. The solid curve represents the best fit of the experimental data to the Langmuir isotherm equation 158 Figure 5.3 Chromatogram for CBM9-GFP purification on Perloza\u00E2\u0084\u00A2 MT100/G15 composite media. Clarified E. coli cell lysate containing CBM9-GFP was loaded at a superficial velocity of 4.25 x 10\"3 cm s\"1 onto a 10 mL column. Fractions were collected and analyzed both by absorbance at 280 nm (line) and by fluorescence intensity (solid diamond) as shown 159 Figure 5.4 SDS-PAGE documentation of CBM9-GFP purification on a 10 mL Perloza\u00E2\u0084\u00A2 MT100/G15 column. A l l samples were dissolved in sample buffer containing 10% SDS. Lane M : molecular mass markers; Lane 1: clarified cell lysate prior to column loading; Lane 2: column flow through; Lane 3: high salt wash; Lane 4: low salt wash; Lane 5: CBM9-GFP eluted in low salt buffer containing 1 M glucose 160 Figure 5.5 First moment (pi) analysis for a 10 mL column packed with Perloza\u00E2\u0084\u00A2 MT100/G15. Integrated pi values are reported for pulse injections of a 50 pL solution of blue dextran (MW 2000 kDa) over the interstitial velocity range 4.25 x 10\"3 cm s\"1 to 2.02 x 10\"2 cm s\"1. The mobile phase consisted of 50 m M phosphate buffer, 150 m M NaCl (pH 7, 4\u00C2\u00B0C) 161 Figure 5.6 Measured average porosity ep of composite media as a function of molecular weight of standard proteins. Pulse injections of standard molecular weight protein markers were used over the interstitial velocity range 2.1 x 10\"3 cm s\"1 to 8.5 x 10\"3 cm s\"1. The porosity was determined from equation 5.15 using the measured void fraction of 0.425 162 Figure 5.7 Determination of axial dispersion coefficient Di and overall solute mass-transfer coefficient K m for injection of CBM-GFP under nonbinding conditions onto a 10 mL column packed with Perloza\u00E2\u0084\u00A2 MT100/G15. Pulse injections of a 50 pL solution of CBM9-GFP was used over the interstitial velocity range 4.25 x 10\"3 cm s\"1 to 1.38 x 10\"2 cm s\"1. The mobile phase consisted of 50 m M potassium phosphate, 150 m M NaCl (pH 7, 4\u00C2\u00B0C) with 2 M glucose added to achieve nonbinding conditions. 163 Figure 5.8 Time course fluorescent intensity profile of CBM9-GFP uptake into Perloza\u00E2\u0084\u00A2 MT100 particle. Protein uptake monitored by an inverted Zeiss L S M 510 confocal laser scanning microscope with the center of the particle used as the focal plane. Excitation and emission wavelengths were 488 nm and 505 nm, respectively. The initial CBM9-GFP o concentration C\ outside the particle was 5.4 p M 164 xiv Figure 5.9 Time dependent radial profiles of CBM9-GFP uptake into a Perloza\u00E2\u0084\u00A2 MT100 particle for a feed concentration of 5.4 uM. Predicted uptake rates using (Figure 9A) pore-diffusion model and (Figure 9B) two-zone model are compared with experiment. The rapid drop in fluorescence intensity at radial positions above ca. 39 urn indicates the position of the outer radius of the bead 165 Figure 5.10 Comparison of T Z M model predictions of rc(t) with values computed from C L S M data. C L S M determined core radius reported as the radius at which the measured fluorescence intensity falls below 3X the standard deviation of the background fluorescence. Load conditions same as stated in Figure 5.9 166 Figure 5.11 Time dependent radial profiles of CBM9-GFP uptake into a Perloza\u00E2\u0084\u00A2 MT100 particle for a feed concentration of 49 uM. Predicted uptake rates using (Figure 11 A) pore-diffusion model and (Figure 11B) two-zone model are compared with experiment 167 Figure 5.12 Comparison of T Z M (solid curve) and P D M (dashed curve) predictions with experimental (points) breakthrough curves. Pure CBM9-GFP loaded at a superficial velocity of 8.5 x 10\"3 cm s\"1 onto a Perloza\u00E2\u0084\u00A2 MT100/G15 composite media column: (A) frontal load of 3.4 x 10\"3 mol m\"3 CBM9-GFP, (B) frontal load of 1.85 x 10\"2 mol m\"3 CBM9-GFP 168 Figure 5.13 T Z M predicted (line) and experimental (points) breakthrough curves for pure CBM9-GFP loaded onto a Perloza\u00E2\u0084\u00A2 MT100/G15 composite media column at three different feed concentrations: c ; = 4.77 x 10\" mol m\" (triangles), 2.2 x 10\"2 mol m' 3 (squares), and 8.0 x 10\"3 mol m\"3 (circles). Mobile phase loaded at a flow rate of 0.4 mL min\"1 169 Figure 5.14 T Z M predicted (line) and experimental (points) breakthrough curves as a function of interstitial velocity. Pure CBM9-GFP loaded onto a Perloza\u00E2\u0084\u00A2 MT100/G15 composite media column: u = 1.7 x 10' cm s\" (triangles), 8.5 x 10\"3 cm s\"1 (circles), and 4.2 x 10\"3 cm s\"1 (squares) 170 Figure 5.15 Predicted (line) and experimental (points) breakthrough curves for frontal loading of a clarified cell extract onto a Perloza\u00E2\u0084\u00A2 MT100/G15 composite media column. The clarified cell extract contained 25 u.M CBM9-GFP 3 1 and was loaded at a superficial velocity of 8.5 x 10\" cm s\" . Eluent absorbance data at 280 nm (open squares) are also shown to indicate total protein as a function of time 171 Figure 6.1 Picture of an MT100 column pre- and post-compression. Picture shows bed compression at a superficial velocity of 1.06 x 10\"4 m/s 197 xv Figure 6.2 Schematic representation of the proposed epoxide-based reaction for cross-linking cellulose with epichlorohydrin 198 Figure 6.3 Imageplot (A) and wireframe surface (B) for a locally weighted polynomial regression fit to the measured w c ri t (in m s\"1 x 104) data set reported in Table 1. Created with the loess, levelplot, and wireframe functions in R (Young et al. 1980) and the add-on package Lattice (Lattice Graphics R package, version 0.14-16), with a span of 0.65 and a degree of 2 (i.e. local fits are quadratic). The location of the maximum is indicated by (\u00C2\u00AE) 199 Figure 6.4 Wireframe surface of the results obtained from the standard quadratic response surface model (6.2) when applied to the measured \u00C2\u00AB c rj t (in m s\"1 x 10*) data set reported in Table 1 200 Figure 6.5 Pressure curve for solvent flow through an MT100 (open circles) or CRL100-7 (filled circles) column monitored over a range of superficial velocities. The pressure drop curve for C R L 100-7 was compiled from two independent experiments 201 Figure 6.6 Equilibrium adsorption isotherms for binding of CBM9-GFP to MT100 (filled circles) and C R L 100-7 (open squares) at 21\u00C2\u00B0C. Control experiment shows ovalbumin binding to CRL 100-7 (filled squares) at 21\u00C2\u00B0C. The mobile phase buffer consisted of 50 m M phosphate buffer, 100 m M NaCl, pH 7. The curve represents the best fit of the experimental data to the Langmuir adsorption isotherm equation where q\ is the protein concentration bound to the media surface and c\ is the equilibrium concentration of protein free in solution 202 Figure 6.7 Measured first central moment ( p i ) for blue dextran (filled squares) and CBM9-GFP (filled circles) on a column (I.D. 1.0 cm) packed with CRL100-7 203 Figure 6.8 Second moment analysis under non-binding conditions for a column (I.D. 1.0 cm) packed with CRL100-7 media. Pulse injections of CBM9-GFP were used over the superficial velocity range 4.25 x 10\"5 m/sto 4.25 x 10\"4 m/s. The mobile phase consisted of 2 M glucose in 50 m M potassium phosphate, 100 m M NaCl (pH 7, 21\u00C2\u00B0C) 204 Figure 6.9 Van Deemter plot describing CBM9-GFP transport within the CRL100-7 column (I.D. 1.0 cm X 11.4 cm); the reduced plate height h a 2 L^y / ^p^d p j is plotted as a function of the reduced linear velocity, given by (UCIV)IDM) 205 xvi Figure 6.10 Chromatogram for purification of CBM9-GFP on a 60-mL CRL100-7 column at 21\u00C2\u00B0C. Clarified cell extract from E. coli BL21 were loaded onto a CRL100-7 column (2.6 cm I.D. X 11.8 cm) at a superficial velocity of 1.5 x 10\"2 m/s, washed with ca 3 column volumes (CV) of high salt buffer, ca 2.5 C V of low salt buffer and bound CBM9-GFP was eluted with I M glucose in low salt buffer. 10 ml fractions were collected and analyzed by both absorbance at 280 nm (solid line) and fluorescence emission at 509 nm (filled circles) 206 Figure 6.11 12% SDS-PAGE documentation for preparative-scale affinity purification of CBM9-GFP on a CRL100-7 column (I.D. 2.6 cm). A l l samples were dissolved in sample buffer containing 10%) SDS. Lane M : molecular weight markers in kg/mole; Lane 1: clarified cell extract prior to column loading; Lane 2: column flow through; Lane 3: high salt wash; Lane 4: low salt wash; Lane 5: pure CBM9-GFP eluted in low salt buffer containing 1 M glucose 207 xvii Nomenclature As peak asymmetry factor ct total concentration of free solute / in the pore liquid ( mol m\" ) Cj * equilibrium concentration of free solute / in the pore liquid (mol m\" ) Cj total concentration of solute i in the interstitial mobile phase liquid ( mol m\" ) C\u00C2\u00B0 total concentration of solute i in the interstitial liquid at time t=0 ( mol m\"3 ) Cfeed total concentration of solute / in the injected sample (mol m\"3) Cp. heat capacity (kcal\u00C2\u00B0C mol\"1) dc column diameter ( cm) dp particle diameter ( p m ) DL axial dispersion coefficient (m s\" ) 2 1 DM molecular coefficient in bulk liquid ( m s\" ) 2 1 DP intraparticle diffusivity ( m s\" ) 2 1 D s surface diffusivity (m s\" ) E efficiency of energy transfer ET total enzyme concentration ( n M ) fo corrected fluorescence intensity of donor (arbitrary units) HETP height equivalent to a theoretical plate ( m ) AHcai calorimetric enthalpy of denaturation ( J mol\"1) I fluorescence emission intensity at specific wavelength (arbitrary units) xviii J spectral overlap between donor fluorescence and acceptor absorbance (nm 4 M \" 1 cm\"1) k aspect factor kads sorption rate constant ( m 3 mol\"1 s\"1) kai Langmuir binding constant (M\"1) kf fluid film mass transfer coefficient - ( m s\"1 ) kcat catalytic constant (s\" 1) k0bs observed rate constant (s\"1 ) KM Michaelis Menten kinetic constant ( p M ) KM overall solute mass transfer coefficient ( s\"1) L column length ( m ) M protein molecular weight ( g mol' 1) n refractive index of medium NTU number of theoretical units P pressure qt* equilibrium concentration of bound solute / <7, concentration of bound solute i in pore liquid ql\"ax maximum capacity of sorbent to bind solute i q?at sorbent saturation capacity (pqjmax) QD quantum yield of donor chromophore rc core radius ( m ) (Pa) (umol g\"1 resin) (mol m\"3) (umol g\"1 resin) (mol m\"3) xix rp particle radius ( m ) R average Forster distance (A) R0 Forster distance at E = 50% (A) RM overall resistance to solute mass transfer ( s ) 5 , average solute concentration within sorbent particle (mol m\"3) St solute concentration within sorbent particle (mol m\"3) [s] substrate concentration ( p M ) [s]0 initial substrate concentration ( p M ) t time ( s ) ti/2 half-life of substrate ( h ) T temperature ( \u00C2\u00B0 C ) Tm melting temperature ( K ) u superficial fluid velocity ( m s\"1) u0 interstitial fluid velocity (m s\"1) Ucrit critical superficial velocity for column compression (m s\"1) v rate of enzyme cleavage ( y M h\"1) v0 initial reaction rate ( p M h\"1) vmax maximum reaction rate ( p M h\"1) z axial coordinate ( m ) xx Qimensionless numbers Bi Biot Number (kjdp/6Dp) Da Damkohler Number (C\u00C2\u00B0kadsdp/4DP) h reduced plate height (cr2L)/(juf dp) Re Reynolds number (dpup/ju) Sc Schmidt number (JU/PDM) Sh Sherwood number (kjd/DM) Pe Peclet Number (udp/Dp) ured Reduced Linear Velocity (udp/DM) Greek letters Po regression coefficient for total error in fit of equation (m s\"1) 2 1 pi linear regression coefficient for epichlorohydrin concentration (gm\" s\" ) 2 1 p 2 linear regression coefficient for DMSO concentration ( g m\" s\" ) 2 5 1 Pn quadratic regression coefficient for epichlorohydrin concentration ( g m\" s\" ) 2 5 1 P12 regression coefficient for linear interaction between and ^2 (g rn\" s\" ) a, y, 5 fitted parameters (Equation 5.17) input variable for epichlorohydrin concentration ( ml g\"1) ^2 input variable for DMSO concentration ( ml g\"1) *F random variable accounting for total error in measurement ( m s\"1) s interstitial void fraction of column s a extinction coefficient (M\" 1 cm\"1) s p average stationary phase porosity K2 orientation factor n fluid viscosity (g.m\"1 s\"1) fraction of substrate remaining p fluid density (g m\"3) 7 2 a second central moment of an eluent peak ( s ) Ida lifetime of donor signal in presence of acceptor chromophore (ms) t d lifetime of donor signal in absence of acceptor chromophore ( ms ) xxii xjb lifetime of free terbium (III) signal (ms) p fluid viscosity (g m\"1 s\"1) p i first central moment of an eluent peak ( s ) xxiii Acknowledgements First and foremost, I would like to thank my research supervisor, Dr. Charles Haynes for his invaluable guidance, for allowing me the opportunity to gain experience in a, variety of research areas and for teaching me to see the forest beyond the trees. I would also like to thank Drs. Doug Kilburn and Tony Warren for giving me the opportunity to become a member of their group. Their kindness and guidance will always be remembered and appreciated. I'm grateful to Drs. Steve Withers and Chris Overall for allowing me use of their resources. Without their generosity, certain aspects of my research would have been harder to fulfill. A special thanks goes to Dr. Richard Fellmen for his advice on enzyme kinetics. I would like to thank all the members of the Michael Smith Laboratories and especially my colleagues in both the Haynes' Lab and the Cellulase Group for providing a fun and friendly environment to work in. To Dr. Louise Creagh, thank you for all your advice. Last but not least, I would like to thank my parents for their love and support. This thesis is dedicated to them. xxiv To my parents, Houshang and Parizad Kavoosi xxv Co-Authorship Statement This work was done under the guidance of Dr. Charles A . Haynes and Dr. Douglas G. Kilburn. Dr. Louise A . Creagh helped with the analysis of the differential scanning calorimetry results reported in Chapter 3 and performed the isothermal titration calorimetry experiments reported in Chapter 4. Dr. Jenny Bryan assisted with the statistical analysis reported in Chapter 6. Nooshafarin Sanaie wrote the computer programs used to predict the breakthrough curves reported in Chapter 5. Florian Dismer from Dr. Jiirgen Hubbuch's group did the confocal laser scanning microscopy experiments reported in Chapter 5. Emily Kwan provided the immobilized CBM2a-Factor X a fusion protein used for experiments reported in Chapter 2. Julia Meijer and Dexter Lam were undergraduate students under my supervision. I performed all other experiments and data analysis reported in this thesis. I am also the principle author on all the publications reported in this thesis and had primary responsibility for both their conceptual and practical aspects. xxvi 1 Introduction, Background and Thesis Objectives 1.1 Overview In 2004, the global biotechnology industry reported annual revenues of $54.6 billion and accounted for over 20% of all venture capital investments. This included 365 products in the pipeline from Phase I to Phase III clinical trials and 55 new drug applications submitted for review by the Food and Drug Administration (FDA) (Source: Ernst and Young). On average, the cost of product development is now estimated at $800 million per new drug. This rapidly rising cost and the increasing number of new protein therapeutics reaching the market are placing significant pressure on a global healthcare industry already challenged, particularly in North America, by an increasingly aging population. Sustainable growth of the biotechnology industry is therefore dependent on discovery of new ways to increase efficiency and productivity in order to keep both development and manufacturing costs in check. The process of manufacturing a biologic involves a number of unit operations, including (i) cell-line development and clone selection, (ii) fermentation/cell culture, (iii) cell harvest/removal, (iv) cell disruption (if needed), (v) product concentration, (vi) product purification, and (vii) final product formulation. Each of the final six processing steps generally involves one or more unit operations. Due to technological advancements over the past few decades, the cost of producing a biologic has decreased considerably. Refinements in expression system design and in fed-batch and perfusion-culture protocols now enable the production of the target biomolecule at g/L quantities at relatively little cost ($100-$500/g total cost of goods). As a result, the overall economics of therapeutic protein production now tends to be dominated by the cost of product purification (Lowe et al. 2001). New strategies to purify recombinant proteins from complex high-density cultures must therefore be developed with the aim of maximizing yield and purity while minimizing costs. A key objective is to streamline downstream operations into a smaller 1 number of unit operations, with each step employing a separation strategy highly selective for the target protein. The bioselective adsorption process known as affinity chromatography is one such unit operation. Affinity chromatography takes advantage of highly selective interactions found in nature, such as the tight arid specific binding (K a ~ 106 to 109 M\" 1) between an antibody and its antigen. For example immobilized IgG can be used to purify protein A from Staphylococcus aureus (Moks et al. 1987; Yarnall and Boyle 1986). Conversely, protein A covalently coupled to activated-Sepharose media is now widely used to remove and isolate immunoglobulins from serum of different species (Chen et al. 2006; Jungbauer and Hahn 2004) and in the purification of monoclonal antibodies (mAbs), the fastest growing class of recombinant protein therapeutics (Follman and Fahrner 2004; Hahn et al. 2006). Advances in recombinant D N A technology have allowed the power of high-affinity interactions to be generically applied to the concentration and purification of recombinant proteins. Recombinant hybrids containing an N - or C-terminally fused affinity polypeptide tag that selectively binds to a complementary ligand immobilized onto a suitable chromatographic matrix are now widely used to facilitate the purification of target proteins or peptides. Numerous affinity tags have been developed over the years (Hearn and Acosta 2001) with many of them successfully commercialized for laboratory scale applications (Nilsson et al. 1997; Terpe 2003). However, their application at the preparative scale is sparse, due in large part to the high cost of the associated affinity media (Ladisch 2001). This high cost can arise from the complex chemical modifications needed to immobilize the ligand to the resin surface. In addition, immobilized ligands can experience fouling and structural damage, particularly at the ligand-resin junction, resulting in a reduced column capacity and a short column lifetime. This latter sensitivity often leads to a low tolerance of many affinity media to repeated processing and sanitization cycles, which limits their use in production-scale purifications. Thus, although affinity chromatography has the potential to simplify downstream processing by achieving higher yields and purities than conventional methods, it is not without problems. This doctoral thesis aims to alleviate ligand stability and cost limitations associated with conventional affinity chromatography by introducing an 2 affinity tag that binds to a very inexpensive and stable cellulose-based matrix. The research focuses on the development of a generic and inexpensive affinity purification technology based on high-level recombinant expression in E. coli of fusion proteins containing the family 9 carbohydrate-binding module (CBM9) attached to either the N -or C-terminus of the target protein or polypeptide. A linker sequence containing a specific processing site is designed to facilitate rapid and quantitative removal of the tag, and to thereby recover the desired target sequence following affinity purification. The development, modeling and commercialization of this technology are the central objectives of this thesis. 1.2 Literature Review 1.2.1 Downstream Processing in Biotechnology Recombinant D N A technology has provided effective protein therapeutics for many of our deadliest and most debilitating diseases, such as various forms of cancer, asthma, arthritis, heart disease, diabetes and AIDS, to name just a few. Yet despite the great benefit of recombinant proteins to human health, the $800 million dollar price tag to bring each drug to market (Source: Ernst and Young) is serving to limit the technology largely to the production of blockbuster therapeutics - i. e. protein therapeutics expected to achieve greater than $100 million in sales per annum. Many important but less profitable indications are therefore receiving little attention. In order to overcome this ethical problem and to meet the challenges created by rising healthcare costs, industry must focus on reducing development and manufacturing costs of novel and generic biopharmaceutical proteins. Lowering the number of processing steps and increasing the yield at each step would serve this objective (Ladisch 2001). Many manufacturers are therefore looking towards redesigning and improving the efficiency and yields of existing production processes to lower costs (Narayanan 1994; Thomson 1996). Manufacturing of biopharmaceuticals must comply with quality assurance and cGMP (current good manufacturing practice). The FDA and other regulatory agencies such as the E M E A (European Medicines Agency) require that all finished biologies be manufactured under full cGMP compliance. This means that the final therapeutic 3 product must be a well-characterized biologic with defined purity, efficacy, potency, stability, pharmacokinetics, pharmacodynamics, toxicity and immunogenicity. The product must also be exhaustively tested for contaminants such as nucleic acids, viruses, pyrogens, residual host cell proteins, cell culture media, and leachates from the separation media. In addition, many therapeutic proteins are naturally and recombinantly produced in different isoforms originating from post-translational modifications such as glycosylation, sulphation, oxidation, end termination modifications, misfolding, aggregation, misaligned disulphide bridges and nicked or truncated variants. The process by which a therapeutic protein is produced is known to influence all of these product quality characteristics. Thus, the F D A views the process as the product, and great care must be taken in the selection and operation of the purification process. 1.2.2 Affinity Chromatography in Preparative-Scale Protein Purification Regulatory compliance, process economics and strict quality assurance have instigated a rethinking of the design and operation of purification processes, with effectiveness, robustness and economics largely determining the choice of the purification strategy. Traditional, multi-step purification protocols based on non-specific physio-chemical properties such as pH, temperature, and solubility, are being replaced with highly selective and more sophisticated strategies, including those based on affinity chromatography (Bonnerjea et al. 1986). Affinity chromatography exploits natural biological binding processes for the selective separation of the target protein. Affinity separation techniques have the power to reduce the total number of processing steps, thereby increasing yields and downsizing capital equipment leading to improved process economics. However, these techniques do suffer from problems, particularly with respect to regulatory and other issues discussed below (Stevenson 1997). Most affinity ligands used in research and industry originate from natural sources. The library of affinity ligands commonly used today includes immobilized sugars and polysaccharides, nucleic acids, dyes, chelating agents, proteins, protein fragments, monoclonal antibodies, peptides and amino acids (Hearn and Acosta 2001; Hopp 1988). While these ligands often show good selectivity, their use in affinity chromatography can 4 be challenged by a number of factors. For example, they must be produced and purified, may be contaminated with host D N A and viruses, tend to be fragile, show lot-to-lot variation and are often costly to produce. In addition, production of therapeutics requires stringent column sterilization typically involving clean-in-place protocols that utilize strong caustic solutions which can degrade the immobilized ligand, shortening column life and contaminating the end product with potentially toxic or immunogenic leachates. Biological ligands can also suffer from low binding capacities, limited life cycles and low scale-up potential. 1.2.3 \"Natural\" Affinity Chromatography In natural affinity chromatography the binding motif is endogeneous to the target protein, allowing for direct and specific purification through selective binding to its cognate ligand immobilized onto an appropriate chromatography media. Examples of natural affinity chromatography are abundant in the literature (Jones 1990). Many glycoproteins are currently purified using immobilized lectins (Cummings 1997; Yang and Hancock 2004), while conversely many lectins have been efficiently isolated by specific adsorption on immobilized carbohydrates (Chen and Billingsley 1999; Lee et al. 1995). Other examples include the industrial purification of human plasminogen from blood plasma using immobilized lysine (Deutsch and Mertz 1970), the purification of ATP-dependent kinases and NAD+-dependent dehydrogenases using immobilized 5'-A M P (Mulcahy et al. 2002), and the use of the sulfated polysaccharide heparin to affinity purify coagulation factors (Levine 1976). Aromatic compounds, particularly natural and synthetic textile dyes, have also been used as ligands in natural affinity chromatography. Cibacron blue is the most popular of these dyes and has been widely used in the research setting to purify albumin and other blood proteins, including oxido-reductases, carboxylases, glycolytic enzymes, nucleases, hydrolases, lyases, synthetases and transferases (Clonis et al. 2000; Denizli and Piskin 2001). The application of synthetic aromatics and dyes to the purification of therapeutic proteins has however been limited by concerns over the purity, selectivity, leakage and toxicity of these capture agents. Finally, for immunogenic proteins that do 5 not possess a known binding motif, protein-specific chromatography can be carried out using an immobilized monoclonal antibody (Ostermann 1990; Wells et al. 1993). However, the exceptionally high cost of monoclonal antibody production effectively limits the large-scale application of the technology to purification of rare and very high value proteins. 1.2.4 Affinity Fusion-Tag Technology Affinity fusion-tag technology takes advantage of an existing and well-characterized interaction between a unique and infrequent binding motif and its corresponding ligand to purify a target protein. Recombinant D N A technology is used to genetically fuse the binding motif (henceforth called the affinity tag) to either the N - or C-terminus of a target protein. The recombinant fusion protein can then be expressed and purified using the selective interaction between the tag and the complementary ligand immobilized onto a suitable chromatographic support. Affinity tag chromatography therefore allows for good selectivity and greater predictability in the separation process, and the general ability to be applied to a range of target proteins without the need for developing a new protocol for each new protein of interest (Uhlen and Moks 1990). Affinity tags range in complexity and molecular size from short, unstructured peptides to large binding domains. In addition to their application in the detection and purification of the protein of interest, affinity tags can offer a number of advantages. In cases where the target protein is conformationally unstable or requires complex disulfide bond formation, some affinity tags can stabilize the target protein, allowing it to fold more efficiently into the correct \"near native\" conformation in vivo by functioning as pseudo-chaperones or secondary structure nucleation centers (Bardwell et al. 1991; Hammarberg et al. 1989; Murby et al. 1991; Nygren et al. 1994; Tucker and Grisshammer 1996). Others can act as reporters, allowing for easy detection and quantification of the fusion protein using high-sensitivity detection assays such as an enzyme-linked immunosorbent assay (ELISA). Positioning the tag at the C-terminus ensures the production and recovery of a fully intact fusion protein. Truncated variants brought about by errors in translation will not have the affinity tag at the C-terminus and 6 therefore will not bind to the column containing the corresponding immobilized ligand (Jones et al. 1995). Some fusion tags have also been shown to extend the half-life and increase the pharmacokinetic activity of therapeutic proteins. For example, CD4 fused to the Fc region of IgG has an activity 200X greater than native CD4 (Capon et al. 1989). In addition, fusion tags, due to their highly selective interaction, can be used to purify low abundance proteins present in a complex feedstock (Brewer et al. 1991). Finally, certain affinity tags are an attractive tool for structure-function studies, functioning as specific reporters for use in systems such as in vitro two-hybrid protein-protein binding systems (Hoffmann and Roeder 1991; Jonasson et al. 1996; Lu et al. 1993; Nieba et al. 1997; Nilsson et al. 1997). Therefore, additional benefits may be realized through selection of an appropriate affinity tag based on the characteristics and intended application of the target protein. 1.2.5 Factors to Consider in Designing the Affinity Tag System 1.2.5.1 The Choice of Peptide Versus Protein Tags Affinity tags can be grouped into either peptide tags or protein tags. Tags less than 25 amino acids in size are typically classified as peptide tags whereas small globular modules with unique binding properties are considered protein tags. Each class of affinity tag offers unique properties that must be considered during the selection process. For example, peptide affinity tags, due to their small size, are less likely to interfere with the structure or alter the catalytic activity of the target protein, and hence, their removal may not be necessary after purification (Tucker and Grisshammer 1996). For cases where tag removal is required, their simple structure generally permits enhanced accessibility of the processing enzyme to any introduced cleavage site (Jonasson et al. 1996). On the other hand, protein affinity tags have been known to help enhance structural stability and solubility by promoting the correct three-dimensional folding of the target protein. For example, both insulin-like growth factor II (IGF-II) (Hammarberg et al. 1989) and T-cell receptor (Murby et al. 1991) have been shown to be more stable when expressed as a fusion protein. Protein tags such as protein A (Abrahmsen et al. 7 1986) and maltose-binding protein (MBP) (di Guan et al. 1988) not only stabilize the protein, but are generally more compatible with leader peptide insertion to direct the fusion protein to the periplasm where the oxidizing environment can promote disulfide bond formation. In addition, protein tags can facilitate protein folding by acting as covalently linked pseudochaperones (LaVallie 1995) or \"protein enhancers\" (Nygren et al. 1994; Stahl et al. 1997a; Stahl et al. 1997b). For example, the ZZ domain protein derived from Staphylococcal protein A has been shown to dramatically improve the correct folding of recombinant IGF-I (Samuelsson et al. 1996; Samuelsson et al. 1994). 1.2.5.2 Position of the Affinity Tag The positioning of an affinity tag can influence the expression of the resulting fusion protein. For example, a hexahistidine tag is seldom placed at the N-terminus of proteins destined for cellular export since the multiple positively charged imidazoyl side chain groups can interfere with signal sequence processing, trafficking and export biorecognition, resulting in arrest of protein secretion or incorrect cleavage of the signal peptide (Seidler 1994). In general, N-terminal tags offer the advantage of yielding a target protein with its authentic N-terminal sequence, provided an appropriate specific protease recognition site is inserted between the tag and the target protein (Jones et al. 1995). Disadvantages of N-terminal tagging include the potential to copurify truncated products with the full-length protein, as well as the aforementioned potential for the tag to interfere with translation initiation. A C-terminal tag, on the other hand, does select only full-length protein, but generally adds unwanted amino acids to the end of the target protein upon protease cleavage (Jones et al. 1995). Finally, in certain circumstances, regardless of their position, fusion tags have been found to be buried within the internal structure of the folded protein (Kronina et al. 1999; LaVallie 1995; Lindner et al. 1992; Sharrocks 1994; Uhlen and Moks 1990) or to partially cover a binding/catalytic site on the target protein (Casey et al. 1995). For example, a hexahistidine tag placed at the C-terminus of a tumor-associated single chain Fv sequence was found to partially cover the antigen binding site of the protein (Goel et al. 2000). Therefore, care must be taken in tag selection and positioning to ensure that it 8 functions independently and does not affect the activity or cause any perturbation in the conformation of the target protein (Hearn and Anspach 1990; Hearn and Gomme 2000). 1.2.5.3 Expression of Fusion Proteins in Host Cells A large number of organisms have now been specifically engineered to express recombinant proteins (Makrides 1996). They include E. coli and other prokaryotic cells (Ford et al. 1991; Nilsson et al. 1997), Saccharomyces cervesiaie (Hopp 1988; Shimabuku et al. 1991), Pichia pastoris (Andrade et al. 2000; Higgins and Cregg 1998; Rajamohan et al. 2000) and other yeast strains, mammalian cells (Chubet and Brizzard 1996; Lo et al. 1998; Witzgall et al. 1994), insect cells based on the baculovirus expression system (Allet et al. 1997; Kuusinen et al. 1995; Yet et al. 1995), egg sack yolk (Fassina et al. 1998; Hansen et al. 1998), and even transgenic plants and animals (Hennig and Schafer 1998; Hermanson and Turchi 2000; Kleiner et al. 1999). Higher eukaryotic organisms possess the machinery to perform post-translational modifications and as such, are often used to express proteins that require complex modifications including O- or N -linked glycosylations, phosphorylation/methylation and complex disulfide bond formation (Marino 1989). These systems can however, suffer from low expression of recombinant proteins (Ford et al. 1991), hyperglycosylation or alternative sites of glycosylation (Higgins and Cregg 1998). Bacterial expression systems can provide higher expression levels but often at the expense of increased protein loss during the subsequent refolding (if necessary) and purification steps (Brewer et al. 1991). New to the field of recombinant protein expression, the egg yolks and the milk of lactating transgenic animals are gaining much interest for offering a fast, simple and inexpensive method for manufacturing biologies. . Recombinant proteins can sometimes be expressed in the form of inclusion bodies (Kane and Hartley 1988; Krueger et al. 1989). Within these inclusion bodies, proteins exist in a biologically inactive, partially unfolded and reduced state. The recovery protocol needed to refold the protein back to its native state is unique for each individual protein and must be determined experimentally. This adds to the cost and complexity of the bioprocess since each protocol must be optimized with respect to temperature, pH and 9 buffer composition (Flaschel and Friehs 1993; Rudolph 1996). However, with a robust renaturation protocol predetermined for a given protein, incorporating inclusion body formation as part of a bioprocess offers the advantage of ease of recovery from the cell homogenates (Thatcher 1990) and protection from cellular proteolysis (Sassenfeld 1990). For problematic cases, inclusion body formation can be curbed by directing the fusion protein into an oxidative environment such as the periplasm of E. coli or through the rough endoplasmic reticulum and golgi apparatus and into the culture medium in eukaryotic cells (Abrahmsen et al. 1986; Carter et al. 1992). Trafficking the fusion protein to a specific compartment of the cell or secretion into the culture supernatant offers a number of benefits, including (i) limited exposure to cytoplasmic proteases, which could degrade the target protein (Murby et al. 1991), and (ii) a simplified purification process by secreting the fusion protein into an environment with less contamination from other cellular components (Hansson et al. 1995; Hockney 1994). 1.2.5.4 Tag Removal Post-Purification Recombinant proteins destined for pharmaceutical/therapeutic application typically must have their tag removed to minimize adverse side effects, including immunogenic responses, and to meet regulatory standards (Pedersen et al. 1999). This is achieved by introducing, at the cloning stage, a site-specific cleavage site between the tag and the target protein. Once the fusion protein is purified, cleavage at this specific processing site can be accomplished by either enzymatic or chemical means (Brewer et al. 1991; Casey et al. 1995). Chemical cleavage is most often achieved with cyanogen bromide (Itakura et al. 1977) which cleaves only after methionine residues. Although inexpensive, it is toxic and the harsh reaction conditions can cause denaturation and degradation of the target protein. Cleavage using site specific endopeptidases such as enterokinase (Hosfield and Lu 1999; Prickett et al. 1989), thrombin (Hakes and Dixon 1992), Factor X a (Mcpherson et al. 1992), and the two relatively new processing enzymes, tobacco etch virus (TEV) (Shih et al. 2005) and human rhinovirus 3C protease (HRV 3C) (Libby et al. 1988), offers a gentler alternative. Although enzymatic methods are safer and more specific, they are generally less efficient since their rate of cleavage is 10 affected by the accessibility of the cleavage site and the chemistry of the adjacent amino acids, particularly those lying in the linker sequence immediately upstream of the recognition site. Some endopeptidases such as thrombin, trypsin and H R V 3C also leave behind extra amino acids on the termini of the target protein upon cleavage (Wyborski et al. 1999). Self-cleavable systems have attracted much interest in recent years. Intein-based systems are based on natural protein splicing sequences fused to affinity tags such as the chitin-binding domain (Derbyshire et al. 1997; Wood et al. 1999). Cleavage of the intein can be achieved by addition of thiols, shifting pH, or temperature. Although replacing specific processing recognition sites with inteins would simplify the cost and complexity of a bioprocess, inteins do have some significant limitations: (i) the large size of the intein and its associated catalytic machinery places a greater metabolic burden on the cells; (ii) the processing efficiency is dependent on the amino acid sequences at the junctions; (iii) inteins generally have a very slow autoprocessing rate; and (iv) inteins do not generally enhance the solubility or facilitate the purification of the fusion partner. 1.2.5.5 Affinity Tag-Ligand Interaction To provide for a robust and selective separation process, the interaction between the fusion tag and the ligand should be a reversible through a mild change in solution conditions. That is, the association should be sufficiently strong and specific, yet under the appropriate condition, allow for an efficient, yet gentle, dissociation. Ideally, the association constant (K a ) should be in the range of 104 to 108 M \" ' (Chao et al. 1998). With a K a less than 104 M \" 1 , the interaction will be too weak, resulting in product loss, while with a K a greater than 108 M \" 1 the complex is sufficiently strong to make elution difficult (Lindner et al. 1992). While identifying the ideal adsorption conditions is often relatively straightforward, determining the elution conditions that result in a high yield with a large purification factor can require extensive trials (Thiele and Fahrenholz 1994). The difficulty lies in identifying a condition that disrupts the tag-ligand interaction without causing degradation or conformational changes in the target protein. Low pH buffers 11 containing high concentrations of salts or chaotropic chemicals are often used to elute bound fusion proteins, but their application can be detrimental to the target protein (Hofmann et al. 1980; Schmidt and Skerra 1993; Yarmush et al. 1992). The use of excess ligand is a milder alternative, but in some cases, it is inefficient, expensive and provides low yields (Thiele and Fahrenholz 1994). Other protocols include elution based on irradiation with light (Olejnik et al. 1998; Thiele and Fahrenholz 1994), C a + 2 dependent binding (Hentz et al. 1996; Schauer-Vukasinovic and Daunert 1999) and even on-column use of proteases (Graslund et al. 1997; Nilsson et al. 1997). 1.2.6 Affinity Tag Sys tems The number of binding motifs designed for use as affinity tags has increased significantly over the past three decades. They include small antigentic epitope tags that bind to antibodies immobilized on columns, polyionic tags that adsorb onto ion exchange resins, tags that chelate with immobilized metal ions, biotinylated and non-biotinylated tags, tags derived from protein-protein or enzyme-substrate interactions, and finally, tags that bind to carbohydrate moieties (Table 1.1). A l l have been used at the laboratory scale, but only a few have succeeded beyond the laboratory that pioneered the given tag and become part of a commercial product (marked with asterisk in Table 1.1). The wide availability and relative low cost of polysaccharide matrixes might explain the commercially available affinity systems derived from protein tags that bind to carbohydrate based affinity resins. These tags vary in size from ~40 kDa for the maltose binding protein (MBP) (Blondel and Bedouelle 1990; di Guan et al. 1988; Riggs 2000) to about 119 amino acids for the starch binding domain (SBD) (Chen et al. 1991b). Currently, the M B P tag is the best studied and most popular of these carbohydrate-based affinity tags. Fusion proteins containing the MBP affinity tag can be purified on columns containing cross-linked amylose and the bound protein can be desorbed by competitive elution with 10 m M maltose in physiological buffer (Kellermann and Ferenci 1982). The MBP tag has been shown to act as a solubility enhancer (Kapust 1999) but only when fused to the N-terminus of the target protein (Sachdev and Chirgwin 1999). 12 A number of commercially available affinity systems also take advantage of peptide tags. These tags have flexible structures that minimize steric interference with the correct folding of the target protein and allow for enhanced accessibility of processing enzymes to any introduced cleavage sites. For example, the Flag tag (Hopp 1988) is a hydrophilic, 8 amino acid peptide that binds to immobilized monoclonal antibodies. Elution can be achieved either by the addition of chelating agents such as EDTA or by a decrease in the effective pH of the buffer. The Flag tag offers the unique advantage of encoding for a specific protease processing site at its C-terminus that can subsequently be used to remove the Flag peptide tag and obtain the purified target protein. Of the tag technologies listed, the glutathione S-transferase and polyhistidine tag technologies have found the greatest use in preparative-scale purifications. Both of these affinity tags were among the first to become commercially available; therefore, their timing, more than performance, might account for their more frequent use. Immobilized metal affinity chromatography (IMAC) is generally used to purify recombinant proteins containing a polyhistidine tag. First described in 1975 (Porath et al. 1975), IMAC is based on the interaction between an immobilized divalent transition-metal ion (Ni , Co + 2 , C u + 2 , Zn + 2 ) and solvent exposed aromatic residues on proteins or peptide tails. Histidine, or more specifically the electron-donating nitrogen of the acylimidazole ring of histidine, binds particularly tight through favorable coordination with the electron-accepting transition metal. A number of expression vectors are available that fuse a polyhistidine peptide, typically containing between 5 and 10 histidine residues, to the N -or C-terminus of any recombinant protein of interest (Hochuli et al. 1987). The resulting chimeric protein can then be adsorbed onto traditional I M A C columns containing an immobilized divalent metal ion, usually N i + 2 , and eluted either by adjusting the pH of the buffer or by adding free imidazole. The use of imidazole can be problematic, as it can influence N M R experiments, competition studies and crystallographic trials, and its presence can result in protein aggregation (Hefti et al. 2001). In addition, the polyhistidine tag is not recommended for purification of metal containing proteins as the bound metal ion can dissociate, resulting in elution of an apo-protein with potential for denaturation and/or protein aggregation. The I M A C columns designed to bind target proteins tagged with the polyhistidine tail are particularly susceptible to this problem 13 because the nickel cation that serves as the polyhistidine coordination site is physically adsorbed to the column through binding to immobilized ligands with high affinity for metal ions. Finally, the potential for dissociation and leakage of the divalent metal ion, even at ppm levels, typically prohibits its use in the purification of FDA approved proteins. The GST affinity tag system, developed by Smith and Johnson (Smith and Johnson 1988), takes advantage of the natural affinity between the enzyme glutathione S-transferase (GST) and its substrate, glutathione. Fusion proteins containing the GST affinity tag can therefore be purified on columns displaying immobilized glutathione. Being the first protein-based affinity tag system commercially available for protein purification, the GST tag technology has gained widespread use and an almost exclusive market share of large-scale purifications of therapeutic proteins that cannot be purified by conventional methods. The GST tag is, however, a homodimer (Kaplan et al. 1997), making it unsuitable for the purification of oligomeric proteins. A l l of the affinity tags listed in Table 1.1 exhibit one or more of a number of characteristics that have proven disadvantageous for production-scale purification of therapeutic proteins. Most require a ligand that is expensive to produce and/or to immobilize. For example, the coiled-coil affinity tag system requires synthesis and immobilization of a long peptide sequence (35 amino acids), composed of five consecutive repeats of a specific heptad sequence. Robust chemistries for coupling these ligands to the stationary phase are often not available, resulting in gradual leaching of the ligand into the mobile phase during column processing. The potential for dissociation and leakage of the ligand, even at ppm levels, typically prohibits their use in the purification of F D A approved proteins. Other, more tag specific problems have also been reported. For example, the GST tag has the propensity to produce fusion proteins in inclusion bodies (denatured protein aggregates), possibly due to the four solvent exposed cysteine residues present in each subunit (Kaplan et al. 1997). Thus, recovery of the purified protein in its native, fully active state requires a robust refolding strategy. Difficulties in eluting chitin-binding 14 protein tagged fusion proteins have been reported (Vaillancourt et al. 2000), and low binding efficiencies have been observed with M B P fusions (Pryor 1997). However, the dominant impediment to the use of these affinity tags in large-scale bioprocessing is cost. Table 1.2 lists the costs associated with using some of the common commercial affinity tags. Production-scale affinity chromatography typically requires column volumes greater that 10 L and up to 5000 L. For the systems reported in Table 1.2, this represents columns where the resin cost alone is in excess of $30,000 and can be as high as $11,000,000. A new generic affinity tag that binds to an inexpensive, chemically and hydrodynamically robust chromatography media is therefore highly desirable. 1.2.7 Carbohydrate Binding Modules Carbohydrate binding modules (CBMs) are discrete protein modules found in a large number of hydrolases involved in the degradation of biomass (Gilkes et al. 1991; Tomme et al. 1995b). Within these carbohydrolases, the CBMs function to bind the enzyme to its substrate and in so doing, increase the effective substrate concentration proximal to the enzyme (Warren 1996). CBMs are also found in a few non-hydrolytic proteins (Goldstein et al. 1993; Shoseyov and Doi 1990; Tomme et al. 1995b) where they function as part of a scaffolding subunit that organizes synergistic enzymes into a cohesive multienzyme complex called a cellulosome, which then binds strongly to cellulose to catalyze its degradation (Beguin and Aubert 1994). i More than 200 CBMs are known and can be classified into forty eight unique families, seventeen of which are known to contain members that bind cellulose [Carbohydrate Active Enzymes database: (http://www.cazy.org/) (Coutinho and Henrissat 1999)]. CBMs from different families have different binding affinities and binding specificities (Tomme et al. 1998). For example, CfXynl0A-CBM2a (hereafter referred to as CBM2a), a family 2 C B M from xylanase 10A of Cellulomonas fimi, binds crystalline regions of insoluble cellulose with a dissociation constant in the low micromolar range (McLean 2000). In contrast, the family 4 C B M , CfCel9B-CBM4-l from endoglucanase 9B of C. fimi, binds amorphous cellulose and water-soluble cello-15 oligosaccharides, but shows no affinity for crystalline cellulose (Abou-Hachem et al. 2000). CBMs come in a wide range of sizes from larger domains of up to 180 amino acids (family III) to small (~35 amino acids) fungal domains (family 1 CBMs) and are found in various positions within the parent carbohydrolase. As shown in Figure 1.1, Cel9A from Cellulomonas fimi has an internal family 3 C B M and a C-terminal family 2a C B M , while Cel9B, also from Cellulomonas fimi, has two family 4 CBMs in tandom located on the N-terminal side of its catalytic module. CfCel6A has an N-terminal family 2a C B M binding module while CfXynlOA has a family 2a C B M at the C-terminus. In addition, naturally occurring N-terminal CBMs have been translocated to the C-terminus and used as C-terminal affinity tags in purification of recombinant protein without loss of binding affinity (Tomme et al. 1994). The ability to use a C B M as either a N - or C-terminal tag offers flexibility in the purification of recombinant proteins, especially in cases where the protein of interest can only tolerate foreign sequences at one end to remain active. CBMs have been used as tags for immobilization or purification of a number of recombinant proteins. For example, CBMs have been fused to a number of bioprocessing enzymes such as E. coli alkaline phosphatase (Greenwood et al. 1992), Agrobacterium P-glucosidase (Ong 1991) and human factor X (Assouline et al. 1993), as well as to medically relevant enzymes such as heparinase (Shpigel 1999) and interleukin-2 (Ong 1995). Capture agents such as Streptomyces streptavidin (Le 1994) and Staphylococcus protein A (Ramirez et al. 1993) have been fused to a C B M to permit their immobilization on cellulose for binding of the target ligand. CBMs have also been expressed in a number of different hosts, including E. coli (Greenwood et al. 1992; Le 1994; Ong et al. 1989; Ramirez et al. 1993; Shpigel 1999; Shpigel 2000), the yeast P. pastoris (Doheny et al. 1999; Ong 1995), mammalian cells (Assouline et al. 1993; Guarna et al. 2000; Ong 1995), and insect cells (Pfeifer and Theilmann 2001). Although not well-understood, periplasmic expression of fusion proteins with a C B M from C. fimi often result in their non-specific leakage into the culture supernatant (Hasenwinkle et al. 1997; Ong et al. 1993; Tomme et al. 1995a). This property offers great potential not only for expression 16 of proteins with complex disulfide bond formation but would also simplify any purification process by directing the protein into an environment with less protein contaminant. 1.2.8 The Family 9 Carbohydrate-Binding Module (CBM9) While CBMs from some families can and have served as affinity tags for recombinant protein purification at the laboratory scale (Ong 1995; Sakka 1998; Shpigel 1999), the family 9 carbohydrate-binding module (CBM9) of xylanase 10A from the thermophilic bacterium Thermotoga maritima (Winterhalter et al. 1995) has not; yet it offers some unique functional characteristics that appear to make it a particularly useful tag for preparative-scale affinity purification applications. In particular, CBM9 is able to bind to both soluble sugars and insoluble cellulose (amorphous cellulose) (Boraston et al. 2001). As a result, CBM9 bound to insoluble preparations of cellulose can be dissociated from the solid substrate by the addition of a soluble sugar, such as glucose. The binding specificity of CBM9 has recently been characterized (Boraston et al. 2001) and its crystal structure determined up to 1.9 A resolution (Notenboom et al. 2001), providing a sound fundamental basis for designing an efficient purification technology based on the use of CBM9 as an affinity fusion tag. CBM9 is an extremely thermostable, 22 kDa globular protein (Wassenberg et al. 1997) with the unique ability to bind to the reducing ends of cellulose and soluble polysaccharides. It binds three C a + 2 ions, thought to contribute to the unusually high thermal stability of this protein (Notenboom et al. 2001). Its structure consists of two twisted anti-parallel P-sheets, each made up of five P-strands that form a P-sandwich fold (Figure 1.2). The crystal structure of cellobiose bound CBM9 indicates that two tryptophan residues function to sandwich the substrate within the binding pocket (Figure 1.2). Thermodynamic (Boraston et al. 2001) and structural analyses (Notenboom et al. 2001) confirm that the disaccharide cellobiose is sufficient to fully occupy the binding site. 17 1.2.9 Cellulose-Based Chromatography Resins Beaded cellulose and various derivatives of cellulose are currently used as stationary phase media for a number of commercial liquid chromatography applications. Size-exclusion chromatography media comprised of cellulose include Whatman CC31, which is a pure microgranular cellulose powder, Whatman CF1, a particulate microcystalline cellulose optimized for batch chromatography, Perloza\u00E2\u0084\u00A2 MT, a beaded cellulose gel filtration media, Perloza\u00E2\u0084\u00A2 ST, a dry regenerated cellulose that can be used as initial material for the preparation of all basic types of ion exchange and for a number of further modified derivatives of spherical cellulose, Perloza\u00E2\u0084\u00A2 SF, a pure dry beaded cellulose that has been sterilised by radiation, and Prolinx Versalinx\u00E2\u0084\u00A2, a cross-linked cellulose network. Most of these media are available in a range of bead sizes and porosities to facilitate optimization of the chromatographic separation for a given feedstock. More importantly, these cellulose-based media are generally very inexpensive, can withstand stringent clean-in-place (CIP) protocols, and have good mechanical properties. For example, the bulk cost of Perloza\u00E2\u0084\u00A2 MT100 is $35 USD per liter, or about 1/100 the bulk cost of an equivalent amount of the glutathione affinity resin used in conjunction with the GST tag system. 1.3 Thesis Objectives Generic affinity chromatography based on affinity tag technology has the potential to simplify downstream processing by achieving higher yields and purities than conventional methods. However, the high cost of current affinity tag technologies, due mainly to the expense of their associated affinity chromatography media, limits their application at production scales. This thesis aims to alleviate this cost limitation by introducing an affinity tag that binds to a very inexpensive cellulose-based matrix. Development of this technology will involve the following research objectives: (i) design of a generic CBM9 expression vector for production of chimeric fusions containing an N -terminal CBM9, a linker region containing a suitable processing site at its C-terminus for efficient removal of the affinity tag following affinity purification, and a C-terminal target protein; (ii) validate the performance of the affinity system in terms of product 18 yield, purity and concentration factor; (iii) develop a useful strategy to design an effective linker sequence to stably connect the CBM9 tag to the target protein and to permit efficient tag removal through enzyme-catalyzed cleavage; (iv) derive and validate a mathematical model to predict binding and elution behavior of CBM9 fusion proteins on a Perloza\u00E2\u0084\u00A2 MT100 column; and finally, (v) briefly address technology scale-up issues, including the synthesis of a mechanically stable stationary phase. The first objective will provide industry with a ready-made expression vector that can be used to express any target protein as a CBM9 fusion protein. The second objective will provide evidence that my proposed technology provides for efficient production and purification of a target recombinant protein with an overall level of performance that is either superior or comparable to current commercially available fusion-tag technologies. Although often overlooked, the linker is an important component of any affinity tag technology. The third objective will look at a number of strategies for selecting an optimal linker that will provide for in vivo stability, efficient processing with the serine protease rhFactor X a or enterokinase, and independent functioning of the two domains. The forth and fifth objectives will help industry with the application of the CBM9 fusion tag technology. 19 1.4 Tables Table 1.1 Affinity tag systems described in the literature. Tag size and complementary ligand immobilized on the stationary phase are given. Tags that are currently offered commercially are denoted by *. Affinity tag Size Immobilized Ligand or Matrix Reference Polypeptide-binding proteins Staphylococcal Protein* A (SpA) ZZ domains Albumin-binding protein Phosphate-binding domain 14-31 kDa 7 kDa 5-25 kDa 200-263 kDa IgG IgG HSA Hydroxyapatite (Moks et al. 1987; Nilsson and Abrahmsen 1990; Uhlen etal. 1983) (Hansson etal. 1995; Nilsson etal. 1996) (Nygren et al. 1988) (Anbaetal. 1987) Carbohydrate-binding domains Maltose-binding protein* (MBP) Starch-binding domain Chitin-binding domain* Cellulose-binding domain (families 1, 2a* and 3*) 40 kDa 119aa 51 aa 27- 189 aa Cross-linked amylose Starch Chitin Cellulose (Bedouelle and Duplay 1988; di Guanet al. 1988) (Chen etal. 1991a; Chen etal. 1991b) (Chongetal. 1997; Mathys et al. 1999) (Greenwood etal. 1989; Ong 1995; Tomme et al. 1998) Enzymes P-Galactosidase Glutathione-S-transferase* (GST) Chloramphenicol acetyltransferase 116 kDa 26 kDa 24 kDa A P T G Glutathione chloramphenicol (Germino and Bastia 1984; Oliaroetal. 2000) (Guan and Dixon 1991; Smith and Johnson 1988) (Dekeyzer et al. 1994; Dykes etal. 1988) 20 Affinity tag Size Immobilized Ligand or Matrix Reference Biotin-binding domains Biotin-binding domain Avidin 8kDa 63 kDa Streptavidin 2-iminobiotin-agarose (Cronan 1990) (Airenne etal. 1999) Metal-affinity tags Poly-Histidine* ProHissPro HAT* 6-10 aa 7 aa 19 aa N i + 2 - N T A ; C o + 2 - C M A N i + 2 - N T A ; Z n + 2 - N T A N i + 2 - N T A (Crowe etal. 1994; Porathetal. 1975) (Skerra et al. 1991) (Chagaet al. 1999) Non-biotinylated affinity tags Strep-tag II* Streptavidin-binding peptide* (SBP) 8 aa 38 aa streptavidin Streptavidin (Schmidt 1996; Voss and Skerra 1997) (Keefe2001; Wilson et al. 2001) Antigenic epitopes (Ca+2 dependent) F L A G * F L A G II* HPC4-binding peptide 8 aa 4 aa 12 aa Anti-FLAG M l mAb Anti -FLAG M2 mAb Anti-KT3 mAb (Einhauer and Jungbauer 2001; Hopp 1988) (Brizzard et al. 1994) (Martinet al. 1990) Antigenic epitopes (Ca+2 independent) Myc T7 AU5 Btag 8-tag 10 aa 11 aa 6 aa 6 aa 12 aa CA5 mAb Anti-T7 mAb Anti-AU5 mAb Anti-Btag mAb Anti-e-tag mAb (Chenet al. 1993; Munro and Pelham 1986) (Borjigin and Nathans 1994) (Lutzfreyermuth et al. 1990) (Rubinfeld et al. 1991; Wang etal. 1996) (Olahetal. 1994) 21 Affinity tag Size Immobilized Ligand or Matrix Reference V S V tag 11 aa Anti-VSV mAb (Green etal. 1996) RecA 144 aa Anti-recA mAb (Krivietal . 1985) Charged amino acid tags Poly-Arginine 5-15 aa Anionic resins (Sassenfeld 1984; Smith 1984) Poly-Phenylalanine 11 aa Phenyl-superose (Persson et al. 1988) Other tags Calmodulin-binding peptide* (CBP) 2.96 Calmodulin (Stofko-Hahn et al. 1992; Vaillancourt et al. 1997; Zheng etal. 1997) S-peptide 15 aa S-protein (Kim and Raines 1993; Raines et al. 2000) Heterodimeric coiled-coil tags 35 aa Immobilized hexa-peptide (Tripet et al. 2000; Tripetetal. 1996) 22 Table 1.2 Estimated costs of several commercially available affinity tag technologies. Prices provided by respective vendors. Estimated Cost (USD) System Resin (100 mL) Eluent ($/L) CBP $2200 $0.86 (2 m M EGTA) GST $1067 $10.07 (10 m M reduced glutatione) Polyhistidine $656 $2.87 (250 m M imidazole) M B P $300 $0.71 (10 m M maltose) 23 1.5 Figures Cel6B Cel9A Cel9B Cel5A 111111 2a Cel6A 2a Cel48A XynlOA 2a Catalytic Module Fn3 Module Binding Module Other Module | Linker Figure 1.1 Schematic representation of the domain structure of Cellulomonas fimi cellulases and xylanases. Note: the domain structure of xylanase 10A of T. maritima is homologous to that of XynlOA of C. fimi with the family 2a CBM9 substituted by CBM9. 24 T r P 1 7 5 Asn 172 T r P 7 1 Arg 161 Gin 151 Gin 96 Figure 1.2 Backbone secondary structure of C B M 9 from T. maritima, presented in ribbon diagram form, determined from the x-ray crystallographic data of Notenboom (Notenboom et al. 2001). Trp71 and T ip 175 (solid arrow) involved in ligand binding, and Glu77, Gln96, Arg98, G l n l 5 1 , A r g l 6 1 , and A s n l 7 2 (dashed arrow), involved in hydrogen bonding, with the ligand are shown. 25 1.6 References Abou-Hachem, M . ; Nordberg-Karlsson, E.; Bartonek-Roxa, E.; Raghothama, S.; Simpson, P. J.; Gilbert, H. J.; Williamson, M . P.jHolst, O. (2000). \"Carbohydrate-binding modules from a thermostable Rhodothermus marinus xylanase: cloning, expression and binding studies\". Biochemical Journal 345(l):53-60. Abrahmsen, L . ; Moks, T.; Nilsson, B.;Uhlen, M . (1986). \"Secretion of Heterologous Gene-Products to the Culture-Medium of Escherichia-Coli\". Nucleic Acids Research 14(18):7487-7500. Airenne, K. J.; Marjomaki, V . S.;Kulomaa, M . S. (1999). \"Recombinant avidin and avidin-fusion proteins\". Biomolecular Engineering 16(l-4):87-92. Allet, B.; Bernard, A . R.; Hochmann, A. ; Rohrbach, E.; Graber, P.; Magnenat, E.; Mazzei, G. J.;Bernasconi, L. (1997). \"A bacterial signal peptide directs efficient secretion of eukaryotic proteins in the baculovirus expression system\". Protein Expression and Purification 9(l):61-68. Anba, J.; Baty, D.; Lloubes, R.; Pages, J. M . ; Joseph-Liauzun, E.; Shire, D.; Roskam, W.;Lazdunski, C. (1987). \"Expression vector promoting the synthesis and export of the human growth-hormone-releasing factor in Escherichia coli\". Gene 53(2-3):219-226. Andrade, E. V. ; Albuquerque, F. C ; Moraes, L. M . ; Brigido, M . M.;Santos-Silva, M . A . (2000). \"Single-chain Fv with Fc fragment of the human IgGl tag: construction, Pichia pastoris expression and antigen binding characterization\". Journal of Biochemistry 128(6):891-895. Assouline, Z.; Shen, H . ; Kilburn, D. G.;Warren, R. A . J. (1993). \"Production and Properties of a Factor-X Cellulose-Binding Domain Fusion Protein\". Protein Engineering 6(7):787-792. Bardwell, J. C. A . ; Mcgovern, K.;Beckwith, J. (1991). \"Identification of a Protein Required for Disulfide Bond Formation Invivo\". Cell 67(3):581-589. Bedouelle, H.;Duplay, P. (1988). \"Production in Escherichia coli and one-step purification of bifunctional hybrid proteins which bind maltose. Export of the Klenow polymerase into the periplasmic space\". European Journal of Biochemistry 171(3):541-549. Beguin, P.; Aubert, J. P. (1994). \"The Biological Degradation of Cellulose\". FEMS Microbiology Reviews 13(l):25-58. 26 Blondel, A. ; Bedouelle, H . (1990). \"Export and Purification of a Cytoplasmic Dimeric Protein by Fusion to the Maltose-Binding Protein of Escherichia-Coli\". European Journal of Biochemistry 193(2):325-330. Bonnerjea, J.; Oh, S.; Hoare, M . ; Dunnill, P. (1986). \"Protein purification: the right step at the right time\". Bio/Technology 4(11):954, 956, 958. Boraston, A . B.; Creagh, A . L. ; Alam, M . M . ; Kormos, J. M . ; Tomme, P.; Haynes, C. A. ; Warren, R. A.;Kilburn, D. G. (2001). \"Binding specificity and thermodynamics of a family 9 carbohydrate-binding module from Thermotoga maritima xylanase 10A\". Biochemistry 40(21):6240-6247. Borjigin, J.; Nathans, J. (1994). \"Insertional Mutagenesis as a Probe of Rhodopsins Topography, Stability, and Activity\". Journal of Biological Chemistry 269(20): 14715-14722. Brewer, S. J.; Haymore, B. L. ; Hopp, T. P.; Sassenfeld, H. M . (1991). \"Engineering proteins to enable their isolation in a biologically active form\". Bioprocess Technology 12:239-266. Brizzard, B. L. ; Chubet, R. G.;Vizard, D. L. (1994). \"Immunoaffinity purification of F L A G epitope-tagged bacterial alkaline phosphatase using a novel monoclonal antibody and peptide elution\". Biotechniques 16(4):730-732, 734-735. Capon, D. J.; Chamow, S. M . ; Mordenti, J.; Marsters, S. A . ; Gregory, T.; Mitsuya, H. ; Byrn, R. A. ; Lucas, C ; Wurm, F. M . ; Groopman, J. E.; Broder, S.; Smith, D. H. (1989). \"Designing Cd4 Immunoadhesins for Aids Therapy\". Nature 337(6207):525-531. Carter, P.; Kelley, R. F.; Rodrigues, M . L. ; Snedecor, B.; Covarrubias, M . ; Velligan, M . D.; Wong, W. L. T.; Rowland, A . M . ; Kotts, C. E.; Carver, M . E.; Yang, M . ; Bourell, J. H. ; Shepard, H. M . ; Henner, D. (1992). \"High-Level Escherichia-Coli Expression and Production of a Bivalent Humanized Antibody Fragment\". Bio/Technology 10(2): 163-167. Casey, J. L. ; Keep, P. A . ; Chester, K. A. ; Robson, L.; Hawkins, R. E.; Begent, R. H. J. (1995). \"Purification of Bacterially Expressed Single-Chain Fv Antibodies for Clinical-Applications Using Metal Chelate Chromatography\". Journal of Immunological Methods 179(1): 105-116. Chaga, G.; Bochkariov, D. E.; Jokhadze, G. G.; Hopp, J.; Nelson, P. (1999). \"Natural poly-histidine affinity tag for purification of recombinant proteins on cobalt(II)-carboxymethylaspartate crosslinked agarose\". Journal of Chromatography A 864(2):247-256. 27 Chao, H. M . ; Bautista, D. L. ; Litowski, J.; Irvin, R. T.; Hodges, R. S. (1998). \"Use of a heterodimeric coiled-coil system for biosensor application and affinity purification\". Journal of Chromatography B 715(l):307-329. Chen, C ; Billingsley, P. F. (1999). \"Detection and characterization of a mannan-binding lectin from the mosquito, Anopheles stephensi (Liston)\". European Journal of Biochemistry 263(2):360-366. Chen, C ; Huang, Q.-L.; Jiang, S.-H.; Pan, X.;Hua, Z.-C. (2006). \"Immobilized protein ZZ, an affinity tool for immunoglobulin isolation and immunological experimentation\". Biotechnology and Applied Biochemistry 45(2):87-92. Chen, L. J.; Ford, C ; Kusnadi, A.;Nikolov, Z. L. (1991a). \"Improved Adsorption to Starch of a Beta-Galactosidase Fusion Protein Containing the Starch-Binding Domain from Aspergillus Glucoamylase\". Biotechnology Progress 7(3):225-229. Chen, L. J.; Ford, C.;Nikolov, Z. (1991b). \"Adsorption to Starch of a Beta-Galactosidase Fusion Protein Containing the Starch-Binding Region of Aspergillus Glucoamylase\". Gene 99(1):121-126. Chen, Y . T.; Holcomb, C.;Moore, H. P. H. (1993). \"Expression and localization of two molecular weight GTP-binding proteins, Rab8 and RablO, by epitope tag\". Proceedings of the National Academy of Sciences of the United States of America 90(14):6508-6512. Chong, S.; Mersha, F. B.; Comb, D. G.; Scott, M . E.; Landry, D.; Vence, L. M . ; Perler, F. B.; Benner, J.; Kucera, R. B.; Hirvonen, C. A. ; Pelletier, J. J.; Paulus, H.;Xu, M . Q. (1997). \"Single-column purification of free recombinant proteins using a self-cleavable affinity tag derived from a protein splicing element\". Gene 192(2):271-281. Chubet, R. G.;Brizzard, B. L . (1996). \"Vectors for expression and secretion of F L A G epitope-tagged proteins in mammalian cells\". Biotechniques 20(1): 136-141. Clonis, Y . D.; Labrou, N . E.; Kotsira, V . P.; Mazitsos, C ; Melissis, S.;Gogolas, G. (2000). \"Biomimetic dyes as affinity chromatography tools in enzyme purification\". Journal of Chromatography, A 891(l):33-44. Coutinho, P. M.;Henrissat, B. (1999). \"Carbohydrate-active enzymes: an integrated database approach\". Special Publication - Royal Society of Chemistry 246(Recent Advances in Carbohydrate Bioengineering):3-12. Cronan, J. E., Jr. (1990). \"Biotination of proteins in vivo. A post-translational modification to label, purify, and study proteins\". Journal of Biological Chemistry 265(18):10327-10333. 28 Crowe, J.; Dobeli, H. ; Gentz, R.; Hochuli, E.; Stuber, D.;Henco, K. (1994). \"6xHis-Ni-N T A chromatography as a superior technique in recombinant protein expression/purification\". Methods in Molecular Biology 31:371-387. Cummings, R. D. (1997). \"Lectins as tools for glycoconjugate purification and characterization\". Glycosciences: 191-199. Dekeyzer, N . ; Engelborghs, Y.;Volckaert, G. (1994). \"Cloning, expression and purification of a sarcoplasmic calcium-binding protein from the sandworm Nereis diversicolor via a fusion product with chloramphenicol acetyltransferase\". Protein Engineering 7(1): 125-130. Denizli, A.;Piskin, E. (2001). \"Dye-ligand affinity systems\". Journal of Biochemical and Biophysical Methods 49(1-3):391-416. Derbyshire, V . ; Wood, D. W.; Wu, W.; Dansereau, J. T.; Dalgaard, J. Z.;Belfort, M . (1997). \"Genetic definition of a protein-splicing domain: Functional mini-inteins support structure predictions and a model for intein evolution\". Proceedings of the National Academy of Sciences of the United States of America 94(21): 11466-11471. Deutsch, D. G.;Mertz, E. T. (1970). \"Plasminogen: purification from human plasma by affinity chromatography\". Science 170(962): 1095-1096. di Guan, C ; L i , P.; Riggs, P. D.;Inouye, H. (1988). \"Vectors that facilitate the expression and purification of foreign peptides in Escherichia coli by fusion to maltose-binding protein\". Gene 67(l):21-30. Doheny, J. G.; Jervis, E. J.; Guarna, M . M . ; Humphries, R. K. ; Warren, R. A . J.;Kilburn, D. G. (1999). \"Cellulose as an inert matrix for presenting cytokines to target cells: production and properties of a stem cell factor-cellulose-binding domain fusion protein\". Biochemical Journal 339(2):429-434. Dykes, C. W.; Bookless, A. B.; Coomber, B. A. ; Noble, S. A. ; Humber, D. C.;Hobden, A . N . (1988). \"Expression of atrial natriuretic factor as a cleavable fusion protein with chloramphenicol acetyltransferase in Escherichia coli\". European Journal of Biochemistry 174(2):411-416. Einhauer, A.;Jungbauer, A . (2001). \"The F L A G peptide, a versatile fusion tag for the purification of recombinant proteins\". Journal of Biochemical and Biophysical Methods 49(l-3):455-465. Fassina, G.; Verdoliva, A. ; Palombo, G.; Ruvo, M.;Cassani, G. (1998). \"Immunoglobulin specificity of TGI9318: a novel synthetic ligand for antibody affinity purification\". Journal of Molecular Recognition 11 (1 -6): 128-133. 29 Flaschel, E.;Friehs, K . (1993). \"Improvement of Downstream Processing of Recombinant Proteins by Means of Genetic-Engineering Methods\". Biotechnology Advances ll(l):31-77. Follman, D. K.;Fahrner, R. L. (2004). \"Factorial screening of antibody purification processes using three chromatography steps without protein A\" . Journal of Chromatography, A 1024(1-2):79-85. Ford, C. F.; Suominen, I.;Glatz, C. E. (1991). \"Fusion tails for the recovery and purification of recombinant proteins\". Protein Expression and Purification 2(2-3):95-107. Germino, J.;Bastia, D. (1984). \"Rapid Purification of a Cloned Gene-Product by Genetic Fusion and Site-Specific Proteolysis\". Proceedings of the National Academy of Sciences of the United States of America-Biological Sciences 81(15):4692-4696. Gilkes, N . R.; Henrissat, B.; Kilburn, D. G.; Miller, R. C , Jr.;Warren, R. A . J. (1991). \"Domains in microbial beta-1, 4-glycanases: sequence conservation, function, and enzyme families\". Microbiological Reviews 55(2):303-315. Goel, A. ; Colcher, D.; Koo, J. S.; Booth, B. J. M . ; Pavlinkova, G.;Batra, S. K. (2000). \"Relative position of the hexahistidine tag effects binding properties of a tumor-associated single-chain Fv construct\". Biochimica Et Biophysica Acta-General Subjects 1523(l):13-20. Goldstein, M . A . ; Takagi, M . ; Hashida, S.; Shoseyov, O.; Doi, R. H.;Segel, I. H . (1993). \"Characterization of the Cellulose-Binding Domain of the Clostridium-Cellulovorans Cellulose-Binding Protein-A\". Journal of Bacteriology 175(18):5762-5768. Graslund, T.; Nilsson, J.; Lindberg, A . M . ; Uhlen, M.;Nygren, P. A . (1997). \"Production of a thermostable D N A polymerase by sites-specific cleavage of a heat-eluted affinity fusion protein\". Protein Expression and Purification 9(1): 125-132. Green, E. A . ; Botting, C ; Webb, H . M . ; Hirst, T. R.;Randall, R. E. (1996). \"Construction, purification and immunogenicity of antigen-antibody-LTB complexes\". Vaccine 14(10):949-958. Greenwood, J. M . ; Gilkes, N . R.; Kilburn, D. G.; Miller, R. C , Jr.;Warren, R. A. J. (1989). \"Fusion to an endoglucanase allows alkaline phosphatase to bind to cellulose\". FEBS Letters 244(1):127-131. Greenwood, J. M . ; Ong, E.; Gilkes, N . R.; Warren, R. A . J.; Miller, R. C.;Kilburn, D. G. (1992). \"Cellulose-Binding Domains - Potential for Purification of Complex Proteins\". Protein Engineering 5(4):361-365. 30 Guan, K. L.;Dixon, J. E. (1991). \"Eukaryotic proteins expressed in Escherichia coli: an improved thrombin cleavage and purification procedure of fusion proteins with glutathione S-transferase\". Analytical Biochemistry 192(2):262-267. Guarna, M . M . ; Cote, H. C ; Kwan, E. M . ; Rintoul, G. L. ; Meyhack, B.; Heim, J.; MacGillivray, R. T.; Warren, R. A . J.;Kilburn, D. G. (2000). \"Factor X fusion proteins: improved production and use in the release in vitro of biologically active hirudin from an inactive alpha-factor-hirudin fusion protein\". Protein Expression and Purification 20(2): 133-141. Hahn, R.; Shimahara, K. ; Steindl, F.;Jungbauer, A . (2006). \"Comparison of protein A affinity sorbents III. Life time study\". Journal of Chromatography, A 1102(1-2):224-231. Hakes, D. J.;Dixon, J. E. (1992). \"New Vectors for High-Level Expression of Recombinant Proteins in Bacteria\". Analytical biochemistry 202(2):293-298. Hammarberg, B. ; Nygren, P. A. ; Holmgren, E.; Elmblad, A. ; Tally, M . ; Hellman, U . ; Moks, T.;Uhlen, M . (1989). \"Dual affinity fusion approach and its use to express recombinant human insulin-like growth factor II\". Proceedings of the National Academy of Sciences of the United States of America 86(12):4367-4371. Hansen, M . B.; Lihme, A . ; Spitali, M.;King, D. (1998). \"Capture of human Fab fragments by expanded bed adsorption with a mixed mode adsorbent\". Bioseparation 8(1-5):189-193. Hansson, L. ; Noppa, L. ; Nilsson, A . K. ; Stromqvist, M.;Bergstrom, S. (1995). \"Expression of Truncated and Full-Length Forms of the Lyme-Disease Borrelia Outer Surface Protein-a in Escherichia-Coli\". Protein Expression and Purification 6(l):15-24. Hasenwinkle, D.; Jervis, E.; Kops, O.; Liu, C ; Lesnicki, G.; Haynes, C. A.;Kilburn, D. G. (1997). \"Very high-level production and export in Escherichia coli of a cellulose binding domain for use in a generic secretion-affinity fusion system\". Biotechnology and Bioengineering 55(6):854-863. Hearn, M . T.;Acosta, D. (2001). \"Applications of novel affinity cassette methods: use of peptide fusion handles for the purification of recombinant proteins\". Journal of Molecular Recognition 14(6):323-369. Hearn, M . T. W.;Anspach, B. (1990). \"Chemical, physical, and biochemical concepts in isolation and purification of proteins\". Bioprocess Technology 9(Separation Processes Biotechnology): 17-64. 31 Hearn, M . T. W.;Gomme, P. T. (2000). \"Molecular architecture and biorecognition processes of the cystine knot protein superfamily: Part I. The glycoprotein hormones\". Journal of Molecular Recognition 13(5):223-278. Hefti, M . H.; Van Vugt-Van der Toorn, C. J. G.; Dixon, R.;Vervoort, J. (2001). \"A novel purification method for histidine-tagged proteins containing a thrombin cleavage site\". Analytical Biochemistry 295(2): 180-185. Pfennig, L.;Schafer, E. (1998). \"Protein purification with C-terminal fusion of maltose binding protein\". Protein Expression and Purification 14(3):367-370. Hentz, N . G.; Vukasinovic, V.;Daunert, S. (1996). \"Affinity chromatography of recombinant peptides proteins based on a calmodulin fusion tail\". Analytical Chemistry 68(9): 1550-1555. Hermanson, I. L.;Turchi, J. J. (2000). \"Overexpression and purification of human X P A using a Baculovirus expression system\". Protein Expression and Purification 19(1):1-11. Higgins, D. R.;Cregg, J. M . (1998). \"Introduction to Pichia pastoris\". Methods in Molecular Biology 103(Pichia Protocols): 1-15. Hochuli, E.; Dobeli, H.;Schacher, A . (1987). \"New metal chelate adsorbent selective for proteins and peptides containing neighbouring histidine residues\". Journal of Chromatography 411:177-184. Hockney, R. C. (1994). \"Recent Developments in Heterologous Protein-Production in Escherichia-Coli\". Trends in Biotechnology 12(ll):456-463. Hoffmann, A. ; Roeder, R. G. (1991). \"Purification of His-Tagged Proteins in Nondenaturing Conditions Suggests a Convenient Method for Protein-Interaction Studies\". Nucleic Acids Research 19(22):6337-6338. Hofmann, K. ; Wood, S. W.; Brinton, C. C ; Montibeller, J. A.;Finn, F. M . (1980). \"Iminobiotin affinity columns and their application to retrieval of streptavidin\". Proceedings of the National Academy of Sciences of the United States of America 77(8):4666-4668. Hopp, T., Pricket, KS , Price, V L , Libby, RT, March, CJ, Ceretti, DP, Urdal, DL, Conlon, PJ. (1988). \"A short polypeptide marker sequence useful for recombinant protein identification and purification\". Bio/Technology 6:1204-1210. Hosfield, T.; Lu, Q. (1999). \"Influence of the amino acid residue downstream of (Asp)(4)Lys on enterokinase cleavage of a fusion protein\". Analytical Biochemistry 269(1): 10-16. 32 Itakura, K.; Hirose, T.; Crea, R.; Riggs, A . D.; Heyneker, H . L. ; Bolivar, F.;Boyer, H. W. (1977). \"Expression in Escherichia-Coli of a Chemically Synthesized Gene for Hormone Somatostatin\". Science 198(4321):1056-1063. Jonasson, P.; Nilsson, J.; Samuelsson, E.; Moks, T.; Stahl, S.;Uhlen, M . (1996). \"Single-step trypsin cleavage of a fusion protein to obtain human insulin and its C peptide\". European Journal of Biochemistry 236(2):656-661. Jones, C ; Patel, A . ; Griffin, S.; Martin, J.; Young, P.; Odonnell, K . ; Silverman, C ; Porter, T.;Chaiken, I. (1995). \"Current Trends in Molecular Recognition and Bioseparation\". Journal of Chromatography A 707(l):3-22. Jones, K. (1990). \"Affinity chromatography: a technology update\". American Biotechnology Laboratory 8(13):26-30. Jungbauer, A.;Hahn, R. (2004). \"Engineering protein A affinity chromatography\". Current Opinion in Drug Discovery & Development 7(2):248-256. Kane, J. F.;Hartley, D. L. (1988). \"Formation of Recombinant Protein Inclusion-Bodies in Escherichia-Coli\". Trends in Biotechnology 6(5):95-101. Kaplan, W.; Husler, P.; Klump, H. ; Erhardt, J.; SluisCremer, N.;Dirr, H . (1997). \"Conformational stability of pGEX-expressed Schistosoma japonicum glutathione S-transferase: A detoxification enzyme and fusion-protein affinity tag\". Protein Science 6(2):399-406. Kapust, R. B., Waugh, D. S. (1999). \"Escherichia coli maltose-binding protein is uncommonly effective at promoting the solubility of polypeptides to which it is fused\". Protein Science 8(8):1668-1674. Keefe, A . D., Wilson, D. S., Seelig, B., Szostak, J. W. (2001). \"One-step purification of recombinant proteins using a nanomolar-affinity streptavidin-binding peptide, the SBP-Tag\". Protein Expression and Purification 23(3):440-446. Kellermann, O. K.;Ferenci, T. (1982). \"Maltose-Binding Protein from Escherichia-Coli\". Methods in Enzymology 90:459-463. Kim, J. S.;Raines, R. T. (1993). \"Ribonuclease S-Peptide as a Carrier in Fusion Proteins\". Protein Science 2(3):348-356. Kleiner, O.; Butenandt, L. ; Carell, T.;Batschauer, A . (1999). \"Class II D N A photolyase from Arabidopsis thaliana contains F A D as a cofactor\". European Journal of Biochemistry 264(1): 161-167. Krivi, G. G.; Bittner, M . L.; Rowold, E., Jr.; Wong, E. Y . ; Glenn, K. C ; Rose, K. S.;Tiemeier, D. C. (1985). \"Purification of recA-based fusion proteins by 33 immunoadsorbent chromatography. Characterization of a major antigenic determinant of Escherichia coli recA protein\". Journal of Biological Chemistry 260(18):10263-10267. Kronina, V . V. ; Wirth, H . J.;Hearn, M . T. W. (1999). \"Characterisation by immobilised metal ion affinity chromatographic procedures of the binding behaviour of several synthetic peptides designed to have high affinity for Cu(II) ions\". Journal of Chromatography A 852(1 ):261-272. Krueger, J. K. ; Kulke, M . H. ; Schutt, C.;Stock, J. (1989). \"Protein inclusion body formation and purification\". BioPharm Journal 2(3):40, 42-45. Kuusinen, A . ; Arvola, M . ; Okerblom, C.;Keinanen, K. (1995). \"Purification of Recombinant Glur-D Glutamate-Receptor Produced in Sf21 Insect Cells\". European Journal of Biochemistry 233(3):720-726. Ladisch, M . R. (2001). Bioseparations engineering : principles, practice, and economics. New York: Wiley. LaVallie, E. R., McCoy, J. M . (1995). \"Gene fusion expression systems in Escherichia coli\". Current Opinion in Biotechnology 6(5):501-506. Le, K. D., Gilkes, N . R., Kilburn, D. G., Miller, R. C , Jr., Saddler, J. N . , Warren, R. A . (1994). \"A streptavidin-cellulose-binding domain fusion protein that binds biotinylated proteins to cellulose\". Enzyme and Microbial Technology 16(6):496-500. Lee, W . - C ; Hsiao, C.-C.;Ruaan, R.-C. (1995). \"Affinity chromatography of glucose-specific lectin using silica-based support\". Journal of Chemical Technology and Biotechnology 64(l):66-72. Levine, S. P., Wohl, H . (1976). \"Human platelet factor 4: Purification and characterization by affinity chromatography. Purification of human platelet factor 4\". Journal of Biological Chemistry 251(2):324-328. Libby, R. T.; Cosman, D.; Cooney, M . K. ; Merriam, J. E.; March, C. J.;Hopp, T. P. (1988). \"Human Rhinovirus 3C Protease - Cloning and Expression of an Active Form in Escherichia-Coli\". Biochemistry 27(17):6262-6268. Lindner, P.; Guth, B.; Wuelfing, C ; Krebber, C ; Steipe, B. ; Mueller, F.;Plueckthun, A . (1992). \"Purification of native proteins from the cytoplasm and periplasm of Escherichia coli using IMAC and histidine tails: A comparison of proteins and protocols\". Methods 4(l):41-56. 34 Lo, K. M . ; Sudo, Y . ; Chen, J.; L i , Y . ; Lan, Y . ; Kong, S. M . ; Chen, L . L . ; An, Q.;Gillies, S. D. (1998). \"High level expression and secretion of Fc-X fusion proteins in mammalian cells\". Protein Engineering 11(6):495-500. Lowe, C. R.; Lowe, A . R.;Gupta, G. (2001). \"New developments in affinity chromatography with potential application in the production of biopharmaceuticals\". Journal of Biochemical and Biophysical Methods 49(1-3):561-574. Lu, T.; Vandyke, M.;Sawadogo, M . (1993). \"Protein-Protein Interaction Studies Using Immobilized Oligohistidine Fusion Proteins\". Analytical Biochemistry 213(2):318-322. Lutzfreyermuth, C ; Query, C. C.;Keene, J. D. (1990). \"Quantitative-Determination That One of 2 Potential Rna-Binding Domains of the a-Protein Component of the U l Small Nuclear Ribonucleoprotein Complex Binds with High-Affinity to Stem Loop-Ii of U l Rna\". Proceedings of the National Academy of Sciences of the United States of America 87(16):6393-6397. Makrides, S. C. (1996). \"Strategies for achieving high-level expression of genes in Escherichia coli\". Microbiological Reviews 60(3):512-538. Marino, M . H . (1989). \"Expression systems for heterologous protein production\". BioPharm Journal 2(7): 18-33. Martin, G. A. ; Viskochil, D.; Bollag, G.; McCabe, P. C ; Crosier, W. J.; Haubruck, H. ; Conroy, L. ; Clark, R.; O'Connell, P.;et al. (1990). \"The GAP-related domain of the neurofibromatosis type 1 gene product interacts with ras p21\". Cell 63(4):843-849. Mathys, S.; Evans, T. C , Jr.; Chute, I. C ; Wu, H.; Chong, S.; Benner, J.; Liu, X.-Q.;Xu, M.-Q. (1999). \"Characterization of a self-splicing mini-intein and its conversion into autocatalytic N - and C-terminal cleavage elements: facile production of protein building blocks for protein ligation\". Gene 231(1-2): 1-13. McLean, B. W., Bray, M . R., Boraston, A . B., Gilkes, N . R., Haynes, C. A. , Kilburn, D. G. (2000). \"Analysis of binding of the family 2a carbohydrate-binding module from Cellulomonas fimi xylanase 10A to cellulose: specificity and identification of functionally important amino acid residues\". Protein Engineering 13(11):801-809. Mcpherson, D. T.; Morrow, C ; Minehan, D. S.; Wu, J. G.; Hunter, E.;Urry, D. W. (1992). \"Production and Purification of a Recombinant Elastomeric Polypeptide, G-(Vpgvg)19-Vpgv, from Escherichia-Coli\". Biotechnology Progress 8(4):347-352. 35 Moks, T.; Abrahmsen, L. ; Holmgren, E.; Bilich, M . ; Olsson, A. ; Uhlen, M . ; Pohl, G.; Sterky, C ; Hultberg, H. ; Josephson, S.; Holmgren, A. ; Jornvall, H.;Nilsson, B. (1987). \"Expression of Human Insulin-Like Growth Factor-I in Bacteria - Use of Optimized Gene Fusion Vectors to Facilitate Protein-Purification\". Biochemistry 26(17):5239-5244. Mulcahy, P.; O'Flaherty, M . ; Jennings, L.;Griffin, T. (2002). \"Application of kinetic-based biospecific affinity chromatographic systems to ATP-dependent enzymes: studies with yeast hexokinase\". Analytical Biochemistry 309(2):279-292. Munro, S.;Pelham, H . R. B. (1986). \"An Hsp70-Like Protein in the Er - Identity with the 78 Kd Glucose-Regulated Protein and Immunoglobulin Heavy-Chain Binding-Protein\". Cell 46(2):291-300. Murby, M . ; Cedergren, L. ; Nilsson, J.; Nygren, P. A. ; Hammarberg, B. ; Nilsson, B.; Enfors, S. 0.;Uhlen, M . (1991). \"Stabilization of recombinant proteins from proteolytic degradation in Escherichia coli using a dual affinity fusion strategy\". Biotechnology and Applied Biochemistry 14(3):336-346. Narayanan, S. R. (1994). \"Preparative Affinity-Chromatography of Proteins\". Journal of Chromatography A 658(2):237-258. Nieba, L.; NiebaAxmann, S. E.; Persson, A. ; Hamalainen, M . ; Edebratt, F.; Hansson, A. ; Lidholm, J.; Magnusson, K. ; Karlsson, A . F.;Pluckthun, A . (1997). \"BIACORE analysis of histidine-tagged proteins using a chelating N T A sensor chip\". Analytical Biochemistry 252(2):217-228. Nilsson, B.;Abrahmsen, L. (1990). \"Fusions to staphylococcal protein A\" . Methods in Enzymology 185(Gene Expression Technology): 144-161. Nilsson, J.; Jonasson, P.; Samuelsson, E.; Stahl, S.;Uhlen, M . (1996). \"Integrated production of human insulin and its C-peptide\". Journal of Biotechnology 48(3):241-250. Nilsson, J.; Stahl, S.; Lundeberg, J.; Uhlen, M.;Nygren, P.-A. (1997). \"Affinity fusion strategies for detection, purification, and immobilization of recombinant proteins\". Protein Expression and Purification 11(1): 1-16. Notenboom, V. ; Boraston, A . B.; Kilburn, D. G.;Rose, D. R. (2001). \"Crystal structures of the family 9 carbohydrate-binding module from Thermotoga maritima xylanase 10A in native and ligand-bound forms\". Biochemistry 40(21):6248-6256. Nygren, P. A. ; Eliasson, M . ; Abrahmsen, L. ; Uhlen, M.;Palmcrantz, E. (1988). \"Analysis and use of the serum albumin binding domains of streptococcal protein G\". Journal of Molecular Recognition l(2):69-74. 36 Nygren, P. A. ; Stahl, S.;Uhlen, M . (1994). \"Engineering Proteins to Facilitate Bioprocessing\". Trends in Biotechnology 12(5): 184-188. Olah, Z.; Lehel, C ; Jakab, G.;Anderson, W. B. (1994). \"A Cloning and Epsilon-Epitope-Tagging Insert for the Expression of Polymerase Chain Reaction-Generated Cdna Fragments in Escherichia-Coli and Mammalian-Cells\". Analytical Biochemistry 221(1):94-102. Olejnik, J.; Krzymanska-Olejnik, E.;Rothschild, K. J. (1998). \"Photocleavable affinity tags for isolation and detection of biomolecules\". Methods in Enzymology 291 (Caged Compounds): 135-154. Oliaro, J.; Johnson, R. D.; Chen, W. X . ; Chadwick, V . S.;Murray, A . (2000). \"Identification of an immunogenic 18-kDa protein of Helicobacter pylori by alkaline phosphatase gene fusions\". Journal of Medical Microbiology 49(7):643-650. Ong, E., Alimonti, J. B., Greenwood, J. M . , Miller, R. C , Jr., Warren, R. A. , Kilburn, D. G. (1995). \"Purification of human interleukin-2 using the cellulose-binding domain of a prokaryotic cellulase\". Bioseparation 5(2):95-104. Ong, E.; Gilkes, N . R.; Miller, R. C ; Warren, R. A . J.;Kilburn, D. G. (1993). \"The Cellulose-Binding Domain (Cbdcex) of an Exoglucanase from Cellulomonas-Fimi - Production in Escherichia-Coli and Characterization of the Polypeptide\". Biotechnology and Bioengineering 42(4):401-409. Ong, E., Gilkes, N . R., Miller, R. C , Jr., Warren, A . J., Kilburn, D. G. (1991). \"Enzyme immobilization using a cellulose-binding domain: properties of a beta-glucosidase fusion protein\". Enzyme and Microbial Technology 13(l):59-65. Ong, E.; Greenwood, J. M . ; Gilkes, N . R.; Kilburn, D. G.; Miller, R. C.;Warren, R. A . (1989). \"The Cellulose-Binding Domains of Cellulases - Tools for Biotechnology\". Trends in Biotechnology 7(9):239-243. Ostermann, C , Dickmann, U . , Muley, T., Mader, M . (1990). \"Large-scale purification of choline acetyltransferase and production of highly specific antisera\". European Journal of Biochemistry 192(1):215-218. Pedersen, J.; Lauritzen, C ; Madsen, M . T.;Dahl, S. W. (1999). \"Removal of N-terminal polyhistidine tags from recombinant proteins using engineered aminopeptidases\". Protein Expression and Purification 15(3):389-400. Persson, M . ; Bergstrand, M . G.; Bulow, L.;Mosbach, K. (1988). \"Enzyme purification by genetically attached poly cysteine and polyphenylalanine affinity tails\". Analytical Biochemistry 172(2):330-337. 37 Pfeifer, T. A. , Guarna, M . M . , Kwan, E. M . , Lesnicki,, .;Theilmann, D. A. , Grigliatti, T. A. , Kilburn, D. G. (2001). \"Expression analysis of a modified factor X in stably transformed insect cell lines\". Protein Expression and Purification 23(2):233-241. Porath, J.; Carlsson, J.; Olsson, I.;J3elfrage, G. (1975). \"Metal chelate affinity chromatography, a new approach to protein fractionation\". Nature 258(5536):598-599. Prickett, K. S.; Amberg, D. C.;Hopp, T. P. (1989). \"A calcium-dependent antibody for identification and purification of recombinant proteins\". Biotechniques 7(6):580-589. Pryor, K. D., Leiting, B. (1997). \"High-level expression of soluble protein in Escherichia coli using a His6-tag and maltose-binding-protein double-affinity fusion system\". Protein Expression and Purification 10(3):309-319. Raines, R. T.; McCormick, M . ; Van Oosbree, T. R.;Mierendorf, R. C. (2000). \"The S tag fusion system for protein purification\". Applications of Chimeric Genes and Hybrid Proteins, Pt A 326:362-376. Rajamohan, F.; Doumbia, S. O.; Engstrom, C. R.; Pendergras, S. L. ; Maher, D. L.;Uckun, F. M . (2000). \"Expression of biologically active recombinant pokeweed antiviral protein in methylotrophic yeast Pichia pastoris\". Protein Expression and Purification 18(2): 193-201. Ramirez, C ; Fung, J.; Miller, R. C ; Antony, R.; Warren, J.;Kilburn, D. G. (1993). \"A Bifunctional Affinity Linker to Couple Antibodies to Cellulose\". Bio/Technology 11(13):1570-1573. Riggs, P. (2000). \"Expression and purification of recombinant proteins by fusion to maltose-binding protein\". Molecular Biotechnology 15(l):51-63. Rubinfeld, B.; Munemitsu, S.; Clark, R.; Conroy, L . ; Watt, K. ; Crosier, W. J.; McCormick, F.;Polakis, P. (1991). \"Molecular cloning of a GTPase activating protein specific for the Krev-1 protein p21 rap 1\". Cell 65(6): 1033-1042. Rudolph, R. (1996). \"Successful protein folding on an industrial scale.\" Protein Engineering: 2 83 -298. Sachdev, D.;Chirgwin, J. M . (1999). \"Properties of soluble fusions between mammalian aspartic proteinases and bacterial maltose-binding protein\". Journal of Protein Chemistry 18(1):127-136. Sakka, K. , Karita, S., Kimura, T., Ohmiya, K. (1998). \"Purification of a fusion protein using the family VI cellulose-binding domain of Clostridium stercorarium XynA\". Annals of the New York Academy of Sciences 864:485-488. 38 Samuelsson, E.; Jonasson, P.; Viklund, F.; Nilsson, B.;Uhlen, M . (1996). \"Affinity-assisted in vivo folding of a secreted human peptide hormone in Escherichia coli\". Nature Biotechnology 14(6):751-755. Samuelsson, E.; Moks, T.; Nilsson, B.;Uhlen, M . (1994). \"Enhanced in-Vitro Refolding of Insulin-Like Growth-Factor-I Using a Solubilizing Fusion Partner\". Biochemistry 33(14):4207-4211. Sassenfeld, H. M . (1990). \"Engineering Proteins for Purification\". Trends in Biotechnology 8(4):88-93. Sassenfeld, H. M . , Brewer, S. J. (1984). \"A polypeptide fusion designed for purification of recombinant proteins\". Bio/Technology 2:76-81. Schauer-Vukasinovic, V.;Daunert, S. (1999). \"Purification of recombinant proteins based on the interaction between a phenothiazine-derivatized column and a calmodulin fusion tail\". Biotechnology Progress 15(3):513-516. Schmidt, T. G., Koepke, J., Frank, R., Skerra, A . (1996). \"Molecular interaction between the Strep-tag affinity peptide and its cognate target, streptavidin\". Journal of Molecular Biology 255(5):753-766. Schmidt, T. G. M.;Skerra, A . (1993). \"The Random Peptide Library-Assisted Engineering of a C-Terminal Affinity Peptide, Useful for the Detection and Purification of a Functional Ig Fv Fragment\". Protein Engineering 6(1):109-122. Seidler, A . (1994). \"Introduction of a Histidine Tail at the N-Terminus of a Secretory Protein Expressed in Escherichia-Coli\". Protein Engineering 7(10): 1277-1280. Sharrocks, A . D. (1994). \"A T7 Expression Vector for Producing N-Terminal and C-Terminal Fusion Proteins with Glutathione-S-Transferase\". Gene 138(1-2): 105-108. Shih, Y . P.; Wu, H. C.; Hu, S. M . ; Wang, T. F.;Wang, A . H. J. (2005). \"Self-cleavage of fusion protein in vivo using TEV protease to yield native protein\". Protein Science 14(4):936-941. Shimabuku, A . M . ; Saeki, T.; Ueda, K.;Komano, T. (1991). \"Production of a Site Specifically Cleavable P-Glycoprotein-Beta-Galactosidase Fusion Protein\". Agricultural and Biological Chemistry 55(4): 1075-1080. Shoseyov, 0.;Doi, R. H. (1990). \"Essential 170-Kda Subunit for Degradation of Crystalline Cellulose by Clostridium-Cellulovorans Cellulase\". Proceedings of the National Academy of Sciences of the United States of America 87(6) :2192-2195. 39 Shpigel, E., Goldlust, A. , Efroni, G., Avraham, A. , Eshel, A . , Dekel, M . , Shoseyov, O. (1999). \"Immobilization of recombinant heparinase I fused to cellulose-binding domain\". Biotechnology and Bioengineering 65(1): 17-23. Shpigel, E., Goldlust, A. , Eshel, A. , Ber, I. K., Efroni, G., Singer, Y . , Levy, I., Dekel, M . , Shoseyov, O. (2000). \"Expression, purification and applications of staphylococcal protein A fused to cellulose-binding domain\". Biotechnology and Applied Biochemistry 31 ( Pt 3): 197-203. Skerra, A. ; Pfitzinger, I.;Plueckthun, A . (1991). \"The functional expression of antibody Fv fragments in Escherichia coli: improved vectors and a generally applicable purification technique\". Bio/Technology 9(3):273-278. Smith, D. B.;Johnson, K. S. (1988). \"Single-step purification of polypeptides expressed in Escherichia coli as fusions with glutathione S-transferase\". Gene 67(1):31-40. Smith, J. C , Derbyshire, R. B., Cook, E., Dunthorne, L. , Viney, J., Brewer, S. J., Sassenfeld, H . M . , Bell, L. D. (1984). \"Chemical synthesis and cloning of a poly(arginine)-coding gene fragment designed to aid polypeptide purification\". Gene 32(3):321-327. Stahl, S.; Nygren, P. A.;Uhlen, M . (1997a). \"Detection and isolation of recombinant proteins based on binding affinity of reporter: protein A \" . Methods in Molecular Biology 63:103-118. Stahl, S.; Nygren, P. A.;Uhlen, M . (1997b). \"Strategies for gene fusions\". Methods in Molecular Biology 62:37-54. Stevenson, R. (1997). \"Affinity interactions: it's more than affinity chromatography\". American Biotechnology Laboratory 15(10):24, 26, 28. Stofko-Hahn, R. E.; Carr, D. W.;Scott, J. D. (1992). \"A single step purification for recombinant proteins. Characterization of a microtubule associated protein (MAP 2) fragment which associates with the type II cAMP-dependent protein kinase\". FEBS Letters 302(3):274-278. Terpe, K. (2003). \"Overview of tag protein fusions: from molecular and biochemical fundamentals to commercial systems\". Applied Microbiology and Biotechnology 60(5):523-533. Thatcher, D. R. (1990). \"Recovery of Therapeutic Proteins from Inclusion-Bodies -Problems and Process Strategies\". Biochemical Society Transactions 18(2):234-235. Thiele, C.;Fahrenholz, F. (1994). \"Photocleavable Biotinylated Ligands for Affinity-Chromatography\". Analytical Biochemistry 218(2):330-337. 40 Thomson, A . R. (1996). \"Product modification to assist isolation (facilitated processing)\". Downstream Processing of Natural Products: 105-121. Tomme, P.; Boraston, A . ; McLean, B.; Kormos, J.; Creagh, A . L. ; Sturch, K. ; Gilkes, N . R.; Haynes, C. A. ; Warren, R. A . J.;Kilburn, D. G. (1998). \"Characterization and affinity applications of cellulose-binding domains\". Journal of Chromatography B 715(l):283-296. Tomme, P.; Driver, D. P.; Amandoron, E. A . ; Miller, R. C ; Antony, R.; Warren, J.;Kilburn, D. G. (1995a). \"Comparison of a Fungal (Family-I) and Bacterial (Family-Ii) Cellulose-Binding Domain\". Journal of Bacteriology 177(15):4356-4363. Tomme, P.; Gilkes, N . R.; Miller, R. C , Jr.; Warren, R. A . J.;Kilburn, D. G. (1994). \"An internal cellulose-binding domain mediates adsorption of an engineered bifunctional xylanase/cellulase\". Protein Engineering 7(1): 117-123. Tomme, P.; Warren, R. A . J.;Gilkes, N . R. (1995b). \"Cellulose hydrolysis by bacteria and fungi\". Advances in Microbial Physiology 37:1-81. Tripet, B.; Wagschal, K. ; Lavigne, P.; Mant, C. T.;Hodges, R. S. (2000). \"Effects of Side-chain Characteristics on Stability and Oligomerization State of a de Novo-designed Model Coiled-coil: 20 Amino Acid Substitutions in Position \\"d\\"\". Journal of Molecular Biology 300(2):377-402. Tripet, B.; Yu, L. ; Bautista, D. L. ; Wong, W. Y . ; Irvin, R., T.;Hodges, R. S. (1996). \"Engineering a de novo-designed coiled-coil heterodimerization domain off the rapid detection, purification and characterization of recombinantly expressed peptides and proteins\". Protein Engineering 9(11):1029-1042. Tucker, J.;Grisshammer, R. (1996). \"Purification of a rat neurotensin receptor expressed in Escherichia coli\". Biochemical Journal 317:891-899. Uhlen, M.;Moks, T. (1990). \"Gene fusions for purpose of expression: an introduction\". Methods in Enzymology 185(Gene Expression Technology): 129-143. Uhlen, M . ; Nilsson, B.; Guss, B.; Lindberg, M . ; Gatenbeck, S.;Philipson, L. (1983). \"Gene fusion vector based on the gene for staphylococcal protein A\" . Gene 23(3):369-378. Vaillancourt, P.; Simcox, T. G.;Zheng, C. F. (1997). \"Recovery of polypeptides cleaved from purified calmodulin-binding peptide fusion proteins\". Biotechniques 22(3):451-453. 41 Vaillancourt, P.; Zheng, C. F.; Hoang, D. Q.;Breister, L. (2000). \"Affinity purification of recombinant proteins fused to calmodulin or to calmodulin-binding peptides\". Methods in Enzymology 326:340-362. Voss, S.;Skerra, A . (1997). \"Mutagenesis of a flexible loop in streptavidin leads to higher affinity for the Strep-tag II peptide and improved performance in recombinant protein purification\". Protein Engineering 10(8):975-982. Wang, L. F.; Yu , M . ; White, J. R.;Eaton, B. T. (1996). \"BTag: A novel six-residue epitope tag for surveillance and purification of recombinant proteins\". Gene 169(l):53-58. Warren, R. A . J. (1996). \"Microbial hydrolysis of polysaccharides\". Annual Review of Microbiology 50:183-212. Wassenberg, D.; Schurig, H. ; Liebl, W.;Jaenicke, R. (1997). \"Xylanase XynA from the hyperthermophilic bacterium Thermotoga maritima: Structure and stability of the recombinant enzyme and its isolated cellulose-binding domain\". Protein Science 6(8):1718-1726. Wells, P. A. ; Beiderman, B.; Garlick, R. L. ; Lyle, S. B.; Martin, J. P., Jr.; Herberg, J. T.; Meyer, H. F.; Henderson, S. L.;Eckenrode, F. M . (1993). \"Large-scale immunoaffinity purification of recombinant soluble human antigen CD4 from Escherichia coli cells\". Biotechnology and Applied Biochemistry 18 ( Pt 3):341-357. Wilson, D. S.; Keefe, A . D.;Szostak, J. W. (2001). \"The use of mRNA display to select high-affinity protein-binding peptides\". Proceedings of the National Academy of Sciences of the United States of America 98(7):3750-3755. Winterhalter, C ; Heinrich, P.; Candussio, A. ; Wich, G.;Liebl, W. (1995). \"Identification of a novel cellulose-binding domain within the multidomain 120 kDa xylanase XynA of the hyperthermophilic bacterium Thermotoga maritima\". Molecular Microbiology 15(3):431-444. Witzgall, R.; Oleary, E.;Bonventre, J. V . (1994). \"A Mammalian Expression Vector for the Expression of Gal4 Fusion Proteins with an Epitope Tag and Histidine Tail\". Analytical Biochemistry 223(2):291-298. Wood, D. W.; Wu, W.; Belfort, G.; Derbyshire, V.;Belfort, M . (1999). \"A genetic system yields self-cleaving inteins for bioseparations\". Nature Biotechnology 17(9):889-892. Wyborski, D. L . ; Bauer, J. C ; Zheng, C. F.; Felts, K.;Vaillancourt, P. (1999). \"An Escherichia coli expression vector that allows recovery of proteins with native N -42 termini from purified calmodulin-binding peptide fusions\". Protein Expression and Purification 16(1): 1-10. Yang, Z.;Hancock, W. S. (2004). \"Approach to the comprehensive analysis of glycoproteins isolated from human serum using a multi-lectin affinity column\". Journal of Chromatography, A 1053(l-2):79-88. Yarmush, M . L. ; Antonsen, K. P.; Sundaram, S.;Yarmush, D. M . (1992). \"Immunoadsorption - Strategies for Antigen Elution and Production of Reusable Adsorbents\". Biotechnology Progress 8(3): 168-178. Yarnall, M.;Boyle, M . D. P. (1986). \"Isolation and partial characterization of a type II Fc receptor from a group A Streptococcus\". Molecular and Cellular Biochemistry 70(l):57-66. Yet, S. F.; Moon, Y . K.;Sul, H . S. (1995). \"Purification and Reconstitution of Murine Mitochondrial Glycerol-3-Phosphate Acyltransferase - Functional Expression in Baculovirus-Infected Insect Cells\". Biochemistry 34(22):7303-7310. Zheng, C. F.; Simcox, T.; Xu , L.;Vaillancourt, P. (1997). \"A new expression vector for high level protein production, one step purification and direct isotopic labeling of calmodulin-binding peptide fusion proteins\". Gene 186(l):55-60. 43 2 Inexpensive One-Step Purification of Polypeptides Expressed in E. Coli as Fusions With the Family 9 Carbohydrate-Binding Module of Xylanase 10A from T. Maritima 2.1 Introduction The continued maturation of the pharmaceutical and biotechnology industries has created an increasing need for practical and economical large-scale processing techniques. Production methods such as fed-batch fermentations of recombinant microbes have advanced to a level where target biomolecules can be produced in g/L concentrations at relatively modest cost. As a result, downstream processing often accounts for more than 60% of the total operating cost, and as much as 70% of the capital cost of current biochemical production processes (Ladisch 2001). Purification of a target protein during manufacturing usually requires several chromatographic steps in series due to the relatively non-specific physico-chemical interactions that drive separations in these columns. Although product purities are often quite high, overall yields from multi-step chromatographic procedures are generally low due to the accumulated loss of product (Chase 1984; Hearn and Acosta 2001). The challenge therefore is to reduce costs and increase overall yields by process simplification through elimination or combination of purification steps. * A version of this chapter is published in the Journal of Chromatography B. [Reference: Mojgan Kavoosi, Julia Meijer, Emily Kwan, A . Louise Creagh, Douglas G. Kilburn, Charles A . Haynes, Inexpensive one-step purification of polypeptides expressed in E. coli as fusions with the family 9 carbohydrate-binding module of xylanase 10A from T. maritime. J. Chromatography B. 807(1); 87-94; (2004)] 44 Toward this goal, a number of affinity separation systems have been developed in the past two decades to replace difficult multi-step chromatographic procedures with a highly selective binding step that serves to both purify and concentrate the product (Lowe et al. 2001; Wilchek and Chaiken 2000). Polypeptide fusion tags that selectively bind a complementary ligand immobilized onto a suitable chromatographic matrix are now widely used at the laboratory scale to facilitate recombinant expression and purification of target proteins (Nilsson et al. 1997; Terpe 2003). In addition to allowing rapid purification, affinity fusion tags have been shown in certain cases to increase in vivo proteolytic stability of the target protein, improve product solubility, and control product localization in or secretion from the expression host (Jonasson et al. 2002). Skillful engineering of the fusion-tag/immobilized-ligand pair can therefore provide a robust and generic method for efficient production and high-resolution affinity purification of recombinant protein targets. Commercially available affinity tag systems include the calmodulin binding peptide (Vaillancourt et al. 1997; Zheng et al. 1997), the glutathione S-transferase (GST) tag from Schistosoma japonicum (Guan and Dixon 1991; Smith and Johnson 1988), and various polyamino-acid affinity tags such as the polyhistidine tag (Crowe et al. 1994; Hochuli et al. 1987; Porath et al. 1975) and its associated immobilized metal-ion affinity chromatography (IMAC) capture column. Each of these affinity tag systems has been used extensively at the laboratory scale, but only the GST and polyhistidine tag systems have found any appreciable use in manufacturing. More extensive use of affinity tags systems by industry has been thwarted in part by the complex chemical modifications required to cross-link the solid support or to graft the affinity-tag receptor to the resin surface, and by the relatively low tolerance of many of these affinity resins to repeated processing and sanitization-in-place cycles. However, the dominant impediment to the use of current commercially available generic affinity tags in large-scale bioprocessing is cost. A new affinity tag that binds to an inexpensive, chemically and hydrodynamically robust chromatographic resin is therefore highly desirable. Here, we present a generic and inexpensive affinity purification technology based on high-titer recombinant expression in E. coli of fusion proteins containing the 45 carbohydrate-binding module CBM9 attached to the N-terminus of the target protein or polypeptide. TmXynlOACBM9-2 (henceforth referred to as CBM9), the C-terminal family 9 carbohydrate-binding module of xylanase 10A from Thermotoga maritime (Winterhalter et al. 1995), binds specifically to the reducing ends of cellulose and soluble polysaccharides, a property that is currently unique to this C B M . Measured association constants (K a) for adsorption of CBM9 to insoluble allomorphs of cellulose are between 2 x 105 to 2 x 106 M \" 1 . CBM9 also binds a range of soluble sugars (Boraston et al. 2001), including glucose, such that a 1-M glucose solution is effective in quantitatively eluting CBM9 and CBM9-tagged fusion proteins from a cellulose-based capture column. The presence of the CBM9 tag therefore allows for affinity capture and purification of a fusion protein on an inexpensive cellulose-based chromatography resin. A unique processing site is encoded at the C-terminus of the tag to facilitate rapid arid quantitative removal of the tag by Factor X a to recover the desired target protein sequence following affinity purification (Nagai and Thogersen 1984). Validation of the technology is provided by fusing the CBM9 affinity tag to the N-terminus of green fluorescent protein (GFP) from the jellyfish, Aquorin victoria (Crameri et al. 1996; Shimomura et al. 1962). The use of GFP as the target protein has the advantage that the natural fluorescence of GFP measured at 509 nm (excitation at 395 nm) offers a direct and convenient means of tracking the target fusion protein throughout the production and affinity purification process. The generic CBM9 affinity-tag technology proposed here involves five distinct processing steps: (i) recombinant production (cytoplasmic) of the properly folded fusion protein in recombinant E. coli BL21 (DE3) cells, (ii) cell lysis and lysate resuspension, (iii) affinity purification (including elution) of the CBM9-tagged fusion protein on a suitable commercial cellulose-based chromatography resin, (iv) cleavage of the CBM9-linker-IEGR affinity tag sequence using immobilized recombinant Factor X a , and finally, (v) removal of CBM9 to obtain the purified target (GFP). Each of these essential processing steps is evaluated in terms of product yield, purity and concentration factor to provide a measure of the overall performance of the technology. 46 2.2 Materials and Methods 2.2.1 Reagents Kanamycin, glucose, and all other chemicals were purchased from Sigma Chemicals (St. Louis, MO, USA). All reagents were analytical grade unless stated otherwise. Restriction enzymes were purchased from New England Biolabs (Beverly, MA). T4-DNA ligase was obtained from Roche Molecular Biochemicals (Laval, Quebec). Perloza\u00E2\u0084\u00A2 MT100 and MT500 chromatography resins having a nominal particle diameter distribution of 50-80 pm and 100-250 pm, respectively, were purchased 4-9 from Iontosorb Inc. (Czech Republic). E. coli BL21 (DE3) and Ni -Sepharose resin were obtained from Novagen (Milwaukee, MI). 2.2.2 Cloning of CBM9-GFP Fusion Protein All cloning procedures were performed using standard molecular biology techniques (Sambrook et al. 1989). The GFP and CBM9 coding regions were amplified from the vectors pGFPuv (Clontech, Palo Alto, CA) and pETCBM9, respectively. A Bsp HI restriction endonuclease site (underlined) was introduced at the 5' end of CBM9 gene fragment, using the oligonucleotide 5' - T T G C T A G C T T C A T G A C T A G C G G A A T A A T G GTAGC-3' as primer. The sequence encoding for the S3N10LA linker (italic) (henceforth referred to as S3N10 linker) and a Pvu I site (underlined) were introduced at the 3' end of the CBM9 coding region using the oligonucleotide 5'-TCCCTCGATCGCG/J GGTTGTTGTTA TTGTTA TTGTTGTTGTTGT TCGAGCTCGAAA GCTTGATGAGCCTGAGGTTACC-3 ' as primer. For the GFP gene fragment, the sequence encoding for the Factor X a recognition site (IEGR) (italic) and a Pvu I restriction endonuclease site (underlined) were placed at the 5' end, using the oligonucleotide 5 - C C G ^ T C G A G G G T C G T A T C A T G A G T A A A G G A G A - 3 ' as primer. For the 3' end, a Not I site (underlined) was introduced using the oligonucleotide 5'-T G C G G C C G C T T T G T A G A G C T C A T C C A T G C C A T G T G T A A T C C C - 3 ' as primer. Each PCR mixture (50 pL total volume) contained 50 ng of template, 30 pmol of each primer, 5% DMSO, 0.4 mM 2'-deoxy-nucleoside 5'-triphosphates, and 1 unit of PWO DNA 47 polymerase in buffer (Roche Molecular Biochemicals, Laval, Quebec). The following protocol for 25 successive PCR cycles was followed: denaturation at 94\u00C2\u00B0C for 30 s, annealing for 2 min by linearly increasing the temperature from 55\u00C2\u00B0C to 72\u00C2\u00B0C, and primer extension at 72\u00C2\u00B0C for 45 s. The resulting C B M 9 - S 3 N i 0 and FXa-GFP coding regions were digested with Bsp HI/Pvu I and Pvu I/Not I, respectively, and ligated (16\u00C2\u00B0C, 16 h) into the pET28b vector (Novagen) previously digested with Nco I and Not I to give the appropriate pET28-CBM9-S 3 N 1 0 -IEGR-GFP construct (hereafter referred to as pET28-CBM9-GFP). D N A sequencing was then completed to verify the construct (NAPS Unit, Biotechnology Laboratory, University of British Columbia). 2.2.3 Protein Production Overnight cultures of E. coli strain BL21/pET28-CBM9-GFP were diluted 100-fold in tryptone-yeast extract-phosphate medium (TYP) supplemented with 50 ug/mL of kanamycin and grown at 37\u00C2\u00B0C to a cell density (OD600 nm) of ~ 1.0 absorbance units. Isopropyl-l-thio-P-D-galactoside (IPTG) was added to a final concentration of 0.3 mM. Incubation was then continued at 30\u00C2\u00B0C for a further 10-12 h. The cells were harvested by centrifugation (8500g) at 4\u00C2\u00B0C for 20 min and then resuspended in high salt buffer (1 M NaCl, 50 mM potassium phosphate, pH 7.0) by gentle mixing. Cells were ruptured by two passages through a French pressure cell (21000 lb/in2) and the cell debris removed by centrifugation for 30 min at 27000 g and 4\u00C2\u00B0C. CBM9-GFP fusion protein was purified by affinity chromatography as described below. The stability of the fusion protein against proteolysis was assayed as follows. At 18 hours post-induction, the culture was divided into two equal volumes and cells were harvested as described above. The cells in one container were resuspended in high salt buffer while the cells in the second container were resuspended in high salt buffer containing 1 m M phenylmethylsulfonyl fluoride (PMSF) (Sigma). Cells were disrupted (in the presence of fresh 1 m M PMSF or buffer) and the clarified cell extract (1.5 mL) incubated with 200 uL of Perloza\u00E2\u0084\u00A2 MT100 (94.5 mg dry weight/mL). The mixture was mixed at room temperature for 3.5 hours by rotating end-over-end. The resin was collected by centrifugation at 8500g for 8 min, washed three times with 1 mL high salt 48 buffer, 2X with low salt buffer (50 m M potassium phosphate pH 7.0) and I X with TBS8 (15 mM NaCl, lOmM TrisHCl, pH 8.0). The protein was desorbed with 400 uL of 1-M glucose in TBS8 and analyzed by SDS-PAGE. 2.2.4 Affinity Chromatography A Pharmacia XK-16 column (8.5 cm X 1.6 cm I.D.) was packed by standard inclined pouring with Perloza\u00E2\u0084\u00A2 MT100 resin to give a final bed volume of ca. 17 mL. A l l purification chromatograms were completed on a Pharmacia P-500 FPLC system (Amersham Biosciences) at 4\u00C2\u00B0C and with a flow rate of 0.5 ml/min. The Perloza\u00E2\u0084\u00A2 MT100 affinity column was equilibrated with ~ 10 column volumes (CV) of high salt buffer. Clarified cell extract (50 mL) was loaded onto the column and unbound protein was removed by washing the column with 10 C V of high salt buffer, 5 C V of low salt buffer and 4 C V of TBS8 buffer. CBM9-GFP was then eluted from the column with 5 CV of 1-M glucose in TBS8. Eluted protein fractions were analyzed for purity by 12% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) using 20% SDS sample buffer. Column regeneration was completed using 10 C V water followed by 10 CV of high salt buffer for equilibration. 2.2.5 Tag Cleavage by Factor X a In this experiment carried out at 21\u00C2\u00B0C and pH 7, a purified chimeric protein comprised of Factor X a fused to CBM2a, the family 2a carbohydrate binding module (CBM2a) of xylanase 10A of the soil bacterium Cellulomonas fimi, was immobilized onto a Perloza\u00E2\u0084\u00A2 MT500 column (henceforth called CBM2a-FXa i m ) (Kwan et al. 2002) and used to enzymatically remove the CBM9 affinity tag. Pure CBM9-GFP fractions were pooled and incubated with CBM2a-FXaj m at 21\u00C2\u00B0C, rotating end-over-end. After 108 hours, CBM2a-FXaj m was removed by centrifugation (8500g, 15 min) and washed extensively to collect all cleaved product. The cleaved products were buffer exchanged into low salt buffer and concentrated in a stirred ultrafiltration (UF) unit (Amicon, Beverly, M A ) on a 1 KDa cutoff filter (Filtron, Northborough, M A ) . The concentrated protein solution was applied to a column (24 cm X 0.9cm I.D.) packed with Perloza\u00E2\u0084\u00A2 49 MT100 and washed with 15 C V of low salt buffer. Free GFP was collected in the flow through and concentrated by UF. The processing time required for Factor X a cleavage of the IEGR-terminal affinity tag was also assayed. 3 mg of purified CBM9-GFP fusion protein was incubated with 3 pL CBM2a-FXa i m (Kwan et al. 2002) at 21\u00C2\u00B0C (final [Factor X a ] to [fusion protein] ratio of 1:1000). The control experiments contained buffer in place of CBM2a-FXai m . Samples were taken at the following time points (0, 0.5, 1, 2, 4, 6, 8, 10, 12, 14, 25.5, and 35 hours post-incubation) and analyzed by SDS-PAGE. 2.2.6 Fluorescence Calibration Curves Fluorescence measurements were used to quantify target protein (CBM9-GFP) concentrations in buffer and in complex cell lysate feed stocks. Highly pure CBM9-GFP was obtained by sequentially purifying CBM9-GFP on the Perloza\u00E2\u0084\u00A2 MT100 affinity +2 column followed by immobilized-metal affinity chromatography (IMAC) on a N i -Sepharose I M A C resin (according to manufacturer's instructions). This highly pure protein was buffer exchanged into low salt buffer and concentrated as described above. Concentrations of the purified protein were determined by U V absorbance (280 nm) using a calculated molar extinction coefficient of 62870 cm' M \" 1 (Machetal. 1992). Varying concentrations of highly pure protein were mixed with either loading buffer (TBS8), elution buffer (1 M glucose in TBS8) or BL21 cell extract (A 2 g 0 = 6 or A280 = 0.6) and the fluorescence measured (395 nm for excitation; 509 nm for emission) using a Cary Eclipse fluorescence spectrophotometer (Varian, Palo Alto, CA). Linear calibration curves (0 to 0.35 p M CBM9-GFP) for measured fluorescence as a function of CBM9-GFP concentration in buffer and in BL21 cell extract were constructed from each data set. 2.2.7 Measurement of Binding Isotherms Samples containing purified fusion protein at concentrations ranging from 1 to 30 p M were incubated with resin (1 mg (dry weight) of Perloza\u00E2\u0084\u00A2 MT100, 5 mg CF31, 5 mg 50 CF1 and 5 mg Avicel) in high salt buffer to a final volume of 1 ml. The samples were then incubated for 30 hours at 4\u00C2\u00B0C (25\u00C2\u00B0C for Avicel samples) while mixing end-over-end. The cellulose was removed by centrifugation at 27000g for 16 min at 4\u00C2\u00B0C. The supernatant was collected and the concentration of unbound protein was determined by U V absorbance (280 nm) using a Cary 100 Spectrophotometer (Varian). An isotherm was generated by plotting the concentration of bound protein (umol/g of resin) against the concentration of unbound protein (uM). The binding parameters were determined by a non-linear fitting of the Langmuir-type adsorption isotherm equation to the experimental data using GraphPad Prism 3.0 software. Binding isotherms were also measured for CBM9-GFP in the presence of bacterial cell extract (A2so of 5.3). Samples were incubated for 16 hours at 4\u00C2\u00B0C while mixing end-over-end. The cellulose was removed by centrifugation as described above and the fluorescence of the supernatant was measured. The concentration of unbound protein in the supernatant was determined from the calibration curve. Bound CBM9-GFP concentrations were then computed by mass balance. 2.3 Results and Discussion Figure 2.1 shows a block diagram for the C B M 9 - G F P fusion protein construct used in this work to validate the utility and performance of our C B M 9 tag technology for inexpensive affinity purification of recombinant proteins and peptides in E. coli. In the pET28-CBM9-GFP vector, the coding sequence for the N-terminal C B M 9 is followed by the gene fragment encoding an S3N10 linker that serves to separate the C B M 9 fusion tag from the target protein. The synthetic S3N10 linker was used in this study because it has proven useful in our laboratory in the stable expression of a range of fusion proteins. The combined C B M 9 - S 3 N 1 0 fusion tag is separated from the n-terminal amino acid of GFP by the four amino acid IEGR processing site for the endoprotease recombinant human Factor X a . The presence of the IEGR processing site allows Factor X a catalyzed removal of the affinity tag following fusion protein purification to recover the pure target protein with its natural N-terminus. 51 2.3.1 Binding Isotherms and Thermodynamics Equilibrium adsorption isotherms at 4\u00C2\u00B0C for CBM9-GFP binding to the porous cellulose-based chromatography resin Perloza\u00E2\u0084\u00A2 MT100 are shown in Figure 2.2. As shown in Table 2.1, Perloza\u00E2\u0084\u00A2 MT100 stationary phase resin binds pure CBM9-GFP with a capacity of 9.93(\u00C2\u00B10.31) pmol/g MT100. For the 53 kg mol\"1 CBM9-GFP fusion protein, this equates to a saturation loading capacity of 527 mg protein bound per gram of resin, or ca. 115 mg/mL of column. In pure buffer at 4\u00C2\u00B0C, CBM9-GFP binds Perloza\u00E2\u0084\u00A2 MT100 with an affinity of 1.5(\u00C2\u00B10.21) x 106 M \" 1 . As shown in Table 2.2, CBM9-GFP also binds to a number of other commercially available cellulose-based resins. However, in each case, the resin capacity (and to a lesser extent the binding affinity) is significantly lower than observed for binding to Perloza\u00E2\u0084\u00A2 MT100, indicating that the Perloza\u00E2\u0084\u00A2 resin offers a relatively high concentration of entropically unhindered reducing ends for CBM9 binding. CBM9 binding to cellulose is exothermic (Boraston et al. 2001). Thus, the affinity characterizing binding to Avicel will increase with decreasing temperature. However, q i m a x for this resin is appreciably lower than for Perloza\u00E2\u0084\u00A2 MT100, making it a less desirable matrix for affinity chromatography applications. The equilibrium adsorption isotherm at 4\u00C2\u00B0C for binding of pure CBM9 to Perloza\u00E2\u0084\u00A2 MT100 is also shown (Figure 2.2, Table 2.1). The binding properties (K a and qi m a x ) of the isolated fusion tag (CBM9) are similar to those of the fusion protein (CBM9-GFP), indicating that the presence of the target protein does not significantly affect the performance of the CBM9 affinity tag. The intrinsic fluorescence of GFP allowed us to also measure the binding isotherm for the CBM9-GFP fusion protein in the presence of the E. coli cell lysate from which it is purified. Although errors in CBM9-GFP fluorescence measurements are large when cell lysate components are present in the solution phase, the data suggest that neither the binding capacity of the resin nor the affinity of the CBM9-GFP fusion protein for the resin is significantly altered by the presence of a large concentration of 52 contaminant proteins (data not shown), indicating the specificity of the Perloza\u00E2\u0084\u00A2 MT100 cellulose-based resin for CBM-tagged proteins. 2.3.2 Fusion Protein Expression and Stability Unoptimized batch fermentation yields of soluble CBM9-GFP in recombinant E. coli BL21 cultures were typically around 210 mg/L of culture, which represents a 40% increase in GFP yield over more standard expression systems (Chalfie and Kain 1998; Chalfie et al. 1994). The tendency for C B M fusion tags, including the more commonly used maltose binding protein, to increase soluble expression of otherwise low expressing proteins is well documented (Fox et al. 2001; Fox and Waugh 2003). This ability to enhance titers of soluble protein is likely due, at least in part, to the relatively high solubility of CBMs, which allows them to serve as effective solubilizing agents for aggregation-prone polypetides. In certain cases, fusion to a C B M can also promote the proper folding of the attached protein into its biologically active conformation. This chaperone-like quality distinguishes CBMs such as CBM9 and M B P from other affinity tags and greatly enhances their value as a fusion partner. The performance of a fusion tag technology depends not only on the properties of the tag, but also on the stability of the amino-acid sequence that links the tag to the target protein. Spiking and incubation of purified CBM9 in an E. coli BL21 culture lysate resulted in no detectable degradation of the C B M during its purification as measured by SDS PAGE. The stability of the S3N10IEGR linker against degradation by endogenous E. coli proteases present in the cytoplasm and cell lysate was therefore analyzed by SDS PAGE following cell disruption and lysate clarification, either in the presence or absence of the protease inhibitor PMSF. As the C B M is not degraded significantly by endogenous proteases, proteolytic degradation within the linker results in the appearance of a band on an SDS P A G E gel corresponding to (or close to) the molecular mass of CBM9. As shown in Figure 2.3, when PMSF is added to the washed cell suspension, a very small amount of proteolytic degradation of the S3N10 linker occurs, either in vivo or during the cell processing and affinity purification steps. In the absence of a protease inhibitor, a slightly larger fraction of the recombinantly expressed CBM9-GFP fusion 53 protein is lost due to degradation within the linker region. Under both processing conditions, however, the vast majority of expressed fusion protein remains intact through the induction, cell lysis and affinity chromatography steps. 2.3.3 Affinity Purification on Perloza\u00E2\u0084\u00A2 MT100 Column A typical chromatogram for affinity purification on a Perloza\u00E2\u0084\u00A2 MT100 capture column of CBM9-GFP from an E. coli BL21 clarified cell lysate is shown in Figure 2.4. No protease inhibitor (PMSF) was added to the cell suspension or lysate. The corresponding SDS P A G E gel documentation of the purification process is shown in Figure 2.5, and a summary of the fusion protein yield, purity, and concentration factor following elution from the Perloza\u00E2\u0084\u00A2 MT100 column is provided in Table 2.3. The intrinsic fluorescence of GFP allows us to monitor simultaneously the elution of contaminating proteins (UV absorbance @ 280 nm) and the concentration (fluorescence intensity @ 509 nm) of CBM9-GFP and its degradation products in each elution fraction. A small amount of CBM9-GFP or GFP within the clarified lysate load is lost in the column flowthrough. It is likely that most if not all of this fluorescent material represents the small amount of fusion protein that is degraded within the S3N10 linker region, as shown in Figure 2.3. Weakly bound contaminating proteins are sequentially removed in the column flowthrough and the two column wash steps. No loss of CBM9-GFP is detected in either wash step (Figure 2.5, lanes 3 and 4). A 1-M glucose solution (in TBS8) is effective in quantitatively eluting all specifically bound fusion protein (Figure 2.5, lane 5). CBM9-GFP elutes from the column in a single sharp peak, as is evident from the overlapping A280 and fluorescence intensity peaks in the chromatogram. The purity of CBM9-GFP in the pooled fractions of the elution peak was greater than 95% at an average yield of 86%. Both values are competitive with (in fact superior to) the published performance of other commercially available affinity tag systems, including the GST and poly-His fusion-tag technologies (Appa Rao et al. 1997; Baek et al. 1997; Chang et al. 1999; Littlejohn et al. 2000; Modesti etal. 1995). 54 In these experiments, the Perloza\u00E2\u0084\u00A2 MT100 column was loaded to less than half saturation capacity to guarantee capture of all CBM9-GFP from the clarified cell lysate. Despite operating the column at under-loaded conditions, a remarkably high concentration factor of ca. 46 was achieved, indicating that the fusion protein loads, binds and elutes from the column in a reasonably tight band. 2.3.4 Column Reusability As the cost of any affinity chromatography technology is largely determined by the purchase price and reusability of the capture resin, we investigated the ability of the Perloza\u00E2\u0084\u00A2 MT100 resin to provide acceptable and predictable purification performance with repeated column use. Six consecutive purifications were performed on a single Perloza\u00E2\u0084\u00A2 MT100 column to identify any changes in column performance with increasing number of purification cycles. Very high product purity (>95%) was achieved in all six purification cycles. As shown in Table 2.4, product yield and concentration factor, however, were affected by repeated column use. An average yield of 86(\u00C2\u00B13.6)% was observed for the affinity purification of CBM9-GFP from clarified cell lysate on a clean, freshly poured Perloza\u00E2\u0084\u00A2 MT100 column. Slightly lower yields of ca. 79% were then consistently observed for each purification cycle thereafter. The product concentration factor followed the same trend, with a measured concentration factor of 46(\u00C2\u00B19.9) for the first column cycle falling to a consistent value of ca. 28 for each subsequent cycle. The source of these modest changes is unclear. However, the repeatable good performance (> 95% purity, 79% yield, concentration factor of 28) of the column following the first column cycle suggests that our C B M 9 fusion tag technology can provide a robust platform for affinity purification of recombinant proteins. 2.3.5 Removal of the CBM9-S 3N 1 0-IEGR Affinity Tag Using an Immobilized Factor X a Column In certain cases, such as in the production of a human therapeutic protein, removal of the fusion tag following purification is required to recover the desired target protein with its natural N-terminus. We therefore have incorporated a Factor X a processing site adjacent to the N-terminus of the target protein to facilitate tag removal by specific 55 enzymatic cleavage. Figure 2.6 is an SDS PAGE gel showing the kinetics of tag cleavage when purified CBM9-GFP is processed at 21 \u00C2\u00B0C and pH 8 with CBM2a-FXa i m at a fusion-protein to Factor X a concentration ratio of 1000 to 1. To avoid stagnant settling of the CBM2a-FXai m Perloza\u00E2\u0084\u00A2 MT500 resin, the reaction mixture, which also contained 1 - M glucose in the liquid phase, was mixed end-over-end in an orbital mixer. In the presence of 1-M glucose, CBM9 does not bind to Perloza\u00E2\u0084\u00A2 MT500, while binding of CBM2a-FXa is irreversible at these conditions. Complete cleavage of the CBM9-S3N10-IEGR fusion tag was observed after 28 h. The Factor X a treated solution was then diafiltered on a 1 K cut-off filter to remove the 1-M glucose and loaded onto a second Perloza\u00E2\u0084\u00A2 MT100 column to capture the cleaved CBM9 tag. Pure, N-terminally correct GFP was collected in the flow through with a yield of 98% and a purity of greater than 95%. This resulted in an overall yield of the purified target protein (GFP) of 84% when a fresh Perloza\u00E2\u0084\u00A2 MT100 column was used, or 77% when the same column was used for multiple purification cycles. 2.4 Conclusions We have shown that proteins expressed in E. coli as fusions with the family 9 carbohydrate-binding module of xylanase 10A from T. maritima can be affinity purified on a cellulose-based Perloza\u00E2\u0084\u00A2 MT100 column. The performance of our technology is competitive with all commercial fusion tag systems, and may offer advantages with respect to improving the expression of the target protein in a soluble form. Acceptance and use of affinity tag systems in manufacturing of recombinant proteins have been slowed, at least in part, by the associated costs of the technology, particularly the cost of the resin. Perloza\u00E2\u0084\u00A2 MT100 is a simple, highly porous regenerated cellulose/cellulose xanthate of uniform particle size and flow characteristics. The polymer bead structure is stabilized by hydrogen bonds only; there are no covalent cross-links within the resin. As a result, it is a durable and surprisingly inexpensive resin. When bought in bulk quantities, the cost of Perloza\u00E2\u0084\u00A2 MT100 is ca. $35 U.S. per liter of resin, which, for example, is close to 1/100 the cost of an equivalent volume of 56 glutathione affinity resin used to purify GST tagged proteins. The cost of Perloza\u00E2\u0084\u00A2 MT100 resin also compares very favorably with the costs of those resins designed to capture fusion proteins tagged with GST or calmodulin binding protein. Direct capture on a packed column of Perloza\u00E2\u0084\u00A2 MT100, which binds CBM9-tagged proteins with extraordinarily high capacity (in excess of 500 mg/g resin), therefore appears to offer a robust and inexpensive strategy for affinity purification of proteins expressed in soluble form as fusions with the CBM9 tag. 57 2.5 Tables Table 2.1 Langmuir adsorption parameters (equilibrium association constant K a and binding capacity qi m a x ) for binding of CBM9 and CBM9-GFP to Perloza\u00E2\u0084\u00A2 MT100 at 4\u00C2\u00B0C. Solvent contains pure protein in high-salt buffer. P r o t e i n K a ( M \" 1 ) . m a x ( u , m o l p r o t e i n / g r e s i n ) CBM9-GFP CBM9 1.5 (\u00C2\u00B10.21)* x 106 1.2 (\u00C2\u00B10.06) x 106 9.93 (\u00C2\u00B10.31) 11.20 (\u00C2\u00B10.15) Reported errors represent 2a (i.e. 95% confidence interval) Table 2.2 Binding affinity and capacity of CBM9-GFP on various cellulosic resins R e s i n K a ( M 1 ) q . m a x ( u m o l p r o t e i n / g r e s i n ) Perloza\u00E2\u0084\u00A2 MT100 3 1.5 (\u00C2\u00B10.21)* x 106 9.93 (\u00C2\u00B10.31) C F l a 2.5 (\u00C2\u00B10.11) x 105 0.12 (\u00C2\u00B10.01) CF31 a 4.5 (\u00C2\u00B10.19) x l O 5 0.30 (\u00C2\u00B10.03) Avicel b 4.1 (\u00C2\u00B10.68) x 105 0.52 (\u00C2\u00B10.02) Reported errors as in Tab e2.1 a: Binding performed at 4\u00C2\u00B0C in high-salt buffer b: Binding performed at 25\u00C2\u00B0C in high-salt buffer 58 Table 2.3 Summary of purification of CBM9-GFP on Perloza\u00E2\u0084\u00A2 MT100 at 4\u00C2\u00B0C Protein1 (mg) Yield Purity 2 Concentration factor Cell extract 71.5 100% Elution 61.5 86 (\u00C2\u00B13.6)% > 95% 45.7 (\u00C2\u00B19.9) Free GFP (after tag removal) 60.2 84% > 95% 1 Protein concentration was quantified by fluorescence (Xex = 395 nm, A, e m = 510 nm) 2 Purity determined by SDS-PAGE Table 2.4 CBM9-GFP yield and purity for consecutive purification runs through the same Perloza\u00E2\u0084\u00A2 MT100 column Column cycle1 Yield 2 Concentration factor 1 86 (\u00C2\u00B11.7)% 45.7 2 78.4 (\u00C2\u00B11.6)% 28.8 3 81.2 (\u00C2\u00B11.6)% 29.2 4 79 (\u00C2\u00B11.6)% 28.8 5 78.8 (\u00C2\u00B11.6)% 26.8 6 78 (\u00C2\u00B11.6)% 27.5 1 A l l runs gave a CBM9-GFP purity of > 95% as determined by SDS-PAGE 2 Protein concentrations were quantified by fluorescence as noted in Table 2.3 59 2.6 Figures C B M 9 S 3 N I 0 F X a G F P His-6 Figure 2.1 Schematic representation of gene fragment coding for the CBM9-S3N10-IEGR-GFP fusion protein C, (tiM) Figure 2.2 Equilibrium adsorption isotherms for binding of CBM9 and CBM9-GFP to Perloza\u00E2\u0084\u00A2 MT100 at 4\u00C2\u00B0C. CBM9-GFP binding to Perloza\u00E2\u0084\u00A2 MT100 at 4\u00C2\u00B0C in high salt buffer (solid circle). CBM9 binding at 4\u00C2\u00B0C in high-salt buffer (open circle), where qi is the bound protein concentration and Cj is the equilibrium concentration of protein free in solution. 60 + PMSF M -PMSF M KDa \u00E2\u0080\u0094 20 Figure 2.3 Proteolytic stability of the S3N10 linker in CBM9-GFP. 12% SDS-PAGE of CBM9-GFP purified with Perloza\u00E2\u0084\u00A2 MT100 in a small batch system. PMSF treated cell extract containing CBM9-GFP was mixed end-over-end, washed with buffer and desorbed with 1-M glucose in TBS8. Despite its molecular weight of 22 kDa, CBM9 runs as a 26-31 kDa protein depending on the length of the linker fragment attached to it (Wassenberg et al. 1997) 61 70000 - , E 60000 c - , 3.0 200 300 Volume (mL) Figure 2.4 Chromatogram of CBM9-GFP purification from an E. coli BL21 clarified cell lysate on Perloza\u00E2\u0084\u00A2 MT100 at 4\u00C2\u00B0C. 50 mL of clarified cell extract was loaded at 0.2 mL min\"1 on a 17 mL column packed with Perloza\u00E2\u0084\u00A2 MT100 resin, and then washed with 10 column volumes (CV) high salt buffer and 5 C V low salt buffer. Bound fusion protein was desorbed with 1 M glucose in TBS8. 10 mL fractions were collected and analyze by fluorescence (509 nm) (solid circle) and absorbance at 280 nm (line). 62 Figure 2.5 SDS-PAGE documentation of the affinity purification of CBM9-GFP. 12% SDS-PAGE of CBM9-GFP purified on a 17 mL Perloza\u00E2\u0084\u00A2 MT100 column. A l l samples dissolved in sample buffer containing 10% SDS. Lane M : molecular mass markers in kg mol*1. Lane 1: clarified cell extract prior to column loading. Lane 2: column flow through. Lane 3: high salt wash. Lane 4: low salt wash. Lane 5: pure CBM9-GFP eluted in TBS8 containing 1-M glucose. Lane 6: purified GFP after affinity-tag removal by immobilized Factor X a . 63 Time (hours) kDa 0 0.5 1 2 4 6 8 10 12 14 25.5 35 97. 66-45-30-20-* mm \u00E2\u0080\u00A2 * - \u00E2\u0080\u00A2 Figure 2.6 Time course of CBM9-GFP cleavage by Factor X a at 23\u00C2\u00B0C as shown on a 12% SDS-PAGE. A fusion protein to Factor X a concentration ratio of 1000:1 was used. 64 2.7 References Appa Rao, K. B.; Garg, L. C ; Panda, A . K.;Totey, S. M . (1997). \"High-level expression of ovine growth hormone in Escherichia coli: single-step purification and characterization\". Protein Expression and Purification 11(2):201-208. Baek, M . C ; Choi, K. H. ; Oh, T. G.; Kim, D. H.;Choi, E. C. (1997). \"Overexpression of arylsulfate sulfotransferase as fusion protein with glutathione S-transferase\". Protein Expression and Purification 11(3):257-262. Boraston, A . B.; Creagh, A . L.; Alam, M . M . ; Kormos, J. M . ; Tomme, P.; Haynes, C. A. ; Warren, R. A.;Kilburn, D. G. (2001). \"Binding specificity and thermodynamics of a family 9 carbohydrate-binding module from Thermotoga maritima xylanase 10A\". Biochemistry 40(21):6240-6247. Chalfie, M.;Kain, S. (1998). Green fluorescent protein : properties, applications, and protocols. New York: Wiley-Liss. p 385 Chalfie, M . ; Tu, Y . ; Euskirchen, G.; Ward, W. W.;Prasher, D. C. (1994). \"Green fluorescent protein as a marker for gene expression\". Science 263(5148):802-805. Chang, M . ; Bolton, J. L.;Blond, S. Y . (1999). \"Expression and purification of hexahistidine-tagged human glutathione S-transferase Pl-1 in Escherichia coli\". Protein Expression and Purification 17(3):443-448. Chase, H. A . (1984). \"Affinity separations utilising immobilised monoclonal antibodies-A new tool for the biochemical engineer\". Chemical Engineering Science 39(7-8):1099-1125. Crameri, A . ; Whitehorn, E. A . ; Tate, E.;Stemmer, W. P. (1996). \"Improved green fluorescent protein by molecular evolution using D N A shuffling\". Nature Biotechnology 14(3):315-319. Crowe, J.; Dobeli, H . ; Gentz, R.; Hochuli, E.; Stuber, D.;Henco, K . (1994). \"6xHis-Ni-N T A chromatography as a superior technique in recombinant protein expression/purification\". Methods in Molecular Biology 31:371-387. Fox, J. D.; Kapust, R. B.;Waugh, D. S. (2001). \"Single amino acid substitutions on the surface of Escherichia coli maltose-binding protein can have a profound impact on the solubility of fusion proteins\". Protein Science 10(3):622-630. Fox, J. D.;Waugh, D. S. (2003). \"Maltose-binding protein as a solubility enhancer\". Methods in Molecular Biology 205:99-117. 65 Guan, K. L.;Dixon, J. E. (1991). \"Eukaryotic proteins expressed in Escherichia coli: an improved thrombin cleavage and purification procedure of fusion proteins with glutathione S-transferase\". Analytical Biochemistry 192(2):262-267. Hearn, M . T.;Acosta, D. (2001). \"Applications of novel affinity cassette methods: use of peptide fusion handles for the purification of recombinant proteins\". Journal of Molecular Recognition 14(6):323-369. Hochuli, E.; Dobeli, H.;Schacher, A . (1987). \"New metal chelate adsorbent selective for proteins and peptides containing neighbouring histidine residues\". Journal of Chromatography 411:177-184. Jonasson, P.; Liljeqvist, S.; Nygren, P. A.;Stahl, S. (2002). \"Genetic design for facilitated production and recovery of recombinant proteins in Escherichia coli\". Biotechnology and Applied Biochemistry 35(Pt 2):91-105. Kwan, E.; Guarna, M . M . ; Boraston, A . B.; Gilkes, N . R.; Haynes, C. A. ; Kilburn, D. G.;Warren, R. A . (2002). \"Self-activating factor X derivative fused to the C-terminus of a cellulose-binding module: Production and properties\". Biotechnology and Bioengineering 79(7):724-732. Ladisch, M . R. (2001). Bioseparations engineering : principles, practice, and economics. New York: Wiley. Littlejohn, T. K . ; Takikawa, O.; Skylas, D.; Jamie, J. F.; Walker, M . J.;Truscott, R. J. (2000). \"Expression and purification of recombinant human indoleamine 2, 3-dioxygenase\". Protein Expression and Purification 19(l):22-29. Lowe, C. R.; Lowe, A . R.;Gupta, G. (2001). \"New developments in affinity chromatography with potential application in the production of biopharmaceuticals\". Journal of Biochemical and Biophysical Methods 49(1-3):561-574. Mach, H. ; Middaugh, C. R.;Lewis, R. V . (1992). \"Statistical determination of the average values of the extinction coefficients of tryptophan and tyrosine in native proteins\". Analytical Biochemistry 200(l):74-80. Modesti, A . ; Taddei, N . ; Bucciantini, M . ; Stefani, M . ; Colombini, B. ; Raugei, G.;Ramponi, G. (1995). \"Expression, purification, and characterization of acylphosphatase muscular isoenzyme as fusion protein with glutathione S-transferase\". Protein Expression and Purification 6(6):799-805. Nagai, K.;Thogersen, H . C. (1984). \"Generation of beta-globin by sequence-specific proteolysis of a hybrid protein produced in Escherichia coli\". Nature 309(5971):810-812. 66 Nilsson, J.; Stahl, S.; Lundeberg, J.; Uhlen, M.;Nygren, P.-A. (1997). \"Affinity fusion strategies for detection, purification, and immobilization of recombinant proteins\". Protein Expression and Purification 11(1):1-16. Porath, J.; Carlsson, J.; Olsson, I.;Belfrage, G. (1975). \"Metal chelate affinity chromatography, a new approach to protein fractionation\". Nature 258(5536):598-599. Sambrook, J.; Fritsch, E. F.;Maniatis, T. (1989). Molecular cloning : a laboratory manual. 2nd ed. New York: Cold Spring Harbor Laboratory Press. Shimomura, O.; Johnson, F. H.;Saiga, Y . (1962). \"Extraction, purification and properties of aequorin, a bioluminescent protein from the luminous hydromedusan, Aequorea\". Journal of Cellular and Comparative Physiology 59:223-239. Smith, D. B.;Johnson, K. S. (1988). \"Single-step purification of polypeptides expressed in Escherichia coli as fusions with glutathione S-transferase\". Gene 67(1):31-40. Terpe, K. (2003). \"Overview of tag protein fusions: from molecular and biochemical fundamentals to commercial systems\". Applied Microbiology and Biotechnology 60(5):523-533. Vaillancourt, P.; Simcox, T. G.;Zheng, C. F. (1997). \"Recovery of polypeptides cleaved from purified calmodulin-binding peptide fusion proteins\". Biotechniques 22(3):451-453. Wassenberg, D.; Schurig, H. ; Liebl, W.;Jaenicke, R. (1997). \"Xylanase XynA from the hyperthermophilic bacterium Thermotoga maritima: Structure and stability of the recombinant enzyme and its isolated cellulose-binding domain\". Protein Science 6(8):1718-1726. Wilchek, M.;Chaiken, I. (2000). \"An overview of affinity chromatography\". Methods in Molecular Biology 147:1-6. Winterhalter, C ; Heinrich, P.; Candussio, A. ; Wich, G.;Liebl, W. (1995). \"Identification of a novel cellulose-binding domain within the multidomain 120 kDa xylanase XynA of the hyperthermophilic bacterium Thermotoga maritima\". Molecular Microbiology 15(3):431-444. Zheng, C. F.; Simcox, T.; Xu , L.;Vaillancourt, P. (1997). \"A new expression vector for high level protein production, one step purification and direct isotopic labeling of calmodulin-binding peptide fusion proteins\". Gene 186(l):55-60. 67 3 Strategy for Selecting and Characterizing Linker Peptides for CBM9-Tagged Fusion Proteins Expressed in E. Coli 3.1 Introduction Fusion tag technology provides an efficient and generic strategy for high titre production and subsequent purification of recombinant proteins from culture media or clarified cell extracts. The technology is based on genetically linking a short peptide or small protein domain to either the amino- or carboxy-terminus of the target protein and then using the fusion tag to facilitate efficient capture and purification of the chimeric protein on an affinity matrix. Certain fusion partners such as the maltose binding protein (MBP) (di Guan et al. 1988), other carbohydrate binding modules (CBMs) (Ahn et al. 2004; Murashima et al. 2003), thioredoxin (Trx) (LaVallie et al. 1993), and the E. coli protein N-utilizing substance A (NusA) (Davis et al. 1999), have been shown to improve product solubility and expression by stabilizing the mRNA (Makrides 1996) or by providing a psuedo-chaperone effect (Samuelsson et al. 1994). Fusion tags have also been used to increase in vivo proteolytic stability (Jansson et al. 1990; Martinez et al. 1995; Staahl and Nygren 1997), to serve as effective expression and localization reporters (Gerdes and Kaether 1996; Ozawa 2006), and, when combined with their corresponding leader peptide, to control product localization in or secretion from the expression host (Ford et al. 1991; Johnson et al. 1992; Uhlen and Moks 1990). Skillful selection and positioning of the fusion tag therefore provides a flexible platform for recombinant protein processing. * A version of this chapter has been published in Biotechnology and Bioengineering. [Reference: Mojgan Kavoosi, A . Louise Creagh, Douglas G. Kilburn, Charles A . Haynes, Strategy for selecting and characterizing linker peptides for CBM9-tagged fusion proteins expressed mE. coli. Biotechnology and Bioengineering, 98(3); 599-610; (2007)] 68 The polyhistidine tag (His-tag) (Hochuli et al. 1987; Porath et al. 1975), which permits product capture and purification by immobilized metal affinity chromatography (IMAC), and the Schistosoma japonicum glutathione S-transferase (GST) tag (Smith and Johnson 1988), which has specific affinity for immobilized glutathione, are the two most popular affinity tags. However, other affinity tags offer some important advantages (Glatz et al. 1995). For target proteins that express poorly or exhibit low solubility, an MBP or C B M tag is often used to improve expression while at the same time permitting affinity purification on a complimentary carbohydrate resin (Kapust 1999). A number of reviews (Hearn and Acosta 2001; Terpe 2003) discuss several other useful affinity tags, including Staphylococcus aureus protein A and its synthetic two-domain variant (Moks 1986; Nilsson 1987), and the increasingly popular F L A G peptide (Einhauer and Jungbauer 2001; Hopp 1988). The presence of an affinity tag can compromise the function and intended application of the target protein. This problem is generally overcome by incorporating a specific cleavage site for an endopeptidase that may be used to remove the tag either following elution from the column or while the fusion protein is bound to the affinity matrix (Arnau et al. 2006). Two relatively new commercially available endopeptidases, HRV 3C (Libby et al. 1988) and TEV (Shih et al. 2005) are thought, due to their relatively long recognition sites, to minimize non-specific cleavage. However, Factor X a and enterokinase are still widely used because they allow for recovery of the natural N -terminus of the target. The specific cleavage site is typically introduced as part of a longer linking peptide connecting the fusion tag and the target protein (Carter 1990; Jenny et al. 2003). Poly-glycine and glycine-rich sequences have been by far the most frequently used linkers because they enhance solubility, are thought to be resistant to proteolysis, and presumably confer conformational flexibility that enhances accessibility of the processing enzyme to its recognition site and allows the adjoining domains to function independently (Branden and Tooze 1999). For example, the (G4S)X motif of Huston et al. (Huston et al. 69 1988) is widely used as the linker between single-chain Fv fragments (scFv) (Huston et al. 1993). Proline-rich peptides, particularly Pro-Thr repeats, are also frequently employed as linkers due to their apparent resistance to proteolytic degradation and their frequent presence as peptides connecting domains in multidomain proteins (Beck et al. 1997; Wootton and Drummond 1989). Many microbial cellulases and xylanases connect their carbohydrate binding modules to their catalytic module via linkers rich in proline and hydroxylamines (Beguin and Aubert 1994; Gilkes et al. 1991; Gilkes et al. 1988; Knowles et al. 1987; Tomme et al. 1994; Tomme et al. 1988). These natural linkers have been shown to enhance the functional properties of the associated carbohydrate binding module (Ong 1995) and the overall proteolytic stability (Greenwood et al. 1989) of these glycolytic enzymes. Although comprehensive studies on linker performance are lacking, data drawn from the fusion protein literature suggest that the utility of an affinity tag technology can be compromised by a number of factors sensitive to the chemistry and length of the linker (Brinkmann et al. 1992; Tang et al. 1996), including poor product expression (Furusawa 1976; Sauer 2001; Tang et al. 1996), interference with target protein function (Arvidson et al. 2003; Maeda et al. 1997), susceptibility to proteolysis (Fukuoka et al. 1993; Hampe et al. 2000), low binding efficiency of the affinity tag (Jalaguier et al. 1996; Lucius et al. 1992; Sheffield et al. 1999), and poor enzymatic processing in situations where removal of the affinity tag is required (Davis et al. 1999; Fassina et al. 1994; Guan and Dixon 1991). Here, we use a complimentary set of experiments to evaluate the properties of linkers as part of fusion proteins to the C-terminal family 9 carbohydrate binding module (CBM9) of xylanase 10A from Thermotoga maritima (Figure 3.1) (Boraston et al. 2001; Winterhalter et al. 1995). This affinity tag system is used to characterize the performance of the popular poly-glycine and PT-repeat linker chemistries. It is then used to evaluate an alternative strategy for selecting a linker chemistry based on the application of MEROPS\u00E2\u0084\u00A2 (Rawlings et al. 2006), a comprehensive bioinformafics database of peptidase (protease) specificities, to predict linker sequences that are resistant to 70 proteolysis during fusion protein expression and processing in the host organism, which in this case, is the gram negative bacteria Escherichia coli. Previously (Kavoosi et al. 2004), we introduced a novel affinity separation technology based on CBM9 and showed that it has performance characteristics competitive with other commercial affinity tag systems (Vaillancourt et al. 1997). It utilizes a high-capacity (static binding capacities between 90 and 150 mg mL\"1) cellulosic resin that is ca. l /50 t h the cost of an equivalent volume of N i 2 + - N T A (IMAC) resin used for polyhistidine tag systems, making the CBM9 affinity tag system a cost-competitive and potentially attractive technology for industrial bioprocessing. In this study, the CBM9 affinity tag is fused to the N-terminus of a convenient reporter, the green fluorescent protein (GFP) from the jellyfish Aquorin victoria (Crameri et al. 1996; Shimomura et al. 1962), through a selected linker sequence and the Factor X a (FX a) recognition sequence. Luminescence resonance energy transfer (LRET) with terbium as the donor and GFP as the acceptor is used to determine a characteristic relative distance of separation between the tag and target and its dependence on linker chemistry. Differential scanning calorimetry (DSC) is used to determine the influence of the linker on the thermodynamic stability of the fusion partners. These results are combined with MEROPS\u00E2\u0084\u00A2 predictions and tandem mass-spectrometry (MS) data to interpret the effect of linker chemistry and length on production rates, resistance to proteolysis, independent folding and functioning of the two domains, binding isotherms, and the rate of enzymatic cleavage of the CBM9 tag using Factor X a . 3.2 Materials and Methods 3.2.1 Reagents Isopropyl-l-thio-(3-D-galactoside (IPTG), glucose, kanomycin and all other chemicals were purchased from Sigma (St. Louis, M O , USA). A l l reagents were analytical grade unless stated otherwise. Restriction enzymes were purchased from New England Biolabs (Beverly, M A ) . T4-DNA ligase and PWO D N A polymerase were obtained from Roche Molecular Biochemicals (Laval, Quebec). Perloza\u00E2\u0084\u00A2 MT100 chromatography resin having a nominal particle diameter distribution of 50-80 pm was 71 purchased from Iontosorb Inc. (Czech Republic). N i -Sepharose resin and E. coli BL21 (DE3) cells were obtained from Novagen (Milwaukee, MI). 3.2.2 Cloning of CBM9-Linker-FXa-GFP Fusion Proteins A l l cloning procedures were performed using standard molecular biology techniques (Sambrook and Russell 2001). The GFP and CBM9 coding regions were amplified from pGFPuv (Clontech, Palo Alto, CA) and pETCBM9 (Boraston et al. 2001), respectively, using the priming oligonucleotides listed in Table 3.1 and following the PCR protocol previously described (Kavoosi et al. 2004). The construction of the pET28-CBM9-P-FX a -GFP expression vector for production of the CBM9-P-FX a -GFP fusion protein, where the linker is comprised of a single proline residue P preceding the F X a processing site comprised of the 4 amino acid sequence IEGR, was achieved as follows: primers 1 and 3, encoding for restriction endonuclease sites Bsp HI and Pvu I, respectively, were used to amplify the CBM9-P coding region. The F X a - G F P a coding region was amplified using primers 7 and 8, which encode for restriction sites Pvu I and Not I, respectively. The resulting CBM9-P and F X a - G F P a coding regions were digested with Bsp HI/Pvu I and Pvu I/Not I, respectively, and ligated (16\u00C2\u00B0C, 16 h) into pET28 vector (Novagen) previously digested with Nco I and Not I to give pET28-CBM9-P-FX a -GFP. Other fusion constructs were prepared in a similar manner using the primer sets shown in Table 3.1. D N A sequencing was performed to verify all constructs (NAPS Unit, The Michael Smith Laboratories, The University of British Columbia). 3.2.3 Fusion Protein Production and Purification Fusion protein was produced in overnight cultures of E. coli BL21 (DE3) containing a pET28-CBM9-linker-FX a-GFP expression vector diluted 100-fold in tryptone-yeast extract-phosphate medium (TYP) supplemented with 50 ug mL\"1 of kanamycin. To allow for meaningful comparison of productivities, the cells were grown at 37\u00C2\u00B0C to a cell density ( O D 6 0 0 nm) of 1.0 \u00C2\u00B1 0.1, induced with IPTG to a final concentration of 0.3 m M , and incubated for a further 16 \u00C2\u00B1 0.2 h at 30\u00C2\u00B0C. The cells were harvested by centrifugation (8,500 x g) at 4\u00C2\u00B0C for 20 min, resuspended in high salt buffer (1 M NaCl, 50 m M potassium phosphate, pH 7.0), and ruptured by two passages through 72 a French pressure cell (21000 lb in\"2). The cell debris was then removed by centrifugation (27,000 x g) at 4\u00C2\u00B0C for 30 min. The final CBM9-linker-FX a-GFP fusion protein concentration in the cell extract was determined by SDS-PAGE analysis following purification of the fusion by affinity chromatography as described previously (Kavoosi et al. 2004). Cell growth and fusion protein production were also measured in real time by continuously monitoring OD620nm and fluorescence emission at 510 nm (400 nm excitation wavelength) in a SpectraFluorPlus spectrophotometer (Tecan, USA). A l l experiments were done in replicates of five. 3.2.4 Binding Isotherm Measurement and Analysis Standard solutions of purified CBM9-linker-FX a-GFP fusion protein in high salt buffer, ranging in protein concentration from 1 to 30 pM, were mixed with 1 mg (dry weight) of Perloza\u00E2\u0084\u00A2 MT100 in high salt buffer to reach a final volume of 1 ml. The samples were incubated at 4\u00C2\u00B0C for 30 hours (time course experiments show equilibrium reached within 3-5 hours) while rotating end over end. The loaded resin was separated by centrifugation at 27,000 x g for 16 min at 4\u00C2\u00B0C. The supernatant was collected and the equilibrium concentration of unbound protein cx (pM) was determined by U V absorbance at 280 nm using a Cary 100 Spectrophotometer (Varian). The concentration of bound fusion protein q, (pmol g\"1 of Perloza\u00E2\u0084\u00A2 MT100) was computed by mass balance. The equilibrium binding constant Ka (M\"1) and saturation capacity (<7;max) were then determined by non-linear regression of the Langmuir adsorption isotherm equation to the experimental isotherm data using GraphPad Prism 3.0 software. 3.2.5 Linker Stability Analysis To allow clear visualization of differences in rates of linker hydrolysis, clarified cell lysate was incubated for different periods of time at room temperature in the absence of any protease inhibitors, and then analyzed as follows. Following the incubation period, clarified cell lysate (1.5 ml) was mixed with 200 pL of Perloza\u00E2\u0084\u00A2 MT100 (94.5 mg dry weight mL\"1) for 3.5 h at room temperature while rotating end-over-end. The resin was collected by centrifugation at 8,500 x g for 8 min, washed three times with high salt buffer, 2X with low salt buffer (50 m M potassium phosphate, pH 7.0) and I X with 73 TBS8 (15 m M NaCl, 10 m M Tris-HCl, pH 8.0). The fusion protein was then eluted and collected in 400 uL of 3 M glucose in TBS8 and analyzed by SDS-PAGE to quantify the relative amounts of intact and CBM9 byproducts generated by proteolytic cleavage within the linker region of the fusion. Intact CBM9-GFP and fusion protein degradation byproducts were sequence analyzed by LC/MS/MS in the Genome BC Proteomics Centre at the University of Victoria (Victoria, British Columbia, Canada) to verify protein sequence and to identify cleavage positions within the linker region. Briefly, 1 to 10 ug of affinity-purified protein was isolated by SDS-PAGE, excised and extracted into trypsin cleavage buffer, digested overnight with trypsin, dissolved in 0.2% acetic acid, and transferred to autosampler vials for LC/MS/MS analysis. Tandem MS was performed using an L C Q D E C A IT mass spectrometer (Thermo Finnigan, San Jose, CA) with an in-house fabricated microelectrospray source and an HP 1100 solvent delivery system (Agilent, Palo Alto, CA) . Peptide sequences were determined using SEQUEST\u00E2\u0084\u00A2 (Thermo Finnigan), and Peptide-Prophet\u00E2\u0084\u00A2 (Keller et al. 2002) was used to verify correctness of peptide assignments. 3.2.6 Factor X a Processing Analysis To simplify our hydrolysis analysis, site-specific proteolytic removal of the CBM9 fusion tag was carried out using Factor X a immobilized onto Perloza\u00E2\u0084\u00A2 MT500, a highly porous (porosity of 0.97) cellulosic media that permits faster mass transfer than Perloza\u00E2\u0084\u00A2 MT100 and is therefore better suited as stationary phase substrate for the immobilized enzyme reaction. Immobilization was achieved by producing and purifying Factor X a as a chimeric fusion to CBM2a, a family 2a carbohydrate-binding module of xylanase 10A of Cellulomonas fimi (CBM2a-FXai m , where \"im\" indicates immobilized) (Kwan et al. 2002). Unlike CBM9, CBM2a and thus the CBM2a-FX a fusion bind irreversibly to Perloza\u00E2\u0084\u00A2 media, permitting processing of purified CBM9-linker-FX a-GFP in the solution state by including 1 M glucose in the mobile phase. Purified CBM9-linker-FX a-GFP (3 mg) was mixed with CBM2a-FX a (3 u.g) in a TBS8 solution (final [Factor X a ] to [fusion protein] ratio of 1:1000) containing 1 M glucose and maintained at 74 21\u00C2\u00B0C. Buffer was used in place of CBM2a-FXai m in control experiments. The 300 pL sample was continuously rotated end-over-end with 2 pL samples of the supernatant taken 0, 0.5, 1, 2, 4, 6, 8, 10, 12, 14, 25.5, and 35 h after initial mixing. Each 2 pL sample was diluted 2X in loading buffer and analyzed by SDS-PAGE to determine the percentage of cleaved fusion protein with time, from which was determined the half-life ha (h) of each hydrolysis reaction under identical reaction conditions. 3.2.7 Differential Scanning Calorimetry Studies Melting thermograms for all CBM9-linker-FX a-GFP fusion proteins were measured by differential scanning calorimetry (DSC) using a VP-DSC Extended Range MicroCalorimeter (MicroCal Incorporated, USA). Each 1.5281 mL sample contained 90 p M of purified protein in 50 m M potassium phosphate buffer (pH 7). Samples were dialyzed against pure buffer with at least five volume exchanges to achieve a stable and reproducible thermogram baseline. Degassed samples were syringe-loaded into the DSC and thermally scanned from 5\u00C2\u00B0C to 110\u00C2\u00B0C at a rate of 1\u00C2\u00B0C min\"1. Excess heat capacity data for the protein melting transitions were calculated by subtracting the heat capacity baseline for the solvent obtained by loading buffer in both the sample and reference cells of the calorimeter (Creagh et al. 2005). Melting transitions for the CBM9 and GFP domains of each fusion were analyzed using software provided by MicroCal Inc. to determine the melting temperature (Tm) and the enthalpy of denaturation (MJcai) at Tm. 3.2.8 LRET Studies Hogue et al. (Hogue et al. 1992) have demonstrated that calcium-binding sites on proteins can bind Tb(III) and thereby serve as a lanthanide-based luminescent probe for measuring molecular distances. The inherently low absorbance of Tb(III) in solution is overcome in the bound state by the sensitizing effect of the chelating groups which, in the case of the strong calcium binding site of CBM9, bind Tb(III) in a close to perfect octahedral geometry through coordination with Asp, Glu and Val residues (Notenboom et al. 2001). Time-resolved energy transfer from excited Tb(III) in the metal-ion binding site therefore allows determination of the relative distance between the bound terbiums and an acceptor using Forster theory. Since lanthanide emission does not result from a 75 singlet-to-singlet transition characteristic of fluorescence, the process is termed luminescence resonance energy transfer (LRET). Each purified CBM9-linker-FX a-GFP fusion protein was incubated with Chelex 100X resin, mixing end-over-end at 4\u00C2\u00B0C for 48 h, to strip all bound metal from the calcium binding sites of CBM9. The apo-protein was then passed over a Sephadex G10 size-exclusion column to remove any trace contaminants. Chelex-treated fusion protein was diluted to a concentration of 3 p M in nano-pure water containing excess terbium (15 pM) and the solution was mixed for 15-30 min at 4\u00C2\u00B0C, centrifuged (27,000 x g, 10 min) to remove any precipitates, and used directly in the fluorescence and LRET measurements. Steady-state fluorescence measurements and time-resolved luminescence decay measurements were made on a Cary Eclipse fluorescence spectrophotometer (Varian) equipped with a pulsed excitation source and gating system that takes advantage of the millisecond lifetime of the lanthanide to eliminate any background fluorescence from direct excitation of the acceptor fluorophore (GFP). Tb(III) binding to CBM9 in each CBM9-linker-FX a-GFP fusion was confirmed by monitoring sensitized luminescence intensity at 512 nm during titration of the apo-protein with Tb(III), which indicated stoichiometric binding to the calcium binding sites of CBM9 (Notenboom et al. 2001). Fluorescence lifetime experiments were conducted at an excitation wavelength of 222 nm with collection of the emission intensity /(f) at 512 nm, a wavelength near the peak maximum in the GFP fluorescence emission spectra (Chalfie and Kain 1998; Kain et al. 1995) that does not overlap with any of the emission peaks for Tb(III) (Jiao et al. 2003). Emission intensities were recorded following a 200 ps delay after the excitation flash to ensure that any signal detected at 512 nm is entirely due to resonance energy transfer and not from direct excitation. A l l readings were an average of 1000 cycles. Analysis of /(f) data using Forster theory followed the procedure of Selvin et al. (Selvin 2002; Selvin and Hearst 1994). Measured /(f) curves were fit to an exponential of the form /(f) = 1(0) exp(-t/xda), where id a is the lifetime of the donor in the presence of the acceptor. The fits were always characterized by an r 2 > 0.999, confirming that the data follow a single 76 exponential decay and therefore, that the donor alone makes no significant contribution to the intensity signal recorded at 512 nm. Forster theory relates the distance between donor and acceptor R to the resonance energy transferred E, given by (1 - T d a / x d ) , R = R, fl_] 1 / 6 -1 (3.1) where t d is the lifetime of the donor in the absence of the acceptor, and R0 is the Forster distance, where there is 50% energy transfer (E = 0.5), given by R =0.211 1* (3.2) with J, the spectral overlap, given by J = lfDWsaWdA \fD(X)di (3.3) In Equation 3.2, K2 is an orientation factor having a value of 2/3 in these experiments, Qo is the quantum yield of the donor (Tb(III) bound CBM9) (Root et al. 1999). The refractive index r\ is 1.33 for biological samples in water and, in equation 3.3,/D(A,) is the corrected fluorescence intensity of the donor, and sa is the extinction coefficient of the acceptor. The value of Qu taken from (Root et al. 1999) may not fully incorporate the contribution of the sensitizing chromophore, CBM9. As a result, the value of R0 is likely smaller than its true value. Moreover, as noted previously, C B M 9 is capable of binding three Tb(III) ions at the reaction conditions used. As a result, energy transfer is the sum of E for each bound Tb(III). Nevertheless, since reaction conditions and R0 were held 77 constant in these studies, we do not report absolute R values, but rather values, / P where Rp is the measured separation distance between donor and acceptor in the CBM9-P-IEGR-GFP fusion. 3.3 Results and Discussion Since we first introduced the CBM9 affinity tag (Kavoosi et al. 2004), the technology has been applied to the production and purification of over a dozen fusion proteins harboring the tag at the N-terminus and utilizing the different linkers characterized in this study. Here, we report results obtained for CBM9-linker-FX a-GFP fusion proteins (Figure 3.1), taking advantage of the unique fluorescence of the GFP reporter to quantify the results. These results are consistent with those obtained for other fusion constructs. Table 3.2 lists the linker sequences studied. The lengths of the poly(G) and poly(PT) sequences were varied to assess the impact of linker length on performance. The poly(PT) linker used in this study is derived from the natural 23-amino acid linker found in endoglucanase A (Cen A) of the bacterium Cellulomonas fimi (Shen 1991). 3.3.1 Linker Selection and Screening Using MEROPS\u00E2\u0084\u00A2 Although proteolysis is observed within protein domains, it is known to occur far more frequently within flexible or less compact regions of the folded polypeptide chain (i.e., those with large crystallographic B-factors and poorly defined electron density) (Fontana et al. 1997). It is thought that these flexible regions can more easily adopt conformations compatible with the active site of the attacking endopeptidase (Huber and Bennett 1983; Huyton et al. 2003). Inter-domain linkers are therefore particularly susceptible to proteolysis, a feature frequently exploited by crystallographers to identify and isolate domains of interest (Orengo et al. 1997; Wheelan et al. 2000). A number of bioinformatics tools have recently been developed to identify potential cleavage sites using genome-based assignment of peptidases and their associated putative specificities (Boyd et al. 2005). Among the most comprehensive is MEROPS\u00E2\u0084\u00A2, a protease specificity database developed by Rawlings et al. (Rawlings et al. 2006). The use of 78 MEROPS\u00E2\u0084\u00A2, which includes the complete inventory of known and putative E. coli peptidases (Barrett et al. 1998), therefore establishes a potential strategy for in silico design of an effective linker for fusion tags to be expressed within a particular host organism. For example, the S3N10 region of the linker sequence shown in Table 3.2 is predicted by MEROPS\u00E2\u0084\u00A2 to be completely resistant to E. coli endopeptidases. This MEROPS\u00E2\u0084\u00A2-based screening strategy may be applied to any putative or existing linker to assess its proteolytic stability. For each investigated linker, Table 3.2 reports putative cleavage sites that are predicted by MEROPS\u00E2\u0084\u00A2 for all known and assigned endopeptidases of E. coli, as well as the cleavage sites observed experimentally by tandem mass spectrometry after production and purification of the fusion protein in E. coli. MEROPS\u00E2\u0084\u00A2 identifies a number of potential cleavage sites within each poly(G)-IEGR sequence, predicting the potential for cleavage at both Gly/Gly (through the action of the E. coli peptidases colicin V processing peptidase, cytotoxin SubA, HtpX peptidase, as well as a number of unassigned peptidases from the M23, M48, and S8 families of E. coli peptidases) and Gly/Ile (Lit peptidase, an E. coli subfamily U49 peptidase) junctions. Tandem MS data for processed CBM9-poly(G)-FX a-GFP fusion proteins reveals that cleavage does in fact occur at each of these junction types (Table 3.2). Thus, although widely used for fusion-protein bioprocessing, glycine-rich linker sequences are susceptible to hydrolysis by a number of E. coli peptidases. In contrast, poly(PT) sequences are predicted by MEROPS\u00E2\u0084\u00A2 and confirmed by tandem MS results to be proteolytically stable in E. coli, supporting the dominant rationale for utilizing these \"natural\" sequences as linking agents for chimeric protein expression in bacteria (Gustavsson et al. 2001). The F X a recognition sequence is predicted by MEROPS\u00E2\u0084\u00A2 to be stable against E. coli peptidases. Tandem MS results, however, revealed cleavage at the Gly/Arg junction within the F X a recognition sequence, possibly due to endopeptidase mediated hydrolysis upstream of the F X a cleavage site and subsequent activity by carboxypeptidases to the Gly/Arg junction. As a result, for all linkers studied, a small amount of cleavage within the F X a recognition sequence was observed in the absence of the F X a processing enzyme. The use of H R V 3C or TEV in place of the F X a would not alleviate this problem and in 79 fact might exacerbate it since both H R V 3C and TEV have potential sites of E. coli endopeptidase-catalyzed hydrolysis predicted by MEROPS\u00E2\u0084\u00A2. Table 3.2 also reports sites of hydrolysis within the linker sequence observed by tandem MS but not predicted by MEROPS\u00E2\u0084\u00A2. A l l sites of cleavage within the poly(G) and the poly(PT)x-P linkers correspond to ones predicted by MEROPS\u00E2\u0084\u00A2. For the S3N10 linker, tandem MS data indicate cleavage within the poly(S) region which is not captured by MEROPS\u00E2\u0084\u00A2 predictions. However, the amount of cleavage is extremely small, as the MS Data reveals that the S3N10 linker remains intact for at least 99.2% of the fusion protein population. Therefore, MEROPS\u00E2\u0084\u00A2 appears to be a useful tool in designing and selecting linkers resistant against endogeneous peptidases for expression in a given host. 3.3.2 Impact of the Linker on Fusion-Tag Performance Our ongoing studies on CBM9-tagged fusion proteins indicate that the choice of linker chemistry and length can have a significant impact on fusion-tag performance. While similar specific growth rates [average doubling time of 0.5614 h\"1 (\u00C2\u00B1 0.0256 h\"1) as monitored by optical density at 620 nm] were observed for all constructs, fusion protein production rates and yields showed a dependence on linker chemistry (Figure 3.2). A significant drop in the expression rate is observed when tryptophans are introduced into the linker, with the drop in productivity proportional to the number of aromatic amino acids in the sequence, which might explain why tryptophan and phenylalanine are seldom found in natural linkers (Argos 1990). A weaker dependence on linker composition is observed for sequences devoid of aromatic amino acids or significant hydrophobic amino acid content. In addition to tandem MS analysis, the stability of the linker against long-term exposure to endogeneous E. coli peptidases was examined by SDS-PAGE analysis (Figure 3.3). The results confirm that proteolytic stability is strongly correlated to linker composition and length. As both the CBM9 (melting temperature @ pH 7 of 97.9 \u00C2\u00B0C) and GFP (T m @ pH 7 of 83.7 \u00C2\u00B0C) domains are highly stable, hydrolysis of the fusion protein is largely restricted to the linker region, resulting in a band of hydrolyzed products covering a relatively narrow molecular-weight range lying slightly above the 80 molecular weight at which CBM9 alone runs on the gel (ca. 28 kDa). The results observed by SDS-PAGE analysis (Figure 3.3) are consistent with MEROPS\u00E2\u0084\u00A2 predictions and tandem MS results reported in Table 3.2. The poly(PT)x-P linkers and the synthetic S3N10 linker are significantly more stable than the two poly(G) linkers against hydrolysis by E. coli peptidases. Densitometry data on gels from independent samples indicate that about 36\u00C2\u00B15% of the expressed CBM9-(G)3-IEGR-GFP fusion protein is lost due to hydrolysis within or near the linker region during the fermentation and purification process; these losses are significantly higher (45\u00C2\u00B15%) for the CBM9-(G) 15-IEGR-GFP fusion protein. For the (PT)X-P-IEGR linker series, the size of the CBM9-containing fragment released by proteolysis increases with x, consistent with MS data that show no significant cleavage within the (PT)X-P sequence while detecting cleavage within the F X a processing site. In contrast, the dominant CBM9-containing fragment released by hydrolysis is the same size for the (G)3-IEGR and (G)i5-IEGR constructs, indicating a common and preferred set of hydrolysis sites within both poly(G) sequences. The S3N10 linker was the most proteolytically stable linker studied, with combined losses due to hydrolysis within the linker-region and Factor X a processing site during long-term incubation studies with E. coli proteases constituting less than 7\u00C2\u00B14% of the purified CBM9-S 3 Ni 0 - IEGR-GFP product. Virtually all of that loss was due to hydrolysis at the Gly/Arg junction within the F X a processing site as determined by tandem MS results. Therefore, despite its relatively long length, the S3N10 construct maintained good proteolytic resistance. Interdomain linkers can also influence the structure and activity of the connected domains (Arvidson et al. 2003; Deyrup et al. 1999). Equilibrium binding isotherms were therefore performed to determine the influence of the linker on the performance of CBM9 in purifying the fusion protein. Determined by nonlinear regression of the Langmuir equation (qx = q^axKaci/(l + Kact)) to adsorption isotherm data (Figure 3.4), association constants (Ka) near 106 M - 1 (Table 3.3) were measured for binding of all CBM9-linker-FX a -GFP fusion proteins to Perloza\u00E2\u0084\u00A2 MT100. The Ka values are essentially equivalent 81 to that of free CBM9 (Ka = 1.2 X 106 M\" 1), indicating that the linker and the GFP fusion partner do not interfere with the function of the affinity tag. Similarly, saturation capacities (q\u00E2\u0084\u00A2x) of ca. 10 pmol protein/g Perloza\u00E2\u0084\u00A2 MT100 resin were measured for all fusion proteins (Table 3.3), representing a saturation binding capacity of ca. 530 mg protein/g resin (75 mg protein/mL column), which is significantly higher than almost any commercially available affinity matrix. Half lifes (t\a, representing the time where 50% of the original fusion protein has been cleaved) for F X a processing of CBM9-Linker-FX a -GFP fusion proteins were determined by SDS-PAGE analysis (Figure 3.5) and were found to depend strongly on linker chemistry but not on linker length (Table 3.4). Relatively rapid cleavage is observed for fusion proteins containing either a poly(G) linker or the S3N10 linker. The introduction of a proline residue or, more dramatically, a PT repeat sequence upstream of the F X a processing site was found to dramatically reduce the rate of cleavage, suggesting that F X a function is inhibited by these upstream elements. The proteolytically susceptible poly-glycine linkers had the fastest F X a processing rates, with a t\a of 1.4(\u00C2\u00B14.0) x 104 s observed for both the G 3 and G15 linkers. These results are not surprising since glycine is often placed directly upstream of the N -terminus of a protease recognition sequence to enhance its rate of cleavage (Guan and Dixon 1991). However, for fusion proteins expressed and localized in the cytoplasm of E. coli, the approximately 2-fold enhancement in the rate of F X a processing relative to the S3N10 linker is offset by the significantly higher loss of poly(G)-linked fusion proteins due to linker proteolysis by endogenous E. coli peptidases. 3.3.3 Determination of Characteristic Distances Using Luminescence Resonance Energy Transfer Increasing the length of the (PT)X-P linker had relatively little impact on susceptibility to proteolysis. In contrast, increasing the length of the poly(G) linker from 3 to 15 amino acids resulted in a dramatic increase in linker proteolysis, reflecting both an increase in the number of potential cleavage sites as predicted by MEROPS\u00E2\u0084\u00A2 (Table 3.2), and the possibility of an increase in the accessibility of those sites to E. coli 82 peptidases. To explore this latter issue, we used LRET to measure relative distances of separation between the donor chromophore, Tb(III) bound to the calcium-binding sites of CBM9, and the acceptor chromophore, GFP. Fluorescence lifetime I(t) data, such as that shown in Figure 3.6 for the CBM9-(PT) 2P-IEGR-GFP fusion protein, were collected at 512 nm (222 nm excitation wavelength) and fit to the exponential decay equation to determine T d a , the lifetime of the donor in the presence of the acceptor. Forster theory (equations 3.1 to 3.3) was then used to compute the characteristic distance R between donor and acceptor (Table 3.5). As one would expect, the shortest linker sequence P-IEGR is characterized by the smallest R (32.6 \u00C2\u00B1 1.0 A) value, hereafter called Rp. R/R is therefore greater than unity for all for the poly(G)-IEGR linker, suggesting that this linker adopts a hydrated, extended configuration that makes it accessible to both E. coli peptidases, leading to an enhancement in the proteolytic degradation rate, and F X a , leading to an improvement in site specific tag-removal kinetics. Thus, the chemistry of the linker is seen to strongly influence overall linker performance and stability in a manner that is not fully captured by MEROPS\u00E2\u0084\u00A2. As a result, we find that MEROPS\u00E2\u0084\u00A2 can be used as an initial but not an exhaustive linker selection strategy. . 3.3.4 Influence of the Linker on the Thermodynamic Stability of CBM9-Linker-FXa-GFP Fusion Proteins The thermodynamic stability of the native protein fold is known to be a key determinant in the expression of recombinant proteins in E. coli, with a decrease in the thermodynamic stability generally resulting in a decrease in the half-life of the protein (Mclendon and Radany 1978). For example, expression yields progressively increase and proteolysis rates decrease for increasingly higher Tm variants of either T4 lysozyme (Inoue and Rechsteiner 1994) or the N-terminal domain of X-repressor protein (Parsell other fusion proteins studied. For a given linker increasing linker length. The increase in R with linker length is particularly strong 83 and Sauer 1989). Studies by Kwon et al. (Kwon et al. 1996) on barnase variants indicate that these correlations also hold for recombinant proteins exported to the periplasm of E. coli, where a 4.3 \u00C2\u00B0C increase in the Tm of barnase resulted in a 50% increase in protein yield. As their length affects spatial positioning of the fused domains (Table 3.5), linkers therefore may also influence intracellular rates of proteolysis by modulating the thermodynamic stabilities of the fusion partners. We therefore used the proline-containing set of linkers to explore the connection between linker length and the thermodynamic stabilities of the fused domains. Melting thermograms for CBM9 on its own, for CBM9-P-IEGR-GFP and for CBM9-(PT) 7P-IEGR-GFP (Figure 3.7) show that formation of the fusion does not alter the Tm of GFP, but results in a decrease in the Tm and in AHca\, the calorimetric enthalpy of denaturation, of CBM9 (Table 3.6). Both results indicate that fusing CBM9 to GFP through the linker-IEGR sequence fhermodynamically destabilizes CBM9, with the degree of destabilization increasing with decreasing length of the (PT)XP linker such that ATm = - 8.4 \u00C2\u00B0C when the (PT)oP linker (i.e., the P linker) is used. This result correlates well with intracellular protein yield data reported in Figure 3.2, which show a decrease in fusion protein yield with decreasing length of the (PT)XP linker. 3.4 Conclusions Linker design and its importance to the performance of fusion tag technology has received relatively little attention (Carlsson et al. 1996; Gustavsson et al. 2001; Sauer 2001). Here, we have shown that the level of expression, proteolytic stability, thermodynamic stability, and Factor X a processibility of fusion proteins containing an N -terminal CBM9 tag and expressed in recombinant E. coli can depend on the chemistry and length of the linker. This points to the need for methods for designing or selecting an effective linker that take into account the specific cellular environment in which the fusion protein is produced and localized, and from which it must be purified. For example, glycine-rich linker sequences, which are currently by far the most commonly reported linker chemistry, were found in this study to be quite susceptible to hydrolysis by E. coli peptidases, emphasizing that linkers should be specifically designed to resist 84 hydrolysis catalyzed by the peptidases of the host organism. Our strategy for meeting this important design criterion was to exploit bioinformatics tools (MEROPS\u00E2\u0084\u00A2) for predicting protease specificities to screen putative linker sequences for hydrolytic resistance to endogenous peptidases. This approach builds on earlier efforts to improve linker design by selecting linker sequences derived from naturally occurring linkers (Gustavsson et al. 2001). Both approaches are shown in this work to identify linkers with a considerably higher resistance to endogenous E. coli proteolysis than observed for poly(G)-based linkers. The S3N10 linker selected by MEROPS\u00E2\u0084\u00A2-based screening was slightly more resistant to proteolysis than those designed from naturally occurring linkers (the (PT)X-P linker series). However, in both cases, the majority of degradation occurred within the Factor X a processing site and not within the linker itself. While MEROPS\u00E2\u0084\u00A2-based screening was effective in selecting a stable linker, tandem MS based sequencing of the C-terminal peptide released by tryptic digestion of the CBM9-containing hydrolysis product indicated that certain sites of potential hydrolysis predicted by MEROPS\u00E2\u0084\u00A2 were not cleaved, while one site not identified by MEROPS\u00E2\u0084\u00A2 was cleaved, albeit to a very small degree. Thus, the screening method is not exact, possibly due to a dependence of proteolytic susceptibility on local structure (Parsell and Sauer 1989; Wriggers et al. 2005). Nevertheless, it is shown in this work to provide a novel and potentially useful strategy for linker design. Our results further show that design of an effective linker should not be based on proteolytic stability considerations alone. The thermodynamic stability and spatial positioning of the fused proteins/domains may also influence the overall performance of the fusion tag by altering expression levels and the rate at which the tag can be removed from the target protein following affinity capture and purification. The additional experimental methods (DSC, LRET, SDS-PAGE, etc.) described in this work therefore constitute a more comprehensive approach to linker selection and performance analysis. 85 3.5 Tables Table 3.1 Oligonucleotides used in the construction of CBM9-Linker-IEGR-GFP fusion proteins. Restriction sites are underlined and the Factor X a recognition sequence is in bold. P r i m e r O l i g o n u c l e o t i d e S e q u e n c e ( a l l s e q u e n c e s a r e 5'-3') 1 T T G C T A G C T T C A T G A C T A G C G G A A T A A T G G T A G C 2 A G C G G C C G C A G G C C T A C C C T C G A T C G G A G T C G G A G T C G G C G T C G G -A G T C G G A A G C T T G A T G A G C C T G A G G T T A C C 3 A G C G G C C G C C T C G A T C G G A A G C T T G A T G A G C C T G A G G T T A C C 4 C T C G A T C G G A C C A C C A C C A A G C T T G A T G A G C C T G A G G T T 5 T C C C T C G A T C G C G C C G C C A C C G C C G C C A C C A C C A C C A C C G C C G C C A -C C G C C A C C G C C A A G C T T G A T G A G C C T G A G G T T A C C 6 TCCCTCGATCGCGAGGTTGTTGTTATTGTTATTGTTGTTGTTGTTCG-A G C T C G A A A G C T T G A T G A G C C T G A G G T T A C C 7 C C G A T C G A G G G T C G T A T C A T G A G T A A A G G A G A 8 T G C G G C C G C T T T G T A G A G C T C A T C C A T G C C A T G T G T A A T C C C 9 A A G A A T T C A A G C T T C C G A C G C C G A T C G A G G G T C G T A T G A G T A A A G -G A G A A G A A C T T T T C A C 10 T T C A A G C T T C C G A C T C C G A C T C C G A C G C C G A C T C C G A C C C C G A C T C -C A A C T C C G A T C G A C G G T C G T A T C A T G A G T A A A G G A G A 86 Table 3.2 Linker lengths, sequences, and susceptibility to proteolytic cleavage as predicted by MEROPS\u00E2\u0084\u00A2 and confirmed experimentally by LC/MS/MS L i n k e r L e n g t h ( a a ) C l e a v a g e s i t e s p r e d i c t e d b y M E R O P S \u00E2\u0084\u00A2 ( U ) a n d o b s e r v e d b y t a n d e m M S ( T ) C l e a v a g e S i t e s P r e d i c t e d b y M E R O P S \u00E2\u0084\u00A2 G3-IEGR 3 G-rG-fG-f \u00E2\u0080\u0094 IEGf R G/G (C39; M23; M48; S8); G/I (U49) G 1 5 - IEGR 15 Lr-f Lr-f Kjf KJI Cj-f t j f L i | (jf Lr-f Ixf (J-J. (jf Lr-f Lr-j. \u00E2\u0080\u0094 l iHa^ \u00E2\u0080\u0094 K . G/G (C39; M23; M48; S8); G/I (U49) P-IEGR 1 P - I E G t R (PT) 2P-IEGR 5 P T P T P - I E G t R (PT) 4P-IEGR 9 PTPTPTPTP - I E G t R (PT) 7P-IEGR 15 PTPTPTPTPTPTPTP- IEG t R S3N10-IEGR 15 S t S S N N N N N N N N N N L y A - I E G t R L / A (Family S8) Symbol ( U ) denotes putative cleavage site predicted by MEROPS Symbol ( T ) denotes cleavage site observed by tandem mass spectrometry 87 3.3 Regressed Langmuir isotherm parameters for binding of CBM9-Linker-IEGR-GFP on Perloza\u00E2\u0084\u00A2 MT100 at pH 7 (4\u00C2\u00B0C). Standard deviations (a) computed from triplicate measurements. Linker Ka ( M 1 ) qmax (umol/g resin) P-IEGR 2.0 (\u00C2\u00B10.26) x 106 10.4 (\u00C2\u00B10.18) (PT) 2P-IEGR 2.4 (\u00C2\u00B10.36) x 106 10.4 (\u00C2\u00B10.25) (PT) 4P-IEGR 6.0 (\u00C2\u00B10.99) x 106 9.9 (\u00C2\u00B10.30) (PT) 7P-IEGR 3.1 (\u00C2\u00B10.24) x 106 10.3 (\u00C2\u00B10.14) G3 - IEGR 4.6 (\u00C2\u00B10.71) x 106 9.9 (\u00C2\u00B10.27) G15-IEGR 5.2 (\u00C2\u00B10.82) x 106 9.5 (\u00C2\u00B10.26) S3N10-IEGR 1.5 (\u00C2\u00B10.21) x 106 9.9 (\u00C2\u00B10.31) CBM9 1.2 (\u00C2\u00B10.06) x 106 11.2 (\u00C2\u00B10.15) * reported errors represent \u00C2\u00B12a (i.e. 95% confidence intervals) 88 Table 3.4 Half-lifes for Factor X a cleavage of CBM9-Linker-IEGR-GFP fusion proteins. A l l experiments were conducted in 50 m M potassium phosphate buffer (pH 7, 21 \u00C2\u00B0C) at a fusion protein to Factor X a concentration ratio of [1000] to [1]. Linker tm (h) P-IEGR 8 (\u00C2\u00B11) (PT) 2P-IEGR 25 (PT) 4P-IEGR 35 (PT) 7P-IEGR >35 G3 - IEGR 4 G15-IEGR 4 S 3 N 1 0 - IEGR 6 89 Table 3.5 Average relative distance of separation R/Rp between the bound terbiums on CBM9 and the fluorescent chromophore of GFP determined by LRET at21\u00C2\u00B0C. Linker A P-IEGR 1.00 (\u00C2\u00B10.03) (PT) 2P-IEGR 1.08 (\u00C2\u00B1 0.02) (PT) 4P-IEGR 1.11 (\u00C2\u00B1 0.04) (PT) 7P-IEGR 1.14 (\u00C2\u00B1 0.02) G3 - IEGR 1.09 (\u00C2\u00B10.02) G15-IEGR 1.51 (\u00C2\u00B10.10) S3N10-IEGR 1.04 (\u00C2\u00B1 0.05) * reported errors represent \u00C2\u00B1a (i.e. 60% confidence interval) 90 Table 3.6 Melting temperatures (Tm) and denaturation enthalpies (Aif c ai) for CBM9 on its own and as part of various CBM9-linker-IEGR-GFP fusion proteins. Linker Tm (\u00C2\u00B0C) AHcal (J mol\"1) None (isolated CBM9) 97.9 1.34 (\u00C2\u00B10.06) x 105 P-IEGR 89.5 0.71 (\u00C2\u00B10.05) x 105 (PT) 2P-IEGR 89.8 0.83 (\u00C2\u00B10.05) x 105 (PT) 4P-IEGR 91.6 0.96 (\u00C2\u00B10.04) x 105 (PT) 7P-IEGR 93.4 1.12 (\u00C2\u00B10.05) x 105 * reported errors represent \u00C2\u00B12a (i.e. 95% con idence intervals) 91 3.6 Figures CBM9 Linker FXa GFP His6 Figure 3.1 Schematic representation of the CBM9-Linker-FX a -GFP fusion protein. F X a indicates the recognition site (IEGR) for Factor X a processing. Time (hour) Figure 3.2 Intracellular Expression of CBM9-Linker-FX a -GFP fusion proteins in E. coli. Cells expressing a CBM9-Linker-FX a -GFP fusion protein were grown at 30\u00C2\u00B0C to an OD620 of ~0.9 and protein expression was induced with 0.1 m M IPTG. CBM9-Linker-FX a -GFP expression was continuously monitored at 510 nm with excitation at 400 nm. 92 kDa P M tf\u00C2\u00B0W\u00C2\u00A9\u00C2\u00BB \u00C2\u00BB\u00E2\u0080\u00A2 , ^ ^ H g M \u00C2\u00BB J^gt ^ ^ ^ ^ \u00E2\u0080\u00A2 P l ^ ^ l P f \u00C2\u00BB P \u00C2\u00BB < f \" C B M 9 - L i n !||g|jp: '^ ^^ ^^ \u00E2\u0080\u0094 ^ i ^ ^ ^ ' ^ M M M M M M B -97 66 45 30 <~ mm ~ * U t t M ^ 4 - C B M 9 20 Figure 3.3 Long-term proteolytic stability of the linkers in CBM9-Linker-FX a -GFP fusion proteins. Clarified cell lysates were incubated at 23\u00C2\u00B0C for 4 hours and then the intact and degraded CBM9-Linker-FX a -GFP fusion proteins were purified on Perloza\u00E2\u0084\u00A2 MT100 in the absence of protease inhibitors. Each lane represents proteins eluted off the Perloza\u00E2\u0084\u00A2 MT100 resin with elution buffer containing 3 M glucose. Lane M contains the molecular weight markers. 93 C . (uM) Figure 3.4 Equilibrium adsorption isotherm for binding of CBM9 -G3 - IEGR-GFP to Perloza\u00E2\u0084\u00A2 MT100. Isotherm measured at 4\u00C2\u00B0C in high salt buffer (pH7; 50 m M potassium phosphate, 1 M NaCl). 94 Time (hours) kDa 0 0.5 1 2 4 6 8 10 12 14 25.5 35 Figure 3.5 F X a cleavage kinetics. CBM9-P-IEGR-GFP fusion protein was incubated at 23\u00C2\u00B0C with Factor X a at a fusion protein/FXa concentration ratio of 1000:1. Samples were taken at each time point indicated and analyzed by SDS-PAGE. 95 10GS 0 1 2 3 4 5 T i m e ( m s ) ure 3.6 Sensitized fluorescence emission lifetime data for CBM9-(PT)2P-IEGR-GFP at 23\u00C2\u00B0C. Solid line indicates fit to the first-order exponential decay equation (r2 = 0.999). 96 Figure 3.7 Differential scanning calorimetry thermograms for CBM9, GFP, CBM9-P-IEGR-GFP and CBM9-(PT) 7P-IEGR-GFP in 50 m M potassium phosphate buffer (pH 7). Area under the denaturation peak(s) provides the denaturation enthalpy through the relationship AHcal = J C P (T)dT, where Cp is the heat capacity. 97 3.7 References Ahn, J. O.; Choi, E. S.; Lee, H. W.; Hwang, S. H. ; Kim, C. S.; Jang, H. W.; Haam, S. J.;Jung, J. K . (2004). \"Enhanced secretion of Bacillus stearothermophilus LI lipase in Saccharomyces cerevisiae by translational fusion to cellulose-binding domain\". Applied Microbiology and Biotechnology 64(6):833-839. Argos, P. (1990). \"An Investigation of Oligopeptides Linking Domains in Protein Tertiary Structures and Possible Candidates for General Gene Fusion\". Journal of Molecular Biology 211(4):943-958. Arnau, J.; Lauritzen, C ; Petersen, G. E.;Pedersen, J. (2006). \"Current strategies for the use of affinity tags and tag removal for the purification of recombinant proteins\". Protein Expression and Purification 48(1): 1-13. Arvidson, B.; Seeds, J.; Webb, M . ; Finlay, L.;Barklis, E. (2003). \"Analysis of the retrovirus capsid interdomain linker region\". Virology 308(1):166-177. Barrett, A . J.; Rawlings, N . D.;Woessner, J. F. (1998). Handbook of Proteolytic Enzymes. London: Elsevier. Beck, C ; Marty, R.; Klausli, S.; Hennecke, H.;Gottfert, M . (1997). \"Dissection of the transcription machinery for housekeeping genes of Bradyrhizobium japonicum\". Journal of Bacteriology 179(2):364-369. Beguin, P.;Aubert, J. P. (1994). \"The Biological Degradation of Cellulose\". FEMS Microbiology Reviews 13(l):25-58. Boraston, A. B.; Creagh, A . L. ; Alam, M . M . ; Kormos, J. M . ; Tomme, P.; Haynes, C. A. ; Warren, R. A.;Kilburn, D. G. (2001). \"Binding specificity and thermodynamics of a family 9 carbohydrate-binding module from Thermotoga maritima xylanase 10A\". Biochemistry 40(21):6240-6247. Boyd, S. E.; Pike, R. N . ; Rudy, G. B.; Whisstock, J. C.;De La Banda, M . G. (2005). \"PoPS: A computational tool for modeling and predicting protease specificity\". Journal of Bioinformatics and Computational Biology 3(3):551-585. Branden, C.-I.;Tooze, J. (1999). Introduction to protein structure. New York: Garland Publishing. Brinkmann, U . ; Buchner, J.;Pastan, I. (1992). \"Independent Domain Folding of Pseudomonas Exotoxin and Single-Chain Immunotoxins - Influence of Interdomain Connections\". Proceedings of the National Academy of Sciences of the United States of America 89(7):3075-3079. 98 Carlsson, H. ; Ljung, S.;Bulow, L. (1996). \"Physical and kinetic effects on induction of various linker regions in beta-galactosidase/galactose dehydrogenase fusion enzymes\". Biochimica et Biophysica Acta 1293(1): 154-160. Carter, P. (1990). \"Site-Specific Proteolysis of Fusion Proteins\". ACS Symposium Series 427:181-193. Chalfie, M.;Kain, S. (1998). Green fluorescent protein : properties, applications, and protocols. New York: Wiley-Liss. Crameri, A. ; Whitehorn, E. A. ; Tate, E.;Stemmer, W. P. (1996). \"Improved green fluorescent protein by molecular evolution using D N A shuffling\". Nature Biotechnology 14(3):315-319. Creagh, A . L . ; Tiong, J. W. C ; Tian, M . M . ; Haynes, C. A.;Jefferies, W. A . (2005). \"Calorimetric studies of melanotransferrin (p97) and its interaction with iron\". Journal of Biological Chemistry 280(16):15735-15741. Davis, G. D.; Elisee, C ; Newham, D. M.;Harrison, R. G. (1999). \"New fusion protein systems designed to give soluble expression in Escherichia coli\". Biotechnology and Bioengineering 65(4):382-388. Deyrup, A . T.; Krishnan, S.; Singh, B.;Schwartz, N . B. (1999). \"Activity and stability of recombinant bifunctional rearranged and monofunctional domains of ATP-sulfurylase and adenosine 5'-phosphosulfate kinase\". Journal of Biological Chemistry 274(16):10751-10757. di Guan, C ; L i , P.; Riggs, P. D.;Inouye, H. (1988). \"Vectors that facilitate the expression and purification of foreign peptides in Escherichia coli by fusion to maltose-binding protein\". Gene 67(l):21-30. Einhauer, A.;Jungbauer, A . (2001). \"The F L A G peptide, a versatile fusion tag for the purification of recombinant proteins\". Journal of Biochemical and Biophysical Methods 49(l-3):455-465. Fassina, G.; Merli, S.; Germani, S.; Ciliberto, G.;Cassani, G. (1994). \"High-Yield Expression and Purification of Human Endothelin-1\". Protein Expression and Purification 5(6):559-568. Fontana, A. ; Zambonin, M . ; deLaureto, P. P.; DeFilippis, V. ; Clementi, A.;Scaramella, E. (1997). \"Probing the conformational state of apomyoglobin by limited proteolysis\". Journal of Molecular Biology 266(2):223-230. Ford, C. F.; Suominen, I.;Glatz, C. E. (1991). \"Fusion tails for the recovery and purification of recombinant proteins\". Protein Expression and Purification 2(2-3):95-107. 99 Fukuoka, Y . ; Tachibana, T.;Yasui, A . (1993). \"Expression of Biologically-Active C3a as Fusion Proteins\". Immunology Letters 38(2):153-158. Furusawa, T., Suzuki, M , Smith, J. M . . (1976). \"Rate parameters in heterogeneous catalysis by pulse techniques.\" Catalysis Reviews - Science and Engineering 13(l):43-76. Gerdes, H.-H.;Kaether, C. (1996). \"Green fluorescent protein: applications in cell biology\". FEBS Letters 389(l):44-47. Gilkes, N . R.; Henrissat, B.; Kilburn, D. G.; Miller, R. C.;Warren, R. A . J. (1991). \"Domains in Microbial Beta-1,4-Glycanases - Sequence Conservation, Function, and Enzyme Families\". Microbiological Reviews 55(2):303-315. Gilkes, N . R.; Warren, R. A . J.; Miller, R. C.;Kilburn, D. G. (1988). \"Precise Excision of the Cellulose Binding Domains from 2 Cellulomonas-Fimi Cellulases by a Homologous Protease and the Effect on Catalysis\". Journal of Biological Chemistry 263(21): 10401-10407. Greenwood, J. M . ; Gilkes, N . R.; Kilburn, D. G.; Miller, R. C , Jr.;Warren, R. A . J. (1989). \"Fusion to an endoglucanase allows alkaline phosphatase to bind to cellulose\". FEBS Letters 244(1): 127-131. Guan, K. L.;Dixon, J. E. (1991). \"Eukaryotic proteins expressed in Escherichia coli: an improved thrombin cleavage and purification procedure of fusion proteins with glutathione S-transferase\". Analytical Biochemistry 192(2):262-267. Gustavsson, M . ; Lehtio, J.; Denman, S.; Teeri, T. T.; Hult, K.;Martinelle, M . (2001). \"Stable linker peptides for a cellulose-binding domain-lipase fusion protein expressed in Pichia pastoris\". Protein Engineering 14(9):711-715. Hampe, W.; Voss, R. H . ; Haase, W.; Boege, F.; Michel, H.;Reilander, H. (2000). \"Engineering of a proteolytically stable human beta(2)-adrenergic receptor/maltose-binding protein fusion and production of the chimeric protein in Escherichia coli and baculovirus-infected insect cells\". Journal of Biotechnology 77(2-3):219-234. Hearn, M . T.;Acosta, D. (2001). \"Applications of novel affinity cassette methods: use of peptide fusion handles for the purification of recombinant proteins\". Journal of Molecular Recognition 14(6): 3 23-3 69. Hochuli, E.; Dobeli, H.;Schacher, A . (1987). \"New metal chelate adsorbent selective for proteins and peptides containing neighbouring histidine residues\". Journal of Chromatography 411:177-184. 100 Hogue, C. W. V . ; Macmanus, J. P.; Banville, D.;Szabo, A . G. (1992). \"Comparison of Terbium(Iii) Luminescence Enhancement in Mutants of Ef Hand Calcium-Binding Proteins\". Journal of Biological Chemistry 267(19): 13340-13347. Hopp, T., Pricket, K S , Price, V L , Libby, RT, March, CJ, Ceretti, DP, Urdal, DL, Conlon, PJ. (1988). \"A short polypeptide marker sequence useful for recombinant protein identification and purification\". Bio/Technology 6:1204-1210. Huber, R.;Bennett, W. S., Jr. (1983). \"Functional significance of flexibility in proteins\". Biopolymers 22(1)261-279. Huston, J. S.; Levinson, D.; Mudgetthunter, M . ; Tai, M . S.; Novotny, J.; Margolies, M . N . ; Ridge, R. J.; Bruccoleri, R. E.; Haber, E.; Crea, R.;Oppermann, H. (1988). \"Protein Engineering of Antibody-Binding Sites - Recovery of Specific Activity in an Anti-Digoxin Single-Chain Fv Analog Produced in Escherichia-Coli\". Proceedings of the National Academy of Sciences of the United States of America 85(16):5879-5883. Huston, J. S.; McCartney, J.; Tai, M . S.; Mottola-Hartshorn, C ; Jin, D.; Warren, F.; Keck, P.;Oppermann, H. (1993). \"Medical applications of single-chain antibodies\". International Reviews of Immunology 10(2-3): 195-217. Huyton, T.; Pye, V . E.; Briggs, L . C ; Flynn, T. C ; Beuron, F.; Kondo, H. ; Ma, J.; Zhang, X.;Freemont, P. S. (2003). \"The crystal structure of murine p97/VCP at 3.6 .ANG\". Journal of Structural Biology 144(3):337-348. Inoue, I.;Rechsteiner, M . (1994). \"On the Relationship between the Metabolic and Thermodynamic Stabilities of T4 Lysozymes - Measurements in Escherichia-Coli\". Journal of Biological Chemistry 269(46)29241-29246. Jalaguier, S.; Mesnier, D.; Leger, J. J.;Auzou, G. (1996). \"Putative steroid binding domain of the human mineralocorticoid receptor, expressed in E-coli in the presence of heat shock proteins shows typical native receptor characteristics\". Journal of Steroid Biochemistry and Molecular Biology 57(l-2):43-50. Jansson, B.; Palmcrantz, C ; Uhlen, M.;Nilsson, B. (1990). \"A dual-affinity gene fusion system to express small recombinant proteins in a soluble form: expression and characterization of protein A deletion mutants\". Protein Engineering 3(6):555-561. Jenny, R. J.; Mann, K. G.;Lundblad, R. L. (2003). \"A critical review of the methods for cleavage of fusion proteins with thrombin and factor Xa\". Protein Expression and Purification 31(1): 1-11. 101 Jiao, FL; Liao, F. FL; Tian, S. J.;Jing, X . P. (2003). \"Luminescent properties of Eu3+ and Tb3+ activated Zn3Ta208\". Journal of the Electrochemical Society 150(9):H220-H224. Johnson, K. ; Murphy, C. K.;Beckwith, J. (1992). \"Protein export in Escherichia coli\". Current Opinion in Biotechnology 3(5):481-485. Kain, S. R.; Adams, M . ; Kondepudi, A. ; Yang, T. T.; Ward, W. W.;Kitts, P. (1995). \"Green Fluorescent Protein as a Reporter of Gene-Expression and Protein Localization\". Biotechniques 19(4):650-655. Kapust, R. B., Waugh, D. S. (1999). \"Escherichia coli maltose-binding protein is uncommonly effective at promoting the solubility of polypeptides to which it is fused\". Protein Science 8(8):1668-1674. Kavoosi, M . ; Meijer, J.; Kwan, E.; Creagh, A . L. ; Kilburn, D. G.;Haynes, C. A . (2004). \"Inexpensive one-step purification of polypeptides expressed in Escherichia coli as fusions with the family 9 carbohydrate-binding module of xylanase 1 OA from T-maritima\". Journal of Chromatography B 807(l):87-94. Knowles, J.; Lehtovaara, P.; Penttila, M . ; Teeri, T.; Harkki, A.;Salovuori, I. (1987). \"The Cellulase Genes of Trichoderma\". Antonie Van Leeuwenhoek Journal of Microbiology 53(5):335-341. Kwan, E.; Guarna, M . M . ; Boraston, A . B.; Gilkes, N . R.; Haynes, C. A . ; Kilburn, D. G.;Warren, R. A . (2002). \"Self-activating factor X derivative fused to the C-terminus of a cellulose-binding module: Production and properties\". Biotechnology and Bioengineering 79(7):724-732. Kwon, W. S.; Da Silva, N . A.;Kellis, J. T., Jr. (1996). \"Relationship between thermal stability, degradation rate and expression yield of barnase variants in the periplasm of Escherichia coli\". Protein Engineering 9(12): 1197-1202. LaVallie, E. R.; DiBlasio, E. A . ; Kovacic, S.; Grant, K. L. ; Schendel, P. F.;McCoy, J. M . (1993). \"A thioredoxin gene fusion expression system that circumvents inclusion body formation in the E. coli cytoplasm\". Bio/Technology 11 (2): 187-193. Libby, R. T.; Cosman, D.; Cooney, M . K. ; Merriam, J. E.; March, C. J.;Hopp, T. P. (1988). \"Human Rhinovirus 3C Protease - Cloning and Expression of an Active Form in Escherichia-Coli\". Biochemistry 27(17):6262-6268. Lucius, R.; Kern, A . ; Seeber, F.; Pogonka, T.; Willenbucher, J.; Taylor, H. R.; Pinder, M . ; Ghalib, H. W.; Schulzkey, H.;Soboslay, P. (1992). \"Specific and Sensitive Igg4 Immunodiagnosis of Onchocerciasis with a Recombinant 33kd Onchocerca-Volvulus Protein (Ov33)\". Tropical Medicine and Parasitology 43(3):139-145. 102 Maeda, Y . ; Ueda, H. ; Kazami, J.; Kawano, G.; Suzuki, E.;Nagamune, T. (1997). \"Engineering of functional chimeric protein G - Vargula luciferase\". Analytical Biochemistry 249(2): 147-152. Makrides, S. C. (1996). \"Strategies for achieving high-level expression of genes in Escherichia coli\". Microbiological Reviews 60(3):512-538. Martinez, A. ; Knappskog, P. M . ; Olafsdottir, S.; Doskeland, A . P.; Eiken, H. G.; Svebak, R. M . ; Bozzini, M . ; Apold, J.;Flatmark, T. (1995). \"Expression of recombinant human phenylalanine hydroxylase as fusion protein in Escherichia coli circumvents proteolytic degradation by host cell proteases. Isolation and characterization of the wild-type enzyme\". Biochemical Journal 306(2):589-597. Mclendon, G.;Radany, E. (1978). \"Protein Turnover Thermodynamically Controlled\". Journal of Biological Chemistry 253(18):6335-6337. Moks, T., Abrahmsen, L. , Nilsson, B., Hellman, U . , Sjoquist, J., Uhlen, M . (1986). \"Staphylococcal protein A consists of five IgG-binding domains\". European Journal of Biochemistry 156(3):637-643. Murashima, K. ; Kosugi, A.;Doi, R. H. (2003). \"Solubilization of cellulosomal cellulases by fusion with cellulose-binding domain of noncellulosomal cellulase EngD from Clostridium cellulovorans\". Proteins-Structure Function and Genetics 50(4):620-628. Nilsson, B., Moks, T., Jansson, B., Abrahmsen, L. , Elmblad, A. , Holmgren, E., Henrichson, C , Jones, T. A. , Uhlen, M . (1987). \"A synthetic IgG-binding domain based on staphylococcal protein A\" . Protein Engineering 1(2): 107-113. Notenboom, V . ; Boraston, A . B.; Kilburn, D. G.;Rose, D. R. (2001). \"Crystal structures of the family 9 carbohydrate-binding module from Thermotoga maritima xylanase 10A in native and ligand-bound forms\". Biochemistry 40(21):6248-6256. Ong, E., Alimonti, J. B., Greenwood, J. M . , Miller, R. C , Jr., Warren, R. A. , Kilburn, D. G. (1995). \"Purification of human interleukin-2 using the cellulose-binding domain of a prokaryotic cellulase\". Bioseparation 5(2):95-104. Orengo, C. A. ; Michie, A . D.; Jones, S.; Jones, D. T.; Swindells, M . B.;Thornton, J. M . (1997). \" C A T H - a hierarchic classification of protein domain structures\". Structure 5(8): 1093-1108. Ozawa, T. (2006). \"Designing split reporter proteins for analytical tools\". Analytica Chimica Acta 556(l):58-68. 103 Parsell, D. A.;Sauer, R. T. (1989). \"The Structural Stability of a Protein Is an Important Determinant of Its Proteolytic Susceptibility in Escherichia-Coli\". Journal of Biological Chemistry 264(13):7590-7595. Porath, J.; Carlsson, J.; Olsson, I.jBelfrage, G. (1975). \"Metal chelate affinity chromatography, a new approach to protein fractionation\". Nature 258(5536):598-599. Rawlings, N . D.; Morton, F. R.;Barrett, A . J. (2006). \"MEROPS: the peptidase database\". Nucleic Acids Research 34:D270-D272. Root, D. D.; Shangguan, X . ; Xu , J.;McAllister, M . A . (1999). \"Determination of Fluorescent Probe Orientations on Biomolecules by Conformational Searching: Algorithm Testing and Applications to the Atomic Model of Myosin\". Journal of Structural Biology 127(l):22-34. Sambrook, J.;Russell, D. W. (2001). Molecular cloning : a laboratory manual. New York: Cold Spring Harbor Laboratory Press. Samuelsson, E.; Moks, T.; Nilsson, B.;Uhlen, M . (1994). \"Enhanced in-Vitro Refolding of Insulin-Like Growth-Factor-I Using a Solubilizing Fusion Partner\". Biochemistry 33(14):4207-4211. Sauer, J., Christensen, T., Frandsen, T. P., Mirgorodskaya, E., McGuire, K . A. , Driguez, H. , Roepstorff, P., Sigurskjold, B. W., Svensson, B. (2001). \"Stability and function of interdomain linker variants of glucoamylase 1 from Aspergillus niger\". Biochemistry 40(31):9336-9346. Selvin, P. R. (2002). \"Principles and biophysical applications of lanthanide-based probes\". Annual Review of Biophysics and Biomolecular Structure 31:275-302. Selvin, P. R.;Hearst, J. E. (1994). \"Luminescence energy transfer using a terbium chelate: improvements on fluorescence energy transfer\". Proceedings of the National Academy of Sciences of the United States of America 91(21):10024-10028. Sheffield, P.; Garrard, S.;Derewenda, Z. (1999). \"Overcoming expression and purification problems of RhoGDI using a family of \"parallel\" expression vectors\". Protein Expression and Purification 15(l):34-39. Shen, H., Schmuck, M . , Pilz, I., Gilkes, N . R., Kilburn, D. G., Miller, R. C , Jr., Warren, R. A . (1991). \"Deletion of the linker connecting the catalytic and cellulose-binding domains of endoglucanase A (CenA) of Cellulomonas fimi alters its conformation and catalytic activity\". Journal of Biological Chemistry 266(17):11335-11340. 104 Shih, Y . P.; Wu, H . C ; Hu, S. M . ; Wang, T. F.;Wang, A . H . J. (2005). \"Self-cleavage of fusion protein in vivo using TEV protease to yield native protein\". Protein Science 14(4):936-941. Shimomura, O.; Johnson, F. H.;Saiga, Y . (1962). \"Extraction, purification and properties of aequorin, a bioluminescent protein from the luminous hydromedusan, Aequorea\". Journal of Cellular and Comparative Physiology 59:223-239. Smith, D. B.;Johnson, K. S. (1988). \"Single-step purification of polypeptides expressed in Escherichia coli as fusions with glutathione S-transferase\". Gene 67(1):31-40. Staahl, S.;Nygren, P. A . (1997). \"The use of gene fusions to protein A and protein G in immunology and biotechnology\". Pathologie Biologie 45(l):66-76. Tang, Y . ; Jiang, N . ; Parakh, C.;Hilvert, D. (1996). \"Selection of linkers for a catalytic single-chain antibody using phage display technology\". Journal of Biological Chemistry 271 (26): 15682-15686. Terpe, K. (2003). \"Overview of tag protein fusions: from molecular and biochemical fundamentals to commercial systems\". Applied Microbiology and Biotechnology 60(5):523-533. Tomme, P.; Gilkes, N . R.; Miller, R. C , Jr.;Warren, R. A . J.;Kilburn, D. G. (1994). \"An internal cellulose-binding domain mediates adsorption of an engineered bifunctional xylanase/cellulase\". Protein Engineering 7(1): 117-123. Tomme, P.; Vantilbeurgh, Ff.; Pettersson, G.; Vandamme, J.; Vandekerckhove, J.; Knowles, J.; Teeri, T.;Claeyssens, M . (1988). \"Studies of the Cellulolytic System of Trichoderma-Reesei Qm-9414 - Analysis of Domain Function in 2 Cellobiohydrolases by Limited Proteolysis\". European Journal of Biochemistry 170(3):575-581. Uhlen, M.;Moks, T. (1990). \"Gene fusions for purpose of expression: an introduction\". Methods in Enzymology 185(Gene Expression Technology): 129-143. Vaillancourt, P.; Simcox, T. G.;Zheng, C. F. (1997). \"Recovery of polypeptides cleaved from purified calmodulin-binding peptide fusion proteins\". Biotechniques 22(3):451-453. Wheelan, S. J.; Marchler-Bauer, A.;Bryant, S. H. (2000). \"Domain size distributions can predict domain boundaries\". Bioinformatics 16(7):613-618. Winterhalter, C ; Heinrich, P.; Candussio, A . ; Wich, G.;Liebl, W. (1995). \"Identification of a novel cellulose-binding domain within the multidomain 120 kDa xylanase XynA of the hyperthermophilic bacterium Thermotoga maritima\". Molecular Microbiology 15(3):431-444. 105 Wootton, J. C.;Drummond, M . H. (1989). \"The Q-Linker - a Class of Interdomain Sequences Found in Bacterial Multidomain Regulatory Proteins\". Protein Engineering 2(7):535-543. Wriggers, W.; Chakravarty, S.;Jennings, P. A . (2005). \"Control of protein functional dynamics by peptide linkers\". Biopolymers 80(6):736-746. 106 4 Direct Measurement of the Kinetics of CBM9 Fusion-Tag Bioprocessing Using Luminescence Resonance Energy Transfer (LRET) 4.1 Introduction Affinity-tag technology, where a target protein or peptide is produced as a recombinant fusion to an N - or C-terminal affinity tag possessing a highly specific binding partner that can serve as a stable ligand for affinity chromatography, is a general and potentially powerful approach to streamline production of recombinant therapeutics and other biologies (Lowe et al. 2001). In addition to facilitating a highly selective separation, affinity tags, particularly those comprised of carbohydrate binding modules, can improve product solubility, increase in vivo proteolytic stability, and control product localization in or secretion from the host organism (Hearn and Acosta 2001). Recently we introduced the CBM9 affinity tag (Kavoosi et al. 2004), which is based on the C-terminal family 9 carbohydrate binding module of xylanase 10A from Thermotoga maritima (Winterhalter et al. 1995). CBM9 exhibits strong and specific affinity for both insoluble cellulose and soluble sugars such as glucose (Boraston et al. 2001). Affinity purification of CBM9-tagged recombinant proteins on a commercially available cellulosic media, Perloza\u00E2\u0084\u00A2 MT100, results in highly pure, concentrated products recovered at high yields (Kavoosi et al. 2004). Due to the low cost and high binding capacity of Perloza\u00E2\u0084\u00A2 MT100 media, the CBM9 affinity system provides a considerable economic advantage over other commercially available affinity tag technologies. * A version of this chapter has been prepared for submisstion to Biotechnology and Bioengineering. [Reference: Mojgan Kavoosi, A . Louise Creagh, Douglas G. Kilburn, Charles A . Haynes, Direct measurement of the kinetics of CBM9 fusion-tag bioprocessing using Luminescence Resonance Energy Transfer (LRET). 107 However, the cost competitiveness of a given affinity-tag technology, particularly at preparative scales, is not solely determined by the cost and maximum cycle number of the capture column. It also depends on the cost and efficiency of the post-purification step required to remove the affinity tag and obtain the desired final product with its natural N - and C-termini. This can be an expensive step, with its economics and yield known to depend on the choice of processing enzyme and the local structure of the cleavage site, most notably the peptide sequence (often denoted the linker sequence) flanking the recognition site on the tag side (Girard et al. 2006; Jenny et al. 2003; Lien et al. 2001). For example, thrombin-catalyzed processing of glutathione-S-transferase (GST) tagged tyrosine phosphatase is greatly enhanced by the incorporation of a glycine linker immediately upstream of the recognition site (Guan and Dixon 1991). The lack of studies on and related methods for selecting the optimal processing enzyme and linker sequence for a particular affinity-tagged protein product is therefore surprising, and points to the need for a simple and effective spectroscopic method to screen a candidate library of processing-enzyme/linker-sequence combinations or, in cases where regulatory or IP constraints predetermine the processing enzyme, a library of linker sequences. A pre-validation screening method of this type could guide bioprocess design and speed translation to the manufacturing scale in a manner that increases downstream process throughput while decreasing cost of goods. Several recent studies have reported on the use of fluorescence (or Forster) resonance energy transfer (FRET) to measure enzyme activity, and on the potential of FRET in high-throughput kinetic screens (Boskovic et al. 2004; Chen et al. 2005; Cotrin et al. 2004; Medintz et al. 2006; Nishikata et al. 2006; Thomas et al. 2006). In FRET, distances between 15 and 80 A can be measured through the distance-dependent energy transfer between a donor and an acceptor fluorophore. This energy transfer leads to measurable changes in donor intensity or excited-state lifetime, as well as to spectral changes in the acceptor dye. Placement of the donor and acceptor fluorophores on opposite sides of an enzymatic-cleavage site on a substrate thereby provides a method for real-time monitoring of hydrolysis kinetics. However, while FRET has proven useful, it suffers from well-documented limitations. The maximum distance that can be measured, 108 which is a direct result of the small and relatively inefficient energy transfer between donor and acceptor, is less than optimal for many biological applications (Selvin 1996). Measurement of the sensitized emission of the acceptor dye is difficult due to a low signal-to-background ratio resulting from interfering fluorescence emission from the donor and direct excitation of the acceptor fluorophore (Selvin 1995). Finally, donor fluorophores used in FRET generally exhibit short (nanosecond) multiexponential lifetimes, necessitating the use and careful synchronization of fast electronics and detectors, which can be difficult and expensive (Selvin 1995). Luminescence resonance energy transfer (LRET), which overcomes many of the limitations of FRET (Selvin 2002), is a related spectroscopic method that has not been applied to real-time monitoring of hydrolytic enzyme activity. In LRET, a luminescent lanthanide chelate is used in place of a traditional fluorophore as the donor, while the acceptor remains a standard organic dye. Lanthanide emission arises from high-spin to high-spin transitions, and is therefore not related to the singlet-to-singlet transitions characteristic of fluorescence. Excitation of lanthanide ions produces very narrow, non-polarized and strongly Stokes-shifted emission bands with extremely long luminescence lifetimes (milliseconds) that can serve to excite the acceptor fluorophore over a sustained period. As in FRET, direct excitation of the acceptor fluorophore can occur. However, the lifetime of this fluorescence emission is on the order of nanoseconds, and any acceptor emission after this very short-lived event will only be due to energy transfer received by the acceptor from the long-lived lanthanide donor. The resulting high signal-to-background ratio of the LRET measurement provides very significant advantages over the more common FRET experiment. In particular, sub-femtamolar detection limits are possible (Sammes and Yahioglu 1996), allowing highly accurate kinetic data to be obtained. Although it has not been used to screen for peptidase specificity and activity, LRET has been used in a number of other applications, including the measurement of conformational changes on the nanometer scale (Callaci et al. 1999; Kapanidis et al. 2001; Xiao et al. 1998) during formation of protein-protein (Kolb et al. 1997; Mathis 1995; Root 1997) and protein-DNA complexes (Churchich 1997; Heyduk et al. 1997). It 109 has also been used to measure inter- and intramolecular distances between protein domains (Burmeister-Getz et al. 1998), and to characterize structures of duplex D N A (Kapanidis et al. 2001; Selvin and Hearst 1994). Finally, the lanthanides terbium and europium have been extensively used to explore the mechanism of metal-ion binding both to D N A and to metalloproteins (Chaudhuri et al. 1997; Veenstra et al. 1995). Here, we introduce a LRET-based assay designed to rapidly and accurately measure CBM9 fusion-tag bioprocessing kinetics and their dependence on the choice of linker sequence and processing enzyme. The assay is based on the nonradiative energy transfer between lanthanide-based donors specifically bound to the C B M 9 tag and an acceptor fluorophore presented on the target protein or peptide, and thus, on the opposite side of the processing site. Enzyme-catalyzed cleavage of the fusion tag then terminates the resonance energy transfer, resulting in a change in the fluorescence intensity of the sensitized acceptor that can be continuously monitored to quantify substrate concentration over time. CBM9 is particularly well suited for this assay, as it is a calcium (II) binding metalloprotein (Notenboom et al. 2001) capable of binding the terbium (III) lanthanide ion with high affinity in the absence of C a 2 + . This Tb3 +-binding property can therefore be combined with a number of elegant and efficient methods that have recently been developed to specifically label the C-terminus of a protein with a fluorophore (Becker et al. 2006; Chao et al. 1998; Dursina et al. 2005; Tripet et al. 1996) to provide a general LRET-based assay of the kinetics of C B M 9 fusion-tag bioprocessing. Our aim in this work is to prove that such a LRET-based assay can provide accurate kinetic data sets for fusion-tag bioprocessing, allowing one to identify both major and subtle differences in rates of hydrolysis. To do this, we chose as the fusion partner the green fluorescent protein (GFP) from the jellyfish Aquorin victoria (Shimomura et al. 1962), thereby exploiting the fact that the target protein, GFP, endogenously provides the required fluorescent acceptor group. This permitted the rapid preparation of a library of chimeric proteins (Figure 4.1) comprised of an N-terminal Tb3+-loaded CBM9 (hereafter referred to as Tb 3 + -CBM9) fused to GFP via a cleavage-site terminated linker sequence specific to the popular processing enzyme enterokinase. Our LRET-based screen was applied to the library to quantify differences in enterokinase 110 cleavage activity due to changes in the linker sequence immediately upstream of the enterokinase recognition element, thereby providing proof-of-concept of the assay. 4.2 Methods and Materials 4.2.1 Materials A l l chemicals and reagents, including TbCl3, were purchased from either Sigma-Aldrich (St. Louis, MO) or Fisher (Pittsburgh, PA) Chemicals and of analytical grade unless otherwise stated. Recombinant enterokinase (EKmax\u00E2\u0084\u00A2) was purchased from Invitrogen Corp. (Carlsbad, CA). The concentration of EKmax\u00E2\u0084\u00A2 was determined using the extinction coefficient provided by the vendor and verified using the commercial Bradford Protein Assay of BioRad Laboratories (Hercules, CA) using B S A as the standard. 4.2.2 Cloning and purification of fusion protein substrates A l l cloning was performed using standard molecular biology techniques (Sambrook and Russell 2001). Previously prepared pET28-CBM9-GFP expression vectors (Kavoosi et al. 2007) were digested with Pvu I and Not I to excise the GFP coding sequence, which was then amplified using the vector pGFPuv (Clontech, Palo Alto, CA). The enterokinase recognition sequence (D 4 K) (bold) and a Pvu I restriction endonuclease site (italic) were introduced at the 5' end of GFP using the oligonucleotide 5 ' - A A A G C G ^ 7 T G A T G A C G A C G A C A A G G C C A T G A G T A A - A G G A G A A G A A C T - 3 ' as primer. A Not I site (italic) was introduced at the 3'end using the oligonucleotide 5'-T G C G G C C G C T T T G T A G A G C T C A T C C A T G C C A T G T G T A A T C C C - 3 ' as primer using a PCR protocol previously described (Kavoosi et al. 2004). The compatible pET28-CBM9-Linker expression vectors comprising the library of CBM9-Linker coding sequences were prepared in a similar primer-assisted manner to allow the D4K-GFP coding region to be digested with Pvu I/Not I and ligated (16\u00C2\u00B0C, 16 h) into the digested expression vectors to give the appropriate pET28-CBM9-linker-D4K-GFP expression vector. D N A sequencing was performed to verify all constructs (NAPS Unit, Biotechnology Laboratory, The University of British Columbia). The vectors were 111 transformed into E. coli BL21 (DE3) cells and fusion protein expression and purification were performed as previously described (Kavoosi et al. 2007). 4.2.3 Preparation of apo-fusion protein Each purified fusion protein was incubated with Chelex 100X resin (Sigma-Aldrich; St. Louis, MO), mixing end-over-end at 4\u00C2\u00B0C for 48 h to strip all bound metal from the calcium binding sites of CBM9. The apo-protein was then passed over a Sephadex G10 size-exclusion column using nano-pure water as solvent and mobile phase to remove any trace contaminants. Each purified apo-fusion protein was stored at 4\u00C2\u00B0C in nano-pure water prior to use in the LRET-based assay. 4.2.4 Isothermal titration calorimetry studies Isothermal titration calorimetry (ITC) was performed using a VP ITC (MicroCal, Inc., Northampton M A ) . A l l samples were pH 6.9 in 20-mM Hepes or Tris buffer, 50-mM NaCl. Titrations were performed by injecting consecutive 5-uL aliquots of TbCb solution (0.60 mM) into an ITC cell (volume = 1.3528 mL) containing apo-CBM9-GFP (20 uM). The ITC data were corrected for the heat of dilution of the titrant by subtracting mixing enthalpies for 5-u.L injections of TbCb solution into protein-free buffer. Six and three independent titration experiments were performed for linkers PID4K and (PT)2PID4K, respectively; single titrations were also performed for linkers (PT) 4PID 4K, (PT) 7PID 4K, G3ID4K, G i 5 I D 4 K , and S 3 N 1 0 L A I D 4 K . A l l titrations were at 25 \u00C2\u00B0C. Binding stoichiometry, enthalpy and equilibrium association constants were determined by fitting the corrected data to a model for two independent binding sites on CBM9-GFP (software provided with instrument). 4.2.5 LRET-based assay of fusion-tag cleavage kinetics Absorbance spectra for solvated and CBM9-bound lanthanide ion (Tb(III)) were measured in a quartz cuvette over the range of wavelengths 300 to 700 nm on a Cary Eclipse fluorescence spectrophotometer equipped with a xenon pulsed lamp and a Czerny-Turner monochromator. Wavelength scans were performed at a rate of 1 nm s\"1. 112 Enterokinase-catalyzed cleavage of the CBM9 fusion tag can be described by the Michaelis-Menten equation (Boulware and Daugherty 2006; Gasparian et al. 2003), where the reaction rate (v) is related to the substrate (Tb 3 +-CBM9-linker-ID 4K-GFP) concentration [S\ by v = _ i ^ = j v j s ] ( 4 1 ) dt KU+[S] where v m a x is the maximum reaction rate, and KM is the Michaelis constant. At low substrate concentrations, where [5] \u00C2\u00AB K M , equation 4.1 reduces to an apparent first order reaction K [ET][S] (4.2) where [1ST] is the total enzyme concentration, & c a t is the catalytic constant, and the ratio hJK-u is a measure of substrate specificity. Equation 4.2 can be solved by integration to give the change in [S\ with time ln(^) = ln , J=_UMt 0J v0A.s' \u00E2\u0080\u0094 \u00E2\u0080\u009E * (4-3) where A:0bS'is the observed rate constant, and is the fraction of substrate remaining, which is directly measured by our LRET-based assay and which can be used to determine the initial reaction rate vQ at each initial substrate concentration [S]0. Our LRET-based assay was used to acquire initial reaction-rate (v0 and $/)) data for [S]0 values ranging from 1 uM to 48 uM. For each [S]0, kcat/KM was estimated by linear regression of equation 4.3 to individual ln($ versus t data sets, and the results (duplicates) were averaged to provide the reported kcJKu and associated standard deviation. A secondary check of the kcJKu value was obtained from the slope of a Michaelis-Menten plot as [S] approaches zero. If the two estimates of kcJKu were consistent, a Lineweaver-Burke plot was constructed to obtain a rough estimate of KM to 113 confirm that [5] \u00C2\u00AB KM. The basic LRET-based assay proceeded as follows. Chelex-prepared apo-fusion protein was incubated with 5x excess terbium in reaction buffer (20 mM Hepes, 50 m M NaCl, pH 6.9) for 15-30 min at room temperature, then centrifuged at 27,000 x g for 10 min to remove any precipitates. Tb(III)-loaded fusion protein (40 uL) at a concentration 2 x [S]0 was then mixed with an equivalent volume of EKmax\u00E2\u0084\u00A2 (0.74 pmole) solution to reach a final reaction volume of 80 uL. A l l reactions were carried out at 30\u00C2\u00B0C in a 384-well microplate (Corning Life Sciences; Big Flats, N Y ) and monitored on a POLARstar Optima microplate reader (BMG Lab Technologies, Inc.; Germany), a high-performance microplate reader designed for luminescence/fluorescence-based applications (note that the instrument used in this chapter differs from that of chapter 3). The system integrates a high-energy xenon flash lamp with an excitation and emission filter system to generate a clean reproducible signal and maximal sensitivity (filter sets for this instrument required excitation at 235 nm and emission at 520 nm). Black, low binding, 384-well plates (Corning Life Sciences; Big Flats, N Y ) were used. Samples were excited at 235 nm and sensitized emissions were measured at 520 nm for 1500 u.s following a 200 us post-excitation delay to eliminate background noise introduced by direct excitation of the acceptor fluorophore, GFP. Integration of the resulting fluorescence emission signal was then used to obtain the average fluorescence intensity at the given reaction time point. For each well in the microtitre plate, a reading was taken every 15 seconds. Raw fluorescence data were baseline corrected for losses in the fluorescence intensity of the unreacted Tb 3 +-CBM9-linker-ID4K-GFP control due to bleaching and other factors. The fluorescence intensity corresponding to [S]0 was determined from EKmax\u00E2\u0084\u00A2-free control wells to permit construction of the required ln(^) versus t plots. The kinetic parameters reported are the means of two independent experiments with the error computed as the propogation of this average error and the error in the linear fit of the data. The half-life t]/2 (h) of each reaction was determined from ty2 = ln2/k o b s. 114 4.2.6 Gel-based analysis of EKmax\u00E2\u0084\u00A2 cleavage reaction Tb 3 + -CBM9-PID 4 K-GFP fusion protein (32 uM) was prepared as described above, mixed with EKmax\u00E2\u0084\u00A2 (0.74 pmole) in reaction buffer to a final volume of 80 uL and incubated at 30\u00C2\u00B0C in an Eppendorf Thermomixer R (Hamburg, Germany) with shaking (5 s every hour). Reaction buffer was used in place of EKmax\u00E2\u0084\u00A2 in the control experiment. At each time point (0, 0.25, 0.5, 0.75, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5 and 24 h after initial mixing), 2 uL of the supernatant was removed and diluted with 2x loading buffer for analysis by SDS-PAGE to determine the percentage of cleaved fusion protein with time, from which the half-life tm (h) of each cleavage reaction was estimated. 4 . 3 Results and Discussion Terbium (III), like all lanthanide ions, is characterized by an emission spectrum comprised of sharp, fully-resolved peaks with exceptionally long decay times (Sammes and Yahioglu 1996). Unfortunately, the molar absorptivity of T b 3 + is extremely low. Excitation of water-solvated Tb 3 + can therefore only be accomplished with an intense laser light source. This problem can be overcome by binding the lanthanide ion to a sensitizing agent, such as CBM9, possessing a high molar absorptivity and an excited triplet state comparable to that of the emissive state of the lanthanide metal ion (Selvin 1996). Energy absorbed by the sensitizing agent may then be transferred to the terbium ion. The sensitizer must also function to shield the ion from the quenching effects of water (Parker and Williams 2003). Through its Ca -ion binding sites (Notenboom et al. 2001), the CBM9 affinity tag provides coordination sites for binding of up to three terbium (III) ions. Tb 3 + ion binding to each of these sites on the apo-CBM9-PID4K-GFP fusion protein was characterized by isothermal titration caloriometry (ITC) experiments, which reveal the presence of a single strong TV binding site (Ka = 2 (\u00C2\u00B13) x 10\u00C2\u00B0 M\" 1) and two somewhat weaker Tb 3 + sites (Ka = 2 (\u00C2\u00B12) x 105 M\" 1). Essentially identical ITC results were obtained for the binding of Tb 3 + to apo-CBM9-linker-ID4K-GFP fusion proteins containing each of the other linker sequences. Based on these Ka values, all three binding sites are saturated with Tb under the loading conditions used in this study. 115 Figure 4.1 shows the schematic diagram of the fusion protein analyzed in this LRET-based assay, with the linker sequences that make up the fusion-protein library listed in Table 4.1. The bound terbium and GFP correspond to the LRET donor and acceptor, respectively. For resonance energy transfer to occur between the donor and the acceptor, two conditions must be met: (i) the emission spectrum of the donor must overlap with the absorbance spectrum of the acceptor, and (ii) the donor/acceptor pair must be separated by a distance no greater than the wavelength of light A, emitted by the sensitized donor that is frequency matched to the corresponding absorbance peak of the acceptor (in this case meaning a distance less than 490 nm). Under these conditions, the excited terbium transfers a fraction of its energy to GFP via a long-range dipole-dipole coupling mechanism (Heyduk 2001), allowing real time monitoring of the decrease in fluorescence intensity of the acceptor resulting from cleavage of the CBM9-linker-ID4K affinity tag. 4.3.1 Characterization of energy transfer from bound Tb(lll) to GFP The spectral characteristics of both the donor (Tb 3 + -CBM9 on its own) and acceptor (GFP on its own) are shown in Figure 4.2. Excitation of CBM9-bound Tb 3 + at 222 nm results in four sharp emission peaks with maxima at 490 nm, 546 nm, 585 nm and 622 nm, respectively. Comparing this luminescence emission spectrum with the absorbance spectra of GFP (acceptor fluorophore) reveals an overlap between the Tb emission peak at 490 nm and the minor GFP absorption peak at 475 nm. The spectral overlap between CBM9-bound Tb and GFP indicates that the energy levels of the two match in frequency, allowing internal transfer of energy from excited terbium to GFP. Figure 4.3 shows the fluorescence emission spectrum for the acceptor resulting from energy transfer from the sensitized donor. Excitation of CBM9-bound terbium at 235 nm produces an oscillating electric dipole field. In the absence of a suitable acceptor, the field energy decays in the form of radiative energy, shown by the four characteristic luminescent emission peaks of Tb(III)-CBM9 in Figure 4.3. However, in the intact fusion protein (donor and acceptor), a fraction of the electric field energy of the sensitized donor is nonradiatively transfered to the acceptor (Selvin 1995). Excitation at 116 235 nm then produces five emission peaks (Figure 4.3): four corresponding to the characteristic emission signal of CBM9-bound Tb 3 + while the fifth, with a broad maximum at 510 nm, corresponding to sensitized emission from GFP. The average characteristic distance between bound terbiums and GFP in the various fusion proteins characterized in this study was determined (by a method previously described (Kavoosi et al. 2007)) to range from 35.9 \u00C2\u00B1 0.3 A to 37.2 \u00C2\u00B1 0.4 A. These values are much less than X (~ 500 nm), indicating that within the fusion protein, a significant fraction of the donor-derived energy is transferred to fused GFP in a nonradiative manner. 4.3.2 Fusion-tag bioprocessing kinetics Hydrolysis of the fusion protein by EKmax\u00E2\u0084\u00A2 separates the donor/acceptor pair such that R \u00C2\u00BB X, leading to a decrease in energy transfer and a corresponding decrease in the sensitized fluorescence emission of the acceptor. Figure 4.4 reports the time-resolved baseline-corrected decrease in GFP fluorescence emission for EKmax\u00E2\u0084\u00A2-catalyzed hydrolysis of Tb 3 + -CBM9-PID4K-GFP at an initial substrate concentration of 32 uM. The fluorescence emission intensity was monitored at 520 nm, where interfering signals from donor emission are minimized (241:1 maximum signal intensity ratio), as opposed to at 510 nm, where some signal overlap is observed. A n exponential decay in fluorescence intensity is observed for the hydrolysis reaction, reaching a minimum intensity of ca. 10000 after 5 hours of reaction. No reduction in this minimum intensity was observed when monitoring the same reaction for up to 48 hours, indicating that the background fluorescence is not related to unreacted substrate. This fact was further verified through SDS-PAGE documentation of the reaction. Instead, the background fluorescence is likely due to stochastic interactions between the cleavage products, Tb(III)-CBM9 and GFP. The resulting ln(^) versus t data (Figure 4.5) are linear (R2 = 0.991) at these reaction conditions for which, as will be shown later, [S]0 \u00C2\u00AB KM-Regression of equation 4.3 to the 520 nm data therefore provides an accurate measure of &obs, from which k^JKu can be directly determined (Table 4.1). Similarly, regression of equation 4.2 to the corresponding initial reaction rate v 0 versus [S]0 data (Figure 4.6) also 117 provides an estimate of kcJKu, which, for all fusion proteins studied, was statistically indistinguishable from the average value regressed from the ln($ versus t data. KM can be estimated from the x-intercept of a traditional Lineweaver-Burk plot. However, as is often noted, the error in . K M values obtained with this approach can be large since significant extrapolation to the x-intercept is generally required. This extrapolation bias can be reduced by fixing the slope of the Lineweaver-Burk plot at the previously determined value of &0bS and then positioning this slope to best fit the experimental data (Figure 4.7). The resulting KM estimates (Table 4.1) are then sufficiently accurate to confirm that KM is indeed much greater than [S]0 for each data set used to obtain estimates of kcJKu through regression of equation 4.3. Finally, confirmation that the decrease in the LRET signal is due to EKmax\u00E2\u0084\u00A2 mediated hydrolysis was obtained by reacting a fusion protein lacking the enterokinase recognition sequence with EKmax\u00E2\u0084\u00A2. No significant decrease in LRET signal was detected compared to the control reaction containing no enzyme (Figure 4.4). The accuracy of the kinetic parameters determined using our LRET-based assay was verified by comparing reaction half-lifes (t\n) measured using the LRET-based assay with those computed from an independent SDS-PAGE experiment. Figure 4.8 shows an SDS-PAGE documentation of the hydrolysis of Tb 3 + -CBM9-PID 4 K-GFP at a fusion protein concentration of 32 pM. The two methods provide very similar reaction half lifes (SDS-PAGE tm = 0.75 (\u00C2\u00B10.25) h; LRET tm = 0.61 (\u00C2\u00B10.02) h), validating the use of the LRET-based assay as an alternative and much more convenient method for monitoring fusion-tag processing, particularly in cases where the reaction kinetics are too rapid to permit visualization by SDS-PAGE. The LRET assay confirms that changes in linker composition and length can impact CBM9 fusion-tag bioprocessing kinetics (Table 4.1). A ca. 2-fold difference in t\a values, and thus bioprocessing times, is observed for relatively small library of Tb -CBM9-linker-ID4K-GFP fusion proteins included in this study. The assay is simple, fast and accurate, providing kcat/KM values that typically contain standard errors of less than 3%. As a result, both substantial and more subtle differences in bioprocessing kinetics 118 can be measured. For example, a 9.2% increase in kcJKu was observed when the length of the poly-glycine linker was increased from 3 to 15 amino acids. The 384-well micro-titre plate format of the assay should therefore allow for a number of useful high-throughput studies, including selection of an optimal linker chemistry as well as the preferred processing enzyme. 119 4.4 Tables Table 4.1 Michaelis Menten kinetic constants and reaction half-lives determined by the LRET-based assay EKmax\u00E2\u0084\u00A2 recognition site and linker sequence ( i v rY 1 x lfr4) hn (hr) (MM) PIDDDDK 3.43 \u00C2\u00B10.12 0.61 \u00C2\u00B1 0.02 4301 \u00C2\u00B1712 PTPTPIDDDDK 6.62 \u00C2\u00B10.10 0.31 \u00C2\u00B10.01 1439\u00C2\u00B11203 PTPTPTPTPIDDDDK 6.54 \u00C2\u00B10.12 0.32 \u00C2\u00B10.01 1328 \u00C2\u00B1850 PTPTPTPTPTPTPTPIDDDDK 5.22 \u00C2\u00B10.13 0.40 \u00C2\u00B10.01 626 \u00C2\u00B1 299 GGGIDDDDK 3.82 \u00C2\u00B10.11 0.55 \u00C2\u00B1 0.02 2119 \u00C2\u00B11071 G G G G G G G G G G G G G G G I D D D D K 4.17\u00C2\u00B1 0.12 0.50 \u00C2\u00B10.01 541 \u00C2\u00B1653 SSSNNNNNNNNNNLAIDDDDK 4.07 \u00C2\u00B10.11 0.51 \u00C2\u00B10.01 595 \u00C2\u00B1334 reported errors represent \u00C2\u00B12o (i.e. 95% confidence intervals) 120 4.5 Figures Figure 4.1 Schematic representation of the terbium bound fusion protein used for the LRET-based peptidase assay (not to scale). Excitation of bound Tb(III) at 235 nm results in the transfer of energy at 490 nm to GFP, leading to the emission of a signal at 510 nm. D4K indicates the recognition site for the serine peptidase, enterokinase. 121 100 t/> c 73 O N E 25000 - 20000 k15000 > U\" m o fi) O 10000 0 5000 | 300 350 400 450 500 550 600 650 700 Wavelength (nm) Figure 4.2 Individual spectrum of donor and acceptor. Solid line is the emission spectrum of CBM9-bound terbium (donor) obtained after excitation at 235 nm and a post-excitation delay of 200 us. Dashed line is the absorbance spectrum of GFP (acceptor). The spectral overlap (480 nm to 505 nm) between CBM9-bound terbium and GFP allows for energy transfer to take place. 122 400 450 500 550 600 650 700 Wavelength (nm) Figure 4.3 Emission spectra of CBM9-bound terbium (donor domain only) (solid line) and Tb 3 +-CBM9-(PT) 2PID4K-GFP fusion protein (dashed line). Samples were excited at 235 nm followed by a 200 (as delay prior to measurement of emission signal at each wavelength. Scans were performed at a rate of 1 nm/s. 123 50000 10000 J o -\ 1 1 \" 1 1 1 \u00E2\u0080\u00A2 1 > 1 0 1 2 3 4 5 Time (Hours) Figure 4.4 Real-time LRET-based signal from EKmax\u00E2\u0084\u00A2 mediated hydrolysis reaction. 32 u M of Tb 3 + -CBM9-PID 4 K-GFP fusion protein was incubated at 30\u00C2\u00B0C with 9.3 nM EKmax\u00E2\u0084\u00A2 (solid squares). The control reaction has buffer in place of enzyme (open squares). Sample was excited at 235 nm and emission was measured at 520 nm following a 200 u.s delay. Each measurement is an average of 10 excitations. 124 0.36788 -3 0.13534-3 0.04979 -i 0.01832 4 0.00674-i 0.00248 4 i\u00E2\u0080\u0094\u00E2\u0080\u00A2\u00E2\u0080\u0094i\u00E2\u0080\u0094'\u00E2\u0080\u0094i\u00E2\u0080\u00941\u00E2\u0080\u0094i\u00E2\u0080\u00941\u00E2\u0080\u0094i\u00E2\u0080\u00941\u00E2\u0080\u0094i\u00E2\u0080\u00941\u00E2\u0080\u0094i\u00E2\u0080\u00941\u00E2\u0080\u0094i\u00E2\u0080\u00941\u00E2\u0080\u0094r 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 Time (Hours) Figure 4.5 Baseline corrected logarithmic decay in ln(0 z = L, all t > 0 (5.14) Time and space domains were discretized by a Crank-Nicolson scheme (Crank and Nicolson 1996) to approximate differentials by a central difference in time and an average central difference in space. The column was meshed in the z dimension into N (at least 400) volume elements to match (or slightly exceed) the number of theoretical units (NTUs) within the column, and the number of radial volume elements within the stationary phase particle was set equal to 30. Finer meshing within the stationary phase increased computational time without a noticeable improvement in model accuracy. This 140 discretization of the model equations yields a set of tridiagonal linear algebraic equations that were solved by the Thomas algorithm (Holland and Liapis 1983) and the application of a first-order upwind-corrected power-law scheme (Patankar 1980) to ensure diagonal dominant matrices. Time increments for solution of equations 5.3 and 5.8 were set at 0.0\L/Nu and 0.002Z/M/, respectively, where L is the column length and u is the superficial velocity. 5.4 Materials and Methods 5.4.1 Chromatographic Media and Reagents Perloza\u00E2\u0084\u00A2 MT100 was purchased from Iontosorb Inc. (Czech Republic). Perloza\u00E2\u0084\u00A2 MT100 is a porous, spherical media derived from regenerated cellulose. The particle diameter (dp) distribution of Perloza\u00E2\u0084\u00A2 MT100 as well as the average dp was determined by light scattering using a Malvern Mastersizer 2000 (Malvern Instruments, UK). The average particle volume was calculated assuming a spherical geometry from which the mean particle diameter of the sphere was obtained using MIE theory (Barber and Hil l 1990). Perloza\u00E2\u0084\u00A2 MT100/G15 has a particle diameter (dp) distribution of 56 urn to 159 um with a mean value of 84 \u00C2\u00B1 0.6 um. Sephadex G15 and Ni+ 2-Sepharose IMAC media were obtained from Sigma-Aldrich (Mississauga, ON, Canada) and Novagen (Milwaukee, MI), respectively. 5.4.2 Scanning Electron Micrographs Scanning electron micrographs of Perloza\u00E2\u0084\u00A2 MT100 media and its associated pore structure were obtained using a Hitachi S-4700 Field Emission scanning electron microscope operating at an accelerating voltage of 5 kV with a working distance of 5 to 15 mm. Samples were prepared by loading a high-pressure freezing hat with approximately 3 uL of concentrated Perloza\u00E2\u0084\u00A2 MT100 in nano-pure water. The hat was immersed for 5-7 seconds in subcooled liquid nitrogen (-210 \u00C2\u00B0C), then fractured open by microtome cleavage. The exposed resin surface was mounted and then sputter coated with gold for 30 s to generate the appropriate phase contrast for imaging. 141 5.4.3 Protein Production The cloning of CBM9 and CBM9-GFP is reported elsewhere (Kavoosi et al. 2004). CBM9 or CBM9-GFP was produced in a 60 L fermentation as follows. BL21 (DE3) cells containing the pET28-CBM9-GFP expression vector were grown at 37\u00C2\u00B0C in Laural broth (LB) to a cell density (OD600 nm) between 0.8 and 1.0. Protein expression was induced with isopropyl-l-thio-P-D-galactoside (IPTG) to a final concentration of 0.1 mM and the cells allowed to incubate for a further 10-12 h at 30\u00C2\u00B0C. The cells pellet was resuspended in high salt buffer (1 M NaCl, 50 m M potassium phosphate, pH 7.0), ruptured by two passages through a French pressure cell (21000 lb in\"2), and the cell debris was removed by centrifugation (27,000 x g) for 30 min at 4\u00C2\u00B0C. Highly pure CBM9 or CBM9-GFP was obtained by first passing the clarified cell extract over a Pharmacia XK-16 column (10 cm x 1.6 cm I.D.) packed with the Perloza\u00E2\u0084\u00A2 MTlOO-based composite media. Contaminating' proteins were removed by washing the column with 10 column volumes (CV) of high salt buffer, followed by 5 C V of low salt buffer (150 m M NaCl, 50 m M potassium phosphate, pH 7.0). CBM9 or CBM9-GFP was then eluted from the column with 2 C V of 1 M glucose in TBS8 (15 m M NaCl, 10 m M Tris-HCl, pH 8.0) and the eluent peak was injected into a column packed with Ni+ 2-Sepharose I M A C media and purified according to the manufacturer's instructions. The protein eluted from the IMAC column was buffer exchanged into low salt buffer, concentrated in a stirred ultrafiltration (UF) unit (Amicon, Beverly M A ) and stored at 4\u00C2\u00B0C until use. The concentration of the purified protein was determined by U V absorbance (280 nm) using a calculated molar extinction coefficient of 43100 M \" 1 cm\"1 (CBM9) or 62870 M \" 1 cm\"1 (CBM9-GFP) (Mach etal. 1992). 5.4.4 Equilibrium binding isotherms Equilibrium isotherms for binding of CBM9 and CBM9-GFP to Perloza\u00E2\u0084\u00A2 MT100 and to the Perloza\u00E2\u0084\u00A2 MT100/G15 composite media were measured at pH 7.0 and 4\u00C2\u00B0C. Purified protein at concentrations ranging from 1 to 30 u M was mixed with stationary phase media (1 mg dry weight) in low salt buffer to a final volume of 1 ml. Samples were then incubated overnight under continuous end-over-end rotation. The 142 media was removed by centrifugation (27,000 x g) for 16 min. The supernatant was collected and the concentration of free protein was determined by U V absorbance (280 nm) using a Cary 100 Spectrophotometer (Varian). The resulting isotherm was generated by plotting the concentration of bound protein (pmol g\"1 dry resin), determined by total mass balance, against the concentration of free protein (mol L\" 1). Binding parameters were then determined by non-linear regression of the Langmuir adsorption isotherm equation to the experimental data using GraphPad Prism 3.0 software. 5.4.5 Confocal Laser Scanning Microscopy (CLSM) Time-course fluorescence intensity profiles for uptake of CBM9-GFP into Perloza\u00E2\u0084\u00A2 MT100 media packed into an optically transparent two-dimensional chromatography column were measured according to the procedures described in (Hubbuch et al. 2002) using an inverted Zeiss L S M 510 confocal laser scanning microscope equipped with a water immersion 63x/NA1.2 C-Apochromat (Zeiss) objective and an argon laser. The GFP chromophore has an absorbance maximum at 475 nm. Excitation was conducted at 488 nm, on the red-shifted shoulder of the absorbance peak, to attenuate the resulting fluorescence emission at 505 nm to avoid saturation of the signal. The laser intensity was kept constant for all experiments. Slight adjustments (\u00C2\u00B120 V) of the photomultiplier (PMT) detector gain were necessary to account for signal attenuation effects from neighboring particles in the packed bed. A n 80/20 filter and no-band or long-pass filter before the PMT were used to improve signal-to-noise. The fluorescence intensity profile within the central particle was measured every second with 300-500 frames in total. A l l profiles were stored as eight-bit single scans with a resolution of 512 x 512 pixels representing an area of 146.2 x 146.2 m. The chosen time interval allowed monitoring of the diffusion both into and out of the particle in one run. Bleaching of GFP was not observed. Before each individual run, the reflection mode of the microscope was used to verify that the focal plane went through the center of the particle. The image frame consisted of the particle of interest, other particles (focal plane 143 not necessarily through particle center), and interstitial areas between the particles, from which the average bulk protein concentration was determined. 5.4.6 Characterization and Application of Perloza\u00E2\u0084\u00A2 MT100/G15 Composite Media Column A Pharmacia Inc. (GE Healthcare) FPLC system with two P-500 reciprocating pumps, an 8-port mixing and injection valve, a UV-MII flow spectrophotometer (monitoring absorbance at 280 nm), and a Frac-200 fraction collector was used to measure all chromatograms. Columns 7 to 10 mL in volume (Pharmacia Inc. HR-10 column (1.0 cm I.D.) were packed with either degassed Perloza\u00E2\u0084\u00A2 MT100 or a 50% Perloza\u00E2\u0084\u00A2 MT100 (by mass)/50% Sephadex G15 degassed slurry under a superficial velocity u of 4.25 x 10\"4 m s\"1 and then equilibrated with degassed loading buffer (50 m M potassium phosphate, 150 m M NaCl, 0.02% NaN 3 , pH 7.0) prior to use. In all columns used, the column length L to diameter dc ratio (L/dc) was maintained well above 2 to minimize end effects (Jungbauer 1993). In addition, the volume of solute pulses used for moment analysis was kept well below 0.5% of the column void to minimize any precolumn solute dispersion effects (Ladisch et al. 1984). Wall effects could be ignored since the column diameter to particle diameter ratio was much higher than 30 (Knox et al. 1976). 5.5 Results and Discussion 5.5.1 Geometric and Sorption Properties of Perloza\u00E2\u0084\u00A2 MT100/Sephadex G15 Composite Media Perloza\u00E2\u0084\u00A2 MT100 is a highly porous, hydrophilic media derived from regenerated cellulose with its structural elements comprised of partially microcrystalline regions stabilized by interchain hydrogen bonds (Iontosorb, CZ). Scanning electron micrographs (SEM) of the spherical MT100 particles (Figure 5.1 A) show that this media offers a network of primarily submicron pores with a small percentage of larger pores up to 2 um in nominal diameter (Figures 5.IB and 5.1C). Simple hydrodynamic calculations 144 proposed by Liapis (Heeter and Liapis 1997) predict for the 80-u.m particles and range of linear solvent velocities used in this work that intraparticle convective flow requires pores greater than ca. 4 to 5 um in nominal diameter. Thus, diffusion is the primary solute transport process within this stationary phase media. Adsorption of CBM9 and CBM9-tagged fusion proteins to Perloza\u00E2\u0084\u00A2 MT100 has previously been shown to follow Langmuir-type adsorption behavior (Kavoosi et al. 2004). Efficient purification from clarified E. coli cell lysates of CBM9-tagged fusion proteins on analytical-scale MT100 columns (i.e. less than 5 ml in volume) has also been demonstrated (Kavoosi et al. 2004). Due to the compressible gel-like nature of the MT100 matrix, scale-up of a packed MT100 column to a volume above ca. 10 to 15 mL is compromised by stationary-phase compression effects that degrade column performance. Mechanical stabilization of the stationary phase is therefore required and can be achieved through addition of an inert support media, in this case Sephadex G15, that provides a rigid mechanical scaffold stabilizing the active stationary phase to allow stable columns to be prepared on the semi-preparative to preparative scale. Here, however, our focus is on validation of our proposed generalized two-zone model of adsorptive chromatography. Equilibrium isotherms for pure CBM9-GFP binding to a 50/50 (by dry mass) composite media of MT100/G15 (Figure 5.2), and regression of the Langmuir isotherm parameters (Table 5.1) to those data show that the addition of the G15 mechanical support does not affect the equilibrium association constant (Ka) for CBM9-GFP binding to MT100 but reduces the static capacity of the composite media (q{sat) to half its value when the stationary phase consists of pure Perloza\u00E2\u0084\u00A2 MT100. This is expected since q\u00E2\u0084\u00A21 is reported in terms of the total volume of the stationary phase, only half of which is MT100 in the composite media suitable for scale-up. Nevertheless, due to the extraordinarily high binding capacity of MT100 for CBM9 tagged fusion proteins (ca. 11-12 pmol g\"1 dry resin), the MT100/G15 composite column is capable of binding 45 to 60 mg of CBM9-GFP per column mL, making it highly competitive with popular commercial affinity chromatography media.. Finally, control experiments confirm that the binding interaction is between CBM9 and cellulose (Table 5.1). Interactions between 145 CBM9-GFP and Sephadex G15 and between untagged GFP and Perloza\u00E2\u0084\u00A2 MT100 are insignificant. 5.5.2 Purification of CBM9-GFP on Perloza\u00E2\u0084\u00A2 MT100/G15 Composite Media Column The chromatogram and gel documentation for purification of CBM9-GFP from a clarified cell lysate are shown in Figure 5.3 and Figure 5.4, respectively. Eluent from the column was continuously monitored both for total protein using absorbance at 280 nm (A280) and for CBM9-GFP using the intrinsic fluorescence of GFP (excitation at 395 nm and emission at 509 nm). Initial breakthrough of contaminating proteins was observed just after 1 column void volume. The small fluorescence peak observed within the contaminant breakthrough peak is due to the presence in the feed mixture of free GFP, released as a result of a small amount of degradation (no protease inhibitors were used during the purification) within the linker region connecting CBM9 to GFP (Kavoosi et al. 2007). Two wash steps were used to effectively remove most of the contaminating proteins present in the column. As evident by the overlapping A280 and fluorescent signals (Figure 5.3), addition of 1 M glucose to the mobile phase elutes CBM9-GFP as a single sharp peak. Table 5.2 compares the yield, purity and concentration factor for the affinity purification of CBM9-GFP on the MT100/G15 composite column relative to that achieved on the pure MT100 column (Kavoosi et al. 2004). The performance of the two columns is very similar at the relatively small column scales used in this study (10 mL composite column, 7 mL pure MT100 column). 5.5.3 Characterization of Solute Mass Transfer Within Perloza\u00E2\u0084\u00A2 MT100/G15 Composite Columns Measured solute (CBM9-GFP) mass-transfer and column-geometry parameters for our proposed generalized two-zone model for mass transport in a Perloza\u00E2\u0084\u00A2 MT100/G15 composite media column are listed in Table 5.3. Good column packing uniformity, defined according to the guidelines proposed by Bristow and Knox (Bristow 146 and Knox 1977), is indicated both by a peak asymmetry factor (As) close to unity and a reduced plate height value (h) close to 3 for pulse injection of 100 uL at a superficial velocity of 2.1 x 10\"3 cm s\"1. The column void fraction \u00C2\u00A3was determined by pulse injection of blue dextran (MW = 2,000,000 Da) as a function of the superficial velocity, u (m s\"1). Elution peaks for this large, non-binding solute were Gaussian or very nearly Gaussian in shape. Direct computation of the first moment ( fix) of each elution peak and application of the theory of Haynes and Sarma (Haynes and Sarma 1973) yielded sfrom the slope of a plot of versus u'1 (Figure 5.5) under the approximation that sp = 0. The estimated value of s is slightly higher than the value expected for a column packed uniformly with spherical beads (Ladisch et al. 1984), but agrees well with an independent measure of s obtained from application of the Blake, Kozeny and Carmen (Allen 1997) equation for pressure drop data across the column measured as a function of u, where rj is the fluid viscosity (g m\"1 s\"1). For spherical packing, the aspect factor k is assumed equal to 5 (Janson and Ryden 1998). Protein-based probes were used to determine the effective porosity Sp of the MT100/G15 composite media as a function of protein molecular mass (Figure 5.6). Each data point in the figure was determined by measuring /u\ as a function of u for the respective marker and application of equation 5.15 using the measured void fraction of 0.425. Interpolation between measured ep data was facilitated by fitting to an exponential decay type equation of the form: (5.15) AP = 36k (5.16) 147 -M/ ep = ae / y -8 (5.17) where a, 5 and y are fitted parameters and M is the protein molecular weight (kg mol\"). The solid curve in Figure 5.6 represents the best fit, for which a = 0.815, 8 = 0.033 and y = 207.8. Numerical determination of the second moment (o2) of the elution peak as a function of the interstitial velocity, u0, was combined with the Laplace transform results of Haynes and Sarma (Haynes and Sarma 1973) for operation within the linear region of the adsorption isotherm u\u00E2\u0080\u009Ecr L 2rf = D L + K 1 ( \u00C2\u00A3 ^ 1 + -2 (5.18) to obtain values for the parameters characterizing mass transfer of the CBM9-GFP fusion protein within the Perloza\u00E2\u0084\u00A2 MT100/G15 composite media column. In equation 5.18, which has been used in many similar parameter estimation studies (e.g. (Giddings 1965; Guiochon et al. 1994; Ladisch 2001)), KM is the overall solute mass transfer coefficient, given by K M 3\u00C2\u00A3y rP 1 \5sPDp k ads V 2 1 + f (5.19) where RM is the overall resistance to solute mass transfer, k/ is the film mass-transfer coefficient (m s\"1) and kacis is the sorption rate constant (M\" 1 s\"1). Equations 5.18 and 5.19 can be applied under both binding and nonbinding conditions to estimate the axial dispersion coefficient (DL) (m 2 s\"1) and the overall mass transfer coefficient (%) from the y-intercept and slope, respectively, of a plot of u0r^L/(2jui2) versus u02. Results for CBM9-GFP under nonbinding conditions (i.e., in the presence of 2-M glucose) are shown in Figure 5.7, from which the parameters in Table 5.3 were determined following estimation of the film mass transfer coefficient using the correlation of Wilson and Greankoplis (Wilson and Geankoplis 1966) 148 Sh = (5.20) V 8 J for solute mass transfer in a packed bed of porous, spherical particles at conditions where 0.0016 < Re < 55 and 165 < Sc < 70600. The correlation of Young and Carroad (Young et al. 1980) was used to estimate the bulk molecular diffusivity DM (m 2 s\"1) of CBM9-GFP where T is the temperature (K) and M is the molecular mass of the solute in units of g estimated using either the correlation of Wakao et al. (Wakao et al. 1958) or of Goto et al. (Goto etal. 1983). First and second moments analysis of the first derivative of breakthough curve data for frontal loading of pure CBM9-GFP under binding conditions indicates that the last term on the right-hand side of equation 5.19 makes less than a 1% contribution to RM-This indicates that kads > 0.33 m 3 mole\"1 s\"1, which is consistent with the value of kads determined by Jervis et al (Jervis et al. 1997) for the binding of the family 2a carbohydrate binding module (CBM2a) to the surface of crystalline cellulose. The measured mass-transfer and adsorption-kinetics parameters allow calculation of the column Peclet (Pe = 2rpua/Dp), Biot (Bi = kfrp/3Dp) and Damkohler (Da = C\"kadsrp/DP) numbers (Table 5.3), which together indicate that pore diffusion limits CBM9-GFP mass transport within the composite media column. As a result, the local equilibria approximation can be invoked for the column loading process, with elution band profiles and the rate of solute uptake within the stationary phase best described by an appropriately formulated pore diffusion model of chromatography. As noted previously, relatively slow rates of intraparticle diffusion often limit affinity chromatography processes as well as other forms of adsorptive chromatography (Jungbauer 1996; Ladisch 2001), particularly under conditions where the concentration of solute in the feed is significantly greater than XIKa (Hall et al. 1966), as was the case in (5.21) mole\"1. The estimate of fu s ing this approach is in close agreement (\u00C2\u00B110%) with that 149 our experiments. The CBM9-GFP affinity capture process therefore provides a suitable system for investigating the advantages of treating intraparticle solute uptake in pore-diffusion-limited adsorptive chromatography processes using the proposed generalized two-zone model of affinity chromatography relative to using the traditional P D M for spherical sorbent particles encoded in equations 5.1 through 5.6. 5.5.4 CLSM-Derived Rates of CBM9-GFP Uptake Time-course fluorescence-intensity profiles (Figure 5.8) for batch uptake of CBM9-GFP into a Perloza\u00E2\u0084\u00A2 MT100 particle routinely show a two-zone behavior characterized by a region within the interior of the sorbent particle containing no protein, and a second region in which S[ is nonzero and increases with r. In this experiment, the interstitial volume of the 100 uL viewing chamber was rapidly flooded with CBM9-GFP feed solution to an initial concentration of 5.4 u M to permit fluorescence intensities both within the central particle and within the surrounding interstitial volume to be monitored as a function of time. Fluorescence intensities in the outer protein-containing shell of the stationary phase particle do not exhibit the square-wave characteristics predicted by the traditional shrinking-core model, where adsorption equilibrium is described by the rectangular isotherm. Instead, finite concentration gradients within the shell region are observed, whose description is better described through use of a more realistic equilibrium relationship such as the Langmuir-type isotherm used in the traditional P D M and the two-zone model proposed here. Time-dependent radial profiles of CBM9-GFP uptake within a Perloza\u00E2\u0084\u00A2 MT100 particle predicted by the P D M (Figure 5.9A) and by the new T Z M (Figure 5.9B) are compared with normalized fluorescence intensities measured by C L S M for an interstitial o feed concentration (Cj ) of 5.4 uM. This feed concentration represents conditions within the linear portion of the adsorption isotherm, and both models accurately capture initial rates and profiles of CBM9-GFP uptake under this moderate loading condition. Divergence of the traditional P D M from the measured solute uptake profiles is observed after ca. 10 to 18 minutes of exposure, with the model predicting a faster rate of protein uptake due to an unrealistically rapid accumulation of solute near the center to the 150 stationary phase particle - an error created by the continuous nature of the model that requires both c\(r=0,t) and q\(r=Q,i) to become nonzero immediately upon contact of the stationary phase particle with the mobile phase liquid. By eliminating this approximation and incorporating a realistic isotherm model (relative to the traditional shrinking core model), the T Z M accurately predicts the experimentally observed rates and profiles of solute uptake within the Perloza\u00E2\u0084\u00A2 MT100 particle throughout the loading process. This includes model predictions of rc(t), which agree with C L S M estimated values to within experimental error (Figure 5.10). Errors in rates of CBM9-GFP uptake predicted by the P D M decrease with 0 increasing protein load in the feed. When C\ is increased to 49 pM, adsorption equilibrium lies in the nonlinear region of the isotherm, resulting in a more rapid penetration of solute into the sorbent particle and improved agreement of P D M predictions with both experiment and T Z M predictions (Figure 5.11). Indeed, under nonlinear or overload conditions, model predictions are more sensitive to the choice of isotherm model than to improvements provided by the T Z M . 5.5.5 Simulation of Breakthrough Curves Figures 5.12A and 5.12B compare model predictions to normalized breakthrough curve data for the cases where binding equilibrium is within the linear and non-linear regions of the isotherm, respectively. Under linear binding conditions, the T Z M agrees with experiment while the traditional P D M over-predicts the dynamic capacity of the column, resulting in a significant delay in the predicted onset of breakthrough. Both models provide a reliable prediction of breakthrough behavior when the concentration of CBM9-GFP in the feed is increased to nonlinear loading conditions due to the more rapid rate of protein uptake in the stationary phase. Model agreement, however, is not exact. The experimental breakthrough curve is slightly asymmetric, such that the leading edge of the breakthrough transition is a bit sharper than the approach to saturation. This slight asymmetry, present in the other breakthrough curves measured at low linear velocities, could be due to a number of factors not fully captured in either of the two models, 151 including non-uniform mixing in some regions of the bed, non-specific adsorption, or conformational changes in the adsorbed protein (Johnston and Hearn 1990). T Z M predictions are compared in Figure 5.13 to raw breakthrough data to show that the model captures changes in elution band profiles and dynamic capacity over a wide range of feed concentrations. The T Z M also predicts the dependence of the elution band profile on u (Figure 5.14) and thus the expected changes in solute dispersion and dynamic capacity (Mao et al. 1991; Mao et al. 1995). 5.5.6 CBM9-GFP Breakthrough from a Clarified Cell Extract Feed Figures 5.12 to 5.14 present breakthrough data for a binary buffer solution containing pure CBM9-GFP. We have previously shown that our custom expression vectors for expression of CBM9-tagged fusion proteins direct high-level production of the chimeric protein, typically yielding concentrations of soluble product between 0.4 to 5 grams per liter (Kavoosi et al. 2004). This high level expression is exemplified in Figure 5.4, where soluble CBM9-GFP comprises approximately half of the total protein within the clarified cell extract. Nonbinding contaminant effects on mass transport and affinity binding of the target protein are therefore likely to be relatively minor, suggesting that T Z M predictions using the one-component Langmuir isotherm and pure-component transport parameters (Table 5.3) may be sufficient to predict product elution bands for a clarified cell extract feed containing CBM9-GFP. Figure 5.15 compares predictions of this simplified T Z M to breakthrough data for frontal loading of a clarified cell extract containing 25 \u00C2\u00B1 0.7 u M CBM9-GFP onto a 7.5 3 1 ml Perloza\u00E2\u0084\u00A2 MT100/G15 composite column at a superficial velocity of 8.5 x 10\" cm s\" . The breakthrough of CBM9-GFP was continuously monitored using the intrinsic fluorescence of GFP (excitation at 395 nm, emission at 510 nm). The results confirm that our pseudo-binary solution assumption is accurate provided the fusion-protein makes up a high percentage of the total protein in the feed and the total protein load in the feed is not greater than ca. 3 g L\" 1 . 152 5.6 Conclusions We have introduced a novel generalized two-zone model for adsorptive chromatography and compared it to the traditional pore-diffusion model for protein uptake within a porous stationary phase. The T Z M divides each stationary phase particle into two zones, an inner protein-free core and an outer zone into which a finite mass of protein has penetrated. The T Z M replaces the rectangular isotherm of the traditional shrinking-core model with Langmuir theory, thereby allowing for simultaneous intraparticle mass transport and sorbent loading within the outer region of the porous particle by accounting for the presence of a protein concentration gradient within this outer region. The model therefore improves upon the P D M by eliminating a boundary condition that forces the unrealistic prediction of a nonzero protein concentration at the center of the particle immediately upon contact of the empty particle with a protein-containing mobile phase. Confocal laser scanning microscopy results indicate that under linear loading conditions, both models accurately predict the initial rate of solute uptake. As loading time increases, the P D M deviates from experimental results due to errors associated with the continuous nature of the model, while the T Z M continues to accurately predict experimental data. Under nonlinear loading conditions, both models performed well, with the choice of isotherm model becoming the more critical factor in determining the accuracy of the prediction. Finally, the T Z M was able to predict product breakthrough over a range of feed concentrations and superficial velocities, including accurate prediction of product breakthrough during frontal loading of a clarified cell extract. 153 5.7 Tables Table 5.1 Measured Langmuir isotherm parameters for binding of CBM9-GFP and each of its fusion partners to Perloza\u00E2\u0084\u00A2 MT100, Sephadex G15, and the composite MT100/G15 media at 4 \u00C2\u00B0C. The solvent consisted of 50 mM potassium phosphate, 150 m M NaCl, pH 7. N B indicates that no binding was observed.. P e r l o z a \u00E2\u0084\u00A2 M T 1 0 0 M T 1 0 0 / G 1 5 C o m p o s i t e S e p h a d e x G 1 5 P r o t e i n Ka x 1 0 6 ( M 1 ) \u00E2\u0080\u009E max ( u m o l / g ) Ka x 1 0 6 ( M - 1 ) \u00E2\u0080\u009E max Ii ( u m o l / g ) Ka x 1 0 6 ( M 1 ) ( u m o l / g ) CBM9-GFP 1.5 (\u00C2\u00B10.21) 9.93 (\u00C2\u00B10.31) 1.2 (\u00C2\u00B10.15) 5.27 (\u00C2\u00B10.18) NB NB CBM9 1.2 (\u00C2\u00B10.06) 11.2 (\u00C2\u00B10.15) 0.85 (\u00C2\u00B10.15) 5.67 (\u00C2\u00B10.32) GFP NB NB * reported errors represent \u00C2\u00B12o (i.e. 95% confidence interval) 154 Table 5.2 Yield, purity and concentration factor for the affinity purification of CBM9-GFP on a MT100/G15 composite column and on a pure MT100 column. Clarified E. coli cell lysate containing CBM9-GFP was loaded onto each column at a superficial velocity of 4.25 x 10\"3 cm s\"1. Fractions were collected and analyzed both by absorbance at 280 nm and by fluorescence intensity to obtain reported data. The total lysate volume loaded was 65 mL and contained a CBM9-GFP concentration of 5.7 uM. MT100/G15 Column MT100 Column Purity* > 95% > 95% Yield 80 \u00C2\u00B1 3% 82 \u00C2\u00B1 3% Concentration Factor (Q/C\u00C2\u00B0) 29 \u00C2\u00B1 2 31 \u00C2\u00B1 2 *Determined from SDS-PAGE analysis 155 Table 5.3 Measured mass-transfer and column-geometry parameters for CBM9-GFP transport in a Perloza\u00E2\u0084\u00A2 MT100/G15 composite media column. Column Property Value Units L 7.9-11.4 cm dc 1.0 cm M 83.7 \u00C2\u00B10 .6 pm PP 1.33 \u00C2\u00B10.73 kg m\"3 8 0.425 \u00C2\u00B10.015 sp 0.65 \u00C2\u00B1 0.03 h (reduced plate height) 2.9 \u00C2\u00B10 .3 As (peak asymmetry factor) 1.20 \u00C2\u00B10.03 DL 3.0 x 10\"5 m V 1 kf 8.45 x 10\"6 m s\"1 DM 1.54 x 10\"11 m V 1 Dp 7.17 x 10\"12 m V Pe 500-2000 Bi 16.4 Da >82 * reported errors represent \u00C2\u00B12o (i.e. 95% confidence interval) 156 5.8 Figures Figure 5.1 Scanning electron micrographs of Perloza\u00E2\u0084\u00A2 MT100 beaded media. Figures A , B and C show an MT100 particle at 1.0k, 13.0k and 80.0k magnification, respectively. The pore structure shown at the center of Figure B is further magnified and shown in Figure C. 157 6 c.(M\"1) (x10 6 ) Figure 5.2 Equilibrium adsorption isotherm for batch binding of CBM9-GFP to Perloza\u00E2\u0084\u00A2 MT100/G15 composite media at 4\u00C2\u00B0C. The solvent consisted of 50 m M potassium phosphate, 150 m M NaCl, pH 7. The solid curve represents the best fit of the experimental data to the Langmuir isotherm equation. 158 0 40 80 120 160 200 240 280 Volume (mL) Figure 5.3 Chromatogram for CBM9-GFP purification on Perloza\u00E2\u0084\u00A2 MT1007G15 composite media. Clarified E. coli cell lysate containing CBM9-GFP was loaded at a superficial velocity of 4.25 x 10\"3 cm s\"1 onto a 10 mL column. Fractions were collected and analyzed both by absorbance at 280 nm (line) and by fluorescence intensity (solid diamond) as shown. CBM9-GFP Figure 5.4 SDS-PAGE documentation of CBM9-GFP purification on a 10 mL Perloza\u00E2\u0084\u00A2 MT100/G15 column. A l l samples were dissolved in sample buffer containing 10% SDS. Lane M : molecular mass markers; Lane 1: clarified cell lysate prior to column loading; Lane 2: column flow through; Lane 3: high salt wash; Lane 4: low salt wash; Lane 5: CBM9-GFP eluted in low salt buffer containing 1 M glucose. 160 1400 L If1 (S) Figure 5.5 First moment (pi) analysis for a 10 mL column packed with Perloza\u00E2\u0084\u00A2 MT100/G15. Integrated pi values are reported for pulse injections of a 50 pL solution of blue dextran (MW 2000 kDa) over the interstitial velocity range 4.25 x 10\"3 cm s\"1 to 2.02 x 10\"2 cm s\"1. The mobile phase consisted of 50 m M phosphate buffer, 150 m M NaCl (pH 7, 4\u00C2\u00B0C). 161 0 400 800 1200 1600 2000 Molecular Weight (kDa) Figure 5.6 Measured average porosity ep of composite media as a function of molecular weight of standard proteins. Pulse injections of standard molecular weight protein markers were used over the interstitial velocity range 2.1 x 10\"3 cm s\"1 to 8.5 x 10\"3 cm s\"1. The porosity was determined from equation 5.15 using the measured void fraction of 0.425. 162 0 -) 1 1 r 1 1 1 i 1 T 1 \u00E2\u0080\u00A2 1 0 2 4 6 8 10 12 u o 2 (x10 8 ) (ms\"1)2 Figure 5.7 Determination of axial dispersion coefficient \u00C2\u00A3>L and overall solute mass-transfer coefficient KM for injection of CBM-GFP under nonbinding conditions onto a 10 mL column packed with Perloza\u00E2\u0084\u00A2 MT100/G15. Pulse injections of a 50 uL solution of CBM9-GFP was used over the interstitial velocity range 4.25 x 10\"3 cm s\"1 to 1.38 x 10\"2 cm s\"1. The mobile phase consisted of 50 m M potassium phosphate, 150 m M NaCl (pH 7, 4\u00C2\u00B0C) with 2 M glucose added to achieve nonbinding conditions. 163 gure 5.8 Time course fluorescent intensity profile of CBM9-GFP uptake into Perloza\u00E2\u0084\u00A2 MT100 particle. Protein uptake monitored by an inverted Zeiss L S M 510 confocal laser scanning microscope with the center of the particle used as the focal plane. Excitation and emission wavelengths were 488 nm and 505 nm, respectively. The initial CBM9-GFP 0 concentration Cx outside the particle was 5.4 pM. 164 Figure 5.9 Time dependent radial profiles of CBM9-GFP uptake into a Perloza\u00E2\u0084\u00A2 MT100 particle for a feed concentration of 5.4 uM. Predicted uptake rates using (Figure 9A) pore-diffusion model and (Figure 9B) two-zone model are compared with experiment. The rapid drop in fluorescence intensity at radial positions above ca. 39 pm indicates the position of the outer radius of the bead. 165 0 1 0 2 0 3 0 4 0 5 0 6 0 Load Time (min) ure 5.10 Comparison of T Z M model predictions of rc(t) with values computed from C L S M data. C L S M determined core radius reported as the radius at which the measured fluorescence intensity falls below 3X the standard deviation of the background fluorescence. Load conditions same as stated in Figure 5.9. 166 - I \u00E2\u0080\u00A2 1 1 1 \u00E2\u0080\u00A2 1 ' 1 1 J\u00E2\u0080\u0094 0 10 20 30 40 50 Radial Position (um) 10 20 30 40 50 Radial Position (um) Figure 5.11 Time dependent radial profiles of CBM9-GFP uptake into a Perloza\u00E2\u0084\u00A2 MT100 particle for a feed concentration of 49 pM. Predicted uptake rates using (Figure 11 A) pore-diffusion model and (Figure 11B) two-zone model are compared with experiment. 167 0 0 0 3 5 -0 0 0 3 0 -0 0 0 2 5 -i 0 a, 0 0 0 2 0 -0 0 0 1 5 -0 0 0 1 0 -0 0 0 0 5 -0 0 0 0 0 --0 0 0 0 5 -400 800 1200 1600 Elution Volume (mL) 2000 o a 0.020 0.0154. 0.0104 O 0.005 4 0.000 4 -50 0 50 100 150 200 250 300 350 400 Elution Volume (mL) Figure 5.12 Comparison of T Z M (solid curve) and P D M (dashed curve) predictions with experimental (points) breakthrough curves. Pure CBM9-GFP loaded at a superficial velocity of 8.5 x 10\"3 cm s\"1 onto a Perloza\u00E2\u0084\u00A2 MT100/G15 composite media column: (A) frontal load of 3.4 x 10\"3 mol m\"3 CBM9-GFP, (B) frontal load of 1.85 x 10\"2 mol m\"3 CBM9-GFP. 168 0.050 -j 0.045-0.040 -0.035-CO 1 0.030-E 0.025-o -E 0.020 -o\" 0.015-0.010-0.005 -0.000-100 200 300 400 Elution Volume (mL) 500 Figure 5.13 T Z M predicted (line) and experimental (points) breakthrough curves for pure CBM9-GFP loaded onto a Perloza\u00E2\u0084\u00A2 MT100/G15 composite media 2 3 column at three different feed concentrations: C\u00C2\u00B0 = 4.77 x 10\" mol m\" (triangles), 2.2 x 10\" mol m (squares), and 8.0 x 10\" mol m\" (circles). Mobile phase loaded at a flow rate of 0.4 mL min\"1. 169 1.2 0 40 80 120 160 200 240 280 Elution Volume (mL) Figure 5.14 T Z M predicted (line) and experimental (points) breakthrough curves as a function of interstitial velocity. Pure CBM9-GFP loaded onto a Perloza\u00E2\u0084\u00A2 MT100/G15 composite media column: u = 1.7 x 10\"2 cm s\"1 (triangles), 8.5 x 10\" cms\" (circles), and 4.2 x 10\" cms\" (squares). 170 0.000 100 150 200 250 Elution Volume (mL) H O w TJ o CD a 3 to 00 o 0.0 300 Figure 5.15 Predicted (line) and experimental (points) breakthrough curves for frontal loading of a clarified cell extract onto a Perloza\u00E2\u0084\u00A2 MT100/G15 composite media column. The clarified cell extract contained 25 u M CBM9-GFP and was loaded at a superficial velocity of 8.5 x 10\" cm s\" . Eluent absorbance data at 280 nm (open squares) are also shown to indicate total protein as a function of time. 171 5.9 References Allen, T. (1997). Powder Sampling and Particle Size Measurement. Particle Size Measurement. New York: Chapman & Hall, p 528 Arnold, F. H. ; Blanch, H . W.;Wilke, C. R. (1985a). \"Analysis of affinity separations. I: Predicting the performance of affinity adsorbers.\" Chemical Engineering Journal 30(2):B9-23. Arnold, F. H. ; Blanch, H . W.jWilke, C. R. (1985b). \"Analysis of affinity separations. II: The characterization of affinity columns by pulse techniques.\" Chemical Engineering Journal 30(2):B25-36. Arve, B. H.;Liapis, A . I. (1987). \"The modeling and analysis of the elution stage of biospecific adsorption in fixed beds.\" Biotechnology and Bioengineering 30(5):638-649. Barber, P. W.;Hil l , S. C. (1990). Light scattering by particles: computational methods. Singapore World Scientific, p 261. Boraston, A . B.; Creagh, A . L. ; Alam, M . M . ; Kormos, J. M . ; Tomme, P.; Haynes, C. A. ; Warren, R. A.;Kilburn, D. G. (2001). \"Binding specificity and thermodynamics of a family 9 carbohydrate-binding module from Thermotoga maritima xylanase 10A\". Biochemistry 40(21):6240-6247. Bristow, P. A.;Knox, J. H . (1977). \"Standardization of Test Conditions for High-Performance Liquid-Chromatography Columns\". Chromatographia 10(6):279-289. Carta, G.; Ubiera, A . R.;Pabst, T. M . (2005). \"Protein mass transfer kinetics in ion exchange media: Measurements and interpretations\". Chemical Engineering & Technology 28(11):1252-1264. Crameri, A. ; Whitehorn, E. A. ; Tate, E.;Stemmer, W. P. (1996). \"Improved green fluorescent protein by molecular evolution using D N A shuffling\". Nature Biotechnology 14(3):315-319. Crank, J.;Nicolson, P. (1996). \"A practical method for numerical evaluation of solutions of partial differential equations of the heat-conduction type\". Advances in Computational Mathematics-6(3-4):207-226. Crowe, J.; Dobeli, H . ; Gentz, R.; Hochuli, E.; Stuber, D.;Henco, K . (1994). \"6xHis-Ni-N T A chromatography as a superior technique in recombinant protein expression/purification\". Methods in Molecular Biology 31:371-387. 172 Deutsch, D. G.;Mertz, E. T. (1970). \"Plasminogen: purification from human plasma by affinity chromatography\". Science 170(962): 1095-1096. Einhauer, A.;Jungbauer, A . (2001). \"The F L A G peptide, a versatile fusion tag for the purification of recombinant proteins\". Journal of Biochemical and Biophysical Methods 49(l-3):455-465. Farnan, D.; Frey, D. D.;Horvath, C. (1997). \"Intraparticle mass transfer in high-speed chromatography of proteins\". Biotechnology Progress 13(4):429-439. Farnan, D.; Frey, D. D.;Horvath, C. (2002). \"Surface and pore diffusion in macroporous and gel-filled gigaporous stationary phases for protein chromatography\". Journal of Chromatography A 959(l-2):65-73. Giddings, J. C. (1965). Dynamics of Chromatography. New York: Dekker. Goto, M . ; Hayashi, N.;Goto, S. (1983). \"Separation of Electrolyte and Non-Electrolyte by an Ion Retardation Resin\". Separation Science and Technology 18(5):475-484. Guan, K. L.;Dixon, J. E. (1991). \"Eukaryotic proteins expressed in Escherichia coli: an improved thrombin cleavage and purification procedure of fusion proteins with glutathione S-transferase\". Analytical Biochemistry 192(2):262-267. Guiochon, G. (2002). \"Preparative liquid chromatography\". Journal of Chromatography A965(l-2):129-161. Guiochon, G.; Shirazi, S. G.;Katti, A . M . (1994). Fundamentals of preparative and nonlinear chromatography. Boston: Academic Press. Hall, K . R.; Eagleton, L . C ; Acrivos, A.;Vermeulen, T. (1966). \"Pore- and solid-diffusion kinetics in fixed-bed adsorption under constant-pattern conditions\". Industrial and Engineering Chemistry Fundamentals 5(2):212-223. Haynes, H. W.;Sarma, P. N . (1973). \"A model for the application of gas chromatography to measurements of diffusion in bidisperse catalysis\". AIChE Journal 19:1043-1046. Heeter, G. A.;Liapis, A . I. (1997). \"Estimation of pore diameter for intraparticle fluid flow in bidisperse porous chromatographic particles\". Journal of Chromatography A761(l-2):35-40. Holland, C. D.;Liapis, A . I. (1983). Computer methods for solving dynamic separation problems. New York: McGraw-Hill, p xvi, 475. Hopp, T. P.; Pricket, K . S.; Price, V . L.; Libby, R. T.; March, C. J.; Ceretti, D. P.; Urdal, D. L.;Conlon, P. J. (1988). \"A short polypeptide marker sequence useful for 173 recombinant protein identification and purification\". Bio/Technology 6:1204-1210. Hubbuch, J.; Linden, T.; Knieps, E.; Thoemmes, J.;Kula, M.-R. (2003). \"Mechanism and kinetics of protein transport in chromatographic media studied by confocal laser scanning microscopy Part II. Impact on chromatographic separations\". Journal of Chromatography, A 1021 (1 -2): 105-115. Hubbuch, J.; Linden, T.; Knieps, E.; Thommes, J.;Kula, M . R. (2002). \"Dynamics of protein uptake within the adsorbent particle during packed bed chromatography\". Biotechnology and Bioengineering 80(4):359-368. Janson, J. C.;Ryden, L. (1998). Protein purification : principles, high-resolution methods, and applications. 2nd ed. New York: Wiley, p x, 695. Jervis, E. J.; Haynes, C. A.;Kilburn, D. G. (1997). \"Surface diffusion of cellulases and their isolated binding domains on cellulose\". Journal of Biological Chemistry 272(38):24016-24023. Johnston, A.;Hearn, M . T. W. (1990). \"High-performance liquid chromatography of amino acids, peptides and proteins. O i l . Mass transfer resistances in ion-exchange and dye-affinity chromatography of proteins\". Journal of Chromatography 512:101-114. Johnston, A.;Hearn, M . T. W. (1991). \"High-performance liquid chromatography of amino acids, peptides and proteins. CXIV. Protein interactions with porous coulombic sorbents: comparison of experimental findings with predictions of several adsorption models.\" Journal of Chromatography 557(l-2):335-358. Jungbauer, A . (1993). \"Preparative Chromatography of Biomolecules\". Journal of Chromatography 639(1):3-16. Jungbauer, A . (1996). \"Insights into the chromatography of proteins provided by mathematical modeling\". Current Opinion in Biotechnology 7(2):210-218. Jungbauer, A.;Hahn, R. (2004). \"Engineering protein A affinity chromatography\". Current Opinion in Drug Discovery & Development 7(2):248-256. Katoh, S.; Kambayashi, T.; Deguchi, R.;Yoshida, F. (1978). \"Performance of affinity chromatography columns.\" Biotechnology and Bioengineering 20(2):267-280. Kavoosi, M . ; Creagh, A . L. ; Kilburn, D. G.;Haynes, C. A . (2007). \"Strategy for selecting and characterizing linker peptides for CBM9-tagged fusion proteins expressed in E. coli\". Biotechnology and Bioengineering In Press. 174 Kavoosi, M . ; Meijer, J.; Kwan, E.; Creagh, A . L. ; Kilburn, D. G.;Haynes, C. A . (2004). \"Inexpensive one-step purification of polypeptides expressed in Escherichia coli as fusions with the family 9 carbohydrate-binding module of xylanase 1 OA from T-maritima\". Journal of Chromatography B 807(l):87-94. Keefe, A . D.; Wilson, D. S.; Seelig, B.;Szostak, J. W. (2001). \"One-step purification of recombinant proteins using a nanomolar-affinity streptavidin-binding peptide, the SBP-Tag\". Protein Expression and Purification 23(3):440-446. Knox, J. H. ; Laird, G. R.;Raven, P. A . (1976). \"Interaction of Radial and Axial-Dispersion in Liquid-Chromatography in Relation to Infinite Diameter Effect\". Journal of Chromatography 122(Jul7): 129-145. Ladisch, M . R. (2001). Bioseparations engineering : principles, practice, and economics. New York: Wiley. Ladisch, M . R.; Voloch, M.;Jacobson, B. (1984). \"Bioseparations: column design factors in liquid chromatography\". Biotechnology and Bioengineering Symposium 14(6th Symposium on Biotechnology for Fuels and Chemicals):525-541. L i , P.; Xiu , G.;Rodrigues, A . E. (2003). \"Modeling separation of proteins by inert core adsorbent in a batch adsorber\". Chemical Engineering Science 58(15):3361-3371. Linden, T.; Ljunglof, A . ; Hagel, L. ; Kula, M . R.;Thommes, J. (2002). \"Visualizing patterns of protein uptake to porous media using confocal scanning laser microscopy.\" Separation Science and Technology 37(1): 1-32. Ljunglof, A.;Hjorth, R. (1996). \"Confocal microscopy as a tool for studying protein adsorption to chromatographic matrices\". Journal of Chromatography A 743(l):75-83. Lowe, C. R.; Lowe, A . R.;Gupta, G. (2001). \"New developments in affinity chromatography with potential application in the production of biopharmaceuticals\". Journal of Biochemical and Biophysical Methods 49(1-3):561-574. Ma, Z.; Whitley, R. D.;Wang, N . H. L. (1996). \"Pore and surface diffusion in multicomponent adsorption and liquid chromatography systems\". AIChE Journal 42(5): 1244-1262. Mach, H. ; Middaugh, C. R.;Lewis, R. V . (1992). \"Statistical determination of the average values of the extinction coefficients of tryptophan and tyrosine in native proteins\". Analytical Biochemistry 200(1 ):74-80. Mao, Q. M . ; Johnston, A. ; Prince, I. G.;Hearn, M . T. W. (1991). \"High-performance liquid chromatography of amino acids, peptides and proteins. CXIII. Predicting 175 the performance of non-porous particles in affinity chromatography of proteins.\" Journal of Chromatography 548(1-2):147-163. Mao, Q. M . ; Prince, I. G.;Hearn, M . T. W. (1995). \"High-Performance Liquid-Chromatography of Amino-Acids, Peptides and Proteins .139. Impact of Operating Parameters in Large-Scale Chromatography of Proteins\". Journal of Chromatography A 691(l-2):273-283. Martin, C ; Iberer, G.; Ubiera, A.;Carta, G. (2005). \"Two-component protein adsorption kinetics in porous ion exchange media\". Journal of Chromatography A 1079(1-2):105-115. Mulcahy, P.; O'Flaherty, M . ; Jennings, L.;Griffin, T. (2002). \"Application of kinetic-based biospecific affinity chromatographic systems to ATP-dependent enzymes: studies with yeast hexokinase\". Analytical Biochemistry 309(2):279-292. Patankar, S. V . (1980). Numerical heat transfer and fluid flow. New York: McGraw-Hill, p xii i , 197. Patwardhan, A . V. ; Ataai, M . M.;Zenouzi, M . (1995). \"A mathematical model of the immobilized metal affinity chromatography for protein separations: simulation of loading and elution conditions.\" American Society of Mechanical Engineers, Heat Transfer Division 322:61-68. Pinto, N . G.;Graham, E. E. (1987). \"Application of the shrinking-core model for predicting protein adsorption\". Reactive Polymers, Ion Exchangers, Sorbents 5(l):49-53. Porath, J.; Carlsson, J.; Olsson, I.;Belfrage, G. (1975). \"Metal chelate affinity chromatography, a new approach to protein fractionation\". Nature 258(5536):598-599. Pritzker, M . D. (2003). \"Model for parallel surface and pore diffusion of an adsorbate in a spherical adsorbent particle\". Chemical Engineering Science 58(2):473-478. Shimomura, O.; Johnson, F. H.;Saiga, Y . (1962). \"Extraction, purification and properties of aequorin, a bioluminescent protein from the luminous hydromedusan, Aequorea\". Journal of Cellular and Comparative Physiology 59:223-239. Sirotti, D. A.;Emery, A . (1983). \"Mass transfer parameters in an immobilized glucoamylase column by pulse response analysis.\" Biotechnology and Bioengineering 25(7): 1773-1779. Skidmore, G. L.; Horstmann, B. J.;Chase, H. A . (1990). \"Modeling Single-Component Protein Adsorption to the Cation Exchanger S Sepharose F f . Journal of Chromatography 498(1): 113-128. 176 Smith, D. B.;Johhson, K. S. (1988). \"Single-step purification of polypeptides expressed in Escherichia coli as fusions with glutathione S-transferase\". Gene 67(l):31-40. Stofko-Hahn, R. E.; Carr, D. W.;Scott, J. D. (1992). \"A single step purification for recombinant proteins. Characterization of a microtubule associated protein (MAP 2) fragment which associates with the type II cAMP-dependent protein kinase\". FEBS Letters 302(3):274-278. Vaillancourt, P.; Simcox, T. G.;Zheng, C. F. (1997). \"Recovery of polypeptides cleaved from purified calmodulin-binding peptide fusion proteins\". Biotechniques 22(3):451-453. Wakao, N . ; Oshima, T.;Yagi, S. (1958). \"Mass transfer from packed beds of particles to a fluid\". Journal of Chemical Engineering of Japan 22:780-785. Weisz, P. B.;Goodwin, R. D. (1963). \"Combustion of carbonaceous deposits within porous catalyst particles. I. Diffusion-controlled kinetics\". Journal of Catalysis 2(5):397-404. Wilson, D. S.; Keefe, A . D.;Szostak, J. W. (2001). \"The use of mRNA display to select high-affinity protein-binding peptides\". Proceedings of the National Academy of Sciences of the United States of America 98(7):3750-3755. Wilson, E. J.;Geankoplis, C. J. (1966). \" Liquid mass transfer at very low Reynolds numbers in packed beds.\" Industrial and Engineering Chemistry Fundamentals 5(1):9-14. Winterhalter, C ; Heinrich, P.; Candussio, A. ; Wich, G.;Liebl, W. (1995). \"Identification of a novel cellulose-binding domain within the multidomain 120 kDa xylanase XynA of the hyperthermophilic bacterium Thermotoga maritima\". Molecular Microbiology 15(3):431-444. Young, M . E.; Carroad, P. A.;Bell , R. L. (1980). \"Estimation of Diffusion-Coefficients of Proteins\". Biotechnology and Bioengineering 22(5):947-955. Zheng, C. F.; Simcox, T.; Xu , L.;Vaillancourt, P. (1997). \"A new expression vector for high level protein production, one step purification and direct isotopic labeling of calmodulin-binding peptide fusion proteins\". Gene 186(l):55-60. 177 6 A Mechanically Stable Porous Cellulose Media for Affinity Purification of CBM9-Tagged Fusion Proteins 6.1 Introduction Recombinant D N A technology and affinity chromatography have advanced the field of protein purification such that many proteins and peptides can now be purified without extensive a priori knowledge of the physicochemical properties of the desired product. By designing a chimeric protein incorporating a known affinity tag, the protein of interest can be purified from a culture supernatant in a single, highly selective capture step (Ford et al. 1991; Lowe et al. 2001). Affinity tags have been used extensively at the laboratory scale (Hearn and Acosta 2001; Nilsson et al. 1997), yet their application at the preparative scale has been limited, due in large part to the cost and stability (both chemical and mechanical) of the associated affinity matrices (Terpe 2003). For example, IMAC technology, based on the (His)6 affinity tag, relies on a stationary phase bearing iminodiacetate or imidazole groups to which a transition metal ion may complex. The complexed metal ions act as the coordination sites for selective binding of (His)6-tagged proteins (Porath et al. 1975; Rulisek and Havlas 2000). Although a popular and effective technology for laboratory-scale applications, the cost associated with the synthesis of the stationary phase along with concerns related to leaching of the bound metal ion into the eluent make this technology less attractive at preparative scales. Other popular affinity-tag systems such as the glutathione S-transferase tag (Guan and Dixon 1991; Smith and Johnson 1988) and the F L A G tag (Einhauer and Jungbauer 2001; Hopp 1988) likewise * A version of this chapter has been accepted for publication in the Journal of Chromatography A. [Reference: Mojgan Kavoosi, Dexter Lam, Jenny Bryan, Douglas G. Kilburn, Charles A . Haynes, A mechanically stable porous cellulose media for affinity purification of CBM9-tagged fusion proteins. J. Chromatography A. In Press] 178 exhibit properties that limit their scalability to preparative applications (Linhult et al. 2005). The successful scale-up of tag-based affinity chromatography would therefore benefit from an affinity tag that provides efficient capture and elution on an inexpensive, yet chemically and mechanically robust stationary phase. Kavoosi et al. (Kavoosi et al. 2004) recently introduced novel vectors for high-level production in E. coli of soluble chimeric proteins comprised of a target protein or peptide fused to CBM9, the family 9 carbohydrate binding module (CBM9) of xylanase 10A of T. maritima. CBM9 strongly binds both insoluble cellulose and soluble polysaccharides (Boraston et al. 2001), allowing for the specific capture of CBM9-tagged fusion proteins on a cellulose-based capture column and quantitative elution of the purified protein using an inexpensive soluble sugar such as glucose. The CBM9-cellulose affinity cassette is potentially attractive for preparative-scale applications because cellulose is inexpensive, chemically inert, and can be machined into well defined geometries offering high specific surface area (Rodriguez et al. 2004). Cellulose has also been approved for many pharmaceutical and human uses and is unlikely to introduce any harmful agents into the purified product (Hoenich et al. 1997; Varela et al. 2001). Cellulose beads have therefore been widely utilized as packing materials in chromatography, including serving as the base chemistry for a wide range of size-exclusion and ion-exchange media (Motozato and Hirayama 1984). Kuga (Kuga 1984) has reviewed preparation methods of cellulose beads and their separation properties. CBM9 binds directly to the reducing ends of cellulose chains (Boraston et al. 2001), permitting most commercially available cellulose-based stationary phases to serve as the affinity media without the need for additional chemical processing. As an example, the porous cellulosic media Perloza\u00E2\u0084\u00A2 MT100 (henceforth referred to as MT100) binds CBM9 strongly (Ka ~ 106 M\" 1), with very high capacity (^,max~ 500 mg protein/g dry media), and without significant fouling, so that repeated purification cycles from a clarified bacterial cell extract consistently result in high product purities and yields (Kavoosi et al. 2004). However, a limitation of highly porous cellulose-based matrixes such as MT100 is the tendency for bed compression at elevated flow rates or at larger bed dimensions (Colby et al. 1996b; Ostergren et al. 1998). Conventional cellulose 179 beads, including MT100, are characterized by relatively large particle-size distributions, resulting in an additional flow resistance in columns packed with these media that is associated with smaller particles occupying void spaces among larger ones. Pure cellulose media is also susceptible to compressive deformation, which can not only increase flow instabilities within the packed bed, reducing dynamic capacity and column performance, but can also significantly increase band broadening in extreme cases (Colby et al. 1996a; Mohammad et al. 1992). Scale-up of columns utilizing these media therefore requires a strategy to mechanically stabilize the packed bed. Sephadex and other popular hydrophilic gel-like chromatographic media typically achieve mechanical strength through interchain cross-linking (Hjerten et al. 1987; Laas 1976; Leonard 1997). In contrast, the mechanical stability of Perloza\u00E2\u0084\u00A2 MT100, a highly porous, hydrophilic media derived from regenerated cellulose, is only from interchain hydrogen bonds and dispersion forces between disordered and partially crystalline regions of its regenerated cellulose architecture (Iontosorb, CZ). This suggests that the mechanical stability of MT100 may be improved through cross-linking using an appropriate epoxide or halohydrin (Narayanan and Crane 1990). A particularly promising cross-linking agent for this purpose is epichlorohydrin, which yields hydrophilic linkages, is industrially proven, and is readily available and inexpensive. The resulting epoxide bond is stable at high pH, permitting stringent clean-in-place protocols using caustic. Epichlorohydrin is therefore the preferred crosslinking agent in the preparation of Sephadex (Holmberg et al. 1995) and its reaction chemistry is compatible with cellulose. Ruckenstein and Guo (Guo and Ruckenstein 2001), for example, have shown that cross-linking with epichlorohydrin can be used to increase the mechanical stability of cellulose-based filter paper. This paper focuses on the use of response surface methodology (RSM) to identify optimal reaction conditions for synthesis of a mechanically stable, porous cellulose affinity chromatography media for use with CBM9 fusion-tag technology. Intraparticle cross-links are introduced into MT100 using epichlorohydrin-based epoxide chemistry with the concentrations of epichlorohydrin and dimethyl sulfoxide (DMSO) representing the two input variables. Treated and untreated media are compared both by monitoring 180 pressure drop across the column and by identifying the critical superficial velocity w c ri t for column compression, with the latter being used as the response variable. A mechanically enhanced, cross-linked media is identified and moment analysis, equilibrium binding isotherms and HETP analysis are used to compare the binding and transport properties of this mechanically stabilized media against untreated MT100. The application of this cross-linked media to the affinity purification of a recombinant protein from E. coli is also demonstrated. 6.2 Materials and Methods 6.2.1 Reagents Epichlorohydrin, glucose and all other reagents were analytical grade and purchased from Sigma-Aldrich (Mississauga, ON, Canada) unless otherwise stated. Perloza\u00E2\u0084\u00A2 MT100 chromatographic media with a nominal particle diameter distribution of 50 to 80 urn was purchased from Iontosorb Inc. (Czech Republic). E. coli BL21 (DE3) was obtained from Novagen (Madison, WI, USA). 6.2.2 Cross-Linking of MT100 Cross-linking reactions were carried out in a 200 mL round-bottom flask immersed in a water-filled jacket. A PolyScience 910 constant-temperature water pump (Preston Ind, USA) circulated heated water through the jacket, maintaining the reaction temperature at 50\u00C2\u00B0C. The reaction flask was charged with a magnetic stir bar and appropriate volumes (mL per gram MT1Q0 cellulose) of 5.0 M sodium hydroxide, epichlorohydrin and DMSO were added. The reagents were allowed to mix and thermally equilibrate before the addition of an aliquot of MT100 slurry equivalent to 2 g of dry media. The resulting mixture was well stirred throughout the reaction, which was allowed to proceed for 2 hours before being quenched with acid addition to neutralize pH and immediate immersion in an ice bath. The cross-linked media was then filtered and repeatedly washed with 200 mL of water before being degassed as a dilute slurry. The cross-linking reactions were continuously mixed to prevent formation of interparticle linkages and to ensure that all cross-linking occurred inside the resin 181 particle. Lack of interparticle cross-linking was verified by optical microscope imaging of the reaction products. 6.2.3 Hydrodynamic Characterization MT100 or cross-linked MT100 media were packed into a Pharmacia HR10 column (I.D. = 1.0 cm) at an elevated flow rate using standard inclined pouring techniques. A silica filter pad was positioned on top of the bed and the flow distributor lowered and secured to achieve finger-pressure compression of the bed. The final bed heights ranged between 8.5 and 10.6 cm. The flow curve for each column was measured by pumping buffer (50 mM potassium phosphate, 100 m M NaCl, pH 7.0) through the column at specified superficial velocities and measuring the pressure drop at steady state. Compression of the stationary phase, measured as the separation of the bed from the flow distributor was also monitored at each flow rate by recording time-lapse images of the top surface of the bed using a Logitech Quickcam Pro 4000 web camera. 6.2.4 Measurement of Binding Isotherms Equilibrium adsorption isotherms for binding of CBM9-GFP to MT100 and each cross-linked MT100 preparation were performed at 22\u00C2\u00B0C in low salt buffer (50 m M potassium phosphate, 100 m M NaCl, pH 7.0). Purified protein at concentrations ranging from 1 to 30 uM was mixed with media (1 mg dry weight) in low salt buffer to a final volume of 1 ml. Samples were then incubated overnight at 22\u00C2\u00B0C while mixing end-over-end. The supernatant was collected after centrifugation at 27,000 x g for 20 min at 22\u00C2\u00B0C and the concentration of unbound protein was determined by U V absorbance (280 nm) using a Cary 100 Spectrophotometer (Varian, Palo Alto, CA) and a calculated molar extinction coefficient of 62870 M \" 1 cm\"1 (Mach et al. 1992). The above protocol was repeated for binding of ovalbumin to identify any non-specific binding that may arise due to the cross-linking reaction. Langmuir-type binding parameters were determined by non-linear regression of the Langmuir isotherm equation to the experimental data using GraphPad Prism 3.0 software. 182 6.2.5 Affinity Purification of CBM9-GFP An overnight 5 mL culture of E. coli BL21 (DE3) cells carrying the pET28-CBM9-GFP vector (Kavoosi et al. 2004) was used to inoculate 500 mL of Laural Broth (LB). Cells were then grown at 37\u00C2\u00B0C to an ODeoo nm of about 0.8 and protein production was induced by the addition of 0.1 m M isopropyl-l-thio-(3-D-galactoside (IPTG). The culture was further incubated at 30\u00C2\u00B0C for another 10-12 hours after which the cells were harvested by centrifugation for 20 min at 8,500 x g and 4\u00C2\u00B0C. The cells were resuspended in high salt buffer (50 m M potassium phosphate, 1 M NaCl, pH 7.0) and ruptured by two passes through a French pressure cell (21,000 lb/in ). The cell debris was removed by centrifugation at 27,000 x g for 30 min and 4\u00C2\u00B0C. The clarified cell extract was frontal loaded onto an XK26 column (GE Healthcare) packed with MT100 or a cross-linked MT100 preparation. Unbound proteins were washed through the column using 210 mL of high salt buffer followed by 150 mL of low salt buffer. Bound CBM9-GFP was isocratically eluted with 100 mL of 1 M glucose in low salt buffer; 10 mL fractions were collected throughout the process. The chromatogram for the purification process was detected using in-line U V absorbance at 280 nm and off-line measurement of the fluorescence intensity (395 nm excitation of GFP, 509 nm emission) of each collected fraction using a Cary Eclipse fluorescence spectrophotometer (Varian). Process yields were determined by comparing the fluorescence intensity of appropriately diluted elution fractions to that of the original cell lysate. The purity of the combined product peak was analysed by 12% SDS-PAGE. 6.3 Results and Discussion 6.3.1 Response Surface Methodology to Improve Mechanical Properties Both flow curve and bed-compression visualization data were used to evaluate the mechanical stability of unmodified MT100 and of each chemically modified cellulose stationary phase. As illustrated in Figure 6.1, significant compression of a packed bed of unmodified Perloza\u00E2\u0084\u00A2 MT100 is observed at all superficial velocities above 0.64 x 10\"4 m 183 s'1, indicating the need for mechanical stabilization of this media prior to column scale-up to permit acceptable throughput without column compression. The epichlorohydrin-based cross-linking chemistry proposed to stabilize the mechanical properties of MT100 against pneumatic compression is described in Figure 6.2. Cross-linking reactions involving epoxides such as epichlorohydrin are typically carried out under strong caustic conditions to deprotonate all solvent-exposed hydroxyl groups and increase their reactivity. Previous studies have shown that the reaction product is relatively insensitive to NaOH provided the concentration of NaOH is sufficient to bring the reaction pH above 12 (Guo and Ruckenstein 2001). A polar organic co-solvent such as methanol, or in this case DMSO, is generally used to swell the matrix and improve accessibility of reactive hydroxides to the cross-linking reagent, and the mechanical stability of the reaction product is therefore likely sensitive to the concentration of the co-solvent ([DMSO]) used. The concentration of epichlorohydrin ([Epi]) is also expected to be a determinant of the characteristics of the reaction product. We investigated the dependence of the mechanical stability (response variable Went, detailed below) of epichlorohydrin cross-linked media on the MTlOO-normalized concentration of epichlorohydrin (input variable in mL Epi/g MT100) and DMSO (input variable \u00C2\u00A3 2 in mL DMSO/g MT100) used in the formulation reaction. For each column prepared at a given , \u00C2\u00A3 2 condition, pressure drop was measured as a function of superficial velocity u to determine wcrj t, the value of u at which the flow curve deviates from linearity, which was used as our response variable quantifying mechanical stability. In all Cases, Wcrit differed by less than 15% from the linear velocity at which incipient bed collapse was observed, suggesting that the onset of flow curve non-ideality is directly related to bed compression. A general response model for this system can be written as: u c n l ^ y ^ f ( ^ 2 ) + ^ (6.1) where ^represents random variation (e.g., measurement error or background noise) and is assumed to have a normal distribution with mean zero and variance o 2. To identify a 184 range of , \u00C2\u00A3 2 values for which model (6.1) is applicable, we first determined a set of , \u00C2\u00A3 2 pairs that yielded a well-defined reaction product and for which a value of wCrit could be measured reproducibly. The resulting values suggested an input variable space of 0,0 < \u00C2\u00A3,,\u00C2\u00A3 2 - 12,12 mL/g, which is consistent with the range of epichlorohydrin and DMSO concentrations previously used by Ruckenstein and Guo (Guo and Ruckenstein 2001) to optimize cross-linking of cellulose filters. As the true response function / was unknown, we first performed a nonparametric exploration of the observed response values. To obtain a flexible and smooth representation of the response surface, we used locally weighted polynomial regression, in which the fitted response surface for a target point results from a weighted, least squares fit of a polynomial to the data points in a neighborhood of the target (Cleveland and Devlin 1988). The resulting estimate of the response surface / depicted in Figure 6.3, showed strong evidence for substantial curvature; specifically, a mound-shaped response surface was suggested by the data. This fit implies that the mechanical stability w c ri t is maximized when [epichlorohydrin] and [DMSO] are 4.08 and 8.2 mL/g respectively, which is consistent with maximum observed in our experiment at input concentrations of \u00C2\u00A3,= 2.78 mL/g and \u00C2\u00A3 2 = 8.33 mL/g. The plots of the response surface also suggest that mechanical stability at or near this maximum is only achieved in a relatively small set of DMSO and epichlorohydrin concentration pairings. For the purposes of conducting inference and interpreting model parameters, it was also desirable to pose a more structured, global model. Interestingly, the standard second-order response surface model can be seen as simply a global implementation of the polynomial fits employed above (Box et al. 2005): \u00C2\u00AB\u00C2\u00AB* = /fc>&) = A + +Mx + Aitf +P*\u00C2\u00A3 +012^2 +v (6-2) where fi\ and f5\ \ are the linear and quadratic regression coefficients for the main effect of [Epi] on M C r i t , and fin is the regression coefficient for linear interaction effect between and \u00C2\u00A3 2 . The results of fitting model (6.2) are presented in Table 6.2 and Figure 6.4, and the interaction effect /?12 is seen to be highly insignificant (p = 0.8664 \u00C2\u00BB 0.05), whereas all other non-intercept terms are highly significant (maximum p = 0.0134 < 0.05). The 185 results of fitting a version of model (6.2) without the interaction term are also presented in Table 6.2 and we find that both the estimates and the strong statistical significance of the remaining non-intercept terms are essentially unchanged. The simplified model therefore permits a direct and unambiguous analysis of the dependence of wcrjt on the input variables and \u00C2\u00A3 2 , each of which was found to influence the mechanical stability. At low DMSO concentrations, the observed improvement in ucrn with increasing \u00C2\u00A3 2 is consistent with the known solvating effects of DMSO. The parabolic dependence of \u00C2\u00AB c rj t on \u00C2\u00A3 2 is then observed because DMSO concentrations greater than 8.33 mL/g result in significant swelling of the particles, weakening interchain hydrogen bonds between disordered and partially crystalline regions of the regenerated cellulose architecture of MT100 even in the presence of a large concentration of crosslinking agent. Similarly, excessive crosslinking of the media at high epichlorohydrin loads (i.e., > 3 mL/g) results in significant swelling and mechanical breakdown of the media, presumably through the high osmotic pressures and new interchain forces created through the crosslinking reaction. Thus, the experimental and R S M results suggest that chemical crosslinking can improve the mechanical stability of MT100 only under conditions where the reaction does not disrupt the interchain hydrogen-bonding forces that naturally provide mechanical stability to the unmodified MT100 media. Finally, a two-fold change in the volume of 5.0 M NaOH added was also applied at two randomly chosen reaction conditions (\u00C2\u00A3,= 0.56 mL/g,\u00C2\u00A3 2 = 8.33 mL/g; and 11.11 mL/g,\u00C2\u00A3 2 = 11.11 mL/g) to verify, as reported by Ruckenstein and Guo (Guo and Ruckenstein 2001), that the NaOH concentration does not influence the reaction product provided the concentration of NaOH is sufficient to bring the reaction pH above 12. As quantified by \u00C2\u00AB c rj t , the mechanical stability of C R L 100-7 (wcrjt = 6.0(\u00C2\u00B10.15) x 10\"4 m s\"1) is ca. an order of magnitude greater than that of untreated Perloza\u00E2\u0084\u00A2 MT100 (wCrit = 0.64 x 10\"4 m s\"1). This is verified in Figure 6.5, which compares the measured flow curve for a packed bed of each media to show the significant extension of the desired linear response in pressure drop to increases in flow. As a result, mechanically permissible linear velocities within the CRL100-7 column extend deep onto the linear region of the van Deemter plot (as will be shown later in this manuscript), allowing 186 column operating conditions to be selected based on the traditional trade-off between throughput and mass-transfer related reductions in column efficiency. Thus, cross-linked media CRL100-7 was selected and further characterized as a potentially scalable matrix for affinity capture and purification of CBM9-tagged fusion proteins. 6.3.2 Comparison of Sorption Properties Equilibrium adsorption studies were performed to determine the influence of cross-linking on the binding properties of the media. Adsorption isotherms for CBM9-GFP binding to untreated MT100 and to the cross-linked media are shown in Figure 6.6. Both data sets are well described by the Langmuir equation, which shows that the epichlorohydrin cross-linking reaction does not alter binding affinity,, but reduces static capacity by ca. 25% (Table 6.3). The epichlorohydrin cross-linking chemistry is expected to target free hydroxyls (Figure 6.2), including the reducing ends of solvent-exposed cellulose chains within the media, leading to the observed decrease in the ligand density for CBM9 binding. Nevertheless, the static capacity, which equates to 370 mg CBM9-GFP bound per g cross-linked media, remains significantly higher than that reported for most commercially available media used for fusion-tag based affinity chromatography (Lichty et al. 2005). Finally, negligible binding of ovalbumin was observed (Figure 6.6), indicating that cross-linking of the media produced no chemical alterations that enhance non-specific binding of proteins. 6.3.3 Comparison of Column Parameters and Properties Using methods described previously (Rodriguez et al. 2004), moments analysis (Kubin 1965; Kucera 1965) was combined with van Deemter theory (Van Deemter et al. 1956) to determine parameters characterizing the geometric properties of columns packed with CRL100-7 media and the mass-transfer behavior of CBM9-GFP within those columns. Experiments were conducted under nonbinding conditions with the column length to diameter (L/dc) ratio maintained above 5 to minimize end effects (Jungbauer 1993). A l l wall effects were ignored since the ratio of dc to particle diameter (dp) was well above that needed to safely ignore column wall contributions to band broadening (Knox etal. 1976). 187 Measured parameters and properties are listed in Table 6.4 for both an unmodified MT100 column and our C R L 100-7 cross-linked media column. A peak asymmetry factor (As) close to unity and a reduced plate height (h) close to 2 are both good indicators (Bristow and Knox 1977) that the cross-linked media packs uniformly. The first moment Oi) for CBM9-GFP elution (nonbinding conditions) from the C R L 100-7 column shows a linear dependence on solute residence time (Figure 6.7) that, together with a regressed e value lower than measured for the untreated MT100 column, indicates the more mechanically stable cross-linked media packs somewhat more efficiently than the compressible media. A decrease in stationary-phase porosity sp with protein molecular weight is expected for media with pore diameters within the size range of the partitioning proteins. Stationary-phase porosities were therefore determined with respect to partitioning of CBM9-GFP, a 53 kDa fusion protein. Cross-linking decreases \u00C2\u00A3p for this protein probe from its original value of 0.81 for untreated MT100 to 0.67, indicating a significant change in intraparticle pore architecture that is consistent with the 25% drop in static binding capacity reported in Table 6.3. Such changes in stationary-phase architecture might also alter solute mass transport such that gains in mechanical stability achieved through the cross-linking reaction occur at the expense of overall column performance. Second moment (a2) data for CBM9-GFP elution peaks were therefore determined as a function of interstitial velocity u0 and combined with the moments theory of Haynes and Sarma (Haynes and Sarma 1973) for operation under nonbinding conditions 2/4 L \\-ej 1 + rp , - rp 3kf \5\u00C2\u00A3pDpj (6.3) to obtain values for the parameters characterizing mass transfer of the CBM9-GFP fusion protein within a column packed with cross-linked MT100 media. The second moment data are plotted in Figure 6.8 according to equation 6.3, where D L is the axial dispersion coefficient, and kf and Dp are the film mass-transfer coefficient and pore diffusivity, respectively. Standard correlations for the bulk molecular diffusivity D M (Young et al. 1980) and kf (Wilson and Geankoplis 1966) therefore permit estimation of D L and Dp 188 through regression of equation 6.3. The subsequent calculation of the column Peclet (Pe) and Biot (Bi) numbers then provides a sound basis for determining i f the cross-linking reaction has negatively impacted solute mass transport within the column. In a previous study (Kavoosi et al. 2007b), we determined that uptake of CBM9-tagged proteins within a MT100 column is limited by the rate of solute diffusion within the pores of the stationary phase, with axial dispersion providing no significant contribution to elution band broadening. As shown in Table 6.4, cross-linking of the MT100 media changes this result very little. Axial dispersion remains insignificant, mass transfer across the fluid film is enhanced by the small reduction in column voidage, and the pore diffusivity decreases only slightly compared to the column utilizing untreated MT100. As a result, both Pe and Bi are largely unchanged by the cross-linking reaction, indicating that rates of solute mass transfer and overall mass transfer behavior are quite similar in the two columns. Finally, as noted previously, the van Deemter plot (Figure 6.9) characterizing CBM9-GFP transport within the CRL 100-7 column confirms that this cross-linked media can be reliably operated far into the flow regime where reduced plate height is controlled by solute mass transfer kinetics within the stationary phase. 6.3.4 Preparative Scale Affinity Purification of CBM9-GFP on CRL100-7. Figure 6.10 shows the chromatogram for the purification of CBM9-GFP from clarified E. coli BL21 cell lysate on a 60 mL CRL 100-7 column, which represents an 8-fold column scale-up by volume. No protease inhibitors were added to the cell extract and the purification was performed at room temperature (21\u00C2\u00B0C). Although u in this separation is ca. three times larger than wcrjt for untreated Perloza\u00E2\u0084\u00A2 MT100, the purification on the C R L 100-7 column proceeded in a similar manner to that on the MT100 column (Kavoosi et al. 2004) when both chromatograms are reported in terms of elution volume. Initial breakthrough of contaminating proteins was observed after ca. one column void volume. The small fluorescence peak observed within the breakthrough profile for contaminating proteins indicates the presence in the clarified cell lysate of free GFP, released as a result of a small amount of degradation within the linker region 189 connecting CBM9 to GFP (Kavoosi et al. 2007a). Two wash steps were used to effectively remove most of the contaminating proteins present in the column. As evident by the overlapping A280 and fluorescent signals, addition of 1 M glucose to the mobile phase elutes CBM9-GFP as a single sharp peak. The product yield obtained in the pooled elution peak for the first cycle was 75 \u00C2\u00B1 2 % (Table 6.5). Most of the lost product can be attributed to CBM9-GFP degradation products in the feed. Nevertheless, the yield from the initial run on the cross-linked column is lower than the 86% yield previously recorded for the same separation performed on an untreated MT100 column at a 3.6-fold lower linear velocity of 0.41 x l C W . SDS-PAGE gel documentation of the scaled-up purification (Figure 6.11) shows quantitative flowthrough (Lane 2) of contaminating proteins with minimal presence of CBM9-GFP, resulting in elution of the 53 kDa CBM9-GFP product (Lane 5) characterized by a purification factor greater than 95%. The weak 26 kDa elution band in Lane 5 corresponds to CBM9 affinity tag released by proteolytic degradation of the fusion protein within the linker region. As the purification was performed at 21\u00C2\u00B0C with no protease inhibitors, activity from endogeneous proteases was likely responsible for the observed product degradation. Higher product yields may therefore be realized through application of a suitable protease inhibitor. 6.3.5 Column Reusability The ability of CRL100-7 to provide acceptable and predictable purification performance with repeated column use was also investigated. Three consecutive purifications were performed at a superficial velocity of 1.5 x 10\"4 m s\"1 to identify any changes in column performance with increasing number of purification cycles. In all three purification cycles, very high product purity (>95%) was observed. Interestingly, product yield increased after the first column use (Table 6.5) to a value comparable to that observed for the untreated MT100 column. This suggests that the cross-linking reaction introduces a small number of sites that exhibit irreversible binding to CBM9-190 GFP. These sites are blocked in the first cycle, leading to the higher observed yields in the second and subsequent cycles. 6.4 Conclusions The mechanical stability of porous Perloza\u00E2\u0084\u00A2 MT100 cellulosic matrix was enhanced through an R S M optimized cross-linking reaction using epichlorohydrin to create a chromatographic media for affinity purification of recombinant CBM9-tagged fusion proteins. The chemistry used to cross-link MT100 is simple to conduct and inexpensive, yielding a stable affinity media that is ca. 1750th - l/100 t h the cost of popular affinity media currently used for industrial processing. The mechanically enhanced CRL100-7 media can withstand mobile-phase superficial velocities up to 6 x 10\"4 m/s, representing an order of magnitude improvement in throughput over unmodified MT100. The cross-linking reaction alters pore architecture, reducing the porosity of the stationary phase by 20% and thereby lowering binding capacity by ca: 25%. Despite this loss in capacity, CRL100-7 is capable of binding 7.1 umol CBM9-GFP/g media (376 mg/g), a value superior to most commercial affinity-tag systems. The mechanical stability of the cross-linked media is maintained during (at least) modest column scale-up. Consecutive purifications on a scaled-up column showed yields equal to or greater than 82%, demonstrating reliable column performance from cycle to cycle. 191 6.5 Tables Table 6.1 Cross-linked resins and reaction conditions Resin Epichlorohydrin (ml/g cellulose) D M S O (ml/g cellulose) Merit (m/s) MT100 0 0 0.64 (\u00C2\u00B10.10) x 10\"4 CRLlOO-1 0.28 0 1.3 (\u00C2\u00B10.20) x lO\" 4 CRL 100-2 2.78 0 1.1 (\u00C2\u00B10.17) x 10\"4 CRL100-3 11.11 0 1.1 (\u00C2\u00B10.17) x 10\"4 CRL 100-4 0 8.33 0.64 (\u00C2\u00B10.10) x 10\"4 CRL100-5 0.28 8.33 1.2 (\u00C2\u00B10.18) x 10\"4 C R L 100-6 0.56 8.33 1.5 (\u00C2\u00B10.23) x 10\"4 CRL100-7 2.78 8.33 6.0 (\u00C2\u00B10.90) x 10\"4 CRL100-8 0 11.11 0.74 (\u00C2\u00B10.11) x 10\"4 CRL100-9 0.28 11.11 1.1 (\u00C2\u00B10.17) x 10\"4 CRL100-10* 11.11 11.11 1.2 (\u00C2\u00B10.18) x 10\"4 *The reaction for CRL100-10 had 2.8 ml/g of cellulose of 5 M NaOH. A l l other reactions had 1.4 ml/g of cellulose of 5 M NaOH. 192 Table 6.2 Results of fitting quadratic response surface models to the data set reported in Table 6.1 Regression Equation \u00C2\u00AB** =A>+ PA + + Pu?x + Putt + + * C o e f f i c i e n t E s t i m a t e S t a n d a r d E r r o r f - s t a t i s t i c p - v a l u e 0.548 0.425 1.289 0.205 Pi 0.710 0.163 4.353 9.79 x 10\"5 Pi 0.432 0.140 3.094 0.0037 Pn -0.068 0.013 -5.085 1.02 x 10\"5 Pn -0.033 0.013 -2.595 0.0134 Pn 0.002 0.010 0.169 0.8664 Residual standard error: 1.079 on 38 degrees of freedom Multiple R-Squared: 0.5281 Adjusted R-squared: 0.466 F-statistic: 8.506 on 5 and 38 DF p-value: 1.784 x 10\"5 Regression Equation C o e f f i c i e n t E s t i m a t e S t a n d a r d E r r o r / - s t a t i s t i c p - v a l u e Po 0.512 0.363 1.410 0.167 P\ 0.720 0.148 4.851 2.00 x 10\"5 Pi 0.436 0.135 3.223 0.0026 Pn -0.068 0.013 -5.152 7.75 x 10\"6 P22 -0.033 0.013 -2.622 0.0124 Residual standard error: 1.066 on 39 degrees of freedom Multiple R-Squared: 0.5278 Adjusted R-squared: 0.4793 F-statistic: 10.9 on 4 and 39 DF p-value: 4.997 x 10\"6 193 Table 6.3 Equilibrium adsorption parameters for MT100 and CRL100-7 affinity media determined from two independent experiments Resin Affinity Capacity (pmol/g resin) MT100 CRL100-7 0.71 (\u00C2\u00B10.05) x 106 1.04 (\u00C2\u00B10.21) x 106 9.53 (\u00C2\u00B10.23) 7.10 (\u00C2\u00B10.33) * reported errors represent \u00C2\u00B12a (i.e. 95% confidence interval) 194 Table 6.4 Column properties and parameters Parameter MT100 CRL100-7 Units L 9.3 11.4 cm h 2.3 (\u00C2\u00B1 0.02) 2.6 (\u00C2\u00B1 0.03) A, 1.1 (\u00C2\u00B10.03) 1.01 (\u00C2\u00B10.09) s 0.53 (\u00C2\u00B1 0.07) 0.49 (\u00C2\u00B10.01) ep 0.81 (\u00C2\u00B10.02) 0.67 (\u00C2\u00B1 0.01) DL 2.9 (\u00C2\u00B1 1.8 ) x 10\"9 5.0 (\u00C2\u00B1 1.4) x 10\"8 m2/s kf* 7.2 (\u00C2\u00B1 0.13) x 10'6 7.8 (\u00C2\u00B10.23)x 10\"6 m/s DM 6.54 x 10\"11 6.54 x 10~\" m2/s Dp* 4.5 (\u00C2\u00B10.21)x 10\"12 1.7 (\u00C2\u00B1 0.07) x 10\"12 m2/s Pe* 613 (\u00C2\u00B129) 1668 (\u00C2\u00B169) Bi* 17 (\u00C2\u00B10.8) 51 (\u00C2\u00B12.3) * Properties determined for superficial velocity of 4.25 x 10\"5 m/s at 294 K; two independent columns were packed to determine the error in the packing procedure. * reported errors represent \u00C2\u00B12o (i.e. 95% confidence interval) 195 Table 6.5 Yield and purity for consecutive purification cycles of CBM9-GFP on a C R L 100-7 column Column cycle Yield Purity* 1 74.8 (\u00C2\u00B12)% >95% 2 88.7 (\u00C2\u00B13)% >95% 3 83.5 (\u00C2\u00B13)% >95% *Determined by SDS-PAGE analysis 196 6.6 Figures Figure 6.1 Picture of an MT100 column pre- and post-compression. Picture shows bed compression at a superficial velocity of 1.06 x 10\"4 m/s. 197 c -H I - C - C l o I H O H ' Epichlorohydrin 0\"\u00C2\u00AB.. . H C C C \u00E2\u0080\u0094 C l ' * i I I H H H Cellulose surface H I ' C -I H 0 \u00E2\u0080\u00A2 c \u00E2\u0080\u0094 c H , OH\" H 0\" H 1 ^ C -- C C H H H H OH H I I I \u00E2\u0080\u00A2 c \u00E2\u0080\u0094 c \u00E2\u0080\u0094 c -I I I H H H H OH H t c C \u00E2\u0080\u0094 - C 1 H H H ^ure 6.2 Schematic representation of the proposed epoxide-based reaction for cross-linking cellulose with epichlorohydrin. 198 -, , , n p 2 4 6 8 10 Epi mL/g Figure 6.3 Imageplot (A) and wireframe surface (B) for a locally weighted polynomial regression fit to the measured wcrjt (in m s\"1 x 104) data set reported in Table 1. Created with the loess, levelplot, and wireframe functions in R (Young et al. 1980) and the add-on package Lattice (Lattice Graphics R package, version 0.14-16), with a span of 0.65 and a degree of 2 (i.e. local fits are quadratic). The location of the maximum is indicated by (\u00C2\u00AE). 199 200 Superficial Velocity X 104 ( m / s ) Figure 6.5 Pressure curve for solvent flow through an MT100 (open circles) or CRL 100-7 (filled circles) column monitored over a range of superficial velocities. The pressure drop curve for C R L 100-7 was compiled from two independent experiments. 201 0 2 4 6 8 10 12 14 16 18 20 C. (uM) Figure 6.6 Equilibrium adsorption isotherms for binding of CBM9-GFP to MT100 (filled circles) and C R L 100-7 (open squares) at 21\u00C2\u00B0C. Control experiment shows ovalbumin binding to CRL 100-7 (filled squares) at 21\u00C2\u00B0C. The mobile phase buffer consisted of 50 m M phosphate buffer, 100 m M NaCl, pH 7. The curve represents the best fit of the experimental data to the Langmuir adsorption isotherm equation where q\ is the protein concentration bound to the media surface and c{ is the equilibrium concentration of protein free in solution. 202 0 400 800 1200 1600 2000 2400 2800 3200 Liu ( s ) Measured first central moment for blue dextran (filled squares) and CBM9-GFP (filled circles) on a column (I.D. 1.0 cm) packed with CRL 100-7. 203 (72 (x 10 7 ) ( m / s ) 2 Figure 6.8 Second moment analysis under non-binding conditions for a column (I.D. 1.0 cm) packed with C R L 100-7 media. Pulse injections of CBM9-GFP were used over the superficial velocity range 4.25 x 10\"5 m/s to 4.25 x 10\"4 m/s. The mobile phase consisted of 2 M glucose in 50 m M potassium phosphate, 100 m M NaCl (pH 7, 21\u00C2\u00B0C). 204 300 1000 reduced Figure 6.9 Van Deemter plot describing CBM9-GFP transport within the CRL100-7 column (I.D. 1.0 cm X 11.4 cm); the reduced plate height h = (a2L)/(jU\dP) is plotted as a function of the reduced linear velocity, given by (udv)IDM)-205 3500 ~ 3000J 25000 20000 TJ 0) 15000 JJ c o CD 10000 o "Thesis/Dissertation"@en . "10.14288/1.0059015"@en . "eng"@en . "Chemical and Biological Engineering"@en . "Vancouver : University of British Columbia Library"@en . "University of British Columbia"@en . "For non-commercial purposes only, such as research, private study and education. Additional conditions apply, see Terms of Use https://open.library.ubc.ca/terms_of_use."@en . "Graduate"@en . "The CBM9 fusion tag : a new technology for inexpensive production and affinity purification of recombinant proteins"@en . "Text"@en . "http://hdl.handle.net/2429/31363"@en .