CHEMBIOCHEM COMMUNICATIONS DOI: 10.1002/cbic.201400059
Derivatization of Antibody Fab Fragments: A Designer Enzyme for Native Protein Modification Sandra Liebscher,[a] Petra Kornberger,[b] Gerhard Fink,[c] Eva-Maria Trost-Gross,[c] Eva Hçss,[c] Arne Skerra,[b] and Frank Bordusa*[a] Bioconjugates, such as antibody–drug conjugates, have gained recent attention because of their increasing use in therapeutic and diagnostic applications. Commonly used conjugation reactions based upon chemoselective reagents exhibit a number of drawbacks: most of these reactions lack regio- and stereospecificity, thus resulting in loss of protein functionality due to random modifications. Enzymes provide an obvious solution to this problem, but the intrinsic (natural) substrate specificities of existing enzymes pose severe limitations to the kind of modifications that can be introduced. Here we describe the application of the novel trypsin variant trypsiligase for site-specific modification of the C terminus of a Fab antibody fragment via a stable peptide bond. The suitability of this designed biocatalyst was demonstrated by coupling the Her2-specific Fab to artificial functionalities of either therapeutic (PEG) or diagnostic (fluorescein) relevance. In both cases we obtained homogeneously modified Fab products bearing the artificial functionality exclusively at the desired position.
Monoclonal antibodies and their derivatives constitute the fastest-growing class of therapeutic molecules.[1] Combination of the unique specificity of monoclonal antibodies for tumor-associated antigens with the pharmacological potency of cytotoxic molecules or radionuclides results in powerful drugs, especially in oncology. In such antibody–drug conjugates (ADCs), the cytotoxic drug or radionuclide–chelate complex is covalently linked to an antibody, thereby resulting in a variety of new functional protein reagents. Currently, more than 140 clinical trials involving novel ADCs are listed by the U. S. National Institute of Health. Despite the extraordinary potential of these modified antibodies, there is a lack of generally usable and versatile methods that are both regiospecific and sufficiently mild to retain the full biological activity of the modified antibody [a] Dr. S. Liebscher, Prof. Dr. F. Bordusa Institute of Biochemistry/Biotechnology Martin Luther University Halle–Wittenberg Kurt-Mothes-Strasse 3, 06120 Halle/Saale (Germany) E-mail: [email protected] [b] Dr. P. Kornberger, Prof. Dr. A. Skerra Munich Center for Integrated Protein Science (CIPS-M) and Lehrstuhl fr Biologische Chemie, Technische Universitt Mnchen Emil-Erlenmeyer-Forum 5, 85350 Freising–Weihenstephan (Germany) [c] G. Fink, Dr. E.-M. Trost-Gross, Dr. E. Hçss Roche Diagnostics GmbH Nonnenwald 2, 82372 Penzberg (Germany) Supporting information for this article is available on the WWW under http://dx.doi.org/10.1002/cbic.201400059.
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
product. Although a few approaches have demonstrated regiospecific labeling of full length antibodies (e.g., through protein trans-splicing[2] or by using an expanded genetic code),[3] currently all known ADCs in clinical trials were synthesized by using chemically activated compounds, such as amine- or thiol-reactive derivatives. Amine-reactive reagents such as Nhydroxysuccinimidyl esters and isothiocyanates target aliphatic amino groups, including the N termini of proteins and the eamino groups of lysines.[4] Most proteins, especially antibodies, contain multiple lysine residues, thereby resulting in heterogeneous product mixtures. Alternative antibody derivatization protocols use maleimides, which react with the sulfhydryl groups of less-frequent cysteines through Michael addition.[5] This procedure involves partial reduction of disulfide bonds, which are essential for the structural integrity of multi-chain antibodies. Usually, one to four of these disulfide bonds becomes broken, thus resulting in up to eight reactive sites for alkylation, thus impairing antibody stability and provoking aggregation.[6] Despite the overall high degree of modification, up to 20 % of the antibodies can remain unconjugated.[5] Beside their use in ADC production, these chemical modification procedures also play a role in the preparation of polymermodified protein drugs.[7] For example, attachment of poly(ethylene glycol) (PEG) leads to polymer–protein conjugates that exhibit prolonged pharmacokinetics. PEG-modified adenosine deaminase and l-asparaginase are examples of proteins with extended plasma half-lives that have proven successful in clinical therapy.[8] Because of their inherent stereo- and regiospecificity, enzymes are ideal for mediating selective protein derivatization. Sortase A (SrtA) and bacterial transglutaminase (TG) are presently the only enzymes of this type that have been investigated with regard to suitability for derivatization of antibodies or related proteins. SrtA is a Staphylococcus aureus transpeptidase that recognizes the amino acid sequence LPXTG (X = D, E, A, N, Q, K) and catalyzes transpeptidation between threonine and glycine in the presence of oligo-glycine compounds as acyl acceptors. C-terminal derivatization, in this case biotinylation, of a single-chain Fv fragment of an antibody was recently demonstrated.[9] For fluorescent labeling of an Fab fragment, undesired hydrolysis and random crosslinking catalyzed by sortase were minimized.[10] Whether the modified Fab fragment had retained its binding functionality was uncertain; the remarkably high required excess of enzyme and substrate (relative to the target protein) is the most serious drawback for application of this biocatalyst at the preparative scale.
ChemBioChem 2014, 15, 1096 – 1100
1096
CHEMBIOCHEM COMMUNICATIONS Transglutaminases (TGs) catalyze acyl transfer reactions between the g-carboxamide group of glutamine and the e-amino group of lysine to form catabolically stable isopeptide bonds in vivo. Most TGs are promiscuous with respect to the lysine substrate and accept even 5-aminopentyl compounds. In contrast, the criteria for recognition of a glutamine residue are much more stringent: it must be located in a flexible region of the protein and flanked by specific amino acid residues.[11] Recently, bacterial TG was used for the conjugation of an antiCD20 antibody (Rituximab) with a chelator for radionuclides.[12] The modification reaction occurred at Glu295 in the Fc region; deglycosylation of Asn297 was required to enhance flexibility. The well-known high specificity of this enzyme for its protein substrate limits its general suitability. Given the reversibility of enzymatic reactions, proteinases are also suitable for ligation reactions, as demonstrated for bacterial IgA protease[13] and subtiligase.[14] Nevertheless, natural proteases suffer from a major drawback: hydrolysis is generally highly favored over ligation, thus requiring careful kinetic or thermodynamic control of the reaction. A variety of approaches have been explored to improve the balance between the competing reactions, including solvent, substrate, and proteinase engineering.[15] As a contribution to this approach, we recently described a trypsin variant carrying four amino acid exchanges (K60E/N143H/E151H/ D189K, “trypsiligase”), which was specifically designed for selective protein modification.[16] Its application for N-terminal labeling of proteins on a preparative scale with quantitative product yields was demonstrated. The designed biocatalyst is characterized by its highly selective proteolytic activity towards the very short triplet recognition sequence YRH, which has an occurrence of just 0.5 % in the human proteome. This substrate specificity combined with the enzyme’s high synthesis capacity and low hydrolytic activity enables site-directed modification of proteins such as antibody fragments by a transpeptidase reaction (Scheme 1). In order not to affect the antigen-binding activity of the Nterminal (variable) domains, the target Fab fragment was equipped with the biocatalyst’s recognition sequence YRH at the C terminus (at either the heavy or light chain), and complemented with a purification tag. Trypsiligase selectively cleaves between tyrosine and arginine of the recognition sequence,
Scheme 1. General procedure for selective modification of antibody Fab fragments catalyzed by trypsiligase. Initially, the Fab fragment is equipped with the recognition sequence YRH at one of its C termini and, optionally, with a peptide tag for purification. In the presence of trypsiligase and substrates with an N-terminal Arg-His moiety, transpeptidation results in the selective modification of the Fab fragment.
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org thereby resulting in the formation of a covalent acyl–enzyme intermediate, which is then aminolyzed by the functional group to be introduced into the protein of interest. To this end, diverse modification reagents with N-terminal Arg-His residues were used as efficient nucleophiles, thus resulting in covalent attachment to the Fab C terminus by a stable peptide bond. As proof of concept, we applied trypsiligase for the selective modification of the biopharmaceutically relevant anti-Her2 Fab fragment (4D5v8),[17] which is derived from the therapeutic antibody Trastuzumab; this is used in the treatment of patients with breast cancer overexpressing human epidermal growth factor receptor 2 (Her2). The recombinant Fab fragment was extended at the C terminus of its heavy chain by the recognition sequence YRH followed by the Strep-tag II[18] (amino acid sequence in Figure S1 in the Supporting Information). The modified Fab fragment was expressed in the periplasm of Escherichia coli JM83 cells and affinity-purified to homogeneity (see the Supporting Information and Experimental Section). Finally, the protein was dialyzed and concentrated to 150 mm, notably without formation of aggregates (Figure S2). The acceptance of the recognition sequence by trypsiligase was ascertained in initial hydrolysis experiments (Figure S3). Accordingly, the Strep-tag II peptide was quantitatively cleaved within 30 min without the formation of detectable byproducts, thus confirming the potential of the previously reported restrictive cleavage activity of the engineered enzyme through a substrate-activated catalysis mechanism.[16]
Scheme 2. Structures of trypsiligase substrates. The peptide RHAK served as scaffold for functionalization at the lysine side chain: 1) Ne-6-carboxyfluorescein, RHAK(CF); 2) Ne-7-methoxycoumarin-4-acetic acid/carbonyl C PEG 20000, RHAK(PEG20k).
ChemBioChem 2014, 15, 1096 – 1100
1097
CHEMBIOCHEM COMMUNICATIONS
www.chembiochem.org
Next, we focused on modification of the anti-Her2 Fab fragment with a fluorescent dye and with a pharmaceutically relevant polymer (PEG). To this end, we tested the RHAK peptide scaffold for functionalization with either 6-carboxyfluorescein (CF) or PEG 20000 by using the lysine Ne-amino function as the attachment site (Scheme 2, for synthesis details see the Supporting Information). Based on the optimized S’ subsite specificity of trypsiligase we achieved high acceptance of the H2N-Arg-His nucleophile, thus enabling efficient transpeptidation within the Cterminal peptide moiety of the Fab fragment. In the presence of variable concentrations of RHAK(CF), trypsiligase formed the desired modified Fab fragment within a few minutes, as confirmed by Coomassie Blue staining and fluorescence detection after SDS-PAGE of the reaction mixture at different time points (Figure 1 A, B). Furthermore, the identity of the desired product was proven by mass Figure 1. Trypsiligase-catalyzed modification of the anti-Her2 Fab fragment with 6-carboxyfluorescein (CF). A) Coomassie staining, and B) fluorescence detection of nonreducing SDS-PAGE of the reaction mixture. 1) PageRuler spectrometry (Figure 1 C). Beprestained marker, 2) t = 0 min, 3) 5 min, 4) 10 min, 5) 15 min, 6) 20 min, 7) 30 min. C) ESI mass spectrometric analside the main peak (carboxy- ysis of the reaction mixture at t = 20 min: Fab-CF (Mcalcd = 48 749 Da; Mfound = 48 750 Da), Fab-YRH-StrepII (Mcalcd = fluorescein-modified Fab frag- 49 412 Da; Mfound = 49 413 Da), Fab-Y (Mcalcd = 47 880 Da; Mfound = 47 881 Da). D) UPLC analysis of the time-course ment; Mcalcd = 48 749 Da; Mexp. = of product formation in the presence of different RHAK(CF) substrate concentrations: *: 200 mm, ~: 500 mm, &: 1 mm. Conditions: 100 mm Fab-YRH-StrepII, 0.2–1 mm RHAK(CF), 10 mm trypsiligase, 100 mm HEPES/NaOH pH 7.8, 48 750 Da), we only detected the 0.1 mm ZnCl2, 100 mm NaCl, 10 mm CaCl2. educt (Fab-YRH-StrepII; Mcalcd = 49 412 Da; Mexp. = 49 413 Da) and the hydrolysis product (Fabthan that with RHAK(CF), due to the high molecular weight of Y; Mcalcd = 47 880 Da; Mexp. = 47 881 Da), which appeared in simiPEG 20000. To simplify reaction analysis, the PEG substrate was lar amounts. The remaining educt (Fab, ~ 30 %) may in princiequipped with a coumarin group as an optical probe, thus enple be reused. MS analysis of the reaction mixture under reabling sensitive detection at 320 nm and exact quantification. ductive conditions indicated that only the heavy chain of the The separation of the PEGylated Fab fragment was simplified Fab was modified (Figure S4 B). To determine the yields of the by the high molecular weight of the PEG moiety (see the sigmodified Fab fragment, the reaction mixtures were further annificant size increase of the conjugate in SDS-PAGE; Figure 2 A, alyzed by ultra-performance liquid chromatography (UPLC) at B). Reductive cleavage of the intermolecular disulfide bond 440 nm (Figure S4 A). Typically, the product yields increased that links the heavy and light chains of the Fab fragment with the concentration of the RHAK(CF) peptide (~ 40 % yield further indicated a size shift exclusively for the heavy chain, as in the presence of 1 mm labeled peptide after 15 min reaction expected for the enzymatic PEG modification (cf. Figure S5). time; Figure 1 D). The slow decrease of product concentration UPLC-based separation and quantification of the PEG-conjugatafter longer reaction times was attributed to re-formation of ed Fab fragment revealed a maximum of about 35 % product the trypsiligase recognition sequence YRH in the modified Fab yield of PEG 20000-modified Fab (Figure 2 C, D). Similar results products, subsequently resulting in hydrolysis to Fab-Y after were obtained with a second Fab fragment, anti-DKK1;[19] this consumption of the educt Fab-YRH-StrepII. To achieve PEGylation of the anti-Her2 Fab we used RHAKdemonstrates the versality of the novel derivatization approach (PEG20k) as the transamidating nucleophile (Scheme 2). This (Figure S6). PEGylation reaction appeared to be by far more ambitious 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2014, 15, 1096 – 1100
1098
CHEMBIOCHEM COMMUNICATIONS
www.chembiochem.org
the unmodified versions; in contrast, koff values were similar (Table 1). Finally, we investigated the purified anti-Her2 Fab-PEG20k fragment for its biological binding activity toward the Her2 receptor displayed on the surface of SK-BR-3 tumor cells (Figure S9). The Kd value of the enzymatically modified Fab fragment (Table 1) did not reveal significant differences from those of the non-PEGylated Fab fragments. Taken together, the data obtained by surface plasmon resonance and flow cytofluorimetry clearly indicate that the functionality of the anti-Her2 Fab fragment was essentially retained after its enzymatic modification with PEG 20000 by using trypsiligase as a biocatalyst. The overall higher Kd values in the Figure 2. Trypsiligase-catalyzed PEGylation of the anti-Her2 Fab fragment. A) Coomassie staining, and B) subsequent iodine staining after nonreducing SDS-PAGE of the reaction mixture at different time points. 1) PageRuler FACS measurements (compared prestained marker, 2) reference anti-Her2 Fab fragment (Fab-YRH-StrepII), 3) t = 0 min, 4) t = 20 min, 5) t = 40 min, to those from the Biacore analy6) t = 60 min, 7) trypsiligase, 8) substrate RHAK(PEG20k). C) UPLC-analysis: b RHAK(PEG20k) (320 nm), b Fab sis) might be explained by re(280 nm), reaction mixture t = 10 min (c320 nm), (c280 nm). Peak 1: Fab, Peak 2: trypsiligase, Peak 3: Fabduced accessibility of the native PEG20k, Peak 4: RHAK(PEG20k). Acquity BEH130 C18 column (2.1 100 mm), 10–90 % ACN in 5 min. D) Time-course of product formation. Conditions: 100 mm anti-Her2 Fab fragment (Fab-YRH-StrepII), 1 mm RHAK(PEG20k), 10 mm Her2 receptor on the surface of trypsiligase, 100 mm HEPES/NaOH pH 7.8, 0.1 mm ZnCl2, 100 mm NaCl, 10 mm CaCl2. the eukaryotic cells surrounded by the glycocalyx. In summary, our results demonstrate that trypsiligase represents a universal and novel bioNext, we investigated the functionality of the PEG-conjugatcatalytic tool, not only for mediating efficient and highly seleced anti-Her2 Fab fragment. For this, the derivatized product tive N-terminal protein modification[16] but also for the regiowas isolated by size-exclusion chromatography on Superdex 75 from a preparative reaction at t = 30 min, after inactivation specific derivatization of antibody fragments at their C termini. of trypsiligase by addition of 2 mm EDTA (Figure 3 A). SDSThe enzymatic modification reactions proceed under native PAGE analysis of the relevant gel filtration fractions revealed conditions for the target protein, are fast, and are absolutely successful separation of the PEGylated Fab fragment from specific for the recognition sequence. The result is a homogenon-PEGylated Fab species (Figure 3 B, C). To remove residual neously modified protein product. This method does not refree PEG from the PEGylated Fab we additionally performed quire extended recognition sequences, expensive cofactors, or ion exchange chromatography (Figure S10). laborious substrate syntheses; it is fully compatible with estabNext, the modified anti-Her2 Fab fragment was compared lished standard expression protocols, and does not mediate with other versions prepared in this study (Fab-YRH-StrepII, undesired degradation reactions, in spite of the fact that trypsiFab-StrepII, and Fab-His6). We analyzed these samples by SDSligase was derived from a highly active protease. PAGE under reducing and nonreducing conditions (Figure S7). A size shift for the heavy chain conjugated with PEG 20000 was clearly detectable. Surface plasmon resonance measurements Table 1. Apparent Kd values for modified anti-Her2 Fab fragments meaperformed on a Biacore 2000 instrument with the covalently sured in Biacore and FACS experiments with the recombinant Her2 ectoimmobilized recombinant Her2 ectodomain (Figure S8) redomain and Her2-positive SK-BR-3 cells, respectively. vealed Kd = 0.442 nm for the PEG-conjugated Fab fragment, that is, just slightly higher than for Fab-YRH-StrepII (0.132 nm), Biacore FACS Analyte kon [106 m 1 s 1] koff [10 6 s 1] Kd [nm] Kd [nm] Fab-StrepII (0.143 nm), and Fab-His6 (0.110 nm; Table 1). This was mostly a result of decreased kon for the PEGylated Fab Fab-His6-tag 1.47 0.003 162 2.4 0.110 10.9 2.14 Fab-StrepII 1.25 0.003 179 2.6 0.143 3.43 1.31 fragment (0.34 106 m 1 s 1), thus indicating slightly slower Fab-YRH-StrepII 1.22 0.002 161 2.6 0.132 1.68 0.54 binding of the PEGylated Fab fragment to its antigen than for Fab-PEG 20000
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
0.34 0.001
152 2.1
0.442
4.17 0.47
ChemBioChem 2014, 15, 1096 – 1100
1099
CHEMBIOCHEM COMMUNICATIONS
www.chembiochem.org the DFG priority program SPP1623 “Chemoselective reactions for the synthesis and application of functional proteins”. The authors thank Christa Langer for technical assistance. Keywords: antibody conjugation · biocatalysis · proteases · protein modifications · transamidation
Acknowledgements
[1] A. Beck, J.-F. Haeuw, T. Wurch, L. Goetsch, C. Bailly, N. Corvaia, Discovery Med. 2010, 10, 329 – 339. [2] S. Mçhlmann, P. Bringmann, S. Greven, A. Harrenga, BMC Biotechnol. 2011, 11, 76. [3] F. Tian, Y. Lu, A. Manibusan, A. Sellers, H. Tran, Y. Sun, T. Phuong, R. Barnett, B. Hehli, F. Song, M. J. DeGuzman, S. Ensari, J. K. Pinkstaff, L. M. Sullivan, S. L. Biroc, H. Cho, P. G. Schultz, J. DiJoseph, M. Dougher, D. Ma, et al., Proc. Natl. Acad. Sci. USA 2014, 111, 1766 – 1771. [4] E. Basl, N. Joubert, M. Pucheault, Chem. Biol. 2010, 17, 213 – 227. [5] S. O. Doronina, B. E. Toki, M. Y. Torgov, B. A. Mendelsohn, C. G. Cerveny, D. F. Chace, R. L. DeBlanc, R. P. Gearing, T. D. Bovee, C. B. Siegall, J. A. Francisco, A. F. Wahl, D. L. Meyer, P. D. Senter, Nat. Biotechnol. 2003, 21, 778 – 784. [6] P. J. Burke, P. D. Senter, D. W. Meyer, J. B. Miyamoto, M. Anderson, B. E. Toki, G. Manikumar, M. C. Wani, D. J. Kroll, S. C. Jeffrey, Bioconjugate Chem. 2009, 20, 1242 – 1250. [7] a) C. Delgado, G. E. Francis, D. Fisher, Crit. Rev. Ther. Drug Carrier Syst. 1992, 9, 249 – 304; b) S. Shaunak, A. Godwin, J.-W. Choi, S. Balan, E. Pedone, D. Vijayarangam, S. Heidelberger, I. Teo, M. Zloh, S. Brocchini, Nat. Chem. Biol. 2006, 2, 312 – 313. [8] a) M. S. Hershfield, R. H. Buckley, M. L. Greenberg, A. L. Melton, R. Schiff, C. Hatem, J. Kurtzberg, M. L. Markert, R. H. Kobayashi, A. L. Kobayashi, A. Abuchowski, N. Engl. J. Med. 1987, 316, 589 – 596; b) M. J. Keating, R. Holmes, S. Lerner, D. H. Ho, Leuk. Lymphoma 1993, 10, 153 – 157. [9] a) M. P. Madej, G. Coia, C. C. Williams, J. M. Caine, L. A. Pearce, R. Attwood, N. A. Bartone, O. Dolezal, R. M. Nisbet, S. D. Nuttall, T. E. Adams, Biotechnol. Bioeng. 2012, 109, 1461 – 1470; b) H. T. Ta, S. Prabhu, E. Leitner, F. Jia, D. von Elverfeldt, K. E. Jackson, T. Heidt, A. K. N. Nair, H. Pearce, C. von zur Muhlen, X. Wang, K. Peter, C. E. Hagemeyer, Circ. Res. 2011, 109, 365 – 373. [10] S. Mçhlmann, C. Mahlert, S. Greven, P. Scholz, A. Harrenga, ChemBioChem 2011, 12, 1774 – 1780. [11] A. Fontana, B. Spolaore, A. Mero, F. M. Veronese, Adv. Drug Delivery Rev. 2008, 60, 13 – 28. [12] S. Jeger, K. Zimmermann, A. Blanc, J. Grnberg, M. Honer, P. Hunziker, H. Struthers, R. Schibli, Angew. Chem. Int. Ed. 2010, 49, 9995 – 9997; Angew. Chem. 2010, 122, 10191 – 10194. [13] M. Lewinska, C. Seitz, A. Skerra, F. P. Schmidtchen, Bioconjugate Chem. 2004, 15, 231 – 234. [14] D. Y. Jackson, J. Burnier, C. Quan, M. Stanley, J. Tom, J. A. Wells, Science 1994, 266, 243 – 247. [15] F. Bordusa, Chem. Rev. 2002, 102, 4817 – 4868. [16] S. Liebscher, M. Schçpfel, T. Aumller, A. Sharkhuukhen, A. Pech, E. Hçss, C. Parthier, G. Jahreis, M. T. Stubbs, F. Bordusa, Angew. Chem. Int. Ed. 2014, 53, 3024 – 3028. [17] M. Schlapschy, I. Theobald, H. Mack, M. Schottelius, H.-J. Wester, A. Skerra, Protein Eng. Des. Sel. 2007, 20, 273 – 284. [18] T. G. M. Schmidt, A. Skerra, Nat. Protoc. 2007, 2, 1528 – 1535. [19] M. Fulciniti, P. Tassone, T. Hideshima, S. Vallet, P. Nanjappa, S. A. Ettenberg, Z. Shen, N. Patel, Y.-t. Tai, D. Chauhan, C. Mitsiades, R. Prabhala, N. Raje, K. C. Anderson, D. R. Stover, N. C. Munshi, Blood 2009, 114, 371 – 379.
This work was supported by the BMBF network ProNet-T3 “Protein competence center Halle: tools, targets & therapeutics” and
Received: January 24, 2014 Published online on April 29, 2014
Figure 3. Isolation of the PEG-conjugated anti-Her2 Fab fragment. A) Sizeexclusion chromatography on a Superdex 75 column (100 mm HEPES/NaOH pH 7.5, 150 mm NaCl; flow rate 1 mL min 1). B) Iodine and Coomassie staining, and C) Coomassie staining after nonreducing SDS-PAGE of the elution fractions: 1) PageRuler Plus prestained standard ladder (Fermentas/Thermo Scientific), 2) elution fraction 9, 3) fraction 11, 4) fraction 12, 5) fraction 13, 6) fraction 15, 7) fraction 18, 8) fraction 20, 9) fraction 22, 10) fraction 24.
Furthermore, as a matter particularly important for the modification of disulfide-bridged antibodies or their fragments, trypsiligase does not need the addition of reducing thiols for enzyme activation (as is regularly recommended for cysteinedependent enzymes). Future studies will focus on further optimization of trypsiligase to improve the product yield of the reaction.
Experimental Section Full experimental details are given in the Supporting Information.
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2014, 15, 1096 – 1100
1100
Derivatization of Antibody Fab Fragments: A Designer Enzyme for Native Protein Modification Sandra Liebscher,[a] Petra Kornberger,[b] Gerhard Fink,[c] Eva-Maria Trost-Gross,[c] Eva Hçss,[c] Arne Skerra,[b] and Frank Bordusa*[a] Bioconjugates, such as antibody–drug conjugates, have gained recent attention because of their increasing use in therapeutic and diagnostic applications. Commonly used conjugation reactions based upon chemoselective reagents exhibit a number of drawbacks: most of these reactions lack regio- and stereospecificity, thus resulting in loss of protein functionality due to random modifications. Enzymes provide an obvious solution to this problem, but the intrinsic (natural) substrate specificities of existing enzymes pose severe limitations to the kind of modifications that can be introduced. Here we describe the application of the novel trypsin variant trypsiligase for site-specific modification of the C terminus of a Fab antibody fragment via a stable peptide bond. The suitability of this designed biocatalyst was demonstrated by coupling the Her2-specific Fab to artificial functionalities of either therapeutic (PEG) or diagnostic (fluorescein) relevance. In both cases we obtained homogeneously modified Fab products bearing the artificial functionality exclusively at the desired position.
Monoclonal antibodies and their derivatives constitute the fastest-growing class of therapeutic molecules.[1] Combination of the unique specificity of monoclonal antibodies for tumor-associated antigens with the pharmacological potency of cytotoxic molecules or radionuclides results in powerful drugs, especially in oncology. In such antibody–drug conjugates (ADCs), the cytotoxic drug or radionuclide–chelate complex is covalently linked to an antibody, thereby resulting in a variety of new functional protein reagents. Currently, more than 140 clinical trials involving novel ADCs are listed by the U. S. National Institute of Health. Despite the extraordinary potential of these modified antibodies, there is a lack of generally usable and versatile methods that are both regiospecific and sufficiently mild to retain the full biological activity of the modified antibody [a] Dr. S. Liebscher, Prof. Dr. F. Bordusa Institute of Biochemistry/Biotechnology Martin Luther University Halle–Wittenberg Kurt-Mothes-Strasse 3, 06120 Halle/Saale (Germany) E-mail: [email protected] [b] Dr. P. Kornberger, Prof. Dr. A. Skerra Munich Center for Integrated Protein Science (CIPS-M) and Lehrstuhl fr Biologische Chemie, Technische Universitt Mnchen Emil-Erlenmeyer-Forum 5, 85350 Freising–Weihenstephan (Germany) [c] G. Fink, Dr. E.-M. Trost-Gross, Dr. E. Hçss Roche Diagnostics GmbH Nonnenwald 2, 82372 Penzberg (Germany) Supporting information for this article is available on the WWW under http://dx.doi.org/10.1002/cbic.201400059.
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
product. Although a few approaches have demonstrated regiospecific labeling of full length antibodies (e.g., through protein trans-splicing[2] or by using an expanded genetic code),[3] currently all known ADCs in clinical trials were synthesized by using chemically activated compounds, such as amine- or thiol-reactive derivatives. Amine-reactive reagents such as Nhydroxysuccinimidyl esters and isothiocyanates target aliphatic amino groups, including the N termini of proteins and the eamino groups of lysines.[4] Most proteins, especially antibodies, contain multiple lysine residues, thereby resulting in heterogeneous product mixtures. Alternative antibody derivatization protocols use maleimides, which react with the sulfhydryl groups of less-frequent cysteines through Michael addition.[5] This procedure involves partial reduction of disulfide bonds, which are essential for the structural integrity of multi-chain antibodies. Usually, one to four of these disulfide bonds becomes broken, thus resulting in up to eight reactive sites for alkylation, thus impairing antibody stability and provoking aggregation.[6] Despite the overall high degree of modification, up to 20 % of the antibodies can remain unconjugated.[5] Beside their use in ADC production, these chemical modification procedures also play a role in the preparation of polymermodified protein drugs.[7] For example, attachment of poly(ethylene glycol) (PEG) leads to polymer–protein conjugates that exhibit prolonged pharmacokinetics. PEG-modified adenosine deaminase and l-asparaginase are examples of proteins with extended plasma half-lives that have proven successful in clinical therapy.[8] Because of their inherent stereo- and regiospecificity, enzymes are ideal for mediating selective protein derivatization. Sortase A (SrtA) and bacterial transglutaminase (TG) are presently the only enzymes of this type that have been investigated with regard to suitability for derivatization of antibodies or related proteins. SrtA is a Staphylococcus aureus transpeptidase that recognizes the amino acid sequence LPXTG (X = D, E, A, N, Q, K) and catalyzes transpeptidation between threonine and glycine in the presence of oligo-glycine compounds as acyl acceptors. C-terminal derivatization, in this case biotinylation, of a single-chain Fv fragment of an antibody was recently demonstrated.[9] For fluorescent labeling of an Fab fragment, undesired hydrolysis and random crosslinking catalyzed by sortase were minimized.[10] Whether the modified Fab fragment had retained its binding functionality was uncertain; the remarkably high required excess of enzyme and substrate (relative to the target protein) is the most serious drawback for application of this biocatalyst at the preparative scale.
ChemBioChem 2014, 15, 1096 – 1100
1096
CHEMBIOCHEM COMMUNICATIONS Transglutaminases (TGs) catalyze acyl transfer reactions between the g-carboxamide group of glutamine and the e-amino group of lysine to form catabolically stable isopeptide bonds in vivo. Most TGs are promiscuous with respect to the lysine substrate and accept even 5-aminopentyl compounds. In contrast, the criteria for recognition of a glutamine residue are much more stringent: it must be located in a flexible region of the protein and flanked by specific amino acid residues.[11] Recently, bacterial TG was used for the conjugation of an antiCD20 antibody (Rituximab) with a chelator for radionuclides.[12] The modification reaction occurred at Glu295 in the Fc region; deglycosylation of Asn297 was required to enhance flexibility. The well-known high specificity of this enzyme for its protein substrate limits its general suitability. Given the reversibility of enzymatic reactions, proteinases are also suitable for ligation reactions, as demonstrated for bacterial IgA protease[13] and subtiligase.[14] Nevertheless, natural proteases suffer from a major drawback: hydrolysis is generally highly favored over ligation, thus requiring careful kinetic or thermodynamic control of the reaction. A variety of approaches have been explored to improve the balance between the competing reactions, including solvent, substrate, and proteinase engineering.[15] As a contribution to this approach, we recently described a trypsin variant carrying four amino acid exchanges (K60E/N143H/E151H/ D189K, “trypsiligase”), which was specifically designed for selective protein modification.[16] Its application for N-terminal labeling of proteins on a preparative scale with quantitative product yields was demonstrated. The designed biocatalyst is characterized by its highly selective proteolytic activity towards the very short triplet recognition sequence YRH, which has an occurrence of just 0.5 % in the human proteome. This substrate specificity combined with the enzyme’s high synthesis capacity and low hydrolytic activity enables site-directed modification of proteins such as antibody fragments by a transpeptidase reaction (Scheme 1). In order not to affect the antigen-binding activity of the Nterminal (variable) domains, the target Fab fragment was equipped with the biocatalyst’s recognition sequence YRH at the C terminus (at either the heavy or light chain), and complemented with a purification tag. Trypsiligase selectively cleaves between tyrosine and arginine of the recognition sequence,
Scheme 1. General procedure for selective modification of antibody Fab fragments catalyzed by trypsiligase. Initially, the Fab fragment is equipped with the recognition sequence YRH at one of its C termini and, optionally, with a peptide tag for purification. In the presence of trypsiligase and substrates with an N-terminal Arg-His moiety, transpeptidation results in the selective modification of the Fab fragment.
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org thereby resulting in the formation of a covalent acyl–enzyme intermediate, which is then aminolyzed by the functional group to be introduced into the protein of interest. To this end, diverse modification reagents with N-terminal Arg-His residues were used as efficient nucleophiles, thus resulting in covalent attachment to the Fab C terminus by a stable peptide bond. As proof of concept, we applied trypsiligase for the selective modification of the biopharmaceutically relevant anti-Her2 Fab fragment (4D5v8),[17] which is derived from the therapeutic antibody Trastuzumab; this is used in the treatment of patients with breast cancer overexpressing human epidermal growth factor receptor 2 (Her2). The recombinant Fab fragment was extended at the C terminus of its heavy chain by the recognition sequence YRH followed by the Strep-tag II[18] (amino acid sequence in Figure S1 in the Supporting Information). The modified Fab fragment was expressed in the periplasm of Escherichia coli JM83 cells and affinity-purified to homogeneity (see the Supporting Information and Experimental Section). Finally, the protein was dialyzed and concentrated to 150 mm, notably without formation of aggregates (Figure S2). The acceptance of the recognition sequence by trypsiligase was ascertained in initial hydrolysis experiments (Figure S3). Accordingly, the Strep-tag II peptide was quantitatively cleaved within 30 min without the formation of detectable byproducts, thus confirming the potential of the previously reported restrictive cleavage activity of the engineered enzyme through a substrate-activated catalysis mechanism.[16]
Scheme 2. Structures of trypsiligase substrates. The peptide RHAK served as scaffold for functionalization at the lysine side chain: 1) Ne-6-carboxyfluorescein, RHAK(CF); 2) Ne-7-methoxycoumarin-4-acetic acid/carbonyl C PEG 20000, RHAK(PEG20k).
ChemBioChem 2014, 15, 1096 – 1100
1097
CHEMBIOCHEM COMMUNICATIONS
www.chembiochem.org
Next, we focused on modification of the anti-Her2 Fab fragment with a fluorescent dye and with a pharmaceutically relevant polymer (PEG). To this end, we tested the RHAK peptide scaffold for functionalization with either 6-carboxyfluorescein (CF) or PEG 20000 by using the lysine Ne-amino function as the attachment site (Scheme 2, for synthesis details see the Supporting Information). Based on the optimized S’ subsite specificity of trypsiligase we achieved high acceptance of the H2N-Arg-His nucleophile, thus enabling efficient transpeptidation within the Cterminal peptide moiety of the Fab fragment. In the presence of variable concentrations of RHAK(CF), trypsiligase formed the desired modified Fab fragment within a few minutes, as confirmed by Coomassie Blue staining and fluorescence detection after SDS-PAGE of the reaction mixture at different time points (Figure 1 A, B). Furthermore, the identity of the desired product was proven by mass Figure 1. Trypsiligase-catalyzed modification of the anti-Her2 Fab fragment with 6-carboxyfluorescein (CF). A) Coomassie staining, and B) fluorescence detection of nonreducing SDS-PAGE of the reaction mixture. 1) PageRuler spectrometry (Figure 1 C). Beprestained marker, 2) t = 0 min, 3) 5 min, 4) 10 min, 5) 15 min, 6) 20 min, 7) 30 min. C) ESI mass spectrometric analside the main peak (carboxy- ysis of the reaction mixture at t = 20 min: Fab-CF (Mcalcd = 48 749 Da; Mfound = 48 750 Da), Fab-YRH-StrepII (Mcalcd = fluorescein-modified Fab frag- 49 412 Da; Mfound = 49 413 Da), Fab-Y (Mcalcd = 47 880 Da; Mfound = 47 881 Da). D) UPLC analysis of the time-course ment; Mcalcd = 48 749 Da; Mexp. = of product formation in the presence of different RHAK(CF) substrate concentrations: *: 200 mm, ~: 500 mm, &: 1 mm. Conditions: 100 mm Fab-YRH-StrepII, 0.2–1 mm RHAK(CF), 10 mm trypsiligase, 100 mm HEPES/NaOH pH 7.8, 48 750 Da), we only detected the 0.1 mm ZnCl2, 100 mm NaCl, 10 mm CaCl2. educt (Fab-YRH-StrepII; Mcalcd = 49 412 Da; Mexp. = 49 413 Da) and the hydrolysis product (Fabthan that with RHAK(CF), due to the high molecular weight of Y; Mcalcd = 47 880 Da; Mexp. = 47 881 Da), which appeared in simiPEG 20000. To simplify reaction analysis, the PEG substrate was lar amounts. The remaining educt (Fab, ~ 30 %) may in princiequipped with a coumarin group as an optical probe, thus enple be reused. MS analysis of the reaction mixture under reabling sensitive detection at 320 nm and exact quantification. ductive conditions indicated that only the heavy chain of the The separation of the PEGylated Fab fragment was simplified Fab was modified (Figure S4 B). To determine the yields of the by the high molecular weight of the PEG moiety (see the sigmodified Fab fragment, the reaction mixtures were further annificant size increase of the conjugate in SDS-PAGE; Figure 2 A, alyzed by ultra-performance liquid chromatography (UPLC) at B). Reductive cleavage of the intermolecular disulfide bond 440 nm (Figure S4 A). Typically, the product yields increased that links the heavy and light chains of the Fab fragment with the concentration of the RHAK(CF) peptide (~ 40 % yield further indicated a size shift exclusively for the heavy chain, as in the presence of 1 mm labeled peptide after 15 min reaction expected for the enzymatic PEG modification (cf. Figure S5). time; Figure 1 D). The slow decrease of product concentration UPLC-based separation and quantification of the PEG-conjugatafter longer reaction times was attributed to re-formation of ed Fab fragment revealed a maximum of about 35 % product the trypsiligase recognition sequence YRH in the modified Fab yield of PEG 20000-modified Fab (Figure 2 C, D). Similar results products, subsequently resulting in hydrolysis to Fab-Y after were obtained with a second Fab fragment, anti-DKK1;[19] this consumption of the educt Fab-YRH-StrepII. To achieve PEGylation of the anti-Her2 Fab we used RHAKdemonstrates the versality of the novel derivatization approach (PEG20k) as the transamidating nucleophile (Scheme 2). This (Figure S6). PEGylation reaction appeared to be by far more ambitious 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2014, 15, 1096 – 1100
1098
CHEMBIOCHEM COMMUNICATIONS
www.chembiochem.org
the unmodified versions; in contrast, koff values were similar (Table 1). Finally, we investigated the purified anti-Her2 Fab-PEG20k fragment for its biological binding activity toward the Her2 receptor displayed on the surface of SK-BR-3 tumor cells (Figure S9). The Kd value of the enzymatically modified Fab fragment (Table 1) did not reveal significant differences from those of the non-PEGylated Fab fragments. Taken together, the data obtained by surface plasmon resonance and flow cytofluorimetry clearly indicate that the functionality of the anti-Her2 Fab fragment was essentially retained after its enzymatic modification with PEG 20000 by using trypsiligase as a biocatalyst. The overall higher Kd values in the Figure 2. Trypsiligase-catalyzed PEGylation of the anti-Her2 Fab fragment. A) Coomassie staining, and B) subsequent iodine staining after nonreducing SDS-PAGE of the reaction mixture at different time points. 1) PageRuler FACS measurements (compared prestained marker, 2) reference anti-Her2 Fab fragment (Fab-YRH-StrepII), 3) t = 0 min, 4) t = 20 min, 5) t = 40 min, to those from the Biacore analy6) t = 60 min, 7) trypsiligase, 8) substrate RHAK(PEG20k). C) UPLC-analysis: b RHAK(PEG20k) (320 nm), b Fab sis) might be explained by re(280 nm), reaction mixture t = 10 min (c320 nm), (c280 nm). Peak 1: Fab, Peak 2: trypsiligase, Peak 3: Fabduced accessibility of the native PEG20k, Peak 4: RHAK(PEG20k). Acquity BEH130 C18 column (2.1 100 mm), 10–90 % ACN in 5 min. D) Time-course of product formation. Conditions: 100 mm anti-Her2 Fab fragment (Fab-YRH-StrepII), 1 mm RHAK(PEG20k), 10 mm Her2 receptor on the surface of trypsiligase, 100 mm HEPES/NaOH pH 7.8, 0.1 mm ZnCl2, 100 mm NaCl, 10 mm CaCl2. the eukaryotic cells surrounded by the glycocalyx. In summary, our results demonstrate that trypsiligase represents a universal and novel bioNext, we investigated the functionality of the PEG-conjugatcatalytic tool, not only for mediating efficient and highly seleced anti-Her2 Fab fragment. For this, the derivatized product tive N-terminal protein modification[16] but also for the regiowas isolated by size-exclusion chromatography on Superdex 75 from a preparative reaction at t = 30 min, after inactivation specific derivatization of antibody fragments at their C termini. of trypsiligase by addition of 2 mm EDTA (Figure 3 A). SDSThe enzymatic modification reactions proceed under native PAGE analysis of the relevant gel filtration fractions revealed conditions for the target protein, are fast, and are absolutely successful separation of the PEGylated Fab fragment from specific for the recognition sequence. The result is a homogenon-PEGylated Fab species (Figure 3 B, C). To remove residual neously modified protein product. This method does not refree PEG from the PEGylated Fab we additionally performed quire extended recognition sequences, expensive cofactors, or ion exchange chromatography (Figure S10). laborious substrate syntheses; it is fully compatible with estabNext, the modified anti-Her2 Fab fragment was compared lished standard expression protocols, and does not mediate with other versions prepared in this study (Fab-YRH-StrepII, undesired degradation reactions, in spite of the fact that trypsiFab-StrepII, and Fab-His6). We analyzed these samples by SDSligase was derived from a highly active protease. PAGE under reducing and nonreducing conditions (Figure S7). A size shift for the heavy chain conjugated with PEG 20000 was clearly detectable. Surface plasmon resonance measurements Table 1. Apparent Kd values for modified anti-Her2 Fab fragments meaperformed on a Biacore 2000 instrument with the covalently sured in Biacore and FACS experiments with the recombinant Her2 ectoimmobilized recombinant Her2 ectodomain (Figure S8) redomain and Her2-positive SK-BR-3 cells, respectively. vealed Kd = 0.442 nm for the PEG-conjugated Fab fragment, that is, just slightly higher than for Fab-YRH-StrepII (0.132 nm), Biacore FACS Analyte kon [106 m 1 s 1] koff [10 6 s 1] Kd [nm] Kd [nm] Fab-StrepII (0.143 nm), and Fab-His6 (0.110 nm; Table 1). This was mostly a result of decreased kon for the PEGylated Fab Fab-His6-tag 1.47 0.003 162 2.4 0.110 10.9 2.14 Fab-StrepII 1.25 0.003 179 2.6 0.143 3.43 1.31 fragment (0.34 106 m 1 s 1), thus indicating slightly slower Fab-YRH-StrepII 1.22 0.002 161 2.6 0.132 1.68 0.54 binding of the PEGylated Fab fragment to its antigen than for Fab-PEG 20000
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
0.34 0.001
152 2.1
0.442
4.17 0.47
ChemBioChem 2014, 15, 1096 – 1100
1099
CHEMBIOCHEM COMMUNICATIONS
www.chembiochem.org the DFG priority program SPP1623 “Chemoselective reactions for the synthesis and application of functional proteins”. The authors thank Christa Langer for technical assistance. Keywords: antibody conjugation · biocatalysis · proteases · protein modifications · transamidation
Acknowledgements
[1] A. Beck, J.-F. Haeuw, T. Wurch, L. Goetsch, C. Bailly, N. Corvaia, Discovery Med. 2010, 10, 329 – 339. [2] S. Mçhlmann, P. Bringmann, S. Greven, A. Harrenga, BMC Biotechnol. 2011, 11, 76. [3] F. Tian, Y. Lu, A. Manibusan, A. Sellers, H. Tran, Y. Sun, T. Phuong, R. Barnett, B. Hehli, F. Song, M. J. DeGuzman, S. Ensari, J. K. Pinkstaff, L. M. Sullivan, S. L. Biroc, H. Cho, P. G. Schultz, J. DiJoseph, M. Dougher, D. Ma, et al., Proc. Natl. Acad. Sci. USA 2014, 111, 1766 – 1771. [4] E. Basl, N. Joubert, M. Pucheault, Chem. Biol. 2010, 17, 213 – 227. [5] S. O. Doronina, B. E. Toki, M. Y. Torgov, B. A. Mendelsohn, C. G. Cerveny, D. F. Chace, R. L. DeBlanc, R. P. Gearing, T. D. Bovee, C. B. Siegall, J. A. Francisco, A. F. Wahl, D. L. Meyer, P. D. Senter, Nat. Biotechnol. 2003, 21, 778 – 784. [6] P. J. Burke, P. D. Senter, D. W. Meyer, J. B. Miyamoto, M. Anderson, B. E. Toki, G. Manikumar, M. C. Wani, D. J. Kroll, S. C. Jeffrey, Bioconjugate Chem. 2009, 20, 1242 – 1250. [7] a) C. Delgado, G. E. Francis, D. Fisher, Crit. Rev. Ther. Drug Carrier Syst. 1992, 9, 249 – 304; b) S. Shaunak, A. Godwin, J.-W. Choi, S. Balan, E. Pedone, D. Vijayarangam, S. Heidelberger, I. Teo, M. Zloh, S. Brocchini, Nat. Chem. Biol. 2006, 2, 312 – 313. [8] a) M. S. Hershfield, R. H. Buckley, M. L. Greenberg, A. L. Melton, R. Schiff, C. Hatem, J. Kurtzberg, M. L. Markert, R. H. Kobayashi, A. L. Kobayashi, A. Abuchowski, N. Engl. J. Med. 1987, 316, 589 – 596; b) M. J. Keating, R. Holmes, S. Lerner, D. H. Ho, Leuk. Lymphoma 1993, 10, 153 – 157. [9] a) M. P. Madej, G. Coia, C. C. Williams, J. M. Caine, L. A. Pearce, R. Attwood, N. A. Bartone, O. Dolezal, R. M. Nisbet, S. D. Nuttall, T. E. Adams, Biotechnol. Bioeng. 2012, 109, 1461 – 1470; b) H. T. Ta, S. Prabhu, E. Leitner, F. Jia, D. von Elverfeldt, K. E. Jackson, T. Heidt, A. K. N. Nair, H. Pearce, C. von zur Muhlen, X. Wang, K. Peter, C. E. Hagemeyer, Circ. Res. 2011, 109, 365 – 373. [10] S. Mçhlmann, C. Mahlert, S. Greven, P. Scholz, A. Harrenga, ChemBioChem 2011, 12, 1774 – 1780. [11] A. Fontana, B. Spolaore, A. Mero, F. M. Veronese, Adv. Drug Delivery Rev. 2008, 60, 13 – 28. [12] S. Jeger, K. Zimmermann, A. Blanc, J. Grnberg, M. Honer, P. Hunziker, H. Struthers, R. Schibli, Angew. Chem. Int. Ed. 2010, 49, 9995 – 9997; Angew. Chem. 2010, 122, 10191 – 10194. [13] M. Lewinska, C. Seitz, A. Skerra, F. P. Schmidtchen, Bioconjugate Chem. 2004, 15, 231 – 234. [14] D. Y. Jackson, J. Burnier, C. Quan, M. Stanley, J. Tom, J. A. Wells, Science 1994, 266, 243 – 247. [15] F. Bordusa, Chem. Rev. 2002, 102, 4817 – 4868. [16] S. Liebscher, M. Schçpfel, T. Aumller, A. Sharkhuukhen, A. Pech, E. Hçss, C. Parthier, G. Jahreis, M. T. Stubbs, F. Bordusa, Angew. Chem. Int. Ed. 2014, 53, 3024 – 3028. [17] M. Schlapschy, I. Theobald, H. Mack, M. Schottelius, H.-J. Wester, A. Skerra, Protein Eng. Des. Sel. 2007, 20, 273 – 284. [18] T. G. M. Schmidt, A. Skerra, Nat. Protoc. 2007, 2, 1528 – 1535. [19] M. Fulciniti, P. Tassone, T. Hideshima, S. Vallet, P. Nanjappa, S. A. Ettenberg, Z. Shen, N. Patel, Y.-t. Tai, D. Chauhan, C. Mitsiades, R. Prabhala, N. Raje, K. C. Anderson, D. R. Stover, N. C. Munshi, Blood 2009, 114, 371 – 379.
This work was supported by the BMBF network ProNet-T3 “Protein competence center Halle: tools, targets & therapeutics” and
Received: January 24, 2014 Published online on April 29, 2014
Figure 3. Isolation of the PEG-conjugated anti-Her2 Fab fragment. A) Sizeexclusion chromatography on a Superdex 75 column (100 mm HEPES/NaOH pH 7.5, 150 mm NaCl; flow rate 1 mL min 1). B) Iodine and Coomassie staining, and C) Coomassie staining after nonreducing SDS-PAGE of the elution fractions: 1) PageRuler Plus prestained standard ladder (Fermentas/Thermo Scientific), 2) elution fraction 9, 3) fraction 11, 4) fraction 12, 5) fraction 13, 6) fraction 15, 7) fraction 18, 8) fraction 20, 9) fraction 22, 10) fraction 24.
Furthermore, as a matter particularly important for the modification of disulfide-bridged antibodies or their fragments, trypsiligase does not need the addition of reducing thiols for enzyme activation (as is regularly recommended for cysteinedependent enzymes). Future studies will focus on further optimization of trypsiligase to improve the product yield of the reaction.
Experimental Section Full experimental details are given in the Supporting Information.
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2014, 15, 1096 – 1100
1100
