Covalent Immunoglobulin Labeling through a Photoactivable

Article pubs.acs.org/bc

Covalent Immunoglobulin Labeling through a Photoactivable Synthetic Z Domain Anna Konrad,† Amelie Eriksson Karlström,‡ and Sophia Hober*,† †

Division of Proteomics, AlbaNova University Center, ‡Division of Molecular Biotechnology, School of Biotechnology, Royal Institute of Technology (KTH), SE-106 91 Stockholm, Sweden ABSTRACT: Traditionally, labeling of antibodies has been performed by covalent conjugation to amine or carboxyl groups. These methods are efficient but suffer from nonspecificity, since all free and available amine/carboxyl groups have the possibility to react. This drawback may lead to uncontrolled levels and locations of the labeling. Hence, the labeled molecules might behave differently and, possibly, the binding site of the antibody will also be affected. In this project, we have developed a highly stringent method for labeling of antibodies by utilizing an immunoglobulin-binding domain from protein A, the Z domain. Domain Z has been synthesized with an amino acid analogue, benzoylphenylalanine, capable of forming covalent attachment to other amino acids upon UV-exposure. This feature has been used for directed labeling of immunoglobulins and subsequent use of these in different assays.

■

desirable due to stabilizing molecules often being added to the antibody mixture for stabilizing purposes. Several Ig-binding molecules have been reported in literature.3−5 Among these, staphylococcal protein A (SpA), binding to VH and Fc, is one of the best characterized. Staphylococcal protein A is frequently used in many different applications, such as affinity chromatography, where its ability to bind antibodies is utilized. The protein is used both for purification of IgG molecules6 and as affinity tag for protein purification.7 The five homologous domains, EDABC, that constitute protein A, each consists of approximately 58 aa, and share the Ig-binding feature.8 NMR analysis of the structure of the B domain shows a three-helix bundle with the helices ordered in an antiparallel fashion. The domains of protein A exhibit binding to both the Fc and Fab regions of immunoglobulins. By X-ray crystallography, the structure of the B domain in complex with the Fc region of IgG has been solved, and it revealed an interaction that mostly involved amino acids of hydrophobic character.9 Binding seems to occur in the interface between CH2 and CH3 of IgG, where 11 residues from helix 1 and helix 2 of the B domain are suggested to participate.9,10 Also, the structure of the interaction between protein A and the Fab region has been solved. The crystal structure of the D domain binding the Fab region of human IgM disclosed the involvement of 11 residues of helices 2 and 3 from the D domain and an interaction of a polar character with the variable heavy chain.11 By altering two positions in the B-domain of protein A, an engineered variant called the Z

INTRODUCTION

Immunoglobulins, or antibodies, are a group of molecules that are extensively used as affinity reagents in many applications in research, clinical diagnostics, and therapy. With the ability to bind their ligands with high affinity and in a selective manner, they are of great importance, and are, by far, the most commonly used affinity reagents.1 As when employing any affinity reagent, the method used for detection needs to be considered. The antibodies’ excellent capacities to bind their ligands need to be combined with high selectivity in order to be of great use. As a consequence, techniques for labeling of antibodies necessarily need to be of high quality. To make detection of the antibody and its binding events possible, a reporter group is normally attached to the antibody. Methods most commonly used when conjugating reporter groups to the antibody are based on the exploration of amine or carboxyl groups in the protein for coupling.2 Another application that ordinarily takes advantage of the surface-exposed functional groups in the protein is the immobilization to a solid support. When applying amine- or carboxyl-based chemistry, normally a high degree of labeling or immobilization is obtained, but unfortunately, the binding site might be compromised, since the control of level and location of the labeling/coupling is limited. This means that optimization of the protocol is needed for every antibody and every conjugation/immobilization. Hence, a method for labeling, where specific and controlled conjugation can be achieved, would be a great advantage. To obtain specific conjugation, molecules that have a natural, specific, and defined binding to antibodies could be employed. Moreover, to be able to conjugate the antibodies in a complex environment, with other proteins in the conjugation mixture, is © 2011 American Chemical Society

Received: January 26, 2011 Revised: October 12, 2011 Published: October 26, 2011 2395

dx.doi.org/10.1021/bc200052h | Bioconjugate Chem. 2011, 22, 2395−2403

Bioconjugate Chemistry

Article

Table 1. Amino Acid Sequences of Z5BPA and Z18BPAa ZF5BPA (denoted Z5BPA) VENKXNKEQQNAFYEILHLPNLNEEQRNAFIQSLKDDPSQSANLLAEAKKLNDAQAPKMtt ZH18BPA (denoted Z18BPA) VENKFNKEQQNAFYEILXLPNLNEEQRNAFIQSLKDDPSQSANLLAEAKKLNDAQAPKMtt a

Underlined amino acids were coupled twice during synthesis. X = benzoylphenylalanine.

domain has been made. In the N-terminal of the Z domain, an alanine residue was replaced by valine. A glycine-to-alanine substitution was made for the removal of a hydroxylamine cleavage site,12 which also resulted in loss of binding to the Fab region.11,13 The Z domain is small (6.7 kDa), easy to produce, and has a stable three-dimensional structure and also the capacity to refold.14 It was previously proven to be suitable for chemical peptide synthesis, thereby making the introduction of synthetic active groups possible, extending the usability of the domain.15 Benzoylphenylalanine (BPA) is a synthetic amino acid that can be incorporated in a peptide during synthesis. Benzophenone (BP), which is part of BPA, is a photoreactive group that forms covalent bonds to other amino acids upon UV-exposure. BPA is considered to be efficient, stable, and also easy to handle, 16 and it is primarily used when mapping protein−ligand interactions. When mapping interactions, the strategy is to produce variants of a protein with BPA incorporated at different positions, and then allow the protein to bind its interaction partner.17 When subjecting the complex to UV light, BPA forms a diradical, which renders the generation of a covalent bond between the protein and its interaction partner possible. The utilization of an antibody-binding protein domain with an incorporated photoreactive group for attachment to antibodies has previously been reported by Jung et al. By using a modified C2 domain from streptococcal Protein G and covalently attaching BP at positions 21 and 29 via two incorporated cysteines, a photoreactive procedure could be used for modification of the antibodies. It was demonstrated that this approach could be used for directed antibody immobilization onto a surface. The strategy included recombinant production of the domain and by site-directed mutagenesis two cysteines were incorporated. The cysteines were labeled by maleimide chemistry for attachment of the photoreactive group, BP, into the domain.18 In this study, we report the development of a stringent and effective method for specific covalent labeling of immunoglobulins. By the use of a synthetic Z domain with the photoreactive probe BPA incorporated in the amino acid sequence, covalent conjugation to the antibody has been achieved. In this report, SPPS has been used as a means for production, and in this manner, we are able to incorporate the BPA as an unnatural amino acid specifically in the peptide backbone. Thereby, production and modification could be made in a single manufacturing process. Moreover, a detection-handle, biotin, could be incorporated in a specific position into the protein. By combining the inherent affinity of the Z domain and the Fcfragment with the ability of BPA to create a covalent bond, specifically labeled antibodies were successfully achieved, characterized, and tested in different platforms.

Several antibodies used were kind gifts; FITC-BSA-specific antibody (human IgG1 and mouse IgG2a, BioInvent International AB) and antibody specific for His6-ABP (polyclonal, Atlas Antibodies AB). Production of Peptides. The two variants ZF5BPA (Z5BPA) and ZH18BPA (Z18BPA) were produced by solidphase peptide synthesis using Fmoc/t Bu protection strategy. The syntheses were performed on a 433 A Peptide synthesizer (Applied Biosystems) using an acid-labile Fmoc-amide resin (substitution 0.67 mmol g−1, Applied Biosystems). Cleavage of the Fmoc group was performed using 20% (v/v) piperidine in NMP. The coupling reactions were performed with a 10-fold molar excess of amino acid activated with 2-(1H-benzotriazole1-yl)-1,1,3,3-tetramethyluronium hexafluorophosphate (HBTU) and 1-hydroxybenzotriazole (HOBt) (both from Iris Biotech GmbH) and DIEA in NMP. Apart from the amino acids underlined in Table 1, single couplings were carried out. Unreacted amino groups were capped with acetic anhydride. Standard side-chain protecting groups were used except in position 58, where Fmoc-Lys(Mtt)-OH (NovaBiochem) was introduced. The photactivable probe Fmoc-BPA-OH (Peptech Corporation) was incorporated in position 5 in Z5BPA and in position 18 in Z18BPA. Incorporation of biotin in Z5BPA and Z18BPA was performed with the peptides still bound to the resin. The peptide-resin was treated with TFA/TIS/DCM (1:5:94) for the removal of the 4-methyltrityl (Mtt) group protecting the ε-amine of Lys58. Biotin was coupled with 5 equiv of D-biotin (Sigma), activated with HBTU/HOBt/DIEA in NMP for 2 × 1 h, with the reactions monitored using Kaiser test.19 The removal of protecting groups and release from the resin were achieved by treatment with TFA/TIS/H2O (95:2.5:2.5) for 2 h at room temperature. The peptides were extracted in H2O/tert-butyl methyl ether (1:1) and lyophilized. HPLC and MS. The products from the syntheses of Z5BPA and Z18BPA were analyzed and purified by RP-HPLC, using a Silica-C18 column with 3.5 μm particle size and 4.6 × 150 mm length (Agilent Technologies). A flow rate of 0.9 mL min−1 and a gradient of solvent B (0.1% TFA/CH3CN, vol/vol) in solvent A (0.1% TFA/H2O, vol/vol) were used. Fractions obtained from RP-HPLC were analyzed by MS, and the correct products were pooled and lyophilized. The protein concentration of the samples was determined by amino acid analysis (Aminosyraanalyscentralen). Mass spectrometry was used to verify that correct products were obtained. The products from the syntheses of Z5BPA and Z18BPA were analyzed by ESI-MS, performed on a Q-TOF II (Waters Corporation, Micromass MS Technologies). For Z5BPA-bio and Z18BPA-bio, the fractions collected in RPHPLC were analyzed using a MALDI-MS Biflex IV (BRUKER Daltonics). As reference and for external calibration myoglobin, carbonic anhydrase II (CA II) and insulin (all obtained from Sigma-Aldrich) were used. General Procedure for Photoconjugation. 100 nM antibodies and 1 μM Z5BPA-bio or Z18BPA-bio in PBS were incubated in 20 °C for 1 h. Cross-linking was achieved by exposure to light, 365 nm (Spectronics Corporation) for

■

MATERIAL AND METHODS General. In this study, recombinantly produced Zwt was used as a reference.12 The protein domain was randomly biotinylated resulting in an average of four biotin moieties/ domain, as measured by MALDI-MS (data not shown). 2396



Article

45 min, or 1 h on ice. For the analysis of conjugated antibodies by Biacore and Western Blot, Fc and Fab fragments as well as full-length IgG were used. 1.3 μM of the antibody/antibody fragment was mixed with 26 μM Z5BPA-biotin in PBST and incubated at room temperature for 1 h. The antibody solutions were exposed to light at 365 nm for 2 h on ice. For buffer exchange spin concentrators with a 10 kDa cutoff (Vivaspin, Sartorius Stedim Biotech) were used, centrifugation at 15 000 rcf for 10 min. Buffers used were 0.2 M HAc (VWR), pH 3.2, for lowering the pH and PBST for restoring the pH to 7. The protein content was analyzed by measuring absorbance at 280 nm (280 nm, ε = 210 00 M−1 cm−1). Luminex. Bead Coupling. Proteins and antibodies were coupled to beads according to the manufacturer’s recommendation (COOH Microspheres, Luminex Corporation). 1.6 μg antibody to FITC-BSA was used for coupling to approximately 5 × 105 beads. Protein FITC-BSA (BioInvent International AB) and His6-ABP (Atlas Antibodies AB) were coupled using 2 μg for approximately 2.5 × 105 beads and 10 μg for 1 × 106 beads, respectively. Coupled beads were kept at 4 °C in a storage buffer (Blocking Reagent for ELISA, Roche Applied Science). Analysis of Photoconjugated Antibodies. Antibodies subjected to photoconjugation were diluted with PBST to a concentration of 100 nM. Filter plates (0.45 μm MSHVN45 MultiScreen HTS, Millipore) were used for incubation of 45 μL diluted antibody with 5 μL bead solution (200 beads/μL) at 23 °C with mixing for 1 h. Washing with 3 × 50 μL PBST was performed before PhycoLink Streptavidin-R-Phycoerythrin (2.2 μg mL−1) (Prozyme) was supplied. After incubation (23 °C for 20 min) and washing (3 × 50 μL PBST), the fluorescence was measured with Luminex Lx200. When performing the sandwich assay, the first incubation with beads was made with target protein at 23 °C for 1 h; thereafter, a wash step (3 × 50 μL PBST) was introduced, followed by the procedure described above. SPR-Analysis. Analysis of the binding kinetics of Z5BPA and Z18BPA to different IgG molecules was performed by the use of SPR technology (BIAcore 2000 instrument, Biacore). Antibodies and human serum albumin (HSA) were immobilized onto a CM5 sensor chip resulting in approximately 2000 response units (RU) for the antibodies, and 700 RU for HSA. A flow rate of 30 μL min−1 at 25 °C was used during the analysis. HBS-EP (HEPES 100 mM, NaCl 1.5 M, EDTA 34 mM (Merck) and 0.05% (v/v) surfactant p20 (VWR)) was used as a running buffer, and for regeneration of surfaces 10 mM HCl. The proteins were analyzed in concentrations ranging from 1.6 nM to 203.5 nM for Z5BPA-bio, 11.9 nM to 1560 nM for Z18BPA-bio, and 5.0 nM to 635 nM for Zwt. All samples were run in duplicate. The software BIAevaluation 3.2 (BIAcore AB) was employed to determine the dissociation constants based on the Langmuir 1:1 model. The antigen His6-ABP was immobilized onto a CM5 sensor chip resulting in 280 response units (RU). Polyclonal rabbit IgG targeting the HisABP-construct was subjected to the photolabeling procedure either with photoreactive Z5BPAbiotin or without the reagent, both performed in duplicate. The samples were diluted to a concentration of 100 nM and flown over the Biacore surface with the antigen for analysis of the ability of the antibody to bind the antigen. The resulting Biacore curves were referenced against a blank chip surface. To obtain KD values for the photolabeled and unlabeled antibodies, GraphPad Prism was used for nonlinear regression analysis of binding data from the Biacore. The curve fits were calculated

from an average curve of duplicate samples, conjugated and unconjugated antibodies, respectively, flown over the chip surface. Western Blot. Protein or antibodies (2−3 μg) were separated on SDS-PAGE gradient gels (NuPAGE 4−12% BisTris SDS-PAGE, NuPAGE 3−8% Tris-Acetate (Invitrogen) or Criterion 10−20% Tris-HCl (Bio-Rad) under reducing conditions, followed by transfer to PVDF membranes (Invitrogen or Bio-Rad) according to the manufacturer’s recommendations. Membranes were soaked in methanol and blocked (0.5% casein in 1×PBST vol/vol) for 1 h at 20 °C. To directly detect biotinylated proteins, the membrane was incubated with peroxidase-conjugated streptavidin (diluted 1:70 000, DakoCytomation). For detection of specific proteins, the membranes were first incubated with antibodies (of rabbit, goat species, respectively) specific for HisABP or human albumin for 1 h, followed by washing (1×PBST). Detection was made possible by either peroxidase-conjugated streptavidin (diluted 1:5000, DakoCytomation) or peroxidase-conjugated IgG directed to goat antibodies (diluted 1:100 000, SigmaAldrich). Detection was carried out with Immobilion Western Chemiluminescent HRP substrate (Millipore) according to the manufacturer.

■

RESULTS Design and Synthesis of Z-Variants. The procedure of labeling and covalent attachment of reporter groups to antibodies is traditionally done through the chemistry of amine or carboxyl groups in the antibody. This is an efficient method, but since the labeling occurs randomly, many different groups in the protein have the possibility to react and the conjugated group might influence the binding of the antibody to its antigen. Hence, an optimization of the labeling conditions is a necessity for every antibody, and therefore, an alternative method for specific and directed labeling would be beneficial for many applications. Directed labeling would also be beneficial for antibodies in complex environment, where other proteins surround the antibody. Here, we report a novel method that enables a specific labeling of antibodies where a photoactivable probe together with a specific binding event is utilized to achieve covalent and specific attachment to immunoglobulins. To achieve a specific and reliable covalent labeling, the IgG-binding domain Z was utilized. After studying the structural data available on Protein A (domain B) and its interaction surface with IgG, the positions Phe5 and His18 were chosen to be exchanged for BPA. Phenylalanine 5 is positioned in the N-terminal part of the domain, close to the first helix, and it is also claimed to participate in the binding between the B-domain and IgG (Deisenhofer 1981). The other position, histidine 18, is the last amino acid in the first helix, and it is not suggested to be part of the binding to IgG, but is situated in close proximity to the binding surface. Hence, two variants of the Z domain were produced, Z5BPA and Z18BPA. To achieve specific incorporation of the photoreactive probe, the two molecules were produced using solid-phase peptide synthesis. In Table 1, the sequences of the synthesized molecules are shown, where underlined amino acids were coupled twice during synthesis to ensure a high yield of the product. In both Z5BPA and Z18BPA, a D2E substitution was made to avoid aspartimide formation during synthesis. This substitution had previously been shown not to interfere with the structural or functional behavior of the Z domain.15

2397



Article

The synthesis products, Z5BPA and Z18BPA, were purified using RP-HPLC, and the purified correct products were verified using MS (data not shown). To enable easy and flexible detection and to be able to analyze the efficiency of the conjugation, the two variants were specifically biotinylated in at the C-terminus of the protein domain. This was made by applying an orthogonal strategy, through the incorporation of a lysine protected with a 4-methyltrityl (Mtt) group in the last position in the sequence of the domains. Also, these protein products were successfully purified to homogeneity and analyzed by MS (Figure 1a and b). Analysis of the Binding Characteristics of the ZVariants. The ability of the Z domain to refold after synthesis has been shown previously,15 and therefore, retained binding to IgG was expected. However, with the introduction of BPA in different positions in the proximity of, or in the binding surface, there is a potential risk of influencing the ability to bind IgG. Therefore, an analysis of the binding kinetics of the two Z variants using surface plasmon resonance (SPR) was made. The ability of the two novel Z variants to bind to different IgG molecules was analyzed and compared to the parental Z domain. The analysis revealed an affinity of Z5BPA-bio/to IgG complex comparable to the parental Z/IgG. However, a considerably lower affinity was detected when analyzing the binding of Z18BPA-bio to IgG (Table 2). Evaluation of Covalent Coupling of Z-Variants to Immunoglobulins. For the evaluation of covalent coupling of the two modified Z domains to IgG, two approaches were employed. In the first approach, polyclonal rabbit IgG was coupled to Luminex beads and then incubated with the synthetic Z domains, followed by light-induced activation of the covalent coupling (Figure 2A, setup 1). Thereafter, the beads were washed with a low-pH buffer in order to break all noncovalent interactions. By comparing the biotinylation of beads washed with low pH, with beads washed with neutral buffer, we were able to show that Z5BPA-bio could be covalently attached to IgG. Different concentration of the participating molecules and also different times for the UV light activation was used, and it was concluded that the most effective ratio between IgG and Z was 1:10 (IgG:Z). Moreover, the time for illumination of the reaction vessel was set to 1 h. Also, data show that the variant Z18BPA-bio was not covalently linked to IgG by this treatment (Figure 2B). To ensure that this behavior was not concentration-dependent, a higher concentration of Z18BPA-bio was used (1:500, IgG:Z), but still no covalent coupling was obtained (data not shown). The successful conjugation of Z5BPA-bio to IgG was further confirmed in the second approach, where cross-linking was performed in solution and the photoconjugated antibodies were evaluated in the Luminex system. In this experiment, Z5BPAbio and IgG were incubated in solution and subjected to UV light for cross-linking. To remove excess unbound Z molecules, a spin filter column (10 kDa cutoff) was used. This also allows for buffer exchange and lowering of pH, which enable release of noncovalently bound Z molecules from IgG. Hence, the bound and cross-linked Z domains will stay attached to the IgG molecules, while noncovalently bound Z domains will be released and washed away. In order to investigate the efficiency of covalent coupling of the bound Z5BPA-bio to the antibody, a sample of the photoconjugated antibodies was taken out before lowering the pH. Also, as negative control recombinantly produced and biotinylated Z was used. For evaluation of the covalent conjugation, antigens were linked to Luminex beads

Figure 1. Mass spectrometry analysis of the synthetic Z domains. (A) The mass spectrum of Z5BPA-bio showing a molecular weight of 6981 Da (theoretical molecular weight: 6984 Da). (B) The mass spectrum of Z18BPA-bio showing a molecular weight of 6991 Da (theoretical molecular weight: 6994 Da).

and subsequently incubated with corresponding IgG molecules from the conjugation experiment (Figure 2A, setup 2). In the negative control, biotinylated Z, the signal diminishes when 2398



Article

the paratope of the antibody is intact. Evaluation of the conjugation efficacy was made by comparing the achieved signal before and after treatment with low pH. Therefore, the conclusion that more than 80% of the bound Z-molecules were efficiently cross-linked to the immunoglobulins could be drawn (Figure 2C). Specificity of Labeling. Since many manufactured antibodies are stabilized by the addition of other proteins, commonly albumin, it is important to be able to perform the conjugation in complex solutions. To investigate the selectivity of the labeling, a solution containing eight times more HSA than specific monoclonal antibodies was prepared. The efficiency of the conjugation was analyzed and compared with conjugation without HSA (assay number 1 in Figure 3A). The efficiency of the labeling is not affected by the presence of HSA (Figure 3B). To assess any unwanted biotinylation of HSA, a Western blot experiment was performed. In the first experiment, all biotinylated proteins were detected (Figure 3C, left panel), and in the second setup, the present HSA molecules were detected (Figure 3C, right panel). This experiment indicates that no covalent coupling of the biotinylated Z-domains to HSA has occurred. However, since the migration of HSA in SDS-PAGE is rather similar to the heavy chain of IgG in complex with the Z domain, another experiment was performed. This time, the biotinylation was assessed using Luminex analysis. An HSA-binding protein was covalently attached to Luminex beads. Thereafter, the UV-exposed IgGZ5BPA-bio or HSA-Z5BPA-bio mixture was mixed with the beads. Now, HSA or IgG is able to bind to the beads, by using either the covalent coupled protein’s inherent ability to bind HSA or the conjugated antibody’s ability to bind to the antigen. Thereafter, biotinylation was assessed through fluorescently labeled streptavidin. By utilizing the ability of the antibodies to selectively bind to the antigen-conjugated beads (assay number 1 in Figure 3A), high signals from the antigen binding IgGmolecules could be detected. Since similar signal intensity was detected both before and after washing with low pH, we could conclude that the conjugation of Z5BPA-bio to the antibodies was very effective. Furthermore, the conclusion that no biotinylation of HSA was obtained could be drawn since no signal was detected from the HSA-binding beads mixed with fluorescently labeled streptavidin (assay number 2 in Figure 3A). However, the ability of HSA to bind to the beads was confirmed by HSA-recognizing antibodies conjugated with biotin (setup 3 in Figure 3A,D). These results also show that the covalent attachment of Z to IgG is specific since no biotinylation of HSA could be detected. To investigate the influence of labeling on the antigen binding capacity of the antibody an SPR-analysis of conjugated and unconjugated antibodies, binding was performed. As can be seen in Figure 4A−C, conjugated and unconjugated antibodies show the same affinity to its antigen, which confirms that the antigen-binding site is preserved after complete conjugation (Figure 4). The similar KD values obtained for conjugated and unconjugated, 6.0 nM and 6.4 nM, respectively, clearly show that the antigen site of the antibody remains unaffected after being subjected to the conjugation process (Figure 4B,C). To further show the specificity of the conjugation, Z5BPAbiotin was allowed to react with full-length antibodies, Fc fragments, and Fab fragments. The obtained conjugations were analyzed by Western blot. Figure 5 shows that conjugation occurs at the Fc part of the antibody, since no biotinylation could be detected on the Fab fragments. However, biotinylation

Table 2. KD values for Z5BPA-bio, Zwt, and Z18BPA-bio Binding to Human IgG1 and Mouse IgG2a Human IgG1 Mouse IgG2a Rabbit IgG poly

Z5BPA-bio

Z18BPA-bio

Zwt

10 nM 650 nM 30 nM

60 μM -

20 nM 550 nM 60 nM

Figure 2. Covalent conjugation of the Z-variants was evaluated by using the Luminex platform. (A) The two different strategies used for analysis of covalent coupling are shown. In strategy 1, the antibodies were covalently coupled to Luminex beads. Thereafter, the beads were incubated together with the different Z variants and illuminated with UV-light. After washing, the biotinylation was detected by fluorescently labeled streptavidin. In strategy 2, the IgG and Z-variants were mixed and illuminated in solution. After illumination, the protein mixtures were washed, with a buffer with pH 3.2 or 7.0. To enable this, spin concentrators were used. After further washing and increasing the pH to neutral, the antibody solution was mixed with Luminex beads with already covalently attached antigens. Also, here the biotinylation was detected by fluorescently labeled streptavidin. (B) The crosslinking of Z5BPA-bio and Z18BPA-bio to polyclonal rabbit IgG. As negative control, uncoupled beads were used. Strategy 1 in (A) was used. Mean value of two different experiments is shown.

lowering the pH; hence. the noncovalent interaction between Z and IgG is possible to break with low pH (Figure 2C). Moreover, Z5BPA-bio is efficiently linked to the IgG molecules since the IgG molecules after treatment with low pH still give a strong signal. Also, the ability to bind the antigen indicates that 2399



Article

Use of Conjugated Antibodies. The photoconjugated antibodies were used in a sandwich assay setup where capture antibodies were coupled to beads and incubated with target protein in various concentrations (Figure 5A). After thorough washing, target-specific photoconjugated antibodies were added. The beads were washed, and the fluorescence from streptavidin-R-phycoerythrin was detected by using the Luminex platform. Analyses using both human monoclonal antibodies and mouse monoclonal antibodies recognizing FITC-BSA were successfully performed, and the antigen could be detected down to a concentration of 0.1 ng/mL within a concentration window spanning 4 orders of magnitude (Figure 5). Also, the conjugated antibodies were used for detection in a Western blot assay. By conjugation of the antibodies with Z5BPA-bio, the protein targeted by the antibodies could be detected by streptavidin-HRP. In Figure 5C, two Western blot membranes are shown. On the rightmost membrane, proteins are detected with covalently conjugated antibodies, and on the left membranes, as a negative control, the same antibodies were used, but in the conjugation step, Z-Bio, lacking the BPA group, was used. The molecular weights of the protein bands detected are as expected; hence, the system works both in a sandwich setup with the antigen in solution and in detection of proteins bound to a membrane.

■

DISCUSSION

To be able to covalently attach the Z domain to immunoglobulins, the structural information on the binding site was thoroughly studied. Two amino acids in close proximity to the binding surface were exchanged for the photoactivable probe, BPA. Both suggested Z variants were successfully synthesized with high yield and the C-terminal biotinylation was efficiently and specifically made through an orthogonal protection strategy (Figure 1). The interaction between IgG and the two synthesized Z variants was analyzed revealing a retained affinity for the Z5BPA-bio molecule, while Z18BPA-bio showed very low affinity (∼100 μM, Table 2). For the Z5BPA-bio molecule, both the on- and off-rates are in the same range as for the parental Z domain (data not shown). The retained affinity of the Z5BPA-bio variant could be explained by steric similarity of phenylalanine and BPA making the inherent structure of Z intact. The addition of the extra benzoyl group seems to fit well between the two molecules upon binding. On the other hand, replacement of histidine 18 for BPA is deleterious and destroys the interaction with IgG. This could be due to the change of charge in the position of amino acid 18. Also, the larger side chain of the unnatural amino acid could stericallly inhibit the ability to bind IgG. This photoactivable molecule has previously been used to covalently label antibodies through incorporation in an antibody binding molecule from protein G,18 but a different strategy for production of the binding domain was used. In this publication, they have recombinantly produced the IgG-binding domain and after purification covalently attached a BPA-group via a cysteine. Moreover, the area of application in the referred study was to covalently attach the antibodies onto a solid support in an oriented way. In this project, we have focused on labeling of the antibodies for detection purposes. A very important characteristic of a molecule used for selective labeling is the efficiency of the covalent linking. Hence, a thorough characterization of this was performed. Different IgG molecules were used for this analysis, and in

Figure 3. Degree of covalently linked Z5BPA-bio domain in a complex sample was analyzed. Samples of antibodies to be photoconjugated with Z5BPA-biotin were supplemented with 800 nM HSA (Albumina Kabi). (A) Three different strategies used for analysis of covalent coupling are shown. An HSA-binding protein was covalently attached to the Luminex beads. These beads could be used in three different ways. According to strategy 1, the attached protein is used as an antigen, since the labeled antibody is directed to this protein. In strategy 2, the protein is used to bind to HSA. In these two experiments, streptavidin-R-phycoerythrin is used to detect any biotinylation. In strategy 3, antibodies with biotin are used to detect the bound HSA. (B) Degree of cross-linking was analyzed both with and without HSA in the sample. Covalent linking of Z5BPA-bio was unaffected by the HSA present in the sample. Setup according to strategy 1 in Figure (A) was used. (C) To assess if Z5BPA-bio was able to covalently bind to the HSA molecule, a Western blot was performed. In the left panel, the amount of covalently linked Z5BPAbio in three different samples was analyzed using streptavidin-HPR. No attachment to the HSA molecule could be detected. In the right panel, the presence of HSA in the samples was analyzed by using goat antibodies targeting HSA, visualized with HRP-conjugated antigoat antibodies. (D) To confirm the conclusion from the previous experiment, a Luminex experiment was made where both the amount of biotinylated IgG molecules and HSA molecules were analyzed using strategy 1 in (A). The number of biotinylated IgG molecules was shown to be high, both after and before treatment with low pH. This indicates a high degree of covalent coupling by Z5BPA-bio. The amount of biotinylated HSA-molecules was analyzed using strategy 2 in (A). No biotinylation could be detected before or after treatment with low pH. To ensure the presence of HSA, a third strategy was used (number 3 in (A)); hence, detecting HSA with an HSA-binding antibody and thereby the presence of HSA was confirmed.

was detected on both the full-length antibodies and the Fc fragments (Figure 4D). 2400



Article

Figure 4. Investigation into the performance of an antibody was made by analyzing the affinity constant for antigen binding before and after conjugation. Moreover, specificity of the conjugation was assessed by Western blot analysis. (A) SPR analysis of antibodies targeting their antigen was performed, before and after conjugation. Antigen was immobilized on the chip and the antibodies were flowed over the surface. Average of duplicate samples is shown. Nonlinear regression analysis were made of the binding curve of (B) photolabeled antibodies giving a KD value of 6.0 nM and (C) control antibodies giving a KD value of 6.4 nM. Hence, the labeling procedure does not influence the performance of the antibody. (D) Western Blot analysis of duplicate samples of conjugated and unconjugated full-length antibodies (lanes 1−4), Fc fragments (lanes 5−8), and Fab fragments (lanes 9−12). Z5BPA conjugates specifically to the Fc part of the antibodies.

streptavidin using conjugated antibodies from both mouse and human. Also, an ordinary Western blot was made showing that here also the conjugated antibodies successfully detect the anticipated proteins on the membrane. These data show that the covalent linkage between IgG and Z5BPA-bio is stable and usable in different well-known assays. When performing the sandwich assay, antigen concentrations down to 0.1 ng/mL could be detected. The conjugated antibodies in this study could possibly have one or two biotin on each antibody since only one biotin is incorporated in each Z domain. Hence, by introducing more than one biotin or even by directly introducing a number of fluorescent probes in the Z domain the detection level will likely be improved. The comparison of the ability of conjugated and unconjugated antibodies to bind their antigen by SPR analysis shows that the procedure allows the antibody to retain its binding to the antigen and that the covalently attached photoreactive Z5BPA-biotin does not influence the reactivity of the antibody (Figure 4). The specificity of the labeling is further demonstrated by the Western blot experiment where biotin is detected only if the antibody or antibody fragment is covalently conjugated. The data clearly show that conjugation is achieved on the full-length antibody and also on Fc fragment. However, no conjugation could be detected on the Fab fragment (Figure 4D). Here, we have presented a stringent and effective method for labeling of antibodies by utilizing an IgG-binding protein domain, Z. By introducing a photoactivable group in the

the experiment, the amount of covalently attached Z-domains were compared with the amount of Z-domains able to bind to the IgG molecules. All immunoglobulins with affinity for Z were successfully covalently labeled, with an efficiency of more than 80% when the ratio 1:10 (IgG:Z) was used (Figure 2). When using the coupling strategy in complex solution, we could conclude that no unwanted covalent labeling could be detected, despite high concentration of both HSA and Z5BPAbio (Figure 3). This is of utmost importance and shows that the benzophenone (BP) group needs to be in very close proximity to create a covalent link, thereby ensuring specific and directed covalent attachment. Hence, specific and stable interaction as well as UV light of correct wavelength is needed. The necessity of UV light makes this photoactivable probe convenient and easy to handle since no light protection is needed during the synthesis, purification, or other experimental steps where no conjugation is desired. Moreover, if a covalent bond not is created, the probe relaxes to the ground state after excitation and thereby the uncoupled Z molecules are possible to use in a next round of labeling (data not shown). To assess the functionality of the covalently linked antibodies, three different methods were used. A sandwich assay was successfully designed by taking advantage of capturing antibodies covalently linked to Luminex beads. The prepared beads were incubated with different concentrations of antigen (Figure 5A). Hence detection could be made by 2401



Article

Figure 5. Covalently cross-linked antibodies were used in two common assays. (A) Sandwich analysis where the capturing antibodies were attached to Luminex beads. (B) Incubation of beads with different concentrations of antigen and subsequent washing. The amount of captured antigen was analyzed by using detection antibodies covalently linked to Z5BPA-bio in combination with streptavidin-R-phycoerythrin. The diagram to the left shows human monoclonal IgG1 targeting FITC-BSA, and in the right diagram, mouse monoclonal IgG2a is used; the mean values of three experiments are shown. (C) Covalently cross-linked antibodies were used in a Western blot analysis. In the left panel, detection was made by using an antibody covalently linked to Z5BPA-bio, and in the right panel, the same antibody but with Zwt-bio was used as control. Protein samples on the gel are as follows: MW, a molecular weight marker; lane 1, His6ABP (19 kDa); lane 2, His6ABP-PrEST (25.5 kDa).

■

protein scaffold during synthesis, a covalent linkage between IgG and the synthesized protein domain can be created with very high efficacy. This linkage has been shown to be efficiently formed and stable under different conditions. Here, we have been using biotin as the reporting group, but a large variety of different groups could be introduced in the protein domain to tailor-make the antibodies for a certain purpose. The new approach for labeling of antibodies presented here is both flexible and reliable and would be suitable for a wide range of applications where antibodies are used in the capture or detection step.

AUTHOR INFORMATION

Corresponding Author *Tele: +46 5537 8330. Fax: +46 5537 8481. E-mail: sophia.hober@ biotech.kth.se.

■

ACKNOWLEDGMENTS The authors would like to thank P-Å. Nygren and J. Schwenk for fruitful discussions. G. Sundqvist is acknowledged for assistance with the MS analyses. The ProNova VINN Excellence Centre for Protein Technology financially supported 2402



Article

this work with Atlas Antibodies AB, BioInvent International AB, GE Healthcare Bio-Sciences AB, Gyros AB, Mabtech AB and Olink AB as active partners in the project.

■

REFERENCES

(1) Borrebaeck, C. A. (2000) Antibodies in diagnostics - from immunoassays to protein chips. Immunol. Today 21, 379−82. (2) Brinkley, M. (1992) A brief survey of methods for preparing protein conjugates with dyes, haptens, and cross-linking reagents. Bioconjugate Chem. 3, 2−13. (3) Bjorck, L. (1988) Protein L. A novel bacterial cell wall protein with affinity for Ig L chains. J. Immunol. 140, 1194−7. (4) Bjorck, L., and Kronvall, G. (1984) Purification and some properties of streptococcal protein G, a novel IgG-binding reagent. J. Immunol. 133, 969−74. (5) Forsgren, A., and Sjoquist, J. (1966) ″Protein A″ from S. aureus. I. Pseudo-immune reaction with human gamma-globulin. J. Immunol. 97, 822−7. (6) Langone, J. J. (1982) Applications of immobilized protein A in immunochemical techniques. J. Immunol. Methods 55, 277−96. (7) Ståhl, S., Nilsson, J., Hober, S., Uhlén, M., and Nygren, P.-Å. (1999) in The Encyclopedia of Bioprocess Technology: Fermentation, Biocatalysis and Bioseparation (Fleckinger, M. C., and Drew, S. W., Eds.) pp 8−22, John Wiley and Sons Inc., New York. (8) Uhlén, M., Guss, B., Nilsson, B., Gatenbeck, S., Philipson, L., and Lindberg, M. (1984) Complete sequence of the staphylococcal gene encoding protein A. J. Biol. Chem. 259, 1695−1702. (9) Deisenhofer, J. (1981) Crystallographic refinement and atomic models of a human Fc fragment and its complex with fragment B of protein A from Staphylococcus aureus at 2.9- and 2.8-A resolution. Biochemistry 20, 2361−70. (10) Gouda, H., Shiraishi, M., Takahashi, H., Kato, K., Torigoe, H., Arata, Y., and Shimada, I. (1998) NMR study of the interaction between the B domain of staphylococcal protein A and the Fc portion of immunoglobulin G. Biochemistry 37, 129−36. (11) Graille, M., Stura, E. A., Corper, A. L., Sutton, B. J., Taussig, M. J., Charbonnier, J.-B., and Silverman, G. J. (2000) Crystal stucture of a Staphylococcus aureus protein A domain complexed with the Fab fragment of a human IgM antibody: Structural basis for recognition of B-cell receptors and superantigen activity. Proc. Natl. Acad. Sci. 97, 5399−5404. (12) Nilsson, B., Moks, T., Jansson, B., Abrahmsen, L., Elmblad, A., Holmgren, E., Henrichson, C., Jones, T. A., and Uhlen, M. (1987) A synthetic IgG-binding domain based on staphylococcal protein A. Protein Eng. 1, 107−13. (13) Jansson, B., Uhlén, M., and Nygren, P.-Å. (1998) All individual domains of staphylococcal protein A show Fab binding. FEMS Immunol. Med. Microbiol. 20, 69−78. (14) Ståhl, S., and Nygren, P. A. (1997) The use of gene fusions to protein A and protein G in immunology and biotechnology. Pathol. Biol. (Paris) 45, 66−76. (15) Engfeldt, T., Renberg, B., Brumer, H., Nygren, P. A., and Karlstrom, A. E. (2005) Chemical synthesis of triple-labelled threehelix bundle binding proteins for specific fluorescent detection of unlabelled protein. ChemBioChem 6, 1043−50. (16) Dorman, G., and Prestwich, G. D. (2000) Using photolabile ligands in drug discovery and development. Trends Biotechnol. 18, 64−77. (17) Kawamura, A., Hindi, S., Mihai, D. M., James, L., and Aminova, O. (2008) Binding is not enough: flexibility is needed for photocrosslinking of Lck kinase by benzophenone photoligands. Bioorg. Med. Chem. 16, 8824−9. (18) Jung, Y., Lee, J. M., Kim, J. W., Yoon, J., Cho, H., and Chung, B. H. (2009) Photoactivable antibody binding protein: site-selective and covalent coupling of antibody. Anal. Chem. 81, 936−42. (19) Kaiser, E., Colescott, R. L., Bossinger, C. D., and Cook, P. I. (1970) Color test for detection of free terminal amino groups in the solid-phase synthesis of peptides. Anal. Biochem. 34, 595−8. 2403


Covalent Immunoglobulin Labeling through a Photoactivable

Recommend Documents