Covalent and Oriented Immobilization of scFv Antibody Fragments via

Dec 6, 2012 - ‡Microbiology and §Network of Excellence for Functional Biomaterials (NFB), NUI Galway, University Road, Galway, Ireland...
0 downloads 0 Views 2MB Size
Article pubs.acs.org/Biomac

Covalent and Oriented Immobilization of scFv Antibody Fragments via an Engineered Glycan Moiety Xuejun Hu,*,† María J. Hortigüela,‡ Sylvain Robin,‡ Heng Lin,† Yajie Li,† Anthony P. Moran,‡ Wenxin Wang,§ and J. Gerard Wall*,‡,§ †

Medical College, Dalian University, Xuefu Avenue No.10, Dalian Economical and Technological Development Zone, Liaoning 116622, China ‡ Microbiology and §Network of Excellence for Functional Biomaterials (NFB), NUI Galway, University Road, Galway, Ireland S Supporting Information *

ABSTRACT: Antibody-derived fragments have enormous potential application in solid-phase assays such as biomarker detection and protein purification. Controlled orientation of the immobilized antibody molecules is a critical requirement for the sensitivity and efficacy of such assays. We present an approach for covalent, correctly oriented attachment of scFv antibody fragments on solid supports. Glycosylated scFvs were expressed in Escherichia coli and the C-terminal, binding pocket-distal glycan tag was oxidized for covalent attachment to amine-functionalized beads. The glycosylated scFvs could be immobilized at salt concentrations that precluded nonspecific adsorption of unglycosylated molecules and the covalently attached antibody fragments exhibited 4-fold higher functional activity than ionically adsorbed scFvs. The glyco-tethered scFvs were stable in NaCl concentrations that removed greater than 90% of adsorbed scFvs and they exhibited improved stability of antigen binding over both adsorbed scFvs and soluble, nonimmobilized scFvs in accelerated degradation tests. The simple expression and immobilization approach reported is likely to find broad application in in vitro antibody tests.



or to introduce novel functions that create fit-for-purpose molecules.6−8 The immobilization of scFvs on solid surfaces is a key step in many scFv-based immunotechniques. The binding activity of scFvs, and antibody molecules in general, is often impaired, however, upon direct immobilization onto solid supports.9 Controlled orientation of the antibody molecules is a critical requirement, therefore, for the sensitivity and efficacy of solidphase assays.10,11 Ideally, the immobilization method should yield molecules that are correctly oriented on the substrate, stably (preferably covalently) attached for enhanced performance duration and to allow regeneration and reuse in, for example, protein purification, form a monolayer and are unmodified by the reaction chemistry used to generate the linkage. Antibody immobilization via simple adsorption

INTRODUCTION ScFv antibody fragments are one of smallest recombinant antibody fragments that typically retain the binding properties of their parent antibody molecules.1 They are composed of the antibody variable heavy (VH) and light (VL) chain domains, covalently joined by a synthetic polypeptide linker to yield a monovalent fragment 25−30 kDa in size and of approximate dimensions 5 × 4 × 4 nm, compared with 150 kDa monoclonal antibodies of 15 × 7 × 4 nm. ScFvs can be obtained by gene cloning from hybridomas or by selection from large synthetic, ̈ or immune libraries that mimic the human immune naive, repertoire.2 They are the most widely exploited antibody fragments in bioengineering applications and their small size allows dense packing of ligand-binding pockets on surfaces, leading to potential for improved detection sensitivities or purification efficiencies.3−5 A key advantage of using scFvs in immunoassays and protein purification applications is that they can be inexpensively produced in high yields in Escherichia coli, as well as easily engineered to improve their binding properties © 2012 American Chemical Society

Received: September 27, 2012 Revised: November 20, 2012 Published: December 6, 2012 153

dx.doi.org/10.1021/bm301518p | Biomacromolecules 2013, 14, 153−159

Biomacromolecules

Article

same antibiotic concentrations, followed by incubation for 12 h at 25 °C with shaking. After harvesting cells by centrifugation, the scFv was purified by using immobilized metal affinity chromatography (IMAC), as described previously.26 Purified scFvs were analyzed by Coomassie blue staining of SDS-PAGE gels and immunodetection of the FLAG tag using a monoclonal anti-FLAG M1 antibody (Sigma-Aldrich).27 For production of the unglycosylated 4M5.3 scFv, the above procedure was carried out with cells containing only the pIG64M5.3-Gly vector in E. coli CLM37. Measurement of Fluorescein Binding by Competitive Enzyme Immunoassay (EIA). The wells of a multiwall polystyrene plate (Nunc) were coated overnight at 4 °C with 50 μL of 10 μg/mL fluorescein conjugated to bovine serum albumin (BSA). After three washes with phosphate-buffered saline (PBS) containing 0.05% Tween 20, wells were blocked with 2% BSA in PBS for 2 h at room temperature. After washing five times with PBS-Tween, wells were incubated for 1 h at room temperature with 50 μL of 1 μg/mL glycosylated or unglycosylated 4M5.3 scFv that had been preincubated for 1 h with free fluorescein at concentrations ranging from 0 to 100 ng/mL. After three washes with PBS-Tween, 50 μL of antipolyhistidine horse radish peroxidase-conjugated antibody, diluted 1:3000 in PBS containing 2% BSA, was added to wells, followed by five washes in PBS. The reaction was developed using 50 μL of 3,3′,5,5′tetramethylbenzidine (TMB) and stopped by the addition of 50 μL of 1 M H2SO4. Reactions were read at OD450. Data were fitted to a dose− response curve using a four parameter Hill equation in OriginPro 8 software. Oxidation of Glycosylated scFv. Purified scFv proteins [100 μg/ mL; measured using a Bradford protein assay kit (Sigma-Aldrich)] were incubated with 10 mM meta-sodium periodate in 0.1 M sodium acetate buffer, pH 5.5, for 30 min at room temperature. Oxidized glycosylated scFv or a 2:1 mixture of oxidized glycosylated/ unglycosylated scFvs, resulting from expression of the glycosylated scFv under nonoptimized conditions, was dialyzed against the desired final concentration of NaCl in 5 mM sodium phosphate at pH 7.0 and used immediately after preparation. Qualitative Analysis of Glycosylated scFv. Hydrazide groups react readily with aldehydes created upon oxidation of cis-glycol groups in polysaccharide moieties of N-linked glycoproteins to form stable hydrazone bonds. Therefore, this was used to confirm glycosylation of the 4M5.3 scFv, as described by Zhang and co-workers.28 Oxidized scFvs described above (0.4 mL of 100 μg/mL) were incubated with 40 mg of wet UltraLink hydrazide beads for 24 h at room temperature with gentle rocking. Noncovalently bound scFvs were removed by incubating the beads with an equal volume of urea solution (8 M urea in 0.4 M NH4HCO3, pH 8.3) and supernatants were analyzed in 12% SDS-PAGE gels stained with Coomassie brilliant blue. Immobilization of 4M5.3 scFvs on Aminated Beads. Ethylenediamine (EDA)-agarose with the desired degree of activation was prepared as previously described.29,30 Immobilization of oxidized or nonoxidized 4M5.3-Gly scFv was carried out by mixing 0.4 mL of scFv solutions (100 μg/mL in 5 mM phosphate buffer containing 5 to 300 mM NaCl, pH 7.0) with 40 mg of wet EDA-agarose (15 ± 3 μmol EDA groups per wet g). The suspension was incubated with gentle rocking at room temperature, followed by centrifugation at 12000 rpm for 1 min to pellet the beads and measurement of absorbance of the scFv-containing supernatant at 280 nm. The concentration of scFv in the supernatant was determined using a standard curve prepared with purified 4M5.3-Gly whose concentration had been determined by Bradford assay using BSA as standard. Measurement of Fluorescein Binding by Immobilized scFvs. Following covalent immobilization (in 5 mM phosphate buffer containing 150 mM NaCl, pH 7.0) or adsorption (5 mM NaCl, phosphate buffer, pH 7.0) of the 4M5.3-Gly scFv at concentrations ranging from 5 to 20 μg/mL (0.15−0.6 μM), aminated beads were washed five times with the relevant immobilization buffer, followed by blocking of uncovered surfaces with 3% BSA in phosphate buffer, pH 7.0 for 1 h at 37 °C. Beads were washed three times, and fluorescein (0.2−0.6 μM sodium salt in 5 mM sodium phosphate buffer, pH 7.0) was added. The suspension was incubated with gentle rocking for 12 h

reactions can lead to protein denaturation and the occurrence of disordered or inaccessible binding sites in as many as 90% of adsorbed molecules due to their uncontrolled orientation.12 The use of antibody-binding proteins, such as proteins G and A that specifically target the Fc region of an antibody,13,14 improves orientation but increases the dimensions of the immobilized layer and the potential for cross-reactivities, while it also targets the Fc stem lacking in most recombinant antibody fragments. Similarly, fragments such as scFvs that lack the glycosylated Fc domain cannot benefit from directed immobilization approaches developed for whole antibodies that are based on interactions with oxidized glycans.15 A variety of methods for oriented immobilization of scFvs that utilize fused histidine or cysteine tags have been developed.16 Incorporation of additional cysteine(s) can lead to lower yields of some scFvs in recombinant bacterial systems, however, or extensive protein aggregation mediated by intermolecular disulfides.17 The incorporation of genetically encoded peptide tags such as hexahistidine peptides8 holds great promise as it allows oriented immobilization of antibody fragments and facilitates purification via generic affinity chromatographies but applications of this approach are limited by the noncovalent nature of the antibody-support interaction. In this work, we describe a method for covalent, correctly oriented attachment of scFvs on aminated supports based on incorporation of a glycan tag into the recombinant antibody molecules. The protein N-glycosylation machinery from the εproteobacterium Campylobacter jejuni was previously transferred into E. coli and offers the opportunity to produce recombinant homogeneous glycoproteins in the bacterial periplasm.18 The present study, utilizing the well-characterized 4M5.3 antifluorescein scFv19,20 as a model antibody, demonstrates that the engineered glycan can be used for covalent, oriented immobilization of scFvs. The immobilization approach will have broad usefulness in sensing, diagnostic, and a variety of surface functionalization applications.



EXPERIMENTAL SECTION

Materials. All chemicals were obtained from Sigma-Aldrich unless otherwise specified. Sodium meta-periodate and UltraLink hydrazide beads were from Thermo Fisher Scientific. Cross-linked 4% agarose beads were supplied by Agarose Bead Technologies. Restriction enzymes were from TaKaRa Biotechnology (Dalian, China). E. coli CLM37 (rph-I IN(rrnD-rrnE)1, ΔwecA) and the pACYCpgl vector containing the pgl protein glycosylation locus from Campylobacter jejuni21 were provided by Professor Markus Aebi, ETH Zurich, Switzerland. The pIG6 periplasmic expression vector22 was provided by Professor Andreas Plückthun, University of Zurich, Switzerland. Cloning and Expression of Antifluorescein 4M5.3 scFvs. The hexahistidine-4M5.3-N-glycosylation sequence scFv fragment was synthesized by Nanjing GenScrip Tech Ltd. (Nanjing, China) and subcloned into pIG6 at EcoRV and HindIII sites. The resulting plasmid (pIG6−4M5.3-Gly) encoded the antifluorescein 4M5.3 scFv23 in an ompA leader-FLAG-His6-scFv-glycosylation site format (Supporting Information, S1). To produce glycosylated 4M5.3 scFv (4M5.3-Gly), the pIG6−4M5.3-Gly plasmid was cotransformed with the pACYCpgl plasmid into E. coli CLM37 cells. E. coli CLM37 is unable to synthesize native glycan chains due to the deletion of its wecA gene and therefore produces only the bacillosamine-containing glycan chains dictated by the plasmid-encoded C. jejuni glycosylation machinery.24 A 25 mL volume of LB medium containing 34 or 68 μg/mL chloramphenicol (Cam) and 100 or 200 μg/mL ampicillin (Amp) was inoculated with a single, freshly transformed colony from a LB-Cam-Amp plate and grown overnight in a shaking incubator at 37 °C. This culture was used to inoculate 1 L of ZYM-5052 autoinducing medium25 containing the 154

dx.doi.org/10.1021/bm301518p | Biomacromolecules 2013, 14, 153−159

Biomacromolecules

Article

Figure 1. Glycosylation of the 4M5.3 scFv. (A) Crystal structure of the 4M5.3 scFv (PDB entry 1X9Q) showing the locations of a bound fluorescein molecule in the binding pocket and the N- and C-termini of the scFv molecule. (B) Immunoblot and (C) Coomassie blue-stained SDS-PAGE gel analysis of expression of glycosylated (“Gly”) and unglycosylated (“Ungly”) 4M5.3 scFvs in E. coli CLM37. ScFvs were detected using M1 anti-FLAG antibody. Lanes 1 and 2, 4M5.3 scFv proteins purified from the periplasm of E. coli CLM37 cells containing the pIG6-4M5.3-Gly and the pACYCpgl plasmids and expressing the scFv in the presence of (1) standard and (2) double antibiotic concentrations; lane 3, unglycosylated 4M5.3 scFv purified from the periplasm of E. coli CLM37 cells containing the pIG6-4M5.3-Gly plasmid alone; lane 4, molecular weight protein markers. at 25 °C, following which bead mixtures were centrifuged at 12000 rpm for 1 min. Supernatants were excited at 490 nm and emission was measured at 514 nm. The concentration of fluorescein in the supernatant was determined using a standard curve and subtracted from the original fluorescein concentration to determine the amount of bound fluorescein. Values were corrected for binding to the beads and data were fitted in Prism 5.0 using log(inhibitor) versus response. Investigation of Leaching of Immobilized 4M5.3-Gly. The composites containing the adsorbed 4M5.3-Gly (in 5 mM sodium phosphate buffer, pH 7.0) or covalently immobilized 4M5.3-Gly (in 150 mM NaCl, phosphate buffer, pH 7.0) were incubated in a range of NaCl concentration (5 to 300 mM) in 5 mM sodium phosphate, pH 7.0. The amount of protein released into the supernatant at intervals up to 5 h was determined as described above. Analysis of Thermal Stability of Soluble or Immobilized scFvs. The thermal stability of immobilized and nonimmobilized scFvs was compared by monitoring their temperature-induced denaturation.31 ScFv beads (covalently and noncovalently bound) with bound fluorescein were incubated at 46 °C and released fluorescein was measured at intervals from 10 to 120 min, with emission at 514 nm and excitation at 490 nm. The released fluorescein was measured as described above. To compare the thermal stability of soluble, nonimmobilized 4M5.3-Gly scFv, an equimolar concentration of fluorescein and scFv (100 nM) was incubated in 5 mM phosphate buffer, pH 7.0, and analyzed from 10 to 120 min, as described above.

molecules could be detected. Immunoblotting using the M1 antibody revealed approximately 60−70% of 4M5.3 scFv molecules expressed in E. coli CLM37 cells exhibited an increase in molecular weight when expressed in the presence of the pACYCpgl protein glycosylation vector and standard antibiotic concentrations (100 μg/mL ampicillin, 34 μg/mL chloramphenicol; Figure 1B). After doubling of antibiotic concentrations, the lower molecular weight, unglycosylated scFv was undetectable by SDS-PAGE and immunoblot analysis in cells containing the pACYCpgl vector (Figure 1B,C). The results indicate that loss of the pACYCpgl plasmid from scFvexpressing cells was likely to be responsible for the initial, inefficient glycosylation of the target scFv. The 4M5.3 scFv protein, glycosylated scFv and mixture of the glycosylated and unglycosylated proteins were purified using IMAC from the periplasm of E. coli cells induced under the relevant conditions and the homogeneity of the glycosylated 4M5.3 scFv was confirmed by SDS-PAGE (Figure 1C). The purified yields of both the glycosylated and unglycosylated scFvs were 1−2 mg/L E. coli culture. Qualitative Analysis of Glycosylated scFv. The presence of the oxidized glycan chain was confirmed by immobilization of the purified scFvs on inert, low protein adsorption hydrazide beads which exhibit binding specificity for oxidized glycans.28 Following expression of the glycosylated scFv under nonoptimized conditions (see Figure 1), the resultant approximate 2:1 ratio of oxidized glycosylated and unglycosylated 4M5.3 scFvs was incubated with hydrazide beads. Washing the beads with 8 M urea to remove noncovalently bound protein molecules yielded only the smaller molecular weight, unglycosylated scFv, indicating the glycosylated protein was covalently attached to the beads (Figure 2). We (not shown) and others18,32 have previously confirmed the GalNAc-α1,4GalNAc-α1,4-[Glc-β1,3-]GalNAc-α1,4-GalNAc-α1,4-GalNAcα1,3-Bac-β1,N-Asn sequence of glycan chains attached by the C. jejuni glycosylation machinery to scFvs and other recombinant proteins expressed in E. coli. This single-branched heptasaccharide is characteristic of N-linked protein glycosylation in C. jejuni35 and is unchanged in recombinant proteins expressed in E. coli. Fluorescein-Binding Analysis of 4M5.3 scFvs. Binding analysis of the glycosylated and unglycosylated scFvs in solution revealed that the attached glycan chain had no significant effect on fluorescein binding by the 4M5.3 scFv, with IC50s of the unglycosylated and glycosylated scFvs of 2.89 ± 0.77 and 3.4 ± 0.5 pM, respectively (Figure 3). While maturation of antibodies in vivo is thought to reach a Kd



RESULTS AND DISCUSSION Production of Glycosylated 4M5.3 scFv. Glycosylation of scFvs in E. coli using the N-glycosylation machinery of C. jejuni has previously been reported.32 N-linked glycosylation of recombinant proteins in E. coli has been demonstrated to occur in flexible, accessible regions of already folded proteins,21 unlike in eucaryotes where protein translocation and folding are coupled to the glycosylation process. In this work, therefore, we introduced a glycan tag, incorporating the DQNAT oligosaccharide acceptor sequence of the C. jejuni PglB oligosaccharyl transferase,33 on a flexible linker (GGGGS) at the C-terminal end of the 4M5.3 scFv. The DQNAT glycosylation motif was based on previous work that established the optimal acceptor sequence for transfer of oligosaccharides by the C. jejuni PglB oligosaccharyl transferase to the asparagine side chain of acceptor proteins.33 The scFv C-terminus was chosen to maximize access of the C. jejuni glycosylation machinery and, based on the crystal structure of the scFv, its distal location from the antigen binding site (Figure 1A), which should ensure that N-glycosylation would not affect the antigen binding properties of the scFv. To investigate glycosylation of the 4M5.3 scFv, the M1 antiFLAG antibody, specific for N-terminal FLAG sequences,34 was utilized to ensure that only signal-processed periplasmic scFvs 155

dx.doi.org/10.1021/bm301518p | Biomacromolecules 2013, 14, 153−159

Biomacromolecules

Article

techniques have been utilized in vitro to create improved antibody variants for application in biotechnology and tumor targeting, where binding strength can limit efficacy. The 4M5.3 scFv was generated by directed evolution of the 4−4−20 scFv against fluorescein−biotin, resulting in an 1800-fold increase in affinity for the antigen and one of the highest affinity antibodies engineered to date.23 Protein Immobilization on Aminated Beads. After activation of agarose beads with EDA, scFvs were immobilized according to the schematic in Figure 4. EDA-agarose has been well-characterized in covalent-oriented immobilization of whole antibodies through their glycosidic chains, where oxidation of their sugar chains using sodium periodate is followed by immobilization of the antibodies on supports carrying primary amino groups.15 As an inert surface was required to reduce false positive signals, beads with a low concentration of amino groups and incubation in high salt concentrations15 were used to minimize nonspecific binding of the scFv. The 4M5.3 glycosylated scFv (100 μg/mL) was incubated with aminated beads, with or without oxidation of the attached glycan, in the presence of varying concentrations of sodium phosphate to inhibit noncovalent interactions between the scFvs and the agarose support. The results demonstrate that scFv immobilization occurs via two different mechanism: ionic adsorption and covalent attachment.15,37 While approximately 80% of both scFvs attached to the beads within 1 h in the presence of 5 mM sodium phosphate, at higher salt concentrations where ionic interactions were interrupted almost exclusively oxidized protein remained attached to the support (Figure 5): no immobilized, nonoxidized scFv was detectable after up to 5 h incubation in 150−200 mM sodium phosphate, whereas 80− 90% of the oxidized protein remained attached after 5 h incubation in up to 200 mM sodium phosphate, indicating a covalent association between the oxidized, glycosylated scFvs and the beads. The investigation also yielded experimental

Figure 2. Analysis of glycosylation of 4M5.3 scFv using hydrazide beads. After incubation of 40 μg of oxidized scFv with 40 mg of wet beads, proteins were separated by SDS-PAGE and stained using Coomassie blue. Lane 1, 2:1 mixture of oxidized glycosylated/ unglycosylated scFv prior to immobilization on hydrazide beads; lane 2, protein removed from beads upon washing with 8 M urea in 0.4 M NH4HCO3, pH 8.3; lane 3, molecular weight markers.

Figure 3. Competitive EIA to compare IC50s for fluorescein of purified glycosylated (squares) and unglycosylated (circles) 4M5.3 scFvs.

ceiling at approximately >0.1 nM due to physiological limitations in selecting for higher affinities,36 affinity maturation

Figure 4. Schematic of scFv immobilization on functionalized beads. The 4M5.3 scFv is shown as a ribbon structure, with α helices in red, β sheets in yellow and flexible loops in green. The glycan chain, attached to the C-terminus of the scFv, is shown as seven sugar monomers, including the scFvproximal bacillosamine in red, five N-acetylglucosamine units in yellow and the chemical structure of the β1,3-branched glucose that undergoes oxidation shown for clarity of the immobilization chemistry. The scFv binding pocket is arrowed. 156

dx.doi.org/10.1021/bm301518p | Biomacromolecules 2013, 14, 153−159

Biomacromolecules

Article

Figure 5. Investigation of attachment of (A) nonoxidized and (B). oxidized 4M5.3 glyco-scFv to aminated beads in the presence of 5 mM (●), 50 mM (▲), 100 mM (▼), 150 mM (◆) and 200 mM (*) sodium phosphate buffer, pH 7.0. Protein remaining in the reaction supernatant was measured after removal of the beads by centrifugation.

of the noncovalently adsorbed scFv, incubation in the presence of 30−60 mM NaCl over 5 h led to desorption of 30−60% of the immobilized scFv molecules, while exposure to 120 mM or higher concentrations of NaCl led to removal of greater than 90% of the scFv over the 5 h analysis period. This leaching of adsorbed scFv from the support has the potential to greatly impact the performance and sensitivity of EIA assays or the efficacy of protein purification in the case of immunomatrices. The increased stability of the covalently attached scFvs, however, could enable applications such as the development of reuseable immunosensors in areas such as water monitoring, where ongoing, low cost monitoring in situ could be complemented by elution of the target analyte and sensor reuse by simple rinsing in water or a low pH buffer. Stable, durable attachment of scFvs also provides significant advantages in protein purification, where covalently immobilized scFvs greatly reduce the required amounts of the one of the most expensive and potentially process-limiting components of the production pipeline, the scFv, and significantly decrease process downtime necessitated by column regeneration. The antigen-binding ability of the immobilized scFvs was investigated by comparing the amount of fluorescein bound by differently tethered scFvs in fluorescence immunoassay. After incubation of 5−20 μg/mL of the glycosylated/oxidized and unglycosylated scFvs with 100 mg beads under optimized conditions, up to 5.45 nmol/g beads of unglycosylated scFv and 5.56 nmoles of glycosylated protein was found to be retained on the beads. Upon incubation with ligand, the covalently attached glycosylated scFv exhibited approximately 4-fold higher fluorescein-binding activity than their noncovalently adsorbed counterparts at all protein concentrations investigated (Figure 7A). Normalization of the amount of fluorescein bound per immobilized scFv molecule revealed binding efficiencies of one fluorescein molecule for every 12.1 to 18.8 noncovalently adsorbed scFvs, compared with one ligand per 3.0 to 5.0 covalently attached scFvs. As the immobilized scFv reached an equimolar ratio with fluorescein at just 1.5 nmol/g, the binding of only 1 fluorescein molecule per 3 to 5 covalently attached scFvs was due to limitations of diffusion at lower scFv concentrations and the excess of scFv at higher antibody concentrations. While the percentage of available fluorescein bound by the covalently immobilized scFvs ranged from 31% at 1.4 nmol/g scFv to 73% at 5.6 nmol/g, however, the adsorbed scFvs captured only 7% (1.3 nmol scFv/g) to 19% (5.5 nmol/ g) of the fluorescein in solution. As the affinities of the glycosylated and unglycosylated scFvs are similar, this approximately 4-fold difference in the “specific activity” or functional activity of the immobilized scFvs is most likely due

conditions (150 mM sodium phosphate buffer, pH 7.0, 5 h) that minimized nonspecific adsorption events for subsequent analyses of covalent immobilization of the glycosylated scFv via its oxidized glycan. This ability to covalently immobilize analyte-specific scFvs on solid supports, while greatly reducing nonspecific protein adsorption will find particular application in the investigation of low concentration analytes using bioanalytical techniques such as surface plasmon resonance and quartz crystal microbalance. One of the primary parameters to be considered in immunosensor or immunomatrix development is the stability of the antibody protein on the biological matrix. Many immunoassays are carried out over periods of 3−6 h, incorporating at least hour-long incubations of supportimmobilized antibodies with blocking agents, ligand and reporter antibody or antibodies, extensive washing to remove nonspecifically bound materials and development of the signal. After adsorption or oxidation and covalent immobilization of 810 ± 33 μg (25.3 ± 1.03 nmoles) or 868 ± 33 μg (27.2 ± 0.4 nmoles), respectively, of unglycosylated or glycosylated 4M5.3 scFv under conditions based on Figure 5, the scFv removed from the beads was measured over 5 h and calculated as a percentage of the initial immobilized scFv concentrations. In the case of the covalently attached scFv in this work, loss of protein from the support was minimal over the 5 h period, even at NaCl concentrations as high as 300 mM (Figure 6), indicating suitability of the immobilization approach for immunoassays and multiplex screening platforms. In the case

Figure 6. Removal of adsorbed (●) or covalently immobilized (▲) 4M5.3 scFv from aminated beads. Unglycosylated and glycosylated 4M5.3 scFv (100 μg/mL) was adsorbed or oxidized and covalently immobilized to beads for 5 h in the presence of 5 mM sodium phosphate buffer or 150 mM sodium phosphate buffer, respectively, based on Figure 4. The amounts of scFv removed under varying conditions were measured over 5 h and are presented as a percentage of the initial immobilized scFv. 157

dx.doi.org/10.1021/bm301518p | Biomacromolecules 2013, 14, 153−159

Biomacromolecules

Article

Figure 7. Fluorescein binding of 4M5.3 scFvs. (A) Comparison of fluorescein binding ability of scFvs covalently immobilized (■) and adsorbed (▲) onto the aminated beads. The initial fluorescein concentration was 0.6 μM. (B) Thermal stability of fluorescein binding of soluble (●), adsorbed (▲), and covalently immobilized (▼) 4M5.3 scFvs. Fluorescein binding was measured before and during incubation at 46 °C. All values are corrected for binding of fluorescein to beads coated with BSA. Error bars represent standard deviations obtained from triplicate samples.

over the use of whole antibodies, the additional increase in analyte detection sensitivity and reduction in nonspecific protein adsorption demonstrated using this immobilization approach will facilitate detection of low concentration analytes in diagnostic and immunosensing applications. The technology also has potential application in the development of improved, reusable protein purification matrices, multifunctional and biocompatible surfaces, and cell-targeted delivery of polymeric drug carriers.

to increased accessibility of the binding pocket of the latter molecules resulting from their improved, glycan-mediated orientation on the beads and reduced multilayering of the scFvs, and lower distortion of the conformation of the oriented compared with the ionically adsorbed antibody molecules. This improvement in binding is considerably greater than the 62% increase in signal level previously reported from biotin-based orientation of scFvs on EIA plates or orientation via polystyrene-binding polypeptides,38,39 though correct orientation of scFvs on carbon nanotubes, using a hexahistidine tag, has also been found to be essential for detection of any scFv− Ag interaction.40 A comparative analysis of stability of fluorescein antigen binding was investigated using an accelerated degradation test carried out at 46 °C.31 This revealed similar reductions in fluorescein binding of approximately 40% after 10 min and more than 80% after a 2 h incubation in nonimmobilized and adsorbed, randomly oriented scFvs (Figure 7B). Meanwhile, the covalently attached, glycosylated scFv exhibited higher binding stability, with in excess of 50% of its fluorescein binding ability retained over the same 2-h analysis period. This increased stability of antigen binding by the covalently attached scFv, in combination with its increased functional activity, indicates broad potential application of the scFv immobilization approach in immunoassay development or applications such as protein purification.



ASSOCIATED CONTENT

S Supporting Information *

Schematic of expression vector encoding the 4M5.3 scFv and relevant sequence motifs. This material is available free of charge via the Internet at http://pubs.acs.org.



AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]; [email protected]. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS We thank Markus Aebi and Andreas Plückthun for the kind gifts of E. coli CLM37 and pACYCpgl plasmid, and pIG6 vector, respectively. Funding was received from National Natural Science Foundation of China (NSFC Grant 31070822 to X.H.), Enterprise Ireland Science and Technology Agency (Grant PC/2007/021 to S.R.), and Irish Research Council for Science, Engineering and Technology (IRCSET Grant PD/2010/1689 to M.J.H.).



CONCLUSION In summary, a procedure for covalent, correctly oriented attachment of scFv antibody fragments to solid supports was developed. A glycan chain was engineered onto a model scFv, distal from the antigen binding pocket, for immobilization of scFvs in a binding-compatible orientation. The scFv glycosylation efficiency was improved by increasing antibiotic concentrations in the culture medium and the glycosylated scFv was purified in a single step using metal affinity chromatography. The glycosylated scFv exhibited unchanged antigen binding compared with the unmodified scFv and could be covalently attached after oxidation of the glycan to aminefunctionalized beads under conditions that minimized noncovalent protein adsorption. Nonspecifically adsorbed protein molecules, but not the covalently attached scFvs, could be removed by washing in high salt concentrations while the covalently immobilized scFvs bound 4-fold more antigen than adsorbed protein due to their oriented immobilization via the binding site-distal glycan. While the use of scFv fragments offers increased yields, reduced costs, and ease of protein modification



REFERENCES

(1) Padlan, E. A. Mol. Immunol. 1994, 31, 169−217. (2) McCafferty, J.; Griffiths, A. D.; Winter, G.; Chiswell, D. J. Nature 1990, 348, 552−554. (3) Shen, Z.; Stryker, G. A.; Mernaugh, R. L.; Yu, L.; Yan, H.; Zeng, X. Anal. Chem. 2005, 77, 797−805. (4) Hil’chuk, P. V.; Okuniev, O. V.; Pavlova, M. V.; Irodov, D. M.; Horbatiuk, O. B. Ukr. Biokhim. Zh. 2006, 78, 52−61. (5) Hu, X.; Spada, S.; White, S.; Hudson, S.; Magner, E.; Wall, J. G. J. Phys. Chem. B 2006, 110, 18703−18709. (6) Ferrer-Miralles, N.; Domingo-Espin, J.; Corchero, J. L.; Vazquez, E.; Villaverde, A. Microb. Cell Fact. 2009, 8, 17. (7) Kolaj, O.; Spada, S.; Robin, S.; Wall, J. G. Microb. Cell Fact. 2009, 8, 9. 158

dx.doi.org/10.1021/bm301518p | Biomacromolecules 2013, 14, 153−159

Biomacromolecules

Article

(8) Shen, Z.; Yan, H.; Zhang, Y.; Mernaugh, R. L.; Zeng, X. Anal. Chem. 2008, 80, 1910−1917. (9) Torrance, L.; Ziegler, A.; Pittman, H.; Paterson, M.; Toth, R.; Eggleston, I. J. Virol. Methods 2006, 134, 164−170. (10) Hu, X.; O’Connor, I. B.; Wall, J. G. In Biological Interactions with Surface Charge in Biomaterials; Syed, T. A., Ed. Royal Society of Chemistry: London, 2012; pp 90−104. (11) Jung, Y.; Jeong, J. Y.; Chung, B. H. Analyst 2008, 133, 697−701. (12) Cho, I.-H.; Paek, E.-H.; Lee, H.; Kang, J. Y.; Kim, T. S.; Paek, S.H. Anal. Biochem. 2007, 365, 14−23. (13) Danczyk, R.; Krieder, B.; North, A.; Webster, T.; HogenEsch, H.; Rundell, A. Biotechnol. Bioeng. 2003, 84, 215−223. (14) Jung, Y.; Lee, J. M.; Kim, J. W.; Yoon, J.; Cho, H.; Chung, B. H. Anal. Chem. 2009, 81, 936−942. (15) Batalla, P.; Fuentes, M.; Grazu, V.; Mateo, C.; FernandezLafuente, R.; Guisan, J. M. Biomacromolecules 2008, 9, 719−723. (16) Hernandez, K.; Fernandez-Lafuente, R. Enzyme Microb. Technol. 2011, 48, 107−122. (17) Shen, Z.; Mernaugh, R. L.; Yan, H.; Yu, L.; Zhang, Y.; Zeng, X. Anal. Chem. 2005, 77, 6834−6842. (18) Wacker, M.; Linton, D.; Hitchen, P. G.; Nita-Lazar, M.; Haslam, S. M.; North, S. J.; Panico, M.; Morris, H. R.; Dell, A.; Wren, B. W.; Aebi, M. Science 2002, 298, 1790−1793. (19) Midelfort, K. S.; Wittrup, K. D. Protein Sci. 2006, 15, 324−334. (20) Midelfort, K. S.; Hernandez, H. H.; Lippow, S. M.; Tidor, B.; Drennan, C. L.; Wittrup, K. D. J. Mol. Biol. 2004, 343, 685−701. (21) Kowarik, M.; Numao, S.; Feldman, M. F.; Schulz, B. L.; Callewaert, N.; Kiermaier, E.; Catrein, I.; Aebi, M. Science 2006, 314, 1148−1150. (22) Ge, L.,; Knappik, A.; Pack, P.; Freund, C.; Plückthun, A. In Antibody Engineering: Expressing Antibodies in Escherichia coli, 2nd ed.; Borrebaeck, C. A. K., Ed.; Oxford University Press, Inc.: New York, 1995; pp 229−266. (23) Boder, E. T.; Midelfort, K. S.; Wittrup, K. D. Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 10701−10705. (24) Linton, D.; Dorrell, N.; Hitchen, P. G.; Amber, S.; Karlyshev, A. V.; Morris, H. R.; Dell, A.; Valvano, M. A.; Aebi, M.; Wren, B. W. Mol. Microbiol. 2005, 55, 1695−1703. (25) Studier, F. W. Protein Exp. Purif. 2005, 41, 207−234. (26) Hu, X.; O’Hara, L.; White, S.; Magner, E.; Kane, M.; Wall, J. G. Protein Exp. Purif. 2007, 52, 194−201. (27) Hu, X.; O’Dwyer, R.; Wall, J. G. J. Biotechnol. 2005, 120, 38−45. (28) Zhang, H.; Li, X. J.; Martin, D. B.; Aebersold, R. Nat. Biotechnol. 2003, 21, 660−666. (29) Fernandez-Lafuente, R.; Rosell, C. M.; Rodriguez, V.; Santana, C.; Soler, G.; Bastida, A.; Guisan, J. M. Enzyme Microb. Technol. 1993, 15, 546−550. (30) Fuentes, M.; Pessela, B. C.; Mateo, C.; Palomo, J. M.; Batalla, P.; Fernandez-Lafuente, R.; Guisan, J. M. Biomacromolecules 2006, 7, 1357−1361. (31) Orr, B. A.; Carr, L. M.; Wittrup, K. D.; Roy, E. J.; Kranz, D. M. Biotechnol. Prog. 2003, 19, 631−638. (32) Lizak, C.; Fan, Y. Y.; Weber, T. C.; Aebi, M. Bioconjugate Chem. 2011, 22, 488−496. (33) Chen, M. M.; Glover, K. J.; Imperiali, B. Biochemistry 2007, 46, 5579−5585. (34) Knappik, A.; Plückthun, A. BioTechniques 1994, 17, 754−761. (35) Young, N. M.; Brisson, J. R.; Kelly, J.; Watson, D. C.; Tessier, L.; Lanthier, P. H.; Jarrell, H. C.; Cadotte, N.; St Michael, F.; Aberg, E.; Szymanski, C. M. J. Biol. Chem. 2002, 277, 42530−42539. (36) Foote J, E. H. Proc. Natl. Acad. Sci. U.S.A. 1995, 92, 1254−1256. (37) Batalla, P.; Fuentes, M.; Mateo, C.; Grazu, V.; FernandezLafuente, R.; Guisan, J. M. Biomacromolecules 2008, 9, 2230−2236. (38) Garcia-Suarez, M. M.; Cron, L. E.; Suarez-Alvarez, B.; Villaverde, R.; Gonzalez-Rodriguez, I.; Vazquez, F.; Hermans, P. W.; Mendez, F. J. Clin. Microbiol. Infect. 2009, 15, 443−453. (39) Kumada, Y.; Hamasaki, K.; Shiritani, Y.; Nakagawa, A.; Kuroki, D.; Ohse, T.; Choi, D. H.; Katakura, Y.; Kishimoto, M. Anal. Bioanal. Chem. 2009, 395, 759−765.

(40) Lo, Y. S.; Nam, D. H.; So, H. M.; Chang, H.; Kim, J. J.; Kim, Y. H.; Lee, J. O. ACS Nano 2009, 3, 3649−3455.

159

dx.doi.org/10.1021/bm301518p | Biomacromolecules 2013, 14, 153−159