Self-Directed and Self-Oriented Immobilization of Antibody by Protein

A versatile biolinker for efficient antibody immobilization was prepared by site-specific coupling of protein G to DNA oligonucleotide. This protein G...
0 downloads 0 Views 450KB Size
Download PDF
Anal. Chem. 2007, 79, 6534-6541

Self-Directed and Self-Oriented Immobilization of Antibody by Protein G-DNA Conjugate Yongwon Jung,† Jeong Min Lee,† Hyungil Jung,‡ and Bong Hyun Chung*,†

BioNanotechnology Research Center, Korea Research Institute of Bioscience and Biotechnology, P.O. Box 115, Yuseong, Daejeon 305-600, Korea, and Department of Biotechnology, Yonsei University, Seoul, 129-749, Korea

A versatile biolinker for efficient antibody immobilization was prepared by site-specific coupling of protein G to DNA oligonucleotide. This protein G-DNA conjugate ensures the controlled immobilization of an antibody to the intended area on the surface of bioassay chips or particles, while maintaining the activity and orientation of the bound antibody. Streptococcus protein G tagged with a cysteine residue at the N-terminus was chemically linked to aminemodified, single-stranded DNA. SPR analysis indicated that the protein G-DNA conjugates sequence-specifically bind to complementary surface-bound DNA probes. More importantly, the resulting protein G, which is hybridized onto the DNA surface, possesses a greater antibody/ antigen binding ability than even properly oriented protein G linked on the chip surface by chemical bonding. Antibody targeting on glass slides could also be achieved by using this linker system without modifying or spotting antibodies. Moreover, the protein G-DNA conjugate provided a simple but effective method to label DNAfunctionalized gold nanoparticles with target antibodies. The DNA-linked protein G construct introduced in this study offers a useful strategy to manage antibody immobilization in many immunoassay systems. A highly specific and strong binding ability of an antibody to its target antigen has been extensively employed in a wide range of biosensors as well as in proteomic analyses.1,2 A common key step for the development of most immunoassays is the efficient immobilization of an antibody on the solid surface of assay devices. Solid supports include those of gold and glass chips and, more recently, various nanostructures such as nanoparticles and nanotubes.3 Primary problems in the antibody immobilization process to these surfaces are instability of bound antibodies and a loss of the binding activity due to the random orientation and chemical modification of the antibodies. Considerable efforts have been made to circumvent these difficulties in the field of antibody-based bioassays. * To whom correspondence should be addressed. E-mail: [email protected], Fax: +82-42-879-8594. † Korea Research Institute of Bioscience and Biotechnology. ‡ Yonsei University. (1) Murphy, L. Curr. Opin. Chem. Biol. 2006, 10, 177-184. (2) Hultschig, C.; Kreutzberger, J.; Seitz, H.; Konthur, Z.; Bussow, K.; Lehrach, H. Curr. Opin. Chem. Biol. 2006, 10, 4-10. (3) Jianrong, C.; Yuqing, M.; Nongyue, H.; Xiaohua, W.; Sijiao, L. Biotechnol. Adv. 2004, 22, 505-518.

6534 Analytical Chemistry, Vol. 79, No. 17, September 1, 2007

DNA-directed immobilization (DDI) has been used to spatially assemble target proteins on DNA-functionalized assay surfaces.4 DNA surfaces are robust and stable, and also easy to be fabricated compared to protein surfaces. DNA-directed immobilization of proteins allows us to avoid long-term storage and harsh incubation and spotting processes of unstable proteins on various assay surfaces. Previously, several strategies for antibody immobilization on DNA surfaces were reported. The first example employed streptavidin-DNA conjugates, which were used to immobilize biotinylated antibodies.5 Antibodies could also be directly linked to DNA and addressed to DNA-modified gold surfaces.6,7 Both methods, however, require antibody modifications with either small molecules or DNA itself, which leads to the random antibody orientation and possibly chemical modifications at the antigen binding sites. Protein G has been widely used to immobilize different types of antibodies in numerous immunoassays. The protein specifically binds the Fc region of an antibody and therefore provides proper orientation of the bound antibody, resulting in the antigen binding sites optimally exposed to the assay solution.8 In addition, protein G-mediated antibody immobilization usually does not require any antibody modifications. Thus, this method offers significantly improved antigen detection by bound antibodies compared to conventional antibody immobilization methods such as covalent coupling or physical adsorption. In recent years, investigators have focused on improving the orientation of protein G itself on assay surfaces to enhance the antibody binding ability of the protein even further.9-11 Here we introduce a novel linker system that possesses strengths of both DNA-directed immobilization and protein G for antibody immobilization. Antibodies can be easily addressed onto the designed spots of assay surfaces without the need of any antibody modifications. Amine-modified, single-stranded DNA was linked to the N-termius of recombinant protein G, which contains two Fc binding domains. Antibody assembly assisted by the (4) Niemeyer, C. M. Trends Biotechnol. 2002, 20, 395-401. (5) Wacker, R.; Niemeyer, C. M. Chembiochem 2004, 5, 453-459. (6) Boozer, C.; Ladd, J.; Chen, S.; Yu, Q.; Homola, J.; Jiang, S. Anal. Chem. 2004, 76, 6967-6972. (7) Boozer, C.; Ladd, J.; Chen, S.; Jiang, S. Anal. Chem. 2006, 78, 1515-1519. (8) Guss, B.; Eliasson, M.; Olsson, A.; Uhlen, M.; Frej, A. K.; Jornvall, H.; Flock, J. I.; Lindberg, M. EMBO J. 1986, 5, 1567-1575. (9) Ha, T. H.; Jung, S. O.; Lee, J. M.; Lee, K. Y.; Lee, Y.; Park, J. S.; Chung, B. H. Anal. Chem. 2007, 79, 546-556. (10) Gao, D.; McBean, N.; Schultz, J. S.; Yan, Y.; Mulchandani, A.; Chen, W. J. Am. Chem. Soc. 2006, 128, 676-677. (11) Fowler, J. M.; Stuart, M. C.; Wong, D. K. Anal. Chem. 2007, 79, 350-354. 10.1021/ac070484i CCC: $37.00

© 2007 American Chemical Society Published on Web 08/02/2007

Figure 1. (A) Cysteine-tagged protein G construct and amine-modified oligonucleotide. Antibody binding domain (BD), a 6× His tag (HIS6), and a cysteine residue (Cys) of the protein G construct are depicted. (B) Schematic representations of DNA-directed antibody immobilization by the protein G-DNA conjugate.

protein G-DNA conjugate and subsequent antigen interaction were closely investigated on the DNA surfaces of gold and glass chips by SPR and fluorescence scanning analyses, respectively. We were also able to manipulate the level of antibody coverage on DNA-functionalized gold nanoparticles (AuNPs) by controlling the nature of the surface DNA. Well-developed DNA immobilization techniques can be effectively utilized for antibody immobilization by using our system. EXPERIMENTAL SECTION Materials. Bovine serum albumin (BSA), immunoglobulin Gs (IgGs) from various sources, sodium citrate tribasic dehydrate, and hydrogen tetrachloroaurate(III) trihydrate were all purchased from Sigma. Anti-human kallikrein 3/PSA antibody and recombinant human kallikrein 3/PSA were purchased from R&D Systems, Inc. Sulfosuccinimidyl 4-(N-maleimidomethyl)cyclohexane-1-carboxylate (sulfo-SMCC) and disuccinimidyl suberate (DSS) were purchased from Pierce. Oligo(ethylene glycol) thiol HS(CH2)11(OCH2CH2)6OH and carboxylated oligo(ethylene glycol) thiol HS(CH2)11(OCH2CH2)6OCH2CO2H were supplied by Cos Biotech (Daejon, Korea). All modified guide and complementary oligonucleotides are obtained from Bioneer (Daejon, Korea). Guide DNA A (gA: 5′-TTCTGTGTGAAATTGTTATCCGCT-3′) and B (gB: 5′-TGAATCATGGTCATAGCTGTTGGC-3′) are modified by an amine group at the 5′ end. Complementary DNA A (cA: 5′-AGCGGATAACAATTTCACACAGG-3′) and B (cB: 5′GCCAACAGCTATGACCATGATTC-3′) are modified by a thiol or amine group also at the 5′ end for the immobilization on the gold chip or glass slide surface, respectively. Bare Au sensor chips for SPR measurement and 2-(2-pyridinyldithio)ethaneamine (PDEA) were supplied by Biacore AB (Uppsala, Sweden). Deionized water (Satorius) was used for the preparation of all solutions. Preparation of Modified Protein G Construct. The gene of Streptococcal protein G with two Fc binding domains (Figure 1) was amplified by using a polymerase chain reaction (PCR) from

the genomic DNA of Lancefield’s group G Streptococcus (KCCM, Korean Culture Center of Microorganisms). The forward primer (5′-GGGAATTCCATATGGATTGCGGCGGCGGCGGCAGCAAAGGCGAAACAACTACTGAAGCT) was designed to contain an NdeI restriction enzyme site and additional seven amino acids (AspCysGlyGlyGlyGlySer) at the N-terminus of protein G. The reverse primer (5′-GAGCTCGAGTTCAGTTACCGTAAAGGTCTTAGTC-3′) contains an XhoI restriction enzyme site and the sequence for a 6× His fusion peptide for protein purification. The PCR product was cloned into the expression vector pET21a (Novagen), and the protein was expressed in BL21(DE3) cells. Cell cultures were shaken at 37 °C until the optical density, OD600, was 0.6. The protein expression was induced by 1 mM isopropyl β-D-thiogalactopyranoside and continued for an additional 12 h at the same temperature. The protein G construct was first purified with an Immobilized Metal Affinity Chromatography column (Bioprogen Co.) according to the manufacturer’s procedure. The resulting protein was further purified by anionexchange chromatography on a Q Exellose column (Bioprogen Co.) with a continuous sodium chloride gradient from 0 to 1 M. Purified protein was dialyzed against a phosphate-buffered saline (PBS, pH 7.4) buffer containing 1 mM EDTA and 10 mM DTT and stored at -80 °C before being used. Synthesis and Purification of Protein G-DNA Conjugates. Sixty nanomoles of amine-modified guide DNA oligonucleotides in 400 µL of 0.25 M phosphate buffer (pH 7.1) was reacted with 1.5 mg of sulfo-SMCC (3400 nmol) dissolved in 75 µL of DMF. Following incubation at room temperature for 1 h, the activated oligonucleotides were separated from the excess sulfo-SMCC by using a Sephadex G25 gel filtration column (PD-10; Amersham) with conjugation buffer (20 mM Tris, 50 mM NaCl, and 1 mM EDTA pH 7.0). During the DNA activation, cysteine-tagged protein G (55 µM) was further reduced by 20 mM DTT for 30 min at room temperature to completely free sulfhydryl groups. QuantiAnalytical Chemistry, Vol. 79, No. 17, September 1, 2007

6535

tation of free sulfhydryl groups in protein G solution was performed by using Ellman’s reagent (Pierce) according to the manufacturer’s protocol, yielding 0.96 ( 0.03 of SH groups per protein. The reduced protein G was transferred to the conjugation buffer by using a desalting column (PD-10). The same molar ratio of protein G was mixed with maleimido-activated oligonucleotides, followed by 2-h incubation at room temperature. The unreacted DNA was first removed from the reaction mixture by a His tag affinity column, which captures only unreacted protein G and protein G-DNA conjugates. The protein G-DNA conjugate was further purified by anion-exchange chromatography as described above. The collected conjugates were concentrated, and the buffer was changed to PBS by repeated ultrafiltration (3000 cutoff; Millipore). The conjugate concentration was determined by absorbance measurement at 260 nm (the absorbance coefficient of the conjugate, 260 000 M-1 cm-1). The protein-DNA conjugate was characterized on 12% SDS and 10% native polyacrylamide gels. Oligonucleotides in gels were stained with GelRed (Biotium, Inc.). Surface Plasmon Resonance (SPR) Analysis on Gold Chip. The SPR measurements were performed with DNAmodified gold chips on a Biacore 3000 device (Biacore AB) at 25 °C using a PBS buffer as a running solution. The gold chip surface was cleaned with a concentrated “piranha” solution (70% (v/v) H2SO4, 30% (v/v) H2O2) and was thoroughly rinsed with ethanol and deionized water prior to the DNA modification. A mixed self-assembled monolayer (SAM) of thiolated complementary oligonucleotides and oligo(ethylene glycol) thiol was formed on a gold surface as previously described.6 Briefly, a precleaned gold chip was immersed in a 1.0 M KH2PO4 solution of 100 nM DNA and 10 µM oligo(ethylene glycol) thiol for 24 h. A solution with 100 nM DNA and 5 µM oligo(ethylene glycol) thiol was used to obtain the DNA-modified gold surface with a higher DNA density, where this DNA gold surface was employed for the antibody/antigen binding study. For PDEA-protein G immobilization, a mixed SAM of oligo(ethylene glycol) thiol and carboxylated oligo(ethylene glycol) thiol (3:1) was formed on a gold surface by emersing a precleaned gold chip into 1 mM mixed thiol solution in ethanol for 12 h. The modified gold chip was washed with deionized water and dried by nitrogen. Assay chips were stored at 4 °C before the analysis. For SPR measurements, sample solutions were introduced into the instrument channel at a rate of 10 µL/min. DNA and protein G-DNA conjugates were immobilized on the DNA-modified gold surface at a concentration of 100 nM, unless indicated otherwise. Source-dependent antibody binding was carried out at a concentration of 50 µg/mL. Cysteine-mediated chemical conjugation of protein G on the carboxylated gold surface was performed on a Biacore 3000 instrument at a flow rate of 8 µL/min. Following the activation of the sensor chip with PDEA according to the manufacturer’s protocol, cysteine-tagged protein G (50 µg/mL) was introduced until the desired degree of protein immobilization was obtained. Subsequent anti-PSA and PSA bindings were carried out at a protein concentration of 100 nM for 10 min. Upon each protein immobilization, the SPR binding response is measured in resonance units (RUs), where 1 RU equals 1 pg of protein/mm2 of a protein density on the surface. 6536

Analytical Chemistry, Vol. 79, No. 17, September 1, 2007

Fluorescence Scanning Analysis on Glass Surface. For the construction of DNA array, 10 µM amine-modified complementary oligonucleotides were spotted onto an epoxide glass slide (Corning) by using a pin-type protein arrayer (CM1000, Proteogen). The slides were incubated for 16 h at room temperature and stored at 4 °C until use. Immediately before the hybridization experiment, the slides were treated with blocking solution containing 5× SSC (0.75 M NaCl and 75 mM sodium citrate pH 7.0), 0.1% SDS, and 0.1 mg/mL BSA for 1 h at 42 °C. The blocked slides were washed with 0.1× SSC and deionized water and dried by blowing nitrogen. In order to immobilize antibodies onto the constructed DNA glass surface, 100 nM protein G-DNA conjugate was mixed with 150 nM antibody in the PBS buffer supplemented with 0.01% Tween 20 and 0.1 mg/mL BSA, and the mixture was treated to the DNA surface for 1 h. The slides were washed three times with PBS containing 0.01% Tween 20 (PBST). Following antigen interaction was performed by adding the indicated concentration of Cy3-labeled antigen onto the antibody surface and incubating for 1 h. The slides were sequentially rinsed with PBST, PBS, and last 0.1× PBS for 5 min and briefly dried with nitrogen before fluorescence measurement. The fluorescence images were obtained by using a GenPix 4200 (Axon). Antibodies were chemically cross-linked to the protein G-DNA conjugate by using bifunctional amine-specific reagent DSS. The 0.1 nmol of protein-DNA conjugate was mixed with 0.2 nmol of the antibody (anti-PSA) and the resultant mixture incubated for 1 h at 4 °C. To this mixture were added different amounts of DSS cross-linker (4, 20, and 100 nmol). Cross-linked antibody/protein G-DNA was purified by a His affinity column as described above. Antibody Labeling of DNA-Functionalized Gold Nanoparticles. Citrate-stabilized gold nanoparticles (13 nm ( 1 nm) were prepared by citrate reduction of chloroauric acid.12 The particles were coated with thiol-modified complementary oligonulceotides (cA) as previously reported.13 Complementary DNA coverage on gold nanoparticles was reduced by using a mixture of complementary oligonucleotides with short thiol-modified DNA (5′-SHTTTTTTTT-3′) (1:3). Fluorescein-labeled guide DNA A (Fl-gA) was used to determine the extent of hybridization on cAfunctionalized gold nanoparticles.14 Hybridization of the protein G-guide oligonucleotide A conjugate (gA-G) and subsequent antibody binding to cA-modified 13-nm gold nanoparticles was performed in a PBS buffer supplemented with 0.01% Tween 20 and 0.1 mg/mL BSA at room temperature. Following the incubation of gA-G (500 nM) with 10 nM nanoparticles for 30 min, excess gA-G was removed twice by centrifugation for 30 min at 14 000 rpm. The resulting protein G-covered particles were reacted with 200 nM human IgG. Gold nanoparticles with various modifications were analyzed on 1-2% agarose gels with 0.5× TAE (10 mM acetic acid, 0.5 mM EDTA, and 20 mM Tris pH 8.0) as a running buffer. RESULTS AND DISCUSSION Preparation of Protein G-DNA Conjugates. A single cysteine residue was introduced at the N-terminus of protein G (12) Frens, G. Nat. Phys. Sci. 1973, 241, 20-22. (13) Rosi, N. L.; Giljohann, D. A.; Thaxton, C. S.; Lytton-Jean, A. K.; Han, M. S.; Mirkin, C. A. Science 2006, 312, 1027-1030. (14) Demers, L. M.; Mirkin, C. A.; Mucic, R. C.; Reynolds, R. A.; Letsinger, R. L.; Elghanian, R.; Viswanadham, G. Anal. Chem. 2000, 72, 5535-5541.

Figure 2. PAGE analysis of protein G-DNA conjugate. (A) 12% SDS-PAGE analysis of cysteine-tagged protein G construct (G) and the protein G-DNA conjugate (gA-G) with protein size markers (M). Schematic representation of gA-G is shown in the right box. (B) 10% native PAGE analysis of gA-G. Lane 1, protein G (75 pmol); lane 2, gA-G (15 pmol); lane 3, gA-G (15 pmol) + cA (150 pmol); lane 4, gA-G (15 pmol) + cB (150 pmol); lane 5, cA (150 pmol); M, 100-base pair DNA size marker. The gel was first stained with GelRed for DNA staining, followed by protein staining by Comassie.

containing two repeated antibody binding domains. Without any cysteine residues present on native protein G, amine-modified DNA can be specifically conjugated to the N-terminal of this protein G construct by using a heterobifunctional amine/thiolreactive cross-linker such as sulfo-SMCC. Similar strategies have been successfully used to conjugate DNA to various thiolcontaining biomolecules including peptides and proteins.15-17 Here 24-mer amine-modified oligonucleotides were coupled to the protein G construct. Two consecutive chromatographic purifications were carried out to isolate the protein G-DNA conjugates from unreacted proteins and DNA. Denaturing as well as native gel analyses confirmed the highly pure protein-DNA conjugate, which is essentially free from uncoupled protein G and oligonucleotides (Figure 2). Under the native condition, the protein G-DNA conjugate migrates along with 200-base pair DNA as shown in Figure 2B. The conjugate can be stained by both DNA (GelRed) and protein (Commassie) specific dyes. Complementary oligonucleotides were sequence-specifically hybridized to the single-stranded DNA of the protein G-DNA conjugate, yielding the retarded migration of the conjugate on a native gel (lane 2 versus lane 3). The presence of the hybridized complementary oligonucleotide on the protein G conjugate was additionally confirmed by the increased band intensity shown only by DNA staining (lane 3). (15) Kukolka, F.; Niemeyer, C. M. Org. Biomol. Chem. 2004, 2, 2203-2206. (16) Bongartz, J. P.; Aubertin, A. M.; Milhaud, P. G.; Lebleu, B. Nucleic Acids Res. 1994, 22, 4681-4688. (17) Tung, C. H.; Rudolph, M. J.; Stein, S. Bioconjugate Chem. 1991, 2, 464465.

The results indicate that the prepared gA-G conjugate is highly homogeneous, containing a single guide oligonucleotide. SPR Analysis of Protein G-DNA Hybridization on DNAGold Surface. Purified protein G-DNA conjugate gA-G as well as both guide oligonucleotides gA and gB were treated to the SPR gold chip surface modified by complementary oligonucleotide cA. As shown in Figure 3A, the cA-covered surface shows good sequence specificity, capturing only oligonucleotide gA, not gB. Based on the RU value, the hybridization density of gA (7.5 kDa) is calculated to be 1.8 × 1010 molecules/mm2, which lies within the density range obtained from previous reports.6,18 The proteinDNA conjugate gA-G (21.5 kDa) yielded a higher SPR signal change (564 RU) than gA (231 RU) upon hybridization due to the higher molecular weight. The calculated hybridization density of gA-G is ∼1.6 × 1010 molecules/mm2, slightly lower than that of gA. The protein G constructs conjugated to oligonucleotides likely cause a low level of steric hindrance during the hybridization process. The protein G layer formed on the DNA-modified gold chip surface was able to capture various types of antibodies (Figure 3B, see Supporting Information for the binding profile of protein G to IgGs from various sources; Table S1). Human and goat IgGs were captured on the protein G surface slightly better than mouse and rat IgGs, where this pattern is consistent with previously reported binding constants of these IgGs against protein G.19 Although binding affinity of rat IgG (Kd ) 0.7 nM)19 is nearly 60-fold weaker than that of human IgG, the binding is still strong enough that a high level of rat IgG was immobilized on the protein G surface (Figure 3B). In addition, a high concentration of BSA (1 mg/mL) was introduced to the surface without any SPR responses, showing a good resistance against nonspecific interactions other than antibody-protein G interaction. Antibodies can also be directly immobilized on the DNA-modified gold surface in one step by preincubating IgGs with protein G-DNA conjugates (data not shown). Furthermore, sequencespecific hybridization of protein G-DNA conjugates with captured antibodies was observed in array formats by using SPR imaging analysis (Supporting Information; Figure S1). SPR Analysis of Protein G-DNA Conjugates for Antibody/ Antigen Binding. Antibody and subsequent antigen binding to the hybridized protein G on the DNA-gold surface was also investigated by SPR analysis. Human prostate-specific antigen (PSA) was used as a target analyte. Different levels of protein G-DNA conjugates were hybridized on the DNA surface, and ensuing anti-PSA/PSA immobilization was examined in detail with a particular attention to immobilization densities upon each interaction. The surface density of the protein G-DNA conjugate was controlled by varying the conjugate concentration and incubation time (Figure 4). Introduction of 50 nM gA-G for 10 min resulted in 707 RU increases of SPR signals, which corresponded to the binding of 2.0 × 1010 molecules/mm2 on the chip surface (Figure 4). About 2.4-fold lower surface coverage of gA-G (0.84 × 1010 molecules/mm2) was achieved by injecting 10 nM gA-G for 7 min to the same surface. SPR responses during anti-PSA (150 kDa) and PSA (36 kDa) immobilization on these gA-G surfaces are shown in Figure 4A, and the resulting RU shift values with calculated surface densities are summarized in (18) Herne, T. M.; Tarlov, M. J. J. Am. Chem. Soc. 1997, 119, 8916-8920. (19) Akerstrom, B.; Bjorck, L. J. Biol. Chem. 1986, 261, 10240-10247.

Analytical Chemistry, Vol. 79, No. 17, September 1, 2007

6537

Figure 3. SPR responses upon (A) gA-G, gA, and gB bindings to the cA-modified gold surface and (B) the immobilization of various antibodies to the gA-G hybridized gold surface. Antibodies are human immunoglobulin G 1 (HIgG1), goat IgG (GIgG), mouse IgG2 (MIgG2), and rat IgG (rIgG). 1 mg/mL BSA was treated to the gA-G hybridized gold surface to test nonspecific interaction.

Figure 4C. At the high gA-G surface, an average of ∼0.44 molecule of goat anti-PSA IgG was captured by 1 gA-G molecule while ∼0.58 anti-PSA was immobilized by 1 gA-G at the low conjugate surface. Since an antibody has two protein G binding sites, two gA-G conjugates might participate for one antibody immobilization.20 In addition, there are likely more steric hindrances between bound antibodies at a high protein density, contributing to the low antibody capturing efficiency by gA-G on the high conjugate density surface. Interestingly, however, the antigen (PSA) binding efficiency of anti-PSA was not reduced by increased antibody surface density (0.75-0.71 PSA/anti-PSA; Figure 4C). The antigen binding efficiency varies with the nature of the antigen and antibody.9 At a surface density of 0.89 × 1010 molecules/mm2 (high gA-G) for anti-PSA, single antibody occupies 10 nm × 10 nm of the gold surface on average while 14 nm × 14 nm of the surface can be used for one anti-PSA at the low antibody density (0.50 × 1010 molecules/mm2). Retained PSA binding ability at the even high anti-PSA density suggests that bound anti-PSA IgGs are uniformly oriented regardless of IgG densities, and 10 nm × 10 nm of the gold surface provides enough space for this orientation. Considering the molecular size of the

antibody (15 nm × 7 nm × 3.5 nm) and Fc-specific binding of protein G, antibodies linked on the gold surface by our protein G-DNA conjugate are likely to be oriented for efficient antigen binding (nearly straight upward immobilization) as depicted in Figure 1B. The anti-PSA/PSA binding efficiency of the present protein G-DNA conjugate surface was next compared with that of a conventional protein G layer formed on a gold surface. We previously reported that a well-oriented protein G layer can be obtained by chemically linking cysteine-tagged protein G constructs to a carboxylated gold surface with a carboxyl/thiol-specific reagent PDEA through the disulfide bond formation.21 Our protein G constructs containing an N-terminus cysteine residue can therefore be immobilized on a gold surface with a uniform orientation. For proper comparison, a gold surface was modified with oligo(ethylene glycol) thiols, similar to the DNA-modified surface. Oriented protein G surfaces were formed with protein densities comparable to those of gA-G surfaces (Figure 4C) by using the PDEA method as described in the Experimental Section. Anti-PSA and PSA were subsequently treated to these PDEAprotein G surfaces (Figure 4B) as performed with the gA-G

(20) Sauer-Eriksson, A. E.; Kleywegt, G. J.; Uhlen, M.; Jones, T. A. Structure 1995, 3, 265-278.

(21) Lee, J. M.; Park, H. K.; Jung, Y.; Kim, J. K.; Jung, S. O.; Chung, B. H. Anal. Chem. 2007, 79, 2680-2687.

6538

Analytical Chemistry, Vol. 79, No. 17, September 1, 2007

Figure 4. SPR responses upon anti-PSA/PSA binding. The 100 nM anti-PSA and subsequently the same concentration of PSA were treated to (A) the gA-G-modified gold surfaces and (B) the oriented protein G surfaces formed by PDEA. (C) SPR responses and calculated surface densities for each protein immobilization. The values of anti-PSA/gA-G or protein G were obtained from surface densities of anti-PSA divided by surface densities of gA-G or protein G. The values of PSA/anti-PSA were obtained similarly with surface densities of PSA. The values are the means from three independent experiments with standard deviations.

surfaces. Anti-PSA IgG capturing efficiencies of these protein G constructs are, however, over 2-fold lower than those of protein G-DNA conjugates (Figure 4C). For instance, at the nearly same protein G density, one gA-G captures 0.44 anti-PSA while only 0.17 anti-PSA was captured by protein G. In addition, PSA immobilization by bound anti-PSA is also more effective on the protein G-DNA surface. Enhanced antibody/antigen binding ability of the protein G-DNA conjugate compared to chemically attached protein G, where protein G molecules are properly oriented for both cases, is possibly due to the rigid doublestranded DNA spacer between the gold surface and protein G. Protein G binds to the middle of the Fc region of IgG,20 requiring at most 5 nm of space for the straight upward antibody immobilization. The 24-base pair of the current DNA linker can provide almost 8 nm of space, which will offer enough room to accommodate the upward orientations for antibody immobilization (Figure 1B). Antibody Targeting on a Glass Surface by Protein G-DNA Conjugate. SPR studies clearly showed that the present protein G-DNA conjugates afford simple but highly efficient antibody immobilization on a DNA-modified gold surface without antibody modification. We next explored possible applications of the conjugate on various other surface platforms. Antibody patterning on a glass surface was first achieved from a simple model DNA array. Amine-modified complementary oligonucleotides cA and cB were spotted on an epoxide glass surface followed by BSA blocking. Cy3-labeled mouse IgG was directed only to the cA spot by the gA-G conjugate as shown in Figure 5A. Moreover, fluorescence intensity at the cB spot was even lower than the background signal, indicating that nonspecific IgG interaction to

Figure 5. Fluorescence imaging analyses. (A) Antibody immobilization. Cy3-labeled mouse IgG (1 nM) was preincubated with gA-G (100 nM) and treated to the DNA-spotted glass slide. The fluorescence profile of the area covering both cA and cB DNA spots is plotted in the right box. (B) Anti-PSA/PSA immobilization. Anti-PSA (150 nM) was preincubated with gA-G or gB-G (100 nM) and treated to the DNA-spotted glass slide. The 1 nM Cy3-labeled PSA was subsequently added to the glass slide.

the DNA-modified glass surface is weaker than to the BSA-blocked glass surface. Anti-PSA capturing on the complementary DNA spot assisted by the protein G-DNA conjugate and subsequent PSA Analytical Chemistry, Vol. 79, No. 17, September 1, 2007

6539

Figure 6. Anti-PSA/gA-G cross-linking. (A) Anti-PSA was immobilized onto DNA-modified glass slides by introducing either anti-PSA preincubated with gA-G (left side panel) or anti-PSA/gA-G cross-linked with increasing amounts of DSS (right side panels) as described in the Experimental Section. (B) Purified anti-PSA/gA-G complex (left side panel) and anti-PSA/gA-G cross-linked with DSS (right side panel) were incubated with gB-G prior to the treatment onto DNA-modified glass slides. The 1 nM Cy3 labeled PSA was subsequently added to each glass slide.

interaction was also confirmed by using Cy3-labeled PSA protein (Figure 5B). The data suggest that well-developed DNA immobilization techniques on glass surfaces can be transformed to IgG immobilization tools in conjunction with our protein G-DNA conjugate. One of main advantages of DNA-directed protein immobilization is the sequence diversity of available DNA linkers. Multiple proteins, where each protein is modified by a different guide oligonucleotide, can be simultaneously targeted to the complementary oligonucleotide spots.7 Unlike single antibody immobilization (Figure 1), antibodies and protein G-DNA conjugates must be incubated together prior to the treatment to DNA arrays to achieve multiple antibodies addressing (see Supporting Information for strategic explanations; Figure S2). To test whether the present conjugate can be employed for this multiplexed protein targeting, anti-PSA/gA-G complex was first formed and unbound anti-PSA was removed by a His affinity column. Resulting complex was co-treated with gB-G (the protein G-guide oligonucleotide B conjugate) to the cA/cB-modified surface. Antibody once bound to one conjugate (gA-G), however, was able to move to another conjugate containing different DNA (gB-G) during the DNAdirected immobilization as demonstrated in Figure 6B (left panel). It is likely because the protein G-DNA conjugate interacts with an antibody through noncovalent bonding. To prevent this shuffling, the anti-PSA/gA-G complex was covalently locked by bifunctional amine/amine-specific cross-linker DSS. DSS has been successfully used to covalently attach antibodies onto protein G beads for immunoprecipitation.22,23 Anti-PSA/gA-G complexes treated by varying concentration of DSS (see Experimental Section) were purified again by a His affinity column and (22) Kerkela, R.; Pikkarainen, S.; Majalahti-Palviainen, T.; Tokola, H.; Ruskoaho, H. J. Biol. Chem. 2002, 277, 13752-13760. (23) Perez, R. G.; Waymire, J. C.; Lin, E.; Liu, J. J.; Guo, F.; Zigmond, M. J. J. Neurosci. 2002, 22, 3090-3099.

6540

Analytical Chemistry, Vol. 79, No. 17, September 1, 2007

successfully directed to the cA spot as shown in Figure 6A. Moreover, anti-PSA linked on gA-G by DSS did not shuffle to gB-G, resulting in anti-PSA/Cy3-PSA addressing only to the cA spot (Figure 6B; right panel). Although cross-linked anti-PSA/ gA-G complex can still maintain its orientation, the antibody can also be overly modified by DSS during the process, lowering the antigen binding ability. Alternative strategies are needed to improve multiplexed antibody addressing, which is currently under investigation in our laboratory. Controlled Antibody Labeling of DNA-Functionalized Gold Nanoparticles. AuNPs have attracted great attention for the past several years. They are used as core building blocks for nanostructure assemblies, probes for numerous biosensor systems, and recently even antisense agents for gene regulation.13,24 Antibody labeling of DNA-gold nanoparticles is essential for most diagnostic sensing systems, which employ these nanoparticles. Moreover, controlled antibody modification will greatly expand the possible applications of these particles. Here the protein G-DNA conjugate was used to modify DNA-functionalized gold nanoparticles with IgG. As shown in Figure 7A, cA-modified gold nanoparticles (AuNP-cA) are hybridized only with gA and gAG, showing good sequence specificity on the particle surface. As expected from the SPR data, gA-G (21.5 kDa) hybridization causes a larger band shift than that by gA (7.5 kDa) hybridization to cA-gold nanoparticles. The DNA loading process onto gold nanoparticles has been extensively studied, and multiple strategies were reported to control the nature of loaded DNA.25,26 Here we prepared two cAmodified gold nanoparticles AuNP-cA-I and AuNP-cA-II, where (24) Rosi, N. L.; Mirkin, C. A. Chem. Rev. 2005, 105, 1547-1562. (25) Hurst, S. J.; Lytton-Jean, A. K.; Mirkin, C. A. Anal. Chem. 2006, 78, 83138318. (26) Hazarika, P.; Giorgi, T.; Reibner, M.; Ceyhan, B.; Niemeyer, C. M. Methods Mol. Biol. 2004, 283, 295-304.

Unconjugated proteins such as antibodies remaining in the reaction mixture compete with the labeled particles for binding to their specific antigen, reducing the sensing signals. Although gold nanoparticles can be separated from unreacted proteins by simple centrifugation, it frequently suffers from particle aggregation, caused by metal-metal, metal-protein, or protein-protein bindings. DNA-functionalized AuNP, however, can be repeatedly collected by centrifugation and resuspended in an aqueous solution.26 In the present study, gold nanoparticles hybridized with protein G-DNA conjugates were also stable during washing steps via centrifugation despite the presence of protein G on the particle surface, which is a very useful property for particle modification. Antibody-labeled AuNPs, however, quickly aggregated by centrifugation as many other antibody-modified nanoparticles. Here the protein G conjugate, however, contains a 6× His tag, allowing simple isolation of IgG-labeled AuNPs from unconjugated IgGs by using His binding resins (data not shown). For the preparation of IgG-conjugated AuNPs, the protein G-DNA linker developed in the present study provides a good antibody orientation, flexible IgG surface density, and stable purification methods.

Figure 7. Agarose gel analysis of the formation of antibody-labeled AuNP. (A) 2% agarose gel analysis of cA-modified gold nanoparticles (AuNP-cA) treated with gA, gB, gA-G, and gB-G. AuNP-cA and gAG-hybridized AuNP-cA (AuNP-cA-G) are indicated by arrows. (B) 1% agarose gel analysis of cA-modified gold nanoparticles (AuNP-cA-I and -II) treated with gA-G and subsequently human IgG. gA-G hybridized nanoparticles were isolated from uncoupled gA-G before the IgG binding. (C) Schematic representation of IgG-labeled AuNPcA-I and AuNP-cA-II.

each particle can accommodate an average of 21 or 9.5 molecules of hybridized gA oligonucleotides, respectively (Figure 7C). An estimated diameter of 13-nm AuNP covered with cA is ∼28 nm. On the surface of AuNP-cA-I, one molecule of hybridized gA can occupy ∼11 nm × 11 nm of the surface dimension of the particle, which is enough space for gA-G and subsequent antibody binding as discussed above. cA-modified AuNPs were hybridized with gA-G and subsequently labeled with human IgGs. Differently loaded nanoparticles with protein G conjugates and IgGs were again identified on an agarose gel (Figure 7B). AuNP-cA-I clearly shows higher levels of gA-G and human IgG labeling on the particle surface than AuNP-cA-II as shown by more retarded migrations. The last step of protein conjugation to nanoparticles is often to remove unconjugated proteins from the modified nanoparticles.

CONCLUSION In the present study, single-stranded DNA was coupled to the N-terminus of the protein G construct, bearing two repeated antibody binding domains. DNA-directed antibody immobilization was achieved on the surfaces on gold chips, glass slides, and also nanoparticles by using this protein G-DNA conjugate. The conjugate allowed DDI of IgGs without any IgG modifications, preserving the antibody activity. More importantly, detailed SPR analysis demonstrated that protein G constructs linked to the surface through hybridized double-stranded DNA showed excellent spacing and orientation for the antibody/antigen binding, which were even better than those of protein G site-specifically linked to the surface by chemical coupling. Gold nanoparticles were easily manipulated with antibodies by using the protein G-DNA conjugate as a versatile handle for antibody labeling and purification. The present protein G-DNA construct provides a novel linker system to direct antibody immobilization on the surface of various immunosensors. ACKNOWLEDGMENT This research has been supported by grants from BioGreen21 Program funded by the Rural Development Administration and the Collaborative R&D Program of KRCF, Republic of Korea. We thank Jinyoung Jeong for providing gold nanoparticles. SUPPORTING INFORMATION AVAILABLE Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org.

Received for review March 9, 2007. Accepted July 2, 2007. AC070484I

Analytical Chemistry, Vol. 79, No. 17, September 1, 2007

6541