Open-Sandwich Enzyme Immunoassay for One-Step Noncompetitive

Sep 24, 2009 - Department of Chemistry and Biotechnology, School of Engineering, The ... of Pharmaceutical Sciences, Tohoku University Hospital, 1-1 ...
0 downloads 0 Views 688KB Size
Download PDF
Anal. Chem. 2009, 81, 8298–8304

Open-Sandwich Enzyme Immunoassay for One-Step Noncompetitive Detection of Corticosteroid 11-Deoxycortisol Masaki Ihara,† Tatsuya Suzuki,‡ Norihiro Kobayashi,§ Junichi Goto,| and Hiroshi Ueda*,†,‡,⊥ Department of Chemistry and Biotechnology, School of Engineering, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8656, Japan, Department of Bioengineering, School of Engineering, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8656, Japan, Kobe Pharmaceutical University, 4-19-1, Motoyama-Kitamachi, Higashinada-ku, Kobe 658-8558, Japan, Department of Pharmaceutical Sciences, Tohoku University Hospital, 1-1 Seiryo-machi, Aoba-ku, Sendai 980-8574, Japan, and PRESTO, JST, Sanbancho Building 5-Sanbancho, Chiyoda-ku, Tokyo 102-0075, Japan A noncompetitive immunoassay has the potential for improved sensitivity and working range compared with corresponding competitive assays. However, monovalent antigens with less than 1000 in molecular weight are not susceptible to sandwich assays due to their small size. As a noncompetitive immunoassay that can be performed with a clone of an antibody, an open-sandwich immunoassay (OS-IA) based on the antigen-dependent stabilization of the antibody variable region (VH + VL) was applied to the quantification of 11-deoxycortisol (11DC; Mr 346.5), a corticosteroid serving as a diagnostic index for pituitary-adrenal function, as a model target hapten. By one step OS-IA detection of enzyme-labeled VH fragment bound to immobilized VL in the presence of sample in microplate wells, 11-DC was measured with a femtomolar detection limit and the working range was wider than that with corresponding competitive assay. In addition, the selectivity against analogues was found almost identical to that of conventional assays. The effect of the mutagenesis of a VH residue at the VH/VL interface to reduce background signal was also shown, implying the wider application of OS-IA in small molecule analyses. The characterization of trace amounts of physiologically active substances with low molecular weights, including steroids, catecholamines, eicosanoids, thyroid hormones, and synthetic drugs, is critical in biomedical analyses. Immunoassays are now widely used for this purpose, due to their excellent specificity, feasibility, and higher sensitivity than other methods. The measurement of such small molecules, immunochemically classified as haptens, is typically dependent on competitive reactions between an unlabeled antigen (analyte) and a labeled antigen in the presence * Corresponding author. Hiroshi Ueda, e-mail: [email protected]. Phone/fax: +81-3-5841-7362. † Department of Bioengineering, School of Engineering, The University of Tokyo. ‡ Department of Chemistry and Biotechnology, School of Engineering, The University of Tokyo. § Kobe Pharmaceutical University. | Tohoku University Hospital. ⊥ PRESTO, JST.

8298

Analytical Chemistry, Vol. 81, No. 20, October 15, 2009

of limiting amounts of antihapten antibody. The sensitivity of such competitive immunoassays, however, is limited by the affinity of the antihapten antibodies.1-4 Noncompetitive immunoassays, also called “immunometric assays”, are potentially more sensitive than competitive assays.2 Immunometric assays are classified into two categories, two-site immunometric assays5 and single-antibody (single-epitope) immunometric assays.6 Both are based on the reaction of an analyte in the presence of excess amounts of a relevant antibody labeled with a signal-generating group. The use of excess antibody promotes the antigen-antibody reaction to capture even trace amounts of antigen; the formation of “antigen-antibody complexes” is selectively and directly monitored by signal intensity. Higher precision and a more broad working range of detection, both of which result from the use of excess antibody, are additional advantages of immunometric assays. However, conventional immunometric assays as described have common problems such as the need of more than single step and tend to be complex and time-consuming assays.4,7,8 As a novel type of immunometric assay approach, previously we proposed the open sandwich-immunoassay (OS-IA).9-11 Briefly, this assay exploits reassociation of the antibody variable region fragments VH-VL by a bridging antigen. By OS-ELISA (OS-IA performed in microplates) using immobilized VL and enzymetagged VH fragments, one can measure less than 10 ng/mL of antigen in a shorter time period than using a conventional sandwich assay, due to the omission of an incubation/washing cycle. Also, the assay was found to be compatible with a (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11)

Berson, S. A.; Yalow, R. S. J. Clin. Invest. 1959, 38, 1996–2016. Ekins, R. Nature 1980, 284, 14–15. Jackson, T.; Ekins, R. J. Immunol. Methods 1986, 87, 13–20. Kobayashi, N.; Goto, J. Adv. Clin. Chem. 2001, 36, 139–170. Maiolini, R.; Masseyeff, R. J. Immunol. Methods 1975, 8, 223–234. Schuurs, M.; Van Weemen, B. K. U.S. Patent 3,654,090, 1972. Giraudi, G.; Anfossi, L.; Rosso, I.; Baggiani, C.; Giovannoli, C.; Tozzi, C. Anal. Chem. 1999, 71, 4697–4700. Gonzalez-Techera, A.; Vanrell, L.; Last, J. A.; Hammock, B. D.; GonzalezSapienza, G. Anal. Chem. 2007, 79, 7799–7806. Suzuki, C.; Ueda, H.; Tsumoto, K.; Mahoney, W.; Kumagai, I.; Nagamune, T. J. Immunol. Methods 1999, 224, 171–184. Ueda, H. J. Biosci. Bioeng. 2002, 94, 614–619. Ueda, H.; Tsumoto, K.; Kubota, K.; Suzuki, E.; Nagamune, T.; Nishimura, H.; Schueler, P. A.; Winter, G.; Kumagai, I.; Mahoney, W. C. Nat. Biotechnol. 1996, 14, 1714–1718. 10.1021/ac900700a CCC: $40.75  2009 American Chemical Society Published on Web 09/24/2009

Table 1. Nucleotide Sequences of the Primers Useda

Figure 1. Structures of 11-DC and the BSA conjugate (11-DC-BSA).

number of antihapten antibodies and could attain a similar or lower detection limit as well as a wider working range.12-14 However, the applicability of OS-ELISA to more clinically important compounds such as steroids has been elusive, and more importantly, the selectivity of the assay to structurally similar analogues has not been compared to the corresponding competitive assay yet. Here we report the development of an OS-ELISA for 11deoxycortisol (11-DC; Mr 346.5), which is a corticosteroid used as a diagnostic index for pituitary-adrenal function as a clinically important hapten molecule, which lead to the detection of 11DC with a similar or higher sensitivity (detection limit, 9 fmol/ assay) than conventional competitive immunoassays.15,16 After the cloning of VH/VL genes by phage display technology, we generated fusion proteins VH-AP and MBP-VL, which combined the VH domain specific for 11-DC with alkaline phosphatase (AP), and the VL domain with maltose binding protein (MBP), respectively, both derived of Escherichia coli (E. coli). After reaction of VH-AP with the surface immobilized with MBP-VL, the enzymatic activities generated by the immune complexes, which increased with increasing 11-DC amounts during the reaction, were measured by chemiluminescence. Furthermore, by using the MBP-VH mutant fusion protein labeled with horseradish peroxidase (HRP), we could evaluate the 11-DC concentration in the serum of normal individuals. The present OS-IA afforded an extraordinarily broad measurable range and a high practical specificity. EXPERIMENTAL SECTION Reagents. Unlabeled steroids including 11-DC (Figure 1) were purchased from Sigma (St. Louis, MO). CDP star substrate and reaction buffer were obtained from New England Biolabs (Tokyo, Japan). Block Ace and Immuno-Block (blocking solutions) were purchased from DS Pharma (Osaka, Japan). The 11-DC and bovine serum albumin (BSA) conjugate (11-DC-BSA) linked via the 4-position on the steroid nucleus was prepared previously.17 All other reagents and solvents were of analytical grade. (12) Aburatani, T.; Sakamoto, K.; Masuda, K.; Nishi, K.; Ohkawa, H.; Nagamune, T.; Ueda, H. Anal. Chem. 2003, 75, 4057–4064. (13) Suzuki, C.; Ueda, H.; Mahoney, W.; Nagamune, T. Anal. Biochem. 2000, 286, 238–246. (14) Suzuki, T.; Munakata, Y.; Morita, K.; Shinoda, T.; Ueda, H. Anal. Sci. 2007, 23, 65–70. (15) Fiet, J.; Giton, F.; Boudou, P.; Villette, J. M.; Soliman, H.; Morineau, G.; Boudi, A.; Galons, H. J. Steroid Biochem. Mol. Biol. 2001, 77, 143–150. (16) Kobayashi, N.; Shibahara, K.; Ikegashira, K.; Shibusawa, K.; Goto, J. Steroids 2002, 67, 733–742. (17) Hosoda, H.; Miyairi, S.; Kobayashi, N.; Nambara, T. Chem. Pharm. Bull. (Tokyo) 1982, 30, 2127–2132.

primer name

nucleotide sequence

MuIgGVH3′-2 MH2Back MkCfor Vk4BkFL2 DOCVHBack VH1For-2 DOCVLBack JK2Not10 OlinkBack2 LinkForA H39Back SplitVLseqFor M13RV H39For MkBackEcoSal MycForHind DOCHE5RV VH1For2Hind

5′-cccaagcttccagggrccarkggataracigrtgg-3′ 5′-sargtnmagctgsagsagtcwgg-3′ 5′-ggatacagttggtgcagcatc-3′ 5′-gacgacgattaaaggagatatcatatggayattgtgmtsacmcarwctmc-3′ 5′-gccggccatggcccaaattcagctggtgcagtcaggacctgag-3′ 5′-tgaggagacggtgaccgtggtcccttggcccc-3′ 5′-caagctcagtcgacgtccattgtgatgacccagactcccaaattcctgc-3′ 5′-gagtcattctgcggccgcccgttttatttccagcttggtccc-3′ 5′-ggggccaagggaccacggtcaccgtctcgagcggtgctgctgactac-3′ 5′-tccgtcgactgagcttgcgcaac-3′ 5′-ctgggtgaagarggctccaggaa-3′ 5′-cctcttctgagatgagtttttgttct-3′ 5′-caggaaacagctatgaccatgattacg-3′ 5′-ttcctggagccytcttcacccag-3′ 5′-ggggggaattcgcgtcgacggayattgtgmtsacmcarwctmca-3′ 5′-gggggcaagcttgcttatgcggccccattcagatcc-3′ 5′-ggggggatatccaaattcagctggtgcagtca-3′ 5′-gggggaagcttgctcgagacggtgaccgtggt-3′

a The underline shows the recognition sites for restriction enzymes. The mutated codons are shown in bold.

Buffers. PBS, 10 mM sodium phosphate buffer containing 137 mM NaCl and 2.68 mM KCl, pH 7.4. MPBS, PBS containing skim milk. PBST, PBS containing 0.05 (v/v)% Tween 20. IPBS, PBS containing Immuno-Block. Materials. Mouse hybridoma S.CET.M8.1.1 secreting anti-11DC IgG CET-M818 was used for RNA extraction. Escherichia coli strain TG1 (GE Healthcare, Tokyo, Japan) was used for phage production and also for the production of MBP-VL, HB2151 (GE Healthcare) for VH-displaying phage/VL production, XL-10 Gold (Stratagene, La Jolla, CA) for recombinant DNA preparation, and BL21(DE3, pLysS) (Novagen, Merck, San Diego, CA) for the production of VH-AP and MBP-VH. Cloning of Antibody V-Genes. The cDNAs of variable regions were prepared from the total RNA derived from hybridomas using OneStep RT-PCR Kit (Qiagen) according to the manufacturer’s instructions, with primers MuIgGVH3′-2 (IgPrime, Novagen) and MH2Back19 for the amplification of VH, and MkCfor19 and Vk4BkFL2 for the amplification of VL (Table 1). The reaction conditions were once at 50 °C for 45 min and 94 °C for 15 min and 35 cycles of 94 °C for 1 min, 50 °C for 1 min and 72 °C for 2 min, followed by 72 °C for 10 min. For the subsequent splice overlap extension (SOE) PCR, the gel-purified bands were reamplified with primers DOCVHBack and VH1For-212 for VH, and with DOCVLBack and JK2Not1020 for VL. The gel-purified fragments were then assembled by SOE PCR using a split Fv linker including gene 9 that was preamplified with primers OlinkBack2 and LinkForA essentially as described.12 The purified split Fv fragments were then digested with restriction enzymes NcoI and NotI, gel purified again, and ligated with phagemid pKST212 digested with the same using ligation high (Toyobo, Osaka, Japan). The ligation product was then used to transform the TG1 or HB2151 strain and plated on 2YTAG agar (16 g/L tryptone, 10 g/L yeast extract, 5 g/L NaCl, pH 7.2 (18) Hosoda, H.; Kobayashi, N.; Nambara, T.; Sawada, J.; Terao, T. Chem. Pharm. Bull. (Tokyo) 1984, 32, 381–382. (19) Wang, Z.; Raifu, M.; Howard, M.; Smith, L.; Hansen, D.; Goldsby, R.; Ratner, D. J. Immunol. Methods 2000, 233, 167–177. (20) Pope, A. R.; Embleton, M. J.; Mernaugh, R. In Antibody Engineering; McCafferty, J., Hoogenboom, H. R., Chiswell, D. J., Eds.; IRL Press: Oxford, U.K., 1996; Vol. 169, pp 1-40.

Analytical Chemistry, Vol. 81, No. 20, October 15, 2009

8299

supplemented with 100 µg/mL ampicillin, 1% glucose, and 1.5% agar) plates overnight at 37 °C. Split-Fv Phage Display. Single colonies of E. coli TG1 or HB2151 strain were inoculated to 4 mL of 2YTAG and cultured at 37 °C, before confluent culture was 100-fold diluted with 2YTAG and incubated at 37 °C until OD600 reached approximately 0.5. After helper phage M13KO7 was added to multiplicity of infection (m.o.i.) of 20, the cells were left undisturbed for 30 min at 37 °C and centrifuged at 2 500g for 15 min. The cells were resuspended in 2YTAK (2YT containing 100 µg/mL ampicillin and 50 µg/mL kanamycin) and vigorously shaken for 16 h at 30 °C. The overnight culture containing either split Fv-displaying phage (in the case of TG-1 as a host) or VH-displaying phages as well as soluble VL (in the case of HB2151 as a host) were centrifuged at 10 800g for 10 min and the supernatants stored at 4 °C. Split-Fv Phage Enzyme-Linked Immunosorbent Assay (ELISA). Falcon 353914 (BD Biosciences) 96-well microplate was incubated with or without 100 µL per well of 1 µg/mL 11-DC-BSA conjugate in PBS overnight at 4 °C, blocked with 2% skim milk in PBS (MPBS) for 2 h at 25 °C and washed once with PBS containing 0.1% Tween-20 (Sigma) (PBST). Then the dilutions of 11-DC prepared in 95-97.5 µL of 1% MPBS, and 2.5-5 µL of split Fv phages from TG1 were added to respective wells (total 100 µL) in duplicate and incubated for 90 min at 25 °C. Afterward, the wells were washed five times with PBST and incubated with 100 µL of 5000-fold diluted HRP-conjugated mouse anti-M13 monoclonal antibody (GE Healthcare) in 1% MPBS for 1 h at 25 °C. The microplate was then washed three times with PBST and developed with 100 µL of 100 µg/mL 3,3′,5,5′-tetramethylbenzidine (TMB, Sigma) and 0.04 µL/mL H2O2 in 100 mM NaOAc, pH 6.0. The reaction was stopped with 50 µL of 1 M H2SO4 after incubation between 5 and 30 min, and the absorbance was read at 450 nm with a reference at 655 nm using a microplate reader (model 680, Bio-Rad, Tokyo, Japan). Open Sandwich Phage ELISA. The 96-well microplate was incubated with 100 µL per well of 1 µg/mL mouse antimyc 9E10 monoclonal antibody (Sigma) overnight at 4 °C and blocked with 2% MPBS for 2 h at 25 °C. The microplate was then washed once with PBST and 10 µL of 10-times 11-DC solutions originally dissolved at 1 mg/mL in 50% ethanol and diluted to various concentrations in 2% MPBS were added to respective wells. Then 20-50 µL of culture supernatant from HB2151 was added to respective wells and gently mixed after adjusting the total volume to 100 µL. The microplate was incubated for 2 h at 25 °C and washed three times with PBST, and bound phages were detected as mentioned above. Site-directed mutagenesis of H39 position was performed by SOE PCR using primers H39Back and H39For, where the semirandomized codon is shown in bold. Briefly, the two fragments amplified with primer pairs M13RV/H39For and H39Back/ SplitVLseqFor, using pKST2(11-DC) as a template, were assembled by SOE PCR using primers M13RV and SplitVLseqFor. The amplified fragment was digested by NcoI and NotI, inserted to pKST2(11-DC) digested by the same, and determined for the sequence by Shimadzu Corp. (Kyoto, Japan). The obtained phagemids were used to prepare phage culture supernatant to perform OS-ELISA as above. 8300

Analytical Chemistry, Vol. 81, No. 20, October 15, 2009

Preparation of the Fusion Proteins. To construct MBP-VL expression vector pMAL-VL(11-DC), the VL fragment was amplified with primers MkBackEcoSal and MycForHind from phagemid pKST2(11-DC), digested by EcoRI and HindIII, gelpurified, and ligated with pMAL-p2 (New England Biolabs, Ipswich, MA), which has been digested with the same enzymes. E. coli TG1 was transformed with the ligation product, plated on 2YTAG agar, and incubated overnight at 37 °C. Several colonies were picked, and the extracted plasmids were confirmed for their nucleotide sequence. To obtain VH-AP expression vector pVH(11DC)-AP, the VH gene was PCR amplified from phagemid pKST2(11-DC) using primers DOCHE5RV and VH1For2Hind. The fragment was digested with EcoRV and HindIII, gel purified, and ligated with pVH-PhoA(D101S),9 which has been digested with the same enzymes. An objective recombinant was screened as above. To obtain pET-based expression vector for MBP-VH(11-DC/ 39K), pET-MBPp-VH(BPA)21 and pKST2(11-DC/H39K) were used. A periplasmic expression vector pET-MBPp-VH(BPA) to secrete MBP-VH (for bisphenol A, BPA) protein was digested with NcoI and XhoI to remove the fragment encoding the VH for BPA and ligated with the NcoI-XhoI fragment of pKST2(11DC/H39K) encoding the VH for 11-DC harboring a Q39K mutation, resulting in pET-MBPp-VH(11-DC/39K). With the use of the obtained expression vectors, MBP-VL, VH-AP, and MBP-VH(39K) proteins were expressed and purified as described.22 The purified MBP-VH was labeled with HRP using a peroxidase-labeling kit-NH2 (Dojindo, Kumamoto, Japan) according to the manufacturer’s instructions. After the coupling reaction, the reaction mixture was purified with Talon affinity resin (Clontech, Mountain View, CA) according to the manufacturer and buffer-exchanged by using a PD-10 column equilibrated with PBS (GE Healthcare). The quantified protein solution was added with an equal volume of glycerol and stored at -20 °C until use. Open Sandwich ELISA Using Purified Proteins. OS-ELISA with MBP-VL and VH-AP proteins was performed as described.22 Briefly, a white 96-well microplate (Maxisorp, Nunc, Denmark) was incubated with 100 µL of 2.5 µg/mL MBP-VL protein in PBS overnight at 4 °C and blocked with 30% Immunoblock (DS Pharma, Osaka, Japan) in PBS (IPBS) for 1 h at 25 °C. The microplate was washed three times with PBST and 50 µL of 2× sample at various concentrations diluted in PBS, and 50 µL of 5 µg/mL purified VH-AP protein in IPBS was added to each well. The microplate was incubated for 90 min at 25 °C, washed five times with PBST, once with CDP star buffer, and 100 µL of CDP star substrate solution (New England Biolabs) was added to each well and incubated for 30 min at 25 °C in the dark, before being measured for chemiluminescence for 5 s using a microplate luminometer AB-2100 (ATTO, Tokyo, Japan). Preparation of Serum Samples. The extraction of 11-DC from human serum was carried out by methylene chloride. Human serum (10 mL) was mixed with 3 mL of methylene chloride and vortexed, and the organic phase was transferred to a new tube. (21) Sakata, T.; Ihara, M.; Makino, I.; Miyahara, Y.; Ueda, H. Anal. Chem. 2009, 81, 7532–7537. (22) Lim, S.-L.; Ichinose, H.; Shinoda, T.; Ueda, H. Anal. Chem. 2007, 79, 6193– 6200.

The extract was dried in vacuo on a rotating wheel, suspended into 10 µL of dimethylsulfoxide (DMSO), with 90 µL of ethanol added and diluted to 9.9 mL of PBST. For 10× concentrated extracts, the dried pellet was suspended similarly to 1 µL of DMSO, 9 µL of ethanol, and 0.99 mL of PBST. To make 11-DC-free serum, pooled human serum (DS Pharma) was treated with Norit activated charcoal (neutralized powder, 100400 mesh, Nacalai Tesque, Kyoto, Japan) at final concentration of 1% (w/v) and gently rotated at 4 °C overnight. Subsequently, the serum was centrifuged and filtered through a 0.45 µm hydrophilic poly(vinylidene fluoride) (PVDF) filter (Millipore, Billerica, MA). Estimation of 11-DC Concentration in Serum. A 96 well microplate (Costar, Corning, NY) was incubated with 2 µg/mL MBP-VL in PBS either for 16 h at 4 °C or at 25 °C for 1 h and blocked with 30% IPBS at 25 °C for 2 h. After the microplate was washed with PBST, each well was added with the mixture of 25 µL of Norit-treated serum extract, 12.5 µL of 5% IPBS containing 11-DC at various concentrations, and 12.5 µL of 0.5 µg/mL HRP-labeled MBP-VH in 5% IPBS at 25 °C for 1 h. Subsequently, the plate was washed with PBST six times and 50 µL of ECL was added plus the Western blotting detection system (GE Healthcare, Buckinghamshire, U.K.). After approximately 8 min of the reaction, the luminescence was measured for 2 s. To detect intrinsic 11-DC level, 10× concentrated serum extract without Norit treatment was used as above without adding 11-DC. RESULTS Cloning of Anti-11-DC Antibody Variable Region cDNA. To perform OS-ELISA for 11-DC, we cloned the anti-11-DC antibody variable region genes using a phage display system. Since the isotype of the antibodies was IgG1 with κ light chain, the VH and Vκ cDNAs were prepared from the hybridoma S.CET.M8.1.1, which produces an anti-11-DC antibody (CETM8) with a high affinity (Ka ) 2 × 1010 M-1) and very low crossreactivities with cortisol and cortisone (0.2 and 0.3%, respectively).23 With the use of the split-Fv phage display system12 that exploits M13 protein 9 and protein 7 for the display of VH and VL fragments, respectively, the cloning of the correct VH/ VL genes were performed (Figure 2). With the use of an amber suppressor strain TG-1 as a host to produce the Fv-displaying phages, a phage clone showing strong binding signal to 11-DCBSA immobilized plate but negligible signal to BSA was obtained. To validate the cloned variable region fragments, indirect competitive phage ELISA was performed. With the use of the phage-displayed Fv fragment, competition between 11-DC in sample and immobilized 11-DC-BSA was probed with HRP-labeled anti-M13 antibody. As a comparison, binding to immobilized BSA was tested. As a result, a working range within 2 orders of magnitude (∼10-500 ng/mL), similar to that reported previously for the radioimmunoassay (RIA) using scFv16 was obtained (Figure 3A). The nucleotide sequences of the cloned V regions showed an exact match to the reported scFv sequence.16 Open Sandwich Phage ELISA Using the Wild-Type and Mutant VH-Displaying Phages. By taking advantage of the splitFv system that allows a facile switch of the display/secretion of (23) Hosoda, H.; Kobayashi, N.; Tamura, S.; Mitsuma, M.; Sawada, J.; Terao, T.; Nambara, T. Chem. Pharm. Bull. (Tokyo) 1986, 34, 2914–2918.

Figure 2. Schematic structure of split Fv phage display vector pKST222 (A) and the produced phages derived of the compounds in parts B and C. The V-genes were amplified from total RNA by RTPCR and assembled into split-Fv with split-Fv linker through overlap extension PCR, cloned into phagemid vector pKST2, and transformed into E. coli TG1 for the selection and competitive phage ELISA (B) and HB2151 strain for OS-phage ELISA (C).

the myc-tagged VL fragment by changing the host phenotype, open sandwich ELISA was performed using the phage culture supernatant prepared with a nonsuppressing HB2151 strain. To immobilize the secreted VL fragment in the media, antimyc antibody was coated on microplate wells. The amount of phage particles bound to the wells through VH/VL interaction in the presence of 11-DC in the sample was evaluated using HRPlabeled anti-M13 antibody. The result with the wild-type VHdisplaying phage (Figure 3B) showed a characteristic noncompetitive dose-response curve only in the presence of immobilized antimyc 9E10 antibody. However, at least in the first trial, a high background signal without added antigen and a modest antigendependent increase was also observed. To improve the doseresponse by lowering the background signal, site-directed mutagenesis of a heavy chain VH/VL interface residue H39 was attempted, since this residue (glutamine) was identified as one of most important residues that determine the VH/VL interaction strength through hydrogen bonding with the neighboring glutamine residue at L38,24,25 and the mutation at this residue was proven effective to lower the VH/VL interaction strength.26 The resultant competitive phage ELISA using the H39QK and H39QR mutant Fv-displaying phages, as well as the wild-type Fv phage, is shown in Figure 4A. In spite of some reduction in binding signal, both mutants showed similar 11-DC dose-dependent competition curves. Then we attempted OS phage ELISA using the H39QK mutant that showed higher antigen binding. As shown in Figure 4B, it showed considerably lower background signal and the dose-dependent increase in signal for more than 3 orders of magnitude. While the sensitivity attained was around 1 ng/ mL, another measurement with different volume of supernatant gave the detection limit of final 0.01-0.1 ng/mL (1-10 pg/assay) (Figure 4C), which is arguably superior than those of competitive RIA or ELISA using the same antibody.23,27 This assay-dependent difference in detection limit might reflect a different VH/VL concentration in the media, which affects the background and positive signal strengths. (24) Masuda, K.; Sakamoto, K.; Kojima, M.; Aburatani, T.; Ueda, T.; Ueda, H. FEBS J. 2006, 273, 2184–2194. (25) Essen, L. O.; Skerra, A. J. Mol. Biol. 1994, 238, 226–244. (26) Sasajima, Y.; Aburatani, T.; Sakamoto, K.; Ueda, H. Biotechnol. Prog. 2006, 22, 968–973. (27) Hosoda, H.; Tamura, S.; Tsukamoto, R.; Kobayashi, N.; Sawada, J.; Terao, T.; Nambara, T. Chem. Pharm. Bull. (Tokyo) 1987, 35, 1497–1502.

Analytical Chemistry, Vol. 81, No. 20, October 15, 2009

8301

Figure 3. Phage ELISA using split Fv vector encoding the wild-type Fv gene. (A) Competitive ELISA performed with Fv-displaying phage (2.5 µL) prepared with the TG1 strain. (B) OS-ELISA performed with HB2151 culture supernatant (20 µL) containing VH-displaying phage and soluble VL. Development was terminated at 20 min. The vertical bars indicate the SDs (n ) 2).

Figure 4. Phage ELISA using split Fv vectors encoding the mutant Fv gene. (A) Competitive ELISA performed with Fv-displaying phages (5 µL) prepared with the TG1 strain. Half the volume (2.5 µL) of phage was used for the H39QK mutant. (B) OS-ELISA performed with HB2151 culture supernatant (20 µL) containing VH-displaying phage and soluble VL. Development was terminated at 40 min. Blank wells without immobilized antibodies were taken as controls. (C) The same assay as in part B but with increased culture supernatant (50 µL). Development was terminated at 30 min. The vertical bars indicate the SDs (n ) 2).

Open Sandwich ELISA Using MBP-VL and VH-AP Proteins. To further confirm the results with phage OS-ELISA, OSELISA with purified proteins was performed. To immobilize VL on microplate wells, E. coli maltose binding protein (MBP) was chosen as a fusion partner. MBP has been widely used as a fusion tag for both cytoplasmic and periplasmic/secretory expression of various proteins.28 Fusion with MBP has been known to enhance the solubility of the protein and also enhance scFv folding in the reductive cytoplasm.29 We used MBP not only for the ease of VL expression but also for passive immobilization to avoid the possible interference of antitag antibody to the VH/VL interaction of CET-M8. Also, direct immobilization was expected to result in an overall reduction in assay time and steps. For the detection of bound VH, first (28) di Guan, C.; Li, P.; Riggs, P. D.; Inouye, H. Gene 1987, 67, 21–30. (29) Shaki-Loewenstein, S.; Zfania, R.; Hyland, S.; Wels, W. S.; Benhar, I. J. Immunol. Methods 2005, 303, 19–39.

8302

Analytical Chemistry, Vol. 81, No. 20, October 15, 2009

the fusion protein with the E. coli alkaline phosphatase mutant with higher Vmax than the wild-type enzyme (D101S)9 was employed. The resultant dose-response curve using purified immobilized MBP-VL and VH-AP, together with the detection by chemiluminescent substrate CDP star, is shown in Figure 5. As seen, a wide working range of more than 3 orders of magnitude was obtained. It is worth noting that although we used the wild-type VH gene to prepare the fusion protein, the background signal was not too high and the resultant sensitivity based on the signal above 3SD of the background signal was as low as the final 0.03 ng/mL (3 pg/assay). Selectivity Against Analogues. To confirm whether the specificity of original CET-M8 Ab was retained by the cloned Fv and also in the OS-ELISA using its derivatives, OS-ELISA with the purified proteins was also performed for the 11-DC analogues. Among the compounds with analogous structure, the selectivity against cortisol is considered most important because it can exist

Figure5.(A)Schematicrepresentationoftheassay.(B)Dose-response curves of OS-ELISA with MBP-VL and VH-AP for 11-DC and cortisol. The value in the parentheses (0.2) shows the cross-reactivity in percent obtained by competitive RIA using CET-M8 antibody. The vertical bars indicate the SDs (n ) 3).

in serum at a larger excess than 11-DC. When the dose-response curve for cortisol was drawn, a similar but significantly rightshifted curve was obtained (Figure 5). The concentration difference between cortisol and 11-DC that gave the same counts of ∼2350 and ∼4400 per 5 s were 300- and 1000-fold, respectively. Since these values correspond well to the previously determined cross-reactivity (0.2%) of this antibody by competitive RIA, we would say that the selectivity of OS-IA is preserved as well as that of conventional competitive IAs. In addition, similar doseresponse curves with reduced sensitivity were also observed for the all other analogues tested (Figure 6A). Stability of the Reagents. Isolated antibody variable domains and its derivatives are sometimes regarded not as stable as fullsized antibodies. To evaluate this possibility, VH-AP after storage at 4 °C for 6 months were compared for its performance in OS-ELISA with the same protein stored at -80 °C. In addition, the same batch of MBP-VL stored at 4 °C for 6 months was used to capture both proteins on the microplate wells. As shown in Figure 6B, we could see no major differences in the two dose-responses except a marginal difference at low (