Journal of Clinical Laboratory Analysis 4:208-212 (1990)
Novel and Sensitive Noncompetitive (Two-Site) Enzyme lmmunoassay for Haptens With Amino Groups Koichiro Tanaka, Takeyuki Kohno, Seiichi Hashida, and Eiji lshikawa Department of Biochemistty Medical College of Miyazaki, Kiyotake, Miyazaki, Japan -
~
A novel and sensitive noncompetitive (twosite)enzyme immunoassay for haptens with
amino ClrouDs is described. L-Thyroxine (T,) was used as a model hapten. T, was indirectly biotinylated with glutathione as spacer between T4and biotin molecules and trapped onto anti-(T,-bovine serum albumin) IgGcoated polystyrene balls. After washing the polystyrene balls to eliminate unreacted biotin and other biotinylated substances, biotinylated T4was eluted from the polystyrene balls with HCI and was reacted with anti-(T,-bovine serum albumin) Fab’-horseradishperoxidase conjugate. The complex formed was trapped onto streptavidin-coated polystyrene balls. Key words:
Thyroxine, peroxidase, biotinylation
INTRODUCTION A highly sensitive two-site enzyme immunoassay was described for a-human atrial natriuretic peptide (a-hANP), a 28-amino acid single-chain peptide with an intramolecular disulfide bridge, secreted from the atrium (1). Three different antibodies were used. One was specific for the C-terminus of the peptide (C-terminus antibody), the second was specific for the N-terminus (N-terminus antibody), and the last recognized the N-terminal half of the ring structure of the peptide (ring antibody). A ring antibody-coated or N-terminus antibodycoated polystyrene ball was incubated with a-hANP and subsequently with C-terminus antibody-peroxidase conjugate. The detection limit of a-hANP was 10-30 amol(30-90 fg)/assay or 0.6-2.3 ngiL using 40-50 p,l of plasma. It is now possible to measure, directly without extraction, the lowest level of a-hANP in plasma (about 5 ng/L) of healthy subjects, whereas extraction and concentration are essential for radioimmunoassay ( 2 ) . The distance between the two epitopes recognized by the C-terminus antibody and the ring antibody or the N-terminus antibody is equivalent to a single-chain peptide consisting of approximately 12-15 amino acids. The results strongly suggest that single-chain peptides consisting of more than 12- 15 amino acids can be measured at attomole levels by the above technique. However, it is obviously impossible to measure, by two-site immunoassays, smaller haptens that cannot be simultaneously bound by two different antibodies specific for the corresponding two different epitopes on the
0 1990 Wiley-Liss, Inc.
Peroxidase activity bound to the polystyrene balls was assayed by fluorimetry. The detection limit of T, was 78 fg (0.1 fmol)/tube, which was 50-fold lower-than the limit by competitiveenzyme immunoassay using the same antibody and T,-peroxidase conjugate. This noncompetitive enzyme immunoassay was applied to the measurement of total T4 in serum. Serum levels of total T4in 10 healthy subjects aged 25-40 yr were 94 * 13 (SD) Fg/L (range, 78-1 14). The principle of this noncompetitive enzyme immunoassay may allow it to measure other haptens with amino groups more sensitively than can competitive immunoassays.
hapten molecules. For the last 30 yr, such smaller haptens have been measured mainly by competitive radioimmunoassay and partly by competitive nonisotopic immunoassays. The detection limit of haptens, (e.g., a-hANP, arginine vasopressin, angiotensin-I, and L-thyroxine) by competitive immunoassays is at femtomole levels (1.7-6.4 fmol) (2-5). This article describes a novel and sensitive noncompetitive (two-site) enzyme immunoassay for haptens with amino groups that is more sensitive than competitive immunoassays using hapten-enzyme conjugates. L-Thyroxine (T4) was used as a model hapten.
MATERIALS AND METHODS Treatment of Bovine Serum Albumin With Charcoal Bovine serum albumin (100 mg, crystallized, Miles Laboratories, Elkhart, IN) in 1 ml of deionized water was mixed with 110 pl of 1 M HCI to adjust the pH to 2.5 and incubated with 2% (w/v) charcoal (Norit A, Nakarai Tesque, Kyoto, Japan) at 30°C for 10 min with continuous shaking. The mixture was centrifuged at 3,000 rpm for 5 min, and the supernatant was treated with charcoal in the same way. The
Received December 13, 1989; accepted January 2, 1990. Address reprint requests to E. Ishikawa, M.D., Ph.D., Department of Biochemistry, Medical College of Miyazaki, Kiyotake, Miyazaki 889- 16, Japan.
Noncompetitive Enzyme lmmunoassay for Haptens
209
supernatant was neutralized by addition of 1 M sodium phosphate buffer, pH 7.5. The amount of bovine serum albumin was calculated from the absorbance at 280 nm by taking the extinction coefficient to be 0.63 L/g . cm (6).
on a column (1 X 30 cm) of Ultrogel AcA 54 (LKB Produkter AB, Stockholm-Bromma, Sweden) using 0.1 M sodium phosphate buffer, pH 6.5. The average number of T4 molecules introduced per bovine serum albumin molecule was 2.6.
Buffers
Protein-Sepharose 4B
The regularly used buffers were (a) 0.1 M sodium phosphate buffer, pH 7.0, containing 0.1 M NaCl, 1 g / L bovine serum albumin (crystallized, Miles Laboratories) treated with charcoal, and 0.6 mM 8-anilino- 1-naphthalenesulfonic acid ammonium salt (7) (Nakarai Tesque) (buffer A), and (b) 10 mM sodium phosphate buffer, pH 7.0, containing 0.1 M NaCl and 1 g / L bovine serum albumin (crystallized, Miles Laboratories), treated with charcoal (buffer B).
T4-bovine serum albumin conjugate (0.4 mg) and nonspecific rabbit IgG (1 .O mg) were coupled to cyanogen bromide (CNBr)-activated Sepharose 4B (0.2 g, Pharmacia Fine Chemicals AB, Uppsala, Sweden) according to the instructions of Pharmacia.
Haptens and Antibodies L-Thyroxine (3,3’,5,5’-tetraiodo-~-thyronine; T4) and L-triiodothyronine (3,3’,5-triiodo-~-thyronine; T3) were obtained from Sigma Chemical Company, St. Louis, MO. Rabbit anti-T4-bovine serum albumin serum was obtained from Miles Laboratories, Napeville, IL. Goat antirabbit IgG serum was obtained from Medical and Biological Laboratories, Nagoya, Japan. IgG was prepared from serum by fractionation with Na2S04 followed by passage through a column of diethylaminoethyl cellulose (8). F(ab‘)z was prepared by digestion of IgG with pepsin, and Fab’ was prepared by reduction of F(ab’)* (8). The amount of IgG and its fragments was calculated from the absorbance at 280 nm (8).
T4-BovineSerum Albumin Conjugate Mercaptosuccinylated bovine serum albumin Thiol groups were introduced into bovine serum albumin (fraction V; Armour Pharmaceutical, Kankakee, IL) using S-acetylmercaptosuccinic anhydride (8). The average number of thiol groups introduced per bovine serum albumin molecule was 10 (8).
Maleimide-T4 An aliquot (0.19 ml) of 1 mM T4 in N,N-dimethylformamide was incubated with 19 p.1 of 5 mM N-succinimidyl6-maleimidohexanoate (Dojindo Laboratories, Kumamoto, Japan) in N,N-dimethylformamide and 0.36 ml of 0.1 M sodium phosphate buffer, pH 7.0, at 30°C for 30 min.
T4-bovineserum albumin conjugate The mercaptosuccinylated bovine serum albumin (0.57 mg) in 0.2 ml of 0.1 M sodium phosphate buffer, pH 6.0, containing 5 mM ethylenediaminetetraacetate (EDTA) was incubated with 0.55 ml of the maleimide-T4 solution at 30°C for 30 min. The reaction mixture was subjected to gel filtration
Affinity-Purified Anti-T,-Bovine Fab’-Peroxidase Conjugate
Serum Albumin
Anti-T,-bovine serum albumin Fab‘ was conjugated to horseiradish peroxidase (grade I , RZ = 3.0, Boehringer Mannheim GmbH, Mannheim, West Germany) using N succir1imidyl-6-maleimidohexanoate(Dojindo Laboratories) (9). The conjugate was affinity-purified by elution from a column of T,-bovine serum albumin-Sepharose 4B at pH 2.5 (10). ‘The amount of the conjugate was calculated from peroxidase activity (8).
Protein-Coated Polystyrene Balls Polystyrene balls (3.2-mm diameter, Immunochemical, Okayama, Japan) were coated with rabbit anti-T4-bovine serum albumin IgG (0.1 g/L), affinity-purified goat (antirabbit IgG) IgG (0.1 g/L), and biotinyl nonspecific rabbit IgG (0.1 g/L) by physical adsorption ( 1 1). Goat (antirabbit IgG) IgG was affinity-purified by elution from a column of nonspecific rabbit IgGSepharose 4B at pH 2.5 (10). Biotinyl nonspecific rabbit IgG was prepared by the reaction of maleimidenonspccific rabbit IgG with N-biotinyl-2-mercaptoethylamine ( 12). Streptavidin-coated polystyrene balls were prepared by incubation of biotinyl nonspecific rabbit IgG-coated polystyrene balls with streptavidin (0.1 g/L) (Bethesda Research Laboratories, Bethesda, MD) at 30°C for 4 hr. The protein-coated polystyrene balls were stored in 10 mM sodium phosphate buffer, pH 7.0, containing 0.1 M NaCl, 1 g i L bovine serum albumin (fraction V, Armour Pharmaceutical), and 1 g / L NaN, at 4°C.
Dilution of Serum Samples and T4 Serum samples from healthy subjects aged 25-40 yr were diluted with buffer A 10,000-fold. T4 (2.0 mg) was dissolved in 0.5 ml of 0.1 M NaOH and diluted with buffer A. The amount of T4 was calculated by taking the molecular weight as 771.
Biotinylation Biotinylation was performed in two different ways.
210
Tanaka et al.
Direct biotinylation
T4-PeroxidaseConjugate
A 100-p1 aliquot of the diluted T4 or the diluted serum was incubated with 10 (1.1 of 66 mM N-hydroxysuccinimidobiotin (Zymed Laboratories, San Francisco, CA) in dimethylsulfoxide at 20°C for 1 hr. After incubation, the reaction mixture was incubated with 10 pl of 1M glycine-NaOH, pH 7.0, at 20°C for 30 min, followed by addition of 30 pl of buffer A containing 5 g/L NaN3 (biotinylated mixture).
MercaptosuccinylatedT4 An aliquot (480 pl) of 5.2 mM T4in 0.1 M sodium phosphate buffer, pH 7.5, containing 90% (viv) dimethylsulfoxide was incubated with 20 p1 of 25 mM N-succinimidyl-S-acetylthioacetate (Boehringer Mannheim) in N,N-dimethylformamide at 30°C for 1 hr. To the reaction mixture were successively added 160 pl of 125 mM glycine-NaOH, pH 7.0, 15 p1 of 100 mM EDTA, and 40 p1 of 1 M hydroxylamine.
Indirect biotinylation A 100-pl aliquot of the diluted T4 or the diluted serum was incubated with 5 p1 of 63 mM N-succinimidyl-6maleimidohexanoate (Dojindo) in dimethylsulfoxide at 20°C for 1 hr. After incubation, the reaction mixture was incubated with 5 pl of 99 mM glutathione (reduced form, Dojindo) in 0.1 M sodium phosphate buffer, pH 7.0, containing 1 mM EDTA at 20°C for 1 hr; it was subsequently reincubated with 5 pl of 138 mM N-hydroxysuccinimidobiotin (Zymed) in dimethylsulfoxide at 20°C for 1 hr. Finally, 5 pl of 2 M glycine-NaOH, pH 7.0, was added, and the incubation was continued at 20°C for 30 min, followed by addition of 30 pl of buffer A containing 5 g/L NaN3 (biotinylated mixture).
Noncompetitive Enzyme lmmunoassay for T4 The biotinylated mixture (150 pl) was incubated with an anti-T4-bovine serum albumin IgG-coated polystyrene ball at 4°C overnight. After incubation, the polystyrene ball was washed twice by addition and aspiration of 2 ml of 10 mM sodium phosphate buffer, pH 7.0, containing 0.1 M NaCl and incubated with the mixture of 100 pl of buffer B containing 0.72 mM 8-anilino- 1-naphthalenesulfonic acid ammonium salt and 20 p1 of 1 M HCl at 4°C for 1 hr. After removal of the polystyrene ball, the remaining solution was neutralized by addition of the mixture of 10 pl of 1 M sodium phosphate buffer, pH 7.0, and 20 pl of 1 M NaOH. The neutralized mixture was incubated with affinity-purified anti-T4- bovine serum albumin Fab'-peroxidase conjugate (100 fmol) and nonspecific rabbit F(ab'), (0.1 mg) in 20 pl of buffer B at 4°C overnight. Subsequently, two streptavidin-coated polystyrene balls were added, and the incubation was continued at 4°C for 5 hr. After removal of the reaction mixture, the polystyrene balls were washed twice as described above, and peroxidase activity bound to the polystyrene balls was assayed at 30°C for 60 min using 3-(4-hydroxyphenyl)propionic acid as substrate (13). Fluorescence intensity was measured relative to 1 mg/L quinine in 50 mM H2S04 using 320 nm for excitation and 405 nm for emission with a Shimadzu fluorophotometer (RF-5 10, Shimadzu Seisakusho, Kyoto, Japan).
Maleimide-peroxidase Maleimide groups were introduced into horseradish peroxidase (grade I, RZ = 3.0, Boehringer Mannheim) using N-succinimidyl-6-maleimidohexanoate (Dojindo) (8). The average number of maleimide groups introduced per peroxidase molecule was 1.4.
T4-peroxidaseconjugate The mercaptosuccinylated T4 solution (650 p1) was incubated with the maleimide-peroxidase (1.9 mg) in 300 pl of 0.1 M sodium phosphate buffer, pH 6.0, at 4°C for 12 hr. The reaction mixture was subjected to gel filtration on a column (1.5 x 45 cm) of Ultrogel AcA 54 (LKB Produkter AB) using 0.1 M sodium phosphate buffer, pH 6.5. The amount of the conjugate was calculated from peroxidase activity (8).
Competitive Enzyme lmmunoassayfor T4 Rabbit anti-T4-bovine serum albumin IgG was diluted with buffer A to a concentration of 2 nM. T4-peroxidase conjugate was diluted with buffer A to a concentration of 0.6 nM. The diluted T4 (50 p1) was incubated with the diluted T4-peroxidase conjugate (50 pl) and the diluted rabbit anti-T4-bovine serum albumin IgG (50 pl) at 4°C for 8 hr. After incubation, an affinity-purified goat (antirabbit IgG) IgGcoated polystyrene ball was added to the reaction mixture, and the incubation was continued at 4°C overnight. The polystyrene ball was washed twice by addition and aspiration of 2 ml of 10 mM sodium phosphate buffer, pH 7.0, containing 0.1 M NaCl. Peroxidase activity bound to the polystyrene ball was assayed as described above.
Expression of the Detection Limit of T4 The detection limit of T4 by noncompetitive enzyme immunoassay was taken as the minimal amount of T4 that gave a bound peroxidase activity significantly in excess of that nonspecifically bound in the absence of T4 (background). The detection limit of T4 by competitive enzyme immunoassay was taken as the minimal amount of T4 that gave a bound peroxidase activity significantly below that in the absence of T4. The existence of a significant difference was confirmed
Noncompetitive Enzyme lmmunoassay for Haptens
by the t test (p < 0.001, n = 5 for noncompetitive immunoassay; p < 0.01, n = 4 for competitive immunoassay).
RESULTS AND DISCUSSION
Detection Limit of T4 To was biotinylated directly with N-hydroxysuccinimidobiotin and trapped onto anti-T4-bovine serum albumin IgG-coated polystyrene balls. The polystyrene balls were washed to eliminate unreacted biotin and other biotinylated substances. Biotinylated T4 was eluted from the polystyrene balls with HCl and was measured using anti-To-bovine serum albumin Fab'-peroxidase conjugate and streptavidincoated polystyrene balls. Bound peroxidase activity in the presence of 78 pg (100 fmol)/tube To was not significantly higher than that in the absence of T4. This suggested that biotin residues bound to To molecules through amino groups might have been fairly close to the epitopic sites recognized by anti-T4-bovine serum albumin Fab'-peroxidase conjugate. Therefore, the biotinylated T4 molecules, which had been bound to the conjugate molecules, might not have efficiently reacted with streptavidin-coated polystyrene balls due to steric hindrance. In order to test this possibility, T4 molecules were bound indirectly to biotin molecules with glutathione molecules as spacers between the two molecules (Fig. 1). Maleimide groups were introduced into To molecules using Nsuccinimidyl-6-maleimidohexanoate and subsequently reacted with thiol groups of glutathione molecules. Finally, amino groups of glutathione residues bound to T4 molecules were reacted with N-hydroxysuccinimidobiotin.The detection limit of To by the indirect biotinylation was 78 fg (0.1
21 1
fmol)/tube (Fig. 2). This limit was %-fold lower than that by cornpetitive enzyme immunoassay using the same antibody and T4-peroxidase conjugate (Fig. 2 ) . This result supports the hypothesis above and suggests that other haptens with amino groups could also be measured with high sensitivity in a similar manner. In the subsequent experiments, therefore, enzyme immunoassay was performed after the indirect biotinylation.
Specificity Bound peroxidase activity in the presence of 2 pg (3 fmol) of T3 was not significantly higher than that in the absence of T4 (nonspecifically bound peroxidase activity) and si,gnificantly lower than that in the presence of 78 fg (0.1 finol, the detection limit) of T4. The cross-reaction with T3 on a molar basis was 2%.
Recovery of T4 Added to Serum The recoveries of T4 added to serum samples were 96.1 k 6.9% (SD) (range, 86%-103%; n = 10) when T4 at two different levels (155 and 310 pg/L) was added to five serum samples (0.01 pl) containing 78-106 p g i L of To.
Assay Precision The assay precision was examined at 10 different T4 levels over the range of 78-412 p g / L prepared with and without addition of T4 to serum samples. The coefficient of withinassay variation was 4.9%-9.9% (n = 5).
0 @CH:,C03H
HS-CY-CH-NH-~-CH2CH2-CHNH:, 0 Glutathione U
1
0.1
N-hydroxysuccinimidobiotln
1
5 10
100
1O(
T4 (fmolltube) 0
0
y y
S-CHiCH NH-F CH2-CHz-CH NH-C (CH )
2 4
0
Fig. 1. Indirect biotinylation of Tq.
NH
Fig. 2. Standard curves of T, by noncompetitive enzyme immunoassay after indirect biotinylation (open circles) and competitive enzyme immunoassay (closed circles). Each point is the mean of five (open circles) or four (closed (circles)determinations.
212
Tanaka et al.
Application The present enzyme immunoassay was applied to the measurement of total T4 in serum from healthy subjects. Serum levels of total T4 in 10 healthy subjects aged 25-40 yr were 94 ? 13 (SD) pg/L (range, 78-114). These results were consistent with the previous reports of competitive immunoassays (7,14). The present enzyme immunoassay is being applied to the measurement of other haptens with amino groups. The detection limits of arginine vasopressin and a-human atrial natriuretic peptide are 5-10 amol, and the volume of plasma or serum that can be used with satisfactory recoveries of haptens added to plasma or serum is 5- 10 pl. These results, which will be described in detail elsewhere, strongly suggest that the present enzyme immunoassay will be a powerful tool in future studies on peptides and other haptens.
REFERENCES Hashida S, lshikawa E, Mukoyama M, Nakao K, lmura H: Direct measurement of a-human atrial natriuretic polypeptide in plasma by sensitive enzyme immunoassay. JClin LabAna12:161-167, 1988. Sugawara A, Nakao K , Morii N, et al: a-Human atrial natriuretic polypeptide is released from the heart and circulates in the body. Biochem Biophys Res Commun 129:439-446, 1985. Morton JJ, Padfield PL, Forsling ML: A radioimmunoassay for plasma arginine-vasopressin in man and dog: Application to physiological and pathological states. J Endocrinol 65:411-424, 1975.
4. Fyhrquist F, Soveri P, Puutula L, Stcnman U-H: Radioimmunoassay of plasma renin activity. Clin Chem 22:250-256, 1976. 5. Maeda M, Ito K, Arakawa H, Tsuji A: An enzyme-linked immunosorbent assay for thyroxine in dried blood spotted on filter paper. J Immunol Methods 82:83-89, 1985. 6. Webster GC: Comparison of direct spectrophotometric methods for the measurement of protein concentration. Biochem Siophys Actu 207: 371-373, 1970. 7. Chopra IJ: A radioimmunoassay for measurement of thyroxine in unextracted serum. J Clin Endocrinol Metab 34:938-947, 1972. 8. Ishikawa E, Imagawa M , Hashida S, Yoshitake S , Hamaguchi Y, Ueno T: Enzyme-labeling of antibodies and their fragments for enzyme immunoassay and immunohistochemical staining. J Immunoassq 4: 209-327, 1983. 9. Hashida S, Imagawa M, Inoue S, Ruan K-h, Ishikawa E: More useful maleimide compounds for the conjugation of Fab‘ to horseradish peroxidase through thiol groups in the hinge. J Appl Biochem 6: 56-63, 1984. 10. Ruan K-h, Hashida S , Yoshitake S, et al: A micro-scale affinitypurification of Fab’-horseradish peroxidase conjugates and its use for sandwich enzyme immunoassay of insulin in human serum. Clin Chim Acta 147:167-172, 1985. 1 I . Ishikawa E, Kato K: Ultrasensitive enzyme immunoassay. Scand J Immunol 8(suppl7):43-55, 1978. 12. Kohno T, lshikawa E, S u g i y a n ~S, Nakamura S: Novel enzyme immunoassay ofanti-insulin IgG in human serum. JClinLabAna/2:19-24, 1988. 13. Imagawa M, Hashida S , Ishikawa E. et al: A highly sensitive sandwich enzyme immunoassay for insulin in human serum developed using capybara anti-insulin Fab’-horseradish peroxidase conjugate. Anal Lett 16(B 19):1509- 1523, 1983. 14. Schall RF, Jr, Fraser AS, Hansen HW, Kern CW, Tenoso HJ: A sensitive manual enzyme immunoassay for thyroxine. Clin Chem 24: 1801-1804. 1978.
Novel and Sensitive Noncompetitive (Two-Site) Enzyme lmmunoassay for Haptens With Amino Groups Koichiro Tanaka, Takeyuki Kohno, Seiichi Hashida, and Eiji lshikawa Department of Biochemistty Medical College of Miyazaki, Kiyotake, Miyazaki, Japan -
~
A novel and sensitive noncompetitive (twosite)enzyme immunoassay for haptens with
amino ClrouDs is described. L-Thyroxine (T,) was used as a model hapten. T, was indirectly biotinylated with glutathione as spacer between T4and biotin molecules and trapped onto anti-(T,-bovine serum albumin) IgGcoated polystyrene balls. After washing the polystyrene balls to eliminate unreacted biotin and other biotinylated substances, biotinylated T4was eluted from the polystyrene balls with HCI and was reacted with anti-(T,-bovine serum albumin) Fab’-horseradishperoxidase conjugate. The complex formed was trapped onto streptavidin-coated polystyrene balls. Key words:
Thyroxine, peroxidase, biotinylation
INTRODUCTION A highly sensitive two-site enzyme immunoassay was described for a-human atrial natriuretic peptide (a-hANP), a 28-amino acid single-chain peptide with an intramolecular disulfide bridge, secreted from the atrium (1). Three different antibodies were used. One was specific for the C-terminus of the peptide (C-terminus antibody), the second was specific for the N-terminus (N-terminus antibody), and the last recognized the N-terminal half of the ring structure of the peptide (ring antibody). A ring antibody-coated or N-terminus antibodycoated polystyrene ball was incubated with a-hANP and subsequently with C-terminus antibody-peroxidase conjugate. The detection limit of a-hANP was 10-30 amol(30-90 fg)/assay or 0.6-2.3 ngiL using 40-50 p,l of plasma. It is now possible to measure, directly without extraction, the lowest level of a-hANP in plasma (about 5 ng/L) of healthy subjects, whereas extraction and concentration are essential for radioimmunoassay ( 2 ) . The distance between the two epitopes recognized by the C-terminus antibody and the ring antibody or the N-terminus antibody is equivalent to a single-chain peptide consisting of approximately 12-15 amino acids. The results strongly suggest that single-chain peptides consisting of more than 12- 15 amino acids can be measured at attomole levels by the above technique. However, it is obviously impossible to measure, by two-site immunoassays, smaller haptens that cannot be simultaneously bound by two different antibodies specific for the corresponding two different epitopes on the
0 1990 Wiley-Liss, Inc.
Peroxidase activity bound to the polystyrene balls was assayed by fluorimetry. The detection limit of T, was 78 fg (0.1 fmol)/tube, which was 50-fold lower-than the limit by competitiveenzyme immunoassay using the same antibody and T,-peroxidase conjugate. This noncompetitive enzyme immunoassay was applied to the measurement of total T4 in serum. Serum levels of total T4in 10 healthy subjects aged 25-40 yr were 94 * 13 (SD) Fg/L (range, 78-1 14). The principle of this noncompetitive enzyme immunoassay may allow it to measure other haptens with amino groups more sensitively than can competitive immunoassays.
hapten molecules. For the last 30 yr, such smaller haptens have been measured mainly by competitive radioimmunoassay and partly by competitive nonisotopic immunoassays. The detection limit of haptens, (e.g., a-hANP, arginine vasopressin, angiotensin-I, and L-thyroxine) by competitive immunoassays is at femtomole levels (1.7-6.4 fmol) (2-5). This article describes a novel and sensitive noncompetitive (two-site) enzyme immunoassay for haptens with amino groups that is more sensitive than competitive immunoassays using hapten-enzyme conjugates. L-Thyroxine (T4) was used as a model hapten.
MATERIALS AND METHODS Treatment of Bovine Serum Albumin With Charcoal Bovine serum albumin (100 mg, crystallized, Miles Laboratories, Elkhart, IN) in 1 ml of deionized water was mixed with 110 pl of 1 M HCI to adjust the pH to 2.5 and incubated with 2% (w/v) charcoal (Norit A, Nakarai Tesque, Kyoto, Japan) at 30°C for 10 min with continuous shaking. The mixture was centrifuged at 3,000 rpm for 5 min, and the supernatant was treated with charcoal in the same way. The
Received December 13, 1989; accepted January 2, 1990. Address reprint requests to E. Ishikawa, M.D., Ph.D., Department of Biochemistry, Medical College of Miyazaki, Kiyotake, Miyazaki 889- 16, Japan.
Noncompetitive Enzyme lmmunoassay for Haptens
209
supernatant was neutralized by addition of 1 M sodium phosphate buffer, pH 7.5. The amount of bovine serum albumin was calculated from the absorbance at 280 nm by taking the extinction coefficient to be 0.63 L/g . cm (6).
on a column (1 X 30 cm) of Ultrogel AcA 54 (LKB Produkter AB, Stockholm-Bromma, Sweden) using 0.1 M sodium phosphate buffer, pH 6.5. The average number of T4 molecules introduced per bovine serum albumin molecule was 2.6.
Buffers
Protein-Sepharose 4B
The regularly used buffers were (a) 0.1 M sodium phosphate buffer, pH 7.0, containing 0.1 M NaCl, 1 g / L bovine serum albumin (crystallized, Miles Laboratories) treated with charcoal, and 0.6 mM 8-anilino- 1-naphthalenesulfonic acid ammonium salt (7) (Nakarai Tesque) (buffer A), and (b) 10 mM sodium phosphate buffer, pH 7.0, containing 0.1 M NaCl and 1 g / L bovine serum albumin (crystallized, Miles Laboratories), treated with charcoal (buffer B).
T4-bovine serum albumin conjugate (0.4 mg) and nonspecific rabbit IgG (1 .O mg) were coupled to cyanogen bromide (CNBr)-activated Sepharose 4B (0.2 g, Pharmacia Fine Chemicals AB, Uppsala, Sweden) according to the instructions of Pharmacia.
Haptens and Antibodies L-Thyroxine (3,3’,5,5’-tetraiodo-~-thyronine; T4) and L-triiodothyronine (3,3’,5-triiodo-~-thyronine; T3) were obtained from Sigma Chemical Company, St. Louis, MO. Rabbit anti-T4-bovine serum albumin serum was obtained from Miles Laboratories, Napeville, IL. Goat antirabbit IgG serum was obtained from Medical and Biological Laboratories, Nagoya, Japan. IgG was prepared from serum by fractionation with Na2S04 followed by passage through a column of diethylaminoethyl cellulose (8). F(ab‘)z was prepared by digestion of IgG with pepsin, and Fab’ was prepared by reduction of F(ab’)* (8). The amount of IgG and its fragments was calculated from the absorbance at 280 nm (8).
T4-BovineSerum Albumin Conjugate Mercaptosuccinylated bovine serum albumin Thiol groups were introduced into bovine serum albumin (fraction V; Armour Pharmaceutical, Kankakee, IL) using S-acetylmercaptosuccinic anhydride (8). The average number of thiol groups introduced per bovine serum albumin molecule was 10 (8).
Maleimide-T4 An aliquot (0.19 ml) of 1 mM T4 in N,N-dimethylformamide was incubated with 19 p.1 of 5 mM N-succinimidyl6-maleimidohexanoate (Dojindo Laboratories, Kumamoto, Japan) in N,N-dimethylformamide and 0.36 ml of 0.1 M sodium phosphate buffer, pH 7.0, at 30°C for 30 min.
T4-bovineserum albumin conjugate The mercaptosuccinylated bovine serum albumin (0.57 mg) in 0.2 ml of 0.1 M sodium phosphate buffer, pH 6.0, containing 5 mM ethylenediaminetetraacetate (EDTA) was incubated with 0.55 ml of the maleimide-T4 solution at 30°C for 30 min. The reaction mixture was subjected to gel filtration
Affinity-Purified Anti-T,-Bovine Fab’-Peroxidase Conjugate
Serum Albumin
Anti-T,-bovine serum albumin Fab‘ was conjugated to horseiradish peroxidase (grade I , RZ = 3.0, Boehringer Mannheim GmbH, Mannheim, West Germany) using N succir1imidyl-6-maleimidohexanoate(Dojindo Laboratories) (9). The conjugate was affinity-purified by elution from a column of T,-bovine serum albumin-Sepharose 4B at pH 2.5 (10). ‘The amount of the conjugate was calculated from peroxidase activity (8).
Protein-Coated Polystyrene Balls Polystyrene balls (3.2-mm diameter, Immunochemical, Okayama, Japan) were coated with rabbit anti-T4-bovine serum albumin IgG (0.1 g/L), affinity-purified goat (antirabbit IgG) IgG (0.1 g/L), and biotinyl nonspecific rabbit IgG (0.1 g/L) by physical adsorption ( 1 1). Goat (antirabbit IgG) IgG was affinity-purified by elution from a column of nonspecific rabbit IgGSepharose 4B at pH 2.5 (10). Biotinyl nonspecific rabbit IgG was prepared by the reaction of maleimidenonspccific rabbit IgG with N-biotinyl-2-mercaptoethylamine ( 12). Streptavidin-coated polystyrene balls were prepared by incubation of biotinyl nonspecific rabbit IgG-coated polystyrene balls with streptavidin (0.1 g/L) (Bethesda Research Laboratories, Bethesda, MD) at 30°C for 4 hr. The protein-coated polystyrene balls were stored in 10 mM sodium phosphate buffer, pH 7.0, containing 0.1 M NaCl, 1 g i L bovine serum albumin (fraction V, Armour Pharmaceutical), and 1 g / L NaN, at 4°C.
Dilution of Serum Samples and T4 Serum samples from healthy subjects aged 25-40 yr were diluted with buffer A 10,000-fold. T4 (2.0 mg) was dissolved in 0.5 ml of 0.1 M NaOH and diluted with buffer A. The amount of T4 was calculated by taking the molecular weight as 771.
Biotinylation Biotinylation was performed in two different ways.
210
Tanaka et al.
Direct biotinylation
T4-PeroxidaseConjugate
A 100-p1 aliquot of the diluted T4 or the diluted serum was incubated with 10 (1.1 of 66 mM N-hydroxysuccinimidobiotin (Zymed Laboratories, San Francisco, CA) in dimethylsulfoxide at 20°C for 1 hr. After incubation, the reaction mixture was incubated with 10 pl of 1M glycine-NaOH, pH 7.0, at 20°C for 30 min, followed by addition of 30 pl of buffer A containing 5 g/L NaN3 (biotinylated mixture).
MercaptosuccinylatedT4 An aliquot (480 pl) of 5.2 mM T4in 0.1 M sodium phosphate buffer, pH 7.5, containing 90% (viv) dimethylsulfoxide was incubated with 20 p1 of 25 mM N-succinimidyl-S-acetylthioacetate (Boehringer Mannheim) in N,N-dimethylformamide at 30°C for 1 hr. To the reaction mixture were successively added 160 pl of 125 mM glycine-NaOH, pH 7.0, 15 p1 of 100 mM EDTA, and 40 p1 of 1 M hydroxylamine.
Indirect biotinylation A 100-pl aliquot of the diluted T4 or the diluted serum was incubated with 5 p1 of 63 mM N-succinimidyl-6maleimidohexanoate (Dojindo) in dimethylsulfoxide at 20°C for 1 hr. After incubation, the reaction mixture was incubated with 5 pl of 99 mM glutathione (reduced form, Dojindo) in 0.1 M sodium phosphate buffer, pH 7.0, containing 1 mM EDTA at 20°C for 1 hr; it was subsequently reincubated with 5 pl of 138 mM N-hydroxysuccinimidobiotin (Zymed) in dimethylsulfoxide at 20°C for 1 hr. Finally, 5 pl of 2 M glycine-NaOH, pH 7.0, was added, and the incubation was continued at 20°C for 30 min, followed by addition of 30 pl of buffer A containing 5 g/L NaN3 (biotinylated mixture).
Noncompetitive Enzyme lmmunoassay for T4 The biotinylated mixture (150 pl) was incubated with an anti-T4-bovine serum albumin IgG-coated polystyrene ball at 4°C overnight. After incubation, the polystyrene ball was washed twice by addition and aspiration of 2 ml of 10 mM sodium phosphate buffer, pH 7.0, containing 0.1 M NaCl and incubated with the mixture of 100 pl of buffer B containing 0.72 mM 8-anilino- 1-naphthalenesulfonic acid ammonium salt and 20 p1 of 1 M HCl at 4°C for 1 hr. After removal of the polystyrene ball, the remaining solution was neutralized by addition of the mixture of 10 pl of 1 M sodium phosphate buffer, pH 7.0, and 20 pl of 1 M NaOH. The neutralized mixture was incubated with affinity-purified anti-T4- bovine serum albumin Fab'-peroxidase conjugate (100 fmol) and nonspecific rabbit F(ab'), (0.1 mg) in 20 pl of buffer B at 4°C overnight. Subsequently, two streptavidin-coated polystyrene balls were added, and the incubation was continued at 4°C for 5 hr. After removal of the reaction mixture, the polystyrene balls were washed twice as described above, and peroxidase activity bound to the polystyrene balls was assayed at 30°C for 60 min using 3-(4-hydroxyphenyl)propionic acid as substrate (13). Fluorescence intensity was measured relative to 1 mg/L quinine in 50 mM H2S04 using 320 nm for excitation and 405 nm for emission with a Shimadzu fluorophotometer (RF-5 10, Shimadzu Seisakusho, Kyoto, Japan).
Maleimide-peroxidase Maleimide groups were introduced into horseradish peroxidase (grade I, RZ = 3.0, Boehringer Mannheim) using N-succinimidyl-6-maleimidohexanoate (Dojindo) (8). The average number of maleimide groups introduced per peroxidase molecule was 1.4.
T4-peroxidaseconjugate The mercaptosuccinylated T4 solution (650 p1) was incubated with the maleimide-peroxidase (1.9 mg) in 300 pl of 0.1 M sodium phosphate buffer, pH 6.0, at 4°C for 12 hr. The reaction mixture was subjected to gel filtration on a column (1.5 x 45 cm) of Ultrogel AcA 54 (LKB Produkter AB) using 0.1 M sodium phosphate buffer, pH 6.5. The amount of the conjugate was calculated from peroxidase activity (8).
Competitive Enzyme lmmunoassayfor T4 Rabbit anti-T4-bovine serum albumin IgG was diluted with buffer A to a concentration of 2 nM. T4-peroxidase conjugate was diluted with buffer A to a concentration of 0.6 nM. The diluted T4 (50 p1) was incubated with the diluted T4-peroxidase conjugate (50 pl) and the diluted rabbit anti-T4-bovine serum albumin IgG (50 pl) at 4°C for 8 hr. After incubation, an affinity-purified goat (antirabbit IgG) IgGcoated polystyrene ball was added to the reaction mixture, and the incubation was continued at 4°C overnight. The polystyrene ball was washed twice by addition and aspiration of 2 ml of 10 mM sodium phosphate buffer, pH 7.0, containing 0.1 M NaCl. Peroxidase activity bound to the polystyrene ball was assayed as described above.
Expression of the Detection Limit of T4 The detection limit of T4 by noncompetitive enzyme immunoassay was taken as the minimal amount of T4 that gave a bound peroxidase activity significantly in excess of that nonspecifically bound in the absence of T4 (background). The detection limit of T4 by competitive enzyme immunoassay was taken as the minimal amount of T4 that gave a bound peroxidase activity significantly below that in the absence of T4. The existence of a significant difference was confirmed
Noncompetitive Enzyme lmmunoassay for Haptens
by the t test (p < 0.001, n = 5 for noncompetitive immunoassay; p < 0.01, n = 4 for competitive immunoassay).
RESULTS AND DISCUSSION
Detection Limit of T4 To was biotinylated directly with N-hydroxysuccinimidobiotin and trapped onto anti-T4-bovine serum albumin IgG-coated polystyrene balls. The polystyrene balls were washed to eliminate unreacted biotin and other biotinylated substances. Biotinylated T4 was eluted from the polystyrene balls with HCl and was measured using anti-To-bovine serum albumin Fab'-peroxidase conjugate and streptavidincoated polystyrene balls. Bound peroxidase activity in the presence of 78 pg (100 fmol)/tube To was not significantly higher than that in the absence of T4. This suggested that biotin residues bound to To molecules through amino groups might have been fairly close to the epitopic sites recognized by anti-T4-bovine serum albumin Fab'-peroxidase conjugate. Therefore, the biotinylated T4 molecules, which had been bound to the conjugate molecules, might not have efficiently reacted with streptavidin-coated polystyrene balls due to steric hindrance. In order to test this possibility, T4 molecules were bound indirectly to biotin molecules with glutathione molecules as spacers between the two molecules (Fig. 1). Maleimide groups were introduced into To molecules using Nsuccinimidyl-6-maleimidohexanoate and subsequently reacted with thiol groups of glutathione molecules. Finally, amino groups of glutathione residues bound to T4 molecules were reacted with N-hydroxysuccinimidobiotin.The detection limit of To by the indirect biotinylation was 78 fg (0.1
21 1
fmol)/tube (Fig. 2). This limit was %-fold lower than that by cornpetitive enzyme immunoassay using the same antibody and T4-peroxidase conjugate (Fig. 2 ) . This result supports the hypothesis above and suggests that other haptens with amino groups could also be measured with high sensitivity in a similar manner. In the subsequent experiments, therefore, enzyme immunoassay was performed after the indirect biotinylation.
Specificity Bound peroxidase activity in the presence of 2 pg (3 fmol) of T3 was not significantly higher than that in the absence of T4 (nonspecifically bound peroxidase activity) and si,gnificantly lower than that in the presence of 78 fg (0.1 finol, the detection limit) of T4. The cross-reaction with T3 on a molar basis was 2%.
Recovery of T4 Added to Serum The recoveries of T4 added to serum samples were 96.1 k 6.9% (SD) (range, 86%-103%; n = 10) when T4 at two different levels (155 and 310 pg/L) was added to five serum samples (0.01 pl) containing 78-106 p g i L of To.
Assay Precision The assay precision was examined at 10 different T4 levels over the range of 78-412 p g / L prepared with and without addition of T4 to serum samples. The coefficient of withinassay variation was 4.9%-9.9% (n = 5).
0 @CH:,C03H
HS-CY-CH-NH-~-CH2CH2-CHNH:, 0 Glutathione U
1
0.1
N-hydroxysuccinimidobiotln
1
5 10
100
1O(
T4 (fmolltube) 0
0
y y
S-CHiCH NH-F CH2-CHz-CH NH-C (CH )
2 4
0
Fig. 1. Indirect biotinylation of Tq.
NH
Fig. 2. Standard curves of T, by noncompetitive enzyme immunoassay after indirect biotinylation (open circles) and competitive enzyme immunoassay (closed circles). Each point is the mean of five (open circles) or four (closed (circles)determinations.
212
Tanaka et al.
Application The present enzyme immunoassay was applied to the measurement of total T4 in serum from healthy subjects. Serum levels of total T4 in 10 healthy subjects aged 25-40 yr were 94 ? 13 (SD) pg/L (range, 78-114). These results were consistent with the previous reports of competitive immunoassays (7,14). The present enzyme immunoassay is being applied to the measurement of other haptens with amino groups. The detection limits of arginine vasopressin and a-human atrial natriuretic peptide are 5-10 amol, and the volume of plasma or serum that can be used with satisfactory recoveries of haptens added to plasma or serum is 5- 10 pl. These results, which will be described in detail elsewhere, strongly suggest that the present enzyme immunoassay will be a powerful tool in future studies on peptides and other haptens.
REFERENCES Hashida S, lshikawa E, Mukoyama M, Nakao K, lmura H: Direct measurement of a-human atrial natriuretic polypeptide in plasma by sensitive enzyme immunoassay. JClin LabAna12:161-167, 1988. Sugawara A, Nakao K , Morii N, et al: a-Human atrial natriuretic polypeptide is released from the heart and circulates in the body. Biochem Biophys Res Commun 129:439-446, 1985. Morton JJ, Padfield PL, Forsling ML: A radioimmunoassay for plasma arginine-vasopressin in man and dog: Application to physiological and pathological states. J Endocrinol 65:411-424, 1975.
4. Fyhrquist F, Soveri P, Puutula L, Stcnman U-H: Radioimmunoassay of plasma renin activity. Clin Chem 22:250-256, 1976. 5. Maeda M, Ito K, Arakawa H, Tsuji A: An enzyme-linked immunosorbent assay for thyroxine in dried blood spotted on filter paper. J Immunol Methods 82:83-89, 1985. 6. Webster GC: Comparison of direct spectrophotometric methods for the measurement of protein concentration. Biochem Siophys Actu 207: 371-373, 1970. 7. Chopra IJ: A radioimmunoassay for measurement of thyroxine in unextracted serum. J Clin Endocrinol Metab 34:938-947, 1972. 8. Ishikawa E, Imagawa M , Hashida S, Yoshitake S , Hamaguchi Y, Ueno T: Enzyme-labeling of antibodies and their fragments for enzyme immunoassay and immunohistochemical staining. J Immunoassq 4: 209-327, 1983. 9. Hashida S, Imagawa M, Inoue S, Ruan K-h, Ishikawa E: More useful maleimide compounds for the conjugation of Fab‘ to horseradish peroxidase through thiol groups in the hinge. J Appl Biochem 6: 56-63, 1984. 10. Ruan K-h, Hashida S , Yoshitake S, et al: A micro-scale affinitypurification of Fab’-horseradish peroxidase conjugates and its use for sandwich enzyme immunoassay of insulin in human serum. Clin Chim Acta 147:167-172, 1985. 1 I . Ishikawa E, Kato K: Ultrasensitive enzyme immunoassay. Scand J Immunol 8(suppl7):43-55, 1978. 12. Kohno T, lshikawa E, S u g i y a n ~S, Nakamura S: Novel enzyme immunoassay ofanti-insulin IgG in human serum. JClinLabAna/2:19-24, 1988. 13. Imagawa M, Hashida S , Ishikawa E. et al: A highly sensitive sandwich enzyme immunoassay for insulin in human serum developed using capybara anti-insulin Fab’-horseradish peroxidase conjugate. Anal Lett 16(B 19):1509- 1523, 1983. 14. Schall RF, Jr, Fraser AS, Hansen HW, Kern CW, Tenoso HJ: A sensitive manual enzyme immunoassay for thyroxine. Clin Chem 24: 1801-1804. 1978.
