Journal of Clinical Laboratory Analysis 6:201-208 (1 992)

Novel and Ultrasensitive Noncompetitive Enzyme lmmunoassay (Hetero-Two-Site Complex Transfer Enzyme Immunoassay) for

a-Human Atrial Natriuretic Peptide Seiichi Hashida and Eiji lshikawa

Department of Biochemistry, Medical College of Miyazaki, Kiyotake, Miyazaki, Japan

A novel and ultrasensitive noncompetitive enzyme immunoassay (hetero-two-site com- plex transfer enzyme immunoassay) for a- human atrial natriuretic peptide (a-hANP) in plasma is described. a-hANP was biotinylated using N-hydroxysuccinimidobiotin and trapped onto an anti-a-hANP [6-281 IgG-coated poly- styrene ball. After washing, biotinylated a- hANP was eluted from the polystyrene ball with HCI and was reacted with 2,4-dinitrophenyl- fluorescein-bovine serum albumin-disulfide- rabbit anti-a-hANP [6-281 IgG conjugate. The complex formed was trapped onto (anti-2,4- dinitrophenyl group) IgG-coated polystyrene balls and, after washing, reacted with avidin- p-D-galactosidase conjugate. The polysty- rene balls were washed, and the complex of the three components was eluted with EN-2,

4-dinitrophenyl-L-lysine and transferred to anti-fluorescein IgG-coated polystyrene balls. After washing, the complex was released from the polystyrene balls by reduction with 2-mercaptoethylamine and transferred to (anti-rabbit IgG) IgG-coated polystyrene balls. p-D-Galactosidase activity bound to the last polystyrene balls was assayed by fluorome- try. The detection limit of a-hANP [l-281 was 3 fg (1 amol)/tube. Interference by plasma proteins was eliminated by separation of peptides from proteins using a molecular sieve. The assay range of plasma a-hANP [l-281 was 0.04-120 ng/L, and plasma levels of hANP in healthy subjects (1 1-56 ng/L) were measured without concentra- tion. o 1992 Wiley-Liss, Inc.

Key words: p-D-galactosidase, 2,4-dinitrophenyl group, fluorescein, disulfide bond

INTRODUCTION

On the basis of pioneering studies on atrial granules by de Bold and his coworkers, a 28-amino acid polypeptide with potent diuretic and natriuretic activity was isolated from the rat atrium, and its amino acid sequence was determined (1). Later, three distinct polypeptides called a-, p-, and y-human atrial natriuretic polypeptides (a-hANP, P-hANP, and y-hANP) were identified in the human atrium (2,3). a-hANP is a 28-amino acid polypeptide with an intramolecular disulfide bridge. P-hANP is an antiparallel dimer of a-hANP with inter- molecular disulfide bridges. y-hANP is a 126-amino acid polypeptide carrying the a-hANP sequence at the C-terminus. The molecular form of plasma hANP is almost exclusively

a-hANP has been measured with high sensitivity by two- site (sandwich) enzyme immunoassay (5). A polystyrene ball coated with antibody IgG for N-terminal half of the ring struc- ture of a-hANP [ 1-28] was incubated with a-hANP and sub- sequently with enzyme-labeled antibody fragment Fab’ for the C-terminus of a-hANP [l-281. The detection limit of a-hANP [ 1-28] was 10 amol, which was 100-fold lower than that by competitive radioimmunoassay. The assay range of plasma a-hANP [l-281 was 0.6-600 ng/L using 50 ~1 of

0 1992 Wiiey-Liss, Inc.

a-hANP (4).

plasma. Plasma levels of hANP in healthy subjects were mea- sured without concentration. However, a disadvantage of this method is that two antibodies, which can simultaneously bind to a-hANP molecules, are required.

This paper describes a novel and ultrasensitive noncom- petitive enzyme immunoassay (hetero-two-site complex trans- fer enzyme immunoassay) for a-hANP, which uses only one antibody for the C-terminus of a-hANP [ 1-28].

MATERIALS AND METHODS

Buffers

The regularly used buffers were 10 mmol/L sodium phos- phate buffer, pH 7.0, containing 0.1 mol/L NaC1, 1 mmol/L MgC12, 1 g/L NaN3, and 1 g/L bovine serum albumin (crys- tallized, Miles Inc., Diagnostics Division, Kankakee, IL) (buffer A), 0.1 mol/L sodium phosphate buffer, pH 7.0, con-

Received December 16, 1991; accepted February 19, 1992.

Address reprint requests to Dr. E. Ishikawa, Department of Biochemistry, Medical College of Miyazaki, Kiyotake, Miyazaki 889-16, Japan.

202 Hashida and lshikawa

taining 0 .3 mol/L NaCI, 0.1 g/L bovine serum albumin (crystallized, Miles Inc.), and 1 mmol/L EDTA (buffer B), and 10 mmol/L sodium phosphate buffer, pH 7.0, containing 0.3 mol/L NaCl, 1 mmol/L MgC12, 1 g/L NaN3, and 1 g/L bovine serum albumin (crystallized, Miles Inc.) (buffer C).

Peptides

[7-281, and P-hANP were obtained from Peptide Institute, Inc., Osaka, Japan. a-hANP [6-281 was generously supplied from Dr. K. Kondo (Chemistry Research Laboratories, Research and Development Division, Takeda Chemical Industries, Ltd., Yodogawaku, Osaka, Japan).

a-hANP[1-28],a-hANP[4-28],a-hANP[5-28],a-hANP

Antibodies

Anti-a-hANP [6-281 serum, which was raised in New Zea- land white rabbits by subcutaneous injections of a-hANP [6-28]-bovine-thyroglobulin conjugate prepared using glutar- aldehyde, was supplied from Dr. K. Kondo (Takeda Chemi- cal Industries, Ltd.) (6).

Rabbit (anti-2,4-dinitrophenyl bovine serum albumin) serum was obtained from Shibayagi Co., Ltd., Shibukawa, Gumma, Japan. Rabbit anti-fluorescein IgG was obtained from Dakopatts ak, Glostrup, Denmark. Goat (anti-rabbit IgG) IgG 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 (7). The amount of IgG was calculated from the absorbance at 280 nm (7).

2,4-Dinitrophenyl-Bovine Serum Albumin

Mercaptoacetyl-bovine serum albumin Bovine serum albumin (12 mg, 180 nmol, crystallized,

Miles Inc.) in 0.45 ml of 0.1 mol/L sodium phosphate buffer, pH 7.0, was incubated with 50 p.1 of 40 mmoYL N-succinimidyl- S-acetylthioacetate (Boehringer Mannheim GmbH, Mann- heim, Germany) in N,N-dimethylformamide at 30°C for 30 min. The reaction mixture was further incubated with 50 p.1 of 1 mol/L Tris-HC1 buffer, pH 7.0, 30 ~1 of 0 .1 mol/L EDTA, pH 7.0, and 60 p.1 of 1 mol/L hydroxylamine, pH 7.0, at 30°C for 5 min, and was subjected to gel filtration on a column (1.0 X 30 cm) of Sephadex G-25 (Pharmacia LKB Biotechnology AB, Uppsala, Sweden) using 0.1 mol/L sodium phosphate buffer, pH 6.0, containing 5 mmoliL EDTA. The amount of bovine serum albumin was calculated from the absorbance at 280 nm by taking the extinction co- efficient and the molecular weight to be 0.63 g ~ '. 1 .cm- ' (8) and 66,200 (9), respectively. The average number of thiol groups introduced per bovine serum albumin molecule was 5.9 (7).

aN-Maleimidohexanoyl-~N-2,4-dinitrophenyl-L-lysine

An aliquot (0.9 ml) of 11.1 mmol/L ~N-2,4-dinitrophenyl- L-lysine-HC1 (Tokyo Kasei Kogyo, Co., Ltd., Tokyo, Japan) in 0.1 moliL sodium phosphate buffer, pH 7 .O, containing 5 mmol/L EDTA was incubated with 0.1 ml of 20 mmol/L N-succinimidyl-6-maleimidohexanoate (Dojindo Laboratories, Kumamoto, Japan) in N,N-dimethylformamide at 30°C for 30 min (7).

2,4-Dinitrophenyl bovine serum albumin

Mercaptoacetyl bovine serum albumin (1 1 mg, 170 nmol) in 0.9 ml of 0.1 mol/L sodium phosphate buffer, pH 6.0, containing 5 mmol/L EDTA was incubated with an aliquot ( 1.1 ml) of the aN-maleimidohexanoyl-~N-2,4-dinitrophenyl- L-lysine solution at 30°C for 30 min. The reaction mixture was subjected to gel filtration on a column (1 .O x 30 cm) of Sephadex (3-25 (Pharmacia LKB Biotechnology AB) using 0.1 mol/L sodium phosphate buffer, pH 7.0. The number of 2,4-dinitrophenyl groups was calculated from the absorbance at 360 nm by taking the molar extinction coefficient to be 17,400 mol-' .l .cm-' (lo), and the ratio of the extinction coefficient at 360 nm to that at 280 nm of 2,4-dinitrophenyl groups was 1:0.32. The average number of 2,4-dinitrophenyl groups introduced per bovine serum albumin molecule was 5.6, which was calculated from the absorbance at 280 nm and 360 nm (8-10).

2,4-Dinitrophenyl-FIuorescein-Bovine Serum Albumin-Disulfide-Rabbit Anti-a-hANP [6-281 IgG Conjugate

6-Maleimidohexanoyl-2,4-dinitrophenyl-bovine serum albumin

2,4-Dinitrophenyl-bovine serum albumin (10 mg, 150 nmol) in 0.45 ml of 0.1 mol/L sodium phosphate buffer, pH 7.0, was incubated with 50 p.1 of 30 mmol/L N-succinimidyl- 6-maleimidohexanoate (Dojindo Laboratories) in N,N- dimethylformamide at 30°C for 30 min. After incubation, the reaction mixture was subjected to gel filtration on a column (1 .O x 30 cm) of Sephadex (3-25 (Pharmacia LKB Biotech- nology AB) using 0.1 mol/L sodium phosphate buffer, pH 7 .O. The average number of maleimide groups introduced per bovine serum a1 bumin molecule was 5.8 (7).

N-Fluorescein-2-mercaptoethylamine

An aliquot (0.2 ml) of 100 mmol/L fluorescein isothio- cyanate (Sigma Chemical Company, St. Louis) in N,N- dimethylsulfoxide was incubated with 1.8 ml of 1 .O mmol/L 2-mercaptoethylamine (Nacalai Tesque, Inc., Kyoto, Japan) in 0.5 mol/L sodium carbonate buffer, pH 9.5, at room tem- perature for 1 h. After incubation, 0.2 ml of 2 mol/L glycine- NaOH buffer, pH 8.0, was added to the reaction mixture to

Enzyme lmmunoassay of a-hANP 203

2,4-Dinitrophenyl-fluorescein-bovine serum albumin-disulfide-rabbit anti-a-hANP [6-281 IgG conjugate

Mercaptoacetyl anti-a-hANP [6-281 IgG (0.7 mg, 5 nmol) in 0.2 ml of 0.1 mol/L sodium phosphate buffer, pH 6.0, containing 5 mmol/L EDTA was incubated with 2-pyridyl- disulfide-2,4-dinitrophenyl-fluorescein-bovine serum albumin (1.7 mg, 25 nmol) in 0.2 ml of 0.1 mol/L sodium phosphate buffer, pH 7.5, at 4°C for 20 h. After incubation, the mixture was subjected to gel filtration on a column (1.5 X 60 cm) of Ultrogel AcA 34 (IBF biotechnics, Villeneuve-la-Garenne, France) using 0.1 mol/L sodium phosphate buffer, pH 7.0. The molar ratio of bovine serum albumin to anti-a-hANP [6-281 IgG in the conjugate was 1.8, which was calculated from the absorbance at 280 nm and 360 nm (7-10).

eliminate remaining fluorescein isothiocyanate, and the incubation was continued at room temperature for 1 h. After incubation, the pH was adjusted to 7 . 0 with 1 mol/L HCl .

2,4-Dinitrophenyl-fluorescein-bovine serum albumin

6-Maleimidohexanoyl-2,4-dinitrophenyl-bovine serum albu- min (9.6 mg, 145 nmol as bovine serum albumin) in 0.4 ml of 0.1 mol/L sodium phosphate buffer, pH 7.0, was incu- bated with an aliquot (1.6 ml, 1.6 pmol) of the N-fluorescein- 2-mercaptoethylamine solution at 30°C for 30 min. After incubation, the mixture was subjected to gel filtration on a column (1 .O x 45 cm) of Sephadex G-25 (Pharmacia LKB Biotechnology AB) using 0.1 mol/L sodium phosphate buffer, pH 7.0. The number of fluorescein residues was calculated from the absorbance at 495 nm by taking the molar extinction coefficient to be 68,000 molpl . l .cm- ' (11), and the ratio of the extinction coefficient at 495 nm to that at 280 nm of fluorescein residues was 1:0.30. The average number of flu- orescein residues introduced per bovine serum albumin mol- ecule was 5.6, which were calculated from the absorbance at 280 nm, 360 nm, and 495 nm (8- 11).

2-Pyridyl disulfide-2,4-dinitrophenyl-fluorescein- bovine serum albumin

2,4-Dinitrophenyl-fluorescein-bovine serum albumin (8.9 mg, 135 nmol) in 0.45 ml of 0.1 mol/L sodium phosphate buffer, pH 7.0, was incubated with 50 pl of 30 mmol/L N-succinimidyl-3-(2-pyridyl-dithio)propionate (Pharmacia LKB Biotechnology AB) in ethanol at 30°C for 30 min. After incu- bation, the reaction mixture was subjected to gel filtration on a column (1 .O X 30 cm) of Sephadex G-25 (Pharmacia LKB Biotechnology AB) using 0.1 mol/L sodium phosphate buffer, pH 7.5. The average number of 2-pyridyl disulfide residues introduced per bovine serum albumin molecule was 8.2 (7).

Mercaptoacetyl anti-a-hANP [6-281 IgG

Anti-a-hANP [6-281 IgG (1 .O mg, 7 nmol) in 0.45 ml of 0.1 mol/L sodium phosphate buffer, pH 7.0, was incubated with 50 I J . ~ of 1 mmol/L N-succinimidyl-S-acetyl-thioacetate (Boehringer Mannheim GmbH) in N,N-dimethylformamide at 30°C for 30 min. After incubation, the mixture was incu- bated with 50 p.1 of 1 mol/L Tris-HC1, pH 7.0, 30 pl of 0.1 mol/L EDTA, pH 7.0, and 60 pl of 1 mol/L hydroxylamine, pH 7.0, at 30°C for 5 min. The reaction mixture was sub- jected to gel filtration on a column (1 .O X 30 cm) of Sephadex G-25 (Pharmacia LKB Biotechnology AB) using 0. I mol/L sodium phosphate buffer, pH 6.0, containing 5 mmoVL EDTA. The average number of thiol groups introduced per IgG mol- ecule was 1.4 (7).

Protein-Sepharose 4B

2,4-Dinitrophenyl-bovine serum albumin (10 mg) and non- specific rabbit IgG (10 mg) were coupled to CNBr-activated Sepharose 4B (1 g, Pharmacia LKB Biotechnology AB) according to the instruction of Pharmacia.

Affinity-Purification of Antibodies

(Anti-2,4-dinitrophenyl bovine serum albumin) IgG and (anti-rabbit IgG) IgG were affinity-purified by elution from columns of 2,4-dinitmphenyl-bovine serum albumin-Sepharose 4B and nonspecific rabbit IgG-Sepharose 4B at pH 2.5, respectively (12).

Protein-Coated Polystyrene Balls

Polystyrene balls (3.2 mm in diameter, Immuno Chemical Inc., Okayama, Japan) were coated by physical adsorption with 0.1 g/L of anti-a-hANP [6-281 IgG, affinity-purified (anti- 2,4-dinitrophenyl-bovine serum albumin) IgG, anti-fluorescein IgG, and affinity-purified (anti-rabbit IgG) IgG (1 3). Protein- coated polystyrene balls were stored in buffer A at 4°C. Polystyrene balls coated with affinity-purified (anti-2,4- dinitrophenyl-bovine serum albumin) IgG and affinity-purified (anti-rabbit IgG) IgG had been colored pink for discrimina- tion from other polystyrene balls.

Avidin-p-D-Galactosidase Conjugate

6-Maleimidohexanoyl-avidin Avidin (1.4 mg or 20 nmol, Avidin D, Vector Laboratories,

Inc., Burlingame, CA) was dissolved in 0.48 ml of 0.1 mol/L sodium phosphate buffer, pH 7.0, and incubated with 20 I J . ~

of 5 mmol/L N-succinimidyl-6-maleimidohexanoate (Dojindo Laboratories) in N,N-dimethylformamide at 30°C for 30 min. The reaction mixture was subjected to gel filtration on a col- umn (1 .O x 45 cm) of Sephadex G-25 (Pharmacia LKB Bio- technology AB) using 0.1 mol/L sodium phosphate buffer,

204 Hashida and lshikawa

pH 6.0. The amount of avidin was calculated from the ab- sorbance at 280 nm by taking the extinction coefficient and the molecular weight to be 1.4 g - ' . l *cm- ' and 68,000, respectively (14). The average number of maleimide groups introduced per avidin molecule was 1.5 (7).

Avidin-P-D-galactosidase conjugate

6-Maleimidohexanoyl-avidin (about 0.4 mg, 6 nmol) in 1 .O ml of 0.1 mol/L sodium phosphate buffer, pH 6.0, was incubated with P-D-galactosidase (Escherichiu coli, Boehringer Mannheim GmbH) (2.0 mg, 4 nmol) in 1 .O ml of 0.1 mol/L sodium phosphate buffer, pH 6.0, containing 5 mmoYL EDTA at 4°C for 20 h. After incubation, the reaction mixture was subjected to gel filtration on a column (1.5 x 60 cm) of Sepharose 6B (Pharmacia LKB Biotechnology AB) using 10 mmol/L sodium phosphate buffer, pH 6.5 containing 0.1 mol/L NaCl, 1 mmoVL MgC12, 1 g/L NaN3, and 1 g/L bovine serum albumin (crystallized, Miles Inc.). Absorbance at 280 nm and P-D-galactosidase activity of each fraction were measured. An aliquot of each fraction was diluted 2,000-fold with buffer A, and 10 pl of the diluted sample was subjected to P-D- galactosidase assay at 30°C for 10 min using 4-methylumbellif- eryl P-D-galactoside as substrate (7). Fluorescence intensity was measured relative to 10- mol/L 4-methylumbelliferone in 0.1 mol/L glycine-NaOH buffer, pH 10.3, using a Shimadzu spectrofluorophotometer (RF-5 10, Shimadzu Corporation, Kyoto, Japan). Fractions containing the conjugate were stored at 4°C.

a-hANP Standards, Plasma, and Plasma Filtrates

a-hANP [ 1-28] (0.59 mg, Peptide Institute, Inc.) was dis- solved in the mixture of 1.1 ml of 0.1 moliL HCl and 1.1 ml of acetonitrile, and diluted with buffer B. The amount of a-hANP was calculated by taking the molecular weight as 3,081 ( 2 ) .

Blood was collected from the antecubital vein of healthy male subjects aged 23-41 yr with ad libitum water intake and normal activity approximately 4 h after breakfast. Blood with- drawn into chilled glass tubes containing EDTA (Venoject VT-070NA EDTA-2Na, Terumo Corp., Tokyo, Japan) was centrifuged at 1,500 xg for 20 min to separate plasma. The final concentration of EDTA in blood was 1.5 g/L.

Plasma (0.3 ml) was mixed with 0.1 ml of 0.4 mol/L sodium phosphate buffer, pH 7.0 and filtered by centrifuga- tion in a microconcentrator with polysaccharide membrane (CENTRICON-100, Amicon Division, W.R. Grace & Co., Beverly, MA) at 4°C at 5,000 xg for 30 min. The filtrates (0.15 ml) were mixed with 1/99 volume of 10 g/L bovine serum albumin (crystallized, Miles Inc.).

Measurement of Amino Groups

A 10 pl aliquot of samples was mixed with 2.0 ml of 0.2 mol/L sodium borate buffer, pH 8.0, and reacted with 0.8 ml

of 0.1 g/L fluorescamine (Sigma Chemical Company) in ace- tone. The fluorescence intensity of the reaction mixture was measured using 390 nm for excitation and 475 nm for emis- sion (15). L-Leucine was used as standard.

Biotinylation

A 100 pl aliquot of the diluted a-hANP or plasma filtrates was incubated with 5 pl of 63 mmol/L N-hydroxysuccinimi- dobiotin (Zymed Laboratories, Inc., San Francisco, CA) in N,N-dimethylformamide at 4°C for 1 h. The reaction mix- ture was further incubated with 5 pl of 2 moVL glycine-NaOH, pH 7.0, at 4°C for 1 h, followed by addition of 40 p1 of buffer B containing 2.7 mmol/L 8-anilino- l-naphthalenesulfonic acid ammonium salt (Nacalai Tesque, Inc.) and 5 g/L NaN3.

Hetero-Two-Site Complex Transfer Enzyme lmmunoassay

The biotinylated mixture (0.15 ml) was incubated with an anti-a-hANP [6-281 IgG-coated polystyrene ball at 4°C over- night. After removal of the biotinylated mixture, the polysty- rene ball was incubated with 2 ml of buffer A at 30°C for 10 min, washed by addition and aspiration of 2 ml of buffer A, and incubated with 75 pl of buffer C and 15 pl of 1 moliL HCl at 4°C for 1 h. After removal of the polystyrene ball, the remaining solution was neutralized by addition of 15 pl of 1 mol/L sodium phosphate buffer, pH 7.0, and 15 p1 of 1 mol/L NaOH. The neutralized mixture was incubated with 100 fmol of 2,4-dinimphenyl-fluorescein-bovine serum albumin-disulfide- rabbit anti-a-hANP [6-281 IgG conjugate in 30 pl of buffer C at 20°C for 4 h. Subsequently, two affinity-purified (anti-2,4- dinitrophenyl-bovine serum albumin) IgG-coated polystyrene balls, which had been colored pink, were added, and the in- cubation was continued at 4°C overnight. After removal of the incubation mixture, the polystyrene balls were washed twice by addition and aspiration of 2 ml of buffer A and incubated with 250fmolofavidin-P-D-galactosidaseconjugate inO. 15 ml of buffer C at 20°C overnight. After removal of the incubation mixture, the polystyrene balls were washed twice as described above, and incubated with 0.15 ml of buffer C containing I mmoVL ~N-2,4-dinitrophenyl-Llysine and two anti-fluorescein IgG-coated polystyrene balls with no color at 20°C for 3 h. After removal of the incubation mixture, the polystyrene balls with no color were washed twice as described above and in- cubated with 0.15 ml of buffer C containing 10 mmol/L 2- mercaptoethylamine and two afflnity-purified (anti-rabbit IgG) IgG-coated polystyrene balls, which had been colored pink, at 20°C for 3 h. After removal of the incubation mixture, the polystyrene balls colored pink were washed twice as described above. Bound P-D-galactosidase activity was assayed by fluoro- metry at 30°C for 20 h as described above. In some experiments, P-D-galactosidase activity bound to (anti-2,4-dinitrophenyl group) IgG-and anti-fluorescein IgG-coated polystyrene balls before elution of the complex was also assayed for 1 h.

Enzyme lmmunoassay of a-hANP 205

Immobilized Biotinylated anti-hANP IgG hANP

DNP-FL-BSA-SS- Rabbit anti-hANP IgG

Immobilized anti-DNP IgG

/ A

Avidin-Gal 1% 0 DNP-L-lysine

Mercaptoethylamine

Immobilized L (anti-rabbit IgG) IgG

HS

Fig. 1. orescein; BSA, bovine serum albumin; SS, disulfide bond; Gal, P-D-galactosidase,

Hetero-two-site complex transfer enzyme immunoassay for a-hANP. DNP, dinitrophenyl group; FL, flu-

Expression of the Detection Limit of a-hANP [l-281 The detection limit of a-hANP [1-281 was taken as the

minimal amount of a-hANP [ 1-28] which gave a bound P-D- galactosidase activity significantly in excess of that nonspe- cifically bound in the absence of a-hANP [ 1-28] (background). The existence of a significant difference from the background was confirmed by the t-test (P < .001, n = 5) .

RESULTS

a-hANP [ 1-28], a 28 amino acid peptide with an intramo- lecular disulfide bridge, was biotinylated and measured by a noncompetitive enzyme immunoassay (hetero-two-site com- plex transfer enzyme immunoassay) as schematically shown in Figure I .

Detection Limit of a-hANP [l-281 The detection limit of a-hANP [ 1-28] was 3 fg (1 amol)/tube

(Fig. 2). This was 10-fold lower than that by a fluorometric two-site enzyme immunoassay (5) and 80-fold lower than that by a colorimetric two-site enzyme immunoassay (6).

Specificity The specificity was examined using a-hANP-related

peptides (Fig. 3). The reactivity significantly decreased by

deletion of N-terminal and C-terminal amino acids. The cross- reactions with a-hANP [4-281, a-hANP [5-281, a-hANP [6-281, a-hANP [7-281, and a-hANP [5-271 on a molar basis were 13.9%, 29.4%, 55.8%, 1.0%, and 0.6%, respectively. The low cross-reaction with a-hANP [5-271 indicated that the anti-a-hANP [6-281 antibody used recognized the C-terminus of a-hANP [ 1-28]. The low cross-reaction with a-hANP [7-28) might have been due to an insufficient distance between bio- tin residues and the C-terminus of a-hANP [7-281 in biotinylated a-hANP [7-281 molecules for simultaneous binding of 2, 4-dinitrophenyl-fluorescein-bovine serum albumin-disulfide- rabbit anti-a-hANP [6-281 IgG conjugate and avidin-P-D- galactosidase conjugate. The cross-reaction with PhANP was 28.3% on a molar basis.

Plasma Interference

As the volume of plasma subjected to biotinylation in- creased, the recovery of a-hANP added to plasma decreased. This was due to the presence of amino groups at high con- centrations in plasma (approximately 60 mmol/L), while the concentration of N-hydroxysuccinimido-biotin added to the reaction mixture for biotinylation was 3 mmol/L. The maxi- mal volume of plasma that could be used with a satisfactory recovery was only 5 pJ.

206 Hashida and lshikawa

Measurement of ANP in Plasma

The assay range of a-hANP in plasma using 100 p1 of plasma filtrates corresponding to 75 p1 of plasma was 0.04- 120 ng/L, since 100 pl of plasma filtrates corresponding to 75 pl of plasma could be used without interference as described above. The concentrations of hANP in plasma of 12 healthy subjects aged 23-41 yr with ad libitum water intake and nor- mal activity approximately 4 h after breakfast were 30 & 11 (SD) ng/L (range, 11-56 ng/L), which were similar to those previously reported by two-site enzyme immunoassay (5).

P

L

P x a C Q) c

c .-

c - 8 8 8 ii

C Q)

Q)

3

I 1 I L

1 1 0 100 1000

a-hANP [ 1-28 ] ( arnol/ tube )

Fig. 2. Standard curves of whANP [l-281 by hetero-two-site complex transfer enzyme immunoassay. Open and solid circles indicate bound P-D- galactosidase activity (specifically bound P-D-galactosidase activity plus nonspecifically bound P-D-galactosidase activity) and specifically bound P-D-galactosidase activity, respectively. Vertical bars indicate standard devi- ations of 3-5 determinations.

In order to overcome this difficulty, plasma was diluted 1.3-fold and filtered by centrifugation in a microconcentrator with polysaccharide membrane to separate peptides from plasma proteins. The concentrations of amino groups in the plasma filtrates of healthy subjects aged 28-37 yr at 9:OO a.m. before breakfast (n = 4) and approximately 4 h after break- fast (n = 8) were both 1.7-2.1 mmollL. When 100 p1 of plasma filtrates corresponding to 75 p1 of plasma was sub- jected to biotinylation, the recoveries of a-hANP (0.25-25 ng/L) added to 6 plasma samples containing 0.17-21.7 ng/L of hANP were 8 1- 109%.

Assay Variation

The assay variation was examined at 3 different plasma levels over the range of 0.22-29 ng/L for within-assay and at 3 different plasma levels over the range of 0.36-27 ng/L for between-assay. The volume of plasma filtrates used was 100 p1 corresponding to 75 p1 of plasma, and the num- bers of determinations at each level were 10 for within-assay and 5 for between-assay. The variation coefficients of within- assay and between-assay were 6.0-1 1% and 2.5-9.79'0, respectively.

DISCUSSION

In the present method, the background, that is, the non- specific binding of avidin-P-D-galactosidase conjugate, was decreased by transfer from solid phase to solid phase of the complex consisting of 2,4-dinitrophenyl-fluorescein-bovine serum albumin-disulfide-rabbit anti-a-hANP [6-281 IgG con- jugate, biotinylated a-hANP, and avidin-P-D-galactosidase conjugate with less decrease in the specific binding. This improved the detection limit of a-hANP [ 1-28] to 1 amol as shown by the following calculations (Fig. 4). When the amount of a-hANP [l-281 subjected to biotinylation was 1 fmol, P-D-galactosidase activity specifically bound to (anti-2,4- dinitrophenyl group) IgG-coated polystyrene balls was equiv- alent to 0.070 fmol (7.0% of a-hANP [l-281 subjected to biotinylation on a molar basis), and that nonspecifically bound was 0.044% of avidin-P-D-galactosidase conjugate added. The detection limit of a-hANP [ 1-28] was approximately 0.3 fmol. In the next step, the complex was transferred to anti-fluorescein IgG-coated polystyrene balls. When the amount of a-hANP [ 1-28] subjected to biotinylation was 30 amol, P-D-galacto- sidase activity specifically bound to anti-fluorescein IgG- coated polystyrene balls was equivalent to 0.5 1 amol ( I .7% of a-hANP [ 1-28] subjected to biotinylation on a molar basis), and that nonspecifically bound was 0.00069% of avidin-P-D- galactosidase conjugate added. The detection limit of a-hANP [ 1-28] was improved 15-fold to approximately 20 amol. In the final step, the complex was transferred to (anti-rabbit IgG) IgG-coated polystyrene balls. When the amount of a-hANP [ 1-28] subjected to biotinylation was 3 amol, P-D-galactosidase activity specifically bound to (anti-rabbit IgG) IgG-coated polystyrene balls was equivalent to 0.030 amol (1.0% of a-hANP [ 1-28] subjected to biotinylation on a molar basis), and that nonspecifically bound was 0.000016% of avidin-P-D- galactosidase conjugate added. The detection limit was im- proved 20-fold to 1 amol.

The smaller decreases in the specific binding with the larger decreases in the nonspecific binding might have been due to the following facts. Not only the complex of the three com- ponents but also free 2,4-dinitrophenyl-fluorescein-bovine serum albumin-disulfide-rabbit anti-a-hANP (6-281 IgG con- jugate was transferred, and the affinity between biotinylated a-hANP [ 1-28] and avidin-P-D-galactosidase conjugate was

1 I0000

t

Enzyme lmmunoassay of a-hANP

P

a-hANP [I281

I I I

10 100 1000

hANP ( amol /tube )

a-hANP

P-hANP a-hANP

a-hANP

a-hANP

a-hANP

' [6-281

[5-281 [4-281

[7-281 [5-271

Fig. 3. Dose-response curves of a-hANP f 1-28] and related peptides by hetero-two-site complex transfer enzyme immunoassay

extremely high (16). Both factors might have been effective in minimizing the dissociation of the complex consisting of the three components. However, P-D-galactosidase activity specifically bound was very low as compared with the amount of a-hANP [ 1-28] subjected to biotinylation (7.0% or less of a-hANP [ 1-28] subjected to biotinylation on a molar basis). This remains to be improved by increasing the efficiency of

L ' 10 I ' 100 I ' 1000 I ' 4 0 ' I

a-hANP [I-281 ( arnol I tube )

Fig. 4. Dose-response curves of a-hANP [ 1-28] by hetero-two-siteenzyme immunoassay with different numbers of complex transfer steps. Triangles, squares, and circles indicate P-D-galactosidase activity bound to (anti- 2.4-dinitrophenyl group) IgG-, anti-fluorescein IgG-, and (anti-rabbit IgG) IgG-coated polystyrene balls, respectively. Namely, triangles, squares, and circles indicate the assays with no transfer step, one transfer step, and two transfer steps, respectively. Details of the three methods are described in the Discussion. Vertical bars indicate standard deviations of 3-5 determinations.

207

each process involved (biotinylation, trap and elution of biotinylated a-hANP [ 1-28] and complex formation).

a-hANP has been measured with a detection limit of 10 amol by two-site enzyme immunoassay using two antibodies specific for the ring structure and C-terminus of a-hANP (5). In general, however, it is not always possible to prepare two high affinity antibodies, which can bind simultaneously to relatively short chain peptide molecules. In contrast, the pres- ent method detected 1 amol of a-hANP [l-281 using only one antibody as described above. By the same method, 1 am01 of Arg'-vasopressin, a nine amino acid single chain peptide with a intramolecular disulfide bridge, has also been detected, which will be described in detail elsewhere. This together with the results described above suggests usefulness of the present method for the detection of other haptens including peptides which are present at low concentrations in plasma.

REFERENCES

1. Flynn TG, de Bold ML, de Bold AJ: The amino acid sequence of an atrial peptide with potent diuretic and natriuretic properties. Biochem Biophys Res Commun 117359-865, 1983.

2. Kangawa K, Matsuo H: Purification and complete amino acid sequence of a-human atrial natriuretic polypeptide (a-hANP). Biochem Biophys ResCommun 118:131-139, 1984.

3. Kangawa K, Fukuda A, Matsuo H: Structural identification of P- and y-human atrial natriuretic polypeptides. Nature 313:397-400, 1985.

4. Nakao K, Sugawara A, Shiono S , Saito Y, Morii N, Yamada T, Itoh H, Mukoyama M, Arai H, Sakamoto M, Imura H: Secretory form of atrial natriuretic polypeptide as cardiac hormone in humans and rats. Can J PhysiolPharmacol65:1756-1761, 1987.

5. Hashida S , Isbikawa E, Mukoyama M, Nakao K, Imura H: Direct mea- surement of a-human atrial natriuretic polypeptide in plasma by sensi- tive enzyme immunoassay. JCl in Lab Anal 2:161-167, 1988.

6. Watanabe H, Wakimasu M, Kondo K: A sensitive enzyme immunoas- say for atrial natriuretic polypeptide. J Immunof Methods I24:25-28, 1989.

7. Ishikawa E, Imagawa M, Hashida S , Yoshitake S , Hamaguchi Y, Ueno T: Enzyme-labeling of antibodies and their fragments for enzyme immu- noassay and immunohistochemical staining. J Immunoassay 4:209- 327, 1983.

208 Hashida and lshikawa

8. Webster GC: Comparison of direct spectrophotometric methods for the measurement of protein concentration. Biochim Biophys Actu 207:371- 373, 1970.

9. Peters Jr. T: Serum albumin. In The Plasma Proteins. Structure, Func- tion, and Genetic Control, Vol. 1, FW Putnam, ed. Academic Press, NewYork, 1975,pp 133-181.

10. Carsten ME, Eisen HN: The interaction of dinitrobenzene derivatives with bovine serum albumin. JAm Chem Soc 75:4451-4456, 1953.

1 1 . Kawamura Jr. A: Fluorescent Antibody Techniques and Their Appli- cations, University of Tokyo Press, Tokyo, 1977, pp 13-76.

12. Ruan K-h, Hashida S , Yoshitake S, Ishikawa E, Wakisaka 0, Yama- moto Y, Ichioka T, Nakajima K: A micro-scale affinity-purification of

Fab’-horseradish peroxidase conjugates and its use for sandwich en- zyme immunoassay of insulin in human serum. Clin Chim Acra 147:

13. lshikawa E, Kato K: Ultrasensitive enzyme immunoassay. Scand J

14. Green NM: Avidin. AdvProrein Chem 29:85-133, 1975. 15. Udenfriend S, Stein S, Bohlen P, Dairman W, Leimgruber W, Weigele

M: Fluorescamine: A reagent for assay of amino acids, peptides, pro- teins, and primary amines in the picomole range. Science 178237 1-872, 1972.

16. Wilchek M, Bayer EA: The avidin-biotin complex in bioanalytical appli- cations. Anal Biochem 171:l-32, 1988.

167-172, 1985.

Immunol 8(Supp1.7):43-55, 1978.