A tetrazolium method for staining peroxidase labels in blotting assays

Journal of Immunological Methods, 95 (1986) 71-77 Elsevier

71

JIM 04145

A tetrazolium method for staining peroxidase labels in blotting assays Kazuhisa Taketa 1,,, Eriko Ichikawa 1 and Toshiro H a n a d a 2 I Health Research Center, Kagawa University, 1-1 Saiwai-cho, Takamasu 760, and 2 Diagnostic Division, Research Center, Wako Pure Chemical Industries, Ltd., 1633 Matoba, Kawagoe-shi, Saitama 350, Japan

(Received 28 May 1986, accepted 28 July 1986)

A sensitive staining method of horseradish peroxidase-labeled immunoglobulins on nitrocellulose membrane was established by employing a reaction chain leading to formazan formation with phenol as a substrate of peroxidase and NADH as a hydrogen donor to reduce nitro blue tetrazolium. Higher concentrations of NADH relative to phenol were necessary to increase the intensity of staining and to ensure a wide dose-response range of color production with respect to the applied enzyme activities. By an optimized tetrazolium method in combination with antibody-affinity blotting, as low as 4 ng/ml a-fetoprotein was detected and 3-4-fold greater color intensities in a working assay range as compared with those of existing methods were obtained. The present technique of peroxidase staining may prove to have a wide application for the enzyme immunoassay using blotting modalities. Key words: Horseradish peroxidase; Nitro blue tetrazolium; Phenol; Enzyme immunoassay; Western blotting; ts-Fetoprotein

Introduction The sensitivity of enzyme immunoassays for the detection of proteins using blotting techniques is largely dependent on the intensity of the color produced by final enzyme reaction as well as the quality of the antibodies used in immunoassay (Shields and Turner, 1986). In peroxidase immunoassay of a-fetoprotein (AFP) by antibodyaffinity blotting (Taketa et al., 1985), the sensitivity was limited at the level of color production even if 3,3'-diaminobenzidine, a highly sensitive substrate for horseradish peroxidase (HRP), was used.

* To whom correspondence should be addressed. Abbreviations: HRP, horseradish peroxidase; NBT, nitro blue tetrazolium; AFP, a-fetoprotein; NC, nitrocellulose

A peroxidase-mediated system of formazan formation for determination of lipoprotein lipids has been reported by Hanada et al. (1985) in abstract form. In this system, the color is produced by the reduction of tetrazolium salt, unlike the ordinary color reactions of peroxidase involving the oxidation of substrate to form colored products. Since formazan is a highly colored and stable product, the reactions leading to formazan are widely used not only for the staining of NAD + or NADP+-linked enzymes but also for the localization of alkaline phosphatase (McGadey, 1970). In the present study, the reductive system of peroxidase staining was established using HRP-labeled immunoglobulins blotted to nitrocellulose (NC) membranes. In detection of AFP by an optimized tetrazolium method in combination with the antibody-affinity blotting, 3-4-fold higher sensitivities over the routinely employed methods of peroxidase staining were attained.

0022-1759/86/$03.50 © 1986 Elsevier Science Publishers B.V. (Biomedical Division)

72 Materials and methods

Goat anti-rabbit IgG (H + L)-HRP conjugate (IgG-HRP), NC membranes, 4-choloro-l-naphthol and Bio-Dot microfiltration apparatus were purchased from Bio-Rad Laboratories, Richmond, CA. Nitro blue tetrazolium (NBT), NADH, NADPH, 3,3'-diaminobenzidine tetrahydrochloride and bovine albumin, fraction V, were the products from Sigma Chemical Co., St. Louis, MO. Affinity-purified horse antibodies to human AFP were kindly donated by H. Taga and H. Hirai, Tumour Laboratory, Tokyo. Japanese AFP standard (1000 ng/ml) was obtained from Nippon Bio-Test Laboratories, Tokyo. F(ab')2 fragments of rabbit immunoglobulins to human AFP were prepared from rabbit immunoglobulins to human AFP (DAKO-immunoglobulins, Copenhagen) by pepsin digestion according to Erlich et al. (1978). A hemolysate prepared by lysis of washed human red blood cells in water was used as a source of hemoglobin after determination of its concentration at 541 nm. All other compounds and reagents obtained were of the highest grade available.

Dot-blotting To wells of Bio-Dot apparatus assembled with a sheet of NC membrane, 0.4 ml samples of serially diluted IgG-HRP solutions were applied, leaving the flow valve of the apparatus open. After 30 min, the NC membrane was washed twice with 0.05% Tween 20, taken out by disassembling the apparatus, blocked for 2 min with 2.0% Tween 20 and washed. Solutions for the dilution, washing and blocking were made in Tris-buffered saline (20 mM Tris-HC1, pH 7.5, 500 mM NaC1). Strips of NC membrane, each with a lane for the serially diluted samples, were made and treated with color developing solutions of varied compositions. Japanese AFP standard was serially diluted with 10 mM phosphate buffer, pH 7.2, containing 150 mM NaC1 and 3.0 mg/ml bovine albumin, and 0.04 ml samples were dot-blotted to NC membranes which were precoated with affinity-purified horse immunoglobulins to human AFP (the antibody-affinity blotting of Taketa et al. (1985)). The AFP blots were treated overnight with F(ab')2 fragment of rabbit anti-human AFP (12 /xg/ml) and allowed to react with IgG-HRP for 60 min

under conditions described previously (Taketa et al., 1985).

Color development Staining solutions were prepared immediately before use by adding aqueous solutions of phenol and NBT to 5 ml of 100 mM phosphate buffer, pH 7.0, containing NADH, and adjusting the final volume to 10 ml with water to give indicated concentrations of the additions. Reaction was started by adding 0.02 ml of 10% H202, unless otherwise noted, to the reaction mixture and by bringing the complete staining system into contact with NC strips. After 30 rain of incubation, the NC strips were washed with water, dried, treated with decalin and subjected to densitometric scanning with a Bio-Rad Model 1650 transmittance/ reflectance densitometer equipped with a 150 nm range of wavelength with a yellow-biased light source, set at the lowest sensitivity in the transmittance mode and connected with a Linear recorder, Model 156 (Linear Instruments Corp., Irvine, CA). For evaluation of the test results, the color intensity was expressed in mm of peak heights. The entire procedure of dot-blotting and staining was carried out at 25 _+ 2°C, unless otherwise indicated. Methods for peroxidase staining with 3,3'-diaminobenzidine and 4-chloro-l-naphthol are given elsewhere (Taketa et al., 1985). Colorimeteric and kinetic assays Colorimetric determination of HRP activities was carried out according to Hanada et al. (unpublished observation) with a final reaction mixture containing 50 mM phosphate buffer, pH 7.0, 0.5 mg/ml NADH, 0.5 mg/ml phenol, 0.25 mg/ml NBT, 0.05% Triton X-100, 0.02% H202 and 0.1 ml of 1000-fold diluted IgG-HRP solutions in a volume of 3.0 ml. After incubation, 1.0 ml of 20 mg/ml sodium dodecyl sulfate was added and the color was read at 560 nm against a blank solution without IgG-HRP. HRP activity of IgG-HRP was also assayed in 10 mM phosphate buffer, pH 6.0, using o-dianisidine as substrate according to the procedure given in Worthington Enzyme Manual - 1972 (Worthington Diagnostics, Freehold, N J). 1 U was defined as equvalent to 1 #mol of substrate hydrolyzed per rain employing the molecular absorbancy of 11.3 x 106/cm at 460 nm. The

73

IgG-HRP preparation used in this study had a peroxidase activity of 1.06 U / m l 1000-fold diluted solution. The rate of NADH oxidation was determined by following the absorbance change at 340 nm in a reaction mixture containing 50 mM phosphate buffer, pH 7.0, 0.2 mg/ml NADH, 0.375 mg/ml phenol, 0.015% H202 and 0.1 ml of 1000-fold diluted IgG-HRP in a final volume of 4 ml, unless otherwise indicated. The colorimetric and kinetic assays of peroxidase activity were carried out at 25°C with a Hitachi digital spectrophotometer, Model 624 (Hitachi, Tokyo).

50

40

30

20

i0

I0

20

30

Time (min)

Fig. 2. Time courses of color production. IgG-HRP concentration of diluted samples: o, 0.50; and A, 0.25 lal/ml. Staining reagents contained 0.025% H202, 0.5 m g / m l NBT, 0.5 m g / m l NADH and 0.2 mg/ml phenol.

Results IgG-HRP was serially diluted, dot-blotted to NC membranes and stained by formation of formazan under varying conditions. The highest values of densitometric peak height were obtained at pH 7.0, as shown in Fig. 1, and this buffer system was used in the subsequent analyses. Time courses of color development are presented in Fig. 2. The rate of color production decreased slightly as the incubation period of time increased. The color intensity in 60 min was 1.5 times that in 30 min and the background stain became apparent. Incubation at 37°C gave 1.4 times as intense a color

100

80 %.

60

E

50

\

40

\ \ \ 20 \

20

[/"

10

0

0

0,25

0,50

0.75

1,00

Concentration of IgG-HRP ( ~ l / m l ) I

I

I

I

6.0

6,5

7.0

7.5

pH Fig. 1. Effect of pH on the color production. IgG-HRP concentration of diluted samples: o, 0.50; and A, 0.25 /~l/ml. Staining solutions contained 0.5 mg/ml NBT, 0.5 m g / m l NADH and 0.2 m g / m l phenol in 50 mM phosphate buffers, pHs 6.0, 6.7 and 7.0, and in 50 mM Tris-HCl, pH 7.5.

Fig. 3. Effect of NADH and IgG-HRP concentration on the color production. NADH concentration: O, 0.75; O, 0.50; It, 0.25; and A, 0.10 mg/ml. . . . . . . , the lowest peak height of tinged stains (see Fig. 4). IgG-HRP concentrations of diluted samples before application to NC membranes are given. Developing solutions contained 0.75 mg/ml NBT and 0.2 m g / m l phenol.

74

as at 25 o C and also a slightly darker background. Changes in concentrations of H 2 0 2 from 0.006 to 0.025% and of NBT from 0.1 to 0.75 mg/ml had little effect on the resulting color intensity. The effect of NADH concentration on the formazan formation with varied concentrations of IgG-HRP is shown in Fig. 3. As the activity of IgG-HRP dot-blotted to NC membrane was increased, the color intensity increased to some extent and then decreased. The peak height of colored dots increased and the peak shifted to higher enzyme levels as the concentration of NADH was increased. The suppression of color development by higher activities of IgG-HRP was less marked in the periphery of dots, resulting in the formation of colored rings (Fig. 4). When NC blots with stained dots were further incubated in a staining solution without NADH, the color of formazan did not faint. Increasing the concentration of phenol in the presence of 0.5 mg/ml NADH enhanced the color production in a lower range of IgG-HRP activity and suppressed the color production in a higher range of IgG-HRP activity (Fig. 5B). The suppressing effect of phenol was counteracted by increasing the concentration of NADH to 1.0

200

I

160

120

8O

4O

0

4O

I

0 0

i 0.25

l 0.50

I 0.75

I 1,00

C o n c e n t r a t i o n o f IgG-HRP ()Jl/ml)

Fig. 5. Effect of phenol and NADH concentration on the color production. A: 1.0 m g / m l NADH; and B: 0.5 m g / m l NADH. Phenol concentration: O, 0.25; A, 0.15; and II, 0.10 mg/ml. ...... and IgG-HRP concentration, see the legend to Fig. 3. The concentration of NBT was 0.25 mg/ml.

60

v

i t.-

40

"F~

mg/ml (Fig. 5A). At concentrations of 0.1 mg/ml phenol and 1.0 mg/ml NADH, a linear or slightly sigmoidal relationship was obtained between the color intensity and the concentration of IgG-HRP. NADH could be replaced by NADPH with identical results. The interaction of phenol with NADH was studied by following the rate of absorbance change at 340 nm and the results are presented in Fig. 6. The oxidation of NADH by peroxidase with H202 was negligible, while it was markedly increased by the addition of phenol (Fig. 6A and B). The rate of NADH oxidation was not reduced by lowering the NADH concentration from 0.20 to 0.04 mg/ml (or from 0.28 to 0.056 mM), while it was significantly reduced by lowering the phenol concentration from 0.375 to 0.075 mg/ml (or from 4.0 to 0.8 mM) (Table I). When IgG-HRP was preincubated with phenol and H202, the rate of NADH oxidation was markedly reduced (Fig. 6C), al-

1

,v. (1.1

20

1

2

3

4

5

6

7

8

Fig. 4. Effect of IgG-HRP level on the pattern of dot staining. IgG-HRP concentration of diluted samples; 1, 0; 2, 0.016; 3, 0.031; 4, 0.063; 5, 0.13; 6, 0.25; 7, 0.50; and 8, 1.00 #l/ml. A staining mixture contained 0.5 mg/ml NBT, 0.5 mg/ml NADH and 0.2 mg/ml phenol.

75

1.5 A

NADH H202

NADH phenol

1,0

B

.,_,...,,_,.,

[ H202

o

[ ~

i\

phenol

0(ldlt 10n01

IgG-HRP

0,5 NADH

phenol l H202 0

I 5

L I0

0

!

I

5

I0

Tlme

~

0

5

!

!

I0

15

(mln)

Fig. 6. Effect of the order of addition of reactants on the oxidation of NADH by IgG-HRP. Arrows indicate the addition of solutions.

though a newly added IgG-HRP showed an absorbance change comparable to the control rates (Fig. 6A and B). The results indicated that the peroxidase was inactivated during preincubation with phenol and H202 and that the presence of N A D H prevented the enzyme from inactivation. Similar results were obtained when IgG-HRP blots were preincubated with phenol, NADH and H202 for 15 min and the blots were transferred to a complete staining mixture and incubated for 15 min (Fig. 7). Preincubation with phenol alone slightly enhanced the color development for lower levels of IgG-HRP. However, when phenol and TABLE I RATES OF NADH OXIDATION AT DIFFERENT CONCENTRATIONS OF NADH AND PHENOL a Concentrations (mg/ml) NADH

Phenol

0.20 0.20 0.04 0.04

0.375 0.075 0.375 0.075

A A34o/min

0.038 0.010 0.053 0.014

a Reaction was initiated by addition of H202.

present during preincubation, the resuiting color intensities were markedly reduced. The reductions in color intensity were partially eliminated by the presence of NADH. In cases of marked reductions in the formation of formazan due to the absence or the insufficient concentration of NADH, increasing intensities of brown discoloration of dots in proportion to the applied peroxidase activities were observed at the end of preincubation. When IgH-HRP, phenol and H 2 0 z werepreincubated in the absence of NADH, NBT and Triton X-100 in a colorimetric assay system, a brown color with an absorption maximum around 410 nm developed and the production of formazan was markedly suppressed in subsequent incubation in the complete system. The absorption maximum shifted to longer wave length as the preincubation was prolonged. N A D H partially inhibited the formation of the brown color and markedly reduced the suppressive effect of phenol and HzO z o n the formazan production. When Triton X-100 was present during the preincubation, the shift in absorption maximum was blocked, and the brown color yield and the suppression of H202 were

76 140

200

120 150

g

100

J

N 100 \\\ j-.--o

80

J

50

,I

6O

i 50

lO0

150

200

AFP concentrotion (nglml)

40

20

I

I

I

0.25

0.50

0,75

|

1.00

Fig. 8. Color production of NC membrane-boundAFP stained by peroxidase immunoassaywith different substrates. AFP was bound to NC membranes by antibody-affinity blotting, o, a tetrazolium method with 2.0 mg/ml NADH and 0.2 rag/m1 phenol; A, a tetrazolium method with 1.0 mg/ml NADH and 0.1 mg/ml phenol; O, 4-chloro-l-naphthol; and It, 3,3'-diaminobenzidine.

Concentrotion of IgG-HRP OJllml)

Fig. 7. Effect of preincubation of IgG-HRP blots with phenol, NADH and H202 on the color production in a complete staining mixture. Preincubation for 15 min with the following additions: O, buffer alone; O, 0.5 mg/ml phenol; A, 0.5 mg/ml phenol and 0.02% H202; and II, 0.5 mg/ml phenol, 0.02% H202 and 1.0 mg/ml NADH. The complete system contained 0.1 mg/ml phenol, 1.0 mg/ml NADH and 0.25 mg/ml NBT. - . . . . . and IgG-HRP concentration, see the legend to Fig. 3. formazan formation were diminished. The results on the effects of N A D H and Triton X-100 in preincubation on the production of formazan are summarized in Table II. TABLE II EFFECT OF PREINCUBATIONOF PEROXIDASE,NADH AND H202 WITH AND WITHOUTTRION X-100 ON THE COLOR PRODUCTIONIN THE COMPLETEASSAYSYSTEM a Additions during preincubation other than the buffer and IgG-HRP

As6o

Phenol, H202 : Phenol, H202, NADH

0.061 0.471 0.850 0.880 0.745 1.018

H202 Triton X-100, phenol, H202 Triton X-100, phenol, H202, NADH Triton X-100, H202

a 15 min preincubation, followedby 15 re.in incubation.

In colorimetric assay of peroxidase activity, the absorbance increase at 560 nm caused by the addition of 0.1 ml of 1.0 m g / m l hemoglobin was 13% of that produced by the addition of 0.1 ml of 1000-fold diluted IgG-HRP. The same concentration of hemoglobin inhibited the peroxidase-dependent color production by 54%. However, the hemoglobin solution showed no or negligible color production when it was dot-blotted to N C membrane. Serially diluted A F P was dot-blotted to N C membranes precoated with horse antibodies to AFP. Four strips o f blot were made and treated with F(ab')2 fragment of rabbit anti-AFP, followed by IgG-HRP. Two of the strips were stained by the present tetrazolium method with 0.3 m g / m l NBT, one with 0.1 m g / m l phenol and 1.0 m g / m l N A D H and the other with 0.2 m g / m l phenol and 2.0 m g / m l N A D H , and the rest two were stained each with 3,3'-diaminobenzidine and 4-chloro-1naphthol. Results are shown in Fig. 8. The tetrazolium method gave darker dots of A F P stain than the methods employing 3,3'-diaminobenzidine and 4-chloro-l-naphthol. With 0.2 m g / m l phenol and 2.0 m g / m l N A D H , a lower detection limit of 4 n g / m l was obtained and the staining intensity increased linearly up to 100 n g / m l ; thus the intensities of color production were 3-4-fold greater in the working assay range covering up to

77 100 n g / m l as compared with the two routinely used methods of peroxidase stain.

Discussion The exact mechanism of the peroxidase reaction leading to the formation of formazan is not known. The involvement of superoxide anion in this reaction has been ruled out based on the observation that superoxide dismutase had no effect on the color production (unpublished observation of Hanada et al.). The phenoxyl radical produced by 1-electron oxidation of phenol by H R P (Nakamura et al., 1985) and p,p'- and o,o'biphenoxy radicals resulting from its further oxidation (Subrahmanyam and O'Brien, 1985) might well serve as active intermediates, since the radicals can be readily reduced by N A D H , thus preventing the formation of colored polymers which were ineffective in transferring hydrogen to N B T and rather inactivated HRP. Although the recycling of the active intermediates as catalysts is theoretically possible, formazan appears to be produced stoichiometrically by the cleavage of H202 (unpublished observation of Hanada et al.). Although higher concentrations of N A D H relative to phenol are essential in order to assure a wider linear-response range of color production to given activities of peroxidase, the following assay system may be recommended as an optimization: 2.0 m g / m l N A D H , 0.2 m g / m l phenol, 0.02% H202, 0.3 m g / m l NBT in 50 mM phosphate buffer, p H 7.0, for 30 min incubation at 25°C. When a central fainting or edge effect is noted in stained dots or bands, the concentration of N A D H should be further increased. Hemoglobin, at concentrations less than 1.0 mg/ml, produced no

color on N C membrane, although it may interfere with the color production by peroxidase. The presence of diaphorase and superoxide-forming systems, which cause peroxidase-unrelated formation of formazan, could be checked by omitting H202 from the staining system. In application of the tetrazolium method to visualization of AFP separated by lectin affinity electrophoresis (Taketa et al., 1985), the splitting of AFP bands or the presence of extra bands has not been noted as compared with the results obtained with 3,3'-diaminobenzidine as substrate for peroxidase reaction (Taketa et al., unpublished observation). Thus, the present technique of peroxidase staining has a wide application for localization of peroxidase labels in enzyme amplified blotting assays.

Acknowledgements The authors wish to thank Drs. H. Taga and H. Hirai, Tumour Laboratory, Tokyo, for their active cooperation in this study by providing us with purified antibodies to human AFP.

References Erlich, H.A., S.N. Cohen and H.O. McDevitt, 1978, Cell 13, 681. Hanada, T., S. Nobara, K. Yamanishi, K. Arai and Y. Sakagishi, 1985, Physico-Chem.Biol. 29, 300. McGadey, J., 1970, Histochemie23, 180. Nakamura, M., I. Yamazaki, T. Kotani and S. Ohtaki, 1985, J. Biol. Chem. 260, 13546. Shields, J.G. and M.W. Turner, 1986,J. Immunol. Methods 87, 29. Subrahmanyam, V.V. and P.J. O'Brien, 1985, Xenobiotica 15, 873. Taketa, K., E. Ichikawa, H. Taga and H. Hirai, 1985, Electrophoresis 6, 492.