187,89-93
ANALYTICALBIOCHEMISTRY
(1990)
Ultrasensitive Enzymatic Radioimmunoassay Fusion Protein of Protein A and Neomycin Phosphotransferase II in Two-Chamber-Well Microtiter Plates H.-D.
Hunger,’
Chr. Flachmeier,
G. Schmidt,
Academy of Sciences of the GDR, Central Institute Berlin DDR-1115, German Democratic Republic
Received
November
G. Behrendt,
of Molecular
and Ch. Coutelle
Biology, Department
of Human
Molecular
Genetics,
6,1989
A new sensitive method for antigen detection employing a phosphorylation reaction is described using human serum albumin as a model. The antigen is initially bound to the surface of polystyrene microtiter plates and reacted with an antibody (rabbit). A microbiologitally produced bifunctional fusion protein of protein A and neomycin phosphotransferase II (NPT II) serves as a second immunological reagent by virtue of its protein A component. The detection is based on the phosphorylation of an aminoglycoside antibiotic by the NPT II moiety of the fusion protein using [T-~‘P]ATP as a cosubstrate. This reaction is performed in solution and the evaluation is accomplished by dotting aliquots of the reaction mixture onto phosphocellulose paper, washing with water, and autoradiography. Microtiter plates with a specially designed lo-&volume reaction chamber are particularly advantageous for this procedure. The sensitivity of detection is currently 10 fg (1 pg/ml) of antigen. 0 1990 Academic Press. Inc.
Enzyme-linked immunosorbent assays (ELISA)’ (1) are among the most widely used techniques for detection of antigens or antibodies from biological sources. The enzymes peroxidase, alkaline phosphatase, and P-D-galactosidase are used mainly for labeling, in most cases of a second antibody. The enzyme reaction with chro1 To whom correspondence should be addressed. ’ Abbreviations used: NPT II, neomycin phosphotransferase II; HSA, human serum albumin; PBS, 0.14 M NaCl, 2.6 mM KCl, 7 mM Na/K phosphates, pH 7.2; PBST, 0.05% (v/v) Tween 20 in PBS; PBSM, 5% (w/v) emulsion of nonfat dry milk powder in PBS; PBSTM, 5% (w/v) emulsion of nonfat dry milk powder and 0.05% (v/v) Tween 20 in PBS; ELISA, enzyme-linked immunosorbent assay; SIA, slide immunoassay; RIA, radioimmunoassay. 0003.269’7/90
Using a
$3.00
Copyright 0 1990 by Academic Press, Inc. All rights of reproduction in any form reserved.
mogenic or fluorogenic substrates such as o-phenylenediamine, p-nitrophenyl phosphate disodium salt, and pnitrophenyl+?-D-galactopyranoside or 4-methyl-umbelliferyl-P-D-galactoside allows the identification and/or quantitative determination of the reaction products and thereby of the biomolecule by spectrophotometry or fluorometry. The ELISA is usually carried out in microtiter plates and its principle has been extended to small sample volumes down to drops of 10 ~1 by the slide immunoassay (SIA) (2) performed on flat glass slides. However, to obtain extremely high sensitivities the probes or biomolecules themselves must be radioactively labeled, generally with iz51, and are detected by a radioimmunoassay (RIA) (3). In ultrasensitive enzymatic immunoassays (4) the amplification reaction of an enzyme is used to produce a radioactive detection signal. We effectively applied this principle to the phosphorylation reaction of aminoglycoside antibiotics. This is achieved by a protein A-NPT II fusion protein (5). Its protein A moiety acts as a second immunological reagent recognizing the antibody while its NPT II enzyme activity, using [y-32P]ATP as a cosubstrate, phosphorylates neomycin which is detected by autoradiography. This paper describes the successful attempt to combine the advantages of a microtiter plate assay with the small reaction volumes of the SIA. MATERIALS
AND
METHODS
Kanamycin sulfate and neomycin sulfate were obtained from Medexport (USSR); HSA (68 kDa) was from Serva Feinbiochemica (FRG); dithiothreitol was from Boehringer-Mannheim (FRG); [T-~~P]ATP (110 TBq/mmol) was from ZfK Rossendorf (GDR); HS 11 Xray film was from VEB Fotochemisches Kombinant Wolfen (GDR); nonfat dry milk was from VEB Dauermilchwerke Stendal (GDR); rabbit antiserum against 89
90
HUNGER
ET
HSA was from VEB Germed (GDR); thin-layer plastic plates and PEI cellulose F were from Merck (FRG); polystyrene microtiter plates, flat-bottom type, volume per well 0.4 ml, were from VEB Polyplast Halberstadt (GDR); and phosphocellulose paper P81 was from Whatman Chemical Separation, Ltd. (UK). Construction
of Two-Chamber-
Well Microtiter
Plates (6)
A cone-shaped 12-pl-volume reaction chamber was drilled into the center of each well of a flat-bottom polystyrene microtiter plate using a 4.8-mm-diameter drill (see Fig. 1). Construction of the Gene Coding for the Protein II Fusion Protein and Expression in Escherichia coli Cells (7)
of the Antigen
Specially made two-chamber microtiter plates were used. Ten microliters of a dilution series of HSA containing 1 pg, 100 ng, 10 ng, 1 ng, 100 pg, 10 pg, 1 pg, and 0 pg per milliliter of PBS and 1 pug/ml of nonfat dry milk were each pipetted into the reaction chambers of six rows of 9 wells each and allowed to bind for 4 h. The wells were then washed three times for 2 min with PBST (12). For quantitative determination, a dilution series of HSA containing 10 ng, 1 ng, 100 pg, 10 pg, 1 pg, and 0 pg per milliliter PBS and 1 pg/ml of nonfat dry milk was used and the procedure carried out as described, the controls being the same as above. Immunological
Reaction
Blocking of the antigen-loaded plates was performed using 400 ~1 of PBSTM for 30 min. One hundred microliters of rabbit antiserum against HSA (Fig. 2, +) diluted 1:5000 in PBSTM was pipetted into the wells of the first two rows. The wells of the third and fourth rows (Fig. 2, -) were filled only with 100 ~1 of PBSTM and those of the fifth and sixth rows (Fig. 2, CS) with a rabbit control serum (without the specific antibody, diluted 1:lOOO in PBSTM). The antibody reaction was allowed to proceed for 1 h, after which the wells were washed 10 times for 5 min with PBST. The wells were blocked again for 30 min with 400 ~1 of PBSM. One hundred microliters of the E. coli cell lysate containing the protein A-NPT II fusion protein was added to 10 ml of PBSM. One hundred microliters of this emulsion was added to each well and allowed to react for half an hour. Thereafter the wells were washed 5 times for 3 min with PBS. Enzyme
Reaction
The enzyme reaction was carried out in 10 ~1 of enzyme buffer (20 mM Tris-HCl, pH 7.2; 125 mM NH&l; 20 mM MgCl,; 10 mM dithiothreitol) containing [T-~“P]ATP (5 X lo6 cpm/ml) and 1 ~1 of a kanamycin stock solution (50 mg/ml) .
the Evaluation
A pellet of lOlo E. coli cells was resuspended in 1 ml of cold lysis buffer (25 mM Tris-HCl, pH 7.4) and sonicated three times for 5 s (Branson sonifier, output 50 W, range 5). Between sonications the tube was cooled on ice. After centrifugation (Eppendorf centrifuge, 4”C, 5 min) the clear supernatant was pipetted into a second Eppendorf tube. For immediate use the lysate was stored on ice; otherwise, it was frozen at -20°C. Microtiter
Immobilization
A-NPT
The DNA coding for NPT II (EC 2.7.1.95) from a derivative of the plasmid pSV2-neo (8) was inserted into the EcoRI/BamHI sites of the multicloning sequence of the plasmid pRIT2T (9) which contains the sequence for the Fc binding part of the protein A gene. E. coli NFl cells (10) were transformed with the resulting plasmid (pSAL1). Molecular cloning procedures were performed according to Maniatis et al. (11). E. coli cells were grown at 30°C for 4 h to a density of about lo* cells/ml. The cells were cultured for a further 2 h at 42”C, the temperature at which the protein ANPT II fusion protein becomes expressed due to the Ps promoter at the 5’-terminus of this gene. The cells were then harvested by centrifugation (4000 rpm, 10 min). They were resuspended in Tris buffer (25 mM Tris-HCl, pH 8.0; 50 mM glucose; 10 mM EDTA) to a cell density of lOlo cells/ml, and l-ml aliquots were filled into Eppendorf tubes and centrifuged (5 min, 4°C). The supernatant was decanted, and the pellets were frozen and stored at -20°C. They can be stored in this form for at least 1 year without loss of immunological or enzymatic activity. Preparation of E. coli Cell Lysates Containing Protein A-NPT II Fusion Protein
AL.
Plate Assay
All steps (unless otherwise stated) are performed room temperature with gentle shaking.
at
of the Enzyme
Reaction
One mi(a) Dotting on negatively charged matrices. croliter of the reaction mixture of each well was dotted onto P81 paper. After drying for 30 s the paper was washed three times for 5 min with bidistilled water and then twice for 15 min at 70 to 80°C. The paper was then exposed to an X-ray film. (b) Absorbing the reaction mixture tophosphocellulose paper strips. Five microliters of the reaction mixture was absorbed by 30 X 2-mm phosphocellulose paper strips with a tip at the end. The strips were treated as described in (a) and evaluated by autoradiography or by scintillation counting.
ULTRASENSITIVE
i FIG. 1. Two-chamber-well (3) with a volume of 5-15 of a flat-bottom polystyrene
ENZYMATIC
RADIOIMMUNOASSAY
I microtiter plate. The reaction chamber ~1 is drilled into the bottom (4) of a well (2) microtiter plate (1).
USING
A FUSION
PROTEIN
91
Already after 3 h of enzyme reaction at 37°C (Fig. 2A, +) all seven dilutions of HSA are visible on the autoradiogram. The signals of detection for HSA (+) are significantly enhanced after the enzyme reaction is allowed to proceed further, e.g., 16 h at room temperature. The procedure allows a quantitative determination of the antigen in the femtogram range (Fig. 3). For this purpose, larger volumes (5 ~1) of the reaction mixture are absorbed by phosphocellulose paper strips (Fig. 3A). The distance of capillary suction of the liquid is about 2 cm in this case but the [32P]kanamycin phosphate is concentrated on the first 3 mm of the phosphocellulose paper strips. The evaluation can be made both by autoradiography (A) and by Cerenkov counting of the phosphocellulose paper strips. The calibration curve should be obtained by processing standard amounts of the antigen on the-same microtiter plate and plotting the corresponding Cerenkov countings. The shape of the measured curve is determined mainly by the absorption characteristics of the antigen to the polystyrene surface (13,14).
Thin-Layer Chromatography PEI Cellulose
of the Reaction Mixture
on
One microliter of the sample was dotted onto the starting point of the PEI cellulose plates. Chromatography was performed with B-cm plates developed in 0.75 M sodium phosphate buffer, pH 3.5. The evaluation was performed by autoradiography.
An additional way of analyzing the reaction mixture is by its thin-layer chromatographic separation on polyethylene imine cellulose plates (Fig. 4). The application of this method is both a control for the completeness of reaction of [T-~~P]ATP forming [32P]kanamycin phosphate (lane 2) and a control for the stability of the [T-~~P]ATP (lane 1). Lane 3 shows the position of [3”P]orthophosphate for comparison.
RESULTS
Our enzymatic radioimmunoassay is based on the reaction of a bifunctional detection probe, the protein ANPT II fusion protein, and a specially designed twochamber-well polystyrene microtiter plate with a lo-p1 reaction chamber, shown in Fig. 1. The antigen is bound to the small surface of the reaction chamber and reacted with the specific antibody. This antibody is then recognized by the protein A moiety of the bifunctional probe, while its NPT II enzyme activity is used for detection by 32P phosphorylation of the aminoglycoside antibiotic kanamycin. Blocking, washing, and the specific antibody reaction are carried out with excess volumes (up to 400 ~1) in the larger incubation chamber, while the enzyme reaction is again performed in the small volume (10 ~1) of the reaction chamber. Figure 2 demonstrates the sensitivity of detection of this procedure. As shown, we are able to detect reliably 10 fg (1 pg/ml) of HSA (line 7) by dotting l-p1 aliquots of the reaction mixture onto phosphocellulose paper. Both the negative controls without HSA in PBS (line 8) or in 10 ~1 PBS containing 10 ng of nonfat dry milk (line 9) as well as the negative controls using a rabbit antiserum without specific antiHSA antibodies (CS) or omitting the specific antiserum (-) remain clearly below the level of the detection signals with HSA after reaction with rabbit antiserum containing specific HSA antibodies (+).
0
*
l
9 FIG. 2.
Sensitivity of the two-chamber-well microtiter plate assay. (A and B) Autoradiographs of l-al dots of the reaction mixture after enzyme reaction. Exposure time, 16 h, -70°C. (A) Enzyme reaction at room temperature for 3 h; (B) enzyme reaction performed for a further 16 h in the same microtiter plate as that in A. Line 1,10 ng (1 pg/ml); line 2, 1 ng (100 rig/ml); line 3, 100 pg (10 rig/ml); line 4, 10 pg (1 ng/ ml); line 5, 1 pg (100 pg/ml); line 6, 100 fg (10 pg/ml); line 7, 10 fg (1 pg/ml) of HSA; line 8,O pg of HSA in PBS; line 9,0 pg of HSA and 10 ng (1 pg/ml) of nonfat dry milk in PBS. Lanes +, rabbit antiserum containing specific antibodies against HSA, lanes -, without antiserum; lanes CS, rabbit antiserum without specific antibodies against HSA. The duplicates were obtained from two different rows of the microtiter plate.
92
HUNGER
B I\
2000-
1500E 8
.i 1000-l
\
5oo;
k----T
\
0.1
0.01
pg HSA FIG. 3.
Quantitative evaluation of the microtiter plate assay. 5 ~1 of the reaction mixture is sucked up by phosphocellulose paper strips. (A) Autoradiograph of the strips (exposure time 10 h at room temperature). Lane 1, 100 pg (10 rig/ml); lane 2, 10 pg (1 rig/ml); lane 3, 1 pg (100 pg/ml); lane 4,100 fg (10 pg/ml); lane 5,10 fg (1 pg/ml) of HSA; lane 6, PBS without HSA; lane 7, 10 ng (1 @g/ml) of nonfat dry milk powder in PBS without HSA. (B) Quantitative evaluation of the strips by plotting the Cerenkov countings. The individual points are means of duplicates obtained from two different rows of the microtiter plate. The values of lanes 6 and 7 are 145 and 122 cpm, respectively.
ET
AL.
structed similarly. Further, on the basis of its properties, (i) high specific activity and turnover rate, (ii) good stability at room temperature and higher temperatures, (iii) ready availability because of its ease of preparation, (iv) low molecular weight (30 kDa), and (v) renaturation after treatment with detergents, the enzyme NPT II is suitable for chemical coupling to other molecular probes. In this paper we report on the utilization of the phosphorylation reaction with the substrate kanamycin and the cosubstrate [T-~‘P]ATP in solution. We have also developed a solid-phase assay (“Contact-Copy” method) using this principle (15). At present, the sensitivity of the procedure demonstrated here is limited only by the background arising from nonspecific binding of the first antibody to the material of the microtiter plates. An advantage of our model system is the antigen fixation out of a volume of 10 ,ul. Any application should be checked if a reduction of the sample volume before antigen fixation is possible. For comparison purposes, all sensitivities were recalculated to a volume of 1 ml (given in parentheses in the legend to Fig. 3 and below). The method discussed here exhibits an extraordinary sensitivity of detection (cf. Figs. 2 and 3); i.e., the limit of detection is currently 10 fg (0.15 amol) of HSA (1 pg/ml). In experiments for comparison using a microtiter plate assay and peroxidase second antibody conjugates (data not shown) we could detect 10 pg of HSA (1 rig/ml). The very different setup and performance of the various ELISA and RIA methods as well as the great variety of antigens investigated make a direct comparison of sensitivities somewhat difficult. Methods employing mi-
DISCUSSION
This work was aimed at demonstrating the principle and sensitivity of detection of our enzymatic radioimmunoassay. Using HSA as a test model, we report here the application of the phosphorylation reaction as a general method for the detection of antigens. An advantage of HSA is that it is available as an electrophoretically pure, accurately weighed antigen, making it possible to produce solutions with exact protein concentrations. Our protein A-NPT II fusion protein was constructed by genetic engineering. It may be easily and reproducibly produced microbiologically, and the assay may be performed using nonpurified E. coli lysates. It offers the opportunity for general use of the enzymatic phosphorylation reaction for detecting antigens because protein A binds to the Fc region of IgG from many species (e.g., human, rabbit, guinea pig) and it does not interfere with the antigen-antibody reaction. Of course, the application of the protein A-NPT II fusion protein is limited to problems of analysis in which protein A can be used. However, other fusion proteins with NPT II and different molecular probes such as protein G may be con-
1
2
3
FIG. 4. Analysis of the NPT II enzyme reaction mixture by thinlayer chromatography on PEI cellulose, autoradiograph of the PEI cellulose after 30 min. Lane 1, enzyme buffer containing [y-32P]ATP before the enzyme reaction; lane 2, reaction mixture after nearly complete reaction of [y-32P]ATP to [32P]kanamycin phosphate; lane 3, [32P]phosphoric acid: 1 pl(5000 cpm) of each solution was dotted onto PEI cellulose.
ULTRASENSITIVE
ENZYMATIC
RADIOIMMUNOASSAY
crotiter plates for the ELISA technique (e.g., urease, peroxidase) reach a sensitivity of detection of 50 pg of antigen (reaction mixture of 100 ~1, i.e., 500 pg/ml) (16). The most sensitive ELBA procedure detects 6 pg of HIV p24 core antigene in a volume of 200 ~1 (30 pg/ml) of tissue culture supernatant, plasma, or serum (17). One picogram of human pancreatic secretory trypsin inhibitor (18) has been detected in lOO-~1 aliquots of the sample (10 pg/ml). The SIA technique has been used to detect 50 pg of purified immunoglobin in 10 ~1 (2) and even 5 pg of this protein in 5 ~1 of solution (1 rig/ml) (19). Competitive radioimmunoassays, e.g., for human urokinase, employing ‘251-labeledtracers reach a limit of detection of 10 pg of antigen/ml sample (20). The protein A part of our fusion protein acts as a second immunological reagent, thereby allowing detection of a great variety of antibodies. The design of the microtiter plate wells combines a very small reaction chamber (10 ~1) with the relatively large volume of the incubation chamber (400 ~1). This is particularly advantageous for reduction of the background in the detection reaction because it reduces the amount of nonspecifically bound antibodies. This method certainly may be used for the determination of low concentrations of antibodies from biological samples. Furthermore, the microtiter plate may be constructed with several reaction chambers (6) in one well which may carry different immobilized antigens or antibodies. This would make a simultaneous testing of different biomolecules from the same sample possible. The utilization of 32P as the isotope for detection is especially advantageous for increasing the sensitivity of detection. Our experiments have not come close to exhausting the possibilities of an enhancement of the signals. We believe that such approaches together with efforts to decrease the nonspecific background will be the basis for a further increase in sensitivity. One microliter of the reaction mixture dotted onto phosphocellulose paper (Fig. 2) or 5 ~1absorbed by phosphocellulose paper strips (Fig. 3) is used here for quantitative evaluation of the lo-p1 reaction mixture. The evaluation by dotting of aliquots of the reaction mixture onto and the absorption by phosphocellulose represent very good opportunities for automation of the detection procedure, where large numbers of test samples are used in routine analysis. The application of the method is aimed mainly at the detection of biomolecules which are contained in samples in very low amounts. Since an additional separation step of the unreacted cosubstrate from the reacted substrate is necessary, the technical processing is more complex than that of the state-of-the-art ELISA. Our system exhibits extremely favorable preconditions compared to other enzymatic radioimmunoassays (4), in
USING
A FUSION
93
PROTEIN
which, e.g., [3H]adenosine must be eluted from DEAESephadex columns and scintillation cocktails must be added for evaluation: selective binding of the reacted substrate to the matrix phosphocellulose, simple washing of the matrix with water as the separation step, and detection by Cerenkov counting without scintillation cocktails. The biohazard associated with 32P is less than that with RIA which generally uses 1251.This hazard may be decreased by use of suitable means of protection, especially in the case of automation. We are convinced that the enzymatic radioimmunoassay demonstrated here is currently one of the most sensitive methods for detection of antigens and its application may be extended to a variety of fields. REFERENCES 1. Engvall, E., and Perhnan, P. (1971) Immunochemistry 8,871-874. 2. Conway de Macario, E., Jovell, R. J., and Macario, A. J. L. (1985) BioTechniques 3,138-145. 3. Mayer, R. J., and Walker, J. H. (1987) Immunochemical Methods in Cell and Molecular Biology, pp. 49-58, Academic Press, Boca Raton, FL. 4. Harris, C. C., Yolken, R. H., Krokan, H., and Hsu, I. C. (1979) Proc. Nutl. Acad. Sci. USA 76,5336-5339. 5. Merkel, D., Schmidt, G., Flachmeier, C., Behrendt, G., Coutelle, Ch., and Hunger, H.-D. (1989) J. Biochem. Biophys. Methods 18,
277-286. 6. Hunger, H. D., Schmidt, 7. 8. 9.
(1989) Gefaesskombination DD-WP Application WP Hunger, H.-D., Behrendt, lecular probes, processes Pat. Application 304.934; Southern, P. J., and Berg, 341. Nilsson, B., Abrahmsen, 1075-1080.
G., Behrendt, G., and Flachmeier, Chr. zum Nachweis von Biomolekuelen, G 01 N/330 196 8. G., and Schmidt, G. (1989) Labeled mofor their preparation, and their use, Eur. Cl. G 01 N 33/58. P. (1982) J. Mol. Appl. Genet. 1, 327L., and Uhlen,
M.
(1985)
EMBO
J. 4,
10. Bernard, H.-U., Remaut, E., Hershfield, M. V., Das, H. K., Helinski, D. R., Yanofsky, C., and Franklin, N. (1979) Gene 5,59-76. 11. Maniatis, T., Fritsch, E. F., and Sambrook, J. (1982) Molecular Cloning. A Laboratory Manual, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY. 12. Mohammad, K., and Esen, A. (1989) J. Zmmunol. Methods 117, 141-145. 13. Walsh, J., and Gosling, J. P. (1986) Anal. Biochem. 158,413-423. 14. Cantarero, L. A., Butler, J. E., and Osborne, J. W. (1980) Anal. Biochem. 105,375-382. 15. Hunger, H.-D., Schmidt, G., Flachmeier, Chr., Behrendt, G., and Coutelle, Ch. (1989) Anal. Biochem., in press. 16. Gerber, M., and Sarkar, S. (1988) J. Plant Dk. Protect. 95, 544-
550. 17. Improved HIV p24 core antigen ELISA has greater sensitivity, known specificity (1989) DuPont Biotech Update 4(2), 16. 18. Kurobe, M., Kono, M., Yoshida, N., and Hayashi, K. (1988) Clin. Chim.Actu 178,205-214. 19. Conway de Macario, E., Macario, A. J. L., and Jovell, R. J. (1983) J. Zmmunol. Methods 39,39-47. 20. Huber, K., Kirchheimer, J., and Binder, B. R. (1984) J. Lab. Clin. Med. 103,684-694.
ANALYTICALBIOCHEMISTRY
(1990)
Ultrasensitive Enzymatic Radioimmunoassay Fusion Protein of Protein A and Neomycin Phosphotransferase II in Two-Chamber-Well Microtiter Plates H.-D.
Hunger,’
Chr. Flachmeier,
G. Schmidt,
Academy of Sciences of the GDR, Central Institute Berlin DDR-1115, German Democratic Republic
Received
November
G. Behrendt,
of Molecular
and Ch. Coutelle
Biology, Department
of Human
Molecular
Genetics,
6,1989
A new sensitive method for antigen detection employing a phosphorylation reaction is described using human serum albumin as a model. The antigen is initially bound to the surface of polystyrene microtiter plates and reacted with an antibody (rabbit). A microbiologitally produced bifunctional fusion protein of protein A and neomycin phosphotransferase II (NPT II) serves as a second immunological reagent by virtue of its protein A component. The detection is based on the phosphorylation of an aminoglycoside antibiotic by the NPT II moiety of the fusion protein using [T-~‘P]ATP as a cosubstrate. This reaction is performed in solution and the evaluation is accomplished by dotting aliquots of the reaction mixture onto phosphocellulose paper, washing with water, and autoradiography. Microtiter plates with a specially designed lo-&volume reaction chamber are particularly advantageous for this procedure. The sensitivity of detection is currently 10 fg (1 pg/ml) of antigen. 0 1990 Academic Press. Inc.
Enzyme-linked immunosorbent assays (ELISA)’ (1) are among the most widely used techniques for detection of antigens or antibodies from biological sources. The enzymes peroxidase, alkaline phosphatase, and P-D-galactosidase are used mainly for labeling, in most cases of a second antibody. The enzyme reaction with chro1 To whom correspondence should be addressed. ’ Abbreviations used: NPT II, neomycin phosphotransferase II; HSA, human serum albumin; PBS, 0.14 M NaCl, 2.6 mM KCl, 7 mM Na/K phosphates, pH 7.2; PBST, 0.05% (v/v) Tween 20 in PBS; PBSM, 5% (w/v) emulsion of nonfat dry milk powder in PBS; PBSTM, 5% (w/v) emulsion of nonfat dry milk powder and 0.05% (v/v) Tween 20 in PBS; ELISA, enzyme-linked immunosorbent assay; SIA, slide immunoassay; RIA, radioimmunoassay. 0003.269’7/90
Using a
$3.00
Copyright 0 1990 by Academic Press, Inc. All rights of reproduction in any form reserved.
mogenic or fluorogenic substrates such as o-phenylenediamine, p-nitrophenyl phosphate disodium salt, and pnitrophenyl+?-D-galactopyranoside or 4-methyl-umbelliferyl-P-D-galactoside allows the identification and/or quantitative determination of the reaction products and thereby of the biomolecule by spectrophotometry or fluorometry. The ELISA is usually carried out in microtiter plates and its principle has been extended to small sample volumes down to drops of 10 ~1 by the slide immunoassay (SIA) (2) performed on flat glass slides. However, to obtain extremely high sensitivities the probes or biomolecules themselves must be radioactively labeled, generally with iz51, and are detected by a radioimmunoassay (RIA) (3). In ultrasensitive enzymatic immunoassays (4) the amplification reaction of an enzyme is used to produce a radioactive detection signal. We effectively applied this principle to the phosphorylation reaction of aminoglycoside antibiotics. This is achieved by a protein A-NPT II fusion protein (5). Its protein A moiety acts as a second immunological reagent recognizing the antibody while its NPT II enzyme activity, using [y-32P]ATP as a cosubstrate, phosphorylates neomycin which is detected by autoradiography. This paper describes the successful attempt to combine the advantages of a microtiter plate assay with the small reaction volumes of the SIA. MATERIALS
AND
METHODS
Kanamycin sulfate and neomycin sulfate were obtained from Medexport (USSR); HSA (68 kDa) was from Serva Feinbiochemica (FRG); dithiothreitol was from Boehringer-Mannheim (FRG); [T-~~P]ATP (110 TBq/mmol) was from ZfK Rossendorf (GDR); HS 11 Xray film was from VEB Fotochemisches Kombinant Wolfen (GDR); nonfat dry milk was from VEB Dauermilchwerke Stendal (GDR); rabbit antiserum against 89
90
HUNGER
ET
HSA was from VEB Germed (GDR); thin-layer plastic plates and PEI cellulose F were from Merck (FRG); polystyrene microtiter plates, flat-bottom type, volume per well 0.4 ml, were from VEB Polyplast Halberstadt (GDR); and phosphocellulose paper P81 was from Whatman Chemical Separation, Ltd. (UK). Construction
of Two-Chamber-
Well Microtiter
Plates (6)
A cone-shaped 12-pl-volume reaction chamber was drilled into the center of each well of a flat-bottom polystyrene microtiter plate using a 4.8-mm-diameter drill (see Fig. 1). Construction of the Gene Coding for the Protein II Fusion Protein and Expression in Escherichia coli Cells (7)
of the Antigen
Specially made two-chamber microtiter plates were used. Ten microliters of a dilution series of HSA containing 1 pg, 100 ng, 10 ng, 1 ng, 100 pg, 10 pg, 1 pg, and 0 pg per milliliter of PBS and 1 pug/ml of nonfat dry milk were each pipetted into the reaction chambers of six rows of 9 wells each and allowed to bind for 4 h. The wells were then washed three times for 2 min with PBST (12). For quantitative determination, a dilution series of HSA containing 10 ng, 1 ng, 100 pg, 10 pg, 1 pg, and 0 pg per milliliter PBS and 1 pg/ml of nonfat dry milk was used and the procedure carried out as described, the controls being the same as above. Immunological
Reaction
Blocking of the antigen-loaded plates was performed using 400 ~1 of PBSTM for 30 min. One hundred microliters of rabbit antiserum against HSA (Fig. 2, +) diluted 1:5000 in PBSTM was pipetted into the wells of the first two rows. The wells of the third and fourth rows (Fig. 2, -) were filled only with 100 ~1 of PBSTM and those of the fifth and sixth rows (Fig. 2, CS) with a rabbit control serum (without the specific antibody, diluted 1:lOOO in PBSTM). The antibody reaction was allowed to proceed for 1 h, after which the wells were washed 10 times for 5 min with PBST. The wells were blocked again for 30 min with 400 ~1 of PBSM. One hundred microliters of the E. coli cell lysate containing the protein A-NPT II fusion protein was added to 10 ml of PBSM. One hundred microliters of this emulsion was added to each well and allowed to react for half an hour. Thereafter the wells were washed 5 times for 3 min with PBS. Enzyme
Reaction
The enzyme reaction was carried out in 10 ~1 of enzyme buffer (20 mM Tris-HCl, pH 7.2; 125 mM NH&l; 20 mM MgCl,; 10 mM dithiothreitol) containing [T-~“P]ATP (5 X lo6 cpm/ml) and 1 ~1 of a kanamycin stock solution (50 mg/ml) .
the Evaluation
A pellet of lOlo E. coli cells was resuspended in 1 ml of cold lysis buffer (25 mM Tris-HCl, pH 7.4) and sonicated three times for 5 s (Branson sonifier, output 50 W, range 5). Between sonications the tube was cooled on ice. After centrifugation (Eppendorf centrifuge, 4”C, 5 min) the clear supernatant was pipetted into a second Eppendorf tube. For immediate use the lysate was stored on ice; otherwise, it was frozen at -20°C. Microtiter
Immobilization
A-NPT
The DNA coding for NPT II (EC 2.7.1.95) from a derivative of the plasmid pSV2-neo (8) was inserted into the EcoRI/BamHI sites of the multicloning sequence of the plasmid pRIT2T (9) which contains the sequence for the Fc binding part of the protein A gene. E. coli NFl cells (10) were transformed with the resulting plasmid (pSAL1). Molecular cloning procedures were performed according to Maniatis et al. (11). E. coli cells were grown at 30°C for 4 h to a density of about lo* cells/ml. The cells were cultured for a further 2 h at 42”C, the temperature at which the protein ANPT II fusion protein becomes expressed due to the Ps promoter at the 5’-terminus of this gene. The cells were then harvested by centrifugation (4000 rpm, 10 min). They were resuspended in Tris buffer (25 mM Tris-HCl, pH 8.0; 50 mM glucose; 10 mM EDTA) to a cell density of lOlo cells/ml, and l-ml aliquots were filled into Eppendorf tubes and centrifuged (5 min, 4°C). The supernatant was decanted, and the pellets were frozen and stored at -20°C. They can be stored in this form for at least 1 year without loss of immunological or enzymatic activity. Preparation of E. coli Cell Lysates Containing Protein A-NPT II Fusion Protein
AL.
Plate Assay
All steps (unless otherwise stated) are performed room temperature with gentle shaking.
at
of the Enzyme
Reaction
One mi(a) Dotting on negatively charged matrices. croliter of the reaction mixture of each well was dotted onto P81 paper. After drying for 30 s the paper was washed three times for 5 min with bidistilled water and then twice for 15 min at 70 to 80°C. The paper was then exposed to an X-ray film. (b) Absorbing the reaction mixture tophosphocellulose paper strips. Five microliters of the reaction mixture was absorbed by 30 X 2-mm phosphocellulose paper strips with a tip at the end. The strips were treated as described in (a) and evaluated by autoradiography or by scintillation counting.
ULTRASENSITIVE
i FIG. 1. Two-chamber-well (3) with a volume of 5-15 of a flat-bottom polystyrene
ENZYMATIC
RADIOIMMUNOASSAY
I microtiter plate. The reaction chamber ~1 is drilled into the bottom (4) of a well (2) microtiter plate (1).
USING
A FUSION
PROTEIN
91
Already after 3 h of enzyme reaction at 37°C (Fig. 2A, +) all seven dilutions of HSA are visible on the autoradiogram. The signals of detection for HSA (+) are significantly enhanced after the enzyme reaction is allowed to proceed further, e.g., 16 h at room temperature. The procedure allows a quantitative determination of the antigen in the femtogram range (Fig. 3). For this purpose, larger volumes (5 ~1) of the reaction mixture are absorbed by phosphocellulose paper strips (Fig. 3A). The distance of capillary suction of the liquid is about 2 cm in this case but the [32P]kanamycin phosphate is concentrated on the first 3 mm of the phosphocellulose paper strips. The evaluation can be made both by autoradiography (A) and by Cerenkov counting of the phosphocellulose paper strips. The calibration curve should be obtained by processing standard amounts of the antigen on the-same microtiter plate and plotting the corresponding Cerenkov countings. The shape of the measured curve is determined mainly by the absorption characteristics of the antigen to the polystyrene surface (13,14).
Thin-Layer Chromatography PEI Cellulose
of the Reaction Mixture
on
One microliter of the sample was dotted onto the starting point of the PEI cellulose plates. Chromatography was performed with B-cm plates developed in 0.75 M sodium phosphate buffer, pH 3.5. The evaluation was performed by autoradiography.
An additional way of analyzing the reaction mixture is by its thin-layer chromatographic separation on polyethylene imine cellulose plates (Fig. 4). The application of this method is both a control for the completeness of reaction of [T-~~P]ATP forming [32P]kanamycin phosphate (lane 2) and a control for the stability of the [T-~~P]ATP (lane 1). Lane 3 shows the position of [3”P]orthophosphate for comparison.
RESULTS
Our enzymatic radioimmunoassay is based on the reaction of a bifunctional detection probe, the protein ANPT II fusion protein, and a specially designed twochamber-well polystyrene microtiter plate with a lo-p1 reaction chamber, shown in Fig. 1. The antigen is bound to the small surface of the reaction chamber and reacted with the specific antibody. This antibody is then recognized by the protein A moiety of the bifunctional probe, while its NPT II enzyme activity is used for detection by 32P phosphorylation of the aminoglycoside antibiotic kanamycin. Blocking, washing, and the specific antibody reaction are carried out with excess volumes (up to 400 ~1) in the larger incubation chamber, while the enzyme reaction is again performed in the small volume (10 ~1) of the reaction chamber. Figure 2 demonstrates the sensitivity of detection of this procedure. As shown, we are able to detect reliably 10 fg (1 pg/ml) of HSA (line 7) by dotting l-p1 aliquots of the reaction mixture onto phosphocellulose paper. Both the negative controls without HSA in PBS (line 8) or in 10 ~1 PBS containing 10 ng of nonfat dry milk (line 9) as well as the negative controls using a rabbit antiserum without specific antiHSA antibodies (CS) or omitting the specific antiserum (-) remain clearly below the level of the detection signals with HSA after reaction with rabbit antiserum containing specific HSA antibodies (+).
0
*
l
9 FIG. 2.
Sensitivity of the two-chamber-well microtiter plate assay. (A and B) Autoradiographs of l-al dots of the reaction mixture after enzyme reaction. Exposure time, 16 h, -70°C. (A) Enzyme reaction at room temperature for 3 h; (B) enzyme reaction performed for a further 16 h in the same microtiter plate as that in A. Line 1,10 ng (1 pg/ml); line 2, 1 ng (100 rig/ml); line 3, 100 pg (10 rig/ml); line 4, 10 pg (1 ng/ ml); line 5, 1 pg (100 pg/ml); line 6, 100 fg (10 pg/ml); line 7, 10 fg (1 pg/ml) of HSA; line 8,O pg of HSA in PBS; line 9,0 pg of HSA and 10 ng (1 pg/ml) of nonfat dry milk in PBS. Lanes +, rabbit antiserum containing specific antibodies against HSA, lanes -, without antiserum; lanes CS, rabbit antiserum without specific antibodies against HSA. The duplicates were obtained from two different rows of the microtiter plate.
92
HUNGER
B I\
2000-
1500E 8
.i 1000-l
\
5oo;
k----T
\
0.1
0.01
pg HSA FIG. 3.
Quantitative evaluation of the microtiter plate assay. 5 ~1 of the reaction mixture is sucked up by phosphocellulose paper strips. (A) Autoradiograph of the strips (exposure time 10 h at room temperature). Lane 1, 100 pg (10 rig/ml); lane 2, 10 pg (1 rig/ml); lane 3, 1 pg (100 pg/ml); lane 4,100 fg (10 pg/ml); lane 5,10 fg (1 pg/ml) of HSA; lane 6, PBS without HSA; lane 7, 10 ng (1 @g/ml) of nonfat dry milk powder in PBS without HSA. (B) Quantitative evaluation of the strips by plotting the Cerenkov countings. The individual points are means of duplicates obtained from two different rows of the microtiter plate. The values of lanes 6 and 7 are 145 and 122 cpm, respectively.
ET
AL.
structed similarly. Further, on the basis of its properties, (i) high specific activity and turnover rate, (ii) good stability at room temperature and higher temperatures, (iii) ready availability because of its ease of preparation, (iv) low molecular weight (30 kDa), and (v) renaturation after treatment with detergents, the enzyme NPT II is suitable for chemical coupling to other molecular probes. In this paper we report on the utilization of the phosphorylation reaction with the substrate kanamycin and the cosubstrate [T-~‘P]ATP in solution. We have also developed a solid-phase assay (“Contact-Copy” method) using this principle (15). At present, the sensitivity of the procedure demonstrated here is limited only by the background arising from nonspecific binding of the first antibody to the material of the microtiter plates. An advantage of our model system is the antigen fixation out of a volume of 10 ,ul. Any application should be checked if a reduction of the sample volume before antigen fixation is possible. For comparison purposes, all sensitivities were recalculated to a volume of 1 ml (given in parentheses in the legend to Fig. 3 and below). The method discussed here exhibits an extraordinary sensitivity of detection (cf. Figs. 2 and 3); i.e., the limit of detection is currently 10 fg (0.15 amol) of HSA (1 pg/ml). In experiments for comparison using a microtiter plate assay and peroxidase second antibody conjugates (data not shown) we could detect 10 pg of HSA (1 rig/ml). The very different setup and performance of the various ELISA and RIA methods as well as the great variety of antigens investigated make a direct comparison of sensitivities somewhat difficult. Methods employing mi-
DISCUSSION
This work was aimed at demonstrating the principle and sensitivity of detection of our enzymatic radioimmunoassay. Using HSA as a test model, we report here the application of the phosphorylation reaction as a general method for the detection of antigens. An advantage of HSA is that it is available as an electrophoretically pure, accurately weighed antigen, making it possible to produce solutions with exact protein concentrations. Our protein A-NPT II fusion protein was constructed by genetic engineering. It may be easily and reproducibly produced microbiologically, and the assay may be performed using nonpurified E. coli lysates. It offers the opportunity for general use of the enzymatic phosphorylation reaction for detecting antigens because protein A binds to the Fc region of IgG from many species (e.g., human, rabbit, guinea pig) and it does not interfere with the antigen-antibody reaction. Of course, the application of the protein A-NPT II fusion protein is limited to problems of analysis in which protein A can be used. However, other fusion proteins with NPT II and different molecular probes such as protein G may be con-
1
2
3
FIG. 4. Analysis of the NPT II enzyme reaction mixture by thinlayer chromatography on PEI cellulose, autoradiograph of the PEI cellulose after 30 min. Lane 1, enzyme buffer containing [y-32P]ATP before the enzyme reaction; lane 2, reaction mixture after nearly complete reaction of [y-32P]ATP to [32P]kanamycin phosphate; lane 3, [32P]phosphoric acid: 1 pl(5000 cpm) of each solution was dotted onto PEI cellulose.
ULTRASENSITIVE
ENZYMATIC
RADIOIMMUNOASSAY
crotiter plates for the ELISA technique (e.g., urease, peroxidase) reach a sensitivity of detection of 50 pg of antigen (reaction mixture of 100 ~1, i.e., 500 pg/ml) (16). The most sensitive ELBA procedure detects 6 pg of HIV p24 core antigene in a volume of 200 ~1 (30 pg/ml) of tissue culture supernatant, plasma, or serum (17). One picogram of human pancreatic secretory trypsin inhibitor (18) has been detected in lOO-~1 aliquots of the sample (10 pg/ml). The SIA technique has been used to detect 50 pg of purified immunoglobin in 10 ~1 (2) and even 5 pg of this protein in 5 ~1 of solution (1 rig/ml) (19). Competitive radioimmunoassays, e.g., for human urokinase, employing ‘251-labeledtracers reach a limit of detection of 10 pg of antigen/ml sample (20). The protein A part of our fusion protein acts as a second immunological reagent, thereby allowing detection of a great variety of antibodies. The design of the microtiter plate wells combines a very small reaction chamber (10 ~1) with the relatively large volume of the incubation chamber (400 ~1). This is particularly advantageous for reduction of the background in the detection reaction because it reduces the amount of nonspecifically bound antibodies. This method certainly may be used for the determination of low concentrations of antibodies from biological samples. Furthermore, the microtiter plate may be constructed with several reaction chambers (6) in one well which may carry different immobilized antigens or antibodies. This would make a simultaneous testing of different biomolecules from the same sample possible. The utilization of 32P as the isotope for detection is especially advantageous for increasing the sensitivity of detection. Our experiments have not come close to exhausting the possibilities of an enhancement of the signals. We believe that such approaches together with efforts to decrease the nonspecific background will be the basis for a further increase in sensitivity. One microliter of the reaction mixture dotted onto phosphocellulose paper (Fig. 2) or 5 ~1absorbed by phosphocellulose paper strips (Fig. 3) is used here for quantitative evaluation of the lo-p1 reaction mixture. The evaluation by dotting of aliquots of the reaction mixture onto and the absorption by phosphocellulose represent very good opportunities for automation of the detection procedure, where large numbers of test samples are used in routine analysis. The application of the method is aimed mainly at the detection of biomolecules which are contained in samples in very low amounts. Since an additional separation step of the unreacted cosubstrate from the reacted substrate is necessary, the technical processing is more complex than that of the state-of-the-art ELISA. Our system exhibits extremely favorable preconditions compared to other enzymatic radioimmunoassays (4), in
USING
A FUSION
93
PROTEIN
which, e.g., [3H]adenosine must be eluted from DEAESephadex columns and scintillation cocktails must be added for evaluation: selective binding of the reacted substrate to the matrix phosphocellulose, simple washing of the matrix with water as the separation step, and detection by Cerenkov counting without scintillation cocktails. The biohazard associated with 32P is less than that with RIA which generally uses 1251.This hazard may be decreased by use of suitable means of protection, especially in the case of automation. We are convinced that the enzymatic radioimmunoassay demonstrated here is currently one of the most sensitive methods for detection of antigens and its application may be extended to a variety of fields. REFERENCES 1. Engvall, E., and Perhnan, P. (1971) Immunochemistry 8,871-874. 2. Conway de Macario, E., Jovell, R. J., and Macario, A. J. L. (1985) BioTechniques 3,138-145. 3. Mayer, R. J., and Walker, J. H. (1987) Immunochemical Methods in Cell and Molecular Biology, pp. 49-58, Academic Press, Boca Raton, FL. 4. Harris, C. C., Yolken, R. H., Krokan, H., and Hsu, I. C. (1979) Proc. Nutl. Acad. Sci. USA 76,5336-5339. 5. Merkel, D., Schmidt, G., Flachmeier, C., Behrendt, G., Coutelle, Ch., and Hunger, H.-D. (1989) J. Biochem. Biophys. Methods 18,
277-286. 6. Hunger, H. D., Schmidt, 7. 8. 9.
(1989) Gefaesskombination DD-WP Application WP Hunger, H.-D., Behrendt, lecular probes, processes Pat. Application 304.934; Southern, P. J., and Berg, 341. Nilsson, B., Abrahmsen, 1075-1080.
G., Behrendt, G., and Flachmeier, Chr. zum Nachweis von Biomolekuelen, G 01 N/330 196 8. G., and Schmidt, G. (1989) Labeled mofor their preparation, and their use, Eur. Cl. G 01 N 33/58. P. (1982) J. Mol. Appl. Genet. 1, 327L., and Uhlen,
M.
(1985)
EMBO
J. 4,
10. Bernard, H.-U., Remaut, E., Hershfield, M. V., Das, H. K., Helinski, D. R., Yanofsky, C., and Franklin, N. (1979) Gene 5,59-76. 11. Maniatis, T., Fritsch, E. F., and Sambrook, J. (1982) Molecular Cloning. A Laboratory Manual, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY. 12. Mohammad, K., and Esen, A. (1989) J. Zmmunol. Methods 117, 141-145. 13. Walsh, J., and Gosling, J. P. (1986) Anal. Biochem. 158,413-423. 14. Cantarero, L. A., Butler, J. E., and Osborne, J. W. (1980) Anal. Biochem. 105,375-382. 15. Hunger, H.-D., Schmidt, G., Flachmeier, Chr., Behrendt, G., and Coutelle, Ch. (1989) Anal. Biochem., in press. 16. Gerber, M., and Sarkar, S. (1988) J. Plant Dk. Protect. 95, 544-
550. 17. Improved HIV p24 core antigen ELISA has greater sensitivity, known specificity (1989) DuPont Biotech Update 4(2), 16. 18. Kurobe, M., Kono, M., Yoshida, N., and Hayashi, K. (1988) Clin. Chim.Actu 178,205-214. 19. Conway de Macario, E., Macario, A. J. L., and Jovell, R. J. (1983) J. Zmmunol. Methods 39,39-47. 20. Huber, K., Kirchheimer, J., and Binder, B. R. (1984) J. Lab. Clin. Med. 103,684-694.
