Detection of Clostridium botulinum neurotoxin type Ausing immuno-PCR

H.C. Wu, Y.L. Huang, S.C. Lai, Y.Y. Huang and M.F. ShaioInstitute of Preventive Medicine, National Defense Medical Center, Taipei, Taiwan, Republic of China

2000/37: received 29 January 2001 and accepted 12 February 2001

H.C . WU, Y .L . HUANG, S .C. LA I , Y .Y . HUANG AND M.F. SHAIO . 2001.

1 Aims: An immuno-polymerase chain reaction (immuno-PCR) has been developed for the

sensitive detection of antigens, which greatly extends the detection limits of immunoassays. In

the current study, the method was applied to the detection of Clostridium botulinum neurotoxin

type A (BTx-A).

Methods and Results: Anti-BTx-A antibody-DNA conjugates were synthesized using a

heterobifunctional cross-linker reagent to covalently link the reporter DNA and the antibodies.

The antibody-DNA conjugates with antigens were ampli®ed by PCR, and dose-dependent

relationships for each analyte were demonstrated. Detection limits of immuno-PCR for BTx-A

(3á33 ´ 10)17 mol) exceeded the conventional enzyme-linked immunosorbent assay

(3á33 ´ 10±14 mol) by a 1000-fold enhancement in detection sensitivity.

Conclusions: Detection of BTx-A antigens by immuno-PCR demonstrated 100% sensitivity

and 100% speci®city in 100-fold magnitude below the detection limit of ELISA.

Signi®cance and Impact of the Study: It is concluded that the immuno-PCR method could

be used to detect a very low level of BTx-A for clinical diagnosis.

INTRODUCTION

Botulinum neurotoxin type A (BTx-A) produced by the

anaerobic bacterium, Clostridium botulinum, is one of the most

potent toxins known to humans (Gill 1982; Franz et al.1997). The neurotoxin acts as a zinc-dependent endoprotease

to cleave proteins that are essential for the release of the

neurotransmitter, acetylcholine, and lead to paralysis (Simp-

son 1986). Although there are several routes through which

the toxin can enter the body, most cases involve ingestion of

toxin or ingestion of bacteria that produce the toxin.

Recently, botulinum neurotoxins have become an extremely

useful therapeutic drug in the treatment of segmental

movement disorders, such as adult onset spasmodic torti-

collis, spasmodic dysphonia, oral mandibular dystonia and

blepharospasm (Tsui 1996; Jankovic 1998; Lopez and

Morales 1998). Consequently, sensitive and speci®c detec-

tion of BTx-A is very important in the investigation of

suspected food-borne and therapeutic formulations.

Immuno-PCR, using the speci®city of an antibody, and a

reporter DNA molecule ampli®ed by PCR, enabled the

development of a sensitive assay to detect BTx-A antigen.

This method combines the ampli®cation power of PCR and a

method, similar to enzyme-linked immunosorbent assay

(ELISA), which detects an antigen-antibody reaction; how-

ever, instead of an enzyme being conjugated to an antibody, a

reporter DNA was used which could be ampli®ed by PCR.

Several immuno-PCR methods have been developed for

the detection of antigens. In comparison with ELISA, an

enhancement in detection sensitivity, from 102 (Joerger et al.1995) to 105-fold (Sano et al. 1992), was obtained by

immuno-PCR. This methodology had been claimed to have

the potential to detect very low concentrations of antigens,

such as tumour markers (Zhang et al. 1998; Ren et al. 2000),

cytokines (Sanna et al. 1995; Saito et al. 1999), hormones

(Hendrickson et al. 1995; Joerger et al. 1995) and viral

antigens (Maia et al. 1995; Mweene et al. 1996).

In the present study, reporter DNA-antibody conjugates

were prepared and immuno-PCR assays developed for

botulinum neurotoxin. The results demonstrate that double-

stranded reporter DNA can be covalently linked to anti-

BTx-A antibodies through amine and sulphhydryl groups

on the reporter DNA and antibodies. The DNA-antibody

conjugates can be used to detect BTx-A antigens directly,

without the addition of biotinylated reagents, bindingCorrespondence to: Dr Han-Chung Wu, Institute of Preventive Medicine,

National Defense Medical Center, PO Box 90048±700, San-Hsia, Taiwan,

R.O.C. (e-mail: hancw@pchome.com.tw).

ã 2001 The Society for Applied Microbiology

Letters in Applied Microbiology 2001, 32, 321±325

proteins and numerous wash steps. The immuno-PCR

assays described here may prove valuable for antigen

detection in clinical diagnosis.

MATERIALS AND METHODS

Reagents

The murine monoclonal antibodies, anti-BTx-A (BT57±1),

that were used to covalently couple the double-stranded

DNA (dsDNA) to form the reporter conjugates, were

developed in this laboratory (unpublished data). BTx-A

(90% purity by SDS-PAGE and quantitative densitometric

analysis) was a gift from H.-Y. Chao. Sulphosuccinimidyl

4-(maleimidomethyl) cyclohexane-1-carboxylate (sulpho-

SMCC) was purchased from Pierce Chemical Co. PCR

reagents and Taq DNA polymerase (AmpliTaqÔ) were

obtained from Perkin Elmer Corp. The b-cyanoethyl-

phosphoramidite amino-modifying reagent (Aminolink 2Ô)

was purchased from Applied Biosystems.

Primers and synthesis of amino-modi®edreporter DNA

Double-stranded reporter DNA for antibody conjugates was

generated by PCR from a DNA template encoding a

Luciferase gene (Promega, Madison, WI, USA). The

primers used for the ampli®cation reaction were forward

primer 5¢amino-modi®ed (N-TFA-C6 Aminomodi®er,

CruaChem, Taipei, TW) Amin498: 5¢GTTCGTCACATC-

TCATCTAC-3¢ and reverse primer 1061(±): 5¢TCGGG-

TGTAATCAGAATAGC-3¢. The PCR reaction was

carried out in a 96-well microtitre plate under the following

conditions: 94°C for 30 s, 54°C for 30 s, 72°C for 30 s for 35

cycles. The PCR reaction volumes from each well were

pooled, and the amino-modi®ed reporter DNA was puri®ed

using Qiagen tip-500 and Sephacryl S-400 chromatography.

Synthesis of the DNA-antibody conjugates

The amino-modi®ed reporter DNA primary antibody

conjugates were synthesized as described previously

(Hendrickson et al. 1995; Joerger et al. 1995). Brie¯y, the

monoclonal antibody anti-BTx-A was reduced with

2-mercaptoethylamine-HCl (2-MEA), and the amino-mo-

di®ed reporter DNA was activated with sulpho-SMCC. The

reduced antibody and activated reporter DNA were mixed.

The cross-linking reaction was allowed to proceed for 2 h at

room temperature, in the dark with gentle shaking, after

which the product was stored at 4 °C. The ®nal conjugate

solution was stored at 4°C in 100 mmol l)1 sodium phos-

phate, 1 mmol l)1 EDTA and 0á1 mg ml)1 acetylated BSA,

pH 7á0.

Immuno-PCR assay

The test analyte was immobilized on a 96-well, U-bottom,

polystyrene microtitre plate (Falcon 3911, Becton Disknson,

Oxnard, CA, USA). Negative control wells received PBS/

Tween (PBS containing 0á1% Tween-20) buffer without

antigen. The microplate was washed three times with

washing buffer, PBSTE (PBS containing 0á1% Tween-20

and 0á1 mmol l)1 EDTA). Non-adsorbed sites in the micro-

titre wells were blocked with blocking buffer (Boehringer

Mannheim). After washing with PBSTE, 50 ll of 1:100 000

diluted reporter DNA-antibody conjugates were added to

the test wells and incubated at room temperature for 1 h.

Conjugate solutions were removed and the wells were

washed with PBSTE seven times and once with PBS as

described.

The microtitre plate was inserted into the 96-well sample

block of a Perkin Elmer Gene Amp thermocycler (Norwalk,

CT, USA). Ampli®cation of the reporter DNA and

separation of the PCR products was according to the

method previously published, with some modi®cations

(Wu and Lee 1997). Brie¯y, the ampli®cation was performed

in 30 cycles using the following thermal cycling conditions:

94°C for 30 s, 54°C for 45 s and 72°C for 45 s. The ®nal

chain extension was made at 72°C for 5 min. The ethidium

bromide-stained DNA bands were visualized with a u.v.

transilluminator and recorded on a Digital Imaging System

IS-1000 (Alpha Innotech Corp., San Leandro, CA, USA).

The digitized image was further analysed to measure the

intensity of the PCR product bands.

Enzyme-linked immunosorbent assay (ELISA)

The ELISA assays for BTx-A were carried out with the

same steps as described for immuno-PCR except that the

reporter DNA-primary antibody conjugates were replaced

by 1 : 1000 diluted anti-BTx-A and horseradish peroxidase-

conjugated anti-mouse antibody (Jackson ImmunoResearch

Laboratories, Inc., West Grove, PA, USA2 ). After washing, a

horseradish peroxidase-mediated colour reaction was carried

out utilizing hydrogen peroxide and the chromogen,

o-phenylenediamine3 dihydrochloride. The optical density

(O.D.) of the samples was determined at 490 nm in an

ELISA reader (THERMOmaxÔ, Molecular Device Cor-

poration, Sunnyvale, CA, USA).

RESULTS AND DISCUSSION

Immuno-PCR assays were ®rst shown to detect BTx-A

using microtitre plates coated with BTx-A antigen. The

BTx-A reporter antibody was conjugated covalently to

reporter DNA and the assay was performed as described

(Fig. 1a). Amino-modi®ed, double-stranded DNA for

322 H.C. WU ET AL .

ã 2001 The Society for Applied Microbiology, Letters in Applied Microbiology, 32, 321±325

antibody conjugates was generated by PCR from a DNA

template encoding a Luciferase gene. The ampli®ed reporter

DNA was further puri®ed, using Qiagen tip-500 and

Sephacryl S-400 chromatography to remove primers, and

a 563 bp reporter DNA was obtained (Fig. 1b, lane 2). For

these studies, the reporter DNA to BTx-A antibodies were

conjugated using a heterobifunctional cross-linker sulpho-

SMCC. Under these conditions, as much as 40% of the ds

DNA in the reaction became cross-linked to BTx-A

antibodies (Fig. 1b, lanes 3 and 4). The conjugation

procedure used takes advantage of the presence of certain

disulphide bonds in the heavy chain of IgG that could be

reduced with mild reducing agents, without the loss of

antibody activity (Hendrickson et al. 1995; Joerger et al.1995). Potential loss of test sensitivity by steric hindrance,

caused by the attachment of relatively large DNA molecules

near the antigen-binding site, could also be reduced.

Immuno-PCR is a very sensitive method and the elim-

ination of non-speci®c reaction is essential. The non-speci®c

reaction is generated when a higher concentration of

reporter DNA-antibody conjugates is used. Therefore, the

optimal concentration of reporter DNA-antibody conjugates

for the immuno-PCR method was determined. The data

clearly demonstrate that the non-speci®c PCR signal was

generated when the concentration of the reporter DNA-

antibody conjugates was higher than 1/16 000 dilution

(Fig. 1c, lanes 1 and 2). If the concentration of the reporter

DNA-antibody conjugates was diluted to 1/64 000 and

1/256 000, the PCR reaction signal was only generated in

those samples containing 5 ng well)1 BTx-A, and no non-

speci®c signals were generated in the sample without

BTx-A. The concentration of the reporter DNA-antibody

conjugates that were used in subsequent immuno-PCR

assays was 1/100 000. These results also indicate that the

synthesized DNA-antibody conjugate formed an effective

complex in response to BTx-A.

The sensitivity of the BTx-A immuno-PCR assay was

compared with a BTx-A ELISA assay (Fig. 2). In both assay

formats, dose-dependent relationships were observed over a

range of BTx-A concentration. From the ELISA data

(Fig. 2b), the cut-off value was calculated as the mean of

control ascites (NA) (0á075) + 3 standard deviations, which

was set as 0á225. Therefore, the detection limit for BTx-A

using the ELISA assay was set as 5 ng. A speci®c 375 bp

product was observed in the lanes that contained 5 ng,

0á5 ng, 0á05 ng, 5 pg and 0á5 pg BTx-A (Fig. 2a, lanes 1±5).

However, the 375 bp product was not visible in the absence

of BTx-A or template DNA. This result indicates that the

synthesized DNA-antibody conjugate was speci®cally bound

to the BTx-A antigens. The detection limit was 5 ng

(3á33 ´ 10)14 mol) for the ELISA format (Fig. 2b) and 5 pg

(3á33 ´ 10)17 mol) for the immuno-PCR assay (Fig. 2a).

Fig. 1 (a) The principle of the immuno-PCR method employed in this study. Antigens and reporter DNA-antibody conjugates were stepwise-

immobilized as in ELISA assays. The amount of antigen present in the sample was then quanti®ed by PCR ampli®cation of the conjugated reporter

DNA. (b) Analysis of puri®ed reporter DNA and DNA-antibody conjugates. Lane 1: 100 bp DNA ladder marker; lane 2: 50 ng 563 bp reporter

DNA; lane 3: 0á5 ll DNA-antibody complex reaction mixture; lane 4: 1 ll DNA-antibody complex reaction mixture. The reporter DNA-antibody

conjugates were retarded in their migration by covalently coupled antibody molecules. (c) Determination of the optimal concentration of reporter

DNA-antibody conjugates for the detection of BTx-A antigen by immuno-PCR. The microtitre plate was immobilized with 5 ng BTx-A antigen in

lanes 5±8, and without BTx-A antigen in lanes 1±4. The working concentration of the reporter DNA-antibody conjugates was 1/4000 dilution in

lanes 1 and 5, 1/16 000 dilution in lanes 2 and 6, 1/64 000 dilution in lanes 3 and 7, and 1/256 000 dilution in lanes 4 and 8. Lane M: 100 bp DNA

ladder marker; lane +: PCR positive control (10±5 ng reporter DNA); lane ±: PCR negative control (no reporter DNA)

DETECTION OF ANTIGENS USING IMMUNO-PCR 323

ã 2001 The Society for Applied Microbiology, Letters in Applied Microbiology, 32, 321±325

Detection of BTx-A antigens by immuno-PCR exhibited a

three orders of magnitude increase in sensitivity over

comparable ELISA assays. Figure 2(c) shows the results of

the intensity of PCR products (arbitrary units) in Fig. 2(a).

The amount of PCR products increased with increasing

amount of BTx-A, from 0á0005 to 5 ng, and revealed a

linear relationship between 0á005 and 0á5 ng of BTx-A.

To evaluate the ef®cacy of detection of BTx-A antigens by

immuno-PCR, 12 samples were prepared which included six

samples containing 0á05 ng (100-fold magnitude below the

detection limit of ELISA) of BTx-A and six samples

without BTx-A, as controls for the immuno-PCR assay. The

12 samples were detected blindly by two researchers (Fig. 3,

samples 1±6 and samples 7)12) using immuno-PCR. The

results revealed that only those samples containing 0á05 ng

BTx-A antigen could be detected by immuno-PCR (Fig. 3,

lanes 2, 4, 5, 8, 9, and 12), and those samples without BTx-A

antigen showed no any positive signals (Fig. 3, lanes 1, 3, 6,

7, 10, and 11). Detection of BTx-A antigens by immuno-

PCR demonstrated 100% sensitivity and 100% speci®city in

unknown sample solutions (Fig. 3).

The major difference in the immuno-PCR assay described

here was the use of covalent attachment of the DNA label to

reporter antibody. The DNA-labelled reporter reagents

used in the earlier reports were assembled by non-covalent

attachment such as biotin±avidin or biotin±streptavidin

(Sano et al. 1992; Sanna et al. 1995; Zhang et al. 1998; Saito

et al. 1999; Ren et al. 2000). These methods involve

numerous steps for the addition of reporter reagents, and

need more than 20 washing steps to remove excess reagents.

The reagent additions and washing steps increase complex-

ity and add time to the immuno-PCR procedure. By

contrast, the reporter DNA covalently conjugated to the

primary antibody only required washing after the addition of

sample antigen and after the addition of the DNA-antibody

conjugates. Furthermore, the use of DNA-conjugated

antibodies could detect more than two analytes at the same

Fig. 2 Comparison of the detection limit of immuno-PCR and ELISA

to BTx-A antigen. (a) Detection of BTx-A antigen by immuno-PCR

and analysis of PCR products by agarose gel electrophoresis. Lane M:

100 bp DNA ladder marker; lanes 1±5: PCR products of samples from

assay wells with 10-fold decreasing amounts of BTx-A antigen: lane 1,

5 ng well)1; lane 2, 0á5 ng well)1; lane 3, 0á05 ng well)1; lane 4,

5 pg well)1; lane 5, 0á5 pg well)1; and lane 6, 0 pg well)1. Lane +:

PCR positive control (10)5 ng reporter DNA); lane ±: PCR negative

control (no reporter DNA). (b) Detection of BTx-A antigen by ELISA

assay. Wells contained twofold serial dilutions of the BTx-A from

50 ng well)1 to 0á02 ng well)1. Anti-BTx-A: monoclonal antibody

against BTx-A (h) antigen; NA (j): control ascites. (c) Quanti®cation

of the PCR product. The band intensity (arbitrary units) in Fig. 2(a)

was plotted against amount of BTx-A

Fig. 3 Detection of 0á05 ng BTx-A antigen by immuno-PCR in blind

samples and analysis of PCR products by agarose gel electrophoresis.

Lane M: 100 bp DNA ladder marker; lanes 2, 4, 5, 8, 9 and 12: samples

with 0á05 ng BTx-A antigen; lanes 1, 3, 6, 7, 10, and 11: samples

without BTx-A antigen; lane +: PCR positive control; lane ±: PCR

negative control

324 H.C. WU ET AL .

ã 2001 The Society for Applied Microbiology, Letters in Applied Microbiology, 32, 321±325

time by the use of differential size or sequence of reporter

DNA (Hendrickson et al. 1995).

In conclusion, the method described here demonstrates

that immuno-PCR technology greatly extends the sensitivity

of immunoassays. This hybrid technology exhibited analyte

detection from 100- to 1000-fold better than the ELISA

method performed with the same antibodies. Immuno-PCR

technology, in principle, provides the basis for a new

generation of sensitive immunoassays, and may be useful in

clinicopathological assays as well as detection of low level

antigens.

ACKNOWLEDGEMENTS

The authors thank H.-Y. Chao for his kind gift of

Clostridium botulinum toxin Type A. They also thank Dr

L.-K. Chen for valuable discussions and advice. This

research was supported in part by research Grant NSC

89±2320-B-016-027 from the National Science Council,

R.O.C. to HCW, and a Grant 89-0303 from the Institute of

Preventive Medicine, National Defense Medical Center,

Taipei, R.O.C. to HCW.

REFERENCES

Franz, D.R., Jahrling, P.B., Friedlander, A.M. et al. (1997) Clinical

recognition and management of patients exposed to biological warfare

agents. Journal of the American Medical Association 278, 399±411.

Gill, D.M. (1982) Bacterial toxins: a table of lethal amounts.

Microbiological Review (March) 86±94.

Hendrickson, E.R., Truby, T.M.H., Joerger, R.D., Majarian, W.R.

and Ebersole, R.C. (1995) High sensitivity multianalyte immuno-

assay using covalent DNA-labeled antibodies and polymerase chain

reaction. Nucleic Acids Research 23, 522±529.

Jankovic, J. (1998) Medical therapy and botulinum toxin in dystonia.

Advances in Neurology 78, 169±183.

Joerger, R.D., Truby, T.M., Hendrickson, E.R., Young, R.M. and

Ebersole, R.C. (1995) Analyte detection with DNA-labeled anti-

bodies and polymerase chain reaction. Clinical Chemistry 41,

1371±1377.

Lopez del Val, L.J. and Morales Asin, F. (1998) New applications of

botulinum toxin. Neurologia 13, 388±394.

Maia, M., Takahashi, H., Adler, K., Garlick, R.K. and Wands, J.R.

(1995) Development of a two-site immuno-PCR assay for hepatitis B

surface antigen. Journal of Virological Methods 52, 273±286.

Mweene, A.S., Ito, T., Okazaki, K., Ono, E., Shimizu, Y. and Kida, H.

(1996) Development of immuno-PCR for diagnosis of bovine

herpesvirus 1 infection. Journal of Clinical Microbiology 34, 748±750.

Ren, J., Chen, Z., Juan, S.J., Yong, X.Y., Pan, B.R. and Fan, D.M.

(2000) Detection of circulating gastric carcinoma-associated antigen

MG7-Ag in human sera using an established single determinant

immuno-polymerase chain reaction technique. Cancer 88, 280±285.

Saito, K., Kobayashi, D., Sasaki, M. et al. (1999) Detection of human

serum tumor necrosis factor-alpha in healthy donors, using a highly

sensitive immuno-PCR assay. Clinical Chemistry 45, 665±669.

Sanna, P.P., Weiss, F., Samson, M.E., Bloom, F.E. and Pich, E.M.

(1995) Rapid induction of tumor necrosis factor alpha in the

cerebrospinal ¯uid after intracerebroventricular injection of lipo-

polysaccharide revealed by a sensitive capture immuno-PCR assay.

Proceedings of the National Academy of Sciences USA 92, 272±275.

Sano, T., Smith, C.L. and Cantor, C.R. (1992) Immuno-PCR: very

sensitive antigen detection by means of speci®c antibody-DNA

conjugates. Science 258, 120±122.

Simpson, L.L. (1986) Molecular pharmacology of botulinum toxin and

tetanus toxin. Annual Review of Pharmacology and Toxicology 26,

427±453.

Tsui, J.K. (1996) Botulinum toxin as a therapeutic agent. Pharmacology

and Therapeutics 72, 13±24.

Wu, H.C. and Lee, E.H.Y. (1997) Identi®cation of a rat brain gene

associated with aging by PCR differential display method. Journal of

Molecular Neuroscience 8, 13±18.

Zhang, Z., Irie, R.F., Chi, D.D. and Hoon, D.S. (1998) Cellular

immuno-PCR. Detection of a carbohydrate tumor marker. American

Journal of Pathology 152, 1427±1432.

DETECTION OF ANTIGENS USING IMMUNO-PCR 325

ã 2001 The Society for Applied Microbiology, Letters in Applied Microbiology, 32, 321±325