Improved microplate immunoenzymatic assay of PCR products for rapid detection of Mycoplasma pneumoniae

Molecular and CellularProbes(1995) 9, 25-32

Improved microplate immunoenzymatic assay of PCR products for rapid detection of Mycoplasmapneumoniae Antoine

Vekris, 1 F r e d e r i c B a u d u e r 7 S o p h i e M a i l l e t , z C h r i s t i a n e Jacques B o n n e t 2.

B~b~ar 1 and

1Laboratoire de Bact~riologie, Universit~ de Bordeaux II, 146, rue L~o Saignat, 33076 Bordeaux cedex, and 21BGC-CNRS, 1, rue Camille Saint SaEns, 33077 Bordeaux cedex, France (Received 4 May 1994, Accepted 20 September 1994) We developed a microtitre hybridization assay for the detection of polymerase chain reaction (PCR) amplified sequences. For this, cloned Mycoplasma pneurnoniae DNA containing a sequence complementary to the PCR products is first covalently bound to microtitre wells. These coated microplates can be stored for several months. Then, an aliquot of the PCR product, labelled with digoxigenin-dUTP during its synthesis is hybridized to the immobilized DNA. The use of a rapid hybridization buffer makes this step very short (5 min). Finally, the hybridization signal is detected by an anti-digoxigenin antibody conjugated with alkaline phosphatase. Compared to Southern or other microplate hybridization techniques, this method is cheaper, involved fewer steps and allows easy handling of a large number of samples. This method was used for detection of M. pneumoniae in a series of clinical specimens.

KEYWORDS: Mycoplasma pneumoniae, detection, PCR, microplate, non-isotopic labelling. INTRODUCTION Polymerase chain reaction (PCR) is a powerful tool for the detection and identification of infectious agents. Generally, a two-step method is used: first, amplification and second, identification of the amplified products. The most commonly used identification procedure is gel electrophoresis. However, when dealing with clinical specimens non-specific amplification is a common occurrence, a further analysis involving a hybridization step must be done to assess the specificity of the amplified products. This can be achieved by Southern or dot-blot hybridization. However, these techniques are not convenient for clinical laboratories because they are time-consuming and difficult to apply to the processing of a large number of samples. As the PCR technology is currently finding a wide application in routine medical diagnosis, procedures for mass screening must be

developed. Several techniques involving microplates and taking advantage of the specificity of hybridization have been proposed. A first format is based on the immobilization .of a capture nucleic acid molecule: oligonucleotide I or DNA 2'3 on a solid support and the hybridization of the PCR product labelled during its synthesis. A second format consists in immobilizing the PCR product and hybridizing it with a specific probe. 4 We describe here a fast, sensitive and very simple hybridization assay. Plasmid DNA containing a specific sequence, covalently bound to microplate wells is used to capture PCR products labelled during their synthesis by incorporation of Dig-dUMP. We used it for detecting Mycoplasma pneumoniae sequences in clinical samples. After completion of this work, Gibellini et aLs

* Author to whom correspondenceshould be addressedat: Laboratoired'lmmunologie et de Parasitologie,Universit~de Bordeaux II, Bat lb, 146, rue L~o Saignat, 33076 Bordeauxcedex, France. 0890-8508195/010025+ 07 $08.0010

25

© 1995 Academic Press Limited

26

A. Vekris et

published the description of a similar assay for detection of parvovirus. But our assay has a few advantages which will be discussed.

MATERIALS AND METHODS

al.

performed by incorporation of Dig-dUMP at the 3'-OH and using terminal deoxynucleotide transferase (Boehringer Mannheim) as indicated by the manufacturer. The reaction time was limited to 10 min in order to incorporate only about 15 digoxigenin residues per molecule as estimated by sizing the labelled products on polyacrylamide gel.

PCR target sequence and primers

The PCR target was a randomly selected, cloned sequence shown to be specific to M. pneumoniaeand not present in several bacterial species (Mycoplasma

genitalium, Mycoplasma hominis, Acholeplasma ladlawii, Mycoplasma salivarium, Mycoplasma orale, Mycoplasma buccale, Escherichia coil, Staphylococcus aureus, Haemophilus influenzae, Streptococcus faecalis, Peptococcus anaerobius, Pseudomonas aeruginosa, Klebsiella pneumoniae and several species of Lactobacillus) and human DNA. The primers used are MPS-I: GAAGCTTATGGTACAGGTTGG and MP5-2: ATTACCATCCFIGTTGTAAGG.

DNA amplification and labelling

The source of DNA was either 10 ng of an EcoR1 linearized plasmid (pA021) containing a M. pneumoniae specific genomic sequence6or a 10-p.I aliquot of broncho-alveolar lavage concentrate. PCR buffer composition was: 25 mM [3-(Tris-hydroxymethyl)methylamino]-l-propanesulfonic acid (TAPS), pH 9.3; 2 mM magnesium chloride; 50 mM potassium chloride; 1 mM dithiothreitol; 0.05% (v/v) W-l; 0.2 mM dATP, dCTP, dGTP, dTTP and 1.25 IJM of each primer MP5-1, MP5-2; 6 final volume: 501~1. The reaction mixture was incubated at 100°C for 10 rain and then ice cooled. One unit of Taq polymerase was added and the final mixture was overlaid with 10 I11 of mineral oil. PCR Was taken through 30 cycles of amplification (except when indicated), each cycle consisting in 95°C for 1 min, 55°C for 1 min and 72°C for 3 min. We increased the length of the extension step because Dig-dUTP partially inhibits Taq polymerase. Temperature cycling was carried out in a programmable thermostat (DNA Thermal Cycler, Perkin Elmer/ Cetus). Amplified samples were stored at -20°C. PCR products were labelled either during the amplification reaction by adding Dig-dUTP (digoxigenin-dUTP) at a final concentration of 650 nM or after PCR. In that case, the amplified products were first purified on Qiagen tips (Diagen) according to the suppliers' instructions and concentrated by ethanolic precipitation. Labelling was

Immobilization of capture DNA

We used Primaria 96-well plates (Falcon) coated with primary amines. The plates were activated with 1% glutaraldehyde in 0.5M phosphate buffer, pH 7, 100111 per welt, for 15rain. The capture DNA (the l inearized pA021 plasmid or the 144-base-pair amplification product) was dissolved in PBS (at a final concentration of 10 l~g m1-1 and 1O0 Ill (1 I~g) of the diluted DNA were added to each well to cover a 1 cm 2 surface. After shaking for 15 min, the wells were washed three times with 3501~1 washing buffer (50 mM phosphate buffer, pH 7). The remaining non-reacted aldehyde groups were reacted with 1% ethylenediamine in phosphate buffer (300 I~1 per well) for 15 min with agitation, followed by three washes. Non-specific binding sites were then blocked with 200 I~1 per well of blocking buffer (100 mM Tris-HCI, pH 7.5, 150 mM sodium chloride 0.5% (w/v) blocking reagent (Boehringer Mannheim) for 1 h. Up to this point all the reactions were performed at room temperature. When dried, these plates could be stored at -20°C for at least 6 months without loss of their properties.

Hybridization

Mictotitre plates coated with capture DNA were prehybridized at 42°C for 1 h with each well containing 200p.I of the following prehybridization buffer: 100 mM Tris-HCI, pH 7.5, 50 mM magnesium chloride, 0-5% (w/v) blocking reagent. For incubation, plates were sealed with Parafilm. A 4-1~1aliquot of the PCR sample was diluted to a final volume of 100 i~1 in the 'rapid hybridization buffer' [i.e. the prehybridization buffer containing 1% (w/v) polyethylene glycol 8000 (Pharmacia) and 50% (v/v) formamide (Aldrich)]. The DNA to be hybridized was denatured at 100°C for 10min and added quickly to the wells after removing the prehybridization buffer. Samples were hybridized at 42°C for 5 min. After hybridization, the wells were washed twice with 350 p.I of 2 x SSC (1 x SSC is 15 mM sodium citrate and 15 mM sodium chloride) for 15 min and were blocked for 30 min with 200 I~1

Fast microplateassayfor PCR detection of M. pneumoniae of blocking buffer and this was followed by a 30min incubation with 2001~1 alkaline phosphatase-anti-digoxigenin IgG conjugate (Boehringer Mannheim), diluted to 150Um1-' in the same blocking buffer. Wells were washed three times and incubated for 1 h with 2001~1 of 2 mgml-' para-nitrophenyl phosphate in 100 mM Tris-HCI, pH 9.3, 150 mM sodium chloride, 50 mM magnesium chloride. Absorbance was read at 405 nm. The readings were made against wells containing only buffer or against wells treated the same way as the others except that no PCR products were added.

Determination of the amount of actually hybridized products After hybridization, the PCR Dig-labelled products were denatured and removed from the microwells and spotted on nylon membrane. The DNA was detected with anti-Dig alkaline phosphatase conjugate using 5-bromo-4-chloro-3-indolyl-phosphate and nitroblue tetrazolium as described by the manufacturer and densitometric data were obtained by an LKB Ultroscan XL Laser Densitometer.

Gel electrophoresis PCR products were also analysed by gel electrophoresis as previously described. 6

27

The strategy of the assay is outlined in Fig. 1. The capture DNA is denatured and covalently bound to the wells of activated microplates2 The haptenlabelled amplified products are incubated in the wells and hybridized to the capture DNA. Finally, the amount of hybridized products is measured by a classical immunoassay procedure. The experimental conditions were optimized for low background and high sensitivity of detection of the PCR products. The sensitivity of the assay for biological samples has been found to be equal to that of the procedure described by Bernet et al.; 6 that is to say, 10 ccu (ccu: colour changing unit; 1 ccu is estimated to be equivalent to 10 to 100 Mycoplasma cells). The method was not directly compared to the culture assay, but our previous PCR assay was at least as sensitive as the culture. This allows the use of only a small aliquot of the PCR sample, so that enough material can be saved for further analysis of the products (detection of RFLP, for instance).

Background The influence of various factors affecting the background was evaluated by omitting any one of the steps of the protocol or various reagents leaving unchanged all the other conditions (Fig. 2). When all the blocking steps were performed, the background was very low and came only from the spontaneous dephosphorylation of the substrate. To evaluate the cut-off value, assays were made using for the hybridization step either non-specific labelled DNA or hybridization buffer alone. An absorbance of 0-06 (i.e. the mean value for these assays plus twice their standard deviation) was used as the cut-off value.

Clinical specimens

Binding of the capture DNA

Broncho-alveolar lavages were centrifuged and the pelleted cells were resuspended in 500111 of the supernatant and stored at -20°C until use. Ten microlitres of the samples were used for amplification.

We checked the efficiency of the assay as a function of glutaraldehyde concentration used for the binding of the capture DNA. As shown in Fig. 3, a concentration of about 1% was optimal. Higher concentrations led to a reduction of sensitivity when small-sized DNA fragments were used for capture. Probably, an excessive crosslinking of the capture DNA, due to too high a density of reactive groups, inhibits its hybridization properties. To check the stability of the covalent links between the wells and the capture DNA, the same plates were used several times. After development of the colour, the captured DNA was removed by washing each well with 200 I11of formarnide at 42°C for 15 mira In

RESULTS AND DISCUSSION Our assay for NI. pneumoniae DNA detection combines PCR amplification and labelling of a 144-basepair mycoplasmal sequence with non-radioactive detection in microtitre wells after specific hybridization.

28

A. Vekris et a/.

Enzyme Antibody

I-I,,pten.~

~

[I ~ biCovalent nding l " Sotid [ Jsupport

Capture molecule

fffffffffffffflfJI#rf ff~fff

ff~ffffffffffffff

Fig. 1. Schematic diagram of the assay. (Denaturation of the capture DNA; covalent binding to aldehyde-activated microwells; blocking with ethylene diarnine; saturation of non-specific sites with 'blocking reagent'; hybridization with PCR-labelled products; saturation of non-specific sites with 'blocking reagent'; reaction with the antibody-enzyme conjugate; addition of the enzyme substrate and development of the coloured product,)

0.070

-

8 0.050

0.030

0.010

fl-"--"-] I

I

I

I

Step

or reagent

Fig. 2.

II

I

omited

Effect of the omission of various steps or reagents on the background.

these conditions we could use the microplates at least three times without loss of sensitivity.

Hybridization

ll'"

reaction

After PCR, the labelled products were hybridized to the pAo21 plasmid bound to the microtitre plate and the hybridized products were detected immunoenzymatically. We found that several hours were necessary to obtain a good signal intensity in

the standard hybridization medium (data not shown), while hybridization was very fast if polyethylene glycol and magnesium ions were included in the hybridization buffer (Fig. 4, open circles). It can be seen that the minimum practical hybridization time, .5 rain, was enough to get a good optical signal. We noticed that longer incubation led to a decrease in the signal even though the amount of hybridized PCR products increased (Fig. 4, closed circles). This could be explained by the formation of networks 7 leading to a poor accessibility of the conjugate to the hapten.

Fast microplate assay for PCR detection of M. pneumoniae

29

Application to the detection of M. pneumoniae in clinical specimens

~0.8 ~0.4

0

1 2 3 4 Glutaraldehyde concentration (%)

5

Fig. 3. Effectof the concentration of glutaraldehyde used for activation on the hybridization signal. The capture DNA was either the linearized pAo21 plasmid (0) of 3460 bp (containing a sequence homologous to the PCR product), or the 144-bp PCR product (O).

For this particular application, 5-rain hybridization time was optimum; however, this parameter has to be optimized for each sequence and PCR product size (data not shown). For a negative control, we used the non-recombinant linearized Bluescribe plasmid (pBS) as the capture molecule. No signal above the background was recorded.

Simultaneous amplification and labelling were done on aliquots of broncho-alveolar lavages. Then hybridizations were performed with the amplified products obtained after either 12 or 30 cycles (Fig. 5B). As expected, the signal was stronger when using the 30-cycles PCR products, but the 12cycles PCR products can also be used for the detection. This allows a reduction in the duration of the assay. The same 12- and 30-cycles PCR products were also analysed by gel electrophoresis. The results were in agreement with those of the colorimetric assay. We then wanted to determine that the labelling with digoxigenin during the PCR step did not introduce any losses of specificity or sensitivity. So aliquots of the clinical samples were amplified without Dig-dUTP. The resulting products were then labelled by digoxigenin incorporation at the 3' end and analysed either by gel electrophoresis (Fig. 5D) or by microplate hybridization (Fig. 5B). There was no significant difference between the results obtained if the PCR reaction was conducted in the presence or absence or Dig-dUTP.

0.4

0.4

0-3

0.3

0.2

o.2

0.1

0.1

8 v

I

10

I

I

8~

I

20 30 40 Hybridization time (rain)

50

Fig. 4. Kinetics of hybridization in 'rapid buffer'. The hybridization and the measure of the amount of DNA actually hybridized were conducted as described in Materials and Methods. Absorbance yielded by the PCR products hybridized to the microplates: (0). Amount of actually hybridized PCR products measured after removal by denaturation and spotting on a membrane (0).

30

A. Vekris et a/.

A

I :I +

+ + -l + - l - I + - - I - I ? I

- -l+l+l+I-I

I

!

I

- -

,

1

-

r-~

°i., 5

2

['

F

•

I

!

6

7

8

9

K

I0

11 12 13 14 15 16 17 18 20 21 22 23 24 25 28 29 43

I-l-I?J?l?l-i+li

:

I

!

1.5 8

!

D O

i I

I

_

0-5 0

H 5

6

IN

7

8

9

I~11

N

IMI

IN

INI

10 11 12 13 14 15 16 17 18 20 21 22 23 24 25 28 29 43

Sample number

Fig. 5. Analysis of PCR products from clinical samples. Gel electrophoresis: (A and C). Not done: ND. Absence of ethidium-bromide-detectable product: - . Presence of the expected 144-bp product: +. Presence of an uninterpretable pattern: ?. Microplate hybridization analysis: (13 and D). Optical readings from products obtained after 12 cycles (open bars) or 30 cycles (hatched bars) of PCR. The products were labelled during (A and B) or after (C and D) PCR. The revelation time was 30 rain (B) or 240 rain (D) and the readings were made against wells without added PCR products.

PCR inhibition in clinical samples

CONCLUSION

We noticed, wh.en running different aliquots of the same sample, that the signal intensity was often not correlated to the aliquot size, suggesting that there is an inhibition of the amplification reaction by components contained in the sample. This was confirmed by running a model reaction set up by introducing variable quantities of linearized A ~ I plasmid in a reaction mixture containing 10 I~1of clinical samples. Figure 6 shows that there is actually a partial inhibition of the reaction. This inhibition was also observed when a different PCR target sequence was used (i.e. pBS plasmid and T3/T7 oligonucleotides) (Fig. 6). This inhibition precludes the use of PCR for quantification of M. pneumoniae in clinical samples without prior DNA purification.

Detection of fastidious bacteria such as M. pneumoniae in clinical specimens by culture is slow and less sensitive than PCR assays. However, for PCR to become a routine procedure, the assay must be simple, isotope-free and rapid. Our microplate assay meets these criteria. Furthermore, it leaves the majority of the PCR product available for further analysis. Our assay is very simple, fast and cheap and does not use any synthetic DNA molecule since the capture DNA is plasmidial. Indeed, we first showed that our covalent binding of capture DNA makes it possible to store plates for a long time and reuse them several times, so that it is not necessary to prepare the plates for each series of assays. Second, the use of a rapid hybridization buffer allows us to get an optimal signal

Fast microplate assay for PCR detection of M. pneumoniae

31

(a) 1 ng

10 ng

II

100 ng

I000 ng

II

I

1-5

Nn

1

0

~0-5

6

6

6

6

7

In 8

9

(b) 1 ng

o

100 ng

I

1000 ng

I

I

7

2 8

10 ng

I

I

7 m

1.5

\

1

\ \

0.5

\

\ \ \

\ \ \

\ \ \

\ \

\ \

\ \

\

,

6

\

,

6 Sample number

"1

\

6

7

8

9

Fig. 6. Presenceof PCR inhibitors in clinical samples. Amplifications using 1, 10, 100 and 1000 ng of A021 plasmid as a template and MP5-1 and MP5-2 as primers (a), or 1, 10, 100 and 1000 ng pBS plasmid as a template and T3 and T7 as primers (b), were run in the absence (hatched bars), or in the presence (open bars), of lO-p.I aliquots of the indicated clinical samples. Revelation time: 30 min. in 5 min; we also showed that our assay was sensitive enough for using only 12 PCR cycles. These elements allow us to shorten the time necessary for the assay. The microtitre plate format makes the assay compatible with standard ELISA pipetting, washing and reading equipment, already present in clinical laboratories, and can be automated.

3. 4.

5. REFERENCES

1. Saiki, R. K., Walsh, P. S., Levenson, C. H. & Erlich, H. A. (1989). Genetics of amplified DNA with immobilised sequence specific oligonucleotide probes. Proceedings of the National Academy of Science, USA 86, 6230-4. 2. Keller, G. H., Huang, D.-P., Shih, J. W. K. & Manak, M. M. (1990). Detection of hepatitis B virus DNA in serum by polymerase chain reaction amplification and

6.

7.

microtiter sandwich hybridization. Journal of Clinical Microbiology 28, 1411-16. Inouye, S. & Hondo, R. (1990). Microplate hybridization of amplified viral DNA segment. Journal of Clinical Microbiology 28, 1469-72. L0nenberg, E., Jensen, J. S. & Frosch, M. (1993). Detection of Mycoplasma pneumoniae by polymerase chain reaction and nonradioactive hybridization in microtiter plates. Journal of Clinical Microbiology 31, 1088-94. Gibellini, D., Zerbini, M., Musiani, M., Venturoli, S., Gentilomi, G. & La Placa, M. (1993). Microplate capture hybridization of amplified parvovirus B19 DNA fragment labelled with digoxigenin. Molecular and Cellular Probes 7, 453-8. Bernet, C., Garret, M.~ de Barbeyrac, B., B~b~ar, C. & Bonnet, J. (1989). Detection of Mycoplasma pneumoniae by using the polymerase chain reaction. Journal of Clinical Microbiology 27, 2492-6. B0nneman, H. (1982). Immobilisation of denatured DNA to macroporous supports: II Steric and kinetic parameters of heterogeneous hybridization reactions. Nucleic Acids Research22, 7181-96.