- Email: [email protected]
Detection of human cytomegalovirus polymerase chain reaction products using oligonucleotide probes directly conjugated to alkaline phosphatase
Journal of viiological Methods Journal of Virological Methods 61 (1996) 145-150
Detection of human cytomegalovirus polymerase chain reaction products using oligonucleotide probes directly conjugated to alkaline phosphatase Dehila Gor, David Lee, Vincent C. Emery Department of Virology, Royal Free Hospital and School of Medicine, Rowland Hill Street, Hampstead, London NW3 2QG, UK
Accepted 1 May 1996
Abstract
A 24 base pair oligonucleotide probe directly conjugated to alkaline phosphatase has been used to detect immobilised amplicons derived from a cytomegalovirus specific polymerase chain reaction (PCR). The sensitivity of detection using a hig,hly amplified alkaline phosphatase detection system was four genome equivalents and was comparable to the limit of detection using agarose gel methods. The mean optical density at 492 nm of samples not known to contain cytomegalovirus DNA was 0.085 + 0.006 and was well separated from the optical density generated from four genome equivalents (absorption at 492 nm: 0.132). The assay was used to identify the presence of cytomegalovirus in blood DNA extracts from immunocompromised patients in whom conventional ethidium bromide stained agarose gel ellectrophoresis revealed the presence of multiple amplicons. Samples yielding an uninterpretable result at both neat and diluted 1 in 20 in the PCR gave rise to the highest proportion of positive results (68%) whilst samples that produced uninterpretable results neat but were negative at 1 in 20 and vice versa gave positive rates of 33.6 and 21.7%, reslpectively. The use of this assay for identifying cytomegalovirus specific PCR products in problematic samples .is discussed. Keywords:
Alkaline phosphatase;
Cytomegalovirus;
Herpesvirus; DNA
1. Introduction The polymerase chain reaction come established as one of the techniques for the identification of agents (Whitley and Lakeman, 0166-0934/96/$15.00
0 1996
PII SOl66-0934(96)02081-2
(PCR) has bemost powerful viral infectious 1995; Garson,
1994; Gravitt and Manos, 1992; Birkenmeyer and Mushahwar, 1994). Detection of PCR amplicons may involve direct visualisation on agarose gels containing ethidium bromide, oligonucleotide probing of immobilised products following Southern blotting, or capture of amplicons via oligonu-
Elsevier Science B.V. All rights reserved
146
D. Gor et al. /Journal
of Virological Methods 61 (1996) 145-150
cleotide probes immobilised on the solid phase, followed by enzymatic detection. We described previously a sensitive and specific PCR methodology for the detection of human cytomegalovirus in clinical samples from transplant patients and AIDS patients (Kidd et al., 1993). While the PCR protocol is highly reproducible in the majority of samples so far analysed (in excess of 15 000 samples of blood or urine), occasionally non-specific products, i.e. multiple amplicons of varying molecular size, are generated which require further investigation. In order to improve the rapidity of detection of the PCR product and to alleviate the problems of non-specific banding, we have investigated the use of an oligonucleotide probe directly conjugated to alkaline phosphatase for the detection of the PCR amplicon. Alkaline phosphatase-labelled oligonucleotide probes have been used for the identification of infectious agents, either directly (Huang and Jungkind, 1991) or following PCR (Cano et al., 1992, 1993; Rasmussen et al., 1994). In some cases, the alkaline phosphatase is conjugated to an internal base within the probe (Bollet et al., 1992; Thorne et al., 1990), whilst in other studies direct conjugation of the enzyme to the 5’ terminus of the oligonucleotide has been utilised (Cano et al., 1993; Whitby and Garson, 1995; Rasmussen et al., 1994). In the current study, we report the use of an oligonucleotide probe with alkaline phosphatase directly conjugated to the 5’ terminus for the detection of immobilised amplicons derived from the cytomegalovirus-specific PCR method used in our laboratory.
2. Materials and methods 2.1. Streptavidin coating of microtitre plates
The production of streptavidin coated microtitre plates was essentially via the method described by Kaye et al., 1992. Briefly, streptavidin (25 ~1 of a 25 pg/ml solution in Tris-HCl, pH 7.6) was added to each well of a 96-well Nunc Maxisorb U-well plate and incubated overnight at room temperature in a humidified chamber. The wells were washed twice with PBS and subse-
quently filled with blocking solution consisting of 1% bovine serum albumin in PBS containing 0.1% sodium azide. After 1 h at room temperature, the blocking solution was aspirated to leave approximately 100 ,~l in each well. 2.2. Polymerase chain reaction The conditions for the amplification of HCMV using primers specific for the glycoprotein B sequence have been previously described (Darlington et al., 1991). In the present study, PCR was performed under identical conditions except that primer gB2 was labelled at its 5’ terminus with biotin. In all PCR analyses of clinical samples, 5 ~1 of blood DNA extract equivalent to 25 ~1 of whole blood and 5 ~1 of urine were analysed together with 1 in 20 dilutions. Details of this procedure have been outlined in Kidd et al. (1993). 2.3. Detection of PCR products The remaining blocking solution from the streptavidin coated plates (see above) was aspirated and 25 pl of the PCR product (total volume 100 ~1) added to the appropriate wells and incubated at room temperature for 0.5 h. The wells were washed three times with washing solution (PBS containing 0.05% Tween-20) and 40 pi/well of 0.15 M NaOH added. After 5 min incubation at room temperature, the wells were washed a further four times with washing solution. The alkaline phosphatase coupled oligonucleotide probe (alkaline phosphatase-5’-GGTACTCGTAGGCCGAGTT(GC)CCGGCGATGA) was diluted in PBS containing 2% dried milk powder (Marvel), 20% sheep serum (SeraLab) and 0.5% Tween-20 and 5 ~1 added to each well of the plate. The plate was sealed and incubated at 37°C for 0.5 h. Subsequent detection of the alkaline phosphatase conjugated hybrid was performed using an AMPAK kit as specified in the manufacturer’s instructions (DAK0 Diagnostics, Cambridge, UK). The reaction was terminated by addition of 50 pi/well of 2 M sulphuric acid and the optical density of the reaction products determined at 492 nm using a conventional plate reader.
D. Gor et al. /Journal
0
200
400
of Virological Methods 61 (1996) 145-150
600
800
1000
147
1200
Number of input HCMV genomes Fig. 1. Standard curve relating the optical density at 492 nm to the quantity of input target DNA included in the HCMV PCR. The ODdQ2for 1, 2, 4, 5, 8, 16, 32, 64, 128, 256, 512 and 1000 genome copies is shown in the inset table. The results represent the mean values obtained from three replicas. 3. Results
Clinical samples (147) of blood obtained from organ transplant recipients in whom no active HCMV replication had been identified post-transplant by PCR were used to determine the negative cut-off for the assay. Using the alkaline phosphatase probe method the mean ( f 1 SD.) The ODdg2 for these samples was 0.085 f 0.006. The results followed a normal distribution. Thus, samples not known to contain HCMV DNA produced a consistently low OD,,, under the assay conditions. As a consequence of these data, the negative cut-off for the assay was set at 0.104 OD,,, (mean + 3 SD.) and was used in all future analyses. Subsequently, known copy numbers of amplicon cloned into pUC18 were used in the range of l-1000 copies to generate a standard curve. The results shown in Fig. 1 demonstrate that the method was capable of detecting amplicons derived from four copies of HCMV target. The sensitivity of the alkaline phosphatase labelled probe method was marginally better than gel electrophoresis and had the advantage of providing a more objective measure of low levels of amplicon. Intra-assay variability was between 2-5%, however, interassay variability was higher
and usually ranged from 10 to 20% depending upon copy number of input DNA. Variability between 15 and 20% was observed with low copy numbers ( < 10 genomes). It should be noted that the optical density obtained for five copies of target HCMV (OD,,, = 0.157; see Fig. 1) was well separated from the negative cut-off. The use of this assay to identify true positive reactions in 168 samples yielding multiple bands within our existing PCR method was evaluated. An example of the amplicons produced in these samples is shown in Fig. 2 together with the alkaline phosphatase-oligonucleotide probe results. The distribution of results for all samples analysed is shown in Table 1. Of the 44 samples in which a non specific result was obtained at both dilutions of sample (5 pl/ 100 ~1 PCR mixture and 0.25 pl/lOO ~1 PCR mixture), 68% were positive following hybridization. In contrast, samples in which the undiluted sample yielded non-specific bands but which were negative for HCMV DNA at the 1 in 20 dilution (n = 101) yielded only 33.6% positive results. Finally, in samples that were negative neat but yielded non specific bands at 1 in 20 (n = 23), the true positive rate was 21.7%. The mean OD,,, for samples that were positive was 0.63 indicating that these samples
D. Gor et al. /Journal
148
of Virological Methods 61 (1996) 145-150
6
A
nnNnnNn --_-_++ ++
nNnnNn -I++
II
--_
_
++
++
n _I
a
I_
nn +++--
NP
a
Fig. 2. Results of gel analysis of HCMV amplicons compared with the results using the alkaline phosphatase conjugated oligonucleotide probe. In panel A the results for a selection of positive and negative clinical samples are shown. Each sample was analysed neat (5 ~1) and at a 1 in 20 dilution (see Materials and methods). The results from probing are shown by the positive and negative signs above each lane. The open arrow indicates the migratory position of the 149 bp amplicon, N, negative control. In panel B the results for a selection of samples yielding non-specific amplicons are shown analysed by gel electrophoresis and by the alkaline phosphatase method. The descriptors are the same as those used in panel A, p, positive control (50 genomes).
contained a reasonable quantity of target HCMV molecules. Overall, only 2% of samples yielded results that were within 4 standard deviations from the mean OD,,, of the negative samples.
4. Discussion The sensitivity of the PCR depends upon many variables including primer sequences, length of amplified sequence and genetic composition of the sequence. We have previously described the utility of two oligonucleotides specific for the glycoprotein B gene of HCMV that are highly sensitive and result in the amplification of the majority of HCMV strains present in the population (Kidd et al., 1993; Darlington et al., 1991; Roy et al., 1993). However, due to the inherent nature of a minority of samples, especially blood DNA ex-
tracts, the resulting amplicons following PCR comprise a range of molecular sizes. In such instances, it is difficult to ascertain whether the sample contained the target sequence and hence to provide a diagnostic result. In order to circumvent this problem we have used a probe oligonudirectly conjugated to alkaline cleotide phosphatase to detect biotinylated amplicons immobilised on streptavidin coated microtitre plates. The results from plasmid reconstruction experiments demonstrated that this process resulted in the detection of amplicons derived from as little as four copies of target sequence. Importantly, the optical density obtained from either four or five copies of target was substantially higher than the mean optical density obtained for clinical samples not known to contain HCMV. The direct detection of amplicons using alkaline phosphatase conjugated oligonucleotide probes has been described
D. Gor et al. /Journal
of Virological Methods 61 (1996) 145-150
149
Table 1 Distribution of results using the alkaline phosphatase labelled oligonucleotide probe Primary results
Number of samples
Neat
l/20
NSB”
NSB
44
NSB
Neg
10’1
Neg
NSB
23
Alkaline phosphatase probe results Neat
l/20
Posh Pos Neg’ Neg Pos Pos Neg Neg Pos Pos Neg Neg
Pos Neg Pos Neg Pos Neg Pos Neg Pos Neg Pos Neg
Number of samples
Mean OD,,, (range)
8 14 6 16 0 34 0 67 0 4 2 17
1.414 (0.107-3.424) 0.842 (0.105-2.900) 0.626 (0.105-3.596) 0.052 (0.001~0.104) N/Ad 0.564 (0.1 l-2.972) N/A 0.029 (0.006-0.080) N/A 0.486 (0.116-0.863) 0.131 (0.109-0.153) 0.022 (0.002-0.047)
Samples (168) yielding non specific banding results at either neat, 1 in 20 or both concentrations using conventional agarose gel electrophoresis/ethidium Ibromide staining of amplicons (primary result). The mean optical densities in the Pos/Neg and Neg/Pos combinations were calculated for the positive samples only. “Non-specific bands. bPositive. “Negative. dNot applicable.
previously (Cano et al., 1993; Whitby and Garson, 1995) although in some cases the sensitivity of detection has been limiting (for example, Rasmussen et al., 1993). This may have been a consequence of an insensitive PCR process used in some of these studies, although a more likely explanation may reiside in the methods used for detecting the hybrild product. In our assay we have utilised an amplification system based on the conversion of NADlPH to NADH which is used in the subsequent amplification step using diaphorase and alcohol dehydrogenase. The resulting redox cycle prolduces coloured formazan and leads to a signal amplification of approximately lOO-fold. Under these conditions, amplicons from four copies of target sequence were reliably detected. Such quantities of product are at the limit of detection in conventional gel electrophoresis systems which, in our hands, requires amplicons from 5-10 copies of target sequence to produce easily visible products after ethidium bromide staining. Whilst the: dynamic range of this assay does not approach chemiluminescent detection of
alkaline phosphatase conjugated oligonucleotides (Whitby and Garson, 1995) it has the advantage of requiring no new instrumentation over existing enzyme linked immunoassays whilst still maintaining sensitivity. In order to evaluate the usefulness of this assay we analysed a series of clinical specimens that yielded indeterminate results in a conventional gel electrophoresis based amplicon detection system. The results confirmed that in the majority of cases a finite result could be obtained using the alkaline phosphatase conjugated oligonucleotide approach. Interestingly, 68% of samples in which the primary analyses at both neat and 1 in 20 dilutions yielded non-specific bands were true positive reactions whereas only 33.6 and 21.7% were true positives in the combination of non-specific neat, negative at 1 in 20 or vice versa, respectively. Hence, samples that were negative for HCMV either undiluted or diluted 1 in 20 in our existing assay were more likely to be true negatives than samples yielding non-specific bands at both concentrations of analyte. These samples account for
150
D. Gor et al. /Journal
of Virological Methods 61 (1996) 145-150
approximately 85% of samples yielding non-specific reaction products. As a consequence of these data we incorporated this methodology into the laboratory for the assessment of samples yielding non-specific results. At present, the comparative costs of performing gel analysis versus the alkaline phosphatase method precludes the more widespread use of the latter approach in routine PCR surveillance for HCMV in transplant recipients and AIDS patients.
Acknowledgements This work was supported by the UK Medical Research Council and the Wellcome Foundation.
References Birkenmeyer, L.G. and Mushahwar, I.K. (1994) Detection of hepatitis A, B and D virus by the polymerase chain reaction. J. Virol. Methods 49, 101-l 12. Ballet, C., Vignoli, C., Freney, J., Gevaudan, M.J., Galicher, V. and de Micco, P. (1992) Identification of Mycobacterium tuberculosis and Mycobacterium avium by non-radioactive probes: evaluation of the snap syngene system. Annal. Biol. Clin. 50, 599-601. Cano, R.J., Torres, M.J., Klem, R.E., Palomares, J.C. and Casadesus, J. (1992) Detection of salmonellas by DNA hybridization with a fluorescent alkaline phosphatase substrate. J. Appl. Bacterial. 72, 393-399. Cano, R.J., Rasmussen, S.R., Sanchez Fraga, G. and Palomares, J.C. (1993) Fluorescent detection-polymerase chain reaction (FD-PCR) assay on microwell plates as a screening test for salmonellas in foods. J. Appl. Bacterial. 75, 247-253. Darlington, J., Super, M., Patel, K., Grundy, J.E., Griffiths, P.D. and Emery, V.C. (1991) Use of the polymerase chain reaction to analyse sequence variation within a major
neutralizing epitope of glycoprotein B (gp58) in clinical isolates of human cytomegalovirus. J. Gen. Virol. 72, 198551989. Garson, J.A. (1994) The polymerase chain reaction and hepatitis C virus diagnosis. FEMS Microbiology Reviews 14, 2299239. Gravitt, P.E. and Manos, M.M. (1992) Polymerase chain reaction-based methods for the detection of human papillomavirus DNA. IARC Scientific Publications 121- 133. Huang, C.H. and Jungkind, D.L. (1991) Non-radioactive DNA probe for the rapid identification of Mycobacterium avium complex from clinical specimens. Mol. Cell. Probes 5, 2777280. Kaye, S., Loveday, C. and Tedder, R.S. (1992) A microtitre format point mutation assay: application to the detection of drug resistance in human immunodeficiency virus type-1 infected patients treated with zidovudine. J. Med. Virol. 37, 241-246. Kidd, I.M., Fox, J.C., Pillay, D., Charman, H., Griffiths, P.D. and Emery, V.C. (1993) Provision of prognostic information in immunocompromised patients by routine application of the polymerase chain reaction for cytomegalovirus. Transplantation 56, 867-871. Rasmussen, S.R., Rasmussen, H.B., Larsen, M.R., Hoff-Jorgensen, R. and Cano, R.J. (1994) Combined polymerase chain reaction-hybridization microplate assay used to detect bovine leukemia virus and Salmonella. Clin. Chem. 40, 200&205. Rasmussen et al., 1993. Author please provide details. Roy, D.M., Grundy, J.E. and Emery, V.C. (1993) Sequence variation within neutralizing epitopes of the envelope glycoprotein B of human cytomegalovirus: comparison of isolates from renal transplant recipients and AIDS patients. J. Gen. Virol. 74, 2499-2505. Thorne, G.M., Macone, A. and Goldmann, D.A. (1990) Enzymatically labelled nucleic acid (NA) probe assays for detection of Campylobacter spp. in human faecal specimens and in culture. Mol. Cell Probes 4, 133-142. Whitby, K. and Garson, J.A. (1995) Optimisation and evaluation of a quantitative chemiluminescent polymerase chain reaction assay for hepatitis C virus RNA. J. Virol. Methods 51, 75-88. Whitley, R.J. and Lakeman, F. (1995) Herpes simplex virus infections of the central nervous system: therapeutic and diagnostic considerations. Clin. Infect. Dis 20, 414-420.
Detection of human cytomegalovirus polymerase chain reaction products using oligonucleotide probes directly conjugated to alkaline phosphatase Dehila Gor, David Lee, Vincent C. Emery Department of Virology, Royal Free Hospital and School of Medicine, Rowland Hill Street, Hampstead, London NW3 2QG, UK
Accepted 1 May 1996
Abstract
A 24 base pair oligonucleotide probe directly conjugated to alkaline phosphatase has been used to detect immobilised amplicons derived from a cytomegalovirus specific polymerase chain reaction (PCR). The sensitivity of detection using a hig,hly amplified alkaline phosphatase detection system was four genome equivalents and was comparable to the limit of detection using agarose gel methods. The mean optical density at 492 nm of samples not known to contain cytomegalovirus DNA was 0.085 + 0.006 and was well separated from the optical density generated from four genome equivalents (absorption at 492 nm: 0.132). The assay was used to identify the presence of cytomegalovirus in blood DNA extracts from immunocompromised patients in whom conventional ethidium bromide stained agarose gel ellectrophoresis revealed the presence of multiple amplicons. Samples yielding an uninterpretable result at both neat and diluted 1 in 20 in the PCR gave rise to the highest proportion of positive results (68%) whilst samples that produced uninterpretable results neat but were negative at 1 in 20 and vice versa gave positive rates of 33.6 and 21.7%, reslpectively. The use of this assay for identifying cytomegalovirus specific PCR products in problematic samples .is discussed. Keywords:
Alkaline phosphatase;
Cytomegalovirus;
Herpesvirus; DNA
1. Introduction The polymerase chain reaction come established as one of the techniques for the identification of agents (Whitley and Lakeman, 0166-0934/96/$15.00
0 1996
PII SOl66-0934(96)02081-2
(PCR) has bemost powerful viral infectious 1995; Garson,
1994; Gravitt and Manos, 1992; Birkenmeyer and Mushahwar, 1994). Detection of PCR amplicons may involve direct visualisation on agarose gels containing ethidium bromide, oligonucleotide probing of immobilised products following Southern blotting, or capture of amplicons via oligonu-
Elsevier Science B.V. All rights reserved
146
D. Gor et al. /Journal
of Virological Methods 61 (1996) 145-150
cleotide probes immobilised on the solid phase, followed by enzymatic detection. We described previously a sensitive and specific PCR methodology for the detection of human cytomegalovirus in clinical samples from transplant patients and AIDS patients (Kidd et al., 1993). While the PCR protocol is highly reproducible in the majority of samples so far analysed (in excess of 15 000 samples of blood or urine), occasionally non-specific products, i.e. multiple amplicons of varying molecular size, are generated which require further investigation. In order to improve the rapidity of detection of the PCR product and to alleviate the problems of non-specific banding, we have investigated the use of an oligonucleotide probe directly conjugated to alkaline phosphatase for the detection of the PCR amplicon. Alkaline phosphatase-labelled oligonucleotide probes have been used for the identification of infectious agents, either directly (Huang and Jungkind, 1991) or following PCR (Cano et al., 1992, 1993; Rasmussen et al., 1994). In some cases, the alkaline phosphatase is conjugated to an internal base within the probe (Bollet et al., 1992; Thorne et al., 1990), whilst in other studies direct conjugation of the enzyme to the 5’ terminus of the oligonucleotide has been utilised (Cano et al., 1993; Whitby and Garson, 1995; Rasmussen et al., 1994). In the current study, we report the use of an oligonucleotide probe with alkaline phosphatase directly conjugated to the 5’ terminus for the detection of immobilised amplicons derived from the cytomegalovirus-specific PCR method used in our laboratory.
2. Materials and methods 2.1. Streptavidin coating of microtitre plates
The production of streptavidin coated microtitre plates was essentially via the method described by Kaye et al., 1992. Briefly, streptavidin (25 ~1 of a 25 pg/ml solution in Tris-HCl, pH 7.6) was added to each well of a 96-well Nunc Maxisorb U-well plate and incubated overnight at room temperature in a humidified chamber. The wells were washed twice with PBS and subse-
quently filled with blocking solution consisting of 1% bovine serum albumin in PBS containing 0.1% sodium azide. After 1 h at room temperature, the blocking solution was aspirated to leave approximately 100 ,~l in each well. 2.2. Polymerase chain reaction The conditions for the amplification of HCMV using primers specific for the glycoprotein B sequence have been previously described (Darlington et al., 1991). In the present study, PCR was performed under identical conditions except that primer gB2 was labelled at its 5’ terminus with biotin. In all PCR analyses of clinical samples, 5 ~1 of blood DNA extract equivalent to 25 ~1 of whole blood and 5 ~1 of urine were analysed together with 1 in 20 dilutions. Details of this procedure have been outlined in Kidd et al. (1993). 2.3. Detection of PCR products The remaining blocking solution from the streptavidin coated plates (see above) was aspirated and 25 pl of the PCR product (total volume 100 ~1) added to the appropriate wells and incubated at room temperature for 0.5 h. The wells were washed three times with washing solution (PBS containing 0.05% Tween-20) and 40 pi/well of 0.15 M NaOH added. After 5 min incubation at room temperature, the wells were washed a further four times with washing solution. The alkaline phosphatase coupled oligonucleotide probe (alkaline phosphatase-5’-GGTACTCGTAGGCCGAGTT(GC)CCGGCGATGA) was diluted in PBS containing 2% dried milk powder (Marvel), 20% sheep serum (SeraLab) and 0.5% Tween-20 and 5 ~1 added to each well of the plate. The plate was sealed and incubated at 37°C for 0.5 h. Subsequent detection of the alkaline phosphatase conjugated hybrid was performed using an AMPAK kit as specified in the manufacturer’s instructions (DAK0 Diagnostics, Cambridge, UK). The reaction was terminated by addition of 50 pi/well of 2 M sulphuric acid and the optical density of the reaction products determined at 492 nm using a conventional plate reader.
D. Gor et al. /Journal
0
200
400
of Virological Methods 61 (1996) 145-150
600
800
1000
147
1200
Number of input HCMV genomes Fig. 1. Standard curve relating the optical density at 492 nm to the quantity of input target DNA included in the HCMV PCR. The ODdQ2for 1, 2, 4, 5, 8, 16, 32, 64, 128, 256, 512 and 1000 genome copies is shown in the inset table. The results represent the mean values obtained from three replicas. 3. Results
Clinical samples (147) of blood obtained from organ transplant recipients in whom no active HCMV replication had been identified post-transplant by PCR were used to determine the negative cut-off for the assay. Using the alkaline phosphatase probe method the mean ( f 1 SD.) The ODdg2 for these samples was 0.085 f 0.006. The results followed a normal distribution. Thus, samples not known to contain HCMV DNA produced a consistently low OD,,, under the assay conditions. As a consequence of these data, the negative cut-off for the assay was set at 0.104 OD,,, (mean + 3 SD.) and was used in all future analyses. Subsequently, known copy numbers of amplicon cloned into pUC18 were used in the range of l-1000 copies to generate a standard curve. The results shown in Fig. 1 demonstrate that the method was capable of detecting amplicons derived from four copies of HCMV target. The sensitivity of the alkaline phosphatase labelled probe method was marginally better than gel electrophoresis and had the advantage of providing a more objective measure of low levels of amplicon. Intra-assay variability was between 2-5%, however, interassay variability was higher
and usually ranged from 10 to 20% depending upon copy number of input DNA. Variability between 15 and 20% was observed with low copy numbers ( < 10 genomes). It should be noted that the optical density obtained for five copies of target HCMV (OD,,, = 0.157; see Fig. 1) was well separated from the negative cut-off. The use of this assay to identify true positive reactions in 168 samples yielding multiple bands within our existing PCR method was evaluated. An example of the amplicons produced in these samples is shown in Fig. 2 together with the alkaline phosphatase-oligonucleotide probe results. The distribution of results for all samples analysed is shown in Table 1. Of the 44 samples in which a non specific result was obtained at both dilutions of sample (5 pl/ 100 ~1 PCR mixture and 0.25 pl/lOO ~1 PCR mixture), 68% were positive following hybridization. In contrast, samples in which the undiluted sample yielded non-specific bands but which were negative for HCMV DNA at the 1 in 20 dilution (n = 101) yielded only 33.6% positive results. Finally, in samples that were negative neat but yielded non specific bands at 1 in 20 (n = 23), the true positive rate was 21.7%. The mean OD,,, for samples that were positive was 0.63 indicating that these samples
D. Gor et al. /Journal
148
of Virological Methods 61 (1996) 145-150
6
A
nnNnnNn --_-_++ ++
nNnnNn -I++
II
--_
_
++
++
n _I
a
I_
nn +++--
NP
a
Fig. 2. Results of gel analysis of HCMV amplicons compared with the results using the alkaline phosphatase conjugated oligonucleotide probe. In panel A the results for a selection of positive and negative clinical samples are shown. Each sample was analysed neat (5 ~1) and at a 1 in 20 dilution (see Materials and methods). The results from probing are shown by the positive and negative signs above each lane. The open arrow indicates the migratory position of the 149 bp amplicon, N, negative control. In panel B the results for a selection of samples yielding non-specific amplicons are shown analysed by gel electrophoresis and by the alkaline phosphatase method. The descriptors are the same as those used in panel A, p, positive control (50 genomes).
contained a reasonable quantity of target HCMV molecules. Overall, only 2% of samples yielded results that were within 4 standard deviations from the mean OD,,, of the negative samples.
4. Discussion The sensitivity of the PCR depends upon many variables including primer sequences, length of amplified sequence and genetic composition of the sequence. We have previously described the utility of two oligonucleotides specific for the glycoprotein B gene of HCMV that are highly sensitive and result in the amplification of the majority of HCMV strains present in the population (Kidd et al., 1993; Darlington et al., 1991; Roy et al., 1993). However, due to the inherent nature of a minority of samples, especially blood DNA ex-
tracts, the resulting amplicons following PCR comprise a range of molecular sizes. In such instances, it is difficult to ascertain whether the sample contained the target sequence and hence to provide a diagnostic result. In order to circumvent this problem we have used a probe oligonudirectly conjugated to alkaline cleotide phosphatase to detect biotinylated amplicons immobilised on streptavidin coated microtitre plates. The results from plasmid reconstruction experiments demonstrated that this process resulted in the detection of amplicons derived from as little as four copies of target sequence. Importantly, the optical density obtained from either four or five copies of target was substantially higher than the mean optical density obtained for clinical samples not known to contain HCMV. The direct detection of amplicons using alkaline phosphatase conjugated oligonucleotide probes has been described
D. Gor et al. /Journal
of Virological Methods 61 (1996) 145-150
149
Table 1 Distribution of results using the alkaline phosphatase labelled oligonucleotide probe Primary results
Number of samples
Neat
l/20
NSB”
NSB
44
NSB
Neg
10’1
Neg
NSB
23
Alkaline phosphatase probe results Neat
l/20
Posh Pos Neg’ Neg Pos Pos Neg Neg Pos Pos Neg Neg
Pos Neg Pos Neg Pos Neg Pos Neg Pos Neg Pos Neg
Number of samples
Mean OD,,, (range)
8 14 6 16 0 34 0 67 0 4 2 17
1.414 (0.107-3.424) 0.842 (0.105-2.900) 0.626 (0.105-3.596) 0.052 (0.001~0.104) N/Ad 0.564 (0.1 l-2.972) N/A 0.029 (0.006-0.080) N/A 0.486 (0.116-0.863) 0.131 (0.109-0.153) 0.022 (0.002-0.047)
Samples (168) yielding non specific banding results at either neat, 1 in 20 or both concentrations using conventional agarose gel electrophoresis/ethidium Ibromide staining of amplicons (primary result). The mean optical densities in the Pos/Neg and Neg/Pos combinations were calculated for the positive samples only. “Non-specific bands. bPositive. “Negative. dNot applicable.
previously (Cano et al., 1993; Whitby and Garson, 1995) although in some cases the sensitivity of detection has been limiting (for example, Rasmussen et al., 1993). This may have been a consequence of an insensitive PCR process used in some of these studies, although a more likely explanation may reiside in the methods used for detecting the hybrild product. In our assay we have utilised an amplification system based on the conversion of NADlPH to NADH which is used in the subsequent amplification step using diaphorase and alcohol dehydrogenase. The resulting redox cycle prolduces coloured formazan and leads to a signal amplification of approximately lOO-fold. Under these conditions, amplicons from four copies of target sequence were reliably detected. Such quantities of product are at the limit of detection in conventional gel electrophoresis systems which, in our hands, requires amplicons from 5-10 copies of target sequence to produce easily visible products after ethidium bromide staining. Whilst the: dynamic range of this assay does not approach chemiluminescent detection of
alkaline phosphatase conjugated oligonucleotides (Whitby and Garson, 1995) it has the advantage of requiring no new instrumentation over existing enzyme linked immunoassays whilst still maintaining sensitivity. In order to evaluate the usefulness of this assay we analysed a series of clinical specimens that yielded indeterminate results in a conventional gel electrophoresis based amplicon detection system. The results confirmed that in the majority of cases a finite result could be obtained using the alkaline phosphatase conjugated oligonucleotide approach. Interestingly, 68% of samples in which the primary analyses at both neat and 1 in 20 dilutions yielded non-specific bands were true positive reactions whereas only 33.6 and 21.7% were true positives in the combination of non-specific neat, negative at 1 in 20 or vice versa, respectively. Hence, samples that were negative for HCMV either undiluted or diluted 1 in 20 in our existing assay were more likely to be true negatives than samples yielding non-specific bands at both concentrations of analyte. These samples account for
150
D. Gor et al. /Journal
of Virological Methods 61 (1996) 145-150
approximately 85% of samples yielding non-specific reaction products. As a consequence of these data we incorporated this methodology into the laboratory for the assessment of samples yielding non-specific results. At present, the comparative costs of performing gel analysis versus the alkaline phosphatase method precludes the more widespread use of the latter approach in routine PCR surveillance for HCMV in transplant recipients and AIDS patients.
Acknowledgements This work was supported by the UK Medical Research Council and the Wellcome Foundation.
References Birkenmeyer, L.G. and Mushahwar, I.K. (1994) Detection of hepatitis A, B and D virus by the polymerase chain reaction. J. Virol. Methods 49, 101-l 12. Ballet, C., Vignoli, C., Freney, J., Gevaudan, M.J., Galicher, V. and de Micco, P. (1992) Identification of Mycobacterium tuberculosis and Mycobacterium avium by non-radioactive probes: evaluation of the snap syngene system. Annal. Biol. Clin. 50, 599-601. Cano, R.J., Torres, M.J., Klem, R.E., Palomares, J.C. and Casadesus, J. (1992) Detection of salmonellas by DNA hybridization with a fluorescent alkaline phosphatase substrate. J. Appl. Bacterial. 72, 393-399. Cano, R.J., Rasmussen, S.R., Sanchez Fraga, G. and Palomares, J.C. (1993) Fluorescent detection-polymerase chain reaction (FD-PCR) assay on microwell plates as a screening test for salmonellas in foods. J. Appl. Bacterial. 75, 247-253. Darlington, J., Super, M., Patel, K., Grundy, J.E., Griffiths, P.D. and Emery, V.C. (1991) Use of the polymerase chain reaction to analyse sequence variation within a major
neutralizing epitope of glycoprotein B (gp58) in clinical isolates of human cytomegalovirus. J. Gen. Virol. 72, 198551989. Garson, J.A. (1994) The polymerase chain reaction and hepatitis C virus diagnosis. FEMS Microbiology Reviews 14, 2299239. Gravitt, P.E. and Manos, M.M. (1992) Polymerase chain reaction-based methods for the detection of human papillomavirus DNA. IARC Scientific Publications 121- 133. Huang, C.H. and Jungkind, D.L. (1991) Non-radioactive DNA probe for the rapid identification of Mycobacterium avium complex from clinical specimens. Mol. Cell. Probes 5, 2777280. Kaye, S., Loveday, C. and Tedder, R.S. (1992) A microtitre format point mutation assay: application to the detection of drug resistance in human immunodeficiency virus type-1 infected patients treated with zidovudine. J. Med. Virol. 37, 241-246. Kidd, I.M., Fox, J.C., Pillay, D., Charman, H., Griffiths, P.D. and Emery, V.C. (1993) Provision of prognostic information in immunocompromised patients by routine application of the polymerase chain reaction for cytomegalovirus. Transplantation 56, 867-871. Rasmussen, S.R., Rasmussen, H.B., Larsen, M.R., Hoff-Jorgensen, R. and Cano, R.J. (1994) Combined polymerase chain reaction-hybridization microplate assay used to detect bovine leukemia virus and Salmonella. Clin. Chem. 40, 200&205. Rasmussen et al., 1993. Author please provide details. Roy, D.M., Grundy, J.E. and Emery, V.C. (1993) Sequence variation within neutralizing epitopes of the envelope glycoprotein B of human cytomegalovirus: comparison of isolates from renal transplant recipients and AIDS patients. J. Gen. Virol. 74, 2499-2505. Thorne, G.M., Macone, A. and Goldmann, D.A. (1990) Enzymatically labelled nucleic acid (NA) probe assays for detection of Campylobacter spp. in human faecal specimens and in culture. Mol. Cell Probes 4, 133-142. Whitby, K. and Garson, J.A. (1995) Optimisation and evaluation of a quantitative chemiluminescent polymerase chain reaction assay for hepatitis C virus RNA. J. Virol. Methods 51, 75-88. Whitley, R.J. and Lakeman, F. (1995) Herpes simplex virus infections of the central nervous system: therapeutic and diagnostic considerations. Clin. Infect. Dis 20, 414-420.