341
Journal of Virological Methods, 40 (1992) 341-356 0 1992 Elsevier Science Publishers B.V. / All rights reserved / 0166-0934/92/$05.00 VIRMET 01368
A new method for measuring reverse transcriptase activity by ELISA Josef Eberle” and Rudolf
Seiblb
“Max volt Pettenkofer-Institut, Ludwig-Maximilians-University of Mtinchen, ‘Boehringer Mannheim GmbH, Biochemical Research Centre, Penzberg (Germany) (Accepted
12 March 1992)
Summary A new and sensitive assay of reverse transcriptase (RT) activity of retroviruses measures the incorporation of digoxigenin-labelled dUTP in newly synthesized DNA instead of radioactively labelled (3H- or 32P-)dTTP. To avoid difficulties associated with separation of non-incorporated nucleotides from the newly synthesized DNA, biotin-labelled dUTP is added to the reaction mixture in very low concentrations. After reverse transcription, the newly synthesized, doubly labelled DNA is immobilized on streptavidin-coated ELISA wells and evaluated photometrically by binding of peroxidaseconjugated anti-digoxigenin-antibodies (sheep) and subsequent colour development with 2,2’-azino-di[3-ethylbenzthiazolin-sulfonate(6)] (ABTSR) as substrate. For better standardization, it is suggested that RT activity is given in units (one unit of RT is the amount of enzyme incorporating one nanomole of labelled dNTP in 10 min at 37°C into an acid precipitable DNA) rather than in cpm (counts per minute). The method is specific and easy to perform. Reverse transcriptase;
ELISA; Retrovirus; DNA-polymerase;
HIV
Introduction Assays for reverse transcriptase (RT) activity have become accepted techniques for the detection and quantification of retroviruses in cell cultures (Poeisz et al., 1980; Barre-Sinoussi et al., 1983). RT testing has been used to Correspondence to: J. Eberle, Max von Pettenkofer-Institut, Miinchen 2, Germany.
HIV-Labor, Pettenkoferstrasse
9a, D-8000
348
detect residual infectivity after virus inactivation (Barre-Sinoussi et al., 1985) and to screen inhibitors (Wu et al., 1988). Although reaction conditions and methods have been improved (Hoffman et al., 1985, Somogyi et al., 1990) RT testing in general is laborious, poorly standardized and expensive. Recently, new non-isotopic methods for testing RT have been described (Lee et al., 1990, Porstmann et al., 1991). Our own experience with isolating HIV from peripheral blood lymphocytes of infected patients and with monitoring permanent lymphocytic cell cultures productively infected with HIV-l or HIV2 led us also to develop a new RT ELISA capable of detecting RT activity with at least the sensitivity of standard isotopic RT tests and with a minimum of handling steps.
Materials and Methods Supernatants of HIV-l- or HIV-2-infected H9 cells were the source for HIV virions, either directly or after pelleting by ultracentrifugation (1 h at 85 000 x g; 36000 rpm in a 50 Ti Beckman rotor). As an alternative to ultracentrifugation, concentrating virions by PEG precipitation gives comparable results (data not shown). To determine the sensitivity of the new RT test, avian myeloblastosis virus reverse transcriptase (AMV-RT, Boehringer Mannheim, Germany) with an activity of 25 U/p1 was used. Serial dilutions of the enzyme were prepared in lysis buffer. To correlate the standard 3H-dTTP RT isotopic test and the RT ELISA, 204 supernatants of peripheral blood lymphocyte cultures of anti-HIV-negative and anti-HIV-positive patients were used. Special reagents required for the RT ELISA (poly(rA) . (dT)r5, digoxigeninl l-2’-desoxyuridin-5’-triphosphate, biotin-16-2’-desoxyuridin-5’-triphosphate, anti-digoxigenin-POD, BM-blocking reagent and streptavidin-coated ELISA plates) were supplied by Boehringer. RT ELISA
procedure
In addition to deoxythymidine-5’-triphosphate (dTTP), digoxigenin- 11-2’deoxyuridin-5’-triphosphate (DIG-dUTP, Miihlegger et al., 1990) and biotin16-2’-deoxyuridin-5’-triphosphate (BIO-dUTP) were included in the classical RT reaction. After up to 24 h RT reaction time, the newly synthesized DNA was trapped on streptavidin-coated ELISA plates, and was then detected immunologically by binding peroxidase-labelled anti-digoxigenin sheep antibodies (anti-DIG-POD, Kessler, 1991) with subsequent colour development with ABTSR (2,2’-azino-di[3-ethylbenzthiazolin-sulfonate(6)]) as substrate (Fig. 1). Virions from cell culture supernatants were pelleted by ultracentrifugation. Concentrated virus particles were lysed in 40 ~1 lysis buffer (Tris-HCl 50 mM,
349
ELfA
well,
created
Fig. I. The principle of the RT ELISA. Symbols DIG i‘ _C : DIG-dUTP BJO : BIO-dUTP T : dTTP J.L : rATP I : streptavidin + : ABTSR (substrate) POD 1 : anti-digoxigenin-POD (conjugate).
wrth
streptavldrn
used in the figure have the following
meaning:
pH 7.8; Triton X-100 0.5%; DTT 2.5 mM, KC1 80 mM, EDTA 0.75 mM). For the direct examination of virus-containing liquids without prior concentration, 35 ~1 of supernatant were mixed with 5 11 of eight times concentrated lysis buffer. After 30 min at room temperature or up to 1 wk at 4”C, 40 ~1 of lysed material was mixed with 20 ~1 of RT reaction mix (poly(rA) . (dT)i5, 750 AzGO U/l, where one AzeO U means absorbance of 1.O at 260 nm 2: 47 pg/ml DNA; dTTP 8.3 MM; DIG-dUTP 2.5 PM; RIO-dUTP 125 nM; Tris-HCl.50 mM, pH 7.8; DTT 10 mM; KC1 290 mM; MgC& 30 mM) and incubated for up to 24 h at 37°C in a moist chamber. After the RT reaction, the whole volume was transferred to a str~ptavidin-moated microtitre plate and allowed to adhere for 1 h. After 5 washings with 0.15 M NaCl, 200 ~1 of conjugate dilution (200 U/l anti-DIG-POD in conjugate buffer, i.e. 0.1 M sodium phosphate (pH 7.4), NaCl 0.I5 M, Tween 20 0.1%, EDTA 0.5 mM, BM-Blocking reagent 1“A) were added to each well and incubation was continued for another h at 37°C. Unbound antibody was removed by five further washings. For the substrate step, 200 ~1 of ABTSR solution (1 g/l of ABTSR in ABTS buffer) were added, and colour development was quantified at 405 nm in an ELISA reader (Behring ELISA Processor II) after 1 h of incubation at 37°C {Fig. 2). The final concentrations during the RT reaction in 60 ~1 volume were: Tris-WC1 50 mM (pH 7.8); DTT 5 mM; KC1 150 mM; Triton X-100 0.33%; EDTA 0.5 mM; MgCI2 10 mM; poly(rA) +(dT)is, 250 A260 U/l ;eu z 700 ng; dTTP 2.77 PM; DIG-dUTP 833 nM; BIO-dUTP 42 nM.
Ultrac~ntrifuge-pelleted virions from cell culture supernatants were treated with lysis buffer (Tris-HCl 10 mM (pH 7.3); Triton X-100 2%; EDTA I mM; DTT 5 mM; KC1 0.6 M) for at least 30 min at room temperature, and the other reagents for the RT reaction were then added to obtain the following final concentrations: MgC12 5 mM; DTT 2.4 mM; KC1 60 mM; [methyl-3H]dTTP 2 PM, with a specific activity of 30 Ci/mmol; poly(rA) f (dT)is 200 AzhO U/l ;eu
350
RT reaction 4O&lL i2OpL
37OC, 24 h incubation
lysed virus material reaction mix
RT
transfer to streptavidin-coated ELISA plate i
RT ELISA 37*C, 1 h incubation
binding of total volume to streptavidincoated wells
5 washings *
0.15M NaCl
37*C, 1 h incubation
2OOpL anti-DIG-POD (200 U/L in O.lM sodium phosphate pH 7.4, NaCf 0.15M EDTA 0.5mM, Tween-20 O.l%, BM-Blocking reagent 1%)
5 washings *
0.15M NaCl
37OC, 1 h incubation
200~L substrate (1 g/L ABTSR in ABTSR
reaction
buffer) 405 nm
OD measuring l
the plate must be almost
dry after the fifth
washing Fig. 2. Detailed
procedures
fortheRT ELISA.
560 ng/60 ~1. After 90 min at 37°C 50 ~1 were pipetted onto DEAE ion exchange filter paper disks (Whatman DE81) and dried for 20 min at 80°C. Non-incorporated nucleotides were removed by three washes with an excess volume of 0.5 M disodium hydrogen phosphate, followed by one wash with distilled water and a final washing step with 70% ethanol. Thereafter, the filter disks were completely dried at 80°C transferred to a plastic scintillation tube, AquasolTM (NEN research products, Boston) was added and the amount of incorporated “H-dTTP in the synthesized DNA was quantized in a p-counter. p24 antigen HA For comparison of the sensitivity for monitoring HIV cell cultures, a commercial p24 antigen EIA (Abbott Labs, North Chicago) was used according to the manufacturer’s recommendations.
351
Results For initial evaluation of the new RT ELISA, serial dilutions of known positive materials were used. We started with concentrated material representing pelleted virions from 1 ml culture supernatant, at a reaction time of 90 min. Dilutions down to l/25 gave a maximal signal, followed by an almost linear decrease of the signal at higher dilutions and with slight tailing for high dilutions (Fig. 3). Undiluted mock pellets from negative materials gave signals of below 0.15 OD at 405 nm. OD405 nm
3.0 2.0
1.0 0.5 0.2.
1’5
1125
l/50
l/l25
11250 11500 l/1000
dilution
Fig. 3. RT activities given as optical densities (OD) at 405 nm (after 90 min RT reaction followed by the RT ELISA) of different dilutions of pelleted HIV-l virions from 1 ml cell culture supernatant. The open squares mark the highest and lowest values of 8 independent measurements on at least 4 test occasions. As negative control, a mock pellet preparation of non-infected H9 cell cultures was used at a l/25 dilution, and gave consistent values below 0.15 OD.
In the next step, 204 cell culture supernatants from virus isolation attempts were monitored in parallel with the RT ELISA and the standard RT assay, both for a RT reaction time of 90 min. Ultracentrifuge pellets from 3-5 ml of supernatants were resuspended in lysis buffer and equal amounts used for both tests. The results of both tests agreed well, i.e. all positives by one test were positive by the other and vice versa. For quantitative comparison of the two tests, the Pearson correlation coefficient was calculated as 0.943 (Fig. 4). As prolongation of the RT reaction time for up to 24 h resulted in no loss of specificity and increased the sensitivity of the RT ELBA by a factor of about ten (data not shown), further experiments were carried out with 24 h reaction time. Under these conditions, the RT ELISA detected as few as 1 mU of commercially available AMV-RT, corresponding to 1 ~1 of a l/25 000 dilution of the supplied stock solution (Table I). Encouraged by this high sensitivity, RT activity was examined in supernatants of HIV-producing cell cultures (35 ~1 supernatant mixed with 5 ~1 of eight times lysis buffer), and the RT ELISA
352
RT ELISA OD‘lwJnm 2.0
-
L
n
I
AA
1.5
-
1.0
-
0.8
-
06
-
0.4
-
0.1
-
A A
1
A
A
* AA
I 0.2
x 10s STANDARD
ocr”~rS RT
. 5.0
. 2.0
I 1.0
1 0.5
PER
I 10.0
. 20.0
MINUTE
ASSAY
Fig. 4. Comparison between RT ELISA signals (OD) and standard RT test (cpm), with 90 min RT reaction for both assays. ‘A’ means I observation, ‘B’ 2 observations, ‘C’ 3 observations, ‘D’ 4 observations, ‘E’ 5 observations. Pearson correlation calculated from 204 materials tested in parallel was 0.943.
detected RT activity both from virus-producing permanent cell lines and in most of the virus-positive peripheral lymphocyte cultures without prior concentration. During this series, the signal seemed to be lower than expected in some instances, and the signal reduction was found to be caused by fetal calf I
TABLE Serial
IO-fold
dilutions
Units
of AMV-RT
10 IU 100 mU 10 mU 5mU 2mU 1 mU Negative
of avian
myeloblastosis
virus
RT (AMV-RT)
RT ELISA
were tested
by RT ELISA
(OD)
>2.5 >2.5 >2.5 2.11 1.05 0.77 0.46 control
0.20
OD results given are the mean of duplicate determinations of at least two independent One unit of T7 DNA polymerase (Pharmacia) was used as a negative control.
experiments.
353 TABLE 2 Serial IO-fold dilutions of cell culture supernatants of two HIV-l and one HIV-2 isolates were tested by p24 antigen EIA (200 ~1) and compared to RT activities measured by RT ELISA of serial IO-fold dilutions of pelleted virions from 800 ~1 of the same supernatants HIV-I (M-899)
OD IO0
10-l IOF 10F3 10F4 10-s
RT ELISA
p24 antigen EIA
> 2.0 > 2.0 > 2.0 1.637 0.207 0.065
pg/ml
% 330 z 33
OD > 2.5 > 2.5 > 2.5 0.545 0.120 nd.
mU RT/ml
C 1.5
HIV-l (M-5180) 100 > 2.5 10-1 10-’ 10-s lO-4
3, 2.0 1.662 0.375 0.069
> 2.5 > 2.5 0.296 0.108
HIV-2 (M-l 1971) 10° lo-’ 10-2 lop3
1.239 0.811 0.265 0.058
1.431 > 2.5 1.162 0.189
> 2.0
The cutoff OD for p24 antigen EIA was 0.09, for RT ELISA 0.3.
serum in the culture medium which led to a marked signal reduction at concentrations of above 3% (data not shown). To establish whether RT testing is comparable to the sensitivity of a p24 antigen assay, another series of experiments (Table 2) were carried out. Serial IO-fold dilutions of supernatants of two different HIV-l isolates (one with growth characteristics similar to HTLV-IIIB/LAV called MVP-899 and the other, MVP-5180, an isolate from Cameroon) and one HIV-2 isolate (MVP11971) were tested with a p24 antigen EIA (200 ~1 sample/assay, cutoff OD 0.09). Ultracentrifuge pellets of 2 ml of supernatants of the same cultures were lysed in 100 ~1 of lysis buffer, and serial lo-fold dilutions prepared. 40 ,ul of these materials were tested in the RT ELISA (cutoff OD 0.3). For MVP-899, the p24 antigen EIA was reactive until lop4 with a signal/cutoff ratio (S/CO) of 2.3 whereas the RT ELISA reacted until IOK (S/CO 1.8). For MVP-5180, the p24 antigen EIA was reactive until lop3 (S/CO 4.2) and the RT ELISA was positive until lop2 (S/CO > 8.3) and still gave a borderline signal with the 1O-3 dilution (S/CO 0.99). With the HIV-2 isolate MVP-11971, both tests were positive until 10p2, whereas the RT ELISA showed a slightly higher S/CO compared to the p24 antigen EIA (3.9 versus 2.9).
354
Discussion The advantage of the RT ELISA is that a laborious and time-consuming additional separation between incorporated and free labelled nucleotides is not necessary. The competition of free BIO-dUTP with incorporated BIO-dUTP in the newly synthesized DNA is avoided by the low concentration of BIO-dUTP present in the reaction mixture and the excess binding capacity of the streptavadin-coated ELISA well. A further important aspect is the relation between ‘cold’ nucleotides (dTTP) and labelled nucleotides (DIG-dUTP and BIO-dUTP): The intensity of the signal is proportional to the amount of incorporated DIG-dUTP but if an excess of labelled nucleotides is present during the RT reaction, the RT does not work, probably due to sterical hindrance, so the given concentrations and ratios of the nucleotides are critical. These have been assessed with an artificial poly(rA)-template and somewhat different ratios might be needed with other templates (Kessler et al., 1990). With this new RT ELISA, RT activities of several hundred samples can be determined with comparatively little work, although of course the preparation of samples must also be considered. The test is reliable, and the RT ELISA can be used by laboratories without permission to handle radioactive reagents. In addition, nothing more than the standard equipment available in a routine virology laboratory is required, and reagents are less harmful and cheaper than those for the standard assay. The RT ELISA may be suitable for large-scale screening of potential RT inhibitors because it can easily be automated. Under certain circumstances, where RT activity has to be quantified, the short linear range of the RT ELISA (where the amount of RT is directly proportional to the OD) might seem disadvantageous but the low costs, easy handling and high sensitivity overcome the problem of testing serial dilutions. Alternatively, the concentration of DIG-dUTP could be slightly reduced, resulting in an extension of the linear range of the RT ELISA but reduction of sensitivity. Under conditions where specific antigen assay has not been established, RT testing may be important for further characterization of virus isolates or the quantitation of virions, and the RT ELISA makes widespread use of RT testing possible. The RT ELISA is an interesting supplement or even an alternative to p24 antigen testing for HIV research and diagnosis. In most instances, however, concentration of virus particles by a pelleting step is needed to achieve a high level of sensitivity. For calculating the relative sensitivity of the two assays, the different sample volumes must be considered (200 ~1 for the p24 antigen EIA and 40 ~1 for the RT ELISA, representing pelleted virions from 800 ~1 of cell culture supernatant). With some HIV-1 isolates, the titre of the p24 antigen EIA is about one order of magnitude higher than for the RT ELISA, so the sensitivity of the latter is lower by a factor of about 40, but with other HIV-I isolates the difference in sensitivity between the test systems was smaller. The ability to detect HIV-2 in cell culture is about the same with both tests. In practice, however, p24 antigen EIA results from supernatants will be compared
355
with RT signals of virus pellets from at least one or more millilitres of supernatant, and then sensitivity of RT ELISA is only about one order of magnitude or even less below the sensitivity of p24 antigen EIA for HIV-l isolates, and is at least as sensitive as EIA in detecting HIV-2. This difference in sensitivity of the p24 antigen test for HIV-l compared to HIV-2 is due to its type specificity. Feorino et al. (1987) compared the sensitivities of a p24 antigen EIA with a standard RT assay and found a lOO-fold lower sensitivity of the latter. This observation agrees well with the results with the RT ELISA, keeping in mind the about IO-fold increased sensitivity of the 24 h RT ELISA over the standard (90 min) RT assay and the about lo-fold higher sensitivity of the p24 antigen EIA with some HIV-I isolates. For some experiments, e.g. the quantitation of inhibition of HIV-l retroviral protease during virus assembly and maturation, measuring RT activity of released particles rather than quantifying p24 antigen content is a much better surrogate marker for remaining infectivity. Many time-consuming and expensive titration experiments could be replaced by this cheaper and more rapid alternative.
Acknowledgements The repeated complaints of our laboratory technicians about some tedious features of the standard RT assay provided the impulse to look for a better solution. Staff of the Biomedical Research Centre (Werk Penzberg and Werk Tutzing) from Boehringer Mannheim GmbH contributed ideas during the development of the test and supplied the required reagents. We thank Bernd Lorbeer for the statistical calculations. This paper is dedicated in memory of Professor Friedrich Deinhardt, our highly esteemed teacher and mentor.
References Barre-Sinoussi, F., Chermann, J-C., Rey, F., Nugeyre, M.T., Charmare, S., Gruest, J., Dauguet, C., Axler-Blin, C., Brun-Vezinet, F., Rouzioux, C., Rozenbaum, W. and Montagnier, L. (1983) Isolation of a T-lymphotropic retrovirus from a patient at risk for acquired immune deficiency syndrome (AIDS). Science 220, 868-871. Barre-Sinoussi, F., Nugeyre, M-T. and Chermann, J-C. (1985) Resistance of AIDS virus at room temperature. Lancet 2, 721-722. Feorino, P., Forrester, B., Schable, C., Warfield, D. and Schochetman, G. (1987) Comparison of antigen assay and reverse transcriptase assay for detecting human immunodeficiency virus in culture. J. Clin. Microbial. 25, 23442346. Hoffman, A.D., Banapour, B. and Levy, J.A. (1985) Characterization of the AIDS-associated retrovirus reverse transcriptase and optimal conditions for its detection in virions. Virology 147, 3266335. Kessler, C. (1991) The digoxigenin: anti-digoxigenin (DIG) technology - a survey on the concept and realization of a novel bioanalytical indicator system. Mol. Cell. Probes 161-205.
356 Kessler, C., Hiiltke, H.J., Seibl, R., Burg, J. and Miihlegger, K. (1990) Non-radioactive labelling and detection of nucleic acids, Part I. Biol. Chem. Hoppe-Seyler 371, 917-927. Miihlegger, K., Huber, E., von der Eltz, H., Riiger, R. and Kessler, C. (1990) Non-radioactive labelling and detection of nucleic acids. Part IV. Biol. Chem. Hoppe-Seyler 371, 953-965. Lee. M., Kusakabe, H., Reisinger, D., Takasaki, T., Sano, K. and Imagawa, D. (1990) Detection of HIV by a non-radioactive sensitive reverse transcriptase assay. Sixth International Conference on AIDS, San Francisco, USA, 2, 321, Abstract 1028. Poiesz, B.J., Ruscetti, R.W., Gazdar, A.F., Bunn, P.A., Minna, J.D. and Gallo, R.C. (1980) Detection and isolation of type C retrovirus particles from fresh and cultured lymphocytes of a patient with cutaneous T-Cell lymphoma. Proc. Nat]. Acad. Sci. USA 77, 7415-7419. Porstmann, T., Meissner, K., Glaser, R., Diipel, S.H. and Sydow, G. (1991) A sensitive nonradioactive assay specific for HIV-I associated reverse transcriptase. J. Viral. Methods 31, 181188. Somogyi, P.A., Gyuris, A. and FBldes, I. (1990) A solid phase reverse transcriptase micro-assay for the detection of human immunodeficiency virus and other retroviruses in cell culture supernatants. J. Virol. Methods 27, 269-276. Wu. J.C., Chernow, M., Boehme, R.E., Suttmann, R.T., McRoberts, M.J., Prisbe, E.J., Matthews, T.R., Marx, P.A., Chuang, R.Y. and Chen, MS. (1988) Kinetics and inhibition of reverse transcriptase from human and simian immunodeficiency viruses. Antimicrob. Agents Chemother. 32, 1887-1890.
Journal of Virological Methods, 40 (1992) 341-356 0 1992 Elsevier Science Publishers B.V. / All rights reserved / 0166-0934/92/$05.00 VIRMET 01368
A new method for measuring reverse transcriptase activity by ELISA Josef Eberle” and Rudolf
Seiblb
“Max volt Pettenkofer-Institut, Ludwig-Maximilians-University of Mtinchen, ‘Boehringer Mannheim GmbH, Biochemical Research Centre, Penzberg (Germany) (Accepted
12 March 1992)
Summary A new and sensitive assay of reverse transcriptase (RT) activity of retroviruses measures the incorporation of digoxigenin-labelled dUTP in newly synthesized DNA instead of radioactively labelled (3H- or 32P-)dTTP. To avoid difficulties associated with separation of non-incorporated nucleotides from the newly synthesized DNA, biotin-labelled dUTP is added to the reaction mixture in very low concentrations. After reverse transcription, the newly synthesized, doubly labelled DNA is immobilized on streptavidin-coated ELISA wells and evaluated photometrically by binding of peroxidaseconjugated anti-digoxigenin-antibodies (sheep) and subsequent colour development with 2,2’-azino-di[3-ethylbenzthiazolin-sulfonate(6)] (ABTSR) as substrate. For better standardization, it is suggested that RT activity is given in units (one unit of RT is the amount of enzyme incorporating one nanomole of labelled dNTP in 10 min at 37°C into an acid precipitable DNA) rather than in cpm (counts per minute). The method is specific and easy to perform. Reverse transcriptase;
ELISA; Retrovirus; DNA-polymerase;
HIV
Introduction Assays for reverse transcriptase (RT) activity have become accepted techniques for the detection and quantification of retroviruses in cell cultures (Poeisz et al., 1980; Barre-Sinoussi et al., 1983). RT testing has been used to Correspondence to: J. Eberle, Max von Pettenkofer-Institut, Miinchen 2, Germany.
HIV-Labor, Pettenkoferstrasse
9a, D-8000
348
detect residual infectivity after virus inactivation (Barre-Sinoussi et al., 1985) and to screen inhibitors (Wu et al., 1988). Although reaction conditions and methods have been improved (Hoffman et al., 1985, Somogyi et al., 1990) RT testing in general is laborious, poorly standardized and expensive. Recently, new non-isotopic methods for testing RT have been described (Lee et al., 1990, Porstmann et al., 1991). Our own experience with isolating HIV from peripheral blood lymphocytes of infected patients and with monitoring permanent lymphocytic cell cultures productively infected with HIV-l or HIV2 led us also to develop a new RT ELISA capable of detecting RT activity with at least the sensitivity of standard isotopic RT tests and with a minimum of handling steps.
Materials and Methods Supernatants of HIV-l- or HIV-2-infected H9 cells were the source for HIV virions, either directly or after pelleting by ultracentrifugation (1 h at 85 000 x g; 36000 rpm in a 50 Ti Beckman rotor). As an alternative to ultracentrifugation, concentrating virions by PEG precipitation gives comparable results (data not shown). To determine the sensitivity of the new RT test, avian myeloblastosis virus reverse transcriptase (AMV-RT, Boehringer Mannheim, Germany) with an activity of 25 U/p1 was used. Serial dilutions of the enzyme were prepared in lysis buffer. To correlate the standard 3H-dTTP RT isotopic test and the RT ELISA, 204 supernatants of peripheral blood lymphocyte cultures of anti-HIV-negative and anti-HIV-positive patients were used. Special reagents required for the RT ELISA (poly(rA) . (dT)r5, digoxigeninl l-2’-desoxyuridin-5’-triphosphate, biotin-16-2’-desoxyuridin-5’-triphosphate, anti-digoxigenin-POD, BM-blocking reagent and streptavidin-coated ELISA plates) were supplied by Boehringer. RT ELISA
procedure
In addition to deoxythymidine-5’-triphosphate (dTTP), digoxigenin- 11-2’deoxyuridin-5’-triphosphate (DIG-dUTP, Miihlegger et al., 1990) and biotin16-2’-deoxyuridin-5’-triphosphate (BIO-dUTP) were included in the classical RT reaction. After up to 24 h RT reaction time, the newly synthesized DNA was trapped on streptavidin-coated ELISA plates, and was then detected immunologically by binding peroxidase-labelled anti-digoxigenin sheep antibodies (anti-DIG-POD, Kessler, 1991) with subsequent colour development with ABTSR (2,2’-azino-di[3-ethylbenzthiazolin-sulfonate(6)]) as substrate (Fig. 1). Virions from cell culture supernatants were pelleted by ultracentrifugation. Concentrated virus particles were lysed in 40 ~1 lysis buffer (Tris-HCl 50 mM,
349
ELfA
well,
created
Fig. I. The principle of the RT ELISA. Symbols DIG i‘ _C : DIG-dUTP BJO : BIO-dUTP T : dTTP J.L : rATP I : streptavidin + : ABTSR (substrate) POD 1 : anti-digoxigenin-POD (conjugate).
wrth
streptavldrn
used in the figure have the following
meaning:
pH 7.8; Triton X-100 0.5%; DTT 2.5 mM, KC1 80 mM, EDTA 0.75 mM). For the direct examination of virus-containing liquids without prior concentration, 35 ~1 of supernatant were mixed with 5 11 of eight times concentrated lysis buffer. After 30 min at room temperature or up to 1 wk at 4”C, 40 ~1 of lysed material was mixed with 20 ~1 of RT reaction mix (poly(rA) . (dT)i5, 750 AzGO U/l, where one AzeO U means absorbance of 1.O at 260 nm 2: 47 pg/ml DNA; dTTP 8.3 MM; DIG-dUTP 2.5 PM; RIO-dUTP 125 nM; Tris-HCl.50 mM, pH 7.8; DTT 10 mM; KC1 290 mM; MgC& 30 mM) and incubated for up to 24 h at 37°C in a moist chamber. After the RT reaction, the whole volume was transferred to a str~ptavidin-moated microtitre plate and allowed to adhere for 1 h. After 5 washings with 0.15 M NaCl, 200 ~1 of conjugate dilution (200 U/l anti-DIG-POD in conjugate buffer, i.e. 0.1 M sodium phosphate (pH 7.4), NaCl 0.I5 M, Tween 20 0.1%, EDTA 0.5 mM, BM-Blocking reagent 1“A) were added to each well and incubation was continued for another h at 37°C. Unbound antibody was removed by five further washings. For the substrate step, 200 ~1 of ABTSR solution (1 g/l of ABTSR in ABTS buffer) were added, and colour development was quantified at 405 nm in an ELISA reader (Behring ELISA Processor II) after 1 h of incubation at 37°C {Fig. 2). The final concentrations during the RT reaction in 60 ~1 volume were: Tris-WC1 50 mM (pH 7.8); DTT 5 mM; KC1 150 mM; Triton X-100 0.33%; EDTA 0.5 mM; MgCI2 10 mM; poly(rA) +(dT)is, 250 A260 U/l ;eu z 700 ng; dTTP 2.77 PM; DIG-dUTP 833 nM; BIO-dUTP 42 nM.
Ultrac~ntrifuge-pelleted virions from cell culture supernatants were treated with lysis buffer (Tris-HCl 10 mM (pH 7.3); Triton X-100 2%; EDTA I mM; DTT 5 mM; KC1 0.6 M) for at least 30 min at room temperature, and the other reagents for the RT reaction were then added to obtain the following final concentrations: MgC12 5 mM; DTT 2.4 mM; KC1 60 mM; [methyl-3H]dTTP 2 PM, with a specific activity of 30 Ci/mmol; poly(rA) f (dT)is 200 AzhO U/l ;eu
350
RT reaction 4O&lL i2OpL
37OC, 24 h incubation
lysed virus material reaction mix
RT
transfer to streptavidin-coated ELISA plate i
RT ELISA 37*C, 1 h incubation
binding of total volume to streptavidincoated wells
5 washings *
0.15M NaCl
37*C, 1 h incubation
2OOpL anti-DIG-POD (200 U/L in O.lM sodium phosphate pH 7.4, NaCf 0.15M EDTA 0.5mM, Tween-20 O.l%, BM-Blocking reagent 1%)
5 washings *
0.15M NaCl
37OC, 1 h incubation
200~L substrate (1 g/L ABTSR in ABTSR
reaction
buffer) 405 nm
OD measuring l
the plate must be almost
dry after the fifth
washing Fig. 2. Detailed
procedures
fortheRT ELISA.
560 ng/60 ~1. After 90 min at 37°C 50 ~1 were pipetted onto DEAE ion exchange filter paper disks (Whatman DE81) and dried for 20 min at 80°C. Non-incorporated nucleotides were removed by three washes with an excess volume of 0.5 M disodium hydrogen phosphate, followed by one wash with distilled water and a final washing step with 70% ethanol. Thereafter, the filter disks were completely dried at 80°C transferred to a plastic scintillation tube, AquasolTM (NEN research products, Boston) was added and the amount of incorporated “H-dTTP in the synthesized DNA was quantized in a p-counter. p24 antigen HA For comparison of the sensitivity for monitoring HIV cell cultures, a commercial p24 antigen EIA (Abbott Labs, North Chicago) was used according to the manufacturer’s recommendations.
351
Results For initial evaluation of the new RT ELISA, serial dilutions of known positive materials were used. We started with concentrated material representing pelleted virions from 1 ml culture supernatant, at a reaction time of 90 min. Dilutions down to l/25 gave a maximal signal, followed by an almost linear decrease of the signal at higher dilutions and with slight tailing for high dilutions (Fig. 3). Undiluted mock pellets from negative materials gave signals of below 0.15 OD at 405 nm. OD405 nm
3.0 2.0
1.0 0.5 0.2.
1’5
1125
l/50
l/l25
11250 11500 l/1000
dilution
Fig. 3. RT activities given as optical densities (OD) at 405 nm (after 90 min RT reaction followed by the RT ELISA) of different dilutions of pelleted HIV-l virions from 1 ml cell culture supernatant. The open squares mark the highest and lowest values of 8 independent measurements on at least 4 test occasions. As negative control, a mock pellet preparation of non-infected H9 cell cultures was used at a l/25 dilution, and gave consistent values below 0.15 OD.
In the next step, 204 cell culture supernatants from virus isolation attempts were monitored in parallel with the RT ELISA and the standard RT assay, both for a RT reaction time of 90 min. Ultracentrifuge pellets from 3-5 ml of supernatants were resuspended in lysis buffer and equal amounts used for both tests. The results of both tests agreed well, i.e. all positives by one test were positive by the other and vice versa. For quantitative comparison of the two tests, the Pearson correlation coefficient was calculated as 0.943 (Fig. 4). As prolongation of the RT reaction time for up to 24 h resulted in no loss of specificity and increased the sensitivity of the RT ELBA by a factor of about ten (data not shown), further experiments were carried out with 24 h reaction time. Under these conditions, the RT ELISA detected as few as 1 mU of commercially available AMV-RT, corresponding to 1 ~1 of a l/25 000 dilution of the supplied stock solution (Table I). Encouraged by this high sensitivity, RT activity was examined in supernatants of HIV-producing cell cultures (35 ~1 supernatant mixed with 5 ~1 of eight times lysis buffer), and the RT ELISA
352
RT ELISA OD‘lwJnm 2.0
-
L
n
I
AA
1.5
-
1.0
-
0.8
-
06
-
0.4
-
0.1
-
A A
1
A
A
* AA
I 0.2
x 10s STANDARD
ocr”~rS RT
. 5.0
. 2.0
I 1.0
1 0.5
PER
I 10.0
. 20.0
MINUTE
ASSAY
Fig. 4. Comparison between RT ELISA signals (OD) and standard RT test (cpm), with 90 min RT reaction for both assays. ‘A’ means I observation, ‘B’ 2 observations, ‘C’ 3 observations, ‘D’ 4 observations, ‘E’ 5 observations. Pearson correlation calculated from 204 materials tested in parallel was 0.943.
detected RT activity both from virus-producing permanent cell lines and in most of the virus-positive peripheral lymphocyte cultures without prior concentration. During this series, the signal seemed to be lower than expected in some instances, and the signal reduction was found to be caused by fetal calf I
TABLE Serial
IO-fold
dilutions
Units
of AMV-RT
10 IU 100 mU 10 mU 5mU 2mU 1 mU Negative
of avian
myeloblastosis
virus
RT (AMV-RT)
RT ELISA
were tested
by RT ELISA
(OD)
>2.5 >2.5 >2.5 2.11 1.05 0.77 0.46 control
0.20
OD results given are the mean of duplicate determinations of at least two independent One unit of T7 DNA polymerase (Pharmacia) was used as a negative control.
experiments.
353 TABLE 2 Serial IO-fold dilutions of cell culture supernatants of two HIV-l and one HIV-2 isolates were tested by p24 antigen EIA (200 ~1) and compared to RT activities measured by RT ELISA of serial IO-fold dilutions of pelleted virions from 800 ~1 of the same supernatants HIV-I (M-899)
OD IO0
10-l IOF 10F3 10F4 10-s
RT ELISA
p24 antigen EIA
> 2.0 > 2.0 > 2.0 1.637 0.207 0.065
pg/ml
% 330 z 33
OD > 2.5 > 2.5 > 2.5 0.545 0.120 nd.
mU RT/ml
C 1.5
HIV-l (M-5180) 100 > 2.5 10-1 10-’ 10-s lO-4
3, 2.0 1.662 0.375 0.069
> 2.5 > 2.5 0.296 0.108
HIV-2 (M-l 1971) 10° lo-’ 10-2 lop3
1.239 0.811 0.265 0.058
1.431 > 2.5 1.162 0.189
> 2.0
The cutoff OD for p24 antigen EIA was 0.09, for RT ELISA 0.3.
serum in the culture medium which led to a marked signal reduction at concentrations of above 3% (data not shown). To establish whether RT testing is comparable to the sensitivity of a p24 antigen assay, another series of experiments (Table 2) were carried out. Serial IO-fold dilutions of supernatants of two different HIV-l isolates (one with growth characteristics similar to HTLV-IIIB/LAV called MVP-899 and the other, MVP-5180, an isolate from Cameroon) and one HIV-2 isolate (MVP11971) were tested with a p24 antigen EIA (200 ~1 sample/assay, cutoff OD 0.09). Ultracentrifuge pellets of 2 ml of supernatants of the same cultures were lysed in 100 ~1 of lysis buffer, and serial lo-fold dilutions prepared. 40 ,ul of these materials were tested in the RT ELISA (cutoff OD 0.3). For MVP-899, the p24 antigen EIA was reactive until lop4 with a signal/cutoff ratio (S/CO) of 2.3 whereas the RT ELISA reacted until IOK (S/CO 1.8). For MVP-5180, the p24 antigen EIA was reactive until lop3 (S/CO 4.2) and the RT ELISA was positive until lop2 (S/CO > 8.3) and still gave a borderline signal with the 1O-3 dilution (S/CO 0.99). With the HIV-2 isolate MVP-11971, both tests were positive until 10p2, whereas the RT ELISA showed a slightly higher S/CO compared to the p24 antigen EIA (3.9 versus 2.9).
354
Discussion The advantage of the RT ELISA is that a laborious and time-consuming additional separation between incorporated and free labelled nucleotides is not necessary. The competition of free BIO-dUTP with incorporated BIO-dUTP in the newly synthesized DNA is avoided by the low concentration of BIO-dUTP present in the reaction mixture and the excess binding capacity of the streptavadin-coated ELISA well. A further important aspect is the relation between ‘cold’ nucleotides (dTTP) and labelled nucleotides (DIG-dUTP and BIO-dUTP): The intensity of the signal is proportional to the amount of incorporated DIG-dUTP but if an excess of labelled nucleotides is present during the RT reaction, the RT does not work, probably due to sterical hindrance, so the given concentrations and ratios of the nucleotides are critical. These have been assessed with an artificial poly(rA)-template and somewhat different ratios might be needed with other templates (Kessler et al., 1990). With this new RT ELISA, RT activities of several hundred samples can be determined with comparatively little work, although of course the preparation of samples must also be considered. The test is reliable, and the RT ELISA can be used by laboratories without permission to handle radioactive reagents. In addition, nothing more than the standard equipment available in a routine virology laboratory is required, and reagents are less harmful and cheaper than those for the standard assay. The RT ELISA may be suitable for large-scale screening of potential RT inhibitors because it can easily be automated. Under certain circumstances, where RT activity has to be quantified, the short linear range of the RT ELISA (where the amount of RT is directly proportional to the OD) might seem disadvantageous but the low costs, easy handling and high sensitivity overcome the problem of testing serial dilutions. Alternatively, the concentration of DIG-dUTP could be slightly reduced, resulting in an extension of the linear range of the RT ELISA but reduction of sensitivity. Under conditions where specific antigen assay has not been established, RT testing may be important for further characterization of virus isolates or the quantitation of virions, and the RT ELISA makes widespread use of RT testing possible. The RT ELISA is an interesting supplement or even an alternative to p24 antigen testing for HIV research and diagnosis. In most instances, however, concentration of virus particles by a pelleting step is needed to achieve a high level of sensitivity. For calculating the relative sensitivity of the two assays, the different sample volumes must be considered (200 ~1 for the p24 antigen EIA and 40 ~1 for the RT ELISA, representing pelleted virions from 800 ~1 of cell culture supernatant). With some HIV-1 isolates, the titre of the p24 antigen EIA is about one order of magnitude higher than for the RT ELISA, so the sensitivity of the latter is lower by a factor of about 40, but with other HIV-I isolates the difference in sensitivity between the test systems was smaller. The ability to detect HIV-2 in cell culture is about the same with both tests. In practice, however, p24 antigen EIA results from supernatants will be compared
355
with RT signals of virus pellets from at least one or more millilitres of supernatant, and then sensitivity of RT ELISA is only about one order of magnitude or even less below the sensitivity of p24 antigen EIA for HIV-l isolates, and is at least as sensitive as EIA in detecting HIV-2. This difference in sensitivity of the p24 antigen test for HIV-l compared to HIV-2 is due to its type specificity. Feorino et al. (1987) compared the sensitivities of a p24 antigen EIA with a standard RT assay and found a lOO-fold lower sensitivity of the latter. This observation agrees well with the results with the RT ELISA, keeping in mind the about IO-fold increased sensitivity of the 24 h RT ELISA over the standard (90 min) RT assay and the about lo-fold higher sensitivity of the p24 antigen EIA with some HIV-I isolates. For some experiments, e.g. the quantitation of inhibition of HIV-l retroviral protease during virus assembly and maturation, measuring RT activity of released particles rather than quantifying p24 antigen content is a much better surrogate marker for remaining infectivity. Many time-consuming and expensive titration experiments could be replaced by this cheaper and more rapid alternative.
Acknowledgements The repeated complaints of our laboratory technicians about some tedious features of the standard RT assay provided the impulse to look for a better solution. Staff of the Biomedical Research Centre (Werk Penzberg and Werk Tutzing) from Boehringer Mannheim GmbH contributed ideas during the development of the test and supplied the required reagents. We thank Bernd Lorbeer for the statistical calculations. This paper is dedicated in memory of Professor Friedrich Deinhardt, our highly esteemed teacher and mentor.
References Barre-Sinoussi, F., Chermann, J-C., Rey, F., Nugeyre, M.T., Charmare, S., Gruest, J., Dauguet, C., Axler-Blin, C., Brun-Vezinet, F., Rouzioux, C., Rozenbaum, W. and Montagnier, L. (1983) Isolation of a T-lymphotropic retrovirus from a patient at risk for acquired immune deficiency syndrome (AIDS). Science 220, 868-871. Barre-Sinoussi, F., Nugeyre, M-T. and Chermann, J-C. (1985) Resistance of AIDS virus at room temperature. Lancet 2, 721-722. Feorino, P., Forrester, B., Schable, C., Warfield, D. and Schochetman, G. (1987) Comparison of antigen assay and reverse transcriptase assay for detecting human immunodeficiency virus in culture. J. Clin. Microbial. 25, 23442346. Hoffman, A.D., Banapour, B. and Levy, J.A. (1985) Characterization of the AIDS-associated retrovirus reverse transcriptase and optimal conditions for its detection in virions. Virology 147, 3266335. Kessler, C. (1991) The digoxigenin: anti-digoxigenin (DIG) technology - a survey on the concept and realization of a novel bioanalytical indicator system. Mol. Cell. Probes 161-205.
356 Kessler, C., Hiiltke, H.J., Seibl, R., Burg, J. and Miihlegger, K. (1990) Non-radioactive labelling and detection of nucleic acids, Part I. Biol. Chem. Hoppe-Seyler 371, 917-927. Miihlegger, K., Huber, E., von der Eltz, H., Riiger, R. and Kessler, C. (1990) Non-radioactive labelling and detection of nucleic acids. Part IV. Biol. Chem. Hoppe-Seyler 371, 953-965. Lee. M., Kusakabe, H., Reisinger, D., Takasaki, T., Sano, K. and Imagawa, D. (1990) Detection of HIV by a non-radioactive sensitive reverse transcriptase assay. Sixth International Conference on AIDS, San Francisco, USA, 2, 321, Abstract 1028. Poiesz, B.J., Ruscetti, R.W., Gazdar, A.F., Bunn, P.A., Minna, J.D. and Gallo, R.C. (1980) Detection and isolation of type C retrovirus particles from fresh and cultured lymphocytes of a patient with cutaneous T-Cell lymphoma. Proc. Nat]. Acad. Sci. USA 77, 7415-7419. Porstmann, T., Meissner, K., Glaser, R., Diipel, S.H. and Sydow, G. (1991) A sensitive nonradioactive assay specific for HIV-I associated reverse transcriptase. J. Viral. Methods 31, 181188. Somogyi, P.A., Gyuris, A. and FBldes, I. (1990) A solid phase reverse transcriptase micro-assay for the detection of human immunodeficiency virus and other retroviruses in cell culture supernatants. J. Virol. Methods 27, 269-276. Wu. J.C., Chernow, M., Boehme, R.E., Suttmann, R.T., McRoberts, M.J., Prisbe, E.J., Matthews, T.R., Marx, P.A., Chuang, R.Y. and Chen, MS. (1988) Kinetics and inhibition of reverse transcriptase from human and simian immunodeficiency viruses. Antimicrob. Agents Chemother. 32, 1887-1890.
