Diagnostic reverse-transcription polymerase chain reaction kit for filoviruses based on the strain collections of all European biosafety level 4 laboratories

Abstract

A network of European biosafety level 4 laboratories has designed the first industry-standard molecular assay for all filoviruses species, based on the strain collections of all participants. It uses 5 optimized L gene primers and 3 probes, as well as an internal control with a separate detection probe. Detection limits (probit analysis, 95% detection chance) were as follows: Zaire ebolavirus, 487 copies/mL of plasma; Sudan ebolavirus Maleo, 586 copies/mL; Sudan ebolavirus Gulu, 1128 copies/mL; Cote d'Ivoire ebolavirus, 537 copies/mL; Reston ebolavirus, 4546 copies/mL; Lake Victoria marburgvirus Musoke, 860 copies/mL; and Lake Victoria marburgvirus Ravn, 1551 copies/mL. The assay facilitates reliable detection or exclusion screening of filovirus infections.

Figure 1

Design of the in-house assay serving as a basis for the test kit. A, Binding sites of oligonucleotides. Letters in the left column represent the species (Z, Zaire ebolavirus [ZEBOV]; IC, Cote d'Ivoire ebolavirus [CIEBOV]; RE, Reston ebolavirus [REBOV]; S, Sudan ebolavirus [SEBOV]; and M, Lake Victoria marburgvirus [MARV]), followed by a representative strain designation (GenBank accession nos. or sequences from in-house strain collections). Note that only representative sequences are shown. All sequences occurring more than once at each of the oligonucleotide binding sites have been deleted from the alignment. The alignment was updated in October 2006 with latest GenBank entries. The complete alignment can be retrieved from http://www.bni-hamburg.de/ebovalign. B, A probe matching ZEBOV (top panel), modified to adapt a single nucleotide mismatch as present in REBOV (probe EBOg, lower panel). No loss of signal occurred for ZEBOV (bottom panel), and the signal for REBOV improved (not shown; REBOV was not yet available at the time these experiments were performed). C, top panel, A probe containing 2 inosin residues at positions of variability but otherwise matching SEBOV (probe EBOSu), yielding a good detection signal (similar to the signal from a corresponding RNA concentration of ZEBOV; compare with panel B). C, middle panel, A probe of similar design, adapted to CIEBOV, yielding good signal for CIEBOV. However, because CIEBOV was also detectable with good efficiency with the EBOSu probe (C, bottom panel), EBOSu was used for both SEBOV and CIEBOV. D, Detection signals with oligonucleotides, as shown in panel A, in reactions containing low concentrations of virus RNA (flat lines denote controls; the figure shows signals with all probes and primers mixed in 1 reaction). E, Determination of working concentration for competitive internal control. A constant small amount of target gene RNA (solid line; 50 copies of ZEBOV transcript per assay) was amplified in the presence of increasing concentrations, as plotted on the X-axis, of the internal control (dotted line). Observed crossing point values were plotted on the Y-axis for both target genes (note that the virus and internal control carried different reporter dyes). Increasing values on the Y-axis indicate decreasing amplification efficiencies. From 40 copies of internal control onward, stable amplification was possible. From 400 copies/reaction (cps/rxn) onward, amplification of ZEBOV RNA lost efficiency, as is represented by the increasing crossing point value.

Figure 2

Performance of the ready-made filovirus test kit. A, Testing of plaque-quantified virus stock solutions. Virus was diluted in virus-negative human serum to the listed concentrations and tested in 4 replicate assays each. +, positive result; − negative result; RT-PCR, reverse-transcription polymerase chain reaction. B, Probit analysis on replicated tests (typically 6 tests per concentration) of virus-negative human serum inoculated on the level of the lysis buffer with given concentrations (X-axis) of filovirus RNA in vitro transcripts. The observed rates of positive results per total no. of tests performed are plotted on the Y-axis. Nos. in each panel are the concentrations beyond which the statistical chance of detection exceeds 95%, according to the probit model (dose-response relationship). Nos. in parenthesis are 95% confidence intervals for these numbers. cps, copies.