11. A. Bukreyev, V. E. Volchkov. V. M. Blinov, S. V. Netesov, FEBS Lett. 323, 183 (1993). 12. C. Wi l l et al.. 1. Virol. 67. 1203 (1993). 13. See Web Fig. 1 on Science Online at www.sciencemag. org/cgi/content/full/1057269/DC1. 14. V. E. Volchkov et dl., 1. Cen. Virol. 80. 355 (1999). 15. V. E. Volchkov, V. M. Blinov. S. V. Netesov, FE8S Lett. 305. 181 (1992). 16. The plasmid pFL-EBOVe' was designed t o place the T7 promoter adjacent t o the viral leader region, and the viral trailer region was constructed t o be adjacent t o a ribozyme sequence followed by tandem terminators of T7 transcription. In this case, the correct 3' end of the transcribed EBOV antigenome, free of additional nucleotides, was generated by self-cleavage of the ribozyme (24). To increase the transcriptional activity of the T7 RNA polymerase, an additional guanosine residue was introduced between the promoter of the T7 polymerase and the first residue of the EBOV genome (25). The length of the encoded antigenome (FL-EBOVet) is therefore increased by 1 n t compared t o that of the wild-type virus (GenBank accession number, AF086833). The size of the full-length antigenome encoded by the plasmid pFL-EBOVe was 1 n t longer than that encoded by pFL-EBOVet and 2 n t longer than the size of the genome of the wild-type virus because of the insertion of additional adenosine residues at the editing site. Mutations at the GP gene editing site were introduced into the plasmid pKSS25 by site-directed mutagenesis using the primers 5'-GG GAAACTAAGAAGAAACCTCACTAG and 5'-CTAGTGAGGTTTCTTCTTAGTTTCCC. 17. The Sal I restriction site (GTCGAC) located in the GP gene at position 6180 was mutated by site-directed mutagenesis using the pair primers 5'-GGTTAGTGATGTAGATAA ACTAGTTTG and 5'-CAAACTAGTTTATC TACATCACTAACC. This mutation is silent. In addition. an accidental mutation in a nontranslated region of the L gene (A -+ U at position 18227) was found after complete sequence analysis of the final plasmid clones. 18. U. J. Buchholz, S. Finke. K.-K. Conzelmann.]. Virol. 73. 251 (1999). 19. E. Muhlberger, M. Weik. V. E. Volchkov. H.-D. Klenk. S. Becker,]. Virol. 73, 2333 (1999). 20. BSR T7/5 cells were grown overnight t o about 60 t o 80% confluency in 25-cm2 flasks in l x Glasgow medium supplemented wi th 10% NCS (newborn calf serum). One hour before transfection, cells were washed twice wi th medium without NCS. Cells were transfected w i th a plasmid mixture containing 2 k g of full-length plasmid (pFL-EBOVe+). 0.2 p g of pT/ VP30EBOV, 0.5 p g of pT/VP35EBOV, 0.5 k g of pT/ NPEBOV. and 1 p,g of pGEM-LEBOV (74, 79). Transfection experiments were carried out wi th a Fusion 6 reagent (transfection protocol supplied by Roche). The transfection medium was removed at 8 hours after transfection; cells were washed and maintained in Glasgow medium containing 2.5% NCS for 6 t o 9 days after transfection. O n average, in each rescue experiment, approximately 100 foci of rounded cells were observed in the cell culture flask (about 1 X l o s t o 2 X l o s cells). That means that one in approximately l o3 cells allowed the formation of viral particles. However, virus release from BHK cells was extremely low, and amplification of recombinant EBOV on Vero cells was necessary before further analysis. Determination of the virus titers by plaque formation showed that about 200 infectious particles were recovered from the average rescue experiment. 21. E. Muhlberger. B. Lotfering, H.-D. Klenk, S. Becker. I . Virol. 72, 8756 (1998). 22. For RT-PCR, RNA from culture supernatants of Vero E6 cells infected with individual plaques of recEBOV-et, recEBOV-e-, or wild-type EBOV was purified (with the Rneasy Kit, Qiagen) when an extensive CPE was observed. To verify the identity of the recombinant viruses, the region containing the marker restriction site Sal I shown in Fig. 1 was amplified by RT-PCR using primers 5'-AGTCATCCACAATAGCACAT and 5'-TCGTGGCAGAGGGAGTGT. The PCR products were only seen when the RT step was performed, which confirms that they were derived solely from viral RNA and not from contaminating cDNA. PCR products were consistent with the predicted size of 1298 bp. Demonstration of the presence or absence of a Sal I site in authentic EBOV and recEBOV was performed on 1% agarose gel. Sal I digestion products were consistent with the predicted sizes of 1130 and 168 bp. In addition. the sequences at the restriction site marker and at the GP gene editing site were confirmed by partial nucleotide sequencing of RT-PCR productr 23. S. Vidal, J. Curran, D. Kolakofsky, EMBO 1. 9, 2017 (1990). 24. A. T. Perrotta, M. D. Been, Nature 350, 434 (1991). 25. A. K. Pattnaik. L. A. Ball, A. W. LeGrone. G. W. Wertz, Cell 69, 1011 (1992). 26. V. E. Volchkov et dl., Virology 277, 147 (2000). 27. For electron microscopy, 72 hours after infection, control and virus-infected cells were fixed wi th Hanks' balanced salt solution (HBSS) containing 2.5% glutaraldehyde, postfixed wi th HBSS containing 1% osmium tetraoxide, dehydrated, and embedded in Epon. Ultrathin sections were cut, placed on 200mesh copper electron microscopy grids, stained wi th uranyl acetate and lead citrate, and examined wi th a Zeiss 109 transmission electron microscope a t 80 kV. 28. See Web Fig. 2 on Science Online at www.sciencemag. org/cgi/content/full/1057269/DC1. 29. For immunofluorescent staining, Vero E6 cells were infected a t a multiplicity of infection (MOI) of l o ' and were processed 3 days later for indirect immunofluorescence analysis. Cells grown on coverslips were washed w i th phosphate-buffered saline (PBS) solution, fixed wi th 4% paraformaldehyde at 4°C for 24 hours, and permeabilized wi th 0.1% Triton X-100 in PBS for 5 min. The nonspecific binding was blocked by blocking buffer (2% bovine serum albumin. 5% glycerol, and 0.2% Tween-20 in PBS), and cells were then incubated wi th 100 m M glycine for 10 min, washed wi th PBS, and incubated at 4°C for 18 hours wi th the respective antibodies [human monoclonal antibody (mAb) KZS2, which is specific for the EBOV GP (dilution 1:100 in blocking buffer); and mouse mAb B6C5, which is specific t o EBOV NP (dilution 1:10 in blocking buffer)]. Subsequently, cells were washed three times wi th PBS and stained wi th fluorescein isothiocyanate-conjugated goat anti-human immunoglobulin G (IgG) (diluted 1:SO in blocking buffer) or wi th rhodamine-conjugated goat antimouse IgG (1:100 in blocking buffer) for 1 hour at room temperature. Nuclear counterstaining was performed wi th 4',6'-diamidino-2-phenylindole (DAPI) (0.1 pg/ml). Finally, cells were washed three times wi th PBS, dipped into dH20, covered wi th mounting medium, coverslipped, and examined wi th a fluorescence microscope (Axiomat, Zeiss) wi th digital photographic equipment for taking images (Spot camera system, version 2.1.2, Diagnostic Instruments). 30. See Web Fig. 3 on Science Online at www.sciencemag. org/cgi/content/full/1057269/DC1. 31. All experiments involving infectious EBOV were carried out in biosafety level 4 (BSL4) laboratories at the lnstitute o f Virology in Marburg, Germany, and at the Jean Merieux P4 Research Center in Lyon. France. We thank D. Burton and P. Parren for providing mAb KZ52A; S. Becker for mAb B6C5; K.-K. Conzelmann for the BSR T7/5 cell Line; A. Sergeant and E. Derrington for critical reading of the manuscript; and C. Laukel and M. Rossi for expert technical help. M.W. was supported as a recipient of a fellowship from the Boehringer lngelheim Fonds. This work was supported in pan by the Deutsche Forschungsgemeinschaft (SFB 286). the Fonds der Chemischen Industrie. and INSERM; and by a grant from the Fondation pour la Recherche Medicale t o V.E.V.