Given the growing number of diseases caused by emerging or endemic viruses, original strategies are urgently required: (1) for the identification of new drugs active against new viruses and (2) to deal with viral mutants in which resistance to existing antiviral molecules has been selected. In this context, antiviral peptides constitute a promising area for disease prevention and treatment. The identification and development of these inhibitory peptides require the high-throughput screening of combinatorial libraries. Phage-display is a powerful technique for selecting unique molecules with selective affinity for a specific target from highly diverse combinatorial libraries. In the last 15 years, the use of this technique for antiviral purposes and for the isolation of candidate inhibitory peptides in drug discovery has been explored. We present here a review of the use of phage display in antiviral research and drug discovery, with a discussion of optimized strategies combining the strong screening potential of this technique with complementary rational approaches for identification of the best target. By combining such approaches, it should be possible to maximize the selection of molecules with strong antiviral potential.

We wanted to develop a therapeutic approach against rabies disease by targeting the lyssavirus transcription/replication complex. Because this complex (nucleoprotein N-RNA template processed by the L polymerase and its cofactor, the phosphoprotein P) is similar to that of other negative-strand RNA viruses, we aimed to design broad-spectrum antiviral drugs that could be used as a complement to postexposure vaccination and immunotherapy. Recent progress in understanding the structure/function of the rabies virus P, N, and L proteins predicts that the amino-terminal end of P is an excellent target for destabilizing the replication complex because it interacts with both L (for positioning onto the N-RNA template) and N (for keeping N soluble, as needed for viral RNA encapsidation). Thus, peptides mimicking various lengths of the aminoterminal end of P have been evaluated, as follows: (i) for binding properties to the N-P-L partners by the two-hybrid method; (ii) for their capacity to inhibit the transcription/replication of a rabies virus minigenome encoding luciferase in BHK-21-T7 cells; and (iii) for their capacity to inhibit rabies virus infection of BHK-21-T7 cells and of two derivatives of the neuronal SK-N-SH cell line. Peptides P60 and P57 (the first 60 and first 57 NH 2 residues of P, respectively) exhibited a rapid, strong, and long-lasting inhibitory potential on luciferase expression (>95% from 24 h to 55 h). P42 was less efficient in its inhibition level (75% for 18 to 30 h) and duration (40% after 48 h). The most promising peptides were synthesized in tandem with the Tat sequence, allowing cell penetration. Their inhibitory effects were observed on BHK-21-T7 cells infected with rabies virus and Lagos bat virus but not with vesicular stomatitis virus. In neuronal cells, a significant inhibition of both nucleocapsid inclusions and rabies virus release was observed.