The chemokine receptors CCR5 and CXCR4 are promising non-virus-encoded targets for human immunodeficiency virus (HIV) therapy. We describe a selection procedure to isolate mutant forms of RANTES (CCL5) with antiviral activity considerably in excess of that of the native chemokine. The phage-displayed library of randomly mutated and N-terminally extended variants was screened by using live CCR5-expressing cells, and two of the selected mutants, P1 and P2, were further characterized. Both were significantly more potent HIV inhibitors than RANTES, with P2 being the most active (50% inhibitory concentration of 600 pM in a viral coat-mediated cell fusion assay, complete protection of target cells against primary HIV type 1 strains at a concentration of 10 nM). P2 resembles AOP-RANTES in that it is a superagonist of CCR5 and potently induces receptor sequestration. P1, while less potent than P2, has the advantage of significantly reduced signaling activity via CCR5 (30% of that of RANTES). Additionally, both P1 and P2 exhibit not only significantly increased affinity for CCR5 but also enhanced receptor selectivity, retaining only trace levels of signaling activity via CCR1 and CCR3. The phage chemokine approach that was successfully applied here could be adapted to other chemokine-chemokine receptor systems and used to further improve the first-generation mutants reported in this paper.Despite the success of highly active antiretroviral therapy, new human immunodeficiency virus type 1 (HIV-1) inhibitors are still needed and among the most promising new approaches is the blockade of viral entry into target cells (20). HIV-1 entry into target cells is initially dependent on the interaction of its envelope glycoproteins with CD4 and a coreceptor, with the chemokine receptors CCR5 and CXCR4 being by far the most commonly used by HIV-1 (5).HIV entry is inhibited by the natural chemokine ligands of the coreceptors, including MIP-1␣ (CCL3), MIP-1 (CCL4), and RANTES (CCL5) for CCR5 (8) and SDF-1 (CXCL12) for CXCR4 (6,23). Certain N-terminal modifications have been shown to increase the anti-HIV activity of native chemokines (21,32,33,36), and the most potent of these molecules owe their anti-HIV activity to their ability to induce prolonged intracellular sequestration of coreceptors (18,31).Up until now, chemokine structure-activity relationships have been studied via either scanning or truncation mutagenesis (14,16,19,24), peptide scanning of primary sequence (22), or semirational design of chemokine analogues (21,32,36). A more-rapid, bioengineering-based approach for the selection of useful chemokine variants has yet to be described.We decided to apply current knowledge of the structureactivity relationship of chemokines and the mechanism by which they inhibit HIV entry (2,7,18) to the design of a phage display strategy for the discovery of N-terminally mutated RANTES variants with improved anti-HIV activity. Selection led to the isolation of around 40 clones that exhibited a consensus sequence, and two clones were chosen...