Background: Few HIV-1fusion and replication inhibitors were developed, with limited
clinical applications because of their short half-life, drug resistance and cross-reactivity with
preexisting antibodies in HIV-infected patients.
Objective: These limitations call for new strategies in the development of next anti-HIV-1 drugs.
Among anti-gp41HIV-1 inhibitors, short-peptides exhibit high antiviral activity but the mechanism of
action at molecular level has not been sufficiently addressed.
Method: We report potent QSAR (Quantitative Structure-Activity Relationship) models, used for
biological activity prediction of novel short HIV-1 gp41 inhibitor peptides in order to: (i) validate the
anti-HIV-1 activity of MT-sifuvirtide, MT-SC34EK, MT-C34 and HP23, expressed as IC50fusion and
IC50replication; (ii) predict inhibitory activity of SC24EK and its MT-derivative expressed as IC50resistant
HIV-1 NL4-3 variant; (iii) propose new derivatives DMT-SC22EK, DMT-SC29EK and DMT-sifuvirtide
through addition of aspartic acids by induced-mutagenesis; (iv) use molecular similarity established by
fingerprint models to correlate molecular spatial features with predicted biological activity of newly
generated inhibitors over parent compounds.
Results: We obtained good QSAR statistic parameters, demonstrating that our QSAR models are able
to predict biological activity of new HIV-1 inhibitors with suitable accuracy.
Conclusion: Despite acknowledged drawbacks of a reduced dataset, our results may enhance the
evaluation of biological activity of new and classical synthetic peptides as anti-HIV-1agents and
represent a good start for further studies in developing new antiviral drugs.