Computationally Designed Peptides for Zika Virus Detection: An Incremental Construction Approach

Abstract

Herein, and in contrast to current production of anti-Zika virus antibodies, we propose a semi-combinatorial virtual strategy to select short peptides as biomimetic antibodies/binding agents for the detection of intact Zika virus (ZIKV) particles. The virtual approach was based on generating different docking cycles of tetra, penta, hexa, and heptapeptide libraries by maximizing the discrimination between the amino acid motif in the ZIKV and dengue virus (DENV) envelope protein glycosylation site. Eight peptides, two for each length (tetra, penta, hexa, and heptapeptide) were then synthesized and tested vs. intact ZIKV particles by using a direct enzyme linked immunosorbent assay (ELISA). As a reference, we employed a well-established anti-ZIKV antibody, the antibody 4G2. Three peptide-based assays had good detection limits with dynamic range starting from 105 copies/mL of intact ZIKV particles; this was one order magnitude lower than the other peptides or antibodies. These three peptides showed slight cross-reactivity against the three serotypes of DENV (DENV-1, -2, and -3) at a concentration of 106 copies/mL of intact virus particles, but the discrimination between the DENV and ZIKV was lost when the coating concentration was increased to 107 copies/mL of the virus. The sensitivity of the peptides was tested in the presence of two biological matrices, serum and urine diluted 1:10 and 1:1, respectively. The detection limits decreased about one order of magnitude for ZIKV detection in serum or urine, albeit still having for two of the three peptides tested a distinct analytical signal starting from 106 copies/mL, the concentration of ZIKV in acute infection.

Keywords

• virtual screening
• molecular modelling
• peptides
• Zika virus
• Dengue virus
• antibody mimetics
• ELISA