Rational Design of Multivalent Synthetic Binders for the Development of Ultrasensitive Ebola Virus Diagnostics

Abstract

Orthoebolaviruses represent a significant threat to global public health due to their high case fatality rate and the alarming rise in the frequency of outbreaks recently observed. In 2014, the World Health Organization (WHO) issued a call for the development of rapid, accurate, sensitive, as well as cost-effective diagnostic tests that can be utilized in resource-limited settings. Recently, multiple efforts have focused on the development of rapid diagnostic tests (RDTs) capable of detecting a range of viral proteins, including the secreted glycoprotein (sGP) — an early biomarker that is uniquely specific to orthoebolaviruses. However, most of these RDTs rely on the use of monoclonal antibodies (mAbs), which are expensive, difficult to manufacture in large scale, and require cold chain storage, thus rendering them unsuitable for basic healthcare settings. The focus of this dissertation work is therefore to develop the next-generation RDTs, by replacing the traditional mAbs with synthetic binding proteins (SBPs), which are cost-effective, stable, and easy to manufacture.To achieve this objective, nanobodies (Nbs) were selected as the preferred SBP scaffold. First, a high-affinity anti-sGP Nb was isolated from a synthetic yeast display library and then subjected to affinity maturation through autonomous hypermutation. The resulting matured Nb was found to bind to the antigen with a 2:1 stoichiometry and displayed low nanomolar affinity. Next, novel cryo-EM data were integrated with AlphaFold3 modeling, to map the Nb's epitope on the glycoprotein isoform. This structural analysis also provided insights into the intramolecular distance separating the two Nb binding sites on the sGP homodimer. This estimated distance was used to guide the design of multivalent Nb constructs aimed at improving affinity through avidity. To better understand the relationship between avidity and apparent binding affinity, a series of bivalent Nb constructs with varying flexible linker lengths were produced. A novel statistical thermodynamic model based on the theory of worm-like chain from polymer physics was then developed to further elucidate the binding behavior of these multivalent nanobody constructs. This model revealed the optimal linker length that maximizes the true intramolecular binding and was able to distinguish it from the apparent affinity enhancement arising from intermolecular crosslinking between two different sGP dimer molecules. The resulting rationally designed bivalent Nb construct demonstrated picomolar affinity for its target, rendering it appropriate for incorporation into an Ebola RDT. Altogether, by combining yeast display, AlphaFold modeling, and our statistical thermodynamic model, this work establishes a framework to guide the rational design of multivalent synthetic binders.