UC Berkeley

Self-assembling proteins are emerging as compelling solutions in both drug delivery and vaccine development. Typically, a self-assembling protein, such as a virus-like particle, is repurposed as a drug delivery vehicle by decorating a native, well-characterized viral capsid through chemical or genetic modification. While this was successful in several applications, it does not allow the physical properties of the particle itself to be adjusted in a rational manner. Indeed, there are many instances in which native capsid properties—such as stability, size, binding, chemical reactivity, among others—are non-ideal for the application at hand.

Herein, we developed and employed a new technique to study the mutability of self-assembling virus-like particles. This technique allows facile generation of highly targeted libraries, as well as simultaneous evaluation of many variants in a single pool. Furthermore, this technique is compatible with many different direct functional selections, such as heat, acid, or chemical challenges, enabling granular insight into how mutations affect chemical and physical properties. We evaluated how single amino acid mutations affect self-assembly of a model virus-like particle (Chapter 2). We then applied this technique to study the two-amino acid mutability of a small and flexible loop (Chapter 3). We also studied how N-terminal extensions alter the stability and chemical reactivity of the virus-like particle (Chapter 4). Finally, we sought to understand how mutations can affect the quaternary geometry of a self-assembling particle (Chapter 5). In sum, by allowing the simultaneous evaluation of many variants in a single pool, this work has generated the most systematic data available regarding the effects of individual amino acid substitutions on the resulting properties of virus-like particles.