(549d) Rational and Combinatorial Design of Peptides for Ss-mRNA/dsRNA Separation and Purification

The development of a generalizable, fast, effective, scalable, translatable and economic purification platform is critical for the technical and economic success of mRNA-based therapeutics. Our goal in this work is to develop downstream bioseparation strategies to separate double-stranded RNA (dsRNA) impurities from single-stranded mRNA (ss-mRNA). To achieve this goal, we aim to identify peptide ligands that are selective towards dsRNA and couple these peptides to polymer filtration membranes to demonstrate the feasibility of purifying labile ss-mRNA from dsRNA at high yield and purity in a scalable fashion.

Two strategies were employed for the design and selection of peptide ligands. The first, a rational and focused approach, was based on mimicking the active binding pocket of naturally occurring dsRNA specific binding proteins. Protein design templates were selected based on the following criteria: (i) binding affinity to dsRNA; (ii) binding selectivity/specificity to dsRNA versus ss-mRNA; (iii) structural information available; and (iv) length/size of the protein. Four proteins/domains were identified as templates including a double stranded RNA binding domain and three RNA silencing suppressors. A library of 16-mer peptide candidates was generated using linear epitope mapping of the target proteins with particular focus on the peptides that cover the key interacting regions of the protein. A second approach based on combinatorial phage display was employed to screen a vast pool of randomized peptide sequences. Pure ss-mRNA and dsRNA were challenged against a commercial phage library for negative and positive selections. Lead peptide candidates identified from the two design strategies were synthesized and screened for dsRNA binding affinity and selectivity in an immobilized format on microarrays.

Peptides that exhibited strong binding to dsRNA while displaying weak or no binding to ssRNA were identified through high-throughput screening and further validated in an ELISA or microarray format. Lead peptides were mapped on to their parent protein templates and revealed a strong consensus with the dsRNA binding site. Interestingly, a few ssRNA-selective peptides were also identified, originating from the dsRNA-selective binding proteins but from different regions of the protein when compared to the dsRNA binding site. Lead sequences shared several common motifs and/or amino acid propensity.