RNA-binding proteins (RBPs) play central roles in the dynamic regulation of RNA metabolism, especially during cellular stress that threatens proteostasis and gene expression. One of these, hnRNP A2/B1, is involved in a wide variety of RNA processes such as alternative splicing, mRNA transport, and microRNA sorting into extracellular vesicles. Nevertheless, the complete set of cellular functions of hnRNP A2/B1, including its stress-induced changes in localization and interaction with nucleic acids, remains poorly understood, in part due to technical obstacles to tracking its endogenous forms in living cells.
To fill this need, we screened a yeast display library of Nbs against hnRNPA2/B1 using a peptide sequence derived from the protein. From this screen, we identified an anti-hnRNP A2/B1 Nb we named Nb19. This Nb showed specific binding to the hnRNP A2/B1 peptide and strong colocalization to EGFP-fused hnRNP A2/B1 in cell lines. Also, Nb19 colocalized to endogenous hnRNP A2/B1, as assessed using antibody staining. In addition to visualization, Nb19 enabled acute and specific modulation of hnRNP A2/B1 abundance via proteasomal degradation. Nb19 fused to an E3 ligase domain IpaH9.8 caused an approximately 40% reduction in EGFP-hnRNP A2 levels in a cell line. These results show that Nb19 specifically interacts with hnRNP A2/B1 in the cytoplasm of mammalian cells.
To improve the binding of Nb19, we conducted directed evolution using yeast display. We generated mutants of Nb19 using error-prone PCR and conducted fluorescence-activated sorting (FACS) followed by DNA shuffling and further mutagenesis. After 3 rounds of FACS, we identified Nb19 mutants with improved affinity. We identified a mutant named Nb19.3 that contains all the high-frequency mutations in the high-affinity pool, which showed 23-fold higher affinity (Kd = 126 nM) than the wild-type Nb19 (~3 µM). However, Nb19.3 showed punctate aggregates in the cytoplasm of mammalian cells that don’t co-localize with anti-hnRNP A2/B1 antibody staining. This suggests that the mutations that improve affinity destabilize the Nb in the cytoplasm. These results highlight the challenges of engineering Nbs for improved binding.
Finally, we also demonstrate the robustness of our screen by identifying an anti-tau Nb. Using the same Nb screening pipeline, we found a clone we named tauNb1. The clone showed specific binding to the peptide antigen sequence used for screening and colocalized with EGFP-fused tau. The colocalization pattern significantly changed when EGFP-tau was fused to a nuclear localization signal, further confirming its binding. We also show that tauNb1 fused to IpaH9.8 causes degradation of EGFP-tau. Taken together, our screening approach allows the identification of Nbs to challenging targets such as intrinsically disordered proteins, providing a useful tool in studying these proteins under physiological conditions.