Combinatorial recognition of clustered RNA elements by the multidomain RNA-binding protein IMP3

Abstract

How multidomain RNA-binding proteins recognize their specific target sequences, based on a combinatorial code, represents a fundamental unsolved question and has not been studied systematically so far. Here we focus on a prototypical multidomain RNA-binding protein, IMP3 (also called IGF2BP3), which contains six RNA-binding domains (RBDs): four KH and two RRM domains. We establish an integrative systematic strategy, combining single-domain-resolved SELEX-seq, motif-spacing analyses, in vivo iCLIP, functional validation assays, and structural biology. This approach identifies the RNA-binding specificity and RNP topology of IMP3, involving all six RBDs and a cluster of up to five distinct and appropriately spaced CA-rich and GGC-core RNA elements, covering a >100 nucleotide-long target RNA region. Our generally applicable approach explains both specificity and flexibility of IMP3-RNA recognition, allows the prediction of IMP3 targets, and provides a paradigm for the function of multivalent interactions with multidomain RNA-binding proteins in gene regulation.

The structural basis for RNA selectivity by the IMP family of RNA-binding proteins

“…These studies 15,16,40 of the IMP family as well as the data presented here show the necessity of the GG motif for all IMP family members. This provides an important consideration when performing next generation sequencing of IMP family targets.…”

“…Both CA and GG dinucleotides are necessary for binding to all the IMPs 15,16,40 , and were enriched by SELEX in this study. KH domains recognize 4–6 nt sequence elements 6 ; the IMP KH domains have been shown to recognize 4 nt sequences 15,16,40 , therefore the nucleotides flanking the CA and GG motifs are likely important for binding. Their selective enrichment also changed across the last three rounds of SELEX (i.e., rounds 7, 8 and 9) (Fig.…”

“…Recently structural studies have investigated the binding preferences for the remaining family member, IMP3, and found an analogous bipartite consensus sequence of CGGC—12–25—CA 40 . Given that the amino acid sequence of IMP3 is closer to ZBP1 than IMP2, one may expect its consensus sequence to more closely resemble ZBP1’s CGG-(A/U) motif than IMP2’s (A/U)-GG-(A/U) motif 16 .…”

“…Given that the IMP family, like most RNA-binding proteins rely on multiple RNA-binding domains where each domain has its specific sequence preference, it is necessary for motif searches to consider possible randomized linker regions between specific binding motifs. Alongside previous work 16,40 , this study incorporates randomized variable linker regions between different domain consensus sequences when performing in silico target searches. The increased flexibility between consensus sequence elements helps discover targets of the IMP family and may inform future genome wide searches of other RBP targets.…”

IGF2BP3 From Physiology to Cancer: Novel Discoveries, Unsolved Issues, and Future Perspectives

“…Nevertheless, IGF2BP3 binds to coding regions with a higher frequency than either IGF2BP1 or IGF2BP2 (Conway et al, 2016). In addition, a recent analysis of RNA recognition by multidomain IGF2BP proteins indicated that while IGF2BP1 associates with the CGGAC RNA motif, IGF2BP3 recognizes two related GGC-core elements (GGCA and CGGC), further supporting the existence of differences during the recognition of RNA (Schneider et al, 2019). The data described in the literature suggest that multiple RNA motifs, including CACA, UACA, AACA (Conway et al, 2016), GCAC (Palanichamy et al, 2016), and GGAC (Huang et al, 2018b), are recognized by the IGF2BP family, but the extent to which these sequences are specific to each RBP is still unclear.…”