RNA-protein interactions in an unstructured context

Abstract

Despite their importance, our understanding of noncovalent RNA-protein interactions is incomplete. This especially concerns the binding between RNA and unstructured protein regions, a widespread class of such interactions. Here, we review the recent experimental and computational work on RNA-protein interactions in an unstructured context with a particular focus on how such interactions may be shaped by the intrinsic interaction affinities between individual nucleobases and protein side chains. Specifically, we articulate the claim that the universal genetic code reflects the binding specificity between nucleobases and protein side chains and that, in turn, the code may be seen as the Rosetta stone for understanding RNA-protein interactions in general.

Keywords: RNA-protein granules; RNA-protein interactions; intrinsically disordered proteins; long noncoding RNAs; nucleobase/amino acid interaction affinity scales.

Figure 1

RNA binding and protein disorder. (A) Distribution of the degree of structural disorder among human proteins according to IUPRED 42, with the most enriched GO functional term for each bin indicated in color; (B) Structural disorder among poly(A) RNA‐binding proteins in human according to IUPRED 42. Figure 1A was reproduced with permission (Oxford University Press) from Ref. 13.

Figure 2

1D physicochemical profiles of FUS. Disorder, charge, hydrophobicity and GUA‐affinity profiles of human FUS in relation to its domain structure. The charge, hydrophobicity and GUA‐affinity profiles were determined by using a running‐average window of 21‐residues.

Figure 3

Robustness of affinities between nucleobases and amino acid side chains. (A) A close correlation between the knowledge‐based (KNB) scales of GUA/side‐chain affinity derived from two largely independent sets of structures of RNA–protein complexes (NAR2013 50 and NDB2017). The NDB2017 set was generated using a representative dataset from the Nucleic Acid Database (http://ndbserver.rutgers.edu) 110 with the resolution cutoff of 2.5 A using the identical method as described by Polyansky et al. 50. The overlap between the two sets is given by the Venn diagram. Inset: anticorrelation between the GUA and ADE side‐chain affinity scales is observed for two sets of knowledge‐based scales (NAR2013 50 and NDB2017) and the MD‐based scales of nucleobase–amino acid affinity in methanol 81. The bars indicate the Pearson R coefficients between the GUA and ADE scales. (B) ADE and URA amino acid‐binding free energy scales are up to a constant largely insensitive to local dielectric constant, while those for GUA and CYT strongly depend on it (Pearson Rs given in the inset) 81. (C) Andrews et al. 80 nucleobase–amino acid‐binding ΔG scales (x‐axis) correlate closely with de Ruiter et al. scales 81 (y‐axis) Figure 3C was reproduced with permission (American Chemical Society) from Ref. 80.

Figure 4

The complementarity hypothesis. (A) Codon PYR content correlates with the cognate amino acid affinity for PYR mimetics 100, while codon PUR content correlates with the cognate amino acid affinity for GUA 50. (B) Right: profile calculation method together with a typical pair of mRNA PYR density and protein PYR‐mimetic affinity profiles in human; left: Pearson Rs for mRNA PYR density/protein PYR‐mimetic affinity profiles in 15 species. (C) Left: Location of top matches for human mRNAs and cognate proteins, including UTRs and transition regions (violet/olive) for mRNA PUR density/protein ADE affinity, and mRNA PYR density/protein PYR affinity cases. Numbers of top matches are given above bars. Right: An example of a top match between mRNA PUR density and cognate protein ADE affinity. Figures 4A and 4B were reproduced with permission (Oxford University Press) from Ref. 102.

Figure 5

The universal genetic code as the Rosetta stone for understanding RNA–protein interactions.