Proteome-wide search reveals unexpected RNA-binding proteins in Saccharomyces cerevisiae

Abstract

The vast landscape of RNA-protein interactions at the heart of post-transcriptional regulation remains largely unexplored. Indeed it is likely that, even in yeast, a substantial fraction of the regulatory RNA-binding proteins (RBPs) remain to be discovered. Systematic experimental methods can play a key role in discovering these RBPs--most of the known yeast RBPs lack RNA-binding domains that might enable this activity to be predicted. We describe here a proteome-wide approach to identify RNA-protein interactions based on in vitro binding of RNA samples to yeast protein microarrays that represent over 80% of the yeast proteome. We used this procedure to screen for novel RBPs and RNA-protein interactions. A complementary mass spectrometry technique also identified proteins that associate with yeast mRNAs. Both the protein microarray and mass spectrometry methods successfully identify previously annotated RBPs, suggesting that other proteins identified in these assays might be novel RBPs. Of 35 putative novel RBPs identified by either or both of these methods, 12, including 75% of the eight most highly-ranked candidates, reproducibly associated with specific cellular RNAs. Surprisingly, most of the 12 newly discovered RBPs were enzymes. Functional characteristics of the RNA targets of some of the novel RBPs suggest coordinated post-transcriptional regulation of subunits of protein complexes and a possible link between mRNA trafficking and vesicle transport. Our results suggest that many more RBPs still remain to be identified and provide a set of candidates for further investigation.

Figure 1. Methods for identifying novel RNA-binding proteins.

A. A schematic representation of the protein microarray method. Specific mRNA of interest is synthesized in vitro, while polyA-selected mRNA is isolated from mid-log phase cells grown in YPD. The samples are labeled with Cy-dyes, pooled together and bound to a protein microarray. In subsequent analyses, proteins interacting with total mRNA or preferentially with individual mRNA are identified. B. Proteins associating with mRNAs immobilized on oligo-dT beads were isolated and identified using LC-MS/MS.

Figure 2. Enrichment of annotated RBPs in the protein microarray data.

The red line represents proteins annotated as RBPs, the blue line represents proteins not annotated as RBPs. Protein microarrays were probed with 500 nM polyA-selected RNA labeled with Cy5. Proteins were ranked by the ratio of normalized mean Cy5 signal to median microarray background. The Wilcoxon rank test was used to evaluate the significance of enrichment of known RBPs relative to proteins not annotated as RBPs.

Figure 3. GO Term enrichment of targets for novel RBPs identified by IP-microarrays.

Enrichment was determined using GO Term search by on target genes with FDR≤0.01% (with the exception of Smy1 and Mtq2, which are not counted as RBPs, for which targets with FDR≤1% were considered). Hypergeometric p-value<0.05 (corrected for multiple hypothesis testing) was used as a cut-off. Complete target datasets are available in Table S5. In blue, are shown enriched “Biological Process” and in red- “Cellular Component”.