Integrative regulatory mapping indicates that the RNA-binding protein HuR couples pre-mRNA processing and mRNA stability

Abstract

RNA-binding proteins coordinate the fates of multiple RNAs, but the principles underlying these global interactions remain poorly understood. We elucidated regulatory mechanisms of the RNA-binding protein HuR, by integrating data from diverse high-throughput targeting technologies, specifically PAR-CLIP, RIP-chip, and whole-transcript expression profiling. The number of binding sites per transcript, degree of HuR association, and degree of HuR-dependent RNA stabilization were positively correlated. Pre-mRNA and mature mRNA containing both intronic and 3' UTR binding sites were more highly stabilized than transcripts with only 3' UTR or only intronic binding sites, suggesting that HuR couples pre-mRNA processing with mature mRNA stability. We also observed HuR-dependent splicing changes and substantial binding of HuR in polypyrimidine tracts of pre-mRNAs. Comparison of the spatial patterns surrounding HuR and miRNA binding sites provided functional evidence for HuR-dependent antagonism of proximal miRNA-mediated repression. We conclude that HuR coordinates gene expression outcomes at multiple interconnected steps of RNA processing.

Figure 1

Identification of HuR regulatory elements. A) For binding sites detected in FOS 3′ UTR (top) and SFRS9 3′UTR (bottom), the KDE of conversion events (black line) and percent conversion (heatmap) were superimposed on read depth information (grey bars) for each nucleotide in the initial group. B) RREs identified applying cERMIT to clusters ranked by CLIs (left). Comparison of quantile plots, number of occurrences, and p-values for each RRE shown (right).

Figure 2

Landscape of HuR binding sites in transcript regions. A) The proportion of densities for HuR binding clusters derived from each region of an mRNA as a function of cross-linking index (CLI). B) Histogram (log2) and quantiles of the distribution for the number of HuR binding sites per transcript. C) Pie-chart depicting transcripts with specific regions of HuR binding sites.

Figure 3

Comparison of PAR-CLIP and RIP-chip HuR interaction data. A) Venn diagram of the overlap between transcripts defined as HuR targets using PAR-CLIP and RIP-chip, which were also functionally responsive (see figure S5C). B) Comparison of the cumulative distribution of transcript abundance for targets detected by RIP-chip only (blue), PAR-CLIP only (green), and RIP-chip and PAR-CLIP (red) using microarray data to quantify transcript abundance. C) Comparison of the HuR LOD score distributions of PAR-CLIP data categorized by number of binding sites per transcript. D) Comparison distribution of PAR-CLIP binding sites/transcript for targets identified only by PAR-CLIP or by both PAR-CLIP and RIP-chip.

Figure 4

HuR-dependent mRNA stabilization. A) Comparison of cumulative distribution of mRNA differential expression after HuR knockdown for targets identified by RIP-chip (black), PAR-CLIP (dashed), and neither method (grey). The depletion of RIP-chip and PAR-CLIP targets were statistically significantly compared non-targets (p < 0.001). B) Protein classes enriched in HuR PAR-CLIP, RIP-chip, and knockdown targets. After each dataset was percentile rank transformed, the enrichment of significant Panther defined gene sets was reported using the −log10(enrichment p-value).

Figure 5

Degree of HuR binding and functional outcome. A) Transcripts with more HuR binding sites exhibit a greater decrease in response to HuR knockdown. B) Transcripts were divided into 4 categories based on their percentile of decrease after knockdown, 75–100th representing mRNAs that decreased the most. Transcripts with higher HuR LOD scores correlated with a greater decrease in mRNA levels following HuR knockdown. C) Transcripts with binding sites only in introns (green) or only in 3′ UTRs (red, behind green line) were significantly destabilized by HuR knockdown. D) More intronic HuR binding sites exhibit greater destabilization after HuR knockdown. E) Transcripts with HuR binding sites in both 3′ UTRs and introns are significantly more destabilized by HuR knockdown than transcripts with HuR binding sites only in 3′ UTRs or introns.

Figure 6

HuR regulated pre-mRNA processing. A) Transcripts (restricted to those with only intronic sites) with more intronic binding sites exhibited more pre-mRNA destabilization upon HuR depletion. B) Transcripts with both intronic and 3′ UTR sites were more destabilized upon HuR depletion than transcripts with either intronic or 3′ UTR sites. C) Effect of HuR depletion on individual pre-mRNAs were validated using pre-mRNA specific primers (* indicates paired t-test p < 0.05, actual p-values listed above bar). D) Overrepresentation of binding sites in the Py-tract exons (red lines represent 1^st, 10^th, 50^th, 90^th, and 99^th percentiles of the null distribution). E) HuR RNA splicing map exhibiting consistent binding pattern for both included and excluded exons with adjacent HuR binding sites.

Figure 7

Combinatorial regulation by HuR and miRNAs. A) Distribution of distances between HuR and Ago sites restricted to the 3′ UTR identified by PAR-CLIP. Red lines indicate distribution of null model (1st, 10th, 50th, 90th, and 99th percentile). B) Based on the distribution of distances, transcripts were classified by HuR and miRNA binding profiles depicted. Only transcripts exclusively belonging to one of the classes depicted and targeted by one of the depleted miRNAs were considered. C) Transcript classes with HuR binding sites were destabilized upon HuR knockdown regardless of miRNA sites. D) De-repression of miRNA and HuR overlap transcripts compared to no-overlap transcripts upon miRNA depletion (see ST5 for p-values).