RC3H1 post-transcriptionally regulates A20 mRNA and modulates the activity of the IKK/NF-κB pathway

The RNA-binding protein RC3H1 (also known as ROQUIN) promotes TNFα mRNA decay via a 3′UTR constitutive decay element (CDE). Here we applied PAR-CLIP to human RC3H1 to identify ∼3,800 mRNA targets with >16,000 binding sites. A large number of sites are distinct from the consensus CDE and revealed a structure-sequence motif with U-rich sequences embedded in hairpins. RC3H1 binds preferentially short-lived and DNA damage-induced mRNAs, indicating a role of this RNA-binding protein in the post-transcriptional regulation of the DNA damage response. Intriguingly, RC3H1 affects expression of the NF-κB pathway regulators such as IκBα and A20. RC3H1 uses ROQ and Zn-finger domains to contact a binding site in the A20 3′UTR, demonstrating a not yet recognized mode of RC3H1 binding. Knockdown of RC3H1 resulted in increased A20 protein expression, thereby interfering with IκB kinase and NF-κB activities, demonstrating that RC3H1 can modulate the activity of the IKK/NF-κB pathway.

(g) Mouse RC3H1 target mRNAs identified by Leppek and colleagues are compared to human PAR-CLIP RC3H1 target mRNAs. Out of 95 genes, 91 genes are converted to orthologous human genes, and divided into two groups based on FPKM expression value in HEK293 cells. For each group, number of mRNAs that are overlapping in human PAR-CLIP RC3H1 target mRNAs is shown.
Number of mouse CDE containing mRNAs is shown in parentheses.
(h) Distribution of consensus RC3H1 binding cluster along 3'UTRs of mRNA.
(i) Density of predicted conserved miRNA target sites around crosslink sites in 3'UTRs. RC3H1 crosslink sites and miRNA target sites display no tendency for direct overlap but the larger context (10-50 nt) shows mildly elevated seed density. The gray envelope represents the standard error of the mean. RC3H1 target sites identified by the 4SU-1 PAR-CLIP were used in this analysis.

Supplementary Figure 2
(a) log10 frequencies of 7mers occurring in the 41 nt window around the RC3H1 preferred crosslink sites are shown for 4SU-1 PAR-CLIP and 4SU-2 PAR-CLIP libraries, showing a good correlation of 7mer occurrence between the libraries.
(b) A scatter plot showing 7mers log10 frequencies in the 41 nt window around the preferred crosslink sites of consensus 3'UTR RC3H1 binding sites versus 7mers log10 frequencies in all 3'UTR sequences. 7mers comprising of only A/U or G/U are plotted in red or blue, respectively. U-rich sequences with A contents (red) are more frequent and enriched over the background frequency compared to control 7mer U-rich sequences with G contents.
(c) A scatter plot showing 5mers log10 frequencies in the 41 nt window around the preferred crosslink sites of consensus 3'UTR RC3H1 binding sites versus 5mers log10 frequencies in all 3'UTR sequences.
(d) RC3H1 binding sites tend to have stem-loop secondary structure. 41 nt sequences centered around RC3H1 crosslink site were computationally folded. Base pairing probability for each position around crosslink sites are averaged over all 3'UTR binding site (red) and control 3'UTR sequences (gray).
(e) A RNA structure dot plot for top1000 RC3H1 consensus targets, ranked by number of T to C transition events in 3'UTR (top right triangular), and control sets of random 3'UTR sequence from RC3H1 target genes (bottom left triangular) demonstrates the stem-loop structure of RC3H1 binding sites. A dot placed in the ith row and jth column of a triangular array represents the base pair between the ith base with jth base, and the size of dot is proportional to the square root of average base paring probability for each base paring.

Supplementary Figure 3
(a) A scatter plot of the log2 fold changes of "heavy" to "light" SILAC ratios (H/L) versus "heavy" to "light" SILAC ratios (H/L) in label swap experiment in Figure 3a. (b) RC3H1 target transcripts have shorter half-lives. Red or black dots represent the 3'UTR RC3H1 consensus targets or non-targets, respectively. log2 expression levels and half-lives (min) are blotted on x-and y-axis, respectively.
(c) A cumulative distribution function (CDF) plot of mRNA half-lives shown in (b). The mean mRNA half-lives of RC3H1 targets and non-targets are 269.9 min and 311.1 min, respectively. The difference is significant with a p-value smaller than 2.2e-16 (Wilcoxon's rank sum test).
(d) IGF2BP1 target transcripts are not as short half-lived as RC3H1 bound mRNAs. Red or black dots represent the 3'UTR IGF2BP1 targets or non-targets, respectively. Log2 expression levels and halflives (min) are blotted on x-and y-axis, respectively.   Synthesized IBα protein binds to NF-B to form the NF-B/IBα complex and thus inactivates NF-B. Synthesized A20 protein inhibits IKK activity. RC3H1 destabilizes the IκBα and A20 mRNAs.
Right: The simulated dynamics of 6 components of the NF-B model are shown for wild type cells (black lines), RC3H1 overexpression (orange lines) and siRC3H1/2 expression (blue lines). The simulations qualitatively fit to the experimental findings for the three conditions (compare Fig. 6c-f).
RC3H1 overexpression (orange lines) leads to a decrease in A20 expression and increases IKK activation (top panels), while marginally effecting IκBα expression and NF-κB activity (bottom panels) compared to wild type cells (black lines). In contrast, siRC3H1/2 treated cells (blue lines) exhibit higher levels of A20 and IBα and decreased IKK as well as NF-B activity compared to wild type cells (black lines).      Reads from individual PAR-CLIP experiment were independently mapped to the human genome (hg18) using tophat2. Mapped all or unique reads in exons of UCSC genes as well as tRNA and rRNA regions were counted using quasR. tRNA and rRNA regions were obtained from the RepeatMasker track of the UCSC genome browser. Reads containing T-to-C transitions for 4SU library (indicated as TC-reads) or G-to-A transitions for 6SG library (indicated as GA reads) were counted and shown in the table.