Zcchc11-dependent uridylation of microRNA directs cytokine expression

Abstract

Mounting an effective host immune response without incurring inflammatory injury requires the precise regulation of cytokine expression. To achieve this, cytokine mRNAs are post-transcriptionally regulated by diverse RNA-binding proteins and microRNAs (miRNAs) targeting their 3' untranslated regions (UTRs). Zcchc11 (zinc-finger, CCHC domain-containing protein 11) contains RNA-interacting motifs, and has been implicated in signalling pathways involved in cytokine expression. The nature of the Zcchc11 protein and how it influences cytokine expression are unknown. Here we show that Zcchc11 directs cytokine expression by uridylating cytokine-targeting miRNAs. Zcchc11 is a ribonucleotidyltransferase with a preference for uridine and is essential for maintaining the poly(A) tail length and stability of transcripts for interleukin-6 (IL-6) and other specific cytokines. The miR-26 family of miRNAs targets IL-6, and the addition of terminal uridines to the miR-26 3' end abrogates IL-6 repression. Whereas 78% of miR-26a sequences in control cells contained 1-3 uridines on their 3' ends, less than 0.1% did so in Zcchc11-knockdown cells. Thus, Zcchc11 fine tunes IL-6 production by uridylating miR-26a, which we propose is an enzymatic modification of the terminal nucleotide sequence of mature miRNA as a means to regulate gene expression.

Figure 1. Zcchc11 is a uridyltransferase

a, Key structural and catalytic domains of Zcchc11. b, The catalytic core of Zcchc11 is highly conserved among mammals and lower organisms. Three highlighted aspartates (two of which were mutated to alanines in the DADA construct, as indicated) are essential for pol β nucleotidyltransferase activity. c, Zcchc11 but not the DADA mutant (immunopurified using a Flag tag) transferred adenosines onto RNA substrates. Total RNA was visualized by SYPRO Green II staining (left) and labeled RNA was visualized using autoradiographic film (right). nc, non-coding. d, Nucleotide specificity, showing ability of endogenous Zcchc11 (immunoprecipitated from A549 cells) to transfer ³²P-labeled ATP, CTP, GTP, or UTP to yeast tRNA substrate. e, Flag-tagged Zcchc11 transferred ³²P-uridine to diverse substrates including a 100 bp RNA ladder, total RNA isolated from control or Zcchc11 siRNA treated A549 cells, or yeast tRNA. Total and radiolabeled RNA were visualized as in 1c.

Figure 2. Zcchc11 knockdown alters cytokine mRNA stability and expression

a, siRNA-targeted depletion of Zcchc11 (Z) but not β-actin (β) in A549 and MLE-15 cells was confirmed by immunoblot. b, A549 cells transfected with nonspecific or Zcchc11-targeting siRNA were stimulated with TNF-α and supernatants were analyzed using a multiplex bead assay. The effect of Zcchc11 knockdown was indicated as “% Control,” with 100% equal to that observed after nontargeting transfection. “ns,” not significantly different (P>0.50 for each; N=5) from control (100%). *, significantly different from control (Factorial ANOVA and Bonferroni post hoc test, P<0.000001 for each; N=5). c, IL-6 protein concentrations in the supernatant of A549 cells treated with nonspecific and Zcchc11 siRNA was assessed by ELISA at 0, 1, 6, and 24 hours after stimulation with 10 ng/ml TNF-α. *significant effects of Zcchc11 siRNA on IL-6 at both 6 (P=0.01) and 24 (P=0.0002) hours (Factorial ANOVA and Bonferroni post hoc test; N=3–4). d, IL-6 protein concentrations were also assessed in control and Zcchc11 knockdown groups in MLE-15 cells by ELISA after 24 hour incubations with or without mouse TNF-α. *, significant effect of Zcchc11 siRNA (P=0.001; t test; N=3). e, IL-6 mRNA levels were assessed by real-time RT-PCR at the indicated time points in TNF-α (10 ng/ml) stimulated A549 cells treated with nonspecific and Zcchc11 targeting siRNA. Each time point represents the mean and SEM from N=3 independent experiments. Asterisks indicate statistical significance of P < 0.05. f, IL-6 mRNA stability in A549 cells stimulated with TNF-α for 1 hour, using actinomycin D to halt transcription. *, P=0.0007 (Factorial ANOVA; N=3). g, IL-6 poly(A) tail length was determined by LM-PAT to be decreased in Zcchc11 knockdown A549 cells stimulated with TNF-α for 1 hour. PCR products were electrophoresed in an agarose gel and visualized with ethidium bromide staining.

Figure 3. Mir-26 family regulates IL-6 expression

a, The mir-26 target sequence in the 3′ UTR of IL-6 is conserved across multiple mammalian species. The human mir-26a and mir-26b indicated was consistent across clones from A549 cells (i.e., with no non-genomic modifications shown). Seed sequence pairing is indicated by lines. b, mir-26b content was unaffected by Zcchc11 siRNA. RNA was collected 0, 1, and 6 hours after TNF-α stimulation of A549 cells, and mir-26b was detected using Northern hybridization and quantified by densitometry (P=0.95, t test; N=3). c, Transfection with mir-26b mimetics decreased IL-6 from A549 cells stimulated with TNF-α for 24 hours. *, P=0.005 (paired t test; N=3). d, Oligonucleotide inhibition of endogenous mir-26b increased TNF-induced IL-6 expression from A549 cells. *, P=0.004 (paired t test; N=3). e, Zcchc11 but not the DADA mutant (immunopurified using a Flag tag) preferred to uridylate small single stranded RNA versus double stranded RNA. Total RNA was visualized by SYPRO Green II staining (left) and labeled RNA was visualized using autoradiographic film (right). f, Increasing uridylation of mir-26 caused enhanced IL-6 production from TNF-stimulated A549 cells. * P=0.042 vs. mir-26b, and ** P=0.025 vs. mir-26b-U and P=0.001 vs. mir-26b (Factorial ANOVA and Bonferroni post hoc test; N=3).

Figure 4. Zcchc11-dependent uridylation of mir-26 abrogates IL-6 mRNA silencing

a, Zcchc11 knockdown decreased luciferase expression from a reporter construct containing the IL-6 3′UTR. * P=0.037 (t test; N=3). b, Overexpression of Zcchc11 increased luciferase expression driven by the IL-6 3′UTR reporter vector and this increase was attenuated by the nucleotidyltransferase null Zcchc11 mutant (DADA). * P=0.0019 (paired t test; N=5). c, mir-26 drove down IL-6 3′UTR-dependent luciferase expresssion, which was inhibited by mir-26 uridylation. * P=0.006 vs. control mimetic, † P=0.043 vs. mir-26 (one-way ANOVA and Bonferroni post hoc test; N=4). d, mir-26 did not affect expression from a construct containing a version of the IL-6 3′UTR in which 2 nucleotides targeted by the mir-26 seed sequence were changed. Neither the addition of mir-26b nor its uridylation had any significant effect (P=0.73, one-way ANOVA; N=4). e, Zcchc11 knockdown resulted in a loss of uridines at the 3′ termini of mir-26a. All sequences included in mir-26a analyses share 100% fidelity through the first 21 nucleotide mature miRNA sequence.