ATM Signals miRNA Biogenesis Through KSRP

Abstract

In this issue of Molecular Cell, Zhang and colleagues describe a critical link between the DNA damage response and the miRNA pathway, in which DNA double strand breaks (DSBs) induce ATM-dependent KSRP phosphorylation to facilitate pri-miRNA processing.

Genomic integrity is essential for normal cellular function, tumor suppression, and organism survival. Eukaryotic cells have evolved elaborate DNA damage response (DDR) mechanisms to arrest cell cycle and activate DNA repair to maintain genomic integrity or initiate programmed cell death (apoptosis) if the damage is beyond repair (Harper and Elledge, 2007). Ataxia-telangiectasia mutated (ATM) kinase is a key DDR transducer that phosphorylates many target proteins to initiate various cellular responses (Shiloh, 2003). On page ### of this issue, Zhang et al. describe a critical link between DDR and the micro-RNA (miRNA) pathway, whereby ATM kinase signals miRNA biogenesis through phosphorylation of KH-type splicing regulatory protein (KSRP or KHSRP).

miRNAs are ∼ 22-nt cellular RNAs that control expression of target mRNAs through mRNA decay and/or translational repression (Chekulaeva and Filipowicz, 2009). As miRNAs regulate numerous biological and disease processes, miRNA expression is tightly controlled transcriptionally and post-transcriptionally. Biogenesis of miRNA is catalyzed by two multidomain RNaseIII enzymes, Drosha and Dicer (Bartel, 2004). In the nucleus, Drosha processes the primary (pri-) miRNA transcript into ∼ 60-nt stem-loop precursor (pre-) miRNA. In the cytoplasm, Dicer further cleaves pre-miRNA into mature miRNA. KSRP associates with both Drosha and Dicer and post-transcriptionally regulate the biogenesis of miRNAs (Trabucchi et al., 2009).

It has been reported that DNA damage (e.g. DSBs) induces changes in miRNA expression (Suzuki et al., 2009).To understand the mechanism underlying this phenomenon, Zhang et al. used a radiomimetic drug, neocarzinostatin (NCS), to generate DSBs in Atm+/+ and Atm-/- mouse embryonic fibroblasts (MEFs). miRNA microarray analysis revealed that 71 miRNAs, approximately one fourth of the identified mouse miRNAs, are significantly up-regulated (≥ 2 fold) in the NCS-treated Atm+/+ MEFs, but not in Atm-/- MEFs. This ATM-dependent miRNA induction is likely controlled by a post-transcriptional mechanism as suggested by the lack of significant change in pri-miRNA levels after NCS treatment.

Remarkably, all of the previously reported KSRP-dependent miRNAs (29 miRNAs) were induced in the Atm+/+ MEFs upon NCS treatment (Trabucchi et al., 2009). Knocking down KSRP in U2OS cells abolished NCS-induced up-regulation of these miRNAs. Focusing on this group of ATM- and KSRP-dependent miRNAs, Zhang et al. postulated that KSRP might be a direct substrate of ATM kinase. Indeed, in vivo phosphor-labeling showed that DSBs induce KSRP phosphorylation in an ATM-dependent manner in human fibroblast cells. ATM co-immunoprecipitated KSRP with or without NCS treatment. Immunopurified wild-type, but not kinease-dead, ATM phosphorylated purified KSRP in vitro, suggesting that ATM directly phosphorylate KSRP in response to DSBs (Figure 1).

Figure 1. ATM-dependent KSRP phosphorylation promotes pri-miRNA processing.

DNA double strand breaks (DSBs) activate the ATM kinase, which in turn phosphorylates one of its substrates, KSRP. Phosphorylation of KSRP enhances its binding to target pri-miRNAs and the recruitment of these pri-miRNAs to Drosha-DGCR8 complex for processing. In addition, ATM may phosphorylate other target proteins that also contribute to DSB-induced miRNA biogenesis. By regulating target gene expression, the induced miRNAs may facilitate DNA damage response (DDR) or provide a negative feedback loop to dampen DDR.

Since ATM phosphorylates its substrates at the consensus BXBS/TQ site, bioinformatic analysis predicted three serine residues in KSRP, which are highly conserved in mammals, as the potential ATM phosphorylation sites. To functionally validate these phosphorylation sites, individual and triple phospho-mutant (serine to alanine) or phospho-mimic (serine to aspartic acid) KSRP variants were generated and overexpressed in the KSRP-deficient cells. Only wild-type, not phospho-mutant, KSRP rescued the DNA damage-induced miRNA biogenesis. On the other hand, phospho-mimic KSRP induced miRNA biogenesis in the absence of DNA damage. These studies indicate that ATM regulates miRNA biogenesis through phosphorylation of KSRP. Surprisingly, expression of phospho-mimic KSRP failed to up-regulate its target miRNAs in the ATM-deficient cells. Thus, besides phosphorylation of KSRP, ATM appeared to play another unidentified role, such as phosphorylation of other unknown targets, for the induction of KSRP-dependent miRNAs (Figure 1).

Previous studies showed that KSRP facilitates the recruitment of specific pri-miRNAs to Drosha by recognizing the GGG triplets in the terminal loop region (Trabucchi et al., 2009). Phosphorylation of KSRP significantly enhanced its activity in recruiting target pri-miRNAs to Drosha for processing (Figure 1). It will be interesting to examine if phospho-mimic KSRP still displays enhanced affinity for pri-miRNA in the ATM-depleted cells. Biochemical and biophysical studies of recombinant wild-type and phospho-mimic KSRP will provide further mechanistic insight into how phosphorylation of KSRP regulates its pri-miRNA binding activity. Modulating activities of specific RNA binding proteins that recruit pri-miRNA to Drosha likely represent a general mechanism for how different signaling systems target miRNA pathway to achieve biological responses.

This elegant study by Zhang et al. not only provides a clear mechanism for how DSBs regulate miRNA expression, but also raises many interesting questions for future studies. First, since the KSRP-dependent miRNAs are only a subset of the DSB-induced miRNAs, alternative mechanisms must exist to account for the induction of the KSRP-independent miRNAs. For instance, upon DNA damage, p53 up-regulates the transcription of pri-miR34a-c (He et al., 2007) and the processing of a subset of pri-miRNAs through association with the p68 RNA helicase (Suzuki et al., 2009). It will be equally interesting to study the post-transcriptional mechanisms of DNA damage-induced down-regulation of miRNAs. Second, it is important to understand the functions of these ATM-dependent miRNAs by identifying their target genes. Two possible functions of the DDR-regulated miRNAs might be: 1) contribute to DDR by facilitating DNA repair, cell cycle arrest, and apoptosis (Figure 1). 2) act as a negative feedback loop to dampen DDR when the repair is completed (Ivanovska et al., 2008; Le et al., 2009; Wang et al., 2009). Furthermore, eukaryotic cells are constantly challenged by a variety of environmental and intracellular genotoxic stress. The current study represents a nice template for studying how other types of DNA damage regulate the miRNA machinery. Finally, miRNAs govern many biological and disease processes, therefore, connecting the miRNA pathway to different signaling systems will be an exciting area for future research.