MOV10L1 Binds RNA G-Quadruplex in a Structure-Specific Manner and Resolves It More Efficiently Than MOV10

摘要

class="head no_bottom_margin" id="sec1title">IntroductionRNA helicases are ubiquitous motor enzymes that participate in nearly all aspects of RNA metabolism (, ). Despite their importance, only a few eukaryotic RNA helicases have been functionally and kinetically characterized in vitro (). The mammalian paralogs Moloney leukemia virus 10 (MOV10) and MOV10-like 1 (MOV10L1) are RNA helicases belonging to superfamily 1 (SF1), and they exhibit 5′ to 3′ RNA helicase activity in unwinding RNA duplex in vitro (, ). MOV10 is expressed in multiple tissues, and its diverse functions, such as retroelement inhibition and mRNA modulation, have been widely reported (, , ). On the other hand, MOV10L1 is a testis-specific RNA helicase with a critical function restricted to male reproduction (, , ). The MOV10L1 helicase has been demonstrated to associate with PIWI proteins and control PIWI-interacting RNA (piRNA) biogenesis for retrotransposon silencing and protect the genome integrity of post-meiotic germ cells (, ). Its point mutations K778A in the ATP-binding motif and DE888AA in the ATP hydrolysis motif cause loss of function of MOV10L1 in piRNA biogenesis and male fertility (, ). However, the molecular mechanisms and functions of MOV10L1 in piRNA biogenesis are still obscure, although studies on mammalian piRNAs have provided a framework as to how piRNAs are generated (href="#bib21" rid="bib21" class=" bibr popnode">Fu and Wang, 2014, href="#bib29" rid="bib29" class=" bibr popnode">Hirakata and Siomi, 2016, href="#bib33" rid="bib33" class=" bibr popnode">Ku and Lin, 2014, href="#bib64" rid="bib64" class=" bibr popnode">Weick and Miska, 2014). Primary precursor transcripts require at least two critical steps to generate piRNAs. A first step with endonucleolytic cleavages on the primary piRNA precursor generates piRNA precursor intermediate fragments (PPIFs), which are loaded onto PIWI proteins that stabilize their 5′ ends (href="#bib63" rid="bib63" class=" bibr popnode">Vourekas et al., 2012, href="#bib62" rid="bib62" class=" bibr popnode">Vourekas et al., 2015). This is followed by a second step with 3′ to 5′ exonucleolytic trimming by PNLDC1 (href="#bib15" rid="bib15" class=" bibr popnode">Ding et al., 2017, href="#bib45" rid="bib45" class=" bibr popnode">Nishida et al., 2018, href="#bib65" rid="bib65" class=" bibr popnode">Zhang et al., 2017). Although MOV10L1 was proposed to mediate the initial step in piRNA processing when it binds the primary precursor transcripts to facilitate their endonucleolytic cleavage (href="#bib62" rid="bib62" class=" bibr popnode">Vourekas et al., 2015), a deepened mechanistic understanding of its molecular function as a helicase is crucial in piRNA biogenesis.An intriguing feature shared by MOV10 and MOV10L1 from cross-linking and immunoprecipitation (CLIP)-seq analyses is that its preferable binding sequences harbor clusters of guanine (G) residues (href="#bib32" rid="bib32" class=" bibr popnode">Kenny et al., 2014, href="#bib62" rid="bib62" class=" bibr popnode">Vourekas et al., 2015). Their difference lies in the fact that MOV10L1-bound piRNA precursor transcripts that originate from genomic piRNA clusters are more enriched in G resides compared with those from other areas like MOV10-bound mRNAs (href="#bib62" rid="bib62" class=" bibr popnode">Vourekas et al., 2015). It has been long known that G-rich sequences in DNA and RNA have a propensity to fold into stable secondary structures termed G-quadruplexes (G4s), which are based on the stacking of several G-quartets with each layer consisting of four guanine bases held together by Hoogsteen-type hydrogen bonding (href="#bib4" rid="bib4" class=" bibr popnode">Bochman et al., 2012, href="#bib44" rid="bib44" class=" bibr popnode">Millevoi et al., 2012). Increasing evidence indicates that intramolecular RNA G-quadruplex (RG4) motifs are biologically relevant structures, and their occurrence can play vital roles in many key cellular functions, including telomere homeostasis, gene expression, and pre-mRNA processing (href="#bib1" rid="bib1" class=" bibr popnode">Agarwala et al., 2015, href="#bib8" rid="bib8" class=" bibr popnode">Bugaut and Balasubramanian, 2012, href="#bib9" rid="bib9" class=" bibr popnode">Cammas and Millevoi, 2017, href="#bib17" rid="bib17" class=" bibr popnode">Fay et al., 2017, href="#bib44" rid="bib44" class=" bibr popnode">Millevoi et al., 2012, href="#bib49" rid="bib49" class=" bibr popnode">Rhodes and Lipps, 2015, href="#bib53" rid="bib53" class=" bibr popnode">Simone et al., 2015). Even though several helicases and nucleases have been shown to remove DNA G-quadruplex (DG4) structure and regulate cellular processes (href="#bib3" rid="bib3" class=" bibr popnode">Baran et al., 1997, href="#bib12" rid="bib12" class=" bibr popnode">Chen et al., 2018, href="#bib28" rid="bib28" class=" bibr popnode">Harrington et al., 1997, href="#bib43" rid="bib43" class=" bibr popnode">Mendoza et al., 2016, href="#bib52" rid="bib52" class=" bibr popnode">Sauer and Paeschke, 2017, href="#bib58" rid="bib58" class=" bibr popnode">Sun et al., 1998, href="#bib57" rid="bib57" class=" bibr popnode">Sun et al., 1999), only a few RNA helicases, so far, have been reported to be capable of unwinding RG4 structures in vitro (href="#bib6" rid="bib6" class=" bibr popnode">Booy et al., 2012, href="#bib11" rid="bib11" class=" bibr popnode">Chakraborty and Grosse, 2011, href="#bib14" rid="bib14" class=" bibr popnode">Creacy et al., 2008, href="#bib23" rid="bib23" class=" bibr popnode">Gao et al., 2019, href="#bib42" rid="bib42" class=" bibr popnode">McRae et al., 2017, href="#bib50" rid="bib50" class=" bibr popnode">Ribeiro de Almeida et al., 2018). One of the reasons is that RG4 is a thermodynamically stable structure compared with other forms of RNA, thereby in requirement of specialized RNA helicase to resolve it (href="#bib27" rid="bib27" class=" bibr popnode">Hardin et al., 2000). The abundance of RG4s located within piRNA precursor along with an intimate coupling of piRNA precursor processing with RG4 raises the possibility that RG4 may exist as a structural mediator for the endonucleolytic cleavage of piRNA precursors, and MOV10L1 may take responsibility for resolving RG4 to facilitate the cleavage. However, whether and how the MOV10L1 helicase resolves the bona fide RG4 structures is unknown. In addition, in the unified model of PIWI-guided phased piRNA biogenesis, a question also remains how such a phasing process gets smoothly through a primary piRNA precursor bearing multiple RG4 obstacles (href="#bib22" rid="bib22" class=" bibr popnode">Gainetdinov et al., 2018, href="#bib61" rid="bib61" class=" bibr popnode">Vourekas and Mourelatos, 2018).In this study, we employed robust biochemical assays to test the capability of MOV10L1 and MOV10 in binding and unwinding of RG4 structure in vitro. We found that even though both of them can unwind RG4, there are a few striking differences between them. MOV10L1 could take advantage of its unique features in RG4 binding and unwinding to mediate piRNA biogenesis. Last, a proof-of-concept assay with MOV10L1 and an endonuclease supports a model that they might cooperatively unwind and cleave RG4 structures in the initial step of piRNA biogenesis.