It remains unclear how post-transcriptional gene silencing (PTGS) in plants discriminates aberrant RNAs from canonical messenger RNAs (mRNAs). The key step of plant PTGS is the conversion of aberrant RNAs into double-stranded RNAs by RNA-DEPENDENT RNA POLYMERASE6 (RDR6). Here, we show that RDR6 itself selects aberrant poly(A)-less mRNAs over canonical polyadenylated mRNAs as templates at the initiation step of complementary strand synthesis. This mechanism can be viewed as an innate safeguard against 'self-attack' by PTGS.

The RNA-guided endonuclease Cas9 generates a double-strand break at DNA target sites complementary to the guide RNA and has been harnessed for the development of a variety of new technologies, such as genome editing. Here, we report the crystal structures of Campylobacter jejuni Cas9 (CjCas9), one of the smallest Cas9 orthologs, in complex with an sgRNA and its target DNA. The structures provided insights into a minimal Cas9 scaffold and revealed the remarkable mechanistic diversity of the CRISPR-Cas9 systems. The CjCas9 guide RNA contains a triple-helix structure, which is distinct from known RNA triple helices, thereby expanding the natural repertoire of RNA triple helices. Furthermore, unlike the other Cas9 orthologs, CjCas9 contacts the nucleotide sequences in both the target and non-target DNA strands and recognizes the 5′-NNNVRYM-3′ as the protospacer-adjacent motif. Collectively, these findings improve our mechanistic understanding of the CRISPR-Cas9 systems and may facilitate Cas9 engineering.

PIWI-clade Argonaute proteins associate with PIWI-interacting RNAs (piRNAs) and silence transposable elements in animal gonads. Here, we report the crystal structure of a silkworm PIWI-clade Argonaute, Siwi, bound to the endogenous piRNA, at 2.4 Å resolution. Siwi adopts a bilobed architecture consisting of N-PAZ and MID-PIWI lobes, in which the 5' and 3' ends of the bound piRNA are anchored by the MID-PIWI and PAZ domains, respectively. A structural comparison of Siwi with AGO-clade Argonautes reveals notable differences in their nucleic-acid-binding channels, likely reflecting the distinct lengths of their guide RNAs and their mechanistic differences in guide RNA loading and cleavage product release. In addition, the structure reveals that Siwi and prokaryotic, but not eukaryotic, AGO-clade Argonautes share unexpected similarities, such as metal-dependent 5'-phosphate recognition and a potential structural transition during the catalytic-tetrad formation. Overall, this study provides a critical starting point toward a mechanistic understanding of piRNA-mediated transposon silencing.

Paraspeckles are nuclear bodies built on the long noncoding RNA Neat1, which regulates a variety of physiological processes including cancer progression and corpus luteum formation. To obtain further insight into the molecular basis of the function of paraspeckles, we performed fine structural analyses of these nuclear bodies using structural illumination microscopy. Notably, paraspeckle proteins are found within different layers along the radially arranged bundles of Neat1 transcripts, forming a characteristic core-shell spheroidal structure. In cells lacking the RNA binding protein Fus, paraspeckle spheroids are disassembled into smaller particles containing Neat1, which are diffusely distributed in the nucleoplasm. Sequencing analysis of RNAs purified from paraspeckles revealed that AG-rich transcripts associate with Neat1, which are distributed along the shell of the paraspeckle spheroids. We propose that paraspeckles sequester core components inside the spheroids, whereas the outer surface associates with other components in the nucleoplasm to fulfill their function.

Pluripotent stem cells can be classified into two distinct states, naïve and primed, which show different degrees of potency. One difficulty in stem cell research is the inability to distinguish these states in live cells. Studies on female mice have shown that reactivation of inactive X chromosomes occurs in the naïve state, while one of the X chromosomes is inactivated in the primed state. Therefore, we tried to distinguish the two states by monitoring X chromosome reactivation. Thus far, X chromosome reactivation has been analysed using fixed cells; here, we inserted different fluorescent reporter gene cassettes (mCherry and eGFP) into each X chromosome. Using these knock-in "MOMIJI" mice, we detected X chromosome reactivation accurately in live embryos, and confirmed that the pluripotent states of embryos were stable ex vivo, as represented by embryonic and epiblast stem cells in terms of X chromosome reactivation. Thus, MOMIJI mice provide a simple and accurate method for identifying stem cell status based on X chromosome reactivation.

The mammalian cell nucleus contains membraneless suborganelles referred to as nuclear bodies (NBs). Some NBs are formed with an architectural RNA (arcRNA) as the structural core. Here, we searched for new NBs that are built on unidentified arcRNAs by screening for ribonuclease (RNase)-sensitive NBs using 32,651 fluorescently tagged human cDNA clones. We identified 32 tagged proteins that required RNA for their localization in distinct nuclear foci. Among them, seven RNA-binding proteins commonly localized in the Sam68 nuclear body (SNB), which was disrupted by RNase treatment. Knockdown of each SNB protein revealed that SNBs are composed of two distinct RNase-sensitive substructures. One substructure is present as a distinct NB, termed the DBC1 body, in certain conditions, and the more dynamic substructure including Sam68 joins to form the intact SNB. HNRNPL acts as the adaptor to combine the two substructures and form the intact SNB through the interaction of two sets of RNA recognition motifs with the putative arcRNAs in the respective substructures.

Alternative splicing generates protein diversity essential for neuronal properties. However, the precise mechanisms underlying this process and its relevance to physiological and behavioural functions are poorly understood. To address these issues, we focused on a cassette exon of the Caenorhabditis elegans insulin receptor gene daf-2, whose proper variant expression in the taste receptor neuron ASER is critical for taste-avoidance learning. We show that inclusion of daf-2 exon 11.5 is restricted to specific neuron types, including ASER, and is controlled by a combinatorial action of evolutionarily conserved alternative splicing factors, RBFOX, CELF and PTB families of proteins. Mutations of these factors cause a learning defect, and this defect is relieved by DAF-2c (exon 11.5+) isoform expression only in a single neuron ASER. Our results provide evidence that alternative splicing regulation of a single critical gene in a single critical neuron is essential for learning ability in an organism.