Functional Analysis of lncRNAs

To investigate the physiological significance of the acquisition of species-specific nuclear abundant lncRNAs, we have created knockout (KO) mice that lack Neat1, Gomafu, and Malat1 expression. Phenotypic analyses of these mice have revealed that none of these abundant nuclear lncRNAs are essential for viability, but all of them played important biological roles when the animals were placed in a specific conditions/environments.

Neat1 - an architectural lncRNA that supports female reproduction

Neat1 is an lncRNA that serves as an architectural component of the nuclear body paraspeckle (reviewed in Nakagawa and Hirose, 2012). Neat1 is highly expressed in the corpus luteum in the female ovary (Nakagawa et al., 2011), and the fertility of Neat1 KO mice was severely decreased due to the lack of corpus luteum formation and the consequent decrease in levels of serum progesterone, a steroid hormone essential for the establishment and maintenance of pregnancy (Nakagawa et al., 2014).

Interestingly, only half of the copulation events lead to conception failure, and even the same female mouse that failed to become pregnant can become pregnant from a subsequent copulation. These observations suggested that Neat1 and nuclear body paraspeckle is stochastically required for the formation of the pregnant corpus luteum under certain environments, of which the particular condition still remains unknown.

Gomafu - a neuronal lncRNA controls animal behavior

Gomafu is a neuron-specific lncRNA that forms unique nuclear bodies (Sone et al., 2007). Gomafu is strongly expressed in neurons, and expression is controlled by neuronal activities. The Gomafu KO mice exhibit marginal hyper-locomotive activities in a series of behavior tests, whereas anxiety-related behavior or learning and memory are not significantly affected (Ip et al., 2016).

Identification of hnRNP U as the chromosomal localization factor of Xist

Xist is the master regulator of X chromosome inactivation, an epigenetic mechanism that silences one of the two X chromosomes in mammalian females. Notably, Xist accumulate on the entire inactive X chromosome (Xi), resulting in the recruitment of chromatin-modifying complexes and subsequent heterochromatin formation. While basic molecular processes leading to the formation of Xi have been well described, it has long remained elusive how Xist RNA transcripts associate with the chromosome. To identify the chromosomal localization factor of Xist, we hypothesized that a certain RNA-binding protein tethers the transcripts to the chromosome. Thus, we prepared a custom siRNA library designed against abundant RNA-binding proteins and examined if the localization of Xist RNA was affected upon knockdown of each gene. We found that the siRNA-mediated knockdown of hnRNP U leads to dissociation of Xist RNA into the nucleoplasm. This effect was rescued by introduction of full-length hnRNP U, but not the deletion mutant molecules of hnRNP U lacking either its DNA- or RNA-binding domains, suggesting that hnRNP U regulates the chromosomal localization of Xist RNA through its bipartite nucleic-acid binding domains. We also found that hnRNP U is required for the formation of an inactive X chromosome, suggesting that hnRNP U-mediated chromosomal localization of Xist RNA is an essential step during X chromosome inactivation.

Gomafu regulates splicing kinetics via association with SF1

To obtain insight into the molecular function of Gomafu, we tried to identify protein that interact with the noncoding RNA. We initially searched sequence motif conserved between chicken, mouse, and human Gomafu, and found that all of the three genes contained a tandem repeat of UACUAAC, which is the essential and conserved intron branch point sequence. We then identified a general splicing factor SF1 as a Gomafu binding protein through affinity purification using the repeat sequence as a bait, and subsequent in vitro splicing assays revealed that the Gomafu repeat cause a distinct delay in the formation of the early spliceosome when using a IgM intron as a substrate. Notably, this effect was not observed with an intron substrate from adenovirus, suggesting that Gomafu regulates kinetics of splicing reaction of a subset of introns. Taken together, we propose that the Gomafu RNA regulates splicing efficiency by changing the local concentration of splicing factors within the nucleus (Tsuiji et al. 2011).

Formation of a novel nuclear body by Gomafu-associating proteins

To identify factor that can regulate localization or stability of Gomafu, we screened custom siRNA library designed against abundant RNA binding proteins. We found knockdown of Celf3, previously identified as a member of CUGBP1 (CUG repeat binding protein 1) protein family, leads to marked decrease of Gomafu expression. Cross-link immunoprecipitation analysis using antibodies raised against Celf3 confirmed specific interaction between Gomafu and Celf3. Interestingly, Celf3 formed novel nuclear bodies (CS bodies) in a neuroblastoma cell line Neuro2A, which localization was coincided with SF1, another Gomafu-interacting protein. On the other hand, Gomafu did not accumulate in the CS bodies but separately distributed throught the nucleus. We thus speculate that Gomafu indirectly modulate the function of RNA binding proteins in CS bodies by sequestering them in separate regions in the nucleus.

Analyses of molecular functions of clade-specific noncoding RNA 4.5SH

As described above, lnRNAs are more species-specific compared to protein-coding mRNAs. We have recently performed a series of functional analyses on 4.5SH, an lncRNA specific to the Myodonta clade (a group of rodents with small sizes and short lives) and highly homologous to the retrotransposon SINE B1. 4.5SH is localized in the nuclear compartment called nuclear speckles, which contain various types of splicing factors as well as pre- and processed mRNAs. We speculated that nuclear 4.5SH may display base pairing with target poly-A (+) mRNAs that contain an antisense insertion of SINE B1, which is highly homologous to 4.5SH. We also hypothesized that the formation of inter-molecular double-stranded RNA (dsRNA) structures lead to decreased gene expression via nuclear retention, as has been proposed for intra-molecular dsRNA structures. We found that 4.5SH inhibits the nuclear export of mRNAs containing an antisense insertion of the retrotransposon SINE B1 and thereby downregulates the expression of these mRNAs. Considering that the emergence of 4.5SH cluster genes occurred later than the expansion of the retrotransposon SINE B1, we propose that the emergence of the 4.5SH gene cluster in specific rodent species rapidly changed the cryptic genetic variations into phenotypic variations via regulation of the expression of genes containing antisense insertions of SINE B1s.