Oligonucleotide-targeted degradation of U1 and U2 snRNAs reveals differential interactions of simian virus 40 pre-mRNAs with snRNPs

Genetics and biochemistry remain essential in the structural era of the spliceosome

The spliceosome is not a single macromolecular machine. Rather it is a collection of dynamic heterogeneous subcomplexes that rapidly interconvert throughout the course of a typical splicing cycle. Because of this, for many years the only high resolution structures of the spliceosome available were of smaller, isolated protein or RNA components. Consequently much of our current understanding of the spliceosome derives from biochemical and genetic techniques. Now with the publication of multiple, high resolution structures of the spliceosome, some question the relevance of traditional biochemical and genetic techniques to the splicing field. We argue such techniques are not only relevant, but vital for an in depth mechanistic understanding of pre-mRNA splicing.

Mitotic noncoding RNA processing promotes kinetochore and spindle assembly in Xenopus

Grenfell et al. show that transcription and RNA processing occur in metaphase-arrested egg extracts and that noncoding RNA biogenesis is important for centromere, kinetochore, and mitotic spindle assembly.

Transcription at the centromere of chromosomes plays an important role in kinetochore assembly in many eukaryotes, and noncoding RNAs contribute to activation of the mitotic kinase Aurora B. However, little is known about how mitotic RNA processing contributes to spindle assembly. We found that inhibition of transcription initiation or RNA splicing, but not translation, leads to spindle defects in Xenopus egg extracts. Spliceosome inhibition resulted in the accumulation of high molecular weight centromeric transcripts, concomitant with decreased recruitment of the centromere and kinetochore proteins CENP-A, CENP-C, and NDC80 to mitotic chromosomes. In addition, blocking transcript synthesis or processing during mitosis caused accumulation of MCAK, a microtubule depolymerase, on the spindle, indicating misregulation of Aurora B. These findings suggest that co-transcriptional recruitment of the RNA processing machinery to nascent mitotic transcripts is an important step in kinetochore and spindle assembly and challenge the idea that RNA processing is globally repressed during mitosis.

Complementation of U4 snRNA in S. cerevisiae splicing extracts for biochemical studies of snRNP assembly and function

Pre-messenger RNA splicing is a surprisingly complex and dynamic process, the details of which remain largely unknown. One important method for studying splicing involves the replacement of endogenous splicing components with their synthetic counterparts. This enables changes in protein or nucleic acid sequence to be tested for functional effects, as well as the introduction of chemical moieties such as cross-linking groups and fluorescent dyes. To introduce the modified component, the endogenous one must be removed and a method found to reconstitute the active splicing machinery. In extracts prepared from S. cerevisiae, reconstitution has been accomplished with the small, nuclear RNAs U6, U2, and U5.We describe a comparable method to reconstitute active U4 small, nuclear RNA (snRNA) into a splicing extract. In order to remove the endogenous U4 it is necessary to target it for oligo-directed RNase H degradation while active splicing is under way, i.e., in the presence of a splicing transcript and ATP. This allows complete degradation of endogenous U4 and subsequent replacement with an exogenous version. In contrast to the procedures described for depletion of U6, U2, or U5 snRNAs, depletion of U4 requires concurrent active splicing. The ability to reconstitute U4 in yeast extract allows a variety of structural and functional studies to be carried out.

Mechanistic insights into human pre-mRNA splicing of human ultra-short introns: potential unusual mechanism identifies G-rich introns

It is unknown how very short introns (<65 nt; termed 'ultra-short' introns) could be spliced in a massive spliceosome (>2.7 MDa) without steric hindrance. By screening an annotated human transcriptome database (H-InvDB), we identified three model ultra-short introns: the 56-nt intron in the HNRNPH1 (hnRNP H1) gene, the 49-nt intron in the NDOR1 (NADPH dependent diflavin oxidoreductase 1) gene, and the 43-nt intron in the ESRP2 (epithelial splicing regulatory protein 2) gene. We verified that these endogenous ultra-short introns are spliced, and also recapitulated this in cultured cells transfected with the corresponding mini-genes. The splicing of these ultra-short introns was repressed by a splicing inhibitor, spliceostatin A, suggesting that SF3b (a U2 snRNP component) is involved in their splicing processes. The 56-nt intron containing a pyrimidine-rich tract was spliced out in a lariat form, and this splicing was inhibited by the disruption of U1, U2, or U4 snRNA. In contrast, the 49- and 43-nt introns were purine-rich overall without any pyrimidine-rich tract, and these lariat RNAs were not detectable. Remarkably, shared G-rich intronic sequences in the 49- and 43-nt introns were required for their splicing, suggesting that these ultra-short introns may recruit a novel auxiliary splicing mechanism linked to G-rich intronic splicing enhancers.

Splicing-independent recruitment of spliceosomal small nuclear RNPs to nascent RNA polymerase II transcripts

In amphibian oocytes, most lateral loops of the lampbrush chromosomes correspond to active transcriptional sites for RNA polymerase II. We show that newly assembled small nuclear ribonucleoprotein (RNP [snRNP]) particles, which are formed upon cytoplasmic injection of fluorescently labeled spliceosomal small nuclear RNAs (snRNAs), target the nascent transcripts of the chromosomal loops. With this new targeting assay, we demonstrate that nonfunctional forms of U1 and U2 snRNAs still associate with the active transcriptional units. In particular, we find that their association with nascent RNP fibrils is independent of their base pairing with pre–messenger RNAs. Additionally, stem loop I of the U1 snRNA is identified as a discrete domain that is both necessary and sufficient for association with nascent transcripts. Finally, in oocytes deficient in splicing, the recruitment of U1, U4, and U5 snRNPs to transcriptional units is not affected. Collectively, these data indicate that the recruitment of snRNPs to nascent transcripts and the assembly of the spliceosome are uncoupled events.

Identity elements used in export of mRNAs

Different classes of RNA are exported from the nucleus by distinct factors. We demonstrate that U1 snRNA is exported like an mRNA on insertion of a pre-mRNA intron or either sense or antisense mRNA exon sequences. mRNA-specific factors are recruited onto the spliced or elongated U1 RNA whereas U snRNA-specific factors are not, suggesting that an unstructured region of sufficient length in an RNA acts as a dominant determinant of mRNA identity. After export, spliced U1 RNA undergoes cytoplasmic maturation but is not reimported into the nucleus. These data provide insight into mechanisms for discrimination of different classes of nuclear RNA and demonstrate that two RNAs of identical sequence can have distinct cytoplasmic fates depending on their mode of export.

Three small nucleolar RNAs that are involved in ribosomal RNA precursor processing

Three small nucleolar RNAs (snoRNAs), E1, E2 and E3, have been described that have unique sequences and interact directly with unique segments of pre-rRNA in vivo. In this report, injection of antisense oligodeoxynucleotides into Xenopus laevis oocytes was used to target the specific degradation of these snoRNAs. Specific disruptions of pre-rRNA processing were then observed, which were reversed by injection of the corresponding in vitro-synthesized snoRNA. Degradation of each of these three snoRNAs produced a unique rRNA maturation phenotype. E1 RNA depletion shut down 18 rRNA formation, without overaccumulation of 20S pre-rRNA. After E2 RNA degradation, production of 18S rRNA and 36S pre-rRNA stopped, and 38S pre-rRNA accumulated, without overaccumulation of 20S pre-rRNA. E3 RNA depletion induced the accumulation of 36S pre-rRNA. This suggests that each of these snoRNAs plays a different role in pre-rRNA processing and indicates that E1 and E2 RNAs are essential for 18S rRNA formation. The available data support the proposal that these snoRNAs are at least involved in pre-rRNA processing at the following pre-rRNA cleavage sites: E1 at the 5' end and E2 at the 3' end of 18S rRNA, and E3 at or near the 5' end of 5.8S rRNA.

Disruption of pre-mRNA splicing in vivo results in reorganization of splicing factors

We have examined the functional significance of the organization of pre-mRNA splicing factors in a speckled distribution in the mammalian cell nucleus. Upon microinjection into living cells of oligonucleotides or antibodies that inhibit pre-mRNA splicing in vitro, we observed major changes in the organization of splicing factors in vivo. Interchromatin granule clusters became uniform in shape, decreased in number, and increased in both size and content of splicing factors, as measured by immunofluorescence. These changes were transient and the organization of splicing factors returned to their normal distribution by 24 h following microinjection. Microinjection of these oligonucleotides or antibodies also resulted in a reduction of transcription in vivo, but the oligonucleotides did not inhibit transcription in vitro. Control oligonucleotides did not disrupt splicing or transcription in vivo. We propose that the reorganization of splicing factors we observed is the result of the inhibition of splicing in vivo.

Localization of pre-messenger RNA at discrete nuclear sites

We have studied the nuclear localization of rhodamine-labeled pre-mRNA after microinjection into nuclei of cultured rat kidney epithelial cells. Intranuclear localization of the injected RNA was followed in the living cells by fluorescence microscopy and digital image processing. Injected human beta-globin pre-mRNA became localized in 30-60 discrete nuclear sites that were coincident with loci defined by monoclonal antibodies against small nuclear ribonucleoproteins (Sm) or another spliceosome component (SC-35) in parallel immunocytochemical studies on the same nuclei. Similar patterns of nuclear localization were observed with a rat proenkephalin pre-mRNA. Nuclear microinjection of an intronlacking beta-globin RNA, a splicing-defective beta-globin mutant pre-mRNA, or an antisense beta-globin pre-mRNA did not result in localization at discrete sites. These results indicate that pre-mRNA binds preferentially to nuclear Sm and SC-35 antibody-reactive sites in vivo and that the binding requires intron sequences.

Trans-activating rev protein of the human immunodeficiency virus 1 interacts directly and specifically with its target RNA

The 20-kDa phosphorylated rev protein from human immunodeficiency virus 1 has been shown to transactivate posttranscriptionally the expression of viral structural proteins by selective stabilization and nuclear export of unspliced and incompletely spliced viral mRNA. We could demonstrate in gel-mobility and immunoprecipitation assays that the recombinant rev protein purified from a baculovirus expression system forms a distinct and specific complex with its target RNA (rev-responsive element), a 234-nucleotide sequence within the envelope coding region of human immunodeficiency virus 1. No complex formation could be observed using RNAs with similar secondary structure nor with other human immunodeficiency virus 1 recombinant proteins. Deletion analysis mapped this specific binding to the first 90 nucleotides of this rev-responsive element, which contains a U2 small nuclear RNA homologous region. We propose that the specific binding of rev to its target RNA sequence plays an essential part in releasing an incompletely spliced viral mRNA containing this target sequence to the cytoplasm.