Molecular consequences of specific intron mutations on yeast mRNA splicing in vivo and in vitro

We have altered the TACTAAC sequence in the yeast CYH2m gene intron to TACTACC. This mutation changes the nucleotide at the normal position of the branch in intron RNA lariats produced during pre-mRNA splicing, and it prevents splicing in vivo. In a yeast pre-mRNA splicing system, CYH2m pre-mRNA carrying the TACTACC mutation is not specifically cut or rearranged in any way. Substitution of an A for the first G of the CYH2m intron, converting the highly conserved GTATGT 5' splice site sequence to ATATGT, also blocks intron excision in vivo and in vitro: pre-mRNA carrying this mutation was still cut normally at the mutant 5' splice site in vitro, to give authentic exon 1 and an intron-exon 2 lariat RNA with an A-A 2'-5' phosphodiester linkage at the branch point. This lariat RNA is a dead-end product. The subsequent cleavage at the 3' splice site is therefore sensitive to the sequence of the 5' end of the intron attached at the branch point.     

Trans-complementation of the second step of pre-mRNA splicing by exogenous 5' exons.

During splicing of nuclear pre-mRNAs, the first step liberates the 5' exon (exon 1) and yields a lariat intron-3'exon (intron-exon 2) intermediate. The second step results in exon ligation. Previous results indicated that severe truncations of the 5' exon of the actin pre-mRNA result in a block to the second splicing step in vitro in yeast extracts, leading to an accumulation of intron-exon 2 lariat intermediates. We show that exogenous exon 1 RNA oligonucleotides can chase these stalled intermediates into lariat intron and spliced exons. This reaction requires some of the cis elements and trans-acting factors that are required for a normal second step. There is no strong sequence requirement for the exon 1 added in trans, but oligonucleotides with complementarity to the U5 snRNA conserved loop perform the chase more efficiently. Using a dominant negative mutant of the DEAH-box ATPase Prp16p and ATP depletion, we show that the stalled intermediate is blocked after the Prp16p-dependent step. These results show that exogenous RNAs with various sequences but containing no splicing signals can be incorporated into spliceosomes and undergo RNA recombination and exon shuffling during the second step of pre-mRNA splicing.     

Group II intron RNA-catalyzed recombination of RNA in vitro.

We report the first evidence for a novel reaction mediated by the self-splicing yeast mitochondrial group II intron bl1; the site-specific recombination of RNA molecules in vitro. Upon incubation of the intron lariat with two different RNAs, each harbouring a short sequence complementary to exon binding site 1 (EBS1) of the intron, novel recombined RNAs are formed. As a result of this intron-mediated shuffling of gene segments, the 5' part of RNA1 is ligated to the 3' part of RNA2 and, reciprocally, the 5' part of RNA2 to the 3' part of RNA1. Sequence analysis of the recombinant junction shows that the site of recombination is precisely located 3' to intron binding site 1 (IBS1). The hypothesized mechanism of recombination involves exchange of RNA 5' parts after the first step of a reverse splicing reaction. The possible role of this mechanism in vivo and during prebiotic evolution is discussed.     

Ribonucleoprotein complex formation during pre-mRNA splicing in vitro.

The ribonucleoprotein (RNP) structures of the pre-mRNA and RNA processing products generated during in vitro splicing of an SP6/beta-globin pre-mRNA were characterized by sucrose gradient sedimentation analysis. Early, during the initial lag phase of the splicing reaction, the pre-mRNA sedimented heterogeneously but was detected in both 40S and 60S RNP complexes. An RNA substrate lacking a 3' splice site consensus sequence was not assembled into the 60S RNP complex. The two splicing intermediates, the first exon RNA species and an RNA species containing the intron and the second exon in a lariat configuration (IVS1-exon 2 RNA species), were found exclusively in a 60S RNP complex. These two splicing intermediates cosedimented under a variety of conditions, indicating that they are contained in the same RNP complex. The products of the splicing reaction, accurately spliced RNA and the excised IVS1 lariat RNA species, are released from the 60S RNP complex and detected in smaller RNP complexes. Sequence-specific RNA-factor interactions within these RNP complexes were evidenced by the preferential protection of the pre-mRNA branch point from RNase A digestion and protection of the 2'-5' phosphodiester bond of the lariat RNA species from enzymatic debranching. The various RNP complexes were further characterized and could be distinguished by immunoprecipitation with anti-Sm and anti-(U1)RNP antibodies.     

Self-splicing of a group II intron in yeast mitochondria: dependence on 5'' exon sequences.

It has been previously suggested that self-splicing of group II introns starts with a nucleophilic attack of the 2' OH group from the branchpoint adenosine on the 5' splice junction. To investigate the sequences governing the specificity of this attack, a series of Bal31 nuclease deletion mutants was constructed in which progressively larger amounts of 5' exon have been removed starting from its 5' end. The ability of mutant RNAs to carry out self-splicing in vitro was studied. Involvement of 5' exon sequences in self-splicing activity is indicated by the fact that a mutant in which as many as 18 nucleotides of 5' exon remain is seriously disturbed in splicing, while larger deletions eliminate splicing entirely. Mutants containing a truncated 5' exon form aberrant RNAs. One of these is a 425-nucleotide RNA containing the 5' exon as well as sequences of the 5' part of the intron. Its 3' end maps at position 374 of the 887-nucleotide intron. The other is a less abundant lariat RNA probably originating from the remainder of the intron linked to the 3' exon. We interpret this large dependence of reactivity of the intron on 5' exon and adjoining intron sequences as evidence for base-pairing interactions between the exon and parts of the intron, leading to an RNA folding necessary for splicing. Possible folding models are discussed.     

Effect of 5'' splice site mutations on splicing of the preceding intron.

Three exon constructs containing identical intron and exon sequences were mutated at the 5' splice site beginning intron 2 and assayed for the effect of the mutation on splicing of the upstream intron in vitro. Alteration of two or six bases within the 5' splice site reduced removal of intron 1 at least 20-fold, as determined by quantitation of either spliced product or released lariat RNA. The prominent product was skip splicing of exon 1 to exon 3. Examination of complex formation indicated that mutation of the 5' splice site terminating exon 2 depressed the ability of precursor RNAs containing just the affected exon to direct assembly in vitro. These results suggest that mutation at the end of an internal exon inhibits the ability of the exon to be recognized by splicing factors. A comparison of the known vertebrate 5' splice site mutations in which the mutation resides at the end of an internal exon indicated that exon skipping is the preferred phenotype for this type of mutation, in agreement with the in vitro observation reported here. Inhibition of splicing by mutation at the distal and of the exon supports the suggestion that exons, rather than splice sites, are the recognition units for assembly of the spliceosome.     

In vitro splicing of simian virus 40 early pre mRNA.

The products of splicing of simian virus 40 early pre mRNA in HeLa cell nuclear extracts have been characterized. Of the two alternative splicing patterns exhibited by this precursor in vivo, which involve the use of alternative large T and small t 5' splice sites and a single shared 3' splice site, only one, producing large T mRNA, was found to occur in vitro. A number of possible intermediates and byproducts of splicing of large T mRNA were observed, including free large T 5' exon, lariat form intron joined to 3' exon and free lariat and linear forms of large T intron. The formation of these products argues strongly for a basic similarity in the mechanism underlying large T and other, non-alternative splices. A collection of RNAs resulting from protection of early pre mRNA at specific points from an endogenous 5' to 3' exonuclease activity in vitro have also been observed. The regions of the precursor RNA protected map to positions immediately upstream of the 5' splice sites of large T and small t and the lariat branchpoint, and may represent interaction of these regions with components of the splicing machinery.     

Impairment of yeast pre-mRNA splicing by potential secondary structure-forming sequences near the conserved branchpoint sequence.

The absolutely conserved TACTAAC box within introns of RNA polymerase II-transcribed genes of the yeast Saccharomyces cerevisiae serves an indispensable role in lariat formation. We show in this report that rather short palindromic sequences inserted into the yeast actin gene intron immediately 3' to the TACTAAC box block the second but not the first splicing step. In contrast, a palindromic sequence inserted some 23 bp 3' of the TACTAAC box did not affect correct and efficient splicing. The data suggest that hairpin structures that might form adjacent to the branchsite sequence interfere with some necessary alteration of the spliceosome required for 3' intron cleavage and exon ligation.     

DNA polymerization catalysed by a group II intron RNA in vitro.

The excised group II intron bI1 from Saccharomyces cerevisiae can act as a ribozyme catalysing various chemical reactions with different substrate RNAs in vitro . Recently, we have described an editing-like RNA polymerization reaction catalysed by the bI1 intron lariat that proceeds in the 3'-->5'direction. Here we show that the bI1 lariat RNA can also catalyse successive deoxyribonucleotide polymerization reactions on exogenous substrate molecules. The basic mechanism of the reaction involved interacting cycles between an alternative version of partial reverse splicing (lariat charging) and canonical forward splicing (lariat discharging by exon ligation). With an overall chain growth in the 3'-->5' direction, the 5' exon RNAs (IBS1dN) were elongated by successive insertion of deoxyribonucleotides derived from single deoxyribonucleotide substitutions (dA, dG, dC or dT). All four deoxyribonucleotides were used as substrates, although with different efficiencies. Our findings extend the catalytic repertoire of group II intron RNAs not only by a novel DNA polymerization activity, but also by a DNA-DNA ligation capacity, supporting the idea that ribozymes might have been part of the first primordial polymerization machinery for both RNA and DNA.     

Fate of the junction phosphate in alternating forward and reverse self-splicing reactions of group II intron RNA.

The RNA-catalysed self-splicing reaction of group II intron RNA is assumed to proceed by two consecutive transesterification steps, accompanied by lariat formation. This is effectively analogous to the small nuclear ribonucleoprotein (snRNP)-mediated nuclear pre-mRNA splicing process. Upon excision from pre-RNA, a group II lariat intervening sequence (IVS) has the capacity to re-integrate into its cognate exons, reconstituting the original pre-RNA. The process of reverse self-splicing is presumed to be a true reversion of both transesterification steps used in forward splicing. To investigate the fate of the esterified phosphate groups in splicing we assayed various exon substrates (5'E-*p3'E) containing a unique 32P-labelled phosphodiester at the ligation junction. In combined studies of alternating reverse and forward splicing we have demonstrated that the labelled phosphorus atom is displaced in conjunction with the 3' exon from the ligation junction to the 3' splice site and vice versa. Neither the nature of the 3' exon sequence nor its sequence composition acts as a prominent determinant for both substrate specificity and site-specific transesterification reactions catalysed by bI1 IVS. A cytosine ribonucleotide (pCp; pCOH) or even deoxyoligonucleotides could function as an efficient substitute for the authentic 3' exon in reverse and in forward splicing. Furthermore, the 3' exon can be single monophosphate group. Upon incubation of 3' phosphorylated 5' exon substrate (5'E-*p) with lariat IVS the 3'-terminal phosphate group is transferred in reverse and forward splicing like an authentic 3' exon, but with lower efficiency. In the absence of 3' exon nucleotides, it appears that substrate specificity is provided predominantly by the base-pairing interactions of the intronic exon binding site (EBS) sequences with the intron binding site (IBS) sequences in the 5' exon. These studies substantiate the predicted transesterification pathway in forward and reverse splicing and extend the catalytic repertoire of group II IVS in that they can act as a potential and sequence-specific transferase in vitro.     

Bimolecular exon ligation by the human spliceosome bypasses early 3' splice site AG recognition and requires NTP hydrolysis

Here we report further characterization of an in vitro assay system for exon ligation by the human spliceosome in which the 3' splice site AG is supplied by a different RNA molecule than that containing the 5' splice and branch sites. By varying the time during splicing reactions when the 3' splice site AG is made available to the splicing machinery, we show that AG recognition need not occur until after lariat formation. Thus an early AG recognition event required for spliceosome formation and lariat formation on some mammalian introns is not required for exon ligation. Depletion/add-back studies and cold competitor challenge experiments reveal that commitment of a 3' splice site AG to exon ligation requires NTP hydrolysis. Because it both physically and kinetically uncouples exon ligation from spliceosome assembly and lariat formation, the bimolecular system will be a valuable tool for further mechanistic analysis of the second step of splicing.     

Integration of group II intron bI1 into a foreign RNA by reversal of the self-splicing reaction in vitro

Group II intron bI1, the first intron of the COB gene in the mitochondria of S. cerevisiae, is able to self-splice in vitro with the basic pathway similar to nuclear pre-mRNA splicing. We show that incubation of the intron lariat with ligated exons bE1 and bE2 leads to a complete reversal of the splicing reaction. The integration of the intron into the ligated exons is correct; the reconstituted preRNA of the reverse reaction can undergo a self-splicing reaction anew. When incubated with a foreign RNA species bearing a sequence motif that is complementary to exon binding site 1, the lariat can integrate into this RNA with the position of insertion immediately downstream of this sequence. This result implies that transposition of group II introns on the RNA level by reversal of the splicing reaction is, in principle, conceivable.     

Intron sequences involved in lariat formation during pre-mRNA splicing

We have shown that lariat formation during in vitro splicing of several RNA precursors, from Drosophila to man, occurs at a unique and identifiable but weakly conserved site, 18 to 37 nucleotides proximal to the 3' splice site. Lariat formation within an artificial intron lacking a normal branch-point sequence occurs at a cryptic site a conserved distance (approximately 23 nucleotides) from the 3' splice site. Analysis of beta-thalassemia splicing mutations revealed that lariat formation in the first intron of the human beta-globin gene occurs at the same site in normal and mutant precursors, even though alternate 5' and 3' splice sites are utilized in the mutants. Remarkably, cleavage at the 5' splice site and lariat formation do not occur when the precursor contains a beta-thalassemia deletion removing the polypyrimidine stretch and AG dinucleotide at the 3' splice site. In contrast, a single base substitution in the AG dinucleotide blocks cleavage at the 3' splice site but not at the 5' site.     

Multiple exon-binding sites in class II self-splicing introns

Partial deletion of the exon 5' to S. cerevisiae intron a5, a self-splicing mitochondrial class II intron, reveals the existence of several sites of intron-exon interaction. We have identified two of the corresponding exon-binding sites in intron a5 by comparative sequence analysis and RNAase H digestion of the intron complexed to a DNA version of its 5' exon. Introduction of mutations in either the intronic sites or the complementary exonic sequences affects splicing in vitro, whereas double mutants in which intron-exon pairings have been restored show normal activity. Some of the mutants accumulate a product that was shown to be the intron-3' exon lariat, a postulated splicing intermediate. The possible role of one of the intronic sites in aligning exons for the ligation step is discussed.     

Spatial organization of protein-RNA interactions in the branch site-3' splice site region during pre-mRNA splicing in yeast

A series of efficiently spliced pre-mRNA substrates containing single 4-thiouridine residues were used to monitor RNA-protein interactions involving the branch site-3' splice site-3' exon region during yeast pre-mRNA splicing through cross-linking analysis. Prior to the assembly of the prespliceosome, Mud2p and the branch point bridging protein cross-link to a portion of this region in an ATP-independent fashion. Assembly of the prespliceosome leads to extensive cross-linking of the U2-associated protein Hsh155p to this region. Following the first step of splicing and in a manner independent of Prp16p, the U5 small nuclear ribonucleoprotein particle-associated protein Prp8p also associates extensively with the branch site-3' splice site-3' exon region. The subsequent cross-linking of Prp16p to the lariat intermediate is restricted to the 3' splice site and the adjacent 3' exon sequence. Using modified substrates to either mutationally or chemically block the second step, we found that the association of Prp22p with the lariat intermediate represents an authentic transient intermediate and appears to be restricted to the last eight intron nucleotides. Completion of the second step leads to the cross-linking of an unidentified approximately 80-kDa protein near the branch site sequence, suggesting a potential role for this protein in a later step in intron metabolism. Taken together, these data provide a detailed portrayal of the dynamic associations of proteins with the branch site-3' splice site region during spliceosome assembly and catalysis.     

U5 snRNA interacts with exon sequences at 5' and 3' splice sites.

U5 snRNA is an essential pre-mRNA splicing factor whose function remains enigmatic. Specific mutations in a conserved single-stranded loop sequence in yeast U5 snRNA can activate cleavage of G1----A mutant pre-mRNAs at aberrant 5' splice sites and facilitate processing of dead-end lariat intermediates to mRNA. Activation of aberrant 5' cleavage sites involves base pairing between U5 snRNA and nucleotides upstream of the cleavage site. Processing of dead-end lariat intermediates to mRNA correlates with base pairing between U5 and the first two bases in exon 2. The loop sequence in U5 snRNA may therefore by intimately involved in the transesterification reactions at 5' and 3' splice sites. This pattern of interactions is strikingly reminiscent of exon recognition events in group II self-splicing introns and is consistent with the notion that U5 snRNA may be related to a specific functional domain from a group II-like self-splicing ancestral intron.