The 3' splice site of pre-messenger RNA is recognized by a small nuclear ribonucleoprotein

A compensatory base change in human U2 snRNA can suppress a branch site mutation

We have developed an assay to test whether U2 snRNA can base-pair with the branch site during mammalian mRNA splicing. The beta 110 point mutation (GG----AG) within the first intron of human beta-globin generates a new 3' splice site that is preferentially used. We show here that use of the normal 3' splice site can be restored either by improving the match of a cryptic branch site to the branch site consensus or by introducing mutant U2 snRNAs with greater complementarity to the cryptic branch site. These data indicate that human U2 snRNA can form base pairs with the mRNA precursor; however, base pairing appears to be optional because some mammalian branch sites do not match the consensus.

Hydroxyl radical "footprinting" of RNA: application to pre-mRNA splicing complexes

We present an adaptation of the hydroxyl radical DNA "footprinting" technique that permits high-resolution mapping of protected regions of RNA. Hydroxyl radical cleaves RNA independently of base sequence and secondary structure of the RNAs examined and allows resolution of protected regions at the single nucleotide level. By using this technique, we show that several regions of the 3' splice site of mRNA precursors are protected during the formation of splicing-specific ribonucleoprotein complexes in an in vitro splicing system. These regions include the 3' intron/exon junction and a portion of the adjacent exon, the polypyrimidine tract, and the site of branch formation. These protections appear to be due to splicing specific complexes since their formation is sensitive to point mutations at crucial residues and requires ATP and incubation. The formation of these protected regions is independent of the presence of a 5' splice site.

Mammalian pre-mRNA branch site selection by U2 snRNP involves base pairing

SV40 early pre-mRNA is alternatively spliced to produce large T and small t mRNAs by use of different 5'-splice sites and a shared 3'-splice site. The large T splicing pathway uses multiple lariat branch sites, whereas small t splicing, constrained by its small intron size, can use only one. We exploited this situation to test the hypothesis that RNA-RNA base pairing between U2 snRNA and the branch site sequence is important in mammalian pre-mRNA splicing by constructing and analyzing several mutations in the small t pre-mRNA branch site (UUCUAAU). All of the mutations resulted in substantial decreases in small t splicing relative to large T. To test whether these effects resulted from decreased base pairing with U2 snRNA, compensatory mutations were introduced at the appropriate positions (nucleotides 34-36) in a cloned human U2 gene. All branch site mutations tested (four separate single base substitutions and two triple mutations) were suppressed (i.e., small t splicing was increased) by the appropriate U2 mutations. These results establish that recognition of the poorly conserved mammalian pre-mRNA branch site sequence by U2 snRNP can involve base-pairing.

U3 small nuclear RNA can be psoralen-cross-linked in vivo to the 5' external transcribed spacer of pre-ribosomal-RNA

U3 small nuclear RNA is hydrogen-bonded to high molecular weight nucleolar RNA and can be isolated from greater than 60S pre-ribosomal ribonucleoprotein particles, suggesting that it is involved in processing of ribosomal RNA precursors (pre-rRNA) or in ribosome biogenesis. Here we have used in vivo psoralen cross-linking to identify the region of pre-rRNA interacting with U3 RNA. Quantitative hybridization selection/depletion experiments with clones of rRNA-encoding DNA (rDNA) and cross-linked nuclear RNA showed that all of the cross-linked U3 RNA was associated with a region that includes the external transcribed spacer (ETS) at the 5' end of the human rRNA precursor. To further identify the site of interaction within the approximately 3.7-kilobase ETS, Southern blots of rDNA clones were sandwich-hybridized with cross-linked RNA and then probed for cross-linked U3 RNA. These experiments showed that U3 RNA was cross-linked to a 258-base sequence between nucleotides +438 and +695, just downstream of the ETS early cleavage site (+414). The localization of U3 to this region of the rRNA precursor was not expected from previous models for a base-paired U3-rRNA interaction and suggests that U3 plays a role in the initial pre-rRNA processing event.

PRP4 (RNA4) from Saccharomyces cerevisiae: its gene product is associated with the U4/U6 small nuclear ribonucleoprotein particle

The PRP4 (RNA4) gene product is involved in nuclear mRNA processing in yeast cells; we have previously cloned the gene by complementation of a temperature-sensitive mutation. Sequence and transcript analyses of the cloned gene predicted the gene product to be a 52-kilodalton protein, which was confirmed with antibodies raised against the PRP4 gene product. These antibodies inhibited precursor mRNA splicing in vitro, demonstrating a direct role of PRP4 in splicing. Immunoprecipitations with the antibodies indicated that the PRP4 protein is associated with the U4/U6 small nuclear ribonucleoprotein particle.

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

We have investigated the roles of U1 and U2 snRNP particles in SV40 pre-mRNA splicing by oligonucleotide-targeted degradation of U1 or U2 snRNAs in Xenopus laevis oocytes. Microinjection of oligonucleotides complementary to regions of U1 or U2 RNAs either in the presence or absence of SV40 DNA resulted in specific cleavage of the corresponding snRNA. Unexpectedly, degradation of U1 or U2 snRNA was far more extensive when the oligonucleotide was injected without, or prior to, introduction of viral DNA. In either co-injected or pre-injected oocytes, these oligonucleotides caused a dramatic reduction in the accumulation of spliced SV40 mRNA expressed from the viral late region, and a commensurate increase in unspliced late RNA. When pre-injected, two different U2 specific oligonucleotides also inhibited the formation of both large and small tumor antigen spliced early mRNAs. However, even when, by pre-injection of a U1 5' end-specific oligonucleotide, greater than 95% degradation of the U1 snRNA 5' ends occurred in oocytes, no reduction in early pre-mRNA splicing was observed. In contrast, the same U1 5' end oligonucleotide, when added to HeLa splicing extracts, substantially inhibited the splicing of SV40 early pre-mRNA, indicating that U1 mRNP is not totally dispensable for early splicing. These findings confirm and extend our earlier observations which suggested that different pre-mRNAs vary in their requirements for snRNPs.

The AU-rich sequences present in the introns of plant nuclear pre-mRNAs are required for splicing

Plant cells do not in general process the introns of transcripts expressed from introduced vertebrate genes. By studying the processing of model introns in transfected plant protoplasts, we have investigated the special requirements for intron recognition by plant cells. Our results indicate that the requirements for intron recognition in plants are different from those of both metazoa and yeast. A synthetic intron of arbitrary sequence but incorporating splice site consensus sequences and a high proportion of U and A nucleotides, a characteristic feature of plant introns, was efficiently spliced in protoplasts. We have studied the effects of various sequence alterations and conclude that AU-rich sequences are necessary for intron recognition. In addition, we find that the criteria for branch site selection are relaxed, as they are in vertebrates, but a polypyrimidine tract is not necessary.

20S small nuclear ribonucleoprotein U5 shows a surprisingly complex protein composition

U5 small nuclear ribonucleoprotein (snRNP), purified from HeLa nuclear extracts (splicing extracts), shows a complex protein composition. In addition to the snRNP proteins B', B, D, D', E, F, and G, which are present in each of the major snRNPs U1, U2, U4/U6, and U5, U5 snRNP contains a number of unique proteins characterized by apparent molecular masses of 40, 52, 100, 102, 116, and 200 (mostly a double band) kDa. The latter set of proteins may be regarded as U5-specific for the following reasons. They are not only eluted specifically, together with snRNP particles, from anti-2,2,7-trimethylguanosine immunoaffinity columns by 7-methylguanosine, they also cofractionate with U5 snRNP during chromatography and, most importantly, in glycerol gradient centrifugation. These U5 snRNP particles show a high sedimentation constant of about 20S. U5 snRNPs that lack the U5-specific proteins are also found in nuclear extracts but have (in comparison) a lower sedimentation value of only 8-10S. Autoimmune sera from patients with systemic lupus erythematosus were identified that, on immunoblots with purified U5 snRNP proteins, reacted selectively with the 100- or 200-kDa proteins. This indicates that at least the high molecular mass U5-specific proteins are structurally distinct and not derived one from the other by proteolytic degradation. The existence of so many unique proteins in the U5 snRNP suggests that this snRNP particle may exert its function during splicing mainly by virtue of its protein components.

Probing the structure and function of U2 snRNP with antisense oligonucleotides made of 2'-OMe RNA

We have used oligonucleotides made of 2'-OMe RNA to analyze the role of separate domains of U2 snRNA in the splicing process. We show that antisense 2'-OMe RNA oligonucleotides bind efficiently and specifically to U2 snRNP and demonstrate that masking of two separate regions of U2 snRNA can inhibit splicing by affecting different steps in the spliceosome assembly pathway. Masking the 5' terminus of U2 snRNA does not prevent U2 snRNP binding to pre-mRNA but blocks subsequent assembly of a functional spliceosome. By contrast, masking of U2 sequences complementary to the pre-mRNA branch site completely inhibits binding of pre-mRNA. Hybrid formation at the branch site complementary region also triggers a specific change which affects the 5' terminus of U2 snRNA.

Purified U5 small nuclear ribonucleoprotein can relieve the inhibition of spliceosome assembly and splicing by snRNP-free nuclear proteins

As demonstrated by RNase T1 protection assays at 0 degrees C without ATP, U1 and U5 snRNPs purified by isopycnic centrifugation in cesium chloride bind to the 5' and 3' splice sites of human beta-globin pre-mRNA, respectively. We also devised a saturation-complementation assay and have found that this purified U5 snRNP, unlike U1, successfully competes with snRNP-free fractions of nuclear proteins which inhibit spliceosome assembly and splicing. Restoration of activity requires intact U5 snRNA and correlates with the presence of the 100 Kd intron binding protein (IBP) which we have previously characterized (Tazi et al., 1986, Cell 47, 755-766). Our results are compatible with a model in which the recognition of the 3' splice site by IBP-U5 snRNP is one of the earliest events of the spliceosome assembly. It could organize the structure of the 3' splice site region of the human beta-globin like pre-mRNAs. However, on the basis of results showing that beta-globin and major late adenovirus seem to have different requirements with respect to IBP-U5 snRNP, it appears that some pre-mRNAs could have a native structure that necessitates less if at all IBP-U5.

Pattern analysis of RNA secondary structure similarity and consensus of minimal-energy folding

We describe an automated procedure to search for consensus structures or substructures in a set of homologous or related RNA molecules. The procedure is based on the calculation of optimal and sub-optimal secondary structures using thermodynamic rules for base-pairing by energy-minimization. A linear representation of the secondary structures of the related RNAs is used so that they can be compared and classified using standard alignment and clusterings programs. We illustrate the method by means of two sets of homologous small RNAs, U2 and U3, and a set of alpha-globin mRNAs and show that biologically interesting consensus structures are obtained.

Genetic analysis of the role of human U1 snRNA in mRNA splicing: I. Effect of mutations in the highly conserved stem-loop I of U1

The 5' splice site mutation known as hr440 can be suppressed efficiently in vivo by a compensatory base change in U1 small nuclear RNA (snRNA). We have now begun a second-site reversion analysis of this suppressor U1-4u snRNA (which has a C----U change at position 4) to identify U1 nucleotides that are essential for mRNA splicing. Point mutations in U1-4u that disrupt the structure of stem-loop I or alter phylogenetically conserved nucleotides within the loop cause loss of suppression. The level of suppressor activity observed for most mutants correlated with the abundance of the corresponding suppressor RNA, suggesting that mutations in stem-loop I cause loss of suppression by destabilizing U1 snRNA or the U1 snRNP (small nuclear ribonucleoprotein particle). We favor the interpretation that incompletely or improperly assembled U1 snRNPs are unstable, because two severe point mutations in stem-loop I were found to decrease the binding of U1 snRNP-specific proteins in vitro. In a separate set of experiments, we found that increasing the distance between stem-loop I and the 5' end of U1 snRNA also inhibited suppression but did not affect assembly or stability of the U1 snRNP. This suggests that the relationship between the 5' splice site and the body of the U1 snRNP is important for mRNA splicing.

A block in mammalian splicing occurring after formation of large complexes containing U1, U2, U4, U5, and U6 small nuclear ribonucleoproteins

The assembly of mammalian pre-mRNAs into large 50S to 60S complexes, or spliceosomes, containing small nuclear ribonucleoproteins (snRNPs) leads to the production of splicing intermediates, 5' exon and lariat-3' exon, and the subsequent production of spliced products. Influenza virus NS1 mRNA, which encodes a virus-specific protein, is spliced in infected cells to form another viral mRNA (the NS2 mRNA), such that the ratio of unspliced to spliced mRNA is 10 to 1. NS1 mRNA was not detectably spliced in vitro with nuclear extracts from uninfected HeLa cells. Surprisingly, despite the almost total absence of splicing intermediates in the in vitro reaction, NS1 mRNA very efficiently formed ATP-dependent 55S complexes. The formation of 55S complexes with NS1 mRNA was compared with that obtained with an adenovirus pre-mRNA (pKT1 transcript) by using partially purified splicing fractions that restricted the splicing of the pKT1 transcript to the production of splicing intermediates. At RNA precursor levels that were considerably below saturation, approximately 10-fold more of the input NS1 mRNA than of the input pKT1 transcript formed 55S complexes at all time points examined. The pKT1 55S complexes contained splicing intermediates, whereas the NS1 55S complexes contained only precursor NS1 mRNA. Biotin-avidin affinity chromatography showed that the 55S complexes formed with either NS1 mRNA or the pKT1 transcript contained the U1, U2, U4, U5, and U6 snRNPs. Consequently, the formation of 55S complexes containing these five snRNPs was not sufficient for the catalysis of the first step of splicing, indicating that some additional step(s) needs to occur subsequent to this binding. These results indicate that the 5' splice site, 3' and branch point of NS1 and mRNA were capable of interacting with the five snRNPs to form 55S complexes, but apparently some other sequence element(s) in NS1 mRNA blocked the resolution of the 55S complexes that leads to the catalysis of splicing. On the basis of our results, we suggest mechanisms by which the splicing of NS1 is controlled in infected cells.

Factor interactions with the simian virus 40 early pre-mRNA influence branch site selection and alternative splicing

To study the interaction of splicing factors with the simian virus 40 early-region pre-RNA, which can be alternatively spliced to produce large T and small t mRNAs, we used an in vitro RNase protection assay that defines the 5' boundaries of factor-RNA interactions. Protection products reflecting factor interactions with the large T and small t 5' splice sites and with the multiple lariat branch site region were characterized. All protection products were detected very early in the splicing reaction, before the appearance of spliced RNAs. However, protection of the large T 5' splice site was detected well before small t 5' splice site and branch site protection products, which appeared simultaneously. Oligonucleotide-targeted degradation of small nuclear RNAs (snRNAs) revealed that protection of the branch site region, which occurred at multiple sites, required intact U2 snRNA and was enhanced by U1 snRNA, while protection of the large T and small t 5' splice sites required both U1 and U2 snRNAs. Analysis of several pre-RNAs containing mutations in the branch site region suggests that factor interactions involving the multiple copies of the branch site consensus determine the selection of branch points, which is an important factor in the selection of alternative splicing pathways.

In vitro reconstitution of snRNPs: a reconstituted U4/U6 snRNP participates in splicing complex formation

We have reconstituted in vitro the four snRNPs known to be involved in pre-mRNA splicing: U1, U2, U5, and U4/6. Reconstitution involves adding either authentic or in vitro-synthesized snRNAs to extracts enriched in snRNP structural polypeptides. The reconstituted snRNPs have the same buoyant density and are immunoprecipitated by the same antibodies as authentic snRNPs. Thus, the polypeptide composition of reconstituted snRNPs is similar, if not identical, to that of authentic snRNPs. We show further that a reconstituted U4/U6 particle is fully functional in forming splicing complexes with pre-mRNA. As is the case for the authentic U4/U6 snRNP, the reconstituted U4 snRNP, but not the U6 snRNA, dissociates from the complex prior to formation of the mature spliceosome. The ability to reconstitute snRNPs and assay their activity in spliceosome formation should provide a powerful approach to study these particles.

UACUAAC is the preferred branch site for mammalian mRNA splicing

The conserved branch-site sequence UAC-UAAC is known to form base pairs with the complementary sequence GUAGUA in U2 small nuclear RNA (snRNA) during mRNA splicing in the yeast Saccharomyces cerevisiae. Although the GUAGUA element is conserved in mammalian U2 snRNA, mammalian branch sites conform only weakly to a YURAC consensus and can even be deleted without obvious effects on the efficiency of splicing in vivo. To understand why the GUAGUA element of U2 is conserved in evolution but the branch site is not, we have devised two different competitive assays for branch-site selection using the first intron of the human beta-globin gene. We find that a sequence resembling UACUAAC is the most efficient branch site for mammalian mRNA splicing both in vivo and in vitro. Our results suggest that in mammals U2 snRNA can form base pairs with the branch site and the interaction between U2 and the branch site can be augmented or replaced by an interaction between the spliceosome and some other element of the intron or exons, perhaps the conserved polypyrimidine tract located immediately upstream from the 3' splice site.

A general model for the evolution of nuclear pre-mRNA introns

We present an overview of the evolution of eukaryotic split gene structure and pre-mRNA splicing mechanisms. We have drawn together several seemingly conflicting ideas and we show that they can all be incorporated in a single unified theory of intron evolution. The resulting model is consistent with the notion that introns, as a class, are very ancient, having originated in the "RNA world"; it also supports the concept that introns may have played a crucial role in the construction of many eukaryotic genes and it accommodates the idea that introns are related to mobile insertion elements. Our conclusion is that introns could have a profound effect on the course of eukaryotic gene evolution, but that the origin and maintenance of intron sequences depends, largely, on natural selection acting on the intron sequences themselves.

Autonomous splicing and complementation of in vivo-assembled spliceosomes

We have used an in vivo system generating assayable amounts of a specific pre-mRNA to study the relationship between splicing and an operationally defined nuclear matrix preparation (NM). When NM is prepared by extraction of DNase I-treated nuclei with an approximately physiological concentration of KCl (0.1 M), a portion of NM-associated precursor can be spliced in vitro in the presence of ATP and Mg2+ and in the absence of splicing extract ("autonomous splicing"). We propose that the autonomous reaction, which does not exhibit a temporal lag and is half-complete in 5 min, occurs in fully assembled, matrix-bound ribonucleoprotein complexes (in vivo spliceosomes). Extraction of the NM with concentrations of KCl greater than 0.4 M eliminates autonomous splicing but leaves behind preassembled complexes that can be complemented for splicing with HeLa cell nuclear extract. The splicing complementing factor, representing one or more activities present in the nuclear extract and also in the cytoplasmic S100 fraction, is relatively heat resistant, devoid of an RNA component, and does not bind to DEAE-Sepharose in 0.1 M KCl. It exists in the nucleus in two forms; bound to autonomous spliceosomes and free in the nucleoplasm. Biochemical features of the complementation reaction, and conditions for reversible uncoupling of the two splicing steps are described and discussed.

Characterization of an SNR gene locus in Saccharomyces cerevisiae that specifies both dispensible and essential small nuclear RNAs

A genetic locus is described that specifies two Saccharomyces cerevisiae small nuclear RNAs (snRNAs). The genes specifying the two snRNAs are separated by only 67 base pairs and are transcribed in the same direction. The product RNAs contain 128 and 190 nucleotides and are designated snR128 and snR190, respectively. These RNAs resemble snRNAs of other eucaryotes in nuclear localization and possession of a 5' trimethylguanosine cap. Neither snRNA is related in sequence to previously described vertebrate or yeast snRNAs. Both RNAs exhibit properties consistent with nucleolar organization and hydrogen bonding to pre-rRNA species, suggesting possible roles in ribosome biogenesis. The snR128 species cosediments with deproteinized 27S pre-rRNA, whereas snR190 is associated with a 20S intermediate. Gene disruption in vitro followed by replacement of the chromosomal alleles reveals that SNR128 is essential, whereas SNR190 is not.

Competition between splicing and polyadenylation reactions determines which adenovirus region E3 mRNAs are synthesized

Complex transcription units encode multiple mRNAs which arise by alternative processing of a common pre-mRNA precursor. It is not known how the pre-mRNA processing pathways are determined or controlled. We are investigating this problem by using the E3 complex transcription unit of adenovirus as a model. Our approach is to construct virus mutants with lesions in E3 and then determine how the mutation affects the accumulation of E3 mRNAs in vivo. We report results which indicate that competition between splicing reactions and polyadenylation reactions occurs in vivo and that this plays an important role in alternative pre-mRNA processing.

A 5' splice-region G----C mutation in exon 1 of the human beta-globin gene inhibits pre-mRNA splicing: a mechanism for beta+-thalassemia

We have characterized a Mediterranean beta-thalassemia allele containing a sequence change at codon 30 that alters both beta-globin pre-mRNA splicing and the structure of the hemoglobin product. Presumably, this G----C transversion at position -1 of intron 1 reduces severely the utilization of the normal 5' splice site since the level of the Arg----Thr mutant hemoglobin (designated hemoglobin Kairouan) found in the erythrocytes of the patient is very low (2% of total hemoglobin). Since no natural mutations of the guanine located at position -1 of the CAG/GTAAGT consensus sequence had been isolated previously, we investigated the role of this nucleotide in the constitution of an active 5' splice site by studying the splicing of the pre-mRNA in cell-free extracts. We demonstrate that correct splicing of the mutant pre-mRNA is 98% inhibited. Our results provide further insights into the mechanisms of pre-mRNA maturation by revealing that the last residue of the exon plays a role at least equivalent to that of the intron residue at position +5.

The effect of estramustine, nor-nitrogen mustard and tauromustine on macromolecular labelling in the human prostatic tumour cell line 1013L

To further clarify the mode of action of estramustine, the influence on macromolecular synthesis in the human prostatic tumour cell line 1013L was investigated. Cell treatment with estramustine, nor-nitrogen mustard and tauromustine, followed by radioactive nucleotide and leucine incorporations, as a measure of RNA, DNA and protein labelling, were carried out. The initial effect of estramustine clearly differed from that obtained after treatment with nor-nitrogen mustard and tauromustine. No inhibition of DNA synthesis was found whereas an inhibition of overall RNA synthesis was predominant. Adaption of an established RNA separation method was used in an indepth study of RNA labelling after estramustine treatment. An inhibition of 29S, 18S and 4-7S RNA was found after estramustine treatment, indicating disturbances in either RNA processing or RNA transport. The lack of 45S RNA labelling additionally indicates pre-ribosomal inhibition.

A block in mammalian splicing occurring after formation of large complexes containing U1, U2, U4, U5, and U6 small nuclear ribonucleoproteins

The assembly of mammalian pre-mRNAs into large 50S to 60S complexes, or spliceosomes, containing small nuclear ribonucleoproteins (snRNPs) leads to the production of splicing intermediates, 5' exon and lariat-3' exon, and the subsequent production of spliced products. Influenza virus NS1 mRNA, which encodes a virus-specific protein, is spliced in infected cells to form another viral mRNA (the NS2 mRNA), such that the ratio of unspliced to spliced mRNA is 10 to 1. NS1 mRNA was not detectably spliced in vitro with nuclear extracts from uninfected HeLa cells. Surprisingly, despite the almost total absence of splicing intermediates in the in vitro reaction, NS1 mRNA very efficiently formed ATP-dependent 55S complexes. The formation of 55S complexes with NS1 mRNA was compared with that obtained with an adenovirus pre-mRNA (pKT1 transcript) by using partially purified splicing fractions that restricted the splicing of the pKT1 transcript to the production of splicing intermediates. At RNA precursor levels that were considerably below saturation, approximately 10-fold more of the input NS1 mRNA than of the input pKT1 transcript formed 55S complexes at all time points examined. The pKT1 55S complexes contained splicing intermediates, whereas the NS1 55S complexes contained only precursor NS1 mRNA. Biotin-avidin affinity chromatography showed that the 55S complexes formed with either NS1 mRNA or the pKT1 transcript contained the U1, U2, U4, U5, and U6 snRNPs. Consequently, the formation of 55S complexes containing these five snRNPs was not sufficient for the catalysis of the first step of splicing, indicating that some additional step(s) needs to occur subsequent to this binding. These results indicate that the 5' splice site, 3' and branch point of NS1 and mRNA were capable of interacting with the five snRNPs to form 55S complexes, but apparently some other sequence element(s) in NS1 mRNA blocked the resolution of the 55S complexes that leads to the catalysis of splicing. On the basis of our results, we suggest mechanisms by which the splicing of NS1 is controlled in infected cells.

Additional low-abundance human small nuclear ribonucleoproteins: U11, U12, etc

Two-dimensional gel fractionation has revealed the existence of a number (greater than or equal to 8) of additional species of HeLa cell small RNAs that have 5' trimethylguanosine cap structures and are bound by proteins containing Sm epitopes. Therefore, these low-abundance (10(3)-10(4) per cell) RNAs belong to the Sm class of small nuclear ribonucleoproteins (snRNPs), whose best-known members are the four highly abundant (approximately 10(6) per cell) particles required for pre-mRNA splicing. The complexity of Sm snRNPs in mammalian cells is thus not greatly different from that previously established for lower eukaryotes. Two of the new RNAs, designated U11 (131 nucleotides) and U12 (150 nucleotides), have been sequenced. The U11 and U12 snRNPs have been characterized further by examining their nuclease sensitivity and their possible interactions with other snRNPs. Potential roles for the low-abundance snRNPs in aspects of pre-mRNA processing are discussed.

An in vitro interaction between the human U3 snRNP and 28S rRNA sequences near the alpha-sarcin site

Model transcripts containing mammalian pre-rRNA sequences were incubated with a HeLa cell extract, digested with T1 RNase, and immunoprecipitated with anti-(U3)RNP or control antibodies. Two overlapping fragments derived from the 3' domain of human 28S rRNA were specifically immunoprecipitated although transcripts which spanned the transcription initiation site, the ETS processing site, the 5' end of 18S, and both termini of 5.8S yielded no protected fragments. The sequence of these fragments was determined using a novel technique in which the [32P]-labeled fragment was co-finger-printed with [3H]-labeled total transcript serving as an internal marker. The fragments immunoprecipitated derive from nucleotides 4570-4590 and 4575-4590 of human 28S and are adjacent to the alpha-sarcin site. Protection most likely involves the U3 RNA since it is sensitive to pretreatment of the extract with micrococcal nuclease. Complementarity between U3 and this rRNA region is phylogenetically conserved in species ranging from human to S. cerevisiae. The possible significance of this finding is discussed.

Categorical discriminant analysis of 3'-splice site signals of mRNA precursors in higher eukaryote genes

With regard to the signals that direct excision of introns from mRNA precursors in higher eukaryote genes, a consensus sequence, (sequence; see text); has been proposed for the 3'-splice site, but actual 3'-splice site sequences differ from it to a greater or lesser degree. In the present paper, nucleotide sequences were transformed into categorical data, and quantification analysis (class II), as proposed by Hayashi, was applied to the system. Categorical weights given to variables related to position and the species of nucleotide were estimated so that the two classes of 3'-splice site sequences and sequences other than 3'-splice site might be discriminated most distinctly. The 3'-splice site signals were then characterized in terms of these categorical weight values. We also calculated partial correlation coefficient values, which explain the relative importance of each position in the 3'-splice site signal sequence.

Alternative splicing of SV40 early pre-mRNA is determined by branch site selection

Splicing of SV40 early pre-mRNA to alternative large-T and small-t mRNAs involves the utilization of multiple lariat branch sites. To determine the functional significance of these sites, we constructed and analyzed a set of base substitution mutants in which the major branch acceptors were altered, either singly or in combination. The ratio of large-T to small-t mRNAs produced in vivo was found to vary by over 100-fold between different mutants. When splicing was assayed in vitro with a standard pre-RNA, which results in splicing almost exclusively to large-T mRNA, the patterns of branch site utilization were altered dramatically, although the mutations were essentially without effect on splicing efficiency. However, use of a 5' truncated pre-RNA, which results in a splicing pattern that reflects the in vivo alternative splicing potential of this pre-RNA, revealed a strong correlation between the effects of the base substitutions on branch site selection and alternative splice-site utilization. An RNase protection analysis of factor interactions with the 5' splice sites and branch sites in wild-type and mutant pre-RNAs suggests that a competition for different branch sites plays a crucial role in the assembly of 'alternative' spliceosomes, thereby controlling alternative splice-site selection.

Pre-mRNA splicing by complementation with purified human U1, U2, U4/U6 and U5 snRNPs

The four major nucleoplasmic small nuclear ribonucleoprotein particles U1, U2, U4/U6 and U5 can be extensively purified from HeLa cells by immunoaffinity chromatography using a monoclonal anti-trimethylguanosine antibody. The snRNP particles in active splicing extracts are selectively bound to the immunoaffinity matrix, and are then gently eluted by competition with an excess of free nucleoside. Biochemical complementation studies show that the purified snRNPs are active in pre-mRNA splicing, but only in the presence of additional non-snRNP protein factors. All the RNPs that are necessary for splicing can be purified in this manner. The active snRNPs are characterized with respect to their polypeptide composition, and shown to be distinct from several other activities implicated in splicing.

5' splice site selection in yeast: genetic alterations in base-pairing with U1 reveal additional requirements

Using a strategy of compensatory nucleotide changes between yeast U1 and a 5' splice site, we have analyzed the contribution of base-pairing to the efficiency and fidelity of pre-mRNA splicing in vivo. Watson-Crick base-pairing interactions with U1 can be demonstrated at intron positions 1 and 5 but not at position 4. Moreover, restoration of the ability to pair with U1 is not sufficient to restore activity in the second step of splicing to intron position 1 mutants. Finally, in contrast to recent observations in mammalian systems, we find that the precise position of 5' splice site cleavage is not determined solely by the base-pairing interaction with U1. Rather, the presence of a G residue at position 5 is required for the correct localization of the nucleolytic event. Taken together, these results indicate that the demands for 5' splice site selection and utilization are more complex than a simple maximization of Watson-Crick interactions with U1.

3' cleavage and polyadenylation of mRNA precursors in vitro requires a poly(A) polymerase, a cleavage factor, and a snRNP

We have separated and purified three factors from HeLa cell nuclear extracts that together can accurately cleave and polyadenylate pre-mRNAs containing the adenovirus L3 polyadenylation site. One of the factors is a poly(A) polymerase with a molecular weight of approximately 50-60 kd. The second activity is a cleavage factor with a native molecular weight in the range of 70-120 kd. The third component is a factor (cleavage and polyadenylation factor, CPF) that is needed for the cleavage reaction and, in addition, confers specificity to the poly(A) polymerase activity; the native molecular weight of CPF is approximately 200 kd. Poly(A) polymerase together with CPF is sufficient to specifically polyadenylate pre-mRNA substrates that have been precleaved at the poly(A) addition site. In contrast, all three components are required for accurate cleavage and polyadenylation of pre-mRNA substrates. Further purification of CPF by buoyant density centrifugation, ion exchange, and affinity column chromatography or by gel filtration demonstrates that CPF activity resides in a ribonucleoprotein and copurifies with U11 snRNP.

Presplicing complex formation requires two proteins and U2 snRNP

Six fractions derived from a HeLa cell nuclear extract are necessary for the generation of spliced mRNA in vitro. To establish a function for the protein factors present in these fractions, their role in the formation of splicing complexes was analyzed by electrophoresis in native polyacrylamide gels. Two of the fractions are sufficient to assemble the adenovirus major late mRNA precursor into a presplicing complex with characteristics similar to the presplicing complex assembled in nuclear extract. One fraction supplies splicing factor (SF) 1 and at least one small nuclear ribonucleoprotein particle, U2 snRNP. The other fraction contains SF3. Extensive fractionation of this protein has revealed that it is essential for presplicing complex assembly and the splicing reaction.

Four novel U RNAs are encoded by a herpesvirus

Marmoset T lymphocytes transformed by herpesvirus saimiri contain the first virally encoded U RNAs (called HSURs) to be identified. HSURs assemble into small nuclear ribonucleoproteins of low abundance (less than or equal to 2 x 10(4) copies/cell). They bind proteins with Sm determinants and acquire a 5' trimethylguanosine cap structure. The sequences of HSUR 1 (143 nucleotides), HSUR 2 (115 nucleotides), HSUR 3 (76 nucleotides), and HSUR 4 (106 nucleotides) are related to each other but are distinct from any previously characterized cellular U RNA. The viral genes encoding the HSURs possess conserved enhancer, promoter, and 3' end formation signals unique to U RNA genes. HSUR 1 and HSUR 2 have a similar 5' end sequence that exhibits perfect complementarity to the highly conserved AAUAAA polyadenylation signal. Oligonucleotide directed RNAase H degradation indicates that this 5' end region is available for base pairing interactions within the HSUR 1 and HSUR 2 snRNP particles.

The C. elegans trans-spliced leader RNA is bound to Sm and has a trimethylguanosine cap

mRNA splicing in C. elegans is unusual: most introns are very short (approximately 50 bases), and many mRNAs receive a leader by trans-splicing. The donor in trans-splicing is a 94 nucleotide molecule, termed the leader RNA, that contributes its 5' 22 nucleotides to a variety of mRNAs. We show here that C. elegans has the usual snRNAs, which presumably catalyze the splicing reactions. As expected, they are bound to the Sm antigen and have 2,2,7-methylguanosine caps. Remarkably, the trans-spliced leader RNA is also Sm-associated and has this special cap. Hence, a molecule discovered as a substate of splicing has properties of molecules heretofore known only to facilitate splicing of other RNAs. Mature mRNAs that have received the leader evidently lack 2,2,7-methylguanosine caps, suggesting that these caps are removed or altered during processing.

Competition between splicing and polyadenylation reactions determines which adenovirus region E3 mRNAs are synthesized

Complex transcription units encode multiple mRNAs which arise by alternative processing of a common pre-mRNA precursor. It is not known how the pre-mRNA processing pathways are determined or controlled. We are investigating this problem by using the E3 complex transcription unit of adenovirus as a model. Our approach is to construct virus mutants with lesions in E3 and then determine how the mutation affects the accumulation of E3 mRNAs in vivo. We report results which indicate that competition between splicing reactions and polyadenylation reactions occurs in vivo and that this plays an important role in alternative pre-mRNA processing.

Characterization of an SNR gene locus in Saccharomyces cerevisiae that specifies both dispensible and essential small nuclear RNAs

A genetic locus is described that specifies two Saccharomyces cerevisiae small nuclear RNAs (snRNAs). The genes specifying the two snRNAs are separated by only 67 base pairs and are transcribed in the same direction. The product RNAs contain 128 and 190 nucleotides and are designated snR128 and snR190, respectively. These RNAs resemble snRNAs of other eucaryotes in nuclear localization and possession of a 5' trimethylguanosine cap. Neither snRNA is related in sequence to previously described vertebrate or yeast snRNAs. Both RNAs exhibit properties consistent with nucleolar organization and hydrogen bonding to pre-rRNA species, suggesting possible roles in ribosome biogenesis. The snR128 species cosediments with deproteinized 27S pre-rRNA, whereas snR190 is associated with a 20S intermediate. Gene disruption in vitro followed by replacement of the chromosomal alleles reveals that SNR128 is essential, whereas SNR190 is not.

Unusual branch point selection in processing of human growth hormone pre-mRNA

Intron A of the human growth hormone gene does not contain an A residue within 56 nucleotides preceding the 3' splice site. The analysis of the excised intron lariat revealed a C residue 28 nucleotides upstream from the 3' splice site as the major branch acceptor nucleotide. Two additional minor branched nucleotides were identified as U residues at positions -22 and -36. An adenosine substitution at position -22 results in lariat formation solely to this nucleotide. Therefore, C and U residues can function efficiently as natural branch acceptors, but an A residue is preferred if available in the proper region. In addition, the data strongly reinforce the importance of the distance constraint for lariat formation. To explain selection of the branch acceptor nucleotide, potential base-pairing interactions of branch point sequences with the U2 RNA are discussed.

Tissue-specific expression and cDNA cloning of small nuclear ribonucleoprotein-associated polypeptide N

Sera from some patients with systemic lupus erythematosus and other autoimmune diseases have antibodies against nuclear antigens. An example is anti-Sm sera, which recognize proteins associated with small nuclear RNA molecules [small nuclear ribonucleoprotein (snRNP) particles]. In this paper anti-Sm sera were used to probe immunoblots of various rat tissues. A previously unidentified Mr 28,000 polypeptide was recognized by these anti-Sm sera. This polypeptide, referred to as "N," is expressed in a tissue-specific manner, being most abundant in rat brain, less so in heart, and undetectable in the other tissues examined. Immunoprecipitation experiments using antibodies directed against the cap structure of small nuclear RNAs have demonstrated that N is a snRNP-associated polypeptide. Anti-Sm serum was also used to isolate a partial cDNA clone (lambda rb91) from a rat brain phage lambda gt11 cDNA expression library. On RNA blots, the 450-base-pair cDNA insert of this clone hybridized to a 1600-nucleotide mRNA species with an identical tissue distribution to N, suggesting that lambda rb91 encodes at least part of N. A longer cDNA clone was obtained by rescreening the library with lambda rb91. In vitro transcription and subsequent translation of this subcloned, longer insert (pGMA2) resulted in a protein product with the same electrophoretic and immunological properties as N, confirming that pGMA2 encodes N. The tissue distribution of N and the involvement of snRNP particles in nuclear pre-mRNA processing may imply a role for N in tissue-specific pre-mRNA splicing.

Sequence and genetic analysis of a dispensible 189 nucleotide snRNA from Saccharomyces cerevisiae

The structure of a Saccharomyces cerevisiae gene that encodes a small nuclear RNA (snRNA) of 189 nucleotides is described. This gene, designated SNR189, is located 400 base pairs upstream of the CRY1 gene on yeast chromosome III. Gene replacement analysis revealed the SNR189 gene to be dispensable for growth under a variety of culture conditions. The snR189 sequence lacks homology with other sequenced yeast or metazoan snRNAs.

Purification and visualization of native spliceosomes

Mammalian spliceosomes were purified in preparative amounts by gel filtration chromatography and shown to be functional by in vitro complementation experiments. The column fractions containing spliceosomes are enriched in the snRNAs U1, U2, U4, U5, and U6 and a subset of proteins present in the nuclear extract. Splicing intermediates, the entire set of snRNAs, and the enriched proteins can be immunoprecipitated with three different monoclonal antibodies that recognize snRNP determinants. At least one U1 snRNP is present in each spliceosome since the particles are quantitatively immunoprecipitated by an anti-U1 snRNP monoclonal antibody. Examination of the spliceosome fractions by EM revealed a relatively homogeneous population of 40-60 nm particles with a striking morphology. Evidence that these particles are spliceosomes is their sensitivity to micrococcal nuclease, their ATP-dependent assembly, and their immunoprecipitation with a trimethyl cap monoclonal antibody. In addition, pre-mRNA was visualized in the particles by EM.

Functional analysis of the sea urchin U7 small nuclear RNA

U7 small nuclear RNA (snRNA) is an essential component of the RNA-processing machinery which generates the 3' end of mature histone mRNA in the sea urchin. The U7 small nuclear ribonucleoprotein particle (snRNP) is classified as a member of the Sm-type U snRNP family by virtue of its recognition by both anti-trimethylguanosine and anti-Sm antibodies. We analyzed the function-structure relationship of the U7 snRNP by mutagenesis experiments. These suggested that the U7 snRNP of the sea urchin is composed of three important domains. The first domain encompasses the 5'-terminal sequences, up to about nucleotides 7, which are accessible to micrococcal nuclease, while the remainder of the RNA is highly protected and hence presumably bound by proteins. This region contains the sequence complementarities between the U7 snRNA and the histone pre-mRNA which have previously been shown to be required for 3' processing (F. Schaufele, G. M. Gilmartin, W. Bannwarth, and M. L. Birnstiel, Nature [London] 323:777-781, 1986). Nucleotides 9 to 20 constitute a second domain which includes sequences for Sm protein binding. The complementarities between the U7 snRNA sequences in this region and the terminal palindrome of the histone mRNA appear to be fortuitous and play only a secondary, if any, role in 3' processing. The third domain is composed of the terminal palindrome of U7 snRNA, the secondary structure of which must be maintained for the U7 snRNP to function, but its sequence can be drastically altered without any observable effect on snRNP assembly or 3' processing.

Intron sequences and the length of the downstream second exon affect the binding of hnRNP C proteins in an in vitro splicing reaction

The proteins that are in direct contact with the pre-mRNA in an in vitro splicing reaction were analyzed by UV cross-linking experiments. Six major proteins (120, 55, 44, 42, 39 and 38 KD) and three minor polypeptides (84, 72 and 63 KD) were detected. The predominant proteins 44, 42 KD belong to the class of hnRNP C proteins since they were immunoprecipitated by monoclonal antibodies directed against hnRNP C proteins. The cross-linked proteins were not detected in the absence of Mg2+, ATP or when RNA lacking introns were used as substrates in the splicing reactions. The effect of exon sequences on the binding efficiency for the photocrosslinked proteins was investigated. Transcripts containing a second exon of 24 nucleotides for the beta-globin or 107 nucleotides for the mouse insulin, yielded a reduced amount of cross-linked proteins when compared with "full length" pre-mRNAs. Sequences within the first exon of the beta-globin pre-mRNA did not affect the binding efficiency of these proteins. The reduced binding efficiency of the cross-linked proteins for the truncated beta-globin or mouse insulin pre-mRNAs correlated with the lower efficiency for in vitro splicing. Substitutions with unrelated sequences in the beta-globin second exon restore the binding of the cross-linked proteins indicating that the length of the second exon and not specific sequences are relevant for the binding efficiency of these proteins. The SP6/mouse insulin oligonucleotides cross-linked to the hnRNP C proteins were isolated and sequenced. A 17-mer was located in the second exon (134 nucleotides downstream from the 3' splice site) and a 14-mer in the intron region (25 nucleotides downstream the 5' splice site). The beta-globin oligonucleotides cross-linked to the hnRNP C proteins were a 13-mer in the second exon (28 nucleotides downstream the 3' splice site) and an 8-mer in the first exon (81 nucleotides downstream the 5' end of the pre-mRNA). Our results indicate that the hnRNP C proteins interact with those oligonucleotides located in different regions of the pre-mRNA. The binding efficiency of those proteins, however, depends on the length of the second exon and the presence of intron sequences (secondary and/or tertiary pre-mRNA structure).

Splice site selection, rate of splicing, and alternative splicing on nascent transcripts

Based on ultrastructural analysis of actively transcribing genes seen in electron micrographs, we present evidence that pre-mRNA splicing occurs with a reasonable frequency on the nascent transcripts of early Drosophila embryo genes and that splice site selection may generally precede polyadenylation. The details of the process observed are in agreement with results from in vitro splicing systems but differ in the more rapid completion of in vivo splicing. For those introns that are removed cotranscriptionally, a series of events is initiated following 3' splice site synthesis, beginning with ribonucleoprotein (RNP) particle formation at the 3' splice site within 48 sec, intron loop formation within 2 min, and splicing within 3 min. The initiation of the process is correlated with 3' splice site synthesis but is independent of 5' splice site synthesis, the position of the intron within the transcript, and the age or length of the transcript. In some cases, introns are removed from the 5' end of a transcript before introns are synthesized at the 3' end, supporting a possible role for the order of transcription in splice site pairing. In general, our observations are consistent with the 'first-come-first-served' principle of splice site selection, although an observed example of exon skipping indicates that alternative splicing possibilities can be accommodated within this general framework.

Gel electrophoretic isolation of splicing complexes containing U1 small nuclear ribonucleoprotein particles

Assembly of splicing precursor RNAs into ribonucleoprotein particle (RNP) complexes during incubation in in vitro splicing extracts was monitored by a new system of RNP gel electrophoresis. The temporal pattern of assembly observed by our system was identical to that obtained by other gel and gradient methodologies. In contrast to the results obtained by other systems, however, we observed requirements of U1 small nuclear RNPs (snRNPs) and 5' splice junction sequences for formation of specific complexes and retention of U1 snRNPs within gel-fractionated complexes. Single-intron substrate RNAs rapidly assembled into slow-migrating complexes. The first specific complex (A) appeared within a minute of incubation and required ATP, 5' and 3' precursor RNA consensus sequences, and intact U1 and U2 RNAs for formation. A second complex (B) containing precursor RNA appeared after 15 min of incubation. Lariat-exon 2 and exon 1 intermediates first appeared in this complex, operationally defining it as the active spliceosome. U4 RNA was required for appearance of complex B. Released lariat first appeared in a complex of intermediate mobility (A') and subsequently in rapidly migrating diffuse complexes. Ligated product RNA was observed only in fast-migrating complexes. U1 snRNPs were detected as components of gel-isolated complexes. Radiolabeled RNA within the A and B complexes was immunoprecipitated by U1-specific antibodies under gel-loading conditions and from gel-isolated complexes. Therefore, the RNP antigen remained associated with assembled complexes during gel electrophoresis. In addition, 5' splice junction sequences within gel-isolated A and B complexes were inaccessible to RNase H cleavage in the presence of a complementary oligonucleotide. Therefore, nuclear factors that bind 5' splice junctions also remained associated with 5' splice junctions under our gel conditions.

The length of the downstream exon and the substitution of specific sequences affect pre-mRNA splicing in vitro

We have shown previously that truncation of the human beta-globin pre-mRNA in the second exon, 14 nucleotides downstream from the 3' splice site, leads to inhibition of splicing but not cleavage at the 5' splice site. We now show that several nonglobin sequences substituted at this site can restore splicing and that the efficiency of splicing depends on the length of the second (downstream) exon and not a specific sequence. Deletions in the first exon have no effect on the efficiency of in vitro splicing. Surprisingly, an intron fragment from the 5' region of the human or rabbit beta-globin intron 2, when placed 14 nucleotides downstream from the 3' splice site, inhibited all the steps in splicing beginning with cleavage at the 5' splice site. This result suggests that the intron 2 fragment carries a "poison" sequence that can inhibit the splicing of an upstream intron.