Complex alternative RNA processing generates an unexpected diversity of poly(A) polymerase isoforms

Multiple forms of poly(A) polymerase (PAPs I, II, and III) cDNA have previously been isolated from bovine, human, and/or frog cDNA libraries. PAPs I and II are long forms of the enzyme that contain four functional domains: an apparent ribonucleoprotein-type RNA-binding domain, a catalytic region that may be related to the polymerase module, two nuclear localization signals (NLSs I and 2), and a C-terminal Ser/Thr-rich region. PAP III would encode a truncated protein that lacks the NLSs and the S/T-rich region. To investigate further the structure and expression of these forms, we isolated the mouse PAP gene and an intronless pseudogene from a mouse liver genomic library. The structure of the gene indicates that different forms of PAP are produced by alternative splicing (PAPs I and II) or by competition between polyadenylation and splicing (PAP III). The pseudogene appears to reflect yet another form of long PAP, which we call PAP IV. Mouse PAP III and two additional truncated forms, PAPs V and VI, which would be produced by use of poly(A) sites in adjacent introns, were also isolated from a mouse brain cDNA library. RNase protection and reverse transcription-PCR analyses showed that PAP II, V, and VI are expressed in all tissues tested but that PAP I and/or IV and III are tissue specific. However, immunoblot analysis detected only the long forms, raising the possibility that the short-form RNAs are not translated. Purified recombinant baculovirus-expressed PAPs were tested in several in vitro assays, and the short forms were found to be inactive. We discuss the possible significance of this complex expression pattern.

Mechanisms of transcriptional activation and repression can both involve TFIID

Regulation of transcription involves the activities of activators and repressors. Recent experiments have provided evidence that the function of both types of regulators can involve interactions with one or more component of the basal transcription machinery. A principal target appears to be TFIID, which consists of the TATA binding protein (TBP) and associated factors (TAFs). Here we describe experiments that provide added support for the idea that interactions affecting TFIID can play important roles in both activation and repression. We show, using transfection assays in Drosophila Schneider cells, that recruitment of TBP to a promoter as a GAL4-TBP fusion protein can provide a substantial activation of transcription. The conserved core of TBP is necessary and sufficient for this effect, which was observed with both TATA-containing and TATA-lacking basal promoters. These findings extend experiments performed in yeast, and strengthen the idea that recruitment of TBP (TFIID) can be an important mechanism of activation. We also provide further support for the idea that TBP can be a target for a transcriptional repressor, the Drosophila Even-skipped protein (Eve). We present evidence that the homeodomain, which is necessary for binding TBP in vitro, can also be required for repression in vivo, independent of its role in DNA binding. On the other hand, deletion of the alanine/proline-rich region that is essential for repression in vivo and TBP binding in vitro does not significantly affect DNA binding by the purified protein. These results strengthen the view that TBP, either directly or indirectly as a component of TFIID, can be a target of both activators and repressors.

Functional interactions between the pelle kinase, Toll receptor, and tube suggest a mechanism for activation of dorsal

A complex signal transduction pathway functions in the early Drosophila embryo to establish dorsal-ventral polarity. Activation of this pathway results in the nuclear transport of the protein dorsal (dl), a member of the rel/NF-kappaB family of transcription factors. Genetic studies have identified three intracellular components whose activity is required for activation of dl: Toll, a transmembrane receptor; pelle (pll), a serine/threonine protein kinase; and tube, a protein of unknown function. Here we examine the activities of these proteins when coexpressed in Drosophila Schneider cells. Coexpression of pll with dl enhanced dl nuclear localization and resulted in a modest increase in transcriptional activity. However, when pll was coexpressed with a specific mutant derivative of Toll (TlNaeI), although not with wild-type Toll, a striking synergistic activation of dl was detected. Unexpectedly, coexpression of pll plus TlNaeI, in the absence of dl, resulted in a similar synergistic activation of a GAL4-tube fusion protein. Based on these and other results, we propose a model in which pll receives a signal from activated Toll and phosphorylates tube, which then participates directly in dl activation.

Interaction between the U1 snRNP-A protein and the 160-kD subunit of cleavage-polyadenylation specificity factor increases polyadenylation efficiency in vitro

We have previously shown that the U1 snRNP-A protein (U1A) interacts with elements in SV40 late polyadenylation signal and that this association increases polyadenylation efficiency. It was postulated that this interaction occurs to facilitate protein-protein association between components of the U1 snRNP and proteins of the polyadenylation complex. We have now used GST fusion protein experiments, coimmunoprecipitations and Far Western blot analyses to demonstrate direct binding between U1A and the 160-kD subunit of cleavage-polyadenylation specificity factor (CPSF). In addition, Western blot analyses of fractions from various stages of CPSF purification indicated that U1A copurified with CPSF to a point but could be separated in the highly purified fractions. These data suggest that U1A protein is not an integral component of CPSF but may be able to interact and affect its activity. In this regard, the addition of purified, recombinant U1A to polyadenylation reactions containing CPSF, poly(A) polymerase, and a precleaved RNA substrate resulted in concentration-dependent increases in both the level of polyadenylation and poly(A) tail length. In agreement with the increase in polyadenylation efficiency caused by U1A, recombinant U1A stabilized the interaction of CPSF with the AAUAAA-containing substrate RNA in electrophoretic mobility shift experiments. These findings suggest that, in addition to its function in splicing, U1A plays a more global role in RNA processing through effects on polyadenylation.

The Clk/Sty protein kinase phosphorylates SR splicing factors and regulates their intranuclear distribution

Mammalian Clk/Sty is the prototype for a family of dual specificity kinases (termed LAMMER kinases) that have been conserved in evolution, but whose physiological substrates are unknown. In a yeast two-hybrid screen, the Clk/Sty kinase specifically interacted with RNA binding proteins, particularly members of the serine/arginine-rich (SR) family of splicing factors. Clk/Sty itself has an serine/arginine-rich non-catalytic N-terminal region which is important for its association with SR splicing factors. In vitro, Clk/Sty efficiently phosphorylated the SR family member ASF/SF2 on serine residues located within its serine/arginine-rich region (the RS domain). Tryptic phosphopeptide mapping demonstrated that the sites on ASF/SF2 phosphorylated in vitro overlap with those phosphorylated in vivo. Immunofluorescence studies showed that a catalytically inactive form of Clk/Sty co-localized with SR proteins in nuclear speckles. Overexpression of the active Clk/Sty kinase caused a redistribution of SR proteins within the nucleus. These results suggest that Clk/Sty kinase directly regulates the activity and compartmentalization of SR splicing factors.

The 160-kD subunit of human cleavage-polyadenylation specificity factor coordinates pre-mRNA 3'-end formation

Cleavage-polyadenylation specificity factor (CPSF) is a multisubunit protein that plays a central role in 3' processing of mammalian pre-mRNAs. CPSF recognizes the AAUAAA signal in the pre-mRNA and interacts with other proteins to facilitate both RNA cleavage and poly(A) synthesis. Here we describe the isolation of cDNAs encoding the largest subunit of CPSF (160K) as well as characterization of the protein product. Antibodies raised against the recombinant protein inhibit polyadenylation in vitro, which can be restored by purified CPSF. Extending previous studies, which suggested that 160K contacts the pre-mRNA, we show that purified recombinant 160K can, by itself, bind preferentially to AAUAAA-containing RNAs. While the sequence of 160K reveals similarities to the RNP1 and RNP2 motifs found in many RNA-binding proteins, no clear match to a known RNA-binding domain was found, and RNA recognition is therefore likely mediated by a highly diverged or novel structure. We also show that 160K binds specifically to both the 77K (suppressor of forked) subunit of the cleavage factor CstF and to poly(A) polymerase (PAP). These results provide explanations for previously observed cooperative interactions between CPSF and CstF, which are responsible for poly(A) site specification, and between CPSF and PAP, which are necessary for synthesis of the poly(A) tail. Also supporting a direct role for 160K in these interactions is the fact that 160K by itself retains partial ability to cooperate with CstF in binding pre-mRNA and, unexpectedly, inhibits PAP activity in in vitro assays. We discuss the significance of these multiple functions and also a possible evolutionary link between yeast and mammalian polyadenylation suggested by the properties and sequence of 160K.

Chromosomal localization of mouse and human genes encoding the splicing factors ASF/SF2 (SFRS1) and SC-35 (SFRS2)

The mammalian SR-type splicing factors ASF/SF2 and SC-35 play crucial roles in pre-mRNA splicing and have been shown to shift splice site choice in vitro. We have mapped the ASF/SF2 gene in mice and humans and the SC-35 gene in mice. Somatic cell hybrid mapping of the human ASF/SF2 gene (SFRS1 locus) reveals that it resides on chromosome 17, and fluorescence in situ hybridization refines this localization to 17q21.3-q22. Recombinant inbred mapping of the mouse ASF/SF2 gene (Sfrs1 locus) and the mouse SC-35 gene (Sfrs2 locus) demonstrates that both genes are located in a part of mouse chromosome 11 that is homologous to human chromosome 17. Mapping of Sfrs1 using F1 hybrid backcross mice between the strains C57BL/6 and DDK places Sfrs1 very near the marker D11Mit38 and indicates that the ASF/SF2 gene is closely linked to the Ovum mutant locus.

The transcriptional repressor even-skipped interacts directly with TATA-binding protein

The Drosophila homeodomain protein Even-skipped (Eve) has previously been shown to function as a sequence-specific transcriptional repressor, and in vitro and in vivo experiments have shown that the protein can actively block basal transcription. However, the mechanism of repression is not known. Here, we present evidence establishing a direct interaction between Eve and the TATA-binding protein (TBP). Using cotransfection assays with minimal basal promoters whose activity can be enhanced by coexpression of TBP, we found that Eve could efficiently block, or squelch, this enhancement. Squelching did not require Eve DNA-binding sites on the reporter plasmids but was dependent on the presence of the Eve repression domain. Further support for an in vivo interaction between the Eve repression domain and TBP was derived from a two-hybrid-type assay with transfected cells. Evidence that Eve and TBP interact directly was provided by in vitro binding assays, which revealed a specific protein-protein interaction that required an intact Eve repression domain and the conserved C terminus of TBP. The Eve homeodomain was also required for these associations, suggesting that it may function in protein-protein interactions. We also show that a previously characterized artificial repression region behaves in a manner similar to that of the Eve repression domain, including its ability to squelch TBP-enhanced expression in vivo and to bind TBP specifically in vitro. Our results suggest a model for transcriptional repression that involves an interaction between Eve and TBP.

U1 snRNP-ASF/SF2 interaction and 5' splice site recognition: characterization of required elements

Members of the SR family of proteins, can collaborate with U1 snRNP in the recognition of 5' splice sites in pre-messenger RNAs. We have previously shown that purified U1 snRNP and ASF/SF2 form a ternary complex with pre-mRNA, which is dependent on a functional 5' splice site. In this manuscript we dissect the requirements for the formation of this complex. Sequences in the pre-mRNA, domains in ASF/SF2 and components of the U1 snRNP particle are shown to be required for complex formation. We had shown that sequences at the 5' splice site of PIP7. A are necessary and now we show these are sufficient for complex formation. Furthermore, we show that one functional RNA binding domain and the RS domain are both required for ASF/SF2 to participate in complex formation. The RNA binding domains were redundant in this assay, suggesting that either domain can interact with the pre-messenger RNA. Finally, our experiments show no function for the U1-specific A protein in complex formation, whereas a function for U1-specific C protein was strongly suggested. The study of the earliest interactions between pre-mRNA and splicing factors suggests a model for 5' splice site recognition.

The human splicing factors ASF/SF2 and SC35 possess distinct, functionally significant RNA binding specificities

ASF/SF2 and SC35 belong to a highly conserved family of nuclear proteins that are both essential for splicing of pre-mRNA in vitro and are able to influence selection of alternative splice sites. An important question is whether these proteins display distinct RNA binding specificities and, if so, whether this influences their functional interactions with pre-mRNA. To address these issues, we first performed selection/amplification from pools of random RNA sequences (SELEX) with portions of the two proteins comprising the RNA binding domains (RBDs). Although both molecules selected mainly purine-rich sequences, comparison of individual sequences indicated that the motifs recognized are different. Binding assays performed with the full-length proteins confirmed that ASF/SF2 and SC35 indeed have distinct specificities, and at the same time provided evidence that the highly charged arginine-serine region of each protein is not a major determinant of specificity. In the case of ASF/SF2, evidence is presented that binding specificity involves cooperation between the protein's two RBDs. Finally, we demonstrate that an element containing three copies of a high-affinity ASF/SF2 binding site constitutes a powerful splicing enhancer. In contrast, a similar element consisting of three SC35 sites was inactive. The ASF/SF2 enhancer can be activated specifically in splicing-deficient S100 extracts by recombinant ASF/SF2 in conjunction with one or more additional protein factors. These and other results suggest a central role for ASF/SF2 in the function of purine-rich splicing enhancers.

Overexpression of the SR proteins ASF/SF2 and SC35 influences alternative splicing in vivo in diverse ways

The SR proteins are a family of essential splicing factors highly conserved throughout metazoa. Here we examine the effects of two prototypical SR proteins, ASF/SF2 and SC35, when overexpressed by transfection in cultured cells together with plasmids encoding alternatively spliced model transcripts. As expected from past work, both proteins were found to affect alternative splicing, but differences as well as similarities in their behavior were observed. With adenovirus E1a pre-mRNA, ASF/SF2 caused shifts in alternative splicing similar to those observed previously, and the effects of mutations, in the protein and the pre-mRNA, were largely consistent with in vitro results. For example, the C-terminal RS domain was not required to alter splice site selection. SC35 overexpression also altered E1a splicing, but the pattern was distinct from that detected with ASF/SF2, indicating that the two proteins can function differently in vivo. Unexpectedly, with SV40 early pre-mRNA, overexpression of either protein resulted in a marked inhibition of splicing, with the downstream small t 5' splice site more sensitive than the upstream large T 5' splice site. This is essentially the opposite of what has been observed when the concentration of these proteins is increased in vitro. The RS domain was necessary but not sufficient for this effect. Finally, overexpression of SC35, but not ASF/SF2, resulted in substantial accumulation of the unspliced SV40 pre-mRNA, which was efficiently transported to the cytoplasm. This finding suggests that SC35 may play an unanticipated role in mRNA stability and/or transport.

A novel U2-U6 snRNA structure is necessary for mammalian mRNA splicing

Splicing of mRNA precursors requires a complex and dynamic set of RNA-RNA base-pairing interactions in which the U2 and U6 snRNAs play central roles. Using a genetic suppression assay, we refine and extend a U2-U6 snRNA structure that may comprise the catalytic center of the spliceosome. We first show that a critical U2-U6 helix proven in yeast, helix Ia, is also essential for mammalian splicing. Mutations in the adjacent helix Ib, however, cannot be similarly suppressed, and relevant residues in both U2 and U6 are shown to participate in intramolecular, rather than intermolecular, base-pairing. We next demonstrate the requirement for a novel U2-U6 helix, helix III, which involves bases extending 3' from the branch site recognition sequence in U2 and 5' from an evolutionarily invariant sequence in U6 implicated previously in 5' splice site recognition. This configuration suggests that helix III may help juxtapose the pre-mRNA 5' splice site and branch site. We provide evidence for this by demonstrating that a branch site mutation can be suppressed by a mutation in the 5' splice site, provided that compensatory changes are made in the appropriate bases in U2 and U6. Our results provide new insights into how U2 and U6 snRNAs interact with each other and with the pre-mRNA to initiate the first catalytic step in splicing.

A direct interaction between a glutamine-rich activator and the N terminus of TFIIB can mediate transcriptional activation in vivo

Studies examining the mechanism by which transcriptional activators function have suggested that the general transcription factor IIB (TFIIB) can be a target for certain regulatory proteins. For example, we showed previously that expression of a mutant form of TFIIB can specifically inhibit activation in vivo mediated by the strong, glutamine-rich activator protein GAL4-ftzQ. Using transient cotransfection assays, we have defined the regions in both GAL4-ftzQ and TFIIB that are required for activity in vivo and provide evidence that a potential zinc finger structure at the N terminus of TFIIB is necessary for the observed functional interaction between the two proteins. Using a protein binding assay, we have demonstrated that GAL4-ftzQ can specifically interact with TFIIB in vitro. This interaction requires the same regions in both molecules necessary for function in vivo and is reduced or eliminated by mutations predicted to disrupt the zinc finger in TFIIB. These results support the idea that a direct interaction between a regulatory protein and TFIIB can be important for transcriptional activation in vivo and, combined with previous data of others, suggest that different activators can function by contacting distinct regions of TFIIB.

A complex protein assembly catalyzes polyadenylation of mRNA precursors

Nearly every eukaryotic mRNA contains a poly(A) tail at its 3' end. Unlike the factors that catalyze other RNA processing reactions, the basal polyadenylation machinery consists entirely of proteins. As with other steps in gene expression, however, this machinery is complex, consisting of multiple separable factors. Two multisubunit factors are required for specification of the poly(A) site and formation of a stable protein-RNA complex. Two additional proteins, as well as poly(A) polymerase, join the complex and are required for cleavage of the pre-mRNA and synthesis of the poly(A) tail, respectively. These and other properties suggest at least a superficial similarity between transcription initiation and polyadenylation.

Cooperation between core promoter elements influences transcriptional activity in vivo

Core promoters for RNA polymerase II frequently contain either (or both) of two consensus sequence elements, a TATA box and/or an initiator (Inr). Using test promoters consisting of prototypical TATA and/or Inr elements, together with binding sites for sequence-specific activators, we have analyzed the function of TATA and Inr elements in vivo. In the absence of activators, the TATA element was significantly more active than the Inr, and the combination of elements was only slightly more effective than the TATA-only promoter. In the presence of any of several coexpressed activator proteins, the TATA elements was again most active, but here addition of the Inr allowed significant increases in activity, indicating a cooperative interaction between the two elements. An interesting exception was observed with the activator Sp1, which was more effective with the Inr-only promoter, and addition of a TATA box did not enhance activity. Finally, in all cases the TATA plus Inr promoters were found to be partially or completely resistant to the dominant negative effects of a transcription factor TFIIB mutant previously shown to interfere with expression from TATA-only promoters. This result strengthens the conclusion that TATA and Inr elements can cooperate in vivo.

Regulation of dorsal in cultured cells by Toll and tube: tube function involves a novel mechanism

We described previously a transient cotransfection assay that allows us to study regulation of the Drosophila Dorsal protein (dl) in cultured cells. For example, we showed that over-expression of the Toll transmembrane receptor was sufficient to cause relocalization of dl from the cytoplasm to the nucleus. Here we present data that the tube protein, shown previously by genetic studies to act downstream of Toll, can function in a novel way to enhance dl activity. In the absence of dl, or when dl is cytoplasmic, tube is also found in the cytoplasm of transfected cells. But when dl is localized to the nucleus, so is tube. tube can then function to enhance reporter gene expression, either by cooperation with dl or as a GAL4-tube fusion protein. tube thus appears capable of acting both as a chaperon or escort for dl as it moves to the nucleus, and then as a transcriptional coactivator. We also show that the intracytoplasmic domain of Toll, and specifically the region sharing homology with the interleukin-1 receptor, is sufficient to induce dl-tube nuclear translocation.

A polyadenylation factor subunit is the human homologue of the Drosophila suppressor of forked protein

Polyadenylation of messenger RNA precursors is a complex process that requires multiple protein factors (for reviews, see refs 1, 2). Cleavage stimulation factor (CstF) is one of these, functioning together with cleavage-polyadenylation specificity factor, two cleavage factors, and poly(A)+ polymerase. CstF is composed of three subunits of M(r) 77, 64 and 50K. The 64K and 50K subunits contain, respectively, an RNP-type RNA-binding domain that contacts the pre-mRNA and transducin repeats characteristic of G-protein beta-subunits. Here we report the cloning and characterization of the 77K subunit of human CstF (referred to as 77K). We show that the 77K subunit is required for formation of active CstF and bridges the 64K and 50K subunits. Sequence analyses indicate that the 77K subunit is the homologue of the protein encoded by the Drosophila melanogaster suppressor of forked (su(f)) gene. Mutations in su(f) can enhance or suppress the effects of transposable element insertions, and our data indicate that this is due to changes in polyadenylation. Both the 77K subunit and the su(f) protein share homology with Saccharomyces cerevisiae RNA14, previously shown to be involved in mRNA metabolism. Our results thus also indicate that components of the complex polyadenylation machinery are conserved from yeast to man.

Poly(A) polymerase contains multiple functional domains

Poly(A) polymerase (PAP) contains regions of similarity with several known protein domains. Through site-directed mutagenesis, we provide evidence that PAP contains a functional ribonucleoprotein-type RNA binding domain (RBD) that is responsible for primer binding, making it the only known polymerase to contain such a domain. The RBD is adjacent to, and probably overlaps with, an apparent catalytic region responsible for polymerization. Despite the presence of sequence similarities, this catalytic domain appears to be distinct from the conserved polymerase module found in a large number of RNA-dependent polymerases. PAP contains two nuclear localization signals (NLSs) in its C terminus, each by itself similar to the consensus bipartite NLS found in many nuclear proteins. Mutagenesis experiments indicate that both signals, which are separated by nearly 140 residues, play important roles in directing PAP exclusively to the nucleus. Surprisingly, basic amino acids in the N-terminal-most NLS are also essential for AAUAAA-dependent polyadenylation but not for nonspecific poly(A) synthesis, suggesting that this region of PAP is involved in interactions both with nuclear targeting proteins and with nuclear polyadenylation factors. The serine/threonine-rich C terminus is multiply phosphorylated, including at sites affected by mutations in either NLS.

The human splicing factor ASF/SF2 can specifically recognize pre-mRNA 5' splice sites

ASF/SF2 is a human protein previously shown to function in in vitro pre-mRNA splicing as an essential factor necessary for all splices and also as an alternative splicing factor, capable of switching selection of 5' splice sites. To begin to study the protein's mechanism of action, we have investigated the RNA binding properties of purified recombinant ASF/SF2. Using UV crosslinking and gel shift assays, we demonstrate that the RNA binding region of ASF/SF2 can interact with RNA in a sequence-specific manner, recognizing the 5' splice site in each of two different pre-mRNAs. Point mutations in the 5' splice site consensus can reduce binding by as much as a factor of 100, with the largest effects observed in competition assays. These findings support a model in which ASF/SF2 aids in the recognition of pre-mRNA 5' splice sites.

Protein-protein interactions and 5'-splice-site recognition in mammalian mRNA precursors

Exactly how specific splice sites are recognized during the processing of complex precursor messenger RNAs is not clear. Small nuclear ribonucleoprotein particles (snRNPs) are involved, but are not sufficient by themselves to define splice sites. Now a human protein essential for splicing in vitro, called alternative splicing factor/splicing factor 2, is shown to cooperate with the U1 snRNP particle in binding pre-mRNA. This cooperation is probably achieved by specific interactions between the arginine/serine-rich domain of the splicing factor and a similar region in a U1 snRNP-specific protein.

Functional domains of the human splicing factor ASF/SF2

The human splicing factor ASF/SF2 displays two predominant activities in in vitro splicing assays: (i) it is an essential factor apparently required for all splices and (ii) it is able to switch utilization of alternative 5' splice sites in a concentration-dependent manner. ASF/SF2 is the prototype of a family of proteins typified by the presence of one or two RNP-type RNA binding domains (RBDs) and a region highly enriched in repeating arginine-serine dipeptides (RS regions). Here we describe a functional analysis of ASF/SF2, which defines several regions essential for one, or both, of its two principal activities, and provides insights into how this type of protein functions in splicing. Two isoforms of the protein, which arise from alternative splicing, are by themselves inactive, but each can block the activity of ASF/SF2, thereby functioning as splicing repressors. Some, but not all, mutations in the RS region prevent ASF/SF2 from functioning as an essential splicing factor. However, the entire RS region can be deleted without reducing splice site switching activity, indicating that it is not absolutely required for interaction with other splicing factors. Experiments with deletion and substitution mutants reveal that the protein contains two related, but highly diverged, RBDs, and that both are essential for activity. Each RBD by itself retains the ability to bind RNA, although optimal binding requires both domains.

Functional domains of the Drosophila Engrailed protein

We have studied the transcriptional activity of the Drosophila homeodomain protein Engrailed (En) by using a transient expression assay employing Schneider L2 cells. En was found to very strongly repress promoters activated by a variety of different activator proteins. However, unlike another Drosophila homeodomain-containing repressor, Even-skipped (Eve), En was unable to repress the activity of several basal promoters in the absence of activator expression. These findings indicate that En is a specific repressor of activated transcription, and suggest that En may repress transcription by a different mechanism than Eve, perhaps by interfering with interactions between transcriptional activators and the general transcription machinery. By analyzing the properties of a variety of En mutants, we identified a minimal repression domain composed of 55 residues, which can function when fused to a heterologous DNA binding domain. Like repression domains identified in the Drosophila repressors Eve and Krüppel, the En repression domain is rich in alanine residues (26%), but unlike these other domains, is moderately charged (six arginine and three glutamic acid residues). Separate regions of En that may in some circumstances function in transcriptional activation were also identified.

Identification of an snRNP-associated kinase activity that phosphorylates arginine/serine rich domains typical of splicing factors

The U1 snRNP-specific 70K protein is one of the few snRNP proteins from higher eukaryotic cells that is phosphorylated in vivo (1,2). Immunoaffinity purified spliceosomal snRNPs (U1, U2, U5, and U4/U6) were tested for their ability to phosphorylate in vitro the U1-specific 70K protein. An snRNP-associated kinase activity which phosphorylates all U1-70K isoelectric variants was identified. Like its in vivo counterpart, this snRNP-associated enzyme phosphorylates solely serine residues of the 70K protein, preferentially utilizing ATP as a phosphodonor. Tryptic phosphopeptide analysis revealed an overlapping set of at least four radiolabeled peptides in the in vivo and in vitro phosphorylated protein, suggesting that the snRNP-associated serine kinase is responsible, at least in part, for the 70K protein phosphorylation observed in vivo. Chymotryptic digestion of in vitro, 32P-labeled 70K protein and in vitro phosphorylation studies with a synthetic peptide, indicated that the multiple 70K phosphorylation sites are limited to a highly charged, C-terminal domain of the protein. In vitro phosphorylation studies with the splicing factor ASF/SF2 and several deletion mutants demonstrated that, similar to the U1-70K protein, the snRNP-associated serine kinase phosphorylates the carboxy terminal RS-rich domain of ASF/SF2. A potential general role for this enzyme in the phosphorylation of splicing factors and its consequences for pre-mRNA splicing regulation are discussed.

Interaction between a transcriptional activator and transcription factor IIB in vivo

Transcription of messenger RNA-encoding genes in vitro requires many protein factors. Transcription factor IID, possibly with the cooperation of TFIIA, binds to the TATA element of the promoter, forming a complex that can bind TFIIB (refs 6, 7) followed by RNA polymerase II (refs 6, 8) and other factors. One or more of these steps is thought to be facilitated by gene-specific transcriptional activation proteins; this seems to require TFIID-associated auxiliary factors and may involve direct contact between the activator and TFIID and/or TFIIB. If such contact is necessary in vivo, activation might conceivably be blocked by a TFIIB derivative containing the sequences necessary for this interaction, but lacking those necessary for binding to the rest of the transcriptional apparatus, an effect similar to that referred to as squelching or transcriptional interference. Here we show that the activity of the glutamine-rich fushi tarazu activation domain is indeed blocked by truncated TFIIB derivatives in Drosophila Schneider L2 cells, suggesting that it is mediated by interactions with TFIIB.

Transcriptional repression by the Drosophila even-skipped protein: definition of a minimal repression domain

We have used a transient expression assay employing Drosophila tissue culture cells to study the transcriptional repression activity of the homeo domain protein Even-skipped (Eve). Eve was found to repress all promoters that contained Eve-binding sites, including both TATA-containing and TATA-lacking minimal promoters, as well as promoters activated by several different classes of activator proteins. These findings suggest that the general transcription machinery can be a target of Eve. By analyzing properties of a variety of Eve mutants and chimeric fusion proteins, we have identified several features important for efficient repression. In addition to the DNA-binding domain, a potent repressor requires a repression domain, which can be as small as 27 residues. The minimal 57-residue Eve repression domain, as well as several others studied here, were all found to be proline rich and to contain a high percentage of hydrophobic residues. An intriguing feature of the strong repressors was that their DNA-binding activities, measured by gel retention assays with nuclear extracts, were significantly less than those of derivatives inactive in repression.

The Drosophila RNA-binding protein RBP1 is localized to transcriptionally active sites of chromosomes and shows a functional similarity to human splicing factor ASF/SF2

An RNA-binding protein gene (rbp1) from Drosophila melanogaster, encoding an RNA recognition motif and an Arg-Ser rich (RS) domain, has been characterized. The predicted amino acid sequence of rbp1 is similar to those of the human splicing factor ASF/SF2, the Drosophila nuclear phosphoprotein SRp55, and the Drosophila puff-associated protein B52. Northern and immunohistochemical analyses showed that rbp1 is expressed at all stages in all tissues and that the RBP1 protein is localized to the nucleus. Consistent with a role in mRNA metabolism, indirect immunofluorescence reveals that the RBP1 protein colocalizes with RNA polymerase II on larval salivary gland polytene chromosomes. RBP1 protein made in Escherichia coli was tested for splicing activity using human cell extracts in which ASF has been shown previously both to activate splicing and to affect the choice of splice sites in alternatively spliced pre-mRNAs. In these assays, RBP1 protein, like ASF, is capable of both activating splicing and switching splice site selection. However, in each case, clear differences in the behavior of the two proteins were detected, suggesting that they have related but not identical functions. The general nuclear expression pattern, colocalization on chromosomes with RNA polymerase II, the similarity to ASF/SF2, SRp55, and B52, along with the effect on alternative splicing shown in vitro, suggest that rbp1 is involved in the processing of precursor mRNAs.

Multiple functional domains of human U2 small nuclear RNA: strengthening conserved stem I can block splicing

We showed previously that a branch site mutation in simian virus 40 early pre-mRNA that prevented small t antigen mRNA splicing could be efficiently suppressed by a compensatory mutation in a coexpressed U2 small nuclear (sn) RNA gene. We have now generated second-site mutations in this suppressor gene to investigate regions of U2 RNA required for function. A number of mutations in a putative stem at the 5' end of the molecule inhibited splicing, indicating that bases in this region are important for activity. However, several lines of evidence suggested that formation of the entire stem is not essential for splicing. Indeed, mutations that strengthen the stem actually inhibited splicing, and evidence that this prevents a required base-pairing interaction with U6 snRNA is presented. These results suggest that the relative stabilities of competing intra- and intermolecular base-pairing interactions play an important role in the splicing reaction. Mutations in a conserved single-stranded region immediately 3' to the branch site recognition sequence all inhibited splicing, indicating that this region is required for U2 function, although its exact role remains unknown. Finally, two mutations in the loop of stem IV at the 3' end of the molecule, which destroy the binding site of U2 sn ribonucleoprotein B", prevented small t splicing; this finding contrasts with previous studies which utilized different assay systems. Analysis of the accumulation and subcellular localization of all of the mutant RNAs showed that they were similar to those of the parental suppressor U2 RNA, indicating that the effects observed indeed reflect defects in splicing.

A human polyadenylation factor is a G protein beta-subunit homologue

Cleavage stimulation factor (CstF) is one of the multiple factors required for polyadenylation of mammalian pre-mRNAs in vitro. We have shown previously that this factor is composed of three distinct subunits of 77, 64, and 50 kDa, and that the 64-kDa subunit can be UV-cross-linked to RNA in a polyadenylation signal (AAUAAA)-dependent manner. By molecular cloning, the 64-kDa subunit was shown to contain a ribonucleoprotein-type RNA binding domain and a novel repeat structure. To study the functions of the other subunits, we have now isolated cDNAs encoding the 50-kDa subunit of human CstF. This subunit shares extensive homology with mammalian G protein beta-subunits and has a characteristic repeat structure (transducin repeat), in which an approximately 44-amino acid-long sequence is repeated seven times. To our knowledge, the 50-kDa subunit is the first example of a functional beta-subunit-like protein in vertebrates. Possible roles of the transducin repeat, both in CstF function specifically and in other beta-subunit homologues more generally, are discussed.

Selective nuclear transport of the Drosophila morphogen dorsal can be established by a signaling pathway involving the transmembrane protein Toll and protein kinase A

Establishment of dorsal-ventral polarity in the early Drosophila embryo requires a concentration gradient of the maternal morphogen dorsal (dl). This concentration gradient is established by selective nuclear transport of dl so that dl protein is present only in ventral nuclei. The activity of 11 genes is required for dl nuclear localization. One of these genes, Toll, encodes a transmembrane protein that appears to play the most direct role in regulating dl localization. We have examined the effects of Toll on dl in cotransfected Schneider cells to gain insight into the nature of the interaction between these proteins. We have found that Toll can enhance the nuclear localization of dl and, independently, the ability of dl to activate transcription once in the nucleus. We present evidence that the signaling pathway from Toll to dl involves protein kinase A (PKA) and that nuclear transport and activation of dl results from phosphorylation of dl by PKA. We discuss the significance of these results with respect both to Drosophila embryogenesis and to the regulation of the mammalian transcription factor NF-kappa B.

Ectopic expression of the Drosophila tramtrack gene results in multiple embryonic defects, including repression of even-skipped and fushi tarazu

The tramtrack (ttk) gene of Drosophila encodes 69-kDa and 88-kDa proteins through alternative splicing of the primary ttk transcript. The two proteins share a common amino-terminal sequence, but contain different carboxy-terminal portions, each of which has a distinct zinc finger domain with a unique DNA binding specificity. The 69-kDa ttk protein has been shown to bind multiple sites within important regulatory elements of the pair-rule genes even-skipped (eve) and fushi tarazu (ftz), and it has been suggested that this protein may function as a repressor of ftz transcription. Here we present evidence that the 69-kDa ttk protein can indeed repress expression not only of ftz, but also of eve. Ectopic expression of the 69-kDa protein, but not of the 88-kDa form, was found to nearly abolish the striped patterns of expression of both eve and ftz in transgenic embryos. These findings, coupled with our detection of significant levels of ttk protein in ovaries and 0-2-h embryos, support the idea that maternally supplied ttk protein serves to prevent premature activation of eve and ftz, thereby helping to establish the timing of the onset of zygotic expression of these two genes. Furthermore, gross defects in the larval cuticle resulting from misexpression of the 69-kDa protein suggest that this protein performs additional functions in the early embryo.

Characterization of the multisubunit cleavage-polyadenylation specificity factor from calf thymus

Cleavage-polyadenylation specificity factor (CPSF) is one of five separable factors known to be required for 3' cleavage and polyadenylation of mRNA precursors in vitro. Previous studies have shown that the cleavage and poly(A) addition reactions can be uncoupled in vitro and have suggested that CPSF may be the only factor essential for both of these subreactions. Here we report the purification of CPSF to near homogeneity from calf thymus and show that the purified factor contains three polypeptides of 165, 105, and 70 kDa. These polypeptides cosediment precisely with CPSF activity, which has a sedimentation coefficient of 11.5 S. Consistent with previous reports from our laboratory, purified CPSF does not contain a detectable RNA component, indicating that it is a multisubunit protein and not a small nuclear ribonucleoprotein. Extensively purified bovine CPSF can function with human poly(A) polymerase to bring about AAUAAA-dependent poly(A) addition or with human cleavage factors to catalyze accurate 3' cleavage of a pre-mRNA substrate. UV cross-linking and gel retention analyses demonstrate that highly purified CPSF interacts with one of these cleavage factors, the multisubunit cleavage-stimulation factor, to facilitate stable binding of both to an AAUAAA-containing pre-mRNA. Likewise, evidence is presented indicating that poly(A) polymerase and CPSF can interact directly.

Alternatively spliced transcripts of the Drosophila tramtrack gene encode zinc finger proteins with distinct DNA binding specificities

A protein present in nuclear extracts of Drosophila embryos binds multiple sites in the promoter and genetically defined autoregulatory element of the pair-rule gene even-skipped (eve). We reported here the isolation of a cDNA encoding this binding activity, the sequence of which identifies it as the 69 kDa zinc finger tramtrack (ttk) protein. As ttk was previously implicated in controlling the expression of another pair-rule gene, fushi tarazu (ftz), our findings suggest that ttk plays a role in the regulation of at least two developmentally important genes. An additional ttk-related cDNA clone was isolated which gives rise to an 88 kDa protein with an alternative set of zinc fingers having a DNA binding specificity distinct from that of the 69 kDa protein. Both proteins were shown to be encoded by the ttk gene through alternative splicing, providing the first example of the use of this mechanism to generate related proteins with distinct DNA binding specificities. Whole mount in situ hybridization analysis revealed different patterns of embryonic expression of the two ttk mRNA isoforms.

The human 64-kDa polyadenylylation factor contains a ribonucleoprotein-type RNA binding domain and unusual auxiliary motifs

Cleavage stimulation factor is one of the multiple factors required for 3'-end cleavage of mammalian pre-mRNAs. We have shown previously that this factor is composed of three subunits with estimated molecular masses of 77, 64, and 50 kDa and that the 64-kDa subunit can be UV-crosslinked to RNA in a polyadenylylation signal (AAUAAA)-dependent manner. We have now isolated cDNAs encoding the 64-kDa subunit of human cleavage stimulation factor. The 64-kDa subunit contains a ribonucleoprotein-type RNA binding domain in the N-terminal region and a repeat structure in the C-terminal region in which a pentapeptide sequence (consensus MEARA/G) is repeated 12 times and the formation of a long alpha-helix stabilized by salt bridges is predicted. An approximately 270-amino acid segment surrounding this repeat structure is highly enriched in proline and glycine residues (approximately 20% for each). When cloned 64-kDa subunit was expressed in Escherichia coli, an N-terminal fragment containing the RNA binding domain bound to RNAs in a polyadenylylation-signal-independent manner, suggesting that the RNA binding domain is directly involved in the binding of the 64-kDa subunit to pre-mRNAs.

TFIID can be rate limiting in vivo for TATA-containing, but not TATA-lacking, RNA polymerase II promoters

We have studied the effect of exogenous expression of the basal transcription factor TFIID on the activities of several different TATA-containing and TATA-lacking promoters. Overexpression of TFIID from a transfected plasmid in Drosophila Schneider cells resulted in substantial concentration-dependent increases in expression from a cotransfected minimal TATA-containing promoter. Overexpression of TFIID activated expression from all TATA-containing promoters tested, with the maximum level of activation being inversely proportional to the strength of the promoter. In contrast, expression from TATA-less promoters was not enhanced, and could in fact be reduced, by increased expression of TFIID. Consistent with these findings overexpression of TFIID had opposite effects on Sp1-mediated activation observed from minimal synthetic promoters consisting of Sp1-binding sites and either a TATA box or initiator element. We discuss the significance of these results in terms of the role of TFIID in the initiation of transcription and as a possible regulatory target for expression from TATA-containing promoters, as well as the role TFIID may play in expression from TATA-less promoters.

SV40 T-antigen-binding sites within the 5'-flanking regions of human U1 and U2 genes

The 5' flanking regions of the genes (U1 and U2) encoding the human U1 and U2 small nuclear RNAs (snRNAs) each contain sequences that bind specifically to the simian virus (SV40) large tumor antigen (T.Ag). Substitution of these sites with sequences that lack T.Ag-binding sites did not block accumulation of U1 or U2 snRNA in a variety of cell types, but deletion of these regions resulted in the total loss of expression. Thus, these sequences may serve only a spacing function, and the T.Ag-binding sites appear not to be necessary for expression. However, coexpression of T.Ag markedly reduced expression of a U1 gene containing a high-affinity T.Ag-binding site (from the SV40 genome) in place of the U1 T.Ag-binding site. In contrast, coexpression of T.Ag enhanced synthesis of U2, but not U1, snRNA, independent of the presence of the T.Ag-binding sites. Thus, while the consensus T.Ag-binding sites within the U1 and U2 promoter regions do not appear to influence expression, the binding of SV40 T.Ag to a high-affinity site can lead to significant repression of a strong snRNA promoter, and T.Ag can enhance expression of another in the absence of a known binding site.

A novel protein factor is required for use of distal alternative 5' splice sites in vitro

Adenovirus E1A pre-mRNA was used as a model to examine alternative 5' splice site selection during in vitro splicing reactions. Strong preference for the downstream 13S 5' splice site over the upstream 12S or 9S 5' splice sites was observed. However, the 12S 5' splice site was used efficiently when a mutant pre-mRNA lacking the 13S 5' splice site was processed, and 12S splicing from this substrate was not reduced by 13S splicing from a separate pre-mRNA, demonstrating that 13S splicing reduced 12S 5' splice site selection through a bona fide cis-competition. DEAE-cellulose chromatography of nuclear extract yielded two fractions with different splicing activities. The bound fraction contained all components required for efficient splicing of simple substrates but was unable to utilize alternative 5' splice sites. In contrast, the flow-through fraction, which by itself was inactive, contained an activity required for alternative splicing and was shown to stimulate 12S and 9S splicing, while reducing 13S splicing, when added to reactions carried out by the bound fraction. Furthermore, the activity, which we have called distal splicing factor (DSF), enhanced utilization of an upstream 5' splice site on a simian virus 40 early pre-mRNA, suggesting that the factor acts in a position-dependent, substrate-independent fashion. Several lines of evidence are presented suggesting that DSF is a non-small nuclear ribonucleoprotein protein. Finally, we describe a functional interaction between DSF and ASF, a protein that enhances use of downstream 5' splice sites.

Primary structure and expression of bovine poly(A) polymerase

Poly(A) polymerase has a critical role in the synthesis of messenger RNA in eukaryotic cells. The isolation and characterization of complementary DNAs encoding bovine poly(A) polymerase is described here. The predicted sequences of the mRNA and protein reveal features that provide insights into how the enzyme functions and how it might be regulated. Poly(A) polymerase expressed from a cloned cDNA is fully functional in in vitro assays, and mutational analyses have identified a putative regulatory domain that enhances, but is not essential for, activity.

Base pairing between U2 and U6 snRNAs is necessary for splicing of a mammalian pre-mRNA

Splicing of pre-messenger RNA in eukaryotic cells occurs in a multicomponent complex termed the spliceosome, which contains small nuclear ribonucleoprotein particles (snRNPs), protein factors and substrate pre-mRNA. Assembly of the spliceosome involves the stepwise binding of snRNPs and protein factors to the pre-mRNA through a poorly understood mechanism which probably involves specific RNA-RNA, RNA-protein and protein-protein interactions. Of particular interest are the interactions between snRNPs, which are likely to be important not only for assembly of the spliceosome but also for catalysis. U1 snRNP interacts with the 5' splice site and U2 snRNP with the branch site of the pre-mRNA; both of these interactions involve Watson-Crick base pairing. But very little is known about how other factors such as the U4/U6 and U5 snRNPs reach the spliceosome and function in splicing. Here we report evidence that U6 snRNA interacts directly with U2 snRNA by a mechanism involving base-pairing, and that this interaction can be necessary for splicing of a mammalian pre-mRNA in vivo.

Primary structure of the human splicing factor ASF reveals similarities with Drosophila regulators

We described previously the purification of a human protein, called alternative splicing factor (ASF), that can switch utilization of alternative 5' splice sites in an SV40 early pre-mRNA. We now report the isolation of a cDNA, designated ASF-1, that encodes this protein. ASF-1 consists of 248 amino acid residues, including an 80 residue RNA-binding domain at its N-terminus and a 50 residue C-terminal region that is 80% serine plus arginine. ASF-1 produced in E. coli can activate splicing in vitro and switch 5' splice-site utilization, establishing that the recombinant protein is sufficient to supply these activities. Analysis of additional cDNAs revealed that ASF pre-mRNA can itself be alternatively spliced, surprisingly, by utilization of a shared 5' splice site and two closely spaced 3' splice sites. Use of the upstream site results in a second mRNA (ASF-2) in which translation of the downstream exon occurs extensively in an alternative reading frame distinct from ASF-1.

A nuclear micrococcal-sensitive, ATP-dependent exoribonuclease degrades uncapped but not capped RNA substrates

We have developed an assay for an exoribonuclease present in HeLa cell nuclear extracts that degrades capped but not uncapped RNA substrates, and used it to partially purify and characterize such an activity. Capped and uncapped transcripts of different sizes (37-317 nt) were incubated with fractionated nuclear extracts, and in all cases the capped RNAs were stable while their uncapped counterparts were completely degraded. No changes in activity were detected when cap analogs were included in reaction mixtures, suggesting that the stability of capped RNAs was not due to a cap binding protein. The exoribonuclease was shown to be specific for RNA, and to function processively with either substrates containing 5'-hydroxyl or 5'-phosphorylated ends. The products were predominantly 5'-mononucleotides, and no detectable intermediates were observed at any reaction time points. Sedimentation analysis suggests that the native size of the nuclease is 7.4S or approximately 150 kDa. Interestingly, a nucleoside triphosphate was found to be necessary for specific and complete degradation of the uncapped RNAs. Finally, micrococcal nuclease (MN) pretreatment of the partially purified enzyme inhibited its activity. As several controls indicated that this was not due to non-specific effects of MN, this finding suggests that the exoribonuclease contains an essential RNA component.

Activation and repression of transcription by the gap proteins hunchback and Krüppel in cultured Drosophila cells

We have studied the ability of the Drosophila gap proteins Krüppel and hunchback to function as transcriptional regulators in cultured cells. Both proteins bind to specific sites in a 100-bp DNA fragment located upstream of the segment polarity gene engrailed, which also contains functional binding sites for a number of homeo box proteins. The hunchback protein is a strikingly concentration-dependent activator of transcription, capable of functioning both by itself and also synergistically with the pair-rule proteins fushi tarazu and paired. In contrast, Krüppel is a transcriptional repressor that can block transcription induced either by hunchback or by several different homeo box proteins. While repression of the homeo box protein activators requires a Krüppel-binding site on the DNA, repression of hunchback can occur efficiently in the absence of a Krüppel-binding site. We discuss the possible molecular mechanisms underlying these activities, as well as the potential significance of these results with respect to segmentation in Drosophila.

A multisubunit factor, CstF, is required for polyadenylation of mammalian pre-mRNAs

We have purified and characterized a factor required for accurate polyadenylation of mammalian pre-mRNAs in vitro. This factor, called cleavage-stimulation factor (CstF), is composed of three distinct polypeptide subunits of 77, 64, and 50 kD. Using monoclonal antibodies directed against the 64- and 50-kD subunits, we show that CstF is required for efficient cleavage of polyadenylation substrates. Furthermore, CstF present in unfractionated nuclear extracts interacts with pre-mRNAs containing the signal sequence AAUAAA, but not AAGAAA, in such a manner that the 64-kD subunit can be cross-linked to the RNA by UV light. This polypeptide is thus identical to the previously described 64-kD nuclear protein that binds to AAUAAA-containing RNAs. Finally, indirect immunofluorescence of fixed cells indicates that CstF is distributed diffusely throughout the nucleus in a granular pattern distinct from the "speckled" pattern displayed by factors involved in pre-mRNA splicing, but similar to that of heterogeneous nuclear ribonucleoproteins. A model is presented in which CstF binds specifically to nascent RNA polymerase II transcripts and, by interacting with other factors, results in a rapid initiation of 3'-end processing of pre-mRNAs.

Sequence specificity of the human mRNA N6-adenosine methylase in vitro

N6-adenosine methylation is a frequent modification of mRNAs and their precursors, but little is known about the mechanism of the reaction or the function of the modification. To explore these questions, we developed conditions to examine N6-adenosine methylase activity in HeLa cell nuclear extracts. Transfer of the methyl group from S-[3H methyl]-adenosylmethionine to unlabeled random copolymer RNA substrates of varying ribonucleotide composition revealed a substrate specificity consistent with a previously deduced consensus sequence, Pu[G greater than A]AC[A/C/U]. 32-P labeled RNA substrates of defined sequence were used to examine the minimum sequence requirements for methylation. Each RNA was 20 nucleotides long, and contained either the core consensus sequence GGACU, or some variation of this sequence. RNAs containing GGACU, either in single or multiple copies, were good substrates for methylation, whereas RNAs containing single base substitutions within the GGACU sequence gave dramatically reduced methylation. These results demonstrate that the N6-adenosine methylase has a strict sequence specificity, and that there is no requirement for extended sequences or secondary structures for methylation. Recognition of this sequence does not require an RNA component, as micrococcal nuclease pretreatment of nuclear extracts actually increased methylation efficiency.

The Drosophila even-skipped promoter is transcribed in a stage-specific manner in vitro and contains multiple, overlapping factor-binding sites

To investigate the factors contributing to regulation of expression of the Drosophila segmentation gene even-skipped (eve), we have analyzed both the in vitro transcription and eve-promoter-binding proteins in embryo extracts. We show that the eve promoter is accurately and efficiently expressed in nuclear extracts derived from Drosophila embryos and that transcription is more efficient in extracts prepared from embryos at early stages of development than in those from older embryos, broadly reproducing the temporal pattern of expression observed in vivo. This stage-specific expression is dependent on sequences upstream of the eve transcription start site which contain multiple binding sites for at least two distinct proteins present in embryo nuclei. One of these proteins, the binding sites for which correspond to the sequences required for stage-specific expression, appears to be the previously described GAGA factor. Although the binding activity of the GAGA factor remains constant, the level of the binding activity of the other protein, which we have called the TCCT factor, changes during the course of embryogenesis. Activity is first detected 3 to 5 h after fertilization and decreases during later stages of embryogenesis. We discuss the possibility that the TCCT factor plays a role in the maintenance or refinement of the eve expression pattern.

A protein factor, ASF, controls cell-specific alternative splicing of SV40 early pre-mRNA in vitro

SV40 early pre-mRNA is alternatively spliced by utilization of two different 5' splice sites and a shared 3' splice site to produce large T and small t mRNAs. The ratio of small t to large T mRNAs produced in human embryonic kidney 293 cells is 10- to 20-fold greater than in other mammalian cells, suggesting the existence of a 293 cell-specific factor that modulates alternative splicing. Here we show that nuclear extracts from 293 cells give rise to significantly more small t splicing than do extracts from HeLa cells. Using an in vitro complementation assay, we have characterized and extensively purified a factor from 293 extracts that brings about striking increases in small t splicing with concomitant decreases in large T splicing. The factor is heat sensitive and micrococcal nuclease resistant, suggesting that it is a protein lacking an accessible RNA component. Purification of the alternative splicing factor indicates that the activity is contained in one of several possibly related polypeptides of 30-35 kd.

Polyoma virus small tumor antigen pre-mRNA splicing requires cooperation between two 3' splice sites

We have studied splicing of the polyoma virus early region pre-mRNA in vitro. This RNA is alternatively spliced in vivo to produce mRNA encoding the large, middle-sized (MTAg), and small (StAg) tumor antigens. Our primary interest was to learn how the 48-nucleotide StAg intron is excised, because the length of this intron is significantly less than the apparent minimum established for mammalian introns. Although the products of all three splices are detected in vitro, characterization of the pathway and sequence requirements of StAg splicing suggests that splicing factors interact with the precursor RNA in an unexpected way to catalyze removal of this intron. Specifically, StAg splicing uses either of two lariat branch points, one of which is located only 4 nucleotides from the 3' splice site. Furthermore, the StAg splice absolutely requires that the alternative MTAg 3' splice site, located 14 nucleotides downstream of the StAg 3' splice site, be intact. Insertion mutations that increase or decrease the quality of the MTAg pyrimidine stretch enhance or repress StAg as well as MTAg splicing, and a single-base change in the MTAg AG splice acceptor totally blocks both splices. These results demonstrate the ability of two 3' splice sites to cooperate with each other to bring about removal of a single intron.