INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Many genes are interrupted by introns that are removed in the process of making a functional mRNA and, ultimately, a protein product. This splicing process takes place in a megadalton RNA-protein complex termed the spliceosome. The spliceosome consists of many components, including five small nuclear ribonucleoprotein particles (snRNPs) each consisting of a small nuclear RNA and a complement of proteins (17). Additionally, 100 or more non-snRNP proteins are present in the spliceosome (23).

One group of non-snRNP proteins that has been studied intensively since their discovery is the SR protein family (8). SR proteins have a similar protein structure, with at least one RNA recognition motif (RRM) at the N terminus and a region rich in arginine-serine dipeptides (RS domain) at the C terminus. The RRM consists of 80 to 90 amino acids; is present in many RNA-binding proteins, including members of the SR protein family; and confers RNA binding specificity (13, 20). The RS domain has been postulated to mediate protein-protein interactions between SR family members (15, 39, 40). The SR protein family includes proteins with sizes of 20, 30, 40, 55, and 75 kDa, and the high degree of conservation from Drosophila to humans indicates the importance of these proteins (8). This family of SR proteins, purified by a two-step salt precipitation procedure, has a well-defined biochemical activity, such that each member can complement a splicing-deficient cytoplasmic (S100) extract, indicative of a functional redundancy among SR proteins.

In Drosophila melanogaster, only two SR proteins, RBP1 and B52, have been cloned and characterized to date. RBP1, homologous to human SRp20, is expressed in all tissues throughout development and colocalizes with RNA polymerase II on larval salivary gland polytene chromosomes. RBP1 also tested positive for splicing activity in a cell-free human splicing system (14). B52, the Drosophila homolog of human SRp55 is present during all stages of development and was initially characterized because of its localization to heat shock puffs on polytene chromosomes. B52 brackets RNA polymerase at the largest heat shock puffs and stains within developmental puffs (4, 18). The proper level of expression of B52 is critical for development of the animal. Overexpression of B52 in various tissues results in developmental arrest in almost all cases (18). The generation of a B52-null mutant showed that B52 expression was essential, given that these animals do not develop past the second-instar larval stage. Surprisingly, several genes analyzed thus far show proper splicing patterns in arrested larvae of the B52-null mutant (22, 24). Thus, a global RNA processing defect may not explain the lethality of the B52 deletion mutation.

In order to understand the cause of the lethality of the null mutation, we sought to characterize more thoroughly the splicing activity of B52 in vivo and in vitro. Here, we show that B52 is essential for splicing an ftz pre-mRNA in Drosophila extracts in which B52 is the predominant SR protein. In vivo, other SR proteins appear to complement the B52 deficiency in most tissues. However, by analyzing RNA isolated from the brain, a tissue in which B52 is the major SR protein expressed, we showed that the production of ftz mRNA is dependent on B52. These results support the hypothesis that the lethality seen in the B52-null mutant is a consequence of splicing defects in tissues that lack sufficient concentrations of other SR protein family members.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Northern blot analysis. RNAs were transferred to a Gene Screen Plus membrane (Dupont) by using a Trans-Blot apparatus (Bio-Rad). The membrane was prewashed and hybridized at 42°C, in a solution containing 20% formamide, 10% dextran sulfate, and 2× SSC (1× SSC is 0.15 M NaCl and 0.015 M sodium citrate), with 250 µl of salmon testis DNA (10 mg/ml) and 5 × 10⁶ cpm of a ³²P-labeled, kinase-treated oligonucleotide (5'-AGTCTTTGCAATCTGATGCCAAAG). The membrane was washed twice in 6× SSC-0.1% sodium dodecyl sulfate (SDS) for 20 min at room temperature and once in 6× SSC prior to exposure for autoradiography.

In vitro splicing assay. The ftz pre-mRNA substrate was synthesized in 50-µl reaction volumes containing the following: 0.04 mg of template DNA/ml; T7 RNA polymerase; 0.5 mM diguanosine triphosphate; 0.4 mM each ATP, CTP, and GTP; 0.3 mM UTP; and 5 µl of [³²P]UTP (NEN 007H). The reaction mixtures were incubated at 37°C for 30 min. The template DNA was removed by a DNase digestion, and free nucleotides were removed by using a Sephadex G50 spin column. The RNA was extracted with phenol and chloroform and then precipitated with ethanol. Nuclear and cytoplasmic (S100) extracts were made from Kc cells (6). In vitro splicing reaction mixtures, consisting of nuclear extract or S100 supplemented with B52, were assembled as described previously (25), and reactions were carried out at 20°C for 90 min. The resulting RNAs were treated with proteinase K (0.6 mg/ml) for 10 min at 30°C and purified by a phenol-chloroform extraction followed by an ethanol precipitation. The RNAs were separated on a 6% polyacrylamide-7 M urea gel.

Immunoblot analysis. Proteins separated on a sodium dodecyl sulfate (SDS)-polyacrylamide gel (10 or 12.5%) were transferred to nitrocellulose filters (Schleicher & Schuell). The filters were blocked in Tris-buffered saline containing 3% gelatin and 10% calf serum, washed, and incubated with primary antibody (0.15 µg of the RRM antibody [18] or the mAb104 antibody [27]/ml) in 1% gelatin containing 10% calf serum. The filters were washed again, incubated with horseradish peroxidase-conjugated anti-rabbit or anti-mouse immunoglobulin G (Amersham), and processed by chemiluminescence according to the directions of the manufacturer. Frozen Drosophila larvae or dissected tissues were ground to a fine powder and sonicated in isolation buffer (41). The proteins from whole larvae were centrifuged at 8,000 × g for 20 min, while the proteins from tissues were spun at 13,000 × g for 5 min. The supernatants were collected, and aliquots were loaded on a gel.

D. melanogaster transformation. The ftz pre-mRNA substrate was subcloned into the pUAST transformation vector (1) to produce the plasmid UAS_gal-ftz. UAS_gal-ftz was introduced into the B52 deletion fly line (B52²⁸/TM6b Tb e) by P-element-mediated germ line transformation (26, 29, 33). An insertion into the B52²⁸ third chromosome was identified and confirmed by segregation analysis of chromosomes. To produce ftz pre-mRNA, these flies were crossed with a strain that uses the hsp70 promoter to express GAL4.

RT-PCR. RNA isolation and reverse transcription (RT)-PCR of whole larvae were performed as described previously (24). The following gene-specific primer pairs were used for: ftz, 5'-AACTGCAGGTCGACTACTTGGACGTCTACTCGC and 5'-CGGGATCCCTCGAGCTCCAGGGTCTGGT; for B52, 5'-AAGGAAGCGTTATCATGGTGGGA and 5'-GGATGTCGCGTGTGCGGCCGTATC; and for Hsp70, 5'-AAGGACAACAATGCATTGGGC and 5'-GCTGGCCGCAGTTTGCTCCCGG. All samples were treated with DNase to eliminate any possible DNA contamination of the RNA preparation. Larval dissections were performed in Ringer's solution (110 mM NaCl, 4 mM KCl, 1 mM MgCl₂, 5 mM CaCl₂, 10 mM HEPES [pH 7.2]), and the isolated tissues were frozen immediately on dry ice. Carrier tRNA (10 µg) was added, and RNA was isolated as described above. The RNA was reverse transcribed as described above except that random hexamers were used as primers. Ten percent of the RT product served as the template for the PCR, which was kept in the linear range of amplification using primers for the ftz gene as well as 18S rRNA internal standards (Ambion). The products were separated on a 4% polyacrylamide-7 M urea gel.

Protein preparations. (i) B52. B52 was prepared as described previously (32).

(ii) RBP1. RBP1 was purified by Ni²⁺ chelate chromatography as described elsewhere (14).

(iii) SR proteins. SR protein preparation was performed as described elsewhere (41) with the following modification. After ammonium sulfate precipitation, gel filtration was carried out instead of dialysis, with G25 resin (Sigma) being used to rapidly desalt the proteins.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

B52 activity in a Drosophila in vitro splicing system. We characterized the splicing activity of B52 by using a homologous Drosophila in vitro splicing system composed of a mini-pre-mRNA substrate from the Drosophila fushi tarazu gene (ftz) (25) and nuclear extract prepared from a Drosophila cell line (Kc). Northern blotting with an exon junction oligonucleotide probe was used to identify the spliced ftz RNA produced in the splicing reaction. This probe consists of 12 nucleotides from the 3' end of exon 1 joined to 12 nucleotides from the 5' end of exon 2. Therefore, it will label the mRNA preferentially over the pre-mRNA, other splicing intermediates, or the 5' exon product due to the increase in hybridization possible with the mRNA and the radiolabeled oligonucleotide. The junction oligonucleotide hybridized to the mRNA and, weakly, to the pre-mRNA during a splicing time course (Fig. 1A, lanes 2 to 4). Production of the mRNA was dependent on the presence of ATP and incubation at the splicing temperature (lanes 5 and 1), and the product was of the expected size. This assay identifies the mRNA not only by size but by sequence identity and demonstrates the fidelity of the in vitro Drosophila splicing system used in our laboratory.

View larger version (50K): [in this window] [in a new window]   FIG. 1.   B52 functions as an in vitro splicing factor in a homologous Drosophila splicing system. (A) Identification of the ftz mRNA by Northern blotting with an exon junction oligonucleotide as the probe. A splicing time course is shown, with either 0 (lane 1), 30 (lane 2), 60 (lane 3), or 90 (lane 4) min under splicing conditions. A control reaction without ATP (ATP) is also shown (lane 5). The junction oligonucleotide probe identifies the mRNA and also the pre-mRNA (although much more weakly). Even though the signals look comparable, the amount of pre-mRNA on the blot is much larger than the amount of mRNA (see Fig. 2 for an example). (B) The ftz pre-mRNA substrate was uniformly radiolabeled for these reactions. The splicing activities of nuclear extract (NE) and uncomplemented S100 extract () are shown (lanes 1 and 2, respectively). S100 extract was complemented with 0.25 (lane 3), 0.5 (lane 4), or 1 (lane 5) µg of B52 produced in a baculovirus expression system. Diagrams of the splicing products and intermediates are shown on the right, where boxes represent exons and lines represent introns. (C) Western blot prepared by using the RRM antibody, indicating the relative amounts of B52 present in the nuclear extract and S100.

B52 is an essential splicing factor. Antibodies have been useful tools in the study of splicing factors. In the initial characterization of two mammalian SR proteins (9G8 and SC35), antibodies were used to show that these proteins are essential splicing factors (3, 9). These authors showed that titrating increasing amounts of antibody against one of these specific splicing factors decreased splicing activity, presumably because the antibody bound to the corresponding splicing factor and prevented it from functioning. We used the RRM antibody to test whether B52 could be functionally inhibited in a similar fashion. Although the splicing efficiency in this experiment was less than we generally observe, it was clear that splicing activity could be seen in a reaction using nuclear extract and in a control reaction in which a nonrelevant antibody was added (Fig. 2, lanes 1 and 5, respectively). The addition of increasing amounts of antibody resulted in a gradual reduction of both the spliced products and splicing intermediates until splicing inhibition became virtually complete (lanes 2 to 4). The inhibition of splicing intermediates suggests that the antibody inhibits B52 function early in the splicing reaction, as proposed previously for SR proteins (35). The addition of increasing amounts of baculovirus-produced B52 to the antibody-inhibited reaction restored splicing activity (lanes 6 to 8). This recovery of splicing activity indicates that the inhibitory action of the antibody is specific for B52 and that B52, like mammalian SR proteins, plays an important role in the splicing reaction.

View larger version (76K): [in this window] [in a new window]   FIG. 2.   B52 is required for ftz pre-mRNA splicing in vitro. The resultant RNAs from a splicing reaction using uninhibited nuclear extract () are shown (lane 1). Increasing amounts of RRM antibody (B52) were added to splicing reactions (lanes 2 to 4), causing a loss of splicing activity compared to a reaction in which a nonspecific antibody (NS Ab) was used (lane 5). The RRM antibody was added to splicing reactions, as well as increasing amounts of B52 (0.25, 0.5, and 1 µg) (lanes 6 to 8, respectively). The size markers to the left are the 1-kb ladder (Gibco BRL).

B52 is not essential for splicing in whole larvae. Since B52 is a necessary splicing factor when assayed in vitro, we took advantage of the B52 deletion fly strain to analyze the RNA splicing requirement for B52 in the organism. The same ftz pre-mRNA was joined to a promoter that contains a regulatory protein binding site (UAS_gal) for the positive yeast activator protein GAL4 (1) and introduced into the B52 deletion fly line (B52²⁸/TM6b Tb e) by germ line transformation (see Materials and Methods). Crossing the UAS_gal-ftz-containing transformant line with a GAL4-expressing line provided a means of producing the ftz pre-mRNA in the larvae. The B52 deletion, balanced over the dominant balancer (Tm6B), allows the hand selection of larvae containing one copy of the complete B52 gene (for simplicity, these larvae will be called B52⁺). These larvae have a tubby phenotype and are easily separable from the nontubby larvae homozygous for the B52 deletion (B52 larvae) (24).

View larger version (23K): [in this window] [in a new window]   FIG. 3.   Significant splicing of the ftz pre-mRNA occurs in larvae lacking the B52 gene. (A) RT-PCR analysis of RNA isolated from larvae expressing the ftz transgene and containing (ftz, B52+) (lanes 1 and 2) or lacking (ftz, B52) (lanes 3 and 4) a complete B52 gene. Results of a similar analysis for RNA isolated from larvae without the transgene are shown to determine the endogenous ftz levels (lanes 5 to 8). The RNA in lanes 1, 3, 5, and 7 was treated with RNase and then subjected to RT-PCR. A control reaction without RNA (No) (lane 9) and a reaction using a plasmid containing the ftz pre-mRNA to indicate the position of the pre-mRNA (DNA) (lane 10) are also shown. The band below the mRNA is not dependent on the ftz transgene and is not always observed. , no RNase added. (B) Western blot prepared by using the RRM antibody, indicating the amount of B52 present in ftz B52+ (lane 2) and ftz B52 (lane 3) larvae; nuclear extract (NE) (lane 1) is shown for comparison.

Other SR proteins may complement the B52 deletion to retain splicing activity. One simple hypothesis to explain ftz mRNA production in B52 larvae is that in larvae there are other SR proteins present that can complement the B52 deletion for general splicing activity. To test this hypothesis, we prepared SR proteins from hand-selected B52⁺ and B52 larvae. Equal amounts of protein from the two SR protein preparations were titrated in the splicing reaction. Both SR preparations complemented the S100 extract such that splicing of the ftz pre-mRNA occurred in vitro (Fig. 4A, lanes 2 to 7). The B52 SR preparation even showed slightly more splicing activity than the B52⁺ preparation, probably due to subtle differences in the specific activities of the two preparations. Defective splicing activity of the S100 extract was evident (lane 1). Since both SR preparations were functional in this splicing assay, it was necessary to verify that the SR proteins from B52 larvae did not contain B52. Western blotting with the RRM antibody showed that the SR preparation from B52 larvae was devoid of B52 (Fig. 4B; compare lanes 1 and 2). Another Drosophila SR protein (RBP1), prepared in Escherichia coli, also complemented the S100 extract to generate splicing activity (Fig. 4A, lanes 8 and 9). These experiments demonstrate that other Drosophila SR proteins can complement the B52-deficient (S100) extract to splice the ftz pre-mRNA and provide a possible explanation for the presence of ftz mRNA in B52 larvae.

View larger version (38K): [in this window] [in a new window]   FIG. 4.   Larval SR proteins other than B52 rescue ftz splicing in vitro. (A) RNAs from splicing reactions of uncomplemented S100 () (lane 1) or S100 complemented with SR proteins purified from B52+ larvae (200, 400, or 600 ng) (lanes 2, 3, and 4, respectively) or B52 larvae (200, 400, or 600 ng) (lanes 5, 6, and 7, respectively) or RBP1 (250 or 500 ng) (lanes 8 and 9, respectively) are shown. (B) The left panel shows a Western blot of the SR protein preparations (4 µg) from B52+ (lane 1) and B52 (lane 2) larvae, prepared by using the RRM antibody (Ab). The right panel shows a Western blot, prepared using the SR protein-specific antibody mAb104, of nuclear extract (NE) prepared from Kc cells (lane 3), Kc cells (Kc) (lane 4), or proteins from B52+ larvae (Lar) (lane 5). Multiple proteins are present in the B52+ larvae, including different SR proteins and, most likely, their degradation products. The positions of protein molecular mass standards (in kilodaltons) are shown to the right.

B52 is essential for splicing activity in the larval brain. Given that in mouse, calf, and rat tissues, SR proteins are differentially expressed (11, 31, 42), we were prompted to examine whether there is in Drosophila larvae a specific tissue in which B52 is the major SR protein. In such a tissue, the function of B52 can be assayed without the background of other SR proteins, which may complement B52 function. Various tissues from 15 developing Drosophila larvae were dissected, and from them, proteins were prepared and then analyzed by Western blotting with the mAb104 antibody (Fig. 5A). B52 was the predominant SR protein found in the brain and imaginal discs. The other SR proteins identified in whole larvae were present in different tissues. The gut contained large amounts of the 30-kDa protein, fat bodies had the 75-kDa protein, while the 40-kDa protein was prevalent in the cuticle (data not shown). Since B52 is the only SR protein present in detectable quantities in the brain, and the tissue is not difficult to dissect, brain was the tissue chosen to assay the requirement for B52 in ftz pre-mRNA splicing.

View larger version (59K): [in this window] [in a new window]   FIG. 5.   B52 is necessary for the production of ftz mRNA in the B52 larval brain. (A) Fifteen larvae were dissected, like tissues were pooled, and proteins were prepared and separated on an SDS-12.5% polyacrylamide gel. Western analysis with the mAb104 antibody (Ab) shows the SR protein complement of brain (Br), imaginal disc (ID), salivary gland (SG), gut, and fat body (Fat). The mobilities of B52 and protein molecular mass standards (in kilodalton) are indicated. (B) (Left panel) Brain or salivary gland RNA was isolated from single B52+ (+) and B52 () larvae, and RT-PCR was performed. Shown and labeled are the PCR products for rRNA, ftz, and B52 separated on a 4% denaturing gel. The template DNA for all three PCRs came from the same RT reaction. Asterisks indicate nonspecific PCR products. MW, 32P-labeled 1-kb ladder (Gibco-BRL). (Right panel) Either the B52+ or B52 PCR product was digested with PvuII (P), NaeI (N), or SspI (S), and the digestion products were separated on a 8% denaturing gel. The expected sizes (in nucleotides) of the digestion products of ftz mRNA (shown to the right) are 162 and 163 with PvuII, 212 and 113 with NaeI, and 327 (uncut) with SspI. The 113-nucleotide NaeI digestion product was run off the bottom of the gel. (C) Transcription of newly induced Hsp70 mRNA occurs in B52 larvae. RT-PCR was performed with Hsp70-specific primers, using RNA isolated from the brain or salivary glands of B52+ and B52 larvae that were (+) or were not () subjected to heat shock.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Many in vitro studies have shown that SR proteins function in pre-mRNA splicing (8, 35). Although some of the interactions of SR proteins with splicing factors have been defined in recent years, the role of these specific interactions needs further investigation. In addition to functioning as essential splicing factors, SR proteins can regulate alternative splice site choice. In splicing substrates with multiple splice sites, some SR proteins promote proximal splicing while others, or hnRNP A1, shift splicing to distal splice sites (10, 21, 43). Similar effects have been reproduced in tissue culture by overexpressing ASF/SF2 and hnRNP A1 in HeLa cells (2, 36). In addition, studies using the chicken B-cell line DT40 showed the SR protein ASF/SF2 to be essential, since targeted disruption of the gene resulted in a loss of cell viability (37). The depletion of ASF/SF2 in this cell line was shown to affect the splicing of a few specific transcripts (38). However, splicing studies in multicellular organisms have lagged behind these studies in cell culture and certainly those performed in vitro.

In this study, we characterized the splicing activity of B52 to investigate the lethal phenotype of the B52-null mutant. This is the only mutant of an SR family member studied in a whole organism. Using a Drosophila homologous in vitro splicing system, we demonstrated that B52 is an essential splicing factor. Given that finding a natural target gene has proven difficult, we introduced into Drosophila the ftz minigene splicing substrate, which we used successfully in vitro, to assay the splicing activity of B52 in vivo. Other SR proteins appear to complement the lack of B52 for splicing activity in whole B52 larvae. However, we found that B52 was the major SR protein expressed in the brain. In this tissue from B52 larvae, we detected no ftz mRNA, although the mRNA was observed in tissues in which SR proteins other than B52 are expressed. mRNA of the newly induced Hsp70 gene was detected, which demonstrates that RNA transcription functions in the brain of the B52 larva. These results, taken together, show that B52 functions in vivo as a critical splicing factor in tissues in which it is the predominant SR protein. The lethality of the B52 deletion mutant may be a consequence of splicing defects for specific genes in the brain, where other SR proteins cannot complement for the loss of B52.

In a previous study, flies homozygous for a P-element insertion in another Drosophila splicing factor showed severely reduced viability and bristle defects but no obvious splicing defects (30). Analogously, we also did not observe an effect on splicing in the B52 mutant when whole larvae were analyzed. However, since B52 is an essential gene, it has at least one particular function that other SR proteins cannot complement. We screened several genes and found that the B52 larvae were modestly defective (twofold) in the splicing of one of these, the dopa decarboxylase pre-mRNA (B. E. Hoffman and J. T. Lis, unpublished results). A distinct pre-mRNA substrate specificity (3, 7) and different alternative splicing patterns (42, 43) have been found for different SR proteins in mammalian systems. Such assays using synthetic RNA substrates are informative, but the identities of genes requiring a specific SR protein for proper processing remain elusive.

To analyze the issue of SR protein redundancy versus specificity more closely, in vitro selection was performed to identify the RNA binding sequences for two Drosophila SR proteins, RBP1 (12) and B52 (32). The target sequences identified by the two proteins are not similar and presumably identify different genes. A functional selection procedure also resulted in different consensus sequences for three different SR proteins (19). More specifically, the Drosophila doublesex gene contains multiple copies of the RBP1 sequence but no B52 target sequences. A specific endogenous target gene for B52 has not been identified. Binding site analysis of two different mammalian SR proteins agree with the findings for Drosophila and suggest that different SR proteins interact with a set of distinct RNA sequences (34). However, both B52 and RBP1 can activate splicing of the same ftz pre-mRNA. We found that ftz pre-mRNA contains a partial match to the B52 consensus binding site (32), but we have not found an RBP1 consensus binding site (12). Whether one or both of these SR proteins interact with the ftz pre-mRNA directly, possibly at different sites, or function indirectly through other splicing factors is not known. It has been shown that SR proteins can act directly on pre-mRNAs containing their target sequences, since placing the selected binding sites of B52 or ASF/SF2 in a pre-mRNA increases the splicing efficiency of the substrate (34; H. Shi, B. E. Hoffman, and J. T. Lis, unpublished results). Alternatively, SR proteins have been proposed to bridge the gap across introns to facilitate splicing through protein-protein interactions (39).

Understanding the roles of the SR family of proteins will require a battery of in vitro and in vivo studies. While some functions may be shared, others are clearly specific to particular members of the family. The differences in their RNA binding specificity, in the strength of their interactions with other proteins (39), and in their patterns of expression in whole organisms could all contribute to more-specialized roles for individual SR proteins. Finally, while it is clear that SR proteins function as splicing factors in vitro and that their levels influence splicing in vivo, we cannot rule out the possibility of other roles for these proteins.