In trypanosomes all mRNAs are generated through trans mRNA splicing, requiring the functions of the small nuclear RNAs U2, U4 and U6. In the absence of conventional cis mRNA splicing, the structure and function of a U5-analogous snRNP in trypanosomes has remained an open question. In cis splicing, a U5 snRNP-specific protein component called PRP8 in yeast and p220 in man is a highly conserved, essential splicing factor involved in splice-site recognition and selection. We have cloned and sequenced a genomic region from Trypanosoma brucei, that contains a PRP8/p220-homologous gene (p277) coding for a 277 kDa protein. Using an antibody against a C-terminal region of the trypanosomal p277 protein, a small RNA of ∼65 nucleotides could be specifically co-immunoprecipitated that appears to be identical with a U5 RNA (SLA2 RNA) recently identified by Dungan et al. (1996). Based on sedimentation, immunoprecipitation and Western blot analyses we conclude that this RNA is part of a stable ribonucleoprotein (RNP) complex and associated not only with the p277 protein, but also with the common proteins present in the other trans-spliceosomal snRNPs. Together these results demonstrate that a U5-analogous RNP exists in trypanosomes and suggest that basic functions of the U5 snRNP are conserved between cis and trans splicing.
Trans splicing is an essential step in the expression of all protein-coding genes in trypanosomes and involves the addition of a short, non-coding miniexon derived from the spliced leader (SL) RNA onto each protein-coding exon sequence present within long polycistronic precursor transcripts. Besides trypanosomes, this SL-type trans splicing occurs also in nematodes, trematodes and Euglena (reviewed by Agabian, 1990; Nilsen, 1993, 1995). Although discovered in trypanosomes 10 years ago (Murphy et al., 1986; Sutton and Boothroyd, 1986), trans mRNA splicing is still not well understood in comparison to conventional cis mRNA splicing. In analogy to cis splicing, we assume that trans splicing requires the assembly of the two precursor RNAs, the SL RNA and the polycistronic pre-mRNA, into a complex RNA–protein structure, a trans spliceosome. Only a few RNA and protein components have been identified in trypanosomes that are relevant for trans splicing. In particular, U2, U4 and U6 snRNAs are essential co-factors (Tschudi and Ullu, 1990) and occur in the form of snRNPs in the cell (Michaeli et al., 1990; Cross et al., 1991; Günzl et al., 1992). The SL RNP is an additional, trans-splicing specific factor and has been proposed as a U1 snRNP equivalent (Bruzik et al., 1988; discussed in Steitz, 1992); although it contributes its miniexon during the trans-splicing reaction, it also resembles an snRNP on the basis of its protein composition (Palfi et al., 1991). Using affinity purification of Trypanosoma brucei snRNPs, five common low-molecular-weight polypeptides have been identified that are shared by the U2, U4/U6 and SL RNPs; in addition, this analysis revealed several snRNP-specific protein components (Palfi et al., 1991). In contrast to the yeast and mammalian cis-splicing systems, only one snRNP protein gene has been cloned so far from the trypanosome trans-splicing system, coding for a U2-specific 40 kDa protein with homology to the cis-spliceosomal U2 A′ protein (Cross et al., 1993).
An important open question has been whether there is a trans-spliceosomal homologue of the U5 RNA. U5 RNA is the least conserved among the spliceosomal RNAs, and phylogenetic comparisons showed that essentially only the 11-nucleotide 5′ loop sequence is conserved (Guthrie and Patterson, 1988; Frank et al., 1994). The 5′ loop of U5 RNA is functionally important in cis splicing, in particular during 5′ and 3′ splice-site recognition and selection; this has been established in yeast using in vivo mutational studies (Newman and Norman, 1991, 1992) and also in the mammalian splicing system (Wyatt et al., 1992; Cortes et al., 1993; Sontheimer and Steitz, 1993). A recent in vitro snRNP reconstitution study in yeast, however, demonstrated that the U5 loop sequence is not essential for the first step of splicing, but critical for holding the splicing intermediates together (O'Keefe et al., 1996). The latter study pointed to an important role for U5-associated proteins.
Based on the functional correlation between the U5 RNA and the PRP8/p220 protein and the importance of these two components of the cis spliceosome, it would in fact be very surprising if there were no equivalents in the trans spliceosome. Very recently Dungan et al. (1996) reported the RNA sequence of a 62-nucleotide ‘U5-like RNA’ from T.brucei containing the 5′ loop sequence conserved in all known spliceosomal U5 RNAs. Additional evidence for this RNA being the trans-spliceosomal U5 counterpart was derived from crosslinking experiments, showing a contact between the U5 loop sequence and the free SL miniexon trans-splicing intermediate, an interaction analogous to that between the cis-spliceosomal U5 RNA loop and the free first-exon splicing intermediate. In nematodes, where both cis and trans splicing occur, a recent in vitro study demonstrated that U5 RNA is present in cis as well as in trans spliceosomes (Maroney et al., 1996); in trans splicing one role of U5 appears to be to hold onto the free SL exon after the first step of the reaction.
Our approach to address the question of U5 snRNP function was to identify a PRP8/p220 homologous gene in trypanosomes and to search for an associated RNA, which would represent the bona fide U5-analogous RNA in the trans-splicing system. We report here the sequence of a PRP8-homologous gene of T.brucei coding for a 277 kDa protein; using specific antibodies against the trypanosomal protein, we have identified a small RNA of ∼65 nucleotides that appears to be identical to the U5-like RNA identified by Dungan et al. (1996). We demonstrate that the trypanosome U5 RNA is part of a stable ribonucleoprotein (RNP) complex and associated not only with the p277 protein, but also with the common proteins present in the other trans-spliceosomal snRNPs. In sum, a U5-analogous RNP exists in trypanosomes suggesting that basic functions of the U5 snRNP are conserved between cis and trans splicing.
Results and discussion
Genomic cloning and expression of a PRP8-homologous gene (p277) from T.brucei
In a polycistronic transcription unit encoding two catalytic subunits of the T.brucei protein kinase A (T.Klöckner and M.Boshart, manuscript in preparation) an open reading frame homologous to the C-terminal part of the PRP8 gene from Saccharomyces cerevisiae was found by database searches. Using this region as a probe, several phage clones were isolated and mapped from a genomic T.brucei λDASH II library (see Materials and methods). Sequencing of ∼8 kb of a λ clone revealed a long open reading frame coding for a PRP8-homologous protein of 2403 amino acids with a predicted molecular weight of 276 921 and pI of 8.35 (designated p277 in the following; see accession number Y12638 for genomic nucleotide sequence and data not shown). Southern blot hybridization analysis of T.brucei genomic DNA indicated that p277 is encoded by a single copy gene (data not shown). A single transcript of 7.8–8 kb in length was detected at equal abundance in two major stages of the trypanosome life cycle (slender bloodstream forms and procyclic forms; data not shown). The major 3′ splice site (spliced leader acceptor site) was mapped by reverse transcription (RT)–PCR and sequencing of the products to an AG located 112 nucleotides upstream of the putative AUG start codon. This 3′ splice site is preceded by a pyrimidine-rich tract (between positions −7 and −52 relative to the 3′ splice site with 76% pyrimidines). The major polyadenylation sites were mapped by RACE–PCR between 169 and 178 nucleotides downstream of the stop codon. Thus, the calculated length of the transcript without poly(A) tail (7530 bp) is compatible with the observed length on Northern blots, and use of the first ATG downstream of the spliced leader acceptor site is very likely. In addition, Western blot analysis with specific antibodies confirmed that a protein of the expected size is detectable in T.brucei extract (see below and Figure 3).
Comparison of the T.brucei p277 protein sequence with known cis-spliceosomal homologues
Since PRP8 conservation is exceptionally high (e.g. 61% identity between the yeast PRP8 protein and the corresponding C.elegans sequence; see Hodges et al., 1995), we compared the new sequence of the T.brucei p277 protein with the known PRP8 homologues (Figure 1 and data not shown). To maximize the alignment between p277 and PRP8, several gaps had to be introduced into the trypanosome protein sequence (two of them spanning more than 10 amino acids); the sizes of the three proteins are strikingly similar (T.brucei, S.cerevisiae and C.elegans: 2403, 2413 and 2329 amino acids respectively). Significantly, the trypanosomal p277 protein shares only 40% identical and 61% similar amino acids with the yeast PRP8 sequence (T.brucei/C.elegans: 41% identical and 64% similar amino acids). The overall sequence conservation of the T.brucei sequence is more pronounced in the central region of the protein (approximately amino acids 500–2000) than in the N- and C-terminal regions (see Figure 1 for a schematic representation). There is an area with a particularly high degree of identity (62%) between amino acids 535 and 703. Towards the C-terminus the proportion of identical positions strongly declines, from 56% (between amino acids 1551 and 1600) to values of between 20 and 30% at the C-terminal region (amino acids 2100–2403). Yet all three sequences are almost colinear at their carboxy termini. As previously noted (Hodges et al., 1995), there are two extended highly conserved regions in the yeast and C.elegans protein sequences, corresponding to yeast amino acids 1600–1660 (97% identity) and amino acids 1889–1947 (93% identity); significantly, the corresponding regions of the trypanosome p277 protein do not exhibit this high level of homology. In contrast to C.elegans, where both cis and trans splicing occur, trypanosomes had to retain only trans splicing. Therefore the function implied by these conserved regions may have adapted during evolution to the particular requirements of the trans-splicing machinery.
Similarly as for the yeast and C.elegans homologues, no known protein motifs in the trypanosomal p277 sequence could be identified. The only exception is the N-terminal sequence of yeast PRP8, where four runs of proline-rich regions interspersed by acidic residues occur (amino acids 1 to ∼80, Hodges et al., 1995). This region appears to be yeast-specific, since it is missing not only in the C.elegans (Hodges et al., 1995), but also in the T.brucei sequence (this study).
The mutational analysis of the yeast PRP8 protein has revealed two separate domains involved in 3′ splice-site fidelity and in uridine-tract recognition (Umen and Guthrie, 1996; see Figure 1). These two domains are close to each other, and the two highly conserved regions (see above) fall into the same area. First, most of the yeast PRP8 mutations that affected 3′ splice-site fidelity cluster within a small internal region and concern positions that are identical in the C.elegans sequence (Umen and Guthrie, 1996). Significantly, three of these five amino acid positions in yeast are identical in the trypanosome sequence (yeast M1399/T.brucei M1386; yeast T1565/T.brucei T1554; yeast W1609/T.brucei W1598), one is similar (yeast W1575/T.brucei F1564), and only one differs (yeast E1576/T.brucei N1565). A second striking result of the mutational analysis in yeast by Umen and Guthrie (1996) was that a collection of U-tract recognition mutations were all mapped to two positions of yeast PRP8, F1834 and E1960, which were found to be conserved in various species. The former position is not conserved in the trypanosome sequence (yeast F1834→T.brucei G1823); the latter position is replaced by a similar amino acid (yeast E1960→T.brucei D1951).
Taken together, the high overall conservation of the PRP8/p277 protein sequences implies that basic functions of this protein splicing factor are conserved between cis and trans splicing. This argues in favour of the same catalytic mechanism operating in both splicing systems. However, not all of the amino acid positions critical for function of the cis-spliceosomal PRP8 protein are conserved in p277, suggesting that certain differences exist between PRP8 functions in cis and trans splicing. For example, one can envision that the same mechanism that ensures 3′ splice-site fidelity operates in cis and trans splicing, whereas the U-rich region is recognized in a different manner.
The trypanosomal p277 protein is associated with a small RNA
Based on its similarity with PRP8, we expected p277 to be associated with a trans-spliceosomal homologue of the U5 RNA. In order to study RNA and protein interactions of p277 in T.brucei, in particular to identify a putative trans-spliceosomal U5 homologue, we raised polyclonal antibodies against a C-terminal portion of p277 (for a characterization by Western blot analysis, see below). Affinity-purified anti-p277 antibodies and the corresponding non-immune control were used for immunoprecipitation from T.brucei extract; in parallel, immunoprecipitation was done with anti-common protein antiserum, which is specific for five protein components shared by the U2 and U4/U6 snRNPs as well as the SL and SLA RNPs (Palfi and Bindereif, 1992; Palfi et al., 1994). RNA was purified from the immunoprecipitate and labelled with [32P]pCp and RNA ligase. As Figure 2 shows, anti-p277 antiserum specifically precipitated a small RNA (‘RNA X’) migrating as a triplet with a major species of ∼65 nucleotides (see lane anti-p277). The other RNA species in the 80-nucleotide range most likely represent non-specifically precipitated tRNAs also seen with the control non-immune serum. Significantly, an RNA of the same mobility as RNA X was recognized by anti-common protein antibodies (see lane anti-CP), in addition to a collection of RNAs, the major species of which have previously been identified as U2, SL, U4, and U6 (Palfi and Bindereif, 1992). Based on partial RNase T1 digestion patterns of gel-purified material, RNA X co-precipitated by either anti-p277 or anti-CP antibodies appeared to be identical (data not shown). In sum, we conclude that a small RNA of ∼65 nucleotides is specifically associated with both p277 and common proteins; therefore this RNA represents a likely candidate for the U5-analogous trans-spliceosomal snRNA of T.brucei.
Stable association of the T.brucei U5 RNA with p277 protein and common proteins: a trans-spliceosomal U5-analogous RNP
Recently Dungan et al. (1996) identified a novel RNA called SLA2/U5-like RNA as a good candidate for the U5-homologue of T.brucei. Therefore we have tested whether the p277 protein is associated with this RNA, using primer-extension assays and a primer specific for the 3′ terminal region of the SLA2/U5-like RNA (hereafter referred to as U5 RNA).
To characterize the putative trypanosomal U5 RNP, T.brucei S100 extract was fractionated through glycerol gradient sedimentation. RNA was prepared from gradient fractions and analysed by primer extension for U4, U5 and U6 RNAs. In parallel, protein was purified, and the p277 distribution was tested by immunoblotting with affinity-purified p277 antibodies. Figure 3 shows that most of the U5 RNA is concentrated in fractions #7–11, corresponding to the 20S region. The distribution of U4 and U6 RNAs across fractions #5–9 (corresponding to the 16S region) overlaps with the U5 region, is slightly shifted towards the top of the gradient, and reflects the occurrence of both U4/U6 snRNP and free U6 RNP. As the Western blot analysis shows, p277 was detected above the 220 kDa marker protein as the only protein band, with a peak in fraction #9. This fractionation of p277 reflects the distribution of the U5 RNA at ∼20S.
To assess whether U5 RNA and p277 are associated with each other, fractions containing U4, U5, and U6 RNAs were pooled, followed by anti-p277 and anti-common protein immunoprecipitations and U4, U5, and U6 RNA primer extensions. In parallel, non-immune serum was used in control reactions. Figure 4 (gradient fraction lanes) clearly shows that the U5 RNA could be efficiently and specifically immunoprecipitated with anti-p277 antibodies; however, neither U4 nor U6 RNA could be detected in the anti-p277 immunoprecipitate. This confirms that the anti-p277 immunoprecipitated RNA species (‘RNA X’) we have detected by 3′ end-labelling (see above and Figure 2) and by primer extension with a U5 RNA-specific oligonucleotide (Figure 4) are identical. Using anti-common protein antiserum, U5 RNA was also efficiently immunoprecipitated, as well as U4 and U6 RNAs. Taken together, these data establish that the trypanosomal U5 RNA is stably and specifically associated with both p277 protein and common proteins in an RNP representing the trans-spliceosomal U5 snRNP. Somewhat surprisingly U4 and U6 RNAs were not found in this complex; therefore under our conditions the major fraction of U5 RNA (at least 95%) appears not to be present in a U4/U5/U6 triple-snRNP form. Since we might have missed a small fraction of U5 in the triple-snRNP form sedimenting outside of the pooled region of the gradient, the same set of immunoprecipitations was carried out with unfractionated T.brucei S100 extract, yielding identical results as with the pooled gradient fractions (Figure 4, S100 extract lanes).
Core particles of snRNPs are characteristically stable during isopycnic centrifugation in CsCl (Lelay-Taha et al., 1986), as shown previously for the RNPs of SL, U2, U4/U6, and SLA RNAs (Michaeli et al., 1990; Cross et al., 1991; Günzl et al., 1992; Palfi et al., 1994). To characterize the U5 RNP of T.brucei further, S100 extract was fractionated through a CsCl density gradient, RNA from the fractions was prepared, and the distribution of U4, U5, and U6 RNAs was analysed by primer extension. As Figure 5A shows, U5 core snRNPs are concentrated in fractions #3–6 and are clearly separated from U4/U6 core snRNPs (fractions #8–10). The peak fractions of the U5 core RNP (#4) and the U4/U6 core snRNP (#9) correspond to densities of 1.36 g/ml and 1.51 g/ml respectively; based on an empirical formula (Spirin, 1969), the RNA–protein mass ratio of the U5 core snRNP was calculated to be 18:82, corresponding to a total protein mass of ∼93 kDa per 62-nucleotide U5 RNA. For the U4/U6 core snRNP, an RNA–protein mass ratio of 43:57 would reflect a total protein mass of ∼89 kDa per U4/U6 RNA molecule. The close correspondence between the total mass of the U4/U6 and U5 core protein components (90 kDa) suggests that they share the same set of proteins; however, since the common proteins we have identified previously add up to a total mass of 60 kDa (Palfi et al., 1991), either their stoichiometry is not equimolar, or there are more, unidentified core protein components.
To identify the protein components of the trypanosomal U5 core snRNP immunoprecipitations were carried out (Figure 5B). Pooled CsCl gradient fractions containing U5 core snRNPs were dialysed and reacted with anti-p277 and anti-common protein antibodies, and—as a control for specificity—with non-immune serum. Precipitates and supernatants of the immunoreaction were then analysed by primer extension for their U5 RNA content. The U5 core snRNP was efficiently precipitated by anti-common protein antibodies, but not detectably by anti-p277 antibodies, demonstrating clearly that common proteins, but not the U5-specific p277 protein, are present in the U5 core snRNP.
Comparison between cis and trans-splicing: functional role and interactions of the U5 snRNP
The U5-specific p277 protein of T.brucei represents only the second protein factor identified and cloned in the SL-type trans-splicing system from trypanosomatids; the first case was the U2-specific 40K protein of T.brucei, the N-terminal half of which is homologous to the cis-spliceosomal U2 A′ protein (Cross et al., 1993). In contrast, homology between the trans-spliceosomal p277 protein and its cis-spliceosomal counterparts extends throughout their entire length, although the level of homology changes considerably across the length of the protein (see Figure 1). It is of interest that not all functionally important positions, which were identified in the yeast PRP8 protein and which are conserved in the other cis-spliceosomal homologues, occur in the trypanosome p277 sequence. However, there is no evidence at present for an additional, potentially trans-splicing specific domain in the p277 protein. Since all known PRP8 homologues lack any recognizable protein motifs (with the exception of the yeast PRP8-specific proline-rich repeats), a mutational analysis of the trypanosomal p277 protein should help to define the domain structure conserved between cis- and trans-splicing. Finally, because of the large size of this protein factor it might be a key component to identify other interacting protein and snRNA elements in the trans spliceosome.
Based on the results of this study and that of Dungan et al. (1996) it was surprising to find that conservation of the U5 snRNP is quite different on the protein and RNA level: whereas with the exception of the 5′ loop the trypanosomal U5 RNA sequence greatly differs in its overall secondary structure and sequence from all the cis-spliceosomal U5 RNAs, at least one protein component (PRP8/p220/p277) is highly homologous. This indicates that important functional properties reside in the protein components of the splicing machinery and are conserved between cis and trans splicing.
One of the most important open questions in the trans-splicing field is how both splice sites on the two separate precursor RNAs are recognized and how they interact stably with each other. Although it is now established that a U5-analogous RNP exists in the trypanosome system, we still know only very little about how it acts during trans-spliceosome assembly and trans splicing. Since a model has been proposed that incorporates both U1 and U5 RNA functions into the trypanosomal SL RNA (Steitz, 1992), the question remains how the U5 RNA—in conjunction with its associated p277 protein—functions in splice-site recognition. In cis splicing U5 enters the spliceosome in the form of the U4/U5/U6 triple-snRNP; our analysis in the trypanosome trans-splicing system gave no evidence for such a stable snRNP complex in significant abundance, although Dungan et al. (1996) showed evidence for a small fraction of U5 existing together with U4 and U6 snRNAs as a triple-snRNP. These results do not rule out that U5 forms in vivo—in analogy to the cis-spliceosomal complex—a triple-snRNP that functions as a precursor in trans-spliceosome assembly, yet is not stable under extract conditions.
In the absence of an in vitro system derived from trypanosomes it will be particularly important to establish functional in vivo systems to carry out mutational studies of the p277 protein and the U5 RNA components. A detailed comparison between results from such studies and those from cis-splicing systems as well as the use of heterologous systems promises to provide novel insights not only into the pathway of trans-spliceosome assembly and the basic trans-splicing mechanism, but also into the cis- and trans-specific characteristics of these two important spliceosome components.
The C-terminus of p277 was identified by sequencing the genomic clone pC3-1, which harbours a protein kinase A catalytic subunit gene of T.brucei (T.Klöckner and M.Boshart, manuscript in preparation). A genomic T.brucei (strain AnTat 1.1) library of partially digested Sau3A fragments (R.Krämer and M.Boshart, unpublished results) constructed in λDASH II (Stratagene) was screened with a riboprobe covering amino acid positions L2017–I2323 of p277. Four phage clones were isolated and mapped, and λ#6 carrying a 12.4 kb insert with the entire p277 reading frame was selected for subcloning and construction of nested deletions (Pharmacia). Both strands of the region covering the p277 reading frame were sequenced by automated cycle sequencing. Nucleotide sequences were compiled and analysed using the GCG package (version 8); protein sequence alignment and statistical analysis were performed using the GAP program (Devereux et al., 1984).
Northern blot and RACE analysis
Slender bloodstream forms of T.brucei strain AnTat 1.1 were purified from infected rat blood. Procyclic forms were differentiated in vitro from bloodstream forms as described (Vassella and Boshart, 1996). Total RNA was isolated using the guanidinium isothiocyanate/CsCl cushion method (Ausubel et al., 1987) and 5 μg were separated on a 1% agarose–formaldehyde gel, blotted, and hybridized as described in Boshart et al. (1991) with a riboprobe covering amino acid positions L2017–I2323. The polyadenylation site was mapped by RACE techniques using the Marathon cDNA amplification kit (Clontech), poly(A)+-RNA from procyclic forms and the p277-specific primer 5′-GAGGCTGTGTCTGTGCGTCTG-3′ (corresponding to amino acid positions E2040–L2046). Amplification products were cloned using the internal SalI site at amino acid position S2330 and sequenced. The 5′-end was mapped by RACE techniques, using a p277-specific primer (5′-GGGTACTTCATCCTTTTG-3′; corresponding to amino acids Q58 to P64) and an SL-specific primer, 5′-AACTAACGCTATTATTAGAA-3′ (nucleotides 1–20 of the T.brucei SL RNA), followed by direct sequencing of the PCR product.
Antibodies and bacterial expression of proteins
The C-terminal region of p277 from position P2019 to T2331 was PCR-amplified and cloned in BamHI/SalI cut bacterial expression vectors pGEX4T3 (Pharmacia) and pQE30 (Qiagen) via a primer-derived BamHI and the genomic SalI site at amino acid position S2330. Fusion proteins were expressed using the protocols provided by the suppliers and extracted from the insoluble fraction with 8 M urea followed by affinity purification on Ni-NTA columns (Qiagen) or by SDS–PAGE and blotting onto nitrocellulose. Rabbits were immunized with powdered or DMSO-dissolved nitrocellulose strips (200 μg fusion protein) using either Freund's adjuvant or Titermax™ (Vaxcel, Inc.) and boosted four times at 4–5 week intervals. Sera and preimmune controls were affinity purified on nitrocellulose strips (Pringle et al., 1991) using 200 μg His-tagged fusion protein per ml of serum.
Glycerol gradient centrifugation
One millilitre of T.brucei S100 extract was mixed with 0.5 ml of buffer G (20 mM KCl, 20 mM HEPES pH 8.0, 1 mM MgCl2) and fractionated through an 11 ml glycerol gradient in buffer G (10–30% glycerol; SW 40 rotor, 6 h, 40 000 r.p.m., 4°C). Fractions of 0.5 ml were collected from the top of the gradient. RNA was prepared from 100 μl aliquots of every second fraction and assayed for U4, U5 and U6 RNA by primer extension with DNA oligonucleotides complementary to nucleotides 50–62 of U5 RNA (Dungan et al., 1996), 53–72 of U4 RNA and 40–59 of U6 RNA (Cross et al., 1991). For primer extension reactions, RNA and primers were co-precipitated with ethanol and dissolved in 1× Expand™ reverse transcriptase buffer containing 0.5 mM of each dNTP. Following denaturation for 10 min at 65°C and cooling on ice, the reactions were supplemented with RNasin (0.5 U/μl), 10 mM DTT and 25 U of Expand™ reverse transcriptase (Boehringer Mannheim). Incubation was for 10 min at 30°C followed by 45 min at 42°C. Primer-extension products were ethanol precipitated and fractionated by denaturing polyacrylamide gel electrophoresis (10%).
CsCl density gradient centrifugation
CsCl was dissolved in buffer D (Dignam et al., 1983) containing 15 mM MgCl2 to a density of 1.48 g/ml. Two hundred microlitres of T.brucei S100 extract were fractionated through a 1 ml gradient (centrifugation in Beckman TL 100 tabletop ultracentrifuge; TLA-100.2 rotor; 90 000 r.p.m. for 20 h at 4°C). Fractions of 100 μl were collected from the top of the gradient. The density distribution across the gradient was measured in a gradient run in parallel to range from 1.31 g/ml (fraction #1) to 1.54 g/ml (fraction #10). RNA prepared from 50 μl of each fraction was assayed by primer extension with U4, U5 and U6 RNA-specific primers. U5 RNA peak fractions were pooled, dialysed against buffer D supplemented with 1 mM DTT and 1 mM PMSF, and used for immunoprecipitation reactions.
Immunoprecipitation and primer-extension analysis
Fifty microlitres of a rabbit serum containing antibodies against the common snRNP proteins of T.brucei ('anti-CP′; Palfi and Bindereif, 1992) and the corresponding non-immune serum were coupled to 25 μl of pre-swollen protein A–Sepharose beads (Pharmacia) in 400 μl NET-150 buffer [50 mM Tris–HCl, pH 8.0, 150 mM NaCl, 0.01% Nonidet P-40 (v/v), 0.5 mM DTT] at 4°C overnight. For anti-p277 immunoprecipitations affinity-purified rabbit anti-p277 antibodies were used. The pre-coated beads were washed four times with 0.5 ml of NET-150 each and resuspended in 400 μl of the same buffer. One hundred microlitres of T.brucei S100 or gradient fraction respectively, were incubated with the protein A–Sepharose bound antibodies for 2 h at 4°C; then the beads were washed extensively with NET-150 buffer. RNA was released from the beads by phenol/chloroform extractions and precipitated from the aqueous phase with ethanol, followed by 3′ end-labelling with 10 μCi [32P]pCp and T4 RNA ligase, as described by England and Uhlenbeck (1978), and analysed by denaturing polyacrylamide gel electrophoresis. For primer extension analysis RNA was purified from both pellet and supernatant of the immunoprecipitation reactions as well as from 10% of the input and assayed with U4, U5 and U6 RNA-specific primers.
Protein was prepared from glycerol gradient fractions (300 μl of every second fraction) by phenolization and acetone precipitation, separated on an 8.5% SDS–polyacrylamide gel, and transferred electrophoretically to Hybond ECL nitrocellulose membrane (Amersham). Blots were blocked in PBS, pH 8.0, containing 1× Roti-Block solution (Roth) and 0.2% Tween-20 overnight at 4°C, followed by probing for 4 h with anti-p277 antibodies (diluted 1:250 in blocking solution). Blots were washed in 1× PBS containing 0.2% Tween-20. Immunocomplexes were detected with sheep anti-rabbit IgG-horseradish peroxidase conjugate diluted 1:20 000 in blocking solution using ECL developing reagents from Amersham.
We thank Andreas Langer and Shahri Raasi for help with the construction of deletion derivatives and acknowledge Miriam van den Bogaard for technical assistance. This work was supported by the Deutsche Forschungsgemeinschaft (A.B.) and German Federal Ministry for Science and Technology (BMFT; grant BEO 21/0316211A; M.B.).