Biochemical characterization of a U6 small nuclear RNA-specific terminal uridylyltransferase Ralf Trippe, Holger Richly* and Bernd-Joachim Benecke Department of Biochemistry, Ruhr University Bochum, Germany The HeLa cell terminal uridylyltransferase (TUTase) that specifically modifies the 3¢-end of mammalian U6 small nuclear RNA (snRNA) was characterized with respect to ionic dependence and substrate requirements. Optimal enzyme activity was obtained at moderate ionic strength (60 m M KCl) and depended on the presence of 5 m M MgCl 2 . In vitro synthesized U6 snRNA without a 3¢-terminal UMP residue was not accepted as substrate. In contrast, U6 snRNA molecules containing one, two or three 3¢-terminal UMP residues were filled up efficiently, generating the 3¢-terminal structure with four UMP residues observed in newly transcribed cellular U6 snRNA. In this reaction, the addition of more than one UMP nucleotide depended on higher UTP concentrations. The analysis of internally mutated U6 snRNA revealed that the fill-in reaction by the U6-TUTase was not controlled by opposite-strand nucleo- tides, excluding an RNA-dependent RNA polymerase mechanism. Furthermore, electrophoretic mobility-shift analyses showed that the U6-TUTase was able to form stable complexes with the U6 snRNA in vitro. On the basis of these findings, a protocol was developed for affinity purifi- cation of the enzyme. In agreement with indirect labeling results, PAGE of a largely purified enzyme revealed an apparent molecular mass of 115 kDa for the U6-TUTase. Keywords:3¢ uridylation; affinity chromatography; terminal uridylyltransferase; U6 snRNA. Nuclear pre-mRNA splicing is a process by which introns are removed from primary transcripts. This two-step trans- esterification mechanism is performed by the spliceosome, an RNP complex of five small nuclear RNA molecules (snRNA U1, U2, U4, U5 and U6), and more than 60 proteins (reviewed in refs [1,2]). Spliceosomes are newly formed on each intron in a well-defined manner. Initially, the 5¢ splice site of the pre-mRNA is recognized by the U1 snRNP. Then, U2 snRNP binds to the branchpoint sequence located near the 3¢-end of the intron. Finally, the spliceosome is completed by incorporation of the U4/U6/ U5 tri-snRNP. The splicing reaction requires extensive rearrangements of RNA–RNA interactions, including the unwinding of the U4/U6 snRNA duplex and the formation of a U2/U6/pre-mRNA structure [3–5]. The precise mech- anism by which proteins control these RNA–RNA inter- actions within the catalytic spliceosome remains to be elucidated. However, evidence exists that intermediate structural variants of U6 snRNA are involved in RNA– RNA rearrangements that take place in the center of the spliceosome [6]. U6 snRNA differs from the other spliceosomal snRNAs in several ways. Unlike other snRNAs, it is transcribed by RNA polymerase III [7,8] and also has a different cap structure [9]. Furthermore, U6 snRNA was unusually well conserved during the evolution of eukaryotes [10] and has no binding site for Sm proteins [11,12]. Instead, U6 snRNPs contain LSm-proteins (ÔlikeÕ Sm) which seem to recognize the 3¢-end oligouridylic structure of the U6 snRNA and which are thought to be involved in U4/U6 snRNP formation [13–16]. Finally, U6 snRNA molecules have a remarkable heterogeneity, resulting from extensive post- transcriptional modification of their respective 3¢-termini. Most ( 90%) cellular U6 snRNA molecules are blocked by a cyclic 2¢,3¢-phosphate (> p) 3¢-end group, and  10% of the U6 snRNA has been found to contain 3¢-oligo(U) stretches up to 20 nucleotides long [17–19]. Recently, two highly specific U6 snRNA-modifying enzyme activities have been identified: a 3¢-exonuclease [20] and a terminal uridylyltransferase (TUTase) [21]. Both enzymes exclusively accept the 3¢-terminus of U6 snRNA as substrate for the addition or deletion of UMP residues. The functional significance of these structural variants of U6 snRNA molecules remains to be elucidated. It is conceivable, however, that all of these modifications together form a cyclic process of regeneration of U6 snRNA, which in turn may be essential for the assembly and catalytic function of spliceosomes. In this report, we present a detailed characterization of the reaction catalyzed by U6-TUTase in vitro,withitshigh selectivity for distinct structural requirements of U6 sub- strate RNA. This study also includes an analysis of the U6 snRNA–TUTase interaction in vitro, which provided the basis for further purification to near-homogeneity of this low-abundancy protein. Correspondence to B J. Benecke, Department of Biochemistry NC6, Ruhr-University, D-44780 Bochum, Germany. Fax: + 49 234321 4034, Tel.: + 49 234322 4233, E-mail: bernd.benecke@ruhr-uni-bochum.de Abbreviations: TUTase, terminal uridylyltransferase; snRNA, small nuclear RNA; EMSA, electrophoretic mobility-shift analysis. *Present address: Department of Molecular Cell Biology, Max Planck Institute of Biochemistry, D-82152 Martinsried, Germany. (Received 12 November 2002, revised 13 January 2003, accepted 17 January 2003) Eur. J. Biochem. 270, 971–980 (2003) Ó FEBS 2003 doi:10.1046/j.1432-1033.2003.03466.x Materials and methods Protein fractionation Cytoplasmic S100 extracts (15 mgÆmL )1 protein) were isolated from HeLa cells as described [22]. For purification of the U6-TUTase, 200 mg extract proteins were separated in a 40-mL Q-Sepharose FF (Pharmacia) column and fractions obtained by step elution. The QS3 fraction (200– 400 m M KCl) was dialyzed against phosphate buffer [25 m M K 2 HPO 4 /KH 2 PO 4 (pH 7.9); 0.2 m M EDTA; 20% glycerol (v/v)] and applied to a 10-mL hydroxyapatite (Bio- Rad) column. The HA2 fraction was obtained by step elution with 150 m M potassium phosphate in the buffer system described above and was further fractionated in a 26/60 Superdex G200 column (Pharmacia). For affinity chromatography, 3 lgoligoA/U6-3RNAwereheatedto 70 °C, cooled down slowly, and incubated for 20 min at 4 °C with 100 mg washed oligo(dT)–cellulose (Roche Molecular Biochemicals) in buffer D [20 m M Hepes/KOH (pH 7.9); 100 m M KCl; 5 m M MgCl 2 ;0.2m M EDTA; 2m M dithiothreitol, 20% glycerol (v/v)]. This suspension waspackedintoacolumn,loadedwith10mLfraction HA2, washed with buffer D, and eluted with the same buffer containing 600 m M KCl. Electrophoresis of proteins in denaturing 7.5% polyacrylamide gels in the presence of SDS was essentially as described by Laemmli [23]. Templates Construction of the U6-3, U6-4 and U6-5 cDNA templates under control of the bacteriophage T7 promoter has been described in detail previously [21]. Based on the U6-5 construct, U6-0, U6-1 and U6-2 cDNA templates were generated by PCR with 5¢-TTTAATACGACTCACTAT AGGGTGCTCGCTTCGGCA-3¢ as upstream primer and 5¢-TATGGAACGCTTCACGAATT-3¢ (U6-0), 5¢-ATAT GGAACGCTTCACGAATT-3¢ (U6-1) or 5¢-AATATGG AACGCTTCACGAATT-3¢ (U6-2) as downstream primer, respectively. PCR fragments were purified by agarose gel electrophoresis and used for in vitro transcription by T7 RNA polymerase as described previously [24]. DNA templates with internal mutations (U6-2/C and U6-1/A) were obtained as follows: in a first PCR, U6 snRNA-coding sequences were amplified with a 5¢-primer carrying the desired mutation and a wild-type 3¢-primer. These frag- ments were used in a second PCR with the same 3¢-primer but a longer 5¢-primer carrying the T7 promoter as upstream flanking sequence element. The amplified PCR fragments were cloned blunt-end into the EcoRV site of the Bluescript KS + vector (Stratagene). PCR fragments suit- able for in vitro transcription and carrying different numbers of 3¢-end UMP residues (i.e. U6-1 and U6-2) were then obtained with the T7 promoter-specific upstream primer and one of the downstream primers described above. The oligo(A)/U6-3 RNA construct was cloned from the previ- ously described U6-3 RNA gene [21]. From this gene, an AspHI (+5 of U6-DNA)–HindIII (downstream of coding sequence) fragment was recovered. Two synthetic oligonu- cleotides were hybridized and gave rise to a double-stranded element containing the oligoA (20) flanked by an upstream SacII overhang and the first 5 bp of the U6-coding sequence, in the form of a downstream AspHI overhang. These two DNA fragments were inserted simultaneously into the KS + vector, restricted with SacII and HindIII and purified by agarose gel electrophoresis before use. After linearization of this plasmid with the DraIenzyme,T7RNA polymerase transcription from the neighboring KS + pro- moter yielded suffciently large quantities of oligo(A)/U6-3 RNA for coupling to oligo(dT)–cellulose. Primary struc- tures of all constructs were confirmed by sequencing. Assay of TUTase Standard TUTase reaction assays were performed in buffer containing 60 m M KCl, 12 m M Hepes/KOH (pH 7.9), 5m M MgCl 2 ,2m M dithiothreitol, 0.1 m M EDTA, 12% (v/v) glycerol and 5 lCi [a- 32 P]UTP, in a total volume of 50 lL. Substrate RNAs were either 1 lg total cellular RNA or 50 ng in vitro synthesized U6 snRNA, and incubations were performed for 60 min at 30 °C. Phenol-extracted RNAs were analysed in 6% polyacrylamide gels in TEB buffer [90 m M Tris base, 90 m M boric acid, 2 m M EDTA (pH 8.3)] containing 6 M urea. Electrophoresis was for 50 min at constant 30 W. To discriminate between RNA molecules differing in length by only 1 nucleotide, high- resolution gels were used in some experiments. These gels contained 7% polyacrylamide, were twice as long (50 cm), and were run for 210 min at constant 37.5 W. Autoradio- graphy of the dried gels was at )70 °C for 16 h, using a Cronex intensifier screen. Electrophoretic mobility-shift assay (EMSA) ForEMSAs,U6-3RNAsweresynthesizedin vitro with T7 RNA polymerase (Fermentas), in the presence of [a- 32 P]UTP (800 CiÆmmol )1 ; New England Nuclear). Labe- led U6-3 RNA (100 000 c.p.m.) was incubated with Super- dex G200 fractions for 10 min at 4 °Cin12m M Hepes/ KOH (pH 7.9), containing 60 m M KCl, 5 m M MgCl 2 , 2m M dithiothreitol, 0.1 m M EDTA, 1 lgyeasttRNA (Roche Molecular Diagnostics) and 12% (v/v) glycerol. Electrophoresis in nondenaturing 6% polyacrylamide gels with 0.25 · TEB buffer was at 6 VÆcm )1 for 4 h at 4 °C. Indirect labeling analysis Affinity-purified TUTase was incubated with labeled U6-3 RNA (100 000 c.p.m.) as described for shift analyses. UV cross-linking (Hoefer UVC 500) was at 0.3 JÆcm )2 Æmin )1 for 10 min on ice. Subsequently, half of the reaction mixture was incubated with 10 lg RNase A for 10 min at room temperature. After acetone precipitation, proteins from both fractions were analysed in 7.5% polyacrylamide/SDS Laemmli gels. Results Ionic requirements of HeLa cell TUTases Earlier experiments had established that, aside from an unspecific enzyme, HeLa cells contain a highly specific TUTase that exclusively modifies the 3¢-ends of U6 snRNA molecules. To separate this activity from the unspecific 972 R. Trippe et al.(Eur. J. Biochem. 270) Ó FEBS 2003 enzyme and also to remove contaminating RNases, a protocol had been developed for the preparation of partially purified enzyme fractions [21]. This purification started with phosphocellulose P11 chromatography of HeLa cell S100 extracts. Subsequently, the unspecific TUTase was separ- ated from the U6-TUTase by gel filtration in Superdex G200. These two crude enzyme fractions were analysed in parallel for magnesium and salt (KCl) dependence of their respective transferase reactions, with 1 lgtotalRNAfrom HeLa cells added as substrate. As shown in Fig. 1A,B, the two activities separated by gel filtration were clearly distinguishable by their acceptance of substrate RNA. In agreement with our previous results [21], the unspecific enzyme (unspec; left panels of Fig. 1A,B) modified a variety of cellular RNA molecules, whereas the U6-specific TUTase (spec; right panels) exclusively accepted U6 snRNA as substrate. However, the two enzymes revealed similar salt optima. As seen in Fig. 1A, both activities showed a clear optimum at 5 m M MgCl 2 (lanes 2 and 6), with no activity detectable in the absence of Mg 2+ ions (lanes 4 and 8). Corresponding results for the two enzymes were also observed with respect to ionic strength (Fig. 1B). With both enzymes, optimal conditions for the transfer of UMP residues were obtained in the presence of 60 m M KCl (Fig. 1B, lanes 1 and 6). This is the amount of KCl already provided by the protein fractions, with no further salt added to the reaction. It should be noted that lower amounts of KCl (40 m M ), obtained after dialysis of protein fractions, did not further increase the enzyme activity (not shown). However, both enzymes were inhibited significantly by higher salt concentrations (150 m M KCl; lanes 3 and 8) and showed no detectable activity at and above 200 m M KCl (lanes 4 and 5 and 9 and 10). Therefore, although slight differences between the two enzymes were observed in response to higher salt conditions, these two TUTase activities basically depended on similar reaction conditions, but with clearly different substrate requirements. These results on Mg 2+ and KCl dependence were primarily required to establish the affinity-purification protocol for the U6-TUTase, described below. Therefore, we did not analyse in more detail the suitability of other bivalent cations and/or salts such as manganese or ammonium sulfate. Restoration of authentic 3¢-ends by the U6 snRNA-specific TUTase Next, we wanted to determine RNA substrate requirements of the U6-TUTase. For this, a U6-TUTase fraction was used that was prepurified by Q-Sepharose and hydroxy- apatite chromatography, followed by gel filtration in Superdex G200 (see Materials and methods section). U6 snRNA molecules were synthesized as substrates that differed with respect to the number of UMP residues at their 3¢-ends respectively. After run-off transcription by T7 RNA polymerase of mutant U6 genes, in vitro synthesized U6 snRNA molecules were obtained that contained either no (U6-0) or up to five (U6-5) 3¢-terminal UMP residues. To ensure homogeneity of the RNA molecules applied, in vitro synthesized RNA was first separated in high-resolution polyacrylamide gels, and transcripts of the correct length were recovered, before their use as substrate in a standard TUTase reaction with [a- 32 P]UTP. Subsequently, labeled U6 snRNA molecules were analysed again in high-resolu- tion gels. In this analysis, U6-3 marker RNA, labeled by T7 RNA polymerase transcription of the corresponding tem- plate, was included as a size standard (ÔmÕ in Fig. 2A). From the results presented in Fig. 2A it is evident that U6-1, U6-2 and U6-3 RNA molecules (lanes 2–4) were efficient substrates for the U6-TUTase reaction. Furthermore, the high-resolution capacity of our gel system allowed the resolution of closely related molecules, differing in length by only one nucleotide. A close examination of the bands in lanes 2–4 of Fig. 2A revealed that concomitant with increasing length of the UMP tails of the substrate RNA, a decreasing number of distinguishable labeled RNA products was obtained: three bands in lane 2, two in Fig. 1. Ionic requirements of HeLa cell TUTases. The U6 snRNA- specific (spec.) and the unspecific (unspec.) TUTase of HeLa cell extracts were separated in a Superdex G200 column [21] and analysed with 1 lg total cellular RNA as substrate. (A) Peak fractions of the two activities were tested in the presence of various concentrations of MgCl 2 (Mg ++ ): 10 m M (lanes 1,5), 5 m M (lanes 2,6), 2.5 m M (lanes 3,7) and 0.0 m M MgCl 2 (lanes 4,8). (B) Same analysis as in (A), except that standard TUTase reactions (5 m M MgCl 2 ) were performed in the presence of increasing amounts of KCl: 60 m M (lanes 1,6), 110 m M (lanes 2,7), 150 m M (lanes 3,8), 200 m M (lanes 4,9) and 250 m M KCl (lanes 5,10). Analysis of labeled RNA products was in 6% poly- acrylamide gels containing 6 M urea. Exposure of the dried gels to Kodak X-ray films was for 16 h using a Cronex intensifier screen. The position of labeled U6 snRNA is indicated by an arrow and ÔmÕ represents labeled marker DNA. Ó FEBS 2003 Characterization of U6 terminal uridylyltransferase (Eur. J. Biochem. 270) 973 lane 3, and a single band in lane 4. In comparison with the labeled U6-3 marker RNA, these results indicate that the U6-TUTase has a clear preference to fill up the 3¢-ends of U6 snRNA molecules to the four UMP residues found in newly transcribed cellular U6 snRNA. The 3¢-ends were not elongated further, as evidenced by the absence of labeled products observed with U6-4 (lane 5) and U6-5 (lane 6) as substrate RNA. Furthermore, the U6-TUTase did not accept U6-0 RNA (lane 1) as substrate, indicating that at least one pre-existing UMP residue at the 3¢-end of U6 snRNA is a prerequisite for this modification reaction to take place. In addition, the finding of intermediates of the transferase reaction (Fig. 2A, lanes 2 and 3) seemed to point to dependence on the nucleotide concentration of the chain elongation rate. Therefore, TUTase reactions were per- formed as before with U6-1 RNA and [a- 32 P]UTP, but this time in the presence of increasing amounts of unlabeled UTP (Fig. 2B). Again, U6-3 RNA labeled by T7 transcrip- tion in vitro was included as size marker (m). The comparison in Fig. 2B of the standard reaction products, i.e. in the presence of labeled UTP only (lane 1), with those obtained in the presence of increasing amounts of unlabeled UTP (lanes 2–5) shows a clear shift to the full-length modification product (U6-4 RNA, lane 5), at the expense of the intermediate bands seen in lanes 2–4 of Fig. 2B. It should be noted that the significant reduction in overall signal intensity (lanes 3–5) is simply due to the competition for incorporation of labeled nucleotides by the excess of unlabeled UTP. Together, we conclude that the U6 snRNA-specific TUTase reaction depends on both the 3¢-end structure of the template RNA and the concentration of substrate nucleotides present in the reaction mixture. Structural requirements for the U6 substrate RNA The finding that the U6-TUTase preferentially fills in the 3¢-end of U6 snRNA only to four UMP residues raises the question of how this elongation reaction is controlled. This is even more intriguing because the proposed secondary structure of U6 snRNA (Fig. 3A [25]) has an extended 3¢–stem–loop structure with exactly four internal AMP nucleotides (+27/+30) opposing the four 3¢-terminal UMP residues (+103/106) found in newly transcribed cellular U6 snRNA. Therefore, it is tempting to speculate that the U6-TUTase may be guided by the sequence of the opposite strand, establishing some sort of substrate-specific RNA- dependent RNA polymerase reaction. To test this hypo- thesis, we analysed two U6 mutant genes. One (U6-2/C) consisted of a U6-2 clone that contained one additional CMP nucleotide, inserted between positions +27/+28 (two of the four AMP residues mentioned above). The corres- ponding mutant U6-2/C substrate RNA was tested either in a standard TUTase reaction (with labeled UTP) or in the presence of 32 P-labeled GTP, supplemented with unlabeled UTP. The analysis of the modified RNA is shown in Fig. 3B, with unmodified U6-2/C RNA, labeled during T7 transcription, as size marker (lanes 1 and 4). The result shown in lane 2 of Fig. 3B confirmed that, in a standard TUTase reaction, U6-2/C RNA still served as an efficient substrate for the transferase reaction. Surprisingly, however, the TUTase reaction stopped after the addition of one UMP residue and did not fill-up this U6 mutant RNA to the four UMP residues observed previously (compare with lane 3 of Fig. 2A). This is evident from the size comparison between the modified U6-2/C RNA (Fig. 3B, lane 2) and the unmodified control RNA (lanes 1 and 4). In contrast, the TUTase reaction performed with the same substrate RNA und unlabeled UTP, but in the presence of labeled GTP as tracer, did not give rise to any labeled RNA product (Fig. 3B, lane 3). This indicates that the newly introduced CMP nucleotide at position +28 of the mutant gene was not functional as a Ôtemplate-strandÕ nucleotide, able to direct the incorporation of GMP residues into the opposite 3¢-end of the mutant RNA. A second mutation of the U6 substrate RNA consisted of the introduction of two additional AMP nucleotides in front of the oligo(A) stretch, beginning at +27 of the wild-type sequence. This mutation was introduced into a U6-1 construct (resulting in U6-1/A) and aimed to extend the internal oligo(A) sequence from four to six AMP residues. The analysis of this mutant RNA as substrate for the U6-TUTase is shown in Fig. 3C. We have shown above Fig. 2. RNA substrate requirements of the U6-TUTase. (A) A partially purified U6-TUTase was tested under standard reaction conditions with 50 ng in vitro synthesized U6 snRNA substrate molecules containing zero (lane 1), one (lane 2), two (lane 3), three (lane 4), four (lane 5) or five (lane 6) 3¢-terminal UMP residues. A size standard consisting of unmodified U6-3 RNA, labeled by T7 RNA polymerase transcription in vitro, is indicated (m). Labeled RNA products were analysed in high-resolution gels as described in Materials and methods. (B)Thesameenzymeasin(A)wastestedwith[a- 32 P]UTP and U6-1 RNA as substrate under standard reaction conditions. In this case, however, increasing amounts of unlabeled UTP were added to the reaction: 0.0 l M (standard reaction; lane 1), 0.25 l M (lane 2), 0.5 l M (lane 3) 1.0 l M (lane 4) and 2.0 l M (lane 5). As above, unmodified labeled U6-3 RNA was included as a size standard (m). 974 R. Trippe et al.(Eur. J. Biochem. 270) Ó FEBS 2003 that, in these TUTase reactions, the nucleotide concentra- tions may be a limiting factor for the prolongation of U6 substrate RNA. Therefore, as in Fig. 2B, the TUTase reactions were performed either under standard conditions, i.e. in the absence of unlabeled UTP (lane 1), or with increasing amounts of unlabeled UTP added to the reaction (lanes 2–5). As before, increasing amounts of unlabeled nucleotides led to the disappearance of modified inter- mediates and to an overall reduction in signal intensities. However, comparison of the longest labeled products in lanes 2–5 with the two different unmodified marker RNAs (m 1 ,m 2 ) revealed that the TUTase-catalysed reaction again stopped exactly at a position corresponding to four 3¢-terminal UMP residues. In this analysis, the unmodified marker RNAs, labeled during T7 transcription, were U6-3 RNA (m 1 )andU6-2/ARNA(m 2 ). In its unmodified form, the length of U6-3 RNA exactly matches that of the unmodified U6-1/A sequence. The U6-2/A marker RNA corresponds to the U6-1/A mutant analysed here, but containing two instead of one 3¢-terminal UMP residues. Consequently, migration of this unmodified marker RNA should correspond exactly to the smallest labeled band of the modified U6-1/A RNA. This is confirmed by compar- ison of the lower bands of lanes 1–4 with the labeled m 2 band (Fig. 3C). These data provide evidence that the enzyme was not able to ÔreadÕ as template the two additional AMP nucleotides of the opposite RNA strand. Therefore, we conclude that the TUTase does not act as an RNA- dependent RNA polymerase. Rather the enzyme catalyses a strictly selective modification reaction, solely to regenerate the authentic 3¢-structure of newly transcribed cellular U6 snRNA, constituted by four UMP residues. Complex formation of the U6-TUTase with substrate RNA The U6 snRNA-specific TUTase described here differs from the unspecific enzyme of HeLa cells by its high selectivity for its substrate RNA. Therefore, we wanted to know whether this highly specific RNA–protein interaction might provide a useful tool for affinity purification of the enzyme. To test this possibility, we first studied binding of the TUTase to U6 snRNA by EMSA. For this, a partially purified U6-TUTase was run in a preparative Superdex G200 gel filtration column and individual fractions tested with 1 lg total cellular RNA (Fig. 4A). As indicated by the labeled U6 snRNA products, the specific enzyme was obtained as a broad peak with fractions 27 through 51, corresponding to a size range from 70 to 130 kDa for elution from this column. Fig. 3. Substrate analysis of the TUTase with structural mutants of the U6 snRNA. (A) Proposed secondary structure of U6 snRNA [25]. Mutant RNAs were cloned by insertion of CMP or AMP nucleotides into the internal oligo(A) stretch (+27/+30) of the wild-type sequence. (B) U6-2/C substrate RNA represents a U6 snRNA con- taining two 3¢-terminal UMP residues and one additional internal CMP nucleotide, inserted at position +28 of the wild-type sequence. 50 ng of this RNA were tested with labeled UTP under standard reaction conditions (lane 2) or with labeled [a- 32 P]GTP in the presence of 2.0 l M unlabeled UTP (lane 3). The position of unmodified U6-2/C RNA is indicated on the right. This marker RNA (lanes 1,4) was labeled during T7 transcription in vitro and served as a size standard. (C) U6-1/A RNA contains one 3¢-terminal UMP residue and an insertion of two additional AMP residues at position +27 of the wild- type sequence. U6-1/A RNA was analysed either under standard conditions (lane 1) or in the presence of increasing amounts of unlabeled UTP: 0.25 l M (lane 2), 0.5 l M (lane 3), 1.0 l M (lane 4) and 2.0 l M (lane 5). T7 RNA polymerase-labeled marker RNAs were: U6-3 (m 1 ) and U6-2/A (m 2 ); the latter corresponds to U6-1/A, but with two 3¢-terminal UMP residues (see text). Ó FEBS 2003 Characterization of U6 terminal uridylyltransferase (Eur. J. Biochem. 270) 975 A clear maximum of enzyme activity was detected in fractions 39–45. Subsequently, aliquots of the gel filtration fractions were incubated with U6-3 RNA, labeled by T7 RNA polymerase transcription in vitro. Complexes were separated in 6% polyacrylamide gels, as described in Materials and methods. As shown in Fig. 4B, the fractions with maximum enzyme activity, mainly fractions 39–45, clearly shifted the free U6 snRNA (arrowhead) to a complex of higher molecular mass (arrow). Furthermore, the distribution between fractions of the major shifted complex paralleled that of the enzyme activity observed in Fig. 4A. Additional minor complexes did not correspond to the pattern of enzyme activity, and may represent other cellular proteins capable of binding U6 snRNA, either specifically (such as LSm proteins) or in an unspecific way. These results confirm that in vitro stable complexes may be obtained between U6 snRNA and the corresponding U6-TUTase. In a second step, we wanted to generate a U6 snRNA- based affinity column for purification of the enzyme. As proteins have a significant tendency to bind nonspecifically to a variety of matrices, the choice of carrier is important in affinity purification. Previous experiments had shown that cellulose may be superior to other materials (unpublished observation). Another important point is the accessibility of the immobilized RNA. TUTase is expected to primarily recognize the 3¢-terminal structure of U6 snRNA. However, bulky compounds such as biotinylated nucleotides may interfere with the correct folding of the target RNA, thereby changing the structural motif recognized by the TUTase. Therefore, we decided to couple U6 snRNA to oligo(dT)– cellulose via an oligo(rA) linker, fused to the 5¢-end of the wild-type sequence. For this, a mutant gene was cloned with an oligo(A) (20) linker preceding the U6-3 RNA sequence. In vitro transcription of this template by T7 RNA poly- merase gave rise to U6-3 transcripts 149 nucleotides in length. A control experiment confirmed that the presence of the oligo(rA) linker and the joining element did not interfere with the TUTase reaction. This is shown in Fig. 5A. Comparison of lanes 2 and 3 shows that the amount of modified U6 RNA labeled in a standard TUTase reaction stayed the same, irrespective of whether U6-3 RNA (lane 2) or oligo(A)/U6-3 RNA (lane 3) was applied. In this standard U6-TUTase reaction, a control was included with 1 lg total cellular RNA (Fig. 5A, lane 1). It should be noted that the slightly slower migration observed with the modified U6-3 RNA (lane 2), compared with cellular U6 snRNA (lane 1), is due to two additional 5¢-terminal GMP residues in the U6-3 RNA. The resulting three 5¢-terminal GMP nucleotides of such U6 constructs are required for efficient initiation of transcription by T7 RNA polymerase. Fig. 5. Affinity chromatography of U6-TUTase with immobilized 5¢-adenylated U6 snRNA. (A) In vitro transcribed oligo(A)/U6-3 RNA (149 nucleotides in length; 50 ng) was tested in a standard TUTase reaction (lane 3) in comparison with the authentic U6-3 sequence (lane 2). Lane 1 shows a control reaction with 1 lg total cellular RNA. Labeled DNA fragments are included as marker (m). Products were analysed in 6% polyacrylamide gels containing 6 M urea. (B) After dialysis against 100 m M KCl, a partially purified U6-TUTase fraction (hydroxyapatite step, see Materials and methods) was applied to an oligo(dT)–cellulose column loaded with oligoA/U6-3 RNA. The load fraction (lane 1), the flow-through material (lane 2), and the fractions eluted with 600 m M KCl (lane 3) or 2000 m M KCl (lane 4) were analysed in a standard TUTase reaction with 1 lg total cellular RNA. Electrophoretic analysis of labeled RNA products in 6% polyacryl- amide gels and autoradiography were as before. Fig. 4. Complex formation of U6-TUTase with substrate RNA. (A) Prepurified U6-TUTase was run in a Superdex G200 column and the fractions indicated above each lane tested under standard conditions with U6-3 RNA as substrate. The loaded material is indicated by ÔlÕ and labeled DNA used as a marker indicated (m). (B) EMSA with the same TUTase fractions as in (A) and T7-transcribed labeled U6-3 RNA as substrate. Electrophoresis was in nondenaturing 6% polyacrylamide gels (see Materials and methods). Lane 1 shows a minus-protein control, with the free U6-3 RNA marked by an arrowhead. The shifted complex is indicated by an arrow (left side). 976 R. Trippe et al.(Eur. J. Biochem. 270) Ó FEBS 2003 This system was used for affinity chromatography of the U6-TUTase. For this, in vitro synthesized oligoA/U6-3 RNA was coupled to an oligo(dT)–cellulose matrix via A/T base-pairing. In an initial experiment, a fairly crude protein fraction (hydroxyapatite step) was applied to the column. At this level of purification, residual amounts of the nonspecific cellular TUTase are still present. This is evident from the standard TUTase reaction with total cellular RNA as substrate (Fig. 5B). As seen in lane 1 of Fig. 5B, the load fraction of the affinity column was able to transfer UMP residues to more than just U6 snRNA, although the latter was by far the most abundant labeled reaction product. In contrast, very little (if any) TUTase activity was associated with the flow-through fraction (lane 2). It should be noted that most of the minor bands seen with the load material (lane 1) were obtained in this flow-through fraction. As these labeled bands probably reflect residual activity of the nonspecific TUTase, it appears that the nonspecific enzyme did not bind to the affinity matrix. Subsequently, two more fractions were step-eluted with 600 m M KCl (lane 3) and 2 M KCl (lane 4). Before being analysed for TUTase activity, these fractions were dialysed against 100 m M KCl. This confirmed that most TUTase molecules were eluted from the affinity column at 600 m M KCl(lane3).Further- more, the upper section of lane 3 in Fig. 5B indicates that a small amount of the oligo(A)/U6-3 target RNA was also eluted from the affinity column. One has to keep in mind, however, that in this analysis the TUTase reaction was performed in the presence of 1 lg total cellular RNA as substrate. Taking into account the low concentration of U6 snRNA in total cellular RNA, even the smallest amounts of oligo(A)/U6-3 RNA coeluted from the column would attract considerable labeling by the U6-TUTase. Such artificial leakage of a relatively small proportion of the immobilized target RNA, however, would not affect the general suitability of this purification step. With this information to hand, a purification scheme was developed for the U6 snRNA-specific TUTase. Starting with HeLa cell S100 extracts, the enzyme was prepurified by Q-Sepharose, hydroxyapatite chromatography, and gel filtration in Superdex G200. This partially purified U6-TUTase fraction was subjected to affinity chromato- graphy on the oligo(A)/U6-3 RNA column described above. As proteins tend to bind to any matrix in an unspecific way, the combined peak fractions of the G200 column were split and run in parallel in two affinity columns: one consisting of oligo(dT)–cellulose only (Ô–Õ RNA column) and a second one loaded with poly(A)/U6-3 RNA (Ô+Õ RNA column). The TUTase assay performed with the material eluted from both columns at 600 m M KCl confirmed that binding of the enzyme strictly depended on the presence of the substrate RNA (Fig. 6A). As seen in lane 2 of Fig. 6A, the TUTase activity was exclusively eluted from the Ô+Õ column. In contrast, virtually no enzyme activity was detectable in lane 1, representing the elution fraction of the mock column (Ô–Õ). In agreement with these findings, the distribution of enzyme activities associated with the two flow-through fractions was exactly the other way around, i.e. full activity in the case of the mock column and less than 5% of TUTase passing through the Ô+Õ column (data not shown). To determine more precisely the molecular mass of the U6-TUTase, indirect labeling experiments were performed with the affinity-purified enzyme. Labeled U6-3 RNA was incubated with the protein under shift conditions, followed by UV cross-linking. Analysis of labeled proteins in SDS/poly- acrylamide gels was either directly (–) or after RNase A digestion (+). Figure 6B shows that a distinctly labeled RNP complex  145 kDa in size was obtained without RNase A treatment (lane 1). As expected, RNase treatment of the cross-linked material resulted in an overall loss of radioactivity associated with proteins (lane 2). Furthermore, albeit low in intensity, now one additional new band appeared that was not observed in lane 1. This band (arrow) had an apparent molecular mass of 115 kDa. All other labeled bands visible in lane 2 of Fig. 6B (mainly corres- ponding to a size range of 45–60 kDa) were already detectable in the absence of RNase treatment. These bands probably represent cross-linking products labeled by resid- ual amounts of free UTP. Finally, the protein composition of the affinity-purified material was analysed in silver-stained SDS/polyacrylamide gels. Initial results showed that high-salt elution from the affinity matrix still resulted in a very complex spectrum of polypeptides, not allowing unambiguous identification of the TUTase (data not shown). The high-salt conditions obviously mobilized a large number of proteins unspecifi- cally bound to the matrix. Therefore, a more gentle mode of elution was applied, avoiding changes in ionic strength. This approach consisted of RNase A treatment of the affinity columns, and indeed resulted in recovery of vastly reduced Fig. 6. Affinity purification and indirect labeling of U6-TUTase. (A) U6-TUTase prepurified by Q-Sepharose, hydroxyapatite and Super- dex G200 was loaded in parallel to an oligo(dT)–cellulose column (Ô–Õ) or an oligo(dT) column coupled with oligoA/U6-3 RNA (Ô+Õ). After being washed with 100 m M KCl, material eluted at 600 m M KCl was tested for TUTase activity with U6-3 substrate RNA in a standard reaction. Labeled products obtained with the elution fraction of the Ô–Õ column (lane 1) and of the Ô+RNAÕ column (lane 2) were analysed as before. DNA marker fragments are shown on the left (m). (B) Affinity- purifiedTUTasewasincubatedwithlabeledU6-3RNAundershift conditions, followed by UV cross-linking. Half of the material was analysed directly (without nuclease digestion, Ô–Õ, lane 1) and half after RNase A treatment (Ô+Õ, lane 2) in SDS/polyacrylamide gels (see Materials and methods). Numbers on the right indicate the positions of unlabeled marker proteins (kDa). Ó FEBS 2003 Characterization of U6 terminal uridylyltransferase (Eur. J. Biochem. 270) 977 numbers of proteins. Electrophoretic analysis of proteins obtained from the Ô+Õ and Ô–Õ RNA affinity columns is presented in Fig. 7. As is evident from lanes 1 and 2, a simple protein spectrum was obtained after RNase A elution (lane 3 shows a control with the RNase A alone). Most importantly, the fractions obtained from the Ô+Õ (lane 1) and the Ô–Õ (lane 2) columns differed by one prominent polypeptide only (arrow). Apart from a few minor quan- titative differences, all other bands were identical between the two fractions. The prominent band selectively eluted from the Ô+Õ column showed an apparent molecular mass of 115 kDa, which is in full agreement with the size obtained previously for the U6-TUTase by indirect labeling (Fig. 6B). Together, these lines of evidence suggest that the 115-kDa protein is probably the human U6 snRNA- specific TUTase. Discussion By several criteria, U6 snRNA is remarkable among the small stable RNA molecules of eukaryotic cells (see the Introduction). Furthermore, it seems to play a major role in the center of the active spliceosome. The finding of accompanying enzymes responsible for the highly specific modification of this particular RNA, such as a 3¢-nuclease and a 3¢-terminal uridylyltransferase [20,21], supports the structural significance of U6 snRNA. Consequently, one would expect that the combined action of these two proteins (and presumably others) is closely associated with the biological function of U6 snRNA in pre-mRNA splicing. Therefore, detailed information on the reactions catalysed by these two U6-specific enzymes may provide a valuable tool for studying internal events within the spliceosome. Here, we have reported on the catalytic properties and substrate requirements of the U6 snRNA-specific TUTase. With respect to its general properties, it seems to correspond closely to the nonspecific TUTase (Fig. 1) [26]. Both enzymes were inhibited by high salt concentrations and showed maximal stimulation by 5 m M Mg 2+ , but with clearly different RNA substrate specificities. In contrast with the U6-TUTase, however, the reaction catalysed by the nonspecific enzyme was found to be strictly limited to the transfer of only one UMP residue [26], irrespective of whether or not higher concentrations of UTP were applied. The transfer of more than one UMP residue by the U6-TUTase (Fig. 2B) was not observed in our initial study [21], because it depends on the presence of higher UTP concentrations. In addition, these experiments were per- formed with substrate RNAs that contained three (or more) 3¢-terminal UMP residues. At first glance, the finding that the U6-TUTase is able to transfer more than one UMP may classify it as a common poly(U) polymerase. However, for this class of enzyme, the length of the poly(U) product is a function of the incubation time. Certainly, this was not the case here. The addition of UMP residues was strictly limited to four, thereby restoring the 3¢-terminal structure of newly transcribed cellular U6 snRNA. Therefore, the mode of action of the U6-TUTase appears to be tightly controlled by the structure of the RNA substrate. This notion is supported by the observation that a synthetic U6-0 RNA, containing no 3¢-UMP at all, was not accepted as substrate. As U6-0 RNA carries an AMP nucleotide at its 3¢-end, this finding was reminiscent of results obtained with various substrate RNAs in unfractionated HeLa cell extracts and frog oocytes [27]. One has to bear in mind, however, that the uridylating activity analysed in those experiments did not show any substrate RNA specificity. The four 3¢-terminal UMP residues of newly tran- scribed U6 snRNA exactly match four internal AMP residues. Consequently, it is tempting to speculate that the U6-TUTase functions as an RNA replicase. How- ever, analysis of constructs with internal Ôopposite strandÕ mutations definitely excluded such a mode of action. Surprisingly, even the introduction of two additional AMP residues into the internal oligo(A) stretch did not give rise to an extended TUTase reaction product, now containing six instead of four complementary 3¢-terminal UMP residues. Therefore, it appears that the underlying principle is not simply to ensure a double strand at the basis of the 3¢-terminal stem-loop structure of U6 snRNA. Rather, these results suggest that a sophisticated mechanism controls a highly restrictive elongation pro- cess, only allowing restoration of the authentic 3¢-end of U6 snRNA. Such a scenario attributes a special import- ance to the 3¢-terminal structure of this RNA, with four UMP residues being involved in base-pairing. Apart from a more general contribution to the overall folding of U6 snRNA, this A/U RNA duplex element may ensure the appropriate stability for melting and reasso- ciation of this spliceosomal RNA, a prerequisite to the Fig. 7. SDS/PAGE of proteins recovered from affinity chromato- graphy. U6-TUTase was affinity-purified as described in Fig. 6. In this case, however, elution of the bound material from the plus (+, lane 1) and the minus (–, lane 2) column was by treatment with 100 lg RNase A. Lane 3 shows the control analysis of the RNase A used for elution. Protein bands were visualized by silver staining of the gel. The molecular mass (kDa) of marker proteins (m) is indicated on the left. 978 R. Trippe et al.(Eur. J. Biochem. 270) Ó FEBS 2003 extensive RNA rearrangements that occur during the splicing procedure. Several uridylating enzyme activities have been charac- terized [26,28–30]. However, the TUTase analysed here differs from those by its pronounced RNA substrate specificity [21]. This RNA selectivity is superimposed by the highly specific control of the elongation reaction described above. In this context, it is interesting to note that the molecular mass of the U6-TUTase activity obtained under native conditions in the gel-filtration column (Fig. 4A) exactly corresponded to that of the specific polypeptide observed in denaturing gels, after affinity chromatography (Fig. 7). This supports the notion that the catalytic activity of the U6-TUTase, together with its two specificities outlined above, is associated with a single polypeptide chain. Therefore, binding of the U6-TUTase to its target RNA will certainly establish an interesting model system for studying a very specific but transient RNA– protein interaction. For such a detailed analysis, however, a recombinant TUTase will be required. Such a recombinant protein would also allow a structural comparison of the enzyme with other previously cloned TUTases [30,31]. In addition, the availability of a recombinant U6-TUTase would give access to monoclonal antibodies, which in turn should provide a valuable tool to study the functional significance of this U6 snRNA modification. Such func- tional studies would provide clues to why evolution allowed this small stable U6 snRNA the unique luxury of acquiring its ÔownÕ modifying enzyme. Acknowledgements We thank Dr Andre ´ Frontzek for skilful introduction into the RNA electrophoretic mobility-shift analysis technique, and Nadine Pieda for expert technical assistance. Thanks are also due to Klaus Grabert for the photographs. This work was supported by a grant from the Deutsche Forschungsgemeinschaft (Be 531/19-1). References 1. Will, C.L. & Lu ¨ hrmann, R. (1997) Protein functions in pre- mRNA splicing. Curr. Opin. Cell Biol. 9, 320–328. 2. Burge, C.B., Tuschl, T.H. & Sharp, P.A. (1999) Splicing of precursors to mRNAs by the spliceosomes. In RNA World II (Gesteland, R.F., Cech, T.R. & Atkins, J.F., eds), pp. 525–560. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, USA. 3. Nilsen, T.W. (1998) RNA–RNA interactions in nuclear pre- mRNA splicing. In RNA Structure and Function (Simons, R. & Grunberg-Manago, M., eds), pp. 279–307. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, USA. 4. Staley, J.P. & Guthrie, C. (1998) Mechanical devices of the spliceosome: motors, clocks, springs and things. Cell 92,315– 326. 5. Kuhn, A.N., Li, Z. & Brow, D.A. (1999) Splicing factor Prp8 governs U4/U6 RNA unwinding during activation of the spliceosome. Mol. Cell 3, 65–75. 6. Kuhn, A.N. & Brow, D.A. (2000) Suppressors of cold-sensitive mutation in yeast U4 RNA define five domains in the splicing factor Prp8 that influence spliceosome activation. Genetics 155, 1667–1682. 7. Kunkel, G.R., Maser, R.L., Clavet, J.P. & Pederson, T. (1986) U6 small nuclear RNA is transcribed by RNA polymerase III. Proc. Natl Acad. Sci. USA 83, 8575–8579. 8. Reddy,R.,Henning,D.,Das,G.,Harless,M.&Wright,D.(1987) The capped U6 small nuclear RNA is transcribed by RNA polymerase III. J. Biol. Chem. 262, 75–81. 9. Singh, R. & Reddy, R. (1989) a-Monomethyl-phosphate: a cap structure in spliceosomal U6 snRNA. Proc. Natl Acad. Sci. USA 86, 8280–8283. 10. Brow, D.A. & Guthrie, C. (1988) Spliceosomal RNA U6 is remarkably conserved from yeast to mammals. Nature (London) 334, 213–218. 11. Branlant,C.,Krol,A.,Ebel,J.P.,Lazar,E.,Haendler,B.&Jacob, M. (1982) U2 RNA shares a structural domain with U1, U4 and U5 RNAs. EMBO J. 1, 1259–1265. 12. Liautard, J.P., Sri-Widada, J., Brunal, C. & Jeanteur, P. (1982) Structural organization of ribonucleoproteins containing small nuclearRNAsfromHeLacells.Proteinsinteractcloselywitha similar structural domain of U1, U2, U4 and U5 small nuclear RNAs. J. Mol. Biol. 162, 623–643. 13. Cooper, M., Jonston, L.H. & Beggs, J.D. (1995) Identification and characterization of Uss1p (Sdb23p): a novel U6 snRNA asso- ciated protein with significant similarity to core proteins of small nuclear ribonucleoproteins. EMBO J. 14, 2066–2075. 14. Se ´ raphin, B. (1995) Sm and Sm-like proteins belong to a large family: identification of proteins of the U6 as well as the U1, U2, U4 and U5 snRNPs. EMBO J. 14, 2089–2098. 15.Achsel,T.,Brahms,H.,Kastner,B.,Bachi,A.,Wilm,M.& Lu ¨ hrmann, R. (1999) A doughnut-shaped heteromer of human Sm-like proteins binds to the 3¢-end of U6 snRNA, thereby faci- litating U4/U6 duplex formation in vitro. EMBO J. 18, 5789–5802. 16. Mayes, A.E., Verdone, L., Legrain, P. & Beggs, J.D. (1999) Characterization of Sm-like proteins in yeast and their association with U6 snRNA. EMBO J. 18, 4321–4331. 17. Lund, E. & Dahlberg, J.E. (1992) Cyclic 2¢,3¢-phosphates and nontemplated nucleotides at the 3¢ end of spliceosomal U6 small nuclear RNAs. Science 255, 327–330. 18. Rinke, J. & Steitz, J.A. (1985) Association of the lupus antigen La withasubsetofU6snRNAmolecules.Nucleic Acids Res. 13, 2617–2629. 19. Gu, J., Shumyatsky, G., Makan, N. & Reddy, R. (1997) Formation of 2¢,3¢-cyclic phosphates at the 3¢end of human U6 small nuclear RNA in vitro. Identification of 2¢,3¢-cyclic phos- phates at the 3¢ ends of human signal recognition particle and mitochondrial RNA processing RNAs. J. Biol. Chem. 272, 21989– 21993. 20. Booth, B.L. & Pugh, B.F. (1997) Identification and characteriza- tion of a nuclease specific for the 3¢ end of the U6 small nuclear RNA. J. Biol. Chem. 272, 984–991. 21. Trippe, R., Sandrock, B. & Benecke, B.J. (1998) A highly specific terminal uridylyl transferase modifies the 3¢-end of U6 small nuclear RNA. Nucleic Acids Res. 26, 3119–3126. 22. Weil,P.A.,Segall,J.,Harris,B.,Ng,S Y.&Roeder,R.G.(1979) Faithful transcription of eukaryotic genes by RNA polymerase III in systems reconstituted with purified DNA templates. J. Biol. Chem. 254, 6163–6173. 23. Laemmli, U.K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature (London) 227, 680–685. 24. Melton, D.A., Krieg, P.A., Rebagliati, M.R., Maniatis, T., Zinn, K.&Green,M.R.(1984)Efficientin vitro synthesis of biologically active RNA and RNA hybridization probes from plasmids con- taining a bacteriophage SP6 promoter. Nucleic Acids Res. 12, 7035–7056. 25. Fortner, D.M., Troy, R.G. & Brow, D.A. (1994) A stem/loop in U6 RNA defines a conformational switch required for pre mRNA splicing. Genes Dev. 8, 221–233. 26. Zabel, P., Dorssers, L., Wernars, K. & VanKammen, A. (1981) Terminal uridylyl transferase of Vigna unguiculata:purification Ó FEBS 2003 Characterization of U6 terminal uridylyltransferase (Eur. J. Biochem. 270) 979 and characterization of an enzyme catalyzing the addition of a single UMP residue to the 3¢-end of an RNA primer. Nucleic Acids Res. 11, 2433–2453. 27. Chen, Y., Sinha, K., Perumal, K. & Reddy, R. (2000) Effect of 3¢ terminal adenylic acid residue on the uridylation of human small RNAs in vitro andinfrogoocytes.RNA 6, 1277–1288. 28. Andrews, N.C., Levin, D. & Baltimore, D. (1985) Poliovirus replicase stimulation by terminal uridylyl transferase. J. Biol. Chem. 260, 7628–7635. 29. Tazi,J.,Forne,T.,Janteur,P.,Cathala,G.&Brunel,C.(1993) Mammalian U6 small nuclear RNA undergoes 3¢ end modi- fications within the spliceosome. Mol. Cell. Biol. 13, 1641–1650. 30. Aphasizhev, R., Sbicego, S., Peris, M., Jang, S H., Aphasizh- eva, I., Simpson, A.M., Rivlin, A. & Simpson, L. (2002) Try- panosome mitochondrial 3¢ terminal uridylyl transferase (TUTase): the key enzyme in U-insertion/deletion RNA editing. Cell 108, 537–648. 31. Perumal, K., Sinha, K., Henning, D. & Reddy, R. (2001) Puri- fication, characterization and cloning of the cDNA of human SRP RNA-3¢ adenylating enzyme. J. Biol. Chem. 276, 21791– 21796. 980 R. Trippe et al.(Eur. J. Biochem. 270) Ó FEBS 2003 . upstream primer and 5¢-TATGGAACGCTTCACGAATT-3¢ (U6- 0), 5¢-ATAT GGAACGCTTCACGAATT-3¢ (U6- 1) or 5¢-AATATGG AACGCTTCACGAATT-3¢ (U6- 2) as downstream primer, respectively Biochemical characterization of a U6 small nuclear RNA-specific terminal uridylyltransferase Ralf Trippe, Holger Richly* and Bernd-Joachim Benecke Department