TIBS 1 5 -
SEPTEMBER1 9 9 0
RNA polymerase II: subunit structure and function
RNA POLYMERASEI1, the general transcription factors with which it associates, and at least some of the mechanisms involved in transactivation are highly conserved in eukaryotes ~-3.Some components of the yeast and the mamRNA polymerase II is the core of the complex apparatus that is malian mRNA.transcription apparatus, responsible for the regulated synthesis of mRNA. A comprehensive including the TATA-binding factor, are knowledge of RNA polymerase II is essential to our understanding of the even functionally interchangeable. RNA molecular mechanisms through which a variety of transcription factors polymerase II has a conserved multiregulate eukaryotic gene expression. The recent cloning of genes for all subunit architecture in eukaryotes and ten subunits of yeast RNA polymerase II has revealed intriguing 40-50% amino acid sequence identity is similarities and differences between the eukaryotic RNA polymerase and often found when subunit sequences from different organisms are compared. its simpler prokaryotic counterpart. Epitope tagging and other experiments The vast arsenal of genetic techniques made possible by the cloning of these genes have provided a clearer available in Saccharomyces cerevisiae, picture of RNA polymerase II subunit composition, stoichiometry and and the ease with which the investifunction, and set the stage for further investigating the dialogue between gator can combine powerful genetic RNA polymerase II and transcription factors. and biochemical methods using this organism, makes the yeast transcription apparatus an excellent eukaryotic ten yeast RNA polymerase lI subunit they are related to the prokaryotic core prototype for study. genes, RPBI-RPBIO, in the haploid RNA polymerase subunits. Three other genome and, with one exception, all are essential subunits, RPB5, RPB6 and necessary for yeast cells to grow at RPB8, are also found in RNA polymerRNA polymerase II subunits and genes Yeast RNA polymerase II is composed wild-type rates and for survival within ases I and Ill, which are responsible for of ten polypeptides with apparent the normal temperature range for cell transcribing ribosomal RNA and small molecular weights that range from growth (Table I). The ten genes were RNAs, respectivelyTM. Surprisingly, 220 kDa to 10 kDa, whether purified by isolated by screening a ~.gtl 1 recombi- three of the four remaining small subconventional techniques ~or by immuno- nant DNA library with subunit-specific units, RPB4, RPB7 and RPB9, are not precipitation with a monoclonal anti- antibodies 7,8, and through the use of essential for yeast cell viability at modbody directed against an epitope-tagged oligonucleotide probes 9-~4. Both to erate temperatures; these subunits may subunit 4. Until recently, eukaryotic RNA design gene-specific oligonucleotides nonetheless contribute to fine tuning of polymerases were defined solely as a and to confirm that the correct genes the transcription apparatus 1°J3. The set of proteins that copurify with tran- had been isolated, the amino acid smallest subunit, RPB10, is vital for the scriptional activity; whether these sequences of tryptic peptides from all yeast cell despite its small size (46 polypeptides are all true subunits and ten subunits of the conventionally amino acid residues) relative to the holoenzyme (4329 residues) 12. have a role in transcription in vivo was purified enzyme were obtained. A picture of the enzyme has emerged not entirely clearL In contrast, the prokaryotic RNA polymerases, com- from structural and functional studies RNA polymerase II subunit composition and posed of three core subunits ([3', ]3 and of the RNA polymerase lI genes and stoichiometry With the genes for the ten putative it) and a specificity factor (¢~), are bet- their products. The three largest subter defined both biochemically and units, RPB1, RPB2 and RPB3, are essen- RNA polymerase II subunits cloned and geneticallys,6. The isolation and charac- tial for yeast cell viability and are prob- sequenced, a powerful tool, epitope tagterization of the genes that encode the ably responsible for RNA catalysisS,~4,~s; ging4, could be used to investigate the ten yeast RNA polymerase II subunits, the immunoprecipitation of epitopeTable I. Yeast RNA polymerase II subunit genes tagged subunits to deduce subunit composition and stoichiometry, and the isoGene SDS-PAGE mobility Genecopy Chromosomal Deletion References (mol. wt x 10-3) number location viability lation and analysis of RNA polymerase II mutants have substantially improved RPB1a 220 1 IV inviable 7,14 our understanding of the eukaryotic RPB2 150 1 XV inviable 8 RPB3 45 1 IX inviable 9 enzyme. RPB4 32 1 X conditional 10 There are single copies of each of the RPB5 27 1 II inviable 11 RPB6 23 1 XVl inviable 11 RPB7 17 1 Xll viable 13 N.A. Woychik and R.A. Young are at the RPB8 14 1 XV inviable 11 Whitehead Institute for Biomedical Research, RPB9 13 1 VII conditional 13 Nine CambridgeCenter, Cambridge, MA RPBIO 10 1 XV inviable 12 02142, USA and Department of Biology, Massachusetts Institute of Technology, aAISOknownas RPO21,RPB220. Cambridge, MA 02139, USA. 347 © 1990,ElsevierSciencePublishersLtd,(UK) 0376-5067/90/$02.00
TIBS 15 - SEPTEMBER 1 9 9 0 (b)
S5 S I
Crude Extract Polymin P P r e c i p i t a t i o n
RPB7 . . . . RPB8-
Figure 1 (a) Analytical purification of RNA polymerase II. The RPB3 subunit was modified by the addition of a nine amino acid sequence recognized by a monoclonal antibody9. Extracts from cells labeled with [35S]methionineor 32p were fractionated by polymin P precipitation, followed by immunoprecipitationwith a monoclonal antibody specific for the epitope tag. (b) An autoradiograph of an SDS polyacrylamidegel containing the [3SS]methionineand 32p-labeledproteins that are immunoprecipitatedfrom cells that contain RPB3 with (+) or without (-) the epitope tag recognized by monoclonal antibody 12CA5. Reproduced,with permission, from Ref. 4.
enzyme from a new perspective (Fig. la). The epitope-tagging approach has several advantages over immunoprecipitation with standard anti-subunit antibodies for investigating a complex multisubunit enzyme: the addition of the epitope to the end of a polypeptide minimizes the potential for disruption of the multi-subunit complex by antibody binding; the experiments can be carried out with a well characterized monoclonal antibody; and, most importantly, a negative control can be prepared from cells that do not carry the epitopetagged protein. RPB3 was selected as a target for immunoprecipitation because it had been demonstrated to copurify with RNA polymerase lI activity in vitro and to be essential for mRNA synthesis in vivo 9. Moreover, the addition of a nine amino acid epitope to its N-terminus did not adversely affect cell growth, suggesting that RNA polymerase II function was not significantly affected. The ten polypeptides which are immunoprecipitated from radiolabeled 348
yeast cell extracts containing the epitope-tagged RPB3 subunit (Fig. lb) are identical in size and number to those obtained for RNA polymerase II purified by conventional column chromatography. The epitope-tagging result thus confirms that RNA polymerase II is composed of these ten subunits. The relative stoichiometry of the subunits was calculated by combining knowledge of the methionine content of the subunits deduced from sequence analysis and the relative extents of labeling with [35S]methionine (Table II). The average RNA polymerase II molecule appears to be composed of one copy each of RPB1, 2, 6, 8 and 10, and two copies of RPB3, 5 and 9. The RPB4 and 7 subunits are present at less than one copy per enzyme molecule; approximately half of the RNA polymerase II molecules have RPB4 and RPB7. An implication of this observation is that two forms of RNA polymerase lI, differing in their subunit composition, may exist in vivo. The in vivo phosphorylation state of
RNA polymerase II subunits has also been investigated using the immunoprecipitation approach (Fig. lb). The RPB1 and RPB6 subunits are clearly phosphorylated. RPB1 appears to be present in both phosphorylated and unphosphorylated forms in vivo, with approximately half of the RNA polymerase II molecules containing one form, and half the other form. Relatively low levels of phosphate can be detected on RPB2 with longer exposures than that shown in Fig. 1. The presence of different forms of RNA polymerase II in the cell raises intriguing questions about their function. RNA I~lymetase U subunit functions Clues to the functions of the subunits have been obtained from sequence similarities with prokaryotic relatives, from studies of DNA and nucleotide binding in vitro, and from the analysis of mutant defects. The eukaryotic 'core' subunits.The amino acid sequences of the RNA polymerase II subunits revealed that some of them have relatives in prokaryotic RNA polymerase (Table II). The two largest RNA polymerase II subunits, RPB1 and RPB2, are clearly related to the two largest subunits of E. coil RNA polymerase, [Y and [38J4(Table II). The sequence similarity occurs in eight or nine segments that span the two large subunits (Fig. 2). The two large subunits of prokaryotic and eukaryotic RNA polymerases are functional homologs. The [Y subunit binds DNA, the [3 subunit binds nucleoside triphosphate substrates and interacts with the transcription factor ~, and both may contain portions of the catalytic site for RNA synthesis s& The eukaryotic RPB1 apparently has a DNA binding site, RPB2 binds nucleotide substrates, and both large subunits may contribute to the active site for catalysis ~6,~¢.Highly conserved lysine and histidine residues in RPB2 have been implicated in nucleotide binding TM. These lysine and histidine residues are located at positions 965 and 1095, respectively, in RPB28. Several features of RPB3 and the E. coli RNA polymerase (z subunit indicate that they are probably functional homologs. The two subunits are very similar in size, both polypeptides occur twice in the RNA polymerase molecule, and mutations in both subunits can affect assembly of the enzyme6,gJs. When the two proteins are aligned, there is a 20 amino acid segment
TI BS 1 5 - SEPTEMBER 1990
Table II. Yeast RNA polymeraseII subun~s Subunit
SDS-PAGEmobili~ (tool. wt x 10-3)
Proteinmass (mol. wt x 10-3)
C~2-C-X6-C-X2-H~zH~23~-X2-Ca C-X2-C-X15~2-C LKFVKVRVR]-I'~X12o-QRLRHMVDDKb
RPB3 RPB4 RPB5 RPB6 RPB7 RPB8 RPB9
45 32 27 23 17 14 13
35 25 25 18 18 17 14
2.1 0.5 2.0 0.9 0.5 0.8 2.0
4.3 4.6 10.2 5.2 4.2 4.2 9.6
(~)c (G7o)c phosphate
C-X2-C-X18-C-X2-C C-X2-C-X24-C-X2-C C-X2-C-G
arhe metal binding sequence C-X2-C-Gis found within this region. bputative purine nucleotide binding site. CLimited sequence similarity.
assembly-defective mutants in bacterial RNA polymerase which accumulate specific subassembly complexes for a particular core subunit mutation. The subcomplexes of RNA polymerase II subunits that accumulate in the RPB1, RPB2 and RPB3 assembly-defective mutants correspond to those found in ~', ![3 and ~ assembly mutants. The resemblance in size and stoichiometry, the presence of limited amino acid sequence similarity, and the shared assembly mutant phenotypes together argue that RPB3 and (z play similar roles in eukaryotic and prokaryotic RNA polymerases.
(residues 29-48 in RPB39) that exhibits considerable sequence similarity. To ascertain whether RPB3 and (z are related, we immunoprecipitated epitopetagged RPB3 from RPB1, RPB2, or RPB3 assembly-defective mutant strains to investigate whether subunit interactions among the three largest S. cerevisiae RNA polymerase II subunits parallel those known for the three E. coli core enzyme subunits '5. The order of assembly of E. coli RNA polymerase subunits is well defined, starting with the dimerization of two ~xsubunits, followed by the sequential addition of the and 13' subunits. In addition, there are
Although the three large eukaryotic RNA polymerase II subunits share many features with the three prokaryotic core subunits, the RPB1 subunit has a unique C-terminal domain (CTD) that is not shared with its prokaryotic homolog. This CTD contains 26 or 27 repeats, depending upon the strain, of the consensus sequence Pro-Thr-SerPro-Ser-Tyr-Ser]4,]9. Except for proline residues, the amino acid side chains are all hydrophilic, suggesting that the CTD is exposed to solvent and projects out from the globular fold of the rest of the polypeptide. The CTD is essential for yeast cell viability]9, is highly
A103 A104 A101
I 11 6 9
100 a m i n o acids
Rgu/e 2 Segments of RPB1 and RPB2 that are similar in sequence to the prokaryotic RNA polymerase subunits I~' and 13, respectively. The eight regions of substantial homology between RPB1 protein and the I~' subunit of E. coli RNA polymerase and the nine regions conserved between RPB2 and the 13 subunit of E. coli RNA polymerase are shown as black boxes. The box with diagonal lines represents the heptapeptide repeat domain (CTD). The position of RPB1 and RPB2 conditional mutations that affect RNA polymerase II function or assembly are also indicated; these mutations generally affect amino acid residues that are invariant among homologous subunits from other eukaryotes 3°.
TIBS15- SEPTEMBER1990 phosphorylated in a substantial portion of the RNA polymerase II molecules in the cell4,2°, and protein kinases have been purified that appear to be involved in CTD phosphorylation2L22. Mutations in the C-terminal repeat of RNA polymerase II may help elucidate the function of this unusual domain. Certain partial deletions of the CTD of S. cerevisiae RNA polymerase II cause cells to exhibit temperature-sensitive (ts), cold-sensitive (cs) and auxotrophic phenotypes~9,23.The phenotypes associated with these CTD mutations are not a consequence of instability of the large subunit; rather, they appear to reflect some functional deficiency of the enzyme. The possibility that RPB1 CTD responds to regulatory factors during initiation of transcription is a particularly attractive hypothesis, and is one of several that are currently under investigation. [A review in next month's issue of TIBSwill focus on the CTD.] The common subunlts. RPB5, RPB6 and RPB8 are essential components of all three nuclear RNA polymerases (see Ref. 11 and references therein). These subunits could conceivably be involved in nuclear localization, the maintenance of transcriptional efficiency, or the coordinate regulation of rRNA, mRNA and tRNA synthesis. Although their sequences have not provided clues to their functions, conditional mutations already constructed in two of the common subunit genes may help elucidate the role of these subunits in all three nuclear RNA polymerases. The nonessential subunits. RPB4, RPB7 and RPB9 subunits are not essential for mRNA synthesis, as cells lacking any one of these proteins are viable, at least at moderate temperatures. RPB4 and RPB7 appear to form a subcomplex within RNA polymerase II that can dissociate from purified RNA polymerase 1I under partially denaturing conditions (see Ref. 1 and references therein). Both RPB4 and RPB7 are missing from RNA polymerase II immunoprecipitated from cells lacking RPB4, indicating that RPB4 provides a crucial link between the RPB4-RPB7 complex and the rest of the enzyme4. RPB4-deficient cells grow slowly and are temperature sensitive, and RNA polymerase I1 lacking RPB4 has reduced activity in vitro, suggesting that RNA polymerase II requires the RPB4 subunit for maximal efficiency~°. It is intriguing that 102 amino acids in the middle of RPB4 are 30% identical to residues in a portion of the prokaryotic RNA polymerase (~70 subunit, but the
function of this portion of c 7° is not yet Drosophila proteins are identical, and defined, and the implications of this the cysteine repeat motifs are consequence similarity are not yet clear ~°. served. This degree of sequence conIn contrast to the effect of deleting servation supports the idea that this RPB4, the absence of RPB7 does not 'nonessential' subunit plays an importhave a marked effect on cell growth ~3. ant accessory role in transcription. The Together, these results indicate that the observation that eukaryotic RNA polymRPB4-RPB7 complex does not have a erase II subunit sequences are highly central role in mRNA synthesis, but conserved indicates that studies of the does contribute to the efficiency of the yeast enzyme will continue to provide mRNA transcription apparatus. insights relevant to all eukaryotic RNA Like the RPB4-RPB7 complex, RPB9 polymerases. is not necessary for mRNA catalysis, but is essential for cell growth at tem- Perspectives perature extremes and probably also A well-defined RNA polymerase II has a role in fine tuning the efficiency of enzyme provides an important starting the transcription apparatus 13. RPB9 is point for genetic and biochemical unusual among RNA polymerase lI sub- experiments that will improve our units in that it contains two cysteine understanding of transcription, and will repeat motifs, C-X2-C-X~8-C-X2-Cand C-X2- help elucidate how RNA polymerase C-X24-C-X2-C, which may bind some of recognizes and responds to regulatory the zinc that is associated with the signals. For example, conditional muenzyme24,2s. Sequences that have been tations in seven of the ten RNA polymimplicated in metal binding are also erase II subunit genes have been isolatfound in RPB1, RPB2 and RPB10 (Table ed and are being used to identify subunit functions and to search for interI0. acting proteins through second-site Eukaryotic RNA polymerase II subunit suppressor analysis. In second-site supsequences are highly conserved pressor analysis, the phenotype of a Sequence analysis of several RNA primary mutation (e.g. temperature senpolymerase II subunit genes from sitivity) in a subunit gene is reverted to eukaryotes other than yeast reveals a wild type by a mutation in another remarkable degree of amino acid gene, thus implicating a direct or insequence conservation. Comparison of direct interaction between the proteins the sequences of RPB1 subunits from S. encoded by the two genes. Second-site cerevisiae TM, Caenorhabditis elegans26, suppressor genetics has already identDrosophila melanogaster~7 and the ified previously unknown proteins that mouse28 reveal that almost 40% of the appear to interact with the RPB1 amino acid residues are invariant. The C-terminal repeat domain23. The funchighly conserved residues are not tional defects of RNA polymerase II conrestricted to the segments of the sub- ditional mutations and the transcripunit that are related in sequence to the tional role of proteins identified through prokaryotic RNA polymerase 13' subunit suppressor analysis can now be investi(Fig. 2). Similar observations have been gated using a variety of biochemical made with homologs of RPB28,29. Of the assays, including assays for the forapproximately two dozen RPB1 and mation of a TATA-associated transcripRPB2 conditional mutations that have tion complex and for specific, factorbeen mapped thus far, it is interesting dependent transcription initiation in that most affect amino acid residues vitro. This potent combination of genetthat are invariant among the ics and biochemistry will surely yield eukaryotes3°. valuable new insights into the molecuThe smaller subunits of S. cerevisiae lar mechanisms that govern transcripRNA polymerase 1I also appear to have tion. closely related counterparts in higher eukaryotes. Thirty per cent of the References 1 Sentenac, A. (1985) Crit. Rev. Biochem. 18, amino acid residues of RPB5, the largest 31-91 of the common subunits, are identical 2 Paule, M. R. (1981) Trends Biochem. Sci. 6, to those of the 23 kDa subunit of human 128-131 3 Guarente, L. (1988) Cell 52, 303-305 RNA polymerase IIu. The yeast RPB9 4 Kolodziej, P. A., Woychik, N. A., Liao, S-M. and subunit appears to have a homolog in Young, R. A. (1990) Mol. Cell. Biol. 10, D. melanogaster which is encoded by 1915-1920 DNA upstream of the suppressor of 5 Chamberlin, M. J. (1982) in The Enzymes, Vol. 15 (Boyer, P., ed.), pp. 61-86, Academic Press Hairy wing gene ]3. Almost 50% of the 6 Yura, T. and Ishihama, A. (1979) Annu. Rev. amino acid residues of the yeast and
TIBS 15- SEPTEMBER1990 Genet. 13, 59-97 7 Young, R. A. and Davis, R. W. (1983) Science 222,778-782 8 Sweetser, D., Nonet, M. and Young, R. A. (1987) Proc. Natl Acad. Sci. USA 84, 1192-1196 9 Kolodziej, P. A. and Young, R. A. (1989) Mol. Cell Biol. 9, 5387-5394 10 Woychik, N. A. and Young, R. A. (1989) Mol. Cell Biol. 9, 2854-2859 11 Woychik, N. A., Liao, S-M., Kolodziej, P. and Young, R. A. (1990) Genes Dev. 4, 313-323 12 Woychik, N. A. and Young, R. A. J. Biol. Chem. (in press) 13 Woychik, N. A., Lane, W. S., and Young, R. A. (submitted) 14 Allison, L. A., Moyle, M., Shales, M. and Ingles, C. J. (1985) Ce1142, 599-610 15 Kolodziej, P. A. and Young, R. A. (in press)
INTRONS, or intervening sequences, are sequence stretches which interrupt coding sequences of genes. They are transcribed along with the exons, the functional sequences of the mature RNA molecules, but are excised from the primary transcripts by a process called RNA splicing. On the basis of structural features such as nucleotide sequence or potential RNA folding, introns have been classified into four distinct groups: group I, group II, nuclear premRNA and nuclear pre-tRNA introns. Group I and lI introns are characterized by short conserved sequences and the potential ability of their RNA to fold in complex secondary structures that are conserved despite a high degree of primary sequence divergence TM. Thus far, group II introns have been found only in organelles of fungi and plants. Despite this relatively restricted distribution, more than 70 distinct group lI intron sequences have already been published (see Ref. 5 for review). Individual members of both group 16 and group II7-9 intron families are able to undergo efficient and accurate selfsplicing in vitro, in the absence of proteins. In both groups of introns, this splicing occurs via two successive phosphate transfers (i.e. transesterification reactions) (see Fig. 1). The main difference between group I and group II splicing lies in the nature of the hydroxyl group that initiates the first reaction. In group I introns it is the 3' OH of a free guanosine that serves as A. Jacquler is at the Laboratoire de G6n~tique des Levures, D6partement de Biologie Mol6culaire, Institut Pasteur, 25 rue du Dr Roux, 75724 Paris Cedex 15, France.
16 Carroll, S. B. and Stollar, B. D. (1983) J. Mol. Biol. 170, 777-790 17 Riva, M., Schaffner, A., Sentenac, A.,
Hartmann, G., Mustaev,A., Zaychikov, E. and Grachev, M. (1987) J. Biol. Chem. 262, 14377-14380 18 Berghofer, B., Krockel, L., Kortner, C., Truss, M., Schallenberg, J. and Klein, A. (1988) Nucleic Acids Res. 16, 8113-8128 19 Nonet, M., Sweetser, D. and Young, R. A. (1987) Cell 50, 909-915 20 Cadena, D. and Dahmus, M. (1987) J. Biol. Chem. 262, 12468-12474 21 Cisek, L. and Corden, J. (1989) Nature 339, 679-684 22 Lee, J. M. and Greenleaf,A. L. (1989) Proc. Natl Acad. Sci. USA 86, 3624-3628 23 Nonet, M. and Young, R. A. (1989) Genetics 123, 715-724
24 Solaiman, D. and Wu, F. Y-H. (1984) Biochemistry 23, 6369-6377 25 Lewis, M. K. and Burgess, R. R. (1982)in The Enzymes, Vol. 15, (Boyer, P., ed.), pp.
109-153, Academic Press 26 Bird, D. M. and Riddle, D. L. (1989) Mol. Cell. Biol. 9, 4119-4130 27 Jokerst, R. S., Weeks, J. R., Zehring, W. A. and Greenleaf,A. L. (1989) Mol. Gen. Genet. 215,
266-275 28 Ahearn, J. M., Jr, Bartolomei, M. S., West,
M. L., Cisek, L. J. and Corden, J. L. (1987) J. Biol. Chem. 262, 10695-10705 29 Falkenburg,D., Dworniczak,B., Faust, D. and Bautz, E. K. F. (1987) J. Mol. Biol. 195,
929-937 30 Scafe, C., Martin, C., Nonet, M., Podos, S., Okamura, S. and Young, R. A. (1990) Mol. Cell. Biol. 10, 1270-1275
Self-splicing group II and nuclear pre-mRNA introns: how similar are they?
The splicing pathway of pre-mRNA introns bears similarities to that of the group II introns, some members of which undergo self-splicing. The snRNAs may provide the pre-mRNA introns with RNA structures in trans comparable to those available in cis in group II introns. This article examines the available evidence for the hypothesis that the catalysis of these two splicing pathways is fundamentally equivalent. the attacking group, while in group II introns it is the 2' OH of an adenosine within the intron (7 or 8 nucleotides from its 3' end). The reaction produces an intermediate molecule comprising the intron and exon 2. In group II this intermediate is in a lariat form (a circular tailed molecule) with the first nucleotide of the intron (commonly a G) ligated via an unusual 2'-5' phosphodiester bond, to an A (designated the branch site) close to the intron-exon 2 boundary. In both group l and group II introns, the 3' OH of exon I, which is released by the first reaction, initiates the second transesterification reaction by nucleophilic attack of the intron-exon 2 junction. This results in ligation of exons 1 and 2, and release of the intron, which in group II is in a lariat form (see Fig. 1). Most interestingly, the splicing of nuclear pre-mRNA introns gives rise, via the same intermediates, to end products similar to those of group II intron
© 1990,ElsevierSciencePublishersLtd,(UK) 0376-5067/90/$02.00
splicing: the ligated exon and the released intron in a lariat form, with the first nucleotide of the intron (also a G) ligated, via a 2'-5' phosphodiester bond, to an A near the 3' end of the intron. This strongly suggests that pre~ mRNA splicing also occurs by two successive transesterification reactions. Moreover, the similarity between the two systems is reinforced by the fact that the consensus sequences at the 5' and 3' ends of group II introns (GUGYG ... and ... AY, respectively) are somewhat reminiscent of the corresponding sequences in pre-mRNA introns (GURAGU ... AG in higher eukaryotes and GUAUGU ... AG in yeast). This leads to the speculative but highly attractive hypothesis that pre-mRNA introns and group II introns may be evolutionarily related. Pursuing the speculation further, it suggests that pre-mRNA splicing is also an RNA-catalysed reaction in which most of the RNA catalyst would be provided in trans by the snRNAs