RNA Molecules Containing Exons Originating from Different Members of the Cytochrome P450 2C Gene Subfamily (CYP2C) in Human Epidermis and Liver

RNA molecules containing exons originating from different members of the cytochrome P450 2C gene subfamily (CYP2C) in human epidermis and liver

Peter G. Zaphiropoulos*

Department of Bioscience, Center for Nutrition and Toxicology, Karolinska Institute, Novum, 141 57 Huddinge, Sweden

Received April 22, 1999; Revised and Accepted May 17, 1999

ABSTRACT is difficult to explain by the scanning model. More recently it has been proposed that splice sites are initially recognized as Reverse transcription–PCR analysis in human epider- closely spaced pairs, positioned across an exon (or possibly mis, using primers from CYP2C18 and CYP2C19, across small introns) and subsequent exon juxtaposition events revealed products containing combinations between allow the defined exons to be joined together (4,5). Although canonically defined exons of these two genes. The the detailed mechanisms of exon juxtaposition are not fully major RNA species identified contained 2C18 exon 8 understood, a prominent role for the non-snRNP class of spliced with 2C19 exon 2. However, the terminal splicing factors, the SR proteins and the hnRNPs, is emerging exons 1 and 9 were never detected in any of these (6–15). composite molecules. When similar experiments In line with the proposed exon definition/exon juxtaposition were performed with liver RNA, exons 1 and 9 of both model of splicing is the finding of RNA molecules that have an 2C18 and 2C19 were readily identified in composite inverted order of exons (exon scrambling) (16,17). Once the 2C18/2C19 RNAs. Moreover, molecules containing exons are defined, it is plausible that some juxtaposition events may allow exons that are not in the same order as in genomic 2C9 sequences spliced with 2C18 exons were also DNA to be brought together. It is also possible that splice sites detected. These findings suggest that during the across a single exon could be joined, resulting in exon circles process of RNA splicing of the 2C transcripts, various (18–21). Moreover, the RNA species containing scrambled exon juxtaposition events may occur, including com- exons also have properties that are consistent with circular binations between exons of distinct genes. However, molecules, as exemplified by the lack of both a poly(A)+ tail the frequency of these events is quite low and the and a cap structure. Furthermore, in some cases, the exons levels of the composite RNA molecules are generally composing the scrambled RNAs were found to be absent in estimated at less than one molecule per cell. Since certain alternative spliced mRNA molecules (22,23). This the order of these genes on chromosome 10q24 is observation suggested that the processes of exon skipping and CYP2C18–CYP2C19–CYP2C9, it is conceivable that circular RNA formation may be interrelated and could arise from the same mechanism. the composite RNAs may result from multiple canonical To further investigate the variations that may occur during and inverse splicing events of a long pre-mRNA that the process of exon juxtaposition, it was examined whether encompasses the three genes. However, these mol- exons from distinct genes could be spliced together. As a ecules could also be rationalized as being the prod- model system, the cluster of the human CYP2C genes was ucts of trans splicing phenomena between distinct selected. These are four single copy genes that have been pre-mRNAs. mapped in chromosome 10q24, in a region of ~500 kb (24). The gene order and orientation in the cluster is Cen–CYP2C18– CYP2C19–CYP2C9–CYP2C8–Tel. The human CYP2C18 and INTRODUCTION CYP2C9 genes span >55 kb and possess nine exons, the The splicing process is a major event in eukaryotic gene intron–exon boundaries of which are conserved with those of expression allowing non-contiguous segments of pre-mRNAs, CYP2C genes of other species (25). the exons, to be joined together. The mechanism underlying Surprisingly, and contrary to expectations, epidermal RNA this phenomenon was originally thought to be rationalized by a molecules containing various combinations of exons from the scanning model mechanism, which, in its simplest form, suggests CYP2C18 and the CYP2C19 genes, joined at their canonical that the spliceosome moves along the pre-mRNA and excises splice sites, were detected. Moreover, liver RNA molecules the introns as it encounters them (1–3). However, the fact that containing exon combinations between the CYP2C18 and within large introns there are sequences with a good match to a either the CYP2C19 or the CYP2C9 gene, including the terminal single splice site consensus that nevertheless remain non-utilized exons 1 and 9, were also identified.

*Tel: +46 8 6089151; Fax: +46 8 6081501; Email: [email protected] 2586 Nucleic Acids Research, 1999, Vol. 27, No. 13

MATERIALS AND METHODS Table 1. Sequence of the PCR primers used in the RT–PCR amplification experiments Total human epidermal RNA was isolated by the guanidinium isothiocyanate method (26) as described before (23). Total Primer Sequence liver RNA was purchased from Clontech. cDNA synthesis was performed using either random hexamers or oligo(dT) primers, CYP2C18 following previously established protocols (23). Both the 1f2C18 5'-TGG ATC CAG CTG TGG CT initial and the nested PCR amplifications were performed for 1F2C18(nested) 5'-GTG CTC TGT CTC TCC TG 30 cycles, with 1 min at 94°C, 1 min at 52–54°Cand1minat 2f2C18 5'-TGT GTA TTT TGG CCT GAA G 72°C, using Perkin-Elmer model 480 and 2400 thermocyclers 2F (nested) 5'-ATA TGA AGC AGT GAA GGA G or a MJ Research model PTC-200 DNA engine. For the nested 2C18 amplifications, 1 µl of the first amplification mix was used 3f2C18 5'-TGG GGA AGA GGA GCA TC directly. When Taq polymerase was substituted for Expand 3F2C18(nested) 5'-GGA GTT GAG AAA AAC CAA T (Boehringer), the extensions were performed for 3 min at 4f 5'-ACT TGA TGG AAA AAT TCA ATG ° 2C18 72 C, using the buffers supplied by the manufacturer. The 4F (nested) 5'-CTC AGG ATT CTG AGC TCT PCR products were analyzed by gel electrophoresis and cloned 2C18 using the pGEM-T vector system (Promega). Plasmid DNA 5f2C18 5'-GGA GAG AAT AAA AGA ACA TC was extracted from randomly chosen clones using the JetQuick 5F2C18(nested) 5'-AAG AAT CCC TGG ACA TGA A kit (Genomed) and sequenced with dideoxynucleotides at the 7f2C18 5'-CAG AGA TAC ATT GAC CTC C facilities of Cybergene AB. Sequencing comparisons were 7F (nested) 5'-GAC CTG TGA TGT TAA ATT CAA performed with the BLAST programs of NCBI. 2C18 5r2C18 5'-ATA CAT AAC TTT TAA TGT AAG C

5R2C18(nested) 5'-TTT TAT TAT GAC TTC CTG GG RESULTS 9r2C18 5'-GCA GCC AAA CTA TCT GCC

Non-polyadenylated RNA molecules containing 2C19 9R2C18(nested) 5'-CAG ACA GGA ATG AAG CAG exons spliced 5' of 2C18 exons

During the RT–PCR analysis of the 2C18 scrambled RNA CYP2C19 molecules (23), some products were identified that contained 1f 5'-ATG TTT GCT TCT CCT TTC AA segments of 2C19 exons. These preliminary findings prompted 2C19 the initiation of a more thorough analysis of RNA species that 1F2C19(nested) 5'-GAA ATA TCC TAC AGA TAG ATA might contain combinations of 2C19 and 2C18 exons. A set of 4f2C19 5'-ACT TGA TGG AAA AAT TGA ATG

2C18 exon 5 antisense primers was used with sets of either 4F2C19(nested) 5'-ACA TCA GGA TTG TAA GCA C 2C19 exon 4 or exon 5 sense primers (Table 1) in RT–PCR 5f2C19 5'-GAA AGT AAA AGA ACA CCA AG amplification of oligo(dT) and random primed cDNA from 5F (nested) 5'-ATC GAT GGA CAT CAA CAA C human epidermis. Several different sized products were 2C19 detected from the random primed but not from the oligo(dT) 3r2C19 5'-ATC CCA AAA TTC CGC AGC primed cDNA with both sets of PCR primers. This finding 3R2C19(nested) 5'-CTC TTT CCA TTG CTG AAA AC suggests that the products obtained originate from non- 5r2C19 5'-CCA AAA TAT CAC TTT CCA TAA polyadenylated RNAs, as oligo(dT) priming is unable to result 5R (nested) 5'-GTA ATT TGT TAT GGG TTC CC in detectable amplification. Cloning and sequencing of these 2C19 products established that they were specifically initiated with 9r2C19 5'-GGA ATG AAG CAC AGC TGA the primers used and contained several combinations of 2C19 9R2C19(nested) 5'-GGT TAA AGT TCT GTA AAA TGA exons spliced together with 2C18 exons at their canonical splice sites (Table 2). Primers are numbered according to the exons, with f and F representing sense and r and R antisense primers. Polyadenylated and non-polyadenylated RNA molecules containing 2C18 exons spliced 5' of 2C19 exons To obtain additional evidence that the species detected originate Since in the experiment described above the design of the from polyadenylated RNAs, the experiment described above primers used was such that 2C19 sense primers were combined was repeated using new cDNA preparations that were either with 2C18 antisense primers, RT–PCR analysis was subsequently primed or not primed by oligo(dT). Only the oligo(dT) primed performed using sense primers from 2C18 and antisense primers cDNA resulted in detectable PCR products of the expected from 2C19. Sets of PCR primers from 2C18 exon 7 and from size, suggesting that an oligo(dT) priming event has to occur 2C19 exon 3 were combined (Table 1) and used to amplify during reverse transcription in order to allow efficient amplifi- oligo(dT) and random primed cDNA (Fig. 1). Interestingly, the cation of these species (data not shown). patterns of the PCR products obtained from both the oligo(dT) To further analyze the RNA molecules that contained 2C18 andtherandomprimedcDNAwerequitesimilar.Twomajor exons spliced upstream of 2C19 exons, the exon 7 sense species were revealed (Fig. 1, lanes 5 and 6) and by cloning primers of 2C18 were substituted with exon 4 sense primers and sequencing were found to contain 2C18 exons spliced (Table 1). Again, the pattern obtained with oligo(dT) primed together with 2C19 exons at their canonical splice sites (Table 2). cDNA had similarities to the one with random hexamers. Nucleic Acids Research, 1999, Vol. 27, No. 13 2587

Table 2. Schematic diagram of the various products obtained with the use of RT–PCR analysis

Figure 1. Agarose gel electrophoresis of the PCR products obtained using exon 7 sense primers from 2C18 with exon 3 antisense primers from 2C19 on human epidermal cDNA. Lane 1, molecular weight markers; lane 2, PCR products using the initial primers 7f2C18 and 3r2C19 with oligo(dT) primed cDNA; lane 3, as lane 2 but with random hexamer primed cDNA; lane 4, as lane 2 but without addition of cDNA; lane 5, PCR products using the nested primers 7F2C18 and 3R2C19 on an aliquot of the products shown in lane 2; lane 6, as lane 5 but with an aliquot of the products shown in lane 3; lane 7, as lane 5 but with an aliquot of the products shown in lane 4. The major species cloned and sequenced are indicated by arrows.

these products established that they contained combinations of 2C18 and 2C19 exons spliced at their canonical splice sites (Table 2). Only the longest 2C18/2C19 species, containing exons (4–5–6–7–8)2C18–(2–3)2C19, was detected in both oligo(dT) and random primed cDNAs. Worth noting is that the first 2C19 exon present in these products was always found to be exon 2. RT–PCR products containing repeats of 2C19 exons: evidence supporting a circular nature of these RNA molecules To investigate whether RNA molecules containing combinations of 2C18 and 2C19 exons could be detected not only with sets of PCR primers from each of these two CYP2C genes, but also with primers originating from a single CYP2C gene, PCR was performed using primers in both orientations of 2C19 exon 5 (Table 1). It was anticipated that most of the products would represent typically scrambled molecules, containing exclusively 2C19 exons. Oligo(dT) primed cDNA did not result in specifically amplified PCR products, however, random primed cDNA resulted in numerous products that were White boxes represent 2C18 exons, shaded boxes 2C19 exons and hatched cloned and sequenced (Fig. 2 and Table 2). As expected, most boxes 2C9 exons. The forward or reverse primers of individual exons are of these products contained exclusively 2C19 exons, with the indicated by the number of the exon and the letters F or R, respectively. Note most abundant species having a 5–4–5 order of exons. How- that some products contained sequences representing complete introns ever, some products were found to contain a 2C18 exon, exon (intron), segments of introns (intronic) or were of unknown origin (?). 8, that was flanked on both sides with 2C19 exons (three out of 50 independent clones; order of exons: 5–82C18–2–3–4–5). Moreover, three additional independent clones were identified However, the latter was characterized by a higher complexity. that contained repeats of 2C19 exons 5 and 4 (order of exons: This finding suggests that the products common in the 5–4–5–4–5), indicating that scrambling had occurred twice oligo(dT) and the random primed cDNA preparations originate between exon 5 and exon 4. To rationalize the mechanism of from polyadenylated RNAs. However, the products present generation of these molecules, one has to assume that not only exclusively in the random primed cDNA are likely to result one but two distinct trans splicing events had occurred. from non-polyadenylated RNAs. Cloning and sequencing of Alternatively, a simpler interpretation for the latter molecule is 2588 Nucleic Acids Research, 1999, Vol. 27, No. 13

mRNA, serial 10-fold dilutions of epidermal oligo(dT) cDNA were amplified with 2C18 exon 7 sense primers combined with either 2C19 exon 3 antisense or 2C18 exon 9 antisense primers. The 10–1 dilution allowed the synthesis of equivalent amounts of the 2C18 product as the 2C18/2C19 product generated from the undiluted cDNA, suggesting that the level of expression of the 2C18/2C19 species is about an order of magnitude less than the 2C18 mRNA (data not shown). Lack of RNA molecules containing combinations of 2C18 exons with either 2C8 or 2C9 exons in human epidermis To determine whether the finding of RNA molecules containing exons from two genes of the CYP2C cluster might not be limited to the neighboring CYP2C18 and CYP2C19 genes, RT–PCR amplification was performed using primers from Figure 2. Agarose gel electrophoresis of the PCR products obtained using exon 5 sense and antisense primers from 2C19 on human epidermal cDNA. 2C18 exons combined with primers from exons of either 2C8 Lane 1, molecular weight markers; lane 2, PCR products using the initial or 2C9. However, in no case could a specifically amplified primers 5f2C19 and 5r2C19 with oligo(dT) primed cDNA; lane 3, as lane 2 but PCR product be detected (data not shown). This finding with random hexamer primed cDNA; lane 4, as lane 2 but without addition of suggests that the proximity of the CYP2C18 and CYP2C19 cDNA; lane 5, PCR products using the nested primers 5F2C19 and 5R2C19 on an aliquot of the products shown in lane 2; lane 6, as lane 5 but with an aliquot of genes in chromosomal locus 10q24 might be an essential factor the products shown in lane 3; lane 7, as lane 5 but with an aliquot of the products that allows the generation of epidermal RNA molecules shown in lane 4. The single product seen from the oligo(dT) amplification containing exon combinations from these two genes. (lane 5) was sequenced and found to represent a misprimed product. The major species from lane 6, containing exons 5–4–5, is indicated by an arrow. 2C18/2C19 RNA molecules containing 2C18 exon 1 or 2C19 exon 9 in human liver that a circular RNA template composed of exons 4 and 5 has The unexpected finding that in epidermis the 2C18 exon 1 and been reverse transcribed several times during cDNA synthesis the 2C19 exon 9 are not present in the 2C18/2C19 RNA molecules prompted an investigation as to whether these exons might be generating linear 5–4–5–4–5–4–… species. Indeed, the capability present in composite RNAs from the liver. Indeed, by RT–PCR of polymerases to follow a rolling circle model of DNA/RNA analysis using primer combinations either from 2C18 exon synthesis is well documented (28,29). Moreover, when the 1 and 2C19 exon 3 or from 2C18 exon 7 and 2C19 exon 9 same experiment was performed with Expand, a combination of (Table 1), 2C18/2C19 products containing the terminal exons thermostable polymerases that is optimized for the amplification were detected from both oligo(dT) and random primed cDNA of long DNA templates, a PCR product containing four repeats (Table 2). Interestingly this experiment revealed that 2C18 of the scrambled 5–4 core sequence could be detected (data not exon 1 can also be spliced with exons from 2C9. This 2C18/ shown). 2C9 product was generated because the exon 3 antisense The polyadenylated 2C18/2C19 RNA molecules do not primer used also amplifies 2C9 sequences (complementary contain the expected terminal exons except for the 5' nucleotide to 2C9). What is worth noting is that the 2C9 exon spliced with 2C18 exon 1 is not a canonical When sense primers from exon 5 of 2C18 were combined with exon but originates from the 5' flanking sequence of the antisense primers from exon 5 of 2C19 (Table 1) a major product CYP2C9 gene. It is enclosed between positions –246 and –111 composed of exons (5–6–7–8)2C18–(2–3–4–5)2C19 and present and has AG and GT dinucleotides at its borders (25). in both the oligo(dT) and the random primed cDNAs was revealed (data not shown). This continuous detection of apparently poly- 2C18/2C19 RNA molecules containing 2C18 exon 9 or adenylated RNA molecules, which were mostly characterized 2C19 exon 1 in human liver by the joining of 2C18 exon 8 to 2C19 exon 2, prompted exam- Since a terminal exon from each of the CYP2C18 and ination of the likely exon borders of these molecules. For that CYP2C19 genes was found to be present in composite 2C18/ purpose sense primers from exons 1–3 of 2C18 were combined 2C19 RNA molecules in liver, it was investigated whether the with the antisense primers from exon 3 of 2C19 that were remaining terminal exons, i.e. exon 9 from CYP2C18 or exon previously used (Table 1). The exon 3/exon 3 combination 1fromCYP2C19, could also be present in composite 2C18/ resulted in the identification of two major polyadenylated 2C19 species. For that purpose RT–PCR analysis was per- species composed of exons (3–4–5–6–7–8)2C18–(2–3)2C19 and formed using primer combinations either from 2C19 exon 1 (3–5–6–7–8)2C18–(2–3)2C19, however, the exon 1/exon 3 com- and 2C18 exon 5 or from 2C19 exon 5 and 2C18 exon 9 bination, as well as the exon 2/exon 3 combination, did not (Table 1). Indeed, RNA molecules containing exon combin- result in detection of PCR products from the cDNA preparations. ations between 2C18 and 2C19 that included either 2C19 exon Similarly, no products could be identified when sense primers 1 or 2C18 exon 9 were detected (Table 2). Moreover, a species from exon 7 of 2C18 were combined with antisense primers containing 2C9 exon 5 spliced to 2C18 exon 8 was also from exon 9 of 2C19 (data not shown). identified. This product was generated because the exon 5 To examine the level of expression of the species containing sense primer used also amplifies 2C9 sequences (complementary 2C18 exon 8 spliced with 2C19 exon 2 relative to the 2C18 except for position 4 to 2C9). Nucleic Acids Research, 1999, Vol. 27, No. 13 2589

DISCUSSION transcript hypothesis has to include an additional novel element, namely that inverse splicing events are not limited to internal The presence of various combinations of 2C18 exons with exons but may also encompass terminal exons. either 2C19 or 2C9 exons in contiguous RNA molecules is Inclusion of the 2C18 and 2C19 terminal exons in composite intriguing and suggests that juxtaposition events of defined 2C18/2C19 RNA molecules can also be interpreted as an argument exons might not be limited to single genes. Splicing in trans is that these species represent products of authentic trans splicing known to occur in lower eukaryotes, the trypanosomes and events. Canonical pre-mRNAs of individual CYP2C genes, at also in nematodes and flatworms (30,31). Moreover, recent low frequency, undergo splicing events in trans,simply experimental findings have suggested that this phenomenon because of the proximity of the newly synthesized primary may also be occurring in higher eukaryotes (32–34). transcripts to the active spliceosomal complex. In fact, this The RNA molecules identified in this report containing would be in agreement with the increasing body of evidence combinations of exons from different CYP2C genes are not that suggests a tight linkage between transcription and pre- abundant. Northern analysis, in our hands, is not sensitive mRNA processing (39). However, the identified RT–PCR enough to allow their detection and efforts to show their product that contains 2C18 exon 8 embedded within 2C19 presence by RNase mapping were also unsuccessful. However, exons, i.e. the exon 5–82C18–2–3–4–5 molecule, cannot be the fact that the exons present in these molecules are spliced at rationalized by a single trans splicing event. To account for correct sites provides a strong argument that they pre-exist in this RNA species one has to assume that two distinct trans the RNA samples analyzed and therefore do not result from splicing events have occurred, one that joined exon 5 of 2C19 any type of PCR artifact. Certainly these RNA species are not withexon8of2C18andanotherthatjoinedexon8of2C18 abundant and may, in some cases, be present at levels lower with exon 2 of 2C19. An alternative and simpler interpretation than one copy per cell. On the other hand it should be noted though is that this molecule was generated by inverse splicing that in yeast, 80% of expressed genes exist at levels of 0.1– events from an extended primary transcript that encompasses 2 molecules per cell, implying that low abundance transcripts the two CYP2C genes. On the other hand, an extensive analysis represent the majority of expressed RNAs in biological systems of scrambled RNAs containing exclusively 2C8 exons revealed (35). products that can only be rationalized by trans splicing events (C.Finta and P.G.Zaphiropoulos, unpublished observations). Trans splicing versus splicing from a multigene primary transcript The lack of the canonical terminal exons in the 2C18/2C19 RNA molecules in epidermis was not an anticipated finding. What is a mechanism that may account for the finding of RNA These observations suggest that in this tissue, alternative molecules containing exon combinations from tandemly terminal exons are used, since it is difficult to assume that arranged genes? At the genomic level, although the order of internal exons can substitute for terminal exons. In this context these genes has been determined as CYP2C18, CYP2C19 and it is worth noting that the penultimate exon with a segment of CYP2C9, their relative orientation is unknown, i.e. whether the 3' flanking intron can substitute for the last exon in certain after the end of the CYP2C18 gene there is the start of the transcripts of the putative tumor suppressor gene Patched 2 CYP2C19 gene or vice versa. Moreover, it is not known if these (40), providing some clues as to the variations in the definition human CYP2C genes are in the same phase, i.e. transcribrd from of terminal exons that may occur. In liver, though, 2C18/2C19 the same DNA strand, although judging from the genomic RNA species containing terminal exons have been detected not organization of other CYP subfamilies, this is anticipated (36). only by the RT–PCR technology, but also by means of cDNA However, whatever their relative orientation may be, as long as library screening (41). This 2C18/2C19 cDNA contained 2C19 these CYP2C genes are in the same phase, it is conceivable that exon9andhadexon8of2C18joinedtoexon2of2C19, a primary RNA transcript could be generated that encompasses i.e. the most frequent joining of 2C18 and 2C19 exons observed not only the 5'-most gene (CYP2C18 or CYP2C9), but also by RT–PCR. additional genes of the cluster. This would support the notion In summary, the present work provides evidence that, at low that a transcription unit in higher eukaryotes may encompass a frequency, exons from distinct CYP2C genes can be spliced neighboring gene, as is the case with Caenorhabditis elegans together. Whatever the mechanism of generation of these RNA (37). Moreover, since splicing is known to have the capability molecules may be, authentic trans splicing events, splicing to join exons flanking introns that are >100 kb, one may assume from a multigene primary transcript or a combination of both, that, in this primary transcript, exon joining of distantly localized these findings substantiate the proposal that the splicing exons originating from different CYP2C genes could take process is not restricted to the linear order of exons of single place. In fact, the finding of a transcript containing 2C18 exon genes. At low frequency, events that may juxtapose exons 1 joined to a 2C9 exon-like sequence present in the promoter of from neighboring genes, irrespective of their relative orien- the CYP2C9 gene may support the postulation of splicing tations, can occur, implying an additional complexity in the events from a putative multigene primary transcript. However, expression of the informational content of genes. to account for the observation that 2C18 exons were found both upstream and downstream of 2C19 exons, one has to ACKNOWLEDGEMENTS assume that inverse splicing events also take place in processing of the extended primary transcript. This type of splicing, with Dr Csaba Finta is acknowledged for helpful discussions and downstream exons joined 5' of upstream exons, has been critical reading of the manuscript. This work was supported by demonstrated to occur only within internal exons (22,38). grants from the Swedish Natural Science Research Council, Therefore, to rationalize the finding of 2C18/2C19 RNA molecules the foundations Lars Hiertas Minne and Åke Wiberg and the containing the terminal exons 1 and 9, the multigene primary Karolinska Institute. 2590 Nucleic Acids Research, 1999, Vol. 27, No. 13

REFERENCES 20. Braun,S., Domdey,H. and Wiebauer,K. (1996) Nucleic Acids Res., 24, 4152–4157. 1. Lewin,B. (1980) Cell, 22, 324–326. 21. Pasman,Z., Been,M.D. and Garcia-Blanco,M.A. (1996) RNA, 2, 603–610. 2. Sharp,P. (1981) Cell, 23, 643–646. 22. Zaphiropoulos,P.G. (1996) Proc. Natl Acad. Sci. USA, 93, 6536–6541. 3. Lang,K.M. and Spritz,R.A. (1983) Science, 220, 1351–1355. 23. Zaphiropoulos,P.G. (1997) Mol. Cell. Biol., 17, 2985–2993. 4. Berget,S.M. (1995) J. Biol. Chem., 270, 2411–2414. 24. Gray,I.C., Nobile,C., Muresu,R., Ford,S. and Spurr,N.K. (1995) 5. Sterner,D.A., Carlo,T. and Berget,S.M. (1996) Proc. Natl Acad. Sci. USA, Genomics, 28, 328–332. 93, 15081–15085. 25. de Morais,S.M.F., Schweikl,H., Blaisdell,J. and Goldstein,J.A. (1993) 6. Gallego,M.E., Gattoni,R., Stevenin,J., Marie,J. and Expert-Bezancon,A. Biochem. Biophys. Res. Commun., 194, 194–201. (1997) EMBO J., 16, 1772–1784. 26. Chomczynski,P. and Sacchi,N. (1987) Anal. Biochem., 162, 156–159. 7. Chabot,B., Blanchette,M., Lapierre,I. and La Branche,H. (1997) 27. Romkes,M., Faletto,M.B., Blaisdell,J.A., Raucy,J.L. and Goldstein,J.A. Mol. Cell. Biol., 17, 1776–1786. (1991) Biochemistry, 30, 3247–3255. 8. Chandler,S.H., Mayeda,A., Yeakley,J.M., Krainer,A.R. and Fu,X.-D. 28. Daubendiek,S.L. and Kool,E.T. (1997) Nat. Biotechnol., 15, 273–277. (1997) Proc. Natl Acad. Sci. USA, 94, 3596–3601. 29. Diegelman,A.M. and Kool,E.T. (1998) Nucleic Acids Res., 26, 3235–3241 9. Cogan,J.D., Prince,M.A., Lekhakula,S., Bundey,S., Futrakul,A., 30. Blumenthal,T. (1995) Trends Genet., 11, 132–135. McCarthy,E.M.S. and Phillips,J.A. (1997) Hum. Mol. Genet., 6, 31. Evans,D., Zorio,D., MacMorris,M., Winter,C.E., Lea,K. and Blumenthal,T. 909–912. (1997) Proc. Natl Acad. Sci. USA, 94, 9751–9756. 10. Hertel,K.J. and Maniatis,T. (1998) Mol. Cell, 1, 449–455. 32. Eul,J., Graessmann M. and Graessmann,A. (1995) EMBO J., 14, 3226–3235. 11. Stark,J.M., Bazett-Jones,D.P., Herfort,M. and Roth,M.B. (1998) 33. Eul,J., Graessmann M. and Graessmann,A. (1996) Nucleic Acids Res., 24, Proc. Natl Acad. Sci. USA, 95, 2163–2168. 1653–1661. 12. Tacke,R., Tohyama,M., Ogawa,S. and Manley,J.L. (1998) Cell, 93, 34. Caudevilla,C., Serra,D., Miliar,A., Codony,C., Asins,G., Bach,M. and 139–148. Hegardt,F.G. (1998) Proc. Natl Acad. Sci. USA, 95, 12185–12190. 13. Graveley,B.R. and Maniatis,T. (1998) Mol. Cell, 1, 765–771. 35. Holstege,F.C.P., Jennings,E.G., Wyrick,J.J., Lee,T.I., Hengartner,C.J., 14. Blencowe,B.J., Issner,R., Nickerson,J.A. and Sharp,P.A. (1998) Green,M.R., Golub,T.R., Lander,E.S. and Young R.A. (1998) Cell, 95, Genes Dev., 12, 996–1009. 717–728. 15. Liu,H.-X., Zhang,M. and Krainer,A.R. (1998) Genes Dev., 12, 36. Matsunaga,E., Umeno,M. and Gonzalez,F.J. (1990) J. Mol. Evol., 30, 1998–2012. 155–169. 16. Nigro,J.M., Cho,K.R., Fearon,E.R., Kern,S.E., Ruppert,J.M., Oliner,J.D., 37. Blumenthal,T. (1998) Bioessays, 20, 480–487. Kinzler,K.W. and Vogelstein,B. (1991) Cell, 64, 607–613. 38. Jarrell,K.A. (1993) Proc. Natl Acad. Sci. USA, 90, 8624–8627. 17. Cocquerelle,C., Mascrez,B., Hétuin,D. and Bailleul,B. (1993) FASEB J., 39. Neugebauer,K.M. and Roth,M.B. (1997) Genes Dev., 11, 3279–3285. 7, 155–160. 40. Zaphiropoulos,P.G., Undén,A.B., Rahnama,F., Hollingsworth,R.E. and 18. Capel,B., Swain,A., Nicolis,S., Hacker,A., Walter,M., Koopman,P., Toftgård,R. (1999) Cancer Res., 59, 787–792. Goodfellow,P. and Lovell-Badge,R. (1993) Cell, 73, 1019–1030. 41. Romkes,M., Faletto,M.B., Blaisdell,J.A., Raucy,J.L. and Goldstein,J.A. 19. Bailleul,B. (1996) Nucleic Acids Res., 24, 1015–1019. (1993) Biochemistry, 32, 1390 (correction).