Title: Excision of the Sinorhizobium meliloti Group II Intron RmInt1 as Circles in Vivo
Abstract: Excision of group II introns as circles has been described only for a few eukaryotic introns and little is known about the mechanisms involved, the relevance or consequences of the process. We report that splicing of the bacterial group II intron RmInt1 in vivo leads to the formation of both intron lariat and intron RNA circles. We determined that besides being required for the intron splicing reaction, the maturase domain of the intron-encoded protein also controls the balance between lariat and RNA intron circle production. Furthermore, comparison with in vitro self-splicing products indicates that in vivo, the intron-encoded protein appears to promote the use of a correct EBS1/IBS1 intron-exon interaction as well as cleavage at, or next to, the expected 3′ splice site. These findings provide new insights on the mechanism of excision of group II introns as circles. Excision of group II introns as circles has been described only for a few eukaryotic introns and little is known about the mechanisms involved, the relevance or consequences of the process. We report that splicing of the bacterial group II intron RmInt1 in vivo leads to the formation of both intron lariat and intron RNA circles. We determined that besides being required for the intron splicing reaction, the maturase domain of the intron-encoded protein also controls the balance between lariat and RNA intron circle production. Furthermore, comparison with in vitro self-splicing products indicates that in vivo, the intron-encoded protein appears to promote the use of a correct EBS1/IBS1 intron-exon interaction as well as cleavage at, or next to, the expected 3′ splice site. These findings provide new insights on the mechanism of excision of group II introns as circles. Group II introns are large catalytic RNAs with a conserved secondary structure consisting of six domains, one of which (dIV) may incorporate the coding sequence of a reverse transcriptase (RT) 3The abbreviations used are: RT, reverse transcriptase; IEP, intron-encoded protein; RNP, RNA-protein; IBS, intron-binding sites; EBS, exon-binding sites; nt, nucleotide. 3The abbreviations used are: RT, reverse transcriptase; IEP, intron-encoded protein; RNP, RNA-protein; IBS, intron-binding sites; EBS, exon-binding sites; nt, nucleotide. (1Michel F. Ferat J.L. Annu. Rev. Biochem. 1995; 64: 435-461Crossref PubMed Scopus (488) Google Scholar). Although some group II introns self-splice in vitro, this reaction requires nonphysiological conditions, and in vivo, proteins are required to fold the intron RNA into a catalytically active structure. Group II intron-encoded proteins (IEPs) promote both splicing and mobility of the intron RNA through formation of a specific RNA-protein (RNP) complex (2Lambowitz A.M. Belfort M. Annu. Rev. Biochem. 1993; 62: 587-622Crossref PubMed Scopus (532) Google Scholar, 3Lambowitz A.M. Caprara M.G. Zimmerly S. Perlman P.S. Cech T.R. Atkins J.F. 2nd Ed. The RNA World. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY1999: 451-485Google Scholar, 4Lambowitz A.M. Zimmerly S. Annu. Rev. Genet. 2004; 38: 1-35Crossref PubMed Scopus (351) Google Scholar). The IEPs have two conserved domains, an N-terminal RT domain and domain X, a putative RNA-binding domain associated with maturase (RNA splicing) activity. In some cases the IEP also includes a C-terminal DNA-binding and a DNA-endonuclease domain (5San Filippo J. Lambowitz A.M. J. Mol. Biol. 2002; 324: 933-951Crossref PubMed Scopus (77) Google Scholar). Group II introns splice typically by the same two sequential transesterification reactions used in nuclear mRNA splicing (1Michel F. Ferat J.L. Annu. Rev. Biochem. 1995; 64: 435-461Crossref PubMed Scopus (488) Google Scholar). In a first step, the 2′-OH group of a branch point nucleotide residue, usually a bulged adenosine in dVI, attacks the 5′ splice junction resulting in cleavage of the 5′ exon and the formation of an intron-3′ exon-branched lariat intermediate. The released 5′ exon remains associated to the intron via base pairing of the intron-binding sites (IBS1 and IBS2) to the exon-binding sites (EBS1 and EBS2) located in domain dI. In the second step, the free 3′-OH of the 5′ exon attacks the 3′ splice junction leading to the release of the intron lariat and the ligation of the 5′ and 3′ exons. There also exists an alternate pathway, in which the first splicing step is initiated by a nucleophilic attack of water or an OH– ion, resulting in the formation of a linear intron-3′ exon intermediate that subsequently participates in a normal second step reaction. This hydrolytic pathway is observable in vitro and has been shown to be used in vivo by yeast mitochondrial introns carrying branch-site mutations (6Podar M. Chu V.T. Pyle A.M. Perlman P.S. Nature. 1998; 391: 915-918Crossref PubMed Scopus (86) Google Scholar). Moreover, even though most of the plant chloroplast group II introns follow the typical lariat-generating pathway, hydrolytic splicing has been reported for intron trnV, which lacks the conserved bulged A in dVI, whereas both the hydrolytic and branching pathways seem to coexist in the case of the barley trnK intron (7Vogel J. Börner T. EMBO J. 2002; 21: 3794-3803Crossref PubMed Scopus (72) Google Scholar). Circular DNA versions of group II introns are known to exist in Podospora anserina where they are somehow associated with cellular senescence (8Osiewacz H.D. Esser K. Curr. Genet. 1984; 8: 299-305Crossref PubMed Scopus (160) Google Scholar, 9Schmidt W.M. Schweyen R.J. Wolf K. Mueller M.W. J. Mol. Biol. 1994; 243: 157-166Crossref PubMed Scopus (30) Google Scholar, 10Begel O. Boulay J. Albert B. Dufour E. Sainsart-Chanet A. Mol. Cell. Biol. 1999; 19: 4093-4100Crossref PubMed Scopus (43) Google Scholar). In this particular case a model was proposed, by which the circular DNA molecules would be generated via transposition of the intron in front of itself, followed by excision of one of the tandem copies by homologous recombination (11Sellem C.H. Lecellier G. Belcour L. Nature. 1993; 366: 176-178Crossref PubMed Scopus (98) Google Scholar, 12Müeller M.W. Allmaier M. Eskes R. Schweyen R.J. Nature. 1993; 366: 174-176Crossref PubMed Scopus (85) Google Scholar, 13Sainsard-Chanet A. Begel O. Belcour L. J. Mol. Biol. 1994; 242: 630-643Crossref PubMed Scopus (16) Google Scholar). More recent data indicate that yeast intron aI5γ can be also excised as a true circular form in vitro (14Murray H.L. Mikheeva S. Coljee V.W. Turczyk B.M. Donahue W.F. Bar-Shalom A. Jarrell K.A. Mol. Cell. 2001; 8: 201-211Abstract Full Text Full Text PDF PubMed Scopus (44) Google Scholar). The circle seems to result from formation of a 2′-5′ phosphodiester bond between the last and first residues of the intron and it has been proposed that this reaction requires the 3′ exon to have been released first from precursor molecules by a trans-splicing reaction triggered by 5′ exon molecules previously generated by the so-called SER (spliced-exon reopening) reaction. For the yeast intron aI2, both RNA and DNA circles have been detected in vivo and it is thought that circular aI2 RNAs are copied into DNA, presumably by reverse transcription (14Murray H.L. Mikheeva S. Coljee V.W. Turczyk B.M. Donahue W.F. Bar-Shalom A. Jarrell K.A. Mol. Cell. 2001; 8: 201-211Abstract Full Text Full Text PDF PubMed Scopus (44) Google Scholar). In addition, the presence of circular forms of excised intron molecules in plant mitochondria has been recently reported (15Li-Pook-Than J. Bonen L. Nucleic Acids Res. 2006; 34: 2782-2790Crossref PubMed Scopus (58) Google Scholar). However, circular intron molecules have not been described so far for bacterial group II introns. RmInt1 is a bacterial group II intron identified in Sinorhizobium meliloti, the nitrogen-fixing symbiont of alfalfa (Medicago sativa). The RmInt1 IEP is required for intron splicing in vivo (16Muñoz-Adelantado E. San Filippo J. Martínez-Abarca F. García-Rodríguez F.M. Lambowitz A.M. Toro N. J. Mol. Biol. 2003; 327: 931-943Crossref PubMed Scopus (36) Google Scholar), but like those of many other bacterial group II introns, it lacks C-terminal DNA endonuclease and DNA-binding domains (5San Filippo J. Lambowitz A.M. J. Mol. Biol. 2002; 324: 933-951Crossref PubMed Scopus (77) Google Scholar, 17Martínez-Abarca F. García-Rodríguez F.M. Toro N. Mol. Microbiol. 2000; 35: 1405-1412Crossref PubMed Scopus (58) Google Scholar, 18Zimmerly S. Hausner G. Wu X-C. Nucleic Acids Res. 2001; 29: 1238-1250Crossref PubMed Scopus (152) Google Scholar, 19Dai L. Zimmerly S. Nucleic Acids Res. 2002; 30: 1091-1102Crossref PubMed Scopus (144) Google Scholar, 20Toro N. Environ. Microbiol. 2003; 5: 143-151Crossref PubMed Scopus (73) Google Scholar). RmInt1 is nevertheless an efficient mobile element that has two retrohoming pathways for mobility, with predominant use of the nascent lagging strand at DNA replication forks for priming (21Martínez-Abarca F. Barrientos-Durán A. Fernández-López M. Toro N. Nucleic Acids Res. 2004; 32: 2880-2888Crossref PubMed Scopus (51) Google Scholar). Recently, we reported that RmInt1 self-splices in vitro in the absence of the IEP, but the in vitro activity of the intron is atypical in that the second step of splicing is unusually inefficient and several unconventional products are generated alongside the expected excised intron and ligated exons (22Costa M. Michel F. Molina-Sánchez M.D. Martinez-Abarca F. Toro N. Biochimie (Paris). 2006; 88: 711-717Crossref PubMed Scopus (17) Google Scholar). In this study, we have investigated the excision of the S. meliloti RmInt1 intron in vivo. Our data indicate that the RmInt1 group II intron is excised in vivo both as intron lariat and intron circles and that the IEP seems to determine the balance between these two excision products. Bacterial Strains, Media, and Growth Conditions—The strain used in this work was S. meliloti RMO17. This strain was cultured at 28 °C on TY medium for RNA extraction and RNP particle isolation. Escherichia coli DH5α was used for the cloning and maintenance of plasmid constructs. For plasmid maintenance the antibiotic kanamycin was added at 200 μg/ml–1 for rhizobia and 50 μg/ml–1 for E. coli. RmInt1 and Mutant Derivatives—The pKG2.5, pKG2.5-YAHH, and pKG2.5D5-CGA constructs have been previously reported (16Muñoz-Adelantado E. San Filippo J. Martínez-Abarca F. García-Rodríguez F.M. Lambowitz A.M. Toro N. J. Mol. Biol. 2003; 327: 931-943Crossref PubMed Scopus (36) Google Scholar, 17Martínez-Abarca F. García-Rodríguez F.M. Toro N. Mol. Microbiol. 2000; 35: 1405-1412Crossref PubMed Scopus (58) Google Scholar) and their main features are specified under “Results.” The RmInt1 IEP maturase mutants were generated by site-directed mutagenesis. For the construction of the YY → AA mutant, two pairs of primers were designed to amplify the 5′ and 3′ sections of the IEP, respectively. A 5′ end primer mut UP (5′-GTCAGCGGTGCTGGAAGTATG-3′) and a 3′ end primer YY-AA/DN (5′-CTGTACCGTCCCGCGGCGGCAATCCATCCCCGAAGG-3′) were used to generate the upstream fragment containing 690 bp; a 5′ end primer YY-AA/UP (5′-GATGGATTGCCGCCGCGGGACGGTACAGTCGTTCGG-3′) and a 3′ end primer mut DN (5′-GCGCGCGTAATACGACTCAC-3′) were used to generate the downstream 829-bp fragment. The mutagenic primers contain an overlapping region of 28 bp and promote the replacement of two conserved tyrosines (YY) at IEP positions 354–355 by two alanines (AA) through changing TACTAT to GCCGCG in these positions. The final 1492-bp amplified fragment was digested with EcoRI and SpeI and used to replace the corresponding wild-type fragment in pKG2.5, which generated plasmid pKG2.5-A354A355. The K381A maturase mutant was constructed by site-directed mutagenesis using the Altered Sites II in vitro Mutagenesis pAlter-1 System (Promega). In pAL2.5-A381, which was derived from pAL2.5 (16Muñoz-Adelantado E. San Filippo J. Martínez-Abarca F. García-Rodríguez F.M. Lambowitz A.M. Toro N. J. Mol. Biol. 2003; 327: 931-943Crossref PubMed Scopus (36) Google Scholar), RmInt1 nucleotides AAG located at positions 1687 to 1689 are replaced by GCG, so that the conserved IEP residue Lys381 is changed to Ala381. The final construct, pKG2.5-A381, was generated by cloning the RmInt1-containing fragment resulting from BamHI/SpeI digestion of pAL2.5-A381 into pKG0 (17Martínez-Abarca F. García-Rodríguez F.M. Toro N. Mol. Microbiol. 2000; 35: 1405-1412Crossref PubMed Scopus (58) Google Scholar). Primer Extension Assays—These assays were carried out on both total RNA and RNP particle preparations. Primer extension reactions were essentially as described previously (16Muñoz-Adelantado E. San Filippo J. Martínez-Abarca F. García-Rodríguez F.M. Lambowitz A.M. Toro N. J. Mol. Biol. 2003; 327: 931-943Crossref PubMed Scopus (36) Google Scholar). Samples were resolved on a denaturing 6% polyacrylamide gel. RNA Isolation—RNA was extracted from free-living cultures of S. meliloti strain RMO17 containing plasmids expressing wild-type or mutant RmInt1 grown in 10 ml of TY medium supplemented with kanamycin until they reached an A600 of 0.6 units. The cells were pelleted, resuspended, and incubated for 10 min at 65 °C in 300 μl of a lysis solution (1.4% SDS, 4 mm EDTA). Proteins were removed by adding 150 μl of NaCl (5 m) at 4 °C for 10 min. After centrifugation, the nucleic acids present in the supernatant were precipitated by adding 1 ml of ethanol (100%). The pellet was resuspended in 85 μl of nucleasefree water and digested with 50 units of RNase-free DNase I (Roche). The RNA was extracted with 1 volume of a 25:24:1 mixture of phenol (pH 4.5)/chloroform/isoamyl alcohol and then extracted with 1 volume of a 24:1 mixture of chloroform/isoamyl alcohol. Finally the RNA was precipitated with 3 volumes of ethanol (100%) and 75 mm NaOAc (pH 5.2). The RNA pellet was washed with 70% ethanol, dried, and resuspended in 20 μl of nuclease-free water. For the RT-PCR assays shown in Fig. 5B, RNA preparations were set up with a few modifications. Proteinase K (25 μl of a solution of 5 mg/ml) was added to the lysis solution followed by the procedure described above, or after DNase I digestion the sample was treated with 25 μlof5 m guanidine thiocyanate, 2.5 μlof3 m NaOAc (pH 4.8), 93 μlof phenol (pH 4.5), and 30 μl of chloroform/isoamyl alcohol (49:1) for 15 min on ice and then extracted with chloroform/isoamyl alcohol (24:1). Preparation of RNP Particles—RNP isolation was carried out essentially as described previously (16Muñoz-Adelantado E. San Filippo J. Martínez-Abarca F. García-Rodríguez F.M. Lambowitz A.M. Toro N. J. Mol. Biol. 2003; 327: 931-943Crossref PubMed Scopus (36) Google Scholar). The clarified lysate (5 ml) was layered on a 5 ml of 1.85 m sucrose cushion and centrifuged for 20 h in a Beckman Ti50 rotor at 50,000 × g. A supernatant with two different fractions according to its density (S1, light and S2, dense) and a pellet (P) containing the RNPs was obtained. For the RT-PCR assays shown in Fig. 5A, the two fractions of the supernatant (∼5 ml each) were collected and precipitated. The S2 fraction was precipitated with 3 volumes of ethanol (100%) and 0.8 m LiCl. The resulting pellet was resuspended in water and precipitated again with 3 volumes of ethanol and 0.25 m NaOAc (pH 5.2). The S1 fraction was mixed with 1.5 volumes of a solution containing 0.8 m guanidine thiocyanate, 50 mm NaOAc (pH 4.8), and phenol (pH 4.5), chloroform, isoamyl alcohol (150:49:1). The mixture was kept on ice for 15 min and centrifuged at 13,000 × g for 15 min. The supernatant was extracted with chloroform/isoamyl alcohol (24:1) followed by precipitation with 1 volume of isopropyl alcohol. The resulting pellet was resuspended in water and digested with 100 units of RNase-free DNase I (Roche) followed by phenol (pH 4.5)/chloroform/isoamyl alcohol (25:24:1) extraction and isopropyl alcohol precipitation. RT-PCR and cRT-PCR Assays—First strand cDNA synthesis was started by annealing 6.5 μg of total cellular RNA or an equivalent quantity of RNPs (0.1625 A260 units) and 25 pmol of Ect1 primer complementary to a sequence 188 nt from the 5′ end of RmInt1 in the presence of 12.5 nm dNTPs equimolar mixture. The mixture was first heated at 90 °C for 2 min, then slowly cooled for 15 min prior to being put on ice for 15 min. cDNA synthesis was triggered by addition of 5× first strand reverse transcriptase buffer (Invitrogen), 0.2 mol of dithiothreitol, 30 units of RNAguard™ RNase inhibitor (Amersham Biosciences), and 400 units of SuperScript II RNase H– reverse transcriptase (Invitrogen) for 120 min. After heat inactivation of the enzyme (at 70 °C for 15 min), RNase H digestion was carried out for 20 min at 37 °C. One-fifteenth of that reaction was used as a template in the PCR with 30 pmol of LL primer (5′-GAGGTTCACGCACCGTTCTG; designed complementary to a sequence 59–40 nt from the 3′ end of RmInt1), 30 pmol of 5′-radiolabeled P primer (around 200,000 cpm), 12.5 mmol of dNTPs mixture, 50 mm HEPES (pH 7.9), 1.5 mm MgCl2, 50 mm KCl, and 2 units of Taq polymerase in 50 μlasa final volume. After the preincubation step at 94 °C for 3 min, 35 cycles was performed: 45 s at 94 °C, 30 s at 63 °C, 30 s at 72 °C, and a final extension at 72 °C for 10 min. One μl of RT-PCR product was separated on a denaturing 6% polyacrylamide gel. Subsequent cloning in pGEM-T easy vector (Promega) and sequence analysis of the RT-PCR products was carried out by isolation of the corresponding bands from agarose gels. The cRT-PCR experiments were carried out as the RT-PCR assays, but using previously ligated RNA samples to circularize any linear intron molecules that contain a 5′ monophosphate. Twenty-five μg of total RNA or 0.625 A260 units of RNP particle preparations from wild-type intron pKG2.5 cells were added to a solution containing 50 mm Tris-HCl (pH 7.8), 10 mm MgCl2, 10 mm dithiothreitol, 1 mm ATP, and 50 units of T4 RNA ligase (New England Biolabs) in a final volume of 20 μl and the mixture was incubated at 37 °C for 2 h. Subsequently, T4 RNA ligase was inactivated by boiling for 2 min and the reaction underwent phenol and chloroform extractions followed by ethanol precipitation for 1 h at –80 °C. Finally, samples were resuspended in 5 μl of diethyl pyrocarbonate/water and 2.5 μl were used for reverse transcription. RT Assays—Exogenous RT activity in RNP particles was measured as previously described (16Muñoz-Adelantado E. San Filippo J. Martínez-Abarca F. García-Rodríguez F.M. Lambowitz A.M. Toro N. J. Mol. Biol. 2003; 327: 931-943Crossref PubMed Scopus (36) Google Scholar). RmInt1 Excised Products in Vivo as Determined by Primer Extension—Because the RmInt1 ribozyme has a bulged A located seven nucleotides upstream of the 3′ splice site, we expected the intron to splice primarily as a lariat in vivo. Initially, we characterized the RmInt1 splicing reaction by primer extension using a primer P (Fig. 1) complementary to a sequence located 80–97 nt from the 5′ end of the intron (16Muñoz-Adelantado E. San Filippo J. Martínez-Abarca F. García-Rodríguez F.M. Lambowitz A.M. Toro N. J. Mol. Biol. 2003; 327: 931-943Crossref PubMed Scopus (36) Google Scholar). Interestingly, primer extension assays using RNA and RNP particles from S. meliloti cells expressing wild-type intron pKG2.5 revealed two extension products (along with larger bands) that differ by one nucleotide (97–98 nt; Fig. 1) relative to the position expected for the 5′ end of the intron. Note that the shorter 97-nt extension product is clearly a major product in both RNA and RNP particle preparations. None of these two extension products were detected with RNA or RNPs isolated from cells harboring the splicing-defective intron pKG2.5D5-CGA, which has a mutation (GUU → GCA) in the AGC-GUU critical conserved pairing of ribozyme domain dV (Fig. 1, lanes 2 and 4). Therefore, we conclude that the 97- and 98-nt extension products were derived from excised intron RNA molecules. Whereas primer extension data do not by themselves make it possible to clarify the nature of the excised intron RNA, it is interesting to note that primer extension performed on a circular intron has been reported to result in two extension products that differ by 1 nucleotide, the longer product being the major one (14Murray H.L. Mikheeva S. Coljee V.W. Turczyk B.M. Donahue W.F. Bar-Shalom A. Jarrell K.A. Mol. Cell. 2001; 8: 201-211Abstract Full Text Full Text PDF PubMed Scopus (44) Google Scholar). Reverse transcriptase is known to pause after reading through the 2′-5′ linkage in a nonlariat RNA (23Lorsch J.R. Bartel D.P. Szostak J.W. Nucleic Acids Res. 1995; 23: 2811-2814Crossref PubMed Scopus (57) Google Scholar), whereas with lariat RNA, the pause occurs before reading through the 2′-5′ linkage (14Murray H.L. Mikheeva S. Coljee V.W. Turczyk B.M. Donahue W.F. Bar-Shalom A. Jarrell K.A. Mol. Cell. 2001; 8: 201-211Abstract Full Text Full Text PDF PubMed Scopus (44) Google Scholar). Thus, one possible interpretation of our data were that in addition to intron lariat, RmInt1 might be excised in vivo as intron circles. Detection of Two Different RmInt1 Excision Products in Vivo by Reverse Transcription and PCR—To further investigate the nature of RmInt1 excision products, RNA and RNP particle preparations were analyzed by reverse transcription with primer Ect1 (Fig. 2A), located downstream of the 5′ splice site, followed by PCR with primers P and LL, the latter located 5′ of the branch site (Fig. 2A). The PCR products were radiolabeled using a 5′ end-labeled P primer and the products were separated on a denaturing, 6% (w/v) polyacrylamide gel (Fig. 2B). When RNA samples were used (Fig. 2B), two major products of 265 and 156 nt were observed in cells expressing the wild-type intron pKG2.5 (lane 2) or the defective homing mutant pKG2.5-YAHH (lane 4), which has a YADD to YAHH mutation in the RT domain and therefore cannot generate duplicated linear introns joined head-to-tail. The 265-nt product was also present in the splicing-defective mutant pKG2.5D5-CGA (lane 6), indicating that it was not derived from the spliced intron; as shown later, this product derives from the RNA precursor. Strikingly, the 156-nt product, whose size is that expected from a circular RNA substrate in which the first G residue of the intron is linked to the last C residue, was absent when RNA from the splicing-defective mutant pKG2.5-D5-CGA was used (lanes 6) or when reverse transcription was omitted (Fig. 2, lanes 1, 3, and 5). These results indicate that the RT-PCR product of 156 nt obtained using RNA samples is primarily derived from an excised intron RNA product. The 156-nt RT-PCR products were also detected when preparations of RNP particles from cells expressing the wild-type intron pKG2.5 were used (Fig. 2B, lane 8). Consistent with the data reported above, the 156-nt product was not seen when mutant pKG2.5-D5-CGA RNP particles were used (lane 12). This product was clearly reduced in RNP particles from the RT mutant pKG2.5-YAHH (lane 10) that has ∼60% of the wild-type splicing activity (not shown). Interestingly, in the absence of reverse transcription and the primer Ect1, a PCR product of 156 nt was still clearly detected in wild-type intron pKG2.5 RNP particles (lane 7), but was not observed in the splicing defective mutant pKG2.5-D5-CGA (lane 11). These results appear to indicate the existence of DNA molecules generated from spliced intron circles. In addition to the aforementioned products, in the RNP particles, but not in the RNA samples (Fig. 2B), we detected another major RT-PCR product of 150 nt, whose size is that expected from a substrate consisting of an intron lariat in which the branching point is the bulged A in dVI. This product was observed in both wild-type intron pKG2.5 (lane 8) and the RT mutant pKG2.5-YAHH (lane 10), but was absent in the splicing-defective mutant pKG2.5D5-CGA (lane 12) or when reverse transcription was omitted (lanes 7, 9, and 11). These results imply that the RT-PCR product of 150 nt is generated from the amplification of a second intron RNA excision product. RmInt1 Excision in Vivo Produces Both Lariat and Intron Circles—The amplified RT-PCR products from both RNA and RNP particle preparations were gel-extracted, cloned, and sequenced. All 19 clones generated by the 265-nt RT-PCR product from wild-type intron pKG2.5 and RT mutant pKG2.5-YAHH, and either RNA or RNP particles, showed 168 nucleotides of the 5′ exon joined to 97 nucleotides of the 5′-end of the intron. A similar result was obtained by sequencing nine clones from a 265-nt PCR product obtained in the absence of reverse transcription using RNP particle preparations from RT mutant pKG2.5-YAHH. Therefore, we conclude that as stated above, this amplification product is generated from intron RNA precursor and/or contaminant DNA. The RT-PCR products of 156 nt generated from RmInt1-excised intron using RNA samples (Fig. 2B) were also cloned and sequenced. Ten of 11 clones from the RT mutant pKG2.5-YAHH were derived indeed from intron RNA circles and displayed the expected point of circular ligation between the first and last residues of the intron (Fig. 2C), whereas the remaining clone (157 nt) corresponded to a molecule with an extra cytosine at the site of circularization. Similar results were obtained when the same RT-PCR product was sequenced from wild-type intron pKG2.5. Five of 9 clones corresponded to the 156-nt intron circle product, whereas the remaining four displayed an additional cytosine (157 nt). Together, these findings are consistent with the production of intron RNA circles in vivo. The PCR product of 156 nt obtained in the absence of reverse transcription and primer Ect1 when wild-type intron pKG2.5 RNP particles were used (Fig. 2B, lane 7) was also cloned and sequenced. All six clones analyzed had the same sequence as the RT-PCR product of 156 nt generated from the RNA samples. We interpret the latter molecules as evidence that in addition to RmInt1 intronic circular RNA, there exist intronic DNA circles. In agarose gels, the coexisting RT-PCR products of 150 and 156 nt (Fig. 2B, lanes 8 and 10) generated by amplification of RNP particle samples from either wild-type intron pKG2.5 or RT mutant pKG2.5-YAHH cannot be resolved (data not shown). Therefore, the amplified products were isolated together, cloned, and sequenced. Eleven of 13 clones from wild-type intron pKG2.5 showed the expected linkage between the bulged A of dVI and the first residue of the intron (Fig. 2C), indicating that they arose from a lariat intron, whereas the remaining two correspond to intron circles. Note that in the products amplified from lariat intron (Fig. 2C), there was misincorporation of adenosine instead of thymidine as reverse transcriptase encountered the branched nucleotide (7Vogel J. Börner T. EMBO J. 2002; 21: 3794-3803Crossref PubMed Scopus (72) Google Scholar, 24Vogel J. Hess W.R. Börner T. Nucleic Acids Res. 1997; 25: 2030-2031Crossref PubMed Scopus (78) Google Scholar). Consistent with the higher proportion of the 150-nt RT-PCR product in the RT mutant pKG2.5-YAHH (Fig. 2B, lane 10), all 10 derivative clones had the sequence expected from a lariat intron. Together, these findings imply that excision of RmInt1 in vivo produces intron lariat as well as intronic RNA and DNA circles. RmInt1 RNA Molecules Detected by cRT-PCR—It has been suggested that intron circle formation by intron aI5γ could result from the previous release of the 3′ exon by free 5′ exon molecules arising from the exon reopening reaction (splicedexon reopening) (14Murray H.L. Mikheeva S. Coljee V.W. Turczyk B.M. Donahue W.F. Bar-Shalom A. Jarrell K.A. Mol. Cell. 2001; 8: 201-211Abstract Full Text Full Text PDF PubMed Scopus (44) Google Scholar). RmInt1 circles might also be originated by ligation of the ends of a previously linear intron molecule produced by the hydrolytic pathway in which the first splicing step results in the formation of a linear intron-3′ exon intermediate. To test for the possible presence of such intermediates in the RNA population, the RNA was circularized using T4 RNA ligase prior to reverse transcription with primer Ect1 and PCR with primers P and LL (cRT-PCR) (7Vogel J. Börner T. EMBO J. 2002; 21: 3794-3803Crossref PubMed Scopus (72) Google Scholar). Because the resulting products from both total RNA and RNP particles form a smear in polyacrylamide gels (not shown), they were cloned directly from the PCR. As shown in Fig. 3, sequence analysis of pKG2.5 samples (8 clones from total RNA and 8 clones from RNP particles) revealed a predominance of molecules in which the 5′ exon was ligated either to the exact 3′ end of the intron (C-1884) or to one of the nucleotides neighboring the 3′ splice site (9 of 16 clones). Moreover, the 5′ end of the RNA precursor seems to be primarily processed around nucleotide –30. Five other clones contained the intron flanked by 5′ and 3′ exon sequences, presumably representing intron RNA precursor molecules, whereas the remaining two clones likely corresponded to molecules degraded either at the 5′ or 3′ end. These results are consistent with the suggestion that the release of the 3′ exon might be the first step for RmInt1 circle formation. The Maturase Domain Is Required for RmInt1 Splicing and Intron Circle Formation—In group II IEPs, the RT domain is followed by a region denoted domain X, affecting the maturase (RNA splicing) activity. To assess the involvement of the maturase domain of RmInt1 in the splicing reaction in vivo and intron circle formation, we created the mutant YY → AA, which has two alanine residues instead of the conserved Tyr354 and Tyr355 residues, and the mutant K381A, in which the conserved Lys381 residue is replaced by an alanine residue (Fig. 4A). Primer extension analysis and RT-PCR assays using both RNA and RNP particle preparations indicate that the YY → AA mutant shows no detectable splicing (Fig. 4, B and D). Moreover, the RNP particles from cells expressing pKG2.5-A354A355 had strongly decreased RT activity similar to that of the RT YAHH mutant (Fig. 4C), consistent with the inability of these mutants to form active RNP particles. As expected from these data, mobility is abolished in the YY → AA mutant (data not shown). These findings indicate that as expected the RmInt1 maturase X domain is required for the splicing reaction in vivo, and also for RNA intron circle formation. As shown in Fig. 4B, the single point mutation K381A inhibits, but does not abolish splicing (∼30% wild-type). Also, the RNP particles of cells expressing pKG2.5-A381 have strongly decreased RT activity (Fig. 4C) and the K381A mutant has no detectable mobility (data not shown). Interestingly, RT-PCR assays using RNP particle preparations from this mutant (Fig. 4D) showed an apparent bias to intron circle formation (compare the levels of products derived from the lariat and intron circles between the mutant, lane 10, and the wild type, lane 12). Moreover, the RT-PCR product of the K381A mutant corresponding to intron circles appears to be primarily one nucleotide longer than that obtained from the wild-type intron. This was confirmed by sequencing the cloned RT-PCR product, which showed an extra C residue at the circle ligation point. The apparent shift from lariat to intron circles might only reflect a decreased affinity of the mutant RT-maturase for the lariat product leading to a reduction of lariat intron in the RNP particles. To assess this possibility the RNA from supernatant fractions S1 and S2 after the RNPs (P) pelleting step were also analyzed by RT-PCR (Fig. 5A). In the case of the wild-type intron (pKG2.5), the RT-PCR product derived from the lariat was the prominent product in both the RNPs (P) and the dense fraction of the supernatant (S2), whereas in the light fraction of the supernatant (S1) similar levels of RT-PCR products derived from lariat and intron circles was observed. Interestingly, in the last fraction (S1) intronic DNA circles were detected (reverse transcription and primer Ect1 omitted, lane 1). On the contrary, in the K381A mutant the major RT-PCR product in the RNPs (P) and in both supernatant fractions (S1 and S2) was derived from intron circles. These findings are consistent with a genuine shift from lariat to intron circle production in the K381A mutant. Nevertheless, this conclusion was still hampered by the fact that in the total RNA preparations we had only detected intron circles by RT-PCR, hence we could not differentiate the mutant from the wild-type intron. A possible explanation is that the lariat RNA could be degraded or the RNP particles eliminated in some way during the RNA isolation. To assess this possibility we also obtained total RNA using methods that included specific steps of denaturation or removal of proteins. As shown in Fig. 5B either the use or guanidine thiocyanate or proteinase K in the RNA isolation procedure allowed us to detect both lariat and intronic circles in the RNA samples. As expected the RT-PCR product derived from lariat molecules is the predominant product in the wild-type intron pKG2.5, but again the RT-PCR product derived from intron circles was the major product in the K381A mutant. Thus, taken together these findings indicate that the maturase domain X of the IEP controls the balance between lariat and RNA intron circle production. We show here that S. meliloti group II intron RmInt1 is excised in vivo both as intron lariat and intron circles and that the maturase domain of the IEP is not only required for intron RNA splicing, but plays a role in the mechanism chosen for intron excision. RmInt1 Is Excised as a Lariat but Also as Intron Circles in Vivo—By RT-PCR and sequencing, we found that excision of the RmInt1 wild-type intron from its RNA precursor produces both lariat and RNA circular molecules. RNA intron circles are also produced by a mutant in the RT active site (YAHH) and by the maturase mutant K381A. Because these mutant introns are not mobile and therefore, cannot generate intron dimers arranged head-to-tail, we believe that PCR products in which the intron 5′ and 3′ ends are joined together were not generated by transcription of intron duplications, but rather reflect the presence of genuine RNA circles resulting from intron excision. It should be added that RmInt1 also generates RNA circles in a RecA– background (data not shown), which adds further support to the above conclusion. Our data also suggest that not only RNA circles, but DNA circular molecules are produced as well by the wild-type RmInt1 intron; whether the latter molecules are generated through reverse transcription of the RNA circles by the IEP remains to be determined. The circle produced in vitro by the yeast intron a5γ seems to result from formation a 2′-5′ bond and this may also be the case for RmInt1 RNA circles. Although the absolute ratio of intron circle over lariat cannot be estimated from these assays because RT-PCR is bound to be influenced by the differing ability of reverse transcriptase to read through 2′-5′ linkages depending on whether these are part of branched (lariat) or unbranched (intron circle) structures (14Murray H.L. Mikheeva S. Coljee V.W. Turczyk B.M. Donahue W.F. Bar-Shalom A. Jarrell K.A. Mol. Cell. 2001; 8: 201-211Abstract Full Text Full Text PDF PubMed Scopus (44) Google Scholar, 23Lorsch J.R. Bartel D.P. Szostak J.W. Nucleic Acids Res. 1995; 23: 2811-2814Crossref PubMed Scopus (57) Google Scholar), the RT-PCR assays suggest that the lariat is the predominant excision product of wild-type intron RmInt1. The detection of lariat intron molecules in the RNA preparations, but not the circular forms depends on the use of specific steps of denaturation or removal of proteins during the RNA isolation. In addition, the RT-PCR analysis of the supernatant fractions and the pellet after the RNPs pelleting step suggests that the lariat associated with the IEP (RNPs) and the intronic circles have different hydrodynamic properties. These properties may be influenced by both the type of folding of the different excised intron molecules and a distinctive interaction of these forms with the IEP. The Role of the RmInt1 IEP in Splicing and Intron Excision as Circles in Vivo—A mutation in which the conserved Tyr354– Tyr355 residues of the RmInt1 IEP maturase domain were replaced by two alanine residues abolished formation of both RNA circles and intron lariat, thus implying that the maturase domain X is required for the two excision mechanisms in vivo. The mutation Y354A, Y355A may inhibit RNA splicing directly by altering some of the contacts of the IEP with the intron RNA or by decreasing the stability of the IEP. In the maturase mutant K381A splicing was reduced to 30% of the wild-type and there was a marked decrease of intron lariat molecules. This mutation may also alter the IEP interaction with the intron RNA, which in turn may affect the intron excision mechanism. Our results suggest that the maturase domain X controls in some way the balance of intron excision as lariat or intron circles. Moreover, we found that some of the amplified products presumably derived from RmInt1 RNA circles contain an extra cytosine at the circle ligation point. Noticeably, those products predominate among molecules amplified from the maturase mutant K381A. The addition of non-encoded nucleotides at the circle ligation point was observed for in vitro-generated circular aI5γ RNA (14Murray H.L. Mikheeva S. Coljee V.W. Turczyk B.M. Donahue W.F. Bar-Shalom A. Jarrell K.A. Mol. Cell. 2001; 8: 201-211Abstract Full Text Full Text PDF PubMed Scopus (44) Google Scholar) and the inserted residues were hypothesized to having been incorporated by the reverse transcriptase used in the assays when it encounters the putative 2′-5′ phosphodiester bond. However, a cytosine residue (IBS3) is also present at position +1 of the RmInt1 3′ exon. Thus, the extra C residue in the RmInt1 RNA circular molecules might arise from cleavage at position +1 (IBS3) in the 3′ exon, rather than being an artifact of the reverse transcription reaction. This plausible interpretation of our results implies that the maturase would help to determine the precise site of excision at the 3′ splice junction during intron circle formation. The Mechanism of RmInt1 Excision as Circles in Vivo—Whereas RmInt1 is also able to form RNA intron circles in the absence of protein by using a pseudo-IBS1 (IBS1*) sequence located close to the correct 3′ splice site in the 3′ exon (22Costa M. Michel F. Molina-Sánchez M.D. Martinez-Abarca F. Toro N. Biochimie (Paris). 2006; 88: 711-717Crossref PubMed Scopus (17) Google Scholar), the IEP seems to promote the correct EBS1-IBS1 interaction in vivo. Analysis of intron-containing RNA using T4 RNA ligase assays revealed a predominance of intermediates with a site of 3′ exon cleavage located at or next to the 3′ splice site. Thus, in vivo the presence of the IEP seems to ensure that cleavage occurs either at the correct 3′ splice site or at position +1 of the 3′ exon before intron RNA circle formation, rather than at position +10 or +11 as occurs in vitro (22Costa M. Michel F. Molina-Sánchez M.D. Martinez-Abarca F. Toro N. Biochimie (Paris). 2006; 88: 711-717Crossref PubMed Scopus (17) Google Scholar). After the release of the 3′ exon, which may result from a trans-splicing reaction as suggested for intron aI5γ (14Murray H.L. Mikheeva S. Coljee V.W. Turczyk B.M. Donahue W.F. Bar-Shalom A. Jarrell K.A. Mol. Cell. 2001; 8: 201-211Abstract Full Text Full Text PDF PubMed Scopus (44) Google Scholar), the 2′-OH of the terminal cytosine will attack the 5′ splice site, releasing the 5′ exon and an intron circle. Excision of group II introns as circles had been shown to occur in vivo for yeast introns like aI2 (14Murray H.L. Mikheeva S. Coljee V.W. Turczyk B.M. Donahue W.F. Bar-Shalom A. Jarrell K.A. Mol. Cell. 2001; 8: 201-211Abstract Full Text Full Text PDF PubMed Scopus (44) Google Scholar) and for some plant mitochondria introns (15Li-Pook-Than J. Bonen L. Nucleic Acids Res. 2006; 34: 2782-2790Crossref PubMed Scopus (58) Google Scholar), but is reported for the first time in this work for a mobile bacterial group II intron. It is now apparent that this peculiar mode of excision is more widespread in nature than it seemed and we regard it as likely that it will be found to have a biological role, perhaps with respect to intron spread and dissemination. We thank María Costa and François Michel for discussions and critical reading of the manuscript. We are grateful to Asunción Martos Tejera for technical assistance.