The Brown Algae Pl.LSU/2 Group II Intron-Encoded Protein Has Functional Reverse Transcriptase and Maturase Activities

Group II introns are self-splicing mobile elements found in prokaryotes and eukaryotic organelles. These introns propagate by homing into precise genomic locations, following assembly of a ribonucleoprotein complex containing the intron-encoded protein (IEP) and the spliced intron RNA. Engineered group II introns are now commonly used tools for targeted genomic modifications in prokaryotes but not in eukaryotes. We speculate that the catalytic activation of currently known group II introns is limited in eukaryotic cells. The brown algae Pylaiella littoralis Pl.LSU/2 group II intron is uniquely capable of in vitro ribozyme activity at physiological level of magnesium but this intron remains poorly characterized. We purified and characterized recombinant Pl.LSU/2 IEP. Unlike most IEPs, Pl.LSU/2 IEP displayed a reverse transcriptase activity without intronic RNA. The Pl.LSU/2 intron could be engineered to splice accurately in Saccharomyces cerevisiae and splicing efficiency was increased by the maturase activity of the IEP. However, spliced transcripts were not expressed. Furthermore, intron splicing was not detected in human cells. While further tool development is needed, these data provide the first functional characterization of the PI.LSU/2 IEP and the first evidence that the Pl.LSU/2 group II intron splicing occurs in vivo in eukaryotes in an IEP-dependent manner.


Introduction
Prokaryotic and eukaryotic organelle introns are mobile elements able to integrate specifically in the exon junction of an intronless genome [1][2][3]. This property, called ''homing'', contributes to the spreading of introns and has been used for precise in vivo genome engineering [4][5][6][7][8][9][10][11][12][13]. Two general homing mechanisms have been described depending on the type of intron: group I introns encode a very specific nuclease (meganuclease) to produce a double-strand break (DSB) in the intronless genome at the junction of exons. The DSB is then repaired by homologous recombination (HR) with a template DNA coming from the ''invader'' genome [14]. Group II introns are ribozymes that selfsplice from precursor RNA yielding excised intron lariat RNAs. The lariat splicing intermediate recognizes the DNA junction between the exons of the intronless genome and integrates the genome forming a DNA-RNA hybrid. After reverse transcription, the template intronic RNA is degraded and the gap is repaired by a DNA polymerase. In spite of very divergent mechanisms used for homing, both group I or group II introns rely on the expression of a protein coded by the intron itself (IEP, Intron-Encoded Protein) for the homing process [2]. These IEPs often carry different activities: maturase (to help the proper splicing of the intron) [15][16][17][18][19][20][21][22], double strand endonuclease (in group I introns) [17,18,20,22], single strand endonuclease and reverse transcriptase (in group II introns) [23][24][25][26][27][28]. Homing always results in the incorporation of a copy of the intron into the intronless genome.
The ability of group I and group II introns to recognize and to integrate into a specific genomic site has been exploited to generate various knock-out or knock-in models in mammalian cells [4,12,13,29,30], plants [31][32][33], and bacteria [5,[34][35][36][37][38]. Specific genomic targeting is obtained by changing the native target recognition sequences using rational engineering or directed molecular evolution [13,39]. At present, group II intron-derived genomic targeting strategies are only used in bacteria. However, using group II introns in mammalian cells could represent some advantages over the currently existing technologies. In mammalian cells, meganucleases and strategies based on FokI restriction endonuclease coupled to engineered Zinc fingers [32] or Transcription Activator-Like Effectors [40,41] are used for specific genomic insertion. Both of these approaches utilize DSB repair processes which occur mainly by non-homologous end joining (NHEJ) or by HR in the presence of a DNA template [42][43][44]. These approaches are limited by safety concerns and by low efficiency in some cell types. The radically different homing mechanism of group II introns could be an alternative to DSB mediated gene engineering.
There are only a few examples of the use of group II introns in eukaryotes. An initial proof of concept in mammalian cells has been reported by Guo et al. [4] using the Lactococcus lactis Ll.LtrB intron to target two genes carried into plasmids in HEK 293 human cells. These initial experiments showed homing into the plasmid as detected by PCR with specific oligonucleotides but efficacy was not measured. The splicing of the Ll.LtrB intron inside the HEK 293 cells was also not demonstrated since the authors introduced directly the purified LtrA-lariat ribonucleoparticles into the cell to obtain homing. However, it is known that the Ll.LtrB intron can acquire its correct tertiary structure in bacteria with the help of LtrA as used in commercial gene knockout systems [5,[45][46][47].
Several elements may contribute to the limitations in the use of group II introns in eukaryotes. Ribozyme activity of group II introns depends on the correct folding of RNA into a specific tertiary structure [48]. This activity is necessary for both the intron splicing and its insertion into target DNA. Most of group II introns are able to self-splice in vitro in the absence of proteins but have to be ''chaperoned'' by proteins in vivo [48,49]. These proteins may not function optimally in eukaryotic cells. Self-splicing of group II introns depends on the correct tertiary structure folding of the intronic RNA, and Mg 2+ cations contribute to this folding by stabilizing the RNA tertiary structure [48,50]. Mg 2+ is also required for catalysis of group II introns [51] and its concentration is critical for proper self-splicing in vitro. It is therefore possible that most group II intron cannot function optimally in eukaryotic cells where the free Mg 2+ concentration is estimated to be around 1-2 mM [52]. For example, the optimal in vitro reaction conditions for the well-known bacterial Lactococcus lactis Ll.LtrB intron IEPassisted RNA splicing use 5 mM Mg 2+ and 10 mM of Mg 2+ are required for an optimal reverse splicing reaction of RNPs into DNA target site [53]. The importance of Mg 2+ concentration is emphasized by a recent study showing that Ll.LtrB RNP particles were able to insert efficiently into a plasmid target in eukaryotic nuclei by providing MgCl 2 during the microinjection of RNPs in Xenopus laevis oocytes, and Drosophilia melanogaster and zebrafish (Danio renio) embryos [29]. In this work, we have chosen to study the Pl.LSU/2 intron from the mitochondrial large subunit rRNA gene of the brown algae Pylaiella littoralis because of its ability to self-splice in vitro at unusually low Mg 2+ concentrations (0.1 mM) [54]. We reasoned that this characteristic could make the Pl.LSU/ 2 intron a good candidate to be used as a tool for genome manipulation in eukaryotic cells. The Pl.LSU/2 intron structure presents a highly canonical secondary structure model consisting in six helical domains (I to VI) radiating from a central wheel [55,56]. Although the Pl.LSU/2 intron tertiary structure has been characterized [57][58][59], there is no report on the homing property of this intron nor on its potential biochemical activities in vivo. Moreover, the putative intron-encoded protein of PI.LSU/2 has not been described. The fact that the Pl.LSU/2 intron sequence presents in its domain IV an open reading frame containing putative domains of most group II intron IEPs [56] suggests that at least one of the Pl.LSU/2 IEP putative activities has been preserved by evolution.
We have therefore characterized the biochemical activities of the Pl.LSU/2 IEP as well as the Pl.LSU/2 intron activity in vivo in eukaryotic cells in order to evaluate the possibility of using this intron in genome engineering. Here, we show that active recombinant Pl.LSU/2 IEP can be produced and purified in Escherichia coli. The purified IEP shows a reverse transcriptase activity in vitro both alone or when expressed together with the intronic RNA. The Pl.LSU/2 intron is able to splice properly in yeast cells helped by the maturase activity of its IEP.

Plasmids and Oligonucleotides
Plasmids and oligonucleotides used in this study are listed in Supplemental information (Table S1 A and Table S1 B, respectively).
Expression of the Pl.LSU/2 Intron-encoded Protein (IEP) in E. coli E. coli strain Rosetta-gami B (DE3) was transformed with the appropriate expression plasmid and single colonies were inoculated into 4 ml of LB containing appropriate antibiotics. Precultures were shaken at 170 rpm at 32uC overnight, inoculated into 100 ml of LB medium without antibiotics and grown at 32uC for 3-6 hr, until OD 600 nm reached 0.4-0.8. Induction was started by addition of IPTG (1 mM final), and the incubation was continued for 3 hr at 30uC. Cells were then collected by centrifugation (1,900 g for 10 min at 4uC), and washed once with 20 ml of 150 mM NaCl. The washed cell pellet was stored at 280uC overnight.

Purification of the Pl.LSU/2 IEP and RNP Particles by Sucrose Cushion Centrifugation
The IEP, tagged in N-terminus with an histidine stretch (6xHis) and a V5 epitope (GKPIPNPLLGLDST) was purified by sucrose cushion centrifugation, as described [23]. The washed cell pellet was resuspended in 4 ml of ice-cold buffer A (50 mM Tris-HCl at pH 7.4, 1 mM EDTA, 1 mM DTT, 10% (v/v) glycerol), and lysozyme was added to a final concentration of 4 mg/ml. After 45 min of incubation on ice, cells were lysed by three cycles of freeze/thawing between 270uC and 37uC, followed by addition of 2.5 volumes of HKCTD buffer (25 mM Tris-HCl at pH 7.4, 500 mM KCl, 50 mM CaCl 2 , 5 mM DTT). Lysate was then centrifuged (14,000 g for 15 min at 4uC), and supernatant was layered over 5 ml of 1.85 M sucrose containing HKCTD and centrifuged in a Beckman Type 70 Ti rotor (50,000 g for 17 h at 4uC). The resulting pellet was gently washed with 1 ml of ice-cold Milli-Q water and then dissolved in 25 ml of ice cold 10 mM Tris-HCl at pH 8.0, and 1 mM DTT. IEP mtDD-and RNP particles containing IEP WT or IEP mtDD-were purified according to the same procedure. Purified proteins and RNPs preparations were stored at 280uC. The yield of RNP particles was 25-90 OD 260 nm units per 100 ml of culture, with 1 OD 260 nm unit of RNP containing 0.8460.12 mg of IEP. The yield of purified proteins was 55-150 mg per 100 ml of culture. Quantification of proteins was performed by using Bio-Rad DC Protein Assay Kit 1 (BioRad Laboratories) according to the manufacturer's instructions. All protein fractions were analyzed by Coomassie-blue staining of SDS-PAGE and western blotting. Nucleic acids in IEP and RNPs (WT and mtDD-) purified fractions were quantified by spectrophotometry (NanoDrop 8000) and analyzed by electrophoresis on agarose gel.

RNA Extraction
Total cellular RNA from yeast cells was extracted from 5 ml of mid-log phase cultures in SD minimal medium (minimal SD base or minimal SD base Gal/Raf, Clontech). After centrifugation (3,800 g for 5 min at 4uC), the cell pellet was washed with 1 ml of Milli-Q water and resuspended in 2 ml of buffer Y1 (1 M sorbitol, 0.1 M EDTA) supplemented with 0.1% b-mercaptoethanol and 300 U of lyticase (from Arthrobacter luteus, .2,000 units/mg protein, Sigma). After 30 min of incubation at 30uC on a rotary shaker, total cellular RNA was extracted from the resulting spheroplasts using the RNeasy Mini Kit (Qiagen) according to the manufacturer's instructions. Total cellular RNA from human cells was also isolated with RNeasy Mini Kit (Qiagen). In both cases, a DNase I treatment (RNase-free DNase set, Qiagen) was performed oncolumn during RNA purification according to the manufacturer's instructions.

Protein Extraction
To obtain yeast total proteins, 20 ml of culture were centrifuged (3,800 g for 5 min at 4uC). The cell pellet was washed with 5 ml of Milli-Q water, resuspended in CelLytic Y reagent (2.5 ml/g cell pellet, Sigma-Aldrich) supplemented with 7 mM DTT and incubated at 25uC for 20 min on a rotary shaker. The resulting lysate was centrifuged at 18,000 g for 10 min at 4uC. Proteins were then precipitated with 2 volumes of acetone for 1 hr at 4uC and recovered by centrifugation at 18,000 g for 15 min at 4uC. Protein pellet was then washed with ethanol and resuspended in 0.5 volume of 50 mM Tris-HCl at pH 8.5, 8 M Urea, and 10 mM DTT.
Yeast nuclear proteins were extracted from 200 ml of culture after centrifugation (4,000 g for 5 min at 4uC). Cell pellets were washed first with 25 ml of Milli-Q water and then with 3 ml of spheroplasting buffer (1 M Sorbitol, 50 mM K 2 HPO 4 at pH 6.5, 0.018% b-mercaptoethanol). Cells were resuspended in 3 ml of spheroplasting buffer containing 500 U of lyticase and incubated 30 min at 30uC on a rotary shaker. The spheroplasts were collected by centrifugation (4,500 g for 5 min at 4uC), washed with 3 ml of ice-cold spheroplasting buffer, resuspended in 8 ml of icecold buffer L (18% Ficoll 400, 20 mM K 2 HPO 4 at pH 6.8, 1 mM MgCl 2 , 0.5 mM EDTA, 2 mM PMSF, 1 mg/ml aprotinin) and lysed on ice by 20 strokes of a dounce homogeneizer. The resulting lysate was then centrifuged (3,500 g for 10 min at 4uC). Supernatant was then centrifuged in a Beckman SW 55 Ti rotor (58,000 g for 35 min at 4uC) and pelleted nuclei were resuspended in 200 ml of ice-cold buffer NP (0.34 M sucrose, 20 mM Tris-HCl at pH 7.4, 50 mM KCl, 5 mM MgCl 2 , 2 mM PMSF, 1 mg/ml aprotinin). Nuclear protein extracts were stored at 280uC.
Protein concentrations were determined with the Bio-Rad DC Protein Assay kit 1.

Western Blot Analyses
Proteins were resolved on a denaturing 10% polyacrylamide-SDS gel (Criterion XT Bis-Tris gels, Biorad), transferred onto a nitrocellulose membrane (Hybond ECL, Amersham) and probed with the appropriate antibodies.
Yeast nuclear proteins (80 mg/lane) and human total proteins (30 mg/lane) were probed with an HRP-conjugated mouse anti-c-Myc antibody (1:3,000 dilution, Life technologies, Invitrogen) to detect the myc-tagged NLS-IEP. For yeast nuclear proteins, a mouse anti-TATA binding protein (TBP) antibody (final concentration of 2 mg/ml, Abcam) was added and HRP-conjugated goat anti-mouse antibody (1:20,000 dilution) was used as secondary antibody. Bands were revealed by chemiluminescence.

Reverse Transcription
Reverse transcription of yeast or human cellular RNA was carried out in 20 ml of reaction medium using the Verso cDNA Kit (Thermo Scientific). One microgram of total RNA was incubated for 5 min at 70uC and then mixed with 1X reverse transcription buffer, 0.5 mM of each dNTPs, 2 mM of RM-R oligonucleotide (yeast RNA; Table S1 B) or 300 ng of random hexamers and 125 ng of anchored oligo dT (human RNA), 1 ml of RT Enhancer enzyme, and 1 ml of Verso reverse transcriptase. After incubation at 42uC for 1 hr, reaction was stopped by heating 2 min at 95uC. A control reaction without Verso reverse transcriptase was performed to ensure the absence of DNA contamination in the RNA samples (minus-Verso RT samples).

PCR
Two ml of yeast cDNA were used as a template for PCR in 50 ml of reaction medium containing 1 unit of Phusion high-fidelity DNA polymerase (Thermoscientific), 62.5 mM of each dNTPs, PCR buffer, and 0.2 mM of each p1 and p2 or p3 and p4 primers. The amplification consisted of 30 cycles at 98uC for 30 sec, 70uC for 30 sec and 72uC for 30 sec. PCR products were analyzed on 12% polyacrylamide gels.
Two ml of human cDNA were used as a template for PCR in 40 ml of reaction medium containing 0.8 unit of Phusion highfidelity DNA polymerase, 200 mM of each dNTPs, PCR buffer, and 0.25 mM of each p5 and p6 primers. The amplification consisted of 30 cycles at 98uC for 10 sec, 58uC for 20 sec and 72uC for 1 min 30 sec. PCR products were analyzed on 1.2% agarose gels.

Quantitative PCR
The qPCR consisted of a SYBR Green-based detection of cDNA with specific primers hybridizing E2 and the E2-E3 junction or primers hybridizing E2 and the E2-Intron junction. Amplification reactions (25 ml) contained 5 ml of a 1/10 dilution of cDNA and 12.5 ml of qPCR buffer (Power SYBR Green PCR master mix, Applied Biosystems), 0.3 mM of each primers and consisted of 40 cycles at 95uC for 15 sec then at 65uC for 1 min on a 7900 HT Real-Time PCR System (Applied Biosystems). To ensure the absence of non-specific amplifications, a dissociation step was added consisting in a final cycle at 95uC (15 sec) then 65uC (15 sec) and finally 95uC (15 sec). To quantify cDNA copy number obtained respectively from spliced or precursor mRNA, standard amplification curves were made by serial dilutions of appropriate linearized plasmids. All PCR measures were performed at least in duplicate. All qPCR experiments include samples from mock vector-transformed cells and minus-Verso RT samples as negative controls.
Data were edited using the Sequence Detection Systems 2.3 software (Applied Biosystems) and interpreted in the linear portion of the standard curve. The following test acceptability criteria were used: linear regression coefficient of the standard curve .0.98; less than 0.5 CT variation for duplicate samples; CT comprised between 35 and 40 for H 2 O; mock vector-transformed cells and minus-Verso RT samples.

Generation and Titration of Lentiviral Vectors
Recombinant lentiviral vectors (LVs) using the pRRL backbone plasmid [63] (Table S1 A) were constructed to express the indicated transgenes under the control of the human phosphoglycerate kinase (PGK) promoter. VSV-G-pseudotyped particles were produced by quadritransfection of HEK 293 T cells (Table  S1 A) as previously described [63,64]. Harvested virus particles were concentrated by ultracentrifugation and titered in infectious genome particles (IG) as previously described [64].

Lentiviral Vector Transduction of Human Cell Lines
HCT 116 cells were seeded at 10 5 cells per well in 12-well plates 16-24 hrs prior transduction. Cells were transduced with LVs using concentrations ranging from 10 6 to 10 8 IG/ml in the presence of 6 mg/ml of polybrene. Transduction medium was replaced with fresh medium 6 hrs after transduction. A second hit of transduction was performed the following morning in the same conditions using LVs encoding proteins. Cells were removed and analyzed 48 hrs later. For LVs stably expressing different forms of Pl.LSU/2 intron, cells were expanded for about 10 days. Vector copy numbers per cell were calculated by quantitative PCR on cell lines genomic DNA as previously described [64] (0.9 for intron-DDIV; 0.3 for intron-DIVa; 4.7 for intron-DIVab, and 1.2 for full length intron).

Pl.LSU/2 IEP Contained in RNP Particles Presents an RT Activity in vitro
The Pylaiella littoralis Pl.LSU/2 group II intron is located in the mitochondrial gene encoding the large ribosomal RNA (Fig. 1A; LSU rRNA gene). This intron contains in its domain IV an openreading frame presenting the predicted conserved domains of group II intron-encoded proteins which are the reverse transcriptase (RT), DNA-binding domain (D), maturase (X) and endonuclease (En) [15,23,25,28,53] (Fig. 1A). Subsequently, recombinant Pl.LSU/2 intron-encoded protein (IEP) was expressed and purified in order to measure its predicted biochemical activities.
Expression of IEP in RNP particles. Some group II intronencoded proteins are known to be fully active either alone [53,65] or when intron RNA is copurified with the expressed IEP [23]. To express the Pl.LSU/2 IEP in RNP particles, we designed the plasmid p151-E+I+IEP for the induction of both the Pl.LSU/2 intron and IEP expression in E. coli (Fig. 1A). This plasmid contains the Pl.LSU/2 intron and its flanking exons (50 last nucleotides of exon 2 and 71 first nucleotides of exon 3) [54] cloned downstream of the phage T7 promoter in the expression vector pET151/D-TOPO (Table S1 A). The IEP ORF is fused to a 6xHis and a V5 epitope tags in its N-terminus used to the detection of the protein by western blot. This vector could potentially allow the expression of both the intron RNA and the wild-type IEP ORF (IEP WT) with its own Shine-Dalgarno sequence for translation. We also constructed a derivative negative control plasmid (p151-E+I+IEPmtDD-) in which the catalytic YADD motif of the IEP RT domain is mutated to YAAA (Fig. 1A, position indicated by a diamond; Table S1 A) [66]. This plasmid is expected to express a mutant RT-defective IEP (IEP mtDD-).
The expression plasmids p151-E+I+IEP and p151-E+I+IEPmtDD-were transformed in E. coli Rosetta-gami B (DE3). After induction, soluble protein fractions were used to purify RNP particles by sucrose cushion centrifugation [23]. Bands at the expected size of IEP WT and mtDD-are observed in the Coomassie-stained SDS-PAGE gel ( Fig. 2A; Coomassie, 69 kDa, black arrow). Other bands are also detected by Coomassie staining and correspond to E. coli contaminant proteins and/or IEP (WT and mtDD-) degradation products. The western blot analysis confirms the presence of IEP WT and mtDD-at the expected size ( Fig. 2A; WB) and shows the presence of some degradation products. A non-specific band is also detected by western blot at 150-kDa in RNPs WT and mtDD-preparations ( Fig. 2A; WB). The analysis of these RNPs fractions by agarose gel electrophoresis shows the presence of ribosomal RNAs (Fig. S1). RNP particles containing either IEP WT or IEP mtDD-can thus be partially purified using sucrose cushion centrifugation.
RT activity. The open reading frame of the intron Pl.LSU/2 contains a conserved RT domain. The RT activity of Pl.LSU/ 2 IEP (WT and mtDD-) in RNP particle preparations was assayed with the artificial template-primer substrate poly(rA)-oligo(dT) [12][13][14][15][16][17][18] . We show that RNP particles from cells expressing p151-E+I+IEP and purified by sucrose cushion centrifugation have an RT activity (Fig. 2B; RNP WT). Quantitatively, the RT activity of 0.1 OD 260 nm units of RNP particles is similar to that of 0.03 units Pl.LSU/2 Group II Intron-Encoded Protein Activity PLOS ONE | www.plosone.org of the commercial SuperScriptH II reverse transcriptase (SSII RT; Life technologies, Invitrogen). The time course of RT reactions shows that the RT activity of WT IEP-containing RNPs increases over incubation time (Fig. 2C) and is positively correlated with the amount of RNP particles (Fig. 2D). As expected, the RT activity of the Pl.LSU/2 IEP is abolished by point mutations in the conserved YADD motif (Fig. 2B, 2C and 2D; RNP mtDDconditions) and is similar to that of the background condition ( Fig. 2C; No protein). The 6xHis tag was also used to purify RNP particles by immobilized metal-ion affinity chromatography (IMAC) on a Ni 2+ -charged column but no RT activity was ever detected in those RNP particle preparations (data not shown), suggesting that this purification process destabilizes the complex. These results indicate that the Pl.LSU/2 IEP in RNPs particles can be expressed in E. coli and purified by sucrose centrifugation thereby preserving its RT activity and thus its active conformation.
Pl.LSU/2 IEP is Active in vitro without the Help of the Intron RNA Isolated Pl.LSU/2 IEP has in vitro RT activity. To analyze the intrinsic property of the Pl.LSU/2 IEP, we expressed the IEP in E. coli without coexpressing the intron RNA. We used the plasmid p151-IEP (Fig. 1A) which places the IEP ORF fused to the 6xHis and V5 epitope tags immediately downstream of the phage T7 promoter and Shine-Dalgarno sequence of the vector. This plasmid version was also constructed with a mutation of the YADD motif ( Fig. 1A; indicated by a diamond). Both WT and mutant proteins were expressed in E. coli and then purified by sucrose cushion centrifugation. SDS-PAGE shows that the WT and the RT-defective IEPs (mtDD-) are successfully purified from E. coli lysates ( Fig. 3A; Coomassie). The identity of the purified proteins is verified by western blot (Fig. 3A; WB) and by mass spectrometry (data not shown). It is worth noting that these purified protein fractions contain substantial amount of nucleic acids, in particular ribosomal RNAs (Fig. S1).
RT assays with Pl.LSU/2 IEP purified by sucrose cushion centrifugation show that the protein displays an intrinsic in vitro RT activity (Fig. 3B; IEP WT). This activity is abolished by mutations in the RT domain as expected ( Fig. 3B; IEP mtDD-). The positive control SuperScriptH II Reverse transcriptase (0.03 U) shows a number of pixels per spot of 18.1 x 10 6 65.78 (not shown). It is noteworthy that no RT activity is detected following IMAC purification of IEP (WT and mtDD-, data not shown), which is consistent with previous findings obtained with RNPs. These data suggest that purification in sucrose cushion preserves the activity of the IEP and thereby its correct folding. It also demonstrates that recombinant IEP has intrinsic RT activity and does not necessarily require the help of intron RNA to become active. In vitro self-splicing of Pl.LSU/2 has been demonstrated [54] using a deleted form of the intron in the domain IV (See Fig. 1B; intron-DDIV). However, most group II introns need the maturase activity of their IEP to achieve in vivo splicing. In Lactococcus lactis Ll.LtrB group II intron, the LtrA intron-encoded protein interacts Pl.LSU/2 Group II Intron-Encoded Protein Activity PLOS ONE | www.plosone.org with a substructure located in the beginning of the domain IV, and other contacts are made with the conserved core regions of the intron [27,67]. In order to evaluate the influence of the IEP on the splicing of the Pl.LSU/2 intron in yeast, we included a part of the domain IV in our construction. This domain corresponds to DIVab indicated in the secondary structure of the domain IV (1870 nucleotides) predicted by the S-fold software (See Fig. 1B;   Fig. S2). This DIVab domain retains putatively most of the predicted secondary structure of domain IV (S-fold prediction, not shown).
To study the expression and splicing of the Pl.LSU/2 intron in yeast, we developed a URA3-based intron-splicing reporter expressed from a 2-m plasmid ( Fig. 4A; pEgpIIE-URA3; Table S1 A). The Pl.LSU/2 (DIVab) intron flanked by its natural exons (50 last nucleotides of exon 2 and 71 first nucleotides of exon 3) [54] was fused to the coding sequence of URA3, which encodes a orotidine 5-phosphate decarboxylase (Ura3p). The URA3 gene is read in-frame only upon precise Pl.LSU/2 splicing, which should lead to Ura3p expression and growth of an ura3-(ura3D0) strain of S. cerevisiae on minimal media lacking uracil (Fig. 4A). Expression of the Ura3p can be detected with an HA tag added upstream of the exon 2, generating the fusion HA-E2-E3-Ura3p. A control plasmid (pEE-URA3) was also used in this assay to ensure that the hybrid HA-E2-E3-URA3 protein is able to complement the ura3-mutation.
Previous studies have shown that optimal splicing of the bacterial Ll.LtrB group II intron in bacteria and yeast requires the maturase activity of its encoded protein, LtrA, by promoting the folding of the intron RNA into its catalytically active structure. The open reading frame of Pl.LSU/2 intron contains a conserved X domain similar to that of LtrA. Therefore, to determine if Pl.LSU/2 IEP has maturase activity facilitating the splicing of its intron, we first attempted to demonstrate intron splicing through functional restoration of yeast growth in the URA3-based intronsplicing reporter assay. In this system, the IEP protein was conditionally expressed using an inducible GAL10 promoter plasmid ( Fig. 4B; pNLS-IEP co ) and intron sequences were delivered in trans. For these experiments, we used an IEP sequence that was codon-optimized for translation in human cells (IEP co ) tagged in C-terminus with c-Myc epitopes, and containing nuclear localization signals (NLS) of the SV-40 T-antigen (Fig. 4B) to address the protein to the nucleus. The inducible/repressible expression of the NLS-IEP co was confirmed respectively in galactose (Gal) and glucose (Glc) media and nuclear localization of the protein was verified by western blot analysis on yeast nuclear protein extracts (data not shown).
Yeasts were thus transformed with either the control pEE-URA3 plasmid or the pEgpIIE-URA3 plasmid. Subsequently, the NLS-IEP co expressing plasmid was transformed or not in yeast carrying pEgpIIE-URA3 and the number of colonies on medium containing or not uracil was determined for each condition (Fig. 4C). Functional restoration of growth could not be demonstrated in the URA3-based intron splicing reporter assay. Yeasts expressing the EgpIIE-URA3 cassette repeatedly fail to grow on the -ura selective medium, even when EgpIIE-URA3 is coexpressed with NLS-IEP co (Fig. 4C). However, yeasts grow on minimal medium lacking uracil when the cells are transformed with the pEE-URA3 plasmid encoding the hybrid E2-E3-URA3 protein proving the orotidine 5-phosphate decarboxylase activity of the hybrid protein, as it confers the ability to complement the ura-mutation (Fig. 4C). Thus, the splicing of the Pl.LSU/2 intron could not be demonstrated through this assay into yeast cells.

Pl.LSU/2 Group II Intron can Splice in Yeast in an IEPdependent Manner
Since the assay reads-out both RNA splicing and translation of the spliced mRNA, further investigation was conducted to determine if RNAs were transcribed, spliced and translated. To determine if the Pl.LSU/2 intron was spliced from the precursor mRNA, an RT-PCR analysis was performed on RNA extracted from yeast cells harboring either pEE-URA3 or pEgpIIE-URA3  Numbers of yeast colonies growing on appropriate minimal medium containing (+ura) or not (2ura) uracil. Strain carrying pEE-URA3 was used as a control. Strain carrying pEgpIIE-URA3 splicing reporter was transformed or not with the NLS-IEP co expressing plasmid (pNLS-IEP co ). Three resulting transformants were cultured until OD 600 nm reached 2.5 and applied on the appropriate minimal medium using a 10 24 dilution of the culture. Data indicate the numbers of yeast colonies on the plate per ml of the dilution. doi:10.1371/journal.pone.0058263.g004 expressed in the presence or absence of NLS-IEP co . Two different pairs of primers were used to specifically amplify cDNA obtained from precursor or spliced mRNA ( Fig. 5A; Table S1 B). We first show that the pEgpIIE-URA3 splicing reporter allows the expression of precursor mRNA in all conditions ( Fig. 5B; precursor cDNA, pEgpIIE-URA3). As expected, the control plasmid pEE-URA3 allows the expression only of a mRNA corresponding to the spliced mRNA ( Fig. 5B; pEE-URA3). In absence of NLS-IEP co , a poorly efficient splicing of Pl.LSU/2 can be detected from yeast harbouring the pEgpIIE-URA3 reporter ( Fig. 5B; spliced cDNA, pEgpIIE-URA3). It is worth noting that splicing efficiency is significantly increased when NLS-IEP co is expressed into yeast cells (Fig. 5B; spliced cDNA, Glucose2/ Galactose+). Accurate Pl.LSU/2 splicing was verified by sequencing across the splice junction of the RT-PCR spliced products (data not shown). These results were quantified using an RT-qPCR analysis performed on four independent experiments (Fig. 5C). Precursor and spliced cDNA copy numbers were calculated and ratios of spliced/precursor cDNA were determined to measure splicing efficiency. We confirm that in absence of NLS-IEP co expression (Glc+/Gal2), the splicing of Pl.LSU/2 is poorly efficient while the Pl.LSU/2 intron splicing efficiency in yeast cells is significantly increased upon NLS-IEP co expression ( Fig. 5C; Glc2/Gal+; between 2.1 and 7.9 fold change compared to background; p,0,032, t-test using a unilateral pair-wise comparison). These results indicate first that the Pl.LSU/2 intron can be expressed in yeast and also that the IEP presents a maturase activity in vivo.

Spliced Pl.LSU/2 Group II Transcripts are not Translated
The results of the phenotypic analysis showing an absence of clones on medium lacking uracil (Fig. 4C) and the level of splicing observed in yeast harboring the pEgpIIE-URA3 reporter and expressing the NLS-IEP co (Fig. 5C) suggested a blockade in expression of the spliced messenger. Indeed, we had not been able to detect by western blot the Ura3p hybrid protein resulting theoretically from the Pl.LSU/2 spliced mRNA ( Fig. 5D; pEgpIIE-URA3, Ura3p), even upon robust NLS-IEP co expression (Fig. 5D). In contrast, high levels of Ura3p hybrid protein are expressed in yeast harboring pEE-URA3 (Fig. 5D). The absence of detectable level of Ura3p in yeast carrying pEgpIIE-URA3 thus explains the failure of functional growth restoration in the URA3based intron splicing reporter assay.

Splicing Assay of Pl.LSU/2 Intron in Human Cells
To study if the splicing of the Pl.LSU/2 intron could also occur in human cells, we first aimed to determine if the Pl.LSU/2 intron could be expressed in the colon carcinoma HCT 116 cell line. We tested various forms of the intron in which the domain IV was more or less extensively deleted (See Fig. 1B; DDIV, DIVa, DIVab and full length intron). To stably express the various intron sequences in human cells, we transduced HCT 116 cells with lentiviral gene transfer vectors (LVs) expressing the different forms of the PI.LSU/2 intron flanked by the last 50 nucleotides of exon 2 and the first 71 nucleotides of exon 3 ( Fig. 6A; upper panel; pRRL-intron +/2DIV). Transduced HCT 116 cells were expanded for at least ten days to establish stable cell lines expressing the different forms of the intron. RT-PCR using primers p5 and p6 hybridizing E2 and E3, respectively ( Fig. 6A; upper panel; Table S1 B) shows expression of the appropriate precursor transcripts in each stable cell lines ( Fig. 6A; lower panel; amplicons of 736 bp for DDIV intron, 1020 bp for DIVa intron, 1472 bp for DIVab intron and 2534 bp for full length intron). Nevertheless, we did not find any trace of spliced mRNA in any of the stable cell lines analyzed (data not shown). These results suggest that Pl.LSU/2 intron RNA can be expressed but does not splice in human cells.
To determine if the intron-encoded protein could promote splicing of Pl.LSU/2 intron in human cells, we induced the expression of Pl.LSU/2 IEP co or GFP-IEP co fusion protein in human HCT 116 cells by transduction with protein-expressing LVs ( Fig. 6B; upper panel). Forty eight hours following transduction of HCT 116 cells with the LV, western blot analysis demonstrates expression of the IEP co or GFP-IEP co in the cells (Fig. 6B; lower panel). GFP expression was also confirmed by western blot using an anti-GFP antibody and by evidence of nuclear localization following microscopy analysis (data not shown).
The previously described Pl.LSU/2 intron expressing stable lines were then transduced by these LVs. Western blot analysis showed the expression of IEP co and GFP-IEP co in these cells (data not shown). To determine the Pl.LSU/2 intron splicing capacity in the presence of IEP, RNAs were extracted and analyzed by RT-qPCR. Primers p7 and p4 (Table S1 B), hybridizing E2 and E2-E3 junction respectively, were used to detect the spliced mRNA. Primers p1 and p2 (Table S1 B), hybridizing E2 and E2-Intron junction respectively, were used to detect the precursor mRNA. Ratios of spliced/precursor cDNA copy number, which determine the splicing efficiency, are less than 2.10 24 in every condition tested and for every form of the Pl.LSU/2 intron tested (data not shown). This result shows that, in contrast to yeast, the Pl.LSU/2 intron is unable to splice efficiently in human cells in this context, even in presence of the intron-encoded protein that was codonoptimized for translation in human cells.

Discussion
This study provides the first functional description of the ribozyme activity of the Pl.LSU/2 group II intron in vivo. A catalytically-active recombinant PI.LSU/2 intron-encoded protein can be produced in E. coli and purified. This protein presents a reverse transcriptase activity in vitro both alone or when coexpressed with the intronic RNA. It also displays a maturase activity which facilitates the splicing of the Pl.LSU/2 intron in vivo in yeast in a IEP dose-dependent manner.
These results contribute to characterize the poorly studied Pl.LSU/2 intron. Prior to our study, no information was yet available on the in vivo properties of this intron, neither in its natural environment nor in experimental systems. The description of the RT and maturase activities of the IEP confirms some of the predicted properties encoded by the domain IV of the intron which was known to contain an open reading frame theoretically encoding a protein presenting the four characteristic domains (RT, X/maturase, D and En) of other group II intron-encoded proteins [55,56].
One limitation of the use of group II introns in eukaryotes could be the requirement for a correct folding of both the IEP and the intron RNA, therefore it is important to assess the cooperation between the intron and IEP. Here, we show for the first time that the Pl.LSU/2 IEP displays a RT activity when complexed as RNP. In addition, the PI.LSU/2 IEP has an intrinsic RT activity in vitro in the absence of the intron RNA. However, one can postulate that the presence of nucleic acids in the sucrose cushion purification fractions may participate in the Pl.LSU/2 IEP stabilization and activity. In previous studies of the Lambowitz group [23,53,65], the Lactococcus lactis Ll.LtrB IEP (LtrA) correct folding was demonstrated by its RT activity either alone or when coexpressed with its intron RNA. Moreover, highly active LtrA primer; p1 and p2: forward and reverse primers amplifying cDNA derived from the precursor mRNA (111 bp product); p3 and p4: forward and reverse primers amplifying cDNA derived from the spliced mRNA (77 bp product). (B) Acrylamide electrophoresis of RT-PCR products.Total RNA were extracted from yeast carrying pEE-URA3 or pEgpIIE-URA3 and transformed (+) or not (2) with the NLS-IEP co expressing plasmid (pNLS-IEP co ). Cells were grown in presence (+) or absence (2) of glucose or galactose (inducer of the NLS-IEP co expression). P: amplification product from precursor cDNA; S: amplification product from spliced cDNA. Numbers at left indicate molecular mass marker in base pair (bp). (C) Quantification of the Pl.LSU/2 in vivo splicing by RT-qPCR. cDNA copy number obtained from spliced and precursor mRNA were calculated using standard amplification curves made by serial dilutions of SphI-linearized pEE-URA3 and pEgpIIE-URA3 plasmids, respectively. Ratios of spliced/precursor cDNA copy number were determined and data were normalized on condition 1 (pNLS-IEPco (2); Glc+; Gal2). Four independent experiments are represented (Exp 1 to Exp 4). (D) Western blot analysis of Ura3p made from pEE-URA3 or pEgpIIE-URA3. Yeast strains were cultivated in glucose (Glc+; Gal2) or in galactose (Glc2; Gal+) and in absence (2) or presence (+) of pNLS-IEP co . Tubulin 1 protein (Tub1p) expression was also determined, as well as nuclear expression of NLS-IEP co and TATA-Binding protein (TBP). doi:10.1371/journal.pone.0058263.g005 was recently obtained after nucleic acids removal prior to purification [65], showing that this protein does not require bounded nucleic acids to be stabilized. Similarly, Vellore et al. expressed in E. coli the G.st.I1 intron-encoded protein (called trt) from Geobacillus stearothermophilus fused to a 6xHis tag without coexpressing the intron RNA [68]. The authors showed an RT activity using partially purified protein fraction. In the case of Pl.LSU/2, the IEP reverse transcriptase activity could be clearly demonstrated in the absence of intronic RNA, showing the PI.LSU2 IEP self-activation properties.
The biochemical properties of this IEP may also facilitate its use in vivo. The PI.LSU/2 intron can be engineered to splice in vivo in yeast and this is facilitated in a dose-dependent manner by the Pl.LSU/2 IEP. Although the Pl.LSU/2 IEP encoding gene used here is a codon-optimized sequence for translation in human cells, IEP translation in yeast is enough efficient to improve splicing of Pl.LSU/2. Presumably the maturase activity of the PI.LSU/2 IEP promotes the folding of the intron RNA into its catalytic tertiary structure in vivo. Interestingly, a residual splicing could be detected by RT-PCR and RT-qPCR using RNA extracts from yeasts that do not express the IEP. This IEP-independent splicing could have occur in vivo (which would be consistent with the ability of this intron to splice in vitro even under not optimal Mg 2+ conditions [54]) or in vitro during the RT-PCR reactions precisely because this ribozyme is highly active in vitro.
In spite of evidence of functional properties of the PI.LSU/2 intron, we failed to develop a productive system for genomic targeting at this stage. While we demonstrated the presence of a spliced mRNA in yeast cells, there was no translation of the spliced mRNA since restoration of the URA3 ORF did not lead to the expected growth on a minimal medium without uracil. It is possible that the yield of spliced mRNA in yeast cells is not sufficient to the expression of a detectable amount of Ura3p proteins by western blot and that the yield of expressed Ura3p does not reach the required threshold for yeast growth on medium lacking uracil. However, the possibility of a translation blockade is reminiscent of results already reported by Chalamcharla et al. with the Ll.LtrB intron using different reporter genes [69]. The authors demonstrated that Ll.LtrB intron splicing occured predominantly in the cytoplasm of yeast cells and that precursor mRNA was subjected to nonsense-mediated mRNA decay (NMD). They also showed that the spliced mRNA was subjected to an NMD-independent translation blockade. However, the mechanism involved remains unknown. The authors speculated that the pairing mechanism between the intron lariat and the spliced mRNA via EBS (exon binding sites)/IBS (intron binding sites) interactions [6] could impede the spliced mRNA translation [69]. One would also postulate that the translation blockade could result from RNA sequestration to cytoplasmic microdomains such as P-bodies or stress granules that play important role in mRNA processing, including repression of translation and mRNA decay [70][71][72]. Anyway, the fact that this translation blockade also occurs through the use of Pl.LSU/2 group II intron supports the Chalamcharla et al. hypothesis that the mechanism involved is transposable to other group II introns and may has impeded their spread in eukaryotic nuclear genes.
We then tested the ability of the intron to splice in human cells. Here, we used different forms of the Pl.LSU/2 intron with deletions of various sizes of the domain IV in order to determine if one of these parts are required for intron high-affinity binding to the IEP. The domain IV of the Ll.LtrB intron is known to bind its IEP [73], but is not needed for in vitro splicing, as in absence of it, a residual splicing occurs [27,67]. In the same way, the domain IV of the yeast coxI-I2 intron is required for stable binding of its IEP and additional contacts with the catalytic core of the intron promote the splicing [74]. The domain IV of group II introns appears to guide the interactions or anchor the IEP to catalytic core regions of the intron. In spite of the use of various domain IV coding regions, in spite of using a humanized codon-optimized IEP, in spite of evidence of expression of the IEP and intron in human cells and in spite of the expected nuclear localization of the recombinant IEP in human cells, we failed to detect any splicing of Pl.LSU/2 intron in human cells. Misfolding of IEP and/or intron RNA could explain this splicing defect, but alternatively, defective nuclear/cytoplasmic compartimentalization or inadapted environnement could also be implicated.
The adaptation of group II introns homing mechanism could be considered in the context of a gene repair strategy approach in gene therapy. In current gene therapy assays, the transgene integration with retroviral vectors is not site-specific. Expression of therapeutic transgenes is sometimes unregulated due to cis-acting elements present in the neighbouring of the insertion site. Insertions near oncogenes also present a risk of activation by promoters or enhancers carried by the vector [75][76][77][78][79][80]. All of these issues could be circumvented by targeting the insertion into the original site (gene repairing). The repaired wild type copy will be under the control of original cis-regulating sequences. Today, strategies based on genome double strand breaks (DSB) in order to achieved reparation by homologous recombination are thouroughly investigated. The use of group II intron could be an alternative approache avoiving the putative genotoxicity due to off-target DSB.
We present encouraging data that suggest that PI.LSU/2 group II intron could have advantageaous qualities for engineering genomic targeting strategies. This intron and its IEP function in vivo in yeast. However, the use of Pl.LSU/2 and other group II introns in human genomic engineering will require further optimizations.