Here we investigate the potential of group IC1 and group IE intron ribozymes, derived from the myxomycetes Didymium and Fuligo, in addition to the Tetrahymena ribozyme, for RNA reprogram
Trang 1in vitro by myxomycete group IC1 and IE ribozymes
Tonje Fiskaa1,*, Eirik W Lundblad1,2,*, Jørn R Henriksen1,3, Steinar D Johansen1
and Christer Einvik1,3
1 Department of Molecular Biotechnology, RNA Research group, Institute of Medical Biology, University of Tromsø, Norway
2 Department of Microbiology, University Hospital of North Norway, Tromsø, Norway
3 Department of Pediatrics, University Hospital of North Norway, Tromsø, Norway
Group I ribozymes, which normally perform intron
splicing reactions within the nucleus of many
unicellu-lar eukaryotes, can be modified to trans-splice 3¢
exons into separate RNA molecules in a
sequence-specific reaction Group I intron trans-splicing is
initiated by the binding of an exogenous guanosine
(exoG) into the ribozyme guanosine-binding site
Sub-sequently, base pairing between the internal guide
sequence (IGS) at the 5¢ end of the ribozyme and a
target RNA sequence creates a pseudo-P1 structure
containing the 5¢-splice site, which becomes attacked
by the bound exoG The splicing reaction proceeds
through two consecutive transesterification steps When targeting mutated messenger RNAs, trans-spli-cing may lead to chimerical reprogrammed transcripts
of biochemical or therapeutic interest Despite the fact that more than 2000 group I introns are known
by sequence [1], only a few have been applied in RNA reprogramming approaches The Tetrahymena ribozyme has been used in almost all reported cases
of RNA reprogramming (including RNA repair) [2– 7], except for a few studies using the Pneumocystis ribozyme [8,9] and the Didymium myxomycete ribo-zyme DiGIR2 [10,11]
Keywords
a-mannosidase mRNA; group I intron; RNA
repair; RNA reprogramming; trans-splicing
Correspondence
S D Johansen, Department of Molecular
Biotechnology, RNA Research Group,
Institute of Medical Biology, University of
Tromsø, N-9037 Tromsø, Norway
Fax: +47 776 45350
Tel: +47 776 45367
E-mail: Steinar.Johansen@fagmed.uit.no
*These authors contributed equally to this
study
(Received 17 March 2006, accepted 27 April
2006)
doi:10.1111/j.1742-4658.2006.05295.x
Trans-splicing group I ribozymes have been introduced in order to mediate RNA reprogramming (including RNA repair) of therapeutically relevant RNA transcripts Efficient RNA reprogramming depends on the appropri-ate efficiency of the reaction, and several attempts, including optimization
of target recognition and ribozyme catalysis, have been performed In most studies, the Tetrahymena group IC1 ribozyme has been applied Here we investigate the potential of group IC1 and group IE intron ribozymes, derived from the myxomycetes Didymium and Fuligo, in addition to the Tetrahymena ribozyme, for RNA reprogramming of a mutated a-mannosi-dase mRNA sequence Randomized internal guide sequences were intro-duced for all four ribozymes and used to select accessible sites within isolated mutant a-mannosidase mRNA from mammalian COS-7 cells Two accessible sites common to all the group I ribozymes were identified and fur-ther investigated in RNA reprogramming by trans-splicing analyses All the myxomycete ribozymes performed the trans-splicing reaction with high fidelity, resulting in the conversion of mutated a-mannosidase RNA into wild-type sequence RNA protection analysis revealed that the myxomycete ribozymes perform trans-splicing at approximately similar efficiencies as the Tetrahymena ribozyme Interestingly, the relative efficiency among the ribozymes tested correlates with structural features of the P4–P6-folding domain, consistent with the fact that efficient folding is essential for group I intron trans-splicing
Abbreviations
exoG, exogenous guanosine; IGS, internal guide sequence; RPA, RNA protection analysis.
Trang 2tion, and recently reported RNA reprogramming of a
mutated glycosylasparaginase mRNA sequence by a
Didymiummyxomycete group IE ribozyme deficient in
hydrolysis [10] Both the specificity and the efficiency
of trans-splicing were improved by extending the IGS
and by the removal of unwanted hydrolysis side
reac-tions by ribozyme modificareac-tions Here we report the
mapping of accessible ribozyme target sites in a
mutant version of a-mannosidase mRNA isolated from
mammalian COS-7 cells A strategy based on
random-ized IGS [3,11], using four distinct group I ribozymes,
the Didymium group IE ribozyme (named DiGIR2),
two Fuligo group IC1 ribozymes (named Fse.L569 and
Fse.L1898), and the prototype Tetrahymena group IC1
ribozyme (Tth.L1925), were used Two accessible sites,
common to all four ribozymes, were identified and
included in site-directed RNA reprogramming
Results and Discussion
Structural features of the myxomycete group IC1
and group IE ribozymes
The intron secondary-structure diagrams presented in
Fig 1A show that all four ribozymes included in this
study have an overall similar structural organization of
the catalytic domain (P3–P7–P8) The different
myx-omycete ribozymes were selected as a result of their
pronounced in vitro splicing activities and because of
their distinct structural features [14,15] Whereas
DiGIR2, derived from the twin-ribozyme intron
Dir.S956-1 in D iridis [14,16–18] represents the group
IE ribozymes, the two F septica ribozymes, Fse.L569
and Fse.L1898, represent group IC1 ribozymes [15]
Finally, the prototype T thermophila group IC1
ribo-zyme, Tth.L1925, was included as a reference control
A dramatic variation in both sequential and
struc-tural features is noted among the folding domains
pre-sented in Fig 1B (P4–P6) This domain varies in size
from only 94 nucleotides in DiGIR2 to 681 nucleotides
in Fse.L569, with the intermediate-sized Tth.L1925
and Fse.L1898 folding domains (157 nucleotides and
weak) L9b–P5 interdomain interaction [14], features typical of the IE subclass of nuclear group I ribo-zymes Fse.L569, on the other hand, harbours a group IC1 folding domain with branched P5 (P5abc) and the A-bulge in P5a Large extensions are located in P5b and P6, as well as in the highly unusual P5d region The latter region is more than 300 nucleotides long and contains 17 identical copies of a 16-nucleotide tan-dem repeat motif [15] The repeat has probably no important function in splicing as a mutant ribozyme with only seven copies performs the self-splicing reac-tion at similar rates, and cognate self-splicing intron ribozymes in Badhamia and Diderma (Bgr.L569 and Dni.L569, respectively) lack the repeat (S D Johansen
et al., unpublished results)
Group IC1 and group IE ribozymes select the same accessible sites in an a-mannosidase mRNA sequence
Deficiency of lysosomal activity of human a-mannosi-dase, an exoglycosidase enzyme involved in the ordered degradation of N-linked oligosaccharides, results in the autosomal-recessive lysosomal storage disorder, a-mannosidosis [24] The most frequent mutation in a-mannosidase is the R750W substitution, and affected individuals accumulate partially degraded oligosaccharides in the lysosome [25] No causal treat-ments are currently available for a-mannosidosis, and there is thus a need for developing alternative gene therapy approaches Group I ribozyme-based mRNA repair may represent an interesting new approach in gene therapy by reprogramming RNA molecules carry-ing disorder mutations To investigate whether RNA reprogramming could be applied on human a-mannos-idase mRNA, total RNA isolated from COS-7 cells expressing R750W mutant a-mannosidase was mixed with ribozyme libraries designed to detect accessible target sites within the messenger RNA GN4⁄ 5 ribo-zyme-tag libraries (see Experimental procedures) were constructed for each of the group I ribozymes (Fig 1A) During incubation at trans-splicing
Trang 3condi-tions, a unique 3¢ exon tag (Fig 1A) was trans-spliced
into various mRNA sequences, based on the
accessibil-ity to the ribozymes [11] The resulting RNA
recombi-nants were detected and identified by an RT-PCR
approach, cloned into plasmid and subsequently DNA
sequenced Here, a total number of 19 distinct clones,
all representing true trans-splicing events within an
1 kb region of the mRNA, were identified This
region was chosen because it contains sequences
upstream of the R750W substitution (Fig 1C) and is
thus suitable for RNA reprogramming by group I
ribozyme trans-splicing Surprisingly, only three
access-ible sites were detected (Fig 1C) Whereas the
Tetra-hymena ribozyme detected all three sites (U1357,
U1381, and U1732), the Fuligo and Didymium
ribo-zymes detected the same two sites (U1357, U1381)
Several interesting findings are noted from this
experi-ment, namely that (a) sites U1357 and U1381 appear
to be particularly accessible because they were detected
by all four ribozymes, (b) U1357, U1381 and U1732
could not be predicted as unambiguous accessible
regions by the mfold computer program (data not
shown), stressing the importance of determining
accessible sites within target RNAs experimentally,
(c) no obvious sequence similarities were seen between
the selected target sites and the natural 5¢ splice sites
of the ribozymes, indicating that the accessible site
detection was based on true selection, (d) two of the
selected target sites (U1357 and U1732) were identical
in sequence, but the latter was only detected by the
Tetrahymena ribozyme, and (e) the selected target
sequences GCACCU(1357⁄ 1732) and ACGACU1381
generate GC-rich P1 pairings, suggesting that increased
stability between ribozyme and target RNA is an
addi-tional selective factor [26] In summary, we found that
group IC1 and group IE ribozyme-tag libraries are
able to select the same accessible sites within an
endog-enously expressed human a-mannosidase mRNA
Increasing the trans-splicing specificity at
a-mannosidase RNA sites U1357 and U1381
The two accessible a-mannosidase RNA sites (U1357
and U1381) were selected for more detailed analysis in
RNA reprogramming because they were recognized by
all four ribozymes tested In order to obtain more
opti-mal ribozyme targeting, several modifications in the
ri-bozyme structures were performed These include IGSs
complementary to the sequences flanking U1357 and
U1381, as well as EGSs, 35 nucleotides in length,
complementary to the target RNA sequences 3¢ of
U1357 and U1381 (Fig 2A) These modifications, along
with the short P10 base pairing important in the second
step of trans-splicing, were included to increase the spe-cificity of the reaction according to previously published work on RNA trans-splicing optimizations [4,5,10–12] Furthermore, full-length a-mannosidase mRNA sequences ( 1660 nucleotides), corresponding to the regions 3¢ of U1357 and U1381, were added as trans-splicing 3¢ exons in the ribozyme constructs It is important to note that these 3¢ exons harbour the RNA sequence corresponding to the wild-type arginine resi-due at position 750 (R750), and thus have to be consid-ered as restorative 3¢ exon sequences (Fig 2B) Finally,
to avoid strong intermolecular base pairing between the EGS and the restorative 3¢ exon during the trans-spli-cing reaction [10,11], the corrected (wild-type) a-man-nosidase sequences were degenerated by alternative codons for the first 16 and 15 triplets following the tar-get sites U1357 and U1381, respectively (Fig 2A) All eight ribozyme constructs (DiGIR2, Fse.L569, Fse.L1898, and Tth.L1925 targeting both U1357 and U1381) were incubated at trans-splicing conditions (see Experimental procedures) with in vitro-transcribed a-mannosidase target RNA in a 2 : 1 (ribozyme⁄ tar-get) molar ratio In an RT-PCR approach, the trans-ligated exon products were amplified as the expected
390 bp and 437 bp products for positions U1357 and U1381, respectively (Fig 3A) Representative ampli-cons for all eight reactions were DNA sequenced and confirmed to result from a correct and accurate trans-splicing reaction (Fig 3B) A minor RT-PCR product, shorter in size than the expected 390 bp, was observed
at U1357 for all four ribozyme reactions (Fig 3A) However, after gel purification and DNA sequencing, this product was found to be a result of oligonucleo-tide mispriming during the RT-PCR reaction In sum-mary, all four ribozymes were designed to target the two most accessible sites in a-mannosidase mRNA Several modifications that increase the specificity and efficiency of the reaction were included, and all the ribozymes were found to perform the trans-splicing reaction in a highly accurate manner
Determination of trans-splicing efficiencies at a-mannosidase RNA sites U1357 and U1381
To determine the efficiency of the trans-splicing reac-tions and to compare the different ribozyme con-structs, the same reactions described above were performed but analysed by different experimental approaches In the first experiment, unlabelled tran-scripts of each of the eight ribozymes and [35 S]CTP-labelled target RNA were mixed (2 : 1 molar ratio), incubated at trans-splicing conditions at various time points (0, 5 15, 30, 60 and 90 min), subjected to
Trang 4polyacrylamide gel analysis, and finally visualized by
autoradiography A representative time course
trans-splicing analysis at U1381 is presented in Fig 3C
Here, trans-spliced RNA (RNA1) is found to
accumu-late RNA1 from all eight reactions was gel purified
and DNA sequenced by an RT-PCR approach, which
confirmed that all the ribozyme transcripts were able
to trans-splice target RNA in vitro in both sites (data
not shown) We note that free 5¢ exons (RNA3), but
not free 3¢ exons, are readily detected in the gel
analysis (see Fig 3C), an observation explained by the
strong intermolecular base pairings between
the exchanged 3¢ exons and EGS Interestingly, one of the ribozymes (Fse.L1898) was apparently more effi-cient in trans-splicing at both sites compared with the other myxomycete ribozymes tested
In the second experiment we performed an RNA protection analysis (RPA) on the trans-spliced prod-ucts detected above in order to quantify the reactions and compare the efficiencies among DiGIR2, Fse.L569, Fse.L1898 and the Tetrahymena ribozyme, Tth.L1925 The RPA probes were designed to hybrid-ize to 351 and 385 nucleotides of target RNA 5¢ exon sequences, and 36 and 52 nucleotides of restorative 3¢
Fig 1 Group I ribozymes and mutant a-mannosidase target RNA (A) Secondary structure diagrams of trans-splicing ribozymes in accessible site selection The paired segments P2–P9 and P13 are indicated The randomized internal guide sequence regions (IGS; GN 5 in Tth.L1925 and DiGIR2, and GN4in Fse.L569 and Fse.L1898) are boxed at the 5¢ end of the ribozymes The DiGIR2 splicing ribozyme is derived from the twin-ribozyme intron Dir.S956-1 [16] The unique TAG sequence used in RT-PCR detection is indicated at the 3¢ end of the ribozymes (B) Secondary structure diagrams of P4–P6 folding domains of the group I ribozymes Tth.L1925, DiGIR2, Fse.L569 and Fse.L1898 The DiGIR2 splicing ribozyme is derived from the twin-ribozyme intron Dir.S956-1 [16] Intradomain tertiary interactions (A–bulge ⁄ P4 interactions and L5b–P6 tetraloop receptor interactions) are indicated by arrows The 16-nucleotide direct-repeat motif present in 17 identical copies at P5d in Fse.L569 is boxed (C) Schematic presentation of the a-mannosi-dase cDNA expressed in COS-7 cells The selected accessible sites are indicated as T1357, T1381, and T1732 The gene mutant corresponding to the R750W substitution resulting in a-mannosidosis is shown.
Trang 5exon sequences at sites U1357 and U1381, respectively.
The protected regions for the trans-spliced RNAs
cor-respond to 387 nucleotides and 437 nucleotides Gel
analysis of RPA products (Fig 4A) confirmed the
above experiments of in vitro trans-splicing The
relat-ive efficiencies of the trans-splicing reactions were cal-culated in comparison to the Tetrahymena reference ribozyme, and the corresponding values are shown in Fig 4B Here, the average amounts of trans-spliced RNAs in four parallel experiments performed by
Fig 1 (Continued).
Trang 6Fig 2 Design of trans-splicing ribozyme constructs targeting specific sites within mutant a-mannosidase RNA sequences (A) Design of trans-splicing ribozyme (Rz) constructs targeting a-mannosidase U1357 and U1381 The ribozyme contains an internal guide sequence (IGS) and an extended guide sequence (EGS), which base pair to the complementary sequence in a-mannosidase mRNA upstream of the R750W (C to T at position 2248) mutation The ribozyme constructs used contain silent mutations (underlined) introduced by alternative codons in the restorative 3¢ exon (B) Schematic presentation of the group I ribozyme-mediated trans-splicing reaction resulting in RNA reprogramming
of a-mannosidase mRNA The trans-splicing ribozyme construct base pairs to mRNA sequences upstream of the mutation (Mut) and cata-lyses the coupled cleavage of mutated mRNA and the ligation of a restorative 3¢ exon containing wild-type sequences.
Trang 7DiGIR2, Fse.L569, and Fse.L1898 are 81%, 79% and
104% for U1357, and 39%, 58% and 96% for U1381,
respectively The fact that Fse.L1898 is noted as the
most efficient of the myxomycete ribozymes tested
correlates well with the gel analysis presented in
Fig 3C The efficiency of the trans-splicing ribozymes
at both U1357 and U1381 target sites (Fig 4C)
appears to correlate to a putative folding problem of
the P4–P6 domain, either as a result of the lack of
intradomain stabilization or by misfolding of the
com-plex sequence features Fse.L1898 possesses a P4–P6
folding domain similar to that of the Tetrahymana
ribozyme (Fig 1B), both in size, organization, and
pre-dicted intra- and interdomain interactions Consistent
with the above argument, we suggest that RNA
fold-ing advantages in the P4–P6 domain make Fse.L1898
the most efficient of the myxomycete trans-splicing
ribozymes tested (Fig 4C)
In summary, our analyses confirmed that all three
myxomycete ribozymes tested perform the
trans-spli-cing reaction as accurately as the Tetrahymena
ribo-zyme Furthermore, one of the ribozymes (Fse.L1898)
was more efficient in trans-splicing than the other
myxomycete ribozymes tested
Experimental procedures
Mapping accessible sites within a-mannosidase
mRNA
Mapping of accessible sites in a-mannosidase mRNA by
GN4 ⁄ 5 ribozyme-tags was performed as previously
des-cribed [3,11] The IGS, preceding the UG wobble pair in
Fig 3 Reprogramming a-mannosidase RNA by group I ribozyme
trans-splicing (A) RT-PCR products from in vitro trans-splicing
experiments with mutant target a-mannosidase RNA and the
trans-splicing ribozymes The RT-PCR products correspond to
trans-spliced RNAs of the expected sizes (390 bp and 437 bp) at
positions U1357 and U1381, respectively The controls (Ctrl)
con-tain first-strand synthesis master mix with Tth.L1925 only, or target
RNA only (B) Representative results of correct trans-spliced
a-man-nosidase mRNA sequences at positions U1357 and U1381 obtained
from RT-PCR amplifications (C) Representative time-course
analy-sis of in vitro trans-splicing experiments a-Mannosidase RNA and
trans-splicing ribozymes targeting U1381 were in vitro transcribed
with and without [ 35 S]CTP labelling, respectively Trans-splicing
ribozymes and target RNA were incubated at a 2 : 1 molar ratio, at
50 C for 3 h under splicing conditions Samples were collected at
0, 5, 15, 30, 60 and 90 min Trans-splicing products were analyzed
by PAGE and visualized by autoradiography The major RNA
spe-cies detected were trans-spliced RNA (RNA1), a-mannosidase
tar-get RNA (RNA2) and free 5¢ exon tartar-get RNA (RNA3) M, RNA size
marker.
Trang 8Fig 4 Ribonuclease protection analyses (RPA) of trans-spliced a-mannosidase RNA sequences (A) Schematic presentation of the RPA experimental approach See the legends to Fig 2B for details (B) Represen-tative results of the major RNA species (numbered 1–4) detected in RPA RNA1, undigested probe; RNA2, trans-spliced a-mannosidase mRNA; RNA3, a-mannosi-dase target RNA; RNA4, ribozyme RNA M, RNA size marker (C) Quantification of the RPA of trans-spliced a-mannosidase mRNA generated by the different trans-splicing ribozymes Comparative quantitative data were collected from six independent RPA experiments The trans-splicing efficiency (percentage) was calculated as previously described [10], except that values were normalized in respect to the Tth.L1925 ribozyme activity (100%) The raw yields
of trans-spliced target RNA for the Tetrahymena ribozyme were found to vary from 3 to 15% and from 2 to 8% for the sites U1357 and U1381, respectively, in six independent experiments However, the relative yields among the four ribozymes are similar for each of the experiments.
Trang 9the P1 helix (see Fig 1), was used in GN4⁄ 5 mapping
libraries (n¼ 4 or n ¼ 5; G is the invariant guanosine
resi-due engaged in the UG wobble pair) The libraries were
PCR amplified using the primer combinations designed for
each ribozyme, including OP757⁄ OP483 for Tth.L1925,
OP758⁄ OP479 for DiGIR2, OP759 ⁄ OP761 for Fse.L569,
and OP760⁄ OP762 for Fse.L1898 (Table 1) A human
lyso-somal a-mannosidase cDNA, cloned into the pcDNA3.1(–)
vector (Invitrogen, Oslo, Norway) under the control of the
cytomegalovirus (CMV) promoter, was transfected into
COS-7 cells and total RNA isolated after 24 h by the Trizol
reagent (Invitrogen) Total RNA (1 lg) was mixed with
GN4 ⁄ 5ribozyme-tag libraries (2 lm) were mixed in low-salt buffer, added to the total COS-7 RNA and incubated for trans-splicing at 37C for 3 h The trans-splicing reaction was reverse transcribed with primer OP299 using Super-script II RNaseH Reverse TranSuper-scriptase (Invitrogen) and amplified by PCR using different forward primers (OP764, OP765, or OP766) and a nested reverse primer (OP763) RT-PCR products were subsequently sequenced by the ABI PRISM BigDyeTerminator Cycle Sequencing Ready Reac-tion Kit (Perkin-Elmer, Oslo, Norway) running on an ABI Prism 377 system (Perkin-Elmer)
Table 1 Oligonucleotide primer sequences used in this study.
Name 5¢- to 3¢ sequence
Trang 10mentary oligonucleotide primers (primer combinations
respectively) The annealed primers were digested with NotI
and EcoRI and inserted into the corresponding sites in the
pcDNA3.1-D vector, generating pAS1357 and pAS1381
Wild type a-mannosidase 3¢-exon sequences, following
U1357 and U1381, were PCR amplified with primers
carry-ing the restriction enzyme sites BmgBI and BamHI (primer
combinations OP933⁄ OP890 and OP889 ⁄ OP890 for U1357
and U1381, respectively) The forward primers were used to
generate alternative codons for the first 16 and 15 triplets
following the target sites U1357 and U1381, respectively
The corresponding PCR products were digested with
BmgBI and BamHI and inserted between the BmgBI sites
generated by the insertion of EGS regions and the BamHI
sites in the multiple cloning sites of the vectors, resulting in
pAS1357a and pAS1381a Ribozymes were amplified and
inserted into pAS1357⁄ 1381a between the XbaI or NheI
(for U1357 and U1381) site and the BmgBI site The primer
Tth.L1925, OP936⁄ OP937 for DiGIR2, OP938 ⁄ OP939 for
Tth.L1925, OP893⁄ OP894 for DiGIR2, OP895 ⁄ OP896 for
products were then digested with XbaI and BmgBI and
ligated into the corresponding vector sites, generating
trans-splicing ribozyme constructs targeting U1357 and U1381 in
a-mannosidase mRNA All plasmid constructions were
con-firmed by automatic DNA sequence analysis
Oligonucleo-tide primer sequences are listed in Table 1 The in vitro
transcription plasmid, pMannRNAa, was generated by
amplifying a 1 kb region within the a-mannosidase cDNA
by using OP764 and mph306R, and subsequently the PCR
product was ligated downstream of the T7 promoter in the
pGEM-T easy vector (Promega, Madison, WI, USA)
pMannRNAa contains the T1357 and T1381 target
posi-tions used in this study
In vitro RNA trans-splicing and time-course
analysis
Precursor RNAs were transcribed from T7 promoters off
linearized ribozyme plasmids [35S]CTP[aS] (10 lCiÆlL )1;
GTP) As both molar ratio reactions for all four ribozymes worked equally well, the lowest molar ration (2 : 1) was selected for use in further analyses Samples were collected
at 0, 5, 15, 30, 60 and 90 min, and the reaction was termin-ated by the addition of an equal volume of STOP-solution (95% formamide, 50 mm EDTA, 0,02% xylene cyanol, 0.05% Bromophenol Blue) Reactions were denatured for
2 min at 88C and separated on 7 m urea ⁄ 5% polyacryla-mide gels, followed by autoradiography The PAGE-puri-fied RNAs [28], corresponding to the correct trans-spliced transcripts, were confirmed by RT-PCR and sequencing analysis using the primers OP1079 (for U1357) or OP948 (for U1381) for first-strand cDNA synthesis The
respectively
Trans-splicing and RPA
The trans-splicing ribozyme RNAs were treated with Turbo DNase (Ambion, Huntingdon, UK), according to the manufacturer’s instructions, phenol⁄ chloroform extracted, ethanol precipitated and dissolved in diethyl pyrocarbo-nate-treated water Unlabeled trans-splicing ribozymes and PAGE-purified a-mannosidase RNA were mixed at a 5 : 1 molar ratio under splicing conditions (40 mm Tris⁄ HCl
pH 7.5, 200 mm KCl, 2 mm spermidine, 5 mm dithiothrei-tol, 10 mm MgCl2, 0.2 mm GTP) and incubated at 50C for 3 h RPA was performed on 5 lL of trans-splicing RNA mix by using the RNase protection kit (Roche Applied Science, Penzberg, Germany), according to the manufacturer’s instructions The RPA probes were gener-ated from the RT-PCR products of in vitro trans-spliced a-mannosidase RNA at positions U1357 and U1381 (see above) cloned into the pGEM-T easy vector (Promega) These plasmids were linearized and transcribed, then labelled with [35S]CTP, as described above, to obtain RPA probes of larger sizes than probe fragments protected by trans-spliced RNAs in analysis by RPA RPA samples were separated on 7 m urea⁄ 5% polyacrylamide gels, followed
by autoradiography and phosphoimager quantification (Fuji BAS 5000 system; image gauge 4.0 software) The cytosine contents in the parts of the RPA probes protected
by the differently sized RNAs were calculated and included