RNA editing status of
nad7
intron domains in wheat mitochondria
RNA editing status of nad7 intron domains in wheat mitochondria
Catherine
Carrillo
and
Linda
Bonen*
Biology Department, University of Ottawa, 30 Marie Curie,
Ottawa
K1N 6N5,
Canada
Received August 13, 1996;
Revised and Accepted November 22, 1996
DDBJ/EMBL/GenBank accession no. X75036
ABSTRACT
The most highly conserved structures of group II introns are the helical domains
V and VI near the 3
'
splice site. Within this region of each of the four introns in the wheat
mitochondrial
nad7
gene encoding NADH dehydrogenase subunit 7, there are A-C mispairs. To determine whether C-to-U type RNA editing restores conventional A-U pairing, we sequenced RT-PCR products from partially-spliced
nad7
template RNA and gel-fractionated, excised intron RNA. We examined transcripts from germinating wheat embryos and seedlings because these
two stages of development show pronounced differences in steady state levels of
nad7
intronic RNAs. We observed editing at only two of the six predicted sites, and
they were located at homologous positions within domain V of the third and
fourth introns. A third site was found to be edited within the unmodelled
domain VI loop of the fourth intron. Similar patterns of RNA editing were seen
in wheat embryos and seedlings. These observations, and the presence of other
non-conventional base pairs particularly within domain V of plant
mitochondrial introns, indicate weaker helical core structure than in ribozymic
group II introns. Moreover, the incompleteness or absence of editing in wheat
nad7
excised intron RNA suggests that, although editing may contribute to splicing
efficiency, it is not essential for splicing.
INTRODUCTION
Virtually all plant mitochondrial introns identified to date have been
classified as group II and most are located within genes encoding subunits of the NADH dehydrogenase complex (reviewed in
1
,
2
). A subset of these have undergone DNA rearrangements so that mRNA maturation
requires splicing
in trans
(reviewed in
3
,
4
). Designation as group II is based on the presence of distinctive secondary-tertiary structural features and such introns are usually depicted as six
helical domains radiating from a central wheel (reviewed in
5
,
6
). Over 100 group II introns have been identified in organellar and bacterial
genomes and a few, although none in plant mitochondria, have been shown to self-splice
in vitro
(reviewed in
6
,
7
). The catalytic core is comprised of domain V and sequences within domain I, and splicing proceeds through two
transesterification steps, the first being initiated by the 2' hydroxyl group of a bulging adenosine within domain VI. A conformational
change between the two steps involves tertiary interactions between sequences
in domains VI and II (
8
). Group II intron excision is mechanistically similar to nuclear pre-mRNA splicing and it has been suggested that these two classes of introns
may share a common evolutionary origin.
Domain V is the most diagnostic structure of group II introns and it is
essential for splicing
in vitro
and
in vivo
. It is highly conserved in both sequence and length, with a 14 bp helix
interrupted by a 2 nucleotide (nt) bulge and closed by a purine tetraloop
(reviewed in
5
,
6
). The most conserved bases in this structure, numbered from the 5'-end of the helix, are A2, G3 and C4 which pair with U33, U32 and
G31 respectively. The bases A2, G3 and C4 have been shown by mutational
analysis to be crucial for splicing (
9
,
10
) and the terminal tetraloop, which is usually GAAA, interacts with a motif
within domain I (
11
). The nucleotide sequence of domain V is conservative to such an extent that
the identification of group II introns in data bank searches has been possible
(
12
). Domain VI, although less well-conserved and having a shorter helix, almost always possesses a bulging
adenosine which is located 7-8 nt from the 3' splice site and involved in lariat formation.
In addition to intron excision, either
in cis
or
in trans
, the maturation of plant mitochondrial mRNAs also involves C-to-U type RNA editing (reviewed in
2
,
13
,
14
). It occurs predominantly within protein coding sequences where it typically
changes the amino acid specified to one that improves similarity with
homologous proteins from other organisms. RNA editing is an early processing
event and fully-spliced transcripts are usually fully edited whereas unspliced ones may show partial editing (
15
,
16
).
There have been a few RNA editing positions identified within group II intron
sequences and it has been suggested that RNA editing may be required to improve
base pairing for proper folding needed in splicing (
2
,
3
). This model is conceptually appealing because intron editing would then serve
an important biological role, as it does in coding regions. Some, but not all,
of the RNA editing events observed within plant mitochondrial introns do
improve base pairing within the core structure. The conversion of A-C mispairs to A-U pairs has been observed at single sites within the domain I, IV or
VI helices of several Oenothera
nad
introns (
17
-
21
) and domain I of the potato
rps10
intron (
22
). Interestingly, editing at a site adjacent to the bulging adenosine in domain
VI of the Oenothera
nad1
intron 3 has been found to be necessary for the self-splicing of a chimeric yeast autocatalytic aI5c intron containing the
domain VI stem of this Oenothera intron (
21
). In several other cases, RNA editing weakens helical stability by converting G-C to G-U pairs (
17
,
23
), appears not to affect it (
24
), or editing sites are located outside the defined helical core structure and
cannot be assessed (
18
,
24
). It should be noted that a number of RNA editing sites have also been observed
in the
mat-r
gene which is located within the non-core part of
nad1
intron 4 and they follow editing patterns seen in other plant mitochondrial
coding sequences (
17
,
25
). No editing sites have previously been reported within the highly conserved
domain V.
In our characterization of the wheat mitochondrial
nad7
gene which contains four group II introns, we had predicted that C-to-U type editing would convert A-C mispairs observed within domain V/VI helices to Watson-Crick A-U base pairs (
26
). In this study, we have used RT-PCR sequencing analysis to examine the RNA editing status of these
intronic RNAs in wheat embryos and seedlings.
MATERIALS AND METHODS
Isolation of wheat mitochondrial RNA
Mitochondrial RNA was isolated from 24-h germinating wheat embryos and 6-day etiolated wheat seedlings (
Triticum aestivum
var. Frederick) as previously described (
27
). To remove residual DNA, RNA preparations were repeatedly precipitated for 2 h
with 2 M LiCl and treated with RNase-free DNAse I prior to phenol extraction, phenol-chloroform extraction and ethanol precipitation (
28
).
cDNA synthesis and PCR amplification
Synthetic oligonucleotide primers were designed from the wheat mitochondrial
nad7
gene sequence (
26
) and their positions and orientations are shown in Figure
1
A. Coordinates are according to the EMBL database accession number X75036.
Sequencing of RT-PCR products and cloned cDNAs
Samples were sequenced by dideoxynucleotide chain-termination using Sequenase version 2.0 (US Biochemicals). PCR products were
sequenced directly using a modification of the Sequenase protocol in which 75
ng primer were annealed with ~500 ng template in the presence of 0.5% Nonidet P-40. Samples were boiled for 3 min, quick-cooled in a -80oC ethanol bath, and then gradually warmed to room
temperature. The labelling cocktail was also modified to include four times
more labelling mix, ~0.5% NP-40, the supplier's Mn buffer and 1.5 times the suggested amount of
dATP. The primers used in sequencing reactions were the same intron primers as
in the corresponding PCR reactions, except in the case of the fully-spliced
nad7
mRNA, where an
nad7
exon c primer 5'-GCACAGCAAGCAAAGGATTG-3' (3361-3380, coordinates according to EMBL accession
no. X75036) was used.
Isolation and sequence analysis of excised intron RNA
Approximately 10 [mu]g wheat mitochondrial RNA, which had been repeatedly LiCl precipitated, were
electrophoresed on a 1.2% agarose gel containing 6.6% formaldehyde in 20 mM
MOPS buffer and RNA fractions of interest were isolated using the RNAid kit
(BIO 101). The purified RNA was treated with DNase I as above.
cDNA for the intron C region was synthesized using primer 5 and PCR was then
performed with primers 4 and 5 as described above except with an annealing
temperature of 42oC. To assess the contribution of residual contaminating DNA template in the
PCR reaction, we included a set of
nad1a
primers, designated in ref.
31
as A2 (5'-AAGGCTACTCCTAGTAGAAG-3') and A3 (5'-TCGGGTCGACCAGGTCAGGC-3'), with coordinates 198-179 and 86-105
respectively, according to EMBL accession number X57968. RT-PCR products were gel-purified for direct sequencing as described above. To assess the
presence of other
nad7
intron C-containing transcripts, primer 6 was included in RT-PCR reactions with primers 4 and 5, so that any contaminating
intron C/
nad7d
-containing transcripts (or DNA) would generate a 263 bp product.
RNA blot analysis
Northern hybridization experiments were performed using standard procedures (
28
). Approximately 4 [mu]g wheat mitochondrial RNA were electrophoresed on a denaturing formaldehyde
gel in MOPS buffer as described above. Hybridization probes were radioactively- labelled by strand-specific primer extension using M13 cloned DNAs, except in the case
of introns C and D respectively where the synthetic oligomers 5'-ATGCATGCTTTTGTAGGGTC-3' and 5'-CGTGTCAGCTTAGTTATC-3' (EMBL accession no. X75036 coordinates 3494-3475 and 4977-4960
respectively) were radioactively-labelled with T4 polynucleotide kinase and [[gamma]-
32
P]ATP.
Calculation of predicted thermodynamic stabilities of intron helices
The potential folding energies of domain V helices, constrained to the
conventional 14 bp helix with a 2 nt bulge (
5
), were calculated using the programs available on the `efn server' web page (http://www.ibc.wustl.edu/~zuker/rna/energy/form1.cgi) which were derived from algorithms described by
Zuker
et al
. (
32
).
RESULTS
Editing status of the domain V/VI regions of wheat
nad7
introns
We had previously established (
26
) that the wheat mitochondrial
nad7
gene contains four group II introns (Fig.
1
A) and has a mature mRNA of ~1.6 kb (Fig.
2
F) with 32 C-to-U edits in the coding sequence. When
nad7
intron-specific probes were used in RNA blot analysis (Fig.
2
B-E), we observed markedly higher levels of precursor transcripts and
excised intron RNAs (Fig.
2
, asterisks) in 24-h germinating wheat embryos than in 6-day etiolated seedlings, relative to constant levels of 18S
ribosomal RNA (Fig.
2
A). In addition, the RNA profiles from these two stages show qualitative
differences, most notably in the region of excised intron D RNA species (Fig.
2
E asterisk, lanes 1 versus 2). These developmentally-specific differences in the abundance and apparent complexity of
nad7
intronic transcripts may reflect differential RNA stabilities or RNA processing
pathways, and raise the possibility of a role for intron editing if such events
are important for proper folding for splicing as has been proposed (
2
,
3
).
Editing within the
nad7
terminal exon in precursor and fully-spliced RNA
The RT-PCR product used in the analysis of intron D (see above) contained four
exon
nad7e
editing sites (
26
). One of these, located three nucleotides downstream from the intron-exon border appeared to be unedited as judged by direct sequence analysis
(Fig.
5
A, open arrow) in contrast to the full editing seen in completely-spliced
nad7
mRNA (Fig.
5
B, black arrow) and the nearly-complete editing seen at the intronic site within the loop of domain VI
(Fig.
5
A, black arrow). Moreover, none of 22 clones examined from embryo or seedling
RNAs showed editing at this position. In contrast, the other three exon sites
showed >50% editing in these RNAs, which would include early unspliced
precursors, but all of which were known to have the
Nhe
I site generated by editing. These observations are consistent with editing
being a relatively `late' RNA processing event at that exon site, and suggest
that the context of sequences around the exon junction may be important for
editing and/or splicing.
DISCUSSION
We observed C-to-U type RNA editing at only three sites within the domain V/VI
regions of the four wheat
nad7
introns. One occurs within the unmodelled domain VI loop in intron D, and the
other two convert A-C mismatches to A-U pairs at identical positions within the domain V helix of introns
C and D. This is the first report of RNA editing within domain V of any plant
mitochondrial intron. Although these two events improve helicity and so may be
important for splicing, they do not generate the structure that is invariably
present in this catalytic domain of ribozymic group II introns. Moreover,
incomplete editing of excised intron RNA indicates that such editing events are
not essential for splicing. In addition, four other A-C mispairs within these helices show no editing, so Watson-Crick base pairing at those positions must not be required for proper folding for
splicing.
In the case of intron C, the single editing event shifts the calculated free
energy of domain V from -0.6 to -3.4 kcal/mol, which is still markedly less stable than those of
ribozymic introns, which have values of -13.0 to -17.9 kcal/mol. Editing at the other two A-C mismatches (Fig.
1
B, open arrows) would have resulted in a free energy of -11.6 kcal/mol. Interestingly, this intron which has such weak domain V
structure in wheat mitochondria, is absent from the
nad7
gene of tobacco although the other three are present (
34
). For intron D, the edited helix is considerably more stable (free energy of -7.0 kcal/mol) than the unedited form (-1.7 kcal/mol). The domain V sequences of introns A and B, which showed no
editing, have somewhat more stable helices, with free energies of -13.6 and -9.4 kcal/mol respectively.
ACKNOWLEDGEMENTS
This work was supported by funding from the Natural Sciences and Engineering Research Council of Canada. The expert technical assistance of K. Williams is acknowledged for data shown in Figure 5B.
REFERENCES
1 Bonen,L. and Brown,G.G. (1993) Can. J. Bot.71, 645-660.
