ABSTRACT
Larch mitochondria contain a `native' tRNA
His
which is absent from angiosperms. Sequence comparisons of genomic DNA and cDNA
obtained from unprocessed primary transcripts of the larch mitochondrial gene
trnH
encoding this tRNA revealed three nucleotide discrepancies. These three
nucleotide alterations, in the acceptor stem, D stem and anticodon stem
respectively, are conversions of genomic cytidines to thymidines in the cDNA
(uridines in the tRNA) and thus resemble the RNA editing events observed in
nearly all plant mitochondrial mRNAs. Two cases of editing affecting
mitochondrial tRNAs from angiosperms have already been described, but we
present here the first example of such events in a gymnosperm mitochondrial
tRNA. All three editing events correct mismatched C[middot]A base pairs which appear when folding the gene sequence into the
standard cloverleaf structure, thereby improving the secondary structure of the
tRNA
.
When incubated with a heterologous potato mitochondrial processing extract,
only the edited form of the larch mitochondrial tRNA
His
precursor was efficiently processed
in vitro.
These data strongly suggest that editing of larch mitochondrial tRNA
His
is a prerequisite for its processing.
RNA editing is a widespread phenomenon, occurring in organelles of many
organisms, that is defined as the alteration of a transcript such that its
sequence differs from that of the gene from which it was transcribed (reviewed
by
1
). In plant mitochondria and chloroplasts, RNA editing nearly always consists in
C to U transitions, probably by a cytosine deaminase-like activity (
2
-
5
). In other organisms, other types of editing exist, such as insertion/deletion
of U or C residues (
6
) and the mechanisms almost certainly differ from those at work in plants (
1
). Most editing events occur in mRNA coding sequences and serve to improve
sequence similarity with the corresponding sequences from other organisms.
However, examples do exist of editing of rRNAs (
7
) and tRNAs (
8
,
9
). Editing of structural RNAs is difficult to define, as rRNAs and particularly
tRNAs undergo many post-transcriptional base modifications, many of which are important for
correct structure and function. Generally, the term `RNA editing', when applied
to tRNAs, is restricted to base changes which greatly resemble those seen in
mRNAs in the same organism and which are therefore probably brought about in
the same manner. By these criteria, only two cases of tRNA editing have come to
light in plant mitochondria; a C to U alteration in tRNA
Phe
, observed in potato, bean (
9
) and
Oenothera
(
8
), and a C to U alteration in tRNA
Cys
(
8
). The small number of examples makes it difficult to draw any general conclusion about the process.
Most studies on higher plant mitochondria have involved angiosperms for the good
reason that they are easy to cultivate and are relatively good sources of
mitochondria and mitochondrial nucleic acids. Gymnosperms, although widespread
in nature and economically important in many regions, have received little
attention because of the technical problems involved in the extraction of
mitochondria. Nevertheless, RNA editing of presumably mitochondrial transcripts
in conifers has been confirmed by PCR amplification of DNA and cDNA from total
nucleic acid preparations (
10
,
11
). The finding that embryogenic cell suspension cultures of larch make
reasonable sources of pure mitochondria (
12
) has allowed more detailed studies to be performed. During an attempt to
catalogue the tRNAs present in larch mitochondria (
13
), we came across a larch mitochondrial gene coding for tRNA
His
. This gene is homologous to the tRNA
His
gene from liverwort (
14
), and is presumably derived from the genome of the original endosymbiont
ancestor of plant mitochondria. This is the first time a `native' mitochondrial
tRNA
His
gene has been found in a higher plant, and the study of its sequence and
expression has given interesting insights into RNA editing in plants, and the
reasons why plant mitochondria contain a varied set of tRNAs from three
different genetic origins.
Isolation of larch (
Larix x leptoeuropaea
) mitochondrial DNA as well as `Southern blot' hybridization were performed
under conditions previously described (
12
). To clone the tRNA
His
gene, a partial larch mtDNA library was prepared and screened using standard
techniques (
15
). Five micrograms of larch mtDNA were cut with
Bam
HI, and fragments between 3 and 4 kb were isolated from an agarose gel and
cloned into the
Bam
HI site of pUC19. The resulting clones were screened with an [[alpha]-
32
P]UTP-labelled RNA probe prepared by
in vitro
transcription of a clone of the liverwort
trnH
gene. Two hybridizing clones were retained and subclones containing the larch
trnH
gene were sequenced using fluorophore-labelled primers in a semi-automated sequencing system (Applied Biosystems, Foster City, CA).
Larch mitochondria purified according to (
12
) were used for RNA extraction. This extraction was performed as described in (
16
). For cDNA synthesis, 2 [mu]g of the larch mitochondrial RNA were used for reverse transcription (
17
) in the presence of oligonucleotide 1, containing a
Bam
HI restriction site (underlined) attached to a sequence complementary to the
downstream flanking region of the larch mitochondrial tRNA
His
gene, GTACT
Uniformly labelled RNA transcripts were synthesized by standard
in vitro
transcription with T7 RNA polymerase in the presence of 20 [mu]Ci [[alpha]-
32
P]UTP. The cloned DNA fragments obtained by reverse transcriptase-PCR
amplification (see above) of the edited or non-edited precursor forms of larch mitochondrial tRNA
His
were used as templates after linearization with
Pst
I. Following transcription, the synthesized RNAs were fractionated by
electrophoresis on a sequencing gel. After identification by autoradiography, the bands corresponding to the expected products were excised
from the gel and the labelled RNAs eluted (
20
).
To perform primer extension analysis of the processing products and to analyse
incorporation of [[alpha]
32
P]CTP by tRNA nucleotidyl transferase during
in vitro
processing of tRNA precursors, unlabelled transcripts were prepared as
described above except that [[alpha]
32
P]UTP was replaced by 250 [mu]M UTP.
To prepare enzymatic processing extracts, mitochondria were purified from 2.5 kg
of potato tubers according to (
21
). Mitochondrial lysates able to perform tRNA processing were obtained
essentially as described by Hanic-Joyce and Gray (
22
) and processing assays were carried out according to (
16
).
To test for incorporation of the CCA end into the processed products, processing
assays were performed as described above but with unlabelled precursor RNAs as
substrates and in the presence of 40 [mu]Ci [[alpha]
32
P]CTP and 40 [mu]M CTP.
The mature tRNA-sized product obtained from unlabelled edited tRNA
His
precursor was recovered from a sequencing gel (
20
) using as a reference the migration of the same product derived from the
corresponding labelled tRNA
His
precursor. Primer extension was carried out as described in Remacle and Maréchal-Drouard (
17
) using one fifth of the gel-purified processing product as a template and oligonucleotide 3
(GGCGAATAACGGGATTCG) complementary to the 3'-end of larch mitochondrial tRNA
His
as a primer.
A labelled RNA probe obtained by
in vitro
transcription of the mitochondrial
trnH
gene from liverwort (
Marchantia polymorpha
) hybridized to mitochondrial tRNA from larch but not to mito- chondrial tRNAs from angiosperms (
13
). The same probe also hybridizes to larch mitochondrial DNA (mtDNA) (Fig.
1
). Only one hybridizing fragment was obtained with a variety of restriction enzymes, so we assume that there is only one copy of the gene. A plasmid
library of size-selected
Bam
HI fragments from larch mtDNA was screened with the liverwort
trnH
probe, and a plasmid containing a hybridizing insert was recovered and
sequenced (Fig.
2
A, EMBL accession no. Z70031). By comparing this sequence to the
trnH
sequence from liverwort and by folding it into the typical cloverleaf tRNA
structure (Fig.
2
B), it appeared that the folded larch sequence contains three mismatches, C
6
[middot]A
67
in the acceptor stem, C
12
[middot]A
23
in the D stem, and A
29
[middot]C
41
in the anticodon stem, whereas in these three locations the liverwort sequence
contains U[middot]A Watson and Crick pairs. These three mismatches could potentially be
repaired by classical C to U RNA editing. In addition, the usual R
15
[middot]Y
48
Levitt tertiary base pair (
23
) is predicted to be A
15
[middot]C
48
in the larch structure, which again could be potentially corrected by C to U
editing. The corresponding base pair in the liverwort tRNA is G
15
[middot]C
48
, as in almost all other tRNA
His
from bacteria to higher eukaryotes (
24
).
Larch mitochondrial tRNA
His
precursors were amplified by RT-PCR, cloned and sequenced (Fig.
3
). Some sequenced molecules were identical to the gene sequence, but six out of
11 presented C to T transitions at positions corresponding to residues C
6
, C
12
and C
41
in the tRNA
His
sequence. Such alterations are indicative of RNA editing. No evidence of
partially edited precursors was found, nor was there any evidence found for
editing of C
48
or any of the other C residues in the sequence, for example C
13
or C
21
where U is found in the corresponding liverwort mt tRNA
His
sequence.
In plant mitochondria, only two cases of single editing event affecting tRNAs
have been described so far, namely for tRNA
Phe
(GAA) in potato, bean and
Oenothera
and tRNA
Cys
(GCA) in
Oenothera
(
8
,
9
). The larch mitochondrial tRNA
His
studied here displays the most extensive RNA editing yet seen in a tRNA from a
higher plant. Although edited tRNAs have been discovered in mitochondria from
other organisms (
28
-
32
), the pattern and type of editing is not the same and the mechanisms
responsible are certainly different. The C to U editing seen here is very
reminiscent of the editing seen in plant mitochondrial and chloroplast mRNAs.
As the mechanism by which RNA editing occurs in plant organelles is still
unclear, it is interesting to note some features of the editing of larch tRNA
His
. Firstly, all three editing sites are at C[middot]A mismatches in potentially double-stranded regions. Similarly, editing of tRNA
Phe
in potato, bean and
Oenothera
takes place at a C[middot]A mismatch (
8
,
9
) and tRNA editing in
Acanthamoeba
also appears to involve mismatch repair in double-stranded regions, although editing in
Acanthamoeba
is limited to the acceptor stem, but is not limited to C to U changes (
30
). RNA editing in trypanosomes involves guide RNAs; editing relies on mismatches
(between the guide RNA and the transcript being edited) to direct insertion or
removal of uridines at specific sites. Mismatch repair is also an attractive
hypothesis for the targeting of C to U editing in plant organelles, although no
evidence has been found for the existence of guide RNAs or local regions of
complementarity which could direct the editing machinery. The editing system of
apolipoprotein B transcripts in mammals (
33
,
34
), which biochemically most closely resembles that seen in plants, does not
involve complementary base pairing. Larch mitochondrial tRNA
His
is the best example to date in the plant kingdom where complementary base
pairing might be involved. It is also interesting to compare the larch tRNA
His
editing sites to the consensus obtained from a systematic study of 61 non-homologous editing sites (
35
). The first editing site is preceded by a G, and the third by an A which is
rather unusual as plant organellar editing sites nearly always follow a
pyrimidine, and a G preceding the editing site is extremely rare (for a review
see
36
).
Transfer RNA
His
is unique among tRNAs in that it contains an extra base pair in the acceptor
stem between G
-1
and C
73
, the discriminator base. This feature is essential for recognition by histidyl-tRNA synthetase (
37
,
38
), and is conserved between prokaryotes and eukaryotes, with the difference that
G
-1
is added post-transcriptionally in eukaryotes (
39
), whereas in prokaryotes and in chloroplasts it is present in the precusor
transcript. G
-1
is retained in the mature tRNA because the bacterial and plastid RNase P cuts
tRNA
His
precursors between -2 and -1 rather than between -1 and +1 as in all other precursor tRNAs (
40
,
41
). Angiosperm mitochondrial `cp-like'
trnH
genes (
42
) and both the liverwort and larch `native' mitochondrial
trnH
genes also contain a G at the -1 position. Although these data could suggest that, as in bacteria and
chloroplasts, the mature mitochondrial tRNA
His
contains G
-1
due to an unusual cleavage by RNase P the primer extension results on the
processed product from the larch mitochondrial tRNA
His
precursor leave open the possibility that the extra G is added post-transcriptionally. Two termini, corresponding to G
1
and C
2
were found and no product ending at G
-1
was detected. This may be due to the use of a heterologous (potato) processing
extract; potato mitochondria contain a `cp-like' tRNA
His
which is quite different from the larch mitochondrial tRNA
His
, and the motifs needed for correct processing by RNase P may differ somewhat
between the two species. Therefore, the absence of a product ending at G
-1
does not exclude the possibility that,
in vivo
, the larch mitochondrial RNase P cuts tRNA
His
precursors between -2 and -1. Indeed, chloroplast RNase P does not process bacterial tRNA
His
correctly, and it has been suggested that the highly conserved extended
acceptor stem (G
-2
[middot]C
74
, G
-3
[middot]C
75
) characteristic of chloroplast tRNA
His
precursors (and mitochondrial `cp-like' tRNA
His
precursors) is necessary for correct processing (
40
). Neither the larch nor the liverwort mitochondrial
trnH
transcripts can form such an extended acceptor stem.
In mRNAs, RNA editing is found almost exclusively in coding sequences, where it
generally serves to improve the similarity of the encoded protein sequence with
those found in other organisms. There are also examples of the creation of
start or stop codons in transcripts, and even examples of editing in non-coding sequences where it can be assumed that the editing is essential for
correct expression (for reviews, see
1
-
5
). We have shown that, as in the case of the
Oenothera
and potato mitochondrial tRNA
Phe
(
16
,
27
), editing of larch mitochondrial tRNA
His
precursors is an absolute prerequisite for processing and thus for the
accumulation of a functional tRNA. Given that RNA editing is now essential for
the expression for many plant mitochondrial genes, it has been suggested that
plant mitochondria have become `addicted' to the process (
4
), and a hypothesis to explain how this might have occurred has been proposed (
43
). At least as regards tRNA genes, it seems that plant mitochondria might be
trying to `kick the habit'. The native gene found in larch is absent from
angiosperms (
13
,
44
,
45
), and the edited tRNA
Phe
and tRNA
Cys
genes have also been lost from many plant lineages, replaced by tRNA genes
carried on inserted fragments of plastid DNA, or by nuclear genes coding for
imported cytosolic tRNAs (
13
,
44
,
46
). There is a clear evolutionary tendency towards functional replacement and
subsequent loss of native mitochondrial tRNA genes. The primitive plant
Marchantia polymorpha
contains 29 mitochondrial tRNA genes (
14
), whereas maize contains only 10 (
47
) and sunflower 11 (of which 10 are expressed) (
48
). It seems that for tRNAs natural selection prefers alternative modes of
expression which avoid RNA editing.
We wish to thank Prof. J. H. Weil and A. Dietrich for critical comments and
careful reading of the manuscript and A. Cosset for her expert technical
assistance. The gift of the
M.polymorpha trnH
clone by Dr Ohyama is gratefully acknowledged. Claire Remacle was supported by
the Fonds National de la Recherche Scientifique (Belgium) as a Senior Research
Assistant.
REFERENCES
Return