OUP user menu

The reverse transcriptase encoded by ai1 intron is active in trans in the retro-deletion of yeast mitochondrial introns

Ali Gargouri
DOI: http://dx.doi.org/10.1016/j.femsyr.2004.11.012 813-822 First published online: 1 June 2005


Genomic mitochondrial intron deletion occurs frequently during the reversion of mitochondrial intronic mutations in Saccharomyces cerevisiae. The multiplicity as well as the apparent polarity of intron deletion led us to propose the implication of reverse transcription in this process. The two first introns of the COX1 (cytochrome oxidase I) gene, ai1 and ai2, are known to be homologous to viral reverse transcriptase and to encode such activity. We have tested the involvement of these introns in the deletion process by constructing three isogenic strains. They contain the same reporter mutation in the second intron of the CYTb (cytochrome b) gene but differ from each other by the presence or the absence of the ai1 and/or ai2 introns in the other gene encoding the COX1 subunit. Only the strain lacking ai1 and ai2 introns is no more able to revert by intron deletion. The strain retaining only the ai1 intron was able to revert by intron deletion. We conclude that the reverse transcriptase activity, even when encoded by only ai1 intron, can act in trans in the intron deletion process, during the reversion of intronic mutations.

  • Mitochondrial introns
  • Reversion
  • Deletion
  • Reverse transcriptase

1 Introduction

Many mitochondrial intronic mutants can revert frequently by clean deletion of the mutated intron, a process we have called in the past “DNA splicing”[1]. Introns can be deleted from the gene during the reversion of intronic mutations in Saccharomyces cerevisiae[15] and in Schizosaccharomyces pombe[6,7]. In the majority of revertants, we isolated from CYTb intron mutants, neighbouring safe introns are also deleted in a somewhat polar fashion: single bi1 intron deletions characterise revertants of bi1 mutants; (bi1, bi2) or (bi1, bi2, bi3) deletion characterise revertants of bi2 mutants; and the deletion of (bi1, bi2, bi3) introns is found exclusively in revertants of bi3 mutants. This polar fashion resembles the polarity of the splicing of the CYTb pre-mRNA (see in [8]). Because of these two features of reversion by intron deletion, “multiplicity and polarity”, we have proposed that the process requires an RNA intermediate, and we have therefore argued for the involvement of a hypothetical mitochondrial reverse transcriptase activity [1]. Such an activity has been discovered in Neurospora associated with mitochondrial plasmids [9], in some strains of Escherichia coli[10] and in yeast mitochondria [11]. The second intron of the COX1 gene, ai2, has been shown in vitro and in vivo to encode reverse transcriptase activity while, surprisingly, its very closely homologous intron, ai1, encodes an inactive protein in S. cerevisiae, according to [11]. However, ai1 intron protein from other yeast strains has been shown to be slightly active as a reverse transcriptase [11]. More recent results showed that the ai1 protein was involved in the retro-transposition of its own intron by reverse splicing directly into an ectopic DNA site [12]. On the other hand, ai2 reverse transcriptase was found to be involved in the mobility of both introns ai1 and ai2 [14,15]. The ai2 protein was also shown to be able to use a non-cognate template in a reverse transcription process, a result which has great implications in precise intron deletion processes [16].

Before the biochemical demonstration, indirect evidence already existed, consisting of the finding of similarity between some mitochondrial introns and viral reverse transcriptase genes [17]. The highest similarity found concerned the two first introns of the mitochondrial COX1 gene: the ai1 and ai2 introns [18]. The group-II introns sharing reverse transcriptase domains have been found in some bacterial, chloroplastic and mitochondrial genomes (see [19] for a review).

We have made use of these findings by testing the direct involvement of the ai1 and ai2 introns in the process of reversion by intron deletion. In this study, we have addressed the following question: is the reversion, by intron deletion, of a reporter mutation in the second intron of the CYTb gene still possible in the absence of the two first introns ai1 and ai2 of the COX1gene? We show that this is not possible, we also make clear that the ai1 intron alone is able to re-establish reversion by intron deletion, demonstrating that it encodes a fully functional reverse transcriptase.

2 Materials and methods

2.1 Strains and media

All the strains used derive, mitochondrially, from the strain 777-3A: (α) adel, op1[20]. W303-1B is a rho+, (α) leu2, trp1, ura3, his3, CanR (constructed by Rothstein) and JC8/55 is a rho° (a) leu1, CanR[21].

Media: YPDif: 1% Yeast extract, 1% Bacto – peptone, 2% glycerol, 0.1% glucose. YPGA: as YPD, without glycerol but with 2% glucose, 20 μg ml−1 adenine. N3: as YPD but without glucose.

2.2 Origin of the CKV277 strain

This strain has the 777-3A/V277 mitochondria and the JC8/55 nucleus. V277 is a cis dominant mutation localised in the ai1 intron, the first intron of the COXI gene (4). In order to eliminate the effect of op1 mutation (in the gene AAC2, [22]) of the 777-3A nucleus [20], V277 was passed by cytoduction in JC8/55, a rho° strain having the op1+ and kar1-1 alleles, and named CKV277.

2.3 Construction of the TIM3041 rho-strain

This strain was isolated by ethidium bromide (EtBr) treatment (according to [23]) of the CK247 strain (cytoductant with the 777-3A/M3041 mitochondria in a JC8/55 nuclear background). This “petite” has lost the COX2, COX3 and COX1 genes since it cannot restore, by recombination in crosses, the respiration of mutants in each of these genes (V25, V276 and V277 mutations, respectively). On the other hand, it covers neighbouring bi2 mutations like M2075 [24] and W91 [25], situated on either sides of the M3041 mutation.

2.4 Independent revertant selection

About 100 cells of a fresh mutant culture were spread on YPDif plates and incubated at 28 °C. On such rich medium with 0.1% glucose and 2% glycerol as carbon sources, respiratory-deficient cells formed colonies by consuming the glucose by fermentation (we refer to these colonies as mutant colonies). These colonies stopped growing after a few days and contained ca. 106 to 107 cells per colony. In some mutant colonies, revertant cells “emerged” as papillae and continued to grow up by using the glycerol as a respiratory carbon substrate. The revertants were picked and disposed in grids on N3 plates. Further replica on the same medium were carried out to discard the “residual” parental cells. The revertant colonies were then analysed by “in-situ hybridisation”[26] or by a rapid minilysate DNA extraction (see below).

2.5 Yeast DNA extraction by minilysate procedure

This involved minor modifications of the method of mit DNA extraction [3]. The cells were cultured overnight in 2 ml YPGA, 1 ml was centrifuged, cells were collected, washed in water and then resuspended in 1 ml protoplast buffer: 1 M sorbitol, 50 mM citric acid, 150 mM K2HPO4, 10 mM EDTA, 0.1% mercaptoethanol and 0.3 mg ml−1 of Zymolyase. After 20 min at 37 °C, protoplasts were centrifuged in an “eppendorf” tube and resuspended in 400 μl lysis buffer: 150 mM NaCl, 10 mM Tris–HCl, pH 7.5, 5 mM EDTA and 1% sarkosyl detergent. Cells were broken by vigorous shaking and the proteins extracted twice by Tris-saturated phenol. Traces of phenol were removed by two diethyl ether treatments. Total DNA was precipitated with 1/10 volume of 3 M sodium acetate, pH 5.2, and 2.5 volumes of ethanol. After freezing, DNA was centrifuged, vacuum dried and suspended in 50 μl of water or 10 mM Tris–HCl, 1 mM EDTA, pH 7.5. Routinely, 5–10 μl were used in restriction analysis.

2.6 Restriction, electrophoresis, transfer and hybridisation

All of these procedures were carried out as described by [27] with minor modification, as for the “minilysate” DNA restriction higher NaCl concentration were needed. Gels were made with 0.7% agarose in Tris–acetate–EDTA. Gel transfers on nitrocellulose filter were carried up with 20 × SSC buffer. Filters were hybridised at 65 °C overnight in: 6 × SSC, 1 × Denhardt, 0.1% SDS and 100 ug/ml of sonicated Salmon sperm DNA. Filters are washed at 65 °C three times 20 min in 2 × SSC, 0.1 SDS and once in 0.1 × SSC.

2.7 Probes

Four pure intronic probes were used. The first, pYJL5 is bi2 specific and consists of the B231 petite [25] cloned in a pBR322 plasmid. The second, pCH1 donated by C. Jacq, is a bi4 fragment inserted in pBR322 plasmid. The last two probes donated by P. Netter are respectively ai1 and ai2 specific. Radioactive probes were prepared by nick translation [27].

3 Results

3.1 Construction of the strains

The first step was to obtain a strain devoid of the first two introns (ai1 and ai2) in the mitochondrial COX1 gene (Fig. 2). This was achieved by isolating a revertant from strain CKV277 strain containing a cis-dominant mutation, 777-3A/V277, of the mitochondrial ai1 intron (see Section 2). Indeed, the CKV277 mutant reverts frequently and one revertant named AV1 was isolated in which the ai1, ai2 and ai3 introns were deleted. Restriction analysis and hybridisation checked this deletion, with pure intronic probes. In order to perform further crosses, the AV1 mitochondria were passed by cytoduction in an α-nuclear background, that of W303-1B, and named RV1.

Figure 2

Strategy of strain construction. The strategy of construction of two isogenic strains both containing the “reporter” M3041 mutation in the CYTb gene and differing by the absence (GH1) or presence (WG1) of one intron in the COX1 gene is shown schematically. The cell is represented by a rectangle, the nuclear genotype by a smaller one (at the corner) and, from the mitochondrial genome, only two bars represent the CYTb and COX1 genes. The introns are shown as triangles; the black ones represent the mutated introns. Note that the two constructed strains have the same nuclear genotype, by virtue of the cytoduction procedure.

The second step consisted of the introduction of a tester mutation for reversion by genomic intron deletion in the RV1 mitochondria. Such a tester mutation should be able to revert frequently by genomic intron deletion as well as by other mechanisms like back mutation or extragenic suppression, which would constitute an internal control. Therefore, we have chosen a frameshift mutation in the bi2 maturase, 777-3A/M3041 [2] which reverts frequently by genomic intron deletion (80% of revertants are simultaneous deletions of introns bi1 and bi2 or introns bi1, bi2 and bi3 of the CYTb gene [3]) and more rarely by extragenic suppression [3,28] (see Fig. 2).

Genetic crosses between these two strains performed the introduction of the M3041 mutation in the RV1 strain, in two distinct manners:

(1) Introduction via rhostrain: Firstly, a rho“petite” containing the M3041 mutation was isolated, after EtBr mutagenesis of the CK247 strain, and called TIM3041 (“Section 2, Fig. 2). TIM3041 was then crossed to RV1 and the respiratory-deficient α-cytoductants were identified; one of them was isolated and called GH1. The absence of ai1, ai2 and ai3 introns was checked once more by the same restriction blotting procedure (Fig. 3, note that the GH1 pattern is the same as AV1 or RV1). The presence of the M3041 tester mutation was checked genetically by crossing with another “petite”, B231, which has a basic repeat length of only 420 bp (data of J. Lazowska in [3]) and known to cover the M3041 mutation [2]. Crossing of GH1 with B231 gave respiratory-competent diploids.

Figure 3

Restriction and hybridization demonstrate the presence or absence of introns ai1 and ai2 in the COX1 gene of constructed strains. (a) The mitochondrial DNA of the two control strains (wt and CW51) and the two constructed strains (GH1 and WG1) was isolated and digested by the HhaI enzyme. The agarose gel was blotted in “sandwich” to nitrocellulose filters. Each filter transfer was hybridized either to the pure ai1 or ai2 intronic probe and auto-radiographied. Panel (b) gives the restriction map of the gene containing all COX1 introns and of those deleted for some introns. Exons are in black, intronic ORF and BRF (blocked reading frame) are in grey and white, respectively; the COX1 surrounding region is in dark grey.

(2) Introduction via a rho+ strain: The second way was to cross CK247 (containing the M3041 mitochondria in the JC8/55 nuclear background) with RV1 and isolate the α-cytoductant recombinant strains which were restored by the B231 petite (Fig. 2). Among ten such recombinants tested by HhaI restriction analysis, one strain, WG1, was isolated. It has the M3041 mutation in the bi2 intron (checked by genetic crossing with the B231 petite) but has acquired the ai1 intron. Fig. 3 shows in fact that the ai1 probe gave a hybridisation signal with a new HhaI fragment (D) of 7.7 kbp in the WG1 profile. The ai2 and ai3 were still absent since no hybridisation was seen with the ai2 and ai3 intronic probes. We can also note the reappearance of the 8.6 kbp fragment (comprising the beginning of the COX1 gene) due to the HhaI site in the 5′ end of the ai1 intron (fragment C on Fig. 3).

So from these crosses we have constructed two isonuclear strains GH1 and WG1, both having acquired the same tester mutation in the bi2 intron and lost the ai2 and ai3 introns; the only difference resides in the ai1 intron which was also lost in GH1 and was retained in WG1. These strains are isonuclear since they resulted from cytoduction of the recombinant mitochondria in cells containing the same W303-1B nucleus.

As a control, a third isonuclear strain was constructed by cytoducting the 777-3A/M3041 mitochondria into W303-1B; it was called CW51. Mitochondrially, it had the M3041 mutation and a full complement of introns in the COX1 gene. In order to test the effect of the nuclear background, a second control strain, named BA1, was also constructed by cytoduction of GH1 mitochondria into the JC8/55 nuclear background.

3.2 Reversion of the constructed strains: the intron deletion annihilated in strains lacking ai1 and ai2 introns

To test the involvement of the ai1 and/or ai2 reverse transcriptases in the DNA splicing process, we analysed independent revertants from each constructed strain. This independence is essential to estimate the real frequency of each type of reversion event. Therefore, we adopted a method of plating individual cells which ensures the independence of revertants (see Section 2 and Fig. 1(a)). It is important to note that each revertant clone emerged from a mutant colony originating from a single cell, so that revertant clones obtained from different mutant colonies necessarily represent independent reversion events.

Figure 1

(a) Independent revertants isolation on YPDif plates. After about 10 d incubation of the mutant colonies (about 100 colonies per plate) on YPDif medium, revertant colonies, shown by arrows, appeared and grew. They were then picked randomly and arranged on a glycerol medium (N3 plates), their DNA extracted and subjected to the analysis shown in panel (b) or in Fig. 4. (b) A routine example of HhaI restriction analysis of minilysate total DNA extracted from independent revertants.

We tested about 30 independent revertant clones per genetic construction, which provided a statistically significant sample, allowing to disclose the occurrence of the intron deletion phenomenon. The presence/absence of introns in mt.DNA was visualised by a direct HhaI restriction analysis of the total DNA extracted by a minilysate procedure from independent revertants (a routine gel is shown in Fig. 1(b)). Note that the HhaI restriction site (CGCG) is very rare in yeast mit. DNA but very frequent in nuclear DNA; this enabled us to distinguish easily between mitochondrial fragments and nuclear ones, amongst which the repetitive rDNA are well discerned (Fig. 1(b)). Since the CYTb gene is contained in a single HhaI fragment of about 13 kb, each intron deletion decreases the length of this fragment (Fig. 1(b)). The different fragments could also be visualised by hybridising the transferred gel with the bi4 intron probe (Fig. 4). Notice that the bi4 intron is never deleted since it is required in the RNA-splicing of the ai4 intron in the COX1 gene [29,30]. The bi4 genomic deletion was exceptionally obtained in two cases: in a strain having lost the ai4 intron or in a strain bearing the Nam2-1 active suppressor in the nucleus, which alleviates the absence of bi4 (for a more detailed account of this interplay, see [23]).

Figure 4

Restriction and hybridization analysis of revertants. Total yeast DNA from independent revertants derived from strain CW51 (panel a) and those derived from strain GH1 (panel b) was restricted with HhaI, migrated on agarose gel and transferred to nitrocellulose filter. Filters were hybridized with the bi4 pure intronic probe and autoradiographied. In (a), specific mt DNA fragments of various lengths (see panel c for the expected fragment lengths) which correspond to the deletion of CYTb introns are found in independent revertants. In (b), all independent revertants have retained CYTb introns. In (c), the restriction map of the gene containing all CYTb introns (noted +bi 1, 2, 3) and of those deleted for some introns. Exons are in black, intronic ORF and BRF (blocked reading frame) are in grey and white, respectively; the Cytb surrounding region is in dark grey.

The revertant clones were picked randomly without selecting for fast- or slow-growing colonies, as we did not know the effect of the different intron configurations on the growth of the revertants. From Table 1 showing the results of revertant analysis, we can conclude:

View this table:
Table 1

Frequency of clean genomic deletion of mitochondrial introns in the CYTb gene as a function of the presence/absence of the ai1 and ai2 in the COX1 gene

StrainMitochondrionReversionIntron deletion (%)
CK247M3041, (ai1, ai2, ai3)+5040 (80)
CW51//, //1813 (72)
GH1//, Δ(ai1, ai2, ai3)220 (0)
BA1//, Δ(ai1, ai2, ai3)450 (0)
WG1//, ai1+Δ(ai2, ai3)386 (16)
  • 173 independent revertants of the CYTb intronic mutation M3041 in the strains containing or not the COX1 introns were analyzed by various procedures: by HhaI restriction analysis (Fig. 1(b)) that could be followed by Southern hybridization (as in Fig. 4). Deletions in CYTb introns did not occur in the absence of introns ai1, ai2 and were less frequent when only ai1 was present. The deleted revertants were of only two types in all cases: Δ(bi1, bi2) or Δ(bi1, bi2, bi3) in the ratio 2 to 1, respectively. The deletion of bi2 or (bi2 + bi3) was never found among these revertants.

● The positive control strains, CW51 and CK247, which contain introns ai1 and ai2, revert frequently by genomic intron deletion as expected. The frequency of deletion is the same in both strains (χ2 homogeneity test is not significant: 0.4 for 1 degree of freedom); 78% of all reversion events are due to the genomic intron deletion.

● The GH1 and BA1 strain, both lacking the ai1, ai2 and ai3 introns, do not revert at all by intron deletion. The revertants still contain the intron bi2 as shown by restriction and hybridisation analyses. They can revert by other means such as back mutations as well as by intragenic or extragenic suppression.

3.3 The ai1 intron alone is sufficient for the reversion by intron deletion

The WG1 strain, which has ai1 but lacks both ai2 and ai3 introns, gave an interesting result: 16% of reversion events are due to genomic intron deletions (Table 1). This means that the presence of ai1 intron restores the deletion of introns during reversion. Interestingly, this deletion frequency is much lower than in the CW51 isogenic control strain that has both ai1 and ai2 introns (χ2 homogeneity test is highly significant: 39 for 1 degree of freedom).

4 Discussion

In this work, we have shown the involvement of the two first intron products of the COX1 gene in the reversion by intron deletion. Indeed, when ai1 and ai2 introns were absent, no bi2 deletion was obtained among a total of 67 analysed revertants of the tester mutation, M3041, situated in the bi2 intron. More important is the finding that the presence of only ai1, in a M3041 strain deleted for the intron ai2 and ai3, was sufficient to permit reversion by intron deletion.

In addition, the frequency of reversion by intron deletion in this strain was very low compared to that in the control strain. This could be explained either by a gene dosage effect (two similar ai1 and ai2 products in control, compared to only the ai1 product in WG strain) or by the fact that ai2 protein is more active. It has been shown that, in S. cerevisiae, only the ai2 protein is active and, in other species, the ai1 protein has a very weak activity [11]. Furthermore, the frequency of genomic intron deletion of ai5α in the (ai2+, ai1 deleted) pet strain has been shown not to be lower than in the (ai1+, ai2+) pet strain [31,32].

Concerning the ai3 intron, we have shown that its absence in the WG1 strain (which contains the intron ai1) did not hamper the intron deletion process. This indicates that the endonuclease encoded by the intron ai3 is not involved in the genomic intron deletion process and similar results have been obtained by [32]: the pet strains retaining ai1 and ai2 but devoid of ai3 intron revert at a normal frequency by ai5α intron deletion.

In this work, we have demonstrated the involvement in trans of the reverse transcriptase, encoded by two group-II introns of the COX1 gene, ai1 and ai2, in the genomic deletion of introns located in another mosaic gene, of the cytochrome b. This is the most important difference with the work of [32] demonstrating the involvement in cis of ai1 and/or ai2 reverse transcriptase in genomic deletion of the ai5α intron of the same COX1 gene.

What is the mechanism of intron deletion and how does the ai1 (and ai2) protein act in this process? In our previous paper [1], based on genetic evidence, we have proposed three possible mechanisms for genomic intron deletion. The first was reverse transcription of a partially spliced pre-mRNA and subsequent integration of the cDNA in the genome. The second was recombination between a partially processed pre-mRNA intermediate and the anti-complementary DNA strand of the resident gene. The third was the transposition-like capacity of introns (an “auto-deletion”-like process).

We shall retain the first hypothesis claiming an involvement of a reverse transcribed pre-mRNA intermediate in a yet undefined recombination process with the DNA as in some pseudogenes or retroposons. We can refer to the genomic intron deletion as a “retrodeletion” process. The difference with a pseudogene arising by reverse transcription is the absence of polyA and the formation of short direct repeats boarding the pseudogene. The leakiness of the majority of bi2 mutations argues for the existence of some RNA intermediates, having spliced-out the mutated bi2 intron [1,3]. Indeed, the leakiness of the majority of bi2 mutants has been quantified by measuring the incorporation of a tritiated valine aminoacid in mutant yeast cells growing on synthetic glycerol medium, selective for respiration [3]. The results showed that the majority of these mutants are indeed leaky, meaning that they must have a certain amount of mature cyt.b mRNA, otherwise they cannot respire glycerol [1,3].

Our results rule out the “auto-deletion” since the process needs a factor provided in trans by the ai1 or ai2 intron. Taking in account our present results, the suggested existence of reverse transcriptase in yeast mitochondria [1,32], the demonstrated similarity of the ai1 and ai2 proteins to the reverse transcriptase family [18], and finally the biochemical demonstration of reverse transcriptase activity [11], we are convinced that such activity is involved and essential in the “retro-deletion” of mitochondrial introns.

This deletion leads to the “retro-homing” of the processed gene, see the model proposed in Fig. 5. Notice that we do not yet know how these proteins might initiate this process. For instance, is a tRNA required as a primer as in retrovirus [33] or is it the end of pre-mRNA itself which serves as initiator as in the case of the mitochondrial plasmids of Neurospora[9], or is it an internal sequence in the pre-mRNA, as in the case of the demonstrated ai2 reverse transcription both in vitro and in vivo [11]. Nevertheless, we suggest that the accumulation of unspliced RNA in the mutant cells could serve as a “signal” for the initiation of the reverse transcription and the deletion process.

Figure 5

A hypothetic model of genomic intron deletion. First step: some pre-mRNA intermediates correctly processed of the mutated intron are postulated to exist among the large amount of intermediates blocked in the RNA splicing step, due to the intronic mutation. This implies that the mitochondrial RNA splicing is somehow leaky in some intronic mutants that revert by intron deletion (1, 3). Second step: some of the matured intermediates are reversely transcribed by the ai1 and/or ai2 protein into cDNAs. Third step: integration of this cDNA into the mitochondrial genome by recombination with the “long” gene version: the cDNA hybridises (for instance, during the progression of a replication fork) to the anti-complementary DNA strand, looping out the introns that do not have their counterpart in the cDNA. The displaced DNA strands are then degraded. Finally, the cDNA is ligated to its collinear strand.

Concerning the model, one can also imagine that step 3 comes before step 2, i.e. reverse transcription follows the recombination between the spliced RNA precursor and the DNA. In this context, integration of linear intron RNA directly into DNA has already been shown [14]. It is worth to note that the ai2 protein has been shown to be able to use non-cognate template in a reverse transcription process [16].

Mitochondrial reverse transcriptase activity in yeast mitochondria has been attributed mostly to ai2 intron [11,13,14]. Even though recent results have shown that the ai1 protein is involved in the retro-transposition of its own intron by reverse splicing directly into an ectopic DNA site [12], its implication in intron deletion has not been clearly established. Our results demonstrate the involvement of the ai1 intron in the retro-deletion process, suggesting firmly that the ai1 intron is really endowed with reverse transcriptase activity and is surely implicated in the precise intron deletion process.

It is known that viral reverse transcriptases share other activities: endonuclease, RNAseH, protease and unwindase [34]. We have to note, however, that one of these proteins, the ai1, was shown to be an mRNA maturase [35]. This activity resides in a p68 protein, derived from an initial p90 product supposed to be proteolysed into the p68 and a p20 peptide [35,36]. The p90 is translated in frame from the beginning of the first exon to the end of the ORF of the first intron [35]. The reverse transcriptase homology was found in fact in the p68 peptide [18]. One can predict that other activities shared by the ORF(s) of group-I introns could be involved in the integration mechanism, as some of them are shown to be endonucleases [37,38]. In this context, it was noted that both ai1 and ai2 ORF share a zinc finger motif at the C-terminus, with similarity to the N-terminal region of retroviral integrase domains [19]. Both ai1 and ai2 were shown to possess endonuclease activities involved in the mobility of these introns [14]. The origin and the presence of a group-II intron, containing an ORF with similarity to reverse transcriptase, in mitochondria [7,18], chloroplast [39] and bacteria [40], is very intriguing. They should date from before the supposed endosymbiotic origin of the eukaryotic cell, even if one cannot exclude that they have been introduced in organelles later from a nuclear retroposable element like Ty. They probably have been maintained by an insertion spreading mechanism that can move by homing [41] or transposition [4244]. Both phenomena are supposed and shown to pass through reverse splicing and/or reverse transcription [1215]. Finally, one can ask if such a phenomenon of intron retro-deletion is restricted to mitochondria, or could also be found in the nucleus, giving rise to functional intron-less genes and not solely to pseudogenes. We recently discovered a like phenomenon in the human nucleus (Baccouche et al., in preparation).


We are very indebted to E. Petrochilo and Ch.J. Herbert for helpful discussions and critical reading of the manuscript. We thank B. Poirier for her help in writing the text. This work was supported by grants from INSERM, CNRS, ATP Biologie Moléculaire du Gène and la Ligue Nationale Française contre le Cancer and la Fondation pour la Recherche Médicale.


  1. [1].
  2. [2].
  3. [3].
  4. [4].
  5. [5].
  6. [6].
  7. [7].
  8. [8].
  9. [9].
  10. [10].
  11. [11].
  12. [12].
  13. [13].
  14. [14].
  15. [15].
  16. [16].
  17. [17].
  18. [18].
  19. [19].
  20. [20].
  21. [21].
  22. [22].
  23. [23].
  24. [24].
  25. [25].
  26. [26].
  27. [27].
  28. [28].
  29. [29].
  30. [30].
  31. [31].
  32. [32].
  33. [33].
  34. [34].
  35. [35].
  36. [36].
  37. [37].
  38. [38].
  39. [39].
  40. [40].
  41. [41].
  42. [42].
  43. [43].
  44. [44].
View Abstract