Mitochondrial Group II Introns, Cytochrome c Oxidase, and Senescence in Podospora anserina

Abstract

Podospora anserina is a filamentous fungus with a limited life span. It expresses a degenerative syndrome called senescence, which is always associated with the accumulation of circular molecules (senDNAs) containing specific regions of the mitochondrial chromosome. A mobile group II intron (α) has been thought to play a prominent role in this syndrome. Intron α is the first intron of the cytochrome c oxidase subunit I gene (COX1). Mitochondrial mutants that escape the senescence process are missing this intron, as well as the first exon of the COX1 gene. We describe here the first mutant of P. anserina that has the α sequence precisely deleted and whose cytochrome c oxidase activity is identical to that of wild-type cells. The integration site of the intron is slightly modified, and this change prevents efficient homing of intron α. We show here that this mutant displays a senescence syndrome similar to that of the wild type and that its life span is increased about twofold. The introduction of a related group II intron into the mitochondrial genome of the mutant does not restore the wild-type life span. These data clearly demonstrate that intron α is not the specific senescence factor but rather an accelerator or amplifier of the senescence process. They emphasize the role that intron α plays in the instability of the mitochondrial chromosome and the link between this instability and longevity. Our results strongly support the idea that in Podospora, “immortality” can be acquired not by the absence of intron α but rather by the lack of active cytochrome c oxidase.

The filamentous fungus Podospora anserina has limited vegetative growth (36). After several divisions, the apical cells stop growing and die. Earlier studies showed that the transition from the juvenile to the senescent state is caused by the recurrent appearance of a specific cytoplasmic factor (25, 48). Subsequently, molecular analysis of the mitochondrial DNA (mtDNA) revealed that the senescent state is always correlated with the accumulation of circular multimeric DNA molecules called senDNAs. Several groups of senDNAs (α, β, and γ, etc.), which originate from separate regions of the mitochondrial chromosome, can be recovered from independently growing senescent cultures. Strikingly, one senDNA (senDNA α) is present in all senescent cultures of wild-type strains; it corresponds exactly to the first intron (intron α) of the COX1 gene, which encodes subunit I of cytochrome c oxidase (for a review, see reference 10).

Intron α is a mobile group II intron that is able to transpose into the mitochondrial chromosome (45). Recent studies with yeast have shown that mobility of group II introns involves intron-encoded reverse transcriptase and DNA endonuclease activities that are needed for the site-specific insertion of the introns into DNA (12, 17, 53–56). Because of its intronic properties and because of the systematic accumulation of senDNA α during senescence of wild-type strains, intron α has been thought to have a prominent role in the natural senescence process of P. anserina. However, this idea has been questioned recently by the analysis of some long-lived nuclear mutants that undergo the senescence process and accumulate mitochondrial rearrangements other than senDNA α (6, 47).

Nevertheless, the role of the senDNAs and, more specially, the role of intron α in the natural senescence process in wild-type strains remains unsolved. The only way to clarify this question is to compare the properties of two strains that differ only by the presence or absence of this intron. Until now the only available mutants lacking this sequence are the mex mutants, which carry a deletion of their mitochondrial chromosome that covers part of the intron and its upstream exon. As a consequence, mutants that escape the senescence process are deficient not only for the intron and its encoded functions but also for cytochrome c oxidase activity (2, 39, 50); i.e., they use a cyanide-insensitive respiration pathway. Such alternative pathways exist in many eukaryotic organisms (27) and in particular in the related fungus Neurospora crassa (11, 24, 29). Thus, it was not clear whether the “immortality” of the mex mutants was due to a lack of intron α or to their respiratory defect.

Here we report the selection and analysis of a new mitochondrial mutant (the mid26 mutant) that has the intronic sequence precisely deleted. Moreover, the sequence of the integration site of the intron is slightly modified to prevent the reinvasion of the site by the intron, as would otherwise occur with a mobile element. This modification does not disrupt the reading frame of the COX1 gene, and the encoded protein differs by only two amino acids from the wild-type protein. The mutation has no effect on the cytochrome c oxidase activity, and the growth rate and fertility of the mid26 strain are identical to those of the wild-type strain. We show that this strain displays a senescence syndrome whose characteristics are similar to those of the wild-type strain, with systematic amplification of β and γ senDNAs. However, its life span is about twofold longer. The introduction of another group II intron, COX1i4, into the mitochondrial genome does not restore a wild-type life span.

This comparative study of two strains which differ by the presence or absence of intron α allows us for the first time to define precisely and directly the role that this intron plays in the senescence process of P. anserina. The results show clearly that deletion of intron α in the mitochondrial chromosome does not prevent senescence but rather delays its appearance. These data emphasize the role of this mobile intron in the instability of the mitochondrial genome and the link, in an obligate aerobe, between this instability and the longevity. On the other hand, these data indicate that the immortality of the mex mutants is not due to the lack of the intron but probably is due to the lack of cytochrome c oxidase and its metabolic consequences.

MATERIALS AND METHODS

Strains and growth conditions.

The genetic and biological properties of P. anserina were first described by Rizet and Engelmann (38) and have been reviewed by Esser (13). All of the strains used in this study are derived from the s (35) or (A)s strain. The (A)s strain has the nuclear genome of strain s and the mitochondria from race A. The AS1-4 mutation was identified as an antisuppressor mutation (32). AS1-4 mat− strains display the premature death syndrome, and mycelial death occurs a few centimeters after ascospore germination. This syndrome is linked to the accumulation of mtDNA site-specific deletions covering about one-third of the genome (3, 7). Most often, the deletion chromosomes carry a specific deletion, Δ1, in which one endpoint corresponds exactly to the 5′ end of the mobile intron α and the other endpoint corresponds to a transposition site of the intron. The deletion is assumed to arise from an intramolecular recombination event between the two repeats of the intron (45). Other types of deletions can accumulate within the mitochondrial chromosome of dying AS1-4 mycelia; they also involve intron α at one end (42).

All media (corn meal extract [MR], minimal synthetic [M2], and germination [G] media) were as described by Esser (13). Longevity analysis was performed on M2 medium at 27°C as previously described (1).

Life spans were measured for four or five subcultures from two to five strains exhibiting a given genotype. Determination curves for a given culture were obtained as follows. As soon as the spore had germinated, 1 implant was grown in a culture tube and 15 implants were regularly taken at 2 cm downstream of the growth edge. These implants, taken at various distances from the initial implant, were analyzed in culture tubes for their longevity.

Contamination experiments, which allow the transfer of mobile mitochondrial sequences between a donor and a recipient strain, were performed as previously described (41). A small inoculum of the donor growing mycelium was put onto the recipient mycelium that had grown about 3 cm on solid medium. Generally, at about 6 to 8 cm of growth after the donor was put on the recipient, all of the mobile optional sequences, but no other markers from the donor, were introduced into the mtDNA of the recipient.

mtDNA analysis, PCR conditions, and sequencing of the PCR products.

The characteristics and the complete sequence of the mitochondrial chromosome of P. anserina, race A, were described by Cummings et al. (9). mtDNA was extracted by standard methods either by purification on a CsCl gradient (8) or by a minipreparation method (23). The specific probes used to reveal the α intron, the COX1i4 intron, and the β and γ senDNAs are, respectively, Pα (intron α cloned in pBR322), a cloned fragment of 1,300 bp that covers the circular junction of the COX1i4 intron (41), a cloned monomer of one β senDNA (20), and the cloned EcoRI fragment 1 that covers a part of the γ region (see reference 52, in which the γ region is named β). PCRs were done as described by Sambrook et al. (43). The mid26 fragment with intron α deleted was obtained after the cloning of the PCR product amplified with oligonucleotides 1456 (GGATTACTAGGTACAGCG) and 2432 (GGATTATTTTTAATACATCTTCACTA) flanking the α sequence. Sequencing was performed on four independent clones derived from four independent PCR experiments. Amplification of molecules derived from the Δ1 deletion chromosome and lacking the α sequence was performed with oligonucleotide 2432, located inside COX1i2, and oligonucleotide 8456 (AGTCAGTTCACTGTAGCAGG), located about 300 nucleotides upstream from the Δ1 deletion endpoint. For amplification of the junction sequences of the β senDNAs, pairs of divergent oligonucleotides, which were previously shown to hybridize with the senDNA, were used. Sequencing was performed after cloning of the PCR products.

Respiration.

The percentages of cyanide (CN)-insensitive and salicyl hydroxamic acid (SHAM)-insensitive respiration were measured on mycelia harvested from liquid cultures that were grown for 2 days. Measurements were done polarographically with a Gilson oxigraph in 0.6 M sorbitol–7.5 mM MgCl₂–10 mM KH₂PO₄–10 mM imidazole (pH 7.4)–0.2% bovine serum albumin. KCN and SHAM were added to final concentrations of 1 and 2.5 mM, respectively.

Cytochrome c oxidase activity was determined on mitochondria isolated as described by Sellem et al. (46) by monitoring cytochrome c oxidation at 550 nm (18, 31).

RESULTS

Selection of the mid26 mutant devoid of the α sequence.

The mid26 (for mitochondrial intron deletion) mutant was isolated from a strain displaying the premature death syndrome (3, 7) in an attempt to obtain spontaneous mutations able to overcome the syndrome. One sector (sector 26) that was able to escape growth arrest was recovered from a culture grown on corn meal medium. After 10 cm of growth, it was crossed with a wild-type strain, and a 1:1 segregation was observed for the germination phenotype corresponding to the AS1⁺/AS1-4 alleles. However, AS1-4 mat− progeny showed the premature death syndrome only when the growing sector 26 was used as a male, suggesting that the suppressor mutation of the arrest of growth of the AS1-4 mat− strain was mitochondrial. The mid26 strain was isolated after three generations of backcrosses of the strain having the mutant cytoplasm and a wild-type nuclear genome (as the female parent) with a wild-type strain (as the male parent).

As shown in Fig. 1, restriction analysis and hybridization of the mtDNA of the mid26 strain revealed that this mtDNA carried a deletion of the α sequence. This was demonstrated by the absence of the restriction fragments containing the intron (HaeIII 1,900- and 840-bp fragments, BamHI 3,280-bp fragment, and HindIII 1,330-, 1,250-, and 5,400-bp fragments) and the absence of hybridization with an α probe. The sequence of the modified region was amplified with primers flanking the α sequence, and the amplified product was cloned and sequenced. The results are shown in Fig. 2, and they require several comments. First, in the mid26 strain, the α sequence was completely deleted, in contrast to the case for the mex mutants previously selected (2, 39, 50). Second, there were five nucleotide substitutions, at E-4, E-8, E-11, E-12, and E-13 (indicating the nucleotide position in the spliced transcript, relative to the site of integration of the intron), in the COX1e1 open reading frame sequence. Third, the modified sequence exactly corresponded to the 13 to 16 nucleotides containing a potential intron binding site (IBS′1) and behind which intron α is able to transpose. They constitute one endpoint of the junction of the deleted Δ1 chromosome that accumulates in an AS1-4 strain (Fig. 2) (3, 45). Finally, the rearrangement did not interrupt the reading frame of gene COX1, and the encoded mutant protein was expected to differ from the wild type by only two amino acids, Q to I (glutamine to isoleucine) and A to S (alanine to serine).

FIG. 1.

Restriction and hybridization analysis of the mtDNA of the mid26 strain. (A) mtDNAs of the mid26 (lanes 2, 4, and 6) and wild-type (lanes 1, 3, and 5) strains were digested with restriction enzymes HaeIII (lanes 1 and 2), BamHI (lanes 3 and 4), and HindIII (lanes 5 and 6). The fragments lacking in the mid26 mtDNA are indicated by circles. (B) A corresponding gel was hybridized with a ³²P-labeled probe, Pα (sequence α inserted into pBR322). The expected restriction fragments (HaeIII 1,900 and 840 bp, BamHI 3,280 and 40,000 bp, and HindIII 1,330, 1,250, and 5,400 bp) are revealed only in the wild-type lanes. (C) Restriction sites of the COX1 region. ●, HaeIII; ◊, BamHI; ★, HindIII. CYTb, cytochrome b gene.

FIG. 2.

Nucleotide sequence of the COX1e1-COX1e2 junction of the mid26 strain. The upper line shows the wild-type (WT) sequence of the COX1e1 and COX1e2 exonic regions flanking the α intronic sequence, the middle line shows the sequence of the COX1e1-COX1e2 junction of the α deletion chromosome of the mid26 strain, and the lower line shows the sequence of the junction of the Δ1 deletion chromosome. The wild-type exonic sequence is shaded. The IBS1, IBS′1, and IBS2 motifs are indicated. The deduced amino acid sequence is shown below the nucleotide sequence. Variable bases E-4, E-8, E-11, E-12, and E-13 (indicating nucleotide positions in the spliced transcript, relative to the integration site of the intron) and variable amino acids are in boldface.

As just pointed out, COX1e2 is joined in the mid26 mutant to a short sequence identical to the Δ1 deletion endpoint (Fig. 2). The mid26 chromosome could originate from a recombination event between the wild-type chromosome and reverse transcripts of α-spliced RNAs derived from the Δ1 deletion chromosome, as diagrammed in Fig. 3. To test for the occurrence of such reverse transcripts, PCR experiments were done with DNA extracted from a AS1-4 mat− culture, using primers surrounding the α sequence in the Δ1 chromosome. One was located about 300 nucleotides upstream of the Δ1 junction, and the other one was located about 400 nucleotides downstream of the 5′ end of COX1e2. A product of about 700 bp was obtained. Its sequence, shown in Fig. 4, indicated that it corresponded to molecules joining COX1e2 to the Δ1 endpoint with a precise deletion of the α sequence.

FIG. 3.

Model of formation of the mid26 chromosome. The dark and hatched rectangles symbolize, respectively, the IBS1 sequence of intron α (located at the 3′ end of COX1e1) and the IBS′1 sequence (located in an intergenic region about 37,000 bp upstream of the COX1 gene [3]). The crosses symbolize recombination events. WT, wild type.

FIG. 4.

DNA sequence of the amplified product obtained with primers surrounding the α sequence in the Δ1 chromosome. The amplification primers were located inside COX1i2 and inside the intergenic region upstream of the Δ1 endpoint, respectively. The junction between the 5′ end of COX1e2 and the IBS′-Δ1 deletion endpoint is indicated by an arrow. The positions of the nucleotides shown at the top and the bottom of the sequence are indicated by arrowheads.

Physiological and respiratory properties of the mid26 strain.

We compared the growth rates and fertilities of the mid26 and wild-type strains on corn meal and synthetic media. In contrast to the case for the mex mutants, which are female sterile and whose growth rates are reduced, the growth rates and fertilities of the mid26 and wild-type strains were identical (data not shown). Oxigraphic measurements indicated that the respiration of exponential-phase cultures of the mid26 strain was almost completely sensitive to inhibition by KCN, as was the case for the wild type, whereas the respiration of the mex mutants was completely KCN insensitive but was sensitive to hydroxamic acids (SHAM) (Table 1). The levels of cytochrome c oxidase activity of the mid26 and wild-type strains were measured on purified mitochondria. Table 1 shows that they were identical in the two strains.

TABLE 1.

Respiration properties of mid26, mex16, and wild-type strains

Strain	% of respiration (mean ± SD) following addition of:		Cytochrome c oxidase activitya (mean ± SD)
	KCN	SHAM
Wild type	8 ± 3.5	92 ± 3.5	3,230 ± 120
mex16	100	0	NA b
mid26	8.2 ± 2	91.8 ± 2	3,149 ± 91

^aMicromoles of cytochrome c oxidized per minute per milligram of mitochondrial protein. 

^bNA, not applicable. 

The amino acid sequences of subunits I of cytochrome c oxidases of numerous organisms are related. By comparison with the Paracoccus and the bovine cytochrome c oxidases, whose crystal structures have been established (19, 49), it appeared that the two modified residues, I at position 53 and S at position 56, were located in a short interhelix region between the transmembrane helices I and II. These two changes had no effect on the activity of cytochrome c oxidase. Overall, no difference in the growth and respiratory characteristics of the mid26 and wild-type strains was detectable.

Senescence properties of the mid26 mutant.

The longevities of mid26 cultures (mat− and mat+) obtained from different crosses between the mid26 strain used as female and the wild type used as male are shown and compared with those of reference wild-type s cultures in Table 2. These experiments demonstrated two striking points concerning the mid26 mutant. First, the mid26 cultures systematically underwent a senescence process. Second, they had an increased longevity that extended the life span by a factor of about 1.5 for mid26 mat− and about 2 for mid26 mat+. The control of the timing of senescence by the mat haplotype of strain s and the longer life spans of mat+ versus mat− strains were established long ago (37). These data, reported as the frequency of living subcultures with respect to growth length, are presented in Fig. 5. They show that the survival curves were quite similar for the mutant and the wild-type strains with the exception of the plateau, which is longer in the mutant, suggesting an increased incubation delay.

TABLE 2.

Life spans of wild-type and mid26 strains

Expta	Mating type	Mean life span (cm) ± SD for strain:
		s	mid26	s(A)	mid26(A)
I	mat−	9.5 ± 1.05 (5 × 4)b	15 ± 2.95 (6 × 4)	7.7 ± 1.7 (2 × 5)	11.5 ± 0.95 (3 × 4)
	mat+	10.4 ± 1.45 (3 × 4)	19 ± 1.65 (2 × 4)	9.2 ± 1.4 (2 × 5)	19.7 ± 3.7 (2 × 4)
II	mat−	8.6 ± 0.95 (5 × 4)	13 ± 1.45 (5 × 4)		11.6 ± 1.2 (3 × 5)
	mat+	11.3 ± 1 (5 × 4)	22.6 ± 4.3 (3 × 4)		16.9 ± 3.1 (2 × 5)
III	mat−		12.9 ± 1.5 (4 × 4)	7 ± 0.9 (5 × 5)	11.6 ± 1.3 (6 × 5)
	mat+		18.3 ± 2.8 (4 × 4)	8.5 ± 1.1 (7 × 5)	15.9 ± 1.9 (6 × 5)

^aThree experiments were performed. 

^bNumbers in parentheses indicate the number of subcultures for each genotype (number of strains × number of subcultures for each strain). 

FIG. 5.

Life span curves of the mid26 and wild-type strains curves for s mat− (▵), s mat+ (▴), mid26 mat− (○), and mid26 mat+ (●) strains are shown. The data are reported as the frequency of living subcultures with respect to growth length. Data are taken from Table 2 (experiment I).

Early studies had shown that senescence in wild-type strains of Podospora is promoted by the appearance and spreading of a cytoplasmic factor called the determinant (25, 48). In order to test whether the senescence process occurring in the mid26 mutant was similar to the one occurring in wild-type strains, determination curves were performed as described by Marcou (25). For a given culture, groups of 15 implants, taken at various distances from the germination point, were analyzed for their longevity. Most of the determination curves obtained for the wild-type cultures showed a sharp slope and a plateau that decreased as a function of the distance at which the implants were taken (data not shown). In contrast, as shown in Fig. 6, the determination curves obtained for the mid26 strain indicated that implants taken at 0.3 and 3 cm from the germination point had the same plateau and had curves that were similar to the life span curve of the mutant. For the implants taken after 3 cm from the germination point, the curves showed a plateau that decreased as a function of the distance from the senescent edge. These data were in agreement with Marcou’s model and indicated that the senescence process in the mid26 strain had the same characteristics as that in wild-type strains. mid26 mycelia behaved as if a single random event was responsible for the transformation from a nondetermined to a determined state. Under our culture conditions, wild-type cultures were determined as soon as or very early after germination, whereas most of the mid26 cultures were determined after 3 cm of growth. The slopes for the mutant and wild type were similar, indicating that the frequency of occurrence of the determinant was unaltered in the mutant and that, as in the wild type, there was no escape once the primary event had occurred.

FIG. 6.

Determination curves for a mid26 culture. Immediately after germination of a mid26 mat− spore (mid26 11−), one implant was grown in a culture tube. Fifteen implants were taken at various distances from the initial implant and analyzed for their longevity. Results for implants taken at 0.3 (■), 3 (□), 6 (▵), 9 (○), and 11.5 (×) cm are shown. For this mid26 11− culture, arrest of growth occurred at 18 cm. Longevity curves for implants taken at 0.3 and 3 cm are similar to each other and to the longevity curve for the mid26 mat− strain shown in Fig. 5. In contrast, longevity curves for implants taken after 3 cm from the initial implant show a linear decrease of longevity, as they were taken closer to the senescent edge.

In P. anserina, all cases of senescence reported until now are correlated with the accumulation of senDNAs. Accumulation of senDNA α is systematic in wild-type senescent cultures, and it is frequently accompanied by additional senDNAs (β and γ, etc.). The mtDNA contents of 15 independent mid26 senescent cultures were therefore examined. All cultures tested displayed gross mtDNA rearrangements involving the amplification of a variety of β and γ senDNAs, as shown on the gel and the autoradiograph in Fig. 7. The sequences of the junction sites of four different β senDNAs obtained for the mid26 strain were established. The breakpoints of one of them were bound by a short direct repeat of 6 bp, but those of the other three senDNAs did not occur within repeated sequences. For one of them, the 3′ termini occurred a few nucleotides upstream of the 5′ end of the tRNA₂^Arg gene. All of these properties agreed with those of the previously characterized β senDNAs (21) and showed that the β senDNAs obtained for the mid26 strain were similar to those obtained for the wild type when they were amplified together with senDNA α.

FIG. 7.

mtDNA content of senescent cultures from the mid26 strain. (A) HaeIII restriction patterns of the mtDNAs of nine senescent cultures of the mid26 strain (lanes 1 to 9) and of a young culture of the wild-type strain (lane 10). (B) Corresponding hybridization with a specific probe for region β. In the young wild-type culture (lane 10), this probe reveals the three encompassed HaeIII chromosomal fragments 1 (8,600 bp), 10 (3,400 bp), and 16 (2,100 bp) (∗). In the senescent mid26 cultures (lanes 1, 3, 6, 7, 8, and 9), it reveals the intact fragment 16 present in large amounts plus additional fragments corresponding to the junction fragments of β senDNAs; these are constituted by parts of fragments 1 and 10. In senescent cultures 4 and 5, the junction fragments involve the chromosomal fragment 16; in senescent culture 2, the gross amplification seen in ethidium bromide staining (panel A) does not correspond to a β senDNA. (C) Corresponding hybridization with a specific probe for region γ. In the young wild-type culture (lane 10), this probe reveals the four encompassed HaeIII fragments 2 (6,700 bp), 4 (4,900 bp), 17 (2,000 bp), and 30 (930 bp) (∗). In the mid26 senescent cultures 5, 6, 7, 8, and 9, the probe identifies an additional junction fragment whose size is greater than that of chromosomal fragment 2. In senescent culture 2, it reveals only a γ monomer whose size is about 5,000 bp; therefore, this culture contains a nonidentified gross amplification in addition to this senDNA (see panel A).

Intron α mobility was ineffective in the mid26 mutant, and the absence of intron α was not compensated for by another group II intron.

The mitochondrial genome of P. anserina contains three group II introns: α (or COX1i1), COX1i4, and ND5i4 (9). COX1i4 is an optional intron not present in race s, from which the mid26 strain was obtained. It is similar to intron α, and its open reading frame also encodes a protein with putative endonuclease and reverse transcriptase activities. It has been postulated that its presence in race A is involved in the decreased longevity of this race (26). We have demonstrated that, as with intron α, intron COX1i4 is able to form circular DNA molecules that can be amplified in senescent cultures, although to a lesser degree. In contrast, intron ND5i4 has never been observed as circular DNA molecules (41). It differs from the two others in that the sequence for its putative reverse transcriptase is split by an insertion (9, 22), and it is possible that this characteristic is responsible for the absence of ND5i4 DNA circles.

Contamination experiments that allowed the transfer of introns from donor strain to recipient strain by homing (see Materials and Methods) were performed between the mid26 strain and an s(A) strain, which contains, in addition to intron α, three optional sequences: COX1i4, CYTbi3, and insert C (4). Fifteen mid26 mycelia were independently contaminated, and their mtDNAs were analyzed by restriction and Southern experiments 6 cm after contact with the donor inoculum. Probes that hybridized inside and upstream of the COX1i4, CYTbi3, insert C, and α sequences revealed that the 15 mycelia contained only one type of mitochondrial chromosome, where the three optional sequences of the donor, but not the α sequence, invaded the recipient (data not shown). These results indicated that although the homing of group I and II introns was quite effective in the mid26 strain, no homing of intron α was detectable in this mutant.

The mid26 mutant that contained the three optional sequences was called mid26(A). Two successive crosses between the mid26(A) mutant used as a female and wild-type used as a male were performed, and the longevities of the mat− and mat+ progenies were recorded and compared to those of reference wild-type s(A) and mat+ strains (Table 2). Comparison of the longevities of the s and s(A) strains on one hand and of the mid26 and mid26(A) strains on the other (Table 2) clearly showed that the presence of one or more of the three optional sequences in the mitochondrial chromosome of isonuclear strains accelerated the senescence process. However, comparison of the longevities of the s (α⁺ COX1i4⁻) and mid26(A) (α⁻ COX1i4⁺) strains indicated that the presence of intron COX1i4 did not restore the characteristic longevity of a strain carrying intron α. Eight independent senescent cultures of the mid26(A) strain were examined. Six of them accumulated large quantities of circular (or tandemly arranged) copies of intron COX1i4 in addition to β and γ senDNAs. One accumulated only a β senDNA, and another accumulated only circular copies of COX1i4 (data not shown).

DISCUSSION

We describe here a new mitochondrial mutant that differs from the wild-type only by the absence of intron α and by the benign change of two amino acids in subunit I of cytochrome c oxidase. We show that this mutant, whose respiratory properties are identical to those of the wild type, displays a senescence syndrome correlated with the accumulation of large amounts of β and γ senDNAs and that its longevity is increased by a factor of 1.5 to 2. This study demonstrates that the α sequence is not necessary for the senescence syndrome but that its presence accelerates the process.

The mid26 strain behaves as a pseudo-wild-type strain devoid of the α intron.

The mitochondrially modified chromosome was isolated from an AS1-4 mat− strain as a suppressor of the premature death syndrome. The suppressor effect of the mid26 mutation will be described elsewhere. The mid26 chromosome is issued from an heteroplasmic cytoplasm containing a low concentration of wild-type chromosomes and abundant Δ1 deletion chromosomes in which intron α is joined to an IBS-like motif (IBS′1) located 37,000 bp upstream of the 5′ end of the intron in the wild-type chromosome (3). The occurrence of DNA molecules derived from Δ1 and with the α sequence precisely deleted strongly suggests that these molecules probably result from a reverse transcription mechanism (40). This means that Δ1 transcripts do exist and that intron α is able to be spliced from them. One model that could explain the formation of the mid26 allele relies on recombinational events between the wild-type chromosome and these Δ1 α-spliced reverse transcripts (Fig. 3).

The data presented here strongly support the idea that the few changes in the upstream sequence of the integration site eliminate or drastically reduce α homing in the mid26 strain, whereas they do not block splicing of the intron in the Δ1 chromosome. The homing process of group II introns occurs by a targeted, DNA-primed reverse transcription mechanism in which the intron RNA reverse splices into the recipient DNA (12, 17, 53–56). Our data agree with those of Eskes et al. (12), which showed that base pairing between the EBS1 (RNA) of the intron and the IBS1 of the DNA target site is essential for homing and that a single base change can block intron mobility, whereas it has little effect on splicing. Moreover, in vitro experiments indicate that the target site for reverse splicing and sense cleavage extends from E−21 to E+1 for the yeast ai2 intron (17). It is therefore not surprising that α homing is abolished in the mid26 mutant, since the EBS1-IBS1 pairing is disturbed by one mismatch (T/C at E−4 [Fig. 2]) and it also carries substitutions at E−8, E−11, E−12, and E−13. These characteristics, in addition to the suppressor effect of the mid26 mutation on the premature death syndrome, probably accounts for the selection and stability of the chromosome devoid of intron α.

The upstream sequence substitutions flanking the intron insertion site change only two amino acids, and they do not impair the reading frame of the gene COX1. This modification has no effect on the activity of the cytochrome c oxidase. The mid26 strain can be therefore considered a pseudo-wild-type strain that differs from the wild-type s strain of P. anserina only by the absence of the α intron.

Role of the α intron in the senescence process of P. anserina.

Intron α has two characteristic properties. First, it is a mobile group II intron, which is a source of instability of the mitochondrial chromosome (45). Second, it corresponds to a senDNA. The role of this intron in the senescence process of P. anserina has been a much-debated question. On one hand, it has been hypothesized that the senDNAs (20), and especially senDNA α (5, 14, 30), correspond to the senescence factor that Marcou (25) called the determinant. On the other hand, it has been speculated that senDNA α and the other senDNAs have only a secondary role in the senescence process (47).

We show here that a strain lacking only intron α displays a senescence process caused by a single random event (the determination event) quite similar to that occurring in a wild-type strain. This directly demonstrates that intron α is not the primary determinant. Nevertheless, this strain does show an increased longevity. It has been shown that modifications in the mitochondrial functions modulate longevity (1). Since no change in the cytochrome c oxidase activity of the mid26 strain is detectable, its increased longevity is probably not due to a change in its mitochondrial metabolism but rather is due to the absence of the intron α. We show furthermore that the senescence process in the mid26 strain is also correlated with mitochondrial rearrangements and the accumulation of senDNAs, as is the case for all senescence events in P. anserina described until now.

These results can be formally interpreted in two ways. First, as suggested in early studies, a threshold concentration of senDNA molecules may function as the determinant of senescence. The different senDNAs (α, β, and γ, etc.) would then play equivalent roles in the senescence process, although they would differ in their rates of genesis and/or spreading. In this hypothesis, the increased longevity of the mid26 strain would result from a longer time for senDNAs, other than α, to reach the necessary amount for senescence expression. It is known that senDNA α and the other senDNAs originate and accumulate by different mechanisms (21), and we have also shown that the amplification of senDNA α precedes that of the other senDNAs in wild-type senescent cultures (unpublished results). A second possibility is that the determinant does not correspond to senDNAs. In that case, the effect of this factor would be different according to whether it occurs in a strain with or without intron α. In other words, its mode of expression or action would depend in part on the quality of the mitochondrial chromosome. It is generally accepted that group II introns, due to their mobility and their enzymatic properties, are a source of mitochondrial instability (28, 44, 45). Thus, the effect of the determinant would be modulated by the intrinsic properties of the stability of the mitochondrial chromosome. We show here that intron α and intron COX1i4, although closely related, cannot substitute for each other, and this underscores the peculiar role of intron α in the instability of the mtDNA.

Whatever the hypothesis on the nature of the senescence determinant, our data emphasize the link between longevity and stability of the mitochondrial chromosome, and they strengthen the idea that there is a causal relationship between these two parameters. Cases of senescence in the fungal genera Aspergillus and Neurospora have been described. Their study has revealed heterogeneous sets of molecular processes. However, the great majority of them lead to the degeneration of mitochondrial function and are related to mtDNA instability (for a review, see reference 16).

Respiratory metabolism, mitochondrial instability, and senescence in P. anserina.

Our data allow us for the first time to choose between the two hypotheses currently used to explain the immortality of the mex mutants. They strongly support the idea that the immortality of these mutants is due to the lack of active cytochrome c oxidase rather than to the lack of intron α. Recent data obtained from a strain with a disruption in a nuclear gene coding for a subunit of cytochrome c oxidase confirm this conclusion (9a). Thus, the initial challenge in understanding the link between senescence and mitochondrial rearrangements is now coupled with a new challenge of understanding the link between senescence and respiratory metabolism. It is tempting to think that these two questions are connected and that the respiratory metabolism is involved in the control of the mitochondrial stability.

Strains of P. anserina lacking cytochrome c oxidase use an alternate cyanide-resistant oxidase. It has been demonstrated that life spans of cultures of P. anserina are greatly increased by the action of the mitochondrial mutation capR1 or by growth in the presence of some inhibitors of the mitochondrial functions. All of these situations correspond to a deficiency in detectable cytochrome c oxidase, and it was proposed as early as 1980 that the increased life span in P. anserina is correlated with the functioning of the alternate oxidase (1). Several reports state that in plants the alternate oxidase constitutes an efficient antioxygen defense system of mitochondria (33, 51). However the links between the functioning of the cyanide-sensitive or -insensitive respiratory pathways and longevity are far from being clear. Indeed, it can be observed that senescence in P. anserina (1), as in Neurospora (34), is paralleled by switching from cytochrome c oxidase-mediated respiration to cyanide-resistant respiration. Moreover, the long-living nuclear iviv double mutant has been reported to respire only via the cyanide-sensitive pathway and to be incapable of alternate respiration (15). Taken together, these data suggest that the respiratory metabolism plays a major role in the control of the mitochondrial DNA integrity. Experiments to elucidate the nature of this control are in progress.