Abstract
The mitochondrial rotenone-sensitive NADH:ubiquinone oxidoreductase (complex I) comprises more than 30 subunits, the majority of which are encoded by the nucleus. In Chlamydomonas reinhardtii, only five components of complex I are coded for by mitochondrial genes. Three mutants deprived of complex I activity and displaying slow growth in the dark were isolated after mutagenic treatment with acriflavine. A genetical analysis demonstrated that two mutations (dum20 and dum25) affect the mitochondrial genome whereas the third mutation (dn26) is of nuclear origin. Recombinational analyses showed that dum20 and dum25 are closely linked on the genetic map of the mitochondrial genome and could affect the nd1 gene. A sequencing analysis confirmed this conclusion: dum20 is a deletion of one T at codon 243 of nd1; dum25 corresponds to a 6-bp deletion that eliminates two amino acids located in a very conserved hydrophilic segment of the protein.
THE rotenone-sensitive NADH:ubiquinone oxidoreductase, traditionally called complex I, catalyzes electron transfer from NADH to ubiquinone in a process coupled to proton transport across the inner mitochondrial membrane. Complex I is the most complicated enzyme of the respiratory chain and remains the least understood of the respiratory complexes. It comprises more than 30 subunits, the majority of which are encoded by nuclear genes and imported from the cytoplasm (reviewed by Weisset al. 1991; Walker 1992; Rasmussonet al. 1998). Seven subunits are coded for by mitochondrial genes (nd1, -2, -3, -4, -4L, -5, and -6) in mammals and Neurospora crassa and two additional ones (nd7 and nd9) in plants. The NADH:ubiquinone oxidoreductase equivalent of complex I is much smaller in bacteria. It contains 13 or 14 proteins homologous to the 13 or 14 major subunits of the mitochondrial complex I, including the seven mitochondrially encoded ND subunits (Dupuiset al. 1998a; Friedrich 1998; Yagiet al. 1998). The reason for the complexity of complex I in eukaryotes is unknown. Electron microscopic analysis of the Neurospora enzyme has shown that complex I has an l-shaped structure with one arm buried in the mitochondrial membrane and the other protruding in the matrix (Hofhauset al. 1991; Guénebautet al. 1997). The hydrophobic membrane arm contains all the mitochondrial ND subunits and bears the quinone binding sites as well as the proton pumping machinery. The matrix arm is made up entirely of nuclear encoded subunits, possesses the NADH dehydrogenase activity, and contains most of the prosthetic groups involved in electron transfer (Weisset al. 1991).
Despite the knowledge of the primary structure of the mitochondrial DNA-encoded subunits, little is known about their functions, except for ND1, which binds rotenone and has been proposed to carry the ubiquinone-binding domain(s) (Earleyet al. 1987). Useful information about their roles comes from the analysis of complex I-deficient mutants in different organisms. In Neurospora, the E35 “stopper” mutant is deficient for the nd2 and nd3 genes and assembles a subform of complex I, which seems to lack the membrane domain (Alves and Videira 1998). In humans, point mutations in nd1, nd4, and nd6 genes are responsible for complex I deficiencies and cause optic nerve degeneracy (Shapira 1998). Activity of complex I is dramatically reduced or totally abolished in human or mouse cell lines that carry frameshift mutations in nd4, nd5, or nd6 genes (Hofhaus and Attardi 1993, 1995; Bai and Attardi 1998). In plants, mitochondrial nd mutations have been described in a nonchromosomal stripe mutant (NCS2) of Zea mays (Marienfeld and Newton 1994) and in two cytoplasmic male sterile mutants (CMSI and CMSII) of Nicotiana sylvestris (Plaet al. 1995). Both types of mutants display a reduced complex I activity. Their analysis is made complicated because mutations occur through extensive genomic rearrangements or because they generally cannot be maintained in pure homoplasmic lines, except in the case of a rare NCS2 nd4 deletion mutant plant of maize (Yamato and Newton 1999).
The unicellular green alga Chlamydomonas reinhardtii can be used as a model system to investigate mitochondrial gene function in plant cells. Its small linear 15.8-kb mitochondrial genome has been totally sequenced and all the genes identified (Vahrenholzet al. 1993). Moreover, several nonlethal mutations (dum) altering mitochondrial genes encoding apocytochrome b (cob gene) or subunit 1 of cytochrome oxidase (cox1 gene) have been characterized (Matagneet al. 1989; Dorthuet al. 1992; Randolph-Andersonet al. 1993; Colinet al. 1995; Remacle and Matagne 1998). Mutant cells are homoplasmic for the mutation and do not retain any wild-type mitochondrial DNA copies. The mutants lack the cytochrome pathway of respiration but their respiratory activity is partially retained via the nonphosphorylating alternative pathway that drives the electrons from reduced ubiquinone to oxygen. Phenotypically, the mutants have lost the capacity to grow under heterotrophic conditions (Dark– or Dk– phenotype) whereas the photoautotrophic growth is barely affected. An extreme phenotype is obtained when mutant cells possess mitochondrial DNA copies deleted for cob, nd4, and the 3′ end of nd5 (Duby and Matagne 1999). This dum24 mutant lacks the cyanide-sensitive cytochrome pathway of respiration and the activity associated with complex I. A low respiratory rate of mutant cells is maintained via the activities of rotenone-resistant NADH dehydrogenase, complex II, and alternative oxidase.
Until now, mutants of Chlamydomonas altered only in complex I have never been characterized at the molecular level. We describe here three mutants from Chlamydomonas that lack complex I activity. In contrast to the other respiratory mutants characterized so far, they grow slowly under heterotrophic conditions (Dk+/– phenotype), probably because the two phosphorylation sites associated with complex III and complex IV are preserved. One mutation is of nuclear origin whereas the two others affect the nd1 mitochondrial gene. A recombinational analysis has allowed us to draw a more complete genetic map of the mitochondrial genome and to confirm that the recombination frequency is ∼3%/kb.
MATERIALS AND METHODS
Strains and growth conditions: Strains used in this work are derived from the 137c strain of Chlamydomonas reinhardtii. The following mitochondrial mutants have been used (Table 1): dum15, double base pair substitution in the cob gene, lacking complex III activity; dum19, one T deletion in cox1, lacking complex IV activity (Colinet al. 1995); dum20, mitochondrial mutation (not characterized at molecular level), lacking complex I activity (Remacle and Matagne 1998); dum24, deletion encompassing cob, nd4, and the 3′ end of nd5, lacking activities of complex I and complex III (Duby and Matagne 1999). Arg7, arg7-3, arg7-7, and arg7-8 are nuclear mutations at the ARG7 nuclear locus. Some of these arg mutations complement (Matagne 1978), allowing diploids to be selected on Trisminimal phosphate (TMP) minimal medium.
Cells were routinely grown under light (75 PAR) on TAP (Tris-acetate phosphate) or TMP agar medium supplemented with 100 mg/liter arginine when required (Harris 1989). Rotenone, antimycin A, and myxothiazol were dissolved in ethanol and added to TAP agar medium after autoclaving at 35, 1, and 7.5 μm final concentrations, respectively.
Mutagenesis and genetical analysis: Mutagenesis with acriflavin was performed on the wild-type mt– (2) or dum19 mt– (239) strains as described in Duby and Matagne (1999). The transmission pattern of the mutations in crosses was determined by random analysis of the meiotic products.
Recombinational analysis: To determine the recombination frequencies between mitochondrial markers, strains carrying different dum mutations and complementary arg7 nuclear markers were crossed and the diploid colonies were selected on minimal TMP medium. After 10 days, when the mitotic segregation had produced near homoplasmic diploid progeny, the recombination frequencies were determined, assuming that the percentages of Dk+ recombinant cells corresponded to the recombination rates (Remacleet al. 1995).
Whole-cell respiration: Measurements of whole-cell respiration were made using a Clark Electrode (Hansatech Instruments, King's Lynn, England) as described in Duby and Matagne (1999). Rotenone dissolved in ethanol (15 mm) was used at a final concentration of 100 μm. To determine the reduction of whole-cell respiration by rotenone, the inhibitory effect of 0.7% ethanol alone (∼10%) was deduced.
Enzyme activity analyses: NADH:ubiquinone oxidoreductase activity (complex I) was assayed on membrane fractions by using duroquinone as an electron acceptor. Membranes were prepared as follows. Cells from 300-ml TAP cultures (2–3.106 cells/ml) were collected by centrifugation (700 × g for 10 min). They were resuspended in 2 to 3 ml of MET buffer (280 mm mannitol, 100 μm EDTA, 10 mm Tris-HCl pH 7, and 0.1% BSA) and then disrupted by sonication (two times for 30 sec; Vibra Cell Sonicator, Danbury, CT). The suspension was centrifugated at low speed (10 min at 480 × g followed by 4 min at 3000 × g). The supernatant was then centrifugated at high speed (27,000 × g for 15 min) and the final pellet containing membranes was suspended in 600 μl of MET. A short pulse of ultrasounds (5 sec) was used to homogenize the suspension. A total of 100 to 200 μg proteins of the membrane fraction were added to the assay buffer (20 mm Tris HCl pH 8.0, 100 μm NADH, and 100 μm duroquinone) in a final volume of 1 ml. Enzyme activity was monitored by recording NADH oxidation at 340 nm, using the extinction coefficient of 6.22 mm–1 cm–1. Rotenone-sensitive NADH:ubiquinone oxidoreductase activity (complex I) was determined by adding rotenone (15 mm in ethanol) at a final concentration of 10 μm in the mixture. In each case, the inhibitory effect of ethanol alone (∼10%) was deduced. The specific activities were expressed in nanomoles NADH oxidized per minute per milligram per protein.
Succinate:cytochrome c oxidoreductase (complexes II + III) was assayed on membrane fractions prepared as above, except that MOPS buffer (280 mm mannitol, 10 mm MOPS-KOH pH 7.4, 0.1% BSA) was used instead of MET. The assay conditions were the following: 10 mm MOPS-KOH pH 7.4, 53 μm ferricytochrome c, 20 mm succinate, 1 mm KCN, and 50–200 μg proteins in a final volume of 1 ml. Enzyme activity was monitored by recording reduction of cytochrome c at 550 nm, using an extinction coefficient of 19.6 mm–1 cm–1. The specific activity was expressed in nanomoles cytochrome c reduced per minute per milligram protein.
Cytochrome c oxidoreductase (complex IV) was assayed on membrane fractions prepared as for measurements of complexes II + III activity. The assay conditions (adapted from Wisemanet al. 1977) were the following: 10 mm MOPS-KOH pH 7.4, 53 μm ferrocytochrome c, 1% Triton, 20–50 μg proteins in a final volume of 1 ml. Ferrocytochrome c was prepared according to Fritz and Beevers (1955). Enzyme activity was monitored by recording oxidation of cytochrome c at 550 nm, using an extinction coefficient of 19.6 mm–1 cm–1. Reaction was totally inhibited by 1 mm KCN. The specific activity was expressed in nanomoles cytochrome c oxidized per minute per milligram per protein.
Mutations affecting the mitochondrial genome
Protein content was determined according to the method of Bradford (1976).
Sequencing analyses: C. reinhardtii total DNA was prepared according to the procedure of Newman et al. (1990). Four segments of the C. reinhardtii mitochondrial genome (GenBank CRU03843: 1629–3301, 3236–5233, 6636–8593, and 10282–11283) containing nd4, nd5, nd2-nd6, and nd1, respectively, were amplified by PCR and fully sequenced on both strands. PCRs were performed according to standard protocol using a Taq DNA polymerase from QIAGEN (Hilden, Germany) in a thermocycler GeneAmp PCR System 9700 (PE Applied Biosystems, Foster City, CA). Amplified products were sent to GenomeExpress (Paris) for automated sequencing. Overlapping segments were assembled with the aid of the GCG sofware package (Genetics Computer Group, Madison, WI) and final consensus sequences were aligned with the MAP alignment tool freely available through the Internet (Human Genome Center, Baylor College of Medicine, Houston, TX).
RESULTS
Isolation and phenotypical analysis of mutants 169 and 300: Mutants 169 and 300 were isolated after acriflavine treatment of wild-type (strain 2) and dum19 (strain 239) cells, respectively. Their growth on TAP agar medium in darkness and in light was compared to the growth of wild-type and of previously characterized mutants (Table 1): dum19, which lacks cytochrome c oxidase (complex IV) activity; dum20, which lacks rotenone-sensitive NADH:ubiquinone oxidoreductase (complex I) activity; and dum24 deprived of complex I and complex III activities (see materials and methods). Mutants 169 and dum20 display the same phenotypes: both strains grow very slowly under heterotrophic conditions (darkness + acetate) and produce smaller colonies than wild-type and dum19 cells in the light. Mutant 300 does not grow in the dark and grows slowly in the light, as does the dum24 mutant.
The growth of the different mutants in the presence of rotenone, an inhibitor of complex I, and of myxothiazol + antimycin A, inhibitors of complex III, was also tested (Table 1). Both mutants 169 and dum20 were insensitive to rotenone and sensitive to myxothiazol + antimycin A, whereas the reverse situation was found for dum19. This suggests that, like dum20, mutant 169 has a defect in complex I whereas its cytochrome pathway of respiration is still functional. Mutants 300 and dum24 are both insensitive to the inhibitors. This is an indication that mutant 300, which lacks complex IV activity (due to the dum19 mutation), is also deprived of complex I, as is dum24. The phenotypes of mutants 169 and 300 in the absence of inhibitors are similar to those of dum20 and dum24, respectively, which reinforces these hypotheses.
Whole-cell respiration and enzyme activity analyses: Dark respiration of mutant cells from strains 169 and 300 grown under mixotrophic conditions (on TAP medium in the light) was measured with a Clark electrode. Cells from wild type, dum19, dum20, and dum24 were used as controls (Table 2). The total respiratory rate of dum20 and mutant 169 was about the same as the respiratory rate of dum19 and represents about half the rate of the wild type. The total respiratory rate of mutant 300 was very low, similar to that of dum24. The sensitivity of respiration to rotenone was also checked. The addition of rotenone induced a reduction of 56 and 50% of the cell respiration of wild-type and dum19 cells, respectively, whereas only a low effect was observed for dum20, dum24, and the two new mutant strains. This suggests again that mutants 169 and 300, like dum20 and dum24, have lost the activity of complex I.
The NADH:duroquinone oxidoreductase activity was measured using membrane fractions from wild-type and various mutant strains. The enzyme activity sensitive to rotenone represents the activity of complex I (Table 3). In wild-type and dum19 extracts, the complex I activity was 15–17 nmol NADH oxidized min–1 mg protein–1 whereas it was null in dum20, dum24, 169, and 300 mutant strains.
Respiratory rates of wild-type and mutant strains
The activities of complexes II + III and complex IV were also determined (Table 3). Except in dum24, the activity of complex II + III was not altered in the mutant strains and was even higher in 169 and 300 than in the other strains. As expected, cytochrome c oxidase (complex IV) activity was null in strains dum19 and 300 and was not modified in the other mutant strains.
Taken together, these data indicate that mutants dum20, 169, and 300 all lack complex I activity and have a reduced respiratory rate, insensitive to rotenone. Strain 300 also lacks complex IV activity, as does mutant dum19, from which it derives.
Genetical analysis: All the respiratory-deficient mutants previously isolated in our laboratory have been induced by treatment with acriflavine or ethidium bromide. Most of the mutations induced by these mutagens were shown to affect the mitochondrial genome (Remacle and Matagne 1998). The distinction between nuclear mutations and mitochondrial mutations can be easily made by genetical analysis: the former are transmitted in a Mendelian fashion whereas the latter are transmitted uniparentally, the meiotic progeny inheriting almost exclusively the mitochondrial genome from the mating-type minus (mt–) parent.
Activities of respiratory complexes
In crosses between dum20 and wild-type cells, the meiotic progeny inherited the phenotype of the mt– parent (Table 4), confirming that the mutation responsible for the Dk+/– phenotype (slow growth in the dark) affects the mitochondrial genome (Remacle and Matagne 1998). In contrast, in the cross 169 mt– × WT mt+, the meiotic segregation was Mendelian. A Mendelian transmission was also observed when a Dk+/– mt+ progeny clone (strain 163 isolated from previous cross 169 × WT) was crossed with WT mt– cells (Table 4). Thus, the mutation responsible for the Dk+/– phenotype of strains 169 and 163 affects a nuclear gene. It was called dn26 (Dk+/– phenotype, nuclear origin).
Strain 300 mt–, Dk– (no growth in the dark) and insensitive to rotenone due to the inactivation of both complex IV and complex I (see here above), was crossed to wild-type mt+ and gave rise to meiotic products, 97% of which were Dk– Rotr. In the reciprocal cross (301 × WT), all the meiotic progeny were Dk+ Rots (Table 4). These results indicate that strains 300 and 301 possess, in addition to dum19, another mitochondrial mutation (hereafter called dum25) that is responsible for the inactivation of complex I.
Isolation of the mitochondrial dum25 mutation: It has been previously shown that the few “vegetative” zygotes that divide mitotically to produce a stable diploid progeny transmit the mitochondrial genomes from both mt+ and mt–parents. In these zygotes and their diploid progeny, the mitochondrial genomes recombine at high rate to generate various types of mitochondrial DNA copies. The different mitochondrial DNA copies segregate during the successive mitotic divisions and, after 15–20 divisions, most cells are homoplasmic for all the markers (Boyntonet al. 1987; Remacleet al. 1990; Remacle and Matagne 1993).
Meiotic segregations obtained in crosses between mutant and wild-type strains
To isolate a recombinant diploid strain carrying only the dum25 mutation, an arg7-8 dum19 dum25 mt+ haploid mutant (strain 303) was first constructed. Strain 303 was then crossed to arg7-3 dum15 mt– (246). The dum15 mutation present in strain 246 affects the mitochondrial cob gene and determines the absence of complex III activity (Colinet al. 1995; Table 1). The arg7-8/arg7-3 diploid clones (phenotypically Arg+) resulting from the cross were selected on minimal (TMP) medium and after 10 days (when most of the cells are homoplasmic), the progeny cells were grown under heterotrophic (TAP, darkness) conditions. Individual colonies growing slowly in the dark and suspected to carry only the dum25 mutation were tested for their sensitivity in the light to rotenone and to antimycin A + myxothiazol. Diploid clones insensitive to rotenone and sensitive to antimycin A + myxothiazol were selected. Measurements of complex I, complexes II + III, and complex IV activities confirmed that these clones were recombinant, bearing only a defect in complex I (data not shown). Two arg7-8 haploid clones (212 mt+ and 213 mt–), growing slowly in the dark, were then selected from a cross between one diploid clone (phenotypically mt–) and the wild-type mt+ haploid strain. They showed no complex I activity and a whole-cell respiration of 5.5 nmol O2 min–1 × 10–7 cells (close to the values found for strains 169 and dum20; Table 2), nearly insensitive to rotenone. The haploid character of these two clones (data not shown) was confirmed by determining the number of nucleoids in the chloroplast by 4′6-diamidino-2-phenylindole staining (Matagneet al. 1991).
Genetic mapping of the dum25 mutation site on the mitochondrial genome: Vegetative zygotes produced from crosses between different Dk– mitochondrial mutants segregate Dk+ wild-type recombinant cells whose proportion depends on the distance separating the mitochondrial markers. For example, the dum18 and dum19 mutations, which are separated by ≈20 bp, recombine at a rate of 0.04% whereas the percentage of recombination between dum15 and dum18 or dum19, which are 4.2 kb apart, is 13.7% (Figure 1). The dum20 mitochondrial mutation, which determines the inactivation of complex I and must thus affect one of the nd genes encoding subunits of complex I, had been located at the right side of cox1 on the genetic map of the mitochondrial genome (Remacleet al. 1995; Remacle and Matagne 1998).
To determine the position of the dum25 mutation site on the genetic map, the arg7-8 dum25 mt+ mutant (strain 212) isolated above was crossed to three different mitochondrial mutants carrying a complementary arginine-auxotrophic marker (Table 5). A cross between dum19 and dum20 was also performed as a control. The percentages of Dk+ recombinants were determined when most diploid cells had become homoplasmic. The low percentage of recombination between dum20 and dum25 indicates that the two mutations are closely linked and thus probably affect the same gene. Moreover, despite a certain variability found in the recombination experiments, it can be deduced that dum20 and dum25 are more distant from dum15 (mutation in cob) than from dum19 (mutation in cox1) and are thus located at the right side of cox1 on the genetic map. It should, however, be noted that for an unknown reason, the recombination rate obtained in cross 247 × 206 (Table 5) is higher (14–17.5%) than that previously found for the same mutations (2–7%; Remacleet al. 1995).
From the data of Table 5 and those obtained previously (Remacleet al. 1995), the genetic map presented in Figure 1 can be drawn. Considering the high recombination rates found in crosses between dum19, on one hand, and dum20 or dum25, on the other hand (212 × 248 and 247 × 206), it can be postulated that both dum20 and dum25 mutations affect the nd1 gene.
Molecular characterization of dum20 and dum25 mutations and relationship between the physical map and the genetic map: To identify the dum20 and dum25 mutations at the molecular level, the nd1, nd2, and nd6 genes from wild type (strain 2), dum20 (strain 234), and dum19 dum25 (strain 300) were amplified by PCR and sequenced. The nd4 and nd5 sequences from strains 2 and 300 and the nd4 sequence from strain 234 were also determined.
—Physical map of the 15.8-kb linear mitochondrial genome of C. reinhardtii. The positions of the genes encoding subunits of respiratory complexes are indicated: cob, cytochrome b (component of complex III); cox1, subunit 1 of cytochrome c oxidase or complex IV; nd1, nd2, nd4, nd5, and nd6, subunits 1, 2, 4, 5, and 6 of complex I. Arrows indicate the directions of transcription. The positions (mutated codons in parentheses) of dum15, dum18, dum19, dum20, and dum25 mutations in cob, cox1, and nd1 genes, as well as the physical distances (base pairs or kilobase) between the different mutation sites and the recombination frequencies (percentages), are indicated (from Remacleet al. 1995; and data from Table 5 and Figure 2).
A single frameshift mutation, corresponding to the deletion of one T at codon 243 of nd1, was found in strain 234 (Figure 2). The same deletion was present in two other dum20 isolates derived from crosses (data not shown). No other difference was found between the nd1, nd2, and nd6 sequences from strains 2 and 234.
In the case of strain 300, the mutation was a deletion of six contiguous base pairs of the nd1 sequence. The deletion occurred in a segment with short repeated motifs, GAG GCT GAG GCT GAG, corresponding to codons 199–203 (Figure 2). The same deletion was found in the diploid clone and in the two haploid strains 212 and 213 carrying only the dum25 mutation (see above). No deletion was found in the nd1 sequence from dum19 (strain 239), which was used to generate the mutated strain 300 (Figure 2). The consequence of the dum25 mutation is the loss of two amino acid residues (Glu-Ala or Ala-Glu) in a highly conserved polar segment separating helices E and F (Figure 3).
Frequencies of Dk+ recombinant diploid cells obtained in four different crosses
In the course of sequencing the five nd genes from wild-type and mutant strains, a few differences were found compared to the “standard” C. reinhardtii mitochondrial genome (GenBank accession no. U03843). Surprisingly, the 6-bp deletion corresponding to the dum25 mutation is also present in the nd1 sequence published by Boer and Gray (1988). Since the nd1 sequence from wild type (strain 2), dum19 (strain 239), and dum20 (strain 234; Figure 2) but also from two other strains (strains 25 and cc-277, cw15 and cw15-2, respectively; data not shown) did not have the deletion, it is possible that the nd1 sequence published by Boer and Gray corresponds to a mutant sequence or that it contains errors. U. Kück (personal communication) also found in a wild-type strain the nd1 sequence identical to our wild-type sequence. An additional argument in favor of our conclusion is the conservation of the motif E(A/G)E(A/S)ELV in all ND1 protein sequences examined (Figure 3). The same EAEAELV conserved amino acid sequence, interrupted by an intron, can also be found in the ND1 sequence from Chlamydomonas eugametos (GenBank accession no. AF008237).
Another difference with the sequence from Boer and Gray (1988) is observed at codon 193, changing GAC (D) into GTA (V) (Figure 3). As a valine is present in all strains we sequenced and also in Chlorogonium elongatus, we assume that our sequence is probably the correct one.
—Mutations detected in the nd1 sequence from dum20, dum19 dum25, and dum25 mutant strains (controls: WT and dum19). Numbers below the sequence refer to codons. The deletions are underlined. In the case of dum25, the position of the 6-bp deletion is arbitrary since the deletion can affect codons 199 to 203.
The molecular characterization of dum20 and dum25 allows us to position the two mutations on the physical map and thus to re-examine the relationship between the frequencies of recombination relative to the distances separating the mutation sites on the mitochondrial genome (Figure 1). From the data presented in Figure 1, one can conclude that there is a good correlation between the physical distances, ranging from a few base pairs to ∼10 kb, and the genetic distances corresponding to the recombination rates. Our data show that the recombination rate is ∼3%/kb in the present case.
DISCUSSION
Two mutations, dn26 and dum25, both leading to the inactivation of complex I, have been isolated after mutagenic treatment with acriflavine. The two mutant strains have the same phenotype as dum20, a mitochondrial mutant previously shown to lack complex I activity but not further characterized (Remacle and Matagne 1998). The inactivation of complex I determines a strong reduction of cell growth in darkness but also affects the growth rate under mixotrophic conditions (light + acetate). The three complex I-deficient mutants are clearly distinguished from mutants inactivated in complex III or complex IV by their capacity to grow slowly in the dark. The use of inhibitors of complex I (rotenone) or complex III (antimycin A and myxothiazol) in growth tests and in measurements of whole-cell respiration also allowed the discrimination between complex I- and complex III- or complex IV-deficient mutants (Tables 1 and 2).
—Partial alignment of ND1 amino acid sequences from R. capsulatus (AAC24997), N. crassa (P08774), Homo sapiens (P03886), Triticum aestivum (Q01148), Marchantia polymorpha (P26845), C. elongatum (CAA73993), and C. reinhardtii. Solid lines above the sequences mark the positions of hydrophobic segments E and F (Fearnley and Walker 1992). Conserved amino acid residues are shown by a shaded background. The positions of the mutations in the C. reinhardtii ND1 sequence are underlined. The asterisk shows the valine residue found at position 193 of the C. reinhardtii ND1 protein.
The absence of complex I activity determines a substantial reduction in the respiratory rate of the mutant cells. The oxidation of NADH produced in the tricarboxylic cycle probably must occur through the activity of a nonproton pumping NAD(P)H-dehydrogenase. Such an enzyme bound to the inner membrane has been identified in mitochondria from higher plants (Siedow 1995; Soole and Menz 1995) and is also present in mitochondria from Chlamydomonas (Atteia 1994; Remacle and Matagne 1998). The drastic reduction of heterotrophic growth displayed by complex I mutant cells points out the importance of the phosphorylating enzyme in the respiratory metabolism.
A sequencing analysis allowed the characterization of the two mitochondrial mutations, dum20 and dum25,at the molecular level. The dum20 mutation corresponds to a deletion of one T in a context where several base pairs are repeated whereas dum25 corresponds to the deletion of six contiguous base pairs in a short segment containing two different repeated motifs (Figure 2). It is interesting to note that two other mutations previously characterized (dum18 and dum19) correspond to the deletion or the addition of one T in a run of three or four T (Colinet al. 1995). Because all these mutations were induced by treatment with acriflavine, it is likely that they result from replication errors favored by distortion of the double helix structure when the intercalating dye is bound to DNA.
Both the dum20 and dum25 mutations affect the nd1 gene, which points out the essential role of the ND1 subunit in the activity of complex I. As mentioned in the Introduction, ND1 binds rotenone and has been proposed to carry the ubiquinone-binding domain(s). To our knowledge, very few mutations affecting only nd1 have been described. In humans, the ND1/3460 mutation changes an alanine to threonine in a highly conserved region of the subunit and causes Leber's hereditary optic neuropathy (Huoponenet al. 1991). The mutation provokes a marked decrease in the specific activity of complex I (Majanderet al. 1991; Carelliet al. 1997). In Rhodobacter capsulatus, disruption of the gene equivalent to nd1 totally suppresses the activity of complex I (Dupuiset al. 1998b). In the complex CMSII mitochondrial mutant of N. sylvestris, the absence of ND1 and ND7 and the important reduction of other complex I subunits lead to the quasi-loss of complex I activity (Gutierres et al. 1997, 1999).
The nd1 gene of C. reinhardtii encodes a polypeptide of 294 amino acids (taking into account the two additional codons in comparison to the sequence published by Boer and Gray 1988) and is homologous to other ND1 subunits, as judged by similarities in primary sequence and hydropathy (Boer and Gray 1988). The ND1 proteins contain eight defined hydrophobic stretches corresponding to transmembrane helices (Fearnley and Walker 1992; Kurkiet al. 2000). The deletion of one T corresponding to dum20 mutation occurs at codon 243, just after helix F of the protein (Figure 3). It changes the reading frame downstream of the mutation site and induces a stop signal at codon 273 of the nd1 sequence. This region of the protein must thus be essential for enzyme activity since its modification inactivates complex I.
The deletion of 6 bp corresponding to the dum25 mutation occurs in a very conserved segment located between helices E and F, at the matrix side of the inner membrane (Figure 3). As pointed out by Fearnley and Walker (1992), the two best conserved regions of ND1 are in polar segments linking hydrophobic spans A and B, and E and F, respectively. The dum25 deletion occurs in a short segment that contains repeated motifs and transforms the amino acid EAEAELV into EAELV, without changing the reading frame (Figure 3). The sequence E(A/G)E(A/S)ELV is invariant in all sequences examined so far. Interestingly, the three-glutamate cluster bears some resemblance to the dicyclohexylcarbodiimide (DCCD)-binding sequence of the c subunit of ATPase F0 (Hassinen and Vuokila 1993). DCCD, an inhibitor of the mitochondrial proton pumping complexes, binds notably to the ND1 subunit (Yagi and Hatefi 1988) and competes for the same site as piericidin A (Hassinen and Vuokila 1993). This complex I inhibitor was shown to bind to a domain that overlaps the rotenone-binding site (Okunet al. 1999). Moreover, it has been recently proposed that many of the inhibitors of complex I share a common wide binding domain located between the two ubiquinone reaction centers (Tormo and Estornell 2000). On the other hand, site-specific mutagenesis in a bacterial homologue of ND1 suggests that residues located close to helices E and F on the cytoplasmic side of the membrane (the matrix side in mitochondria) are intimately involved in ubiquinone binding and reduction (Kurkiet al. 2000). It is thus tempting to propose that the amino acid residues modified by the dum25 mutation play a major role in the binding and reduction of ubiquinone.
The recombination analysis involving the dum20 and dum25 mutations has allowed the completion of the genetic map of the mitochondrial genome from C. reinhardtii. Until now, only mutations affecting the mitochondrial genes cob and cox1 had been characterized at the molecular level and used in parallel for a recombinational analysis (Remacleet al. 1995). The results obtained in that work led to the conclusion that the frequency of recombination was 3.2% (±0.7%)/kb. The present data extend the genetic map to a third gene (nd1) and confirm that the recombination rate per kilobase is close to 3%. Moreover, it shows that the recombination frequency found for a distance of ∼10 kb (i.e., the distance separating dum15 from dum20 or dum25) is higher than the frequencies found for distances of ∼5 kb (Figure 1). This means that the recombinational analysis can be extended to mutation sites separated by distances as high as 10 kb (which corresponds to two-thirds of the genome) and thus constitutes a powerful tool to position any mutation on the algal mitochondrial genome. In this respect, Chlamydomonas constitutes a unique model system since, in Saccharomyces cerevisiae, the maximum recombinational rate (20–25%) is reached for a distance of only 1.5 kb (Dujon 1981), which represents a very small part of the yeast mitochondrial genome.
Acknowledgments
We thank J. Vaassen for technical assistance and M. Dejace for manuscript preparation. This research was supported by grants from the Belgian Fonds National de la Recherche Scientifique (1.5.211.99 and 2.4552.01). D.B. and P.C. are Research Fellows and C.R. is a Research Associate from FNRS.
Footnotes
-
Communicating editor: K. J. Newton
- Received October 2, 2000.
- Accepted April 17, 2001.
- Copyright © 2001 by the Genetics Society of America
