INTRODUCTION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Besides the cyanide-sensitive cytochrome pathway, mitochondria from higher plants, some protists, and many fungi possess an alternative pathway that is resistant to cyanide but sensitive to salicylhydroxamic acid and n-propyl gallate. Cyanide-resistant respiration is mediated by a single non-phosphorylating enzyme, the alternative oxidase (AOX), which bypasses proton-translocating complexes III and IV of the cytochrome pathway to directly transfer electrons from reduced ubiquinone to molecular oxygen. In the thermogenic spadix of Sauromatum guttatum and other Araceae, the free energy of the alternative pathway is involved in heat production during anthesis (Moore and Siedow, 1991). Although its precise function in other tissues is still not fully understood, the AOX is often considered to be a regulatory enzyme balancing carbon metabolism and electron transport. According to the energy overflow hypothesis (Lambers, 1982), shunting electrons to the alternative pathway would allow continued operation of glycolysis and tricarboxylic acid cycle when the cytochrome pathway is impaired or restricted by a high adenylate charge (for review, see Wagner and Krab, 1995; Vanlerberghe and McIntosh, 1997). The alternative pathway is also thought to prevent over-reduction of respiratory chain components that might otherwise result in the generation of harmful reactive oxygen species (for review, see Moller, 2001).

Enhanced alternative respiration is observed following various developmental or environmental stimuli, and especially in stress conditions. Regulation of AOX activity is complex and occurs at both transcriptional and posttranslational levels. In isolated mitochondria from higher plants, AOX activity strongly increases upon reduction of an intersubunit disulphide bridge, yielding a non-covalently linked dimeric protein. Under its reduced form, the enzyme is activated by -keto acids, including pyruvate, through the formation of a thiohemiacetal. Both regulatory mechanisms occur at the same highly conserved Cys residue of the enzyme (for review, see Affourtit et al., 2002). In fungi and protozoa, most studies so far point to a monomeric AOX that is only stimulated by ADP, AMP, and GMP, whereas the AOX from higher plants is unaffected by these purine nucleotides (for review, see Joseph-Horne et al., 2001). Comparisons of AOX protein sequences reveal a domain of about 40 amino acids surrounding the regulatory Cys in the plant sequences that is not found in the fungal sequences (Umbach and Siedow, 2000). This conserved domain may be involved in dimerization and posttranslational regulation of the plant enzyme. These discrepancies in structural and regulatory properties between AOXs from both groups suggest potential differences in their physiological roles (for review, see Joseph-Horne et al., 2001).

In higher plants, where AOX is generally encoded by small multigene families, many conditions are known to enhance Aox gene transcription and/or AOX protein accumulation. Some of these conditions include saline (Hilal et al., 1998) and oxidative stresses (Wagner, 1995; Vanlerberghe and McIntosh, 1996), exposure to heavy metals (Padua et al., 1999), inhibition of mitochondrial protein synthesis (Zhang et al., 1996), cytochrome pathway restriction (Vanlerberghe and McIntosh, 1992, 1994; Saisho et al., 1997; Wagner and Wagner, 1997; Tanudji et al., 1999; Ducos et al., 2001), and incubation with metabolites such as acetate, Cys, or salicylic and citric acids (Elthon et al., 1989; Rhoads and McIntosh, 1992; Vanlerberghe and McIntosh, 1996; Lennon et al., 1997; Potter et al., 2000). In fungi, AOX induction is especially dependent on cytochrome pathway restriction (Lambowitz et al., 1989; Sakajo et al., 1991; Minagawa et al., 1992; Li et al., 1996; Yukioka et al., 1998; Huh and Kang, 1999, 2001) and triggering by reactive oxygen species (Yukioka et al., 1998; Huh and Kang, 2001).

The alternative pathway of the unicellular green alga Chlamydomonas reinhardtii has been mainly investigated in terms of cell respiratory capacities. Under photoautotrophic conditions, the addition of cyanide only weakly reduces (about 20%) the cell respiratory rate (Husic and Tolbert, 1987; Goyal and Tolbert, 1989), indicating that the alternative pathway capacity is constitutively high in C. reinhardtii. This capacity considerably decreases during the log growth of cells cultivated at high CO₂, but returns as the culture enters the stationary phase. At high CO₂, when the capacity of the alternative pathway is low, addition of cyanide or salicylic acid induces its reappearance (Goyal and Tolbert, 1989). The presence of acetate in the medium determines a 2- to 3-fold stimulation of respiration in air-grown cells (Fett and Coleman, 1994; Endo and Asada, 1996), due to an increase of the cytochrome pathway capacity without significant modification of the alternative pathway capacity (R.F. Matagne, D. Bauzain, and M. Dinant, unpublished data). The AOX enzyme from C. reinhardtii is recognized by a monoclonal antibody raised against the AOX of S. guttatum (Derzaph and Weger, 1996).

The algal protein is encoded by two different genes, the Aox1 gene being much more transcribed than Aox2. Both genes lack the 40-amino acid conserved domain and the regulatory Cys typical of the plant sequences and are thus more similar to the Aox genes described in fungi (Dinant et al., 1998, 2001). In the genome of C. reinhardtii, the Aox1 gene, previously named Nar5 (Quesada et al., 1998), lies in opposite orientation by the side of Nrt2;3, a gene coding for a nitrite/nitrate transporter (Quesada et al., 1998; Rexach et al., 1999; Navarro et al., 2000). Both Aox1 and Nrt2;3 genes are under control of the nitrogen source. Their expression is down-regulated by ammonium and is considerably stimulated by nitrate and nitrite in a deletion mutant that lacks several nitrate-related assimilation genes, including the Nia1 gene encoding nitrate reductase (Quesada et al., 1998, 2000).

In this work, we have fused the Aox1 promoter region to the promoterless arylsulfatase (Ars) reporter gene to study the transcriptional regulation of the Aox1 gene in C. reinhardtii. We show that the expression of the Aox1/Ars chimeric construct does not respond to the classical inducers for AOX. The expression is primarily regulated by the nitrogen source, being down-regulated by ammonium and stimulated by nitrate. Moreover, in nitrate-grown cells, inactivation of nitrate reductase by azide, tungstate, or mutation of the Nia1 gene increases the expression of the chimeric construct. This kind of regulation also occurs at the AOX protein and respiratory levels. A deletion analysis of the promoter region allowed us to define segments required for basal expression and regulation of the Aox1 gene.

	RESULTS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

The Aox1 Promoter Efficiently Drives the Expression of the Ars Reporter Gene

To investigate the transcriptional regulation of the Aox1 gene, we fused its 1.4-kb promoter region to the Ars-encoding Ars reporter gene deprived of its own promoter. The Aox1 promoter segment extends from 1364 to +60 with respect to the transcription start site and encompasses the ATG start codon of Nrt2;3 (see "Materials and Methods" and Fig. 5A). This chimeric construct called Aox1/Ars was introduced into the wall-less Arg-requiring 325 strain of C. reinhardtii by cotransformation with the pASL plasmid bearing the wild-type Arg7 gene as the selectable marker. The pJD54 plasmid harboring the promoterless Ars gene was used as a control. The percentage of transformants expressing ARS was determined by in situ staining of approximately 500 Arg-independent clones. No colony transformed with pJD54 displayed ARS activity, whereas 39% of the colonies transformed with Aox1/Ars produced the enzyme. Sixteen transformants exhibiting various levels of ARS activity and two negative controls were selected for further analyses.

To verify the status of the chimeric construct in the genome of the 16 transformants, their DNA was purified and submitted to PCR amplification. By using two different sense oligonucleotides in the Aox1 promoter region and an antisense primer in the Ars coding sequence, we concluded that 10 transformants bore at least one copy of the 1.4-kb promoter region, whereas five of the six others carried one or more 5'-truncated fragments containing at least approximately 650 bp of the 1.4-kb segment. Southern-blot hybridizations with a 1.2-kb probe located in the promoter region showed that most transformants had received several copies of the Aox1/Ars construct (data not shown). As expected, control and untransformed strains only displayed the signal corresponding to the endogenous Aox1. Clone 6, bearing a unique full copy of the chimeric construct, as well as clone 2, carrying two truncated copies, and clone 3, having at least one full copy among about four integration events, were selected for further study.

In Contrast to a Number of Known AOX Inducers, the Nitrogen Source Strongly Modulates the Expression of the Chimeric Construct

The expression of AOX at transcriptional and/or protein level was reported to increase upon exposure to a number of stress or inhibitory conditions (see the introduction). To analyze the expression of the Aox1 gene from C. reinhardtii under different experimental conditions, cells from clone 6 were cultivated for 8 h in Tris-acetate-phosphate (TAP)_NH4 liquid medium containing various agents susceptible to stimulate the alternative pathway. As shown in Table I, most treatments did not markedly modify the expression of the Aox1/Ars construct. Only paraquat and the three respiratory inhibitors, rotenone, antimycin A, and myxothiazol, very moderately stimulated the expression of the chimeric construct. Similar results were obtained with clone 3, as well as in 24-h cultivations of clones 3 and 6 (data not shown).

Because the expression of the chimeric construct was much higher in medium containing nitrate as nitrogen source (see below), ARS activities were reexamined in TAP_NO3 medium, except for saline and osmotic stresses (Table I). Only copper was able to slightly enhance expression of the chimeric construct. Such an effect of copper on Aox transcription was previously reported for sycamore (Acer pseudoplatanus) suspension cells and interpreted as a part of a general stress response, as the consequence of a copper-induced inhibition of the cytochrome pathway, or as the result of some interaction between iron and copper pools in cells (Padua et al., 1999). Otherwise, none of the tested agents was able to significantly increase the expression of the Aox1/Ars construct, hinting that the transcriptional regulation of Aox1 in C. reinhardtii is different from that of many other organisms.

To study the impact of mitochondrial mutations on the expression of the chimeric construct, we crossed clone 6 with several dum mitochondrial mutant strains isolated in our laboratory (for review, see Remacle et al., 2001b). The mutants used were deprived of complex I (dum20 and dum25), complex III (dum2), complex IV (dum19), or complex I + III (dum24) activities. Meiotic products carrying both the dum mutation and the Aox1/Ars construct were selected on nitrate-containing TAP agar medium. None of these progeny clones expressed an ARS activity significantly higher in TAP_NH4 or TAP_NO3 medium than clone 6 (data not shown). Aox1 transcript level similarly was not significantly modified in mutant cells (data not shown).

To examine the influence of the nitrogen status on the expression of the chimeric construct, ARS activities of transformants 2, 3, and 6 were analyzed after transfer for 8 h from TAP_NH4 into TAP liquid medium deprived of nitrogen or containing different nitrogen sources. As shown in Figure 1A, ammonium removal (N-free medium) led to a nearly 2-fold stimulation in the expression of the Aox1/Ars construct, whereas substitution of 4 mM nitrate for equimolar ammonium resulted in a 4- to 5-fold increase of expression. Moreover, cultivation in the presence of both ammonium and nitrate gave rise to an intermediate expression level. Taken together, these results suggest that ammonium and nitrate mediate competing effects on the expression of the Aox1 promoter.

View larger version (13K): [in this window] [in a new window]   Figure 1.   Effect of different nitrogen sources and concentrations on the expression of the Aox1/Ars chimeric construct. A, Ammonium-grown cells from clones 2, 3, and 6 were transferred for 8 h into TAP liquid media deprived of nitrogen (N-free) or containing 4 mM ammonium chloride (NH4), 4 mM sodium nitrate (NO3), or 4 mM ammonium nitrate (NH4NO3). ARS activities (nanomoles of naphthol per hour per milligram of protein) are the means of independent experiments ± SE (n = 4 for clones 2 and 3; n = 6-12 for clone 6). B, Ammonium-grown cells from clone 6 were transferred for 8 h into TAP liquid media deprived of nitrogen (black circles) or containing 1 to 32 mM ammonium chloride (white triangles) or 0.05 to 32 mM sodium nitrate (gray triangles). Each symbol is the ARS activity determined in an independent culture.

To analyze the expression of the chimeric construct in response to the nitrogen-source concentration, cells from clone 6 were transferred into TAP liquid media containing increasing molarities of ammonium or nitrate, and ARS activities were determined after 8 h (Fig. 1B). Whereas expression in ammonium slowly decreased with concentration down to one-third of the N-free level, ARS activity steeply increased and peaked (5-fold stimulation) around 1 mM nitrate. Beyond that point, ARS activity grew much more slowly to finally reach a 6-fold increase. Such a biphasic dose response curve to nitrate was also observed with cells from clone 3 (data not shown).

Azide Enhances the Expression of the Chimeric Construct without Blocking the Cytochrome c Oxidase

Searching for treatments able to stimulate the expression of the Aox1/Ars construct, we also examined the effect of the metalloprotein inhibitor sodium azide. Given its extensive coordination chemistry with metals, it is a competitive inhibitor of many heme- or copper-containing enzymes, including the mitochondrial cytochrome c oxidase (Smith and Wilcox, 1994; Saisho et al., 2001). As shown in Figure 2A, ARS activity was notably higher when cells from clone 6 were cultivated in the presence of 200 µM sodium azide. In absolute value, the increase of enzyme activity was much higher in nitrate-grown than in ammonium-grown cells. Because the addition of azide during the enzymatic incubation did not modify ARS activity, the observed stimulation actually resulted from a higher accumulation of the ARS enzyme. The same treatment applied to cells carrying both the dum19 mutation (lack of a functional cytochrome c oxidase) and the chimeric construct also induced a 2-fold increase of ARS activity in cells grown in the presence of nitrate (Fig. 2B). These results thus indicate that the higher expression of the Aox1/Ars construct following azide addition is particularly prominent for cells grown in nitrate-containing medium and is unrelated to the inhibition of cytochrome c oxidase.

View larger version (8K): [in this window] [in a new window]   Figure 2.   Effect of sodium azide on the expression of the Aox1/Ars chimeric construct. Ammonium-grown cells from clone 6 (A) or from a dum19 progeny clone lacking a functional cytochrome c oxidase (B) were transferred for 8 h into TAP liquid media containing 4 mM ammonium chloride (NH4) or 4 mM sodium nitrate (NO3), each in the absence () or in the presence (+) of 200 µM sodium azide. ARS activities (nanomoles of naphthol per hour per milligram of protein) are the means of four or five independent experiments ± SE.

Inactivation of Nitrate Reductase Further Enhances the Expression of the Chimeric Construct in Nitrate-Containing Medium

Nitrate reductase, the first enzyme of the nitrate assimilation pathway, is a heme-containing protein that is also inhibited by sodium azide (Guerrero and Gutierrez, 1977). In C. reinhardtii, this key enzymatic component of nitrogen metabolism is only present when cells are cultivated in a medium deprived of ammonium and containing nitrate (for review, see Fernandez et al., 1998). Because the positive azide effect on the expression of the Aox1/Ars construct was markedly high for cells grown in nitrate-containing medium, we wondered whether part of this stimulation was related to the inhibition of nitrate reductase. As shown in Figure 3A, the in vitro nitrate reductase activity was dramatically reduced by 200 µM sodium azide and was completely abolished by about 300 µM. Interestingly, these kinetic data are in accordance with the important azide stimulation of the chimeric construct expression in clone 6 cultivated for 16 h in TAP_NO3 liquid medium (Fig. 3B). Similar observations were made on clone 3 and in 8-h cultivations of both clones (data not shown). Because growth of clone 6 in these culture conditions was reduced by 85%, it can be thought that 200 µM sodium azide also markedly reduces the in vivo activity of nitrate reductase. Taken together, these results suggest that a large part of the positive azide effect in the presence of nitrate is related to the inhibition of nitrate reductase.

View larger version (18K): [in this window] [in a new window]   Figure 3.   Expression of the Aox1/Ars chimeric construct in relation to nitrate reductase inactivation. A, Nitrate reductase (NR) activity in crude extracts from strain 83, measured in the presence of increasing concentrations of sodium azide. The activity is expressed relative to the uninhibited activity Vmax (nanomoles of NO2 per hour per milligram of protein). B, Ammonium-grown cells from clone 6 were transferred for 16 h into TAPNO3 liquid medium containing increasing concentrations of sodium azide. ARS activity is expressed in nanomoles of naphthol per hour per milligram of protein. C, Cells from clone 6 pregrown in the presence of 1 µM molybdate or 1 µM tungstate were transferred for 16 h into TAP liquid media containing 4 mM ammonium chloride, 4 mM sodium nitrate, or 4 mM sodium nitrate with 200 µM sodium azide, each supplemented with 1 µM molybdate or 1 µM tungstate. ARS activities are the means of six independent experiments ± SE. D, Ammonium-grown cells from nia1 + or nia1 (lacking a functional nitrate reductase) strains derived from clone 6 were transferred for 8 h into TAP liquid media containing 4 mM ammonium chloride, 4 mM sodium nitrate, or 4 mM sodium nitrate with 200 µM sodium azide. ARS activities are the means of three independent experiments ± SE.

The nitrate reductase enzyme uses molybdenum as a metal cofactor and is inactivated by substituting tungstate for molybdate in the growth medium (Vega et al., 1971; Galvan et al., 1992; Loppes et al., 1999). To analyze by another approach the effect of nitrate reductase inactivation on the expression of the Aox1/Ars construct, cells from clone 6 were maintained for 10 generations on TAP agar medium supplemented with 1 µM molybdate or 1 µM tungstate. They were then transferred for 16 h into TAP liquid media containing 4 mM ammonium, 4 mM nitrate, or 4 mM nitrate with 200 µM sodium azide, each supplemented with molybdate or tungstate. As shown in Figure 3C, tungstate did not affect ARS activity in TAP_NH4 medium nor in azide-supplemented TAP_NO3 medium. In contrast, expression in TAP_NO3 medium was increased more than twice by tungstate to reach the sodium azide level, whereas both nitrate reductase activity and cell growth were reduced by about 90% (data not shown). These results were repeated on clone 3, as well as in 8-h cultivations. Here again, the data strongly suggest that in nitrate-grown cells, the inactivation of nitrate reductase markedly enhances the expression of the chimeric construct.

To further test this hypothesis, clone 6 was crossed with a nia1 mutant strain lacking a functional nitrate reductase, and a meiotic product carrying both the nia1 mutation and the Aox1/Ars construct was isolated. ARS activities were measured in cells cultivated for 8 h in TAP liquid media containing ammonium, nitrate, or nitrate with sodium azide. As shown in Figure 3D, ARS activity of nia1 cells was about 12-fold higher in TAP_NO3 than in TAP_NH4 medium. Moreover, the ARS activity recorded in TAP_NO3 medium was nearly twice higher with the nia1 mutant than with a nia1⁺ wild-type clone isolated from the same cross. Finally, azide barely increased the expression of the chimeric construct in the nia1 mutant, whereas the nia1⁺ clone showed the familiar stimulation pattern. These results were repeated in 24-h cultivations and following direct cotransformation of a nia1 strain with the Aox1/Ars construct (data not shown). Taken with the observations reported above, these data indicate that any kind of nitrate reductase inactivation enhances the nitrate-mediated stimulation of the chimeric construct expression. Moreover, because sodium azide also enhances the expression of the Aox1/Ars construct in ammonium-grown wild-type cells (Fig. 2A), it might marginally act on another enzymatic system unrelated to nitrate reductase.

The AOX Protein and Alternative Respiration Levels Parallel the Expression of the Chimeric Construct

Having demonstrated that the chimeric construct is transcriptionally down-regulated by ammonium and stimulated by nitrate, it was of interest to examine the AOX protein levels of the wild-type strain cultivated in the same growth conditions. A polyclonal antibody raised against the AOX of C. reinhardtii was tested for this purpose and was found to be 100-fold more sensitive than the monoclonal antibody raised against the AOX of Sauzomatun gutattum. This specific antibody preparation could thus be used to study the AOX abundance in membrane fractions from strain 83 cultivated for 16 h in TAP liquid media deprived of nitrogen or containing different nitrogen sources. A band of approximately 40 kD was detected in all culture conditions (Fig. 4A). This band most probably corresponds to the product of Aox1 because the Aox2 gene is only poorly transcribed (Dinant et al., 2001) and because the Aox2 promoter (GenBank accession no. AF537324) is unable to drive significant expression of the Ars reporter gene under a broad range of growth conditions (data not shown). However, the mass of this protein is higher than the predicted mass (33.4 kD) of the mature AOX1 enzyme (Dinant et al., 2001). A similar discrepancy between the molecular masses calculated from the deduced amino acid sequence and determined by SDS-PAGE was observed for AOX proteins from soybean (Glycine max; Finnegan et al., 1997). A second, less intense band of approximately 35 kD was sometimes detected on the immunoblots (Fig. 4A). Reyes-Prieto et al. (2002) also detected the 40-kD polypeptide and a minor band around 36 kD using the same antibody. Considering that both bands disappear in an antisense strain deprived of alternative respiration (M. Dinant and R.F. Matagne, unpublished data), we presume that the lower band should correspond to an isoform (maybe a degradation product) of the AOX1 protein.

View larger version (24K): [in this window] [in a new window]   Figure 4.   Effect of nitrogen source and sodium azide on the AOX protein accumulation (western blots). Except in A, each lane was loaded with 30 µg of membrane proteins. A, Ammonium-grown wild-type cells (strain 83) were transferred for 16 h into TAP liquid media deprived of nitrogen (N-free) or containing 4 mM ammonium chloride (NH4), 4 mM sodium nitrate (NO3), or 4 mM ammonium nitrate (NH4NO3). Membrane proteins (15 µg for cells maintained in N-free medium; 30 µg for other conditions) were loaded on the gel to get similar signals for subunit III of cytochrome c oxidase. Ammonium-grown cells from wild-type strain (B) or from a dum19 mutant (strain 193) lacking a functional cytochrome c oxidase (C) were transferred for 16 h into TAP liquid media containing 4 mM ammonium chloride (NH4) or 4 mM sodium nitrate (NO3) in the absence () or in the presence (+) of 200 µM sodium azide. Time exposure of the screen was longer for A than for B and C.

The data presented in Figure 4A show that AOX accumulation was low in TAP_NH4 medium, higher in N-free medium, and still increased in nitrate-containing media. In the same loading conditions, we did not detect any change in the amount of subunit III of cytochrome c oxidase. The stimulation by 200 µM sodium azide was also investigated at the protein level on cells grown for 16 h in TAP_NH4 or TAP_NO3 liquid medium (Fig. 4B). AOX accumulation in the presence of azide was higher on both nitrogen sources, the abundance of the protein being particularly high in nitrate-containing medium. Moreover, another band with an apparent mass of 75 kD was systematically detected in the latter condition. This additional band could correspond to a homodimeric form of AOX associated with its overexpression. Similar responses to azide in both ammonium and nitrate media were obtained with membrane fractions from a dum19 mutant (strain 193) deprived of cytochrome c oxidase (Fig. 4C), thus confirming that the positive azide effect is unrelated to the cytochrome c oxidase inhibition. Because in ammonium-grown dum19 cells, azide treatment increased the amount of AOX protein but not ARS expression (compare Figs. 2B and 4C), one could suggest that, at least in the mutant, AOX protein accumulation does not require transcriptional activation of the Aox1 gene.

To further assess the physiological relevance of the AOX regulation by the nitrogen source, we examined the respiratory rates of wild-type cells from strain 83 grown for 16 h in TAP liquid media containing 4 mM ammonium or nitrate. Capacities of the alternative and cytochrome pathways of respiration were determined by the addition of potassium cyanide (inhibitor of cytochrome c oxidase) and n-propyl gallate (inhibitor of AOX), respectively. As shown in Table II, total respiratory rates were not affected by the nitrogen source. However, the alternative pathway capacity (i.e. the oxygen uptake inhibited by n-propyl gallate in the presence of cyanide) was at least twice higher in nitrate-containing medium than in ammonium-containing medium (15.0 versus 6.5), whereas the cytochrome pathway capacity (oxygen uptake inhibited by cyanide in the presence of n-propyl gallate) was barely increased (21.5 versus 15.7). Together, these results show that there is a close correlation among the transcriptional regulation of the Aox1/Ars construct, the AOX protein accumulation, and the capacity of the alternative pathway.

A 0.3-kb Proximal Upstream Segment Is Sufficient for Regulated Expression of the Chimeric Construct, But Distant Elements Are Required for Full Expression

To define a minimal promoter still able to mediate Aox1 transcriptional regulation, a series of 5'-deleted fragments was produced by PCR and fused to the Ars reporter gene. Six new chimeric constructs were generated, named Aox11/Ars, Aox12/Ars, Aox13/Ars, Aox14/Ars, Aox15/Ars, and Aox16/Ars, that contained portions of the Aox1 upstream region up to positions 1129, 935, 694, 466, 253, and 59, respectively (Fig. 5B). After linearization, these constructs were introduced with pASL into C. reinhardtii, and Arg-independent colonies were selected on TAP agar medium. In a first step, 25 clones from each cotransformation were picked up at random and submitted to in situ staining to determine the percentage of transformants expressing ARS. With constructs Aox11/Ars to Aox15/Ars, the co-expression rate was high (56%-80%), whereas it fell to 8% with Aox16/Ars (Fig. 5B). This indicates that the 253 to +59 segment of the Aox1 promoter is sufficient to drive the expression of the Ars reporter gene. For three different constructs, the total DNA from the 25 clones was purified and used in PCR amplification of the Aox1-Ars junction to determine the percentage of cotransformed cells. The rates obtained (79%-87%; Fig. 5B) were high and homogenous enough to examine ARS activities in pools of transformants. As shown by Ohresser et al. (1997), all of the Arg-independent colonies from one cotransformation experiment can be harvested and assayed for ARS activity as a single population without prior selection of the transformants harboring the chimeric construct. This procedure bypasses the so-called "position effects" and obviates the need for molecular characterization of integration events in dozens of individual transformants. About 300 colonies from each cotransformation were thus harvested as a pool and plated onto fresh TAP agar medium. After 3 d, these cells were transferred for 16 h into TAP liquid media containing 4 mM ammonium, 4 mM nitrate, or 4 mM nitrate with 200 µM sodium azide, before determination of ARS activities. The results of two independent cotransformation experiments are presented in Figure 6A. Except for Aox16/Ars that gave rise to very low ARS activities, not significantly different from that of a pool bearing either pJD54 (promoterless reporter gene) or no construct, ARS expression could be analyzed with all of the constructs. The general trend in TAP_NO3 and azide-supplemented TAP_NO3 media was a decrease of expression from Aox11/Ars to Aox15/Ars. However, ARS activities of these five constructs remained higher in TAP_NO3 than in TAP_NH4 medium and were still subject to azide stimulation. This indicates that proximal elements ensuring basal transcription and regulation lie in the short promoter segment present in Aox15/Ars and not in Aox16/Ars (253 to 59), but that distant elements are required for full expression.

View larger version (20K): [in this window] [in a new window]   Figure 5.   A, Schematic drawings of the Aox1 and Nrt2;3 genes arranged in opposite orientation and of the Aox1/Ars chimeric construct. Nrt2;3, Aox1, and Ars coding sequences are shown as black boxes. B, Constructs Aox11/Ars to Aox16/Ars are derivatives of pJD54; K, unique KpnI restriction site used for linearization. C, Constructs Aox151/Ars to Aox153/Ars are derivatives of pJD100. Gray boxes represent the TubB2 minimal promoter of constructs Aox151/Ars to Aox153/Ars. Numbering is relative to the transcription start site, which is preceded by a thin black area symbolizing the putative TATA box. Whenever determined, the co-expression rate on TAP agar medium and the cotransformation percentage are given in B and C (n = 25; see text for details). nd, Not determined.

View larger version (12K): [in this window] [in a new window]   Figure 6.   Deletion analysis of the Aox1 promoter. Ammonium-grown cells from pools of transformants obtained with six linear constructs (A) or with five circular constructs (B) were transferred for 16 h into TAP liquid media containing 4 mM ammonium chloride (white circles), 4 mM sodium nitrate (gray circles), or 4 mM sodium nitrate with 200 µM sodium azide (black circles). Each symbol is the average ARS activity (nanomoles of naphthol per hour per milligram of protein) determined from a single cotransformation experiment. AA, Aox1/Ars construct; 1 to 5, constructs Aox11/Ars to Aox15/Ars; 51 to 53, constructs Aox151/Ars to Aox153/Ars.

To further demonstrate that critical elements are present within the 253 to 59 promoter region, three 3' deletions of the Aox15 fragment were cloned into pJD100. This plasmid encodes the Ars reporter gene driven by a 35-bp minimal promoter derived from the C. reinhardtii TubB2 gene (Davies and Grossman, 1994). The new chimeric constructs, named Aox151/Ars, Aox152/Ars, and Aox153/Ars, contained portions of the Aox1 upstream region from position 267 to positions 39, 133, and 187, respectively (Fig. 5C). Because these constructs could not be linearized, they were introduced under circular form into C. reinhardtii by cotransformation with pASL. Circular Aox1/Ars and Aox15/Ars were used as controls. After selection of Arg-independent colonies, the percentage of transformants expressing ARS was determined as above. Though lower than in the previous experiment, probably because circular plasmids are more subject to disruption during integration (Davies and Grossman, 1994; Loppes and Radoux, 2001), the co-expression rates were maximal with controls (27% and 31%) and minimal with Aox152/Ars and Aox153/Ars (4% and 8%), suggesting that these two constructs were unable to promote significant Ars transcription (Fig. 5C). The cotransformation rate for three different constructs were still highly homogenous (75%-79%; Fig. 5C), thus allowing the recording of ARS activities in pools of transformants as previously done.

In two independent cotransformation experiments, about 200 colonies from each cotransformation were pooled, plated onto fresh TAP agar medium, and then transferred for 16 h into TAP liquid media containing ammonium, nitrate, or nitrate with sodium azide. ARS activities with Aox152/Ars and Aox153/Ars were negligible and not significantly different from those of a pool bearing either pJD100 or no construct (Fig. 6B). In contrast, the Aox151/Ars construct lacking the 39 to +59 upstream segment of Aox15/Ars was still expressed but no more regulated. As previously, the deleted Aox15/Ars construct showed a lower expression level than the full Aox1/Ars construct, but basically responded to both nitrate and azide stimulations.

Together, the results obtained with the different chimeric constructs indicate that (a) elements between 133 and 39 are needed for sufficient transcription of Aox1; (b) elements conferring a response to nitrate and sodium azide lie in the 39 to +59 segment; and (c) elements in the distal part of the Aox1 promoter are required to ensure full gene expression and regulation.

	DISCUSSION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

To study the transcriptional regulation of the main AOX gene from C. reinhardtii, we fused the full 1.4-kb upstream region of Aox1 to the Ars reporter gene. We showed that expression of the chimeric construct was generally not enhanced by a number of known AOX inducers, including stress agents, respiratory inhibitors, and metabolites. A slight stimulation of the ARS expression was mostly observed when cells with less abundant AOX content (i.e. ammonium-grown cells) were exposed to respiratory inhibitors and paraquat (Table I). This contrasts with the situation described in many higher plants and fungi where the tested agents strongly stimulate the expression of the Aox gene and/or the accumulation of the AOX protein (see the introduction). However, it is worth noting that in the absence of inducers, the capacity of the alternative pathway is constitutively high in C. reinhardtii (see Table II). In several fungi, the AOX activity conversely is negligible under normal growth conditions and becomes detectable only when the cytochrome pathway of respiration is blocked by chemical inhibition or by mutation (Lambowitz et al., 1989; Sakajo et al., 1991; Minagawa et al., 1992; Li et al., 1996; Yukioka et al., 1998; Huh and Kang, 1999, 2001).

It has also to be pointed out that AOX isoforms unresponsive to cytochrome pathway inhibition and chemical induction have been described in several organisms (Saisho et al., 1997; Tanudji et al., 1999; Huh and Kang, 2001). In eudicot plant species in particular, the AOX is encoded in two discrete gene subfamilies: The Aox1 subfamily is responsive to stress agents and respiratory inhibitors, whereas the Aox2 subfamily shows tissue and developmental stage specificity but does not respond to Aox1-inducing agents (Considine et al., 2002).

Apparently peculiar to C. reinhardtii is the regulation of the amount of Aox1 transcript and AOX protein by the nitrogen source (Quesada et al., 1998, 2000; this work). Simple removal of ammonium from the culture medium resulted in a significant increase of expression at both transcriptional and protein levels (Figs. 1A and 4A). Addition of nitrate further enhanced the expression, whereas the combination of ammonium and nitrate led to an intermediate expression level. In this alga, four high-affinity nitrate/nitrite transport systems are responsible for the entry of nitrate and nitrite into the cell. These transporters are differentially regulated by the nitrogen source and exhibit variable affinity for their substrates (for review, see Galvan and Fernandez, 2001). As shown in Figure 1B, the expression of the Aox1/Ars construct in relation to the nitrate concentration displays a biphasic curve. The change between the two response curves might reflect a switch between two of these transport systems. Moreover, one transporter (system IV) is known to be resistant to the ammonium inhibition affecting the other transport systems (Rexach et al., 1999; Navarro et al., 2000), which could account for the competing effects of ammonium and nitrate on the expression of Aox1 transcript and AOX protein accumulation. The same hypothesis has been proposed by Llamas et al. (2002) to explain the transcriptional regulation of the Nia1 gene by nitrate and ammonium.

In nitrate-grown cells, the inactivation of nitrate reductase by sodium azide, tungstate, or mutation of the Nia1 gene further enhances the expression of the chimeric construct and the amount of AOX protein. Hence, intracellular accumulation of nitrate would provide a permanent positive signal, as observed for the expression of nitrate-regulated genes (Rexach et al., 2002). Such a model is in accordance with the up-regulation of the C. reinhardtii Aox1 transcript in the G1 mutant lacking Nia1 and five other nitrate-related assimilation genes but retaining a functional system IV transporter (Quesada et al., 1998; Navarro et al., 2000).

The 75-kD polypeptide visualized after azide treatment of nitrate-grown cells could correspond to a dimeric form of the AOX enzyme. However, the dimerization domain surrounding the regulatory Cys typical of the plant sequences is missing in both AOX isozymes from C. reinhardtii as well as in all fungal sequences (Umbach and Siedow, 2000; Dinant et al., 2001). Besides, the algal AOX enzyme is not stimulated by pyruvate (M. Dinant, personal observation). Because treatment with dithiothreitol did not eliminate the band at 75 kD in the SDS gel (data not shown), the detected signal could represent an artifact, possibly created occasionally during heating of the sample, as observed for the AOX of Acanthamoeba castellanii (Jarmuszkiewicz et al., 1997).

Although the major effect of sodium azide occurs through nitrate reductase inhibition in nitrate-containing medium, stimulation was also observed in the presence of ammonium, mainly at protein level (Fig. 4, B and C). This stimulation, independent from cytochrome c oxidase inhibition, must result from the inactivation of another unidentified enzymatic system.

In C. reinhardtii, regulatory sequences appear in most cases to be located within 200 to 300 bp from transcription start sites (for review, see Kindle, 1998). This rule is in agreement with our 5'-deletion analysis, demonstrating that a shortened promoter corresponding to the 253 to +59 segment is sufficient to ensure basal transcription and regulation of the Aox1 gene. More specifically, elements lying between 133 and 39 are required for gene expression, whereas elements conferring response to the nitrogen source and azide should be found in the 39 to +59 region. However, the most expressed and regulated constructs extend to the Nrt2;3 promoter range, which suggests that distant elements are probably also involved in the expression of the Aox1 gene.

The 1.4-kb segment lying between Aox1 and Nrt2;3 was examined for conserved elements susceptible to mediate transcriptional regulation of the Aox1 gene. Various 4- to 6-bp motifs previously identified in other C. reinhardtii promoters were found, but none concerned a nitrogen-related element described in the Nia1 gene promoter from either C. reinhardtii (Loppes and Radoux, 2001, 2002) or Chlorella vulgaris (Cannons and Shiflett, 2001). A linker-scanning mutagenesis is necessary to identify responsive elements in the Aox1 promoter (Quinn et al., 2000; Loppes and Radoux, 2002).

Our data raise the question of the physiological signification of the AOX regulation by the nitrogen source. The photosynthetic reduction of nitrate into nitrite requires the oxidation of NAD(P)H. Light reactions inside the chloroplast drive the coupled production of NADPH and ATP. In illuminated green tissues and microalgae, oxygen evolution is stimulated when nitrogen is supplied as nitrate rather than ammonium (Bloom et al., 1989; de la Torre et al., 1991), which suggests that nitrate assimilation enhances the noncyclic electron flow by acting as an alternative electron acceptor in the chloroplast. As hypothesized by Quesada et al. (2000), the nitrate reduction might result in an increased production of photogenerated ATP. In this context, the specific stimulation of AOX production by nitrate and not by ammonium might indicate the requirement for the cell to modify the partitioning of electron flow to adapt mitochondrial ATP synthesis during assimilation of these different nitrogen sources. Alternatively, the regulation observed at transcriptional, protein, and alternative pathway capacity levels might simply be due to the side effect of a nitrogen-related Nrt2;3 enhancer on the Aox1 gene expression. However, it is worth mentioning that the two genes partly differ in their regulation pattern: Although both are down-regulated by ammonium and better expressed in nitrate-containing medium, Nrt2;3 is only transcribed when the regulatory Nit2 gene is functional, which is not the case for Aox1 (Quesada et al., 1998; R.F. Matagne and M. Dinant, unpublished data). Further investigations on other microalgae and higher plants would clear up this intriguing question.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Chlamydomonas reinhardtii Strains and Culture Conditions

The host strain used in cotransformation experiments was the cell wall deficient, Arg-requiring cw15 arg7-8 mt⁺ mutant (strain 325, lacking argininosuccinate lyase). Two other strains were used in additional experiments: cw15 mt⁺ (strain 83) and cw15 dum19 mt (strain 193, lacking cytochrome c oxidase). These three strains were able to assimilate nitrate as sole nitrogen source. Mutant strains used in genetic crosses were the following: dum20 mt and dum25 mt (strains 235 and 228, lacking NADH-ubiquinone oxidoreductase), dum2 mt (strain 166, lacking ubiquinone-cytochrome c oxidoreductase), dum19 mt (strain 239, lacking cytochrome c oxidase), dum24 mt (strain 283, lacking both NADH-ubiquinone and ubiquinone-cytochrome c oxidoreductases), and cw15 nia1 mt (strain 69, lacking nitrate reductase). The dum strains inactivated in mitochondrial genes and the nia1 strain (also known as nit1) are reviewed by Remacle et al. (2001b) and Fernandez et al. (1998), respectively.

Cells were grown on solidified agar medium at 25°C under continuous illumination (70 µE m² s¹). The routinely used medium was TAP containing 7.5 mM NH₄Cl as nitrogen source (Harris, 1989) and 100 mg L¹ Arg when required. For liquid cultures, we used TAP medium containing 4 mM NH₄Cl, 4 mM NaNO₃, or 4 mM NH₄NO₃, or TAP deprived of any nitrogen source. Stress agents, inhibitors, and metabolites were sterilely added to the TAP liquid media after autoclaving. In the experiments devoted to the effects of tungstate, media were prepared according to Loppes et al. (1999).

General Methods

Crosses and isolation of meiotic products were performed as described by Harris (1989). Manipulation of nucleic acids was achieved by standard procedures (Sambrook et al., 1989; Ausubel et al., 1993). Plasmid DNA for sequencing and cotransformation was purified using Concert Rapid Plasmid Miniprep System (Invitrogen, Bethesda, MD) and Nucleobond AX (Macherey-Nagel, Düren, Germany). Total C. reinhardtii DNA for Southern blotting was obtained by the procedure of Newman et al. (1990). A faster cell lysis protocol communicated by L. Ellis was used in the preparation of total DNA for PCR amplification. In brief, 0.5 cm² of cells resuspended in 100 µL of PCR buffer containing 4 mg mL¹ proteinase K was incubated for 1 h at 58°C and then transferred for 1 h at 95°C. Dimethyl sulfoxide (6%, v/v) was required for optimal PCR amplification. Oligonucleotide synthesis and automated sequencing were ordered from GENOME Express (Paris). Restriction enzymes were purchased from Roche Diagnostics (Mannheim, Germany), whereas calf intestinal alkaline phosphatase, T4 DNA polymerase, T4 DNA ligase, and RadPrime DNA labeling system were ordered from Invitrogen. Amplification of DNA was carried out using either a standard Taq DNA polymerase from Amersham Biosciences (Piscataway, NJ) or a Super Taq Plus DNA polymerase from HT Biotechnology Ltd. (Cambridge, UK), an enzyme mixture displaying both high processing and proofreading activities. The buffers were those recommended by the suppliers, and the thermocycler was a GeneAmp PCR System 9700 (Applied Biosystems, Foster City, CA). All other chemicals were purchased from Sigma-Aldrich (St. Louis).

Chimeric Constructs

A 1,522-bp DNA fragment extending from Aox1 to Nrt2;3 coding sequences was amplified by PCR using forward primer 5 from Dinant et al. (2001) with the MDRP4 reverse oligonucleotide (5'-TATGGAGCCCAGAAGACGAGAAGC-3'), which is complementary to amino acids 32 to 25 of the Aox1 coding sequence (Dinant et al., 2001), and cloned into the pGEM-T Easy vector (Promega, Madison, WI) for sequencing (GenBank accession no. AF537323). Taking advantage of a unique NruI site located 3 bp before the Aox1 initiation codon, a 1,424-bp EcoRI-NruI fragment (from positions 1364 to +60 relative to the transcription initiation site) was gel-purified, blunt-ended, and inserted into the SalI restriction site of the pJD54 plasmid harboring a promoterless Ars gene (Davies et al., 1992). This construct named Aox1/Ars would produce a transcript with a 160-nt leader sequence (including 60 nt from Aox1 and 69 nt from Ars; see Fig. 5A).

The XhoI-1423-R reverse oligonucleotide (5'-AAGCCTCGAGATATCGAACGGTGAC-3'), which is complementary to the Aox1 5'-untranslated region sequence from positions +68 to +44 relative to the transcription initiation site, was combined with forward primers XhoI-235-F (5'-GTTACTCGAGGTGCACGGAAAAGCG-3'), XhoI-429-F (5'-AACCCTCGAGGCGCTGCCAGAATGA-3'), XhoI-670-F (5'-TTGGCTCGAGGTTTTCGGCATCCCC-3'), XhoI-898-F (5'-TGTCCTCGAGATGACAGCGACCTAG-3'), XhoI-1111-F (5'-CGGCCTCGAGGTGCAACCCGATGCG-3'), and XhoI-1305-F (5'-CCTGCTCGAGTCGTCTACACTGGAA-3') to amplify upstream regions between positions +59 and 1129, 935, 694, 466, 253, and 59, respectively. These PCR products were XhoI-digested and inserted into the SalI restriction site of the dephosphorylated pJD54 plasmid to generate chimeric constructs Aox11/Ars to Aox16/Ars. These deleted constructs would produce a transcript with a 159-nt leader sequence (lacking one more base from the Aox1 5'-untranslated region than original Aox1/Ars; see Fig. 5B).

Similarly, the KpnI-1097-F forward primer (5'-GCTCGGTACCCAGGGGCTCGGCCTC-3'), which is complementary to the Aox1 upstream sequence from positions 276 to 252, was used with the reverse oligonucleotides KpnI-1326-R (5'-TCGAGGTACCAGTGTAGACGATTCG-3'), KpnI-1232-R (5'-CACTGGTACCGACCGCCGCCGCAAC-3'), and KpnI-1178-R (5'-TATGGGTACCAACTTCAAGCACAGA-3') to amplify upstream regions between positions 267 and 39, 133, and 187, respectively. Likewise, the PCR products were KpnI-digested and inserted into the KpnI restriction site of the dephosphorylated pJD100 plasmid harboring a 35-bp -tubulin (TubB2) minimal promoter fused to the Ars gene (Davies and Grossman, 1994). These constructs, named Aox151/Ars to Aox153/Ars, would produce a transcript with a 170-nt leader sequence (including 65 nt from TubB2 and 69 nt from Ars; see Fig. 5C).

All junctions were checked by automated sequencing. Computer analyses were performed with the aid of the GCG software package (Genetics Computer Group, Madison, WI; available at the Belgian EMBL Node, ULB/VUB).

C. reinhardtii Cotransformation

Nuclear cotransformation of C. reinhardtii strain 325 was performed using the glass bead method (Kindle, 1990) with minor modifications (Loppes and Radoux, 2001) using 1 µg of pASL (Adam and Loppes, 1998) bearing the wild-type Arg7 gene as the selectable marker (Debuchy et al., 1989) and 5 µg of cotransforming DNA. Except otherwise stated, plasmids were linearized by BamHI (pASL) or KpnI (chimeric constructs) before cotransformation. Cells were then spread onto TAP agar plates for selection of Arg-independent transformants.

To determine the percentage of transformants expressing ARS, a number of Arg-independent clones were transferred and maintained for 3 d on TAP agar medium and then in situ stained for ARS activity (Ohresser et al., 1997). Total DNA from these clones was also purified to estimate the percentage of cotransformed cells by PCR amplification of the Aox1-Ars junction using the reverse oligonucleotide Aox1-Ars-R (5'-TTCTGAATGGCGTCCTGGTC-3'), which is complementary to amino acids 42 to 36 of the Ars coding sequence (de Hostos et al., 1989), either with forward primers XhoI-1305-F (Aox1/Ars and Aox11/Ars to Aox16/Ars) or with KpnI-1097-F (Aox151/Ars to Aox153/Ars). To examine the status of the chimeric construct in the genome of the original Aox1/Ars transformants, the Aox1-Ars-R reverse oligonucleotide was combined either with forward primer 5 from Dinant et al. (2001) or with forward primer in-pAox1-F (5'-TACGTACCGTGCTGCGAGGAGC-3'), which is complementary to the Aox1 upstream region from positions 607 to 586 relative to the transcription initiation site. The copy number of the chimeric construct was also investigated in these clones through Southern-blot hybridizations following restriction of total DNA by HinfI. The radiolabeled DNA probe was a 1.2-kb SalI fragment of the Aox1 upstream region covering positions 1359 to 140 with respect to the transcription initiation site. In deletion studies, several hundreds transformants from each experiment were pooled and maintained on TAP agar medium for further analyses.

Isolation of Membrane Fractions

Crude fractions containing mitochondrial and chloroplast membranes were prepared by cell sonication (Vibra Cell, Sonics and Materials, Newtown, CT) in MOPS buffer (280 mM mannitol, 10 mM MOPS-KOH, pH 7.4, and 0.1 M bovine serum albumin) as described by Remacle et al. (2001a).

Enzymatic Assays

Cells from isolated clones or pools of transformants were transferred from fresh TAP agar plates into glass tubes containing 2 mL of liquid medium at a density of 6 × 10⁶ cells mL¹ and maintained for 8, 16, or 24 h under continuous light with agitation. After overnight freezing, ARS activity was determined on 50 to 400 µL of the thawed cell suspension using -naphthyl sulfate as a substrate and tetrazotized-o-dianisidine as a post-coupling agent (Ohresser et al., 1997).

Cytochrome c oxidase activity was monitored by recording reoxidation of ferrocytochrome c in membrane fractions (2 µg proteins) as described by Remacle et al. (2001a). NADH-nitrate reductase activity was assayed in crude cell extracts according to Hageman and Reed (1980) by measuring the amount of nitrite formed after 15 min in the presence of potassium nitrate and NADH. Crude cell extracts were prepared by cell sonication in Tris-HCl buffer (50 mM Tris-HCl, pH 7.5, 0.1 mM dithiothreitol, 0.1 mM EDTA, and 20 µM FAD) as described by Loppes et al. (1999).

Total protein content was determined according to Bradford (1976). The spectrophotometer apparatus used in all assays was a computer-controlled UV/Vis Lambda 20 running UV WinLab software (PerkinElmer Life Sciences, Wellesley, MA).

Whole-Cell Respiration

Cells from cultures in liquid medium were harvested by centrifugation and resuspended into 1 mL of fresh medium at a density of 5 × 10⁶ cells mL¹. Respiratory rates were measured as oxygen uptake in the dark at 25°C using a Clark-type oxygen electrode (Hansatech, King's Lynn, UK). Capacities of the alternative and cytochrome pathways were estimated by addition of 1 mM potassium cyanide (in aqueous solution) and 1 mM n-propyl gallate (in ethanol), respectively.

Western-Blot Analysis

Membrane fractions were solubilized in an equal volume of denaturation buffer (125 mM Tris-HCl, pH 6.8, 10% [w/v] SDS, 20% [v/v] glycerol, and 0.1% [w/v] bromophenol blue) and boiled for 5 min. The proteins were separated by SDS-PAGE according to Laemmli (1970) using a 12% (w/v) polyacrylamide resolving gel and subsequently electrotransferred onto a Hybond C Super membrane (Amersham Biosciences) in transfer buffer (25 mM Tris, 192 mM Gly, and 20% [v/v] methanol). Prestained standard (Invitrogen) was used to estimate molecular mass and the efficiency of electrotransfer. Immunodetection of the AOX protein was carried out using the BM chemiluminescence western-blot kit (Roche Diagnostics) according to the manufacturer's instructions. Polyclonal antibodies raised against the AOX1 protein of C. reinhardtii (a kind gift from Dr. S. Merchant, University of California, Los Angeles) was used at dilution of 1:54,000. Antibodies against the subunit III of yeast cytochrome c oxidase were obtained from Molecular Probes (Eugene, OR) and used at concentration of 1 µg mL¹.