“Labile” heme critically regulates mitochondrial biogenesis through the transcriptional co-activator Hap4p in Saccharomyces cerevisiae

Heme (iron protoporphyrin IX) is a well-known prosthetic group for enzymes involved in metabolic pathways such as oxygen transport and electron transfer through the mitochondrial respiratory chain. However, heme has also been shown to be an important regulatory molecule (as “labile” heme) for diverse processes such as translation, kinase activity, and transcription in mammals, yeast, and bacteria. Taking advantage of a yeast strain deficient for heme production that enabled controlled modulation and monitoring of labile heme levels, here we investigated the role of labile heme in the regulation of mitochondrial biogenesis. This process is regulated by the HAP complex in yeast. Using several biochemical assays along with EM and epifluorescence microscopy, to the best of our knowledge, we show for the first time that cellular labile heme is critical for the post-translational regulation of HAP complex activity, most likely through the stability of the transcriptional co-activator Hap4p. Consequently, we found that labile heme regulates mitochondrial biogenesis and cell growth. The findings of our work highlight a new mechanism in the regulation of mitochondrial biogenesis by cellular metabolites.

Mitochondria are organelles that have a critical role in the energy and intermediary metabolism in eukaryotic cells. They are well-known for their involvement in the synthesis of ATP, the currency in charge of fulfilling cell energy demand. Mitochondrial ATP synthesis mostly relies on the oxidative phosphorylation system (OXPHOS), 3 which involves enzymatic complexes that form the mitochondrial respiratory chain and the phosphorylation system. It is well-established that mitochondrial enzymatic content within cells varies to match ATP synthesis to ATP demand. This phenomenon has been observed in different species and experimental models, ranging from microorganisms, such as yeast, to mammals, and is associated with diverse pathologies (1)(2)(3)(4)(5)(6).
Although mitochondria possess their own genome (mtDNA), they are genetically semiautonomous. Indeed, the mtDNA encodes a few (8 in yeast and 13 in mammals) of the numerous proteins constituting the OXPHOS. All of the remaining OXPHOS components and mitochondrial proteins (protein machineries involved in mtDNA replication, transcription and translation, assembly factors, mitochondrial protein import, and intermediary metabolism) are encoded by the nuclear genome. Thus, regulation of the expression of both mitochondrial and nuclear genes encoding mitochondrial proteins is controlled by a defined network of nuclear transcription factors.
In the yeast Saccharomyces cerevisiae, the transcription factors Hap2p/3p/4p/5p (HAP complex) are the master regulators of mitochondrial biogenesis (7,8). This heteromeric complex relies on the binding to the promoter of target genes by the DNA-binding subcomplex Hap2p/Hap3p/Hap5p. These three subunits are constitutively expressed. The activation of transcription is mediated by Hap4p, the co-activator subunit (9 -13) and the only subunit whose expression is known to be regulated by the carbon source. Deletion of any of the subunits of the HAP complex impairs yeast growth on nonfermentable substrate, such as lactate, on which mitochondria are the only source of ATP production (11). In mammals, due to multicellular organization, more transcriptional regulators are involved, although the general mechanism is the same, with the well-known co-activator PGC-1␣ acting like a functional counterpart of the yeast Hap4p (14,15). Extensive work is still being carried out to identify the molecular regulators of the activity of PGC-1␣ and Hap4p. In this regard, it is interesting to point out recent results that highlight that both the yeast and mammalian co-activators share common regulatory signals. For example, it has been recently demonstrated that PGC-1␣ is regulated by the GSH redox state, as we previously showed for Hap4p (16,17). Previous studies also established the involvement of reactive oxygen species and cAMP-dependent signaling in negative and positive regulation of the transcriptional regulators of mitochondrial biogenesis in both yeast and mammals (5,(18)(19)(20)(21).
Heme (iron protoporphyrin IX) is well-known as a prosthetic group for enzymes involved in the metabolic pathway, such as oxygen transport and electron transfer through the mitochondrial respiratory chain. However, heme has also been shown to be an important regulatory molecule ("labile" heme) for diverse processes, such as translation, kinase activity, and transcription in mammals, yeast, and bacteria (22)(23)(24). Moreover, the hypothesis that heme can be a regulatory molecule for the HAP complex has been previously formulated (11,(25)(26)(27). However, no study has been able to establish molecular details surrounding heme regulation of the HAP complex, although a heme-dependent regulation of Hap4p transcription has been described recently (28 -32). Thus, although the name HAP historically stands for "heme activator protein," the HAP complex is sometimes designated like a heme-independent transcription factor (33). Taking advantage of a yeast strain deficient for heme production, we tackled this question. Our results are the first to show that cellular labile heme is critical for the HAP complex activity, most likely through an enhancement of the stability of the co-activator Hap4p. Consequently, labile heme regulates mitochondrial biogenesis and cell growth. This work represents a new contribution to the field of mitochondrial biogenesis regulation by cell metabolites.

Modulation of cellular labile heme pool
Our cell growth experiments were carried out on a nonfermentable substrate: lactate. In this condition, the mitochondrial compartment is well-differentiated, and mitochondrial oxidative phosphorylation processes and the HAP complex are mandatory for cell growth and proliferation. The heme biosynthesis pathway is localized within two distinct compartments: the mitochondria and the cytosol (Scheme 1). To modulate the cellular labile heme pool, we made use of a strain deleted for the first enzyme of this pathway: Hem1p. This enzyme, which is localized in the mitochondrial matrix, catalyzes the condensation of glycine with succinyl-CoA to generate ␦-aminolevulinate (ALA). A strain deleted for this enzyme is auxotrophic for ALA (35). In an attempt to modulate cellular labile heme, the ⌬hem1 strain was supplemented with increasing concentrations of ALA, and the cellular labile heme pool was assessed thanks to the reliable HMG2-lacZ reporter gene (35,36). In this system, expression of the ␤-gal is under the regulation of the HMG2 promoter, which is strongly repressed by the heme level. Fig. 1 shows that whereas this promoter's activity is low in WT cells, this activity is strongly increased in ⌬hem1 cells grown in the presence of low (5 and 10 g/ml) ALA concentrations. Increasing extracellular ALA concentrations lead to a decrease in this promoter's activity. This shows that modulation of extracellular ALA in a ⌬hem1 strain allows modulation of the cellular labile heme pool. Extracellular ALA concentration above 100 g/ml in both WT and ⌬hem1 cells did not increase the repression of HMG2-lacZ reporter gene (data not shown).

Modulation of mitochondrial compartment
As stipulated above, cells were grown on nonfermentable substrate, where growth is strictly dependent on mitochondrial activity (i.e. energy conversion processes (ATP synthesis) take place at the mitochondrial level). Further, mitochondrial oxidative phosphorylation requires heme biosynthesis to generate mitochondrial cytochromes. We thus investigated mitochondrial activities in ⌬hem1 cells supplemented with various concentrations of ALA. Fig. 2A shows that there is a 50% decrease in ⌬hem1 -ALA5-cellular respiratory rate compared with the WT Scheme 1. Overview of the heme synthesis pathway in yeast. Heme is synthesized from glycine and succinyl-CoA in a pathway initiated (Hem1, 5-aminolevulinic acid synthase) and terminated inside mitochondria (Hem14, protoporphyrinogen oxidase; Hem15, ferrochelatase) with a number of intermediary reactions taking place in the cytosol (Hem2, ALA dehydratase; Hem3, porphobilinogen deaminase; Hem4, uroporphyrinogen III synthase; Hem12, uroporphyrinogen III decarboxylase; Hem13, coproporphyrinogen III oxidase). Heme can then be complexed in mitochondrial or cytosolic hemoproteins or act in signaling pathways. Deuteroporphyrin IX is an analogue of protoporphyrin IX that is much more soluble and stable and can be converted to deuteroheme (an analogue of heme) by the ferrochelatase enzyme but cannot be incorporated into hemoproteins.

Labile heme regulates mitochondrial biogenesis
cells. Increasing extracellular ALA up to 100 g/ml allowed a full restoration of the cellular respiratory rate. Fig. 2B shows that this modulation of the cellular respiratory rate affected the cellular growth rate, and the relationship between these two parameters is linear (Fig. 2B, inset), as shown previously (3). In addition, Fig. 2C shows that the activity of cytochrome c oxidase (one of the most controlling enzymes of the oxidative phosphorylation system (37-39)) exhibits a 50% decrease in ⌬hem1 -ALA5-compared with the WT cells. Increasing extracellular ALA up to 100 g/ml allows a full restoration of this activity. Furthermore, cytochrome c oxidase activity perfectly correlates with respiratory rate (Fig. 2C, inset).
The above-mentioned results relate to the oxidative phosphorylation amount/activity. To determine whether we observed an overall quantitative regulation of the mitochondrial compartment, we assessed citrate synthase activity, a well-accepted marker of mitochondrial amount within cells (40 -42). Fig. 2D shows that there is a 60% decrease in ⌬hem1 -ALA5-citrate synthase activity compared with the WT cells. Increasing extracellular ALA up to 100 g/ml allows a full restoration of citrate synthase activity. The inset in Fig. 2D shows a very good correlation between respiratory rate and citrate synthase activity, reinforcing the hypothesis of an overall regulation of the mitochondrial compartment upon the addition of ALA in ⌬hem1 cells.
Within the cell, mitochondria are highly dynamic organelles that form a network. They undergo fusion and fission events continuously, leading to a diverse range of mitochondrial morphologies, from fragmented states to continuous networks (43)(44)(45). It remains unclear how the diverse morphologies interact with bioenergetic properties. To determine whether the mitochondrial structure/network was altered when mitochondrial oxidative phosphorylation activities decreased, we performed both fluorescence and electronic microscopy under our experimental conditions. Fig. 3A shows that when visualized in fluorescence microscopy, mitochondria exhibit a well-differentiated network, comparable with that of the WT for any of the extracellular ALA concentrations tested. Fig. 3B shows that when visualized in EM, mitochondria exhibit a coherent ultrastructure, comparable with that of the WT for any of the extracellular ALA concentration tested. However, for the lowest (⌬hem1 -ALA5-and ⌬hem1 -ALA10-) ALA concentrations, a decrease in mitochondrial diameter was observed. For higher ALA concentrations, the mitochondrial diameter was comparable with that of the WT.

Cellular labile heme pool regulates mitochondrial biogenesis
The amount of mitochondria within a cell is controlled by its turnover (i.e. the respective rates of mitochondrial biogenesis and mitochondrial degradation). The HAP complex has been shown to be involved in the specific induction of genes involved in gluconeogenesis, metabolism of alternate carbon sources, respiration, and mitochondrial development. The disruption of any subunit of this complex renders the cells unable to grow on nonfermentable carbon sources (10,12,46,47). Moreover, many genes involved in energy metabolism have been shown to be regulated by this complex (8,48). We thus assessed the activity of the HAP complex with a widely used reporter gene, pCYC1-lacZ (49). Fig. 4A shows that under low cellular labile heme condition (⌬hem1 -ALA5-), the activity of the HAP complex is highly decreased, and this activity increases proportionally with the amount of cellular labile heme, being slightly higher than in the WT for the ⌬hem1 -ALA100-condition. The master regulator of the activity of the HAP multicomplex is the subunit Hap4p (11). Cellular amounts of Hap4p were thus assessed under our experimental conditions. Fig. 4B shows that the amount of Hap4p is highly increased proportionally with the amount of cellular labile heme, being higher than in the WT for the ⌬hem1 -ALA100-condition.

Hap1p is not involved in labile heme-induced regulation of mitochondrial biogenesis through Hap4p
A number of studies have established that Hap1p is the heme sensor that controls the expression of a number of genes encoding heme-and oxygen-dependent processes (34). In addition, Hap1p controls the expression of Hap4p (32), which is the activation domain of the Hap2/3/4/5 transcription complex (HAP complex). To assess whether Hap1p could play a role in hemeinduced regulation of mitochondrial biogenesis through Hap4p, we generated a ⌬hap1 and a ⌬hap1⌬hem1 mutant in which most of the experiments mentioned above were repeated. As shown in Fig. 5, Hap1p deletion did not modify labile heme titration by ALA in the absence of Hem1p (A), the increase in respiratory rate induced by an increase in labile heme (B), the relationship between growth rate and respiratory rate (C), the increase in cytochrome oxidase activity induced by an increase in labile heme (D), the increase in HAP complex activity induced by an increase in labile heme (E), and the increase in Hap4p induced by an increase in labile heme (F). These results clearly show that Hap1p is not involved in the labile heme-induced regulation of mitochondrial biogenesis through Hap4p.

Labile heme regulates mitochondrial biogenesis Hemin interacts with Hap4p
As shown above, intracellular Hap4p amount is strongly correlated to the cellular labile heme amount. This led us to formulate the hypothesis of an eventual interaction between this protein and cellular labile heme. GSH S-transferase (GST)-Hap4p(330 -554) was thus purified, and in vitro interaction between this Hap4p fragment and hemin was tested through spectrophotometric analysis. We focused on the fragment Figure 2. Effect of ALA addition on respiratory rates, growth, and cytochrome c oxidase and citrate synthase activities in WT and ⌬hem1 cells. WT and ⌬hem1 cells were grown aerobically in the presence of increasing concentrations of ALA for the latter as described in the legend to Fig. 1. Each culture was subjected to several analyses during exponential phase, all repeated five times over a 4-h period. The experiments were reproduced at least three times with independent cultures. A, spontaneous respiratory rates measured on cultures in the different conditions. The given values for each condition are means of all data points Ϯ S.D. (error bars). B, growth rate measured on cultures in the different conditions. The given values for each condition are means of all data points Ϯ S.D. Inset, a linear regression line between respiratory rate and growth rate is presented for which the determination coefficient is indicated (r 2 ). Growth rates for each condition were calculated in each experiment from measured OD 600 nm of cultures over a 4-h period. C, cytochrome c oxidase activities measured on cultures in the different conditions are shown. The given values for each condition are means of all data points Ϯ S.D. Inset, a linear regression line between respiratory rate and cytochrome c oxidase activity is presented for which the determination coefficient is indicated (r 2 ). D, citrate synthase activities measured on cultures in the different conditions. Inset, a linear regression line between respiratory rate and citrate synthase activity is presented for which the determination coefficient is indicated (r 2 ). The given values for each condition are means of all data points Ϯ S.D. **, p Ͻ 0.01; ***, p Ͻ 0.001; ****, p Ͻ 0.0001.

Labile heme regulates mitochondrial biogenesis
330 -554 because it contains the only triplet of histidine that can allow an interaction between Hap4p/heme. This fragment and its relevance have already been described (54). Further, to this day, none of the groups working on Hap4p have been able to purify a full-length protein. It has been shown that the hemin spectrum is strongly modified upon its interaction with a protein (50). Fig. 6A shows that GST-Hap4p(330 -554) does not exhibit any absorbance spectra, whereas hemin exhibits a strong absorbance signal around 390 nm, characteristic of the Soret band. Adding GST-Hap4p(330 -554) (4 M) to hemin (4 M) at a stoichiometric concentration induces a shift in the hemin maximal absorbance signal. Further increasing the GST-Hap4p(330 -554) concentration (up to 8 M) did not modify the GST-Hap4p(330 -554)-hemin absorbance spectrum. Fig.  6A (inset) shows the differential spectrum between hemin and GST-Hap4p(330 -554)-hemin. Furthermore, Fig. 6B shows that GST alone does not modify the hemin absorbance spectrum. Fig. 6B (inset) shows the differential spectrum between hemin and hemin in the presence of GST spectra. These results strongly indicate an interaction between hemin and GST-Hap4p(330 -554).
To determine whether the full-length Hap4p is also able to interact with hemin, we performed co-precipitation of Hap4p on whole-cell lysates using hemin-agarose beads. Fig. 6C clearly shows that even in different amounts, hemin and Hap4p (from both WT and ⌬hem1 -ALA100-) co-precipitate, strongly reinforcing the possibility of an interaction between hemin and Hap4p. The specificity of this interaction was further confirmed by performing a competition experiment between hemin and heminagarose. When cell extracts were preincubated in the presence of hemin, there was no detectable interaction between Hap4p and hemin-agarose (Fig. 6D). To further characterize this inter-action in vitro, a difference spectroscopy titration of hemin binding to Hap4p was performed, and an apparent K d of about 1.4 M was measured (Fig. 6E).

Cellular labile heme pool regulates Hap4p stability
Our results show not only that cellular labile heme amount correlates with the amount of Hap4p, but also that Hap4p and hemin can interact in vitro. Based on these results, we hypothesized that cellular labile heme could play a role in regulating the amount of Hap4p through a stabilization of this protein.
Consequently, Hap4p turnover was assessed through cycloheximide addition to cells under conditions where cellular labile heme amount is modulated (see Fig. 2). Fig. 7 clearly shows that increasing cellular labile heme increases Hap4p half-life. Indeed, with the lowest concentrations of ALA (⌬hem1 -ALA5and ⌬hem1 -ALA10-), Hap4p underwent a decrease of 60% in 3 min and 70% in 7 min, whereas with high concentrations of ALA (⌬hem1 -ALA25-and ⌬hem1 -ALA100-), Hap4p is only reduced by 30% in 3 min and 50% after 7 min, similarly to the WT, with a low variability between the experiments (Table 1).

A nonmetabolizable analogue of heme up-regulates mitochondrial biogenesis
The experiments stated above point to a regulation of mitochondrial biogenesis by cellular labile heme. However, in these experiments, labile heme levels were manipulated through ALA addition to a ⌬hem1 strain. ALA is the second metabolite of the heme biosynthesis pathway and can be incorporated in hemoproteins. To strengthen our observations, we decided to reproduce our experiments with deuteroporphyrin IX (DP IX), a so-called gratuitous inducer (51). DP IX is a more soluble analog of protoporphyrin IX, which is the penultimate product of the heme

Labile heme regulates mitochondrial biogenesis
biosynthesis pathway (Scheme 1) and can be maturated by the ferrochelatase. Hence, DP IX is another way to increase intracellular heme level and to emphasize the involvement of the heme molecule in regulatory processes. However, once maturated by ferrochelatase, this molecule cannot be incorporated into hemoproteins (51). Consequently, a ⌬hem1 strain cannot grow with DP IX alone (data not shown), and experiments performed with DP IX were done in the presence of the lowest concentration of ALA used all throughout our experiments. Fig. 8A shows that DP IX supplementation of ⌬hem1 -ALA5-cells restores cellular labile heme to the WT level. Fig. 8B shows that DP IX supplementation of ⌬hem1 -ALA5-cells strongly increases the respiratory rate in such a way that it reaches 90% of the WT respi-ratory rate. Growth and cytochrome c oxidase activity are partially restored by DP IX supplementation of ⌬hem1 -ALA5-cells (Fig. 8, C  and D). Fig. 8E shows that DP IX supplementation of ⌬hem1 -ALA5cells restores citrate synthase activity to the WT level.
We then assessed whether DP IX was able to regulate HAP complex activity. Fig. 9A shows that DP IX supplementation of ⌬hem1 -ALA5-cells significantly increases HAP complex activity. Further, in these conditions, Hap4p levels are similar to the ones assessed in the WT cells (Fig. 9B). Last, DP IX supplementation of ⌬hem1 -ALA5-cells strongly decreases Hap4p turnover (Fig. 9C). Taken together, the results obtained by DP IX supplementation of ⌬hem1 -ALA5-cells clearly show that cellular labile heme regulates mitochondrial biogenesis.  The wavelengths of major peaks are indicated above them. B, spectra of hemin in the presence or absence of the indicated concentrations of purified recombinant GST. Inset, difference spectrum between hemin with GST (ϩGST, 4 M) and hemin alone (Hemin, 4 M). C, co-precipitation of Hap4p in hemin pulldown experiments. WT or ⌬hem1 cells were cultured aerobically in the absence or in the presence of 100 g⅐ml Ϫ1 of ALA, respectively. Cell extracts were subjected to a co-precipitation assay with hemin-agarose beads and analyzed by SDS-PAGE and Western blotting using antibodies directed against Hap4p or PGK1. A negative control experiment was also performed with WT cell extracts and unconjugated agarose beads (WT ϩ agarose beads). Representative Western blotting results of at least three replicates are shown. D, co-precipitation of Hap4p in hemin pulldown experiments. Cell extracts were preincubated with 100 g⅐ml Ϫ1 hemin for 2 h and then subjected to a co-precipitation assay with hemin-agarose beads and analyzed by SDS-PAGE and Western blotting using antibodies directed against Hap4p. Representative Western blotting results of at least three replicates are shown. E, difference absorption spectra and titration of GST-Hap4p(330 -554) at 100 nM with increasing concentrations of hemin (from 0 to 15 M). The curves were generated from fits to an equation describing a single binding site (Y ϭ B max ⅐X/(K d ϩ X)) with GraphPad Prism. Error bars, S.D.

Discussion
In this paper, making use of a strain deleted for the first enzyme of the heme biosynthesis pathway, we were able to modulate the intracellular labile heme concentration. On nonfermentable growth substrate, this labile heme concentration modulation positively regulates cellular mitochondrial content. We investigated the origin of this regulation and were able to show that intracellular labile heme regulates mitochondrial biogenesis. Indeed, the activity of the master regulator of this process, the HAP complex, positively correlates with the cellu-lar labile heme concentration. Moreover, both the amount and the stability of the co-activator of this complex, Hap4p, positively correlate with the cellular labile heme concentration. Further, we were able to show that hemin and Hap4p interact directly in vitro as well as by co-precipitation. In vitro experiments allowed us to assess an apparent K d of about 1.4 M. Although that K d seems quite weak, especially relative to current estimates for the labile heme concentration in various nonerythroid cell lines, 25-300 nM, there may be active processes that increase the flux of heme into the nucleus to activate

Table 1 Mean percentages of Hap4p protein stability assay
For each condition and each time, mean percentages Ϯ S.D. of T0 (shown in Fig. 7) are reported; data were extracted from Fig. 7.

Labile heme regulates mitochondrial biogenesis
Hap4p. For instance, peroxide and NO are two potential signals that can labilize heme to activate Hap4p. Another possibility is that the Hap4p-heme complex stability has been underestimated because heme-binding measurements were done on a peptide fragment of Hap4p and not the full-length protein, which is technically challenging to purify. Even a modest 3-fold boost in heme-binding affinity, which may be accounted for in the context of full-length Hap4p, would place the heme affinity of Hap4p as more similar to that of Mss51, a heme-sensing translational activator with K d ϳ300 nM. It is tempting to speculate that the Hap4p-labile heme interaction allows an increase in Hap4p stability, which itself will increase mitochondrial biogenesis.
It should be pointed out here that despite an important modulation of mitochondrial activities in our experimental conditions, no alteration of the mitochondrial network was assessed, which shows that there is no simple correlation between cellular mitochondrial activities and the mitochondrial network.
It has previously been shown that in WT cells, heme is a limiting factor for Hap4p transcription (32). However, this is only the case in glucose-grown cells, where heme intracellular concentration is increased by supplying the cells with either ALA or hemin. In galactose-grown cells, a situation closer to our study in nonfermentable medium (no glucose repression of oxidative metabolism), neither heme nor ALA (at a concentration 3 times higher than the one used in our study) regulated Hap4p transcription (32). Such a regulation is thus unlikely under our experimental conditions. Further, our assessment of Hap4p halflife clearly shows that this parameter is positively regulated by labile heme. To our knowledge, our results are the first ones evidencing a post-translational regulation of the HAP complex.
A crucial point in our study is the link between heme and mitochondrial oxidative phosphorylation. Indeed, heme is

Labile heme regulates mitochondrial biogenesis
mandatory for mitochondrial biogenesis and the generation of mitochondrial cytochromes. Consequently, to demonstrate the regulation of mitochondrial biogenesis by cellular labile heme implies a distinction between heme that will lead to the generation of cytochromes and labile heme that is involved in cell signaling. To do so, we made use of deuteroporphyrin IX, a gratuitous inducer analogous to labile heme (51). This allowed us to show the following. (i) Even at low ALA concentration, an increase in deuteroheme-thanks to DP IX-allows an increase in mitochondrial biogenesis. This indicates that low ALA concentration is not controlling mitochondrial biogenesis through cytochrome generation. (ii) An increase in deuteroheme increases mitochondrial biogenesis through an increase in HAP activity and Hap4p stability.
In conclusion, we show that in the yeast S. cerevisiae, mitochondrial biogenesis is positively regulated by labile heme, via an increase of the HAP4 protein's stability, most likely through its binding to labile heme (Scheme 2).

Oxygen consumption assays
The oxygen consumption was measured polarographically at 28°C using a Clark oxygen electrode in a 1-ml thermostatically controlled chamber. 1 ml of culture was transferred to the chamber, and respiratory rates were determined from the slope

␤-Gal activity measurement
Cells were harvested from cultures at OD 600 nm of 0.5-1 and permeabilized using a standard procedure as described (57). After a 30-min preincubation period, 0.4 mg/ml O-nitrophenyl-␤-D-galactopyranoside was added, and the tube was briefly vortexed. The reaction was stopped after a yellow color had developed by the addition of 0.5 M Na 2 CO 3 . The samples were centrifuged for 30 s at 14,000 ϫ g, and the absorbance of the supernatant was read at 420 nm. ␤-Gal activity is expressed in arbitrary units per OD 600 nm and per minute (57).

Protein extraction, electrophoresis, and Western blotting
Cells were lysed using a mixture of 7.5% ␤-mercaptoethanol in 1.85 M NaOH. After a 10-min incubation on ice, proteins were precipitated by the addition of an equal volume of 3 M TCA for 10 min on ice. After a rapid centrifugation at 4°C, the protein pellet was suspended in a mixture of 10% SDS and sample buffer (0.06 M Tris, 2% SDS, 2% ␤-mercaptoethanol, 5% glycerol, 0.02% bromphenol blue). Protein amounts corresponding to 0.5 OD units of cells were separated by 10% SDS-PAGE performed according to the method of Laemmli. After electro-transfer onto nitrocellulose membranes (Amersham Biosciences), proteins were probed with the desired primary antibodies: ␣-Hap4p and ␣-PGK1 (mAb, Invitrogen) and detected using peroxidase-conjugated secondary antibodies (Jackson ImmunoResearch) and ECL Prime reagent (Amersham Biosciences), according to the manufacturers' instructions. Signal quantifications were done using the ImageJ software.

Protein degradation assay
To quantify Hap4p life span, 1 ml of each cell culture was centrifuged at 4°C, corresponding to T0. Then 0.4 mg/ml cycloheximide was added to the cultures. After 3 and 7 min (T3 and T7), 1 ml of the cultures was centrifuged as previously. The dry pellets were conserved at Ϫ20°C until protein extraction and Western blotting, as described above.

Citrate synthase activity
Cells were lysed in 50 mM Tris-HCl (pH 7.5) buffer and glass beads with five cycles (20 s at 4 m/s) of FastPrep TM . Citrate synthase (EC 4.1.3.7) activity was determined spectrophotometrically (Safas Monaco) by monitoring at 412 nm the oxidation of CoA (produced by citrate synthase activity) by 5,5Ј-dithiobis(2nitrobenzoic acid) (DTNB) over time in the following buffer: 50 mM Tris-HCl (pH 7.5), 0.1 mM acetyl-CoA, 0.2 mM DTNB, and 0.5 mM oxaloacetate. The enzyme activity was calculated using an extinction coefficient of 13,600 M Ϫ1 ⅐cm Ϫ1 at 412 nm for DTNB. One citrate synthase unit was equal to 1 mol of DTNB reduced per minute per mg dry weight.

Cytochrome c oxidase activity
Cells were lysed in 50 mM potassium phosphate (pH 7.5) buffer and glass beads with five cycles (20 s at 4 m/s) of Fast-Prep TM . Cytochrome c oxidase activity (1 mM potassium cyanide-sensitive) was determined by monitoring spectrophotometrically (550 nm) the rate of disappearance of reduced cytochrome c at 28°C in the following buffer: 50 mM P i K, 100 M reduced cytochrome c. The enzyme activity was calculated using an extinction coefficient of 18,500 M Ϫ1 ⅐cm Ϫ1 at 550 nm for reduced cytochrome c (52, 58).

Scheme 2. Heme-induced regulation of mitochondrial biogenesis in the yeast S. cerevisiae.
A, in the presence of low concentrations of heme or deuteroheme, the HAP4 turnover rate is high, and mitochondrial biogenesis is low. B, an increase in heme of deuteroheme concentration within the cell stabilizes HAP4, and the mitochondrial biogenesis rate is higher.

Labile heme regulates mitochondrial biogenesis Hemin absorption spectra
Recombinant GST and GST fused to the C-terminal region of Hap4p (residues 330 -554, containing the transcriptional activation domain) were purified from bacterial lysate after a 3-h induction with 1 mM isopropyl 1-thio-␤-D-galactopyranoside. Bacterial cells were washed once in the following buffer: 25 mM HEPES, 50 mM KCl, 10% glycerol, 0.5 mM EDTA, 10 mM ␤-mercaptoethanol, 1 mM phenylmethylsulfonyl fluoride, pH 7.7. After three freeze-thaw cycles, cells were sonicated on ice and centrifuged for 15 min at 16,000 ϫ g. Purification was performed on the supernatant using GST-Sepharose 4B resin (GE healthcare) using the manufacturer's recommendations. The spectrum of hemin (Sigma-Aldrich) was recorded in the absence and presence of different concentrations of the recombinant proteins in potassium phosphate buffer (0.1 M; pH 7) using a spectrophotometer (Safas).
Heme titrations were performed by difference spectroscopy at room temperature and aerobically using the Safas spectrophotometer. Heme working stock solution was 1 mM in 0.1 M NaOH. A concentration of 100 nM of the recombinant GST-Hap4p was used in potassium phosphate buffer (0.1 M; pH 7). After the baseline was set (Hap4p 100 nM against buffer), equal concentrations of hemin were added to the reference and sample cuvette (0.025-15 M), and the cuvettes were incubated for 10 min to reflect the binding of recombinant protein and hemin and not the absorbance of the free ligand. To determine the binding parameters, the data obtained from the heme titrations were plotted and fit to an equation describing a single binding site using the GraphPad Prism software.

Hemin co-precipitation assay
Cells were lysed in the following buffer: 50 mM Tris, 100 mM NaCl, 5 mM EDTA, 5% glycerol, 1 mM phenylmethylsulfonyl fluoride, pH 7.5 (IP buffer) and an anti-protease mixture (cOmplete, Roche Applied Science), with four cycles (20 s at 4 m/s) of FastPrep TM . For the co-precipitation, 50 l of 50% (v/v) hemin-agarose or agarose beads (Sigma) were added to 500 l of supernatant, corresponding to 80 OD units of cells, and incubated for 2 h at 4°C. Beads were then washed with the IP buffer, and proteins were eluted with a mixture of 10% SDS and sample buffer. Inputs were samples without incubation with heminagarose beads and precipitated in 35% TCA. Protein pellets were resuspended in a mixture of 10% SDS and sample buffer. After quantification (Bio-Rad), proteins were separated by SDS-PAGE using precast Nupage 4 -12% Bis-Tris Plus Gels (Invitrogen).

Epifluorescence microscopy
For each condition, 900 l of cell cultures were added to 100 l of 37% paraformaldehyde. After 8 min of room temperature incubation, samples were pelleted and washed with 1 ml of PBS, and pellets were resuspended in PBS. Cells were observed in a fully automated inverted microscope (Olympus) using a ϫ100, 1.4 numerical aperture Plan-Apochromat objective. Stacks of fluorescence images were collected automatically at 0.2-m Z-intervals. The deconvolution of each Z-series was automatically computed with two plugins of ImageJ: PSF generator and DeconvolutionLab. To obtain a three-dimensional image, the deconvolutions were treated with the software Chimera (University of California, San Francisco).

EM microscopy
The yeast pellets were placed on the surface of a copper EM grid (400 mesh) that had been coated with Formvar. Each loop was very quickly submersed in liquid propane precooled and held at Ϫ180°C by liquid nitrogen. The loops were then transferred in a precooled solution of 4% osmium tetroxide in dry acetone in a 1.8-ml polypropylene vial at Ϫ82°C for 72 h (substitution) and warmed gradually to room temperature, followed by three washes in dry acetone. Specimens were stained for 1 h in 1% uranyl acetate in acetone at 4°C in a darkroom. After another rinse in dry acetone, the loops were infiltrated progressively with araldite (epoxy resin, Fluka). Ultrathin sections were contrasted with lead citrate and observed with a Hitachi 7650 electron microscope (Bordeaux Imaging Center, Electron Microscopy Pole of the University of Bordeaux Segalen).

Statistical analysis
Results are expressed as mean Ϯ S.D. Statistical analysis was carried out using analysis of variance for all results. Prism software (GraphPad, San Diego, CA) was used for all tests. A p value of less than 0.05 was considered significant.