Yeast flavohemoglobin, a nitric oxide oxidoreductase, is located in both the cytosol and the mitochondrial matrix: effects of respiration, anoxia, and the mitochondrial genome on its intracellular level and distribution.

Yeast flavohemoglobin, YHb, encoded by the nuclear gene YHB1, has been implicated in both the oxidative and nitrosative stress responses in Saccharomyces cerevisiae. Previous studies have shown that the expression of YHB1 is optimal under normoxic or hyperoxic conditions, yet respiring yeast cells have low levels of reduced YHb pigment as detected by carbon monoxide (CO) photolysis difference spectroscopy of glucose-reduced cells. Here, we have addressed this apparent discrepancy by determining the intracellular location of the YHb protein and analyzing the relationships between respiration, YHb level, and intracellular location. We have found that although intact respiration-proficient cells lack a YHb CO spectral signature, cell extracts from these cells have both a YHb CO spectral signature and nitric oxide (NO) consuming activity. This suggests either that YHb cannot be reduced in vivo or that YHb heme is maintained in an oxidized state in respiring cells. By using an anti-YHb antibody and CO difference spectroscopy and by measuring NO consumption, we have found that YHb localizes to two distinct intracellular compartments in respiring cells, the mitochondrial matrix and the cytosol. Moreover, we have found that the distribution of YHb between these two compartments is affected by the presence or absence of oxygen and by the mitochondrial genome. The findings suggest that YHb functions in oxidative stress indirectly by consuming NO, which inhibits mitochondrial respiration and leads to enhanced production of reactive oxygen species, and that cells can regulate intracellular distribution of YHb in accordance with this function.

Indeed, whereas most vertebrate hemoglobins consist of two types of polypeptide subunits that have single heme domains and form tetrameric oligomers, microbial, invertebrate, and protozoan hemoglobins are far more varied. Bacterial, fungal, and protozoan hemoglobins fall into two general types: (i) dimeric hemoproteins composed of two single heme domain polypeptides; and (ii) monomeric flavohemoproteins containing a single heme binding domain at the amino terminus and a carboxyl-terminal FAD-binding domain that is related to those found in proteins in the ferredoxin-NADPϩ reductase family (3). From DNA sequence analysis, these flavohemoglobins appear to form a distinct subgroup within the hemoglobin family (1,4).
The role of vertebrate and invertebrate hemoglobins in the transport of O 2 and CO 2 has been known for some time. Recently, it has been discovered that the hemoglobins of both eucaryotes and prokaryotes have additional functions relating to nitric oxide (NO). 1 For example, in mammals hemoglobin functions to transport NO to effectors that regulate blood flow and, therefore, the rate of delivery of oxygen to tissue cells (5)(6)(7)(8). In contrast, the function of the Escherichia coli hemoglobin HMP is to protect cells against nitrosative stress brought about by NO or NO-donating S-nitrosothiols (9 -11). This hemoglobin functions in both aerobic and anaerobic cells.
In the presence of air it functions as an NO oxygenase that generates mainly nitrate, and in the absence of air it functions as an NO reductase that generates nitrous oxide. HMP may also play a role in oxidative stress (9). The flavohemoglobin of Ralsonia eutropha is involved in denitrification with the production of N 2 O (12). The genes that encode the flavohemoglobins of R. eutropha, and Bacillus subtilis are up-regulated under oxygen-limiting conditions (13,14). Together, these findings suggest a multiplicity of roles for hemoglobins.
The physiological role(s) of the flavohemoglobin found in Saccharomyces cerevisiae cells is (are) still somewhat unclear. The expression of its gene, YHB1, is induced by oxygen, which is the opposite of the effect oxygen has on the expression of most bacterial hemoglobin genes. This observation has been taken to indicate that the hemoglobins of yeast and bacteria may have different functions (15). Several observations suggest physiological connections between the expression/function of this protein, mitochondrial respiration, and oxidative and nitrosative stress. First, respiring S. cerevisiae cells have low levels of reduced YHb pigment as detected by carbon monoxide (CO) photolysis difference spectroscopy (16). The intracellular levels of a reduced YHb CO photoproduct pigment are in-creased in cells in which the mitochondrial respiratory chain has been compromised by either mutation (16) or respiration inhibitors (17,18). Its level also increases in cells expressing the hypoxic isoform Vb of cytochrome c oxidase subunit V under aerobic conditions (16). Second, yeast cells carrying a deletion in YHB1, the structural gene for YHb, become sensitive to some conditions that promote oxidative (19) and other forms of stress (15). It has also been reported that some conditions promoting oxidative stress increase the expression of YHB1 (19), but YHB1 expression under some of these conditions has been questioned (15). Finally, it has been reported recently that YHb metabolizes NO and functions as an NO oxygenase under aerobic conditions and an NO reductase under anaerobic conditions (4).
To better understand the physiological function of YHb, we have examined here the relationship between respiration and the level of assembled YHb, established the intracellular location of YHb, and looked at the effects of oxygen, respiration, oxidative stress, and the mitochondrial genome on intracellular YHb localization.
Preparation of Whole Cell Lysates-Whole cell extracts were prepared from cells grown to mid-exponential phase in YPGal or YPD. Cells were harvested by centrifugation (5,000 ϫ g for 10 min), washed twice with ice-cold distilled water, and resuspended at 0.5 g (wet weight)/ml in 0.5 ml of radioimmune precipitation assay buffer (0.2 M NaCl, 0.125 M NaPO4, pH 7, 0.0125% (v/v) Nonidet P-40, 0.125% SDS, and 0.03 M sodium deoxycholate) in a conical microfuge tube containing 1 g of glass beads (0.25-0.5-mm diameter). Cells were broken by vortexing in two 1-min intervals separated by 1 min on ice. Cell debris was removed by centrifugation at 5000 ϫ g max in a Beckman Microfuge 18. The supernatant was saved and frozen at Ϫ80°C until ready for use.
Isolation of Mitochondrial and Cytosol Fractions-For the preparation of mitochondrial and cytosolic fractions, cells were grown to midexponential phase, harvested, and spheroplasted as described (24). All subsequent steps were performed at 4°C. Spheroplasts were harvested by centrifugation (5 min at 3,000 ϫ g), washed gently in post-spheroplast buffer (1.5 M sorbitol, 1 mM Na 2 EDTA, and 0.1% bovine serum albumin, pH 7), and sedimented at 3,000 ϫ g for 5 min. Washed spheroplasts were resuspended in lysis buffer (0.6 M mannitol, 2 mM Na 2 EDTA, and 0.1% bovine serum albumin, pH 7.4), lysed in a Sorvall Omnimixer (Newton, CT) at low speed for 3 s and full speed for 45 s, and then centrifuged for 5 min at 1,900 ϫ g to pellet unbroken cells, nuclei, and debris. The resulting supernatant was decanted and centrifuged for 10 min at 12,100 ϫ g to pellet mitochondria. The mitochondrial pellet was washed by resuspension in mitochondrial lysis buffer minus bovine serum albumin (pH 7), homogenized with a glass/Teflon homogenizer, and centrifuged at 1,651 ϫ g for 5 min. The resulting supernatant was decanted and centrifuged at 23,500 ϫ g for 10 min to pellet the mitochondria. The post-mitochondrial supernatant, collected after the 23,500 ϫ g centrifugation, was used as the cytosolic fraction. Protein in the mitochondrial fractions from aerobically grown JM43 cells, anaerobically grown JM43 cells, aerobically grown JM43 0 cells, and aerobi-cally grown JM43GD5ab cells accounts for 9. 8, 9, 8.5, and 9.8%, respectively, of the combined protein in mitochondrial and cytosolic fractions.
Subfractionation of Mitochondria-Mitochondrial subfractions were prepared by a modification of the digitonin fractionation procedure described (25). In this procedure, freshly isolated mitochondria are converted to mitoplasts, during which the outer membrane and intermembrane space fractions are removed. The mitoplasts are then disrupted, and the inner membrane and matrix fractions are separated. To prepare mitoplasts, mitochondria were suspended at a protein concentration of 10 mg/ml mitoplast suspension buffer (0.6 M mannitol and 10 mM NaPO 4 , pH 7), and a 1-ml portion of the suspension was treated for 1 min on ice with 0.35 mg of digitonin per milligram of mitochondrial protein. All subsequent steps were performed at 4°C. The suspension was diluted immediately with 5 volumes of mitoplast suspension buffer and centrifuged for 10 min at 12,000 ϫ g max . The pellet was the mitoplast fraction. The supernatant was carefully separated from the pellet and centrifuged for 60 min at 144,000 ϫ g max . The resulting pellet and the supernatant were used as the outer mitochondrial membrane and the intermembrane space fractions, respectively. The inner membrane and matrix fractions were prepared from the mitoplasts collected above as follows. Mitoplasts were resuspended at 10 mg of protein per milliliter in 100 mM NaPO 4 , pH 7, and sonicated for 20 s at 50 watts with a Branson Sonifier (Model W 185) equipped with a microtip. Sonicated mitoplasts were centrifuged for 20 min at 95,000 ϫ g max . The pellet was the inner membrane fraction, and the supernatant was the matrix fraction. The pellet was washed by resuspension in 100 mM NaPO 4 , pH 7, and recovered by centrifugation at 95,000 ϫ g max for 20 min.
Partial Purification of Cytosolic YHb-A freshly isolated cytosolic fraction was subjected a series of ammonium sulfate fractionations, as modified from the work of Mok et al. (26). Briefly, ammonium sulfate was added to 30% saturation, and the sample was incubated at 4°C for 15 min with stirring and then centrifuged at 27,000 ϫ g max for 15 min at 4°C. The pellet was suspended in 100 mM NaPO 4 , pH 7, and saved. The supernatant was adjusted to 45% ammonium sulfate, and the suspension was incubated at 4°C for 15 min with stirring and then centrifuged at 27,000 ϫ g max for 15 min at 4°C. The pellet was suspended in 100 mM NaPO 4 , pH 7, and saved. The supernatant was adjusted to 65% ammonium sulfate, and the suspension was incubated at 4°C for 15 min with stirring and then centrifuged at 27,000 ϫ g max for 15 min at 4°C. The pellet was suspended in 100 mM NaPO 4 , pH 7, and saved. The supernatant was adjusted to 95% saturation, incubated at 4°C for 15 min with stirring, and then centrifuged at 27,000 ϫ g max for 15 min at 4°C. The pellet was suspended in 100 mM NaPO4, pH 7, and saved. The remaining supernatant was saved. All pellets were resuspended to a volume of 3 ml per 100 ml of starting volume. All fractions were stored at 4°C until spectra were recorded and then frozen at Ϫ70°C for immunoblot analysis.
Difference Spectroscopy-Whole cell suspensions were prepared for CO photolysis spectra as follows. A concentrated cell suspension (0.45 g (wet weight)/ml) was adjusted to 1% (w/v) glucose in a sealed small test tube and incubated without agitation for 5 min at room temperature to achieve anaerobiosis. The suspension of reduced cells was then adjusted to 30% ethylene glycol and bubbled with CO for 10 min prior to transfer to a cuvette. The cuvette was frozen in liquid nitrogen, and CO photolysis difference spectra on the frozen samples were recorded as described (16). Reduced CO-ligated difference spectra of cell fractions were recorded at room temperature as follows. Each cell fraction was reduced with sodium dithionite and a spectrum was recorded. Each fraction was then bubbled slowly with CO gas (99.5%) for 5 min. The visible/Soret spectra of the same sample were recorded before and after exposure to CO. The baseline spectrum, obtained from the difference between the two spectra of the same reduced sample, was subtracted from the reduced minus CO-ligated difference spectra. All spectra were recorded using an SLM AMINCO DW2000 dual wavelength scanning, UV lightvisible spectrophotometer (SLM Instruments, Inc.) with the following settings: dual beam, wavelength acquisition, filter slow, 0.8-nm slit width, recording monochrometer (monochrometer 2) from 39 to 700 nm, and the reference monochromator (monochrometer 1) set to a wavelength of 577 nm.
Preparation of Antibodies-Polyclonal antibodies to YHb, PET100p, and cytochrome c oxidase subunit IV were prepared. For YHb, an antibody was obtained against a peptide synthesized to correspond to amino acid residues 379 -397 (17); for PET100p, an antibody was obtained against a peptide synthesized to correspond to residues 95-111 (27). The peptides were synthesized with an amino-terminal cysteine and coupled through this cysteine to maleimide-activated keyhole limpet hemacyanin (KLH). The peptide-KLH conjugate was mixed with Freund's complete adjuvant, and 1 mg of conjugate was injected sub-cutaneously into New Zealand White rabbits. Booster injections (0.5 mg) were given with Freund's incomplete adjuvant at 2 and 8 weeks. Antiserum was collected at 2-week intervals, with the terminal collection at 12 days after the final boost. Antibodies to HPLC purified cytochrome c oxidase subunit IV were prepared as described (28). Monoclonal antibodies to yeast phosphoglycerate kinase (and porin) were obtained from Molecular Probes, Inc. (Eugene, OR).
Measurement of NO Consumption-NO consumption was measured using a Clark-type NO electrode (World Precision Instruments, Sarasota, FL). All measurements were done in triplicate in a water-jacketed chamber with stirring at 30°C. The electrode was calibrated with a 50 M solution of KNO 2 to generate a series of known concentrations of NO (100, 200, and 400 nM) in 10 ml of a 0.1 M potassium iodide and 0.1 M H 2 SO 4 solution. NO consumption of whole cell extracts or mitochondrial or cytosolic fractions was then measured in 2 ml of phosphate-buffered saline (80 mM Na 2 HPO 4 , 20 mM NaH 2 PO 4 , and 100 mM NaCl), 250 M NADH, and 0.1 mM EDTA after the addition of 2 M NO from an NO-saturated solution of distilled water. Preparation of the NO-saturated solution was as described by the manufacturer of the NO electrode. Using a buffer control, it was possible to determine the amount of NO injected into the chamber as well as its stability in the absence of cell extract.
Northern Blotting-For Northern analysis, RNA samples (30 g) were separated on 1.5% agarose gels containing 0.22 M formaldehyde (29), transferred to Schleicher & Schuell Nytran membranes, and hybridized according to the manufacturer's suggestions. DNA probes were prepared by random primer labeling of double-stranded DNA fragments using [␣-32 P]dCTP. Probes were a 600-bp SspI fragment for YHB1 and a 500-bp StyI fragment for ACT1. Blots were quantitated with an Ambis radio-analytic imaging system.
SDS-PAGE and Immunoblot Analysis-SDS-PAGE was performed either on 16% SDS-polyacrylamide gels containing 10% glycerol and 3.6 M urea (24) or 10% SDS-polyacrylamide gels (30) as indicated in the legends to Figs. 2-4, 6, and 8. Following electrophoresis, proteins were electroblotted to nitrocellulose, and immunoblots were quantitated from multiple exposures using Kodak 1D image analysis software.
Miscellaneous Methods-Protein determination was performed using either the Lowry assay (31) or the BCA assay (Pierce Biotechnology) with bovine serum albumin as a standard.

RESULTS
Expression of YHb in Rho ϩ and Rho 0 Cells-Previously, we have reported that the level of the reduced YHb CO photoproduct is low in respiration-proficient cells and increased in respiration-deficient mutants (16). This can be seen in Fig. 1, which compares the CO photolysis difference spectra (CO-reduced minus reduced) of whole suspensions of JM43, a respirationproficient rho ϩ strain, and its isochromosomal derivative, JM43 0 , a respiratory-deficient rho 0 strain. The major COliganded pigment in JM43 is cytochrome c oxidase. It absorption maxima are at 447 and 610 nm, and absorption minima are at 434 and 590 nm. In contrast, the main CO-liganded pigment in JM43 0 is YHb. Its absorption maxima are at 441, 557, and 590 nm, with the absorption minima at 423, 538, and 574 nm. From Fig. 1 and previous studies that have examined the CO recombination kinetics of JM43 (16), it is clear that JM43 lacks a spectrally detectable CO-binding pigment that has the absorption characteristics of YHb. In contrast, JM43 0 cells contain a CO-binding pigment with all of the spectral signatures of YHb. Despite the absence of a CO spectral signature assignable to YHb in JM43 cells, we have found that YHB1 mRNA levels in JM43 and JM43 0 are essentially identical (Fig. 1B).
The presence of YHb mRNA but the absence of a CO-liganded YHb in JM43 cells suggested one of three things. First, JM43 cells may not translate YHB1 mRNA and, hence, lack the YHb apoprotein. Second, JM43 cells may lack assembled and functional YHb. Third, JM43 cells may express and assemble YHb but maintain the heme of YHb in an oxidized state, which is incapable of binding CO. To examine these possibilities, we first asked if the YHb apoprotein is present in rho ϩ cells. A polyclonal antibody (anti-YhbC) to a 19-amino acid synthetic peptide whose sequence corresponds to that predicted from the carboxyl terminus of the YHB1 gene ( Fig. 2A) was produced to address this question. The reactivity and specificity of this antibody was analyzed by immunoblot analysis of whole cell extracts prepared from JM43 and DR11, a derivative of JM43 that carries a null mutation in the YHB1 gene. From Fig. 2B it is clear that the antibody detects an antigen in whole cell extracts from JM43 but not from DR11. This antigen has an apparent molecular mass of 47,000 Da, which is consistent with the molecular mass predicted for YHb from the sequence of its gene (17). Because the antigen recognized by anti-YHbC is present in JM43 but not in DR11 and has a size similar to that predicted from the DNA sequence of YHB1, we conclude that anti-YHbC recognizes and is specific for YHb. We also conclude that the YHb apoprotein is expressed in JM43 cells. To compare the levels of the YHb apoproteins in JM43 and JM43 0 , cells we subjected whole cell extracts from both to immunoblot analysis. From Fig. 2C it can be seen that YHb levels in JM43 0 cells are 57% of the level in JM43 cells. The finding that the YHb protein is present in JM43 and at higher levels than in JM43 0 allows us to conclude that the absence of a spectrally detectable YHb protein in JM43 cells cannot be explained by a low level of expression of the YHb apoprotein.
Assembled YHb is Present in Both the Cytosol and Mitochondrial Matrix of Respiration-proficient Cells-The results presented above make it clear that the YHb apoprotein is expressed in JM43 cells even though no CO difference spectral signature for assembled YHb can be detected in these cells. To gain further insight concerning this apparent discrepancy, we examined the intracellular location of YHb. We were particularly interested in determining whether YHb is a mitochondrial protein, because rho 0 cells lack both a mitochondrial genome and mitochondrial respiration. By immunoblotting cellular fractions against antisera to YHb, Pet100p (an inner mitochon-FIG. 1. Expression of YHb in rho ؉ and rho 0 cells. Panel A, low temperature CO photolysis difference spectra of strains JM43 and JM43 0 . Glucose reduced whole cell suspensions (0.8 g (wet weight)/ml) were bubbled for 10 min with CO, frozen in liquid nitrogen, placed in a pre-cooled cryostat, equilibrated to Ϫ120°C, and a spectrum was recorded. The cells were then subjected to photolysis, and a second spectrum was recorded. The pre-photolysis CO ligated baseline spectrum is subtracted from that of the photoproduct. Panel B, Northern blot analysis of total RNA isolated from JM43 (lane 1) and JM43 0 (lane 2). Total RNA was prepared from both strains and blotted with both YHB1 and ACT1 probes as described under "Experimental Procedures." The YHB1/ACT1 ratios were determined by quantitation with a phosphorimaging device. drial membrane protein), and 3-phosphoglycerate kinase (a cytosolic protein), we find that YHb is present in both mitochondrial and cytosolic fractions (Fig. 3A), whereas, as expected, PET100p is present only in the mitochondrial fraction, and 3-phosphoglycerate kinase is present mainly in the cytosolic fraction. Immunoblot analysis reveals the complete lack of detectable YHb in both the cytosolic and mitochondrial fractions (Fig. 3B) from strain DR11, indicating that the disruption of the YHB1 gene alone is sufficient to abolish both cytosolic and mitochondrial forms of YHb.
To localize YHb within mitochondria, we prepared mitochondrial subfractions and immunoblotted them against antisera to YHb, Pet100p, and COX IV (subunit IV of cytochrome c oxidase). Pet100p is an integral inner membrane protein, whereas COX IV is a peripheral protein subunit that binds holocytochrome c oxidase on the matrix side of the inner membrane (32,33). The assembled form of COX IV is located at the inner membrane, whereas the unassembled form is located in the matrix (34). From Fig. 4 it can be seen that YHb is present in mitoplasts (lane 2), which contain both the inner membrane and matrix. It is not present in the intermembrane space (Fig.  4, lane 5) or outer membrane (lane 6). Upon fractionating mitoplasts into their two subfractions, the inner membrane (Fig. 4, lane 3) and matrix (lane 4), we find that most of the YHb present in mitochondria and mitoplasts resides in the mitochondrial matrix. As expected for a matrix-located protein, mitochondrial YHb is resistant to proteolysis by externally added trypsin (data not shown) Further evidence for the presence of YHb in the mitochondrial matrix comes from CO difference spectra of this fraction (Fig. 5). Both whole mitochondria (Fig. 5A) and the inner mitochondrial membrane fraction ( Fig. 5B) exhibit major absorption maxima at 447 nm, the CO spectral signature of cytochrome c oxidase. In contrast, the mitochondrial matrix fraction (Fig. 5C) exhibits major absorption maxima at 441 nm, the CO spectral signature of YHb. It is interesting that YHb is not observed by CO difference spectroscopy of intact mitochondria but is observed when the matrix fraction derived from intact mitochondria is assayed.
To determine whether the cytosolic YHb antigen has an YHb CO spectral signature, we subjected the cytosol to ammonium sulfate fractionation and performed immunoblots and CO difference spectroscopy on the fractions obtained. From Fig. 6A it can be seen that the YHb antigen is not precipitated until the ammonium sulfate concentration exceeds 45%. It is most abundant in fraction 4, which comes down between 45 and 65% ammonium sulfate. CO difference spectral analysis of this fraction as well as unfractionated cytosol reveals that each contains a pigment with the spectral signature of YHb. These findings permit three conclusions. First, the YHb antigen and the pigment with a CO spectral signature copurify. Second, the YHb antigen present in the cytosolic fraction is assembled with its prosthetic groups. Third, even though the YHb pigment cannot be observed in whole cells, it is nonetheless present and is easily observed in broken cell extracts.
YHb Is Active in NO Consumption in Both Mitochondria and the Cytosol-Previously, it has been shown that yeast cell ly-

FIG. 2. Generation and specificity of anti-YHbC and comparison of YHb levels in rho ؉ and rho 0 cells. Panel A, a polyclonal
antibody was made to a 19-amino acid synthetic peptide whose sequence corresponds to amino acids 379 -397 of YHb. Panel B, whole cell extracts were prepared from the strains JM43, a strain with a wild-type YHB1 gene, and DR11, which carries a yhb1 null mutation. An aliquot of each was subjected to SDS-PAGE (16% gels) and immunoblotted with anti-YHbC. Lane 1, 10 g of total cell protein from JM43; lane 2, 10 g of total cell protein from DR11. The migration positions of molecular mass markers, given in kilodaltons, are shown at the left. Panel C, aliquots of whole cell extracts from JM43 and JM43 0 were analyzed on 16% SDS-polyacrylamide gels and immunoblotted with anti-YHbC. sates consume NO and that this is attributable to YHb (4). We have been able to confirm this finding here by assaying whole cell extracts from JM43 and DR11 for their ability to consume NO (Fig. 7A). In these studies, NO is injected into an airtight chamber, and NO consumption is followed with an NO electrode. Using a buffer control it is possible to determine the amount of NO injected into the chamber as well as its stability in the absence of cell extract. The buffer control trace and that from the DR11 extract are similar (Fig. 7A), indicating that the disruption of the YHB1 gene nearly completely eliminates NO consumption in whole cell extracts. In contrast, NO consumption in a comparable amount of JM43 whole cell extract is extremely rapid. Indeed, it is so rapid that the initial peak of added NO observed in the buffer control is absent and cannot be observed until the amount of whole cell extract added is reduced 10-fold (data not shown). From Fig. 7, B and C, it can be seen that both mitochondrial and cytosolic fractions from JM43 but not DR11 are also capable of consuming added NO. The initial peak of NO observed in the buffer control is greatly diminished when the JM43 mitochondrial protein or cytosolic protein is assayed. These findings clearly establish that the YHb proteins in both the mitochondrial matrix and cytosol fractions are capable of NO consumption.
YHb Levels in sod1 and sod2 Mutants-The localization of functional YHb to both mitochondrial and cytosolic fractions is surprising and led us to ask if the relative levels of YHb in these two compartments is regulated. To address this question, we first examined the intracellular distribution in cells grown under conditions known to regulate the expression of YHB1. First, we examined the intracellular location of YHb in mutants defective in one or both forms of superoxide dismutase. Previously, we have reported that the YHB1 transcript levels are increased nearly 3-fold in a strain lacking both copper-zinc superoxide dismutase (encoded by SOD1) and manganese superoxide dismutase (encoded by SOD2) (19). Immunoblots of whole cell extracts from sod1, sod2, and sod1sod2 mutants have shown that YHb protein levels are increased in sod1 or sod1sod2 mutants but are unaffected in a sod2 mutant. From Fig. 8A it is clear that YHb concentrations are increased in both mitochondrial (lanes 3 and 4) and cytosolic fractions (lanes 7 and 8) from sod1 or sod1sod2 mutants. Mitochondrial YHb concentrations in these mutants are approximately three times higher than in the wild-type. Cytosolic YHb concentrations are ϳ5-fold higher in sod1 mutants and 7-fold higher in the sod1sod2 mutant, compared with the wild-type. In contrast, YHb concentrations in mitochondrial (Fig. 8A, lanes 1 and 2) FIG. 5. CO-ligated difference spectra of isolated mitochondria, the mitochondrial inner membrane, and matrix fractions. Shown are CO-ligated difference spectra of aliquots (1.7 mg of protein per milliliter) of whole mitochondria (panel A), the mitochondrial inner membrane (panel B), and the mitochondrial matrix (panel C) from JM43. Each fraction was reduced with sodium dithionite, and a spectrum was recorded. Each fraction was then bubbled slowly with CO gas for 5 min. The visible/Soret spectra of the same sample were recorded before and after exposure to CO. The baseline spectrum, obtained from the difference between the two spectra of the same reduced sample, was subtracted from the reduced minus CO-ligated difference spectra.
FIG. 6. CO-ligated difference spectra of the unfractionated cytosol and partially purified cytosolic YHb. The cytosolic fraction from JM43 was subjected to a series of sequential ammonium sulfate precipitations (30,45,65, and 95%) in order to partially purify the YHb protein. Purification was followed by SDS-PAGE (16% gel) and immunoblot analysis using anti-YHbC and CO difference spectroscopy. Panel A, immunoblot analysis of total cytosol (lane 1), 30% ammonium sulfate precipitate (lane 2), 45% ammonium sulfate precipitate (lane 3), 65% ammonium sulfate precipitate (lane 4), 95% ammonium sulfate precipitate (lane 5), and the supernatant that was removed from the 95% precipitate (lane 6). Each ammonium sulfate precipitate was resuspended in 100 mM NaPO 4 , pH 7. An aliquot containing 10 g of protein was loaded onto each lane. Panel B, CO-ligated difference spectra of aliquots (4 mg of protein per milliliter) of the unfractionated cytosol and the resuspended 65% ammonium sulfate precipitate. Spectra were obtained as described in the legend to Fig. 5. Top, cytosol; bottom, resuspended 65% ammonium sulfate precipitate. and cytosolic fractions (lanes 5 and 6) are equivalent in the wild-type and sod2 strains. These findings indicate that the increased expression of YHb observed in a sod1sod2 mutant is brought about by the lack of copper-zinc superoxide dismutase, but not of manganese superoxide dismutase, and that in the absence of copper-zinc superoxide dismutase, YHb levels increase in both the mitochondrion and the cytosol.
Intracellular Distribution of YHb in Rho 0 , Nuclear pet, and Anoxic Cells-To examine further the relationship between respiration and Yhb, we have determined the total levels of YHb in mitochondrial and cytosolic fractions from JM43 0 cells, JM43GD5ab, and JM43 cells grown under anoxic and normoxic conditions. These three strains were chosen to examine the effects of respiration, the mitochondrial genome, and the presence of oxygen on the intracellular localization of YHb. They are all isochromosomal. JM43 0 differs from JM43 in lacking both a mitochondrial genome and the mitochondrially encoded proteins of the respiratory chain. JM43GD5ab differs from JM43 in lacking cytochrome c oxidase and, hence, respiration. It contains a wild-type mitochondrial genome. Anoxic JM43 cells also retain a mitochondrial genome and the components of the respiratory chain 2 but differ from normoxic JM43 cells by being grown in the absence of oxygen, the terminal electron acceptor of the respiratory chain.
Because YHb is present only in the mitochondrial and cytosolic fractions, it is possible to determine the relative total content of YHb in each fraction from the relative concentration of YHb in each fraction and the relative total protein content of each fraction (Table I). The majority of YHb in aerobically grown JM43 cells is present in the cytosolic fraction; this represents 36.3 arbitrary units of the protein. As seen in Fig. 1 and Table I, the level of YHb in JM43 0 cells is ϳ57% of its level in JM43 cells. Surprisingly, nearly all of the YHb protein in JM43 0 cells is found in the cytosol. (Fig. 8B and Table I). JM43 0 cells differ from JM43 cells both genotypically and phenotypically; they lack a mitochondrial genome and respiration. To address whether the intracellular distribution of YHb is affected by the lack of the mitochondrial genome or the lack of respiration, we examined mitochondrial and cytosolic YHb levels in JM43GD5ab, a nuclear pet mutant that retains a wild-type mitochondrial genome but carries gene disruptions in both subunit V isoforms of cytochrome c oxidase and does not respire. This strain distributes its YHb between mitochondrial and cytosolic fractions (Fig. 8, lanes 5 and 6; Table I) in the same way as JM43, its wild-type parent. Hence, the altered distribution of YHb in JM43 0 cannot be attributed to the lack of respiration per se. Further evidence for this comes from an examination of the intracellular distribution of YHb in anoxically grown JM43 cells. Anoxic growth has only a slight effect on total YHb protein levels in JM43 cells (Table I). Interestingly, however, it does affect the intracellular distribution of YHb in a way that is opposite from that observed in rho 0 cells. All of the YHb in anoxic JM43 cells is located in the promitochondrial fraction (Fig. 8, lanes 7 and 8, and Table I). None can be observed in the cytosol. Interestingly, the number of arbitrary units of YHb protein in promitochondria of anoxic JM43 cells is nearly identical to the number of arbitrary units of YHb protein in the cytosolic fraction of aerobic JM43 cells.
When considered together, these findings suggest that the altered distribution of YHb observed in rho 0 cells is the result of the lack of the mitochondrial genome and not the lack of respiration. These results and those obtained with the sod mutants indicate that yeast cells can differentially regulate the intracellular level of YHb and its intracellular distribution between cytosolic and mitochondrial fractions. DISCUSSION This study presents three novel findings concerning the YHb flavohemoglobin in yeast. First, although CO-liganded YHb is not detectable in intact reduced rho ϩ cells, it is detectable in cell fractions from rho ϩ cells. Second, functional YHb localizes to two distinct intracellular compartments, the cytosol and mitochondrial matrix. Third, the distribution of YHb between the cytosol and mitochondrial matrix is affected by anoxia and the mitochondrial genome.
Our findings that YHB1 mRNA and the YHb protein are expressed in rho ϩ cells, and that a CO-liganded YHb is not detectable in intact rho ϩ cells but is detectable in cell fractions from rho ϩ cells seem to be paradoxical. There appear to be at least two plausible explanations. One explanation for this apparent discrepancy is that the accessibility of intracellular YHb to NAD(P)H or NADH, its physiological reductants (35), is limited in vivo most likely due to competition between YHb and the mitochondrial respiratory chain. To bind CO, the heme group in YHb must be reduced. In the absence of reduced heme, CO cannot bind to YHb in vivo, and a spectral signature would not be observed. According to this explanation, CO-liganded YHb would be observable in cell fractions and rho 0 cells because the competition for NAD(P)H and/or NADH would be absent. A second and more likely explanation is that YHb heme is maintained in the oxidized state in intact cells because it has a high turnover rate with NO (35). As such, electrons would be released from the heme group and act on NO as soon as it binds to the heme. According to this explanation, CO-liganded YHb should be observed when levels of its substrate, NO, decrease. Because NO is very unstable, its level should decrease dramatically when cells are broken. This decreased level of NO in broken cell extracts could explain why CO-liganded YHb is observed in cell fractions from rho ϩ cells but not in intact rho ϩ cells. Because CO-liganded YHb is observed in rho 0 cells, this explanation would also imply that the respiratory chain and/or the mitochondrial genome is required for NO production. Involvement of the mitochondrial respiratory chain in NO production could explain our observation that a CO-liganded YHb pigment cannot be observed in intact respiring mitochondria but can be observed when the mitochondrial matrix subfraction is isolated.
The presence of YHb that is functional in NO consumption in both cytosolic and mitochondrial fractions of JM43 implies that NO is produced in both cellular compartments in yeast cells. This is not surprising, given a recent report that yeast cells express proteins that cross-react with all three mammalian nitric oxide synthase isoforms, namely neuronal (NOS1), inducible (NOS2), and endothelial (NOS3) (36). Although the intracellular location of the yeast isoforms has not been established to date, at least one of the mammalian nitric oxide isoforms (NOS3) is associated with mitochondria (37,38).
Localization of YHb to mitochondria can help explain how it functions both in oxidative stress (19) and nitrosative stress (4). Indeed, it is likely that YHb protects yeast cells from oxidative stress primarily by controlling levels of mitochondrial NO, which is a potent inhibitor of cytochrome c oxidase and respiration (39) and a reactant that, together with superoxide, can form peroxynitrite (40). Peroxynitrite participates in some forms of protein nitration (41). In the absence of YHb, NO levels would increase (4) and inhibit cytochrome c oxidase, leading to the production of superoxide and, hence, enhanced levels of peroxynitrite. Thus, YHb can reduce superoxide levels not by dismuting superoxide, like superoxide dismutase, but rather by reducing its formation via the respiratory chain. Our finding that YHb is present mainly in the cytosol of rho 0 cells is consistent with this hypothesis, because these cells lack important components of the mitochondrial respiratory chain (cytochrome bc 1 and cytochrome c oxidase) and, hence, are respiration-deficient and incapable of producing superoxide. Our finding that YHb is present in both the cytosol and the mitochondrion of JM43GD5ab cells is also interesting, because although these cells lack cytochrome c oxidase they still retain the bc 1 complex and a partial respiratory chain. Hence, they are capable of superoxide production. The presence of YHb in the promitochondria of anoxic cells is somewhat surprising, considering that these cells are grown in the absence of oxygen and have greatly reduced levels of oxidative stress (42). However, this finding implies that NO is present in anoxic mitochondria, which is consistent with recent studies that have demonstrated that the concentration of NO increases in rat liver mitochondrial suspensions subjected to low oxygen tensions (43) and that mitochondrial protein S-nitrosylation occurs in anoxic rat liver mitochondria (44).
Our finding that the ratio of YHb in mitochondrial and cytosolic fractions differs in rho ϩ , rho 0 , and anoxic cells indicates that yeast cells can differentially regulate the intracellular distribution of YHb between the cytosol and the mitochondrion. Precedence for the regulated differential targeting of a single yeast protein between two different intracellular locations comes from studies on the dual targeting of catalase A to a Values given are the average of at least two determinations of YHb levels in mitochondrial and cytosolic fractions. b Total units are determined relative to those in JM43 cells grown under aerobic conditions, which is taken as 100%. Levels in JM43 grown under aerobic conditions are calculated from the percentage of concentration in mitochondrial and cytosolic fractions (determined by quantitating Western immunoblots such as those in Fig. 8B) multiplied by the ratio of total protein accounted for by each fraction (see "Experimental Procedures"). Total units in mitochondrial and cytosolic fractions from JM43 cells grown under anaerobic conditions, JM43 0 cells, and JM43GD5ab cells were calculated as described above and normalized to their whole cell level. mitochondria and peroxisomes (45). The level of respiration and/or growth on oleic acid influences the intracellular distribution of this protein. The presence or absence of respiration does not appear to influence the intracellular distribution of YHb, because the mitochondrial fraction to cytosolic fraction ratio of YHb is identical in JM43, a respiration-proficient strain, and in JM43GD5ab, a respiratory-deficient derivative of JM43. Moreover, the ratios of mitochondrial/cytosolic YHb in JM43 0 and anoxic JM43 cells are nearly the converse of one another, even though both cells are respiration-deficient. These findings clearly demonstrate that the absence of respiration per se is not a determinant of the intracellular localization of YHb and indicate that anoxia and the mitochondrial genome influence the intracellular localization of YHb.
Previously, we have shown that the mitochondrial genome, acting independently of respiration and oxidative phosphorylation, affects the expression of a subset of nuclear genes (46), and we have termed this type of mitochondrial-nuclear crosstalk "intergenomic signaling." Genes that are regulated by "intergenomic signaling" are specifically down-regulated in cells that lack a mitochondrial genome (46). More recently, using microarray analysis we have been able to show that this pathway affects the expression of Ͼ100 nuclear genes in yeast. 3 The target genes of this pathway are different from the target genes of the better understood "retrograde regulation" pathway, which involves mitochondrial respiration and connects altered mitochondrial respiration to altered carbon and nitrogen metabolism (47). In considering how the mitochondrial genome affects the intracellular distribution of YHb, it seems likely that one of the nuclear target genes affected by intergenomic signaling encodes a protein that functions in the intracellular partitioning of YHb.
Two lines of evidence indicate that the mitochondrial and cytosolic forms of YHb are derived from a single nuclear gene, YHB1. First, a null mutation in YHB1 simultaneously abolishes both proteins. Second, transformation of a yhb1 null mutant with a plasmid containing the YHb protein leads to the appearance of YHb in both the mitochondrial matrix and the cytosol (19). 4 The presence of a single nuclear gene product in multiple intracellular compartments is not without precedence. Indeed, a growing number of mitochondrial proteins are being found in both the cytosol and the mitochondrion (48,49). These include the histidyl-tRNA synthase (50), fumarase (51,52), copper-zinc superoxide dismutase (53), and dihydroxybutanone phosphate synthase (54) from yeast, thioredoxin from Drosophila (55), the thioredoxin glutathione reductase from Echinococcus (56), and several human proteins (48) including aspartate aminotransferase (57) and the Wilson's copper transport protein (58). Close examination of several of these proteins has revealed a multiplicity of mechanisms for the localization of proteins to the mitochondrion and the cytosol (48). These include the use of multiple transcription start sites, multiple translation start sites, and mitochondrial import followed by retrograde movement out into the cytosol (52). Currently, it is not clear which of these mechanisms is used by YHb. Moreover, because the DNA sequence of YHB1 does not predict the presence of a mitochondrial signal peptide, it is also not clear how YHb is targeted to the mitochondrial matrix. One possibility is that, like some other nuclear coded mitochondrial proteins, YHb uses an internal targeting sequence (59,60) for import into the mitochondrion. Alternatively, the YHb gene may produce an alternative transcript that puts a mitochondrial targeting sequence on the YHb protein. These possibilities are currently under study.