Flavohemoglobin expression and function in Saccharomyces cerevisiae. No relationship with respiration and complex response to oxidative stress.

The yeast Saccharomyces cerevisiae contains a flavohemoglobin, encoded by the gene YHB1, whose function is unclear. Previous reports presented evidence that its maximal expression requires disruption of mitochondrial respiration and that it plays a role in the response to oxidative stress. We have studied the expression of YHB1 in respiratory deficient cells and in cells exposed to various compounds causing oxidative stress. Several different strains and approaches (spectroscopic detection of the oxygenated form of Yhb1p, beta-galactosidase activity of a YHB1-lacZ fusion, and Northern blot analysis) were used to demonstrate that YHB1 expression and Yhb1p production are not increased by respiration deficiency. YHB1 expression was unchanged in cells challenged with antimycin A or menadione, while it decreased in cells exposed to H2O2, diamide, dithiothreitol, and Cu2+. Transcription of YHB1 is not under the control of the transcriptional factor Yap1p. These results do not support a participation of YHB1 in the genetic response to oxidative stress. We also analyzed the growth phenotypes associated with altered Yhb1p production using YHB1-deleted strains and strains that greatly overproduced Yhb1p. Yhb1p appears to protect cells against the damage caused by Cu2+ and dithiothreitol, while sensitizing them to H2O2. Yhb1p overproduction in a glucose-6-phosphate dehydrogenase-deficient mutant decreased its growth rate. These data indicate that there is a complex relationship(s) between Yhb1p function(s) and cell defense reactions against various stresses.

The yeast Saccharomyces cerevisiae contains a flavohemoglobin, encoded by the gene YHB1, whose function is unclear. Previous reports presented evidence that its maximal expression requires disruption of mitochondrial respiration and that it plays a role in the response to oxidative stress. We have studied the expression of YHB1 in respiratory deficient cells and in cells exposed to various compounds causing oxidative stress. Several different strains and approaches (spectroscopic detection of the oxygenated form of Yhb1p, ␤-galactosidase activity of a YHB1-lacZ fusion, and Northern blot analysis) were used to demonstrate that YHB1 expression and Yhb1p production are not increased by respiration deficiency. YHB1 expression was unchanged in cells challenged with antimycin A or menadione, while it decreased in cells exposed to H 2 O 2 , diamide, dithiothreitol, and Cu 2؉ . Transcription of YHB1 is not under the control of the transcriptional factor Yap1p. These results do not support a participation of YHB1 in the genetic response to oxidative stress. We also analyzed the growth phenotypes associated with altered Yhb1p production using YHB1-deleted strains and strains that greatly overproduced Yhb1p. Yhb1p appears to protect cells against the damage caused by Cu 2؉ and dithiothreitol, while sensitizing them to H 2 O 2 . Yhb1p overproduction in a glucose-6-phosphate dehydrogenase-deficient mutant decreased its growth rate. These data indicate that there is a complex relationship(s) between Yhb1p function(s) and cell defense reactions against various stresses.
Hemoglobins (Hbs) 1 are very ancient hemoproteins that are widely distributed in nature; they are found in all groups of organisms, including many invertebrates, plants, fungi, and bacteria (1). They all bind reversibly molecular oxygen, but their avidities vary greatly. The hemoglobin from the worm Ascaris binds oxygen 4 orders of magnitude more tightly than does human Hb (2). Some hemoglobins bind other molecules besides oxygen; the Hbs from the mollusc Lucina pectinata (3) and the deep-sea hydrothermal vent worm Riftia pachyptila (4) bind hydrogen sulfide and transport it to the sulfur-oxidizing bacterial symbionts which feed them. Hbs also vary greatly in size and composition, from the single-domain monomeric myoglobin to the complex heteromultimeric Riftia Hbs (4); but their heme-binding domains all seem to have retained the characteristic "globin fold," although there are subtle variations in the amino acid residues lining the heme pocket, and these are responsible for the differences in their properties (Refs. 1-3 and references therein).
Hb is normally involved in oxygen transport, storage, and facilitated diffusion, as is well established in animals. In plants, the specialized leghemoglobins transport oxygen in the nodules during symbiotic nitrogen fixation, while the nonsymbiotic Hbs presumably facilitate oxygen diffusion in actively respiring cells (5). These Hbs are expressed in several tissues with high metabolic activity (5) and, in barley, they are induced by hypoxia and by inhibitors of respiration and oxidative phosphorylation (6). The homodimeric Hb from the obligately aerobic bacterium Vitreoscilla, whose synthesis is strongly stimulated in hypoxic conditions, probably plays a role in facilitation of oxygen diffusion or in oxygen delivery (7,8). The overproduction of Vitreoscilla Hb in such hosts as Escherichia coli (9), yeast (10), and plants (11), enhanced growth metabolism by increasing the availability of oxygen and/or the supply of ATP in the cells.
A class of intriguing Hbs, the flavohemoglobins, has recently been described in yeast and in several bacteria. These "twodomain" globins contain an N-terminal globin module fused to a reductase module that is homologous to the ferredoxin-NADPH reductase family, with one FAD and one NAD(P)Hbinding domains (12)(13)(14)(15)(16)(17)(18). Their physiological function remains uncertain and the published data on them are often confusing. The flavoHb in Erwinia chrysanthemi is essential for pathogenicity (16). The flavoHbs of the denitrifier Alcaligenes eutrophus, Bacillus subtilis, and E. coli are all induced by nitrate and nitrite under anaerobic conditions (15,17,19); nitric oxide also induces expression of the E. coli flavoHb (19). These findings suggest a role of these flavoHbs in the anaerobic metabolism of nitrogen oxide compounds, but the way in which they act is not known. In vitro studies on purified E. coli flavoHb suggest that it may function as an NADH oxidase (20), with the release of superoxide anions (21), or also as a reductase (22), acting perhaps as an oxygen sensor in vivo. E. coli flavoHb has also been implicated in oxidative stress: it produces superoxide in vivo when overproduced (21) and its expression is induced by the redox cyclic agent, paraquat, which generates intracellular superoxide (23).
Hb was first identified in yeasts by Keilin (24). The flavoHb of Candida mycoderma has a very high affinity for oxygen, similar to that of cytochrome oxidase but a thousand times higher than that of human Hb (but with same oxygen dissociation rate constant); there is no evidence that it facilitates oxygen diffusion in the cells (25,26). In Saccharomyces cerevi-siae, the expression of the gene YHB1, encoding the flavoHb in this yeast, is regulated by the growth phase and oxygen in a manner opposite to that found in most bacteria, suggesting that it has a different function in this yeast (27). The fact that cells deleted of YHB1 become sensitive to some conditions promoting oxidative stress, and that these conditions also enhance YHB1 expression suggest that Yhb1p contributes to the response to oxidative stress (28). This study explores the regulation of YHB1 gene expression by respiration and by different agents that cause oxidative stress (H 2 O 2 , diamide, menadione, DTT, Cu 2ϩ ), and the effect of a lack or excess of Yhb1p on sensitivity of cells to these various stresses.

EXPERIMENTAL PROCEDURES
Yeast Strains and Plasmids-The S. cerevisiae strains used in this study are described in Table I. The rho Ϫ derivatives were prepared by growing the cells in the presence of ethidium bromide (35). The YHB1 gene (reading frame YGR234w of chromosome VII (36)) was disrupted by "␥-transformation" (31) using routine DNA manipulation procedures (37). Two DNA fragments (nucleotides Ϫ679 to ϩ7 and ϩ990 to ϩ1202; the A of the initiating ATG is ϩ1) flanking the coding region were amplified by polymerase chain reaction using yeast genomic DNA, and cloned into the plasmid pRS403 (31) in tandem, but in reverse order, leaving an EcoRI site between them. After linearization at this site, the construct was transformed into yeast cells by a modified lithium acetate method (38). Integration of the deleted/disrupted yhb1-⌬1::HIS3 allele at the proper locus was confirmed by polymerase chain reaction analysis of whole cells (39).
The plasmids pADE/YHB1 and pPGE/YHB1 for overproducing Yhb1p were constructed as follows. The coding region of YHB1 was polymerase chain reaction-amplified from yeast genomic DNA using forward (5Ј-CGCGGATCCATTCATTatgCTAGCCG-3Ј) and reverse (5Ј-CCGGAATTCCGCctaAACTTGCACGG-3Ј) primers containing sequences for BamHI and EcoRI restriction sites (underlined) and the initiating ATG and stop TAG codons (in lowercase type), respectively. The resulting DNA fragment was inserted into the polylinker of the multicopy expression vectors pYADE4 and pYPGE2 (40). This placed the expression of YHB1 under the control of the glucose-regulatable ADH2 or the constitutive PGK1 promoters. The entire YHB1 sequence was verified by sequencing using the Sequenase kit (U. S. Biochemical Corp.). The YHB1-lacZ fusion was constructed by ligating a polymerase chain reaction-amplified DNA fragment encompassing 679 nucleotides of the 5Ј-untranslated region and the first two codons of YHB1 to the lacZ gene of the episomal YEp357 plasmid (41).
Respiration-The oxygen consumption of whole cells was monitored at 30°C with a Clark-type electrode in 67 mM KH 2 PO 4 , using glucose or ethanol as the respiratory substrate. KCN was used at a concentration of 2 mM final. Respiration is expressed as nanomole of O 2 /min/mg dry weight of cells.
␤-Galactosidase Assay-␤-Galactosidase activity was measured on chloroform-permeabilized cells grown in selective minimal medium (SCglu -ura) at OD 600 ϭ 0.3-0.5 (35). It is expressed as Miller units (A 420 nm ϫ 1000/min/OD 600 ). The values reported are the averages of duplicate determinations of at least two separate cultures, each inoculated with 10 -12 independent transformants.
Northern Blot Analysis-Total RNA (15 g), isolated from cells in exponential growth (42), was electrophoresed on agarose-formaldehyde gels, transferred to nylon membranes, and probed with 32 P-labeled DNA fragments under standard conditions (35,37). The probes were a 1.2-kilobase BamHI-EcoRI fragment from plasmid pPGE/YHB1 for the YHB1 gene, and a 1.1-kilobase XhoI-HindIII fragment of the actin gene ACT1 used to monitor RNA loading. The autoradiograms were quantified by scanning densitometry.
Total Heme and Cytochrome Content-The whole cell cytochrome content was estimated by low temperature (liquid nitrogen) spectrophotometry of cells (40 mg dry weight) filtered from liquid cultures in exponential growth phase (43), using a Leres S-66 spectrophotometer. The filtered cells were also resuspended in 5 ml of water, bubbled with CO for 5 min, or treated with sodium dithionite (a few grains) or KCN (5 mM final), and the treated cells processed for spectrophotometric analysis as above. This rapid, reliable method allowed the direct, quantitative comparison of the cytochrome and hemoglobin contents of different cells. The total heme contents of whole cells and mitochondrial membranes were quantified from the reduced-minus-oxidized difference spectra of the pyridine hemochromes (44). Membranes were prepared after mechanical disruption of the cells (45).

RESULTS AND DISCUSSION
Assignment of the Absorption Band at 575 nm to Oxyhemoglobin (HbO 2 )-Low temperature spectra of whole cells of S. cerevisiae often have an absorption band near 575-577 nm which we attributed previously to porphyrins and Zn-porphyrins (46). We have now shown that this band is in fact due to the oxygenated form of the flavohemoglobin Yhb1p. It disappears after cells are reduced with sodium dithionite and is As for SEY6210, yap1-⌬1::HIS3 replaced by a more diffuse band at 568 -570 nm in cells treated with carbon monoxide (CO) (Fig. 1), which is characteristic of a hemoprotein that combines reversibly with molecular oxygen. The band at 575 nm is also absent from S/⌬HB cells lacking the YHB1 gene (⌬Yhb1), and reappears when these cells are transformed with a plasmid expressing YHB1 (pADE/YHB1) (Fig. 1).
Identical results were obtained with the strains YPH499 and W303-1B and their ⌬Yhb1 derivatives. Control experiments with hem12 mutant cells partially deficient in uroporphyrinogen decarboxylase (47) showed that the peak at 575 nm due to the accumulated Zn-porphyrins does not change when cells are treated with dithionite or CO (data not presented).
Respiratory Activity and Expression of YHB1-Yhb1p is found mainly in rho Ϫ cells and in cells growing in the presence of antimycin A. It was therefore proposed that its expression is activated by the disruption of mitochondrial respiration (13,28,48). We have examined the effects of respiratory deficiency on the presence of Yhb1p, detected by the absorption at 575 nm of its oxygenated form, in parallel with the expression of the YHB1 gene estimated by Northern blot and the activity of a YHB1-lacZ fusion. Strains S150-2B and W303-1B were grown in a glucose medium in the presence of antimycin A, an inhibitor of the mitochondrial electron transport chain, or chloramphenicol, an inhibitor of mitochondrial protein synthesis (for S150-2B, cells in the rho Ϫ state were used, as these are phenotypically similar to cells grown with chloramphenicol). Full derepression of the respiratory chain was achieved by growth with glycerol ϩ ethanol. In contrast to S150-2B cells (Fig. 1), the peak at 575 nm was very small or absent in the low temperature spectra of W303-1B grown in glucose, unless the cells were treated with KCN ( Fig. 2A, curves 1). This is because the W303-1B cells have a greater respiratory activity than S150-2B cells, and "cyanide, by inhibiting the reoxidation of the reduced components of cytochrome, prevents the utilization of oxygen in yeast respiration and thus protects oxyhemoglobin from deoxygenation" (24). KCN must also be added to observe the band at 575 nm in fully derepressed cells grown in glycerol ϩ ethanol ( Fig. 2A, curve 4, and Fig. 3B), and the further addition of dithionite or CO abolished or displaced it as was shown in Fig. 1.
Disrupting the mitochondrial respiration with antimycin A or chloramphenicol did not appear to increase the intensity of the absorption band at 575 nm ( Fig. 2A, curves 2 and 3), indicating no activation of the YHB1 gene. This is confirmed by the lack of any significant change in the steady-state amount of YHB1 mRNA (Fig. 2B, lanes 1-3), or in the activity of a YHB1-lacZ fusion (Table II). YHB1 expression was decreased to about 60% of the control value in the rho Ϫ derivative of strain S150-2B (Fig. 2B, lane 3 and Table II). Growing the cells in glycerol ϩ ethanol slightly increased the amounts of YHB1 mRNA (Fig. 2B, lane 4) and caused a 2.5-3.5-fold increase in ␤-galactosidase activity (Table II), confirming that glucose exerts very little, if any, repression on YHB1 expression (27). Heme-deficiency and anaerobiosis both reduced considerably the expression of YHB1, but to different extents (Table II), as described previously (27,28); this is in contrast to most bacterial Hb genes, which are induced during growth under oxygen Where noted, the cells were treated with KCN. B, Northern blot analysis of the expression of YHB1 in the S150-2B and W303-1B strains. Total RNA was prepared and probed for YHB1 and ACT1 mRNA levels. The YHB1/ACT1 ratios were determined by densitometry and normalized to 1 for normal growth in glucose (lane 1). Lane ⌬HB, strain S/⌬HB (yhb1-⌬1::HIS3) used as control for YHB1 mRNA; lanes 1-4, same growth conditions as those described above for curves 1-4, except that strain S150-2B/rho Ϫ was used in lane 3. limitation (7,8,15,17).
Our results clearly show that mitochondrial respiratory activity need not be disrupted for optimal expression of YHB1 and the production of the holoprotein. The reported absence of Yhb1p in respiratory-proficient cells (13) is probably due to the fact that the cells were in late stationary phase: because the concentration of YHB1 mRNA drops sharply upon exit from the exponential growth (27), Yhb1p is probably diluted out during the oxidative diauxic growth, what would not occur in respiratory deficient cells.
Effects of Overexpressing YHB1-The effects of deleting the flavohemoglobin gene on cell growth and sensitivity to oxidative stress have been described (28). But the effects of Yhb1p overproduction have not previously been reported. Cells overexpressing Yhb1p from the plasmid pPGE/YHB1 exhibited a red-brown tint when collected. They grew a little slower than control cells carrying the vector pYPGE2 (Table III), but reached the same final cell density. For instance, the increase in doubling time was 14% for the strain S150-2B and 7% for the strain W303-1B growing in glucose synthetic medium. Overproducing Yhb1p might be detrimental to the cells, and this is consistent with plasmid loss experiments showing that 80 -85% of the cells contained the plasmid pYPGE2, but only 32-36% contained pPGE/YHB1 at the end of the culture on glucose rich medium.
Low temperature spectra of whole cells overexpressing YHB1 showed the presence of large amounts of hemoglobin absorbing at 575 and 541 nm in its oxygenated form, and at 568 and 538 nm in its CO-liganded form (Fig. 3, A and B). The amount of Yhb1p, which was comparable to that of the mitochondrial cytochromes in the control cells, was estimated to be 1-2% of the total cellular proteins, similar to that reported for the production of various human Hbs in S. cerevisiae (49,50). This represented a doubling of the heme synthesis capacity (Table III), which was not increased further by adding the heme precursor 5-aminolevulinic acid to the culture. It is interesting that the synthesis of large amounts of holohemoglobin occurred at the expense of mitochondrial cytochromes. The decrease in their concentrations, which can be seen in the low temperature spectra (Fig. 3, A and B), was estimated to be 2-4-fold by spectrophotometric analysis and measuring the total heme contents of mitochondrial membranes (Table III). Glucose-grown cells also showed a concomitant decrease in cell respiration (Table III). Cyanide-insensitive respiration was very low in glucose-grown cells (0.5 nmol of O 2 /min/mg dry wt), and zero in cells grown in glycerol ϩ ethanol. These data indicate that the accumulated flavohemoglobin does not contribute significantly to the oxygen consumption of whole cells, and, consequently that Yhb1p probably has no oxidase activity in vivo. This is consistent with the fact that the flavoHb purified from the yeast C. mycoderma is relatively stable in its oxygenated form (26).
Thus, the cells appear to produce more heme in response to an increased demand made by the apohemoglobin, and this heme is used by the cytosolic globin rather than by the mitochondrial cytochromes. This implies that the heme newly synthesized in the mitochondria is first used in the cytosol before being incorporated into the mitochondrial cytochromes. Alternatively, overproduction of hemoglobin may reduce the transcription of the genes encoding the apocytochromes, by disturbing heme traffic or by some other unknown means. Our results (Table IV) suggest that this might be the case.
We then asked whether overproduction of Yhb1p occurs in anaerobically grown cells. Although both coproporphyrinogen and protoporphyrinogen oxidases (sixth and seventh enzymes in the heme biosynthetic pathway) require molecular oxygen for their activity in vitro, a limited amount of heme is synthesized in vivo by cells growing in anaerobiosis, probably at the expense of traces of contaminating air that are virtually impossible to avoid (46). This heme is associated with the microsomal cytochromes b 5 and P450 and with Yhb1p. Yhb1p is overproduced in the anaerobic cells containing pPGE/YHB1, but its concentration varied from one culture to another ( Fig.  3C and Table III). This most likely reflects the differences in the amount of air that leaked into the flasks. The possibility of significantly increasing the amount of heme made by anaerobic cells led us to examine the transcriptional activity of the HEM13 and CYC1 genes in these cells, as their expression is controlled by oxygen and heme in an opposite fashion, negative for HEM13 and positive for CYC1 (see Ref. 51 for review). Since there is less heme synthesized in anaerobic cells, heme is  Tables III and IV.   TABLE II Effects of different physiological conditions on the expression of a YHB1-lacZ fusion Strains S150 -2B and W303-1B carrying the plasmid-borne YHB1-lacZ fusion were grown overnight in synthetic medium supplemented with glucose or glycerol ϩ ethanol to OD 600 ϭ 0.3-0.5. Antimycin A, when added, was 1 mg/liter. For respiration deficiency, the S150 -2B/ rho Ϫ strain was used, and strain W303-1B was grown with 4 g/liter chloramphenicol. For heme deficiency, the strain S/⌬H1 (hem1-⌬) was used. ␤-Galactosidase activity was assayed in duplicate on cultures inoculated with 10 -12 independent transformants. Values reported (A 420nm ϫ 1000/min/OD 600 of cells) are the mean Ϯ S.E.

Yeast Hemoglobin Expression and Function
believed to mediate oxygen regulation (51). Overproduction of Yhb1p, and hence of endogenous heme, should thus reverse the effects of oxygen deficiency and lead to aerobic levels of expression. This is obviously not the case (Table IV). The activities of the HEM13-lacZ and CYC1-lacZ fusions were unchanged in the anaerobic cells, whatever the amount of heme made, indicating that oxygen regulation is not simply mediated by the heme concentration. In contrast, a 2-3-fold difference in the expressions of the two fusions was consistently found in aerobic cells overproducing Yhb1p, suggesting that Yhb1p overproduction influenced the aerobic transcription of HEM13 and CYC1. The expression of SOD2, encoding the mitochondrial manganese-superoxide dismutase, was not induced by Yhb1p overexpression (Table IV), as would have been expected (30) if Yhb1p had generated superoxide anions during NAD(P)H oxidation, as has been described for E. coli flavoHb (21). Expression of YHB1 in Response to Agents Causing Oxidative Stress-The expression of YHB1 was not significantly altered in cells grown in the presence of antimycin A, or in rho Ϫ cells (Fig. 2), conditions that enhance (antimycin A (see Ref. 52 and references therein and also, Ref. 53)) or decrease (rho Ϫ state (54,55)) the mitochondrial production of superoxide anions and other reactive oxygen species. These results differ from those of a recent report (28) indicating that YHB1 is induced in cells exposed to reagents that promote oxidative stress, including antimycin A. We therefore analyzed the expression of YHB1 in response to oxidative stress generated by antimycin A, H 2 O 2 , the thiol-oxidizing agent diamide (which decreases the cell glutathione content), the redox-cycling agent menadione bisulfite (that generates superoxide anions and H 2 O 2 extracellularly (56)), DTT, which is a reductant in a metal-catalyzed oxidation system (Ref. 57 and references therein), and Cu 2ϩ , a transition metal ion that can initiate Fenton chemistry. Cultures of the strains S150-2B and W303-1B, with or without the plasmid-borne YHB1-lacZ fusion, were grown in glucose minimal medium to exponential phase (OD 600 ϭ 0.3-0.4), the drugs were added, and the cells were incubated for an additional 1 h. The ␤-galactosidase activity was then assayed and RNA extracted for Northern blot analysis. OD 600 measurements indicated that the growth of the cells treated with H 2 O 2 , diamide, DTT, and Cu 2ϩ was slowed. As shown in Fig. 4A for the strain S150-2B, the ␤-galactosidase activity in cells treated with antimycin A or menadione was similar to that of untreated cells. Enzyme activity was slightly decreased after treatment with H 2 O 2 , and this was more pronounced (30 -35%) in cells exposed to diamide, DTT, or Cu 2ϩ . Identical results were obtained with strain W303-1B (not shown). Thus none of the stresses tested substantially increased the expression of YHB1, which was confirmed by the analysis of the steady state levels of YHB1 mRNA in both S150-2B and W303-1B cells (Fig. 4B). The amounts of YHB1 mRNA were not significantly altered by exposure to antimycin A or menadione. H 2 O 2 caused a slight decrease, while diamide and DTT led to dramatic falls in YHB1 mRNA content. The responses of the two strains to Cu 2ϩ appeared to be different; this probably reflects the difference(s) in their sensitivity (and Cu 2ϩ metabolism?), S150-2B being much more sensitive than W303-1B. The fact that there was a much greater drop in mRNA than the reduction in ␤-galactosidase activity is probably due to the great stability of the ␤-galactosidase fusion protein (t1 ⁄2 Ͼ 20 h (58)). We also analyzed the response of YHB1 to oxidants using the experimental procedure described by Zhao et al. (28). Again, there was no induction with either the W303-1B strain (Fig. 4B, lower part), or the S150-2B strain (not shown), indicating that our results are not peculiar to one strain. Last, we explored the possibility that YHB1 was under the control of Yap1p, a transcriptional activator that plays a central role in the oxidative stress response of cells by mediating the induction of the genes for a number of TABLE III Effect of Yhb1p overexpression on heme and cytochrome synthesis Strain S150 -2B carrying plasmid pYPGE2 (control) or pPGE/YHB1 (Yhb1p overproduction), was grown aerobically (O 2 ) in SC medium containing glucose (Glu) or glycerol ϩ ethanol (Gly), and anaerobically (N 2 ) in SC glucose medium. Cells were collected in exponential growth phase at OD 600 ϭ 1-1.2. The respiration of whole cells and the total heme content of whole cells and membranes were measured. Results are the averages of duplicate assays done in two separate experiments. Standard errors were Ϯ7% for respiration and Ϯ5-15% for heme content. The values reported for anaerobic cells are the ranges of values determined in the four separate experiments described in Table IV    Thus our results indicate that Yhb1p does not participate in the genetic response to oxidative stress. Not only was the expression of YHB1 not induced by any of the oxidants tested, but it was greatly reduced by some of them, like diamide and DTT. Whether the discrepancies between our results and those reported recently (28) are due to differences in strain, experimental procedures, or other factors is not clear at present.
Sensitivity to Oxidants and the Production of Yhb1p-The drop in YHB1 mRNA in cells challenged with H 2 O 2 , diamide, DTT, and Cu 2ϩ suggested that Yhb1p was harmful for cells enduring the stresses caused by these compounds. To test the possibility that Yhb1p plays a role in the antioxidant enzyme system, we analyzed the effects of various oxidants on the growth of cells containing different amounts of Yhb1p. The strains S150-2B or W303-1B and their derivatives deleted of YHB1 or overexpressing it from the plasmid pPGE/YHB1 were used. The results obtained in the context of strain W303-1B are presented in Fig. 5. Yhb1p seems to protect the cells from damage(s) caused by Cu 2ϩ and DTT. The YHB1-deleted cells were sensitive to lower concentrations of Cu 2ϩ and DTT than were wild type cells. Conversely, overexpressing YHB1 increased the resistance of cells to these compounds; in particular, the cells became much less sensitive to killing by high concentrations of Cu 2ϩ . In contrast, Yhb1p appears to be detrimental to cells exposed to H 2 O 2 : YHB1-deleted cells were more resistant, while YHB1 overexpression sensitized cells to H 2 O 2 . The response to diamide is less clear, although it was checked repeatedly in replicate experiments. Yhb1p appears to have both beneficial (at low concentrations of diamide) and detrimental (at high concentrations of diamide) effects on the resistance of the cells to this compound. Essentially the same correlations between the resistance of cells to these drugs and the cell Yhb1p contents were obtained with strain S150-2B (not shown). Increased sensitivity to diamide and diethylmaleate (a thiol oxidant), but not to H 2 O 2 , upon deletion of YHB1 has been reported (28). One possible explanation for this difference from our results is that cells in the rho Ϫ state were used, and these cells are much more sensitive that rho ϩ cells to oxidative stress FIG. 4. Effects of oxidants on YHB1 expression. A, expression of the fusion YHB1-lacZ in response to oxidative stress. S150-2B cells transformed with pHB-Z were grown overnight in glucose synthetic medium (SCglu) to OD 600 ϭ 0.4. Aliquots of the culture were then incubated for 1 h with 1 mg/liter antimycin A, 0.4 mM H 2 O 2 , 2 mM diamide, 2 mM menadione sodium bisulfite, 5 mM DTT, or 1 mM CuSO 4 , and ␤-galactosidase activity (A 420 nm ϫ 1000/min/OD 600 of cells) was assayed. B, Northern blot analysis of YHB1 mRNA levels in S150-2B and W303-1B cells growing exponentially in SCglu and exposed to the different drugs for 1 h as described in panel A. W303-1B cells were also grown in rich galactose medium (YPgal) and treated as described in Ref. 28. At OD 600 ϭ 1, the cells were collected by centrifugation, suspended in an equal volume of fresh YPgal containing 2.5 mg/liter antimycin A, 1 mM H 2 O 2 , 1.5 mM diamide, 5 mM DTT, or 1 mM CuSO 4 , and incubated for 1 h at 30°C. Total RNA were analyzed as described in the legend to Fig. 2. The YHB1/ACT1 ratios were normalized to 1 for expression in untreated cells.
FIG. 5. Effect of hemoglobin deletion and overexpression on the sensitivity of cells to oxidative stress. SCglu medium was inoculated at 2 ϫ 10 3 cells/ml from fresh exponential precultures and distributed as 10-ml aliquots in 50-ml flasks containing the indicated drugs at the concentrations shown. Following incubation at 30°C for 15-17 h (10 -12 generations), growth was monitored by OD 600 measurements for 6 h. Values were averaged and normalized to the OD of control cultures without oxidative agents. The whole experiment was repeated 3 times for confirmation. Data from a representative experiment are shown. q, control strain W303-1B transformed with pYPGE2; ⌬, strain W/⌬HB deleted of YHB1 and transformed with pYPGE2; Ⅺ, strain W303-1B transformed with pPGE/YHB1 overexpressing YHB1.
(Ref. 61 and references therein), thus precluding the detection of small, specific effects.
Last, we examined the effects of varying the amounts of Yhb1p on the growth of strains made more sensitive to oxidants by deleting the genes YAP1 (33,59,60) or ZWF1/MET19 encoding glucose-6-phosphate dehydrogenase (32,62). This enzyme catalyzes the first rate-limiting NADPH-producing step of the pentose phosphate pathway, which is considered to be an important source of reducing power for cells. The strains wildtype for YAP1, overexpressing it (from the plasmid YEp351/ YAP1 (33)), or YAP1-deleted, all grew about 10% more slowly (1.8 h) when overproducing Yhb1p from the plasmid pPGE/ YHB1, compared with 1.5 h for controls. This is within the range (7-14%) mentioned for strains S150-2B and W303-1B. But it was significantly larger for the mutant ⌬zwf1(met19), as its doubling time in glucose minimal medium (1.9 h) was 50% longer when overexpressing Yhb1p from pPGE/YHB1 (2.9 h); the increase was smaller (25%) for cells growing in glycerol ϩ ethanol medium, and dropped to ϳ10% in cells grown under anaerobic conditions. However, we can hardly conclude from this experiment that oxygen is required for the deleterious effect of large production of Yhb1p in this mutant, since less holoYhb1p is made under anaerobiosis.
Although our results clearly show a relationship between the cell Yhb1p content and their resistance to various stresses, they are complex, for the most part unexpected, and often contradictory. For example, one would not have expected Yhb1p to protect cells against damage caused by DTT, given that YHB1 expression drops in cells exposed to it. Also surprising are the contrasting effects of Yhb1p on the resistance of cells to H 2 O 2 and Cu 2ϩ . The relatively complicated situation, compared with that described previously (28), the paucity of information on the effects of DTT in cells (Ref. 57 (53)), all make it difficult to imagine just how Yhb1p can participate in the defense of cells against oxidative stress. This is all the more difficult because hemoproteins in general and hemoglobins in particular are known for their inherent ability to produce reactive oxygen species and to induce peroxide-mediated damage in various biological systems (52, 64 -67). Clearly, better knowledge of the properties of Yhb1p itself, and further phenotypic studies of YHB1-deleted and overexpressing strains are required to clarify the function(s) of this enigmatic ancient form of hemoglobin.