Molecular cloning, phylogenetic analysis and three-dimensional modeling of Cu,Zn superoxide dismutase (CnSOD1) from three varieties of Cryptococcus neoformans.

Cryptococcus neoformans (Cn), causal agent of fungal meningoencephalitis, has three varieties with variable host predilection. To explore mechanisms for these pathogenic differences, we have characterized Cu,Zn SOD gene (CnSOD1). A Saccharomyces cerevisiae sod1Delta mutant was complemented with Cn var. grubii yeast expression library. The complementing clone had an ORF of 462 bp and the deduced 154 aa sequence showed 61% identity with S. cerevisiae SOD1 and 53-65% with other eukaryotic SOD1s. Cn var. grubii CnSOD1 cDNA was used to clone corresponding cDNAs from var. neoformans and var. gattii. ORFs from three varieties revealed 20-29% differences in deduced aa (s) with a significant 6% non-synonymous aa substitution between Cn var. grubii and Cn var. gattii. Cosmid library screening and PCR cloning were used to obtain genomic SOD1, which was split by five introns with identical placements and a typical 5' splice junction sequence, GTNNGY. These introns also showed a large nt variation among the three Cn varieties. Phylogenetic analyses revealed CnSOD1 to be in a group distinct from other eukaryotic SOD1s and with a significant divergence of the var. grubii from var. gattii. The CnSOD1 -deduced protein was modeled based on the crystal structure of S. cerevisiae SOD1, which showed an excellent fit. Most of the non-synonymous aa substitutions occurred on the outside of the molecule and these may contribute to differences in antigenicity among the three varieties. Notably, Cn var. neoformans and var. gattii Cu,Zn SOD had three substitutions of glycine (Gly26, Gly92 and Gly123 for Asn26, Ser92 and Ser123) that may contribute to the observed lower thermostability of this enzyme vis-a-vis Cn var. grubii. This is the first nucleotide and structural comparison of a protein-encoding gene from the three Cn varieties, which may provide a framework for future studies on the role of Cu,Zn SOD in Cn pathogenesis.

Reciprocal regulation of anaerobic and aerobic cell wall mannoprotein gene expression in Saccharomyces cerevisiae.

The DAN/TIR genes encode nine cell wall mannoproteins in Saccharomyces cerevisiae which are expressed during anaerobiosis (DAN1, DAN2, DAN3, DAN4, TIR1, TIR2, TIR3, TIR4, and TIP1). Most are expressed within an hour of an anaerobic shift, but DAN2 and DAN3 are expressed after about 3 h. At the same time, CWP1 and CWP2, the genes encoding the major mannoproteins, are down-regulated, suggesting that there is a programmed remodeling of the cell wall in which Cwp1 and Cwp2 are replaced by nine anaerobic counterparts. TIP1, TIR1, TIR2, and TIR4 are also induced during cold shock. Correspondingly, CWP1 is down-regulated during cold shock. As reported elsewhere, Mox4 is a heme-inhibited activator, and Mot3 is a heme-induced repressor of the DAN/TIR genes (but not of TIP1). We show that CWP2 (but not CWP1) is controlled by the same factors, but in reverse fashion-primarily by Mot3 (which can function as either an activator or repressor) but also by Mox4, accounting for the reciprocal regulation of the two groups of genes. Disruptions of TIR1, TIR3, or TIR4 prevent anaerobic growth, indicating that each protein is essential for anaerobic adaptation. The Dan/Tir and Cwp proteins are homologous, with the greatest similarities shown within three subgroups: the Dan proteins, the Tip and Tir proteins, and, more distantly, the Cwp proteins. The clustering of homology corresponds to differences in expression: the Tip and Tir proteins are expressed during hypoxia and cold shock, the Dan proteins are more stringently repressed by oxygen and insensitive to cold shock, and the Cwp proteins are oppositely regulated by oxygen and temperature.

Regulatory mechanisms controlling expression of the DAN/TIR mannoprotein genes during anaerobic remodeling of the cell wall in Saccharomyces cerevisiae.

The DAN/TIR genes of Saccharomyces cerevisiae encode homologous mannoproteins, some of which are essential for anaerobic growth. Expression of these genes is induced during anaerobiosis and in some cases during cold shock. We show that several heme-responsive mechanisms combine to regulate DAN/TIR gene expression. The first mechanism employs two repression factors, Mox1 and Mox2, and an activation factor, Mox4 (for mannoprotein regulation by oxygen). The genes encoding these proteins were identified by selecting for recessive mutants with altered regulation of a dan1::ura3 fusion. MOX4 is identical to UPC2, encoding a binucleate zinc cluster protein controlling expression of an anaerobic sterol transport system. Mox4/Upc2 is required for expression of all the DAN/TIR genes. It appears to act through a consensus sequence termed the AR1 site, as does Mox2. The noninducible mox4Delta allele was epistatic to the constitutive mox1 and mox2 mutations, suggesting that Mox1 and Mox2 modulate activation by Mox4 in a heme-dependent fashion. Mutations in a putative repression domain in Mox4 caused constitutive expression of the DAN/TIR genes, indicating a role for this domain in heme repression. MOX4 expression is induced both in anaerobic and cold-shocked cells, so heme may also regulate DAN/TIR expression through inhibition of expression of MOX4. Indeed, ectopic expression of MOX4 in aerobic cells resulted in partially constitutive expression of DAN1. Heme also regulates expression of some of the DAN/TIR genes through the Rox7 repressor, which also controls expression of the hypoxic gene ANB1. In addition Rox1, another heme-responsive repressor, and the global repressors Tup1 and Ssn6 are also required for full aerobic repression of these genes.

Induction and repression of DAN1 and the family of anaerobic mannoprotein genes in Saccharomyces cerevisiae occurs through a complex array of regulatory sites.

The DAN/TIR mannoprotein genes of Saccharomyces cerevisiae (DAN1, DAN2, DAN3, DAN4, TIR1, TIR2, TIR3 and TIR4) are expressed in anaerobic cells while the predominant cell wall proteins Cwp1 and Cwp2 are down-regulated. Elements involved in activation and repression of the DAN/TIR genes were defined in this study, using the DAN1 promoter as a model. Nested deletions in a DAN1/lacZ reporter pinpointed regions carrying activation and repression elements. Inspection revealed two consensus sequences subsequently shown to be independent anaerobic response elements (AR1, consensus TCGTTYAG; AR2, consensus AAAAATTGTTGA). AR1 is found in all of the DAN/TIR promoters; AR2 is found in DAN1, DAN2 and DAN3. A 120 bp segment carrying two copies of AR1 preferentially activated transcription of lacZ under anaerobic conditions. A fusion of three synthetic copies of AR1 to MEL1 was also expressed anaerobically. Mutations in either AR1 site within the 120 bp segment caused a drastic loss of expression, indicating that both are necessary for activation and implying cooperativity between adjacent transcriptional activation complexes. A single AR2 site carried on a 46 bp fragment from the DAN1 promoter activated lacZ transcription under anaerobic conditions, as did a 26 bp synthetic AR2 fragment fused to MEL1. Nucleotide substitutions within the AR2 sequence eliminated the activity of the 46 bp segment. Ablation of the AR2 sequences in the full promoter caused a partial reduction of expression. The presence of the ATTGTT core (recognized by HMG proteins) in the AR2 sequence suggests that an HMG protein may activate through AR2. One region was implicated in aerobic repression of DAN1. It contains sites for the heme-induced Mot3 and Rox1 repressors.

Telomerase activity in benign and malignant cytologic fluids.

BACKGROUND: Telomerase is a ribonucleoprotein that maintains telomeric base pair repeats at the ends of mammalian chromosomes during DNA replication. Telomerase is expressed in various human tumors, some normal tissues, and immortalized cell lines. The assay of telomerase activity has potential as an adjunct for cancer detection in cytologic fluids. METHODS: Twenty-four unfixed cytologic fluids, including 13 ascitic fluids, 7 pleural fluids, 3 pelvic washings, and 1 bronchial washing, were prepared for polymerase chain reaction (PCR)-based telomeric repeat amplification protocol (Oncor, Inc., Gaithersburg, MD). Telomerase activity was determined by PCR. The presence of a ladder of products with 6 base pair increments, separated by polyacrylamide gel electrophoresis and detected by phosphoimaging, was considered a positive result. Results were compared with cytologic evaluation of alcohol fixed, Papanicolaou stained smears. RESULTS: Of the 14 cytologically malignant specimens, 11 (79%) contained detectable telomerase activity. Two cytologically malignant samples could not be evaluated for telomerase activity due to the presence of inhibitory substances of PCR reaction. Of the 10 cytologically negative specimens, 1 (10%) was positive for telomerase activity; this specimen was from a patient with history of both endometrial and lung carcinomas. CONCLUSIONS: Telomerase activity can be detected in malignant cytologic specimens and thus has potential as a diagnostic adjunct in cytopathology.

Telomerase activity in human benign prostate tissue and prostatic adenocarcinomas.

Telomerase adds a hexanucleotide telomeric sequence to the chromosomal ends during replication and is postulated to play a role in cellular senescence and immortalization. Thirty-four human prostate tissues (18 malignant; 16 benign) were analyzed for telomerase activity by a sensitive nonradioactive polymerase chain reaction (PCR)-based method using the TRAP-eze telomerase detection kit (Oncor, Inc., Gaithersburg, MD). Telomerase activity in the homogenized tissue extracts was correlated with tumor grade, pathologic stage, and DNA ploidy. Specimens that exhibited the 36 bp internal control band and a ladder of products with 6-base increments starting with 50 nucleotides were considered positive. Fourteen (78%) of 18 prostatic adenocarcinomas (PACs) and only 2 (13%) of 16 benign prostate tissues exhibited telomerase activity. Our results indicate that, in contrast to most benign prostate tissues, telomerase activity can be detected in the majority of PACs and appears to be independent of tumor grade, stage, or DNA ploidy. Telomerase expression is occasionally detected in benign prostatic tissues bordering PACs and may result from either the presence of undetected tumor foci in these stored specimens or the proliferative response of the benign elements to adjacent cancer.

Telomerase activity in benign endometrium and endometrial carcinoma.

Telomerase, a ribonucleoprotein associated with synthesis of telomeric DNA, is postulated to play a role in cellular senescence and immortalization. Telomerase adds a hexonucleotide telomeric sequence to the chromosomal ends during replication and is preferentially expressed in most malignant and germ-line tissues but is usually undetectable in normal somatic cells. In the current study, 34 human endometrial tissues (20 malignant and 14 benign) were analyzed for telomerase activity by a nonradioactive PCR-based method using the TRAP-eze telomeric repeat amplification detection kit (Oncor). Nineteen of 20 (95%) endometrial carcinomas and 8 of 8 (100%) benign endometrial tissues from premenopausal women exhibited strong telomerase activity, whereas 6 of 6 (100%) benign endometrial tissues from postmenopausal women showed only weak telomerase activity. There was no correlation of telomerase activity with tumor grade, depth of invasion, or DNA content. Benign cycling endometrium, a rapidly proliferating tissue, features positive telomerase activity, although expression in nonneoplastic tissues has only rarely been previously reported. Only weak activity is detected in endometrial tissues after menopause, but telomerase activity can be strongly reactivated in patients who develop endometrial cancer.

The DAN1 gene of S. cerevisiae is regulated in parallel with the hypoxic genes, but by a different mechanism.

The DAN1 gene is expressed under anaerobic conditions in yeast and completely repressed during aerobic growth. The function of the gene is unknown, and genetic disruption had no effect on fitness which could be detected, even upon prolonged anaerobic growth. Expression of DAN1 was constitutive in a heme-deficient strain, indicating that heme participates in repression. Expression was blocked by heme in anaerobic medium, suggesting that heme acts as a negative co-effector rather than through its metabolic functions, i.e., in the production of a co-effector. Expression of DAN1 was regulated in parallel with the hypoxic gene ANB1, showing identical kinetics of induction and dose response to heme. However, unlike ANB1, DAN1 is not regulated by the repressor of the hypoxic regulon, ROX1, as shown by observation of normal aerobic repression of DAN1 in a strain carrying a deletion of ROX1. These results indicate the existence of a parallel regulatory system which produces an identical response to oxygen by a different mechanism than that controlling the hypoxic regulon.

Transient adaptation to oxidative stress in yeast.

Adaptive responses to the oxidative stress of hydrogen peroxide (H2O2) were studied in the yeast Saccharomyces cerevisciae strain RZ53. Our results show that the growth of naive cells is readily arrested by H2O2 challenge. In contrast, cells that have been preexposed to relatively low H2O2 priming treatments (i.e., cells that have first been pretreated with low H2O2 concentrations) are able to survive a subsequent challenge dose and continue to divide at normal rates. The most effective adaptation was observed with the following conditions: 5 x 10(6) cells/ml at pretreatment, pretreatment or priming peroxide addition of 0.4 mM H2O2, interval between pretreatment and challenge of 45 min, challenge peroxide concentration of 0.8 mM H2O2 for 2 h. Under these conditions cells that were challenged without pretreatment exhibited a 90% loss of plating efficiency. In contrast, peroxide-pretreated cells grew and divided at rates that were actually 15-30% faster than those of nonpretreated cells, and some 90-100% of such pretreated cells continued to divide at normal rates even following exposure to the H2O2 challenge concentration. The increased H2O2 resistance of pretreated cells was transient, being readily reversed during 60-90 min of growth in the absence of H2O2. Furthermore, cells that were allowed to deadapt over a 4-h period again exhibited a transient adaptive response when reexposed to H2O2 pretreatment. These results, plus the high survival rates (90-100%) of H2O2 pretreated and challenged cells, demonstrates that our results represent a true transient adaptation, rather than a selection for any preexisting peroxide resistant subpopulation. H2O2 adaptation required protein synthesis as evidenced by studies with the translational inhibitor, cycloheximide. At least 21 proteins exhibited increased expression following H2O2 adaptation, while the expression of some 8 other proteins was decreased. Adaptation is now widely reported in bacterial strains and has also been observed in some mammalian cell lines. We propose that the basis for such adaptive responses rests in increased expression of genes that encode protective enzymes and repair enzymes.

The Rox1 repressor of the Saccharomyces cerevisiae hypoxic genes is a specific DNA-binding protein with a high-mobility-group motif.

The ROX1 gene encodes a repressor of the hypoxic functions of the yeast Saccharomyces cerevisiae. The DNA sequence of the gene was determined and found to encode a protein of 368 amino acids. The amino-terminal third of the protein contains a high-mobility-group motif characteristic of DNA-binding proteins. To determine whether the Rox1 repressor bound DNA, the gene was expressed in Escherichia coli cells as a fusion to the maltose-binding protein and this fusion was partially purified by amylose affinity chromatography. By using a gel retardation assay, both the fusion protein and Rox1 itself were found to bind specifically to a synthetic 32-bp DNA containing the hypoxic consensus sequence. We assessed the role of the general repressor Ssn6 in ANB1 repression. An ANB1-lacZ fusion was expressed constitutively in an ssn6 deletion strain, and deletion of the Rox1 binding sites in the ANB1 upstream region did not increase the level of derepression, suggesting that Ssn6 exerts its effect through Rox1. Finally, ROX1 was mapped to yeast chromosome XVI, near the ARO7-OSM2 locus.

Regulation of gene expression by oxygen in Saccharomyces cerevisiae.

The oxygen regulation of two broad categories of yeast genes is discussed in this review. The first is made up of genes regulated by heme, and the second is made up of genes whose regulation is heme independent. Heme-regulated genes fall into two classes: heme-activated and heme-repressed genes. Activation is achieved through one of two transcriptional activators, the heme-dependent HAP1 protein or the heme-activated, glucose-repressed HAP2/3/4 complex. Some of the properties and the DNA-binding sites of these activators are discussed. Heme repression is achieved through the action of the ROX1 repressor, the expression of which is transcriptionally activated by heme. Once ROX1 is synthesized, its function is heme independent. Evidence that ROX1 binds to DNA or is part of a DNA-binding complex is described. Factors which modulate the function of these regulatory proteins are discussed, and a schematic of heme activation and repression is presented. The mitochondrial subunits of cytochrome c oxidase are induced by oxygen in a heme-independent fashion. The translation of one, cytochrome c oxidase subunit III, is dependent upon three nucleus-encoded initiation factors. One of these, PET494, is itself translationally regulated by oxygen in a heme-independent fashion. The expression of at least four other mitochondrially encoded cytochrome subunits is dependent upon specific translation factors, raising the potential for translational regulation as a general mechanism. Finally, a number of anaerobic genes that show heme-independent, oxygen-repressed expression have been identified. These fall into two kinetic classes, suggesting that there are at least two different regulatory circuitries.

A yeast protein with homology to the beta-subunit of G proteins is involved in control of heme-regulated and catabolite-repressed genes.

The product of the Saccharomyces cerevisiae AER2 gene is responsible for maintaining repression of at least two distinct regulatory pathways: heme activation/repression and catabolite repression. Mutations in the gene caused an eightfold increase in the expression of the heme-activated CYC1 gene in the absence of heme, a substantial increase in the expression of the heme-repressed ANB1 gene in the presence of heme, and a 13-fold increase in the expression of the catabolite-repressed GAL1 gene in the presence of glucose. Lesser or no increases in the expression of these genes were observed under derepressed or activation conditions. The aer2 mutations also caused a large increase in CYC7 gene expression under all conditions; this gene is subject to heme activation/repression, as well as catabolite repression. The AER2 gene was cloned and the sequence determined. The large open reading frame contiguous with the transcript from the complementing region encoded a 713-amino acid polypeptide chain with extensive homology to the beta-subunit of G proteins. The sequence revealed that AER2 is the TUP1 gene. A deletion mutation was constructed and the null phenotype was the same as the original mutants. The aer2 null mutant was shown to have increased aerobic and anaerobic levels of RNA encoding the ROX1 repressor, normally expressed only aerobically and responsible for the aerobic repression of ANB1 expression. The increase in both ROX1 and ANB1 RNAs aerobically in this mutant suggests that the repressor is nonfunctional in the mutant.

A hypoxic consensus operator and a constitutive activation region regulate the ANB1 gene of Saccharomyces cerevisiae.

We have identified a consensus operator sequence, YYYATTGTTCTC, which mediates the repression imposed by the ROX1 factor upon the members of the hypoxic gene regulon, which includes ANB1, HEM13, COX5b, and CYC7. The members of the regulon were repressed with widely varying stringency, and the variation was correlated with the number and fidelity of operator sequences observed. ANB1 had two operators operating with unequal efficiency, each containing two copies of the operator sequence. Synthetic operator sequences introduced into an operator deletion were effective as monomers but much more so as dimers, consistent with cooperativity. The native operators both imposed ROX1 repression on the GAL1 gene, in either orientation, but the synthetic operators did not, indicating that the sequence context may be important. The repression and activation of ANB1 are independent spatially and functionally, since deletion of the operators did not reduce expression and since both the operator and activation regions functioned separately in the GAL1 UAS. The ANB1 UAS was constitutive, containing several elements distributed over a 300-bp region. There were two dT-rich segments, one of 51 bp and one of 165 bp, the latter capable of activating transcription by itself. Flanking segments containing GRF2 (REB1) and ABF1 (GF1) sites may contribute to activation but were not essential. The UAS showed a strongly preferred orientation.

ROX1 encodes a heme-induced repression factor regulating ANB1 and CYC7 of Saccharomyces cerevisiae.

The ROX1 gene encodes a product implicated in the regulation of heme-repressed and heme-induced genes in Saccharomyces cerevisiae. The gene has been cloned and shown to code for a 1.4-kilobase transcript. The cloned gene was used to construct a null mutant to determine the role of ROX1 in regulating the expression of several heme-regulated genes. Constitutive expression of ANB1 (a heme-repressed gene) was observed in the null strain, indicating that ROX1 codes for a repressor or a facilitator of repression. Enhancement of expression of CYC7 in the null strain indicated that the ROX1 factor is required for repression of CYC7 to its normal low level of expression, consistent with evidence that CYC7 has a hybrid heme-induced, heme-repressed regulatory mechanism. The null mutation had only a slight negative effect on expression of the heme-induced genes CYC1 and tr-1 (a heme-induced homolog of ANB1), suggesting that the ROX1 factor is not directly involved in their regulation despite the existence of an unusual rox1 mutation (rox1-a1) causing constitutive expression of this group. The respiratory competence of the null mutant indicates that ROX1 is not a respiratory factor. ROX1 expression was found to be induced by heme, indicating that the heme repression of ANB1 and its family is the result of a cascade in which heme induces a repression factor which keeps the family of heme-repressed genes inactive during aerobic growth. The rox1-a1 allele had earlier been shown to cause constitutive expression of the family of heme-induced respiratory genes. This allele was found to cause constitutive expression of the ROX1 transcript itself, indicating that ROX1 is in the major heme-induced regulon.

Elements involved in oxygen regulation of the Saccharomyces cerevisiae CYC7 gene.

The CYC7 gene of Saccharomyces cerevisiae encodes the minor species, iso-2, of the cytochrome c protein. Its expression is governed by two regulatory sequences upstream from the gene: a positive site which stimulates transcription 240 base pairs 5' from the protein-coding sequence (-240) and a negative site which inhibits transcription at -300. In this study, the nature of the positive site and its relationship to the negative site has been investigated. Expression of the CYC7 gene is weakly inducible by oxygen. This effect was greatly enhanced by the semidominant CYP1-16 mutation in the trans-acting gene CYP1. The weak oxygen regulation in wild-type cells and the enhanced induction in CYP1-16 mutants were found to be mediated through the positive site. A mutational analysis of this site implicated at least part of a tandem, direct repeat of 9 base pairs as essential for the functioning of this site. The relationship between the positive and negative sites was investigated by comparing the expression of the intact gene with that of derivatives lacking either one or the other site. The expression of the gene containing only the negative site was actually stimulated anaerobically, while the gene containing the positive site alone, although having higher expression aerobically than anaerobically, had higher anaerobic expression than did the intact gene. Thus, it appeared that the combination of the positive and negative sites suppressed anaerobic expression. A model which attempts to explain these properties of the two sites and account for the regulation of the expression of the intact gene is presented.

Negative regulation of the Saccharomyces cerevisiae ANB1 gene by heme, as mediated by the ROX1 gene product.

In Saccharomyces cerevisiae the anaerobic (oxygen-repressed) ANB1 gene and a group of aerobic (oxygen-induced) genes are coordinately regulated by the ROX1 gene. We report here that heme, known as an inducer of aerobic genes, also causes inhibition of ANB1 expression. Thus, in combination with the ROX1 gene product heme has an opposite effect on the expression of anaerobic and aerobic genes. Accumulation of ANB1 mRNA was sharply decreased in anaerobic cells grown in the presence of heme. This effect must operate at the level of transcription since heme also inhibited accumulation of CYC1 mRNA from an ANB1-CYC1 fusion. Heme precursors did not appear to function either as inhibitors or as activators. Oxygen itself also had no effect on transcription of ANB1. Repression by heme cannot be attributed to the respiratory competence conferred by heme since both ANB1 and the aerobic genes tr-1 and CYC1 were regulated normally in [rho 0] mutants. The results are consistent with a classical allosteric coeffector function for heme, although more indirect explanations are tenable. A role for the ROX1 gene product in transcriptional regulation can be inferred from the observation that there was no inhibition of ANB1 expression by heme in rox1 mutants. Judging from this epistasis the rox1 phenotype is not due to a defect in heme production; this would indicate that the ROX1 factor functions by mediating the effect of heme on transcription.

Oxygen regulation of anaerobic and aerobic genes mediated by a common factor in yeast.

The expression of a number of yeast genes is regulated by oxygen levels. While many of these are known to be induced in the presence of oxygen, we have described a gene, ANB1, that responds in the opposite fashion, being expressed only under anaerobic conditions. To identify genes involved in regulation of ANB1 and other oxygen-regulated genes, we selected mutations causing constitutive expression of ANB1, using a fusion of the ANB1 modulator segment to the CYC1 gene. A number of trans-acting mutations affecting a gene designated ROX1 caused constitutive expression of both the fused and wild-type genes, indicating that the ROX1 gene product operates through the ANB1 modulator sequence at the level of transcription. The mutant alleles of ROX1 fall into two phenotypic classes. The rox1-a class is semi-dominant, and the rox1-b class is recessive. One mutant, rox1-a1, is pleiotropic and causes constitutive expression of three oxygen-induced genes--CYC1, SOD (superoxide dismutase), and tr-1 (an oxygen-induced gene with homology to ANB1)--as well as constitutive expression of the oxygen-repressed ANB1 gene. Alleles of the rox1-b class cause constitutive expression of ANB1 but do not affect expression of the oxygen-induced genes tested. The pleiotropy of the rox1-a1 mutant indicates that the ROX1 gene product is involved in coordinate expression of both oxygen-induced and oxygen-repressed genes.

Modulator sequences mediate oxygen regulation of CYC1 and a neighboring gene in yeast.

Three transcripts from Saccharomyces cerevisiae--CYC1 mRNA (transcribed from the iso-I cytochrome c gene) and two RNAs of unknown function, designated tr-1 and tr-2-were identified by reverse Southern blot analysis and found to be regulated in response to oxygen. CYC1 mRNA and tr-1 accumulation occurred only in the presence of oxygen while tr-2 appeared only under anaerobic conditions. tr-2 was transcribed from a region approximately 1 kilobase 5' from the CYC1 coding sequence and in the opposite direction. tr-1 showed homology to the same region as tr-2 but was transcribed from elsewhere in the genome. Expression of tr-2 and CYC1 was observed to be normal in cells transformed with centromeric plasmids carrying the two genes. Mutant transforming plasmids were constructed in which a 400-base-pair region between tr-2 and CYC1 was either deleted or inverted. The deletion led to low-level nearly unregulated expression of both the CYC1 and tr-2 genes, suggesting that sequences upstream from both genes are important for their expression and regulation. The inversion mutation produced a reversed pattern of CYC1 regulation in which the mRNA was present in anaerobically grown cells but absent in the presence of oxygen, mimicking wild-type tr-2 regulation and suggesting that the CYC1 transcription unit is under the control of the translocated tr-2 modulator sequences. Models for the function of these modulators are discussed.