Vol. 11, No. 11

MOLECULAR AND CELLULAR BIOLOGY, Nov. 1991, p. 5639-5647

0270-7306/91/115639-09$02.00/0 Copyright © 1991, American Society for Microbiology

The ROX3 Gene Encodes an Essential Nuclear Protein Involved in CYC7 Gene Expression in Saccharomyces cerevisiae LYNNE S. ROSENBLUM-VOS,t LINDA RHODES,: CARLOS C. EVANGELISTA, JR., KEN A. BOAYKE, AND RICHARD S. ZITOMER* Department of Biological Sciences, State University of New York at Albany, Albany, New York 12222 Received 5 March 1991/Accepted 22 August 1991 The ROX3 gene was identified during a hunt for mutants with increased expression of the heme-regulated CYC7 gene, which encodes the minor species of cytochrome c in the yeast Saccharomyces cerevisiae. The rox3 mutants caused a 10-fold increase in CYC7 expression both in the presence and absence of heme, had slightly increased anaerobic expression of the heme-activated CYCI gene, and caused decreases in the anaerobic expression of the heme-repressed ANBI gene and the aerobic expression of its heme-induced homolog. The wild-type ROX3 gene was cloned, and the sequence indicated that it encodes a 220-amino-acid protein. This protein is essential; deletion of the coding sequence was lethal. The coding sequence for j-galactosidase was fused to the 3' end of the ROX3 coding sequence, and the fusion product was found to be localized in the nucleus, strongly suggesting that the wild-type protein carries out a nuclear function. Mutations in the rox3 gene showed an interesting pattern of intragenic complementation. A deletion of the 5' coding region complemented a nonsense mutation at codon 128 but could not prevent the lethality of the null mutation. These results suggest that the amino-terminal domain is required for an essential function, while the carboxy-terminal domain can be supplied in trans to achieve the wild-type expression of CYC7. Finally, RNA blots demonstrated that the ROX3 mRNA was expressed at higher levels anaerobically but was not subject to heme repression. The nuclear localization and the lack of viability of null mutants suggest that the ROX3 protein is a general regulatory factor.

The expression of many of the nuclear genes encoding respiratory functions in the yeast Saccharomyces cerevisiae is regulated in response to both the availability of oxygen and the presence of fermentable sugars. Positive oxygen regulation serves to minimize the expression of these genes, whose products play no role during anaerobic growth. Transcription of these genes is activated through the function of at least two different transcriptional activators, HAP1 and the HAP2/3/4 complex, both of which respond to heme (5, 7, 8, 25-27, 46), the biosynthesis of which is dependent on molecular oxygen as a reactant (20). Interestingly, there are oppositely regulated isoforms for some of the respiratory proteins; one of each pair is heme activated, and the other is heme repressed. It is likely that these heme-repressed isoforms represent hypoxic forms, proteins that function more efficiently in limiting oxygen. Such isoforms have been found for subunit V of cytochrome oxidase (10, 42), cytochrome c (7, 13, 16, 39, 53), and for reasons that are not yet clear, the translation factor eIF-5A (15, 16, 22, 35). The HEMB3 gene, which encodes an oxidase during heme biosynthesis, is also subject to heme repression, ensuring that enzyme levels rise as the levels of its substrate, molecular oxygen, fall (50, 51). Similarly, the ERGJJ gene, encoding cytochrome P-450, is heme repressed (44). Heme repression in all these cases is mediated by the ROX1 repressor, whose transcription is heme activated (10, 17, 18). Thus, in aerobic cells, heme accumulates and activates the expression of respiratory

functions as well as the repressor of hypoxic functions. In cells grown in limiting oxygen, heme levels fall, causing decreased transcription of the heme-induced genes, including the repressor; as a consequence, the hypoxic genes are transcribed. Catabolite repression is superimposed on heme regulation for many genes, serving to limit the expression of respiratory proteins. Thus, even at high oxygen tensions, most of the carbon flows through fermentation when cells are provided with glucose, resulting in the accumulation of ethanol, which may serve to toxify the environment for other microorganisms. The mechanisms of catabolite repression are not well characterized, although some of the components have been identified. The expression of both cytochrome c protein genes and the COX6 gene, encoding a subunit of cytochrome oxidase, is affected by mutations in the SNFJ (COX6 and CYCJ) (48), SSN6/CYC8 (CYCI, CYC7, and COX6) (32, 48), and TUPJICYC9 (CYCI and CYC7) (32, 52) genes, which also regulate the expression of the catabolite-repressed invertase and the galactose catabolic enzymes (36, 43, 47), suggesting a common link between catabolite repression among these systems. However, these regulatory genes seem to play a much more general role in regulating cellular functions. Mutations in SSN6 and TUPI render MATt cells sterile (14, 23, 34), and mutations in TUPI cause derepression of heme-repressed and heme-activated genes (52). Thus, these genes may be involved either in circuitry which processes the input from multiple signals and transmits them to the appropriate DNA-binding proteins or in mediating interactions between specific repressors and the general transcriptional machinery. The CYC7 gene provides one vehicle for identifying genes involved in heme regulation, since there is a strong positive

* Corresponding author. t Present address: Howard Hughes Medical Institute, University of Pennsylvania School of Medicine, Philadelphia, PA 19104. t Present address: Merck Research Laboratories, Rahway, N.J.

07065.

5639

5640

ROSENBLUM-VOS ET AL.

selection for its overexpression (3, 32). This gene encodes the minor species of cytochrome c, iso-2 (4), and is transcribed at low levels both aerobically and anaerobically because it contains only a weak heme activation site for HAP1 and a heme repression site for ROX1 (18, 26, 53). Thus, mutations in the HAP] gene that generate a protein with altered specificity and null mutations in ROX1 were, in hindsight predictably, isolated by causing CYC7 overexpression (18, 45). In addition, mutations in two of the abovementioned genes, SSN6 and TUPI, were identified during hunts for mutants with increased expression of the CYC7 gene (3, 32, 52). In the case of tup) mutations, overexpression appears to be in part due to the lack of activity of the ROX1 repressor (52). We have continued the search for genes involved in CYC7 regulation and report here the identification of a gene encoding an essential nuclear protein. The lethality of null mutations suggests that it also plays a wider regulatory role than a simple involvement in heme regulation or catabolite repression. MATERIALS AND METHODS Strains, growth conditions, and transformations. The rox mutants were isolated in aGH1 (MATa trpl-289 leu2-3,112 gallA152) and aLR1 (MATa trpl-289 his3-11,15, galJAl52). aGHl is congenic with aGH1. It was derived by generating diploids from aGH1 by transformation with a plasmid carrying the HO gene, isolating segregants that lost the plasmid, and then sporulating the diploid cells. lys2 derivatives of aGH1 and oaGH1 were derived by the method of Chattoo et al. (2). TP3-17Z (MATa ura3-52 his4-519 cycl::CYC7/lacZ) contained the CYC7/lacZ fusion from YCp7Z integrated into the CYCI locus, disrupting the CYCI gene. i182-248 (MATa trpl-289 ura3-52 his4-519 cycl::CYC7/lacZ), i182-363 (MATa trpl-289 ura3-52 his4-519 cycl::CYC71lacZ rox3-182) and i182-504 (MATa trpl-289 ura3-52 leu2-3,112 cycl::CYC7/ lacZ rox3-182) were derived from the diploid generated from a cross of TP3-17Z by aGH1-182 (aGH1 containing the rox3-182 mutation). The diploid dRZ68 was derived from a mating of RZ53-6 (MATa trpl-289 leu2-3,112 ura3-52 adel-100) and RZ49-4 (MATa trpl leu2-3,112 adel-100 gallA152). RZ53-6heml was derived from RZ53-6 by one-step gene replacement using a HindIII fragment containing the HEM) gene with an insertion of the PstI LEU2 fragment from YEp13 into the PstI site within the coding sequence. Genetic manipulation of yeast cells (38) and transformations (11) were carried out as described previously. S. cerevesiae was grown in YPD (2% glucose) or YPR (2% raffinose) for the preparation of RNA and 3-galactosidase assays as described previously (53). The general manipulations of Escherichia coli cells used for the recombinant DNA procedures have been previously described by Maniatis et al. (19). Plasmids. The vectors YCpSO (30); YEp13 (1); YEplacll2, YEplacl95, YCplac33, and YCplac22 (6); pBSM13+ (Stratagene) and YIpS (41); and the lacZ fusion vector pMC1871 (37) have been described previously. The CYC7/lacZ fusion, YCp7Z, has been described previously (53). The HO-containing plasmid YEpHO was obtained from R. Jensen and I. Herskowitz. pACTI contained the yeast actin gene (24). The original ROX3 clone was obtained from a yeast genomic library in YCpSO (30) by complementation and was designated YCpROX3. A map of this clone is presented in Fig. 2; in restriction sites referred to below, numbers in

MOL. CELL. BIOL.

parentheses are the positions of the sites (in kilobases) measured from the left border of the insert. The deletion derivatives shown in Fig. 2, AB, AC, and AP, were obtained by cleavage with BamHI, ClaI, and PvuII, respectively, and ligation at low dilution. The BamHI (5.2)-SphI (5.9) fragment was inserted into the homologous sites of YCp5O to generate YCpROX3BSp. YEpROX3H contained the 2.7-kb HindlIl (3.6 to 6.3) fragment inserted into the HindIII site of YEplac112. YEpROX3Sp contained the 2.2-kb SphI (3.7 to 5.9) fragment inserted into the SphI site of YEplacl95. YCpROX3Pv contained the PvuII (3.9 to 4.4) fragment inserted into the SmaI site of YCplac33. YIpROX3 contained the BamHI (5.2)-SalI (vector site) fragment inserted into the homologous sites of a YIpS derivative. The following fragments from YCpROX3 were inserted into compatible sites in the polylinker of pBSM13 + and resulted in the designated plasmids: HindlIl (3.6 to 6.3), pBSROX3H; BamHI (5.2)-DraI (6.0), pBSROX3BD; BamHI (5.2)-EcoRI (9.6), pBSROX3BR; and BamHI (5.2)-HindIII (3.6), pBSROX3BH. The ROX311acZ fusion was constructed in pBSROX3H by introducing an XhoI site into the ROX3 gene 1 bp after the last codon by site-directed mutagenesis (12) using the oligonucleotide 5'-GGAGGCTGGAGTCTCGAGCTTTACGCCA and then inserting the lacZ-containing Sail fragment of pMC1871 into the new XhoI site. The in-frame fusion junction was confirmed by sequence analysis. The HindIll fragment containing the fusion was subcloned into the HindlIl site of the yeast vectors YEp13 and YCplac22, generating the plasmids YEpR3Z and YCpR3Z, respectively. The null allele of ROX3 was constructed by replacing the pBSROX3H BglII (5.1)-SphI (5.9) fragment, which contains the ROX3 coding sequence (see Fig. 4), with the 2.2-kb XhoI-SalI LEU2 fragment from YEp13. The null construct was excised with HindIII, generating a fragment with ends homologous to the ROX3 locus, and transferred into the yeast genome by one-step gene replacement (31). Cloning the ROX3 mutant alleles. Recovery of the mutant alleles from the yeast genome was achieved by gap repair, using the plasmid YEpROX3H-X. This plasmid contained the ROX3 HindIll fragment in which the XhoI site had been placed at the last codon subcloned into YEplacll2. Digestion with XhoI and BglII excised the ROX3 coding sequence, and this gapped plasmid was transformed into yeast cells, selecting for the TRPI marker on the plasmid. Genomic DNA was prepared from transformants and used to transform E. coli cells to ampicillin resistance (41), thereby recovering those plasmids in which the gap had been repaired by gene conversion from the mutant ROX3 locus. Nucleic acid blots and enzyme assays. RNA blots using total cellular yeast RNA were carried out as described previously (53). The amount of hybridization was quantitated with a Betascan to determine the amount of radioactivity in each band normalized to an actin control. For Southern blots, genomic yeast DNA was prepared (41) and blots were carried out (40) as described previously. RNA probes were made by using the Stratagene kit and the recommended conditions. P-Galactosidase assays were carried out as described previously (29). All data presented represent the average activities for extracts prepared from several independent transformants. DNA sequence analysis and primer extension. The sequences of both strands of the ROX3 gene were determined by the dideoxy-chain termination method of Sanger et al.

ROX3 GENE OF S. CEREVISIAE

VOL . 1 l, 1991

(33). Synthetic oligonucleotides were used as primers. Primer extension was carried out with poly(A) RNA as described previously (9, 21). Immunofluorescence. Immunofluorescence was carried out with permeabilized cells as described previously (28). The primary antibody was mouse anti-3-galactosidase antiserum (Sigma); the secondary antibody was goat anti-mouse immunoglobulin G conjugated to fluorescein isothiocyanate (Cappel). Materials. Enzymes were purchased from Boehringer Mannheim, U.S. Biochemicals, and Stratagene and were used as recommended by the vendors. Nucleotide sequence accession number. The sequence of the ROX3 locus presented here is entered in the EMBL data bank under accession number X58300. RESULTS

Isolation of CYC7 overexpressing mutants. The CYC7 gene encodes the iso-2-cytochrome c protein, the minor cytochrome c species in aerobically growing yeast cells (4). This gene is expressed at low levels both aerobically and anaerobically because it contains regulatory elements for both heme activation via the HAP1 transcriptional activator and heme repression through the ROX1 repressor (18, 26, 46, 53). Mutations in either of these regulatory proteins or the factors that modulate their activities may cause increased CYC7 expression, and selections for CYC7 overexpressing mutants have been carried out previously (3, 32). However, with the exception of semidominant hap] mutants (3, 45), roxi mutants (18), and the cyc8/ssn6 (32, 36, 43) and cyc9ltupl (32, 52) mutants, the genes defined in these studies have not been characterized. For our studies of the heme activation/repression regulatory circuitry, we repeated this selection with a modification. To avoid excluding respiratory deficient mutants that might result from mutations in these pathways, we used a CYC71galK fusion vector (49) in a gall deletion strain and selected for growth on galactose rather than selecting directly for increased CYC7 expression. Seventy-seven mutants in two strains of opposite mating type were isolated, and genetic analysis revealed five major, unlinked groups. Two of these groups identified new alleles of the ROX1 (17) and TUPJ (52) genes; the others were designated ROX3, 5, and 6. Forty-five roxi mutants were isolated, making this the most abundant group, while the 15 rox3 mutants represented the second most abundant group. For this reason, rox3 mutants were chosen for further study. The effect of a typical rox3 mutation, rox3-182, on the accumulation of CYC7 mRNA is shown in Fig. 1. Increased RNA levels in preparations from cells grown on glucose (catabolite repressed) both in the presence (aerobically) and, even more dramatically, the absence (anaerobically) of heme were observed. Under catabolite-derepressing conditions, the effect of the rox3 mutation was less severe. To determine whether this mutation affected the expression of the hemeactivated CYC) gene and the heme-repressed ANB1 gene, the same RNA blot was probed for these transcripts. As seen in Fig. 1, the rox3 mutation had little visible effect on the expression of the CYC1 gene but caused a threefold decrease in the anaerobic expression of ANBI. The mRNA for the heme-activated ANBI homolog, transcript trl, is also visible on this blot because of cross hybridization with the ANB1 probe, and it showed a twofold decrease of aerobic expression in the rox3 mutant. Thus, the rox3 mutation caused increased expression of CYC7 under all growth conditions

WT -

R

02 N2

H-

rox3

R

02

-d

N2

S4

*

_

_-w

_ts *

-*

-

la

_

5641

YC7

CYCI

totrI ANBI

ACTI

FIG. 1. Effect of the rox3 mutation on RNA levels of hemeregulated genes. RNA blots with total cellular RNA prepared from cells grown aerobically in 2% raffinose (R) or 2% glucose (02) and anaerobically in 2% glucose (N2) were performed. The RNAs for which the probe was specific are indicated to the right of the resulting autoradiograph. WT and rox3 represent RNA prepared from wild-type (aGH1) and rox3-182 (aGH1-182) cells, respectively. These panels represent a single blot that was sequentially probed for each set of RNAs.

but caused a decrease in the activated expression of both ANB1 and trl. Isolation of the ROX3 gene. The wild-type allele of the ROX3 gene was isolated by complementation of the rox3-182 mutation. For this purpose, a strain was constructed containing the rox3-182 allele and the CYC71acZ fusion integrated at the CYCI locus. This strain gave dark blue colonies on 5-bromo-4-chloro-3-indolyl-o-D-galactopyranoside (X-gal) plates, reflecting the increased CYC7 expression resulting from the rox3 mutation. Cells were transformed with a yeast library constructed in the URA3-containing centromeric vector YCp5O, and the resulting Ura+ transformants were replated on X-gal plates. Pale blue colonies were picked, and four contained identical plasmids capable of transforming rox3-182 cells to the wild-type CYC7 phenotype. A restriction map of the yeast insert in the complementing clone is presented in Fig. 2. Deletions within the insert and subcloning localized the complementing region to an 800-bp region within the BamHI and SphI sites (Fig. 2). To confirm that the complementing DNA contained the ROX3 gene, the URA3 gene was integrated into the genomic locus represented by the putative ROX3 clone, and this marker was then mapped with respect to the original rox3 mutation. For this approach, the plasmid YIpROX3, which contained the URA3 gene and a 5.2-kb BamHI-Sall fragment containing the ROX3 complementing region, was constructed. The plasmid was cleaved at the unique ClaI site within the cloned sequences which directed integration to the cloned locus in i182-248, a ROX3 wild-type strain containing a ura3 mutation and the CYC711acZ fusion integrated at the CYCJ locus. The proper integration event was confirmed by Southern analysis (data not shown). This strain was mated with i182-504, which contained the rox3-182 allele, a ura3 mutation, and the CYC7/lacZ fusion integrated at the CYC1 locus. Fifteen tetrads were dissected and tested for the segregation of the ROX3 phenotypes by using the colony color assay on X-gal plates. The validity of the color

ROSENBLUM-VOS ET AL.

5642

E

YCpROX3

X G

ETG

CH

GAG

T

)II

.1.... \I C B CO

NA

E S

COMPLEMENT

I(

.j

+

--YCpROX3P--

YCpROXK3C

_

-

--

..

-

-

-

-

+

w.

... ...

-

-

.. ':;.;'..:-.'..; '. . '. '. EXt'. . ....

Subclones YCpROX3BSp

3

. _i _|.

Deletions

YCpROX3LB

.:- .

2

3

2

H

HSp P B PSpH H H CP

MOL. CELL. BIOL.

:,

s

S

'.

..

H-H~

r.: .:sffl

YEpROX3Sp

e. .:

+

_

YEpROX3H lKb

FIG. 2. Complementation analysis of the ROX3 locus. The top line represents the region of plasmid YCpROX3 carrying the yeast insert. The thin line represents the insert, and the thick lines represent plasmid sequences. The various deletions and subclones of this plasmid are represented below, and the ability (+) or lack of ability (-) of these plasmids to complement the rox3-182 mutant is indicated on the right. The symbols for the restriction sites are as follows: A, XbaI; B, BamHI; C, ClaI; E, EcoRI; G, BglII; H, Hindlll; N, NcoI; P, PvuII; S, Sall; Sp, SphI; T, PstI; X, XhoI. assay was confirmed by carrying out on all the haploid segregants from

P-galactosidase assays

five tetrads (data not shown). All the tetrads showed the 2:2 segregation of rox3 ura3:ROX3 URA3 expected if the cloned sequences were allelic with ROX3. Thus, the findings that the cloned DNA on a low-copy-number plasmid complemented the rox3 mutation and that the clone was derived from a locus tightly linked to the ROX3 gene strongly indicate that the clone contains the wild-type ROX3 allele. ROX3 transcript. The complementation analysis indicated that the ROX3 gene was contained within the BamHI-SphI fragment (Fig. 2). A single transcript of 900 nucleotides hybridized to probes from this region (data not shown). This RNA was transcribed from left to right on the map in Fig. 2 as determined by hybridizing single-stranded RNA probes to yeast RNA; only the probe synthesized from the DraI (800 bp rightward from the BamHI site) to the BamHI site hybridized, as seen in Fig. 3. Sequence of the ROX3 gene. The 1.4-kb sequence of DNA spanning the complementing region was determined and is presented in Fig. 4. An open reading frame beginning with a methionine codon and extending 220 codons in the correct orientation to be contained within the transcript visualized from this region was found, but the 5' end extended beyond the BamHI site (at bp 52), which comprised one border of the complementing region. If the first methionine codon 3' to this BamHI site were the true initiator, then a protein of only 121 residues would be synthesized. The presence of the open reading frame and the length of the transcript suggested that a 220-amino-acid protein was encoded by this region, while the complementation data using several different constructs containing the fragments extending from the BamHI site (Fig. 2) indicated that the ROX3 protein might be smaller. A number of experiments presented below were carried out to resolve this question. First, a null mutation was constructed to determine whether a complete deletion of the entire open reading frame would give the same complementation pattern as the uncharacterized rox3-182 mutation. Second, the 5' end of the transcript was mapped to determine whether it could encode the larger protein. The results of these experiments and further analyses of other rox3 mutants are

6ii~ T B

T

B T B Probes FIG. 3. Direction of transcription of the ROX3 gene. RNA was prepared from aGH1 cells grown aerobically in 2% glucose and subjected to Northern (RNA) analysis. Duplicate blots were prepared with 8 ,ug (lane 1) and 20 ,ug (lane 2) of RNA and 1 ng of pBSROX3H plasmid DNA as a control (lane 3). The probes were prepared from the plasmid pBSROX3BD either by digestion with PstI (the site of which is adjacent to the DraI end of the insert) and transcription with T7 RNA polymerase (left panel) or by digestion with BamHI and transcription with T3 RNA polymerase (right panel). The 0.8-kb ROX3 insert in this plasmid is indicated by the thick line. T, PstI; B, BamHI.

presented below and clearly implicate the larger open reading frame as encoding the ROX3 protein. This putative protein is highly charged and quite basic in the carboxyl-terminal region. It contains no significant homology with any protein or open reading frame in the GenBank library, as determined by scanning with the tfasta program in the Wisconsin package. ROX3 encodes an essential product. To determine the phenotype of a cell completely lacking the ROX3 gene product, we attempted to construct a null allele by a one-step gene replacement (31). A plasmid was constructed in which the region from the BgIIl site at -58 to the SphI site at 839 (176 bp 3' to the end of the coding sequence) was replaced with the LEU2 gene. Several attempts to displace the genomic wild-type ROX3 gene with this construct in haploid cells failed, suggesting that the gene might be essential. To test this hypothesis, the null allele was transformed into a diploid cell homozygous for the Ieu2 mutation and for the wild-type ROX3 gene, and a ROX31rox3::LEU2 heterozygote was isolated. Southern analysis confirmed the correct construct (data not shown). This diploid was sporulated, and the resulting tetrads were analyzed with respect to the segregation of the Leu phenotype. From a total of 29 tetrads, 21 contained only two viable spores, 7 contained one viable spore, and 1 contained three viable spores. In every case, the colonies derived from the viable spores were Leu-; no viable spores containing the rox3::LEU2 allele were observed (Table 1). The original diploid was also heterozygous for the URA3 and GAL) markers, and these segregated as

ROX3 GENE OF

11, 1991

VOL.

-511 AC

S.

AT

T

No. of colonies

Haploid segregantb

A60T

-420

=m60TCCAGGCT66CTTA60

CTTC6TCTGATGTGACG

lCCC

-360

CAAAMCGACCTTTC CT6TTCtTTTTTTTI

CM

T

rox3::LEU2 2b 2c 7c 9a ROX3 Sb 9b

CTA

-300

AACAATCTGGAAAGATAC AICAAAT -240 AGCCTGTTTGlTCTCGTC6

TCTTCAMTTCTMCTM

TAGAT6CATGATTGTCTC TTTTGATAGAGACMTMAT

-15

AGACGGAAMGGAAAAAAMAA

EAGCAAATT AGAAA#4-AEATGTATT

-120 -60

T U^C AACMGGGT60 CATTCCFTTTCTGATCFGAC

TlTT-

T60TTAATAGGCCCTT

BgU 1|9111

ATAGATCTGTAGTGAGTGCG MATCAATATAAGTGMC 1

T

A

48

GCT TCT MA GTG GC GM ACT MA GTC CCC TCA TAC TAC TAT TAC Net Ala Ser Arg VaL Asp GLu Thr Thr VaL Pro Ser Tyr Tyr Tyr Tyr 16 ATG

BaII

5643

TABLE 1. Plasmid loss in a rox3 null mutanta

ACTCT

-480

60TCCGCiTGGCTGAATTGT

CEREVISIAE

96

GTG GAT CCG GMA ACT ACA TAT ACG TAC CMA CMA CCA MT CCT CTA CAG Val Asp Pro Giu Thr Thr Tyr Thr Tyr Gln Gin Pro Asn Pro Leu Gln 32

Trp+

Trp-

50 50 50 50

0 0 0 0

29 41

21 9

a Plasmid loss was determined by growing lightly inoculated 25-ml overnight cultures on nonselective medium (YPD), plating the cells on YPD plates, and then testing individual colonies for the Trp phenotype. b The haploid segregants were derived from dRZ68-1, a ROX31rox3::LEU2 heterozygote, homozygous for mutations in TRPI, LEU2, and URA3.

144

GAC TTG ATA TCG GTG TAT GGC TTG GAT GAC ATC TCC AGG CA GTG 6CA Asp Leu Ile Ser VaL Tyr Gly Leu Asp Asp Ile Ser Arg Gin VaL ALa

48

192 TAC

AGA ACA AAT TTG GAC 6GC ACT AM 6CC GTG MG CTA AGA AMA TCT Arg Thr Asn Leu Asp GLy Thr Lys ALa VaL Lys Leu Arg Lys Ser Tyr 240 MAG AAC CAG ATA GCA GAT CTT TCA 66T AMA TTC TCC ACC ATA CCG ACC Lys Asn Gin Ile Ala Asp Leu Ser GLy Lys Phe Ser Thr Ile Pro Thr 288 AMA GMA AAT GGT MAA 6T GGT CA ATA GCA CAT ATT CTT TTC CAA AAT Arg GLu Asn Giy LyV GLy Giy Gin Ile Ala Nis Ile Leu Phe Gin Asn 336 AC CCA GAC ATG ATG ATA CA CCA CCT CAG CMG 6T CM AC ATG TCA Asn Pro Asp Net Net Ile Gin Pro Pro Gin Gln GLy Gin Asn Net Ser

64

80 96 112

384

GAG CMA CMA TGG CGC GAM CAG CTG CGC AAT AGA GAC ATA GCA TTG m Gtu Gin Gin Trp Arg Giu Gin Leu Arg Asn Arg Asp Ile Aia Leu Phe 128

*-202 CAG CCT CCA AMT TTC GAT TGG GMC Gin Pro Pro Asn Phe Asp Trp Asp *-182 m GMA AGG TCA TAT CCA AGC GAG Phe Gtu Arg Ser Tyr Pro Ser Giu

432 CTT T6C TCT TCG GTA CTA TCG CAG Leu Cys Ser Ser Vat L.u Ser Gin

14

40

TTC 6CA AAC CAM AAC CMA G6 GA Phe Ala Asn Gin Asn Gin Gly Gty 160

528 6CC CA GCG CCG m GAT ATA GAC GAC TTG 6CG m GAT CTA GMC 6GT Ala Gin Ala Pro Phe Asp lie Asp Asp Leu ALa Phe Asp Leu Asp Giy 176 576 AC GGA MA AGC CA TCC GGC TCA AAT TCA GGT AAC AAT T MAG MAA Thr Gty Lye Ser Gin Ser Giy Ser Asn Ser Giy Asn Asn Ser Lys Lye 192 624 AGG MG AAC A TCT AGT GGA AGT TCG ATG GCT ACA CCAACA CAT AT Arg Lye Asn Lye Ser Ser Gly Ser Ser Net Ala Thr Pro Thr Nis Ser 208 674 GAC AMT CAT GAG 6AT ATG AM AGA AG MGG CTG GAG TAM CTTTGCCAT Asp Ser lNis Gu Asp Net Lys Arg Arg Arg Leu Giu 734 CGCATATCACCCCATATA MATAAMGTGCAT 6GGTGWATTTGT6CT

794 TATATATGTAACTTTMTT GTATAGMTATTATTAA a54 SphI CM60MTA60CTATTTTGA GCAGGCATGCTTTTCAM ATATCAAGCATMAAAA

TG;TAAGTGCA

914 TAMGAACMGTAGTC AAGATAMTTGG

TGGCATTTTTTGCCATC

FIG. 4. Sequence of the ROX3 locus. The sequence of the coding strand of the ROX3 locus is presented. The bases are numbered with the A in the translational initiation codon of the long open reading frame as 1; bases toward the 3' end are designated as positive integers, and bases toward the 5' end are designated as negative integers. The amino acid residues are numbered to the right of each line. The BglII and SphI sites deleted in the null mutant are indicated, as is the BamHI site. The sites of the point mutations in the rox3-182 and 202 alleles are each indicated by an asterisk below the mutated base; in both cases, the C was changed to a T, generating a nonsense codon.

Since the original rox3 mutations affected the expression of heme-regulated genes, we investigated the possibility that the failure of rox3 null alleles to germinate was due to a defect in heme biosynthesis, which would result in a requirement for sterol. We attempted to germinate spores on medium supplemented with Tween 80 and ergosterol. Each of the nine tetrads analyzed in this manner gave only two viable spores, both of which were Leu-. Thus, the inability to obtain spores with the null allele was not due to sterol auxotrophy. To determine whether the rox3 null allele would prevent germination or would not support vegetative growth, the heterozygote containing the null allele was transformed with the plasmid YEpROX3H (containing the TRPJ selectable marker); spores containing this plasmid and the rox3::LEU2 allele germinated. By using a haploid strain derived in this manner, the stability of the plasmid was tested. As shown in Table 1, after a long period of nonselective growth, Trpauxotrophs were isolated from the ROX3 wild-type transformant, but no such auxotrophs could be isolated from the rox3::LEU2 mutant, indicating that ROX3 is essential for vegetative growth on rich medium. This same system was used to test the ability of the incomplete ROX3 gene contained within the BamHI subclones described above to complement the null allele. Cells containing the rox3::LEU2 null allele and the YEpROX3H plasmid were transformed with a second URA3-CEN4 plasmid containing either the original ROX3 clone or the BamHISphI fragment capable of complementing the rox3-182 allele (Fig. 2). Several independent transformants carrying both the URA3 and TRPI plasmids were grown and plated on nonselective medium, and individual colonies were then tested for the presence of the plasmids. As seen in Table 2, segregants lacking the TRPI plasmid (Trp-) were obtained TABLE 2. Plasmid complementation of the rox3 null allelea TRPI TRPI

expected; approximately equal numbers of Ura+ and Uraand Gal' and Gal- spores were obtained. The original diploid from which the ROX3 heterozygote was obtained was sporulated, and six tetrads were analyzed; five of them gave four viable spores, and the sixth gave three viable spores.

No. of colonies

Plasmid

YEpROX3H YEpROX3H

URA3

rTrp+ ~Ura+

Trp+ Ura-

TrpUra4

Ura-

YCpROX3 YCpROX3BSp

11 65

1

25

78 0

0 0

Trp-

a Complementation was measured in the rox3::LEU2 mutant 9a (see Table 1). Cells carrying two plasmids were grown under nonselective conditions and then tested for the frequency with which cells lacking either or both plasmids segregated out, as described in footnote a of Table 1.

5644

MOL. CELL. BIOL.

ROSENBLUM-VOS ET AL. GA

R

ATG.

B-

FIG. 5. Mapping of the 5' end of the ROX3 RNA. Primer extension of a synthetic oligonucleotide complementary to the ROX3 coding strand from bases 333 to 352 and labeled with 32P at the 5' end was carried out with poly(A) RNA prepared from anaerobically grown cells. The figure shows the autoradiograph of the denaturing gel, in which lane R contains the primer extension products and lanes G and A contain the dideoxy G and A sequencing reactions carried out with the same primer. B indicates the position of the BamHI site within the coding sequence (see Fig. 4), and the position of the first ATG in the open reading frame is shown.

at a high frequency in cells containing the YCpROX3 plasmid. Clones lacking YCpROX3 (Ura-) appeared less frequently, presumably because of the presence of CEN4. As

expected, no clones lacking both plasmids were obtained. On the other hand, no Trp- clones were obtained from cells carrying YCpROX3BSp, indicating that this plasmid could not complement the inviability of the rox3 null allele. These results demonstrated that while the BamHI-SphI fragment was capable of complementing the rox3-182 allele, it could not complement the null allele and therefore must not contain the intact gene. The 5' end of the ROX3 transcript. The 5' end of the ROX3 transcript was mapped by primer extension. To obtain sufficient material, poly(A) RNA was prepared from cells transformed with YEpROX3Sp, a ROX3-containing multicopy plasmid. The transcript synthesized from this plasmid is identical in size to that in untransformed cells (data not shown). A primer complementary to a region 3' to the methionine codon downstream from the BamHI site was used to ensure that if this AUG codon were the initiator, the 5' end could be mapped. The result of this experiment is shown in Fig. 5 and clearly indicates that the ROX3 transcript began approximately 60 nucleotides upstream from

the first AUG of the large open reading frame. (Because the purpose of this experiment was to determine whether the 5' end of the mRNA was upstream from the first AUG 3' to the BamHI site or 5' from it, the primer was made complementary to a region well within the coding sequence, and the resulting cDNA was too long to map to a single nucleotide.) This result, combined with the requirement for a fragment containing the entire open reading frame for complementation of the null allele, strongly suggest that this 220-codon open reading frame encodes the ROX3 gene. Nature of the rox3 mutations. Because the ROX3 gene product was essential and the rox3-182 mutation was complemented by a fragment of the coding sequence, we cloned and sequenced two of the original rox3 mutations, hoping to obtain some insight into the functional domains of the gene. The rox3-182 allele contained a CG-to-TA transition at bp 385 which changed codon 129 from a glutamine codon to a nonsense codon (Fig. 4). A second allele, rox3-202, was found to have a CG-to-TA transition at bp 355 which also generated a nonsense codon; this allele encoded a protein containing only the first 117 amino acids of the putative ROX3 protein. Cells carrying these mutant alleles are clearly viable; therefore, the shortened forms of these proteins are probably synthesized and are at least partially functional. The possibility that the original strains harbored a nonsense suppressor is unlikely, since these two mutant alleles were isolated in different strains. Nonetheless, to formally rule this out, the rox3-202 mutant was mated with a strain carrying a number of amber suppressible markers, and no suppression was observed in the diploid. Finally, to rule out the possibility that a low level of translational readthrough of the nonsense mutations was responsible for the viability of these mutants, a plasmid, YCpROX3Pv, was constructed. This plasmid contains the ROX3 sequences from -432 to 357; thus, it lacks the coding sequence beyond codon 119. When transformed into the strain carrying the chromosomal null allele and a plasmid carrying the wild-type gene, isolates lacking the wild-type gene could be obtained, demonstrating that the amino-terminal half of the coding sequence is all that is required to support growth. It should be noted that the rox3-182 mutation was complemented by the BamHI-SphI fragment containing only part of the ROX3 gene. Given that this mutant probably synthesized only the amino half of the protein and that the BamHI-SphI fragment could not support vegetative growth and was missing part of the amino terminus, these results indicate that intragenic complementation occurred. ROX3 encodes a nuclear protein. To verify that the open reading frame encodes an authentic yeast protein and to determine the cellular localization of that protein, the lacZ coding sequence was fused in frame to the 3' end of the ROX3 coding sequence. ,3-Galactosidase activity in cell extracts was detected (Table 3), indicating that the open reading frame could be translated in yeast cells and, combined with the sequence data and the transcript mapping data, strongly implicating the 220-amino-acid protein as the product of the ROX3 gene. Antibody against ,-galactosidase was used for in situ immunofluorescence to determine the location of the fusion protein. 4',6-Diamidino-2-phenylindole (DAPI) counterstaining was used to visualize the nuclei, and as can be seen in Fig. 6, the immunofluorescence was localized to the same region as the DAPI stain for cells carrying the ROX3/lacZ fusion plasmid (Fig. 6a and c), while in cells carrying the vector alone, no specific immunofluorescence was observed

VOL. 11, 1991

ROX3 GENE OF S. CEREVISIAE

TABLE 3. Effect of heme and ROX1 on ROX3 expression Straina

RZ53-6 RZ53-6heml

R

02

5645

N2

P-Galactosidase activity' Aerobic

Anaerobic

0.66 0.68

4.1 3.8

a Cells were transformed with YCpR3Z, a centromeric plasmid containing the ROX311acZ fusion. I Cells were grown aerobically and anaerobically, and 0-galactosidase assays were performed as described in Materials and Methods. Anaerobic cultures and the aerobic culture for the hem] strain were supplemented with 0.2% Tween 80 and 20 ±g of ergosterol per ml. The enzyme assays were performed with three different independent transformants for each strain, and the error was less than 20%o.

(Fig. 6b and d). Thus, the ROX3 region directed the fusion protein to the nucleus. Regulation of ROX3 gene expression. The expression of the ROX3 gene was investigated. RNA blot analysis indicated that the mRNA accumulated at slightly higher levels in anaerobically grown cells (Fig. 7). This finding was confirmed by p-galactosidase assays carried out with extracts from cells containing the ROX311acZ fusion plasmid (Table 3). To test whether heme played a role in the oxygen repression of ROX3 expression, ,-galactosidase assays were carried out on extracts of hem] (deficient in heme biosynthesis) mutant transformed with the ROX31lacZ fusion plasmid contained on a centromeric plasmid. The results presented in Table 3 clearly indicate that the low aerobic expression of the ROX3 gene is not a result of heme repression. These results have been confirmed by RNA blots (data not shown). It is not clear what role this regulation plays in ROX3 function. When a multicopy plasmid carrying the ROX3 gene was placed in wild-type cells, the RNA accumulated at high levels under both aerobic and anerobic conditions but there was no effect on the accumulation of the CYC7, ANB1, or trl transcripts (data not shown). Thus, if the ROX3 protein also accumulated at higher levels in the aerobically grown trans-

FIG. 6. Cellular localization of the ROX31lacZ fusion protein. Antiserum directed against 3-galactosidase was used to perform in situ immunofluorescence with aGH1 cells grown aerobically in 2% glucose. The cells were also counterstained with the nucleusspecific stain DAPI. (a and b) The same field photographed to specifically reveal the immunofluorescence and DAPI staining, respectively, of cells transformed with YEpR3Z, the ROX31lacZ fusion; (c and d) the same field photographed to specifically reveal the immunofluorescence and DAPI staining, respectively, of cells transformed with the control plasmid, YEp13.

ROX 3

_~E

ACT 1

Ama.

AL-.&-,

CYCI

FIG. 7. Regulation of ROX3 RNA levels. RNA blots with RNA prepared from aGHl cells grown as described in the legend to Fig. 1 were performed. The 0.8-kb DraI-BamHI fragment containing the ROX3 gene was used as a probe for ROX3 RNA. The same blot was subsequently probed for CYCI and ACT) to confirm that equal amounts of RNA were loaded on the gel and that anaerobic conditions were achieved.

formants, it did not alter aerobic expression of the ROX3 target genes. DISCUSSION We describe here the initial characterization of the ROX3 gene. Mutations in this gene were identified as causing increased expression of the CYC7 gene, and the wild-type allele was cloned by complementation. The protein-coding region was identified by a combination of complementation and transcript mapping, and we demonstrated that the open reading frame in this region was translated in yeast cells by the use of a lacZ fusion. This fusion protein was found to be localized in the nucleus, strongly implying that the native ROX3 protein functions there. The protein has no significant homology to any other protein in the GenBank data base. While the results presented here do not precisely define the function of the ROX3 protein, some inferences can be drawn. ROX3 is localized in the nucleus, and viable mutations affect the expression of some genes. These findings suggest that the ROX3 protein is a transcription factor. Viable rox3 mutations affect the expression of different genes in opposite ways; CYC7 expression is increased, while that of the oppositely regulated eIF5A homologs are decreased. These results imply that ROX3 does not function in a simple way in heme regulation, nor does it function solely as a repressor or activator of transcription. This last conclusion is also suggested by an inspection of the ROX3 protein sequence, which contains none of the motifs associated with DNA-binding proteins or transcriptional activators. In addition, unpublished gel retardation studies using wild-type and rox3 mutant extracts revealed no size differences in the complexes formed on the CYC7 ROX1 or HAP1 binding sites, further suggesting that ROX3 is not a sequencespecific DNA-binding protein. These conclusions point to a role for ROX3 as a general transcription factor, perhaps involved in mediating responses of the general transcription machinery to a family of regulatory proteins. Experiments are in progress to explore this possibility. One of the more intriguing results obtained in this study is

5646

ROSENBLUM-VOS ET AL.

the complementation data. Although the ROX3 gene product is essential, the two point mutations sequenced contained nonsense codons halfway through the coding sequence, indicating that a protein containing only the first 118 (rox3-202) or 128 (rox3-182) amino acids is capable of carrying out the essential function of ROX3. Also, a deletion construct lacking the coding sequences 3' from codon 119 supported growth of a null mutant, thus eliminating the possibility that very low levels of translational readthrough were responsible for the viability of the nonsense mutations. At the same time, a variety of BamHI subclones which do not contain the 5' end of the ROX3 coding sequence (Fig. 4) complemented the rox3-182 mutant phenotype but did not alter the inviability of a rox3 deletion. The nature of the protein synthesized from this fragment is not known; the first in-frame methionine codon 3' to the BamHI site is at codon 100, but we cannot rule out the possibility that vector initiation sites were used or that a low level of initiation at a non-AUG codon occurred. Nonetheless, these results clearly indicate that intragenic complementation occurred and that either the ROX3 protein has distinct domains that can function independently or a functional multimeric unit can be assembled from the truncated peptides. These possibilities are under investigation. ACKNOWLEDGMENTS We thank Paula Wick and Marjorie A. Zitomer for their technical assistance. This research was supported by an NIH grant to R.S.Z. and by the Sterling Research Group.

REFERENCES 1. Broach, J. R., J. N. Strathern, and J. B. Hicks. 1979. Transformation in yeast: development of a hybrid cloning vector and isolation of the CAN] gene. Gene 8:121-133. 2. Chattoo, B. B., F. Sherman, D. A. Azubalis, T. A. Fjellstedt, D. Mehnert, and M. Ogur. 1979. Selection of lys2 mutants of the yeast Saccharomyces cerevisiae by the utilization of a-amino adipate. Genetics 93:51-65. 3. Clavilier, L., G. Pere-Aubert, and P. P. Slonimski. 1969. Mise en evidence de plusieurs loci independants impliques dans la synthese de l'iso-2-cytochrome c chez la levure. Mol. Gen. Genet. 104:195-218. 4. Downie, J. A., J. W. Stewart, N. Brochman, A. Schweingruber, and F. Sherman. 1977. Structural gene for iso-2-cytochrome c. J. Mol. Biol. 113:369-384. 5. Forsburg, S. L., and L. Guarente. 1989. Identification and characterization of HAP4: a third component of the CCAATbound HAP2/HAP3-responsive heteromer. Genes Dev. 3:11661178. 6. Gietz, R. D., and A. Sugino. 1988. New yeast-Escherichia coli shuttle vectors constructed with in vitro mutagenized yeast genes lacking six-base pair restriction sites. Gene 74:527-534. 7. Guarente, L., and T. Mason. 1983. Heme regulates the transcription of the CYCI gene of S. cerevisiae via an upstream activation site. Cell 32:1279-1286. 8. Hahn, S., and L. Guarente. 1988. Yeast HAP2 and HAP3: transcriptional activators in a heteromeric complex. Science 240:317-321. 9. Healy, A. M., T. L. Helser, and R. S. Zitomer. 1987. Sequences required for transcriptional initiation of the Saccharomyces cerevisiae CYC7 gene. Mol. Cell. Biol. 7:3785-3791. 10. Hodge, M. R., G. Kim, C. Christian, and M. G. Cumsky. 1989. Inverse regulation of the yeast COX5 genes by oxygen and heme. Mol. Cell. Biol. 9:1958-1964. 11. Klebe, R. J., J. V. Harriss, Z. D. Sharp, and M. G. Douglas. 1983. A general method for polyethylene-glycol-induced genetic transformation of bacteria and yeast. Gene 25:333-341. 12. Kunkel, T. A. 1985. Rapid and efficient site-directed mutagene-

MOL. CELL. BIOL.

13. 14.

15. 16.

17. 18. 19. 20.

21. 22.

23.

24. 25. 26.

27. 28.

sis without phenotypic selection. Proc. Natl. Acad. Sci. USA 82:488-492. Laz, T. M., D. F. Pietras, and F. Sherman. 1984. Differential regulation of the duplicated iso-cytochrome c genes in yeast. Proc. Natl. Acad. Sci. USA 81:4475-4479. Lemontt, J. F., D. R. Fugit, and V. L. MacKay. 1980. Pleiotropic mutations in the TUPI locus that affect the expression of mating-type dependent functions in Saccharomyces cerevisiae. Genetics 94:899-920. Lowry, C. V., and R. Leiber. 1986. Negative regulation of the Saccharomyces cerevisiae ANBI gene by heme, as mediated by the ROXI gene product. Mol. Cell. Biol. 6:4145-4148. Lowry, C. V., J. L. Weiss, D. A. Walthall, and R. S. Zitomer. 1983. Modulator sequences mediate the oxygen regulation of CYCI and a neighboring gene in yeast. Proc. Natl. Acad. Sci. USA 80:151-155. Lowry, C. V., and R. S. Zitomer. 1984. Oxygen regulation of aerobic and anaerobic genes mediated by a common regulatory factor in yeast. Proc. Natl. Acad. Sci. USA 81:6129-6133. Lowry, C. V., and R. S. Zitomer. 1988. ROXI encodes a heme-induced repression factor regulating ANBI and CYC7 of Saccharomyces cerevisiae. Mol. Cell. Biol. 8:4651-4658. Maniatis, T., E. F. Fritsch, and J. Sambrook. 1982. Molecular cloning: a laboratory manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. Mattoon, J. R., W. E. Lancashire, H. K. Saunders, E. Carvajal, D. R. Malamud, G. R. C. Braz, and A. D. Panek. 1979. Oxygen and catabolite regulation of hemoprotein biosynthesis in yeast, p. 421-435. In W. S. Caughey (ed.), Biochemical and clinical aspects of oxygen. Academic Press, Inc., New York. McNeil, J. B., and M. Smith. 1986. Transcriptional initiation of the Saccharomyces cerevisiae iso-1-cytochrome c gene. Multiple independent TATA sequences. J. Mol. Biol. 187:363-378. Mehta, K. D., D. Leung, L. Lefebvre, and M. Smith. 1990. The ANB1 locus of Saccharomyces cerevisiae encodes the protein synthesis initiation factor eIF-4D. J. Biol. Chem. 265:88028807. Muksi, Y., S. Harashima, and Y. Oshima. 1991. AAR1/TUP1 protein, with a structure similar to that of the P subunit of G proteins, is required for al-a2 and a2 repression in cell type control of Saccharomyces cerevisiae. Mol. Cell. Biol. 11:37733779. Ng, R., and J. Abelson. 1980. Isolation and sequence of the gene for actin in Saccharomyces cerevisiae. Proc. Natl. Acad. Sci. USA 77:3912-3916. Pfeiffer, K., B. Arcangoli, and L. Guarente. 1987. Yeast HAP1 activator competes with the factor RC2 for binding to the upstream activation site UAS1 of the CYC1 gene. Cell 49:9-18. Pfeiffer, K., T. Prezant, and L. Guarente. 1987. Yeast HAP1 activator binds to two upstream activation sites of different sequence. Cell 49:19-27. Pinkham, J. L., J. T. Olesen, and L. Guarente. 1987. Sequence and nuclear localization of the Saccharomyces cerevisiae HAP2 protein, a transcriptional activator. Mol. Cell. Biol. 7:578-585. Redding, K., C. Holcomb, and R. S. Fuller. Submitted for

publication. 29. Rose, M., and D. Botstein. 1983. Construction and use of gene fusions to lacZ (IGal) that are expressed in yeast. Methods Enzymol. 101:167-180. 30. Rose, M. D., P. Novick, J. H. Thomas, D. Botstein, and G. R. Fink. 1987. A Saccharomyces cerevisiae genomic plasmid bank based on a centromere-containing shuttle vector. Gene 60:237243. 31. Rothstein, R. J. 1983. One-step gene disruption in yeast. Methods Enzymol. 101:202-211. 32. Rothstein, R. J., and F. Sherman. 1980. Genes affecting the expression of cytochrome c in yeast: genetic mapping and genetic interactions. Genetics 94:871-889. 33. Sanger, F., S. Nicklen, and A. R. Coulson. 1977. DNA sequencing with chain terminating inhibitors. Proc. Natl. Acad. Sci. USA 74:5463-5467. 34. Schamhart, D. H. J., A. M. A. Ten Berge, and K. W. Van De Poll. 1975. Isolation of a catabolite repression mutant of yeast as

ROX3 GENE OF S. CEREVISIAE

VOL. 11, 1991

35.

36. 37.

38. 39. 40.

41.

42.

43. 44. 45.

a revertant of a strain that is maltose negative in the respiratorydeficient state. J. Bacteriol. 121:747-752. Schnier, J., H. G. Schwelberger, Z. Smet-McBride, H. A. Kang, and J. W. B. Hershey. 1991. Translation initiation factor 5A and its hypusine modification are essential for cell viability in the yeast Saccharomyces cerevisiae. Mol. Cell. Biol. 11:3105-3114. Schultz, J., and M. Carlson. 1987. Molecular analysis of SSN6, a gene functionally related to the SNF1 protein kinase of Saccharomyces cerevisae. Mol. Cell. Biol. 7:3637-3645. Shapira, S. K., J. Chou, F. V. Richaud, and M. J. Casadaban. 1983. New versatile plasmid vectors for expression of hybrid proteins coded by a cloned gene fused to lacZ gene sequences encoding an enzymatically active carboxy-terminal portion of 3-galactosidase. Gene 25:71-82. Sherman, F., G. R. Fink, and J. B. Hicks. 1985. Methods in yeast genetics. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. Sherman, F., and J. W. Stewart. 1971. Genetics and biosynthesis of cytochrome c. Annu. Rev. Genet. 5:257-2%. Southern, E. M. 1975. Detection of specific sequences among DNA fragments separated by gel electrophoresis. J. Mol. Biol. 98:503-517. Struhl, K., D. T. Stinchcomb, S. Scherer, and R. W. Davis. 1975. High frequency transformation of yeast: autonomous replication of hybrid DNA molecules. Proc. Natl. Acad. Sci. USA 76:1035-1039. Trueblood, C. E., R. M. Wright, and P. 0. Poyton. 1988. Differential regulation of the two genes encoding Saccharomyces cerevisiae cytochrome c oxidase subunit V by heme and the HAP2 and REOJ genes. Mol. Cell. Biol. 8:4537-4540. Trumbly, R. J. 1988. Cloning and characterization of the CYC8 gene mediating glucose repression in yeast. Gene 73:97-111. Turi, T. G., and J. C. Loper. 1990. Expression of cytochrome P450 lanosterol 14 a-demethylase (CYPSI) is regulated by multiple cis and trans elements. Yeast 6:S234. Verdiere, J., F. Creusot, and L. Guarente. 1986. The overpro-

46.

47.

48.

49.

50.

51. 52.

53.

5647

ducing CYPI mutation and the underproducing hapi mutations are alleles of the same gene which regulates in trans the expression of the structural genes encoding iso-cytochromes c. Curr. Genet. 10:339-342. Verdiere, J., F. Creusot, and M. Guerineau. 1985. Regulation of the expression of iso-2-cytochrome c gene in S. cerevisiae: cloning of the positive regulatory gene CYPI and identification of the region of its target sequence on the structural gene CYP3. Mol. Gen. Genet. 199:524-533. Williams, F. E., and R. J. Trumbly. 1990. Characterization of TUPI, a mediator of glucose repression in Saccharomyces cerevisiae. Mol. Cell. Biol. 10:6500-6511. Wright, R. W., and R. 0. Poyton. 1990. Release of two Saccharomyces cerevisiae genes, COX6 and CYCI, from glucose repression requires the SNF1 and SSN6 gene products. Mol. Cell. Biol. 10:1297-1300. Wright, C. F., and R. S. Zitomer. 1985. Point mutations implicate repeated sequences as essential elements of the CYC7 negative upstream site in Saccharomyces cerevisiae. Mol. Cell. Biol. 5:2951-2958. Zagorec, M., J.-M. Buhler, I. Treich, T. Keng, L. Guarente, and R. Labbe-Bois. 1988. Isolation, sequence, and regulation by oxygen of the yeast HEM13 gene coding for coproporphyrinogen oxidase. J. Biol. Chem. 263:9718-9724. Zagorec, M., and R. Labbe-Bois. 1986. Negative control of yeast coproporphyrinogen oxidase synthesis by heme and oxygen. J. Biol. Chem. 261:2506-2509. Zhang, M., L. S. Rosenblum-Vos, C. V. Lowry, K. A. Boayke, and R. S. Zitomer. A yeast protein with homology to P-subunits of G proteins is involved in control of heme-regulated and catabolite-repressed genes. Gene, in press. Zitomer, R. S., J. W. Seilers, D. W. McCarter, G. A. Hastings, P. Wick, and C. V. Lowry. 1987. Elements involved in oxygen regulation of the Saccharomyces cerevisiae CYC7 gene. Mol. Cell. Biol. 7:2212-2220.