MOLECULAR AND CELLULAR BIOLOGY, OCt. 1991, p. 4934-4942

Vol. 11, No. 10

0270-7306/91/104934-09$02.00/0 Copyright © 1991, American Society for Microbiology

Regulation of the Yeast CYTI Gene Encoding Cytochrome cl by HAP1 and HAP2/3/4 JANE C. SCHNEIDERt AND LEONARD GUARENTE* Department of Biology, Massachusetts Institute of Technology, Cambridge, Massachusetts 021394307 Received 11 March 1991/Accepted 5 July 1991

Mitochondrial biogenesis requires the coordinate induction of hundreds of genes that reside in the nucleus. We describe here a study of the regulation of the nuclear-encoded cytochrome cl of the b-cl complex. Unlike cytochrome c, which is encoded by two genes, CYC1 and CYC7, cl is encoded by a single gene, CYTI. The regulatory region of the CYTI promoter contains binding sites for the HAP1 and HAP2/3/4 transactivators that regulate CYCI. The binding of HAP1 to the CYTI element was studied in detail and found to differ in two important respects from binding to the CYCI element. First, while CYCi contains two sites that bind HAP1 cooperatively, CYTI has a single high-affinity site. Second, while the CYTI site and the stronger HAPl-binding site of CYCI share a large block of homology, the HAP1 footprints at these sites are offset by several nucleotides. We discuss how these differences in HAP1 binding might relate to the difference in the biology of cytochrome c and cytochrome cl. The mitochondrion of eukaryotic cells is the site of elec-

and heme biosynthetic enzymes (9, 29). The primary signal that regulates the activity of the HAP2/3/4 complex is the carbon source; the activity is low in glucose and high in lactate (12). The HAP2/3/4 complex has been conserved in eukaryotes that range from yeasts to mammals, although its cellular role has apparently been altered over evolution (1, 3, 28). The importance of the HAP complex to mitochondrial function in yeast cells is illustrated by the phenotype of hap2, or -3, or 4 mutants, an inability to grow on nonfermentable carbon sources. In contrast, hapi mutants do grow on such carbon sources, because the genes activated by HAP1 can also be activated by the HAP2/3/4 system. The regulatory circuitry of genes regulated by these HAP pathways has been partly delineated. These regulators were initially identified by their effects on CYCI (iso-l-cytochrome c). Both HAP1 and HAP2/3/4 bind to UASs in the CYCJ promoter to activate transcription. A second gene, CYC7, encodes an isoform of cytochrome c, which is expressed at a low level even when cells grow under oxygen limitation (24, 35). While HAP1 activates CYC7 under aerobic growth, an independent, poorly characterized system keeps the gene on at a basal level under oxygen limitation (35). CYC7 is not regulated at all by HAP2/3/4. A situation similar to that of CYCI and CYC7 occurs in the case of the cytochrome c oxidase complex, which is encoded by several nuclear and mitochondrial loci. Subunit 5 of the complex is encoded by two nuclear genes, COXSA and COX5B (4, 5), which are expressed under aerobic growth and oxygen limitation, respectively (16). COX5A is regulated by both HAP1 and HAP2/3/4, while COX5B responds to HAP1 and not HAP2/3/4 (17, 40). It has been suggested that the anaerobic forms of these cytochromes may function more efficiently than their aerobic counterparts under conditions of oxygen limitation. Less is known about the regulation of the cytochrome b-cl complex, which passes electrons to cytochrome c in the inner membrane of the mitochondria. This complex contains one mitochondrially encoded cytochrome, b (encoded by COB), and one nuclear-encoded cytochrome, c1 (encoded by CYTI) (26, 36). The complex also contains numerous other nuclear-encoded subunits, including the Rieske iron-sulfur protein (encoded by RIPI), and other subunits encoded by

tron transport and energy production by oxidative phosphor-

ylation. It is also the location of numerous biosynthetic and degradative pathways. While the organelle has its own genome which encodes about 10 proteins, the vast majority of the hundreds of proteins that reside in mitochondria are encoded in the nucleus (8). These include proteins that are necessary for macromolecular synthesis within the organelle, cytochromes and related proteins that carry out electron transport, and enzymes in anabolic and catabolic pathways that reside in mitochondria. All of these nuclearencoded proteins must be imported into mitochondria after their synthesis in the cytoplasm. The electron transport chain and related proteins are highly regulated in the yeast Saccharomyces cerevisiae. This organism will grow anaerobically, but under these conditions, the cytochrome cofactor, heme, is not synthesized (25). The activity of heme-requiring enzymes is muted under such conditions, resulting in several auxotrophies, such as for ergosterol, unsaturated fatty acids, and methionine. Further, the synthesis of the cytochromes themselves is repressed in anaerobically grown cells (13). When cells grow in the presence of oxygen, the level of synthesis of cytochromes depends on the carbon source in the medium. The level of synthesis of cytochromes is low when cells grow in glucose and high when they grow in a nonfermentable carbon source such as lactate (42). Two trans-acting regulatory systems that govern cytochrome synthesis have been described. The first is a protein, HAP1 (CYP1), which is a transcriptional activator that responds to oxygen (12). The oxygen signal is registered by synthesis of heme (13), which activates the ability of HAP1 to bind to the upstream activation sequence (UAS) sites of particular cytochrome and related genes (30, 31, 41). The second system is composed of the HAP2/3/4 heteromeric complex that binds to CCAAT boxes of many genes involved in mitochondrial function, including cytochromes

Corresponding author. t Present address: Plant Research Laboratory, Michigan State University, East Lansing, MI 48824. *

4934

VOL . 1 l, 1991

CORI and CORII involved in the assembly of the complex (2, 21, 22). In this report, we begin an analysis of the regulation of this complex by studying CYTl. A detailed picture of the UASs and trans-acting factors that govern CYT1 expression is derived. The regulation of this gene is compared and contrasted with that of CYCI. MATERIALS AND METHODS Yeast strains. BWG1-7a is MATa ura3-52 leu2-3 leu2-112 his4-519 adel-100 (12). Strains hapJA::LEU2 (32), hap2A:: LEU2 (27), and TPhl (hemi::LEU2) (34) contain a disruption of the specified locus made by inserting a LEU2 gene in strain BWG1-7a. Strain JS101 contains a lacZ fusion inserted into the CYTI locus of BWG1-7a; pJCS105, described below, was partially digested with EcoRV to direct integration at -120 upstream of the CYTI initiation codon. (All position numbers are relative to the start of translation.) This resulted in a fusion of all cognate upstream sequences as well as 90 bp of coding sequence 5' to the lacZ fusion. Strain JS105 is disrupted by LEU2 at the CYTI locus by transformation and recombination with a HindIII-BamHI fragment from pJCS108, described below. Media and assays. Undefined rich medium for yeast cultures was 2% yeast extract-1% peptone (Difco) (YEP) with 2% glucose (YPD) or lactate (YEPL). Plates were made by adding agar to 2%. Selective medium was 0.67% yeast nitrogen base without amino acids (Difco). Appropriate amino acid supplements were added as needed to 0.004%. For growth of strain TPhl (heml::LEU2), &-aminolevulinic acid (8-ALA) was added to 50 jig/ml (heme sufficient or high ALA) or to 0.5 jig/ml (heme deficient or low ALA). ,-Galactosidase units were measured as described by Guarente and Mason (13). Plasmid constructions and DNA sequencing. (i) pJCS105. BamHI linkers (8 bp) were ligated to a 441-bp Hindlll fragment from CYTl clone pll3A (36); this piece contains sequence from -352 to +90. The fragment was ligated into the BamHI site of pSX A-178 (13) in order to fuse CYTJ in frame to lacZ. CYCI sequences, which lay upstream of the CYTJ-lacZ fusion in pJCS101, were removed by a partial BamHI and SmaI digest, and the plasmid was reclosed with 8-bp SmaI linkers to make pJCS105. (ii) pJCS109 and pJCS110. Plasmid JCS109 is based on pJCS105 with the following changes: the BamHI site at the lacZ junction contains a linker that destroyed the site; a 2.8-kb HindIII ARS1 CEN4 fragment was inserted into the HindIII site; and upstream sequence to bp -1800 was added from the new CYTI clone. To make pJCS110 (also known as deletion [del] 201), a 1-kb SmaI-to-EcoRV fragment was removed from pJCS109, leaving 736 bp of upstream sequence. To make deletions from the 5' end (del 202, 203, and 204), a 1.3-kb PvuII-BamHI fragment from the CYTI clone was cloned into a plasmid derived from pEMBL-18. Bal3l was used to generate deletions from the unique EcoRV site (-737). Deletions were sequenced and religated back into the SmaI-BamHI backbone of pJCS109. To make del 206, Bal31 deletions were commenced from the BamHI site and SmaI-BamHI adapters were attached to reconstruct the lacZ plasmid. (uii) pJCS108. A 493-bp HindIll-HinclI fragment from the Sadler clone was subcloned into HindIII-EcoRV-digested pBR322. A 2.2-kb SalI-XhoI fragment containing the LEU2 gene was inserted into the site resulting from removal of a 148-bp BstEII fragment within the CYTI clone.

REGULATION OF THE YEAST CYCI GENE

4935

Strain TG1 was used to grow single-stranded M13 mpl8 or mp19 recombinant phage, which were purified according to protocols from New England BioLabs. Double-stranded plasmid templates were purified according to Kraft et al. (20). Single-stranded DNA from pEMBL-18-derived vectors was made by superinduction with IR1 as according to Dente et al. (6). The lacl primer (also known as the ATG primer) hybridizes just downstream of the lacZ fusion junction. The sequence is 5'-TCACCAGTGAGACGGGCAAC-3'. Sequencing was carried out by the dideoxy-chain termination method (37), using a modified form of T7 polymerase (Sequenase [39] kit from U.S. Biochemical). Primer extension. Total RNA was isolated as described by Pikielny and Rosbash (33). Reverse transcripts were extended from the radioactively labeled lacI primer (see above) with reverse transcriptase as described previously (14). Transcripts were separated on an 8% sequencing gel. Footprinting. Fragments were labeled by filling in recessed 3' ends with a-32P-labeled deoxynucleoside triphosphates and Klenow fragment as described by Maniatis et al. (23). Fragments were then cleaved from a larger piece to generate fragments with one 3'-labeled end. Gel shift analysis. DNA fragments were purified from acrylamide gels by either the crush-and-soak or the electroelution method. The crush-and-soak procedure entailed mashing the gel slice and soaking it in 5 ml of 0.5 M sodium chloride-10 mM Tris (pH 8)-10 mM EDTA-0.5% sodium dodecyl sulfate (SDS) overnight; DNA was then ethanol precipitated. Alternately, gel slices were electroeluted for 50 min at 150 V into 3 M sodium acetate in an IBI electroelution chamber, run in a buffer made of 20 mM Tris (pH 8) 3 mM sodium chloride, and 0.2 mM EDTA. Extracts were prepared, bound to end-labeled DNA fragments, and electrophoresed through 4% acrylamide gels as described previously (30, 31). [35S]Met-labeled HAP1 was a gift from Karl Pfeifer. DNase I protection. Labeled fragments were incubated with protein extract from BGW1-7a pSDSHAP1 and treated with DNase I as described previously (10). Sufficient HAP1 extract was used to drive all of the DNA into a bound complex, as assessed by a gel shift assay. Ethylation interference footprinting. DNA fragments were reacted with diethylsulfate (DES) before being used in a binding reaction. The base specificity of this chemical is the same as for dimethylsulfate (38), but ethyl groups are attached instead of methyl groups. Fragments in 0.1 ml of DES reaction buffer (500 mM sodium cacodylate, pH 8) were treated with 2 pul of DES for 1 h at 37°C, ethanol precipitated three times, dried, and resuspended in Tris-EDTA. The ethylated fragment was bound to protein extract and electrophoresed as described above to separate bound from free DNA. DNA was isolated from the gel by electrophoresis onto a sheet of NA45 DEAE-paper (Schleicher & Schuell) in 1 x TBE at 4°C for 2 h at 0.5 A. Fragments were eluted from the paper in 1 M NaCl-0.1 M EDTA-29 mM Tris (pH 8) for 20 min at 65°C. The eluate was extracted twice with chloroform and ethanol precipitated. Fragments were resuspended in 43 ,ul of 20 mM sodium acetate-1 mM EDTA and cleaved at the ethylation sites at 90°C for 30 min by addition of 7.5 ,ul of 1 M NaOH. Samples were neutralized by adding 7.5 pl of 1 M HCI and 50 ,ul of 20 mM Tris, pH 7.5. DNA was precipitated with 4 ,ug of calf thymus DNA, 150 p.1 of 70% ethanol, and 1 ml of n-butanol. After resuspension in 150 p.l of 1% SDS, DNA was precipitated again and washed with butanol, dried, and loaded on a sequencing gel. Binding rates. The on-minus-off rate was measured by

4936

MOL. CELL. BIOL.

SCHNEIDER ET AL.

TABLE 1. Activity and regulation of the CYTI-lacZ fusiona ,B-Galactosidase activity (U) Plasmid

HAP] HAP2

hap) HAP2

HAP) hap2, glucose

Glucose Lactate Glucose Lactate

CYTI CYCI a

21 164

142 1,024

2 6

78 192

3 66

gel was loaded at various times after addition of the competitor.

hem) Low High

8-ALA 8-ALA 2 ND

26 ND

P-Galactosidase was assayed in strain BWG1-7a, or mutant derivatives,

bearing pJCS109 (CYTI-lacZ) or pLG-312 (CYC)-lacZ) (13). The mutant strains bearing the hap)-), hap2-), or hem)-) mutation have been described elsewhere (12, 35). Strains were grown in glucose or lactate. The hem) strain was grown in glucose in the presence of low (0.5 ,ug/ml) or high (50 ,Lg/ml) concentrations of B-ALA. The HAP) hap2 strain was grown only in glucose because the hap2 mutation prevents growth in nonfermentable carbon sources. ND, not determined.

mixing crude extracts with fragment and then loading the gel at various times after mixing. This is not a strict on rate, since dissociation of the complex is possible during the reaction. The off rate was measured by allowing binding of crude extract and fragment to proceed to equilibrium (30 min) and then adding unlabeled competitor (0.5 ,ug of XhoI-digested pLGA312 at time zero, an amount previously determined to be enough to eliminate binding to the labeled fragment.) The

RESULTS Cloning a genomic CYTI fragment with extensive 5' DNA. The yeast CYTI clone described by Sadler et al. (36) contained only 352 nucleotides of 5' flanking DNA. To isolate the gene with a more extensive upstream region, we generated a strain in which a CYTl clone could be selected, by using the clone of Sadler et al. to disrupt the CYTI coding sequence. An internal 148-base BstEII fragment was replaced with the LEU2 gene to generate pJCS108, and a DNA fragment encompassing this replacement was used to transform BWG1-7a to Leu+. Disruption of CYTI was confirmed by Southern blotting (data not shown). The disrupted strain could not grow on glycerol, nor could revertants that would grow on glycerol be isolated, even after ethyl methanesulfonate mutagenesis. This finding indicates that CYTI is the only gene in S. cerevisiae encoding cytochrome c1. The CYTI clone was isolated in this strain from a 2-pum library by selecting for growth on glycerol. Construction of a CYTI-lacZ reporter. To examine the regulation of the CYTJ nuclear gene of S. cerevisiae, we initially used the clone of Sadler et al. (36) to construct a fusion with 352 nucleotides of CYTI 5' DNA (pJCS105). However, when this plasmid was transformed into S. cere-

BglH

EcoRV

GATCTIATrAGr:CkTtGTTCAATCA7CAlCAACAGTACGATATCGA

-730

AccI A CA r IDCTIUA"G IMCGAGXrMGTIr;GCGAGAlrSICwIrCCAlAGACG

-670

CGGLrcTAGCGCATCATACATmAGTrCTAACTrACAGiTGCTG'iNw

-610

Haan

A$FI'C1I'GAGCCAATCCACAnIrTrWATAAGGACGGGATAAGACGGAGCGCCACG

-550

492 -510 XuaI ICGG:GG YC"AG GATIwIAA=GGIWCCGGlA -478 -4 -464 TaqI

-490

-528

-430 1GAAMCATkA CGAGI@AGACGICMC IGGCCCCCviGaa TAAGGW7CfCTAGTkTATAAACWCA: T'TGCC

-370

HindIII

BamHI

GGAICCATGTACTAATLGCTlAGTAGAGGCCAl&T=TTr=%TAATCGGT A

A

-310

A

TICACMAT7ITFAGCAkCATrrATACAIALVAC77rJ7TrCMA"TAT

-250

TAAGGACAI LG 1GIAAAGAGAATCCATATICTATATrAG rMCACTATrmr

-190

C TAAATMTTTCCAGTArCTCCCCCAGTrGATA

kGCA; T1'CALTGMATTITrCCCA

-130

CATr ATTACATrTAT1GATC-AG CTAICrrGTFGAAALGGAAGM'ITAAAT

-70

TIr;CA¶TrrcATrACACTATA¶crTATITrI CTGGAGGGALGTrATAACT

-10

£

Scal SCI

£

IkoRV

AT1TT'ACMTG

FIG. 1. Sequence upstream of the CYTI ATG initiator codon. The sequence of approximately 800 bp of upstream DNA was determined described in Materials and Methods. Nucleotides below triangles are sites of mRNA initiation. The sites homologous to the HAP1 consensus (single underline) and HAP2/3/4 consensus (double underline) are indicated. Also shown are restriction sites referred to in the text.

as

REGULATION OF THE YEAST CYCI GENE

VOL . 1 l, 1991

.-

aim

.A

...-i,-.

'. -.- -I

-

4-9-

404-0-

mom 4moom

1

2

3

4

5

FIG. 2. Sites of CYTI mRNA initiation. Primer extension was used to determine the initiation sites as described in Materials and Methods. Strains contained either no CYTI-lacZ fusion, an integrated copy of CYTI-lacZ (strain JS101), or a multicopy plasmid with CYTI-lacZ, pJCS110. The initiation sites as revealed in

pJCS11O are indicated by arrows. Their position relative to the CYTI coding sequence was deduced by alignment to a sequencing ladder.

,-galactosidase activity in the CYT1

detected, suggesting missing. Using the genomic clone that we isolated, a CYTJ-lacZ fusion containing about 1,800 bases of CYTI 5' flanking DNA was constructed in a vector containing an autonomously replicating sequence and a centromere (pJCS109). The regulation of the reporter vector, pJCS109, was studied under various conditions and in different mutant strains (Table 1). Several important aspects of regulation were noted. First, like CYCI, CYT1 was regulated both by carbon source and by heme. Activity was very low under heme limitation, intermediate in glucose in the presence of heme, and fully induced in lactate. The full range of regulation was about 70-fold. Second, mutations in either HAP1 or HAP2 reduced expression, indicating that CYTI, like CYCI, responds to both the HAP1 and HAP2/3/4 systems. Mapping of the HAPl- and HAP2/3/4-responsive elements of CYTI. The sequence of CYTI 5' flanking DNA was determined to about -800 from the ATG (Fig. 1). Inspection of this sequence revealed strong similarity to both the HAP1 site in CYCI, UAS1B (to be described in detail later), and the HAP2/3/4 consensus sequence (TNA/GTTGG) between -464 and -512 (each underlined in Fig. 1). The 5' ends of CYT1 mRNA were mapped by extension of a primer which binds to lacI-Z (Fig. 2). Strains from which RNA was obtained contained either no reporter plasmid (lane 1), the CYTJ-lacZ fusion in pJCS105 inserted into the chromosome visiae,

no

that key

sequences

was

promoter were

4937

(strain JS101) (lanes 2 and 3), or the multicopy plasmid pJCS110 (a 5' deletion of pJCS109 ending at -736) (lanes 4 and 5). Because pJCS110 contains a small insertion of DNA linker nucleotides at the CYTI-lacZ junction relative to pJS105, the five prominent extension products are shifted slightly to a slower mobility in the gel. These products correspond to initiation sites at -140, -182, -268, -284, and -297 (indicated by triangles above these sites in Fig. 1) relative to the CYTI ATG. Deletions were constructed to determine the possible significance of the HAP1 and HAP2/3/4 binding-site homologies mentioned above (Fig. 3). A 5' deletion to -510 did not affect activity in glucose or lactate. Deletion to -492, which removed the HAP1 binding-site homology, greatly reduced the activity in glucose but only mildly lowered the induced activity in lactate. This is expected for loss of HAP1 responsiveness, because HAP1 is much more active than HAP2/3/4 when cells grow in glucose. Deletion of the HAP1 site in CYCJ, UAS1, also has a large effect on activity of the promoter in glucose (12). A further deletion to -464, which removed the HAP2/3/4 homology, abolished activity in lactate as well as glucose. A corresponding deletion of both sites in CYCI has similar effects on activity. An internal deletion removing the HAP2/3/4 but not the HAP1 homology gave a low to intermediate level of activity in glucose and lactate. This intermediate level of activity required HAP1 because it was greatly reduced in a hapl mutant strain but was unaffected in a hap2 mutant (data not shown). These findings provide evidence that the HAP1 and HAP2/3/4 homologies are both functional and respond to the two regulators in vivo. The HAP1 element was studied in detail and is described below. The HAP2/3/4 element was not investigated in detail, in part because of the difficulty in obtaining in vitro binding to sequences that do not contain precisely the canonical CCAAT. It is extremely likely, however, that HAP2/3/4 functions at the -492 to -464 element. First, the region contains the sequence TTGGTTGGTGGA, which matches perfectly the UAS2 of CYCI, a known HAP2/3/4 element (7). Second, transcription of CYTl is greatly reduced in the hap2 mutant strain (Table 1). Attempts to study the response of this element to HAP2/ 3/4 in vivo was complicated by the strong apparent synergy between the HAP1 and HAP2/3/4 homologies in the CYT1 promoter. Thus, deleting either the HAP1 homology (del -492) or the HAP2/3/4 homology (del -478 to -368) greatly reduced expression. This level of synergy is somewhat greater than for UAS1 and UAS2 in CYCI. Nevertheless, we were able to show that the low level of activity in del -492 was reduced further in a hap2 strain (from 0.4 to 0.1 U of activity). In summary, our in vivo analysis suggests that the CYTI promoter bears functional HAP1 and HAP2/3/4 sites that are arranged in a manner similar to that of CYCJ. Binding of HAP1 to the CYTI UAS. We wished to obtain evidence that HAP1 itself bound to the HAPl-responsive element of CYTI. Accordingly, we carried out gel shift analysis of a 77-bp fragment encompassing bases -528 to -455 in CYTI 5' DNA. A protein bound to this probe that was absent in a strain with a hapi disruption (Fig. 4, lane 1) and increased in concentration in strains that expressed increased amounts of HAP1. As found for UAS1 or CYC7, a strain bearing HAP1 on a high-copy-number plasmid (lane 3) displayed a moderate increase in the levels of the complex compared with a wild-type strain (lane 2), and a strain with HAP1 under GAL control (pSD5HAP1) showed a large increase (lane 4). This protein-DNA complex required heme

4938

MOL. CELL. BIOL.

SCHNEIDER ET AL.

C',

u

31

co,g

m

W-

0 '-4

-o

04

V-

C'.

9-

ci

0

U'co

cq

0

c

0 oo

0

m

C5 o2C

ci

ci

o0

0c

cI

cq

=*0 o

co

oo

*0

0

0

1

u

1

0

0

10

10

c *o

r-_

N

'-",,

CZ~

-o3 ruY

Is

0 1W.0

0 -

U

'

an 0

U

rA

-o U Go

U,0 U ,

o 0CA o-a

0

'-I

1cr

I?

Si

I

IT

Q o,

ho

EU 0

R .) U

0f

-

CI

I1

11%, I's

I's "I.,

sL0-4h (U

VOL.

11, 1991

REGULATION OF THE YEAST CYCl GENE

4939

-41

.;-

;....i ll.-

-Itl

C',

-.;,

11

I

1-11

C,~,

.-,

."

IN

I, .-

HAP1 .N-

1

Free

1

2

3

4

FIG. 4. Binding of HAP1 in yeast extracts to the CYT1 site by gel shift. Extracts were prepared from BWG1-7a bearing a hapl null mutation (hapl::LEU2), a single copy of HAP1 (hap2::LEU2), a 2pLm multicopy HAP1 plasmid (pHAP1) (30), or a plasmid with HAP1 under control of the GAL promoter (pSD5HAP1) (18). The probe was a SmaI-TaqI fragment of the del 202 plasmid from -528 to -455 in CYT1 upstream DNA. The binding mix contained 0.5 p.g of salmon sperm DNA, 15 pLg of heme per ml, and 0.5 ng of labeled CYTI fragment. The complex with the mobility characteristic of a HAP1-DNA complex is indicated by the arrow.

for its formation (data not shown), a requirement also seen for HAP1 binding to UAS1 or CYC7. Although these data made it extremely likely that this complex represented HAP1 binding to the CYT1 site, we confirmed this conclusion by synthesizing the protein in vitro labeled with [35S]Met. As shown in Fig. 5, the DNAbinding domain of HAP1 (residues 1 to 245) synthesized in vitro bound to probes containing CYTI (as well as UAS1 or CYC7) but not UAS2 of CYCl. A band seen in all lanes corresponds to HAP1 (residues 1 to 245), while only those lanes containing fragments to which HAP1 binds showed the additional HAP1-DNA complex, indicated by the arrow. We conclude from this binding analysis that CYTI contains a single HAPl-binding site which mediates activation by the protein

2

3

4

I Im1

5

FIG. 5. Binding of HAP1 synthesized in vitro to the CYTI site. A [35S]Met-labeled HAP1 fragment (residues 1 to 245) was synthesized in vitro in a rabbit reticulocyte system and used in a binding reaction with 3 ng of unlabeled DNA. The binding mix contained 0.3 Rg of salmon sperm DNA and 4 pl of translation extract. The DNA fragments used are a 160-base SmaI-BamHI fragment from del 202 (-528 to -368) containing the CYTI UAS1 and UAS2, a 90-base SmaI-XhoI fragment containing CYCI UAS1, a 125-base SmaI-XhoI fragment from del 205 containing the CYI1 UAS2, and an 80-base BglII-XhoI fragment containing the CYC7 HAP1 site. The band present in all lanes corresponds to labeled HAP1 fragment. The HAP1-DNA complexes are indicated by the arrow.

by several bases, a point to which we return in Discussion. The CYT1 site also contains a strong dyad symmetry (indicated by inverted arrows in Fig. 7) that is not nearly as evident in the UAS1 site (30, 31). This dyad predicts that HAP1 binds to its target sites as a dimer. Affinity of HAP1 for the CYTI site versus UAS1 of CYCI. The UAS1 of CYCl contains two sites, the A site and the B site, to which HAP1 binds cooperatively (19, 30). UAS1B is a much stronger binding site than UAS1A, and it binds HAP1 with an affinity comparable to that for the single HAPl-binding site of CYC7 (31). To compare the relative affinities of HAP1 for UAS1 versus CYTI, we determined on and off rates of the protein for UAS1B and the CYT1 site (Fig. 8). While on times were rapid and comparable for the two sites, the off time was three- to fivefold longer for the CYT1 site. These results suggest that HAP1 has a higher affinity for the CYT1 site than for UAS1B. This higher affinity could give rise to a higher activity in vivo. The intact UAS1 may compensate for this difference by containing the second HAP1-binding site, UAS1A.

in vivo.

The HAPl-binding site in CYT1 was pinpointed by DNase I and ethylation interference analysis (Fig. 6). In the DNase I experiment (lanes 1 to 5), we detected protection by an extract that contained sufficient HAP1 to bind to all of the probe. The protected sequences are indicated with a bracket. In the ethylation interference experiment (lanes 6 to 10), the DNA was pretreated with DES as described in Materials and Methods and then subjected to gel shift after incubation with an extract containing HAP1. The HAPlbound and free DNA was excised and cleaved prior to analysis on a sequencing gel (Fig. 6). Bases which when ethylated interfere with binding are indicated by arrows. A summary of the footprinting data along with a comparison of similar data for UAS1 (30, 31) is shown in Fig. 7. Best alignment of the sequences reveals a 15-nucleotide block in which 12 bases match. Amazingly, this alignment offsets the DNase and alkylation interference footprints of the two sites

DISCUSSION In this report, we describe an analysis of the transcriptional regulation of the yeast nuclear gene, CYT1, encoding cytochrome c1 of the b-cl complex. In several respects, the regulation of this gene mirrors that of the CYCl gene encoding the major isoform of cytochrome c. Both genes are regulated by the two signals of heme and carbon source. Activity is low or missing in the absence of heme, intermediate in glucose in the presence of heme, and fully induced in lactate. In the case of CYT1, this regulation spans a 70-fold range. The organization of the CYT1 promoter is also similar to that of the CYCI promoter. In both cases, transcription is driven by a UAS region that contains binding sites for the two activators, HAP1 and HAP2/3/4. The UAS region in CYT1 lies between -510 and -464 from the ATG. Tran-

4940

SCHNEIDER ET AL.

MOL. CELL. BIOL.

G/A

-

- +

+ HAP1 extract

B

CYTi

C.

B. CYT1 bottom strand

CYTI bottom strand

A.

top strand

B

F

F

F

a _t i. :. -492 4-

496

-507/8

-504 -503

4-- -502 4o- -505/6

-513 -516

1

2 3 4 5

6

7

-502

8

9

-501

10

FIG. 6. DNase and ethylation interference footprinting of the HAP1 site in CYTI. (A) A fragment in which the bottom strand of CYTI DNA was labeled was prepared by digesting the del 202 plasmid (Fig. 3) with AccI (which cleaves in the URA3 gene), filling in with DNA polymerase and a-32P-labeled nucleotides, and cleaving with BamHI. The fragment bearing CYTI sequences from -528 to -368 was gel isolated for DNase I protection. Protein extract from BWG1-7a bearing the GAL promoter-HAP1 plasmid was added to levels that complexed all of the labeled fragment. Heme (15 ,ug/ml) and salmon sperm DNA (0.5 ,ug/ml) were also added. DNAse treatment was as described in Materials and Methods. The samples were run on standard sequencing gels. Lanes: 1, size marker; 2, free DNA, 10 ng of DNase; 3, free DNA, 1 ng of DNAse; 4, extract plus DNA, 120 ng of DNase; 5, extract plus DNA, 60 ng DNase. (B and C) For ethylation interference, the fragment described above (bottom strand) as well as a fragment labeled at a Taq site at -456 in CYTI (top strand) were used. The latter fragment was subcut at the AccI site as described above. The fragment contains CYTJ sequences from -528 to -456. DNA fragments were modified with DES and then combined with extract as described above. Bound (B) and free DNA (F) were separated by gel shift, isolated separately, and treated with sodium hydroxide to cleave at modified bases. The samples were run on sequencing gels, as shown. Positions in CYTI DNA that interfere with binding are missing in the bound lanes, compared with the free lanes, and are indicated by arrows. Lanes: 6 and 8, bound DNA; 7, 9, and 10, free DNA.

scripts initiate at five predominant sites that span a 160nucleotide region, between -300 and -140. Both the HAP1 site and the HAP2/3/4 site were deemed functional by analyzing deletion-bearing constructs in wild-type and hap mutant strains. As for CYCI, the HAP1 site is promoter distal from the HAP2/3/4 site. Since HAP2/3/4 sites are asymmetrical, they may be assigned an orientation relative to the promoter. The orientation at CYTI is the same as that of CYCJ. However, there are several important differences between the CYCI and CYTI UAS regions. First, the CYCI UAS contains two adjacent HAPl-binding sites, which bind the protein cooperatively (19, 30, 31), while CYTI has a single site to which HAP1 binds with a higher affinity than to either

GGCGGCCGGTATTTCCGGCGGC CCGCCGGCCATAAAGGCCGCCG

TGGCCGGGGTTTACGGACGATGA ACCGGCCCCAAATGCCTGCTACT

CYTi

UAS1B

FIG. 7. Comparison of HAP1 sites of CYTI and CYCI. The DNase I footprints of the HAPl-binding sites of CYTI and CYCI (UAS1B) are shown. Bases that are contact sites by ethylation

(CYTI) or methylation (CYCI) interference are in bold. The inverted repeat sequence in the CYTI site is indicated by the arrows. The best alignment of the sequences yields a 15-nucleotide region (pluses and minuses) in which 12 bases are identical (pluses). Note that the footprints are offset by several nucleotides.

UAS1 site. We believe that this difference may reflect the biological context of the expression of these two genes, a point on which we elaborate below. A second difference, also discussed below, is that the HAP1 DNase I footprints on the CYTI site and on the stronger CYCI site are offset with respect to the DNA homology element. A third difference between the CYTI and CYCI UAS regions is that the distance between the HAP1 footprint and the HAP2/3/4 CCAAT box in CYTI (20 bases) is closer than in CYCJ (42 bases). This closer spacing may facilitate synergy between HAP1 and HAP2/3/4 at CYTI, but this point has not yet been

firmly established. A major consequence of the two cooperative sites in UAS1 is that the HAP1 activation profile of CYCI should be sigmoidal. In fact, the activation curve of CYCI-lacZ as a function of concentration of the heme supplement B-ALA is sigmoidal (15). Thus, the activity of UAS1 will drop off sharply under conditions of reduced HAP1 activity (due to reduced synthesis of heme when cells grow under oxygen limitation). In contrast, the single UAS site in CYTI should provide that gene with a more linear reduction in activity as HAP1 activity is lowered. Why should CYCJ and CYTI respond differently under conditions of oxygen limitation? One salient difference between the biology of cytochromes c and c1 is that there exist two forms of cytochrome c, encoded by CYCI and CYC7. Under conditions of oxygen limitation, cells preferentially express the CYC7 form of cytochrome c (24). We speculate that a precipitous decline in UAS1 activity as oxygen levels fall ensures that CYC7 will direct the bulk of cytochrome c synthesis under oxygen

REGULATION OF THE YEAST CYCI GENE

VOL. 1 l, 1991

A 100

80

.5a

60

4)

I6..

4)

Q.

40

20

0

10

20 minutes

30

40

B

4941

region in the DNase I footprint at C YTI. These protection profiles suggest that HAPI can bind to these related sequence elements in qualitatively different ways. The differences in the nature of the protein-DNA interactions at the two sites may account for the observed difference in HAP1 binding affinity for CYTI versus CYCI. This kind of flexibility in the HAP1 DNA-binding domain was suggested from several earlier observations. First, the CYC7 site, to which HAP1 binds with high affinity, bears minimal sequence similarity to UAS1B or CYTI (31). Second, the response of the CYC7 and UAS1 elements to HAP1 differs quantitatively in vivo. UAS1 is more active than CYC7 (11, 18). Third, the effects of internal deletions in HAP1 are opposite at UAS1B and CYC7 (19). These deletions increase activity at UAS1 and decrease activity at CYC7. To address these observations, we have proposed models that invoke different conformations in the DNAbinding domain of HAP1 when the protein is bound at the dissimilar UAS1 and CYC7 sequences (18). Such differences would impose an additional layer of regulation to adjust the activity level of HAP1 at different UAS elements. ACKNOWLEDGMENTS This work was supported by NIH grant GM30454 to L.G. We thank Bernard Turcotte, Neal Silverman, and Ellora Young for comments on the manuscript and Karl Pfeifer for the [35S]Metlabeled HAP1.

4)

r_ 3..4)

I.-

0

10

2D

30

4

minutes

FIG. 8. On and off rates for HAP1 binding to CYTI and CYCI. Fragments used were the Sma-Taq fragment from CYTI del 202 (-528 to -456) and the XhoI-Hinfl fragment from pLG-312 AAX (13) containing UAS1B. On (A) and off(B) rates were determined as described in Materials and Methods. Heme (15 ,ug/ml) and salmon sperm DNA (0.5 ,ug) were included in the binding reactions. The intensity of the bands on the autoradiograms derived from gel shift CYC1 UAS1; assays was quantitated by densitometry. Symbols: 0, CYTI UAS1. E,

limitation. For its part, CYC7 has a single HAP1 site and an additional site which ensures that the gene is not turned off under anaerobic conditions (35). Our finding that a CYTl gene disruption is a tight mutation which does not revert is consistent with the view that CYTI is the only gene encoding cytochrome cl in S. cerevisiae. The linear reduction in CYTI expression and cytochrome c1 synthesis as oxygen levels fall would match the change in levels of total cytochrome c driven by CYCI and CYC7 combined. A comparison of the HAPl-binding site at CYTI and the strong site at UAS1 of CYCI, UAS1B, reveals an unusual feature. The best alignment of these two sequences shows a 12-of-15-base match which is within the DNase I footprints of both sites (Fig. 7). Surprisingly, although the footprints overlap along most of their lengths, including the homology block, they are offset by several nucleotides at the ends. The contacts as determined by methylation interference at UAS1B and ethylation interference at CYTI show some similarities but also many differences. Most striking is the fact that UAS1B contains several contacts that lie at the end of the DNase I footprint but are outside of the corresponding

REFERENCES 1. Becker, D. M., J. D. Fikes, and L. Guarente. 1991. A cDNA encoding a human CCAAT-binding protein cloned by functional complementation in yeast. Proc. Natl. Acad. Sci. USA, in press. 2. Beckmann, J. D., P. 0. Ljungdahl, L. Lopez, and B. L. Trumpower. 1987. Isolation and characterization of the nuclear gene encoding the Rieske iron-sulfur protein (RIP1) from Sac-

charomyces cerevisiae. J. Biol. Chem. 262:8901-8909. 3. Chodosh, L. A., J. Olesen, S. Hahn, A. S. Baldwin, L. Guarente, and P. A. Sharp. 1988. A yeast and human CCAAT binding protein have heterologous subunits that are functionally interchangeable. Cell 53:25-35. 4. Cumsky, M. G., C. Ko, C. E. Trueblood, and R. 0. Poyton. 1985. Two nonidentical forms of subunit V are functional in yeast cytochrome c oxidase. Proc. Natl. Acad. Sci. USA 82:2235-2239. 5. Cumsky, M. G., C. E. Trueblood, C. Ko, and R. 0. Poyton. 1986. Structural analysis of two genes encoding divergent forms of yeast cytochrome c oxidase subunit V. Mol. Cell. Biol. 7:3511-3519. 6. Dente, L., G. Cesareni, and R. Cortese. 1983. pEMBL: a new family of single-stranded plasmids. Nucleic Acids Res. 11:16451655. 7. Forsburg, S., and L. Guarente. 1988. Mutational analysis of the

upstream activation sequence 2 of the CYCI gene of Saccharomyces cerev'isiae: a HAP2-HAP3-responsive site. Mol. Cell. Biol. 8:647-654. 8. Forsburg, S., and L. Guarente. 1989. Communication between mitochondria and the nucleus in regulation of cytochrome genes

in the yeast Saccharomyces cerevisiae. Annu. Rev. Cell Biol. 5:153-180. 9. Forsburg, S., and L. Guarente. 1989. Identification and characterization of HAP4, a third component in the CCAAT-bound HAP2/3 heteromer. Genes Dev. 3:1166-1178. 10. Galas, D. S., and A. Schmitz. 1978. DNAse I footprinting: a simple method for the detection of protein-DNA binding specificity. Nucleic Acids Res. 5:3157-3170. 11. Guarente, L. Unpublished data. 12. Guarente, L., B. Lalonde, P. Gifford, and E. Alani. 1984. Distinctly regulated tandem upstream activation sites mediate

catabolite repression of the C YCI gene of Saccharomyces

4942

SCHNEIDER ET AL.

cerevisiae. Cell 36:503-511. 13. Guarente, L., and T. Mason. 1983. Heme regulates transcription of the CYCi gene of S. cerevisiae via an upstream activation site. Cell 32:1279-1286. 14. Hahn, S., E. T. Hoar, and L. Guarente. 1985. Each of three "TATA elements" specifies a subset of the transcription initiation sites at the CYCI promoter of Saccharomyces cerevisiae. Proc. Natl. Acad. Sci. USA 82:8562-8566. 15. Haldi, M., and L. Guarente. Unpublished data. 16. Hodge, M. R., G. Kim, K. Singh, and M. G. Cumsky. 1989. Inverse regulation of the yeast COX5 genes by oxygen and heme. Mol. Cell. Biol. 9:1958-1964. 17. Hodge, M. R., K. Singh, and M. G. Cumsky. 1990. Upstream activation and repression elements control transcription of the yeast COX5b gene. Mol. Cell. Biol. 10:5510-5520. 18. Kim, K. S., and L. Guarente. 1989. Mutations that alter transcriptional activation but not DNA-binding in the zinc finger of yeast activator HAP1. Nature (London) 342:200-203. 19. Kim, K. S., K. Pfeifer, L. Powell, and L. Guarente. 1990. Internal deletions in the yeast activator HAP1 have opposite effects at two sequence elements. Proc. Natl. Acad. Sci. USA 87:4524-4528. 20. Kraft, R., J. Tardiff, K. S. Krauter, and L. A. Lainwand. 1988. Using mini-prep plasmid DNA for sequencing double stranded templates with Sequenase. BioTechniques 6:544-549. 21. Loon, A. P. G. M. V., R. J. DeGroot, M. DeHaan, A. Dekker, and L. A. Grivell. 1984. The DNA sequence of the nuclear gene coding for the 17-kDa subunit VI of the yeast ubiquinolcytochrome c reductase: a protein with an extremely high content of acidic amino acids. EMBO J. 3:1039-1043. 22. Maarse, A. C., M. DeHaan, A. Bout, and L. A. Grivell. 1988. Demarcation of a sequence involved in mediating catabolite repression of the gene for the 11 kDa subunit VIII of ubiquinolcytochrome c oxidoreductase in Saccharomyces cerevisiae. Nucleic Acids Res. 16:61-75. 23. Maniatis, T., E. F. Fritsch, and J. Sambrook. 1982. Molecular cloning: a laboratory manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. 24. Matner, R. R., and F. Sherman. 1982. Differential accumulation of two apoiso-cytochromes c in processing mutants in yeast. J. Biol. Chem. 257:9811-9821. 25. Mattoon, J. R., W. E. Lancashire, H. K. Sanders, E. Carvajal, D. R. Malamud, G. R. C. Brag, and A. D. Panek. 1979. Oxygen and catabolite regulation of hemoprotein biosynthesis in yeast, p. 421-435. In W. J. Caughey (ed.), Biochemical and clinical aspects of oxygen. Academic Press, New York. 26. Nobrega, F. G., and A. Tzagoloff. 1980. DNA sequence and organization of the cytochrome b gene in Saccharomyces cerevisiae D273-1OB. J. Biol. Chem. 255:9828-9837. 27. Olesen, J., and L. Guarente. 1990. The HAP2 subunit of yeast CCAAT transcriptional activator contains adjacent domains for subunit association and DNA recognition: model for the HAP2/

MOL. CELL. BIOL.

3/4 complex. Genes Dev. 4:1714-1729. 28. Olesen, J., J. Fikes, and L. Guarente. 1991. The fission yeast homolog of budding yeast HAP2 reveals selective and stringent conservation of the small essential core protein domain. Mol. Cell. Biol. 11:611-619. 29. Olesen, J., S. Hahn, and L. Guarente. 1987. Yeast HAP2 and HAP3 activators both bind to the CYCI activation site, UAS2, in an interdependent manner. Cell 51:953-961. 30. Pfiefer, K., B. Arcangioloi, and L. Guarente. 1987. Yeast HAP1 activation competes with the factor RC2 for binding to the upstream activation site UAS1 of the CYCI gene. Cell 49:9-18. 31. Pfiefer, K., T. Prezant, and L. Guarente. 1987. Yeast HAP1 activator binds to two upstream activation sites of different sequences. Cell 49:19-27. 32. Pfiefer, K. 1988. Coordinate regulation of cytochrome c genes in Saccharomyces cerevisiae by heme and HAP1. Ph.D. thesis. Massachusetts Institute of Technology, Cambridge. 33. Pikielniy, C. W., and M. Rosbash. 1985. mRNA splicing efficiency in yeast and the contribution of nonconserved sequences. Cell 41:119-126. 34. Prezant, T. 1986. Regulation of expression of CYC7 in Saccharomyces cerevisiae. Ph.D. thesis. Massachusetts Institute of Technology, Cambridge. 35. Prezant, T., K. Pfeifer, and L. Guarente. 1987. Organization of the regulatory region of the yeast CYC7 gene: multiple factors are involved in regulation. Mol. Cell. Biol. 7:3252-3259. 36. Sadler, I., K. Suda, G. Schatz, F. Kaudewitz, and A. Haid. 1984. Sequencing of the nuclear gene for the yeast cytochrome cl precursor reveals an unusually complex amino-terminal presequence. EMBO J. 3:2137-2143. 37. Sanger, F., S. Nicken, and A. R. Coulson. 1977. DNA sequencing with chain-terminating inhibitors. Proc. Natl. Acad. Sci. USA 74:5463-5467. 38. Sun, L., and G. Singer. 1975. The specificity of different classes of ethylating agents towards various sites of HeLa cell DNA in vitro and in vivo. Biochemistry 14:1795-1802. 39. Tabor, S., and C. C. Richardson. 1987. DNA sequence analysis with a modified bacteriophage T7 DNA polymerase. Proc. Natl. Acad. Sci. USA 84:4767-4771. 40. Trueblood, C. E., R. M. Wright, and R. 0. Poyton. 1988. Differential regulation of the two genes encoding Saccharomyces cerevisiae cytochrome c oxidase subunit V by heme and HAP2 and REOJ genes. Mol. Cell. Biol. 8:4537-4540. 41. Winkler, H., G. Adam, E. Mattes, M. Schanz, A. Hartig, and H. Ruis. 1988. Coordinate control of synthesis of mitochondrial and non-mitochondrial hemoproteins: a binding site for the HAP1 (CYP1) protein in the UAS region of the yeast catalase T gene (CTT1). EMBO J. 7:1799-1804. 42. Zitomer, R. S., D. L. Montgomery, D. L. Nichols, and B. D. Hall. 1979. Transcriptional regulation of the yeast cytochrome c gene. Proc. Natl. Acad. Sci. USA 76:3627-3631.

4.

Mitochondrial biogenesis requires the coordinate induction of hundreds of genes that reside in the nucleus. We describe here a study of the regulation...
2MB Sizes 0 Downloads 0 Views