MOLECULAR AND CELLULAR BioLOGy, Dec. 1992, p. 5455-5463

Vol. 12, No. 12

0270-7306/92/125455-09$02.00/0 Copyright © 1992, American Society for Microbiology

Histone H3 Transcription in Saccharomyces cerevisiae Is Controlled by Multiple Cell Cycle Activation Sites and a Constitutive Negative Regulatory Element KATIE B. FREEMAN,t LARRY R. KARNS,4 KEVIN A. LUTZ, AND M. MITCHELL SMITH* Department of Microbiology, School of Medicine, University of Virginia, Charlottesville, Virginia 22908 Received 1 June 1992/Returned for modification 15 July 1992/Accepted 31 August 1992

The promoters of the Saccharomyces cereviswae histone H3 and H4 genes were examined for cis-acting DNA sequence elements regulating transcription and cell division cycle control. Deletion and linker disruption mutations identified two classes of regulatory elements: multiple cell cycle activation (CCA) sites and a negative regulatory site (NRS). Duplicate 19-bp CCA sites are present in both the copy I and copy H histone H3-H4 promoters arranged as inverted repeats separated by 45 and 68 bp. The CCA sites are both necessary and sufficient to activate transcription under cell division cycle control. A single CCA site provides cell cycle control but is a weak transcriptional activator, while an inverted repeat comprising two CCA sites provides both strong transcriptional activation and cell division cycle control. The NRS was identified in the copy I histone H3-H4 promoter. Deletion or disruption of the NRS increased the level of the histone H3 promoter activity but did not alter the cell division cycle periodicity of transcription. When the CCA sites were deleted from the histone promoter, the NRS element was unable to confer cell division cycle control on the remaining basal level of transcription. When the NRS element was inserted into the promoter of a foreign reporter gene, transcription was constitutively repressed and did not acquire cell cycle regulation.

The eukaryotic cell division cycle comprises a complex set of interrelated pathways (35). Cell cycle progression is controlled through the actions of both positive and negative regulatory functions resulting in the orderly growth and duplication of the cell with high fidelity (17). Although most cellular mRNAs and proteins are expressed throughout the course of the cell division cycle, a few genes are sensitive to cell cycle position and are only expressed during a specific limited period of the division cycle (for a review, see reference 2). These periodically expressed genes include the major S-phase histone genes (18, 32), genes involved in DNA replication (10, 13, 24, 27), DNA repair (21), mating type control in yeasts (1, 4, 29), and cell cycle transcription control (3, 31). Often the periodic expression of these genes is controlled at the level of transcriptional regulation; however, the signals controlling this expression are largely unknown. Understanding these temporal division cycle signals and the mechanisms by which they are transduced to an active gene promoter remain important and challenging problems for molecular genetics. Several cis-regulatory sequences mediating cell cycle control have been identified. In the yeast Saccharomyces cerevisiae, a conserved hexamer sequence, 5'-ACGCGT-3', is found in the promoter regions of many genes required for DNA synthesis. Deletions, disruptions, and point mutations within these elements in the TMPI, CDC9, and POLl genes abolished the transcriptional activation and periodicity of the promoter. Furthermore, when cloned in front of a reporter gene, copies of this element were able to promote transcription of the reporter gene with a cell cycle periodicity that was coincident with the expression of the native regulated gene *

Corresponding author.

t Present address: Department of Microbiology, Columbia Uni-

versity, New York, NY 10032. 4: Present address: Division of Cardiology, Veterans Administration Medical Center, San Francisco, CA 94121.

5455

(13, 24, 27). The POLI gene also contains an upstream negative regulatory site (NRS) capable of repressing maximal expression from the activation site (13). A second regulatory sequence element, 5'-CACGAAAA-3', is repeated 10 times within the 1,400-bp promoter region of the cell cycle-regulated yeast HO gene (30). When cloned in front of a heterologous reporter gene, this element is sufficient to periodically activate transcription coincident with expression of the HO gene (4). In higher eukaryotes histone subtype-specific regulatory sequences have been identified and implicated in the periodic expression of the replicationdependent histone genes. Factors have been isolated that bind to the H2B, H4, and Hi subtype-specific elements (7, 11, 22), and Hi and H2B subtype-specific factors have been shown to be responsible for cell cycle control (8, 22). By analogy it is predicted that periodic expression of other histone subtypes will be regulated by their subtype-specific elements (18). As in other eukaryotic cells, histone gene expression in actively growing yeast cells is subject to cell cycle regulation. Both the accumulation of histone mRNA and the synthesis of histone protein are tightly controlled with respect to the cell division cycle and are restricted to the late G1- and S-phase periods of the division cycle (6, 19, 20, 23, 25). Osley et al. (33) identified two classes of cis-acting DNA regulatory elements within the promoter region of one of the two yeast histone H2A-H2B gene pairs (TRT1). The first class was a single negative cell cycle regulatory (CCR) region of approximately 70 bp. Deletion of the negative CCR region resulted in constitutive derepression of both the H2A and H2B genes and loss of cell cycle periodicity in synchronous cultures. The CCR was also able to confer periodicity on a reporter gene construct with the CYCI promoter. Thus, the CCR was both necessary and sufficient for periodic cell cycle transcription. The second class of regulatory element was a 16-bp upstream activation sequence (UAS) repeated three times within the promoter. The UAS elements had a

5456

MOL. CELL. BIOL.

FREEMAN ET AL.

positive and additive effect on H2A and H2B gene transcription. These UAS elements were unable to promote cell cycle control of H2A transcription in the absence of CCR function. Thus, cell cycle regulation of the TRT1 promoter is predominantly through the negative control of the CCR. However, multiple copies of a synthetic UAS consensus sequence were able to confer periodic activation on a heterologous gene (33, 34). In the experiments reported here we have sought to identify the cis-acting regulatory sites responsible for periodic cell cycle transcription of the yeast histone H3 and H4 genes. These experiments were designed to define the minimal DNA sequences necessary for gene activation and cell cycle control of transcription, to identify specific regulatory sites within this minimal promoter, and to determine the role of each element in the regulation of histone H3 transcription. The results of these experiments suggest that transcription of the yeast H3 and H4 histone genes is regulated through an NRS and multiple interacting cell cycle activation (CCA) sites. MATERIALS AND METHODS Strains and media. The two yeast strains used in these studies were DBY747 (MA4Ta ura3-52 leu2-3,112 trpl-289 his3) (provided by D. Botstein) and YM147 (MA4Ta gallAI52 ura3-52 trpl-289) (provided by M. Johnston). When assaying for expression from the CYC1 UAS elements, cells were grown in 2% raffinose. In all other cases growth was in 2% glucose. MV medium has been described previously (20). Construction of reporter fusion. An 810-bp AluI fragment spanning the 5' coding and intergenic sequences of the HHT1 and HHFI genes was cloned into a lacZ fusion plasmid, pMC1403 (5), which had been cut with XmaI and had its overhang filled in by using DNA polymerase I. This should have resulted in a fusion of the first 21 amino acids of H3 in frame with the eighth amino acid of lacZ. However, the fusion junction was missing one base from the filled XmaI site. To correct the reading frame, the construct was digested with BamHlI, and the sticky ends were repaired with DNA polymerase I and religated. This resulted in the insertion of an additional 4 bp of linker sequence between the two gene fragments and restored the reading frame. To facilitate manipulation of mutated promoter constructs, the SmaI site 157 bases upstream of the H3 initiation codon was converted to a HindIII site by the insertion of a synthetic linker (Fig. 1A and B). The HindIII-SalI fragment containing the HHTI::lacZ fusion was moved into a singlecopy YIp5-derived vector containing CEN3 and the H4 autonomously replicating sequence to yield plasmid pLK3dl. This construct formed the backbone of the specialized shuttle vector and contained the presumptive HHTJ TATA sequences and the known transcription initiation sites (6). Intergenic sequences from the SmaI site at -157 to the RsaI site at -589 (Fig. 1A) were cloned into the SmaI site of pUC8 in which the Sall site had been destroyed. The intergene sequences could then be subcloned as an EcoRIHindIII fragment into the pLK3-dl shuttle vector to recreate an intact H3 promoter (Fig. 1B). Deletion and linker scanning mutations. The base coordinates of promoter deletion and linker disruption (LD) mutants are summarized in Table 1. Deletions were created 3' to 5' with respect to H3 by cutting the intergene subclone at HindIII and then partially digesting with exonuclease III. The resulting fragments were then blunted by using nuclease Si and digested with EcoRI. Fragments were eluted from

HHF1

CCA Site

Homologies _ __

TATA

Ie

-379 HinPi

am

O

-157 Smal

a

_

U

TATA

CCA

-_

I'* -157 Hindlil

-379

HinPi

A

TATA

_-

m

*' -589 EcoRI

HHTI ::IacZ

CCA Site Homologies

TATA

HHF2

HHTI

-

*m

-589 Real

a

TATA

Site

_X,

+63

Sall

HHT2

Homologies

TATA

-16 BsamHI

-402

HindilI

FIG. 1. Organization of the histone H3-H4 loci. The structures of the promoter regions of the histone genes and fusion constructs are shown for the copy I locus (A), the copy I HHTl::lacZ fusion gene (B), and the copy II locus (C). The locations and directions of the transcription units are indicated by the labeled arrows for each gene, and the positions of coding sequences are shown by the shaded boxes (not to scale). The relative positions and orientations of sequences with homology to the upstream CCA sites are indicated by the open squares and short arrows. Relevant restriction endonuclease sites are indicated on the maps, and their coordinates are numbered relative to the ATG initiation codon of their respective histone H3 (HHT) genes. The TATA sites shown were deduced from sequence analysis and their positions relative to transcription initiation, but they have not been directly tested by experiment.

acrylamide and subcloned into M13mpl8. Endpoints were determined by dideoxynucleotide sequencing (37). Deletions of interest were subcloned into the reporter fusion plasmid pLK3-dl between the HindlIl and EcoRI restriction sites. Deletions in the 5'-to-3' direction were constructed in a similar manner except that the intergene clone was first cut with EcoRI. After resection, XbaI linkers were ligated to the blunt ends and the resulting HindIII-XbaI fragments were cloned into mpl9 for analysis. Linker substitution mutations were constructed by joining HindIII-XbaI 5'-to-3' deletions with appropriate XbaIEcoRI 3'-to-5' deletions. These ligations resulted in the TABLE 1. HHT1 promoter mutations Deletion Deletion

A6 A7

Al A8 A36

A33 A26

A24 A42 AlO A23 A9

All A13 A12

Sequences ~deleted'

LD

None -379 to -589 -157 to -589 -157 to -379 -157 to -166 -157 to -204 -157 to -237 -157 to -274 -157 to -297 -157 to -314 -157 to -319 -157 to -324 -157 to -331 -157 to -356 -157 to -385

LD5 LD22 LD4 LD21 LD2 LD1 LD3 LD33

Sequences disrupted-"' -215 to -241 to -260 to -270 to -304 to -340 to -373 to -270 to

-227 -271 -274 -294 -319 -356 -386 -356

a Base pairs deleted relative to the translational initiation codon of the H3 gene. b Base pairs deleted and replaced by a 15-bp synthetic oligonucleotide linker.

VOL. 12, 1992

replacement of promoter DNA sequences with a 15-bp polylinker that inserted BamHI and XbaI restriction sites at the deletion junctions. The CYCl::galK fusion gene promoter constructs were derived from plasmid YCpR2 (a gift from B. Rymond) (36). Plasmid YCpR2AX was made by deleting the XhoI fragment containing the natural CYCI UAS promoter elements (16). To test the function of the histone H3-H4 UAS elements, the fragments were adapted with synthetic XhoI linkers and inserted into the unique XhoI site of YCpR2AX. Plasmid pKT45 contained the copy I CCA1 sequence (nucleotides 947 to 995 in GenBank accession number X00724), plasmid pKT44 contained the copy I CCA1+CCA2 fragment (nucleotides 886 to 995 in X00724), and plasmid pKT18 contained the copy II CCA1+CCA2 fragment (nucleotides 1035 to 1153 in X00725). To test the function of the HHTJ NRS, the DNA from -157 to -241 was adapted with XhoI linkers and cloned into the XhoI site downstream of the CYCI UAS elements in YCpR2 (16, 36), placing the NRS between the CYC1 UAS elements and the start of transcription. This fusion gene construct was then transferred as a BamHI-KI>nI fragment from the YCpR2 derivative into vector pRS316 (38), creating plasmid pKT103. Sequence comparison. The DNA sequences of the histone promoters were compared by using the SEQH program of Goad and Kanehisa (12). Within the intergene regions of the two promoters the probability of the sequence similarity shared by the copy I and copy II CCA1 elements occurring by chance is less than 8 x 10-8. Cell synchrony. Synchronous cultures were obtained either by induction with a-factor or by selection with centrifugal elutriation. For induced synchrony, cells were arrested in G1 by treatment with a-factor as previously described (19). The a-factor was then removed, and cells were resuspended in fresh MV medium at 3 x 106 cells per ml. Fractions of 1 x 108 to 2 x 108 cells were removed at 15-min intervals following release. For selected synchrony, cell cycle separation was performed by centrifugal elutriation as previously described (14). Small unbudded cells in early G1 were collected by elutriation and then resuspended in fresh medium at 28°C to establish a synchronously growing culture. Fractions of 0.5 x 108 to 1.0 x 108 cells were removed for analysis as the cells progressed through the cell cycle. Fractions were analyzed for bud morphology, cell volume, and DNA content to monitor their synchrony. DNA content was measured by flow cytometry of propidium iodidestained cells (41). Analysis of promoter mutations. Levels of 3-galactosidase were assayed by the method of Guarente (15) and normalized to the number of cells in the reaction and the duration of the reaction. The numbers reported are the averages of triplicate assays which had an average variability of about +9.5%. In separate experiments the ,B-galactosidase activities for a single mutant could vary as much as twofold; however, the relative activities between different mutants were consistent. For RNA analysis, total RNA was extracted from either log-phase or synchronous cells, and mRNA levels were determined either by Northern (RNA) blot hybridizations or primer extension analysis. Northern blot assays were hybridized with DNA fragments specific for the histone H3 or H4 genes, the Escherchia coli galK gene, the CYCI gene, and the ACTI gene. For quantitative primer extension assays, single-strand primers were labeled with [.y-32PJATP by using T4 polynucleotide kinase and then hybridized to single-stranded M13 DNA templates containing subclones of HHT1, URA3,

CELL CYCLE REGULATION OF H3

5457

or lacZ fragments. Following polymerase extension with cold deoxynucleoside triphosphates, the resulting doublestranded product was digested with appropriate restriction endonucleases. The released radiolabeled single strand probes were isolated from 8 M urea-6% polyacrylamide gels. These single-stranded probes were then hybridized to 5 ,g of total RNA in 60% formamide, 0.4 M NaCl, 40 mM piperazine-N,N'-bis(2-ethanesulfonic acid) (PIPES; pH 6.6), and 1 mM EDTA at 80°C for 20 min and then overnight at 42°C. After hybridization, the samples were precipitated and resuspended in 50 mM Tris (pH 8.3), 75 mM KCl, 3 mM MgCl2, 10 mM dithiothreitol, and 400 mM deoxynucleoside triphosphates. Primer extension reactions were carried out with 200 U of murine leukemia virus (MLV) reverse transcriptase (Bethesda Research Laboratories) at 42°C for 90 min. Reaction products were separated on denaturing 8 M urea-6% polyacrylamide gels and visualized by autoradiography. Quantification was performed by using a BioImage Visage 2000 high-resolution video densitometer. Nucleotide sequence accession numbers. All of the DNA sequences used in this report have been published previously (42) and are contained in the GenBank data base. The DNA sequence of the copy I histone H3-H4 locus is filed under accession number X00724 and the copy II DNA sequence is filed under accession number X00725.

RESULTS H3 promoter activity. The genome of haploid S. cerevisiae contains two copies of each of the core histone genes arranged as nonallelic sets of gene pairs: two H2A-H2B gene pairs (TRT1 and TRT2) and two H3-H4 gene pairs (copy I and copy II). The genes in each pair are divergently transcribed from a common promoter region (39). The copy I H3 gene is designated HHT1, and the copy I H4 gene is designated HHFJ (Fig. 1A); the copy II genes are similarly designated HH7I2 and HHF2 (Fig. 1C). We have focused on expression of the HHTI gene and the role of DNA sequence elements within its promoter in transcriptional control. To study the transcriptional regulation of the H3 gene in the absence of posttranscriptional controls involving 3' mRNA sequences (26, 44), the H3-H4 promoter region was placed in front of a foreign reporter gene. This reporter construct was made by fusing nucleotides -589 to +63 of the histone H3 gene to the coding region of the E. coli lacZ gene (Fig. 1B). The histone gene sequences in the fusion included the H3-H4 promoter region and the DNA encoding the first 21 amino acids of the histone H3 protein. The expression of the HHTl::lacZ fusion construct was first analyzed to ensure that transcription initiated correctly and that cell cycle regulation was maintained in the absence of the HHTJ 3' sequences. From the size of the primer extension product produced we calculate that transcription of the HHT1::lacZ fusion gene initiated about 30 bp upstream of the ATG translational initiation codon, in agreement with previous results for the wild-type HHTJ gene (data not shown) (6). Expression of the fusion gene was analyzed in cell cycle fractions selected by centrifugal elutriation and in synchronous cultures induced by a-factor block and release. In both kinds of experiments the cell cycle pattern of expression of the experimental HHTI ::lacZ fusion mRNA was identical to that of the native H3 mRNA (data not shown). From these results we conclude that the 5' sequences included in the fusion construct are sufficient to confer proper initiation and cell cycle-specific transcription. Identification of cis-acting regulatory sites. The positions of

5458

MOL. CELL. BIOL.

FREEMAN ET AL.

Deletion A6 A7 Al A8 A36 A 33 A 26 A 24 A42 AlO A 23 A9 All A13 A12

0-gal

CCA1 CCA2 NRS I

Ii l -

~W

.. -.. .

16.7 10.6 9.1 9.2 6.7 4.6 0.8 0.6

......

I0

Copy-I (CCAl) Copy-I (CCA2) Copy-I

Copy-l1 (CCAI) Copy-Il (CCA2) H3-H4 Consensus

42.0

MJ'.

600

7.1 4.0 0.2 0.2 10.8 20.9

-1 00

FIG. 2. Promoter deletion mutations. The structures of deletion mutations in the promoter of the HHT1::lacZ fusion gene are shown. Coordinates relative to the ATG translational initiation codon of the fusion gene are indicated at the bottom of the diagram. The sequence coordinates of the mutations are given in detail in Table 1. Promoter A6 contains the complete copy I H3-H4 promoter region through the RsaI (EcoRI) site at -589 (see Fig. 1B). The structures of other deletion derivatives are indicated below, with the solid bars representing the sequences retained in the promoter. The relative levels of p-galactosidase activity (P-gal) driven by each promoter are shown to the right of the structure. The relative positions of regulatory sites are indicated by the vertical boxes spanning all constructs. The three sites with sequence homology to the UAS consensus element are shown by the narrow boxes, with the arrows at the top indicating their relative orientations. Two of these sites (CCAl and CCA2) have CCA function and are indicated by the dark shading. The third site has not been shown to have activation function and is unshaded (see text). The approximate location of the NRS is shown by the wider lightly shaded box; the right-hand boundary of the NRS is located between the endpoints of A36 and A33, while the left-hand boundary is located between the endpoints of A26 and A24.

regulatory sequences within the promoter of the HHT1::lacZ fusion gene were localized by restriction fragment deletions, with 0-galactosidase activity as a measure of expression (Fig. 2). A plasmid containing the complete H3-H4 intergene sequence (A6) gave strong expression of the HHTI::lacZ fusion gene, producing approximately 7 U of p-galactosidase activity. A deletion derivative (Al) that lacked the promoter sequences upstream of the SmaI site at -157 produced very low levels of activity, approximately 0.2 U. Since the presumptive TATA element and the known transcription start sites (6) are retained in the Al derivative, additional sequences must be required to activate transcription. A deletion (A8) that carried the far upstream sequences from the HinPI site at -379 to the RsaI site at -589 failed to activate transcription; however, a derivative (A7) that contained sequences from the SmaI site at -157 to the HinPI site provided strong transcriptional activation. These results show that the promoter sequences between nucleotides -157 and -349 contain the major regulatory sequences required for HHTI gene activation. A series of internal deletions through the promoter region were constructed to analyze the major control region (Table 1). These deletions and their levels of expression are illustrated in Fig. 2 (A36-A12). A complex pattern of expression was observed when these deletions were analyzed for n-galactosidase activity. The deletions in A33 and A26 removed

TRT1 Consensus H4TF-2

AAA-1- -.. aC c TG AAAC

FIG. 3. Histone promoter DNA sequence homologies. A comparison is shown of the five sites in the copy I and copy II histone promoters with sequence homology to the copy I UAS element. UAS elements that confer CCA are labeled CCAl and CCA2 for each promoter. The 21 bp initially identified as conserved between the copy I CCAl and the copy II CCA1 sites are shown, and bases that are conserved among the five sites are boxed and shaded. The H3-H4 consensus sequence deduced from these five sites is a 19-bp sequence. The alignments of two other histone promoter elements are shown at the bottom of the diagram: the H2A-H2B UAS consensus sequence (32) and the H4TF-2 binding site from the human histone H4 genes (7). The H4TF-2 binding site is shown as the reverse complement of its traditional orientation to permit alignment with the yeast UAS consensus sequence.

48 and 81 bp beyond the SmaI site, respectively, and both resulted in increased levels of ,B-galactosidase activity. Further deletion of the intergenic sequences (A24 and A42) resulted in a decrease of 3-galactosidase activity, returning to a level of expression comparable to that of the wild-type promoter. Finally, removal of the sequences between nucleotides -332 and -357, from All to A13, reduced 3-galactosidase activity to background levels. These results suggested that the H3 promoter region contained both positive and negative control sequences: an NRS spanning the region defined by A33 and A26 and a UAS defined by the All and A13 constructs. In addition, the decrease in activity seen in going from A26 to A42 suggested the presence of additional positive elements within this region. Promoter sequence comparisons. The promoter sequences of the nonallelic copy I and copy II H3-H4 loci are highly divergent, suggesting that the duplication of the H3-H4 gene pairs occurred early during the evolution of S. cerevisiae (40). We compared the two H3-H4 promoter regions for the presence of putative regulatory sites, guided by the results of the deletion analyses indicating the positions of NRS and UAS elements. Starting with the region between the All and A13 endpoints, computer analysis revealed a 21-bp region of homology with high statistical significance shared between the copy I and the copy II intergenic regions. Further analysis revealed that these sites belonged to a set of 19-bp homologies repeated within both intergenic regions (three times in the copy I intergene region and twice within the copy II intergene region). The locations of the sites in the copy I promoter corresponded to the A26 to A42 and All to A13 activation regions. The sequences of these sites (Fig. 3) are very similar to those of the UAS elements previously identified in the TRT1 H2A-H2B locus (33) and predicted to be in the H3-H4 promoters on the basis of computer sequence analysis (32, 33). Comparison of the histone UAS site consensus sequence with those of the promoters of other cell cycle-regulated genes in S. cerevisiae showed little if any homology. How-

_~ ~

VOL. 12, 1992

CELL CYCLE REGULATION OF H3

the H4 subtype-specific element of higher eukaryotes has an intriguing sequence similarity to these yeast UAS sites (Fig. 3). This sequence homology is especially strong with the human H4 histone clone pHuH4 and corresponds to the sequences protected by the factor H4TF-2 (7). Thus, it is interesting to speculate that the general CCA sequence of the histone genes in Saccharomyces spp. may have evolved to become one of the histone subtype-specific elements in higher eukaryotes. No obvious sequence similarities of high statistical significance were detected between the copy I NRS region and the copy II H3-H4 promoter region. Either the HHTJ NRS is specific for the copy I promoter or the conserved sequence requirements are not easily recognized by the algorithms used for comparison. However, a shorter sequence homology, 5'-ACGCTAAA-3', between the copy I and copy II H3-H4 promoters has been noted to resemble the sequence 5'-ACGCTCAA-3' in the negative CCR region of the TRT1 H2A-H2B promoter (32). This sequence is located from nucleotides -177 to -184 in the copy I promoter, a site towards the right-hand boundary of the NRS within the region deleted in the A33 promoter. LD mutants. Although the previous results were consistent with the presence of UAS and NRS elements, the differences in transcriptional activity might have been caused by changes in the relative positions of other elements. To minimize promoter context effects, LD mutations were targeted to the putative regulatory sites identified by the deletion analysis and DNA sequence comparisons. Specific sequences were replaced by a 15-bp synthetic oligonucleotide linker fragment, thus removing the regulatory site without major alterations in the spacing of the elements. The locations of these linker substitutions and the promoter activities of the mutants are summarized in Fig. 4. The levels of 3-galactosidase were assayed as before, and in addition, for selected mutants the relative levels of the fusion mRNAs were assayed by quantitative Northern blot analysis. The results of the two different assays were in agreement. The linker sequences themselves had no significant effect on the promoter activity since the substitutions in LD2 and LD3 (Fig. 4) did not change the levels of 3-galactosidase expression. The substitutions in LD1, LD21, and LD22 were each designed to remove one UAS homology site. A significant decrease in promoter activity was observed in two of these: the LD1 and LD21 mutations. The LD1 substitution disrupts the consensus sequence in the All to A13 region, and thus the decreased level of expression is consistent with the loss of a UAS element. Similarly, the LD21 substitution disrupts the UAS homology within the A26 to A42 region. This result, together with that for the A42 deletion, is consistent with the presence of a second activation site in this region. Disruption of the third consensus sequence by the linker in LD22 had little effect on transcription, and it may not play a major role in transcriptional activation. Consistent with this interpretation, an internal deletion (LD33) simultaneously removing the other two active UAS elements reduced expression of 3-galactosidase and fusion mRNA to a level less than that of either disruption alone. Thus, the third homology site does not provide strong transcription activation on its own within the H3 promoter. The NRS occupies a region of 36 to 72 bp as defined by the deletion mutations A33 to A24. The presence of the NRS was confirmed by linker substitution LD5. This promoter mutant showed a large increase in the level of transcription of the fusion gene, consistent with the disruption of a negative

CCAI

ever,

CCA2

PI-ai

NRS

__ F-| .-'.5SS_ - _ _ 6im _

A6

LD5

LD22 LD4

LD21

-

LD1

-

_ _

=

1

I

LD3

_

LD33

1.00

1.00

6.58

9.82

0.76

-

047 0.79 0.48

0.43

0.96 0.25

400

mRNA

~~0.82

_~~~~~ _~ .

_

LD2

Zw -

5459

0.32

-200

FIG. 4. LD mutations. The structures of LD mutations in the promoter of the HHTIT::lacZ fusion gene are shown. The region in the diagram begins at -400 and extends to the HindIII site at -157 (Fig. 1B) relative to the ATG translational initiation codon of the fusion gene. The sequence coordinates of the mutations are given in detail in Table 1. The solid bars represent the sequences retained in the promoter, the sawtooth lines represent sequences deleted, and the black boxes represent the 15-bp synthetic linkers. The relative positions of regulatory sites are indicated by the vertical boxes spanning all the constructs. The NRS and the three sites with sequence homology to the UAS consensus element are indicated as described in the legend to Fig. 2. The relative levels of P-galactosidase, normalized to the level expressed by the A6 mutant, are shown to the right of each construct. The relative HHTI::lacZ fusion mRNA levels are shown for selected mutants. The mRNA levels were determined from quantitative Northern blot assays and normalized to the level expressed by the A6 mutant.

regulatory element. Thus, transcription of the histone H3 gene is controlled by both positive and negative cis-acting regulatory sites. The histone UAS elements can activate a foreign gene. The results of the deletion and LD experiments showed that the UAS elements are necessary for activation of histone H3 expression. To determine whether they are sufficient to activate transcription, we subcloned isolated UAS fragments in front of a CYCJ::galK fusion construct (36). A 49-bp subclone of the copy I intergene region containing one of the H3 UAS elements was cloned into the CYCI promoter replacing the normal CYCI UAS (16) to produce plasmid pKT45. This single UAS element activated transcription of the fusion gene and permitted gall yeast cells to grow on galactose medium (36). However, the level of mRNA with this UAS was less than that produced by the normal CYCI promoter (Table 2). We next investigated the effect of the copy I UAS pair on expression. A 110-bp fragment containing the two UAS elements was placed in front of the TABLE 2. Activation of CYCI::galK Promoter

No UAS

Plasmid

YCpR2AX

CYCI UAS YCpR2 pKT45 Copy I CCA1 pKT43 Copy I CCA1+CCA2 pKT18 Copy II CCA1+CCA2 No plasmid None a Levels of mRNA were determined by quantitation

Relative mRNA0

1.0 8.2 3.7 42.0 23.8 0.0

of Northern blots. Expression was normalized to YCpAX, a CYCI promoter mutant lacking any UAS element.

5460

MOL. CELL. BIOL.

FREEMAN ET AL.

CYCJ:.galK fusion gene forming plasmid pKT43. This promoter produced high levels of CYCIJ::galK fusion message, approximately five times greater than that produced from the normal CYCI promoter (Table 2). These results suggested that the inverted repeat UAS elements functioned significantly better than the single UAS. However, the double UAS fragment also contained an additional 47 bp between the two sites, and the high level of transcription could have resulted from the presence of this sequence. Two experiments suggest that the strong activation of the double UAS derivative is the result of the multiple elements and not the DNA sequence between the sites. First, the linker in LD2 (Fig. 4) did not affect transcription of the HHTI ::lacZ fusion gene. This disruption replaced 16 bp of DNA centrally located between the two UAS consensus elements. Second, a 119-bp fragment containing the inverted-repeat UAS elements from the copy II H3-H4 promoter also activated CYClJ::galK transcription to high levels (plasmid pKT18 [Table 2]). Therefore, it is likely that the strong activation observed in these constructs is conferred by the inverted repeat UAS elements and not by the unrelated intervening sequence. The UAS elements are CCA sites. The role of the UAS elements in cell cycle control was investigated by analyzing the activity of promoter mutations in synchronized cultures. Synchronous cultures were established by two independent procedures. In the first case, small unbudded, early-G1 cells were selected by centrifugal elutriation and resuspended in fresh medium to produce a synchronously dividing culture. In the second case, synchronous cultures were induced by blocking cell cycle progression in late G1 with the mating pheromone a-factor and then washing out the pheromone. The DNA histograms for synchronous cultures, either selected by elutriation or induced with a-factor, are shown in Fig. 5. To assay cell cycle control of transcription, samples were removed from synchronous cultures at timed intervals, and RNA from each fraction was subjected to primer extension analysis. Three UAS mutants were analyzed: the two disruptions of the single UAS elements in LD1 and LD21 and the double UAS disruption in LD33. Although either single UAS disruption decreased the overall levels of HHTJ::lacZ fusion transcripts, the LD1 and LD21 promoters both retained some cell cycle control of transcription (data not shown). However, the LD33 promoter lacking both UAS elements was unable to exert cell cycle control over transcription of the fusion mRNA (Fig. 6). These results show that both UAS elements contribute to transcriptional activation and cell division cycle control of HHT1 expression. Although the H2A-H2B UAS elements are unable to provide cell cycle control to the TRT1 promoter in the absence of a functional CCR-negative site, multiple copies of a synthetic consensus oligonucleotide are able to confer cell cycle control on a heterologous lacZ fusion gene (33). To determine whether the H3-H4 UAS elements were sufficient for cell cycle control, we examined their effect on the promoter activity of the CYCIJ::galK fusion gene. Synchronous cell cultures were established by selecting early-G1 cells by using centrifugal elutriation as before. Cell cycle fractions were collected from the cultures at timed intervals, and RNA was extracted for analysis. The levels of mRNA were quantified by Northern blot hybridization with probes for the E. coli galK portion of the fusion mRNA. The transcriptional activity of the promoter with the single histone UAS element is shown as a function of the division cycle in Fig. 7A. Although the single element was a weak

195

FIG. 5. DNA histograms of synchronous cultures. Cell samples at times during the growth of synchronous cultures were stained with propidium iodide, and histograms of DNA content were collected by flow cytometry as described in Materials and Methods. Sets of histograms showing the profile of relative fluorescence for DNA staining are plotted as a function of time after establishing the culture. (A) A synchronous culture selected by centrifugal elutriation; (B) a synchronous culture induced by a-factor block and release.

transcriptional activator (Table 2), it displayed cell cycle control. Transcription was induced with the same cell cycle timing as the internal histone H4 control, although the amplitude of the induction was less than that for the histone H4 control. The transcriptional activity of the promoter containing the two copy I UAS elements is shown in Fig. 7B. This promoter was able to exert strong cell cycle regulation over the transcription of the CYCJ::galK fusion gene, with an amplitude that was indistinguishable from that of the histone H4 control. Identical results were obtained with the copy II double UAS promoter derivative (data not shown). Thus, the UAS elements from either histone H3-H4 promoter region were sufficient to activate transcription in a cell cycle-specific manner. Since these elements are required for histone H3 cell cycle control and are sufficient to confer periodic activation of a foreign gene, we have designated the two elements in each promoter CCA sites, for cell cycle activation. Function of the NRS. In the experiments described above, cell division cycle control was abolished when both the CCA1 and CCA2 sites were deleted from the H3 promoter (Fig. 6). Thus, the NRS element was unable to exert cell cycle control over the remaining basal level of transcription in the absence of the CCA elements. To determine whether the NRS was necessary for H3 cell cycle control, the promoter activity of NRS disruption mutants was examined.

CELL CYCLE REGULATION OF H3

VOL. 12, 1992 Time (min)

A >

URA3

FIG. 9. Promoter activity of the LD5 mutant in synchronous cultures. The transcriptional activity of the LD5 promoter mutant was assayed by primer extension. The LD5 promoter is disrupted for the NRS regulatory element. RNA samples were prepared from cells harvested from synchronously dividing cultures at the times shown, and the levels of mRNA were determined for the histone H3 genes, the HHTI::lacZ fusion gene, and the URA3 gene. (A) A synchronous culture selected by centrifugal elutriation; (B) a synchronous culture induced by a-factor block and release.

DISCUSSION The results of these experiments support several conclusions: (i) transcription of the copy I histone H3 gene is controlled by both positive and negative cis-acting regulatory sites, (ii) the positive CCA sites are both necessary and sufficient to activate periodic cell cycle transcription, (iii) the NRS cannot confer cell cycle regulation on its own in the absence of the CCA sites, (iv) disruption of the NRS causes increased transcription but does not alter cell cycle periodicity, and (v) the NRS represses transcription of a foreign reporter gene fusion but does not confer cell cycle regulation. The TRT1 H2A-H2B and copy I H3-H4 promoters both contain positive and negative regulatory elements; however, Time (min',

**

H3 CYC 1:::galK

ACT1

"I

'

4.

a, ,,

*

FIG. 10. Activity of the NRS in a heterologous promoter. The plasmid pKT103 contains the HI-IT] NRS cloned at the XhoI site downstream of the CYCl UAS, between the UAS and the start of transcription of the CYCI::galK fusion gene. Early-G1 cells containing pKT103 were selected by centrifugal elutriation and used to establish a synchronously dividing culture. RNA samples were prepared from cells harvested at the times shown, and the levels of mRNA were determined by Northern blot analysis for the histone H3 genes, the CYCl::galK fusion gene, and the ACTI gene.

the mechanisms of cell cycle control are different for the two promoters. Regulation of the TRT1 locus is complex. Although multiple copies of a synthetic UAS consensus sequence are able to activate the periodic transcription of a heterologous gene, deletion analysis of the TRT1 promoter shows that the UAS sites do not make a major contribution to the cell cycle regulation of the H2A and H2B genes (33). Deletions that remove one or two of the three TRT1 UAS sites do not alter cell cycle control (for example A4-17 and A6-8 in reference 33). Furthermore, slightly larger deletions that still retain the same one or two UAS sites have nevertheless lost cell cycle control (for example, A10-18 and A8-5 in reference 33). Finally, an internal deletion of 54 bp within the 70-bp CCR (A16' in reference 33) leads to a loss of cell cycle control despite all three UAS sites remaining intact in this promoter. Thus, cell cycle regulation of the TRT1 promoter is exerted by periodic repression through the CCR site and, in the absence of this negative function, the UAS elements are unable to provide this control. In contrast, the CCA sites in the HHTI promoter are both necessary and sufficient for periodic activation. Simultaneous deletion of both CCA1 and CCA2 abolishes cell cycle regulation of the H3 promoter. Conversely, in the absence of NRS function the promoter still retains cell cycle control. There are several possible explanations for the functional differences between the H2A-H2B and the H3-H4 positive regulatory sites. For example, the TRT1 promoter may contain additional unidentified regulatory functions that suppress the cell cycle periodicity of the UAS elements (see reference 33 for discussion). Alternatively, the H2A-H2B UAS and the H3-H4 CCA elements may bind different members of a related family of transcription factors that are able to discriminate between similar DNA binding sites. Although similar in size, the functions of the negative regulatory elements in the H2A-H2B and the H3-H4 promoters are also different. The TRT1 CCR is essential for cell cycle transcription of the H2A and H2B genes. In addition, when the negative CCR site is inserted into the CYCl promoter, it is sufficient to produce periodic transcription of a CYCI::lacZ fusion gene (33). In contrast, disruption of the H3 NRS function, either by deletion or by linker substitution, has no effect on the periodicity of HHT1 transcription within the cell cycle. Furthermore, when the NRS is inserted into the CYCI promoter, it represses transcription of the CYCJ::lacZ fusion gene but does not establish cell cycle periodic control. The TRT1 CCR site has also been shown to be important for the dosage compensation response of the promoter to H2A-H2B gene copy number (28). The copy I H3-H4 promoter does not show dosage compensation at the level of promoter strength in response to deletions of the copy II H3-H4 genes (6). Therefore, we currently favor the hypothesis that the histone H3 NRS element is involved in controlling the basal level of expression of the copy I genes, perhaps functioning in quiescent cells to prevent expression (9, 22), or controlling the level of histone expression necessary for chromatin maintenance, a role played by specialized histone subtype genes in higher eukaryotes (43). ACKNOWLEDGMENTS We thank our colleagues for help and discussions during the course of this work, B. C. Rymond and R. Zitomer for the gift of their CYCI :galK fusion construct, Chris Eichman for expert technical assistance with the flow cytometry, and D. Engel and B. Spoth for critical reading of the manuscript. This work was supported by National Institutes of Health grant GM28920, awarded to M.M.S.

VOL. 12, 1992

REFERENCES 1. Andrews, B. J., and I. Herskowitz. 1989. The yeast SW14 protein contains a motif present in developmental regulators and is part of a complex involved in cell-cycle-dependent transcription. Nature (London) 342:830-833. 2. Andrews, B. J., and I. Herskowitz. 1990. Regulation of cell cycle-dependent gene expression in yeast. J. Biol. Chem. 265: 14057-14060. 3. Breeden, L., and G. E. Mikesell. 1991. Cell cycle-specific expression of the SWI4 transcription factor is required for the cell cycle regulation of HO transcription. Genes Dev. 5:1183-1190. 4. Breeden, L., and K. Nasmyth. 1987. Cell cycle control of the yeast HO gene cis- and trans-acting regulators. Cell 48:389-397. 5. Casadaban, M. J., J. Chou, and S. N. Cohen. 1980. In vitro gene fusions that join an enzymatically active ,-galactosidase segment to amino-terminal fragments of exogenous proteins: Escherichia coli plasmid vectors for the detection and cloning of translational initiation signals. J. Bacteriol. 143:971-980. 6. Cross, S. L., and M. M. Smith. 1988. Comparison of the structure and cell cycle expression of mRNAs encoded by two histone H3-H4 loci in Saccharomyces cerevisiae. Mol. Cell. Biol. 8:945-954. 7. Dailey, L., S. M. Hanly, R G. Roeder, and N. Heintz. 1986. Distinct transcription factors bind specifically to two regions of the human histone H4 promoter. Proc. Natl. Acad. Sci. USA 83:7241-7245. 8. Dalton, S., and J. R. Wells. 1988. Maximal binding levels of an Hi histone gene-specific factor in S-phase correlate with maximal Hi gene transcription. Mol. Cell. Biol. 8:4576-4578. 9. Drebot, M. A., L. M. Veinot-Drebot, R A. Singer, and G. C. Johnston. 1990. Induction of yeast histone genes by stimulation of stationary-phase cells. Mol. Cell. Biol. 10:6356-6361. 10. Elledge, S. J., and R. W. Davis. 1990. Two genes differentially regulated in the cell cycle and by DNA-damaging agents encode alternative regulatory subunits of ribonucleotide reductase. Genes Dev. 4:740-751. 11. Gallinan, P., F. LaBella, and N. Heintz. 1989. Characterization and purification of H1TF2, a novel CCAAT-binding protein that interacts with a histone Hi subtype specific consensus element. Mol. Cell. Biol. 9:1566-1575. 12. Goad, W. B., and M. I. Kanehisa. 1982. Pattern recognition in nucleic acid sequences. I. A general method for finding local homologies and symmetries. Nucleic Acids Res. 10:247-263. 13. Gordon, C. B., and J. L. Campbell. 1991. A cell cycle-responsive transcriptional control element and a negative control element in the gene encoding DNA polymerase a in Saccharomyces cerevisiae. Proc. Natl. Acad. Sci. USA 88:6058-602. 14. Gordon, C. N., and S. G. Elliot. 1977. Fractionation of Saccharomyces cerevisiae cell populations by centrifugal elutriation. J. Bacteriol. 129:97-100. 15. Guarente, L. 1983. Yeast promoters and lacZ fusions designed to study expression of cloned genes in yeast. Methods Enzymol. 101:181-191. 16. Guarente, L., and M. Ptashne. 1981. Fusion of Escherichia coli lacZ to the cytochrome c gene of Saccharomyces cerevisiae. Proc. Natl. Acad. Sci. USA 78:2199-2203. 17. Hartwell, L. H., and T. A. Weinert. 1989. Checkpoints: controls that ensure the order of cell cycle events. Science 246:629-634. 18. Heintz, N. 1991. The regulation of histone gene expression during the cell cycle. Biochim. Biophys. Acta 1088:327-339. 19. Hereford, L., S. Bromley, and M. A. Osley. 1982. Periodic transcription of yeast histone genes. Cell 30:305-310. 20. Hereford, L. M., M. A. Osley, J. R Ludwig, and C. S. McLaughlin. 1981. Cell-cycle regulation of yeast histone mRNA. Cell 24:367-375. 21. Kupiec, M., and G. Simchen. 1986. Regulation of the RAD6 gene of Saccharomyces cerevisiae in the mitotic cell cycle and in meiosis. Mol. Gen. Genet. 203:538-543. 22. LaBella, F., H. L. Sive, R. G. Roeder, and N. Heintz. 1988. Cell-cycle regulation of a human histone H2B gene is mediated by the H2B subtype-specific consensus element. Genes Dev. 2:32-39.

CELL CYCLE REGULATION OF H3

5463

23. Lorincz, A. T., M. J. Miller, N. Xuong, and E. P. Geiduschek. 1982. Identification of proteins whose synthesis is modulated during the cell cycle of Saccharomyces cerevisiae. Mol. Cell. Biol. 2:1532-1549. 24. Lowndes, N. F., A. L. Johnson, and L. H. Johnston. 1991. Coordination of expression of DNA synthesis genes in budding yeast by a cell-cycle regulated trans factor. Nature (London) 350:247-250. 25. Ludwig, J. R, and C. S. McLaughlin. 1982. Periodic synthesis of histone proteins through the cell cycle of Saccharomyces cerevisiae as determined by centrifugal elutriation, p. 113-121. Proceedings of the Berkeley Workshop on Recent Advances in Yeast Molecular Biology: Recombinant DNA. University of California, Berkeley. 26. Lycan, D. E., M. A. Osley, and L. M. Hereford. 1987. Role of transcriptional and posttranscriptional regulation in expression of histone genes in Saccharomyces cerevisiae. Mol. Cell. Biol. 7:614-621. 27. McIntosh, E. M., T. Atkinson, R. K. Storms, and M. Smith. 1991. Characterization of a short, cis-acting DNA sequence which conveys cell cycle stage-dependent transcription in Saccharomyces cerevisiae. Mol. Cell. Biol. 11:329-337. 28. Moran, L., D. Norris, and M. A. Osley. 1990. A yeast H2A-H2B promoter can be regulated by changes in histone gene copy number. Genes Dev. 4:752-763. 29. Nasmyth, K. 1983. Molecular analysis of a cell lineage. Nature (London) 302:670-676. 30. Nasmyth, K. 1985. At least 1400 base pairs of 5' flanking DNA is required for the correct expression of the HO gene in yeast. Cell 42:213-223. 31. Nasmyth, K., A. Seddon, and G. Ammerer. 1987. Cell cycle regulation of SWI5 is required for mother-cell-specific HO transcription in yeast. Cell 49:549-558. 32. Osley, M. A. 1991. The regulation of histone synthesis in the cell cycle. Annu. Rev. Biochem. 60:827-861. 33. Osley, M. A., J. Gould, S. Kim, M. Kane, and L. Hereford. 1986. Identification of sequences in a yeast histone promoter involved in periodic transcription. Cell 45:537-544. 34. Osley, M. A., and D. Lycan. 1987. trans-acting regulatory mutations that alter transcription of Saccharomyces cerevisiae histone genes. Mol. Cell. Biol. 7:4204-4210. 35. Pringle, J. R., and L. H. Hartwell. 1981. The Saccharomyces cerevisiae cell cycle, p. 97-142. In J. N. Strathern, E. W. Jones, and J. R. Broach (ed.), The molecular biology of the yeast Saccharomyces cerevisiae: life cycle and inheritance. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. 36. Rymond, B. C., R. S. Zitomer, D. Schumperli, and M. Rosenberg. 1983. The expression in yeast of the Escherichia coli galK gene on CYCI ::galK fusion plasmids. Gene 25:248-262. 37. Sanger, F. S., S. Nicklen, and A. R. Coulson. 1977. DNA sequencing with chain-terminating inhibitors. Proc. Natl. Acad. Sci. USA 74:5463-5467. 38. Sikorski, R. S., and P. Hieter. 1989. A system of shuttle vectors and yeast host strains designed for efficient manipulation of DNA in Saccharomyces cerevisiae. Genetics 122:19-27. 39. Smith, M. M. 1984. The organization of the yeast histone genes, p. 3-33. In G. S. Stein, J. L. Stein, and W. F. Marzluff (ed.), Histone genes. John Wiley & Sons, Inc. New York. 40. Smith, M. M. 1987. Molecular evolution of the Saccharomyces cerevisiae histone gene loci. J. Mol. Evol. 24:252-259. 41. Smith, M. M. 1991. Mutations that affect chromosomal proteins in yeast. Methods Cell Biol. 35:485-523. 42. Smith, M. M., and 6. S. Andr6sson. 1983. DNA sequences of yeast H3 and H4 histone genes from two non-allelic gene sets encode identical H3 and H4 proteins. J. Mol. Biol. 169:663-690. 43. Wu, R. S., and W. M. Bonner. 1981. Separation of basal histone synthesis from S-phase histone synthesis in dividing cells. Cell 27:321-330. 44. Xu, H., L. Johnson, and M. Grunstein. 1990. Coding and noncoding sequences at the 3' end of yeast histone H2B confer cell cycle regulation. Mol. Cell. Biol. 10:2687-2694.