Functional Domains of the Yeast Transcription/Replication Factor MCM1

Downloaded from genesdev.cshlp.org on October 6, 2021 - Published by Cold Spring Harbor Laboratory Press

Functional domains of the yeast transcription/replication factor MCM1

Chantal Christ and Bik-Kwoon Tye Section of Biochemistry, Molecular and Cell Biology, Comell University, Ithaca, New York 14853 USA

MCM1 is an essential yeast DNA-binding protein that affects both minichromosome maintenance, in a manner suggesting that it has DNA replication initiation function, and gene expression. It activates a-specific genes together with MATod, and represses a-specific genes together with MAToL2. Alone, MCMI can activate transcription. To determine whether different domains of the protein mediate these diverse functions, we constructed and analyzed several mcml mutants. The gene expression and minichromosome maintenance phenotypes of these mutants suggest that the role of MCM1 in DNA replication initiation may not involve transcriptional activation. However, both transcription and replication activities require only the 80-amino-acid fragment of MCM1 homologous to the DNA-binding domain of the serum response factor (SRF). This small fragment is also sufficient for cell viability and repression of a-specific genes. A polyacidic amino acid stretch immediately adjacent to the SRF homologous domain of MCM1 was found to be important for activation of a-specific genes in ,v cells. Mutants lacking the acidic stretch confer higher expression from an a-specific UAS in a cells in addition to lower expression in oL cells, suggesting that negative regulation at this site occurs in a cells, in addition to the well-documented positive regulation in oL cells. [Key Words: Replication factor MCM1; transcription activation; replication initiation; regulation] Received January 15, 1991; revised version accepted February 15, 1991.

It has become increasingly clear that for many of the at particular binding sites has been proposed to be mod- sequence-specific DNA-binding proteins, the same pro- ulated by binding site context and by interaction with tein can have multiple, diverse activities. The Droso- other DNA-binding proteins. Genetic and biochemical phila homeo box proteins and the steroid hormone re- evidence suggests that MCM1 activity at mating-type- ceptors activate transcription at some promoters and re- specific genes is determined by cooperative binding with press transcription at others (Levine and Manley 1989). cofactors, either with MAT~I at c~-specific genes, which The NFI and NFIII proteins, which enhance replication results in activation, or with MAT~2 at a-specific genes, initiation of adenoviruses, are also transcriptional acti- which results in repression (Jarvis et al. 1989; Keleher et vators (Challberg and Kelly 1989). Two yeast proteins, al. 1989; Passmore et al. 1989; Ammerer 1990). The com- RAP1 and ABF1, bind to silencers as well as upstream bination of these activities allows ~ cells to express the activating sequences (UASs), and may contribute either gene set required for mating with a cells. The mecha- to repression or activation of transcription depending on nism by which the different complexes act to either re- the binding site context (Shore and Nasmyth 1987; press or activate and whether MCM1 activity at other Buchman et al. 1988). In addition, RAP1 binds to se- sites is also modulated by cofactor interactions are un- quences found at telomeres, whereas ABF1 binds to sev- known. eral autonomously replicating sequences (ARSs) (Buch- MCM1 was first identified as a putative replication man et al. 1988). Both proteins may be involved in mul- factor. The rectal-1 allele was isolated in a screen for tiple regulatory processes. The yeast MCM 1 protein is a mutants defective in minichromosome maintenance transcription activator that also affects DNA replication (Maine et al. 1984). Stability of a minichromosome, and acts both as a corepressor and a coactivator in regu- which contains only one ARS, a centromere, and select- lation of mating-type-specific genes (Passmore et al. able markers, is extremely sensitive to mutations that 1988; 1989; Jarvis et al. 1989; Keleher et al. 1989; Am- affect chromosome replication or segregation. Minichro- meter 19901. The diverse activities, coupled with the mosomes carrying different ARSs, which have similar small size, of MCM1 make it a particularly interesting stability in wild-type strains, are affected differently in protein for structure and function analysis. Analysis of the rectal-1 mutant. The loss rate of a minichromosome MCM1 and other multifunctional proteins may reveal varies >10-fold, depending on the ARS sequence used, common mechanisms by which the activity of these pro- suggesting that rectal-1 affects ARS function (Maine et teins is regulated. al. 1984). Because ARSs have been shown to be the in The activity of multifunctional DNA-binding proteins vivo origins of DNA replication (Brewer and Fangman

GENES & DEVELOPMENT 5:751-763 © 1991 by Cold Spring Harbor Laboratory Press ISSN 0890-9369/91 $3.00 751 Downloaded from genesdev.cshlp.org on October 6, 2021 - Published by Cold Spring Harbor Laboratory Press

Christ and Tye

1987; Huberman et al. 1987; 1988), the rectal-1 pheno- acidic or polyglutamine domains, have substitutions of type suggests that MCM1 is important for replication portions of other transcription factors, or have insertions initiation. MCM1 may affect ARS function by binding to into the putative DNA-binding domain, as shown in Fig- the origin as part of the replication initiation complex or ure 1. We made mutations that alter the coding sequence it may act indirectly by affecting the expression of an- with minimal changes in the transcript to avoid alter- other protein that acts at ARSs. MCM1 is essential for ations that might affect transcript stability. The cell division, although whether it is required for replica- mcml-AQ allele codes for a protein with no polyglu- tion initiation or for regulation of other essential genes is tamine domain, mcml-ADE removes only the acidic unknown. stretch, and mcml-ADEQ codes for a protein lack- Although MCM1 is a small protein with only 286 ing both acidic and polyglutamine domains. The amino acids (Passmore et al. 1988), it may be similar to mcml-ANlzDEQ mutant has an additional deletion of other eukaryotic transcription factors in having separate amino acids 2-17, leaving only the SRF-homologous do- DNA-binding, transcription activation, and regulatory main of MCM1. The mcml-SRF/DE and mcml-gcn4/ domains. If so, different domains may be important for DE(Q) alleles were constructed to assay whether the its different activities. MCM1 is a member of a gene fam- acidic stretch could be replaced with a less acidic se- ily including the serum response factor (SRF)(Norman et quence or with a known acidic activating sequence. The al. 1988), another yeast gene, ARG80 (ARG RI) (DuBois portion of SRF immediately following the SRF/MCM1- et al. 1987; Passmore et al. 1988), and several plant ho- homologous domain was inserted in place of the acidic meotic genes (Sommer et al. 1990; Yanofsky et al. 1990). stretch in mcm 1-SRF/DE, resulting in 7/34 acidic amino The proteins all share homology in a region that was acids, as compared to 20/23 acidic amino acids in identified as the DNA-binding domain of SRF. The MCM1. A portion of GCN4 characterized previously as amino-terminal portion of MCM1 {amino acids 18-98) is an acidic activation domain (Hope et al. 1988) was sub- 70% identical to the portion of SRF that was shown to be stituted for the acidic stretch and part of the polyglu- essential for DNA binding and dimerization {Norman et tamine region of MCM1 in mcml-gcn4/DE(Q) (see leg- al. 1988). The two proteins share similar DNA-binding end to Fig. 1). The mcml-Xho92 and rectal-Barn92 al- specificity, each binding to the palindromic recognition leles have four amino acids (either AlaArgAlaAla or sequence CC(A/T)6GG {Norman et al. 1988; Passmore et ArgIleArgAla) inserted in-frame after amino acid 92, in al. 1989). Therefore, the homologous domain is likely to the region of MCM1 homologous to SRF. be the DNA-binding domain of MCM 1 as well. Immedi- Because MCM1 is an essential gene, we tested whether ately following the SRF homologous domain of MCM1 yeast with each of the mutant genes alone is viable, us- are 20 amino acids, of which 19 are either aspartate or ing the plasmid shuffle assay (Boeke et al. 1987)in a glutamate. This region may be an acidic activation se- strain with a chromosomal deletion of MCM1. We found quence, although the acidic residues are much more that all of the mutants except mcml-Xho92 and rectal- highly clustered than acidic activation domains of other Barn92 provide sufficient MCM1 function for viability. transcription factors such as GAL4 and GCN4 (Hope and This result localizes the essential domain of MCM1 to Struhl 1986; Keegan et al. 1986). The carboxy-terminal the 80 amino acids homologous to SRF. half of the protein is 50% glutamine and may behave like Because mcml-Xho92 and rectal-Barn92 are unable the glutamine-rich activation domains of Spl and AntP to rescue lethality of an mcml deletion, even in high (Courey and Tjian 1988; Courey et al. 1989). Although copy, we tested whether these mutations affect the truncated genes that lack much of the polyglutamine amount or the activity of the resulting mutant MGM1 domain are functional (Passmore et al. 1988; Jarvis et al. protein. Immunoblot analysis of a wild-type strain con- 1989), the effect of removing the entire polyglutamine taining these high-copy constructs shows that the mu- domain has not been tested previously. tant proteins are present at similar levels as wild-type In this study we specifically alter or delete portions of MCM1 on a similar plasmid (Fig. 2), suggesting that MCM1 which, as described above, may be separate func- these mutations do not affect the stability or expression tional domains. We analyze the effect of these mutations of MCM1. Therefore, these mutations must affect on the known activities of MCM1 in vivo: transcription MCM1 activity, perhaps by interfering with dimeriza- activation, minichromosome maintenance, and regulation, by analogy to the role of the homologous region of tion of mating-type-specific genes. This mutant analysis SRF {Norman et al. 1988). should reveal whether different functions of MCM1 require different domains, and provide insight into how MCMl-dependent gene expression and MCM1 affects replication initiation and transcription minichromosome maintenance in mcml mutants are activation and how it may interact with cofactors to me- uncorrelated diate gene regulation. Because MCM1 affects plasmid stability, we analyzed Results the phenotypes of the viable mutants by replacing the wild-type MCM1 gene on chromosome XIII with each of The SRF-homologous domain of MCM1 is sufficient the mutated genes (as described in Materials and meth- for viability. ods). We used a haploid MAT~ strain, into which we The MCM1 mutant proteins constructed are missing the similarly introduced the mcml-1 point mutation. The

752 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on October 6, 2021 - Published by Cold Spring Harbor Laboratory Press

Functional domains of MCM1

Provides viability 17 97 120 286 I I I [ MCM1 + i .ihN Nx \ xx*X~xxx SRF related acid ~'polyglutamine

MCM 1-AQ + ~ r, ,-, ...... ~ RIGRIR

MCM1-ADE + I ixx-\\ \\\ \\ -~/A'l~ Figure 1. Schematic of mutant proteins. For MCM1 sequence, see Passmore et al. {19881]. The numbers above the MCM1 se- MCM1-ADEQ + , ~\x~xxxxxxxx,Vy quence refer to the last amino acid in the boxed section. MCM 1-AQ is altered after Gly 118, with the ArgIleGlyArgIleArg non- MCM 1-AN17DEQ + ~ ...... \--,VY native sequence added followed by a stop codon. MCM1-ADE has amino acids 98- P97 ~L 120 replaced with ValAlaThr. MCM1- ADEQ introduces ValTyrSTOP after MCMI-1 + , Pro97, whereas MCM1-ANxTDEQ has an additional deletion of amino acids 2-17. MCMI-1 is a change of Pro97 to Leu. MCM1-SRF/DE + , MCM1-SRF/DE introduces Val after Pro97, followed by amino acids 223-251 of SRF {Norman et al. 19881, and then Asp- MCM1-GCN4/DE(Q) + , SerThr and amino acids 121-286 of MCM1. MCM1-GCN4/DE(Q) inserts ARAA ValAspAlaPro after Pro97, followed by amino acids 85-150 of GCN4 {Hope et al. MCM 1-Xho92 - 1988} and amino acids 153-286 of MCM1. MCM1-Xho92 and MCM1-Bam92 are RIRA both four amino acid in-frame insertions MCM 1-Bam92 - , ~.. ~~ ~ -, -, -, -,~'~ after amino acid 92 of MCM1, as shown.

final result is a set of isogenic strains differing only in the mcml-1 mutant has the lowest expression of ~- MCM1 allele. Growth curves showed that each mutant galactosidase at 6% wild-type levels, whereas the SRF has a doubling time of -1.5 hr at 30°C, similar to wild substitution for the acidic domain has greater activity type, and none of the mutants are heat or cold sensitive than wild-type MCM1 and the GCN4 substitution re- for growth (data not shown I. Therefore, even when the duces activity to intermediate levels (22% of wild typel. mutant genes are present in single copy, no growth de- These results suggest that the acidic stretch is not re- fect is observed. quired for transcription activation, because its deletion The ability of each mutant MCM1 protein to activate has no effect, whereas the polyglutamine domain may transcription was monitored using the DSEI4-1acZ re- contribute to activation, because expression is reduced porter gene shown in Figure 3, integrated in single copy twofold in this mutant. The even lower activity in mu- on chromosome III. Expression of []-galactosidase is ab- tants lacking both domains may be due to instability of solutely dependent on MCM1 (R. Elble, in prep.I and is these truncated proteins, which will be addressed below. not cell type specific. The level of B-galactosidase in Because the mutants vary widely in expression level wild-type a or ~ cells is 40-fold higher from the reporter from an MCMl-dependent promotor, we tested them for containing the MCMl-binding site than from an other- minichromosome maintenance (Mcm), to determine wise identical construct lacking it. The mcml mutants whether the two phenotypes are equally affected in each have varying effects on expression from this promoter mutant. We transformed each strain with YCpl01 (Table 11. Removal of only the polyglutamine domain (ARS1, CENS, LEU21 and calculated the loss rate of the (mcml-AQ} results in a decrease in ~-galactosidase ex- minichromosome per cell per generation of nonselective pression to 40% of wild-type levels, whereas removal of growth. The loss rates are also shown in Table 1. Al- the acidic domain (mcml-ADE} results in wild-type ex- though the two smallest mutant proteins, MCM1- pression levels. Removal of both acidic and polyglu- ADEQ and MCM1-AN 17DEQ, confer a somewhat higher tamine domains {mcml-ADEQI results in a further drop loss rate than wild-type (cf. 0.07 and 0.06 with 0.02}, only in B-galactosidase activity to 18% of wild-type levels. rectal-1 and mcml-gcn4/DE(Q) have severe Mcm de- Additional removal of amino acids 2-17 has no effects fects Iloss rates of 0.30 and 0.21, respectively). activity in mcml-ANlzDEQ is 20% wild type. The There is no correlation between the gene expression

GENES & DEVELOPMENT 753 Downloaded from genesdev.cshlp.org on October 6, 2021 - Published by Cold Spring Harbor Laboratory Press

Christ and Tye

do not all merely affect MCM1 protein stability, because the two phenotypes show different severity in the same mutant. We confirmed this prediction by testing whether any of the mutations result in lower levels of the mutant protein. Immunoblot analysis of a wild-type

oo strain in which the mutant proteins are expressed from [.v ,o high-copy 2-micron plasmids is shown in Figure 4. The MCM1-1 protein is present at greater than wild-type levels, and MCM1-ADE, MCM1-SRF/DE, and 45 ld) MCM1-GCN4/DE(Q) are present at slightly lower levels. MCM1-AQ appears less abundant, whereas MCM1-ADEQ is not detectable under these conditions. These small mutant proteins may be less abundant or may have different transfer properties than the larger o 29kD -- mutants. MCM1-AN17DEQ was not tested because mcml-AN~ 7DEQ and mcml-ADEQ have identical phenotypes, and our antiserum, which was generated against a fusion protein with the first 92 amino acids of MCM1, may not bind quantitatively to the smaller fragment. The apparently low level of MCM1-ADEQ protein is consistent with the previous result that mcml-ADEQ has a more severe phenotype than mcml-AQ, even though deletion of the acidic stretch Figure 2. MCM1-Xho92 and MCM1-Bam92 are present at alone, in mcml-ADE, has no effect. wild-type levels. Immunoblot using anti-f3-gal-MCMll_92 is To distinguish whether the mutations primarily affect shown of strain BJ2168 transformed with high-copy YEp351- based plasmids. Transformants were grown in SC-uracil media binding affinity or the activity of the protein once bound to OD6oo 1.5. Equal volumes of yeast crude extracts were loaded to DNA, we tested the effect of overproducing the mu- in each lane; however, staining of the gel reveals that the tant proteins on expression from DSEIa-lacZ. DSE14 is a YEp351-mcml-Bam92 lane had approximately twice the pro- high-affinity binding site for MCM1. Overproduction of tein as the others. MCM1 runs at -40 kD on this gel and is not wild-type protein, by expression from a 2-micron vector, visible in the YEp351 lane, which has extract from the strain results in at least fivefold more MCM1 protein (Fig. 4) with only the single-copy chromosomal MCMI gene. but only 1- to 1.7-fold higher levels of 13-galactosidase from a DSE14-1acZ reporter (Passmore et al. 1989), sug- and minichromosome maintenance defects: Although gesting that this binding site is saturated at single-copy the mutants with high levels of expression from DSE14 gene dosage of MCM1. If transcription activation is re- also maintain minichromosomes well, the mutants that duced in a mutant because of reduced occupancy of the affect both phenotypes affect them with different sever- DSE14-binding site, due to either reduced DNA-binding ity. For example, mcml-ADEQ, mcml-AN~TDEQ, and affinity or lower protein level, overproduction of the mu- mcml-gcn4/DE(Q) have similar levels of 13-galacto- tant protein should result in increased B-galactosidase sidase expression from DSE~4-1acZ (-20% wild type), activity. Conversely, if the mutant protein is less active yet the deletion mutants maintain minichromosomes but binds well, J3-galactosidase expression should remain much better than the GCN4 subsitution mutant (loss low, regardless of overproduction of the mutant protein. rate of YCpl01 is 0.07 vs. 0.21). Furthermore, because Three of the mutants tested show increased activity mcml-1 is temperature sensitive for expression of when overproduced on 2-micron plasmid (Table 2). The a-specific genes (C. Christ, unpubl.), we tested gene ex- largest increase occurs with overproduction of MCMI-1 pression and minichromosome maintenance at room protein in the mcml-1 strain. Expression from DSE14 is temperature as well as at 30°C. At the lower tempera- increased over eightfold, nearing wild-type levels. This ture, expression of DSE14-1acZ in mcml-1 improves to a result is consistent with an observation by Keleher et al. level similar to that in mcml-ADEQ (16% and 19% wild (1988) that mcml-1 mutant extracts bind the STE2 op- type), but the two mutants are still very different in their erator in DNA band-shift assays approximately fivefold ability to maintain minichromosomes (loss rate of less efficiently than wild-type extracts and suggests that YCpl01 is 0.23 for mcml-1 and 0.08 for mcml-ADEQ). the mcml-1 mutation affects binding affinity. These results suggest that the minichromosome main- J3-Galactosidase expression in the mcml-ADEQ strain tenance defect of mcml mutants is not simply due to increases fourfold when MCM1-ADEQ is overproduced, reduced gene expression by the mutant MCM1 protein. reaching 58% of wild-type expression levels, consistent with the previous result that this truncation affects MCM1 protein level. Activity also increases threefold in Restoration of MCMt activity by overproduction of the GCN4-substituted mutant with overproduction of mutant proteins that construct, to even higher levels than wild-type The preceding experiments suggest that the mutations MCM1, suggesting that the insertion of the GCN4 acti-

754 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on October 6, 2021 - Published by Cold Spring Harbor Laboratory Press

Functional domains of MCM1

Non-cell-type-specific: DSE14-1acZ

Hind I11 STE2-1acZ

TCGATI'TCCTAATTAGGAAAAGCT c~-specific: MFct1PQ-lacZ

SalI HindIII STE2-1acZ ! ,

GTCGACGGAACACC'IWCCTAATTAGGI~ATI~AACGACAGTAAATrCCCAAGCTT P Q

a-specific: STE2-1acZ Figure 3. Tester genes for measuring MCMl-dependent transcription. The binding sites DSEt4 and MF~IPQ are de- Sal I HindlII HindlII STE2-1acZ scribed in Passmore et al. (1989), and STE2-1acZ is described in Smith (1986). URS~ MCMl-binding sites are shown in bold, ~ and c,1 or ~2 recognition elements are un- derlined. All constructs are cloned into the 0.3kb...ACCATGTAAATrTCCTAATTGGGTAAQTACATGATGAAACACATATGAAGAAAAAAGCIT SalI site downstream of LEU2 and inte- ~x2 o~2 grated in single copy on chromosome III. vation domain into MCM1 has two opposing effects on whereas the polyglutamine domain also contributes to gene expression from an MCMl-binding site: It appar- the full activity of the protein. ently lowers the DNA-binding affinity while increasing The effect of MCM1 mutant protein overproduction the transcription activation potential of the mutant pro- on gene expression showed that the same mutants that tein. f~-Galactosidase activity remains approximately the have a minichromosome maintenance defect are the same regardless of overproduction of the mutant MCM 1 ones that reduce DSE14-binding site occupancy. If the protein in the remaining strain that showed reduced ac- minichromosome maintenance defect is also due to re- tivity, mcml-AQ, as well as with the fully active alleles, duced DNA binding, then overproduction of the mutant MCM1, mcml-ADE, and mcml-SRF/DE, showing that protein should suppress this defect as well. Mutant in these strains, DSE14 is fully bound by the MCM1 pro- strains containing YCpl01 and either the high-copy vec- tein in single-copy gene dosage. tor YEp24 (2~ ARS, URA3) or YEp24 bearing the mutant These results demonstrate that the polyglutamine do- mcml gene were tested for maintenance of the YCpl01 main contributes to the activity of MCM1, because its plasmid. Cells were grown in synthetic complete (SC) removal reduces MCMl-dependent gene expression two- medium lacking uracil to retain the YEp24 derivative fold, regardless of overproduction of the mutant protein. while allowing loss of the YCpl01 tester plasmid. The In contrast, the acidic stretch is not important for loss rates of YCpl01 (Table 3) show that in all three MCM1 function at the DSE14 site by two criteria: First, Mcm-defective mutants tested, minichromosome main- mcml-ADE has wild-type or very near wild-type activ- tenance is restored to near wild type when the mutant ity; second, overproduction of MCM1-ADEQ or gene is present on the YEp24 ptasmid. These results sug- MCM1-AQ yields the same activity from DSE~4 (Table gest that in each mutant, the Mcm defect is also due to 2); both give approximately half wild-type levels of reduced rather than altered function of the mutant pro- f~-galactosidase. The ability of the MCM1-ADEQ and tein. MCM1-AN~TDEQ proteins to activate transcription equally well (Table 1) localizes the remaining transcrip- The acidic domain is important for activation of tional activation activity to the SRF-homologous do- a-specific genes in a cells main of MCM1. These results suggest that the 80- amino-acid SRF-homologous domain of MCM1 is capa- We anticipated that because many of these mutations ble of transcription activation as well as DNA binding, affect expression from a non-cell-type-specific MCM1-

GENES & DEVELOPMENT 755 Downloaded from genesdev.cshlp.org on October 6, 2021 - Published by Cold Spring Harbor Laboratory Press

Christ and Tye

Table 1. MCMl-dependent transcription activation and a-specific construct poorly (13% wild type) while being a Mcm phenotype of mcml mutants good activator of the same reporter with the DSE14-bind- Expression Minichromosome ing site instead (100% wild type). The mutants contain- from DSE14 maintenance ing the acidic stretch, mcml-AQ and mcml-1, activate (% wild-type loss rate of transcription from the a-specific promoter to the same Allele activity) YCpl01 level relative to wild type as they do the nonspecific promoter with DSE14 (-50% for mcml-AQ and 5% for MCM1 100 0.02 mcml-1). These results suggest that in contrast to ex- m cm 1-A Q 40 0.02 mcm I-ADE 100 0.02 pression from the nonspecific promoter, the acidic mcml-ADEQ 18 (191 0.07 (0.08) stretch of MCM1 is important for expression from an m cm 1-AN17DiEQ 20 0.06 a-specific promoter. rectal-1 6 (16) 0.30 (0.23) m cm I-SRF/DE 144 0.03 mcml-gcn4/DE(Q) 22 0.21 MCM1 DNA-binding domain is sufficient for repression of a-specific genes in a cells Expression from the integrated DSEI4-1acZ tester gene was measured by [3-galactosidase activity assay of two separate cul- MCM1 has been implicated not only in activation of tures from each strain. Assays were done with 1 ml of each a-specific genes but in repression of a-specific genes in a culture, grown in YPD at 30°C to OD6oo 1.0, as described in cells. The cooperative binding of MCM1 and MATa2 at Guarante (1983). Activity varied by 10% or less between dupli- the operators of a-specific genes is believed to be respon- cate cultures. Results shown are from one experiment, normal- sible for their repression in a cells. We tested whether ized to percent wild-type activity. Similar experiments gave the these mcml mutants affect repression. We again studied same pattern of activity. Wild-type cells generally had 12 Miller expression from an integrated STE2-1acZ reporter gene, units of [3-galactosidase. Activities in parentheses are those of now containing 700 nucleotides upstream of STE2 (Fig. cells grown at 23°C. Minichromosome maintenance assays 3). This region contains the MCM1/a2 operator as well were done at 30°C, using YPD as the nonselective growth medium. Loss rates were calculated for two independent transfor- as UASs and is sufficient for a-specific expression (Smith mants from each strain; the average is shown here. Variation in 1986). Expression from this promoter is 80-fold higher in loss rate within a strain was <20%. A similar experiment using a cells than in isogenic a cells (Fig. 6). YCpl as the tester plasmid gave identical loss rates. Loss rates Results of [3-galactosidase assays in the MATa mcml in parentheses are from cells grown at 23°C. mutant strains (Fig. 6) show that most of the mutants are able to repress the a-specific gene. Only mcml-gcn4/ DE(Q) shows substantial derepression, with 28% of wild-type MATa activity, whereas rectal-1 has a slight dependent promoter, they would also affect expression of effect, with approximately threefold higher [3-galactosi- a-specific genes, which are activated by MCM1 in con- dase levels than wild-type a cells, consistent with the junction with MATal. The two types of promoters reduced DNA-binding affinity of the MCMI-1 protein. might be affected differently in the different mutants The derepression observed in the mcml-gcn4/DE(Q) because of involvement of this cofactor in a-specific mutant is most likely due to structural alterations that gene activation. In vitro-binding studies have shown that affect the ability of the hybrid protein to bind to DNA al increases the binding affinity of MCM1 to the "PQ" with the a2 protein, because no portion of MCM1 is sites of a-specific genes (Passmore et al. 1989). Therefore, missing from this mutant that is important for DNA we tested the effect of each mutant on a-specific expres- binding or corepression with a2 (cf. with mcml-ADEQ). sion using the same lacZ reporter gene as before, but The mcml-ADE, mcml-AQ, and mcml-ADEQ mu- with an a-specific upstream sequence, as shown in Fig- tants show no derepression of a-specific genes, suggest- ure 3. This reporter, MFalPQ-lacZ, contains a 60-mer ing that neither the acidic nor the polyglutamine domain oligonucleotide corresponding to the most proximal is required for their repression. However, if an mcml MCM1/al-binding site of the MFal promoter (Inokuchi mutant were unable to activate transcription of STE2- et al. 1987) in place of the DSE14. The reporter gene was lacZ in this genetic background, no conclusion could be again integrated into each mutant strain in single copy at reached concerning the ability of the mutant protein to the LE U2 locus on chromosome III. Expression from this repress transcription of a-specific genes. Therefore, we a-specific reporter gene was 17-fold higher in wild-type a tested expression of STE2-1acZ in isogenic MATa cells cells than in isogenic a cells, demonstrating that it be- derived from two of the mcml mutants. We found no haves as expected in the two cell types. difference in 13-galactosidase expression from this pro- Results of [3-galactosidase assays in the MATa mcml motor in mcml-ADE or mcml-ADEQ a strains com- mutant strains (Fig. 5A) show that the mutants affect pared to wild-type a strains: Each strain expresses -100 a-specific expression differently than they do nonspe- Miller units of f~-galactosidase (Fig. 6). The low level of cific expression. Expression from the two promoters is f~-galactosidase in the a strains is therefore due to core- compared in Figure 5B. All mutants lacking the acidic pression with a2, which must only require the amino- stretch show greatly reduced activity from the a-specific terminal portion of MCM1. promoter compared to DSE14. In particular, mcml-ADE, The dramatic effects of these mcml mutants on ex- which lacks only the acidic stretch, expresses the pression of mating-type-specific genes in a cells should

756 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on October 6, 2021 - Published by Cold Spring Harbor Laboratory Press

Functional domains o| MCMI

Figure 4. Effect of the mcml mutations on the level of the resulting protein. Im- munoblots using anti-B-gal-MCMll_92 are shown. Transformants of strain BJ2168 were grown in SC-uracil media to OD6oo 1.5, and crude extracts were loaded onto a A B 10% gel (A) or a 15% gel (B). (A) Approxi- mately 3 ~g of total protein was loaded in each lane. A 2-min exposure is shown. The 45kD- wild-type MCM1 protein from the chromosomal MCM1 gene runs as the ~40-kD 45 kD band of approximately equal intensity in each lane. (B) Approximately 1.5 ~g of to- 29 kD -- tal protein was loaded in each lane. A 5- min exposure is shown. Lanes 1 and 2 were loaded with the same extracts as in 29kD 71- -! A. With overexposure, bands of the ex- 18kD-- pected size of MCM1 (~40 kD) and 14kD -- MCM1-ADEQ (~ll kD) appear in the YEp24--mcml-ADEQ lane, with the smaller band having lower intensity than the band from the chromosomal MCM1 gene.

result in mating defective phenotypes. We assayed the some of the MCM1 mutations may affect this regula- ability of each of the MATa mcml mutant strains to tion. We tested this hypothesis by assaying expression of mate with wild-type tester strains, as shown in Figure 7. the a-specific reporter gene, MFalPQ-lacZ, in a wild- The mating defects correspond well with the observed defects in mating-type-specific gene expression. Only the mcml-AQ and mcml-SRF/DE mutants are able to Table 2. Effect of MCM1 mutant protein overproduction on mate at all temperatures, consistent with the finding expression from DSE14-1acZ that only these mutants have at least 25 % wild-type ex- Expression pression from the a-specific reporter gene. Only the from DSE14 mcml-gcn4/DE(Q) mutant mates with other a cells {% wild-type Fold rather than with a cells, an a-like faker phenotype sim- Allele activity) difference ilar to that of mat~l mat~2 double mutants (Strathem et al. 1981), which is consistent with the high level of a- MCM1 + YEp24 100 specific gene expression and low level of a-specific gene + YEp24 MCM1 98 1.0 expression in this mutant. Mating is improved and ex- mcml-AQ pression from MFalPQ-lacZ increases 3- to 10-fold with + YEp24 46 overproduction of wild-type or mutant proteins in their + YEp24 mcml-AQ 59 1.3 respective strains. However, mcml-gcn4/DE(Q) be- m cm 1-ADE comes a weak bi-mater, rather than mating only with a + YEp24 71 cells, and expression from STE2-1acZ suggests that core° + YEp24 m cm 1-ADE 80 1.1 pression with a2 is not restored with overproduction of m cm 1-ADE Q the MCM1-GCN4/DE(Q) protein (data not shown). + YEp24 15 + YEp24 mcml-ADEQ 58 3.9 rectal-1 The acidic domain reduces MCM1 activity at an + YEp24 10 a-specific UAS in a ceils + YEp24 rectal-1 84 8.4 m cm 1-SRF/DE The current model for a-specific gene regulation is that + YEp24 82 only positive regulation occurs, by cooperative binding + YEp24 mcml-SRF/DE 102 1.2 of MCM1 and MATal in a cells and that binding of m cm 1-gcn4/DE(Q) MCM1 alone to the upstream elements of these genes is + YEp24 60 too weak to allow their expression in a cells. However, + YEp24 mcml-gcn4/DE(Q) 193 3.2 MCM1 can bind to PQ elements of MFal and STE3 in Duplicate cultures of each strain were grown at 30°C in vitro with high affinity in the absence of al (Tan et al. SC-uracil media to OD6oo 1.5, and f~-galactosidase activity was 1988; Passmore et al. 1989; Ammerer 1990). Therefore, assayed by using 1 ml of each culture as described {Guarante we wondered whether a-specific gene expression is reg- 1983). Activity varied by 20% or less between the duplicates. ulated in a cells, as well as in a cells, in this case to The experiment was repeated yielding similar results to those decrease MCM1 binding or activity at PQ elements. If so, shown.

GENES & DEVELOPMENT 757 Downloaded from genesdev.cshlp.org on October 6, 2021 - Published by Cold Spring Harbor Laboratory Press

Christ and Tye

Table 3. Effect of MCM1 mutant protein overproduction on stretch--MCM1-ADE, MCM1-SRF/DE, and MCM1- minichromosome stability GCN4/DE(Q)--significantly increase expression from Loss rate of the a-specific UAS in a cells when overproduced, as Allele YCpl01 shown in Figure 8. Expression from the PQ site increases five- to eightfold with overproducion of any of these mu- mcml-ADEQ tant proteins, whereas activity only increases twofold or + YEp24 0.25 less when wild-type MCM1 or MCM1-AQ is overpro- + YEp24 mcml-ADEQ 0.04 duced. This result is unlikely to be due to differences in mcml-1 the amount of MCM1, because immunoblots indicate + YEp24 0.38 + YEp24 mcml-1 0.03 that MCM1 wild-type protein is present at a similar, mcrn I-gcn4/DE(Q) perhaps even higher, level than these mutant proteins + YEp24 0.16 when expressed from a YEp24 plasmid in a wild-type + YEp24 rncrn 1-gcn4/DE(Q) 0.05 strain (Fig. 4A). Therefore, the acidic domain of MCM1 appears to be important not only for activation at the Minichromosome maintenance assays were done at 30°C in a-specific UAS in ~ cells but also for maintaining a low SC + leucine media, lacking uracil. Dilutions from initial and final cultures were plated on YPD and replicated to SC-uracil level of expression from this site in a cells. These results and SC/uracil/leucine. Loss rate of YCpl01 was calculated by suggest that MCMl-dependent expression of ~-specific using the percentage of cells containing YEp24 (-+ mcmlJ that genes may be regulated in a cells as well as in ~ cells. also contain YCpl01. Loss rates shown are averages from testing two independent transformants, whose loss rates varied by <20%. Discussion Functional domains of MCM1 type a strain overproducing MCM1 proteins, and asking DNA-binding domain This study revealed that the whether activity is increased more when a mutant rather three domains of MCM 1 identified by similarity to other than a wild-type protein is overproduced. This experi- proteins do provide different functions, as summarized ment was possible because expression of MF~ 1PQ-lacZ, in Figure 9. However, the 80-amino-acid region homol- unlike DSE14-1acZ , is low in wild-type a cells and in- ogous to SRF was found to be sufficient for most MCM1 creases with overproduction of MCM1. We used an a functions, including cell viability, minichromosome strain isogenic to the ot strain used in the previous ex- maintenance, transcription activation from a non-cell- periments, containing the integrated MFalPQ-lacZ type-specific MCMl-dependent promoter, and corepres- tester. We transformed this strain with YEp24 or YEp24 sion with ~2 of a-specific genes. MCM1 is therefore un- bearing either a wild-type or mutant MCM1 gene and like most previously studied eukaryotic transcription assayed for f~-galactosidase expression. factors, which have separate DNA-binding and activa- Surprisingly, only proteins missing the acidic tion domains (for review, see Mitchell and Tjian 1989),

A. B. Allele MFc~I PQ-lacZ expression 200 (% activitu in a MCPll)

MATo MCM I 10 0 mcm I-AQ 5 1 ir rncm ! -,4DE 13

mcm I-.4DEO 4 [] DSEI4 mcm I- I 5 [] MF~IPQ o. • "- mcm I-SRFIDE 25

mcm l-gcn41DE(O) 6

MATa MCM l 6 Figure 5. Transcriptional activation from MFc~IPQ-lacZ and comparison with activation from the non-cell-type-specific DSE14-1acZ reporter gene in MATs mcml mutants. (A) Duplicate cultures of each strain were grown at 30°C in YPD media to OD6oo 1.5, and ~3-galactosidase activity from the integrated MFedPQ-lacZ reporter gene was assayed by using 1 ml of each culture as described. Activity varied by 20% or less between the duplicates and is expressed as a percentage of the activity in wild-type MATs cells. The same experiment was repeated several times yielding similar results. B-Galactosidase activity was usually 6 Miller units from the MFctlPQ promoter in wild-type ct cells. (B) Light crosshatching indicates DSE14-1acZ expression relative to wild type (from Table 1); dark crosshatching indicates MFcdPQ-lacZ expression relative to wild type.

758 GENES& DEVELOPMENT Downloaded from genesdev.cshlp.org on October 6, 2021 - Published by Cold Spring Harbor Laboratory Press

Functional domains of MCM1

120 " whether it is mediated by interaction of each protein MATo~ MATa with the DNA. IOO In light of our finding that the DNA-binding domain of BO MCM1 is sufficient for its function, it is interesting that

~ 6o ARGS0 (ARG RII, which has 70% identity to MCM1 in these 80 amino acids {Passmore et al. 1988), cannot sub- ~ ,40 stitute for MCM1. Even overproduction of ARG RI can- I:: 20 not rescue the lethality of an mcml deletion {R. Elble, unpubl.1. ARG RI has not been demonstrated to bind DNA and has no apparent transcription activation ability (Qiu et al. 19901, suggesting that ARG RI and MCM1 may differ in amino acids that are important for DNA- binding affinity or transcriptional activation. Interest- ingly, the mcml-1 point mutant at Progz has the most severe phenotype of all the viable mutants we examined, Figure 6. Repression of an a-specific reporter gene in mcml suggesting that Pro97 is important for MCM1 function. mutant strains. Duplicate cultures of each strain were grown at SRF also has a proline at the homologous position, 30°C in YPD media to OD6oo 1.S, and B-galactosidase activity whereas ARG RI does not. was assayed using 1 ml of each culture as described. Activity varied by 20% or less between the duplicates. The experiment was repeated twice yielding similar results. Results are shown Acidic domain The acidic domain of MCM1 is impor- as percent activity in wild-type a cells. B-Galactosidase activity tant for regulation of a-specific genes. All mutants lack- was usually 100 Miller units in a cells. ing the acidic domain show low expression of an oL-specific reporter gene in oL cells, even if they have no defect in expressing a non-cell-type-specific gene. When suggesting that transcriptional activation by MCM1 may a mutant protein lacking the acidic domain is overpro- occur by a different mechanism from that of most other duced in a cells, the a-specific reporter gene is expressed transcription factors. Further experiments will be re- at an abnormally high level compared to the level with quired to determine whether transcriptional activation overproduction of MCM1 proteins containing an acidic requires an activity of MCM1 beyond DNA binding, domain. These results suggest that the acidic domain is such as making specific contacts with other proteins or important for both positive and negative regulation of possibly affecting the DNA or chromatin structure. The MCM1 activity at a-specific promoters. Neither substi- MCM1/a2/DNA complex may be similar to the ternary tution of a portion of SRF or of an acidic activation do- complex of SRF, p62TCF, and DNA, which was sug- main of GCN4 could replace the function of the MCM1 gested to require only the DNA-binding domain of SRF acidic stretch in either cell type. (Schr6ter et al. 1990). It will be interesting to determine In a cells, the acidic stretch may be necessary to form whether the cooperativity of DNA binding by MCM1 a ternary complex of MCM1, oL1, and the PQ DNA. Tan and a2 involves specific protein-protein interactions or and Richmond (19901 showed that only the smallest pro-

Figure 7. Mating of MATa mcml mutants. Strains were patched onto YPD plates covered with a lawn of either the MATa met4 or MATa met4 mating-type tester strain. Plates were incubated overnight at 23, 30, or 37°C and replica-plated to SD (media containing no added amino acidsl, so that only the diploids could grow. The SD plates were incubated at 30°C ovemight. The density of growth gives a relative indication of the mating efficiency of the mutant strains. As a reference, quantita- tive mating tests had shown previously that mcml-1 mates at 1% wild-type efficiency at 30°C (Passmore et al. 1988].

GENES & DEVELOPMENT 759 Downloaded from genesdev.cshlp.org on October 6, 2021 - Published by Cold Spring Harbor Laboratory Press

Christ and "rye

50 MCM1-ADE proteins should distinguish these possibil- xt (,4 ities. la.I 20 I..-.. Polyglutamine domain The polyglutamine domain was found to be dispensable for all MCM1 functions.

o However, mutants lacking the polyglutamine domain -, |0 had at most 60% of wild-type MCM1 transcriptional ac- > tivation activity, on both a nonspecific and a-specific promoter. The reduced activity at the nonspecific pro- 0 moter was shown not to result from an effect on DNA- binding site occupancy. The polyglutamine domain ~. ~-- .~,~,- y,-- ~ therefore contributes to the transcription activity of MCM 1. In addition, the polyglutamine domain may contribute to the stability of the MCM1 protein, because MCM1-AQ and MCM1-ADEQ appear less abundant by Western blots. Unlike the polyglutamine domains of Figure 8. Expression of an a-specific gene in an a strain over- Spl, which were shown by Courey and Tjian (1988) to be producing MCM1 mutant proteins. Two transformants of each required for its transcription activation ability in Droso- plasmid into the MATa MCM1 strain with the integrated MFalPQ-lacZ reporter gene were grown at 30°C in SC-uracil phila cells, the MCM1 polyglutamine domain contrib- media to OD6o o 1.5. [3-Galactosidase activity was assayed by utes only weakly to its total activity. Polyglutamine do- using 1 ml of each culture as described (Guarante 1983) and is mains are common among yeast proteins but have not expressed as a percentage of the activity of an isogenic MATer yet been demonstrated to be important for any function. strain containing YEp24, tested simultaneously. Activity varied It is not clear, therefore, whether the MCM1 polyglu- by 20% or less between the two transformants. tamine domain could be acting similarly to that of Sp 1 in transcriptional activation. teolytic fragments of MCM1 that are still able to bind Role of transcription factors in yeast DNA replication DNA are unable to form a ternary complex with ~ 1, sug- initiation gesting that the portion of MCM1 required for oL1 interaction is close to, but not within, the minimal DNA- In yeast, both MCM1 and another transcription factor, binding domain. Their results, combined with the func- ABF1, have been implicated in replication initiation. tional analysis presented here, suggest that the acidic ABF1 was identified as a factor that binds DNA within stretch of MCM1 is likely to be the site of interaction several ARSs {Buchman et al. 1988; Diffley and Stillman with od. We are currently testing whether or not pro- 1988), and ABFI-binding sites were shown to increase the teins lacking the acidic stretch are able to bind cooper- efficiency of ARS function in a minichromosome main- atively with od to the PQ site. tenance assay (Walker et al. 1990). The identification of Although negative regulation at the a-specific UAS in these transcription factors as potential DNA replication a cells has not been reported previously, results of an initiation factors in yeast suggests that chromosomal earlier study support this hypothesis. Jarvis et al. (1989) replication initiation may be similar to that of the eu- found that activity from a reporter gene with the STE3 karyotic viruses. Viral replication initiation is enhanced PQ element in a cells was half that from the same pro- by the direct binding of mammalian transcription factors moter with only the P element. When they overproduced to the replication origin. This enhancement was shown MGM1, activity from the P element increased to 560 recently to require only the DNA-binding domain of the units, whereas activity from the PQ element increased transcription factor (Mermod et al. 1989; Verrijzer et al. only to 150 units. This result suggests that the Q ele- 1990). ment negatively modulates the ability of MCM1 to ei- Although both transcription and replication activity of ther bind or activate transcription at the neighboring P MCM1 are contained in an 80-amino-acid domain, the element in a cells. With these data and our observation phenotypes of the mcml mutants described in this study that removal of the acidic stretch increases MCM1 activity at PQ in a cells, and by analogy with regulation in oL cells, it seems likely that regulation of MCMl-depen- 17 97 120 286 dent transcriptional activation of (x-specific genes in a I I I I cells may also be mediated by protein(s) binding at the Q MCM1 ...... element. An alternative possibility is that the PQ site is \DNA binding4 ~ ti°n activati°n ''''/ normally not bound with protein in a cells and that de- Transcription activation u-specific gene regulation Repression of a-specific genes letion of the acidic stretch could increase MCMl-bind- Minichromosome maintenance ing affinity to PQ in the absence of cofactors, resulting in Cell viability a level of e~-specific gene expression intermediate to that Figure 9. Functional domains of MCM1. A summary of the of wild-type a cells and the od-induced level of oL cells. In activities attributed to the different parts of the protein is vitro-binding studies with the wild-type and shown. The boxed regions correspond to those in Fig. 1.

760 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on October 6, 2021 - Published by Cold Spring Harbor Laboratory Press

Functional domains of MCM1

suggest that the role of MCM1 in replication initiation is + 185 to +200 and +468 to +450 from the ATG. The deduced probably not indirectly mediated through affecting tran- amino acid sequence, altered following Gly~ s, is shown in Fig- scription of another gene. Comparison of the phenotypes ure 1. The mcml-ADE, mcml-hDEQ, mcml-SRF/DE, rncml- of those mutants with a minichromosome maintenance gcn4/DE(Q), and rectal-AN 17DEQ mutations are derived from (Mcm) defect, rectal-l, mcml-gcn4/DE(Q), and the same oligonucleotide mutagenesis experiment. The SphI- mcml-hDEQ, with those that are wild type for Mcm EcoRI fragment with part of MCM1 (Passmore et al. 1988) was (mcml-ADE, mcml-AQ, and mcml-SRF/DE) shows cloned into KS - {Stratagene), which had the SalI site destroyed. that the only phenotype yet found that differentiates Oligonucleotide mutagenesis was performed by using the Bio- these two groups is DSE14 DNA-binding site occupancy. Rad reagents and the mutagenic oligonucleotide 5'-GTCTTAA- The Mcm-defective class showed lower occupancy of CGCCCCTGTCGACTATGCAACAGC-3', which replaces 60 DSE14 , whereas the Mcm-proficient class showed full nucleotides coding for the acid stretch with a SalI-HincII-AccI occupancy of DSE14. All other phenotypes, that is, pro- site. mcml-aDE was derived by cutting with AccI, filling in the tein level, gene expression from DSE14 , coactivation ends, and religating, mcml-ADEQ was derived by cutting with with ~1, and corepression with ~2, varied within these HincII and cloning in an XbaI linker with stop codons in all three open reading frames, 5'-TACTAGTCTAGACTAGTA-3'. groups. Therefore, the Mcm phenotype correlates with mcml-SRF/DE was derived by cutting with AccI, filling in the DNA-binding site occupancy but not with gene expres- ends, and ligating in an 80-nucleotide Hinfl fragment of the SRF sion or binding with cofactors. These results, together gene from pG3.5 (Norman et al. 1988; from R. Treisman). with the discovery of MCMl-binding sites in ARSs (V. mcml-gcn4/DE(Q) was derived by cutting with SalI and KpnI Chang and S. Passmore, unpubl.), are most consistent and inserting a 195-nucleotide SalI-KpnI fragment from the with the direct model for the action of MCM1 in repli- GCN4 derivative YCp88-1exA-gcn4-D 19 {Hope et al. 1988; from cation, that MGM1 affects minichromosome mainte- K. Struhl). The resulting construct was put back into flame by nance and DNA replication by binding to replication or- cutting with SalI, filling in, and religating. Each construct was igins. Clearly, however, they do not rule out the possi- verified to have the expected junction sequence by restriction bilities that MCM1 acts indirectly to affect replication, mapping and double-stranded dideoxy sequencing, using a primer from + 185 to +200 of MCM1. mcml-ANlzDEQ was perhaps in combination with cofactors that interact with derived from mcml-ADEQ by oligonucleotide-mediated mu- it in ways unlike etl or et2, or by affecting expression of tagenesis, using 5'-CACCCAGCAAAAATGAGAAGAAA- a number of genes, or even that the total effect of mcml GATAGAA-3', which loops out amino acids 2-17. Transfor- mutations on minichromosome maintenance may result mants were screened by sequencing with a primer from -47 to from both direct and indirect activities. Further study of -27 of the MCM1 gene. mcml mutants, combined with mutational analysis of All of the mutated fragments were subcloned into the yeast the ARS elements will be important in resolving this shuttle vectors YIp5 and YEp24 containing the entire MCM1 issue. gene on a 3.4-kb XhoI-BamHI fragment, by replacing the 1.2-kb SphI-KpnI wild-type fragment with the mutagenized fragment. Correct constructs were confirmed by restriction mapping, and for mcrnl-1 and mcml-ANlzDEQ by sequencing. YEp351 (Hill Materials and methods et al. 1986) constructs containing a mutant gene were made by Strains using the SphI-BamHI fragment from the YIp5 constructs. mcml-Xho92 and rectal-Barn92 were made by cutting a Escherichia coli strains used were DHSe* (BRL) for routine clon- YEp351 plasmid containing the wild-type MCM1 gene with StuI ing, GM2163 (hsdR2 mcrB1 daml3::Tn9 dcm61 from New En- and ligating in either an XhoI or BamHI 12-nucleotide linker. gland Biolabs, for preparation of DNA that can be cut at the StuI DNA from resulting transformants was sequenced by using the site in MCM1, and CJ236 [dut ung thi relA pCJ105 (Cmr)] for + 185 to + 200 primer and determined to have only one linker oligonucleotide-directed mutagenesis. Yeast strains 8534-8C inserted in each case. {MATe, his4A34, leu2-3,112, ura3-521 and an isogenic MATa Tester plasmids were derived from pCDH or pCDH-DSE~4 strain made by transient transformation with HO and sporula- (Passmore et al. 1989). DSE14-1acZ was made by cutting pCDH- tion of the resultant diploid, were used for constructing and DSE14 with BamHI and BglII and religating to remove the cen- analyzing mcml mutants; 6697/1 (MATa met4) and 6697/3 tromere so that the plasmid could be integrated into a yeast (MATe* met4) were used for mating tests; BJ2168 (MATa leu2 chromosome. The control plasmid with no MCMl-binding site trp ura3-52 prbl-1122 prcl-407 pep4-3), from David Shore, was was made by cutting pCDH with SalI and HindIII, filling in, and used in the Western blot analyses; and C2/501 [MATe* religating. MFulPQ-lacZ was made by cloning MFe*lp60 (Pass- mcml&Xho/Bam his4 1eu2-3,112 ura3-52 trpl-289 YCp501 more et al. 1989) into pCDH at the SalI and HindIII sites. STE2-- (MCM1 ARS1 CEN5 URA3}] was used in the plasmid shuffle lacZ was constructed by replacing the SalI-SacI fragment of experiments. pCDH with the corresponding fragment of pCD14 (Smith 1986). Centromeres were removed from all plasmids to allow integration, as described for DSE14 -lacZ. Construction of mcml mutants and tester plasmids The mcml-AQ and mcml-1 alleles were derived from previ- Yeast methods ously cloned mutations on XhoI-EcoRI fragments containing part of MCM1 in YIp5 (Passmore et al. 1988). mcml-AQ was Plasmid shuffle assays were done as described (Boeke et al. derived from YIp5 B453 {Maine 1984), which had a BamHI 1987}. In each case, strain C2/501 was transformed with a pos- linker that mapped near the acidic stretch of MCM1. We se- itive control plasmid {the vector containing MCM1 ), a negative quenced through the linker insertion site by double-strand control plasmid (the vector alone), and the vector containing the dideoxy sequencing using primers within the MCM1 gene at mutated mcml gene. Ability of the mutant to provide MCM1

GENES & DEVELOPMENT 761 Downloaded from genesdev.cshlp.org on October 6, 2021 - Published by Cold Spring Harbor Laboratory Press

Christ and lye function was analyzed by growth on 5-fluoro-orotic acid (5- A 0.3-kb BamHI-StuI fragment from pETMCMl(1-188) (Pass- FOA), which selects for cells that have lost YCpS01. The exper- more et al. 1989) was inserted into pUR278 (Rfither and Miiller- imental sample and positive and negative controls were Hill 1983). E. coli strain DHSa containing this plasmid was streaked on the same 5-FOA plate to ensure that growth re- induced with IPTG to express the 13-gal-MCM1 fusion protein. flected the ability of the mutant gene to complement the mcml Cells were lysed by boiling in SDS sample buffer (Laemmli deletion, rather than variation in 5-FOA concentration. 1970), and total protein was separated on preparative 6% acry- The MCM1 gene was replaced with each of the mutant al- lamide SDS-PAGE. The [3-gal-MCM1 fusion protein was cut leles, using a two-step method. YIp5 containing the mutated out, electroeluted, and used to immunize rabbits. The rabbits mcml gene was digested with SphI to target plasmid integration were boosted twice, and antiserum was collected 2 weeks after to the MCM1 locus. Transformants of 8534-8C were selected on the final boost. SC-uracil media, and recombinants retaining only one copy of Immunoblots were performed by standard methods. Yeast ex- the MCM1 gene were selected on 5-FOA (Boeke et al. 1984). tracts were made by pelleting log phase yeast (OD6oo 1.5) and These isolates were tested for retention of the wild-type or mu- resuspending in SDS sample buffer with glass beads, followed by tant allele by genomic Southern blots, except for mcml-1. Yeast repeated boiling and vortexing. Protein concentration in the ex- genomic DNA was digested with appropriate restriction en- tracts was estimated by using the Bio-Rad assay on diluted ex- zymes to differentiate mutant from wild type [mcml-AQ has a tracts, with protein standards to which we added similar BamHI site, mcml-ADE has an NruI site, mcml-hDEQ and amounts of sample buffer. Proteins were separated by SDS- mcml-ANzDEQ have an XbaI site, mcmI-SRF/DE has a SalI PAGE on 10% or 15% acrylamide gels and transferred to nitro- site, and mcml-gcn4/DE(Q) has a PvuI site at the mutation). cellulose by using a semidry blotter (Hoefer) and the three- DNA was separated on agarose gels, blotted to nylon mem- buffer system (Kyhse-Anderson 1984), adding SDS to 0.1% to all branes, and probed with random-prime labeled MCM1 fragment buffers. Nonfat dry milk (5%) was used to block the mem- -14 to +274. Replacement of wild type with mcml-1 was branes, and primary antibody was diluted 1 : 1000 and detected screened by the sterile phenotype and confirmed by the mini- by using horseradish peroxidase-conjugated goat anti-rabbit IgG chromosome maintenance defect, which were both identical to (BRL, 1 : 6000). Visualization was by enhanced chemilumines- those of the previously described isolate of mcml-1 (Passmore cence (Amersham). et al. 1988). Tester plasmids DSE14-1acZ, MFet 1PQ-lacZ, and STE2-1acZ were integrated into each mutant strain at LEU2 by cutting Acknowledgments with BstEII. Transformants were tested for having a single copy We thank Kevin Struhl for the gift of the GCN4 derivative of the tester plasmid by Southern blot. Yeast genomic DNA was YCp88-1exA--gcn4--D 19, which we used to make m cm 1-gcn4/ cut with PstI, which cuts once in the vector and again 5 kb DE(Q), and Tom Fox, for critical reading of the manuscript. upstream of LE U2. DNA was separated on agarose gels, blotted This work was supported by a National Science Foundation to nylon membranes, and probed with a random-prime-labeled Graduate Fellowship to C.C. and by National Institutes of fragment of lacZ. Transformants with only the expected 13-kb Health grant GM 34190. band, lacking a plasmid-sized fragment, were chosen for The publication costs of this article were defrayed in part by B-galactosidase assays. payment of page charges. This article must therefore be hereby 13-Galactosidase activity was measured according to Guarante marked "advertisement" in accordance with 18 USC section {1983). Units were calculated as (1000 x OD42o)/[time 1734 solely to indicate this fact. (min) x vol (ml) x OD6oo) ] (Miller 1972). Cultures were grown to the same OD6o o (ranging from 1-1.5 in different sets of assays), and activity was normalized to the activity of the wild- References type strain, measured simultaneously. Cultures were grown in either YPD or SC media, as stated in the table footnotes. Ammerer, G. 1990. Identification, purification, and cloning of a Minichromosome maintenance assays were done as follows: polypeptide (PRTF/GRM) that binds to mating-specific pro- A colony grown on selective media was suspended in 0.2 ml of moter elements in yeast. Genes & Dev. 4: 299-312. water. A 0.1-ml aliquot was used to inoculate 5 ml of nonselec- Boeke, J.D., F. LaCroute, and G.R. Fink. 1984. A positive selective media (either YPD or SC --+ uracil), and cultures were grown tion for mutants lacking orotidine-5'-phosphate decarboxyl- with aeration until saturated (-10 generations). Dilutions of the ase in yeast: 5-Fluoro-orotic acid resistance. Mol. Gen. initial suspension were plated on YPD plates, and colonies were Genet. 197: 345-346. counted to determine the initial concentration of viable cells; Boeke J.D., J. Trueheart, G. Natsoulis, and G.R. Fink. 1987. these plates were then replica-plated to SC-leucine to deter- 5-Fluoroorotic acid as a selective agent in yeast molecular mine the initial percentage of plasmid-bearing cells. Final sam- genetics. Methods Enzymol. 154: 164--175. ples were treated similarly. The number of generations of non- Brewer, B.J. and W.L. Fangman. 1987. The localization of repli- selective growth (n) was calculated as log(final concentration/ cation origins on ARS plasmids in S. cerevisiae. Cell initial concentration)/log2. The loss rate per cell per generation 51: 463-471. was calculated as 1- (final % plasmid-bearing cells/initial % Buchman, A.R., W.J. Kimmerly, J. Rine, and R. D. Komberg. plasmid-bearing cells) lm. Each assay set was done simulta- 1988. Two DNA-binding factors recognize specific se- neously to eliminate variations due to temperature or media. quences at silencers, upstream activating sequences, auton- Loss rates were found to be twofold higher in synthetic com- omously replicating sequences, and telomeres in Saccharo- pared to YPD media for all strains except for mcml-1 and myces cerevisiae. Mol. Cell. Biol. 8: 210--225. rncm 1-gcn4/DE(Q). Challberg, M.D. and T.J. Kelly. 1989. Animal virus DNA replication. Annu. Rev. Biochem. 58: 671-717. Courey, A.J. and R. Tjian. 1988. Analysis of Spl in vivo reveals Antiserum preparation and immunological techniques multiple transcriptional domains, including a novel glu- Antiserum was generated against a [3-gal-MCM1 fusion protein tamine-rich activation motif. Cell 55: 887-898. containing only the amino-terminal 92 amino acids of MCM1. Courey, A.J., D.A. Holtzman, S.P. Jackson, and R. Tjian. I989.

762 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on October 6, 2021 - Published by Cold Spring Harbor Laboratory Press

Functional domains of MCM1

Synergistic activation by the glutamine-rich domains of hu- Mermod, N., E.A. O'Neill, T.J. Kelly, and R. Tjian. 1989. The man transcription factor Spl. Cell 59: 827-836. proline-rich transcriptional activator of CTF/NF-I is distinct Diffley J.F.X. and B. Stillman. 1988. Purification of a yeast pro- from the replication and DNA binding domain. Cell tein that binds to origins of DNA replication and a transcrip- 58:741-753 tional silencer. Proc. Natl. Acad. Sci. 85: 2120-2124. Miller, J.H. 1972. Experiments in molecular genetics. Cold Dubois, E., J. Bercy, and F. Messenguy. 1987. Characterization Spring Harbor Laboratory, Cold Spring Harbor, New York. of two genes, ARGRI and ARGRIlI required for specific reg- Mitchell, P.J., and R. Tjian. 1989. Transcriptional regulation in ulation of arginine metabolism in yeast. Mol. Gen. Genet. mammalian cells by sequence-specific DNA binding pro- 207: 142-148. teins. Science 245: 371-378. Guarante, L. 1983. Yeast promoters and lacZ fusions designed Norman, C., M. Runswick, R. Pollock, and R. Treisman. 1988. to study expression of cloned genes in yeast. Methods Enzy- Isolation and properties of cDNA clones encoding SRF, a mol. 101: 181-187. transcription factor that binds to the c-los serum response Hill J.E., A.M. Myers, T.J. Koemer, and A. Tzagoloff. 1986. element. Cell 55: 989-1003. Yeast/E. coli shuttle vectors with multiple unique restric- Passmore, S., G.T. Maine, R. Elble, C. Christ, and B.K. Tye. tion sites. Yeast 2: 163-167. 1988. Saccharomyces cerevisiae protein involved in plasmid Hope, I.A. and K. Struhl. 1986. Functional dissection of a eu- maintenance is necessary for mating of MATot cells. ]. Mol. karyotic transcriptional activator protein, GCN4 of yeast. Biol. 204: 593-606. Cell 46: 885-894. Passmore, S., R. Elble, and B.K. Tye. 1989. A protein involved in Hope, I.A., S. Mahadevan, and K. Struhl. 1988. Structural and minichromosome maintenance in yeast binds a transcrip- functional characterization of the short acidic transcriptional enhancer conserved in eukaryotes. Genes & Dev. tional activation region of yeast GCN4 protein. Nature 3: 921-935. 333: 635-640. Qiu, H., E. Dubois, P. Bro6n, and F. Messenguy. 1990. Func- Huberman, J.A., L.D. Spotila, K.A. Nawotka, S. M. E1-Assouli, tional analysis of ARGRI and ARGRIII regulatory proteins and L.R. Davis. 1987. The in vivo replication origin of the involved in the regulation of arginine metabolism in Saccha- yeast 2~xm plasmid. Cell 51: 473-481. romyces cerevisiae. Mol. Cen. Cenet. 222: 192-200. Huberman, J.A., J. Zhu, L.R. Davis, and C.S. Newlon. 1988. Rfither, U. and B. Mfiller-Hill. 1983. Easy identification of Close association of a DNA replication origin and an ARS cDNA clones. EMBO ]. 2:1790-1794. element on chromosome III of the yeast Saccharomyces cer- Schr6ter, H., C.G.F. Mueller, K. Meese, and A. Nordheim. 1990. evisiae. Nucleic Acids Res. 16: 6373-6384. Synergism in ternary complex formation between the Inokuchi, K., A. Nakayama, and F. Hishinuma. 1987. Identifi- dimeric glycoprotein p67 sRr, polypeptide p62 TcF and the c- cation of sequence elements that confer cell-type-specific fos serum response element. EMBO L 9:1123-1130. control of MFal gene expression in Saccharomyces cerevi- Shore, D. and K. Nasmyth. 1987. Purification and cloning of a siae. Mol. Cell. Biol. 7: 3185-3193. DNA binding protein from yeast that binds to both silencer Jarvis, E.E., D.C. Hagen, and G.F. Sprague, Jr. 1988. Identifica- and activator elements. Cell 51: 721-732. tion of a DNA segment that is necessary and sufficient for Smith, P.P. 1986. The control of the a specific gene for the alpha a-specific gene control in Saccharomyces cerevisiae: Impli- factor receptor in Saccharomyces cerevisiae: Studies on a cations for regulation of a-specific and a-specific genes. Mol. lacZ fusion. Ph.D. thesis. Cornell University, Ithaca, New Cell. Biol. 8: 309-320. York. Jarvis, E.E., K.L. Clark, and G.F. Sprague Jr. 1989. The yeast Sommer, H., J.-P. Beltran, P. Huijser, H. Pape, W.-E. L6nnig, H. transcription activator PRTF, a homolog of the mammalian Saedler, and Z. Schwarz-Sommer. 1990. Deficiens, a ho- serum response factor, is encoded by the MCM1 gene. Genes meotic gene involved in the control of flower morphogenesis & Dev. 3: 936-945. in Antirrhinum majus: The protein shows homology to tran- Keegan, L., G. Gill, and M. Ptashne. 1986. Separation of DNA scription factors. EMBO ]. 9: 605--613. binding from the transcription-activating function of a eu- Strathem, J., J. Hicks, and I. Herskowitz. 1981. Control of cell karyotic regulatory protein. Science 231: 699-704. type in yeast by the mating type locus: the al-a2 hypothesis. Keleher, C.A., C. Goutte, and A.D. Johnson. 1988. The yeast ]. Mol. Biol. 147: 357-372. cell-type-specific repressor a2 acts cooperatively with a non- Tan, S. and T.I. Richmond. 1990. DNA binding-induced confor- cell-type-specific protein. Cell 53: 927-936. mational change of the yeast transcriptional activator PRTF. Keleher, C.A., S. Passmore, and A.D. Johnson. 1989. Yeast re- Cell 62: 367-377. pressor a2 binds to its operator cooperatively with yeast pro- Tan, S., G. Ammerer, and T.J. Richmond. 1988. Interactions of tein Mcml. Mol. Cell. Biol. 9: 5228-5230. purified transcription factors: binding of yeast MATal and Kyhse-Anderson, J. 1984. Electroblotting of multiple gels: A PRTF to cell-type-specific, upstream activating sequences. simple apparatus without buffer tank for rapid transfer of EMBO ]. 7: 4255-4264. proteins from polyacrylamide gels to nitrocellulose mem- Verrijzer, C.P., A.J. Kal, and P. C. Van der Vliet. 1990. The DNA branes. ]. Biophys. Biochem. Methodol. 10: 203-209. binding domain {POU domain) of transcription factor oct-1 Laemmli, U.K. 1970. Cleavage of structural proteins during the suffices for stimulation of DNA replication. EMBO ]. assembly of the head of bacteriophage T4. Nature 277: 680- 9: 1883-1888. 685. Walker, S.S., S.C. Francesconi, and S. Eisenberg. 1990. A DNA Levine, M. and J.L. Manley. 1989. Transcriptional repression of replication enhancer in Saccharomyces cerevisiae. Proc. eukaryotic promoters. Cell 59: 405-408. Natl. Acad. Sci. 87: 4665-4669. Maine, G.T. 1984. Mutants of Saccharomyces cerevisiae defec- Yanofsky, M.F., H. Ma, J.L. Bowman, G.N. Drews, K.A. Feld- tive in the maintenance of minichromosomes. Ph.D. Thesis. mann, and E.M. Meyerowitz. 1990. The protein encoded by Cornell University, Ithaca, New York. the Arabidopsis homeotic gene agamous resembles tran- Maine, G.T., P. Sinha, and B.K. Tye. 1984. Mutants of S. cere- scription factors. Nature 346: 35-39. visiae defective in the maintenance of minichromosomes. Genetics 106: 365-385.

GENES & DEVELOPMENT 763 Downloaded from genesdev.cshlp.org on October 6, 2021 - Published by Cold Spring Harbor Laboratory Press

Functional domains of the yeast transcription/replication factor MCM1.

C Christ and B K Tye

Genes Dev. 1991, 5: Access the most recent version at doi:10.1101/gad.5.5.751

References This article cites 42 articles, 11 of which can be accessed free at: http://genesdev.cshlp.org/content/5/5/751.full.html#ref-list-1

License

Email Alerting Receive free email alerts when new articles cite this article - sign up in the box at the top Service right corner of the article or click here.