Most genes involved in DNA replication in the yeast Saccharomyces cerevisiae are transcribed transiently during late Gi as cells undergo START. Their promoters all contain one or more versions of an 8-base pair motif (ACGCGTNA) called the Mlul cell cycle box (MCB). MCBs have been shown to be both necessary and sufficient for the late Gi-specific transcription of the TMP1 thymidylate synthase and POLI DNA polymerase genes. A different late Gi-specific transcription element called the SCB (CACGAAAA) is bound by a factor containing the SWI4 and SWI6 proteins. We describe here the formation in vitro of complexes on TMP1 MCBs that contain the SWI6 protein and, we suggest, a 120 kDa protein that is distinct from SWI4. Transcription due to SCBs and MCBs occurs in the absence of SWI6 but it is no longer correctly cell cycle regulated. We suggest that SWI6 is an essential regulatory subunit of two different START-dependent transcription factors. One factor (SBF) contains SWI4 and binds to SCBs whereas the other (MBF) contains p120 and binds MCBs.

In many eukaryotic organisms, proliferation is regulated by controls on cells’ progression through Gi phase. In the budding yeast Saccharomyces cerevisiae, commitment to the mitotic cell cycle is made at a stage in late Gi called START (Pringle and Hartwell, 1981). START only takes place if cells have reached a certain size and it is prevented by pheromones that promote conjugation. Gene products required for START include a protein kinase encoded by CDC28 (Reid and Hartwell, 1977; Lorincz and Reed, 1984) and three Gi-specific cyclin-like proteins encoded by CLN1, CLN2 and CLN3 (Nash et al., 1988; Richardson et al., 1989). There are indications that an increase in the kinase activity of a complex containing CDC28 and CLN proteins may trigger passage through START (Wittenberg et al., 1990).

Transcriptional controls have an important role at START. CLN1 and CLN2 transcripts are absent in early Gi cells but appear at START, suggesting that CDC28 kinase activity in Gi may be regulated at least in part by the transcriptional activation of CLNI and CLN2 (Wittenberg et al., 1990). The full activation of CLN! and CLN2 requires the activity of Gi cyclins and that of the CDC28 protein kinase, implying that it occurs via a positive feedback loop (Cross and Tinkelenberg, 1991; Dirick and Nasmyth, 1991). Transcription of many genes for DNA replication enzymes, such as thymidylate synthase (TMP1) and ribonucleotide reductase (RNR1), also occurs at START and requires the CDC28 protein kinase (reviewed by Andrews and Herskowitz, 1990).

Two different types of START-dependent cz.s-acting regulatory elements have been identified: one initially identified within the promoter of the cell cycle-regulated HO endonuclease (Nasmyth, 1983) whose sequences resemble CACGAAAA (Nasmyth, 1985), and a second found in almost all the cell cycle-regulated genes involved in DNA replication (Pizzagalli et al., 1988) whose consensus ACGCGTNA contains an Mlul restriction site. We refer to these as SCB (swi cell cycle box) and MCB (Mlul cell cycle box) respectively. Both types of element can alone confer START-dependent transcriptional activation on heterologous reporter genes (Breeden and Nasmyth, 1987a; McIntosh et al., 1991; Gordon and Campbell, 1991). A complex containing the SWI4 and SWI6 proteins binds specifically to SCBs found in the HO and CLN2 promoters (Andrews and Herskowitz, 1989; Taba et al., 1991; Nasmyth and Dirick, 1991; Ogas et al., 1991) but it is unclear what factor is responsible for the biological activity of MCBs. Lowndes et al. (1991) have reported the binding of a high molecular weight complex to tandem repeats of ACGCGT (but the same sequence did not give rise to START-dependent transcription in vivo), whereas Verma et al. (1991) have purified a 17 kDa protein that binds to MCBs.

Since both SCBs and MCBs are subject to similar cell cycle control and have some sequence resemblance, we have investigated whether either SWI4 or SWI6 are involved in MCB function. We describe the partial characterization and purification of an activity (named MBF) containing SWI6 and a 120 kDa protein, distinct from SWI4, that binds to both MCBs sites within the TMP1 promoter. MBF must have an important role in the biological activity of MCBs in vivo since we find that the transcription of several genes regulated by MCBs is no longer cell cycle regulated in swi6 mutants. Thus, SWI6 is a common component of distinct complexes that exert START-dependent transcription from MCBs and SCBs.

A factor containing SWI6 binds to MCBs

In the course of studying the role of SWI4 and SWI6 in the activation of Gi cyclins, we noticed that swi6 mutations caused subtle but reproducible changes in the pattern of RNR1 transcription (see below). Since RNR1 belongs to the group of genes whose promoter contains Mini sites (MCBs) and is transcribed at START (Elledge and Davis, 1990; Price et al., 1991), we investigated whether SWI4 or SWI6 proteins can bind to MCBs.

The best characterized MCBs are the two required for efficient TMP1 transcription (McIntosh et al., 1991). We therefore chose as a probe for DNA binding a natural oligonucleotide from the TMP1 promoter (MCB-TMP1) that contains both these MCBs. Upon mixing crude extracts with MCB-TMP1, we were able to detect the formation of a very slow migrating gel retardation complex that is competed by a fifty-fold molar excess of unlabelled oligonucleotide (Mlu 2x) or by oligonucleotides containing multiple SCBs (CACGA4 and sRS2) but not by an unrelated one (Fig. 1 A, lanes 7-12). To test whether the formation of this complex requires SWI4 or SWI6, we analysed whether it can also be formed using extracts from swi4 or swi6 mutants. We find that the complex can be formed by swi4 but not by swi6 mutant extracts (Fig. 1A, lanes 2 and 3). Treatment with SWI6-specific antibodies led to the disappearance of the complex and caused the appearance of a yet slower migrating complex that was barely able to enter the gel (Fig. 1A, lane 5). Neither non-specific nor SWI4-specific antibodies had any effect (Fig. 1A, lanes 4 and 6). We conclude that the SWI6 protein is part of an MCB-binding factor (MBF) but that SWI4 is not.

Fig. 1.

SWI6 is part of a complex (MBF) that can bind to MCBs at all stages of the cell cycle. (A) Complexes formed on MCB-TMP1 (see Fig. 3) in crude extracts analysed by gel electrophoresis. Whole-cell extracts were prepared from wild-type (K1268), swi4 mutant (R/H1071) and swi6 mutant (K1354) strains (Taba et al., 1991). MBF points to the slow migrating complex formed that is present in swi4 but not in swi6 mutants (lanes 1-3). Deletion of SWI4 is lethal for K1268 but such deletions can be obtained if CLN2 is expressed from the spADH promoter. This strain (K2527) also contains wild-type levels of MBF (data not shown). Lanes 4-6 show the effect of adding antibodies (1:20 dilution of sera) directed against DHFR (non specific), SWI6 and SWI4 proteins to extracts from the WT. Lanes 7-12 show competition with cold oligonucleotides (for the sequence of CACGA4 and sRS2, see Taba et al., 1991). Note that oligonucleotides containing several SCBs as well as MCBs can compete. (B) The amount of MBF in a factor-synchronized cells. The graph shows the ratio of MBF to nonspecific (N) complexes in cycling, a factor-arrested, and released cells. In this culture, the START-dependent HO gene is fully activated by 70 minutes, buds emerge (BE) by 80 minutes, and the frequency of anaphase cells peaks at 150 minutes (A) (Taba et al., 1991).

Fig. 1.

SWI6 is part of a complex (MBF) that can bind to MCBs at all stages of the cell cycle. (A) Complexes formed on MCB-TMP1 (see Fig. 3) in crude extracts analysed by gel electrophoresis. Whole-cell extracts were prepared from wild-type (K1268), swi4 mutant (R/H1071) and swi6 mutant (K1354) strains (Taba et al., 1991). MBF points to the slow migrating complex formed that is present in swi4 but not in swi6 mutants (lanes 1-3). Deletion of SWI4 is lethal for K1268 but such deletions can be obtained if CLN2 is expressed from the spADH promoter. This strain (K2527) also contains wild-type levels of MBF (data not shown). Lanes 4-6 show the effect of adding antibodies (1:20 dilution of sera) directed against DHFR (non specific), SWI6 and SWI4 proteins to extracts from the WT. Lanes 7-12 show competition with cold oligonucleotides (for the sequence of CACGA4 and sRS2, see Taba et al., 1991). Note that oligonucleotides containing several SCBs as well as MCBs can compete. (B) The amount of MBF in a factor-synchronized cells. The graph shows the ratio of MBF to nonspecific (N) complexes in cycling, a factor-arrested, and released cells. In this culture, the START-dependent HO gene is fully activated by 70 minutes, buds emerge (BE) by 80 minutes, and the frequency of anaphase cells peaks at 150 minutes (A) (Taba et al., 1991).

We have measured the level of MBF activity in a synchronous culture produced by release from a factor-induced Gi arrest. In this culture (Taba et al., 1991), START-dependent transcription peaks at 75 minutes, bud formation occurs at 80 minutes, and the frequency of anaphase cells peaks at 150 minutes. There do not appear to be any major changes in the level of MBF during this interval (Fig. IB).

We also find the same level of MBF in cycling cells compared to a population of unbudded Gi cells purified by elutriation (data not shown). These data suggest that the level of MBF does not fluctuate noticeably during the cell cycle.

We have partially purified MBF by fractionation on a heparin-agarose column (Ammerer, 1990). An activity that gives rise to a very similar gel retardation complex and contains SWI6 but not SWI4 elutes at 0.3 M ammonium sulphate (Fig. 2). This partially purified form of MBF protects specifically both MCB sites in the TMP1 promoter from cleavage by copper phrenanthroline (Fig. 3A). We have analysed which G residues when methylated by DMS (Fig. 3B) and which A or G residues when carbethoxylated by diethyl pyrocarbonate (Fig. 3C), interfere with MBF binding. Modification of bases in both MCBs interferes with complex formation, suggesting that MBF binding is cooperative. Consistent with this, oligonucleotides containing only the 5’ MCB cannot compete for binding (Fig. 1 A, lane 10) and point mutations in either of the MCBs in MCB-TMP1 abolish complex formation without giving rise to faster migrating complexes in which only a single MCB would be occupied (Fig. 4). The two sites bound are at least two turns of the double helix apart, but this separation is not obligatory since MBF can also bind to MCBs that are separated by only a single turn of the helix (data not shown). The pattern of interference at the two MCBs is similar if we assume that both sites are asymmetric and have opposing orientations (see Fig. 3D). MBF binding requires two A-T base pairs that are outside the Mlul restriction site (positions 3, 4, 8 and 9 in Fig. 3D), as is also found for MCB function in vivo (McIntosh et al., 1991). MBF makes contacts along an extensive stretch of the major groove of each MCB and must therefore bind to both sides of the double helix at each site.

Fig. 2.

MBF partially purified by heparin agarose chromatography contains SWI6 but not SWI4. Whole-cell extracts (from a MATa, prbl, prcl, pep4 strain; KI 142) were prepared and fractionated over a heparin-agarose column (Ammerer, 1990). The fractions were then analysed for the formation of gel retardation complexes with the MCB-TMP1 oligonucleotide, as described in Fig. 1. MBF activity elutes from the column at 0.3 M (NFUhStT) (data not shown). Addition of CC-SWI4 (lane 2) and 0C-SWI6 antibodies (lane 3) to the binding assays confirmed that this complex contains SWI6 but not SWI4.

Fig. 2.

MBF partially purified by heparin agarose chromatography contains SWI6 but not SWI4. Whole-cell extracts (from a MATa, prbl, prcl, pep4 strain; KI 142) were prepared and fractionated over a heparin-agarose column (Ammerer, 1990). The fractions were then analysed for the formation of gel retardation complexes with the MCB-TMP1 oligonucleotide, as described in Fig. 1. MBF activity elutes from the column at 0.3 M (NFUhStT) (data not shown). Addition of CC-SWI4 (lane 2) and 0C-SWI6 antibodies (lane 3) to the binding assays confirmed that this complex contains SWI6 but not SWI4.

Fig. 3.

Footprint and interference analysis of MBF binding to the TMP1 promoter. (A) Protection from phrenanthroline copper cleavage (Kuwabara and Sigman, 1987); (B) methylation interference (Sturm et al., 1987); (C) carbethoxylation interference (Sturm et al., 1987); (D) summary of interference data and the sequence of MCB-TMP1. Upper and lower refer to the strand analysed and F and B refer to free and bound probe respectively. Base pairs conserved in MCBs are in bold. Note that the two MCBs can be interpreted as having opposite orientations. Greek letters refer to bases whose methylation interferes with binding and numbers refer to bases whose carbethoxylation interferes with binding.

Fig. 3.

Footprint and interference analysis of MBF binding to the TMP1 promoter. (A) Protection from phrenanthroline copper cleavage (Kuwabara and Sigman, 1987); (B) methylation interference (Sturm et al., 1987); (C) carbethoxylation interference (Sturm et al., 1987); (D) summary of interference data and the sequence of MCB-TMP1. Upper and lower refer to the strand analysed and F and B refer to free and bound probe respectively. Base pairs conserved in MCBs are in bold. Note that the two MCBs can be interpreted as having opposite orientations. Greek letters refer to bases whose methylation interferes with binding and numbers refer to bases whose carbethoxylation interferes with binding.

Fig. 4.

MBF binding requires both MCBs. Partially purified MBF was mixed with wild-type and mutant oligonucleotide probes in the presence or absence of anti-SWI6 antibodies. The 5 ′ and 3 ′ MCBs were mutated from ACGCGTTA to ACTCATTA and from ACGCGTTT to ACTCATTT respectively (top strand sequences). Note that both mutations eliminate complexes marked MBF without causing the appearence of faster migrating SWI6-dependent complexes due to the binding of MBF to the remaining single site. Note also that the mutations appear to create a binding site for a new factor.

Fig. 4.

MBF binding requires both MCBs. Partially purified MBF was mixed with wild-type and mutant oligonucleotide probes in the presence or absence of anti-SWI6 antibodies. The 5 ′ and 3 ′ MCBs were mutated from ACGCGTTA to ACTCATTA and from ACGCGTTT to ACTCATTT respectively (top strand sequences). Note that both mutations eliminate complexes marked MBF without causing the appearence of faster migrating SWI6-dependent complexes due to the binding of MBF to the remaining single site. Note also that the mutations appear to create a binding site for a new factor.

MBF also contains a 120 kDa protein

To determine whether MBF contains proteins other than SWI6, we have analysed which proteins can be cross-linked to BrdU-substituted DNA following UV irradiation. BrdU-substituted 32P-labelled MCB-TMP1 was mixed with partially purified MBF and complexes separated from unbound DNA by gel retardation electrophoresis. The gel was then irradiated, the protein-DNA complexes were digested with DNase 1 and eluted from the gel, and the 32P-labelled proteins analysed by SDS-PAGE gel electrophoresis. A single protein of approximately 120 kDa (pl20) was detected (Fig. 5A1, lane 2), whose recovery is greatly reduced by adding SWI6-specific antibodies to the binding reaction (Fig. 5A1, lane 1). pl20 is the wrong size for SWI6, which normally migrates as a 90-95 kDa protein (Taba et al., 1991). To confirm whether or not the cross-linked protein is SWI6, the gel shown in Fig. 5Al was rehydrated and SWI6 detected by western blotting. SWI6 from a crude extract (Fig. 5A2, lane 3) and from the cross-linked DNase I-treated complexes (Fig. 5A2, lane 2; though barely visible after reproduction) runs alongside the 92.5 kDa marker, well in front of pl20. SWI4 is predicted to be 123 kDa but it cannot be pl20 because it is not part of MBF (Figs 1A, 2) and because it was shown by western blotting to migrate as a 150 kDa protein in the same gel (data not shown), pl20 is also cross-linked to MCBs by UV irradiation when partially purified MBF is mixed with BrdU-substituted MCB-TMP1 in solution in the absence, but not in the presence, of specific competitor (Fig. 5B). In addition, pl20 can be co-immunoprecipitated from the reaction with SWI6-but not with SWI4-specific antibodies (Fig. 5C). We therefore suggest that MBF is composed of at least two proteins: SW16 and a 120 kDa protein that is not SWI4. The detection by silver staining of similar amounts of SWI6 and a 120 kDa protein in highly purified preparations of MBF (data not shown) corroborates the above assertion. We were not able to cross-link SWI6 to the TMP1 oligonucleotide, suggesting that SWI6 might not be in intimate contact with the DNA.

Fig. 5.

MBF contains a 120 kDa protein in addition to SWI6. (A) Binding reactions using partially purified MBF were performed in the presence and absence of a-SWI6 antibodies, except that a bromodeoxyuridine (BrdU)-substituted MCB-TMPl oligonucleotide was used (Barberis et al., 1990). After gel electrophoresis, the protein-DNA complexes were covalently cross-linked by UV irradiation, DNase I-digested, eluted from the gel and covalently bound protein analysed by SDS-PAGE and autoradiography. A single band is detected at approximately 120 kDa (lane 2) that is strongly reduced if (X-SWI6 antibodies are present during the binding (lane I). Yeast whole cell extracts (lane 3) were separated on the same SDS-PAGE gel. An identical UV cross-linking result was obtained using a crude extract (data not shown). The very same gel was rehydrated after autoradiography, proteins were transferred to nitrocellulose and SW16 was assayed using CX-SWI6 antibodies and alkaline phosphatase conjugated to anti-rabbit antibodies. SW16 is detected not only in the whole cell extract (lane 3’) but also in the cross-linked sample (lane 2 ′; very faint but visible in original prints). In similar gels, SWI6 from the heparin fraction was shown to co-migrate with that from crude extracts and SWI4 from both the heparin fraction and from crude extracts shown to migrate as a 150 kDa protein (data not shown). The positions of the molecular weight markers (kDa) run in lanes 4 and 4 ′ are indicated. (B) Binding reaction mixes were directly UV-irradiated, DNase I-digested and analysed by SDS-PAGE and autoradiography. Binding reactions were carried out both in the absence of competitor (lane 2) and in the presence of a 50-fold molar excess of cold specific competitor (MCB-TMPI; lane 3). Addition of a 50-fold molar excess of a cold point-mutated MCB-TMPl (lane 4; 5 ′ and 3 ′ MCBs mutated, see Fig. 4) does not lead to a significant reduction in the amount of p 120 detected. The positions of the molecular weight markers (kDa) run in lane I are indicated. (C) As in B, except that cross-linked complexes were immunoprecipitated with SWI6-specific (lanes 2-4), SWI4-specific (lane 5) or DHFR-specific (lane 6) antibodies after digestion with DNase I. No competitor (lanes 2, 5 and 6), 50-fold molar excess unlabelled MCB-TMPl (lane 3) and 50-fold molar excess of unlabelled MCB-TMPl, both of whose MCBs are mutated (lane 4), were added during the binding reactions.

Fig. 5.

MBF contains a 120 kDa protein in addition to SWI6. (A) Binding reactions using partially purified MBF were performed in the presence and absence of a-SWI6 antibodies, except that a bromodeoxyuridine (BrdU)-substituted MCB-TMPl oligonucleotide was used (Barberis et al., 1990). After gel electrophoresis, the protein-DNA complexes were covalently cross-linked by UV irradiation, DNase I-digested, eluted from the gel and covalently bound protein analysed by SDS-PAGE and autoradiography. A single band is detected at approximately 120 kDa (lane 2) that is strongly reduced if (X-SWI6 antibodies are present during the binding (lane I). Yeast whole cell extracts (lane 3) were separated on the same SDS-PAGE gel. An identical UV cross-linking result was obtained using a crude extract (data not shown). The very same gel was rehydrated after autoradiography, proteins were transferred to nitrocellulose and SW16 was assayed using CX-SWI6 antibodies and alkaline phosphatase conjugated to anti-rabbit antibodies. SW16 is detected not only in the whole cell extract (lane 3’) but also in the cross-linked sample (lane 2 ′; very faint but visible in original prints). In similar gels, SWI6 from the heparin fraction was shown to co-migrate with that from crude extracts and SWI4 from both the heparin fraction and from crude extracts shown to migrate as a 150 kDa protein (data not shown). The positions of the molecular weight markers (kDa) run in lanes 4 and 4 ′ are indicated. (B) Binding reaction mixes were directly UV-irradiated, DNase I-digested and analysed by SDS-PAGE and autoradiography. Binding reactions were carried out both in the absence of competitor (lane 2) and in the presence of a 50-fold molar excess of cold specific competitor (MCB-TMPI; lane 3). Addition of a 50-fold molar excess of a cold point-mutated MCB-TMPl (lane 4; 5 ′ and 3 ′ MCBs mutated, see Fig. 4) does not lead to a significant reduction in the amount of p 120 detected. The positions of the molecular weight markers (kDa) run in lane I are indicated. (C) As in B, except that cross-linked complexes were immunoprecipitated with SWI6-specific (lanes 2-4), SWI4-specific (lane 5) or DHFR-specific (lane 6) antibodies after digestion with DNase I. No competitor (lanes 2, 5 and 6), 50-fold molar excess unlabelled MCB-TMPl (lane 3) and 50-fold molar excess of unlabelled MCB-TMPl, both of whose MCBs are mutated (lane 4), were added during the binding reactions.

The effect o/swi6 mutations on transcription from MCBs

To test whether MBF also binds MCBs in vivo, we have analysed the dependence on SWI6 of a LacZ reporter gene whose transcription is due to the fusion of the MCB-TMP1 oligonucleotide upstream of the CYC1 TATA box (Guáreme and Mason, 1983). We find that the modest activation caused by MCB-TMP1 is noticeably.STWó-dcpendent (Fig. 6). This result tentatively suggests that the MBF:DNA complexes observed by us in vitro also form in vivo and may play a role in transcriptional activation.

Fig. 6.

The levels of β -galactosidase (PGal) activity in wild type (WT) and swi6 Δ mutant strains containing plasmids expressing LacZ either from a UAS-less CYC1 promoter (pLGA178; Guarente and Mason, 1983) or from a derivative in which MCB-TMP1 with Xhol sticky ends was cloned in a reverse orientation in the Xhol site ofpLG Δ 178.

Fig. 6.

The levels of β -galactosidase (PGal) activity in wild type (WT) and swi6 Δ mutant strains containing plasmids expressing LacZ either from a UAS-less CYC1 promoter (pLGA178; Guarente and Mason, 1983) or from a derivative in which MCB-TMP1 with Xhol sticky ends was cloned in a reverse orientation in the Xhol site ofpLG Δ 178.

It is more relevant, however, to analyse the effect of mutating SWI6 on the transcription of the TMP1 gene itself. Using S1 mapping, we find no effect on either the level or the start sites of TMP1 transcripts (Fig. 7). This is surprising because TMP1 transcription is completely dependent upon the MCBs within its promoter (McIntosh et al., 1991). In contrast, we do see effects on the RNR1 gene (Fig. 7), which is transcribed at START and contains MCBs. Mutation of SWI6 but not of SWI4 causes an increase in the number of RNR1 mRNAs that start upstream of the normal cluster of START sites. The latter are somewhat reduced. A similar phenomenon has been observed at the SWI5 promoter, where mutating either trans-acting factors or their binding sites not only reduces correct transcription but also increases abnormal upstream initiation (Lydall et al., 1991). These data suggest that a SWI6-containing factor (MBF) has a definite role in RNRI transcription but leave its role at TMP1 unresolved.

Fig. 7.

TMPI and RNRl RNA levels in swi4 and swi6 mutants. TMPI and RNRl RNAs were measured by an S1 protection assay (Nasmyth, 1983) and quantified using a Molecular Dynamics phospho-imager. The levels of TMPI expression (normalized against MATal (al), which was used as an internal control) were 102% for Swis4 Δ (K1899) and 90% for swi6 Δ (K1970) relative to wild type (WT) (KI 107; 100%). The ratios for “normal” RNRl transcripts were 104% for Δ and 59% for swi6 Δ. Strains are as described by Nasmyth and Dirick (1991).

Fig. 7.

TMPI and RNRl RNA levels in swi4 and swi6 mutants. TMPI and RNRl RNAs were measured by an S1 protection assay (Nasmyth, 1983) and quantified using a Molecular Dynamics phospho-imager. The levels of TMPI expression (normalized against MATal (al), which was used as an internal control) were 102% for Swis4 Δ (K1899) and 90% for swi6 Δ (K1970) relative to wild type (WT) (KI 107; 100%). The ratios for “normal” RNRl transcripts were 104% for Δ and 59% for swi6 Δ. Strains are as described by Nasmyth and Dirick (1991).

A regulatory role for SWI6

P120 appears to be the primary DNA-binding moiety of MBF and it may therefore be capable of binding to MCBs without SWI6. Though SWI6 seems necessary for MBF to bind to MCBs in vitro and to activate a LacZ reporter gene, it is possible that conditions within nuclei allow pl20 to bind without SWI6 in vivo. Thus, TMP1 may be activated by a SWI6-containing MBF in wild-type cells but by pl20 alone in swi6 mutants. This raises the possibility that, although the overall level in cycling cells is not largely altered, transcription of TMP1 in the absence of SWI6 might be abnormally regulated.

Addressing the regulatory role of SWI6 is complicated by swió mutants being grossly altered in their cell cycle control. They are partially defective in Gi cyclin function (Nasmyth and Dirick, 1991) and in some aspects of bud formation, cytokinesis or cell separation, with the result that it is difficult to synchronize swi6 mutants by conventional means. Release from an a factor-induced Gi arrest is slow and asynchronous and it is difficult to isolate a pure population of unbudded Gi cells by centrifugal elutriation. We have therefore synchronized cells by directly manipulating Gi cyclin levels. We have made congenie SWI6+ and swi6A strains whose three Gi cyclin genes CLN1, 2 and 3 have been deleted and whose growth is dependent on the conditional expression of CLN1 from the GALI-10 promoter. Removal of galactose from the medium causes such cells to arrest in Gi and its subsequent re-addition leads, at least in SWI6+ cells, to their synchronous passage through START (Cross and Tinkelenberg, 1991). As a measure of synchrony, we have analysed SWI5 transcription because its S/G2/M phase-specific transcription is not directly dependent upon SWI6 (Lydall et al., 1991). SW/5 transcripts decline upon galactose starvation and transiently increase 40 minutes following its restoration in both SWI6+ and swi6A cells (Fig. 8). Similar results were obtained for CLB2, which encodes a B-type cyclin that is expressed at the same stages of the cell cycle as SW15 (Surana et al., 1991; Ghiara et al., 1991) (Fig. 8). Histone H2B RNAs which appear somewhat earlier are also cell cycle regulated in both strains (data not shown). By these criteria, then, the synchrony of swi6 mutants is almost, but not quite, as good as wild type.

Fig. 8.

The effect of mutating SW16 on the cell cycle periodicity of genes containing MCBs and SCBs. Congenie SW16+ and swi6 Δ strains whose passage through START is dependent upon expression of CLN I from the GALl-lO promoter were grown in YEP medium containing 2% raffinose and 2% galactose at 30°C. At time=0, both cultures were harvested by filtration, washed in YEP raff medium and re-inoculated in YEP raff medium (—Gal). After 150 min, galactose was re-added to the medium (+Gal). SWI5, TMP1, CLN2, RNR1 and CLN1 RNAs were measured by SI protection (Nasmyth, 1983), whereas CLB2 RNAs were measured by northern blotting. All RNAs were quantified using a Molecular Dynamics phospho-imager and normalized using a MATal internal control. The ratio of these RNAs to MATal multiplied by 100 is plotted. Autoradiographs of SI-protected fragments due to SWI5, TMP1 and CLN2 RNAs are shown above (WT; K2771) and below (swi6A) the relevant graphs. K2771 (clnl Δ, cln2 Δ, cln3 Δ, YCpGAL-CLNl; congenie with KI 107) was a meiotic segregant from the diploid K2342 (Nasmyth and Dirick, 1991) transformed with YCpGAL-CLNl (Cross and Tinkelenberg, 1991). The swi6::TRPl derivative (K2786) was made by transformation. Both strains grow only in the presence of galactose. c

Fig. 8.

The effect of mutating SW16 on the cell cycle periodicity of genes containing MCBs and SCBs. Congenie SW16+ and swi6 Δ strains whose passage through START is dependent upon expression of CLN I from the GALl-lO promoter were grown in YEP medium containing 2% raffinose and 2% galactose at 30°C. At time=0, both cultures were harvested by filtration, washed in YEP raff medium and re-inoculated in YEP raff medium (—Gal). After 150 min, galactose was re-added to the medium (+Gal). SWI5, TMP1, CLN2, RNR1 and CLN1 RNAs were measured by SI protection (Nasmyth, 1983), whereas CLB2 RNAs were measured by northern blotting. All RNAs were quantified using a Molecular Dynamics phospho-imager and normalized using a MATal internal control. The ratio of these RNAs to MATal multiplied by 100 is plotted. Autoradiographs of SI-protected fragments due to SWI5, TMP1 and CLN2 RNAs are shown above (WT; K2771) and below (swi6A) the relevant graphs. K2771 (clnl Δ, cln2 Δ, cln3 Δ, YCpGAL-CLNl; congenie with KI 107) was a meiotic segregant from the diploid K2342 (Nasmyth and Dirick, 1991) transformed with YCpGAL-CLNl (Cross and Tinkelenberg, 1991). The swi6::TRPl derivative (K2786) was made by transformation. Both strains grow only in the presence of galactose. c

In contrast to the above, mutation of SW16 has a dramatic effect on the behaviour of genes whose transcription is dependent upon MCBs. The sharp periodicity of TMP1 RNAs seen in the wild type is abolished in a swi6 mutant, which has neither a drop following removal of galactose nor any increase following its restoration (Fig. 8). A similar result is seen for the RNRl gene whose normal (Fig. 8) and abnormal transcripts (data not shown) are de-regulated. It appears therefore that the cell cycle regulation of several MCB-activated genes is dependent upon SWI6.

SW16 and SCB regulation

SWI6 participates not only in complexes that bind MCBs, but also in complexes with SWI4 that bind SCBs (Andrews and Herskowitz, 1989; Taba et al., 1991). Might the cell cycle periodicity of SCB-driven transcription also depend on SWI6? The CLN2 promoter contains SCBs and its transcripts are severely reduced in swi4 mutants and more modestly reduced in swi6 mutants (Nasmyth and Dirick, 1991). To assess the dependence on SW14 of CLN2 transcription in swi6 mutants, we have measured the level of CLN2 transcripts in wild type and swi4 and swi6 mutant strains that express CLN2 from the 5. pombe ADH promoter (spADH-CLN2). This promoter/gene fusion has no effect on the level of endogenous CLN2 mRNAs (data not shown) and it helps the growth of swi mutants, rescuing lethality in the case of swi4;swi6 double mutants and ameliorating the growth defects of both single mutants. Fig. 9 shows that swi4;swi6 double mutants have much less endogenous CLN2 RNA than either single mutants (RNAs from the exogenous and endogenous CLN2 genes can be distinguished by SI mapping). Thus, SWI4 is responsible for most of the CLN2 transcription seen in swi6 mutants. Fig. 8 shows that the cell cycle periodicity of CLN2 RNAs is reduced but not entirely

Fig. 9.

The levels of CLNl and CLN2 RNAs as measured by S1 mapping in wild-type (WT; K2832), swi4A (K2833), swi6A (K2831)and swi4 Δ;swi6 Δ (K2392; Nasmyth and Dirick. 1991) mutants carrying two copies of spADH-CLN2. The spADH-CLN2 fusion has little or no effect on the level of endogenous CLN2 transcripts. The graph plots the ratio of endogenous CLNl and CLN2 RNAs relative to the unregulated spADH-CLN2 RNA.

Fig. 9.

The levels of CLNl and CLN2 RNAs as measured by S1 mapping in wild-type (WT; K2832), swi4A (K2833), swi6A (K2831)and swi4 Δ;swi6 Δ (K2392; Nasmyth and Dirick. 1991) mutants carrying two copies of spADH-CLN2. The spADH-CLN2 fusion has little or no effect on the level of endogenous CLN2 transcripts. The graph plots the ratio of endogenous CLNl and CLN2 RNAs relative to the unregulated spADH-CLN2 RNA.

START-dependent transcription factors in S. cerevisiae absent in swi6 cells, suggesting that SWI6 is required for regulating SWI4 activity. Cell cycle-regulation of SWI4-dependent (Fig. 9) CLN1 transcription is also dependent on SWI6 (Fig. 8).

Most genes involved in DNA replication in yeast are transcribed transiently during late Gi as cells undergo START. Their promoters all contain one or more versions of an 8-base pair motif (ACGCGTNA) called the Mhd cell cycle box (MCB; Pizzagalli et al., 1988). MCBs have been shown to be both necessary and sufficient for the late G1-specific transcription of the TMP1 and POLI genes (McIntosh et al., 1991; Gordon and Campbell, 1991). A different late G1-specific transcription element called the SCB (CAC-GAAAA; Nasmyth, 1985) is bound by a complex containing the SWI4 and SWI6 proteins (Andrews and Herskowitz, 1989; Taba et al., 1991). We describe here the formation in vitro of complexes on MCBs within a TMP1 promoter fragment. These complexes contain the SWI6 protein and a 120 kDa protein (pl20) that is distinct from SWI4. The 17 kDa protein claimed to bind MCBs (Verma et al., 1991) could be either a proteolytic fragment of pl20 or an unrelated protein. Lowndes et al. (1991) have described a factor that binds to an artificial oligonucleotide composed of three tandem Mhd restriction sites. This oligonucleotide did not however contain all the sequences sufficient for the cell cycle regulation characteristic of MCB-containing promoters. The complexes formed on it might nevertheless be the same that we have characterized because they have similar electrophoretic properties.

Three lines of evidence suggest that SWI6/pl20 complexes also bind MCBs in vivo and are responsible for their transcriptional regulation. First, bases whose methylation or carbéthoxylation interfers with SWI6/pl20-binding in vitro all lie within the minimal functional MCB (McIntosh et al., 1991). Second, SWI6 is required in vivo for the transcriptional activation by TMP1 MCBs of a LacZ reporter gene. Third and most important, transcription of endogenous genes whose promoters contain MCBs is in some cases modestly reduced and in all cases de-regulated in swi6 mutants. Our demonstration that SWI6/pl20 complexes bind cooperatively in vitro to the two TMP1 MCBs is consistent with the conclusion of McIntosh et al. (1991) that mutation of either MCB interferes with TMP1 transcription in vivo.

SWI6 may play a central role in late Gi or START-dependent transcription in yeast because it is the common component of the two different START-dependent transcription factors characterized (Fig. 10). SWI6 forms complexes with SWI4 on SCBs, but with p120 on MCBs. Since SWI4 alone can bind to SCBs in vitro (Primig et al., 1992) and since pl20 but not SWI6 can be cross-linked to MCBs, we propose that the slightly different DNA sequence specificities of the two types of SWI6-containing complexes are due to the site-specific DNA-binding domains of SWI4 and pl20. These domains may prove to be related because the SCB and MCB are similar sequences and can cross-complete in binding assays (Fig. IA). SWI6 appears to help SWI4 and pl20 in binding SCBs and MCBs in vitro, but this function may be critical in vivo only for some promoters such as HO (Breeden and Nasmyth, 1987a), CLNl (Fig. 9) and the artificial MCB-TMPl/CYC-LacZ promoter fusion (Fig. 6). To explain why the level of TMPl transcripts in cycling cells is unaffected by the deletion of SWI6 whereas extracts from swi6 mutants fail to form any specific complexes on TMPl MCBs in vitro, we propose that pl20 alone may be capable of binding to MCBs, but with a reduced affinity. Thus, it may bind and activate transcription when present at high concentration within intact nuclei but can no longer bind to MCBs in vitro when diluted in extracts. Such a phenomenon would be analogous to the activation of CLN2 by SWI4 in the absence of SWI6.

Fig. 10.

SWI6 has a central role in modulating START dependent transcription. SWI6 participates in the function of two different START-dependent promoter elements. It forms one type of complex with SWI4 on SCBs found in the HO and CLN2 promoters and another type of complex with a 120 kDa protein on MCBs within the TMP1 promoter. The activity of both types of complex are dependent upon the CDC28 kinase.

Fig. 10.

SWI6 has a central role in modulating START dependent transcription. SWI6 participates in the function of two different START-dependent promoter elements. It forms one type of complex with SWI4 on SCBs found in the HO and CLN2 promoters and another type of complex with a 120 kDa protein on MCBs within the TMP1 promoter. The activity of both types of complex are dependent upon the CDC28 kinase.

Whereas SWI6 may be dispensable for the binding of p120 and SWI4 to many promoters, it seems to fulfill an essential role in regulating transcription exerted by them. Genes that contain either MCBs (TMP1 RNR1) or SCBs (CLN2) are no longer properly cell cycle regulated in swi6 mutants. The former become totally aperiodic but the latter merely less periodic. CLN2 transcription is only modestly reduced in swi6 mutants but drastically reduced in swi4;swi6 double mutants, suggesting that SWI6 is required to regulate transcriptional activation by SWI4. We propose by analogy that SWI6 performs the same function for pl20 (Fig. 11). This SWI6 regulation must involve not only the repression of transcription at stages of the cell cycle when MCBs and SCBs are normally inactive, but also the activation of transcription during late G1. SWI6 is somewhat analogous to the GAL80 protein which binds to the GAL4 transcription factor and confers galactose dependence (Lue et al., 1987). Neither GAL80 nor SWI6 alone appear to be capable of binding DNA specifically. An important difference is that GAL80 has a negative role only, whereas SWI6 acts as both repressor and activator. We suggest therefore that it be called a modulator, a name that seems apt for a factor that imposes periodicity on transcription. The tumour-suppressing retinoblastoma gene in humans may perform an analogous function in modulating transcription due to the E2F transcription factor (Devoto et al., 1992).

Fig. 11.

SWI6 is not essential for transcription due to MCBs but is necessary for its correct regulation. We propose that pl20 can bind MCBs without SWI6 in vivo but that it activates transcription at a low level constitutively. SWI6 is required for cell cycle modulation of MCB-driven transcription.

Fig. 11.

SWI6 is not essential for transcription due to MCBs but is necessary for its correct regulation. We propose that pl20 can bind MCBs without SWI6 in vivo but that it activates transcription at a low level constitutively. SWI6 is required for cell cycle modulation of MCB-driven transcription.

It is likely that cell cycle changes in the factors that bind SCBs and MCBs are responsible for their cell cycle regulation. The aperiodicity of swi6 mutants suggests that cell cycle changes in the SWI6 protein could impose cell cycle control. An obvious possibility is that phosphorylation of SWI6 by the CDC28 (or another) kinase transforms it from a repressor to an activator, around the time of START. It is also possible that the changes triggering transcription at START occur on other components of the complexes (eg. SWI4 and pl20) and that SWI6 is required to prevent these proteins adopting an active conformation in the absence of such changes. Our detection of SWI6/pl20-binding activity throughout the cell cycle suggests that changes in an “activation” function rather than binding per se may trigger MCB transcription at START; a hypothesis that is consistent with our proposal that pl20 activates TMP1 constitutively in the absence of SWI6. Lowndes et al. (1991) have claimed that variations in an M/t/I-binding factor are responsible for cell cycle-regulated MCB transcription. Neither our nor their data warrant this conclusion. In vivo foot-printing will nevertheless be required to determine whether SWI6/pl20 complexes are bound to MCBs throughout the cell cycle.

Further analysis of the phenotype of swi6 mutants may provide insight into the importance of regulated MCB function. swi6 mutant cells are large and misshapen even when G1 cyclins are expressed ectopically (data not shown). This phenotype could stem from de-regulation of MCB transcriptional activation. The proposal (Nasmyth and Dirick, 1991) that the cdc10 protein from Schizosaccharomyces pombe, whose C-terminal half is very similar to that of SWI6 (Breeden and Nasmyth, 1987b), may perform similar functions, has recently been confirmed by the observation that cdc10 is a component of an MCB-binding factor (Lowndes et al., 1992). There is, however, at least one crucial difference between the functions of cdc10 and SWI6. Cdc10 is an essential gene whereas SWI6 is not. Cdc10 could have a more crucial role in DNA-binding than SWI6 and may therefore be indispensable for the transcription of an essential cell cycle gene.

Transcriptional activation by SWI4/SW16 complexes binding to SCBs appears essential for the accumulation of unstable G1 cyclins and thereby for further cell cycle progression. In contrast, most of the genes regulated by MCBs in yeast encode stable proteins that do not need to be synthesised immediately prior to S phase (Byers and Sowder, 1980). There may nevertheless be some MCB-regulated genes whose expression in late Gi is required for subsequent S phase entry. Transcriptional periodicity confered by SWI6 could prove to have an important role in coordinating cell cycle events.

We would like to thank Fred Cross for the YCpGAL-CLNl plasmid, Gustav Ammerer and Peter Sorger for heparin agarose-purified fractions, Rita Taba for yeast extracts, Hannes Tkadletz for preparing the figures, Robert Kurzbauer and Gotthold Schaffner for oligonucleotide synthesis and DNA sequencing, and Beverley Errede, Christian Koch and Meinrad Busslinger for their comments on the manuscript.

Ammerer
,
G.
(
1990
).
Identification, purification and cloning of a polypeptide (PRTF/GRM) that binds to mating-specific promoter elements in yeast
.
Genes Dev
.
4
,
299
312
.
Andrews
,
B. J.
and
Herskowitz
,
I.
(
1989
).
The yeast SWI4 protein contains a motif present in developmental regulators and is part of a complex involved in cell-cycle dependent transcription
.
Nature
342
,
830833
.
Andrews
,
B. J.
and
Herskowitz
,
I.
(
1990
).
Regulation of cell-cycle dependent gene expression in yeast
.
J. Biol. Chem
.
265
,
14057
14060
.
Barberis
,
A.
,
Widenhorn
,
K.
,
Vitelli
,
K.
and
Busslinger
,
M.
(
1990
).
A novel B-cell lineage-specific transcription factor present at early but not late stages of differentiation
.
Genes Dev
.
4
,
849
859
.
Breeden
,
L.
and
Nasmyth
,
K.
(
1987a
).
Cell cycle control of the yeast HO gene: cis- and tran.s-acting regulators
.
Cell
48
,
389
397
.
Breeden
,
L.
and
Nasmyth
,
K.
(
1987b
).
Similarity between cell-cycle genes of budding yeast and fission yeast and the Notch gene of Drosophila
.
Nature
329
,
651
654
.
Byers
,
B.
and
Sowder
,
L.
(
1980
).
Gene expression in the yeast cell cycle
.
J. Cell Biol
.
87
,
6a
.
Cross
,
F.
and
Tinkelenberg
,
A. H.
(
1991
).
A potential positive feedback loop controlling CLN1 and CLN2 gene expression at the start of the yeast cell cycle
.
Cell
65
,
875
883
.
Devoto
,
S. M.
,
Mudryj
,
M.
,
Pines
,
J.
,
Hunter
,
T.
and
Nevins
,
J. R.
(
1992
).
A cyclin A-protein kinase complex possesses sequence specific DNA binding activity: p33cdk2 is a component of the E2F-cyclin A complex
.
Cell
66
,
167
176
.
Dirick
,
L.
and
Nasmyth
,
K.
(
1991
).
Positive feedback in the activation of Gi cyclins in yeast
.
Nature
351
,
754
757
.
Elledge
,
S. J.
and
Davis
,
R. W.
(
1990
).
Two genes differentially regulated in the cell cycle and by DNA damaging agents encode alternative regulatory subunits of ribonucleotide reductase
.
Genes Dev
.
4
,
740751
.
Ghiara
,
J. B.
,
Richardson
,
H. E.
,
Sugimoto
,
K.
,
Henze
,
M.
,
Lew
,
D. J.
,
Wittenberg
,
C.
and
Reed
,
S. I.
(
1991
).
A cyclin B homolog in S. cerevisiae: chronic activation of the CDC28 protein kinase by cyclin prevents exit from mitosis
.
Cell
65
,
163
174
.
Gordon
,
C. B.
and
Campbell
,
J. L.
(
1991
).
A cell cycle-responsive transcriptional control element and a negative control element in the gene encoding DNA polymerase apha in S. cerevisiae
.
Proc. Natl. Acad. Sci. USA
88
,
6058
6062
.
Guarente
,
L.
and
Mason
,
T.
(
1983
).
Heme regulates transcription of the CYC-1 gene of S. cerevisiae via an upstream activation site
.
Cell
32
,
1279
1286
.
Kuwabara
,
M. D.
and
Sigman
,
D. S.
(
1987
).
Foot-printing DNA protein complexes in situ following gel retardation assays using 1, 10-phenanthroline-copper ion: Escherichia coli RNA polymerase - lac promoter complexes
.
Biochemistry
26
,
7234
7238
.
Laemmli
,
U. K.
(
1970
).
Cleavage of structural proteins during the assembly of the head of bacteriophage T4
.
Nature
227
,
680
685
.
Lorincz
,
A. T.
and
Reed
,
S. I.
(
1984
).
Primary structure homology between the product of yeast division control gene CDC28 and vertebrate oncogenes
.
Nature
307
,
183
185
.
Lowndes
,
N. F.
,
Johnson
,
A. L.
and
Johnston
,
L. H.
(
1991
).
Coordination of expression of DNA synthesis genes in budding yeast by a cell-cycle regulated trans factor
.
Nature
350
,
247
250
.
Lowndes
,
N. F.
,
Mclnerny
,
C. J.
,
Johnson
,
A. L.
,
Fantes
,
P. A.
and
Johnston
,
L. H.
(
1992
).
Control of DNA synthesis genes in fission yeast by the cell-cycle gene cdcUr
.
Nature
355
,
449
453
Lue
,
M. F.
,
Chasman
,
D. L
,
Buchman
,
A. R.
and
Kornberg
,
R. D.
(
1987
).
Interaction of GAL4 and GAL80 gene regulatory proteins in vitro
.
Mol. Cell Biol
.
7
,
3446
3451
.
Lydall
,
D.
,
Ammerer
,
G.
and
Nasmyth
,
K.
(
1991
).
A new role for MCM1 in yeast: cell cycle regulation of SWI5 transcription
.
Genes Dev
.
5
,
24052419
.
Maxam
,
A. M.
and
Gilbert
,
W.
(
1980
).
Sequencing end-labeled DNA with base-specific chemical cleavages
.
Methods Enzymol
.
65
,
499
560
.
McIntosh
,
E. M.
,
Atkinson
,
T.
,
Storms
,
R. K.
and
Smith
,
M.
(
1991
).
Characterization of a short, cis-acting DNA sequence which conveys cell cycle stage-dependent transcription in 5
.
cerevisiae. Mol. Cell Biol
.
11
,
329
337
.
Nash
,
R.
,
Tokowa
,
G.
,
Anand
,
S.
,
Erickson
,
K.
and
Futcher
,
A. B.
(
1988
).
The WHI+ gene of Saccharomyces cerevisiae tethers cell division to cell size and is a cyclin homolog
.
EMBO J
.
7
,
4335
4346
.
Nasmyth
,
K.
(
1983
).
Molecular analysis of a cell lineage
.
Nature
302
,
670676
.
Nasmyth
,
K.
(
1985
).
A repetitive DNA sequence that confers cell-cycle START (CDC28)-dependent transcription of the HO gene in yeast
.
Cell
42
,
225
235
.
Nasmyth
,
K.
and
Dirick
,
L.
(
1991
).
The role of SWI4 and SW16 in the activity of G, cyclins in yeast
.
Cell
66
,
995
1013
.
Ogas
,
J.
,
Andrews
,
B. J.
and
Herskowitz
,
I.
(
1991
).
Transcriptional activation of CLN1, CLN2 and a putative new Gi cyclin (HCS26) by SWI4, a positive regulator of Gi-specific transcription
.
Cell
66
,
10151026
.
Pizzagalli
,
A.
,
Valsasnini
,
P.
,
Plevani
,
P.
and
Lucchini
,
G.
(
1988
).
DNA polymerase I gene of S. cerevisiae: nucleotide sequence, mapping of a temperature-sensitive mutation and protein homology with other DNA polymerases
.
Proc. Natl. Acad. Sci. USA
85
,
3772
3776
.
Price
,
C.
,
Nasmyth
K.
and
Schuster
T.
(
1991
).
A general approach to the isolation of cell cyle-regulated genes in the budding yeast, Saccharomyces cerevisiae
.
J. Mol. Biol
.
218
,
543
556
.
Primig
,
M.
,
Sockanathan
,
S.
,
Auer
,
H.
and
Nasmyth
,
K.
(
1992
).
Anatomy of a transcription factor important for the START of the cell cycle in Saccharomyces cerevisiae
.
Nature
358
,
593
596
.
Pringle
,
J. R.
and
Hartwell
,
L. H.
(
1981
).
The Saccharomyces cerevisiae cell cycle
.
In The Molecular Biology of the Yeast Saccharomyces: Life Cycle and Inheritance
, (ed.
J. N.
Strathem
,
E. W.
Jones
and
J. R.
Broach
) pp.
97
142
.
Cold Spring Harbor
.
Reid
,
B.
and
Hartwell
,
L. H.
(
1977
).
Regulation of mating in the cell cycle of S. cerevisiae
.
J. Cell Biol
.
75
,
355
365
.
Richardson
,
H. E.
,
Wittenberg
,
C.
,
Cross
,
F.
and
Reed
,
S. I.
(
1989
).
An essential Gi function for cyclin-like proteins in yeast
.
Cell
59
,
1127
1133
.
Sturm
,
R.
,
Baumruker
,
T.
,
Franua
,
B. R.
and
Herr
,
W.
(
1987
)
A 100-kD HeLa cell octamer binding protein (OBP100) interacts differently with two separate octamer-related sequences within the SV40 enhancer
.
Genes Dev
.
1
,
1147
1160
.
Surana
,
U.
,
Robitsch
,
H.
,
Price
,
C.
,
Schuster
,
T.
,
Fitch
,
L
,
Futcher
,
A. B.
and
Nasmyth
,
K.
(
1991
).
The role of CDC28 and cyclins during mitosis in the budding yeast S
.
cerevisiae. Cell
65
,
145
161
.
Taba
,
M. R.
,
Muroff
,
I.
,
Lydall
,
D.
,
Tebb
,
G.
and
Nasmyth
,
K.
(
1991
).
Changes in a SWI4, 6-DNA-binding complex occur at the time of HO gene activation in yeast
.
Genes Dev
.
5
,
2000
2013
.
Verma
,
R.
,
Patapoutian
,
A.
,
Gordon
,
C. B.
and
Campbell
,
J. L.
(
1991
).
Identification and purification of a factor that binds to the Mlul cell cycle box of yeast DNA replication genes
.
Proc. Natl. Acad. Sci. USA
88
,
71557159
.
Wittenberg
,
C.
,
Sugimoto
,
K.
and
Reed
,
S. I.
(
1990
).
Gi-specific cyclins of S. cerevisiae: cell cycle periodicity, regulation by mating pheromone and association with the p34CDC28 protein kinase
.
Cell
62
,
225
237
.