Mcm1 is required for the expression of many constitutively transcribed genes and a subset of the M- and M/G1
-specific transcripts. As a result, some fraction of the Mcm1 in a cell must be present in the nucleus and functional throughout the cell cycle. The restriction of its activity to either M or the M/G1
transition must be determined by the promoter context in which it binds. The factors which interact with Mcm1 and confer M-specific transcription have been identified recently (17
). In this paper we provide evidence that the M/G1
-specific and Mcm1-dependent complexes that form on ECB elements are also large and heterogeneous and contain at least one other protein. Moreover, we identify sequences flanking the Mcm1 binding site that affect the binding and activity of these complexes. We also characterize the binding of Mcm1 to ECBs in vivo during the cell cycle and in different carbon sources.
The Mcm1 binding site has been exhaustively studied. Site selection identified the minimal 10-bp sequence required for Mcm1 binding in vitro (34
). In addition, the larger 16-bp palindromic sequence that is required for function in vivo has been mutagenized at every position (1
) and the crystal structure of an Mcm1 fragment/Matα2/DNA complex has been determined (48
). As a result, the bases within the 16-bp palindrome that make contact with the Mcm1 and are required for Mcm1 binding and activity are known in at least a few contexts. In early studies, it was noted that the α-specific genes that are induced by Mcm1 and α1 show a striking divergence from the canonical Mcm1 binding site on the side adjacent to the α1 binding site and that the presence of α1 provides necessary stability to these complexes (21
). Thus, the degenerate Mcm1 binding site serves to make Mcm1 binding and activation of these promoters dependent upon the accessory factor α1, which is only present in α cells. In the case of the M-specific transcripts, many of the Mcm1 binding sites contain noncanonical residues on the side opposite to that to which the Fkh proteins bind. However, in the few cases tested, Fkh binding requires the presence of Mcm1 (2
) and both are bound constitutively through the cell cycle (23
). Activation requires the cell-cycle-regulated association of a third protein called Ndd1 (23
-specific promoter elements show a high degree of conservation across the 16-bp palindrome. In addition, there are symmetrically placed binding sites for the homeobox protein α2, which represses these genes in α cells. Activation of these genes in a
cells requires the Ste12 protein (12
). Unlike the other Mcm1 accessory proteins, Ste12 binding sites are not necessarily adjacent to the Mcm1 site; rather, they are often found in multiple copies and at variable distances from the Mcm1 binding site (24
). The best studied of the a
-specific genes is STE2
. Mcm1 can bind the STE2
element in the absence of Ste12, but it only weakly activates transcription (18
). This suggests that Mcm1 cannot activate transcription on its own; rather, it relies on associated proteins that confer this property to the complex. Interestingly, some a
-specific genes (including STE2
) are also cell cycle regulated and peak at the M/G1
), so they may have some regulatory elements in common with the ECB-regulated genes.
Alignment of the M/G1-specific promoter elements shows that they are symmetrically conserved across the palindrome, but little other sequence conservation is evident. Based upon the similarity between the Mcm1 binding sites in the M/G1-specific genes and the other well-studied sites, we expected that Mcm1 could bind these sites in the absence of accessory factors. Consistent with this, we have shown that in vitro-translated Mcm1 binds to these elements (Fig. and data not shown). Moreover, Mcm1 is bound to ECB elements throughout the cell cycle. So, just as with the M-specific and a-specific genes, the binding of Mcm1 to the promoter element is not sufficient to activate transcription. Other proteins or modifications of Mcm1 activity must be involved, and the DNA context of the ECB element must be responsible for their specification as M/G1-specific transcription elements.
The additional sequence information restricting ECB activity to the M/G1
boundary of the cell cycle could be distal to the Mcm1 binding site, as is the case with Ste12, or it could be embedded within it. In the case of CDC6
, both distal and proximal sequence elements may be in play because there is at least one Swi5 binding site near the fourth ECB element (11
). Swi5 is required for maximal transcription of this M/G1
-specific gene (40
), but Swi5 is also required for transcription of genes, like HO
, which are expressed at a later stage of the cell cycle (6
). At the HO
promoter, Swi5 has been shown to recruit chromatin remodeling factors that in turn enable the late G1
-specific transcription factors, Swi4 and Swi6, to bind and activate transcription at distal SCB elements (10
). Swi5 may act in an analogous fashion at a subset of the M/G1
-specific promoters. However, Swi5 is not responsible for the cell cycle specificity of ECB elements, because we have shown that small DNA fragments including the two tandem ECBs from CDC47
or the single SWI4
ECB, cloned into a lacZ
reporter construct, are sufficient to confer M/G1
-specific transcription (32
). Neither of these constructs includes a Swi5 binding site, so there is no reason to think that Swi5 is involved. Rather, the sequence information required to restrict ECB activity to the M/G1
boundary is likeliest to be embedded within the 16-bp palindrome.
In order to identify the cis
- and trans
-activators of ECB elements, we have carried out a series of experiments. Simple alignment of the elements shows that, in addition to maintaining preferred residues for Mcm1 binding within the palindrome, there is further conservation extending a few bases beyond the palindrome and at positions −3 and +3 within the palindrome, where mutagenesis and crystallographic studies indicate that Mcm1 should have no base-specific contacts (1
). Not all putative ECBs contain these additional conserved residues, but all but one of the M/G1
-specific promoters under study contain more than one putative ECB. Thus, we do not know which of these sites are active. It could be that M/G1
-specific regulation involves binding of another protein to a subset of these sites via conserved bases adjacent to and/or embedded within the otherwise palindromic Mcm1 binding site.
Our studies of the ECB binding complex verify the importance of Mcm1 in ECB activation, but they also reveal additional complexities that are indicative of the presence of other proteins in the ECB complex. DNase I protection studies showed that all four of the M/G1-specific promoters analyzed have complex patterns of protection of the ECB elements extending about 10 bp on one or both sides of the palindrome to which the Mcm1 dimer is known to bind. Gel retardation assays show that ECB complexes from crude cell extracts are highly heterogeneous compared to those formed with in vitro-translated Mcm1. Moreover, the nature and stability of ECB complexes are influenced by the sequence of the flanking DNA. Mutation of the flanking sequences from the fourth CLN3 ECB results in a threefold-higher dissociation constant for the binding complexes and reduces the variety of complexes that can be formed on the ECB.
The bases critical for complex formation and stability have not been exhaustively analyzed; however, we have noted a region of limited homology in the protected flanking region and shown that substitutions at the most conserved positions affect transcriptional activity of the ECB element in vivo. Cell cycle regulation persists in spite of these changes, but the activity is elevated and possibly extended for a broader interval of time. This suggests that the flanking sequence may affect the stability, rather than the composition, of the complexes that form on ECB elements through the cell cycle. The possibility that flanking sequences influence the activity of the ECB under specific environmental conditions (e.g., carbon source shifts) is being investigated.
Gel filtration shows that the binding complex on ECB elements is in excess of 200 kDa. This is far larger than expected for a dimer of Mcm1, which would be 70 kDa, so it is likely that other proteins are associated with the complex. In fact, the dimeric and monomeric forms of Mcm1 are not detectable in the gel filtration fractions. This indicates that most of the Mcm1 in the cell is associated with other proteins. The large complex which binds ECB elements in this assay lacks at least one protein of 25 kDa, as that size fraction must be added back to generate a band shift complex of wild-type mobility. We have assayed the formation of the ECB-specific band shift complex from extracts of cells deleted for the DNA binding proteins Swi5 (40
), the related protein Ace2 (31
), and Ste12 (12
). None of these proteins appears to be involved, as the behavior of the Mcm1-specific complex did not change (data not shown).
Mcm1 binding to ECB elements does not change through the cell cycle, but it is affected by changes in the carbon source. CHIP analysis shows that Mcm1 binding complexes on ECBs from the CLN3
, and SWI4
promoters are much more prevalent in cells grown on poor carbon sources like glycerol and raffinose than they are in glucose-grown cells. In spite of the increased binding, ECB elements are less active in poor carbon sources. This suggests that an inactive form of the complex is being stabilized on the ECB under nonoptimal growth conditions. PIS1
, another Mcm1-regulated gene, is also down-regulated in poor carbon sources (4
), so this may be a general property of Mcm1. Chen and Tye have shown that the activity of unstable alleles of MCM1
can be enhanced by reduced glycolytic flux (8
), but the signaling metabolite has not been identified. Mcm1 activity is also affected by osmotic stress and perhaps by other environmental changes mediated by the Sln1 two-component response regulator (50
). Modification of Mcm1 by phosphorylation has also been detected during salt stress (26
Further studies are required to understand the dynamics of Mcm1 activity at ECB elements through the cell cycle and in response to environmental cues. The products of the ECB-regulated genes under study either promote the G1-to-S transition (SWI4 and CLN3) or are involved in the formation of prereplication complexes at which DNA synthesis is initiated (CDC47 and CDC6). Understanding how internal and external signals influence the expression of these genes may provide new insights into the control of the early events of the cell cycle.