Kuang Yu Chen

Department of Chemistry and The Cancer Institute of New Jersey, Rutgers - The State University of New Jersey, Piscataway, NJ. 08855-0939

Received 8/19/97 Accepted 8/25/97

4. TRANSCRIPTIONAL BINDING SITES IN THE G1/S GENE PROMOTERS

4.1 Promoter analysis of G1/S genes

We have previously used the GCG program, FINDPATTERN, to scan the promoter region for trans-acting factor binding sites in nine G1/S genes in order to determine whether there is any common cis-element in the promoter regions of these genes (9). The most commonly present cis-element in these genes is the SP1 (promoter-specific transcription factor-1) binding site. However, SP1 in human normal diploid fibroblasts does not show any cell cycle- or age-dependent change in binding activity (10), suggesting that it could not be the candidate for the age-dependent "master switch". Although we could not find any common "master switch" motif, we noticed that the promoter regions of these G1/S genes share some common characteristics. We found that the age-dependent G1/S genes can be divided into two groups, termed Group 1 and Group 2, based on their promoter organizations. Genes in Group 1 contain multiple tandem CCAAT elements, similar to that in TK promoter whereas genes in Group 2 contain E2F site, similar to that in DHFR. Thus, TK and DHFR can be considered as the founding member for Group 1 and Group 2 genes, respectively.

4.2 Promoter organizations of TK and DHFR genes

Figure 4 schematically shows the features of the promoter organization of TK and DHFR genes. Two promoter regions in human TK gene, termed 28-bp (-155/-128) and 67-bp (-127/-61) fragment, give clear and prominent age-dependent binding activity as measured by gel mobility shift assay (10). The serum-induced CBP/tk binding activity is inversely proportional to the population doubling level (PDL) of IMR-90 cells and appears to be almost absent in cells derived from premature aging syndromes (22). Moreover, CBP/tk activity appears to be deregulated in transformed cells (22). CBP/tk recognizes both 28-bp and 67-bp fragments. Within each fragment there is a 19-bp repeat containing an inverted CCAAT element. The inverted CCAAT element and its flanking sequence match closely with the Y-box consensus sequence CTGATTGGYYRR (10/12 and 11/12 match). A 25-bp fragment (-172/-148) upstream to the distal CCAAT element in TK promoter contains an E2F-like (7/8 match) and Yi-like (9/10 match) element. However, this fragment does not show any binding activity in IMR-90 nuclear extracts (23). Although the significance of the dyad structure of the two 19-bp repeats is unclear, it can be certain that the CBP/tk binding sites represent the key cis-elements for cell cycle- and age-dependent regulation of human TK gene. In the case of DHFR, the functional promoter region contains two E2F sites, arranged in opposing directions as an imperfect palindrome. This region also contains an inverted Yi binding site and an SP1 site upstream to the E2F sites. The E2F site exhibits clear cell cycle- and age-dependent binding activities. In contrast, SP1 is constitutively active whereas Yi is absent in normal human cells (23). E2F complex is a heterodimer, consisting of one E2F family protein and one DP family protein (16, 17). Among the E2F and DP family proteins, E2F1appears to be the one responsible for the down-regulation of DHFR in senescent human IMR-90 cells (11).

Figure 4. Organization of the promoter sequence in human thymidine kinase (TK) and dihydrofolate reductase (DHFR) gene. Human TK promoter: The distal CBP/tk site (28-bp) is indicated in the box. The 19-bp repeat within the CBP/tk site is underlined. The two inverted CCAAT elements (ATTGG) are shown in bold letters. The putative E2F site (7/8 match) does not show binding activity (23). Human DHFR promoter: The E2F sites in DHFR are arranged in two opposing directions as an imperfect palindrome. The arrows indicate the directions of E2F sites which are shown in bold letters. The SP1 (GGGGCGGG) and Yi site (CCCGCCCCCT) are also indicated with arrows.

4.3 Group 1 G1/S genes with CCAAT repeats within their promoters

TK promoter contains two Y-box arranged in tandem. This characteristic feature is shared by six other G1/S genes as shown in Figure 5. Thus, RNR R2 subunit, histones H2A, H2B, and H3, PCNA, and cdc2 all contain two to four CCAAT elements arranged close to each other in sense or in antisense direction. Figure 6 compares the flanking regions of each CCAAT or inverted CCAAT element in these genes. Almost all of the CCAAT elements in these genes are part of a sequence identical or similar to Y-box consensus sequence with a sequence match ranging from 9/12 to 11/12. Except for cdc2, which also contains E2F sites, the tandem CCAAT element (Y-box) represents the only prominent motif structure within the ~400 bp region upstream relative to the ATG codon in all these Group 1 genes. In some cases, the CCAAT elements and its flanking regions form long repeats, similar to that in TK promoter. The most striking example is the three almost identical 20-bp repeats in the RNR R2 subunit gene. Within each of the 20-bp repeat of the RNR R2 promoter, the sequence match with the Y-box consensus sequence is only 8/12 or 9/12. Nevertheless, the putative binding complex for R2 promoter has been suggested to be NF-Y proteins (24).

Figure 5. Promoter of the Group 1 genes that contain tandem CCAAT element in their promoters. The accession number of each gene is indicated in the parenthesis. The numbers above the shaded boxes indicate the position of cis-elements relative to the ATG codon since the transcription initial site for some of the genes has not been determined. The CCAAT element is indicated with a forward arrow and the inverted CCAAT element a backward arrow. The cdc2 gene also contains E2F site at positions -243/-236 and -142/-135.

Figure 6. Sequence comparison of the flanking regions of inverted CCAAT elements. Sequence match with Y-box is indicated in the parenthesis. Rev in the parenthesis indicates that the antisense strand is used for comparison.

4.4 Group 2 G1/S genes with E2F site within their promoters

The most prominent feature of Group 2 gene promoters is that they all contain E2F site, similar to that in the DHFR promoter. These genes can be further divided into two subgroups, one contains only E2F site and the other contains a CCAAT element in addition to the E2F site. Figure 7 shows that DHFR and TS contain only E2F binding site, but no CCAAT element in their promoters. In contrast, the promoter regions of DNA polymerase alpha, histone H1, cyclin A and cdc2 contain both E2F site and CCAAT element. The E2F site in TS, cdc2, DNA polymerase alpha and DHFR, match perfectly with the E2F consensus sequence. The putative E2F site in H1 and cyclin A has one mismtch. The CCAAT element in DNA polymerase alpha, H1, and cyclin A also appear to be part of Y-box with a sequence match of 9/12 to 10/12 (data not shown).

Figure 7. Promoter analysis of the Group 2 genes that contain E2F site in their promoters. The accession number of each gene is indicated in the parenthesis. The numbers above the boxes indicate the position of E2F site (8-bp) or CCAAT element (5-bp) relative to the ATG codon. The arrow on or below the shaded boxes indicates the direction of CCAAT element.

4.5 Late G1/S genes containing other cis-elements within their promoters

Among the G1/S genes that we have examined, histone H4 is the only one that contains neither CCAAT nor E2F element. Instead, it contains a CRE element with one mismatch at position -473/-466 relative to ATG codon. Although CREBP binding activity appears to be serum-responsive and age-dependent (9), it is not clear whether CREBP is responsible for the age-dependent attenuation of H4 gene expression in senescent cells.

4.6 Promoter analysis of NF-YA, NF-YB and E2F1

Promoter analysis as described above suggests that transcription factors such as Y-box binding protein CBP/tk (or NF-Y) and E2F1 may play a key role in controlling the expression of other G1/S genes, in addition to TK and DHFR, during cell senescence. Thus, the study of the regulation of these transcription factor genes during senescence would represent a further step in the understanding of the regulatory network controlling cell aging. We, therefore, have also examined the promoter organization of these transcription factor genes. Figure 8A shows the putative transcription factor binding sites in the promoter regions of NF-YA, NF-YB and E2F1. Both NF-YA and NF-YB contain two CCAAT elements in tandem, similar to that found in Group 1 gene promoters. The sequences of 12-bp containing CCAAT (-320/-309 and -292/-281) in NF-YA are almost identical to that in TK gene (Figure 8B). In the case of E2F1 gene, its promoter contains four E2F-binding sites arranged in two imperfect palindromes. In addition to E2F sites, E2F1 promoter also contains two CCAAT element and one inverted CCAAT element upstream to E2F sites. Studies are currently underway to determine (i) whether NF-Y and E2F1 are transcriptionally regulated and (ii) how the cis-elements such as E2F site and multiple CCAAT elements contribute to the down-regulation of these transcription factors.

Figure 8. (A) Promoter analysis of transcription factors NF-Y and E2F1. The numbers indicate the positions of CCAAT and E2F element relative to ATG initiation codon. (B) Sequence comparison of the CCAAT elements and their flanking regions in NF-YA, NF-YB and E2F1 genes. Sequence match with Y-box is indicated in the parenthesis. Rev in the parenthesis indicates that the antisense strand is used for comparison.