The major components of the Escherichia coli DNA replication machinery have been identified and largely characterized (Kornberg and Baker, 1992, DNA Replication). However, many questions concerning the cell cycle regulation of initiation at the origin of replication (oriC ) remain unanswered. In particular, it is not clear whether co-ordinated expression of genes involved in DNA replication directly contributes to synchronous initiation at oriC.
DnaA, the initiator, is the key protein involved in initiation of DNA replication. The expression of the dnaA gene has been studied extensively, and several models for DnaA controlled initiation of replication have been proposed (see Herrick et al., 1996, Mol Microbiol19: 659–666, and references therein). In these models, tight control of dnaA transcription was postulated to be the determinant for firing of the origins at precise time intervals, once per division cycle. The dnaA regulatory region consists of two promoters (dnaA1P, dnaA2P ) and a DnaA box (binding site for DnaA protein) located roughly midway between the promoters. In vivo and in vitro studies suggested that the expression of DnaA is autoregulated. It was reasonable to assume that DnaA regulates its own synthesis by binding to the DnaA box in the promoter region. However, mutations introduced into the DnaA box did not have the expected effect of dnaA derepression (Polaczek and Wright, 1990, New Biol2: 564–582; Smith et al., 1997, Mol Microbiol23:1303–1315). Furthermore, a mutant variant of this box was introduced into the chromosome with no apparent effects on cell growth or initiation synchrony (Smith et al., 1997, Mol Microbiol23:1303–1315). The role of this DnaA box remains unclear.
Many proteins involved in initiation of replication at E. coli oriC appear to have binding sites in the dnaA promoter region. IciA and Fis have been shown to bind at the dnaA promoter region in vitro (Lee et al., 1996, Mol Microbiol19: 389–396; Froelich et al., 1996, J Bacteriol178: 6006–6012). Mutations in IciA or Fis have only modest effects on DnaA transcript/protein levels. In addition, no defects on chromosomal initiation synchrony were detected in an IciA null mutant (Lee et al., 1996, ibid.). Fis protein has a highly degenerate consensus recognition sequence and IciA binds to AT-rich regions with no apparent consensus sequence. Therefore, the observed binding of Fis and IciA may be a consequence of their relaxed sequence recognition and have limited regulatory implications. A putative IHF site, identical to a site in oriC with respect to conserved bases (CAGnnnnTTGATC), was identified in the dnaA promoter region (Polaczek, 1990, New Biol2: 265–271). Despite this striking homology, IHF does not bind to this site. Binding of IHF is determined both by the recognition sequence and by the local DNA structure (Rice et al., 1996, Cell87: 1295–1306.). Therefore, the IHF site in the dnaA promoter region may be non-functional because of unfavourable DNA structure.
The dnaA regulatory region also contains GATC sites, and the dnaA gene is regulated at the level of methylation in that Dam methylation moderately stimulates dnaA gene expression (Braun and Wright, 1986, Mol Gen Genet202: 246–250) and hemimethylation inactivates the gene (Campbell and Kleckner, 1990, Cell62: 967–979). The latter may suggest the involvement of SeqA, a protein shown to bind to hemimethylated oriC sequences, in dnaA regulation (Slater et al., 1995, Cell82: 927–936). A twofold increase in the level of DnaA protein was found in seqA mutants (von Freiesleben et al., 1994, Mol Microbiol14: 763–772).
In summary, the expression of the DnaA protein does not appear to be tightly regulated at the level of transcription. Lack of complex transcriptional regulation of dnaA may not be surprising given that an increase in transcription of dnaA does not necessarily lead to increased protein levels, suggesting translational control (Polaczek and Wright, 1990, New Biol2: 564–582). Also, DnaA protein exists in the cells in ADP- and ATP-bound forms, the ATP form being active in initiation of replication (Sekimuzu et al., 1987, Cell50: 259–265), suggesting further post-translational regulation.
Yoshikawa and Ogasawara (1991, Mol Microbiol5: 2589–2597) have proposed that the dnaA regulatory region in E. coli was part of an ancestral origin of replication that lost the function of an origin through the deletion of all but one of the DnaA boxes. They further argue that this region lost all the characteristics of an origin, except for autogenous regulation of the dnaA gene. We agree that the dnaA locus has indeed lost the function of an origin, but we propose that it retains most of the signatures of a canonical prokaryotic replicator. It looks increasingly plausible that the regulatory sequences within and surrounding the dnaA promoters constitute an evolutionary fossil of no significant physiological relevance.
Identification of the multitude of regulatory elements in the dnaA promoter region may suggest that the gene is masterfully designed for precise and fine-tuned expression, but this scenario is lacking experimental support. Definitive answers await analysis of the effects of chromosomal mutations in the dnaA promoter region on initiation synchrony.