Vol. 50 No. 4/2003 941–945 QUARTERLY Communication

The SeqA protein of Escherichia coli is not only the main negative regulator of DNA replication initiation but also a specific transcription factor. It binds to hemimethylated GATC sequences and, with somewhat different specificity, to fully methylated GATC regions. Recently, a microarray analysis was reported, in which transcriptomes of wild-type and DeltaseqA strains were compared. Although in the seqA mutant the levels of some transcripts were significantly decreased while certain transcripts were evidently more abundant relative to wild-type bacteria, no correlation between the presence of GATC motifs in promoter sequences and transcription activity was found. However, here we show that when larger DNA fragments, encompassing positions from -250 to +250 relative to the transcription start site, are analyzed, some common features of GATC distribution near the promoters activated by SeqA can be demonstrated. Nevertheless, it seems that the GATC pattern is not the only determinant of SeqA-dependence of promoter activity.

1995; Boye et al., 1996;Wold et al., 1998).The SeqA protein is involved in sequestration of the origin of chromosome replication, oriC, in the hemimethylated state (Lu et al., 1994;Slater et al., 1995).This regulatory process is crucial for precise control of bacterial cell cycle.
SeqA binds specifically to hemimethylated GATC sequences located on oligonucleotide fragments but no specific binding to analogous fully methylated or unmethylated DNA was reported (Brendler & Austin, 1999;Kang et al., 1999).However, when longer DNA fragments were employed, SeqA did bind a fully methylated DNA fragment bearing oriC strongly and specifically (Slater et al., 1995).Moreover, hemimethylated oriC and non-oriC sequences were bound by SeqA with similar affinity (Slater et al., 1995).Investigation of SeqA-binding sites on hemimethylated oriC revealed the same preferential binding to each side of the DnaA box R1, as was found for fully methylated oriC (Skarstad et al., 2000).Thus, this binding has the same specificity to certain GATC sites irrespective of whether they are fully or hemimethylated.According to this statement, it was demonstrated that SeqA interacts specifically with certain bacteriophage l DNA fragments containing GATC sequences both in the fully methylated and hemimethylated state (Słomińska et al., 2001).Perhaps surprisingly, recent studies indicated that apart from being the main negative regulator of DNA replication initiation, SeqA is also a specific transcription factor.Namely, this protein specifically stimulates transcription from bacteriophage l p R , p I and p aQ promoters (p I and p aQ require also the action of the l-encoded CII protein) both in vivo and in vitro, whereas some other l promoters (like p L and p E ) appear to be SeqA-independent (Słomińska et al., 2001;2003a;2003b).Very recently, effects of DseqA mutation on the expression of all E. coli genes were studied using microarray analysis (Lobner-Olsen et al., 2003).Transcription of certain genes was ei-ther decreased or increased in the seqA mutant, while levels of transcripts of most genes were similar (less than two-fold difference) in the wild-type and DseqA strains.
Since SeqA binds GATC sequences, a correlation between the presence of such motifs in the promoter sequence and stimulation or repression of transcription by the seqA gene product was analyzed.Lobner-Olsen et al. (2003) did not find any common features of the promoters whose activities appeared to be significantly changed in the seqA mutant relative to wild-type bacteria.Therefore, those authors excluded the possibility that SeqA acts directly at promoter sequences, and suggested that the decreased or increased transcription of certain genes in the seqA mutant may result from changes in nucleoid organization.However, contrary to that conclusion, it was demonstrated experimentally that SeqA stimulates transcription from certain promoters both in vivo and in vitro, even when relatively short DNA templates were used in in vitro experiments (Słomińska et al., 2001;2003a;2003b).These results strongly suggested that the SeqA protein is directly involved in the regulation of activity of certain promoters.Interestingly, analysis of distribution of GATC motifs near bacteriophage l promoters, as well as the results of studies on SeqA-DNA interactions at these regions, strongly suggest that SeqA-binding sites downstream of a promoter may be important for the SeqA-mediated stimulation of promoter activity (Słomińska et al., 2003a;2003b).These sites can be located as far as 100-200 bp from the transcription start site (Fig. 1).Such regions were not analyzed by Lobner-Olsen et al. (2003).Therefore, here we analyzed the distribution of GATC motifs in regions of promoters (from position -250 to +250) of the E. coli genes activated and repressed in the seqA mutant (reported by Lobner-Olsen et al., 2003), and in randomly selected genes whose transcription was similar in the wild-type and DseqA strains.

MATERIALS AND METHODS
The list of genes whose transcription was either induced or repressed was taken from the work by Lobner-Olsen et al. (2003).Sequences of these genes (with adjacent regions), as well as of regions of randomly selected genes whose transcription was reported to be SeqA-independent (Lobner-Olsen et al., 2003), were taken from the GenBank database (http://www.ncbi.nml.nih.gov).The sequences were aligned to show locations of GATC motifs relative to the transcription start site of each promoter.

RESULTS AND DISCUSSION
Results of our analysis of distribution of GATC motifs in regions of promoters of E. coli genes activated, repressed and unaffected in the E. coli seqA mutant are presented in Figs. 2 and 3.Although it is difficult to find any patterns of the GATC motifs that might by obligatory for SeqA-mediated regulation of promoter activity, the distribution of these sequences near promoters that are depressed in seqA mutants (thus, potentially activated by SeqA) seems to be non-random (Fig. 2A).In a relatively large fraction of these promoters (9 out of 19), one GATC motif occurs in a region near 100 (between 60 and 140) bp downstream of the promoter (promoters of genes: cydD, gltB, trpE, holD, hisG, gapC, asnA, dld, osmE), often followed by at least one more such sequence located either downstream or upstream of the transcription start site (all promoters listed above The binding sites were determined using electron microscopy; the data are taken from Słomińska et al. (2001, 2003a, 2003b).Nucleotide positions (from -250 to +250) are relative to transcription start points (+1; dotted vertical line) and are shown at the top of the panel.except those for genes cydD, trpE and asnA).A similar distribution of GATC motifs occurs downstream of the bacteriophage l p I promoter that is stimulated by SeqA (S³omiñska et al., 2003a).Moreover, there are two neighboring GATC sequences between positions +150 and +200 in the regions of three additional promoters stimulated by SeqA (promoters for fecI, dps and tpr).Such double GATC motifs were found downstream of the SeqA-activated l p R and p aQ promoters (S³omiñska et al., 2001, 2003a, 2003b).These types of distribution of SeqA-binding sequences are significantly less frequent in regions of promoters that are repressed by SeqA (Fig. 2B) or are SeqA-independent (Fig. 3).Therefore, one may speculate that such patterns of GATC motifs are favorable for formation of SeqA-DNA complexes able to stimulate transcription from the nearby promoters.It is worth noting that these GATC motifs found near the SeqA-stimulated promoters generally do not follow the rules for SeqA binding proposed by Brendler & Austin (1999) on the basis of in vitro studies with oligonucleotides, namely the presence of at least two GATC sequences located on the same DNA side and separated by no more than 3 helical turns.However, it was demonstrated experimentally that when using longer DNA fragments, SeqA can bind to GATC regions distributed in a completely different manner than that suggested by Brendler & Austin (1999) as a requirement for SeqA-DNA interactions (Słomińska et al., 2001;2003a;2003b).Interestingly, many promoters that are not significantly affected by SeqA have relatively many GATC motifs in their regions while others are devoid of such sequences.
It is plausible that the location of the GATC motif(s) relative to a promoter, described above, results in stimulation of transcription initiation by SeqA.Nevertheless, other promoters, which do not reveal characteristic patterns of GATC motifs, might be depressed in seqA mutants.There are several possible reasons that could explain this phenomenon.First, SeqA-dependent regulation of transcription due to changes in nucleoid structure, suggested by Lobner-Olsen et al. (2003), cannot be excluded.Second, the regulation of transcription of some genes by SeqA may be indirect, due to putative SeqA-mediated control of expression of other genes, whose products might be involved in transcription of those showed in Fig. 2A and 2B (note that in this figure, only genes whose transcript levels differed between the wild-type and DseqA strains at least two times are shown, and even minor changes in expression of genes coding for regulatory proteins may have significant effects on expression of other genes, controlled by these proteins).Third, although only a few examples of specific SeqA-mediated activation of promoters have been reported to date, it is clear that there are various mechanisms of regulation of transcription by this protein.For example, bacteriophage l p R promoter seems to be directly acti- vated by SeqA (Słomińska et al., 2001;2003b), whereas seqA gene product stimulates the activity of p I and p aQ promoters by facilitating action of another transcription activator, the CII protein (Słomińska et al., 2003a).Therefore, various locations of GATC motifs may be necessary for different mechanisms of SeqA-mediated transcription stimulation.We conclude that it is likely that for direct stimulation of a promoter by SeqA, the presence of a GATC motif (followed by another such motif) in the +100 region, or a couple of GATC sequences in a region between +100 and +200, is required.Activation of other promoters shown in Fig. 2A is perhaps indirect, for instance due to changes in nucleoid structure.
The results of the microarray analysis performed by Lobner-Olsen et al. (2003) have one more implication.It was demonstrated previously that the properties of cellular membranes are significantly changed in seqA mutants (Węgrzyn et al., 1999).One possibility was that expression of genes coding for major membrane proteins, or genes whose products are involved in production or degradation of other membrane compounds, is affected in the absence of active SeqA.However, there are no such genes among those significantly stimulated or repressed by SeqA (Fig. 2).Therefore, other mechanism(s) must be responsible for the changes in cell membranes observed in seqA mutants.

Figure 1 .
Figure 1.Distribution of SeqA-binding sites (dots) in regions of bacteriophage l promoters (p R , p I and p aQ ) stimulated by the seqA gene product.

Figure 2 .
Figure 2. Distribution of GATC sequences (dots) in regions of promoters of genes for which decreased (panel A) or increased (panel B) mRNA levels in the DseqA mutant were demonstrated recently (Lobner-Olsen et al., 2003).Nucleotide positions (from -250 to +250) are relative to transcription start points (+1; dotted vertical line) and are shown at the top of each panel.Quantitative data on the levels of particular transcripts are given in Lobner-Olsen et al. (2003).

Figure 3 .
Figure 3. Distribution of GATC sequences (dots) in regions of randomly chosen promoters of genes for which mRNA levels in the DseqA mutant were found to be unaffected or changes relative to wild-type bacteria were less than two-fold (on the basis of the results reported by Lobner-Olsen et al., 2003); in fact, most E. coli genes belong to this group.