INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

The DnaA protein is an essential factor for initiation of duplication of the bacterial chromosome from a specific site, the origin of replication, oriC. The DnaA protein binds to DnaA boxes in oriC, and in vitro studies indicate that initiation takes place when sufficient DnaA proteinapproximately 20 monomershas been bound to oriC (8). The formation of this so-called initial complex leads to the opening of a region in oriC containing AT-rich 13-mers and allows the entry of DnaB and DnaC proteins to form the pre-prepriming complex, which is followed by several other stages, as detailed by Sekimizu et al. (30). The DnaA protein binding sitesthe DnaA boxeshave the consensus sequence TT^A_TTNCACA (29). There are 308 consensus DnaA boxes on the E. coli chromosome, three of which are located within the minimal oriC locus. Using the DNA binding domain of the DnaA protein to isolate restriction fragments from a digest of total chromosomal DNA, only a handful of fragments could be isolated and characterized as containing high-affinity DnaA protein binding motifs in vitro (27).

Initiation of chromosome replication is a complex process which, besides oriC and the DnaA protein, involves a number of accessory factors. In the period between one initiation and the next, oriC will undergo a number of structural changes to be prepared for the new initiation. The newly replicated (and hemimethylated) GATC Dam methylation sites in oriC facilitate membrane binding (24). In this state, oriC is inaccessible for initiation and it is conceivable that the eclipse period could be defined as the period during which oriC is sequestered (6). Also, other factors have roles in the initiation of chromosome replication (31). Many details of the initiation process are quite well understood due to the impressive work at Arthur Kornberg's laboratory (17). However, how the bacterial cell senses when initiation is supposed to take place in the cell cycle is still under debate.

Phenomenologically, the time of initiation is coupled to the mass increase such that initiation occurs at a critical mass per originthe initiation mass (9). The initiation mass, expressed as units of optical density per amount of DNA, appears to be nearly constant at different growth rates (4). Initiation can be thought to occur either when the mass per origin has increased sufficiently or when the number of origins per unit of mass has decreased sufficiently at a given growth rate. These views were reflected in the autorepressor model (33) and the inhibitor dilution model (25), respectively.

It has previously been shown that changing the DnaA protein concentration or activity changes the initiation mass. Overproduction of DnaA protein decreases the initiation mass (3, 20), i.e., increases the origin-to-mass ratio. Conversely, cells, which have lower DnaA protein activity than normal have a higher initiation mass. A lower DnaA protein concentration or activity has been obtained, for example, in dnaA(Ts) strains grown at nonpermissive temperatures, at which suboptimal DnaA protein concentrations were achieved by induction of normal DnaA protein from a plasmid system (20). Also, a dnaA(Ts) strain which was grown at semipermissive temperatures, at which it was expected that the activity of the DnaA protein would decrease (13), showed increasing initiation masses at increasing growth temperatures. These changes in initiation mass were paralleled by changes in dnaA gene expression; i.e., low DnaA protein activity led to derepression of the dnaA promoter, and high DnaA protein activity led to repression of the dnaA promoter (2, 5).

The initiator titration model (11) combines the views of the two models mentioned above. The essence of this model is that there is a long period of the cell cycle where newly synthesized DnaA proteins are titrated by binding to high-affinity binding sites. When the cell cannot titrate any more DnaA protein, there will be free DnaA protein molecules which can participate in a postulated lower-affinity reaction, namely, making the initiation complex. Thus, the high-affinity binding sites are postulated to be inhibitor elements preventing the DnaA protein from forming the initiation complex. In addition, the initiator titration model postulates that DnaA protein released at the time of initiation from one origin will increase the free-DnaA-to-origin ratio, thereby increasing the probability of initiation of the remaining origins. This will result in the observed synchronous initiation at multiple origins in fast-growing cells.

The experimental basis of the formulation of the initiator titration model was a study in which dnaA gene expression was determined in cells carrying additional DnaA boxes (12). In the present study, we have extended this analysis to further characterize the regulatory role of the DnaA boxes in the control of initiation of chromosomal replication. We introduced additional DnaA boxes carried on pBR322-derived plasmids into cells and studied their effect on initiation mass, cell size, and DNA content in balanced bacterial cultures growing at different rates. We have also investigated the effect of mutating one or more of the DnaA boxes in oriC with respect to the titrating ability by using the mutations described by Holz et al. (14).

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains and plasmids. The E. coli BBC119 is an LJ24 derivative (thi-1 leu-6 lacY1 lacIZ(MluI) supE44 tonA21 rpsL rfbD1) (26) which has been lysogenized with RB1 (5) that carries a dnaA'-'lacZ fusion. Southern blotting was used to check that the strain was a single lysogen of RB1 (data not shown). The plasmids used (Fig. 1) are deletion derivatives of plasmid pFHC271 that contain a complete oriC region cloned into pBR322. None of the deletion derivatives have a functional oriC locus (see reference 12 for details). Plasmids with mutations in the oriC DnaA boxes were constructed by exchanging a ClaI-BglII fragment in plasmid pFHC496 with the same fragment from the mutated DnaA box plasmid (14).

View larger version (22K): [in this window] [in a new window]   FIG. 1.   Plasmids carrying DNA from the oriC region of E. coli. The relevant genes are indicated above the map of plasmid pFHC271, which is a chimeric plasmid carrying oriC and neighboring sequences cloned into the HindIII site of pBR322; the positions of the tet and bla genes are shown. Plasmid pTAC909 (2) is identical to pBR322, except that the HindIII site has been destroyed, rendering the tet gene inactive. The remaining plasmids are deletion derivatives of pFHC271 and carry different numbers of DnaA boxes (represented by black boxes below the respective plasmids) from the oriC region. The restriction enzyme sites used to make the deletions are indicated at the different junctions: B, BamHI; Bg, BglII; C, ClaI; H, HindIII. The chromosomal oriC locus is inactivated on all of the plasmids except pFHC271. Plasmids pFHC496 and pFHC1425 carry all of the DnaA boxes of the minimal oriC locus but lack the small BglII fragment, which carries sequences from the 13-mer region of oriC, that is essential for oriC function.

Bacterial growth experiments. The host strain and its plasmid-containing derivatives were kept in balanced growth for more than 10 mass doublings in A+B medium (7) supplemented with 1% Casamino Acids-0.2% glucose, 0.2% glucose, 0.2% glycerol, or 0.4% succinate. Thiamine was always present at 2 µg/ml, and leucine was present in the minimal medium at 20 µg/ml.

Flow cytometric procedures. Samples were prepared and flow cytometry was performed as summarized in reference 32. Average cell mass was determined as average light scatter, and the average amount of DNA per unit of mass was determined as fluorescence per unit of light scatter of samples taken directly from exponentially grown cultures. The average number of origins per cell was determined from parallel samples incubated for more than 3 h with rifampin (300 µg/ml) to block initiation of replication and cephalexin (36 µg/ml) to block cell division (20). This treatment normally results in fully replicated chromosomes, which will be equivalent to the number of origins per cell at the time of drug addition and can be visualized directly by flow cytometry. The percentage of asynchronous cells was calculated as 100*N_async/(N_total  N₁), where N_async is the sum of all cells which do not contain 2ⁿ origins (n = 0, 1, 2, 3, 4, etc.), N_total is the total number of analyzed cells, and N₁ is the number of cells carrying 1 origin. The number of cells that have one chromosome is subtracted from the total number of cells because we cannot say anything about synchrony of initiation in such cells.

Enzyme measurements. Cell extracts prepared by treatment with toluene were used to determine -galactosidase activity as previously described (23).

Determination of plasmid copy number per oriC locus by Southern blotting. Total (plasmid and chromosome) DNA was prepared as previously described (10) with the modifications described previously (3). The DNA was restricted with EcoRI and HindIII, and Southern blot analysis was carried out as previously described (3) by using a [³⁵S]dATP-labeled probe mixture. Probes for hybridization were prepared by labeling DNA with [³⁵S]dATP using DNA polymerase I Klenow fragment and hexanucleotide random priming. We used probes, one made from a PCR-derived fragment of 295 bp from the tet gene of pBR322 (from position 723 to position 1017), and another prepared from a PCR-derived fragment of 1,197 bp, which will hybridize to the 2.1-kb HindIII fragment carrying most of the gidA gene from the oriC region, to estimate copy numbers per oriC locus in Southern hybridization experiments. The Southern blots were quantified by using the Instant Imager (Packard). We included samples of plasmid pFHC271 digested with HindIII to obtain exact plasmid copy numbers per oriC locus by normalizing the plasmid and chromosomal hybridization signals of the total DNA samples from the experiment to the hybridization signals of pFHC271 where the fragments representing the plasmid part and the chromosomal part are present at a 1:1 ratio.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

It was previously shown that the presence of extra DnaA boxes in strains carrying pBR322-derived plasmids with different combinations of the DnaA boxes from the oriC region (Fig. 1) would titrate DnaA protein to various degrees, depending on the number, and apparently also the quality, of the DnaA boxes (12). This titration was measured as the derepression of the autoregulated dnaA promoter of a dnaA'-'lacZ fusion gene positioned at the att site. In the present study, we have investigated how the introduction of additional DnaA boxes affects initiation of chromosome replication.

Mutations in DnaA boxes decrease titration and decrease changes in cell size. The plasmids we used in previous studies were different in structure and contained different parts of chromosomal DNA from the oriC region. Therefore, to prove that the effects we observed were a consequence of the presence of the DnaA boxes on the plasmids, we constructed a number of plasmids which had the same structure as plasmid pFHC496 (Fig. 1) but had mutations in the different DnaA boxes. Strains carrying these plasmids and control strains carrying plasmids with oriC fragments containing no, two, three, or four DnaA boxes were used to determine DnaA--galactosidase activity (Fig. 2A) as a measure of the derepression of the dnaA gene. We also determined cell size by determining the average light scatter of cells by flow cytometry (Fig. 2B). Changes in cell size should be expected if addition of extra DnaA boxes to cells changes the initiation mass.

View larger version (25K): [in this window] [in a new window]   FIG. 2.   Mutations in DnaA boxes cause phenotypes which are similar to the phenotypes caused by plasmids in which the corresponding DnaA boxes are absent. Expression of the dnaA'-'lacZ gene (measured as DnaA--galactosidase activity) is shown in panel A, and cell size estimated as light scatter in the flow cytometer is shown in panel B. These parameters are expressed relative to the values obtained in the control strain carrying plasmid pTAC909. The pBBC plasmids with mutations in different DnaA boxes are derivatives of plasmid pFHC496 (Fig. 1). The DnaA boxes present on the different plasmids (Fig. 1) are indicated to the left. The R1 box in plasmid pBBC166 (boxed number) contains a neutral change of the DnaA box from the R1 and R4 type (TTATCCACA) to the R2 type (TTATACACA), which has the same in vitro binding affinity as the R1 type (14). The remaining mutant DnaA boxes (circled numbers) all had the sequence TTTCCCACA. This sequence shows no affinity for DnaA protein in in vitro binding assays (14). Cultures of strains carrying the respective plasmids were grown in glucose minimal medium. In this growth medium, changes in relative cell size caused by the DnaA box plasmids were more readily observed than on a richer growth medium (see Fig. 3).

Cell size, origins, and total chromosomal DNA at different growth rates. To study how the addition of extra DnaA boxes affects the control of initiation of chromosome replication in more detail, we extended the flow cytometric analysis. We varied the intracellular DnaA box concentration by using strains carrying plasmids with different DnaA titration activities (pTAC909, pFHC496, and pFHC1425) and by growing these strains in media which would give different plasmid copy numbers due to the pBR322 copy number increase seen with a decreasing growth rate (1, 19).

View larger version (31K): [in this window] [in a new window]   FIG. 3.   Cell size and DNA distribution are changed in strains carrying plasmids supplying extra DnaA boxes. Cell size and DNA distribution were determined by flow cytometry as the light scatter and fluorescence distributions, respectively, of cell populations. Cell number is normalized to the total number of cells analyzed. Strains carrying plasmids pTAC909, pFHC496, and pFHC1425 were grown in minimal growth medium supplemented with glucose-Casamino Acids (A and B), glucose (C and D), or glycerol (E and F). The light scatter and fluorescence distributions obtained for strains carrying the three plasmids are shown in different colors as follows: pTAC909, light gray; pFHC496, dark gray; pFHC1425, black.

View larger version (115K): [in this window] [in a new window]   FIG. 4.   Fluorescence microscopy of plasmid-carrying strains without (pTAC909) and with additional (pFHC1425) DnaA boxes. The different growth media used are indicated. The same ethanol-fixed cells which were used for flow cytometry in Fig. 3 were stained with a fluorescent probe hybridizing to the 16S rRNA. Computer images of the fluorescent cells were collected via a charge-coupled device camera.

View larger version (28K): [in this window] [in a new window]   FIG. 5.   Three-dimensional flow cytogram of the strain carrying plasmid pFHC1425 grown in glycerol minimal medium (the same sample as shown in Fig. 3E and F and 4).

View larger version (37K): [in this window] [in a new window]   FIG. 6.   Decrease in the origin/mass ratio with increasing load of additional DnaA boxes. The origin/mass ratio was determined as the fluorescence/light scatter ratio (in populations of cells treated with rifampin and cephalexin for a time sufficient to finish ongoing rounds of replication) and expressed relative to the origin/mass ratio obtained in the strain carrying pTAC909 grown in the different growth media.

Additional DnaA boxes disturb the synchrony of initiation. Finally, we studied the effect of the extra DnaA boxes on initiation synchrony by flow cytometry. Figure 7 shows the flow cytometric fluorescence distributions of rifampin-cephalexin-treated samples of strains carrying control plasmid pTAC909, which did not contribute extra DnaA boxes, and plasmids pFHC496 and pFHC1425, which did. These distributions showed a significant increase in initiation asynchrony caused by the plasmids carrying additional DnaA boxes. There was a correlation between the level of derepression of the dnaA promoter and the initiation asynchrony. Thus, the plasmids carrying the most effective sets of oriC DnaA boxes with respect to derepression of the dnaA promoter were also those which affected the initiation asynchrony most. The strain carrying plasmid pFHC1425 exhibits approximately 60% asynchronous cells, in contrast to the strain without a plasmid (or with plasmid pTAC909), which shows only 17% asynchronous cells (Fig. 7). The maximal asynchrony which can be obtained in cells with the number of origins per cell obtained in this experiment is approximately 75%.

View larger version (21K): [in this window] [in a new window]   FIG. 7.   Increase in initiation asynchrony due to the introduction of pBR322-derived plasmids carrying extra DnaA boxes into cells. Origin distributions were obtained as the fluorescence distribution of a population of cells grown in glucose-Casamino Acids medium and treated with rifampin and cephalexin for a time sufficient to finish ongoing rounds of replication. The light gray curve shows the control strain BBC119 carrying plasmid pTAC909. The other two curves demonstrate the increase in initiation asynchrony caused by plasmids which introduce extra DnaA boxes into the cells.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Previous studies using dnaA(Ts) mutants showed that low DnaA protein activity or concentration increased the initiation mass and led to derepression of the dnaA promoter. Here we have studied how extra DnaA protein binding sites affect the initiation mass by using strains containing plasmids carrying DnaA boxes to compete for DnaA protein binding with the normal target DnaA binding sites of the cell. The main finding of this study is that the introduction of extra DnaA protein binding sites leads to an increase in initiation mass. In general, we found proportionality between the number of extra DnaA boxes and the effect on initiation mass. This, in turn, suggests that the DnaA protein binding sites situated on the chromosome are important elements in the setting of the initiation mass through titration of DnaA protein as proposed in our initiator titration model (11).

By using plasmids with mutations in different oriC DnaA boxes, we could show that the previously reported titration of DnaA protein and, thus, derepression of the dnaA gene (12), as well as changes in cell size, observed in this study, were caused by the increased cellular content of DnaA boxes.

The good correlation between the presence of DnaA boxes and cell size, indicating that titration of DnaA protein affected chromosome initiation control, prompted us to use the flow cytometer in a more thorough study to determine the DNA-to-mass ratio, as well as the origin-to-mass ratio, which is inversely proportional to the initiation mass. We varied the cellular DnaA box content by introducing plasmids carrying different numbers (and qualities) of DnaA boxes in our strain and by growing the DnaA box plasmid containing strains in media giving different generation times and, therefore, different plasmid copy numbers. In complete agreement with our expectations, we found that the higher the number of extra DnaA boxes, the greater was the effect on the origin-to-mass ratio and cell size.

We observed increasing asynchrony of initiation in individual cells containing increasing numbers of extra DnaA boxes. In all of these cases, the overall regulation of chromosome replication was relatively unaffected, as the initiation mass was only moderately changed. It has been proposed that when a fast-growing cell starts to initiate chromosome replication, initiation at the first oriC locus will release DnaA protein and thus increase the free DnaA protein concentration, making it more likely to initiate at the next oriC locus, etc. (11, 22). The initiation cascade also works in strains containing minichromosomes, which are present at 5 to 10 copies per chromosomal origin. Minichromosomal origins are initiated at the same time as the chromosomal origins (18) and do not disturb initiation synchrony (21). We suggest that the initiation asynchrony we observed was caused by partial interference with the initiation cascade, because the extra DnaA boxes were accumulated in concert with the replication of the pBR322-derived plasmids; i.e., they were accumulated during the cell cycle in proportion to the mass increase and, thus, independently from any cell cycle-related controls. This might present plasmid DnaA boxes at a time relative to chromosome initiation such that the initiation cascade is disrupted by the plasmid(s) titrating the DnaA protein released from the origins initiated first in the cascade.

The present work clearly shows that the introduction of extra DnaA binding sites from the high-affinity DnaA protein binding region oriC causes a significant increase in initiation mass. Our data complement previous studies in which changes in the intracellular concentration or activity of DnaA protein were shown to change the initiation mass. Recently, it was shown that plasmids carrying the datA locus, one of the other high-affinity DnaA protein binding regions on the chromosome (15), had very similar effects on chromosome initiation control (16). The authors of that report also showed that deletion of the datA locus from the chromosome caused overinitiation, i.e., a decrease in initiation mass.

It should be mentioned that the results from our work and the results of Kitagawa et al. (16) are in full agreement with computer simulations of the initiator titration model (data not shown). We are fully aware that a number of other factors are also actors in the play of initiation of chromosome replication; e.g., there might be factors that alter DnaA protein activity or compete with DnaA protein binding to DnaA boxes. However, it is generally agreed that the DnaA protein is the main actor in initiation. Our data, which show that (extra) DnaA boxes are negatively acting elements which delay initiation, i.e., change the initiation mass, lend support to the initiator titration model which is based on very simple law of mass action considerations and where we postulate that the DnaA boxes in oriC and at other places on the chromosome are the main inhibitory elements defining the initiation mass.