The FIS protein fails to block the binding of DnaA ... - Europe PMC

7 downloads 0 Views 502KB Size Report
ABSTRACT. The Escherichia coli chromosomal origin contains several bindings sites for factor for inversion stimulation. (FIS), a protein originally identified to be ...
 1998 Oxford University Press

5170–5175 Nucleic Acids Research, 1998, Vol. 26, No. 22

The FIS protein fails to block the binding of DnaA protein to oriC, the Escherichia coli chromosomal origin Carla Margulies and Jon M. Kaguni* Department of Biochemistry, Michigan State University, East Lansing, MI 48824-1319, USA Received July 6, 1998; Revised September 16, 1998; Accepted September 25, 1998

ABSTRACT The Escherichia coli chromosomal origin contains several bindings sites for factor for inversion stimulation (FIS), a protein originally identified to be required for DNA inversion by the Hin and Gin recombinases. The primary FIS binding site is close to two central DnaA boxes that are bound by DnaA protein to initiate chromosomal replication. Because of the close proximity of this FIS site to the two DnaA boxes, we performed in situ footprinting with 1,10-phenanthroline-copper of complexes formed with FIS and DnaA protein that were separated by native gel electrophoresis. These studies show that the binding of FIS to the primary FIS site did not block the binding of DnaA protein to DnaA boxes R2 and R3. Also, FIS appeared to be bound more stably to oriC than DnaA protein, as deduced by its reduced rate of dissociation from a restriction fragment containing oriC. Under conditions in which FIS was stably bound to the primary FIS site, it did not inhibit oriC plasmid replication in reconstituted replication systems. Inhibition, observed only at high levels of FIS, was due to absorption by FIS binding of the negative superhelicity of the oriC plasmid that is essential for the initiation process. INTRODUCTION A major component of the bacterial nucleoid (1), factor for inversion stimulation (FIS) is a homodimer of 11.5 kDa subunits. As a DNA binding protein, it stimulates DNA inversion by a family of site-specific recombinases Hin, Gin and Cin (reviewed in 2). FIS also acts as a transcriptional regulator to modulate the expression of several genes (reviewed in 3). Comparative sequence analysis of DNA fragments preferentially bound by FIS reveals that it recognizes a highly degenerate DNA sequence (G/T)NNN(A/G)NN(T/A)NNTNNN(C/A). However, DNA topology is as important as nucleotide sequence in determining binding affinity; FIS not only bends DNA upon binding but also binds to naturally bent DNA (3,4). At a site in the Escherichia coli chromosomal origin (oriC), it induces a bend angle of ∼40 (5). At its respective site in attR, it induces a more severe bend of ∼90 (6). In the E.coli chromosomal origin, oriC, several DNA sequence elements serve as the sites for binding of specific DNA binding proteins (reviewed in 7). Four sites termed DnaA boxes R1, R2, R3 and R4 (Fig. 1; 8) are specifically bound by the DnaA protein. A fifth DnaA box (R5) and sites for the binding of integration host

factor (IHF) and FIS have also been described (9–13). The binding of IHF to the IHF site in oriC is believed to be important for an early step in the initiation of DNA replication. This is because IHF aids in the unwinding by DnaA protein of the AT-rich region near the left border of oriC (14,15). Based on the FIS consensus sequence, two FIS binding sites are predicted between DnaA boxes R2 and R3. One site resembles the FIS consensus sequence and the other is more degenerate (herein cited as the consensus and primary FIS sites, respectively). Two studies relying on DNase I and methidiumpropyl EDTA footprinting demonstrated that FIS bound to the primary site (10,16), but DNase I footprinting in a third report indicated that FIS bound to the FIS consensus sequence (5). A FIS binding site to the right of DnaA box R4 and other weaker sites in the left portion of oriC have also been observed (5,10). On the significance of these observations, a number of studies suggest that FIS is involved in initiation from oriC. First, whereas wild-type E.coli initiates chromosomal replication in a synchronous manner, fis mutants are asynchronous (17), suggesting that FIS normally mediates assembly of the nucleoprotein complex for initiation and that this step is inefficient in fis mutants. Second, fis mutants poorly maintain oriC plasmids at temperatures ≥37C (5,10). Third, site-directed mutagenesis of the primary FIS site between R2 and R3 inactivated oriC function when in a minichromosome (16). Fourth, deletion of DnaA box R4 in the chromosomal oriC locus is tolerated only in the presence of the wild-type fis gene (18). Whereas these findings suggest a positive role for FIS at oriC, several reports support a negative role. One study of in vivo footprinting monitored specific sequences within oriC that were altered as the cell progressed through the cell cycle (19). The FIS site was found to be protected throughout most of the cell cycle. At the time of initiation, the FIS site was no longer bound but instead the adjacent DnaA box R3 became protected, suggesting that the binding of FIS occluded the binding of DnaA protein to R3. This observation is supported by a study of in vitro DNase I footprinting in which the authors concluded that the binding of DnaA protein to DnaA box R3 was mutually exclusive to the binding of FIS to the adjacent consensus FIS site (5). Others have examined the effect of FIS on in vitro replication of oriC plasmids in reconstituted replication systems (20,21). These investigators failed to detect a positive role for FIS in replication from oriC. Instead, FIS was inhibitory. In one report, FIS was concluded to block opening of the AT-rich region in oriC by DnaA protein but it was not clear if this was due to the binding of FIS to oriC (20). In the second, inhibition was observed even when the primary FIS site in oriC was mutated (21). The mutant oriC was shown to have

*To whom correspondence should be addressed. Tel: +1 517 353 6721; Fax: +1 517 353 9334; Email: [email protected]

5171 Nucleic Acids Acids Research, Research,1994, 1998,Vol. Vol.22, 26,No. No.122 Nucleic

5171

Figure 1. The oriC region. Escherichia coli oriC (258 bp) is indicated by the open box. The positions of DnaA boxes R1, R3, R3 and R4 (filled triangles), 13mers near the left end of oriC (cross-hatched boxes) and FIS binding sites (stippled boxes) based on the FIS consensus sequence (3) are also shown. The outlined stippled box between DnaA boxes R2 and R3 bound by FIS as observed by Gille et al. (5) are shown as a filled box. The black box represents the FIS site seen by Filutowicz et al. (10).

reduced but measurable FIS binding, leaving open the possibility that FIS bound to oriC accounted for the inhibition. In both reports, inhibition of DNA synthesis required high levels of FIS. We previously found that DnaA protein binds to DnaA boxes in oriC in an ordered manner (22). Optimal replication activity required its binding to all four DnaA boxes. Because FIS may hinder the binding of DnaA protein to R3 to modulate the initiation process, we investigated the effect of FIS on DnaA protein binding and its function in DNA replication by assays of gel mobility shift, chemical footprinting and reconstituted oriC plasmid replication. Our results indicate that FIS does not occlude the binding of DnaA protein to DnaA box R3 of oriC and that the replication inhibition caused by the binding of FIS at high levels is due to absorption of the negative superhelicity in the oriC plasmid that is essential for the initiation process. These results also suggest that FIS bound to the primary FIS site in oriC has no influence on the initiation process. MATERIALS AND METHODS Proteins and plasmids FIS protein was a generous gift of Dr Reid Johnson (Department of Biological Chemistry, UCLA, Los Angeles, CA). Other proteins and plasmid DNA were described elsewhere (23). Assays Gel mobility shift assays and in situ cleavage with 1,10-phenanthroline-copper and reconstituted replication assays lacking RNA polymerase, RNase H and topoisomerase I were performed as described (23). Supercoiling assays (20 µl) contained 90 fmol of supercoiled pBSoriC DNA and the indicated amounts of FIS in 20 mM Tris–HCl (pH 7.5), 10 mM MgCl2, 2 mM DTT, 5% glycerol and 5 mM NaCl. Assays were incubated for 15 min at 30C followed by addition of E.coli topoisomerase I (15 U) and incubation for 30 min at 30C. To each tube, 5 µl of 30% glycerol, 10 mM EDTA and 6% SDS were added and then incubated at 85C for 3 min prior to electrophoresis in a 0.8% agarose gel containing 1 µg/ml chloroquine. Quantitative analysis of gels is described in the respective figure legends. RESULTS FIS forms multiple complexes with a DNA fragment containing oriC and binds to several sites within it Within oriC, two overlapping FIS sites are predicted between DnaA boxes R2 and R3 based on the FIS consensus sequence (Fig. 1). DNase I footprinting studies showed that FIS bound to

Figure 2. FIS bound to oriC forms multiple complexes that are resolved by native polyacrylamide gel electrophoresis. Gel mobility shift assays were performed as described (22) with the indicated amounts of FIS protein and 25 fmol of a 32P-3′-end-labeled SmaI–XhoI oriC fragment. Unbound DNA is noted as Free.

these sites (5,10,16). Additional sites have been observed to the right of DnaA box R4 and in the left portion of oriC. Because no direct evidence has correlated the binding of FIS at oriC to its inhibitory effect on in vitro oriC plasmid replication (20,21), we sought to establish this link. Under this aim, one issue we addressed was whether FIS could bind to the overlapping FIS sites independently. If so, these different complexes may not have been discerned by DNase I footprinting of complexes analyzed in solution, but may be separable by gel mobility shift analysis and distinguished by in situ footprinting with 1,10-phenanthrolinecopper. Indeed, previous reports using gel mobility shift assays showed that FIS forms multiple complexes when bound to an oriC fragment (5,20), At low FIS levels, several discrete complexes were resolved (Fig. 2). At higher FIS levels (≥16 ng), the FIS-bound complex migrated progressively nearer to the sample wells, suggesting non-specific binding. In situ cleavage of the discrete complexes with 1,10-phenanthroline-copper was performed in concert with quantitative scanning densitometry of the resulting autoradiogram to determine the site(s) of FIS binding (Fig. 3). From this analysis, complex 1 is formed by the binding of FIS to the primary site between DnaA boxes R2 and R3, in agreement with previous reports (10,16). An important characteristic of the FIS protection pattern at low FIS levels is the lesser protection near the right FIS boundary (Fig. 3B). The protected site is not the FIS consensus sequence observed to be bound by Gille et al. that overlaps DnaA box R3 (5). This result also indicates that FIS binds to the primary FIS site with greater affinity than to other sites in oriC. Compared with complex 1, we attribute the lower mobility of complex 2 (and other complexes) to the added binding of FIS to DNA near the AccI site. This region includes a FIS consensus sequence (Fig. 1). Complexes 3–5 differed from complexes 1 and 2 by slight protection of DnaA box R2. Because it seems unlikely that the modest reduction in chemical cleavage at DnaA box R2 reflects the binding of FIS to account for the retarded mobility of these complexes, they may have arisen either by an induced topological change in the DNA fragment and/or by the binding of additional FIS molecules to those already bound. At higher FIS levels, additional more slowly migrating complexes (designated complexes 6–8) were observed. By in situ footprint analysis with 1,10-phenanthroline-copper,

5172 Nucleic Acids Research, 1998, Vol. 26, No. 22

A

B

Figure 3. In situ footprinting of FIS bound to oriC with 1,10-phenanthroline-copper. In situ cleavage of electrophoretically separated FIS–oriC complexes (Fig. 2) was performed as described with a SmaI–XhoI oriC fragment that was 3′-end-labeled at the XhoI site (22). The DNA was then recovered from gel slices and separated by electrophoresis on a 6% sequencing gel. (A) The positions of the DnaA boxes and the site bound by FIS are indicated at the left of the autoradiogram. (B) The cleavage pattern of each complex was compared with the cleavage pattern of unbound (free) DNA, analyzed by quantitative β emission scanning with a Molecular Dynamics PhosphorImager.

these differed from the complexes described above by increased protection of DnaA boxes R2 and R3, the region to the right of the AccI site and DNA between DnaA boxes R1 and R2 (Fig. 3B). FIS inhibits oriC plasmid replication only when added at high levels The above results suggest that FIS may affect initiation from oriC by influencing the binding of DnaA protein. To address this possibility, the effect of FIS on replication of an oriC plasmid in a reconstituted replication system was examined. At the same molar amount of oriC in a supercoiled plasmid as in a DNA fragment in Figures 2 and 3, the addition of low levels of FIS that resulted in discrete complexes (Figs 2 and 3) had a negligible effect (Fig. 4). It was inhibitory only when added at much higher levels, confirming previous observations (20,21). Under similar reaction conditions, vector sequences (pBluescript) did not influence the binding of FIS to oriC. This was determined in a gel mobility shift assay by the addition of unlabeled pBluescript as a

competitor to reactions containing a radiolabeled oriC fragment and FIS (data not shown). Hiasa and Marians (20) observed that high levels of FIS inhibited replication of an oriC plasmid in the absence of IHF and HU protein. The presence of either IHF or HU counteracted the inhibition. We examined the effect of FIS on oriC plasmid replication in the absence of HU or with it added at low, intermediate or high levels (Fig. 5). HU did not alleviate FIS inhibition when added at high levels, nor was FIS stimulatory at any level. IHF can functionally replace HU protein in oriC plasmid replication (24). Comparable experiments with IHF showed that FIS was inhibitory only when added at high levels (data not shown). Under no condition was FIS stimulatory. The gidA and mioC genes flank oriC. Transcription from their promoters has been proposed to influence the replication efficiency of oriC (25). Also, promoters in or near oriC have been described (7). As FIS modulates the expression of several genes as a transcription factor (reviewed in 3), its affect on initiation from oriC may be at the level of transcription. Consequently, the effect of FIS in a reconstituted replication system supplemented with

5173 Nucleic Acids Acids Research, Research,1994, 1998,Vol. Vol.22, 26,No. No.122 Nucleic

Figure 4. FIS inhibits in vitro oriC plasmid replication only at high levels. Reconstituted replication assays lacking RNA polymerase, RNase H and topoisomerase I were performed with pBSoriC as a template as described (23). FIS was added prior to DnaA protein. The panel inset indicates the amount of DNA synthesis at low FIS levels.

Figure 5. FIS inhibition of oriC plasmid replication is not influenced by the level of HU. Reconstituted replication assays were performed as described in the legend to Figure 4 with the indicated amounts of HU. FIS was added prior to DnaA protein.

RNA polymerase was examined. As in the above experiments, FIS did not stimulate DNA synthesis (data not shown). It was inhibitory only at high levels and independent of the inclusion of either HU or IHF. Binding of FIS and DnaA protein are not mutually exclusive In synchronized cells, in vivo footprinting of oriC revealed that the primary FIS site between DnaA boxes R2 and R3 was protected throughout most of the bacterial cell cycle (19). Just prior to initiation, the FIS site was no longer protected and DnaA box R3 was bound. This observation and others (22) suggest that the binding of DnaA protein to DnaA box R3 is critical for the initiation process and that FIS might play a regulatory role. In addition, the proximity of the FIS site to DnaA boxes R2 and R3 and the report of mutually exclusive binding of FIS and DnaA protein to DnaA box R3 suggest a negative role for FIS (5,19). We performed in situ footprinting of electrophoretically separated complexes to determine if the binding of FIS blocks the binding of DnaA protein to DnaA box R3. The analysis relied on the characteristic footprint pattern of respective proteins to determine the composition of different complexes. At a level of FIS just sufficient to bind to the primary FIS site (FIS:oriC ratios of 1:1 and 3:1; Fig. 6, lanes 2 and 3), the addition of DnaA protein resulted in formation of novel complexes not seen with either protein alone (Fig. 6). At these low FIS levels, the complexes obtained with FIS only (complexes 1 and 2 and trace amounts of complex 3) did not migrate to the region of the gel containing the novel complexes and, hence, did not interfere with the in situ

5173

Figure 6. Novel complexes form on an oriC fragment by addition of both DnaA protein and FIS. Gel mobility shift assays with DnaA protein, FIS or both together were performed essentially as described (22) with a SmaI–XhoI restriction fragment labeled at the XhoI site with the large fragment of DNA polymerase I and all four [α-32P]dNTPs. Reactions were incubated for 30 min at 25C before electrophoresis on a native polyacrylamide gel.

analysis. Chemical cleavage of complex C and its quantitative analysis revealed that it was formed by the co-binding of FIS and DnaA to R3 as well as to other DnaA boxes (Fig. 7). In this complex, we deduced that FIS was bound to the primary FIS site because of the lesser protection of sequences near the right border of the FIS site. This was a characteristic we noted in the in situ analysis of complexes formed with FIS only (Fig. 3). In comparison, complex A was formed by the binding of DnaA to DnaA boxes R1 and R4 and Fis to the primary FIS site. As controls, complexes formed with FIS or DnaA only showed protection of the expected sites (DnaA complexes II and V and FIS complex 1 of Fig. 7). These results indicate that the binding of DnaA protein to DnaA box R3 and FIS to the adjacent site are not mutually exclusive. FIS dissociates from oriC more slowly than DnaA protein The observations described above appear to contradict a previous study reporting mutually exclusive binding of FIS and DnaA that was based on DNase I footprinting (5). We wondered if the relative dissociation rates of FIS and DnaA bound to an oriC fragment may have affected the DNase I footprinting results. If the FIS complex dissociates more slowly than DnaA protein bound to oriC, it may appear as though FIS blocks the binding of DnaA protein to R3. To explore this possibility, we compared the dissociation rates of FIS bound to oriC with DnaA protein–oriC complexes (Fig. 8). The level of FIS used was sufficient to form two predominant complexes resulting from its binding to the primary FIS site and to a site near the AccI restriction site and a third less abundant product called complex 3 (Figs 2 and 3). For DnaA protein, the level added formed a slowly migrating complex (complex VI) that is formed from the ordered binding of DnaA protein to the DnaA boxes in oriC with DnaA box R3 bound last (22). These complexes were then challenged with a 100-fold molar excess of unlabeled oriC plasmid. At the indicated times, aliquots were removed and free DNA was separated from protein–DNA complexes by electrophoresis in a native polyacrylamide gel. By quantitative densitometry of the autoradiogram, the amount of free DNA at each time point compared with the total DNA was measured to calculate the fraction of DNA remaining in complexes. On addition of the unlabeled oriC plasmid, most of

5174 Nucleic Acids Research, 1998, Vol. 26, No. 22

Figure 7. In situ footprinting with 1,10-phenanthroline-copper of complexes formed by the binding of FIS and DnaA at oriC. In situ cleavage of electrophoretically separated complexes was performed as described with a SmaI–XhoI restriction fragment that was 3′-end-labeled at the XhoI site (22). Complexes of FIS were obtained at molar ratios of FIS dimer to oriC of 1:1 and 3:1 to form complexes 1 and 2 and trace amounts of complex 3 (Fig. 6). The electrophoretic mobilities of these complexes were to positions in the gel that did not interfere with the in situ analysis of the novel complexes nor complicate the interpretation of the data. The molar ratio of DnaA protein to oriC was 10:1 as in Figure 6. Following electrophoresis on a 6% sequencing gel of the cleaved DNA, detection and quantitative β emission scanning was as described in Figure 3.

the radiolabeled oriC fragment bound by DnaA protein was released at the earliest time point (30 s). At a ratio of 1.5 molecules of FIS to oriC fragment, the addition of the unlabeled competitor did not result in the dissociation of FIS from the radiolabeled oriC fragment until 5–10 min had elapsed. These results suggest that FIS dissociates from its respective binding sites including the primary FIS site at an ∼10-fold slower rate than DnaA protein. This result may account for the appearance that FIS blocked the binding of DnaA protein to DnaA box R3 (5). High levels of FIS alter the topology of oriC In vitro replication of oriC-containing plasmids requires a negatively supercoiled template (26). We investigated the biochemical basis of FIS inhibition of oriC plasmid replication when added at high levels, suspecting that its binding to DNA non-specifically or to more degenerate FIS sites outside of oriC absorbs the negative superhelicity of the oriC plasmid. To demonstrate this, FIS was incubated with the oriC plasmid, then

Figure 8. Complexes of FIS bound to oriC are much more stable than DnaA protein–oriC complexes. The volume of the gel mobility shift reaction (normally 10 µl; 22) was scaled up 10-fold. (A) DnaA protein (390 ng) was added at a molar ratio of 30:1 to the 32P-labeled SmaI–XhoI DNA fragment containing oriC and incubation was for 5 min at 20C prior to the addition of unlabeled pBSoriC plasmid at a 100-fold molar excess to the radiolabeled oriC fragment. (B) FIS was included at a protein:oriC fragment ratio of 1.5, then incubated for 90 min at 20C before the unlabeled pBSoriC plasmid was added (100-fold molar excess). In lane 2 of (B), the challenge DNA was added prior to FIS addition. In (A) and (B), portions (10 µl) were removed at the times indicated after addition of pBSoriC and applied to a native polyacrylamide gel that had the current on. The retardation of the unbound (free) DNA with increasing time after addition of the competitor DNA reflects the time difference between samples. (C) The amount of unbound (free) DNA was quantified by β emission scanning, then plotted as a ratio of the amount of unbound DNA to the total DNA in each lane.

the sample was treated with E.coli topoisomerase I to stabilize the topological alteration induced by FIS binding followed by agarose gel electrophoresis in the presence of chloroquine. At levels of FIS that inhibited oriC plasmid replication (Figs 4 and 5), the DNA was more negatively supercoiled than the mock-treated plasmid (Fig. 9). This suggests that the inhibitory effect of FIS at high levels is due to its binding non-specifically to DNA to sequester the negative superhelicity essential for initiation from oriC. DISCUSSION Physiological studies that suggest a positive role for FIS in the initiation process at oriC (10,16,17) contrast with biochemical studies of FIS inhibition of oriC plasmid replication (20,21). In addition, others described that the binding of FIS to oriC blocked

5175 Nucleic Acids Acids Research, Research,1994, 1998,Vol. Vol.22, 26,No. No.122 Nucleic

5175

to bind to DNA at 160 bp intervals assuming two chromosomes per cell. This suggests that FIS acts as a negative regulator of initiation at oriC, but only in cells in early log phase or in response to changes in culture conditions. ACKNOWLEDGEMENTS

Figure 9. FIS at high levels sequesters the negative superhelicity of pBSoriC. Where indicated, FIS was incubated with 90 fmol of supercoiled pBSoriC at the listed molar ratios. Incubation was at 30C for 15 min in 20 mM Tris–HCl, pH 7.5, 10 mM MgCl2, 2 mM DTT, 5% glycerol and 5 mM NaCl in a 20 µl reaction volume. Escherichia coli topoisomerase I (15 U) was then added to fix the change in topology induced upon FIS binding. Reactions were then incubated for 30 min at 30C, then stopped by addition of 5 µl of 30% glycerol, 10 mM EDTA and 6% SDS followed by incubation at 85C for 3 min that was sufficient to remove proteins bound to DNA. Electrophoresis was in a 0.8% agarose gel containing 1 µg/ml chloroquine phosphate. These conditions separate relaxed from negatively supercoiled pBSoriC isolated from E.coli. Note that in lanes 8 and 9, the binding of FIS results in even more negatively supercoiled pBSoriC than the topological forms isolated from E.coli. Lane 10 is the negative control of untreated plasmid.

the binding of DnaA protein to DnaA box R3 (5) and that the primary FIS site is occupied throughout most of the cell cycle except at the time of initiation (19). DnaA box R3 was occupied instead. For these reasons, we performed experiments to explore the effect of FIS on oriC plasmid replication. We found that FIS formed multiple complexes with a DNA fragment containing oriC by gel mobility shift analysis. In these complexes, FIS bound to the primary FIS site between DnaA boxes R2 and R3 to confirm previous observations (10,21). At low levels of FIS, its binding to this site did not inhibit oriC plasmid replication under a variety of conditions (varying HU levels, reconstituted oriC plasmid replication with or without RNA polymerase), nor did it block the binding of DnaA protein to DnaA box R3. We also found that FIS–oriC complexes dissociated more slowly than DnaA–oriC complexes. Lastly, high levels of FIS inhibited oriC plasmid replication by sequestering the negative superhelicity of the oriC plasmid. The failure to see a stimulatory effect of FIS on DNA replication in vitro may be due to the absence of a factor critical for FIS function. Alternatively, the effect of FIS may be indirect and involve its ability to alter DNA topology. Indeed, fis mutants have aberrant nucleoids leading to the proposal that FIS may act to organize the bacterial chromosome (10). The FIS effect may also be due to its role as a transcriptional regulator. In fis null mutants that show a replication asynchrony phenotype (17), expression of the dnaA gene is elevated 2–3-fold relative to wild-type strains (27). In other studies, a 1.5–3-fold increase in dnaA expression stimulated initiation at oriC but the mode of DNA replication that followed was aberrant (28,29). Finally, expression of FIS protein is growth rate dependent (30,31). FIS concentrations increase rapidly when cells move from stationary to log phase growth and decrease through mid to late log phase. It is speculated that FIS acts to couple growth rate to changing nutritional and environmental conditions. In early log phase, the level of FIS (5–10 × 104 molecules/cell) is sufficient

We thank Dr Reid Johnson for the gift of FIS protein. This research was supported by grant GM33992 from the National Institutes of Health. Acknowledgment is also made to the Michigan Agricultural Experiment Station for its support of this research.

REFERENCES 1 Pettijohn,D.E. (1988) J. Biol. Chem., 263, 12793–12796. 2 Nash,H.A. (1996) In Neidhardt,F.C., Ingraham,J.L., Low,K.B., Magasanik,B., Schaechter,M. and Umbarger,H.E. (eds), Escherichia coli and Salmonella typhimurium. Cellular and Molecular Biology, 2nd Edn. American Society for Microbiology, Washington, DC, Vol. 2, pp. 2363–2376. 3 Finkel,S.E. and Johnson,R.C. (1992) Mol. Microbiol., 6, 3257–3265. 4 Bailly,C., Waring,M.J. and Travers,A.A. (1995) J. Mol. Biol., 253, 1–7. 5 Gille,H., Egan,J.B., Roth,A. and Messer,W. (1991) Nucleic Acids Res., 19, 4167–4172. 6 Thompson,J.F. and Landy,A. (1988) Nucleic Acids Res., 16, 9687–9705. 7 Messer,W. and Weigel,C. (1996) In Neidhardt,F.C., Ingraham,J.L., Low,K.B., Magasanik,B., Schaechter,M. and Umbarger,H.E. (eds), Escherichia coli and Salmonella typhimurium. Cellular and Molecular Biology, 2nd Edn. American Society for Microbiology, Washington, DC, Vol. 2, pp. 1579–1601. 8 Fuller,R.S., Funnell,B.E. and Kornberg,A. (1984) Cell, 38, 889–900. 9 Matsui,M., Oka,A., Takanami,M., Yasuda,S. and Hirota,Y. (1985) J. Mol. Biol., 184, 529–533. 10 Filutowicz,M., Ross,W., Wild,J. and Gourse,R.L. (1992) J. Bacteriol., 174, 398–407. 11 Messer,W., Egan,B., Gille,H., Holz,A., Schaefer,C. and Woelker,B. (1991) Res. Microbiol., 142, 119–125. 12 Filutowicz,M. and Roll,J. (1990) New Biol., 2, 818–827. 13 Polaczek,P. (1990) New Biol., 2, 265–171. 14 Bramhill,D. and Kornberg,A. (1988) Cell, 52, 743–755. 15 Hwang,D.S. and Kornberg,A. (1992) J. Biol. Chem., 267, 23083–23086. 16 Roth,A., Urmoneit,B. and Messer,W. (1994) Biochimie, 76, 917–923. 17 Boye,E., Lyngstadaas,A., Loebner-Olesen,A., Skarstad,K. and Wold,S. (1993) In Fanning,E., Knippers,R. and Winnacker,E.L. (eds), DNA Replication and the Cell Cycle. Springer, Mosbacher Kolloquium, Springer, Berlin. Vol. 43, pp. 15–26. 18 Bates,D.B., Asai,T., Cao,Y., Chambers,M.W., Cadwell,G.W., Boye,E. and Kogoma,T. (1995) Nucleic Acids Res., 23, 3119–3125. 19 Cassler,M.R., Grimwade,J.E. and Leonard,A.C. (1995) EMBO J., 14, 5833–5841. 20 Hiasa,H. and Marians,K.J. (1994) J. Biol. Chem., 269, 24999–25003. 21 Wold,S., Crooke,E. and Skarstad,K. (1996) Nucleic Acids Res., 24, 3527–3532. 22 Margulies,C. and Kaguni,J.M. (1996) J. Biol. Chem., 271, 17035–17040. 23 Hwang,D.S. and Kaguni,J.M. (1988) J. Biol. Chem., 263, 10633–10640. 24 Skarstad,K., Baker,T.A. and Kornberg,A. (1990) EMBO J., 9, 2341–2348. 25 Asai,T., Chen,C.P., Nagata,T., Takanami,M. and Imai,M. (1992) Mol. Gen. Genet., 231, 169–178. 26 Funnell,B.E., Baker,T.A. and Kornberg,A. (1986) J. Biol. Chem., 261, 5616–5624. 27 Froelich,J.M., Phuong,T.K. and Zyskind,J.W. (1996) J. Bacteriol., 178, 6006–6012. 28 Atlung,T. and Hansen,F.G. (1993) J. Bacteriol., 175, 6537–6545. 29 Lobner-Olesen,A., Skarstad,K., Hansen,F.G., von Meyenburg,K. and Boye,E. (1989) Cell, 57, 881–889. 30 Ball,C.A., Osuna,R., Ferguson,K.C. and Johnson,R.C. (1992) J. Bacteriol., 174, 8043–8056. 31 Thompson,J.F., Moitoso de Vargas,L., Koch,C., Kahmann,R. and Landy,A. (1987) Cell, 50, 901–908.