SeqA protein stimulates the relaxing and decatenating activities of topoisomerase IV.

The SeqA protein, which prevents overinitiation of chromosome replication, has been suggested to also participate in the segregation of chromosomes in Escherichia coli. Using a bacterial two-hybrid system, we found that SeqA interacts with the ParC subunit of topoisomerase IV (topo IV), a type II topoisomerase involved in decatenation of daughter chromosomes and relief of topological constraints generated by replication and transcription. We demonstrated that purified SeqA protein stimulates the activities of topo IV, both in relaxing supercoiled plasmid DNA and converting catenanes to monomers. The same moderate levels of SeqA protein did not affect the activities of DNA gyrase or topoisomerase I. At higher levels of SeqA, topo IV favored the formation of catenanes, caused by intermolecular strand exchange among plasmid DNA aggregates formed by SeqA. Excess SeqA inhibited the activity of all topoisomerases. We also found that stimulation of topo IV was dependent upon the affinity of SeqA for DNA. Our results suggest that this stimulation is mediated by the specific interaction of topo IV with SeqA. Some of the known phenotypes of mutant cells lacking SeqA, such as deficient chromosome segregation and increased negative superhelicity, support that the SeqA protein is required for topo IV-mediated relaxation and decatenation of chromosomes and plasmids, during and after their replication.

The SeqA protein, which prevents overinitiation of chromosome replication, has been suggested to also participate in the segregation of chromosomes in Escherichia coli. Using a bacterial two-hybrid system, we found that SeqA interacts with the ParC subunit of topoisomerase IV (topo IV), a type II topoisomerase involved in decatenation of daughter chromosomes and relief of topological constraints generated by replication and transcription. We demonstrated that purified SeqA protein stimulates the activities of topo IV, both in relaxing supercoiled plasmid DNA and converting catenanes to monomers. The same moderate levels of SeqA protein did not affect the activities of DNA gyrase or topoisomerase I. At higher levels of SeqA, topo IV favored the formation of catenanes, caused by intermolecular strand exchange among plasmid DNA aggregates formed by SeqA. Excess SeqA inhibited the activity of all topoisomerases. We also found that stimulation of topo IV was dependent upon the affinity of SeqA for DNA. Our results suggest that this stimulation is mediated by the specific interaction of topo IV with SeqA. Some of the known phenotypes of mutant cells lacking SeqA, such as deficient chromosome segregation and increased negative superhelicity, support that the SeqA protein is required for topo IV-mediated relaxation and decatenation of chromosomes and plasmids, during and after their replication.
Besides its role as a negative modulator of chromosome rep-lication, the SeqA protein most likely functions in segregation of replicated DNA. Strains with seqA mutations exhibit aberrant nucleoid distribution, a higher frequency of anucleoid cells, and filamentation (11)(12)(13)(14). Microscopy of immunolabeled or green fluorescent protein-tagged SeqA has revealed that SeqA is predominantly localized in foci situated at the replication forks, presumably forming complexes with the newly replicated DNA and possibly contributing to proper segregation of daughter chromosomes (4,11,15,16). Overproduction of the SeqA protein interferes with the segregation of replicated chromosomes and leads to delayed cell division (14,15). These results indicate that an optimal level of SeqA is required for the segregation of chromosomal DNA. In addition to the above mentioned phenotypes, the seqA mutant exhibits increased negative superhelicity of chromosomal and plasmid DNA as well as multimerization of plasmids (13,17). Negative superhelicity of DNA is necessary for both DNA replication and transcription and is essential for cell viability; therefore, negative superhelicity is tightly maintained (review in Refs. 18 -21). In E. coli, superhelicity is maintained through the balanced actions of topoisomerase I (topo I), 1 DNA gyrase, and topoisomerase IV. Whereas DNA gyrase introduces negative supercoiling, topo I relaxes negative supercoils, and topo IV relaxes both positive and negative supercoils. DNA gyrase and topo IV are type II topoisomerases that change the linking number in steps of two and require ATP binding and hydrolysis for their functions. These topoisomerases introduce a transient double strand break, pass a second duplex segment through the break, and then ligate the break. The parC and parE genes encoding the subunits of topo IV are required for segregation of daughter chromosomes (22)(23)(24)(25). The role of topo IV is to decatenate the replicated chromosome and plasmid DNA prior to cell division (26 -29). Because topo IV is essential for cell viability, the in vivo function of topo IV has been addressed by using conditional mutants or inhibitors. Inhibition of topo IV results in increased negative superhelicity, equivalent to that of a topo I mutant, suggesting that the relaxation activities of both topo I and IV contribute to the maintenance of a certain level of negative supercoiling (27, 30 -32). Together with DNA gyrase, topo IV participates in removal of the positive supercoils that accumulate ahead of replication forks and RNA polymerase transcription (30,33).
In this report, we demonstrate that the SeqA protein stimulates the relaxation and decatenation activities of topo IV. The stimulation of topo IV by SeqA depends on the binding affinity of SeqA for DNA, suggesting that the binding of SeqA to DNA is involved in the stimulation.
Proteins-Topo I, the GyrA and GyrB subunits of DNA gyrase, and the ParC and ParE subunits of topo IV were overexpressed in pBAD expression vectors (35). Overexpressed proteins were purified as described previously (26,36,37). Subunits of gyrase and topo IV were mixed at equimolar concentrations then incubated on ice for 15 min to allow for holoenzyme assembly before assaying. SeqA purification has been previously described (3). Protein concentrations used (unless otherwise indicated) are as follows: SeqA, 1.4 g/l; DNA gyrase, 0.6 g/l; topo I, 0.24 g/l; and topo IV, 0.5 g/l.
Bacterial Two-hybrid Screening-Bacterial two-hybrid screening with the adenylate cyclase protein of Bordetella pertussis has been previously described (5,38). The DHP1 strain (derivative of DH1, cya) was used for detecting protein-protein interactions. The coding region of seqA was inserted into the KpnI/HindIII sites of the pT18 plasmid and used as bait for screening. For constructing a prey library, 0.5-to 3-kb fragments of Sau3AI partial digests of E. coli genomic DNA were ligated to the BamHI site of the pT25 plasmid; the library diversity was determined to be over 10 6 . The pT25 genomic DNA library was electroporated into the DHP1 strain harboring pT18-seqA and plated onto LB-agar plates containing 50 g/ml ampicillin, 34 g/ml chloramphenicol, 10 g/ml phosphomycin, 40 g/ml 5-bromo-4-chloro-3-indolyl-␤-Dgalactopyranoside (X-gal), and 0.4 mM isopropyl-1-thio-␤-D-galactopyranoside. Blue colonies were isolated, re-streaked onto plates of the same composition, and incubated at 37°C. Single blue colonies were again selected and cultured in 3 ml of LB containing 34 g/ml chloramphenicol to isolate prey DNA. ␤-Galactosidase activities were analyzed as described previously (39).
Generation of Un-, Hemi-, and Fully Methylated Plasmids-The Dam methylation reaction has been previously described (3). For obtaining hemi-methylated DNA, 10 g of unmethylated pBS or pFToriC plasmids isolated from E. coli strain GM3819 was added to 200 l of Dam assay solution (0.1 M Tris-HCl at pH 8.0, 10 mM EDTA, 2.5 mM dithiothreitol, 1.6 M S-adenosylmethionine) containing 0.5 g/l SeqA. Dam methylase (4 g) was then added, and the mixture was incubated at 37°C for 5 min. The DNA was recovered by phenol/chloroform extraction and ethanol precipitation. Fully methylated plasmids were obtained by omitting SeqA from the Dam assay solution and extending the reaction time to 30 min.
Topoisomerase Assays-For topoisomerase assays, the 10-l reactions contained 50 mM Tris⅐HCl at pH 7.5, 10 mM MgCl 2 , 0.1 mM EDTA, 70 mM KCl, 1 mM ATP, 5 g of bovine serum albumin, 1 mM dithiothreitol, 5% glycerol, and 10 ng of indicated DNA. Following addition of the topoisomerase, the reaction was incubated at 37°C for 10 min unless indicated. The reaction was stopped by adding an equal volume of stop solution (3% SDS, 20 mM EDTA, 15% glycerol). Products were separated through 1% agarose at 20 V for 17-20 h, transferred to nitrocellulose membranes for Southern hybridization, and visualized by autoradiogram. For quantification, ImageQuaNT software (Amersham Biosciences) was used.
Relaxed plasmids for the DNA gyrase supercoiling assay were generated by using calf thymus topoisomerase I (Invitrogen). Relaxed plasmids were isolated by phenol/chloroform extraction and ethanol precipitation.
Preparation of Catenanes-The topo IV relaxation mixture (200 l) contained 2 g of hemi-methylated pFToriC, 3.6 g of SeqA protein, and 200 ng of topo IV. The reaction was stopped by adding SDS and EDTA stop solution. DNA products were isolated by phenol/chloroform extraction and ethanol precipitation.

Interaction of SeqA with the C-terminal Domain of the ParC
Subunit of Topo IV-Interaction between proteins can be detected with a bacterial two-hybrid system by using the functional complementation of the N (T25)-and C (T18)-terminal regions of Bordetella pertussis adenylate cyclase (5,38). SeqA proteins cooperatively bind to DNA and form aggregates (5,7). When the SeqA protein is fused to both T25 and T18, SeqA-SeqA interaction allows the reconstitution of active adenylate cyclase and results in reporter (␤-galactosidase) activity 4-fold higher than the activity of the control (pT18-seqA/pT25) ( Table  I) (5). To identify the one or more proteins that interact with the SeqA protein, we used SeqA-T18 expressed from the pT18-seqA plasmid as bait. After screening an E. coli genomic library cloned into the pT25 plasmid, we identified two independent SeqA-interacting clones expressing the C-terminal regions of the ParC subunit of topo IV that interacted with SeqA (Table I). These two clones exhibited 2.3-and 2.5-fold higher ␤-galactosidase activities than the activity for pT25 alone.
SeqA Stimulates the Relaxation Activity of Topo IV-Topo IV relaxes supercoils in DNA (18 -20, 40). To determine the effect of SeqA on the relaxation activity of topo IV, we examined the relaxation activity by using hemi-methylated, negatively supercoiled pBS (Fig. 1A). The unmethylated DNAs, which were used to generate hemi-and fully methylated DNAs in vitro (see "Experimental Procedures"), were isolated from dam mutant cells and therefore contained higher amounts of multimerized DNAs than DNAs obtained from wild-type cells. The relaxation reaction mixtures containing topo IV and hemi-methylated pBS were incubated for the times indicated and were then quenched with stop solution containing EDTA and SDS. The resulting topoisomers were separated by agarose gel electrophoresis and detected by Southern blot hybridization (Fig. 1A). In comparison with the absence of SeqA, the presence of SeqA facilitated topo IV-mediated conversion of supercoiled plasmid to relaxed plasmid DNA. The amount of SeqA required to facilitate topo IV activity was determined by titrating SeqA into a topo IV reaction (Fig. 1B). With increasing amounts of SeqA (up to 10 nM tetramer, or the equivalent to 8 ng), the formation of relaxed DNAs was enhanced. Addition of excess SeqA (equal to or greater than 20 nM) inhibited this stimulation. In the absence of topo IV, SeqA did not alter supercoiled or relaxed DNAs (Fig. 1C). These results show that moderate levels of SeqA stimulate the relaxation activity of topo IV.
SeqA Does Not Stimulate the Activity of Topo I or DNA Gyrase-To determine whether SeqA activates all topoisomerases or topo IV exclusively, we examined the effects of SeqA on the other E. coli topoisomerases, topo I, and DNA gyrase (Fig.  2). For the topo I assay, the reaction mixtures contained supercoiled, hemi-methylated plasmid DNA. The same procedure was used for the DNA gyrase assay, but relaxed, hemi-methylated plasmid was used instead of supercoiled plasmid. The relaxation activity of topo I ( Fig. 2A) and the supercoiling activity of DNA gyrase (Fig. 2B) were not significantly affected with up to 5 nM SeqA. As seen with topo IV, excess SeqA inhibited both DNA gyrase and topo I. This inhibition of topo I agrees with previous studies (41). Thus, stimulation by SeqA is unique to topo IV and does not occur with the other topoisomerases.
SeqA Allows Topo IV to Interlink Plasmid DNAs-In the presence of SeqA, topo IV produced slowly migrating DNAs (SMDs), which were observed in the upper part of gels near the wells (Fig. 1). In contrast, addition of SeqA to either topo I or DNA gyrase reactions did not result in the production of SMDs (Fig. 2). The formation of SMDs was confirmed with hemimethylated pFToriC plasmid DNA, which contains the GATC abundant oriC region cloned into the pBS vector (data not shown). Under the reaction conditions described in Fig. 1, SeqA also stimulated the relaxation of this plasmid by topo IV. As with the relaxation activity of topo IV, the accumulation of SMDs of pFToriC occurred in a SeqA concentration-dependent manner. Topo IV converted negatively supercoiled plasmids to relaxed plasmids at lower concentrations of SeqA and to SMDs at higher concentrations. SMD formation appeared to occur more readily with pFToriC containing the abundant GATC sequences than with pBS. This preference of pFToriC caused that topo IV produced more SMDs than relaxed DNAs (Fig. 3).
It is reasonable to suppose that SMDs are high molecular weight forms of DNA. Topo IV, a type II topoisomerase, cleaves both strands on duplex DNA, transiently generating 5Ј-ends that are covalently linked to the tyrosine residue of the enzyme (19,20,42). This covalent linkage is required for strand passage during relaxation and decatenation. These properties of topo IV indicate two possible ways in which the SMDs might be formed: 1) topo IV might be covalently cross-linked to DNA, forming a so-called protein-DNA adduct, or 2) Topo IV causes catenation, which is the reverse reaction of decatenation. To examine whether SMDs formed with hemi-methylated pFToriC are protein-DNA adducts of topo IV, SeqA, or both proteins covalently linked to DNA, the reaction mixtures containing SMDs were treated with combinations of SDS, EDTA, proteinase K, and phenol (Fig. 3A). The resistance of SMDs to all of these treatments implies that SMDs are not protein-DNA adducts. Digestion of SMDs with XhoI restriction enzyme, which cleaves one site on pFToriC, produced a unit-length of linearized pFToriC (Fig. 3B). Incubation with topo IV resolved the SMDs to relaxed DNAs (Fig. 3C). DNA gyrase, which does not effectively decatenate catenanes in vitro (28,29), did not resolve a significant amount of the SMDs. These results suggest that the SMDs generated by the cooperative action of topo IV and SeqA are interlinked plasmid DNAs, also known as catenanes, which can be resolved by topo IV, but not by DNA gyrase.
Like topo IV, DNA gyrase makes a double-strand break and performs strand passage between adjacent stretches of DNA. If plasmid molecules are aggregated by spermidine, the activity of gyrase leads to intermolecular linking and catenation of the plasmid DNA (43). Electron microscope studies have shown that binding excess SeqA to DNA results in aggregation of DNA molecules (7). This aggregation of DNA by SeqA might explain the catenation of plasmid DNAs found in the presence of topo IV. At a certain level of SeqA, plasmid DNAs form aggregates, allowing topo IV, which in addition might be stimulated by SeqA, to carry out intermolecular strand passages between adjacent plasmids in the aggregates. This intermolecular linking catenates the aggregated plasmid DNAs.
Topo IV Stimulation Is Dependent upon SeqA Affinity for DNA-Next we determined if SeqA binding to DNA is involved in the stimulation of topo IV activity (Fig. 4). Binding of SeqA to un-, hemi-, and fully methylated plasmid DNA was compared in agarose gel-shift assays (Fig. 4, A and B). Each type of plasmid bound SeqA with a different affinity. The hemi-methylated plasmid bound SeqA with the highest affinity, followed by fully methylated and then unmethylated DNA. These differences in affinity are in agreement with previous studies (4). As shown in Fig. 1, SeqA facilitated the topo IV-mediated conversion of negatively supercoiled hemi-methylated DNA to the relaxed state or to catenanes (Fig. 4C). Hemimethylated substrate DNA was converted to relaxed DNA most efficiently, and unmethylated DNA was converted least efficiently. Catenane formation was most pronounced in the presence of hemi-methylated DNA, followed by fully methylated DNA. These differences in the relaxation and the formation of catenanes are comparable to the respective binding affinities of SeqA to DNAs of the three categories methylation. These results suggest that the stimulation of relaxation The Decatenation of Topo IV Is Stimulated by SeqA-Topo IV possesses decatenation activity that resolves interlinked DNAs to monomers. To examine the effect of SeqA on the decatenation activity of topo IV, we treated the catenated pFToriC of Fig. 3 with topo IV and SeqA (Fig. 5A). The catenated DNAs were inefficiently detected by Southern blot analysis, because their transfer to nitrocellulose membranes is inefficient; however, decatenation could be detected by increasing the amounts of the monomer plasmids. Addition of 1.25 and 2.5 nM SeqA stimulated the conversion to monomers by topo IV. With increasing levels of SeqA, the conversion to monomers was inhibited. At more than 10 nM SeqA, further catenane formation, rather than monomerization, seemed to be favored.
Kinetoplast DNA (kDNA), highly interlinked catenanes of 2.5-kb minicircles, of the insect trypanosome Crithidia fasciculate can be used as a substrate for analyzing the decatenating activity of topoisomerases (26,44). Incubation of topo IV with kDNA resolved much of the interlinked DNA to monomers (Fig. 5B). Because the highly interlinked kDNA was not transferable to nitrocellulose membranes, only the decatenated DNA was detected in Southern blot analysis. In the presence of topo IV, SeqA increased the amount of resolved monomers. Without topo IV, SeqA was unable to produce monomers. SeqA seemed to stimulate the rate of decatenation as well as the total amount of monomers produced (Fig.  5C). The optimal amount of SeqA was 10 nM for decatenating 10 ng of kDNA (Fig. 5B) but 2.5 nM for decatenating 10 ng of hemi-methylated pFToriC (Fig. 5A). The insect trypanosome does not contain Dam methylase, therefore kDNA is not methylated at GATC sequences. The reason for the requirement of a greater amount of SeqA for kDNA decatenation might be the reduced affinity of SeqA for unmethylated DNA (Fig. 4, A and B). These results indicate that the process of decatenation is also dependent on the binding of the SeqA protein to DNA. FIG. 4. Stimulation of topo IV depends on SeqA binding to DNA. A, SeqA was added to 10-l topoisomerase reaction mixtures containing 10 ng (4.2 nM) of the un-, hemi-, and fully methylated pFToriC DNAs but not containing topoisomerase, incubated at 37°C, and then separated through 1% agarose. B, quantification of the unbound and unit size substrate DNAs from A. C, SeqA was titrated to the topo IV reaction mixture containing the indicated DNAs and 4.5 or 7.8 nM topo IV.

DISCUSSION
We demonstrate here that both the relaxation and decatenation activities of topo IV are stimulated by the SeqA protein.
Moderate levels of SeqA stimulated topo IV to convert negatively supercoiled DNA to its relaxed state (Figs. 1 and 4) and catenanes to monomers (Fig. 5). Whereas moderate levels of SeqA stimulate topo IV, DNA gyrase and topo I were not affected by similar SeqA levels, implying that the stimulation is specific to topo IV. Dependence of the stimulation upon SeqA affinity to DNA suggests that the stimulation of topo IV is mediated by the binding of SeqA to DNA. Because higher levels of SeqA cause substrate DNA to aggregate, topo IV might be able to carry out the formation of catenanes by intermolecular strand exchange between adjacent plasmid DNAs simply by virtue of aggregation. Therefore, the stimulation of the catenation activity of topo IV by SeqA might be of a passive character and due to aggregation of the DNA by SeqA.
Topo IV is the principal enzyme for decatenation of replicated DNA and mutations in parC or parE, and therefore, the cause of defects in the segregation of daughter chromosomes (23,24,45). Lack of topo IV activity also leads to an increase in the steady-state level of negative supercoiling because of a shift in the balance of topoisomerase activities in the cell (27,28,30,32). SeqA-deficient strains exhibit aberrant nucleoid formation, abnormal segregation of chromosomal DNA, and an increased steady-state level of negative superhelicity (11)(12)(13)(14)17). These until now unexplained parC-and parE-like phenotypes of seqA mutants might be explained by the results we are reporting here, namely that the SeqA protein is required for proper activity of the topo IV enzyme.
During replication, positive supercoils accumulate ahead of the replication fork. Relief of overwinding is necessary for progression of the replication fork, and DNA gyrase releases the increased positive superhelicity by introducing negative supercoils. Studies using topo IV-inhibiting drugs show that topo IV also participates in relieving the positive supercoils accumulated in front of replication forks. It has been proposed that part of the positive supercoils diffuses over to the replicated daughter strands behind the fork and forms positive precatenanes (21,33,46). If the precatenanes are not effectively removed, the daughter strands are suggested to become interwound and cause barriers for progression of the replication fork and segregation of the replicated chromosomes. Topo IV activity seems to be responsible for resolving the precatenanes (20,21,33). In an experiment in which topo IV, but not gyrase, was inhibited by norfloxacin, chromosome replication rapidly stopped in cells lacking SeqA, but did not stop in cells harboring SeqA (30). In this experiment, fork arrest was caused by the formation of topo IV-DNA adducts by the drug. This result therefore indicates that topo IV might be ahead of the forks in the absence of SeqA, but behind the forks in the presence of SeqA. ParC co-localized with the replication factory containing DNA polymerase III (47). The hemi-methylated daughter strands behind the fork are normally bound by the SeqA protein (4). SeqA bound here might contribute to localizing ParC to the area behind the fork. This option suggests that SeqA stimulates decatenation of the positive precatenanes diffused from the front of replication fork in two ways, both by stimulating the activity and affecting the localization of topo IV.
Our results suggest that the SeqA protein is involved in maintaining the integrity of replication forks throughout the replication cycle. At initiation, SeqA is involved in preventing the launch of an excessive number of forks (8, 10). During replication, SeqA stimulates topo IV-mediated removal of precatenanes diffused from ahead of replication forks and, at the FIG. 5. SeqA stimulates the decatenation activity of topo IV. The catenated pFToriC prepared in Fig. 3A was extracted in phenol-chloroform and recovered by ethanol precipitation. This DNA was resuspended and incubated with the indicated amounts of topo IV and SeqA in the topo IV reaction mixture. B, the topo IV reaction mixtures containing 10 ng (0.6 nM) of kDNA, and the indicated amounts of topo IV and SeqA were incubated for 10 min. C, the topo IV reaction mixtures containing 10 ng of kDNA and 1 nM of topo IV were incubated with or without SeqA for the indicated times. end of replication, stimulates topo IV-mediated decatenation of newly replicated daughter chromosomes.