Escherichia coli DNA Topoisomerase I and Suppression of Killing by Tn5 Transposase Overproduction: Topoisomerase I Modulates Tn5 Transposition

Abstract

Tn5 transposase (Tnp) overproduction is lethal to Escherichia coli. The overproduction causes cell filamentation and abnormal chromosome segregation. Here we present three lines of evidence strongly suggesting that Tnp overproduction killing is due to titration of topoisomerase I. First, a suppressor mutation of transposase overproduction killing, stkD10, is localized in topA (the gene for topoisomerase I). The stkD10 mutant has the following characteristics: first, it has an increased abundance of topoisomerase I protein, the topoisomerase I is defective for the DNA relaxation activity, and DNA gyrase activity is reduced; second, the suppressor phenotype of a second mutation localized in rpoH, stkA14 (H. Yigit and W. S. Reznikoff, J. Bacteriol. 179:1704–1713, 1997), can be explained by an increase in topA expression; and third, overexpression of wild-type topA partially suppresses the killing. Finally, topoisomerase I was found to enhance Tn5 transposition up to 30-fold in vivo.

A transposon is a DNA sequence that is able to insert itself into a new location in the genome without requiring any sequence similarity within the target DNA. Transposons cause genome rearrangements and are partially responsible for the spread of bacterial antibiotic resistance, heavy metal resistance, and increased tolerance to other deleterious agents (2). Furthermore, bacterial transposition reactions serve as paradigms for many gene transfer systems, including the processing and integration of the human immunodeficiency virus type 1 provirus (10, 17, 31). Therefore, understanding the mechanism of transposition and the role of host factors in this process is of considerable interest. Nonetheless, the involvement of host factors both in transposition and in retroviral integration is poorly understood.

Tn5 is a composite 5.8-kb transposon that contains 1.5-kb inverted repeats (IS50R and IS50L) (3, 38). IS50R and IS50L each have 19-bp repeats at their ends (the inside end and the outside end). These sequences are recognized by the transposase and by host factors involved in Tn5 transposition (3, 38). IS50R encodes two proteins that are important for transposition: a 476-amino-acid-long cis-acting transposase (Tnp) and a 421-amino-acid-long trans inhibitor (Inh) (3, 38). The coding sequences of Tnp and Inh are identical except that Inh lacks 55 N-terminal amino acids. IS50L encodes two proteins, P3 and P4, that have no known function. P3 and P4 have the same coding sequences as Tnp and Inh, respectively, except that both P3 and P4 lack 26 C-terminal amino acids. The central 2.75-kb region of Tn5 contains three cotranscribed antibiotic (kanamycin, bleomycin, and streptomycin) resistance genes. This region does not have a known function in transposition (3, 38).

Transposition is a very rare event in many organisms due to tight regulation whereby proteins (and/or functions) encoded by the transposon itself and, in some cases, specific host factors play a role (2, 38). For example, regulation of Tn5 transposition is dependent on the ratio between the two proteins, Tnp and Inh, encoded by the element itself (in IS50R). The Tn5 transposition frequency is also affected by a number of host factors (3, 38). Integration host factor (25), HU (3), DNA polymerase I (42), DnaA (55), topoisomerase I (Topo I) (46), and DNA gyrase (21) are probably involved in the positive regulation of Tn5 transposition, while Dam DNA methylase (56), Fis (51), and SulA (43) are negative regulators. However, the underlying mechanisms of the majority of these host factors in transposition remain to be determined.

Tn5 Tnp overproduction kills its host (52). It has been shown that Tnp overproduction causes filament formation and defective nucleoid segregation (52, 54). This phenomenon could be a consequence of Tnp interaction with a host factor(s) involved in transposition and a landmark event of the cell cycle, such as DNA segregation. Hence, Tnp overproduction killing could be used as a tool to study the host factors involved in Tn5 transposition and a possible relationship between transposition and chromosomal DNA segregation. Interestingly, Tnp overproduction killing does not require an active transposase (52). Nevertheless, analysis of N-terminal deletions of Tnp showed that deletion of the first 3 or 11 amino acids (in the Δ3 or Δ11 mutant) from the N terminus blocked killing while the same amount of Tnp accumulated in the cell (52). The wild-type N terminus of Tnp must, therefore, be critical for the cell killing phenotype.

To investigate the Tnp overproduction killing phenotype further, we isolated and localized four host mutations, designated stk, that suppress Tnp overproduction killing. These mutations were localized to four discrete loci in the Escherichia coli genome. The mutations map very close to genes known to be involved in cell division or DNA segregation: stkD10 at 28 min (shown in this report to be in topA, encoding Topo I), stkA14 at 76 min (in rpoH, encoding sigma 32 [54]), stkB33 at 85.5 min (near xerC, uvrD, and recQ), and stkC12 at 99.5 min (near dnaK, dnaJ, dnaT, and dnaC) (52).

Previously we have studied stkA14 and localized this mutation to rpoH. The sigma 32 mutation causes a constitutive induction of heat shock protein levels, suggesting that an induction of some sigma 32-programmed function(s) suppresses Tnp overproduction killing. However, none of the well-known heat shock functions appear to be involved in Tnp-associated killing (54).

In this report, we present evidence that Tnp killing may be due to titration of a specific host factor, DNA Topo I. The evidence comes from a detailed characterization of a second stk mutant, stkD10, localized in topA. This mutation causes a sixfold increase in Topo I abundance, and the mutant Topo I is only partially active. The mutation causes an alteration in gyrase activity. We also show that overproduction of wild-type Topo I partially suppresses Tnp overproduction killing. Additional evidence was obtained by a closer examination of stkA14. topA is transcribed at a higher rate in stkA14, increasing Topo I abundance and activity. These results suggest that the suppressor phenotype of this mutation is probably due to an increase in Topo I abundance. Finally, we present evidence (in partial agreement with Sternglanz et al. [46]) that DNA Topo I stimulates Tn5 transposition up to 30-fold in vivo. These results suggest that there is a relationship between Tnp overproduction killing and Topo I.

MATERIALS AND METHODS

Strains and media.

The E. coli K-12 strains used (Table 1) were grown in Luria broth (LB) (41) or M9-glucose minimal salt medium (41) containing all amino acids (except methionine and cysteine), each supplemented with the following five vitamins at 0.01 mM: p-aminobenzoic acid, p-dihydroxybenzoic acid, p-hydroxybenzoic acid, pantothenate (calcium salt), and thiamine (50). MgSO₄ was added to 0.1 mM (final concentration). LB plates contained 15 g of Bacto Agar per liter. The antibiotic concentrations were as follows: chloramphenicol, 20 μg/ml; ampicillin, 100 μg/ml; kanamycin, 40 μg/ml; streptomycin, 100 μg/ml; nalidixic acid, 5 μg/ml; and tetracycline, 15 μg/ml.

TABLE 1.

E. coli strains used in this study

Strain	Relevant genotype	Source
MC1061	hsdR mcrB araD139 Δ(araABC-leu)7679 ΔlacX174 galU galK rpsL thi	M. Casadaban
XL1-Blue	supE44 hsdR17 recA1 endA1 gyrA46 thi relA1 lac
	F′ [proAB+lacIqlacZΔM15 Tn10(Tetr)]	Lab collection
DM800	Δ(top cysB)204 araC13 gyrB225	J. Wang
SD7	gal-25 λ−topA10 pyrF287 fnr-1 rpsL195
	(Strr) gyrB226 iclR7(Con)trpR72(Am)	ATCCa
RS2	gal-25 λ−topA10 pyrF287 fnr-1 rpsL195
	(Strr) gyrB226 iclR7(Con) trpR72(Am)	ATCC
JTT1	gal-25 λ−pyrF287 fnr-1 rpsL195
	(Strr) gyrB226 iclR7(Con) trpR72(Am)	ATCC
MDW10	Δ(lac-pro) ara polA thi SpcrrpsL Nalr	M. Weinreich
	malT1(λr)
MDW505	MC1061/pRZ4775	M. Weinreich
MDW507	MC1061/pRZ4777	M. Weinreich
MDW565	MC1061stkD10	M. Weinreich
MDW566	MC1061stkC12	M. Weinreich
MDW570	MC1061stkA14	M. Weinreich
MDW573	MC1061stkB33	M. Weinreich
HY116	SD7 F′ [proAB+lacIqlacZΔM15 Tn10(Tetr)]	This study
HY117	RS2 F′ [proAB+lacIqlacZΔM15 Tn10(Tetr)]	This study
HY118	JTT1 F′ [proAB+lacIqlacZΔM15 Tn10(Tetr)]	This study
HY126	DM800/pBAD33 (Kanr) F′ [proAB+lacIqlacZΔM15 Tn10(Tetr)]	This study
HY137	HY116/pJW312-SalI (Ampr)	This study
HY138	HY117/pJW312-SalI (Ampr)	This study
HY139	HY118/pJW312-SalI (Ampr)	This study
HY152	HY126/pJW312-SalI (Ampr)	This study
HY156	MC1061 stkE22/pRZ4775 (Camr)	This study
HY156	MC1061 stkF32/pRZ4775 (Camr)	This study

^aATCC, American Type Culture Collection. 

Plasmids.

Most of the plasmids used in this study are detailed in Fig. 1. pRZ4775 (encoding Tnp under λp_R control) was described by Weinreich et al. (52). pRZ4787 was described by Weinreich (53). pJW312-SalI, carrying topA (encoding Topo I), was obtained from J. Wang (57). Cloning vector pBII was supplied by A. Roca, and pBIP3, a cloning vector for allele exchange experiments, was obtained from R. Maurer (45).

FIG. 1.

Plasmids used in this study. pRZ4775, encoding wild-type Tnp, does not have IS50 end sequences; pRZ4787 encodes Tnp and contains a mini-Tn5 cassette; pRZ4775 does not express Tn5 Inh; pRZ4787 expresses Tn5 Tnp and Inh; pJW312-SalI encodes wild-type Topo I under P_lac control. OE, outside end; IE, inside end; amp, ampicillin resistance; cam, chloramphenicol resistance; kan, kanamycin resistance.

To perform allele replacement studies, a DNA fragment containing a wild-type or mutant topA gene was cloned as follows. Wild-type topA was cloned from pJW312-SalI by digestion with HindIII-PvuII and ligation into HindIII-SmaI sites of pBII to create pRZ8860. An ApaI-NotI fragment of pRZ8860 encoding wild-type topA was then cloned into the same sites of pBIP3, yielding pRZ8861. The same PCR product used for sequence determination of the stkD10 mutation was cut with PvuII-AatII and then cloned into the same sites of pRZ8860 to generate pRZ8862. The ApaI/NotI fragment containing the topA mutation of pRZ8862 was cloned into the same sites of pBIP3 to construct pRZ8864. The presence of the mutation on pRZ8864 was confirmed by sequencing; then pRZ8864 (encoding mutant topA) and pRZ8861 (encoding wild-type topA) were used for the allele replacement experiment to test whether the mutation was sufficient to confer suppression of Tnp-associated killing in an otherwise wild-type background (MC1061).

topA allele replacement.

Mutant and wild-type alleles of topA were subcloned into pBIP3 as described above. The plasmids were then transferred into E. coli, and a phagemid-based system for generating allele replacements was used as described by Slater and Maurer (45). After presumed allele exchange, about 200 to 250 individual colonies for each construct were tested for the transposase overproduction phenotype at 42°C. pRZ4775 was used as the Tnp overproducer in these experiments.

³⁵S labeling.

The cells were grown in M9-glucose medium supplemented with all amino acids (except methionine and cysteine) and five vitamins (50) at 32°C to an optical density at 450 nm of 0.1. The cultures were then shifted to 42°C to induce Tnp overproduction and, after a 1-h induction, pulse-labeled by addition of 20 μCi of Trans³⁵S-label mixture (l-[³⁵S]methionine and l-[³⁵S]cysteine; 1.146 Ci/mmol; ICN Pharmaceuticals, Inc.) for 1 min with shaking; then the label was chased by addition of 100 μl of methionine and cysteine (each at 10 mg/ml) for 1 min. The labeled sample was mixed with 110 μl of 50% trichloroacetic acid (final trichloroacetic acid concentration of 5%). For each sample, (5 × 10⁵ cpm was analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) on a 10% gel.

Western blot analysis of Topo I.

The overnight cultures of the wild type (MDW505) and an stkD10-containing strain (MDW565) were diluted 1:100 and grown to mid-log phase at 32°C (in the absence of Tnp) or at 42°C (in the presence of Tnp). The cells were harvested and sonicated (57). Equal amounts of proteins from all samples were separated by SDS-PAGE (10% gel) and transferred onto nitrocellulose filters (41). Western blots were performed as instructed by the manufacturer (DuPont NEN). The antibody for E. coli DNA Topo I was kindly supplied by J. Wang. The Western blots were scanned in a densitometer, and the intensities of the protein bands were quantitated by using ImageQuant and Excel programs.

Quantitation of Tnp overproduction killing by efficiency of plating experiments.

After transformation of the plasmids (pRZ4775 and/or pJW312-SalI) into CaCl₂-treated wild-type (MC1061) competent cells (41), 12 to 20 individual colonies for each sample were grown overnight at 32 or 37°C in LB containing the appropriate antibiotic(s). Depending on the mode of regulating Tnp synthesis and/or various host proteins in individual experiments, dilutions of each sample were plated onto LB agar plates containing the appropriate antibiotic(s) and the indicated concentrations of isopropyl-β-d-thiogalactopyranoside (IPTG) (see figure legends) or were incubated at 32 and 42°C. After 18 h, the number of colonies formed in the presence of IPTG or at 42°C was compared with the number in the uninduced control samples (grown in the absence of IPTG or at 32°C).

PCR amplification and DNA sequencing of topA.

The coding region topA and its control region were PCR amplified in four partially overlapping fragments from both stkD10 and wild-type (MC1061/pRZ4775) strains (see Fig. 4). The amplifications were carried out as described previously (54). The PCR-amplified fragments were sequenced with dye termination (PRISM; Applied Biosystems Inc.) by following the manufacturer’s protocol. The sequencing reactions were analyzed at the University of Wisconsin Biotechnology Center. The presence of the mutation was confirmed by using four separate PCR stocks: two different stocks of fragment 1, reading the noncoding strand; and two different stocks of fragment 2, reading the coding strand.

FIG. 4.

stkD10 is located in topA. The topA region was PCR amplified and sequenced. The stkD10 mutation was localized in the overlapping region of fragments 1 and 2. The DNA and resulting amino acid sequence changes are shown. The presence of a mutation was confirmed by analyzing four different PCR stocks.

Protein assays.

Protein concentrations in cell extracts were determined by the Bradford protein assay as described by Rossomando (39).

Measurement of plasmid supercoiling.

To determine plasmid DNA supercoiling in different stk mutants, the mutant and wild-type strains were transformed with pUC19. Plasmid DNA was isolated from overnight cultures by the alkaline method as described by Sambrook et al. (41). The plasmid DNAs were analyzed by 1% agarose gel electrophoresis in the presence of chloroquine (12 μg/ml). Electrophoresis was carried out as described by Mizushima et al. (30). The gels were stained with Syber Green II (Molecular Probes) and analyzed with a FluorImager (Molecular Dynamics).

Topo I relaxation assay.

DNA Topo I relaxation activity was assayed by the ability of a crude lysate prepared from the stk and wild-type strains to relax supercoiled pUC19 DNA. The plasmid DNA was prepared by CsCl purification as described previously (41). The crude cell extract was prepared and the relaxation assay was performed exactly as described by Zumstein and Wang (57) except that the gels were stained with Syber Green II (Molecular Probes) and were examined with a FluorImager (Molecular Dynamics).

DNA gyrase activity assay.

DNA gyrase activity was assayed by the ability of a crude lysate prepared from the stk and wild-type strains to introduce negative supercoils into relaxed pUC19 DNA in the presence of ATP. The same CsCl-purified pUC19 DNA as used in the Topo I relaxation assays was used, except that the DNA was relaxed by using wheat germ Topo I (Promega) as described by the manufacturer. The gyrase activity tests were carried out as described by DiNardo et al. (12), but with the same crude cell extracts as prepared for the Topo I relaxation assay. The gels were examined after staining with Syber Green II with a FluorImager.

Primer extension.

The primer extensions were carried out as described by Sambrook et al. (41). The gels were analyzed with a PhosphorImager (Molecular Dynamics).

Transposition assays.

The λ infection assays were carried out as described by Sternglanz et al. (46) by using NK467. The stains were modified to obtain isogenic backgrounds as follows. An F′ plasmid encoding lacI^q from E. coli XL-Blue was introduced into RS2, SD7, and JTT1. Then pJW312-SalI was introduced into the strains containing the F′ plasmid. Six individual colonies for each strain (RS2F′, RS2F′/pJW312-SalI, SD7F′, SD7F′/pJW312-SalI, JTT1F′, and JTT1F′/pJW312-SalI) were used in the λ infection assay. RS2 and SD7 are topA10 strains with the same genotype. To ensure that the observed results were not due to genotypic changes during storage in a given strain, two different topA strains were used. For the same reason, two different ΔtopA strains (DM800 and a derivative) were also used. To determine the transposition frequency under Topo I overproduction conditions, strains containing pJW312-SalI (the plasmid encoding topA) were induced by using 0 to 0.2 mM IPTG. DM800 was transformed by pBAD33, so that the strain would contain cam as a selective marker. The F′ plasmid from E. coli from XL-Blue containing lacI^q was introduced yielding HY126. This strain is ΔtopA. To obtain an isogenic topA⁺ strain, HY126 was then transformed with pJW312-SalI, resulting in HY152. For each strain, 16 individual colonies were used in the λ infection assay.

RESULTS

Localization of stkD10 in topA. (i) stkD10 causes a five- to six-fold increase in the abundance of Topo I.

Previously four stk (suppressor of transposase overproduction killing) mutations were isolated and located to four discrete loci in the E. coli genome (53). We studied these mutants further in order to investigate the host factors involved in Tn5 transposition and its regulation. This strategy might also help to determine whether there is a coupling between Tn5 transposition or its regulation and chromosomal DNA segregation, since Tnp overproduction causes defective DNA segregation (52, 54).

The Hfr crossing and P1 mapping results localized the stkD10 mutation to 28 min near topA, tonB, terA, and terD (52). To determine if there was any change in protein synthesis or accumulation levels in stkD10-containing strains, we examined pulse-labeled samples and compared them to a wild-type strain and an stkA14-containing strain (Fig. 2). These results suggested that stkD10 causes an increased synthesis of a 90-kDa protein and two proteins of approximately 100 kDa (Fig. 2, lane 2). Topo I is one of the two proteins encoded at the 28-min region, having a molecular mass of close to 100 kDa. Thus, we suspected that the stkD10 mutation may cause an increase in topA expression.

FIG. 2.

stkD10 causes higher levels of three proteins of 90 to 100 kDa. Pulse-labeling and SDS-PAGE analyses were done as described in Materials and Methods; autoradiograms of the gels are presented. The autoradiograms were quantitated by densitometric analyses. All strains used, wild type (wt), MDW565 (stkD10), and MDW570 (stkA14), harbor pRZ4775 for Tnp overproduction. The molecular masses were estimated from the known molecular masses of Tnp (53 kDa), DnaK (70 kDa), and Hsp90. The gel shown is representative of three individual experiments.

Next we examined by Western blot analysis whether one of the proteins whose abundance was enhanced in stkD10 was Topo I. The Topo I level in stkD10 was compared to the wild-type (MC1061) level both at 32°C in the absence of Tnp and at 42°C in the presence of Tnp. Equal amounts of protein from all samples were examined by Western blotting as described in Materials and Methods with Topo I polyclonal antibodies. As shown in Fig. 3, two protein bands were detected by the Topo I antibody. We assume that one band is full-length Topo I and the second is a Topo I degradation product. It is possible that these two bands represent the two ∼100-kDa bands seen to increase in the stkD10 strain in Fig. 2. Densitometric examination of the Western blots shows that the Topo I level in the stkD10 mutant increased five- to sixfold at 32°C and fourfold at 42°C (Fig. 3A and B, respectively). These observations together with the results of genetic mapping strongly suggested that the stkD10 mutation could be in the topA regulatory or coding region.

FIG. 3.

stkD10 mutation causes an increase in the level of Topo I. Western blot analysis of cell extracts (32 μg of total cell protein from each strain) prepared from wild-type (wt), stkD10-containing, and stkA14-containing strains was carried out as described in Materials and Methods with Topo I antibodies, and the results were analyzed by densitometry. (A) Cell extracts prepared from the wild-type strain and the stkD10-containing strain at 32°C (in the absence of Tnp); (B) cell extracts prepared from the wild-type strain and the stkD10-containing strain at 42°C (in the presence of Tnp); (C) cell extracts prepared from the stkA14-containing strain at 42°C (in the presence of Tnp) and the stkA14-containing strain at 32°C (in the absence of Tnp) and from the wild-type strain. Comparable Western blot experiments were performed four times with similar results.

(ii) stkD10 is located in topA (encoding Topo I).

To determine the precise nature of the stkD10 mutation, we performed a sequence analysis of PCR-amplified topA and its control region (both from an stkD10-containing strain and from the parental wild-type strain) as shown in Fig. 4. Sequencing of these PCR-amplified fragments as described in Materials and Methods revealed that the stkD10 mutation lies in topA and results in an alanine-to-aspartate change at codon 118 (Fig. 4). We confirmed this mutation by sequencing four individual PCR-amplified stocks.

(iii) The topA mutation is sufficient for the suppressor phenotype caused by the stkD10 mutation.

To investigate whether the topA mutation by itself is sufficient to suppress Tnp killing, we cloned wild-type topA onto a phagemid, pBIP3, and inserted the mutation stkD10 as described in Materials and Methods. The presence of the mutation was confirmed by sequencing the phagemid clone. We used the allele replacement method described by Slater and Maurer (45) to introduce the topA mutation and, as a control, the wild-type allele into a wild-type strain. After selecting appropriate recombinants, we tested the isolates for the Tnp overproduction killing. In four individual experiments in which 60 colonies were tested each time, we found that when the topA mutation was introduced into the wild-type strain, 38 (±2)% of the recombinants became mutant (98% resistant to Tnp overproduction killing) while 62 (±2)% remained wild type (99% sensitive to Tnp overproduction killing). When the wild-type clone was used, 98 (±2)% of the recombinants remained wild type (99% sensitive for Tnp killing). These results clearly show that the topA mutation is sufficient for the suppression of Tnp killing in the stkD10-containing strain.

Properties of the mutant Topo I.

It is critical to examine the effects of the mutation (stkD10) on Topo I activities in order to understand the mechanism(s) for the suppression of Tnp killing. In addition, an understanding of the properties of the Topo I mutant may help elucidate a possible role for Topo I in Tn5 transposition.

(i) The mutant Topo I is defective for the DNA relaxation activity.

Here, we present data showing that the stkD10 mutant version of Topo I is defective in its ability to relax negatively supercoiled DNA. This conclusion was made by examining DNA supercoiling of plasmid DNA isolated from the stkD10-containing strain as well as testing the crude cell extracts prepared from stkD10-containing and wild-type strains for Topo I relaxation activity.

To examine DNA supercoiling in vivo in the stkD10-containing strain, we introduced pUC19 into the wild-type and stkD10-containing strains. The plasmid DNA was then isolated and examined as described in Materials and Methods. Figure 5 shows the resolution of the topoisomers of pUC19 isolated from stkD10-containing and wild-type bacteria, along with a densitometric graph of the gel. From these data, it is clear that plasmid DNA isolated from stkD10 is more negatively supercoiled than the same plasmid DNA isolated from wild-type bacteria. This result suggests that mutant Topo I is not fully active in relaxation activity. The conclusion that DNA isolated from the stkD10-containing strain was negatively supercoiled arose from examining the resolution of topoisomers on gels containing increasing concentrations of chloroquine in comparison with DNA from the wild-type strain; as the concentration of chloroquine increased, the negatively supercoiled DNA first became relaxed and then became positively supercoiled. For simplicity, we only show one gel run at the lowest chloroquine concentration indicating the differences of DNA supercoiling between the parental wild-type strain and the stkD10-containing strain.

FIG. 5.

stkD10 causes a decrease in Topo I relaxation activity. Plasmid (pUC19) DNAs (2 μg) prepared from the wild-type (wt) and stkD10-containing strains were electrophoresed in the dark on a 1% agarose gel containing chloroquine (12 μg/ml). The gel was stained and analyzed with a FluorImager (Molecular Dynamics). Densitometric traces of the two lanes have been overlaid to show the differences in DNA supercoiling.

We also used an in vitro Topo I relaxation assay (57) to study stkD10 Topo I. The results shown in Fig. 6 are consistent with the results of the in vivo assay in that the mutant Topo I is defective for relaxation activity. This experiment was repeated with three different crude cell extracts where at least two individual assays were performed per preparation. This reduction in activity in the stkD10 extracts occurred despite the fact that the extracts contained about six times as much Topo I protein as the wild-type extracts (Fig. 3A and B).

FIG. 6.

stkD10 causes reduced Topo I relaxation activity in vitro. Crude cell extracts (12 μg of crude cell protein/reaction) prepared from wild-type (wt), stkD10-containing and stkA14-containing strains at 32°C were used to relax supercoiled pUC19 DNA (0.5 μg/reaction) in vitro (57). The reactions were stopped (57) and run on a 1% agarose gel. The gel was stained and analyzed with a FluorImager (Molecular Dynamics). Densitometric traces of lanes 2 to 4 have been overlaid to show the differences in DNA supercoiling. Lane 1 contains the unreacted substrate DNA.

(ii) stkD10 confers a modest alteration in gyrase activity.

The same crude cell extracts as used in the Topo I assays were used to determine DNA gyrase activity in both stkD10-containing and wild-type strains as described by DiNardo et al. (12) (Fig. 7). Quantitation of the gels indicated that DNA gyrase activity is reduced four- to fivefold in the stkD10 mutant. This was calculated by measuring the percentage of relaxed pUC19 DNA converted into supercoiled DNA. A densitometric scan of an overlaid graph of the gel is also shown in Fig. 7 to demonstrate the differences between the wild type and the stk mutants.

FIG. 7.

stkD10 and stkA14 mutations cause a modest alteration in gyrase activity. The same crude cell extracts (12 μg of crude cell protein/reaction) as used for Fig. 6 were used to measure gyrase activity in vitro. The percentage of pUC19 converted to supercoiled DNA was measured (12). One percent agarose gel electrophoresis was used, and the topoisomers were quantitated by densitometry. Lanes: 1, control DNA; lanes 2 and 5, DNA treated with a wild-type (wt) cell extract; 3 and 6, DNA treated with a cell extract from the stkD10-containing strain; 4 and 7, DNA treated with a cell extract from the stkA14-containing strain. The cell extracts used in lanes 5 to 7 were fourfold diluted. Densitometric traces of lanes 5 to 7 have been overlaid to show the differences in the DNA supercoiling.

The reduced gyrase activity in stkD10 could be explained by a reduction in the expression of gyrase due to supercoiling-dependent transcription regulation (13).

Co-overproduction of wild-type Topo I partially suppresses Tnp killing.

As mentioned above, there is in vivo evidence that Topo I is increased in stkD10 and stkA14. Therefore, we wanted to directly assess whether there is a relationship between suppression of Tnp killing and an increase in Topo I quantities. To test this possibility, we overproduced wild-type Topo I from P_lac by varying the concentration of IPTG (0 to 0.2 mM) to achieve different levels of expression while overproducing Tnp under the control of lambda p_R. The results in Fig. 8 (averages from separate experiments [the differences between experiments were less than 2%]) show that Topo I co-overproduction suppresses Tnp killing 100-fold. Although this level of suppression is substantially less than observed for the stkA14 or stkD10 mutant, its occurrence supports a link between Tn5 Tnp killing and Topo I level.

FIG. 8.

Co-overproduction of wild-type Topo I suppresses Tnp killing. Overnight cultures of MDW505/pJW312-SalI (+ Topo I) and MDW505 (wild type) were diluted and plated on LB agar containing various concentrations of IPTG (0 to 0.2 mM). After 18 to 14 h of incubation at 32°C (no Tnp induction) and at 42°C (Tnp induction), the colonies were counted and the percentage of CFU of the Tnp-induced cultures relative to the uninduced cultures was calculated. In general, the variation between different experiments for the same data points was less than 1%. IPTG induces overproduction of Topo I. For instance, Western blot analysis indicated a 10- to 12-fold increase in Topo I when cells were grown in the presence of 0.06 mM IPTG.

Possible connection between stkA14 and stkD10. (i) stkD10 and stkA14 both lead to an increase in total topA mRNA.

Our previous studies regarding stkA14 (in rpoH) suggested that a sigma 32-regulated function other than induction of the well-known heat shock response proteins is responsible for the suppressor phenotype of stkA14. Since it has been shown that the P1 promoter of topA is sigma 32 dependent (23), we examined Topo I RNA levels in an stkA14-containing strain and an stkD10 containing strain. We used primer extension to determine the levels of the topA transcripts and to determine which promoters are responsible for the transcription in a given strain. Equal amounts of total RNA were subjected to primer extension (Fig. 9). The primer extension studies show that in the stkA14-containing strain, 70% of the mRNA is driven by the P1 promoter and the P1 RNA level is 10-fold higher than in the wild-type strain (Fig. 9, lanes 1 and 3). In addition there was two- to threefold increase from the P2 and P3 transcripts in the stkA14-containing strain. The total topA mRNA level in the stkD10-containing strain is about six- to sevenfold increased (lanes 1 and 2) and is driven mostly from P2 and P3. This could be explained by supercoiling-dependent regulation of topA transcription (47, 48). This finding of increased transcription of topA in the stkD10-containing strain clearly explains the increased abundance of the mutant Topo I.

FIG. 9.

stkD10 (Topo I mutant) and stkA14 (sigma 32 mutant) alter topA mRNA levels. Primer extensions were done with equal amounts of total RNA (2 μg) isolated from wild-type (wt), stkD10-containing, and stkA14-containing strains. The P4 promoter of topA was used as an internal control. The gels were quantitated with a PhosphorImager (Molecular Dynamics). P1, P2, P3, and P4 indicate the positions of known topA transcripts. The sequencing standard (G, A, T, and C) was a dideoxy primer extension analysis of wild-type topA.

We also determined by Western blot analysis (Fig. 3C) that the Topo I protein level is increased four- to fivefold in the stkA14-containing strain.

(ii) stkA14 causes an increase in both DNA gyrase and Topo I activities.

Crude cell extracts prepared from the stkA14-containing strain were used to determine Topo I relaxation activity in vitro as described by Zumstein and Wang (57) (Fig. 6, lane 3). Figure 6 also shows a densitometric scan of lane 3. The results (averages from three separate experiments) showed that there is about a four- to fivefold increase in the topoisomerase relaxation activity in the stkA14-containing strain. This result correlates with the Topo I protein level in that strain.

We also assayed gyrase activities in the crude cell extracts prepared from the stkA14-containing strain (Fig. 7, lanes 4 and 7). As seen in lane 7, stkA14 causes a modest (two- to threefold) increase in gyrase activity. This increase is most probably in response to the increase in Topo I activity (13, 26).

These results suggest that the suppressor phenotypes of both stkA14 and stkD10 mutations are due to an increase in the level of Topo I.

DNA topology is not critical for suppression.

If an alteration of DNA topology is required to suppress Tnp overproduction killing, one would expect to see a similar pattern of in vivo DNA topology in all suppressor strains. We compared the in vivo plasmid DNA supercoiling in all stk mutants that were isolated (52, 53) and with that of the same plasmid DNAs isolated from the wild type.

The results clearly show that an alteration in DNA topology is not required for the suppression of Tnp killing by stk mutants. Figure 10 shows the distribution of topoisomers of plasmid DNAs (pUC19) isolated from the wild type and several stk mutants. Of the six mutant strains isolated, only the stkD10 mutant shows a significant difference in DNA supercoiling in vivo. These results suggest that the suppression in stkD10 is not due simply to an alteration in DNA topology.

FIG. 10.

DNA topology is not critical for suppression. Plasmid (pUC19) DNAs (0.5 to 2 μg) prepared from wild-type (wt) and stk mutants were electrophoresed in the dark on a 1% agarose gel containing chloroquine (12 μg/ml). The gel was stained and analyzed with a FluorImager (Molecular Dynamics).

Topo I is a stimulator of Tn5 transposition.

If Topo I plays an essential role in Tn5 transposition, then topA-defective mutations should dramatically affect the frequency of Tn5 transposition. Sternglanz et al. (46) have reported that Tn5 transposition decreases 70- to 90-fold in ΔtopA strains and about 40-fold in topA10 strains. These strains were subsequently shown to have compensatory mutations in DNA gyrase. Therefore, we have reexamined the effects of both ΔtopA and topA10 on Tn5 transposition, using control strains that are as nearly isogenic with the topA mutants as possible.

The frequency of Tn5 transposition was tested in topA strains and the isogenic topA⁺ backgrounds (Table 2). The ΔtopA strain HY126 showed a 10- to 30-fold decrease in Tn5 transposition frequency, but neither of the topA10 strains (HY116 and HY117) showed more than a 2-fold difference in transposition frequency. Also, the frequency of Tn5 transposition did not change more than twofold in the stkD10 mutant (data not shown). These results suggest that the presence of Topo I modulates Tn5 transposition but it is not absolutely required; they also suggest that the presence of Topo I but not its relaxation activity is critical for its stimulation of transposition. We also examined Tn5 transposition in a strain overproducing wild-type topA (data not shown). The results suggested that overproduction of topA does not further stimulate Tn5 transposition.

TABLE 2.

Tn5 transposition in topA strains^a

E. coli strain	topA allele	Mean Tn5 transposition frequencyb (SD)	Relative frequency
HY118	topA+	1.3 × 10−3 (±1)	1
HY139	topA++	4 × 10−4 (±1.2)	0.31
HY116	topA10	2.6 × 10−4 (±0.8)	1.3
HY137	topA+	2 × 10−4 (±0.4)	1
HY117	topA10	9.3 × 10−4 (±2)	2.3
HY138	topA+	4 × 10−4 (±0.54)	1
HY126	ΔtopA	0.96 × 10−4 (±1)	0.033
HY152	topA+	2.8 × 10−3 (±1)	1

^aStrains HY139, HY137, HY138, and HY152 contain pJW312-SalI, encoding wild-type topA under the control of P_lac. topA⁺⁺ indicates that a diploid strain was used. 

^bTested by λ infection assay in the absence of IPTG and calculated as described in Material and Methods. 

DISCUSSION

Tnp killing is due to a titration of Topo I.

Detailed examination of stkD10 and stkA14 mutants strongly suggests that their suppressor phenotypes are due to an increase in Topo I levels. stkD10 is mutated for topA. The mutant Topo I has reduced relaxation activity (Fig. 5 and 6). This reduction is likely responsible for the increase in topA transcription, which then causes the increased level of Topo I. This is due to the supercoiling-dependent regulation of topA transcription (47, 48). The reduced DNA gyrase activity (Fig. 7) in this mutant can be explained by the supercoiling-dependent regulation of gyrase transcription (16, 26).

stkA14 is a sigma 32 mutant (54). Recent studies have suggested that some host factors other than the major heat shock proteins are also regulated by sigma 32 (6, 7, 9, 36). It has been shown that topA is one of the factors regulated by sigma 32 (23). Our studies demonstrate that stkA14 results in elevated levels of topA mRNA (Fig. 9) and Topo I protein (Fig. 3). As a result, the activity of Topo I in the stkA14 mutant is increased (Fig. 6). Probably in response to an increase in Topo I activity, the activity of DNA gyrase is also increased in the stkA14 mutant (Fig. 7).

The above results strongly suggest that Tnp overproduction killing could be overcome by an increase in the level of Topo I. This model was supported by the finding that Tnp overproduction killing can be at least partially suppressed by co-overproduction of wild-type Topo I (Fig. 8).

Overproduction of Topo I by itself is deleterious to the cell (49a). This might explain why only a partial recovery is seen when wild-type Topo I is co-overproduced. In the cases of stkA14 and stkD10, there may have been a selection for achieving the best level of Topo I synthesis to balance these two phenomena.

Altogether, these results strongly suggest that Tnp overproduction killing is due to a titration of Topo I. From the studies done by Wang (49) and DiNardo et al. (12), it is clear that Topo I is essential for cell survival. Consequently, a titration of Topo I by Tnp overproduction could be sufficient for the killing. In support of this conclusion, we have shown that Tnp and Topo I copurify and Tnp (but not Δ37, whose overproduction is not lethal) can inhibit Topo I in vitro (54a).

Topo I modulates Tn5 transposition.

The examination of stkD10 and stkA14 mutants suggested that Topo I is involved in Tnp killing. Sternglanz et al. (46) presented evidence that Topo I enhances Tn5 transposition; however, it was later suggested that these results were due to the compensatory DNA gyrase mutants in these backgrounds (3).

Here, we show that Topo I enhances Tn5 transposition in vivo. Unlike the case for DNA gyrase, in which the activity of gyrase is required (21), the presence but perhaps not the activity of Topo I is required (Table 2). The conclusion that the activity is not critical comes from an examination of Tn5 transposition in the stkD10- and topA10-containing strains. The stkD10-containing strain is defective for Topo I relaxation activity yet the frequency of Tn5 transposition is not reduced but rather is enhanced two- to threefold (Table 2). topA10 has been extensively studied (57). This allele encodes a Topo I that is defective for relaxation activity but is partially functional for Topo I-DNA covalent complex formation. In a ΔtopA strain, we found that the transposition frequency is decreased 10- to 30-fold (Table 2). Therefore, the decrease in Tn5 transposition frequency correlates with the absence of Topo I protein. However, this decrease in the transposition frequency does not reflect the accumulation of more negatively supercoiled DNA in vivo. Additionally, the comparative analysis of ΔtopA and ΔtopA/pJW312-SalI strains clearly eliminates the possibility that the decrease in the transposition in the ΔtopA background is due to a compensatory gyrase mutation in this background.

The results obtained here differ from those of Sternglanz et al. (46). In their study, a parental strain, wild-type for both gyrB and topA, was used for comparison to a ΔtopA strain later shown to have a compensatory mutation in gyrB (12). In this comparison, an 80- to 90-fold depression in Tn5 transposition was observed in the ΔtopA strain. Additionally, a 40-fold difference was observed in a topA10 strain that was later shown to have a compensatory mutation in gyrB (12). But the results of Isberg and Syvanen (21), showing that the activity of gyrase is a stimulator for Tn5 transposition, suggest that the difference between our results and the results of Sternglanz et al. (46) may be due to a gyrB or gyrA mutation.

The involvement of Topo I in Tn5 transposition could be explained by one or more of the following models. (i) Topo I could be involved in target capture. Involvement of a secondary protein encoded by the transposon itself or by the host in target recognition in transposition and retroviral integration has been shown (10). For example, it has been shown for bacteriophage Mu transposition that MuA (transposase) can select a target DNA by itself, but MuB bound DNAs are preferred targets (28, 33). (ii) Alternatively Topo I could be involved in resolving the transposition products that are catenated. These models are currently under investigation.

Tnp and Topo I copurify on Ni-nitrilotriacetic acid when Tnp is histidine tagged (54a). These recent studies strongly suggest that the involvement of Topo I in Tn5 transposition is through a direct protein-protein interaction. In addition, we have discovered that Tnp inhibits Topo I relaxation activity in vitro as well as in vivo (54a). Finally, we have recently discovered that purified Topo I stimulates Tn5 transposition in a defined in vitro system and that this stimulation is only partially due to its relaxation activity (54a).

Could Topo I titration by Tnp cause a defect in DNA segregation?

The results presented here suggest that Tnp overproduction killing could be due to a specific interaction between Tnp and Topo I. We have also shown that Tnp overproduction causes defective DNA segregation and an increase in anucleated cell formation (52, 54). It has been shown that anucleated cells form in par (gyrB [40] and parC [1, 40]), mukA (19, 20), mukB (15, 34), minB (6), and xerC (4, 5) mutants. These gene products have been suggested to be involved in the bacterial chromosome segregation processes. Recently it has been shown that topA mutants can suppress mukB mutants (44) and overproduction of topA can suppress parC mutants (22, 35). These results suggest that Topo I could be involved in chromosomal DNA segregation, even though studies of the four topoisomerases of E. coli have suggested that Topo I is incapable of catalyzing decatenation in vitro (18, 49).

Many of the processes involved in DNA segregation may be membrane associated (14, 20, 27, 40); 34% of Topo I (29) and 20% of Tnp (54) are associated with the inner cell membrane, and Tnp membrane association correlates with the killing effect and the defective nucleoid segregation (54). Unfortunately, neither the mechanism nor the role of Tnp-membrane or Topo I-membrane association is known.

We hypothesize that the membrane-bound form of the Topo I is involved in the chromosomal DNA segregation process. Perhaps Tnp is localized in the membrane via Topo I. This interaction could then block proper DNA segregation and result in cell killing when Tnp is overproduced. The possible benefit of this interaction for transposition could be explained in combination with the role of Dam DNA methylase in Tn5 transposition. Dam DNA methylase regulates the synthesis of Tnp and (for IS50) the transposition event itself, suggesting that transposition preferentially takes place soon after the DNA replication fork passes when the donor DNA is hemimethylated (2, 38). This would allow the use of the sister chromosome to fill in the gap left by the excision of a Tn5. The membrane-bound Topo I-Tnp might delay the segregation process until transposition and the required repair in the donor are completed. This could be critical for the transposon to be able to be inherited in both cells when it is located in the chromosomal DNA (2, 38).