Prophage Induces Terminal Recombination in Escherichia coli by Inhibiting Chromosome Dimer Resolution: An Orientation-Dependent cis-Effect Lending Support to Bipolarization of the Terminus

Corresponding author: Jean-Michel Louarn, Laboratoire de Microbiologie et de Génétique Moléculaires du CNRS, 118 route de Narbonne, 31062 Toulouse Cedex, France., louarn{at}ibcg.biotoul.fr (E-mail)

Communicating editor: R. MAURER

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

A prophage inserted by homologous recombination near dif, the chromosome dimer resolution site of Escherichia coli, is excised at a frequency that depends on its orientation with respect to dif. In wild-type cells, terminal hyper- (TH) recombination is prophage specific and undetectable by a test involving deletion of chromosomal segments between repeats identical to those used for prophage insertion. TH recombination is, however, detected in both excision and deletion assays when dif, xerC, or ftsK mutations inhibit dimer resolution: lack of specialized resolution apparently results in recombinogenic lesions near dif. We also observed that the presence near dif of the prophage, in the orientation causing TH recombination, inhibits dif resolution activity. By its recombinogenic effect, this inhibition explains the enhanced prophage excision in wild-type cells. The primary effect of the prophage is probably an alteration of the dimer resolution regional control, which requires that dif is flanked by suitably oriented (polarized) stretches of DNA. Our model postulates that the prophage inserted near dif in the deleterious orientation disturbs chromosome polarization on the side of the site where it is integrated, because DNA, like the chromosome, is polarized by sequence elements. Candidate sequences are oligomers that display skewed distributions on each oriC-dif chromosome arm and on DNA.

THE terminus of the Escherichia coli chromosome may be the venue of two different recombination events: chromosome dimer (CD) resolution by XerCD recombinases acting at dif sites (KUEMPEL et al. 1991 ) and local homologous hyperrecombination. Terminal hyper- (TH) recombination was discovered and analyzed using an assay based on the excision of a prophage inserted between direct repeats located at various positions in the terminus. It affects a large region of several hundred kilobases, centered on the resolution site dif, and is a RecABC-dependent phenomenon (LOUARN et al. 1991 , LOUARN et al. 1994 ; CORRE et al. 1997 ). Strong polarization characterizes TH recombination, since the frequency of excisive exchanges was found to depend on the orientation of the tested sequences (CORRE et al. 1997 ). Orientations yielding frequent excisions are opposite on either side of dif, which suggests some interaction between TH recombination and CD resolution (CORRE et al. 1997 ).

CD resolution at dif is also subject to regional control. Catalyzed by XerCD recombinases (BLAKELY et al. 1997 ) and dependent on FtsK protein (STEINER et al. 1999 ), CD resolution between dif sites is coupled with cell division (STEINER and KUEMPEL 1998A ) and operates only when the sites are located in the terminus region (CORNET et al. 1996 ; KUEMPEL et al. 1996 ). We have shown recently (K. PERALS, F. CORNET, Y. MERLET and J. M. LOUARN, unpublished results) that resolutive exchanges at dif depend on the polarity of the region harboring dif since the orientation but not the genetic content of the segments flanking dif controls CD resolution. Resolution proficiency requires that the dif site be located between sufficiently large chromosome segments displaying opposite polarities. Since replichores, the chromosome arms corresponding to the regions normally replicated unidirectionally from oriC to dif (BLATTNER et al. 1997 ), display a polarization imprinted at the sequence level by numerous small skewed oligomers (SALZBERG et al. 1998 ), it is possible that regional polarization vis à vis CD resolution is exerted by the multiple oligomers distributed with an opposite skew on either side of dif.

In this article, additional correlations between TH recombination and CD resolution are described. We first confirmed that in the excision assay the orientation yielding TH recombination changes abruptly at dif. We then demonstrated that the sequences controlling this polarity are carried by the prophage. Crucial information stemmed from the use of a deletion assay. This prophage-free assay failed to reveal TH recombination in wild-type backgrounds. In contrast, TH recombination was consistently observed when CD resolution was inhibited, whatever the assay. Finally, we found that a prophage inserted close to dif inhibits CD resolution when its orientation favors TH recombination. This observation suggests strongly that it is the prophage-induced turn-off of CD resolution that turns on TH recombination. Inactivation of CD resolution by polar elements carried by the prophage emphasizes the role of chromosome polarization in the control of dif activity (K. PERALS, F. CORNET, Y. MERLET and J. M. LOUARN, unpublished results).

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Bacterial strains, plasmids, and bacteriophages:
All bacterial strains used in this work derive directly from strain CB0129 (F^- W1485 leu thyA deoB or C supE; BIRD et al. 1972 ). Therefore, a list giving the derivation of the 120 strains is not needed. Most tet or Tn10 insertions have been described previously or were constructed as described previously (LOUARN et al. 1994 ; CORRE et al. 1997 ). The recD::Ap^r mutation constructed for this work is derived from the recD1903::minitet of BIEK and COHEN 1986 , by (i) cloning this allele on plasmid pLN135, (ii) generating a NdeI deletion of the tet genes, (iii) tagging a 500-bp EcoRV deletion internal to recD by an Ap^r interposon (FELLAY et al. 1987 ), and (iv) reinserting the new allele into the chromosome using the integration-excision properties of vector pLN135, as described in CORNET et al. 1996 . This mutation confers a typical RecD^- phenotype, tested by all available genetic criteria including the absence of recognition of sites in crosses between suitable bacteriophage mutants (STAHL et al. 1980 ). Other mutations used are xerC2::Ap^r (LOUARN et al. 1994 ), xerCY17::Cm^r (COLLOMS et al. 1990 ), and ftsKa' (LIU et al. 1998 ; kindly provided by W. Donachie). Tagged insertions and mutations were moved by phage P1-mediated transduction. Phages TSK (REBOLLO et al. 1988 ) and TSKinv (this work) are presented in Fig 1. Phage TSKinv derives from TSK by several steps (not presented), the final result being the inversion of the TSK insertion between the EcoRI sites flanking the phage DNA, plus a silent 103-bp deletion just downstream of the J gene.

View larger version (23K): In this window In a new window Download PPT slide   Figure 1. Measuring indigenous recombination. (A) The phages TSK and TSKinv derive from phage vector 413L (KOURILSKY et al. 1978 ) and differ from each other by the orientation of the cloned material and by a 103-bp silent deletion right of gene J in TSKinv. The whole integration-excision system is deleted, the cI857 ind double mutant allele codes for a temperature-sensitive repressor insensitive to SOS induction, and the cII and cIII genes are wild type. These phages enter normally in lysogeny at 30°, but cannot integrate at the chromosomal att site. The cloned material is a 2.8-kb tet region coming from the transposon Tn10 and interspersed with a 2-kb Smr determinant and a 3.4-kb Knr determinant inserted within the Tcr tetA gene ~0.9 kb from each other. (B) Prophage integration by homologous recombination between the 2.8-kb tet regions present in prophage and chromosome. The double arrow indicates where integrative recombination must occur to yield a lysogen Ts Tcs Smr Knr. (C) The excision test. In a lysogenic cell as described above, a recombination event between the repeated sequences flanking the prophage results in prophage loss and recovery of a Tr phenotype. Due to the Smr and Knr insertions, excision events occurring between regions a leave a tes insertion in the chromosome, those occurring between regions b reconstitute the original tet insertion, and those occurring between regions c leave a tek insertion in the chromosome. (D) The deletion test. A strain has been created that harbors a tes and a tek insertion in direct repeats. The intervening chromosome segment must be nonessential, and the Knr and Smr insertions must be proximal with respect to the segment to be deleted. A recombination event between b regions can reconstitute the tet locus and delete the intervening segment with simultaneous loss of Knr and Smr markers.

Media and general methods:
Tryptone medium and Luria-Bertani (LB) rich medium are described in MILLER 1992 . Antibiotic concentrations used are: ampicillin (Ap), 25 µg/ml; chloramphenicol (Cm), 25 µg/ml; kanamycin (Kn), 25 µg/ml; streptomycin (Sm), 15 µg/ml; tetracycline (Tc), 15 µg/ml. All in vivo or in vitro genetic experiments were performed according to standard procedures (SAMBROOK et al. 1989 ; MILLER 1992 ).

Excision and deletion assays:
The general outlines may be found in FRANCOIS et al. 1987 and LOUARN et al. 1991 . Bacteria carrying the tet insertion of interest and plated in Tryptone soft agar are inoculated by a droplet of either TSK or TSKinv phage stocks and incubated at 30° overnight. Bacteria from the center of the lysed area are picked, subcloned on L-Kn plates, and incubated at 30°. Separate clones were tested for temperature sensitivity (Ts) and antibiotic resistances. Only Kn^r Sm^r Tc^s Ts clones are retained, after subcloning and retesting. They conform to the structure drawn in Fig 1C. The excision assay, carried out using -resistant derivatives, consists of the determination of the frequency of bacteria forming colonies on LB plates at 42°. In general, at least five independent colonies each containing between 5 x 10⁷ and 10⁸ bacteria were tested. Average values are reported.

The deletion assay requires additional steps:

1. The cured bacteria found above are tested for antibiotic resistances; the Kn^r Tr clones carry tek substitutions, the Sm^r Tr ones carry tes substitutions.

2. The tes and tek strains are constructed by transducing the tek insertion into a tes carrier. The resulting genetic organization must be as shown in Fig 1D, and this was frequently checked by Southern analysis.

3. The frequency of Tc^r derivatives is determined on three to five independent colonies (each ~5 x 10⁷ bacteria), and 20 of these derivatives are tested for antibiotic resistances. Since >90% of the Tc^r are Kn^s Sm^s, the frequency of deletion derivatives is assimilated to the frequency of Tc^r derivatives. Average values are reported.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The assays for chromosomal indigenous recombination:
Phenotypic detection of indigenous recombination requires repeated sequences. The repeats used here are derived from tetA, the tetracycline resistance gene (Tc^r) of transposon Tn10. The tes sequence carries a spectinomycin/streptomycin resistance (Sp^r/Sm^r) determinant inserted near the beginning of tetA, and the tek sequence carries a kanamycin resistance (Kn^r) determinant inserted near the end of tetA, the insertion sites being separated by ~0.9 kb (FRANCOIS et al. 1987 ). In the excision assay, a TSK or a TSKinv prophage (Fig 1A) has recombined into a chromosomal tet insertion (Fig 1B), so that the tes and tek sequences are separated by the prophage, which carries a temperature-sensitive mutant cI repressor. Prophage excision can be monitored by selecting for temperature-resistant (Tr) or Tc^r derivatives (Fig 1C). The excision assay was used exclusively in our earlier studies (LOUARN et al. 1991 , LOUARN et al. 1994 ; CORRE et al. 1997 ). In the deletion assay (Fig 1D), the tes and tek sequences are separated by chromosomal DNA that must be nonessential. Deletion of this material is monitored by selecting for Tc^r derivatives, with concomitant loss of the Sm^r/Sp^r and Kn^r determinants. Occurrence of deletions in the terminus is not limited by loss of function since the whole region is deletable (HENSON and KUEMPEL 1985 ). Both prophage excisions and chromosomal deletions involve recombination between the same repeated sequences identically positioned (the Sm^r/Sp^r and Kn^r determinants being proximal to the segment to be lost; Fig 1C and Fig D). Thus, both assays should yield the same results, unless the eliminated sequences participate in or interfere with recombination.

Prophage orientation leading to TH recombination changes abruptly at dif and this polarity does not implicate RecD:
The previous observation (CORRE et al. 1997 ) that prophage orientation with respect to dif controls TH recombination intensity has been confirmed and extended. We measured the frequency of bacteria cured of the prophage in a number of lysogens, all wild type for the CD resolution system and differing only in the location of the resident tet segment in the vicinity of dif. At most locations, this segment was inserted in either orientation. The data are reported in Fig 2. TH recombination is maximum in the vicinity of dif and displays clear-cut orientation dependence, one orientation, designated V by convention, yielding many more cured bacteria than the other, designated M. Strikingly, a given orientation of the resident tet results in the V behavior on one side of dif and in the M behavior on the other side. For instance, at position zdc338, 7 kb to the left of dif, the prophage inserted in the tet_ccw insertion is frequently lost, whereas at position hipA, ~1 kb to the right of dif, the same orientation yields low excision frequency.

View larger version (27K): In this window In a new window Download PPT slide   Figure 2. Prophage loss in wild-type and RecD- backgrounds. Semi-log plot of cured bacteria frequency vs. map position of the tet integration locus. In contrast with our previous reports, we have not calculated excision frequencies per cell generation since selective values of lysogens vs. nonlysogens have not been reestimated. Map positions are in minutes (RUDD 1998 ). Arrowheads, dotted or open, refer to the orientation of the tet unique site in wild-type strains, as indicated in the insets. Closed arrowheads refer to RecD- mutants. A vertical dotted line joins insertions at the same locus but in opposite orientations. Some measurements are from CORRE et al. 1997 . Curves are drawn symmetric with respect to a vertical axis at dif. tet or Tn10 insertions (distances in kilobases from dif, negative to its left, positive to its right, and cloning sites of tet insertions are indicated within brackets): (1) zda192 [-151; BamHI], cw and ccw; (2) zdc310 [-35.3; BamHI], cw and ccw; (3) zdc330 [-15.8; Ecl136], cw and ccw; (4) zdc338 [-7.0; BglII], cw and ccw; (5) hipA [+1.0; EcoRV], cw and ccw; (6) zdd355 [+8.9; BamHI], cw and ccw; (7) zdd365 [+17.8; BglII], ccw; (8) zdd370 [+23.2; BamHI], cw and ccw; (9) zde381 [+34], Tn10cw; (10) zde395 [+48], Tn10cw; (11) zde406 [+58], Tn10ccw; (12) zdf237 [+100], Tn10cw; (13) zdg232 [+125] Tn10cw.

TH recombination is abolished when the RecBC system is inactivated, and our previous model for its polarity control pointed to the possible roles of RecD and sites present in the tet repeats (CORRE et al. 1997 ). Polarity was proposed to be due to the modulation at the tes and tek sites of the exonuclease activities of the RecBCD complex acting systematically in the dif-to-oriC direction. Since these exonuclease activities depend on RecD⁺ activity (KOWALCZYKOWSKI et al. 1994 ), we tested the effect of a recD null mutation on TH recombination. The orientation dependence was totally conserved in the RecD^- background (Fig 2). Thus, direct tests failed to detect any indication that TH recombination polarity depends on a RecBCD mode of action. It should be noted that the prophage carries the SOS-insensitive allele of cI repressor, so that the excision assay is not affected by SOS induction.

The prophage is responsible for the orientation dependence of TH recombination:
To determine whether the elements controlling the polarity of TH recombination map in the prophage or in the flanking repeats, we have constructed a new recombinant phage, named TSKinv, in which the cloned TSK sequences are inverted with respect to the phage sequences (Fig 1A). Prophage excision assays repeated with TSKinv yielded clear and unambiguous results (Table 1). Again, one orientation yielded many more cured bacteria than the other, and the determining factor appeared clearly to be the orientation of DNA with respect to dif, not the orientation of the tes and tek sequences. Thus, although excisive recombination requires the latter sequences, its frequency is controlled by polar determinants carried by the prophage. This crucial observation led us to reject the model, already weakened by the absence of a role for RecD, that polarity of TH recombination is determined by the properties of recombination enzymes. Why only a single orientation of the prophage triggers local action of RecBC(D) appeared an enigma.

View this table: In this window In a new window   Table 1. Frequencies of cured bacteria in wild-type lysogens as a function of DNA orientation

Deletion assays by tes and tek recombination fail to detect TH recombination:
A comparative analysis of deletion frequencies resulting from exchanges between tes and tek insertions inside and outside the terminus has confirmed the role of sequences in TH recombination. Five segments of the terminus not containing dif and bordered by tes and tek sequences in the two possible orientations were assayed (Fig 3A). Deletions, scored as Tc^r Sm^s Km^s derivatives, constituted on average >95% of the Tc^r indigenous recombinants, and none of these deletions had noticeable effects on growth rates in LB medium. Southern analyses, when performed, always confirmed the rearrangement (data not shown). The results reported in Fig 3B (first two lines) indicate that deletions were much less frequent than excision of V-oriented prophages in the region, and that their frequencies displayed no marked dependence on the orientation of the tes and tek sequences, ranging between 2 x 10^-4 and 2 x 10^-3. In RecBC^- conditions (data not shown), deletion events were at least 10-fold less frequent, indicating the strong contribution of the RecABCD pathway to these rearrangements also observed in other systems (GALITSKI and ROTH 1997 ).

View larger version (33K): In this window In a new window Download PPT slide   Figure 3. Occurrence of deletions in the vicinity of dif. (A) The segments tested are indicated by hatched rectangles. They are bordered by tes and tek insertions as presented in Fig 1D. The arrowheads indicate the orientations of these insertions, using the convention of Fig 2. For a given segment, the first number refers to ccw orientation and the second to cw orientation. Segments (numbers within brackets indicate their sizes in kilobases): (1) zci1315-zci1325 (10), includes the trp operon, the external control; (2 and 3) zdc310-zdc338 [28]; (4 and 5) zdc330-zdc338 [9]; (6 and 7) hipA-zdd355 [8]; (8 and 9) hipA-zdd370 [22]; (10 and 11) zdd355-zdd370 [14]. (B) Occurrence of deletions in different genetic backgrounds. In each case, the upper lines refer to ccw orientation, the lower lines to cw orientation. First and second lines, wild-type background; third and fourth, xerC17::Cmr mutants; fifth and sixth, recD::Apr mutants; seventh and eighth, xerC17::Cmr recD::Apr double mutants. In all cases, deletion and parental strains displayed similar growth abilities, at least on LB agar medium.

For comparison, we have measured prophage excision frequency and deletion frequency in the trp operon region at 28 min, taking advantage of three Tn10 insertions, one in trpB at 1317 kb of genomic E. coli sequence (BLATTNER et al. 1997 ) and the two others flanking the trp operon at 1315 and 1325 kb (Fig 3A). This 10-kb region is clearly outside the region affected by TH recombination, since in excision tests bacteria cured of a TSK prophage inserted in trpB::Tn10 (orientation V) or in any of the flanking Tn10 (orientation M) were in all cases found at a low frequency of 2–5 x 10^-4. Tc^r recombinant derivatives of strains carrying zci1315::Tn10-tes and zci1325::Tn10-tek were somewhat more frequent (Fig 3B) and 85% were deletions, displaying a Trp^- Sm^s Kn^s phenotype. Thus, the deletion frequency in the control region is very similar to that observed in the dif region.

When the TSK prophage was recombined into any of the deletions in the terminus described above, we again observed an orientation-dependent TH recombination: these deletions do not affect the recombination events specific to the terminus (data not shown). In bacteria genetically proficient for homologous and dif-specific recombination, TH recombination is thus detected only with the excision assay, and only with a certain prophage orientation with respect to dif. The data do not support the conclusion that a high frequency of homologous recombination affects the terminus of wild-type chromosomes.

Inactivation of the dif/XerCD system always turns on TH recombination:
Crossing the xer2::Ap^r allele into the lysogenic strains presented in Fig 2 resulted in strains displaying constitutive TH recombination. As in wild-type bacteria, the dif region was the venue of the highest frequencies of cured bacteria (Fig 4), and these excision events were controlled largely by the RecBC pathway (data not shown). The major difference from wild-type cells was the absence of polarity, especially when the prophages were inserted in the vicinity of dif, where both prophage orientations gave rise to similar excision rates. However, some polarity was detected when the prophage was located away from dif, in regions where TH recombination is much less active. In these cases, the frequencies of cured bacteria were lower for the M orientation than for the V.

View larger version (19K): In this window In a new window Download PPT slide   Figure 4. Prophage loss in the XerC- or (dif) backgrounds. Same conventions as in Fig 2. Insertions: (1) zda192, cw and ccw; (2) zdc310, cw and ccw; (3) zdc330, cw and ccw; (4) zdc338, cw and ccw; (5) (dif58), cw and ccw (data from CORRE et al. 1997 ); (6) hipA, cw and ccw; (7) zdd355, cw and ccw; (8) zdd365, ccw; (9) zdd370, cw and ccw; (10) zde395, cw; (11) zde406, ccw.

CD resolution deficiency (here due to the xerCY17::Cm^r allele) also results in a considerable increase in deletion frequencies (between 4 and 50 times) for segments of the terminus (Fig 3B, lines 3 and 4). However, the observed values remain lower than those for V-oriented prophage excision. The effects of RecD inactivation on deletion frequency have also been studied (Fig 3B, lines 5 and 6). An increase in the terminus region was easily detected. This effect is cumulative with that of XerC inactivation, and in XerC^- RecD^- double mutants, bacteria deleted for a segment of the terminus represented several percent of the population (Fig 3B, lines 7 and 8), becoming about as frequent as cured bacteria in excision tests conducted on similar double mutants [strain (dif)::tetTSK recD::Ap^r; data not shown]. In striking contrast, the deletion frequency in the trp region remained unaffected by these mutations (Fig 3B, lane 1).

Recently, it has been shown that the cytoplasmic domain of FtsK protein is involved in CD resolution (STEINER et al. 1999 ). We have tested the effect on TH recombination of a mutation eliminating the FtsK cytoplasmic domain. As reported in Table 2, this mutation increases the rates of excision of a prophage in M orientation and of deletion of a segment near dif to the same extent as a XerC^- mutation (compare with Fig 3 and Fig 4) and has no effect at positions remote from dif (lacZ, trp).

In conclusion, the terminus region is the venue of intense homologous recombination when CD resolution is inactivated, an effect also elicited by concomitant inactivation of RecD in the deletion assay. Under these conditions, the region surrounding dif may be considered as a fragile area, highly susceptible to recombinogenic damage.

Prophage in the V orientation near dif weakens the resolution activity:
This novel phenomenon provides the clue for understanding the role of prophage in TH recombination. We were intrigued by the high frequency of cured bacteria when the prophage TSK is inserted in hipA::tetcw [the V orientation; designated hipA::TSK(V) hereafter], which is the tet insertion closest to dif (Fig 2). This frequency was clearly above that measured at neighboring positions, and this difference between prophage positions was not observed when the strains were XerC^- (Fig 4). Microscopical observation of the hipA::TSK(V) bacteria revealed a high proportion of long bacteria, bacterial chains, and filaments, with abnormalities in nucleoid distribution (Fig 5C), which were reminiscent of the phenotype of dif/xerC mutants (Fig 5B).

View larger version (49K): In this window In a new window Download PPT slide   Figure 5. Microscopic observations of bacteria undergoing CDR attenuation. Bacteria exponentially growing in LB medium at 30° were stained with DAPI and examined under combined fluorescence and phase-contrast microscopy, using a Leica DMRB microscope. Images were recorded with a CCD camera and processed by use of Adobe Photoshop LE software. (A) strain LN3542, hipA::tetcw (wild-type); (B) strain LN2911, (dif58)::tet (CD resolution-minus); (C) strain LN3561, hipA::tetcwTSK (prophage in orientation V); (D) strain LN3588, hipA::tetccwTSK (prophage in orientation M); (E) strain LN3609, zdc330::tetccwTSK (prophage in orientation V). Magnification, x2450.

No difference was detected under the microscope between hipA::TSK(V) xerC lysogens and their nonlysogen ancestors (data not shown), and competition experiments (co-culture of both strains for 40 generations in LB medium at 30°) indicated that they display similar generation times. In this Xer^- background, the prophage exerts no deleterious influence. This suggests that the negative effect under Xer⁺ conditions of this particular prophage insertion is not an alteration of the expression of nearby genes (in case of insertions in hipA, this was a concern since this gene may have a function related to cell cycle control; BLACK et al. 1994 ). The presence of the prophage near dif seems rather to inhibit the CD resolution function. We named this phenomenon CDR attenuation.

CDR attenuation was not observed when the prophage has recombined into a tet sequence inserted in hipA at the very same position but in the opposite orientation [the hipA::TSK(M) strain, which does not undergo TH recombination; Fig 2]: very few abnormal cells were detectable (Fig 5D). When the prophage TSKinv was recombined into the same tet insertions, again CDR attenuation was observed in one orientation only and again when the prophage was oriented so as to induce hyperrecombination in the excision assay (data not shown). Thus, CDR attenuation appears to be controlled by sequences in the same polar fashion as TH recombination. We have extended these observations to prophage insertions at other positions. Another clear example of CDR attenuation was observed when prophage TSK was recombined into a tet pif insertion tagging a (dif) deletion (pif is a hybrid between the resolution site of the pSC101 plasmid and dif which substitutes nearly perfectly for dif in this construction; CORNET et al. 1994 ), and the prophage was in orientation V with respect to the nearby pif. We also detected chains and filaments in cultures of other lysogenic strains displaying TH recombination (Fig 5E) and very few of them in cultures of strains harboring the prophage in M orientation (not shown). Thus, CDR attenuation is associated with the presence of the prophage in V orientation, but the frequency of abnormal cells decreases when the distance between the prophage and the dif site increases (data not shown). This explains why the phenomenon was previously unnoticed.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Our results lead to a better understanding of the causes of TH recombination. The data show absence of homologous hyperrecombination in the terminus when resolution of chromosome dimers occurs normally. TH recombination is observed in two situations: when mutations in the dif/XerCD/FtsK system hinder dimer resolution, and when a prophage is inserted in a certain orientation in the region. Since we also observed that presence of the prophage in the same orientation tends to inhibit CD resolution, a unifying explanation is that TH recombination in the excision assay is a direct consequence of CDR attenuation by the prophage.

Additional information on CD resolution explains how its inactivation may trigger TH recombination. CD resolution occurs late in the cell cycle, at about division time and long after completion of chromosome replication (STEINER and KUEMPEL 1998A , STEINER and KUEMPEL 1998B ). The recent observation that FtsK controls CD resolution (STEINER et al. 1999 ) supports the view that the various components of the process are coordinated in the peculiar environment of the division plane, since FtsK is preferentially located at the septum (WANG and LUTKENHAUS 1998 ; YU et al. 1998 ). At division time, the nucleoid has been largely rebuilt, so that the chromosome dimer is distributed in two bodies, each corresponding to a condensed and partitioned chromosome. There must be two DNA stretches running between these nucleoid bodies through the plane of the growing septum. Recently we (K. PERALS, F. CORNET, Y. MERLET and J. M. LOUARN, unpublished results) observed that a control is exerted over CD resolution by virtue of a polarization in opposite directions of the regions flanking dif. We proposed that this polarization, imprinted in the chromosome sequence, serves to position the dif loci in the septum plane so that their recognition by XerC/XerD/FtsK proteins is facilitated. The subsequent resolution provides two chromosome monomers and allows the cell division process to continue to completion in wild-type bacteria. In bacteria unable to carry out CD resolution, the dif regions are trapped by the septum at division (assuming that the positioning process is independent of the resolution machinery). Eventually, a process affecting the trapped DNA generates targets for the RecBCD complex (TH recombination requires RecABCD proficiency; CORRE et al. 1997 ). Perhaps hydrodynamic shearing generates double-strand breaks (DSBs; Fig 6A), or the next replication forks arriving at the trapped zone are transformed in Holliday junctions with generation of dsDNA ends as described by SEIGNEUR et al. 1998 . In this model, TH recombination does not necessarily ensue from RecBCD action. If the trapped dimer is evenly distributed between the Siamese twin cells, no segment of the terminus region can be present twice in one twin (and absent from the other), so that homologous recombination cannot reconstitute a complete free circular chromosome. Only DNA degradation can occur, and these twins are doomed since they cannot cope with the chromosome lesion. Hence the phenotype of xerC/xerD/ftsK/(dif) mutants is low viability, apparent increase of generation time, and strong cell size heterogeneity with a majority of normal cells frequently mixed with filaments or chains. TH recombination with reconstitution of a viable chromosome is a happier end, which may come about when the distribution of dimer DNA between the Siamese twins is uneven, with one twin receiving more than one genome. Recombination may restore a free circular chromosome in such twins.

View larger version (42K): In this window In a new window Download PPT slide   Figure 6. The link between inhibition of CD resolution and induction of TH recombination. Models. (A) Inhibition of CD resolution may provide sites of entry for RecBCD in the terminus. The possible situation just after division in a xerC mutant harboring a chromosome dimer is illustrated. Normal positioning of dif sites is postulated. The septum just formed between the sister cells has pinched the dimeric DNA abreast of these sites. In some way, a free DNA end (here a DSB for easy drawing) is generated near the trapped zone, to be the target for future RecBCD action. In this case, the even DNA distribution precludes eventual recombinational repair. (B) CDR attenuation by the prophage. The prophage is drawn inserted in V orientation, with its putative polar elements interfering with the ambient chromosome polarity. The positioning process has located the junction under the septum between oppositely polarized phage and chromosome but not the dif sites. These sites are no longer accessible to their enzymatic partners, hence inhibition of dimer resolution occurs. (C) When TH recombination occurs in the absence of dimer resolution, one of the DNA stretches passing through the septum in this V-lysogen is pinched in the dif region, and the second at the prophage border distal from dif. The result is that the twin on the right harbors two copies of the region between dif and the prophage distal end. This terminal duplication will favor recombinational repair and reformation of a circular chromosome when a site of entry for RecBCD is provided. Prophage excision is an obvious possible outcome of the repair event. In the left twin, the absence of terminal duplication makes repair impossible. This cell will filament and die.

The observation that the presence near dif of the TSK prophage in a certain orientation attenuates CD resolution may be connected to observations made in this laboratory that inversions of chromosomal segments located near dif may hinder the CD resolution activity of dif (K. PERALS, F. CORNET, Y. MERLET and J. M. LOUARN, unpublished results). A coherent interpretation for CDR attenuation by the prophage is that the polar elements carried by DNA are functionally similar to those generating the dif activity zone (DAZ). When the prophage is inserted so that its polar elements are in the same orientation as the surrounding chromosomal ones, no interference is observed: the DAZ is formed normally, CD resolution may occur at dif, and no hyperrecombination is induced in the region. On the contrary, prophage insertion in the opposite orientation generates a disturbance of the DAZ similar to that induced by deleterious inversions. As CD resolution is turned off, the conditions for TH recombination are created. This model is presented in Fig 6B.

The model that the distribution of specific polar sequences controls dif activity by determining the cellular location of the site also has the virtue of explaining why in dif/xerC mutants, in which TH recombination is constitutive, deletion frequencies are lower than excision frequencies. When the prophage present in V orientation disturbs the positioning process, the situation depicted in Fig 6C may be frequent: normal positioning of one DNA stretch running between the segregated nuclear bodies (dif under the septum) and aberrant positioning of the second stretch (the junction between the prophage and the chromosome under the septum). In these circumstances, the DNA is not evenly distributed, and one twin receives more DNA than the other. DSBs are in consequence easily repaired by a RecBCD-dependent event in this twin, hence TH recombination. In contrast, the inversion of polarity occurs normally near dif in bacteria undergoing tes and tek deletion events or harboring the prophage in M orientation. In these cells, the fragile regions trapped under the septum are the dif regions, and the dimer DNA is evenly distributed between the twins, as drawn in Fig 6A. This, as mentioned above, does not help recombinational repair of DSBs. There should be some unevenness in the dimer DNA distribution, otherwise no terminal recombination could occur. However, redundancies under these conditions might be too small to favor efficient recombination, except perhaps when RecD inactivation permits the use of ends as recombinogenic material, which could explain the elicitor effect of recD mutation (Fig 3B). One important postulate in this model is that the mechanism positioning dif in the septal plane must operate independently from XerCD/dif activity. Interestingly, TH recombination was found to have similar characteristics in xerC and ftsK mutants. This suggests that the positioning of the dif region does not involve FtsK, in spite of the homology of part of this protein with the Bacillus subtilis SpoIIIE protein, which is involved in nucleoid migration towards the prespore (WU and ERRINGTON 1994 ).

The sequences that generate a functional polarity on either side of dif have not yet been determined. Such sequences might be distributed with a preferential orientation along the regions flanking dif, this orientation being inverted at dif. Sequences of this type exist. SALZBERG et al. 1998 have described several families of skewed octamers along each replichore from oriC to dif. Among these, the degenerate motif RRRAGGGY, where R is any purine and Y any pyrimidine, would be a suitable candidate. It is found eight times in the 40 kb leftward of dif, all in the same orientation, and nine times in the 40 kb to the right of dif, all but one in inverse orientation. The motif is not present within an 8-kb region to the immediate left of dif, whose inversion has no effect on dif resolution activity (K. PERALS, F. CORNET, Y. MERLET and J. M. LOUARN, unpublished results). In the sequence, six copies of the motif in the same orientation are present to the left of gene J (Fig 1A), i.e., in that half of the prophage that is distal to dif when the prophage induces TH recombination. In this situation, the RRRAGGGY motifs carried by this prophage half are polarized in opposite orientation to the motifs of the chromosome, resulting in an arrangement similar to that normally present on either side of dif. This could disturb the positioning at division of the dif sites carried by a chromosome dimer. Whether the RRRAGGGY motif is actually involved in determining the dif activity zone is presently an open question, but at least its distribution in the terminus and in DNA makes our model plausible.

Finally, the role proposed here for the numerous sequences oppositely oriented on each side of dif may not be limited to the terminus region. Most sequence skews discovered along the chromosome seem to invert at oriC and dif (SALZBERG et al. 1998 ). It has been suggested that some of these skews may be relevant to replication (LOBRY 1996 ; SALZBERG et al. 1998 ). Some may favor chromosome repair after replication fork breakage, as proposed for the skewed distribution of sites (KUZMINOV 1995 ). Some may be consequences of the preferred gene orientation along the chromosome arms, as suggested also for sites by BIAUDET et al. 1998 and COLBERT et al. 1998 . The present work leads to the prediction that some of the skewed sequences may also be involved in nucleoid condensation and positioning of DNA within the cell. Such a role could be important not only after replication, as illustated here, but throughout the replication process by helping chromosomal DNA to move through replisomes that are probably at fixed positions within the cell (LEMON and GROSSMAN 1998 ).

	ACKNOWLEDGMENTS

We are grateful to Agammenon Carpoussis and David Lane for careful reading and correction of the manuscript. We thank K. Perals and F. Cornet for constant cooperation. This work was supported by the Association pour la Recherche sur le Cancer (contract ARC no. 9823).

Manuscript received May 24, 1999; Accepted for publication September 17, 1999.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

BIAUDET, V., M. EL KAROUI, and A. GRUSS, 1998  Codon usage can explain GT-rich islands surrounding Chi sites on the Escherichia coli genome. Mol. Microbiol. 29:666-669[Medline].

BIEK, D. P. and S. N. COHEN, 1986  Identification and characterization of recD, a gene affecting plasmid maintenance and recombination in E. coli.. J. Bacteriol. 167:594-603[Abstract/Free Full Text].

BIRD, R. E., J. LOUARN, J. MARTUSCELLI, and L. CARO, 1972  Origin and sequence of chromosome replication in Escherichia coli.. J. Mol. Biol. 70:549-566[Medline].

BLACK, D. S., B. IRWIN, and H. S. MOYED, 1994  Autoregulation of hip, an operon that affects lethality due to inhibition of peptidoglycan or DNA synthesis. J. Bacteriol. 176:4081-4091[Abstract/Free Full Text].

BLAKELY, G. W., A. O. DAVIDSON, and D. J. SHERRATT, 1997  Binding and cleavage of nicked substrates by site-specific recombinases XerC and XerD. J. Mol. Biol. 265:30-39[Medline].

BLATTNER, F. R., G. PLUNKETT, III, C. A. BLOCH, N. T. PERNA, and V. BURLAND et al., 1997  The complete genome sequence of Escherichia coli K12. Science 277:1453-1474[Abstract/Free Full Text].

COLBERT, T., A. F. TAYLOR, and G. R. SMITH, 1998  Genomics, chi sites and codons: "islands of preferred DNA pairing" are oceans of ORFs. Trends Genet. 14:485-488[Medline].

COLLOMS, S. D., P. SYKORA, G. SZATMARI, and D. J. SHERRATT, 1990  Recombination at ColE1 cer requires the Escherichia coli xerC gene product, a member of the lambda integrase family of site-specific recombinases. J. Bacteriol. 172:6973-6980[Abstract/Free Full Text].

CORNET, F., I. MORTIER, J. PATTE, and J. M. LOUARN, 1994  Plasmid pSC101 harbors a recombination site, psi, which is able to resolve plasmid multimers and to substitute for the analogous chromosomal Escherichia coli site dif.. J. Bacteriol. 176:3188-3195[Abstract/Free Full Text].

CORNET, F., J. LOUARN, J. PATTE, and J. M. LOUARN, 1996  Restriction of the activity of the recombination site dif to a small zone of the Escherichia coli chromosome. Genes Dev. 10:1152-1161[Abstract/Free Full Text].

CORRE, J., F. CORNET, J. PATTE, and J. M. LOUARN, 1997  Unraveling a region-specific hyper-recombination phenomenon: genetic control and modalities of terminal recombination in Escherichia coli.. Genetics 147:979-989[Abstract].

FELLAY, R., J. FREY, and H. K. KRISCH, 1987  Interposon mutagenesis of soil and water bacteria: a family of DNA fragments designed for in vitro insertional mutagenesis of Gram-negative bacteria. Gene 52:147-154[Medline].

FRANCOIS, V., J. LOUARN, and J. M. LOUARN, 1987  A system for in vivo selection of genomic rearrangements with predetermined endpoints in Escherichia coli using modified Tn10 transposons. Gene 56:99-108[Medline].

GALITSKI, T. and J. R. ROTH, 1997  Pathways of homologous recombination between chromosomal direct repeats in Salmonella typhimurium.. Genetics 146:751-767[Abstract].

HENSON, J. M. and P. L. KUEMPEL, 1985  Deletion of the terminus region (340 kilobase pairs of DNA) from the chromosome of Escherichia coli.. Proc. Natl. Acad. Sci. USA 82:3766-3770[Abstract/Free Full Text].

KOURILSKY, P., M. PERRICAUDET, D. GROS, A. GARAPIN, and M. GOTTESMAN et al., 1978  Description and properties of bacteriophage lambda vectors used for the cloning of EcoRI fragments. Biochimie 60:183-187[Medline].

KOWALCZYKOWSKI, S. C., D. A. DIXON, A. K. EGGLESTON, S. D. LAUDER, and W. M. REHRAUER, 1994  Biochemistry of homologous recombination in Escherichia coli.. Microbiol. Rev. 58:401-465[Abstract/Free Full Text].

KUEMPEL, P. L., J. M. HENSON, L. DIRCKS, M. TECKLENBURG, and D. F. LIM, 1991  dif, a recA-independent recombination site in the terminus region of Escherichia coli.. New Biologist 3:799-811[Medline].

KUEMPEL, P. L., A. HOGGARD, M. NIELSEN, O. NAGAPAN, and M. TECKLENBURG, 1996  Use of a transposon Tndif to obtain suppressing and non-suppressing insertions of the dif resolvase site of Escherichia coli.. Genes Dev. 10:1162-1171[Abstract/Free Full Text].

KUZMINOV, A., 1995  Breakage of arrested replication forks in E. coli.. BioEssays 17:733-741[Medline].

LEMON, K. P. and A. D. GROSSMAN, 1998  Localization of bacterial DNA polymerase: evidence for a factory model of replication. Science 282:1516-1519[Abstract/Free Full Text].

LIU, G., G. C. DRAPER, and W. D. DONACHIE, 1998  FtsK is a bifunctional protein involved in cell division and chromosome localization in Escherichia coli.. Mol. Microbiol. 29:893-903[Medline].

LOBRY, J., 1996  Asymmetric substitution patterns in the two DNA strands of bacteria. Mol. Biol. Evol. 13:660-665[Abstract].

LOUARN, J., F. CORNET, V. FRANCOIS, J. PATTE, and J. M. LOUARN, 1994  Hyperrecombination in the terminus region of the Escherichia coli chromosome: possible relation to nucleoid organization. J. Bacteriol. 176:7524-7531[Abstract/Free Full Text].

LOUARN, J. M., J. LOUARN, V. FRANCOIS, and J. PATTE, 1991  Analysis and possible role of hyperrecombination in the termination region of the Escherichia coli chromosome. J. Bacteriol. 173:5097-5104[Abstract/Free Full Text].

MILLER, J. H., 1992 A Short Course in Bacterial Genetics. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

REBOLLO, J. E., V. FRANCOIS, and J. M. LOUARN, 1988  Detection and possible role of two large nondivisible zones on the Escherichia coli chromosome. Proc. Natl. Acad. Sci. USA 85:9391-9395[Abstract/Free Full Text].

RUDD, K. E., 1998  Linkage map of Escherichia coli K12, edition 10: the physical map. Microbiol. Mol. Biol. Rev. 62:985-1019[Abstract/Free Full Text].

SALZBERG, S. T., A. J. SALZBERG, A. R. KERLAVAGE, and J. T. TOMB, 1998  Skewed oligomers and origins of replication. Gene 217:57-67[Medline].

SAMBROOK, J., E. F. FRITSCH and T. MANIATIS, 1989 Molecular Cloning: A Laboratory Manual, Ed. 2. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

SEIGNEUR, M., V. BIDNENKO, S. D. EHRLICH, and B. MICHEL, 1998  RuvAB acts at arrested replication forks. Cell 95:419-430[Medline].

STAHL, F. W., M. M. STAHL, R. E. MALONE, and J. M. CRASEMAN, 1980  Directionality and nonreciprocality of chi-stimulated recombination in phage lambda. Genetics 94:235-248[Abstract/Free Full Text].

STEINER, W., G. LIU, W. D. DONACHIE, and P. KUEMPEL, 1999  The cytoplasmic domain of FtsK is required for resolution of chromosome dimers. Mol. Microbiol. 31:579-583[Medline].

STEINER, W. W. and P. L. KUEMPEL, 1998a  Cell division is required for resolution of dimer chromosomes at the dif locus of Escherichia coli.. Mol. Microbiol. 27:257-268[Medline].

STEINER, W. W. and P. L. KUEMPEL, 1998b  Sister chromatid exchange frequencies in Escherichia coli analyzed by recombination at the dif resolvase site. J. Bacteriol. 180:6269-6275[Abstract/Free Full Text].

WANG, L. and J. LUTKENHAUS, 1998  FtsK is an essential cell division protein that is localized to the septum and induced as part of the SOS response. Mol. Microbiol. 29:731-740[Medline].

WU, L. J. and J. ERRINGTON, 1994  Bacillus subtilis SpoIIIE protein required for DNA segregation during asymmetric cell division. Science 264:572-575[Abstract/Free Full Text].

YU, X. C., A. H. TRAN, Q. SUN, and W. MARGOLIN, 1998  Localization of cell division protein FtsK to the Escherichia coli septum and identification of a potential N-terminal targeting domain. J. Bacteriol. 180:1296-1304[Abstract/Free Full Text].

This article has been cited by other articles:

				E. K. O'Reilly and K. N. Kreuzer Isolation of SOS Constitutive Mutants of Escherichia coli J. Bacteriol., November 1, 2004; 186(21): 7149 - 7160. [Abstract] [Full Text] [PDF]

				E. P. C. Rocha The replication-related organization of bacterial genomes Microbiology, June 1, 2004; 150(6): 1609 - 1627. [Abstract] [Full Text] [PDF]

				J. Corre and J.-M. Louarn Evidence from Terminal Recombination Gradients that FtsK Uses Replichore Polarity To Control Chromosome Terminus Positioning at Division in Escherichia coli J. Bacteriol., July 15, 2002; 184(14): 3801 - 3807. [Abstract] [Full Text] [PDF]

				M. I. Guijo, J. Patte, M. del Mar Campos, J.-M. Louarn, and J. E. Rebollo Localized Remodeling of the Escherichia coli Chromosome: The Patchwork of Segments Refractory and Tolerant to Inversion Near the Replication Terminus Genetics, April 1, 2001; 157(4): 1413 - 1423. [Abstract] [Full Text]