Localized Remodeling of the Escherichia coli Chromosome: The Patchwork of Segments Refractory and Tolerant to Inversion Near the Replication 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

The behavior of chromosomal inversions in Escherichia coli depends upon the region they affect. Regions flanking the replication terminus have been termed nondivisible zones (NDZ) because inversions ending in the region were either deleterious or not feasible. This regional phenomenon is further analyzed here. Thirty segments distributed between 23 and 29 min on the chromosome map have been submitted to an inversion test. Twenty-five segments either became deleterious when inverted or were noninvertible, but five segments tolerated inversion. The involvement of polar replication pause sites in this distribution was investigated. The results suggest that the Tus/pause site system may forbid some inversion events, but that other constraints to inversion, unrelated to this system, exist. Our current model for deleterious inversions is that the segments involved carry polar sequences acting in concert with other polar sequences located outside the segments. The observed patchwork of refractory and tolerant segments supports the existence of several NDZs in the 23- to 29-min region. Microscopic observations revealed that deleterious inversions are associated with high frequencies of abnormal nucleoid structure and distribution. Combined with other information, the data suggest that NDZs participate in the organization of the terminal domain of the nucleoid.

TWO sites on the circular Escherichia coli chromosome play a major role during the cell cycle: oriC, where replisomes are assembled for bidirectional replication (MESSER and WEIGLER 1996 , for review), and the diametrically opposite dif site, where chromosome dimers, resulting from an odd number of exchanges between sister chromosomes, are resolved (KUEMPEL et al. 1991 ; STEINER and KUEMPEL 1998 ). The existence of a long-range nucleoid organization on each of the chromosome arms joining these sites has been long suspected (LOUARN et al. 1982 ; REBOLLO et al. 1988 ) and is now supported by an increasing body of evidence. Cytological studies have shown that origin and terminus regions move within the cell with different choreography as nucleoids are replicated and partitioned, suggesting the existence of two macrodomains in the nucleoid, the Ori and Ter domains (NIKI et al. 2000 ). Factors involved in the organization into macro-domains have not yet been directly approached. However, the region corresponding to the Ter domain is characterized by two phenomena that might contribute to this organization: a regional control on the activity of the dif site that probably involves chromosome migration within the cell and the existence of nondivisible zones (NDZs) whose disruption by inversion is deleterious.

Genetic studies of requirements for dif activity have revealed that dimer resolution is efficient only when dif is located within a limited region between replication arms, named DAZ for Dif activity zone, and when the extensive regions flanking the site remain in their natural orientation with respect to replication (CORRE et al. 2000 ; PERALS et al. 2000 ). The orientation requirement is presumably related to the mechanism that positions sister dif sites near the growing septum where resolution is supposed to occur (STEINER et al. 1999 ). Our data suggested that multiple oriented short sequences act in a concerted manner to locate the DAZ at the right place (in the septal plane) and at the right moment of the cell cycle (when the septum closes) and that such sequence elements generate a functional long-range polarization of a significant fraction of each chromosome arm on either side of dif. The functional polarization of the terminus region in opposite directions from dif may be correlated with the fact that a variety of small sequences are preferentially found on the chromosome strands running 5' to 3' from oriC to dif. First described by SALZBERG et al. 1998 , skewed oligomers of the RRNAGGGS family (R = purine, N = any base, S = G or C) display all topographical requirements expected for motifs determining dif activity, though direct proof of their involvement is still lacking (CAPIAUX et al. 2001 ).

KONRAD 1977 was first to notice that the segment between the lac operon and att80 on the E. coli chromosome (a region we further analyze here) is refractory to inversion. REBOLLO et al. 1988 extended this observation and analyzed the consequences of inversion of many segments covering the entire E. coli chromosome. They described three behaviors: segments may be either (i) invertible and tolerant to inversion (type T, the inverted state is stable over generations); (ii) invertible but refractory (type R, the inverted state is detrimental); or (iii) nonpermissive (type N, either the inversion is not feasible or so strongly disabling that it is never detected). The striking observation was that deleterious inversions have one endpoint located in the 20–30% of the E. coli chromosome flanking the replication terminus; these regions behaved differently from the rest of the chromosome and were termed nondivisible zones. The NDZs also harbor multiple replication pause sites that, when occupied by the Tus protein, are polar inhibitors of replication fork movement (HILL 1996 ). In earlier experiments Tus and pause sites appeared not to be involved in the NDZ phenomenon because inversions ending in the NDZs remained nonpermitted or deleterious in tus-deleted strains (FRANCOIS et al. 1990A , FRANCOIS et al. 1990B ). It was proposed that the detriment caused by inverted segments of the NDZs could be an alteration of a higher-order nucleoid organization involving interactions between ordered sequences (REBOLLO et al. 1988 ). This was the first reported example of long-range interactions in the chromosome.

In this work, we further analyze the nondivisible zone NDZ1 (located between 20 and 30 min on the genetic map) to determine more precisely the positions of segments whose inversion is detrimental and to examine possible connections between NDZ and DAZ phenomena. Inversions between predetermined positions were sought using a genetic system based on modified Tn10 insertions (REBOLLO et al. 1988 ). The results indicate the existence of several regions harboring elements whose concerted arrangement is functionally important, irrespective of gene content. Splitting such clusters of sequences causes abnormalities in the shape and distribution of nucleoids that are the probable reasons for the disturbed growth of such bacteria. These observations further support the model that inversions may be deleterious because they disturb the interplay of multiple sequence elements governing chromosome organization and its processing.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Bacterial, plasmid, and bacteriophage strains:
Inversions were engineered in derivatives of strain CB0129 (W1485 F^- leu thy deoB or C supD; BIRD et al. 1972 ) and are referred to by their Tes and Tek insertions. One exception is INV(aroA-fadR), which was engineered in Hfr KL19. Hfr LN2085, also a CB0129 derivative, was described earlier (FRANCOIS et al. 1990C ); it promotes chromosome transfer from the trp operon in a counterclockwise direction. The recipient (F^-) strain for conjugation assays, LN1902, is W945 ara leu lacY1 purE gal pyrC76::Tn10 (Tc^R) trp zdd230::Tn9 (Cm^R) his gyrA (Nal^R) argG rpsL (Sm^R) xyl mtl ile metA/B. Plasmid pVF16, constructed by V. Francois to facilitate the creation of new Hfr's, is composed of a 2.7-kb fragment from pSC101 carrying a thermosensitive replication origin (HASHIMOTO-GOTO and SEKIGUCHI 1977 ) associated with a 1-kb fragment from plasmid F carrying the origin of transfer oriT, a multilinker for cloning, and a 1.85-kb interposon providing resistance to chloramphenicol (FELLAY et al. 1987 ). Plasmids pGR1 to pGR4 are derivatives of pVF16 by cloning in the chromosomal segments listed in Table 1. Plasmid pOX38::Ap^R (GUYER et al. 1980 ) was kindly provided by M. Chandler. Bacteriophage TSK was used to replace Tn10 insertions by their Tes or Tek derivatives, as previously described (FRANCOIS et al. 1987 ; REBOLLO et al. 1988 ). P1vir bacteriophage was used for generalized transduction of chromosomal markers, as described in CARO and BERG 1971 .

Growth media and genetic procedures:
Luria L rich medium and Vogel and Bonner E synthetic medium are described in MILLER 1992 . Antibiotic concentrations used are as follows: ampicillin (Ap), 50 µg/ml; chloramphenicol (Cm), 25 µg/ml; kanamycin (Km), 25 µg/ml; nalidixic acid (Nal), 20 µg/ml; streptomycin (Sm), 20 µg/ml (E medium) or 10 µg/ml (L medium); and tetracyclin (Tc), 10 µg/ml (E) or 50 µg/ml (L). All in vivo or in vitro genetic experiments were performed according to standard procedures (SAMBROOK et al. 1989 ; MILLER 1992 ).

Tn10 mutagenesis:
New Tn10 insertions in NDZ1 were isolated from a pool of random Tn10 insertions in strain CB0129, kindly provided by C. Labi and constructed as described in WAY et al. 1984 . The Tc^R resistance character was cotransduced with prototrophy to pyrC (24.2 min), trp (28.3 min), and pyrF (28.9 min) mutant strains via bacteriophage P1. Other Tn10 insertions were gifts from B. Bachmann or C. Gutierrez.

Tn10-Tes or -Tek substitutions:
Tn10 derivatives carrying a streptomycin/spectinomycin resistance (Sm^R/Sp^R) interposon in the proximal part of the tetA locus (the Tes transposon) or a kanamycin resistance (Km^R) fragment in the distal part of tetA (the Tek transposon) were constructed in vivo using recombinant phage TSK (FRANCOIS et al. 1987 ; REBOLLO et al. 1988 ). Strains harboring a Tes plus a Tek transposon were systematically constructed by adding, via P1 transduction, the Tek transposon to an already present Tes transposon. Since the presence of a single copy of the Sp^R/Sm^R interposon in the terminus region confers weak resistance to streptomycin, and virtually no resistance to spectinomycin, we avoided as much as possible direct selection for Sp^R/Sm^R bacteria, which may favor duplication of the resistance locus or other undesired rearrangements.

Selection of chromosomal inversions:
Fig 1A illustrates the genetic selection used in the inversion search. In a Tc^S Km^R Sm^R bacterium harboring a Tes plus a Tek transposon inserted in opposite orientations, a globally reciprocal homologous exchange occurring between the 0.9-kb segments separating the Km^R and Sm^R insertion positions of the Tes and Tek sequences may yield a selectable Tc^R recombinant with inversion of the chromosomal segment separating the transposons. Inversions being preferentially found among Tc^R Sm^R Kn^R triply resistant recombinants (FRANCOIS et al. 1987 , FRANCOIS et al. 1990A ; REBOLLO et al. 1988 ), selection for Tc^R recombinants was in general done in the presence of kanamycin, then screening for the Sm^R character and finally a search for inversions in this class of indigenous recombinants were performed. Triply resistant clones were picked up, resuspended in 0.5 ml of Luria broth (LB) medium, and incubated without aeration for 1 hr at 37° before the inversion test.

View larger version (28K): In this window In a new window Download PPT slide   Figure 1. Genetic tools for selection and analysis of chromosomal inversions. (A) The Tes and Tek system. Tes and Tek transposons (large arrows), inserted in opposite orientations in the chromosome, can recombine to give an easily selectable TcR determinant. The exchange indicated by a double arrow may result, if reciprocal, in the inversion of the chromosomal segment separating the Tes and Tek insertions, with reconstitution of a wild-type Tn10. Inversion may be detected in conjugation experiments by the inverted order of transfer from oriT of markers X and Y external to the segment inverted. (B) Mapping and orientation of Tn10 insertions and of oriT-carrier plasmid integration sites. The upper line shows a simplified version of the E. coli chromosome Ecomap7 between 1050 and 1400 kb (BERLYN et al. 1996 ). Flags indicate the locations of the three pause sites of the region, all inhibiting leftward moving replication forks. Tn10 insertions are indicated by arrows oriented from IS10L to IS10R. Vertical dotted lines indicate their mapping, estimated from both physical and genetic data. They are designated in the "z" naming system assorted from a number recalling their position in kilobases on Ecomap7. The map positions of Tn10 insertions in known genes are given between brackets. Small rectangles with arrowheads superimposed on the map indicate the positions of the integrated pVF16 (oriT) derivatives, with directions of transfer symbolized by an arrowhead (Hfr's SV1 to SV4; see text). (C) Typical Hfr organization. Here Hfr SV1 is represented, resulting from integrative recombination of pGR1 into the chromosome, with vector pVF16 DNA flanked by direct repeats of the chromosomal segment cloned in pGR1.

The inversion test:
A conjugation test requiring no extensive growth after the isolation procedure was favored. It demands that an F plasmid origin of transfer is located within the segment analyzed (Fig 1). This was obtained using plasmid pVF16, a repA(Ts) derivative of pSC101 that carries an F plasmid origin of transfer, oriT, and a gene for resistance to chloramphenicol. Four segments of the 23- to 30-min region were cloned into this vector (Table 1 and Fig 1), and the resulting plasmids were integrated into the chromosome by homologous recombination, selecting for Cm^R clones at 42° (the expected structure is shown in Fig 1C). This provided insertion of oriT at convenient positions. The trans-acting transfer functions were eventually provided by the conjugation-proficient F factor derivative pOX38::Ap^R (GUYER et al. 1980 ). These Hfr's form an isogenic strain family and the small size of the oriT plasmid minimizes local perturbation, but they are relatively unstable and cryosensitive (FRANCOIS et al. 1990C ). In the normal chromosome configuration, Hfr strain SV1 transfers its chromosome clockwise and the three others (SV2, SV3, SV4) transfer counterclockwise. For the conjugation test, the triply resistant Hfr bacteria were mixed with F^- LN1902 bacteria grown in LB (one male for five females), incubated for 1 hr, and then plated after convenient dilution on minimal medium selecting for PurE⁺ Nal^R or His⁺ Nal^R exconjugants. The selected markers are located on either side of the region analyzed (purE at 12 min and his at 45 min).

Microscopy:
Nucleoids of living bacteria were visualized by 4',6-diamidino-2-phenylindole (DAPI) staining. DAPI was added at a final concentration of 2 µg ml^-1 to cells exponentially growing in LB medium at OD_{560 nm} of 0.3. The cultures were chilled on ice 20 min later, centrifuged, and concentrated 10-fold in DAPI-containing (2 µg ml^-1) LB medium. A 1- to 2-µl aliquot was then spread on a microscope slide and observed under combined phase-contrast and fluorescence emission microscopy (Leica DRMB microscope; final magnification, x2450).

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Establishment of inversions in the 23- to 29-min region:
New Tn10 insertions in this region were selected from a pool of random Tn10 insertions by phage P1-mediated cotransduction with either pyrC (24.2 min), trp (28.3 min), or pyrF (28.9 min) markers. Fourteen out of 70 isolates were retained for further analysis after mapping by measurements of cotransduction frequencies with the pyrC, trp, or pyrF markers and by Southern analysis of PstI, HindIII, and EcoRI segments responding to a Tn10 probe. Some suitably positioned Tn10 insertions previously isolated were added to this list. Tn10 orientation within the chromosome was determined by chromosome mobilization using a Tn10-carrying conjugative plasmid (REBOLLO et al. 1988 ) and was confirmed by Southern analysis. This set of Tn10 insertions is presented in Fig 1B. The procedure for generating inversions, presented in MATERIALS AND METHODS and Fig 1A, required that the Tn10 insertions were eventually converted into Tes and Tek derivatives. An additional requirement was that an oriT plasmid must be present between the Tes and Tek insertions in opposite orientations. The set of oriT insertions used, presented in Table 1 and Fig 1B, has restricted our analysis to the 30 segments shown in Fig 2.

View larger version (44K): In this window In a new window Download PPT slide   Figure 2. Segment behavior in the inversion assay. In general, TcR KnR clones were first selected and then screened for Sm resistance. The inversion assay was carried out on triply resistant clones TcR KmR SmR, the recombinant class displaying the highest frequency of inversions when the segment is tolerant. The upper horizontal line represents the physical map of the chromosome region analyzed here, as in Fig 1B. The different chromosomal segments submitted to the inversion assay are represented by horizontal bars, the endpoints being indicated on the physical map by vertical dotted lines. Green bars, T segments; yellow bars, R segments; red bars, N segments; yellow-to-red graded bars, N/R segments. Class A and class B refer to the presence (A) or absence (B) of a pause site whose inversion might cause serious replication delay (see text and Fig 3). Segment limits (the Tes insertion is indicated first): Class A: 1, pyrC–zci-1325; 2, fadR-trpB83; 3, zce-1137-trpB83; 4, zce-1134-trpB83; 5, narK–zci-1315; 6, narI-treA; 7, zcc-1054–zch-1282; 8, zcc-1054–zce-1137; 9, zcc-1054–zce-1134. Class B: 1, zcj-1349–zcj-1378; 2, zci-1325–zcj-1378; 3, narI–zcj-1378; 4, zch-1282–zcj-1378; 5, zch-1275–zcj-1378; 6, fad–zcj-1378; 7, zce-1137–zcj-1378; 8, zce-1134–zcj-1378; 9, zce-1137–zcj-1353; 10, pyrC–zci-1344; 11, zch-1282-pyrF; 12, fadR-pyrF; 13, zce-1137-pyrF; 14, zce-1134-pyrF; 15, fadR-narK; 16, tre–zch-1275; 17, zce-1117–zce-1137; 18, zce-1098–zce-1137; 19, zcd-1092–zce-1137; 20, zcd-1098–zce-1134; 21, zcd-109 –zce-1134. Hfr's used for inversion tests are as follows: SV1 for segments B1 to B5; SV2 for segments A1 to A5, B6 and B8 to B14; SV3 for segments A6, B7, B15, and B16; and SV4 for segments A7 to A9 and B17 to B21.

Bidirectional transfer indicates a deleterious inversion:
For nearly half the segments assayed, a variable but important fraction of the triply resistant recombinants displayed bidirectional transfer in the conjugation test: both purE⁺ and his⁺ markers were transferred to the female with similar frequencies (Table 2). Subcloned immediately, such recombinants gave rise to two types of colonies in roughly equal proportions: one type promoted bidirectional transfer like the original clone, while the other transferred the chromosome in the same direction as the original Hfr ancestor. This suggested that the triply resistant clones contained, at the time of the test, a mixed population composed of roughly equal numbers of inversion-type and wild-type bacteria. Southern analyses performed on wild-type bacteria isolated from mixed clones confirmed the view that these bacteria arose from a second exchange between the Tn10 sequences that restored a normal gene order and eliminated the initial inversion (data not shown). The mixed populations might have arisen in either of two ways: either the inversion was highly unstable and reverted at a high frequency to the wild-type genome order, or the inversion was deleterious and rare wild-type revertants overgrew the inversion strain more or less rapidly. The second possibility was supported by analyses of the rates of evolution of the recombinant populations. When mixed clones were incubated for longer periods of growth after identification, their transfer properties evolved toward a pattern of transfer identical to that of the noninverted ancestor at rates that, although variable, did not fit in the no-selection model. If the two types grew with similar generation times, the initial data of Table 2 (see also REBOLLO et al. 1988 ) would be compatible with a frequency of return to wild-type order as high as 1–2% per cell generation, yielding populations containing roughly equal numbers of wild-type and inversion-type bacteria at the time of testing the recombinant colonies (30 generations after the initial inversion event). Even with such high recombination frequencies, additional growth for a few generations would not change significantly the proportion of each type in this model. The clone evolution reported in Table 2 is clearly much too rapid and cannot be explained without assuming a growth handicap of the bacteria carrying the inversion. The fastest evolution was observed for INV(fadR-trp) (A2 in Fig 2). Computer simulations indicate that this evolution rate may be observed when the wild-type revertant grows 2.5 times faster than the original inversion-type bacterium and occurs at a frequency of 10^-4 per cell generation of the inversion type. In this example, most of the recombinant clones tested were classified as noninverted (Table 2). In the other example analyzed (Table 3, segment B13 in Fig 2), the accumulation of wild-type bacteria, though less rapid than for A2, also could not fit in the no-selection model. For other inversions, such as INV(narI–zci-1385) (B3), the results reported in Table 2 suggest that clone evolution is even slower, since some clones still showed the inversion phenotype. Growth defects attached to inversions may thus have variable intensities, but, since the frequency of mixed clones depends both on the difference in generation times between inversion-type and wild-type bacteria and on the frequency of return to wild-type order, we cannot estimate precisely the selective disadvantage due to a given inversion. In the following, the existence among Tc^R recombinants of an elevated ratio of clones displaying bidirectional transfer was taken as indicating that the inversion of the tested segment was deleterious (R segments in Table 3).

The mosaic of permissive tolerant, permissive refractory, and nonpermissive segments:
Tc^R indigenous recombinants between Tes and Tek insertions have always been detected, but their frequencies varied over two orders of magnitude. The Tc^R Sm^R Kn^R triply resistant clones, among which inversion events are preferentially observed, also occurred with variable frequencies among the Tc^R recombinants (data not shown). A low frequency of triply resistant clones, added to the fact that in general we avoided direct selection for the Sm^R character, explains why the number of candidate clones tested is low for some segments. In some cases, interchanging the Tes and Tek positions provided identical results (not shown). This is consistent with the idea that the exchanges generating inversions, entirely restricted to Tn10 sequences, should not interfere with the expression of nearby genes.

Results of inversion assays are summarized in Fig 2, where segments are classified according to two criteria: (i) the behavior in the inversion assay (T, R, or N), with reference to the examples presented in Table 3, and (ii) the presence within the segment assayed of pslB and/or pslC but not of pslA pause sites (class A) or all other situations (class B). The reason for considering class A and B inversions separately is presented in the next section.

T segments: Five of the segments were tolerant to inversion. They represent a novel feature of this region, since they were not observed in our previous analyses (REBOLLO et al. 1988 ; FRANCOIS et al. 1990A ). Their sizes range from 27 (fadR-narK) to 247 kb (zce1138-zcj1385). They overlap or encompass many refractory or noninvertible segments, and in some cases they share one endpoint with refractory segments. Only one belongs to Class A (A6).

R segments: Thirteen of the segments gave rise to deleterious inversions, according to the test presented above (bidirectional transfer). They constitute the major category. They are found all over the region analyzed, do not necessarily overlap, and their sizes are variable (from 30 to 245 kb). They belong mostly to class B.

N segments: For five segments inversion was never detected. Some do not overlap, and their sizes vary from 39 to 200 kb. Three belong to class A, and those belonging to class B are among the smallest segments analyzed.

R/N segments: Five segments, characterized by very few mixed clones and a majority of noninverted clones, were assigned to this ambiguous class. In the case of segment A2, the evolution analysis reported in Table 3 allowed an assignment to the R class.

Segment behavior in Tus^-conditions:
The present set of Tn10 insertions was exploited to complete our previous finding that constraints to inversion exist independently of the Tus/pause site system activity (FRANCOIS et al. 1990A ). We chose to focus on segments classified A in Fig 3, since inversion of pslB and/or pslC without inversion of pslA may hinder completion of chromosome replication, as depicted in Fig 3C. By P1 transduction, a tus::Ap^R deletion was crossed into strains carrying Tes and Tek insertions flanking segments A1 to A4 and B12, B17, and B18 as controls. Except for R segments A2 and B12, they were classified N in Tus⁺ conditions (Table 3). Tc^R Sm^R Kn^R recombinants of the Tus^- derivatives were analyzed for presence of inversions. Table 3 shows that the behavior of segments A2, A3, and A4 was unchanged (they remained N or R/N), but that of segment A1 was dramatically altered: it shifted from N state (no inverted clone among 50 triply resistant colonies tested) to T state (80% inverted clones among 80 triply resistant clones). The behavior of B segments, either harboring pause sites (B12, R) or not (B17, R, and B18, N), was not affected by the (tus)::Ap^R mutation.

View larger version (18K): In this window In a new window Download PPT slide   Figure 3. Predicted changes in chromosome replication after pause site inversion. The simplified map shows only the positions of the polar pause sites (pslA, pslB, and pslC), oriented as indicated in Fig 1. Vertical thin arrows indicate inversion endpoints. Inverted segments are indicated by dotted lines. Horizontal large arrows represent replication forks. (A) Wild-type organization: Replication forks travel through the region from left to right. The replication fork trap (RFT) is located to the right of pslA = TerA, and termination occurs normally outside the region explored in the vicinity of psrA = TerC (1607 kb; LOUARN et al. 1994 ). (B) Class B inversion: The inverted segment includes pslA but not pslB or pslC. The arrest of the rightward moving fork at inverted pslA must force termination to occur at this site but the overall replication time is not dramatically increased. A de novo RFT is generated. (C) Class A inversion: The inverted segment includes pslB but not pslA. The inverted pause site stops the rightward moving replication fork, and the leftward moving fork is blocked at pslA. Replication of the region between these terminators must be delayed.

Filament formation and abnormal nucleoids in bacteria harboring deleterious inversions:
Filaments were routinely detected in mixed clones issuing from bacteria that had undergone a deleterious inversion. To characterize this phenotype, we examined the aspect of nucleoids in filaments by fluorescence microscopy after staining the DNA by DAPI. The analysis was not performed directly on the Hfr strains used above because we noted that bacteria carrying the repA(Ts) plasmid integrated in their chromosome tend to form filaments even when cultivated at a temperature nonpermissive for plasmid replication. Rather, it was carried out on inversion strains derived from Hfr strains that do not normally display filaments (Hfr's KL19 and LN2085). The examinations also included F^- bacteria, picked up from the smallest colonies found after selection for Tc^R Sm^R Kn^R characters. We expected such colonies to issue from a bacterium carrying a disabling inversion. Bacteria with abnormal morphology were easily detected for every strain in which the Tes and Tek insertions flanked an R segment. In all these strains, the abnormal cells (in general a few percent of the clone population) looked very similar, with filaments and abnormal distribution of nucleoids that remained grouped in elongated masses located at one or few positions in the filaments (Fig 4, A–C, F). Note that the three R segments examined are not overlapping. For the Hfr strains, the bacteria carrying the inversion represented 30–70% of the mixed clones analyzed, as deduced from conjugation data. This frequency was much greater than that of filaments. Abnormal bacteria could also be observed from time to time in strains having undergone inversion of a T segment but at a much lower frequency than for R segments (an example is provided in Fig 4E for segment A6).

View larger version (39K): In this window In a new window Download PPT slide   Figure 4. Micrographs of bacteria undergoing inversions of T or R segments. Cultures of the strains indicated below were plated on L medium with tetracycline, streptomycin, and kanamycin. The smallest colonies arising after 24 hr of incubation at 37° were picked up, resuspended in 1 ml of LB, and incubated for 2–3 hr. Bacteria were then stained by DAPI and examined as indicated in MATERIALS AND METHODS. For the two Hfr strains used, subsequent conjugation tests correlated the presence of abnormal bacteria with bidirectional transfer. (A) HfrKL19 aroA::Tes fadR::Tek (an R class A segment located between map positions 962 and 1236 kb; REBOLLO et al. 1988 ). (B) Hfr LN2085 pyrF::Tes fadR::Tek (R segment B12; FRANCOIS et al. 1990A ). (C) CB0129 pyrF::Tes fadR::Tek (same R segment B12). (D) CB0129 narK::Tes fadR::Tes (T segment B15). (E) CB0129 treA::Tes narI::Tek (T segment A6). (F) CB0129 zci-1349::Tes zcj-1378::Tek (R segment B1).

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Experimental study of long-range interactions in the bacterial chromosome is hampered by methodological obstacles. Our approach has been to use precisely located genetic inversions to look for orientation-dependent cis interactions between unknown elements remote from each other on the chromosome. It has recently proved its usefulness for understanding the regional control on dif activity and for definition of the mechanism generating the dif activity zone (DAZ; PERALS et al. 2000 ), as well as for localizing the DNA regions responsible for orientation of the chromosome in the prespore during sporulation in Bacillus subtilis (J. ERRINGTON, personal communication). Extended analyses of inversions have also been performed in Salmonella typhimurium (SCHMID and ROTH 1983 ; SEGALL et al. 1988 ; MIESEL et al. 1994 ), where examples of true noninvertible segments have been reported. The genetic system used here allowed the generation of many conservative inversions. We have analyzed the reaction of the cell to inversion of 30 segments within the 23- to 29-min region to further explore a chromosome region (NDZ1) previously identified as refractory to inversion (REBOLLO et al. 1988 ; FRANCOIS et al. 1990A , FRANCOIS et al. 1990B ). The data confirm the existence of constraints to inversion in this region since most segments appeared either of N or R types and illustrate a degree of complexity of cis effects responsible for constraints to inversion in that (i) inversion of certain segments of this region is well tolerated and (ii) at least one inversion is detectable in Tus^- conditions only.

The fact that segment A1, the second longest of the A series, shifted from the N state to the T state when the tus locus was deleted indicates that the Tus/pause site system is one phenomenon determining segment behavior in the inversion tests. However, smaller segments A2 and A3, included in A1, remained N type in Tus⁺ as well as in Tus^- conditions. Whether the inhibitory effect of Tus on inversion of segment A1 directly implicates the pause site it harbors is an open question. Since the smallest segments of the A series, A5 and A6, may be inverted in Tus⁺ conditions, it appears that (i) a pause site present within the segment to be inverted may not be an obstacle to the inversion event and (ii) an inverted pause site is in itself not very detrimental, consistent with previous observations (DE MASSY et al. 1987 ; HORIUCHI and FUJIMURA 1995 ). Segments A2, A3, A4, B12, B17, and B18, for instance, are R or N type in Tus⁺or Tus^- conditions, indicating that Tus was not involved in the behavior of these segments. The peculiar behavior of segment A1 can be related to similar observations made independently in S. typhimurium: MIESEL et al. 1994 have observed segments that, although noninvertible by indigenous recombination, could give rise to well-tolerated inversions when the recombining sequences consisted of exogenous DNA introduced by phage-mediated transduction. Eventually, J. ROTH and co-workers (personal communication) discovered that the Tus protein was the factor inhibiting the indigenous recombination required for inversion of these segments. In these closely related organisms, it appears that some chromosome segments are nonpermissive to inversion because the recombination event responsible for their formation aborts systematically for some as yet undetermined activity of Tus.

At least in E. coli, the R or N character of many segments is not determined by Tus activity. We focus in the following only on segments for which inversion has been detected (T and R segments). The existence of T segments encompassing R ones, plus the fact that many R segments do not overlap, make it unlikely that the deleterious effects are due to reorientation of a single locus or to a change in direction of replication. Our general interpretation is that NDZ1 is not a continuum but a succession of several smaller NDZs each harboring sequences whose relative orientation is physiologically important. Long T segments might contain entire NDZs, and the R segments might split and inactivate one of these NDZs, whereas small T segments might contain no or only few polar determinants. The succession of seemingly independent and possibly cooperative NDZs in the 10-min region left of dif leads to the proposal that a given NDZ may be initiated and/or terminated at specialized sequence elements. An NDZ might include such initiator/terminator elements, plus polarization factors. The nature of elements generating an NDZ must for the moment remain uncertain, since the present data do not allow precise determination of number and positions of NDZs in the 23- to 30-min region.

The severity of growth defects due to deleterious inversions varies with the segments, but it is difficult to estimate in the absence of precise information on frequencies of return to normal gene order. We suspect that the growth defect is the consequence of a high risk of fatal problems leading to abnormal nucleoid and cell morphology. The abnormal bacteria are probably destined to die, and the apparent generation time of deleterious inversion strains is consequently increased. The altered nucleoid morphology closely resembles that observed in topoisomerase IV Par^- mutants (KATO et al. 1988 ), as if the rearranged chromosomes have difficulty in partitioning during the cell cycle. This suggests the interesting novelty that the function of NDZs is to participate in spatial separation of sister chromosomes by facilitating or directing decatenase action. It may be that the NDZs revealed here all engage in a similar phenomenon and participate independently but additively in nucleoid dynamics.

A point of interest is the relation between the two phenomena typical of the terminus, the organization within NDZs and the generation of the DAZ. These phenomena are probably independent of one another. The regions where NDZs are observed are peripheral to the terminus and do not cover the region governing dif activity: the map in Fig 5 shows that the NDZ region is separated from the DAZ-controlling region by a zone (roughly between 1400 and 1565 kb on the map) that seems to tolerate inversions, even of very large segments (the largest inversion ever characterized starts at IS5F; LOUARN et al. 1985 ). Polarization of NDZs is probably not achieved by the polar elements influencing dif activity (CORRE et al. 2000 ; PERALS et al. 2000 ), since these elements are not necessarily deleterious when inverted: the long segment between positions 1440 and 1565 kb (Fig 5) is a T segment, although this segment harbors polar elements able to substitute for those normally controlling dif (PERALS et al. 2000 ). However, the alteration of cell morphology due to NDZ splitting, if strongly reminiscent of that observed in Par^- mutants, is not very different from that due to distortion of the DAZ (PERALS et al. 2000 ).

View larger version (33K): In this window In a new window Download PPT slide   Figure 5. Distribution of tolerated and deleterious inversions in the terminus region. This map combines data redrawn from Fig 2 for T and R segments (between positions pyrC and zcj-1378) with previously published data to show the respective positions of the NDZ region, of the DAZ and the surrounding region involved in its formation, and of the seemingly inversion-tolerant intervening region. The large IS5F-IS5T inversion (Inv29-78) is described in LOUARN et al. 1985 , the large inversion from position 1397 kb to pheA is described in REBOLLO et al. 1988 , and other inversions to the right of position 1430 kb are described in PERALS et al. 2000 . Green bars, tolerated inversions; yellow bars, deleterious inversions. Deleterious effects of inversions are probably different in the NDZ region (defects in chromosome separation?) and in the dif-controlling region (defects in dimer chromosome resolution).

Other possible leads to the nature of NDZs have appeared recently. First, the sequence of the region where NDZs are found reveals a high potential for DNA curvature and for low helix stability compared with the rest of the chromosome (PEDERSEN et al. 2000 ). These properties should be examined, in conjunction with skewed sequences, for their relationship to NDZ organization. Second, N. P. HIGGINS (personal communication) has recently discovered that the ability of two res sites inserted in the terminus region of the S. typhimurium chromosome to recombine with each other depends upon their relative locations, and he has located several barriers hindering communication between the sites. Taking into account the requirements for recombination between res sites, he concluded that such barriers must behave as elements preventing supercoiling diffusion (HIGGINS et al. 1996 ), a property that would make them good candidates for limits between NDZs.

Finally, the Ter macrodomain, described recently by NIKI et al. 2000 and characterized by co-location in the cell of all tested segments of a large region (20–30% of the chromosome) during the cell cycle, coincides well with the region where NDZs are found. One may speculate that the organization within NDZs involves interactions between ordered sequences and some scaffolding apparatus that not only participate in spatial separation of sister nucleoids but also help compact this region of the chromosome. The behavior of the Ter domain in nucleoid dynamics might be due in part to the NDZ-dependent compaction and in part to the DAZ-associated mobilization of terminus DNA.

	ACKNOWLEDGMENTS

We are indebted to David Lane for numerous improvements to the manuscript. We thank Rafael Camacho and John Roth for discussions and sharing of unpublished information. We are grateful to Koryn Perals for help with microscopy, and to Vincent Francois for pVF16 construction. This work has been supported by contract no. SC1*-CT91-0713 of the Science Program of the European Community, by contract CE92-0003 of the Direccion General de Investigacion Cientifica y Técnica (Spain), and by the France-Spain "Mercure" Cooperation Program.

Manuscript received October 5, 2000; Accepted for publication November 30, 2000.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

BERLYN, M. K. B., K. B. LOW, K. E. RUDD and M. SINGER, 1996 Linkage map of Escherichia coli K12, pp. 1715–1902 in Escherichia coli and Salmonella Cellular and Molecular Biology, Ed. 2, edited by F. C. NEIDHARDT, R. CURTISS III, J. L. INGRAHAM, E. C. C. LIN, B. MAGASANIK et al. American Society for Microbiology, Washington, DC.

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

CAPIAUX, H., F. CORNET, J. CORRE, M. GUIJO, and K. PERALS et al., 2001  Polarization of the Escherichia coli chromosome: a view from the terminus. Biochimie in press.

CARO, L. and C. BERG, 1971  P1 transduction. Methods Enzymol. 21:444-458.

CORRE, J., J. PATTE, and J. M. LOUARN, 2000  Prophage lambda induces terminal recombination in Escherichia coli by inhibiting chromosome dimer resolution: an orientation-dependent cis effect lending support to bipolarization of the terminus. Genetics 154:39-48[Abstract/Free Full Text].

DE MASSY, B., S. BEJAR, J. LOUARN, J.-M. LOUARN, and J.-P. BOUCHE, 1987  Inhibition of replication forks exiting the terminus region of the E. coli chromosome occurs at two loci separated by 5 min. Proc. Natl. Acad. Sci. USA 84:1758-1763.

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, J. PATTE, 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].

FRANCOIS, V., J. LOUARN, J. PATTE, J. E. REBOLLO, and J. M. LOUARN, 1990a  Constraints in chromosomal inversions in Escherichia coli are not explained by replication pausing at inverted terminator-like sequences. Mol. Microbiol. 3:995-1002.

FRANCOIS, V., J. LOUARN, J. E. REBOLLO and J. M. LOUARN, 1990b Replication termination, nondivisible zones and structure of the Escherichia coli chromosome, pp. 351–359 in The Bacterial Chromosome, edited by K. DRLICA and M. RILEY. American Society for Microbiology, Washington, DC.

FRANCOIS, V., A. CONTER, and J. M. LOUARN, 1990c  Properties of new Escherichia coli Hfr strains constructed by integration of pSC101-derived conjugative plasmids. J. Bacteriol. 172:1436-1440[Abstract/Free Full Text].

GUYER, M. S., R. R. REED, J. A. STEIZ, and K. B. LOW, 1980  Identification of a sex factor affinity site in Escherichia coli as gamma delta. Cold Spring Harbor Symp. Quant. Biol. 45:135-140.

HASHIMOTO-GOTO, T. and M. SEKIGUCHI, 1977  Mutations to temperature sensitivity in R plasmid pSC101. J. Bacteriol. 131:405-412[Abstract/Free Full Text].

HIGGINS, N. P., X. YANG, Q. FU, and J. R. ROTH, 1996  Surveying a supercoil domain by using the gamma delta resolution system in Salmonella typhimurium. J. Bacteriol. 178:2825-2835[Abstract/Free Full Text].

HILL, T. M., 1996 Features of the chromosomal terminus region, pp. 1602–1614 in Escherichia coli and Salmonella Cellular and Molecular Biology, Ed. 2, edited by F. C. NEIDHARDT, R. CURTISS III, J. L. INGRAHAM, E. C. C. LIN, B. MAGASANIK et al. American Society for Microbiology, Washington, DC.

HORIUCHI, T. and Y. FUJIMURA, 1995  Recombination rescue of the stalled replication fork: a model based on analysis of an Escherichia coli strain with a chromosome region difficult to replicate. J. Bacteriol. 177:783-791[Abstract].

KATO, J.-I., Y. NISHIMURA, M. YAMADA, H. SUZUKI, and Y. HIROTA, 1988  Gene organization in the region containing a new gene involved in chromosome partition in Escherichia coli. J. Bacteriol. 170:3967-3977[Abstract/Free Full Text].

KONRAD, E. B., 1977  Method for the isolation of Escherichia coli mutants with enhanced recombination between chromosomal duplications. J. Bacteriol. 130:167-172[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 Biol. 3:799-811[Medline].

LOUARN, J., J. PATTE, and J.-M. LOUARN, 1982  Suppression of E. coli dnaA46 mutation by integration of plasmid R100.1 derivatives; constraints imposed by the replication terminus. J. Bacteriol. 151:657-667[Abstract/Free Full Text].

LOUARN, J.-M., J. P. BOUCHE, F. LEGENDRE, J. LOUARN, and J. PATTE, 1985  Characterization and properties of very large inversions of the E. coli chromosome along the origin-to-terminus axis. Mol. Gen. Genet. 201:467-476[Medline].

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].

MESSER, W., and C. WEIGLER, 1996 Initiation of chromosome replication, pp. 1579–1601 in Escherichia coli and Salmonella Cellular and Molecular Biology, Ed. 2, edited by F. C. NEIDHARDT, R. CURTISS III, J. L. INGRAHAM, E. C. C. LIN, B. MAGASANIK et al. American Society for Microbiology, Washington, DC.

MIESEL, L., A. SEGALL, and J. R. ROTH, 1994  Construction of chromosomal rearrangements in Salmonella by transduction: inversions of non-permissive segments are not lethal. Genetics 137:919-932[Abstract].

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

NIKI, H., Y. YAMAICHI, and S. HIRAGA, 2000  Dynamic organization of chromosomal DNA in Escherichia coli.. Genes Dev. 14:212-223[Abstract/Free Full Text].

PEDERSEN, A. G., L. J. JENSEN, S. BRUNAK, H.-H. STAERFELDT, and D. W. USSERY, 2000  A DNA structural Atlas for Escherichia coli. J. Mol. Biol. 299:907-930[Medline].

PERALS, K., F. CORNET, Y. MERLET, I. DELON, and J.-M. LOUARN, 2000  Functional polarization of the Escherichia coli chromosome terminus. The dif site acts in chromosome dimer resolution only when located between long stretches of opposite polarity. Mol. Microbiol. 36:1-12[Medline].

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

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

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

SCHMID, M. B. and J. R. ROTH, 1983  Selection and endpoint distribution of bacterial inversion mutations. Genetics 105:539-557[Abstract/Free Full Text].

SEGALL, A., M. J. MAHAN, and J. R. ROTH, 1988  Rearrangements of the bacterial chromosome: forbidden inversions. Science 241:1314-1318[Abstract/Free Full Text].

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

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].

WAY, J. C., M. A. DAVIS, D. MORISATO, D. E. ROBERTS, and N. KLECKNER, 1984  New Tn10 derivatives for transposon mutagenesis and for construction of lacZ operon fusions by transposition. Gene 32:369-379[Medline].

This article has been cited by other articles:

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

				G. Achaz, E. Coissac, P. Netter, and E. P. C. Rocha Associations Between Inverted Repeats and the Structural Evolution of Bacterial Genomes Genetics, August 1, 2003; 164(4): 1279 - 1289. [Abstract] [Full Text] [PDF]

				S. Alokam, S.-L. Liu, K. Said, and K. E. Sanderson Inversions over the Terminus Region in Salmonella and Escherichia coli: IS200s as the Sites of Homologous Recombination Inverting the Chromosome of Salmonella enterica Serovar Typhi J. Bacteriol., November 15, 2002; 184(22): 6190 - 6197. [Abstract] [Full Text] [PDF]