Recombination Enhancement by Replication (RER) in Rhizobium etli

Corresponding author: David Romero, Programa de Genética Molecular de Plásmidos Bacterianos, Centro de Investigación sobre Fijación de Nitrógeno-UNAM, Apartado Postal 565-A, 62210 Cuernavaca, Morelos, México., dromero{at}cifn.unam.mx (E-mail)

Communicating editor: G. R. SMITH

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Studies in several organisms show that recombination and replication interact closely. Recombinational repair usually requires associated replication at some stage; moreover, additional replication can induce recombination through either homologous or illegitimate events. In prokaryotes, stimulation of recombination by replication is more dramatic when rolling circle replication is employed. In contrast, -type replication induces only a modest increase in recombination frequency. In this article, we show that induction of -type replication from a supernumerary origin in the symbiotic plasmid (pSym) of Rhizobium etli leads to a 1000-fold increase in deletion formation on this plasmid. These deletions span 120 kb (the symbiotic region) and have as endpoints the reiterated nitrogenase operons. We have named this phenomenon RER, for recombination enhancement by replication. RER is not affected by the position of the replication origin in the pSym, the direction of advance of the replication fork, or the distance from the origin to the recombining repeats. On the other hand, RER is dependent on an active recA allele, indicating that it is due to homologous recombination. RER displays a strong regionality restricted to the symbiotic region. The similarities and differences of RER with the recombination process observed at the terminus of replication of the Escherichia coli chromosome are discussed.

TRADITIONALLY, DNA replication and recombination are studied separately. However, work with bacteriophages (KREUZER et al. 1995 ; KUZMINOV and STAHL 1999 ), bacteria (LOUARN et al. 1982 , LOUARN et al. 1991 ; NOIROT et al. 1987 ; PETIT et al. 1992 ; HORIUCHI and FUJIMURA 1995 ; CORRE et al. 1997 ; KOGOMA 1997 ; SAVESON and LOVETT 1997 ), and eukaryotic cells (RATTRAY and SYMINGTON 1993 ; MALKOVA et al. 1996 ) shows that recombination may be induced during the replication process. Rearrangement formation by replication arises either through recA-dependent (PETIT et al. 1992 ; SAVESON and LOVETT 1997 ) or recA-independent mechanisms (KREUZER et al. 1995 ; LOVETT and FESCHENKO 1996 ; BIERNE et al. 1997 ; SAVESON and LOVETT 1997 ). Recombination triggered by replication can be either reciprocal (NOIROT et al. 1987 ) or nonreciprocal (NOIROT et al. 1987 ; MALKOVA et al. 1996 ).

Currently accepted models for recombination involve DNA replication (STAHL 1994 , STAHL 1996 ). DNA replication primed by recombination intermediates is particularly relevant when chromosome integrity is affected or when the chromosomal origin is damaged and/or inactivated. Under these circumstances, replication induced by recombination restores genome integrity, allowing bacterial survival (ASAI et al. 1994 ; BIERNE and MICHEL 1994 ; KUZMINOV 1995 ; KOGOMA 1997 ).

Rolling circle-type replication usually provokes a strong stimulation in recombination (NOIROT et al. 1987 ; PETIT et al. 1992 ). This is reasonable, since the strand generated and displaced during replication may become a good substrate for recombination proteins, such as RecA, RecFOR complex, RecOT complex, and SsbB (KOWALCZYKOWSKI et al. 1994 ), or their counterparts in other systems, such as UvsX and UvsY in phage T4 (KREUZER et al. 1995 ). In contrast, a report suggests that -type replication has only a minor effect on recombination when tested on a bacterial system (MOREL-DEVILLE and EHRLICH 1996 ). This proposal is in conflict with results obtained in eukaryotic systems, where apparent -type replication strongly induces recombination (RATTRAY and SYMINGTON 1993 ; MALKOVA et al. 1996 ).

A particular drawback for most of the studies in bacterial systems is the use of artificial repeats created on the chromosome, in which a supernumerary replication origin has also been placed. Moreover, in many cases in which a new replication origin is placed into the chromosome, coexisting with a functional, vegetative origin, bacterial cells display altered nucleoid partition, aberrant cell morphology, and a decrease in viability (YAMAGUCHI and TOMIZAWA 1980 ; LOUARN et al. 1982 ; NOIROT et al. 1987 ; ELIASSON et al. 1996 ). Therefore, a system in which the connections between replication and recombination can be studied under less restrictive conditions would contribute to unraveling the links between these processes.

We have been using the nitrogen-fixing bacterium Rhizobium etli as a model for the study of genome dynamics. One of the unusual features of Rhizobium is the presence of abundant DNA reiteration, comprising about 18% of the genome (reviewed by ROMERO et al. 1997 ). Reiterated elements in Rhizobium spp. reside either in the chromosome or in plasmids (FLORES et al. 1987 ; GIRARD et al. 1991 ; FREIBERG et al. 1997 ), and include IS elements, housekeeping genes, and genes for nodulation and nitrogen fixation. Genomic rearrangements are facilitated by the presence of these repeated sequences (reviewed by ROMERO and PALACIOS 1997 ; ROMERO et al. 1997 ).

The genome of R. etli comprises the chromosome and six plasmids (pa to pf), ranging in size from 150 to 600 kb (BROM et al. 1992 ; GARCIA-DE LOS SANTOS et al. 1996 ). Plasmid d is the symbiotic plasmid (pSym, 390 kb) since it carries almost all the genes needed for symbiosis with beans. Either the pSym can be eliminated (BROM et al. 1992 ) or the copy number of several regions in this plasmid can be increased (ROMERO et al. 1991 , ROMERO et al. 1995 ) without detrimental effects for free-living growth. The pSym shows amplification and deletion events at high frequency (10^-4 per viable cell), affecting a 120-kb region termed the symbiotic region; these events are mediated by homologous recombination between flanking reiterated operons (nifHDK regions a and b) encoding the nitrogenase enzyme complex (ROMERO et al. 1991 ). A third, incomplete copy of these repeats is located inside the symbiotic region (nifH region c), in an inverted orientation vis à vis regions a and b. Recombination between the direct repeats is facilitated by their size (5 kb) and high homology. The nitrogenase reiterations participate in gene conversion-like events that preserve their identity (RODRIGUEZ and ROMERO 1998 ). Other reiterations participate in the generation of rearrangements, but at lower frequencies (~10^-6; ROMERO et al. 1995 ).

Studies aimed to identify processes that may trigger rearrangement formation in the pSym are yet absent. The main goal of this work was to evaluate the role of inducing additional replication on rearrangement formation in the pSym of R. etli. Our data show that additional -type replication in the pSym induces a dramatic increase in the formation of rearrangements, leading to loss of the symbiotic region through homologous recombination. Other zones of the genome are not destabilized by the additional replication process in this plasmid.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Bacterial strains, plasmids, and media:
The bacterial strains employed are listed in Table 1. Escherichia coli strains were grown at 37° in Luria-Bertani medium (MILLER 1972 ). R. etli and Agrobacterium tumefaciens strains were grown at 30° in PY medium (NOEL et al. 1984 ). When indicated, sucrose was added to PY plates at 12.5% (w/v). Antibiotics were added at the following concentrations (in micrograms per milliliter): (1) E. coli, carbenicillin (Cb) 100, chloramphenicol (Cm) 15, kanamycin (Km) 30, spectinomycin (Sp) 100, streptomycin (Sm) 100, tetracycline (Tc) 10; (2) R. etli, Cm 20, gentamycin (Gm) 80, Km 15, nalidixic acid (Nal) 20, neomycin (Nm) 75, rifampicin (Rif) 100, Tc 2.

Filter blot hybridization and determination of plasmid profiles:
Total DNA was digested with BamHI, electrophoresed in 1% agarose gels, blotted onto nitrocellulose (Hybond N⁺, Amersham, Arlington Heights, IL), and hybridized under stringent conditions using Rapid-Hyb buffer (Amersham) as specified by the manufacturer. Plasmid profiles were obtained by an in-gel lysis method (ECKHARDT 1978 ), blotted onto nitrocellulose, and hybridized similarly. Most probes were linearized by digestion with restriction enzymes and labeled with [³²P]dCTP by a random priming procedure (FEINBERG and VOGELSTEIN 1983 ) using the Rediprime DNA labeling system (Amersham). PCR products used as probes were not digested.

Plasmid construction:
The plasmids used in this work are shown in Table 2. All DNA manipulations were carried out under standard protocols (SAMBROOK et al. 1989 ). Restriction enzymes, T4 polynucleotide ligase, and bacterial alkaline phosphatase were purchased from Amersham and used as suggested by the provider.

To construct pEYM1, a 4.28-kb HindIII fragment from pTn5-luxB (C. KONCZ, unpublished results), containing both the replication and transfer origins (oriV and oriT, respectively) from plasmid RK2 (THOMAS and SMITH 1987 ; PANSEGRAU et al. 1994 ), was cloned into pIC20R (MARSH et al. 1984 ) in the desired orientation. Plasmid pEYM1 was digested with BglII and the 3.9-kb fragment containing oriV-oriT was then ligated into the unique BamHI site of the Tn5-containing plasmid pSUP5011 (SIMON 1984 ), replacing the mob sequence. The resulting plasmid (pEYM4) harbors a new Tn5 derivative designated as Tn5-GDYN2 (acronym for genome dynamics; Fig 1).

View larger version (15K): In this window In a new window Download PPT slide   Figure 1. Experimental design. (A) Schematic representation of the GDYN1 element. Arrows show the transcriptional direction of the sacB gene and the antibiotic resistance determinants. (B) Organization of Tn5-GDYN2. The IS50 elements are shown as open ovals; the central part shows the position of the Kmr and oriV-oriT determinants. (C) Schematic structure of the pSym of R. etli CFNX5, showing the location of different insertions of Tn5-GDYN2; these are indicated as a circle (strain CFNX600), inverted triangle (strain CFNX601), or small square (strain CFNX602), labeled as oriV-RK2. These strains carry the nifH region c::GDYN1 insertion, useful for determination of deletion frequencies. (D) Plasmid pEYM5, used to achieve replication from the supernumerary origin; the trfA gene is under the control of PTaclacUV5 (arrow).

Plasmid pEYM2 contains the oriV-oriT fragment from pEYM1 ligated to the 1.3-kb BamHI fragment from pSV16 (DURLAND et al. 1990 ), containing the Km^r gene from Tn903. Since the only active replication origin in pEYM2 is oriV from RK2, this plasmid must be maintained in a strain producing TrfA in trans, such as E. coli S17-1.

Activation of replication from oriV-containing elements was achieved using plasmid pEYM5. pEYM5 is a derivative from pMMB206 (MORALES et al. 1991 ) in which a 1.2-kb EcoRI-PstI fragment from pFF1-34wt (DURLAND et al. 1990 ), containing a promoterless trfA gene, was cloned under the control of the PTaclacUV5 promoter.

Plasmids pEYM9 and pEYM10 are integrative plasmids for R. etli derived from plasmid pSUP202 (SIMON et al. 1983 ), containing either an internal 4.5-kb EcoRI fragment from the BamHI band 47 or an internal 2.5-kb EcoRI fragment from the BamHI band 81 of the pSym map, respectively (GIRARD et al. 1991 ). Both plasmids confer a Tc^r phenotype when integrated into pSym.

For the construction of additional integrative elements, the 3.9-kb BglII fragment from pEYM1 (oriV-oriT) was ligated to a 2-kb BamHI Km cassette from pBSL128 (ALEXEYEV et al. 1995 ), generating plasmid pEYM11. Since this plasmid replicates using oriV-RK2, it was maintained in E. coli strain S17-1. Digestion of pEYM11 with HindIII and ligation into the unique HindIII site of the nifH c-containing vector, pCQ23 (QUINTO et al. 1982 ), generates integrative plasmids pEYM12 and pEYM13. These plasmids differ in the orientation toward the oriV replication fork advances.

Construction and characterization of R. etli strains carrying Tn5-GDYN2 on the symbiotic plasmid:
For introduction of Tn5-GDYN2, a biparental mating was done, using E. coli S17-1/pEYM4 as donor and R. etli CFNX5 (ROMERO et al. 1991 ) as recipient, selecting for Nal^r Nm^r transconjugants. To identify transconjugants with the transposon inserted in the pSym, we took advantage of the GDYN1 cassette, conferring Km/Gm, Sp/Sm resistance (ROMERO et al. 1991 ), present in the pSym of CFNX5. Triparental matings were made with A. tumefaciens GMI9023 as a recipient and individual R. etli colonies as donors; E. coli HB101/pRK2013 was used as a conjugation helper. A. tumefaciens transconjugants that had received a pSym with GDYN2 inserted in it were identified by their Nm^r Sp^r phenotype. Once we obtained transconjugants with the desired phenotype, we recovered the R. etli donors that had originated them. To verify that the Tn5-GDYN2 was inserted in the pSym of these R. etli derivatives (CFNX600, CFNX601, and CFNX602), their plasmid profiles were hybridized against the 3.9-kb BglII fragment containing oriV-oriT from pEYM1 or with an internal nifH probe (a 455-bp PCR product, covering from nucleotides 423 to 878, RODRIGUEZ and ROMERO 1998 ). To locate the position of the Tn5-GDYN2 in the pSym, total DNA was digested with BamHI and analyzed by Southern blot hybridization using as probes the pSym cosmid collection (GIRARD et al. 1991 ).

Construction of recA::Sp derivatives carrying Tn5-GDYN2 in the pSym:
To generate recA::Sp derivatives from the different insertions, the pSyms from strains CFNX600, CFNX601, and CFNX602 were transferred by conjugation into strains CFN2001 or its isogenic recA::Sp derivative, CFNX107 (see Table 1), using HB101/pRK2013 as a conjugation helper. Selection for Rif^r Nm^r transconjugants gave rise to strains CFNX603 to CFNX608. As a verification, plasmid profiles of selected transconjugants were hybridized against oriV and nifH specific probes (see above).

Construction and characterization of R. etli strains harboring oriV in alternative orientations on the pSym:
To create R. etli strains with the oriV replication fork on the pSym advancing either clockwise or counterclockwise, plasmids pEYM12 and pEYM13 were introduced by conjugation into strain CE3, selecting for Nal^r Nm^r transconjugants; this generates strains CFNX609 and CFNX610. Cointegrates between the pSym and either pEYM12 or pEYM13 resulted from single-exchange events, among the homologous region in all these plasmids (nifHc). The presence of the cointegrated plasmid was confirmed by Southern blot hybridization of plasmid profiles of selected transconjugants, using pBR322 as a probe. Formation of cointegrates in the nifHc region was verified by Southern blot hybridization of BamHI-digested total DNA, probed against an internal nifH fragment.

Construction and characterization of R. etli strains harboring the Tn5-GDYN2 in plasmid pGM1:
We introduced Tn5-GDYN2 into a strain containing pGM1, CFNX250 (Table 1) through a biparental mating using E. coli S17-1 as donor, selecting for a Nal^r Nm^r phenotype. We identified those transconjugants in which pGM1 received Tn5-GDYN2 through biparental matings against A. tumefaciens using individual R. etli colonies as donors and selecting for a Nm^r Sp^r phenotype. The original R. etli colonies that gave rise to these transconjugants were identified. Hybridization of plasmid profiles of selected transconjugants against the oriV-oriT probe confirmed the presence of Tn5-GDYN2 in pGM1. Location of Tn5-GDYN2 on pGM1 was established through BamHI-digested total DNA hybridization against the cosmids covering the symbiotic region. For derivatives not showing an altered band pattern, we assumed that the GDYN2 was located in the part corresponding to plasmid a. Strains CFNX614, CFNX616, and CFNX620 were characterized in this way.

Construction and characterization of R. etli derivatives harboring integrated plasmids at different genomic locations:
To generate strains carrying an additional deletable zone into a separate endogenous plasmid, we employed plasmid pLPS42b (Table 2). This plasmid harbors a 7.5-kb fragment encompassing the lpsß region from plasmid b (GARCIA-DE LOS SANTOS and BROM 1997 ) cloned into an integrative plasmid. We introduced pLPS42b into strains CFNX600, CFNX601, and CFNX602 through biparental mating, using E. coli S17-1 as a donor. Nal^r Tc^r transconjugants arise through a single crossover between pLPS42b and its homologous target. Strains CFNX611 to CFNX613 were obtained in this way.

To obtain strains with integrated plasmids in the pSym, we introduced by conjugation either plasmid pEYM9 or pEYM10 into strains CFNX600 and CFNX602 using E. coli S17-1 as donor, selecting for Nal^r Tc^r transconjugants. Integration of pEYM9 creates a cointegrate in band 47 of the pSym, whose borders are 4.5-kb long repeats (strains CFNX622 and CFNX624). Integration of pEYM10 generates a cointegrate in band 81 of the pSym, whose borders are 2.5-kb long repeats (strain CFNX623 and CFNX625). Presence of the cointegrated plasmids in all the derivative strains was confirmed by Southern hybridization of plasmid profiles using pBR322 as a probe.

To search for the zone that destabilizes the symbiotic region under active replication from oriV, pEYM12 was introduced by biparental matings with E. coli S17-1, into either strain CFNX55 or strain CFNX88 (harboring different deletions of the pSym; see Table 1), selecting for transconjugants in media containing Nal and Nm; these crosses generated strains CFNX627 and CFNX630.

Evaluation of deletion frequencies:
The GDYN1 element (Fig 1A; ROMERO et al. 1991 ) contains a dose-dependent Km/Sp determinant (positive selection for amplifications) as well as the sacB gene of Bacillus subtilis, which confers sucrose-dependent lethality to gram-negative bacteria. Plating of cells carrying the GDYN1 element on media containing sucrose has been useful for the positive selection of deletions affecting the symbiotic zone (ROMERO et al. 1991 , ROMERO et al. 1995 ). For evaluation of deletion frequencies, at least six single-colony isolates were obtained from each cross done to introduce pEYM5 into the different strains carrying oriV. Since transconjugants were obtained as single-colony isolates, each transconjugant colony represents an independent activation event. These colonies were diluted in MgSO₄ (10 mM)-Tween 40 (0.01%) and plated onto medium with Nal and Cm (for viable cell count) and onto medium containing sucrose to select for deleted derivatives. Deletion frequency is the number of bacteria growing in sucrose-containing medium divided by the number of bacteria growing in NalCm plates. For low copy number plasmids, such as the pSym (one to two copies per cell; RAMIREZ-ROMERO et al. 1997 ), evaluating the recombinant formation frequency by this technique is a satisfactory way to determine the real frequency of the event (CHEDIN et al. 1997 ), since a low number of plasmids per cell reduces not only the probability of having preformed recombinant plasmids, but also competition between parental and recombinant plasmids.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Experimental design:
To evaluate the role of additional rounds of replication in the formation of rearrangements on the pSym, we devised a strategy that is based on the use of: (i) an element that allows a positive selection for the formation of deletions (GDYN1; see MATERIALS AND METHODS and Fig 1A), (ii) a Tn5 derivative that contains a conditional origin of replication (Tn5-GDYN2), and (iii) a plasmid (pEYM5) that permits the activation of the newly introduced origin of replication (Fig 1).

The Tn5-GDYN2 transposon (Fig 1B) contains both the replication origin and the conjugal transfer origin (oriV-oriT) from plasmid RK2 (PANSEGRAU et al. 1994 ). The replication origin of RK2 allows low copy number replication (five to six copies per cell) in a -type, unidirectional mode (MEYER and HELINSKI 1977 ) in many hosts, including Rhizobium (THOMAS and SMITH 1987 ). The main determinant for replication from this origin is the RK2-encoded TrfA protein (THOMAS and SMITH 1987 ). Several broad-host-range vectors have been constructed using only the oriV-RK2 and the trfA gene (THOMAS and SMITH 1987 ; BLATNY et al. 1997 ). Thus, activation of the oriV carried by Tn5-GDYN2 can be controlled by the presence or absence of TrfA. Presence of TrfA was achieved by introduction of pEYM5, which carries a promoterless trfA gene under the control of the PTaclacUV5 promoter (Fig 1D).

To ascertain the functionality of this replication system, plasmid pEYM2, harboring oriV-RK2 as the only replication origin, was introduced by conjugation into R. etli recA⁺ or recA::Sp derivatives carrying pEYM5. Transconjugants carrying pEYM2 were readily obtained (1 x 10^-3 per recipient cell); the presence of both plasmids as separate molecules in these transconjugants was verified by examination of their plasmid profiles (data not shown). In contrast, no transconjugants were observed in similar experiments using recipient strains lacking pEYM5. Transconjugants were seen in pEYM5-harboring recipients despite the absence of the inducer IPTG, suggesting that synthesis of TrfA in absence of the inducer is high enough to activate replication, as observed previously (DURLAND and HELINSKI 1990 ).

Induction of replication from oriV induces a dramatic increase in rearrangement formation in the pSym:
Strains CFNX600, CFNX601, and CFNX602 carry the Tn5-GDYN2 transposon at different locations in the pSym (Fig 1C); pEYM5 was introduced by conjugation into these strains. The resulting transconjugants (at least 30 from each cross) were analyzed for plasmid profile (see MATERIALS AND METHODS).

Upon introduction of pEYM5 into strain CFNX5 (lacking a supernumerary origin), no alteration in plasmid size or abundance was detected. In contrast, upon introduction of this plasmid into strains containing a supernumerary origin on the pSym either inside (CFNX600) or outside (CFNX602) the symbiotic region, 62 out of 70 transconjugant colonies (CFNX600) or 62 out of 66 colonies (CFNX602) analyzed showed an altered plasmid pattern (Fig 2A). The main differences are the absence of a normally sized pSym and an increase in the intensity of the band corresponding to plasmid c. These differences are consistent with a large, 120-kb deletion in the pSym, generating a deleted plasmid that comigrates with plasmid c (ROMERO et al. 1991 ). Southern blotting of the plasmid profile, hybridized against an internal nifH probe, confirmed the presence of a deleted pSym (comigrating with plasmid c) in our strains (Fig 2B). The size of the deleted plasmids (270 kb), suggests that these were generated by recombination between nifHDK region a and nifHDK region b (ROMERO et al. 1991 ), removing the symbiotic region. Confirming this expectation, Southern blots of BamHI-restricted total DNA, probed with an internal nifH fragment, revealed that deleted strains lack the three nifH reiterations observed in the parental strain, having only a single band of altered size (Fig 2C). Its size (9.9 kb) is consistent with a recombination between nifHDK region a and nifHDK region b, thus generating a hybrid band that carries the 5' end of nifHDK region a and the 3' end of nifHDK region b (ROMERO et al. 1991 ).

View larger version (32K): In this window In a new window Download PPT slide   Figure 2. Molecular characterization of recombinant products obtained after supernumerary replication. In each panel, lane 1 is strain CFNX601, while lanes 2–7 show CFNX601 sucrose-resistant derivatives obtained upon introduction of pEYM5. (A) Plasmid profile, stained with ethidium bromide. Location of each plasmid is marked on the left side of the panel. Approximate sizes are as follows: plasmid a (160 kb), b (150 kb), c (270 kb), Sym (390 kb), e (500 kb), and f (600 kb). (B) Southern blot hybridization of the plasmid profile, probed with an internal nifH fragment. Position of the pSym (390 kb) and its deleted derivative (270 kb) is indicated at the left. (C) Southern analysis of total DNA of deleted derivatives. Total DNA of each strain was digested with BamHI, electrophoresed on 1% agarose gels, and probed with an internal nifH fragment. Position of each nif region is shown at the left: region a, 9.0 kb; region b, 5.6 kb; region c::GDYN1, 10.3 kb. On the right, the position of the joint point for the deletions is marked (5'a-3'b, 9.9 kb). Band sizes are according to ROMERO et al. 1991 .

These data were further extended by analyzing over 400 colonies for plasmid profile, coming from at least 20 independent experiments employing strains CFNX600, CFNX601, and CFNX602. In every case, upon activation of supernumerary replication of the pSym, most of the resulting colonies (127 out of 162, strain CFNX600; 89 out of 116, strain CFNX601; 100 out of 144, strain CFNX602) were resistant to sucrose and showed a plasmid profile consistent with a deletion of the symbiotic zone in the pSym. A diagrammatic representation of the results is shown in Fig 3. In these experiments, no selection was applied for strains that had undergone a deletion in the pSym; selection was made only for presence of the replication-activating plasmid. Similar proportions of derivatives lacking the symbiotic region were observed in strains harboring the supernumerary origin inside the symbiotic region (strains CFNX600 or CFNX601) or outside of it (CFNX602). The high frequency of derivatives carrying deletions of the symbiotic region cannot be attributed to selection against the supernumerary origin, because in derivatives of strain CFNX602 the additional origin is preserved on the deleted pSym (Fig 3). These data are best explained as suggesting that supernumerary replication stimulates deletion formation. We dubbed this phenomenon RER, for recombination enhancement by replication.

View larger version (25K): In this window In a new window Download PPT slide   Figure 3. RER is not affected by the position of the supernumerary origin. (Top) Activation of a supernumerary origin (stick and ball symbol, labeled as oriV-RK2) located inside the symbiotic region by introduction of pEYM5 leads to enhanced deletion by recombination between nif regions a and b; the joint point on the resulting plasmid is indicated. (Bottom) Activation of a supernumerary origin located outside the symbiotic region also leads to hyperrecombination between nif regions a and b.

To quantitate the deletion frequency, we collected individual transconjugants from the crosses, scoring the frequency of sucrose-resistant cells in each colony (see MATERIALS AND METHODS). As shown in Table 3, introduction of pEYM5 into a strain lacking the Tn5-GDYN2 element does not alter the frequency of deletions. In contrast, after introduction of pEYM5 into strains harboring a supernumerary replication origin, there is a 1000-fold increase in deletion frequency (Table 3).

The observed increase in deletion frequency was similar regardless of the location of the Tn5-GDYN2 element on the pSym, ruling out positional effects of the new replication origin with respect to the deleted zone. Similar deletion frequencies were found when the additional origin was located at 3.2 kb (strain CFNX601), 32 kb (strain CFNX600), or at 79 kb (strain CFNX602) from the nearest recombining repeat (Table 3). Thus, these results indicate that the magnitude of RER is independent of the distance of the supernumerary origin to the recombining repeats. In fact, strains with insertions in bands 43 or 84 (65 or 54 kb away from the nearest recombining repeat, respectively; GIRARD et al. 1991 ), which are located outside the symbiotic region, also showed a high frequency of deletion (data not shown). In these strains, as well as in CFNX602, the additional origin is still preserved on the deleted pSym.

RER cannot be attributed to incompatibility between the replication origins of the different plasmids. As shown in Table 3, introduction of pEYM5 into CFNX5 (lacking the supernumerary origin) did not increase deletion frequency. Similarly, introduction of a plasmid lacking trfA (pMMB206) into strains CFNX600, CFNX601, or CFNX602 or introduction of an RK2 derivative (pRK7813, which uses oriV-RK2 and TrfA to replicate) into CFNX5 did not increase deletion frequency (Table 3).

RER is independent of the direction of advance of the supernumerary replication fork:
To test whether the enhanced frequency of deletion formation was dependent on the direction of travel of the replication fork, we employed plasmids pEYM12 and pEYM13, which harbor alternative orientations of the unidirectional oriV-RK2 origin. These plasmids also contain nifH region c, as a target for homologous recombination. Cointegration of these plasmids on the pSym generates strains where the oriV-RK2 element is flanked by direct repeats of nifH region c. Depending on the integrative plasmid used, advance of the supernumerary replication fork can occur either counterclockwise (CFNX609) or clockwise (CFNX610) relative to nifHDK region a (Fig 4). Supernumerary replication was primed by the introduction of pEYM5.

View larger version (11K): In this window In a new window Download PPT slide   Figure 4. RER is not affected by the orientation of advance of the supernumerary replication fork. (Top) Linear map of the pSym of R. etli, showing the BamHI sites (small vertical lines); some bands are numbered for reference. The symbiotic region spans from band 1 to 25. (Middle and bottom) In both diagrams (strains CFNX609 and CFNX610), rectangles labeled a, b, and c indicate the nif regions. Cointegrates are shown as solid rectangles joined by lines over each map; arrow, orientation of oriV-RK2. RER effect upon introduction of pEYM5 in these strains is shown at the right, evaluated by analysis of plasmid profiles. Note that in these strains, the symbiotic region or just the supernumerary oriV-RK2 can be deleted independently.

Since these strains lack the GDYN1 element, RER was evaluated by examining pSym size by an in-gel lysis electrophoretic procedure (Eckhardt-type gels, see MATERIALS AND METHODS) and scoring for retention of the oriV-RK2 origin by hybridization against a specific probe. About 50 transconjugants from each construction, from three independent experiments were evaluated. For the construction with oriV in the clockwise direction, ~50% of the transconjugant colonies analyzed had lost the symbiotic zone, while the remaining 50% had lost only the cointegrated plasmid carrying oriV-RK2 (Fig 4). Similar results were obtained for the construction in the counterclockwise orientation (Fig 4). Thus, RER is independent of the direction of movement of the additional replication fork.

RER is dependent on a functional recA allele:
To evaluate the participation of recA in RER, the symbiotic plasmids from strains CFNX600, CFNX601, and CFNX602 were mobilized into strains CFN2001 or its isogenic recA::Sp derivative, CFNX107, as described in MATERIALS AND METHODS. As shown in Table 4, in the recA::Sp derivatives deletion formation under active replication is completely abolished (compare CFNX603 and CFNX605 vs. CFNX606 and CFNX608, respectively). Similar results were obtained with CFNX604 vs. CFNX607 (data not shown). The strict dependence on an active recA product suggests that RER operates through homologous recombination.

RER is not a generalized (i.e., SOS-like) response:
A plausible explanation for RER might be that replication intermediates participate in mounting a SOS-like response, which in turn activates recombination of the whole genome. To test this idea, we cointegrated pLPS42b (Table 2) into the pb plasmid of strains CFNX600, CFNX601, and CFNX602, activating replication afterward by introducing pEYM5 (see MATERIALS AND METHODS and Fig 5, top). If additional replication provokes an SOS-like response (and hence genome-wide recombination), the cointegrated plasmid should become as unstable as the symbiotic region. Retention of the cointegrated plasmid in these experiments was evaluated by the frequency of Tc^r derivatives, while deletion frequency of the symbiotic region was scored by the relative frequency of sucrose-resistant derivatives. As shown in Table 5, although the pSym of strain CFNX611 loses the symbiotic zone at a high frequency upon replication activation, the cointegrate in pb remained stable. Stability of the cointegrated plasmid was further ascertained by testing 100 derivatives from strain CFNX611 for loss of the Tc^r marker; no Tc^s isolates were found. Similar results were obtained for strains CFNX612 and CFNX613 (no Tc^s isolates among 100 derivatives from each strain). Thus, RER does not induce a generalized recombinogenic effect in R. etli, but one restricted to the pSym.

View larger version (21K): In this window In a new window Download PPT slide   Figure 5. Schematic representation of the strains analyzed in Table 5. Most symbols are as in Fig 1 and Fig 3, except cointegrate plasmids, which are depicted as open rectangles (marked as Tcr) joined by lines over each map. Note that in strain CFNX611 the cointegrate is located on plasmid b, whereas in strains CFNX622 and CFNX623 the cointegrate is located on different positions on the pSym.

RER is restricted to the symbiotic zone of the pSym:
In some cases where replication activates recombination, hyperrecombination occurs in a preferred region (HORIUCHI and FUJIMURA 1995 ; CORRE et al. 1997 ). Moreover, the susceptibility to high recombination is preserved even if the specific region is translocated to a new chromosomal position (CORRE et al. 1997 ). Our data suggest that RER displays some regionality, since high recombination upon supernumerary replication is seen only for the symbiotic region. However, this apparent regionality might be attributed to the paucity of appropriately oriented repeats outside the symbiotic region.

To evaluate whether RER displays regionality toward the symbiotic region, integrative plasmids pEYM9 or pEYM10 (harboring a Tc^r marker) were used to generate properly oriented repeats at two different locations in the pSym outside the symbiotic region (strains CFNX622 to CFNX625; see MATERIALS AND METHODS and Table 1). These strains also carry the Tn5-GDYN2 element either inside the symbiotic region (strains CFNX622 and CFNX623; Fig 5, bottom) or outside of it (CFNX624 and CFNX625, not shown) and carry the GDYN1 element in the middle of the symbiotic region. Upon replication activation, the symbiotic region was lost at a high frequency while the cointegrated plasmid in each strain remained stable (Table 5). The cointegrated plasmid appears to be rather stable in these constructions, since no Tc^s isolates were detected upon direct testing of 264 derivatives (CFNX622) or 293 derivatives (CFNX623). Similar results were also obtained with strains CFNX624 and CFNX625, thus ruling out any effect of Tn5-GDYN2 position (data not shown). These results clearly indicate that RER displays regionality, i.e., it is restricted to the symbiotic region.

A further indication of regionality was obtained using plasmid pGM1 (Table 2), a 320-kb self-transferable plasmid in which the whole symbiotic region of the pSym (bands 82–27 of the pSym map; see Fig 6, top) was translocated into plasmid a of R. etli. This plasmid was modified by the introduction of Tn5-GDYN2 at different locations, thus generating strains CFNX614, CFNX616, and CFNX620 (Fig 6). Since these strains lack the GDYN1 element, scoring for deletion was done by analyzing pSym size on Eckhardt-type gels. Upon introduction of pEYM5 into strains CFNX614 and CFNX616, pGM1 loses the symbiotic region at high frequency through recombination at the nifHDK operons, irrespective of the location of the Tn5-GDYN2 element (Fig 6). Interestingly, no deletions were found for strain CFNX620 upon replication activation (Fig 6). In this strain, Tn5-GDYN2 (9.6 kb) is located in nifHDK region a, close to a Km/Gm, Sp/Sm cassette (7.2 kb). Thus, absence of deletions may be attributed to reduced pairing between the recombining repeats, due to these large, heterologous insertions. These results indicate that (i) RER is restricted to the symbiotic region and (ii) all the determinants needed for the observed regionality, besides the additional replication origin, are contained within the symbiotic region.

View larger version (19K): In this window In a new window Download PPT slide   Figure 6. Effect of translocations or deletions of the symbiotic region on RER. (Top) Linear map of the pSym of R. etli, showing the BamHI sites (small vertical lines). Rectangles (labeled a, b, and c) indicate the nif regions; note that strain CFNX630 harbors a single, hybrid nif region (5'a-3'b). For strains with translocations (CFNX614, CFNX616, and CFNX620) the diagrams show plasmid pGM1 (solid line, symbiotic region; hatched sector, plasmid a) carrying Tn5-GDYN2 (stick and triangle symbol, labeled as oriV-RK2) and nifHDK::Km/Gm, Sp/Sm (stick and circle symbol). Additional constructions either on a wild-type pSym (strain CFNX609) or on pSyms harboring different deletions (strains CFNX630 and CFNX627) are also shown (dotted lines, deleted zones). Cointegrates are shown as solid rectangles joined by lines over each map; arrow, orientation of oriV-RK2. RER effect upon introduction of pEYM5 in these strains is shown at the right, evaluated by analysis of plasmid profiles; NA, not applicable.

Deletions that affect the symbiotic region suppress RER:
To determine whether the whole symbiotic region is needed for the regionality characteristic of RER, we took advantage of two deletions of the pSym that affect the symbiotic region (ROMERO et al. 1995 ). Strains carrying these deletions were modified by incorporating oriV-RK2 into the deleted pSym, via a cointegrative plasmid (pEYM12; Fig 6). Since the deletions analyzed remove one of the repeats (nifHDK) that flank the symbiotic region, RER was evaluated by checking the stability of pEYM12 upon activation of the supernumerary origin. Deletions that remove the whole symbiotic region (bands 1–25; Fig 6, strain CFNX630) or a part of it (bands 85–21; Fig 6, strain CFNX627) do not show enhanced instability of pEYM12 upon supernumerary replication (compare with strain CFNX609, Fig 6). This suggests that a determinant necessary for regionality is located between bands 1 and 21.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

In this article we show that -type replication from a supernumerary origin provokes a 1000-fold increase in the formation of deletions in the pSym of R. etli. This effect displays a strong regionality, invariably removing the whole symbiotic region through homologous recombination. This phenomenon, called RER, was unaffected by the position or relative orientation of the introduced origin vis à vis the symbiotic region. RER may also occur in the pSym of other rhizobial species, since induction of replication in the pSym of R. tropici generates deletions at levels comparable to those seen in R. etli (P. MAVINGUI and R. PALACIOS, personal communication); in this case, the recombining repeats (insertion sequences) are 60 kb apart (MAVINGUI et al. 1998 ).

A plausible explanation for RER could be that additional replication increases the copy number of a hypothetical locus within the symbiotic region, with adverse effects for pSym maintenance; this might lead to a strong selection for strains with deletions of the symbiotic region. We believe that such a hypothesis is unlikely because: (i) copy number of the symbiotic region can be increased (up to 12 copies) by tandem amplification without adverse effects on pSym maintenance or cell survival (ROMERO et al. 1991 ; FLORES et al. 1993 ), (ii) in a recA::Sp strain, stable coexistence of the additional origin and the symbiotic region is seen (see Table 4), without curing of the pSym or loss of viability (data not shown), and (iii) stable maintenance of both the symbiotic region and the additional origin is seen, even in a wild-type background, whenever pairing of the recombining repeats is hindered (strain CFNX620; Fig 6). Therefore, we favor the hypothesis that RER is acting by stimulation of regional recombination, caused by additional replication.

One enigmatic aspect of RER that remains to be clarified is related to its magnitude. Under normal replication conditions (1–2 copies per cell; RAMIREZ-ROMERO et al. 1997 ), recombination on the pSym occurs at moderate frequency (10^-4 per viable cell); as shown in this work, overreplication starting from an additional origin (6–8 copies per cell; THOMAS and SMITH 1987 ), provokes a huge increase in homologous recombination, at frequencies approaching 60% per viable cell. Although the reasons for this increase are still unknown, the occurrence of RER suggests that control of copy number of plasmids belonging to the repABC family (RAMIREZ-ROMERO et al. 1997 ), such as the pSym, must be finely tuned, in order to avoid enhanced instability. In this sense, RER is akin to other situations found in mammalian cells, where overreplication may provoke recombinational instability in specific genomic regions (SCHIMKE et al. 1986 ).

It is interesting that RER leads to the preferential recovery of only one of the possible recombination products. Invariably, the huge increase in deletion frequency characteristic of RER is not accompanied by a similar increase in the frequency of amplifications (ROMERO et al. 1991 , ROMERO et al. 1995 ), products to be expected to increase in frequency if RER is acting through a reciprocal recombination process (data not shown). Similarly, high levels of excised circles (FLORES et al. 1993 ) were not observed, even though some of them, originating from strains with insertions of Tn5-GDYN2 in the symbiotic region, would carry an active origin. This finding could be explained assuming that both products (deletions and amplifications) are generated during the initial stages of RER; however, further rounds of recombination among the amplified products would tend to diminish their abundance, concomitantly increasing the frequency of deletions. Alternatively, it could be argued that recombination induced by RER is not reciprocal. In this regard, recombination through nonreciprocal models, such as the half crossing-over model (MAHAN and ROTH 1989 ; KOBAYASHI 1992 ) does not generate an amplification with every deletion. Work is in progress to try to settle the issue of reciprocality by analyzing the products obtained at early stages after activation of replication.

Previous evaluations of the effect of additional -type replication, using artificial constructs on the chromosome of B. subtilis, showed only a modest effect (two- to fourfold) on recombination. RER is more comparable to the effects caused by inducing additional rolling circle-type replication in the chromosome of B. subtilis (NOIROT et al. 1987 ; PETIT et al. 1992 ). In both cases, however, recombination was diminished by factors that do not affect RER, such as increasing the distance from the additional replication origin (either - or rolling circle-type) to the recombining repeats or by modifying the relative orientation of the origin (NOIROT et al. 1987 ; PETIT et al. 1992 ; MOREL-DEVILLE and EHRLICH 1996 ). Additionally, none of these examples show the regionality characteristic of RER. These differences could be due to the nature of the system, i.e., chromosome vs. plasmid, the type and size of repeats, or both.

The characteristics observed with RER are reminiscent of the hyperrecombination observed along the replication terminus zone of the E. coli chromosome (LOUARN et al. 1991 ; HORIUCHI and FUJIMURA 1995 ; CORRE et al. 1997 ). Similarities between RER and hyperrecombination in the terminus zone include the following: (i) both depend on a -type replication process (unidirectional for RER, bidirectional for the terminus zone), (ii) both exhibit a similar increase in recA-dependent recombination frequency, (iii) both display clear regionality, extending for at least 280 kb in the E. coli chromosome (HORIUCHI and FUJIMURA 1995 ) and 120 kb for RER, and (iv) regionality is maintained after translocating the corresponding region to a new genomic environment (CORRE et al. 1997 ). It has been proposed that structural features of the termination region favor the formation of catenation links, which are later on resolved by RecBCD-mediated recombination; this is the so-called terminal strand invasion model (CORRE et al. 1997 ).

Although the terminal strand invasion model may explain RER, an alternative possibility would be the existence of a site within the symbiotic region that is prone to replication-mediated generation of persistent single- or double-strand breaks. Both lesions would impede the advance of a replication fork; correction of these lesions is frequently achieved by recombination (LOUARN et al. 1991 ; HORIUCHI and FUJIMURA 1995 ; KUZMINOV 1995 ; CORRE et al. 1997 ). Under this view, additional replication instigates recombination by provoking a persistent cut on a site (within the symbiotic region) other than the additional oriV. The high frequency of recombination associated with RER can be easily explained, since double-strand breaks would allow the loading of recombination enzymes, such as RecBCD, thus making the molecule highly susceptible to homologous recombination. Both regionality and the preferential recovery of deletions affecting the symbiotic region are expected under this model, because one of the products of reciprocal excisive recombination (a circle containing the symbiotic region) would be frequently cut. Resolution between these alternatives must wait for the identification and characterization of the site that conditions RER.

	ACKNOWLEDGMENTS

We are indebted to Csaba Koncz, Donald Helinski, and Victor Morales for providing strains and plasmids, to Susana Brom, Jaime Martínez-Salazar, Patrick Mavingui, and Rafael Palacios for critical reviewing of the manuscript, and to Laura Cervantes, Javier Rivera and María de la Paz Salas for skillful technical assistance. Partial financial support was provided by grants 4321-N9406 from the Consejo Nacional de Ciencia y Tecnología (México), IN203897 from the Programa de Apoyo a Proyectos de Investigación e Innovación Tecnológica, and 030331 from the Programa de Apoyo a las Divisiones de Estudios de Posgrado (both from Universidad Nacional Autónoma de México). E.V.-M. received scholarships from the Consejo Nacional de Ciencia y Tecnología (México) and Dirección General de Estudios de Posgrado (Universidad Nacional Autónoma de México).

Manuscript received July 28, 1999; Accepted for publication November 22, 1999.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ALEXEYEV, M. F., I. N. SHOKOLENKO, and T. P. CROUGHAN, 1995  Improved antibiotic resistance cassettes and elements for Escherichia coli vector construction and in vitro deletion/insertion mutagenesis. Gene 160:63-67[Medline].

ASAI, T., D. B. BATES, and T. KOGOMA, 1994  DNA replication triggered by double-stranded breaks in E. coli: dependence on homologous recombination functions. Cell 78:1051-1061[Medline].

BIERNE, H. and B. MICHEL, 1994  When replication forks stop. Mol. Microbiol. 13:17-23[Medline].

BIERNE, H., S. D. EHRLICH, and B. MICHEL, 1997  Deletions at stalled replication forks occur by two different pathways. EMBO J. 16:3332-3340[Medline].

BLATNY, J. M., T. BRAUTASET, H. C. WINTHER-LARSEN, P. KARUNAKARAN, and S. VALLA, 1997  Improved broad-host-range RK2 vectors useful for high and low regulated gene expression levels in gram-negative bacteria. Plasmid 38:35-51[Medline].

BOYER, H. W. and D. ROULLAND-DUSSOIX, 1969  A complementation analysis of restriction and modification of DNA in Escherichia coli.. J. Mol. Biol. 41:459-472[Medline].

BROM, S. and BROM, S.A. GARCÍA-DE LOS SANTOS, T. STEPKOWSKI, M. FLORES, G. DÁVILA et al., 1992  Different plasmids of Rhizobium leguminosarum bv. phaseoli are required for optimal symbiotic performance. J. Bacteriol. 174:5183-5189[Abstract/Free Full Text].

CHÉDIN, F., R. DERVYN, S. D. EHRLICH, and P. NOIROT, 1997  Apparent and real recombination frequencies in multicopy plasmids: the need for a novel approach in frequency determination. J. Bacteriol. 179:754-761[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].

DURLAND, R. H. and D. R. HELINSKI, 1990  Replication of the broad-host-range plasmid RK2: direct measurement of intracellular concentrations of the essential TrfA replication proteins and their effect on plasmid copy number. J. Bacteriol. 172:3849-3858[Abstract/Free Full Text].

DURLAND, R. H., A. TOUKDARIAN, F. FANG, and D. HELINSKI, 1990  Mutations in the trfA replication gene of the broad-host-range plasmid RK2 result in elevated plasmid copy numbers. J. Bacteriol. 172:3859-3867[Abstract/Free Full Text].

ECKHARDT, T., 1978  A rapid method for the identification of plasmid deoxyribonucleic acid in bacteria. Plasmid 1:584-588[Medline].

ELIASSON, A., K. NORDSTRÖM, and R. BERNANDER, 1996  Escherichia coli strains in which chromosome replication is controlled by a P1 or F replicon integrated into oriC.. Mol. Microbiol. 20:1013-1023[Medline].

FANG, F. C., R. H. DURLAND, and D. R. HELINSKI, 1993  Mutations in the gene encoding the replication-initiation protein of plasmid RK2 produce elevated copy numbers of RK2 derivatives in Escherichia coli and distantly related bacteria. Gene 133:1-8[Medline].

FEINBERG, A. P. and B. VOGELSTEIN, 1983  A technique for radiolabelling DNA restriction endonuclease fragments to high specific activity. Anal. Biochem. 132:6-13[Medline].

FIGURSKI, D. H. and D. R. HELINSKI, 1979  Replication of an origin-containing derivative of plasmid RK2 dependent on a plasmid function provided in trans.. Proc. Natl. Acad. Sci. USA 76:1648-1652[Abstract/Free Full Text].

FLORES, M., V. GONZÁLEZ, S. BROM, E. MARTÍNEZ, and D. PIÑERO et al., 1987  Reiterated DNA sequences in Rhizobium and Agrobacterium spp. J. Bacteriol. 169:5782-5788[Abstract/Free Full Text].

FLORES, M., S. BROM, T. STEPKOWSKI, M. L. GIRARD, and G. DÁVILA et al., 1993  Gene amplification in Rhizobium: identification and in vivo cloning of discrete amplifiable DNA regions (amplicons) from Rhizobium leguminosarum biovar phaseoli.. Proc. Natl. Acad. Sci. USA. 90:4932-4936[Abstract/Free Full Text].

FREIBERG, C., R. FELLAY, A. BAIROCH, W. BROUGHTON, and A. ROSENTHAL et al., 1997  Molecular basis of symbiosis between Rhizobium and legumes. Nature 387:394-401[Medline].

GARCÍA-DE LOS SANTOS, A. and S. BROM, 1997  Characterization of two plasmid-borne lpsß loci of Rhizobium etli required for lipopolisaccharide synthesis and for optimal interaction with plants. Mol. Plant-Microbe Interact. 7:891-902.

Rhizobium plasmids in bacteria-legume interactions. (1996) World J. Microbiol. Biotechnol. 12:119-125.

GIRARD, M. L., M. FLORES, S. BROM, D. ROMERO, and R. PALACIOS et al., 1991  Structural complexity of the symbiotic plasmid of Rhizobium leguminosarum bv. phaseoli. J. Bacteriol. 173:2411-2419[Abstract/Free Full Text].

HANAHAN, D., 1983  Studies on transformation of Escherichia coli plasmids. J. Mol. Biol. 166:407-428[Medline].

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

JONES, J. D. G. and N. GUTTERSON, 1987  An efficient mobilizable cosmid vector, pRK7813, and its use in a rapid method for marker exchange in Pseudomonas fluorescens HV37a. Gene 61:299-306[Medline].

KOBAYASHI, I., 1992  Mechanisms for gene conversion and homologous recombination: the double-strand break repair model and the successive half crossing-over model. Adv. Biophys. 28:81-133[Medline].

KOGOMA, T., 1997  Stable DNA replication: interplay between DNA replication, homologous recombination, and transcription. Microbiol. Mol. Biol. Rev. 61:212-238[Abstract].

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

KREUZER, K. N., M. SAUNDERS, L. J. WEISLO, and H. W. E. KREUZER, 1995  Recombination-dependent DNA replication stimulated by double-strand breaks in bacteriophage T4. J. Bacteriol. 177:6844-6853[Abstract/Free Full Text].

KUZMINOV, A., 1995  Collapse and repair of replication forks in Escherichia coli.. Mol. Microbiol. 16:373-384[Medline].

KUZMINOV, A. and F. W. STAHL, 1999  Double-strand end repair via the RecBC pathway in Escherichia coli primes DNA replication. Genes Dev. 13:345-356[Abstract/Free Full Text].

LOUARN, J., J. PATTE, and J.-M. LOUARN, 1982  Suppression of Escherichia coli dnaA46 mutations 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. LOUARN, V. FRANÇOIS, 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].

LOVETT, S. T. and V. V. FESCHENKO, 1996  Stabilization of diverged tandem repeats by mismatch repair: evidence for deletion formation via a misaligned replication intermediate. Proc. Natl. Acad. Sci. USA. 93:7120-7124[Abstract/Free Full Text].

MAHAN, M. J. and J. R. ROTH, 1989  Role of recBC function in formation of chromosomal rearrangements: a two-step model for recombination. Genetics 121:433-443[Abstract/Free Full Text].

MALKOVA, A., E. L. IVANOV, and J. E. HABER, 1996  Double-strand break repair in the absence of RAD51 in yeast: a possible role for break-induced DNA replication. Proc. Natl. Acad. Sci. USA 93:7130-7136.

MARSH, J. L., M. ERFLE, and E. J. WYKES, 1984  The pIC plasmid and phage vectors with versatile cloning sites for recombinant selection by insertional inactivation. Gene 32:481-485[Medline].

MARTÍNEZ-SALAZAR, J., D. ROMERO, M. L. GIRARD, and G. DÁVILA, 1991  Molecular cloning and characterization of the recA gene of Rhizobium phaseoli and construction of recA mutants. J. Bacteriol. 173:3035-3040[Abstract/Free Full Text].

MAVINGUI, P., T. LAEREMANS, M. FLORES, D. ROMERO, and E. MARTÍNEZ-ROMERO et al., 1998  Genes essential for Nod factor production and nodulation are located on a symbiotic amplicon (AMPRtr CFN299pc60) in Rhizobium tropici. J. Bacteriol. 180:2866-2874[Abstract/Free Full Text].

MEYER, R. J. and D. R. HELINSKI, 1977  Unidirectional replication of the P-group plasmid RK2. Biochim. Biophys. Acta 478:109-113[Medline].

MILLER, J. H., 1972 Experiments in Molecular Genetics. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

MORALES, V. M., A. BACKMAN, and M. BAGDASARIAN, 1991  A series of wide-host-range low copy-number vectors that allow direct screening for recombinants. Gene 97:39-47[Medline].

MOREL-DEVILLE, F. and S. D. EHRLICH, 1996  Theta-type DNA replication stimulates homologous recombination in the Bacillus subtilis chromosome. Mol. Microbiol. 19:587-598[Medline].

NOEL, K. D., F. SÁNCHEZ, L. FERNÁNDEZ, J. LEEMANS, and M. A. CEVALLOS, 1984  Rhizobium phaseoli symbiotic mutants with transposon Tn5 insertions. J. Bacteriol. 158:148-155[Abstract/Free Full Text].

NOIROT, P., M.-A. PETIT, and S. D. EHRLICH, 1987  Plasmid replication stimulates recombination in Bacillus subtilis.. J. Mol. Biol. 196:39-48[Medline].

PANSEGRAU, W., E. LANKA, P. T. BARTH, D. H. FIGURSKI, and D. G. GUINEY et al., 1994  Complete nucleotide sequence of the Birmingham IncP plasmids: compilation and comparative analysis. J. Mol. Biol. 239:623-663[Medline].

PETIT, M.-A., J. M. MESAS, P. NOIROT, F. MOREL-DEVILLE, and S. D. EHRLICH, 1992  Induction of DNA amplification in the Bacillus subtilis chromosome. EMBO J. 11:1317-1326[Medline].

QUINTO, C., H. DE LA VEGA, M. FLORES, L. FERNÁNDEZ, and T. BALLADO et al., 1982  Reiteration of nitrogen fixation gene sequences in Rhizobium phaseoli.. Nature 299:724-726.

RAMÍREZ-ROMERO, M. A., P. BUSTOS, L. GIRARD, O. RODRÍGUEZ, and M. A. CEVALLOS et al., 1997  Sequence, localization and characteristics of the replicator region of the symbiotic plasmid of Rhizobium etli.. Microbiology 143:2825-2831[Abstract/Free Full Text].

RATTRAY, A. J. and L. S. SYMINGTON, 1993  Stimulation of meiotic recombination in yeast by an ARS element. Genetics 134:175-188[Abstract].

RODRÍGUEZ, C. and D. ROMERO, 1998  Multiple recombination events maintain sequence identity among members of the nitrogenase multigene family in Rhizobium etli.. Genetics 149:785-794[Abstract/Free Full Text].

ROMERO, D. and R. PALACIOS, 1997  Gene amplification and genomic plasticity in prokaryotes. Annu. Rev. Genet. 31:91-111[Medline].

ROMERO, D., S. BROM, J. MARTÍNEZ-SALAZAR, M. L. GIRARD, and R. PALACIOS et al., 1991  Amplification and deletion of a nod-nif region in the symbiotic plasmid of Rhizobium phaseoli.. J. Bacteriol. 173:2435-2441[Abstract/Free Full Text].

ROMERO, D., J. MARTÍNEZ-SALAZAR, L. GIRARD, S. BROM, and G. DÁVILA et al., 1995  Discrete amplifiable regions (amplicons) in the symbiotic plasmid of Rhizobium etli CFN42. J. Bacteriol. 177:973-980[Abstract/Free Full Text].

ROMERO, D., G. DÁVILA and R. PALACIOS, 1997 The dynamic genome of Rhizobium, pp. 153–161 in Bacterial Genomes: Physical Structure and Analysis, edited by F. J. DE BRUIJN, J. R. LUPSKI and G. WEINSTOCK, Chapman & Hall, New York.

ROSENBERG, C. and T. HUGHET, 1984  The pATC58 plasmid of Agrobacterium tumefaciens is not essential for tumor induction. Mol. Gen. Genet. 196:533-536.

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

SAVESON, C. J. and S. T. LOVETT, 1997  Enhanced deletion formation by aberrant replication in Escherichia coli.. Genetics 146:457-470[Abstract].

SCHIMKE, R. T., S. W. SHERWOOD, A. B. HILL, and R. N. JOHNSTON, 1986  Overreplication and recombination of DNA in higher eukaryotes: potential consequences and biological implications. Proc. Natl. Acad. Sci. USA 83:2157-2161[Abstract/Free Full Text].

SIMON, R., 1984  High frequency mobilization of gram-negative bacterial replicons by the in vitro constructed Tn5-Mob transposon. Mol. Gen. Genet. 196:413-420[Medline].

SIMON, R., U. PRIEFER, and A. PÜHLER, 1983  A broad host range mobilization system for in vivo genetic engineering: transposon mutagenesis in gram-negative bacteria. Biotechnology 1:784-791.

STAHL, F., 1996  Meiotic recombination in yeast: coronation of the double-strand-break repair model. Cell 87:965-968[Medline].

STAHL, F. W., 1994  The Holliday junction on its thirtieth anniversary. Genetics 138:241-246[Medline].

THOMAS, C. M. and C. A. SMITH, 1987  Incompatibility group P plasmids: genetics, evolution, and use in genetic manipulation. Annu. Rev. Microbiol. 41:77-101[Medline].

YAMAGUCHI, K. and J. TOMIZAWA, 1980  Establishment of Escherichia coli cells with an integrated high copy number plasmid. Mol. Gen. Genet. 178:525-533[Medline].

This article has been cited by other articles:

				G. Santoyo, J. M. Martinez-Salazar, C. Rodriguez, and D. Romero Gene Conversion Tracts Associated with Crossovers in Rhizobium etli J. Bacteriol., June 15, 2005; 187(12): 4116 - 4126. [Abstract] [Full Text] [PDF]

				J. Zuniga-Castillo, D. Romero, and J. M. Martinez-Salazar The Recombination Genes addAB Are Not Restricted to Gram-Positive Bacteria: Genetic Analysis of the Recombination Initiation Enzymes RecF and AddAB in Rhizobium etli J. Bacteriol., December 1, 2004; 186(23): 7905 - 7913. [Abstract] [Full Text] [PDF]