Proc. Natl. Acad. Sci. USA Vol. 90, pp. 4932-4936, June 1993 amplification in Rhizobium: Identification and in vivo cloning of discrete amplifiable DNA regions () from Rhizobium leguminosarum biovar phaseoli (genomic rearrangements/genome dynamics/genome plasticity/symbiosis/nitrogen fixation)

MARGARITA FLORES, SUSANA BROM, TOMASZ STEPKOWSKI, MARiA DE LOURDES GIRARD, GUILLERMO DAVILA, DAVID ROMERO, AND RAFAEL PALACIOS Departamento de Gendtica Molecular, Centro de Investigaci6n sobre Fijaci6n de Nitr6geno, Universidad Nacional Aut6noma de M6xico, Apartado postal 565-A, Cuernavaca, Mor. Mexico Communicated by Donald R. Helinski, January 25, 1993

ABSTRACT A genetic element that allows the positive MATERIALS AND METHODS selection of different genomic rearrangements was used to analyze DNA amplification in Rhizobium leguminosarum bio- Construction of TnS-GDYNl. This transposon derivative var phaseoli. Discrete amplifiable DNA regions (amplicons) was constructed by in vitro manipulation of the TnS insert in were detected in different regions of the genome of the model pSUP2021, which is a mobilizable pBR325 derivative (6). The central Bgl II-Bgl II fragment of TnS in pSUP2021 was strain CFN42, including the chromosome and several large replaced by a 5.7-kb BamHI-BamHI fragment ofpGUS3 (5), plasmids. Amplicons were mobilized into Escherichia coli using carrying the GDYN1 cassette. After ligation, the mixture was a genetic approach that involves the introduction of an origin transformed into E. coli HB101, selecting for recombinant ofreplication active in E. coli and an origin ofconjugal transfer plasmids in LB medium containing chloramphenicol and into the amplifiable DNA regions oftheRhizobium genome. The spectinomycin. Colonies resistant to chloramphenicol and strategy can be a valuable tool for studies on genome organi- spectinomycin that showed sensitivity to sucrose were zation and function. We propose that amplicons define a screened for plasmids with the desired restriction pattern, structural characteristic of the genome that may play an and one of these was named pDR21. important biological role. Strains and Genetic Manipulations. All R. phaseoli strains used were derived from wild-type CFN42, which is naturally Rhizobium species have stimulated scientific interest due to resistant to nalidixic acid (7). Derivatives containing TnS- their ability to interact with plants establishing nitrogen- GDYN1 insertions were obtained by mating E. coli S-17 fixing symbioses. The genome ofdifferent Rhizobium species strain containing plasmid pDR21, which carries the transpo- contains a large amount of reiterated DNA sequences that son, with R. phaseoli strain CFN42 and selecting transcon- include complete operons, specific , regulatory se- jugants resistant to nalidixic acid and spectinomycin. Strains quences, and insertion sequences (for see ref. CFNX201, CFNX207, CFNX209, and CFNX211 contain review, 1). TnS-GDYN1 in plasmid a (pa). Strains CFNX203 and Presumably due to the presence of reiterated DNA se- CFNX205 contain the transposon in plasmid d (pd) and in the quences, the genome of some Rhizobium strains is subjected chromosome, respectively. to frequent genomic rearrangements (2-5). To isolate derivatives carrying DNA amplifications, strains We have recently constructed a genetic element (the containing Tn5-GDYN1 insertions were serially cultured in GDYN1 cassette) that allows the positive selection of differ- liquid PY medium (0.5% peptone/0.3% yeast extract/10 mM ent types of genomic rearrangements (5). This element con- CaCl2) at 30°C for periods of 24 hr as follows: once in the tains the kanamycin/gentamycin and spectinomycin/ presence of spectinomycin at 75 ,.g/ml, twice with kanamy- streptomycin resistance markers from plasmid pSa. In Rhizo- cin at 50 pg/ml, and twice with kanamycin at 100 pg/ml. bium leguminosarum biovar phaseoli (R. phaseoli), the Cells were then plated on PY medium containing kanamycin symbiont of the common bean plant Phaseolus vulgaris, at 125 jig/ml, and individual colonies were isolated. Strains these genes confer resistance to high levels of spectinomycin CFNX202, CFNX204, CFNX206, CFNX208, CFNX210, and to low levels of kanamycin. Because the level of kan- and CFNX212 are amplified derivatives from strains amycin resistance increases with gene dosage, the element CFNX201, CFNX203, CFNX205, CFNX207, CFNX209, provides a positive selection system to detect variants with and CFNX211, respectively. Strain CFNX206A is a deriva- amplified DNA regions. The GDYN1 cassette was used to tive of CFNX206 resistant to kanamycin at 1 mg/ml. demonstrate high-frequency amplification and deletion The procedure to clone in E. coli DNA sequences from R. events of a 120-kb region located in the symbiotic plasmid of phaseoli strains carrying amplifications is presented in Re- R. phaseoli (5). sults. Strains CFNC1, CFNC2, CFNC3, CFNC4, CFNC5, In the present study we used a GDYN1 derivative (TnS- and CFNC6 are E. coli strains containing DNA from Rhizo- GDYN1) to analyze gene amplification events in different bium strains CFNX202, CFNX204, CFNX206, CFNX208, regions ofthe genome ofR. phaseoli strain CFN42, including CFNX210, and CFNX212, respectively. the DNA Electrophoresis, Filter Blot Hybridization, and Quan- chromosome and several large plasmids. Discrete ampli- tifi'cation of DNA Amplification. Total DNA or cosmid DNA fiable DNA regions, referred to here as amplicons, were was digested with BamHI, subjected to electrophoresis in 1% identified and mobilized from R. phaseoli to Escherichia coli agarose gels, blotted onto nitrocellulose, and hybridized as by an in vivo genetic procedure. described (3). Plasmid profiles were obtained by the in-gel lysis method of Eckhardt (8), blotted onto nitrocellulose, and The publication costs of this article were defrayed in part by page charge hybridized similarly. Probes were labeled with 32P by the payment. This article must therefore be hereby marked "advertisement" nick-translation procedure (9). To quantify DNA amplifica- in accordance with 18 U.S.C. §1734 solely to indicate this fact. tion, blots were hybridized against a mixed probe containing 4932 Downloaded by guest on September 29, 2021 Genetics: Flores et al. Proc. Natl. Acad. Sci. USA 90 (1993) 4933

the transposon and a 300-bp fragment of R. phaseoli ribo- A 1 2 3 4 5 6 7 8 somal DNA (10) obtained by PCR. Blots were subjected to Kb autoradiography, and signals were integrated by scanning 30 densitometry. 10 ...... ;K- RESULTS 3 Selection for Gene Amplification Events. The TnS-GDYN1 k4 was randomly introduced into R. phaseoli strain CFN42 as described in Materials and Methods. Derivatives containing the transposon were selected by their resistance to specti- nomycin. The genome of CFN42 contains a chromosome and six large plasmids (pa-pf) ranging in size from 150 to 600 kb; pd (390 kb) has been identified as the symbiotic plasmid (11, U*b... ; 12). The location of the transposon in different spectinomy- B cin-resistant clones was identified by hybridization of plas- Kb mid profiles with a DNA probe that recognizes the transpo- 30 son (data not shown). In two independent experiments a total 10 of 30 derivatives of CFN42 were used to search for DNA amplification events. From these derivatives, 13 carried the transposon in pa, 4 carried the transposon in pb, 2 derivatives 3 , carried it in pc, 1 carried the transposon in pd, 2 derivatives carried it in pe, 2 carried the transposon in pf, and 6 derivatives carried the transposon in the chromosome. To select for amplification events, cells were cultured in liquid medium in the presence of increased concentrations of kan- amycin (see Materials and Methods). To screen for amplification, total DNA from the original TnS-GDYN1 insertions, as well as from the kanamycin- resistant clones derived from them, was digested with restric- C tion endonuclease Southern blots oftotal DNA were Kb BamHI; 30, prepared and hybridized with a mixed probe containing a 10 .1 ribosomal DNA sequence from R. phaseoli and a DNA f * a, :... fragment that recognizes the transposon (see Materials and Aw: do Methods). The ribosomal DNA sequence hybridized to three BamHI bands. Because the TnS-GDYN1 lacks BamHI sites, 3 the insertions appeared as single fragments of different sizes containing the whole transposon and adjacent genomic se- quences. In amplification events, the intensity ofthe fragment corresponding to the insertion increased relative to those f-..I...4 corresponding to ribosomal DNA (see examples in Fig. 1, lanes 3 and 4). The level of amplification was determined by comparative densitometry. From the 30 insertions analyzed, 14 cases ofamplification were detected: 7 cases corresponded FIG. 1. Analysis of DNA amplification events in R. phaseoli. to insertions in pa, 3 cases corresponded to insertions in pb, Total DNA digests or plasmid profiles of strains harboring TnS- GDYN1 insertions and their corresponding amplified derivatives 1 case corresponded to insertions in pd, 1 case corresponded were analyzed by ethidium bromide staining and hybridized against to insertions in pe, 1 case corresponded to insertions in pf, and different probes. (A) Amplification event in pa: strain CFNX201, 1 case corresponded to insertions in the chromosome. original TnS-GDYN1 insertion; strain CFNX202, amplified variant. Characterization of DNA Amplification Events. Fig. 1 pre- (B) Amplification event in pd: strain CFNX203, original insertion; sents the analysis ofthree different amplification events from strain CFNX204, amplified variant. (C) Amplification event in the insertions in pa (A), in pd (B), and in the chromosome (C), chromosome: strain CFNX205, original insertion; strain CFNX206, respectively. In A-C, lanes 3 and 4 show the hybridization of amplified variant. Lanes: 1 and 2, ethidium bromide strain of total genomic Southern blots from the original insertion and the BamHI-digested DNA from original insertion and amplified variants, variant, respectively, against the ribosomal respectively; 3 and 4, hybridization of total DNA of original inser- amplified DNA/ tions and amplified variants, respectively, against a mixed probe TnS-GDYN1 mixed probe; the estimated amplification level containing transposon and ribosomal DNA (arrows show position of is indicated. In all cases the amplified BamHI fragment fragment corresponding to the insertion, and the number indicates corresponding to the insertion of TnS-GDYN1 was of the amplification factor); 5 and 6, hybridization of total DNA of original same size as that of the original insertion. This result suggests insertions and amplified variants, respectively, against either pa (A) that the transposon sequence is not participating as an or pd (B) (arrows indicate amplified fragments); 7, ethidium bromide endpoint of the amplified regions. Lanes 1 and 2 show the stain of plasmid profiles of amplified variants (positions of the total DNA patterns stained with ethidium bromide. The different plasmids are indicated); 8, hybridization of plasmid profile patterns of the original insertions were not distinguishable ofamplified variants, with either pa (A), pd (B), or Tn5-GDYNl (C). from each other (lanes 1) or from the pattern ofthe wild-type strain (data not shown). In contrast, the pattern of clones tion, plasmid profiles were analyzed by Eckhardt-type gels, bearing amplified regions (lanes 2) revealed some differences and blots of such gels were hybridized with the purified specific to each amplified strain. plasmids. The different plasmids were purified from Agro- To further analyze location and extent of the amplified bacterium tumefaciens transconjugants containing pa, pb, genomic regions, purified plasmids of strain CFN42 were pc, pd, or pf, as described (13); pe was not used in these used as probes to hybridize against genomic Southern blots experiments. The original nonamplified strains carrying the of all the original insertions and amplified variants. In addi- insertions showed no differences among them or with the Downloaded by guest on September 29, 2021 4934 Genetics: Flores et al. Proc. Natl. Acad. Sci. USA 90 (1993) wild-type strain, except for the hybridization due to the presence of the transposon (data not shown). Amplified A -- variants showed specific differences in the plasmid profiles or hybridization patterns; these differences were limited to the particular replicon in which the original TnS-GDYN1 inser- tion was present. In the amplifications derived from original insertions in pa B (Fig. 1A) or in pd (Fig. 1B) the Southern blot hybridization using the corresponding plasmid as probe showed an increase in the intensity of some BamHI fragments (see arrows in lanes 6). This result suggests an amplification of a discrete region of pa or pd. The plasmid profile showed the absence of the corresponding plasmid in its original position (lanes 7). -~ 0 - -. ~- When the purified plasmids were used as hybridization probes against the plasmid profiles, structures larger than the corresponding plasmid were revealed in the amplified strains Neo Sp (lanes 8). Interestingly, structures of smaller size than the D t f original plasmids were also revealed (see below). In the amplification derived from an insertion in the chromosome (Fig. 1C), a remarkable difference was evi- denced by hybridization of the plasmid profile against the transposon probe (Fig. 1C, lane 8): a ladder of structures of FIG. 2. In vivo cloning of amplified DNA sequences from Rhizo- different sizes was revealed. These structures, as well as bium into E. coli. Schematic representation of structures involved in the genetic approach (see text) in the Rhizobium strain bearing an those shown in Fig. 1 A and B (lanes 8), might be derived from amplification of a discrete DNA region (A), pSUP5011 (B), the recombination events in tandemly amplified regions that Rhizobium transconjugant that has incorporated pSUP5011 sequences generate monomers or multimers of the amplified unit. The by recombination mediated by the insertion sequences ofTnS (C), the presence of such structures suggested an experimental strat- E. coli transconjugant that has incorporated Rhizobium DNA se- egy to clone Rhizobium amplicons in E. coli. quences (D). , Rhizobium sequences within ; -- -, In Vivo Cloning of Amplified DNA Sequences from R. Rhizobium sequences outside amplicon; -, sequences from TnS- phaseoli into E. coli. The rationale of the strategy to clone GDYN1 or pSUP5011; _, insertion sequences from TnS-GDYN1; a, insertion sequences from pSUP5011. Sp, spectinomycin; Neo, neo- complete amplicons in E. coli is based upon the possibility Ori, origin of replication. that amplified regions are dynamic structures that, by ho- mycin; Clm, chloramphenicol; mologous recombination in any site along the amplified unit, (CFNX204), and in the chromosome (CFNX206A). To de- may generate structures containing the whole unit, either in termine whether the E. coli transconjugants contained ge- a monomeric or in a multimeric form. If an origin of repli- nomic sequences from R. phaseoli, extrachromosomal DNA cation active in E. coli and an origin ofconjugal transfer (Mob from the transconjugants was isolated by an alkaline lysis element) are introduced in such structures, the amplicon method, and the preparations were used as hybridization sequences might be transferred by conjugation and estab- probes against genomic Southern blots of the wild-type lished in E. coli as a plasmid. CFN42 strain and its derivatives. The DNA preparations The structures involved in the approach used to clone R. from E. coli showed specific hybridization patterns depend- phaseoli amplicons are schematized in Fig. 2. Rhizobium ing on the starting Rhizobium strain (see below). strains harboring amplifications of discrete DNA regions Complete Amplicon Sequences Are Transferred to E. coli. were used as recipients for conjugation with an E. coli strain As shown (Fig. 1, lanes 6) the extension ofan amplified DNA containing plasmid pSUP5011 (6). This plasmid has a TnS- can be revealed using the whole replicon where the Mob element, which contains a neomycin-resistance marker, region and pBR325, which has the origin of replication and a amplification occurs as a hybridization probe against ge- chloramphenicol-resistance gene. Each unit in the amplified nomic Southern blots ofthe original strains and the amplified region of the Rhizobium strain contains the TnS-GDYN1 derivatives. The restriction fragments corresponding to the element that confers resistance to spectinomycin; the R. amplified unit increase in intensity in the amplified strain. phaseoli strain used is resistant to nalidixic acid. To obtain Four amplification events derived from different Tn5- Rhizobium derivatives containing both the Mob element and GDYN1 insertions in pa were analyzed. Hybridization the origin of replication in the amplified region, transconju- against pa revealed different sets ofamplified fragments (Fig. gants resistant to nalidixic acid, spectinomycin, neomycin, 3 A-D). This result indicates that different amplicons are and chloramphenicol were selected. Such an antibiotic- present in a single replicon. The protocol described to resistance pattern is compatible with cointegration of the transfer amplicon sequences to E. coli was applied to the four complete pSUP5011 sequence with the TnS-GDYN1 by a strains. The extrachromosomal DNA isolated from each of recombination event mediated by the homologous regions the corresponding E. coli strains was used as hybridization shared by both elements (the insertion sequences of Tn5). probe against the Rhizobium strain containing the amplifica- The Rhizobium transconjugants were then used as donors tion. In all cases the fragments revealed as amplified were the for conjugation into a recipient E. coli strain resistant to same using as hybridization probe either the whole pa or the erythromycin in a triparental cross that included HB101 E. coli DNA (Fig. 3 A-D). These results indicate that containing pRK2013 (14). E. coli transconjugants resistant to complete amplicon sequences were transferred from Rhizo- erythromycin, spectinomycin, neomycin, and chloramphen- bium to E. coli. icol were selected; such transconjugants contain the se- To appreciate the amplified fragments for the amplification quences from both TnS-Mob and TnS-GDYN1, as well as the in the chromosome, strain CFNX206 was cultured in the Rhizobium DNA sequences corresponding to the amplicon. presence of higher concentrations of kanamycin, up to 1 The experimental approach described was applied to mg/ml. A strain resistant to these concentrations strains harboring amplifications of discrete DNA regions in (CFNX206A) presented an amplification factor of =60, and pa (CFNX202, CFNX208, CFNX210, and CFNX212), in pd the amplified fragments were clearly revealed in the total Downloaded by guest on September 29, 2021 Genetics: Flores et al. Proc. Natl. Acad. Sci. USA 90 (1993) 4935 A B C D E

._w ^. ... _ .. -. 4m. 4p _ .f_ E.'...... -' ._, pI* *,,

..it......

tt--r: ,,,,,, WiS

,', :_ .>i7>usi. .: .,:;-,'.:.

f 1 ,.". . .: igLy;;'S

*.:. . 'il'.'-.'.'''.,,::..:-.:.. .. ' . .. '.'.'.''.:'..' .. *_p

1 2 3 1 2 3 1 2 3 1 2 3 1 2 3

FIG. 3. Identification of amplicon sequences cloned in E. coli. (A-D) Amplification events in pa. Lanes 1, Southern blots of nonamplified strains containing the Tn5-GDYN1 insertions hybridized against pa as probe (A, CFNX201; B, CFNX207; C, CFNX209; D, CFNX211). Lanes 2, Southern blots ofamplified strains hybridized against pa as probe (A, CFNX202;B, CFNX208; C, CFNX210; D, CFNX212). Lanes 3, Southern blots of the same strains as in lanes 2 hybridized against the extrachromosomal DNA isolated from the corresponding E. coli strains: A, CFNC1; B, CFNC4; C, CFNC5; D, CFNC6. (E) Amplification event in the chromosome. Lanes: 1, total digested DNA stained with ethidium bromide from original strain containing a TnS-GDYN1 insertion (CFNX205); 2, total digested DNA stained with ethidium bromide from amplified strain (CFNX206A); 3, same DNA as in lane 2 hybridized against extrachromosomal DNA from E. coli strain CFNC3.

DNA fingerprint. The in vivo cloning procedure was applied amplification event in strain CFNX204 is similar to that to this strain. When the DNA isolated from E. coli was used reported by Romero et al. (5) and that the whole region was as probe, the same bands as those that increased in intensity transferred from Rhizobium to E. coli. in the total DNA fingerprint were revealed (Fig. 3E). This result indicates that the whole amplicon sequence was trans- ferred to E. coli. DISCUSSION In Vivo Cloning of a 120-kb nod nif Amplicon from R. Gene amplification is a common phenomenon in both higher phaseoli inE. coli. We have recently reported (5) that a 120-kb organisms (16) and microorganisms (17). The use of the region of the symbiotic plasmid (pSym) of R. phaseoli is GDYN1 element, either as a cassette (5) or as a transposon subjected to amplification and deletion events at high fre- (this paper), greatly facilitates the analysis of DNA amplifi- quency. This region contains several nodulation (nod) and cation events. nitrogen fixation (nif) genes and has as endpoints directly The genetic strategy presented in this paper provides an repeated sequences that correspond to the two nitrogenase experimental approach for the identification and in vivo operons present in this plasmid (15). cloning of DNA regions that have the potential to be ampli- The physical map of the pSym of strain CFN42 has been fied as a unit (amplicons). In the six cases analyzed, complete recently established (15). This map allowed us to analyze the amplicon sequences were transferred to E. coli. In addition DNA region cloned in E. coli from strain CFNX204. Over- to the biological significance that amplicons may have, the lapping cosmids covering the whole structure of the pSym genetic approach presented here might be a valuable tool for were digested with BamHI, subjected to agarose gel electro- the isolation and analysis of discrete DNA regions and for phoresis, and blotted onto nitrocellulose. The blots were studies on genome organization and function. hybridized with the DNA cloned in E. coli from strain The presence of amplicons seems to be a general charac- CFNX204. In this experiment 16 overlapping cosmids cov- teristic of the R. phaseoli genome. We have identified am- ering the whole structure of pSym (15) were analyzed. The plicons in the chromosome and in several large plasmids. structure of the pSym derived from the work of Girard et al. Amplicons vary in size from a few kilobases, as that present (15) is schematized in Fig. 4A. The region revealed by in the chromosome and revealed in strain CFNX206A, to hybridization of the 16 cosmids spans 120 kb, and the >100 kb, as for the nod nifamplicon of the pSym revealed in endpoints (fragments 1 and 25) correspond to the two nitro- strain CFNX204. Furthermore, the analysis of several TnS- genase operons present in this plasmid. The recombination GDYN1 insertions in pa revealed the existence of different event between the two nitrogenase operons (fragments 1 and amplicons in a single replicon. We have also identified 25 of 9.2 and 5.4 kb, respectively) generates a new BamHI amplicons in different regions of the genome of Rhizobium fragment of 5.2 kb (5). The fragments involved were detected leguminosarum biovar trifolii (unpublished data). (Fig. 4C) by hybridization against a probe that corresponds We propose that amplicons define a structural character- to a 300-bp sequence internal to nif gene H. The wild type istic of the genome. The basic elements are the DNA se- (data not shown) and the strain containing the original quences that provide recombination points for amplification insertion (CFNX203) revealed three fragments correspond- events. The frequency at which the amplification event ing to the two nitrogenase operons (fragments 1 and 25) and occurs might be an intrinsic characteristic of the sequences to a third nifH gene (fragment 10 of 3.8 kb). As expected, the involved in the recombination. The level ofamplification that amplified strain (CFNX204) revealed the recombinant frag- a cell can tolerate should be related to the specific sequences ment (labeled R); this fragment and fragment 10 (which is amplified and to the particular environmental conditions internal to the amplicon) were amplified whereas fragments surrounding the cell. An amplified DNA region consists of a 1 and 25 remained in single copy. These data indicate that the series of tandemly repeated sequences in direct orientation. Downloaded by guest on September 29, 2021 4936 Genetics: Flores et al. Proc. Natl. Acad. Sci. USA 90 (1993) A

CGD102 CGD1 5 CGD45 CGD7 CGD28 CGD35 FIG. 4. Identification of the region of pSym plasmid of R. phaseoli corresponding to a nod 11 III1 II I 1 1, _80 85 1 nif amplicon cloned in E. coli. (A) Schematic 5 10 16 21 25 30 representation of pSym of R. phaseoli strain CFN42, as described by Girard et al. (15). Vertical bars indicate BamHI sites; numbers 75 70 64 61 55 52 45 36 _ indicate BamHI fragments. Fragments 1 and 25 v 11 II TT I III III IIHT I lfn 11181O correspond to 10 CGD47 CG332 CGD84 103 nitrogenase operons; fragment CGD180 corresponds to a third nifH gene. Regions cov- 10Kb CGD179 CGD5I CGD17C'O ered by the different cosmids are indicated; cosmids are labeled as reported previously (15); the bar on top indicates the zone revealed by B C hybridization against the DNA isolated from 2 strain CFNC2. (B) Cosmids covering the whole Kb structure of pSym were digested with BamHI, subjected to electrophoresis, and hybridized 30 21- using as probe the DNA isolated from strain 85- CFNC2. Examples of four cosmids are 10 pre- 1 - 1- -~ sented. Lanes: ethidium bromide 46 1, staining; 2, 83- 22- 36- autoradiography. Numbers indicate BamHI 32- 25- _ 1 L- R- 4 fragments in the circular map; * indicates trun- IN cated fragments; L indicates the fragment cor- 10- _ _ responding to the cloning vector used to con- 80- 23- 81- struct the cosmid library; A indicates fragments 82- 31- containing Tn5-Mob sequences from the pSym used to make the cosmid library. For example, 20- 37- a fragment containing sequences from TnS-Mob and from fragment 3 and that is truncated in a 3* cosmid is labeled A3*. Cosmids are indicated at 84- bottom. (C) Southern blots of the nonamplifled 2- 33- strain containing a TnS-GDYN1 insertion 38- 34- (CFNX203) (lane 1) and the amplified strain 1 (CFNX204) (lane 2), hybridized against a probe 35- corresponding to an internal region ofR. phase- 39- oli nifH gene. Fragments are indicated accord- 40- ing to the map position in A. R, fragment gen- erated by recombination of nitrogenase operons cGD7 cGD15 cGD35 cGD103 (see text). Such sequences may recombine, leading to either an increase 4. Brom, S., Garcfa de los Santos, A., Girard, M. L., Divila, G., or a decrease in the amplification factor. In this context, Palacios, R. & Romero, D. (1991) J. Bacteriol. 173, 1344-1346. amplification of a certain region of the genome should be 5. Romero, D., Brom, S., Martinez-Salazar, J., Girard, M. L., considered as a dynamic state, whose copy number may vary Palacios, R. & Davila, G. (1991) J. Bacteriol. 173, 2435-2441. and which may return to the nonamplified state without 6. Simon, R. (1984) Mol. Gen. Genet. 196, 413-420. 7. Quinto, C., De la Vega, H., Flores, M., Fernandez, L., disrupting the structure of the genome. Ballado, T., Sober6n, G. & Palacios, R. (1982) Nature (London) A dynamic structure of the genome might increase its 299, 724-726. plasticity for adaptation. In some prokaryotes, tandem du- 8. Eckhardt, T. (1978) Plasmid 1, 584-588. plications or amplifications of specific DNA regions have 9. Rigby, P. N. J., Dieckmann, M., Rhodes, C. & Berg, P. (1977) been correlated with the capacity of a cell population to J. Mol. Biol. 113, 237-251. survive and grow under particular conditions (18-21). The 10. Segovia, L., Pifiero, D., Palacios, R. & Martfnez, E. (1991) role of gene amplification might be of particular interest in Appl. Environ. Microbiol. 57, 426-433. Rhizobium. Cells harboring amplifications of specific se- 11. Quinto, C., De la Vega, H., Flores, M., Leemans, J., Cevallos, quences could give a selective advantage to the strain, M. A., Pardo, M. A., Azpiroz, R., Girard, M. L., Calva, E. & Palacios, R. (1985) Proc. Natl. Acad. Sci. USA 82, 1170-1174. allowing it to cope with changing environmental conditions or 12. Brom, S., Martinez, E., Davila, G. & Palacios, R. (1988) Appl. to interact with different plants. Environ. Microbiol. 54, 1280-1283. 13. Flores, M., Gonzalez, V., Brom, S., Martfnez, E., Piniero, D., The technical assistance of Virginia Quinto, Rosa Marfa Ocampo, Romero, D., Davila, G. & Palacios, R. (1987) J. Bacteriol. 169, and Angeles Moreno is greatly appreciated. The R. phaseoli ribo- 5782-5788. somal DNA fragment used as hybridization probe was provided by 14. Figurski, D. H. & Helinski, D. R. (1979) Proc. Natl. Acad. Sci. Lorenzo Segovia. Partial financial support for this research was USA 76, 1648-1652. provided by Grant 936-5542.01-523-8.600 from the U.S. Agency for 15. Girard, M. L., Flores, M., Brom, S., Romero, D., Palacios, R. International Development, and by grants from The Rockefeller & Davila, G. (1991) J. Bacteriol. 173, 2411-2419. Foundation and the Consejo Nacional de Ciencia y Tecnologfa, 16. Schimke, R. T. (1988) J. Biol. Chem. 263, 5989-5992. Mexico. 17. Petes, T. D. & Hill, C. W. (1988) Annu. Rev. Genet. 22, 147-168. 1. Martinez, E., Romero, D. & Palacios, R. (1990) Crit. Rev. Plant 18. Goldberg, I. & Mekalanos, J. J. (1986) J. Bacteriol. 165, Sci. 9, 59-93. 723-731. 2. Hahn, M. & Hennecke, H. (1987) Appl. Environ. Microbiol. 53, 19. Anderson, P. & Roth, J. R. (1977) Annu. Rev. Microbiol. 31, 2247-2252. 473-504. 3. Flores, M., Gonzalez, V., Pardo, M. A., Leija, A., Martinez, 20. Rownd, R. H. (1982) in GeneAmplification, ed. Schimke, R. T. E., Romero, D., Pifiero, D., Davila, G. & Palacios, R. (1988) J. (Cold Spring Harbor Lab., Plainview, NY), pp. 273-279. Bacteriol. 170, 1191-1196. 21. Sonti, R. V. & Roth, J. R. (1989) Genetics 123, 19-28. Downloaded by guest on September 29, 2021