Brink, B. A., J. Miller, R. W. Carlson, and K D. Noel. 1990. Expression of Rhizobium leguminosarwm CFN42 genes for lipopolysaccharide in strains derived from ...
JOURNAL OF BACTERIOLOGY, Aug. 1992, p. 5183-5189
Vol. 174, No. 16
0021-9193/92/165183-07$02.00/0 Copyright © 1992, American Society for Microbiology
Different Plasmids of Rhizobium leguminosarum bv. phaseoli Are Required for Optimal Symbiotic Performance SUSANA BROM,* ALEJANDRO GARCIA DE LOS SANTOS, TOMASZ STEPKOWSKY, MARGARITA FLORES, GUILLERMO DAVILA, DAVID ROMERO, AND RAFAEL PALACIOS Departamento de Genetica Molecular, Centro de Investigacion sobre Fijacion de Nitr6geno, Universidad Nacional Aut6noma de Mexico, Apartado Postal 565-A,
Cuemavaca, Morelos, Mexico Received 18 March 1992/Accepted 11 June 1992
Rhizobium leguminosarum bv. phaseoli CFN42 contains six plasmids (pa to pf), and pd has been shown to be the symbiotic plasmid. To determine the participation of the other plasmids in cellular functions, we used a positive selection scheme to isolate derivatives cured of each plasmid. These were obtained for all except one (pe), of which only deleted derivatives were recovered. In regard to symbiosis, we found that in addition to pd, pb is also indispensable for nodulation, partly owing to the presence of genes involved in lipopolysaccharide synthesis. The positive contribution of pb, pc, pe, and pf to the symbiotic capacity of the strain was revealed in competition experiments. The strains that were cured (or deleted for pe) were significantly less competitive than the wild type. Analysis of the growth capacity of the cured strains showed the participation of the plasmids in free-living conditions: the pf strain was unable to grow on minimal medium, while strains cured of any other plasmid had significantly reduced growth capacity in this medium. Even on rich medium, strains lacking pb or pc or deleted for pe had a diminished growth rate compared with the wild type. Complementation of the cured strains with the corresponding wild-type plasmid restored their original phenotypes, thus confirming that the effects seen were due only to loss of plasmids. The results indicate global participation of the Rhizobium genome in symbiotic and firee-living functions.
Bacteria of the genus Rhizobium have the capacity to interact with the roots of leguminous plants, forming nitrogen-fixing nodules. In many Rhizobium species, symbiotic (pSym) plasmids have been defined as the plasmids that carry the essential genes for nodule formation and nitrogen fixation on the appropriate host plants (18, 26, 34). Genetic information present on the chromosome (15, 31, 32, 35) and other extrachromosomal elements (20-22, 24) has also been shown to participate in the symbiotic process, contributing to early or late nodule development, nitrogen fixation, and competitiveness (8, 12, 19, 20, 22, 27, 41, 43, 44). Rhizobium leguminosarum bv. phaseoli CFN42 contains six plasmids (pa to pf) ranging in size from 150 to 600 kb, and pd has been identified as the symbiotic plasmid, as it carries nod and nif genes (36, 37) and confers on anAgrobacterium tumefaciens strain the capacity to nodulate beans (5). The complete physical map of this plasmid has been determined (16). The structural complexity of the R. leguminosarum bv. phaseoli genome has been evidenced by its high content of reiterated sequences (13, 16) and the occurrence of frequent genomic rearrangements (4, 14, 40). Concerning pSym unlinked genes, Noel et al. (30, 33) have shown that in R leguminosarum bv. phaseoli CFN42, sequences involved in lipopolysaccharide (LPS) production are required for formation of completely differentiated nodules. As an approach to study the functional relationship of the different genomic elements, we decided to analyze the participation of each one of the plasmids of CFN42 in the symbiotic and free-living cellular functions. We selected derivatives cured of each plasmid through a strategy that *
Corresponding author. 5183
allows positive selection for insertion as well as for loss of function. A similar approach was used by Hynes and McGregor in a comprehensive study of the plasmids of R. leguminosarum bv. viciae VF39 (20). The strategy employed allowed us to obtain derivatives cured of each of five plasmids and deleted for one of them. The data show that in addition to the physical complexity of the R. leguminosarum bv. phaseoli genome, a functional complexity may also exist, as all of the plasmids affect symbiotic and/or growth functions.
MATERIALS AND METHODS Bacterial strains and growth conditions. The different strains used are listed in Table 1. R. leguminosarum bv. phaseoli strains were grown at 30°C in PY (32) or minimal M9 (1) medium with 10 mM ammonium chloride and succinate as nitrogen and carbon sources, respectively. Eschenchia coli and Agrobacterium strains were grown on LB medium (29) at 37 and 30°C, respectively. For growth rate measurements, overnight cultures of the strains were grown on PY medium and the cells were centrifuged, suspended in sterile water (twice), and diluted in 250-ml Erlenmeyer flasks with 100 ml of either minimal or PY medium. The flasks were incubated at 30°C with shaking (100 rpm). Growth was monitored by measurement of turbidity in a Klett-Summerson photoelectric colorimeter, with a 500- to 570-nm range filter. For the pb- strain, growth was determined by serially diluting the cultures in 10 mM MgSO4-0.01% (vol/vol) Tween 40, plating them on PY, and counting CFU. Media containing sucrose at a final concentration of 12.5% (wt/vol) were prepared as previously described (40). Antibiotics were added in the following concentrations: nalidixic acid, 20 (to select against E. coli) or 100 (to select against A. tume-
5184
BROM ET AL.
J. BACTERIOL.
TABLE 1. Bacterial strains Strain
R. leguminosarum bv. phaseoli CFN42 CFNX182 CFNX183 CFNX184 CFNX89 CFNX185 CFNX186 CFNX187 CFNX188 CFNX189 CFNX190 CFNX191 CFNX192 CFNX193 CFNX194 CFNX180 CFNX181 CFN2001
Agrobacteiium tumefaciens GM19023/pa::TnS-mob GM19023/pb::TnS-mob
GM19023/pc::TnS-mob GM19023/pd::TnS-mob GM19023/pf::TnS-mob Escherichia coli S17/pDR21 HB101/pDel27 HB101/pCosl26 HB101/pRK2013
Relevant characteristic(s)
Source or reference
Wild type CFN42 cured of pa CFN42 cured of pb CFN42 cured of pc CFN42 cured of pd CFN42 deleted for pe CFN42 cured of pf CFNX182 complemented with pa::TnS-mob CFNX183 complemented with pDel27 CFNX183 complemented with pCosl26 CFNX183 complemented with pb::TnS-mob CFNX184 complemented with pc::TnS-mob CFNX89 complemented with pd::TnS-mob CFNX185 complemented with pe::TnS-mob CFNX186 complemented with pf::TnS-mob Chromosomal TnS-mob insertion (referred as CFN42-18 in reference 27) Source of pe::TnS-mob of CFN42 CFN42 cured of pa and pd
This work 23
Source of pa::TnS-mob of CFN42 Source of pb::TnS-mob of CFN42 Source of pc::TnS-mob of CFN42 Source of pd::Tn5-mob of CFN42 Source of pf::TnS-mob of CFN42
13 13 13 13 13
Source of TnS-GDYN1 Source of pDel27 Source of pCosl26
39 3 6 11
Conjugation helper
faciens) jg ml-'; rifampin, 100 jig ml-1; neomycin, 60 ,ug ml-'; spectinomycin, 75 ,ug ml-'; tetracycline, 10 ,g ml-';
streptomycin, 100 ,g ml-'. Plasmid profiles and hybridization analysis. Plasmid patterns were visualized by the procedure of Eckhardt (9), blotted onto nitrocellulose, and hybridized as described previously (13). Genomic DNA was digested with EcoRI, subjected to electrophoresis in 1% agarose gels, blotted onto nitrocellulose, and hybridized as described above. The plasmids used as probes were purified from Agrobactenium transconjugants harboring each one of the plasmids (13) by the method of Hirsch et al. (17) and labelled by nick translation (38). Genetic manipulations. Transposon TnS-GDYN1 contains the GDYN1 cassette, which carries kanamycin-gentamicin and spectinomycin-streptomycin resistance markers and sacR sacB genes (40) flanked by" the IS50 insertion sequences of TnS (39). CFN42 derivatives carrying TnSGDYN1 insertions in each plasmid were obtained by mating E. coli S-17 containing the transposon with CFN42 and selecting Nalr 5pr transconjugants. TnS-GDYN1 was localized by hybridization of'plasmid profiles from the transconjugants by using the transposon as a probe. Plasmid curing was carried out by plating overnight cultures of the transposon-labelled strains on PY plates containing 12.5% sucrose. Sucrose-resistant colonies were selected and verified as SpS. Plasmid profiles of such colonies were analyzed by ethidium bromide staining and hybridization to the purified plasmids of CFN42.
36 This work This work This work This work This work This work This work This work This work This work This work This work This work This work This work
The derivatives cured of pa, pb, pc, pd, and pf were complemented with the corresponding TnS-mob-labelled plasmid by conjugation with Agrobacterium strains that contain these plasmids (13) by using pRK2013 (11) as a helper. Transconjugants were selected on PY with nalidixic acid and neomycin. Strain CFN2001 is a derivative of CFN42 which lacks pd and pa (23). The pe of this strain was labelled with Tn5-mob, and this strain was used as a donor to introduce pe into a derivative deleted for this plasmid, with pRK2013 as a helper and selection for Strr Nmr Nalr transconjugants. Complementation of CFN42 derivatives lacking pb with pCosl26 and pDel27 (3, 6) was done by mating E. coli HB101 containing each clone with the Rhizobium strain and pRK2013 as a helper and selecting for Nalr Tcr transconjugants. Nodulation and nitrogen fixation assay. Overnight cultures were used to inoculate surface-sterilized Phaseolus vulgaris cv. Negro jamapa seeds. Plants were grown in 250-ml Erlenmeyer flasks with Fahraeus agar medium (10), without added nitrogen, at 28°C. At day 15, nodulation was scored and acetylene reduction was assayed as previously described (25). Surface-sterilized nodules were crushed on PY plates, and the plasmid pattern of single colonies was checked by ethidium bromide staining and hybridization to the corresponding plasmids. Competition assays. Seedlings were inoculated with mixtures of parental and derived strains in an approximately equal ratio. One of the strains in the mixture always carried
R. PHA4SEOLI PLASMIDS REQUIRED FOR SYMBIOSIS
VOL. 174, 1992
FIG. 1. CFN42 derivatives with Tn5-GDYN1 insertions in each plasmid. Odd-numbered lanes are ethidium bromide-stained plasmid profiles. Even-numbered lanes are Southern blots of each plasmid profile probed with the transposon. Lanes: 1 and 2, CFN42; 3 to 14, CFN42 derivatives with TnS-GDYN1 insertions in pa (lanes 3 and 4), pb (lanes 5 and 6), pc (lanes 7 and 8), pd (lanes 9 and 10), pe (lanes 11 and 12), and pf (lanes 13 and 14). The letters on the left mark the positions of the different plasmids.
TnS-mob, either on the chromosome or on one of the plasmids, while the other did not. Bacteria for inocula were grown overnight on PY, centrifuged, and diluted in sterile water. Inoculum concentrations were determined by A600. Cell numbers were adjusted in suspensions to formulate the mixtures of inoculum strains. The final cell numbers were verified by serially diluting the inocula in 10 mM MgSO40.01% (vol/vol) Tween 40, plating them on appropriate media for identification of each strain, and counting CFU. To determine which strain was responsible for nodule formation, surface-sterilized nodules were crushed on PY plates and single colonies were picked and tested for growth on selective media. The plasmid pattern of some colonies was also checked. LPS production. Washed cells of R. leguminosarum bv. phaseoli strains were examined for LPS production by sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE) by the method of Cava et al. (6). Gels were fixed and stained by using a Bio-Rad silver staining kit as previously described (6). RESULTS Isolation of cured derivatives and complementation with wild-type plasmids. The six different plasmids of R. leguminosarum bv. phaseoli CFN42 were labelled with transposon TnS-GDYN1 as explained in Materials and Methods. The 5 6 7 8 9 1011 12
i5000
5185
plasmid pattern and hybridization to the transposon of derivatives with insertions in each plasmid are shown in Fig. 1. These derivatives were plated on PY containing sucrose to select for loss of plasmids. Some sucrose-resistant colonies retained the antibiotic resistance markers, indicating that they had acquired sucrose resistance by a mechanism that did not involve loss of the Tn5-GDYN1 element. Sucroseresistant colonies that had also lost the antibiotic resistance markers were screened for loss of plasmids in Eckhardt-type gels stained with ethidium bromide and hybridized to the purified plasmids. The results show that derivatives lacking plasmids pa, pb, pc, pd, and pf were obtained, as they did not present the plasmid band, either by direct visualization (Fig. 2, lanes 1, 5, 13, 17, and 25) or by hybridization to the corresponding plasmid (Fig. 2, lanes 2, 6, 14, 18, and 26). Also, some derivatives with deletions of pc or pd were obtained (data not shown). In the case of pe, only strains with deletions were recovered (peA) (Fig. 2, lanes 21 and 22). We were unable to cure pe, although we screened sucroseresistant derivatives obtained from six different strains with TnS-GDYN1 insertions in pe and varied the growth medium (PY, MY, or PY supplemented with Casamino Acids) or temperature (30 or 37°C). We also selected sucrose-resistant derivatives of a strain that had the transposon inserted in peA&. All of the derivatives obtained still contained this plasmid. To ensure that the phenotypic traits of the derivatives (see below) were due only to loss of plasmids, each strain was complemented with the corresponding TnS-mob-labelled plasmid as described in Materials and Methods. The ethidium bromide-stained plasmid profile (Fig. 2, lanes 3, 11, 15, 19, 23, and 27) and the hybridization against the purified plasmids (Fig. 2, lanes 4, 12, 16, 20, 24, and 28) of these constructions are shown. It should be noted that the strain that had peA lost it when the complete plasmid was introduced. The pb-cured strain was also complemented with clones pCosl26 and pDel27 (3, 6), which carry genes required for LPS production. Their plasmid pattern is shown in Fig. 2, lanes 7 and 9; both clones hybridized with pb (Fig. 2, lanes 8 and 10). These clones complement phenotypic properties of the pb-cured strain (see below). Saprophytic characteristics of cured strains. In Table 2, phenotypic traits of the plasmid-cured strains are presented. It can be seen that the pb- strain has an opaque appearance
13141516
17181920
lt : '?C LU38 ' ;: ' '
FIG. 2. Plasmid-cured, deleted, and complemented derivatives of CFN42. Odd-numbered lanes are ethidium bromide-stained plasmid profiles. Even-numbered lanes are Southern blots of each plasmid profile probed with the following CFN42 plasmids: pa, lanes 2 and 4; pb, lanes 6, 8, 10, and 12; pc, lanes 14 and 16; pd, lanes 18 and 20; pe, lanes 22 and 24; pf, lanes 26 and 28. Strains: CFNX182 (lanes 1 and 2), CFNX187 (lanes 3 and 4), CFNX183 (lanes 5 and 6), CFNX188 (lanes 7 and 8), CFNX189 (lanes 9 and 10), CFNX190 (lanes 11 and 12), CFNX184 (lanes 13 and 14), CFNX191 (lanes 15 and 16), CFNX89 (lanes 17 and 18), CFNX192 (lanes 19 and 20), CFNX185 (lanes 21 and 22), CFNX193 (lanes 23 and 24), CFNX186 (lanes 25 and 26), and CFNX194 (lanes 27 and 28). The asterisk on the left of each panel indicates the cured or deleted plasmid.
BROM ET AL.
5186
J. BACTERIOL.
TABLE 2. Characteristics of plasmid-cured derivatives of CFN42
Relevant genotype
Strain
CFN42 CFNX182 CFNX183 CFNX184 CFNX89 CFNX185 CFNX186 CFNX188 CFNX189 CFNX190 CFNX192
Wild type paPbpCpd-
peA pf-
Colony morphology
Flocculationa
Motilityb
Melanin'
Nodulation
Acetylene reduction
Glossy Glossy Opaque Glossy Glossy Glossy Glossy Glossy Glossy Glossy Glossy
+
+ + + + + + + + +
+ + + + + +
+ +
+ +
+ + + + + + +
+
pb-/pDel27 pb-/PCosl26 pb-/pb+ pd-/pd+ a Checked in liquid minimal or PY medium. b Checked in PY with 0.3% agar. c Checked in PY with tyrosine (50 p,g/ml) and CuS04 (20 ,g/ml).
-
on solid medium, flocculates in liquid medium, and lacks motility. These characteristics have been associated with an LPS-impaired phenotype (6, 20, 35). The pb or either of the LPS clones restored the phenotype of the pb- strain (Table 2). All of the other strains were similar to the wild type, except for the pd- strain, which showed no melanin production. The presence of genes required for melanin synthesis on this plasmid has been previously reported (2). The growth capacities of the different strains in minimal and rich media are presented in Fig. 3. The absence of plasmids pa, pb, pc, pd, and pf or deletion of pe significantly diminished the growth of the strains on minimal medium, and the most dramatic effect was presented by the pf- strain, which was unable to grow in this medium. The addition of a mixture of amino acids, vitamins, or Casamino Acids did not fulfil the growth requirement of the pf- strain (data not shown). Even in rich media, loss of pb or pc or deletion of pe significantly reduced the growth rates of the strains. The growth rates of some plasmid-complemented strains (pa-/
100r
A
+
+ + + + + +
pa+, pb-/Del27, pb-/pCosl26, pb-/pb+, and pf-/pf+) in minimal medium were measured. In all cases, they were indistinguishable from that of the wild-type strain (data not shown). Symbiotic properties of plasmid-cured strains: nodulation and competitiveness. In regard to symbiosis, the absence of nodulation in the pd- strain was confirmed. Surprisingly, strains lacking pb were not capable of inducing nodule formation (Table 2); inoculation of P. vulgaris roots with the pb- strain resulted only in the formation of a few (zero to three) bumps per plant. The nodulation capacity was restored when the plasmidless strain was complemented either with wild-type pb, pCosl26, or pDel27 (Table 2 and Fig. 4). A direct effect on nodulation or nitrogen fixation of derivatives cured of the other plasmids was not observed (Table 2), but when we tested their competitiveness in comparison with that of the wild-type strain (Table 3), it was significantly (99% level) diminished for pc-, pf-, peA, and pb- strains complemented with pCosl26 or pDel27. On the
B
7 i, 50 -C0 .
A
_
1*
*
*p
1,!*
*
1 2 3 4 5 6 7 2 3 4 5 6 7 FIG. 3. Growth of plasmid-cured or deleted derivatives of CFN42 on PY (A) and minimal (B) media. Growth is expressed as the specific growth rate compared with that of the wild-type strain. Strains: CFN42 (lanes 1), CFNX182 (lanes 2), CFNX183 (lanes 3), CFNX184 (lanes 4), CFNX89 (lanes 5), CFNX185 (lanes 6), and CFNX186 (lanes 7). An asterisk indicates a significant difference from the wild-type strain at a 90% confidence level for PY medium and at a 95% confidence level for minimal medium, determined by a Neuman Keuls test. The specific growth rate is defined as the ratio of the increase in turbidity or cell number to time. Each growth measurement was repeated at least three times. 1
D C FIG. 4. P. vulgaris roots inoculated with R leguminosarum bv. phaseoli CFNX183 (panel A), CFNX188 (panel B), CFNX189 (panel C), and CFNX190 (panel D) at 15 days after inoculation.
R. PH4SEOLI PLASMIDS REQUIRED FOR SYMBIOSIS
VOL. 174, 1992
TABLE 3. Competition of CFN42 and its plasmid-cured and complemented derivatives for nodule occupancy in P. No. of nodules observed
Strains in inoculum
5187
vulgarisa
(expected) showing
occupancy by:
Parent"
Derivative
Ratio (parent/ derivative)
CFNX180c CFN42 CFN42 CFN42 CFN42 CFNX180 CFN42 CFNX180 CFN42 CFNX180 CFN42
CFNX182 (pa-) CFNX187 (pa-/pa+) CFNX188 (pb-/pDel27) CFNX189 (pb-/pCosl26) CFNX190 (pb-/pb+) CFNX184 (pc-) CFNX191 (pc-/pc-) CFNX185 (peA) CFNX193 (peA&/pe') CFNX186 (pf-) CFNX194 (pf-/pf+)
1:1 1:0.8 1:0.76 1:0.76 0.8:1 0.74:1 1:1 0.85:1 0.8:1 1:0.8 1:1
Parent
5 (9) 9 (8) 26 (15) 23 (14) 8 (9) 22 (11) 11 (11) 34 (17) 16 (14) 25 (14) 8 (9)
Derivative
Both
13 (9) 6 (7) 0 (11) 1 (10) 13 (12) 3 (14) 11 (11) 3 (20) 16 (18) 1 (12) 10 (9)
10 9 2" 6d 9 od
8 id
6
3d 7
Approximately 30 nodules from three different plants were analyzed. Wild type. c TnS-mob-labelled CFN42 retains its original competitiveness (27). d The derivative was significantly different from the parental strain at a 99% confidence level in a x2 test.
a
b
other hand, the pa- strain seemed to be more competitive than strain CFN42, although this difference was significant only at the 90% level. In all cases, the complementation of each strain with its corresponding plasmid restored its wildtype competitive capacity (Table 3). Localization of LPS production genes. The genes from CFN42 required for LPS production, reported by Cava et al. (6), which are cloned in pCosl26 and subcloned (7.5-kb EcoRI fragment) on pDel27, complement the morphologic characteristics and the nodulation phenotype of the pbstrain. The plasmid localization of the LPS genes was determined by Southern hybridization of restriction enzymedigested DNA of pb- and wild-type cells by using pCosl26 as a probe. The results (Fig. 5) showed no hybridization signal in the pb- strain, while the wild-type or complemented strains presented several bands. It has already been mentioned that when pb was used as a probe on Eckhardttype gels, it hybridized with both cosmids (Fig. 2, lanes 8 and 10). Furthermore, LPS production was determined in the
1
2
3
different strains by using silver-stained SDS-PAGE gels as described in Materials and Methods. Figure 6 shows that the pb- strain does not contain an LPS I band, which corresponds to the complete LPS molecule, having only the LPS II band that corresponds to the LPS core (6), while the wild-type and complemented strains present both bands. The increase of LPS II in extracts of strains that lack LPS I has already been shown (3, 6, 35). DISCUSSION In this work, the contribution to symbiotic and free-living functions of all of the different plasmids present in an R. leguminosarum bv. phaseoli strain was determined. Hynes and McGregor (20) have reported the systematic curing of all plasmids present in a strain of R. leguminosarum bv. viciae. The analysis of the cured derivatives allowed them to show
4
LPS I-~ LPS I1-
FIG. 5. Localization of LPS-coding sequences on pb of CFN42. Autoradiograph of EcoRI digested DNA probed with pCos126. Strains: CFN42 (lane 1), CFNX183 (lane 2), CFNX188 (lane 3), and CFNX189 (lane 4).
II
FIG. 6. LPS profiles on SDS-PAGE gels. Strains: CFN42 (lane 1), CFNX183 (lane 2), CFNX188 (lane 3), CFNX189 (lane 4), and CFNX190 (lane 5).
5188
J. BAcTERiOL.
BROM ET AL.
that the genes necessary for development of efficient nodules are present on two plasmids in addition to pSym. One of these plasmids carries genes involved in LPS production, and the other seems to be important for free-living growth, as well as for nitrogen fixation. Here we report for the first time that a plasmid (pb) distinct from the one that carries nod and nifgenes is indispensable for nodule formation on P. vulgaris roots by R leguminosarum bv. phaseoli. We show that previously reported sequences involved in LPS production (6, 30) are localized on this plasmid. Sequences that are functionally similar to these LPS genes are present in the genome of R. legwminosarum bv. trifolii ANU843 and in a plasmid of R. leguminosarum bv. viciae VF39. This was shown by complementation of LPS mutants of both strains for LPS production and the symbiotic phenotype with R. leguminosarum bv. phaseoli LPS clones (3, 20). Previous reports indicate that R. leguminosarum bv. phaseoli LPS mutants give rise to abortive infection threads, resulting in empty nodules (6, 33). Our results show that nodule formation is completely impaired when pb is lost. This suggests that pb carries other sequences involved in nodule formation, in addition to the LPS genes described by Noel et al. (33). This is further supported by the fact that the symbiotic properties of pb-cured strains were not completely restored by the LPS clones, since the competitive capacity of the complemented derivatives was significantly diminished compared with that of the wild-type strain. We have previously reported that A. tumefaciens GM19023 carrying pd of R leguminosarum bv. phaseoli CFN42 is capable of eliciting the formation of differentiated nodules on P. vulgaris roots (5). Taking into account the results presented here concerning the requirement of pb for nodulation by CFN42, it might be inferred that Agrobacterium sp. contains chromosomal genes that are functionally equivalent to those present in pb of CFN42. A. tumefaciens containing the Ti plasmid induces formation of tumors on dicotyledonous plants (42). LPS purified from Agrobacterium sp. has been shown to inhibit tumor induction by virulent bacteria (45). Additionally, chromosomal mutations in Agrobacterium sp. have been obtained which cause the cells to be avirulent, defective in attachment to plant cells, and altered in cell surface properties (7, 28), thus suggesting a relationship between LPS and virulence. It would be interesting to know whether the same LPS sequences are needed for tumor induction and for nodulation by Agrobacterium sp. carrying either the Ti or a symbiotic plasmid. An approach used in this work for analysis of symbiotic properties was determination of the competitive capacities of the different strains. This approach revealed features which could not be detected solely by analysis of the individual nodulation capacity of each cured strain. Previous studies have correlated the competitive behaviors of different Rhizobium and Bradyrhizobium strains with specific characteristics, such as hydrogen oxidation, bacteriocin production, speed of infection, dicarboxylic acid transport, nitrogen metabolism regulators (nifA and nodD), and growth capacity (19, 43). Our results suggest that a complex interaction of genomic sequences affects the symbiotic properties of a strain, viewed as its competitive capacity. The diminished competitiveness of the plasmid-cured derivatives of CFN42 cannot be ascribed solely to lesser growth capacity. Our results show that equivalent decreases in competitiveness can occur in strains that have dissimilar growth decreases (pc- and pf- strains) and that restoration of wild-type growth capacity in a pb- strain complemented
or pCosl26 does not result in wild4type competitiveness. Plasmid preservation is also important for free-living cells, as demonstrated by the altered growth capacities of plasmidcured derivatives in different media. Furthermore, essential genes present in pe may account for the fact that derivatives completely lacking this plasmid could not be obtained, notwithstanding all of the strategies employed to find such strains. Genes present on pf also have a fundamental role for the cell, as seen by the inability to grow on minimal medium; this inability seems to be due to a complex metabolic function which could not be reversed by addition of amino acids and vitamins. This somewhat resembles the situation found with a strain cured of pe of R leguminosarum VF39, which is also unable to grow on minimal medium, but in this case addition of Casamino Acids partially restored growth (20). This report shows that when the appropriate traits are surveyed, the participation in symbiosis and/or growth functions of all of the plasmids from a strain may be evidenced. The experiments presented in this work were done under laboratory conditions. The survival and symbiotic capacities of plasmid-cured strains in the field should be studied to gain insight into the functional importance of the different plas-
with pDel27
mids under these conditions.
ACKNOWLEDGMENTS We are grateful to Osvaldo L6pez for technical assistance and to Dale Noel for providing LPS probes. Partial financial support for this research was provided by grant 936-5542.01-523-8.600 from the U.S. Agency for International De-
velopment.
REFERENCES
Beringer, J. E. 1974. R factor transfer in Rhizobium leguminosarum. J. Gen. Microbiol. 84:188-198. 2. Borthakur, D., J. W. Lamb, and A. W. B. Johnston. 1987. 1.
Identification of two classes of Rhizobium phaseoli genes required for melanin synthesis, one of which is required for nitrogen fixation and activates the transcription of the other. Mol. Gen. Genet. 207:155-160. 3. Brink, B. A., J. Miller, R. W. Carlson, and K D. Noel. 1990. Expression of Rhizobium leguminosarwm CFN42 genes for lipopolysaccharide in strains derived from different R leguminosarum soil isolates. J. Bacteriol. 172:548-555. 4. Brom, S., A. Garcia de los Santos, M. L Girard, G. Divila, R. Palacios, and D. Romero. 1991. High-frequency rearrangements in Rhizobium leguminosarum bv. phaseoli plasmids. J. Bacteriol. 173:1344-1346. 5. Brom, S., E. Martinez, G. Divila, and R. Palaclos. 1988. Narrow- and broad-host-range symbiotic plasmids of Rhizobium spp. strains that nodulate Phaseolus vulgaris. Appl. Environ. Microbiol. 54:1280-1283. 6. Cava, J. R., P. M. Elfas, D. A. Turowa, and K. D. Noel. 1989. Rhizobium legumnosansm CFN42 genetic region encoding licomplete nodule depopolysaccharide structures essential for171:8-15. velopment on bean plants. J. Bacteriol. Nester. 1982. Agrobac7. Douglas, C. J., W. Halperin, and E. W. tenum tumefaciens mutants affected in attachment to plant cells. J. Bacteriol. 152:1265-1275. 8. Dylan, T., L. Idpi, S. Stafied, L. Kayap, C. Douglas, M. , and G. Ditta. 1986. RhizoYanofsky, E. Nester, D. R. Henodule bium meliloti genes required for development are related to chromosomal virulence genes inAgrobacteriun tumefaciens. Proc. Natl. Acad. Sci. USA 83:4403-4407. 9. Eckhardt, T. 1978. A rapid method for the identification of plasmid deoxyribonucleic acid in bacteria. Plasmid 1:584-588. 10. Fahraeus, G. 1957. The infection of clover root hair by nodule bacteria studied by a single glass slide technique. J. Gen.
VOL. 174, 1992
R. PHASEOLI PLASMIDS REQUIRED FOR SYMBIOSIS
Microbiol. 16:374-381. 11. 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. 12. Finan, T. M., I. Oresnik, and A. Bottacin. 1988. Mutants of Rhizobium meliloti defective in succinate metabolism. J. Bacteriol. 170:3396-3403. 13. Flores, M., V. Gonza4lez, S. Brom, E. Martinez, D. Pifiero, D. Romero, G. Davila, and R. Palacios. 1987. Reiterated DNA sequences in Rhizobium and Agrobacterium spp. J. Bacteriol. 169:5782-5788. 14. Flores, M., V. Gonzalez, M. A. Pardo, A. Leja, E. Martinez, D. Romero, D. Pifiero, G. Davila, and R. Palacios. 1988. Genomic instability in Rhizobium phaseoli. J. Bacteriol. 170:1191-1196. 15. Forrai, T., E. Vincze, A. Banfalvi, G. B. Kiss, G. S. Randhawa, and A. Kondorosi. 1983. Localization of symbiotic mutations in Rhizobium meliloti. J. Bacteriol. 153:635-643. 16. Girard, M. L., M. Flores, S. Brom, D. Romero, R. Palacios, and G. D6vila. 1991. Structural complexity of the symbiotic plasmid of Rhizobium leguminosarum bv. phaseoli. J. Bacteriol. 173: 2411-2419. 17. Hirsch, P. R., M. Van Montagu, A. W. B. Johnston, N. J. Brewin, and J. Schell. 1980. Physical identification of bacteriocinogenic, nodulation and other plasmids in strains of R. leguminosarum. J. Gen. Microbiol. 120:403-412. 18. Hombrecher, G., N. J. Brewin, and A. W. B. Johnston. 1981. Linkage of genes for nitrogenase and nodulation ability on plasmids in Rhizobium leguminosarum and Rhizobium phaseoli. Mol. Gen. Genet. 182:133-136. 19. Hynes, M. F. 1990. The role of plasmids in competition between strains of Rhizobium leguminosarum, p. 262. In P. M. Gresshoff, L. E. Roth, G. Stacey, and W. E. Newton (ed.), Nitrogen fixation: achievements and objectives. Chapman & Hall, Ltd., London. 20. Hynes, M. F., and N. F. McGregor. 1990. Two plasmids other than the nodulation plasmid are necessary for formation of nitrogen fixing nodules by Rhizobium leguminosarum. Mol. Microbiol. 4:567-574. 21. Hynes, M. F., J. Quandt, M. P. O'Connell, and A. Puihler. 1989. Direct selection for the curing and deletion of Rhizobium plasmids using transposons carrying the Bacillus subtilis sacB gene. Gene 78:111-120. 22. Hynes, M. F., R. Simon, P. Muller, K. Niehaus, M. Labes, and A. Pihler. 1986. The two megaplasmids of Rhizobium meliloti are involved in the effective nodulation of alfalfa. Mol. Gen. Genet. 202:356-362. 23. Leemans, J., G. Sober6n, M. A. Cevallos, L. Fernandez, M. A. Pardo, H. de la Vega, M. Flores, C. Quinto, and R. Palacios. 1984. General organization of R. phaseoli nif plasmids, p. 710. In C. Veeger and V. E. Newton (ed.), Advances in nitrogen fixation research. Nijhoff, Junk and Pudoc, The Hague, The Netherlands. 24. Long, S., S. McCune, and G. C. Walker. 1988. Symbiotic loci of Rhizobium meliloti identified by random TnphoA mutagenesis. J. Bacteriol. 170:4257-4265. 25. Martinez, E., M. A. Pardo, R. Palacios, and M. A. Cevallos. 1985. Reiteration of nitrogen fixation gene sequences and specificity of Rhizobium in nodulation and nitrogen fixation in Phaseolus vulgaris. J. Gen. Microbiol. 131:1779-1786. 26. Martinez, E., D. Romero, and R. Palacios. 1990. The Rhizobium genome. Crit. Rev. Plant Sci. 9:59-93. 27. Martfnez-Romero, E., and M. Rosenblueth. 1990. Increased bean (Phaseolus vulgaris L.) nodulation competitiveness of genetically modified Rhizobium strains. Appl. Environ. Micro-
5189
biol. 56:2384-2388. 28. Metts, J., J. West, S. H. Doanes, and A. G. Matthysse. 1991. Characterization of three Agrobacterium tumefaciens avirulent mutants with chromosomal mutations that affect induction of vir genes. J. Bacteriol. 173:1080-1087. 29. Miller, J. H. 1972. Experiments in molecular genetics. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. 30. Noel, K. D. 1992. Rhizobial polysaccharides required in symbioses with legumes, p. 342-357. In D. P. S. Verma (ed.), Molecular signals in plant-microbe communications. CRC Press, Inc., Boca Raton, Fla. 31. Noel, K. D., R. J. Diebold, J. R. Cava, and B. A. Brink. 1988. Rhizobial purine and pyrimidine auxotrophs: nutrient supplementation, genetic analysis and the symbiotic requirement for de novo purine biosynthesis. Arch. Microbiol. 149:499-506. 32. Noel, K. D., A. Sinchez, L. Fernandez, J. Leemans, and M. A. Cevallos. 1984. Rhizobium phaseoli symbiotic mutants with transposon TnS insertions. J. Bacteriol. 158:148-155. 33. Noel, K. D., K. A. Vandenbosch, and B. Kulpaca. 1986. Mutations in Rhizobiumphaseoli that lead to arrested development of infection threads. J. Bacteriol. 168:1392-1401. 34. Nuti, M. P., A. A. Lepidi, R. K. Prakash, R. A. Schilperoort, and F. A. Cannon. 1979. Evidence for nitrogen fixation (nif) genes on indigenous Rhizobium plasmids. Nature (London) 282:533-535. 35. Priefer, U. B. 1989. Genes involved in lipopolysaccharide production and symbiosis are clustered on the chromosome of Rhizobium leguminosarum bv. viciae VF39. J. Bacteriol. 171: 6161-6168. 36. Quinto, C., H. de la Vega, M. Flores, L. Fernindez, T. Ballado, G. Soberon, and R. Palacios. 1982. Reiteration of nitrogen fixation gene sequences in Rhizobium phaseoli. Nature (London) 299:724-728. 37. Quinto, C., H. de la Vega, M. Flores, J. Leemans, M. A. Cevallos, M. A. Pardo, R. Azpiroz, M. L. Girard, E. Calva, and R. Palacios. 1985. Nitrogenase reductase: a functional multigene family in Rhizobium phaseoli. Proc. Natl. Acad. Sci. USA 82:1170-1174. 38. Rigby, P. W. J., M. Dieckman, C. Rhodes, and P. Berg. 1976. Labelling deoxyribonucleic acid to a high specific activity in vitro by nick translation with DNA polymerase I. J. Mol. Biol. 113:237-251. 39. Romero, D., et al. Unpublished data. 40. Romero, D., S. Brom, J. Martinez-Salazar, M. L. Girard, R. Palacios, and G. Divila. 1991. Amplification and deletion of a nod-nif region in the symbiotic plasmid of Rhizobium phaseoli. J. Bacteriol. 173:2435-2441. 41. Rosenberg, C., P. Boistard, J. Denari4, and F. Casse-Delbart. 1981. Genes controlling early and late functions in symbiosis are located on a megaplasmid in Rhizobium meliloti. Mol. Gen. Genet. 184:326-333. 42. Rosenberg, C., and T. Hughet. 1984. The pATC58 plasmid of Agrobacterium tumefaciens is not essential for tumor induction. Mol. Gen. Genet. 196:533-536. 43. Triplett, E. W. 1990. The molecular genetics of nodulation competitiveness in Rhizobium and Bradyrhizobium. Mol. PlantMicrobe Interact. 3:199-206. 44. Watson, R. J., Y.-K. Chan, R. Wheatcroft, A.-F. Yang, and S. Han. 1988. Rhizobium meliloti genes required for C4-dicarboxylate transport and symbiotic nitrogen fixation are located on a megaplasmid. J. Bacteriol. 170:927-934. 45. Whatley, M. H., J. S. Bodwin, B. B. Lippincott, and J. A. Lippincott. 1976. Role for Agrobactenium cell envelope lipopolysaccharide in infection site attachment. Infect. Immun. 13:1080-1083.
