Construction and characterization of a Rhizobium leguminosarum biovar viciae strain designed to assess horizontal gene transfer in the environment

ELSEVIER

FEMS Microbiology

Letters 128 (1995) 255-263

Construction and characterization of a Rhizobium leguminosarum biovar viciae strain designed to assess horizontal gene transfer in the environment Werner Selbitschka a9*, Doris Jording a, Stefan Nieman a, Rainer Schmidt a, Alfred Piihler a, Tom Mendum b, Penny Hirsch b aLehrstuhlfiir Genetik UniuersitiitBielefeld,Postfach 100131,D-33501Bielefel~ Germany b Rothamsted Experimental Station, Harpenden, Herts. AL5 ZiQ, UK Received 11 January 1995; revised 27 February 1995; accepted5 March 1995

Abstract An integration vector was developed which inserts cloned DNA in a non-essential site of the Rhizobium leguminosarum biovar uiciae chromosome. The expression of integrated genes is under the control of the constitutive neomycin phosphotransferase II (nptI1) promotor of transposon Tn5. The design of the vector ensures that loss of vector sequences can be detected, enabling selection of progeny containing only the requisite DNA. The newly constructed vector was employed to insert the Escherichia coli gusA gene conferring GUS activity into R. leguminosarum bv. uiciae strain LRS39401 which is cured of its symbiotic plasmid (pSym). One GUS-positive transconjugant, strain CTO370, was shown to have lost all vector sequences. Conjugal transfer of pSym2004 (a TnS-tagged derivative of symbiotic plasmid pRLlJI, which specifies pea nodulation and symbiotic nitrogen fixation) to CTO370, restored the GUS-positive strain’s symbiotic proficiency. Strain CT0370 is presently being used iu a field release experiment in order to assess the extent of pSym transfer in a natural R. leguminosarum bv. uiciae population under environmental conditions. Keywords: Integration vector; Rhizobium leguminosarum biovar Genetically engineered microorganisms; Deliberate release

1. Introduction Soil bacteria of the species R. leguminosarum bv. uiciae induce the formation of N,-fixing nodules on the roots of plants belonging to the genera Pisum,

* Corresponding author. Tel.: f49 (521) 106 2990, Fax: +49 (521) 106 5626; e-mail: [email protected] 0378-1097/95/$09.50 6 1995 Federation SSDIO378-1097(95)00102-6

of European

Microbiological

uiciae; P-Glucuronidase;

Horizontal gene transfer;

Vi&z, Lathyrus and Lens. The genetic determinants for nodulation (nod genes) and symbiotic nitrogen fixation (nif and fi genes) are located on large indigenous symbiotic plasmids (pSym) which in some cases may also encode genetic functions for their conjugative transfer (for a recent review see 111). Analyses of the genetic structure of natural R. Zeguminosarum bv. viciue populations strongly support the idea that transfer of pSym occurs in the Societies. All rights reserved

W. Selbitschka et al. / FEMS Microbiology

256

environment, since the same symbiotic plasmids were found in strains exhibiting a different genetic background. Conversely, rhizobial strains displaying the same genetic background harbored different symbiotic plasmids (e.g. [2,3]). The forthcoming use of genetically engineered microorganisms (GEMS) in environmental biotechnology (e.g. Rhizobium as biofertilizer with improved symbiotic performance) has raised concern about the ecological consequences of the potential spread of recombinant DNA from GEMS to the resident microbiota and vice versa (for a review see [4]). Therefore, it is of crucial importance to assess the extent to which rhizobia exchange genetic material in nature. Several studies have addressed the issue of plasmid transfer between rhizobial strains as well as between strains of Rhizobium and E. coli under conditions which mimic the natural situation (e.g. [51 and references therein). However, due to the much more complex conditions which exist in the open environment, the relevance of these laboratory-based findings is largely unknown. Consequently, field studies are needed to assess the extent to which pSym transfer occurs in nature. In this paper we report the construction of a R. leguminosarum strain

Letters 128 (1995) 255-263

which is tagged with the Escherichia coli gusA gene conferring GUS activity. The newly constructed strain which lacks its symbiotic plasmid is to be used in a deliberate release experiment to monitor pSym transfer in a natural R. leguminosarum bv. viciae population under environmental conditions.

2. Materials and methods 2.1. Bacterial strains growth conditions

and plasmids,

Escherichia

s17-1

Rhizobium leguminosarum LRS39401 c-l-0370 RSM2004 Plasmids pws35 pSM18 pSM25 pSM26

Relevant characteristics

Reference

coli

HBlOl

F-, h&S (rg mn J, leu, pro, thi, lacY, galK ara, xyl, rpsL, supE, recAl3 F-, rec.4, hsdR, carrying a modified RP4 (ApS, Tc’, Km’) integrated in the chromosome biovar vi&e pSym-cured derivative of VF39, Sm’ Derivative of LRS39401, Sm’, Sp’, GUS’ Derivative of field isolate 248, carries pSym2004 (pRLlJI::TnS), Sm’, Rif’ Derivative of pSUP102, Gm’ R. leguminosarum bv. viciae vector, Gm’ Derivative of pSM18, carries Derivative of pSM25, carries mediating sucrose sensitivity,

and

The bacterial strains and plasmids used in this study are listed in Table 1. E. coli strains were grown at 37” C on Penassay medium (17.5 g 1-l 1. R. leguminosarum bv. viciae strains were grown at 28” C either on the complete media TY or YM [6] or the minimal media Y [6] or VMM [7]. For solid media 15 g of agar per litre of medium was added. The final concentrations of antibiotics per litre medium were: 10 mg gentamicin and 100 mg ampicillin for E. coli, and 500-800 mg streptomycin, 200 mg spectinomycin and 20 mg gentamicin for Rhizobium. In experiments involving the enumeration of bacterial cells in non-sterile soil, the antifungal agents

Table 1 Bacterial strains and plasmids Bacterial strains or olasmids

media

1241

181 [251 This study 191 [I51

integration the gusA gene, Gm’ the sad@ gene Gm’

This study This study This study

W. Selbitschka et al. / FEMS Microbiology Letters 128 (1995) 255-263

benomyl and cycloheximide were added to selection plates at a concentration of 7.5 mg 1-l and 100 mg l-l, respectively. For the selection of sucrose-resistant Rhizobium clones VMM plates were used supplemented with sucrose to a final concentration of 10% (w/v). X-Glut (5-bromo-4-chloro-3-indolyl/?-o-glucuronic cyclohexylammonium salt) was added to solid media to give a final concentration of 50 mg 1-l. 2.2. Bacterial matings Transfer of plasmids between E. coli and Rhizobium to achieve marker exchange was performed according to Simon et al. [8]. Crosses between R. leguminosarum bv. viciae strains were done as described by Hirsch and Spokes [9] 2.3. Plant test and enumeration of bacterial strains in non-sterile soil Rhizobium strains were tested for their ability to form N-fixing nodules on axenic Vicia hirsuta seedlings (John Chambers, Kettering, UK) growing on N-free agar slopes, essentially as described in [lo]. For experiments to assess the symbiotic competence of rhizobia in non-sterile soil, surface-sterilized pea seeds (Pisum sativum cv. Avola) pre-germinated on water agar were transferred to 13-cm pots containing 800 g sieved field soil mixed with 400 g sterile sand. Three seeds were placed in each pot, and each was inoculated with 200 ~1 H,O containing 108-10’ rhizobia (from freshly grown culture in YM spun down, washed and resuspended in sterile H,O). The pots were kept at 18” C with 16 h light, 15” C for the 8 h of darkness. After 30 days, nodules were sampled and screened for GUS activity. Re-isolation and enumeration of bacterial strains in nonsterile soil was essentially done as described 191.

2.4. Test of GUS activity of pea root nodules

257

vided the basis for developing our assay system. Since it proved difficult to recover viable rhizobia after crushing nodules on TY agar following incubation with X-Glut, a MUG assay system was developed to enable screening of large numbers of nodules from field experiments. Batches of nodules, initially surface-sterilized as described [lo] and resterilized immediately before the assay in 70% ethanol for 30 s, followed by a sterile H,O rinse, were placed in the wells of a 96-well microtitre dish with 10 mM MUG, 10 mM sodium phosphate buffer (pH 6.5) and incubated overnight at 28” C. In reconstruction experiments, wells containing only one GUS+ nodule amongst nine normal GUS nodules gave clear and distinct fluorescence when the dish was placed on a 300 nm UV transilluminator. Each nodule from a positive batch could then either be tested singly in MUG or crushed onto non-selective TY agar so that rhizobial growth could be tested for GUS activity and antibiotic resistances. An altemative small-scale screening method involved cutting of surface-sterilized nodules in half, incubating one half of each in 0.2 mM X-Glut, 5 mM sodium phosphate buffer (pH 6.51, 0.1% Triton X-100, 10 mM EDTA overnight at 28” C, and crushing the other half on TY agar. 2.5. Plasmid profiles and DNA methods In order to determine the plasmid profiles of R. bv. viciae strains, the method described previously by Hirsch and Skinner [lo] was used. Plasmid DNA was extracted from E. coli cells following the method of Arnold and Piihler [12]. Restriction endonucleases and other commercially available enzymes were employed according to the manufacturers instructions. Preparation of competent E. coli cells and transformation with plasmid DNA was done as described by Maniatis et al. [13]. leguminosarum

2.6. Oligonucleotide-directed

mutagenesis

and DNA

sequence analysts

Initially, nodules were screened for GUS activity either by incubating the entire root system or excised nodules with X-Glut or with 4-methyl-umbelliferyl j?-D-glucuronide (MUG). The stock solutions were prepared according to [ll] which, together with advice from K. Wilson (personal communication), pro-

Site-directed mutagenesis was performed according to the oligonucleotide-directed in vitro mutagenesis system version 2.1 described by Amersham International, Amersham, UK. The XhoI site in front of the putative R. leguminosarum bv. viciae recA

258

W. Selbitschka et ai. / FEMS Microbiology

Letters 128 (I99.7) 25.5-263

transcription terminator was introduced using the oligonucleotide 5’-ggaatgcggcggctCgaGctgtccgatgagtttagg-3’. The changed bases are given in capital letters. The nucleotide sequence was determined according to the dideoxy method of Sanger et al. [14] using fluorescent primers and the Automated Laser Fluorescent DNA Sequencer (A.L.F., Pharmacia). 2.7. Construction of the R. leguminosarum integration vector pSMl8

bu. viciae

I..,I...,...,...,...~ 1 200

400

600

800

1006

-Q-

Plasmid pWS35 [15] was used as the basic replicon for the construction of the R. leguminosarum bv. viciae-specific integration vector pSM18. As a first step a l.O-kb AluI fragment carrying the C-terminal coding region of the R. leguminosarum bv. viciae recA gene as well as the N-terminal coding region of the downstream located aluS gene was cloned in DraI/PvuII-digested pWS35 resulting in plasmid pSM13. Subsequently, a 220-bp PstI/SulI fragment carrying the recA terminator region was replaced by a mutated fragment in which an XhoI site had been introduced by site-directed mutagenesis (Fig. 1). The resultant plasmid was designated pSM15. As a next step the pnptlI-luc expression cassette described recently [16] was cloned as a Sal1 fragment in XhoIdigested pSM15, yielding plasmid pSM16. The R. leguminosarum bv. viciae integration vector pSM18 (Fig. 2) was obtained from pSM16 by removing the firefly luc gene as a BamHI fragment and subsequent re-ligation.

3. Results 3.1. Construction of R. leguminosarum bv. viciue integration vector pSMl8 and its GUS activity mediating derivative pSM26 DNA sequence analyses of the recA genes of R. leguminosurum bv. viciae and R. meliloti had shown that putative transcription terminators were situated downstream of the rhizobial recA coding regions 1171. Recent analyses of derivatives of the R. meliloti strain 2011 which carry a pnpt ,-luc or a pnpru-gusA expression cassette in front o I the recA transcription terminator, respectively, had revealed that the integration of foreign DNA at the chosen site per se did

alas’

IBd’

Fig. 1. Chromosomal AluI fragment of Rhizobium leguminoswum bv. viciae VF39 into which an XhoI site was introduced by site-directed mutagenesis. The C-terminal coding region of the recA gene (recA * ) as well as the N-terminal coding region of the downstream located alaS gene (aluS * ) are indicated by shaded arrows. Omega indicates the putative Rho-independent transcription terminator (0) situated downstream of the recA coding region. The nucleotide sequence of the putative transcription terminator is shown. Palindromic DNA sequences are indicated by bold arrows. The bases which were introduced by site-directed mutagenesis to create an XhoI site (boxed sequence) are shown above the original bases.

not change the recombinant strains’ properties [161. Therefore, by analogy the sequence immediately in front of the putative recA transcription terminator of R. leguminosurum bv. viciue VF39 was chosen as integration site (see Fig. 1). Since there was no suitable restriction site available, an XhoI site was introduced by site-directed mutagenesis. For this purpose a 1-kb AluI fragment carrying the C-terminal region of the R. leguminosarum bv. viciae recA gene as well as the N-terminal region of the downstream located aluS gene was cloned into phage M13mp19 [18]. Using the oligonucleotide shown in Materials and methods, the XhoI site was introduced by site-directed mutagenesis (Fig. 1). The successful introduction of the XhoI site was verified by DNA sequence analysis (data not shown). Following the cloning strategy used, the R. leguminosurum bv. viciae-specific vector pSM18 was constructed (Fig. 2A). The GUS activity mediating derivative pSM25 of integration vector pSM18 was obtained by cloning a promoterless gusA BumHI fragment [15] into BumHI-digested pSM18. Finally, a SERB contain-

W. Selbitschka et al. / FEMS Microbiology Letters 128 (I 995) 255-263

ing X&I fragment [17] was cloned into the unique &I site of pSM25 resulting in plasmid pSM26 (Fig. 2B). 3.2. Construction of R. leguminosarum bv. viciue strain 00370 which carries the E. coli gusA gene integrated into the chromosome Using E. coli strain S17-1, plasmid pSM26 conferring GUS activity and mediating sucrose sensitivSal1 1

ii

SamHl

Smal ECORI Smd

pSM18 73OObps

\\

mob

/

259

ity was mobilized into R. leguminosarum bv. viciae strain LRS39401. The vector mediated gentamicin resistance served as the selection marker. Gentamicin-resistant transconjugants arose at a frequency of approximately lop4 per donor cell. In order to obtain double recombinants which had integrated the reporter gene by a double crossover event, rhizobial Gm’ clones previously grown non-selectively were plated onto VMM plates containing 10% of sucrose. After selection for sucrose resistance, approximately 5% of the clones tested exhibited a Gm” phenotype. Twenty clones out of several hundred Gm” clones tested displayed a GUS-positive phenotype. Southern hybridization experiments using 11 of these clones revealed that 10 clones had undergone chromosomal rearrangements and that only one clone showed the expected hybridization patterns (data not shown). Since control experiments using strain LRS39401 had indicated that the loss of the Sym plasmid in strain LRS39401 was associated with an increased sensitivity to sucrose the chromosomal rearrangements observed in most of the clones might be a consequence of the strong selection pressure applied. From the clone which had integrated the marker gene at the correct chromosomal site, a spontaneous spectinomycin-resistant mutant was isolated and used for further work. This strain was designated CTO370.

Fig. 2. Restriction maps of the R. leguminosarum bv. viciae integration vector pSM18 (A) and its GUS activity conferring derivative pSM26 (B). The C-terminal coding region of the R. leguminosarutn bv. viciue rec4 gene (recA ) as well as the N-terminal coding region of the downstream located alo.Y gene (al& 1 and the aacC1 gene mediating gentamicin resistance are indicated as shaded and solid arrows, respectively. The non-coding region between the recA and alas sbuctural genes is disrupted by the insertion of a 0.3&b npttl promoter carrying DNA fragment (open arrow). The transcription direction from the nptII promoter is in the same orientation as the rec4 gene transcription direction. The nptIl promoter is flankd downstream and upstream by 446 bp and 560 bp of R. legwninosanun bv. vi&e DNA, respectively. Immediately downstream of the rptIl promoter there are the unique restriction sites .SmoI, EcoRI and BornHI for cloning. Mob (open box) indicates the cloned mobil

l

pSM26 11800 bps

lization functions of plasmid Rp4. Plasmid pSM26 carries the promoterless gusA gene (solid arrow) cloned as a BamHI fragment into pSM18. In addition, pSM26 carries the SUCRB gene (solid box) inserted as an _&II fragment into the unique X&I site of pSM18.

W. Selbitschkn et al. / FEMS Microbiology

260

3.3. Characterization

of prototrophy

tivity of R. leguminosarum

and UV sensi-

3.4. Detection of pSym2004 and CT0370

1 2

3

4

5

6

7

8

9101112

bv. viciae strain CT0370

The integrity of the aZaS gene in CT0370 was demonstrated by normal growth of the strain on minimal Y agar. Tests to check the recA status were conducted by plating liquid TY cultures on TY agar and exposing to 254 nm UV light for different time periods to give a range of total doses from 5 mJ to 16 mJ. The number of colonies appearing after 4 days at 28” C were found to be comparable for LRS39401 and CTO370, indicating an unchanged recA phenotype. Sensitivity to UV light (i.e. reduced colony counts) was apparent on plates receiving doses of 8 mJ and higher, viable counts decreasing with increasing UV doses.

LRS39401

Letters 128 (1995) 255-263

harboring

derivatives of

in root nodules

Fig. 4. Screening method for the detection of GUS-positive nodules amongst nodules induced by GUS-negative strains. Al of the microtitre plate shows the signal that is produced single GUS-positive nodule amongst four GUS-negative root ules using the MUG assay.

root Well by a nod-

and soil

samples

bv. viciae strain LRS39401 and its GUS-positive, spectinomycin-resistant derivative CT0370 were crossed with Rhizobium

1

2

leguminosarum

3

4

5

6

Fig. 3. Plasmid profiles of R. Ieguminosarum bv. viciae strains VF39 (11, CT0370 (21, CT0370 (pSym2004; 3), RSM2004 (4), LRS39401 (pSym2004; 5) and LRS39401 (6).

RSM2004 on filter membranes. Neomycin-resistant transconjugants which had received pSym2004 from RSM2004 arose at a frequency of 10-4-10-5 per parent. The presence of pSym2004 in transconjugants was verified by plasmid profile analysis (Fig. 3). In plant tests, the pSym-lacking strains LRS39401 and CT0370 did not induce any nodules on V. hirsuta, whereas the symbiotically competent strain RSM2004 and the transconjugants LRS39401 (pSym2004) and CT0370 (pSym2004) formed pink nodules which signify N, fixation. These nodules were screened for GUS activity. None was detected from plants inoculated with either RSM2004 or LRS39401 (pSym2004), whereas all nodules from plants inoculated with the transconjugant CT0370 (pSym2004) were GUS+. As shown in Fig. 4, one single GUS+ nodule pooled with four GUS nodules resulted in an unambiguous signal. Hence, this method can be used to screen approximately 500 nodules simultaneously in a single microtitre plate, and therefore it is suitable for screening large numbers of nodules in field studies. Enumeration of indigenous bacteria in soil samples taken from a potential release site of Rothamsted on TY selection plates showed that the total number of cells displaying resistance against both streptomycin and spectinomycin was approximately

W. Selbitschka et al. /FEMS Microbiology Letters 128 (1995) 255-263

2 X lo3 g-’ soil with only 10 of these also being GUS+. Thus, the gusA gene is a suitable marker gene in conjunction with the antibiotic resistances of strain CT0370 to track this strain in the environment. 3.5. Test of the competitive nodulation (pSym2004)

abilities for host plant of strains RSM2004, LSR39401 and CT0370 (pSym2004)

Further assessment of symbiotic competence and competitiveness were made using field soil from the prospective release plot. From six pea plants inoculated with CT0370 (pSym2004), a total of 40 nodules were examined and 12 (30%) were found to contain GUS+ rhizobia, all of which were also resistant to streptomycin and spectinomycin and were thus identical to the inoculant. No nodules with GUS activity were found in peas grown in soil inoculated with RSM2004 or LRS39401 (pSym2004) where checking antibiotic resistance profiles of nodules indicated that the inoculants formed 20% of nodules. Similarly, no GUS+ nodules were detected in soil inoculated with the pSym- strains LRS394Olsp’, CT0370 or uninoculated soil, where all the nodules are presumably formed by the indigenous rhizobial population.

4. Discussion In this paper we report the construction of the R. bv. viciae-specific integration vector pSM18. Using its GUS activity conferring derivative pSM26, the constitutively expressed gusA gene was inserted into the chromosome of pSym- strain LRS39401. The integration of genes using pSM26 and similar plasmids has the advantage that loss of unwanted vector sequences, including the antibiotic resistances necessary for obtaining the initial transconjugants, can be selected so that only the required genes are integrated in a specific site. Plasmid pSM16 which mediates bioluminescence and is also a derivative of pSM18 was successfully employed to integrate the firefly luc gene into the chromosome of strain R. leguminosarum bv. viciae VF39 (our unpublished results). Thus, integration vector pSM18 should be generally applicable for the integration of any trait into the chromosome of R. leguminosarum

261

bv. viciae. In contrast to the random genomic insertion of genetic information by transpositional events mediated by most vector systems, the main advantage of the vector described here is the phenotypic predictability of genetically altered bacterial strains (for a discussion of this topic see also reference [ 151). There are three reports in the literature which have investigated pSym transfer within R. leguminosarum bv. viciae populations in field studies. R. leguminosarum bv. viciae strains which harbor Tn5 tagged self-transmissible symbiotic plasmids were released in 1987 at a site at Rothamsted Experimental Station, UK [9], Bayreuth University, Germany [19] as well as in 1988 at an experimental plot at INRA, Dijon in France [20]. In none of the studies mentioned, pSym transfer to the resident population or to simultaneously released recipient strains was observed under field conditions. However, in contrast to the approach described here, the detection of pSym transfer was rather time-consuming in these studies because the identity of strains re-isolated from root nodules had to be established by selective plating and plasmid profile analyses. In order to detect pSym transfer using the approach described in this paper, donor bacteria harboring conjugative symbiotic plasmids must be present at the release site. Although several studies have demonstrated that natural isolates of R. leguminosarum bv. viciae harbor self-transmissible symbiotic plasmids (e.g. [21]) little is known about the total number of potential donors of pSym present at a given site. The number of donor bacteria present, however, is one parameter which determines the frequency of pSym transfer (e.g. [5]). Therefore, to increase the probability of pSym transfer detection, the experimental plot at Rothamsted Experimental Station where strain RSM2004 (pSym2004) had been previously released was chosen. Continuous monitoring of the released strain had revealed that RSM2004 established as a subpopulation at a density of lo2 to lo3 per g of soil [9]. Hence, strain RSM2004 could act as a donor which provides CT0370 a symbiotic plasmid. The approach to use a pSym- strain as a recipient to investigate the transfer of symbiotic plasmids in a natural Rhizobium population seems to reflect the natural situation of rhizobial life in the environment.

leguminosarum

262

W. Selbitschka et al. / FEMS Microbiology

Several studies indicate the occurrence of non-symbiotic rhizobia in soil. Such isolates gain symbiotic proficiency upon acquisition of the corresponding symbiotic plasmid as shown by in vitro conjugation experiments 122,231. In summary, the use of the well-characterized experimental plot at Rothamsted Experimental Station as a release site of strain CTO370, together with the gene transfer detection system described here should greatly facilitate the analysis of pSym transfer under environmental conditions. According to consent 94\R8\ 2 of the UK Department of the Environment strain CT0370 has been released at a 9-m2 plot at Rothamsted Experimental Station, UK. Presently, experiments are under progress to assess the extent of pSym transfer in an R. leguminosarum bv. uiciae population under natural conditions.

Acknowledgements Parts of this work were supported by grants from the Bundesministerium fi.ir Forschung und Technologie (0319206B) and the EU (BI02 CT92-0370).

References [l] Prakash, R.K. and Atherly, A.G. (1986) Plasmids of Rhizobium and their role in symbiotic nitrogen fixation. int. Rev. Cytol. 104, l-24. [2] Schofield, P.R., Gibson, A.H., Dudman, W.F. and Watson, J.M. (1987) Evidence for genetic exchange and recombination of Rhizobium symbiotic plasmids in a soil population. Appl. Environ. Microbial. 53, 2942-2947. [3] Young, J.P.W. and Wexler, M. (1988) Sym plasmid and chromosomal genotypes are correlated in field populations of Rhizobium Ieguminosurum. J. Gen. Microbial. 134, 27312739. [4] Tiedje, J.M., Colwell, R.K., Grossman, Y.L., Hodson, R.E., Lenski, R.E., Mack, R.N. and Regal, P.J. (1989) The planned introduction of genetically engineered organisms: ecological considerations and recommendations. Ecology 70, 298-315. [5] Kinkle, B.K. and Schmidt, E.L. (1991) Transfer of the pea symbiotic plasmid pJB5JI in nonsterile soil. Appl. Environ. Microbial. 57, 3264-3269. [61 Beringer, J.E. (1974) R factor transfer in Rhiwbium leguminosarum. J. Gen. Microbial. 84, 188-198. [7] Vincent, J.M. (1970) A manual for the practical study of

Letters 128 (I 995) 255-263

root-nodule bacteria (IBP Handbook 15). Blackwell Scientific Publications, Oxford. 181Simon, R., Priefer, U. and Piihler, A. (1983) A broad host range mobilization system for in vivo genetic engineering: transposon mutagenesis in Gram-negative bacteria. Bio/Technology 1, 784-791. 191 Hirsch, P.R. and Spokes, J.D. (1994) Survival and dispersion of genetically modified rhizobia in the field and genetic interactions with native strains. FEMS Microbial. Ecol. 15, 147-160. [lOI Hirsch, P.R. and Skinner, F.A. (1992) The identification and classification of Rhizobium and Bradyrhizobium. In: Identification Methods in Applied and Environmental Microbioiogy (Board, R.G., Jones, K. and Skinner, F.A., Eds.), pp. 45-65. Blackwell Scientific Publications, Oxford. [Ill Gallagher, S.R. (1992) GUS Protocols: Using the GUS Gene as a Reporter of Gene Expression. Academic Press, London. [121 Arnold, W. and Piihler, A. (1988) A family of high-copynumber plasmid vectors with single end-label sites for rapid nucleotide sequencing. Gene 70, 171-179. 1131 Maniatis, T., Fritsch, E.F. and Sambrook, J. (1982) Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY. [141 Sanger, F., Nicklen, S. and Coulson, A.R. (1977) DNA sequencing with chain-terminating inhibitors. Proc. Natl. Acad. Sci. USA 74, 5463-5467. [151 Selbitschka, W., Piihler, A. and Simon, R. (1992) The construction of red-deficient Rhizobium meliloti and R. leguminosarum strains marked with gusA or Zuc cassettes for use in risk-assessment studies. Mol. Ecol. 1, 9-19. 1161Selbitschka, W., Dresing, U., Hagen, M., Niemann, S. and Pihler, A. (1995) A biological containment system for Rhizobium meliloti based on the use of recombination-deficient (red) strains. FEMS Microbial. Ecol. 16, 223-232. [17] Selbitschka, W., Arnold, W., Priefer, U.B., Rottschlfer, T., Schmidt, M., Simon, R. and Plhler, A. (1991) Characterization of recA genes and rec4 mutants of Rhizobium meliloti and Rhizobium leguminosarum biovar viciae. Mol. Gen. Genet. 229, 86-95. [I81 Norrander, J., Kempe, T. and Messing, J. (1983) Construction of improved Ml3 vectors using oligodeoxynucleotide-directed mutagenesis. Gene 26, 101-106. [19] Dohler, K. and Klingmiiller, W. (1988) Genetic interaction of Rhizobium leguminosarum biovar uiciae with Gram-negative bacteria. In: Risk Assessment for Deliberate Releases (Klingmiiller, W., Ed.), pp. 18-28. Springer Verlag, Berlin. [20] Amarger, N. and Delgutte, D. (1990) Monitoring genetically manipulated Rhizobium leguminosarum bv. viciae released in the field. In: The Biosafety Results of Field Tests of Genetically Modified Plants and Microorganisms (MacKenzie, D.R. and Henry, S.C., Eds), pp. 221-228. Agricultural Research Institute Bethesda, Bethesda, MD. I211Geniaux, E. and Amarger, N. (1993) Diversity and stability of plasmid transfer in isolates from a single field population of Rhizobium leguminosarum bv. uiciae. FEMS Microbial. Ecol. 102, 251-260.

W. Selbitschka et al. /FEMS

Microbiology Letters 128 (1995) 255-263

[22] Segovia, L., Pinero, D., Palacios, R. and Martinez-Romero,

E. (1991) Genetic structure of a soil population of nonsymbiotic Rhizobium leguminosarum. Appl. Environ. Microbial. 57,426-433. [23] Laguerre, G., Bardin, M. and Amarger, N. (1993) Isolation from soil of symbiotic and nonsymbiotic Rhizobium Zeguminosarum by DNA hybridization. Can. J. Microbial. 39, 1142-1149.

263

[24] Boyer, H.W. and Roulland-Dussoix, D. (1969) A complementation analysis of the restriction and modification of DNA in Escherichia coli. J. Mol. Biol. 41, 459-472. [25] Hynes, M.F. and McGregor, N.F. (1990) Two plasmids other than the nodulation plasmid are necessary for formation of nitrogen-fixing nodules by Rhizobium Zeguminosurum. Mol. Microbial. 4, 567-574.